qg,fl% s s ST e e ;..']_: e 5 o

y
o

 

RECEIVED BY D“H... ocT 22 I968

 

" NOTICE This document .contains information of -a preliminary noture -
_ and wos prepared primarily for internal use at the Ock Ridge National =~
~ Laboratory. It is subject to revision or correction ond therefore does B

~ not represent a, flnal reporf

OAK RIDGE NATIONAL I.ABORATO
_ © operated by ‘_ S -
. UNION CARBIDE CORPORATION :
_NUCLEAR DIVISION -~ *: B -
B © " for the T |
U S A'I'OMIC ENERGY COMMISSION

. ORNI. TM 2256

  

copY NQ. -

DATE - June 20, 1968

- 'CHEMICAL FEASIBILITY OF FUELING
) MOLTEN SALT REACTORS WITH PuF3

R. E Thoma

 
 

 

 

 

 

 

 

 

 

S
%
..‘;tfl_r
S
g‘
.
] . o , LEGAL NOTICE -
; This report was pfoparod as an account of Government sponsored work. Neither the United States,
? _ ' nor the Commission, nor any person acting on behc_llf of the Commission:
. o A, Makes any warranty or representation, axpressed or implied, with respect to the accuracy,
; ' ’ _completeness, or usefulness of the information contained in this report, or that the use of
eny information, cpparatus, method, or process d:sclosad in this report may not infringe -
L privately owned rights; or :
G 8. Assumes ony liabilities with respact to the use of, or for damages resulting from the use of
: any information, apparatus, method, or process disclosed in this report, 7
f ‘ As vsed in the obove, "‘person acting on behalf of the Commission*’ includes any smployee or
’ conf_rucfor of the Commission, or employes of such contractor, to the extent that such employes
or contractor of the Commission, or employee of such contractor prepares, disseminates, or -
provides access to, any information pursuam to hls employment or contract mih the Commnsflon, i B
or his employment with such contractor, S _ ' : . W
.
b
.
i
!
f
n
i
:
~5
L

 

 

 
ORNL-TM-2256

Contract No. W-7405-eng-26
REACTOR CHEMISTRY DIVISION

CHEMICAL FEASIBILITY OF FUELING
MOLTEN SALT REACTORS WITH PuF,

RE. E. Thoma

LEG AL NOTICE

This repori was prepared 28 an account of Goverurient sponeored work, Neither the United
States, nor the {ommission, nor Any person acting ox behalf of the Commission:

A, Makes way waTranty or mpresentation,exmessed or implied, with respect to the accu-
racy, completeness, or usefainess of the information contained iz this report, or that the use
of any information, ppa ratus, method, or process digeiosed in this report may not infringe
privately owned rights; or

R, Ascumes any liebilities with respect to the use of, or for damagee resuliing from the
uge of any infcrmation. apparatus. method, or process disclosed in this report.

As used in the above, ‘‘person acting on behall of the Commigsion® jncludes any em-
ployee or contractor of the Commission, or employee of such contrector, to the extent that
such employee oT contractor nf the Commission, or employee of such comtractor prepares.
disseminates, or provides access ta, any information pursuant to his employment or coniract
with the Commission, or his amployment with suych vontractor,

OCTOBER 1968

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

5 UNLHATIER
zwmmww%%wwmfi@“wf‘
S

\‘,
 

 
1ii

CONTENTS
page

Abstract . . . . . . 4 0 e e e e e e e e e e e e e e w1
Incentives for Fueling Molten Salt Reactors

with Plutonium Fluoride . . . . . . . . . . . « . . . 2
Previous Evaluations of Plutonium Fluoride

Fueled Reactors . . . . .« « . ¢ « « « « « &« o « o« +« + 4
Chemical Properties of Plutonium Fluoride . . . . . . . 5

