ek
e

© 0 I O Uk WM

Part [

AQUEOUS HOMOGENEOUS
REACTORS

 

JAMES A. LANE, Editor
Oak Ridge N ational Laboratory

. Homogeneous Reactors and Their Development
. Nuclear Characteristics of One- and Two-Region Homogeneous Reactors
. Properties of Aqueous Fuel Solutions

Technology of Aqueous Suspensions

. Integrity of Metals in Homogeneous Reactor Media
. Chemical Processing
. Design and Construction of Experimental Homogeneous Reactors

Component Development

. Large-Scale Homogeneous Reactor Studies

Homogeneous Reactor Cost Studies

AUTHORS
E. G. BOHLMANN H. F. McDvurrFIE
P. R. KASTEN R. A. McNEEs
J. A. LANE C. L. SEGASER
J. P. McBripE I. SPIEWAK
D. G. Taomas
CONTRIBUTORS

B.M

S. E

W. E. BRowNING
W. D. BurcH

R. D. CHEVERTON
E. L. CoMPERE
C. H. GABBARD
J. C. GRIESsS

D. B. HaLL

E. C.
G. H.
J. C.

m@

S. I. KarLaN
N. A. KronN
C. G. LawsoN
R. E. LEuzE
R. N. LyonN
W. T. McDvurrEE
L. E. MoRSE
S. PETERSON
R. C. ROBERTSON
H. C. SAVAGE
D. S.

TooMmB
PREFACE

This compilation of information related to aqueous homogeneous
reactors summarizes the results of more than ten years of research and
development by Oak Ridge National Laboratory and other organizations.
- Some 1500 technical man-years of effort have been devoted to this work,
the cost of which totals more than $50 million. A summary of a program
of this magnitude must necessarily be devoted primarily to the main
technical approaches pursued, with less attention to alternate approaches.
For more complete coverage, the reader is directed to the selected bib-
liography at the end of Part I.

Although research in other countries has contributed to the technology
of aqueous homogeneous reactors, this review is limited to work in the
United States. In a few instances, however, data and references pertaining
to work carried on outside the United States are included for continuity.

Responsibility for the preparation of Part I was shared by the members
of the Oak Ridge National Laboratory as given on the preceding page and
at the beginning of each chapter.

Review of the manuscript by others of the Oak Ridge Laboratory statf
and by scientists and engineers of Argonne National Laboratory and
Westinghouse Electric Corporation have improved clarity and accuracy.
- Suggestions by R. B. Briggs, director of the Homogeneous Reactor Project
at the Oak Ridge Laboratory, and S. McLain, consultant to the Argonne
Laboratory, were particularly helpful.

Others at Oak Ridge who assisted in the preparation of this part include
W. D. Reel, who checked all chapters for style and consistency, W. C.
Colwell, who was in charge of the execution of the drawings, and H. B.
Whetsel, who prepared the subject index.

Oak Ridge, Tennessee James A. Lane, Editor
June 1958
CHAPTER 1
HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT*

1-1. BACKGROUNDT

1-1.1 Work prior to the Manhattan Project. Nuclear reactors fueled
with a solution or homogeneous mixture of fuel and moderator were among
the first nuclear systems to be investigated experimentally following the
discovery of uranium fission. In fact, it was only slightly more than a
year after this discovery that Halban and Kowarski at the Cavendish
Laboratory in England performed experiments which indicated to them
that a successful self-sustaining chain reaction could be achieved with a
slurry of uranium oxide (U3Osg) in heavy water.

In these experiments, reported in December 1940 [1], 112 liters of heavy
water mixed with varying amounts of U3Os powder were used inside an
aluminum sphere 60 cm in diameter, which was immersed in about one ton
of heavy mineral oil to serve as a reflector. (Mineral oil was chosen to
avoid contamination of the D20 in case of a leak in the sphere.) By meas-
uring neutron fluxes at varying distances from a neutron source located in
the center of the sphere, Halban and Kowarski calculated a multiplication
factor of 1.18 4- 0.07 for this system when the ratio of deuterium atoms to
uranium atoms was 380 to 1, and 1.09 4 0.03 when the D/U ratio was
160 to 1.

Other experiments conducted at the same time by Halban and Kowar-
ski [1]1, using U3Os and paraffin wax, indicated that with a heterogeneous
lattice arrangement it would be possible to achieve multiplication factors
as high as 1.37 in a system containing about 100 atoms of deuterium per
atom of uranium.

It is interesting to note that the D20 supply used in the experiments
had been evacuated from France. The D20 originally came from the lab-
oratories of the Norwegian Hydroelectric Company, and with the destruc-
tion of this plant and its D20 stockpile in 1942, this was the sole remaining
supply of purified D20. However, it was not enough to allow a self-
sustalning chain reaction to be established with natural uranium.

*By J. A. Lane, Oak Ridge National Laboratory.

TThis section is based on material supplied by W. E. Thompson, Oak Ridge
National Laboratory.

ISee the list of references at the end of the chapter.

1
2 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT [cHAP. 1

Even earlier (in 1939) Halban and Kowarski, as well as other experi-
mentalists, had fairly well established that self-sustaining chain reactions
with U3Og and ordinary water are not possible [2,3,4]. Homogeneous sys-
tems of uranium with carbon, helium, beryllium, or oxygen were also con-
sidered, and were rejected as not feasible either for nuclear, chemical, or
engineering reasons.

In November 1942, Kowarski, with Fenning and Seligman, reported
more refined experiments which led to the conclusion that neither homo-
geneous nor heterogeneous mixtures of U3Os with ordinary water would
lead to self-sustaining chain reactions, the highest values of the multiplica-
tion factor being 0.79 for the homogeneous system and 0.85 for the hetero-
geneous system.

Because it was clear even by early 1942 that the only feasible homo-
geneous reactor using natural uranium would be one moderated with D20,
and because no D20 was available at that time for use in reactors, interest
in homogeneous reactor systems was purely academic. The atomic energy
program, which was then getting well under way, devoted its attention to
heterogeneous reactors. By using a heterogeneous lattice arrangement
with a core of uranium metal slugs spaced inside graphite blocks and a
periphery containing UsOs slugs (used after the supply of uranium metal
ran out) spaced inside the graphite, the first successful self-sustaining chain
reaction was achieved on December 2, 1942.

1-1.2 Early homogeneous reactor development programs at Columbia
and Chicago universities. Interest in homogeneous reactors lagged until
early in 1943, when it became clear that American and Canadian efforts to
produce large quantities of heavy water would be successful. At that time
the group under H. C. Urey at Columbia University directed its attention
to the development of slurried reactors utilizing uranium oxide and D20.

In March 1943, Urey and Fermi held a conference to review the situa-
tion with respect to homogeneous reactors. They noted the value of 1.18
that Halban and Kowarski had obtained for the multiplication factor in a
U305-D20 slurry reactor and pointed out that the value calculated from
theory was only 1.02. They realized, however, that neither the theory nor
the experiment was free from serious objections, and that insufficient data
were available to allow a trustworthy conclusion to be reached as to the
feasibility of homogeneous systems.

If the results of Halban and Kowarski were correct, then a homogeneous
system containing a few tons of heavy water would be chain reacting. On
the other hand, if the theoretical estimates were correct, the order of
100 tons of D20 would be required.

Urey and Fermi recommended [5] that the earlier U30s-D20O experi-
ments be repeated with the improved techniques then known, and that
1-1] BACKGROUND 3

consideration be given to incorporating a mixture of uranium and heavy
water into the pile at Chicago to determine its effect on the pile reactivity.

From the theoretical considerations of E. P. Wigner and others, it ap-
peared that the most favorable arrangement for a Uz0Os-D20 reactor
would be one in which the slurry was pumped through a lattice of tubes
immersed in D20 moderator. This was especially true because the neutron
absorption cross section assigned to heavy water at that time made it ap-
pear that more than 200 tons of D2O would be required to reach criticality
in an entirely homogeneous system in which the UsOg and moderator were
mixed. With a heterogeneous system it seemed likely that a much smaller
quantity of D20 would suffice and every effort was directed toward pre-
paring a design that would require about 50 tons of D20 [6].

It was estimated by E. P. Wigner that the uranium concentration in the
slurry would have to be 2.5 to 3 grams per cubic centimeter of slurry. It
became apparent immediately that no aqueous solution of a uranium com-
pound could be made with such a density. With pure UF§, 2.48 grams of
uranium per cubic centimeter could be obtained, and piles utilizing this
compound were considered. However, the corrosion problems in such a
system were believed to be so severe that the development of a reactor to
operate at a high power level would be extremely difficult, if not impossible.

Other compounds, such as uranyl nitrate dissolved in D20, were ex-
cluded because in the case of nitrate the neutron absorption of nitrogen
was too high and in other cases sufficient densities could not be obtained.
Thus the initial phase of the research at Columbia was directed toward the
development of high-density slurries [6].

The reactor visualized by the Columbia group was one in which an ex-
tremely dense suspension of uranium in D20 would be pumped through a
large number of pipes arranged inside a heavy-water moderator. It was
planned that both the slurry and the moderator would be circulated
~ through heat exchangers for cooling [6].

Then, in July of 1943, the experiments of Langsdorf [7] were completed,
giving a much lower cross section for deuterium than was known earlier.
As a result, the homogeneous reactor became much more attractive, since
the critical size (neglecting external holdup) could then be reduced to about
30 tons of D20 with about 6 tons of uranium as oxide in an unreflected
sphere [8]. This favorable development allowed emphasis to be shifted to
less dense slurries, greatly simplifying the problems of maintaining a sus-
pension of dense slurry, pumping it, and protecting against erosion. Ex-
periments were directed toward developing a reactor design which would
permit operation without continuous processing of the slurry to maintain
its density [6]. |

By the end of 1943 preliminary designs had been developed at the
University of Chicago Metallurgical Laboratory for several types of heavy-
4 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT [cHAP. 1

water reactors, all using slurry fuel but differing in that one was com-
pletely homogeneous [9], one was a light-water-cooled heterogeneous ar-
rangement [10], and another was a D20-cooled heterogeneous reactor [11].
These reactors were proposed for operation at power levels of 500 Mw or
more (depending on external power-removal systems) and were intended
as alternates to the Hanford piles for plutonium production in case satis-
factory operation of the graphite-natural uranium, water-cooled piles
could not be achieved.

At this point one might ask why it was that homogeneous solution
reactors were not given more serious consideration, especially in view of
the newly discovered cross section for deuterium, which permitted con-
siderably lower concentrations of uranium. The answer is that the only
known soluble salts of uranium which had a sufficiently low cross section to
enable the design of a reactor of feasible size and D20 requirement were
uranyl fluoride and uranium hexafluoride. (Enriched uranium was not
then available.) These were considered, but rejected principally because
of corrosion and instability under radiation. A second factor was the evi-
dence that D20 decomposition would be more severe in a solution reactor
where fission fragments would be formed in intimate contact with the
D20 rather than inside a solid particle as in the case of a slurry.

Research on homogeneous reactors was undertaken at Columbia Uni-
versity in May 1943, and continued with diminishing emphasis until the
end of 1943, at which time most of the members of the homogeneous re-
actor group were transferred to Chicago, where they continued their work
under the Metallurgical Laboratory.

At the Metallurgical Laboratory, the principal motivation of interest in
homogeneous reactors was to develop alternate plutonium production
facilities to be used in the event that the Hanford reactors did not operate
successfully on a suitable large scale, and studies were continued through
1944. With the successful operation of the Hanford reactors, however,
interest in homogeneous plutonium producers diminished, and by the end
of 1944 very nearly all developmental research had been discontinued. The
results of this work are summarized in a book by Kirschenbaum [12].

1-1.3 The first homogeneous reactors and the Los Alamos program.
During the summer of 1943 a group at Los Alamos, under the leadership
of D. W. Kerst, designed a “power-boiler’’ homogeneous reactor, having
as 1ts fuel a uranyl sulfate-water solution utilizing the enriched uranium
which was expected to become available from the electromagnetic process.
However, this design was put aside in favor of a low-power homogeneous
reactor designed by R. F. Christy. The low-power homogeneous reactor
was built and used during the spring and summer of 1944 for the first of a
series of integral experiments with enriched material (see Chapter 7).
1-1] BACKGROUND 5

There were two reasons for choosing U0O2S04 instead of uranyl nitrate
as the fuel: there is less neutron absorption in the sulfate than in the ni-
trate, and the sulfate was thought to be more soluble. The latter reason
was considered important because it was feared that with the maximum-
enrichment material from the electromagnetic process, it might be difficult
to dissolve the critical mass in the desired volume [13]. These objections
to the use of uranyl nitrate, however, were subsequently found to be
invalid.

After gaining experience in operating the low-power reactor, ‘LOPO,”
the Los Alamos group revised its plans for the higher power homoge-
neous reactor, known as the “HYPO,” and after extensive modification
of the design, the reactor was built and put into operation in December 1944
with uranyl nitrate as the fuel.

In April 1949, rather extensive alterations to the HYPO were begun in
order to make the reactor a more useful and safer experimental tool. The
modified reactor, known as “SUPOQO,” is still in operation. The present
SUPO model reached local boiling during initial tests, due to the high
power density. A slight increase in power density above the design level
produces local boiling between cooling coils, even though the average so-
lution temperature does not exceed 85°C.

Interest in solution reactors continued at Los Alamos, and improved
designs of the Water Boiler (SUPO Model II) were proposed [14]. These,
however, have not yet been constructed at Los Alamos, although similar
designs have been built for various universities [15].

The work on water boilers at Los Alamos led to the design of power
reactor versions as possible package power reactors for remote locations.
Construction of these reactors, known as Los Alamos Power Reactor Ex-
periments No. 1 and No. 2 (LAPRE-1 and LAPRE-2), started in early
1955. To achieve high-temperature operation at relatively low pressures,
LAPRE-1 and —2 were fueled with solutions of enriched uranium oxide in
concentrated phosphoric acid. The first experiment reached criticality in
March 1956 and was operated at 20 kw for about 5 hr. At that time
radioactivity was noted in the steam system, and the reactor was shut
down and dismantled. It was discovered that the gold plating on the
stainless steel cooling coils had been damaged during assembly and the
phosphoric acid fuel solution had corroded through the stainless steel.
The cooling coils were replaced and operations were resumed in October
1956. However, similar corrosion difficulties were encountered, and 1t was
decided to discontinue operations. In the meantime, work on LAPRE-2
continued, and construction of the reactor and its facilities was completed
during the early part of 1958. The details of these reactors are given in
Chapter 7.
6 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT [CHAP. 1

1-1.4 Early homogeneous reactor development at Clinton Laboratories
(now Oak Ridge National Laboratory). With the availability of enriched
uranium in 1944, the possibility of constructing a homogeneous reactor
became more attractive because, by using enriched uranium, the DO
requirement could be greatly reduced, or even ordinary water could be
used. The chemists at Clinton Laboratories (now ORNL), notably C. D.
Coryell, A. Turkevich, S. G. English, and H. S. Brown, became interested
in enrichéd-uranium homogeneous reactors primarily as a facility for pro-
ducing other radioisotopes in larger amounts, and a number of reports on
the subject were issued by various members of the Chemistry Division
(D. E. Koshland, Jr., W. J. Knox, and L. B. Werner).

In August 1944 Coryell and Turkevich prepared a memorandum [16]
recommending the construction of a 50-kw homogeneous reactor containing
5 kg of uranium enriched to 1249, U235 or about 500 g of plutonium. The
fuel proposed was to be in the form of salt solution in ordinary water. The
- following valuable uses of such a reactor were listed in this memorandum
and enlarged upon in a later memorandum by Coryell and Brown [17]:

(1) The preparation of large quantities of radioactive tracers.

(2) The preparation of intense radioactive sources.

(3) Studies in the preparation and extraction of U233,

(4) The preparation of active material for Hanford process research.

(5) Study of chemical radiation effects at high power levels.

(6) Accumulation of data on the operating characteristics, chemical
stability, and general feasibility of homogeneous reactors.

The physicists were also interested in the homogeneous reactor, partic-
ularly as a research facility which would provide a high neutron flux for
various experimental uses. The desirability of studying, or demonstrating,
if possible, the process of breeding had been made especially attractive
by the recent data indicating that U232 emitted more neutrons for each
one absorbed than either U23% or Pu?39 and the physicists were quick to
point out the possibility of establishing a U233-thorium breeding cycle
which would create more U233 from the thorium than was consumed in the
reactor. These potentialities were very convincingly presented in No-
vember 1944 by L. W. Nordheim in a report entitled ‘“The Case for an
Enriched Pile” (ORNL-CF—44-11-236).

The power output of such a breeder with a three-year doubling time is
about 10,000 kw, and this was established as a new goal for the homoge-
neous reactor. The reactor, then, was conceived to be a prototype homo-
geneous reactor and thermal breeder; in addition, it was conceived as an
all-purpose experimental tool with a neutron flux higher than any other
reactor.

Work on the 10,000-kw homogeneous reactor was pursued vigorously
through 1945; however, at the end of that year there were still several
1-1] BACKGROUND 7

basic problems which had not been solved. Perhaps the most serious of
these was the formation of bubbles in the homogeneous solution. These
bubbles appear as a result of the decomposition of water into hydrogen and
oxygen by fission fragments and other energetic particles. Because the
bubbles cause fluctuations in the density of the fuel solution, they make it
difficult to control the operating level of the reactor. Nuclear physics
calculations made at the time indicated that under certain conditions it
might be possible to set up a power oscillation which, instead of being
damped, would get larger with each cycle until the reactor went completely
out of control. Minimizing the bubble problem by operating at elevated
temperature and pressure was not considered seriously for two reasons:
first, beryllium, aluminum, and lead were the only possible tank materials
then known to have sufficiently low neutron-absorption characteristics to
be useful in a breeder reactor. Of these metals, only lead was acceptable
because of corrosion, and lead is not strong enough to sustain elevated
temperatures and high pressures. Second, there had been essentially no
previous experience in handling highly radioactive materials under pres-
sure, and consequently the idea of constructing a completely new type of
reactor to operate under high pressure was not considered attractive.

Other major unsolved problems at the end of 1945 were those of corro-
sion, solution stability, and large external holdup of fissionable material.
Because it appeared that the solution of these problems would require
extensive research and development at higher neutron fluxes than were
then available, it was decided to return to the earlier idea of a hetero-
geneous reactor proposed by E. P. Wigner and his associates at the Metal-
lurgical Laboratory. Experimental investigations in this reactor, it was
hoped, would yield data which would enable the homogeneous reactor
problems to be solved. The extensive effort on this latter reactor (later
built as the Materials Testing Reactor in Idaho) forced a temporary
cessation of design and development activities related to homogeneous
breeder reactors, although basic research on aqueous uranium systems
continued.

1-1.5 The homogeneous reactor program at the Oak Ridge National
Laboratory. Early in 1949, A. M. Weinberg, Research Director of Oak
Ridge National Laboratory, proposed that the over-all situation with
respect to homogeneous reactors be reviewed and their feasibility be
re-evaluated in the light of knowledge and experience gained since the
end of 1945. Dr. Weinberg informally suggested to a few chemists, physi-
cists, and engineers that they reconsider the prospects for homogeneous
reactors and hold a series of meetings to discuss their findings.

At the meeting held by this group during the month of March 1949, it
was agreed that the outicok for homogeneous reactors was considerably
8 | HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [CHAP. 1

brighter than in 1945 and that effort directed toward the design of a small
experimental reactor should be resumed. By July 1949, interest in homo-
geneous reactors had increased further as a result of the preliminary studies
which had been started, and it was decided to establish a small develop-
ment effort on homogeneous reactors. A Homogeneous Reactor Com-
mittee, under the direction of C. E. Winters, was formed and reactor
physics and design studies were undertaken on a somewhat expanded
scale. By the latter part of August 1949, a preliminary design of the major
components had been developed.

Construction of the reactor (Homogeneous Reactor Experiment No. 1)
was started in September 1950, and completed in January 1952. After a
period of nonnuclear testing with a natural-uranium fuel solution, HRE-1
reached criticality on April 15, 1952. Early in 1954 it was dismantled
after successfully demonstrating the nuclear and chemical stability of a
moderately high-power-density circulating-fuel reactor, fueled with a
solution of enriched uranyl sulfate.

During the period of construction and operation of HRE-1, conceptual
design studies were completed for a boiling reactor experiment (BRE)
operating at 150 kw of heat and a 58-Mw (heat) intermediate-scale homo-
geneous reactor (ISHR). Further work on these reactors was deferred
late in 1953, however, when it became evident from HRE-1 and the asso-
ciated development program that construction of a second homogeneous
reactor experiment would be a more suitable course of action.

The main reason for this decision was that HRE—-1 did not demonstrate
all the engineering features of a homogeneous reactor required for con-
tinuous operation of a nuclear power plant. Thus a second experimental
reactor (Homogeneous Reactor Test, HRE-2), also fueled with uranyl
sulfate, was constructed on the HRE-1 site to test the reliability of ma-
terials and equipment for long-term continuous operation of a homo-
geneous reactor, remote-maintenance procedures, and methods for the
continuous removal of fission products and insoluble corrosion products.
Construction of the reactor was completed late in 1956 and was followed
by a period of nonnuclear operation to determine the engineering charac-
teristics of the reactor. This testing program was interrupted for six to
nine months by the need for replacing flanges and leak-detection tubing
in which small cracks had developed, owing to stress corrosion induced
by chloride contamination of the tubing. The reactor was brought to
criticality on December 27, 1957, and reached full-power operation at
5 Mw on April 4, 1958. Shortly thereafter, a crack in the core tank de-
veloped which permitted fuel solution to leak into the D20 blanket.
After consideration of the nuclear behavior of the reactor with fuel in both -
the core and blanket, operation was resumed under these conditions in
May 1958.
1-1] BACKGROUND | 9

TABLE 1-1

- LevELs oF EFrrorT ON HOMOGENEOUS
ReAacTror DeEVvELOPMENT AT ORNL

 

 

 

: Millions Man-years
Fiscal year of dollars (technical)
1949 0.15 5
1950 0.54 15
1951 2.2 75
1952 4.1 127
1953 3.4 119
1954 3.9 133
1955 7.7 219
1956 9.1 238
1957 10.0 316
1958 11.5 333

 

 

 

 

 

The ten-year growth of the ORNL effort on homogeneous reactors is
indicated by Table 1-1, which summarizes the costs and man-years de-
voted to the program through fiscal year 1958.

Following the completion of construction and beginning of operation of
HRE-2, the ORNL Homogeneous Reactor Project directed its attention
to the design of a 60-Mw (heat) experimental aqueous thorium breeder
reactor, designated as HRE-3, with the objective of completing the con-
ceptual design during the summer of 1958. Work on slurry development
and component development was accelerated to provide the information
necessary for the start of construction of HRE-3 at the earliest possible
date.

1-1.6 Industrial participation in homogeneous reactor development. In-
dustrial participation in the homogeneous reactor program started with a
number of studies to evaluate the economic potential of such reactors for
large-scale power production [18-22]. The opinion of some who compared
homogeneous breeder reactors with solid-fuel converters is reflected in the
following excerpts from Ref. 19: “The two reactor types that offer the
greatest possibilities for economic production of central station power are
the thermal U233 breeders of the circulating fuel type and fast plutonium
breeders containing fuel easily adaptable to a simple processing system . . .
The self-regulating features of fluid-fuel reactors and low fission-product
inventory due to continuous chemical processing give these reactors the
greatest possibility of safe and reliable operation . . . Both the pressurized
10 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [CHAP. 1

water and sodium-graphite systems suffer from the inability to consume
(in a single cycle) a large fraction of the uranium necessary to result in
low fuel costs that are attainable with breeder systems.”

During late 1954 and early 1955, Westinghouse and Pennsylvania
Power and Light Company, operating under Study Agreements with the
Atomic Energy Commission, made a joint study [21] aimed at determining
the economic feasibility of aqueous homogeneous-type reactor plants. The
study indicated that a two-region solution-slurry plant and a single-region
slurry plant appeared to have excellent long-range possibilities for pro-
ducing competitive electric power. The study also indicated, however,
that considerable development work would be required before the tech-
nical feasibility of either type of plant could be determined with any
degree of certainty. The results of this and other continuing studies led
the two companies to set up the Pennsylvania Advanced Reactor Project
in August 1955. An initial proposal to build a 150-Mw (electric) power
station financed with private funds was made to the A.E.C. by the Pennsyl-
vania Advanced Reactor group at that time. This proposal was later modi-
fied and resubmitted as part of the power demonstration reactor program.

In spite of the formidable development program which appeared to be
associated with the construction of a full-scale homogeneous reactor
power plant, a second industrial group proposed building a homogeneous
reactor as part of the power demonstration program in cooperation with
the government. This proposal (made in response to a request by the
Atomic Energy Commission for small-scale reactors) by the Foster Wheeler
and Worthington Corporations in January 1956, considered construction
of an aqueous homogeneous burner reactor. Plans were for a reactor and
associated oil-fired superheater with a net electrical capacity of 10,000 kw
for the Wolverine Electric Cooperative, Hersey, Michigan. Although this
proposal was accepted in principle by the Atomic Energy Commission in
April 1956, and money was appropriated by Congress for carrying out the
project, in May 1958 the Atomic Energy Commission announced that
plans had been canceled due to increases in the estimated cost of the plant
(from $5.5 million to between $10.7 and $14.4 million).

The second proposal submitted to the Atomic Energy Commission
jointly by the Pennsylvania Power and Light Company and Westinghouse
Electric Corporation was determined by the Commission on February 26,
1958, as acceptable as a basis for negotiation of a contract but was later
“recalled, following a review by the Joint Congressional Committee on
Atomic Energy. The proposal called for the construction of a reactor of
the homogeneous type with a net electrical output of 70,000 to 150,000 kw
to be operated on the Pennsylvania Power and Light Company system.
The reactor would use a thorium-uranium fuel as a slurry in heavy water.
Under the proposal, the Atomic Energy Commission would assume the
1-2] GENERAL CHARACTERISTICS: HOMOGENEOUS REACTORS 11

cost of research and development planned for 1958 and 1959, at which
time a decision would be made either to begin actual construction of a
plant or terminate the project. The cost of the project, scheduled for com-
pletion by December 1963, was estimated at $108 million. The Westing-
house and Pennsylvania Power and Light Company’s share of the cost
included $5.5 million for research since 1955, $57 million for plant con-
struction, and $16 million for excess operating costs during the first five
years of operation. The Atomic Energy Commission was asked to provide
the additional $29 million, including $7 million for research and develop-
ment in 1958-1959, $18 million for research and development following a
decision to construct the plant, and $4 million for fuel charges during the
first five years of operation.

1-2. GENERAL CHARACTERISTICS OF HOMOGENEOUS REACTORS

1-2.1 Types of systems and their applications. Because of the large
number of possible combinations of mechanical systems and compounds
of uranium and thorium which may be dissolved or dispersed in H20 or
D20, there exists in principle an entire spectrum of aqueous homogeneous
reactors. These may be classified according to (a) the type of fissionable
material burned and produced (U235 burners, converters, breeders), (b) the
geometry or disposition of the fuel and fertile material (one-region, two-
region), or (c) the method of heat removal (boiling, circulating fuel, and
fluidized suspension reactors). The possible materials which can be used
in these various reactor types are given in Table 1-2; all combinations are
not compatible.

TABLE 1-2

HoMoGENEOUS REACTOR MATERIALS

 

 

 

 

 

 

 

- Fertile Moderator Corrvosion-re§istant
uel : metals of primary
material coolant )
1nterest
U02804 + HoS04 U238 galt D20 Austenitic stainless
steels
UO2F2 + HF U238 oxide H20 Zircaloy-2
UO2N1503 + HNO3 ThO- Titanium
U0 2SO4 + LizSO4 P latinum
UOs3 + alkali oxide 4+ COq - Gold
UOs+ H3POy4, UO2+ H3PO4
UO3 + H2CrOy4
UOg, UO3, U305

 

 
HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT [cHAP. 1

12

 

UOTISIOAUOD
ezl 0} geg[) 10 JUIPOdIq gez()

(3exuelq) Q% ur
20y, snid (a100) O ur QUL

9100 A1In[s

 

 

 

pue uorjonpoid tomod ofgos-o3rer| | snjd 9PIXO gez() I0 gez() POYOLIUT 0001002 ‘I9p9aIq wWNLIOY) UOLISI-OM]T,
UOISIOAUOD (3oyue[q) \
ggz(l O} geg(l 10 IUIPIOIQ geg() 0% W 2QU, snid (e100) O3 9I00 uOIIN{OS
pug uonjonpoad temod opeos-081eT | *OSP0O[ S© gez[l 10 ¢ez() POYOLIUY 0001—-00Z ‘IopeaIq WNLIOY} UOLIAI-0M]T,
O W 2QYL
uonjonpouid 1omod oeos-081%] | SN[d 9PIXO gez() I0 gez() POYOLIUY 00S1-008S I9PadIQ WNLIOY) UOIFAI-S[SUIY
uoyonpold umru | (39UB[q) OFQ W QSO PereIdeQ
-ojnid snid 1emod osodind-fen(y (8102) O%( uI ¥O]E() peyouuy 00ST1-00S I9onpoad nJ uoi3aI-oMJ,
uorjonpold wnru [(*08)%rT peppe ynoy3im 10 yjim]
-ognd snid temod oesodind-fen( | Q3 ul *QOSEQ) poYyOLIUS A[IY3IS 0002—000T Isonpoad nJg uorsaI-su()
uorjonpoad romod 9[eos-931e] O%( Ut £Q[) poyouue AIYS3I[g 0001-00S SI9)19AU0) 19M0d UOIZAI-9U()
proe orroydsoyd
O/M G UL PRA[OSSIP 2(Q[) PIYILIUY
sjue[d 1omod a[Bos-9)BIpPOW pros ouroydsoyd
-I9JUI PUB -[[8WS Pa)BOO[ A[9j0WRY | O/M (9 Ul PIA[OSSIP Q] poyouuy 001-1 81079801 Jomod ad£) HYJV'I
sjuerd remod 9J1q v
-OW {SUOI}BD0[ }S00-[aNJ-yJIy ur
sjue[d 1omod o[e9s-0318] 0} ~[BWS | O I0 OFH Ul Y]] paysLIuy 00S-0% SIOWIN] gez()
SOXNY UOIJNIU-[BULIY) Ju
UYSIY-8Ij[n 9B YOIBISAI JIBI[ONN 0% ur YOS payoLIuy 000Z2—008 S10908BaI YOIBISAI SNOUSTOWOF
Bururer) O%H u N - .
pue yorxessax Jesponu AysieArun)| (EQN)?Q[ I0 YOSEQN payouuy G0'0-0 Io[Ioq 1938 M
uonyeoriddy 59 AN

 

uorsuadsns 10 uoInos [on g

 

‘@8uea [9A9] 19MOJ

 

UOI}BUTISIP 103083y

 

-1 E1aV],

SNOILVOI'TddYy ANV SHdXA ], 90LOVEY SNOUNIADONOH

 
1-2] GENERAL CHARACTERISTICS: HOMOGENEOUS REACTORS 13

The terms used in classifying homogeneous reactors may be defined as
follows: Burner reactors are those in which fissionable fuel is econsumed
but virtually no new fuel is generated. To this class belong the water
boilers, homogeneous research reactors, U235 burners, and LAPRE-type
reactors. Converter reactors produce a different fissionable fuel than is
destroyed in the fission process, such as in the dual-purpose plutonium
producers or single-region converters, while breeder reactors produce the
same fissionable fuel as that which is consumed. One-region reactors con-
tain a homogeneous mixture of fissionable and fertile materials in a moder-
ator. Generally, these have large reactor diameters, in order to minimize
neutron losses, and contain fuel plus fertile material in concentrations of
100 to 300 g of uranium or thorium per liter of solution or slurry. Two-
region reactors are characterized by a core containing fissionable materials
in the moderator surrounded by a blanket of fertile material In moderator.
These reactors may have comparatively small diameters with dilute core-
fuel concentrations (1 to 5 g of uranium per liter) and a blanket containing
500 to 2000 g of fertile material per liter. Boiling reactors are reactors in
which boiling takes place in the core and/or blanket and heat is removed
by separating the steam from the solution or suspension. Fluidized sus-
pension reactors are those in which solid particles of fuel and fertile ma-
terial are fluidized in the core and/or blanket, but are not circulated
through the cooling system external to the reactor pressure vessel.

A summary of homogeneous reactor types and the primary application
of each is given in Table 1-3.

1-2.2 Advantages and disadvantages of aqueous fuel systems. Aqueous
fuel systems possess certain advantages which make them particularly
attractive for numerous nuclear-reactor applications ranging from small
reactors (for mobile units or package-power plants) to large, high-power
reactors (for large-scale production of plutonium, U233, and/or power).
These advantages stem partly from the fluid nature of the fuel and partly
from the homogeneous mixture of the fuel and moderator ; 1.e., an aqueous
homogeneous reactor combines the attributes of liquid-fuel heterogeneous
reactors with those of water-moderated heterogeneous reactors. If practical
methods for handling a radioactive aqueous fuel system are developed, the
inherent simplicity of this type of reactor should result in considerable
economic gains in the production of nuclear power and fissionable material.

However, many apparently formidable practical problems are associated
with continued operation and maintenance of systems involving radio-
active fuel solutions. It is believed, therefore, that extensive experience in
a series of small- to large-scale reactor installations will be required to
demonstrate the reliability of aqueous homogeneous reactors; this will
necessitate a long-range development program. In addition, the choice of
14 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [CHAP. 1

water as the fuel-bearing medium limits both the fuel concentration and
operating temperature to values which may be less than optimum for pro-
duction of power and fissionable material.

The principal advantages of aquecus fuel systems are:

(1) High power density. Because of the homogeneous nature of the
reactor fuel-fluid, virtually no heat-transfer barrier exists between the fuel
and coolant. Thus reactor power densities of 50 to 200 kw/liter may be
possible, being limited by considerations other than heat transfer, such as
radiation-induced corrosion and chemical reactions.

(2) High burnup of fuel. In heterogeneous reactors, burnup is limited
by radiation damage to fuel elements or loss of reactivity. In liquid-fuel
reactors, continual removal of poisons is possible, as well as continual
additions of new fuel, thereby permitting unlimited burnup.

(3) Continuous plutonium recovery. Continuous removal of neptunium
or plutonium is possible in a liquid-fuel reactor. This yields a product with
a low Pu240 content and increases the value of the plutonium [23].

(4) Simple fuel preparation and reprocessing. The use of aqueous fuel
solutions or slurries eliminates the expensive fuel-element fabrication step
and simplifies the reprocessing of depleted fuel.

(5) Continuous addition or removal of fuel. Charging and discharging fuel
can be accomplished without shutting down the reactor and without the
use of solid-fuel charging machines. |

(6) High neutron economy. Neutron economy is improved by eliminating
absorption of neutrons by cladding and structural material within the
reactor core. Also, there is the possibility of continuously removing
Xe135 and other fission-product poisons. In addition, an aqueous fuel
system lends itself readily to a spherical core geometry, which minimizes
neutron leakage.

(7) Simple control system. Density changes in the moderator create a
sensitive, negative temperature coefficient of reactivity which makes this
system self-stabilizing. This eliminates the need for mechanically driven
regulating rods. In addition, shim control can be achieved by changing the
fuel concentration.

(8) Wide range of core sizes. Depending on concentration and enrich-
ment, critical H2O and D20 homogeneous reactors range from 13 ft to as
large as is practicable. Correspondingly, there is a wide range of applica-
tion for these reactor systems.

The principal problems of aqueous fuel systems are:

(1) Corrosion or erosion of equipment. The acidity of fuel solutions and
abrasiveness of slurries at high flow rates creates corrosion and erosion
1-2] GENERAL CHARACTERISTICS: HOMOGENEOUS REACTORS 15

problems in the reactor and its associated equipment. Special provisions
must therefore be made for maintaining equipment.

(2) Radiation-induced corrosion. The presence of fission radiation in-
creases the rate of corrosion of exposed metal surfaces. This limits the per-
missible wall power density, which in turn restricts the average power
density within the reactor.

(3) External circulation of fuel solution. Removal of the heat from the
reactor core by circulating fuel solution, rather than coolant only, through
external heat exchangers increases the total amount of fuel in the system
and greatly complicates the problems of containment of radioactivity and
accountability of fissionable material. The release of delayed neutrons in
the fuel solution outside of the reactor core reduces the neutron economy
of the reactor and causes induced radioactivity in the external equipment,
resulting in the need for remote maintenance.

(4) Nuclear safety. The safety of homogeneous reactors is associated
with the negative density coefficient of reactivity in such systems ; how-
ever, by virtue of this coefficient, relatively large reactivity additions are
possible through heat-exchanger mishaps and abrupt changes in fuel cir-
culation rate. In boiling reactors changes in the volume of vapor within
the reactor core may lead to excessive reactivity changes.

(5) Limated uranium concentration. In solution reactors, uranium con-
centration is limited by solubility or corrosion effects, and in slurries, by
the effective viscosity and settling characteristics. In H.O-moderated
reactors, in particular, a high uranium or thorium concentration is neces-
sary for a high conversion ratio. Concentrations up to 1000 g/liter, how-
ever, may be considered for solutions and up to 4000 g/liter for fluidized
beds.

(6) Limated operating temperatures. At the present time the operating
temperatures of aqueous solution systems appear limited because of cor-
rosion problems at ~225°C and phase stability problems above 300°C.
Pressures encountered at higher temperatures are also a problem.

(7) Explosive decomposition product. Radiation-induced decomposition
of the moderator can produce an explosive mixture of hydrogen and oxygen
in the reactor system. This hazard means that special precautionary design
measures must be taken. To prevent excessive gas formation and reduce
the requirement for large recombiners, a recombination catalyst such as
cupric lon may be added. Disadvantages associated with this addition are
the neutron poisoning effects and changes in chemical equilibria which
occur.

A comparison of the advantages and disadvantages of specific homo-
geneous reactors is given in Table 1—4.
HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT [cHAP. 1

16

 

quawdinba jo s[iem uo jisodap
Lewl puB UOIIN[OS UL ABIS 10U SOOP WNIUOIN]J
WI)SAS wniue)ij-[[e aanbar LB
swepqold Surjpuey ALn[g
swarqoad umopinys pue dnjrelg
swsqoad Surjpuey ALn[g
HnOWIp
9q A®w 1030831 JO uUMmopInys pue dnjrelq

swajqoid Surpuey L[S

(8100 A1In[s Y}IM paredurod 9109 UoInNjos

)M SNOLISS arowr oq Avw) AJsusp romod
SHIWI] JUB) 9100 WNIUODIIZ JO UOISO.LIOD UOI)BIPBY

SI9)I9AUOD PUR SIFPIAI] SNoduadowroy 0} pared
-Wod (UOIPBISUAFAI OU UM WINIUBIN POYILIUD
jo Suruang o3 onp) §7509 [PN] Y3y APANEBPEY

swarqoad Jurjpury ALIn[s Jo UOI}BUTWI[H

swo[qoad WNIUOIIIZ JO UOTPBUTWII[H
A109U9AUT

[BLI9)BWI-I[I)I9] PUB -9[ISSY MO[ A[QAI}B[dY

[BLIS}BW
UOT}ONIJSUOD B §B WNTUOIIIZ JO UOTJBUTWI[H
£1S09 [9NJ MO[ PUB AWOU0Id UOIINIU YSITH]
uorjnjos
9109 UIOIJ [BAOWIAI 30NPOId-UOISSY 9[qISSOJ
AIOYJUSAUI [BLIF)EW-I[ISSY MO
§]S09 [9NJ MO[ PUB AWOU0Id UOIINAU YSITF]
(109eI9pOW ()3(J)
KIO)UQAUI [BLIS)BW-I[ISSY MO
(103BI9POW ()3F])
JuawaIINbar ()% (T JO UOTYBUTWI[H
querd Jut
-889001d [BOIWOYD JO UOIJBUIWI[® I[QISSOJ

uonn[os ¥)§()[) uorsar-ou()

Aum(s €[] uordai-ouQ

Aunys ()Y ], uoidai-au()

9109 AXIN[S JO UOI}N[OS
‘Iopod1q UOIFAI-0M ],

I0jBIdpOW (O)%(T 10 O%H
‘ToUINg ggz[] UOLIAI-OU()

 

 

 

§938JUBAPBSI(]

 

seSe)uBAPY

 

sod£3 103089y

 

SHdX ], 9OLOVAY SNOANTDOWOJ JO0 NOSIMVIINO()

F—1 @19V],

 
1-3] U235 BURNER REACTORS 17

1-3. U235 BuRNER REACTORS

1-3.1 Dilute solution systems and their applications. One-region re-
actors fueled with a dilute solution of highly enriched uranium or “‘burner
reactors” are ideal as a concentrated source of neutrons, since the critical
mass and size of the core of this type of reactor can be very small. Many
low-power research reactors are in operation which use this fuel system,
and very-high-flux research reactors of this type are being considered [24].
The principal advantages of solution reactors for this latter application are
the small amount of U235 required for criticality and the ability to add
fuel continually.

One-region burner reactors are applicable for both small- and large-
scale nuclear power plants. Such plants can operate for very long periods
of time (20 years or more) without necessity for removal of all the fission
products. Corrosion product buildup, however, must be limited to prevent
uranium precipitation. The fuel concentration would be dilute, increasing
with time of reactor operation if no fuel processing is carried out. Either
light or heavy water can be used as the moderator-coolant; the fuel con-
centrations would always be higher for the light-water-moderated reactors.
An advantage of these systems is that they utilize fuel in the concentration
range which has been studied most extensively. Experience in circulating
such solutions, however, indicates that careful control of operating condi-
tions and the concentrations of the various fuel constituents, such as
H2504, CuSOy4, NiSO4, H202, Og, etc., is necessary to avoid problems of
two-phase separation, uranium hydrolysis, and oxygen-depletion precipi-
tation of uranium.

For power production, homogeneous burner reactors can be considered
as possible competitors to the highly enriched solid-fuel reactors, such as
the Submarine Thermal Reactor and the Army Package Power Reactor.
By eliminating fuel-element fabrication, fuel costs in homogeneous burners
with either D20 or H3O as the coolant-moderator are in the range of
4 mills/kwh at present Atomic Energy Commission prices for enriched
uranium [25].

Possible fuel systems for the dilute, highly enriched burner-type reactors
are UO2SO4 in HzSO4, UOz(NO3)2 in HNO3, UO2oF2 in HF, and UOs3-
alkali metal oxide-CO2 in H20. These fuel systems are compared in
Chapter 3.

1-3.2 High-temperature systems. Fuel systems of enriched uranium
dissolved in highly concentrated phosphoric acid have been suggested for
homogeneous power reactors because of the high thermal stability and low
vapor pressure of such systems. This permits operation at higher tempera-
- tures than is possible with dilute acids, with accompanying higher thermal
efficiencies. Fuel systems of this type include UO3 in 30 to 60 w/o (weight
18 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [cHAP. 1

percent) phosphoric acid, UO2 in 90 to 100 w/o phosphoric acid, and UOs3
in concentrated chromic acid. The UO3-H3PO4 system, used in the Los
Alamos Power Reactor Experiment No. 1 (LAPRE-1), must be pressur-
ized with oxygen to prevent uranium reduction. Solutions containing
phosphate-to-uranium ratios of 4/1 to 10/1 are stable up to 450°C. How-
ever, the neutron economy is poor and these solutions are corrosive to all
metals except platinum and gold. The UO2-H3PO4 systems, pressurized
with hydrogen, have somewhat better corrosion characteristics and copper
may be used at least in regions which are kept below 250°C.

1-4. CONVERTER REACTORS

1-4.1 Purpose of converters. In converter reactors, U?® is burned to
produce U233 or Pu??® by absorption of excess neutrons in fertile material.
Thus the purpose of converter reactors is the production of power, fission-
able material, or both. Since homogeneous reactors have to operate at
temperatures above 225°C and pressures above 1000 psi because of prob-
lems of corrosion and gas production, homogeneous converters are thought
of as dual-purpose reactors for the production of power and fissionable
material or power-only reactors. Such reactors are also considered .mainly
in connection with the U235-U238-Pu239 fuel cycle, whereas the homo-
geneous breeder reactors are associated with the thorium fuel cycle.

1—4.2 One-region converters. One-region converter reactors may be
fueled with a relatively concentrated solution (100 to 300 g/liter D20) of
slightly enriched uranium for plutonium and power production or with a
suspension of slightly enriched uranium oxide for power production only.*
The principal advantage of the solution-type converter for plutonium pro-
duction is the insolubility of plutonium in the high-temperature uranium
sulfate system (see Chapter 6). This opens the possibility of separating the
plutonium by centrifugation rather than by a solvent extraction or ab-
sorption process. The costs of this method of recovering the plutonium,
which contains only small amounts of Pu24%, should be considerably less
than is possible with solid-fuel reactors and conventional processing tech-
niques. Indications are, however, that the plutonium formed in the fuel
solution is preferentially adsorbed on hot metal surfaces in contact with
the solution and is difficult to remove (see Chapter 6). Other problems with
the solution-type converter are the highly corrosive nature of concentrated
uranyl sulfate solutions and the lower temperature at which the two
liquid phases separate. An all-titanium high-pressure system may be

*Early work at Columbia and Chicago was aimed at a low-temperature version
of such a reactor for plutonium production only; however, present-day considera-
tions are limited to high-temperature systems.
1-5] BREEDER REACTORS 19

necessary to contain these solutions, which will lead to considerably higher
equipment and piping costs. The addition of lithium sulfate to the solution
would reduce corrosion and raise the phase-separation temperature so
that i1t might be possible to use stainless steel; however, the neutron
economy with normal lithium is poorer and separated Li? would be costly.

A single-region converter fueled with natural or slightly enriched uranium
oxide as a suspension avoids the problems of plutonium precipitation,
phase separation, and corrosion mentioned above. The advantage of such
a converter reactor for power produoction is the elimination of radiation
damage and fuel burnup problems encountered with solid-fuel elements;
however, the problem of radiation damage to the reactor pressure vessel
must be considered.

1-4.3 Two-region converters. Two-region homogeneous converters may
also be fueled with either D20 solutions or slurries; in these reactors, how-
ever, the U235 is in the core and the fertile material in the blanket. Con-
verters of this type become breeders if the bred fuel is subsequently burned
in the core and there is a net gain in the production of fuel. A two-region
converter with a dilute enriched-uranium core solution and a concentrated
depleted-uranium blanket solution shows promise of producing more eco-
nomical power and plutonium than the one-region converter reactors
mentioned previously. [26] because of the lower inventory charges and
the better neutron economy. Although the power density at the wall of
the titanium-lined pressure vessel is lower in the case of the two-region
machine, which minimizes the possibility of accelerated corrosion rates,
there is some evidence [27] that titanium corrosion will not be severe in
any case. The major materials problem in the dilute-solution core converter
will be that of zirconium corrosion, which may be above 30 mils/year at
power densities necessary for economic production of power and fission-
able material.

Two-region converters fueled with a uranium oxide slurry in the core
may be a possibility as an alternative to the solution-slurry system; how-
ever, not much is known about the corrosion resistance of zirconium in
contact with fissioning uranium oxide or about the engineering behavior
of such a slurry.

1-5. BREEDER REACTORS

1-5.1 The importance of breeding. If present projections [28] for the
growth of the nuclear power industry in the United States are correct, the
installed capacity of nuclear electric plants in 1980 may be as much as
227 million kilowatts and may be increasing by 37 million kilowatts an-
nually. Even assuming optimistic figures for fuel burned in then-existing
plants and fuel plus fertile material for inventories in new plants [29],
20 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [cHAP. 1

the annual requirement of fissionable material will be approximately
420,000 kg in 1980. This fissionable material will have to come from nat-
ural sources (i.e., uranium mined from the ground) or be produced from
neutrons absorbed in fertile material in a reactor (i.e., breeding or con-
version). Since presently known reserves of high-grade ores of uranium
and thorium in the United States [30] contain 148,000 tons of uranium
and 60,000 tons of thorium, respectively, and these in turn contain only
108 kg of fissionable material, it is obvious that conversion of a significant
fraction of the fertile material contained in the reserves will be necessary.

Although such a conversion will not reduce the inventory requirement
of fuel and fertile material for new plants starting up, this amounts to
only about 25%, of the burnup requirement. On this basis, the goal of
nuclear industry should be to develop reactor designs and associated fuel
systems which achieve a consumption of at least 5%, to 109, of the total
fertile material, as well as the initial fissionable material. At this point
the annual burnup requirement would become small compared with the
inventory requirement. This corresponds to a total burnup of about
50,000 Mwd/ton. Although such a fuel consumption might be obtained
in high-neutron-economy converter reactors through recycling of the fuel,
it seems likely that even the best such reactor may fall short of this goal
and that both fast and thermal breeders will be needed.

In the long term, therefore, the development of breeding systems is a
must. In the short term, where emphasis is on fuel costs rather than on
neutron economy and fertile-material utilization, converters rather than
breeders may predominate.

1-5.2 One-region thorium breeders. Since U233 does not occur in nature,
homogeneous thorium breeder reactors will probably start out as con-
verters, with U235 as the fuel and thorium as the fertile material. One-
region reactors of this type utilize a suspension of 100 to 300 g per liter of
thorium oxide plus enriched U235 as oxide and D20 as the moderator. In
order to maintain a breeding ratio greater than 1, fuel processing is neces-
sary to remove fission-product poisons. Also, to reduce losses due to neutron
leakage, the diameter of the reactor should be at least 12 ft.

1-5.3 Two-region breeder reactors. Two-region breeder reactors would
have thorium oxide suspensions (500 to 1500 g/liter) in the blanket re-
gion and could have either a highly enriched uranyl sulfate solution
(1 to 10 g/liter) or a thorium oxide-uranium oxide slurry (200 g ThO2/
liter and 10 g UOg3/liter) in the core region. Use of a solution-type core
permits the continuous removal of insoluble fission-product poisons by
means of hydroclones, while a slurry-type core leads to higher breeding
ratios. Because of these compensating factors, estimated fuel costs are
1-6] MISCELLANEOUS HOMOGENEOUS TYPES 21

- approximately the same in both types of reactors (see Chapter 10). While
the use of a suspension in the core may minimize the problem of radiation-
induced corrosion of the zirconium, not much is yet known about the
behavior of zirconium in a thorium-uranium slurry-fueled reactor. Cal-
culations summarized in Chapter 10 show that both solution- and slurry-
fueled two-region breeders have higher breeding ratios and lower fuel costs
than one-region breeders.

Numerous studies of large-scale two-region breeder reactors have been
carried out [20,26,31-42], some of which are described in detail in Chap-
ter 9.

1-6. MiscELLANEOUS HoMOGENEOUS TYPES

1-6.1 Boiling reactors. In May 1951, following completion of the con-
struction of HRE-1, a group at the Oak Ridge National Laboratory
focused its attention on the possibility of removing heat from a homo-
geneous reactor by boiling, rather than by circulating the fuel solution,
in recognition of the advantages of a boiling reactor. These are: (a) more
rapid response to sudden reactivity increases, minimizing power excur-
sions, (b) elimination of fuel circulating pumps, (c) increase in the tem-
perature of steam delivered to the turbine for a given reactor operating
pressure, and (d) reduction or elimination of problems of corrosion and
induced radioactivity associated with the circulation of fuel and fertile
material through an external heat-removal system. However, at that
time, questions of the nuclear stability of a boiling, liquid-fuel reactor
and the maximum specific power, in terms of kilowatts per liter, that could
be extracted from a given size core remained to be answered.

Experiments on bulk boiling at atmospheric pressures in a 1-ft-diameter
cylindrical tank indicated that power densities up to 5 kw/liter might be
achieved. It soon became apparent, however, that high-pressure power-
density measurements would be required, and the design of a boiling reactor
experiment (BRE) called the “Teapot” was initiated. To answer the
question of the nuclear stability of such a reactor, a combined group from
the Oak Ridge National Laboratory and Los Alamos operated the SUPO
under boiling conditions in October 1951. The reactor was operated at a
total power of 6 kw and solution power densities of 0.5 kw/liter were
obtained.

This removed one of the important obstacles to the construction of an
experimental boiling reactor, and in January 1952, the Oak Ridge National
Laboratory made a proposal to the Atomic Energy Commission to con-
struct the Boiling Reactor Experiment (BRE) to answer the question of
maximum specific power at higher pressures and to investigate the operat-
ing characteristics of boiling reactors. The proposed reactor was to operate
at a power level of 250 kw of heat and pressures up to 150 psi. The re-
22 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [CHAP. 1

actor was estimated to cost approximately $300,000, including the building
to house it. A one-year effort involving nuclear and engineering calcula-
tions, completion of the BRE conceptual design, and experiments on
bubble nucleation in the presence of radiation resulted from the proposal.
In January 1953, however, the problem of maintaining sufficient oxidizing
conditions to prevent reduction and precipitation of the uranium in a boil-
ing uranyl sulfate solution became apparent, and construction of the reactor
was deferred pending outcome of solution-stability experiments. These
experiments, completed in October 1953, indicated that at oxygen concen-
trations likely to be encountered in a boiling reactor (6 to 7 ppm), reduction
of the uranium would occur in solutions in contact with stainless steel.*
Since the metallurgy of titanium or zirconium was not sufficiently advanced
to construct a reactor using these alternate metals, it was decided to
abandon the BRE itself and continue experimental work on the problems
of solution stability, steam separation, and power densities at high pressure.
Work on steam separators and experimental measurements of the move-
ment of air and steam through heated solutions at high pressures were
carried out under contract by the Babcock & Wilcox Company [44].
These results and theoretical calculations of the power removal from boil-
ing reactors [45,46] provide a basis for estimating the obtainable power
density of such reactors under varying core heights, operating pressures,
and moderator density decreases. Values range from 10 to 40 kw/liter
with an average of 18.5 kw/liter for a 15-ft-high core, operating at 2000 psi,
and a density decrease due to steam of 0.4. Although the effect of such
a void fraction on nuclear stability is not known, if tolerable, boiling re-
actors may be able to achieve average power densities comparable to those
estimated for large-scale nonboiling circulating-fuel reactors operating
under a similar pressure [47]. In this latter case, the holdup of solution
in the external circulating system lowers the power density of the core
only, to an average of about 10 to 20 kw/liter. The two systems are
comparable, therefore, in terms of obtainable power densities, and boiling
reactors cannot be excluded on this basis.

The various applications of boiling as a method of heat removal from
homogeneous reactors include a one-region boiling solution or slurry re-
actor, a two-region reactor with a nonboiling core and a boiling blanket,
and a two-region reactor with a boiling core and a boiling blanket. The
problem of maintaining a sufficiently oxidizing solution in a boiling uranyl
sulfate—D20 reactor, although serious in a stainless-steel system, can be
eliminated if all surfaces in contact with the solution are made of titanium
and oxygen is supplied continuously. Solutions containing 10 g of uranium

*In more recent tests [43] with nonboiling solutions, in which oxygen concentra-
tions were held at 2 to 3 ppm, no reduction and precipitation of uranium occurred.
1-6] MISCELLANEOUS HOMOGENEOUS TYPES 23

per liter have been successfully boiled at 325°C in titanium-lined pipe [48].
Experiments have not yet been carried out with higher uranium concen-
trations in titanium. Continued interest in boiling homogeneous reactors
has led to a number of studies of large-scale reactors of this type
[33,36,37,49-51]. The actual construction of a boiling slurry reactor
1s under way in the USSR [52].

Use of boiling as a method for removal of power from the ThOs slurry
blanket of large two-region homogeneous reactors appears to present no
major difficulties [53]. The apparent advantages are that no circulating
pump would be required to handle slurry, containment of the highly active
slurry in the reactor vessel, and the possibility of operating the blanket
at the core pressure and using the blanket power for heating the secondary
steam [54]. One major problem is that of keeping the slurry suspended
during startup.

1-6.2 Gaseous homogeneous reactors. Although this book deals pri-
marily with aqueous systems, some mention should be made of other
types of fluid-fuel homogeneous reactors in which the fuel and moderator
- are mixed and can be circulated. The existence of UFg, which has a low
parasitic capture cross section and is a gas at ordinary temperatures,
makes possible the consideration of gaseous reactors. UFg boils at 56.4°C
at atmospheric pressure and has a critical temperature of 252°C at 720 psi.
Although UF is corrosive to most metals, it can be contained in nickel
and monel. However, the effect of radiation on the integrity of the pro-
tective film has not been studied. Considerable experience has been
ganed in the handling of UF¢ in metal containers at high temperatures
and pressures.

Pure UFg is not a practical possibility for a gaseous homogeneous re-
actor because fluorine is not a good enough moderator. Addition of helium
makes such a reactor possible, and calculations by D. E. Hull in a report,
“Possible Application of UF¢ in Piles” [55], show that the critical mass of
a graphite-reflected, He + UFs, reactor is 84 kg of U235, About 15 tons of
helium in a 60-ft-diameter core would be required. In a recent investiga-
tion [56] of reflector-moderated gaseous reactors (Plasma Fission Reactor),
the critical mass of gaseous U235 in a 2-ft-diameter cavity surrounded by
D20 was calculated to be less than 1 kg. Such reactors would have to
operate at extremely high temperatures (3000°K) which many feel are
beyond the realm of present technology.

Mixed UFs gas and dispersions of solid moderators such as beryllium or
graphite have been suggested, as well as beds of moderator particles
fluidized with circulating UF¢ [55]. However, these proposals have no
apparent advantages compared with gas- or liquid-cooled fluidized systems
described in the following section.
24 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [cHAP. 1

1-6.3 Fluidized systems. A variant of the fixed-bed or solid-moderator
homogeneous reactors consists of subdividing the fuel and/or moderator
to the point where the particles can be fluidized by the flowing gas or
liquid. Gas-cooled reactors of this type have received considerable atten-
tion because of the higher heat-transfer rates obtainable compared with
fixed-bed reactors. .A number of studies of gas-fluidized reactors have been
carried out by ORSORT groups at the Oak Ridge National Laboratory
[57,58] and by other groups [59,60].

In an Oak Ridge study [61] various types of fluidized-bed reactors were
compared. Systems investigated were (a) a sodium-cooled fast reactor,
(b) a gas-cooled system, (c) an organic-moderated reactor, (d). a heavy-
water-moderated reactor, and (e) a light-water system. A detailed study
of this latter system was carried out to compare its characteristics and
performance with solid-fuel heterogeneous pressurized-water reactors.
The results indicated that both the light-water-moderated and organic-
moderated fluidized reactors showed promise, while the gas-cooled, the
D>0-cooled, and the fast (unmoderated) reactors were found to be less
satisfactory for application of the fluidized-bed technique.

Systems using ThO2, fluidized by gas or D20, were described by the
Dutch at the Geneva Conference on Atomic Energy in 1955 [62,63].

Calculations show that a typical, one-region, 400 thermal Mw reactor
having a core diameter of 15 ft and a temperature rise of 50°C would re-
quire particles in the 40- to 60-micron range [64], whereas a two-region
reactor with a liquid-fluidized blanket would require particles in the 200-
to 600-micron range [65], and if the particles were confined to a 6-in.
annulus next to the core the particle size required would be in the 0.10-
to 1.5-cm range [65].

The disadvantages that may be observed with fluidized suspension
systems include the possibility of particle attrition [65], and instabilities
due to channeling during steady-state operation and due to settling if a
circulating pump failed.

Room-temperature attrition tests using 0.1-in.-diameter X 0.1-in.-long
ThO2 and ThO2 + 0.59, CaO cylinders (cylinders prepared by calcination
at temperatures of both 1650 and 1800°C) fluidized in water gave an
attrition rate of 12 to 159 weight loss per week [66]. However, circula-
tion tests using 10- to 20-micron ThOg2 spheres (calcined at 1600°C) in
toroids at superficial velocities up to 26 ft/sec and water temperatures of
250°C showed essentially no attrition for periods up to 200 hr [67].

This appears to indicate that attrition rate is at least a function of par-
ticle size, without giving any indication as to the effect of void fraction,
slip velocity, particle shape and density.
REFERENCES 25

REFERENCES

1. H. HauBaN and L. Kowarski, Cambridge University, March 1941.
Unpublished.

2. H. HAaLBAN et al., Number of Neutrons Liberated in the Nuclear Fission
of Uranium, Nature 43, 680 (1939).

3. H. L. ANDERSON et al., Neutron Production and Adsorption in Uranium,
Phys. Rev. 56, 248 (1939).

4. H. HaLBAN et al., Mise en Evidence d’une Reaction Nucléaire en Chaine
au Sein d’'une Masse Uranifére, J. phys. radium 10, 428 (1939).

5. BE. FerMt and H. C. Urey, Memorandum of Conference Between Prof. E.
Ferma and Prof. H. C. Urey on March 6, 7, and 8, 1 943 USAEC Report A-554,
Columbia University, Mar. 8, 1943.

6. C. F. Hiskey and M. L. EmiNorr, The Heavy-Water Homogeneous Pile:
A Review of Chemical Researches and Problems, USAEC Report CC-1383,
Argonne National Laboratory, Feb. 28, 1944.

7. A. LaNGsDORF, Slow Neutron Cross Section of Deuterium, USAEC Report
CP-902, Argonne National Laboratory, Aug. 30, 1943.

8. I. KaPLAN, Theory and Calculations of Homogeneous P-9 Piles, USAEC
Report A-1203, Columbia University, Sept. 10, 1943.

9. L. A. OHLINGER, Argonne National Laboratory, 1943. Unpublished.

10. L. A. OHLINGER, Design Features of Pile for Light-Water Cooled Power
Plant, USAEC Report CE-805, Argonne National Laboratory, July 16, 1943.

11. H. D. SmitH et al., Argonne National Laboratory, 1943. Unpublished.

12. I. KirsHENBAUM et al. (Eds.), Utilization of Heavy Water, USAEC
Report TID-5226, Columbia University, 1957.

13. C. P. BaxER et al., Water Boiler, USAEC Report AECD-3063, Los Alamos
Scientific Laboratory, Sept. 4, 1944.

14. L. D. P. King, Proceedings of the International Conference on the Peaceful
Uses of Atomic Energy, Vol. 3. New York: United Nations, 1956. (P/ 488)

15. J. W. Cuastrain (Ed.), U. S. Research Reactors, USAEC Report TID-
7013, Battelle Memorial Institute, August 1957.

16. C. D. CoryEeLL and A. TurkEvicH, Oak Ridge National Laboratory, 1944.
Unpublished.

17. C. D. CoryeLL and H. S. Brown, Oak Ridge National Laboratory, 1944,
Unpublished.

18. ComMONWEALTH EpisoN CoMPANY (NUCLEAR POWER Grour) AND PuBLIC
SERVICE CoMPANY OF NORTHERN ILLINOIS, 1952. Unpublished.

19. A Survey of Reactor Systems for Central Station Power Production, USAEC
Report NEA-5301(Del.), Foster Wheeler Corporation—Pioneer Service and
Engineering Company, October 1953.

20. H. G. Carson and L. H. Lanorum (Eds.), Preliminary Design and Cost
Estvmate for the Production of Central Station Power from a Homogeneous Reactor
Utilizing Thortum— Uranium-233, USAEC Report NPG-112, Commonwealth
Edison Company (Nuclear Power Group), February 1955.

21. WESTINGHOUSE ELEcTRIC CORPORATION and PENNSYLVANIA POWER AND
Ligar Company, 1955. Unpublished.
26 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [cHAP. 1

22. Single-flutd Two-region Agqueous Homogeneous Reactor Power Plant—
Conceptual Design and Feasibility Study, USAEC Report NPG-171, Common-
wealth Edison Company (Nuclear Power Group) and Babcock & Wilcox Com-
pany, July 1957.

23. AEC Plutonium Price Schedule, AEC Release, June 1957, Federal Register,
June 6, 1957.

24. P. R. KaASTEN et al., Aqueous Homogeneous Research Reactor—Feastbility
Study, USAEC Report ORNL-2256, Oak Ridge National Laboratory, April 1957.

295 AEC Price Schedule for Enriched Uranium, Nucleonics 14(12), (1956).

26. R. B. Brigas et al., Oak Ridge National Laboratory, 1954. Unpublished.

27. H. F. McDurriE, Corrosion by Aqueous Reactor Fuel Solutions, USAEC
Report CF-56-11-72, Oak Ridge National Laboratory, November 1956.

28 W. K. Davis and L. H. Roopis, AEC’s Fast Reactor Program, Nucleonics
15(4), 67 (1957).

29. J. A. Lang, Determining Nuclear Fuel Requirements for Large Scale
Industrial Power, Nucleonics 12(10), 65-67 (1954).

30. J. C. JounsoN, Uranium Production To Match Needs, Chem. Eng. News
35, No. 48, 70-75 (1957).

31. J. A. LaNE et al., Oak Ridge National Laboratory, 1950. Unpublished.

32. J. A. LANE et al., Oak Ridge National Laboratory, 1951. Unpublished.

33. L. C. WippoEs et al., Oak Ridge National Laboratory, 1951. Unpublished.

34. G. PurnaM et al., Reactor Design and Feastbility Problem; U-233 Power
Breeder, USAEC Report CF-51-8-213, Oak Ridge National Laboratory, 1951.

35. H. F. Kamack et al., Boiling Homogeneous Reactor for Producing Power
and Plutonium, USAEC Report CF-54-8-238(Del.), Oak Ridge National Lab-
oratory, 1954.

36. H. C. CrarBorNE and M. ToBias, Some Economic Aspects of Thortum
Breeder Reactors, USAEC Report ORNL-1810, Oak Ridge National Laboratory,
October 1955.

37. D. C. Hamirton and P. R. KasTEN, Some Economic and Nuclear Character-
istics of Cylindrical Thorium Breeder Reactors, USAEC Report ORNL-2165,
Oak Ridge National Laboratory, October 1956.

38. M. F. Durer and I. L. WiLsoN, A Two-zone Slurry Reactor, Report
AECL-249, Atomic Energy of Canada Limited, December 1953.

39. ComMoNWEALTH EpisoN CompaNY (NucLEArR Power Group), A Third
Report on the Feasibility of Power Generation Using Nuclear Energy, USAEC
Report CEPS-1121(Del.), June 1953.

40. Paciric NorTHWEST PoweEr Groupr, Aqueous Homogeneous Reactors,
USAEC Report PNG-7, February 1956.

41. CoMmmoNWEALTH EprsoN CompaNY (NUcLEAR PowER GROUP) AND BaB-
cock & Wircox Company, Evaluation of a Homogeneous Reactor, Nucleonics
15(10), 64-71 (1957).

42. American Standard and Sanitary Corporation.

43. J. C. Griess and H. C. SAVAGE, in Homogeneous Reactor Project Quarterly
Progress Report for the Period Ending July 31, 19566, USAEC Report ORNL-
2148(Del.), Oak Ridge National Laboratory, 1956. (p. 77)

44. T. A. HucHEs, Steam-Water Mixture Density Studies in a Natural Circula-
REFERENCES _ 27

tton High-pressure System, USAEC Report BW-5435, Babcock & Wilcox Com-
pany, February 1958.

45. P. C. Zmora and R. V. BaiLey, Power Removal from Boiling Nuclear
Reactors, USAEC Report CF-55-7-43, Oak Ridge National Laboratory, 1955.

46. L. G. ALExANDER and S. JAYE, A Parametric Study of Rate of Power
Removal from Homogeneous Botling Reactors, USAEC Report CF-55-9-172,
Oak Ridge National Laboratory, 1955.

47. J. A. LANE et al., Aqueous Fuel Systems, USAEC Report CF-57-12-49,
Oak Ridge National Laboratory, 1957.

48. I. SPIEWAK, in Homogeneous Reactor Project Quarterly Progress Report for
the Period Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak Ridge
National Laboratory, 1957. (p. 14)

49. J. M. StEIN and P. R. KAsTEN, Boiling Reactors: A Preliminary Investi-
gation, USAEC Report ORNL-1062, Oak Ridge National Laboratory, December
1951.

50. R. J. RickEerT et al., A Preliminary Destgn Study of a 10-Mw Homogeneous
Boiling Reactor Power Package for use in Remote Locations, USAEC Report
CF-53-10-23, Oak Ridge National Laboratory, 1953.

51. H. R. ZertLiN et al., Boiling Homogeneous Reactor for Power and U-233
Production, USAEC Report CF-55-8-240, Oak Ridge National Laboratory,
August 1954.

52. A. I. AvuicuaNow et al.,, A Boiling Homogeneous Nuclear Reactor for
Power, in Proceedings of the International Conference on the Peaceful Uses of
Atomic Energy, Vol. 3. New York: United Nations, 1956. (p. 169)

93. P. C. Zmora, Comments on Boiling Slurry Blankets for Homogeneous
Reactors, USAEC Report CF-55-9-125, Oak Ridge Natlonal Laboratory, Sept.
27, 1955.

54. P. C. Zmora, Boiling Blanket for TBR-Power Utilization, USAEC Report
CF-55-3-57, Oak Ridge National Laboratory, Mar. 9, 1955.

55. D. E. HurL, Possible Applications of UFg, USAEC Report MonN-336,
Oak Ridge National Laboratory, 1947.

56. S. A. CorcaTE and R. L. AamopT, Plasma Reactor Promises Direct
Electric Power, Nucleonics 15(8), 50-55 (1957).

97. R. R. HaLix et al., Oak Ridge National Laboratory, 1952. Unpublished.

58. H. W. Graves et al., Oak Ridge National Laboratory, 1953. Unpublished.

59. C. C. SiLvERSTEIN, Fluidized Bed Reactor Cores, Nucleonics 15(3), 101
(1957).

60. J. B. Morris et al.,, The Application of Fluidization Techniques to
Nuclear Reactors, Trans. Inst. Chem. Engrs. London 34, 168-194 (1956).

61. C. L. TEETER et al., Fluidized Bed Reactor Study, USAEC Report CF-57-
8-14, Oak Ridge National Laboratory, August 1957.

62. J. J. WENT and H. pE BruyN, The Design of a Small Scale Prototype of
a Homogeneous Power Reactor Fueled with Uranium Oxide Suspension, in
Proceedings of the International Conference on the Peaceful Uses of Atomic Energy,
Vol. 3. New York: United Nations, 1956. (P/936)

63. J.J. WENT and H. pE BruyYN, Power Reactors Fueled with ‘“Dry”’ Suspen—
sions of Uranium Oxide, in Proceedings of the International Conference on the
28 HOMOGENEOUS REACTORS AND THEIR DEVELOPMENT  [cHAP. 1

Peaceful Uses of Atomic Energy, Vol. 3. New York: United Nations, 1956
(P/938)

64. J. A. Lane, Oak Ridge National Laboratory, personal communication
Oct. 22, 1957.

65. P. R. CrowLEY and A. S. Kirzes, Feasibility of a Fluidized Thorium
Ozide Blanket, USAEC Report CF-53-9-94, Oak Ridge National Laboratory,
Sept. 2, 1953.

66. I. SpiEwak and J. A. Harrorp, Abrasion Test of Thoria Pellets, USAEC
Report CF-54-3-44, Oak Ridge National Laboratory, Mar. 9, 1954.

67. S. A. Reep, Summary of Toroid Run No. 153: Circulation of Slurries of
Thoria Spheres at 260°C and 26 Fps Using Pins of Type 347 S.S., SA-2120-B
Steel, Titanium-75A, and Zircaloy-3A, USAEC Report CF-57-11-46, Oak Ridge
National Laboratory, Nov. 11, 1957.
CHAPTER 2

NUCLEAR CHARACTERISTICS OF
-~ ONE- AND TWO-REGION
HOMOGENEOUS REACTORS*

Nuclear characteristics refer to the conditions and material concentra-
tions under which reactor systems will remain critical, the relative changes
in concentration of materials within the system as a function of reactor
operation, and the time behavior of variables in the reactor system which
occur when deviations from criticality take place. The material concen-
trations are closely connected with fuel costs in power reactors, while
reactor behavior under noncritical conditions is closely related to the safety
and control of the reactor system. These nuclear characteristics are de-
termined from the results obtained from so-called ‘‘reactor criticality’’
and “reactor kinetic’’ calculations. In such studies, certain parameter
values pertaining to nuclei concentrations and reaction probabilities are
used ; for convenience some of these are listed in Section 2-2.

2—-1. CriTicALITY CALCULATIONS

Criticality studies are also termed ‘‘reactor-statics studies.” In these
studies the concentration of the various nuclei present can vary with
time, but it is assumed that the condition of criticality will be maintained.

The statics of chain reactions in aqueous-homogeneous reactors are of
interest primarily in connection with the estimation of the inventory of
fuel and fertile material, power density at the wall, the flux distribution
nside the reactor, and the rate of production of fissionable isotopes. These
enter into economic calculations pertaining to fuel costs in power reactors
and also into criteria specifying the design of the system. The most im-
portant factors determining criticality dre geometry; nature, concentra-
tion, and enrichment of the fuel; nature and distribution of other com-
ponents in the reactor; and operating temperature and pressure. The
production of fissionable isotopes depends primarily on the neutron economy
of the system and will be a function of the relative competition for neu-
trons between the fertile material and the various other absorbers. The
latter include materials of construction, moderator, fuel components,
fission products, and various nonfissionable nuclei formed by parasitic
neutron capture in the fuel. In designing a reactor for the production of

*By P. R. Kasten, Oak Ridge National Laboratory.
‘ 29
30 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

fissionable isotopes, it is therefore important to choose materials (other
than fuel and fertile material) which have low neutron-capture cross
sections. This, in turn, leads to the selection of D20 as the moderator in
nearly all cases.

Although criticality is assumed at all times, the concentration of fuel
isotopes can change appreciably with time owing to the relative competi-
tion between isotope formation, decay, and neutron-absorption processes.
In many studies it has been assumed that the reactor has operated for
such times that the steady-state conditions apply with regard to the nuclei
concentrations. This simplifies the isotope equations but may not always
give an adequate picture of the actual concentrations which would be
present in an operating reactor.

2-1.1 Calculation methods. Since both light and heavy water are ex-
cellent moderators, the energy of fission neutrons is rapidly degraded, with
the result that most of the fissions are produced by thermal neutrons.
Under these conditions either the modified one-group or the two-group
diffusion equations are usually applicable for criticality calculations. For
simplicity, this discussion will be limited to spherical reactors.

For a bare, spherical reactor the criticality condition (assuming Fermi
age theory) is given by

ZPpme Bt | fu 2p(u)
SE+ DuB? 'Jy  EZ.(W)

 

1=»| [P Er@law}, @)

where B2 = (7w/R)?,

Dy, = thermal diffusion coefficient,

D(u) = diffusion coefficient as a function of lethargy,

D = resonance escape probability to thermal energy,

p(u) = resonance escape probability to lethargy wu,

R =radius of reactor plus extrapolation distance,

u = lethargy of neutrons,

utn = u evaluated at thermal energy,

y = neutrons emitted per fission,

£ = average lethargy increment per neutron collision,

¥ = macroscopic cross section; superscript th refers to thermal

value; subscripts f, a, and ¢ refer to fission, absorption,
and total cross sections, respectively; Z(u) refers to Z
evaluated as a function of lethargy,

7sn = Fermi age to thermal energy,

7(u) = Fermi age to letha =|
(u) ermi age to lethargy u | ES )
2-1] CRITICALITY CALCULATIONS 31

By introducing €, the “fast fission factor,” where

 

. 2 pme P (M 2 B gy
__total fissions __ 2t DwB?2 " Jy E2(u)
~ thermal fissions > P ne— B'Tth ’
2% 4 DyB?
(2-2)
Eq. (2-1) becomes Joo— B?Tth
l=<"psra (2-3)
1+ B2L2%
where &k = nepwfin = infinite multiplication constant,
n = neutrons emitted per neutron absorption in fuel,

fin = fraction of thermal neutrons absorbed in fuel,

Dy, thermal diffusion coeflicient

L2 =—r= . : —
th ™ >th " macroscopic thermal absorption cross section

 

Replacing the exponential term in Eq. (2-3) by (1/1+ B?tw), the two-
group criticality condition is obtained as

k= (14 B2ra)(1+ B2L3). (2-4)

Although Egs. (2-1), (2-3), and (2-4) imply that resonance fissions in
the fuel are considered, in usual practice Eqgs. (2-3) and (2-4) are used on
the basis that (ep)sue is equal to unity. Using this assumption, the values
of € and p to be used in evaluating k are those associated with the fertile
material. In the subsequent results and discussion, calculations based on
Egs. (2-3) and (2-4) consider that (ep)suel is equal to unity, while calcula-
tions based on Eq. (2-1) explicitly consider resonance absorptions and
fissions in fuel based on a 1/FE resonance-energy flux distribution.

The breeding ratio (BR) is defined as the number of fuel atoms formed
per fuel atom destroyed for the reactor system. For a bare reactor, assum-
ing that the resonance flux is independent of lethargy and that absorptions
in fertile material produce new fuel, the BR is given by

u — B21- % -— B21-
‘b th 27~ B7p du i [ “th Ztertile P~ 5T du
e [ 1= v [ SR 4z [ Steie B

 

 

 

 

BR = 1£2t — £24——t ’
th (w2 Bp du] th f“th 2, (fuel) pe~ B du
2g (fuel)[l v f(; i, + v A £,
| (2-5)
where 28 ... = thermal absorption cross section of fertile material,
Ziortile = absorption cross section of fertile material at lethargy u,

2q(fuel) = fuel absorption cross section at lethargy wu.
32 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

If resonance absorptions in fuel are neglected, the conventional two-group
formula. is obtained as

— E}"tl.alrtile 77(1 - pfertlle)
BR =38 Gue) T 1T Bra - 0

In the subsequent discussion, reference to one-region, two-group calcula-
tions implies use of Eqs. (2-4) and (2-6) to calculate critical mass and
breeding ratio, respectively.

The two-group diffusion equations were used for two-region reactor
calculations. The effect of a thin shell between the two regions upon
reactor criticality and breeding ratio was taken into account by considering
the shell absorptions by means of a boundary condition; the effect of a
pressure-vessel wall was taken into account by using an “‘effective’’ extra-
polation distance in specifying the radius at which the flux was assumed
to be zero. The two-group equations were written as:

D12y — Znbre+ o Duehre =, 2-7)
DycV2Poc — ZiscPsc + PeiscPre = 0, (2-8)
D2 — Zntn+ 22 Zabp =0, (2-9)
D V2P — ZpPsb + PeissPrs = O. (2-10)

The subscripts f, s, ¢, and b refer, respectively, to the fast flux, slow flux,
core region, and blanket region; D is the diffusion coefficient; ¢ is the
neutron flux; 2, refers to the effective cross section for removing neutrons
from the fast group; and 2, refers to the thermal absorption cross section.
Other symbols have the same meaning given previously, with £ calculated
on the basis that (ep)qe = 1.

The boundary conditions used assume that the fast flux has the same
value on the core side of the core-tank wall as on the blanket side; the
same is also true of the slow flux. It is also specified that the fast flux and
slow flux become zero at some extrapolated reactor radius. At the core-
tank wall, the net fast-neutron current on the blanket side is assumed equal
to that on the core side, while the net slow-neutron current on the core
side is assumed equal to the flux rate of neutron absorptions in the core
tank plus the net slow-neutron current on the blanket side.

A multigroup formulation can be obtained by adding neutron groups
with energies intermediate between the fast and slow groups specified in
Eqgs. (2-7) through (2-10). These intermediate groups would be essentially
of the form

DiV2p;, — Zipi+ pi-12:i-19;-1=0, (2-11)

where ¢ represents the 7th group of neutrons, and ¢ increases with decreas-
2-1] CRITICALITY CALCULATIONS 33

ing neutron energy. Boundary conditions analogous to those specified
above would apply. Such a formulation assumes that neutrons in the <th
group are always slowed down into the ¢4 1 group, corresponding to a
relatively small number of fast groups. Using about 30 groups or more,
multigroup methods [1] used in homogeneous reactor calculations consider
that fast neutrons are born in the various fast groups in accordance with
the fraction of fission neutrons generated in the particular group; that for
all materials but hydrogen, neutrons are ‘‘slowed down’ from one energy
group to the group immediately below; and that with hydrogen, the pos-
sibility exists for a neutron to pass from one group to any group below as a
result of one scattering collision. Other multigroup methods [2] of calcu-
lation have been devised which consider that a scattering collision degrades
a neutron into the groups below in accordance with the probability for
degradation into a particular group. |

2-1.2 Results obtained for one-region reactors. The majority of critical
calculations for large one-region reactors have been based on Eqgs. (2-3)
and (2-4), in which all the fissions are effectively assumed to take place in
the thermal group [3]. However, if the effective value of n(fuel) for the
resonance region is less than the value for the thermal region, the above
models may not be adequate. Some multigroup calculations [4] have
been done for uranium-water systems at an average temperature of 260°C;
in Fig. 2—-1 are shown critical-mass requirements for light water-uranium sys-

1000

 

500 [—

200

 
  
 
   

100

50

RN

EEER

L Lt

N
I

|

 

 

L ] L1 1 L1 it |
10 20 50 100 200

Radius of Bare Sphere, cm

 

Fic. 2-1. U235 mass and critical size of uranium light-water mixtures at 260°C.
Assumed densities: U235 = 18.5 g/cm3, U238 = 18.9 g/cm3, H20 = 0.8 g/cm3.
34 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

 

SITT R I LT

Moles HZO

  
   

()

1/40)

 

—

100

A O N0

  
 

 

 

~ Pure D20 Moderator

 

I

 

u?, kg
N
|

 

 

—

LTl

Pure HoO Moderator

 
 

 

| L[]

 

I

|

 

 

1L R | | | 14
78910 2 3 4 56789100 2 3 4 56789

Radius of Bare Sphere, cm

 

Fic. 2-2. U235 mass and critical size of bare spherical reactors moderated by
H20—D20 mixtures at 260°C. Assumed densities: U235 = 18.5 g/cm3, H20 = 0.8
g/cm3, D20 = 0.89 g/cm3.

tems, while Fig, 2-2 gives the critical-mass requirements for U235-D;0-
H20 systems. Initial conversion ratios for light-water systems are given
in Ref. [4].

Nuclear calculations have also been performed [5] using Eqgs. (2-1)
and (2-5), with the 1/E component of the flux starting at energies of 6 kT'.
The calculations were for single-region reactors containing only ThOq,
U2330g2, and D20 at 300°C with the value of 52?3 in the resonance region
considered to be a parameter. Critical concentrations thus calculated are
given in Fig. 2-3 for zero neutron leakage. The value for 2?3 in the thermal
energy region was taken as 2.25. The value for %23 in the resonance region
has not been firmly established; based on available data, 723/%% lies
between 0.9 and 1.

When the neutron leakage is not negligible and 723 /9% is less than 1, a
finite thorium concentration exists for which the breeding ratio is a maxi-
mum. This is indicated in Fig. 2—4, in which the initial breeding ratio is
2-1] CRITICALITY CALCULATIONS 35

 

 

 

 

 

20
|
10— —
5 - _
£ | _
o
o -
&
S5 S| —
K
c - —
2
3
€
8 _— —_
c
o
O
2 — —
1 l | | | | l | l
0 200 400 600 800 1000

Concentration of Thorium, g/liter

Fic. 2-3. Fuel concentration as a function of thorium concentration and value
of 7res/Ntn for an infinite reactor.

 

 

 

 

 

 

 

 

1.3 T
| | | |23
1.0

1.1 — —
.0
S 10— ]
o2
o
£
T 09 —
o
o
:..g 0.8 |— — e e o e =m == Reaictor Diameter Infinite —
= ——————— Reactor Diameter = 14 ft

07l @ mmmmmm—e—e- Reactor Diameter = 10 ft

ot | | | |

0 200 400 600 800 1000

Thorium Concentration, g/liter

Fic. 2-4. Breeding ratio as a function of thorium concentratlon and nres/n th 1N
a one-region reactor. Reactor temperature = 300°C, nth = 2.25.
36 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

given as a function of thorium concentration, reactor diameter, and rela-
tive value of #23. The breeding ratio goes through a maximum owing to
the increase in resonance absorption in fuel as the thorium concentration
is increased. |

If the above reactors were fueled initially with U235, the initial breeding
ratio would have a maximum value of 1.08 rather than 1.25; however, the
curves would have about the same shape as those presented in Fig. 24,
and the value for 922 /92 would be between 0.8 and 0.9.

Comparison of the above results with those obtained using a two-group
model shows that if 7,.,/7, 1s equal to 1, the breeding ratio obtained by
the two methods is about the same; however, the critical concentration is
about 309, higher when the two-group model is used. If 5% /92 < 1, the
value for the breeding ratio obtained using the two-group model will tend
to be higher than the actual value; however, if the fertile-material con-
centration is low (about 200 g/liter or less) and the reactor size large, little
resonance absorption occurs in fuel. Under these conditions the two-group
model should be adequate for obtaining the breeding ratio and conservative
with respect to estimating the critical concentration.

2-1.3 Results obtained for two-region reactors. Most two-region re-
actors have been calculated on the basis of the two-group model. Results
have also been obtained using multigroup calculations which indicate that
the two-group method is valid so long as the value of y(fuel) is independent
(or nearly so) of energy, or so long as nearly all the fissions are due to ab-
sorption of thermal neutrons.

To compare results obtained by different calculational methods, breed-
ing ratios and critical fuel concentrations were obtained [6] for some two-
region, D20-moderated thorium-blanket breeder reactors using a multi-
group, multiregion Univac program (“Eyewash”) [1] and a two-group,
two-region Oracle program [7]. In these calculations operation at 280°C
was assumed; a z-in.-thick Zircaloy—2 core tank separated the core from the
blanket; a 6-in.-thick iron pressure vessel contained the reactor; and ab-
sorptions occurred only in U233 and thorium in the core and in thorium in
the blanket. Twenty-seven fast groups, one thermal group, and four
regions (core, Zircaloy-2 core tank, blanket, and pressure vessel) were
employed in the multigroup model. The two-group parameters were com-
puted from the multigroup cross sections by numerical integration [8].
In the two-group, two-region calculations a ‘‘thin-shell” approximation [9]
was used to estimate core-tank absorptions, while the effect of the pressure
vessel was simulated by adding an extrapolation distance to the blanket
thickness. -

In the multigroup studies, various values for #23 in the resonance region
were considered. In one case the value of %23, was assumed to be constant
2-1] CRITICALITY CALCULATIONS 37

and equal to the thermal value.* In another the variation of %23 in the
resonance region was based on cross sections used by Roberts and Alex-
ander [10], which resulted in an 723 /%% of about 0.95; in the third case the
effective value for 23 in the resonance region was assumed to be essen-
tially 0.8 of the thermal value of 2.30.

The initial breeding ratios and U233 critical concentrations obtained from
the Eyewash and two-group, two-region calculations are given in Fig. 2-5
for slurry-core reactors (zero core thorium concentration also corresponds
to a solution-core reactor). With solution-core reactors the effect of the
value of 723 upon breeding ratio was less pronounced than for slurry-core
reactors, since fewer resonance absorptions take place with the lower fuel
concentrations. The blanket thorium concentration had little influence
upon the above effect for blanket thorium concentrations greater than
250 g/liter.

As indicated in Fig. 2-5, the breeding ratio is rather dependent upon
the value of 723; the loss in breeding ratio due to a reduced value of 73
in the resonance region is most marked for the slurry-core systems. For
these reactors, relatively more fissions take place in the resonance-energy
region as the core loading is increased, owing to the “hardening” of the
neutron spectrum. Figure 2-5 also shows that the two-group model gives
breeding ratios which are in good agreement with those obtained from
the multigroup model, so long as 725, does not deviate significantly from
n2. Reported measurements [11-13] of 9?2 as a function of energy indi-
cate that for the reactors considered here, the value of % /7% would be
about 0.95; the results given by curve “‘a” in Fig. 2-5 are based effectively
on such a value of 723 /92 and indicate that two-group results are valid.

In general, for the cases studied it was found that for a heavily loaded
blanket (or core), the two-group values of total neutron leakage were
larger than the total leakages obtained from the multigroup calculation
(the multigroup model allowed for competition between fast absorptions
in fuel and fast leakage, while the two-group model assumed that fast
leakage occurred before any resonance absorption occurred). The multi-
group results were also used to calculate the fast effect, e, previously
defined in Eq. (2-2). It was found that resonance fissions accounted for
10%, to 409, of the total fissions in those reactors containing from 0 to
300 g Th/liter in the core region. With no thorium in the core region,
changing from a 4-9 reactor (4-ft-diameter core and a 9-ft-diameter pres-
sure vessel) to a 6-10 reactor decreased core resonance fissions from about
149, to 10%,.

If the reactor core size is small, the two-group method does not ade-
quately treat leakage of fast neutrons; for this case two-group results may

*The thermal value for 723 was assumed to be equal to 2.30 instead of 2.25 used
in more recent calculations (the 2.25 value is believed to be more accurate).
38 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

 

 

   

 

 

 

 

 

 

 
    

 

1.30 I |
) e — — — = ==
Two-Group, Two-Region g
v , 23
o, o”
'~:'-: _fo-" a b “Eyewash” <___r2e; > =0.8
21.20 - —
3  meee——
Q
g
o
1.10
a. Estimated from Averaged Cross Sections
“Eyewash” used by Roberts and Alexander (11)
) b. Estimated for 723 Independent of Energy
10
Two-Group, Two-Region
23
“Eyewash” Tres
23
T th / ef
5+

Critical Concentration, g of y233 per liter

 

 

l |
0 100 200 300
Core Thorium Concentration, g/liter

 

Fig. 2-5. Breeding gain and critical concentration for slurry-core reactors vs.
core thorium concentration. Core diameter = 6.0 ft, pressure vessel diameter = 10.0
ft, blanket thorium concentration = 1000 g/liter, blanket U233 concentration = 3.0
g/kg of Th.

not be adequate even though 7, is equal to 9,,. This is indicated by the
experimental [14] and calculated results for the HRE-2 given in Fig. 7-15.
As 1llustrated, there is excellent agreement between the experimental
data and the data calculated by a multigroup method and by a ‘‘har-
~ monics” method, but not with the results from the two-group model.
The harmonics calculation [15] referred to in Fig. 7-15 does not take into
account fast fissions but does treat the slowing-down of neutrons in a more
realistic manner than does the two-group calculation. The multigroup
result [16] indicated that about 139, of the fissions were due to neutrons
having energies above thermal.

A comparison [15] of the U235 critical concentrations predicted by the
2-2] NUCLEAR CONSTANTS USED IN CRITICALITY CALCULATIONS 39

 

0T 17T T 1 T 1 T T T T 1

800 — -

700 — —

600 — Age —

Age-Yukawa

500 |— | —
Yukawa

400 |— ' —

3, (Fuel) /3, (D90) in Core

300 — —

200 — —

 

 

100 | 1 | L+
50 60 70 80 90 100 110 120
Outer Radius of Reflector, cm :

 

Fic. 2-6. Comparison of critical concentrations obtained for various slowing-
down kernels in D20. Core radius = 39 cm, total age = 237 cm? for all kernels,
diffusion length of pure moderator, Lo? = 40,200 cm2. Fuel only in core.

use of different slowing-down kernels in D2O-moderated reactors is shown
in Fig. 2-6. In the age-Yukawa kernel [given by (e™™/4™1/(4mw71)%/2) X
G \/;2-/47r7'2r)], the “‘age” parameters were taken to be 158 cm? for 73
and 79 ecm? for 72. For both the age kernel (given by e~"/47/(4mwT)3/2)
and the Yukawa kernel (given by e~"/ VT AnT r), T was taken to be 237 cm?.
The results show that in small reactors the calculated critical concentration
obtained using the two-group method (Yukawa kernel) is substantially
lower than that obtained using either an age-Yukawa or an age kernel to
represent the neutron distribution during the slowing-down process. Since
the age-Yukawa kernel is believed to be the proper one to use for D20,
and the HRT is a “small” reactor (in a nuclear sense), it is not surprising
that the two-group results given in Fig. 7-15 are appreciably different
from the experimental results.

2-92 NucLEAR ConsTANTS UsED IN CRITICALITY CALCULATIONS

In obtaining the nuclear characteristics of reactors, it is necessary to
know the probabilities with which different events occur. These reaction
40 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2
probabilities are usually given on an atomic basis in terms of cross sec-
tions [17]. Because of their diverse applications, it is necessary to present
reaction probabilities in this manner; however, in calculating the nuclear
characteristics of homogeneous reactors, it is convenient to combine funda-
mental data concerning atomic density and reaction probabilities so as to
facilitate critical calculations. This has been done to a limited extent in
this section. Listed here are some nuclear data and physical properties of
uranium isotopes, uranyl sulfate, heavy water, thorium oxide, and Zirca-
loy—2 used iIn two-group calculations for thorium breeder reactors [18].

TABLE 2-1

THERMAL MIicroscoric ABSORPTION CROSS
SECTIONS AT VARIOUS TEMPERATURES

(Corrected for Maxwell-Boltzmann distribution and also non-(1/») correction)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Element N e‘;‘gg&;’i‘;fi‘ty | 20°C 100°C 280°C
04, barns [17]
U233 588 526 460 376
U234 92 82 72 59
U235 689 595 515 411
U236 6 5.3 4.7 3.9
238 2.73 2.42 2.15 1.76
Pa233* [19] 60 130 130 130
Th232 7.45 6.60 5.85 4.81
Pu239 1025 975 905 950
Py240* 250 600 700. 1000
Pu24 1399 1240 1118 952
S 0.49 0.43 0.39 0.32
Li7 (99.98%,) [20] 0.23 0.20 0.18 0.15
o s, barns [17]

233 532 472 412 337
235 582 506 438 350
Pu239 748 711 660 693
Pu24 970 860 776 660

 

 

*Estimates of the effective cross section in typical homogeneous-reactor neutron
spectrums (except for 2200 m/sec value); these values include contributions due to
resonance absorptions. (Although these values were not used in the calculations
presented, they are believed to be more accurate than the ones employed. Values
used for Pa?33 were 133, 118, and 97 barns at 20, 100, and 280°C, respectively.)
2-2] NUCLEAR CONSTANTS USED IN CRITICALITY CALCULATIONS 41

2-2.1 Nuclear data. Table 2—1 lists thermal microscopic absorption and
fission cross sections for various elements and for various temperatures.
Table 2-2 lists thermal macroscopic absorption cross sections for H20,
D20, and Zircaloy—2, and the density of HoO and D20 at the various

TABLE 2-2

THERMAL MAcroscoPic ABSORPTION (CROSS SECTIONS
AND DENSITIES AT VARIOUS TEMPERATURES [18]

(Corrected for Maxwell-Boltzmann distribution on basis of 1/v cross section)

 

 

 

 

 

 

 

Element 20°C 100°C 280°C
2 .(H20) 0.0196 0.0167 0.0107
2,(99.759, D20) 8.02 X 105 6.85 X 10° 4.44 X 105
2 (Zircaloy-2) 0.00674 0.00598 0.00491
0(D30) 1.105 1.062 0.828
o(H;0) 1.000 0.962 0.749

 

 

temperatures. All cross sections listed under the columns headed by °C
have been corrected for a Maxwell-Boltzmann flux distribution.

Values of 1 and v for the various fuels, and the fast and slow diffusion
coefficients for Zircaloy are given in Table 2-3.

TABLE 2-3

SoME NUCLEAR CONSTANTS FOR
UraNIUM, PLUTONIUM, AND ZIRCALOY—2

Values of n and » for U and Pu [21]

 

 

 

Element 7 v
U233 2.25 2.50
U235 2.08 2.46
Pu?3° 1.93 3.08
Pu?4 2.23 3.21

 

 

 

Diffusion coefficients for Zircaloy-2: [22]
Dy = D2 =10.98 for all temperatures,
where D; = fast diffusion coefficient,
D2 = slow diffusion coefficient.

 

 

 
42 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

Data for two-group calculations are summarized in Table 2-4 for 7, D,
D2, and p as functions of fertile-material concentration in mixtures of

TABLE 24

Two-GrRouP NUCLEAR CONSTANTS* FOR D2O-MODERATED
SysTEMs AT 280° C [18]

 

 

 

Fertile-material concentration, T, D, Do,
g/liter cm? cm cm p
Th (in ThO2-D20)
0 212 1.64 1.24 1.000
100 212 1.62 1.23 0.909
250 213 1.60 1.22 0.825
500 213 1.56 1.20 0.718
1000 215 1.50 1.16 0.554
U238 (in U02804“D20)
0 212 1.64 1.24 1.000
100 200 1.57 1.20 0.875
250 189 1.49 1.15 0.801
500 179 1.40 1.10 0.720
1000 173 1.28 1.04 0.595
U238 (in UOzSO4—List4—D20)T
0 212 1.64 1.24 1.000
100 198 1.55 1.19 0.873
250 185 1.45 1.13 0.797
500 173 1.33 1.07 0.705
1000 165 1.18 0.99 0.525

 

 

 

 

 

 

 

*r = Fermi age; D; = fast diffusion coefficient; D2 = slow diffusion coefficient;
p = resonance escape probability.
TLi2SO4 molar concentration equal to UO2SO4 molar concentration.

fertile material and heavy water (99.759, D20) at 280°C. Materials con-
sidered are ThO2-D20, U0O2504-D20, and U02504-Li2S04~D20 where
the molar concentration of Li2SO4 is the same as the UO»SO4 molar con-
centration. Reference [18] gives corresponding data at-other temperatures
and also gives some values for the case of H20O as the moderator.

The diffusion coefficients and ages were calculated by a numerical inte-
gration procedure [8]. The fast diffusion constant, D;, and the Fermi
age, 7, are based on a 1/F flux distribution, and the slow diffusion con-
stant, Do, 1s based on a Maxwellian flux distribution.
2-3] STEADY-STATE FUEL CONCENTRATIONS 43

2-2.2 Resonance integrals. Formulas used in calculating resonance in-
tegrals (RI) are given below.

 

 

For U238,
>, \0-471 >, .

RI = 2.69 <]V2§> ’ 0= VE = 4% 103, (2-12)
In RI = 5.64 — (2—7}%)0—65 ];?2—8 > 4 X 103, (2-13)
RI( ) = 280 barns. (2-14)

For Th232;

>, \0-253 >,

RI=8.33 <W> ’ 0= A0z = 4500, (2-15)
RI = 70 barns, -]%2- > 4500, | (2-16)
RI(« ) = 70 barns. (2-17)

9-3. FueL CONCENTRATIONS AND BREEDING RATIOS UNDER INITIAL
AND STEADY-STATE CONDITIONS

The relationships between breeding ratio and reactor-system inventory
determine the fuel costs in homogeneous reactors. The breeding ratio
depends on neutron leakage as well as relative absorptions in fuel fertile
material and other materials present, while material inventory is a function
of reactor size and fuel and fertile-material concentrations; thus a range
of parameter values must be considered to aid in understanding the above
relationships. Based on results given in Section 1-1.3, it appears that the
two-group method gives satisfactory results for critical concentration and
breeding ratio for most of the aqueous-homogeneous systems of interest.
This permits survey-type calculations to be performed in a relatively short
time interval. The results given below are based on the conventional two-
group model.

In steady-state operation, the concentration of the various nuclides
within the reactor system does not change with time. During the initial
period of reactor operation this situation is not true, but is approached
after some time interval if neutron poisons are removed by fuel processing.
Under steady-state operation it is necessary to consider the equiltbrium
isotope relationships. In thorium breeder reactors this involves rate ma-
terial balances on Th, Pa233, U233 [J234 235 U236 fission-product poisons,
and corrosion products. (The uranium isotope chain is normally cut off at
U236 since this is a low-cross-section isotope, and neutrons lost to the suc-
cessors of U236 would tend to be compensated for by fission neutrons gen-
44 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

erated by some succeeding members of the chain.) In uranium-plutonium
reactors, steady-state rate material balances were made on U235 1236
U238, Pu?39 Pu?%0) and Pu?*!; all other higher isotopes were assumed either
to be removed in the fuel-processing step or to have a negligible effect
upon the nuclear characteristics of the reactor.

Although equilibrium results give the isotope ratios which would be
approached in a reactor system, much of the desired nuclear information
can be obtained by considering “‘clean’ reactors, i.e., reactors in which
zero poisons exist, corresponding to initial conditions, or to criticality con-
ditions at reactor startup. This is a result of the rather simple relationships
which exist between breeding ratio, critical concentration, and fraction
poisons, and the ability to represent the higher isotopes by their fraction-
poison equivalent.

2-3.1 Two-region reactors. In order to estimate the minimum fuel costs
in a two-region thorium breeder reactor, it is important to determine the
relation between breeding ratio and the concentrations of fuel (U233) and
fertile material (Th?32) in the core and blanket. Similar considerations
apply to uranium-plutonium converter reactors. The breeding or con-
version ratio will depend on neutron leakage as well as relative neutron
absorptions in fuel, fertile material, and the core-tank wall; therefore a
range of core and pressure-vessel sizes must be considered.

Since fabrication problems and the associated cost of pressure vessels
capable of operating at 2000 psi increase rapidly for diameters above 12 ft,
and since the effect of larger diameters on the nuclear characteristics of the
two-region reactors is relatively small, 12 ft has been taken as the limiting
diameter value. Actually, in most of the calculations discussed here, the
inside diameter of the pressure vessel has been held at 10 ft and the core
diameter allowed to vary over the range of 3 to 9 ft.

In addition to the limitation on the maximum diameter of the pressure
vessel, there will also be a limitation, for a given total power output, on
the minimum diameter of the core vessel. This minimum diameter is de-
termined by the power density at the core wall, since high power densities
at the wall will lead to intolerable corrosion of the wall material (Zircaloy—2).
In order to take this factor into consideration, the power densities, as well
as critical concentrations and breeding ratios, were calculated for the
various reactors.

2-3.2 Two-region thorium breeder reactors evaluated under initial con-
ditions. The results given here are for reactors at startup; although the
trends indicated apply to reactors in steady-state operation, the values
given here for the breeding ratio and fuel concentration would be some-
what different than those for steady-state conditions.
2-3] STEADY-STATE FUEL CONCENTRATIONS 45

Calculations of breeding ratio, the power density at the inside core
wall, and the maximum power density were carried out for some spherical
reactors with 200 g Th/liter in the core. The blanket materials considered
were heavy water (99.759, D20), beryllium, and ThOs-heavy water sus-
pensions. The inside diameter of the pressure vessel was fixed at 10 ft for
one set of calculations and at 12 ft for a second set; core diameters ranged
from 6 to 9 ft in the first set and from 6 to 11 ft in the second set. The
average temperature of all systems was taken as 280°C, and for the purpose
of calculating power densities at the core wall, the total thermal power
was taken as 100 Mw. A 3-in-thick Zircaloy—2 core tank was assumed to
separate the core and blanket in all reactors, and the value of %23 was
taken as 2.32. (A more accurate value of 9?3 is presently considered to be
7= 2.25.) No account was taken of fission-product-poison buildup, pro-
tactinium losses, or fuel buildup in the blanket. The results obtained [23]
indicate that the breeding ratio increases for any core diameter by re-
placing either a D20 or Be blanket with one containing ThOg2; no sig-
nificant increase in breeding ratio is obtained by increasing the blanket
thorium concentration above 2 kg Th/liter; for reactors with fertile ma-
terial in the blanket, the breeding ratio and wall power density increase
with decreasing core diameter.

 

 

 

 

   
  

 

 

 

 

 

 

 

1.3 [ I l [
Thorium in Blanket, g/liter

o L2000 _ ,6 ft Dia. Core
2 ].2 / '7 fl’ Dial. COre
- 1000 ) I
._g 500< 8 fl' Dic. Core
@
o 1.1
o

1 0 | | | |
» 15
3T T T | 500
T NN i
3 §\ \\ 1000 éé ft Did' Core
- \\\ \\
S 10 A \‘\://
‘@ NN - 500
g N NN
a N \\_//]_QQO_— 7 ft Dia! Core
g 5PN\ _~3IZo
3 Nao 2ooo<
fi Seo 8 ft Didl Core
s | | | | | | B

 

0 100 200 300

Thorium in Core, g/liter

Fiag. 2-7. Effect of core thorium concentration on breeding ratio and wall power
density of two-region slurry reactors. 7= 2.25, total reactor power= 100 Mw
(heat), pressure vessel = 10 ft ID, U233 in blanket = 3 g/liter, poison fraction = 0,
temperature = 280°C.
46 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

Additional results [24] obtained on the same bases given above except
that 23 was considered to be 2.25, and that the U233 concentration in the
blanket region was varied, are given in Figs. 2-7 through 2-9. Results
were obtained for 10-ft-diameter pressure vessels and for core diameters
of 6, 7, and 8 ft; the blanket thorium concentration was 500, 1000, or
2000 g/liter, while the core thorium concentration was 0, 100, 200, or
300 g/liter. Generally, the results in Figs. 2-7 through 2-9 are complete
for the 7-ft-diameter-core reactors, while for the 6- and 8-ft-diameter-core
reactors results are shown only for the parameter-value extremes. The
variation of results with parameter value is practically the same for all
three core diameters, and so all the essential results are presented in the
figures.

In all cases slurries of D2O-ThO2-U?3303 are assumed in both the core
and blanket regions, and all power densities are based on the assumption
that the total reactor power is 100 thermal Mw. Whenever the power
density on the blanket side of the core-tank wall was greater than that on
the core side (owing to a fuel coneentration which was higher in the blanket
than in the core), the greater value was plotted. This situation is indicated
by the dashed lines in Fig. 2-7.

Typical information obtained for these slurry reactors is given In
Table 2-5 for two of the cases considered. The values of breeding ratio for

 

TABLE 2-5
SLURRY-REACTOR CHARACTERISTICS

Pressure-vessel inside diameter, ft 10 10
Core inside diameter, ft 7 7
Core thorium concentration, g/liter 100 200
Blanket thorium concentration, g/liter 1000 2000
Blanket U233 concentration, g/liter 3 5
Total power, Mw 100 100
Blanket power, Mw 8.5 4.2
Critical core concentration, g U233 /liter 3.0 6.0
Breeding ratio 1.18 1.21
Power density at core center, kw/liter 43 45
Power density at core wall, kw/liter 5.4 4.9
Flux at core center X 10714 5.5 2.9
Neutron absorptions in core wall, 7, 0.7 0.3

 

 

 

 

 

the two-region reactors given may be compared with those in Table 2-6
for a one-region reactor having the same size and same diameter pressure
vessel. These results indicate that for a given size, two-region reactors
have significantly higher breeding ratios than do one-region reactors.
2-3] STEADY-STATE FUEL CONCENTRATIONS 47

TABLE 2-6

BreEDING RATIO IN 10-FT-DIAMETER
ONE-REGION REACTORS (ZERO POISONS)

 

 

 

Thorium concentration, : :
: Breeding ratio
g/liter
100 0.875
200 0.993
300 1.037

 

 

 

 

For all reactors having thorium concentrations of at least 100 g/liter
in the core and 500 g/liter in the blanket, the neutron absorption by the
core-tank wall was less than 19, of the total absorptions. The variations
of breeding ratio and wall power density with core size, thorium concentra-
tion, and blanket U233 concentration are plotted in Figs. 2-7 and 2-8 for
reactors of 10-ft over-all diameter. The variations of critical concentration,
fraction of total power generated in the blanket region, and the ratio of
fuel to thorium required in the core for criticality are given in Fig. 2-9 for
different thorium concentrations and blanket fuel concentrations. The
curves for the ratio of U233/ Th versus core thorium concentration are of
value in determining the reactivity which would occur if there were a rapid
change in core thorium concentration. If the reactor operating conditions

 

 

 

 

 

 

A | | | | Th in Core,
% Thorium in Blanket, g/liter 5504 g/liter
o N\ 300
E12 [ ~
'§ 1000 200
& 500

1.1 - 100

 

 

 

1.0
%‘ 10
® 300
o5
v = 200
330 100
o. X
<
3

 

0 ] 2 3 4 5
U233 in Blanket, g/liter

Fic. 2-8. Effect of blanket U233 concentration on breeding ratio and wall power
density of two-region slurry reactors. 7= 2.25, total reactor power =100 Mw
(heat), pressure vessel=10 ft ID, core diameter=7 ft, poison fraction= 0,

temperature = 280°C.
48 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

were such that the flat region of the appropriate curve applied, small
uniform changes in core thorium concentration would have a negligible
effect upon reactor criticality. The location of the minimum in these
curves did not vary appreciably with changes in blanket thickness, blanket
thorlum concentration, and blanket fuel concentration. The results given
in Figs. 2-7 through 2-9 indicate that for large spherical reactors the
breeding ratio increases when the core size decreases, when the blanket 1233
concentration decreases, and when the thorium concentration is increased
in either the core or blanket region. The core-wall power density decreases
when the thorium concentration is increased in the blanket region and
when the core size increases, but is relatively insensitive to changes in the
blanket fuel concentration for thorium concentrations greater than 100 and
500 g/liter in the core and blanket, respectively. The critical concentration
of U?33 in the core decreases with decreasing core thorium concentration
and with increasing core diameter, and varies only slightly with changes
in blanket thickness, blanket U233 concentration, and blanket thorium

 

 

 

 

 

 

 

 

 

 

 

15 -
l l | ! l l |
= Thorium in Blanket, g/liter 8 ft Core Dia.
< :\"—- 2000 } 7 ft Core Dia.—
s & 6 ft Core Dia.
)
5 -
o
I
0.044 |
6 ft Core Dia.
g
6 —
c 7 ft Core Dia.
™ 8 ft Core Dia.
3s o
>
020 | | | L | |
0.4
| | | l | l |
0.3 —
§ .
o
23
o | 0.2 |— —
T|- _
K 500
g — 500 6 ft Core Dia.
0.1 _
\ ;ggg 7 ft Core Dia.
| 2000 l ———t——x—i /8 ft Core Dia.

 

0 100 200 300

Thorium in Core, g/liter

Fia. 2-9. Effect of core thorium concentration on U233 critical concentration,
ratio of U233 to Th required for criticality, and fraction of total power generated in
blanket for some two-region slurry reactors. 923 = 2.25, pressure vessel = 10 ft
ID, poison fraction =0, U233 in blanket = 3 g/liter, temperature = 280° C.
2-3] STEADY-STATE FUEL CONCENTRATIONS 49

concentration. The fraction of total power generated in the blanket increases
nearly linearly with increasing blanket U233 concentration, increases with
decreasing core diameter, and also increases with decreasing core and
blanket thorium concentrations. Finally, the U?233-to-thorium ratio re-
quired in the core for criticality passes through a minimum when the core
thorium concentration is permitted to vary. These variations show that
desirable features are always accompanied by some undesirable ones. For
example, increasing either the core or blanket thorium concentrations
results in an increase in breeding ratio, but there is also an accompanying
increase in inventory requirements; decreasing the core radius increases
breeding ratio and possibly decreases inventory requirements but increases
wall power density. These types of variations illustrate that minimum
fuel costs will result only by compromise between various reactor features.

70

 

 

 

 

 

 

 

 

60f—W = 3 —
@ R == Core Tank Radius
6
v _ R=2#
-
2_5 40— o
a=
c 3 30 —
2-—;‘ W =
% 8 20—W = —
= w =
10— —
0
1.25
1.20 — —
2
T
“ 105 ]
o
£
1 —
gno—\x = —
5 ——
W =
1.05— —
W =
1.0
0 10 20 30 40 50

Reactor Length, ft

Fig. 2-10. Gross breeding ratio and maximum power density at core wall for
two-region cylindrical reactors. W =g of U?33/kg of Th?32 in blanket, blanket
radius = 5 ft, power = 450 Mw (heat).

The breeding ratio, the power density at the core wall for a given total
power, and the required fuel concentration have also been evaluated for
cylindrical reactors [25]. The results, plotted in Fig. 2-10, are based on
two-group calculations for cylindrical reactors; the diameter of the pressure
vessel was assumed to be 10 ft, the total reactor power 450 Mw, the reactor
50 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

temperature 300°C, and the core diameter either 4 or 6 ft. The Zircaloy—2
core tank was assumed to be 3-in. thick when the core diameter was 4 ft,
and $ in. when the core diameter was 6 ft. The reactors were assumed to
contain D20, U233, and 69, fraction poisons in the core region and D0,
U233, and 1000 g Th/liter in the blanket region. Figure 2-10 gives the
breeding ratio and the maximum power density at the core wall (core side)
as a function of reactor length for different core radii and blanket fuel con-
centrations. The highest breeding ratios are associated with small core
radi, thick blankets, and long reactors; however, these reactors also have
relatively high power densities at the core wall. Increasing the reactor
length increases the breeding ratio and decreases the wall power density
and the critical fuel concentration but appreciably increases the inventory
of material.

In other studies [26] of cylindrical reactors, results were obtained which
indicated that breeding ratios for cylindrical reactors of interest were about
the same as those obtained for spherical reactors. The required fuel con-
centrations were higher, as expected, so that the average flux was lower
for the cylindrical geometry. Although the maximum core-wall power
density decreased with increasing cylinder diameter, it was always higher
than the wall power density obtained for the spherical reactors of equal
volume. The results of these calculations are given in Table 2-7. These
reactors were assumed to be at 280°C with 79}, core poisons, 3 g of U233
per liter and 1000 g of Th per liter (as ThO2) in heavy water in the blanket,
and operated at a total power of 60 Mw. A cylindrical core was assumed
to be positioned within a cylindrical pressure vessel such that a 2-ft blanket,
thickness surrounded the core.

The results given in Table 2-7 show that increasing the reactor height
had only a slight effect on breeding ratio; also, although the critical con-
centration declined with increasing height, the corresponding total fuel
inventory increased. While not shown, the ratio of blanket power to core
power did not vary significantly with reactor height. Increasing the core
volume caused a pronounced decrease in core-wall power density and an
increase in fuel inventory. Thus, for a 3-ft-diameter core, increasing the
core length from 4.8 to 8 ft decreased the power density by 36%. However,
the blanket and core fuel inventory increased by about 409,

2-3.3 Nuclear characteristics of two-region thorium breeder reactors
under equilibrium conditions. Results [27] of some nuclear computations
assoclated with the conceptual design of HRE-3 are given below for
spherical two-region reactors in which the following conditions were
specified:

(1) The reactor system is at equilibrium with regard to nuclei concen-
trations.
STEADY-STATE FUEL CONCENTRATIONS 51

2-3]

*(389Y) ML 09 JO Iomod 10308l [8)0} UO Paseq ‘Asusp 19mod [[BM-0I0)),

 

 

 

 

 

 

 

 

 

 

8 ol 1 G'¢ el el 9°€ I3 er'l 8¢ Gl
01 ¢l'1 9°¢ i | el'1 L€ 9¢ €r'l 09 01
4! ol'1 LG 81 el 6°€ I€ el 9 8
9¢ €l 1 G'9 89
91 ol'1 0°¢ ¥C ol'1 ¢V 09
63 oIl 9y 8V
G2 0TI I'¥ G'¢€
.m@u:\ M O1)BI IO /3 .HWp:\ | 01381 191 /3 .HMB:\ | o1}8l IO /3
« K1SUSD SUhooL “ « AJISUSP S . « KJISUSP g o
10MO] [poaly ggell 10MO0J Tpootg ggal] 10MO] utpoary gez(l “
15 ‘9109
gg e %z 3O AIOH

 

 

 

 

1} ‘19%9UWIBIP 9I0))

 

 

.—¢ TI9V],

SLHOIAJ] SAOIYVA J0 SHOILOVEY TVOIMANITA)) ¥Od NOILVIINAONO))
cez (] TVOLLIE)) ANV ‘ALISNI(] dEMOJ TIVA\-E¥O)) ‘SOILVY DNIQITUY

 
52 ONE- AND TWO-REGION HOMOGENEQUS REACTORS [cHAP. 2
(2) Hydroclone separation of poisons from the core system is employed
in addition to Thorex processing.
(3) The core Thorex cycle time is a dependent function of the specified
total poison fraction; the blanket cycle time is a function of the blanket

U233 concentration.
TABLE 2-8

CHARACTERISTICS OF INTERMEDIATE-ScALE (HRE-3)
Two-ReEGION REAcTORs (EQUILIBRIUM CONDITIONS)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Core diameter, ft 4 4 4 4 5
Pressure vessel ID, ft 8 8 9 9 9
Blanket thorium, g/liter 500 1000 500 1000 . |1000
Blanket U233, g/kg Th 3.0 3.0 3.0 3.0 3.0
Core poison fraction 0.07 0.07 0.07 0.07 0.07
Concentration of U233
g/liter (core) 3.68 4.04 3.63 4.02 2.19
Concentration of U235,
g/liter (core) 0.41 0.39 0.37 0.37 0.21
Breeding ratio 1.041 1.094| 1.086 1.123 1.089
Core-wall power density
(inside), kw/liter 27 23 27 23 10
Core cycle time, days 833 901 817 893 616
Blanket cycle time, days 220 371 288 504 486
Core power, Mw (heat) 51.5 51.9 51.0 51.7 51.2
Blanket power, Mw (heat) 10.0 9.6 10.5 9.9 10.3
Neutron absorptions and leakages per 100 absorptions in fuel
Absorptions in core by:
U233 74.7 76.4 74.6 76.4 75.3
U234 9.2 8.2 8.4 7.8 8.2
U235 9.1 8.1 8.3 7.6 8.0
U236 0.7 0.4 0.4 0.3 0.4
Poisons 5.2 5.3 5.2 5.3 9.2
Heavy water 0.9 0.8 0.9 0.8 1.6
Absorptions in core tank 2.1 1.6 2.1 1.6 2.4
Absorptions in blanket by: |
U233 16.2 15.5 17.1 16.0 16.6
U234 235 [J236 0.1 0.1 0.1 0.1 0.1
Th 95.8 101.7 [100.9 104.8 101.0
Pa233 0.9 0.5 0.7 0.4 0.4
Poisons 0.4 0.4 0.4 0.4 0.4
Heavy water 0.5 0.2 0.5 0.2 0.2
Fast leakage 4.3 3.6 2.2 1.6 3.0
Slow leakage 3.4 0.9 1.7 0.4 0.8

 

 
2-3] STEADY-STATE FUEL CONCENTRATIONS 53

(4) The poison fraction due to samarium is 0.89%; that due to xenon
is 19}, [poison fraction is the ratio of Z,(poison)/2 s(fuel)].

(5) The external core system has 1.0 liter of volume for every 20 kw of
core power; the blanket external system has 1.0 liter for every 14 kw of
blanket power.

(6) The total reactor power is 61.5 thermal Mw.

(7) The average core and blanket temperatures are 280°C.

The breeding ratio for the reactor variables considered is plotted in Fig. 2-11
as a function of pressure-vessel size for 4- and 5-ft-diameter cores with
several blanket thorium concentrations. More extensive results, including
neutron balances, are given in Table 2-8 for selected reactors. The neutron

 

 
 
   

 

 

 

1.14 l I
Thorium in Blanket,
g/liter
/ /
/
/
1.10 - ]
//
o :
S1.08 / / _
g
£
o
® 106 —
0
1.04 / = 4-ft -Diameter Core o
— w— 5_ft -Diameter Core
1.02 — —
1.00 | | |
8 9 10 11

Inside Diameter of Pressure Vessel, ft

Fic. 2-11. Breeding ratio as function of pressure vessel size for various core
diameters and blanket thorium concentrations. Core poison fraction = 0.07, cor-
rosion products =0, copper concentration =0, blanket U233 =3.0 g/kg of Th,
7?3 = 2.25, mean temperature = 280°C, equilibrium isotope concentrations.

balances are normalized to 100 absorptions in U233 and U?3%; therefore the
numerical values represent approximately the percentage effects of the
various items on the breeding ratio (however, the effect of Pa?33 losses on
breeding ratio would be obtained by doubling the values given).

Some of the materials which act as neutron poisons can be altered by
reactor-system design; these include fission-product poisons, core-tank
material, contaminants such as H20, and additives such as the cupric
ion. The effect of these on breeding ratio is discussed below.

Fission-product poisons. The effect of total core poison fraction, fp
(ratio of absorption cross section of poisons to fission cross section of fuel),
o4 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2
1.14

 

1.12

1.10

o
©

—h
o
o

 

BR, Breeding Ratio

2
"
|

1.02 — —

 

 

00 | | | 1 | l | | |
0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12

 

fpc, Total Core Poison Fraction

Fic. 2-12. Effect of core poison fraction on breeding ratio. Core diameter = 4 ft,
blanket U233 = 3.0 g/kg of Th, 23 = 2.25, mean temperature = 280°C, equilibrium
isotope concentrations.
on breeding ratio is shown in Fig. 2-12. The results indicate that for these
reactors the change in breeding ratio with change in poison fraction can
be estimated from the relation -

ABR = —0.75 Afpe. (2—-18)

Achievement of a xenon poison fraction of 0.01, as postulated in the
computations for Fig. 2-11 and Table 2-8, requires the removal of most of
the xenon or its iodine precursor before neutron capture occurs. If there
is no fast-cycle system for iodine or xenon removal, the poison fraction
resulting from xenon will be about 0.05. According to Eq. (2-18), the
breeding ratio would be reduced by about 0.03 below the values given in
Fig. 2-11 if all xenon were retained in the core system.

Core-tank absorptions. The thickness of the Zircaloy core tank was taken
as 0.42 in. for the 5-ft core and 0.33 in. for the 4-ft vessel. As shown in
Table 2-8, neutron captures in the core tank reduce the breeding ratio
about 0.02 in the 4-ft core and 0.03 in the 5-ft core. If the core-tank thick-
ness were altered, the losses would be changed proportionately.

Absorptions in copper. Copper can be added to act as a recombination
catalyst for decomposed water. No allowance for neutron absorptions in
the copper recombination catalyst was made in these computations. The
poison fraction attributable to copper in various concentrations and the
effects on breeding ratio (ABR) are estimated in Table 2-9. Ifor other
copper concentrations the poisoning effects would be proportionate to the
values in Table 2-9.
2-3] STEADY-STATE FUEL CONCENTRATIONS 55

TABLE 2-9

ErrEcT oF CoPPER ADDITION ON BRrREEDING RATIO

 

 

 

Core Copper Poison ABR
diameter, concentration, fraction
ft g-mole/liter
4 0.01 0.004 — 0.003
5 0.01 0.008 — 0.006

 

 

 

 

 

 

The copper concentration required for 1009, recombination in a 4-ft
core at 61.5 Mw has been estimated to be 0.018 g-mole/liter. For this core
size and copper concentration, the loss of neutrons to copper would reduce
the breeding ratio by 0.005.

H 20 contamination. Any H20 contained in the heavy-water moderator
will act as a poison and reduce the breeding ratio. The above results are
based on the use of heavy water containing 0.259;, H2O. Neutron captures
in the moderator in a 4-ft-core reactor (see Table 2-8) were found to be
about 0.009 per absorption in fuel, of which about 609, were in Hs. Thus
0.259, H20, which is 9 liters in a 3600-liter system, reduced the breeding
ratio by 0.005. Other values are given in Table 2-10. Different concentra-
tions of H2O would cause changes in breeding ratio proportionate to the
values in Table 2-10.

TABLE 2-10

ErreEcT oF HoO CONCENTRATION ON
BreEEDING RATIO

 

 

 

 

 

Core diameter, ft H20 concentration ABR
4 1.09, (36 liters) —0.02
5 1.09, (44 liters) —0.04

 

 

 

A specified volume of H2O added to the blanket has much less effect on
breeding ratio than the same amount added to the core. This is a result of
both the lower flux in the blanket region and the larger volume of the
system. |

Corrosion products. Assuming the surface area of stainless steel in the
core high-pressure system to be 6000 ft2, corrosion to an average depth of
0.001 in. would remove 250 1b of metal. If this were distributed uniformly
throughout the fuel solution, the poison fraction resulting from it would be
956 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

about 0.18 in a 4-ft core and 0.36 in a 5-ft core. However, iron and chro-
mium would precipitate and be removed by hydroclones. If the hydro-
clones were operated on a fast cycle time (several days), neutron capture
in iron and chromium would be unimportant. Nevertheless, the absorption
cross section of the nickel and manganese (which probably remain in solu-
tion) amount to about 159, and 99, respectively, of the total absorption
cross section of type—347 stainless steel. Thus, from 0.001 in. of corrosion,
the nickel and manganese would yield a poison fraction of about 0.04
(ABR = 0.03) in a 4-ft core, and 0.08 (ABR = 0.06) in a 5-ft core.

The actual value of the poison fraction from corrosion products would
depend on the corrosion rate and the chemical processing rate. If the cor-
rosion products are assumed to change to isotopes of the same cross sec-
tion upon neutron capture, the following relations are obtained under
equilibrium conditions: |

f»=10.04 X R X (T,/365) (Core ID = 4 ft), (2-19)
f»=0.08X R X (T./365) (Core ID = 5 ft), (2-20)

where f, is the equilibrium core poison fraction from corrosion products, i
the mean corrosion rate in mils/yr, and T, the core cycle time in days.
An additional point of concern resulting from corrosion of stainless steel
is the adverse effect of high corrosion-product concentrations on the
stability of fuel solution. The concentration of nickel resulting from
0.001 in. of corrosion would be 0.052 g-mole/liter, and that of manganese
would be 0.011 g-mole/liter. Unless adjustments were made to the acid
concentration, the fuel solution would probably form a second phase
before the above concentration of nickel was attained. |
The corrosion products from the Zircaloy—2 core vessel would not appre-
ciably affect the breeding ratio, even if they remained in suspension.
Owing to the dilution effect associated with the large external volume, cor-
rosion of the core tank would result in a slight increase in the breeding ratio.

2-3.4 Equilibrium results for two-region uranium-plutonium reactors.
Initial reactor-fuel materials which have been considered [28] in uranium-
plutonium systems are UO2S04-D20, UO2(NO3)2—D20, and UO3-D20.
Of these, the system which gives the highest conversion ratio is the one
containing UO3-D20O. However, because of the relatively low values for
7n(U235) and n(Pu?3?), it is presently considered that the attainment of a
conversion ratio as great as unity under equilibrium conditions is imprac-
tical because of the high fuel-processing rates and the large reactor sizes
that would be required. However, many uranium-plutonium reactor
systems which will operate on either natural-uranium feed or on fuel of
lower enrichment than natural uranium appear feasible. A two-region
reactor can be operated by feeding natural uranium (or uranium of lower
2-3] STEADY-STATE FUEL CONCENTRATIONS 57

enrichment in U23%) into the blanket region, and plutonium (obtained by
processing the blanket) into the core region.

The reactor system considered here is one containing UOgs, PuO2, and
D:0O; steady-state concentrations of U235 U236 238 Pu239 Pu?t0 and
Pu?#! are considered. Fuel is removed and processed at a rate required to
maintain a specified poison level. The reactor consists of a core region in
which plutonium is burned and of a blanket region containing uranium
and plutonium. Under equilibrium conditions the net rate of production
of plutonium in the blanket is equal to the plutonium consumption in the
core. In Table 2-11 are given [29] some of the nuclear characteristics for

TABLE 2-11

Dara ror Two-RecioN, UO3—PuO2—D2O REAcTORS OPERATING AT 250° C,
HaviNg A CorE DIAMETER OF 6 FT, A BLANKET THICKNESS OF 3 FT, AND
VARIABLE BLANKET-FUEL ENRICHMENT

 

 

 

 

 

U235/1J238 in blanket 0.0026 0.0035 0.0040
Blanket U conc., g/liter 500 500 200
Pu239/U238 in blanket 0.0010 0.0018 0.0022
Pu240/0U238 in blanket 0.00013 0.00041 0.00060
Pu241 /U238 in blanket 0.00002 0.00007 0.00011
Feed enrichment, U23%/U 0.0031 0.0047 0.0058
Blanket power, Mw 247 411 519

Core power, Mw 320 320 320

Core Pu conc., g/liter 1.66 1.48 1.40
Pu?40/Pu249 in core 0.99 0.99 0.99
Pu?41/Pu?49 in core 0.35 0.35 0.35
Fraction of fissions in U235 0.25 0.28 0.30
Fraction of U consumed 0.017 0.016 0.015
Total power, Mw (heat) 567 731 839

 

 

two-region, UO3—PuO2—D20 reactors having a core diameter of 6 ft and an
over-all diameter of 12 ft, and having various U235/U?238 ratios in the blanket
region.

2-3.5 One-region reactors. Single-region reactors have simpler designs
than two-region reactors by virtue of having only a single fuel region; also,
fuel processing costs for one-region reactors are generally lower than for
two-region systems. However, to attain breeding or conversion ratios
comparable to those in a 10-ft-diameter two-region reactor, the diameter
of a one-region reactor has to be about 15 ft or greater. The construction
58 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

of pressure vessels of such diameters is difficult, and relatively little ex-
perience on such construction has been obtained to date.
In the succeeding sections some equilibrium results are given for the
nuclear characteristics of one-region breeder and converter reactors.
1.2

 

I I r I l

—

Diameter
(ft)

1.1 — — ] —_

18

16

14

12 ]
10

0.9

0.8

0.7

Net Breeding Ratio
\\
|

 

 

o | | | | | |
5 100 150 200 250 300 350 400
Thorium Concentration (g/liter)

 

o

F16. 2-13. Breeding ratio vs. thorium concentration for one-region reactors of
various diameters. Poison fraction = 0.08, 523 = 2.25.

2-3.6 Equilibrium results for one-region thorium breeder reactors. Re-
sults have been obtained [30] for one-region thorium breeder reactors
operating under equilibrium conditions. Critical concentrations and
breeding ratios were obtained by means of Eqgs. (2-4) and (2-6). Fig-
ure 2-13 gives the breeding ratio as a function of thorium concentration
and reactor diameter. Comparison of these results with those obtained for
two-region reactors illustrates that reactor diameter influences breeding
ratio to a greater extent in one-region systems than in two-region systems.
Also, increasing the reactor diameter increases the breeding ratio signif-
icantly even for 14-ft-diameter reactors. Although breeding ratio can be
increased by increasing the thorium concentration, there is an accompany-
ing increase in fuel inventory. To keep inventory charges at a reasonably
low level and yet permit a breeding ratio of unity to be attained requires
thorium concentrations between 200 and 300 g/liter and reactor diameters
of about 14 ft.

The equilibrium isotope concentrations as a function of thorium concen-
tration for a 14-ft-diameter reactor are given in Fig. 2-14. This diameter
2-4] UNSTEADY-STATE FUEL CONCENTRATIONS 59

 

    

y235

   

Concentration (g/liter)
O

 

 

| | ' '
o 1 1 ]
50 100 150 200 250 300 350
Thorium Concentration (g/liter)

 

Fig. 2-14. Uranium isotope concentrations under equilibrium conditions vs.
thorium concentration in a one-region reactor of 14-ft diameter.

value has been chosen because it represents the diameter which gives
minimum fuel costs, although it is realized that construction of the cor-
responding vessel may be beyond present technology. The fuel-processing
cycle time which minimized fuel cost was that which corresponded to a
poison fraction (due to fission products) of about 0.08, and so this is the
value used for poison fraction in the results given here. Absorptions In
higher isotopes contributed an additional poison fraction of about 0.03.

2-3.7 Equilibrium results for one-region uranium-plutonium reactors.
The fertile-material concentrations and reactor diameters have essentially
the same effects on conversion ratio and fuel inventory for one-region
uranium-plutonium systems as they do for thorium breeder systems;
however, since the n’s for U235 and Pu?39 are lower than for U233, it is more
difficult to attain a conversion ratio of unity in U-Pu systems than it is in
U233—Th systems. It is still possible, though, for UO3-PuO2-D20 systems
to operate on natural-uranium feed, as evidenced by the results for two-
region reactors. Minimum fuel costs (based on n*l = 1.9; a more accurate
value is now believed to be 2.2) for one-region reactors, however, occur
when the uranium feed is slightly enriched in U235 [32]. Table 2-12 gives
results [32] of some nuclear calculations for these one-region systems
operating under equilibrium conditions. The reactor diameter was taken
to be 15 ft; the fuel-processing rate was such as to maintain a poison frac-
tion of 79, in the reactor core. Fuel feed was considered to be obtained
from an isotope-enrichment diffusion plant. The results indicate that the
uranium feed for these reactors would have to contain between 1 and
1.59, U233,

9—4. UNSTEADY-STATE FUEL CONCENTRATIONS AND BREEDING RATIOS

2-4.1 Two-region reactors. During the period following reactor startup,
there is a buildup of fission-product poisons and higher isotopes with time,
which results in varying nuclear characteristics. This section presents some
calculations relative to the HRE-3 conceptual design for the initial period
of reactor operation.
60 ONE- AND TWO-REGION HOMOGENEOUS REACTORS

TABLE 2-12

ReEacTOrR CHARACTERISTICS FOR SOME ONE-REGION,

UO03-Pu0s—D>0 ReacTors* OPERATING UNDER
EquiLiBriuM CONDITIONS

[cHAP. 2

 

Reactor temperature, °C 250 250 250 300
U conc., g/liter 334 253 170 183
U235 conc., g/liter 2.65 1.47 0.83 1.35
U236 cone., g/liter 0.39 0.22 0.12 0.20
Pu23? conc., g/liter 3.43 1.74 0.84 1.08
Pu240 conc., g/liter 3.39 1.72 0.83 1.17
Pu24! conc., g/liter 1.21 0.61 0.30 0.41
Initial enrichment (no Pu), |

U235/U (total) 0.0116 0.0106 0.0098 0.0140
Steady-state enrichment, |

U235/U (total) 0.0081 0.0059 0.0050 0.0073
Steady-state feed enrichment,

U235/U (total) 0.0153 0.0113 0.0096 0.0136
Fraction of fissions in U233 0.24 0.25 0.29 0.31
Fraction of U consumed 0.018 0.017 0.015 0.014

 

 

 

 

 

 

 

*Reactor diameter = 15 ft; poison fraction = 79}; reactor fuel returned to diffu-
sion plant for re-enrichment; tails from diffusion plant are assumed to have U235
content of U235/U = (0.0025; processing losses are neglected.

Computations have been performed for several spherical reactors using
an Oracle code [31] for two-region, time-dependent, thorium breeder
systems. The variation with time of the breeding ratio and the concentra-
tions of U233 U234 235 236 Pa233 and fission-product poisons were
obtained. Calculations were first confined [32] to solution-core reactors
initially containing either U233 or U235 and generating a core power of
50 Mw. Core and blanket Thorex processing was considered only when the
initial fuel was U233, The use of centrifugal separation (hydroclones) for
core-solution processing was assumed in all cases. The time dependence of
the concentrations of xenon, the samarium group of poisons, and the
poisons removable by hydroclone processing were neglected, since the time
required for these poisons to reach near-equilibrium conditions is relatively
short. Account has been taken of their presence by the use of fixed poison
fractions. The effect of copper added for internal gas recombination was
also included in this way. Core diameters of 4 and 5 ft, pressure-vessel
diameters of 8 and 9 ft and thorium blanket concentrations of 500 and
1000 g/liter were considered. Fixed core poisons in terms of percentage of
2-4] UNSTEADY-STATE FUEL CONCENTRATIONS - 61

core fission cross section were: samarium group, 0.89,; xenon group, 1%,;
copper, 0.89; poisons removable by hydroclones, 19;. Fixed blanket,
poisons were: samarium group, 0.8%; xenon, 19,. It was assumed that
the core solution was processed both by hydroclones and by the Thorex
process described in Chapter 6; the blanket slurry of thorium oxide in
heavy water was processed by Thorex only, and fuel produced in the
blanket was drawn off from the Thorex plant and returned to the core at a
rate sufficient to maintain criticality.

The system was assumed to start “‘clean,” except for the poisons men-
tioned, with either U233 or U233 in the core. Makeup fuel (same as initial
fuel) was fed as needed while the concentration of fuel in the blanket was
increasing. When the blanket fuel concentration reached a predetermined
level, blanket processing was initiated at the rate which would be required
if the reactor were at equilibrium; however, for the U?35-fueled reactors,
calculations were performed only up to the time at which processing
would start. At the start of processing, the fuel feed for the core was
assumed to come from the processed blanket stream. When the core
poison level built up to a predetermined point (89, for the U233 reactors),
processing of the core solution was started. The processed core-fuel stream
was considered mixed with the processed blanket stream; part of the mix-
ture was used as core feed while the excess was drawn off as excess fuel.
The calculations were continued until most of the concentrations ap-
proached equilibrium values. Time lags due to chemical processing holdups
were neglected. The chemical processing rates employed were those cal-
culated earlier for equilibrium reactors [27].

The curves in Fig. 2-15 show results for some representative U233-fueled
reactors. As shown in Fig. 2-15(a), with 1000 g Th/liter in the blanket
region, the breeding ratio falls steadily for about 900 days until core proc-
essing starts. Although blanket processing, begun at 490 days, arrests the
growth of U233 in the blanket rather suddenly, the slope of the breeding-
ratio curve does not change markedly because the buildup of core poisons
is controlling the breeding ratio. When core processing interrupts the
growth of core poisons, the breeding ratio levels off sharply. The variation
of relative leakage and poison losses with time and the variation of blanket
power with time are illustrated in Fig. 2-15(b) for the case of a blanket
thorium concentration of 500 g/liter.

In all cases the time required to reach the 89, core poison level was of
the order of two years. The average breeding ratio during this period was
about 0.02 higher than the equilibrium value. The effect of poisoning due
to buildup of corrosion products is not included in this estimate.

While the above statements concerning breeding ratio are characteristic
of U233-fueled reactors, U235 reactors show quite a different variation of
breeding ratio with time, due to changes in the effective n for the system.
62 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

 

 

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

116 T l I [ o 3 08 T Tppe—
: : >3 : I Core Poison
215\ | 1 2% 06 | | Fraction
S : : —-Q Neutron
| £ o | Leakage
114 \ 4 22 .04 : 2 _
o , ' 86 l
£ | ! “E L :
@ 1.13}— : — £ 5 .02 I Blank?t Poison Fraction —
{ 2L
| l
1.12 l I l S— I n? ndh-: 0 ] l ] l I l
s [*=—Core Processing Starts 15 [«—Blanket Proc. Starts
T | | I I
—_ '<-B|a'nkef Processing Starts | |<—Core Proc. Starts
o ‘e - — 312 —
2 l | Core U233 23 I
6 ! I 23 = Blanket Power at — — —— —
0 3 I ; Blanket U233 - L .
s °I — g9 Equilibrium ]
£ P 2
c ! o |
§ 2~/ ! | - < 6 | —
5 | | X | |
| s |
O | I K] |
1 | — 0 3 : : —_
I 233
. . . Pa | |
= ! + | o l |
0 800 1600 2400 3200 4000 0 500 1000 1500 2000 2500
Time (days) Time (days)
(a) Breeding Ratio, Blanket Conc. of U233 (b) Neutron Leakage, Poison
and Pa233 and Core Conc. of U233 Fraction and Blanket
(Blanket Conc.=1000g Th/1). Power (Blanket Conc.=
500g Th/1).

Fig. 2-15. Nuclear characteristics of a 60-Mw (heat) two-region U233 breeder
during initial operating period. Core diameter =4 ft, pressure vessel diameter
= 9 ft, 280°C, solution core.

Such reactors have appreciably lower breeding ratios than do U233-fueled
systems. This is primarily due to the relatively low value of 2% compared
to 7?3 (9?3 = 2.25 while 725=2.08). As U233 builds up in the blanket of
a U235fueled reactor, the average value of n for the reactor as a whole
increases. This helps compensate for the increase in core poison fraction
and causes the breeding ratio, which initially drops from 1.008 to 1.002
during the first 200 days of operation, to rise again. A maximum of 1.005
in the breeding ratio is reached at about 600 days, after which a slow
decrease follows.

Three types of isotope growth were obtained in this study. First, the
concentration of the main isotope, U233 or U235 remained relatively con-
stant despite sizeable changes in other concentrations. Second, the pro-
tactinium concentration-time curve was found to ‘“‘knee-over’’ even before
processing started. This behavior was due to the radioactive decay rate
being several times larger than the chemical processing rates employed.
Third, the concentration of the heavier isotopes of uranium built up slowly
with time for the assumed power level; even after 10 years’ operation their
concentrations were much less than the equilibrium values. The total core
concentration of uranium was thus significantly less than the equilibrium
value. The isotope concentrations after various operating times are shown
in Table 2-13, along with equilibrium values.
2-4] UNSTEADY-STATE FUEL CONCENTRATIONS 63

TABLE 2-13

CorRE CONCENTRATION OF URANIUM ISOTOPES
AS A FunctmioN oF TiMeE For A U233_-FuerLeEp REACTOR
AND FOR A UZ235-FueLEDp REACTOR

(Core diameter =4 ft; pressure-vessel diameter =9 ft; ThO2 concentration in

blanket = 1000 g/liter; solution core, total power = 60 Mw of heat)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time, days U233 U234 U235 U236 U total

Concentrations of isotopes for U233 fuel, g/liter

| 0 3.75 — — — 3.75

200 3.77 0.33 0.02. — 4.12

500 3.74 0.74 0.07 0.01 4.56

920 3.91 1.29 0.14 0.04 5.38

2000 3.98 2.06 0.26 0.18 6.48

3000 4.02 2.40 0.31 0.34 7.07

Equilibrium 4.02 2.74 0.37 1.57 8.07
Concentrations of isotopes for U235 fuel, g/liter

0 — — 4.21 4.21

200 — — 4.19 0.57 4.76

500 — — 4.12 1.43 5.55

900 — — 4.03 2.54 6.57

 

 

 

 

 

 

 

 

While chemical processing sharply discontinues the growth of U233 in
the blanket, the power level in the blanket may markedly overshoot the
equilibrium level, as shown in Fig. 2-15(b). An overshoot is obtained when
the blanket-U233 concentration reaches its maximum value before core
processing starts. Table 2-14 shows the peak values of blanket power
computed for the reactors studied, along with the equilibrium values.

Results similar to those given above have also been obtained [33] for
two-region breeders having various concentrations of thorium in the core.
The core diameter was set at 4 ft, the pressure-vessel diameter at 9 ft, and
the blanket thorium concentration at 1000 g Th/liter. The core thorium
concentrations studied were 100, 150, and 200 g Th/liter. As before, the
moderator was heavy water in both core and blanket volumes, and both
regions operated at a mean temperature of 280°C; the Zircaloy—2 core tank
was 0.33 in. thick. Calculations were performed at a constant total power
of 60 thermal Mw. The same chemical processing conditions were assumed
64 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

TABLE 2-14

Peax AND EQUILIBRIUM VALUES OoF BLANKET POWER
CoMPUTED FOR U233-FurLED REACTORS

(Reactor power = 60 thermal Mw)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  

 

 

 

 

 

 

 

Core | Reactor | Blanket Equilibrium Peak value of | Time at which
dia., dia., Cone., | value of blanket | blanket power peak value
ft ft g Th/liter power, Mw Mw occurs, days
4 9 1000 9.6 10.8 520
4 9 200 10.4 12.5 350
4 8 1000 9.3 10.8 380
o 9 1000 10.0 115 490
14 | T | T m T T T 1
13 |- U233 Core ; ! 7
12 ! | ~
11 i L
3.2 l :
- 28 I
2 I
= 24 |- :
2 20 | |
c
2 16| '
2
T 12} |
g ]
c 08 |- 233
P Blanket
S 04 I e i an el Pg233 Core
' - = ; 3 —_——
0 fi | | U L | | | |
1.22 T T T T T T T 1 T 0.08
I awot
oo g
o 1.20 ovs° l | —H0.06 %
S co® , | 2
f-; | | Core Processing Starts "c'
£1.18 |- | l —0.04 ¢
3 | | 'S
2 | | Blanket Poison Fraction o
o0 i |
1.16 | i —10.02
Blanket Processing SM Breeding Ratio
1.14 | ] | 1 i ] ] I 1 0

 

 

 

0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time (days)

Fie. 2-16. Variation with time of U233 U234 and Pa?233 in the core, U233 and
Pa233 in the blanket, breeding ratio, and core and blanket poison fractions as func-
tions of time for a two-region 60-Mw thorium breeder reactor. Core thorium con-
centration = 150 g Th/liter, blanket thorium concentration = 1000 g Th/liter,
temperature = 280°C.
2-4] UNSTEADY-STATE FUEL CONCENTRATIONS 65

as before, with the important exception that no hydroclone processing was
employed for the slurry-core cases.

Results given in Fig. 2-16 are for a core thorium concentration of
150 g Th/liter and are typical of all cases studied in their important features.
As with solution-core reactors, there is a small but relatively rapid initial
drop in breeding ratio during the first 100 days following reactor startup.
This is due to the neutron captures in protactinium as its concentration
rises and approaches equilibrium conditions. The buildup of protactinium
is quite similar to the behavior seen in the solution-core cases. KEqui-
librium protactinium levels are reached in both core and blanket long
before chemical processing is started. This period is followed by a more
or less linear fall in breeding ratio with time, due to core poison-fraction
buildup. Initiation of blanket processing produces a relatively minor
change in slope of the breeding ratio-time curve, in contrast to the sharp
break caused by the start of core processing. Linear buildups with time
of the poison fractions and the higher isotopes are also observed, with
U234 in the core being the most important higher isotope. Other isotope
concentrations are comparatively low even at 1500 days.

Figure 2-17 lists some results for the various cases considered (except
for the case already considered in Fig. 2-16). Comparison of the results
shows that the time at which core processing is initiated increases with
increasing core thorium concentration (except for the solution-core case);
this is due to the higher U233 critical concentration associated with higher
thorium concentration, resulting in a longer period to build up to a specified
core poison fraction. For the solution-core case, hydroclone processing was
assumed, so that 759, of the poisons were removed by means other than
by Thorex processing of the core (the time specified for core processing to
begin in Fig. 2-17 corresponds to the initiation of Thorex processing).
It is interesting to note that with slurry cores, despite the absence of hydro-
clone separation, the core poison fraction still takes two or more years to
reach a 79, level.

In these calculations there was no allowance for the poisoning effect of
accumulated corrosion products. Absence of hydroclone processing would
result in retention of all corrosion products in the core system, and for
the same corrosion rate the cross section of corrosion products is about
four times as great as in a solution-core system using hydroclones. How-
ever, the higher critical concentrations in slurry-core systems compensate
for this, and the poison fraction in the core with 200 g Th/liter could be
about the same as in a solution-core reactor.

The critical concentration of U233 in the core displays a small but steady
rise; this contrasts with solution-core reactors, where the critical core
concentration is found to be strikingly constant. Solution-core reactors,
because of their longer thermal diffusion length, are more sensitive to the
66 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

 

    

 

 

 

 

   
   
   
  

 

 

 

 

 

 

 

 

116 ] ] | | ¢
(a) Solution Core Reactor (with Hydroclones) —15
| |
1.14 — : — 4
Blanket Proce;sing Starts ' BR
112 — : :Core Processing Starts —1 2
| [ | | L1 | | 1
. 11T ~
118 I (b) Slurry-Core Reactor— | | I 5
1 I ':
BR 100g of Th/liter in Core as ThO 233 —4 10 <
i 2 U o
ol.16 |- M —J9 §
o= t . -
i r'—Core Processing Starts g g
£ ' : Blanket Processing Starts g
T1.14 » —7 5
g a BR v
o0 I I Lt l l | I 6 o
‘ o
1.20 . ' 20 v
| T T T T o
| 19 &
| =
1.18 |- | 18
17
1.16 |- | 16
| | i 15
i Core Processing Starts |
1.14 I—  Blanket Processing Starts | : — 14
-
| | | L l l L 13
0 200 400 600 800 1000 1200 1400 1600

Operating Time at Reactor Power of 600 Mw(Heat) (days)

F1a. 2-17. Comparison of breeding ratios and U233 core concentrations for two-
region breeder reactors with various core-thorium concentrations. Core diameter
=4 ft, reactor diameter=9 ft, 280°C, blanket thorium concentration = 1000
g/liter, blanket processing started when U233 blanket concentration reached
3 g/liter, core processing started when core poison fraction reached 79, (89, for
solution-core case).

growth of fuel in the blanket than slurry-core machines. The latter have
less thermal leakage from core to blanket, and core-poison and higher-
isotope growth are less well compensated by the rise of fuel level in the
blanket.

