ORNL/TM-5783
Distribution
Category UC-76

Contract No. W-7405-eng-26 )

METALS AND CERAMICS DIVISION

COMPATIBILITY STUDIES OF POTENTIAL MOLTEN-SALT BREEDER
REACTOR MATERIALS IN MOLTEN FLUORIDE SALTS

J. R. Keiser

Date Published: May 1977

NOTICE —m ey
This report was prepared as an account of work
sponsored by the United States Government. Neither
the United States nor the United States Energy
Research and Development Administration, nor any of
their employees, nor any of their contractors,
subcontractors, or their employees, makes any
warranty, express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness
or uscfulness of any information, apparatus, product or
process disclosed, or represents that its use would not
infringe privately owned rights,

 

 

 

NOTICE This document contains information of a preliminary nature,
It is subject to revision or correction and therefore does not represent a

final report.

OAK RIDGE NATIONAL LABORATORY

Oak Ridge, Tennessee 37830 |

operated by ’ MA

UNION CARBIDE CORPORATION AT [ ”
for the

ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION

CUMENT IS UNLIMITED
DISTRIBUTION OF THIS DOCUMENT IS M g{“\
CONTENTS

ABSTRACT . . . ¢« v v ¢ v v 4 4 v & v o W
INTRODUCTION . . . . « « ¢« ¢« ¢ ¢ o« + &
EXPERIMENTAL METHODS . . . . . . .
EXPERIMENTAL RESULTS . . . . . . . .

Thermal-Convection Loop 21A .
Thermal-Convection Loop 23 . . . .
Thermal~Convection Loop 31 . .
Thermal-Convection Loops 18C and 24
Forced-Circulation Loop FCL-2B .
CONCLUSIONS . . . ¢ ¢ v v v v v ¢« o o &
ACKNOWLEDGMENTS . . . . « . . « « + « &

iii
COMPATIBILITY STUDIES OF POTENTIAL MOLTEN-SALT BREEDER
REACTOR MATERTALS IN MOLTEN FLUORIDE SALTS

J. R. Keiser

ABSTRACT

This report summarizes the molten fluoride salt compatibility
studies carried out during the period 1974—76 in support of
the Molten-Salt Reactor Program. Thermal-convection and forced-
circulation loops were used to measure the corrosion rate of
selected alloys. Results confirmed the relationship of time,
initial chromium concentration, and mass loss developed by
previous workers. The corrosion rates of Hastelloy N and
Hastelloy N modified by the addition of 1-3 wt % Nb were well
within the acceptable range for use in an MSBR.

 

INTRODUCTION

The purpose of this report is to summarize the corrosion studies
carried out for the Molten-Salt Reactor Program during the period
September 1974 through May 1976. These studies were intended to determine
the corrosion resistance of potential Molten-Salt Breeder Reactor contain-
ment vessel materials in molten fluoride salt. The nickel-base alloy
Hastelloy N was used successfully for the containment vessel of an
experimental molten-salt reactor, the Molten-Salt Reactor Experiment.
However, the discovery of irradiation embrittlement and grain boundary
embrittlement by the fission product tellurium led to a program to
develop an alloy that would be sufficiently resistant to the conditions.
To ensure that any new or modified alloy would have a high resistance
to corrosion by the fluoride salt, salt-metal corrosion studies were
made. Materials investigated include Hastelloy N, chromium and niobium
modifications of Hastelloy N, Inconel 601, and type 316 stainless steel.
The stainless steel was tested in LiF-BeF,; (65-35 mole %), and the other
alloys were tested in MSBR fuel salt, LiF-BeF;-ThF,-UF,

(72-16-11.7-0.3 mole 7).
o

Previous reserachers!»? have measured the corrosion resistance of
Hastelloy N and two types of stainless steel in various fluoride salt
mixtures., Their work has shown a multicomponent alloy is corroded by
the oxidation and removal of the least noble component. For Hastelloy N,
the least noble component is chromium. Fluoride salts can oxidize
chromium by reaction with impurities in the salt such as HF, NiF,, and
FeF, and by reaction with constituents of the salt. Impurity reactions

expected are

2HF + Cr = Cr¥-> + Hz (l)

and

FeFo + Cr = CrF» + Fe . (2)

The salt constituent UF, can give the reaction:

2UF, + Cr = CrF, + 2UF3; . (3)

If salt containing UF4 and a small amount of impurities is put into
a Hastelloy N system in which the salt circulates nonisothermally,
chromium will initially be removed from all parts of the system because
of reaction with both impurities and UF,. The impurity reactions are
expected to go to completion fairly rapidly so that they will have an
effect only on the short-time corrosion results. On the other hand,
since the equilibrium constant for reaction (3) is a function of tempera-

ture, this reaction provides a means for the continuous transfer of

 

'J. W. Koger, Alloy compatibility with LiF-BeF, Salts Containing
ThFy and UF, ORNL/TM-4286 (December 1972).