Solubility of PuF, in Fluoride Solvent Mixtures . . . 7

Segregation of PuF,; on Crystallization of Fuel Salts. 16

Chemical Compatibility with Fuel Circuit Materials . 16

Solubility of Pu,0; in Fluoride Mixtures . . . . . . 19
Estimation of Effects of Chemical Reprocessing . . . . . 20
Fission Products . . . « .+ « & « & « o + o & « « + & « . 25
Use of the MSRE to Demonstrate Feasibility of Operatlon

of MSR's with PuF, . ., . . . . . . . . . . -
Chemical Development Requirements . . . . . . . . . . . 29

Summary [] ° . ° s e € s & o * s ° o ® & 9 . ° . . ® e & 3 1
 

 
CHEMICAL FEASIBILITY OF FUELING
MOLTEN SALT REACTORS WITH PuF,

R. E. Thoma

ABSTRACT

The fegsibility of starting molten salt reactors with
plutonium trifluoride was evaluated with respect to chemical
compatibility within fuel systems and to removal of plutonium
from the fuel by chemical reprocessing after 239Pu burnout.
Compatibility within reactor containment systems is moderately
well-assured but reguires confirmation of PuF; solubility and
oxide tolerance before tests can be made using the MSRE. Al-
though separation of plutonium and protactinium in the chemi-
cal reprocessing plant, as would be desirable in a large
breedér reactor, has not yet been demonstrated, conceptual
designs of processes for effecting such separations are avail-

able for development.
INCENTIVES FOR FUELING MOLTEN SALT REACTORS WITH PLUTONIUM
FLUORIDE

In a2 recent report, P. R. Kasten described the economic
advantages of using plutonium as a startup fuel in molten salt
reactors,1 The following discussion summarizes his appraisal
of the incentives which are derived from the use of plutonium
in this manner. It is anticipated that large quantities of
plutonium will be produced during the following decades by
light water reactors fueled with slighly enriched uranium.
Sale of the plutonium produced from these reactors at $10/g
of fissile material is an important consideration in the power
cost of these systems. Recycle of plutonium in light water
reactors does not lead to a fuel value of $10/g for fissile
material over many recyclesaz Further, during the first few
vears when the fuel reprocessing industry associated with the
light water reactors is developing, the costs of fabricating
plutonium-fueled elements will be disproportionately high in
comparison with cost for uranium fueled elements, and this
will also tend to discourage recycle of plutonium. Thus, it
appears that within the next several decades the net value
of fissile Pu relative to its use in light water reactors wilil
be less than $10/g, probably about $6/g.

In molten salt reactors the penalty of preparing plutonium
fuels rather than uranium fuels does not appear to be economi-
cally significant. Also as shown1 the value of plutonium in
MSBR systems is about $12/g. Thus, there is a differential

of approximately $6/g between the value of plutonium recycled
N

in light water reactors versus its value in MSBR's. A 1000
MW(e) MSBR requires about 1000 kg fissile plutonium during the
startup period. At a differential of $6/g, this corresponds
to $6 million. Presumably this $6 million advantage for a
1000 MW(e) reactor would not be credited completely to MSBR's
but would be split with light water reactors by using an inter-
mediate Pu value.

One of the reasons for developing fast breeder reactors
is that they can advantageously utilize plutonium as a fuel.
If MSBR's are to serve as an alternative breeder system, it
is desirable that they also utilize plutonium advantageously
as a startup fuel. As indicated above, this appears to be
possible if the technology is favorable. Further, the low
specific inventory in MSBR's permits molten‘salt reactors to
be built in relatively large numbers using plutonium product
fuel from light water reactors. This feature permits MSBR's
to contribute to improved fuel utilization since their opera-
tion would not be limited by the availability of uraniferous
fuels.

The advantage of starting up on plutonium rather than
2357 arises from the fact that a lower concentration of Pu is
required for criticality in the fuel, and also because after
Pu burnout, the higher plutonium isotopes (neutron poisons)
presumably can be separated from the uranium. This operation
leads to slightly better nuclear performance over a 30-year

reactor life when plutonium is the startup fuel than when
4

2357 is the startup fuel and the higher isotopes cannot be
discarded (increase of about 0.01 in the breeding ratio).