2—4.2 One-region reactors. The critical equation for one-region reactors
is not so involved as that for two-region systems, and so the mathematical
system analogous to the one used above for two-region systems is com-
paratively simple. However, relatively few nuclear results for one-region,
time-dependent systems have been obtained specifically. Some results
are given in Reference [29] for a 15-ft-diameter reactor containing 200 g of
uranium per liter and operating at a constant power of 1350 Mw (heat)
at 250°C. Concentrations of U235, Pu?39, Pu?4%, and Pu24!, and the fraction
of fissions due to U235 were obtained as functions of time.
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 67

Under certain conditions the mathematical system involving nuclide
concentrations, criticality, and fuel processing can be solved analytically
[34]; the solutions obtained have been used to calculate fuel costs directly,
but they could also be used to calculate breeding ratio and critical fuel
concentrations explicitly.

2-5. SAFETY AND STABILITY OF HOMOGENEOUS REACTORS
FoLLowING REACTIVITY ADDITIONS

Reactor kinetic studies are usually broken down into investigations of
reactor safety and reactor stability. Both safety and stability are deter-
mined by the generalized equations of motion involved. The two categories
are considered, however, because the time scales in safety studies are
usually much shorter than those involved in stability studies; hence, the
generalized equations of motion can be simplified in accordance with the
specified study. As used here, safety refers to the events which happen as
a result of the initial power excursion following a reactivity addition;
stability refers to the events which occur as a result of subsequent power
surges. By these definitions, it is possible that a reactor can safely with-
stand the first power surge following a reactivity addition, but still not
be stable; under these conditions, the reactor would be safe with respect
to the first power surge but would not be safe to subsequent power surges.
On the other hand, a reactor system can be stable and still not be safe;
i.e., permissible changes in reactor variables can be exceeded as a result
of the first power surge, even though the system would have reached a new
equilibrium condition had it been able to withstand the first power surge.

The majority of work on the kinetics of homogeneous reactors has con-
cerned circulating, pressurized-type systems, and the discussion below
pertains primarily to those systems. There has been relatively little effort
devoted to boiling homogeneous reactors; for these reactors the control
and safety problems may be more difficult than for the nonboiling reactors
[35]. In boiling systems, the most important parameters are those asso-
ciated with vapor formation and bubble growth, vapor removal from the
system, and the control system. With regard to control, boiling reactors
have the disadvantage of producing more power when less power is de-
manded unless a control system is used in which an increase in reactor
pressure leads to ejection of fuel from the core region. Alternatively, power-
demand control can be obtained by permitting the boiling process to
generate only a fraction of the total reactor power, the remainder being
removed by fuel fluid which is circulated through a heat exchanger.

2-5.1 Homogeneous reactor safety. The nuclear safety of homogeneous
reactors will be a function of the maximum permissible reactivity addition
68 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

and the possible reactivity additions. Despite the inherent safety associated
with a large negative temperature coefficient of reactivity, it cannot be
stated a priort that homogeneous reactors will be safe under all operating
conditions. The limiting feature with respect to reactivity addition is the
permissible pressure rise within the reactor system. This is usually either
the maximum permissible core pressure rise (two-region reactor) or the
maximum permissible pressure rise within the pressure vessel (one-region
reactor).

With a given limit on the maximum permissible pressure rise, it is possi-
ble to specify the maximum permissible reactivity addition which can be
added to the reactor. Reactor operations can then be restricted by design,
so that reactivity additions associated with physical events will not exceed
the maximum permissible reactivity addition.

The potential reactivity available in homogeneous systems is inherently
large, because a high operating temperature is coupled with a high negative
temperature coefficient of reactivity. However, all reactivity additions in-
volve a time element. Since it appears desirable to allow continuity of
physical operations, the safety design criteria should be applicable to con-
tinuous, linear rates of reactivity addition. Specifically, the maximum
permissible linear rate of reactivity addition should specify what restrictions
are necessary on the physical system so that this rate is not exceeded.

Equations of motion. The neutron density is the fundamental variable
in homogeneous reactor safety and is influenced primarily by the tempera-
ture and density of the moderator and by the operational changes which
effect a reactivity change. So long as the reactor is not far above prompt
critical, the neutron density is given by the conventional equations of
motion. These may also be used when larger reactivity additions are
considered, if the prompt-neutron lifetime is assumed to be that associated
with the region in which the neutron density is rising most rapidly with
time. Under this condition, the over-all rate of increase in neutron density
1s overestimated, so that a safety factor will exist in reactor designs based
upon these equations.

Reactivity additions which involve homogeneous reactor safety are
considerably In excess of that required for prompt criticality, and for these
cases the reactor power reaches a maximum value in times of the order of
tenths of seconds. Such time intervals are short compared with the average
half-life of the delayed-neutron precursors, and so only a small fraction
of the precursors formed during the power rise decay during that time in-
terval. The delayed neutrons from these precursors therefore contribute
little to the reactor power while the power is rising; rather, they are formed
following the time of peak power and exert a powerful damping influence
on the power oscillation, leading to a single, damped power surge. For
these safety calculations, the delayed-neutron density can be considered
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 69

constant, and so the neutron-density equations can be combined into one
equation. In terms of reactor power, this equation is

%9 _ [lce(l —lB) — 1] P g P, | (2-21)

where k. = effective multiplication constant,
P = reactor power,
Po = P evaluated under initial conditions,

{ = time,
B = effective fraction of fission neutrons which are delayed,
[ = average lifetime of prompt neutrons.

The appropriate value for B is determined on the basis of the time spent
inside and outside the reactor core vessel by a fluid particle.

To complete the mathematical system, the relation between k., and P
i1s required, which requires intermediate relations. For aqueous systems
operating above 200°C, k, is influenced primarily by fluid-density effects,
insofar as inherent reactivity changes are concerned. Since reactivity can
also be added by physical operations, k. can be considered to be given by

k= 1+A+bt+%"”§ (b — po), (2-22)

where = linear rate of reactivity addition to reactor, Ak./sec,

b
ok, : : ..
B0 = density coeflicient of reactivity,
A = instantaneous reactivity addition,
p — po = deviation of the average density of fuel fluid from its in-

1tial value.

The core fluid density is determined from the hydrodynamic equations
of continuity and motion, in conjunction with the equation of state for
the fluid. In most studies to date, a one-dimensional flow model is as-
sumed, gas effects are neglected, the core tank is considered to be rigid,
and the core inlet fluid velocity is considered to be constant. The equation
for the change in fluid density is then

d A
or == po(U — V), (2-23)

where A = cross-sectional area of relief pipe,
V.= volume of core region,
U — Uy = deviation of the fluid velocity in core-relief piping from
its initial value.
70 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

Assuming that the resistance to fluid flow between the core and pres-
surizer is proportional to the square of the average fluid velocity, and that
the fluid is incompressible, the one-dimensional hydrodynamic equation
of motion is given by

 

L
2t [ | & = U0 =p—p, =0T U, (22
where ny = resistance coeflicient associated with velocity U,
A (x) = cross-sectional area of relief pipe,
g = dimensional constant,
D. = core pressure,
P, = pressurizer pressure,
U = average velocity of fluid in relief piping of cross-sectional
area A,

L  =length of relief piping from core to pressurizer.

Use of the absolute value of U in Eq. (2-24) ensures that fluid friction
will always oppose fluid flow.

The relation between pressure, temperature, and density of the fluid is
given by the equation of state:

p=22| (7 — To)(- Z2)+ o= pol- (2-25)

where p = pressure rise in core,
T — Ty = deviation of the average core-fluid temperature from its
initial value.

Relations are still needed between T and P and between p, and p.
Assuming isentropic compression within the pressurizer, p, 1s approxi-
mated by

e PO—P Ve _
Pp = Po="po— = (2-26)

where V » = volume of pressurizing fluid,
Pp — Po = deviation of the pressurizer pressure from its initial value,
n = ratio of heat capacity at constant pressure to heat capac-
ity at constant volume for pressurizing fluid.

The relation between T and P is obtained from an energy rate balance on
the core fluid. Since the rate of energy transport associated with fluid
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 71

flow and thermal diffusion is small during times of interest, the energy
rate balance is approximated by

5.5 =P P, (2-27)
where S, = volume heat capacity of the core fluid.

The mathematical system is now complete, and consists of Eqs. (2-21)
through (2-27). Since an analytic solution to the above system has not
been obtained, numerical integration of the above equations is necessary
in general.

Calculation of reactor pressure rise. The variable of interest is the maxi-
mum value of core pressure for a given set of parameter values. Although
not exact, an analytic expression for the maximum pressure rise, Pmax, can
be derived [36] for the case of an instantaneous reactivity addition. This
is given by

=
Praex = zm%f; [o 385m - w(1+ )]+9-21—” (2-28)
_ponV,.(dp
where Ce= 0 7, <dp>

measure of effect of pressurizer volume upon core pressure rise, dimen-
sionless,

% (vs + m)m

F =1+ ‘%‘[C 2+ T ’

__ A — B = instantaneous prompt reactivity addition divided
1 by I, sec™1,

Pmax = Maximum pressure rise, Ib-force/ft2,

" __ny;Uo  normalized friction coefficient in core relief line,
’
/ A j‘L dx sec” L,

 

A(zx)
_ ( )(510) normalized conversion factor between net core
2 op density change and core pressure rise, ft2/sec-1b
force,
_ Uodpo <3k) 2 o _ g.(dp/dp) p) Y
Y3 = Vi ap:sec , w » Sec™ 4.

h Vf e

Results obtained from Eq. (2-28) have compared favorably with numerical
integration of Eqs. (2-21) through (2-27) [36].
72 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

Although Eq. (2-28) is based on an instantaneous reactivity addition,
it can be applied to rate additions of reactivity in certain circumstances.
For rate additions normally encountered in homogeneous systems, it ap-
pears that the most dangerous amounts of reactivity can be added when
the initial reactor power is low. In these circumstances, reactivity is added
during an interval which is much greater than the time interval in which
the reactivity is decreased as a result of fluid density changes. A simplified
mathematical model can then be used [36] to relate a linear rate of re-
activity addition, b, to its equivalent instantaneous reactivity addition,
Akeq. This relationship is given by the following equation:

£ me/2wnp
14+ =2 =—"T» (2-29)
w2, In(m?/2w2))
where w2, = w2z, sec™2,
1/ ok.\ Po
2 ——f_YteyLl0 .—2
w2 l< 8’1’) S sec™?,
T, = reactor power at prompt critical relative to initial power
level,
Me = _lflgql:_@_, sec™ 1

Akeq = equivalent instantaneous Ak, associated with rate of reac-
tivity addition,

¢ =0b/l, sec™2.
The value for x,. is obtained from the equation
- M M
o= M2 4 V1 —= erf (=) 2-30
= VT e () &

where M, = l_ZfllO)__"':_@,

k.(0) = initial value of k..
Equations (2-29) and (2-30) are plotted in Figs. 2-18 and 2-19 and give

the particular combinations of initial power level and rate of reactivity
addition corresponding to a given equivalent prompt reactivity addition.
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 73

70

 

| | I I I | I I I

(sec'l)

Akeqp
l

me=

 

 

 

I I I I | I | L |

0 10 20 30 40 50 60 70 80 90 100

£, % (sec?)

Fic. 2-18. Relation between equivalent prompt reactivity and rate of reactivity
addition. Ak, = equivalent prompt reactivity, { = mean lifetime of prompt neu-
trons, b = rate of reactivity addition, wnzp = modified nuclear frequency.

 

WO ETT T T T TTIT T T T T 1]

—

LTI
R

|

MS
=
I
]

 

i

100

Ve

I 1T TTIHI

 

M

7
AR

M2
_ 3
2

R
EREE

I
I

 

 

 

I L e

0.1 10 100 1000

 

 

F16. 2-19. Neutron power at prompt critical relative to initial power for differ-
ent rates of reactivity addition.
74 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

Having determined the equivalent prompt reactivity addition, the
maximum pressure rise can be calculated by using Eq. (2-28) with m
replaced by m.. To check the validity of the derived relations, values of
Doax Obtained analytically were compared [36] with those obtained by
numerical integration of KEqs. (2-21) through (2-27). The results obtained
by numerical integration were used to establish the relationship between
rate of reactivity addition and equivalent instantaneous reactivity addition
on the basis of equal pressure rise. Thus, for a given rate of reactivity
addition and initial reactor power level, a particular maximum pressure
was obtained. The amount of prompt reactivity added instantaneously
which gave the same maximum pressure was termed the equivalent prompt
reactivity addition corresponding to a given rate addition and initial
power level. For a given rate of reactivity addition, good agreement was
obtained between the equivalent prompt reactivity addition obtained by
numerical integration and that obtained by use of Eqs. (2-29) and (2-30).

Reactivity additions. Homogeneous reactors usually have relatively
large, negative temperature coefficients of reactivity. A high negative
value for the temperature coefficient is usually associated with a high degree
of reactor safety; however, this is true only if reactivity is introduced by
means other than the temperature coefficient itself. If reactivity is added
by adding cold fuel to the reactor core, e.g., by means of a heat-exchanger
accident, then the smaller the value of the temperature coefficient the
smaller the reactivity addition. From this viewpoint, a small temperature
coefficient is desirable if the most hazardous reactivity addition occurs as a
result of introduction of cold fluid into the reactor. On the other hand, a
large negative temperature coefficient is desirable to compensate for
reactivity added by means other than that associated with decreasing the
core fluid temperature.

Reactivity changes can also occur as a result of changes in fuel con-
centration or distribution. In shurry reactors, large changes in reactivity
can result from settling of the fertile material. The results [37] of two-
group, two-dimensional, two-region calculations are shown in Table 2-15
for a cylindrical reactor containing highly enriched UO2SO4 in a 5-ft-
diameter core region, and ThOg in a 23-ft-thick blanket region. The aver-
age temperature was assumed to be 280°C, and the height was assumed to
be equal to the diameter for both core and pressure vessel.

Although the calculated reactivity additions associated with slurry
settling were quite large, such additions would not occur instantaneously;
rather, they would be a function of time, the rate of reactivity addition
being controlled by the rate at which the slurry settled. Settling data for
slurries containing 1000 or 500 g Th/liter correspond to rates of reactivity
addition of less than 0.029, Ak./sec. These rates are well within permissible
rates of reactivity addition; however, if settling took place over a period
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 75

 

 

 

 

 

 

 

 

 

 

 

 

TABLE 2-15
REAcTIVITY CHANGE AS A FUNCTION OF SLURRY SETTLING
.. A i Percentage of Calculated reactivit
Initial thorium concentration g. y
in blanket, g/lit reactor height change, .
n r
1o blanket, g/1te slurry has settled % Ak,
1000 30 +2.8
1000 50 10.2
500 30 2.1
500 50 8.7
.00
‘ | | l |
0.98 — _
12 ft
- 200g Th/liter
5 | initially
Ec 0.96 10% Poisons ]
B
S
2094} —
o 12 ft
E 200g Th/liter
g initially
< 092 — No Poisons ]
S
o
2 12 ft
& 300g Th/liflclar
- initially
0.90 10% Poisons
10 ft
0.88 |— 300g Th/liter -
initially
10% Poisons
0.86 I I l I l
0 8 16 24 32 40 48

Settled Distance (in.)

F1a. 2-20. Effect of settling on critical mass requirements in one-region reactors.

of time, the reactor temperature and pressure would eventually reach the
point where the fuel system would have to be diluted or dumped.

Settling of fuel and fertile material in large single-region reactors may
cause the reactor to become subecritical. The results [38] of two-group
calculations for D2O-ThO2-U?3303 reactors 10 and 12 ft in diameter and
containing 200 and 300 g Th/liter are given in Fig. 2-20. Cylindrical
76 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

geometry was assumed, with height/diameter ratio equal to one; the slurry
was assumed to be homogeneously distributed in the bottom portion of the
reactor, with a D20 reflector on top. In Fig. 2-20 the amount of fuel
required for criticality (after settling) is compared with the amount of fuel
initially required (before settling). Based on these two-group results,
it appears that large one-region reactors become subcritical following
slurry settling.

In the previous analytical presentation, the effect of decomposition
gases upon reactor safety has been tacitly neglected. However, the genera-
tion of hydrogen and oxygen from the decomposition of water represents a
possible explosion hazard. If the fuel contains no catalyst for recombina-
tion, gas bubbles will form and contain an explosive mixture of hydrogen,
oxygen, and water vapor. Explosions might occur in the core, gas separator,
recombiner system, or connecting piping. If an explosion occurs outside
the reactor core, reactivity can be added to the reactor, resulting in an
undesirably high power surge. Under these conditions, not only must the
gas explosion be considered but also the possible reactivity addition asso-
ciated with the explosion. If the gas explosion occurs in the core, however,
no reactivity addition should be involved, and only the pressure associated
with the explosion need be considered. If no explosion occurs, the presence
of gases will still influence the compressibility of the fuel fluid, and this
effect can be considered in the evaluation of fluid compressibility. Neglect
of gas formation in evaluating reactor safety is justified if complete recom-
bination of the decomposed gases is achieved within the reactor, or if the
reactor is initially operating at a power so low that decomposition gases
are not formed (corresponding to absorption of the decomposition gases by
the fuel solution). If the reactor is at low power and reactivity is added to
the system, then neglect of gas formation is conservative with respect to
safety evaluation, since in this case the formation of gases helps in de-
creasing the reactivity of the system. If the initial power is so high that
undissolved gases are initially present within the core region, then the
effect of these gases must be considered in evaluating fluid compressibility.
The presence of undissolved gases would tend to lower the permissible
reactivity addition; however, an increase in gas volume would imply an
increase in initial reactor power, which aids reactor safety if reactivity can
be added as a rate function only.

The number of fission neutrons which are delayed usually is considered
as an important factor in reactor safety. However, if reactivity 1s added
as a linear rate function at low initial reactor power, the delayed neutrons
have little influence upon the reactivity addition above prompt critical.
They will influence the stability and steady-state operational behavior of
the reactor, though, and are necessary to damp the power oscillation follow-
ing a reactivity addition.
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 77

Reactor safety also involves events which are not directly associated
with reactivity additions; e.g., if the circulating pump failed or if there
were sudden cessation of heat removal in the heat exchanger, after-heat
effects may raise the temperature (and pressure) of the system to unde-
sirably high values. The possibility of an instantaneous rupture of a high
pressure steam line, however, is remote, based on experience in conven-
tional plants.*

2-5.2 Homogeneous reactor stability. The purpose of stability studies
is to determine whether the reactor power will return to a stable equilibrium
condition following a system disturbance. Although inherently connected
with safety, stability studies usually treat small reactivity additions and
concern time intervals long in comparison with those involved in safety
studies. The general stability problem can be broken into simpler problems
by eliminating those parts of the physical system which have only a small
influence upon the time behavior of the variable of interest.

The most familiar of stability studies concerns the reactor core-pressurizer
system, and will be referred to as ‘“‘nuclear stability”’ studies. These
consider the high-pressure system alone and assume that the power demand
is constant. This is a valid assumption, since changes in power demand
and changes in variables resulting from fluid flowing between the low- and
high-pressure systems initiate only low-frequency nuclear power oscilla-
tions in comparison with those sustained by the pressurizer-core system.

Reactor instabilities can also arise due to interactions between the
reactor primary system, heat-removal system, and fuel-storage system.
These are associated with the method of reactor operation, and are
discussed under the heading of Operational Stability.

Under certain conditions it may be desirable to operate reactors with
fission gases retained within the system; in these circumstances the buildup
of Xe'3® during periods of low-power operation may influence subsequent
reactor operation. Although no stability problem is involved, a long time
scale is associated with controlling the reactor in these circumstances.
Reactivity effects and required fuel-concentration changes associated with
Xe!3 buildup and burnup are discussed below under the appropriate
heading.

Nuclear stability. In studying nuclear stability, the equations of motion
for a single-region reactor system are normally used. These should be
adequate if the reactor behavior is controlled by a single region; however,
it should be remembered that the mean lifetime of prompt neutrons should
be that for the reactor as a whole. The general mathematical system is too
complicated to handle analytically, so that it is necessary either to resort

*Conversations with utility people indicate there is no accident on record in
which the high-pressure steam line failed instantly.
78 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

to numerical integration of specific cases or to linearize the equations so
that they can be treated analytically. The linearized approach is valid for
very small power oscillations; since the nonlinear effects appear to intro-
duce damping of the power surges, the reactor system should be stable if
the sability criteria for the linearized equations are satisfied. Not meeting
the linearized stability criteria may not necessarily result in the buildup
of pressure oscillations to proportions where reactor safety 1s concerned;
satisfying the criteria should aid in eliminating the small pressure surges
which may physically weaken the system if they occur in a repetitive
manner.

The conventional equations of motion given in Egs. (2-21) through
(2-27) are used, except that the delayed-neutron precursors are assumed
to decay in accordance with a single effective decay constant, and fluid-
flow effects are considered in which the average fuel-fluid temperature is
the linear average of the core inlet and outlet temperatures. Heat-
exchanger behavior will also affect the temperature of the fuel fluid and
is to be considered. The resulting linearized equations [36] lead to an
equation defining the criteria for stability of the reactor system. If only
the reactor core and pressurizer system are considered, corresponding to a
relatively large residence time for fluid in the core, the stability criteria
reduces to

v Bleta+ oo+ By 8)] > (w+5) >0 @

 

Since all quantities are positive, the last inequality is always satisfied.
By replacing the other inequality with an equal sign and fixing all param-
eters but one, the value of the remaining parameter (which barely fails to
satisfy the inequality) can be obtained. By this procedure, the results
given in Fig. 2-21 were obtained for a specific value of ;. Stable operating
conditions are those lying above the appropriate curve.

The above treatment neglects any effects that fluid transport and heat-
exchanger behavior might have upon the temperature of the fluid entering
the reactor. Results of studies [36] in which these effects were considered
indicate that the resulting stability criteria are less stringent than those
for systems in which fluid transport and heat-exchanger behavior are
neglected. On this basis, the results given in Iig. 2-21 form reasonable
bases for safe design.

A one-region reactor will operate at a power density lower than that at
which the corresponding two-region reactor is operated, if the physical
size of the one-region reactor is large. Based on the stability criteria given
in Fig. 2-21, about the same degree of stability will result if the product of
w? w? remains constant in going from a one- to a two-region reactor. On
the basis of fuel-cost studies, the optimum two-region and one-region
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 79

10°

 

T TTin P T

LTI
| LIl

|
|

N
I

 

 

 

 

 

o
g 104 —
3 —
| &) | —
+ 51— |
S _
3 - —
21— —
103 |— -
— _B -
- Y ¢ —
s 1 | [A]] ] L L L
10 2 5 102 2 5 103 2
2
Wp -2
1+c{(se°)

Fra. 2-21. Stability criteria for homogeneous circulating reactors. 7y, = normal-
ized friction coefficient = 0.5.

reactors do have wj w? which are roughly equal, with the one-region
reactor being somewhat more stable. However, there appears to be no
distinct advantage of one type over the other as regards nuclear stability,
since both normally will operate well within the stable region.

Operational Stability. Other types of stability which might be considered
concern the response of the reactor to changes in load demand and to
changes in rates of flow between the low- and high-pressure systems. These
usually involve power oscillations whose frequencies are much lower than
those sustained by the core-pressurizer system. In these circumstances the
general equations of motion can be simplified; however, load-demand
studies will concern the particular characteristics of the heat exchanger
under consideration.

Changes in fluid flow rates between the high- and low-pressure systems
in reactors can also influence reactor stability. During eperation of the
HRE-1, 1t was noticed that under certain conditions the temperature and
power of the reactor would rise with time. This effect was termed the
“walkaway’’ phenomenon [39]. It was associated with a slow rate of
reactivity addition produced by an increase in core fuel concentration,
leading to an increase in operating power and temperature. This increase
in fuel concentration resulted when water was removed from the core
(as a result of decomposition-gas formation and accompanying vaporization
of water) at a faster rate than it was returned from the low-pressure system.
This process continued until it was stopped by the operator.
[cHAP. 2

ONE- AND TWO-REGION HOMOGENEOUS REACTORS

80

*1omod 03 wInjax

SUIMO[[0] ON[BA WNUWIUIW B Y9I 0} UOI}BIJUSIUOI UOUSX J0] Paarnbal sty oy} jo a3ewnysy (3)
-arnjerodwa) 1090%aa UL [[B] Aq pojesuaduiod 919m

dnp[ing UoUSX YIIM PoIBIIOSSE UOIIPPE AJTAI}OBAI JI 98BII0P dinjeIodurd) ping jo ayewmysy (J)
*UOT}BIJUSIUO0D

UOUSX WNWIXBW JO SN} JB ¢p@X JO [BAOWAI )M DPIIBIIOSSE AJIATORAL JO 9)BWSH (9)

-10m0d 0] WINJSI JO SUII) JB JNOUIN] UOUIX 0} 9N UOHIPPE AJA1308aI Jo ey (P)
*(oanjeradure} ping ut 95uryo ou) 1omod 03 UINJAI

JO oW} 98 A[BIIILI0 UIBJUIBW 0} PAIMbal UOIJBIJUIOUOD ¢op() UL 9FUBYD JO 8998I SAIPB[Y ()

"UMOPINYS I9}J8 PuB 910§9q Iomod 03 dAIpE[al UMOPINYS Juunp dmod (q)
*098. ;WD /SU0IINAU

‘woysAs ounssord-ySy I9A0 PoSeISAE® (UMOpINYS AI0Joq XN [BULIBY} JO ON[BA [eryuf (B)

 

 

 

 

 

 

 

 

 

 

 

Z o31% A ‘%8 5—0I X9°¢ 5-01 X6°9 0 ¢101 o -
e adue[ ‘A 'ee ¢-0I X9'8 5-0I XG'¢ 0 ¢101T ¥
I 09 "Ll 50T X L'T 5-0L X6°9 1°0 ¢10T ¥
G 0¢ el ¢ 0L XT1'% ¢_0I XG'8 0 5101 ¥
g 02 06 ¢ 0T X 2'T ¢_ 0L X8'9 1°0 5101 ¥
01 z 980 10T X% L 9_ 0L X0°'¢ 0 e10T %
1998 1998 "
Iy Do % P w N °d 0 “
any 'LV KAy “yp NPT d ¢ .prwa% .
I B
3) ¥) (o) (p) (0) @ | (® e

 

 

dIMOJ OL NYALAY ANV NMOALAHG HONIMOTION SNOILIANO{) d4OLOVHY NOdA ger¥Y dO0 LOTAIAY
OT—-¢G 414dV],
2-5] HOMOGENEOUS REACTORS AND REACTIVITY ADDITIONS 81

The above type of phenomenon is a function of fuel flow rates between
the low- and high-pressure system and can therefore be controlled. Using
the applicable linearized equations of motion, the criteria for stability
toward walkaway can be obtained. This has been done for the HRE-2
[40]. Design parameters such as operating conditions, fluid-flow rates,
vessel volumes, recombination rate of decomposition gases, and other
conditions will influence the degree of stability. Walkaway will not occur
if decomposition gases are not formed. If gas is formed, a paramount
cause of Instability is insufficient overpressure. Increasing the fuel feed-
pump rate or the ratio of solution volume in the high-pressure system to
that in the low-pressure system, or decreasing the reactor temperature or
power will tend to increase reactor stability against walkaway. Walkaway
can also be prevented by automatic control devices which influence the
behavior of the heat exchanger. Walkaway conditions can be instigated
by reducing overpressure or by increasing the rate of gas formation.

Effect of Xe'35 upon reactor behavior. If reactors are operated with com-
plete recombination of the decomposition gases and no letdown from the
high-pressure system, there will be virtually complete retention of the
fission fragments and the fission-product gases within the reactor system.
Under such conditions, partial or total reactor shutdown will lead to an
increase in Xel!35 concentration, which might, in turn, lead to difficulties
in maintaining criticality or reaching criticality upon subsequent power
demand. To investigate the above effect, calculations were performed
[41] for a 4-ft-diameter spherical reactor operating at 280°C, containing
U235 heavy water, 1135, and Xe!35, and operating at initial fluxes of 103 to
10'° neutrons/(cm?)(sec). For these cases the reactor power was reduced
either to zero or to one-tenth the initial value, and the fuel concentration
required for criticality was obtained for times following the power reduc-
tion. At the time that the xenon concentration reached a maximum value,
the power was returned to its original level, and at this point the maximum
rate of reactivity addition associated with xenon burnout was obtained.
The rate of change in fuel concentration required to maintain criticality
at this point was also obtained. In addition, the time required for the xenon
concentration to reach a minimum value following return to power was
evaluated. The maximum change in xenon concentration was used to
estimate the total reactivity addition associated with xenon buildup and
was Interpreted in terms of fluid temperature changes required to maintain
criticality if the fuel concentration were maintained constant. Table 2-16
summarizes the results obtained and, for comparison purposes, also includes
results for a reactor of infinite diameter. As shown in Table 2-16, the
maximum rate of reactivity addition associated with xenon burnout was
less than 1073 Ak,/sec, which does not appear to constitute a dangerous
rate. However, at an operating flux of 10'5 the increase in fuel concentra-
82 ONE- AND TWO-REGION HOMOGENEOUS REACTORS [cHAP. 2

tion or the decrease in temperature required to maintain criticality following
partial or total shutdown is so large that provisions should be made for
elimination of xenon by external means. At 10'* average flux it appears
that xenon poisoning may be compensated by decreasing the reactor
temperature.

REFERENCES

1. J. H. ALexanDpER and N. D. GiveN, A Machine Multigroup Calculation:
The Eyewash Program for Univac, USAEC Report ORNL-1925, Oak Ridge
National Laboratory, Sept. 15, 1955.

2. G. Saronov, Notes on Multigroup Techniques for the Investigation of
Neutron Diffusion, in Reactor Science and Technology, USAEC Report TID-
2503(Del.), The Rand Corporation, December 1952. (pp. 249-272)

3. S. VisNer, Critical Calculations, Chap. 4.2, in The Reactor Handbook,
Vol. 2, Engineering, USAEC Report AECD-3646, 1955. (pp. 511-533)

4. G. Saronov, The Rand Corporation, 1955. Unpublished.

5. S. Jaye, Oak Ridge National Laboratory, in Homogeneous Reactor Project
Quarterly Progress Report, USAEC Reports ORNL-2222, 1957 (p. 43); ORNL-
2272, 1957. (p. 51)

6. B. E. Prince, Breeding Ratio in Thorium Breeder Reactors, in Homo-
geneous Reactor Project Quarterly Progress Report for the Period Ending Jan. 31,
1958, USAEC Report ORNL-2493, Oak Ridge National Laboratory, 1958.

7. T. B. FowLEr, Oracle Code for a General Two-region, Two-group Spherical
Reactor Calculation, USAEC Report CF-55-9-133, Oak Ridge National Lab-
oratory, Sept. 22, 1955.

8. M. Tosias, Oak Ridge National Laboratory, 1956. Unpublished.

9. M. ToBias, A “Thin-shell’” Approximation for Two-group, Two-region
Spherical Reactor Calculations, USAEC Report CF-54-6-135, Oak Ridge National
Laboratory, June 1954.

10. J. T. RoBERTS and L. G. ALEXANDER, Cross Sections for Ocusol-A-Program,
USAEC Report CF-57-6-5, Oak Ridge National Laboratory, July 11, 1957.

11. I. V. KurcuaTov, Some Aspects of Atomic-power Development in the USSR,
USSR Academy of Sciences, Moscow. (Talk presented at Harwell, England,
1956.)

12. E. H. MacLEBY et al., Energy Dependence of Eta for U-233 in the Region
0.1 to 8.0 Ev, USAEC Report ID0-16366, Phillips Petroleum Co., Nov. 19, 1956.

13. D. E. McMILLAN et al., in Report of the Physics Section for June, July,
August 1956, USAEC Report KAPL-1611(Del.), Knolls Atomic Power Labora-
tory, Dec. 14, 1956. (pp. 14-15)

14. P. N. HauBENREICH, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 81, 19568, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958.

15. M. C. Eprunp and P. M. Woop, Physics of the Homogeneous Reactor
Test-Statics, USAEC Report ORNL-1780(Del.), Oak Ridge National Laboratory,
Aug. 27, 1954.

16. B. E. Prince and C. W. NESTOR, JR., in Homogeneous Reactor Project
REFERENCES 83

Quarterly Progress Report for the Pertod Ending Oct. 31, 1957, USAEC Report
ORNL-2432, Oak Ridge National Laboratory, 1958. (p. 17) -

17. D. J. HucrEs and R. B. Scuwartz, Neutron Cross Sections, USAEC
Report BNL-325, Brookhaven National Laboratory, Jan. 1, 1957.

18. T. B. FowLEr and M. Tosias, Two-group Constants for Aqueous Homo-
geneous Reactor Calculations, USAEC Report CF-58-1-79, Oak Ridge National
Laboratory, Jan. 22, 1958.

19. J. HALPERIN et al., Capture Cross Section of Pa-233 for Thermal Reactor
Neutrons, in Reactor Science and Technology, USAEC Report TID-2504(Del.),
Oak Ridge National Laboratory, 1953. (pp. 345-348)

20. H. C. CraiBorNE and T. B. FowLER, Fuel Cost of Power Reactors Fueled
by UO2804-L19S04-D20 Solution, USAEC Report CF-56-1-145, Oak Ridge
National Laboratory, Jan. 30, 1956.

21. D. E. McMiLLaN et al., A Measurement of Eta and Other Fission Param-
eters for U-233, Pu-239, Pu-241, Relative to U-235 at Sub-Cadmium Neutron
Energies, USAEC Report KAPL-1464, Knolls Atomic Power Laboratory, Dec.
15, 1955.

22. M. Tosias, Certain Nuclear Data and Physical Properties to Be Used in
the Study of Thortum Breeders, USAEC Report CF-54-8-179, Oak Ridge National
Laboratory, Aug. 26, 1954.

23. C. W. NEsToR, Jr.,, and M. W. RoSENTHAL, in Homogeneous Reactor
Project Quarterly Progress Report for the Period Ending Apr. 30, 1956, USAEC
Report ORNL-2096, Oak Ridge National Laboratory, 1956. (pp. 60-62)

24. M. W. RosENTHAL and M. ToBias, Nuclear Characteristics of Two-region
Slurry Reactors, USAEC Report CF-56-12-82, Oak Ridge National Laboratory,
Dec. 20, 1956.

25. D. C. Hamivton and P. R. KASTEN, Some Economic and Nuclear Charac-
teristics of Cylindrical Thorium Breeder Reactors, USAEC Report ORNL-2165,
Oak Ridge National Laboratory, Sept. 27, 1956.

26. M. Tosias, Oak Ridge National Laboratory, in Homogeneous Reactor
Project Quarterly Progress Report, USAEC Reports ORNL-2379, 1957 (p. 43);
ORNL-2432, 1958. (p. 46)

27. M. W. RosentaAL and T. B. FowLER, in Homogeneous Reactor Project
Quarterly Progress Report for Pertod Ending July 31, 1957, USAEC Report ORNL-
2379, Oak Ridge National Laboratory, 1957. (p. 37)

28. J. A. LANE et al., Oak Ridge National Laboratory, 1951. Unpublished.

29. R. B. Brigas, Aqueous Homogeneous Reactors for Producing Central-
station Power, USAEC Report ORNL-1642(Del.), Oak Ridge National Labora-
tory, May 4, 1954.

30. M. W. RoseNTHAL et al., Fuel Costs in Spherical Slurry Reactors, USAEC
Report ORNL-2313, Oak Ridge National Laboratory, Sept. 11, 1957.

3l. T. B. FowLER, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak Ridge
National Laboratory, 1957. (p. 44)

32. M. TosBias et al., in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak Ridge
National Laboratory, 1958. (p. 42)
84 ONE- AND TWO-REGION HOMOGENEOUS REACTORS

33. M. Tosias, in Homogeneous Reactor Project Quarterly Progress Report for
Period Ending Jan. 31, 19568, USAEC Report ORNL-2493, Oak Ridge National
Laboratory, 1958.

34. P. R. KasteN et al., Fuel Costs in Single-region Homogeneous Power
Reactors, USAEC Report ORNL-2341, Oak Ridge National Laboratory, Nov.
18, 1957. (pp. 42-50)

35. J. M. SteiN and P. R. KasTEN, Boiling Reactors: A Preliminary Investiga-
tion, USAEC Report ORNL-1062, Oak Ridge National Laboratory, Nov. 23,
1951. W. MARTIN et al., Density Transtents in Boiling Liquid Systems; Interim
Report, USAEC Report AECU-2169, Department of Engineering, University
of California, Los Angeles, Calif., July 1952; Studies on Density Transients in
Volume-heated Boiling Systems, USAEC Report AECU-2950, Department of
Engineering, University of California, Los Angeles, Calif., July 1953. M. L.
GrEENFIELD et al., Studies on Density Transients in Volume-heated Boiling
Systems, USAEC Report AECU-2950, Department of Engineering, University
of California, Los Angeles, Calif., October 1954. P. R. KastEN, Boiling Reactor
Kinetics, Chap. 4.2, in The Reactor Handbook, Vol. 2, Engineering, USAEC
Report AECD-3646, 1955 (pp. 551-552) ; Operation of Boiling Reactors: Part I—
Power Demand Response, USAEC Report CF-53-1-140, Oak Ridge National
Laboratory, Jan. 14, 1953; Boiling Reactor Operation: Part II— Reactor Governors,
USAEC Report CF-53-2-112, Oak Ridge National Laboratory, Feb. 12, 1953.

36. P. R. KasteN, Dynamics of the Homogeneous Reactor Test, USAEC
Report ORNL-2072, Oak Ridge National Laboratory, June 7, 1956. |

37. H. C. CraiBorNE and P. R. KasTEN, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Jan. 31, 1956, USAEC Report
ORNL-2057(Del.), Oak Ridge National Laboratory, 1956. (pp. 68-69)

38. H. C. CLAIBORNE, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1956, USAEC Report ORNL-1943, Oak
Ridge National Laboratory, 1955. (p. 44 ff)

39. S. VisNER, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending Jan. 81, 1954, USAEC Report ORNL-1678, Oak Ridge
National Laboratory, 1954. (p. 9 ff)

40. M. Tosias, Oak Ridge National Laboratory, in Homogeneous Reactor
Project Quarterly Progress Report, USAEC Reports ORNL-1853, 1955 (pp. 23-27);
ORNL-1895, 1955. (pp. 27-29)

41. M. Tosias, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending July 31, 1956, USAEC Report ORNL-2148(Del.), Oak
Ridge National Laboratory, 1956. (p. 41 ff)
CHAPTER 3

PROPERTIES OF AQUEOUS FUEL SOLUTIONS*

3—1. INTRODUCTION

The chemical and physical properties of aqueous fuel solutions are
important because they affect the design, construction, operation, and
safety of homogeneous reactors in which they are used. This chapter will
discuss primarily those chemical and physical properties, except corrosion,
which are important for reactor design and operation. Special attention
will be given to the properties of solutions of uranyl sulfate, since such
solutions have been the most extensively studied, and at present appear
the most attractive for ultimate usefulness in homogeneous reactors.

Solubility relationships are discussed first, with data for uranyl sulfate
followed by information concerning other fissile and fertile materials.
The effects of radiation on water, the decomposition of water by fission
fragments, the recombination of radiolytic hydrogen-oxygen gas, the de-
composition of peroxide in reactor solutions, and the effects of radiation
on nitrate solutes are then presented. Finally, tables of relevant physical
properties are given for light and heavy water, for uranyl sulfate solutions,
for other solutions of potential reactor interest, and for the hydrogen-
oxygen-steam mixtures which occur as vapor phases in contact with reactor
solutions. |

3—2. SOLUBILITY RELATIONSHIPS OF FISSILE AND FERTILE MATERIALST

3-2.1 General. For the most part, studies of aqueous solutions of fissile
materials for use in homogeneous reactors have dealt with hexavalent
uranium salts of the strong mineral acids. Hexavalent uranium in aqueous
solutions appears as the divalent uranyl ion, UOs*+. Tetravalent uranium
salts in aqueous solutions are relatively unstable, being oxidized to the
hexavalent condition in the presence of air. Other valence states of uranium
either disproportionate or form very insoluble compounds and have not
been seriously proposed as fuel solutes. |

Uranyl salts are generally very soluble in water at relatively low tem-
peratures (below 200°C). At higher temperatures, miscibility gaps appear
in the system. These are manifested by the appearance of a basic salt solid

*By H. F. McDuffie, Oak Ridge National Laboratory.
tTaken from material prepared by C. H. Secoy for the revised AEC Reactor
Handbook.

85
86 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [crAP. 3

phase from dilute solutions and by the appearance of a uranium-rich
second liquid phase from more concentrated solutions. In both the salt
and the second liquid phase, the uranium-to-sulfate ratio is found to be
greater than in the system at lower temperature; this suggests that hy-
drolysis of the uranyl ion is responsible for the immiscibility in each
instance. Hydrolysis can be repressed effectively by increasing the acid-
ity of the solution or, alternatively, by the addition of a suitable complexing
agent for the uranyl ion. Even the anions of the solute itself may be con-
sidered to accomplish this to some degree, since very dilute solutions
hydrolyze much more extensively than more concentrated solutions.

Primary emphasis has been placed on the study of uranyl sulfate solutions
because of the superiority of the sulfate over other anions with respect to
thermal and radiation stability, absorption cross section for neutrons,
ease of chemical processing, and corrosive properties. Other uranyl salts
which have either been used in reactors or studied for possible use include
the nitrate, phosphate, fluoride, chromate, and carbonate. It has been
found possible to improve the solubility characteristics of uranyl salt
solutions at elevated temperatures by the addition of acids or salts of the
chosen anion.

The marked differences between light water and heavy water with re-
spect to moderating ability and thermal neutron absorption cross section
make solutions in both solvents of interest for reactor use. Generally
speaking, the upper temperature limit of solution stability occurs about
10°C lower in heavy-water solutions than in light-water solutions.

Tetravalent uranium can be stabilized by increasing the reduction po-
tential of the solution. However, tetravalent uranium is more readily
hydrolyzed at elevated temperatures than hexavalent uranium, and 1t
probably cannot be kept in solution except by the use of otherwise ex-
cessively high concentrations of acid.

Plutonium, the other fissile material, also forms salts which can be dis-
solved in water. The possibility of using such solutions in aqueous homo-
geneous reactor systems has been examined, and limited experimental
studies have been directed toward this goal but without substantial suc-
cess (see Article 6-6.3).

Uranium-238 and thorium, the fertile materials, have been considered
for use in converter or breeder reactor systems. The solubility of uranium
is such that satisfactory aqueous solutions of uranium can be obtained for
use in the conversion of U238 to plutonium. Substantial efforts have been
made to develop solutions of thorium which could be used as a blanket in
a two-region breeder reactor system.

Thorium appears to be stable in the tetravalent form but has a strong
tendency toward hydrolysis at elevated temperatures. Insoluble thorium
dioxide is ultimately formed as the hydrolysis product. Thorium nitrate
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 87

400 T

 

1

    
   
   
  
 

T 1 1 1
Unsaturated Solution (L9) + Vapor

 

350
Two-liquid Phases

300

 

250 —

200 Unsaturated Solution

Temperature, °C

100 —

Ice + Solution
50

UO2S04°3 HyO + Solution

 

-50 Ice +UO,S0,-3H,0 —

| | | | | | | |

0 1 2 3 4 5 6 7 8 9
m, moles/1000 g HyO

 

 

 

F1a. 3-1. Phase diagram for the system U02S04-H 0.

and thorium phosphate can be maintained in solution at satisfactory con-
centrations by the use of substantial concentrations of nitric or phosphoric
acids to inhibit hydrolysis. However, in the nitrate system an acceptable
breeding ratio could only be obtained by using N15. Thorium phosphate
solutions containing the necessary amount of phosphoric acid are extremely
corrosive to all but the noble metals.

Neptunium and protactinium complete the listing of fissile and fertile
materials, since these are intermediates in the production of plutonium and
U#33 from U238 and thorium. Limited exploratory studies of their solu-
bilities have been carried out primarily in connection with the development
of processes for their continuous removal from blanket systems.

3-2.2 Uranyl sulfate. The solubility of uranyl sulfate in water and the
characteristics of the phase relationships at elevated temperatures, dis-
played in the form of a binary system, U02804-H20, are shown in
Fig. 3-1 [1]. It is necessary, however, to study the ternary system,
UO3-503-H:20, in order to understand the hydrolytic precipitation of the
88 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

UO3-2503-1.5H70
UO3-2503'3H70
UO3-2503°6H70
UO2S04-2H70
UO2504°3H20
UO3-H20

{a & B Polymorphic Forms)
G = G Basic Salt

(5U03-2503-yH,0)?
K = K Basic Salt
(8UO3-3503°xH20)?

H,O

H20

mm g N . >

   
   
   
     
   
   
   
 

\ All Scales Are
Wt %

 

 

 

O3

 

 

 

 

SO
3 (c) 175°C UO3

Fia. 3-2. The system UQ3-S05-H:0.

basic solid phase, 3~-UQO3 - H20, which occurs in very dilute solutions at
elevated temperatures, and the position of the tie lines in the liquid-liquid
miscibility gap. Figure 3-2 shows portions of the ternary isotherms at
25, 100, 175, and 250°C [2,3]. A point of special significance in these dia-
grams is that the solubility of UO3 in uranyl sulfate solutions decreases
with increasing temperature to the extent that at 250°C excess acid 1s re-
quired to maintain homogeneity in solutions of low concentration. Excess
acid also has a marked effect on the liquid-liquid miscibility gap, as shown
by the curves in Fig. 3-3 [4]. In very dilute solutions the surface formed
by these curves intersects the surface representing the liquid compositions in
equilibrium with the hydrolytically precipitated solid phase, 8—~UOs - H20.
Figure 3—4 shows this intersection and three paths on the liquidus surface
at fixed SO3/UO3 mole ratios [5].

Figure 3-5 shows, from the data of Jones and Marshall [6], how the
two-liquid-phase separation temperature is lowered when the solvent is
changed from light water to heavy water. Scattered experiments suggest
that the temperatures for solid-phase separation through hydrolytic pre-
cipitation are also somewhat lower in heavy-water systems than in systems
containing light water as the solvent. '
3-2]

Temperature, °C

SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE)

450

425

400

375

350

325

300

275

 

 

 

 

 

 

 

l ; One Phase | One Phase
at 525°C at 530°C
S-Cfi = 2.955 SO
v 224 = 2481
U
S& = 1.883
U
P4 _ 428
u
P4 _
/
S
—0—4 1.000
U
l | | I
5 10 15 20 25

Uranium, wt %

F1a. 3-3. Coexistence curves for two liquid phases in the system

U02504-H2504-H0.

89

Figure 3—6 shows the liquidus composition isotherms from 150 to 290°C
for dilute sulfuric acid solutions saturated with UO3 [7].
Study of the data leads to the following general conclusions with respect
to the stability of uranyl sulfate solutions of reactor interest:

(1) Stoichiometric uranyl sulfate solutions in light and heavy water are

unstable at temperatures of 280°C and above because of hydrolysis.

(2) Stability up to approximately 325°C is provided at uranium concen-
trations up to 2.5 w/o by the addition of a 50 mole 9} excess of sulfuric

acid.

(3) Stability up to as high as 400°C is provided at uranium concentra-
tions above 20 w/o by the addition of a 100 mole 9, excess of sulfuric acid.

-
90 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

 

375 I T I

 

350 |

w

N

On
|

Temperature, °C

w

o

o
]
T

 

 

 

 

 

250
2.0

Uranium, wt %

Fic. 3-4. Effect of excess H2SO4 on the phase equilibria in very dilute UO2804

solutions.

 

Temperature, °C

 

 

 

270 L1
0 1 2 3 4 5 6
m, moles U02504/1000 gm H,0, DO

 

F1c. 3-5. Two-liquid phase region of uranyl sulfate in ordinary and heavy water.
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 91

 

     
  
    
   
  

 

 

 

 

 

 

 

 

 

 

10 E | [ | | | [ Two-Liquid Phase g
[ Region - —
0.1 = ® Direct UO3 and SO4 —__—_—“
— Analyses —
£ — e Analyses Based on ]
c [ pH at Room Temperature _
o
3 - 7
t
$ 001 — —
c __ p—
o — Z
o, | _
<
O — -
N | -
o~
T | _
0.001 = —
0.0001 I
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
Solubility Mole Ratio UO3/H2504
Fra. 3-6. Solubility of UO3 in H2SO4-H20 mixtures.
400
T T T T R T
380 |— —
360 |— [ —
—
340 [— —_
&
2\ = ——
5
B 320 — —
8 | —
E —
o
= 300 |- —
P— 0.42 - 0.45 m UO2504 Plus LigSO4 ]
2.2 mUO2SO4 Plus LiogSO4
260 |— —
240 L [ L LIl L L 1 11ill] |11
0 0.02 0.05 0.1 0.2 0.5 1.0 2.0 50

LipSO4 , m

F1c. 3-7. Second-liquid phase temperature of UO2804-Li2SO4 solutions. Con-
centrations are uncorrected for loss of water to vapor phase at elevated temper-
atures.
92 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

 

    
 
  

 

 
   

   

 

 

 

 

 

 

 

 

 

15 q
N 325°C”  19.25 mole %
g\\\\\\\ 390 Excess Sulfuric Acid
210 K A i
v
o 310
v
N T N
0 5 10 15 20 O 10 15 20 O 5 10 15 20
_ UO9SO4 ,wt % UO9SO4 . wt % UO9SO4 ,wt %
15 340°C—29.11 mole % Excess 38.75 mole % Excess --- Temperature Cont;)ur Lines (isot}l:erms)
Sulfuric Acid Sulfuric Acid _For Appearance of Two Liquid Phases
o v \V M Region of Blue-Green Crystalline Solids
°10 0\ \ [T Region of Red Solids
3 325—\ 3:;5—\ A Reference Composition of
<r
220 30—~ 0.1 M CuSO4
s \ 3 2N 0.1 M UO2504
o 315 < Reference Composition of HRE-2 Fuel
( | 32 0.005 M CuSO4
| | | N 0.04 M UO2504
0 5 10 15 20 O 5 10 15 20 (Plus ~ 0.025 M H2504) (55%)
UO2504 ,wt % UO2504,wt %

F1g. 3-8. Phase transition temperatures in solutions containing cupric sulfate,
uranyl sulfate, and sulfuric acid.

 

 

 

 

 

 

 

0.04 ® ® — - *
(249) (<249) T(<249) ; T 200 (186) (175)
282 ' 215
220
185
o5 200
(261)
0.03 — @ o ® ® (]8—5)0
(249) (249) (215) (197)
-282/ \-282 229
. 3 CuO-S0O3-2H920
2 UO3-SO3-5H20 2 (+ CuSO4 H,0)
E ( + NiSO4-H0) %
< 7Z )
0.02 —— 7 ® ® —
Q % (240) (211) (186)
Z éé 216
%
Z
----- %,
%,
%
%
% —@
001 %, (215) (189)
Second Liquid Phase //////
330 320 310 ,
\| “ 199)
0 0.01 0.02 0.03 0.04
CuSO4, m

Fig. 3-9. The effect of CuSO4 and NiSO4 on phase transition temperatures
(0.04 m UOzSO4; 0.01 m H2804).
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 93

The addition of lithium sulfate or beryllium sulfate to uranyl sulfate
solutions has been found to elevate the temperatures at which the second
liquid phase appears [8]. Figure 3-7 shows the effect of LioSO4 additions
on the second liquid phase temperature for three uranyl sulfate solutions.
In very dilute uranyl sulfate solutions excess acid must also be added to
prevent hydrolytic precipitation..

The solubility relationships in uranyl sulfate solutions containing cupric
copper are also of interest (see Article 3-3.4). Copper sulfate solutions,
like uranyl sulfate, undergo hydrolytic precipitation at elevated tempera-
tures [9]. Even though the required concentration of cupric ion may be
quite low, its presence influences the phase relationships. This influence
is most significant in dilute uranyl sulfate solutions. A complete phase
diagram for the four-component system, CuO-UO3-S03-H20, has not
been determined, but regions of special interest have been studied. I'ig-
ure 3-8 shows the phase transition temperatures in solutions containing
copper sulfate, uranyl sulfate, and sulfuric acid. The solid phase which
appears at the higher CuSO4 concentrations has been shown to be at least
in part the basic copper sulfate, 3CuO - SO3 - 2H20 [10].

In uranyl sulfate solutions in contact with austenitic stainless steels it
is important to know the effect of the corrosion products upon the solu-
bility relationships. Under most conditions iron and chromium appear as
insoluble hydrolytic products, but nickel appears as a soluble contaminant
of the solution. Studies have been made of the precipitation temperatures
for dilute solutions in the system UQO2S0~—CuS0O4+—NiSO4s+—H2SO4~-H-0,
and the solid phases have been identified [11]. Figure 3-9 summarizes
the information for systems having compositions approximately that of
the fuel solution of the HRE—-2. In this preliminary study the tests were
limited to short time intervals (15 minutes or less of exposure to the ele-
vated temperatures). When solutions containing 0.01 m CuSO4 plus
0.01 m NiSOy4, or 0.02 m CuSO4 with no NiSO4 were heated for longer
periods of time at 300 to 310°C (just below the temperature for the forma-
tion of two liquid phases) green solids were deposited. Thus the results
pictured in Fig. 3-9 should be applied to practical situations with con-
siderable reservation until experiments with the exact composition of
interest have been conducted.

3-2.3 Other uranium compounds. Uranyl nitrate. A phase diagram for
the system uranyl nitrate-water [12] is shown in Fig. 3-10. Although
uranyl nitrate remains very soluble at the elevated temperatures of in-
terest for power-reactor operation, the nitrate group in such solutions
decomposes to yield oxides of nitrogen which appear in the vapor phase.
Although this situation is reversible with the lowering of temperature, it
does introduce corrosion problems with respect to the vapor phase. The
94

400

300

250

N
o
o

Temperature, °C
O
o

o
o

50

-50

PROPERTIES OF AQUEOUS FUEL SOLUTIONS

 

    
 

C Precipitation Limits
@ Vapor Phase Coloration

 

 

 

 

0 10 20 30 40 50 60 70 80 90 100
UOz(NO3)2, Wt %

F1a. 3-10. The UO2(NOj3)2-H20 system.

Temperature, °C

400

350

300

250

200

150

100

50

-50

 

[ [ I

Solution +
[ Satu rafed/

| Solution ¢

- H—A—A—A—A—A—

— Basic Solid -

| ! Solid + Super Critical Fluid ' |
A—b—A—4

v UO2F2:2H20 + Liquid |
G—a—8—B—8—8—8—8—g—

m

  
 

   

F Liquid | + Liquid I

 

+ Saturated
Solution

¥ UO2F9-2 ;

___./.
-9

K
— Unsaturated Solution

— OeBA Ornl Data

A lIce + Saturated So/l'n.
¥ 000-go &

e

© Data of Dean
| UO9F9-2H20
o Data of Kunin ‘£ (:_ Sq%ugatedz

O

Sat'd. Solution
.\:__.

 

7

Q
Se
2\. BUO2%F9:2H20 +

Solution

B

 

[ Ice + @ UD9Fy 2H,0

 

1

 

L 1

 

0 10 20 30

40 50 60
UO9oFg, Wt %

70 80 90

[cHAP. 3

F1a. 3-11. Phase equilibria of aqueous solutions of UO3 and HF in stoichiometric
concentrations.
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 95

 

   
   
 
 
    

 

 

 

2.0 | | l I
Solid Phase Is
UO2(H2PO4)2'3H2O
(25°C)
1.0 —
_ Solid Phase Is ]
(25° — 400°C) ]
2 ]
o
< 0.5 —
=
5 —
0.2 ]
] | I [ b1
0.1 2 5 10 20

PO4, Molarity

F1g. 3-12. Solubility of UO3 in H3PO4 solution.

nitrate ion is, moreover, not completely stable in fissioning solutions;
elemental nitrogen is one of the products of radiation decomposition.
Although uranyl nitrate solutions have proved quite satisfactory in low-
power water-boiler type research reactors, where makeup nitric acid can
be added as needed [13], they do not appear attractive for high-tempera-
ture, high-power aqueous homogeneous reactors.

Uranyl fluoride. Uranyl fluoride is a very attractive fuel solute because
of the low neutron capture cross-section of fluorine. However, at high
temperatures 1t undergoes hydrolysis, which means that excess HF would
be required to maintain homogeneity. Hydrogen fluoride is also a com-
ponent of the vapor phase. Both liquid and vapor are very corrosive
toward zirconium and titanium, but less corrosive toward stainless steel
(see Article 5-3.3). Figure 3—-11 shows the phase relationships in this
system [14].

Uranium phosphate. Neither hexavalent nor tetravalent uranium phos-
phate is sufficiently soluble in water to be of reactor interest, but both
UO2 and UO3 are quite soluble in moderately strong phosphoric acid.
These solutions have been the subject of considerable study at the Los
Alamos Scientific Laboratory [15]. The solubility of UOgs in phosphoric
acid is illustrated by Fig. 3-12 [16]. Although the solubilities of uranyl
96 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cEAP. 3

 

Non-Binary

T
130 |— Solid + Liquid i —
~
120 |— - ]
/I -
~
110 — ~
~
100}— r

90 —

 

B + Liquid

   
    

80 |—

70 —

Temperature , °C

50 |—

 

 

 

 

 

30— A+B __|
20 A + Lliquid _
10 A = UO,CrO4-5% HyO —
= UO2CrO4-X HoO
0 —_—
A +lce
~10 oo
0 10 20 30 40 50 60 70 80 90 100

U02Cr04 ,wt %

Fia. 3-13. The system UO2:CrO4-H20.

phosphate and uranyl sulfate in water are quite different, their respective
solubilities in concentrated phosphoric acid and concentrated sulfuric acid
are analogous; in either case temperatures as high as 450°C can be ob-
tained with no phase separation. Phosphorus has an advantage over sulfur
in possessing a somewhat lower neutron absorption cross section, but both
anions appear to be stable under radiation. Both the phosphate and sul-
fate solutions in concentrated acid at temperatures of 450°C are extremely
corrosive toward most metals and alloys except the noble metals. Attempts
to operate experimental high-temperature reactors using uranium
phosphate-phosphoric acid fuel solutions have failed because of catastrophic
corrosion rates due to imperfections in noble metal plating or cladding of
the reactor core and heat-exchanger tubing [17] (see Section 7-5).

Uranyl chromate. Uranyl chromate solutions also suffer from hydrolysis
at elevated temperatures; excess chromic acid is required for stability [18].
This system is, however, not unattractive insofar as corrosion of stainless
and carbon steels is concerned. The conditions of acidity and oxidation-
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 97

 

1.000 I | [ I

   
   
   

0.800 |—

Predicted 2375 psi CO9

 

 

 

=
™ ¢
S o 875 psi CO9
T o
wed
375 psi CO9
0.400 |— —
® Slopes:
@ 3.37 moles of Li per mole of U
o (LipCO3 Is Solid Phase)
A 4.44 moles of Li per mole of U
o at Isobaric Invariant
0.200 I I 1 | | | | l
0 0.02 0.04 006 008 010 0.12 0.14 0.6 0.18
U02C03, m

Fi1g. 3-14. Variation of LioCOj solubility with UOsCO3 concentration at con-
stant CO3 pressure (250°C).

reduction potential determine the valence state of the chromium, but
present knowledge is not adequate to specify required conditions for re-
liable behavior at elevated temperatures and under reactor radiation.
Figure 3-13 shows the phase diagram insofar as it has been established.

Uranyl carbonate. Uranium trioxide is quite soluble in alkali carbonate
solutions. This solubility can be attributed to the complexing of UO3 or
uranyl ion by the bicarbonate ion to form uranium-containing anions.
In any event, one would not expect the solubility of uranium to be retained
at high temperature unless the carbonate content of the aqueous phase
were kept high. This can be accomplished by retaining an adequately
high partial pressure of CO2 over the solution. The solubility of UO3 in
L12CO3 solutions at 250°C has been studied [19], and the significant re-
sults are shown in Figs. 3-14 and 3-15. Referring to Fig. 3-14, we see that
at a constant CO2 pressure the concentration of uranium increases linearly
with the lithium concentration until a limit is reached at the isobaric in-
variant. The uranium concentration cannot be increased further unless
the COg2 pressure is increased. Figure 3-15 is a projection of the composi-
tions of solutions saturated with respect to lithium and uranium at 250°C
and at a constant, total pressure (COz2 + steam) of 1500 psi. The projection
98 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

 

H,0
100
¢
O
oe <
S Yo
'S5 0 95 .
S I 2

[\
1 \N\\
1 N
1 W

I W\

! \ \\

| VA

! \‘ \

! \

I o

10 \V4 \/ ¢ \\1 \ 90
0 5 10 15 20
UO3IW' %

Fig. 3-15. The system Li20-UQ3-CO2-H20 at 250°C and 1500 psi.

figure gives no information concerning the concentration of carbonate in
the liquid phase. The region in which a single homogeneous liquid phase.
exists is the very narrow shaded region near the H2O apex. Although the
scope of this region is small, there should be no difficulty in maintaining a
homogeneous liquid phase if an adequate pressure of CO2 is kept on the
system and an appropriate composition is selected in preparing the solution.

3-2.4 Solubilities of nonuranium compounds. Thorium. Thorium solu-
tions having concentrations as high as 0.5 m would be useful for one-region
breeder reactors. For the breeder blanket of a two-region reactor con-
centrations of about 6.0 m thorium appear to be optimal, although some-
what lower concentrations would be of interest. At the present time only
thorium nitrate or phosphate solutions in the presence of excess acid have
been demonstrated to have the required solubilities at elevated tempera-
tures; both of these solutions have substantial disadvantages. Thorium
chloride would be expected, by analogy, to show substantial solubility at
elevated temperatures, but this system has not been investigated in
detail. Complex organic salts, such as thorium acetylacetonate, have high
solubilities at relatively low temperatures, but these have not been in-
vestigated for use in aqueous solutions at temperatures above 100°C.

Data from the literature on the solubility of thorium sulfate at low
temperatures both alone and in the presence of other solutes [20] indicate
that such solutions will probably not be satisfactory at elevated tem-
peratures.

Thorium phosphate (or thorium oxide) is very soluble in concentrated
phosphoric acid. Solutions containing up to 1100 g Th/liter with PO4/Th
3-2] SOLUBILITY RELATIONSHIPS (FISSILE AND FERTILE) 99

 

 

  

 

 

 

380 I i
| NO3/U = 2.0
300 ]
NO;/Th = 5.47
|9/
°. 260 —
® NO3/Th = 6.65
=
B
@
Q. — —
£ 220
@
-
NO5/Th = 4.0
180 — —
140 |— —
100 | |
0 1.0 2.0 2.5

Thorium, Uranium, m

Fi1c. 3-16. Hydrolytic stability of thorium nitrate and uranyl nitrate solutions.

ratios of 5 and 7 could be prepared and appeared to be thermally stable but
had high viscosities. Solutions containing 400 g Th/liter at PO4/Th ratios
of 10 were thermally stable at 250 to 300°C with viscosities little higher
than that of concentrated phosphoric acid. Efforts to improve the prop-
erties of thorium phosphate-phosphoric acid systems by the inclusion of
HF, HNO3, H28047, SeO4~, SO4, Lit, or Mg**, alone or in combination,
have not proved encouraging [21].

The thorium nitrate-water system has been reported [22] as having
considerable solubility up to about 225°C, at which point hydrolytic pre-
cipitation occurs. Further investigation [23] revealed a maximal stability
for the 80 w/o0 material (to around 255°C). Increasing the acidity of the
solutions (increasing the NOgsz™ /Th ratio) suppresses hydrolysis and in-
creases the stability of the solutions as indicated by Figure 3-16, which
shows the precipitation temperatures for various solutions [24]. The
intensity of vapor phase coloration at elevated temperatures (rapidly
reversible) increased as the nitrate/thorium ratio was raised above 4.0.

Plutonium. A considerable investigation of the chemistry of plutonium
in aqueous uranyl sulfate solutions has been directed, not toward the
achievement of solubility, but toward the achievement of <nsolubility
in order to provide the basis for continuous processing of a U238 blanket
solution for plutonium production [25] (see Chapter 6).

The possible use of aqueous solutions of plutonium in homogeneous re-
actors has been reviewed by Glanville and Grant [26] in order to determine
[cHAP. 3

100

 

PROPERTIES OF AQUEOUS FUEL SOLUTIONS

UOI)BULIOJUL ON

I9[/NJ JO sweLd Jo IapIo 3y} JO
€00%eN A 20°0 ur 1091 /ng 3§'8<

“ EONH I/ 10
YOd®H I 9°0 ur 1911/ng 8 2'0<

UOI}BULIOJUT ON

UOI}BULIOJUL ON

/N 8 005~

POIOH &nip ut

IN[/NJ JO swels Jo J9pI0 Y} JO
§0IM W ¢'0 Ut 11/ng 3 9°0
UOT}BULIOJUL ON

UOT)BULIOJUL O N

IN[/nd 3 0ge~

AH A 61 ut 181[/nd 3 0§ <

a[qNJos AI9A
AH W T SONH I T
YOOSH W T1°0 Ut 131/ng
3¥70< st pejaodax "XBIA[
BNOODJSHD W 20
“COD%EN 4 30 Ul 131/ng 3 1°0

POS®H 4 T Ul 1091]/0ng 3 GG°Q
SONH W 1°0
‘LQ%ID%eN I T°0 UL 1991/ng
3620 ‘*ONH J OT Ul 9[qn[og
YOS®H J T°0 Ut 1e31[/ng 3 6Z1 <
SEONH /1 & ut 1931/nd 3 009
YOIOH on[Ip Ul
INI/NJ JO SWBIS JO I9PIO 9Y) J()
EONH /1 9 “SOI A 10 Wl
I /D] 3w G'§6 ST pajtodar ‘XBIN

YOI A ST°0

TOSPTH W S'T ur 1¢3/ng 88<
IgH W ¢

ur INI/ng 81 jo I1pIo 8y} j0
[DHNW9

ur 1B_Y1[/nd S (g Jo JIepIo 8y} O
[V 10 17 “3'8 suor Surxardwood
opuiony Jo 9ouesard Ul |[qNOS

a[qn[os A1 A

+H W LE
YOOSH W G0 ur 1931/nd 3 9%°0

€00 %Sy ur s[qnog
IDH W 60 “OdEH W 80 ut

PN[/NJ 3 68°¢ ST paptodor "XBIN

GOE@E&O«GM OZ

YOSPH W 1°0 Ut 1991;/nd 3 ¢ZT

EONH /I 6°0 ut 1311/nd 8 272 <
YOIDH @nIp ut

IdN[/nJ Jo sweld jo Iap10 9y} JO
YOSCH I LT°0

‘OIS JU L100°0 Ul 191]/ng Sw g1

UOT)BULIOJUI ON
I9)BM UI 9[qN[O]
SPIOB 9)N[IP PUB I9)BM UL 9[QN[OS

17, 39 ‘suotl Surxeiduroo
opruonyg jo doussaxd ur 9qniog

9180ZUd(

938[BX()
9)8U0QIB))
a)eydsoyq
9)B8WIOIY))
9YBJINS
9YBINN
9YBIO[YIIS]
938po]
9yBswIoIg
oprwoag
OPLIOTYD)

opLIon|

 

 

 

andg

 

Aaind

 

mnd

 

 

TENLVEAdWE ], WOOY LV SANNOdWO)) WAINOLATJ 40 XLITIENTOS @HT,

1-¢ 414V ],

 
3-3] RADIATION EFFECTS 101

which compounds of plutonium appear most worthy of experimental study
as fuel solutes. Table 3-1 summarizes the available low-temperature solu-
bility information for three valence states of plutonium in the presence of
different anions. |

Limited experimental work has been performed in which the solubilities
of plutonium carbonates, sulfates, and phosphates have been determined at
temperatures up to 300°C [27]. No substantial solubilities have been
established at temperatures above 200°C.

Protactinium. No efforts have been made to achieve high solubilities
of protactinium in order to use it as a component of reactor fuel solutions.
Rather, the chemistry of protactinium has been examined in order to
devise processes for removing Pa232 continuously from thorium breeder
blanket systems. A project was undertaken by the Mound Labora-
tories [28] to separate gram quantities of the longer-lived Pa?3! which
could be used in studies of the chemistry of protactinium.

Considerable information concerning the low-temperature chemical be-
havior of Pa has accumulated as a by-product of the development of
chemical processes for the separation of U233 from irradiated thorium
materials [29].

Neptunium. Np239 is in a class with Pa233; no efforts have been made to
use it as a fuel solute, but consideration has been given to its formation in
and removal from blanket solutions of U238 [30a]. The chemistry of nep-
tunium has been reviewed by Hindman et al. [30b], and the hydrolytic
behavior has been reviewed by Kraus [30c]. Continuous separation of
Np23? would provide a Pu?3? product of high purity by radioactive decay,
whereas plutonium recovered from long-term irradiation of U238 usually
contains appreciable amounts of Pu240, Spectrophotometric cells for use
at elevated temperatures and pressures in the study of the chemistry of
neptunium (and other materials) have recently been developed by Wag-
gener [30d] and have been used to measure the absorption spectra of
dilute neptunium perchlorate in its six-, five-, four-, and three-valence
states, using heavy water as the solvent. Dilute solutions of neptunyl
nitrate in nitric acid have been so studied at temperatures up to 250°C; the
pentavalent state was found to be stable under the test conditions [30e].

3-3. RapiaTioNn ErreEcTs*

.3=-3.1 Introduction. Any aqueous reactor fuel solution will be subjected
to intense fluxes of high-energy radiations. The action of these radiations
both on the water and on the solute is of considerable importance in

*Taken from material prepared by C. J. Hochanadel fo‘r The Reactor Handbook.
102 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

reactor design and operation. Energy will be dissipated in a fuel solution
by the stopping of fast charged particles. These include mainly the fission
recoil particles, the recoil particles such as protons produced by elastic
neutron scattering, and the fast electrons resulting from the absorption of
gamma rays and from the decay of radioactive fission products. The
extent to which each contributes to the total energy absorbed in the fuel
solution depends upon the design of the reactor and the composition of
the solution. -

Water is decomposed by all types of high-energy radiations to give
hydrogen, hydrogen peroxide, and oxygen [31]. If the decomposition
products are confined in solution, a radiation-induced back reaction will
occur and, eventually, steady-state concentrations (pressures) of products
will be attained. The rate of decomposition, the rate of the back reaction,
and hence the steady-state concentrations are sensitive to the conditions
of the system, such as the nature of the radiation, the type and concentra-
tion of solutes present, and the temperature. In particular, the addition of
hydrogen suppresses the decomposition of pure water.

The solutes may also be acted upon by direct absorption of the energy
of the radiations (or by transfer of energy from the solvent) and also by
reactions with the intermediate reactive species produced by the decompo-
sition of the water. |

3-3.2 Primary and secondary reactions in pure water. The fast charged
particles give up energy to the electronic systems of the molecules of the
medium, thereby producing various excited and ionized states. In liquid
water, the ionized and excited molecules are rapidly transformed into the
free radicals H and OH. These are formed in relatively high concentrations
along the tracks of the fission recoils or other charged particles. As a
result, many of the radicals combine before they can diffuse apart, thereby
producing the stable molecules H20, Hz, and H202. The primary chemical
species are therefore considered to be H, OH, Hg, and H2Og2; their yields
per 100 ev of energy absorbed are expressed as G(H), G(OH), G(Hz2), and
G(H202). The primary chemical reaction can be written:

3 HzO —>H —|— OH —|— H2 + H202. (3—1)

Some minor subtleties emerging from recent studies of the radiolysis of
aqueous solutions are: (a) although stoichiometry demands that G(H) +
2G(Hz2) = G(OH) 4 2G(H202), the yields of H and OH and also the
yields of H2 and H202 are not necessarily equal to each other [32]; (b) the
yields of He and H202 are lowered by solutes which scavenge the pre-
cursors in the particle tracks [33]; (¢) HO2 may be another “primary”’
chemical species produced in small yield in the particle tracks [34].
3-3] - RADIATION EFFECTS 103

 

v Wv ¢ v v v
?‘. Wy
5 b3 5 v 3 o7
40} >22 2 3 3 29 —20
o BEEE R ¢ £
— N~ ™M v o
< G(H) G(Hg) —
30 —41.5
>
Q
: g
g
o °
S 20— —{10 E
T N
= I
© o
1.0 —0.5
Se 1=
-------- ::

 

 

 

0.1 1.0 10 102 103 104 109

Initial Linear Energy Transfer,
kev/micron

F1a. 3-17. Yields of atomic and molecular hydrogen from the decomposition of
water by various ionizing radiations.

The ylelds of the primary chemical species depend markedly on the type
of radiation or, more specifically, on the energy transferred to the solution
per unit length along the track of the charged particle. The linear energy
transfer (LET) parameter varies from 5 X 10* kev per micron of path for
fission recoils to 0.2 kev/micron for fast electrons. The yields G(Hz) and
G(H202) are largest for radiations such as fission recoils with large LET,
while the yields G(H) and G(OH) are largest for radiations such as fast
electrons with small LET [31]. This is illustrated in Fig. 3-17, where the
yields [35] G(H2) and G(H) are plotted as a function of LET.

The free radicals which escape immediate combination and diffuse into
the bulk of the solution may react with solutes present, including the
H2 and H202. In water with no added solutes, the principal back reactions
of the free radicals are believed to be:

H + H202 —- Hzo + OH, (3—2)
104 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

OH+ Hs; — H,0+ H, (3-3)
OH + H20; —> H20+ HOs, (3-4)
2HO5; —> H3032 + Oq, (3-5)

H+ 02 — HO3, (3-6)

HOz + H02 — H204 OH + O.. (3-7)

Reactions (3—2) and (3-3) provide a chain mechanism for the back reaction
of Ho and H2032 to reform water [36], thereby leading to steady-state con-
centrations of decomposition products. The steady-state concentration
will depend on the relative yields of molecular products and free radicals
in reaction (3-1). For gamma rays, which produce the free radicals in
high yield and the molecular decomposition products in low yield, the
steady state in pure water is essentially zero. For fission recoils, which
produce essentially no free radicals to promote the back reaction, the
steady-state concentration (pressure) is very high (several thousand psi).
Reactions (3—4) and (3-5) provide a mechanism for decomposing H202 to
02, and reactions (3-3), (3-5), (3-6), and (3-7) provide a mechanism for
combining Hz and O2 to form water at higher temperatures [37].
Dissolved materials may be oxidized or reduced. In general, H atoms
usually reduce the solute and OH radicals reoxidize it. Assuming equal
numbers of H and OH, the net result depends on the action of the H2053.
The peroxide may act in either way (depending on the oxidation-reduction
potentials) but usually oxidation is favored. In the presence of Og, the
H-atom may be converted to HO2, which usually acts as an oxidizing agent.

3-3.3 Decomposition of water in uranium solutions. In Table 3-2 are
listed the hydrogen yields from the decomposition of solutions of various
uranyl salts [38]. The yield depends on the type of radiation and on the
solute concentration, but is independent of temperature. Figure 3-18
shows how the yield of hydrogen produced by fission recoil decomposition,
and by gamma-ray decomposition, decreases with increasing uranium con-
centration. This decrease may result from scavenging of H-atoms in the
particle tracks by the uranyl ions. Decomposition by fission recoil particles
produces mostly Hz (and an equivalent amount of H2O2 plus Og); the
yields G(H) and G(OH) are very small, probably in the range 0 to 0.1
per 100 ev.

In an aqueous homogeneous reactor fuel solution, the water is decom-
posed by fission recoils, proton recoils, and fast electrons. The rate of
hydrogen formation, in moles per liter per minute, can be expressed by
the equation:
3-3] RADIATION EFFECTS 105

 

V | I |

  

Q& 98% Fission Recoil Energy

  

G(H9) , molecules/100 ev

0.5 —

v — Rays

 

 

| | | ]
0 1 2 3 4
Concentration, m |

 

Fi1a. 3-18. The effects of uranium concentration and type of radiation on the
initial Hz yield from irradiated UO2SO4 solutions.

d(H?2)

=2 = 0.00622[Gy X Wy+ Gy X Wy+ G, X W), (3-8)

where G is the hydrogen yield in molecules per 100 ev absorbed; and W is
the power density in kilowatts per liter. The subscripts f, p, and e refer to
the values for fission recoil particles; protons produced by neutron scatter-
ing, and electrons produced by gamma-ray absorption and by radioactive
decay of fission products. For an operating homogeneous reactor, about
969, of the hydrogen gas produced in solution is due to the fission recoil
particles, 29, to the neutrons, and 29, to the gamma rays. Therefore the
last two terms in Eq. (3—8) are usually neglected. The fraction due to
recoils is usually above 0.96, since part of the energy of the prompt neu-
trons, gamma rays, and radioactive decay escapes from the solution. The
value of Gy for a given solute concentration can be obtained from Fig. 3-17
and the value for W, can be calculated from the neutron flux, the concen-
tration of fissionable atoms, the fission cross section, and the kinetic
energy of the fission recoils.

Along with the hydrogen, an equivalent amount of oxidant (either
peroxide or O2) will be formed.
106

TABLE 3-2

PROPERTIES OF AQUEOUS FUEL SOLUTIONS

[cHAP. 3

INITIAL RATES OoF H>o GAs PropUcTION FROM REACTOR-IRRADIATED

URANIUM SOLUTIONS

 

 

 

 

 

 

 

 

 

 

 

 

 

Concentration Fission ener
Solute o oS pH | G(HY)
g U/liter | g U/liter | O ©PETEY

0.399 0.372 0.688 1.61

4.03 3.76 0.957 3.26 | 1.66

18.6 1.63 0.906 2.90 | 1.48

38.1 0.274 0.619 0.95

40.7 37.9 0.995 2.42 | 1.53

102.1 37.4 0.995 2.00 | 1.35

105.2 38.9 0.995 0.10% | 1.20

108.4 40.1 0.995 1.35

202.3 0.063 0.273 0.69

V0,804 202.5 37.6 0.995 1.61 | 1.11
203.4 189.6 0.999 1.11

227.0 1.63 0.906 0.98

310.4 0.096 0.364 0.62

386.0 1.63 0.906 0.80

431.3 37.8 0.995 1.32 | 0.77

436.8 3.10 0.949 0.73

477.2 0.148 0.467 0.56

713.5 33.5 0.995 0.56

796.0 37.4 0.995 1.03 | 0.49

4.25 3.96 0.959 4.25 | 1.63

40.1 37.3 0.995 3.32 | 1.58

118.8 37.1 0.995 2.98 | 1.36

UO,F; 272.0 37.2 0.995 2.64 | 1.11
377.0 39.3 0.996 1.35% | 0.84

405.7 42.3 0.996 2.41 | 0.95

4.24 3.95 0.960 1.63

U05(NO3)2 42.3 39.4 0.995 2.05 | 1.5
318.0 2.29 0.932 1.03 | 0.6

420.1 36.9 0.994 0.60 | 0.55

42.2 39.3 0.995 1.95 | 1.45

U(S04)2 92.5 35.1 0.996 0.1 | 1.25
350.0 32.0 0.995 0.1 | 0.75

 

*pH adjusted by adding acid.

 
3-3] RADIATION EFFECTS 107

3-3.4 Recombination in uranium solutions. For fission recoil particles
the radiation-induced back reaction of Hz, Oz, and H202 is relatively
slow. Also, the thermal recombination rate in the absence of added catalyst
is relatively slow and, as a result, the steady-state pressure of gases is very
high; e.g. at 250°C the pressure is of the order of thousands of psi.

Certain solutes, notably copper salts, have been shown to act as homo-
geneous catalysts [39] in the thermal combination of hydrogen and oxygen
in aqueous uranium solutions. This provides a convenient method for re-
combining the radiolytic hydrogen and oxygen gases.

The reaction rate is first order in hydrogen and in copper concentration,
and independent of the oxygen concentration. The rate-determining step
is the reaction of hydrogen with the catalyst, and the activation energy
is about 24 kcal/mole. For a particular uranium solution, the rate of hy-
drogen removal, in moles/liter/min, can be expressed by

—d(Hz)
dt

 

= kcu(Cu)(Hs), (3-9)

where k¢, is the catalytic constant in liters/mole/min, and concentrations
are given as moles/liter. Some selected values of kg, are listed in Table 3-3.
Increasing the concentration of uranyl sulfate or of sulfuric acid decreases
the catalytic activity of the copper somewhat, possibly as a result of com-
plexing. Also, the rate of reaction with D3 is about 0.6 that with Ho.

TABLE 3-3

SELECTED VALUES OF kcy AT SEVERAL TEMPERATURES AND
UraNIUM CONCENTRATIONS (Cu=10"3 M)

 

 

 

Uranium Temperature, kcy, 103 kcyu/S,
concentration, M °C liters/mole/min psi—!/min
0.17 190 4.3 0.28
0.17 220 26.6 2.3
0.17 250 90.0 12.
0.00 250 83.0* 11.
0.01t00.1 250 | 133 18.
0.01t00.1 275 380 61
0.01t00.1 - 295 850 149

 

 

 

 

 

 

*With 1073 M Cu(ClO4)2 plus HCIO4 in concentrations ranging from 0.005 M
to 0.05 M. <
108 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

The concentration (solubility) [40] of Hs is related to the partial pres-
sure of hydrogen by the proportionality factor

_PH2

S=m)’

(3-10)
where Pg, is given in psi, (Hz) in moles/liter, and S in psi/liter/mole.
Equation (3-9) can then be written

—d(H?2)
dit

Pr,

3 (3-11)

 

At the steady state, the rates of formation and removal of hydrogen are

equal, and from Eqgs. (3-8) and (3-11), the steady-state pressure is given by

P _0.0062 X Gy X WfXS.
(Hg)ss =— (Cu) kCu

 

(3-12)

Application of copper sulfate catalyst to the suppression of gas evolution
during the operation of the Homogeneous Reactor Experiment was dis-
cussed by Visner and Haubenreich [41]. Design calculations for use of
copper catalysts in the HRE-2 and in other reactors have been reported
[42]. The use of internal recombination catalysts in homogeneous reactors
is also discussed in Article 7-3.7. |

3-3.5 Peroxide decomposition in uranium solutions. The hydrogen per-
oxide produced by decomposition of the water can undergo several sec-
ondary reactions: (a) It can react with uranyl ion to form the slightly
soluble peroxide UO4 according to the reaction

U02++ + H:0. '.4__)'_ UO4++ 2H*. (3—13)

The UOQ4 will precipitate if its solubility (~1072 M) is exceeded. (b) It
(or the UO4) can decompose by a radiation-induced reaction via the free
radicals H and OH produced by decomposition of the water. (c) In a re-
actor operating at high temperatures the H2O2 will decompose thermally
at an appreciable rate according to the over-all reaction:

H;0; — H,0 + 1/20, (3-14)

and the UO4 will decompose thermally according to the reactions
UO04 —> UO3 + 1/202, (3-15)
UO3+ 2H* — U0+t + HO. (3-16)
3-3] RADIATION EFFECTS 109

Studies of the kinetics of the thermal decomposition of peroxide in uranyl
sulfate solutions [43] have shown the rate to be first order with respect to
peroxide concentration in the range from 0.4 to 5 X 1073 M, independent
of uranium concentration in the range from 4.5 X 1073 to 0.65 M, and in-
dependent of the acidity from pH 1.6 to 3.3. Traces of certain ions showed
pronounced catalytic effects. The rate of decomposition could be ex-
pressed by

—d(H202)

o = k(H202) + Keat(Cat) (H205), (3-17)

where (H202) represents total peroxide concentration (UO4+ H203) in
moles per liter at time ¢, k is the molar rate constant in the absence of
catalyst, ket is the catalytic constant, and (Cat) is the concentration of
catalyst in moles per liter.

Values of k for uranyl sulfate solutions (4.5 X 1073 M to 0.65 M) with
no added catalyst depended upon the adventitious impurities present, and
were in the ranges as listed in Table 3—4. The indicated activation energy
is 25.5 kcal/mole. For pure water, the rate constant at 78°C was 4.5 X 104
per minute.

Catalytic constants, kcat, for various ions added are listed in Table 3-5.

The net rate of peroxide formation is the difference between the rate of
production and the rate of decomposition. At the steady state the two rates
are equal. The rate of formation of peroxide in an operating reactor in
terms of moles per liter per minute can be expressed in terms of the yield,
(, in molecules per 100 ev of fission recoil energy, and the average fission
recoill power density of the reactor, W, in kilowatts per liter, by the
equation:

d(H202)

= 0.0062X G X W X 0.96. (3-18)

The maximum value of G for the particular solution used is that given
for the Hy yield in Fig. 3-18. During reactor operation the radiation-
induced decomposition of peroxide is negligible, and the rate of decom-
position is essentially the thermal rate given by Eq. (3-17). At the steady
state the peroxide concentration is given by

0.0062 X G X W X 0.96

(H202)88 = k + kca,t(Cat)

 

(3-19)
110 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

TABLE 3-4

MoLAR RATE CONSTANTS FOR PEROXIDE
DecoMPosITION IN UO2S0O4 SOLUTIONS

 

 

 

 

 

 

 

Temperature, °C ~ k, min™!
53 | 1.2X1073to0 12X 10~3
78 1.8X 1072 to 14 X 10—2
100 1.8% 1071 to 12 X 1071
TABLE 3-5

CataLytic ConsTANTs AT 100° C FOR PEROXIDE
DecomMprosITION BY VARIOUS JoNs ADDED TO
U02504 SorLuTIiON

 

 

 

 

Catalyst keat (liters/mole/min)
Fet2 135,000
Rut4 121,000
Agtl 4,500
Nit2 2,600
Cut? 1,200
Fet2 (promoted 502,000
by 793 ppm Cut2)

 

 

 

The maximum allowable power density, Wmax, before precipitation of
uranium peroxide occurs is given by

_ (HzOz)SOl rk + kcat(Cat)]. —
Wmax = "—03062 X G X 0.96 (3-20)

For example, in 0.17 M UO02S03 solution at 100°C with no added catalyst,
the peroxide solubility is =4 X 1073 M, k =~ 1 min~}, and G = 1.5, then
Wmax = 0.4 kw/liter.

Following reactor shutdown, peroxide formation and decomposition will
result from the delayed neutrons and from the 8~ and vy radiation of the
fission products. The yield for peroxide formation will be essentially that
for y-rays, G = 0.46. The yield for radiation-induced decomposition [37]
may be as high as 4.5, but will depend in a complicated way on the amount
of oxygen, hydrogen, and other solutes such as fission products, corrosion
products, etc., present.
3-4] PHYSICAL PROPERTIES 111

3-3.6 Decomposition of water in thorium solutions. Under radiation
thorium nitrate solutions decompose [44] to give Hs, H202, and O2 from
decomposition of the water, and Oz, N, and oxides of nitrogen from de-
composition of the nitrate. Yields of H2 and N2 for several types of radia-
tion, and for several concentrations of thorium nitrate, are given in
Table 3—6. The hydrogen yield decreases with increasing solute concen-
tration, the same as for uranium solutions. The nitrogen yield increases
with increasing nitrate concentration. The N2 is presumably formed by
direct action of radiation on the nitrate. The N3 yield is greater for radia-
tions of greater LET. The N2 yield is independent of temperature, and
little or no radiation-induced back reaction takes place.

Uranyl nitrate solution also decomposes to give N3z in yield comparable
to that for thorium nitrate solution.

TABLE 3-6

THE ErrEcTs OF CONCENTRATION AND TYPE OF RADIATION ON
THE YIELDS OF N2 AND Hs IN THE DECOMPOSITION OF THORIUM
NITRATE SOLUTIONS

 

 

 

 

G(N
G(H,) (Ng)
Th(NOs)4, Fission
molality ok Fission ORNL graphite
recoils ) . . .. Gamma rays
recoils* pile radiation
0.26 1.11 0.002 0.003 0.04 x 10~3
0.55 0.93 0.016 0.003 0.5 x10°3
1.5 0.51 0.047 0.005 1.5 X 1073
2.7 0.33 0.063 0.006 1.1 x10°3
7.2 ~ 0.08% 0.16

 

 

 

 

 

 

 

*0.14 molal enriched UO2804 added to make the energy contributed by fission
recoils > 95%, of the total energy absorbed.
TEstimated by extrapolation.

3—4. PHYSICAL PROPERTIES

3—4.1 Introduction. Knowledge of the physical properties of aqueous
solutions of reactor fuel materials is required for nuclear physics calcula-
tions and analysis of reactor performance, for engineering design, and,
ultimately, for effective reactor operation. The scarcity of information
availlable in 1951 concerning the properties of uranium salt solutions
prompted the Homogeneous Reactor Program at ORNL to sponsor a
physical properties research program at Mound Laboratory beginning in
112 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

July 1951 and carrying through December 1954. The progress of this
effort is discussed in regular reports [45] and in a number of topical reports
dealing with techniques, apparatus, and summarized data [46-56].

A number of compilations of physical properties data for aqueous reactor
solutions have appeared, among which are included those of Van Winkle
[57], Tobias [58], and sections by Briggs, Day, Secoy, and Marshall in
The Reactor Handbook [59].

The properties of light and heavy water are discussed by Lottes [60] as
they relate to reactor heat-transfer problems. Other properties of water
are found in standard reference works [61-62].

The remainder of this section is devoted to particular properties of
reactor solutions which are of interest and to some properties of the vapor
phase above reactor solutions which are important for aqueous homo-
geneous reactors.

 

 

 

 

TABLE 3-7
Liquip AND Varor DEeNsITIES OF D2O
Density, g/cc

T, °C

Vapor Liquid
175 0.004 0.989
180 0.005 0.983
190 0.006 0.970
200 0.007 0.957
210 0.009 0.943
220 0.010 0.929
230 0.013 0.913
240 0.016 0.898
250 0.020 0.881
260 0.024 0.864
270 0.029 0.847
280 0.034 0.829
290 0.040 0.809
300 0.048 0.787
310 0.058 0.763
320 0.070 0.735
330 0.087 0.705
340 0.105 0.668
350 0.129 0.626
360 0.163 0.573
370 0.248 0.462
371.5 0.363 0.363*

 

 

 

 

 

*Critical point.
3-4] PHYSICAL PROPERTIES 113

3—4.2 Density of heavy water and uranyl sulfate solutions. The density of
heavy water has been measured up to 250°C at Mound Laboratory by a
direct method making use of a Jolly balance [51,54]. An indirect method
for determining the density of heavy-water liquid and vapor has been used
to extend the data up to the critical temperature [63]. Table 3-7 gives
density values at convenient temperature intervals. The densities of
uranyl sulfate solutions from 20 to 90°C and at concentrations up to 4.0
molal were measured by Jegart, Heiks, and Orban at Mound Laboratory
[47,56]. The densities of uranyl sulfate solutions were measured by
Barnett et al., of Mound Laboratory [55] at temperatures up to 250°C for
concentrations (at room temperature) of 60.6 and 101.0 g U/liter of solu-
tion in light water and for uranium concentrations (room temperature)
of 20.3, 40.4, and 61.2 g/liter in heavy water. Their data are presented in
Table 3-8.

TABLE 3-8

DENSITIES OF LigHT- AND HEAVY-WATER SOLUTIONS OF
URANYL SULFATE

 

 

 

 

 

Density, g/ml
7 oc | U02804 in H0, g U/liter U02804 in D20, g U/liter
60.6 101.0 20.3 ()?; 40.4 ()?; 61.2
psl psl
30 1.1340 | 1.300 1.1567 1.1842
1.318 | 300 | 1.1598 | 280
45 1.0715 1.1275 | 1.1245 1.1509 1.1781
1.1263 | 305 | 1.1540 | 280
60 1.0649 1.1199 | 1.1170 1.1433 1.1704
1.1198 | 325 | 1.1472 | 300
75 1.0562 1.1111 1.1120 | 340 | 1.1388 | 325 | 1.1628
90 1.0468 1.1008 | 1.1026 | 360 | 1.1297 | 350 | 1.1545
100 1.0395 1.0952 | 1.0954 | 380 | 1.1228 | 365 | 1.1476
125 1.0203 1.0742 | 1.0751 | 415 | 1.1037 | 410 | 1.1278
150 0.9984 1.0528 | 1.0505 | 480 | 1.0822 | 480 | 1.1053
175 0.9735 1.0293 | 1.0221 | 560 | 1.0572 | 530 | 1.0805
200 0.9460 1.0030 | 0.9920 | 680 | 1.0280 | 670 | 1.0540
225 |  0.9156 0.9745 | 0.9578 | 830 | 0.9973 | 840
250 0.9440 | 0.9224 | 1090 | 0.9610 | 1080

 

 

 

 

 

 

 

 

 

 

*Overpressure of oxygen gas as indicated. Otherwise solutions were in contact
with their own vapor.
114 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

The densities of light-water solutions of uranyl sulfate were found by
Marshall [64] to fit the relationship

. 1
~ (78.65/U) — 1.046

 

d, + dn,0,

where d, is the density of the solution in g/cec, U is the weight percent
uranium, and dg.0 is the density of water at the same temperature and
pressure as the solution. The accuracy of this formula is believed to be:

4+ 0.59, from 25 to 300°C for 0-109, U,

+ 1.09,, from 120 to 250°C for 10-509, U,
4 2.09, from 25 to 125°C for 10-509, U,
+ 2.09, from 250 to 280°C for 10-509, U.

The density of heavy-water solutions of uranyl sulfate may be estimated
from an analogous formula:

dy= 1
*= (71.0/0) — 0.944

 

-+ dp,o0.

The densities of heavy-water solutions of uranyl sulfate reported in Table
3-8 were found to be within 39, of values calculated by this formula.
Similar formulas were devised by Tobias [58] for application to the
determination of the densities of solutions of mixed solutes such as uranyl
sulfate-beryllium sulfate and uranyl sulfate-lithium sulfate.
Density information for other uranium salts and for thorium nitrate so-
lutions was compiled by Day, Secoy, and Marshall [59].

3—4.3 Viscosity of DO and uranium solutions. The viscosity of heavy
water was measured from 30 to 250°C by Heiks et al. [564]. Good agree-
ment with four values reported by Hardy and Cottington [65] was found.
The apparatus used has been described by Heiks et al. [51] and the elec-
tronic instrumentation for measuring the time of fall of a plummet con-
taining a radioactive pellet has been described by Rogers et al. [52].

Van Winkle [57] discussed the viscosity of light and heavy water and
the early data for solutions of uranyl sulfate in heavy water.

Heiks and Jegart of Mound Laboratory [50,56] measured the viscosity
of uranyl sulfate solutions in light water over a concentration range of
0.176 to 2.865 molal and over a temperature range of 20 to 90°C by the use
of Ostwald capillary viscometers. Using the falling-body viscometer re-
ferred to above, Barnett et al. of Mound Laboratory have measured the
viscosities of light- and heavy-water solutions of uranyl sulfate at tem-
peratures up to 250°C [55]. Table 3-9 presents a comparison of viscosities
3-4]

ViscosiTiES OF LicHT- AND HEAVY-WATER SOLUTIONS OF

PHYSICAL PROPERTIES

TABLE 3-9

URANYL SULFATE

115

 

T °C

Viscosity, cp

 

UO02804 in H20, g U/liter

U02804 in D20, g U/liter

 

20.2

59.7

100.4

201.4

301.1

400.7

0

20.3

40.4

61.2

 

 

30
45
60
75
90
100
125
150
175
200
225
250

0.857
0.629
0.492
0.403
0.329
0.292
0.238
0.194
0.164
0.142

0.946
0.701
0.536
0.436
0.365
0.327
0.264
0.213
0.178
0.153
0.136
0.125

1.04

0.773
0.605
0.483
0.399
0.351
0.273
0.223
0.188
0.158
0.141
0.130

1.40

1.05

0.820
0.638
0.517
0.454
0.347
0.275
0.226
0.192

1.97
1.39
1.06
0.821
0.675
0.589
0.442
0.335
0.269
0.230
0.209
0.190

S~ ® &

3.
1.
1.
1.
0.

R = O O

84
0.767
0.543
0.437
0.360
0.298
0.264
0.238

0.969
0.713
0.552
0.445
0.365
0.323
0.252
0.208
0.175
0.151
0.135
0.124

0.773
0.586
0.467
0.384
0.342
0.263
0.216
0.182
0.154
0.138
0.125

0.789
0.608
0.487
0.397
0.350
0.269
0.222
0.187
0.160
0.144
0.133

0.827
0.638
0.513
0.415
0.365
0.287
0.233
0.196
0.169
0.151
0.137

 

 

 

 

 

 

 

 

 

 

 

 

 

of light- and heavy-water solutions for selected temperatures and uranium
concentrations.

The viscosity of uranyl nitrate solutions is discussed by Day and
Secoy [59].

3—4.4 Heat capacity of uranyl sulfate solutions. Van Winkle [57] esti-
mated the heat capacity of dehydrated uranyl sulfate by comparison with
uranyl nitrate and with salts of other metals and has estimated the heat
capacity of solutions of uranyl sulfate in heavy water and in light water.
The effect of temperature on the heat capacity of solutions was assumed
to be the same, percentagewise, as the effect on the heat capacity of the
pure solvent. Table 3—10 shows the influence of uranyl sulfate concentra-
tion upon the heat capacity of solutions at 25 and 250°C. No experimental
measurements have been reported for temperatures above 103°C; values
below this temperature differ from the estimates by as much as 109, [57].

3-4.5 Vapor pressure of uranyl sulfate solutions. The vapor pressures
of uranyl sulfate solutions have been measured over the temperature range
24 to 100°C and at uranium concentrations up to 4.8 molal by Day [66].
The ratio of the vapor pressure of the solution to that of pure water at the
116 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3
TABLE 3-10
EsTiMATED HEAT CAPAcCITIES OF URANYL SULFATE SOLUTIONS

Estimated Heat Capacity, cal/g -°C

 

 

 

 

| H20 solutions D20 solutions
UOzSO4, W/O
25°C 250°C 25°C
0 0.998 1.166 1.005
10 0.905 1.070 0.916
20 0.809 0.975 0.830
30 0.714 0.875 0.745
40 0.619 0.770 0.658
50 0.523 0.665 0.568
60 0.428 0.550 0.474
70 0.333 0.425 0.370
80 0.238 0.290 0.252
85.9 (UO2504 - 3D20) — ' — 0.174
87.2 (UO2804 - 3H20) 0.170 — —

 

 

 

 

 

 

same temperature is almost independent of temperature but markedly de-
pendent upon the solute concentration over this range. Table 3-11 shows
the effect of uranium concentration upon the ratios at temperatures up to
100°C. Vapor pressure measurements have been made for light-water
solutions of uranyl sulfate at temperatures up to approximately 200°C and
uranium concentrations (room temperature) of 400 and 500 g/liter. Data
have, however, not yet been published [45]. Other vapor pressure meas-
urements are being made at the Oak Ridge National Laboratory. The
vapor pressure lowering in uranyl sulfate solutions containing added
lithium or berylllum sulfate was stated to be approximately in accord
with Raoult’s law [8b].

The vapor pressures of three solutions of UO3 in phosphoric acid are
shown in Table 3-12 [17], based on values read from the published curve.

3—4.6 Surface tension of uranyl sulfate solutions. The capillary rise
method was adapted to conditions prevailing at elevated temperatures and
pressures for measurements made at Mound Laboratory by Heiks et al.
[51,54]. Briggs [59] has compiled information on the surface tension of
aqueous solutions of uranyl sulfate, including the relationships established
by Van Winkle.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3-4] PHYSICAL PROPERTIES 117
TABLE 3-11
RATios oF VAPOR PRESSURES OF URANYL SULFATE
SOLUTIONS TO THAT OF Pure H-0
p/po (at given temperature)
U02804, M
25°C 50°C 80°C 100°C
0.59166 0990 | 0.990 | 0.991 | 0.993
1.14545 .979 .981 .984
3.17433 .911 .920 .935
4.1821 .863 .881 .900
TABLE 3—-12
Varor PreEssUrRes oF UOsz SorLutions IN HzPO4*
Vapor pressure, 1b/in?
Temperature,
°C 0.76 M UO3in | 0.309 M UO; in 0.75 M UO3 in
0.56 M H3PO4 2.90 M H3PO, 7.5 M H3PO4
110 23 21 21
150 66 64 04
200 200 200 200
250 460 500 500
300 1000 1150 1150
350 1700 2100 2100
400 2900 3600 3700
450 4600 5600 6600
500 6600 8000 —

 

 

 

 

 

 

*Plotted and extrapolated from Los Alamos data.

Barnett and Jegart of Mound Laboratory [53] reported the surface
tension of uranyl sulfate solutions in light water at temperatures up to 75°C
and concentrations up to 2.34 molal.

Barnett et al. [55], using the apparatus described by Heiks et al. [51]
for capillary rise technique at elevated temperatures and pressures, meas-
ured the surface tension of light- and heavy-water solutions of uranyl
sulfate at temperatures up to 250°C.
118 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

TABLE 3-13
Tue pH orF UO3-H2S04-H20 SoLuTioNs AT 25.00°C*

 

Molar ratio, UO3/SO4

 

 

 

Molarity
H.S0
v 0.7006 | 0.8017 | 0.9024 | 1.0029 | 1.1042 | 1.202
0.02432 1.995 | 2.151 | 2.425 | 3.043 | 3.612 | 3.823
0.01214 2250 | 2.422 | 2.692 | 3.267 | 3.731 | 3.922

0.004866 2.619 2.784 3.046 | 3.526 | 3.874 4.038
0.003656 2.730 2.900 3.155 3.604 3.913 4.074
0.002437 2.905 3.065 3.321 3.708 3.977 4.124
0.001216 3.174 3.348 | 3.592 3.906 4.093 4.214
0.0004850 3.587 3.715 3.903 4.130 | 4.250 4.354
0.0003646 3.671 3.822 3.996 4.201 4.304 4.395
0.0002424 3.851 4.004 4.127 4.326 4.411 4.500

 

 

 

 

 

 

 

 

 

*These values are based on a determined pH of 2.075 for 0.01000 molar HCI,
0.900 molar KCl. For pH’s relative to the analytical concentration of HCI subtract
0.075 pH unit from each of the above values.

TABLE 3-14

TaE pH Varues or HCl, H2S804, AND UO2S04 SoLuTioN
AS A FuNctioN oF TEMPERATURE

 

 

 

M t°C pH t°C pH

HCl 0.0100 30 2.13 130 2.17
H2SO04 0.00490 30 1.76 180 1.93
” 0.04615 30 1.28 180 1.62

” 0.09740 30 0.96 180 1.21
UO028047 0.04 30 2.80 150 2.58
» 0.17 30 2.20 180 2.20

» 1.1 30 1.76 180 1.81
U02504* + H2804 30 1.22 180 1.83

 

 

 

 

 

 

 

 

*UO02504 solution which read pH 2.68 at 30°C, to which was added sufficient
H2S04 to bring the pH reading down to 1.22 at 30°C.

TReported values modified and extended by personal communication with M. H.
Lietzke, April 1958. |
34] PHYSICAL PROPERTIES 119

 

 

 

 

22 l I T l I T T
Helium
20 —]
18 - —
.g Nitrogen
6 _
o
o
£
5
‘s 14 —
o
ms
t
2
b ]2 — —
v Hydrogen
t
S
¢ 10 _
o
O
3
S 8
“ " —
> Oxygen
c
o
I
6 ]
4 —
Xenon
2 —
| | ] ] | | |
0 40 80 120 160 200 240 280 320

Temperature, °C

F1c. 3-19. Solubility of gases in water.

3-4.7 Hydrogen ion concentration (pH). Orban of Mound Laboratory
[48,56] studied the acidity of uranyl sulfate solutions from 25 to 60°C.
Orban has also reported the pH values for uranyl sulfate solutions con-
taining excess UO3 at temperatures up to 60°C [49,56]. Secoy has com-
piled information concerning the effects of uranium sulfate concentration
upon pH as reported by a number of investigators [59]. Marshall has
utilized the relationship between pH and concentration to determine the
solubility of UO3 in sulfuric acid at elevated temperatures [7]. Table 3-13
shows the effect of sulfate concentration on pH at 25.00°C for various
ratios of UO3 to sulfate as found by Marshall.

The direct measurement of pH at elevated temperatures and pressures
has been made possible through the development by Ingruber of high-
temperature electrode systems for use in the sulphite pulping process of
the paper industry [68]. Lietzke and Tarrant have demonstrated the
applicability of Ingruber’s electrode system to the measurement of the pH of
 

    
   
 

 

 

 

 

 

260 | | | I [ | 1 |
240 — T = Initial Temperature | |
220 |— 'l _
2 200 [— l', —
o
£ 180 |- /Illl |
3 2
2 160 |— // /) _
s 140 |— 2
@ / [ 5
5 I S I 2
§ 120 T - / II Zo -
Q.
5 M AN
% 80 H / ’I _—
= 1 /1
£ 60 HH / /l —
~\\\ T =350°F < /1
40 —‘\\ //// |
" E\\{\\\\, T =300°F ..-; / 3
0 10 20 30 40 50 60 70 80 90 100

Initial Mole % O2 in HyO Free Gas

Fi1a. 3-20. Approximate explosion limits of spark-ignited gaseous mixtures of O2
and H2 saturated with H2O vapor in 12-in. bomb.

solutions of HCl, UO2S04, and H2SO4 at temperatures as high as 180°C [69].
Table 3-14 lists the values reported. Lietzke has reported the properties
of the hydrogen electrode-silver chloride electrode system at high tempera-
tures and pressures [70].

3-4.8 Solubility of gases. The solubilities of various gases in water and
in reactor solutions are indicated by Fig. 3-19 [71,72].

Composition and PVT data [73,74]. Dalton’s Law of additive partial
pressures has been determined to correlate the P-PVT relationships of
steam-oxygen and steam-helium mixtures in the pressure range of interest
(up to 2500 psi) to within 197,

For approximate calculations the perfect gas laws can be applied when
dealing with the permanent gases under reactor conditions, but values for
the properties of water vapor and D20 vapor should be obtained from
experimental data or standard tables [62].

The compositions of saturated steam-gas mixtures are predictable in a
similar manner, in that the partial density of each constituent is equal to
that of the pure constituent at its partial pressure.

3-4.9 Reaction limits and pressures. Stephen et al. [75] measured the
effect of temperature and Hz/O2 ratio on the lower reaction limits of the
H—O2-steam system in a small 1.5-in.-diameter autoclave. Their investi-
34] PHYSICAL PROPERTIES 121

 

 
   

 

 

 

 

1T T T T T T T T T 1
— —
50 - / \ -
Reflected Detonation I \
- (Calculated) '
20 I' —
Detonation l
(Calculated)
10
£ |
~ b
* |
5
— Explosion | —
(Calculated) /
2 / Observed ]
/
/
-
//
"
1 ] ] | | | | ] ] | | ] ] |
0 12 24 36 48 60 72 84

Mole Percent Knallgas, Yj

Fic. 3-21. Ratio of peak reaction pressure to initial mixture pressure vs. com-
position of Knallgas saturated with water vapor at 100°C. Tube diameter = 0.957
in., spark ignition energy = 180 millijoules.

gation indicated that the composition (mol basis) of the mixture at the
lower limit was not temperature dependent at constant Hz2/O2. Figure 3—20
is a plot of Stephan’s data showing variation in Hz2/O2 [76].

Syracuse University is investigating [77] further the effects of geometry,
temperature, and method of ignition on the reaction and detonation limits
of Ho0—(2H2+ 02). Reaction pressures are also measured. Figure 3-21
shows their reported results in a long 0.957-in.-diameter tube at 100°C;
these are typical of results obtained at 300°C. It is interesting to note
that their lower reaction limit is 6%, Knallgas,* as compared with 20%
Knallgas observed by Stephan.

*Knallgas is the term describing the 2:1 mixture of hydrogen and oxygen ob-
tained electrolytically and considered as a single gas.
122 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

REFERENCES

1. C. H. SEcoy, J. Am. Chem. Soc. 72, 3343 (1950).

2. C. H. Secoy, Survey of Homogeneous Reactor Chemical Problems, in
Proceedings of the International Conference on the Peaceful Uses of Atomw Energy,
Vol. 9. New York: United Nations, 1956. (P/821, p. 377)

3. J. S. GiLL et al., in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending OCtober 31, 19563, USAEC Report ORNL-1658, Oak Ridge
National Laboratory, Feb. 25, 1954. (p. 87)

4. H. W. WriGHT et al., in Homogeneous Reactor Program Quarterly Progress
Report for the Period Ending Oct. 1, 19562, USAEC Report ORNL-1424(Del.),
Oak Ridge National Laboratory, Jan. 10, 1953. (p. 108)

5. W. L. MarsuALL and C. H. Secoy, Oak Ridge National Laboratory, 1954.
Unpublished.

6. E.V.JonNEs and W. L. MARsHALL, in Homogeneous Reactor Project Quarterly
Progress Report for the Period Ending Mar. 15, 1952, USAEC Report ORNL-1280,
Oak Ridge National Laboratory, July 14, 1952. (p. 180)

7. W. L. MaRrsHALL, The pH of UO3s-H2804H20 Mixtures at 25°C and Its
Application to the Determination of the Solubility of UO3z in Sulfuric Acid at
Elevated Temperature, USAEC Report ORNL-1797, Oak Ridge National Lab-
oratory, Nov. 1, 1954.

8. (a) R. S. GreELEY, Oak Ridge National Laboratory, 1956, personal com-
munication. (b) R. S. GREELEY et al., High Temperature Behavior of Aqueous
Uranyl Sulfate— Lithium Sulfate and Uranyl Sulfate—Beryllium Sulfate Solutions,
paper presented at the 2nd winter meeting of the American Nuclear Society,
New York, Oct. 28, 1957.

9. E. PosNnyak and G. TunNeELL, Am. J. Sci. 218, 1 (1929).

10. F. E. CLARK et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 30, 19566, USAEC Report ORNL-2096, Oak
Ridge National Laboratory, May 10, 1956. (p. 130)

11. C. H. Secoy et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak
Ridge National Laboratory, Oct. 10, 1957 (p. 163); F. MoseLEY, Core Solution
Stability in the Homogeneous Agqtueous Reactor: The Effect of Corrosion Product
and Copper Concentrations, Report HARD(C)/P-41, Gt. Brit. Atomic Energy
Research Establishment, May 1957.

12. W. L. MArsHALL et al., J. Am. Chem. Soc. 73, 1867 (1951).

13. L. D. P. King, Design and Description of Water Boiler Reactors, in
Proceedings of the International Conference on the Peaceful Uses of Atomic Energy,
Vol. 2. New York: United Nations, 1956. (P/488, p. 372)

14. W. L. MARsHALL et al., J. Am. Chem. Soc. 76, 4279 (1954).

15. B.J. THAMER et al., The Properties of Phosphoric Acid Solutions of Uranium
as Fuels for Homogeneous Reactors USAEC Report LA-2043, Los Alamos Scien-
tific Laboratory, Mar. 6, 1956. D. FROMAN et al., Los Alamos Power Reactor
Experiments, in Proceedings of the International Conference on the Peaceful Uses
of Atomic Energy, Vol. 3. New York: United Nations, 1956 (P/500, p. 283).
L. D. P. King, Los Alamos Power Reactor Experiment and Its Associated H azards,
REFERENCES 123

- USAEC Report LAMS-1611(Del.), Los Alamos Scientific Laboratory, Dec. 2,
1953; A Brief Description of a One Megawatt Convection-cooled Homogeneous
Reactor—LAPRE II, USAEC Report LA-1942, Los Alamos Scientific Labora-
tory, Apr. 13, 1955. R. P. HammonDp, Los Alamos Homogeneous Reactor Pro-
gram, in HRP Civilian Power Reactor Conference Held at Oak Ridge, March
21-22, 1956, USAEC Report TID-7524, Los Alamos Scientific Laboratory,
March 1957. (pp. 168-176)

16. Figure 3-12 was supplied by R. B. Brigas, Oak Ridge National Labora-
tory. Unpublished.

17. L. D. P. King, Los Alamos Homogeneous Reactor ‘Program, in HRP
Civilian Power Reactor Conference Held at Oak Ridge, March 21-22, 1966, USAEC
Report TID-7524, Los Alamos Scientific Laboratory, March 1951. (pp. 177-209)

18. Frank J. LoprEsT et al., J. Am. Chem. Soc. 77, 4705 (1955).

19. F. J. LopresT et al., Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1955, USAEC Report ORNL-1943, Oak
Ridge National Laboratory, Aug. 9, 1955 (p. 227). W. L. MARsHALL et al,
Homogeneous Reactor Project Quarterly Progress Report for the Period Ending
Jan. 81, 1956, USAEC Report ORNL-2057(Del.), Oak Ridge National Labora-
tory, Apr.17,1956 (p.131). C. A. BLakE et al., J. Am. Chem. Soc. 78, 5978 (1956).

20. W. C. WagGeENER, Oak Ridge National Laboratory, 1953, personal com-
munication.

21. M. H. Lierzge and W. L. MARsHALL, Present Status of the Investigation of
Aqueous Solutions Suitable for Use in a Thorium Breeder Blanket, USAEC
Report ORNL-1711, Oak Ridge National Laboratory, June 1954.

22, W. L. MagrsHALL et al., J. Am. Chem. Soc. 73, 4991 (1951). Jonn R.
FERRARO et al., J. Am. Chem. Soc. 76, 909 (1954).

23. P. G. Jones and R. G. SowpEN, Thortum Nitrate Solution as a Breeder
Blanket in the H.A.R., Report AERE-C/M-298, Part I. Thermal Stability,
Gt. Brit. Atomic Energy Research Establishment, 1956.

24. W. L. MagrsuarLL and C. H. Secoy, Preliminary Exploration of the
Th(NO3)4-HNO3-H20 System at Elevated Temperature, in Homogeneous Re-
"actor Project Quarterly Progress Report for the Period Ending Oct. 31, 1954,
USAEC Report ORNL-1658, Oak Ridge National Laboratory, Feb. 25, 1954.
(pp. 93-96)

25. R. E. LEuzg, Chemistry of Plutonium in Uranyl Sulfate Solutions, in
HRP Civilian Power Reactor Conference Held at Oak Ridge National Laboratory,
May 1-2, 1957, USAEC Report TID-7540, Oak Ridge National Laboratory,
July 1957. (pp. 221-231)

26. D. E. GLanviLLE and D. W. Grant, Gt. Brit. Atomic Energy Research
Establishment, 1956. Unpublished.

27. Mound Laboratory, Interim Monthly Reports, 1957-1938.

28. L. V. JonEs et al., Isolation of Protactinium-231, in Homogeneous Reactor
Project Quarterly Progress Report for the Period Ending July 21, 1956, USAEC
Report ORNL-2148(Del.), Oak Ridge National Laboratory, Oct. 3, 1956.
(pp. 144-145)

29. A. T. GrEsKY, Separation of U233 and Thorium from Fission Products
by Solvent Extraction, in Progress tn Nuclear Energy—Series 3; Process Chem-
124 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

istry, by F. R. Bruck et al.,, New York: McGraw-Hill Book Co., Inc., 1956;
Thorex Process Summary, February 1956, USAEC Report CF-56-2-157, Oak
Ridge National Laboratory, Apr. 25, 1956. A. T. GRrEsky et al., Laboratory
Development of the Thorex Process: Progress Report, Oct. 1, 1952 to Jan. 31, 1953,
USAEC Report ORNL-1518, Oak Ridge National Laboratory, Aug. 4, 1953.
E. D. Ar~owp et al., Preliminary Cost Estimation: Chemical Processing and Fuel
Costs for a Thermal Breeder Reactor Power Station, USAEC Report ORNL-1761,
Oak Ridge National Laboratory, Feb. 23, 1955. R. E. ELsoN, The Chemistry of
Protactinium, in The Actinide Elements, ed. by G. T. SEaBora and J. J. Karz,
National Nuclear Energy Series, Division IV, Volume 14A. New York: McGraw-
Hill Book Co., Inc., 1953. (Chap. 5, p. 103)

30. D. E. FErcusoN, Homogeneous Plutonium Producer Chemical Processing,
in HRP Cwilian Power Reactor Conference Held at Oak Ridge March 21-22, 1956,
USAEC Report TID-7524, Oak Ridge National Laboratory (pp. 224-232). J. C.
HinomaN et al., Some Recent Developments in the Chemistry of Neptunium,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 7. New York: United Nations, 1956 (P/736, p. 345). K. A. Kraus,
Hydrolytic Behavior of the Heavy Elements, in Proceedings of the International
Conference on the Peaceful Uses of Atomic Energy, Vol. 7, New York: United
Nations, 1956. (P/731, p. 245). W. C. WaGGeNER and R. W. STOUGHTON,
Spectrophotometry of Aqueous Solutions, in Chemistry Division Annual Progress
Report for the Period Ending June 20, 1957, USAEC Report ORNL-2386, Oak
Ridge National Laboratory, (pp. 64-71). W. C. WAGGENER, in Chemistry Divi-
sion Seminar—1958, Oak Ridge National Laboratory (to be published).

31. A. O. ALLEN, J. Phys. & Colloid Chem. 52, 479 (1948). C.J. HOCHANADEL,
Radiation Stability of Aqueous Fuel Systems, USAEC Report CF-56-11-54, Oak
Ridge National Laboratory, November 1956.

32. F. 8. Dainron and H. C. SurroN, Trans. Faraday Soc. 49, 1011 (1953).

33. T. J. Sworskr, J. Am. Chem. Soc. 76, 4687 (1954).

34. E. J. Harrt, Radiation Research 2, 33 (1955).

35. Selected values from several literature references.

36. A. O. ALLEN et al., J. Phys. Chem. 56, 575 (1952). C. J. HOCHANADEL,
J. Phys. Chem. 56, 587 (1952).

37. C.J. HocuanapEL, Radiation Induced Reactions in Water, in Proceedings
of the International Conference on the Peaceful Uses of Atomic Energy, Vol. 7.
New York: United Nations, 1956. (P/739, p. 521)

38. J. W. BoYLE et al., Nuclear Engineering and Science Congress, Held in
New York in 1955, Amerlcan Institute of Chemical Engineers (Preprint 222).
H. F. McDurrie, Aqueous Fuel Solutions, Chap. 4.3, in Reactor Handbook,
Vol. 2, Engmeerlng, USAEC Report AECD-3646, Oak Ridge National Labora-
tory, 1955. (p. 571)

39. H. F. McDuvrrIE et al., The Radiation Chemistry of H omogeneous Reactor
Systems: 111—H omogeneous Catalysis of the Hydrogen-Oxygen Reaction, USAEC
Report CF-54-1-122, Oak Ridge National Laboratory, 1954.

40. For solubilities of Hz and Oz in water and in UQ2SQ4 solutions at high
temperatures see: H. A. H. PraY et al.,, Ind. Eng. Chem. 14, 1146 (1952); E. F.
STEPHAN et al., The Solubilities of Gases wn Water and in Aqueous Uranyl Salt
REFERENCES 125

Solutions at Elevated Temperatures and Pressures, USAEC Report BMI-1067
Battelle Memorial Institute 1956; see also See. 4, this chapter.

41. S. VisNer and P. N. HausenNreicH, HRE Experiments on Internal
Recombination of Gas with a Homogeneous Catalyst, USAEC Report CF-55-1-166,
Oak Ridge National Laboratory, Jan. 21, 1955.

42. R. E. AveN and M. C. LAwWRENCE, Calculation of Effects of Copper Catalyst
wn the HRT, USAEC Report CF-56-4-4, Oak Ridge National Laboratory, Apr.
10, 1956.

43. M. D. SiLvERMAN et al., Ind. Eng. Chem. 48, 1238 (1956).

44. J. W. BoyLe and H. A. MAHLMAN, paper presented at the 2nd Annual
Meeting of the American Nuclear Society, Chicago, 1956.

45. M. M. HarriNng, Mound Laboratory, Uranitum Salts Research Progress
Reports, USAEC Report MLM-628, 1951. E. OrBaN, Mound Laboratory, Ura-
nium Salts Research Progress Reports, USAEC Reports MLM-663, 1952;
MLM-680, 1952; MLM-710, 1952; MLM-755, 1952; MLM-790, 1952; MLM-825,
1953; MLM-863, 1953; MLM-928, 1953.

46. E. OrBaN, Mound Laboratory, 1952. Unpublished.

47. J. S. JEGART et al., The Densities of Uranyl Sulfate Solutions Between 20°
and 90°C, USAEC Report MLM-728, Mound Laboratory, Oct. 10, 1952.

48. E. OrBAN, The pH Measurement of Uranyl Sulfate Solutions from 25°
to 60°C, USAEC Report MLM-729, Mound Laboratory, Aug. 1, 1952.

49. E. OrBAN, The pH of Uranyl Sulfate-Uranium Trioxide Solutions, USAEC
Report AECD-3580, Mound Laboratory, Aug. 14, 1952.

50. J. R. HEeiks and J. S. JEGART, The Viscosity of Uranyl Sulfate Solutions
20° to 90°C, USAEC Report MLM-788(Rev.), Mound Laboratory, Feb. 22, 1954.

51. J. R. HEriks et al., Apparatus for Determining the Physical Properties of
Solutions at Elevated Temperatures and Pressures, USAEC Report MLM-799,
Mound Laboratory, Jan. 14, 1953.

52. A. J. RoGEers et al., An Instrument for the Measurement of the Tvme of
Fall of a Plummet in a Pressure Vessel, USAEC Report MLM-805, Mound
Laboratory, Feb. 27, 1952.

53. M. K. BARNETT and J. JEGART, The Surface Tenston of Aqueous Uranyl
Sulfate Solutions Between 20° and 756°C, USAEC Report AECD-3581, Mound
Laboratory, Feb. 15, 1953.

54. J. R. HEiks et al., The Physical Properties of Heavy Water From Room
Temperature to 260°C, USAEC Report MLM-934, Mound Laboratory, Jan. 12,
1954.

55. M. K. BARNETT et al., The Density, Viscosity, and Surface Tension of
Light and Heavy Water Solutions of Uranyl Sulfate at Temperatures to 250°C,
USAEC Report MLM-1021, Mound Laboratory, Dec. 6, 1954.

56. E. OrBaN et al., Physical Properties of Aqueous Uranyl Sulfate Solutions
from 20° to 90°, J. Phys. Chem. 60, 413 (1956).

57. R. VaNn WINKLE, Some Physical Properties of UOg804Dg0 Solutions,
USAEC Report CF-52-1-124, Oak Ridge National Laboratory, Jan. 18, 1952.

58. M. TosBias, Certain Physical Properties of Aqueous Homogeneous Reactor
Materials, USAEC Report CF-56-11-135, Oak Ridge National Laboratory, Nov.
23, 1956.
126 PROPERTIES OF AQUEOUS FUEL SOLUTIONS [cHAP. 3

99. J. A. LANE et al., Properties of Aqueous Solution Systems, Chap. 4.3 in
The Reactor Handbook, Vol 2, Engineering, USAEC Report AECD-3646, Oak
Ridge National Laboratory, 1955.

60. P. A. LorTEs, Physical and Thermodynamic Properties of Light and
- Heavy Water, Chap. 1.3 in The Reactor Handbook, Vol. 2, Engineering, USAEC
Report AECD-3646, Argonne National Laboratory, 1955. (pp. 21-42)

61. N. E. DorsEy, Properties of Ordinary Water-Substance, New York:
Reinhold Publishing Corp., 1940.

62. J. H. KeenaN and F. G. Keves, Thermodynamic Properties of Steam,
New York: John Wiley & Sons, Inc., 1936.

63. G. M. HEBERT et al., The Densities of Heavy-Water Liquid and Saturated
Vapor at Elevated Temperatures, J. Phys. Chem. 62, 431 (1958).

64. W. L. MARrsHALL, Density—Weight Percent—Molarity Conversion Equa-
tions for Uranyl Sulfate—Water Solutions at 25.0°C and Between 100-300°C,
USAEC Report CF-52-1-93, Oak Ridge National Laboratory, Jan. 15, 1952.

65. R. C. Harpy and R. L. CorTINGTON, J. Research Nat. Bur. Standards 42,
573 (1949).

66. H. O. Day and C. H Skecoy, Oak Ridge National Laboratory. Unpub-
lished data.

67. W. L. MARSHALL Oak Ridge National Laboratory, March 1958, personal
communication.

68. O. V. INGRUBER, The Direct Measurement of pH at Elevated Tempera-
ture and Pressure during Sulphite Pulping, Pulp Paper Can. 55 (10), 124-131
(1954).

69. M. H. Lierzke and J. R. TARRANT, Preliminary Report on the Model 904
High-temperature pH Meter, USAEC Report CF-57-11-87, Oak Ridge National
Laboratory.

70. M. H. Lierzke, The Hydrogen Electrode—Silver Chloride Electrode
System at High Temperatures and Pressures, J. Am. Chem. Soc. 77, 1344 (1955).

71. E. F. SteruAN et al.,, The Solubility of Gases in Water and in Aqueous
Uranyl Salt Solutions at Elevated Temperatures and Pressures, USAEC Report
BMI-1067, Battelle Memorial Institute, 1956.

72. H. A. Pray et al., The Solubility of Hydrogen, Oxygen, Nitrogen, and
Helium in Water at Elevated Temperatures, Ind. Eng. Chem. 44, 1146 (1952).

73. J. A. LukEer and THOMAS GNIEWEK, Saturation Composition of Steam—
Helium—Water Muixtures PVT Data and Heat Capacity of Superheated Steam—
Helium Maxtures, USAEC Report AECU-3299, Syracuse University Research
Institute, July 29, 1955.

74. J. A. LukEr and TaoMAs GNIEWEK, Determination of PVT Relationships
and Heat Capacity of Steam—Ozxygen Miztures, USAEC Report AECU-3300,
Syracuse University Research Institute, Aug. 2, 1955.

75. E. F. STePHAN et al., Ignition Reactions in the Hydrogen—Oxygen—W ater
System at Elevated Temperatures, USAEC Report BMI-1138, Battelle Memorial
Institute, Oct. 2, 1956.

76. T. W. LeEvLaNnD, Evaluation of Data from Battelle Memorial Institute on
Solubility of Hg and Og tn Solutions of UO2804 and UOoFs tn Water and on
Explosion Limits in Gaseous M<ixtures of He, Op and Hg0 and in H,, O, He,
REFERENCES 127

and H,0, USAEC Report CF-54-8-215, Oak Ridge National Laboratory, Aug.
23, 1954.

77. IrviN M. MacAFEE, JRr., Detonation, Explosion, and Reaction Liymats of
Saturated Stoichiometric Hydrogen—Ozxygen—Water Mixtures, USAEC Report
AECU-3302, Syracuse University, Research Institute, July 1, 1956. PauL L.
McGiLL and James A. LukgRr, Detonation Pressures of Stotchiometric Hydrogen—
Ozxygen Mixtures Saturated with Water at High Initial Temperatures and Pressures,
USAEC Report AECU-3429, Syracuse University, Research Institute, Dec. 3,
1956.
CHAPTER 4
TECHNOLOGY OF AQUEOUS SUSPENSIONS*
4—1. SUSPENSIONS AND THEIR APPLICATIONS IN REACTORST

4-1.1 Introduction. With the inception of the aqueous homogeneous
power reactor program at Oak Ridge National Laboratory in 1949, the
primary choice of fuel was highly enriched UO2SO4 solution. Use of en-
riched uranium alleviated to some extent the need for strict neutron
economy, but it was found that at high temperature (250 to 300°C) U0O2S04
solutions were more corrosive than pure water and were subject to a high-
temperature instability. As a result, a secondary development effort was
initiated at ORNL to determine the potentialities of suspensions of solid
uranium compounds as reactor fuels. The principal efforts were directed
at forms of UO3, because it was believed that under reactor operating
conditions the trioxide would be the stable oxidation state. Considerable
progress was made In studies of the oxide and its slurries, and in develop-
ment of equipment for circulating the slurries at concentrations of several
hundred grams per liter in 100-gpm loops at 250°C. In addition, a criticality
study was carried out with enriched UO3 - H20 in water to obtain assur-
ance that local fluctuations in concentration or settling would not unduly
affect nuclear stability [1].

In 1955, the UO3 work was set aside so that effort could be concentrated
on ThO2 suspensions, which are at the present time believed to be the only
suitable fluid homogeneous fertile material for use in an aqueous homo-
geneous thorium breeder. The ultimate of this effort at ORNL has been set
at a two-region, ThO2, homogeneous, power breeder.

In the following sections of this chapter a detailed account is given of
the studies on UOj3 slurries and ThOg2 slurries, and a description of the
present state of knowledge of their production, properties, and utilization.
The discussion will be based largely on work done in the United States, but
it should be kept in mind that studies on fuel- and fertile-material sus-
pensions have been conducted in other countries—in particular, in the
Netherlands and in Great Britain—and that exchanges of concepts and
data have aided the U.S. efforts.

*By J. P. McBride and D. G. Thomas with contributions from N. A. Krohn,
R. N. Lyon, and L. E. Morse, Oak Ridge National Laboratory.
1By R. N. Lyon.

128
4-1] SUSPENSIONS: THEIR APPLICATIONS IN REACTORS 129

4-1.2 Types of suspensions and their settled beds. Two-phase systems
of solids in liquids may be classified in several ways. On the basis of size
of particle a phase is said to be colloidal when it is sufficiently finely di-
vided to permit the surface attraction forces of the particles to exert a
strong influence on the mechanical properties of the material as a whole.
If, in addition, the particles are dispersed in a liquid and are sufficiently
small so they diffuse throughout the liquid due to their Brownian motion
in a normal gravitational field, they are referred to as sols. Sols are not
resolved in an ordinary microscope but are usually recognizable in an
ultramicroscope. The particles of a sol are usually less than 0.5 mjcron in
length for materials of density near that of water, while particles in a ThO2
sol are usually less than 0.05 micron. Particles in a sol may join to form a
random network of some strength having a semisolid appearance and called
a gel, or they may coalesce into loose and relatively independent clouds of
joined particles referred to as flocs. Suspensions of flocs or of particles which
are large enough to settle are referred to as slurries.

In some sols the particles are stabilized by the preferential attraction of
the suspending liquid to the particles’ surface. These are referred to as
lyophillic sols. In other sols the thermodynamically stable condition is a
flocculated or a gelled state, but the particles are held apart by electro-
static forces produced by ions which collect on and near the surface of the
particles. Sols of elements, oxides, and salts (including the oxides of
uranium and thorium) are generally of the latter type and are referred to
as lyophobic sols.

Although dispersions having particle sizes greater than 0.5 micron do
not form sols, since they are too large and have too large a mass to be ap-
preciably affected by Brownian motion, the particle surfaces may exhibit
some colloidal properties which are most pronounced when the particles
are very close together. The magnitude of these forces is such that spherical
particles of ThO2 which are 10 to 15 microns in diameter appear to show
only slight tendency to flocculate, while cubic or platelet forms of ThO2 and
UO3 - H20 of 1 or 2 microns on a side do show a marked tendency to
flocculate. »

When slurries having particles that are either relatively large or have a
high ionic charge on their surface (and hence have little tendency to floc-
culate) settle, the settled bed density approaches about 50 to 709, of the
particle density. The bed resuspends only slowly and is not easily de-
formed rapidly. An example of such a bed is settled sand. In general, such
beds may exhibit dilatancy, which means that the bed must expand to be
deformed, and the apparent viscosity of the bed increases as the rate of
shear increases. ThOg2 spheres of more than 5 to 10 microns settle to beds
of this type.

Flocculated slurries settle to a concentration at which the flocs become
130 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

joined, and from that point the particles are in part supported by indirect
contact with the walls and bottom of the container through the floc struc-
ture. The resulting settled bed may continue to compact indefinitely at a
slower and slower rate. Such beds behave more or less in a plastic fashion,
and may even exhibit a pronounced yield stress (i.e., shear stress required
before an appreciable deformation rate is initiated).

In UO3 - H2O slurries of concentrations up to several hundred grams per
liter, the yield stress is less than 0.1 1b/ft? and the slurries are almost of
Newtonian character. A breeding blanket requires ThOg2 slurries con-
- taining 500 to 1500 grams of thorium per liter, and in these concentrations
the yield stress varies from 0 to well over 1 lb/ft?, depending in part on the
concentration, on the shape and form of the oxide particle, and on the
presence or absence of certain additives. Settled beds of both ThO2 and
- UO3 - H20 may be either colloidal and plastic, or much more dense, non-
colloidal, and apparently dilatant.

4-1.3 Engineering problems associated with colloidal properties. The
colloidal behavior of some slurries offers three types of problems: high
yield stress, caking, and sphere-forming tendencies. To these may be
“added a general instability in the colloidal behavior which changes with
time, chemical treatment, and general previous history. A high yield stress,
in turn, offers three main engineering problems: high velocity required to
produce turbulence, a tendency to plug tubes, and a tendency to increase
~ the difficulty of mixing in a large blanket or reactor vessel. |

A cake is defined as an accumulation of particles on part of the surface
of the system in so dense and rigid a form that it cannot be deformed with-
“out fracture. A mud is a similar dense accumulation of greater yield
strength than the circulating slurry but which can be deformed without
" fracture of the accumulation.

Caking and mud formation have occurred occasionally in circulation
loops, causing plugging of tubes, hydraulic or mechanical unbalance in a
centrifugal pump, and drastic reduction in heat transfer to or from the
walls. These phenomena appear to be due to compaction of flocculated
slurry under the influence of stresses due to flow. The rigidity of a cake
or mud appears to be inversely related to the particle size.

Muds have been observed in both ThO2 and UO3 - H20 platelet slurries.
Cakes have been observed in ThO: slurries. In one case, a - to 2-in.
layer of ThO2 cake was built up on essentially all parts of a 3-in. 200-gpm
circulating system. The cake resembled chalk in strength and consistency.
It had a density of about 5.5 g/cc.

Sphere formation occurs when a circulating slurry contains very fine
particles and appears to resemble the formation of a popcorn ball [2].
Spheres ranging in size from about five to several hundred microns in
4-1] SUSPENSIONS: THEIR APPLICATIONS IN REACTORS 131

diameter have been made. Prolonged circulation causes an equilibrium
size to be reached which depends on the circulating conditions and the
starting material. Spheres have been formed from certain types of ThOq
in suspension, but no spheres or cakes have been observed in circulating
UO;3 - H20. This may be due to the fact that the greater solubility of
UO3 - H20 prevents its remaining as extremely fine particles.

Since cake, mud, and sphere formations appear to be a result of colloidal
behavior, effective control of the colloidal behavior of a slurry will prob-
ably control the formation of such aggregates.

Plastic materials which exhibit a high yield stress require a high velocity
‘before they become turbulent. It is not uncommon for ThOg slurries to
require 30 to 40 ft/sec velocity for the onset of turbulence. (It is of interest
to note that velocity by itself appears to be the most important criterion
for whether a given plastic will be in laminar or turbulent flow [3]—as op-
posed to the product of tube diameter and velocity, which is the cor-
responding criterion for a given Newtonian liquid.) Turbulence is, of
course, important in maintaining the suspension and in providing good
heat transfer. |

Control or elimination of colloidal, flocculating properties of slurries
can be accomplished by additives, particle-size control, or particle-shape
control. Electrolyte additives which attach to particle surfaces may pro-
vide so strong a charge that particles cannot approach to form a floc or
gel. In true lyophobic sols, the most effective additives are often those
which produce ions of atoms or radicals of the same type as those com-
posing the particle. For example, ThOz or Th(OH)4 sols of up to
 4000-g/liter concentration are easily made by the addition of Th(NO3)4
solution to freshly prepared Th(OH)4. On a somewhat similar basis, addi-
tives which may form partially ionized or lyophillic surface compounds are
often effective. Very small additions of H2C204, Na2SiO3, NazPOg,
and NaAlQO; have been found effective at room temperatures in producing
free-flowing Newtonian slurries from high-yield-stress ThO2 muds. In
most cases the effect is lost at elevated temperatures. However, coating of
the particles with a silicone compound and firing to convert it to SiO2 has
produced slurries which appear to remain unflocculated at temperatures
up to 300°C [4].

NaoHPO4 or NaHoPO4 added to UOs - HzO platelet slurries has pre-
vented the formation of muds in regions where they normally form. The
latter additive is preferred, since NapHPO4 solution appears to attack
stainless steel in the presence of oxygen at elevated temperatures [5]. ‘

Evidence indicates that as the particle size increases the colloidal effects
become reduced. In the case of UO3 - H20 an equilibrium size is reached
due to the continuous abrasion of the particles and subsequent recrystalli-
zation. In ThOgz, and probably in UOg slurries, the crystals are much more
132 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

resistant to abrasion, but at the same time the recrystallization in solution
is essentially nil. ThO2 produced by calcination of a plate or cube form of
Th(C204)2 retains the plate or cube form, but the particle is composed
of smaller crystals of ThO2. Violent agitation causes crystals in the
particles to break apart, after which they are free to exhibit the col-
loidal behavior of the finer particles. The higher the calcination temper-
ature, the larger the ThO: crystals. By calcination at 1600°C, crystals
in excess of 0.25 micron are produced, whereas calcination at 650°C gives
crystals of 50 to 100 A. In both cases, the over-all particle size may be of
the order of 0.5 to 5 microns. Even the material calcined at the higher
temperature exhibits a discouragingly high yield stress, however, at the
preferred concentration for a two-region breeder blanket (~1500 g/liter
at room temperature). Part of the difficulty may be due to a possible sys-
tematic arrangement of charges on the particle surface which causes actual
attraction and binding of particles into an unusually strong floc in a pre-
ferred orientation [6].

If spherical particles are used, the amount of possible common surface
between particles is limited and the permanence of mutual attachment is
correspondingly limited. In addition, if the surface is essentially uniform,
the charge arrangement might tend to repel rather than attract other
particles. Furthermore, a spherical shape permits the particles to be
larger without excessive abrasion.

Spheres made in circulating systems, as mentioned above, exhibit
Newtonian flow properties at concentrations up to about 3500 to 4000
g/liter at room temperature. At the present time they are rather friable,
although their density is of the order of 8.5. Calcining the spheres at very
high temperatures in furnaces, oxyacetylene flames, or electric arcs gives
them considerably greater integrity while retaining their noncolloidal
properties. However, at high firing temperatures (1800°C) a tendency of
the spheres to break up, owing presumably to internal stresses, has been
noted [7]. Dense spheres have been produced in small quantity by spraying
a Th(OH)4 gel which is subsequently hardened, dried, and fired.

Thus it appears that the colloidal behavior of ThO; slurries may be
minimized through the use of spherical particles, larger particles, coating
the particles with silica or some other compound, or by the use of some as-
yet-unperfected additive.

4-1.4 Engineering problems not associated with colloidal properties.
Sedimentation. One of the principal noncolloidal problems encountered
with suspensions or slurries is sedimentation which, in a flowing system, is
offset by any upward component of liquid velocity. By definition, in
idealized laminar flow in a horizontal conduit there is no upward velocity
component and the rate of settling should proceed at the same rate as in a
4-1] SUSPENSIONS: THEIR APPLICATIONS IN REACTORS 133

stagnant vessel. In a vertical tube with laminar flow, particles tend to be
more concentrated near the center of the tube [8]. One possible explanation
may be that the particles spin in the velocity gradient near the wall in a di-
rection which would cause them to move toward the center of the tube as
the liquid moves past them. It appears possible that a similar effect could
reduce the sedimentation rate in a horizontal tube. In an inclined tube
the solids will collect on the lower side of the tube, while a channel of low
solids content will appear along the upper side. The resulting radial vari-
ation in density and possibly in viscosity distorts the normal parabolic
velocity profile and complicates computation of the local sedimentation
rate.

In turbulent flow, fluctuating radial velocities will tend to cause diffusion
from more concentrated regions to regions of lower concentration in
competition with the settling due to gravity. Although a relatively strong
diffusion tendency exists across the bulk of the conduit, the diffusion rather
suddenly begins to be damped near the wall, although some random radial
velocity fluctuations may occur essentially up to the wall. The distance
from the wall at which damping begins to become pronounced is of the
order of about 1 mm for water for velocities at 1 ft/sec, in tubes larger than
about % to % in. in diameter, and it is generally recognized by hydrody-
namicists as being approximately proportional to v/u, where u, is the
mean velocity of the fluid and v is the kinematic viscosity, about
10~ 5 ft2/sec. Since particles in the slurries under discussion are of micron
size, rather than large fractions of millimeters, those particles which find
themselves well inside this layer above a horizontal surface may tend to
build up a sediment which becomes the solid surface from which the more
or less damped layer must be measured. This process can continue until
the diameter and velocity are reduced to the point where flow becomes
laminar and the tube is choked off completely, or until the local shear
stress becomes high enough to drag the particles along and an equilibrium
bed thickness is approached. At a given distance from the wall in the
damped region the local velocity of the liquid is proportional roughly
to the square of the mean velocity through the conduit, and the radial
velocity fluctuations in the damped regions vary in about the same pro-
portion. Therefore the mean stream velocity at which sediment tends to
accumulate is rather sharply defined for a given slurry.

It follows directly that a slurry which is flowing horizontally cannot
keep its particles in suspension unless the flow is turbulent, or unless it
is an extremely stiff flocculated mud, and even in turbulent flow a minimum
velocity may be required to prevent the accumulation of a sediment along
the bottom of a tube or conduit. Such sedimentation can occur in a reactor
blanket vessel in regions where the net velocity is extremely low.

Abrasion. A second problem is that of abrasion, which is more serious
134 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

in the case of thorium slurries than uranium slurries. UOj3 - H20 crystals
are relatively soft, and when they strike a stainless-steel wall, they tend
to break without damaging the surface of the wall. ThO3 particles, on the
other hand, appear to be sufficiently hard to abrade the protective oxide
film on stainless steel and perhaps to abrade the base metal. Continuous
removal of the film exposes the bare metal to corrosive attack by the hot
water and, in some cases, causes very serious attack. The attack is most
severe in local recirculation regions associated with flow separation and
in regions of sudden acceleration and direction change, such as in orifices,
pump Impellers, and pump seal rings. Materials such as zirconium and
titanium, which form very hard oxide films, and essentially noncorrodible
metals such as gold and platinum show more resistance to ThO2 slurries
than do stainless steels. Reduction of particle size, use of round particles
rather than sharp-cornered particles, and design of components and
piping to avoid regions of high velocity or high acceleration will reduce
attack.

4-1.5 Systems and components for using slurries in reactors. The pre-
ceding discussion implies several general conclusions regarding systems and
components for using slurries. For example, the over-all system should be
kept as free of extraneous circuits and secondary lines as possible. If
possible, the system should always tend to drain into a sump whenever
circulation stops, to prevent plugging by settled beds. Smaller side lines
should, where possible, be attached to the top of a horizontal run of the
main system to allow solids to settle into the main stream, and a minimum
size for smaller lines should be established based on the expected strength
of any reasonably conceivable settled bed. In ThOg2 or UO2 systems, all
elbows should be of at least moderately large radius, and sudden constric-
tions such as orifices should be avoided.

Mechanical pumps for ThO2 or UO2 must be leaktight, and should be
capable of handling the hard abrasive particles; this includes adequate
hydraulic design, the use of particularly resistant materials in regions of
high fluid acceleration or high velocity, and either very abrasion-resistant
bearings or essentially complete isolation of the bearings from the slurry.
Valves must be designed to operate in spite of the abrasive nature and the
settling or compacting properties of the solids; they must be leaktight to
the outside. The trim must be unusually abrasion- and corrosion-resistant
to ensure continued internal leaktightness.

Pressure-sensing instruments should not, in general, include long blind
passages of small diameters, which might easily become plugged; in ThO;
slurries, flowmeters should not include rapidly moving parts in contact
with slurries, unless bearings which are not affected by the abrasive action
of the solids are used.
4-2] URANIUM OXIDE SLURRIES | 135

All vessels should be provided with a means of resuspending solids which
have settled to the bottom. In some vessels, a simple mechanical agitator
is sufficient. In others, steam or gas sparging can be used. In still other
cases, more sophisticated systems may be required involving, for example,
injection of an external liquid or slurry stream to induce strong internal
recirculation currents.

The following sections of this chapter represent a brief, condensed
progress and status report of the work on suspensions. This effort is con-
tinuing at an accelerated rate as their potentialities are becoming more
clearly recognized and as the problems and difficulties are becoming more
rapidly overcome.

4-2. UraNiuM OXIDE SLURRIES*

4-2.1 Introduction. Preliminary studies on uranium oxide slurries for
use in a plutonium-producer reactor were carried out in the period 1940-
1944 by Vernon, Hickey, Huffman, and others, first at Columbia University
and later at the University of Chicago as a part of the Manhattan Project.
This program was discontinued before the feasibility of uranium oxide
slurries could be established, but a large backlog of information on the
properties and slurry behavior of the uranium oxides was obtained. These
studies are reported in detail by Kirschenbaum, Murphy, and Urey [9] in
a still secret volume of the National Nuclear Energy Series (I1I, 4-B)
which should soon be declassified. In 1951 work on the development of
uranium oxide slurries was revived, primarily at the Oak Ridge National
Laboratory. These studies were terminated in 1953 before a satisfactory
slurry was developed. The results are reported by Blomeke [10] and, in a
1955 Geneva paper, by Kitzes and Lyon [11]. Since 1953, emphasis on a
slurry fuel has centered on the development of a thorium-uranium oxide
slurry [12,13].

4-2.2 Chemical stability of uranium oxides. Both the early Manhattan
- Project work [9] and the ORNL work [10] indicated that uranium tri-
oxide would be the probable stable form of uranium oxide under the
radiolytic gas formed by the radiation-induced decomposition of water in a
reactor. Uranium dioxide in an aqueous slurry at 250°C was oxidized to
uranium trioxide in the presence of oxygen overpressure and even in the
presence of excess hydrogen gas. The extent of this oxidation depended on
the oxygen pressure, and seemed to be independent of the partial pressure
of hydrogen (Table 4-1). The extent of oxidation of U3zOs to uranium tri-
oxide depended on both temperature and oxygen pressure. The presence

*Information taken from reports by J. O. Blomeke (Ref. 10) and A. S. Kitzes
and R. N. Lyon (Ref. 11).
136

TECHNOLOGY OF AQUEOUS SUSPENSIONS

TABLE 4-1

[cHAP. 4

OxipATION OF UQO2 SLURRIES UPON HEATING UNDER
VARYING PARTIAL PRESSURE OoF HYDROGEN AND OXYGEN

 

 

 

 

 

 

 

 

 

 

Heating conditions Gas pressure Uranium
» oxidized,
Temp., °C Time, hr Ho, psi Oq, psi %
200 48 63.5 31.8 64.9
250 16 202 74.5
16 378 82.2
24 70 61.5 78 .4
24 70 175 91.4
TABLE 4-2

RepucTioN oF UQO3 SLURRIES AT 250°C UNDER
VARYING PARTIAL PRESSURES OF HYDROGEN AND OXYGEN

 

 

 

 

 

Gas pressure Uranium
Heating time, hr reduced,
Hz, pSi 02, pSi %
20 263 26.3 0.5
2 378 ~— 1.39
24 70 35 0
68 527 26.3 0.4

 

 

 

 

 

 

of a partial pressure of hydrogen did not seem to markedly inhibit the
oxidation (Table 4-2). When a slurry of UO3 - H20 rods prepared by ther-
mal decomposition of uranium peroxide in water was heated at 250°C
under varying pressures of hydrogen and oxygen, it was unchanged in the
presence of a stoichiometric mixture of hydrogen and oxygen in the ratio
of water. It was only very slightly reduced by a tenfold excess of hydrogen
over the stoichiometric (Table 4-3). Reduction of the UO3 and U3Os
under pure hydrogen atmospheres was quite slow, although freshly oxi-
dized uranium species formed by treatment of UOs with peroxide were
rather readily reduced with hydrogen [9].