. ’J. H. DeVan, Effect of Alloying Additions on Corrosion Behavior of
Vickel-Molybdemum Alloys in Fused Fluoride Mixtures M.S. Thesis, University
of Tennessee, August 1960.
chromium from the hotter sections of the system to the cooler sections.
The amount of chromium, AY, removed from a unit area of surface can bhe

shown to be:

 

where B is a temperature-dependent constant, 'y is the initial concentra-
tion of chromium in the alloy, and I is the diffusivity of chromium in
the alloy. The limiting step for this mass transfer has been shown

to be the diffusion of chromium in the metal when Hastelloy N is the
alloy considered. If other strong fluoride formers — such as Ti, Nb,

or Al — are present in the alloy, mass transfer of these elements would

be expected to occur by the same mechanism as discussed above,

EXPERIMENTAL METHODS

Our corrosion studies have been carried out in five thermal-
convection loops and one forced-circulation locp. Figures 1 and 2
show schematic drawings of these loops. These loops circulate salt
around a syvstem across which a temperature gradient is maintained. For
most of these loops the temperature limits were maintained at 704 and
566°C (1300 and 1050°F), the proposed maximum and minimum temperatures
for the fuel salt of an MSBR. Other important features of these loops
are the removable corrosion specimens and the accesses to the salt to
permit insertion of electrodes for controlled-potential voltammetry.
Voltammetric measurements, which were made by members of the Analytical
Chemistry Division, allowed us to make on-line determinations of the
oxidation potential and corrosion product concentration of the salt.
Detailed explanations of voltammetry are available elsewhere.3s"

The operating conditions for each of the loops are described in Table 1.

 

M. W. Rosenthal, P. N. Haubenreich, P. B. Briggs, Comps., 7The
Development Status of Molten Salt Breeder Reactorse., ORNL-4812
(August 1972) pp. 153-57.

“J. R. Keiser, J. H. DeVan, and D. L. Manning, The Corrosion Resist-
ance of Type 31€ Stainless Steel to Li,BeF,, ORNL/TM-5782 (April 1977).
ORNL-DWG 68-398TR3

 
    
     
 
  

STANDPIPE

CLAMSHELL
HEATERS

 

 

 

 

SAMPLER

 
 

VALVES —

 

 

FLUSH
TANK DUMP

TANK

 

 

 

 

 

 

Fig. 1. Schematic of Thermal-Convection
Loop. Scale is 0.15 m. Height shown is 0.76 m.
CRNL-DWG 70-5632R
N

Ry

\i)
FREEZE VALVE |
(TYPICAL) —__

AIR
CORROSION SPECIMENS

  
  
    

(1175°F }
RESISTANCE HEATED SECTION NO.1

 
  
   
  
    
  
   
 
 
 

THERMOCOUPLE
WELL

Yo-in.OD x 0.042-in. WALL

BALL VALVE HASTELLOY N

  

HEATER LUGS (TYPICAL) RESISTANCE HEATED SECTION NO. 2

COOLER NO.4

   

FLOW RATE = ~4gpm
VELOCITY = ~10 fps IN Y%-in. TUBING
REYNOLDS NO. = 6600 TO 14,000

SALT PU

0w i
MP 1] 2

CORROSICON
SPECIMENS

FREEZE VALVES

 
  
   
 

CORROSION

COOLER NO.2 SPECIMENS
v

   
 
 

FiLL. AND DRAIN TANK
705°C (1300°F)

 
 

DRAIN AND FiLL LINE

(&\ s, 1/4-in.0D X 0.035-in. WALL
N

THERMOCOUPLE WELL

Fig. 2. Schematic of Molten-Salt Forced-Convection Corrosion Loop MSR-FCL-2 b.
Flow rate = 0.25 liter/sec; velocity ® 3 m/sec in 13-mm tubing; loop tubing is
13-mm OD by 1.07-mm wall; drain and f£ill line is 6.4~mm OD by 0.9-mm wall.
Table 1.