The incentives described above form the basis for or
justify evaluations of the feasibility of incorporating
plutonium in molten salt reactors. An assessment was made of
current information on chemical properties of PuF; in order
to judge the feasibility of its incorporation in MSR fuel
salts, and to estimate the character and extent of informa-
tion which may be required to demonstrate chemical compati-
bility of PuF,; in the multicomponent environment of fuel-

fertile salt systems.

PREVIOUS EVALUATIONS OF PLUTONIUM-FLUORIDE FUELED REACTORS
During the early stages of the Molten Salt Reactor Program,
the fluorides of plutonium were consi dered for application in
advanced versions of molten salt reactors. The results of one
study3 showed that a PuF; fueled two-region homogeneous fluoride
salt reactor was operable, although its performance was poor.
Further development was not pursued for neither its chemical
feasibility nor methods for improving performance was obvious.
Although the thermochemical properties of the plutonium fluorides
were not well established at that time, it was clear that the
most soluble fluoride, PuF,, would be too strong an oxidant
for use with available structural alloys. The solubility of
PuF,, while sufficient for criticality even in the presence

of fission fragments and non-fissionable isotopes of Pu, was
estimated to limit the amount of ThF, which could be added to
the fuel salt.4 This limitation, coupled with the condition
that the continuous use of 229Pu as a fuel would result in
poor neutron economy in comparison with that of ¢33y~fueled
reactors vitiated further efforts to exploit the plutonium
fluorides for MSBR applications. Recent developments in fuel
reprocessing chemistry and in reactor design have established
the feasibility of a single-fluid MSBR. Consequently, it

now appears that it will be possible to operate a LiF-BeF;-
ThF,-PuF,; single-fluid MSR with lower concentrations of thorium
and plutonium than earlier considerations required, e.g., with
thorium fluoride concentrations of 8 to 12 mole % and with a
plutonium fluoride concentration of approximately 25% less
than required for 233U 1oading,5'i,e,,'% 0.2 mole %. Since
the incentive to use ?3%PuF; in molten salt reactors applies
exclusively to its temporary inclusion in the fuel stream,
prior limitations concerning saturation of the fuel with
respect to 24!PuF, and ?%2PuF,; do not seem to be relevant.

If the chemical properties of plutonium trifluoride prove

that its inclusion in molten salt reactor fuels is economically
and technically feasible, its exploitation in this connection
should be regarded as of significant advantage to the develop-

ment of the United States AEC breeder reactor program.

CHEMICAL PROPERTIES OF PLUTONIUM FLUORIDE

g One characteristic of the actinide elements is that
increasing instability of the higher oxidation states is ob-
served with increasing atomic number. This property is evi-
dent among the compounds of plutonium, particulary the halides.
Three stable fluorides of plutonium are known, whereas among
the other halides, only the trivalent oxidation state is
commonly exhibited. Since PuF, is a gas, only PuF, and PuF;
can be considered for use in molten salt fuel mixtures.
Plutonium tetrafluoride would exhibit higher solubility than
PuF, in fluoride solvents, but would probably prove to be too
strongly oxidizing to be compatible with Hastelloy-N. The
free energy for the following corrosion reaction strongly
favors oxidation of chromium containing alloys:

Cr®(s) + 2PuF,(s) — CrF,(s) + 2PuF,;(s)

 

AFIOOOOK: - 688 kcal = {148 kca% ;miii;i_iii§
-~ 688 keal ~-733.8 kecal

= ~85.8 kcal
The above reaction also shows that it would not be possible
to increase the concentration of plutonium in a fuel salt
which was already saturated with respect to PuF; by addition
of PuF,, since the corrosion reaction would proceed steadily
and produce additional amounts of PuF,;. Plutonium trifluoride
is, therefore, regarded as the only suitable fluoride of Pu
for application as a molten salt reactor fuel constituent..
Current values of the thermochemical properties of PuF; and
PuF, are compared with their uranium analogs and with thorium

tetrafluoride and cerium trifluoride in Table 1. The values -
7
listed here show that PuF, is more stable than UF,;, and suggest
as well that the solubilities of PuF,, UF,, and CeF, in fluoride

solvents might be similar.