4-2.3 Crystal chemistry of UOs. Uranium trioxide in an aqueous slurry
can exist as one of three hydrates, depending on the temperature at which it
4-2] URANIUM OXIDE SLURRIES 137

A

/
N\

/l\

  

L. Bipyramids

Rodlets

Ny
Ny .

W

F1a. 4-1. UO3-H:0 crystal habits.

UOy4 + 4H2O ~° UO3 — 3+ H20

185-200°C Pulvenzed
H2O UOg ++
UOg ++ H20
Pulverlzed 200-300°C

U02++ H20
UO,(NO3), ~ UO3 —o UO3 O3 .H20
185-300°C (Plate|efs) 185-300° C (Blpyramuds)

Fic. 4-2. Preparation of UO3-H20.

is maintained. In the earlier work [Ref. 9, pp. 45—49, 127-131] the mono-
hydrate was shown to be the stable form between 100 and 300°C. Four
crystalline modifications of the monohydrate were described: the a form,
“large six-sided orthorhombic tablets’”; B, “‘small six-sided orthorhombic
tablets”; v, “‘rhombic(?) hexagonal rods’’; and 9, ‘‘triclinic crystals with a
very complicated x-ray pattern.” The o and 3 were stable in water below
185°C and the v and 0 above 185°C.

ORNL studies on preparation of uranium oxide hydrates agreed in
general with those reported in Ref. 9. The most important exception was
the inability to prepare a triclinic crystal resembling the 6—-UO3 - H2O.
Attempts to prepare this modification resulted in the formation of bipyra-
mids or platelets, depending on the conditions. Three allomorphic modi-
fications of the monohydrate were obtained, depending on the mode of
preparation and treatment (Figs. 4-1 and 4-2). A rodlet form of UO3 - H20
was produced when the anhydrous trioxide, formed by heating uranium
peroxide at 300°C, was heated in water at 185 to 300°C (Fig. 4-2). The
rodlets were also prepared by autoclaving at 250°C for at least 16 hr a slurry
of uranium peroxide containing less than 50 ppm uranyl nitrate as an im-
purity. A platelet form of UOs - H2O appeared when the trioxide, made.
138 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

TABLE 4-3

OxipATION OF U3z0g SLURRIES ON HEATING FOR 24
Hours UNDER VARYING P4RTIAL PRESSURES OF
HypROGEN AND OXYGEN

 

 

 

 

 

 

 

 

Gas pressure Uranium
Temp., °C oxidized,
Hz, pSi 02, pSi %

150 51.8 25.9 63.2
170 59.5 29.8 71.7
200 63.5 31.8 90.1
295 67 33.5 92.6
250 — 35 88.0
250 175 35 85.1
25(0) 70 17.5 75.2
250 70 175 96.8

 

 

by decomposition of uranyl nitrate at 300 to 400°C, was hydrated at 185
to 300°C. Pulverized uranium trioxide rodlets digested at 200 to 250°C
converted to the platelet form, whereas pulverized uranium trioxide
platelets digested at 150 to 200°C transformed into rodlets. Crystals which
resemble truncated bipyramids were formed when either rodlets or platelets
were heated with water containing several hundred parts per million of
uranyl ions.

The rodlets were bright yellow in color, normally 1 to 5 microns in di-
ameter and 10 to 30 microns long; the platelets were pale yellow in color,
6 to 50 microns on edge and about 1 micron thick; the bipyramids were
also pale yellow in color and several hundred microns along each edge.

The rodlets appeared to be the same material as a y—UQOg3 - H20 re-
ported in Ref. 9 as having an orthorhombic structure. Unfortunately, cell
dimensions were not given in this reference, and it was impossible to
establish the identity without question. Zachariasen [14] reported the
cell dimensions of two different UOg3 - H20 crystals but gave no information
concerning the chemical history of his samples. He indexed both of these
structures as orthorhombic and called them a— and $-UQO3 - H20, 1nde—
pendently of the nomenclature of Ref. 9. From the cell dimensions given
by Zachariasen the positions of all possible lines in the x-ray diffraction
patterns were calculated, thus permitting a comparison to be made with
material prepared in the present studies. It was established from this that
the rods gave the same x-ray diffraction pattern as Zachariasen’s
a—-UQO3 - H20 and that the platelets had the structure of his S—UO3 - H20.
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 139

4-2.4 UO3 - HyO slurry characteristics. Easily suspended slurries were
prepared of both the rodlets and platelets. On the other hand, the bi-
pyramids, because of their size, required violent agitation to keep them in
suspension. With rodlets or platelets, slurries could be prepared which
were dispersed and kept in suspension, by mild agitation, at both room
temperature and at higher temperatures, even though settling occurred in
stagnant water.

Slurries of the rodlets were pumped satisfactorily at temperatures below
200°C [15]. Slurries of the platelets, although easily pumped, had a tend-
ency to form soft cakes on the pipe walls at temperatures above 200°C [16].
The influence of the trace quantities of nitrate impurities which remained
in the “‘purified”’ oxides was not investigated, however.

The solubility of pure UO3 - H20 in pure water is less than 10 ppm at
room temperature and is also low at high temperatures. As a result, pure
UO3 - H2O slurries were essentially neutral. The presence of soluble
uranyl salts of strong acids lowered the pH of the slurry, however, and
increased the solubility of the oxide. In the preparation of UO3 by the
pyrolysis of UO2(NO3)2, for example, residual nitrate could not readily
be removed by a simple washing step, and slurries of such oxides released
nitrate when the crystals were broken down in a pumping system. Under
extreme conditions the increased uranyl ion concentration in the supernate
caused serious crystal growth and the formation of hard cakes in stagnant
regions [17].

4-2.5 Zero-power reactor tests. The microscopic inhomogeneity of en-
riched rodlet slurry fuel was found to offer no serious difficulty in the
operation of a zero-power homogeneous slurry reactor [9]. In this reactor,
suspension was established by a propeller type of mixer located near the
bottom of the vessel. The reactor was extremely stable at any given stirrer
speed. Changing the stirrer speed produced a change in nuclear reactivity
which was attributed to a redistribution of the oxide when the stirrer
was moving slowly, and to a change in the shape of a vortex type of con-
cavity in the slurry when the stirrer was moving rapidly.

4-3. PREPARATION AND CHARACTERIZATION OF THORIUM OXIDE
AND ITS AQUEOUS SUSPENSIONS*

4-3.1 Selected properties of thorium oxide. Thorium oxide is a white,
granular, slightly hygroscopic solid with a fluorite structure (lattice
constant — 5.5859 4- 0.0005) [18] and an x-ray density of 10.06. The
Chemical Rubber Handbook of Chemistry and Physics [19] gives 10.03 as
the density of thoria. Foex [20] gives pycnometric densities for thorium

*By J. P. McBride.
140 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

powders, prepared by firing the hydroxide, which increased with increasing
firing temperature (8.6 at 450°C; 9.4 at 725°C; 9.7 at 910°C), approaching
the x-ray density asymptotically. Foex also noted that the density of a
compacted bed of the thoria powder (3000 kg/cm? pressures) increased
with firing temperature but remained much lower than the actual powder
densities, the ratio of pycnometric density to bed density changing from
1.46 to 1.40 over the firing-temperature range of 250 to 1000°C. Thoria
powders obtained from oxalate thermal decomposition had pycnometric
densities almost identical with those prepared from the hydroxide for the
same firing temperature [21]. The melting point of thorium oxide has
been reported [22] to be 3050 4 25°C, and the boiling point has been
estimated [23] at 4400°C.

The bibliographies of reports available from the AEC on thorium oxide
in the list appended to this chapter provide sources of more detailed
information.

4-3.2 Preparation of thorium oxide. The principal method of preparing
thorium oxide for use in aqueous slurries has been the thermal decomposition
of the oxalate. Thorium oxalate, precipitated from thorium nitrate solution,
is crystalline, easy to wash and filter, and the oxide product is readily dis-
persed as a slurry. In addition, the oxide particle resulting from oxalate
thermal decomposition retains the relic structure of the oxalate, and hence
thé particulate properties are determined by the precipitation conditions.
The mechanism by which the thermal decomposition takes place has been
quite widely investigated [24-26]. The following is proposed by D’Eye
and Sellman [26] for the thermal decomposition:

ThO2 + 2CO2 4+ 2CO
N\

Th (0204)2 > 30000 \
Th(COs3)2 4+ 2CO

 

ThOz + 2COq

Properties of thorium oxide prepared by the thermal decomposition of
oxalate are discussed in detail in Articles 4-3.3 and 4-3.4.

A satisfactory oxide has also been prepared by the hydrothermal decom-
position of thorium oxalate as an aqueous slurry in a closed autoclave at

300°C[27]: Unidentified organic

and inorganic gases
>175°C
Th(C204)2 + excess H2O — > Th(OH)4 4 2H2C204

ThO: - 2H20
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 141

This preparation is characterized by a very small particle size, approxi-
mately 0.02 micron, and low bulk density, and very closely resembles the
oxide from the thermal decomposition of oxalate after the latter has been
pumped at elevated temperatures.

A third method for the preparation of slurry oxide is the thermal decom-
position of thorium formate [28]. In this procedure, thorium nitrate in
solution is decomposed on adding it to concentrated formic acid at 95°C
[29,30]. The precipitated thorium formate is washed free of excess acid
and decomposed by calcination at 500 to 800°C. The oxide from the formate
procedure is similar in its slurry behavior to that produced by thorium
oxalate thermal decomposition; however, less is known about its handling
characteristics. Because of this, the oxalate preparation method is pre-
ferred at the present time.

Experience on the preparation of oxide by the direct calcination of
Th(NO3)4 is limited. The nitrate decomposes at about 250°C, but firing to
500°C is necessary to remove the last traces of nitrogen oxide decomposition
products. The hydrated salt goes through a plastic stage during calcination,
and the resulting oxide is sandlike and difficult to slurry. In the absence
of a grinding and size-classification step, direct calcination of the nitrate
in a batch process does not appear to be a promising preparation method
for preparing oxide for slurry.

An interesting method for producing submicron-size thorium oxide
directly from thorium nitrate is that developed by Hansen and Minturn
[31]. Their method consisted of the combustion of an atomized solution of
thorium nitrate in an ethanol-acetone mixture and collection of the resulting
thorium oxide smoke. |

Micron-size thorium oxide may also be prepared by the hydrothermal
decomposition of a thorium nitrate solution at 300°C. The product from
the preparation is a free-flowing powder [32]. At temperatures much below
300°C the rate of hydrolysis is quite slow.

Brief studies made with thorium hydroxide indicated [33] that it is
probably not a good source material for the production of slurry oxide. As
precipitated from nitrate solution, the hydroxide formed a bulky precipitate
which was hard to filter and wash, was amorphous to x-rays, and contained
considerable nitrate impurity. Drying at 300 to 500°C yielded a crystalline
oxide product which was difficult to slurry. Autoclaving a slurry of the
hydroxide (without previous drying) at 250°C gave a bulky slurry (settled
volume 300 to 500 g Th/liter) exhibiting a characteristic ThOz x-ray
diffraction pattern.

4-3.3 Large-scale preparation of thorium oxide. In the present method
(Fig. 4-3) [34] for making thorium oxide in a pilot plant operated by the
Chemical Technology Division at Oak Ridge National Laboratory, 1 M
142 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

First Addition 125 gal

10-15°C 1.0 M Thorium Nitrate
Oxalate Addition, 3 hr Second Addition 275 gal

.| Digestion At 75°C—6 hr 1.0 M Oxalic Acid

 

 

Precipitation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

280 Ib/day :—a;sich atio “—1\
Thori | 110 /Ib/ day |
orium 85 % Yield | Su Dewatering
pernate
gxalafe | Redispersion ‘ 65 Ib/day
| Slurry | and 85 % Yield
Demineralizer Water Wash | | Decantation =
3 Times, 25 gal Each | | (2 Times) | Waste Filtrate
{ R(:{cey::e l 170 gal 10-3 g Th/liter THO,
Thorium Oxalate Cake | Supernate | | Cakes
- - |
| [ | Dlsz'e‘:jslon 170 gal. 2.8x10~3M Oxalic Acid
Filtration { Decantation '. 5 liter/kg of THO,, Per
100 Ibs/day l (4 Times) | Dispersion
L

 

Waste Flyzate ]—60._? Z VV_; te Heel 11 % Co;::i;ri\‘:ion
475 gal, 1073 g Th/liter L_S€ eeill /o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O
C:l;:'mezd (Mainly Oversize) 652 Og'll]:.?c{:yhr
First Calcination '
Stepwise i
180°C, 2 hr Second
380°C, 2 hr o ) Calcination
650°C, 1-3/4 hr | _650°C Calcined |1600°C, 4 hr
90 Ib/dOY Thorium Oxide 35 Ib/day SlstdUCt
97 % Yield 96 % Yield 85 '3"33;2: q

Fig. 4-3. Thorium oxide pilot plant chemical flowsheet. Percent yields based
on initial thorium input.

solutions of thorium nitrate and oxalic acid are mixed in an agitated tank
with controlled temperature, addition rate, and order of addition. In the
first step, all the thorium nitrate is added to the precipitator, after which
the oxalic-acid solution is added over a period of 3 hr with the reagents
held at 10°C by external cooling. The slurry of precipitated thorium oxalate
is digested for 6 hr at 75°C and then pumped to a vacuum filter where the
solid is separated from the mother liquor and washed three times with
demineralized water. The oxalate cake on the filter is air-dried and is then
loaded on trays for the first calcination.

In the first calcination the air-dried thorium oxalate is heated succes-
sively at 180°C for 2 hr, at 380°C for 2 hr, and at 650°C for 1.75 hr. The
material is then packed on a tray for the second calcination, and heated
at 1600°C for 4 hr.

The 1600°C-calcined thorium oxide normally  contains about 109, of
particles larger than desired (>5 microns). These oversize particles are
removed by classification, i.e., by suspension of the thorium oxide in oxalic-
acid solution (pH 2.6) to a ThO2 concentration of 100 to 200 g/liter. The
suspension is stirred, and the mixture is then allowed to stand for 5 min
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 143

before the supernate is decanted. Coarse material (5 to 7 microns) is
separated by setting the supernate withdrawal rate at 0.5 in./min liquid
level drop. The withdrawn thorium oxide is dispersed and decanted again
twice to ensure removal of oversized particles. The thorium oxide that
settles to the bottom is also redispersed and decanted three more times to
separate the considerable fraction of the fine particles that settle with the
heel or are imperfectly dispersed. This procedure removes nearly all the
thorium oxide smaller than 5 microns, and the final product contains only
1 or 29, of particles greater than 5 microns. This material is then refired at
650°C to decompose the oxalic-acid dispersant before being used in en-
‘gineering studies.

Oxide prepared in this way has an average particle size of 1 to 3 microns
and has handled well in high-temperature engineering loop tests at slurry
concentrations as high as 1500 g Th/kg H2O. Removal of the oversize
particles has decreased the erosive attack on loop components to essen-
tially what would be observed with water alone (see Section 6-7). At
1500 g Th/kg H20, slurries of average particle sizes =1 micron have moder-
ately high yield stresses (0.5 to 1 1b/ft2). Lower-yield-stress slurries are
obtained with the larger particles.

Previous engineering experience with slurries of oxide prepared similarly
but with final firings at 650 and 800°C [35] showed them to possess an
extremely high yield stress at concentrations greater than 750 g Th/kg H 20,
and an occasionally bad caking characteristic [36]. Firing at 1600°C
appears to have in large part removed or substantially diminished the
caking tendency [37].

4-3.4 Characterization of thorium oxide products. Although thorium
oxide is a very refractory substance, it is well known that such properties
as 1ts density, catalytic activity, and chemical inertness depend on the
conditions of its formation. With particular references to preparation from
the oxalate, the firing temperature has a marked effect on the ease of forma-
tion of colloids [38]. Beckett and Winfield [25] concluded from electron
micrographs of oxide residues that the initial oxalate crystal imposes on
the residual oxide a mosaic structure of thin, spongy, microcrystalline
laminae all oriented in very nearly the same direction. Foex [20], investi-
gating the rate of change in density as a function of the firing temperature
for oxide prepared from the hydrous oxide, associated the density change
with crystallite growth among closely joined crystallites and observed that
no sintering of particles seemed to take place below 1000°C.

Oxide products from thorium oxalate decomposition are normally char-
acterized by their behavior as slurries. In addition, they have been char-
acterized by means of electron micrograph pictures, their nitrogen
adsorption surface areas, particulate properties as measured by sedimenta-
144 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

tion* [39], and average x-ray crystallite size by x-ray diffraction line
broadeningt [41].

Effect of preparation variables on the particulate properties of thortum oxide.
The effects of thorium oxalate precipitation temperature, calcination
temperature, and calcination time on oxide properties were initially in-
vestigated by Allred, Buxton, and McBride [42]. Oxalate was precipitated
at 10, 40, 70, and 100°C from a 1 M thorium-nitrate solution by dropwise
addition of oxalic-acid solution and vigorous stirring. The precipitates
were fired at 400°C for 16 hr and successively at 500, 650, 750, and 900°C
for 24 hr. Electron micrographs of the oxide products showed particles of
the approximate size and shape of the original oxalate particles from which
they were formed. The particles of oxide prepared from 10°C-precipitated
material were approximately 1 micron in size and appeared quite uniform;
those from the 40°C material were 1 to 2 microns in size and less uniform.
A marked increase in particle size was observed for the oxide particles
prepared from the 70°C- and 100°C-precipitated materials, which were
4 to 7 microns in size. There was no change in particle shape or average
particle size and no evidence of sintering as the firing temperature was
increased from 400 to 900°C. Micrographs of shadow-cast oxides showed
that the particles prepared from oxalate precipitated at 10°C were almost
cubic in shape, with an edge-to-thickness ratio of about 3:2, and that those
from the 100°C material were platelets with an edge-to-thickness ratio of
6:1 (Fig. 4-4). The mean particle sizes determined by sedimentation
particle-size analyses were in good agreement with the data from the elec-
tron micrographs (Fig. 4—4).

Table 44 shows typical data obtained with the 10°C-precipitated ma-
terial. Included in Table 4—4 are the results of additional firings up to
1600°C. No increase in average particle size was noted even up to 1600°C.
However, in all oxide preparations, there was about 10 w/o above
5 microns in particle size.

*A radioactivation method for sedimentation particle-size analysis of ThO2 was
developed at ORNL [39]. The oxide was activated by neutron irradiation, dis-
persed at < 0.5 w/o concentration in a 0.001 to 0.005 M NasP,07 solution and
allowed to settle past a scintillation counter connected to a count-rate meter and
a recorder. The scintillation activity, being proportional to thoria concentration,
was analyzed in the usual manner, using Stokes’ law, to give the size distribution
data. Independently, an analogous method for use with UOz powders was devel-
oped at Argonne National Laboratory [40].

+The x-ray crystallite (as opposed to the actual oxide particle, which may be
composed of a great many crystallites in an ordered or disordered pattern) is de-
fined as the smallest subdivision of the solid which scatters x-rays coherently. The
crystallite size can, in principle, be determined from the width of the x-ray diffrac-
tion peak, the width being greater the smaller the average crystallite size [41].
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 145

" () Cubic Shape From 10°C

 

(b) Platelet Shape From 70°C Precipitated Oxalate

F1g. 4-4. Particle shapes of thorium oxide prepared from oxalate thermal de-
composition.

Additional studies on the effects of the chemical and physical variables
of the batch oxalate precipitation step on the particulate properties of
thorium oxide from oxalate thermal decomposition were carried out by
Pearson, et al. [34,43]. In addition to precipitation temperature, the effect
of reagent concentration, stirring rate, reagent addition rate, and digestion
time were investigated. Most of the precipitations were made by adding
oxalic-acid solution to the thorium nitrate solution, which appeared to
give an oxide product of smaller average size and fewer oversize particles
than the reverse. Rather than being added dropwise, the oxalic acid was
146 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

TABLE 44

CHARACTERISTIC PROPERTIES OF THORIUM OXIDE FROM
OXALATE THERMAL DECOMPOSITION

(Precipitation temperature 10°C. Average particle size for all firings was between
' 1.1 and 1.4 microns.)

 

 

 

Final firing e Average x-ray Specific
Firing time o
temperature, by ! crystallite size, surface area,
°C A m2/g
400 16 61 35.0
500 24 78 40.0
650 24 143 25.0
750 24 250 13.0
900 24 550 6.3
1000 12 803 4.3
1200 12 1100 3.0
1400 12 2000 2.4
1600 12 2000 1.0

 

 

 

 

 

 

introduced under pressure into the thorium nitrate solution through a
capillary tube projecting beneath the surface of the nitrate solution, the
tube exit being directly above the agitator blades.

In these experiments [34,43] it was found that the conditions for pro-
ducing oxide with uniform* particles of 1 micron average size and a low
percentage of particles greater than 5 microns were 1 M Th(NOg3)4+ and
H,C>04 concentrations, a 10°C precipitation temperature, a high rate of
oxalic-acid addition, and vigorous stirring. A draft tube with a variable
opening placed around the stirrer permitted a further control over the
average particle size, larger particles (2 to 4 microns) being produced by
increasing the rate of recirculation of the precipitating system through the

*Tt is common practice to present the particle-size distribution in the form of a
plot of the logarithm of the particle size versus the cumulative weight percent
undersize on a scale based on the probability integral. If the particle-size data
follow a logarithmic probability distribution (as they usually do reasonably well
for most ThO. preparations), the resulting plot is a straight line and the steeper
the slope of the line, the less uniform the material. The 509, size in such a plot
is the geometric mean particle diameter (d;). The geometric standard deviation
(g,) is equal to the ratio of the 84.139 size to the 509, size (also 509, size: 15.879%,
size) [44]. For thorium oxide prepared from oxalate precipitated at 10°C, o, was
1.2 to 1.4; og increased with increasing precipitation temperature.
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 147

 

 

 

 

D, micron
0.001 0.01 0.1 1.0 10
O T TIINT T T T T T T T T T
— XA\ —
: IQ.JA R :
L ci\e\’%of_ ]
S,
Z&I.\I;Q')/

o 10 \‘\ '5,.0 —
~ — ]
“E — A .'\%@\, —
§ s=(¢ )2 AN -
< | pF/D A\%, ]
| A Q —_

§ \¢°x)
> | . \60 ]
[ A 100° < 6 1]
| O 40° |
=0 N

A D-16 (Standard Oxide) . \1
01 [P L L L M
10 100 1000 10,000 100,000

Average X-Ray Crystallite Size, A

F16. 4-5. Relationship between average crystallite size and specific surface area
for thorium oxide prepared by the thermal decomposition of thorium oxalate
(400 to 900°C firings).

draft tube. Decreasing the concentration of the oxalic-acid solution in-
creased the average particle size (0.7 to 0.8 micron at 1.5 M and 2 to
3 microns at 0.5 M) but did not affect the size distribution, which appeared
to be primarily temperature-dependent. The shape also changed, with
decreasing reagent concentration, from a cube to a platelet. Only the
particles about 1 micron in size were cubic in shape. Long digestion times
seemed to reduce localized sintering effects in high-fired oxides (1400,
1600°C) and hence lowered the fraction of oversize particles.

Effect of preparation variables on x-ray crystallite size and specific surface
areas of thortum oxide from oxalate thermal decomposition. Specific surface
areas of oxide prepared from oxalate thermal decomposition were much
larger than could be anticipated from the particle size, and decreased with
increasing firing temperatures (Table 4—4). Crystallite sizes as measured
by x-ray diffraction line broadening increased with increasing firing tem-
perature and corresponded very closely to the particle sizes estimated
from the specific surface areas (Fig. 4-5) [42]. The product of the surface
area in m?/g and the crystallite diameter in angstroms, at least for oxide
fired at =900°C, was approximately constant and equal to 3.6 X 103.

While the crystallite size was determined primarily by the firing tem-
perature, a relationship between crystallite size and firing time was also
148 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

established [42]. Log-log plots of crystallite size versus firing time for
oxides at various firing temperatures fit an equation of the form

where D is the crystallite diameter, ¢ the time, and 7' the absolute tem-
perature; «, A, and B are constants. The constant o appears to be char-
acteristic of oxide prepared by the thermal decomposition of thorium ox-
alate. For D in angstroms and ¢ in hours, the equation becomes

D= t0.14e(10.3515— 5482/ T)'

The temperature-dependent function, e(4~8/T is typical for rate processes
requiring an energy of activation, the constant B being equal to AH/R,
where AH is the heat of activation. The heat of activation was determined
to be 10.97 kecal/g-mole [42].

Oxides from the hydrothermal decomposition of thorium oxalate. Oxides
prepared by the hydrothermal decomposition of the oxalate [27] at 300°C in
a closed autoclave were found to be markedly different in their characteristic
properties from the thermally prepared materials. The precipitation tem-
perature of the oxalate had no effect on the final shape or size, and all
evidence of the original oxalate structure had disappeared. Sedimentation
particle-size analyses indicated particle sizes between 0.5 and 1 micron.

 

60 | | | | |

50 |—

  
     
    
 

Thermally Decomposed
Fired at Designated Temp. for 24 hr. ]

8
l

w
o
|

Hydrothermally Decomposed
at 300°C for 24 hr., then Fired
at Temp. for 24 hr. Period.

Specific Surface Area, m2/g

I

 

 

0 | | | | |
200 400 600 800 1000 1200 1400
Temperature,°C
Fia. 4-6. Effect of oxalate decomposition method on thorium oxide surface area.
(Prepared from oxalate precipitated at 100°C.)

 
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 149

The hydrothermal oxides were composed of crystallites 200 A in size which
did not grow on subsequent firing at temperatures up to 900°C. Further-
more, the specific surface area of the hydrothermally prepared oxide
decreased less with increasing firing temperature than did the surface area
of the oxide from oxalate thermal decomposition, and was much larger
for the higher firing temperatures (Fig. 4-6).

Effect of high-temperature water on oxide properties. Experiments have
been carried out [45-48] which show that treatment with high-temperature
water has practically no effect on the characteristic properties of oxide
itself. Prolonged heating in water at temperatures up to 300°C did not
change the crystallite size, bulking properties, or abrasiveness of thorium
hydroxide calcined at 500 and 650°C [45], and tests showed no evidence
of hydrate formation [46,47]. Lack of crystallite growth probably indicates
an extremely low solubility of thorium oxide in water at 300°C. When
slurries of oxide calcined at 650, 750, 900, and 1000°C were heated overnight
at 300°C in the presence and absence of as much as 10,000 ppm of SO4
(pH about 2), there was no increase in the average sedimentation particle
size or x-ray crystallite size [49]. _

Effect of pumping on oxide properties. Oxides with particle size much
greater than 1 micron are degraded to an average particle size much less
than 1 micron on pumping at elevated temperatures,* while oxides com-
posed of cubic particles of about 1 micron show little change [50]. Surface
area increases of from 16 m2/g to 30 m2/g have been noted for some pumped
oxides, the latter figure being almost the theoretical maximum for the
measured x-ray crystallite size [51].

Pumping does not affect the average x-ray crystallite size of slurry
oxides [52]. Also, oxides which have been pumped as slurries, dried, and
then calcined, show relatively little crystallite growth. From these con-
siderations it would seem that the crystallite size as measured by x-ray
diffraction line broadening represents the ultimate limit of the attrition
process due to pumping.

4-3.5 Sedimentation characteristics of thorium oxide slurries. All tho-
rium oxide-water slurries, except the very dilute suspensions, in the absence

*While most of the pumped slurries have been slow settling, composed of small
particles (=1 micron), and have yielded bulky sediments, on occasion the small
particles resulting from the attrition of the original slurry particles reagglomerated
to form large spheres 10 to 50 microns in size. This resulted in a slurry which set-
tled rapidly to a dense but easily resuspended bed. Sphere formation appeared
to be the property of specific oxide preparations and was observed most often with
800°C-fired material. Preparing the oxide in a particle size which does not degrade
on pumping (~ 1 micron) and firing at 1600°C to improve particle stability appears
in the initial pumping studies to have largely removed the problem.
150 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

 

0.14 I

0.10 I~ ‘ —

0.08 |— —

o
o
o
|
l

Initial Settling Rate, cm/sec

0.04 —

0.02 —

     

 

 

0 I I
200 400 600 800
Oxide Concentration, g Th/kg H20

Fig. 4-7. Effect of oxide calcination temperature and slurry concentration on
the room temperature settling rates of aqueous thorium oxide slurries.

of a dispersing agent exhibit “hindered” settling. All particles settle at a
constant rate, and there is a well-defined interface between the suspended
particle and the supernatant. This form of settling is called ‘‘hindered”
because there is mutual interference of the particles in their motion, and
Stokes’ law does not apply to the settling of each particle.

Usually three zones of settling are observed: the initial zone during which
the motion of the interface from its initial level is uniform (hindered settling)
and rapid; a compressive zone during which the motion of the interface
is also uniform, but much slower; and a final stationary state. The first
settling zone terminates when the interface reaches the settled mass of
flocs. The slurry concentration (grams of thorium per liter) at the
point of transition between the initial settling zone and the compressive
zone is termed the critical density or the critical concentration. The slurry
concentration in the final stationary state is called the settled concentration
(sometimes the bulk density). The compressive zone is characteristic of a
flocculated material and shows the rate of compaction of the settled bed
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 151

under the force of gravity. With discrete particles, or with ThOg2 in the
presence of a dispersing agent (0.005 M Na4P207), the particles settle
directly into a permanently settled bed, and there is no compressive zone.

Room-temperature sedimentation characteristics. The room-temperature
hindered-settling rates (for a given slurry concentration), critical concen-
trations, and settled concentrations of thorium oxide slurries (unpumped)
all increased with increasing firing temperature up to 1000°C of the oxide.
In addmon the hindered-settling rates also increased with decreasing
slurry concentration (see Fig. 4-7) [53]. It should be noted in Fig. 4-7
that all slurries are in compaction in the 500 to 800 g Th/kg H20 con-
centration range.

Table 4-5 shows the combined effects of slurry temperature and oxide
calcination temperature on the hindered-settling rate, Uo[50]. From the-
oretical considerations the product Uou, where u is the viscosity of water,
should remain constant for an oxide over a series of settling temperatures,
provided that no change has occurred in the particulate or dispersive char-
acteristics of the oxide. Uou does remain fairly constant over the temper-
ature range 27 to 98°C. That changes do occur in oxide properties,
however, with increasing calcination temperatures up to 1000°C is shown
by the increase in the Ugu product with calcination temperature. The
trend appears to reverse with the 1300°C-fired material, which also shows
a larger change in the Uou product with temperature than do the lower-
fired materials.

High-temperature sedimentation characteristics. Slurry settling rates
at temperatures in excess of 100°C have been obtained in quartz tube
8 mm In diameter [54]. These data, obtained with a slurry of thorium
oxide prepared by a 650°C calcination of thorium formate [55], indicated
that the slurry was already in the compaction zone of settling above
500 g Th/kg H20 at 200 to 300°C. At a concentration of 10600 g Th/kg HO,
no settling occurred at temperatures above 100°C. The small diameter
of the tube probably affected the concentration at which the slurry went into
compaction.

Data on the sedimentation characteristics of thorium oxide slurries at
elevated temperatures in stainless-steel autoclaves, £ in. in inside diam-
eter,* have been obtained by an x-ray adsorption technique [56].
Standard x-ray film was transported at a controlled speed past a vertical
slot in a lead shield behind which a bomb containing the settling slurry
and an x-ray source were placed. A typical radiograph of a settling slurry
at an initial concentration of 250 g Th/kg H2O at 205°C is shown in
Fig. 4-8.

*This diameter should have no effect on slurry hindered-settling rate certainly up
to 400 g Th/kg H20 concentration and possibly even higher.
[cHAP. 4

TECHNOLOGY OF AQUEOUS SUSPENSIONS

152

 

 

 

 

 

 

 

 

 

900 | 020 | OI'0 | ¥¢€0 | W00 | ST'0 | 8¥0'0 9T°0 | 820°0 010 668% 0 86
900 | 9T°0 | OI'0 | 20 | ¥%00 | 10 | ¥%0°0 21’0 | L2800 200 66.€ 0 )
00 | 2000 | 800 | SI'0 | 00 | 600 | 980°0 L0°0 | S20°0 | S¥0°0 767 0 0g
€0'0 | €00 | 800 | OTI'0 00 | 90°0 | 2800 | 8800 | 6200 | T1€0°0 GF$8°0 LT
o) °N o) °n o) °n 0] °n o) °N
os10d1juao Do
0081 060001 0,006 0,008 D059 ‘O%H Jo “dwey
1 AJISOOSTA Aum[g

 

 

 

 

 

 

oInjeradure) PajeoIIPUI 4B POUID[BI SOPIXO JO SOI)SLIAJIBIBYD SUI[})OS

 

 

 

NOILLVNIOTV)) @AIX() ANV TIALVHEAIWE ], AHYNTS 40 IOALAN]

(09s/W0 ‘eI JUI)AS paIapuIy = 0/))
O%H 3q/YJ, 8 00¢ ‘SEIM¥NIS QHJ, 40 SOLLSIMELOVEVH)) ONITLLIS NO HYALVITdWI],

¢—§ A'19V],

 
153

 

Subsidence, cm

O

 

N O

   

 

 

       

 

 

Initial Concentration—250g Th/kg HO

Temperature—205°C

                  

 

10

Time, sec

20

Fia. 4-8. Typical high-temperature settling curve obtained with x-ray apparatus.

Settling Rate, cm/sec

0.1

0.01

Temperature, °C

100 150

200 250

 

 

| |

  

11 microns

2.6 microns

0.7 microns.

 

   

    
   
 

I

I

  

   

Fluidity of Water

L1l

I

 

 

26
177 x 102 (°k-1)

30

22

Fi1c. 4-9. Temperature-particle size effects on the settling rate of thorium oxide
slurries: 250 g Th/kg H-O.
164 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

 

50 | | l | |

Subsidence Rate,cm/sec

 

150°C

100°C

 

 

 

o .| | | | |
0 200 400 600 800 1000 1200
Initial Concentration, g of Th/kg of H,0

Fia. 4-10. Effect of slurry concentration on settling rate.

The combined effects of temperature and particle size on slurry settling
rates for 250 g Th/kg H2O slurries of oxides fired at 800 to 900°C are
shown in Fig. 4-9. Since the systems were flocculated, the hindered-
settling rates are much greater than those predicted from the mean particle
sizes by a simple application of Stokes’ law. The settling rate-temperature
dependence curve for the slurries containing the larger particles closely
- paralleled the curve for the fluidity of water. Hence it may be assumed
that for these slurries little or no change occurred in the flocculating
characteristics with increasing slurry temperature. For the slurries con-
taining the smaller particles the curve is steeper, showing an apparent
increase in agglomerate size or density with increasing temperature.

The effect of slurry concentration on the settling characteristics of a
slurry at elevated temperatures is illustrated in Fig. 4-10. The bulk of this
slurry (>80 w/o) was made up of spherical agglomerates 10 to 15 mi-
crons in size. The data indicate that the slurry settling rate is an exponential
function of the concentration and has the form
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 155

Temperatures,®C
150 200 250 300
| l

W
o
~N
O
8

 

        
   

[ 111

EREEE

|
A
Fluidity of Water

—

Initial Settling Rate, cm/sec
o
O

0.2

0.1

0.05

I

0.02 |—

 

 

0.01 1 | | | 1 1
31 29 27 25 23 21 19 17

1/1 x104, ok !

 

F1a. 4-11. Effect of firing temperature on high-temperature settling rates.
250 g Th/kg H20; particle size 1 micron; prepared from 10°C precipitated oxalate.

Uo _
U,~

where Uy is the measured sedimentation rate, U, is the settling rate at
infinite dilution where Stokes’ law should govern the particulate settling, C
is the slurry concentration, and a is the slope of the logarithmic settling
rate-slurry concentration curve. Extrapolating the straight-line portion
of the curves to zero concentration and assuming an agglomerate density
of 5.0 g/ce, Stokes’ law particle diameters were calculated at the various
temperatures and were found to be approximately the same, again illus-
trating a lack of change in slurry flocculation characteristics with increasing
slurry temperature. The calculated particle diameters at 100, 150, 200,
and 250°C were 40, 45, 45, and 44 microns, respectively, far greater than
those obtained by sedimentation particle-size analysis in dilute suspension.

 

In

aC,
156 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

 

T T T T T T T T
Initial Concentration — 250g Th/Kg H,0

v 250°C

1T

| | |11EA

 

l
l

 
 

 

 

 

o
{’ 1.0 _:— s
€ — —
C; b —_
s .. il
o L —
2 |10 H2504 Titration 200°C ]
£ 8 T~
A 6[-pH
0.1 =4 150°C —
=2 =
— 100°C B
02 _
[ L R R L DL
10 100 1000 10,000

ppm Sulfate
(Based on Dry Solids)

Fia. 4-12. Effect of thorium sulfate on hindered settling rates of oxide slurries.

The effects of oxide firing temperatures of 900 and 1600°C on slurry
settling rates at elevated temperatures are shown in Fig. 4-11. Slurry
concentrations were 250 g Th/kg H2O, and the oxide was prepared from
the 10°C-precipitated oxalate (cubic particles, ~1 micron). Settling data
obtained on the slurries after they had been pumped at elevated tempera-
tures are also included. The slurry of higher-fired material shows much
higher settling rates. Pumping does not greatly affect the settling rates of
slurries of either oxide above 150°C. The temperature dependence of
settling rates roughly follows the change in water fluidity, but the curve
for the 1600°C-fired pumped material deviates considerably below 150°C
and is much flatter.

Effect of additives on settling rates. The fluidity of concentrated non-
Newtonian slurries can be increased by the use of additives. Of particular
importance is a knowledge of the effect of temperature on the action of such
additives. Both sulfate and sodium silicate additions, either as thorium
sulfate or sulfuric acid, were investigated and found to change markedly
the settling rates and handling characteristics [50] of thoria slurries, the
relative effect of any additive concentration on a given slurry depending on
the slurry temperature.

Sodium silicate is a well-known deagglomerator, as well as a wetting
agent. Its addition to thick concentrated slurries at room temperature
increases their fluidity to that approaching water. It also markedly im-
proves their heat-transfer properties for certain flow conditions [58].

The effect of thorium sulfate additions on the high-temperature sedimen-
tation properties of a thoria slurry (250 g Th/kg H20) composed of spheri-
cal agglomerates approximately 15 microns in average size is shown in
4-3] THORIUM OXIDE AND ITS AQUEOUS SUSPENSIONS 157

 

10 | | |
- — 300°C

\

 

L111]

 

250°C

t

150°C

 

|

II IIIIIIII

|

0.1

Settling Rate ,cm/sec

R

I

I IIIHII

 

 

 

100°C
I

 

Slow—sle—0.007 —

1 1000 - 10,000
5i09 Concentration ppm,based on ThO,

0.01

 

F1a. 4-13. Effect of sodium silicate on the hindered settling rates of oxide slurries.
250 g Th/kg H20, d, = 1 micron.

Fig. 4-12. An abnormal increase in the hindered-settling rate in the tem-
perature region of 150 to 200°C was obtained upon the addition of between
500 and 1000 ppm of sulfate (based on ThO32) and again at about 5000 ppm
of sulfate. The concentration region of 2000 to 3000 ppm of sulfate appears
to be one of a relatively low settling rate and good temperature stability.
It is of interest to note that in the operation of a high-temperature loop,
abnormal pump power demands at 200°C were observed when pumping
slurry containing 1000 ppm, and that increasing the sulfate concentration
to between 2000 and 3000 ppm removed the difficulty and permitted
operation at 300°C- [54].

Also appearing in Fig. 4-12 is the sulfuric acid titration curve obtained
with the standard slurry. It should be noted that the sulfate concentration
regions of temperature instability bracket the break in the pH curve and
the region of temperature stability occurs between pH 6 and 7.

Figure 4-13 shows the effect of sodium metasilicate additions on the
settling rate at elevated temperatures of a 250 g Th/kg HoO slurry of
800°C-fired oxide which had been micropulverized to an average particle
size of 1 micron. At silica concentrations of 5000 to 30,000 parts SiO2 per
million ThO3, the settling rates are reduced (by comparison with the pure
slurry) at all temperatures up to 250°C, but the effect is more pronounced
at the lower temperatures.

It would appear from the studies carried out so far that the relative
dispersion effect for any additive concentration depends markedly on the
158 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

slurry temperature. The more pronounced effects are observed at tempera-
tures below 200°C. Above 250°C the effect of the additive becomes less
pronounced and sometimes even negligible from the point of view of its
effect on the settling rate. It may be, however, that at high temperatures
the additive could change the viscous properties of the slurry in a dynamic
system and not affect its settling rate in a quiescent state (essentially zero
shear stress) at the same temperature.

4-3.6 Status of laboratory development of thorium oxide slurries. Tho-
rium oxide prepared by the thermal decomposition of thorium oxalate ap-
pears suitable for a reactor slurry at concentrations up to 1500 g Th/kg H2O.
Study of the preparation variables has indicated that a considera-
ble control can be exercised over the properties of the slurry oxide.
Little is known as to what physical and chemical properties of thorium
oxide are important in determining its handling characteristics in water
at high temperatures, and studies are being made to determine these
properties. Sulfate and silicate additives have been shown in settling studies
to have a marked effect on the dispersion characteristics of slurries at
temperatures below 200°C, but at reactor temperatures the effect of the
additive on the settling rate diminishes and may be negligible. The effect
of additives on the rheological properties of slurries at reactor temperatures
has not yet been determined. Attempts are being made to obtain slurries
of ideal rheological characteristics by the preparation of oxide of controlled
particle size, shape, and surface activity.

4—4. ENGINEERING PROPERTIES*

4-4.1 Introduction. The major difference in descriptions of the engi-
neering properties of aqueous suspensions (compared with aqueous solu-
tions) arises from the fact that suspensions may exhibit either Newtonian
or non-Newtonian laminar-flow characteristics. The consequences of the
possibility of these two different types of behavior modify conventional
heat transfer, fluid flow, and sedimentation correlations, and are important
in the design of large systems for handling slurries.

The magnitude of the effects that can be observed with non-Newtonian
slurries is illustrated in Fig. 4-14, where the critical velocity for the onset
of turbulence is shown to be a strong function of the slurry yield stress and
almost independent of coefficient of rigidity and pipe diameter [59]. The
usefulness of laminar-flow measurements in characterizing different
suspensions, as well as the application of these constants to a variety of
correlations, will be given in the following sections.

*By D. G. Thomas.
4-4]

ENGINEERING PROPERTIES

 

20

Critical Velocity, ft/sec

 

| I

I l

 

 

 

0.5

1.0

Ty, Yield Stress, Ib/ ft2

159

FIG 4-14. Effect of slurry physical properties on velocity for onset of turbulence
= 100 lb/ft3, D=1 to 24 in.

SeECIFIC-HEAT CONSTANTS FOR THE

TABLE 4-6

Oxipes or UraNiuM AND THORIUM [61]

(Cp=a+ (bX 10-3)T + ¢ X 105/T?)

 

 

 

 

 

 

 

 

Material a b ¢ Temp. range, °K
H-0 11.2 7.17 —_ —_
ThO2 16.45. 2.346 —2.124 to 1970
U0 19.20 1.62 —3.957 to 1500
Us0s (65) (7.5) (—10.9) —
UO3 22.09 2.54 —2.973 to 900

 

 
160 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

4-4.2 Physical properties. The heat capacity of suspensions of solids is
commonly taken as the sum of the heat capacities on a weight basis of the
liquid and the solid at the bulk mean suspension temperature, each multi-
plied by its respective weight fraction in the suspension [60]. Specific-heat
values for pure thorium and uranium oxides are given in Table 4-6.

Thermal conductivity data for mixtures of solids have been correlated
[62] using Maxwell’s [63] equation:

 

. 2ko+ky—2¢p(ko—kp)

b= 50 ko kp T blko — ky) @1
for the electrical conductivity of a two-phase system. This equation was
subsequently used to correlate conductivity data for suspensions of solids
in a gel [64]. However, the thermal conductivity of suspensions has not
been shown to be independent of the rate of shear [65]. The thermal
conductivity of sintered thorium oxide having a bulk density of 8.16 g/cc
was found to decrease from 6.0 to 2.5 Btu/(hr) (ft) (°F) as the tempera-
ture was increased from 140 to 500°C [66,67].

Suspensions of solids in liquids may be either Newtonian or non-New-
tonian, depending primarily on particle size and electrolyte atmosphere
around the particles [68]. Newtonian and non-Newtonian materials are
classified and compared by means of shear diagrams in which the rate of
shearing strain, dv./dr, is plotted against the shear stress, 7. Newtonian
fluids are characterized by a shear diagram in which the rate of shearing
strain is directly proportional to the shear stress, as shown in Fig. 4-15, the
viscosity being given by:

p = —g.7/(dv,/dr), (4-2)

where the coefficient of viscosity, u, is independent of the rate of shearing
strain. On the other hand, non-Newtonian fluids have a variable viscosity
that is a function of the rate of shear and in some cases of the duration of
shear. Detailed discussions of non-Newtonian materials are available
elsewhere [69-72].

Einstein [73] has shown that the viscosity of dilute suspensions of rigid
spherical particles is a function of the volume fraction of solids in the
suspension and is independent of particle size, as shown in Eq. (4-3):

ps = p(1 4 2.5¢), (4-3)

where ¢ is the volume fraction of solids in the suspension. It was assumed
that the system was incompressible, that there was no slip between the
particles and the liquid, no turbulence, and no inertia effects, and that the
rglacroscopic hydrodynamic equations held in the immediate neighborhood
4-4] | ENGINEERING PROPERTIES 161

Newtonian
Material

 

   
  

Bingham Plastic
Material
dv _ 9¢

= 7 (‘r—ry)

   

Pseudo-Plastic
Material
dv_9 ,

dr k

 
   
   

  

Dilatant
Material

  

Rate of Shearing Strcin,sec_]

dv
dr’

 

 

 

|

|
|
|
|
l

 

7, Shear Stress,lb/ft2

Fia. 4-15. Classification of Newtonian and non-Newtonian materials by shear
diagram.

of the particles. Einstein’s treatment for dilute suspensions has been
extended by Guth and Simha [74], Simha [75], de Bruyn [76], Saito [77],
Vand [78], and Happel [79] to suspensions of higher concentrations. In
all these treatments the hydrodynamic interaction between particles was
considered, the extension usually taking the form of terms proportional
to ¢? and ¢3, the terms to be added to the Einstein term 2.5¢. These
results are summarized in Table 4-7.

Theoretical treatments have also included such nonspherical particles
as ellipsoids [85] and dumbbells [86] and an empirical relationship has
been determined for rod-shaped particles [80]:

ps = (1l + 2.5F¢ + 8F?¢* + 40F3¢?), (4-4)

- where F is dependent upon the axial ratio, but not upon size or concentra-
tion of the rods. Values of F, determined with a Couette viscosimeter, are
given in Table 4-8.

The viscosities of suspensions of UOgz - H2O rods and platelets with
uranium concentrations of up to 250 g/liter were measured [87] with a
modified Saybolt viscosimeter at temperatures from 30 to 75°C. There was
no detectable difference in viscosity between the slurries of rods and those
of platelets at these uranium concentrations. The viscosity values given in
162 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

TABLE 4-7

COMPARISON OF THEORETICAL AND EMPIRICAL EXTENSION OF
THE EINSTEIN RELATION TO HiGHER CONCENTRATIONS
(All relations are of the form u,= u(l+ A1¢p+ A202+ A3d3)

 

 

 

 

 

 

 

Refer-
Author ez:; A; A Aj Comments
Vand [78] [ 1.5 7.349 0 Considered mutual hydro-
dynamic interaction and
collisions between parti-
cles and pairs of particles
Guth and
Simha [74] | 2.5 14.1 0 Considered mutual hydro-
dynamic interaction; neg-
lected formation of pairs
de Bruyn [76] | 2.5 2.5 2.5 | Considered only mutual hy-
drodynamic interaction
Saito [77] | 2.5 2.5 2.5 | Considered only mutual hy-
drodynamic interaction
Gosting and
Morris [84] | 3.35 0 0 Very dilute
Oden [83] | 2.5 30to60 | O Sulfur sols
Boutaric and | [82] | 2.5 75 0 As2S3 sols
Vuillaume
[80] | 2.5 9to13 | O Glass spheres in water
Eirich
[81] | 2.5 7.17 16.2 | Glass spheres in ZnIs-
Vand water-glycerol solutions
Happel [79] | 5.5¢* — — | Each particle confined to a
cell of fluid bounded by
frictionless envelope

 

 

 

 

 

 

 

 

*y = Interaction factor, 1.00 at ¢ = 0; 4.071 at ¢ = 0.50.
4-4] ENGINEERING PROPERTIES | 163

TABLE 4-8

VALUEs oF CORRECTION FAcTOR, F, FOR
ErrFeEcT OF AXIAL RATIO OoF RoODS
IN VISCOSITY OF SUSPENSIONS

 

 

 

L/D F
5 2.1
11 2.25
17 2.60
23 4.20
25 5.60
32 7.0
50 11.0
75 22
100 32
140 a0

 

 

 

 

Ref. 87 were used to determine the value of the shape factor F in Eq. (4—4),
which was derived for rod-shaped particles. The value of F for the
UOs3 - H20 data is 2.4 4- 0.7, corresponding to an L/D of about 14 (from
Table 4-8). This agrees very well with the dimensions reported for the
rodlets of from 1 to 5 microns diameter and 10 to 30 microns long [87].
(The dimensions for platelets were 6 to 50 microns on edge and about
1 micron thick.)

Powell and Eyring [88] have applied the theory of absolute reaction
rates to arrive at a suggested general relation between the shear stress and
the shear rate for non-Newtonian fluids:

_ldv 1. (lds _
T—cdr bSlnh <a dr) (4-5)

It is found that in the range of most common interest, 103 < (dv/dr) < 109,
sinh~![(1/a)(dv/dr)]= 6.4 4+ 3.5 for a variety of ThO; slurries [89] and does
not change rapidly with changes in dv/dr. Thus 1/b sinh™[(1/a)(dv/dr)]
is in effect 7, in the Bingham equation [90] for an idealized plastic:

dv__gE _ B
8;—77 (7 Ty)- (4-6)
164 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

For convenience, the apparent yield stress, 7,, and the coefficient of
rigidity, n, will be used to characterize different uranium and thorium
oxide slurries [72].

If 1t 1s assumed that the particles in a flocculated non-Newtonian slurry
stick together in the form of loose, irregular, three-dimensional clusters in
which the original particles can still be recognized, and further that the
yield stress is a manifestation of the breaking of these particle-particle
bonds, then for constant isotropic bond strength the yield stress should be
proportional to the cube of the volume fraction solids. Since the shear
forces are exerted in a plane, the yield stress should also be proportional
to the number of particles per unit area, and hence, for constant volume
fraction solids, the yield stress should be proportional to the reciprocal of
the square of the particle diameter.

Data on the yield stress and coefficient of rigidity as a function of
concentration for three particular uranium oxide preparations [91] are
summarized in Table 4-9. The yield stress-volume fraction solids data
may be expressed by a relation of the form

Ty = k1¢4. (4—7)

Values of k; for the three different oxides are shown in Table 4-9. The
data for the coefficient of rigidity may be fitted by a relation of the form

n= 1 explkad], (4-8)

using values of kg given in Table 4-9.

TABLE 4-9

RuEoLOGIC PROPERTIES OF NON-NEWTONIAN
URraNIUM OXIDE SLURRIES

 

 

 

 

 

 

 

 

Particle-size ;
distribution ky=-Y,
Oxide by = R(0/1) ¢
D,, @ Ib/ft?
o microns
UO. 1.7 1.4 1.8 150
U30sg 2.0 1.3 2.2 230
UO3H20 1.9 1.2 2.2 430

 

 

 
ENGINEERING PROPERTIES 165

4-4]

 

 

 

 

ee z'1 ¥'3 LT 08 sassed ¥ I9ZLISA[NA-OIYIIA 0091 2201
¥ 0'1 02 LT — — ouON 0091 1-62-0'T
09 80 i 81 00¢ L8LE dwng 0091 $1-V00Z
001 Al | 9'1 | — — QUON 0091 S-62-0'1
G4l G 1 01 61 0S G0 IOpUd[q SULIBA\ 0091 08-M
0SS 1 GL'0 8'C 00g 0081-006 dung 008 11-V00%
0L 81 8G 0 62 00€ $22 dung 008 1-V002
00TT ¥'g 020°0 A 063 GZe dung 0S9 6G-S
mawhoma 0 “ Do , Iq POt
2N/l 5 ‘a dueg, aonena o JSQ\%MQE&
T =% uo1nsuUl
W% —ey | (/WU 1 EONALYST uorye)I3y OHETRIED

 

 

 

 

 

 

9ZIS-9[01}I8J

 

 

 

 

UOI}BUTISIP 9pIX()

 

SATIYNTS HAIX() WATYOHJ, NVINOLMIN-NON 40 SALLYAJOUJ OIDOTOTHY

0I-% @1av],

 
166 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

 

I T THET T

R

L 11
to Waste

|
I

I

o
n

INERREE
| L 1]l

|
|

o
h
|

Tybg, (Ib/ftz)(micronz)

0.1

T
| L 111l

0.05

0.02

|
I

 

 

 

0.01 : | |
0.01 0.02 0.05 0.1 0.2 0.5 1

Volume Fraction Solids, ¢

 

Fic. 4-16. Effect of particle diameter and volume fraction solids on ThOg slurry
yield stress.

The yield stress and coefficient of rigidity as a function of volume
fraction solids for a variety of different ThO; slurries [89] are given in
Table 4-10. The yield stress-volume fraction solids curves can be fitted
by a relation of the form

Ty = k3d3. (4-9)

The coefficient of rigidity-volume fraction solids curves can be fitted by
a relation of the form

n = u exp(kso). (4-10)

The data given in Table 4-10 were obtained with ThOg2 slurries having
a pH less than 6 and whose rheological constants were relatively insensi-
4-4] ENGINEERING PROPERTIES 167

 

 

 

 

 

 

 

 

= 1.0
Il [ I I I | b === —
/'/- —]
. —o.5
I o~ I _
./4—> _
0.2 l/ I —0.2
| | :
0.1 — —0.1 2
= || | 4 &
— — o
— — A
0.05 '| l —0.05 .
B ] .-
| ] c
D
T 002 | I <
- e
g o001 | =001 =
v | - _— [¥]
: E| | 3 3
@ — — g
= 0.005 |- —{0.005 5
% [ I | 7 §
_— — o
| c
0.002 |— | —0.002 £
|  Region of Newtonian ‘I uE.
0.001 |— |‘ Behavior fi| —{0.001
0.0005 |— | I — 0.0005
0.0002 |- I I —{0.0002
oooor—tL L1 1 | L0 oo
0O 10 20 30 40 50 60 70 80 90 100

Oxalic Acid/mole Th02 ,millimoles

Fic. 4-17. Effect of electrolyte on ThO; slurry yield stress.

tive to dilution and reconcentration and therefore could be considered as
having a similar and reproducible electrolyte atmosphere associated with
the particles [89]. On the basis of the above considerations a plot of k3
(Eq. 4-9) versus mean particle diameter was found to fall on a line of
slope minus two on a log-log plot, as suggested by the plausibility argu-
ments given above. Therefore the data of Table 4-10 were plotted as
7,D2 versus volume fraction solids. All points fell within the two lines
shown in Fig. 4-16. It is believed that the spread of data is largely due
to effects of the electrolyte atmosphere [89], since the deviations from
the mean particle size were similar. The influence of small quantities of
electrolyte on the yield stress of a particular ThO; slurry [92] is shown
in Fig. 4-17. Similar behavior for particulate systems has been described
elsewhere [68,93].
168 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

Preliminary measurements [94] of rheologic properties of ThO2 slurries
at temperatures up to 290°C indicate that the yield stress is essentially
independent of temperature (4309%), whereas the coefficient of rigidity
decreases with temperature, although not to the same extent as water.

4—4.3 Fluid flow. The pressure drop due to friction for viscous flow of
Newtonian fluids through pipes is given by the Poiseuille equation:

Ap=32uLV/g.D%. (4-11)

The same equation may be used for non-Newtonian suspensions, pro-
vided that the “‘apparent’’ viscosity, e, is substituted for the viscosity.

Buckingham [95] has presented a mathematical relationship for the flow
of Bingham plastics in circular pipe:

8V /gD = (1/n)(re —(&/3)7, + (1/3)74/73).  (4-12)

For large values of 7,, the last term of Eq. (4-12) becomes small, and the
resulting expression for the shear stress at the wall, 7, when combined
with Eq. (4-13),
T = DAp/4L, (4-13)
gives
Ap =329 LV /g.D?*+ (16/3)1, L/D (4-14)

for the pressure drop of a Bingham plastic in laminar flow.

Hedstrom [96] has proposed a simple criterion to distinguish between
laminar and turbulent flow of Bingham plastic materials. The Reynolds
number at which turbulence sets in is determined by the intersection of
a parametric curve, defined by

Nue = g.p 7,D?/7?, (4-15)

4
1 _Ji_lfie_.{__l_ Nie (4-16)

and the turbulent Newtonian friction curve on a Fanning friction factor-
Reynolds number plot, provided that the Reynolds number is defined as
DVp/n. The usefulness of the Hedstrom concept has been demonstrated
by several investigators [97,98]. Figure 4-18 is a plot of the solution of
Eqgs. (4-15) and (4-16) superimposed on a friction factor-Reynolds number
diagram for Newtonian fluids flowing in smooth tubes.

In general, non-Newtonian fluids behave similarly to Newtonian fluids
in the turbulent flow region in that they exhibit relatively constant ap-
4-4] ENGINEERING PROPERTIES | 169

 

 

 

 

 

0
T T\ TT T ™\T 11  INT 1T T T 111 ] 1T
D
106 107 108 107 = Hedstrom No. =Ny, = gc—’;;lfi —
“ 1071 —
.'o: —
[¥]
o et
e
=
k5 _
s
£
- _
e
€
Y102 —
w3l L L LUl e L L L
102 103 104 103 108 107

DV
Reynolds Number, Ngp = "n_p

F1e. 4-18. Friction factor-Reynolds number diagram for Bmgham plastic slur-
ries in smooth pipes.

 

   
     
 

 

 

 

[ | 1] l bl I oY

N ]
- \
g 1071 |— 0.124 in. ID —
L — N —
6 B \ —
|- \ —
;E \
2 | Newtonian ——-\\ 1.030 in. ID —
‘e Laminar \
c \
o Flow \
w10-2 |- —
- | Commercial Pipe |

y
L =~ —_
N\ s: ~--L_____
| ~— .
| Smooth TubeT |
10-3 I L 11 | L 1 1 | [ 1 1] l |
102 103 104 10° 106
DVp
NRe'

F1a. 4-19. Friction factor-Reynolds number data for ThO; slurries in turbulent
flow. 7,=0.0751b/ft2, n = 2.9 cp.
170 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

parent viscosity. Alves, Boucher, and Pigford [69] indicate that, in the
absence of data, the conventional friction-factor plot may be used to predict
turbulent pressure drop to within 425%, provided that the Reynolds
number is evaluated by using the coefficient of rigidity or viscosity at
infinite shear and that the density is taken as that of the slurry. However,
it has been shown that a large amount of turbulent pressure-drop data
taken with both Newtonian and non-Newtonian slurries can be correlated
using the usual friction factor-Reynolds number plot provided the density
is taken as that of the slurry and the viscosity as that of the suspending
medium [99-105]. Data obtained [106] with two different ThO2 slurries
using three different tubes are shown in Fig. 4-19. As can be seen, the tur-
bulent friction-factor line is below the smooth-tube Newtonian line at low
Reynolds numbers and approaches the smooth-tube line at high Reynolds
numbers.

Vanoni [107], who reviewed the literature on sedimentation transporta-
tion mechanics through 1953, has pointed out that although a quantitative
description of the phenomena was unavailable at that time, it was clear
that sediment movement is intimately associated with turbulence.
Subsequent work either has been largely of an empirical nature [108-110]
or has unquestioningly accepted and used [111-113] generalized flow rela-
tions which have been developed for homogeneous Newtonian fluids
[114]. Undoubtedly the presence of particulate matter in the flowing
stream will exert a perturbing influence on the flow pattern at velocities
near drop-out, and a quantitative solution to the problem must include at
least an estimate of this effect [115].

It has been proposed [116] that the velocity below which particulate
matter will be deposited on the bottom of horizontal pipes from a Bingham
plastic suspension corresponds to the critical velocity for the onset of tur-
bulence. This velocity may be calculated approximately by setting the
modified Reynolds number (obtained by using the apparent viscosity
us = (1 + g.D7,/69V) from Eq. 4-14 instead of viscosity as usually de-
fined) equal to 2100 and solving for the velocity:

 

_ 21009 210077 (4)(2100)gc7y _
Ve="Dp T \/ - 6 o

Resuspension velocities* for one particular slurry [117] in a $-in. glass pipe
are given in Table 4-11 together with the rheological properties [118] and
the critical Reynolds number calculated with the resuspension velocity.

*The resuspension velocity corresponds to the velocity at which a moving bed
disappears as the mean stream velocity is increased.
4-4] ENGINEERING PROPERTIES 171

TABLE 4-11

ResuspeENSION VELOCITY FOR BiNncHAM Prastic THO2 SLURRIES

 

 

 

 

 

 

Critical Reynolds
Concentration, Resuspension number,
?(’:c g Th Ty m velocity, V.,
g kg H0 Ib/ft2 | <P fps DV,
n[1+ (9. D 74/61V.)]
2.44 1645 0.48 5.9 7.4 2220
2.29 1490 0.35 5.5 6.3 2040
2.17 1325 0.25 5.1 5.5 2050
2.05 1180 0.19 4.6 5.2 2240
1.94 1030 0.12 4.2 4.5 2380
1.84 910 0.098 | 3.9 3.7 1920
1.76 825 0.065 | 3.6 3.5 2350
1.65 690 0.040 | 3.1 3.7 3460

 

 

 

 

 

 

As can be seen, the critical Reynolds number is very close to the proposed
value [116] of 2100. Additional data on a variety of slurries and different
tube diameters are being obtained [119] to further substantiate Eq. (4-17).

4-4.4 Hindered-settling systematics. The hindered-settling velocity of
slurries may be expressed as a coefficient times the Stokes’ law settling
rate. Table 4-12 summarizes [120] typical coefficients obtained in the-
oretical and empirical investigations. A plot of these coeflicients versus
porosity showed that they are substantially in agreement. Steinour [121]
introduced the concept of immobilized water in his treatment of flocculated
suspensions and showed that by defining

__immobilized fluid volume
solid volume

 

(4-18)

and making appropriate corrections to the volume fraction solids term in
his empirical equations, a good correlation could be obtained for all ma-
terials that were studied. Typical values of a for flocculated [120,121]
ThO, slurries are & ~ 1 to 25, which corresponds to floc densities of from
1.3 to ~6g/cc.

Hindered-settling studies [120] with ThOs slurries having yield-stress
characterization constants, k3, from 50 to 500, in containers having diam-
eters from 1.6 to 10.25 cm, showed that for containers having depth greater
than six times the diameter, the onset of compaction was a function of the
[cHAP. 4

TECHNOLOGY OF AQUEOUS SUSPENSIONS

172

"SwaysAs pajeINIv0y 10§ SNIpeI oINBIpAY — Botndurgy (9) ‘sniper sneipAy — eorundur (p)
"l pug d 10y uotsuadsns jo sorjredosd [eorsAyd Jusn me[ soqorg (9) “Aroayy Feig (q)  eyer Surpyges [ejuswLiadxy () -pInbI] UOI}OBIJ SWIN[OA 0} SIOJOI ? ULID) YL,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

srapmod Lrowd Jo £39 —1).(*m —
[1e1] (7¥61) 1nourg (®) -LIBA 10} G€'() 03 L3°0 ="M %
| 28'0-40 g ezI'0
3 —1 i
[oe1] (6€61) s1omog (3) | <Cm —3) 010
(6511 (8961) WY “1[HoOT ®) 0'1-¢°0 fi — 14 i@ =1)
G
mw
[zz1] (F¥61) Inours}g (p) G8L°0-S'0 (G—1)/¢2 8310
[8z1] (LS6T) 'T® 19 S[[eA®(BA (P) 800 (G—1)/£2901°0
[221] (L¥61) wewryuug (@) 0150 T — m.lmH - L@ —-1DF/8) +1
[921] (0S61) Aorsymey o - 0150 [G—DE —1/0 —1)g'g —] dxe ;3
(—1)8189+ 1
[cz1] (T¥61) s193ing (o) 0180 I
[3z21] ($S61) 1{BZ ‘UoSpIByory ) C6°0-S"0 9 (/) w1 § —(q/0)(e/%) — 5/t +1 O 6
yp I — (g /) J 8
_ I
[zz1] (¥¥61) 1nourng (®) 01590 6—1z8'1-01 22
[e31] (7661) 1{eZ ‘uospIByoIy (®) 0°'1-82°0 ¢o-5?
"ON oyny
D Jo 93usa 5
w %k
90UAIRJIY w0 Asolod
1 QT

°00 =71 .umAQ _ anmQ = 0/)
SNOISNAJISAS 40 ONILLLIS dAYAANIH

Z1-¥ @1av],
4-4] ENGINEERING PROPERTIES 173

 

1.25

o
o

Hindered Settling Rate, cm/sec

0.25

 

 

 

 

0 10 20 30 40 50 60 70 80 90
Angle of Inclination from Vertical, deg

Fia. 4-20. Effect of angle of container inclination on ThOg: slurry hindered
settling rate. Slurry concentration 300 g Th/kg H20, container diameter 1.63 in.

slurry yield stress, slurry concentration, and container diameter, the par-
ticular relation being -

 

2
D= (0.07 £ 0.02)ks 7= o v (4-19)

where the concentration term, ¢., is the value for the onset of compaction.

Settling-rate data obtained [122] in inclined tubes showed a maximum
at an angle of about 50° from the vertical, with typical results being given
in Fig. 4-20. The spread in the data is largely due to the difficulty in dis-
cerning an interface due to supernate rushing upward as the thoria
settles out.

4-4.5 Heat transfer. Grigull [131] has presented a theoretical treatment
of heat transfer to pseudoplastic and Bingham plastic non-Newtonian
fluids for laminar flow through tubes. Theoretical treatments of laminar
heat transfer to pseudoplastic materials have been given for a variety of
boundary conditions [132-134]. Pigford [135] has shown that, for Bing-
ham plastics, the laminar isothermal coefficient should increase less rapidly
than the 1/3 power of the mass flow rate and the magnitude of these
174 [cHAP. 4

TECHNOLOGY OF AQUEOUS SUSPENSIONS
coefficients should be increased by the factor [1+4 (1/9)(7./7,)], approxi-
‘ately. Bailey [136] substituted Eqs. (4-13) and (4-14) into Eq. (4-6)
to obtain an approximate expression for the velocity gradient and then
substituted that expression into Leveque’s solution [137] for the case of
constant wall temperature and uniform fluid temperature at the entrance
to the tube, to obtain:
‘ g'r,,D)”3 .

hD Vp.D2\3
— = 1.615( ) (1 + 24V

k kL (4-20)

 

Experimental data [106] are in substantial agreement with Eq. (4-20).
The principal uncertainty in non-Newtonian heat transfer in the tran-
sition and turbulent region is the criterion for the onset of turbulence and
range of the transition region. Data taken in fully developed turbulent
flow with a variety of solid-liquid suspensions may be correlated satisfac-
torily with the Dittus-Boelter equation (or variations of it), with some un-
certainty about the best viscosity to use in the correlation [138-142].

 

  
   
       

 

 

 

   
 

 

 

 

2

10
g: LI/D T T TTTT | L1 T 1T T T
7 == 2% 0.14 -02 —

- 80 h (%7 Tw DVp
6 - —_—
i ()" ()" o007 (22) % =
4 W —

Jj 3
2 e
3]

103 | =
8 — —
7,_,_ —:u
5T~ T
4 |— DIA. L/D P, Ty, n, -
3 |— in L gmfcc Ibfsq t cp __|

00318 378 167 0066 2.8
2 L vV 1.030 175 1.67 0045 29 _|
-2

1 | .

Og — —

P8I ]
6 — ]
5 — —
4 |— ]
3 .

2 |— fi=4.o log (Np, v/ F)-0.407 _

10-3 I L L I L L L

103 2 3 4 5 678 104 2 3 4 5678 105 2 3 4 56 78]0-6

Reynolds Number, Ng_ = P_:.’.’

FiG. 4-21. Heat-transfer and fluid flow characteristics of ThO2 slurries.
4-4] ENGINEERING PROPERTIES 175

 

F1c. 4-22. Slurry blanket mockup.

Figure 4-21 gives experimental data [106] for heat transfer to the ThO»
slurry with which the pressure-drop measurements of Fig. 4-19 were made.
Comparison of the heat-transfer and pressure-drop data shows that al-
though the onset of turbulence in heat transfer occurs at Reynolds numbers
greater than the 2100 expected for Newtonian fluids, it corresponds exactly
with the experimental critical Reynolds number for the onset of turbulence
obtained from the pressure-drop measurement (which can be predicted
approximately by the Hedstrom criterion). The transition region then
extends to Reynolds numbers a factor of four or five greater than the
critical, as is the case with Newtonian materials. Heat transfer to ThO:
slurries in fully developed turbulent flow is the same as that predicted by
the usual Newtonian correlations [143] to within the precision of the ex-
perimental data. The very interesting question as to whether a suspension
has more desirable heat-transfer characteristics than a pure liquid, as
indicated by the results of Orr and DallaValle [139], or whether the yield
stress of a Bingham plastic decreases the heat transfer coefficient, as sug-
gested by Lawson [142], remains unanswered by these results [106] on
ThOq slurries.
176 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

4-5. OPERATING EXPERIENCE WITH THE HRE-2 SLURRY BLANKET
TEsT FaAcIiLiTy*

4-5.1 Introduction. The purpose of constructing the HRE-2 slurry
blanket test facility was to determine whether the blanket system installed
in the HRE-2 for use with solutions could be used with slurry and, if so,
what modifications would be required. The facility, often referred to as
the blanket mockup, was a full-scale replica of the HRT blanket circulating
system with respect to the reactor-vessel dimensions, piping size, circulating
pump, and heat-exchanger tubing size. The pressurizer and piping con-
figurations were not the same, and the high-pressure heat exchanger con-
tained only one-fifth as many tubes as the HRE-2 heat exchanger. A
general view during the last stages of constructions [144] is shown in
Fig. 4-22. ”

The system was divided into a high-pressure and a low-pressure section,
since the reactor vessel available was a steel prototype, good for only a few
hundred psi, and could not be used during circulation and heat-transfer
tests at high pressure. A schematic drawing [145] of the slurry blanket
mockup system is given in Fig. 4-23. The slurry stream entered the
60-1n.-diameter mockup pressure vessel (which contained a stainless steel
replica of the HRE-2 core) through two inlet nozzles arranged to direct the
flow in opposite directions on either side of the core inlet, with the expecta-
tion [146,147] that the flow pattern in the blanket would consist of two
vortices which would rotate in opposite directions in parallel vertical planes,
thus preventing deposition of sediment on the bottom of the pressure
vessel. The slurry was removed at the top of the vessel by putting a shroud
around the core outlet and removing the fluid through this shroud. It was
also expected [146,147] that as the slurry moved across the top of the core
vessel to the shrouded outlet, the rotating motion produced by the inlet
nozzles would set up a free vortex in a horizontal plane, the accompanying
increase in angular velocity of the fluid as it moved toward the axis tending
to sweep the top of the core vessel free of solids. The design of the inlet and
outlet nozzles was based on successful tests with an 18-in. model of the
vessel [146]. Slurry leaving the pressure vessel flowed to the circulating
pump, a Westinghouse type 230A canned-motor pump having a design
point of 230 gpm at 50-ft head and a working pressure of 2000 psi which
produced a pipeline velocity of 10 ft/sec at its nominal capacity. From the
pump, the slurry stream entered the pressurizer, a 6-in. schedule-160
vertical pipe, with an over-all height of 14 ft 0 in. above the inlet. Entering
flow was directed down and out of the pressurizer with a reducing tee, and

*By D. G. Thomas.
4-5] THE HRE-2 SLURRY BLANKET TEST FACILITY 177

Steam or

Vent
Nitrogen 'f ! en
Y m

Pressure Balancing Valves

Level
Controller
Thermal

Flowmeter

!

 
  
 

 

   
  
   
   
 
  

Condensate
Hold
Tank

B

]k Rotameter

  

:||I|
iyt
l':lll

Pressurizer

:::__ ;o Kw

i

 

 

 

Nozzle ¥

 

 

i
'l

 

300 Gpm 10 Kw % ==
Pump

 

 

 

 

Feed, Letdown & Dump ———a—

 

  

 

 

 

F1g. 4-23. Slurry blanket test system.

baffles were installed above the inlet to damp out any vortices. System
pressure was maintained by a 10-kw heater located just below the liquid
line on the top of the pressurizer. Slurry flowed to the bottom of the mockup
pressure vessel from the pressurizer. The main loop auxiliaries consisted of a
slurry feed system and a letdown and dump system, which provided
system versatility during startup and shutdown. A detailed description
of the system is given in reference [145].

4-5.2 Operation of blanket pressure vessel mockup system. Initial ex-
periments [145] in the 60-in.-diameter vessel (duplicating the HRE-2)
at 170 and 200°C showed that a scale-up based on maintaining equal
superficial vertical velocity at the equator and equal inlet nozzle velocities
was inadequate to maintain a uniform suspension. At 170°C, a sharp
concentration gradient was found, and above 180°C most of the slurry
charge remained essentially stagnant in the blanket, with quite dilute
slurry circulating through the piping loop. The increase in settling rate
with temperature is believed to account for the effect of temperature on
slurry distribution.
178 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

As a result of the initial tests the following changes were made in the
mockup system [148,149]:

(1) The circulating pump was reassembled with a 300-gpm stainless-
steel impeller.

(2) The blanket inlet nozzles were redesigned, with the nozzle diameter
being reduced from 2.16 in. to 1.50 in.

(3) A steam sparging system was installed just above the inlet nozzles to
aid in maintaining the slurry in suspension during a dump.

Although these changes are relatively minor in nature, they resulted in
significant improvement in blanket operation [150]. Hydraulic measure-
ments with water showed that the loop flow was increased from 230 to
350 gpm, the head developed was increased from 50 to 56 ft of fluid, the
pressure drop through the blanket went from 7.3 to 33 ft of fluid, and the
loss across the blanket inlet nozzles went from 5.5 to 25 ft of fluid. During
system operation at 200°C with a slurry made from ThO; calcined at 800°C,
samples were withdrawn from 36 different locations in the blanket vessel.
The mean value of the concentration was 610 g Th/kg H2O with a devia-
tion at the 95% level of 75 g Th/kg H20. The average concentration of the
loop circulating stream was 647 g Th/kg H2O, and an average blanket
vessel concentration of 655 g Th/kg HoO was obtained from a gamma-ray
transmission scan of the vessel. Only about 40 kg (out of a total charge of
1000 kg) was unaccounted for by inventory on the system at this time.
This material appeared to be more or less stagnant in the blanket region on
top of the core vessel and at the bottom of the pressure vessel. The evidence
for this was as follows:

(1) The blanket samples indicated high-conéentration regions on the
top of the core vessel and on the bottom wall of the pressure vessel.

(2) Gamma-ray transmission scans of the blanket vessel indicated the
highest concentration at the top of the core vessel.

(3) Addition of uranium tracer to the circulating stream indicated that
10 to 15 kg of ThO2 was immobilized in the blanket and was only slowly
mixing with the circulating stream; this was supported by examination
of the slurry remaining in the system at the conclusion of the test, when a
pure white layer containing no uranium was found to be partially covered
over with a yellow layer containing a substantial quantity of uranium,
both layers being found on the top of the core vessel.

During this operation at a flow rate of 350 gpm the temperature of the
system was raised and lowered at will between 150 and 200°C with no sig-
nificant change in circulating concentration. That this marked improve-
ment in operating characteristics over the initial tests was due to the small
system changes and not due to a change in slurry properties was indicated
by the results from operation with the electrical frequency of the pump
motor reduced from the normal 60 cycles to a value giving pump flow
4-6] RADIATION STABILITY OF THORIUM OXIDE SLURRIES 179

characteristics similar to those in the initial tests. Reduction of the fre-
quency to 42 cycles with a blanket temperature of 190°C gave a circulating
concentration of 265 versus 625 g Th/kg HoO charged. The blanket
samples averaged 495 g Th/kg H20. These results are quite similar to
those observed in the initial tests.

Operation of the steam sparging system during the dumping operation
at the end of the run allowed recovery of all but 50 liters of the slurry from
the blanket (total blanket volume, 1600 liters), compared with about
200 liters remaining at the end of the dump after the initial tests.

During operation with redesigned inlet nozzles, a number of experiments
were run in which the pump was shut down for periods up to 3 hr. In every
case the slurry settled rapidly to a concentration of approximately
800 g Th/liter and then gradually compacted to a concentration of ap-
proximately 1500 g Th/liter. On restarting the pump, no difficulty was
experienced in resuspending the slurry, and the original slurry distribution
was re-established in approximately 2 min. This proved to be one advantage
of operating with a highly flocculated ThOs slurry compared with operation
with a deflocculated slurry that would settle to much more dense beds
that are correspondingly more difficult to resuspend.

Among the problems that have been more clearly defined as a result of
the tests on the blanket mockup are:

(1) The necessity of establishing the effect of settling rate on the slurry
distribution.

(2) Determination of the scale-up laws for applying small- or intermedi-
ate-scale results to full-scale systems.

(3) Evaluation of the danger of possible boiling at the core wall due to
high heat flux and low fluid velocities.

4-6. RADIATION STABILITY OF THORIUM OXIDE SLURRIES*

4-6.1 Introduction. Experiments have been carried out at the Oak
Ridge National Laboratory as part of a continuing program to determine
the effect of radiation on the physical properties of aqueous suspensions of
thorium oxide. Since changes in particle size, surface properties, and vis-
cosity of the suspension might have a deleterious effect on the operability
of a homogeneous-reactor slurry system, these properties were examined
in detail.

Suspensions of thorium oxide and thorium oxide containing 0.5 mole %,
of either natural or highly enriched UO3; were irradiated in the Low In-
tensity Test Reactor (LITR) at a thermal-neutron flux of 2.7 X 10'3
neutrons/(cm?2)(sec). Although more than 40 irradiations were carried
out, no significant changes in the properties studied were noted [151].

*By N. A. Krohn and J. P. McBride.
180 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

 

0 1 2 3 4
Lbnnbnnrdbinnhintnd

Inches

   

F1c. 4-24. Parts and assembly of in-pile autoclave for slurry irradiations.

4-6.2 Experimental technique. The experiments were carried out in
small, cylindrical, stainless steel autoclaves such as shown in Fig. 4-24.
Thermocouples welded into the bottom closure and a 20-mil-ID steel capil-
lary in the top permitted a continuous measurement of temperature and
pressure. Continuous stirring was accomplished by means of a dashpot
type stirrer of Armco iron clad with stainless steel, with a stainless steel stir-
ring head. The capacity of the autoclave with stirrer was approximately
14 ml.

The stirrer was made to reciprocate at 3 to 4 cycles/sec by alternately
energizing two solenoid coils of double glass-insulated aluminum wire
wrapped around the autoclave body. The timing unit consisted of a multi-
vibrator circuit whose frequency of oscillation and cycle time division
could be controlled by varying resistances in the circuit. The timer unit,
in turn, controlled the grids of two pairs of thyratrons which furnished
power to the solenoids. Stirrer operation was monitored by a tickler coil
connected to an oscilloscope.

The radiation facility consisted of a double-walled aluminum tube ~7/ 8
in. in inside diameter which extended from the top of the LITR through
approximately 20 ft of water into the reactor core. The autoclave was
lowered into this tube on a bundle of wires containing the electrical leads,
4-6] RADIATION STABILITY OF THORIUM OXIDE SLURRIES 181

thermocouples, and pressure capillary. During reactor shutdown the auto-
claves were maintained at 300°C by the heat generated by the stirring
coils. Heat was removed during reactor operation by air which flowed
down the tube, over the autoclave, and back through the annulus pro-
vided by the double-walled tube.

The uranium-bearing oxides were prepared by wet-autoclaving mixtures
of ThO2 and UO3 (about 909, enriched uranium) at 300°C or by co-
precipitating thorium and uranous oxalates and calcining to the oxide.
The autoclaves were loaded with 5 ml of slurry at room temperature, which
expanded to fill half of the autoclave at 300°C. When uranium-bearing
slurries were irradiated, either palladium oxide or molybdenum oxide was
added to catalyze the recombination of radiolytic hydrogen and oxygen
(see Section 4-7). In some experiments an oxygen overpressure of 200 to
250 psi was added at room temperature.

Viscosity measurement. The versatility of the timing device described
above made possible the measurement of relative slurry viscosity both
in-pile and out-of-pile. It was found that the time it takes the stirrer to
reach 1ts maximum height for a fixed solenoid current is dependent on the
viscosity of the autoclave contents. The ‘“‘rise time”’ was determined by
adjusting the frequency and load division so that the upper solenoid was
de-energized the instant the oscilloscope indicated that the stirrer had
reached the top of its travel. The rise time in seconds was then obtained
by dividing the load division by the frequency.

Calibration curves of rise time versus viscosity were made using a
silicone oil of known viscosity at various temperatures. Low viscosity
points were obtained using water and air.

4-6.3 Irradiation results. Viscosities, x-ray crystallite sizes, particle
size, and settled concentrations appeared unaffected by the irradiation, as
indicated by the data shown in Table 4-13. Measurements made on ir-
radiated and nonirradiated materials were the same within the limits of
error. The mean particle sizes were in general not changed significantly;
however, in two cases, involving samples cooled for 10 and 11 months, a
large increase in mean particle size was observed. It is possible that the
long cooling period allowed hard agglomerates to form which were not
broken up by the brief shaking prior to opening of the autoclaves. One ex-
periment cooled 12 months did not show an agglomeration effect. In
general, the irradiated slurries poured readily from the autoclave, and no
tendency toward caking was observed even on samples irradiated in settled
condition for 10 days. |

Chemical and radiochemical analyses of both phases’ of the irradiated
slurries showed the bulk of the fission products, protactinium, and uranium
to be associated with the solids. Only cesium appeared in the supernatant
182 TECHNOLOGY OF AQUEOUS SUSPENSIONS

ErrFeEcT OF RADIATION ON THE PROPERTIES

TABLE 4-13

oF THORIUM OXIDE SLURRIES

[cHAP.

Conditions: 300°C, 2.7 X 1013 neutrons/(cm?2)(sec)

 

 

 

 

 

 

 

ThO2+ 0.59% ThO2+ 0.59%
Property ThO nat. UOg(a)O 2350, ® ¢

Precipitation temp., °C 40 40 10
Calcination temp., °C 900 900 900
Radiation time, hr. 175-300 151-172 168-314
Slurry concentration, 750 750 750

g Th/kg H20 |
Viscosity, cenlistokes

Control 16 — 4-7

During irradiation 16 — 5
X-ray crystallite size, A

Original 350-464 350 —

Irradiated 315-540 295-350 —
Settled conc., g Th/liter

Control 1100 1500 1400

Irradiated 1400 1600-1700 1400-1500
Mean particle size, microns®

Control 1-2 2.8 2.1-2.3

Irradiated and cooled up

to 6 months 1-1.8 1.1-314 1.8

Irradiated and cooled

10-12 months — — 15@

 

(a) 650°C-fired ThO2 wet-autoclaved at 300°C with UO3 - H2O; mixtures refired

at 900°C.

(b) Prepared as in (a) and also by thermal decomposition of the coprecipitated

thorium-uranous oxalates.

(c) Measured by sedimentation in dilute suspension dispersed with 0.005 M

Na4P207.

(d) The result in two experiments; a third test showed no change in average

particle size.

 
4-7] CATALYTIC RECOMBINATION OF RADIOLYTIC GASES 183

in significant amounts. Strontium appeared to absorb less on the higher-
fired materials. The ruthenium analyses were inconsistent, the probable
result of the perchloric acid dissolution treatment used.

Total corrosion-product pickup was similar to that obtained in out-of-
pile experiments except with the slurries containing sulfate, which showed
higher corrosion-product pickup under irradiation. All of the iron and
most of the nickel and chromium were associated with the slurry solids.
The greater part of the corrosion-product pickup resulted from abrasive
attack by the slurry solids under the action of the stirrer.

4—7. CATALYTIC RECOMBINATION OF RADIOLYTIC GASES IN AQUEOUS
TaoriuM OXIDE SLURRIES*

4-7.1 Introduction. Radiolytic decomposition of the aqueous phase of a
thorium oxide slurry blanket will produce a stoichiometric mixture of
deuterium and oxygen which must be recombined. Total recombination
may be accomplished external to the blanket system by suitable methods.
It would be advantageous, however, to recombine the gases internally to
minimize both reactor control problems accompanying bubble formation
and engineering problems associated with external recombination. The
magnitude of the internal-recombination reaction desired may be judged
from the estimate that an aqueous thorium oxide-uranium slurry breeding
blanket may produce from 2 to 3 moles of Hz and 1 to 1.5 moles of O2 per
hour per liter of slurry by radiolytic decomposition, assuming a G-value
of 11 and an average blanket flux of 6 X 10'3 neutrons/(cm?)(sec). This
is within a factor of 2 of the decomposition one would expect for a solution
at the same power density.

Work on the development of a catalyst for use in thorium oxide slurries
to recombine the radiolytic gases has been carried out as a part of the
Homogeneous Reactor Project at the Oak Ridge National Laboratory [152].
More recently, Westinghouse (Pennsylvania Advanced Reactor Project)
undertook similar studies. The experimental approach at both labora-
tories has been similar, and the results are in reasonable agreement. While
sufficient data have been obtained to assure that a catalyst can be used 1n
both thorium and thorium-uranium oxide slurries for complete internal
radiolytic-gas recombination, specific conditions for the most efficient
catalyst preparation and use have not as yet been established.

*By L. E. Morse and J. P. McBride.
tMolecules of water decomposed per 100 ev of energy dissipation in the slurry.
184 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

4-7.2 Experimental techniques and method of analysis. The out-of-pile
tests were carried out in ~15-ml stainless-steel autoclaves provided with a
thermocouple well and a capillary pressure connection through the top
closure. The autoclaves were approximately half filled with the slurry con-
taining an appropriate catalyst, closed, and charged with 550 psi oxygen
and then 900 psi hydrogen from regular high-pressure cylinders. To
minimize corrosive attack at temperature, the oxygen charged was in
excess of the stoichiometric 1:2 ratio to hydrogen. The autoclaves were
raised to temperature in an appropriate furnace, mechanically agitated,
and the decrease in pressure with time was followed continuously by ap-
propriate instrumentation.

For a stoichiometric mixture of hydrogen and oxygen, the rate of pres-
sure decrease was proportional to the total gas pressure in excess of a
certain minimum pressure (i.e., steam and inert gases). A plot of pressure
data in the differential form, AP/At versus P, resulted in a straight line,
the slope of which, k., was a measure of the experimentally observed first-
order reaction rate. | |

While the experimental data are insufficient to establish firmly that the
reaction rate is first order, it is convenient to present the data in this form
and assume a first-order dependence on hydrogen partial pressure:*

d Py,
dt

where Pp, is the hydrogen partial pressure, ¢ the time in hours, and k.
the observed reaction rate constant (hr=1). The moles of H; reacted per
hour per liter of slurry, dn/dt, for any given hydrogen partial pressure
may be calculated, assuming the ideal gas law, from

dn _ k:Pr, V,
dt — RT V,

 

= krPHzr

 

where V, and V, are the gas and slurry volumes and R and T are the gas
constant and absolute temperature. The recombination rates calculated
in this way are conservative in that only the hydrogen removed from the
gas phase is taken into account, and dissolved and adsorbed gases are
ignored. The method of analysis is convenient, however, and permits the
evaluation of the relative activity of the various slurry and catalyst
systems.

The results of the out-of-pile tests were evaluated on the ability to at-
tain a reaction rate equivalent to the consumption of 2 moles of H, per
hour per liter of slurry at 280°C and a hydrogen partial pressure of 100 psi.

*The homogeneous catalysis of the hydrogen and oxygen reaction in the case
of solutions is first order with respect to the hydrogen partial pressure (see Ar-
ticle 3-3.4).
4-7] CATALYTIC RECOMBINATION OF RADIOLYTIC GASES 185

4-7.3 Catalytic activity of thorium and thorium-uranium oxide slurries.
Experiments with several different samples of ThO2 alone showed that
none of the aqueous slurries prepared with the pure oxide possessed the
desired degree of catalytic activity for the stoichiometric gas mixture.
Reaction rates in the region of 300°C were equivalent to less than 0.1 mole
of Hs consumed per hour per liter of slurry at a partial pressure of 100 psi
H,. The higher reaction rates appeared to be associated with the slurries
prepared with oxides having higher surface areas.

A small catalytic effect for the stoichiometric gas reaction was obtained
by incorporating uranium in the aqueous ThO: slurries; however, the
reaction rates remained much too slow to be useful. These tests were carried
out with aqueous slurries of simple oxide mixtures, as well as with a mixed
oxide prepared by calcining the coprecipitated thorium-uranous oxalates.
Heating the slurry of the latter preparation under a small hydrogen partial
pressure at 280°C produced a marked but temporary increase in catalytic
activity which progressively diminished as further gas-recombination ex-
periments were carried out at higher temperatures [that is, dn/dt (at Py,
= 100 psi) = 1.42 at 137°C; 0.65 at 154°C; 0.08 at 282°C].

4-7.4 Survey of possible catalysts. The primary criterion, other than a
satisfactory catalytic activity, for a catalyst to be used in an aqueous
thorium oxide slurry blanket is that it have a low thermal-neutron ab-
sorption cross section. A convenient estimate of the allowable concentra-
tion is provided by the rule that the total cross section of elements other
than thorium which are in the blanket slurry should be less than 109, of
that of the thorium itself.

On this basis, scouting experiments were carried out with copper sulfate,
copper chromite, copper oxide, copper and nickel powders, silver carbonate
(reduces to metal at elevated temperature), vanadium oxide, ceric oxide,
palladium oxide, and molybdenum oxide. Of these only the silver, vana-
dium, palladium, and molybdenum oxide showed sufficient activity at
reasonable concentrations. Silver and vanadium were rejected because of
their higher thermal-neutron cross section; palladium appeared susceptible
to poisoning, particularly at lower temperatures, <120°C, and subsequent
development effort was concentrated on molybdenum oxide.

The preliminary catalyst evaluation was carried out with slurries of
thorium oxide fired at 900°C. Subsequent experience with molybdenum
oxide has indicated that it is inactive with low-fired oxides, and its activity
at least at low concentrations is decreased by the presence of uranium
oxide (see the following discussion). Hence in slurry systems using low-
fired thorium oxide or thorium-uranium oxides, silver, palladium, and
platinum, which are active in these slurries, may prove to be useful [154].
186 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

4-7.5 Molybdenum oxide as a catalyst. The molybdenum oxide cata-
lyst used in the initial scouting studies of its catalytic activity was prepared
by calcining ammonium paramolybdate at 480°C for 16 hr. It was added
to the slurry by dry-mixing 900°C-fired ThOg2, UO3 - H20O, and the MoO3
and then slurrying the mixture in water. Prior to its use in recombination
experiments, the slurry was heated under Q2 at 280°C. Recombination
data obtained with slurries prepared in this way indicated that reaction
rates in the range of 1 to 6 moles of Hs per hour per liter were obtained in
slurries which were heated for 1 hr at 280°C with hydrogen at as low a con-
centration as 0.025 m MoOs. Lower reaction rates (0.03 to 1.8 moles of
H2 per hour per liter) were measured for slurries not treated with Ho.

It was found that the catalytic activities of the activated slurries were
independent of the method used to prepare MoOs. Also, experiments
with MoO2 indicated that it was not stable under the experimental con-
ditions and that this form of the oxide is not the very active catalytic
species produced when the slurries were heated with Hs.

The addition of sulfate up to 10,000 ppm SO4= (based on thorium) did
not appear to impair the catalytic activity of the above slurry containing
0.05 m MoO3. Corrosion products resulting from the attack on the
reaction vessel at the highest sulfate concentration greatly decreased the
recombination rate. Similar slurries to which a ferric oxychloride sol
(1500 to 1600 ppm Fe/Th) was added showed decreased catalytic activities
(approximately one-tenth the iron-free rate), but reducing the iron by
treatment with hydrogen gave catalytic activities equal to or greater than
those of the iron-free systems. The addition of 634 ppm rare-earth oxides
(based on Th)* had no effect on the catalytic activity.

Subsequent experiments indicated that in slurries of mixed thorium-
uranium oxides, the order of addition of the slurry solids, the method of
incorporating the uranium, and particularly the oxide firing temperature
and time were important. It was necessary in some cases to fire mixed
oxides containing 0.5 mole 9, uranium as high as 1000°C for as long as
16 hr to give an active slurry with the molybdenum oxide. Lower firing
temperatures or a shorter firing time at 1000°C did not give an active
slurry.

Oxides fired above 1000°C (4 hr at 1200 or 1600°C) gave very active
slurries. Table 4-14 gives the data obtained with slurries of simple mix-
tures of 1600°C-fired thorium oxide and UQO3z - HoO and with a 1600°C-
fired mixed oxide prepared from the coprecipitated oxalates. Both prepara-
tions gave excellent combination results at 0.05 m MoOs both before and

*Estimated steady-state concentration of fission-product oxides to be produced
in the thorium oxide by irradiation at a flux of 5 X 103 neutrons/(cm?)(sec) and
continuous blanket processing on a 250-day cycle [155].
4-7] CATALYTIC RECOMBINATION OF RADIOLYTIC GASES 187

TaBLE 4-14

ReAcTION RATES oF STOICHIOMETRIC H2-O2 MIXTURES
IN THORIUM OXIDE-URANIUM OXIDE SLURRIES

 

Reacti tar) *
Oxide used for makeup of eaction rate, moles Ha/hr (liter)

500 g Th/kg H20 slurry

 

Slurry as prepared | H-activated slurryt

 

 

 

 

Reaction temperature, 250°C

 

10°C coprecipitated oxalates
(U/Th=0.005) calcined at 1600°C
for 24 hr 0.2 0.2
MoQO3; additions, m:
0.012 6.86 11.0
0.024 6.40 11.0
0.036 11.5 18.1
0.048 15.3 17.1
0.06 15.2 15.0

 

 

 

Reaction temperature, 291°C

 

‘ThO2 calcined at 1600°C for 4 hr 0.02 0.02
0.5 mole 9, U added as UO3 - H20 0.02 0.40
MoO3; additions, m:
0.012 0.08 0.05
0.024 0.10 0.05
0.036 0.24. 0.62
0.048 1.04 3.46
0.06 2.53 4.16

 

 

 

 

 

*At 100 psi partial pressure of hydrogen.
#Slurry heated with hydrogen (250 psi at 25°C) for 2 hr at 270°C.

after treatment with Hz, but at low molybdate concentrations the simple
mixture showed lower activity. Apparent over-all activation energies of
13 to 16 kcal/mole were calculated for the hydrogen-treated slurries con-
taining 0.05 m MoO3 from an Arrhenius plot of the observed reaction
rate constants.

Investigations of MoOjs chemistry by the Houdry Process Corpora-
tion [156] (for the PAR Project) have shown that MoO3 reacts in high-
temperature water with 650°C-fired thorium oxide and uranium oxide.
The failure of some lower-fired material and thorium-uranium mixtures
188 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

to give active slurries may be the result of chemical reaction of the molyb-
denum oxide to form catalytically inactive species.

4-7.6 In-pile studies.* Radiolytic-gas production and recombination
rates were determined in the ORNL Graphite Reactor using a slurry of
ThO2 containing approximately 2.8% uranium which was approximately
93% enriched in U235, The mixed oxide was prepared by coprecipitation
of thorium and uranous oxalates at 10°C followed by calcination at 650°C.

Gas production rates were calculated from the initial pressure rise ob-
served in the autoclave at low temperatures, where the reverse reaction
rate could be neglected. The measured production rates were 1.6 X 10~4,
3.3 X 107%, and 4.9 X 10™* moles of Hz per hour, respectively, for slurry
concentrations of 250, 500, and 750 g Th/kg H2O. The G-value for
water decomposition, assuming an average neutron flux of 4.2 X 101!
neutrons/(cm?)(sec), an energy deposition in the slurry of 170 Mev /fission,
and a fission cross section of 580 barns, was 0.8 molecule per 100 ev.

Gas recombination rates were measured by determining the equilibrium
pressures in excess of steam for various temperatures (Fig. 4-25). Table
4-15 lists rate constants, ks (hr—!), calculated from the equilibrium pres-
sures and gas production rates, assuming first-order dependence, and
making the necessary corrections for volume changes with temperature.
From an Arrhenius plot an apparent activation energy of 16.2 kcal/mole
was obtained. There is some evidence which indicates that the action of
the stirrer increases the reaction rate either by forming catalytically active
corrosion products or by providing reactive steel surfaces.

TABLE 4-15

GAs ReEcoMBINATION RATE CONSTANTS FROM
EQUILIBRIUM PRESSURES IN AN IRRADIATED
THORIUM-URANIUM OXIDE SLURRY

 

 

 

Temperature, °C k., hr1
282 0.45
275 0.26
250 0.15
235 0.10
200 0.03

 

 

 

 

 

*Written by N. A. Krohn.
4-7] CATALYTIC RECOMBINATION OF RADIOLYTIC GASES 189

Temperature, °C

 

 

 

 

 

3 300 250 200
2x10
103 —
8
-
£
§ s -
a.
£
2 —
8
5
o ]
W
2 e
1.7 1.8 1.9 2.0 2.1 2.2
1/Tx 103, k-1

Fig. 4-25. Equilibrium radiolytic gas pressure, excluding steam pressure, in
irradiated thorium oxide slurries.

In all the slurry stability tests in the Low Intensity Test Reactor (see
Section 4-6) where enriched uranium was present, sufficient catalyst was
added to prevent any net radiolytic-gas production. Both PdO and MoOs3
were used for this purpose. No radiolytic gas (<25 psi) in excess of steam
pressure was observed in these experiments. In the tests where only ThO2
was used, no catalyst was necessary.

In an LITR experiment [157] carried out with a 1000 g Th/kg H»O slurry
of 1300°C-fired thorium-uranium oxide (U235/Th = 0.005), sufficient
catalyst (0.02 m MoO3s) was added to give a small partial pressure of
radiolytic gas. A G-value (calculated as above) for gas production of 0.6
molecule of Hs per 100 ev was obtained. From the equilibrium gas pres-
sures at 250 and 280°C, first-order rate constants for gas recombination of
4.98 and 7.14 hr~! were calculated. These values compare well with rate
constants of 4.95 and 8.75 hr~! obtained in out-of-pile gas-recombination
experiments made with a similar slurry of the same oxide at the indicated
temperatures.
190 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

BIBLIOGRAPHY

Thorvum and Thorium Oxide

Thortum, A Bibliography of Unclassified Report Literature, USAEC Report
TID-3309, Technical Information Service Extension, AEC, October 1956.

Davip, Lore 8. (Comp.), Thortum, A Bibliography of Unclassified Litera-
ture, USAEC Report TID-3044, Technical Information Service Extension, AEC,
November 1953.

PraTER, W. D. et al. (Comps.), Thorium, A Bibliography of Published Litera-
ture, ed. by R. A. Allen, USAEC Report TID-3044 (Suppl. 1), Mound Labora-
tory, June 1955.

Sacus, Frances (Comp.), Literature Search on Selected Properties of Thortum
Ozide, USAEC Report AECD-3423, Union Carbide Nuclear Co., 1952.

SweeToN, F. H. (Comp.), Thortum Ozxide, Literature Survey of Preparation
and Properties of, USAEC Report CF-55-8-11, Oak Ridge National Laboratory,
1955.

Thorvum and Thorium-Oxide Chemistry

RoppEN, C. J. and WaRrrF, J. C., Thorium, in Analytical Chemistry of the
Manhattan Project, ed. by C. J. Rodden, National Nuclear Energy Series,
Division VIII, Volume 1. New York: McGraw-Hill Book Co., Inc., 1950.
(Chap. 2, p. 160)

REFERENCES

1. D. F. CroNIN and DixoN CarriuaNn, Critical Mass Studies. Part VII.
Aqueous Urantum Slurries, USAEC Report ORNL-1726, Oak Ridge National
Laboratory, 1954.

2. R. N. Lyon, The Choice in Thortum Ozide Slurries for the Prevention of
Caking in Circulating Systems, USAEC Report CF-57-4-77, Oak Ridge National
Laboratory, 1957.

3. D. G. TrOMAS, paper presented at the 3rd Annual Meeting of the American
Nuclear Society, Pittsburgh, Pa., June 1957. (Paper 14-3)

4. JorN P. McBrIDE, personal communication, December 1957.

5. A. S. Kirzes et al., Uranium Slurry Loops, in Homogeneous Reactor Project
Quarterly Progress Report for the Pertod Ending Mar. 31, 1953, USAEC Report
ORNL-1554, Oak Ridge National Laboratory, 1953. (p. 123)

6. R. K. ScrorieLp and H. R. SamsoN, Flocculation of Kaolinite Due to the
Attraction of Oppositely Charged Crystal Faces, Discussions of Faraday Soc.
No. 18, 135-145 (1955).

7. SHERMAN A. ReED, Oak Ridge National Laboratory, personal communica-
tion, April 1958.

8. J. HappeEL and H. BRENNER, A.I.Ch.E. Journal 3, 506-513 (1957). H. E.
WorreE and G. MurprY, Flow of an Aqueous Slurry through a Vertical Tube,
USAEC Report ISC-874, Iowa State College, 1957.
REFERENCES

191

9. I. KirscHENBAUN et al. (Eds.), Utilization of Heavy Water, USAEC Report
TID-5226, Columbia University, Substitute Alloy Materials Labs., 1951.
10. J. O. BroMekE, Aqueous Uranium Slurries, USAEC Report ORNL-1904,

Oak Ridge National Laboratory, 1955.

11. A. S. Kirzes and R. N. LyoN, Aqueous Uranium and Thorium Slurries,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 9. New York: United Nations, 1956. (P/811, p. 414) (cf. Progress in

Nuclear Engineering. Series IV. Technology and Enginee

McGraw-Hill Book Co., Inc., and Pergamon Press, 1956. (p.
12. R. B. Bricas, Aqueous Homogeneous Reactors for Produ

Power, USAEC Report CF-55-11-35, Oak Ridge National

ing. New York:
317)
cing Central-station

Laboratory, 1955;

(paper delivered at the annual meeting of American Society of Mechanical

Engineers, November 1955).

13. Pennsylvania Advanced Reactor (PAR) Project under

joint contract be-

tween Westinghouse Electric Corporation and Pennsylvania Power and Light

Co., Pittsburgh, Pennsylvania.

14. W. H. ZacuariaseN, The Crystal Structure of 8-UO3-
July 1 to Dec. 31, 1946, of Mass Spectroscopy and Crystal
USAEC Report CP-3774, Argonne National Laboratory, 194

H20, in Report for
Structure Division,
7. (p. 20)

15. R. N. Lyon et al., in Homogeneous Reactor Project |Quarterly Progress
Report for the Period Ending Mar. 15, 1962, USAEC Reportt ORNL-1280, Oak
Ridge National Laboratory, 1952. (p. 83)

16. R. N. Lyon et al.,, in Homogeneous Reactor Project |Quarterly Progress
Report for the Period Ending Mar. 31, 1953, USAEC Report ORNL-1554, Oak
Ridge National Laboratory, 1953. (p. 122)

17. R. N. Lyon et al., in Homogeneous Reactor Project Quc
port for the Period Ending Oct. 1, 1952, USAEC Report ORN
National Laboratory, 1952. (p. 24)

18. W. H. ZAcHARIASEN, Phys. Rev. 73, 1104-1105 (1948).

19. C. D. HooemaN (Ed.), Handbook of Chemistry and
Cleveland, Ohio: Chemical Rubber Publishing Co., 1954.

20. M. Forx, Soc. Chim. (France) 55(16), 231-237 (1949).

21. D. E. FErcusoN et al., in Homogeneous Reactor Projec
Report for the Period Ending Oct. 31, 1953, USAEC Report
Ridge National Laboratory, 1954. (p. 116)

22. O. RurF et al., Z. anorg. u. allgem. Chem. 180, 252—-256 (1929).

23. W. R. Morrt, Trans. Am. Electrochem. Soc. 50, 165-175 (1926).

24. C. L. Duvay, Inorganic Thermogravimetric Analysts. New York: Elsevier
Publishing Co., Inc., 1953. (p. 546) |

25. R. BeckerT and M. E. WinFiELD, Australian J. Sct.
644-650 (1951).

26. R. W. M. D’Evye and P. G. SeELLMAN, J. Inorg. & Nuclear Chem. 1,
143-148 (1955).

27. V. D. ALLrED and J. P. McBRripE, in Homogeneous Reactor Project Quar-
* terly Progress Report for the Pertod Ending Oct. 31, 1966, USAEC Report
ORNL-2004(Del.), Oak Ridge National Laboratory, 1956. (p. 172) (cf. paper
delivered at the 2nd Annual Meeting of the American Nuclear Society, Chicago,
I1l., June 1956.)

irterly Progress Re-
1.-1424, Oak Ridge

Physics, 36th ed.

f Quarterlyy Progress
ORNL-1658, Oak

Research Ser. (A)4,

 
192 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

28. R. N. Lyon et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Pervod Endwng Apr. 30, 1964, USAEC Report ORNL-1753(Del.),
Oak Ridge National Laboratory, 1954. (pp. 169-170)

29. J. E. SavoLAINEN, Oak Ridge National Laboratory, private communica-
tion.

30. D. E. FErcuson et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 1954, USAEC Report ORNL-1678, Oak
Ridge National Laboratory, 1954. (pp. 88-89)

31. R.S. HansEN and R. E. MiNTURN, Iowa State College, 1951. Unpublished.

32. J. P. McBrIpE et al., Preparation and Properties of Aqueous Thorium-
Urantum Ozxide Slurries, paper presented at the 2nd Nuclear Engineering and
Science Conference, Philadelphia, Pa., March 1957. (Paper 57)

33. F. R. BruceE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1953, USAEC Report ORNL-1605, Oak
Ridge National Laboratory, 1953. (p. 138)

34. R. L. PrarsonN et al., Preparation of Thorium Ozxide for Homogeneous
Reactor Blanket Use, USAEC Report ORNL-2509, Oak Ridge National Labora-
tory, 1958.

35. W. H. Carg, Pilot Plant Preparation of Thorium Ozxide, USAEC Report
CF-56-1-50, 1956; see also in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1955, USAEC Report ORNL-1943, Oak
Ridge National Laboratory, 1955. (p. 198)

36. D. G. THomas, Engineering Properties of Slurries, in HRP Ciwilian Power
Reactor Conference Held at Oak Ridge National Laboratory May 1-2, 1957,
USAEC Report TID-7540, Oak Ridge National Laboratory, 1957.

37. E. L. CompPERE and S. A. REED, in Homogeneous Reactor Project Quarterly
Progress Report for the Pertod Ending Jan. 31, 19568, USAEC Report ORNL-2493,
Oak Ridge National Laboratory, 1958.

38. V. KonrLscHUTTER and A. Frey, Z. Elektrochem. 22, 145-161 (1916);
J. Soc. Chem. Ind. (London) 35, 668 (1916).

39. G. W. LeppicorrE et al.,, Oak Ridge National Laboratory, to be issued
(cf. Analytical Chemistry Semiannual Progress Report for Oct. 20, 1954, USAEC
Report ORNL-1788, Oak Ridge National Laboratory, 1954). (p. 21)

40. B. M. Asrauam et al., Particle Size Determination by Radioactivation,
Anal. Chem. 7, 1058 (1957).

41. H. P. Kvue and L. E. ALEXANDER, X-Ray D:ziffraction Procedures for
Polycrystalline and Amorphous Materials, New York: John Wiley & Son, Inc.,
1954. (Chap. IX, p. 491)

42. V. D. ALLRED et al., J. Phys. Chem. 61, 117 (1957).

43. J. P. McBriDE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 19568, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1998.

44. J. M. DaLrLavaLLe, Micromeritics, 2nd ed. New York: Pitman Publishing
Co., 1948.

45. D. E. Fercuson, Oak Ridge National Laboratory, 1954. Unpublished.

46. D. E. FErGUsoN et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Oct. 31, 19563, USAEC Report ORNL-1658, Oak
Ridge National Laboratory, 1954. (p. 114)
REFERENCES

193

47. D. M. RicaARDSON, Adsorption of HoO by ThO2 at High Temperatures,
USAEC Report CF-56-1-109, Oak Ridge National Laboratory, 1996.

48. V. D. ALLRED et al., in Homogeneous Reactor Project
Report for the Period Ending July 31, 1956, USAEC Report
Ridge National Laboratory, 1956. (p. 126)

Ruarterly Progress
ORNL-2148, Oak

49. E. V. JonEs, in Homogeneous Reactor Project Quarterly Progress Report
for the Pertod Ending July 31, 1955, USAEC Report ORNL-1943, Oak Ridge

National Laboratory, 1955. (p. 192)
50. V. D. ALLRED et al., in Homogeneous Reactor Project

Quarterly Progress

Report for the Pertod Ending Jan. 31, 1956, USAEC Report ORNL-2057(Del.),

Oak Ridge National Laboratory, 1956. (pp. 115-120)

51. D. E. FErRGguUsoN et al., in Homogeneous Reactor Project
Report for the Period Ending Oct. 31, 1956, USAEC Report
Ridge National Laboratory, 1957. (p. 151)

Quarterly Progress
ORNL-2222, Oak

52. C. E. ScHILLING, in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1955, USAEC Report ORNL-2004(Del.),

Oak Ridge National Laboratory, 1956. (p. 179)

53. E. V. JoNEs, in Homogeneous Reactor Project Quarterly |

Progress Report for

the Pertod Ending Oct. 31, 19565, USAEC Report ORNL-2004(Del.), Oak Ridge

National Laboratory, 1956. (p. 176)
54. S. A. REep and P. R. CrowLEY, Nuclear Sci. and Eng.
55. S. A. REED, in Homogeneous Reactor Project Quarterly ]

the Pertod Ending Apr. 30, 1955, USAEC Report ORNL

National Laboratory, 1955. (p. 93)
56. V. D. ALLRED et al., in Homogeneous Reactor Project

Report for the Pertod Ending Apr. 30, 1956, USAEC Report

Ridge National Laboratory, 1956. (p. 112)
57. R. B. KorsMEYER et al., in Homogeneous Reactor Projec

Report for the Period Ending July 31, 1956, USAEC Report

Oak Ridge National Laboratory, 1956. (p. 58)

58. C. G. LawsoN, Heat Transfer to Bingham Plastics, Th(
Turbulently in Tubes: An Exploratory Study, USAEC Re
Oak Ridge National Laboratory, 1956.

59. D. G. THomMmas, paper presented at 3rd Annual Meetin
Nuclear Society Held in Pittsburgh, Pennsylvania, June 1957

60. CLypE ORR, Jr., and J. M. DavrvravaLLe, Heat Trax
Liquid-Solid Suspensions, in Chemical Engineering Progress
Vol. 50, No. 9, 29-45 (1954).

61. ALvIN GLASSNER, The Thermochemical Properties of the
and Chlorides to 2500°K, USAEC Report ANL-5750, Argonng
tory, 1957. '

62. B. M. Tareer, Colloid J. (USSR) 6, 545 (1940).

63. J. C. MaxweLL, Scientific Papers, ed. by W. D. N
Dover Publications, 1952.

t
ORNL-2148(Del.),

1ven.

1, 511-521 (1956).
Progress Report for
-1895, Oak Ridge

Quarterly Progress
ORNL-2096, Oak

Quarterly Progress

Do Slurries Flowing
port CF-56-9-132,

g of the American
. (Paper 14-3)

1sfer Properties of
Symposium Series,

Ozxrdes, Fluorides,

» National Labora-

New York:

64. CLype ORR, Jr., and J. M. DarvavaLLe, Heat Tra

Liquid-Solid Suspensions, in Chemical Engineering Progress
Vol. 50, No. 9, 29-45 (1954).

nsfer Properties of
Sympostum Series,

 
194 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

65. S. E. Craig, Jr., Fellowship Progress Report, National Science Founda-
tion Grant: G1616, July 23, 1957.

66. F. H. NorToN and W. D. KiNGERLY, The Measurement of Thermal Con-
ductwity of Refractory Materials; Technical Progress Report, USAEC Report
NYO0-601, Massachusetts Institute of Technology, 1952.

67. K. L. JounsoN et al., Thermal Conductwity of Hot-pressed Thortum Ozxide,
USAEC Report M-3475, Battelle Memorial Institute, 1946.

68. H. R. Kruyr, Collotd Science, Vol. 1. Amsterdam: Elsevier Publishing
Co., Inc., 1952.

69. G. E. ALvEs et al., Chem. Eng. Prog. 48, 385-393 (1952). »

70. C. E. LaprprLE et al., Flurd and Particle Mechanics. Newark, Delaware:
University of Delaware Press, 1954.

71. A. B. METZNER, in Advances tn Chemical Engineering, ed. by T. B. Drew
and J. W. Hoopes, Jr. New York: Academic Press, 1956.

72. D. G. Tromas, Solids Dispersed tn Liquids, USAEC Report CF-56-10-35,
1956.

73. A. EINSTEIN, Ann. phys. 19, (1906); 34, 591 (1911); Kolloid-Z. 27, 137
(1920).

74. E. GurH and R. SimuA, Kollotd-Z 74, 266 (1936).

75. R. Simua, J. Research Natl. Bur. Standards 42, 409 (1949).

76. H. pE BrRuYN, Proc. Intern. Congr. Rheol. (Amsterdam) Pt. 2, 95 (1949).

77. N. Sarro, J. Phys. Soc. (Japan) 5, 4 (1950).

78. V. Vanp, J. Phys. & Colloid Chem. 52, 277 (1948).

79. J. HarpEL, J. Appl. Phys. 28, 1288—-1292 (1957).

80. F. E1ricH et al., Kolloid-Z. 75, 20 (1936).

81. V. Vanp, J. Phys. & Colloid Chem. 52, 300 (1948).

82. A. Bouraric and M. VuiLLAUME, J. Chem. Phys. 21, 247 (1924).

83. S. OpEeN, Der Colloide Schwefel, Nova Acta Regiae Soc. Sci. Upsaliensis,
4, 3 (1913).

84. L. J. GostiNng and M. S. Mogrris, J. Am. Chem. Soc. 71, 2005 (1949).

85. F. Eiricu and R. Simua, Monatsh. 71, 67 (1937).

86. R. Simua, J. Colloid Sct. 5, 386 (1950).

87. L. E. MorsE, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending July 1, 1952, USAEC Report ORNL-1318, Oak Ridge
National Laboratory, 1952. (pp. 82-84)

88. R. E. PoweLL and H. EYriNG, Nature 154, 427-428 (1944).

89. D. G. Tuomas, USAEC Report CF-58-6-3, Oak Ridge National Labora-
tory. (In preparation)

90. E. C. BineuaM, Fluidity and Plasticity. New York: McGraw-Hill Book
Co., Inc., 1922.

91. I. KirsHENBAUM et al., eds., Utilization of Heavy Water, USAEC Report
TID-5226, Columbia University, Substitute Alloy Materials Labs., 1951.

92. D. G. THOoMAS, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending July 31, 1954, USAEC Report ORNL-1813(Del.), Oak
Ridge National Laboratory, 1954. (p. 125)

93. J. N. MukHERJEE et al., J. Phys. Chem. 47, 553-577 (1943).

94. C. H. GiBson, Oak Ridge National Laboratory. (In preparation)
REFERENCES

95. E. BuckiNngaAaM, Am. Soc. Testing Materials, Proc. 21,

96. B. O. A. HepstroM, Ind. Eng. Chem. 44, 651-656 (1952).

97. Y. Ovamo and S. Ito, J. Scit. Research Inst. (Tokyo) 48,
98. R. N. WeLTMAN, An Evaluation of Non-Newtonian Fl

National Advisory Committee for Aeronautics Report NAC]

Flight Propulsion Lab., Cleveland, 1955.
99. D. H. CapweLL and H. E. Bassirt, Trans. Am. In
237 (1941).

195
1154 (1921).
1369-1375 (1954).
pow tn Pipe Lines, .

A-TN-3397, Lewis

st. Chem. Inst. 37,

100. G. W. Howarp, Proc. Am. Soc. Ciwvil Engrs. 64, 1377 (1938).
101. A. P. YurIN, Izvest. Akad. Nauk S.S.S.R., Otdel. Tekh. Nauk, No. 8§,

1146 (1949).
102. M. P.O’Brien and R. G. ForsoMm, Unw. of
3(7), 343 (1937).
103. A. Hazen and E. D. Harpy, Trans. Am. Soc. Ciunl En
104. N. S. BrarcH, Trans. Am. Soc. Civil Engrs. 57, 400 (1

Calif. (Berkeley) Publ. Eng.

grs. 57, 307 (1906).
906).

105. H. C. Warp and J. M. DALLAVALLE, in Chemical Engineering Progress
Symposium Series, Vol. 50, No. 10. New York: American Institute of Chemical

Engineers, 1954.

106. D. G. THoMmas and P. H. Haves, Heat-transfer Charac
ThOs Slurries, USAEC Report ORNL-2504, Oak Ridge Na
(In preparation)

107. V. A. Vanoni, A Summary of Sediment Transportaty

Midwestern Conference on Fluid Mechanics. Minneapolis,

of Minnesota Press, 1953. (pp. 129-160)

teristics of Aqueous
tional Laboratory.

on Mechanics, 3rd
Minn.: University

108. D. M. NewrtT et al.,, Hydraulic Conveying of Solids in Horizontal Pipes,

Trans. Inst. Chem. Engrs. (London) 33, 93-113 (1955).
109. K. E. SpeLLs, Correlations for Use in Transport of Ac
of Fine Solids Through Pipes, Trans. Inst. Chem. Engrs. (i
(1955). .
110. R. A. SmrtH, Experiments on the Flow of Sand-Wate
zontal pipes, Trans. Inst. Chem. Engrs. (London) 33, 85-92 (1
111. A. W. Magris, Fluid Turbulence and Suspended
Technol. 33, 470493 (1955).
112. R. J. BuriaN and GLENN Murpny, Rafio of Solid ]
Velocity tn Slurry Flow, USAEC Report ISC-586, Iowa State

jueous Suspensions
[London) 33, 79-84

r Slurries in Hori-
955).
Sediment, Can. J.

Velocity to Muxture
College, 1955.

113. S. L. Soo, Statistical Properties of Momentum Transfer in Two-Phase

Flow, Chem. Eng. Sci. 5, 57-67 (1956).

114. B. A. BakuMETEFF, The Mechanics of Turbulent F
Princeton University Press, 1941.

115. E. M. LAURSEN et al., Proc. Am. Soc. Cwil Engineers

116. R. J. Atkins et al., Some Rheological and Engineeris
Newtontan Slurries, paper presented at the 3rd Annual
Society Meeting, Pittsburgh, Pa., June 1957. (Paper 14.6)

117. J. D. PERRET, in Homogeneous Reactor Project Quarte
for the Period Ending Oct. 31, 1955, USAEC Report ORN
Ridge National Laboratory, 1955. (p. 71)

118. J. D. PerrET, Data Book No. 1, Oak Ridge National ]
1955. (pp. 91-99)

[ow. New Jersey:

78, D56 (1952).
ng Aspects of Non-

American Nuclear

rly Progress Report
L-2004(Del.), Oak

Laboratory, Oct. 3,

 
196 TECHNOLOGY OF AQUEOUS SUSPENSIONS [cHAP. 4

119. D. G. TroMAs, Atmospheric Pressure System for Determining Resuspen-
ston Velocity of Thorium Oxide Slurrives in Round Horizontal Pipes, USAEC
Report CF-56-10-136, Oak Ridge National Laboratory, 1956.

120. D. G. Tromas and R. M. SummERs, Hindered Settling of Flocculated
Sturries. I. Container Wall Effects, USAEC Report ORNL-2541, Oak Ridge
National Laboratory. (In preparation)

121. H. H. STEINOUR, Ind. Eng. Chem. 36, 618-624 (1944).

122. J. F. RicaarpsoN and W. W. Zaki, Trans. Inst. Chem. Engrs. (London)
32, 35-53 (1954).

123. J. F. RicuarpsoN and W. N. Zaxkr, Chem. Eng. Sct. 3, 65B (1954).

124. J. M. BurGErs, Proc. Koninkl. Ned. Akad. Wetenschap. 45, 126 (1942).

125. P. G. W. HawksLey, Some Aspects of Fluird Flow. New York: Arnold
Press, 1950.

126. H. C. BRINKMAN, App. Sct. Research Al, 27-81 (1947).

127. J. M. DaLvavaLLE et al.,, Application of Hindered Settling to Particle
Size Measurement, personal communication, 1956.

128. A. L. LoEFFLER, JR., and B. F. RutH, Mechanism of Hindered Settling
and Fluidization, USAEC Report ISC-468, Iowa State College, 1953.

129. T. C. Powers, Research and Develop. Labs. Portland Cement Assoc.,
Research Dept. Bull. 2 (1939).

130. H. H. STEINOUR, Ind. Eng. Chem. 36, 840-847 (1944).

131. ‘U. GricuLL, Heat Transfer to Non-Newtonian Fluids for Laminar Flow
Through Tubes, Chem.-Ingr.-Tech. 28, (8/9), 553—556 (1956).

132. B. C. Lycue and R. B. Birp, The Graetz-Nussett Problem for a Power
Law Non-Newtonian Fluid, Chem. Eng. Sci. 6, 35—41 (1956).

133. R. E. GeE and J. B. LyoN, Non-Isothermal Flow of Viscous Non-New-
tonian Fluids, Ind. Eng. Chem. 49, 956-960 (1957).

134. A. B. METzNER et al., Heat Transfer to Non-Newtonian Fluids, A.I.Ch.E.
Journal 3, 92-100 (1957).

135. R. L. Pigrorp, Non-Isothermal Flow and Heat Transfer Inside Vertical
Tubes, in Chemical Engineering Progress Sympostum Series, Vol. 51, No. 17.
New York: American Institute of Chemical Engineers, 1955. (pp. 79-92)

136. R. V. BaiLey, Forced Convection Heat Transfer to Slurries in Tubes,
USAEC Report CF-52-11-189, Oak Ridge National Laboratory, 1952.

137. M. A. LEVEQUE, Ann. mines 13, 2 (1928).

138. C. F. Bonirra et al., Heat Transfer to Slurries in Pipe, Chalk and Water
in Turbulent Flow, in Chemical Engineering Progress Symposium Series, Vol. 49,
No. 5. New York: American Institute of Chemical Engineers, 1953. (pp. 127-135)

139. C. OrRr, Jr., and J. M. DaLravaLLE, Heat-Transfer Properties of Liquid-
Solid Suspensions, in Chemical Engineering Progress Symposium Serzes, Vol. 50,
No. 9. New York: American Institute of Chemical Engineers, 1954. (pp. 29-45)

140. J. C. CHuu et al.,, Heat Transfer Coefficients of Pseudo-Plastic Fluids,
Ind. Eng. Chem. 45, 1686-1696 (1953).

141. J. J. SaLamoNE and M. NEwMAN, Heat Transfer Design Data, Water
Suspensions of Solids, Ind. Eng. Chem. 47, 283-288 (1955).

142. C. G. LawsoN, Heat Transfer to Bingham Plastics, ThOg Slurries Flowing
Turbulently tn Tubes: An Exzploratory Study, USAEC Report CF-56-9-132,
Oak Ridge National Laboratory, 1956.
REFERENCES

197

143. W. M. McApawms, Heat Transmission, 3rd ed. New York: McGraw-Hill

Book Co., Inc., 1953.

144. R. N. LyoN, in Homogeneous Reactor Project Quarter
for the Period Ending Oct. 31, 19556, USAEC Report ORN
Ridge National Laboratory, 1955. (p. 76)

ly Progress Report
[.-2004(Del.), Oak

145. L. F. Parsvry, Jr. et al., Report of Slurry Blanket Test Run SM-2, USAEC

Report CF-57-4-87, Oak Ridge National Laboratory, 1957.

146. C. G. LawsoN, in Homogeneous Reactor Project Quarte
for the Period Ending Apr. 30, 1955, USAEC Report ORNI

National Laboratory, 1955. (pp. 149-150)
147. R. N. Lyon, in HRP Civilian Power Reactor Confe

Ridge March 21-22, 1956, USAEC Report TID-7524, Research

Division, AEC, 1956. (p. 90)

148. R. B. KorsMEYER et al., in Homogeneous Reactor
Progress Report for the Period Ending July 31, 1957, USAEC R
Oak Ridge National Laboratory, 1957. (pp. 59-63)

149. R. B. KorsMEYER et al., in Homogeneous Reactor
Progress Report for the Period Ending Oct. 31, 1957, USAEC R
Oak Ridge National Laboratory, 1957. (pp. 62-69)

150. R. B. KorsMEYER et al., in Homogeneous Reactor
Progress Report for the Period Ending Jan. 31, 19568, USAEC R
Oak Ridge National Laboratory, 1958.

151. N. A. Kronn, Radiation Studies of Thorium Oxide
Ciwvilian Power Reactor Conference Held at Oak Ridge Nationc
1-2, 1957, USAEC Report TID-7540, Oak Ridge National
(pp. 128-142)

152. L. E. Morskg, Catalytic Recombination of Radiolytic

rly Progress Report
-1895, Oak Ridge

rence Held at Oak
| and Development

Project Quarterly
eport ORNL-2379,

Progect Quarterly
eport ORNL-2432,

Project Quarterly
eport ORNL-2493,

Slurries, in HRP
1l Laboratory, May
Laboratory, 1957.

Gases in Aqueous

Thorium Oxide Slurries, in HRP Civilian Power Reactor Conference Held at

Oak Ridge National Laboratory, May 1-2, 1957, USAEC
Oak Ridge National Laboratory, 1957. (pp. 143-154)

Report TID-7540,

153. H. F. McDurriE et al.,, The Radiation Chemistry of Aqueous Reactor

Solutions. Part 3. Homogeneous Catalysis of the Hydrogen
in Nuclear Science and Technology (Extracts from Reactor
nology, Vol. 4), 1954, USAEC Report TID-2505(Del.). (pp.

154. W. D. FrLErcaER and D. E. ByrnEs, Westinghouse E
and Pennsylvania Power and Light Company, 1957. Unpubl

-Oxygen Reaction,

Science and Tech-

13-32)

lectric Corporation

ished.

155. A. T. Gresky and E. D. ArNoLp, Products Produced in the Continuous

Irradiation of Thortum, USAEC Report ORNL-1817, Oa
Laboratory, 1956.

156. W. D. FrLercHER et al., Internal Gas Recombinatio
Advanced Reactor Project Quarterly Progress Report, August
WCAP-668, Westinghouse Electric Corporation and Penns]
Light Company, 1957.

157. E. L. CompeERE and L. F. Woo, In-Pile Experiment-I
geneous Reactor Project Quarterly Progress Report for the Pert
1958, Oak Ridge National Laboratory, 1958. (In preparatio:

k Ridge National

n, in Pennsylvania

31, 1957, Report

slvania Power and

,6Z-1225, in Homo-
od Ending Apr. 30,

)

 
CHAPTER 5

INTEGRITY OF METALS IN HOMOGENEOUS REACTOR MEDIA*

5-1. INTRODUCTION

The materials problems of aqueous fluid-fuel reactors are among the
most challenging in modern technology. The temperatures of interest
alone involve considerable departures from those areas for which scientific
data and techniques for obtaining such data are available. However, these
difficulties are minor compared with those encountered in the actual en-
vironment ultimately to be dealt with in an operating nuclear reactor.
The radiation which results from the nuclear reactions may profoundly
alter the chemistry of the fluids through the formation and decomposition
of various chemical species; similarly, the corrosion and physical behavior
of materials may be changed by radiation damage and transmutation.
Consequently, materials for reactor construction cannot be adequately
specified until their behavior in the ultimate reactor environment has
been evaluated through in-pile radiation experiments. Such experiments
can, however, be safely and meaningfully carried out only after extensive,
careful out-of-pile investigation of the systems of interest. Thus a com-
prehensive experimental program is required involving facilities ranging
from conventional laboratory apparatus to complex in-pile experlments
with associated remote-handling and evaluation equipment.

The comprehensive character of the program is further justified by the
high cost associated with reactor component failure. This high cost stems
from the problems involved in repairing or replacing highly contaminated
equipment. These problems are, of course, magnified by a failure which
results in a release of radioactivity from the reactor, even though this
release is only to a leaktight reactor containment chamber. Thus con-
siderable effort to assure unusual reliability of the reactor system is ap-
propriate.

The corrosion behavior of materials in the fluids of interest has also ap-
preciably increased the scope of the program. Corrosion rates showing
complex time dependence for as long as the first several hundred hours
result from changes in fluid characteristics with time and/or the necessity
of forming a protective oxide film. During this period the rate decreases
with time in a quasi-exponential fashion. The corrosion rate then usually

*By E. G. Bohlmann, with contributions from G. M. Adamson, E. L. Compere,
J. C. Griess, G. H. Jenks, H. C. Savage, J. C. Wilson, Oak Ridge National Labo-
ratory.

198
5-2] EQUIPMENT FOR DETERMINING CORROSION RATES 199

becomes linear and when the materials and environment are compatible
1s normally a few orders of magnitude lower than rates for the initial period.
Flow conditions are also very important variables and so must be considered
in all phases of the program; pump loops have fulfilled this requirement
admirably. This time dependence of the corrosion rate complicates the
interpretation of the data because it is a function of fluid flow, temperature,
and fluid composition. Where tests are run long enough to get good in-
cremental measurements, the resultant linear rates provide a good basis
for evaluating the effects of different variables on corrosion. However,
meaningful corrosion rates cannot be obtained in short-term runs, in which
case a better comparison can be made on the basis of the unit weight loss
over the period. This variable is normally selected for comparing results in
runs less than 200 hr.

H—-2. EXPERIMENTAL LQUIPMENT FOR DETERMINING CORROSION RATES*

5-2.1 Out-of-pile equipment. Static auloclaves. A variety of experi-
mental equipment is required to determine the integrity of metals in
aqueous homogeneous reactor media. The simplest apparatus consists of
a glass flask in which various materials may be exposed to reactor fuel
solutions at temperatures up to the atmospheric boiling point. At elevated
temperatures and pressures, a static autoclave designed to withstand the
temperature and pressure requirements is used with conventional or
specially designed furnaces. A large number of tests may be rapidly and
economically made in autoclaves prior to more extensive testing in dy-
namic corrosion test loops. The static autoclaves are used primarily for
solution corrosion tests. Slurry corrosion tests are usually better con-
ducted in equipment which provides agitation or forced circulation in
order to prevent the slurry from settling into a dense bed.

Toroids. The toroid apparatus or rotator [1,2] provides a method of
circulating fluids, as required to obtain data on the effect of fluid flow on
corrosion rates, without the use of a pump. The toroid itself consists of a
length of pipe bent into a circle with its ends joined. Pin-type corrosion
specimens as well as a thermocouple well are inserted into the toroid
through openings around the circumference. Each pin specimen is held
firmly in a fitting, which also seals the opening in the pipe wall. When the
environment is compatible, a Teflon (plastic) bushing is used to assist in
holding the specimen as well as to insulate it from the holder. Continuous
pressure measurements may also be taken during operation by means of
small-diameter tubing connecting the toroid to a pressure gauge or re-
corder.

 

 
  
 
 
 
 

*By H. C. Savage.

 
200 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

’ Rigid Frame |

- Thermocouple
- and Electrical

~ Connections

Counter Weight

oroid

/ Openings

for Pin
Specimens

 

F1g. 5-1. Photograph of toroid rotator.

The fluid medium in the partially filled toroid is circulated by im-
parting a motion to the toroid similar to that by which liquid may be
caused to swirl in a flask.

To attain the necessary circular motion, the toroid is attached to a
horizontal mounting plate which, in turn, is attached at its center (by
means of a bearing) to the vertical shaft of a rotator arm. The rotator arm
is, in turn, mounted on the shaft of a variable-speed motor, as shown in
Fig. 5-1. The toroid is prevented from rotating about its own axis by
means of a connecting rod installed between the toroid plate and an external
rigid frame. Dynamic balancing of the unit is required for operation at
high speeds (500 to 1000 rpm), and a counterweight is provided for this
5-2] EQUIPMENT FOR DETERMINING CORROSION RATES 201

purpose. Flow velocities up to 30 fps are obtained in the unit now in use.
A heating system, which consists of electric heating wire or elements
wrapped around the toroid, provides for operation at elevated temperature
and pressure. The entire apparatus is relatively inexpensive and fills the
need for a laboratory-scale dynamic corrosion test.

In view of the fractional filling of the toroid, which results in ‘“‘slug
flow,” corrosion attack rates based on the elapsed time of operation may
be corrected to account for the time of immersion in the circulating fluid.
This is usually required when corrosion rates obtained in g toroid are being
compared with those obtained in the dynamic corrosion test loops described
below.

The toroid rotator can be used in the study of solutions and slurries.
It has been particularly useful for slurries because of the small amount of
material required. The exploration of many slurry variables has been
possible with amounts of material (50 to 100 g) easily prepared in the

laboratory.
Filler Cap
— -\

Pressurizer ——

 
  
 

Mixing By-Pass

 

 

i d Joint Pressure Taps
— ANGeCIOMII~_ / Heater

—_——
I

| j?
i \
T \ :l]]{ : |
Corrosion- B
a Specimen Holders n
I }[[[ [

;I]Il: / ]EliCooler?
Heater; |:| U
Solution-Sampling
[:' : Valve

F1a. 5-2. Dynamic solution corrosion test loop.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Flow ——»
by

 

 

1

 

 

 

 

 

 

 

Canned Motor Pump

 

 

 

 

 

 

 

 

Dynamic corrosion test loops. Dynamic solution corrosion test loops de-
signed to operate under the various conditions proposed for homogeneous
reactor operation are the principal experimental equipment used for
out-of-pile tests. One such loop is shown diagrammatically in Fig. 5-2, in

 
202 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Clamping Bands,

 
 

 

" Assembled Holder
Tapered Channel Teflon Tapered Channel

 

 

) Cupon Specimen Holder

Teflon

INCHES
6

 

 

Fig. 5-3. Pin and coupon specimen holders.

which the flow is divided between a number of parallel channels in which
corrosion test specimens can be exposed. Several sets of specimens may be
exposed to several flow conditions or a larger variety of specimens to a
given set of flow and solution conditions. Loops of various, designs have
been constructed for other specific purposes [3,4].

The test loop consists of a circulating pump, pipe loop, and a pressur-
izer. The circulating pump is a Westinghouse model 100A canned-motor
centrifugal type [5]. This pump delivers 100 gpm at a 250-ft-head and
1s constructed for operation at pressures up to 2500 psi. The loop and
pressurizer are heated by means of electric elements cast in aluminum
around the outside of the piping.

The loops are usually constructed of 1i-in. schedule-80 pipe. Flanged
joints are used liberally in the loop construction to provide easy access for
inspection and for connection of special experimental equipment. For
simplicity, economy, and elimination of crevices, a simple butt joint using
lap-joint flanges, bearing rings, and a metal gasket [6] is generally used.
The test specimens are exposed in split-channel holders inserted in flanged-
pipe sections as shown.

The pressurizer, which consists of a vertically mounted section of 4- to
6-in.-diameter pipe, serves several functions. It provides system over-
pressure, contains space for solution expansion in the loop during heatup,
5-2] EQUIPMENT FOR DETERMINING CORROSION RATES 203

 

 

 

 

 

 

 

 

  
  
   

 

  

 

 

 

 

   

 

Strip
Pressurizer Condenser Sysvfem\ —— Heater
__£——‘ 2 Strip
S mu=_ Heat
Condenser Cooler \,_fi:i_* o:z:cris\
. Flange
P
Condensate Reservoir
D
Cooler /
F—ERE“TT Pressurizer
esistance
Distilled Water Addition Heater : Heater
Calibrated Metering —
Orifice \‘\\
Rotameter N
By-Pass Mixing Line I .
Pressurizer
J =+R Corrosion Sample L1
Pressure I 1 I Barrels (3) 1
?Goge — Sample -.--llu jm} 1
> >e 4 Valve r Flo s |
I Rotameter 1 oS Sample and o amp'e
* ] Slurry Addition FIndicator L~ Valve
To Drain 4 System
To |
Drain -Illl --IIIIII -|||||||-A|||||
| L oop .
Cooler Venturi| Tee Converter
Pulsafeeder . 100-gpm Centrifugal .
Slurry Addition . Circulating Pump Cooling Letdown
Tank N To Drain Water Sample Unit

 

 

 

 

 

Purge Water to

To Drain Differential-Pressure Cell

F16. 5-4. Dynamic slurry corrosion test loop.

contains space for gas that may be required for solution stability and/or
corrosion studies, and serves as a reservoir for excess solution for samples
removed during operation.

Two types of holders used for exposure of test specimens in the loop are
shown in Fig. 5-3. The pin specimen holder shown is used to test a variety
of materials at a uniform bulk fluid flow velocity. Pin-type specimens
inserted in the holes are exposed to the solution flowing through the chan-
nel when the holder is assembled. Teflon sleeves on the ends of the pins serve
as compressible gaskets to keep the specimens from rattling in the holder
during the test and to insulate the specimens from the holder. Figure 5-3
also shows the specimens and holder used in determining the effect of
velocity on corrosion. In this holder, flat coupon specimens form a comn-
tinuous septum down the center of the tapered channel, so the bulk fluid
flow velocity increases as the solution traverses the holder. Velocity-effect
data thus obtained may be used in the design of reactor piping systems if
the data are confirmed by loop experience.

In the dynamic slurry corrosion test loop [7], shown in Fig. 54, the
pump discharge flow is directed through the bottom portion of the pres-
surizer to minimize settling and accumulation of slurry particles which
would occur in this region if the pressurizer were connected as in the solu-
tion test loop. As shown in this figure, a condenser is installed in the
204 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

To Pressure

Measuring
Device
Specimen Pins ‘
Area = 2.0 cm?2 Per Pin

Thermocouple Well
Pin lliclck
W Tz

 

 

 

 

W s,
— T

- 2.37in.

(a) Type | Autoclave Containing Pin Type Corrosion Specimens

 

 

 

 

 

 

 

 

 

 

   
 

 
   
 
 
    
   

  

    
  

 

Spacer
N\

\\‘\\ L e ';//\\\ , To Measure
§°’ X / S S N N NN N ’/,Mléfié%‘\‘% —=Measuring

U e 1 ? d b .
NEN / “ MN\\\\\\({’”” Device
N\ s 77z
\§§ § Thermocouple Well

v
§ 2 % Coupon Holder

m\§ Coupon Specimen
\

2-7/8in.

(b) Type Il Autoclave Containing Coupon Type Corrosion Specimens

F1c. 5-5. In-pile rocking autoclaves.

pressurizer vapor space to supply clean condensate required to continuously
purge the pump bearings and prevent slurry accumulation, which would
result in excessive bearing wear. The flow of steam or steam-gas mixture
through the condenser is by thermal convection. The condensate thus
produced flows to the rear of the pump as a result of the static pressure dif-
ference between these two points in the system.

A slurry addition device is also incorporated in the loop so that slurry
may be charged at elevated temperature and pressure. This device con-
sists of a reservoir tank connected at its bottom to the main loop piping.
The slurry may be added batchwise to the tank and, after the top flange is
closed, may be forced into the loop by differential pressure, with the slurry
being replaced by an equal volume of water from the condensate system.
A high-pressure rotometer and regulating valve are provided in the addi-
tion system for metering and adjusting this flow of condensate so that flow
5-2] EQUIPMENT FOR DETERMINING CORROSION RATES 205

to the pump bearings is maintained. To ensure purge-water feed to the
pump bearings and to provide positive feed of condensate for the slurry
addition system in the event of a loss of static pressure differential, a
pulsafeeder pump is incorporated in the condensate circuit and may be
used as required.

As an aid in monitoring the density and thus the concentration of the
circulating slurry, venturi type flowmeters with high-pressure-differential
pressure cells are used in the slurry loop.

5-2.2 In-pile equipment. Rocking autoclaves. To obtain information on
the behavior of fuels and materials under reactor conditions, experiments
have been carried out in the Graphite Reactor and Low Intensity Test
Reactor at ORNL and in the Materials Testing Reactor at NRTS [8-12]
using small autoclaves or bombs. The autoclaves used for such tests are
shown in Fig. 5-56 [9]. The autoclave, in a specially designed container
can and shield plug assembly, is rocked by the mechanism at the face of
the reactor shield. The rocking is designed to keep the solution mixed, to
maintain equilibrium between liquid and vapor phases, and to keep all
surfaces wet. The latter provision prevents the formation of local hot
spots by the high gamma fluxes or localized recombination reaction with
resultant explosive reaction of the hydrogen and oxygen formed by radio-
lytic decomposition of water. The assembly is so designed that the auto-
clave can be retracted into a cadmium cylinder, thereby substantially
reducing the flux exposure. This minimizes the necessity for reactor shut-
downs in case of minor experimental difficulties and is useful for obtaining
data. The necessary electrical and cooling lines are carried to the face of
the reactor shield through the container can and a shield plug. In addi-
tion, a capillary tube, filled with water, connects the autoclave to a pres-
sure transducer gauge. By this means a continuous record of the pressure
in the autoclave is obtained. In cases where a quantitative relationship
exists between the corrosion reaction and consumption or production of a
gas, the pressure measurements, suitably corrected, provide a measure of
the generalized corrosion rate.

Pump loops. An in-pile loop [13] is similar to all- forced-circulation
loops; it consists of a pump, pressurizer, circulating lines, heaters and
coolers, and associated control and process equipment. The circulating
‘pump, designed at ORNL [14-16], is a canned-rotor type which delivers
o gpm at a 40-ft head at pressures up to 2000 psig.

The loop assembly and 7-ft-long container are shown in Fig. 5-6. Some
of the physical data for a typical loop are summarized in Table 5-1 [17].
The loop components are usually constructed of type-347 stainless steel,
although one loop has been constructed in which the core section was made
of titanium and another was made entirely of titanium. The loop con-
206 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 5-1

PuaysicaL Data For TypicaL IN-PiLe Loor

I. Loop volume

1. Pump rotor chamber 160 cc
2. Pump scroll 107 cc
3. Core 300 cc
4. Loop pipe (3/8 in. Sch. 40) 440 cc
5. Pressurizer (1-1/2 in. Sch. 80) 750 cc

 

Total fi57 ce

II. Pipe size

1. Main loop 3/8 in. Sch. 40

2. Core 2 in. Sch. 80

3. Pressurizer 1-1/2 in. Sch. 80

4. Pressurizer bypass line 1/4 in. X 0.049 in. wall tubing

5. Pump drain line 0.090 in. OD-0.050 in. ID tubing

6. Loop drain line 0.090 in. OD-0.050 in. ID tubing

7. Gas addition line (pressurizer) 0.060 in. OD-0.020 in. ID tubing

8. Pressure transmitting line 0.080 in. O0D-0.040 in. ID tubing
(pressurizer)

I1I. Flow rates

1. Main loop 5 gpm (8.5 fps)
2. Pressurizer bypass line 6 cc/sec (1.2 fps)
3. Pressurizer 6 cc/sec (0.034 fps)

4. Tapered channel coupon holder 5 gpm (variable: 10-45 fps)

IV. Capacities* of loop heaters and coolers

1. Main loop heater 3000 watts
2. Main loop cooler 6000 watts
3. Pressurizer preheater 1500 watts
4. Pressurizer jacket heater 400 watts

*Values shown are maximum.

tainer, 7 ft long, is 6 in. in diameter at the core end and 8 in. at the pump
end and is designed to withstand a pressure surge of 500 psi in case of a
sudden failure of a loop component. All electrical and process lines are
carried through the pump end of the container through sealed connectors
and then through the shield plug to the “‘valve boxes” at the face of the
reactor. These valve boxes are sealed, shielded containers in which are
located process lines and vessels, valves, samplers, and sensing devices for
5-2] EQUIPMENT FOR DETERMINING CORROSION RATES 207

Welding Plate Pressurizer

   
  

 

Pressurizer
Heater

  
 
 

  

Beam Hole Loop

HB-2 Container

 
      

Loop Cooler

  
 
   

Holder

    

Line Specimen

Outline of Holders Welding

HB-4 Container Fixture Plate

Graphite
Moderator

     
 

Core

Pressurizer Heater
Litr In-Pile Loop

Fig. 5-6. In-pile loop assembly drawing.

the instrumentation [13]. The equipment in the boxes is used with suc-
cessive loops, whereas each loop is built for a single experiment and com-
pletely dismantled by hot-cell techniques thereafter [18]. Samples of the
circulating solution for analysis are withdrawn through a capillary line
during operation, and reagent additions can also be made if desired. A
capillary connection to the pressurizer is used to follow pressure changes in
the loop and to make gas additions when necessary. As in the case of the
in-pile autoclaves, previously described, the pressure data can be used to
follow a generalized average corrosion rate for all the materlals in the loop
as the operation proceeds.

Specimens can be exposed in the pressurizer, in holders in the line just
beyond the pump outlet, and in the core. The specimens in the core in-
side the tapered-channel holder and in the annulus around it are exposed
to the solution in which fissioning is taking place and to direct pile radia-
tion. Duplicate specimens in the in-line holders are exposed to the same
solution in the absence of the fissioning and direct pile radiation. As in
the out-of-pile loops discussed previously, the tapered-channel coupon
holders are used to study velocity effects. The core specimen holder is
shown in place in Fig. 5-7. Figure 5-8 is a photograph of a core corrosion
specimen assembly in which can be seen coupled coupons on a rod assembly,
as well as coupons in the tapered-channel holder and stress specimens in
the core annulus. Although not shown, impact and tensile specimens are
also sometimes exposed in the core annulus.

The loops must be adequately instrumented [19] to provide accurate
measurements of loop temperatures, pressurizer temperatures, and pressures
within the loop. The quantity of oxygen and hydrogen in the pressurizer
is obtained from the pressurizer temperature and pressure measure-
ments. Electrical power demands of the loop heater furnish a measure
of over-all fission power and gamma heating within the loop. Since the
loop, in effect, becomes a small homogeneous reactor operating sub-
critically, when enriched fuel solution is being circulated many automatic
208 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Body Corrosion Specimen Holder Core Channel Specimens

   

Core Annulus Specimens

Fia. 5-7. Core holder for coupon specimens.

| Cou ple
Specimens

 

Fic. 5-8. Core specimen array.

safety interlocks are used to prevent the possible release of radioactive
material.