 

 

. Loop
Loop Material
21A Hastelloy N
23 Inconel 601
31 Type 316 Stainless Steel
24 Hastelloy N
i8C Hastelloy N
FCL-2B Hastelloy N

Loop Operating Conditions

 

 

 

 

Operating
oy Q
Speci?en Salt Yemperature, "C
Material e e e e
Max Min
Thermai-C- ve ! Zoun Leons

Hastelloy N, MSBR /04 366
1%-Nb-mod Hastelloy N Fuel

Inconel 601 MSER S04 566
Fuel

Graphite MSBR 677 550
Fuel

Type 316 Stainless Steel Li,BeF, 649 493

/%-Cr-mod Hastelloy N, MSBR 704 566
127-Cr-mod Hastelloy N, Fuel

3.47-Nb-mod Hastelloy N

10%2-Cr-mod Hastelloy N, MSBR 04 566

15%-Cr-mod Hastelloy N Fuel
Forced-Circualation Loop

Hastelloy N, MSBR 704 566

1%Z-Nb-mod Hastelloy N Fuel

Analytic method
development
Base—-line corrosion data

Measure corrosion rate
ot high-Cr alloy
(Inconel 601)

Investigate raphite
Uy Reaction

Measure corrosion rate
of type 316 stainless
steel in potential
coolant salt

Investigate effect of
Be addition to salt

Measure corrosiun rate
of modified Hastelloy N
alloys

Measure corrosion rate
of modifled Hastelloy N
alloys

Base-line ceorrosion data
in high-velocity salt

 

9
EXPERIMENTAL RESULTS
Thermal-Convection Loop 21A

Hastelloy N loop NCL 21A was the first loop to be put into operativa
when the Molten-Salt Reactor Program was resumed in 1974. As such, the
loop was used to obtain base-line corrosion data for Hastelloy N and to
provide a test bed for voltammetry measurements of MSBR fuel salt. The
voltammetry results showed that the oxidation potential as reflected
by the U(IV)/U(LIL) ratio remained quite high throughout the 17.5 months
of operation. In fact, with a U(IV)/U(III) ratio of about 10* loop 21A
contained the most oxidizing salt of all the loop experiments.

The first specimens used in this loop were made of Hastelloy N
and were removed for examination about every 2500 hr. Figure 3 shows
the weight change as a function of the exposure time for these specimens
for up to 10,000 hr in salt. From this figure the rate of the weight
change clearly decreased with time. Figure 4 shows that the change

in weight varies as the square root of time, as predicted from Eq. (4).

ORNL-DWG 77-3820

 

 

 

 

 

 

 

 

T
566°C i1
_-——'_Q'_-——-—-
1 /.——-—-——_-—F' l
635°C .
0
o~
E 4
)
o 4fi<\\
E N
: ®
Q-2 S
\.
~Jo4°C
\"

 

 

 

 

 

 

0 2000 4000 6000 BOOO 10,000
EXPOSURE TIME (hr)

Fig. 3. Weight Change vs Exposure Time for Hastelloy N Specimens
Exposed to MSBR Fuel Salt at 566, 635, and 704°C in Thermal-Convection

Loop 21A.
ORNL-DWG 76-4843

 

 

 

 

 

AM(mg/cn?)
°

 

 

 

 

 

 

 

 

 

0 20 40 60 80 100
4
(EXPOSURE TIME)”2 (VRp)

Fig. 4. Weight Change vs Square Root of Exposure Time for
Hastelloy N Specimen Exposed to MSBR Fuel Salt at 690°C in Thermal-
Convection Loop 21A.

Specimens that had been exposed for 7500 hr in the hottest and coldest
parts of the loop were examined metallographically. As is evident in
Fig. 5 the pitting on the higher temperature specimen was limited to
about 5 um, indicating that the effect of the salt was relatively mild.
Following the 10,000-~hr exposure of standard Hastelloy N, loop 21A
was used to test specimens of 1Z-Nb-modified Hastelloy N (experimental
heat 522). Only a short exposure was achieved with these specimens
before a power supply malfunction terminated the experiment, but the
short-time corrosion results (Table 2) compare very favorably with

results for standard Hastelloy N.
 