A. Solubility of PuF; in Fluoride Solvent Mixtures

1. LiF-BeF,: The solubility of PuF; in LiF-Bel, solvents
was measured by Bart0n6 for compositions ranging in BeF, from
28.7 to 48.3 mole % and from 450 to 650°C. Solubilities of
PuF; in LiF-BeF, solvents are compared with those for CeF; in
the same composition range in Figure 1. These results imply
that the solubility of PuF; in LiF-BeF, solvents is markedly
temperature- and composition-dependent. Extrapolation of these
data to temperatures which are reasonable for the peritectic
invariant point involving LiF, Li,BeF,, and PuF; (Figure 2)
indicates that the composition of the mixtures at this invariant
point is LiF-BeF,-PuF, (63-37-0.008 mole %), T = 455°C, and
that the Li,BeF,-BeF;-~-PuF,; eutectic occurs af the composition
LiF-BeF, -PuF; (48-52~0.01 mole %), T = 358°C. The composition
dependence of solubility appears to be related to the acid-base
balance of the solvent, as is evident when the data are expressed
as a function of the estimated fraction of "free'" fluorides as
contrasted to "bridging" fluorides. While PuF,; solubility seems
to be minimal in the "neutral" melt, LiF-BeF, (66.7-33.3 mole
% the minimum in the CeF; solubility curves seems to occur in
mixtures which are slightly richer in BeF, (see Figure 3).

Barton investigated the effect of additional solutes on
Free energy of
formation at
1000%K

(kcal/F atom)

m.p. (°C)

Crystal
Structure

Density
(g/cm?)

ay.. Brewer, '""The Chemistry and Metallurgy of Miscellaneous Materials:
Thermodynamics,'" L. L. Quill, ed., McGraw-Hill,

b

Table 1.

ThF, (s)

Comparison of the Properties of PuF; with
ThF¥,, UF,, UF,;, and CekF,.

UF, (s)

 

PuF, (s) UF; (s) PuF; (s) Ce¥F,; (s)
~1012 ~95.3% ~86.0° ~99.9P ~104.3° ~1182
1111 1035 1037 1495 1425 1437

Md Md Md e He He
5.71 6.72 7.0 8.97 9.32 6.16

C. F. Baes, Jr., "Thermodynamics," Vol. I,

and G. Long,

January 31,

1965.

“F. L. Oetting, Chem. Rev., 67, 61 (1967).

dMonoclinic, space group C2/c.

eHexagonal, space group P6/mcm.

New York,

IAEA, Vienna,

1950,

1966, p. 409;

76-192.
o ORNL-DWG 68-5397

 

 

6 |

.4

 

 

o™
N

 

CeF; ~ 650°

o C |

PuF,~ 650°
/

\

\
° ° /
l 3 // Cefz—600° 7

PuF3- 600°

W
w a\\/ ®”  CeF;-550°

PuFB— 550°
- _

 

o
Lo

 

 

0
o\c<

 

CeF; OR PuFy IN FILTRATES (mole %)

 

 

 

 

 

 

 

 

 

 

C4 9//
. L/
\- ® .--/
0.2
O
10 20 30 40 50 e0

Bef, IN SOLVENT (mole %)

Fig. 1. Comparison of CeF3 and PuF3 Solubility im LiF-BeF, Solvent.
10

ORNL-DWG 68-7225

 

750 I k

29.5 49 37 BeFp (mole %)

700 \

245 (g
650 —
\

 

 

 

 

600

 

550

TEMPERATURE (°C)

 

500

 

450

 

 

400
£=358°

 

 

 

 

 

 