Remote-handling equipment. After exposure in the reactor, dismantling
of the loops [18] and autoclaves [20], and subsequent examination of the
test specimens and component parts must be done in hot cells with remote-
handling equipment. Prior to removal from the reactor, the fuel solution
is drained from the loop or autoclave and all radioactive gas is vented.
The loop or autoclave is then withdrawn into a shielded carrier and sepa-
rated from its shield plug to facilitate handling. It is then removed to the
hot cell facilities for dismantling and examination.
52 EQUIPMENT FOR DETERMINING CORROSION RATES 209

Loop Core Clamped in Vise
(Visible Through Window)

 

Fia. 5-9. Exterior view of in-pile loop dismantling facility

 

The operating face of the hot cell for dismantling an in-pile loop is
shown in Fig. 5-9. An abundance of 1- and 2-in-diameter sleeves are in-
cluded for insertion of remote-handling tools; two large zinc-bromide
windows, eight 6-in. zinc-bromide portholes, and six periscope holes are
provided for viewing. The usual hot-cell services, such as hot drain,
metal-recovery drain, hot exhaust, vacuum, air, water, and electricity,
are provided.
210 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

 

Fic. 5-10. Interior view of in-pile loop dismantling facility.

The loop in its container can is lowered from its carrier, through a roof-
plug opening, into the cell, where it is clamped in a chuck which rotates
the assembly while a disk grinder cuts off the rear section of the container.
A jib boom is used to withdraw the loop from the container, and various
sections of the loop are then cut out by means of a disk grinder. Figure 5-10
is a photograph of the dismantling equipment in the hot cell.

The severed loop sections are then removed to a remote-examination
facility [21] where the more exacting tasks of removing, examining, and
photographing loop components, as well as weighing individual test speci-
mens, required to determine corrosion damage, are carried out. Hot-cell
techniques are again used, with the aid of specially designed remote-
handling equipment.
5-3] SURVEY OF MATERIALS 211

5—3. SURVEY OF MATERIALS*

5-3.1 Introduction. To determine the corrosion resistance of many dif-
ferent materials to uranium-containing solutions, a large number of
screening tests have been performed. These tests were carried out either
in pyrex flasks at the boiling point of the solution or in stainless steel
autoclaves or loops at 250°C. Oxygen or air was bubbled through the at-
mospheric boiling solutions, whereas at 250°C the test solutions were
pressurized with 100 to 200 psi oxygen. The corrosion rates of the ma-
terials were determined from weight losses of test specimens after the cor-
rosion products had been removed from their surfaces by an electrolytic
descaling process [22]. The results of the tests are presented in the follow-
ing sections. All tests were carried out in the absence of radiation.

The stainless steel designations are those of the American Iron and Steel
Institute. The composition of most of the other materials is listed in
Engineering Alloys [23] and those not so listed are included in Table 5-2.
All the materials were tested in the annealed condition except in the cases
where parentheses follow the alloy designation, in Table 5-3. The num-
ber thus enclosed is the hardness of the material on either the Rockwell
C (RC) or Rockwell B (RB) scale.

5-3.2 Corrosion tests in uranyl carbonate solutions. Since uranium tri-
oxide is more soluble in lithium carbonate solutions at high temperatures
than in solutions of other carbonates, the corrosion resistance of the
materials was determined only in the lithium carbonate system. All cor-
rosion tests were carried out in stainless steel loops at 250°C. The test so-
lution was prepared by dissolving 0.03 moles of uranium trioxide per liter
of 0.17 m LisCO3 and passing carbon dioxide through the solution until
the uranium was in solution. The solution was then circulated for 200 hr
in a loop pressurized with 700 psi carbon dioxide and 200 psi oxygen. The
flow rate of the solution was 20 fps. Table 5-3 shows the materials that
were tested and the ranges of corrosion rates that were observed.

Thus the corrosion resistance of all the metals and alloys tested was good
with the exception of aluminum, copper, and most of the nickel- or cobalt-
base alloys. The acceptable alloys developed films that would have pre-
vented further corrosion had the tests been continued longer. Data pre-
sented elsewhere [24] show that even the carbon steels and iron would be
satisfactory materials in carbonate systems. The fact that the carbonate
solution is approximately neutral is undoubtedly the reason for the non-
aggressive nature of the solution.

*By J. C. Griess.
[cHAP. 5

INTEGRITY OF METALS IN HOMOGENEOUS REACTORS

212

 

 

 

 

 

 

 

 

 

 

 

17 ‘1eq ‘¢ ug UI}-WINIU0II7
80°0—€0°0 IN

‘eT°0-20°0 ID ‘Z°0-L0°0 °d ‘S'T Ug g Aoreonz
17 °18q ‘GG US 1 Aofeo1rz
LL T8q $°¢ IV 10°0 7 LorreiLy,
LL 89 2T IV 300 X Aoreny,
LLT8q '6°T A ‘9T IV 200 X Aoei,
LL'T8q 7 A 91V AV WnIuEy,
LL 'T8q ‘6’3 US 'S IV LV wnruejy,
LL Teq % U P [V ¥ NV Wniusy,
L 'T8q ‘6 1D ‘€ IV OV Wniueyy,
+66 1L V §F wniuejL,
‘ T8q LT0 CH-1,H 0wy
7P 0D LT M XEUI g G%'% ee ¢¢ Aofrexey
GL'G O 09°0 Teq 800 Z'¥ LT 09¢ Ausysoy
Teq | xeuwrgo'0 | $g'I-08°0 | S 9I-G'¥I [-91 4o[o1)

6%'6-GL'g nD |
GL'GSLT O 01 Teq | x8wW Q0 | 00°9-GL'F L3-6C MOINYAD
¢'T-T OIN Teq | XBW QI°Q g-¢ 08-52 65
g0IV ‘L'0LL g0 Teq 20°0 L L1 M 22¢

“ £1°0-80°0 BL/AN .
¢8'T-00'T M mwwmwm.mosfi XBW ()G 'Teq | XBwWGI'Q | S 9I-G'€T | O'SI-L'¥I N G151 £o[o1)
u ) X 8 BL/AN ¢L'0 Teq | XBW L0°0 62 03 qD 0g Iejuadire)
LON ‘6'g N Teq el 4 2eIT 498
ww ) XOIqN | XBWG'g Teq | XBW 00 G1-€¢1 61-LI 8T
ww ) XgAN | "XBW (g Teq | XBW 800 G131 9%-3% 908 60¢
0162 Teq | XBWGI'Q 9-¥ 61-L1 0%
Y10 UN °d 0 IN 10

uoryBuIIsep Ao[y

o/m ‘uorjisodwo))

 

 

HONVLSISHY] NOISOYYO0() ¥OJd AALSE ], STVIHHLVIN TVIHATS 40 NOILISOdWO) TH],
-G @14v],

 
5-3] SURVEY OF MATERIALS 213

TABLE 5-3

CoORROSION RATES OF SEVERAL ALLOYS IN A SOLUTION
oF 0.17 m Li1pCO3 ConTaINING 0.03 m UOs AT 250° C

Pressurizing gases: 700 psi CO2 and 200 psi Os.
Time: 200 hr. Flow rate: 20 fps.

 

 

 

Metal or alloy Range of avg. corrosion rates,
mpy
Austenitic stainless steels
202, 302B, 304, 304L, 309SCb, 3108, 316,
318, 321, 347 0.8-2.6
Carpenter 20, Carpenter 20 Cb, Incoloy,
Worthite 2.44.0
Ferritic and martensitic stainless steels
322 W, 410, 410 (cast), 414, 416, 420, 430,
431, 440C, 446, 17-7 PH 0.8-2.4
416 (36 RC), 440C (56 RC) 3.6-5.3
Tilanium alloys
55A, 75A, 100A, 150A, AC, AM, AT, AV 0.3-0.6
Other metals and alloys
Zirconium and Zircaloy—2 <0.1
Hastelloy C, Nionel, gold, niobium, plati-
num 0-0.9
Armco Iron, AISI-C-1010, AISI-C-1016 6.4-7.8
Inconel, Inconel X, nickel, Stellite 1, 2, 3,
98M 2, Haynes Alloy 25, copper 16-90
Aluminum > 1000

 

 

 

 

5-3.3 Corrosion tests in uranyl fluoride solutions. In contrast to uranyl
carbonate solutions, uranyl fluoride solutions are acid, and in general the
corrosivity of the fluoride solutions is much greater than that of the car-
bonate solutions. To determine the relative corrosion resistances of many
different metals and alloys, a 0.17 m UO2F2 solution was used. Tests were
performed at 100 and 250°C in static systems and at 250°C in loops. The
static tests were continued for periods up to 1000 and 2000 hr. The dy-
‘namic tests lasted for 200 hr, and the flow rate of the solution past the
specimens was 10 to 15 fps. The results of the dynamic tests are shown in
Table 54.
214 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 54

TaE CORROSION RATES OF SEVERAL ALLOYS IN
0.17m UOFs a1 250° C

Pressurizing gas: 200 psi Oa.
Time: 200 hr. Flow rate: 10 to 15 fps.

 

 

 

. 1 t
Metal or alloy Range of avg. corrosion rates,
mpy
Austenitic starnless steels
304, 304L, 309SCb, 310, 316, 316L, 318,
321, 347 4-13
Ferritic and martensitic stainless steels
322W, 430, 443, 17-4 PH | 3.5-5.0
416 > 2000
Titanium and zirconium alloys
55A, 70A, 1004, 150A 0.01-0.25
AC, AM, AT 5.6-7.7
Zirconium, Zircaloy—-2 > 2000
Other metals
Gold, platinum <0.5
Niobium > 1000

 

 

 

 

All the stainless steels except type 416 showed low corrosion rates under
the conditions of test. Other experiments showed that most of the cor-
rosion occurred in about the first 100 hr of exposure, during which time a
protective coating formed on the steel. After the coating formed, corrosion
rates in the range of 0.1 mpy or less were observed (provided the flow rate
was not too high). The static corrosion tests of longer duration verified
the dynamic results. Thus, had the dynamic tests reported in Table 54
lasted for 1000 hr, the average corrosion rates would have been approxi-
mately one-fifth of those shown.

At high temperatures, all the stainless steels corrode at high constant
rates if the flow rate of the solution exceeds a certain value which depends
on the concentration of the solution and the temperature (to be discussed
in the next section). In 0.17 m UO2F2 at 250°C this critical value is 20
to 25 fps.
5-3] SURVEY OF MATERIALS 215

As expected, zirconium and zirconium alloys showed no resistance to
attack by uranyl fluoride solutions at high temperatures. In fact, other
tests have shown that as little as 50 ppm fluoride ions in uranyl sulfate
solutions leads to appreciable attack of zirconium [25]. On the other hand,
titanium demonstrated high resistance to uranyl fluoride solutions. Tests
with more highly concentrated uranyl fluoride solutions have shown the
corrosion rate of titanium to be low. The titanium alloys showed higher
rates, ranging from 5 to 8 mpy. However, in crevices where oyxgen de-
pletion occurs, titanium and its alloys are severely attacked [26].

No nickel- or cobalt-base alloys were tested under dynamic conditions.
Those tested at 250°C in autoclaves include D Nickel, Chromel P, Stellite
98M2, and nickel; all showed rates in excess of 18 mpy during 1000- to
2000-hr tests. At 100°C nickel corroded at a rate greater than 100 mpy and
Monel corroded at 7 mpy; the other above-listed materials, Elgiloy,
Hastelloys C and D, Illium R, and Inconel, corroded at rates less than
1.5 mpy.

Gold, platinum, and tantalum were practically unattacked by the
uranyl fluoride at 250°C; the resistance of tantalum was unexpected.
Niobium was heavily attacked.

It is seen from the foregoing that several metals, including most of the
stainless steels, have adequate corrosion resistance to uranyl fluoride solu-
tions if the flow rate of the solution is not too great. Uranyl fluoride cannot
be used in a two-region reactor, however, because zirconium and uranyl
fluoride solutions are incompatible. Another possible difficulty of uranyl
fluoride solutions in any system is the fact that the vapor above the acid
uranyl fluoride system may contain some hydrofluoric acid which, if pres-
ent, would present a serious corrosion problem.

5-3.4 Corrosion tests in uranyl sulfate solutions. Many corrosion tests
discussed in detail in HRP progress reports [27] have been carried out in
uranyl sulfate solutions under different conditions of temperature, uranium
concentration, flow rate, etc. Table 5-5 shows representative corrosion
rates of a number of materials obtained in 0.17 m UO2S04 at 250°C in
stainless-steel loops during a 200-hr. exposure. The flow rate of the solu-
~ tion past the specimens was 10 to 15 fps. Static tests, the results of which
are not included in the table, have also been carried out at 100 and 250°C
and generally lasted for 1000 to 2000 hr. The results obtained in static
systems generally confirmed the dynamic results, although the corrosion
rates observed in static systems were less than those measured in dynamic
systems.

The data presented in Table 5-5 show only the relative corrosion re-
sistance of different classes of alloys and need further clarification. From
the results reported in the table, and from many other static and dynamic
216 INTEGRITY OF METALS IN HOMOGENEQUS REACTORS [cHAP. 5

TABLE 5-5

TaE CORROSION RATES OF SEVERAL ALLOYS IN
0.17m UOzSO4 AT 250° C

Pressurizing gas: 200 psi Os.
Time: 200 hr. Flow rate: 10 to 15 fps.

 

Metal or alloy

Range of avg. corrosion rates,

 

 

 

mpy
Austenitic stainless steels
202, 302, 302B, 303, 304, 304L, 309SCb,
3108, 316, 316L, 318, 321, 347, Carpenter
Alloys 10, 20, 20Cb, Croloy 1515N,
Durimet, Incoloy, Multimet, Timken
16-25-6, Worthite 14-65
SRF 1132 190
Ferritic and martensitic stainless steels
Armco 17-4 PH (37 RC), Armco 17-7 PH
(43 RC) 3.14.9
322W, 322W (27-38 RC), 329, 430, 431,
431 (43 RC), 446, Armco 17-4 PH,
CD4MCu, Allegheny 350, Allegheny 350
(38-43 RC), Croloy 16-1, Frogalloy 6—-35
410, 410 (43 RC), 414, 416, 416 (37 ROC),
420, Armco 17-7 PH 46-81
420 (52 RC), 440 C 100-430
Titanrum and zircontum alloys
45A, 55A, 75A, 1004, 150A, AM, Titalloy
X,Y,and Z 0.01
AC, AT, AV 0.03-0.12
Zircaloy-1 and -2, zirconium, zirconium-tin <0.01
Nickel and cobalt alloys
Hastelloy R-235, Inconel X, Stellite 1 77-88
Hastelloy C and X, Haynes Alloy 25,
Inconel, Stellite 3, 6, and 98M2 120-340
Other materials
- Gold, platinum <0.1
Niobium 6.7
Sapphire 17
Quartz 58
Pyrex glass 730

 

 

 
5-3] SURVEY OF MATERIALS 217

tests of longer duration, the following conclusions can be drawn. All the
austenitic stainless steels except SRF 1132, 316, and 316L behaved essen-
tially alike; all corroded rapidly and uniformly for about the first 100 hr,
during which time a protective coating formed (provided the flow rate was
less than 20 fps). Once the film formed, corrosion rates less than 0.1 mpy
were observed. The extent of attack during film formation varied some-
what from run to run, and there was no consistent difference from one
austenitic stainless steel to the other. Thus, even though the rates re-
ported in Table 5-5 are high, the continuing rates are very low; as a class,
the austenitic stainless steels are satisfactory materials for containing
uranyl-sulfate solutions at reasonable flow rates. Types 316 and 316L
showed a tendency toward intergranular attack, and SRF 1132 did not
develop a highly protective coating.

Of the ferritic and martensitic stainless steels, types 410, 416, 420, and
440C in all heat-treated conditions were completely unsatlsfactory The
corrosion rates of these alloys were nearly constant with time, and the
attack was very irregular. Although the precipitation-hardenable steels,
322W, 17—4 PH, and 17-7 PH, showed very low corrosion rates after the
formation of the protective film, all displayed a tendency toward stress-
corrosion cracking in their fully hardened conditions. Croloy 16-1 demon-
strated reasonable corrosion resistance in short-term tests, but long static
tests showed the material to be susceptible to intergranular attack in
uranyl sulfate solutions. The other ferritic and martensitic stainless steels
corroded in the same fashion as did the austenitic stainless steels; that is,
after a protective film formed the corrosion rates were in the range of
0.1 mpy. From a corrosion standpoint these materials have adequate cor-
rosion resistance for use in high-temperature uranyl sulfate solutions.

Titanium, zirconium, and all of their alloys were extremely resistant to
‘attack by uranyl sulfate solutions at all concentrations, temperatures, and
flow rates. In fact, corrosion damage was so small that it was difficult to
detect weight changes of the specimens with a standard analytical balance.
These alloys will be discussed further in Sections 5-5 and 5-6.

Most of the nickel- and cobalt-base alloys listed in Table 5-5 were
rapidly attacked by high-temperature uranyl sulfate solutions. The two
exceptions were Hastelloy R—235 and Elgiloy. Hastelloy R—235 resembled
an austenitic stainless steel in that it developed a film and corroded prac-
tically no further. Although Elgiloy corroded only slightly, it was ex-
tremely susceptible to stress-corrosion cracking in high-temperature
uranyl-sulfate solutions. Thus, of the alloys listed, only Hastelloy R-235
could be considered for use in high-temperature uranyl sulfate solutions.
On the other hand, most of the alloys were resistant to uranyl sulfate solu-
tions at 100°C; corrosion rates less than a few tenths of a mpy were ob-
served. It is particularly significant that many of the very hard alloys,
218 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

such as the Stellites, are resistant at the lower temperature; in many ap-
plications, such as pump bearings, temperatures no greater than 100°C
are required.

In dynamic tests platinum and gold were resistant to attack under all
conditions, but niobium corroded at an appreciable rate, about 7 mpy at
15 fps. The corrosion rate of niobium depended on the flow rate of the so-
lution, and at higher flow rates somewhat higher corrosion rates were
observed. Static tests showed that tantalum and chromium are corroded
only slightly under most conditions. If the solution contained dissolved
hydrogen, tantalum was seriously embrittled; highly oxygenated uranyl-
sulfate solutions at temperatures above 250°C oxidized chromium to the
soluble hexavalent state, and under these conditions the rate of attack
was several mils per year.

A number of nonmetallic substances were statically tested in
0.17 m UO2804 at 100°C. Those tested included various grades of sintered
alumina and graphite, sapphire, silicon carbide, sintered titania and zir-
conia, and quartz. All corroded at rates of less than 3 mpy, except for one
very impure grade of sintered alumina which corroded at 124 mpy. In
fact, the corrosion resistance of the sintered alumina was higher, the higher
the purity. In static tests at 250°C in 0.17 m UO2S0y4, the corrosion rate
of pure sintered alumina was 3 mpy; sapphire corroded at the rate of
3.6 mpy. In dynamic tests under the same conditions, sapphire corroded at
17 mpy, quartz at 58 mpy, and Pyrex glass at 730 mpy. The high corro-
sion rates of quartz and Pyrex glass show why glass-lined equipment can-
not be used for high-temperature experimental work. In addition to the
attack on the glass, the resultant silicates cause precipitation of uranium.

5-3.5 Conclusions. Type—-347 stainless steel has been the basic material
of construction in most homogeneous reactor programs. It has serious
limitations, particularly from the corrosion standpoint, but in considera-
tion of the additional important factors of cost, availability, and experi-
ence with its use, it appears a suitable material. The excellent corrosion
resistance of titanium and several of its alloys makes them very useful in
special applications, particularly where the limitations of the stainless
alloys make their use impractical. Cost, availability, and a recently ob-
served autoignition reaction in oxygen-containing environments are
serious limitations to the use of titanium in pressure-containing equipment
(see Article 5-8.5). Zirconium and zirconium-rich alloys are unique ma-
terials for the core vessel of a two-region breeder.

The behavior of these materials in homogeneous reactor fluids is de-
scribed in detail in subsequent sections of this chapter.
54] CORROSION OF TYPE—347 STAINLESS STEEL 219

5—4. CoRROSION OF TYPE-347 STAINLESS STEEL IN URANYL SULFATE
SoLuTtioNs*

5-4.1 Introduction. The decision to use type—347 stainless steel as the
major material of construction and a uranyl sulfate solution as the fuel
for HRE-1 and HRE-2 was based, at least in part, on the demonstrated
compatibility of the two components and on the fact that the technology
of the austenitic stainless steels was well developed. Articles 5—4.2 through
5-4.7 present the results of an extensive investigation of the corrosion of
type—347 stainless steel in uranyl sulfate solutions in the absence of ionizing
radiation, and in Article 5-4.8 the effect of ionizing radiation on the cor-
rosion of stainless steel is discussed. Further details are reported in the
HRP quarterly progress reports [28].

All the results reported in this section were obtained with solutions con-
taining between 50 and 1000 ppm oxygen to prevent reduction and pre-
cipitation of uranium; the precipitation of uranium as U3Os leads to the
formation of ‘sulfuric acid which attacks stainless steel at a very high rate
at the temperatures of interest. A consequence of these phenomena is the
possibility of localized attack in crevices where oxygen depletion can
occur. Other forms of localized attack have been investigated, but except
for stress-corrosion cracking in the presence of chloride ions (discussed
in Section 5-9) no serious problems have arisen. Thus, even with un-
stabilized stainless steels that have been sensitized by appropriate heat
treatment, no severe intergranular attack in uranyl-sulfate solutions oc-
curs. The coupling of such noble metals as gold and platinum to type-347
stainless steel does not result in accelerated attack of the stainless steel.

5-4.2 Effect of temperature. When type-347 stainless steel is placed in
a uranyl-sulfate solution at temperatures up to 100°C, the steel retains its
metallic luster, and only after long periods of time does it develop a very
thin tarnish film. At higher and higher temperatures the film becomes pro-
gressively heavier, and in the temperature range 175 to 200°C a quite
heavy black scale forms on the surface of the steel in about 100 hr. Up to
about 175°C the film that forms is nonprotective, and the corrosion rate is
dependent on the composition of the solution and independent of the flow
rate past the steel surface. Some typical corrosion rates in this temperature
range are presented in Table 5-6.

*By J. C. Griess.

1Experience indicates such occurrences are minimized by maintaining at least
500 ppm oxygen in solution. Careful attention to elimination of crevices in the
system design is essential.
220 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 56

CoRrRrOSION RATE or TyPE-347 STAINLESS STEEL IN
URANYL SULFATE SOLUTIONS AT 100 To 175°C

 

 

 

 

Solution Temperature, Corrosion rate,
composition °C mpy
0.02 m UO2S04 } 100 0.25
0.006 m H2SO4 150 0.96
0.04 m U02804 150 0.87
0.006 m H2SO4 175 5.4
0.005 m CuSO4
1.3 m UO2S04 100 0.40
125 0.80
150 2.8
175 18.0

 

 

 

 

 

The heavy film that forms in the temperature range 175 to 225°C offers
some protection to the underlying steel, but in most cases the protection
1s poor. - At higher temperatures a heavy scale forms fairly rapidly on the
stainless steel, and once it has been established it affords essentially com-
plete protection against further corrosion, provided the flow rate of the
solution is not too great.

The protectiveness of the film appears to be related to its composition.
At temperatures up to 175°C the scale, as determined by x-ray diffraction,
1s composed of mixed hydrated ferric and chromic oxides. At higher
temperatures the amount of hydrated oxide decreases, and the amount of
anhydrous alpha ferric oxide containing chromic oxide in solid solution
increases. At 250°C and higher, only the anhydrous oxide is found in the
protective scale.

The amount of metal that dissolves during the period of film formation
depends primarily on the flow rate, the composition of the solution, the
temperature, and the presence of additives. If the other variables remain
constant, increasing the temperature decreases the amount of metal that
is corroded during the formation of the protective coating and reduces
the velocity effect.

5-4.3 Effect of solution flow rate. From room temperature to about
175°C the solution flow rate has essentially no effect on the corrosion rate
of the stainless steel. However, at temperatures of 200°C and above, the
corrosion rate is profoundly influenced by flow rate. Figure 5-11 shows
54] CORROSION OF TYPE-347 STAINLESS STEEL 221

160

 

| I l [

140

Q. N
o o
| l

Weight Loss, mg/cm?2
0
o
1

   
 

 

 

 

 

 

60 |- —
40 - —
100 hr
20 —
50 hr
| | | | | |
0 10 20 30 40 50 60 70

Solution Flow Rate, fps

F1e. 5-11. Corrosion of type—-347 stainless steel in 0.17 m U02804 at 250°C as
a function of the flow rate.

typical results obtained using a tapered-channel specimen holder in a
stainless steel loop. The data were obtained from a series of runs in which
0.17 m U02S04 was circulated at 250°C for various times. Since at low
flow rates the corrosion rate depends on time, weight loss rather than cor-
rosion rate is used as the ordinate.

At flow rates up to about 20 to 25 fps all weight losses were nearly the
same regardless of exposure time. Between 25 and 35 fps weight losses in-
creased sharply, and at still higher flow rates the weight loss of the speci-
mens was proportional to the exposure time. An examination of the speci-
mens revealed that up to about 20 fps the specimens were completely
covered with a black, relatively heavy, tenacious scale which, after it
formed, practically prevented further corrosion. In the region of 25 to
35 fps there were areas of the specimens that remained free of scale. At
flow rates greater than 40 fps the specimens did not develop any visible
scale, and upon removal from the holder the specimens had the appearance
of severely etched steel.

Although the results presented in Fig. 5-11 are for only one concentra-
tion of uranyl sulfate, a similar velocity effect is observed at all uranyl-
sulfate concentrations and at all temperatures above 200°C. The tempera-
ture of 200°C is in the middle of the transition region below which velocity
is unimportant and above which a velocity effect is observed. Therefore,
at 200°C a velocity effect is observed but is not well defined. Usually all
specimens, even at the highest flow rate, develop a black coating which
222 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

gives partial protection. In addition, the extent of the velocity effect is
dependent on the exposure time.

The velocity below which a completely protective scale forms is defined
as the critical velocity. The weight loss of stainless-steel specimens above
this velocity increases linearly with time, and the corrosion rate is con-
stant. Below the critical velocity the stainless steel corrodes initially at
the same rate as at the high velocities, but the rate decreases as the pro-
tective coating forms and, generally, after about 100 hr very little, if any,
corrosion occurs. In fact, some specimens have been exposed: continuously
at flow rates less than the critical velocity for periods of time as long as
20,000 hr and the amount of metal corroded was no greater than after
100 hr [29].

5—4.4 Effect of uranyl sulfate and sulfuric acid concentration. All
uranyl-sulfate solutions are acid, and the more concentrated the solution,
the lower the pH as measured at room temperature. The acidity is further
increased by adding sulfuric acid to dilute uranyl sulfate solutions to
prevent hydrolytic precipitation of copper and uranium at high tem-
perature.

Generally, the higher the concentration of uranyl sulfate or free acid in
solution, the greater the extent of metal dissolution during film formation
- below the critical velocity, the lower the critical velocity, and the higher
the film-free corrosion rate above the critical velocity. Table 5-7 shows
how the above three regions change with uranyl-sulfate concentrations
at 250°C.

TABLE 5-7

THE CORROSION OF TyYPE-347 STAINLESS STEEL IN
URANYL SULFATE SOLUTIONS AT 250°C

 

 

 

Concentration Wt. loss Critical Corrosion rate
of UO2S0y, at 10 fps, velocity, at 60 fps,

m mg/cm? fps mpy
0.02 1-2 > 50 ~10
0.11 2-3 25-30 190
0.17 12 20-25 190
0.43 20 10-20 400
0.84 37 10-20 680
1.3 50 10-20 1400

 

 

 

 

 

 
5-4] CORROSION OF TYPE—347 STAINLESS STEEL 223

60

 

Critical Velocity, fps
S 3
| |
|

w
o

 

- N
wn O

 

o

W

Weight Loss, mg/em?2

 

 

 

0 0.01 0.02 0.03
H2504 Concentration, m

Fic. 5-12. The effect of sulfuric acid concentration on the critical velocity and
on the extent of corrosion at low flow rates at 250°C. Solution composition: 0.04 m
U02504 and 0.005 m CuSOs.

Figure 5-12 shows the effect of sulfuric acid added to 0.04 m UO2504
containing 0.005 m CuSOy4, on the amount of metal dissolved during film
formation and on the critical velocity of the system. The effect of adding
sulfuric acid is qualitatively the same at all uranyl sulfate concentrations
and at all temperatures above 200°C and produces a result similar to that
of increasing the .uranyl sulfate concentration. Thus it can be concluded
that the acidity of the solution determines, at least in part, how much
metal dissolves before a protective coating forms, and the critical velocity
of the solution. It has been found that low concentrations of copper sulfate
in uranyl sulfate solutions have no significant effect on the corrosion of
type—347 stainless steel. | |

5-4.5 Temperature dependence of flow effects. It has been stated that
increasing the flow rate of the solution generally produces a detrimental
effect on the corrosion of type—347 stainless steel. This effect is temperature-
dependent, as shown in Fig. 5-13. Although the results are based on a
0.17 m UO2S04 solution, an increase in temperature has a similar effect
at all concentrations. With a given solution composition (i.e., constant
224 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

 

   
  
   

 

12
T T T T 1
100 —
‘E Each Run 100hr
s /5
~
o
£
g\
S
£ S50 |—
D
o
S
25—
275°C
250°C
-
L]

 

 

0 10 20 30 40 50 60
Solution Flow Rate, fps

F1a. 5-13. The corrosion of type—347 stainless steel in 0.17 m U080, at different
temperatures.

uranyl sulfate and sulfuric acid concentrations), increasing the tempera-
ture decreases the amount of metal that dissolves during film formation,
increases the critical velocity, and increases the film-free corrosion rate of
the stainless steel. The reasons for these effects are discussed in Article
5-4.7.

5—4.6 Effect of corrosion inhibitors. The number of substances that
can be added to uranyl sulfate solutions to serve as corrosion inhibitors is
limited for two reasons: (1) all organic inhibitors are oxidized to carbon
dioxide in high-temperature oxygenated uranyl sulfate solutions, and
(2) many inorganic compounds are insoluble in uranyl sulfate solutions.
In spite of the second limitation, a number of inorganic salts have been
added to uranyl sulfate solutions to determine their effectiveness in re-
ducing the corrosion of type—347 stainless steel. Those substances that
have been tested include: Li2SO4, NasSO4, BeSO4, MgSO4, AgsSOy,
CuSO04, Crz(S04)3, NiSO4, FeSO4, MnSO4, Ce(S04)2, [Ru(NO)]2(SO4)s,
NaN03, Na2WO4, Na28i03, NaBiOg, H2M004, H38b04, H7P(MOO4)12,
ASzO5, (UOz)g(PO4)2, NH4TCO4, and K201'207.

Of the compounds listed, only potassium dichromate and certain of the
sulfate salts of the alkali and alkaline earth metals have been effective.
5—4] CORROSION OF TYPE—347 STAINLESS STEEL 225

600

 

 

9,
o
o

A
o
O

Average Corrosion Rate for 100 hr, mpy
S
o

 

 

 

 

 

o
ool 250°C N
275°C
4 [
0 50 100 150 200

Chromium VI Concentration

Fig. 5-14. The effect of Cr(VI) in 0.17 m UO2804 on the corrosion of type—347
stainless steel.

Chromium(VI) was effective even at very low concentrations, whereas the
sulfate salts had to be present in amounts nearly equal in molality to the
uranyl sulfate.

The addition of chromium(VI) to high-temperature uranyl sulfate
solutions reduces the amount of metal corroded during film formation,
greatly increases the critical velocity, but materially increases the corrosion
rate of the stainless steel at flow rates in excess of the critical velocity.
Results obtained using coupon-type corrosion specimens in a loop through
which 0.17 m U02S04 was circulated at 250°C indicated that 200 ppm
chromium(VI) increased the critical velocity from 20 to 25 fps to between
50 and 60 fps [30]. Other tests were carried out in which two pin-type
specimens were exposed for 100 hr at flow rates both above and below the
critical velocity in 0.17 m UO2S04. The results are presented in Fig. 5-14.
The corrosion rates at 70 fps are true rates which would not change on
continued exposure; those at 12 fps are average rates for the 100-hr ex-
posures and would decrease. Had the exposure been for 200 hr, the rates
would have been approximately half those shown. It should be noted that
as the film-free corrosion rate increased, the average corrosion rate at the
low flow rate decreased.

It has been found that the presence of added sulfate salts in nearly
equimolal concentration appreciably reduced the corrosion by concentrated
226 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

uranyl sulfate solutions. The salts that have been studied the most are
beryllium sulfate, lithium sulfate, and magnesium sulfate. The phase
stability [31] of the above solutions has been reported, as have the corrosion
results [32-34]. For example, the weight loss for a 200-hr exposure of
stainless steel in 2.0 m UO2S04 containing 2.0 m Li2S04 at 200°C was
only 5 mg/cm? for velocities up to 60 fps. This may be compared with
100 mg/cm? for exposure to a 1.3 m UO2S04 solution under similar con-
ditions. The effect of lithium sulfate was less pronounced at 250°C and
high velocities; at 250°C and 50 fps, weight losses were approximately 70
and 400 mg/cm?, respectively, for solutions of these same concentrations.
For flow velocities less than 30 to 40 fps initial weight losses were found
to be in the range of 5 to 15 mg/cm? for equimolal concentrations of uranyl
sulfate and lithium sulfate up to 4.4 m and temperatures up to 350°C.

In dilute uranyl sulfate solutions the addition of sulfate salts also reduces
the corrosion of stainless steel, but at temperatures of 250°C and higher
the solutions are chemically unstable and complex hydrolytic precipitates
form. At lower uranyl sulfate concentrations (0.04 to 0.17 m) the solutions
demonstrating the greatest stability are those containing beryllium sulfate,
and of the three sulfates most investigated, the least stable of the solutions
were those with lithium sulfate. Sulfuric acid can be included in such
solutions to prevent precipitation, but in so doing some of the effectiveness
of the sulfate salt is lost. However, addition of both lithium sulfate and
sulfuric acid to dilute uranyl sulfate solutions has been found to result in
improved corrosion resistance of zirconium alloys on in-pile exposure [35].

5—4.7 Qualitative mechanism of the corrosion of stainless steel in
uranyl-sulfate solutions. Although any proposed mechanism for the cor-
rosion of stainless steel in uranyl sulfate solutions at high temperatures
must be considered qualitative, from a study of the effects of several vari-
ables on corrosion and from visual observation of high-temperature solutions
sealed in quartz tubes, certain conclusions can be drawn and the over-all
reaction processes determined.

The austenitic stainless steels and most other metals and alloys of
practical importance in large-scale homogeneous reactors are thermo-
dynamically unstable in aqueous solutions and depend on protective films
for their corrosion resistance. Fortunately, when austenitic stainless
steels are oxidized in high-temperature uranyl sulfate solutions, the steel
oxidizes uniformly so that no element (or elements) is leached preferentially
from the alloy. However, not all the alloying elements contribute to film
formation.

The oxidation and reduction processes of the corrosion reaction can be
considered separately, with the oxidation reactions (considering only the
major alloying elements) represented in the following manner:
5—-4] CORROSION OF TYPE-347 STAINLESS STEEL 227

Fe, Cr, Ni —> Fe(II) + Cr(11I) 4 Ni(II) 4 e, (5-1)
Fe(II) — Fe(IlI) 4 e. (5-2)

Both iron(II) and nickel(II) are soluble in uranyl sulfate solutions to an
appreciable extent at high temperature, but if an oxidizing agent more
powerful than uranyl ions is present, the reaction indicated by Eq. (5-2)
takes place. [Manganese(II) also remains in solution.]

The half-reaction for the reduction can be any of the following, depending
on the conditions:

Oz + 4H* + 46— —> 2H0, (5-3)
UO2(IT) 4 4H+ + 26~ —> UIV) + 2H0, (5-4)
2H* + 2¢7 —> Ho. (5—5)

In most concentrations of uranyl sulfate at temperatures greater than
200°C (and probably even at somewhat lower temperatures), iron (III),
chromium(III), and uranium(IV) are present in solution only briefly
before hydrolyzing in the following manner:

9Fe(IIT) 4 3H20 —> Fes03 + 6H, (5-6)
2CI‘(III) + 3H20 —>- Cro03 4+ 6H+, (5—-7)
QUOQ(II) -+ U(IV) + 4H-0 —- U305+ SHT. . (5—8)

In a system containing oxygen the total reduction process is represented
by Eq. (5-3), although it is highly probable that other ions enter into the
reaction mechanism. Only in the absence of oxygen is either U3Os or
hydrogen found in the system. |

If we examine the equations as written, it is apparent that the over-all
corrosion of stainless steel in uranyl sulfate solutions containing oxygen
can be represented by the sum of Egs. (5-1), (6-2), (5-3), (5-6), and (5-7),
taking into account the percentage of each element in the alloy. From the
over-all reaction it can be seen that the amount of oxygen consumed can
be used to measure the quantity of stainless steel corroded. In addition,
- since the nickel originating from the corrosion process remains soluble, the
total quantity of stainless steel oxidized can be determined from a knowl-
edge of the nickel content of the uranyl sulfate solution. Measurements of
nickel content and oxygen consumption have been found to agree well
with weight loss for determining total corrosion. Since the nickel remains
soluble and replaces hydrogen ions, the pH of the solution slowly increases
as corrosion proceeds.

In a system depleted in oxygen, uranium is reduced and precipitated as
U30s, producing hydrogen ions while ferrous ions remain soluble. Conse-
quently, a pH measurement gives an indirect measure of the presence or
absence of oxygen in the system. Similarly, the presence of ferrous ions
even in very small amounts indicates a deficiency of oxygen and that
uranium is beginning to precipitate from the solution.
228 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Once an anhydrous protective coating has formed on the surface of the
stainless steel, the steel corrodes at an extremely slow rate. However,
since the oxide coatings are thick, it is probable that they are under stress,
and fine cracks or imperfections form. When the film-free metal at the
base of the crack is exposed to the solution, the metal again corrodes actively
until the corrosion products repair the crack. During extended periods of
testing, this process is probably repeated many times over the surface of
the specimen, with the net observable result being a uniform attack of the
specimen.

In the proposed mechanism of film formation, both iron and chromium
form ions before undergoing hydrolysis and crystallization on the surface
of the stainless steel. This mechanism can account for the existence of a
critical velocity, since ions and hydrolyzed particles can diffuse into the
circulating stream before forming the film if the diffusion layer is too thin.
Since the diffusion-layer thickness depends on the flow rate or turbulence,
the process of film formation is, in essence, in competition with diffusion
and turbulence. The critical velocity, then, is that flow rate (or turbulence)
at which the diffusion layer is reduced to such an extent that most of the
corrosion products get into the main circulating stream before they can
form on the surface of the stainless steel. Under conditions where the
rates of dissolution and hydrolysis are fast, the critical velocity would be
expected to be relatively high.

It has been mentioned that solutions of uranyl sulfate are acid and that
the higher the uranyl sulfate concentration, the lower the pH of the solu-
tion. Also, data have been presented to show that the higher the uranium
concentration, the greater the corrosion rate of film-free stainless steel.
These two factors essentially work against each other with regard to film
formation; that is, a high corrosion rate would introduce relatively large
concentrations of corrosion products into the solution in the immediate
region of the stainless steel surface, a factor which should facilitate the
formation of a protective film. But even though more corrosion products
are present, the hydrogen ion concentration is also substantially greater.
Since hydrolysis reactions are very dependent on the hydrogen ion con-
centration, the rate of film formation is actually retarded, so that the net
result is the dissolution of more stainless steel at high uranyl sulfate con-
centrations than at lower concentrations before a protective film is formed.
Because the process of film formation is slower the higher the uranyl sulfate
concentration, the critical velocity is also lower the higher the uranyl
sulfate concentration.

Since the presence of sulfuric acid in uranyl sulfate solutions also de-
creases the rate of hydrolysis, the addition of sulfuric acid has about the
same effect as increasing the uranyl sulfate concentration.

The effect of temperature on the formation of a protective coating can
5-4] CORROSION OF TYPE-347 STAINLESS STEEL 229

be accounted for by considering the effect of temperature on the rate
at which the corrosion products are formed and on their rates of hydrolysis.
Increasing the temperature increases the rate at which corrosion products
enter the solution and also the rate of hydrolysis, with the net result that
films form faster and that the critical velocity is increased. Because the
initial corrosion rate and the rate of precipitation are fast, it would be
expected that the crystallite size of the oxide would be smaller the higher
the temperature of formation, and this has been found to be true. Appar-
ently a compact layer of small crystals forms a more protective and ad-
herent coating than one composed of large crystals.

It is probable, though seemingly paradoxical, that chromium(VI) serves
as a corrosion inhibitor, at least under certain conditions, because it acceler-
ates the film-free corrosion rate of stainless steel. In this respect, the effect
of adding chromium(VI) to uranyl sulfate solutions is similar to that of
increasing the temperature. Because of the increased corrosion rate, large
quantities of corrosion products are formed near the surface of the stainless
steel, and hence the solubility limit of the oxides is exceeded rapidly and
many small crystals form on the surface of the stainless steel. Even though
the rate of hydrolysis is not increased as it is at the higher temperatures,
the diffusion rate is slower, so that the corrosion products are in the vicinity
of the stainless steel longer. Because the film forms faster, the critical
velocity is higher in the presence of chromium(VI) than in its absence, other
variables remaining the same.

The reason why the addition of relatively large quantities of inert
sulfate salts to uranyl sulfate solutions reduces the corrosiveness of the
resulting solutions is not known, but may be due to one of a number of
factors, such as the formation of stable complexes, reduction of acidity,
changes in oxidizing power, increased viscosity and density, or changes in
colloidal properties of the oxide.

5-4.8 Radiation effects.* The effect of radiation on the corrosion be-
havior of stainless steel is important in relation to the use of this material
in construction of the fuel-circulating system and reactor pressure vessel of
a homogeneous reactor. The radiation levels thus encountered, however,
are considerably less than for the zirconium core tank. For example, the
fission power density of the solution in contact with the stainless steel in
the HRE-2 will not exceed 1 kw/liter for operation at 10 Mw, and in a
large-scale two-region breeder or single-region burner reactor would not be
more than 5 kw/liter. These values may be compared with fission power
densities of up to 50 kw/liter at the surface of the zirconium core tank.

The corrosion behavior of type-347 stainless steel in uranyl-sulfate
solutions under irradiation at high temperature has been studied in a

*By G. H. Jenks.
230 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

number of in-pile loop and in-pile autoclave experiments. A few scouting
type experiments of the effect of van de Graaff electrons on steel corrosion
have also been carried out. Although most of the information for the radia-
tion effect on steel has its source in these experiments (in particular, the
loop experiments), information of a general nature has been derived from
the performance of the stainless steel portion of the HRE-1.

With loop experiments, specimens were located in the core portion of
the loop and in portions of the loop external to the core and out of the
high-flux region [36], as shown in Fig. 5-6 (Article 5-2.2). Coupons and
other specimens exposed in these latter positions are designated ‘‘in-line
specimens.”’ Specimen preparation, in general, comprised only cleaning
after machining. However, prior to radiation exposure, each experimental
system, including the specimens, was exposed first to water, with or without
oxygen, and then to oxygenated uranyl sulfate solution at a temperature
near the test temperature [37]. Corrosion attack of a specimen was deter-
mined from weight-loss measurements, visual inspection, and metallo-
graphic examination. The course of corrosion of the system during exposure
in both loops and autoclaves was followed by measuring the rate of oxygen
consumption. The approximate operating conditions and other experi-
mental information for the loop tests are shown in Table 5-9 (Article 5-5.3).
All the loops contained steel specimens except those otherwise indicated.
Experiment L-2-14 contained only a few steel coupons. A detailed de-
sceription of methods and procedures employed with the in-pile tests is
presented in Article 5-5.3.

The behavior of steel under exposure to a solution in which fissioning is
occurring in the immediate neighborhood of the surface appears to change
appreciably with changes in experimental conditions. No comprehensive
picture of the behavior is available as yet, and only the general features of
the experimental results can be reported. Pits of 1 to 2 mils in depth,
which appear to spread laterally and merge (with increasing attack), are
usually found on the surfaces of specimens which have suffered appreciable
attack [38]. After the merging of pits, the attack appears to proceed
fairly uniformly over a surface.

Average corrosion rates which have been determined for core specimens
in 0.17 m UO2S04 solution plus varying amounts of HaSO4 and CuSO4
varied between 0.1 and >180 mpy for solution power densities up to
5 kw/liter and for solution velocities up to 45 fps [39]. These rate values
for specimens and the other rates quoted below are based on exposed
specimen areas and radiation time.

With the exception of the results of two experiments, the rates observed
at power densities below 2 kw/liter were less than 2 mpy. In one of the
experiments, 1.—4-8, for which the results are an exception of this generali-
zation, a specimen exposed to a solution power density of about 2 kw/liter
5—4] CORROSION OF TYPE—-347 STAINLESS STEEL 231

and an average solution velocity of about 40 fps exhibited a rate of 40 mpy.
The pattern of the results in this case indicated that the high solution
velocity was, in part, responsible for the high rate. In the other experi-
ment, DD, rates varying from 2.5 to 57 mpy at solution power densities
from 0.3 to 1.7 kw/liter and solution velocities from 10 to 40 fps were
observed. This experiment was the first of the series of loop experiments.
The higher rates in this loop may have been associated with the exceptional
solution conditions. Only a small amount of excess acid was added initially,
and it was estimated that this excess was consumed in the solution of
nickel produced in corrosion during the course of the experiment. This
solution condition was not repeated in subsequent experiments [40].

No significant acceleration of corrosion has been observed with specimens
exposed at in-line positions in the loops; that is, at positions in which the
sample is not exposed to neutrons but is exposed to solution which has
passed through the core. Average corrosion rates for in-line specimens were
less than 2.3 mpy for all cases considered. These in-line specimens were
exposed to solution velocities in the range of 10 to 40 fps, but no significant
effect of velocity on the corrosion rate was observed [41]. However, in
one experiment, L-4-12, a single pit was observed in the surface of the
recessed shoulder of the volute inlet of the pump, a high-velocity region
[42]. It should be noted that the rates mentioned are based on the loss in
weight of a specimen during exposure as measured after a cathodic defilming
treatment. The oxide is not always removed quantitatively in this treat-
ment. Weight gains were frequently observed for in-line specimens, and
in these cases the rate of attack was assumed to be zero.

The results of the experiments with van de Graaff electrons have not
shown any significant effect of electron irradiation on the corrosion of
type—347 stainless steel by uranyl sulfate solutions at the high temperatures.
One of these experiments, carried out in a type—347. stainless steel thermal
siphon loop, was with an oxygenated 0.17 m UO2SO4 solution at 250°C.
The intensity of electron irradiation was such that the estimated power
density from absorption of electron energy in solution adjacent to the
specimen was 20 to 30 kw/liter. Two exposures, each of 50-hr duration,
were made, and no significant difference was observed between the attack
of irradiated and nonirradiated specimens [43]. The other experiment
was conducted in a titanium thermal loop with an oxygenated solution,
0.04 m UO2S04, 0.025 m HoSO4, and 0.01 m CuSO4 at 280°C. The
estimated power density due to absorption of electron energy adjacent
to the specimen was about 60 kw/liter. Again no significant difference
between the attack of irradiated and nonirradiated specimens was noted
in a 50-hr exposure [44].

It has been suggested that chemical changes in solution due to fissioning
and/or the radiation from fission-product decay, both of which would be
232 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

proportional to the total fission power averaged over the total solution
volume of the high-pressure system, might have an adverse effect on the
behavior of the stainless steel located in the regions external to the core.
In loop experiments at average power densities up to 1.5 kw/liter, no
such effect was apparent. The only available information on the corrosion
of type—347 stainless steel at higher average power densities is that from
the HRE-1. In this case the maximum average power density was about
14 kw/liter, and the over-all stainless steel corrosion rates (including the
core tank) were 6 to 8 mpy as judged from data for total nickel in
solution [45].*

The influences of some of the variables on the corrosion behavior of
stainless steel under exposure to a fissioning solution, as indicated by the
results to date, are listed and summarized as follows:

Fission power density. Corrosion is accelerated by exposure to fissioning
uranyl sulfate solution, and the degree of acceleration increases with
increasing fission power density in solution. In several experiments, reason-
ably good proportionality was found between the fission power density
to which a specimen was exposed and the logarithm of the average corrosion
rate of the specimen during exposure.

Ezxcess H2S0O4. One interpretation of some of the results is that the
optimum concentration of excess H2SO4 is between 0.01 and 0.02m, and
that concentrations above and below these limits may have an adverse
effect on in-pile corrosion. ~

Velocity of solution at specimen. There is some evidence that the in-pile
rate increases with increasing solution velocity.

Radiolytic-gas pressure. A possible interpretation of some of the results
is that the in-pile rate is diminished as the concentration of radiolytic gas
in solution increases.

Galvanic effects. The possibility of galvanic effects on corrosion under
some in-pile conditions has not been ruled out.

Temperature. In general the attack observed at 250°C was greater and
less predictable than that observed at 280°C.

5-5. RADIATION-INDUCED CORROSION OF ZIRCALOY—2 AND ZIRCONIUMT

5-5.1 Introduction. Crystal-bar zirconium and Zircaloy—-2 have been
tested at elevated temperatures in uranyl sulfate solutions under dynamic
and static conditions in the absence and in the presence of radiation. Both

*Specimens exposed in the circulating lines before and after the core experienced
similar attack rates.
1By G. H. Jenks.
5-5] RADIATION-INDUCED CORROSION 233

the metal and alloy are very corrosion resistant in the absence of radiation.
However, under exposure to nuclear radiations, particularly those from
fissioning uranium solution, corrosion rates may be appreciably greater
than those observed out-of-radiation. Although most of this section will
be concerned with the in-pile studies, some of the out-of-pile tests will also
be described.

5-5.2 Corrosion of Zircaloy-2 and zirconium in uranyl sulfate solutions
in the absence of radiation. Most of the radiation-free experiments were
of the type in which specimens of the metal to be tested were exposed in
autoclaves or loops constructed of stainless steel or titanium. The uranyl
sulfate solutions were oxygenated and usually contained excess H2SOg.
CuSO4 was also added in some tests. Cleaned, as-machined specimens
were employed. Under exposure to the uranyl sulfate solutions at high
temperature, the specimens generally formed a black, tightly adhering
film within 100 hr. These films could not be removed without damage to
the metal, and the amount of corrosion was estimated from the weight of a
specimen together with the oxide.

Zirconium and Zircaloy—2 specimens exposed in solutions circulating in
stainless steel systems collected some of the stainless steel corrosion prod-
ucts (iron and chromium oxides) in an outer layer of scale. This outer
layer could be removed partially by a cathodic defilming operation. A
sodium hydride bath treatment was required for complete removal.
Table 5-8 lists values for long-term average corrosion rates observed in a
solution 0.04 m in UO2S04, 0.02 m in HaSO4, and 0.005 m in CuSO4 at
200, 250, and 300°C.

TABLE 58

LoNG-TErRM CoORROSION RATES OF ZIRCONIUM AND ZIRCALOY—2
IN URANYL SULFATE SOLUTIONS

 

 

 

 

 

 

 

 

 

Corrosion rate, mpy
Material
200°C 250°C 300°C
Crystal-bar zirconium <0.01 <0.01 0.13
Zircaloy—2 <0.01 <0.01 0.04

 

These rate values were calculated from the decrease in weight of de-
filmed specimens as measured following an initial exposure period of several
days. Other tests gave similar results at uranyl sulfate concentrations from
234 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

0.02 to 1.3 m. No evidence for appreciable acceleration of attack during
exposure was observed during tests of up to 20,000 hr.

A few experiments with Zircaloy—2 were carried out in autoclaves similar
to the in-pile type [46] described in Article 5-2.2. In these experiments,
the autoclave was constructed of Zircaloy—2 and was charged with the
test solution, oxygen gas, and specimens. The test surfaces were as-
machined and cleaned. Corrosion during exposure was determined by
measuring the rate of oxygen loss within the system. The results of two
of these experiments, H-54 and H-55, are shown graphically in Fig. 5-15,
where the thickness of the layer of metal which was converted to oxide or
removed by oxidation is plotted on a log-log plot versus the exposure time.

Data reported by Thomas [47] on the behavior of Zircaloy—2 in de-
aerated water at temperatures of 290 and 315°C, when converted into
values for the thickness of the layer of oxidized metal, are similar to those
for H-54 and H-55. The water data at 315°C are illustrated in Fig. 5-15
by the dotted line. From these results, it appears that the out-of-pile
behavior of Zircaloy-2 in oxygenated uranyl sulfate solutions is similar to
that in deaerated water in the temperature region investigated.

5-5.3 Methods and procedures employed with in-pile tests. In-pile in-
vestigations were carried out with autoclaves and loops (described in
Article 5-2.2). A majority of the autoclave experiments and all the loop
experiments were conducted in the Low Intensity Test Reactor at ORNL.
The maximum thermal-neutron flux. available for these experiments was
about 2 X 1013 neutrons/(cm?)(sec).

For most experiments, the autoclave and the corrosion specimens were
constructed of the metal to be studied. The autoclave was partly filled
with the solution to be tested, pressurized with oxygen, and stirred by
rocking. Measurements of the oxygen pressure in the system during ex-
posure provided information regarding the course of corrosion.

Most of the in-pile loops were constructed of type—-347 stainless steel;
however, in one experiment the loop was of titanium, and in another the
core portion only was of titanium. Corrosion specimens were placed in the
core and in a portion of the loop which is outside the region of appreciable
neutron flux. The specimens in the latter position were usually duplicates
of the core specimens. The over-all corrosion behavior during exposure
was followed by measuring the rate of oxygen consumption within the
system, and by chemical analysis of solution samples which were with-
drawn from time to time. Specimen examination included visual and me-
tallographic inspection, and weight measurements. The exposure conditions
and composition of test solution which prevailed for each of the in-pile
loop experiments are listed in Table 5-9. Unless otherwise noted, all tests
were run with a light-water solution of 0.17 m UO2SO4 in a type—347
RADIATION-INDUCED CORROSION 235

5-5]

 

 

 

 

 

 

 

 

TOS0N “ %070 (U) "TOSND w ST'0 (3) 'suauIIoads [999s sso[ure)s oN (J) 'doo] YG.-1L TV (9) 'Apoq "S'S L¥E
0100 Y611 (P)  "9wdA[0s O3 (9) '18d 06-01¢ eanssexd 20 FQSON “ F0°0 () *ATuo susuItoads [999s Sse[urelg (%)
Z¥9 1201 zg’ I'1-2°S Gey 003 610 020" —-810° (0) 8T-¥-1
G.8 eeet 8% ¥ 1-4"9 00§ 0p1 ¥20° 820" %20 () 21-2-1
01S 241! 081 e 7807 082 0S 0% gy L8 | (8-9) $1-C-1
GLL 0021 CH- e 1-L°¢ 08z |’ 01T ¥20° 920" —220° 911
42 €8 0L'1 9 7861 8.3 0.8 820" 080" —220° Q121
8L g131 GG* I'T-€°9 053 082 G100 130" —010° (0) €11
GGzl 6981 ¥ 8 0-2'¥ 082 0LT 810" 120" —%10° (qQ) 01-2—'1
cov1 eL13 L9° 6°0-8'F% 053 08T 180 Y0  —L30° (P) 211
GZL e0zT 29 e 1-0"2 052 022 220 0%0° —800° 151
344 6TLT cg- L 0-1'S 052 0LT €20’ W0  —L10° 8——1
L€S LG., 8% 0'1-L°¢ 052 0LT 900" 800" —¥$00° HH
L68 6911 QL e 1-8% 053 033 920" ev0° —S10° (1) D9
L9V 918 qg” L 0-L"¢ 053 A 820" 0%0 0-120° () 44
082 0S¢ 0S80 € 0-L'T 052 0Z1 €00°0 0-800°0 () da
.HQ .HQ hmu:\ 3& h@ufl \ Bum
fiwkom %MEM%W d%w%% ‘suourtoads 9 (48 “19917/00) w w
- * 0 ¢ {
cofimwaw%fi& wnwvw.flmgo.p 18109 m: o onom MWQ@ ‘ureans Edmmww% | ‘umey PduEy zoquinu door]
JO Qu g, [B10L, 93BIOAY i P d Jre orjA[orpelr
‘dura g, ol
payewunIsy
ouII T, Ayisuap 1omod UOISST T YOSCH S$s90xXT

 

 

 

 

 

 

 

 

SLNHWIMHJX d007T ATIJ-N] 404 SNOILIANO)) ONILVYHAd() HE%EHNOMAH&AN

6—C ATAV],
 

 

 

 

 

 

 

 

 

 

 

 

  
  
 
 

 

 

236 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS
Solution Composition
Exp. L. . m Temp |[Slope
No. [UO2SO4|CuSO4| Other °C
H-54 0.34 | 0.04 O9 250 | 0.41
H-55 0.17 | 0.02 o)) 290 | 0.34
1.0 H H-57 0.17 | 0.02 |O9,Hg,Rad| 250 |1 4
E H-58 0.17 | 0.02 | O,Hg,Rad | 250 | 1 =
— H-58 ]
» O01E =
€ — o =
~ — Ho2O 315°C ]
£ F ==
n
o — —
o
v 001E Specimen =
— H-55 Weight Data
- -
0007 bl LUULINL L L IULI L b1 L
0.1 1 10 100 1000
Time, days

[cHAP. 5

F1g. 5-15. Comparison of Zircaloy-2 corrosion in various environments.

stainless steel loop. CuSO4 concentrations from 0.008 m to 0.07 m were
employed. Oxygen pressures ranged from highs of 140 to 195 psi to lows
of 35 to 90 psi. Pressurizer temperatures were normally 15 to 30°C higher
than the temperature of the main stream.

The solution compositions for a majority of the autoclave experiments
are summarized in Table 5-10. In the usual experiment, the fission power

TABLE 5-10

SUMMARY OF SoLUTION COMPOSITION FOR AUTOCLAVE RADIATION
CORROSION EXPERIMENTS WITH ZIRCALOY—2

 

 

 

 

 

Component, Concentration
U02804 (909, enriched uranium) 0.17m
CuSO4 0.01-0.04 m
H2SO4 0-0.04m
Excess Oz 900-20 pst
Radiolytic gas 0-500 psi

 

 

density was 20 w/ml or less. In a few exceptional autoclave experiments,
power densities considerably greater than 20 w/ml were achieved [48].
5-5] RADIATION-INDUCED CORROSION 237

Operating temperatures ranged from 225 to 300°C, with most experiments
at 250 or 280°C.

The method of filling and sealing many of the early autoclave experiments
included air at atmospheric pressure within the system. Solution analyses
after irradiation indicated that most of the nitrogen from this air was
fixed in solution in some oxidized form with possible concentrations (if
all NO3™) up to or about 0.05 m [49].

The usual preparation of specimen surfaces consisted of cleaning after
machining. Before exposure to radiation, the systems, with specimens,
were first subjected to high-temperature water, with or without oxygen,
and then to the uranyl sulfate solution at temperatures near the test tem-
perature. The loop systems were exposed to solutions of normal uranyl
sulfate before the exposure to the U235-enriched uranyl sulfate solutions.

The thermal neutron flux to which a Zircaloy—2 specimen was exposed
while in-pile was determined by measuring and comparing the amount of
the induced activities, Zr®>-Nb%, in the specimen and in a control sample
of Zircaloy—2. The latter was irradiated together with a cobalt monitor in a
separate experiment in which the specimen and cobalt did not contact
solution. For some experiments which contained steel specimens, similar
measurements were made utilizing the Crd! activity. The fission power
density in the uranium solution adjacent to a specimen was calculated
from the resulting value for the thermal flux together with the value for
the fission cross section of uranium, assuming 200 Mev per fission.

Recently, comparisons have been made between values for the flux
indicated by cobalt monitors within the test system and by Zircaloy—2
within the same system. The flux values determined from the cobalt
were about 259, less than those from Zircaloy—2. The cobalt values may
be more valid, but the consistency between Zircaloy—2 results has been
good, and the Zircaloy-2 values are employed throughout in presenting
the results [50].

Experimental conditions other than those summarized above have been
employed, and some of these will be mentioned later.

5-5.4 Results of in-pile tests with Zircaloy-2 and zirconium. The re-
sults of autoclave experiments have demonstrated that the corrosion rate
of Zircaloy-2 under irradiation at a given set of exposure conditions does
not change with the time of exposure, in contrast to out-of-pile behavior.
Two 1n-pile autoclave experiments, H-57 and H-58, plotted in Fig. 5-15,
illustrate this [51]. The lines through the plotted points for tests under
radiation exhibit a slope of 1, and those for out-of-pile experiments, a slope
of about 0.4. In both cases the values for penetration are those calculated
from the results of oxygen consumption measurements. It was assumed in
the calculations that corrosion of the exposed surfaces in the autoclave
238 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

was uniform and that oxygen was lost only in the formation of ZrOs.
These assumptions appear to be generally valid for the in-pile cases. The
corrosion penetration calculated from weight changes of specimens was
usually in near agreement with the penetration calculated from oxygen
data. A pitting type attack was detected in only one of the numerous
specimens which have been exposed [562]. The cause of pitting in this case
is unknown.

The time scale which is employed for the radiation experiments is
cumulative time with the reactor at power. During exposure of a given
experiment, the reactor was frequently at zero power for appreciable
periods. It has been observed in autoclave experiments that a penetration
of from 0.005 to 0.01 mil takes place after shutdown and that, following this,
corrosion essentially stops until power operation is resumed. The corrosion
which takes place during shutdown appears to delay the onset of attack
when irradiation is reinstituted; the amount of delay about balances off the
additional attack which occurs during shutdown. In view of this behavior,
the practice has been adopted of using radiation time only in calculating
corrosion rates [53].

The oxide which is found on Zircaloy—2 surfaces under irradiation 1is
markedly different from that formed outside of radiation. The in-pile
specimens which are removed from the autoclaves are covered with a
relatively heavy brass-colored scale. This scale cracks and flakes off to
some extent upon drying. When a specimen is immersed in acetone and
then air-dried at 100°C, the scale flakes off almost quantitatively [54]; re-
maining scale can be removed by cathodic defilming. The specimen sur-
faces after defilming generally exhibit a dark-gray color and, in some
cases, interference colors. In-pile loop specimens, on the other hand, ap-
pear to be free of heavy scale of the type found in autoclaves. There 1s
usually a dark-brown-colored scale on the surfaces, the amount of which
is greatest for specimens exposed at the lowest power densities. The scale,
of unknown composition, can be removed in a cathodic defilming opera-
tion [55].

Chemical analyses of heavy scale on autoclave specimens revealed that
they contain several percent by weight of uranium and copper in addition
to zirconium. Sulfate is found in some scales [56]. Core specimens from
loop experiment L-2-17 have been analyzed for uranium as removed from
the loop. Twenty micrograms of uranium per square centimeter were
found on the surface of a specimen exposed in a location at which the solu-
tion velocity was about 1 fps. Five micrograms per square centimeter
were found on a specimen exposed at a solution velocity of about 30 fps [57].

Average corrosion rates during exposure to radiation have been calculated
from the observed loss in weight and the exposed specimen area, not In-
cluding areas in close contact with adjacent specimens or covered by the
5-5] RADIATION-INDUCED CORROSION 239

 

4 T T T T T T T T T T T T T T T T

[~ Loop Experiments Represented
. oC
20 L Curve Experiment No. q
A FF, GG, EE, L-4-8, o~
_ L-4-12,1-413, Pl _
(Channel) ~
16 |- B L-4-16,L-2-15 ~
(Channel)
C L-2-15 (Annulus)
12 - D L-4-18 (Channel
and Annulus)

 
 
 
 
 
   
   
 

Corrosion Rate, mpy

 

 

 

— E L-2-14 (Channel =
s L and Annulus) &5\230 C i
— (A) 250°C —
4 —
(D) 235°C .
| | | | | | | | | | | l | | | | |
0 2 4 6 8 10 12 14 16 18 20

Fission Power Density in Solution, kw/liter

F1a. 5-16. Radiation corrosion of Zircaloy-2 in 0.17 m UO2SO4 solutions. Loop
results at 235, 250, and 280°C.

 

24 T T I | l I
Numbers in Parenthesis Refer
to Solution Velocity fps

o
|
OO
so/:
\
|

o® Ny . (10)

    
   
 

Corrosion Rate, mpy
o
1

v ]
V4
(])//
/7
0.04 mUO,SO,
8- Channel Specimens OQSOA");SG’}
(10) (10) (40) m\UOL—T

 

 

 

Fission Power Density in Solution, kw/liter

Fia. 5-17. Radiation corrosion of Zircaloy-2 solutions. Loop results at 280
and 300°C.
240 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

specimen holder. In some cases, where specimens had different ratios of
total to exposed areas, results indicated that the rate of attack of the cov-
ered areas was nearly the same as that of the exposed areas. However, use
of only the exposed area gives more conservative values. For most of the
loop specimens, the ratio of total to exposed area was about 1.6, and in
autoclaves the ratio was nearly 1.

Loop results obtained with 0.17 m UO2SO4 solutions but with other
conditions varying are illustrated in Fig. 5-16 [58]. Except for relatively
minor variations, the solutions for the experiments represented by curves
A, B, and C were the same. Experiment L-2-14, curve E, was with a
solution containing 0.4 m H2SO4 and 0.15 m CuSO4 as additives. KEx-
periment L-4-18, curve D, was with a D20 solution. Channel specimens
were exposed to solution velocities in the range 10 to 40 fps; annulus
specimens to velocities of about 1 fps. Where annulus type specimens were
employed, they usually exhibited rates somewhat greater than the channel
specimens. The annulus type specimens in loop L—2-15 (curve C) represent
the maximum difference observed in 0.17 m UO2SO4 solutions. No dif-
ference between channel and annulus specimens was observed in experi-
ments L.-4-18 and 1.-2-14 (curves D and E). The 280°C data represented
by line B follows the equation

R = 1.04P(1 — =95/, (5-9)

where R = rate in mpy, and P = power density in kw/liter. The data
for 250°C (line A) follow the equation

R =125P(1 — ¢ 6-5/R"%). (5-10)

The derivation of an equation of this general form is given in Article 5-5.6.
Data represented by the other curves are considered insufficient to justify
representation by specific equations.

Results obtained with loop experiments with 0.04 m UO2SO4 solution
(L-2-10 and L-2-17) are illustrated by the curves in Fig. 5-17 [59]. A
portion of line B from Fig. 5-16 is included in Fig. 5-17 for comparison.
Corrosion rates for channel specimens in this solution are greater at the
same power density than those for specimens in 0.17 m UO2SO4 solutions.
The shape of the curve through the channel data is also different. As will be
discussed later, these channel data may be interpreted in terms of a bene-
ficial effect of solution velocity on corrosion. The approximate locations
of specimens in the channel with respect to the power density to which the
specimen was exposed and the average velocity of solution at the specimen
are indicated in Fig. 5-17. The annulus specimens, which were exposed at
low solution velocities, corroded at higher rates than the channel specimens.
5-5] RADIATION-INDUCED CORROSION 241

The results of autoclave experiments at 250°C (curve A, Fig. 5-16) with
solutions of the same general compositions as those listed in Table 5-9,
but with 0.04 m excess H2SO4 [60], followed Eq. (5-10), developed for
250°C loop data. The rates obtained in autoclave experiments with similar
solutions at 280°C [60] followed the equation

R=1.36P(1 — ¢~ 95/R"%) (5-11)

which illustrates that the rates were somewhat greater than those obtained
with loop channel specimens at 280°C (curve B, Fig. 5-16). In autoclave
experiments with solutions of the same general composition but with no
excess HoSO4, the data at 250°C [60] followed the equation

R = 29P(1 — 6_6-5/121.5)’ (5—12)
and the data at 280°C followed the equation
R =2.45P(1 — ¢~ 95/B"), (5-13)

Rates obtained with 0.04 m UO2SO4 solutions in autoclaves were in ap-
proximate agreement with the values observed for annulus coupons in
loop experiments with 0.04 m solutions (Fig. 5-17) [60]. The effects of
changes in some other variables on the in-pile rate are described below:

Time. As mentioned previously, the rates under irradiation appear to
remain fairly constant with time.

Ozxygen and hydrogen. No effect has been noted of changing the pressure
of hydrogen within the test limits, 0 to 350 psi, and of oxygen within 150 to
900 psi.

CuS0O4. CuSO4 concentration apparently has little effect on the rate.
The results of one autoclave experiment, Z-17, in which no copper was
employed confirmed this behavior [61].

Nitrogen. The presence or absence of air at atmospheric pressure in the
~autoclave when sealed apparently has little effect on the rate [62].

Other solution variables. The presence of LiaSO4 as an additive in the
uranyl-sulfate solution has resulted in a decreased radiation effect in some
autoclave experiments [63]. No appreciable effect on the rate has been
noted for other solution additives such as CrOz and KTcO4 [64]. An
experiment to which MoO3s was added indicated an adverse effect for this
material in solutions [64].

Other materials. Crystal-bar zirconium was tested in several loop and
autoclave experiments. Observed corrosion rates were in near agreement
with those for Zircaloy—2.
242 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Chemically polished* specimens of Zircaloy—2 and crystal-bar zirconium
corroded at rates about 109 lower than the as-machined specimens.

Other zirconium alloys have been tested in some experiments, but only
one of these alloys, Zr-15% Nb, has exhibited appreciably better resistance
than Zircaloy—-2. This niobium alloy has been tested in several loop and
autoclave experiments. In the beta-quenched condition, it corroded at
rates lower than those observed for Zircaloy-2 by factors of from 3 (run
L-4-11) to 1.5 (run L.-2-17) [65].

Zircaloy—2 autoclave experiments in the LITR and MTR reactors with
solutions which contained UO2SO4 depleted of U235 have exhibited an
acceleration of corrosion over that expected out-of-radiation. In these
tests, absorption of fast neutron and gamma rays gave rise to a power
density in solution of about 6 kw/liter in MTR experiments and 0.6
kw/liter in the LITR experiments. In each reactor, about one-half of the
power was from gamma-ray and one-half from fast-neutron absorption. At
280 to 300°C the observed rates were about 1 and 4 mpy at the lower and
higher powers, respectively [66].

5-5.5 Tests of the effect of fast-electron irradiation on Zircaloy-2 cor-
rosion. The effect of fast electrons from a van de Graaff accelerator on the
corrosion of Zircaloy-2 by uranyl sulfate solutions has been tested in
experiments at 250 and 300°C. In both cases the specimens were mounted
in a small thermal siphon loop constructed of titanium. Employing elec-
trons with an initial energy of 1.5 Mev and currents in the neighborhood of
2 X 10 /(cm?)(sec), the experimental arrangement was such that an
electron traversed approximately one-half of its range before impinging on
the specimen. The estimated power density due to absorption of beta-ray
energy in the solution adjacent to the specimen was 60 to 90 kw/liter in
each experiment. The duration of exposure was 40 hr at 250°C and 60 hr
at 300°C. No significant acceleration of the corrosion due to irradiation

was found in either case [67].

5-5.6 Discussion of results of radiation corrosion experiments. The
results of the radiation experiments presented above suggest. that the
radiation effect on the corrosion of Zircaloy—-2 by uranyl-sulfate solution
is not directly associated with changes in solution under irradiation. If
the radiation chemistry of these solutions at high temperature is not greatly
different from that of acid solutions near room temperature, the yield of
H-atoms and OH radicals in solution from beta and gamma radiation is
at least as great as that from heavy particles [68]. Since radicals produce
marked changes in reactions, rather than the molecular products H2 and

*509, H20, 459, concentrated HNOs, and 59, HF (489,).
5-5] RADIATION-INDUCED CORROSION 243

Og2, corrosion effects due to changes in solution would be as pronounced
for beta-gamma radiation as for heavy particles. However, no effect of
beta radiation was observed. It was also mentioned that no effect has been
observed of varying Ho and O2 pressures in in-pile autoclave experiments.
Thus the primary action of the radiation is in the metal or its protective
oxide film, or both, probably through the formation of interstitials and
vacancies by heavy nuclear particles. Beta-gamma radiations can also
produce interstitials in such materials, but with considerably less efficiency.
The number of such defects produced by an electron in the energy range
employed in the van de Graaff experiments may be expected to be a factor
of 100 or more less than the number produced by a 1-Mev neutron [69].
Hence the rate of defect formation by the electrons was small compared with
the rate of formation by fast neutrons in the depleted UO2SO4 in-pile tests.

On the basis of these considerations and of the various experimental
observations, a qualitative model of the radiation corrosion of Zircaloy—2
has been developed, from which an equation relating fission power density
in solution and corrosion rate has been derived [70]. The equation has the
general form

R=AP(1 — ¢ BIR™), (5-14)

where R is the corrosion rate, P is the fission power density, and A and B
are constants. This equation is derived as follows.

Under exposure to heavy-particle radiation, the protective oxide film
forms on the metal as it does out-of-radiation, and the kinetics of the pro-
tective oxide formation are about the same in and out of radiation. Under
irradiation, however, the film does not continue to increase in thickness.
Radiation produces defects of unspecified nature in the protective oxide,
and in the presence of these defects, the oxide breaks up and/or reacts
with the solution to form a nonprotective scale. The rate at which the
protective oxide breaks up is proportional to the concentration of defects
in the oxide at the oxide-solution interface. Under these conditions, a
steady state is established in which oxide is removed at a rate equal to the
rate of formation and in which a steady-state thickness of film results.
The corrosion rate is determined by the rate of transfer of reagents across
this film. The defects are produced at a rate proportional to the intensity
of radiation and are removed by thermal annealing at a rate proportional
to the concentration of defects. In the derivation of the general equation,
it was assumed that the rate of oxidation of the metal R, at a given pro-
tective film thickness, X, is given by

R=C/X2, (5-15)

where C is a constant.
244 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

In considering the effect of solution composition and flow velocity on
the radiation corrosion of Zircaloy—2, a number of interesting phenomena,
related apparently to the sorption of uranium on or near the protective
film, were observed. Some of these effects are illustrated in Fig. 5-17,
based on data from runs L-2-10, 1L-2-17, 1-4-16, and L-2-15 listed in
Table 5-9. As shown in Fig. 5-17, at a given solution power density the
corrosion rate was greater in 0.04 m UO2SO4 than in 0.17 m UO2S04,
and the degree of difference was less in high-velocity regions (30 to 40 fps)
than in low-velocity regions (10 fps). The low-velocity (1-fps) annulus
specimens in the 0.04 m solutions were attacked at markedly higher rates
than the channel specimens. The addition of 0.04 m excess H2SOy4 to the
0.17 m UO2S04 solutions in autoclave experiments had a beneficial effect
at a given solution power density. The addition of a greater amount of
excess H2oSO4 (0.4 m) to the 0.17 m UO2804 in loop experiment L-2-14
also produced a beneficial effect.

It appears likely that most of these solution and velocity effects are
associated with the sorption of uranium on the surface of the test speci-
mens. As mentioned previously, uranium is usually found in the scale
from specimens exposed in autoclave experiments. The amount of this
uranium is less for specimens exposed to solution with 0.04 m excess H2504
than for those exposed to solutions with no acid. One to two per cent
uranium by weight is usually found for specimens exposed under the
former conditions, and 5 to 69, under the latter condition. It has been
found that the autoclave results in the different solutions can be inter-
correlated as well as correlated with the loop results if a corrected power
density, P., rather than the power density in solution is employed [71].
In this correlation, which is empirical, the corrected power density is
assumed to be given by the expression

P,=P,+ K'NU,, (5-16)

where U, is the percentage uranium in the scale, N is the thermal neutron
flux, P, is the power density in solution, and K’ is a constant. Since P,
is related to N through the uranium concentration in solution, Us, the
equation may be written:

(5-17)

 

Pc=Ps<1—|—KUf>-

U,

The curve expressed by Eq. (5-9) was assumed for the relationship between
power density in solution and corrosion rate in the absence of scale at
280°C. The constant K in Eq. (5-17) was evaluated from the best fit of
the 280°C autoclave data to the curve. The same value of the constant
was found to apply in fitting the 250°C-autoclave data to Eq. (5-10),
5-6] CORROSION BEHAVIOR OF TITANIUM AND ALLOYS 245
which represents the line drawn through the results of the 250°C-loop
experiments. On the basis of these correlations—that is, when the corrected
power density is employed—it appears that there is no appreciable dif-
ference between the corrosion rate in autoclave solutions containing
0.04 m H2SO4 and in those free of excess acid. There is also no difference
between 0.17 m and 0.04 m UO2S04 solutions in autoclave experiments.

The fact that the autoclave data can be correlated with the loop data
represented by Eqgs. (5-9) and (5-10) is of questionable significance, since
it is not known whether the rates in the 0.17 m UO2S04 loop experiments
were Influenced by sorbed surface uranium. As mentioned previously, the
loop specimens are usually free of heavy scale of the type found in auto-
claves. However, uranium was found on specimens from the L-2-17 ex-
periment, which was with a solution 0.04 m UQO2SO4, and it appears
likely that uranium sorption on a given Zircaloy—2 specimen was respon-
sible for an appreciable fraction of the total corrosion attack on the given
specimen both in this experiment and in experiment L-2-10 [72]. Since a
greater amount of uranium was found on the surface of the low-velocity
annulus specimen than on the high-velocity core channel specimen, it is
also likely that differences in the amount of uranium which is sorbed at
different velocities resulted in, the apparent beneficial effect of increasing
solution velocity [72] in these 0.04 m UO2SO4 experiments.

5—6. CORROSION BEHAVIOR OF TITANIUM AND TITANIUM ALLOYS IN
URANYL SULFATE SOLUTIONS*

5-6.1 Introduction. The corrosion behavior of commercially pure tita-
nium and of titanium alloys has been tested at elevated temperatures in
uranyl-sulfate solutions under dynamic and static conditions in the ab-
sence and 1n the presence of radiation. The equipment and testing pro-
cedures employed were identical with those employed in tests of Zircaloy—2
and type—347 stainless steel described previously.

Of the various in-pile loop experiments listed in Table 5-9, all but those
designated DD, GG, and FF were concerned in part with titanium cor-
rosion. In addition to the in-pile tests, one scouting-type experiment was
made of the effect of fast electrons from a van de Graaff accelerator on the
corrosion of titanium-75A in a small thermal siphon loop constructed of
titanium.

*By J. C. Griess and G. H. Jenks.
246 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

5-6.2 Corrosion of titanium and titanium alloys in uranyl sulfate solu-
tions in the absence of radiation. The commercially pure titanium and
the titanium alloys developed a tightly adhering film under exposure which
exhibited interference colors ranging from tan to deep blue or black, de-
pending upon the alloy and upon conditions and length of exposure. This
film could not be removed without damaging the metal. Titanium speci-
mens exposed in uranyl sulfate solutions circulating in stainless steel loops
collected some of the stainless steel corrosion products (iron and chromium
oxides) in an outer layer of scale which could be removed electrochemically.

The average corrosion rates of commercially pure titanium alloys
(45A, RC-55, 75A, 100A, 150A) in a solution 0.04m UO2504,
0.02 m H2S04, and 0.005 m CuSO4 at 200, 250, and 300°C in the absence
of radiation were less than 0.1 mpy. Under similar conditions, high-strength
titanium alloys (containing 5%, Al, 2%, Sn; 3%, Al, 5%, Cr; 6% Al, 49, V;
49, Al 49, Mn) exhibited corrosion rates up to 0.4 mpy; one alloy
(Ti+ 89, Mn) showed only 0.05 mpy at 300°C. These rate values were
calculated from the decrease in weight of specimens as measured in long-
term tests following an initial exposure period of several days. The speci-
men in each case was descaled electrochemically prior to the final weighing
but, as described above, some film remained on the specimen. As shown
by the results, the commercially pure titanium is very resistant to attack.
The high-strength titanium alloys are slightly less resistant, and corrosion
rates increased with increasing temperature. Other tests have shown
similar corrosion rates at uranyl sulfate concentrations from 0.02 to 1.3 m.

5-6.3 Corrosion of titanium and titanium alloys in uranyl sulfate solu-
tion under irradiation. Specimens exposed in the core of in-pile loops were
coated with an adherent bronze-colored film which could not be removed
without damage to the metal. In some loop experiments, the bronze film
was overlaid with a dark-brown scale similar to that found on Zircaloy-2
specimens and was thickest for those specimens exposed at the lowest
fission power densities [73] as for Zircaloy—2. This scale was partly re-
moved from the titanium and alloys by a cathodic defilming operation
prior to final weighing. The value for the corrosion rate of a core specimen
was calculated from the loss in weight during exposure as indicated by
final weighing, the exposed area, and the radiation time (see Articles 5-5.3
and 5-5.4). |

The results obtained with core specimens from in-pile loop experiments
shown in Fig. 5-18 [74] indicate the spread of experimental data over the
range of conditions investigated (Table 5-9). Some of the available ex-
perimental values at low power density and low corrosion rates have been
omitted from Fig. 5-18. Most of the data shown were obtained with
Ti-55A. For this material, the results indicate that the corrosion rate
5-6] CORROSION BEHAVIOR OF TITANIUM AND ALLOYS 247

 

 

 

 

6
| I I | | l | | l
SH —
>
Q.
£ 41— o —
2\
2 o
X 3 ]
0 °
8 O o
52— o ° ° So ° o ° o
O o o o o,
= %° ° —
§ L8R5 o
°§3’jgflw°j’§°fi8 ol 14111
0 2 4 6 8 10 12 14 16 18 20

Fission Power Density , kw/liter

F1c. 5-18. Radiation corrosion of titanium, loop results.

in-pile is greater than that expected out-of-pile and that the in-pile rate
increases slightly with increasing fission power density. There is no sig-
nificant effect of exposure temperature within the range investigated. A
few results suggest that high solution velocities had an adverse effect on
the in-pile corrosion, but other results show no such adverse effect. It is
possible that fretting between specimens and holder was responsible for
the additional weight loss of those specimens which indicated the adverse
velocity effect. It should be noted that the amount of scale retained by a
specimen after descaling can and may affect the apparent corrosion rate
of the specimen, and differences in the amount of scale retained by different
specimens may influence the apparent corrosion behavior with respect to
the variables power density, velocity, and temperature. The few results
available for titanium alloys indicate that, with the exception of the Ti—-6%
Al-4% V in experiment L—2-14, these materials corroded at rates which
are about the same or only slightly greater than those for Ti—-55A under
the same conditions. The rates for the Ti—69% Al-49% V in L-2-14 were
about double those for Ti—-55A.

The corrosion rate of Ti—-55A in in-line positions outside the radiation
field was found to be in the range of 0 to 2.5 mpy [74]. These results indi-
cate that, in general, the corrosion of titanium exposed in in-line positions
of a loop is greater than expected in the absence of radiation, which out-of-
pile tests showed to be about 0.03 mpy or less. The rates determined from
specimen weights were generally less than those observed with specimens
exposed in the core, although in some experiments the maximum rate ob-
served for in-line specimens was about the same as that observed for core
specimens. Differences between the amount of scale retained by in-line and
core specimens after descaling may account, in part, for the apparently
lower rates of in-line specimens. |
248 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

The results of in-pile autoclave experiments indicated corrosion rates
under irradiation which are in general agreement with those determined in
in-pile loop experiments. The oxygen pressure measurements in autoclave
experiments indicate that the rate of attack remains constant at a given
set of in-pile conditions [75].

The van de Graaff experiment employed a solution 0.04 m UO2SOy4,
0.02 m in excess H2SO4, and 0.01 m CuSO4 at a temperature of 300°C.
Exposure to fast electrons was for a period of 50 hr and at an intensity such
that the estimated power density due to eleétron absorption was about
60 kw/liter in solution adjacent to the specimen. No evidence was ap-
parent that the corrosion of the Ti—75A was accelerated appreciably during
the irradiation [76].

5-7. AQueous SLURRY CORROSION™

The studies reported in this section indicate that corrosion-erosion prob-
lems of high-temperature (250 to 300°C) aqueous oxide slurry systems may
be satisfactorily controlled. However, under unfavorable conditions very
aggressive attack has been noted.

Aqueous slurry corrosion problems in nuclear reactors were studied in
the early days of the Manhattan Project, as reported by Hiskey [77] and
Kirshenbaum [78]; in Great Britain’s Harwell laboratory [79]; and in
the Netherlands’s KEMA laboratory [79]. Slurry corrosion problems are
now being studied for the Pennsylvania Advanced Reactor by Westing-
house Corporation [80], and at Oak Ridge National Laboratory [81-85].

Most of the work described below has been done at the Oak Ridge Na-
tional Laboratory and is summarized in the reports cited above. Most of
the circulating corrosion tests have been carried out at temperatures of
approximately 250 to 300°C in stainless steel equipment.

5-7.1 Nature of attack. Factors involved in the corrosion-erosion problem.
Slurry attack may be regarded as simultaneous abrasion and high-
temperature water corrosion. Information on the high-temperature cor-
rosion of materials by water in nuclear reactors is discussed in the Cor-
rosion and Wear Handbook [86], by L. Scheib [87], and for other systems
in the Corrosion Handbook [88]. This serves as a basis for the considera-
tion of similar aspects of attack by aqueous slurries.

For the practical utilization of circulating aqueous slurries, a suitable
balance must be made between corrosion-erosion attack of materials of
construction and handling properties of the slurry, including viscosity,
heat transfer, settling, resuspension, and caking characteristics. The ma-
terials of construction, operating conditions, slurry characteristics, and

*By E. L. Compere.
5-7] AQUEOUS SLURRY CORROSION , 249

properties of the thoria or other material used to prepare the slurry all
exert important influences on the corrosion-erosion problem. All these
factors are important in certain circumstances, and their effect has fre-
quently been found to depend strongly on the presence or level of other
factors. Considerations of such interactions should not be neglected;
however, the strongest factors in attack by aqueous slurries appear to be
particle size, abrasiveness, and degradation susceptibility, flow velocity
and pattern, and metal composition. Important secondary factors include
temperature, atmosphere, slurry concentration, and additives. The effect
of radiation, particularly fissioning, is expected to interact strongly with
such other variables as flow and type of gaseous atmosphere.

Materials of interest. The materials of interest in aqueous slurry systems
for nuclear reactors include many of those considered for high-purity
aqueous high-temperature systems in the Corrosion and Wear Hand-
book [86]. Although materials found to be unsatisfactory in the purely
aqueous system are usually not suitable for slurry systems, the order of
excellence and the relative importance of different variables has been found
to be altered in slurry systems. Most of the materials and alloys investi-
gated were the same as tested in aqueous uranyl sulfate solutions described
previously. Results for classes of materials are summarized as follows:

The austenitic stainless steels represent the most useful class of con-
struction material for slurry systems. Of the various types available,
type—-347 has been tested most. Differences between various types do
not appear to be significant.

It appears possible to reach corrosion rates considerably below 1 mpy
with flows of 20 fps. The most severe attack of stainless steel has been
localized attack noted on pump impellers and housings, orifice restrictors,
and test specimens in high-velocity regions.

Ferritic and martensitic stainless steels are attractive for certain oper-
ating parts because of their hardness properties. The corrosion resistance
of these materials is about the same as that of the austenitic stainless
steels and slightly better in a number of cases. Because certain alloys
of this class do not appear to be susceptible to stress-corrosion cracking
as are austenitic stainless steels, they remain potentially a very useful
class of materials. |

Carbon steels or low-alloy steels are of interest because of their low
cost. Results based on corrosion specimens in systems at high oxygen
concentrations, or with hydrogen added, indicate comparatively high
attack rates decreasing with time, and it is possible that acceptable rates
will be achieved. However, pitting problems and potential embrittle-
ment by hydrogen or alkali have not been explored. The effect of chro-
mate from corrosion of stainless steel by slurries with high oxygen con-
centrations in a combined carbon steel-stainless steel system has not
yet been determined.
250 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Niuckel alloys appear to perform somewhat better in hydrogenated
systems but generally are not so corrosion resistant as the austenitic
stainless steels in oxygenated slurries. Inconel and Inconel X appear to
give results approaching austenitic stainless steel in some circumstances.

Stellites have performed best in hydrogenated systems. Attack by
high-temperature oxygenated slurry appears to be due to simple aqueous
corrosion.

Titanium and its alloys in oxygenated slurry have shown approximately
the same corrosion rates as the austenitic stainless steels. In alkaline
systems with hydrogen atmosphere, very severe attack has been noted.

Zircontum alloys, especially Zircaloy—2, have shown extremely good
corrosion resistance relative to other materials. Fairly severe attack in
a high-velocity wake region by a very abrasive thoria slurry has been
noted.

Noble metals have been tested along with other types to assist in
evaluating the effects of pure erosion. To & rough approximation the
ratio of attack rates on the platinum and gold is constant over a wide
range of conditions, as suggested by Hiskey [89].

Tantalum, in one test at 21 fps, gave results comparable to platinum.
Bronze and aluminum were severely attacked by high-temperature
aqueous slurries. Synthetic sapphire and aluminum oxide were severely
attacked at high flow velocities by high-temperature aqueous thoria
slurries, and pins of thoria densified by the addition of 0.59, CaO disin-
tegrated under similar conditions.

Types and mechanisms of attack. The attack by aqueous slurries on
metals under nuclear reactor conditions may be treated as a succession of
stages: (a) collision of slurry particles with the surface, (b) damage to the
surface, and (c) reaction of the underlying metal with the aqueous phase.
Particles may be caused to strike the surface by impingement, by eddy
action, or by shear forces.

Impingement attack, which is noticed on upstream surfaces of objects in
line of flow, in piping elbows, etc., is visualized as resulting from a break-
through of flow lines by slurry particles as the direction of flow is changed
on meeting an object. Attack primarily results from impact of the larger
slurry particles with the surface. An example is shown in Fig. 5-19.

Impingement target efficiency [90-93] is estimated as a function of the
dimensionless separation number VoD2%(p; — p)/18uDs, where V is relative
slurry velocity, D, is dimension of obstructing body, D, is particle diam-
eter, u is fluid viscosity, ps is solid density, and p is slurry density. A
sigmoid curve of target efficiency versus the logarithm of separation num-
ber was obtained for various shapes with 55 to 809 target efficiency at a
separation number of 1, and with a limiting value of separation number,
0.06 for cylinders, below which impingement does not occur. Application
5—-7] AQUEOUS SLURRY CORROSION 251

Flow

 

N Ti-75A

347 55

 

F16. 5-19. Impingement erosion of pin specimens by flowing thoria slurry.

of this concept by Thomas [94] to aqueous thoria slurry at 250°C indi-
cated that flow at 26 fps past a cylinder of 0.1 in. diameter would result
in no impingement by particles below about 3 microns.

Corrosion rate would by this model be proportional to concentration,
velocity, target efficiency, and erosivity [95] (the mass of metal removed
per unit mass of particles striking the metal surface). Erosivity would be
expected to be a function of the flow characteristics, particle energy and
abrasiveness, and other properties of the particle and the metal, although
a general formula for it has not been developed.

The results of the impingement concept should apply to attack on all
surfaces obstructing flow lines of the slurry. This type of attack will be
sensitive to equipment size effects.

Eddy attack is noted on pump impellers and housings, high-velocity
test specimens, pipe walls, and other similar regions. It is characterized by
relatively deep-gouged pits in the direction of flow, quite smooth on the
bottom, or by a deep general polishing. The pits are frequently undercut
to the downstream side so that they feel smooth in the direction of flow
and rasplike in the reverse direction. Attack due to wakes, cavitation, or
flow separation is more localized. It is frequently possible to associate the
localization of the attack with some change in the flow pattern. It has
been observed downstream from protruding weld beads, on the down-
stream side of pin test specimens, on the rim of pump impeller shrouds, on
pipe walls downstream from a flow interruption, between adjoining coupon
test specimens, etc. An illustration of such attack on a pump impeller is
given in Fig. 5-20.
252 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

Detail of Eroded Surface

 

Fia. 5-20. Eddy corrosion of stainless-steel pump impeller by thoria slurry.

Pits resulting from eddy attack may become quite deep, with depths of
100 mils having developed in a stainless steel pump-impeller shroud edge
(not shown) in 1000 hr.

Eddies can be caused by high velocity, high rates of shear, cavitation,
wake effects, and flow separation and may result in breakthrough of the
flow boundary layer [96-97].

It would appear that in some sense the impingement model above might
still apply to eddy erosion, with larger particles being thrown to the outer
circumference of an eddy. The edge velocity of the eddy would be at least
equal to the fluid velocity at the point of eddy origin. The postulation of a
lower limit on eddy size has been noted [98]. Small eddies might become
trapped in surface imperfections, leading to continuing localized attack.
Wake effects are similar to the cavitation erosion studied in water tunnels
by Shalnev [99], in which the initiation of cavitation in the wake of circular
objects is attributed to eddies. Eddy attack will be very sensitive to equip-
ment shape and streamlining effects.

Low shear attack 1s a relatively gentle, low-velocity attack such as might
be observed on pipe walls, coupon specimen surfaces parallel to the direc-
tion of flow, ete. It is presumed to result from a reduction in thickness of
the protecting boundary layer as flow velocity increases. As indicated by
Johnstone and Thring [100], this type of attack would be sensitive to
scale-up effects, becoming less severe at equivalent Reynolds number as
the flow-channel dimensions are increased. It is also likely to be sensitive
to water corrosion effects.

Relatively thin colored films (less than 1 mg/cm?2) characteristic of
5-7] | AQUEOUS SLURRY CORROSION 253

water corrosion are normally observed in aqueous slurry systems. In
many cases, especially when a substantial quantity of electrolyte is present,
a bright polished surface has been observed. With slurries having particle
sizes near their degradation minimum, there appeared to be no decrease in
corrosion rate with time, which implies that the outer portions of the pro-
tective oxide film were being removed by the flowing slurry as fast as it
was formed.

Factors involved in surface damage to oxide film or metal proper by im-
pinging particles are complicated. Hardness, elasticity, strength of the
oxide film or metal, and the nature of the metal-to-oxide bond are im-
portant. Impact energy, its localization by particle sharpness, and its
transfer due to particle hardness and strength, are also important. Rosen-
berg and others [101-103] noted that sharp sand was about four times as
erosive to steel (in an air blast) as larger, more rounded sand. IFor many
slurries in toroid or pump loop tests the most severe attack of metal sur-
faces is noted concurrently with the most rapid degradation in the size of
the slurry particles [104].

For all metals of interest, except the most noble, the destruction of a
protective oxide film exposes reactive metal to the aqueous medium. The
rate of reaction is very high, and the rate of film re-formation will control
this reaction rate and the net amount of metal consumed. Changes in the
chemical environment and pH via gaseous atmosphere, additives, or in-
growing corrosion or fission products will strongly affect the rate of metal
reaction and oxide re-formation.

Methods of test. The relative abrasiveness of slurries may be determined
in a laboratory jet-impingement test device developed by McBride [105]
in which a slurry spray is jetted against a thin metal strip and penetration
time noted.

To obtain accurate slurry corrosion data, toroids, pump loops, and in-
pile autoclaves have been used for high-temperature tests with aqueous
slurries. This equipment is described in Section 5-2.

In toroid tests, information is obtained as to corrosion rate of several
metals, slurry particle degradation, and qualitative slurry handling charac-
teristics by examination of metal pin specimens and analysis of the with-
drawn slurry.

In pump loops, a larger amount of quantitative information is obtained
at several different velocities and flow patterns from an examination of
loop components, as well as specimens. Generalized incremental loop cor-
rosion rates and slurry properties are obtained at suitable intervals from
analysis of withdrawn samples.

By use of rocking in-pile autoclaves, tests at negligible flow velocities
have been made at neutron fluxes approaching 10'3 to evaluate corrosion
rates and rates of production and recombination of radiolytic gas.
254 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TaBLE 5-11

TaE ATTACK OF VARIOUS MATERIALS BY 1600°C-
Firep THORIUM OXIDE SLURRY (PuMp-Loop TESTS)

Loop-piping generalized corrosion rate except pump: 0.9 mpy
at ~ 10 fps. Pump-impeller (7.8 in. dia) wt. loss: 27 g at 90 fps
(50-100 mil rim pits)

 

Pin-specimen corrosion rates, mpy:

Velocity 20 fps 40 fps
Austenitic stainless steels 2 25
Titanium alloys 2 6
Zircaloy—2 0.1 1
Gold 0.2 4
Platinum 1 12

 

 

 

5-7.2 Slurry materials. Uranium dioxide. Uranium-dioxide slurry and
quartz slurries were tested in small-scale loops and spinning cylinder sys-
tems in the early days of the Manhattan Project. It was observed [106-
107] that erosion was proportional to the square or cube of the velocity,
varied strongly with particle size, decreased with time as particle size
degradation took place, and exhibited a constant ratio for different ma-
terials under the same conditions. Uranium dioxide has a Moh scale hard-
ness of 6.1 [108] and to a certain extent would be expected to resemble
thoria under a hydrogen atmosphere, as described later.

Uranium trioxide. The attack of stainless steel pump-loop components
by uranium trioxide slurries at 250°C was essentially nil in tests extended
for thousands of hours, even on pump impellers with a tip velocity of
120 fps [109]. The lack of attack by circulating uranium trioxide slurries
under conditions in which impingement certainly occurred is attributed
to the relative softness of the uranium trioxide particles. |

Thortum oxide. The major portion of studies on corrosion-erosion
characteristics of high-temperature aqueous slurries has been carried out
using thorium oxide [110-111]. Most preparations have used the calcina-
tion product of thorium oxalate precipitated under various conditions,
with calcination temperatures from 450 to 1600°C having been used.
Thoria prepared by other procedures, such as formate precipitation, has
also been examined. Thoria calcined at temperatures in the vicinity of
1600°C has low surface area, crystallite sizes approaching particle size, less
tendency to degrade, and exhibits a greater tendency to produce abrasive
sintered particles; consequently, certain of such thoria products have been
5-7] AQUEOUS SLURRY CORROSION 255

TABLE 5-12

RepuctioN IN THORIUM OXIDE SLURRY
CoRROSION BY PARTICLE Size CONTROL
(Pump-loop tests on same batch)

 

 

 

 

 

 

Unclassified Classified

Temperature, °C 200 280
Duration, hr 5 309
Concentration, g Th/kg H20 443 323
Atmosphere Oq Oq
pH (slurry, 25°C) 7.5 5.5
Thoria calcination temperature, ° 1600 1600
Particle size |

Average, u 1.3 1.7

Maximum 289, > 10 u 29, > 5 u
Average crystallite size (x-ray), u > 0.25 > 0.25
Specific surface (N2 adsorption), m?/g 1.2 1.7
Loop-piping generalized corrosion 63 1.5

rate, mpy
Pump-impeller (7.8 in. dia) wt. loss, g 11 1.5
Pin-spectmen corrosion rates, mpy
Velocity 26 fps 40 fps 22 fps 44 fps
Austenitic stainless steels 27-57 150-320 0.8 8
Ferritic stainless steel 34-110 120-180 1 4
Titanium alloys 63 240 0.5 4
Zirconium alloys 8-97 310-700 0 1
Gold 120 630 0 2
Platinum 96 650 0.5 7

 

 

 

classified, using sedimentation procedures, to remove unduly large particles.
Thoria calcined at 650 to 900°C because of its properties was considerably
degraded by continued pumping and was reported [112] to show greatly
reduced corrosion rates as this occurred.

The hardness of thoria minerals has been reported as 6.5 on the Moh
scale. The hardness of thoria densified with 0.59, CaO is 6.8 [113], and
laboratory preparations of thoria have been observed with hardness greater
than 7. The hardness of many mineral oxides, of a composition similar to
the protective film on metals, lies between 6 and 6.5. However, sapphire
attacked by high-temperature aqueous thoria slurry has a hardness of 9.

To evaluate materials under conditions expected for a reactor, pin speci-
256

TABLE 5-13

INTEGRITY OF METALS IN HOMOGENEOUS REACTORS

[cHAP. 5

CoMPARATIVE PIN-CORROSION RATES AND PARTICLE DEGRADATION
AT 300° C; Same TrHoRIA BarcH (THORIUM OXALATE CALCINED
AT 800° C), OxYGEN ATMOSPHERE, PumpP-Loor TESTS

 

 

 

 

 

 

 

 

 

Thoria Unpumped Unpumped Prev. pumped

Hours 49 301 266

g Th/kg H20 (avg.) 543 348 346

fps 11 21 41 10 22 41 13 22 43

Corrosion rate, mpy

Gold 0.4 04 4 0.1 0.1 0.6 | 0.03 0.08 0.9

Zircaloy—2 0.1 1 6 0.1 0.1 2 0 0 1

Titanium 1 4 9 0.2 2 6 0.1 1 5

Stainless steel 3 7 21 1 3 8 0.5 0.5 3

Loop 4.5 1.5 0.8
Particle-size distribution, w/o

Size, 1 >3 13 <1 | >3 18 <1 | >3 13 <1

0 hr 48 27 25 59 26 15 3 32 65

5 hr 8 25 67 —_ = — — = —

30 hr 5 21 74 7 25 69 7 28 65

Final —_ = — 3 32 65 2 21 77

 

 

 

 

 

 

mens of various materials were exposed for 1000 hr to a slurry of 1500 g
Th/kg H20 at 280°C in a pump loop pressurized with steam plus
200 psi O2. The 1600°C-calcined thoria had an average particle size of
1.7 microns, with 39, of the particles greater than 3 microns, and the
slurry had a pH of 5.8 measured at 25°C. The results of the test are sum-
marized in Table 5-11. |

5-7.3 Effect of slurry characteristics. Particle size. Two effects relating
to particle size have been distinguished. These are the effects of largeness
of particles and of particle degradation. Large particles cause erosion-
corrosion. This is in agreement with the impingement model of slurry
attack. Tables 5-12 and 5-13 show comparisons of pump-loop corrosion
tests in which, in the respective tables, the larger-particle-size slurry shows
more aggressive attack. |
5-7] AQUEOUS SLURRY CORROSION 257

The removal of large sintered particles resulted in a substantial reduc-
tion in corrosion rate for all materials in the tests shown in Table 5-12.
This effect for thoria produced by high-temperature calcination was suf-
ficient to justify including similar classification in the regular production
procedure. Particles of large crystallite size (>0.25 micron) from which
the sintered particles had been removed by classification were shown to
produce comparatively low corrosion rates. Particles composed of large
crystallites are not much degraded by continued circulation, and conse-
quently corrosion rates do not diminish much with time.

Low-crystallite-size materials (calcined at relatively low temperatures,
e.g., 800°C) become degraded as circulation proceeds, and aggressiveness
is diminished. Thus, a shorter test showed a higher rate of corrosion, and
a test using previously pumped material showed a lower rate of attack.
This, and the change of particle size with time, is illustrated in Table 5-13.

Effect of calcination temperature. The major effect of increased calcina-
tion temperature is to cause growth of larger crystallites, and under ad-
verse circumstances, sintered particles. Sintered particles, of course, would
cause increased attack. However, unless sintering occurred, particles have
not appeared to increase in size on calcination. At constant particle size,
calcination temperature has not exhibited a direct effect on corrosion
rate. Indirectly, lower calcination temperatures result in lower crystallite
sizes and particles may be more readily degraded to smaller, less erosive
sizes. Table 5-14 illustrates these points [114]. Except for the unac-
countably increased attack by material calcined at 800°C, there appears
to be no general effect of calcination temperature on attack rates in these
short-term tests. It is also noted that the material calcined at higher tem-
peratures is less degraded. In longer tests this would be expected to result
in maintaining the original corrosion rate, rather than resulting in a de-
crease in rate when the particle size becomes smaller.

As crystallite size becomes larger, surface area is reduced. This reduces
the adsorptive capacity of the thoria and thus influences the action of addi-
tives on corrosion. Handling properties are also changed.

Effect of concentration. In general, corrosion-erosion by slurries has been
observed to increase with concentration [115-116], and roughly is indi-
cated to be directly proportional to concentration, in concentration ranges
of reactor interest. As concentration is increased, the effect of rheological
properties on flow characteristics becomes more pronounced, and the effect
on corrosion would become altered.

Effect of atmosphere. It is possible to distinguish between the effects of
different atmospheres resulting from high oxygen concentration, low oxy-
gen concentration, dissolved hydrogen, or other dissolved gases.

High oxygen concentration [117-119] in pump loop and toroid tests has
appeared to result in less aggressive attack on type—347 stainless steel than
258 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 5-14

ErrFeEcT OF THORIA CALCINATION TEMPERATURE ON
" SLURRY CORROSION AND PARTICLE DEGRADATION AT
VARIOUS CIRCULATION VELOCITIES

 

 

 

 

 

 

 

 

 

 

 

ThO» from thorium oxalate. 250°C, toroid tests.
Avg. dia., 2.6 u after calcination. 100 psi oxygen.
1000 g Th/kg water. Test duration: 100-300 hr.
Maximum Average Average attack rate, mpy
Circulation . diameter !
: calcination
velocity, - after
fos temperature, circulation
p °C . | 34788 | T 7r-2
5 650 1.2 0.4 0.4 —
800 1.5 0.8 1.0 +
1200 2.2 0.4 0.6 +
15 650 0.8 5 2 —
800 1.0 4 6 4
1200 1.3 3 4 0.2
1600 2.1 4 2 1
26 650 0.7 7 4 3
800 0.6 14 16 4
1000 1.1 7 3 3
1200 1.0 9 6 3
1400 1.0 8 6 2
1600 1.2 7 6 3
Average attack rates at given velocity:
5 0.6 0.8 +
15 4 3 2
26 8 8 3

 

 

 

 

 

 

 

tests which used no added oxygen. Oxygen concentrations of 250 ppm
(160 cc/kg H20) appear sufficient to protect stainless steel. Toroid tests
[120] in which the oxygen was consumed resulted in more aggressive
attack of this metal. Corrosion products were black rather than tan-brown
in color and were found to contain FesO4. In systems having hydrogen
5T7] AQUEOUS SLURRY CORROSION 259

atmosphere, stainless steel was attacked somewhat (perhaps threefold)
more aggressively than in the presence of sufficient oxygen.

These data are presumably valid also for ferritic and martensitic stain-
less steels. o

The corrosion resistance of carbon steel is improved under hydrogen

atmosphere. Rates in some cases have approached those observed for
stainless steels in the same experiment. Less localized attack is noted
under hydrogen atmosphere.
- Titanium appears to be equally corrosion resistant in oxygenated and
hydrogenated systems, provided the slurry is not strongly alkaline. At
150°C titanium is more readily abraded by slurry with hydrogen atmos-
phere. In alkaline systems with hydrogen atmosphere, a very aggressive
attack has been noted, as would be anticipated from the interpretation of
Schmets and Pourbaix [121].

Zirconium alloys appear to be little affected by atmosphere. They are
possibly more easily abraded under hydrogen atmosphere.

Nickel alloys and also Stellites appear, in agreement with Douglas [122],
to be substantially more corrosion resistant under hydrogen atmosphere,
and several nickel alloys have appeared to be more resistant than stainless
steel when exposed with it in experiments under hydrogen atmosphere.
~ The nature of the atmosphere also affects the oxidation state and solu-
bility of certain corrgsion products and through these may affect the
properties of the slurry. Under an oxygen atmosphere iron exists as FezO3,
and chromium as soluble, acidic CrOs. Under hydrogen atmosphere, iron
1s In the form of Fe3O4, and chromium is insoluble Cr20s.

Effect of additives. Additives include the unavoidable corrosion products
-and fission products; the necessary, or difficultly avoidable, recombination
catalysts; uranium inclusions in the particle and certain impurities; and
optional materials added to modify the properties of the slurry.

Corrosion products have not been observed to affect the course of cor-
rosion in oxygenated systems, although they have built up to levels of a
few grams per liter in certain lengthy tests. Soluble chromic acid may
affect the corrosion rate of such material as carbon steel. No experimental
results are available on the effect of fission products.

Molybdenum oxide (a potential recombination catalyst) has shown a
mild corrosion-inhibiting action. Uranium inclusions in the thoria have not
been found to increase corrosion rates. Certain impurities, i.e., carbonate
and sulfate carried through the production process, have not been found to
affect corrosion rates except when added in large quantity, as described
below. Chloride, which can cause stress-corrosion cracking of stainless
steel, is undesirable in any quantity, as is fluoride.

Certain materials have been found to impart desirable handling proper-
ties to slurries at lower temperatures and have been of interest to test at
260 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

higher temperatures (e.g., 280°C). These have included sulfates (thorium,
sodium, calcium, hydrogen), phosphates (sodium monobasic, dibasic,
tribasic, pyro-), silicates (sodium meta- and more acid compositions), and
fluorides (thorium). In general it has been found that additives imparting
acidity [e.g., H2SO4, Th(SO4)2] tend to attack ferrous-based materials,
especially under low oxygen or reducing atmosphere. There is less effect
of moderate concentrations with oxygen atmospheres. Alkaline materials
tend to increase attack of titanium under reducing atmosphere and to
affect the performance of other metals. Complexing or reacting additives
(e.g., alkaline phosphate with titanium, fluorides with zirconium) tend to
destroy protective films and thereby increase attack rates.

All soluble ionic additives appear initially to be adsorbed on the thoria
surface. Their effect on corrosion is much greater after surface capacity is
exceeded. Surface capacity varies with different thoria preparations, pri-
marily with surface area.

5-7.4 Effect of operation conditions. Flow velocity is a very important
factor in corrosion by slurries. Corrosion rate appears to be approximately
proportional to the square of velocity at moderate velocities and possibly
to a somewhat higher power at higher velocities. The effect of velocity is
shown in Table 5-14 and in Table 5-15.

In addition to increasing regularly with velocity, greater attack in
entrance regions and at the highest velocities was noted. The more severe
attack on cylindrical pins is attributed to greater turbulence effects for
this shape. |

Velocity effects are also shown on the pump impeller in Fig. 5-20.
Velocity increased radially from the hub, and attack was most severe at
the rim. However, the vanes and inner shroud surface were polished,
which indicates the importance of boundary-layer considerations in the
study of such effects.

Cavitation also may be an important factor in the high-velocity attack.
Attack on pump impellers has been found to be more severe in experiments
in which gas bubbles were believed to have been entrained in the slurry.
Erosion patterns similar to those observed by Shalnev [123] have been
found on pin-specimen-holder channel walls, with most severe points of
attack a few pin diameters downstream from a pin. Although this agrees
with Shalnev’s observations, it is not clear whether the effect is due to
cavitation or simply high eddy density at this point.

Shape effects. The importance of shape effects in corrosion by flowing
slurries may be very great, and the most severe attack has been observed
to be a result of eddy erosion-corrosion associated with shape effects. The
attack in some cases has appeared to be self-accelerating as pits develop.
The importance of shape effects will depend strongly on velocity and also
on particle characteristics and chemical environment.
5-7] AQUEOUS SLURRY CORROSION 261
TABLE 5-15
ErrEcT oF VELOCITY ON ATTACK OF DIFFERENT SHAPES

182-hr circulation at 300°C, Oz, 617 g Th/kg H20
Thoria calcined 800°C

 

Stainless steel attack rate, mpy

 

Velocity, Flat coupon o ,
fps Cylindrical pins

across straight channel

 

Tapered channel Straight channel

 

 

1 (entrance)
10 0.8
13 1
16 - 2
24 3
31 4
36 6
41 8
50 10 21

64 50, pitted

 

 

 

 

 

 

Design considerations are quite important. Streamlining of all sensitive
parts and regions should be practiced to the maximum extent possible, in
order to avoid impingement or wake effects. Surface irregularities such as
rough finish, crevices, or protrusions should be avoided. The interior de-
sign of flow channels should be carefully considered in order to avoid im-
pingement, flow separation, and wakes insofar as possible. Conditions con-
ducive to cavitation should be avoided, since vigorous attack would be
anticipated under such conditions.

Temperature. In general, temperature effects are associated with the
reaction between water and the metal. Those materials most sensitive to
water corrosion (e.g., Stellites in oxygenated aqueous systems) exhibit
stronger temperature effects.

For many materials, corrosion rate has appeared to double for every
30 to 60°C increase in temperature.

Time. Changes in attack rate with time generally are not observed after
the slurry has reached its equilibrium condition. Until this condition is
reached, the slurry particle size may be larger than its degradation mini-
mum. Consequently, more aggressive attack is usually observed in the
early exposure period.
262 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS  [CHAP.)

Localized attack which created eddy erosion pits could become more
rapid as these pits developed.

Carbon steel has appeared in oxygenated systems to form its protective
film rather slowly. Diminishing attack rates were observed over a period
of at least 700 hr.

5-7.5 Radiation. Awuloclave tests. Data are available [124] on the at-
tack of Zircaloy—2 under radiation by thoria slurries in gently agitated
autoclaves. No radiation data are available under flow conditions. Com-
parison of unirradiated control experiments with tests conducted in a re-
actor radiation field having a slow-neutron flux of 0.5 to 1.0 X 10!% have
indicated that both experiments had an initial rate of approximately 1 mpy
with thoria slurries in D20 at 280°C (oxygenated). However, no reduction
in rate with time was noted under radiation, while the rate decreased sig-
nificantly with time in a longer unirradiated control. Similar results were
obtained on pure D20 under radiation. Inclusion of enriched uranium in
a thoria preparation permitted an experiment in which a fission power
density of 0.5 to 1 watt/ml was achieved. Only a slight increase in rate,
several tenths of a mill per year, was noted as a result under these con-
ditions.

If radiation affects the protective oxide film, it is likely that more severe
effects will be observed in the presence of radiation with slurry under
sufficiently vigorous flow conditions.

5-8. HoMOGENEOUS REACTOR METALLURGY¥

5-8.1 Introduction. Although many other materials are used for special
applications in homogeneous reactors, in the following section only zirco-
nium, titanium, austenitic stainless steels, and pressure-vessel steel are
considered.

Zircaloy—2 (1.5% Sn, 0.1% Fe, 0.19% Cr, 0.005% Ni) used in the present
reactors is the only commercially available zirconium alloy. It is a single-
phase alloy with moderately good mechanical properties, is not heat
treatable, and is stable under operating conditions.

Of the possible types of titanium and its alloys commercially available
in the United States, only unalloyed titanium and one alpha alloy
A-110AT (5.0% Al, 2.59% Sn) are satisfactory for homogeneous reactor
applications. Beta alloys are not yet available and alpha-beta alloys, often
structurally unstable under high stress and high temperature, cannot be
welded without subsequent heat treatment. Although A-110AT has high
strength and is stable and weldable, it is unavailable as pipe or tubing and

*By G. M. Adamson.
5-8] HOMOGENEOUS REACTOR METALLURGY 263

is difficult to fabricate. For such applications, so called ‘“‘unalloyed”
titanium (A—40 and A-55) is used. These high-purity grades are readily
weldable in contrast to the stronger grades, are stable, and are available
in all forms. Their disadvantage is their low strength at elevated tempera-
tures which, except as linings, makes their use for large high-pressure,
high-temperature equipment doubtful. Extensive use of titanium for
homogeneous reactor use will probably depend upon the eventual develop-
ment of a fabricable high-strength alpha alloy.

The pressure vessel for HRE-2 is constructed of A-212 grade B steel
clad with 347 stainless steel. The serious possibility of radiation damage is
the most important metallurgical limitation of this material. This problem
1s discussed in Article 5-8.8.

5-8.2 Fabrication and morphology of Zircaloy—2. When work was started
on the HRE-2 core tank, very little was known about the physical metal-
lurgy and fabrication of zirconium alloys. A procedure was available for the
fabrication of zirconium fuel elements, but the variables of fabrication had
not been investigated in detail. Since this core tank was a vital portion of
the reactor system and was needed early in order to proceed with the
pressure-vessel fabrication, it was built to a time schedule that permitted
only a limited amount of development work. As a result, much of the
present-day understanding of the physical metallurgy of Zircaloy—-2 was
not obtained in time to be used. During the development work for the tank,
many variations were found in the available Zircaloy—2 plate. These
included variations in mechanical properties, in bend radii, and in the
amount of laminations and stringers.

Fabrication practice for the core-tank plate was based upon procedures
developed by Westinghouse Atomic Power Division [125], but modified
by increased cross-rolling to reduce the preferred orientation. The fabrica-
tion procedure for the core tank* is discussed in detail in Refs. 126 and 127.

This procedure resulted in the fine-grained, equiaxed, but oriented struc-
ture shown by the anodized sample [128] in Fig. 5-21. Two disadvantages
of this material were the stringers found in the structure and the preferred
orientation as shown by uneven extinction in rotation in polarized light.

Since it was suspected that the stringers were responsible for some of
the variation in mechanical properties, they were examined in considerable
detail in the morphological study. The two types of stringers prevalent in
Zircaloy—2 are an intermetallic stringer, usually found in the alpha grain

*(1) Double arc melting, (2) forging to a billet, starting at 850°C, (3) cross rolling
509, to plate, starting at 850°C, (4) straight rolling to finished plate, starting at
770°C, (5) press in one or more steps to the desired shape, at 400-650°C, (6) heat
to 650°C, cool slowly in die, (7) weld subsections, and (8) repeat (6).
264 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

 

Plate 1 (Core Tank) Plate 3 (Present Practice)

Fic. 5-21. Microstructures of as-received Zircaloy-2 plates fabricated by dif-
ferent schedules. Chemically polished, anodized, polarized light, 500 X.

boundaries, as shown in Fig. 5-21, and an elongated void within the grains,
caused by trapped gases. The latter type was found rarely in HRE-2 core-
tank metal, but a small one may be noted in the photomicrograph of the
developed material. This type may be largely eliminated by vacuum
melting of the original ingots. It has been determined that the intermetallic
stringers were formed during fabrication. While the upper fabrication
temperature of 850°C was thought to have been in the all-alpha region,
this is now known to be in error. The correct temperature for the alpha
to alpha-plus-beta transition is 810°C and that for the beta to beta-plus-
alpha transition is 970°C [129]. These temperatures will vary slightly with
ingot composition. Holding at a temperature just above the lower boundary
of the two-phase region resulted in the presence of small amounts of beta in
the grain corners. By quenching samples held at 840°C, it was shown that
approximately 159, of the material had been beta phase at that tem-
perature and that the beta phase had been present in the grain corners.
Iron, nickel, and chromium are beta stabilizers and will partition to the
beta phase, which will be strung out during rolling. On cooling, the beta
phase decomposes, precipitating the alpha zirconium on the neighboring
grains and leaving the intermetallics in the grain boundaries. These inter-
metallic stringers will dissolve on reheating to 1000°C while the stringers
formed by the trapped gases will not.

While mechanical property tests of the core-tank material (see Article
5-8.3) indicated uniform properties in the longitudinal and transverse
directions, it was noted that the fractures were not round but were oval,
5-8] HOMOGENEOUS REACTOR METALLURGY 265

   
  
  
 
  
  

— As-Received
—w-. After Refabrication

\2.0

 

0001 1071 1070
Inverse Pole Figure

Rolling Direction 1130

. As-Received

 

 

 

0.40 | 85
0001 | 1011 1070
Inverse Pole Figure
Transverse Direction
1120
. As-Received
———_ After Refabrication
Lo
4145 A1) 0.33
0001 1014 1013 1010 1070
3.12 1.

Inverse Pole Figure
Normal Direction

Fic. 5-22. Inverse pole figures from HRT core tank Zircaloy-2 before and
after refabrication. Solid lines as-received, dotted lines after refabrication.

indicating anisotropic properties. This was confirmed by a polarized-light
examination. The actual orientation of the material was determined using
a special x-ray diffraction technique [130], with the data being reported in
the form of inverse pole figures [131] shown in Fig. 5-22. Since variations
were found in the intensity of the pole concentrations, it is evident that
the core-tank plate had preferred orientation in all three directions.
Quantitative calculations indicated that the plate had adequate ductility
266 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

and nearly isotropic properties in the rolling plane but very little ductility
in the normal direction to the plate, since in this direction few deformation
systems were available for slip or twinning.

To improve the quality of Zircaloy—2 and to predict what structures
would be encountered in weldments, a study of the effect of heat treatment
and fabrication practice was initiated. Representative photomicrographs
of structures resulting from these studies are presented in Fig. 5-23. The
fine “basket-weave’ alpha structure observed in beta-quenched samples
appeared to offer the most likely starting point. Recrystallization of this
structure by annealing at temperatures between 700 and 800°C resulted
in the formation of very large alpha grains, without stringers or inter-
metallic precipitates. Cold-working of the ‘‘basket-weave’ structure, 15%
or less before annealing, also resulted in the production of large grains;
however, if the amount of cold-working exceeded 20%, improvement in the
form of fine grains (ASTM 6 to 8) with a more nearly random structure
resulted.

Scrap from the core-tank plate was heat treated by beta quenching, cold-
worked 20%, and alpha annealed. The diffraction results after this treat-
ment are also shown on the diagrams in Fig. 5-22. The intensity of the
peaks has been decreased by a factor of two and the peaks shifted 20 deg
from the original position. These changes indicate an increase in duc-
tility in the normal direction.

The results of the morphological study were used to outline an improved
fabrication schedule for Zircaloy—2 [132]. This schedule* has been used
successfully by commercial fabricators while more complete fabrication
studies are being made. The material obtained from this procedure has a
microstructure such as that shown in Fig. 5-21. It is essentially free of
stringers, contains little or no intermetallic precipitate, and has small
equiaxed grains with essentially a random orientation as seen under polar-
ized light.

5-8.3 Mechanical properties of zirconium and titanium. The use of
titanium and zirconium for pressure-containing equipment is complicated
by the fact that both metals are hexagonal, rather than cubic, making
preferred orientation a severe problem. The larger variations of mechanical
properties with crystal orientation make it difficult to achieve the uni-
formity in all directions which is desirable in a pressure vessel. In addition,
the determination of the mechanical properties of hexagonal metals, using
standard tests developed for cubic metals, is questionable because the suita-

*(1) Vacuum arc melt, (2) forge at 970-1050°C, (3) roll at 500-785°C, (4) heat
to 1000°C and water quench or fast air cool, (5) roll 259, at 480-540°C, and (6) an-
neal at 760-790°C and water quench.
5-8] HOMOGENEOUS REACTOR METALLURGY 267

 

(b) Beta Quenched 4 Annealed 800°C 15 min

      

 

+ B>

   

 

p— o

(c) Beta Quenched orkedO | (d) Bétu d;:enched 4+ Cold Worked 20%,
Annealed 800°C 15 min Annealed 800°C 15 min

Fia. 5-23. Microstructure of heat-treated Zircaloy-2. Chemically polished, ano-
dized, polarized light, 100 X.

bility of such tests and the correct interpretation of results have not been
demonstrated.

Zircontum. The anisotropy of Zircaloy-2 is not defined by the usual
tensile test data, given in Table 5-16. Tensile data from both longitudinal
and transverse subsize specimens cut from plates fabricated by various
procedures are given in this table. With the exception of the commercial
plate, the difference in tensile strengths both at room temperature and
300°C are within the experimental limits. Anisotropy in the material is
illustrated, however, by the three columns which tabulate the percentage
reduction in length of both the major and minor axes of the elliptical
[cHAP. 5

IN HOMOGENEOUS REACTORS

INTEGRITY OF METALS

268

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

01 ¢ G¢G ¢ G¢ S vy JANAY 00¥8¥y 00c¥L SUBIJ, G
9°C AR 4! G L8 0'9% G GE 00%99 008¢2 ‘8uor] [eyuouLIadxy
€1 9°'1¢ ¥'8¢C 0'%¥ 1°6¢ 00%6¢ 00cG. SuBd], I
1¢ L9l ¢gg ¥ 0'1¢ 00€€9 00cEL ‘guor] [eyuowILIodxy
L Gl ¢'¢ 90V ¢y 8 G¢ 00E8¢ 00069 SuBlj,
¥ 01 vy G GF L L¥ 1°1¢8 008¥S 006¥9 ‘3uory [BIOIDUIWIO))
8¢ 001 8 LG 129 ¢ ¥¢ 0006¢< 00€€9 "SUBLT,
LG 001 9°9¢ 9°09 L 0€ 0083¢S 00%19 ‘guory JueJ, 910
2UNIDLIAULI] UWO0Y
UOT1109.I UoRoaIIp
[eurIOU /*3UO0] HoSAIp "SuBI) BOIY 1sd 1sd

« [BULION] . A oo (o UOI}IIIP 99.IN0S
o1)BlI I0 *Juor] . g yj3us)s U)3uda)s d

o1do130] UOHEIIOM | prx | epewumyg | TS #¥0d

 

UOI}oNPal 938)JUIID J

 

 

 

 

 

 

Z—XOTVOUI7 J0 SAILITJONJ HTISNA T,

91-¢ @IAV],

 
HOMOGENEOUS REACTOR METALLURGY 269

5-8]

*a1npaooxd sures Suisn £10)8I0qR] UI PIYBILIQR]I JNq ¢ W) S8 )8[d sures :g 9)8[d [ejusurradxyy

"poyPw se3-j1aur 4daoxa ‘gz g—¢ NIy ul ainpadoid padopaasp £q A[erorowInuod pajeoriqe; ;T 9)e[d [ejuswiradxy
*Po[[o1 1YBIeI}s A[Uo Ing 9A0QB 0} IB[IWIS 9)8[d [BI2I8UIMI0) :9)e[d [BIOISUWIWO))

*7'8—G O[PIIIY 998 {UOI)BILIQB] JUB)-0I00 Z—HYH WOIJ [eLIe)ewr deIos :JuB} 910)),

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G'1 ¥ 68 ¢ 8y L 89 I'¢€ 00192 0009€ SUBL, 6
81 g ag €89 G 9. 6 ¢€ 008.L3 008¢€ '3uo] [ByusmILIodx
VI 9°1¢ 8 &Y ¢'19 ¢'I€ 00062 009.€ SUBL], I
¢ v v¢ G ¢S %9 ¥ 0€ 00¥1€ 0098¢ '3u0] [ByuswILIOdX
6% €¢I ¢ 99 ¥ 89 & 66 00¢1¢ 00€6¢ SuBL],
67V 84 ¥ 8¢ G&v Gag 0099¢ 009¢¢€ '3u0] [BIOIOUIUION)
0¥ 8 LI 1L G 9L 6 L¢C 0096¢ 00¢cE SUBL], .
¢'¢c G LG 6 89 9 LL 8 '8¢ 002cé 00¥8¢ *3uo] HuwJ, 9107)
D008
UOI}93IIP UOTORLP
[euLIou/ 3uo| o ‘suBI) BOIY 1sd 1sd
[eULION 9, : : UOT)93IP 90IN0S
‘oryBI I0 *3uory . g ‘Ujsuer)s ‘UI8uar)s - ood o
o1d0130ST UO1)BIUOH DIX oyBwIm uowradg «278d
UOT)ONPaI 938U J
(AIANILNOD)
91-G @IAV],

 
270 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

fracture and the ratio between them. In isotropic material the percentage
reduction would be the same in both directions and the ratio would there-
fore be one. As shown, some reduction in anisotropy was obtained by the
use of the developed procedure described previously.

Anisotropy of these materials may also be shown by V-notch Charpy
impact tests. Since size effects are known to be critical in this type test,
standard size specimens were used, thereby eliminating the thin core tank
plate. Charpy V impact energy curves obtained with specimens and notches
cut from various orientations from commercial plate and from a plate
fabricated by the developed procedure are shown in Fig. 5-24 [133].
Differences in both the energy values and transition temperatures are
apparent for the various orientations. These differences are reduced by the
developed fabrication procedure.

With Zircaloy-2 the fracture appearance of the impact samples was not
obviously indicative of the properties, as with steels. Shear lips, which are
normally a mark of a ductile fracture, were found on samples from some
orientations regardless of breaking temperature, but on specimens from
other orientations from the same plate, shear lips were not found at any
temperature.

The effects of notches and cracks in zirconium were studied using the
drop-weight test as developed for steels by Pellini and others [134] at the
Naval Research Laboratories. This test determines the highest tempera-
ture at which a crack will propagate through a specimen undergoing limited
deformation. For steels this temperature is distinet and reproducible and
has been labeled the NDT (nil ductility transition) temperature. Com-
mercial Zircaloy—2 showed surprisingly good properties from this test.
The NDT temperature was —160°C for the longitudinal direction and be-
tween —100 and —150°C for the transverse direction. However, contrary
to the definition, when the fracture faces were examined, shear lips were
found even at very low temperatures. The NDT temperature was well
below the lower break in the Charpy V impact curve on the flat portion
of the curve.

Titansum. Data available from manufacturers on the mechanical prop-
erties of titanium are generally adequate for present limited uses.

The NDT temperature for A-55 titanium was measured at ORNL on
several plates and found to be below —200°C. Incomplete fractures were
found at this temperature for both longitudinal and transverse specimens.

As a portion of the brittle fracture study a plate of A-70 titanium con-
taining 400 ppm of hydrogen was tested. The results from this plate gave
impact curves that were not affected by orientation and had a sharp
break similar to steels. The NDT temperature for this brittle plate was
100°C and was located at the lower break of the impact curve. With this
5-8] HOMOGENEOUS REACTOR METALLURGY 271

 

 

Impact Strength , ft lbs
o
o

Impact Strength, ft Ibs

 

 

  

 

 

 

 

  

 

10 — 0 | 1t 1
o l | l l | -200—-100 O 100 200 300 400 500
—200 —100 O 100 200 300 400 Testing Temperature , °C
Testing Temperature , °C
FABRICATED BY COMMERCIAL PRACTICE FABRICATED BY DEVELOPED PROCEDURES

Fiac. 5-24. Impact energy curves for Zircaloy-2 fabricated by two techniques.

material, the transition region appears to be a function of hydrogen solution
in the alloy.

Irradiated metals. In spite of the many questions that have been raised
about the adequacy of mechanical property tests for hexagonal metals, the
difficulties of which are magnified in testing irradiated samples, some me-
chanical property tests of irradiated samples have been attempted. Subsize
tensile and impact specimens of Zircaloy-2, crystal-bar zirconium, and
A—40 titanium have been irradiated in fissioning uranyl-sulfate solution,
in in-pile corrosion loops, at temperatures of 250 to 280°C, to total fast
fluxes (>1 Mev) of up to 3 X 10'® nvt for zirconium and 10 for titanium
[133]. Zirconium appears to be resistant to radiation damage under these
conditions. The only changes noted have been small changes in yield point.
Titanium seems to undergo some embrittlement. The tensile and yield
strengths for titanium increased while the reduction in area decreased. The
changes were by about 109, of the unirradiated values.

- 5-8.4 Welding of titanium and zirconium. The welding of titanium and
zirconium and their alloys is complicated by the fact that at elevated
temperatures they will react to form brittle alloys with oxygen, nitrogen,
hydrogen, water vapor, and carbon dioxide. Successful welding is, therefore,
dependent upon protecting the molten metal and adjacent area from all
contaminants. Prior to the construction of the HRE-2 core tank, the only
method used for welding zirconium was fusion welding, with the necessary
protection achieved by conducting the operations inside inert-atmosphere
boxes [135]. The use of 5/16- and 3/8-in. plate for the HRE-2 core required
272 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

 

5/16in.

]

   

3/16-in.R ,
1/16in.
Milled from 5/32-in. Diameter Zr-2 Wire 5/32in.

Zircaloy-2 Trailer Weldment

,+10
% 0Ty
l 1/4-in.R7
0.251 in.

’
to
0.750in.

o~
0.020-0.025 in.J —’. L—O-l /32in.

 

 

 

Titanium Air Weldment

Fia. 5-26. Joint configuration for HRE-2 core tank Zircaloy-2 trailer and ti-
tanium air weldments.
5-8] HOMOGENEOUS REACTOR METALLURGY 273

the development of multipass welding techniques, while the large size of
the vessel made the use of atmosphere boxes impractical.

In a joint effort, the Newport News Shipbuilding & Dry Dock Company
and Oak Ridge National Laboratory developed a machine for multipass
welding of Zircaloy—2 in which the face protection was achieved by the use
of trailers attached to the torch [127]. Root protection was achieved either
by special backup devices or, where possible, by purging closed systems.
In this device, a standard Heliarc torch was mounted on a small metal box,
or trailer, Fig. 5-25. Both the method of supporting the trailer and its
shape were varied with the configuration of the part being welded. The
underside of the trailer was open and the lower edges of the sides were con-
toured to closely match the shape of the part being welded. Inert gas was
distributed through the box by internal copper tubes along both sides.
For all these welds, the torch and trailer were held fixed by a device which
positioned the torch and spring loaded it against the work. The work
pieces were moved under the torch, which remained fixed in a vertical posi-
tion. The welding, with all the necessary adjustments and controls, re-
quired two operators.

The joint configuration of a typical Zircaloy—2 weld is shown in Fig. 5-26.
All root passes were made using the preplaced insert which had been
machined from swaged Zircaloy—-2 wire. The inserts were tacked into place
with a hand torch before the machine welding was started. The first filler
metal pass was made with either 3/32- or 1/8-in. wire and all others with
1/8-in. Careful wire brushing with a stainless steel brush was required after
every pass. Welding details for a typical joint are tabulated in Table
5-17; detailed procedures are given in Ref. 27.

In both this work and that which subsequently developed out of it, one
of the major problems has been to determine the quality of the weld. Welds
were examined visually, with liquid penetrants, and usually with radiog-
raphy; however, none of these provided information about the degree of
contamination. Measurement of the average microhardness of a prepared
sample has proved to be an acceptable and sensitive indication of contami-
nation. While this is a destructive test, it has been adopted as a standard
test for this work. In the actual production welding, reliance had to be
placed upon the strict adherence to procedures previously shown to be
satisfactory by the destructive tests. Average hardnesses of over 500 DPH
(diamond pyramid hardness) are found in lightly contaminated welds
and increase to approximately a thousand diamond pyramid hardness
numbers in contaminated metal. With this welding procedure average
hardnesses of less than 200 DPH were consistently obtained. The absence
of contamination was confirmed by vacuum fusion analyses of weld sections.

Although this welding procedure was slow and awkward, it was used
successfully for the fabrication of the HRE-2 core tank. Many satis-
274 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 5-17

SuMMARY OF CONDITIONS FOR MAKING
TITANIUM AND ZIRCONIUM WELDMENTS

 

 

 

Zircaloy~2 Titanium
core tank
Current, dec, single phase, amp
Tacks 80-100 40-55
Root pass 110 35-50
Filler pass (2) 150 35-90
Other filler passes 185 60-90
Voltage 19-23 10-16
Filler wire 1/8"” 3/32"
Welding speed, in/min 4-6 1-2
Gas flow, cth
Torch 40, helium 10-18, argon
Backup helium
First pass 40 30-60
Others 75 30-60
Trailer helium 65 —

 

 

 

 

 

factory multipass welds were completed in the assembly of the Zircaloy—2
tank.

Construction of a titanium circulating system for a homogeneous reactor
requires many welds of all sizes, some of which must be made in place.
While the procedure discussed above for zirconium can also be used for
titanium, it is impractical for field welding of equipment and piping of
varying sizes and shapes.

It has been found possible to make acceptable weldments in unalloyed
titanium using only conventional tungsten, inert gas, arc-welding equip-
ment [136]. By welding an A-55 base plate (average hardness of about
160 DPH for the annealed plate) with a filler rod of A—40, it is possible to
consistently obtain welds with average hardnesses of less than 190 DPH
(10 kg) for plates 1/4—in. and less in thickness. Hardnesses of less than
220 DPH are achieved for heavier plates requiring more passes. On the
basis of tensile and bend tests, a hardness as high as 240 DPH would be
acceptable.

The recommended conditions and configuration for a typical titanium
weldment are shown in Fig. 5-26 and Table 5-17. A comparison of the
recommended joint design and welding conditions for air welding of ti-
tanium with those used for trailer welding of Zircaloy—2 or for welding of
5-8] HOMOGENEOUS REACTOR METALLURGY 275

austenitic stainless steels shows an obvious trend to lower heat inputs and
the use of small molten pools. The reduced size and temperature of the
pool minimizes the possibilities of contamination during welding and also
permits more rapid cooling after the weld is completed. These changes
reduce the welding speed and thereby make this a precision welding method
rather than a high-speed production procedure. Details of the procedure,
including a comparison with welds made in atmosphere boxes, a discussion
of adequate inert gas coverage, and the use of surface discoloration as an
inspection method, are given in Ref. 136. In an effort to simplify and make
more versatile the procedures used for welding Zircaloy—2 an attempt was
made to adopt the inert gas procedures successfully developed for titanium.
Under similar conditions it was, however, found to be more difficult to
prevent contamination in Zircaloy—2 than it had been with titanium. Where
multipass welds in titanium had resulted in hardness increases of 20 to 30
DPH numbers, similar welds in Zircaloy—-2 increased by 50 to 100 numbers.
It was also necessary to increase the radius used on the bend tests of the
weldments from twice the plate thickness with titanium to four times the
thickness with Zircaloy, primarily because of the lower ductility of the
base metal. While multipass welds of this quality would be acceptable for
many applications, additional improvement will be necessary to make them
generally acceptable.

5-8.5 Combustion of zirconium and titanium. After several equipment
failures in which evidences of melted titanium were found under operating
temperatures not exceeding 250°C, a study of the ignition and combustion
of titanium and zirconium was started. This work has been done by Stan-
ford Research Institute under the direction of E. M. Kinderman [137].

Two types of tests were developed for studying the ignition reactions;
in one, a thin metal disk was fractured either by gas pressure or by a
plunger, while in the other, either a thin sheet or a 1/4-in. rod was broken
in a tensile manner. In either test, both the composition and pressure of the
atmosphere could be varied. With both types of tests, it proved to be
surprisingly easy to initiate combustion of titanium. Ignition and complete
consumption of both disks and rods occurred when titanium was ruptured,
even at room temperature, in a high pressure of pure oxygen.

The limiting conditions for the ignition of A-55 titanium varied with
the total pressure, the percentage of oxygen and the gas velocity. Under
dynamic conditions in pure oxygen, ignition occurred at total pressures as
low as 50 psi but with 50% oxygen the total pressure required was 700 psi
with the reaction curve becoming asymptotic with the pressure axis at
about 35% oxygen, showing no ignition would occur at any total pressure
with this or lower oxygen concentrations. The corresponding values for
static systems are 350 psi (100% O¢2), 1900 psi (50% O2), and 45% oxygen
276 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

(no ignition at any pressure). Similar results were obtained whether the
oxygen was diluted with steam or with helium. In no case was a reaction
noted when samples were fractured under oxygenated water.

As might be expected, lower critical oxygen pressures were required for
the propagation of the combustion than for the ignition. If the molten
spot was formed by an external means, propagation and consumption of the
entire sample occurred with about 20% oxygen.

Because of the autoignition of titanium, various metals and alloys such
as stainless steel, aluminum, magnesium, iron, tantalum, columbium,
molybdenum, Zircaloy—2, and alloys of titanium were tested with oxygen
pressures up to 2000 psi. Only the titanium alloys and Zircaloy—2 reacted.
Within the limits of accuracy of these studies (450 psi) no differences were
found in the behavior of the various titanium alloys.

While zirconium is similar to titanium in that autoignition can occur,
the critical oxygen pressures appear to be considerably higher. Whereas,
under dynamic conditions, titanium ignited with 50 psi oxygen, a zirconium
disk required 500 psi. Under static conditions a 0.015-in. thick strip of
titanium ignited at a pressure of 350 psi but a similar strip of zirconium
required 750 psi. A 1/4-in. titanium rod ignited under the same conditions
as the strip, but a zirconium rod did not ignite at 1500 psi oxygen.

5-8.6 Development of new zirconium alloys. An alloy developnrent
program was started at ORNL to find a radiation-corrosion-resistant
zirconium-base alloy with satisfactory metallurgical properties. These
include weldability, strength, ductility, formability and stability.

A wide variety of binary zirconium alloys was exposed in autoclaves
and in in-pile loops to uranyl-sulfate solutions (see Section 5-5); however,
only alloys of zirconium with niobium, palladium, or platinum showed
greater corrosion resistance than Zircaloy-2. With the exception of the
phase diagrams and a few mechanical property tests on low-niobium al-
loys [138], no information on any of these systems was available in the
open literature. The few results for the mechanical property tests indicated
very brittle alloys.

- Two phase diagrams for the zirconium-niobium system are in the litera-
ture [139,140]. A cursory check of the diagram by determining the eutec-
toid temperature and approximate composition confirmed the first of these
[139]. A few preliminary transformation specimens, with near-eutectoid
compositions, revealed complicated and embrittling transformation struc-
tures. Since the zirconium-niobium alloy system showed promise from a
corrosion viewpoint, these alloys with niobium contents varying from 2 to
33 w/o and many ternary alloys with small additions to the Zr-15Nb
base were studied with the objective of eliminating the undesirable prop-
erties. Information obtained includes the transformation kinetics and
5-8] HOMOGENEOUS REACTOR METALLURGY 277

products, morphologies, in-pile corrosion resistance, fabrication techniques,
metallographic procedures, and some mechanical properties [141]. At
least three transformation reactions occur in the zirconium-niobium binary
system. The transformation sequence is quite complex, with the only
straightforward transformation being a eutectoidal transformation oc-
curring close to the eutectoid temperature. The most troublesome trans-
formation is the formation of an omega phase which occurs in beta-quenched
and reheated samples. Time-temperature-hardness studies for material
heat treated in this manner showed very high hardnesses in short times
with low temperature aging treatments. Aging times of three weeks did
not result in over-aging and softening of such material. This transformation
would make multipass welding of this material very difficult.

Because of the hardness and slow transformations of the binary zirco-
nium-niobium alloys, ternary additions to the Zr-15Nb base of up to
5 w/o Mo, Pd, and Pt, up to 29, Fe, Ni, Cr, Al, Ag, V, Ta, and Th and
0.59, Cu have been studied. In general, the primary effects of the addition
of small amounts of the ternary substitutional alloying elements are the
lowering of the maximum temperature at which the hardening reaction
can occur, an increase in incubation time for the beginning of the hardening
reactions, a lowering of the temperature for the most rapid rate of harden-
ing, and an increase in rate for the higher-temperature conventional
hypoeutectoid reaction [142]. Of the ternary additions, Fe, Ni, and Cr
have the least effect, Pt and Pd an intermediate effect, and Mo the largest.
The additions of Cu and Al drastically reduced the maximum temperature
at which the hardening transformation took place and reduced the
temperature at which the hardening transformation took place at the
maximum rate, but did not increase the incubation period for the reaction
sufficiently to prevent exeessive hardening in the heat-affected zone of a
weldment. The addition of 5 w/o Ta to the Zr—-15Nb base alloy resulted
in the formation of a completely martensitic structure on quenching from
the beta field, and the addition of oxygen by the use of sponge Zr resulted
in an increase in the rate of all transformations. The addition of 29, Pd or
Pt to Zr-15Nb delays the hardening sufficiently to make welding possible;
however, it would be necessary to follow it by a stabilizing heat treatment.

Cursory fabrication studies have been performed during the course of
the alloy development program in the preparation of sheet specimens for
studies of the transformation kinetics [143]. All the Zr-Nb—-X alloys have
been hot-rolled from 800°C quite successfully. A sponge-base Zr—-15Nb arc
casting has been successfully extruded at 950°C to form rods. While the
fabrication techniques developed are adequate for the production of speci-
men material, they are not necessarily optimum for the commercial pro-
duction of plate, sheet, bar, rod, and wire in these alloys.

An ultimate tensile strength of Zr-15Nb at room temperature of 200,000
278 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

psi, with no elongation was obtained on specimens beta-quenched and
aged at 400°C for 2 hr. Similar specimens aged at 500°C for 2 weeks and
measured at room temperature had an ultimate strength of 150,000 psi, -
a yield strength of 135,000 psi, and an elongation of 109;. At 300°C this
material had an ultimate strength of 105,000 psi, a yield of 90,000 psi,
and an elongation of 169} in one inch.

While the Zr-15Nb base ternary alloys are still of interest primarily
for their potential improvement in corrosion resistance, they show promise
of a wider, general use as structural material. These are the first zirconium
alloys that may be heat-treated to high strength and yet are weldable and
fabricable.

5-8.7 Inspection of metals by nondestructive testing methods. Of the
several methods of nondestructive inspection, a visual examination, aided
by contrast dye or fluorescent penetrant, can be used to detect very fine sur-
face cracks or other surface flaws [144]. Radiography with penetrating
radiation 1s used, especially for weldments, to inspect the interior of
materials [145]. Such tests are limited in that unfavorably oriented cracks
and laminations are very difficult to detect and the methods are slow and
expensive. Ultrasonic methods are not subject to the same limitations as
radiography. Defects which are oriented unfavorably for radiography are
often readily detected by pulse-echo ultrasonic inspection.

An 1immersed, pulse-echo, ultrasonic technique developed in the Non-
destructive Test Development Laboratory for inspecting tubular products
[146] utilizes water as a coupling medium for the ultrasound. A very short
pulse of 5-megacycle ultrasonic energy is directed through the water and
into the wall of the tube under inspection. The tube is rotated while the
source of the ultrasound, a lithium-sulfate transducer, is moved along the
tube. The angle at which the sound is incident upon the tube wall is care-
fully adjusted, and echoes from defects are amplified and processed with
commercial equipment. This instrumentation includes an *“A scan,”
which is a cathode-ray tube presentation of echo amplitude on a horizontal
time base, and a "B scan’ which presents time on the vertical sweep, the
horizontal sweep representing the rotation of the tubing being inspected,
and the echo amplitude being indicated by brilliance. The “B scan”
presentation is used as a visual aid in the interpretation of the ultrasonic
reflections. These echoes are compared with the echoes from internal and
‘external notches of known depth in identical tubing in order to estimate
the depth of the flaw which causes the echo.

A remote ultrasonic technique is used to monitor the thickness of the
HRE-2 core vessel without the necessity for access to both sides of the
vessel wall. The method consists of introducing a swept-frequency beam
of ultrasound into the wall of the vessel, which is thereby induced into
5-8] HOMOGENEOUS REACTOR METALLURGY 279

vibration at its resonant frequency and harmonics thereof. The resonances
are sensed by the exciting transducer, and the signals from the resonances
are amplified, processed, and displayed on a cathode-ray tube. In most
cases layers of corrosion products inside the vessel are not included in the
- thickness measured and, therefore, the measurement represents the thick-
ness of sound, uncorroded metal. However, a slight increase (3 to 6 mils) has
been noted in the tank thickness since it was installed and the reactor
operated. It is not known whether this is due to experimental inaccuracies,
to changes in the metal, or to the presence of scale.

Simple shapes such as small-diameter tubing can be tested very rapidly
by eddy-current methods. Such methods have been developed and used
at ORNL to identify and sort various metals, to determine the thickness
of clad or plated layers, to make rapid dimensional measurements of tubing,
and to detect flaws in thin metal sections. Because of the large number of
variables which affect eddy-current inspections, the results of these tests
must be very carefully evaluated [147].

5-8.8 Radiation effects in pressure vessel steels.* The fast-neutron
dose experienced by an aqueous homogeneous reactor pressure vessel
during its lifetime may be greater than 5 X 108 neutrons/cm2. Fast-
neutron doses of this magnitude are capable of causing significant changes
in the mechanical properties of pressure vessel steels, such as loss of tensile
ductility, rise in the ductile-brittle transition temperature, and loss of
energy absorption in the notch-impact test at temperatures at which the
irradiated steel is ductile. |

For several years the Homogeneous Reactor Project has supported in-
vestigations of an exploratory nature to determine the influence of radiation
effects on pressure vessel steels [148, 149, 150]. Although it is not yet
possible to give definitive answers to many of the questions posed, it has
become apparent that radiation effects in steels depend upon a large number
of factors, and the unusual properties of irradiated metals may force a
reappraisal of the usual standards for predicting service performance from
mechanical property data.