 

 

 

 

 

 

 

 

B Y-134311

=

*t

 

Fig. 5. Hastelloy N Exposed to MSBR Fuel Salt for 7500 hr at
(a) 704 and (b) 566°C. 500x.

Table 2. Hastelloy N Corrosion Rate Measurements from Loop 21A

 

 

 

Corrosion Rate, mg cmleyear‘l, at Each
Total - Exposure Temperature
Alloy Exposure . uXxposu npe
X "(hr) 566°C 635°C ,'_704°c
- ~ Standard 10,009 +1.17 +0.39  —3.09

-

1Z Nb Modified 1,004 0.0 —0.25 © =3.30

 
 

e e i iy

 

 

 

10

Théfm&léConvection'Loqp 23

The observation that the high-chromium alloy Inconel 601 (23 wt Z Cr)
resisted intergranular attack by tellurium led to the construction of a
thermalfconvection loop to determine how severe the corrosion by MSBR
fuel salt would be. After the new lobp was filled with salt, voltammetric
techniQues were used to follow the change in the U(IV)/U(III) ratio
as an indication of the extent of the initial reaction between chromium
and.UFu. The U(IV)/U(III) ratio decreased very rapidly, dropping to
about 40 within a fefirdays; meaning that considerable reaction was
probably occurring betweeh the salt and this Inconel 601 loop. Inconel
601 specimens exposed 721 hr all showed a weight loss, and that shown
bythe,hOttthképe¢ipen-was‘véiy iarge_(>30 mg cm~ > year~').. Furthermore,

the material lost by the hottest specimens was not removed umiformly from

the surface, but resulted in the formation of the porous surféce structure
shown in Fig. 6. As shown in Fig. 7; electron microprobe examination

of this specimen showed high thorium concentration in the pores. Since
the only known source of thorium.was the ThFy contained in the salt, the
salt likely penetrated the pores. Continuous line scans made with the
microprobe for the elements Ni, Cr, and Th, shown in Fig. 8, clearly show
the depletion of chromium near the surface. These results provide

further evidence of the presence of thorium in the pores. Diffusion
calculations provide another piece of evidence indicating that salt

must have penetrated the pores. Based on the bulk chromium concentration
of 23 wt %, microprobe measurements® determined a chromium concentration
of 6.6 wt Z at the surface of the specimen in Fig. 6. From these concentra-
tions, a depletion depth of 80 yum taken from Fig. 8, and diffusion values
taken from Evans, Koger, and DeVan,® we calculate that the exposure time
wofild have to be néarly 1000 times greater to attain this concentration
profile as a result of bulk diffusion alone at 704°C. To obtain this

 

SR. S. Crouse, ORNL, Private Communication, July 1975.

, SR. B. Evans III, J. W. Koger, and J. H. DeVan, Corrosion in
Polyt@ermal Loop Systems II. A Solid State Diffusion Mechanism With
and Without Liquid Film Effects, ORNL-4575, Vol. 2 (June 1971).
 

 

11

 

 

 

 

 

 

500x%,

at 704°C.

 Incone1 601 Exposed to MSBR Fuel Salt for 721 hr

6._

Fig.

Y-131294

 

 

 

Bac’ksCaflér‘ed_ Electrons

Xays

Electron-Beam Scanning Images of Inconel 601 Exposed to MSER

720 hr

ThMe

Fig

at 704°C.

~

-

7

Salts for
 

=y

 

i
'
'
t
:

 

 

 

12

ORNL-DWG 77-3819

 

 

NN
sl
N4

 

 

T
»

 

 

 

 

NORMALIZED CONCENTRATION

 

 

Th

L/ P2 e R

 

 

 

 

 

 

 

 

 

 

 

 

0 20 40 €0 80
DEPTH FROM SURFACE (um)

Fig. 8. Microprobe Continuous Line Scan Across Corroded Area in
Inconel 601 Exposed to MSBR Salt for 720 hr at 704°C.

profile with an exposure time of 721 hr the exposure temperature would
have to be about 1000°C, nearly 300°C higher than it was. Thus, to
establish the gradient that was observed, salt was most likely present
in the pores to provide a short—circuit'path for removal of the chromium.
Examination of a specimen from the coldest part of the loop revealed
surface deposits,'shown in Fig. 9, which were identified by microprobe
analysis as chromium. The conclusion from this test is that Ihconé1'601

would be unsuitable for use in a molten-salt breeder reactor.