350
2.0 £.5 1.0 0.5 o

CeF3 OR PuFy {mole %)

Fig. 2. Solubility of CeF3 and PuF3 in LiF-BeF; Solvents Extrapolated
to LiF-BeF,-MF; Invariant Equilibrium Points. 455° = the peritectic,
LiF-Li,BeF,-MF , 358° = the eutectic, Li;BeF,-BeF;-MFj.
11

f@” the solubility of PuF; in LiF-BekF, mixtures;6 using low (£ 1
mole %) concentrations of ThF,, BaF,, and CeF,, and high con-
centrations (20 mole %) of UF,. His results showed that at 1
mole %, ThF, had very little effect on the'solubility of PuF,
in this solvent. The same amount of BaF, diminished the solu-
bility of PuF, in a manner not clearly understood. Barton
speculated that the saturating phase in these experiments was
quite possibly not pure PuF;, but rather was a solid solution
of BaF, and PuF,. As the molar ratios of BaF, and PuF,; were
varied in these experiments the optical properties of the pre-
cipitating phase also varied, such as to indicate that the
solid phase in equilibrium with liquid was a BaF,-PuF; solid
solution. The magnitude of the effect indicated that the con-
centration of divalent fission products anticipated in reactor
operation would probably not significantly affect the solubility
of PuF,.

Data obtained with CeF;-PuF; solute mixtures in the sol-
vent LiF-BeF, (63-37 mole %) are shown in Figure 4. The theo-
retical curves for CeF;-PuF, mixtures shown in Figure 4 were
calculated from the equation NPuF3(d) = S%uF3NPuF3(SS)’ where

the

N d), is the mole fraction of PuF; in solution S%

PuF3( uF;’

mole fraction (solubility) of PuF; in the solvent at a specified
temperature (shown labeled "PuF,; only'" in Figure 4) while
NouF (ss) is the mole fraction of PuF, in solid solution.

3

Agreement between experimental and calculated solubility values

indicates that PuF,; and CeF; form solid solutions.
12

ORNL-OWG 68-7228

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.8
A
/
/
)
16 cd e E 4
4
4
/
//
1.4 4
/
/
/s
4
12 ’4' -
Bt ,
g 1O . o z 4 “,
et =As<‘;52f650[ 7 //
) -~ { L/
L'j -...\ L#a \ A J'/ /
a Pu 650° "~«a ¢ ’
S 08 T~ T~ 7 A 7//
- AN ;:
Lf.;n - \\\ *-.__v/ // A
© T~ Le 600° . o,
06 ) \ l/ : //
: # / 7
o~ Py 600° " ~~< N2 / g
‘h."'s / /’4'
\A"‘s.‘ ,/ //
oa L TTpenCe s50° T~ 7~ A~
. ¢ PU 5500 ,,_.“-... \f A,/'
- -"--.-""-‘ ‘,,’ @
@ N‘/
C.2 - 49 46.5 44 40.5 37* 345 | 32 29 24.5 7
36.75
BeF, (mole%)
|
0 - | :
-50 -40 -30 -20 ~40 0 10 20 30
FREE FLUORIDE {ON BALANCE '
Fig. 3. Effect of LiF-BeF, Sclvent Composition on the Solubility of

CeF3 and PuFj.
13

ORNL-LR-DWG 32944A

TEMPERATURE (°C)
650 600 550

 

500

 

 

|

|

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.05

 

 

 

 

 

PuFz ONLY IN SALT g

 

PuFs + BaF, IN SALT

 

SOLUBILITY OF PuF3 IN LiF -BeF, (63-37 mole o)

1:1 MOLAR RATIO OF CeFz TO
PuF3 IN SALT

o
® PuFz + ThF, IN SALT N
a
A

 

0.02 g 5:1 MOLAR RATIO OF CeF5 TO
PuF3 IN SALT

 

 

 

 

 

 

0.0 L— I
100 105 1.0  #4.5 420

104/7 (°K)

Fig. 4.