Table 5-18 lists the tensile properties of a number of irradiated pressure
vessel steels and one weld. At doses of 5 X 108 fast neutrons/cm? appre-
ciable changes in tensile properties are observed. At doses of 1 X 1020
fast neutrons/cm? very large changes in properties are observed, and the
steels seem unsuited for use in pressure vessels in this condition because of
the limited ductility. Uniform elongation has been used as a measure of
tensile ductility because irradiation reduces the uniform elongation much
more drastically than the necking elongation. In some cases yielding and
necking occur at the same stress or, in some cases, the yield strength exceeds

*This article by J. C. Wilson.
 

 

 

 

 

280 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5
TABLE 5-18
TENSILE PROPERTIES OF IRRADIATED STEELS
Yield | Tensile )
. Dose, fast .Terr.lp..of strength | strength Umfm:m
Line Alloy ’ irradiation, elongation,
neutrons/cm? oF 7,
Thousands of psi

1 A-106 0 — 40 76 18
2 fine 2 X 1019 580 81 102 8
3 grain 2 x 101° 680 55 87 11
4 0.249, C 2 X 1019 760 48 82 12
5 8 X 1019 580 79 106* 6
6 8 X 1019 780 47 79 11
7 1 X 1020 175 97 102 4
8 A-106 0 — 46 80 14
9 | coarse grain 2 X 1019 580 93 115 8
10 0.249, C 2 X 1019 680 67 98 9
11 2 X 1019 760 43 84 14
12 7 X 1019 580 87 103* 3
13 7 %X 101 780 64 94 11
14 1 X 1020 175 116 121 2
15 A-212 0 — 40 75 25
16 0.29, C 2 X 1019 175 92 98 6
17 2 X 1019 560 76 102 9
18 2 X 1019 680 61 90 12
19 2 X 1019 760 56 84 14
20 6 X 1019 700 82 105* 6
21 6 X 1019 780 59 81 13
22 1 X 1020 175 109 116 4
32 E-7016 0 — 59 73 16
33 weld 5 X 1018 175 69 78 11
34 metal 5 X 1018 600 61 77 17
35 2 X 1019 175 108 108 0
36 6 X 1019 700 83 94 12
37 6 X 1019 740 77 85 12
38 6 X 101 780 69 77 15
39 1 X 1020 175 115 115 0

 

 

 

 

 

 

 

 

 

*Broke without necking; work-hardening rate greater than for unirradiated
specimen.
5-8]

HOMOGENEOUS REACTOR METALLURGY

TABLE 5-19

Norcu-ImpacT (SuBsizE 1zop)

PROPERTIES OF STEELS AND WELDS

281

 

 

 

 

Heat Irradia- D fast Increase in | Decrease in
Steel treat- tion ;)se, ?S , | transition “ductile”
ment* | temp., °F | "COUORS/CIT pemp. °F | energy, %
A-212B N 175 5 X 1018 45 0
(No. 18) 575 5 X 1018 15 0
175 5 X 1019 100 35
A-212 B HR 175 5 X 1018 45 20
(No. 43) 175 5 X 1019 272 50
A-212 B N & SR 175 5 X 1019 220 30
(No. 65) HAZ 175 8 X 1019 350 60
E-7016 SR 175 2 X 1019 210 40
Weld Q&T 175 8 X 1019 360 55
Carilloy Q&T 175 5 X 1018 175 20
T-1 575 5 X 1018 100 0
175 7 X 1019 450 50
81% S 175 5 X 108 100 20
Nickel 175 7 X 1019 500 60
A-106 N 175 5 X 1019 85 0
(Fine 175 8 X 1019 250 30
grain)
A-106 N 175 5 X 10'° 30 —
(Coarse 175 -8 X% 1019 300 55
grain)

 

 

 

 

 

 

* N = Normalized

SR = Stress relieved

Q & T = Quenched and tempered

S = Special heat treatment
HAZ = Heat-affected zone near weld

 
282 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

the tensile strength. The loss of uniform elongation is more severe 1n ferritic
steels of lower carbon content (less than 0.10%).

The effect of elevated irradiation temperatures is to reduce the yield
stress increase, but the tensile strength increase may be greater under the
higher temperature conditions. The uniform elongation is usually greater
for elevated temperatures of irradiation. In some cases the reduction of
area is drastically reduced by elevated temperature irradiations, and frac-
ture has occurred without necking. Thus it is not clear whether elevated
irradiation temperatures are always beneficial.

Table 5-19 lists selected data on the effects of irradiation on the notch-
impact properties of a number of steels. The increase in transition tempera-
ture and the percentage of loss of energy absorption are shown as a function
of neutron dose. The data were obtained on subsize I1zod impact specimens.
Limited information on full size Charpy V-notch specimens has indicated
that at doses of the order of 5 X 108 fast neutrons/cm?, larger transition
temperature shifts (by about a factor or two) are observed than with sub-
size Izod specimens. Thus there may be a size effect to be considered.

Preliminary results indicate that elevated irradiation temperatures in-
variably reduce the extent of radiation effects on the notch-impact proper-
ties, although the amount of reduction varies greatly between different
steels of similar composition and heat treatment.

The following suggestions and recommendations for the selection of
pressure vessel steels that will be irradiated in service are based upon the
data obtained: ~

(1) On the basis of tensile ductility (uniform elongation) steels of carbon
content greater than about 0.2% are preferable.

(2) The steel should be aluminum killed to secure a fine-grain material,
but it is not certain that a small grain size per se is preferable to a large
grain size.

(3) The processing and heat treatment should be carried out to attain
the lowest possible transition temperature before the steel is put into
service.

(4) There is some tendency for alloy steels (particularly when heat-
treated to obtain a structure that is not pearlitic) to show more severe
radiation effects than pearlitic steels. This is not to say that alloy steels
are unsuitable; but there are insufficient data to choose between the various
alloy steels.

(5) The radiation effects in steels depend both in kind and degree on
the temperature of irradiation in a very sensitive manner in the tempera-
ture ranges in which steels will be used in reactor vessels.

(6) The effects of static and cyclic stresses during irradiation on dynamic
and static properties have not yet been determined, and it is impossible
to perform a realistic evaluation of engineering properties until some
information is available.
5-9] STRESS-CORROSION CRACKING 283

(7) In the range of neutron doses to be expected in reactor vessels, the
properties of irradiated steels are extremely sensitive to dose (and perhaps
to dose rate). One of the greatest uncertainties is the comparison of the
effectiveness of test reactor fluxes and fluxes to be encountered in service.
Also, nothing is known about the rate of self-annealing at the temperature
of operation.

Currently ASTM type A-212 grade B steel made to satisfy the low-
temperature ductility requirements of ASTM A-300 specification is re-
garded as a good choice for reactor vessels. It is by no means certain that
this is the best grade of steel to use, but there are other types that seem to
be much less desirable.

59. STRESS-CORROSION CRACKING¥

5-9.1 Introduction. Early in the Homogeneous Reactor Project the
susceptibility of the austenitic stainless steels to failure by stress-corrosion
cracking was recognized. As a result, test exposures of stress specimens of
various kinds have been carried out in laboratory glassware, high tem-
perature autoclaves, 100-gpm dynamic loops, and in-pile loops. These
investigations, as well as a great deal of experience with the fluids of
interest in the engineering development programs, have indicated that
stress-corrosion cracking of type—-347 stainless steel is not a serious problem
in these environments in the absence of (<5 ppm) chloride ions. However,
investigations carried out with added chloride ions, and some failures
which have been encountered where the fluids under test were inadvertently
contaminated with chloride ions, have shown that the several oxygenated
aqueous environments present in an operating two-region breeder can
stress-crack the austenitic stainless alloys when the chloride ions are
present. The presence of iodide and bromide ions has also resulted in
localized attack in some tests [151,152]. However, bromide ions are an
unlikely contaminant and are produced in very low yield in fission, and
the 10dide formed can be removed by a silver bed [153]. Whether chloride
and bromide ions are also removed by the silver bed under reactor operating
conditions has not been established.

Thus, the successful utilization of austenitic stainless steels in homo-
geneous reactor construction is dependent on the control of localized attack,
particularly stress-corrosion cracking, by the rigid exclusion or continuous
removal of halide ions. Space does not permit a review of the many complex
factors which influence stress-corrosion cracking by chloride ions in homo-
geneous reactor fluids; however, some of the specific data and experience

*By E. G. Bohlmann.
284 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

which emphasize the necessity for halide control in homogeneous reactor
fluids contained in type—-347 stainless steel are presented. In general, the
effects of the presence of chloride ions are similar to those encountered
in high-temperature water. Stress-corrosion cracking problems in such
environments in water-cooled reactors has been comprehensively reviewed
in the Corrosion and Wear Handbook [154] and so will not be considered
in any detail here.

Titanium and zirconium alloys have been tested in many aggressive
aqueous stress corrodents without ever showing susceptibility to attack.

5-9.2 Fuel systems. An early investigation of stress-corrosion cracking
in boiling uranyl-sulfate solutions involved exposure of stress specimens in
a small, atmospheric pressure, total reflux test evaporator [155] constructed
of type—347 stainless steel. Four constant-strain type—347 stainless-steel
specimens stressed to 20,000 psi by three point loading were exposed in the
solution and vapor phase. The evaporator was run without aeration other
than that resulting from the fact that the condenser was open to the
atmosphere.

No cracking of any of the specimens was observed during the 7630-hr
exposure in HRE-2 composition (0.04 m U02804—0.004-0.015 m H2504—
0.005 m CuSOy,) solution. This solution contained less than 1 ppm chloride.
Two of the specimens, one from the liquid and one from the vapor phase,
were replaced with new specimens and the test was continued, but with
60 ppm of chloride added to the solution as sodium chloride. On examina-
tion of the specimens after 500-hr exposure in this environment, several
small cracks were found in the area of maximum stress of the new solution-
exposed specimen. No cracks were found on the new vapor-exposed speci-
men nor on any of the carryover specimens. These results were unchanged
after an additional 1880-hr exposure to the same solution except that the
cracks in the solution-exposed specimen were larger. Thus it was apparent
that stress-corrosion cracking of type—347 stainless steel at 100°C in uranyl
sulfate solutions containing chloride ions may be profoundly affected by
pretreatment.

Further investigations of this pretreatment effect in boiling uranyl
sulfate solutions in glass equipment with elastically stressed U-bend
specimens have confirmed the results obtained in the evaporator test
[156]. Table 5-20 summarizes some of the information which has been
obtained on the effect of pretreatment in a nonchloride-containing uranyl
sulfate solution on the resistance to cracking of stress specimens in subse-
quent exposure to an environment which is an aggressive crack producer
in new specimens. Thus pretreatment for a period as short as 50 hr has an
appreciable effect; also, stressing the specimen after pretreatment does
not destroy the efficacy of the solution pretreatment. A similar beneficial
5-9] STRESS-CORROSION CRACKING 285

effect was not produced by a pretreatment consisting of heating the stressed
U-bends in air for 1 hr at 677°C.

The results of some laboratory studies on the effect of chloride concen-
tration on stress-corrosion cracking of type-347 stainless steel, not pre-
treated, in aerated, boiling uranyl sulfate solutions are summarized in
Table 5-21 [157]. Under these conditions no cracking of simple beam-type

TABLE 5-20

ErFecT OF PRETREATMENT® ON STREsS CORROSION
CRACKING OF TYPE-347 STAINLESS STEEL IN BoIiLING URANYL
SUuLFATE SOLUTION®® CONTAINING CHLORIDE

 

 

 

Pretreatment No. of Time in test Specimens
time, 0.0 solution ® | cracked,
hr specimens hr o,
50 3 500 0
200 3 500 0
500 3 500 0
500 4 2500 0
500 4©) 2500 0
0 50 400 85@

 

 

 

 

 

 

(a) Pretreatment solution: 0.04 m UO2S04, 0.02 m H2SO04, 0.005 m CuSOg.
(b) Test solution: as in (a) plus 50 ppm chloride. (c) Stressed after pretreatment.
(d) 759, of the specimens cracked in < 50 hr.

specimens was observed with chloride concentrations of 10 ppm or less
over a period of 2500 hr. However, cracking was encountered at chloride
concentrations of 25 to 500 ppm. The exposure-time intervals, after which
microscope examination revealed the first evidence of cracking, ranged
from 100 to 200 hr at 25 ppm to 2000 to 2500 hr at 100 ppm. Results of a
few experiments with added bromide and iodide ions are also given. The
bromide results are suggestive of a relatively unusual stress-accelerated
corrosion rather than cracking.

Other tests [158,159] carried out in autoclaves with oxygen overpressure
have indicated that chloride concentrations as low as 5 to 10 ppm can
cause stress-corrosion cracking of vapor phase specimens at temperatures
of 100 and 250°C. Similar specimens exposed in the solution phase suffered
no comparable attack even with chloride concentrations ranging up to
90 ppm. The quite different results obtained with changes in test condi-
TABLE 5-21

STRESS-CORROSION BEHAVIOR OF TyYPE-347 STAINLESS STEEL IN
BorLiNg AND AERATED 0.04 m UO2804 — 0.02 m H2SO4 — 0.005 m
CuS0O4 SoLuTioN CoNTAINING CHLORIDE, BROMIDE, AND IODIDE

 

 

 

 

 

 

 

 

 

 

ADDITIONS
Additive
Test Total Applied i i
i Incidence of cracking
no. . hr stress, psi
Species | Conc., ppm
P-9 None 0 2500 15,000 None
30,000 None
P-10 | CI- 5 2500 15,000 None
30,000 None
P-11 | CI~ 10 2500 15,000 None
30,000 None
P-12 | CI~ 25 1000 15,000 Cracked (100-200 hr) ©
30,000 Cracked (100-200 hr)
P-13 | CI~ 50 1000 15,000 Cracked (100-200 hr)
30,000 Cracked (100-200 hr)
P-14 | Cl- 100 2500 15,000 Cracked (2000-2500 hr)
30,000 None
S-25 | CI™ 200 1500 15,000 Cracked (200-500 hr)
30,000 Cracked (200-500 hr)
S-26 | Cl~ 500 1500 15,000 Cracked (200-500 hr)
30,000 | Cracked (200-500 hr) ®
S-27 | Br~ 50 1500 15,000 No localized attack
30,000 No localized attack
P-15 | Br~ 100 2500 15,000 Severe subsurface
pitting
30,000 No localized attack
S-28 | Br— 200 1500 15,000 Severe subsurface
pitting
30,000 Severe subsurface
pitting
P-16 | I~ 100® 2500 15,000 | No localized attack
30,000 No localized attack

 

 

(a) No cracking on stress specimen; cracks occurred on stress specimen support
plate stressed at an undetermined value.

(b) Initial iodide concentration adjusted to 100 ppm at start of each 500-hr
run; iodide level at end of 500-hr runs approximately 10 ppm and less.

(¢c) Times given represent exposure interval.
5-9] STRESS-CORROSION CRACKING 287

tions are similar to experience with the phenomenon in other aqueous
environments.

A further seeming inconsistency is the large number of hours (> 300,000)
of operating experience accumulated on 100-gpm dynamic loops without
experiencing a failure due to stress-corrosion cracking, in spite of the fact
that a few ppm of chloride was often present in the solution. In a number
of instances 50 to 200 ppm chloride ion was deliberately added to the solu-
tion used in a particular run. The same loops were run for many thousands
of hours subsequently without stress-cracking failure. Also, in long-term
loop tests with solutions of the same composition as HRE-2 fuel, stress
specimens were exposed in liquid and vapor at 200, 250, and 300°C for
12,000 to 14,000 hr. No evidence of stress cracking was found by subse-
quent microscopic and metallographic examinations [160].

However, as more complex equipment has been operated in connection
with the component development programs a few stress-cracking failures
have been encountered. These were usually associated with crevices
stemming either from the design or formed by accumulations of solid
corrosion products. Comparison of these failures with the lack of difficulty
encountered in the loop experience suggested that the presence of high
concentrations of oxygen helped prevent stress-corrosion cracking by uranyl
sulfate solutions containing chloride ions and that the cracking encountered
in the more complex systems was related to oxygen exhaustion in the
crevices.

To test this hypothesis, two series of loop runs were carried out to study
the effect of oxygen concentration [161].

In the first of these, at 250°C, two sets of five stress specimens, one set
pretreated in situ 98 hr with chloride-free solution, the other set in the as-
machined, degreased condition, were exposed in the loop pressurizer. The
specimens were arranged so the topmost specimen was exposed to vapor
only, the next two were in the solution spray from the pressurizer bypass,
and the bottom two were immersed in the liquid phase. No effect ascribable
to exposure position was observed. Pretreated specimens were not included
in the second series of runs at 200°C. Table 5-22 summarizes the conditions
used in the runs and the results obtained:

At 250°C no cracks were produced in the specimens by varying conditions
from high oxygen concentration to oxygen exhaustion with consequent
uranium precipitation. At 200°C and high oxygen concentration, a crack
was observed underneath the head of a bolt used to fasten the specimen to
a holder instead of in the area of maximum tensile stress. The probability
that this was the crevice-type attack which had been hypothesized was
borne out by the results of the subsequent oxygen exhaustion run. After
this run all the specimens showed transgranular cracks. However, the
cracks were not characteristically related to the pattern of applied stress.
288 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

TABLE 5-22

ConpiTioNs oF RUuNs To INVESTIGATE STRESS-CORROSION CRACKING
IN CHLORIDE-CONTAINING URANYL SULFATE SOLUTION

 

 

 

Run |Time,|{Temp.,|U02804, C.h lo- Oxygen, Cracking
o ride, Remarks :
no. hr C m ppm results
| ppm ,
H-103 98 [ 250 0.17 0 1000 | Pretreatment —
H-104a | 143 | 250 0.17 | 40 (1500-1800 No cracks
H-104b | 250 | 250 0.17 | 40 |1500-1800| Same solution as |No cracks

H-104a
H-104¢ | 260 | 250 0.17 60 20-170 | New solution No cracks
H-104d | 111 | 250 0.17 | 50 0-40 | Oxygen exhaustion| Pits, but

occurred and no cracks
U precipitated
H-105a | 211 | 200 0.17 50 {1000-3000 | One crevice
| | crack
H-105b | 200 | 200 0.17 | 50 | 0-25 [Oxygen exhaustion|All five
occurred and specimens

 

 

 

 

 

 

 

 

U precipitated cracked

 

 

Several of the cracks were parallel rather than normal to the applied tensile
stress, and most were in the regions where the identification numbers were
stamped on the specimens rather than in the areas of maximum elastic
stress. | ~ ' |

The aggressive stress cracking encountered in the oxygen exhaustion
run at 200°C raised the question whether similar effects might be produced
in the absence of chloride. Consequently, oxygen exhaustion studies were
carried out in loop runs with 0.17 M uranyl sulfate solutions containing
<3 ppm of chloride ions. Stress specimens exposed in such runs at 200,
250, and 280°C showed no stress-corrosion - cracks on metallographic
examination [162].

In-pile experience has of necessity been substantially less extensive;
however, the results have been consistent with the negative experience
encountered with out-of-pile loops. Over a period of 23 years, about
17,000 operating hours have been accumulated in fifteen type-347 stainless-
steel in-pile loop experiments without encountering evidence of stress
cracking. Also, stress specimens were exposed in the pressurizer vapor and
liquid locations in one 1700-hr experiment. Subsequent microscopic and
metallographic examination revealed no evidence of cracks. Stress speci-
5-9] STRESS-CORROSION CRACKING 289

mens exposed in the core during a 630-hr experiment also showed no stress
cracks.

Thus, a great deal of experience and the results of many specific investi-
gations indicate stress-corrosion cracking of austenitic stainless alloys is
not a problem in uranyl sulfate solution environments free of chloride.
However, contrariwise, it also seems clear that stress-corrosion cracking

failures will be a problem if these solutions become contaminated with
chloride.

5-9.3 Slurry systems. Substantially less experience with aqueous slur-
ries than with uranyl sulfate solutions has been accumulated in austenitic
stainless steel equipment. However, it appears that stress-corrosion crack-
ing failures manifest themselves about as one might expect from high-
temperature water results. Thus the presence or absence and the concen-
tration of chloride in the slurry are major factors in stress-cracking
incidence.

Stress cracking has been encountered in toroids [163] and loops [164]
operated with oxygenated thorium-oxide slurries containing chloride. No
estimate of the concentration of chloride which is tolerable in oxygenated
slurries can be made from available data. Williams and Eckel [165] have
reported an apparent relationship between oxygen and chloride content
for the development of cracks by alkaline-phosphate treated boiler water.
These data indicate that maintenance of oxygen at concentrations of less
than 0.5 to 1 ppm will provide reasonable assurance against stress-corrosion
cracking at appreciable (10 to ~100 ppm) chloride concentrations. As a re-
sult it has been suggested that elimination of oxygen from slurry systems by
maintenance of a hydrogen overpressure on the system may be a solution
to the stress-cracking problems. A few test results indicate that this may
indeed be true under some out-of-pile conditions [166,167]; however, it is
not likely to be effective under all environmental conditions encountered
in an operating reactor. Thus, it is not clear that a hydrogen overpressure
over the radioactive aqueous fluids encountered in an operating reactor
will entirely repress the formation of oxygen by radiolytic decomposition
of the water or that other species resulting from radiation effects cannot
carry out the function of the oxygen in the stress-cracking mechanism(s)
if the oxygen is eliminated.

As with the uranyl sulfate solutions discussed above, there have been
some 1ndications that stress-corrosion cracking of austenitic alloys in
oxygenated slurry use is more likely to occur in crevices under accumu-
lations of oxide [168].

5-9.4 Secondary systems. Boiler water. Chemical treatment of the
water in the steam side of a fluid fuel reactor heat exchanger is a major
290 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS  [cHAP. 5

unresolved question in the operation of such reactors. McLain [169] has
reviewed the problems attendant on the use of conventional boiler water
treatment. These problems stem from very large radiation fluxes associated
with the radioactive fuel circulating in the primary side of the exchanger.
Radiation decomposition of inhibitors such as hydrazine and sulfite and
the consequent and perhaps unpreventable production of some oxygen by
water decomposition introduce completely novel factors to a long-standing
problem. Thus stress-corrosion cracking failures originating on the second-
ary side of the type-347 heat-exchanger tubes are of concern unless rigid
chloride exclusion can be maintained. The effectiveness of inhibitors in
preventing production of appreciable oxygen in the secondary side water
of a fluid fuel heat exchanger is one of the investigational objectives of the
HRE-2. During this period rigid chloride exclusion is the only evident
preventive measure which can be taken, although the successful accumula-
tion of 700 Mwh experience on the HRE-1 fuel heat exchanger may be
evidence of some cathodic protection by the carbon steel in the shell and
tube sheet. It is probable, however, that in the future duplex 347—Inconel
tubing, with the Inconel in contact with the secondary water environment,
will be used in this application. The comprehensive investigation of stress-
corrosion cracking in chloride-containing boiler water environments carried
out in connection with the Naval Pressurized Water Reactor program
has shown that Inconel is not susceptible to cracking failure.

Other secondary systems. Chloride should also be excluded from other
aqueous environments in contact with the austenitic stainless equipment
not easily replaceable. Thus, failures have been encountered in using
chlorinated, potable water as cooling water [170] and the presence of
marking ink containing from 3000 to 18,000 ppm chloride on the surface of
stress U-bend specimen has been shown to induce cracking on exposure to
saturated oxygenated steam at 300°C [171]. The importance of the exclu-
sion of chloride from aqueous environments contained in austenitic alloys
was clearly demonstrated by the encounter with stress-corrosion cracking in
the HRE-2 leak detector system. The undetected chloride contamination
during manufacture of some of the tubing used in fabrication of the system
resulted in a rapid succession of failures in parts of the system during shake-
down operation. Subsequent penetrant and metallographic examination of
the O-ring flanges to which the system was connected revealed stress-
corrosion cracks in many of those exposed to high temperatures. As a
result, all the high-temperature flanges were replaced before the reactor
operation schedule could be continued [172,173].
201

REFERENCES

1. L. G. DesmonDp and D. R. MosueRr, Lewis Flight Propulsion Laboratory,
National Advisory Committee for Aeronautics, 1951. Unpublished.

2. J. R. CossEY and R. LirrLEwoobp, Feastbility Tests on the Use of Torovds
for Dynamic Corrosion Testing In-pile, Report AERE C/M 281, Gt. Brit.
Atomic Energy Research Establishment, June 1956.

3. H. C. SavaGE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Mar. 15, 1952, USAEC Report ORNL-1280, Oak
Ridge National Laboratory, July 14, 1952. (p. 43)

4. H. C. SavAaGE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Mar. 31, 1953, USAEC Report ORNL-1554, Oak
Ridge National Laboratory, July 10, 1953. (p. 46)

5. C. L. SecasEr and F. C. Zarp, HRE Design Manual, USAEC Report TID-
10082, Oak Ridge National Laboratory, Nov. 18, 1952.

6. H. C. SavAaGE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Mar. 15, 1952, USAEC Report ORNL-1280, Oak
Ridge National Laboratory, July 14, 1952. (pp. 43)

7. E. L. ComPERE et al., HRP Dynamic Slurry Corroswn Studres: Quarter
Ending April 30, 1957, USAEC Report CF-57-4-139, Oak Ridge National
Laboratory, Apr. 30, 1957. (pp. 9-16)

8. F. H. SweEToN et al., The Effect of Irradiation on the Corrosion of Type
347 Stainless Steel by Uranyl Sulfate Solution, USAEC Report ORNL-2284,
Oak Ridge National Laboratory, Apr. 1, 1957.

9. K. S. WarreN and R. J. Davis, In-reactor Autoclave Corrosion Studies—
LITR. Outline of Methods and Procedures, USAEC Report CF-57-5-110, Oak
Ridge National Laboratory, May 22, 1957.

10. See Bibliography following Chapter 10 for Homogeneous Reactor Project
reports.

11. G. H. JeENKs et al., in Homogeneous Reactor Project Quarterly Progress
Report for Period Ending Jan. 31, 1955, USAEC Report ORNL-1853, Oak Ridge
National Laboratory, Feb. 16, 1955. (p. 115)

12. T. L. TrenT and A. F. ZuLLINGER, Experimental Facilities for Multiple
In-Pile Corrosion Tests, in Proceedings of the Fifth Hot Laboratory Operation
and Equipment Conference, Vol. III. New York: Pergamon Press, 1957. (pp-
253-257)

13. G. H. Jenks et al., Circulating In-reactor Loops, in Proceedings of the
Fifth Hot Laboratory Operation and Equipment Conference, Vol. III. New York:
Pergamon Press, 1957. (pp. 237-244)

14. G. H. JENKks et al., in Homogeneous Reactor Proyect Quarterly Progress
Report for the Period Endzng Oct. 31, 1955, USAEC Report ORNL-2004(Del.),
Oak Ridge National Laboratory, Jan. 31, 1956. (p. 124, Fig. 12.4)

15. A. WErTzBeRG and H. C. SavAGE, Performance Test of 220-V Three-phase
Stator for Use with 5-gpm In-pile Loop Pump, USAEC Report CF-57-10-24,
Oak Ridge National Laboratory, Oct. 4, 1957.

16. H. C. SAVAGE, Sintered Alumina as a Pump Bearing and Journal Malerial,
USAEC Report CF-57-11-122, Oak Ridge National Laboratory, Nov. 26, 1957.
292 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

17. G. H. Jenks et al., USAEC Report ORNL-2408, Oak Ridge National
Laboratory, to be issued.

18. D. T. JonEs et al., In-pile Circulating Loop Is Dismantled in This Equip-
ment, Nucleonics 12(11), 76 (November 1954).

19. J. A. RussgLL, Jr., Instrumentation for Solution-loop Experiments Inside
a Nuclear Pile, ISA Journal 9, 390 (September 1957).

20. H. G. Ducean and D. T. Jones, Corrosion Testing Facility and Dis-
assembly Equipment, in Fourth Annual Symposium on Hot Laboratories and
Equipment, Held in Washington, D. C., Sept. 29 and 30, 1955, USAEC Report
TID-5280, Oak Ridge National Laboratory, September 1955. (pp. 64-77)

21. A. M. Trippr and D. T. Jongs, Facility for Underwater Sample Prepara-
tion, in Fourth Annual Symposium on Hot Laboratories and Equipment, Held in
Washington, D. C., Sept. 29 and 30, 19565, USAEC Report TID-5280, Oak
Ridge National Laboratory, September 1955. (pp. 154-172)

22. F. A. Cuamrion, Corroston Testing Procedures, New York: John Wiley &
Sons, Inc., 1952. (p. 185)

23. N. E WoLpmAN, Engineering Alloys 3rd ed. Cleveland: American Society
for Metals, 1954.

24. J. C. GriEss et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Pertod Endmg Apr. 30, 1956, USAEC Report ORNL-2096, Oak Rldge
National Laboratory, May 10, 1956. (pp. 82-83)

25. J. L. EncuisH et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 80, 1954, USAEC Report ORNL-1753(Del.),
Oak Ridge National Laboratory, Sept. 17, 1954. (pp. 97-102) |

26. J. C. Griess et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 19566, USAEC Report ORNL-2148(Del ),
Oct. 3, 1956. (pp. 77-78)

27. See Bibliography following Chap. 10 for Homogeneous Reactor Project
reports.

28. See Bibliography followmg Chap. 10 for Homogeneous Reactor Project
reports.

29. J. C. Griess et al.,, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak
Ridge National Laboratory, Oct. 10, 1957. (pp. 72-73)

30. J. C. GriEss et al.,, Oak Ridge National Laboratory, 1954. Unpublished.

31. J. C. Grigss et al., HRP Dynamic Corrosion Studies for Quarter Ending
July 31, 1956, USAEC Report CF-56-7-52, Oak Ridge National Laboratory,
July 31, 1956. (pp. 28-36)

32. J. C. Grigss et al.,, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Oct. 31, 1955, USAEC Report ORNL-2004(Del.),
Oak Ridge National Laboratory, Jan. 31, 1956. (pp. 98-101)

33. J. C. Griess et al., Solution Corrosion Group Quarterly Report for the
Period Ending Jan. 31, 1956, USAEC Report CF-56-1-167, Oak Ridge National
Laboratory, Jan. 31, 1956. (pp. 29-33) ,

34. J. C. Griess et al., Solution Corrosion Group Quarterly Report for the
Period Ending Apr. 80, 1956, USAEC Report CF-56-4-138, Oak Ridge National
Laboratory, Apr. 30, 1956. (pp. 17-20)
REFERENCES 203

35. G. H. JeNks et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1956, USAEC Report ORNL-2222, Oak
Ridge National Laboratory, Feb. 7, 1957. (pp. 107-111)

36. J. E. Baker et al., HRP Radiation Corrosion Studies: In-pile Loop
L-4-8, USAEC Report ORNL-2042, Oak Ridge National Laboratory, Aug. 21,
1956. (p. 1)

37. G. H. JEnks et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Oct. 31, 19565, USAEC Report ORNL-2004(Del.),
Oak Ridge National Laboratory, Jan. 31, 1956. (p. 126) |

38. . H. Jenks et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2379, 1957
(p. 110); ORNL-2493, 1958.

39. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak
Ridge National Laboratory, Oct. 10, 1957. (p. 112)

40. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Jan. 31, 19565, USAEC Report ORNL-1853, Oak
Ridge National Laboratory, Feb. 16, 1955. (pp. 89, 95)

41. G. H. JENKs et al., Oak Ridge National Laboratory, in Homogeneous Re-
actor Project Quarterly Progress Report, USAEC Reports ORNL-1853, 1955
(p. 102); ORNL-1895, 1955 (p. 115); ORNL-2004(Del.), 1956 (p. 140); ORNL-
2057(Del.), 1956 (p. 95); ORNL-2148(Del.), 1956 (p. 97); ORNL-2222, 1957
(p. 105); ORNL-2272, 1957 (p. 107); ORNL-2331, 1957 (p. 116); ORNL-2379,
1957 (pp. 106-107); ORNL-2493, 1958.

42. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Pertod Ending Oct. 31, 1956, USAEC Report ORNL-2222, Oak
Ridge National Laboratory, Feb. 7, 1957. (p. 103)

43. J. A. GuorMLEY and C. J. HoCHANADEL, in Chemistry Division Semiannual
Progress Report for the Period Ending Dec. 20, 1953, USAEC Report ORNL-1674,
Oak Ridge National Laboratory, Apr. 9, 1954. (p. 76)

44. G. H. Jenks et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak Ridge
National Laboratory, 1958.

45. S. E. BEaLL et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Pertod Ending Jan. 31,1954, USAEC Report ORNL-1678, Oak Ridge
National Laboratory, Apr. 15, 1954. (p. 5)

46. G. H. JENKs et al.,, in Homogeneous Reactor Project Quarterly Progress
Report for the Pertod Ending Oct. 31, 1954, USAEC Report ORNL-1813(Del.),
Oak Ridge National Laboratory, Nov. 10, 1954. (pp. 81-82)

47. D. E. Taomas, in Metallurgy of Zirconium, ed. by B. Lustman and F.
Kerze, Jr., National Nuclear Energy Series, Division VII, Volume 4. New York:
McGraw-Hill Book Co., Inc., 1955. (p. 634) | |

48. G. H. JeNks et al.,, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 30, 1957, USAEC Report ORNL-2331, Oak
Ridge National Laboratory, Sept. 3, 1957. (pp. 120-122)

49. G. H. Jenks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2272, 1957
(p. 110); ORNL-2379, 1957 (pp. 119-121); ORNL-2222, 1957. (pp. 107-109)
294 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

50. G. H. JENks et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958. _

51. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1954, USAEC Report ORNL-1813(Del.),
Oak Ridge National Laboratory, Nov. 10, 1954. (pp. 82-86)

92. G. H. JENKS et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Jan. 31, 1956, USAEC Report ORNL-2057(Del.),
Oak Ridge National Laboratory, Apr. 17, 1956. (pp. 96-98)

93. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1954, USAEC Report ORNL-1813(Del.),
Oak Ridge National Laboratory, Nov. 10, 1954. (pp. 83-84)

54. G. H. JENKks et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 81, 1955, USAEC Report ORNL-1853, Oak
Ridge National Laboratory, Feb. 16, 1956. (pp. 113-114)

95. G. H. JENks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-1895, 1955
(pp. 109-115); ORNL-2004(Del.), 1956 (pp. 131-145); ORNL-2057(Del.), 1956
(p. 94); ORNL-2148(Del.), 1956 (p. 96); ORNL-2222, 1957 (pp. 102-103); .
ORNL-2272, 1957 (pp. 105-106); ORNL-2331, 1957 (p. 115). J. E. BakEr
et al., HRP Radiation Corrosion Studies: In-pile Loop L-4-8, USAEC Report
ORNL-2042, Oak Ridge National Laboratory, Aug. 21, 1956. (pp. 16-17)

56. G. H. Jenks et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 81, 1955, USAEC Report ORNL-1853, Oak
Ridge National Laboratory, Feb. 16, 1956. (p. 113)

57. G. H. JENKs et al.,, in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958.

58. G. H. Jenks et al,, Oak Ridge National Laboratory, in H omogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2379, 1957
(p. 109); ORNL-2432, 1958 (p. 116); ORNL-2004(Del.), 1956 (pp. 146-149);
ORNL-1895, 1955 (pp. 115-118); ORNL-1943, 1955 (pp. 121-124); ORNL-
2057(Del.), 1956 (p. 94); ORNL-2148(Del.), 1956 (p. 96); ORNL-2222, 1957
(pp. 103-104); ORNL-2331, 1957 (pp. 116-118). J. E. BakEr et al., HRP
Radiation Corrosion Studies: In-pile Loop L-4-8, USAEC Report ORNL-2042,
Oak Ridge National Laboratory, Aug. 21, 1956. (pp. 19-20)

99. G. H. JEnNks et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2272, 1957
(pp. 107-108); ORNL-2493, 1958.

60. G. H. JEnNks et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Pertod Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak
Ridge National Laboratory, Jan. 21, 1958. (pp. 120-129)

61. G. H. JeNks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2222, 1957
(pp. 107-111); ORNL-2096, 1956 (pp. 92-94); ORNL-2148(Del.), 1956 (pp.
98-103) ; ORNL-2057(Del.), 1956. (pp. 96-98)

62. G. H. JENKs et al.,, Oak Ridge National Laboratory, in Homogeneous
REFERENCES 205

Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2222, 1957
(pp. 107-109); ORNL-2432, 1958. (pp. 120-129)

63. G. H. JENks et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2222, 1957
(pp. 107-111); ORNL-2493, 1958.

64. G. H. JENKSs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak Ridge
National Laboratory, Jan. 21, 1958. (pp. 124-126)

65. G. H. JENks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2148(Del.),
1956 (pp. 93-96); ORNL-2432, 1958. (pp. 115-116)

66. G. H. JEnks et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-1943, 1955
(pp. 126-131); ORNL-2331, 1957. (pp. 120-123)

67. G. H. JEnks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Quarterly Progress Report, USAEC Reports ORNL-2432, 1958 (p. 130);
ORNL-2222, 1957. (p. 112)

68. C. H. HocuANADEL, Radiation Stability of Aqueous Fuel Systems, USAEC
Report Cf-56-11-54, Oak Ridge National Laboratory, 1956.

69. G. H. KincHIN and R. S. PEASE, Repts. Progr. in Phys. 18, 1 (1955).

70. G. H. Jenks, Effects of Radiation on the Corrosion of Zircaloy-2, USAEC
Report CF-57-9-11, Oak Ridge National Laboratory, 1957.

71. G. H. JENKS et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak Ridge
National Laboratory, Jan. 21, 1958. (pp. 120-126)

72. G. H. JENKSs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Pervod Ending Jan. 81, 1958, USAEC Report ORNL-2493, Oak Ridge
National Laboratory, 1958.

73. G. H. JEnks et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2004(Del.),
1956 (p. 134); ORNL-2222, 1957. (p. 102) |

74. G. H. JeNks et al., Oak Ridge National Laboratory, in Homogeneous Re-
actor Project Quarterly Progress Report, USAEC Reports ORNL-2004(Del.),
1956 (p. 137); ORNL-2148(Del.), 1956 (p. 97); ORNL-2222, 1957 (p. 103);
ORNL-2272, 1957 (p. 107); ORNL-2379, 1957 (pp. 106-107) ; ORNL-2432, 1958
(p. 116); ORNL-2493, 1958. J. E. BAKER et al., HRP Radiation Corrosion
Studies: In-pile Loop L-4-8, USAEC Report ORNL-2042, Oak Ridge National
Laboratory, Aug. 21, 1956. (pp. 21-23)

75. G. H. JENks et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-1895, 1955
(pp. 120-122); ORNL-2148(Del.), 1956. (pp. 99-103)

76. G. H. JENKs et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Endzng Oct. 31, 1957, USAEC Report ORNL-2432, Oak Ridge
National Laboratory, Jan. 21, 1958. (p. 130)

77. C. F. Hiskey, Chemical Research. The Heavy-water Homogeneous Pile:
A Review of Chemical Researches and Problems, USAEC Report CC-1383,
Argonne National Laboratory, 1944.
296 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

78. I. KirsueENBAUM et al., (Eds.), Utilization of Heavy Water, USAEC Report
TID-5226, Columbia University, Substitute Alloy Material Labs., 1951.

79. No corrosion data are available to the writer from this source.

80. J. E. KenToN, Nucleonics 15(9), 166-184 (September 1957).

81. A. S. Krrzes and R. N. Lyon, Aqueous Uranium and Thorium Slurries,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 9. New York: United Nations, 1956. (P/811, p. 414)

82. H. F. McDurriE, Corrosion by Aqueous Reactor Fuel Slurries, USAEC
Report CF-57-4-51, Oak Ridge National Laboratory, 1957.

83. E. L. CompERE, in HRP Cwnlian Power Reactor Conference Held at Oak
Ridge National Laboratory, May 1-2, 1957, USAEC Report TID-7540, Oak
Ridge National Laboratory, 1957. (pp. 249-265)

84. E. L. CompPERE et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-990, 1951;
ORNL-1121(Rev.), 1952; ORNL-1221, 1952; ORNL-1280, 1952; ORNL-1318,
1952; ORNL-1424(Del.), 1953; ORNL-1478(Del.), 1953; ORNL-1554, 1953;
ORNL-1605, 1953; ORNL-1813(Del.), 1954; ORNL-1853, 1955; ORNL-1943,
1955; ORNL-2004(Del.), 1956; ORNL-2057(Del.), 1956; ORNL-2148(Del.),
1956; ORNL-2222, 1957; ORNL-2272, 1957; ORNL-2331, 1957; ORNL-2379,
1957; ORNL-2432, 1958.

85. G. E. Moore and E. L. ComPERE, Small-scale Dynamzc Corroston Studies
in Toroids. Aqueous Thortum Ozide Slurries, USAEC Report ORNL-2502, Oak
Ridge National Laboratory, to be issued.

86. D. J. DePavuL (Ed.), Corrosion and Wear Handbook for Water Cooled
Reactors, USAEC Report TID-7006, Westinghouse Electric Corp., 1957.

87. L. ScuEeiB, Investigation of Materials for a Water-cooled and -moderated
Reactor, USAEC Report ORNL-1915(Del.), Oak Ridge National Laboratory,
1954. (pp. 28-29) |

88. H. H. Unuie (Ed.), Corrosion Handbook, New York: J. Wiley & Sons,
1948.

89. C. F. Hiskry, see reference 77.

90. J. H. Perry (Ed.), Chemical Engineers Handbook, 3rd ed. New York:
McGraw-Hill Book Co., Inc., 1950. (p. 1022)

91. D. G. THOMAS, Comments on the Erosweness of ThOg Slurries, USAEC
Report CF-55-4-36, Oak Ridge National Laboratory, 1955; Attack of Circulating
Aqueous ThOg Slurries on Stainless Steel Systems, USAEC Report CF-56-1-21,
Oak Ridge National Laboratory, 1956.

92. H. F. McDurrIE, see reference 82. |

93. R. V. BaiLey, Erosion Due to Particle Impingement upon Bends in Circular
Conduits, USAEC Report ORNL-1071, Oak Ridge National Laboratory, 1951.

94. D. G. Tuomas, Comments on the Erosiveness of ThOz Slurries, USAEC
Report CF-55-4-36, Oak Ridge National Laboratory, 1955.

95. R. V. BaiLEY, see reference 93.

96. L. PranpTL, Essentials of Flurd Dynamaics, English translatlon New
York: Hafner Publishing Co., Inc., 1952. (pp. 136, 349)

97. J. M. CouLson and J. F. RicaarDsoN, Chemical Engineering, Vol. 1. New
York: McGraw-Hill Book Co., Inc., 1954. (Chap. 9)

98. R. J. HugHEs, Ind. Eng. Chem. 49, 947-955 (1957).
REFERENCES 297

99. K. K. SHALNEV, Experimental Studies of the Intensity of Erosion Due to
Cavitation, in Proceedings of Symposium on Cavitation tn Hydrodynamics Held
at the National Physical Laboratory on Sept. 14—17, 1955. London: H. M. Sta-
tionery Office, 1956. (Paper 22)

100. R. E. JoaxstoNE and M. W. TurING, Pilot Plants, Models, and Scale-Up
Methods in Chemical anmeermg, New York: McGraw-Hill Book Co., Inc.,
1957. (p. 247) ~

101. S. J. ROSENBERG, J. Research Natl. Bur Standards 5, 553 (1930).

102. R. V. BAILEY, see reference 93. |

103. R. L. SToKER, Ind. Eng. Chem. 41, 1196-1199 (1949).

104. G. E. Moore and E. L. CoMPERE, see reference 85.

105. J. P. McBripE, The Abrasive Properties of Thortum Ozide, USAEC
Report CF-53-8-149, Oak Ridge National Laboratory, 1953; and unpublished
work.

106. C. F. Hiskry, see reference 77.

107. H. peE BruYN et al., Homogeneous Oxide Suspensmn Reactor, in Pro-
ceedings of the International Conference on the Peaceful Uses of Atomic Energy,
Vol. 3. New York: United Nations, 1956. (P/936, p. 116)

108. C. E. Currtis, Oak Ridge National Laboratory, personal communication.

109. A. S. Krrzes and R. N. Lvon, Aqueous Uranium and Thorium Slurries,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 9. New York: United Nations, 1956 (P/811, p. 414). A.S. Kitzes
and R. N. Lyon, Oak Ridge National Laboratory, in Homogeneous Reactor
Project Quarterly Progress Report, USAEC Reports ORNL-990, 1951 (p. 143);
ORNL-1121(Rev.), 1952 (p. 160); ORNL-1221, 1952 (p. 131); ORNL-1280,
1952 (p. 83); ORNL-1318, 1952 (p. 88); ORNL-1424(Del.), 1953 (pp. 35-38);
ORNL-1478(Del.), 1953 (p. 107); ORNL-1554, 1953 (p. 122); ORNL-1605, 1953
(pp. 150-153). J. O. BLoMEKE, Aqueous Uranium Slurry Studies, USAEC
Report ORNL-1904, Oak Ridge National Laboratory, 1955.

110. E. L. CompERE et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-990, 1951;
ORNL-1121(Rev.), 1952; ORNL-1221, 1952; ORNL-1280, 1952; ORNL-1318,
1952; ORNL-1424(Del.), 1953; ORNL-1478(Del.), 1933; ORNL-1554, 1953;
ORNL-1605, 1953; ORNL-1813(Del.), 1954; ORNL-1853, 1955; ORNL-1895,
1955; ORNL-1943, 1955; ORNL-2004(Del.), 1956; ORNL-2057(Del.), 1956;
ORNL-2148(Del.), 1956; ORNL-2222, 1957; ORNL-2331, 1957; ORNL-2379
1957; ORNL-2432, 1958.

111. G. E. Moore and E. L. COMPERE, see reference 85. :

112. J. D. KenToN, Nucleonics 15(9), 166-184 (September 1957).

113. C. E. Curtis, Oak Ridge National Laboratory, personal communication.

114. G. E. Moore and E. L. CoMPERE, see reference 89. :

115. E. L. CompERE and S. A. REED, in Homogeneous Reactor Project Quarterly
Progress Report for the Period Ending Jan. 31, 1966, USAEC Report ORNL-
2057(Del.), Oak Ridge National Laboratory, Apr. 17, 1956. (p. 89). G. E. MOORE
and E. L. Compere, USAEC Report ORNL-2502, see reference 8.

116. H. F. McDurriE, see reference 82. D. G. Truomas, Attack of Circulating
Aqueous-ThOg Slurries on Stainless Steel Systems, USAEC Report CF-56-1-21,
Oak Ridge National Laboratory, Jan. 5, 1956.
298 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

117. G. E. Moore and E. L. CompERE, Oak Ridge National Laboratory, in
Homogeneous Reactor Project Quarterly Progress Report, USAEC Reports ORNL-
1813(Del.), 1954 (p. 100); ORNL-1853, 1955. (p. 126)

118. D. G. Tuomas, Attack of Circulating Aqueous-ThOs Slurries on Stainless
Steel Systems, USAEC Report CF-56-1-21, Oak Ridge National Laboratory,
Jan. 5, 1956.

119. H. F. McDurriE, see reference 82.

120. G. E. Moore and E. L. CoMPERE, see reference 85.

121. J. ScumeTs and M. PourBaix, Corrosion of Titanium, Proc. 6th Meeting
Intern. Comm. Electrochem. Thermodynam. and Kinetics, 1955. (pp. 167-179)

122. D. J. DEPavuL (Ed.), Corrosion and Wear Handbook, USAEC Report
TID-7006, Westinghouse Electric Corp., 1957. (p. 17)

123. K. K. SHALNEV, see reference 99.

124. E. L. ComPERE, in Homogeneous Reactor Project Quarterly Progress Re-
port for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958.

125. RicHARDSs et al., Melting and Fabrication of Zircaloy, in Proceedings of
the 2nd International Conference on the Peaceful Uses of Atomic Energy, Geneva,
1958. -

126. L. F. BLEDsoE et al.,, Fabrication of the Homogeneous Reactor Test
Vessel Assembly, Welding J. 35, 997-1005 (October 1956).

127. G. E. ELpER et al., First Zirconium Vessel for HRT Reactor, J. Metals 8,
648-650 (1956).

128. M. L. PickLESIMER, Anodizing as a Metallographic Technique for Zir-
contum Base Alloys, USAEC Report ORNL-2296, Oak Ridge National Labora-
tory, 1957.

129. G. M. ApamsoN and M. L. PicKLESIMER, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Jan. 31, 1956, USAEC Report
ORNL-2057(Del.), Oak Ridge National Laboratory, Apr. 17, 1956. (pp. 101-104)

130. L. K. JETTER and B. S. Borig, Jr., A Method for the Qualitative Deter-
mination of Preferred Orientation, J. Appl. Phys. 24, No. 5, 532-535 (May 1953).

131. L. K. JETTER et al., A Method of Presenting Preferred Orientation Data,
J. Appl. Phys. 27, No. 4, 368-374 (April 1956).

132. M. L. PickrLESIMER and G. M. ApaMmsoN, Development of a Fabrication
Procedure for Zircaloy-2, USAEC Report CF-56-11-115, Oak Ridge National
Laboratory, Nov. 21, 1956.

133. G. M. ApamsoN and J. J. PRISLINGER, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Jan. 31, 19568, USAEC Report
ORNL-2493, Oak Ridge National Laboratory, 1958: See also Homogeneous
Reactor Project Quarterly Progress Report for the Pertod Ending Apr. 31, 1958.

134. P. P. Puzaxk et al., Crack-Starter Tests of Ship Fracture and Project
Steels, Welding J. Res. Supplement 33, 433s-441s (September 1954).

135. B. LustmaN and F. KEeRrzE, Metallurgy of Zirconium, National Nuclear
Energy Series, Division VII, Volume 4. New York: McGraw-Hill Book Co.,
Inc., 1955. (pp. 307-320)

136. G. M. ApamsoN and W. J. LEoNARD, Inert Gas Tungsten Arc Welding
of Titanium for the Nuclear and Chemical Industries, Welding J. (N.Y.) (to be
published). |
REFERENCES 299

137. F. E. LrrtmaN, Stanford Research Institute, Reactions of Titanium with
Water and Aqueous Solutions, Quarterly Report 1 (AECU-3581), 1957; Quar-
terly Report 2 (AECU-3582); see also Quarterly Reports 3 and 4.

138. G. L. MiLLER, Metallurgy of the Rarer Metals-2-Zircontum, New York:
Academic Press, Inc., 1954. (pp. 157-158)

139. B. A. RogErs and D. F. Atkins, The Zirconium-Columbium Phase
Diagram, J. Metals 7, 1034 (1955).

140. Yu Bycuxkov et al., Some Properties of Alloys of Zirconium with Niobium,
Soviet Journal of Atomic Energy (English Translation) 2(2), 165-170 (1957).

141. G. M. ApamsoN and M. L. PickrEsIMER, In Homogeneous Reactor
Project Quarterly Progress Report for the Period Ending July 31, 1957, USAEC
Report ORNL-2379, Oak Ridge National Laboratory, Oct. 10, 1957. (pp.
122-126)

142. G. M. ApamsoN and M. L. PicKLESIMER, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Jan. 31, 1958, USAEC Report
ORNL-2493, Oak Ridge National Laboratory, 1958.

143. G. M. Apamson and P. L. RiTTENHOUSE, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Apr. 30, 1957, USAEC Report
ORNL-2331, Oak Ridge National Laboratory, Sept. 3, 1957. (pp. 127-128)

144. Inspection of M aterval—Fluorescent and Dye-Penetrant Method, U. S. Air
Force Technical Order No. 33B1-2-1-2, U. S. Air Force, June 15, 1955.

145. Radiography in Modern Industry. Rochester, N. Y.: Eastman Kodak
Co., 1957.

146. R. B. OLIVER et al., Immersed Ultrasonic Inspection of Pipe and Tubing,
J. Soc. Non-Destructive Testing 15, No. 3, 140-144 (May-June 1957).

147. J. W. ALLEN and R. B. OLIvER, Inspection of Small Diameter Tubing
by Eddy-Current Methods, J. Soc. Non-Destructive Testing 15, No. 2, 104-109
(March—April 1957).

148. J. C. WiLsoN and R. G. BErGGREN, Effects of Neutron Irradiation in
Steel, Am. Soc. Testing Materials 55, 689 (1955).

149. R. G. BERGGREN and J. C. WiLsoN, Recent Data on the Effects of Neutron
Irradiation on Structural Metals in Alloys, USAEC Report CF-56-11-1, Oak
Ridge National Laboratory, 1957.

150. R. G. BERGGREN and J. C. WiLsoN, in Solid State Semiannual Progress
Report for Pervod Ending Aug. 31, 1957, USAEC Report ORNL-2413, Oak
Ridge National Laboratory, 1957. (p. 75)

151. G. E. Moorg, The Solution and Vapor Phase Corrosion of Type 347 Stain-
less Steel, Titantum 75A and Zircolay-2 Exposed to 0.14m Uranyl Sulfate in the
Absence and Presence of Chlorine or Iodine, USAEC Report CF-55-12-70, Oak
Ridge National Laboratory, Dec. 12, 1955.

152. J. C. Griess et al., Solution Corrosion Group Quarterly Report for the
Period Ending July 31, 1957, USAEC Report CF-57-7-121, Oak Ridge National
Laboratory, July 31, 1957.

1563. I. SpiewaKk et al.,, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Oct. 31, 1957, USAEC Report ORNL-2432, Oak
Ridge National Laboratory, Jan. 21, 1958. (p. 15)

154. D. J. DEPAUL, see reference 86.
300 INTEGRITY OF METALS IN HOMOGENEOUS REACTORS [cHAP. 5

155. E. L. CompeErRE and J. L. EncuisH, Oak Ridge National Laboratory,
1954. Unpublished.

156. J. C. GriEss et al., Quarterly Report of the Solution Corrosion Group for
the Period Ending Jan. 31, 1958, USAEC Report CF-58-1-72, Oak Ridge National
Laboratory, Jan. 31, 1958.

157. J. C. Griess et al., Quarterly Report of the Solution Corrosion Group for
the Period Ending Apr. 30, 1957, USAEC Report CF-57-4-55, Oak Ridge Na-
tional Laboratory, Apr. 30, 1957. (p. 30)

158. G. E. MooORE, see reference 151.

159. T. M. KecLEY, Jr., Metallographic Examination of Type-347 Stainless
Steel and Titanium 76A Corrosion Specimens Exposed to Vapor Above Oxygenated
0.14m U02804 Containing Chlorine or Iodine, USAEC Report CF-56-7-56, Oak
Ridge National Laboratory, July 16, 1956.

160. J. C. Griess et al., see Reference 156.

161. J. C. Grigss et al., Quarterly Report of the Solutton Corrosion Group for
the Period Ending Apr. 80, 1957, USAEC Report CF-57-4-55 Oak Ridge National
Laboratory, Apr. 30, 1957. (p. 19)

162. J. C. Griess et al., Quarterly Report of the Solution Corrosion Group for
the Period Ending July 31, 1957, USAEC Report CF-57-7-121 Oak Ridge Na-
tional Laboratory, 1957. (p. 27)

163. E. L. ComPERE et al., Homogeneous Reactor Project Dynamic Slurry
Corrosion Studies: Quarter Ending Jan. 31, 1957, USAEC Report CF-57-1-146,
Oak Ridge National Laboratory, 1957. (p. 23)

" 164. L. MarTI-BALAYNER, Westinghouse Electric Corporation, Commercial
Atomic Power, 1957. Unpublished. -

165.:.W. L. WiLniams and J. F. Ecker, Stress-Corrosion of Austenitic Stain-
less Steels in High-temperature Waters, J. Am. Soc. Naval Engrs. 68(1), 93-103
(1956). |

166. J. C. GriEss et al., see reference 156.

167. L. MARTI-BALAYNER, see reference 164.

168. L. MARTI-BALAYNER, see reference 164.

169. H. A. McLain, Treatment of HRT Steam System Water, USAEC Report
CF-56-11-132, Oak Ridge National Laboratory, Nov. 29, 1956.

170. J. C. Griess et al., Quarterly Report of the Solution Corrosion Group for
the Period Ending Jan. 31, 1957, USAEC Report CF-57-1-144, Oak Ridge Na-
tional Laboratory, 1957. (p. 35)

171. J. C. GriEss et al., Quarterly Report of the Solution Corroston Group for
the Period Ending Oct. 31, 1957, USAEC Report CF-57-10-80, Oak Ridge
National Laboratory, 1957. (p. 31)

172. E. G. BouLmanN and G. M. Apamson, Stress-Corroston Cracking Prob-
lems in the Homogeneous Reactor Test, USAEC Report CF-57-1-143, Oak Ridge
National Laboratory, Jan. 31, 1957.

173. G. M. ApamsoN et al., Metallurgical Examination of HRT Leak Detector
Tubing and Flanges, USAEC Report CF-57-1-109, Oak Ridge National Labora-
tory, Jan. 31, 1957. | |
CHAPTER 6
CHEMICAL PROCESSING*

6—1. INTRODUCTION

One of the principal advantages of fluid fuel reactors is the possibility of
continually processing the fuel and blanket material for the removal of
fission products and other poisons and the recovery of fissionable material
produced. Such continuous processing accomplishes several desirable
objectives: (a) improvement of the neutron economy sufficiently that the
reactor breeds more fissionable material than it consumes, (b) minimiza-
tion of the hazards associated with the operation of the reactor by main-
taining a low concentration of radioactive material in the fuel, and (c¢) im-
provement of the life of equipment and stability of the fuel solution by
removing deleterious fission and corrosion products. The performance
and operability of a homogeneous reactor are considerably more dependent
on the processing cycle than are those of a solid fuel reactor, although the
objectives of processing are similar.

The neutron poisoning in a homogeneous reactor from Whlch fission
product gases are removed continuously is largely due to rare earths [1],
as shown in Fig. 6-1. In Fig. 61 the rare earths contributing to reactor
poisoning are divided into two groups. The time-dependent rare earths
are those of high yield and intermediate cross section, such as NJ!43
and Nd!#5 Prl4! and Pm!47, which over a period of several months could
accumulate in the reactor and result in a poisoning of about 209,. The
constant rare-earth poison fraction is due primarily to Sm!4® and Sm!5!,
which have very large cross sections for neutron absorption but low yield,
and therefore reach their equilibrium level in only a few days’ operation.
Poisoning due to corrosion of the stainless-steel reactor system was cal-
culated for a typical reactor containing 15,500 ft2 of steel corroding at a
rate of 1 mpy. It isassumed that only the nickel and manganese contribute
to the poisoning, since iron and chromium will hydrolyze and precipitate
and be removed from the reactor system; otherwise, corrosion product
poisoning would be four times greater than indicated in Fig. 6-1. The
control of rare earths and corrosion product elements is discussed in sub-
sequent sections of this chapter. Removal of solids from the fuel solution
also improves the performance of the reactor by diminishing the deposition
of scale on heat-transfer surfaces and reducing the possibility of erosion of
pump impellers, bearing surfaces, and valve seats.

*By R. A. McNees, with contributions from W. E. Browning, W. D. Burch,
R. E. Leuze, W. T. McDuffee, and S. Peterson, Oak Ridge National Laboratory.

301
302

CHEMICAL PROCESSING

[cHAP. 6

 

T

. Total Rare Earths
Time Dependent Rare Earths

. Corrosion Products at 1 mpy

. Constant (98.2% Rare Earths
4+ 1.8% Cd)
Sr89 4 Tc99
. Alkali Metals

. VI B(1131)

. Noble Metals
(Rh103)

  
 
 
 
 
  
   

 

4.5
A
B.
4.0 —C
D
3.5
'E:.
3.0 -4
* H
e 2.5 F
2
'S
e 20
1.5
1.0 |-
5 ,f

 

 

60
Irradiation Time , days

 

90

 

150

Fig. 6-1. Poison effect as a function of chemical group in core of two-region

thermal breeder.

Overflow Returned
to Reactor Core

Feed from
Reactor Core

Slurry from
Blanket

    

DO
Recovery

    
 

U and

Fission
Products

Hydroclone

DO

U, Th, Pq,
Fission Products

 

 

Underflow
Containing
Insoluble
Fission
Products

 

—

 

 

 

Decay
Storage
1{60 days

 

 

 

  

  
 

U

Pa

3

Separation

 
 

Fission Products

 

 

 

 

U

 

ecay
Storage
(120 days)

 

 

To Reactor
Core

To Reactor
Blanket
Th
Fission
Solvent Products
Extraction

U

To Reactor
Core

Fig. 6-2. Conceptual flow diagram for processing fuel and blanket material from

a two-region reactor.

The biological hazards associated with a homogeneous reactor are due
chiefly to the radioactive rare earths, alkaline earths, and iodine [2].
The importance, as a biological hazard, of any one of these groups or nu-
clides within the group depends on assumptions made in describing ex-
posure conditions; however, I'3! contributes a major fraction of the radia-
tion hazards for any set of conditions. While the accumulation of hazardous
materials such as rare earths and alkaline earths will be controlled by the
processing methods to be described, less is known about the chemistry of
6-1] INTRODUCTION 303

Blanket
1.4m UO2SO4
in D20 at
250°C

—==Core I_>D20
\ |

D20 e Dissolution
Recovery in HNO3

    
     
    

 

 

  
 

 

 

 

 

  

 

 

 

4

     

 

 

 

 

 

= e vy e Solvent
|
c 1
2 5 1 £
5 £ I g
L ol O -:
o a : &
1
Solvent == Solvent =gy I
== Uranium Fission Plutonium Uranium
s Plutonium Product Waste

Fic. 6-3. Conceptual flow diagram for processing blanket material from a two-
region plutonium producer.

iodine in the fuel systems and methods for removing it. Existing informa-
tion on 1odine processing is discussed in Section 6-5.

Schematic flowsheets for proposed processing schemes for two types of
two-region aqueous homogeneous reactors are shown in Figs. 6-2 and 6-3.
In both cases, solids are removed by hydroclones and concentrated into
a small volume of solution for further processing. The nature of such
processing will be determined by the exact design and purpose of the
reactor. Thus, for a two-region plutonium producer, the core and blanket,
materials would have to be processed separately to avoid isotopic dilution,
while for a thorium breeder, core and blanket material could be processed
together. However, if an attractive method should be developed for leach-
ing uranium and/or protactinium from a thorium-oxide slurry without
~seriously altering the physical properties of the slurry, the two materials
could be processed separately. In a similar way, the relation between
iodine control and fission product gas disposal is such that neither problem
can be disassociated from the other. A specific, complete, and feasible
chemical processing scheme cannot be proposed for any reactor without
an intimate knowledge of all aspects of design and operation of the reactor.
However, some of the basic chemical knowledge needed to evaluate various
304 CHEMICAL PROCESSING [cHAP. 6

possible processing methods has been developed and is presented in the
following sections.

6—2. CoreE PRrocEssING: SorLips REMOVAL

6-2.1 Introduction. Early in the study of the behavior of fission and
corrosion products in uranyl sulfate solutions at temperatures in the range
250 to 325°C, it was found that many of these elements had only a limited
solubility under reactor conditions. Detailed studies of these elements
were conducted and devices for separating solids from liquid at high
temperature and pressure were constructed and evaluated. Based on this
work, a pilot plant to test a processing concept based on solids separation at
reactor temperature was installed as an adjunct to the HRE-2. These
processing developments are discussed in this section.

6-2.2 Chemistry of insoluble fission and corrosion products. Of the
nongaseous fission products, the rare earths contribute the largest amount
of neutron poison to a homogeneous reactor after a short period of operation
(Fig. 6-1). Therefore, a detailed study of the behavior of these elements

TABLE 6-1

SOLUBILITY OF LANTHANUM SULFATE IN
0.02m UO2804—0.005m HoSO4 As A FuNcTION OF
SOoLUTION TEMPERATURE

 

 

 

 

mg La2(S04)s/kg H20
Temperature,
°C True Concentration required to
solubility initiate precipitation
190 250 760
210 130 360
230 54 167
250 25 77
270 12 36

 

 

 

 

 

has been made. All the rare earths and yttrium showed a negative tem-
perature coefficient of solubility in all the solutions studied and a strong
tendency to supersaturate the solutions, as shown in Table 6-1. With the
exception of praseodymium and neodymium, which are reversed, the solu-
bility at a given temperature and uranyl sulfate concentration increased
with increasing atomic number, with yttrium falling between neodymium
6—2] CORE PROCESSING: SOLIDS REMOVAL 305

and samarium, as shown in Table 6-2. Increasing the uranyl sulfate
concentration increased the solubility of a given rare-earth sulfate, as
shown in Table 6-3.

TABLE 6-2

SOLUBILITY OF VARIOUS RARE-EARTH SULFATES IN
0.02 m U02804—0.005 m HoSO4 AT 280°C

 

 

 

 

 

 

 

 

 

Solubility, Solubility,
Salt mg /kg H0 Salt mg kg H0
Lazb(SO4)3 10 Ndz(SO4)3 110
Ce2(SO4)3 50 Y2(S04)3 240
Pr2(S0O4)3 170 Sm2(SO4)3 420
TABLE 6-3

ErfFEcT OF URANYL SULFATE CONCENTRATION ON THE SOLUBILITY OF
NEODYMIUM SULFATE AT VARIOUS TEMPERATURES

 

 

 

 

Nd2(SO4)3 solubility, mg/kg H20
U, g/kg H:0
250°C 280°C 300°C
5.7 270 115 73
10.8 400 200 120
16.6 770 300 180
22 .4 > 1000 500 300

 

 

 

 

 

 

In a mixture of rare-earth sulfates the solubility of an individual rare
earth is less than it would be if it were present alone. For example, the
solubility of praseodymium sulfate at 280°C is 170 mg/kg H20 with no
other rare earths present, as compared with 12 mg/kg H2O in a solution
made up with a rare-earth mixture containing 6% praseodymium sulfate.
Samples of the precipitating salts isolated from solution at 280°C have
usually been the sulfates and contained no uranium. However, under
special conditions a mixed sulfate salt of neodymium and uranium has been
observed [3].

The alkaline earths, barium and strontium, also show a negative tem-
perature coefficient, but not so strongly as do the rare earths; almost no
effect can be seen when the temperature of precipitating solutions is in-
306 CHEMICAL PROCESSING [cHAP. 6

creased from 250 to 300°C. At 295°C in 0.02 m U02804—0.005 m HoSO4
solution, the solubility of barium sulfate is 7 mg/kg H2O and that of
strontium sulfate is 21 mg/kg H2O. Both the alkaline and rare-earth
sulfates show a strong tendency to precipitate on and adhere to steel
surfaces hotter than the precipitating solutions, and this property can be
used to isolate these solids from liquids at high temperatures.

Other fission and corrosion product elements hydrolyze extensively at
250 to 300°C and precipitate as oxides, leaving very low concentrations
in solution. Iron(III) at 285°C has a solubility of 0.5 to 2 mg Fe/kg H20
and chromium(III), 2 to 5 mg/kg H20O. At 285°C less than 5 mg of zir-
conium or niobium per kilogram of H20 remains in solution.

For other elements of variable valence, such as technetium, the amount
of the element in solution is determined by the stable valence state under
reactor conditions. In general, the higher valence states better resist hy-
drolysis and remain in solution. Thus at 275°C in 0.02 m UO2S04 Te(VII)
is reduced to Tc(IV) if hydrogen is present, and only 12 mg/kg H20 re-
mains in solution. However, a slurry of TcO2 in the same solution but with
oxygen present dissolves to give a solution at 275°C with a technetium
concentration of more than 9 g/kg H20. The same qualitative behavior is
observed with ruthenium. Selenium and tellurium in the hexapositive state
are much more soluble than when in the tetrapositive state [4].

A few elements, e.g., cesium, rubidium, nickel, and manganese, intro-
duced into the fuel solution by fission or by corrosion of the system, are
very soluble under reactor conditions. Their removal and control are dis-
cussed in Section 6—4.

6-2.3 Experimental study of hydroclone performance. It is evident
from the preceding section that the amount of uranium withdrawn from
the reactor diminishes if the collection, concentration, and isolation of the
insolubles can be effected at high temperature. One device capable of
collecting and concentrating solids at high temperature is a solid-liquid
cyclone separator called a “hydroclone,” or “clone.” A diagram of a hydro-
clone 1s shown in Fig. 6—4. In operation, a solids-bearing stream of liquid
1s injected tangentially into the wide portion of a conical vessel. Solids
concentrate in a downward-moving layer of liquid and are discharged from
the bottom of the clone into the underflow receiver. Partially clarified
liquid leaves from the top of the clone through a vortex finder. Use of the
underflow receiver eliminates mechanical control of the discharge flow
rate and, by proper choice of hydroclone dimensions, any desired ratio of
overflow rate to underflow rate can be achieved. The driving force for the
system is provided by a mechanical pump. |

The factors influencing the design of an effective hydroclone for homo-
geneous reactor processing use have been studied, and hydroclone designs
6-2] CORE PROCESSING: SOLIDS REMOVAL 307

 

 

 

 

 

 

 

Overflow
N Re—p
y o
‘i
‘|
o NN
Feed — i
A -4 KL
L
Hydroclone\
T A// 77
- g
D
% u : %
Underflow 1
; Port %
¥ /]
. 7
] /e Underflow
1 L/ Receiver
/| 1
4 /
4 /1
/| /
%
/S S S

 

 

 

 

 

 

Fi1c. 6-4. Schematic diagram of a hydroclone with associated underflow receiver.

based on these studies have been tested in the laboratory and on various
circulating loops [5]. All tests have shown conclusively that such hydro-
clones can separate insoluble sulfates or hydrolyzed materials from liquid
streams at 250 to 300°C. In the HRE-2 mockup loop a mixture of the sul-
fates of iron, zirconium, and various rare earths, dissolved in uranyl-
sulfate solution at room temperature, precipitated when injected into the
loop solution at 250 to 300°C. The solids concentrated into the underflow
receiver of a hydroclone contained 759 of the precipitated rare-earth
sulfates. When the lanthanum-sulfate solubility in the loop solution was
exceeded by 109, the concentration of rare earths in the underflow receiver
was four to six times greater than in the rest of the loop system; some
accumulation of rare earths was observed in the loop heater. A large
fraction of the hydrolyzed iron and zirconium was collected in the gas
separator portion of the loop. In the separator the centrifugal motion
given to the liquid forced solids to the periphery of the pipe and allowed
them to accumulate. Only about 10% of the solids formed in the loop was
recovered by the hydroclone, and examination of the loop system dis-
closed large quantities of solids settled in every horizontal run of pipe.
308 CHEMICAL PROCESSING [cHAP. 6

 

 
 
 
 
 
 
 
 
  

 

 

 

 

 

//
Recombinei Recombinei
% -Condenser -Condenser
Sampler
// I Hydroclone
7 H20 to Waste
D2Q Receiver
10 gdl
| ot L 7
Regecl'r'or Separator Separator [j
Decay Tank (2) Carrier

100 gdl
\

 

  
   

 

 

 

  
 

Solution

 

 

 

 

1 D,O Addition Sanoler
7 - ! s+ Fuel Addition P
‘ Transfer Tcmkl___l

F1c. 6-5. Schematic flow diagram for the HRE-2 chemical processing plant.

 

 

 

 

 

 

TABLE 6—4

DimeENnsioNs o HRE-2 HyprRoCLONES

 

 

 

 

Dimension, in.
Symbol Location i - )
0.25-1n. 0.40-1n. 0.56-1n.
hydroclone hydroclone hydroclone
Dy, Maximum inside 0.25 0.40 0.56
diameter
L Inside length 1.50 2.40 3.20
Dy Underflow port
diameter 0.070 0.100 0.148
Do Overflow port
diameter 0.053 0.100 0.140
Dp Feed port effec-
tive diameter 0.051 0.118 0.159

 

 

 

 

 

 

 
6-2] CORE PROCESSING: SOLIDS REMOVAL 309

Samples taken from the loop after addition of preformed solids and without
the hydroclone operating showed an exponential decrease in solids concen-
tration with a half-time of 2.5 hr; with the hydroclone operating, the half-
time was 1.2 hr. In the HRE-2 chemical plant [5], operated with an aux-
iliary loop to provide a slurry of preformed solids in uranyl sulfate solution
as a feed for the plant, the half-times for solids disappearance and removal
were 11 hr without the hydroclone and 1.5 hr with it. The efficiency of the
hydroclone for separating the particular solids used in these experiments
was about 109%. With gross amounts of solids in the system, concentration
factors have been as large as 1700. -

Correlation of these data with anticipated reactor chemical plant oper-
ating conditions indicates that the HRE-2 chemical plant will hold the
amount of solids in the fuel solution to between 10 and 100 ppm. If neces-
sary, performance can be improved by increasing the flow through the
chemical plant and by eliminating, wherever possible, long runs of hori-
zontal pipe with low liquid Velomty and other stagnant areas which serve
to accumulate solids. ~

6-2.4 HRE-2 chemical processing plant.* An experimental facility to
test the solids-removal processing concept has been constructed in a cell
adjacent to the HRE-2. A schematic flowsheet for this facility is shown
in Fig. 6-5.

A 0.75-gpm bypass stream from the reactor fuel system at 280°C and
1700 psi is circulated through the hlgh—pressure system, consisting of a
heater to make up heat losses, a screen to protect the hydroclone from
plugging, the hydroclene with underflow receiver, and a canned-rotor
circulating pump to make up pressure losses across the system. The
hydroclone is operated with an underflow receiver which is drained after
each week of operation, at which time the processing plant is isolated
from the reactor system.

At the conclusion of each operatmg period 10 liters of the slurry in the
underflow pot is removed and sampled. The heavy water is evaporated
and recovered, and the solids are dissolved in sulfuric acid and sampled
again. The solution is then transferred under pressure to one of two 100-gal
decay storage tanks. Following a three-month decay period, the solution
1s transferred to a shielded carrier outside the cell and then to an existing
solvent extraction plant at Oak Ridge National Laboratory for uranium
decontamination and recovery. The sulfuric acid solution step is incor-
porated in the HRE-2 chemical plant to ensure obtaining a satisfactory
sample. This step would presumably not be necessary in a large-scale
plant.

*Contribution from W. D. Burch.
310 CHEMICAL PROCESSING [cHAP. 6

 

Fic. 6-6. HRE-2 chemical plant cell with equipment.
6-2] CORE PROCESSING: SOLIDS REMOVAL 311

All equipment is located in a 12- by 24- by 21-ft underground cell located
adjacent to the reactor cell and separated from it by 4 ft of high-density
concrete. Other construction features are similar to those of the reactor
cell, with provisions for flooding the cell during maintenance periods in
order to use water as shielding. Figure 66, a photograph of the cell prior
to installation of the roof plugs, shows the maze of piping necessitated by
the experimental nature of this plant.

Dimensions of the three sizes of hydroclones designed for testing in this
plant are shown in Table 6—4. These three hydroclones, which have been

Blind Flange B
Closure ‘l -
ol
i

L

.
.
.
L

Pressure Screw

 

 

AN Y <
/ N\ O\ Y N\ \
N \; \\ N
N\ \ N N N \
\ N\ N\ NN
N N\ O\ Y
N N N
N\ N X
\N T S S e A
N
N\
N \ . For
.
N
\

N\ Leak Detector

 

 

   
 
 

“K / N\ /Overflow
\ R N 222272 — —— - [/

Hydroclone
Retaining Plug

 

   
 

N A

    

 

B % \\\
. ’ oS //\
Gold Wire Gasket ~ s‘é g.
1
.
7 .Hydroclone

Fic. 6-7. HRE-2 chemical plant hydroclone container.

selected to handle the range of possible particle sizes, are interchangeable
at any time during radioactive operation through a unique, specially ma-
chined flange, shown in Fig. 6-7. Removal of the blind closure flange ex-
poses a cap plug and retainer plug. Removal of these with long-handled
socket wrenches permits access to the hydroclone itself. This operation has
been performed routinely during testing with nonradioactive solutions.

In processing homogeneous reactor fuel, a transition from a heavy- to a
natural-water system is desirable if final processing is to be performed in
 conventional solvent extraction equipment. Such a transition must be ac-
complished with a minimum loss of D20 and a minimum contamination of
312 CHEMICAL PROCESSING [cHAP. 6

the fuel solution by H20 in recycled fuel. Initial tests of this step in the
fuel processing cycle have been carried out [6]. In these experiments a
mixture of 59, D20, 959, H20 was used to simulate reactor fuel liquid.
The dissolver system was cycled three times between this liquid and or-
dinary water, with samples being taken during each portion of each cycle.
Isotopic analysis of these samples showed no dilution of the D20 in the
enriched solution and no loss of D20 to the ordinary water system.

At expected corrosion rates, approximately 400 g of corrosion products
will be formed in the reactor system per week, and the underflow receiver
was therefore designed to handle this quantity of solids. The adequacy of
the design was shown when more than three times this quantity of solids
was charged to the underflow receiver and drained in the normal way with-
out difficulty.

Full-scale dissolution procedures have also been tested [6]. To minimize
the possibilities of contaminating the reactor fuel solution by foreign ions,
a dissolution procedure was developed using only sulfuric acid. This con-
sists of a 4-hr reflux with 10.8 M H2SO4 1n a tantalum-lined dissolver fol-
lowed by a 4-hr reflux with 4 M H2SO4, and repeated as required until
dissolution is complete. Decay storage tanks and other equipment required
to handle the boiling 4 M H2SO4 are fabricated of Carpenter—20 stainless
steel. Tests have repeatedly demonstrated more than 99.59, dissolution
of simulated corrosion and fission products in two such cycles.

The HRE-2 hydroclone system has been operated as an integral part
of the reactor system for approximately 600 hr and for an additional
1200 hr with a temporary pump loop during initial solids-removal tests.
During this operating period, in which simulated nonradioactive fuel
solutions were used, the performance of the plant was satisfactory in all
respects.

6-3. FissioN Propuct Gas DisPosAL*

6-3.1 Introduction. To prevent the pollution of the atmosphere by
radioactive krypton and xenon isotopes released from the fuel solution, a
system of containment must be provided until radioactive decay has re-
duced their activity level. This is accomplished by a method based on the
process of physical adsorption on solid adsorber materials. If the adsorber
system is adequately designed, the issuing gas stream will be composed of
long-lived Kr®, oxygen, inert krypton isotopes, inert xenon isotopes, and
insignificant amounts of other radioactive krypton and xenon isotopes. In
case the activity of the Kr33 is too high for dilution with air and discharge
to the atmosphere, the mixture may be stored after removal of the oxygen

*Contribution from W. E. Browning.
6-3] FISSION PRODUCT GAS DISPOSAL 313

 

Or—T—T717 T 17 T T T T T T T

 

Krypton Adsorbed, mg/g of Adsorbent

1.0 3\35 S —
- aneo —_—
0.5 -

 

l/// Decalso :

o GeV-70
silica
0.2 / ‘ wes-AA L —

Zeo-Dur Mo\ecu\d" >

Micro Cel-A

0.1 L L1 011

O 10 20 30 40 50 60 70 80 90 100 110 120 130
Krypton Pressure, mm

 

Fig. 6-8. Adsorption of krypton on various adsorbents at 28°C.

or further separated by conventional methods into an inert xenon fraction
and a fraction containing Kr®> and inert krypton.

6-3.2 Experimental study of adsorption of fission product gases. Evalu-
ation of various adsorber materials based on experimental measurements of
the equilibrium adsorption of krypton or xenon under static conditions is
in progress [7]. Results in the form of adsorption isotherms of various
solid adsorber materials are presented in Fig. 6-8.

A radioactive-tracer technique was developed to study the adsorption
efficiency (holdup time) of small, dynamie, laboratory-scale adsorber
systems [8]. This consists of sweeping a brief pulse of Kr8> through an ex-
perimental adsorber system with a diluent gas such as oxygen or nitrogen
314 CHEMICAL PROCESSING [cHAP. 6
and monitoring the effluent gases for Kr85 beta activity. The activity in
the gas stream versus time after injection of the pulse of Kr3? is recorded.
A plot of the data gives an experimental elution curve, such as shown in
Fig. 69, from which various properties of an adsorber material and ad-
sorber system may be evaluated.

Among the factors which influence the adsorption of fission product
gases from a dynamic system are (1) adsorptive capacity of adsorber ma-
terial, (2) temperature of adsorber material, (3) volume flow rate of gas
stream, (4) adsorbed moisture content of adsorber material, (5) composi-
tion and moisture content of gas stream, (6) geometry of adsorber system,
and (7) particle size of adsorber material. The average time required for
the fission product gas to pass through an adsorber system, imax, 1s influ-
enced by the first five of the above factors. The shapeof the experlmental
elution curve is affected by the last two.

The temperature of the adsorber material is of prime importance. The
lower the temperature the greater will be the adsorption of the fission gases,
and therefore longer holdup times per unit mass of adsorber material will
result. The dependence of adsorptive capacity, k, on temperature as de-
termined by holdup tests with some solid adsorber materials is shown in
Table 6-5.

TABLE 6-5

ApsORPTIVE CAPACITY OF VARIOUS MATERIALS AS A FuNcTION
OF TEMPERATURE

 

 

 

 

 

 

 

 

 

cc gas/g adsorbent™
Gas Diluent Adsorber
| 273°K 323°K 373°K
Xe 02 Charcoal 4.7 X 103 4.0 X 102 80.0
Kr He Charcoal 1.8 X 102 34 9.6
Kr Os or N Charcoal, 68 24 11.0
Kr 0. Linde Molecular 23 9 4.5
Sieve HA
Kr 02 Linde Molecular 11 5.7 3.5
Sieve 10X

 

 

*(Gas volume measured at temperatures indicated.

 
6-3] FISSION PRODUCT GAS DISPOSAL 315

 

l l | | |
Average Holdup Time (t,,) _l

 
   

Kr85 Activity in Effluent Gas Stream —>

|
Breakthrough Time (t},)

      

 

 

| |

Time After Injection of Kr85 —»

I

 

F1c. 6-9. Experimental Kr85 elution curve.

At a given temperature, the average holdup time, ¢y.y, is inversely pro-
portional to the volume flow rate of the gas stream. If the volume flow
rate 1s doubled, the holdup time will be decreased by a factor of two.

All the solid adsorber materials adsorb moisture to some degree. Any
adsorbed moisture reduces the active surface area available to the fission
gases and thus reduces the average holdup time.

The geometry of the adsorber system influences the relation between
breakthrough time, ;, and average holdup time, fax, as shown in Fig. 6-9.
Ideally, for fission product gas disposal, a particular atom of fission gas
should not emerge from the adsorber system prior to the time #yax. Since
this condition cannot be realized in practice, the difference between break-
through and average holdup times should be made as small as possible.
For a given mass of adsorber material a system composed of long, small-
diameter pipes will have a small difference between &, and tnax, whereas a
system composed of short, large-diameter pipes will not.

The particle size of the adsorber material is important for ensuring in-
timate contact between the active surface of the adsorber material and
the fission gases. A system filled with large particles will allow some mole-
316 CHEMICAL PROCESSING [cHAP. 6

cules of fission gases to penetrate deeper into the system before contact is
made with an active surface, while the pressure drop across a long trap
filled with small particles may be excessive. Material between 8 and 14
mesh in size is satisfactory from both viewpoints.

6-3.3 Design of a fission product gas adsorber system. The design of
an adsorber system will be determined partly by the final disposition of
the effluent gas mixture. If ultimate disposal is to be to the atmosphere,
the adsorber system should be designed to discharge only Kr3* plus inert
krypton and xenon isotopes. If the effluent gases are to be contained and
stored, the adsorber system may be designed to allow discharge of other
radioactive krypton and xenon isotopes. In the following discussion it is
assumed that final disposal of the effluent gas mixture will be to the at-
mosphere. The following simple relation has been developed which is use-
ful in finding the mass of adsorber material in such an adsorber system:

F
M —_ Etmax,

where M = mass of adsorber material (grams), F = gas volume flow rate
through adsorber system (cc/min), k= adsorptive capacity under dy-
namic conditions (cc/g), and tmax = average holdup time (min).

The operating characteristics of the reactor will dictate the composition
and volume flow rate of the gas stream; tmax Will be determined by the al-
lowable concentration of radioactivity in the effluent gas; k values for
krypton and xenon must be determined experimentally under conditions
simulating these in the full-scale adsorber system. It should be noted
(Fig. 6-9) that a portion of the fission gas will emerge from the adsorber
system at a time # prior to the average holdup time, imax. The design
should ensure that radioactive gas emerging at time & has decayed suffi-
ciently that only insignificant amounts of activity other than Kr®°> will be
discharged from the bed.

The adsorber system should be operated at the lowest convenient
temperature because of the dependence of adsorptive capacity on tempera-
ture. Beta decay of the fission product gases while passing through the
adsorber system will increase the temperature of the adsorber material
and reduce the adsorptive capacity. Temperature control is especially
critical if the adsorber system uses a combustible adsorber material, such
as activated charcoal, with oxygen as the diluent or sweep gas.

6-3.4 HRE-2 fission product gas adsorber system. The HRE-2 uses a
fission product gas adsorber system containing Columbia G activated
charcoal. Oxygen, contaminated with the fission product gases, is removed
6-4] CORE PROCESSING: SOLUBLES 317

from the reactor, dried, and passed into this system, and the effluent gases
are dispersed into the atmosphere through a stack.

The adsorber system contains two activated charcoal-filled units con-
nected in parallel to the gas line from the reactor. Each unit consists of
40 ft of -in. pipe, 40 ft of 1-in. pipe, 40 ft of 2-in. pipe, and 60 ft of 6-in.
pipe connected in series. The entire system is contained in a water-filled
pit, which is buried underground for gamma shielding purposes. FEach
unit is filled with approximately 520 lb of Columbia G activated charcoal,
8 to 14 mesh.

The heat due to beta decay of the short-lived krypton and xenon iso-
topes is diminished by an empty holdup volume composed of 160 ft of
3-1n. pipe between the reactor and the charcoal adsorber system. This pre-
vents the temperature of the charcoal in the inlet sections of the adsorber
system from exceeding 100°C. Excessive oxidation of the charcoal by the
oxygen in the gas is further prevented by water-cooling the beds.

Before the adsorber system was placed in service, its efficiency was
tested under simulated operating conditions [9]. A pulse of Kr% (25
millicuries) was injected into each unit of the adsorber system and swept
through with a measured flow of oxygen. In this way the krypton holdup
time was determined to be 30 days at an oxygen flow rate of 250 cc/min/
unit. Based on laboratory data from small adsorber systems, the holdup
time for xenon is larger than that for krypton by a factor that varies from
30 to 7 over the temperature range of 20 to 100°C. From these data, it is
~ estimated that the maximum temperature of the HRE-2 adsorber system
will vary between 20 and 98°C after the reactor has been operating at
10 Mw power level long enough for the gas composition and charcoal
temperature to have reached equilibrium through the entire length of the
adsorber unit. The holdup performance of the adsorber system was cal-
culated with corrections for the increased temperature expected from the
fission gases. The calculated holdup time was found to be 23 days for
krypton and 700 days for xenon; this would permit essentially no Xe133
to escape from the trap.

6—4. CorRE PROCESSING: SOLUBLES

6—4.1 Introduction. While the solids-removal scheme discussed in Sec-
tion 6-1 will limit the amount of solids circulating through the reactor sys-
tem, soluble elements will build up in the fuel solution. Nickel and man-
ganese from the corrosion of stainless steel and fission-produced cesium
will not precipitate from fuel solution under reactor conditions until con-
centrations have been reached which would result in fuel instability and
loss of uranium by coprecipitation. Loss of neutrons to these poisons
would seriously decrease the probability of the reactor producing more
318 CHEMICAL PROCESSING [cHAP. 6

fuel than it consumes. Therefore, a volume of fuel solution sufficient to
process the core solution of the reactor at a desired rate for removal of
soluble materials is discharged along with the insoluble materials concen-
trated into the hydroclone underflow pot. This rate of removal of soluble
materials depends on the nature of other chemical processing being done
and on the extent'of corrosion. For example, operation of an iodine re-
moval plant (Section 6-5) reduces the buildup of cesium in the fuel to an
insignificant value by removing .cesium precursors.

6—4.2 Solvent extraction. Processing of the core solution of a homo-
geneous reactor by solvent extraction is the only method presently avail-
able which has been thoroughly proved in practice. However, the amount
of uranium to be processed daily is so small that operation of a solvent ex-
traction plant just for core solution processing would be unduly expensive.
Therefore, the core solution would normally be combined with blanket ma-
terial from a thermal breeder reactor and be processed through a Thorex
plant, but with a plutonium-producing reactor separate processing of core
and blanket materials will be needed. These process schemes are discussed
in detail in Sections 6—6 and 6-7.

The uranium product from either process would certainly be satisfactory
for return to the reactor. .Since solid fuel element refabrication is not a
problem with homogeneous reactors, decontamination factors of 10 to 100
from various nuclides are adequate and some simplification of present
solvent extraction schemes may be possible.

6—4.3 Uranyl peroxide precipitation. A process for decontaminating the
uranium for quick return to a reactor has been proposed as a means of
reducing core processing costs. A conceptual flowsheet of this process,
which depends on the insolubility of UO4 under controlled conditions for
the desired separation from fission and corrosion products, is shown in
Fig. 6-10. A prerequisite for use of this scheme is that losses due to the
insoluble uranium contained in the solids concentrated in the hydroclone
plant be small. However, laboratory data obtained with synthetic solids
simulating those expected from reactor operation indicate that the uranium
content, of the solids will be less than 19, by weight. Verification of the
results will be sought during operation of the HRE-2.

In the proposed method, the hydroclone system is periodically isolated
from the reactor and allowed to cool to 100°C. The hydrolyzed solids re-
main as such, but the rare-earth sulfate solids concentrated in the under-
flow pot redissolve upon cooling. The contents of the underflow pot are
discharged to a centrifuge where solids are separated from the uranium-
containing solution and washed with D20, the suspension being sent to a
waste evaporator for recovery of D20.
6-5] CORE PROCESSING: IODINE 319

 

 

 

 

 

 

 

 

D90 D209 D20 . D250y
sU o
SFC SU o . o
RE 100°C e SFc—|  30°C  |_fd 30°C UO,—pi 100°C
1sU RE RE - _;
IsFC
UO,SO,
IsU SFC
isFC R

020

}

- DoO Recovery

 

 

 

 

 

 

 

 

 

 

SU = Soluble Uranium
SFC = Soluble Fission ~—H,0
and Corrosion Products
RE = Rare Earth Sulfates sy
IsU = Insoluble Uranium e IS FC — To Waste Storage
IsSFC = Insoluble Fission SFC
and Corrosion Products RE

F1g. 6-10. Schematic flow diagram for decontaminating uranium by uranyl
peroxide precipitation.

Uranium in the clarified solution is precipitated by the addition of either
D202 or Naz02. By controlling pD and precipitation conditions, a fast
settling precipitate can be obtained with less than 0.19] of the uranium
remaining in solution. The UOy4 precipitate is centrifuged or filtered and
washed with DO and dissolved in 509, excess of D2SO4 at 80°C before
being returned to the reactor. | -

In laboratory studies uranium losses have been consistently less than
0.1% for this method and decontamination factors from rare earths greater
than 10. Decontamination factors from nickel and cesium have been 600
and 40, respectively. It is estimated that the product returned to the re-
actor would contain about 20 ppm of sodium as the only contaminant in-
troduced during processing. Although either the addition of D202 or use
of D202 generated by radiation from the solution itself appears attractive,
acid liberated by the precipitation of UO4 must be neutralized if uranium
losses are to be minimized. Since the entire operation is done in a D30
system, no special precautions to avoid contaminating the reactor with
ordinary water are needed.

6-5. CorE ProcessinGg: Iopine*

6-5.1 Introduction. The removal of iodine from the fuel solution of a
homogeneous reactor is desirable from the standpoint of minimizing the
biological hazard and neutron poisoning due to iodine and reducing the
production of gaseous xenon and its associated problems. Iodine will also

*Contribution from S. Peterson.
320 CHEMICAL PROCESSING [cHAP. 6

poison platinum catalysts [10] used for radiolytic gas recombination in
the reactor low-pressure system and may catalyze the corrosion of metals
by the fuel solution. For this reason a considerable effort has been carried
out at ORNL and by Vitro [11] to investigate the behavior of iodine in
solution and to develop methods for its removal. In this regard, the iodine
ijsotopes of primary interest are 8-day I'3! and 6.7-hr 1135,

6-5.2 The chemistry of iodine in aqueous solutions. Iodine in aqueous
solution at 25°C can exist in several oxidation states. The stable species
are iodide ion, I~ ; elemental 1odine, I2; 1odate, IO3™ ; and periodate, 104~
or H5IOg. The last of these exists only under very strongly oxidizing con-
ditions, and is immediately reduced under the conditions expected for a
homogeneous reactor fuel. Iodide ion can be formed from reduction of
other states by metals, such as stainless steel, but in the presence of the
oxygen necessary in a reactor system it is readily converted to elemental
iodine; this conversion is especially rapid above 200°C. Thus the only
states of concern in reactor fuel solutions are elemental iodine and iodate.
Under the conditions found in a high-pressure fuel system the iodine is
largely, if not all, in the elemental form.

Volatility of todine. Since the volatile elemental state of iodine 1is pre-
dominant under reactor conditions, the volatility of iodine from fuel solu-
tion is the basis for proposed iodine-removal processes. The vapor-liquid
distribution coefficient [11] (ratio of mole fraction of iodine in vapor to
that in liquid) for simulated fuel solution and for water at the temperatures
expected for both the high-pressure and low-pressure systems of homo-
geneous reactors is given in Table 6-6.

TABLE 6-6

VArPor-LiQuip DISTRIBUTION OF lODINE

 

 

 

 

Distribution coefficient,
vapor/liquid
Solution
High pressure Low pressure
(260-330°C) (100°C)

Clean fuel solution 7.4 0.34

(0.02 m UOzSO4—0.005 m HzSO4——

0.005 m CuSOy4, 1-100 ppm I,
Fuel solution with mixed fission and

corrosion products 2.4
Water (pH 4 to 8, 1-13 ppm I3) 0.29 0.009

 

 

 

 

 
6-5] | CORE PROCESSING: IODINE 321

 

 

 

 

 

 

 

 

 

 

 

r -------- ] I ------------ l
| | 1
| | |
l Adsorb d | |
' sorber Be ' \ l
. l | Silvered Alundum '
| g l
| !
~
>N 1
~
~
~
~
~

‘4 Ei e e a e = -

o jector o

o o Gas

Separator
enemnes (Gas | _J

 

 

Heater

e,

. Sampler

lodine Spike —6

F1a. 6-11. Vitro iodine test loop.

 

Canned Rotor
Pump

 

A number of conclusions are evident from these data. Iodine is much
more volatile from fuel solution than from water at either temperature.
Fission and corrosion products appear to increase the volatility of iodine
from fuel solution at 100°C. Increasing the temperature from 100 to 200°C
increases the volatility of iodine relative to that of water. No systematic
variation of iodine volatility has been found with iodine concentration in
the range 1 to 100 ppm or temperature in the range 260 to 330°C.

The volatility of iodine from simulated fuel solution has been verified
by experiments in a high-pressure loop, shown schematically in Fig. 6-11
[11]. The circulating solution was contacted with oxygen in the ejector;
the separated gas was stripped of iodine by passing through a bed of sil-
vered alundum which was superheated to prevent steam condensation.
Potassium 1odide solution (containing a radioactive tracer, I!3!) was
rapidly injected into the loop to give an iodine concentration of 10 ppm.
The iodine concentration decreased exponentially with time in the circu-
lating solution. Table 6-7 gives the half-times for iodine removal and the
volatility distribution coefficient, calculated from the removal rate and the
flow rates, based on three experiments with clean fuel solution and two
with added iron. Within the accuracy of flow rate measurement, the coef-
322 CHEMICAL PROCESSING [cHAP. 6

" TABLE 6-7

IopiNE REMovAL FrROM A HicH-PRESSURE Loop

 

 

 

 

 

 

Iodine Iodine
Tempera- et s
: removal | distribution
Solution ture, : :
o half-time, | coeflicient,
C : .
min vapor/liquid
Clean fuel solution 230 13.0 7.6
(0.02 m UO2504—0.005 m H2S04— 6.5 16.8
0.005 m CuSOy) 13.0 8.4
Fuel + 30 ppm Fe3™ 220 11.0 10.9
Fuel + 300 ppm Fe3* 225 11.0 9.5

 

 

 

ficient agrees with the average value of 7.4 obtained in numerous static
tests over the high-temperature range. Iron appears to have no eftect.

Oxidation state of odine at high temperatures and pressures. While iodate
ion is quite stable at room temperature, at elevated temperatures it de-
composes according to the equilibrium reaction

4103_ + 4HT z___)' 215 + 509 + 2H20.

The extent of this decomposition in uranyl-sulfate solutions above 200°C
is not known with certainty, since all observations have been made on
samples that have been withdrawn from the system, cooled, and reduced
in pressure before analysis. Although the iodine in-such samples is prin-
cipally elemental, some iodate is always present, possibly because of re-
versal of the iodate decomposition as the temperature drops in the sample
line. Such measurements therefore give an upper limit to the iodate con-
tent of the solution. If periodate is introduced into uranyl-sulfate solution.
at elevated temperatures, it is reduced before a sample can be taken to
detect its presence. Iodide similarly disappears if an overpressure of oxy-
gen is present, although iodide to the extent of 409, of the total iodine has
been found in the absence of added oxygen [11].

Methods that have been used for determining the iodine/iodate ratio in
fuel solutions are (a) analysis of samples taken from an autoclave at 250°C
at measured intervals after injection of iodine in various states [11],
(b) analysis of samples taken from the liquid in liquid-vapor equilibrium
studies at 260 to 330°C [11], (c¢) rapid sampling from static bombs at
250 to 300°C [12], and (d) continuous injection of iodate-containing fuel
solution into the above described ejector loop at 220°C and determining
6—5] CORE PROCESSING: IODINE 323

oxidation states in samples withdrawn [11]. The iodine/iodate ratio in
these samples has varied from slightly over 1 to about 70, with no apparent
relation to variations in temperature, oxygen pressure, and total iodine
concentration. o

The strongest indication of iodate instability was in the loop experi-
ments, which gave the highest observed iodine/iodate ratio, even though
iodine was continuously introduced into the flowing stream as iodate-
and removed by oxygen scrubbing as elemental iodine. The low iodate
content of the samples from these experiments corresponded to a first-
order iodate decomposition rate constant of 6.2 min~—!. Jodate con-
tents averaging about 10% of the total iodine have been observed in -
0.04 m U02804—0.005 m CuS04—H 2S04 solution, rapidly sampled from
a static bomb through an ice-cooled titanium sample line. The observed
iodate content was unrelated to whether the free sulfuric acid concentra-
tion was 0.02 or 0.03 m, whether the temperature was 250 or 300°C, and
whether or not the solution was exposed to cobalt gamma radiation at an
intensity of 1.7 watts/kg.

Ozidation state of iodine at low temperatures. At 100°C the iodate de-
composition and iodine oxidation are too slow for equilibrium to be es-
tablished in reasonable periods of time. Thus both states can persist under
similar conditions. In stainless-steel equipment both states are reduced to
lodide, which is oxidized to iodine if oxygen or iodate is present [12].

In a radiation field the iodide is oxidized, iodine is oxidized if sufficient
oxygen is present, and iodate is reduced [13]. At the start of irradiation,
iodate is reduced, but in the presence of sufficient oxygen, iodine is later
reoxidized to iodate, probably by radiation-produced hydrogen peroxide
which accumulates in the solution. I'inally, a steady state is reached with a
proportion of iodate to total iodine which is independent of total iodine con-
centration from 1076 to 10~ 5 m and temperatures from 100 to 110°C, but
strongly-dependent on uranium and acid concentrations and on the hydro-
gen/oxygen ratio in the gas phase. When the temperature is increased to
120°C there is a marked decrease in iodate stability under all conditions of
gas and solution composition. Experimental data on the effects of radiation
intensity, temperature, and gas composition for the irradiation of a typical
fuel solution containing 0.04 m U02804—0.01 m H2804—0.005 m CuSOy4
are given in Ref. 13. The steady-state iodate percentages are also given in
this reference.

6-5.3 Removal of iodine from aqueous homogeneous reactors. It is
clear that under the operating conditions of a power reactor, iodine in the
the fuel solution is mainly in the volatile elemental state. It can therefore
be removed by sweeping it from the solution into a gas phase, stripping
324 CHEMICAL PROCESSING [cHAP. 6

it from the gas stream by trapping it in a solid absorber or by contacting the
gas with a liquid.

Numerous experiments have shown that silver supported on alundum is
a very effective reagent for removing iodine from gas or vapor systems,
although its efficiency is considerably reduced at temperatures below
150°C. Silver-plated Yorkmesh packing is very effective for removing
iodine from vapor streams in the range 100 to 120°C. In one in-pile ex-
periment [14] 90% of the fission-product iodine was concentrated in a
silvered-alundum pellet suspended in the vapor above a uranyl-sulfate
solution. This method of using a solid iodine absorber, however, would
present difficult engineering problems, since xenon resulting from iodine
decay would be expected to leave the absorber and return to the core unless
the absorbers were isolated after short periods of use and remotely replaced.

Iodine removal by gas stripping requires a continuous fuel letdown. In
case this is not desirable, the vapor can be stripped of iodine in the high-
pressure system by contacting with a small volume of liquid which is sub-
sequently discharged. Liquids considered include water and aqueous
solutions of alkali, sodium sulfite, or silver sulfate [11]. Although the so-
lutions are much more effective iodine strippers than pure water, their use
requires elaborate provision for preventing entrainment in the gas and sub-
sequent contamination of the fuel solution. Thus most of the effort in
design of iodine-removal systems is based on stripping by pure heavy
water. |

One possible iodine-removal scheme uses Oz or Oz + D2 stripping [15].
The iodine is scrubbed from the fuel solution by the gas in one contactor
and then stripped from the gas by heavy water in a second contactor. This
water would then be let down to low pressure and stored for decay or proc-
essed to remove iodine.

In most homogeneous reactors some of the fuel solution is evaporated
to provide condensate for purge of the circulating pump and pressurizer.
Since iodine is stripped from the fuel by this evaporation this operation can
be used for iodine removal. This method, which is illustrated in Fig. 6-12,
has been proposed for the HRE-3 [16]. Here a stream of the fuel solution
is secrubbed with oxygen in the pressurizer. The steam is condensed and the
oxygen recycled. The condensate is distilled to concentrate the iodine into
such a small volume that its letdown does not complicate reactor operation.

Todine removal in the HRE-2. lodine adsorption on the platinized alu-
mina recombination catalyst, such as that used in the HRE-2, poisons the
catalyst severely [10]. Although the catalyst can be restored by operation
at 650°C, this would not be feasible in HRE-2 operation. A method for
removing iodine from the gas stream by contact with alundum or York-
mesh coated with silver was developed in the HRT mockup. Iodine was
introduced into the system and vapor from the letdown stream and dump
6-5] CORE PROCESSING: IODINE - 325

Oxygen

2ft 3/min
- ——————

  
   

Condenser

 

 

7.53 Ib/min

   
 

Condenser

Pressurizer
Holdup Tank

Still | — To High Pressure
D,OStorage 7.46 Ib/min

100kw |3
Core Solution
20 gal/min 536°F To Low Pressure

To Reactor Core 596°F System 0.07 Ib/min

 

 

 

 

 

Fic. 6-12. Iodine removal system proposed for HRE-3.

tank was passed through a silvered alundum bed and the recombiner, and
then to a condenser. Condensate was returned to the high-pressure loop
through a pressurizer and the circulating pump. After injection, the iodine
concentration of the high-pressure loop dropped from 1.8 mg/liter to
0.1 mg/liter in 2 hr. In similar experiments with silvered Yorkmesh, iodine
levels in the condensate and pressurizer were even lower relative to the
high-pressure loop. The Yorkmesh efficiency depended strongly on how
densely it was packed. The iodine removal efficiencies calculated from
these experiments and others are given in Table 6-8. In laboratory ex-
periments with a 1-in.-diameter bed which could not be tightly packed,
Yorkmesh efficiencies were consistently poorer than those of silvered
alundum.

The ability of a bed of silver-plated Yorkmesh to remove iodine from the
reactor system was apparently confirmed during the initial operating
period of the HRE-2 [17]. Here the iodine activity in the reactor fuel
appeared to be even lower than expected when iodine was removed at
the same fractional rate as fuel solution was let down from the high-
pressure system. Less than 3% of the iodine produced during 40 Mwh of
operation was found in the fuel solution. Experience with the HRT
mockup indicates that the iodine not in solution was held on the silvered
bed.
326 CHEMICAL PROCESSING [cHAP. 6

TABLE 6-8

IopiNnE REmovAL ErrFiciENcy orF SiLvErReED Beps 1N HRT Mockup

 

 

 

Absorber Bed .helght, Temp:rature, Efficiency,
in. C Yo
Silvered alundum 8 150 97.7
rings 8 120 81.0
5 110 64.0
Yorkmesh, 22 1b/ft3 10 120 97.0
Yorkmesh, 29 1b/ft3 6 120 99.6

 

 

 

 

 

 

6—6. URANYL SULFATE BLANKET PROCESSING*

6-6.1 Introduction. The uranyl sulfate blanket solution of a plutonium
producer is processed to remove plutonium and to control the neutron
poisoning by corrosion and fission products. Although a modified Purex
solvent extraction process can be used for plutonium removal, the method
shown schematically in Fig. 6-3, based on the low solubility of plutonium
in uranyl sulfate solution at 250°C, appears more attractive. A hydroclone
similar to that used for reactor core processing is used to produce a con-
centrated suspension of PuOg along with solid corrosion and fission prod-
ucts. The small volume of blanket solution carrying the plutonium is
evaporated to recover the heavy water and the solids are dissolved in
nitric acid. After storage to allow Np?3° to decay, plutonium is decon-
taminated by solvent extraction.

~ 6-6.2 Plutonium chemistry in uranyl sulfate solution. The amount of

plutonium remaining dissolved in 1.4 m UO2S04 at 250°C is dependent
on a number of variables, including solution acidity, plutonium valence,
and initial plutonium concentration. Under properly controlled condi-
tions, less than 3 mg/kg H20 has been obtained. Since plutonium is re-
moved from solution by hydrolysis to Pqu, solubilities are increased by
increasing the acidity. Table 6—9 summarizes data on the solublhty be-
havior of plutonium for various acidities.

*Contribution from R. E. Leuze.
6-6] URANYL SULFATE BLANKET PROCESSING 327

TABLE 6-9

SOLUBILITY OF TETRAVALENT PLUTONIUM
IN 1.4m UOzSO4 AT 25000.’

 

Excess sulfuric acid, Pu(IV) solubility,
m mg/kg H20

 

 

3.7
17
39
68

105

Scoooo
> 0 D)

 

 

 

 

Plutonium behavior is difficult to predict because of its complex valence
pattern. In the absence of irradiation, plutonium dissolved in 1.4 m U0,SO4
under a stoichiometric mixture of hydrogen and oxygen at 250°C exists in
the tetrapositive state. However, when dissolved chromium is present or
when an overpressure of pure oxygen is used, part of the plutonium is oxi-
dized to the hexapositive state. Experiments indicate that in the presence
of Co% gamma irradiation [18], reducing conditions prevail even under an
oxygen pressure and plutonium is held in the tetrapositive state. The
valence behavior discussed here is somewhat in question, since actual
valence measurements were made at room temperature immediately after
cooling from 250°C. It is known that tetrapositive plutonium will dispro-
portionate upon heating [19]. The disproportionation in a sulfate system
1s depressed by the sulfate complex formation with tetrapositive plu-
tonium. These results indicate that plutonium in a reactor will be pre-
dominantly in the tetrapositive state.

When the plutonium concentration exceeds the solubility limit, plu-
tonium will hydrolyze to form small particles of PuOs about 0.5 micron
in diameter and in pyrex, quartz, or gold equipment forms a loose preci-
pitate with negligible amounts adsorbed on the walls., However, if these
solutions are contained in type-347 stainless steel, titanium, or Zircaloy,
a large fraction of the PuO2 adsorbs on and becomes incorporated within
the oxide corrosion film. Attempts to saturate these metal surfaces with
plutonium in small-scale laboratory experiments were unsuccessful even
though plutonium adsorption was as much as 1 mg/cm?.

6-6.3 Neptunium chemistry in uranyl sulfate solution. Neptunium dis-
solved in 1.4 m UO2804 at 250°C under air, stoichiometric mixture hy-
drogen and oxygen, or oxygen is stable in an oxidized valence state, prob-
328 CHEMICAL PROCESSING [cHAP. 6

ably Np(V). The solubility is not known, but it is greater than 200 mg/kg
H20. Since the equilibrium concentration is only about 50 mg/kg H-O,
for a 1.4 m UO2S04 blanket with an average flux of 1.8 X 10'* neu-
trons/(cm?2)(sec), all the neptunium should remain in solution in most
reactor designs.

6-6.4 Plutonium behavior under simulated reactor conditions. Pluto-
nium behavior in actual uranyl sulfate blanket systems has not been
studied; however, small-scale static experiments with 100 ml of solution
and circulating loop experiments with 12 liters of solution have been car-
ried out in the absence of irradiation under conditions similar to those ex-
pected in an actual reactor.

In the static experiments, plutonium was added batchwise to 1.4 m
UO0.S04 at a rate of about 6 mg/kg HoO/day. The solution was heated
overnight in a pyrex-lined autoclave at 250°C under 200 psi hydrogen and
100 psi oxygen. The solution was cooled to room temperature for analysis
and for adding more plutonium. This was repeated until a total of 140 mg
of plutonium per kilogram of water was added. Small disks of type—347
stainless steel were suspended in the solution throughout the experiment to
determine the amount of plutonium adsorption. The behavior of plu-
tonium for a stainless-steel surface area/solution volume ratio of 0.6 cm?2/ml
is shown in Fig. 6-13. As the plutonium concentration was gradually in-
creased to 45 mg/kg H20, essentially all the plutonium remained in solu-
tion as Pu(VI). There was a small amount of adsorption, but no precipita-
tion. During the next few additions the amount of plutonium in solution
decreased rapidly to about 5 mg/kg H20. At the same time there was a
rapid increase in plutonium adsorption and in the formation of a loose
PuQ; precipitate. All plutonium added after this was either adsorbed or
precipitated.

Other experiments were made with surface/volume ratios of 0.2 and
0.4 cm2/ml. In all cases, the plutonium remaining in solution and the plu-
tonium adsorption per square centimeter were essentially the same as
that shown in Fig. 6-13. Thus, by decreasing the surface/volume ratio,
it is possible to increase the amount of plutonium in the loose precipitate.
For example, when the total plutonium addition was 130 mg/kg H20, 40%
of the plutonium was as a loose precipitate for a surface/volume ratio of
0.6 cm2/ml, 60% for a ratio of 0.4 cm?/ml, and 68% for a ratio of
0.2 cm?2/ml.

Plutonium behavior under dynamic conditions was studied by injecting
dissolved plutonium sulfate and preformed PuO: into a circulating stream
of 12 liters of 1.4 m U02S04 at 250°C under 350 psi oxygen. This solution
was contained in a type—347 stainless steel loop equipped with a canned
rotor pump, a hydroclone, metal adsorption coupon holders, and a small
6—6] URANYL SULFATE BLANKET PROCESSING - 329

 

 

 

 

 

 

 

 

 

60
| | | | | I |
C\ ON o
o T 40 | \ 1.4 m UO,SO, at 250°C
% g \‘ 100 psiO, + 200 psi H, ,
‘2 g \ Surface to Volume 0.6 cm“/ml
- 20 | \ _
\
|_g 2 4 4 i
0 I | l — | -
1000 | | | | | | |
100 — —
v\
o N
£ £
oY 10 | —
N
29
1 — —]
0.1 | | | | l |
60
l I | l l |
o
3N 40| —
a7
Y x
0
a O
c E 20 —
L | | I l l
0 20 40 60 80 100 120 140 160

Total Plutonium Added, mg/kg H2O

F1g. 6-13. Plutonium behavior in uranyl sulfate solution contained in type—347
stainless steel.

pressure vessel that could be connected and removed while the loop was in
operation. Plutonium was added and circulating-solution samples were
taken through this vessel. Tetrapositive plutonium added to the circulating
solution was completely oxidized to hexapositive in less than 5 min. When
45 mg/kg H»0 of dissolved plutonium was added every 8 hr, the amount
of plutonium circulating in solution increased to a maximum of about
150 mg/kg H20O. As more plutonium was added, it was rapidly adsorbed
on the loop walls. After the last addition of plutonium, the loop was
operated at 250°C for several days. Twelve hours after the last addition
the plutonium concentration had decreased to 100 mg/kg H20, and about
40 hr later the amount of plutonium in solution had dropped to an ap-
parent equilibrium value of 60 mg/kg H20O. Essentially all the plutonium
removed from solution was adsorbed on equipment walls uniformly
throughout the loop. Less than 0.19; of the plutonium was removed in
330 CHEMICAL PROCESSING [cHAP. 6

the hydroclone underflow, and no precipitated solids were circulating.
Even when 850 mg of plutonium as preformed PuO2 was injected into the
loop, no circulating solids were detected 5 min later. Only 209, of this plu-
tonium was removed by the hydroclone, 359, was adsorbed on the stainless
steel, and the rest was distributed throughout the horizontal sections of
the loop as loose solids. The hydroclone was effective for removing solids
that reached it, but the loop walls and low velocity in horizontal pipes
were effective traps for PuOa.

There are several differences in conditions between the loop runs and an
actual reactor, the most important of which are probably the presence of
radiation, the lower surface/volume ratio (0.4 compared with 0.8 ¢m?/ml
for the loop), the slower rate of plutonium growth in the reactor (12 to
15 mg/kg H20/day) and the probability that a plutonium producer
would have to be constructed of titanium and Zircaloy to contain the con-
centrated uranyl-sulfate solution. Based on these laboratory results, how-
ever, it appears that plutonium adsorption on metal walls may be a serious
obstacle to processing for removal of precipitated PuOs..

6-6.5 Alternate process methods. Because of the problem of plutonium
adsorption on metal walls, removal methods based on plutonium concen-
trations well below the solubility limit have been considered. In a full-
scale reactor plutonium will be formed at the rate of up to 12 to 15
mg/kg HeO/day. In order to keep the plutonium concentration below
3 mg/kg H20, the entire blanket solution must be processed at least four
to five times a day. By adding 0.4 m excess H2SO4 (see Table 6-9), the
plutonium solubility is increased to greater than 100 mg/kg H20 and the
blanket processing rate can be decreased to once every 3 or 4 days. Slightly
longer processing cycles can be used if part of the plutonium is removed as
neptunium before it decays.

Of the various alternate processes considered, ion exchange and ad-
sorption methods show the most promise. Dowex-50 resin, a strongly
acidic sulfonic acid resin loaded with UO2**, completely removed tetra-
positive plutonium from 1.4 m UO2S04 containing 20 mg of plutonium
per liter [20]. The resin capacity under these conditions, however, has not
been determined. Because of the high radiation level it may not be feasible
to use organic resins. Sorption of plutonium on inorganic materials shows
some possibilities as a processing method [21]. Although rather low plu-
tonium /adsorber ratios have been obtained, indications are that capacities
will be significantly higher at higher plutonium concentrations. Special
preparation of the adsorbers should also increase capacities. Attempts to
coprecipitate plutonium with tri- or tetrapositive iodates, sulfates, oxalates,
and arsenates were not successful, owing to the high solubilities of these
materials in 1.4 m UO2504.
6-6] URANYL SULFATE BLANKET PROCESSING 331

o= ———

Caustic r—+»To Off Gas
i

  
    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Irridiated
Thorium Scrubber
Additive Condenser
. L}
Dissolvent — : | To Waste
-] 1
' 1
Y !f Lean
” Acid |t Acid
Catch Fractionator
Tank :—_ -
¥ |
Dissolver _____________V_V\.Igs_te..!
f
: To Solvent
| Extraction
Yy l Reboiler
e—Feed Y
Adjustment
Tank Recovered
Acid Tank

 

 

F1a. 6-14. Thorex process, feed preparation flowsheet.

}<———— First Cycle —+——— Second Cycle ————-‘

Aqueous Aqueous Aqueous Aqueous Aqueous
Scrub Strip Scrub Strip Strip

From Feed 'L ‘L‘_, To - J_ i ’ JT—» Organic

Preparation I Solvent I I To
Recovery Th First

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Th Th
U Y c U o U 5 Cycle
| 5 I o Th, U = I g £ o
| 5 5| o 2 &
| 3 2l Feed £ £ 2
& & ‘“ & »
I Reductant I

  

 

 

 

  

Organic
Extractant

 

Th . — ITh :
U Organic Organic | | Uranium

Extract- Scrub |1 Isolation
ant

Waste
or Th
Evaporator Evaporator on
Storage vap = Th Product

Waste

   

 

 

 

 

 

Fiag. 6-15. Thorex process, solvent extraction co-decontamination flowsheet.
332 CHEMICAL PROCESSING [cHAP. 6

‘-————Isolafion ~{= Third Cycle——‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I ElutriantJrElufricnf I Aqueous Aqueous
F Scrub Strip
rom
Solvept Effluent -j—- ._L To
Extraction —'I ™ 10 Waste ——-| _>So|vent

c c Recovery
£l U E tl U g
© o5 Third Cycle '8 2

U O O Feed — v
g _S Preparation 2 CED
£ o o 8
o v U % =

wi n
= D
Organic
> Extractant
I U
U
' | Waste
Waste < U Evaporator U
or
Storage
To St To
© vlorage Shipment

Fia. 6-16. Thorex process, uranium isolation and third cycle flowsheet.

6—7. THoriuM OXIDE BLANKET PROCESSING

6-7.1 Introduction. At the present the only practical method available
for processing irradiated thorium-oxide slurry is to convert the oxide to a
natural water-thorium nitrate solution and treat by the Thorex process.
Although this method is adequate, it is expensive unless one plant can be
built to process thorium oxide from several full-scale power reactors.
Therefore methods for ThO2 reprocessing which could be economically in-
corporated into the design and operation of a single power station have
been considered. Alternate methods that have been subjected to only brief
scouting-type experimentation are discussed in Article 6-7.3.

6-7.2 Thorex process.* The Thorex process has been developed to sep-
arate thorium, U233 fission product activities, and Pa?33; to recover the
thorium and uranium as aqueous products suitable for further direct
handling; and to recover isotopically pure U233 after decay-storage of the
Pa233, The flowsheet includes two solvent-extraction cycles for thorium
and three solvent-extraction cycles plus ion exchange for the uranium.
Although only irradiated thorium metal has been processed, the process is
expected to be satisfactory for recovery of thorium and uranium from
homogeneous reactor fuels.

The Thorex process may be divided into three parts: feed preparation,

*Contribution from W. T. McDuffee.
6-7] THORIUM OXIDE BLANKET PROCESSING 333

solvent extraction, and product concentration and purification. These
three divisions are shown in Figs. 6-14, 6-15, and 6-16.

In the feed preparation step, uranyl sulfate solution from the reactor core
and thorium oxide from the blanket system, freed of D20 and suspended in
ordinary water, are fed into the dissolver tank. The dissolvent is 13 N
nitric acid to which has been added catalytic amounts (0.04 N) of sodium
fluoride. When short-cooled thorium is being processed, potassium iodide
1s added continuously to the dissolver to provide for isotopic dilution of
the large amount of fission-produced I'3! which is present. The dissolver
solution is continuously sparged with air, and the volatilized iodine is re-
moved from the off-gases in a caustic scrubber.

The dissolver solution is transferred to the feed adjustment tank where
aluminum nitrate is added, excess nitric acid recovered, and the resultant
solution made slightly acid-deficient by evaporating until a temperature
of 155°C is reached. During digestion in the feed adjustment tank any
silica present is converted to a form that will not cause emulsion problems
in pulse columns, and fission products generally are converted to forms
less likely to be extracted by the solvent (429, TBP in Amsco).

In the solvent extraction step thorium and uranium are co-extracted in
the first cycle; subsequent partitioning of thorium and uranium in the
second cycle gives two decontaminating cycles to both products while
using only five columns. For short-decayed thorium a reductant, sodium
hydrogen sulfite, is continuously added to the feed streams of both cycles
to decrease the effect of nitrite formed by irradiation. Without the sulfite
addition, the nitrite formed by radiation decomposition of nitrates con-
verts ruthenium to a solvent-extractable form. Acid deficiency in the
second cycle feed is achieved by adding dibasic aluminum nitrate (diban).

The spent organic from the second cycle is recycled to the first cycle as
the organic extractant. The spent solvent from the first cycle is processed
through a solvent-recovery system and reused as the organic extractant in
the second cyecle.

In the uranium product concentration and purification step (Fig. 6-16),
uranium is isolated by ion exchange, using upflow sorption and downflow
elution. In this way a concentrated uranium solution in 6 N HNOs; is ob-
tained. This solution is stable enough for storage or is suitable as a feed
for the third uranium extraction cycle. The third uranium cycle is a
standard extraction-stripping solvent-extraction system using 15% TBP-
Amsco as the organic extractant. Although installed as a part of the com-
plete Thorex flowsheet, the third cycle may be used separately for re-
processing long-stored uranium to free it of objectionable decay daughters
of U232, When used as an integral part of the Thorex scheme, additional
decontamination of the uranium is achieved and the nitrate product is
well adapted for extended storage or future reprocessing.
[cHAP. 6

 

CHEMICAL PROCESSING

334

 

 

 

 

 

 

 

 

0 X ¢ g0l X € 00T X £ 0T X% | o0T X G 01 X6 pakesdp shep 0g
— g0T X 6 ¢0T X 8 0T X G 0T X € ¢0T X € pakeoap sABD (0¥
Qez-wniueI()
g0 X 6 00T X G 50T X € 003 00T X 2 0T X ¥ padeoop sABp 0g
— 00T X T 0T X € 0T X ¥ ;0T X T 0T X T pakeodp sAep Q0¥
Eflfiofl,fi
1 SUJIB 9IBI [B}O ], qN-17 ny B $S0IK)
SI030%] UOI}BUIWIBIUOII(T

 

 

 

 

‘sBp (¢ ATUO POABIOP [BLISYEUI I0] 9[0£D WINIUBIN [BUOI}IPPEB dUO
‘WNLIOY) PUB WINTUBIN [J0q 10§ $9[0L0 93a[dwoo om) ‘u0} Jod ¢EZ-SSBW JO SWRBIZ (J)CE 0} POJBIPBLII WINLIOYT,

INVIJ LOTIJ XHYOH], THL NI
SLONAO¥J WAINVH() ANV WAMMOH]J, ¥0d SYOLOV] NOLLVNINVLNODIE(] EOVHIAY

019 dI1dV],
6-7] THORIUM OXIDE BLANKET PROCESSING 335

For return to an aqueous homogeneous reactor the decontaminated
uranium would probably be precipitated as the peroxide, washed free of
nitrate, and then dissolved in D2SO4 and D20. Product thorium would
be converted to thorium oxide by methods described in Section 4-3.

The adaptability of the Thorex flowsheet just described to processing
thorium irradiated to contain larger amounts of U223 per ton and decayed
a short time has been demonstrated in the Thorex Pilot Plant at Oak Ridge
National Laboratory [22]. Fifteen hundred pounds of thorium irradiated
to 3500 grams of U233 per ton and decayed 30 days was processed through
two thorium cycles and three uranium cycles. The decontamination fac-
tors for various elements achieved with short-decayed material are com-
pared in Table 6-10 with results obtained with longer-decayed material.
While the decontamination factors obtained with the short-decayed ma-
terial compare favorably with the factors for the long-decayed material,
the initial activity in the short-decayed thorium was 1000 times greater
than in the long-decayed. Therefore, while the thorium and uranium
products did not meet tentative specifications after two complete cycles,
the uranium product did meet those specifications after the third uranium
cycle. Since the chemical operations necessary to convert these materials
to forms suitable for use in a homogeneous reactor can be carried out re-
motely, the products are satisfactory for return to a homogeneous reactor
after two cycles.

6-7.3 Alternate processing method.* Attempts to leach protactinium
and uranium produced in ThOg particles by neutron irradiation [23] in-
dicate that both are rather uniformly distributed throughout the mass of
the ThO. particle, and migration of such ions at temperatures up to 300°C
is extremely slow. Since calculations show that the recoil energy of frag-
ments from U233 fission is sufficiently large to eject most of them from a
particle of ThO2 not larger than 10 microns in diameter, this offers the
possibility of separating fission and corrosion products from a slurry of
ThO» without destroying the oxide partic'es. Such a separation, however,
depends on the ability to remove the elements that are subsequently ad-
sorbed on the surface of the ThO2. Adsorption of various cations on ThO2
and methods for their removal are discussed in the following paragraphs.

Trace quantities of such nuclides as Zr°5, Nd!47, Y°!, and Ru'% when
added to a slurry of ThO2 in water at 250°C are rapidly adsorbed on the
oxide particles, leaving less than 107%% of the nuclides in solution. The
tracer thus adsorbed cannot be eluted with hot dilute nitric or sulfuric acid.
The adsorption of macroscopic amounts of uranium or neodymium on
ThO, at 250°C is less for oxide fired to 1600°C than for 650°C-fired oxide,

*Contribution from R. E. Leuze.
336 CHEMICAL PROCESSING [cHAP. 6

TABLE 6-11

ErrEcT OF CALCINATION TEMPERATURE ON
UraNIUM AND NEODYMIUM ADSORPTION ON THOs AT 250°C

0.5 g of ThO2 slurried at 250°C in 10 ml of 0.005 m Nd(NO3)3
or 0.0bm UOzSO4—O.O5 m HzSO4.

 

 

 

 

Adsorption, mg/g Th*
Calcination temperature, P 8/8
o
C
U Nd
650 3.3-4.4 7.4
850 1.9-2.4 6.1
1000 0.72-1.10 | 2.4
1100 0.08-0.19 | 0.5
1600 0.06-0.12 0.3

 

 

 

 

 

*Single numbers represent data from single experiments. In other cases the range
for several experiments is given.

TABLE 6-12

Use or PBO To DECREASE CATION ADSORPTION ON THO»

0.2 g of ThO2 plus various amounts of PbO coslurried in 10 ml of solution at
250°C for 8 hr.

 

 

 

Solids PbO/ Th.Oz Solution Cation adsorbed on ThO.,

wt. ratio ppm
ThO: 0.002 m UO2804 3100
PbO + ThO. 0.2 0.002 m U0O2804 220
ThO: 0.001 m Ce(NO3)3 6200
PbO + ThO. 0.2 0.001 m Ce(NO3)3 10
ThO. 0.01 m Nd tartrate 2700
PbO + ThO- 0.4 0.01 m Nd tartrate 10

 

 

 

 

 

 
6-7] THORIUM OXIDE BLANKET PROCESSING 337

as illustrated in Table 6-11. This change in amount of adsorption may be
almost entirely due to decrease in surface area of ThO2 with increased
firing temperature. The surface area of 1600°C-fired ThO: is only 1 m2/g
ThOg2, while the 650°C-fired ThO3 has a surface area of 35 m2/g ThOs..

The cation adsorption on ThO2 can be decreased by coslurrying some
other oxide with the ThOj;. The added oxide must adsorb fission products
much more strongly than ThO2 and be easily separable from ThOs. The
effectiveness of PbO in decreasing cation adsorption on ThO: is shown in
Table 6-12. When PbOs was used, more than 999% of the cations added
to the ThO2-PbO slurry was adsorbed on the PbO.. However, cations
adsorbed on ThO2 were not transferred to PbOs; when it was added to
slurry in which the cations were already adsorbed on the ThO2 particles.
Addition of dilute nitric acid to the ThO2-PbO coslurry completely dis-
solved the PbO and the cations adsorbed on it without disturbing the ThO..

In all cases, cations adsorbed on ThO2 at 250°C are so tightly held that
dilute nitric or sulfuric acid, even at boiling temperature, will not remove
the adsorbed material. However, the adsorbed ions can be desorbed by
refluxing the ThO; in suitable reagents under such conditions that only a
small amount of 1600°C-fired ThO: is dissolved. Under the same treat-
ment ThO: fired to only 650°C would be 909 dissolved.
338 CHEMICAL PROCESSING [cHAP. 6

REFERENCES

1. A. T. Gresky and E. D. ArNoLp, Poisoning of the Core of the Two-region
Homogeneous Thermal Breeder: Study No. 2, USAEC Report ORNL CF-54-2-208,
Oak Ridge National Laboratory, 1954.

2. E. D. Arnorp and A. T. GreskY, Relative Biological Hazards of Radiations
Expected tn Homogeneous Reactors TBR and HPR, USAEC Report ORNL-1982,
Oak Ridge National Laboratory, 1955.

3. R. A. McNEEs and S. PeTERSON, in Homogeneous Reactor Project Quarterly
Progress Report for the Pertod Ending July 31, 1956, USAEC Report ORNL-1943,
Oak Ridge National Laboratory, 1955. (pp. 201-202)

4. D. E. FErcusoN et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 30, 1956, USAEC Report ORNL-2096, Oak
Ridge National Laboratory, 1956. (p. 118)

5. P. A. Haas, Hydraulic Cyclones for Application to Homogeneous Reactor
Chemacal Processing, USAEC Report ORNL-2301, Oak Ridge National Labora-
tory, 1957.

6. W. D. BurcH et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak
Ridge National Laboratory, 1957. (p. 26)

7. R. A. McNEEs et al.,, Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2379, 1957
(p. 137); ORNL-2432, 1957 (p. 147); ORNL-2493, 1958.

8. W. E. BrownNiNng and C. C. BoLra, Measurement and Analysis of Holdup
of Gas Mixtures by Charcoal Adsorption Traps, USAEC Report ORNL-2116,
Oak Ridge National Laboratory, 1956.

9. W. D. BurcH et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Report ORNL-2432, 1957
(p. 23); ORNL-2493, 1958.

10. P. H. HarrLEYy, HRT Mock-up Iodine Removal and Recombiner Tests,
USAEC Report CF-58-1-138, Oak Ridge National Laboratory, 1958.

11. R. A. KeELER et al., Vitro Laboratories, 1957. Unpublished.

12. S. PETERSON, unpublished experiments.

13. D. E. FErguson et al., Oak Ridge National Laboratory, in Homogeneous
Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2272, 1957
(pp. 133-135); ORNL-2331, 1957 (pp. 142-143, 148-149); ORNL-2379, 1957
(pp. 138-139).

14. R. A. McNEEgs et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 30, 19565, USAEC Report ORNL-1895, Oak
Ridge National Laboratory, 1955. (pp. 175-176)

15. D. E. FErGUsON, Removal of Iodine from Homogeneous Reactors, USAEC
Report CF-56-2-81, Oak Ridge National Laboratory, 1956; Preliminary Design
of an Iodine Removal System for a 460-Mw Thorium Breeder Reactor, USAEC
Report CF-56-7-12, Oak Ridge National Laboratory, 1956.

16. H. O. WEEREN, Preliminary Design of HRE-3 Iodine Removal System
#1, USAEC Report CF-58-2-66 Oak Ridge National Laboratory, 1958.

17. S. PETERSON, Behavior of Iodine in the HRT, USAEC Report CF-58-3-75,
Oak Ridge National Laboratory?® 1958.
REFERENCES 339

18. R. E. LruzE et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Apr. 30, 1957, USAEC Report ORNL-2331, Oak
Ridge National Laboratory, 1957. (p. 151)

19. R. E. CoNNicK, in The Actinide Elements, National Nuclear Energy Series,
Division IV, Volume 14A. New York: McGraw-Hill Book Co., Inc., 1954.
(pp. 221, 238-241)

20. S. S. Kirsuis, Oak Ridge National Laboratory. Unpublished. |

21. R. E. LeuzE and S. S. Kirsuis, Oak Ridge National Laboratory, 1957.
Unpublished.

22. W. T. McDvurree and O. O. YArBRO, Oak Ridge National Laboratory,
1958. Unpublished. o

23. D. E. FERGUSON et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1956, USAEC Report ORNL-1943, Oak
Ridge National Laboratory, 1955. (p. 221)
CHAPTER 7

'DESIGN AND CONSTRUCTION OF EXPERIMENTAL
HOMOGENEOUS REACTORS*

7—1. INTRODUCTION

7-1.1 Need for reactor construction experience. The power reactor de-
velopment program in the United States is characterized by the construc-
tion of a series of experimental reactors which, it is hoped, will lead for each
reactor type to an economical full-scale power plant. Outstanding examples
of this approach are afforded by the pressurized water reactor and boiling
water reactor systems. The development of pressurized water reactors
started with the Materials Testing Reactor, followed in turn by the Sub-
marine Thermal Reactor (Mark I), the Nautilus Reactor (Mark II), and
the Army Package Power Reactor. Experience obtained from the construc-
tion of these reactors was applied to the full-scale plants built by the West-
inghouse Electric Company (Shippingport and Yankee Atomic Electric
Plants) and Babcock & Wilcox Company (Consolidated Edison Plant).

Although many have argued that the shortest route to economic power
will be achieved by eliminating the intermediate-scale plants, most experts
believe that eliminating these plants would be more costly in the long run.
To quote from a speech by Dr. A. M. Weinberg [1], while discussing large-
scale reactor projects: ““The reactor experiment—a relatively small-scale
reactor embodying some, but not all, the essential features of a full-scale
reactor—has become an accepted developmental device for reactor tech-
nology.”’ |

An alternative to the actual construction of experimental nuclear reactors
has been proposed which consists of the development of reactor systems
and components in nonnuclear engineering test facilities, zero-power critical
experiments, and the testing of fuel elements and coolants in in-pile loops.
This approach, although used successfully in the development of various
solid-fuel coolant systems, is not completely applicable to circulating-fuel
reactors because of the difficulty of simulating actual reactor operating
conditions in such experiments. In in-pile loops, for example, the ratio of
the volume of the piping system to the volume of the reacting zone is never
quite the same as in a reactor, making it impossible to duplicate simul-
taneously the conditions of fuel concentration, enrichment, and power
density. In cases where these variables are important, the in-pile loops

*Prepared by J. A. Lane, with contributions from S. E. Beall, S. I. Kaplan, Oak
Ridge National Laboratory, and D. B. Hall, Los Alamos Scientific Laboratory.

340
7-2] WATER BOILERS 341

can at best provide information of an exploratory nature which must be
verified in an operating fluid-fuel reactor.

A second aspect of circulating-fuel reactors, which precludes relying
solely on engineering tests and in-pile loops, is the close interrelation of the
nuclear behavior and the operational characteristics of the fuel circulation
system, which can be determined only through construction and operation
of a reactor. Other aspects of reactor design that can be best determined
In an operating homogeneous reactor are continuous removal of fission
products produced in the nuclear reaction and remote decontamination
and maintenance of reactor equipment and piping.

7-1.2 Sequence of experimental reactors. It is obvious from the fore-
going that the construction of a sequence of experimental reactors has been
an important factor in the development of homogeneous reactors. In this
sequence, which started with nonpower research reactors, seven such
reactors have been built (not including duplicates of the water boilers).
These are the Low Power Water Boiler (LOPO), the High Power Water
Boiler (HYPO), the Super Power Water Boiler (SUPO), the Homogeneous
Reactor Experiment (HRE-1), the Homogeneous Reactor Test (HRE-2),
and the Los Alamos Power Reactor Experiments (LAPRE-1 and -2).
In the sections of this chapter which follow, these reactors are described
in detail, and their design, construction, and operating characteristics are
compared. Their construction covers the regime of homogeneous reactor
technology involving the feasibility of relatively small reactors fueled with
aqueous solutions of uranium. Since their construction and operation does
not include systems fueled with aqueous suspensions of thorium oxide
and/or uranium oxides necessary for the development of full-scale homo-
geneous breeders or converters, additional experimental reactors will un-
doubtedly be built.

7-2. WATER BoOILERS*

7-2.1 Description of the LOPO, HYPO, and SUPO [2-4]. Interest in
homogeneous reactors fueled with a solution of an enriched-uranium salt
was Initiated at the Los Alamos Scientific Laboratory in 1943 through an
attempt to find a chain-reacting system using a minimum of enriched fuel.
The first of a sequence of such reactors, known as LOPO (for low power),
went critical at Los Alamos in May 1954 with 565 grams of U235 as uranyl
sulfate. The uranium, containing 14.59, U235 was dissolved in approxi-
mately 13 liters of ordinary water contained in a type—347 stainless steel
sphere 1 ft in diameter and 1/32 in. in wall thickness. The sphere was sur-

*Prepared from reports published by Los Alamos Scientific Laboratory and
other sources as noted.
342 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

rounded by beryllium oxide as reflector in order to minimize the critical
mass of the U235, The lack of a shield and cooling system limited the heat
power level of LOPO to 50 milliwatts. A cross-sectional drawing of the
LOPO is shown in Fig. 7-1.

Following successful low-power operation of the LOPO, the reactor was
provided with a thicker sphere (1/16 in.), integral cooling coils, and a
shield to permit operation at 6 kw. Also, part of the beryllium oxide
reflector was replaced by a graphite thermal column, and holes through
the shield and reflector were provided for experiments. The critical mass
of the modified reactor was 808 grams of U235 as uranyl nitrate at 14.09,
enrichment, contained in 13.65 liters of solution. The change from uranyl
sulfate to nitrate was made because an extraction method for the removal
of fission products was known only for the latter solution at that time. The
modified reactor, called HYPO (high power), went critical in December
1944 and operated at a normal power of 5.5 kw, producing an average
thermal-neutron flux of 10! neutrons/(cm?)(sec). The temperature of
the solution during operation reached 175°F with cooling water (50 gal/hr)
at 46°F.

Since higher neutron fluxes were desired, as well as more research facilities
than available from HYPO, the reactor was further modified and renamed
SUPO (super power water boiler).' »

The modifications were made in two parts. The first phase, begun in
April 1949 and completed in February 1950, improved the experimental
facilities and increased the neutron flux. The second phase, begun in
October 1950 and completed in March 1951, increased the thermal neutron
irradiation facilities, improved the reactor operation, and removed the
explosive hazard in the exhaust gases.

The first group of alterations consisted of the followmg

(1) The space around the reactor was increased by enlarging the building
so that experiments could be carried out on all four sides instead of only two.

(2) The construction of a second thermal column was made possible by
eliminating a removable portion of the reactor shield. This made available
a neutron beam and irradiation facilities on a previously unused face of the
reactor.

(3) The entire spherical core assembly was replaced as follows:

(a) Three 20-ft-long, 1/4-in.-OD, 0.035-in.-wall-thickness stainless
steel tubes replaced the former single cooling coil. This increased the
operating power level from 5.5 kw to a maximum of 45 kw.

(b) A new removable level indicator and exit gas unit was installed
in the sphere stack tube. The stack tube itself was made more accessible
for future modifications.

(¢) External joints were not welded, but unions of flare fittings
were used to simplify the removal of the sphere or permit pipe replacements.
7-2] WATER BOILERS 343

 

 

|~ Overflow

 

Cadmium Level
Safety Curtain ~ eve
Y " Electrode
Else::fc:)ée | Cd Control Rod

 

Upper Pipe

 

Stainless Steel Sphere
|_—— Containing Enriched

Uranyl Sulfate (UO9SO
BeO Reflector ranyl Sulfate (UO2504)

 

_r/ Graphite Reflector

 

 

 

 

 

 

—_ —

 

 

Level Electrode

 

 

Air Pipe

 

~~

 

 

 

 

 

Dump Basin

 

 

UL

Fig. 7-1. Cross section of LOPO, the first aqueous solution reactor.

(d) An additional experimental hole was run completely through the
reactor tangent to the sphere. This 1y%-in.-ID tube supplemented the
1-in.-ID “‘glory hole’’ running through the sphere.

(4) The beryllium portion of the reflector was replaeed by graphite.
The all-graphite reflector gave a more rapid and complete shutdown of the
reactor and eliminated the variable starting source produced by the (y,n)
reaction on beryllium. A 200-millicurie RaBe source placed in the reflector
was used as a startup neutron source.

(5) Two additional vertical control rods were added which moved into
the sphere in re-entrant thimbles. These consisted of about 120 grams
of sintered B! in the form of 9/16-in. rods about 18 in. long. These
rods gave the additional control required by the change to an all-graphite

reflector. Previously observed shadow effects were eliminated by the in-
ternal position of the rods and by the location of the control chambers
under the reactor.

(6) The reactor solution was changed from 15% U235_enriched uranyl
nitrate to one of 88.79, enrichment. This made possible the continued use
of a low uranium concentration in the solution with the poorer all-graphite
reflector. The gas evolution produced by nitric acid decomposition was
greatly reduced, due to the lower total nitrogen content.
344 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [CHAP. 7

(7) The entire inner reactor shield was improved to permit higher power
operation with a low neutron leakage and also to increase the neutron-to-
gamma-ray Intensity in the thermal columns. Cadmium was replaced by
B4C parathin and additional steel shielding was added.

After operating the reactor with the above modifications for about 10,000
kwh at a power of 30 kw, the following (second group) alterations were
made:

(1) The original south thermal column was completely rebuilt, with
improved shielding to provide many more irradiation facilities.

(2) A recombination system was constructed to handle the off-gases from
the reactor. The use of a closed circulating gas system with a catalyst
chamber of platinized alumina removed any explosive hazard in the ex-
haust gases due to the presence of hydrogen and oxygen. The operating
characteristics of the reactor were greatly improved by returning directly
back to the reactor as water all but a very small fraction of the gases
produced.

(3) A shielded solution-handling system was constructed to simplify
the procedure of routine solution analysis and for the removal or change of
the entire reactor solution.

The average neutron flux in the SUPO during operation at 45 kw is
about 1.1 X 10'? neutrons/(cm?2)(sec), and the peak thermal flux (in the
“glory hole”) is 1.7 X 10'2 neutrons/(cm?)(sec). Estimated values for the
maximum intermediate and fast fluxes at 45 kw are 2.8 and 1.9 X 10'2 neu-
trons/(cm?)(sec), respectively. Calculations made from fast beams emerg-
ing from the north thermal column at this same power level gave the fol-
lowing fast-flux values above 1 Mev in units of neutrons/(cm?) (sec):
(1) at sphere surface, 1.1 X 10'2; (2) at bismuth column, 7 X 101°; and
(3) at a graphite face 1 ft in front of the bismuth column, 2 X 109,

The production of hydrogen plus oxygen due to radiation decomposition
amounts to approximately 20 liters/min during operation of the reactor at
45 kw. These gases leave the reactor core and pass through a reflux con-
denser which removes much of the water vapor and then through a stainless
steel-wool trap for final moisture removal. A blower feeds the gas into one
of two interchangeable catalyst chambers containing platinized alumina
pellets. These chambers, operating at 370 to 470°C, recombine the hy-
drogen and oxygen, and the gas leaving the catalyst contains the water
vapor formed. A second condenser reduces the temperature of the exit
gas to that entering the catalyst chamber. A total of 100 liters/min of gas
18 circulated continuously in the closed gas system at pressures slightly
above atmospheric, and the hydrogen concentration is kept below the
detonation limit at all points of the system. Excess pressures produced in
the gas system can be bled to the atmosphere through a 150-ft-high exhaust
stack.
7-2] WATER BOILERS

345

The characteristics of LOPO, HYPO, SUPO, and the North Carolina

State College Water Boilers are summarized in Table 7-1.

TaABLE 7-1

DEsigN CHARACTERISTICS OF WATER BOILERS

 

 

 

 

 

 

 

LOPO HYPO SUPO NCSR [5]
Power level, kw 5X 107° 5.6 45 10
Solution (in H20) | U02S04 | UO2(NO3)2 UO2(NO3)2 U02504
U235 wt., grams 565 870 870 848
Solution volume, 13 13.65 13.65 15
liters
Enrichment, 9, 14.6 14.0 88.7 90
Maximum thermal- | 3 X 108 | 2.8 X 101! 1.7 x 102 5 X 101
neutron flux
Reflector material BeO Be and Graphite Graphite
graphite
Coolant flow rate None 50 180 240
gal/hr
Solution tempera- 39 85 85 80
ture, °C
Experimental | None 1 thermal 2 thermal 1 thermal
facilities column columns (“‘glory column
hole”” and 12 exposure
tangential hole) ports

 

 

 

7-2.2 Kinetic experiments in water boilers. In August 1953, experi-
ments were performed on the SUPO by a group of scientists from the Oak
Ridge National Laboratory and Los Alamos [6,7] to determine the degree
to which a boiling (and nonboiling) homogeneous reactor automatically

compensates for suddenly imposed supercritical conditions.

Previous

boiling experiments in 1951, unreported in the open literature [8], had
indicated the stability of SUPO under steady-state boiling conditions;
346 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [CHAP. 7

however, there remained considerable doubt as to the adaptability of a
reactor of this type to a sudden introduction of excess reactivity such as
might occur with a sudden increase in pressure above the reactor. The
tests were performed by suddenly ejecting a neutron poison, consisting of
an aluminum rod containing boron carbide at its tip, and simultaneously
measuring the neutron flux level with high-speed recorders connected to
a boron-coated ionization chamber located in the graphite reflector. The
amount of reactivity introduced was determined by the position of a
calibrated control rod. Although the experiments were interrupted by
frequent accidental scrams, caused by the unsuitability of SUPO to boil-
ing at high solution levels in the sphere, the results indicated that both
boiling and nonboiling solution reactors are capable of absorbing reactiv-
ity increases of at least 0.49; k ; added in about 0.1 sec. In both boiling
and nonboiling cases, the reactor power was self-regulating, but excur-
sions were terminated more rapidly under boiling conditions. The average
lifetime of prompt neutrons in the reactor was calculated from the initial
prompt rise in the neutron flux and found to be about 1.7 X 10™% sec.
Following a reactivity addition, the initial rate of reactivity decrease
(0.2 sec after start) was greater than about five times the rate which could
be attributed to core-temperature rise and the associated negative tempera-
ture coefficient (0.0249} k.;/°C). As the gas bubbles left the core region,
reactivity decrease due to core-temperature rise increased in relative
Importance. *

More recent experiments with the Kinetic Experiment for Water Boilers
(KEWB-1), operated by Atomics International for the U. S. Atomic
Energy Commission [9], have verified the self-controlling features of a
solution-type reactor. It was found that automatic shutdown due to the
temperature increase and formation of gas bubbles in the reactor fuel
solution occurs under all abnormal operating conditions tested.

7-2.3 The North Carolina State College research reactor [5]. The sim-
plicity of the Water Boiler reactor has made it of interest as a laboratory
tool for experimental work with neutrons and gamma rays and also to
provide training in reactor operation, and ten such reactors were in opera-
tion or planned in the United States by the end of 1957. The first college-
owned nuclear research reactor, which started operating at 10 kw in Sep-
tember 1953 at North Carolina State College, Raleigh, North Carolina,
was of this type. It was completed after four years of planning, design,
and construction, at a cost of $130,000 for the reactor, plus $500,000 for
the reactor building and associated laboratory equipment. It differs from
the Los Alamos SUPO in that the fuel container is a cylinder 11 in. in
diameter and 11 in. high, rather than a sphere. Its experimental facilities
include 12 access ports and a thermal column. |
7-2] WATER BOILERS 347

In June 1955, the reactor was shut down because of leaks which developed
in the fuel container and permitted the radioactive fuel to contaminate the
inside of the reactor shield. After a major repair job, operation of the
reactor with a new core was resumed in March 1957 at a power level of
500 watts, and the reactor has operated successtully at that level for a year.