 

 

 

Fig. 9. Inconel 601 Exposed to MSBR Fuel Salt for 721 hr at 570°C. 500x,
13

It is expected that the lower limit for the U(IV)/U(III) ratio

in an MSBR will be determined by the conditions under which the reaction
4UF3 + 2C = 3UF, + UC; (5)

proceeds to the right. Because the U(IV)/U(III) ratio of the salt

in loop 23 had decreased to less than 6, we decided to try to reproduce
the results of Toth and Gilpatrick,7 shown in Fig. 10, which predict

that UC, should be stable at the lowest temperatures that could be
maintained in this loop, 545-550°C. However, graphite specimens ex-
posed to the salt for 530 hr did not show any evidence of UC,. Since

the specimens used were made of pyrolytic graphite, the high density of
the material likely limited contact of the salt and graphite. The exper-

iment was repeated by exposing a less dense graphite for 530 hr,

 

L. M. Toth and L. O. Gilpatrick, The Equilibrium of Dilute UFj
Solutions Contained in Graphite, ORNL/TM-4056 (December 1972).

ORNL-DWG 7212321
TEMPERATURE {*C)
500 550 600 650 700

. /

i/

3 /I &

///i ]

 

 

o
o

]
o
o

 

o
N

i

H
- N
N
1
O
[¥7]

 

log,y @

 

R=UFy /{UFy + UF,)

\d
N

 

 

/T

 

 

 

 

 

 

 

 

130 125 420 445 110 105 100
1000/ (ok)

Fig. 10. Equilibrium Quotients, & = (UFa)H/(UFu)a, Versus Temperature
for UC, + 3UFq(d = 4UF3(q) + 2C in the Solvent LiF-BeF,-ThFy,
(72-16-12 mole Z;.
14

then checking the graphite surface for the presence of a new phase by
x-ray diffraction. A new phase was found, and it was tentatively identi-
fied by 0. B. Cavin as UO;. If indeed this compound was U0, it probably
resulted from a uranium fluoride-water reaction, but it could have come
from the hydrolysis of UC, that had been formed by reaction (5). On

the basis of information from L. M. Toth® that nucleation of UC, under
our operating conditions could be very slow, a longer exposure was
undertaken. Two specimens of the less dense graphite were exposed for
about 3000 hr at a minimum temperature of 555°C in salt in which the
U(IV)U/(III) ratio had dropped to about 4. However, x-ray analysis of
these specimens showed no evidence of a phase other than the salt and
graphite. No further investigations were carried out because of the

termination of the program.

Thermal-Convection Loop 31

 

Thermal-convection loop 31 is constructed of type 316 stainless steel,
has type 316 stainless steel specimens, and has been used for corrosion
measurements with one of the altermnative coolant salts, LiF-BeF:

(66=34 mole %). For the first 1000 hr of operation this loop was used

to gather base-line corrosion data with as-received salt, which contained
a relatively high concentration of impurity FeF;. As shown in Fig. 11,
fairly significant weight changes occurred in the specimens, especially
during the first 500 hr. Metallographic examination of specimens from
the hottest and coldest positions showed, respectively, pitting and
deposition, as is apparent in Fig. 12. Electron microprobe examination
of the deposits indicated they were predominately iron, and we expect
that the deposition occurred as a result of reaction (2). Bulk salt
analyses and voltammetric measurement of the FeF, and CrF, concentrations
of the salt during the first 1000 hr support this idea.

To learn if addition of a reductant to the salt would decrease
the impurity level and consequently lower the corrosion rate, beryllium
was added to the salt. New specimens were then inserted, and the corrosion

rate was measured for stainless steel in this "'reducing' salt. As long

 

®L. M. Toth, private communication.
 

 

15

 

Y-137150

ORNL-DWG 76-3496

 

    

q_—_

 

& AS RECEIVED" SALT

X AFTER BE ADDITION

    

—_..u._“__—q-—__-—_—

§ %

C

 

 

i |

 

3.0

 

q <
(bw) IONVHI LHOITM

 

11

Fig.

Weight Change vs Exposure Time for Type 316 Stainless

Steel Exposed to LiF-BeF,.