12.5

13.0

{3.5
14
The solutes were found combined in single-phase materials with it

optical properties intermediate between those of CeF; and PuFj;.

2. LiF-BeF,;-UF,: Barton6 measured the solubility of PuF;
in a LiF-BeF,-UF, melt of the compesition 70-20-1C mole %. The
results which he obtained comprise the only available informa-
tion on the solubility of PuF,; in melts which contain more than
1 mole % of metal tetrafluorides. The values for the solubil-
ity of PuF; in the LiF-BeF,-UF, solvent fall on a straight line
when plotted as logarithm of concentration vs. reciprocal tem-
perature. Considered in terms of '"free fluoride'" ions avail-
ablve, the ion balance in the solvent may vary from -10 to
-30 depending on whether one assumes the predominant anionic
association of uranium ions to be UFy; or UF;3  in the melt.
Tetravalent uranium does not form stable phases of the stoichio-
metries Li,UF,, Li,UF,;, or Li,UF,. Of these, only Li,UF,
exists as an equilibrium crystalline phase, and its tempera-
ture range of stability extends only over 30°C. It seems most
probable that the uranium ions in the solvent exist princi-
pally as UF, . If so, the solubility data from Table 2 fit
closely with those shown in Figure 3. Since "LiF-BeF,-ThF,-
UF, single fluid fuels are likely to be more neutral on the
negative side, we must presume that the solubility will be
near the lowest values. The results of all the measurements
which have been made suggest however that the solubility of

PuF, in MSR sclvent systems will not be lower than 0.25 mole %

at temperatures of 550°C or higher.
15

Table 2. Solubility of PuF, in LiF-BeF,-UF,

(70-10-20 mole %)

 

Filtration Concentration of Pu
Temperature in Filtrate
°C) (wt. %) (mole %)
558 3.43 1.27
600 4.57 1.70

658 | 6.50 2.48
16

The data in Figure 3 indicate that if the '"free fluoride" £
jon balance is negative, the differences in solubilities of ;
CeF; and PuF; are essentially constant. Therefore, the solu-
bility of PuF,; in solvents similar in composition to the MSBR
carrier and MSRE fertile carrier salt mixtures can be deduced
from the results of CeF; solubility measurements, which in
respect to those for PuF,;, can be accomplished with compara-

tive ease.

B. Segregation of PuF; on Crystallization of Fuel Salts

The principal components of MSR fuel mixtures do not form
intermediate compounds with PuF;. From the solubility data
cited above, it can be inferred that if it is employed in fuel
mixtures at concentrations of a few tenths mole percent, PuF;
will tend to crystallize from such mixtures as the primary
phase and in solid solution with UF; and/or the rare earth
trifluorides. The ?LiF/BeF, ratio in ?LiF-BeF,-ThF -PuF;
fuel mixtures could be adjusted to insure that at saturation
other fluorides, such as "Li; (Th,U)F, would coprecipitate with
PuF; at the liquidus. It is anticipated therefore that in
the concentrations at which PuF, would probably be employed,
it would not be deposited preferentially from the bulk salt
during the inadvertent freezing, nor at locations such as in
freeze valves where repeated thawing and freezing would take

place.

C. Chenical Compatibility with Fuel Circuit Materials

S

A considerable amount of theoretical and experimental
17

evidence exists which indicates that as a component of fluoride
fuel mixtures PuF; will be chemically compatible with container
alloys and graphite. Of the actinide fluorides which may be
used to constitute molten salt reactor fuel mixtures, pluton-
ium trifluoride is the most chemically stable. Unlike UF,;,

it shows no tendency to disproportionate to the tetrafluoride
and metal.