7-2.4 Atomics International solution-type research reactors. Several
versions of the Water Boiler are being offered commercially by various
companies. The major supplier is the Atomics International Division of
North American Aviation Company which has built, or is building, 11 such
reactors. Low-power reactors are: the 1-watt Water Boiler Neutron Source
(WBNS) originally at Downey, California, which was moved to Santa
Susana and modified to operate at 2 kw; a new 5-watt laboratory reactor
(L-47) for Atomics International; the 100-watt Livermore Research
Reactor at Livermore, California; and a 5-watt reactor planned for the
Danish Atomic Energy Commission at Risg, Denmark. Higher power
Water Boilers, operating at 50 kw, include the Kinetic Experiment for
Water Boilers (KEWB-1) at Santa Susana; the UCLA Medical Facility
at Los Angeles, California; and reactors for the Armour Research Founda-
tion in Chicago, Illinois; the Japan Atomic Energy Research Institute at
Tokai, Japan; Farbwerke Hoechst A. G. at the University of Frankfurt,

   
    
   
 

Drive Motor
Concrete Shielding
v
:,‘“ % Control Rods

S~

Graphite Reflector

Al
-4 1 R
LV N )
5 KA I
AR s
-y I3
Y =
e, N
o o
v - <
< <N
>. A
N < v 7
| ]
v .
H v
b 'fll -,
'
) Al

Movable Concrete Door]

.

Fig. 7-2. Armour Research Foundation research reactor (courtesy of Atomics
International, a division of North American Aviation Co.)
348 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

West Germany ; the Senate of West Berlin (Institute for Nuclear Research),
Germany; and the Politecnico Enrico Fermi Nuclear Study Center at
Milan, Italy.

The first solution-type research reactor for industrial use went into
operation in June 1956. The general features of this reactor, which Atomics
International built for the Armour Research Foundation at the Illinois
Institute of Technology, Chicago, Illinois, are shown in Fig. 7-2.

The exposure facilities include one 6-in.-diameter beam tube extending
radially to within 2 ft of the core tank and 4 in. in diameter from there
to the core tank; two 4-in. and two 3-in. beam tubes extending radially to
the core tank; two 2-in. through-tubes passing tangentially to the core
tank; one 13-in.-diameter tube passing through the central region of the
core; four 4-in. vertical tubes located in the reflector; one 5 ft X 5 ft graphite
thermal column with a 12-in.-square removable section. The tube facilities
consist of steel sleeves extending through the concrete shield and aluminum
thimbles or liners which reach to the immediate vicinity of the core. Each
tube facility is equipped with a graphite reflector plug and a dense con-
crete-and-steel shielding plug to be installed when the facility is not in use.

The horizontal thermal column is formed by a 5-ft-square column of
graphite, in the center of which are nine removable graphite stringers. A
large volume which may be used for exposures is provided between the
end of the thermal column and the inner face of a movable concrete door.
The thermal column access ports open into this volume.

To take advantage of the 50,000 curies of gamma activity produced by
the fission-product gases circulating through the gas recombiner tank,
exposure facilities are provided which extend from the subpile room into
the exposure room and into the valve room. The facilities listed below
consist of steel sleeves and aluminum thimbles which extend through the
dense concrete walls of the exposure room.

2 gamma, ports, 4 in. diameter
2 gamma ports, 8 in. diameter
1 rectangular gamma slot, 6 in. X 18 in.

In addition, two 4-in.-diameter gamma ports extend from the subpile
room into the valve room. As with the beam tubes, each port is equipped
with a plug to be installed for shielding purposes when the port is not in use.

7-3. THE HoMoGENEOUS REACTOR ExpERIMENT (HRE-1) [10-13]*

7-3.1 Introduction. In 1950 the Oak Ridge National Laboratory under-
- took the task of designing, building, and operating a pilot-plant fluid-
fuel reactor, the Homogeneous Reactor Experiment (HRE-1), shown in

. *Based on a paper by C. E. Winters and S. E. Beall [10].
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 349

 

 

 

 

 

 

 

F1a. 7-3. The homogeneous reactor experiment (HRE-1).

Fig. 7-3. The purpose of this reactor was to investigate the nuclear and
chemical characteristics of a circulating uranium solution reactor at
temperatures and powers sufficiently high for the production of electricity
from the thermal energy released. Specifically, it was designed to operate
with a full-power heat release of 0.6 to 3.5 million Btu/hr (200 to 1000 kw
of heat), and a maximum fuel-solution temperature of 482°F, yielding,
after heat exchange, a saturated-steam pressure of about 200 psi.

During the 24-month period in which the reactor was in operation,
starting in April 1952, liquid was circulated for a total of about 4500 hr.
The reactor was critical a total of 1950 hr and operated above 100 kw for 720
hr. The maximum power level attained was 1600 kw. The reactor was shut
down in the spring of 1954 and dismantled to make room for the Homo-
geneous Reactor Test (HRE-2), having successfully demonstrated the
nuclear stability of a circulating-fuel reactor. The characteristics of the
HRE-1 are summarized in Table 7-2.

7-3.2 The reactor fuel system. The reactor core consisted of a stainless
steel sphere 18 in. in diameter, through which was circulated 100 to 120 gpm
of 93% enriched uranyl sulfate dissolved in distilled water. The temperature
350 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP.7

TABLE 7-2

CHARAcTERISTICS OF HRE-1

 

 

 

Power, heat 1000 kw

Fuel U02804 (939, enriched) in H20

Fuel concentration ~30 g U235 per liter (0.17 m UO02S0,)

U233 in core 1.5— 2 kg

Core 18 in. diameter, stainless steel

Pressure vessel 39 in. ID, 3 in. thick, forged steel

Reflector 10 in. D20, pressurized with He

Specific power 20 kw/liter

Fuel inlet temperature 210°C

Fuel outlet temperature 250°C

System pressure 1000 psi (430 psi above vapor pressure)

Gas removal system Vortex flow through core

Radiolytic gas recombination CuSOy4 (internal); flame and catalytic re-
combination (external)

Control system Reflector level, safety plates, temperature
control

Shielding 7 ft barytes concrete

Steam temperature 382°F

Steam pressure 200 pst

Electrical capacity 140 kw

 

 

rise of the solution passing through the core was about 72°F at a power
level of 1000 kw. The liquid was discharged from the core at a temperature
of 482°F, and cooled to 410°F by evaporating water from the shell side of a
U—-tube heat exchanger, thus generating about 3000 1b/hr of 200 psi steam.
A canned-rotor centrifugal pump returned the fuel to the core to be re-
heated. The total volume of solution in the high-pressure system was about
90 liters, of which 50 liters were in the core. A schematic flow diagram of
HRE-1 is shown in Fig. 7-4.

A total pressure of 1000 psi was maintained in the fuel system by heating
a small volume of fuel to 545°F in a pressurizer chamber directly above the
sphere. The 1000 psi total pressure, which is over 400 psi greater than is
required to prevent boiling of the fuel solution, was necessary to minimize
the volume of decomposition gases.

7-3.3 The reflector system. The reflector of HRE—-1 was a 10-in. layer
of heavy water surrounding the core vessel. The heavy water was pressur-
ized with helium to within 4100 psi of the fuel pressure in order to minimize
stresses in the 3/16-in. wall of the spherical fuel container. Both the
reflector and the concentric core were contained in an outer pressure vessel
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 351

To Stack A ( Steam Drum
{ e 200 psi Steam /2
Gas Blower ) ' (o

=
A
TR, . Heat Exchanger

Condenser " Catalytic | \ Fuel Circulating S
\., { Recombiner i = Pump

  
 
  
 

  
 
      
     
 

. Rod Drive Mechanism 120 gpm Cp |
Fission Gas G
Adsorber

 

  
   
  
 
 
 

    
   

. ) Main Steam Valve .’
O i I Sz
(-20°C) Storage Tanks 1000 psi) 7§} 7
Condensate Weighing Tanks ydrogen-Oxygen E}‘ Pressure Vessel Generator |
Flame Recombiner ‘ -‘ ; ( ( >
Hydrogen, Oxygen, Sl ‘ & \ ’ ‘
Steam, and Fission Gases NN | 2 N
Condener . (7 Q\\/' Neutron Absorbing™~& D,0 Cooler
\ ’f,“:& Plates Turbine
\ PStReactor Core D,0 Circulating Pump

  
 
     

\\
N

 
 
 

High Pressure N
Reducing Valve S .
." ‘. 1000 psi 175°C 15 pSI

  

" Fuel Drain
Valve

1 gpm
Condensate
Return Valve
; c(Close To 1000 psi
oncentrate 15 psi
’l Fuel Feed Pump Fuel) \" D,0Feed Pump

Fiac. 7-4. Schematic flow diagram, HRE-1.

of forged steel, 39 in. in inside diameter, with a 3-in.-thick wall. A 24-in.,
1500-psi standard ring-joint flange at the top of the vessel permitted
removal of the inner core.

In order to limit thermal stresses and to reduce corrosion of the steel
vessel, the reflector temperature was regulated near 350°F. About 50 kw
of heat conducted from the fuel core to the reflector liquid was removed. by
circulating the heavy water with a 30-gpm canned-rotor pump through a
reflector cooler which acted as a boiler feedwater preheater. A jet was
located in this high-pressure circulating loop, the suction of which drew a
continuous stream of gas from the vapor space above the reflector to a
catalytic recombiner so that the concentration of deuterium and oxygen
gases in this vapor space could be kept below explosive limits. . |

Some measure of nuclear control was obtained by changing the level of
the reflector. The level could be lowered by draining liquid through a valve
to storage tanks, or raised by starting a feed pump. This pump, which
employed a hydraulically driven diaphragm with check valves, had a
capacity of approximately 2 gpm against 1000 psi1. Its intake was connected
to supply tanks at atmospheric pressure, located below the reactor. These
tanks also served as degas chambers for the reflector liquid which was dis-
charged from the pressure vessel. Helium, water vapor, and D2 and O3
352 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [CHAP. 7

gas liberated here passed upward through a condenser to a small low-
pressure catalytic bed where the O2 and D4 gases were recombined to D2O.
Cold traps operated at —20°F were included in the reflector-system vent
lines to prevent the loss of D20 or its contamination with H2O vapor.

7-3.4 The fuel off-gas system. When the reactor was operating, with-
out copper sulfate in the fuel solution, the fuel solvent (H20) was decom-
posed by the energy of fission to yield a stoichiometric mixture of hydrogen
and oxygen gas at a rate of 0.28 ¢fm (10 cfm at STP). In addition to this
large volume of decomposition gases, there was also produced a very small
volume (20 cc/day) of intensely radioactive fission gas. If these gases had
not been removed and replaced by more liquid, excessive pressures would
soon result, and since virtually all of this gas was liberated within the core,
the displacement of fuel solution by the gas would make it impossible for
the chain reaction to continue. For this reason, the gas was continuously
separated from the fuel in the core by injecting the main circulating stream
tangentially near the equator of the sphere, which caused the fluid to rotate
and form a vertical cylindrical vortex approximately 1/4 in. in diameter.
The centrifugal action of the rotating fluid served to separate the decom-
position and fission gases from the liquid to the vortex, the axis of which
was aligned with the fuel outlet. A nozzle with a central opening in the
fluid outlet allowed the removal of gas from the vortex. This gas plus
about 0.8 gpm of the fuel solution was passed through the outer annulus of
a countercurrent, concentric-tube heat exchanger, which was partially
cooled by 0.8 gpm flow of fresh makeup liquid being pumped back to the
core. The cooled mixture of gas and liquid was then throttled through a
valve into a gas separator which was connected to the fuel-solution storage
tanks. The gas-steam mixture rose from the gas separator to a condenser
immediately preceding a flame recombiner, so that the gases leaving the
condenser were combustible and reunited to water in the flame of the re-
combiner shown in Fig. 7-4.

The flame recombiner is best described as an oversized Bunsen or Meeker
burner enclosed in a water-jacketed cylinder. In the HRE-1 no attempt
was made to use this 40 kw of high-temperature heat, although in larger
scale reactors this energy might be used to superheat steam about 70°F.

The exit gases from the flame recombiner contained only fission products
and small amounts of unrecombined hydrogen and oxygen. They were
passed through a catalytic recombiner which contained a platinized alumina
catalyst to eliminate the traces of hydrogen and oxygen. Also, the catalytic
bed was used to react the entire gaseous output at low reactor powers when
insufficient gas was being liberated to maintain a steady flame at the burner
of the flame recombiner. The catalytic bed was followed by a condenser
and cold traps to prevent the loss of water from the system. The gas stream
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 353

at this point was composed mainly of excess oxygen plus the highly active
fission gases, mainly xenon, krypton, and their decay products. The ac-
tivity of these gases was many orders of magnitude greater than the activity
which can be discharged directly to the atmosphere without the construc-
tion of a very expensive stack; therefore it was desirable to provide some
inexpensive means of storage for the dissipation of the radioactivity. This
was accomplished by passing the gases through cold traps to remove
moisture and adsorbing them onto water-cooled activated-carbon beds
which were buried underground outside the reactor building. It is estimated
that the equilibrium activity of the gases held on the carbon bed was 400,000
curies. *

The adsorption efficiency of the charcoal, even at ground temperature,
was good enough to prevent a discharge of activity greater than a few
curies per day. However, even this amount of activity had to be diluted
so that the atmospheric concentration at ground level was not greater
than 10713 curies/cc of air. Dilution was accomplished by feeding the
active gas into a 1000-cfm ventilating air stream from the reactor shield
and then to a 100-ft-high stack. During operation the gaseous activity
inside the stack barely exceeded inhalation tolerance.

7-3.5 Fuel concentration control. The condensate which was removed
from the vapor-gas mixture upstream of the recombiner was returned either
to the fuel storage tanks or to weighed holding tanks. The accumulation
of water in the holding tanks provided a means of increasing the concen-
tration of fuel in the storage tanks underneath the reactor. Since fuel was
pumped continuously from the storage tanks to the high-pressure system
by means of a duplex-diaphragm type pump at a rate of 0.8 gpm, it was
possible to vary the concentration of the fuel which circulated through
the reactor. Figure 7-5 shows how the core temperature varied with fuel
concentration, in g/kg H20O. Furthermore, since the operating temperature
of the core was controlled by the fuel concentration as shown in this figure,
the operator had a convenient means of adjusting the solution temperature
to the desired level. This feature of variable concentration was employed
during startup of the reactor when the concentration had to be changed
by large amounts, and also during steady operation for small changes in
temperature. When sudden dilution of the fuel was desired, as in the case
of a complete shutdown, the condensate holdup tanks were quickly emptied
through a drain valve into the fuel-storage tanks, or condensate was pumped
directly into the core.

The steep slope of the curve in Fig. 7-5—i.e., the large negative tempera-
ture coefficient—was a feature which was extremely important from safety

*Qne curie equals 3.7 X 1010 disintegrations per second.
354 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

 

T T T T T T T T

40 |—
39 —
38 |— —
37 }— —
36— .
35—
34—
33—
32—
31—

30—

Fuel Concentration, g U235/kg H,O0

  

29 —
Theoretical s - Experimental

  

28 |—
27 |—
26—

25—

 

 

L

O 25 50 75 100 125 150 175 200 225 250
Temperature, °C

 

24

Fi1a. 7-5. Dependence of critical fuel concentration on temperature in HRE-1.

and power-demand standpoints. For instance, a temperature rise of only
15°C was necessary to overcome a reactivity increase of 1%, an amount
considered very dangerous in most solid-fuel reactors.

7-3.6 Power removal. The steam generated in the fuel heat exchanger
was fed to a conventional multistage condensing turbine generator rated
at 312 kva. With the reactor operating at 1000 kw and 250°C, a sufficient
quantity of steam at 200 psi was produced to generate about 140 kw of
electricity. Steam leaving the heat exchanger was first passed through a
time-delay drum with a radioactivity monitor at the inlet and a quick-
closing valve at the outlet to prevent the escape of activity into the turbine
system in the event of boiler tube failure. Small feed pumps returned the
condensate from the turbine to the boiler.

Upon increase in generator load, the turbine governor opened the
turbine throttle valves, increasing the steam demand, lowering the steam
7-3] THE HOMOGENEOUS REACTOR .EXPERIMENT (HRE-1) 395

pressure and temperature, and reflecting itself into increased cooling of
the uranium solution, which automatically increased the reactivity of the
core and completely compensated for this increased load.

7-3.7 Internal-recombination experiments. The use of copper dissolved
in the reactor fuel for the complete recombination of radiolytic gas was
successfully demonstrated in the HRE-1 [14]. Copper ion was added as
copper sulfate on four occasions, increasing the copper concentration to
10, 25, 75, and 150% of that necessary for complete recombination (i.e.,
6.6 g CuSOy4/liter) in a static system at 250°C, at 1000 psig total pressure,
and at a uniform power density of 20 kw/liter. The investigation was
conducted at temperatures from 185 to 260°C, at pressures from 765 to
1200 psig, and at power levels as high as 1600 kw. In the main, the copper
behaved as expected from static bomb tests. The highest power level for
which all the gas was internally recombined was 1350 kw. In the course of
all copper experiments, including 350 hr of operation at the highest copper
concentration, 0.062 molar, no deleterious effects due to the presence of
copper were observed.

7-3.8 Nuclear safety. Although operating experience later verified early
predictions of the inherent safety of this reactor, at the time of design it
was considered judicious to incorporate conventional safety devices in the
reactor for protection against potentially dangerous situations which might
arise during low-power operation and until the dynamic stability had been
demonstrated by experiment. Safety measures in the order of their auto-
matic action as installed to limit reactor power or power doubling time
were:

(1) Two magnetically coupled safety plates, worth about 45 g of uranium,
which by falling in 0.01 sec caused the reactor temperature to be lowered
from 250°C to approximately 243°C.

(2) Dumping of the reflector.

(3) Dilution of the fuel (this was the normal shutdown procedure).

(4) Stopping of ‘the steam extraction by closing either a steam valve or
the turbine governor.

(5) Draining the fuel solution to the noncritical-geometry tanks below
the reactor.

Experiments demonstrated that the reactor was extremely fast-acting
with respect to limiting power surges and led to the belief that mechanical
control devices were unnecessary [15]. These consisted of a series of kinetic
experiments in which the power responses to reactivity increases were
observed. First, the entire range of normally available reactivity increases
—fuel concentration, rod withdrawal, and reflector level—was tested with
initial power levels as low as 10 watts and reactivity rates up to 0.05% per
356 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP.7

second. Then, in order to provide more drastic tests, the main fuel circu-
lating pump was stopped, the reactor was maintained at a low power and
at high temperature, but the heat exchanger was cooled about 100°C.
When the pump was restarted, the cold fuel from the heat exchanger was
rapidly injected into the core, producing a rate of reactivity increase of as
much as 0.8% k. per second. The results of two experiments in which
only the initial powers differed are shown in Fig. 7—6; the power increased
in a period as short as 35 msec, reaching a peak of 10 Mw in 1 sec, and then
approached the equilibrium power demand within 0.2 sec after the peak.
The calculated pressure rise associated with the peak power of 10 Mw
was only 5 psi.

In these experiments the worst combination of circumstances was im-
posed on the reactor. It was successfully demonstrated that the HRE-1
was sufficiently stable to withstand nuclear transients greater than those
expected from operating errors.

7-3.9 Leak prevention. A major problem in the HRE-1 was to main-
tain absolute leaktightness in all components. The radioactivity of the
solution during operation was about 30 curies/cc. Twenty-four hours
after shutdown the activity was about 3 curies/cc. With these high
activities, the total leakage from the system had to be kept below 1 cc per
day. A much better performance than this was attained through the use
of canned-rotor pumps, double tube-sheet exchangers, and bellows-sealed
valves.

All welded joints were made with extraordinary care and tested by sev-
eral nondestructive methods before being approved for use. Flanged
joints were assembled with stainless steel ring gaskets of oval cross section

TABLE 7-3

HRE-1 ConsTtrUCTION CoOST SUMMARY*

 

 

 

Total 9% of total
Building $ 300,000 27.5
Fuel and reflector equipment and piping 420,000 38.2
Instrumentation 190,000 17.6
Shield 110,000 9.7
Power system 80,000 6.9
Total construction cost for HRE-1 $1,100,000 100.0

 

 

 

*These costs include material, labor, and allocated overhead.
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 357

 

 

 

 

 

 

 

T I —
103 |— ]
3 102 |— —
X
o Exp 25
‘;” °
S 10 |— StartPump =
. (Core Temp 177°C) |
° I Traced From Brush
S 1 — II Recorder —
oc / ——=—- Inferred From Exp 25
Start Pump /I
10°1 |~ (Core Temp ’I —
170°C) /
o 7
alEeeT ) ) | ]
0 1 2 3 4 5 6 7 - 8 9
Time, sec

Fia. 7-6. Power response of HRE-1 during reactivity increase of 0.8%, k./sec.

and each joint contained a leak detection device for constant monitoring
(see Fig. 7-13). In addition, the ventilating air which flowed through each
equipment compartment was monitored constantly with gamma-radiation
detection devices. Although several leaks were experienced in the startup
phases, no leakage was found during the final 12-month period.

7-3.10 Shielding. As with all nuclear reactors, personnel had to be
protected from the high radiation levels which existed in the vicinity of
the reactor core. In the HRE-1 this protection was provided by a 7-ft-
thick shielding wall of high-density concrete. An increase in the density
of the concrete from 2.3 to 3.5 g/cc resulted from the use of barium sulfate
ore as the aggregate material. The shield was a departure from most reactor
shields in that it was constructed of loosely stacked block with only the
outer 16-in. layer of blocks being mortared. |

7-3.11 Construction cost. The total construction cost of the reactor
was $1,100,000, which did not include the cost of fuel and heavy water.
Table 7-3 is a summary showing the cost of various parts of the system.

7-3.12 Maintenance. The maintenance of homogeneous reactors is
greatly complicated by the radioactivity of the parts, and in the HRE-1
it was often necessary to decontaminate equipment and to provide tem-
porary protective shielding before repair work could be done. In most
cases the repair work had to be done with long-handled tools, which made
the job even more difficult. By the use of decontamination, shielding,
and extension tools, it was possible, however, to make a number of major
358 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP.7

repairs in radiation fields as high as 2000 r/hr, without exposing personnel
beyond accepted tolerances. In this regard remote viewing devices such as
mirrors, binoculars, and a Polaroid Land camera were found to be invalu-
able tools. The main circulating pump was replaced or repaired three
times under such conditions and the diaphragm feed pumps twice. In fact,
in no case of a breakdown was it impossible to make the necessary repairs.

7-3.13 Dismantling the HRE-1. After final shutdown-of the HRE-1
the reactor was decontaminated in preparation for disassembly. Over a
period of 30 days, starting with activity levels of the order of 1000 r/hr, the
activity was reduced sufficiently to permit dismantling of the system with
long-handled tools [16]. |

The decontamination treatment consisted of repeated washing alter-
nately with 359, HNOj3 and aqueous solutions of 109, sodium hydroxide,
1.59, sodium tartrate, and 1.59, hydrogen peroxide. The over-all decon-
tamination factors were 22 to 25, including decay, but the factor for
decontamination with a single reagent was only between 1 and 2.25.
Large amounts (of the order of 10° curies) of cerium, zirconium, barium,
lanthanum, strontium, niobium, and ruthenium were removed. The
significant contaminants remaining were zirconium and niobium, which
were bound in the oxide film. Although these could have been removed
from the system by descaling the oxide corrosion film, which would have
given a further decontamination factor of approximately 100, such a
treatment would have made it impossible to determine that no significant
corrosion had taken place during nuclear operation.

7-3.14 Critique of HRE~1 [17]. After the HRE-1 was put into opera-
tion, personnel associated with the Homogeneous Reactor Project were
asked to suggest ways in which the design and construction of the reactor
and the associated development program might have been improved.
More than one hundred specific design changes were recommended, many
of which related to the difficulty of operating the reactor on a continuous
basis and the need for repairing and/or replacing faulty equipment.
Among the many possible improvements recommended were the use
of high-pressure catalytic recombination; external gas separation (Le.,
nonvortex core flow); spacing of equipment for easier inspection and main-
tenance; shielded, waterproof instrument lines; instrumentation for more
accurate fuel accountability; improved feed pumps; and provision for ob-
taining meaningful corrosion data.

The general conclusion reached was that the reactor construction schedule
(16 months) was too accelerated to allow good design and construction
practices to be put fully into effect. The component development and
testing program, in particular, suffered by the short time schedule in that
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 359

pieces of equipment such as valves and pumps were not completely tested
before being used in the reactor.

The thorough analysis of the HRE-1 design provided a basis for the
subsequent design of the HRE-2, in which many of the suggested improve-
ments were incorporated. These included: (a) greater accessibility of
equipment, (b) provision for flooding cells, (¢) better shield construction
with metal walls to permit decontamination, (d) more accurate means for
measuring fuel inventories, and (e) elimination of screw-type fittings.

7-3.15 Summary of results. As a result of the operation of the HRE-1
and of the extensive experimental program conducted with it, several un-
certainties were resolved regarding the nuclear and chemical behavior of
aqueous homogeneous reactors at the high temperatures and high pres-
sures required for power generation. Included were demonstrations of (1)
a remarkable degree of inherent nuclear stability, a result of the very large
negative temperature coefficient of reactivity, (2) the elimination of the
need for mechanical control rods as a consequence of this inherent stability,
(3) flexibility and simplicity of fuel handling, (4) stability of the fuel
(5) the ability to attain and maintain leaktightness in a small high-pressure
reactor system, (6) the safe handling of the hydrogen and oxygen produced
by radiation decomposition of the water, and (7) the direct dependence of
reactor power upon turbine demand.

7-4. THE HomoGENEOUS REAcCTOR TEsT (HRE-2)*

7-4.1 Objectives. The objectives of the Homogeneous Reactor Test
(HRE-2) are (1) to demonstrate that a homogeneous reactor of moderate
size can be operated with the continuity required of a power plant, (2) to
establish the reliability of engineering materials and components of a size
which can be adapted to full-scale power plants, (3) to evaluate equipment
modifications which will lead to simplifications and economy, (4) to test
simplified maintenance procedures and in particular underwater mainte-
nance, and (5) to develop and test methods for the continuous removal of
fission and corrosion contaminants. |

7-4.2 Reactor specifications and description [15,18]. The bases for
the design of the reactor, which are summarized in Table 7-4, were selected
~early in 1954 and were intended to take fullest advantage of the progress
in chemistry, materials and component development, and the experience

*Based on information supplied by S. E. Beall and S. I. Kaplan and reports by
members of the Oak Ridge National Laboratory as noted.
360 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7
with the HRE-1. In order that the objective of a significant test of the
engineering feasibility of a large power station be satisfied, it was necessary
that the physical size of the reactor and its auxiliaries be increased appre-
ciably beyond that of HRE-1. To hold the cost within reasonable bounds
for an experiment, it was decided to limit the power output and thus the
expense for heat-removal equipment, and also to install the reactor in the
building which had previously housed HRE-1, permitting the use of many
of the existing site facilities. The size of the reactor core represents a
compromise between two objectives, attainment of high specific power
required for economy in a large plant and evaluation of the fabrication
and durability of a zirconium-alloy core tank. The power output was,
therefore, set at 5000 kw (heat) with the possibility of a maximum
10,000 kw, and the core diameter at 32 in. Although these together result
in a low specific power of 17 kw/liter in the core at 5000 kw, this was con-
sidered acceptable, since operability at a relatively high specific power of
30 kw/liter had been demonstrated in the HRE-1. Another factor affecting
the selection of core diameter was the opinion of fabricators that current
technology would be exceeded for a zirconium vessel larger than 32 1in.

TABLE 74
- DusigN Bases ror HRE-2

 

 

Power, heat 5000 kw
Temperature, core outlet 300°C
Pressure 750 psi in excess of vapor pressure (see text)

Core diameter

Core solution

Blanket

Fuel circulation rate

Blanket circulation rate

Core flow pattern

Core construction material
System construction material
Radiolytic-gas removal
Radiolytic-gas recombination

Fission-product-gas disposal
Control

Normal

Safety

 

2000 pst maximum total pressure

32 in. |

U02804 in D20 (~10 g U235 per liter)

D20

400 gpm

230 gpm

Straight-through

Zircaloy—2

Type-347 stainless steel

External pipeline separator

Low-pressure system: platinized alumina
catalyst

High-pressure system: CuSOy4 in solution

Decay on activated carbon

Variable solution concentration
Temperature coefficient

 

 
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 361

The increase in fuel temperature from 250°C in the HRE-1 to 300°C was
based upon the more favorable corrosion resistance of both stainless steel
and Zircaloy—2 to dilute uranyl sulfate at the higher temperature and the
possibility of improved thermal efficiencies. Also the temperature at which
the two-liquid phase region appears is higher for the more dilute fuel, per-
mitting this increase.

TABLE 7-5

HRE-2 DEsiGN PARAMETERS

 

 

 

Core Blanket

Power, heat, kw 5000 220
Pressure, psi 2000 2000
Vessel

Inside diameter, in. 32 60

Thickness, in. 5/16 4.4

Material Zircaloy—-2 Stainless-steel-clad

carbon steel

Volume, liters 290 1550
Specific power, kw/liter 17 0.14
SOlutiOIl UO 2SO 4—D 20 D 20
Uranium concentration, g of

U235 per kg D20 9.6 0
Circulation rate, gpm 400 230
Inlet temperature 256°C 278°C
Outlet temperature 300°C 282°C
Volume of gas generated,

ft3/sec at STP 0.96 0.013

ft3/sec at 2000 psi, 280°C 0.015 o

 

 

 

 

 

The design pressure of 2000 psi resulted from the necessity for an over-
pressure on the system to prevent boiling, to reduce the volume of gas in
the core, and to increase the efficiency of the copper catalyst. A maximum
pressure of 750 psi was thus provided in excess of the vapor pressure of
water of about 1250 psi at 300°C.

The thickness of the blanket of 14 in. between the core and pressure
vessel was selected as a compromise between neutron leakage and the use
of a simple pressure vessel of dimensions within the means of standard
fabrication techniques.
362 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [CHAP. 7

The fuel circulation rate of 400 gpm was based upon heat-removal
requirements and the availability of suitable pumps. Pumps capable of
this output had been successfully operated under comparable condi-
tions.

The vortex-type flow through the core of HRE-1 was abandoned in
favor of straight-through flow because hydrodynamic experiments demon-
strated the pressure drop of the former to be excessive for larger cores.
As a result, the extraction of radiolytic gas directly from the core was
excluded, and an external gas separator in the exit pipe from the core was
used to separate gases produced in the core. The amount of these gases
depends on the operating temperature and pressure and the CuSO4 con-
centration. The excess gases are recombined at low pressure but by means
of a platinized alumina catalyst instead of combustion in a flame-type
recombiner as in the HRE-1.

In stagnant fuel lines external to the core, the radiolytically generated
oxygen is insufficient to replace the oxygen consumed in reacting with the
stainless steel to form metallic oxides. Since an oxygen deficiency induces
hydrolytic precipitation of the uranium, approximately 2 liters/min (STP)
of gaseous oxygen are injected into the fuel feed stream, maintaining a
concentration of approximately 500 ppm in the high-pressure system.
The disposal of the fission-product gases which are stripped from the fuel
by the excess oxygen and radiolytic gas is accomplished by adsorption
and subsequent decay on beds of activated carbon. Experience with this
method of disposal was completely satisfactory in the HRE-1.

To reduce the xenon content of the core solution and minimize the
catalyst-poisoning effect of fission-product iodine on the recombiners, an
iodine-absorption bed of silver-coated wire mesh was installed in the off-
gas line between the fuel dump tank and the recombiners. The bed removes
over 99% of the iodine passing through it, and W111 retam the absorbed
iodine at temperatures up to 450°C.

Type—-347 stainless steel was designated as the material of construction
for all of the reactor except the Zircaloy-2 core vessel. Titanium was
chosen to reinforce certain points of high turbulence, such as the pump
impellers and the gas separator. Previous corrosion and welding experience
with this grade of stainless steel, also used for HRE-1, has been excellent.

Based on HRE-1 experience, it was decided to ehmmate mechanical
control devices and depend entirely on varying the fuel concentration for
shim control, on temperature coefficient for transient nuclear changes, and
on dumping the fuel solution for rapid shutdown when required.

The design parameters of HRE-2 are summarized in Table 7-5.

The flow diagram for the reactor is illustrated in Fig. 7-7. Since the fuel
and blanket systems are virtually identical—the only significant differences
being the absence of a blanket iodine separator and the larger vessels
363

7-4]

THE HOMOGENEOUS REACTOR TEST (HRE-2)

‘G-HYUH JO 199USMoly “L-L D14
SA|DA

19ysupaj A N

               

 

Jojpiodpag .

A T

o.”

__1ojpiodpAg

   

s)up) dwn

     

    
 
 

3A|PA dwing i 3AIPA dwng
) 183|007 ¥ T, 18]00D) * e
% 25 - l & ‘ . fi

~

Joypindag
juswuibyu]

 

wajsAg
9INSS3Id MOT
{9yupjg o}

Jojoindag Jauiquoday

P s00)

   

PASNSNNNN

        

 

R
N\

 

, ) i>- - 10jpI8ud0)

’” 034G

    

abpuio}g
9jpsuspuod)

 

SO

L
<&
<

AN

 

 

siaqJospy
§onpouy uoissiyg

   

 

pag uoydiosqy
dUIPO| PUD JBUIqUWOD3Y

 
 

abuny

  
   

   

 

ajpsuspuo) 77777

 

it o1 wos uoyn|os joxunig puc 310> waisAg 4 % oy
1qint o4 IS pinbry pup sooy VL SPO-HO 13qup|g & '
00 [ pois oL €
364 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

necessary to accommodate the greater volume of blanket fluid—the entire
blanket system is not shown. The fuel system is described as follows.

In the high-pressure system the fuel solution is pumped into the bottom
of the reactor core at 256°C and is heated to 300°C as it proceeds upward
to the outlet pipe. At the top of the outlet pipe is attached the pressurizer
in which condensate is electrically heated to a maximum temperature of
335°C to produce 2000-psi steam. The fuel flows past the pressurizer to
the gas separator, where directional vanes cause the fluid to rotate suffi-
ciently to separate the radiolytic gas (D2 and O32), the excess Oz, and fission-
product gas. The separated gas forms a vortex along the axis of the pipe
and is bled to the low-pressure system. The reactor solution continues
from the gas separator to the U-tube primary heat exchanger, where it is
cooled from 300°C to 256°C by transferring heat to the boiler feedwater
surrounding the tube bundle. The 244°C, 520-psi steam produced on the
shell side of the heat exchanger is bled partially to the small (345-kva
output) turbine generator remaining from the HRE-1 and partially to an
air-cooled steam condenser. The uranyl sulfate solution flows next to the
intake of the 400-gpm canned-motor circulating pump and thence is pumped
to the core for reheating. The blanket fluid follows an identical cycle at a
flow rate of 230 gpm. This lower flow rate was based on a pump of the same
horsepower but designed to circulate a thorium oxide suspension, rather
than pure D2O.

The gases and some entrained liquid removed by the gas separator are
transferred to the low-pressure system through a “‘letdown heat exchanger,”
a jacketed pipe which cools the gas-liquid mixture to 90°C. A valve
downstream of the heat exchanger throttles the gas-liquid stream to
atmospheric pressure. The mixture then discharges into the “dump”
tanks, which have sufficient capacity to hold all the reactor liquid. An
evaporator built into the dump tanks provides continuous mixing and,
more important, steam for dilution of the deuterium and oxygen below
the explosive limits. The gas-and-steam mixture flows upward through the
iodine bed to the catalytic recombiner, in which the deuterium and oxygen
react on a bed of platinized alumina pellets to form water vapor. The
heavy water is condensed by the shell-and-tube condenser following the
recombiner and normally flows back to the dump tanks. However, the
water may be diverted to weighed storage tanks in case it is desired to
change the concentration of the fuel solution. Water which is returned
to the dump tanks is mixed with the excess fuel solution stored there (ap-
proximately 25 gal) and then fed to the intake of a sealed-diaphragm
injection pump, which returns the liquid to the high-pressure circulating
system at a rate of about 1 gpm, thus constantly replacing the liquid
removed via the gas separator.

The small volume of intensely radioactive fission gas plus the excess
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 369

     

&Com‘ml Room Floor

   

ter

| Barytes Sand and Wa

    
  
    

54 f0in. A<
pe Colum

-
7~

~

   

1-in. Dia Reinforcing Bars (4 in. OC)

Bulkhead
| F*16-in. Pi

 

  

 

  
 
   

Fi1a. 7-8. HRE-2 shield and vapor container.

| ~*~Steel Plate |,

)
i

i
i

 

Seal Sheet (Between Plug Layers) -

Top Plug Layer (Barytes Concrete)
Stainless Steel Cladding

Lower Plug Layer (Barytes Concrete)y -~ X

Barytes Concrete
366 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP.7

oxygen remaining after condensation of the re-formed heavy water is
dried in cold traps at —23°C and sent to the beds of activated carbon for a
period of decay. Gas leaving the bed is diluted with 1400 c¢fm of air and
discharged to the atmosphere from a 100-ft stack.

Samplers are provided to secure small quantities (5 ml) of the fuel and
blanket liquids from the high- and low-pressure systems for chemical
analysis. These units are located in bypass lines; material circulated
through them is trapped by closing the sampler inlet and exit, after which
the contents are discharged into a portable container through a drain
valve.

All the primary reactor equipment is located in an underground, box-
like, steel tank called the “shield pit,” shown in Fig. 7-8. The design of the
shield pit was influenced by several factors, including a requirement for
accessibility and flexibility because of the experimental nature of the in-
stallation, provision for complete containment of the contents of the re-
actor should a leak develop or should the pressure vessel or heat exchangers
rupture, efficient utilization of the space within an existing structure, and
capability of flooding with water for maintenance or replacement operations.

The reactor shield pit occupies the center high-bay area of the building
and is constructed of 3/4-in. welded steel plate reinforced in such a manner
that an internal pressure of 30 psi will be contained. This pressure corre-
sponds to the instantaneous adiabatic release of the entire contents of the
reactor system. The chemical processing cells, each 12 ft wide by 25 ft
long, are designed similarly.

The upper surface of the blocks forming the roof of the shield pit is at
ground level. This roof is made of high-density concrete 5 {t in total thick-
ness and consists of two layers of removable slabs with a completely welded
steel sheet sandwiched between the layers and extending across the top of the
pit to form a gastight lid. The roof blocks are anchored to the girders and
supporting columns by means of a slot-and-key arrangement, shown In
Fig. 7-8. The vertical columns are embedded in a concrete pad which is
3 ft thick and is heavily reinforced with steel.

The wall between the reactor pit and the control area is a hollow box
5% ft wide, constructed of 1/2-in. steel plate welded to the north side of
the reactor shield tank. It is filled with high-density barytes, sand, and
water. The use of the fluid shield between the reactor and control-room
areas allows flexibility in the locations of service piping and instrument or
electrical conduits. All lines leaving the reactor tank are welded into the
shield wall; conduits are connected into junction boxes inside the pit with
gastight seals on the individual wires. |

The design is such that nowhere outside the shield will the radiation
dosage exceed 10 mrep/hr when the reactor is at 10 Mw. For the purpose
of decreasing the neutron activation of equipment inside the pit, the re-
THE HOMOGENEOUS REACTOR TEST (HRE-2) 367

7-4]

"Ma1A ue[d ‘“yusweduriie jusuodwod Z—qYH ‘6-L PLI

 

 

Jasuapuon)
xn|joy

Jasuapuo)

jup} dung @ (+9xupig) juoj duing % Bm:_%_._flwmmv J05H

19nQO jajup|g 1abupydx3 jpol 434nQ |19n4

Aquol
dwnq ja)up|g

sdoi] pjod
19yup|g
10dD3Y

o)
o
3
>
mum.
37
Q A
® 0
2 3
@ o
T35
o
=

1491 sdpu o7 |an
SHpSUSpUOD 3yuplY 19znssaly LPISS 150

 

 

 

lasuapuon)
Jauiquoday |any

SOA|DA SOA|DA

 

 

 

sdung sdwnyg
Pa34 i3jup|g paag [an4

 

 
368 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

 

Fia. 7-11. Reactor tank with shielding plugs in position during hydrostatic test.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 369

actor vessel 1s surrounded by a thermal neutron shield consisting of a steel
tank with a 2-ft-wide annulus filled with boron ore and water.

The arrangement of reactor components within the main reactor pit is
shown in Fig. 7-9. The reactor is near the center of the pit, enclosed by
the 2-ft-thick thermal shield. To the left of the reactor is all the fuel-
system equipment, and to the right is the blanket equipment.

Most of the high-pressure components are located close to the pressure
vessel. The pressurizers are positioned directly above the reactor; the gas
separators are in the S-shaped outlet lines to the steam generators on the
south side of the pit. The areas to the left and right of the reactor in the
center bay are reserved for future equipment modifications or additions.

To the far left is grouped all of the fuel low-pressure equipment, including
dump tanks, recombiner, condensate tank, and cold traps. These compo-
nents are assembled on a rigid structural-steel frame; the blanket low-
pressure equipment is to the far right.

Insofar as possible, valves are situated close to the control-room wall
so that air lines and leak-detector lines can be kept short. The arrange-
ment of the valves is such that all flanges can be easily disconnected from
above.

7—4.3 Schedule of construction [19,20]. Construction of the HRE-2
was started in July 1954, immediately after the dismantling of HRE-1.
The initial step was the excavation of a large hole beneath the building
(which had previously housed the HRE-1) for the large rectangular steel
tank (60 ft long, 305 ft wide, and 25 ft deep) which contains the reactor and
its associated equipment. Figure 7-10 shows this tank at approximately
509, completion.

The north wall of the reactor tank, seen to the left in Fig. 7-10, 18 com-
mon to both the reactor tank and the control-room area, and it is through
this wall that the many service, instrument, and electrical lines which must
interconnect these two areas pass. The next step in the construction was to
install the approximately 600 lines which penetrate this 53-ft-thick wall.

Figure 7-11 shows the tank after the complete roof structure had been
assembled and welded closed as it was prepared for a hydrostatic test. In
this test the tank was filled with water and then pressurized to give the
equivalent of a 30-1b internal pressure at all points within the tank. Strain
gauges were attached at many points so that the complicated stress pattern
could be studied in some detail for assurance that the tank was safely
within design limits.

While the reactor tank and control-room areas were being constructed,
reactor equipment was being procured and constructed at several places.
Much of the equipment required in the low-pressure system had been
inspected and tested when the container was completed. In November
EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

370

 

main bay.

Ing,

7-12. View of HRE-2 build

Fia

 

 

 

 

 

 

 

 

 

 

 

 

   

 

ey

2\

4

HRE-2

’

ist’s concept of homogeneous reactor test

7-13. Art

Fia
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 371

  
 

  

Dzo
Supply

  
  

To Oxygen System
(3000 psi)

    
 

Vent to Air Control Area Reactor Cell

      
      
 

Pressure Alarm

 
 
  
    
   

Weld

  

Sight Gage
For Liquid
Level Control

   

Seal Weld Test

Point Line

        
 

  

P
ressure Gage North Shield Flanged

Joint

   

Legend:
=—pJ== Valve Normally Open
=P Valve Normally Closed

 

Sectionalizing Headers

Fi1a. 7-14. HRE-2 leak detection system.

Pressurizer Vessel

1955, these parts and the thermal shield surrounding the reactor pressure
vessel were installed. Figure 7-12 shows the reactor core and pressure-
vessel assembly just before installation in January 1956.

The heat exchangers were subsequently installed and followed by the
main circulating pumps so that the high-pressure piping which connected
the pumps, heat exchangers, and pressure vessel could be attached. This
work occupied most of the months of February and March. Construction
of the reactor was completed in May 1956. TFigure 7-13 is an artist’s
concept of the completed reactor.

7-4.4 Nonnuclear testing and operation. Pretesting, operation of the
reactor as a nonnuclear facility, and a lengthy flange-replacement job
occuplied the period from completion of construction in May 1956 to De-
cember 1957. A chronological summary of the events associated with the
nonnuclear operation of the reactor during this period is shown in Table 7-6.

From Table 7-6 it can be noted that preoperational testing of HRE-2
was Interrupted by stress-corrosion cracking difficulties, which were caused
by chloride ion contamination in the stainless steel tubes that are used to
detect and prevent leakage of radioactive solution from flanged joints [20].
Figure 7-14 shows how an individual leak-detector line is attached to the
groove of the ring-joint flange. The tubes from all the flanges terminate
at a valve header station in the control room. Normally this system is
kept pressurized with water to a pressure of 300 to 500 psi above the system
pressure. A leak in any flange results in leakage of water from the header
and a loss in pressure, which actuates an alarm at a fixed level above the
fuel or blanket pressure. This is normally a sensitive and satisfactory means
 

 

 

 

August 1956

August 1956—
October 1956

October 1956—
November 1956

December 1956—
January 1957

January 1957-
March 1957

April 1957-
August 1957

August 1957-
September 1957

September 1957-
October 1957

November 1957—
December 1957

 

372 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7
TABLE 7-6
SuMMARY oF HRE-2 NoNNUCLEAR OPERATION
Period Test or event
May 1956 3000-psig hydrostatic test of high-pressure system and
750-psig test of low-pressure system
June 1956- Cleaning of piping systems with 3%, trisodium phos-
July 1956 phate followed by 5%, nitric acid and initial operation
of pumps. Tests for dump-tank entrainment and
efficiency of catalytic recombiner
July 1956— Initial tests of equipment removal and underwater

maintenance

Flushing of flange leak-detector tubes to remove chlo-
ride contamination

Initial nonnuclear operation of reactor at 280°C and
2000 psig. Thermal cycling of flanged joints

Removal of typical flanges for metallographic examina-
tion to detect possible stress-corrosion cracks. Further
tests of remote-maintenance tools -

Further operation of the reactor with water and with
depleted uranium at various conditions of tempera-
ture and pressure

Removal, inspection, and replacement of flanges and
leak-detection tubing -

Recleaning of system with trisodium phosphate and
nitric acid solutions and hydrostatic testing

Final operation at design conditions with water and
with depleted uranium. Final leak test of reactor
piping with radioactive tracers

Final insulation of reactor piping; installation of new
refrigeration system and new iodine absorption bed,
followed by nonnuclear operations with heavy water.
Criticality achieved December 27, 1957

 

 
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 373

of preventing the leakage of radioactive liquid from the reactor (the leak-
detector fluid leaks out instead) and detecting the leaks when they do occur
(by measuring the pressure or volume loss of the leak-detector fluid).
Volume changes in the header can be read to 41 cc from a graduated scale
next to each header sight-glass. By observing level changes at regular
intervals, noting which lines are isolated from the header, leaks of less than
2 cc per day can be detected.

Since this system is a secondary portion of the reactor, the leak-detector
tubing unfortunately did not receive the same attention from the stand-
point of specification and materials control as the stainless steel used in the
primary piping. The 1/4-in. type-304 stainless steel had been purchased
to standard ASTM tubing specifications, but in 30- to 40-ft lengths instead
of the usual 20-ft lengths. The tubing was received and installed without
difficulty. It was given a hydrostatic test after installation and put in
service with distilled water. After approximately three months it was
observed that some of the water drained from the leak-detector system
was badly discolored. An analysis revealed the liquid to contain approxi-
mately 1000 ppm of chloride. Since conditions of operation up to this
point had been relatively mild, it was thought that the chloride might be
removed simply by flushing, and approximately six weeks were devoted to
disassembling the reactor and washing out all detectable indications of
that contaminant.

After discussions with the tubing manufacturer it was concluded that
the chloride had originated from a die-drawing compound which had not
been removed from the inside of the tubes prior to annealing. The presence
of the chloride-containing hydrocarbon caused carbide precipitation at the
grain boundaries during annealing and created tiny caves into which the
chloride penetrated. The pickling and cleaning treatment which followed
did not remove this material; in fact, it was learned that the manufacturer’s
pickling tanks did not accommodate the full length of the tubing, making
1t necessary to pickle by dipping approximately half the tubing at a time.
The net result was that large quantities of chloride remained inside the
tubing to be leached out later when filled with water.

At the time the stress-cracking damage was discovered late in 1956 the
reactor had been made ready for a series of engineering tests, and for this
reason it was decided to make a brief inspection of the damage resulting
from the chloride contamination before proceeding with the planned ex-
perimentation. This preliminary inspection provided the basis for a
decision to prepare for the replacement of the 259 flanges and the 15,000 ft
of leak-detector tubing in the system. It was further decided that engineer-
ing tests which had been interrupted could proceed for the period of
approximately three months which would be required to procure new
flanges and leak-detector tubing.
374 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [CcHAP. 7

Dismantling of the system was begun in April, and a very careful in-
spection of the 259 flanges was made to determine how many should be
replaced. The Super-Zyglo dye-penetrant method of flaw detection was
chosen as the most sensitive test which could be used practically in the
field.

Of the 259 flanges inspected, 167 were found to be acceptable (i.e., as
good as new), 67 were rejected because of cracks, pits, or other possible
flaws, and 25 were judged questionable. Nearly all of the rejected flanges
were in high-temperature portions of the reactor. While the inspection
method was selected as the best available, it was not judged to be infallible;
e.g., differentiation between mechanical scoring and corrosion pitting was
not always clear-cut, and any cracks covered by smeared metal resulting
from excessive gasket pressure could not be detected. Hence 1t was de-
cided to replace all the high-temperature flanges with new flanges or, where
this was not possible, to remove 0.02 in. of metal from the flange surfaces.
A total of 132 flanges were replaced; 15 were remachined. In addition,
the 1/4-in. stainless steel tubing to all the high-pressure flanges (approxi-
mately 10,000 ft) was replaced in the leak-detector system. This repair
work was completed in August 1957.

To remove any organic material introduced during repairs and to pre-
treat the fresh metal surfaces incorporated into the system, the reactor
piping was subsequently flushed with hot 8% trisodium phosphate solution,
followed by water rinses and a 5% nitric acid wash. After hydrostatic
testing the reactor was test operated with condensate for 150 hr at 280°C,
then charged with depleted uranyl sulfate solution. Test operation with
depleted uranium included: (1) a series of concentration and dilution
experiments to study the transient and equilibrium behavior of ions in the
system, (2) checking of the inventory-control methods by comparing the
fuel analyses and indicated system controls with the quantity originally
charged, and (3) observation of the corrosion behavior by analysis of fuel
samples during a 159-hr run at temperatures above 250°C. At the conclu-
sion of the run the charge was recovered and found to agree well with the
computed inventory, although chemical analyses of high-pressure-system
samples during operation had indicated a uranium concentration 5 to 10%
lower than the amount added would predict. Nickel analyses of the fuel
solution pointed to a system corrosion rate of slightly less than 1/2 mpy.

Before the reactor was charged with enriched fuel, the piping and shield
were subjected to‘careful leak tests. To obviate the presence of helium in
the piping in case further testing with helium became necessary, the reactor
was first pressurized to 500 psig with nitrogen, to which was added 40
curies of Kr8% as a tracer gas. The shield was sealed and the reactor allowed
to stand pressurized for five days, after which air samples were drawn from
the shield and heat-exchanger shells for beta-activity scanning to detect
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 375

‘the presence of any leaking krypton. This test was inconclusive at the time,
however, because of difficulties encountered in purifying the samples for
counting. Large samples of the air being tested were stored in gas cylinders;
then the piping was vented and repressurized with helium. After an addi-
tional waiting period the sealed volumes were again checked using helium
leak detectors sensitive to 1 ppm of helium in air. This test demonstrated
that the piping leakage was less than 0.2 cc/day of helium, with a pressure
differential of 15 psi across the leak. These results were subsequently
confirmed by krypton data after analytical difficulties were resolved.

With the integrity of the reactor piping established, the seal pans were
welded in place on the shield roof, and the roof plugs were locked in place.
By pressurizing the shield and flooding the roof with water, leaks in the
seal pan welds were located for subsequent repair. The major shield leak
was found to be through a 16-in. valve in the ventilating duct; when this
was repaired, the total leakage fell from 25 c¢fm to approximately 1/2 cfm.
The shield was judged sufficiently tight at this point to proceed with
the critical experiments, after which repairs were continued. By pains-
taking individual checking of all shield penetrations and the use of ther-
mosetting resin to seal the metal lips to which the roof seal pans are welded,
the leakage was finally reduced to 4 to 4.5 liters/min at 15 psig.

7-4.5 Nuclear operation [21]. Fuel charging began on December 24,
and criticality was achieved on December 27, 1957, with the core and
blanket near room temperature and at a pressure of about 800 psig.
Nuclear instrumentation for the test consisted of three fission chambers,
viz., two permanent chambers in the instrument tube outside the reactor
pressure vessel and a temporary chamber inside the blanket vessel. An
antimony-beryllium neutron source was suspended in the thimble in the
center of the core (see Fig. 7-18).

The reactor was brought gradually to the critical condition by the in-
jection of enriched uranium into the fuel solution added to the dump tanks
in batches of 100 to 400 g. Fuel feed pumps and purge pumps were operated
continuously to provide mixing between the dump tanks and the high-
pressure system. Iollowing each addition the solution concentration in the
high-pressure system was allowed to reach steady state, as indicated by a
leveling-off of fission count rates. After 2060 g of U235 had been added
and the temperature of the solution lowered to 29°C, the neutron source
was withdrawn. At this point, the fission count rates continued to rise,
indicating that criticality had been achieved. Raising the temperature of
the reactor slightly by pumping warm water through the heat exchangers
stopped the nuclear chain reaction. By further varying the temperature
and concentration of the fuel solution, it was demonstrated that the nu-
clear reaction could be easily and safely controlled in this manner.
376 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP.7

After the initial critical experiment the neutron source was moved from
the core to the blanket thimble, and the reactor was brought to criticality
seven times at successively higher temperatures ranging up to 281°C. In
each experiment the reactor temperature was raised above the desired
point by supplying steam to the heat exchangers, and a batch of fuel so-
lution was injected into the dump tanks. After steady count rates showed
complete mixing of the new fuel, the temperature was slowly lowered until
the critical temperature was reached. It was held at this point for about
1/2 hr before proceeding to the next experiment. I‘igure 7-15 compares
the experimental measurements of the HRE-2 critical concentration as a
function of temperature with concentrations calculated by various methods.
It can be seen that the two-group calculations predicted values about 209
below those observed. The harmonics calculation, which used a convolu-
tion of an age and a Yukawa kernel to represent slowing down in D20,
gave quite satisfactory results. The multigroup calculation also gave
results in agreement with the experimental data.

The first operation of the reactor at significant power levels took place
in February 1958. In April 1958 the power level was raised in steps of
1 Mw to the design power level of 5 Mw. Operation was exceptionally
smooth, and no mechanical difficulties were encountered in the first 500 hr
after charging the reactor with U235, Unfortunately, shortly after reaching
full-power operation a crack developed in the tapered portion of the
zirconium core tank, permitting fuel solution to leak into the blanket,
After a series of tests to determine the magnitude of the leakage and cal-
culations to determine the behavior of the reactor with fuel in the blanket,
it was decided to operate the reactor as a one-region machine (i.e., identical
fuel solution in core and blanket). Operation of the reactor under these
conditions was resumed in May 1958.

7-4.6 Operational techniques and special procedures. Reactor startup.
As the size of homogeneous reactors increases, the use of control-rod neu-
tron absorption to perform a startup becomes progressively less attractive;
e.g., to maintain criticality in HRE-2 while heating from 20 to 280°C re-
quires a reactivity increase of more than 259, Ak because of the large
negative temperature coefficient. It is much more convenient to provide a
means of varying the amount of fuel in the core as required to overcome
temperature and power coeflicients. |

During the initial experimental stage of HRE-2 operation, startup was
begun by evaporating heavy water from the dump-tank solutions, con-
densing, and pumping to the core and blanket circulating loops. The filled
loops were then pressurized, circulation was initiated by starting the
canned-motor pumps, and the circulating stream was preheated to operat-
ing temperature with an auxiliary heat source. The concentrated fuel was
then pumped from the dump tanks to achieve criticality.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 377

 

   
 
 
 
 
 
 
    

 

 

 

12
l | l | | | |

10 — O Gravimetric —
ON 0 Apparent (From Multigroup
0 Inventory Data) Calculation
o
X 8 | .
0
™
N: Harmonics
o Calculation
e 6 - —]
2
B
E With Thimble
9 In Core
s 4 ~
O
©
2
S 2 _

| | | | | I I
0 40 80 120 160 200 240 280 320

Temperature, °C

Fig. 7-15. Critical concentration of HRE-2 as a function of temperature.

An alternate procedure involves starting the reactor without preheating
by varying the fuel concentration. When sufficient fuel has been added to
raise the reactor temperature to its normal operating level, power with-
drawal is begun. Temperature adjustments may be made by removing
pure solvent to temporary storage tanks or adding pure solvent to the
circulating fuel solution.

A variation of this procedure is to fill the reactor slowly with fuel of the
final concentration. In this case the reactor will become chain-reacting at
a low temperature with the core tank only partially filled with fuel solution.
As the quantity of liquid in the core is slowly increased, the temperature
will rise until the desired temperature is attained with the core completely
full.

The time required for startup is determined by the rate at which heating
can be permitted. The same limitations apply to homogeneous reactors as
to other reactors in this respect. Generally, the heating rate of 100°F /hr
is considered reasonable and unlikely to produce excessive stresses. Prob-
ably more important, but more difficult to determine, 1s the temperature
difference which exists across heavy walls. Keeping these temperature dif-
ferences to less than 100°F prevents excessive stresses. Once the tempera-
ture limitations are established, the startup rate of fuel addition can be
set to match them.

Thus, although aqueous homogeneous systems have been demonstrated
to be inherently stable, restrictions are nevertheless placed upon the oper-
378 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

ator to prevent excessive power surges and the resultant excessive pres-
sures or heating rates. These restrictions are generally in the form of elec-
trical interlocks in the control circuit. For example, a typical interlock
might prevent the operator from concentrating fuel if instruments indi-
cated that an excessive concentration was being reached, or an interlock
might stop the addition of fuel solution if the temperature-measuring de-
vices indicated too high a heating rate. Although practically none of these
mistakes 1s serious enough to cause a reactor accident, they are evidence of
poor operating technique, and if permitted might result in more serious
mistakes which could cause damage in spite of the inherent stability of an
aqueous homogeneous system. |

The reactor 1s considered “‘started up” when the desired operating tem-
perature has been achieved and the reactor is ready for power extraction.

Operation. Except for several preliminary preparations, such as
warming-up the steam turbine, power extraction is the simplest part of
the reactor startup routine. This involves merely turning steam to the
turbine, bringing the turbine up to speed, and making the necessary elec-
trical switching changes to distribute the electrical output. Once the gen-
erator is synchronized and feeding into a larger power network, there is
little for the operator to do except to see that the equipment is checked
routinely for proper performance. For HRE-2 this normally takes a crew
which consists of a supervisor, an assistant engineer, and nontechnical
helpers. Checks of all continuously operating equipment are made at 1- to
2-hr intervals to verify that the equipment is performing properly. From
time to time samples of fuel solution are removed from the high-pressure
circulating system to analyze for nickel and other unwanted ions. Samples
must also be removed from the steam system to show that boiler feedwater
treatment is adequate and that oxygen production is not excessive in the
steam generators. Radiation levels must be observed to determine whether
there are leaks of fuel solution or weak places in the shield structure.

In addition to these routine service functions, the operating crew must
also start up and maintain the chemical processing plant associated with
the reactor. One engineer and one technician, in addition to those already
mentioned for the reactor proper, are required to operate the HRE-2
chemical plant.

Shutdown. The shutdown of the reactor is normally accomplished in
two steps.

The first step in the procedure is cessation of power removal, which in-
volves nothing more than closing the steam throttling valve. This might
be accomplished by running the turbine governor down to zero load.
Although this action leaves the reactor critical at 280°C and does not com-
pletely stop heat generation, the power is limited to the normal heat losses
from the system (a few hundred kilowatts for HRE-2). Minor repairs or
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 379

adjustments to equipment which do not require further cooling can now
be made.

The second step is to make the reactor subcritical by diluting the fuel.
This type of shutdown normally requires 3 to 5 hr, since it is desirable to
have the reactor subcritical at the storage temperature of approximately
25°C when the reactor is emptied. The rate at which the temperature can
be reduced is again determined by the permissible cooling rate of the
system components.

Scram. A more rapid shutdown, equivalent to an emergency scram in a
solid-fuel system, is the ‘‘dump.” In this situation the reactor is kept circu-
lating for 2 min to permit recombination of the radiolytic gas in solution
during which time the steam valves are closed to reduce power output.
Then the pressure in excess of the vapor pressure of the core and blanket
is vented to make pressure balancing between core and blanket easier,
and the dump valves are opened, permitting the rapid emptying of the
core and blanket vessels so that the reactor is shut down within minutes.
This type of shutdown is only resorted to in case of emergencies such as
excessive pressures or evidence of a leak of radioactive solution from the
reactor.

Decontamination of equipment. Conventional methods of decontaminat-
ing fuel processing plants have proved to be inadequate for a stainless steel
homogeneous reactor system which has been exposed to uranyl sulfate-
sulfuric acid-fission product solutions at 250 to 300°C. This was demon-
strated with HRE-1, which was decontaminated (without descaling) prior
to disassembly (Article 7-3.13). Since the HRE decontamination was in-
complete, laboratory studies were carried out to explore the nature of the
contamination and develop methods of decontaminating the stainless
steel. It was found that chromous sulfate, a strong reducing agent, would
modify the oxide film and permit dissolution in dilute acids. A 0.4m
CrS04—0.5 m HoSO4 solution has given excellent removal of the film by
modifying and dissolving the oxide corrosion film. Decontamination fac-
tors of 5 X 103 were achieved on specimens from in-pile corrosion loops,
where the activity was reduced to the induced activity of the structural
material, by contacting for 4 hr with the chromous sulfate solution at 85°C.

The solution was also tested satisfactorily on four 22-liter uranyl sulfate
corrosion loops which had run for 22,000 hr at 200 to 300°C. The loops
had a very heavy oxide coating such that in thermal cycles large flakes
broke off the wall and plugged small lines. After a 4-hr contact with the
chromous sulfate-sulfuric acid solution at 85°C, the walls of the loop were
completely free of all clinging oxide. |

The total time involved in the preparation of the chromous sulfate
solution, in descaling the reactor system, and in disposing of the extremely
radioactive scale-waste would probably be at least 48 to 72 hr. If one is
380 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP.7

considering decontamination as an aid to maintenance, time required for
decontamination should be weighed against the reduction in repair time
which would result from the lower levels of activity in the working area.

7-4.7 The HRE-2 Mockup [22].* During the design and construction
of HRE-2 some of its major pieces of equipment were assembled and
operated at design conditions on unenriched uranyl sulfate solution. The
purposes of this engineering mockup were (1) to study the behavior and
removal of gases in the high-pressure system, (2) to study fuel-solution
stability in a circulating system similar to the HRE-2, (3) to establish a
reference corrosion rate for the reactor circulating system, (4) to study the
behavior and removal of corrosion- and fission-product solids in the system,
and (5) to establish the reliability of components, pointing out weaknesses
and possible improvements.

The prototype reactor components tested in the mockup include: (1) a
Westinghouse canned-rotor centrifugal pump, (2) an electrically heated
steam pressurizer, (3) a centrifugal gas separator, (4) a 1/8-scale heat
exchanger similar to the HRE-2 steam generator, (5) a letdown heat ex-
changer to cool the fluid and gas bled from the high-pressure system,
(6) a bellows-sealed letdown valve to throttle the fluid and gas from the
high-pressure system, (7) a liquid-level controller which adjusts the
letdown-valve position, (8) the dump-tank and condensate system for
excess fuel solution and storage of condensate for use as purge to the
pressurizer and circulating pump, (9) diaphragm feed pumps to return
fuel to the high-pressure system from the dump tanks, (10) an oxygen-feed
system to maintain oxidizing conditions in the circulating fuel, and (11) an
alr-injection system for tests of the effectiveness of the gas-separator unit
and letdown system.

In May 1956 a corrosion- and fission-product solids-removal system was
installed in the mockup, consisting of a 5-gpm canned-rotor ORNL pump,
an assembly of a hydroclone and underflow pot, and two 1/2-in. Fulton
Sylphon bellows-sealed air-operated valves separating the system from
the main circulating stream. A commercial pulsafeeder was used to inject
rare earths containing radioactive tracers and corrosion products. In ad-
dition, a through-flow bomb was filled with the required solids and con-
nected into the system. Gamma-ray counting equipment was installed to
detect any buildup of radioactive tracers on the heat-exchanger surfaces,
in the horizontal connecting pipe to the pressurizer, in the gas separator,
and in the underflow collection pot. A multichannel gamma counter was
used for determining activity levels at the various counting stations. The
equipment was removed in November 1956 after demonstrating satis-
factory removal of solids from the system.

*Article 7-4.7 is based on a paper by I. Spiewak and H. L. Falkenberry [22].
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 381

Operation of the mockup for a total of more than 13,000 hr in the three-
year period from Iebruary 1955 to February 1958 provided valuable in-
formation pertinent to the construction and operation of HRE-2. Items
of major importance include: (1) a satisfactory demonstration of the opera-
tion of the equipment for removing gases from the high-pressure system,
(2) the out-of-pile chemical stability of the uranyl sulfate solution over a
long period of time at operating temperatures and pressures, (3) determina-
tion of the oxygen injection and excess sulfuric acid requirements to pre-
vent uranium precipitation in the high-pressure loop, (4) detection of an
unsatisfactory pressurizer design in which excessive corrosion and precipi-
tation of uranium occurred, and tests of a revised model which proved
suitable, (5) a demonstration of the successful removal of injected fission-
product solids and insoluble corrosion products by means of a hydroclone,
and (6) long-term operability of the circulating pump and other pieces of
equipment.

7—4.8 The HRE-2 instrument and control system [23].* The control
system for a homogeneous reactor such as the HRE-2 differs drastically
from that for solid-fuel reactors because control rods and fast electronic
circuitry are not necessary for systems with such large temperature co-
efficients (nearly 0.39, Ak/°C at 280°C for the HRE-2).

Functions similar to those performed by control rods in heterogeneous
reactors, but without such exacting speed-of-response requirements, are
performed by valves which control the concentration of the fuel, which
vary the steam-removal rate from the heat exchangers, or which allow the
fuel to be discharged to noncritical low-pressure storage tanks. Ior these
reasons the nuclear control circuits for a homogeneous reactor are designed
to limit very rapid changes in fuel concentration and steam-removal rates.
Other circuits control the pressures and temperature limits of the circulating
liquid fuel, principally to prevent equipment damage.

In addition to these general considerations, the instrument and control
system for the HRE-2 includes the following special features:

(a) All instrument lines through the shield wall are blocked by valves
on a signal of high shield pressure to prevent the escape of radioactivity.
Electrical leads are sealed by glass-to-metal seals.

(b) Thermocouple, electrical, and air lines may be disconnected by
remotely operated tools for equipment repair or removal. |

(c) All critical core- and blanket-system transmitters, except electric
level transmitters and thermocouples, are located in two shielded instru-
ment cubicles (5 ft diameter X 15 ft high) located adjacent to, but out-
side, the main reactor tank. This arrangement was selected to avoid
opening the main tank to replace instruments, to provide a location for

*Article 7—4.8 is based on a paper by D. S. Toomb [23].
382 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

Pressure
Pressure

 

 

 

 

 

 

 

 

 

 

  

 

Switches
Electri Switches
ectric
L Controller &Recorder Pneumatic LI
"‘ Level .
Pressurizer
Control 3 -
Circuit T — Controller
| X ]
Heater ] Time Prop.
‘ Relay
Heat
C
Exch. .

 

Dump Pulsafeeder
Tank Pump

F1g. 7-16. Key control loops utilizing both pneumatic and electric transmission.

instrument components which could not be easily protected from the
water-flooding of toe main tank during remote-maintenance operations,
and to minimize the radiation exposure of instruments.

(d) The cell air monitors, which provide an alarm in case of a leak of
radioactive vapor from the reactor system, are installed in one of these
instrument cubicles. Cell air is circulated through a 2-in. pipe from the
reactor tank, past the enclosed monitors, and then back to the cell. The
blower is sized so that.only 5 sec is required for cell air to reach the radia-
tion monitors.

Figure 7-16 indicates several key control loops: (a) The pressure of the
core system is controlled from sensed pressure by the proportioning of
power to the pressurizer electric heaters; the blanket pressure is similarly
controlled by a core-to-blanket differential-pressure signal. (b) The
liquid levels in the pressurizers are controlled from sensed levels by pneu-
matic control of the letdown valves. Pneumatic control actions are derived
from transducers which receive signals from electric transmitters. Electric
interlock control of the pneumatic signals to final control elements is
achieved by the use of solenoid-actuated pilot valves.

Other control loops not illustrated are: (¢) The reactor power is con-
trolled by a manual or turbine-governor signal to a valve throttling steam
from the core heat exchanger. Because of the large negative temperature
coefficient and low heat capacity, this reactor system cannot produce more
heat than is being removed, except for very short times. Heat is removed
by opening the steam throttling valve, which causes a decrease in the
temperature of the fuel leaving the heat exchanger. Because of the nega-
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 383
Overflow Pipe

 

Reactor Thermal Shield

T

 

 

 

-12 ft O in.

 

 

 

 

 

.

Fig. 7-17. HRE-2 nuclear instrument thimble.

 

tive temperature coefficient of the system, the cooler fuel entering the
core causes an increase in reactivity and the fuel is reheated until it over-
comes the excess reactivity. (d) The average temperature of the core
system is controlled by varying the concentration of the fuel solution.
The inlet and outlet temperatures will vary with the power extraction,
while the average temperature is a function only of fuel concentration.
(e) The blanket temperature is controlled by a signal, derived from the
difference in average core and blanket temperature, which operates the
steam withdrawal valve of the blanket heat exchanger.

Nuclear instrumentation. Neutron level transmitters are two Westing-
house fission chambers and two gamma-compensated ionization chambers
designed by the Oak Ridge National Laboratory. Neither type of instru-
ment requires a gas purge, and both are amenable to operation in the water-
filled thimbles (Fig. 7-17), which allow the chambers to be positioned or
replaced during reactor operation. Varying the distance of the chambers
from the reactor core affords a means of sensitivity adjustment, which is
needed to accommodate different operating powers.

The gamma-compensated ionization chambers are required to be in a
384 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHaAPp. 7

neutron flux of approximately 10'° at 5-Mw power to utilize their measur-
ing range of 10%. The lower limit of their range of operation is comparable
to the maximum desirable flux for proper operation of the fission chambers.
Therefore it was possible to install the two types of sensing elements in
the same area in the reactor cell. At high fluxes the fission chambers which
are required to follow the neutron flux for five decades during start-up are
withdrawn into a protective boral shield to limit fission-product buildup
in the chamber lining. Proper operation of the compensated ion chambers
is ensured by the lead-shot-and-water fill around the thimbles, which re-
duces the gamma-ray background after power operation from 250,000 r/hr
to 250 r/hr.

The fission-chamber signals are fed to conventional preamplifiers,
A-1 linear amplifiers, logarithmic count-rate meters and a dual-pen re-
corder. For initial reactor startup before gamma-neutron reactions pro-
vided a sizable neutron source, the fission-chamber output was used to drive
a low-range pulse counter.

Control panel. Figure 7-18 shows the main control board and console.
Here are located only those instruments necessary for the safe operation
of the reactor. These are arranged in a “‘visual aid”’ form to reduce opera-
tional errors and to facilitate the training of operators. The graphic section
is essentially a simplified schematic representation of the chemical process
flowsheet with instruments, control switches, and valve-position indicators
located in positions corresponding to their location or function in the actual
system.

Each annunciator is placed in the control board directly over the instru-
ment or portion of the system on the graphic board with which its signal
is associated. |

Key measurements are displayed on ‘‘full-scale’’ recorders in the center
section of the panel and include the fuel temperature, a multipoint tem-
perature recorder, the multiarea radiation monitoring recorder, the re-
actor power, the logarithmic neutron-level and count-rate meter signals.

The patch panel on the extreme left is a “‘jumper board,” which is a
schematic representation of the electrical control circuits. Provision is
made for jumping certain individual contacts in the control circuit with a
plug. Lights indicate the position, open or closed, of the contact in the
system. The board is valuable for making temporary control-circuit altera-
tions necessary for experiments and makes it' unnecessary to use jumper
wires which are normally placed behind the panel and if overlooked and
not removed might introduce a hazardous condition. The board is also an
aid in familiarizing operators with the electrical control circuitry and, since
the lights indicate contact position, as an operations aid during startup.

Switches and controls on the console are restricted to those necessary
for nuclear startup, steady-state power operation, and emergency.
THE HOMOGENEOUS REACTOR TEST (HRE-2) 385

]

7

[oued JUSWINIISUT PUB WOOI [01)UO0D

¢—HdH

81

—L

DI

 
386 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

Data-collection instruments and the transducers that drive the miniature
pneumatic slave recorders on the graphic panel are located in an auxiliary
instrument, gallery beneath the main control room, as are a 548-point
thermocouple patch panel, a relay panel, the nuclear-instrument amplifiers,
and the nuclear-instrument power supplies.

Other panels located near their respective equipment elsewhere in the
building include the steam control station, the turbine control panel, two
sampler control panels, and a control station for the refrigeration system
supplying the cold traps. Standard 2-ft-wide modular cabinets and panels
are used throughout to facilitate design changes. |

Protectiwve interlocking. Extensive protective interlocking of the controls
circuits is provided to prevent unsafe operating conditions.

Examples of interlocked systems include the following: (a) The
pumping-up of fuel to the reactor core instead of condensate is prevented
by several interlocks which keep the fuel-addition valve closed until the
core is full of condensate and has been heated to 200°C. This prevents
power surges and consequent pressure increases. (b) For the same reason
the fuel circulating pump is started in reverse to provide a low flow rate as
protection against pumping cold fuel too rapidly from the heat exchanger
into the core. (¢) To avoid dangerous thermal stresses and abrupt reactivity
surges, the control circuits do not permit the pumping of cold feedwater
into the heat exchangers until the level is above 50%. (d) To give smooth
startup, the fuel injection pump can run only at half-speed until a tem-
perature of 250°C is reached. (e) The fuel feed valve will be closed and the
concentration in the fuel system will be lowered by injecting condensate
if the core outlet temperature becomes excessive, if the circulating pump
stops, or if the power level exceeds normal. (f) The contents of the high-
pressure systems will be automatically emptied to the low-pressure storage
tanks through the “dump’ valves on a signal of extremely high pressure,
or a radiation leak into the steam system. Differential-pressure control
between the core and blanket systems during dumps is by throttling control
of valves from a differential-pressure signal.

Inventory systems. To obtain an accurate inventory of the fuel and
moderator solutions, the storage and condensate tanks are weighed with
pneumatic weigh cells. A pneumatic system was selected primarily because
taring can be done remotely by balancing air pressures, and components
are less susceptible to radiation damage. Piping to the tanks is kept
flexible to compensate for the varying loads which result from thermal
expansion.

The volumes of the fuel and blanket high-pressure systems have been
accurately measured, so that when the pressurizer condensate reservoirs
are filled to capacity, the weight of liquid in the high-pressure systems can
be computed from the core, blanket, and pressurizer temperatures. These
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 387

weights, combined with the respective condensate weights, dump-tank
weights, and experimentally observed holdup in the condensate piping,
yield the total liquid inventory of the reactor.

7—4.9 Remote maintenance [24]. Maintenance of the equipment in the
circulating systems of an aqueous homogeneous reactor is difficult because
of the intense radioactivity emanating from the surfaces following pro-
longed operation at high power levels. This problem, more acute than in
the case of heterogeneous reactors because fission-product activities and
neutrons are not confined to the reactor core, must be successfully resolved
in a practical manner if the homogeneous reactor is to be an economical
power producer. Practical systems for maintenance and repair must pro-
vide adequate personnel shielding to reduce radiations from activated
equipment to tolerable biological values while simultaneously providing
access to the equipment.

Equipment activation. Of the two principal methods by means of which
the equipment is rendered radioactive, activation by fission-product con-
tamination is the more important in the HRE-2. About 30 w/o of the
fission products is normally removed from the fuel solution as gases (xenon,
krypton, iodine) by stripping with Oz, D2, and steam. Of the remaining
fission products, 56 w/o are estimated to have sufficiently low solubilities
to precipitate from the solution as solids. This is particularly true of the
rare earths, which constitute 28 w/o of the fission products. Tellurium,
technetium, and molybdenum, which contribute 12 w/o of the fission
products, are assumed to be soluble. Therefore, so far as activation of
equipment is concerned, the rare gases (~30%) and the soluble products
(~12%), making up 42 w/o of the total, may be considered removed or
easily removable from the equipment. Some portion of the remaining
58 w/o will deposit on hot metallic surfaces or be retained in cracks and
crevices. It is possible by statistical analysis to estimate within a factor of
two the specific gamma and beta activity of the fuel solution. Methods for
calculating this activity have been published [25,26]. The initial fuel ac-
tivity, after prolonged HRE-2 operation, is of the order of 25 curies/ml.

It should be noted that it is proposed to remove the insoluble fission
products continuously with small centrifugal-type separators (hydroclones)
in the HRE-2 fuel processing system. The effectiveness of this procedure
in competition with absorption on the container walls, however, is as yet
unknown.

Induced activation of structural materials by neutron absorptlon is also
an important contributor of radioactivity. During operation of the reactor,
a thermal flux exterior to the equipment items will be present because of
neutron leakage from the reactor vessel. Also, delayed neutrons will be
emitted in the interior of the piping and equipment from fission products
388 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

in the fuel solution. These neutron fluxes may be estimated, and knowing
the surface areas and material constituents of the equipment, the resulting
induced activation may be calculated. Methods of calculating induced
activation are reported in the literature [27,28].

The effect of induced activation may be controlled to some extent by
attenuating neutron leakage from the reactor through an adequate thermal-
neutron shield surrounding the reactor vessel. Also, the constituents of
the structural materials in the cell may be specified in such a manner that
elements of potentially high neutron activation, such as the eobalt normally
present in stainless steels, are minimized. Other materials, such as boron,
which has a high neutron-capture cross section and is an alpha emitter
rather than a gamma radiator, might be used on equipment surfaces to
minimize induced activity.

Radiation dosages. Design criteria followed in the design of shielding and
tooling for remote-maintenance operations are based on tolerance radia-
tion levels that are acceptable for normal operations. In continuous work-
ing areas, radiation levels as low as 1 mr/hr are specified. For certain in-
frequent operations a level of 74 mr/hr is permitted. However, during
periods of intense maintenance it may sometimes be permissible to allow
personnel to work in an area of higher than normal radiation, but the total
weekly radiation dosage must be held to 300 mr or less. In some instances
it 1s necessary to work in relays so that no one person receives a higher than
permissible weekly dosage.

Proposed maintenance concepts [24]. The concepts which have been pro-
posed for the maintenance of equipment in the HRE-2 have been based
on the philosophy of either dry maintenance or underwater maintenance or
combinations of the two. In general, it is assumed that the item requiring
repair will be replaced with a new item and the faulty component removed
to a shielded hot-cell area for maintenance.

Tests of underwater-maintenance procedures proposed for HRE-2 using
special tools and equipment have demonstrated the feasibility of removing
and replacing equipment. In these tests it was demonstrated that con-
tamination of the heavy-water fuel solution by shielding water could be
prevented by freezing a plug of heavy ice in the pipe before disconnecting
equipment. Freeze jackets are placed around piping at flange disconnects
for this purpose. Also in the HRE-2, the insulation around piping was de-
signed to avoid the retention of water after draining the flooded equip-
ment.

Piping joints. Removal of equipment in the HRE-2 depends on flanged
joints adjacent to the equipment which may be disconnected and remade
with suitably designed tooling. The joints in the HRE-2 use American
Standard stainless-steel ring-joint, weld-neck flanges modified to permit
leak detection (Fig. 7-13). These flanges are located so that all bolting is
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 389

 

Fie. 7-19. Universal jig for locating flange positions on HRE-2 replacement
high-pressure valves.

accessible to tools operated from above. The number of loose pieces such
as bolts, nuts, ferrules, and gaskets is minimized by welding nuts to the
flanges such that they become integral with the flange.

Remotely operable tools.* Tools used in the HRE-2 maintenance pro-
cedures are generally simple, rugged units mounted on long handles.
They include manual, air-operated, and water-operated socket wrenches
for turning nuts, hooks for lifting, clamps, knives, magnets, etc. Some
typical maintenance devices are illustrated in Figs. 7-19 through 7-22
and described as follows.

The universal high-pressure valve jig shown in Fig. 7-19 is used for
manufacturing replacement high-pressure flanges for the HRE-2. By
placing brackets in appropriate holes, referenced with respect to the 21
high-pressure HRE-2 valves, the position of the flanged pipe ends for any
one of the valves may be accurately determined. With this device, the re-
quired number of spare valves for the reactor is minimized, and at the
same time any valve which fails can be remotely replaced.

*Other remotely opefable tools are described in Sec. 19.5.6 of the Reactor
Handbook, Vol. II, Section D, Chapter 19, ORNL-CF-57-12-49.
390 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

 

F1g. 7-20. Spinner wrench for rapid removal of HRE-2 flange bolts.

The spinner wrench in Fig. 7-20 is used for removing bolts quickly after
they are loosened with a high-torque wrench, for spinning the bolts snug
before final tightening and, by use of a right-angle bevel-gear attachment
(not shown), for handling bolts in a horizontal position. It has a chain-
drive offset to fit bolts not accessible by direct drive, such as bolts hidden
under pipe fittings and bends.

The flange-spreader tool (Fig. 7-21) is required in remote operations to
spread low-pressure valve flanges to prevent damage to the ring gasket
during removal or insertion of the valves. The tool is shown in place with
the flanges spread and the valve assembly moved back on its cell fixture.

The hydraulic torque wrench in Fig. 7-22 is used to tighten or loosen
bolts on flanges larger than 2-in. pipe size. This wrench is actuated by a
hydraulic (water-operated) cylinder and is capable of applying a 2000 ft-lb
force moment.

Even in the relatively short operating history of HRE-2 the concept of
underwater repairs with tools such as those described above has been
proved to be completely practical. Soon after the reactor attained signifi-
cant power it was found necessary to replace all the electrical power wiring
—nearly 4000 ft of wire. This was accomplished in a period of three weeks,
with the cell completely flooded. During this period the fuel circulating
pump was removed to recover a set of corrosion samples. The removal
and replacement of the pump required only 100 hr working time, which
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 391

 

Fie. 7-21. Flange-spreader tool for preventing damage to gaskets during re-

moval of low-pressure flanges. Shown in place during maintenance operation on
HRE-2.

would have been the total shutdown time if the pump replacement alone
had been performed. Also, the core inspection flange was removed and the
core inner surface inspected by means of a periscope. In these three major
underwater-maintenance operations no delay or difficulty was experienced.

7—4.10 Containment methods [29]. In designing for the containment of
HRE-2, extremely rigid leakage specifications were set, both for the pri-
mary piping system and the shielded tank containing the reactor. In the
case of the piping, the leakage specification was based on the minimum
leak which could be found with the available mass-spectrometer leak-
392 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

 

Fic. 7-22. Hydraulic torque wrench for loosening or tightening flange bolts.
Shown in operation on HRE-2.

detection equipment, namely, 0.1 cc helium (STP) per day. Allowing for
several such minimum leaks, the equivalent loss of liquid at 300°C and at
2000 psig could be as great as 5 cc per day, or a total leakage of approxi-
mately 250 curies/day. Because of its volume and surface area, and be-
cause of the difficulty in measuring small leakage from very large vessels,
the leakage rate for the reactor tank was set at 10 liters/min or less, at a
test pressure of 15 psi. As indicated in Article 7-4.4, actual leakages were
below this value.

The problem of leakage from the reactor vessels and piping can be cata-
logued according to the various mechanisms through which leakage might
result as follows: (1) excessive stresses, (2) defective materials or work-
manship, (3) corrosion, (4) nuclear accidents, (5) hydrogen-oxygen ex-
plosions, and (6) brittle fracture. Each possibility is examined in detail in
the discussion which follows.

Excessive stresses. There are many possibilities for the development of
excessive stresses in a system as complex as the HRE-2. In order to reduce
the likelihood of failure as a result of excessive stress, a maximum allowable
working stress of 12,000 psi was specified for the type—347 stainless steel
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 393

with which the system was fabricated. This permits an additional factor
of safety over the 15,000 psi allowed by the ASME boiler code. As required
by the ASA code for pressure piping, the reactor piping arrangement was
examined for maximum stresses due to pressure, as well as for hoop and
bending stresses resulting from thermal expansion. Equipment was also
studied for determining the magnitude of thermal stresses caused by radia-
tion heating and temperature cycling. Therefore, in order to keep the com-
bined thermal and pressure stresses below the maximum allowable working
stress, the pressure-vessel wall is approximately 2 in. thicker than would
be required otherwise. Heating and cooling rates on the entire system have
been limited to 100°F/hr and 55°F/hr, respectively, and the differential
temperature across heavy metal walls is kept below 100°F. (The cooling
rate is held below the heating rate to minimize rate of flange contraction.)

Cyclic temperature stresses at questionable points in the reactor pres-
sure vessel and steam generators were explored experimentally. Mockups
were fabricated for the testing of the pressure-vessel nozzle joints and the
stainless-steel-to-Zircaloy bolted joint inside the pressure vessel. In each
test the temperature was cycled from approximately 250°F to approxi-
mately 600°F in 1/2 hr and cooled back to 250°F in 1/2 hr. After 100
cycles the joints were found to be sound. The main steam generators were
also cycled in similar tests. Several tube joints cracked open during the
first 50 cycles. They were repaired and the heat exchangers were subjected
to an additional 10 cycles before final acceptance.

Defective materials or workmanship. Defective materials and poor work-
manship constitute another area which required special attention to pre-
vent failures. All materials for the primary systems of HRE-2 were
procured to specifications considerably more rigid than those existing in
commercial practice. Optional requirements such as chemical analyses,
boiling nitric acid tests, and macroetch tests were exercised in all materials
specifications. An additional cost, averaging about 10%, was experienced
in the purchase of materials under the more rigid specifications. In some in-
stances, for example with the heat-exchanger tubes, special ultrasonic and
magnetic eddy-current flaw detectors were employed to indicate defective
parts. Three tubes which might have later failed in operation were thus
eliminated. Dye-penetrant tests were applied to tubing bends and to all
welds throughout the reactor to detect cracks and pinholes. None was dis-
covered in tubing bends, but many were found in welds, especially in the
tube-to-tube-sheet welds.

Special attention was given to the welding of stainless steel butt joints,
of which there are approximately 2000 in the entire reactor. The inert-gas,
nonconsumable-electrode method was used almost entirely. Welds were
inspected to considerably higher standards than required by the ASME
code. In addition to being subjected to dye-penetrant inspection, every
394 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

weld was x-rayed. Although the inspection standards were very rigid, only
3% of the welds were rejected, necessitating rewelding.

Corroston. Although corrosion is an ever-present possibility for leakage,
the HRE-2 was designed to reduce attack rates so far as possible by keep-
ing velocities below 20 fps, temperatures in the range of 250 to 300°C, and
lining some surface areas with titanium for additional resistance.

Nuclear accidents. Nuclear accidents are likely to be rare in homogeneous
reactors because of the large negative temperature coefficient (0.1 to 0.2%
Ak/°C) [30]. As an example, the worst accident considered in the HRE-2
[31] was one in which all the uranium suddenly collects in the reactor core
and results in a reactivity increase of 2.59% Ak/sec. For this rate, starting
at a power of only 0.4 watt, the maximum pressure in the pressure vessel
would be approximately 3900 lb, and the pressure stress in the carbon steel
shell would be less than 30,000 psi.

Hydrogen-oxygen explosions. Since radiolytic gases (deuterium and
oxygen) are produced continuously in the reactor, they are the source of
an ever-present hazard. Explosions may be expected whenever the
deuterium-oxygen-steam mixture is more than 15% gas. For detonations
the required gas fraction is greater. The maximum increase in pressure
from an adiabatic explosion of hydrogen and oxygen is only a factor of
3 to 8, whereas for a detonation the factor might be 23 for an undiluted
mixture. (It is important to note that detonations can occur only in gas
channels that are relatively long and straight.)

In low-pressure areas of HRE-2 the gas was diluted with steam to keep
it noncombustible. Furthermore, a pressure of 500 psi is the basis for
design of low-pressure (atmospheric) equipment. Even for a detonation
the expected peak pressure would be less than the design pressure.

Although an explosion can be tolerated in the high-pressure system with
little danger of vessel rupture, a detonation probably could not be. A
detonation could occur in the gas separator where the gas channel (the
vortex by which the radiolytic gas is collected) is long and straight. It is
calculated [32] that a detonation wave traveling longitudinally along the
vortex would produce impact pressures of the order of 30,000 psi but that
the damage would be limited to the directional vanes inside the separator.
Attenuation of the forces by the solution would limit the pressure rise to
that resulting from combustion of the gas—10,000 psia, which produces a
tolerable wall fiber stress of 35,000 psi. Thus no serious damage is foreseen
from explosions or detonations.

Brittle fracture. 1t is generally known that ferritic steels are subject to the
phenomenon of brittle fracture. Although the likelihood of brittle fracture
in the HRE-2 pressure vessel was known to be small, an investigation was
made [33] in order to determine the consequences of such an accident as
a result of the pressure rise and missile damage.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 395

A sudden release of the liquid contents of the fuel and blanket systems—
3910 1b of solution at 300°C and 2000 psig—would result in a pressure rise
of approximately 30 psi inside the reactor shield. Although it is somewhat
difficult to design a large (25,000 cu ft) rectangular container such as the
HRE-2 container to withstand a 30-psig pressure, it was even more diffi-
cult to design it to withstand heavy missiles. Studies indicated that missile
velocities of 50 to 150 fps could be expected. With the HRE-2 vessel,
which weighs 16,000 lb, 1600-1b fragments might be expected. Such a mass
at 100 fps would lift an unrestrained 5-ft-thick concrete shielding plug
nearly 1 ft. For this reason a blast shield was placed around the vessel.

The blast shield was designed to withstand the 1250-psig pressure which
would result from the 300°C liquid, as well as to absorb the energy of the
expanding steam. A 13-in.-thick wall made of type—-304 stainless steel was
found to be capable of absorbing 2.85 X 108 ft-lb of energy with 297, elonga-
tion. The fragments from the pressure vessel would accumulate this much
energy in traveling across an annulus of 4.8 in. To provide an additional
factor of safety, the blast shield was constructed to surround the pressure
vessel with an annulus not greater than 2 in.

A similar shield made of carbon steel was installed on both fuel and
blanket steam generators as a final precaution against an explosion which
might damage other equipment and the vapor barrier sufficiently to cause
a release of activity. The probability of failure of the steam generator
shells would not be affected by the presence of neutron radiation because
the generators are located in a low flux region.

The vapor-tight container. If leakage of radioactive solution from the
reactor proper occurs through one of the means mentioned above, an all-
welded, vapor-tight tank provides a second barrier to the escape of radia-
tion to the atmosphere. Construction details of this tank, which forms the
liner for the biological shield, are illustrated in Fig. 7-8. After completion
of construction the tank was given a hydrostatic test at an average pressure
of 32 psig, and the welded joints were found to be free of leaks. It was then
reopened for further installation of reactor equipment. Prior to operation
of the reactor at power, the containment vessel was sealed and tested
(Article 7-4.4).

7-4.11 Summary of HRE-2 design and construction experience. Fol-
lowing the completion of HRE-2, personnel associated with the project
recommended design improvements which might be applied to good ad-
vantage in future homogeneous reactors. These recommendations are
summarized in the following paragraphs:

(1) There are many components in the HRE-2 system which must
operate in conjunction with one another for continuous operation of the
reactor. Since a failure of any one of these could cause a shutdown of the
396 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

reactor, the probability of shutdown is higher than it would be with com-
ponents operating independently. In future reactors an attempt should be
made to decrease the dependency upon the simultaneous operation of
essential components. An example of this is a system which can be operated
without continuous letdown from the high-pressure to the low-pressure
system, so that failure of the low-pressure system components will not
necessarily require shutdown of the entire reactor system.

(2) The reactor cell for the HRE-2 contains many structural compo-
nents made of carbon steel. Flooding the cell with water and removing
radioactive contamination by use of acids will eventually do serious damage
to all these carbon steel surfaces. Wherever possible in future reactors,
corrosion-resistant materials should be used for structural components, or
those components should be protected by coatings which are resistant to
radiation, as well as to acids and water.

(3) In laying out equipment and piping, provision should be made for
walkways, stairs, and ladders to provide convenient access during con-
struction without having to climb on the process and service piping. These
walkways, ladders, and stairs may or may not be removable after com-
pletion of construction.

(4) Many items installed inside the reactor cell of the HRE-2 might
have been located outside the radioactive area. These include the cold
traps, which could be located in one of the shielded waste-system compart-
ments, the space coolers, and a number of other nonradioactive components.

(5) In the HRE-2 the sampler is located adjacent to the cell shield, and
it is necessary to remove the sample from the sampler and carry it to the
analytical laboratory in a shielded carrier. The risk of contamination can
be greatly reduced if the hot laboratory is built around the sampler so that
samples may be analyzed without removal from the shield; however, this
would add to the cost of the facility.

(6) Access through the containment seal, which consists of welded pans,
can be made easier through the use of mechanical seals with organic gaskets.
These would not be subject to severe radiation damage since the contain-
ment seal is itself shielded from primary radiation.

(7) Control of all operations affecting the reactor should be located
in one central control room. This includes the controls for the chemical
processing system, as well as those for the reactor system and all main con-
trols for the steam turbine and steam auxiliaries. This does not mean, how-
ever, that all instrumentation must be in the control room, but only those
things which are necessary to keep the reactor operator informed and to
provide him with primary control of the operation.

(8) In the HRE-2 refrigeration system a petroleum distillate, “AMSCO
125-82,” was circulated inside the reactor cell as a secondary coolant be-
cause no primary low-pressure refrigerant could be found which would not
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 397

decompose under irradiation to release chlorides or fluorides detrimental to
the stainless steel piping. Carbon dioxide was not selected because the
piping would have to be designed to withstand about 1500 psi. A solution
to this problem would be to use copper for all tubing and equipment in con-
tact with the refrigerant, in which case I'reon could be used inside the cell.

7-4.12 HRE-2 construction costs. These costs are summarized in Ta-
ble 7-7, including that portion of the costs due to the requirement for pre-
venting the possible escape of radioactive material from the reactor to the
building or surroundings.

7-5. THE Los AraMos PoweER REACTOR EXPERIMENTS
(Larre 1 anD 2) [34, 35]*

7-5.1 Introduction. Homogeneous reactors fueled with uranium oxide
dissolved in concentrated aqueous solutions of phosphoric acid have a
number of advantages compared with dilute aqueous uranyl solutions for
certain power reactor applications. These applications are based on the
fact that concentrated phosphoric acid solutions have high thermal sta-
bility and low vapor pressures. This makes it possible to operate at rela-
tively high temperatures without creating the excessive pressures en-
countered with dilute aqueous solutions. These temperatures are high
enough to take advantage of the back reaction for recombination of gases
produced by radiolytic decomposition and to eliminate the need for an
external gas system or an internal catalyst. A phosphoric acid solution
reactor, therefore, can be a sealed single vessel with no external compo-
nents except a circulator for the steam system. In addition, the high
hydrogen density and the relatively low neutron-absorption cross sections
of the phosphoric acid system' permit the construction of small, compact
reactors with a low inventory of fuel, which may be ideal for remote
package power plants. Other advantages include the possibility of continu-
ous removal of fission-product poisons and a strong negative temperature
coeflicient of reactivity.

Although the advantages of uranium oxide in concentrated phosphoric
acid solutions present a strong case for such a fuel system, the disadvan-
tages, though few in number, are ponderous. In fact, one disadvantage,
the highly corrosive nature of the phosphoric acid to most metals, may well
outweigh all of the advantages. The only metals known to be suitably re-
sistant to phosphoric acid in the concentration and temperature ranges of
interest are gold and platinum. Although with proper design the cost of
using such materials in a reactor can be kept within reason, the problem of
providing and maintaining an impervious noble-metal barrier between the

*Based on material prepared by members of the Los Alamos Scientific Labora-
tories.
398 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

| TABLE 7-7
HRE-2 ConstrUcTION CoST* (TO MAY 1956)

Reactor (HRE-2):

Design
Reactor : $549,000
Structure and services 185,000
Instrumentation 126,000
Total design cost (labor OH @ 65%) $860,000
Reactor installation
Building modification 58,000
Reactor cell and shielding 375,000
Reactor components and piping 1,400,000
Reactor controls (instrumentation and valves) 470,000
Reactor steam system and cooling system ' 161,000
Miscellaneous service piping 69,000
Maintenance tools 54,000
Waste and off-gas system 78,000
Miscellaneous and spare parts 208,000
Division supervisory labor 120,000
Total installation (labor and material) $2,993,000
Total reactor cost $3,853,000

Containment cost estimate for HRE-2:

1. Additions to shield design, required only to meet containment

specification $75,000
2. Cost of sealing conduits and wiring 7,000
3. Cost of leak-tight closures on process service lines 23,000

4. Blast shields on reactor components

Pressure-vessel $35,000
Two steam generators 10,000 45,000
5. Radiation monitoring of service lines 10,000
Total cost $160,000

*Excluding $800,000 cost of supporting research and development labor.
1Building housed HRE-1—initial cost, $300,000.
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 399

TABLE 7-8
PropPERTIES oF LAPRE-1 anpD LAPRE-2 FuEL SYSTEMS

 

 

 

LAPRE-1 LAPRE-2

Type of uranium oxide UO3 UO;
Weight percent HsPO,4 53 95
Vapor pressure, psig 250°C 400 190
Vapor pressure, psig 350°C 1800 300
Vapor pressure, psig 450°C 4500 800
Overpressure gas O2 H-
U235 concentration, g/liter 111 77
H concentration, moles/liter (25°C) 90 62
H/U235 190 190
Fuel volume ratio (430°C to 25°C) 1.33 1.16
Temperature coefficient of

reactivity (per °C) —8 X 104 —5 X 10~¢

 

 

 

 

 

fuel solution and all structural metals with which it might come in contact is
extremely difficult to solve. This problem was not solved in the first ex-
perimental reactor to use phosphoric acid system, the LAPRE-1; however,
considerable improvements in fabrication and testing techniques have been
developed for use in the construction of LAPRE-2.

Of the wide range of possible combinations of uranium oxide, phosphoric
acid, and water, two systems appear to be of special interest. The first,
which consists of UO3 dissolved in an aqueous solution containing between
30 and 60 w/o of phosphoric acid and pressurized with an oxygen over-
pressure, was used for the first Los Alamos Power Reactor Experiment
(LAPRE-1). The properties of this solution are given in Chapter 3 and
summarized in Table 7-8. The vapor pressure of this solution at the design
operating temperature of LAPRE-1 (430°C) is 3600 psi.

The second solution (to be used in LAPRE-2) consists of UO» dissolved
in a 95 w/o phosphoric acid and pressurized with a hydrogen over-
pressure. As seen from Table 7-1, the vapor pressure of this solution at
450°C is only about 800 psi. Because of its reducing properties, the cor-
rosive behavior of this solution is such that anything below hydrogen in the
electromotive series is only slowly attacked. Since the attack rate is pro-
portional to the position of the metals in the series, possible materials
in decreasing order of usefulness are platinum, gold, carbon, silver, and
copper. Although neither silver nor copper is attacked at a rapid rate, both
metals undergo mass transfer, and suitable means for inhibiting this must
be found before they can be considered practical for lining the reactor.
400 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

Solution Steam Discharge Line

Storage /

Tanks 3 Emergency

\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

> N s — |
\
Feed Water \ \ | Solution
Pump Nawswy = =4~ Transfer
] L - i
Exhaust \\ H"T =: a1/~ Pump & Line
& Vent 2 2 —
\\fi g L Valve Control
c ] Panel
§fi=a;=§z:s8 = =0 N fl
2 I 2 o
Auxiliary - z D) N
. -
Equipment ~ ! E{\ Control
Room 1\ I Console
& \ —— =3— And Racks
\ Irradiation Port
Reactor
Vessel  — 1

Lead and Water Shield Walls

Fig. 7-23. Plan view of LAPRE-1 reactor area (courtesy of Los Alamos Scien-
tific Laboratories).

7-5.2 Description of LAPRE-1. The first experimental reactor,
LAPRE-1, was housed in a cell at the Los Alamos site which had been
built previously to handle highly radioactive materials. The cell was modi-
fied by supplementing the shielding with additional concrete on one side
and lead on the other sides and ceiling. A stainless-steel wall was also put
across the reactor end of the cell to permit filling with water for a neutron
shield. Equipment such as the solution transfer pump and sampling system
were located in the cell outside the water-filled portion, while other auxili-
aries were located outside the cell.' These included a 12 gpm, 4000-psi feed-
water pump, three 6-in.-diameter, 32-ft-long underground solution meter-
ing and storage tanks, and an emergency dump tank. Valve controls were
located on the wall of the cell. Figure 7-23 shows a plan view of the
LAPRE-1 reactor area. '

Figure 7-24 shows an artist’s sketch of LAPRE-1 in which the essential
features of the reactor are indicated. As will be seen from this sketch, the
15-in.-ID cylindrical reactor contains a critical zone above the cooling
coils and a reservoir region below. This latter region, which contains a
hollow eylinder of boron to prevent criticality, provided a storage place for
excess fuel and avoided the necessity for transferring the highly radioactive
fuel solution in and out of the reactor as its temperature changed. The
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 401

4.~ Steam Outlet
o 7§y Control

L= Lid

    
  

Vessel
Flow
Baffle

Critical
Region
Heaters &
Insulation

    
   
  

Pump

 

Upper Pump
Cooler

Fic. 7-24. Los Alamos Power Reactor Experiment No. 1 (couftesy of the Los
Alamos Scientific Laboratories).

large negative temperature coefficient of reactivity made it impossible to
compensate for the excess reactivity of the cold critical reactor and it was
not considered practical to change the fuel concentration as is done in the
HRE-2. In the LAPRE-1, shim control was achieved by means of the
thermal expansion of the fuel. This was accomplished by initially filling
the reactor with the cold phosphoric acid fuel solution to a level about
7 to 8 in. above the cooling coils. At this level the reactor would become
critical on removal of the control rods, the solution would heat up, expand,
and gradually fill the critical zone. On cooling, the reverse process took
place, making the reactor self-compensating. The reactor vessel was so
designed that at 430°C, the operating temperature, the fluid level would
reach a point just above the flow baffle. In the actual experiment the
reactor was maintained at zero power while the temperature was raised
402 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [crAP. 7

Water Steam

P. Heat Exchanger
| Graphite
@ Reflector

Platinum Draft
B Tube {Critical
Region)

B BeO Source
= Startup Source

B Reflector Shim

. = Fuel Transfer
Fuel W Line
Storage | ; .
- Tank

 

Fic. 7-25. Los Alamos Power Reactor Experiment No. 2 (courtesy of the Los
Alamos Scientific Laboratories).

to 240°C by using electrical heaters located around the exterior of the
vessel. This was done to minimize the positive reactivity change resulting
from starting up the fuel circulating pump, which caused a rapid injection
of lower temperature fuel solution from the reservoir into the core.

Heat removal from LAPRE-1 was accomplished by circulating the fuel
solution at 600 gpm over the cooling coils by means of the centrifugal
sealed rotor pump shown in Fig. 7-24. The 12-gpm feed-water stream,
heated to 800°F in passing through the cooling coils, was simply discharged
to the atmosphere through a quick-closing throttling valve and steam
silencer as a means of dumping of heat. The possibility of fission products
escaping through the steam discharge line was prevented by installing a
70-sec holdup line, suitably monitored for radioactivity, between the
reactor and throttling valve.

In case of accidental stoppage of the pump during full-power operation,
the reactor could be shut down with the control rods. Provision for dis-
charging the fuel solution by hand or through a rupture disk was included
for added safety.

The entire LAPRE-1 vessel and cell components exposed to fuel solution
were gold-plated, with the exception of the rod thimbles and draft funnel
which were made of, or coated with, platinum to minimize neutron ab-
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 403

TABLE 7-9

CHARACTERISTICS OF Los AvLaMosS
Power REACTOR EXPERIMENTS

 

 

 

LAPRE-1 LAPRE-2
Power level, Mw heat 2.0 0.8
U233 in core, kg 4.1 4.1
U235 total, kg 8.5 6.5
Fuel temperature, °C 430 430
Fuel pressure, psig 4000 700
Steam temperature, °F 800 600
Steam pressure, psig 1800 600
Cost of components $250,000 $120,000

 

 

 

 

 

sorption. The 22 stainless steel cooling coils, 50 ft long, were clad with
6 mils of gold.
The characteristics of LAPRE-1 and -2 are summarized in Table 7-9.

7=5.3 Operation of LAPRE-1. Final tests on the LAPRE-1 system
were made with a 0.51 M UOg3 in 7.25 M H3PO4 fuel solution. Data were
obtained at room temperature in terms of control-rod position at delayed
critical versus volume of fuel injected into the system,* and results were
interpreted in terms of a simplified calculational model to obtain control
rod worths. For the five control rods, four located on a 37%-in. radius
and one central rod, measurements yielded a total worth of 6.39;,. The
latter results were in good agreement with period measurements at cold
critical. Also inferred from the data was an effective delayed neutron
fraction of 0.0091.

With the reactor filled to a predetermined level at room temperature,
initial heating of the system was achieved by means of electrical heaters
disposed around the outside of the vessel. Heat was applied until a core
temperature of 240°C was attained; heating by nuclear power was then
initiated. With nuclear power a core temperature of 340°C was achieved,
at which point the circulator was turned on. Only during the forced con-
vection phase of the operation was a uniform temperature distribution
established in the fuel solution; previotis to starting the circulator, the

*Detailed description of the nuclear data obtained during the test is set forth
in “Control Rod Worths vs. Temperature in LAPRE-1"” by B. M. Carmichael
and M. E. Battat, TID-7532 (Pt. 1), p. 125, U.S.A.E.C., Technical Information
Service Extension, Oak Ridge, Tenn.
404 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

average core temperature was higher than the corresponding reservoir
temperature.

Data were obtained during the course of the experiment in terms of
control rod positions versus core temperature at delayed critical. For the
operation without the circulator, thermocouples located at various points
in the system provided data from which average core and reservoir tem-
peratures could be estimated. With the circulator in use, the data for
control rod position versus temperature were used to infer the temperature
of the solution level at the top of the baffle. Above this point a further
increase in height did not contribute to the reactivity; hence the system
was solely governed by the temperature coefficient of reactivity. Based
on the calculated negative temperature -coefficient for the system
(~8 X 10~%/°C) it was found that the rods were worth 30 to 409, more
at 390°C than at cold critical.

A maximum temperature of 390°C at an operating power of about
150 kw was achieved in the experiment. No steam data were obtained
during the test because of a failure in the heat-exchanger assembly after a
number of hours of operation. Inspection of the assembly after a cooling-off
period indicated that a rupture in the gold cladding of two of the lead-in
tubes was responsible for the failure. This rupture was probably due to
a bonding between the gold cladding of the tubes and the gold plate of the
baffle, in conjunction with vibration of the tubes resulting from the opera-
tion of the circulator. Although no further tests with LAPRE-1 were
conducted after October 1956, careful inspection of the reactor parts
after disassembly provided information valuable to the construction of
LAPRE-2.

7-5.4 Description of LAPRE-2. The second power reactor experiment,
LAPRE-2, was constructed in an underground steel tank 50 in. in di-
ameter and 20 ft long located at a site approximately 100 feet from
LAPRIE-1. A sketch of the reactor is shown in Fig. 7-25. Because of the
lower vapor pressure of the LAPRE-2 fuel solution the design and method
of controlling LAPRE-2 differ from LAPRE-1. As shown in the figure,
the reactor vessel has a simple autoclave shape with relatively thin
(5/8-inch) walls. This compares with the 3-in.-thick vessel required for
LAPRE-1. The cooling coils in LAPRE-2 are located above the reactor
core, which has the advantage of providing a noncritical region for excess
liquid in case of overfilling. Cooling is accomplished by natural convec-
tion circulation of fuel solution over the cooling coils. Fuel circulation is
aided by an inverted platinum cone in the core region. A 1-ft-thick graphite
reflector surrounds the vessel. The inner 6 in. of this reflector can be moved
slowly to adjust the reactivity at varying fuel concentrations.

Cold criticality is achieved by slowly filling the reactor with fuel solution
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 405

from a water-cooled copper tank pressurized with hydrogen. The line
connecting the tank and reactor is gold clad (0.010 in.). After the reactor
becomes critical with cold fuel solution, further additions are made until
the desired operating temperature of 430°C is reached. A separate pancake-
shaped cooling coil, located just above the main tube bundle, serves as a
level indicator. A thermocouple in the discharge of this coil indicates
whether there is power removal and thus if the coil is immersed.

All parts of LAPRE-2 are completely clad with fifteen mils of gold to
provide 1009, protection of the structural metal from the fuel solution.
This gold, as well as the platinum and noble metals used in LAPRE-1,
are recoverable after conclusion of the reactor experiments.

The 600-psi, 600°F steam produced in the LAPRE-2 is of somewhat
lower quality in terms of power production than LAPRE-1 because of
the lower operating pressures. However, at a design level of 1.3 Mw heat,
the reactor could produce 250 kw of electricity which is acceptable for a
remote power station.

The characteristics of LAPRE-2 are summarized in Table 7-9. Con-
struction of the reactor proceeded through 1957 and was essentially com-
pleted in early 1958.
406 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

REFERENCES

1. A. M. WEINBERG, Whither Reactor Development, Inter-American Con-
ference, Brookhaven National Laboratory, May 1957, N ucleonics 15(8), 99
(August 1957).

2. C. P. BAKER et al., Water Boiler, USAEC Report AECD-3063, Los Alamos
Scientific Laboratory, Sept. 4, 1944.

3. F. L. BENTzEN et al., High-power Water Boiler, USAEC Report, AECD-
3065, Los Alamos Scientific Laboratory, Sept. 19, 1945.

4. L. D. P. King, Design and Description of Water Boiler Reactors, in Pro-
ceedings of the International Conference on the Peaceful Uses of Atomic Energy,
Vol. 2. New York: United Nations, 1956. (p. 372)

5. C. K. BEck et al., Nuclear Reactor Project at North Carolina State College,
Nucleonics 7(6), 5-14 (1950); and Further Design Features of the Nuclear Reactor
at North Carolina State College, USAEC Report AECU-1986, North Carolina
State College School of Engineering, January 1952.

6. R. N. LyoN, Preliminary Report on the 19563 Los Alamos Boiling Reactor
Experiments, USAEC Report CF-53-11-210, Oak Ridge National Laboratory,
Nov. 30, 1953.

7 J. R. DietricH et al., Design and Operating Experience of a Prototype
Boiling Water Power Reactor, in Proceedings of the International Conference on
the Peaceful Uses of Atomic Energy, Vol. 3. New York: United Nations, 1956.
(p. 56)

8. L. D. P. King, personal communication, January 1952.

9. Reactor News, Nucleonics 15(2), R10 (February 1957).

10. S. E. BearL and C. E. WinteRs, The Homogeneous Reactor Experiment—
A Chemical Engineering Pilot Plant, Chem. Eng. Progr. 50, 256-262 (May 1954).

11. C. E. Winters and A. M. WEINBERG, Homogeneous Reactor Exzperiment
Feasibility Report, USAEC Report ORNL-730, Oak Ridge National Laboratory,
July 6, 1950.

12. C. E. WinTERs and A. M. WEINBERG, A Report on the Safety Aspects of
the Homogeneous Reactor Ezperiment, USAEC Report ORNL-731, Oak Ridge
National Laboratory, Aug. 29, 1950.

13. C. L. Secaser and F. C. Zapp, HRE Design Manual, USAEC Report
TID-10082, Oak Ridge National Laboratory, Nov. 18, 1952.

14. S. VisNer and P. N. HausenreicH, HRE Experiment on Internal Recom-
bination of Gas with a Homogeneous Catalyst, in Nuclear Sci. Technol., USAEC
Report TID-2506(Del.), Oak Ridge National Laboratory, 1955. (pp. 73-90)

15. R. B. Brigas and J. A. SwarTouT, Aqueous Homogeneous Power Reactors,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 3. New York: United Nations, 1956 (p. 175). S. E. BeaLrL and S.
VisneR, Oak Ridge National Laboratory, 1955. Unpublished.

16. D. O. CampBELL, Decontamination of the Homogeneous Reactor Experi-
ment, USAEC Report ORNL-1839, Oak Ridge National Laboratory, June 1956.
17. C. E. WinTers, Oak Ridge National Laboratory, 1952. Unpublished.

18. S. E. Bearr, The Homogeneous Reactor Test, in HRP Civilian Power
Reactor Conference Held at Oak Ridge, March 21-22, 1956, USAEC Report TID-
7524, Oak Ridge National Laboratory, March 1957. (pp. 11-47)
REFERENCES 407

19. S. E. BeaLr, The Homogeneous Reactor Test, in HRP Civilian Power
Reactor Conference Held at Oak Ridge National Laboratory, May 1-2, 1957,
USAEC Report TID-7540, Oak Ridge National Laboratory, July 1957.
(pp. 11-46)

20. S. E. Bearr, Homogeneous Reactor Experiment No. 2, 4th Annual
Conference of the Atomic Industrial Forum, October 1957.

21. 5. E. BeaLL, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak Ridge
National Laboratory, 1958.

22. I. Sriewak and H. L. FALKENBERRY, The Homogeneous Reactor Test
Mockup, in Proceedings of the Second Nuclear Engineering and Science Conference,
Vol. 2, Advances in Nuclear Engineering. New York: Pergamon Press, 1957.
(pp. 62-69) |
- 23. D. S. Tooms, Jr., Instrumentation and Controls for the HRT, Nucleonics
15(2), 48-52 (1957).

24. 5. E. BeaLL and R. W. JurRGENSEN, Direct Maintenance Practices for
the Homogeneous Reactor Test, Fourth Nuclear Engineering and Science Con-
ference, March 1958 (Paper 82); and USAEC Report CF-58-4-101, Oak Ridge
National Laboratory, April 1958.

25. T. RockweLL III, Reactor Shielding Design Manual. New York: D. Van
Nostrand Co., Inc., 1956.

26. S. GLASSTONE, Principles of Nuclear Reactor Engineering. New York:
D. Van Nostrand Co., Inc., 1955.

27. C. D. Borp and O. Sisman, How To Calculate Gamma, Radiation Induced
in Reactor Materials, Nucleonics 14(1), 46-50 (January 1956).

28. S. GLAssTONE, Principles of Nuclear Reactor Engineering. New York:
D. Van Nostrand Co., Inc., 1955.

29. 8. E. BeaLL, Containment Problems in Aqueous Homogeneous Reactor
Systems, USAEC Report ORNL-2091, Oak Ridge National Laboratory, Aug.
8, 1956.

30. P. R. KasTEN, Operational Safety of the H omogeneous Reactor Test, USAEC
Report ORNL-2088, Oak Ridge National Laboratory, July 1956.

31. 8. E. BeaLL and S. Visner, Oak Ridge National Laboratory, 1955.
Unpublished.

32. S. E. BEaLL and S. VisNer, Oak Ridge National Laboratory, 1955.
Unpublished.

33. P. M. Woop, A Study of the Possible Blast Effects from HRT Pressure
Vessel Rupture, USAEC Report CF-54-12-100, Oak Ridge National Laboratory,
Dec. 14, 1954.

34. D. FroMmAN et al., Los Alamos Power Reactor Experiments, in Proceedings
of the International Conference on the Peaceful Uses of Atomic Energy, Vol. 3.
New York: United Nations, 1956. (p. 283)

35. R. P. HammonDp and L. D. P. King, Los Alamos Homogeneous Reactor
Program, in HRP Civilian Power Reactor Conference Held at Oak Ridge, March
21-22, 1956, USAEC Report TID-7524, Los Alamos Scientific Laboratory,
March 1957. (pp. 168-209)
CHAPTER 8

COMPONENT DEVELOPMENT*

8—1. INTRODUCTION

The reliability of equipment for handling radioactive fuel solutions and
suspensions is considerably more important in homogeneous than in heter-
ogeneous reactors because the residual radioactivity of such equipment
after shutdown of the reactor precludes direct maintenance. The possi-
bility of failures of individual components in a homogeneous reactor,
moreover, is considerably increased by the corrosive or erosive nature of
the media being handled and the temperature fluctuations encountered
during startup and shutdown operations. The technical feasibility of cir-
culating-fuel reactors is so dependent on the behavior and reliability of
mechanical components that there is little likelihood that large-scale plants
will be built before the performance of each piece of equipment has been
adequately demonstrated. In this regard, the development of satisfactory
valves, feed pumps, mechanical joints, and remote-maintenance equipment
for large-scale plants appears to be most difficult.

The component development work at ORNL has been directed primarily
toward equipment for use in the Homogeneous Reactor Experiment
(HRE-1) and the Homogeneous Reactor Test (HRE-2). Although the
HRE-2 has both a core and a blanket, most of the components in these
two systems are identical and designed for use with solutions rather than
suspensions.

Since suspensions, or slurries, have not been used in either of the ho-
mogeneous reactors built by ORNL, the slurry equipment, problems have
received less attention than corresponding solution problems. Much of
the solution technology can be applied to slurries, although additional
difficulties such as the settling tendency of slurries, their less ideal fluid-flow
behavior, and their erosiveness must be taken into consideration.

The following pages give descriptions and illustrations of the aqueous
reactor components which have been selected and developed for use at
ORNL.

*By I. Spiewak, with contributions from R. D. Cheverton, C. H. Gabbard,
E. C. Hise, C. G. Lawson, R. C. Robertson and D. S. Toomb, Oak Ridge National
Laboratory.

408
8-2] PRIMARY-SYSTEM COMPONENTS 409

  
  

Alternate
OQutlet B

|

Diffuser
A\
/ ? \ Alternate

inlet Qutlet A

F1c. 8-1. Conceptual design of two-region reactor with slurry blanket. Arrows
indicate directions of slurry flow.

8-2. PrRiMARY-SYsTEM COMPONENTS

8-2.1 Core and blanket vessel designs. Core hydrodynamics. Flow
tests have been conducted on a variety of spherical vessels simulating
solution-reactor cores which have been selected to meet the following
criteria:

(1) Heat removal from all points must be rapid and orderly to prevent
hot spots from being generated.

(2) Radiolytic gas formed from water decomposition cannot be per-
mitted to collect in the reactor.

(3) The pressure drop should be low.

(4) The core tank should be maintained at a low temperature to prevent
excessive corrosion rates.

Three geometries which satisfy the above requirements have been in-
vestigated. The first, straight-through [1], involves diffusing the inlet flow
through screens or perforated plates [2] to achieve slug flow through the
sphere. The second, mized [3], involves generating a great deal of turbu-
lence and mixing with the inlet jet so that the reactor is very nearly iso-
thermal. The third, rotational [4], is somewhat between the first two; the
fuel is introduced tangentially to the sphere and withdrawn at the center
of a vortex, at the north and south poles.
410 COMPONENT DEVELOPMENT [cHAP. 8

In the straight-through core, used in HRE-2, the flow enters upward
through a conical diffuser containing perforated plates. The number of
perforated plates is determined by the ratio of sphere diameter to inlet-pipe
diameter. In general this ratio will be smaller for a larger reactor, resulting
in fewer plates and better performance. The velocity distribution leaving
the plates can be made to conform approximately to the flux distribution
of the reactor. As a result, the isotherms in the core are horizontal, and the
temperature rises smoothly toward the outlet at the top. The gas bubbles
rise upward at a velocity greater than that of the liquid and are removed
with the liquid. The over-all pressure drop is about 1.5 to 2.0 inlet-velocity
heads. The core tank is cooled by natural convection.

In the mixed core, illustrated in Fig. 8-1, the inlet and outlet are con-
centric at the top of the sphere. The inlet jet coincides with the vertical axis
of the sphere and is broken up when it hits the bottom surface. Except for
the cold central jet, the bulk of the core is at outlet temperature. The
velocity of eddies is great enough so that the gas bubbles travel along with
the liquid. The pressure drop is about 1.0 to 1.5 inlet-velocity heads.
The core-tank surface is maintained at a temperature very close to that
of the core fluid by the high turbulence.

In the rotational core, used in HRE-1, the flow pattern tends to produce
isotherms which are vertical cylinders. These are perturbed by boundary-
layer mixing at the sphere walls. The temperature generally increases in
the direction of the central axis, which is at outlet temperature. The gas
bubbles are centrifuged rapidly into a gas void which forms at the center
axis and from which gas can be removed. The gas void is quite stable 1n
cores up to about 2 ft in diameter, but in larger spheres the pumping re-
quirements to stabilize the void are excessive [5]. The pressure drop
~through a rotational core is a function of the particular system, but is
usually above 5 inlet-velocity heads.

Slurry blanket hydrodynamics. The suspension contained in the blanket
vessel must be sufficiently well dispersed to assure that a maximum of the
core leakage neutrons are absorbed within the blanket, the neutron reflec-
tion from the blanket to the core remains steady, and the transport of
fluids through regions of high heat generation are sufficient for heat re-
moval. The primary flow is taken through a jet eductor where the flow rate
is amplified and forced through a spherical annulus containing the high
heat generation region surrounding the core. It appears that amplification
gains of 2.5 are attainable. The outlet may be located either (1) concentric
with the bottom inlet or (2) at the top. Configuration (1) has the advantage
of high circulation rates in the region outside the shroud. Configuration
(2) has the advantage of better natural circulation in the event of a cir-
culating-pump stoppage.

Also under consideration is a swirling flow pattern similar to the rota-
tional flow which was described under cores.
8-2] PRIMARY-SYSTEM COMPONENTS 411

Reactor pressure vessels. Three principal types of stresses should be con-
sidered in designing the pressure vessels of one- or two-region reactors:

(1) Stresses resulting from the confined pressure.

(2) Thermal stresses resulting from heat production, and consequent
temperature gradients in the metal. |

(3) Stresses introduced by cladding if used. Because of the uncertain
residual stresses introduced during fabrication, this factor has not been
taken into account in the past.

The construction material can be chosen on the basis of corrosion re-
sistance and structural and thermal properties with little regard for nuclear
properties. Carbon steel with a stainless-steel cladding was selected for
use in the HRE-2.

Usually the pressure-vessel wall is thin in comparison with the inner
radius of the vessel; the ‘“‘thin-wall”’ formulas for calculating pressure
stresses are then applicable [6]. For precise calculations the general equa-
tions [7] for vessels with any wall thickness should be used. Thermal
stresses are superposed on the pressure stresses and can be approximated
by conventional formulas for hollow cylinders and spheres [8].

Solution of the stress equations depends upon a knowledge of the radial
temperature distribution, which, in turn, depends upon the manner in
which heat is generated in the metal wall and upon the temperatures at the
inner and outer surfaces. Heat is produced in the metal by the following
processes:

(1) The absorption of gamma rays arising from neutron capture, from
fission products, and from fission within the vessel.

(2) The recoil energy from the scattering of fast neutrons in the shell.

(3) The absorption of gamma rays produced by the inelastic scattering
of fast neutrons in the shell.

(4) The absorption of capture gamma rays produced as neutrons are
captured in the shell.

Although it may be possible to obtain the heat-production function for
the desired cylindrical or spherical geometry, it is simpler and usually
sufficiently accurate to obtain the leakage fluxes of gamma rays and neu-
trons mto the pressure shell for the desired geometry, and then to assume
that the heat-production function in the pressure vessel is the same as it
would be in a plate of the same material. Methods for obtaining the heat-
production function have been summarized by Alexander [9]. The function
can usually be described by the sum and difference of several exponentials.
For some purposes a single exponential can be used as a satisfactory ap-
proximation. The accuracy of the various methods has yet to be determined.
To arrive at a conservative design, reasonable methods indicating the
greatest amount of heat generation should be used. The surface tempera-
tures of the pressure vessel are estimated from a knowledge of the tem-
412 COMPONENT DEVELOPMENT [cHAP. 8

Core Access

  
    
  
 
 
 
 
 
  
  

To Fuel

Fuel Qutlet wrmm == LS e .
- & iy, e ; . Pressurizer

To Blanker

Blanket Outlet Pressurizer

Expansion Joint

Blast Shield

Core Vessel .

Cooling Coils

Diffuser Screens Pressure Vessel

Fuel
inlet

Blanket
Inlet

Fic. 8-2. HRE-2 reactor vessel assembly, fabricated by Newport News Ship-
building & Dry Dock Company.

peratures of the adjacent fluids and the heat-transfer relationships between
metal and fluids.

Chapman [10] has shown in an analysis of thermal stresses in spherical
reactor vessels that minimum thermal stresses are obtained when the
snner and outer vessel wall temperatures are approximately equal. Pressure
stresses decrease and thermal stresses increase as shell thickness 1s increased ;
o minimum combined stress occurs at an optimum wall thickness. Often
this stress is greater than the permissible design stress; thermal shielding
must then be provided between the reactor and pressure vessel to reduce
heat production and obtain a reasonable stress.

HRE-2 core and pressure vessel. The HRE-2 reactor-vessel assembly
presented a number of special design and fabrication problems [11].
Since it was desired to minimize neutron losses, Zircaloy-2 was selected as
material for the core tank, which is 32 in. in diameter and 5/16 in. thick.
The main pressure vessel, 60 in. in inside diameter and 4.4 in. thick, was
constructed of carbon steel with a cladding of type—347 stainless steel.
Because of uncertainties in the long-term irradiation damage of carbon
8-2] PRIMARY-SYSTEM COMPONENTS 413

steel, the pressure vessel was surrounded by a stainless-steel, water-
cooled blast shield which will stop any possible missiles from the reactor
vessel. Thermal radiation from the pressure vessel to the blast shield
permits the pressure vessel to operate at close to an optimum temperature
distribution from the thermal-stress standpoint.

A special mechanical joint was developed to join the Zircaloy core tank
to the stainless-steel piping system. A bellows expansion joint was used
to permit differential thermal expansion between core and pressure vessel.
Welding procedures were developed for joining Zircaloy and for making
the final girth weld in the clad pressure vessel entirely from the outside.

The HRE-2 core and pressure vessel are illustrated in Fig. 8-2.

8-2.2 Circulating pumps.* Pumps are required in aqueous homogeneous
reactors to circulate solutions and slurries at 250 to 300°C and 2000 psi
pressure, at heads of up to 100 psi. The two main considerations for these
pumps are that they must be absolutely leak free and that they must have
a long maintenance-free life.

At this time the only pumps considered capable of meeting these require-
ments are of the hermetically sealed canned-motor centrifugal type.
They consist of a centrifugal pump of standard hydraulic design and an
electric drive motor, built in an integral unit.

To illustrate, the 400A pump used to circulate fuel solution in the HRE-2
18 shown in Fig. 8-3. The HRE-2 blanket pump is identical except for
having a lower-output impeller. The hydraulic end of the pump is separated
from the motor by the thermal barrier, which is used to restrict the transfer
of heat and fluid from the scroll into the motor section of the pump. This
minimizes thermal and radiation damage to bearings and motor insulation.
The thermal barrier is built with sealed air spaces which aid in thermal
insulation. A labyrinth seal around the shaft is used to reduce the fuel
mixing into the motor. Water-lubricated hydrodynamic journal bearings
and pivoted-shoe-type thrust bearings are used to take the radial and
thrust loads, respectively. In the HRE-2, contact of the motor and
bearings with radioactive solutions is minimized by feeding distilled water
continuously into the motor.

The electric drive is a three-phase squirrel-cage induction motor with the
stator and rotor sealed in thin stainless-steel cans which prevent the
process fluid from coming in contact with the stator or rotor windings.
The cans are supported by the laminations to contain the system pressure
of 2000 psi. The motor is enclosed in a heavy pressure vessel which is
designed to hold the full system pressure in the event of a can failure.
The motor and bearings are cooled by the use of a small auxiliary impeller,

*Prepared from material submitted by C. H. Gabbard.
414 COMPONENT DEVELOPMENT [cuAP. 8

mounted on the rotor shaft, which recirculates motor fluid through a heat
exchanger. »

In the HRE-2, it is usually desirable to run the fuel pump at reduced
capacity during startups in order to limit reactivity changes. This 1s
accomplished by starting the 400A pump in reverse, which gives about one-
half of the normal flow. The ability to do this depends on the design of the
impeller and the size of the pump. In larger pumps, it is considered better
to use a two-speed motor to obtain the reduced-capacity operation. The
two-speed motor has an additional advantage in permitting the system to
be heated to operating density at reduced speed, thereby reducing the
required motor size and power consumption.

The service life of the 400A pump, based on out-of-pile tests with solu-
tions, is expected to be two years or more [12]. The slurry pumps currently
being operated have not proved as reliable as the solution pumps, but runs
of up to 3800 hr have been obtained [13]. The hydraulic parts of the pump
are frequently severely eroded, but there has not been a significant change
in the pump output or power requirements during the runs. The pumps
will generally continue to run unless a bearing seizes or breaks down. It
is expected that improvements in bearings and hydraulic design will make
slurry pumps as reliable as solution pumps.

The important problems in solution and slurry circulating pumps are
discussed below.

Stators. Pumps have been built with oil-filled stators to improve heat
removal from the windings and to balance the pressure across the stator
can. These pumps are undesirable for long-term reactor service because
the oil is subject to radiation damage and requires frequent replacement.
Pressure-balanced stator cans have also proved to be unsatisfactory
because of the difficulty in maintaining the proper balance. In pumps of
up to 400-gpm capacity, the problem of cooling the stator windings does
not seem too severe, and the dry-stator design with the can capable of
withstanding the full 2000-psi system pressure seems to be the best and most
commonly used type. In larger pumps, manufacturers are tending to use
a compound of silicone resin and inert filler material to improve heat re-
moval from the windings.

Most manufacturers insulate their motors with class H insulation con-
sisting of Fiberglas cloth impregnated with a silicone varnish binder. This
insulation is probably good for several years’ operation in circulating-
fuel reactors, depending on the radiation level of the pump, but over a
period of time the insulation can be expected to fail because of the decrease
in resistivity and dielectric strength. Hydrogen, released from the silicone
varnish during irradiation, may also build up enough pressure to rupture
the stator can when the system pressure is reduced. The HRE-2 is expected
to yield much information on motor life. Estimates made for the fuel
82] PRIMARY-SYSTEM COMPONENTS 415
Purge Water

Seal Weld Inlet

 Conduit Box

H_— Upper Radial Bearing

Cooling Coil —+— Coola: # Impeller

Stator Lamination “e—— Rotor

[RRA,

 

 

 

Thrust Bearing
i

 

 

t
" # Lower Radial Bearing
J

 

Seal Weld |
| Shaf: Labyrinth Seal
4~ Upper Wear Ring
Thermal Barrier Balance Ports
Seal Weld
. \ Impeller
Thermal Barrier ) .
~ Casing Wear Ring

 

Difruser: | l

F1a. 8-3. The Westinghouse 400A pump used to circulate fuel solution through
the HRE-2.

 

circulating pump of the HRE-2 indicated that the insulation will be sub-
Ject to failure in approximately five years, assuming that the outside of the
motor is protected by a 1-in. lead shield and the inside of the motor is
kept purged free of fuel solution [14]. Tests are being initiated at the
present time to determine the life expectancy more closely by irradiating
stators in gamma and gamma-neutron fields.

The ultimate solution to the insulation problem is probably the use of
ceramic insulation that would be completely radiation resistant. However,
considerably more development work will be required before this type of
insulation becomes usable. There are data available which indicate that
silicone-resin-bonded reconstituted mica (Isomica, trademark name of
Mica Insulator Co.) may have better radiation resistance than Fiberglas
and silicone varnish. If this material proves to be better from a radiation-
damage standpoint, it can probably be incorporated into a pump at a much
earlier date than the ceramic insulation.
416 COMPONENT DEVELOPMENT [cHAP. 8

Bearings. The standard Stellite-vs-Graphitar hydrodynamic bearings
have indicated little or no wear in pressurized-water systems. Presumably
their performance in aqueous homogeneous systems would be comparable
if the motor can could be kept in contact with the water only. In practice,
bearing life of 13,000+ hr has been achieved in actual contact with uranyl
sulfate solutions; however, continuous wear was observed, indicating that
eventually the bearing surfaces will fail.

Laboratory tests in water and tests on small pumps in solutions have
shown that aluminum oxide bearings and journals have superior wear
resistance as compared with the Stellite-vs-Graphitar combination. If
service tests conducted on larger pumps are successful, the aluminum
oxide bearings will be adopted as standard in solution pumps.

There is some doubt whether the hydrodynamic-type bearings currently
being used will be suitable for long-life slurry pumps. There have been very
few runs completed in which the bearings were not badly worn. However,
preliminary tests of small pumps with aluminum oxide bearings have
shown promise. It is planned, also, to evaluate the performance of hydro-
static (pressurized-fluid) bearings in dilute slurries.

In some cases, excessive wear has occurred in the thrust-bearing leveling
linkages of the 400A—type solution and slurry pumps. In this bearing the
thrust load is supported by a linkage system which used 1/8- and 1/4-in.-
diameter pins to transfer the load from link to link. It is uncertain whether
this wear at the contact point is caused by high stresses or by fretting cor-
rosion. A thrust bearing with line-contact linkages and alternate materials
at the contact points will be evaluated in an attempt to correct this problem.

Hydraulic parts. In uranyl sulfate pumps, excellent wear resistance is
obtained by using titanium for impellers, wear rings, and diffusers.
Stainless-steel hydraulic parts have also been used successfully in many
cases [15]. ’

The general design of slurry pumps is similar to that used for uranyl
sulfate pumps. The properties of the slurry are such that only a power
correction for the higher specific gravity is necessary in the hydraulic
design of the impeller. The coefficient of rigidity (viscosity) is generally
not high enough to require a correction to the head-capacity curve. A
most severe problem in slurry pumps is the combination of corrosive and
erosive attack on the hydraulic parts.

The primary difference in the design of a slurry impeller is the use of
radial balancing ribs on the top impeller shroud in place of the top wear
‘ring ‘on a conventional pump. In a conventional pump (Fig. 8-3) there
are small holes which vent the area within the top wear ring to the pump
suction pressure. This is done to balance some of the hydraulic thrust and
therefore reduce the load on the thrust bearing. In certain cases these
balancing holes have become plugged with slurry [16], which upsets the
8-2] PRIMARY-SYSTEM COMPONENTS 417

thrust balance and causes high thrust-bearing wear. The balancing vanes
eliminate one set of wear rings, which are subject to high attack rates,
and also tend to centrifuge the slurry particles to the outside, which aids
in preventing the slurry from entering the motor through the labyrinth
seal.

On the pumps currently in use, the damage to hydraulic parts is usually
limited to the wear rings, the tips of the impeller vanes, and to the volute
“cut water,” which is the point adjacent to the pump discharge where the
volute curve starts. The attack at these points can be reduced by proper
material choice and by using proper design of the flow passages. The best
materials which have been found for the hydraulic parts are Zircaloy—2 and
titanium, with Zircaloy being better in laboratory corrosion tests. There are
no test results for pumps using Zircaloy parts at this time, but vacuum-
~ cast parts have been obtained and placed into service. Other materials are
to be given laboratory corrosion tests, and promising materials will be
service tested.

The wear rings of the present pumps are being redesigned to provide
smooth throttling surfaces rather than the serrated type presently in use.
The smooth surfaces should reduce the turbulence and corrosion consider-
ably, with a very small increase in flow through the rings. One service test
has shown that the damage to this type of wear ring is decreased con-
siderably [17]. A test is being planned to determine whether radial vanes
on the lower impeller shroud similar to the balancing vanes discussed
earlier will reduce the attack rate on the lower wear rings. The radial
vanes will reduce the pressure drop across the wear rings and may reduce
the concentration of slurry flowing through them by centrifugal action.

It is uncertain whether a volute type scroll or a diffuser type scroll is
preferable. The volute type scroll has the advantage of having only the
cut-water subject to high attack, but has the disadvantage of having a
pressure drop across this point, resulting in perpendicular flow across the
cut-water. The diffuser has numerous points which could be eroded, but
~ the flow around these points should be smoother than that at the cut-water
and may not cause excessive damage.

The surface finish on the hydraulic parts is also very critical and the
surface variation should be held to 65 microinches or less. This is especially
evident at areas where the impeller surfaces have been ground during the
dynamic-balancing operation. If these areas are not properly finished,
the scratches will be severely attacked.

Thermal barriers. In pressurized-water pumps, the primary function
of the thermal barrier is to retard the transmission of heat into the motor.
In solution and slurry pumps, another function, that of preventing fluid
mixing from pump to motor, is of critical importance.

This mixing can occur at two places, at the shaft seal and around the
418 COMPONENT DEVELOPMENT [cHAP. 8

Impeller

  
   
  
   
 
 
 
  

 

Double Thrust Bearing

Upper Radial Bearing

Stator Cooling Coil
Stator Laminations
and Windings

 

Rotor Assembly

7f313/16in.

Belleville Spring To Load
Ring Joint Gasket
Lower Radial Bearing

Permanent Seal Weld .

Ring Joint Gasket

Passage to Pump
Suction Pressure

Throttling Surfaces

     
  

Shaft Labyrinth
Seal

Thermal Barrier

   
 

 
 

~

    
    
 
 
  
  

‘ Balancing Ports
Upper Wear Ring
Diffuser

Lower Wear Ring

Part of Pump Permanently
Installed In Piping System

 

Fia. 8-4. 6000-gpm top-maintenance pump for circulating solutions through a
50-Mw reactor, being built by Reliance Electric Company.

outer edge of the barrier. In the 400A pump, the mixing rate at the shaft
labyrinth seal has been reduced to 3 cc/hr by redesign of the seal and by
the use of a 5-gph purge flow through the motor [18]. Further improve-
ments in shaft seals are being attempted.

The seal around the outer edge of the 400A thermal barrier was originally
a mechanical joint, with the head developed by the pump across it. The
purge system did not develop enough pressure to prevent solution leakage
8-2] ' PRIMARY-SYSTEM COMPONENTS 419

through this joint into the motor. Any small leak was rapidly enlarged
by corrosion until excessive motor temperatures were reached. The prob-
lem was solved by seal-welding the joint. However, it would have been
preferable if the joint had originally been designed for welding.

Pump closures. Conventional canned-rotor pumps, such as those used
in the HRE-2, have a large seal-welded closure at the bottom of the stator.
Dismantling this closure for pump maintenance is impractical at the
present time because of the extremely high level of radiation at the
closure.

From a maintenance standpoint, a ‘‘top-maintenance’’ pump appears
to be advantageous. Direct-maintenance practices can be used to bolt and
unbolt the main flange. The pump casing is a permanent part of the
piping system. A top-maintenance pump being developed for the HRE-3
1s 1llustrated in Fig. 8—4.

In a top-maintenance pump, a mechanical thermal-barrier joint cannot
be avoided, since the barrier must be removable from the casing. The
joint must be loaded using the top closure bolts, and the entire mechanical
system must have some flexibility to compensate for differential thermal
expansion of the long motor. A venting system, shown in Fig. 8-4, is used
to eliminate the pressure drop across the thermal-barrier gasket so that
there will not be significant leakage even if the joint is not perfectly tight.
In this case, the purge flow should be effective in preventing leakage of
process fluid into the motor.

8-2.3 Steam generators. The performance of steam generators required
for homogeneous reactor service, measured in terms of undetectable leak-
tightness during long-term operation, considerably exceeds that of similar
units in conventional plants. Unfortunately, no method has yet been
developed of remotely locating and repairing leaks in a radioactive heat
exchanger without removing the entire unit. Failure of the steam generator
in a homogeneous power reactor, therefore, would lead to excessive shut-
down time and must be avoided if at all possible.

HRE-2 steam generators. The heat exchangers used in the HRE-2,
shown in Fig. 8-5, place reliance on the careful welding and inspecting
of tube-to-tube-sheet joints and the extensive thermal-cycle tests which
were carried out prior to actual operation in the reactor. In addition,
thermal gradients which would lead to excessive stresses during reactor
startup and shutdown are held within specified limits. Although the
units fabricated for HRE-2 have been tested with the most advanced
inspection methods available for both materidls and workmanship and have
met initial leaktightness specifications, only through operation of the
reactor will it be possible to judge the adequacy of these precautions.

The characteristics of the HRE-2 steam generators, which were manu-
420 COMPONENT DEVELOPMENT [caAP. 8

TaBLE 8-1

DEesigN DaTA rorR THE HRE-2 HEaT EXCHANGER

 

 

 

 

 

 

 

Shell side Tube side
Circulation rate, 1b/hr 1.62 x 104 1.79 x 105
Temperature in, °F 180 572
Temperature out, °F 471 494 .5
Operating pressure, psia 520 2000
Velocity, fps 67 (in outlet pipe) 11.3
Pressure drop, psi 18.5
Heat exchanged, kw 5000 (1.71 X 107 Btu/hr)
Fouled Ug, Btu/(hr)(ft2)(°F) 670 (based on Up = 3/4 Ug)
Heat-transfer area, ft2 480
Tube outside diameter, in. 0.375

 

factured by the Foster Wheeler Company, are summarized in Table 8-1.
In fabricating these steam generators, all-welded construction was used on
components that were to be exposed to the process solution. Interpass
leakage is controlled by use of a gold gasket. Considerable attention was
given to obtaining the highest quality tubing, which was inspected by
ultrasonic and magnetic flaw detectors capable of detecting imperfections
as small as 0.002 in. Following the bending and annealing operations,
each tube was inspected for surface defects with a liquid penetrant and
subjected to a 4000-psi hydrostatic test. After passing all these tests, the
tubes were rolled into the tube sheet and welded by an inert-gas-shielded
tungsten-arc process. Quality-control welds were made periodically during
the tube-joint welding and were subsequently examined by radiographic
and metallographic methods.

After fabrication, the units were subjected to 50 primary-side thermal
cycles covering temperature changes more severe than those likely to be
encountered in subsequent operation. The units were then helium-leak-
tested at atmospheric pressure with mass-spectrometer equipment capable
of detecting leakage lower than 0.1 cc of helium at STP per day. Leaks
were repaired and the thermal-cycle test and leak test were repeated until
no leakage was detectable.

The HRE-2 steam generators were thermal cycled with diphenyl as the
heating medium. After the test, extensive carbon deposits were found in
the tubes. After considerable difficulty, the deposits were removed by
high-temperature flushing with oxygenated water and uranyl sulfate solu-
tion. Future thermal-cycle tests will be made with steam as the heating
medium.
PRIMARY-SYSTEM COMPONENTS 421

8-2]

"uor)Brodio)) IS[O9Y A\ -19380,] Aq pojeBoLIqe] ‘198uByIXe Y8oY UIBW Z-YH 9YL ‘S-8 'PI]

 

 

ura9Hel

Bl

aun
umopmolg

 

SUOI}>3UU0Y)
|]2A97-pinb1y

e UlV.E $ ¢

 

 

 

 

 

 

,

 

 

 

 

 

 

~— uin1qg

 

 

194A1q wosyg
~~ @

 

 

  

 

4 LA T ]

 

 

 

 

 

LR ____z_____E;__E_.__:

 

\

10}jp21pu|
ainssald 104 do|

\
TR

7

2

L AN

 

SOONNNANANNNNNT

 

 

 

 

 

 

[994G UOQDD
PRID-[834G-SS8juIDIS

             
      

 
422 COMPONENT DEVELOPMENT [cHAP. 8

Level Control Water
Connection Thermal Level

Insulation

     
  
 
 
   
   
 

Blast Shield

Inlet

Feedwater Header

Inlet

Outlet
Header

 

Blowdown
Level Control Thermal
Connections Sleeves

Fia. 8-6. The HRE-2 spare heat exchanger, fabricated by Babcock & Wilcox
Company.

The tube-to-tube-sheet joint is the most damage-sensitive portion of the
steam generator. The primary side is subject to corrosive fuels, and the
secondary side is subject to crevice corrosion and stress-corrosion cracking.
Primary-side corrosion is controlled satisfactorily by maintaining velocities
below 15 fps and minimizing high-velocity turbulence in the headers.
Secondary-side corrosion is limited by strict control of boiler-water chem-
istry, particularly chloride content. One method which has been proposed
for eliminating the stress-corrosion problem is the use of composite tubing
such as stainless steel-Inconel, where the two materials are exposed only
to fuel solution and boiler water, respectively.

Another problem in the operation of steam generators in a radioactive
environment is the generation of radiolytic oxygen in the boiler water.
This oxygen is stripped very rapidly by the steam, which contains about
2 ppm of oxygen. Hydrogen is released at the same time. The corrosivity
of this mixture is not yet known, but it can be controlled by the use of
inhibitors and by proper selection of materials for use in thin metal sections
where pitting attack is undesirable.

HRE-2 spare steam generator. The steam generator shown in Fig. 8-6
was constructed as a possible replacement in case of failure of an HRE-2
steam generator. Although the over-all geometry of this unit, fabricated
by the Babcock & Wilcox Company, conforms to the space requirements of
the present steam generators, the design was changed to minimize the
possibility of stress-corrosion cracking of the tubes on the shell side by
eliminating crevices in contact with boiler water.

The steam generator contains eighty-eight 5/8-in.-OD, 0.095-in.-thick,
type—347 stainless-steel tubes. The tubes have multiple U-bends to provide
8-2] PRIMARY-SYSTEM COMPONENTS 423

the required length for heat-transfer surface. Each tube is brought out
through the shell of the exchanger, and then all the tubes are collected in
the inlet and outlet headers. Thermal sleeves are utilized at every con-
nection of the stainless tubes to the carbon-steel shell wall. Their function
is to prevent high thermal stresses in the tubes by distributing the tem-
perature gradient between shell and tubes along the length of the thermal
sleeves. The normal crevice between tubes and tube sheet, which is the
site of possible corrosion failures, is eliminated. Each sleeve consists of an
austenitic type—-347 stainless-steel section which is welded to the tube on
one end and to a carbon-steel section of the sleeve on the other. The
carbon-steel sleeve is then welded to the carbon-steel shell to seal the
secondary side. Only austenitic type-347 stainless steel is exposed to fuel
solution.

Slurry steam generators. The mechanical design of heat exchangers for
slurry service should not differ greatly from that for solution service.
However, the design must assure that

(1) The pressure drop across all tubes is sufficient to maintain the slurry
in suspension.

(2) The headers have no stagnant regions where sediment can accumu-
late.

(3) The tube-sheet joints are sufficiently smooth to prevent fretting
corrosion by the slurry.

(4) The headers and tubes drain freely.

From the heat-transfer relationships for Bingham plastic slurries,
described in Article 4-4.5, it is evident that for optimum design of steam
generators the flow of slurry through the tubes should be turbulent.

Large heat exchangers. The Foster Wheeler Corporation has prepared
preliminary designs of 50- and 300-Mw heat exchangers [19]. Both single-
drum integral units and units with separate steam drums were considered
in the 50-Mw size; only two-drum units were considered in the 300-Mw
size. Two-drum units, in general, give operational characteristics superior
to those of integral units, but require more shielded volume and reactor
space for installation. The two-drum unit has more stable steam generation
at power and provides greater assurance of high steam quality. The major
problems introduced by increasing size are higher tube-sheet thermal
stress and increased difficulty in the manufacture of large forgings.

The 50-Mw design employs approximately 2200 tubes 3/8 in. in diameter
(5960 ft* of heat-transfer area); the 300-Mw design uses approximately
11,400 tubes of the same size (32,000 ft2). Most of the designs have stainless
steel clad on steel for tube sheets and heads, and steel for steam shells.