)
o
)
~F
o
T
»

 

 
 

 

 

 

Type 316 Stainless Steel Exposed to LiF-BeF, for 1000 hr
500x.

at (a) 650 and (b) 510°C.

Fig. 12,
16

as the beryllium rod was in the salt, the corrosion rate was extremely
low. This is shown in Fig. 11 for the first 500 hr after the addition

of beryllium. After removal of the beryllium, the specimen in the hottest
position showed a pattern of increasing weight loss as a function of

time. This most probably occurred because, once the source of beryllium
was removed, the species in the salt were no longer in equilibrium and

the salt became progressively more oxidizing with increasing time,
especially as moisture would leak into the system. Weight change results

for specimens exposed to the "reducing" salt are also shown in Fig. 11.

Thermal-Convection Loops 18C and 24

 

Hastelloy N loops 18C and 24 were used to determine the effect of
chromium concentration on the corrosion rate of Hastelloy N. Alloys
such as stainless steels and Inconels with a relatively high chromium
content had shown a resistance to grain boundary attack by tellurium
that seemed to be roughly proportional to chromium composition. However,
increasing the chromium content is, according to Eq. (4), expected to
increase the corrosion that will occur because of mass transfer, From
these tests and tellurium exposure tests we hoped to learn if there
is an optimum chromium concentration. Specimens were fabricated from
modified Hastelloy N alloys containing 7, 10, and 12% Cr. Each set of
specimens was exposed in one of the loops for a total of 1000 hr, with
weight change measured after 500 and 1000 hr. Within experimental error
the weight change results (Fig. 13) show the dependence on initial chromium
concentration that would be expected according to Eq. (4).

Following completion of the studies with the chromium-modified
alloys, loop 24 was used to measure the corrosion rate of 3.4%-Nb-modified
Hastelloy N. Results indicate a maximum corrosion rate of less than

2.20 mg cm—? year™! for a 1500-hr exposure.
17

ORNL-DWG 76-7233

 

 

 

 

03

=~ }

E |

s /

e |

: 0.2}

g r /

<

ST /

5 | /

W

* o1} /

i

A

: [/
o;llllgll_lll'lj;

0 3 6 9 12 15

CHROMIUM CONCENTRATION (wt %)

Fig. 13. Effect of Chromium Concentration in Modified Hastelloy N
on Weight Change After 1000 hr in MSBR Fuel Salt.

Forced-Circulation Loop FCL-2B

 

Forced-circulation loop FCL-2B was used initially for fuel salt
chemistry investigations. The first fuel salt corrosion investigations
were made with standard Hastelloy N in salt with a U(IV)/U(III) ratio of
about 100. Thus, the most significant differences between this loop
and thermal-~convection loop 21A were the oxidation potential of the
salt [Vv10% vs V10" in terms of U(IV)/U(III) ratio] and the velocity of
the salt 2.55 m/sec vs 1 m/min). If Eq. (4) and the ideas that led
to its development are correct, we would expect that (1) the only effect
of the low U(IV)/U(III) ratio in FCL-2B would probably be a reduction
in the initial corrosion compared with NCL 21A, and (2) the high salt
velocity would have no effect on mass transfer of chromium since the
limiting factor is the diffusion of chromium to the surface of the
alloy, not transfer of the chromium from the hot to the cold parts of the
system. The results of this study, shown in Table 3, indicate a very
low corrosion rate for the 4000-hr exposure. This test was interrupted
several times for repairs on the loop and for heat transfer measurements,

but that did not seem to have a detrimental effect on the results.
18

A 4000-hr corrosion test of Hastelloy N in fuel salt with a
U(IV)/U(III) ratio of 1000 was planned. Because of the decision to
terminate the Molten-Salt Reactor Program, we decided the remaining
operating time for this loop could be best spent measuring the corrosion
rate of 1/Z-Nb-modified Hastelloy N, an alloy that had shown good resistance
to tellurium attack. Accordinglv, a set of such specimens was prepared
and inserted into the loop. Because of loop cperating difficulties some
of the weight changes measured after 500 and 1000 hr were of questionable
value. Measurements made after 1500~ and 2200-hr exposure were of good
quality and are summarized in Table 3 along with some of the 500- and
1000-hr results. Included in this table for comparison are the results
for the standa~d Hastelloy N specimens exposed for 4309 hr in this same
salt and loop. It should be noted that the niobium-modified specimens
gained more weight than they lost. This is most probably due to mass
transfer of iron from the standard Hastellowv X tubing, which contains
about 5 wt % Fe, to the modified alloy, which has essentially no iron.
Taking into account the value at which the corrosion rate of the modified
Hastelloy N seems to be leveling out, we expect that the corrosion resis-
tance of the 1Z-Nb-modified Hastelloy N is at least as good as that of

standard Hastelloy N.