Fluoride melts containing PuF,; were contained in nickel
vessels in many of the experiments conducted by C. J. Barton
and co-workers. Nickel proved to be an entirely satisfactory
container material for this use. In the nickel based alloy,
Hastelloy-N, the corrosion reaction which is intrinsic to
uraniferous fluoride salt systems is Cr® + 2UF;, == 2UF; + CrF,,
a reaction which has no analog in PuF; fuel systems. The role
of PuF; in corrosion of Hastelloy-N container vessels may
therefore be nil. The possibility that some unidentified
reaction might cause mass transfer in a temperature gradient
cannot be ruled out. Since such corrosion is limited by the
diffusion of chromium in Hasfelloy—N to liquid-solid bound+-
aries,7 the rate of mass transfer could only be extremely low.

The compatibility of PuF; with MSR fuel circuit environ-
ment has, to an extent, already been demonstrated in the MSRE,
where some 100 ppm of plutonium was generated and remained
entirely in the fuel salt. Its stability there was estab-
lished by the results of routine chemical analysis which
were in good agreement with the anticipated values during

2357 operations.
18

It appears highly unlikely that the carbides of plutonium
can form in molten salt reactors which employ PuF; in the fuel
stream. The free energy of formation of the plutonium carbides
is quite low, ~20 kcal/mole at 1000°K.® While the uranium car-
bides have comparably low free energies of formation, the
possibility of carbide formation with moderator graphite exists
only if the activity of U%, formed in disproportionation, is
permitted to rise 2 5x10 ¢, Since disproportionation of PuF,
does not occur, the driving force for the formation of pluton-
ium carbides is entirely absent.

Thermodynamic data suggest that if graphite were to react
with MSR fuel mixtures containing UF,;, the most likely reaction
would be 4UF, + C == CF, + 4UF;, which should come to equili-
brium at CF, pressures of or below 10 & atm. It has been shown
by mass spectrometric analysis9 that the concentrations of CFy
over graphite systems which were maintained for long periods at
elevated temperatures did not exceed the lower detection limits
(<1 ppm) for this compound. Reduction of PuF; by a similar
reaction appears to be very improbable.

From consideration of the thermochemical properties of
PuF,; and from its chemical behavior in the MSRE as described
above, we can anticipate that the compatibility of PuF; with

MSR graphite moderator and containment alloys will be excellent.
19

D. Solubility of Pu,0,; in Fluoride Mixtures

 

Initial demonstration of the application of PuF,; in molten
salt reactors would come appropriately from its inclusion in
MSRE fuel salt. Before embarking on such a demonstration, it
would be necessary to have accurate information about the soclu-
bility of Pu,0; in the MSRE fuel and flush salts. C. F. Baes
appraised the thermochemical data for PuF; and Pu,0; recentlylo
and concluded that there is a distinct possibility of precipi-
tating Pu,0; if PuF, is introduced into the MSRE fuel salt at
a concentration of as high as 0.2 mocle % and if the oxide
level should approach the value for ZrQ, saturation (~500 ppm).
In our previous experience with the MSRE, the total concentra-
tion of oxide in the fuel salt has remained less than 100 ppm.
Although it seems improbable that saturation of the MSRE fuel
salt with Pu;0; could occur at such low oxide concentrations,
the oxide tolerance of such mixtures is currently inestimable
because of the uncertainties which may be present in the thermo-
chemical data. Laboratory experience with PuF; melts has not
suggested that Pu,0; exhibits unusually low solubility in
fluoride mixtures, i.e., that its solubility is lower than
ZrO, or UO,. However, since the possibility exists that Pu,0;
precipitati on might occur, the oxide chemistry of Pu? " in
molten fluorides should be determined experimentally if the