8-2.4 Pressurizers. A pressurizer is required in an aqueous fuel system
to provide (1) sufficiently high pressures to reduce bubble formation and
424 COMPONENT DEVELOPMENT [cHAP. 8

cavitation in the circulating stream, (2) reactor safety by limiting the
pressure rise accompanying a sudden increase in reactivity, and (3) a surge
chamber for relief of volume changes.

Three general methods of pressurizing have been used in test loops
and experimental reactors:

(1) Steam pressurization, such as is used in the HRE-2, where liquid
in the pressurizer is maintained at a higher temperature, hence a higher
vapor pressure, than that of the circulating system.

(2) Gas pressurization, where liquid in the pressurizer is at the same
temperature as the circulating system but excess gas is added to the vapor
above it; if the pressurizing gas is free to diffuse into the circulating liquid,
it reduces the solubility of radiolytic deuterium and enhances bubble
formation.

(3) Mechanical pressurization, where pressure is maintained with a
pump and relief valve; this system is most satisfactory except that it 1s
difficult to relieve sudden large volume changes following a reactivity
change. This system therefore has been limited to nonnuclear test loops.

Solution pressurizers. Solution steam pressurizers must satisfy rather
strict chemical criteria. Stainless-steel surfaces in contact with solutions
must not exceed temperatures at which heavy-liquid-phase solutions form,
giving rise to rapid corrosion [20]. Undesirable reduction of uranium must
be avoided by the presence of some dissolved oxygen. Undesirable hydroly-
sis of uranyl ion must be avoided by control of the chemistry and tem-
perature in pressurizer solutions [21]. The vapor-phase concentration of
deuterium should be maintained below the explosive limit. One solution
to these problems, used in the HRE-2, is the generation of steam from dis-
tilled water rather than from fuel solution. Another solution is the boiling
of solutions in corrosion-resistant titanium. A third solution is the use of
fission-product heating rather than external heating to reach the desired
temperature.

Gas pressurizers using Oz gas are attractive from the solution-stability
standpoint. Care must be exercised to prevent excessive amounts of dis-
solved oxygen appearing as bubbles in the circulating reactor stream. This
can be accomplished either by continuous letdown of fuel solution or by
use of a mixed steam-gas pressurizer where gas supplies only a portion of
the desired overpressure.

Heat may be supplied to pressurizers by several methods. Electrical
heating of pipes, used in the HRE-2, is very convenient but makes con-
trol of surface temperature difficult. Heating media such as steam, Dow-
therm, liquid metals, etc., simplify the temperature-control problem but
introduce costly auxiliaries. Fission-product heating is simple but rather
difficult to regulate. Since the heating problem is so complex, selection of
an optimum system for a specific application is quite difficult.
8-2] PRIMARY-SYSTEM COMPONENTS 425

Purge
Water Level
) Steam Baffle
Dry Pipe
‘ Surge Pipe

Thermocouple
Well Vent ~e— % Drain Pipe

NNNNWZP PP 22D

 

 

 

 

 

   
      
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

! ! ()
Fuel ; ' ' .' |
Solution Level 5 1V ' Purge Water -
| n ‘| Overfl
! g 4 vertiow
d N adl
| g g |
To Reactor 7/ ; éé t
System ™ +—_“—“" : / | ; ,
L :@J | |
G e
/ \
Downcomer Electric
Lower Header Heater

Purge Water
Inlet

F16. 8-7. HRE-2 pressurizer.

Deuterium concentration may be controlled by designing the pressurizer
in a manner which makes buildup of gas improbable, by contacting the
vapor with solid or solution catalyst, or by venting.

HRE-2 pressurizer. Several design configurations were studied for
steam-pressurizing the HRE-2 core system [22]. Both an integral type
unit (Fig. 8-7) incorporating a steam generator and fuel surge volume
within the same vessel, and a two-unit system utilizing separate steam-
generator and surge-volume vessels were considered. The basis of both
systems was the vaporization of a stream of “‘clean’ purge water, pumped
from the reactor low-pressure system, to obtain the required steam over-
pressure.

The integral unit was chosen because of its simpler design and its ability
to maintain a very low dissolved-solids concentration in the boiling water.
Approximately 609, of the purge water overflows and 409, is vaporized.

In determining the internal configuration of the unit, it was necessary
to establish a second basic design criterion. Owing to the nature of the
system selected, continuous operation of the purge pump is essential to
maintaining steam overpressure. Since this type of pump may fail, it was
decided that sufficient water should be stored in the steam generator to
maintain full operating pressure for at least 1 hr after a purge-pump failure.
This appeared to be adequate time for either emergency repair of faults
on the oil side of the purge-pump system or arrangement of an orderly
shutdown.
426 COMPONENT DEVELOPMENT [cHAP. 8

 

—— Vapor i
Antivortex
2 e Baffles
H ——| Liquid
o= b — Water 3 -
- “ i : Jet

Cast Eductor
Electric

Heater

Slurry i
Optional

Nozzle -

Antivortex
Baffles

 

 

 

 

 

 

 

 

ot
v, Q %ot

 

Dowtherm
Heater

 
  
 
  

Steam Separator

 

 

Solution | anal
\w'a’s oy

 

Cooler
Fuel Storage Tank
()

Fic. 8-8. Typical pressurizers. (a) Slurry steam or gas pressurizer. (b) Solution
or slurry gas pressurizer. (c) Boiling-solution pressurizer.

A general description of the basic design and system operation follows:
The low-pressure condensate, pumped to loop operating pressure by a
diaphragm pump, passes through the letdown heat exchanger, where it is
preheated to about 280°C and then enters the lower header of the steam
generator, as shown in Fig. 8-7. Clamshell heaters are attached to four
pipes inclined upward at 55 deg. Heat-load calculations indicate that
32 kw are required under rapid-startup load conditions; however, at steady
state the heat load becomes 12.4 kw. Part of the purge water entering the
heater legs (40 lb/hr) is vaporized to provide the desired steam overpres-
sure. The remainder enters a storage pool in the main pressurizer drum.
Natural recirculation of this water occurs through two downcomers.

The storage-pool level is maintained constant by allowing excess purge
water to be removed through a series of 1/8-in. orifice holes located in the
end plate of the steam generator. A 1/2-in. hole centered 5/8 in. above
the orifice holes is provided as an overflow in case the orifice holes become
plugged. A method of increasing the surge volume, without changing the
8-2] PRIMARY-SYSTEM COMPONENTS 427

storage pool level, is to insert a pipe from the surge chamber through the
storage pool.

A single pipe connects the pressurizer surge chamber to the main core
loop. Liquid level in this pipe is maintained at a point about 10 in. below
the inside diameter of the pressurizer drum by a liquid-level controller.

The clamshell heaters are carefully machined to fit the heater pipes and
strongly clamped to promote contact. There are eight separate Calrod
heaters in each clamshell, so that failure of a few heater elements will not
affect the pressurizer greatly.

Vapor-phase deuterium is kept under control in that during normal
operation it has no way to enter the pressurizer. If a small amount of gas
does enter the pressurizer it will be dissolved in purge water overflowing
into the reactor system. A large amount of gas would be vented.

Slurry pressurizers. The physical problems of slurry pressurizers are
similar to those of solution pressurizers. The chemical problems are fortu-
nately not present. The pressurizer may be designed to promote settling
of solids so that pure supernatant water is available as a working fluid. It
1s necessary, however, that the pressurizer be designed to prevent accumu-
lation of cakes or sludges. This is usually accomplished by flowing all or
part of the circulating stream through the bottom of the pressurizer. This
must be done carefully in steam pressurizers to prevent mixing of cool
circulating fluid with the heated pressurizer fluid above.

Typical pressurizer designs. Several pressurizer designs applicable to
test systems or reactors are illustrated in Fig. 8-8.

In the slurry steam or gas pressurizer (a) slurry at circulating temperature
sweeps the bottom of the pressurizer tank, preventing the formation of
cakes. Two nozzle arrangements are shown. The baffles are used to mini-
mize turbulence and mixing in the system.

In the solution or slurry gas pressurizer (b) a jet 1s used to contact fuel,
which contains a liquid-phase catalyst, with pressurizer vapor. This main-
tains the vapor at a low D2 concentration. The high-velocity regions in
this system should be constructed of special wear-resistant inserts such as
titanium or zirconium.

- The boiling-solution pressurizer (c) has a fuel storage tank where solution
just out of the reactor is heated by its own fission-product decay and
dissolved D2 recombines nearly quantitatively. Additional heating is
supplied, if necessary, by condensing Dowtherm. The mixture of steam
and fuel solution is separated; the steam flows into the pressurizer and the
fuel is cooled to a chemically acceptable temperature before re-entering the
circulating system. The fuel storage tank, Dowtherm heater, steam sepa-
rator, and solution cooler are made of titanium. High-strength alloy
Ti-110-AT is preferred to commercially pure titanium in order to increase
the strength of these parts.
428 COMPONENT DEVELOPMENT [cHAP. 8

8-2.5 Piping and welded joints. The various codes [23] dealing with
pressure piping have proved very satisfactory for determining the strength
required for reactor piping. Pipes are sized on the basis of experimentally
determined maximum velocities for low corrosion and/or erosion rates.
The austenitic stainless steels are used for most piping applications.

Because of the fact that the piping system of a homogeneous reactor must
be absolutely leaktight throughout its service life, care is exercised in
selecting pipe of the highest obtainable quality. The chemical composition
and corrosion resistance are checked. Rigid cleanliness is maintained during
fabrication to prevent undesirable contaminants.

The design of solution piping systems must eliminate stagnant lines
where oxygen depletion may cause solution instability and plugging.
Slurry piping systems should be designed to prevent settling, which can
cause plugging or make decontamination very difficult.

Piping layouts. In laying out the piping system for an aqueous-fuel
homogeneous reactor, sufficient flexibility must be incorporated in the
system to absorb thermal expansions without creating excessive stresses
in the pipe wall, and to avoid high nozzle reaction loads at the equipment.

Equipment must be located where it will be accessible for maintenance,
and the piping adjacent to such equipment must be placed so that it can
be disconnected and reassembled remotely. These requirements may
result in a piping system of excessive length with resultant high fluid
holdup and pressure drop. The final design, therefore, must be a com-
promise between the various conflicting requirements of flexibility, main-
tenance, holdup, and pressure loss in the line.

Methods of piping analysis and evaluation as presented by Hanson and
Jahsman [24] may be used for analyzing the piping layouts in homogeneous
reactor systems. An application of the Kellogg method [25] was used by
Lundin [26] to analyze stresses in the HRE-2 system. Specific rules on
how to absorb the effects of thermal expansion of a piping system by the
provision of a flexible layout are given in the Code for Pressure Piping,
ASA B31.1-1942, Sec. 6.

Welded joints. Welded joints are recommended in preference to mechani-
cal joints for reactor piping. Welds are made approximately equal in
strength and corrosion resistance to the base metal. Pipe and fittings are
designed to utilize full-penetration butt welds throughout the piping sys-
tem. Welds that are to contact process fluids are inspected thoroughly to
ensure that no crevices are present and that penetration is complete. Such
defects could result in crevice corrosion leading to leaks.

The first 1/8 in. on the process side is deposited by use of bare-wire filler
metal and inert tungsten-arc welding techniques. The ferrite content of
the deposit is controlled to minimize the possibility of cracking. This de-
posit is inspected visually and with penetrant. If the weld thus far contains
8-2] PRIMARY-SYSTEM COMPONENTS 429

    
  

. Carbon
Ring Gasket Steel Bolts

Leak Detection
Assembly
(Autoclave Fittings)

F1a. 8-9. HRE-2 ring-joint flange, showing leak-detector connection.

no visible defects, it is radiographed to ensure freedom from all defects.
The balance of the weld is then deposited from either bare or coated wire.
The completed weld is inspected again with dye penetrant and radiography.
A small number of inclusions or porosity are permitted in the final layers
which contact air.

Although welding techniques for clean piping are very satisfactory,
remote-welding procedures for repair of contaminated reactor systems are
only in the development stage. It is desirable to develop methods for
remote cutting, positioning, welding, and inspection of joints, particularly
in large pipes. Possession of these techniques would greatly increase the
maintainability of circulating-fuel reactors.

8-2.6 Flange closures. Piping flanges. Because a practical machine for
remotely rewelding pipe has not yet been developed, equipment which
must be removable from the system for replacement or maintenance must
be connected to the system with mechanical joints. Several types of
mechanical joints applicable to pressure systems have been described in
the literature, but most have been eliminated from consideration for ho-
mogeneous reactor service because their reliability with respect to leak-
tightness following thermal cycling has not been adequately demonstrated.

The ring-joint flange (Fig. 8-9) incorporating American Standard
430 COMPONENT DEVELOPMENT [cHAP. 8

welding-neck flanges and ring-joint gaskets, as described in ASA Standard
B16.20-1956, was used in the HRE-2 and is considered to be the most
reliable closure for reactor piping systems to date. Since there are two
sealing surfaces in this type of joint, it is ideal for leak-detector purge
systems, described below, which prevent even minute leakage of con-
taminated fluids into the shield.

To enforce proper dimensions for HRE-2 gaskets and grooves, special
master rings and grooves were manufactured to measure dimensions to
40.0001 in. Although ASA tolerances of £0.006 in. on pitch diameter are
acceptable, it was convenient to obtain manufacturing tolerances of
+0.001 in. by using these gages and masters. The application of these
tolerances has permitted greater accuracy in fit-up and assures uniform
contact between the ring-joint gasket and the grooves of both mating
flanges. Soft oval or octagonal type-304 ELC stainless-steel rings are used
against the harder type—-347 stainless-steel grooves. Typical leakage ex-
perienced in a 4-in. 2500-1b flange is 6 X 107 ° g of water per day at service
conditions.

The bolting of flanged joints presents a serious problem because the
bolts, under thermal cycling, loosen up after only a very few cycles, thus
threatening the integrity of the joint. Whereas the flange bolts of a con-
ventional pressure piping system may be retightened after a few cycles,
this becomes impractical in a homogeneous reactor system after the reactor
has gone critical. In the HRE-2 it was found desirable to initially stress
the low-alloy steel flange bolts to an average loading of 45,000 psi, as
indicated by a torque wrench. After about three thermal cycles this bolt
loading fell to an asymptotic value of approximately 30,000 psi, which was
found to be adequate to maintain the integrity of the joint indefinitely
throughout further operations; no retightening of the bolts was found to
be necessary [27]. Some deformation of flange grooves and ring-joint
gaskets was found as a result of these high loadings. However, with the
use of flanges and rings machined to the close tolerances noted above, there
was no leakage even after test joints had been opened and reassembled
ten to a hundred times.

Although bolt torque measurements are usually considered very ap-
proximate indications of load, special techniques were developed which
gave reproducibility to 4-10%. Bolts were lubricated with molybdenum
sulfide, and nuts were tightened several times against test blocks which
approximated the flange spacing. Nut-and-bolt combinations were ac-
cepted for use after reproducible compressive stresses were produced in
the test blocks for given torques.

Bolts may be loaded more precisely with the use of pin extensometers.
In this technique, a pin is spot-welded into one end of a hole drilled axially
through the bolt centerline. A depth gage measures quite precisely the
8-2] PRIMARY-SYSTEM COMPONENTS 431

relative strain between the loaded bolt and the unloaded pin. Since both
pin and bolt are at the same temperature, thermal effects are compensated
automatically. Extensometers are inconvenient for contaminated main-
tenance, however.

Because mechanical joints may be expected to leak, some provision must
be made to supply pressure greater than reactor system pressure to the
undersides of the ring-joint grooves. By this means, leaks may bhe detected
by observations of a drop in pressure in the auxiliary system. At the same
time inleakage of a nonradioactive fluid to the reactor system in the event
of a leak prevents radioactive spills. In the case of the HRE-2, D20 is
supplied to the sealed annuli formed in the gasket grooves at a pressure
approximately 500 psi greater than that in the reactor system. A hole is
drilled through one flange at each pipe joint to the annulus of the ring
groove; the ring-joint gasket is also drilled to interconnect the annuli of
the two flanges. Heavy-wall, 1/4-in.—OD stainless-steel tubing connects
each flange pair with a header and pressurizer in the control area. ORNL
experience has indicated that water is more satisfactory as a leak-detector
fluid than gas, because its pressure change is a more sensitive leak indi-
cation and because its surface tension reduces the magnitude of small leaks.

Vickers-Anderson joints.* From a remote-maintenance point of view,
a flange requiring as few bolts as possible is desirable. Adaptation of the
Vickers-Anderson type closure appears to be a possible approach to the
problem, since it obtains uniform circumferential tightening with only
two bolts. In this type of joint, two split clamshell pieces are pressed
together with the two bolts. The clamshells bear on conical flange faces
which transform the tangential bolt forces into forces parallel with the
pipe centerline.

Usually a pressure-seal type of gasket is used with the above type of
closure because it is difficult to exert sufficient axial load to seat a ring-joint
gasket. Unfortunately, the present leak-detector concept is not applicable
to such a gasket. Therefore, to permit use of Vickers—Anderson joints in a
reactor, either a new gasket or a different leak-detector concept would
have to be developed.

Bi-metallic joints. A two-region reactor may have the problem of ob-
taining a leaktight low differential pressure mechanical joint between the
two regions. In the HRE-2, the regions are separated by a zirconium
vessel that must be joined to stainless-steel piping. At this time, welding
and brazing techniques for joining the two materials are unsatisfactory.
Conventional flanges are not satisfactory because of the large difference
in thermal coefficients of expansion for the two materials. A solution to the
problem for the HRE-2 was obtained by using a titanium cylindrical-

*Based on material submitted by R. D. Cheverton.
432 COMPONENT DEVELOPMENT [cHAP. 8

7 .
H i
v - 7
Bottom Pressure Vessel Flange— | | ]
Stainless Steel ' : , |
Bottom Core Vessel ' / 7

Flange—Zircaloy-2

 

 

 

 

 

Concentric Titanium Rings

7
Gold Foil (Top and Bottom) ////

Titanium Studs Z,

 

 

Titanium Ferrules

Stainless Steel Flange /Z » 5 !

Titanium Nuts N ; L 'i~
L ool

—

 

 

 

 

 

 

 

 

 

 

 

 

 

=———1l

 

 

 

 

 

 

 

 

 

 

 

 

 

Fie. 810. Zircaloy-2 type—347 stainless steel transition joint for HRE-2
pressure vessel.

sleeve gasket with gold inserts for sealing, shown in Fig. 8-10. The gasket,
which consists of four concentric rings, flexes radially to absorb thermal
expansion. The gold-capped surfaces of the gasket that make the seal are
not permitted to rotate or slide relative to the flanges. The joint is loaded
with titanium-alloy bolts.

A ring joint for connecting stainless steel and titanium piping at tem-
peratures up to 650°F has also been developed [28] for possible use with
titanium letdown heat exchanger in the HRE-2. The different thermal
coefficients of expansion of the two materials are bridged by use of a
stainless-steel clad carbon-steel flange in the stainless half of the joint. An
HRE type of leak detector is placed in the clad flange.

8-2.7 Gas separators. The problem of removing relatively small amounts
of gas from a stream of liquid is usually solved by using a settling tank
which permits bubbles of gas to rise to a free surface. In applications in
which the amount of liquid holdup is critical, this approach has the serious
drawback of requiring too much liquid. However, a chamber which im-
parts centrifugal force to the liquid and “forces” the gas to a free surface
before the liquid leaves the chamber offers a possible solution. Of the
several types of separators which can be used, one of the most promising
is the pipeline or axial gas separator (Fig. 8-11) used in the HRE-2.

The pipeline gas separator consists of stationary vanes or a volute,
followed by a section of pipe in which gas is centrifuged into a void which
forms at the pipe axis. whence it is removed. The energy of rotation is
8-2] PRIMARY-SYSTEM COMPONENTS 433

Stainless
Steel Pipe

   
 
 

     
   
  
   

4.500-in. OD

    
 

Recover
Vanes y

 

 

 

Titanium Sleeve

 
 

|

30%-in. Overall— A—l
Gas Takeoff Pipe

 
  

 

Fiac. 8-11. HRE-2 gas separator.

partially recovered with vanes or a volute at the discharge end of the
separator.

A model of the gas separator was built and runs were made to test vanes
and volutes of different types for energy conversion and recovery, and
for gas-removal efficiency [29]. It was found possible to control the size
of the gas void by design of the takeoff nozzle. A separator utilizing vanes
was selected for the HRE-2 on the basis of ease of fabrication and because
high-efficiency recovery vanes could be designed. Titanium was used as
a construction material because of its excellent corrosion resistance under
highly turbulent conditions.

The design criteria for vane-type gas separators are discussed in the
following paragraphs.

Pressure distribution. The pressure drop of the liquid stream through a
well-designed gas separator can be approximated by assuming an efficiency
of conversion of pressure to velocity in the rotation system of 909, and a
recovery efficiency of 809,. Frictional drop in the vortex will be about
three times that which would be predicted if the absolute velocity of the
vortex near the wall were in axial flow. The Ap across the HRE-2 separator
is five inlet-velocity heads or 5 psi.

Length of separator. Length is usually selected to be that necessary to
bring a bubble from the periphery into the central void during the time the
bubble 1s moving axially through the separator. For the HRE-2 separator
about two pipe diameters are required, but for larger separators the length-
to-diameter ratio increases.

Vortex stabilsty. The degree of rotation for stable operation is such that
the centrifugal forces on a bubble are greater than the gravitational forces.

The dimensionless group expressing the ratio of these forces is V;/ \/&i,
where V,; is tangential velocity (ft/sec), g is 32.2 ft/sec?, and r is radius (ft).
For a stable vortex this ratio must be greater than 1; for best results it
should be greater than 4.

Entrainment. The minimum amount of entrainment for a given separator
is determined by the void stability and the geometry of the gas takeoff
434 COMPONENT DEVELOPMENT [cHAP. 8

         
     
 
  

 

 

 

D50 Steam
15 HP Motor Gear Reduction Unit lodine . 2
Recombiner Sparge
‘ " hani Remover ‘
Thermal Shields 0 Mecc::r?::::l :eol Entrainment Vapor Steam Outlet To
Vapor Exit =57 9 Separator Condenser Separator
[H{B= Slurry Inlet Section Condensate nd Condenser
Outlet Pipe i Fulllevelyy, Return Full Level
+Steam Heating
Paddle Coil
Steam
Longitudinal Jacket Slurry
Baffles Outlet Slyrry Outlet
(a) (b) ()

Storage Tank

14-in. Pipe Baffle Plates

 

 

   

 

N\ /
= ft_ —_Ti_;_ A_T‘—__‘—T—_ = FU" Level
§ Steam Jacket
-
7260

 

4-in. Pipe

Fig. 8-12. Fuel storage tanks. (a) Mechanically agitated tank (Westinghouse).
(b) Boiling tank. (c) Sparged tank. (d) HRE-2 storage tank-evaporator.

nozzle. It is advantageous to have a high stability number. The nozzle
should be paraboloid facing the stream and have a small takeoff port.
Tests of the HRE-2 separator indicate that entrainment can be limited to
0.1 gpm at liquid and gas throughputs of 400 gpm and 4 gpm, respectively.
Tests of a 5000-gpm separator with 29, gas gave 1 gpm of liquid entrain-
ment, minimum.

Gas-removal efficiency. One hundred percent removal of large gas bubbles
has been achieved in test models. Removal of very small bubbles is con-
siderably less efficient for gas separators of normal length.

Although gas has been removed from slurries in an axial separator, the
design criteria are not known. The primary difficulty lies in the interaction
between small solid particles and bubbles, which may foam. The rates
of bubble rise in slurries have not been measured.

8-3. SupPPORTING-SYSTEM, COMPONENTS

8-3.1 Storage tanks. Solution tank-evaporator. In the HRE-2 the stor-
age tank system has a threefold purpose: (a) it acts as a storage tank for
fuel solution during shutdowns and after removal from the high-pressure
system during the letdown of gaseous decomposition and fission products,
and after emergency dumping; (b) it acts as a generator of diluent steam
8-3] SUPPORTING-SYSTEM COMPONENTS 435

to lower the radiolytic D2 and Oz concentration to a nonexplosive mixture
prior to recombination; and (c) it serves as a purge-water generator.

The storage tank-evaporator is designed to furnish the required amount,
of steam diluent and purge water and also to agitate and mix the solution
stored in the tank. The HRE-2 evaporator is shown in Fig. 8-12(d).

To keep the fuel solution well mixed, it was desired that the solution be
agitated by recirculation through the tank at a high rate. The recirculation
rate is such that the frictional loss in the vaporizing circuit is equal to the
hydrostatic driving force on the vaporizing fluid. For the HRE-2 evapo-
rator, a ratio of 176 1b of liquid circulating for every pound of steam
generated was obtained [30].

The circulation of liquid through this type of long horizontal tank was
found to set up waves which interfered with vapor withdrawal. This inter-
ference was minimized by the use of baffle plates, as shown in Fig. 812(d).

Slurry storage tanks.™ Three approaches are being pursued in the develop-
ment of slurry drain and storage tanks for reactor use: mechanically
agitated tanks, agitation by steam-sparging the tank, and agitation by
surface boiling and consequent vapor transport through the tank.

The development of a mechanically agitated tank accepts the problems
involved in obtaining the necessary reactor-grade mechanical components
of motor, seals, drive shaft, bearings, and agitators. These are related to
the circulating-pump problems which have been discussed previously. A
conceptual design of a mechanically agitated tank proposed by Westing-
house [31] is shown in Fig. 8-12(a).

Agitation by addition of steam and agitation by addition of heat are
essentlially similar, since both rely upon the turbulence created by vapor
transport to keep solids suspended.

Experimental investigations [32,33] of vapor transport through gas-
liquid mixtures have shown that the ratio of vapor to liquid volume may
be related to the vapor transport rate in tanks larger than 6 in. in diameter

< >
‘f‘s

where V, = superficial vapor velocity in ft/sec, fy = volume fraction of
vapor, and f; = volume fraction of slurry. This equation was found appli-
cable for slurries when V, was greater than 0.1 ft/sec and for water when
V , was greater than 0.6 ft/sec. The slurries were suspended at these vapor
velocities.

Conceptual tank designs based on agitation by wvapor transport are
shown in Fig. 8-12 (b) and (c).

*Based on material prepared by C. G. Lawson.
436 COMPONENT DEVELOPMENT [cHAP. 8

Vapor Ovutlet —e— l . \‘

 
 
 
 
 
  

Fine ‘'Yorkmesh'' Demister Element
(0.0045-in. Wire Dia)

Coarse ‘‘Yorkmesh'' Demister Element
(0.011-in. Wire Dia)

 

12-in. Pipe ——— | [E=E2

Corrugated Plate Entrainment
Separator

 

 

 

 

 

 

Vapor
Inlet
Drain To Storage Tank

Fic. 8-13. HRE-2 entrainment separator.

8-3.2 Entrainment separator. In conjunction with the storage tank-
evaporators, an efficient entrainment separator is required to keep the
purity of the condensate at the highest possible level. The entrainment
separator must be designed to function well during normal operation and
also during the emergency dumping operation of the reactor. From the
results of a literature survey and from experimental work, the separator
in Fig. 8-13 was designed for the HRE-2 [34].

The HRE-2 design incorporates three modes of entrainment removal:
centrifugal separation, corrugated plates, and wire-mesh demister elements.
The centrifugal-flow inlet and corrugated plates precede the wire mesh and
remove the main portion of the liquid and the larger particles of entrained
moisture in the letdown from the gas separator. The wire-mesh demister
has a high efficiency for entrainment removal with a very low pressure
drop, and is used as the final separator stage. The wire mesh also serves
to keep the uranium reaching the recombiner below the maximum desirable
limit of 1 ppm.

8-3.3 Recombiners. For safety as well as economic reasons, it is desir-
able to recombine, either at high pressure or at low pressure, the D2 and
O2 which are formed by the decomposition of water in the fuel solution.
In the HRE-2, only low-pressure recombination has been used effectively
- 8-3] SUPPORTING-SYSTEM COMPONENTS 437

   
   
 
 
   
  
   

Thermocouple Wells
Stainless Steel Wire Screen

Catalyst Pellet Basket
12-in. Pipe

Ya-in. Dia Tubes

 

 

 

Cooling Water
Connections

 

Vapor In

 

 

Condensate Out

-t 7 ft 10 in  ———

 

Fiag. 814. HRE-2 recombiner and condenser. The recombiner bed is steam
heated to keep the surface dry at all times.

for external recombination. The two most promising methods of external
recombination are the use of flame and catalytic recombiners.

For catalytic combination of D2 and Oz, platinum black has proved to
be the most satisfactory catalyst. It has been supported on alumina pellets
and on stainless-steel wire mesh. The platinum adheres better to the
alumina pellets, but the wire mesh is less liable to mechanical damage.

Design of the catalytic recombiners. The space-velocity method is a very
convenient basis for recombiner design. Space velocity is defined as the
cubic feet of gas-vapor mixture fed (STP) per cubic foot of catalyst bed
per hour. The maximum allowable space velocity for 100% recombination
is approximately 4.2 X 105 hr~! at atmospheric pressure. However, the
packed bed should be shaped to ensure against channeling and to get a low
pressure drop. The HRE-2 low-pressure recombiner [35] was designed
by the space-velocity method with a safety factor of ten. To ensure against
channeling and to get as low a pressure drop as possible, an annular cylin-
drical bed was designed with a 4-in. inside diameter and 9%-in. outside
diameter (Fig. 8-14).

The controlling mechanisms for catalytic reaction rates are outhned by
Hougen and Watson [36]. One of the important steps is the mass transfer
of the reactant gases to the catalytic surface. Most of the homogeneous-
reactor recombination work at ORNL has been done in the range con-
trolled by mass transfer, at temperatures of 250 to 500°C.

However, experiments conducted at 50 psi indicated [37] mass-transfer
coeflicients lower by a factor of three than the expected values based on
established mass-transfer correlations. This is explained on the basis of
poor bed configuration, channeling, and entrance and exit effects. Tests
run at 500 and 1000 psi have shown values about 60% of the theoretical
[38]. Standard mass-transfer calculations, with a suitable safety factor,
438 COMPONENT DEVELOPMENT [cHAP. 8

Coolant Air
Off-Gas

 
  
  
  
 
  
 

Spark Plugs

  
  
 

—D<>-Off-Gas

Perforated

 

Condenser Coils

Thermowells

_— Platinized

Spray Nozzle Wire Mesh

 

 

Liquid Level :

Control Liquid Level

09, Hy And Control
Steam Mixture

Drain Drain
(a) (b)

Fia. 815. (a) Experimental flame recombiner designed to operate over wide
ranges of gas input. (b) Experimental natural-circulation recombinier used to
recombine HRE-2 off-gas after shutdown.

are believed to be the most accurate method of designing catalytic re-
combiners.

Flame recombiners. In a flame recombiner, the Hz (or D2) and O are
actually ignited and burned to form water. The HRE-1 recombiner was
of this type. It consisted of a combustion chamber of 10-in. pipe 3% ft long
jacketed by a 12-in. pipe through which cooling water was circulated.
The combustible mixture of Hs and O2 was introduced through a many-
holed nozzle upon which a spark impinged from two spark plugs located
along the periphery of the nozzle. The spark impulse was produced by a
magnet, with an ignition transformer on standby. The cooling water re-
moved 70% of the heat of combustion at the design capacity of 15 scfm
of 2Hs 4+ O2. The remainder of the heat was removed in the after-con-
denser. The condensed products of combustion were returned by gravity
to the dump tanks, and the excess Oz and fission-product gases were passed
into the off-gas system to the cold traps and charcoal adsorbers.

At low flow rates the flame burned too close to the nozzle, resulting in
overheating and flashbacks. To prevent flashbacks at low flows, a steamer
pot ahead of the flame recombiner added 2 to 3 sefm of steam to the gas
stream.

In developing flame recombiners to obtain an explosion-proof automatic
load-adjusting unit, the multiple spark-plug model shown schematically
in Fig. 8-15(a) was devised [39]. As the steam-gas mixture traveled up past
the condenser coils, the mixture eventually lost enough steam to become
8-3] SUPPORTING-SYSTEM COMPONENTS 439

combustible. The location at which the mixture became combustible de-
pended on the input concentration of gas.

N atural-circulation recombiner. For some applications it may be desirable
to have a catalytic recombiner which will operate satisfactorily without a
pump or evaporator to circulate diluent to keep the gases below the ex-
plosive limit. The natural-circulation recombiner [40] was developed for
such uses (Fig. 8-15b). Electric heaters or steam coils installed below the
catalyst start the circulation of the diluent and keep the catalyst dry. A
cooling coil located in the annular space around the top of the chimney
completes the convective driving circuit.

High-pressure recombination. The use of high-pressure recombination in
homogeneous reactors would eliminate the need for continuous letdown of
the radiolytic gases and continuous feed-pump operation. To investigate
the possibilities of high-pressure recombination, tests [41] were made with
a loop built at ORNL.

Recombination rates were quite satisfactory. However, stress-corrosion
cracking was a significant problem in operating the stainless-steel loop.
Originally, the chloride content of the loop was high (50 ppm) and was
thought to be the cause of the stress corrosion. However, after the chloride
concentration was lowered to less than 1 ppm, stress corrosion still oc-
curred in the superheated region of the loop. It was established that en-
trained caustic was a contributing factor.

The cracking problem was solved by substitution of Inconel for austenitic
steel; this material would be suitable in a slurry reactor system but not in
a uranyl-sulfate system. One of the ferritic stainless steels might be suitable
for the latter application.

8-3.4 Condenser. A condenser is required in aqueous low-pressure sys-
tems (1) to condense the steam produced in the storage tank-evaporator
which is a source of distilled water, (2) to remove the heat of recombina-
tion, and (3) to cool the reactor contents during and after an emergency
drain. The surface-area requirements are usually determined on the basis
of item (3).

All-stainless-steel shell-and-tube condensers of conventional design
have been used in this application. The quality of construction from the
standpoint of leaktightness should approach that of the main steam gen-
erators. However, since the service conditions in the condenser are rela-
tively mild, its life should be indefinite if it passes acceptance tests. The
condenser used in the HRE-2 is illustrated in Fig. 8-14.

8-3.5 Cold traps.* Cold traps are usually required on fission-product
off-gas lines from homogeneous reactors to conserve D20 and to dry gases

*Based on material from R. C. Robertson, ORNL.
440 COMPONENT DEVELOPMENT [cHAP. 8

prior to adsorption in charcoal beds. Exit gas temperatures should be be-
tween —10 and —30°F. Typical evaporating refrigerant temperature in
the associated primary refrigeration system should be about —50 to —100°F'.

The cold traps may be refrigerated either by a direct-expansion system
or by circulation of a chilled secondary fluid. The secondary type system
offers advantages of the elimination of the expansion valve from the shielded
area, and simpler defrosting procedures when using a warm supply of the
secondary refrigerant. Use of a primary system eliminates the heat ex-
changer and circulating pump, and the sacrifice in about 10°F temperature
difference needed in the heat exchanger, thus affording slightly better co-
efficients of performance for the refrigeration system.

The HRE-2 cold traps are double-pipe stainless-steel heat exchangers.
Flow of refrigerant is countercurrent, with the traps pitched to drain the
D>0 when defrosting. The insulation is in the form of sealed cans of
Santo-Cel (SiO2) fitted around the traps, this material having markedly
better resistance to radiation damage than the more commonly used low-
temperature insulating materials. Cold traps are used in pairs so that
icing and defrosting can be conducted simultaneously.

The major heat load on the HRE-2 cold traps was estimated to be the
internal heat generation due to radioactivity in the off-gases. For the
double-pipe design selected, increasing the heat-transfer surface also in-
creases the mass of metal and the heat generation, so that the size must be
optimized. Over-all heat-transfer coefficients in the HRE-2 cold traps,
using Amsco as the secondary refrigerant, were calculated to be in the range
of 30 to 35 Btu/(hr)(ft2)(°F). Velocities of the gas stream were kept quite
low; less than 5 fpm. Design velocities of the chilled coolant through the
annulus were from 1 to 2 fps.

8-3.6 Charcoal adsorbers. The oxygen off-gas from a homogeneous re-
actor contains the krypton and xenon fission products which are let down
with the radiolytic gas. It is desired to discharge the oxygen to atmosphere,
but the permissible rare-gas discharge is limited by health physics con-
siderations. Charcoal adsorbers are used to hold up krypton and xenon
sufficiently to permit their decay to stable or slightly radioactive daughters.

The HRE-2 charcoal beds were designed [42] on the basis of adsorption
equilibrium data of krypton and xenon from the literature, with a safety
factor of six to compensate for lack of experimental data on the particular
conditions. An HRE-2 bed to process 250 cc/min of off-gas oxygen con-
tains 13.3 ft3 of 8- to 14-mesh activated cocoanut charcoal. There are four
such beds immersed in a water-cooled underground concrete tank. In the
HRE-1, 13.9 ft3 of charcoal were used for a design flow rate of 470 cc /min,
The HRE-1 beds operated successfully.

A more complete treatment of charcoal-adsorber design is given by
Anderson [43].
8-3] SUPPORTING-SYSTEM COMPONENTS 441

— Pshasing
t
ystem To High Pressure

System

Intermediate
Water System

Y
)

Double Wall Shield
Rubber Pulsator

 

 

 

Screen Filters

 

Pulsating _
Oil Supply ——s-=J Stainless Steel

Diaphram

 

 

 

 

 

 

/777X 77V

 

 

<——From Dump
Tanks

F1g. 8-16. Sealed diaphragm feed pump with driving pulsator used to pump
fluids from low pressure to the high-pressure system in HRE-2.

8-3.7 Feed pumps.* High-pressure low-capacity pumps are required to
feed solutions, slurries, and water into aqueous homogeneous reactor
systems. In the HRE-2, which operates at 2000 psi, the requirements are
for from O to 1.5 gpm of fuel solution, 0.25 gpm of purge water to the
pressurizer, and 0.1 gpm of purge water to the circulating pump.

The types of pumps which could possibly be made to meet these require-
ments are the piston or plunger pump, the multistage centrifugal pump,
the turbine regenerative pump, and the hydraulically actuated diaphragm
pump. The piston and plunger pumps are handicapped because most
packing materials are subject to radiation damage and because there is no
absolutely leakproof sliding seal. There is no known commercial centrifugal
pump available in this high-head low-capacity range, and the develop-
ment of such a pump appears difficult. The regenerative turbine pump has
a more suitable head-capacity range, but again there is no existing multi-
stage pump in the range desired. The hydraulically driven diaphragm
pump, as shown in Fig. 8-16, was selected for the HRE-2 because it offers
the following advantages:

(1) The pump head and check valves are of all-welded construction and
are leakproof and maintenance-free for long periods of time.

(2) The only moving parts inside the shield are the diaphragm and the
check valves.

(3) The drive mechanism is outside the shield, where conventional lu-
bricants and maintenance techniques may be used.

(4) The pump output is adjustable by changing the output of the drive
unit.

*Based on material prepared by E. C. Hise.
442 COMPONENT DEVELOPMENT [cHAP. 8

(5) In the event of a diaphragm rupture, radioactive fluid is still retained
within the piping system.

Subsequent development work has demonstrated that diaphragm pumps
will operate satisfactorily for one year or more in this service [44-45].

Construction. The HRE-2 duplex feed pump consists of three main
components: the drive unit, the pulsator assembly, and the diaphragm
heads. The drive unit and pulsator assembly are commercial products
built by Scott & Williams, Inc. The drive unit consists of a high-pressure
positive-displacement oil pump and a slide valve which alternately supplies
oil to one pulsator and then the other at 78 strokes/min. While one pulsator
is being supplied with oil from the pump, the other is being vented back to
the oil reservoir tank. During this venting period the elasticity of the
rubber pulsator forces the oil back to the tank and provides the energy for
the suction stroke of the diaphragm head. Reciprocating oil pumps are
used to drive smaller purge pumps on the HRE-2. The oil pulses are
transferred to the diaphragm head by the column of D20 filling the inter-
mediate system (Fig. 8-16).

The diaphragm head consists of a stainless-steel diaphragm 0.031 1n.
thick operating between two heavy flanges which have carefully machined
contoured surfaces 102-in. in diameter and 0.10 in. deep forming the
diaphragm cavity. The pumped fluid enters the pump through a 3/4-in.
pipe, passes up through the screen tube, and oscillates in and out of the
diaphragm cavity through rows of holes in the contoured surface of the
flange. The driving and pumping flanges are identical except that the
driving flange has only the top pipe connection, since the actuating column
of D20 needs only to oscillate. The screen tube is self-cleaning, since the
flow through it is oscillating. The two flanges are clamped rigidly together
by means of heavy girth welds, which become highly stressed because of
shrinkage during fabrication.

The pump is equipped with all-welded, double-ball, gravity-operated
suction and discharge check valves. The 1-in. balls operate in close-fitting
cages (0.010-in. diametral clearance) which maintain the alignment of the
ball and seat and restrict the ball lift to 0.125 1n.

Operation. The pump output can be varied from 0 to 2 gpm at 2000 psi
discharge pressure by changing either speed or displacement of the drive.
The pump performance is essentially independent of suction head and
temperature so long as cavitation does not occur.

To obtain proper operation of the pump, the amount of water in the
intermediate system between the rubber pulsator and the diaphragm must
be adjusted to ensure that the diaphragm does not bottom solidly against
either contoured face of the head. This procedure, called ‘“‘phasing,” 1s
accomplished manually by adding or venting water as required through
the phasing system shown in Fig. 8-16. For a specific pressure, there 1s a
8-3] SUPPORTING-SYSTEM COMPONENTS 443

fairly wide range of phasing in which the pump will operate properly, since
only one-third of the displacement volume in the head is used. At 2000 psi,
however, the compressibility of the drive and intermediate systems amounts
to another third of the displacement volume of the head. This results in
only a relatively narrow range of phasing in which the pump will operate
properly under all conditions of pressure and capacity.

The volume between the check valves is large compared with the volume
of the stroke, so that the pump is subject to gas-binding. An operational
error or a leaking discharge check valve that permits oxygenated solution
to flow back into the pump may inject sufficient gas so that the head will
not resume pumping against a high discharge pressure. A method of venting
the gas must therefore be provided.

Diaphragm development. The first heads used had a cavity 0.125 in.
deep, a 0.019-in.-thick annealed stainless type-347 diaphragm, and had
no screening. These diaphragms suffered early failure due to irregular
contour machining and dents caused by the trapping of dirt particles
between the contour face and the diaphragm. To reduce the over-all
diaphragm stress level and to reduce or eliminate the localized stress risers,
the contour depth was reduced to 0.010 in., the machining procedure was
changed to produce a smooth, continuous contour, and 40-mesh screens
were installed. These changes increased the average diaphragm life to
about four and a half months. However, some failures occurred in as little
as two months. An intensive program was initiated to develop a head
that would function consistently for one year or more.

The first objective of the program was to reduce or eliminate stress risers
caused by dirt particles. Substitution of 100-mesh screen tubes for the
40-mesh screens reduced denting observed on test diaphragms by an order
of magnitude. A sintered stainless-steel porous tube with 20-micron
openings is being evaluated at the present time in an experimental pump
to reduce the problem further.

A second objective was to investigate possible improvements in contour
in order to minimize diaphragm stress for the desired volumetric displace-
ment. Theoretical and experimental stress analyses showed that the
original contour was nearly optimum, and it was retained [46].

A third objective was to determine the nature of the diaphragm motion
and improve it if necessary. A special spring-loaded magnetic instrument
‘was built to indicate diaphragm position while operating. Three such
indicators were installed in a standard head and recorded simultaneously
on a fast multichannel instrument. It was observed that the diaphragm
was displaced in a wave motion starting at the top of the head, producing a
sharp bend at the bottom where most failures occurred. It was observed
also that there was considerable flutter in the diaphragm, so that it was
being flexed more frequently than anticipated. By increasing the thickness
444 COMPONENT DEVELOPMENT [cHAP. 8

of the diaphragm from 0.019 in. to 0.031 in., symmetrical deflections with
less flutter were obtained. Changes in the drive system which reduced the
noise level were effective in creating smoother diaphragm deflection.
These changes were incorporated into later pumps.

The fourth point of the program involved determining the endurance
limit of annealed 347 stainless steel and other possible diaphragm materials
in fuel solution. A literature review indicated that in a corrosive environ-
ment there may be no endurance limit as such, but that the curve of stress
versus number of cycles would continue its downward slope indefinitely.
The literature also suggested that significant gains in endurance limit may
be achieved by cold-working stainless steel, or by using a precipitation
hardening steel such as Allegheny-Ludlum AM-350. Standard reverse
bending sheet specimens of each material were operated at 2000 cycles/
min in environments of air, distilled water, and fuel solution [47]. Sur-
prisingly, it was found that hardened materials suffered a drastic reduction
in endurance limit in fuel solution but not in water, whereas the annealed
347 stainless-steel endurance at 107 cycles was 39,000, 36,000, and 34,000
psi, respectively, in air, water, and fuel. None of these media produce
appreciable corrosive attack on any of the materials tested.

Check-valve materials. Stellite balls and seats have been operated in fuel
solutions for more than 10,000 hr with no sign of damage. It was rather
surprising when, during preoperational testing of the HRE-2, four sets
of valves failed in oxygenated distilled water in about 500 hr. Further
testing showed that preconditioning by operation in uranyl sulfate made
Stellite suitable for oxygenated-water use. Armco 17-4 PH stainless steel
was also demonstrated to be an excellent seat material in both water and
uranyl sulfate.

HRE-2 fuel pumps now contain Stellite Star J balls and Stellite No. 3
seats. Check valves are pre-run in fuel solution before being welded to the
pump heads. HRE-2 water pumps contain Stellite Star J balls and 17-4
PH seats.

All the standard metals have failed very quickly in check valves pumping
ThO, slurry to high pressure. However, some success has been achieved
with aluminum oxide and other very hard ceramics.

Welding. Considerable difficulty has been experienced in the design of
welds subject to cyclic pressure stresses. Extreme conservatism with regard
to metal thickness is helpful in eliminating fatigue failure of welds. Nozzles
welded to pump heads have heavy sections at the weld. TFull-penetration
welds are used throughout, and butt welds are used if possible. The inside
surfaces of welds are machined smooth when they are accessible.

Slurry diaphragm pumps. Two methods of pumping slurry with the dia-
phragm pump are being tested. In the first, the check valves are located
several feet below the head and connected thereto by a vertical pipe. By
8-3] SUPPORTING-SYSTEM COMPONENTS 445

sizing the vertical leg so as to maintain low oscillatory velocities, a stable
slurry-water interface forms, permitting the diaphragm head to operate in
relatively pure water while slurry pumps through the check valves. A
venting system is necessary. Such a pump has been operated satisfactorily
at low pressure and will be tested at high pressure.

The second method uses a diaphragm head having a contour in the
driving flange, a recessed cavity in the pumping flange, and an arrange-
ment that permits the diaphragm to operate only from the driving contour
to center. This arrangement precludes the possibility of slurry being trapped
between the diaphragm and contour, leading to undesirable diaphragm
deflection patterns. Such a pump head has been built and will be tested

8-3.8 Valves.* Valves are key components in reactor systems, since they
are the means by which process gas and liquid streams are controlled
[48,49]. In the HRE-2 system, which has no control rods, temperature
and reactivity are controlled by valves that control the concentration of the
fuel solution, and the power is controlled by valves that control the rate
of steam removal from the heat exchangers. “Dump’’ valves perform an
emergency scram and normal drain function by controlling the flow of
fuel solution to low-pressure storage tanks. Other valves perform pressure-
control functions, allow noncondensable gases to be bled from the system,
or are used to isolate equipment.

Actuators. The problem of radiation damage to hydraulic fluids, elas-
tomers, or electrical insulations is avoided by utilizing pneumatically
powered metallic bellows for remote actuation of the valves. The actuator
1s a simple linear device which can be controlled with standard pneumatic
controllers or regulators. The bellows may also be stacked to multiply the
forces available. In the HRE-2, pneumatic actuators develop up to 5440 Ib
force.

An actuator capable of developing a thrust of about 12,000 Ib was cycled
four times per minute at a stroke of 1/2 in. and a pressure of 80 psig for
265,000 cycles before developing a small leak in the stem sealing bellows.
Two and three bellows-spring assemblies from these units have been at-
tached to a common shaft and connected in parallel to a source of air pres-
sure in preliminary tests of an even more powerful actuator.

Handwheel operators, with or without extension handles, have been
used successfully in all-welded valves for mildly radioactive service.

Valve designs used in HRE—-2. The valve designs used on the HRE-2 are
all quite similar. Figure 8-17(a) illustrates the ‘“‘letdown’’ valve, which is
typical. This valve throttles a mixture of cooled gas and liquid from the
2000-psi high-pressure system to the low-pressure storage tanks. The
flow is introduced under the seat to keep the bellows on the low-pressure

*Based on material furnished by D. S. Toomb.
 

   

 

  
   

 

 

 

 

 

  
  
 
  
  

 
 
  

 

 

  
  
  
    

  
 

  

   
    

 

446 COMPONENT DEVELOPMENT [cHAP. 8
SN NN
3.!"//15223,5%...‘ Air Exhaust
A=)
o \\E.\Q: !\3\:\\\;}\!‘%\&5 B Port
\ e o N Fy
P e
Loading © /g \ \ ?}
Springs ] ) § ? . Secondary
T N £ ‘
5 S § / s Packing Glond
Actuating g & A / ) Seal
Bellows D B AN D }fi
PSS S {
/A8 %.\{///45%5‘ Actuating ] Bell
NI Air Port ellows
i : :1 Backseating—} Leak-?etechng
I Steam Seal Stem ap
\"E'\‘ Bellows
:EE Bellows N
BN Sealing Weld
) ,
= & “~Primary Bellows Seal
§‘“ — E§ Secondary / imary v
A=t W=l : .
Backseating 1] Ny -//%-2 P§Cklgga(l3land E
Bellows Sealing—ssifm> ellows
THE Y .
Weld % % Leak-IT?aetechng
p
o e
41 1L rimary
101234 g: é Bellows Seal =
g\ 1B
Inches 2N g ; 777777 7<— Integral
; é 7 // ///2 Flanged Body
7 4 v yy
| il A V347 S Body |
Discharge -“‘--Z/A With Integral Seat . |

Stellite No. 6 Plug~”" §.{
2000 psi Inlet Pressure

(a)

 

 

(b)

F1c. 8-17. (a) HRE-2 letdown valve. (b) HRE-2 low-pressure valve.

side of the throttling orifice and thus under less strain. A seat integral
with the valve body is used to avoid the difficult problem of leakage around
removable seats. Stellite No. 6 and type 17-4 PH stainless steel plugs have
been used, since these very hard materials are corrosion resistant in uranyl-
sulfate solutions below 100°C and resist erosion due to flow impingement.
The primary bellows seal, 13-in. OD by 7/8-in. ID and 3% in. long, is mechani-
cally formed of three plies of 0.0085-in. type—347 stainless steel stock. The
bellows seal assembly is in two sections, welded together, because the
bellows length needed for the 1/2-in. stroke cannot be manufactured in a
single section at this time. An average bellows life of 50,000 5/8-in. strokes
has been obtained at 500-psi with this assembly. The stem is of hexagonal
stock and fits in a similarly shaped guide to prevent a torque from being
applied to the bellows. The leak-detecting tap between the bellows and the
secondary graphited-asbestos packing seal affords a means of detecting a
bellows leak, while the asbestos gland prevents gross leakage of process

fluid in case of bellows failure.
8-3] SUPPORTING-SYSTEM COMPONENTS 447

The valve, which was supplied by the Fulton Sylphon Division of Robert-
shaw-Fulton Controls Corporation, is rated for 2500-psi service with the
flow introduced under the seat; however, the downstream pressure is limited
to 500 psi by the bellows seal. The valve has a C, (flow coefficient) of 0.1.
The reversible-action operator, supplied by The Annin Company, has a
50-in? effective area. It is rated for a maximum of 60 psi air operating
pressure. The action illustrated is spring-closed, air-to-open; however,
by a simple interchange of parts, the actuator operation can be reversed.
The actuating bellows is made from type—321 stainless steel and was formed
by the Stainless Steel Products Company. The stem guide bushing is
brass.

The largest valve used in the HRE-2 is the blanket drain valve, which has
a 1-in. port and a C, of 10. The valve and operator were supplied by Fulton-
Sylphon. The operator supplies a maximum force of 5440 lb, and the full
stroke is 3/4 in. B

The only two process valves in the HRE-2 which operate with full system
pressure on the bellows seal are those which are used to isolate the reactor
from the chemical plant. The bellows used here, supplied by Fulton-
Sylphon, are rated at 2000 psi and 300°C. |

The low-pressure HRE-2 valves are novel in that ring-joint grooves are
integral with the valve body, as illustrated in Fig. 8-17(b). Long bolts at
the corners of the valve body hold the companion flanges; the valve is
replaceable with the disassembly of only one set of bolts.

The main problems encountered in HRE-2 valves have been valve stem
misalignment and corrosion of valve plugs.

Valve trim materials. In uranyl sulfate service, stainless steel seats are
used with type 17-4 PH stainless steel or Stellite plugs. The latter material
is useful only below 100°C and where only a small amount of oxygenated-
water service is anticipated with a high pressure differential across the
valve.

In slurry service, metallic trims such as those above have been satis-
factory for low-pressure valves but unsatisfactory for long life in high-
pressure service. Ceramic materials appear promising, but little experience
has been obtained to date.

A gold-gasketed valve has been developed for tight shutoff of gases.
The gasket is placed into a groove machined in the valve plug, which mates
with a tongue machined into the seat. This type of trim has also given ex-
cellent results in one hot uranyl-sulfate loop application.

Slurry service valves. In addition to the erosiveness of slurries, other
problems are introduced by their tendency to settle out in the primary
bellows seal or at stem guiding surfaces, thus interfering with valve action.
This may be avoided by purging slurry from the bellows compartment with
distilled water. It is likely that the hydrodynamic design of slurry valves
448 COMPONENT DEVELOPMENT [cHAP. 8

Asbestos Tape
SS Cover \
N Tantalum Plug

|
| = =SS Shell N

/ /L 5355 yr 7 7 7777
+

  
   

N
Heaters P§

=

[

/A

II
Ao
L .'
-
i =7//

  

 
     
 
  
 
 
  
   
  

L7777 7SS ESS SOt S

S

    

 

 

 

 

; (I ITI IO, N7 >R 77 777
@ //Q(///lflnll(«««q "" TR, 2227777777 -.
e/ —
— 4I.E-hl£1 2 Th | muufl‘.‘lb ‘_
T IN ermocouples S-
NSzrrrrrrrrrrrrrrrrrrrrrrIrr Iz N —
Conax Gland

Fic. 8-18. Differential thermal expansion valve to control gas flow in the HRE-2.

may be revised to make entry of solids into the bellows compartment
improbable.

Slurry throttling has been accomplished by use of long tubes or capil-
laries. These have the disadvantage of fixed orifices, in that continuous
flow control is not possible.

Special gas-metering valve. An ORNL-developed* differential thermal-
expansion metering valve is used to regulate the flow of oxygen gas to the
HRE-2 high-pressure system [50]. The required flow 1s very small and is
difficult to control by conventional mechanical positioning methods. The
valve shown in Fig. 8-18 utilizes the difference in thermal coefficient of
expansion of tantalum and stainless steel to effect flow control. The tan-
talum plug is used to avoid any possibility of an ignition reaction between
the oxygen gas and the metal, the temperature of which for a flow of 2000
std. cc/min with a 400 psi differential can reach 300°C. The design incor-
porates all-welded construction and is covered with a waterproof protective
housing. The resistance heating element and thermocouple are duplicated
to ensure continuity of service. |

8-3.0 Sampling equipment. Operation of an aqueous homogeneous re-
actor requires that numerous samples be taken in maintaining control of
the chemical composition of the solutions. Because of the radioactivity
associated with these fluids, standard sampling equipment must be modi-
ied, or entirely new apparatus must be devised for taking the samples.
Examples of sampling equipment presented here were designed for use on
the HRE-2 at ORNL.

Samples of liquid and suspended solids will be taken from the high- and
low-pressure systems of the HRE-2. Solution from the high-pressure
system is reduced in temperature and pressure from 300°C and 2000 psi to
approximately 80°C and 1 atm by a cooler and throttling valve before

*[J.S. Patent 2,610,300 (1952). [Assigned to the U.S. Atomic Energy Commis-
sion by W. W. Walton and R. C. Brewer.]
8-3]

(With Carrier Removed)

— | 04 8 12
Ll I Inches
N 7

1
1 I

Carrier

Isolation N
Chamber\y
Top of Flask
Holder
Transfer \
Mechanism \
Air
Cylinder \§

   
 
 
 

 

  
  
 
    

   
  
   
  
      
  
  

NN
\: b

SUPPORTING-SYSTEM COMPONENTS

Loadingfl\ \‘

Valve Stems

   

T

 

Positioning Handle

.

Shield Plug

 
    
    
  

o ]
g L

Sample Valve

Tubing Disconnect

\
L]

   

     

To Manifolil!_“t

——

| .Differential Pressure Cell

JE

 

——— .

      

-]

(a)

Positioning Handle

 
  
      
   
  
 
     

   
 

  

  
  
     

449
Universal Joint ¥
RN
:\ Valve Stem
Packing Gland ey \Q:
--------- \\\\!: A'\; .‘ '1',"::"
Vent § ?—*fi's\ \;{, Bellows
VIVA
- SA L)
Overflow —— il Resee] #»;_,_,: =
Stop 4’ ’ [‘f‘ Sample
Isolation Chamber E I Cooling Coil
Ml |
'l E N
s |
1\771' Flask Holder
N O
R} &\"\R‘Qflf; Plunger
IR Needle Holder
, ; Needle
y
8y
Rubber
"""" | Diaphragm

Receptacle

 

Fic. 8-19. (a) HRE-2 sampling facility. Flask holder has just been lowered
through the loading tube. It is then moved under the isolation chamber by the

sample to the receptacle.

transfer mechanism. (b) HRE-2 sampler head, shown in the position of transferring

entering the sample station. There, a sample of 4 to 5 ml is isolated and
removed for analysis [51].

Figure 8-19(a) shows the general assembly of the sampling facility.
Virtually all the mechanism is suspended from a shield plug. Personnel
shielding is provided by a 2-ft depth of lead shot and water in the plug.
The loading tube is sealed by a plug valve to maintain a slight vacuum in
the housing. Threaded backup rods extending through the plug are em-
450 COMPONENT DEVELOPMENT [cHAP. 8

ployed to make the final connections with reactor piping after the plug
assembly is lowered into its housing. Each sampling facility contains two
isolation chambers: one for isolating samples from the high-pressure system
and the other for obtaining samples from the low-pressure system. Each
chamber in the station is served by a common loading and manipulating
device.

When a sample is being taken, solution from the desired system is al-
lowed to flow through its isolation chamber until a representative sample
is obtained. The isolation chamber is then valved off. An evacuated
sample flask is placed in the holder and lowered through the loading tube
to the transfer mechanism. The assembly is then indexed under the proper
isolation chamber, where the flask holder is raised by an air cylinder until
contact is made between the isolation-chamber nozzle and the inverse cone
of the carrier head (Fig. 8-19b). Further lifting of the flask holder causes
the hypodermic needle to puncture the rubber diaphragm. The sample 1s
then discharged into the flask by opening the valve on the chamber. When
the sample is in the flask the procedures are reversed, and the flask holder
is removed into a shielded carrier for transport to the analytical laboratory.
Electrical contacts indicate positive positioning of the flask holder under
the isolation chamber and closure of the isolation-chamber nozzle.

A third sampling station for the HRE-2, identical to the fuel and blanket
facilities except for larger passages and a modified isolation chamber, is
employed for sampling a fuel stream in the chemical processing facility.
This stream has the order of 50 times the solids concentration of the other
streams being sampled.

8-3.10 Letdown heat exchanger. The purpose of the letdown heat ex-
changer is to conserve the sensible and latent heat of the solution-steam-
gas mixture removed in the gas separator prior to discharging it to the
dump-tank system. It is necessary also to cool the letdown stream to below
100°C before it reaches the letdown valve to minimize corrosion of the
valve trim.

The thermal design of the exchanger is conventional [52]. In the HRE-2
stainless-steel triple-pipe unit, 400,000 Btu/hr are removed from the
letdown stream into the countercurrent fuel feed stream, the pressurizer
purge-water stream, and a cooling-water stream.

The unique feature of the design deals. with the flow geometry of the
letdown stream [53]. To promote efficient flow of the two-phase mixture
through the letdown valve, it is necessary to prevent flow separation of
the two phases. This is done by utilizing the annulus of the exchanger,
with weld-bead spacers every 3 in. to promote turbulence. The velocity
of the letdown stream is not permitted to fall below 5 ft/sec for any pipe
lengths above 1 ft anywhere between the gas separator takeoff and the
letdown valve.
8-3] SUPPORTING-SYSTEM COMPONENTS 451

During the transit from reactor operating temperature to 100°C in the
letdown heat exchanger, fuel solution must go through the temperature
range 175 to 225°C at which stainless-steel corrosion resistance passes
through a minimum. This suggests that after several years leakage would
occur between the feed and letdown streams. This problem can be circum-
vented by substitution of titanium for stainless steel.

8-3.11 Freeze plugs. Several reactor installations have employed freeze
plugs on liquid-carrying process lines to assure absolute leaktight shutoff.
Lines up to 4 in. in diameter have been frozen with a simple wrap-around
coil of copper tubing when there was no flow in the pipe other than the
convective currents set up by the freezing process. It is conceivable that
leaktightness in very large lines might be achieved by refrigerating the
passages of valves to freeze a relatively small amount of liquid at the
valve seat. This freezing technique is most helpful in reducing the spread
of contamination during maintenance.

The most efficient freeze-jacket design is one which provides an annular
space around the process pipe and allows direct contact of the refrigerant
with the pipe. This is generally considered undesirable, however, from the
standpoint that if process fluid should leak into the refrigerant, activity
would be carried outside the shielded area. Freeze jackets consisting of
tubing wound around the process pipe perform noticeably better if soldered
or welded to the process pipe; filling the interstitial space with poured
lead also appears to be a worth-while refinement for lines difficult to freeze.
Tubing 5/16-in. in diameter has been used on 1/4- to 1/2-in. standard pipe
sizes; 3/8-in. tubing on 3/4- to 13-in. pipe sizes, and 1/2-in. tubing on sizes
up to 4 in. Clamp-on, or clamshell, types of freeze jackets were developed
for the HRE-2 for temporarily freezing certain lines.

On the HRE-2, stainless-steel refrigerant tubing is used for permanent
freeze jackets on lines which normally operate at or above 350°F. Copper
tubing, which is oxidized more readily in air, is used for lower temperature
lines. A jacket length of 3 to 4 pipe diameters has been demonstrated to be
optimum.

Freezing times of a few minutes for 1/2-in. and smaller lines and up to
several hours for 3- and 4-in. sizes have been observed when the refrigerant
temperature is in the —20 to —40°F range and with flows through the
jacket of 3 to 5 gpm. Insulation outside the freeze jacket materially aids
in the ability to freeze lines with particularly high heat load, such as those
subjected to gamma heating. If the freeze jacket must be operated sub-
merged, such as for underwater maintenance, it has been found that pro-
tecting the jacket from convection water currents by means of aluminum-
foil wrapping aids materially in the freezing process.
452 COMPONENT DEVELOPMENT [cHAP. 8

8—4. AuxiLiArRy COMPONENTS

8—4.1 Refrigeration system.* Refrigeration is required in the HRE-2
for operation of freeze plugs and cold traps. The refrigeration system con-
sists of a primary loop, which is not irradiated, and a secondary liquid cir-
culating system which enters the shield.

A two-stage primary mechanical refrigeration system is employed in the
HRE-2. Refrigerants commonly used in such a system are the halogenated
hydrocarbons, provided that the primary refrigerant remains outside the
reactor shield. Breakdown of this series of refrigerants under radiation has
been observed to have the serious effects of forming phosgene gas and in-
soluble tarry polymers, thus creating conditions corrosive to stainless steel.
Carbon dioxide is probably the best refrigerant for use in an irradiated
direct-expansion system, but it must be used at high pressure.

Choice of a secondary refrigerant to be circulated through radioactive
equipment is difficult in that the fluid must not only meet the obviously
desirable properties of having a low freezing point, suitable viscosity, low
vapor pressure, noncorrosiveness, nontoxicity, and nonflammability, but it
must also be resistant to radiation damage, not contain chloride ions which
might promote stress-corrosion cracking of stainless steels, and not evap-
orate to insoluble residues. Miscibility with water would be advantageous
if underwater maintenance techniques are employed in that if some refriger-
ant escapes, there is less impairment of vision and a film is not left on
equipment when the water is drained.

After considering many possible secondary refrigerants, Amsco 125-82,
an odorless mineral spirit resembling kerosene in its physical properties,
was selected for the HRE-2. Its performance to date has been quite satis-
factory.

In addition to the primary refrigeration system used to maintain a
central supply of chilled Amsco, it was useful for short-term maintenance
operations at the HRE-2 to have also a portable rig, consisting of an in-
sulated tank and circulating pump. Chilling was accomplished by floating
blocks of COs-ice directly in the liquid; secondary refrigerant tempera-
tures of about —75°F were maintained with a circulation rate of about
4 gpm and with an ice consumption rate of 75 to 100 lb/hr.

84.2 Oxygen injection equipment.} Oxygen is needed in the high-
pressure fuel system to maintain chemical stability of the uranyl-sulfate
solution and to reduce corrosion of the stainless steel container. This oxy-

*Based on material furnished by R. C. Robertson.
TMaterial submitted by E. C. Hise.
8—4] AUXILIARY COMPONENTS 453

gen may be introduced most conveniently into the fuel feed stream, at
either the suction or discharge of the feed pump. As a result of operational
experience, high-pressure injection has been found to be more flexible and
to give better feed-pump performance.

The oxygen system requires a high-pressure gas supply and a metering
device. The first supply used in the HRE-2 was a converter manufactured
by Cambridge Corp. of Lowell, Mass. This has been replaced by high-
pressure cylinders, which have considerably lower operating costs. Oxygen
compressors may be desirable to recirculate contaminated oxygen and are
being investigated. Metering is accomplished with a thermal valve (de-
scribed earlier) controlled by a capillary flowmeter.

Ozygen converter. The HRE-2 oxygen generator is designed to convert
liquid oxygen to the gaseous state and deliver it to the fuel and blanket
high-pressure systems at pressures up to 3000 psig. The capacity of the
generator is 0.47 ft3, or 30 Ib of oxygen, when 909, filled with liquid. This
will permit delivery of approximately 21 1b of oxygen gas at 3000 psig and
70°F. This pressure is automatically maintained over a flow range of from
0.01 to 0.7 1b/hr.

The oxygen generator consists of an insulated high-pressure container,
with an electric heater and automatic pressure and temperature controls.
The high-pressure inner vessel is fabricated of type—304 stainless steel.

Charging of the converter with liquid oxygen is a manual operation. The
labor of charging and the inefficient utilization of oxygen are disadvantages
of this unit.

High-pressure cylinders. The HRE-2 is now using 300-liter high-pressure
cylinders which are commercially charged to 2400 psi and are used down
to 2000 psi. A bank of three cylinders will last for about two days of normal
operation. This system involves no waste of gas, since the cylinders are
recharged from 2000 psi to 2400 psi, with very little operator attention or
hazard.

Oxygen compressors. High-pressure low-capacity laboratory-type oxygen
compressors have recently become commercially available.  Pressure
Products Industries, of Hatboro, Pa., produces a compressor having a
stainless-steel diaphragm hydraulically actuated in a contoured chamber
by a reciprocating drive. A single-stage machine capable of compressing
approximately 0.8 scfm of Oz from 500 psi to 2500 psi has been purchased
and placed in service in the HRT mockup. Although there have been some
difficulties with the hydraulic plunger packing, it has been generally satis-
factory. |

A three-stage machine capable of compressing 2 scfm of contaminated
oxygen from atmospheric pressure to 2500 psi is being designed. The
diaphragm heads will be located remotely with respect to the drive, as is
done in diaphragm feed pumps.
454 COMPONENT DEVELOPMENT [cHAP. 8

8-5. INSTRUMENT COMPONENTS*

The instrumentation and controls systems for aqueous homogeneous
reactors are similar to those used in modern high-pressure steam power and
chemical plants. However, problems attendant on radiation damage to
insulations, the difficulty of performing maintenance or replacement
operations, the requirement for the absolute leaktightness and the very
high reliability of components necessary for safety and plant operability
have required considerable development of special components.

8-5.1 Signal transmission systems. In a typical control loop the pri-
mary and final control elements are in a radioactive area isolated from the
control room by a vapor container and a concrete radiation shield.

Electric. Advantages for electric transmission under these conditions
include the ease of readjusting system zeros and spans from the control
room and the ability to sense motion from weld-sealed transmitters without
the use of flexure seals such as bellows and torque tubes. The speed of
information transmission, the ease of switching signals, and the ability of
the sensing elements to operate over wide temperature ranges may also be
important. Disadvantages of the electrical system include the possible
radiation damage to insulations and the present unavailability of a cheap,
reliable linear-power actuator for control valves.

Pneumatic. Advantages of a pneumatic system include the utilization
of all-metallic radiation-resistant construction for the transmitters and
valve actuators by the use of metallic bellows, bourdon tubes, and con-
voluted diaphragms. The advantages of the high state of commercial
development, low cost, reliability, miniaturization, and ease of paralleling
of receiving elements are considerable. A disadvantage of the penumatic
system 1s the tubing transmission line, which affords a path out of the
radiation enclosure for contaminated fluids or vapors in case of a release
of radioactivity coupled with a line break.

HRE-2 system. A combined electric-pneumatic system (described in
Article 7-4.8) is used in HRE-2. In the control room, electric signals from
primary variable sensing elements (temperature, flow, liquid level, pressure,
etc.) are converted by transducers to penumatic signals, and these are used
to actuate miniature pneumatic display instruments and pneumatic valves
in the reactor. The escape of radioactivity through air lines is prevented
by the automatic closure of “block” valves within the vapor-contained
area, on a signal of the release of radiation. Radiation damage to primary
elements is avoided by the use of inorganic electrical insulations such as
glass, ceramics, mica, magnesium oxide, and magnesium silicate. Electric
control actions are derived from the pneumatic signals by pressure switches.
These switches are simple devices in which diaphragm deflection opens or

*Material submitted by D. S. Toomb.
8-5] INSTRUMENT COMPONENTS 455

‘‘‘‘‘

     
  
   
   
      
              
    

Waterproof Coil Housing Piston Position Sensing Coil

(Differential Transformer)

Magnetic Piston With
Stainless Steel Sheath

'

|

,@

Helical
Suspension Springs

 

Permanent Magnet Yoke

 

Copper Damping Ring
Sheathed in Stainless Steel 2w

Vapor Tap \\\\ LENRN
R\ 2

  

 

 

 

 

14 in.

 
  
 

C

SR\

Fig. 8-20. HRE-2 float-type level indicator (covers a 5-in. range at psi-2000
operating pressure).

Low Level Tap

 

 

closes an electric snap-acting switch. Electric interlock control of the
penumatic signals to final control elements is achieved by the use of
solenoid-actuated pilot valves.

8-5.2 Primary variable sensing elements. Liquid-level fransmitiers.
Knowledge of liquid levels in reactor systems and loops 1s critical for main-
taining the proper balance of liquid and vapor in pressurizers and storage
tanks. It is desired also to be able to maintain accurate inventories of the
hazardous and valuable fluids which are contained.
456 COMPONENT DEVELOPMENT fcuaPp. 8

There are a large number of liquid-level sensing devices in use, since no
one device has been developed which satisfies all the criteria of precision,
rapid response, insensitivity to temperature and pressure, and utility of its
signal for control functions. Devices which have been used at ORNL are
described in the following paragraphs.

(1) Displacement or Float Transmitters. The ORNL-developed dis-
placement transmitter, used to control HRE-2 pressurizer level, consists of
a o-in.-long displacer suspended by two helical springs (Fig. 8-20). An
extension rod above the springs positions a magnetic piston in the center
of a differential transformer. Troublesome vibration of the float is damped
by the action of the field from permanent magnets on a one-turn copper
ring. The only nonwelded closure is the ring-joint flange, which makes the
unit easily replaceable.

The differential transformer is a compact, highly sensitive, linear device
which 1s commercially available. The most satisfactory instrument system
for the differential transformer is a high-frequency oscillator-amplifier
phase-sensitive demodulator carrier system which provides the necessary
sensitivity and stability and eliminates phase-motion ambiguity associated
with the null voltage. |

Float transmitters of this type have also been built with cantilever
springs, with floats up to 47 in. long, and with hydraulic damping vanes
attached to the bottom of the float in lieu of the magnetic damping [54].
They have given excellent service in continuous control applications.
However, the displacement transmitter is quite sensitive to fluid densities,
and the springs exhibit some changes in properties with temperature. The
best spring material tested to date is Isoelastic spring alloy supplied by
John Chatillon and Sons, which may be gold-plated for supplementary
corrosion resistance.

Hollow spherical floats, lighter than water, have also been used at pres-
sures below 600 psi. Magnetrol, Inc., makes a unit in which float position is
transmitted magnetically. Moore Products Company supplies a level
alarm where float position is transmitted mechanically through the all-
welded housing by a flexible shaft.

(2) Differential-Pressure Cells. D/P cells have been used successfully
in HRE-1 and in loops as level transmitters. The variable liquid leg is
compared to a reference level maintained by condensation or liquid addi-
tion. Since the density of water is temperature-dependent, the temperature
of the primary system and the lines to the D/P cell must be known for
accurate level measurement.

(3) Weigh Systems. For obtaining an accurate inventory of HRE-2
storage tanks, they are weighed with pneumatic weigh cells. This was
found to be the only feasible method of measuring the quantity of liquid
in long, horizontal storage tanks. Piping to the tanks is kept flexible by
8-5] INSTRUMENT COMPONENTS 457

the use of horizontal L. and U bends. A pneumatic system is selected pri-
marily because taring can be done remotely with balancing air pressures,
and components are less susceptible to radiation damage. The pneumatic
load cells, which are supplied by the A. H. Emery Corporation, have an
accuracy of 1/109,; however, when used in a system with solid pipe con-
nections to the weighed vessels, an accuracy of 1% of full load results.

(4) Heated Thermocouple Wells. Heated thermocouple wells have been
used for liquid-level alarm or control. The thermocouple junction is
normally held a few degrees above the vapor temperature; as the liquid level
surrounds the probe, the increased heat transmission to the fluid from the
probe lowers the thermocouple signal output [55]. Several wells must be
used for control purposes. This system gives rather sluggish response.

(5) Capacitance Probe. An aluminum oxide capacitance probe, manu-
factured by Fielden Instrument Division, has been recently received by
ORNL but has not yet been evaluated. This instrument senses the dielectric
constant of the medium it contacts. Its ceramic-to-metal seal is rated at
2000 psi and 636°F. This type of instrument may prove useful in water or
slurry service.

Differential Transformer

et . Cast Aluminum
Position Adjustment

 

 
  
    
  

   
   
    

     
  
     

   

?\ Electrical Housing
N
§\\ =% 5\‘ Core Guide Bearing
§\ \ 2’:‘ % Terminal Strip
i,
|

2RNRNDOEIRRS
SO
Y

CUCTURUU T
VA
A
LTS

TUXY
2

e

A
Srereeeddd

// 2y a
é/o//

   
    

  
 

  

OO0

Differential Transformer
Core Motion Sensing Element

LRI

    
          
 
   
  
 
 

 

 

 
 

A ISR

N R e

Z DN~
Stainless Steel ‘/é//’ ] w2 ; ‘//’/////////////%M:

N

Bourdon Tubes \\Qi’\!

.

Magnetic Core

NN

1T
A

©

  

N\

N

e

Core Drive Rod

 

 

2_
1 Scale In
1 Inches
37
4

  
   
  

 

 

@ Stainless Steel
@ @ Safety Housing

Fia. 8-21. Bourdon-tube pressure transmitter in safety housing. Sensing ele-
ment, twin Bourdon tubes; range, 0 to 2500 psi; test pressure, 3750 psi; accuracy,
+ 19, of range; transmission, electrical; fabricated by the Swartwout Company.

\
b

[——1
f e

 
458 COMPONENT DEVELOPMENT [cHAP. 8

Magnetic Piston Cased
In Stainless Steel

  
 
 
     
 
 
  

Differential Pressure Sensing Diaphragm &k_

;;;;;;;

1 01 23 4567

Scale In Inches Differential Pressure Transmitter

(Foxboro Company)

  

Autoclave Disconnect Fitting !
Metering Capillary ¢

Snubbing Capillary
N

‘\\\\\\\\\\ Rl BN 772 7) fi\\\‘“

-=_$.'V\'\~\'\‘\'\'\'\'\‘\*\°\'\'\'\'}'\'\\

 
 
 
  

  

 

Y
ey —

AN

Fia. 8-22. Capillary flowmeter, used to meter small gas or liquid streams,
shown with high-pressure seal-welded differential-pressure transmitter.

(6) Fluid Damping Transmitters. The Dynatrol transmitter, manu-
factured by Automation Products, Inc., is an interesting possibility for
use as a fluid damping transmitter. It contains a vane exposed to the
process system and vibrated through a pressure housing by alternating-
current excitation of a solenoid. The degree of damping, which is dependent
on the area of the vane covered with liquid, is measured by a second sensing
coil. No test experience with this transmitter is yet available.

Unusually difficult level-sensing problems are introduced when it is
desired to measure or control the true level of a slurry or a boiling liquid.
Most proven devices are density-sensitive, and the mean density of two-
phase systems is usually unknown. Of the level transmitter types cited
above, none appears adequate for continuous-range indication. For spot
indication, float, capacitance, and Dynatrol transmitters are promising.

Pressure and differential-pressure measurement. Bourdon tubes of weld-
sealed 347 stainless steel are used for pressure transmission in the HRE-2.
Most suitable for reactor use are units contained within secondary pressure
housings, such as the 2500-psi pressure transmitter shown in Fig. 8-21.
Baldwin cells have been widely used for accurate pressure measurement
in loops.

Bellows or diaphragm differential-pressure cells have been used to meas-
ure pressure differentials with full-scale sensitivity of 25 in. of water to 125
psi. A typical D/P cell with electric transmitter is shown in Fig. 8-22.

Pressure transmitters are usually tied into the steam or water portions
of aqueous homogeneous systems to reduce the probability of plugging or
other damage. Where it is necessary to connect a D/P cell into a slurry
system, the pipe connection is regularly purged with 10 to 30 cc/min of
water. Large vertical piping connections with the transmitter mounted
above the primary piping have also been used. Diaphragm transmitters
mounted flush with the pipe surface are being developed for slurry applica-
tions.
8-5] INSTRUMENT COMPONENTS 459

Flow transmatters. Flow measurements are made in high-pressure lines
by sensing the pressure drop across a calibrated orifice or venturi, or by the
transmitting variable-area type of flowmeter. The latter meter resembles a
Rotameter with float position transmitted electrically. It has the ad-
vantage of being an in-line element but is not readily applicable to large
flows.

Another system for metering and controlling small liquid and gas flows
in the HRE-2 is illustrated in Fig. 8-22. The pressure drop across the
metering capillary is measured by the differential-pressure transmitter and
the output signal is calibrated in terms of flow. The *“‘snubbing’ capillary
is used to prevent the sudden application of pressure to the inlet side of
the differential-pressure transmitter, which would cause undesirable zero
shift.

A technique widely used in the HRE-2 for metering purge flows is a
“heat balance’ flowmeter in which a known amount of heat is added or
extracted from the process stream and the temperature change noted.

Temperature measurement. The most commonly used method of tempera-
ture detection in the HRE-2 is the thermocouple measurement of vessel
and pipe wall temperatures; the couples are spot-welded directly to the
wall and then covered with insulation. When faster response is desired,
thermocouples are spring-loaded into thin thermowells. Chromel-Alumel
wire is generally used because its resistance to corrosive attack by moisture
is better than that of iron-constantan alloys.

Thermocouple wire insulated by compressed magnesium oxide powder
and housed in various alloy tubes is available from the Thermo Electric
Company. Another commonly used wire supplied by the Claud S. Gordon
Company is insulated as follows: each strand is coated with phenol formal-
dehyde varnish and Fiberglas-impregnated with a silicone alkyd copolymer,
and the entire wire is Fiberglas-impregnated with a silicone alkyd co-
polymer.

Sound transmitters. Waterproof microphones are attached to pumps to
monitor bearing and check-valve noises.

8-5.3 Nuclear instrumentation in the HRE-2. The purpose of the nu-
clear instrumentation in homogeneous reactors is to provide neutron-level
measurement and the gamma monitoring of auxiliary process lines and
control areas for the detection of radioactive leaks (see Article 7—4.8).

Gamma radiation measurement. Gamma monitors for detecting process
leaks, manufactured by the Victoreen Instrument Company, consist of a
simple one-tube, three-decade logarithmic amplifier sealed within the
chamber head and a remote-contact-making meter and multipoint recorder.
These detectors can be remotely calibrated by exposing a radioactive source
on the actuation of a solenoid-operated shielding shutter. All channels are
460 COMPONENT DEVELOPMENT [cHAP. 8

 

       
 

 

   

 

 
  

RG 59-U Coaxial

L Ni n s
\ \ Nitrogen Ga Fluorethene Insulator
\ .
14$\\\§ '\\5\4& » Aquadag Surface Guard Ring
AN \\\\ i\%‘ “““““““““ I, 7
\ =

   

i
¥
»

 

S
\
\
N\
\
N

  
   

 

  
  

 

    

 

;/A
7
Z
N %

Cables -\§§é
N

e

N N %

»f %

} !

3 ! %

§ N NZ 3 %

§ N——] N= N %

\\§\ gfl-\-\\@ 211 11 g :@é

\ “f %

\ N e A

§\ .................. &% ;@?

 

2

 

  
   

DI T %/////// PRI 7% X
\
Kovar Seals 1/\\\\{\\@\\\§§ Positive Electrode
DY Y DN

Brass Ground Rod
Fluorethene Insulator

Signal Electrode ™ Aluminum Housing

 
   

 

Scale In Inches

Fie. 8-23. High-level gamma ionization chamber. Effective volume, 120 cm3;
electrode spacing, 1/8 in.; performance, 20 ua, chamber saturated at 100 volts at
radiation level of 3 X 107 r/hr; design temperature, 130°F.

duplicated, and control action is initiated only upon a simultaneous signal
from both channels to minimize false “‘scrams.” However, a signal from
either channel is annunciated. For monitoring control areas for personnel
protection, more stable and accurate vibrating-condenser types of elec-
trometers are used.

The cell air monitors, which provide an alarm in case of a leak of radio-
active vapor from the reactor system, are installed in an instrument cubicle.
Cell air is circulated through a 2-in. pipe from the reactor tank, past the
enclosed monitors, and then back to the cell. The blower is sized so that
only 5 sec is required for cell air to reach the radiation monitors.

A high-level gamma ionization chamber, developed at ORNL [56], is
used to measure cell ambient radiation levels up to 107 r/hr (Fig. 8-23).
This measurement is needed to evaluate the effectiveness of shielding, to
assay the rate of radiation damage to reactor components, to measure
radiation levels during maintenance operations, and to provide data for
future reactor designs. The chamber is of inexpensive construction and is
discarded upon failure.

8-5.4 Electrical wiring and accessories. Copper-clad compressed mag-
nesium-oxide spaced and insulated electrical cable is very desirable for
service in extremely high-temperature, radioactive, or wet areas because
no organic material subject to cracking and outgassing is used in the insula-
tion. A waterproof disconnect, designed to be broken remotely to permit
the removal of reactor electrical equipment, is used with this type of cable.
The electrical connectors are terminated inside the disconnect with a
multiple-header ceramic-to-metal seal, voids being filled with magnesium-
8-5] INSTRUMENT COMPONENTS 461

oxide powder. The outside guides are tapered to simplify remote main-
tenance. Long insulators are used on the connecting terminals to minimize
leakage currents after submersion. The cable is available in a varied num-
ber of conductors and sizes, from single to seven conductors in a copper
sheath, as wire sizes No. 16 AWG to 4/0 AWG, from the General Cable
Company. The hermetic end seals are available from the Advanced Vacuum
Products Corporation or Permaseal Corporation.

A compression seal designed around an inorganic material, magnesium
silicate, is used to seal wires at conduit terminations. These seals are
supplied by the Conax Company. A similar device but utilizing a glass-
to-metal seal is manufactured by the Stupakoff Ceramic and Manufactur-
ing Company.

For the windings used on the motion-sensing coils of instruments, 30-
gage anodized aluminum wire supplied by the Sigmund Cohn Company
has successfully withstood temperatures up to 300°C and radiation ex-
posure of 6 X 10!7 nvt fast neutron and 1 X 10® r gamma without failure.
The only electrical insulation on the wire is that afforded by the oxide
film on the aluminum. This wire must be handled carefully to avoid abra-
sion and is suitable only for low-voltage use. For lower temperatures, the
Ceroc magnet wire available from the Sprague Electric Company has been
used very successfully.
462 COMPONENT DEVELOPMENT [cHAP. 8

REFERENCES

1. P. H. Harrwey, Straight Through HRT Core Model Test, USAEC Report
CF-54-9-129, Oak Ridge National Laboratory, Sept. 22, 1954.

2. 1. SP1IEWAK, Preliminary Design of Screens for the Inlet of the ISHR Core,
USAEC Report CF-52-10-181, Oak Ridge National Laboratory, Oct. 18, 1952.
W. D. Baines and E. G. PeTERsoN, Trans. Am. Soc. Mech. Engrs. 73(5), 467
(July 1951).

3. L. B. LeseM and P. H. HarLEY, Scale-up of Alternate HRT Core, USAEC
Report AECD-3971, Oak Ridge National Laboratory, May 7, 1954. L. B.
Lesem and I. Spiewak, Alternate Core Proposal for the HRT, USAEC Report
CF-54-1-80, Oak Ridge National Laboratory, Jan. 28, 1954.

4, F. N. PeeBLES and H. J. GARBER, Studies on the Swirling Motron of Water
within a Spherical Vessel, University of Tennessee, Report S-370, January 1956.

5. L. B. LesEwm et al., Hydrodynamic Studies in an Eight-foot Sphere Utilizing
Rotation Flow, USAEC Report CF-53-7-29, Oak Ridge National Laboratory,
July 20, 1953.

6. S. TiMosHENKO, Strength of Materials, 2nd ed. New York: D. Van Nostrand
Co., Inc., 1940. (Part II, pp. 160, 162)

7. S. TimosaENKO and J. N. GoobpiERr, Theory of Elasticity. 2nd ed. New York:
McGraw-Hill Book Co., Inc., 1951. (pp. 59, 359)

8. S. TimosHENKO and J. N. Goop1ER, Theory of Elasticity. 2nd ed. New York:
MecGraw-Hill Book Co., Inc., 1951. (pp. 412, 418)

9. L. G. ALEXANDER, Estimation of Heat Sources in Nuclear Reactors, A.l.
Ch. E. Journal 2: 177 (June 1956).

10. R. H. CaaprmaN, Analysts of Spherical Pressure Vessel Having an Energy
Source Within the Wall, USAEC Report ORNL-1987, Oak Ridge National Lab-
oratory, Oct. 26, 1954.

11. L. F. BLEDsOE et al., Welding J., N. Y., 35, 997-1006 (October 1956).
W. R. Garr, Nucleonics 14(10), pp. 32-33 (October 1956).

12. J. C. MovErs, Long-term Run of Westinghouse 400A-1 Pump, USAEC
Report CF-57-9-1, Oak Ridge National Laboratory, Sept. 3, 1957.

13. R. B. KorsMEYER et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958.

14. H. A. RunbpELL et al., Investigation of Effect of Seal Configuration on Mixing
Flow and Radiation Damage tn HRT-Type Circulating Pumps, USAEC Report
CF-57-10-48, Oak Ridge National Laboratory, Oct. 10, 1957.

15. J. C. MovErs, Long-term Run of Westinghouse 400A-1 Pump, USAEC
Report CF-57-9-1, Oak Ridge National Laboratory, Sept. 3, 1957.

16. R. B. KorsMEYER et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak
Ridge National Laboratory, Oct. 10, 1957. (p. 59)

17. R. B. KorsMEYER et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 1958, USAEC Report ORNL-2493, Oak
Ridge National Laboratory, 1958.
REFERENCES 463

18. H. A. RunpEeLL et al., Investigation of Effect of Seal Configuration on
Mizing Flow and Radiation Damage in HRT-Type Circulating Pumps, USAEC
Report CF-57-10-48, Oak Ridge National Laboratory, Oct. 10, 1957.

19. W. J. Finan and I. GranNeT, Final Reports on Union Carbide Nuclear
Company Contract No. W35X-31312, Phase 1, Foster-Wheeler Corp., Nov. 15
and Dec. 15, 1956.

20. J. C. Griess et al., Solutton Corrosion Group Quarterly Report for the
Period Ending July 31, 1957, USAEC Report CF-57-7-121, Oak Ridge National
Laboratory, July 31, 1957. (p. 33 ff)

21. C. H. Secoy, Aqueous Fuel Systems, USAEC Report CF-57-2-139, Oak
Ridge National Laboratory, Feb. 28, 1957.

22. C. MicueLsoN, HRT Modified Pressurizer Design, USAEC Report CF-
56-5-165, Oak Ridge National Laboratory, May 25, 1956.

23. Boiler Construction Code, Section I, Power Boilers, American Society of
Mechanical Engineers (1956); ASA Code for Pressure Piping, B31.1-1935.

24. K. L. HansoN and W. E. JausmaN, An Evaluation of Piping Analysis
Methods, USAEC Report KAPL-1384, Knolls Atomic Power Laboratory, Aug.
10, 1955.

25. M. W. KeLrLoacc Company, Design of Piping Systems. 2nd ed. New
York: John Wiley & Sons, Inc., 1956.

26. M. I. LunpiN, HRT High Pressure System Piping Line Deflections and
Reactions on Equipment Nozzles, USAEC Report CF-55-8-83, Oak Ridge Na-
tional Laboratory, Aug. 10, 1955.

27. W. R. GaLL et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Peritod Ending Apr. 30, 1957, USAEC Report ORNL-2331, Oak
Ridge National Laboratory, Aug. 14, 1957. (pp. 22-25)

28. B. DrarEr and H. C. RoLLER, Design and Development of a V4-in. Titanium
to Stainless Flange, USAEC Report CF-57-11-140, Oak Ridge National Labora-
tory, Nov. 27, 1957.

29. J. A. HarrorDp, Development of the Pipe-line Gas Separator, USAEC
Report ORNL-1602, Oak Ridge National Laboratory, Nov. 2, 1953.

30. P. H. HarLEY, Performance Tests of HRT Fuel Solution Evaporator and
Entrainment Separator, USAEC Report CF-54-10-51, Oak Ridge National
Laboratory, Oct. 13, 1954.

31. WESTINGHOUSE ELECTRIC CORPORATION AND PENNSYLVANIA POWER AND
Licar CompaNy, 1957. Unpublished.

32. E. A. FarBER, Bubble and Slug Flow in Gas-Liquid and Gas (Vapor)-
Liquid Solid Mixtures, Research Progress Report on Subcontract N.996 to REED
of Oak Ridge National Laboratory, 1957.

33. R. V. BarLeY et al., Transport of Gases Through Liquid-Gas Maixtures,
USAEC Report CF-55-12-118, Oak Ridge National Laboratory, Dec. 21, 1955.

34. C. L. SEgASER, HRT Entrainment Separator Design Study, USAECReport
CF-54-7-122, Oak Ridge National Laboratory, July 23, 1954.

35. R. E. AveN, HRT Recombiner Condenser Design, USAEC Report CF-54-
11-1, Oak Ridge National Laboratory, Nov. 1, 1954.

36. O. A. Houcen and K. M. WatsoN, Chemical Process Principles, Vol. I1L.
New York: John Wiley & Sons, Inc., 1947. (pp. 902-910)
464 COMPONENT DEVELOPMENT [cHAP. 8

37. J. A. Ransonorr and I. SPiEWAK, in Development of Hydrogen-Oxygen
Recombiners, USAEC Report ORNL-1583, Oak Ridge National Laboratory,
Oct. 22, 1953. (p. 40)

38. P. H. HarLEY, High-pressure Recombination Loop Progress Report, USAEC
Report CF-57-1-90, Oak Ridge National Laboratory, Jan. 4, 1957.

39. J. A. Ransonorr and I. SPIEWAK, in Development of Hydrogen-Oxygen
Recombiners, USAEC Report ORNL-1583, Oct. 22, 1953. (pp. 48-56)

40. I. K. NamBa, Natural Circulation Recombiner Report, USAEC Report
CF-56-9-27, Oak Ridge National Laboratory, Sept. 10, 1956.

41. P. H. HarLEY, High-pressure Recombination Loop Progress Report, USAEC
Report CF-57-1-90, Oak Ridge National Laboratory, Jan. 4, 1957.

42. T. W. LELAND, Design of Charcoal Adsorbers for the HRT, USAEC Report
CF-55-9-12, Oak Ridge National Laboratory, Sept. 6, 1955. |

43. L. B. ANDERsSON, Oak Ridge National Laboratory, 1955. Unpublished.

44, J. S. CuLver and C. B. Granam, High-pressure Diaphragm Pumps for
Reactors, in Safety Features of Nuclear Reactors; Selected Papers from the 1st
Nuclear Engineertng Science Congress, December 12-16, 1955, Cleveland, Ohio.
New York: Pergamon Press, 1957. (pp. 225-230)

45. C. H. GaBBARD, Diaphragm Feed Pumps for Homogeneous Reactors, 4th
Engineering and Science Conference, Held in Chicago, Illinois, March 17-21,
1958. (Preprint 74)

46. R. BLuMBERG et al., Diaphragm Feed Pump Development Program Progress
Report, USAEC Report CF-56-10-114, Oak Ridge National Laboratory, Oct.
29, 1956.

47. Ohio State University, Union Carbide Nuclear Company, Contract
No. 81X-44934.

48. A. M. BiLrings, Control Valves for the Homogeneous Reactor Test, 4th
Nuclear Engineering and Science Conference, Held in Chicago, Illinois, March
17-21, 1958. (Preprint 149)

49. A. M. Biruings, Life Tests of Stem-sealing Bellows for HRT Valves,
USAEC Report CF-58-3-39, Oak Ridge National Laboratory, Mar. 17, 1958.

50. D. S. ToomB et al., in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 31, 1957, USAEC Report ORNL-2272, Oak
Ridge National Laboratory, Apr. 22, 1957. (p. 34)

51. B. A. HanNaForp, HRT Sampler Development, USAEC Report CF-57-
1-87, Oak Ridge National Laboratory, Jan. 22, 1957.

52. R. VAN WINKLE, Fuel Let-down Heat Exchanger, USAEC Report CF-54-
9-143, Oak Ridge National Laboratory, Sept. 20, 1954.

53. C. D. ZerBy, Design of Smoothly Flowing Gas and Liquid Maixtures,
USAEC Report CF-51-10-130, Oak Ridge National Laboratory, Oct. 11,
1951.

54. D. S. ToomB et al., in Homogeneous Reactor Project Quarterly Progress
Report for Pervod Ending Apr. 30, 1956, USAEC Report ORNL-2096, Oak Ridge
National Laboratory, May 10, 1956. (p. 32)

55. D. S. Toowms et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for Period Ending July 31, 1956, USAEC Report ORNL-2148(Del.), Oak
Ridge National Laboratory, Oct. 3, 1956. (p. 67)
REFERENCES 465

56. D. S. ToowMms et al., in Homogeneous Reactor Project Quarterly Progress Re-
port for Period Ending Jan. 31, 19567, USAEC Report ORNL-2272, Oak Ridge
National Laboratory, Apr. 22, 1957. (p. 35)
CHAPTER 9
LARGE-SCALE HOMOGENEOUS REACTOR STUDIES*

9-1. INTRODUCTION

9-1.1 The status of large-scale technology. A large number of groups
in the national laboratories and in industry have prepared detailed designs
of full-scale homogeneous reactors because of the widespread interest in
these reactors and the generally accepted conclusion that they have long-
term potential for central-station power production and other applications.
These designs have, in some cases, been made to compare the economics of
power production in homogeneous reactors with other nuclear plants. In
other cases, the designs have served as the bases for actual construction
proposals. Unfortunately, none of the proposals has yet initiated the con-
struction of a reactor, for it is believed that the gap between the existing
technology of small plants and that necessary for a full-scale plant is too
great to bridge at the present time. Thus the construction of full-scale
plants must await further advances in technology which are expected to
be achieved in the development programs now under way. The extensive
studies of full-scale plants do, however, constitute a body of information
vital to the nuclear industry. It is hoped that the summaries of the large-
- scale homogeneous reactors given in this chapter will serve as a guide to
those contemplating the building of a full-scale nuclear plant.

One of the major problems yet to be solved for a large-scale circulating-
fuel reactor is that of remotely repairing and/or replacing highly radioactive
equipment which fails during operation of the plant.

The various proposed solutions to this problem fall into two categories:

(1) Underwater maintenance, in which all equipment is installed in a
shield which can be filled with water after shutdown of the reactor so that
maintenance operations can be performed from above with special tools
and with visibility provided through the water.

(2) Dry maintenance, in which all operations are done by remote meth-
ods using special remotely operable tools and remote viewing methods
such as periscopes and wired television.

In either case, remote opening and closing of flanged joints or remote
cutting and rewelding of piping must be used to remove and replace equip-
ment. A solution of the problem of maintaining flanged joints in a leaktight
condition in large sizes has not been attempted, the largest pipe in use to

*By C. L. Segaser, with contributions by R. H. Chapman, W. R. Gall, J. A.
Lane, and R. C. Robertson, Oak Ridge National Laboratory.

466
9-1] INTRODUCTION 467

date being approximately 10 in. in diameter. Remote cutting and re-
welding equipment is still in the early stages of development.

The technology of solutions systems is in a more advanced stage of
development than that of slurry systems because of the design and opera-~
tion of two homogeneous reactor experiments and the associated develop-
ment work. Some of the problems remaining to be solved for large-scale
solution reactors include the development of large-scale equipment such as
pumps, valves, feed pumps, and heat exchangers; radiation corrosion of
materials used in the reactor core; high-pressure recombination of hydrogen
and oxygen; and reduction of the number of vital components upon which
reactor operation depends. Instruments for measuring temperature in
high radiation fields and control of inventory and level are some of the
major instrumentation problems for which better solutions are needed.

The achievement of a successful aqueous homogeneous thorium breeder
requires a high-pressure thorium-oxide slurry system. Development work
has been under way for several years to determine the characteristics of
such a system and to develop ways of handling slurries. The technology is
not yet advanced to the point where a large-scale breeder reactor of this
type can be built and operated. Slurry problems under study include
methods of production, circulation through pipes and vessels, storage and
resuspension, evaporation, heat removal, flow distribution, particle size
degradation, internal recombination of deuterium and oxygen, general
information on erosion and corrosion effects, and effects of settling on
maintenance operations.

Extrapolation of small-scale technology to large-scale design presents
several problems of uncertain magnitude, especially in the design of equip-
ment for handling slurries of thorium oxide such as are specified for one-
or two-region breeder reactors. The problems of maintenance of slurry
systems are essentially the same as for solutions, but are complicated by the
erosive nature of the slurry, its relatively high shear strength, and its tend-
ency to cake or settle in regions of low turbulence.

0-1.2 Summary of design studies. The design studies described in this
chapter were made by the national laboratories of the Atomic Energy
Commission and by various industrial study groups for the purpose of
determining the technological and economic feasibility of aqueous homo-
geneous reactor systems as applied to central station power, research
reactors, and the production of plutonium. In general, the design criteria
used in the studies conform as closely as possible to known technology in
order to minimize the scope of new developments required to ensure the
success of the proposals. In all the studies, the importance of over-all
safety and reliability of the reactor complex and individual reactor com-
ponents has been emphasized. Also, considerable attention has been
devoted to the maintenance aspects of the designs.
468 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

The large-scale reactor designs described are grouped according to
the following categories:

(1) One-region solution reactors, typified by the Wolverine Reactor
Study, the Oak Ridge National Laboratory Homogeneous Research
Reactor, and the Aeronutronic* Advanced Engineering Test Reactor.

(2) One-region breeders and converters, such as the Pennsylvania Ad-
vanced Reactor reference design by Westinghouse Electric Corporation,
the Homogeneous Plutonium Producer Study by Argonne National
Laboratory, and the Dual Purpose Feasibility Study by Commonwealth
Edison.

(3) Two-region breeders, represented by the Nuclear Power Group studies,
the Babcock & Wilcox Breeder Reactor, and a sequence of conceptual
designs by the Oak Ridge National Laboratory Homogeneous Reactor
Project.

9—2. GENERAL Prant LAvouT AND DESIGN

9-2.1 Relation of plant layout to remote-maintenance methods. In lay-
ing out a homogeneous reactor plant, the designers, to achieve an optimum
arrangement, must simultaneously consider all aspects of the design,
including the requirements for remote maintenance. It is usual to start
with the high-pressure reactor system (the reactor vessel, circulating pump,
steam generator, and surge chamber and pressurizer), since there exists a
natural relationship between these items in elevation. The layout will
depend primarily on whether a one-region reactor or two-region reactor is
involved, since in the latter case special provision for removing the inner
core may be necessary. If it is feasible to construct the reactor vessel and
core tank as an integral all-welded unit, the layout of the system will be
considerably simplified. Otherwise, provisions will have to be built into
the reactor vessel and the reactor system to remove the vessel and/or the
core tank.

Circulating pumps are vulnerable from the standpoint of long-term re-
liability, and extreme care must be given to their placement and anchorage
in the system layout. Installations and designs to date place the circulating
pumps in a position following the steam generator and the gas separator
and low in the cell in order to provide as low a temperature and least gas-
binding conditions as possible. These pumps, however, will operate at an
overpressure considerably in excess of saturation pressure, and if gas binding
does not prohibit, it may be desirable to place the circulating pumps at a
position more accessible for maintenance.

The placement and design of the steam generators will be dictated to a
major degree by the maintenance philosophy adopted. One general

*Aeronutronic Systems, Inc., a subsidiary of Ford Motor Company.
9-2] GENERAL PLANT LAYOUT AND DESIGN 469

philosophy being considered uses many small steam generators in order to
permit easier removal and replacement when necessary. Another philoso-
phy considers the repair of the steam generators in situ, using remotely
manipulated tooling. One difficulty with this scheme will be the problem
of finding leaky tubes.

The steam generators are usually one of the bulkiest items of equipment
installed in the plant and hence will largely determine the size of contain-
ment vessel and amount of shielding. Their location should be such that
some heat-removal capacity can be obtained by natural convection cir-
culation in the event of failure of the circulating pump.

In considering the layout of the surge chamber (which is normally also
the pressurizer) connecting piping must be as short as possible and the
diameter of the piping should be large for safe control of the reactor. If
a steam generator is used to provide high-pressure DO or H2O vapor, it
should be separated from the surge chamber, and preferably placed in a
location separate from the reactor compartment to facilitate maintenance.

9-2.2 Importance of specifications. To ensure that materials such as
type—347 stainless steel and titanium and zirconium alloys meet the quali-
fications required for homogeneous systems, very rigid specifications cover-
Ing strength, corrosion-resisting properties, impact resistance, etc. must
be prepared. To ensure leaktight integrity, specifications describing
acceptable weld joints and welding procedures are issued. Such specifica-
tions will also describe the welder qualifications required. Since it is im-
perative that the main process piping system shall be absolutely clean and
purged of any material which may poison the reactor or accelerate corro-
sion, cleaning procedures are a necessary part of the specifications.

90-2.3 Approach to an optimum piping system. The cost of the piping
system 1s one of the major items of expense, and its selection and arrange-
ment constitutes one of the major items of design. However, the pipe
diameters are generally determined on a maximum-velocity basis, deter-
mined by corrosion rates rather than from economic considerations. The
welght classification (i.e., pipe wall thickness) is selected on the basis of
pressure, temperature, and corrosion rate for the proposed service life
of the reactor system using the appropriate design stresses from the ASME
Code for the particular metal used. Other factors influencing piping layouts
are (a) provision for drainage, (b) provision for expansion, (¢) accessibility
and convenience of operation, (d) provision for support, and (e) the thick-
ness of insulation. ‘

Long straight runs of high-temperature, high-pressure piping present
the main problem so far as expansion is concerned. Natural anchorages
should be noted, and at the same time, possible locations should be sought
470 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

for special anchors needed to control expansion in accordance with the
design plan. The efficiency of the piping system layout depends largely
on the ability of the designer to visualize the over-all situation and to
select the best arrangement. The design of a piping system for minimum
holdup may be relegated to secondary importance compared with ease of
maintainability of the systems.

Piping joints. Piping joints for homogeneous reactor systems must be
capable of assembly and disassembly by remote methods and must have
essentially zero leakage. The first requirement implies some type of
mechanical joint such as used on the HRE-1 and HRE-2. The second
specification can only be guaranteed by an all-welded piping system, and
consequently an all-welded piping layout may be necessary for large-scale
homogeneous reactor systems. However, such a system requires a reliable
and easily manipulatable remote cutting and welding machine not yet
developed.

9-2.4 Shielding problems in a large-scale plant. Poor shield design can
lead to excessive cost and reduced accessibility for maintenance. Practical
shield designs are developed through the use of methods in the literature [1]
with particular attention to factors pertaining to the shield layout, such
as the arrangement of the piping and heat-exchanger system, materials se-
lection, radioactivity of the shutdown system, effect of radiation streaming
through openings, and the effect of the geometry of the radiation sources.

A number of proposed designs of large-scale homogeneous reactors use
a compartmentalized type of shield. This consists of a primary shield sur-
rounding the reactor pressure vessel to attenuate the neutron flux and re-
duce the radioactivity of auxiliary equipment, and a secondary shield sur-
rounding the coolant system. From a shielding standpoint, the most
highly radioactive sources should be located near the center of the com-
partment, components of lower source strength should be arranged pro-
gressively outward, and equipment with little  radioactivity should be
located to serve a dual purpose as shielding material where possible. High-
intensity sources containing primary coolant, which are poorly located
from a shielding standpoint, may be partially shadow-shielded. Equip-
ment requiring little or no maintenance and which can provide shielding
should be located around the outside of the secondary shield. Considerable
weight can be saved by contouring the secondary (coolant) shield (i.e.,
varying its thickness over the surface) to give closer conformance with the
specified permissible dose pattern. With respect to sample lines which
penetrate the coolant shield and contain radioactive materials of short
half-lives, the transport time from the primary coolant system to the out-
side of the shield should be made as long as practical to take advantage of
the decay of the coolant activity.
9-2] GENERAL PLANT LAYOUT AND DESIGN 471

9-2.5 Containment. Because of the possibility of release of highly
radioactive fuel solution from a homogeneous reactor, such systems are
now being designed to go within a containment vessel or to use doubly con-
tained piping. The containment vessel must be designed to hold the
pressure resulting from expansion of the fluid and vapor contents of the
equipment. Such pressures may be of the order of 50 psi. Although the
best shape of a containment vessel is either a cylinder or a sphere, such con-
figurations present problems with respect to remote maintenance. To pre-
vent the penetration of the containment vessel by flying fragments which
may be released on failure of equipment, a blast shield can be placed around
the periphery of the containment vessel or relatively close to the equip-
ment.

0-2.6 Steam power cycles for homogeneous reactors.* In common with
other pressurized-water types of reactors, homogeneous reactors are handi-
capped by the high pressures required to prevent boiling in comparatively
low-temperature aqueous fluids. Temperatures for steam generation in
homogeneous reactor power systems are limited to practical maximums of
500 to 600°F, and there are no significant opportunities for superheating
the steam with reactor heat. Separately fired superheating equipment,
using conventional fuels, may be expedient in some particular circum-
stances of plant size, load factors, and fuel costs but, in general, superheat-
ing by this means is not justified. Use of low-pressure saturated steam
limits the thermal efficiency obtainable in the heat-power cycle to maxi-
mum values in the range of 25 to 30%.

Homogeneous reactor systems circulate a hot reactor fluid to a steam
generator at essentially constant temperature; the temperature of the fluid
leaving the exchanger is varied with load by changing the temperature dif-
ference for heat transfer by controlling the pressure at which the water
boils in the steam generator. Since the steam pressure falls as the turbine
control valves open on increased electrical generator loads, the negative
temperature coefficient for the reactivity causes the reactor power output
to be self-regulating to match the power demand on the plant.

The full-load steam pressure will be in the order of 450 to 600 psia, and
the near no-load pressure in excess of 1000 psia. The steam piping and
turbine casing must be designed for this maximum pressure rather than
the full-load pressure; design pressures of 1500 psia have been used for the
steam systems of the HRE-1 and HRE-2.

Turbines designed for operation on saturated steam will cost more per
kilowatt of installed capacity than turbines designed for superheated steam
(see Chapter 10). The relatively low energy content, high specific volume
steam supplied to the saturated steam turbine throttle requires greater

*By R. C. Robertson.
472 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

mass flow rates and flow areas for a given power output, adding to the cost
of the governor valves and the high-pressure stages. The low-pressure
stages, which are the most expensive, also must have more flow area, which
may require additional compounding, adding greatly to the cost. The
turbine efficiency will be somewhat less in a saturated steam turbine than
in one using superheated steam because of the greater amount of en-
trained moisture in the steam.

Flow delay tanks in the steam supply mains are considered necessary to
allow time for a stop valve to close in the event radioactivity is detected
in the steam flow from the heat exchangers. Heat losses in this equipment
tend to increase the moisture in the steam, and a separator may be re-
quired at the turbine inlet. The pressure loss in typical separators is about
5% of the inlet pressure and the leaving quality about 99%. Some method
of moisture removal must be provided during the expansion process, either
internally in each of the low-pressure stages, or externally in one or more
separators located between turbine elements. Studies have indicated that
the optimum location for the first stage external moisture separator is at
109, of the throttle pressure. The presence of more moisture in the ex-
panding steam may require that the turbine be an 1800-rpm rather than a
3600-rpm machine.

As with steam power cycles for other reactor types, an emergency by-
pass will probably be required to send the steam directly to the turbine
condenser in event of loss of turbine load. The condenser must be designed
to dispose of the full output energy of the reactor plant. The number of
stages of feedwater heating economically justified is probably limited to
three or four, since the temperature range in the cycle is not great.

Treatment of the water fed to the steam generators is a special problem
in that the water should be essentially free of chloride ions to reduce the
opportunity for stress-corrosion cracking in stainless steel parts of the
system, the water should be deaerated to control corrosion in the steam
system, 1t should be demineralized to reduce the radioactivity pickup of
the steam and in the heat exchanger blowdown, and additives may be
necessary to control the pH and to scavenge oxygen formed by radiolytic
decomposition of the water. Decomposition of these additives under radi-
ation poses problems not yet fully investigated.

Control of the water level in the steam generator involves much the
same problems, due to steam bubbles that are experienced in conventional
boilers, with the added complexity that the steam pressure increases as the
load on the plant decreases. Sizing of the ports in the feedwater regulating
valves must take this into consideration, and the boiler feed pumps must
be designed for the no-load, rather than the full-load head requirements.

Although some superheat can be obtained by recombining the decompo-
sition gases, it is doubtful if such a procedure is economical, owing to the
relatively small amount (59)) of superheat obtained, and also because of
9-3] ONE-REGION U235 BURNER REACTORS 473

 

34lllllllll|llll_

N
o

Thermal Efficiency, %

 

 

 

D03 N I N T I O O
350 400 450 500
Throttle Steam Temperature ,°F

 

Fia. 9-1. Effect of steam conditions on turbogenerator plant efficiency.

the desirability of minimizing gas production within the reactor. Unless
the superheat is more than 100°F, a saturated cycle with moisture separa-
tion may be equally as efficient and practical as a cycle using superheated
steam, provided that in either case the moisture in the turbine exhaust is
kept the same. It is also possible to superheat at the expense of throttle
pressure; while superheat normally is considered to increase the thermal
efficiency, this is not true if the inlet steam temperature is independent of
the amount of superheat. Also, the lower pressure associated with throttling
results in increased turbine costs. Superheating by means of a conventional
plant does not appear economical.

In studies of homogeneous reactors, saturated steam cycles are assumed
in which 129, moisture is permitted in the last stages of the turbine.
Thermal efficiencies of such plants are shown in Fig. 9-1 as a function of
the steam temperature at the turbine throttle [2].

0-3. OnE-REGION U235 BURNER REACTORS

9-3.1 Foster-Wheeler Wolverine Design Study. In response to a re-
quest by the Atomic Energy Commission for small-scale power demon-
stration reactors, the Foster Wheeler Company proposed to construct an
aqueous solution reactor for the Wolverine Electric Cooperative in Hersey,
Michigan [3]. This proposal was rejected by the Atomic Energy Commis-
sion in October 1957 as a basis for negotiation due to increases in the esti-
mated cost of the plant (from $5.5 million to $14.4 million). The project
was canceled in May 1958 following a review. of the design and estimated
costs. This review indicated that the cost of generating electricity would
be several times as great as that in Wolverine’s existing plant.

In December 1957 a group of engineers from the Oak Ridge National
Laboratory and Sargent and Lundy, with the help of Foster—Wheeler, re-
designed the reactor on the basis of recent advances in homogeneous re-
actor technology and re-estimated its costs to be $10.7 million [4]. The
474 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

Sample Valve, Reactor Press. Steam

 

 

 

  

 

 

 

 

 

 

         
 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  
   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

pCondensate Tanks { Vessel /Drum
S N e g 3 / a kA Ry
e jz; @j ] Vg!;re ;|  Gauge Board :
F o LA e TN ] -
e\ | e N e M
Condenser) - v N Yt Ty | I ;
B ; N« AAnalls [Anall 3 N
~ W . X s v nQ o, nG . . _ - | |
. b = ! e | Cellf | Cell| L ;
Cooling Pit o : N . I
] , F_'|._;_~ 1 -n> s 740 L
4y M o 1« ’ ~) Lt
34 ! £ v
A T Ml A | s o, . 3R - .
. : [ 2 o vy 3 3.
O 0 T3 J AN 47D <. L
| i 'a 3N . A 45 o3 ;
A B0 '48 —“H oy TRy SN g * A
> \ e T AT 1
o Nt HO‘(&% 5 zardox Q 4wy [ C__4 D :.1
S N\ Ty e B e T oV
Waste &  NOff Gas Plan Chem. Intermediate Heat
Decont. Cell Process Exchanger
Cell
t
Access Hatch Steam Control Room
1 T
Waste-Decont. ,Off Gas 40 Ft. Dia. Sphere [_!TI H
) 1 §
/Cell Cell Press. Basement\\i | ] i
A/ H 1T T T F
b v . p jo o \ 5
v k N £ Anal. [+ | Anal.}* I_] H ;
: [ .3 O — > . "
, al3 ';_76 Cell | Cell 3
s T b E U
>
o

 

 

 

 

X 991[] HH .

. VA

 

 

 

 
 

Heat Exchanger

 

 

 

Elevation-Section A-A

Fic. 9-2. Plan and sectional elevation of revised Wolverine reactor plant.

following section describes the revised reactor design. Figure 9-2 shows a
plan and elevation sketch of the revised concept.

The fuel solution of highly enriched uranyl sulfate in heavy water is cir-
culated by a canned-motor pump located in the cold leg of the primary
loop and pressurized to prevent boiling and cavitation in the pump. The
steam generated in the heat exchanger is superheated in a gas-fired super-
heater, and the superheated steam drives conventional turbogenerating
equipment for the production of electricity.

The nuclear reactor plant is designed to permit initial operation at 5 Mw
with a single superheater-turbogenerator unit. By adding a second unit,
the capacity can be increased to 10 Mw. Doubling the electrical capacity
1s thus accomplished without making any changes to the reactor other
than adjusting the operating temperatures and uranium concentration.

For 10 MwE operation, 31,000 kw of heat is generated in the reactor
under the following conditions: The hot fuel solution leaves the core at
300°C, is circulated through a heat exchanger, and returns to the reactor
at 260°C. The heat generated in the reactor is transferred to boiling water,
9-3]

ONE-REGION U232 BURNER REACTORS

TAaBLE 9-1

(10-MwE OPERATION)

1. Core

Configuration
Core diameter: inside thermal shields, ft
over-all, ft
Wall thickness, in.
Liquid volume, liters
Power density, kw/liter
Core wall (inner thermal shield)
Average for system
Maximum
Initial fuel concentrations (critical at 300°C), m
U235
CuS04
H2SO4
Steady-state fuel concentrations, m
U235
Total U
CuSO4
HZSO4
NiSO,4

. Pump

Fuel flow rate, gpm at 260°C

Head, ft

Approximate pumping power, hp
(assumes 509, over-all efficiency)

3. Heat exchanger

Shell diameter, in.

Tube diameter, in.

Tube wall thickness, in.

Number

Approximate inside area wetted by fuel solution, ft?
Steam temperature, °F

Log mean average temperature difference, °F
Over-all heat transfer coefficient, Btu/(hr) (ft2)

. Pressurizer

Inside diameter, in.
Wall thickness, in.
Length of cylindrical portion

DesigcN DATA FOR THE REVISED WOLVERINE PRIMARY SYSTEM

Concentric outlet

0.014
0.02
0.02

0.030
0.034
0.02

0.025
0.017

2750
65
80

29
1/2
0.065

1120
4100
480
39
500

56
3
6 ft 9 1n.

475
476 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

TABLE 9-1 (Continued)

Concentric outlet

Volume of solution at low level, liters 150
- Net gas volume (liquid at low level), liters 1400
Normal operating pressure, psia 1900
Normal operating temperature, °F 570
5. Piping
Nominal diameter, in. 10
Wall thickness, in. 1.125
Approximate total volume, liters 950
Maximum velocity, fps 17
6. Estimated power costs (10-MwE plant) Mills/kwh
31-Mw reactor plant ($8,740,000)
Fuel burned 2.83
Fuel inventory @ 49 0.67
(9000 kg D20 4 36.5 kg U235)
Fuel processing 2.46
Fuel preparation 0.62
D20 losses 0.30
Depreciation @ 159 18.72
Operating costs 1.43
Maintenance costs 3.85

10-MwE superheater-turbine generator plant
($1,940,000)

Fuel (oil) 0.69
Depreciation @ 159, 4.17
Operating costs 0.29
Maintenance costs 0.29
Total power costs 36.32

producing 116,000 1b/hr of steam at 600 psia. For operation at 5 MwE,
the hot fuel solution would leave the reactor core at 276°C and return at
257°C, producing 58,000 1b/hr of steam at 600 psi.

A pressurizer is connected to the outlet of the heat exchanger to pressur-
ize the system with oxygen to 1900 psia and to provide a location in the
primary system for the removal of fission-product and other noncon-
densable gases. The layout of the primary system is such as to permit heat
removal by natural circulation in case of pump failure.

A low-pressure system consisting of dump tanks, condenser, and con-
densate tanks is incorporated to handle fluid discharged from the primary
loop and to furnish heavy water required to purge the canned-motor cir-
9-3] ONE-REGION U239 BURNER REACTORS 477

culating pump. TFacilities for adjusting fuel concentration and maintaining
a continuous record of fuel inventory are also included.

Design data. Pertinent design information for the reactor systems and
components is summarized in Table 9-1 and described in the following
paragraphs. Unless otherwise noted, all surfaces in contact with fuel solu-
tion are fabricated of type—347 stainless steel.

Equipment and system descriptions. Reactor vessel. The single-region,
concentric-inlet and -outlet pressure vessel designed for 2500 psia in-
corporates two inner concentric thermal shields to reduce gamma heating
effects in the outer pressure vessel. The thermal shields are constructed of
type—347 stainless steel and are 1 in. and 2 in. thick with inside diameters
of 5 ft 0 in., and 5 ft 5 in., respectively. Backflow through the vessel drain
line during normal operation provides some cooling of the outer thermal
shield.

Primary heat exchanger. The steam generator consists of a horizontal
U-shaped shell-and-tube heat exchanger with a separate steam drum.
These are interconnected with downcomers and risers to provide natural
circulation of the boiling secondary water. Fuel solution is circulated on
the tube side of the heat exchanger, and the boiling secondary water is
circulated on the shell side. Feedwater is introduced into the liquid region
of the steam separating drum. All components in contact with secondary
water and steam are to be fabricated from conventional boiler steels.

Fuel circulating pump. A single, constant-speed, water-cooled, canned-
motor type pump is provided to maintain fuel circulation in the primary
loop. The rotating elements are removable through the top of the unit,
and may be removed without disturbing the piping connections to the
stator casing or the pump volute. Regions of high fluid velocity in the
pump, including the impeller, are titanium or titanium-lined. A purge
flow of condensate is fed into the top end of the pump to reduce erosion and
corrosion of bearings, as well as to prolong the life of the motor windings
by reducing the radiation dose to the electrical installation. In the event
of pump failure, the reactor will undergo a routine shutdown and the
fission-product decay heat will be removed by natural circulation through
the steam generator.

Pressurizer. A small sidestream of fuel solution is continuously directed
into the pressurizer, where it spills through a distribution header and drips
down through an oxygen gas space to the liquid reservoir in the bottom of
the vessel. The pressurizer liquid return line is connected to the suction side
of the primary-loop circulating pump. Oxygen is added batchwise to the
pressurizer to keep the fuel saturated at all times to prevent precipitation
of uranium. As fission-product gases accumulate in the pressurizer, they
are vented to the off-gas system, also in a batchwise operation.

Fuel makeup pump. Two diaphragm-type high-head pumps (one for
478 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

standby) rated at 3 gpm at a pressure of 1900 psi are provided to add ura-
nium to the fuel solution and to fillthe primary system with fluid on
startup.

Dump tanks. The dump tanks, 48 ft long and 28 in. ID, are designed to
remain subcritical while holding the entire contents of the primary system.
An evaporator section underneath each of the vessels is provided to con-
centrate the fuel when necessary, and to aid in mixing the contents of the
tank. |

Containment. The primary coolant system is enclosed in a 40-ft-diameter
spherical carbon-steel vessel, lined with 2 ft of concrete, interconnected
with a 12-ft-diameter by 50-ft-long stainless-clad vessel housing the dump
tanks. The liner serves the dual functions of missile protection and struc-
tural support to withstand the loading of the external concrete. An addi-
tional 1/8-in. stainless steel liner is furnished to permit decontamination
of the primary cell. Since these vessels provide a net containment volume
of approximately 31,000 ft3, the vaporization and release of the reactor
contents results in a maximum pressure of approximately 105 psi. Ac-
cordingly, the primary-cell containment vessel wall thickness is 15/16 in.
and the dump-tank containment vessel wall thickness is 9/16 in. A spray
system 1s incorporated in the design to quickly reduce the pressure within
the containment vessel by condensing the water vapor present.

A bolted hatch 1s provided in the top head of the vessel to allow access
and removal of equipment for maintenance. A bolted manway is also
provided to permit entrance into the containment vessel without removing
the larger auxiliary hatch. In the event of a major maintenance program,
however, the top closure would be cut and removed for free access to the
primary cell.

Brological shielding. The plant biological shielding is indicated on the
general arrangement drawing (I'ig. 9-2). The shielding for the primary
system, including the reactor core, consists of a 2-ft thickness of ordinary
concrete lining the inside of the primary-cell containment vessel and a
minimum of 7 ft of concrete surrounding the outside of the vessel, cooled
by a series of cooling-water coils located in the 2-ft-thick liner. The top of
the primary vessel is shielded with 6 ft of removable blocks of barytes
aggregate concrete (average density of approximately 220 1b/ft3) located
beneath the removable portion of the containment vessel.

A 2-ft-thick water-cooled heavy aggregate thermal shield is placed
around the reactor vessel to reduce the radiation level to approximately
that of the remainder of the primary system. The primary coolant pump
access pit, located inside the containment vessel, is constructed of 3% ft of
barytes aggregate concrete to permit pump removal after the primary cell
has been filled with water and the system drained and partially decon-
taminated. During periods of normal operation, the temperature of the
9-3] ONE-REGION U232 BURNER REACTORS 479

concrete walls and floor of the pit is maintained at 150°I" by cooling-water
coils.

Around each of the analytical and chemical processing cells there will be
a minimum of 4 ft of ordinary concrete with a maintenance gallery between
these facilities for access to, and operation of, the cells. Iach of the two
analytical cells will be provided with thick glass windows adequate for
shielding. The dump-tank cell will be shielded by a 5-ft thickness of
concrete. |

Remote maintenance. Both dry and underwater removal methods are
proposed for remote maintenance of radioactive components in this system,
following practices similar to those developed for HRE-2. All the equip-
ment cells are provided with stainless-steel liners to permit the cells to be
filled with ordinary water during maintenance operations. I'or removal of
the large components it is necessary to move the container vessel cover
through the west end of the building to a temporary storage area. After
the primary vessel cover and top shield are removed, the system com-
ponents are accessible by crane and operations are performed with specially
designed long-handled tools.

9-3.2 Aqueous Homogeneous Research Reactor—feasibility study. A
preliminary investigation of the feasibility of an aqueous homogeneous
research reactor (HRR) for producing a thermal flux of 5 X 10'® neu-
trons/(cm2)(sec) was completed by the Oak Ridge National Laboratory in
the spring of 1957 [5]. The design considered is illustrative of a homogene-
ous reactor capable of producing high neutron fluxes for research and power
for the production of electricity. It consists of a 500-Mw (thermal) single-
region reactor with 8% enriched uranium as the fuel in the form of uranyl
sulfate (10 g of total uranium per kilogram of D20) with sufficient copper
sulfate added to recombine 100% of the radiolytic gases produced and
excess sulfuric acid to stabilize the copper sulfate, uranyl sulfate, and
corrosion-product nickel.

The system operates at solution temperatures of 225 to 275°C, and a
total system pressure of 1400 psia. Under these conditions a maximum
thermal neutron flux of 6.5 X 10!5 neutrons/(cm?)(sec) is achieved n a
10-ft-diameter stainless-steel-lined carbon-steel sphere. = Approximate
power densities are 2 kw/liter at the core wall, 35 kw/liter average, and
110 kw/liter maximum. After correcting for the effect of experiments, a
maximum thermal flux of about 3 X 10!® neutrons/(cm?)(sec) and a fast
neutron flux of about 5 X 10 neutrons/(cm?)(sec) are available.

To minimize corrosion of equipment and piping in the external circuit,
all flow velocities are held to values below the critical velocities. Estimated
corrosion rates are 70 to 80 mpy for the Zircaloy—2 experimental thimbles
and about 10 mpy for the stainless-steel liner of the reactor vessel (based
on a maximum flow velocity of 3 fps).
480 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

TABLE 9-2

HRR STEAM-GENERATOR SPECIFICATIONS
(ONE UnIT)

 

Reactor fluids, forced circulation (tube side)

Inlet temperature, °F 527
Outlet temperature, °F 437
Flow rate, Ib/hr 2,730,000
Pressure, psia 1400
Velocity through tubing, fps 10
Steam, natural recirculation (shell side)
Generation temperature, °F 417
Pressure, psia 300
Generation rate, 1b/hr 351,600
Heat load, Btu/hr 284,300,000
Heat load, Mw 83.3
Steam generator
Number of 3/8-in. 18 BWG tubes 3280
Effective length of tubing, ft 25.9
Heat-transfer surface, ft2 8330
Shell internal diameter, in. 383
Shell thickness, in. 1%
Tube-sheet thickness, in. 5
Steam drum
Internal diameter, in. 36
Length, ft 16
Wall thickness, in. 13
Height above generator, ft 15

 

 

 

 

Fission- and corrosion-product solids, produced at a rate of approxi-
mately 20 lb/day under normal reactor operating conditions, are con-
centrated into 750 liters of fuel solution by means of hydroclones with self-
contained underflow pots and removed from the reactor to limit the buildup
of fission and corrosion products. This solution is subsequently treated for
recovery of uranium and D:O.

The temperature coefficient of reactivity at 250°C is approximately
—2.5 X 1073/°C and at 20°C is approximately —9 X 10~%/°C, which, in
combination with fuel-concentration control, is adequate for operation
without control rods.

Reactor vessel. The 10-ft-ID spherical pressure vessel is designed according
to the ASME Unfired Pressure Vessel Code, with consideration given to
9-3]

ONE-REGION U235 BURNER REACTORS

481

TaBLE 9-3

Ky DESIGN PARAMETERS

 

 

Reactor type

Fuel type
Amount of U232
Uranium concentration
Total uranium
U235
CuSO4 to recombine 1009, of gas
H>S04 to stabilize uranium and copper
Maximum nickel concentration
Fuel-solution temperature
Minimum (inlet to reactor vessel)
Maximum (outlet of reactor vessel)
Average (system)
Fuel system pressure
Neutron flux (experimental)
Maximum thermal
Maximum fast in 1-in. diameter
cylindrical converter
Power density
Maximum (at reactor center)
Average
Minimum (at thermal shield)
Total heat generated
Reactor-vessel key specifications
Inside diameter
Vessel material

Total volume

Net fluid volume (approximate)
Experimental facilities

Horizontal

Vertical

Maximum inside diameter

Material

Minimum wall thickness

Maximum wall thickness

 

Single-region, circulating-fuel, homo-
geneous

UOzSO4 - DzO + CuSO4 + HzSO4

45.8 kg

10 g/liter at 250°C
0.8 g/liter at 250°C
0.02m
0.02 m
0.01m

225°C
275°C
250°C
1400 psi

3-4 X 10'5 n/(cm?)(sec)
4 x 10 to 1 X 10'5 n(cm?)(sec)
110 kw/liter

34 kw/liter
2 kw/liter

500 Mw

10 ft

Carbon steel clad with type-347
stainless steel |

14,800 liters

12,000 lLiters

6

1

6 1n.
Zircaloy—2
3/4 in.

1 in.

 

continued

 
482 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

TABLE 9-3 (Continued)

 

External system
Material Type-347 stainless steel, HRP speci-

fications

Fluid volume (external system only) | 34,000 liters

Allowable velocities

225°C 10-15 fps
250°C 25-35 fps
275°C 30-40 fps
Reactor control Negative temperature coefficient of
reactivity, changes in concentration
of fuel
Heat dissipation Generation of approximately 125 Mw

of electrical power

 

 

 

 

the special problems introduced by the heating of the shell from radiation
absorption and by the necessity of penetrating the shell for insertion of
experimental thimbles. The proposed vessel is fabricated of a carbon-steel
base material with a type-347 stainless steel cladding on all surfaces
exposed to fuel solution.

The fuel solution enters the vessel through two 24-in. nozzles, sized for a
fluid velocity of 10 to 15 fps, flows upward through the vessel, and exits
through two 18-in. nozzle connectors in the top, sized for a fluid velocity
of 30 to 40 fps. A diffuser screen, serving also as part of the thermal shield,
is placed at the entrance to the reactor vessel.

A stainless steel blast shield is placed around the reactor vessel to con-
tain fragments of the vessel in the event of a brittle failure, and cooling coils
are wrapped around the blast shield to control the pressure-vessel tem-
perature.

Heat exchangers (steam generators). Six heat exchangers of 83.3 Mw
capacity each are required to dissipate the 500 Mw of heat generated in the
reactor. The design of these consists of a lower vaporizing shell connected
to a steam drum at a suitable elevation to promote natural circulation by
means of risers and downcomers welded to the shells. Specifications are
summarized in Table 9-2. |

Pressurizer. 'The pressurizer surge chamber, constructed of 24-in.
schedule-100 pipe provides the necessary 1500 liters of surge volume.
Steam is provided in a small high-pressure steam generator physically
separated from the pressurizer surge chamber. Space limitations and ac-
cessibility problems make this separation desirable.
9-3] ONE-REGION U235 BURNER REACTORS 483

Biological Shield Outline

      
 
     
   
  

— Heat Exchangers Crane Rail

/

Cell For Remote
Operated Tools
Concrete Wall
Biological Shield

Shield

     
 
 
  
 
  
 

Circulating Transfer Carriage

Maintenance Building

Reactor

Surge
Chamber
Experimental Thimbles 8 0 8 16 24 32
e —— e — ]

Scale in Feet

O

Fia. 9-3. Homogeneous Research Reactor layout plan view.

System design. Two 17,650-gpm pumps mounted on the outlet pipes of
‘the heat exchanger circulate the reactor solution around the primary cir-
cuit. Saturated steam at 300 psia is generated at a rate of 2.11 X 10° Ib/hr
and is used to generate 133,000 kw of gross electrical power at a cycle
efficiency of 26.5%. A net power generation of 125,000 kw will be delivered
at the station bus bars, approximately 6% being required for station
auxiliaries. Feedwater, consisting of D20 from the condensate tank, is
supplied to the steam generator through an economizer by means of a
0.5-gpm feedwater pump. The reactor does not contain a letdown system
for separating and recombining radiolytic gases, since 100% internal re-
combination will be achieved by means of internal copper catalyst. Key
design parameters are summarized in Table 9-3.

Conceptual layouts of reactor complex. Preliminary conceptual layouts
showing the relation of the items pertaining to the nuclear reactor com-
ponents are given by Figs. 9-3 and 9-4.

Figure 9-3 is a plan view of the reactor complex, 1ndlcat1ng the general
relation of the reactor pressure vessel and its auxiliaries to the heat ex-
changers and circulating pumps. Shielded cubicles around the reactor
provide a means for handling the experimental thimbles. The outer
diameter of the containment vessel around the cubicles is approximately
60 ft. Approximately 6 ft of high-density concrete is placed around the
reactor area, with an additional 3 ft around the periphery of the contain-
ment vessel.
484 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

 

  
 

f:Crcme Sy
/ Le-Shielded Control Cab For
T : .| Circulating D Crane Operator

  

 

 

 

Pump,

 

Removable Plugs
Y F:

: SUI'ge ~. — LP ‘ : : ; E i i I !
1 Chamber-’ ressurizer ¥ il : ; r : : !i 'Concrete
3 . g ’ | I i

e >,
=

 

 

 

H
1

 

 

[ Reactori.

 

 

- {Dump Tanks
{ N
', d.

 

 

 

 

 

8 0 8 16 24 32
Scale in Feet

Fia. 9-4. Homogeneous Research Reactor layout sectional elevation.

A sectional elevation of the reactor complex is shown in Fig. 9-4. The
arrangement of the heat exchangers relative to the reactor vessel is such
that natural circulation through the system will be promoted in the event
of pump failure. Since the centerline of the reactor vessel is located at 30
to 36 in. above the operating-floor level for convenience in experimentation,
the containment vessels for the heat exchangers and circulating pumps
are above ground. The containment vessel for the reactor is a vertical
~cylindrical tank. Two separate horizontally mounted containment vessels,
each 60 ft in diameter, house the heat-exchanger equipment. The dump
tanks are directly below the heat-exchanger containment vessels. Means
for limited access to those portions of the dump-tank system which will
require periodic maintenance, such as dump valves, is provided.

Unique design features. Five horizontal in-pile thimbles spaced equally
on the midplane of the reactor opening into cubicles, and one vertical
nozzle, opening from the top of the reactor, are included in the design.
Figure 9-5 shows the location of the thimbles relative to the containment
vessel and cubicles, and the shield arrangement. As shown by Fig. 9-5,
piping to the heat exchangers passes through one of the hot-cell working
areas. Consequently, this area is not usable for experiments, but contains
the pressurizer and other items which must be adjacent to the reactor but
removed far enough from the reactor cell to permit maintenance.

Maintenance concept. Both dry and underwater removal methods have
been investigated for the HRR; however, both schemes present difficult
design problems. Dry-maintenance philosophy, chosen on a somewhat
arbitrary basis, has been followed in the layouts presented herein.

Maintenance of equipment in the reactor compartment is expected to
9-3] ONE-REGION U232 BURNER REACTORS 485

Experiment Thimble

Thermal Shield

  
   

Blast Shield

Reactor

“/ Equipment

Cubicle

Experimental
Facilities Cell

Fig. 9-5. Plan view of Homogeneous Research Reactor, showing pressure vessel
shielding and cells for remote handling of experiments.

be largely confined to the reactor auxiliary equipment and to the experi-
mental thimbles and equipment. The reactor vessel itself is designed for
the life of the system with thickness for the pressure-vessel wall and corro-
sion liner selected accordingly on the basis of existing corrosion data.

The handling equipment in the reactor containment vessel consists of a
revolving-type crane with a shielded cab for the operator and provision
for remote operation from outside the shielded area using commercial,
remotely operated television cameras. Access from above to any part of
the area is thus possible and all flanges and pipe disconnects are faced
upward to facilitate removal.

A horizontal traveling crane, also with a shielded control cab and remote-
operation control, is provided in each of the containment vessels for
removal of the heat-exchanger equipment. Flanges connecting the circu-
lating pumps and heat exchangers with the main piping are faced hori-
zontally in these installations. All flanges are grouped at one end of the
area and bolts are removed by means of remotely manipulated tools from a
shielded cell. The heat exchangers, mounted on wheeled dollies guided by
486 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

tracks, can be moved horizontally along the track and onto another track
section which can move transversely. From this section, the heat ex-
changer is moved through a large air-lock type of door at the end of the
containment vessel to the maintenance area. The circulating pumps are
designed so that the pump impeller and motor windings may be removed
vertically without removing the pump casing.

During any part of the maintenance procedure, the system is shut down
and drained and the piping and equipment decontaminated as thoroughly
as possible. Shutoff valves of a size and type suitable for the piping of the
HRR have not been developed.

9-3.3 The Advanced Engineering Test Reactor. A study was completed
in March 1957 by Aeronutronic Systems, Inc., to select a reactor system
for an advanced engineering test reactor (AETR), with seven major loop
facilities providing a thermal-neutron flux >2 X 10!5 neutrons/(cm?)(sec)
[6]. To obtain the required flux level while keeping the power density
low, only heavy water-moderated reactors were considered. Comparisons
of two heterogeneous and one homogeneous type, and comparison of single
and multiple reactor installations, led to the conclusion that a single homo-
geneous reactor provides the greatest flexibility and is the most economical
system for research at high neutron fluxes. A description of the homogene-
ous AETR reference design by the Aeronutronic group is given below:

Description of reactor. The 500-Mw reactor consists of a large core
operating at moderate temperature and pressure and containing a D20
solution of 109, enriched uranyl sulfate (10 g total U/liter). The reactor
design, which was based upon the design and operational experience of
the HRE-1 and HRE-2 and upon a design study for a homogeneous
research reactor by ORNL, features continuous fission-product removal
and fuel addition to maintain the total contained excess reactivity at an
essentially constant level. In the center, or loop region, the unperturbed
thermal-neutron flux is approximately 6 X 10!5 neutrons/(cm?)(sec).

The reactor vessel is a spherical, stainless steel container with an internal
diameter of 8 ft and a wall thickness of 3/4 in., contained in a cylindrical
pressure vessel with balanced pressures inside and out. The design is such
that the test loop and coolant circuit tubes emerging through the lid of the
pressure vessel can be disconnected, the packing glands at the bottom of the
pressure vessel removed, and the entire reactor core vessel can be lifted
out of the main container. The thin walls of the core vessel give it a low
gross weight, enabling it to be lifted conveniently.

The cylindrical pressure vessel, 10 ft in diameter, 123 ft high, and 3 in.
thick, 1s constructed of carbon steel to the specifications of the unfired
pressure vessel code for an internal working pressure of 500 psia.

Operating parameters of the AETR are summarized in Table 9—4.
9-4] ONE-REGION BREEDERS AND CONVERTERS 487

TABLE 94
KeYy DEsIGN PARAMETERS (AETR)

 

 

 

Type Thermal, homogeneous
Total heat power 500 Mw
Fuel Aqueous solution of UO2504 in D20
Fuel content
Core 80 kg uranium, enriched to 109,
6.5-8.5 kg U235
7500 liters fuel solution
System 375 kg uranium
37.5 kg U235
37,500 liters fuel solution
Fuel temperature: Inlet 98°C
Outlet 153°C
System pressure 500 psia
Flux (no test loops): Maximum thermal | 6 X 1015 n/(cm?2)(sec)
Power density distribution (with no
loops) : Maximum (center) 220 kw/liter
Average 70 kw/liter
Minimum (wall) 4 kw/liter

 

 

9-4. ONE-REGION BREEDERS AND CONVERTERS

0—4.1 The Pennsylvania Advanced Reactor U?33-thorium oxide refer-
ence design. The Pennsylvania Power and Light Company and the
Westinghouse Electric Corporation joined forces in November 1954 to
survey various reactor types for power generation. The results of the survey
indicated the potential of the aqueous homogeneous reactor to be exceed-
ingly encouraging and led to the formal establishment of the Pennsylvania
Advanced Reactor Project in August 1955 to study the technical and eco-
nomic feasibility of a large aqueous homogeneous reactor plant for central
service application having an electrical output of at least 150,000 kw.

Two reactor plant reference designs were completed, and preliminary
equipment layouts and cost estimates of these two plants were prepared
[7,8]. In the first design it was proposed to use overhead dry maintenance
with the equipment housed in a vertical cylinder 124 ft in diameter and
175 ft long. By incorporating shutoff valves in the system, any one of the
four main coolant loops could be isolated in case of an equipment failure to
permit the remainder of the plant to continue operation. At a convenient
time, the plant would be shut down and the defective item removed and
replaced with remote equipment such as heavy-duty manipulators, special
488 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

Equipment Cab

 
   
 

 

Personnel Access

Corridor Shell

   
   

.| Personnel Access

Primary Door

   
        
  

\
v
'y

 

.7

 

 
     
   

        
 
  
 
  

     

 

S;Jrge T =
Ilieocfor
Pressurizer 0.0 Equipment Cells
: O
O OO
O OO
OOO
o% Lock
O
O Access to

Maintenance
Building

Personnel Access
Corridor

Shielding Window

  
 
   

Access

  
 

Turbine
Steam Line

 

F1g. 9-6. Plan view of Pennsylvania Advanced Reactor Reference Design No.
1A (courtesy of Westinghouse Electric Corp).

Jigs and fixtures, and television viewing equipment lowered into the com-
partment. However, it was concluded that such a scheme would be ex-
tremely expensive. Therefore, a new design (Reference Design 1A) was
prepared based on the specifications embodied in the following recom-
mendations:

(1) Elimination of stop valves in each loop and abandonment of the idea
of partial plant operation.

(2) Compartmentalization of equipment depending on type and level of
radioactivity.

(3) Use of semidirect maintenance techniques wherever possible.

(4) Modification of the vapor container design to permit personnel
access In limited areas during plant operation. |

(5) Increased emphasis on design of components to minimize difficulty
of maintenance. |

Figures 9-6 and 9-7 show a plan and cross-sectional elevation of Refer-
ence Design 1A. In this design, a mixed-oxide slurry of a concentration of
about 260 g/kg of D20, corresponding to a solids concentration of approxi-
mately 3% by volume, is circulated through the reactor vessel releasing
550,000 kw of thermal power, which in turn yields 150,000 kwE. Leaving
the reactor vessel, the slurry branches into four parallel identical loops.
9] ONE-REGION BREEDERS AND CONVERTERS 489

 

 

 

 

100 Ton Crane
Surge Tank
Reactor
. i /N N %fi Steel Shell
Shielded Equipment Cab. /= S
\ Steam To
Shielded Crane Cab Torbine

 

 

 

] Steam Generator
Primary Pump

Gas Separator

Personnel
3 Access
Corridor

 

'<,~2:
sl

 

 

 

TN Y

 

 

K-» -
kvt

 

 

 

 

 

 

 

 

 

 

 

 

| E:‘:’.‘«. . > < . . -"’-i-:fq ]
Aty ke, e RS0 D X

F1a. 9-7. Cross section through main loops of proposed Pennsylvania Advanced
Reactor (courtesy of Westinghouse Electric Corp.).

Each loop contains a circulating pump, a gas separator, and a steam gen-
erator. The system is pressurized with 2000 psia steam generated in a D20
steam generator connected to a surge chamber mounted in close-coupled
position to the reactor vessel. The major portion of radiolytic gases is
recombined internally; the remainder (~109,) is left unrecombined in
order to purge the system of xenon and other gaseous fission products.
These gases are removed from the main stream by a pipeline gas separator
to a catalytic-type recombiner. The recombined heavy water is used to
wash the primary-pump bearings and as makeup water to the steam
pressurizer.

A small bleed stream is concentrated in the slurry letdown system and
delivered to a chemical processing plant where the uranium and thorium
are recovered by a thorex solvent extraction process. The chemical plant
is designed for a small throughput and low over-all decontamination factors.
Although the rates of flow to the auxiliary systems are small compared
with the 18,000,000 lb/hr rate of circulation in the primary system, these
auxiliary systems contribute the major part of the complexity of the plant
and a large fraction of its cost.

The reactor plant layout shown in Figs. 9-6 and 9-7 consists essentially
of a horizontal steel cylinder 125 ft in diameter and 132 ft long with 7-ft-
thick biological shielding walls completely separate from the vapor con-
490 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

 

 

 

 

 

 

 

 
    
 
 
 
 
    

 

 

 

 

 

 

 

 

’———1/ v ' 4€< R
3 ,
*‘f‘:.‘é. A v ML - .;:‘ Pr|m° ry Loop
SRSty L Tt eaAr s I
A 4 Compartment
a2 v-
e .3
> T
3
———-1 X .
—_———— -] Mechanical Arm
<
P ‘l
L o s S it
_________ oo . Y
"% Retractable Rail -7 :l
————————— i — - |
Work - 5Ty
Area B
Heavy <+—— Master-Slave
: Manipulator
5 ".".'-:'::‘ 2D
TR 2 Purge
L . g ) Water
~ | Shielding  Vent Line
Window .
Control Cooling
Console Q Water
For
Mechanical E =
Arm —» 5§
LR
r::
N
—Closed Position
- . Primary Pump —|
;:'3‘;".‘- PR "::4: PREVIEEN AT
Y Y

 

 

 

 

 

 

Pischarge
Suction o) J

Fic. 9-8. Primary circulating pump maintenance, Pennsylvania Advanced
Reactor (courtesy of Westinghouse Electric Corp.).

tainer. The reactor vessel is shielded separately; however, the four primary
coolant loops are contained in one large compartment with no shielding
between the separate loops. Auxiliary equipment is contained in separate
compartments, the equipment being segregated according to the type and
level of radioactivity after shutdown. All four of the primary coolant loops
are designed with polar symmetry to permit any component to be used as a
replacement part in any of the four loops, and any special equipment re-
quired to be equally adaptable to all four loops. In addition, like pieces of
equipment have been grouped to permit the use of relatively permanent
maintenance facilities designed into that particular area. Personnel
access corridors are provided to permit limited access to certain areas inside
of the vapor container during full power operation of the reactor.

Dry-maintenance operations are accomplished primarily through the
use of a 100-ton, shielded-cab crane which traverses the length of the reactor
container. Since the cab can be occupied during operation, the crane
serves as a remote tool for handling heavy shield blocks and removing and
replacing equipment. The design is based on an all-welded piping system
and removal of any item requires a remote cutting and welding machine
not yet developed.
9-4] ONE-REGION BREEDERS AND CONVERTERS 491

/Steam Outlet

 

 

 

 

 

 

 

-
=

LL— Steam Separator

 

[{— Drain

‘.\

 

 

 

Liquid Level
Connections

 

 

 

  

 

 

 

Feedwater Nozzle

 

 

 

 

 

 

o

/U Tubes

 

 

!
J@'

D&
\S
'T_(T—/‘\__HN

 

 

 

 

 

 

Shell Drain

Slurry Qutlet
Slurry Inlet

 

 

 

 

 

 

 

 

 

 

 

 

 

F1g. 9-9. Steam generator for Pennsylvania Advanced Reactor (courtesy of
Westinghouse Electric Corp.).