Table 3. Corrosion Rate Measurement for Hastelloy N
Specimens Exposed in FCL-2B

 

1

Corrosion Rate, mg cm™? year~ ", at Each Exposure Temperature

 

 

Alloy Exposure
566°C 635°C 704°C
Standard 4309 hr +0.02 —0.20 —2.31
1%7-Nb-Modified first 580 hr —0.52
next 497 hr —0.31
next 496 hr +1.69 +2.31 —0.38

next 669 hr —0.46 +(.91 —0.28

 
19

CONCLUSIONS

The following conclusions can be drawn from this work.

1. Voltammetry provides a very convenient means for on-line
measurement of the oxidation potential and impurity concentration of
the salt.

2. Inconel 601 would probably not have sufficient corrosion
resistance to be acceptable for use as a containment vessel material.

3. No evidence of the formation of uranium carbide was found
in studies of the reaction of graphite with very reducing salt.

4., Weight loss of Hastelloy N specimens occurred as a linear
function of the square root of the exposure time indicating diffusion-
controlled corrosion.

5. Type 316 stainless steel has a high initial corrosion rate
in as-received Li;BeF,, but has a very low corrosion rate when beryllium
is added to the salt.

6. Alloys of Hastelloy N modified by the addition of chromium
showed weight losses proportional to the chromium concentration.

7. Limited results indicate that the corrosion rates of 1- and
3.4%-Nb-modified Hastelloy N are at least as good as that of standard

Hastelloy N.

ACKNOWLEDGMENTS

The author wishes to gratefully acknowledge the following persons;
for their contributions; E. J. Lawrence for operation of the thermal-
convection loops, W. R. Huntley and H. E. Robertson for operation of
the forced-circulation loop, H. E. McCoy, J. R. DiStefano, and J. H. DeVan
for their advice and helpful discussions, and S. Peterson for editing

and Gail Golliher for preparing the manuscript.
37—39.

41-50.

79-80.

81—82.

83—196.

INTERNAL DISTRIBUTION

Central Research Library

Document Reference Section

21

Laboratory Records Department

Laboratory Records, ORNL RC

ORNL Patent Office
. F. Baes

. E. Bamberger
S. Bettis

. R. Brinkman

. Brynestad

. A. Canonico

. Cantor

Cavin
Clausing
Crowley
Culler
Cunningham
DeVan
DiStefano
Donnelly
Engel

. Ferris
Goodwin
Grimes
Guymon
Hightower, Jr.
Hill
Huntley
Keiser

gggmwznbuwuuqmuwomugomno

PR PPERRERRARERCCEE

EXTERNAL DISTRIBUTION

51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.

ORNL/TM-5783
Distribution
Category UC-76

A. D. Kelmers

R, E. MacPherson

. Mamantov

. L. Manning

. L. Matthews

. Maya

E. McCoy

J. McHargue

E. McNeese

Postma

Roche

Rosenthal

Savage

Selle

Silverman

Slaughter

Smith

. Toth

Trauger

Weir, Jr.

White

Young

Balluffi (consultant)
Brister (consultant)
Hibbard, Jr. (consultant)
Palmour III (consultant)
Promisel (consultant)
Stein (consultant)

*

?-93:§§q§n:214 OO

REUrpwREEUmOSR

*

URESHYSO0 G

i
M S
D

ERDA OAK RIDGE OPERATIONS OFFICE, P. 0. Box E, Oak Ridge, TN 37830

Director, Reactor Diyision

Research and Technical Support Division

ERDA DIVISION OF NUCLEAR RESEARCH AND APPLICATION, Washington, DC 20545

Director

ERDA TECHNICAL INFORMATION CENTER, Office of Information Services,

P. 0. Box 62, Oak Ridge, TN 37830

For distribution as shown in TID-4500 Distribution Category,
UC-76 (Molten Salt Reactor Technology).

7 U.S. GOVERNMENT PRINTING OFFICE: 1977-748-189/54