Because of the vulnerability of the circulating pumps and steam gener-
ators, special modifications are provided to permit these items to be re-
paired in place. A maintenance facility for repair of the primary circulating
pumps is shown in Fig. 9-8. This consists of two mechanical master-slave
manipulators inserted through the shielding wall adjacent to the pump,
and a mechanical arm which may be placed on two retractable rails canti-
levered from the shielding wall. Visibility is obtained by a glass shielding
492 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

      
 
    
 
 
 
 
    
   

Steam
Generator

Periscope

Tube Bundle

Shielding Window

Leveling

Periscope

Manipulator

Fic. 9-10. Facility for remote maintenance of Pennsylvania Advanced Reactor
steam generator (courtesy of Westinghouse Electric Corp.).

window located beneath the manipulators. The window is designed to be
an effective shield only during plant shutdown, and will be covered by
iron shutters during plant operation to provide neutron and thermal
shielding. A second shielding wall is located behind the work area to make
up for the thin wall at this point.

The pump is provided with flanged joints with bolts and all other con-
nections at the top for easy accessibility. The low-pressure cooling water
connections are easily disconnected with the manipulators. The high-
pressure purge line and vent lines, however, must be disconnected through
flanges or by cutting and rewelding. The large flange bolts on the pump
are provided with centrally drilled holes into which electric resistance
heaters can be inserted with the mechanical master-slave manipulator.
The heated bolts are easily loosened with a power-driven wrench held by
the mechanical arm and removed with the master-slave manipulator.
A lifting fixture is then lowered from the overhead crane and attached to
9-4] ONE-REGION BREEDERS AND CONVERTERS 493

the pump flange and pump internals, which are then pulled from the pump
volute casing. The pump is reinstalled in reverse order.

The steam generator shown in Fig. 9-9 uses inverted vertical U-tubes
and has an Integral steam separator. The unit is 6 ft in diameter and
has an over-all length of about 40 ft. Because of the physical size and cost,
it is not considered practical to use the spare-part replacement philosophy
for this component. Instead, the design of the steam generator and the
over-all plant layout is such that remote maintenance in place is possible
without requiring a prohibitively long shutdown of the plant.

Figure 9-10 illustrates the proposed semiremote method for locating and
repairing a leaky boiler tube.

The facility consists of a manipulator unit mounted on a horizontal
rack which drives the unit through the shielding wall into access holes in
the steam generator head. The manipulator is used to carry and position
a detector for locating a leaky tube and the necessary tools for plugging
and welding the tube. The faulty tube is prepared for welding by a spe-
cially designed grinding machine positioned and supported by the manipu-
lator. The grinder will automatically shape the tube for a plug and the
seal weld which will be made with an automatic welder. This equipment
can be moved from one cell to another as needed; thus all four steam
generator tube sheets can be maintained by semiremote methods.

9-4.2 Large-scale aqueous plutonium-power reactors. Studies of the
feasibility and economics of producing plutonium in homogeneous reactors
fueled with slightly enriched uranium as UO2SO4 in D2O were carried out
by the Oak Ridge National Laboratory [9-11], by the Argonne National
Laboratory [12-13], and by others [14-15]. The studies were all based
on one-region converters constructed of stainless steel utilizing spherical
pressure vessels ranging in size from 15 to 24 ft in diameter. The design
and operating characteristics of typical reactors considered in the studies
are summarized in Table 9-5.

The general conclusion reached was that aqueous homogeneous reactors
are potentially very low-cost plutonium producers; however, considerable
development work remains before large-scale reactors can be constructed.
The major problem is due to the corrosiveness of the relatively concentrated
uranyl sulfate solutions used in such reactors, which requires that all the
equipment in contact with high-temperature fuel be made of titanium, or
carbon-steel lined, or clad with titanium. The development of suitably
strong titanium alloys, bonding methods, or satisfactory steel-titanium
joints has not yet proceeded sufficiently to consider the construction of
full-scale plutonium producers. Alternate approaches, such as the addition
of Li3804 to reduce the corrosiveness of stainless steel by the fuel solution
(see Chap. 5), show promise but also require further development.
LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

494

A,00T pojeayrodng,

 

 

 

CI¥ 09% 2€3 $63 G81 923 HMY/§ ‘59800 JTuf)

88 L6 Q0T L9 .8 2s [8301,

68 A 09 I¢e 18 CT quB[d 10}BI9USSOQINT,

6% 0g G 9¢ 09 L8 querd 10J08Y
(suorf[rux §) ®yBp 9800 [83dEN)
002/28¢ CLT/0LE 003/08¢ 012/98¢ 002/08¢ 003/ x08¥ vsd/ g, ‘SUOIIPUOD WEN)T
0001 0001 0001 0001 0001 0001 1sd ‘aanssard ws£Q
053 053 093 0Sg 053 0S% 0, ‘eInjeradurdy 91xe pmbry
802 012 002 003 903 803 0, ‘@anjerodwoey jaut prnbry
0T eIl G0'T 60T G0'1 260 AMIN/3S ‘@3er uorjonpord wnruoinig
091 e 03T G0T 08T 803 §U0} d11jeW ‘A103URAUL O2(T
8¢S ee GQ Ge 0¢ 7% SU0} O1IJ9W ‘AIOJUSAUIL [aN ]
Z'1 gl'1 801 GL0°1 G0'1 08°0 cez) % ‘YUOWIYDILIUS [oNJ [BIJIUT
06 093 052 053 083 QIT 19911/ () 8 ‘uor}BIIUSOUOD [AN]
— L — . Gy e'c Ul ‘SSOUYOIY} [9SSOA QINSSAIJ
¢l CT CT CT CT ¥ 1} ‘1999WIRBIP 8109 10308y
113 113 Cey 8G% 0L¥ 0€2 AN ‘“mdino 9113099 JoN
7901 8301 0002 8201 0002 - 0001 (IBwIoy)) AN “Tomod
2961 ABIN 2G61 93 GG61 ge61 ~sny 1S61 9o 0961 120 e

- BlR( A9y L—-8-28
T0TT-SddD | 1687~ INV BOTTINYO | A0 INHO 960T~INYO | $S8—~INYO

 

 

 

 

 

 

 

B)B(J JO 92IN0g

 

 

SYAONA0Y J WAINOLATJ SN0ENADY FTVOS-EDUV] J0 SOILSIMILOVIVH))

G—6 AIAV],

 
9-4] ONE-REGION BREEDERS AND CONVERTERS 495

9-4.3 Oak Ridge National Laboratory one-region power reactor studies.
Preliminary designs of intermediate and large-scale one-region reactors
have been carried out at the Oak Ridge National Laboratory for the pur-
pose of establishing the desirability, relative to two-region reactors, of such
plants for producing power [16,17]. A description of the design of a typical
large-scale plant with a capacity of approximately 316 net Mw of electricity
follows. |

The uranium-plutonium or thorium-uranium fuel is pumped at 130,000
gpm through a 15- to 20-ft-diameter core, where the temperature is in-
creased from 213 to 250°C. Slurry leaving the core flows through four large
gas separators, where D2 and O3 are separated and diluted with helium, Os,
and D20 vapor, and then to eight 160-Mw heat exchangers. The slurry is
cooled in the exchangers and returned to the reactor by eight 16,000 gpm,
canned-motor pumps.

Gas and entrained liquid from the separators pass through four parallel
circuits into high-pressure storage tanks, where the entrained liquid is
removed to be returned to the reactor. The D3 and Os are recombined on a
platinized alumina catalyst and cooled in 17 Mw, tubular heat exchangers
which condense the 76 gpm of excess D2O. The cooled gases are recirculated
to the gas separators, and the condensate returns to the fuel through the
rotor cavities of the pumps, the demisters, and the high-pressure storage
tanks. .

The slurry fuel is expected to contain 100 to 300 g/liter of uranium as
either oxide or phosphate, and thorium as either oxide or hydroxide sus-
pended in D20O. Estimates of gas generation rates have been based on the
use of UO;3 platelet particles 1 micron thick and approximately 1 to 5
microns on a side. The Gp,o value was taken as 1.3 molecules of D>O
disintegrated per 100 ev of energy dissipated in the slurry, postulating
that 80% of the fission fragments escape from the oxide particles. It is
possible that much lower G-values will be obtained in representative experi-
ments and that the size of the gas system can thereby be reduced con-
siderably.

A 15-ft-diameter sphere operated at 1000 psi and 250°C requires a
43-in.-thick wall to keep the combined pressure and thermal stress
below 15,000 psi. Carbon steel, clad with stainless steel, is specified as the
material of construction for the vessel. The thermal shield may be stainless
steel or stainless-clad carbon steel, depending on which would be the less
costly. The weight of the vessel and thermal shield is 150 tons, while 75
tons of slurry containing 200 g U/liter are required to fill the vessel.

The estimated cost of the 316-Mw plant was $14-19 million for the
reactor portion and $44 million for the power plant section, which cor-
responds to a unit cost of $185-200/kwE.
496 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

"TABLE 9-6

OPERATING ConDITIONS—180 Mw ELECTRICAL PLANT

 

 

 

 

 

 

 

 

 

 

Core system Blanket system
General
Thermal power, Mw 360 280
Fluid U02504-D20 sol. ThO2-D20 disp.
Concentration, g/liter U233 1.80 8.00
Th232 — 1000.00
Primary system—pressure, psia 1800 1800
Reactor inlet temperature, °C 258 258
Reactor outlet temperature, °C 300 300
System volume, liters 28,760 41,785
Maximum fluid velocity, fps 33.6 28.7
Loop head loss, psig 58 80
Fuel
Total fuel in system, kg, U233 51.8 334.3
Th232 — 41,785
Fuel burnup, g/day,
U233 447 332
Th232 (consumption) — 828
Fuel removed
grams U233/day 210 498
kilograms thorium/day — 62.0
liters/day 117 62.0
Primary circulating pumps
Number 3 _ 3
Capacity, gpm 11,300 9600
Differential pressure, psi 58 80
Estimated efficiency, % 60 60
Horsepower 650 750
(Continued)

9-5. Two-REGION BREEDERS

9-5.1 Nuclear Power Group aqueous homogeneous reactor. A study of
power stations ranging in size from 94 to 1080 megawatts of net electrical
generating capacity was carried out by the Nuclear Power Group [18].
The plants considered utilized a two-region Th-U233 reactor. While
several plants of different electrical capacities were studied, emphasis was
9-5] TWO-REGION BREEDERS 497

TaBLE 9-6 (Continued)

 

 

 

 

 

 

 

 

Core system Blanket system
Steam generators
Number of units ' 3 3
Surface sq. ft./unit 14,800 12,650
Feedwater inlet temperature, °F 405 405
Steam temperature, °F 480 480
Steam pressure, psia 566 566
Thermal capacity/unit, Mw 119 | 93
Gas condenser
Type: Horizontal, straight-tube,
single-pass, shell-and-tube ex-
changers with internal elimi-
nator
Number of units 1 1
Surface area, ft2 800 800
Feedwater inlet temperature, °F 405 405
Steam temperature, °F 480 480
Steam pressure, psia 566 566
Thermal capacity/unit, Mw 3.5 3.5

 

directed toward a plant having a net electrical capacity of 180 MwE.
Pertinent operating conditions of this plant are listed in Table 9-6.

The reactor consists of a 6-ft-diameter spherical core surrounded by a
2-ft-thick blanket enclosed in an 11 ft 4 in. ID stainless-clad carbon steel
pressure vessel with a wall thickness of approximately 6 in. The pressure
vessel has a bolted head to permit removal of the concentric-flow core tank
if necessary. The fuel solution enters the core through a 24-in. inner pipe
and exits through an annulus of equivalent area between two concentric
pipes forming the inlet and outlet connections for the core tank. One
mechanical joint is required to attach the zirconium core tank to the
stainless steel outlet pipe. The slurry enters the blanket through a 24-in.
connection in the bottom of the pressure vessel and exits through three
14-1n. connections located near the top of the vessel.

Thermal shield. The 4-in.-thick thermal shield to protect the pressure
vessel from excessive radiation is provided in the form of two 2-in.-thick
stainless steel plates. A 2-in. space is maintained between these plates and
between the thermal shield and the pressure vessel. Sufficient flow of the
498 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

slurry is maintained between and around the shield segments to ensure
proper cooling.

Vessel closure. The bolted head closure utilizes two Flexitallic gaskets
(asbestos encased in stainless steel), having a low-pressure leakoff between
gaskets. The internal diameter of the closure is slightly greater than 6 ft,
to allow for the core tank removal. Similar bolted joints are provided 1n
the inlet and outlet piping connections to the core as well as in the core
tank dump line.

Steam generators. Six steam generator units each consisting of two heat
exchangers connected to a common steam drum are required, three for the
core system and three for the blanket heat removal. By using U-bend
tubes in the exchangers, the need for an expansion joint in the shell or in
the connection to a floating tube sheet is eliminated. This adds reliability
to the unit, since any expansion joint subject to even infrequent work 1s a
potential and likely source of trouble.

Utilizing the compartmentalized concept in the heat exchangers offers
added reliability, ease of fabrication, and a means by which maintenance
of the units becomes practical. The individual “‘bottles,” consisting of 19
U-tubes attached to their tube sheets in the eccentric pipe reducers by
rolling and welding, can be fabricated and tested as units before installation
in the exchanger. The drilling of the “bottle’”’ tube sheets presents prac-
tically no difficulty because they are only 5§ in. in diameter. Similarly, the
drilling of the exchanger head for insertion of the 2-in. inlet and outlet
pipes to the “‘bottles’” presents no unusual fabrication problems.

Although the goal is theoretically ‘‘leakproof’” heat exchangers, pro-
visions are incorporated for maintenance. This has been done in the
compartmentalized concept. Should a leak occur, it is practical to seal off
the “bottle’” in which the leak occurs by plugging the 2-in. inlet and outlet
pipe connections, rather than remove an entire heat exchanger.

Primary circulating pumps. The hermetically-sealed-motor, centrifugal
pumps required to recirculate the core and blanket fluids are vertical, with
the main impeller mounted on the lower end of a shaft on which also is
mounted the motor rotor. The motor rotor and bearing chamber are
separated from the impeller and volute by means of a labyrinth seal.
D50 from the high-pressure condensate tank is injected into the bearing
chamber and continuously flushed through the labyrinth, thus minimizing
corrosion on the rotor and bearing parts. Both the radial and thrust
bearings are of the fluid piston type. The drive motors are induction type,
suitable for 3-phase 60-cycle 4160-volt power supply. They have im-
pervious liners in the stator bore for hermetic sealing, and an outer housing
totally enclosing the stator as a second safeguard against loss of system
fluid. The motor stator windings are cooled by a liquid passing through
tubular conductors installed in the stator. All material in contact with the
9-5] | TWO-REGION BREEDERS 499

core solution and D3O is stainless steel, except for the impeller, labyrinth
inserts, impeller nut, and wear rings, which are titanium.

By unbolting the top flange, the entire pump mechanism can be re-
moved, leaving only the high-pressure pump casing in the pipeline, thus
facilitating maintenance.

Shielding and containment. The reactor plant is housed in a 175-ft-
diameter steel sphere. Design pressure is 40 psia, which requires a nominal
plate thickness of 3/4 in. The sphere is buried to a depth of 50 ft, allowing
the reactor vessel to be located below grade for natural ground shielding.

Radioactive components are enclosed by a barytes concrete structure
which serves both as a biological and as a blast shield. The top of this
housing is 35 ft above grade elevation. The side walls are 5 ft thick, and
the top shield is 6 ft thick except for an 8-ft-thick section directly over the
reactor vessel. Compartment walls are provided within the housing to
facilitate flooding of individual component sections. The floor and side
walls of each of the compartments are lined with 1/8-in. stainless steel
plate to permit decontamination. Stepped plugs are provided in the top
shield to permit access to the components. The portion of the shielding
around the reactor vessel, which is below grade, is 4 ft thick. A slight
negative pressure is maintained within the container by continuously dis-
charging a small quantity of air to a stack for dispersal. The quantity of
air removed is regulated to control the ambient temperature in the com-
ponent compartments.

Cost analysis

This study indicates that a generating station with a net thermal effi-
ciency of 28.19, might be constructed for approximately $240.00/kw and
$200.00/kw at the 180-Mw and 1080-Mw electrical levels, respectively.
These values result in capital expenses of approximately 4.72 and 3.86
mills/kwh.

9-5.2 Single-fluid two-region aqueous homogeneous reactor power
plant. The feasibility of a 150,000-kw (electrical) aqueous homogeneous
nuclear power plant has been investigated by a joint study team of the
Nuclear Power Group and The Babcock & Wilcox Company [19]. In this
concept, the reactor is a single-fluid two-region design in which the fuel
solution circulates through the thoria pellet blanket as the coolant. Com-
ponents and plant arrangement have been designed to provide maximum
overhead accessibility for maintenance. All components in contact with
reactor fuel at high pressure are themselves enclosed in close-fitting high-
pressure containment envelopes.

General description and operation of plant. The reactor generates 620-psia
steam at the rate of 2.13 X 106 1b/hr.
500 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

The reactor system is contained in a building 196 ft long, 131 ft wide,
and 50 ft high. The equipment is located in a group of gastight cells meas-
uring 196 ft X 131 ft over-all. These cells are equipped with pressure-tight
concrete lids to facilitate overhead maintenance of the various system
components and minimize the radiation shielding required above the floor
and outside of the building. All components in the reactor plant have been
designed in accordance with this overhead maintenance philosophy.

The basic systems comprising the reactor plant are: (1) A primary sys-
tem, containing the reactor, the main coolant loops, the boiler heat ex-
changers, the pressurizer, the surge tank, and the standby cooler; (2) the
Jetdown system; (3) the fuel handling and storage system; (4) the off-gas
system; and (5) the auxiliary systems containing leak-detection and fuel-
sampling facilities.

The blanket consists of 14 cylindrical assemblies arranged around the
periphery of the core region. These assemblies contain thorium-oxide
pellet beds which are cooled by fuel flowing from a ring header below the
reactor vessel. The fuel follows a zigzag path through the pellets and
leaves the assemblies through top outlets, flows through the core region,
and out the bottom of the vessel. By this means, the usual core-tank cor-
rosion and replacement problems and slurry handling problems are mini-
mized. By means of devices located at the tops of the tubes extending out
of the reactor the assemblies are periodically rotated to minimize absorp-
tion of neutrons by protactinium and equalize the buildup of U233 in the
thorium. Replacement of the assemblies is possible through a smaller clo-
sure than would be required for a two-region reactor with a single core tank.

The reactor vessel is surrounded by a high-pressure containment vessel
which forms part of the containment system described below. Over-all
height of the reactor is 28 ft 6 in. and the outer diameter of the containment
shell is 12 ft 1§ in.

The boiler heat exchangers are designed so that by removing the head
there is direct access for plugging tubes or for removing entire tube bundles
if necessary. These exchangers are a once-through type designed to evapo-
rate 95%, of the feedwater flow at full load. The reactor fuel flows counter-
currently through the shell side of the exchangers. Feedwater enters the
baffled heads and passes through the U-tubes where the steam is generated.
The steam-water mixture then leaves the exchangers and flows through a
cyclone separator and scrubber to the turbine. These components, with
their containment, are 34 ft 2% in. high and 4 ft 37 in. OD.

All piping and components holding high-pressure reactor fuel are con-
tained in a close-fitting pressurized envelope capable of withstanding the
total system pressure. These components and piping are further contained
in pressure-tight concrete cells which are vented through rupture disks to a
low-pressure gas holder, as shown in Fig. 9-11. This holder has a liquid-
9-5] TWO-REGION BREEDERS 501

Low Pressure
Heat Exchanger Reactor Gas Holder

—_——

 

 

 

 

 

 

 

 

 

A

3
£

 

 

 

o

1 //

 

 

Relief Pipe

F1a. 9-11. Schematic illustration of containment system (courtesy of the Nu-
clear Power Group and the Babcock & Wilcox Co.). |

sealed roof which moves up and down in a manner similar to the movement
of a conventional gas holder section. The low-pressure components do not
have a close-fitting high-pressure envelope but are contained in pressure-
tight cells and are vented to the low-pressure gas holders in a manner similar
to the high-pressure components.

This type of containment permits operation of components for their full
service life, reduces or eliminates missile formation and fuel losses, reduces
primary system working stresses, and allows equipment arrangement
giving maximum access for maintenance.

Table 9-7 summarizes the characteristics of the proposed plant.

Reactor. The general characteristics of the reactor are illustrated by
Fig. 9-12, which shows the annular arrangement of the Zircaloy-2 blanket
assemblies around the core region. The thorium-oxide pellets within these
assemblies are cooled by the reactor fuel solution, which is pumped up
through the packed beds from the supply header. To reduce the pressure
drop across the pebble bed, the solution is introduced through a tapered
perforated pipe the same length as the assemblies, flows into the bed and
by means of baffles is directed back to the center outlet pipe, which is con-
centric with the inlet. _

The vessel has ellipsoidal heads, is 9 ft 11 in. ID, and has a cylindrical
shell length of 10 ft 6 in. The upper head contains a 3-ft 33-in.-diameter
502 LARGE-SCALE HOMOGENEOUS REACTOR

TABLE 9-7

STUDIES

[cHAP. 9

DEsicN DATA FOR THE SINGLE-FLUID Two-REGioNn REACTOR

 

Over-all plant performance
Thermal power developed in reactor, Mw
Gross electrical power, Mw
Net electrical power, Mw
Station efficiency, 9,

General reactor data
Fuel solution
Operating pressure, psia
Fuel inlet temperature, °F
Fuel outlet temperature, °F

Total volume of primary system, ft3
Area of Zircaloy—2 (in contact with fuel solution), ft2

Core
Fuel flow rate, Ib/hr
Velocity, fps
Volume of core solution, liters
Letdown rate, gpm
Thorex cycle time, days
Hydroclone cycle time, days
Hydroclone underflow removal rate, liters/day

Blanket
Assembly diameter, in.
Fertile material
Thorium loading, kg
Thorium irradiation cycle, days
Thorium processing rate, kg/day
Processing rate of mass-233 elements, g/day

 

Area of stainless steel (in contact with fuel solution), ft2

 

520
158
150

28

U02504-D20

1500
514
572

90,000
2,300
2,240

24 .9 X 108

6
67,000
100
115

1

583

18

ThO2 pellets

17,850
744
23
350

.9

.3

 

 

flanged opening to permit removal of the thorium assemblies. The closure
is a double-gasketed bolted cover having a monitoring or buffer seal con-
nection to the annulus between the gaskets to detect leakage of the fuel
solution. The lower pressure vessel head is penetrated by fourteen 7-in.
openings through which fuel flows upward into the blanket assemblies from
the toroidal supply header. This head has a 3-ft-diameter fuel outlet. The
thermal shielding consists of alternate layers of stainless steel and fuel
solution. The total shielding thickness is 8 in. over the cylindrical portion
and 12 in. at the head ends of the reactor pressure vessel.

Boiler heat exchangers. The boiler heat exchanger is a U-tube, vertical,
9-5] TWO-REGION BREEDERS 503

Double-Gasketed
Bolted Closures

 

Thorium

Oxide Rotation Shaft
Pellets

 

 

Blanket
Assemblies

 

l

Containment
e Vessel

Core Region Reactor

Vessel

 

IOSOOOVOENNANNNN

Thermal
Shielding

 

SO N NN
NN

 

 

Fuel Solution
Supply
Header

 

SOSOOUNNN NN N
SO

 

 

 

N

—

Fuel | Fuel Solution
lif—
Solution Outlet Inlet
i tfre——

Manway

 

 

Fis. 9-12. Single-fluid two-region aqueous reactor (courtesy of the Nuclear
Power Group and the Babcock & Wilcox Co.).

forced-circulation design in which reactor fuel flows on the shell side and
boiling light water on the tube side.

By placing the reactor fuel on the shell side, the tube sheet acts as its own
shield and is subjected to less intense nuclear radiation, minimizing gamma
heating and thermal-stress problems. Such a design permits the use of
thermal shields which would also serve to protect the tube sheet from
thermal shock due to sudden variations of fuel temperatures. In the event
of a failure of tubes or tube sheet connections it is necessary to remove only
the faulty tube bundle and leave the exchanger shell and flanged con-
nections intact. This is accomplished by removing the bolted head and
tube sheet brace and cutting the seal ring weld at the periphery of the
tube sheet. The bundle is then lifted out of the shell by the overhead crane.
o004 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

Reactor building. 'The building which houses the reactor plant will be
airtight and will serve as a containment for radioactive gases which may
be released during maintenance. The building air will be monitored and
filtered and will be vented to the exhaust stack.

Space is provided outside the reactor plant for an off-gas building stack,
gas and vapor holders, gas handling building, hot laboratory and shops,
waste handling building, and chemical processing buildings. The hot shops
and chemical processing building are located as shown to permit mutual
access to a crane bay which extends from the reactor building between the
two buildings. Both “hot” components and blanket assemblies are trans-
ported from the reactor building to the far end of the bay by a low, U-frame
traveling crane. At the end of the bay they are transferred to an overhead
crane running perpendicular to the bay, and transported to either of the
two buildings. With this arrangement, hot materials may be transferred
entirely underwater, thereby eliminating the need for bulky shielding and
mobile cooling systems.

M aintenance considerations. A study of the problem of maintenance of
a large-scale homogeneous reactor indicated the following. It appears im-
possible to accomplish some repair operations remotely in place and under
20 ft of water. Experience to date tends to indicate that the repair of
radioactive equipment may be so difficult that it will be uneconomical to
repalr anything except such small components as valves and pumps. The
larger defective components must be removed from the system and a re-
placement installed. The repairs, if possible, can then be made in a ‘‘hot
machine shop’ after the system is back in operation.

Removal of components from the cells will require shielding, such as
lead easks. Further study is necessary to determine the optimum means
of performing this operation. |

Extensive use of jigs and fixtures in performing maintenance work will
be necessary for rapid and safe work. All the jigs and fixtures should be
designed and constructed before the plant is put into operation. In many
cases 1t will be advantageous to use the jigs during initial construction to
be certain that they will function properly.

The estimated annual maintenance cost for a plant of the size considered
i1s approximately $3,300,000, which includes the capital investment of
maintenance equipment. This amounts to about 4 mills/kwh at 60%
capacity factor or ~3 mills/kwh at 80% capacity factor. The 150,000-kw
nuclear power plant described is estimated to cost $375.00/kw or
$56,400,000, excluding $5,000,000 to $10,000,000 for research and de-
velopment.

9-5.3 Oak Ridge National Laboratory two-region reactor studies. In-
termeduate-scale homogeneous reactor. In October 1952 design studies were
9-5] TWO-REGION BREEDERS 505

 
       

  
 

Chemical Plant Reactor Entrance Turbine Plant Transformer
] Containers Hatch ' Yard
& n\‘ s’ ———x / "" 4;_‘}.‘ —_
N ‘ ‘ ‘:"/ ’ /J!/ .
~ L AN
| | T e ""«i!#'-.‘»y“
_—

Fic. 9-13. Artist’s concept of Thorium Breeder Reactor Power Station.

started for a two-region multipurpose intermediate-scale homogeneous re-
actor as an alternate to the single-region reactors previously studied
[20-22]. Several suggested core vessel arrangements for two-region con-
verters were presented. In all designs the core shape approximates a
4-ft-diameter sphere, and a central thimble is incorporated to permit startup
and shutdown with a full core containing the operating concentration of
fuel. The major interest in the design of this type of reactor is the possi-
bility of converting thorium into U233 in the blanket region of the reactor.

The principal system parameters on which the design of the two-region
intermediate-scale homogeneous reactor is based are presented in Table 9-8.

Large-scale conceptual designs. Design work on the two-region inter-
mediate-scale homogeneous reactor continued through the fall of 1953, with
emphasis being placed on design studies of components and reactor layouts
for an optimum design. In the meantime, conceptual designs of large-scale
two-region reactors described below were carried out as a basis of feasibility
studies.

The first design involved a 1350-Mw (heat) power plant containing three
reactors [17]. Each of these operated at 450 Mw to produce a net of
105 Mw of electricity. The design of this plant, which is reviewed in the
following paragraphs, is representative of the technology as of Septem-
ber 1953.
506

LARGE-SCALE HOMOGENEOUS REACTOR STUDIES

TABLE 9-8

[cHAP. 9

DEsiGN PARAMETERS oF A Two-REGIoN
INTERMEDIATE ScArLE HoMoGENEOUS REACTOR

 

 

 

Core system Blanket system

Power level, Mw 48 9.6
Fluid U02504-D20 sol ThO2-D20 slurry
Concentration, g U/liter 4.8 1000
System pressure, psia 1000 1000
System temperature, °C 250 250
Vessel diameter, ft 4 8
Maximum fluid velocity, fps 22.3 12.3
Pumping requirements, gpm 5000 1000
Steam pressure, psia 215 215
Steam temperature, °F 388 388
Steam generated, 1b/sec 38.7 8.0

 

 

 

 

 

In the proposed arrangement, three large cells are provided for the re-
actors and associated high-pressure equipment and a fourth is provided
for the dump tanks and low-pressure equipment. Each reactor cell is di-
vided into compartments for the reactor, heat exchanger, pumps, and gas-
circulating systems. The low-pressure equipment cell contains compart-
ments for dump tanks, feed equipment, heat and fission-product removal,
D0 recovery, and the limited amount of chemical processing that can be
included in the reactor circulating system. Radiation from the cells is re-
duced to tolerable levels under operating conditions by concrete shielding.
The individual compartments have sufficient shielding to permit limited
access when equipment is being replaced.

Each reactor consists of a 6-ft-diameter spherical core, operated at a
power of 320 Mw (100 kw/liter), surrounded by a 2-ft-thick blanket which
is operated at a power of 130 Mw (11 kw/liter). Under equilibrium con-
ditions, a solution containing 1.30 g of U233 U234 235 7236 (35 uranyl
sulfate dissolved in D20O) is circulated through the core at a rate of
30,000 gpm under a pressure of 1000 psia. Fluid enters the core at
213°C and leaves at 250°C. Decomposition of the D20 moderator by
fission fragments yields 240 c¢fm of gas containing 28 mole 9, D2, 14 mole 9,
O2, and 58 mole 9, D:O.

Liquid leaving the core divides into two parallel circuits, each at
15,000 gpm, which lead into centrifugal gas separators. There the ex-
plosive mixture of deuterium and oxygen is separated from the liquid and
9-5] TWO-REGION BREEDERS 507

diluted below the explosive limit with a recirculated gas stream which con-
tains oxygen, helium, and D20. The gas-free liquid circulates through heat
exchangers and is returned to the core by canned-motor circulating pumps.
Steam is produced in the exchangers at 215 psig and 388°F'.

The gas streams from the separators are joined and flow into a high-
pressure storage tank accompanied by about 500 gpm of entrained liquid.
After the entrainment is removed in mist separators for return to the
liquid system, the D2 and Oz are recombined when the gas passes into a
catalyst bed containing platinized alumina pellets. Heat liberated in the
recombiner increases the temperature of the gas from 250 to 464°C. The
hot gases are cooled to 250°C in a gas condenser which has a capacity of
20 Mw and condenses D20 at a rate of about 8 gpm. Some of the D20
is used to wash the mist separators and to purge the pump bearings and
rotor cavity; the remainder is either returned to the system through the
high-pressure storage tank or held in condensate storage tanks during
periods when the concentration of reactor solution is being adjusted. The
gas is recirculated to the gas separators by an oxygen blower.

Similar gas- and liquid-recirculating systems are used to remove heat
from the blanket, which consists of a thorium-oxide slurry in D20 contain-
ing 500 to 1000 g Th/liter. The slurry is recirculated by means of a
12,400—gpm canned-motor pump through a gas separator and through a
130-Mw heat exchanger.

The reactor is pressurized with a mixture of helium and oxygen which
is admitted as required. It is expected that most of the fission-product
gases will be retained in the high-pressure gas-circulating systems with
only whatever small, daily letdown is required to adjust the pressures.
Calculations for a similar system indicate that enough Xe!3% will be trans-
ferred into the gas stream to reduce the xenon poisoning in the reactor by
a factor of 5 to 10.

The two-region thortum breeder reactor. A later design study was com-
pleted in the fall of 1954 by the Reactor Experimental Engineering Divi-
sion of the Oak Ridge National Laboratory [23] for the purpose of de-
lineating the technical and economic problems which would determine the
ultimate feasibility of an aqueous homogeneous reactor for producing cen-
tral station power.

The concept of the reactor chosen for study was based essentially on
nuclear considerations and consists of a spherical two-region reactor with
dimensions limited by economic considerations.

Table 9-9 presents the principal reactor characteristics for the pre-
liminary design of a 300-Mw station.

The power plant complex, consisting of the reactor plant, the turbo-
generator plant, the chemical processing plant, the cooling system, and
part of the electrical distribution system, is shown in Fig. 9-13.
508

LARGE-SCALE HOMOGENEOUS REACTOR STUDIES

TABLE 9-9

[cHAP. 9

REACTOR CHARACTERISTICS FOR PRELIMINARY
DEsigN oF A 300-Mw STATION

Electrical capability (each of three reactors), 100 Mw
Gross station efficiency, 27.49 |
Net station efficiency, 26.09
Operating pressure, 2000 psia

 

 

 

 

Core Blanket
Material of construction Zircaloy—-2 209, stainless-steel
clad carbon steel
Wall thickness, in. 0.5 5.0
Thermal shield thickness, in. 4.0
Pipe connections Concentric Straight-through
Inside diameter, ft 5 103
Blanket thickness, in. 27
Volume, liters 1855 11,600
Operating temperature, °C
Average 275 280
Inlet 250 245
Outlet 300 315
System volume, liters 9740 13,800
Fluid composition, g/liter D20* | U02S04~D20-CuSO4 UO3ThO2-D20
U233 1.88 3.00
U234 1.90 0.17
U235 0.26 0.01
U236 3.00 0.00
Thorium 1000
Inventory, kg
D20 10,900 12,000
U235 4 U233 26.1 42.9
Thorium 14,300
Flux at core wall, n/(cm?2)(sec) 1.10 X 1015 1 X 1015
Power density at core wall,
kw/liter 70.0
Mean power density in reactor,
kw/liter 193 7.0
Power density in external system,
kw/liter 61 55
Reactor power, Mw (heat) 313 72
Circulation rate, gpm 24,000

 

 

 

 

*At operating conditions.
9-5] TWO-REGION BREEDERS 509

The reactor plant consists of a space 80 ft 0 in. wide, 300 ft 0 in. long,
and 45 ft 6 in. above ground level. At one end of the structure is located a
storage pool for items freshly removed from the shield. A gantry crane
services the reactors and its runway extends a distance beyond the shield
for access to the pool and to provide lay-down space for a reactor-shield
tank dome. Three cylindrical shield tanks are provided to contain each
reactor and components.

In the event of a line rupture or equipment failure resulting in gross
leakage of reactor fluids from the reactor system, no radioactive material
will be released. This is accomplished by placing the reactor and com-
ponents in a cylindrical tank, 66 ft 0 in. diameter X 116 ft O in. high,
capable of withstanding 50 psig. Eight feet of concrete are poured around
this tank to a height of 45 ft 6 in. above ground level for biological shielding.

‘As shown in Fig. 9-13, each of the container buildings has an access
hatch. Items such as circulating pumps, pressurizer heater elements,
evaporators, etc., for which the probability of maintenance is high, are
grouped on one side of the shield and more or less under this hatch for
servicing. No specific procedure for repairing or replacing equipment was
developed; however, consideration was given to both wet and dry main-
tenance methods.

Homogeneous reactor experiment No. 3.* In 1957, conceptual design studies
of HRE-3, a two-region homogeneous breeder reactor fueled with U233
and thorium, were initiated at the Oak Ridge National Laboratory [24].
This reactor, operating at 60 Mw of heat to produce approximately 19 Mw
of electrical energy, will be designed to provide operational and technical
data and to demonstrate the technical feasibility of an intermediate-scale
aqueous homogeneous power breeder. The power plant will be a completely
integrated facility incorporating (1) the nuclear reactor complex, (2) the
electrical generating plant, and (3) the nuclear fuel recycle processing
plant. Preliminary design criteria are given in Table 9-10.

As presently conceived, the reactor will be of the two-region type with
a heavy-water uranyl-sulfate solution being circulated through the inner
(core) region, where 50 Mw of heat are produced. A thorium-oxide slurry
will be circulated through the outer pressure-retaining (blanket) region,
~ where 10 Mw of heat are produced at equilibrium conditions. The fuel
solution and slurry will be circulated through separate steam generators by
the use of canned-motor pumps. The fuel heat exchanger will provide
195,700 1b/hr of saturated steam at 450 psia (456°F) at 50-Mw core power,
and the slurry heat exchanger will provide 39,100 lb/hr of saturated steam
at 450 psia (450°F) at 10-Mw blanket power. The blanket and core regions,
operating at 1500 psia and 275°C and 280°C average temperatures, re-

*By R. H. Chapman.
010 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

TABLE 9-10
"HRE-3 DgxsieN CRITERIA

 

Type Two-region breeder

Core UOzSO4 + CuSO4 + Dst4 n DzO

Blanket ThO2 4+ UO2 + Mo0O3 in DO

Core critical concentration at 50 Mw and | 4.8 g U233/kg D,0
equilibrium 0.45 g U?35/kg D20

Blanket concentration at 10 Mw and | 1000 g Th/kg D20
equilibrium 4.02 g U233 /kg D20

Average core temperature, °C 280

Average blanket temperature, °C 275

Average core power density for 50 Mw,
kw/liter 52.6

Average blanket power density for 10 Mw,
kw/liter 1.04

Estimated breeding ratio 1.05 (minimum)

Saturated steam pressure, psia 450

Gross electrical power, Mw ~19

 

 

 

 

spectively, are interconnected in the vapor region. Inasmuch as oxygen
i1s consumed by mechanisms of corrosion and must be added continuously,
it is currently favored as the pressurizing medium to provide the over-
pressure necessary to prevent boiling and bubble formation. Sufficient
homogeneous catalysts will be provided in the solution and slurry to re-
combine all the radiolytic gases formed in the circulating system during
operation. In this manner it will be unnecessary to operate with continuous
letdown of slurry and/or solution. Purge water for use in the high-pressure
circulating systems will be produced by condensing a portion of the steam
contained in the vapor volume of the pressurizer. Advantage is taken of
the beta and gamma decay energy to maintain the pressurizers at a slightly
higher temperature than the remaining portion of the system.

The electrical generating plant will be essentially of conventional
design. The 20-Mw turbine will operate at 1800 rpm on 450 psia saturated
steam with moisture separation equipment provided. The condensing water
requirements are 31,300 gpm, assuming the water enters at 70°F and leaves
at 80°F. The generating voltage of 13.8 kv is sufficient to permit direct
connection to an existing distribution system:.

The fuel reprocessing for HRE-3 will consist of concentrating insoluble
fission and corrosion products in the underflow pots of hydroclone sepa-
rators, recovery of uranium from the hydroclone underflow by UO4 pre-
cipitation, and recovery of D20 by evaporation. The slurry processing
operation will consist of D20 recovery by evaporation and packaging the
9-5] TWO-REGION BREEDERS oll

irradiated ThO¢ for snipment to the existing Oak Ridge National Labo-
ratory Thorex Pilot Plant, where the uranium and thorium will be sepa-
rated and reclaimed. The thorium will appear from the Thorex process as
a thorium nitrate solution and will be converted to ThO2 before being
returned to the blanket of the reactor.

The maintenance philosophy of the reactor complex has not yet been
established. However, it can be said that the reactor system will be designed
so that all components and equipment will be capable of being removed
and replaced, but with varying degrees of difficulty. The choice of under-
water, dry, or combination thereof, maintenance techniques has not been
made. The major components of HRE-3 are considered to be in the range
of sizes which might be used in.a large-scale Thorium Breeder Power Plant.
Design, development, fabrication, and operational and reliability data are
expected to be gained from HRE-3, in addition to maintenance techniques
for large-scale aqueous homogeneous reactors. A very preliminary cost
study indicates a cost of about $29,000,000 for the reactor complex, the
electrical generating plant, and the fuel reprocessing plant.
512 LARGE-SCALE HOMOGENEOUS REACTOR STUDIES [cHAP. 9

REFERENCES

1. THEODORE ROCKWELL III, Reactor Shielding Design Manual, 1st ed. New
York: D. Van Nostrand Company, Inc., 1956.

2. W. F. Tayrvor, TBR Plant Turbogenerator System Study, USAEC Report
CF-56-7-127, Oak Ridge National Laboratory, June 1956.

3. FosTER-WHEELER CoORPORATION, Wolverine Electric Cooperative Proposal,
Feb. 1, 1956, Report FW-56-004; Sargent and Lundy Estimate No. 3800-2,
November 1957. |

4. M. I. LunpIN and R. Van WiNkLE, Conceptual Design and Evaluation
Study of 10,000 KWE Aqueous Homogeneous Nuclear Power Plant, USAEC
Report CF-57-12-8, Oak Ridge National Laboratory, Dec. 11, 1957.

5. P. R. KasTEN et al., Aqueous Homogeneous Research Reactor—Feasibility
Study, USAEC Report ORNL-2256, Oak Ridge National Laboratory, Apr. 10,
1957.

6. Forp Moror CompaNy, A Selectron Study for an Advanced Engineering Test
Reactor, Document No. U-047, Aeronutronic Systems, Inc., Glendale, Calif.,
Mar. 29, 1957.

7. W. E. Jounson et al., The P.A.R. Homogeneous Reactor Project, Mech.
Eng. 79, 242-245 (1957). |

8. WESTINGHOUSE ELECTRIC CORPORATION, 1956-1957. Unpublished.

9. J. A. LaNE et al.,, Oak Ridge National Laboratory, 1950. Unpublished.

10. J. A. LaNE et al., Oak Ridge National Laboratory, 1951. Unpublished.

11. R. H. BaLL et al., Oak Ridge National Laboratory, 1952. Unpublished.

12. J. J. Katz et al., Argonne National Laboratory, 1952. Unpublished.

13. L. E. LinNk et al., Argonne National Laboratory, 1952. Unpublished.

14. H. A. OurereN and D. J. MaLLoN, Idaho Operations Office, 1952. Un-
published.

15. ComMONWEALTH EDISON CoMPANY AND PuBLIic SERVICE COMPANY OF
NoRTHERN ILLiNols, A Report on the Feastbility of Power Generation Using
Nuclear Energy, 1952. Unpublished.

16. W. E. Tuompson (Comp.), Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Mar. 15, 1952, USAEC Report ORNL-1280, Oak
Ridge National Laboratory, 1952.

17. R. B. Briaas et al., Aqueous Homogeneous Reactors for Producing Central-
station Power, USAEC Report ORNL-1642(Del.), Oak Ridge National Labora-
tory, 1954.

18. H. G. CarsoN and L. H. Lanprum (Eds.), Preliminary Design and Cost
Estimate for the Production of Central-station Power from an Aqueous Homo-
geneous Reactor Utilizing Thortum-Uranium-233, USAEC Report NPG-112,
Commonwealth Edison Company (Nuclear Power Group), Feb. 1, 1955.

19. CommoONWEALTH EpisoN CompaNy, Swngle-flurd Two-region Aqueous
Homogeneous Reactor Power Plant: Conceptual Design and Feastbility Study,
USAEC Report NPG-171, July 1957.

20. W. E. TaomprsoN (Comp.), Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending July 1, 1952, USAEC Report ORNL-1318, Oak
Ridge National Laboratory, 1952.
REFERENCES 513

21. W. E. TaompsoN (Comp.), Homogeneous Reactor Project Quarterly Progress
Report for the Pervod Ending Oct. 1, 1952, USAEC Report ORNL-1424(Del.),
Oak Ridge National Laboratory, 1953.

22. W. E. TrnompsoN (Comp.), Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Jan. 1, 19563, USAEC Report ORNL-1478(Del.),
Oak Ridge National Laboratory, 1953.

23. Oak Ridge National Laboratory, 1954. Unpublished.

24. J. C. BoLGER et al., Prelvminary H RE-3 Design Data (Revised to 11-15-57),
USAEC Report CF-57-11-74, Oak Ridge National Laboratory, Nov. 29, 1957.
CHAPTER 10
HOMOGENEOUS REACTOR COST STUDIES*

10-1. INTRODUCTION

10-1.1 Relation between cost studies and reactor design factors. The
power cost associated with a reactor station may be subdivided into fixed
charges, operating and maintenance costs, and fuel costs. Fixed charges
include interest on investment, depreciation, and taxes; labor, supervision,
and maintenance are included in the operating and maintenance costs;
fuel costs include both variable and fixed chemical processing costs,T cost
of feed materials, and inventory charges. Because of the uncertainty of
these items, it is impossible to determine absolute costs for nuclear power
until large nuclear plants have been built and operated. However, it is
important that a reasonable effort be made to evaluate the cost in order to
compare several fuel or reactor systems of equal technological development,
to point out areas where substantial improvements are required, and to
provide a basis for determining whether economical power can ever be
produced.

Aqueous homogeneous reactors have certain features, such as high neu-
tron economy and continuous fission-product removal, which make them
appear to be potential economic power producers. However, as with all
water-moderated reactors, to attain steam temperatures corresponding to
thermal efficiencies of 25 to 30%, circulating aqueous systems require
operating pressures between 1000 and 2000 psia. Since thermal efficiency
increases relatively slowly with increasing operating pressure, while reactor
costs rise relatively sharply above pressures of about 1500 psi, it; is unlikely
that reactors will be operated at pressures above 2000 psia. In addition,
increasing the reactor temperature tends to decrease the breeding ratio,
which adversely affects fuel costs. Nearly all reactor systems considered
have therefore been assumed to operate at pressures between 1500 and
2000 psia.

In order to optimize the design of a homogeneous reactor of a given
power output and pressure, it is necessary to know how both the fixed and
operating costs vary with the dimensions of the reactor core and pressure
vessel. In this regard, one must take into consideration that the maximum
diameter of the pressure vessel will be limited by fabrication problems, and
the minimum diameter of the core vessel will be limited by corrosion

*By P. R. Kasten, Oak Ridge National Laboratory.
1The fixed costs in a chemical processing plant are those due to plant investment;
variable costs are due to materials, labor, ete.

514
10-1] INTRODUCTION 515

problems. For each combination of core- and pressure-vessel diameters
within these limits, there will be a minimum fuel cost resulting from a
balance of inventory costs, processing costs, and fuel-feed costs; these latter
costs are determined by the breeding ratio, which is a function of fuel con-
centration and processing rate.

Although the fuel-fluid temperature influences power costs, this is nor-
mally limited by the properties of the fuel system or by the above-mentioned
pressure limitations, rather than by economic considerations. However,
the temperature range established on this basis is also close to that which
gives minimum fuel costs. In addition, the power level of the reactor is
usually assumed to be constant, although it is realized that this is a very
important factor influencing the cost of power, since plant investment
charges per unit power constitute a large fraction of the power cost and
change appreciably with power level. The effect of power level on capital
costs is discussed in Section 10-8, and on fuel costs in sections as noted.

The operating and maintenance costs, as well as plant investment costs,
are a function of reactor type and method of maintenance. However, the
exact form of some of the interrelations between design variables is not
known at the present time. For example, the plant investment and main-
tenance costs are undoubtedly different for a burner-type reactor than for
a breeder-type reactor; a cost difference would also exist between one- and
two-region systems. However, because of the lack of information, most
economic studies do not consider such differences, but assume investment
and maintenance charges to be determined primarily by the reactor power
level. The results of such studies are still significant if they are considered
in the light of the assumptions used; as more cost data are accumulated,
the results can be modified as required.

With respect to the fuel cycle costs, established Atomic Energy Commis-
sion prices for thorium, natural uranium, U233, and Pu?3?, and the schedule
of charges for uranium of varying enrichments [1] provide a basis for cost
calculations. Charges for various chemical conversion steps and for proc-
essing spent fuel in a multipurpose chemical plant have also been an-
nounced [2]. Although these charges are applicable to the processing of
aqueous fuels, the possibility of including on-site processing facilities as part
of the homogeneous reactor complex must also be considered, since this
would have an effect on the reactor design.

10-1.2 Parametric cost studies at ORNL. The homogeneous reactor
systems considered include one- and two-region reactors, breeders, con-
verters, and burners. Although no one fuel or type of reactor shows a
marked advantage in power cost over all the others, the superior fuel
utilization of a thorium breeder system suggests that it is potentially the
most economical one for power production. Much effort has been devoted
016 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

to two-region systems primarily because of the relatively high breeding
ratio and low fuel inventory obtainable.

Economic evaluations discussed in this chapter are for the most part
based on a three-reactor station generating a total of 375 Mw of elec-
tricity, where the required chemical processing facilities are shared by the
three reactors. The present choice of reactor dimensions for a given power
capability must be based on engineering judgment and the results of fuel-
cost studies. Based on fuel-cost studies, one-region reactors must be large
(14- to 15-ft diameter) in order to obtain good neutron economy; two-
region systems can have good neutron economy in relatively small sizes
(9- to 10-ft over-all diameter) but require high concentrations (1000 g/liter)
of fertile material in the blanket region. Estimates of near-optimum reactor
sizes for different homogeneous systems are based on fuel-cost studies in
which highly enriched fissionable fuel is valued at $16/g and inventory
charges are 49,. In all cases it is assumed that the particular fuel system is
technologically feasible.

In computing the cost of power, the fixed charges on capital investment
of depreciating items are assumed to be 159 /yr, including depreciation,
interest, return on investment, insurance, and taxes. Fixed charges on
nondepreciating items are assumed to be 49, /yr. Fuel, D20, and fertile
materials are assumed to be nondepreciating materials.

10-2. BaseEs rorR CosT CALCULATIONS

10-2.1 Fuel costs. The fuel costs associated with electrical power pro-
duced from reactors include those charges which are due to replacement of
conventional fuels with nuclear fuel. The fuel cost is considered to be the
sum of the net cost of nuclear-fuel feed; inventory charges for fertile mate-
rial, heavy water, and fuel; material losses; variable fuel-processing
charges; and fixed charges for fuel processing. Fuel-cost studies are pri-
marily for the purpose of investigating the economic importance of the
several parameters; of these parameters the most important are core
diameter, blanket thickness, fertile-material concentrations in core and
blanket, fuel concentration in the blanket, and poison fraction in the core.
In studying fuel costs, aqueous homogeneous power reactors are generally
considered to operate at a temperature of 280°C, a pressure 1500 to 2000
psia, an 809, load factor, and a net thermal-to-electrical efficiency of about
25%,. The thermal power level per reactor is considered to be 450 to 500
Mw. Thorium (as oxide) is valued at $5/lIb with no significant charge for
making ThO2-D20 slurry. Heavy water is valued at $28/1b and highly
enriched fissionable uranium at $16/g. The amount of heavy water re-
quired is estimated on the basis of the total reactor-system volume and the
room-temperature density of heavy water. The makeup rate is taken as
10-2] BASES FOR COST CALCULATIONS 517

59,/yr. The volume of the core circulating system is taken as the volume
of the core plus an external volume of 1 liter for each 20 kw generated in
the core. The blanket external volume (in two-region reactors) is calcu-
lated on the basis of 1 liter for each 14 kw generated in the blanket. The
cost of natural uranium is taken as $40/kg U either as UO3z or UO2504.
The cost of uranium of various enrichments in U235 is obtained essentially
from the AEC price schedule or the equations for an ideal gaseous-diffusion
plant [3-5].

Fuel costs are dependent upon the value assigned to plutonium and U233,
In a power-only economy the value of these fissionable materials can be
no more or less than their fuel value; in these studies their value was taken
to be the same as that for U235 or $16/g. However, if U233 and/or plu-
tonium are assigned different values, the fuel costs can be significantly
affected. This is shown in the results given for plutonium producers, in
which a plutonium value of $40/g was assumed (this value is consistent
with the ability of homogeneous reactors to produce plutonium containing
less than 29 Pu?4°, and the AEC guaranteed fair price schedule for plu-
tonium extending until June 30, 1963).

The fuel values used in these studies are slightly different from those
announced by the AEC; the main difference is associated with the value
used for plutonium. However, the effect of various plutonium values on
fuel cost is indicated in the section on plutonium producers. The AEC-
announced prices for nuclear materials are given in Table 10-1.

Since the rate of fuel burnup is small with respect to the inventory of
fuel required for criticality, the investment and inventory charge for fuel
materials can be appreciable. Unless otherwise specified, inventory charges
for uranium, thorium, plutonium, and heavy water are assumed to be 49,
of their value per year.

An economic consideration in the design of nuclear power plants is the
chemical processing requirement of the spent nuclear fuel. Aqueous
homogeneous reactor fuels can be processed by either the Purex or the
Thorex process. Both involve solvent extraction and require that the
fuels be separated from the D20 for economical processing. The Purex
process 1s used for separating plutonium, uranium, and fission-product
poisons, while the Thorex process is used for the separation of thorium,
uranium, and fission products. Since investment costs in chemical process-
ing plants are presently high, a single processing plant for one good-sized
reactor is not economical. Rather, central processing facilities which serve
many reactors are usually assumed to be available. In the fuel-cost studies
given here, however, the chemical processing plant is considered to serve
a single three-reactor station generating about 1440 thermal Mw (total).
For this size processing plant, the fixed charges (based on 159, of invest-
ment per year) are estimated to be about $5500/day for either Purex or
518 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

TABLE 10-1

U.S. AEC OrriciaL PRICE SCHEDULE FOR
NUcCLEAR MATERIALS [1]

 

(a) Price schedule of U235 as a function of enrichment

 

 

 

: 235
We. (g‘:gfig U $/kg total U $/gm U235 content
0.0072 40.50 5.62
0.010 75.75 8.09
0.020 220.00 11.00
0.030 375.50 12.52
0.10 1,529.00 15.29
0.20 3,223.00 16.31
0.90 15,361.00 17.07

 

 

 

(b) Price schedule of Pu as a function of Pu2?40 content

 

 

 

9 Pu240 Pu price, $/gm
2 41.50
4 38.00
6 34.50
>8.6 30.00

 

 

(¢) Chemical conversion costs

 

Conversion Cost, $/kg

 

 

Uranyl nitrate = UFg

(U containing 59, or less by weight of U233) | 5.60
Uranyl nitrate = UFg

(U containing greater than 59, by weight of U?235) 32.00
Plutonium nitrate - metal buttons 1,500.00
UF¢ (natural U) to oxide Zero

 

 

 

 

(d) Value of U233 (high purity) $15/g

 
10-2] BASES FOR COST CALCULATIONS 519

Thorex (complete decontamination) [6,11]. This corresponds to a power
cost due to chemical plant investment of 0.76 mill/kwh (based on a 375-Mw
net electrical capacity and an 809, load factor), which will be independent,
of the amount of material processed daily.

The variable processing charges arising from labor, materials, and other
factors dependent on the throughput of fuel and fertile material are repre-
sented by Eqgs. (10-1) and (10-2) for processing thorium-uranium mixtures
and uranium-plutonium mixtures, respectively.

Thorex process

Variable daily processing cost = $3.00 W, + 0.50 wy + 0.35 vp,0. (10-1)

Purex process

Variable daily processing cost = $3.50 Wy + $1.00 wpy + 0.35 vp,0. (10-2)
In the above equations:

vp,0 = liters D20 recovered/day,

wpy = g Pu handled separately from U per day,
wy = g U processed per day,

Wrn = kg Th processed per day,

Wu = kg U completely decontaminated per day.

Thus, for Thorex, the variable processing charge is considered to be
$3.00/kg of thorium processed, plus $0.50/g total uranium processed
(U highly enriched in U233 + U?23%) plus $0.35/liter of D3O recovered.

Note that Eqgs. (10-1) and (10-2) take into consideration the effect of
throughput of fissionable material as well as fertile material on the variable
processing charges in a chemical plant of fixed size. This information is
necessary in determining the optimum concentrations of fissionable and
fertile material in a reactor of specified dimensions.

The total processing cost in an on-site chemical plant, such as that
described above, will be 0.76 mills/kwh plus the variable processing charges
and minus the credit for any net fissionable material produced. Thus the
chemical-plant investment costs exert a strong influence on the fuel costs.
These costs are lower if a large multipurpose chemical plant is available to
handle the fuel instead of the on-site plant. Fuel processing charges [2]
have been announced by the United States Atomic Energy Commission
for processing in such a multipurpose plant. These charges amount to
$15,300 per day, and apply to a plant having a daily capacity of 1 ton if
slightly enriched uranium (less than 39 U235 by weight) is processed, or a
daily capacity of 88 kg if highly enriched uranium is considered. In terms
of cost per gram of U235 processed, the above processing charges are
520 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

equivalent to $2.15 per gram of U235 for natural uranium, $0.51 per gram
of U235 for uranium of 3% enrichment, and $0.37 per gram of U235 for
uranium of 479, enrichment. These values can be compared with those
considered previously, namely: $0.50 per gram of total uranium processed
(the enrichment of the uranium in the highly enriched systems is about
50%), and $3.50 per kilogram of natural or slightly enriched uranium
(equivalent to $0.50 per gram of U23% in natural uranium) plus $1.00 per
gram of plutonium. Thus, in most cases, the AEC total processing charges
amount to less than the variable processing charges considered here. In
case of a central processing plant, however, it would be necessary to in-
clude fuel-shipping charges and charges associated with preparing the
processing plant for the specific fuel.

In studying the poisoning effect of fission fragments, three groups of
fission-product poisons are considered. The first group consists of gases;
the second, nongaseous fission fragments having high microscopic cross
sections (greater than ~10,000 barns); and the third is composed of non-
gaseous fission fragments having low microscopic cross sections, and which
transmute to nuclides having the same low cross section. For processing
rates which do not cause excessive variable processing costs, only the third
group of poisons is affected by chemical processing; the first group is re-
moved by means other than Thorex or Purex, while the second group at-
tains equilibrium through neutron capture.

In processing U-Pu systems, a 20-day cooling period takes place before
processing in a Purex plant. Following complete decontamination, the
uranium is permitted to cool 100 more days before being re-enriched in a
diffusion plant. This 120-day holdup and also a 30-day feed supply are
considered in calculating inventory charges. In processing thoria slurries
(core region), the holdup time prior to processing is considered to be 95
days, to permit about 90% of the Pa233 to decay to U233. The processed
material is then held for an additional 110 days to permit the remaining
Pa233 to decay. For thoria slurries in blanket regions, an initial holdup of
55 days is assumed prior to processing, with an additional 150-day holdup
to permit the Pa233 to decay to U233, Thus the protactinium is held up for
205 days, in which time only about 0.5% has not yet decayed. This holdup
time and a 30-day feed supply are considered in calculating material in-
ventories. Protactinium is valued as uranium when outside the reactor,
but no inventory charge is placed against the amount contained in the
reactor system. This procedure is used to take into account that period of
reactor operation between startup and near-equilibrium conditions. Unless
otherwise indicated, the results are based on the assumption that equilib-
rium exists with respect to the nuclei concentrations. The isotope and
fuel concentrations are established by means of material-balance equations
and the critical equation.
10-3] FUEL COSTS IN THO2—-UO3—D20 SYSTEMS 521

In nearly all cases, fuel concentrations required for criticality are ob-
tained using the two-group model [7], in which all fissions are assumed to
occur in the thermal group [see Eqgs. (2-7) through (2-10)]. Resonance
capture is assumed to occur only in fertile material and only when neutrons
are transferred from the fast to the slow group. For the one-region U-Pu
systems, a six-group model is used in the nuclear calculations to allow for
the resonance absorptions in uranium and plutonium. Unless otherwise
specified, the thermal values for the various etas are n(U?33)= 2.25,
n(U235) = 2.08, n(Pu23?) = 1.93, and n(Pu?*!) = 2.23.

10-2.2 Investment, operating, and maintenance costs. The costs con-
sidered here involve capital investment, and those associated with main-
taining and operating the nuclear power plant. Because of the lack of
knowledge and experience in design, construction, and operation of nuclear
power systems, it is difficult to evaluate these costs, and most estimates
are based on the expectation that nuclear reactor plants will have lifetimes
about as long as those of conventional power plants. A 20-year depreciation
rate is assumed for permanent facilities and a 10-year depreciation rate for
all equipment associated with the reactor proper. Preparing a realistic
estimate of the cost and the required maintenance of a large homogeneous
reactor is particularly difficult, since the equipment must handle large
amounts of radioactive material. Little experience has been obtained in
manufacturing the required equipment, and generally the costs are based
upon estimates by manufacturers. These estimated costs of equipment for
a specific reactor power are scaled according to the reactor power raised
to the 0.6 power to obtain the variation of investment cost with power
level. The annual operating and maintenance charges have been estimated
by roughly applying the corresponding charges (based on percentage of
capital investment) in conventional steam plants; however, these estimates
cannot be considered realistic until considerable experience has been ob-
tained by operating actual reactors. These estimates correspond to 3%
of the total capital investment per year.

10-3. ErFEcT OF DESIGN VARIABLES ON THE FUEL CoSsTS
iN THO2-UO3-D20 SysTEMS [8-10]

10-3.1 Introduction. Fuel costs are given here for spherical two-region
reactors with U23303—-ThO2-D20 slurry employed in both the core and
blanket regions. Results are also given for the case of no thorium in the
core and also for one-region reactors. In addition, some results are given
for two-region reactors having cylindrical geometry. In the two-region
systems, materials from the core and blanket are assumed to be fed to a
Thorex plant for chemical processing. Thoria can be returned to both
522 HOMOGENEQUS REACTOR COST STUDIES [cHAP. 10
Feed D,O Makeup Thorium Feed
Feed D,O Supply on
Storage Han
L
______ _.'________._ — _.._____:'___ —
Reckt U Core | <= Core Blanket —p  Blanket
eactor Externa External
System r Volume |ep -- Volume ~

Core Poison
Removal

T

 

ok

£

Recovery

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  

 

Hydroclone Separation
of Fission Products

- D2O Evaporation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pre-Chemical Core ' Blcmkdet HoldtIJp,
P i and Initia
r::’szmg ;I;Igup, Additional Processing
P ays Holdup, 35 Days
150 Days Pa
—_—— e e e e ———— ). U ________ —_———
\ Y
Core Blanket
Chemical Chemical
Chemical Processing Processing
Processing ‘U F’oisons Poisons U Th {

 

Y
Net Uranium
Production
High Purity y233

 

 

Products Fission Products

Fig. 10-1. Schematic fuel-processing flowsheet for a two-region homogeneous
thorium breeder reactor.

regions and fuel is returned to the core as needed to maintain criticality;
the fissionable material produced in excess of that required is considered
to be sold. The fuel product is computed to be a mixture of U233, U235
and other uranium isotopes, as determined by the isotope equations and
the critical equation. The system is assumed to operate under equilibrium
conditions. A schematic flowsheet of the chemical processing cycles for a
two-region reactor having a solution-type core region is given in Fig. 10-1.
The flowsheet for the one-region system would be similar to that for the
blanket of the two-region system, except that processed fuel and thoria
would be returned to the single region.

In the processing cycle shown in Fig. 10-1, essentially two methods of
removing fission-product poisons are considered. One is the removal of
precipitated solids by hydraulic cyclones (hydroclones); by this means
the insoluble fission products are removed from the reactor in a cycle time
of about a day. The second is the removal of essentially all fission products
10-3] FUEL COSTS IN THO2—-UO3—D20 SYSTEMS 523

by processing tne fluid in a Thorex-processing plant. Processing by hydro-
clones can be done only with solution fuels; the associated cycle time is so
short that fission products removed by this method can be considered to
be removed from the reactor as soon as they are formed. Thorex processing,
although removing all fission products that pass through Thorex, is much
more costly than hydroclone processing. Because of this, the associated
cycle time is usually several hundreds of days. In what follows, unless
specified otherwise, the term fuel processing applies only to Thorex or
Purex processing.

The essential difference between the processing cycle shown in Fig. 10-1
and that for solid-fuel reactors is associated with the continuous removal
of fission-product gases and of insoluble fission products (in solution re-
actors). Fuel and fertile material processed by Thorex would undoubtedly
be removed from the reactor on a semibatch basis.

The three groups of fission-product poisons considered previously are
not all affected by Thorex processing; group-1 poisons (the fission-product
gases) are assumed to be physically removed before processing, while
group-2 poisons (nongaseous nuclei having high cross sections) are effec-
tively removed by neutron capture within the reactor system (o, of these
nuclei is of the order of 10,000 barns). The macroscopic cross section of
these two groups of poisons is taken as 1.89, of the fission cross section.
Of this, 0.89 is due to nongaseous high-cross-section nuclei, while 19} is
due to the gaseous high-cross-section nuclei. The concentrations of low-
cross-section nuclei (third group) are affected by Thorex processing (how-
ever, for reactors containing a fuel solution, the nonsoluble group-3 poisons
are assumed to be removed by hydroclone separation). The charge for
hydroclones operating on a one-day cycle is taken as 0.03 mill/kwh, based
upon a charge of $75/day per reactor. For these solution reactors it is
assumed that 759, of the group-3 poisons are insoluble and removed by
hydroclones; in these circumstances only 259, of the generated group-3
poisons are removed by the Thorex process. With slurry-core reactors, all
group-3 poisons which are removed are removed by Thorex processing.

The parameter ranges covered in the spherical reactor calculations are
given in Table 10-2. Values used for 723 and resonance escape probability
are presently accepted values; however, in a few cases they were varied in
order to estimate how the results are affected by these changes. In the
following sections the influence of specific parameters upon fuel cost is
discussed.

10-3.2 Two-region spherical reactors [8]. (1) Concentration of U233 in
blanket and core poison fraction. The optimum values of these variables are
found to be largely independent of other parameters; moreover, there is
little change in fuel cost with changes in either blanket 17233 concentration
524 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

 

| |
Blanket Thorium (g/liter) /2000

 

Net Fuel Cost
(mills/kw-hr)

 

 

 

1.5 2.0 2.5 3.0
Blanket Thickness (ft)

F1g. 10-2. Fuel cost as a function of blanket thickness for various blanket
thorium concentrations. Power per reactor = 480 Mw (heat), core diameter = 5 ft,
core thortum = 200 g/liter, core poison fraction = 0.08, blanket U233 = 4.0 g/kg
Th, n23 = 2.25.

or core poison fraction. For all slurry-core systems, the optimum poison
fraction is about 0.08, independent of the other design parameters. The
optimum poison fraction for the solution core is about 0.07. The lowest
fuel cost occurs at a blanket U233 concentration of about 4.0 g/kg thorium.

 

 

 

TABLE 10-2
PARAMETER VALUES USED IN SLURRY REACTOR STUDIES
Two-region reactors | One-region reactors
Core diameter, ft 815 8-20
Blanket thickness, ft 13-3
Core thorium concentration,
g/liter 0-300 0-400
Core poison fraction, 9 3-20 4-12
Blanket thorium concentration,
g/liter 500-2000
Blanket U232 concentration,
g/kg thorium 1-7

 

 

 

 

 

(2) Blanket thickness and blanket thorium concentration. An example of
the effects of these parameters on fuel cost is presented in Fig. 10-2 for a
slurry-core reactor. Here it is noted that the blanket thorium concentra-
tion has relatively little effect on the minimum fuel costs. The blanket
thickness giving the lowest fuel cost lies between 2 and 2.5 ft. As is ex-
pected, higher thorium loadings are desirable if thin blankets are necessary
on the basis of other considerations. Systems having low concentrations of
thorium in the core require more heavily loaded blankets to minimize fuel
costs. For solution cores, still heavier and thicker blankets are desirable,
particularly if the core diameters are small.
10-3] FUEL COSTS IN THO2-UO3—D20 SYSTEMS H25

 

  
  

 

 

 

1.6 —
'g Core Thorium
X (g/liter)
] 50
= 1.4} e —
§ ~—_ == _ 100
o ~— et I == ~—300
E 1.9 ".13-0-.-:::.:.:::.-—_':__‘.}— —_—— T 200 |
2 7

1.0 | | I

4(2.5) 5(2) 6(2) 7(2)

Core Diameter (Blanket Thickness) (ft)

F1a. 10-3. Fuel cost as function of core diameter and core thorium concentration.
Power per reactor = 480 Mw (heat), blanket thorium = 1000 g/liter, blanket U233
= 4.0 g/kg Th, core poison fraction = 0.08, 23 = 2.25.

(3) Core thorium and core diameter. The effects of these variables upon
fuel costs are shown in Fig. 10-3. These results indicate (on the basis of
fuel cost alone) that the small solution-core reactors (ThOs core concen-
tration equal to zero) have a slight advantage over the slurry reactors.
However, the power density at the core wall is between 160 and 300 kw/liter
for such reactors operating at the given power level of 480 thermal Mw. If
larger cores are required because of power-density limitations, the fuel-cost
advantage moves to the slurry core. The slurry-core systems yield higher
outputs of generated fuel, although all the reactors shown have breeding
ratios greater than unity (see Chapter 2). As illustrated in Fig. 10-3, the
minimum fuel cost is about 1.2 mills/kwh, independent of core diameter.
The fuel cost associated with a core thorium concentration of zero is lower
than that associated with a core thorium concentration of 50 g/liter; this
is due to the ability to use hydroclones to remove fission products only
when a fuel solution is used. The hydroclone installation adds only
0.03 mill/kwh investment cost to the system, while the variable Thorex
processing cost is reduced by two-thirds; this results in the decrease in fuel
costs as shown. The relative flatness of the optimum net fuel cost curve in
Fig. 10-3 is due to compensating factors; i.e., changes in processing charges
and yield of product are offset by accompanying changes in the fuel in-
ventory charge. Similar compensating effects account for the insensitive-
ness of fuel costs to changes in other design parameters.

Table 10-3 presents a breakdown of costs for some typical reactors hav-
ing low fuel costs. The changes which occur when thorium-oxide slurry is
 

 

 

 

 

 

 

526 HOMOGENEQOUS REACTOR COST STUDIES [cHAP. 10
TaBLE 10-3
CosT BREAKDOWN FOR SOME TypricaAL REACTORS
Core diameter, ft 6 5 4 6 14
Blanket thickness, ft 2 2 23 2
Core thorium concentration,

g/liter 200 100 - | O 0 250
Blanket thorium concentration,

g /liter 1000 1000 1000 1000
Blanket U233 concentration,

g/kg thorium 4 4 4 4
Core poison fraction 0.08 0.08 0.08 0.08 0.08
Critical concentration,

g U233/liter 9.4 |64 4.1 1.4 6.8
Net breeding ratio 1.102 | 1.081 | 1.089 | 1.045 | 1.012
Core wall power density, kw/liter | 53 91 170 80
Core cycle time, days 637 418 884 342 1094
Blanket cycle time, days 295 205 176 210
Inventory of U233 and U235 kg | 368 272 200 148 522
Inventory of heavy water, lb 96,100 | 87,400 | 89,600 | 99,500 | 157,000
Net U233 and U235 production,

g/day 49 39 43 21 6
Grams of U233 per g of U pro-

duced 0.67 0.65 0.77 0.72 0.41
Estimated cost, mills/kwh |

Uranium inventory 0.27 0.20 0.15 0.11 0.38

D20 inventory and losses 0.27 0.25 0.25 0.29 0.45

Thorium inventory and feed 0.01 0.01 0.01 0.01 0.01

Fixed chemical processing 0.76 | 0.76 0.76 0.76 0.76

Core processing 0.13 0.13 0.07 0.08 0.19

Blanket processing 0.09 0.12 0.18 0.18

Uranium sale, credit 0.33 0.26 0.29 0.15 0.04

Net fuel cost 1.22 1.22 1.16 1.29 1.76

 

 

used in the core can be seen by comparing results for the two 6-ft-core-
diameter reactors.

(4) Reactor power. The above results are based on the concept of a three-
reactor station of 1440-thermal-Mw capacity, each reactor producing
125 Mw of electricity. The effect of varying power alone is shown in
Table 10—4; the fuel cost is found to be a strong function of power capa-
bility. The greater part of the change is due to variation in the fixed
chemical-processing charge. Since the total fixed processing cost
($5500/day) is assumed to be independent of throughput, this charge on
a mills/kwh basis is inversely proportional to the reactor power.
10-3] FUEL COSTS IN THO2—U03—D20 SYSTEMS 527

TABLE 104

ErrEcT oF PoweErR LEVEL oN FurL Cosrts

 

 

 

Electric power per Net fuel cost, Fixed .chemlcal-
- processing charge,
reactor, Mw mills/kwh mills /cwh
80 1.75 1.19
125 1.22 0.76
200 0.88 0.48

 

 

 

 

 

(5) Nuclear parameters. The values of 7(U233) and the resonance escape
probability of thorium-oxide slurries are not known with certainty. There-
fore the effects of changes in these parameters on the results were com-
puted in order to examine the reliability of the nuclear calculations. The
fuel cost increases by about 0.2 mill/kwh if %23 is changed from 2.25 to
2.18, and is reduced by about the same amount if the 723 value is taken to
be 2.32 rather than 2.25.

The importance of resonance escape probability for thoria-D»O slurries
(p%2) upon fuel cost was studied by using values for (1 — p°2) which are
20% higher or lower than a standard value. With a core thorium concen-
tration of about 200 g/liter, the changes in p°2 have a negligible effect upon
fuel cost. At lower core thorium concentrations, the changes in p°2 result
in fuel-cost changes of about 0.05 mill /kwh.

(6) Xenon removal. For most of the cases studied, the contribution of
xenon to the poison fraction is assumed to be 0.01. To achieve this condi-
tion, about 80% of the xenon must be removed before neutron capture
occurs. Since xenon-removal systems for slurries have not been demon-
~ strated to date, the effect of operating without xenon removal was studied
by increasing the xenon poison fraction to 0.05 (the samarium contribution
was held at 0.008). In the systems examined, when the xenon poison frac-
tion was increased by 0.04, the total poison fraction yielding the lowest fuel
cost also increased by approximately the same amount. The values in
Table 10-5 illustrate this effect by comparing two cases at optimum total
poison fraction, but at different xenon poison levels. Thus the variable
part of the core poison fraction (and the core processing rate) remains about
the same. The higher fuel cost at the higher xenon level appears to be al-
most entirely a result of the reduction in breeding ratio.

10-3.3 One-region spherical reactors [8]. (1) Pozson fraction. The
poison fraction producing the minimum fuel cost for a given system is in
the range from 0.06 to 0.10, the exact value depending on the specific di-
528 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

 

 

 

 

 

2.4
l |
Core Diameter (ft) 18

22—
<
3
&
&
E 201
2
o
o
2
T 18— —
Z

i | | | | | |

50 100 150 200 250 300 350 400

Thorium Concentration, g/liter

Fig. 10-4. Fuel cost as function of thorium concentration in one-region reac-
tors. Power per reactor = 480 Mw (heat), poison fraction = 0.08, 7?3 = 2.25.

ameter and thorium concentration. However, a value of 0.08 gives costs
which are close to the minimum for all cases.

(2) Diameter and thorium conceniration. The fuel costs for some of the
single-region reactors studied are shown in Fig. 10~4. Detailed information

TaBLE 10-5

ErrEct OF XENON PoisoN FractioN oN FueL Costs

 

Xenon poison fraction 0.01 0.05
Optimum total poison fraction 0.08 0.12
Core cycle time, days 637 718
Breeding ratio 1.102 1.070
Fuel inventory charge, mill/kwh 0.27 0.29
Core processing charge, mill/kwh 0.13 0.14
Fuel product (credit), mill/kwh 0.33 0.22
Net fuel cost, mills/kwh : 1.22 1.35

 

 

 

 

 

for a typical one-region reactor is given in the last column in Table 10-3.
In general, for thorium concentrations less than 400 g/liter the reactor
diameter must be greater than 10 ft in order to have a breeding ratio greater
than unity; the 12-ft-diameter reactor is a breeder (breeding ratio = 1.0)
at a thorium concentration of 350 g/liter, while at 250 g Th/liter the
14-ft-diameter reactor is a breeder. For reactors between 10 and 16 ft in
diameter, the thorium concentration yielding the lowest fuel costs is be-
10-3] FUEL COSTS IN THO2—-UO3—D20 SYSTEMS 529

tween 200 and 275 g/liter. The lowest fuel cost is about 1.76 mills/kwh
(for a 14-ft-diameter reactor containing 270 g Th/liter). In the curve for
the 14-ft reactor, the inflection in the neighborhood of 225 g Th/liter is a
result of the reactor changing from a breeder to a nonbreeder. This in-
flection is associated with a marked increase in U236 concentration (concen-
trations are based on equilibrium conditions), which produces an increase
in fuel processing charges (in all cases it is assumed that U233 is fed to
the system).

(3) Power and nuclear parameters. The effect of reactor power on the
fuel costs of one-region reactors is similar to that mentioned earlier for
two-region systems. For example, if the total fuel cost for a three-reactor
station is 1.76 mills/kwh at 125 Mw of electric capability per reactor, it
would be only 1.38 mills/kwh if the output per reactor were increased to
200 Mw.

The importance of changes in nuclear parameter is generally the same
for the one-region reactors as for the two-region systems, although the
effect upon fuel costs of a reduction in 7(U233) is somewhat greater for the
one-region cases.

10-3.4 Cylindrical reactors [9]. The effects of geometry on fuel cost
are due to the associated changes in inventory requirements and breeding
ratio. Accompanying these effects are changes in the average power den-
sities and the wall power densities within the reactor. Because of corrosion
difficulties associated with high power densities, it is desirable to operate
with reasonably large reactor volumes. Cylindrical geometry permits re-
actors to have large volumes without necessitating large reactor diameters.

One-region spherical reactors would have to be large in order to prevent
excessive neutron leakage, and so power densities would not be high
(average power densities of about 30 kw/liter). Also, closure problems
with respect to maintenance of an inside vessel would not exist. Therefore
there is little incentive to increase reactor volume by using cylindrical
geometry for one-region homogeneous reactors.

For two-region reactors, cylindrical geometry may prove advantageous
with respect to feasibility and relative ease of reactor maintenance. How-
ever, the associated larger fuel inventories (in comparison to inventories
for spherical geometry) will increase fuel costs. Comparison of results for
two-region cylindrical reactors with those for spherical two-region reactors
shows that cylindrical geometry gives minimum fuel costs about
0.2 mill/kwh greater than does spherical geometry, if in either case there
were no restrictions on core-wall power density. The difference is even
greater if the core-wall power density influences the reactor size. However,
cylindrical geometry does permit low wall power densities in combination
with relatively small reactor diameters.
530 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

U, Pu Feed Conversion of Enriched U
Feed Preparation UFg to UO3 e ALSL
L—— 30-Day Reserve

A

 

 

 

 

 

 

Pu

.
-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

U, Pu
Partial Decontaminatio Separation of Pu 100-Day Coolin
of All Fuel Sent to 1 . and Complete
: - . el and
U, Pu, F.P. Chemical Plant Decontamination of Fuel[ UF. P )
20-Day Cooling to Diffusion Plant 6 Freparation
F.P. F.P.

 

.
-

Y ’ Waste

 

 

 

Waste

F1ag. 10-5. Fuel cycle for one-region, U-Pu reactor.

10-4. EFrreEcTt OF DESIGN VARIABLES ON FUEL CosTS
IN URANIUM-PLUTONIUM SYSTEMS

Fuel costs in uranium-plutonium systems will depend on whether the
plutonium is removed continually or allowed to remain in the reactor. In
the latter case, there is little difference in fuel costs between a
U05804~Pu0s-D30 system and a UO3—PuO3z-D20 system. Fuel-cost
studies have been made on the basis of either maintaining plutonium
uniformly within the reactor or removing the plutonium immediately after
its formation by means of hydroclones. In addition, the effect upon fuel
costs of LilSO4 addition to UO2S04 solutions is shown. The additive
serves to suppress two-phase separation of the UO2504 solution, and per-
mits reactor operation at temperatures higher than would otherwise be
feasible.

10—4.1 One-region PuO;-UO3-D;0 power reactors [11,12]. Fuel costs
of one-region homogeneous power reactors fueled with PuO2-U0O3-D20
slurries are given as functions of operating conditions, based on steady-
state concentrations of U235 U236 Pu239 Pu24 and Pu2*l. All other
higher isotopes are assumed to be removed in the fuel processing step or to
have zero absorption cross section. Although large reactors require feed
enrichments equal to or less than that of natural uranium, minimum fuel
costs are obtained when the reactor wastes are re-enriched in a diffusion
plant. In the reactor system considered, the plutonium produced is fed
back as reactor fuel after recovery in a Purex plant. The separated uranium
is recycled to a diffusion plant for enrichment. Since in all cases the breed-
ing ratio is less than unity, the additional fuel requirement is met by
uranium feed of an enrichment dictated by the operating conditions. As a
result of the feedback of plutonium, the concentration of plutonium in the
reactor is high enough for resonance absorptions and fissions to have an
appreciable effect upon fuel concentrations.
10-4] FUEL COSTS IN URANIUM-PLUTONIUM SYSTEMS 531

The fuel cycle is shown in Fig. 10-5. Slurry is removed from the reactor
at the rate required to maintain a specified poison level. The fuel is sepa-
rated from the D20, cooled for 20 days while the neptunium decays, and
partially decontaminated in a Purex plant. Part of the fuel is then sent
directly to the reactor feed-preparation equipment. Plutonium is sepa-
rated from the remainder of the uranium and added to the reactor feed.
The uranium is completely decontaminated, stored for 100 days until the
U237 decays, and sent to the diffusion plant. A 30-day reserve of reactor
feed is kept on hand.

TaBLE 10-6

CostT BREAKDOWNS FOR ONE-REGION PuQ2-UQO3-D-0O
ReAacTORS* HAVING A DIAMETER OF 12 FT

 

Process characteristics:
U238 concentration, g/liter 175 200 225
N4l 1.9 2.2 2.4
Reactor poisons, 9, 6.0 6.0 6.0
Total system volume, liters 50,000 50,000
Chemical process cycle time, days 151 138
Initial enrichment (no Pu), U235/U 0.011 0.011 0.011
U235 feed, g/day 579 430
Feed enrichment, 9, U235 0.99 0.70 0.52
U235 concentration, g/liter 1.12 0.82
Pu?3? concentration, g/liter 0.86 1.17
Pu?40 concentration, g/liter 0.49 0.62
Pu24! concentration, g/liter 0.36 0.49
Np?39 concentration, g/liter 0.03 0.03
U236 concentration, g/liter 0.10 0.06
Conversion ratio 0.74 0.84
Fuel costs:
Fuel inventory, mills/kwh (at 49) 0.08 0.07
D20 inventory, mills/kwh (at 49,) 0.15 0.15
D20 losses, mills/kwh (at 59) 0.19 0.19
Net uranium feed cost, mills/kwh 1.15 0.48
Variable chemical processing cost,
mills/kwh 0.27 0.32
Fixed costs for chemical processingT,
mills/kwh 0.76 0.76
Total fuel cost, mills/kwh 2.60 2.20 1.97

 

 

 

 

 

 

*Operating at 280°C; 480 thermal Mw, 125 elec. Mw; 809, load factor; optimum
poisons, 5-79.

TEssentially assumes a chemical processing plant servicing three reactors pro-
ducing 1440 thermal Mw.
532 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

2.3 |

 

I l |

22—

21—

20—

    
 

Reactor Dia.
12 ft

Variable Fuel Cost, mills/kwh

1.8

 

 

 

150 200 250 300 350 400

Uranium Concentration, g/liter

Fic. 10-6. Effect of uranium concentration and reactor diameter on fuel cost in
one-region reactors. 300 Mw of electricity, 1000 Mw of heat, Avg. reactor tem-
perature = 330°C, U02804-Li2804-D20 solution with dissolved Pu, molar ratio
of LisS0O,4 to U02804 = 1, optimum poisons = ~ 5%,.

Some typical results are given in Table 10-6 for a 12-ft-diameter reactor,
in which different values were assumed for %! (the “‘best value’ for n*! is
2.2; therefore, based on these results, natural uranium is adequate feed
material to ensure criticality). Where required, the cost of slightly en-
riched feed is based on the established schedule of charges by the AEC.
Plutonium in the reactor is assumed to have a value of $16/g. Fuel costs
were somewhat lower at larger reactor diameters, but a 12-ft diameter
corresponds to a more feasible reactor size.

10—4.2 One-region U03S04-Li;S04~D20 power reactors [13]. For mini-
mum fuel costs, one-region reactors require fertile-material concentrations
of several hundred grams per liter. The addition of Li2SO4 in molar con-
centration equal to that of the uranium increases the temperature at which
phase separation appears and also acts as a corrosion inhibitor for stainless
steel. Because of the high neutron-capture cross section of natural lithium,
high isotopic purity in Li7 is necessary. Fuel costs for power-only reactor
systems fueled by UO2804-Li2S04-D20 solutions are given here in which
10-4] FUEL COSTS IN URANIUM-PLUTONIUM SYSTEMS 533

TaBLE 10-7

REsuLTs FOR SEVERAL ONE-REGION REACTORS*
NEArR OpTiMmuM CONDITIONS

(Minimum fuel costs)

 

Process characteristics:
Reactor diameter, ft 12 16 20
U238 concentration, g/liter 200 200 200
Feed enrichment, w/o U235 2.11 1.48 1.22
Feed rate, g/day of U235 1078 923 838
Chemical processing cycle, days 359 408 543
Poisons, 9, 5 5 5
Reactor volume, liters 25,600 60,700 119,000
Total system volume, liters 88,000 127,000 181,000
Power density, kw/liter 39.0 16.5 8.4
Isotope concentration, g/liter:
U235 2.11 1.47 1.21
236 0.37 0.26 0.21
Py239 0.97 0.88 0.84
Py240 0.51 0.50 0.49
Np239 0.03 0.02 0.02
Py24! 0.52 0.45 0.43
Conversion ratio 0.67 0.72 0.74
Variable fuel cost, mills/kwh: |
Net feed cost 1.42 1.12 0.96
Variable chemical processing 0.11 0.11 0.10
Fuel inventory 0.10 0.09 0.10
D20 inventory 0.11 0.16 0.23
D20 losses 0.14 0.20 0.29
Lithium losses and inventory 0.006 0.008 0.011
Hydroclones 0.04 0.04 0.04
Total variable fuel cost 1.93 1.73 1.75

 

 

 

 

 

 

*1000 Mw (heat); 300 Mw (electrical); 330°C reactor temperature.

all plutonium is assumed to remain either in solution or uniformly sus-
pended throughout the reactor. The plutonium is returned to the reactor
following fuel processing, and steady-state conditions are assumed. The
same type of fuel cycle as shown in Fig. 10-5 is considered. The results
shown in Fig. 10-6 are for spherical reactors operated at an average tem-
perature of 330°C, producing 1000 Mw of thermal energy, and delivering
300 Mw of electricity to a power grid. The variable fuel costs given do
not include fixed charges for fuel processing; these fixed charges would add
about 0.76 mill/kwh to the fuel costs given. Results are based on a 4%
HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

o34

*SOUO[D0IPAY Aq paAOWIAI J0U PUE J[qN[os A[9)o[dWI0D 8q 0} POWINSSE SeM WNIUOIN[J (d)
'6—0T 9[qBJ, Ul UMOYS aIe sosed 9soy) ut s[resa (p)

-o[qIB1[Seu Surssao0ad [BOTWAYD 10§ SABIBYD PoxXY 9} X BUI

0} s1099%a1 ySnouo Sumiares ore sjue[d Juisseoord [ony Y3} JeY} UOHAUINSSE 9y} UO pIseq (d)

"WNIUBIN

JO UOIJBIUQIUOd IB[OW dY} 0} [enbe aq 03 PIWNSSE SBM SAIIPPE JO UOTIBIJUIIUOD IBJOJA (q)
‘O ~YOSEQ[] ‘UOIIN[OS [aN] {SINY[ (0 LG ‘OUWN[OA TI)SAS
{1 9 “I0JOWBIP ‘MJN [BOLIJO9[d Gg[ ‘MIN [eWdy} (8% ‘emod ‘D 087 ‘arnjeraduwa) 938I0AY (®)

 

 

 

 

 

 

 

 

 

 

 

 

L0°G 91°¢ GE '8 00¥ UON @I 6
91 ¢ ¥¢'¢ 66 ¢ 00€ QUON @1 8
€46 @V ¢ I L 00¢ TOS®r'T % L
69 ¢ G¢9 ¢ a8y 003 TOSerI % 9
€0'¢ 96 ¢ 66 ¢ 001 TOS®rT % g
Vv ¢ ¥E€'¢ GE '8 00v QUON ® v
8Y ¢ PGV G 66 ¢ 00€ SUON % €
L9°¢ 19°C 8TV 002 UON % 6
GL'G 001 UON o I
POIdA0DAI POISA099
ng pue ) ON ng pue 0 ggz(l sgall (skep)
‘ou
uoryBI1ado 1894-(F 1091 /3 (@OMHPPY oflfiwmwowfioMo ose))
‘UO1}BIJUIIUOD POIPAH
UMY /S[[IW (5 ‘S800 [P0 ] wnIueIn [BIIuy

 

(®)

80T TIEV],

SHOLOVEY ¥AMOJ OP(-YOS0)[) SNOENIDOWOJ] CALVHEJI()-HOLVE dOL SISO TAN]

 
10—4] FUEL COSTS IN URANIUM-PLUTONIUM SYSTEMS 535

inventory charge and $28/Ib D20O. Table 10-7 gives a breakdown of costs
for several reactors near optimum conditions. The Li? cost is assumed to be
$100/1b; at this value the effect of lithium cost upon fuel cost is negligible.
However, the added Li;SO4 acts as a neutron poison and diluent which
lowers the conversion ratio; these effects increase fuel costs by 0.2 to 0.3
mill /kwh over those which would exist if no Li;SO4 were required.

The above results are for equilibrium conditions, with continuous fuel
processing. These reactors can also be operated with no fuel processing,
in which case nonequilibrium conditions apply. Under such conditions the
variable fuel cost would be greater than for the case of continuous fuel
processing; therefore any economic advantage associated with batch opera-
tion can exist only if fixed charges for fuel processing (or storage) at the
end of batch operation are effectively lower than fixed charges associated
with continuous processing. Fuel costs are given below for spherical one-
region reactors operating on a batch basis, utilizing an initial loading of
slightly enriched uranium. In all cases, fuel feed is highly enriched UZ23°,
The total reactor power is 480 thermal Mw (125 electrical Mw), and an
0.8 load factor is considered. Inventory charges are 4% /yr, cost of D20
is $28/1b, and the cost of uranium as a function of enrichment is based
on official prices. A summary of the fuel costs is given in Table 10-8 for
6-ft-diameter reactors. More details for two reactors are given in Table
10-9. Credit for plutonium is based on a fuel value of $16/g; it is assumed
that plutonium will remain within the reactor. The shipping costs do not
include fixed charges on shipping containers.

For these reactors (6-ft diameter), the addition of Li2SO4 (99.98% Li7)
(cases 5, 6, and 7) to the fuel solution raises the fuel cost slightly, the max-
imum increase being about 0.1 mill/’kwh. The use of hydroclones (cases 8
and 9) for partial poison removal (assuming no plutonium removal) for
these reactors shows an economic advantage, particularly in the ‘“‘throw-
away’’ fuel costs (no uranium or plutonium recovery). For these latter
costs, the removal of fission-product poisons reduces the fuel costs by 0.3
to 0.4 mill/kwh for 20-year operation. In view of its low solubility, how-
ever, plutonium would be extracted along with the fission-product poisons.
The economic feasibility of hydroclones in these circumstances, in a power-
only economy, is dependent upon the savings effected by posion removal
relative to the costs associated with recovery of the plutonium for fuel
use. Fuel costs for 10-year operation were 0.1 to 0.2 mill/kwh higher than
those for 20-year operation.

104.3 Two-region UO3-PuO2-D;0 power reactors [11]. The major
advantage associated with two-region reactors is that good neutron econ-
omy can be combined with relatively low inventory requirements. None
of the uranium-plutonium fueled aqueous homogeneous reactors. appear
536 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

TaBLE 10-9

IsoroPE CONCENTRATIONS AND CosST BREAKDOWN FOR
BAaTcH-OPERATED HOMOGENEOUS REACTORS*
(20-year operation, no hydroclones, 300 g U238/liter)

 

Initial U235 concentration, g/liter 5.99 7.21
Additive 0 Li2SO4
Average U235 feed rate, kg/yr 136 136
Initial U235 inventory, kg 163 196
Final poison fraction 0.27 0.25
Final isotope concentration, g/liter:
U234 1.22 1.28
U235 12.2 13.2
U236 13.9 14.0
U238 223 223
Np239 0.03 0.03
Pu23? 3.54 3.9
Pu240 3.62 4.04
Pu24 1.24 1.38
Estimated fuel costs, mills/kwh: »
Uranium inventory 0.08 0.11
D20 inventory and losses 0.19 0.19
Uranium feed 2.12 2.12
Chemical processing 0.18 0.18
Hydroclones 0.04 0.04
Shipping 0.04 0.04
Plutonium sale (credit) 0.07 0.08
Diffusion plant (credit) 0.12 0.12
Total (fuel cost) 2.42 2.44

 

 

 

 

 

*Same conditions as listed in Table 10-8.

capable of producing more fuel than is burned; however, it is possible to
operate a two-region reactor with such a low steady-state concentration
of U235 in the blanket that natural uranium can be fed to the reactor and
the waste discarded,

In a power-converter the plutonium is extracted from the blanket and
fed to the core as fuel material. Natural uranium is fed to the blanket,
and the plutonium formed is extracted by Purex processing. By adjusting
the concentrations of fissionable materials in the blanket, the net rate of
production of plutonium in the blanket is made equal to the consumption
in the core. The fuel cycle considered for a two-region reactor is shown in
Fig. 10-7. Plutonium from the core is processed continuously through a
10-5] FUEL COSTS IN PLUTONIUM POWER REACTORS 537

Conversion of
o Feed { Pre;(;?gtion ‘ UF¢ to UO3
30-Day Reserve

 

 

 

 

 

 

Q- Partial Decontamination| Complete 100-Day Coolingj MWN
Y Q\)‘{(' and Pu Separation Decontamination of U and Diffusion Natural U
Blanker, 20-Day Cooling to Diffusion Plant UF¢ Preparation Plant Feed

Core F.P. , l
A Pu

 

 

 

 

 

 

 

 

 

(7
":"? \

 

 

Pu Decontamination, Wast
20-Day Cooling e

b, L_FP.
i

Pu Feed Core Feed L
Preparation \

Waste

 

 

 

 

 

Fia. 10-7. Fuel cycle for two-region U-Pu reactor.

Purex plant to maintain a constant poison fraction. Uranium from the
blanket is also processed through the Purex plant, the rate of processing
being governed by the required plutonium feed to the core. Although it
is possible to operate the reactors without an enriching plant by using
natural-uranium feed to the blanket and discarding the waste, there ap-
pears to be a slight cost advantage in operating the reactors in conjunction
with an enriching plant.

Table 10-10 gives fuel costs and associated information for a two-region
reactor having a core diameter of 6 ft and a 10-ft over-all diameter. The
value of n*! is assumed to be 1.9. (A more recent value of 7*! = 2.2 is be-
lieved to be more accurate.) Comparison of these costs with those for
one-region reactors of 12 to 14 ft diameter indicate that two-region U-Pu
reactors have fuel costs about 0.2 to 0.3 mill/kwh lower than do one-region
reactors.

10=5. FueL Costs IN DuaL-PurrosE PrutoNiuM POowER REACTORS

Since plutonium is quite insoluble in UO2804-D20 solutions and can
be removed by hydroclones, it is possible to produce high-quality plu-
tonium (Pu24® content less than 2%). Based on the AEC price schedule,
this high-quality plutonium has a net value of at least $40 /g after allow-
ing $1.50/g for conversion to metal). At this value it is more economical
to recover plutonium than to burn it as fuel. Fuel costs based on recovery
of the plutonium are given here for one- and two-region U-Pu reactors.
HOMOGENEOUS REACTOR COST STUDIES

TAaBLE 10-10

FueL Costs For A Two-REGION,

U-Pu Power REracToRr*

[cHAP. 10

 

Core diameter, ft 6
Core power, thermal Mw 320
Core Pu concentration, g/liter 1.7
N40/N49 in core 0.99
N4 /N49 in core 0.35
Blanket thickness, ft 2
Blanket power, thermal Mw 180
Blanket U concentration, g/liter 500
N25/N28 in blanket 0.003
N49/N28 in blanket 0.001
N40/N28 in blanket 0.0003
N4 /N28 in blanket v 0.00005
Blanket feed enrichment, N25/NU 0.004
Fraction fissions in U235 0.27
Fraction of U consumed 0.017
Fuel costs, mills/kwh
Core processing (variable) 0.16
Blanket processing (variable) 0.31
D20 recovery 0.07
U 4 Pu losses 0.005
Pu inventory 0.04
U inventory 0.005
D20 inventory plus losses at 99,/yr 0.30
Feed cost minus credit for returned U 0.62
Fixed charges for chemical processingt 0.76
Total fuel costs, mills/kwh 2.27

 

 

 

 

*Average temp., 250°C; total power, 500 thermal Mw; net
thermal efficiency, 25%; load factor, 809,; n* = 1.9.
TAssumes chemical-processing plant servicing three reactors.

10-5.1 One-region reactors [14]. Fuel costs are given for UQ2S04+~D 20
and U02804-LiZS04-D:0 fuel systems operating at 280°C, in which the
plutonium is assumed to be removed by means of hydroclones on about a
one-day cycle. Characteristics of the assumed systems are presented in
Table 10-11.
10-6] FUEL COSTS IN U235 BURNER REACTORS 539

TAaBLE 10-11

CHARACTERISTICS OF REACTOR SYSTEM

 

Electrical power, Mw 125
Heat generation, Mw 480
Reactor diameter, ft 12
Power density in system external to core, kw/liter 20
Average reactor temperature, °C 280
Reactor poisons, 9, 5
Chemical processing rate, g U23%/day 1000
Plutonium removal Instantaneous
Isotopic purity of lithium, %, 99.98
Processing method Purex
Li7 cost, $/1b 40
Inventory charges, 9 4
D30 losses, 9, of inventory/yr 5

Li losses, 9 of inventory/yr 1

 

 

 

 

Because of the high credit for plutonium, the fuel costs are negative even
if the fixed charges for chemical processing (0.76 mill /’kwh) are included.
The total fuel costs obtained are shown in Fig. 10-8. The effect of the
Li2SO4 addition is to decrease the conversion ratio; because of the rela-
tively high value assigned to plutonium, this causes the Li2SO4 addition
(added in the same molal concentration as the UO2S04) to increase fuel
costs by 0.8 mill/kwh over those if no Li2SO4 were required.

Calculations indicate .that batch-type operation of a plutonium pro-
ducer is undesirable if plutonium has a value of $40/g; any savings in fuel-
processing costs using batch-type operation is more than compensated by
a loss in product value associated with the decrease in plutonium pro-
duction.

10-5.2 Two-region reactors [15]. Since the conversion ratio is greater
in a two-region reactor, it is expected that a high plutonium value will
cause this reactor type to have lower fuel costs (greater fuel credits) than
a one-region reactor. Calculations for a reactor having a 6-ft-diameter core
and an over-all diameter of 12 ft, in which the plutonium is recovered with
a fuel cycle similar to that given in Fig. 10-7, indicated fuel costs from
0.5 to 0.6 mill/kwh lower than in one-region reactors.

10-6. FuerL Costs iINn U?23%° BurNER REACTORS

A homogeneous reactor fueled with a dilute, highly enriched U02S04
solution is potentially capable of operating without removing fission
540 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

 

UO2SO4 — LioSO4 — D50 Fuel Solution

 

Net Fuel Cost, mills/kwhr

~1.4 —

 

 

UO 2SO 4 — D9O Fuel Solution
] | |

| |
200 300 400 500 600 700 800

—2.2

 

 

23
U 8 Concentration, g/liter

Fic. 10-8. Effect of LioSO4 and U238 concentration on fuel cost of a one-region
spherical plutonium-producer power reactor. Diameter = 12 ft, avg. reactor tem-
perature = 280°C, electrical power = 125 Mw, heat generation =480 Mw, avg.
lithium cross section = 0.2 barn, plutonium credit = $40/g, inventory charge = 49,
molar ratio of Li2aSO4 to UO2S04 = 1, processing rate = 1000 g U235/day.

products if additional U235 is continually added to offset the buildup of
poisons.

The results of fuel-cost studies [16,17] for spherical one-region reactors
containing dilute, highly enriched UO2S04 in either D20 or H20 are given
in Table 10-12. The noble gases are assumed to be removed continuously.
The removal of poisons and corrosion products by hydroclones is con-
sidered, but the fuel cost is about the same whether hydroclones are used
or not (assuming corrosion rates of about 1 mil/year). The results are for
nonsteady-state conditions.
 

 

 

 

 

 

10-6] FUEL COSTS IN U23% BURNER REACTORS 541
TaBLE 10-12
IsoroPE CONCENTRATIONS AND FUEL CosT BREAKDOWN
FOR SOME U235 BURNERS
(Average temperature-280°C; 125 Mw elect.; 6-ft-diameter core)
Moderator H20 H-20 D20 D20
Hydroclone cycle time, days 1 o0 1 00
Operating time, days 2000 3400 4200 4400
Initial U235 inventory, kg 353 353 34.3 34.3
Total system volume, liters | 27,000 27,000 27,000 27,000
Average U?35 feed rate,
kg/yr 239 233 237
U concentration, g/liter* 24.7 36 18.6 22.8
Fraction poisons,*
2.(p)/2:(25) 0.05 0.17 0.23 0.53
Estimated costs, mills/kwh
Uranium inventory at 4%, 0.29 .03 0.03
D20 inventory and losses at
9% 0 0.19 0.19
Uranium feed cost 3.71 3.64 3.70
Chemical processing cost
(variable cost at end of
operating period) 0.08 0.03 0.03
| Hydroclone cost 0.02 0.02 0
Shipping cost 0.02 0.02 0.02
Plutonium sale (credit, |
$16/gm) 0.0005 0.01 0.0003
Diffusion plant (credit) 0.002 0.02 0.08
Total (fuel cost) T 4.13 4.06 3.91 3.90

 

*At indicated operating time.

tIncludes no fixed charges for fuel processing.

With D20 reactors, it appears that fuel processing can be economically
eliminated if the reactor operating cycle is about 10 years and if the spent
fuel is disposed of cheaply at the reactor site. Such a procedure with the
Hs0-moderated reactor is more expensive because of higher uranium
inventory. For low fuel-shipping and -processing charges, the fuel costs
are nearly independent of the reactor size and moderator, and are about
4 mills/kwh. If fixed charges associated with fuel processing are greater
than ~0.5 mill /kwh, it may be more economical to store the fuel than to

have it processed.

 
542 | HOMOGENEOQOUS REACTOR COST STUDIES [CHAP: 10

TAaBLE 10-13

SUMMARY OF FUEL CosTs oF DIFFERENT SYSTEMS

 

 

 

mills/kwh

(1) One-region burner; power only; UO3SO04~D20 solution

(for 14-year operation) 3.93
(2) One-region burner; power only; UO2SO4—H:0 solution

(for 7-year operation) 4.12
(3) One-region converter; power only; UO3-PuO,-D2O slurry | 2.20
(4) One-region converter; power-plutonium; UO:SO4,—D20

solution —2.00
(5) Two-region converter; power only; core, PuO2—D2O slurry;

blanket, UO3-PuO2-D20 slurry 1.90
(6) Two-region converter; power-plutonium; UQO2SO4-D20

solution . —2.66
(7) One-region breeder; power only; UO3—ThO3-D20 slurry 1.76
(8) Two-region breeder; power only; UO3—ThOs-D20 slurry 1.22
(9) Two-region breeder; power only; core, UO2S04—D20 solu-

tion; blanket, UO3—ThO2-D20 slurry 1.29

 

 

 

 

10-7. SuMMARY OoF HoMoGENEOUS REAcTOR FUEL-Cost CALCULATIONS

10-7.1 Equilibrium operating conditions. Fuel costs in different homo-
geneous systems are summarized in Table 10-13. Appropriate nuclear data
and general reactor characteristics have been given previously.

10-7.2 Nonsteady-state operating conditions [18]. A comparison of fuel
costs is given here for several one-region reactor systems, operating under
nonsteady-state conditions. The reactors are moderated with either H20 or
D20 and fueled with enriched UO3 plus ThOg2, or UO2SO4 of varying en-
richments. The effect of adding Li2SO4 to the UO2SOy4 is also considered.
In the UO3 + ThO: system, it is assumed that the initial fuel is U233 and
that sufficient U233 is available as feed material.

Cost factors which make only a small contribution to the total fuel cost
are fuel processing losses (0.29, of fuel processed), D20 losses (569, of D20
inventory/year), hydroclone costs ($70/day), 30-day inventory supplies,
and shipping charges (it is assumed that the amortization charges for
shipping costs are negligible). In all cases the inventory charges are based
on the volume of the reactor vessel plus the volume of the external system.
This latter volume is calculated on the basis of an average heat-removal
10-7] HOMOGENEOUS REACTOR FUEL-COST CALCULATIONS 543

 

 

 

 

 

4
{4, ,800)
S UO2504 —H20 (6, ,800)
-E [ o e
é 3" {6,»,300)
X UO2504 —D20O
= (10,500,300)
£2 [ (6,%,300) OW
.U-' 0 ! ! ‘b
S UO3 —ThO5 —D20
@
S, (12,500,300 _
@ =
0 | | l | L1 | |
100 200 500 1000

Reactor Power, Thermal Mw

F1a. 10-9. Fuel costs in single-region reactors as a function of power level. Values
in parentheses refer to core diameter (ft), fuel processing cycle time (days), and
fertile material concentration (g/liter) at near-optimum conditions. Inventory
charge = 49, relative chemical processing = 1, relative poisons = 3, 725 = 2.08,
719 = 1.93, 23 = 2.25.

capability of 20 kw/liter of external volume. The enrichment of the heavy
water 1s assumed to be 99.759, D20, and the D3O cost is taken as $28/1b.

Only the optimum or near-optimum reactor conditions are given in
Iig. 10-9. The optimum conditions refer to the diameter, fuel-processing
cycle time, and fertile-material concentration which give the minimum fuel
cost of all the diameters, cycle times, and fertile-material concentrations
studied for the particular case.

A value of unity for relative fuel processing implies a charge of $0.54/g
of U233, U234 U235 and U236 processed; $1/g Pu processed; and $3.50/kg
fertile material processed. A value of two implies processing charges twice
the above values. Fixed charges for chemical processing are not included
in the calculation of fuel cost. A relative poisoning of unity implies repre-
sentation of the fission-product poisons by two effective nuclei having
yields of 0.11 and 1.81 atoms/fission, and thermal absorption cross sections
of 132 and 13.9 barns, respectively (based on values of Robb et al. [19]).
A value of 1/2 implies cross sections 1/2 the above values, corresponding
closely to the values of Walker [20].

Fuel costs were calculated for both 10 and 25 years of reactor operation;
these costs are usually slightly lower after 25 than after 10 years, but the
differences are small, usually between 0 and 0.1 mill/kwh. Therefore, only
results for 10-year operation are given here.

The reactor system and power level influence the fuel cost as indicated
in Fig. 10-9. Changing the power level affects optimum reactor conditions
significantly. The variations in relative poisons and relative chemical
processing considered here did not change fuel costs to a large degree, the
individual effects usually being about 0.1 mill/kwh. The influence of re-
actor composition on fuel cost is due to the values of n for the various
044 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

 

 

25
| | l | |
Relative Poisons=1
= == === Relative Poisons = ,, Relative Chemical Processing=1
(12,200)
Q
>
o
20 < —
-
2
[
>
£
€ (10,300) i

-
\
\

Fuel Cost, mills/kw-hr
o

Y
4% Inventory Charge

   
 

 

 

 

 

é!10:500) (12,200) )

v o

@ ° -

\\ ./ - -
\\ —’.” -
- -
(12,300)
o 1 | | | |
0 0.002 0.004 0.006 0.008 0.010 0.012

Fuel Processing Rate, Fraction of Reactor Inventory Processed/Day

Fia. 10-10. Effects of fuel-processing rate, fuel-processing charge, and poisoning
by fission products upon fuel cost in UO3-ThO2-D20 reactors. Values in paren-
theses refer to core diameter (ft) and thorium concentration (g/liter) at near-
optimum conditions (independent of relative poisons and relative chemical process-
ing values shown). Reactor power = 500 Mw, 723 = 2.25; U233 feed.

fissionable materials and fraction poisons associated with the different
systems. Increasing the inventory charge from 4 to 12% increases fuel
cost by about 0.5 mill/kwh for the D2O systems. Although not shown,
the effect of LioSO4 addition to the U02S04~H>0 system has negligible
effect upon fuel cost, owing to the high poison fraction associated with the
H20. The addition in equimolar proportions of Li2SO4 to the UO2504-D20
system increases fuel costs by about 0.1 mill/kwh.

The influences of fuel-processing rate, fuel-processing charge, and
fission-product poisoning upon fuel cost are shown in Fig. 10-10 for the
UO3-ThO2-D20O system. The fuel-processing rate corresponds to the
fraction of the reactor inventory processed per day. It is seen that doubling
the fission-product poisoning increases the fuel cost about 0.1 mill/kwh for
10-8] CAPITAL COSTS FOR LARGE-SCALE PLANTS 545

the fuel-processing rates considered, and that doubling the processing
charge increases the fuel cost about 0.1 mill/kwh at optimum conditions.

The fuel costs given in Figs. 10-9 and 10-10 include no fixed charges
for fuel processing; the magnitude of these charges would be dependent
upon the number of reactors processed by the processing plant. Increasing
the fuel-processing charges decreases the optimum fuel-processing rate.
Although fuel would undoubtedly be processed at the end of reactor
operation (and this was always assumed in obtaining the results), the per-
missible unit cost for processing at that time is high compared with the
permissible unit cost associated with a high processing rate. If the fixed
charges for processing correspond only to that period required to process
the fuel (for a given processing-plant capacity), then the optimum process-
ing rate is less than the rate obtained on the basis of fixed charges being
independent of fuel processing rate.

10-8. CaritaL CosTs FOR LARGE-SCALE PLANTS

The homogeneous reactors that have been built are small and have in-
vestment costs of over $1000/electrical kw. The capital costs of large
nuclear plants, based largely on paper studies, have been estimated to be
1.2 to 2 times that of conventional power-cost investments. If capital costs
are $400/kw, corresponding fixed charges are 9 mills/kw at 80% load
factor and 12 mills/kw at 60% load factor, considering a 15% annual fixed
charge. It follows that one of the conditions necessary for low-cost power
is a high load factor. The essential problems are to achieve low fuel costs
and to obtain reliability and long plant lifetime. To date, little experience
has been obtained in operating power reactors; therefore it is difficult to
estimate accurately the lifetime and also the reactor reliability over that
lifetime. Although most economic studies are used to measure the relative
economic advantage of different types of reactors, this cannot be firmly
established so long as the investment and maintenance costs remain un-
certain. Consequently, no hard and fast conclusions concerning power
costs can be obtained other than what the ultimate power cost might be.

The initial cost of the reactor building and other buildings associated
with the plant, as well as the site development costs, are strongly influ-
enced by the philosophy behind the design with regard to such factors as
safety and security. Estimates for these costs are usually based on cor-
responding costs for a large conventional steam power plant. Land costs
for nuclear plants are extravagantly high if containment is not assured.
The cost of containment vessels has been estimated to be in the range of
$10 to $40/electrical kw [21]. Equipment costs usually assume that the
desired equipment has been developed and all development costs paid for.
Even so, the depreciation charges on reactor equipment may be considerably
546 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

TaBLE 10-14

PRESENT ESTIMATES OF REACTOR PLANT CosTS FOR A THREE-REACTOR
STATION OPERATING AT A ToraAL PoweRr oF 1350 THERMAL Mw

(315 Electrical Mw) [23]

 

 

 

 

 

 

Min. Cost Max. Cost
High-pressure system for one reactor
Reactor vessel (and core tank) 530,000 970,000
Gas separators (3) 120,000 240,000
Heat exchangers (3) 2,030,000 2,680,000
Fuel and blanket circulating
pumps (3) 670,000 720,000
High-pressure storage tank and
catalytic recombiner, core system 110,000 | 220,000
High-pressure storage tank and
catalytic recombiner, blanket
system 40,000 90,000
20-Mw gas condenser 40,000 70,000
6.5-Mw gas condenser 20,000 30,000
Condensate storage tanks (2) 10,000 20,000
Gas blower for core system 30,000 60,000
Gas blower for blanket system 20,000 40,000
High-pressure process piping and
valves 3,250,000 6,430,000
Steam piping, valves and expansion
joints in cells © ©
Instrumentation 330,000 700,000
Sampling equipment (b) ®)
Installation of high-pressure equip-
ment (foundations, supports,
erection, etc.) 1,430,000 2,210,000
Subtotal for one reactor 8,630,000 14,500,000
Subtotal for three reactors 25,900,000 43,500,000

 

(a) Includes installation, insulation, inspection.

(b) In low-pressure estimate.
(¢) In steam-system estimate.
(d) With no contingency.

 
 

 

 

 

10-8] CAPITAL COSTS FOR LARGE-SCALE PLANTS H47
TABLE 10-14 (continued)
Min. Cost Max. Cost
Low-pressure system for one reactor
Dump tanks (6) and associated
equipment. [Condensers (2),
condensate tanks (2), recombiners
(2), evaporators (2), feed and
circulating pumps, D20 recovery
and fisston-product adsorption
system] 1,260,000 1,800,000
Piping and valves, instrumentation,
sampling equipment 1,080,000 2,270,000
Installation of low-pressure equip-
ment 540,000 770,000
Subtotal for one reactor 2,880,000 4,840,000
Subtotal for three reactors 8,640,000 14,500,000
Reactor structure
Reactor and low-pressure equipment
cells 3,000,000 4,000,000
Equipment transport shield, crane,
maintenance handling equipment 430,000 950,000
Control room, laboratory, and proc-
ess area - 2,000,000 4,000,000
Cell ventilation system (cooling
water, waste disposal) 890,000 1,750,000
Subtotal $6,320,000 $10,700,000
Summary Cost
Direct cost $40,900,000 $68,700,000
Contractor’s overhead and fees
(27%) 11,000,000 18,500,000
Engineering and inspection (10%) 4,100,000 6,900,000
Total® 56,000,000 94,100,000
Reactor plant cost, $/thermal kw 56,000,000 __ 94,100,000 __ 20
1,350,000 1,350,000
Reactor plant cost, $/elec. kw 56,000,000 94,100,000 __
315,000 315,000

 

 

 

 

 
548

Fia.

HOMOGENEOQOUS REACTOR COST STUDIES [caAP. 10

44

 

——t

W

o
w

* ¢ 1120
>
c
9
o
& —{110
o
0
o
&
® 100
Z

Turbogenerator Plant Cost, dollars/elec.k

 

 

 

300 500 700 900 1100
Steam Temperature, °F

 

10-11. Turbine plant cost and net station efficiency vs. steam temperature.

 

T l Tl

   
    

Present Technology

 

Unit Cost, $/elec.kw

  

 

  

L 1 1 1 11]] I L 11 111}

10 30 60 100 300 600 10¢
Net Electrical Capacity, MW

F1a. 10-12. Capital costs of aqueous homogeneous reactors.
10-9] OPERATING COSTS IN LARGE-SCALE PLANTS 549

higher than those for conventional power plants. Also, the insurance costs
for a nuclear plant may be much higher.

Conceptual designs for one- and two-region homogeneous reactors con-
sider stainless steel and stainless-steel-clad carbon steel as materials of con-
struction. Either type reactor will require pressure vessels, gas separators,
steam generators, recombiners, sealed motors, pumps, storage vessels,
valves, and piping. Preliminary estimates [11] of capital investments show
no significant difference in costs between one- and two-region reactor plants.

Engineering considerations place limits upon the operating pressure and
the size of the pressure shell. A spherical reactor shape is desirable, inas-
much as the neutron leakage and the required shell thickness for a given
reactor diameter are thus minimized for a given reactor volume. Fabri-
cators of pressure vessels agree that even though very large shells
can be made, the smaller diameter vessels are more feasible. Because of
the temperature limitation associated with the pressure limitation, satu-
rated steam at relatively low pressures is generated in the heat exchangers
(see Chapter 9).

Although analyses of the turbine plant indicate that thermal efficiency
improves as throttle pressure increases, the reactor system investment
costs rise sharply with increased operating pressure. The effect of saturated-
steam temperature upon turbine plant cost and net station efficiencies is
given [22] in Fig. 10-11.

In estimating reactor plant costs, it is necessary to determine the cost of
the various items of equipment. Present estimates [23] of reactor plant
costs for a large reactor station are given in Table 10-14. The equipment
costs are based on per-pound costs of HRE-2 equipment, cost estimates
obtained from industry, and the assumption that costs are directly propor-
tional to (reactor power) %6, The effect of station size upon reactor station
costs are estimated [23] in Table 10-15, while the possible effect of technical
advances upon capital investment costs is indicated [22] in Fig. 10-12.
In all these estimates, a developed and operable system is postulated.

10-9. OPERATING AND MAINTENANCE CosTs IN LARGE-SCALE PLANTS

Because of the high level of radioactivity associated with nuclear plants,
maintenance and repair of nuclear systems will have to be done remotely.
This type of operation is, in general, more time-consuming and expensive
than methods used in maintaining conventional coal-fired power plants.
To minimize operating and maintenance (O and M) costs, it will be
necessary to design reactor plants so as to simplify maintenance problems;
however, construction costs associated with such design will be higher than
the costs of huilding a reactor plant in which O and M costs are not so
economically significant. These higher costs are associated with the ability
550 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

TaBLE 10-15

PRESENT EsTIMATES oF HoM0GENEOUS REACTOR POWER STATION
CosTs As A FuncTioN oF StaTiON Si1zE [23]

 

Plant size: |
Thermal Mw 1350 240 60
Electrical Mw 315 60 15

 

Min. Max. | Min. Max. | Min. Max.

 

 

Direct cost of reactor
plant,
millions of dollars 41 69 16 27 7.6 13.5

Engineering and design

(at 159%,),
millions of dollars 6 10 2.4 4 1.1 2.0

Contractor overhead and
fees (at 239),
millions of dollars 9 16 3.6 6 1.8 3.1

Total cost reactor plant,
millions of dollars, 56 95 | 22 37| 10.5 18.6
$/elec. kw 180 300 | 370 620 | 700 1240

Turbine-generator plant,
millions of dollars, 33.4 7.0 2.0
$/elec. kw 100 120 130

Reactor station total cost, |
millions of dollars, 90 128 | 29 44 | 12.5 20.6
$/elec. kw 280 400 | 480 730 | 830 1370

 

 

 

 

 

 

to be able to inspect, repair, or replace components in a high radiation
field; this requires component compartmentalization and agccessibility.
Operating and maintenance costs cannot be predicted accurately be-
cause of the lack of knowledge and experience; however, the information
availlable [24] indicates that O and M costs may run as high as 4 mills/kwh
in the first plants constructed. Difficulties associated with predicting O and
M costs concern predicting component lifetimes, component repair/discard
ratio, maintenance procedures, and downtime required for maintenance.
Also, detailed design studies of various maintenance schemes are required
before these various schemes can be fully evaluated. Based on present
551

OPERATING COSTS IN LARGE-SCALE PLANTS

10-9]

"Tomod _@.oflsoofio MIA CZ1
Suronpoad pus D 087 3% Furpeiado swe)sAs SNOSUITOWOY 10§ SUOISUSWIP I030BII PUB SUOIPBIJUIOUOD [9NJ WINLIQIINDY 10T DI

192npoiyd wniuoinid vOmng uoibayj-auQ Aunig wniuoin|4-wniupin) uoibay-auQ Auinig wniioy) uoibay-sup

S1941] 000'0S = SWN|OA WeysAG |D4o | s1941| 000’05 = SWwNn|oA WeysAg |Djo | S1341) 000’99 = sum|oA weisAg (4o |

\ Yos®on /
[ = ———— oy ooy
Yos%i

 

 

 

otqu otq u

z
(*ond wouvo>o.=2 nd) tond so sy ng b g sam1/41 6 06T
oguw

+£0on so seuy/g.,.N B 00Z + 304 /ee N B L0
Y0s%0N s° 4941/g,N B 0OV 8EC ¢

+ 394 /e 1 B | + 91 /ee,N B89
/ + /e N BSL SET \ £€Z

 

 

 

 

 

 

 

 

 

 

- OIp ¥-Z1 DIP K4-Z1 oip H-¥1
(Buiss@d0id |any ou ‘:ado 14-g| 10} 31D "dU0D) 9100 Aunjg 9407 uoyn|og
Jsuing nmND uoibay-auQ ‘19pasag uoibay-om | ‘19pasug uoibay-om |
S1341| 000’/ = dwn|oA waysAg [pio) S1941| 000 'Oy = dwn|oA waysAg |pjo| S1941] 000'0p = SWNjoA weisAg |ojo)

 

 

 
  

0°Q "4941/g¢,N B ¥ o411 41 6 ggo) 0°q “eui/uL B 000l

 

 

z
otq u 0°a "e41/41 6 002 0“a ""os®on s
¥ el
0S“0ON se 434} /ce,N B ¥ +&on so 194 /6N B L0 194 /ee N6 T°0
10 17 + .—Q.———\MMND 6 v'6 47 + ho.—_—\flflND B Tl

 

 

 

 

 

 

 

 

O%H u!
Yoston se B /e 6y . B /e N 6y
IP 2102 §4-9 —‘.||o_1 @102 &-o“_
[ 0P 49— OIP H4-0L -——————OIp Y- ™

 
HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

552

"$9800 douBUUIBW pue Furyerad()l
"9[qIsBOJ oI SWIAYSAS [on] 9y} Jurwnsse ‘AFo[ouyoe) Jussasd U0 PoasBY :PUISAI ],

 

 

 

 

 

 

 

 

 

6'8¥L | ¥II-T'TI -1 7 1'% 8V¢¥% | £01-02 ALmys €0
Uo1391-8U()
8'8-€°L | £9I-0'1I -1 7C 0% 8Vv-€¥% | £01-0L ALmys 20YL,
Uuo13a1-9u()
L OI-1°6 9 L1-9°¢l ¢ 1 V¢ 'y 9V 07 ¢6<¢9 O%H + ge2)
uor3ar-au()
¢0I6'8 ¥ LIV ¢l ¢ 1 V¢ 6 ¢ 9707 G699 0%d + gez)
uor3ar-au()
€8L9 € 91-8'01 1 V¢ eI 0S¥V 0 11-6¢L 9100 Armys
UOI3aI-0M T,
¥ 889 ¥ 91-6 01 ¢ 1 ¥-¢ vl 0¢¥v¥ 0 1I1-¢°L 9100 uonN[OS
UOLFOI-0M T,
arnnyg JUIsdIJ aInng JUIsAL ] arnng «JU9saIJ
Ay wW)SAs ong
UMY /s[[ru yay/s[ru 59500 [Ny ymy/sqrru
‘s1s00 1omod 8307, ‘N pue O ‘0/,GT 78 s9318YD PoXI]

 

 

 

 

 

[22] (D087 - 10308} PBO[ %08 M\ [BWIdY} ()G ‘ MIA [BOLIJOS[d GZT)

SHOLOVAY SNOUNADOWOH $N0FANDY ATVOS-ADUVT NI SLSO) HIMO]

91-01 ¥1dVv],

 
10-10] SUMMARY OF ESTIMATED POWER COSTS 553

technology and a feasible reactor system, it is estimated that these costs
will be 2 to 4 mills/kwh.

As more experience is gained in maintaining plants and in designing for
O and M, it is expected that these costs will decrease; even so, because of
the nature of the problems, O and M costs in future plants will prob-
ably be 1 to 2 mills/kwh, or about two or three times those associated with
a conventional coal-fired plant.

10-10. SumMARY OF EsTIMATED PoweR CosTs

The power cost is the sum of the fixed charges on capital investment,
fuel costs, and operating and maintenance costs. Figure 10-13 specifies the
reactor systems considered, along with typical fuel concentrations and re-
actor dimensions. Estimates of the power cost for these systems are given
in Table 10-16, and are based on operation at 280°C and a power level of
125 electrical Mw (500 thermal Mw). Although operating and maintenance
costs are undoubtedly different for the various systems, it is assumed that
the range considered covers the differences involved.

The design power of a reactor plant will markedly influence power costs,
primarily because the investment cost per unit power is a function of power
level. Table 10-17 indicates the influence of power level on power costs,
by comparing costs for U235 burner-type reactors having power outputs of
125 and 10 electrical Mw, respectively.

TABLE 10-17

INFLUENCE OF PowER LEVEL UPON “PRESENT’ PowkEr CoSTS IN
U235 BURNERS (259, THERMAL EFF.; 809, LoAD FAcCTOR; 280°C)

 

 

 

 

Electrical power level, Mw
10 125
Fixed charges at 159, /yr, mills/kwh 16 8
Operating and maintenance, mills/kwh 5 3
Fuel costs, mills/kwh 5 4
Total power costs, mills/kwh 26 15

 

 

 

 

 
554 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

REFERENCES

1. White House press release and AEC press release, Nov. 18, 1956; AEC
press release No. 1060, May 18, 1957; AEC press release No. 1245, Dec. 27,
1957; AEC press release A-47, Mar. 12, 1958.

2. U. S. Federal Register, Mar. 12, 1957, Vol. 22, p. 1591.

3. K. CoHEN, The Theory of Isotope Separation as Applied to the Large-Scale
Production of U235 National Nuclear Energy Series, Division III, Volume 1B.
New York: McGraw-Hill Book Co., Inc., 1951.

4. J. A. LaNE, The Economics of Nuclear Power, in Proceedings of the Inter-
national Conference on the Peaceful Uses of Atomic Energy, Vol. 1. New York:
United Nations, 1956. (P 476, p. 309)

5. M. Benepict and T. H. Pigrorp, Nuclear Chemical Engineering. New
York: McGraw-Hill Book Co., Inc., 1957. (p. 403)

6. E. D. ArNowp et al., Preliminary Cost Estimation: Chemical Processing
and Fuel Costs for a Thermal Breeder Reactor Station, USAEC Report ORNL-
1761, Oak Ridge National Laboratory, Jan. 27, 1955.

7. S. GrassToNE and M. C. EpLunp, The Elements of Nuclear Reactor Theory,
New York: D. Van Nostrand Company, Inc., 1957. (p. 238 ff)

8. M. W. RoseENTHAL et al., Fuel Costs in Spherical Slurry Reactors, USAEC
Report ORNL-2313, Oak Ridge National Laboratory, Sept. 27, 1957.

9. D. C. HamirLtoN and P. R. KASTEN, Some Economic and Nuclear Character-
wstics of Cylindrical. Thorvum Breeder Reactors, USAEC Report ORNL-2165,
Oak Ridge National Laboratory, Oct. 11, 1956.

10. H. C. CraiBorNE and M. ToBias, Some Economic Aspects of Thorium
Breeder Reactors, USAEC Report ORNL-1810, Oak Ridge National Laboratory,
Oct. 27, 1955.

11. R. B. Briagags, Aqueous Homogeneous Reactors for Producing Central-
statton Power, USAEC Report ORNL-1642(Del.), Oak Ridge National Labora-
tory, May 21, 1954.

12. H. C. CraiBorNE and T. B. FowLkRr, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending July 31, 1956, USAEC Report
ORNL-1943, Oak Ridge National Laboratory, Aug. 9, 1955. (pp. 47-49)

13. H. C. CraiBorNE and T. B. FowLEr, Oak Ridge National Laboratory, in
Homogeneous Reactor Project Quarterly Progress Report, USAEC Reports ORNL-
2057(Del.), 1956 (pp. 63-65); ORNL-2148(Del.), 1956 (pp. 41-43).

14. H. C. CLAIBORNE, in Homogeneous Reactor Project Quarterly Progress
Report for the Period Ending Oct. 31, 1955, USAEC Report ORNL-2004(Del.),
Oak Ridge National Laboratory, Jan. 31, 1956. (pp. 53-59)

15. H. C. CraiBorNE and M. Togias, Oak Ridge National Laboratory, 1954.
Unpublished.

16. H. C. CLAaiBORNE and T. B. FowLER, in Homogeneous Reactor Project
Quarterly Progress Report for the Period Ending Apr. 30, 1956, USAEC Report
ORNL-2096, Oak Ridge National Laboratory, May 10, 1956. (pp. 57-59)

17. P. R. Kasten and H. C. CrLaiBorNE, Fuel Costs in Homogeneous U235
Burners, Nucleonics, 14(11), 88-91 (1956).
REFERENCES 5%%5)

18. P. R. KAsTEN et al., Fuel Costs in One-region Homogeneous Power Reactors,
USAEC Report ORNL-2341, Oak Ridge National Laboratory, Dec. 3, 1957.

19. W. L. RoBs et al., Fission—roduct Buildup in Long-burning Thermal
Reactors, Nucleonics 13(12), 30 (1955).

20. W. H. WALKER, Fission Product Poisoning, Report CRRP 626, Atomic
Energy of Canada, Ltd., Jan. 5, 1956.

21. J. C. Heapr, Cost Estimates for Reactor Containment, (Reactor Engineering
Division) Tech. Memo No. 13 (Revised), Argonne National Laboratory, October
1957.

22. J. A. LANE, Oak Ridge National Laboratory, personal communication.

23. M. I. Lunpin, Oak Ridge National Laboratory, personal communication.

24. Evaluation of a Homogeneous Reactor, Nucleonics 15(10), 64-70 (1957).
557

BIBLIOGRAPHY FOR PART 1

AERONUTRONIC SYSTEMS, INcC., A Selectron Study for an Advanced Engineering
Test Reactor, USAEC Report AECU-3478, Mar. 29, 1957.

ArnoLp, E. D. et al., Exploratory Study: Homogeneous Reactors as Gamma
Irradiation Sources, USAEC CF-56-6-107, Oak Ridge National Laboratory,
July 5, 1956.

BAkER, C. P. et al.,, Water Boiler, USAEC Report AECD-3063, Los Alamos
Scientific Laboratory, Sept. 4, 1944.

Beawn, S. E. et al.,, Status and Objectives—Homogeneous Reactor Project:
Summarties of Presentations to the Reactor Sub-committee of the General Advisory
Committee. USAEC Report CF-56-1-26(Del.), Oak Ridge National Laboratory,
Jan. 10, 1956.

BeaLy, S. E., Decontamination of the Homogeneous Reactor Experiment, USAEC
Report ORNL-1839, Oak Ridge National Laboratory, June 12, 1956.

BeavLL, S. E., Containment Problems in Aqueous Homogeneous Reactor Systems,
USAEC Report ORNL-2091, Oak Ridge National Laboratory, Aug. 8, 1956.

BeavLL, S. E., Homogeneous Reactor Experiment No. 2, presented at the Fourth
Annual Conference of the Atomic Industrial Forum, October 1957.

BeaLr, S. E., and R. W. JURGENSEN, Direct Maintenance Practices for the
HRT, presented at the Nuclear Engineering and Science Congress, March 1958;
also USAEC Report CF-58-4-101, Oak Ridge National Laboratory, 1958.

Brair, S. E., and S. VIsNER, Homogeneous Reactor Test. Summary Report
for the Advisory Committee on Reactor Safeguards, USAEC Report ORNL- .
1834, Oak Ridge National Laboratory, January 1955.

BeNTZzEN, F. L. et al., High-power Water Boiler, USAEC Report AECD-3065,
Los Alamos Scientific Laboratory, Sept. 19, 1945.

Borkowski, C. J., Design. Review Commatiee Report on the Homogeneous
Reactor Experiment, USAEC Report CF-55-6-53, Oak Ridge National Labora-
tory, June 7, 1955.

Brigas, R. B., Aqueous Homogeneous Reactors for Producing Central-station
Power, USAEC Report ORNL-1642(Del.), Oak Ridge National Laboratory,
May 21, 1954.

Bricas, R. B. (omp.), HRP Civilian Power Reactor Conference Held at Oak
Ridge National Laboratory May 1-2, 1957, TID-7540, Oak Ridge National
Laboratory, July 1957.

Brigaes, R. B., and J. A. SwarToUuT, Aqueous Homogeneous Power Reactors,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 3. New York: United Nations, 1956. (P 496, p. 175)

Busey, H. M., and R. P. HamMMmonD, Los Alamos Scientific Laboratory, Test-
Tube Reactor, in Nuclear Science and Technology (Extracts from Reactor Science
and Technology, Vol. 4), USAEC Report TID-2505(Del.), 1954. (pp. 163-166)

Carson, H. G., and L. H. LanorumM (Eds.), Prelvminary Discussion and Cost
Estimate for the Production of Central-station Power from an Aqueous Homogeneous
Reactor Utilvzing Thorium— Uranium-233, USAEC Report NPG-112, Nuclear
Power Group, Feb. 1, 1955.
558 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

CArTER, WiLLiaM L., Design Criteria for the HRT Chemical Plant, USAEC
Report CF-54-11-190, Oak Ridge National Laboratory, Nov. 24, 1954.

Curisty, R. F., Theoretical Discusston of a Small Homogeneous Enriched
Reactor, USAEC Report MDDC-72, Los Alamos Scientific Laboratory, June 18,
1946.

CoLLIER, D. M. et al., The HRE Simulator. An Analog Computer for Solving
the Kinetic Equations of a Homogeneous Reactor, USAEC Report ORNL-1572,
Oak Ridge National Laboratory, Sept. 24, 1954.

CoMMONWEALTH EpisoN CompaNY and Pusric SErvicE Company, A Third
Report on the Feasibility of Power Generation Using Nuclear Energy, USAEC
Report CEPS-1121, June 15, 1953.

CroxtoN, F. E., List of References on Homogeneous Reactors, USAEC TID-299,
Technical Information Service Extension, AEC, Mar. 15, 1950.

DRrAPER, B. D., Mainenance of Vartous Reactor Types, USAEC Report CF-57-
4-92, Oak Ridge National Laboratory, Apr. 8, 1957.

Fax, D. H. (Ed.), Proposed 80,000 kw Homogeneous Reactor Plant. Process
and Plant Description, USAEC Report WIAP-9, Westinghouse Electric Cor-
poration, Industrial Atomic Power Group. February 1955.

FromMman, D. et al., Los Alamos Power Reactor Experiments, in Proceedings of
the International Conference on the Peaceful Uses of Atomic Energy, Vol. 3. New
York: United Nations, 1956. (P 500, p. 283)

GENERAL ELecTRIc CoMmPANY, ATOoMIC POowER EQUIPMENT DEPARTMENT,
Homogeneous Circulation Fuel Reactor Power Plant: Conceptual Design Study
Report, USAEC Report GEAP-2(Del.), May 31, 1955.

GUERON, J. et al., The Economacs of Nuclear Power, Vol. I. London: Pergamon
Press, 1956.

HauseNgrEeIcH, P. N., Calculation for Thorium- and Uranium-fueled Reactors,
USAEC Report CF-53-12-1, Oak.Ridge National Laboratory, Feb. 8, 1954.

HavuBeNREICH, P. N., Esttmation of Most Economical Velocity in Reactor Cir-
culating Systems, USAEC Report CF-54-5-26, Oak Ridge National Laboratory,
May 5, 1954.

Hus, KENNETH A., Discusston of Homogeneous Reactor Possibilities, USAEC
Report AECU-3336, Internuclear Company, May 16, 1956.

JENKS, G. H., Effect of Radiation on the Corrosion of Zircaloy-2, USAEC Report
CF-57-9-11, Oak Ridge National Laboratory, Sept. 30, 1957.

KasTeEN, P. R., Homogeneous Reactor Safety, USAEC Report CF-55-5-51,
Oak Ridge National Laboratory, May 9, 1955.

KasTEN, P. R., Stability of Homogeneous Reactors, USAEC Report CF-55-5-163,
Oak Ridge National Laboratory, May 25, 1955.

KasteEN, P. R., Dyanmics of the Homogeneous Reactor Test, USAEC Report
ORNL-2072, Oak Ridge National Laboratory, June 7, 1956.

KasTeN, P. R., Time Behavior of Fuel Concentrations in Single-region Reactors
Contarning U-233, U-235, Th-232, and Fission-product Poisons, USAEC Report
CF-57-2-120, Oak Ridge National Laboratory, Feb. 26, 1957.

KasTeEN, P. R. et al.,, Aqueous Homogeneous Research Reactor—Feasibility
Study, USAEC Report ORNL-2256, Oak Ridge National Laboratory, May 9,
1957.
BIBLIOGRAPHY FOR PART I 559

King, L. D. P. (Comp.), Los Alamos Power Reactor Experiment and its Assoct-
ated Hazards, USAEC Report LAMS-1611(Del.), Los Alamos Scientific Labora-
tory, Dec. 2, 1953.

King, L. D. P., Design and Description of Water Boiler Reactors, in Proceedings
of the International Conférence on the Peaceful Uses of Atomic Energy, Vol. 2.
New York: United Nations, 1956. (P 488, p. 372)

LANE, J. A., Some Long Range Aspects of Homogeneous Reactors, in Nuclear
Science and Technology (Extracts from Reactor Science and Technology, Vol. 1,
1951) TID-2502(Del.) (pp. 27-42)

LANE, J. A. et al., Aqueous Fuel Systems, USAEC Report CF-57-12-49, Oak
Ridge National Laboratory, December 1957.

LirTLE, INc., A. D., Gas Pressurization of Homogeneous Reactor Test, USAEC
Report AECU-3349, Feb. 14, 1956.

Los Aramos SciEnTiFic LABORATORY, An Enriched Homogeneous Nuclear
Reactor, USAEC Report AECD-3059, Jan. 25, 1951; also Rev. Sci. Instruments
22, p. 489-499 (1951).

LunpiN, M. 1., Economics of Power Generation—Homogeneous Research Reac-
tor, USAEC Report CF-56-6-29, Oak Ridge National Laboratory, June 4, 1956.

MaLLon, R. G. et al., Conceptual Design of an Advanced Engineering Test
Reactor, USAEC Report NY0-4849, Advanced Scientific Techniques Research
Associates, Mar. 1, 1957.

McBripE, J. P., Aqueous Slurries, USAEC Report CF-57-11-120, Oak Ridge
National Laboratory, Nov. 26, 1957.

McDurrig, H. F., Corrosion by Aqueous Reactor Fuel Solutions, USAEC Re-
port CF-56-11-72, Oak Ridge National Laboratory, Nov. 14, 1956.

Oaxk RipGe NaTionaL LaBoraTory, Homogeneous Reactor Project Quarterly
Progress Report: USAEC Reports ORNL-527(Del.), Dec. 28, 1949; ORNL-630,
Apr. 21, 1950; ORNL-730, July 6, 1950; ORNL-826, Oct. 24, 1950; ORNL-925,
Jan. 30, 1951; ORNL-990, May 18, 1951; ORNL-1057, Oct. 10, 1951; ORNL-
1121(Rev.), Jan. 9, 1952; ORNL-1221, Apr. 1, 1952; ORNL-1280, July 14, 1952;
ORNL-1318, Sept. 19, 1952; ORNL-1424(Del.), Jan. 10, 1953; ORNL-1478(Del.),
Mar. 3, 1953; ORNL-1554, July 10, 1953; ORNL-1605, Oct. 20, 1953; ORNL-
1658, Feb. 25, 1954; ORNL-1678, Apr. 15, 1954; ORNL-1753(Del.), Sept. 17,
1954; ORNL-1813(Del.), Dec. 30, 1954; ORNL-1853, Feb. 16, 1955; ORNL-1895,
July 14, 1955; ORNL-1943, Aug. 9, 1955; ORNL-2004(Del.), Jan. 31, 1956;
ORNL-2057(Del.), Apr. 17, 1956 ; ORNL-2096, May 10, 1956; ORNL-2148(Del.),
Oct. 3, 1956; ORNL-2222, Feb. 7, 1957; ORNL-2272, Apr. 22, 1957; ORNL-2331,
Sept. 3, 1957; ORNL-2379, Oct. 10, 1957; ORNL-2432, Jan. 21, 1958; ORNL-
2493, 1958.

Oak RipGe OprEraTIONS OFFICE, RESEARCH AND DEVELOPMENT DIVISION,
Atomic EnNErGgY CommissioN, HRP Ciwvilian Power Reactor Conference Held
at Oak Ridge, March 21-22, 1956, USAEC Report TID-7524, March 1956.

Oak RipGE ScHoOL oF REACTOR TECHNOLOGY, Reactor Studies:
CuarmaN, R. H., An 80-Megawatt Aqueous Homogeneous Burner Reactor,
ORNL-CF-57-8-6, in preparation.

Crark, P. R., Homogeneous Reactor for Ship Propulsion, USAEC Report

CF-55-8-191, August 1955.
060 HOMOGENEOUS REACTOR COST STUDIES [cHAP. 10

Hauspure, A., High-temperature Aqueous Homogeneous Reactor, USAEC Re-
port CF-55-8-191, August 1955.

Kamack, H. F., Botling Homogeneous Reactor for Producing Power and Plu-
tontum, USAEC CF-54-8-238(Del.), August 1954.

MonTgoMERY, D. W. et al., 10-Megawatt Aqueous Homogeneous Circulating
Solution Reactor for Producing Electrical Power tn Remote Locations, USAEC
Report CF-53-10-22, August 1953.

PurNaMm, G. et al., Reactor Design and Feastbility Problem; U233 Power Breeder,
USAEC Report CF-51-8-213, Aug. 6, 1951.

RickerT, R. J. et al., 4 Preliminary Design Study of a 10-Mw Homogeneous
Boiling Reactor Power Package for Use in Remote Locations, USAEC Report
CF-53-10-23, Aug. 21, 1953.

Tuomas, R. A. et al., Ultimate Homogeneous Reactor, USAEC Report CF-54-
8-239, August 1954.

ZeituiN, H. R. et al., Boiling Homogeneous Reactor for Power and U-233
Production, USAEC Report CF-54-8-240, August 1954.

Paciric NorRTHWEST PowER GRouP, Aqueous Homogeneous Reactors, Second
Annual and Final Report to the Atomic Energy Commission, Report PNG-7,
February 1956.

PeeBLEs, FRED N., and HaroLp J. GARBER, The Combination of Hydrogen
and Ozygen in Platinum Catalyzed Flow Reactors, USAEC Report ORNL-1796,
University of Tennessee, Oct. 26, 1954.

Quartes, L. R., and W. P. WaLkER, A Manual of HRE Control and Instru-
mentation, USAEC Report ORNL-1094, Oak Ridge National Laboratory, Jan.
23, 1952.

ReacToR SAFEGUARD CoMMITTEE, AToMIc ENERGY COMMISSION, Report on
the Los Alamos Homogeneous Enriched U?3% Reactor; Meeting Held at Los Alamos
on Feb. 1-2, 1949, USAEC Report TID-5428, June 27, 1949.

RoBErTsON, R. C., Heat-Power Cycles and Prime Movers for Nuclear Power
Plants, USAEC Report CF-56-10-77, Oak Ridge National Laboratory, Oct. 10,
1956.

RoseENTHAL, M. W., Natural Circulation Homogeneous Reactors, USAEC Re-
port CF-56-10-53, Oak Ridge National Laboratory, Oct. 16, 1956.

RosENTHAL, M. W., and M. Tosias, Nuclear Characteristics of Two-region
Slurry Reactors, USAEC Report CF-56-12-82, Oak Ridge National Laboratory,
Dec. 20, 1956.

ScuwarTz, H., Natural Convection Cooling of Liquid Homogeneous Reactors,
North American Aviation, Inc., USAEC Report AECU-706, Dec. 20, 1949.

Secoy, C. H., Aqueous Fuel Systems, USAEC Report CF-57-2-139, Oak Ridge
National Laboratory, Feb. 28, 1957.

Secaser, C. L., and F. C. Zapp, HRE Design Manual, USAEC Report
TID-10082, Oak Ridge National Laboratory, Nov. 18, 1952.

STEIN, J. M., and P. R. KASTEN, Botling Reactors: A Preliminary Investigation,
USAEC Report ORNL-1062, Oak Ridge National Laboratory, Dec. 12, 1951.

TayLor, W. F., TBR Plant Turbogenerator System Study, USAEC Report
CF-56-7-127, Oak Ridge National Laboratory, June 27, 1956.
BIBLIOGRAPHY FOR PART I 561

ToBias, M., Breeding Reactors, USAEC Report CF-55-6-157, Oak Ridge,
National Laboratory, June 7, 1955.

ToBias, M., The HRT and the Walk-Away Problem, in Nuclear Science and
Technology, Vol. 2A. USAEC Report TID-2509, Oak Ridge National Laboratory,
1956. (pp. 83-96) (June 1956).

Toms, D. S., Jr., Instrumentation and Controls for the HRT, Nucleonics
15(2): 48-52 (1957).

TuomMmas, D. G., Solids Dispersed in Liquids, USAEC Report, CF-56-10-35,
Oak Ridge National Laboratory, Oct. 1, 1956.

Tromas, T. H. (Ed.), HRE Operations Manual, USAEC Report CF-53-1-94,
Oak Ridge National Laboratory, Oct. 9, 1952.

TrompsoN, W. E. (Ed.), Notes on the HRP Information Meeting, June 23, 1952,
USAEC Reports CF-52-7-104, Oak Ridge National Laboratory, July 21, 1952.

VIsSNER, S., and P. J. HaAuBeNrEIcH, HRE Experiment on Internal Recombi-
nation of Gas with a Homogeneous Catalyst, in Nuclear Science and Technology,
Vol. 1A, No. 1, USAEC Report TID-2506(Del.), Oak Ridge National Labora-
tory, 1955. (pp. 73-90)

WEEREN, H. O., HRT Hazards Report, Chemical Processing System, USAEC
Report CF-54-12-146, Oak Ridge National Laboratory, Dec. 10, 1954.

WEINBERG, A. M., Outline of Program for Design of Pilot Model of Homog-
eneous Power Reactor, USAEC Report CF-49-7-135(Rev.), Oak Ridge National
Laboratory, July 15, 1949.

WESTINGHOUSE ELECTRIC CORPORATION, INDUSTRIAL ATOMIC POWER GROUP,
Proposed 80,000 kw Homogeneous Reactor Plant. Cost Estimates, USAEC Report
WIAP-11, February 1955.

WinTERS, C. E. et al., Homogeneous Reactor Preliminary Process Design Re-
port, USAEC Reports AECD-3980 and ORNL-527(Del.), Oak Ridge National
Laboratory, Dec. 28, 1949.

Winters, C. E., and A. M. WEINBERG, Homogeneous Reactor Experiment
Feasibility Report, USAEC Report ORNL-730, Oak Ridge National Laboratory,
July 6, 1950. |

WinTERs, C. E., and A. M. WEINBERG, A Report on The Safety Aspects of the
Homogeneous Reactor Experiment, USAEC Report ORNL-731, Oak Ridge
National Laboratory, Aug. 29, 1950.

Zarp, F. C., and P. N. HauBeNREICH, Design Data for ISHR (2nd ed.),
USAEC Report CF-52-11-161, Oak Ridge National Laboratory, Nov. 8, 1952.

Zmora, P. C. et al., Power Removal from Boiling Homogeneous Reactors, in
Nuclear Science and Technology (Extracts from Reactor Science and Technology,
Vol. 4), USAEC Report TID-2505(Del.), Oak Ridge National Laboratory, 1954.
(pp. 47-72).