MSRE were to be used to demonstrate the potential application

of PuF,-based fuels.
20

ESTIMATION OF EFFECTS OF CHEMICAL REPROCESSING

One of the anticipated advantage of starting melten salt
reactors on plutonium rather than on ?35U is that slightly
improved nuclear performance (increase of about 0.01 in the
breeding ratio) over a 30-vear reactor life would result from
its temporary presence in the reactor. The maximum economic
advantage would result from removal of the higher isotopes of
plutonium (neutron poisons) after plutonium burnout. The in-
centives for using PuF,; to start up molten salt reactors are
to some extent enhanced or diminshed relative to the simplicity
(economy) of the fuel reprocessing methods which are employed
in conjunction with its use. For the economic advantage of
employing plutonium in the fuel-fertile salt to be very signifi-
cant, the reprocessing costs associated with removal of ¢%!Pu
and *%2Pu should not add appreciably to the fuel cycle costs.
In order to achieve such economy, it will probably be necessary
to remove plutonium via the same chemical processes which are
to be employed for the 2?3U-Th fuel-fertile stream. At the
current stage of MSBR fuel reprocessing development, it is
anticipated that reductive extraction methods would be em-
ployed. As is shown below, the available electrochemical data
for plutonium and protactinium compounds do not permit us to
deduce whether protactinium is separable from plutonium on a
short cyéle, ~3 days, when plutonium is the fuel. Further,
removal of “%'Pu and #%2Pu after ?3%Pu burnout involves separa-

tion of plutonium from thorium. This separation appears to
21
s be more tractable than the former, but also cannot be assured
at this point. If separation coefficients for plutonium are
found to lie between those for protactinium and the rare
earths, little or no plutonium would be removed concurrent
with protactinium.

Unlike the lanthanides, the actinides exhihit trends
in chemical properties which reach minimum or maximum values
among the first members of the series. Such a trend is shown
in Cunningham and Wallman's values11 of the formal potentials
for the reaction M(s) - M3+ in aqueous solution. (See
Figure 5)

A similar trend is suggested in the reduction potentials
for Th¢™, U3+, and Pu?”’ in the fluoride solvent, LiF-BeF,
(66-34 mole % (Figure 5). These trends might imply that the
€o' for the reduction of plutonium into a bismuth alloy will
be nearly identical to that for prétactinium. We have no
means available at present for estimating €,' for plutonium
reduction with the accuracy required to indicate its posi-
tion in the reduction sequence of the actinides Th to Pu.

Moulton12 has recently evaluated the possibilities of
removing Pu from molten salt reactor fuels by reductive
extraction into bismuth. His conclusions are summarized
as follows. The stability of Pu-Bi intermetallic phases
is not predictable quantitatively. The similarities of the
Th-Bi, U-Bi, and Pu-Bi phase diagrams indicate that the

activity of plutonium will be substantially lower in bismuth.
22

The activity of a metal in bimsuth can be referred to
the pure metal by the use of an activity ccefficient Y
which is <<1 and more or less constant at this value from
infinite dilution up to saturation where the saturating
phase is the solid intermetallic. Then € = €4' - %% 1n WM/VS'
(ys, the ion activity coefficient, goes to 1 at infinite

dilution and can be considered as 1 to a first approximation.)

Literature values of vg, and vy are 5.7x10° % and 1°3X10“4(13’14)
which would give eo%h = -1.58 and €0%'= -1.27 with respect
to the H,-HF electrode at ¢, = 0. The results of experiments

conducted by Moulton and Shaffer show that €, for Pu is
about 0.05V more negative than €,' for U. II Ypy = YTh’

then €,' = -1.20, while if Ypu = -1.40. One can

Yy’ €°éu -
be reasonably sure that Gofiu will fall somewhere within
these limits. Since e¢,' for the rare earths lie about -1.50,
it is likely that Pu can be separated from them and from
thorium. Its position relative to U and Pa is not so clear.
An argument can be made that Ypu will be nearly the
same as vy for either U or th. The solubility of PuBi, in
Bi is greater at any temperature than that of UBi, or ThBi,,
and its congruent melting point is lower (830 vs., 1010 and
1230°C) which suggests that Yp, 1S not very small. On the
other hand, the metal itself melts lower (640 vs. 1132 and
1750°C) so that of the three systems only PuBi, melts higher

than the metal. There is some correlation between electro-

negativity and stability of actinide intermetallic compounds s
23

R with bismuth. Plutonium and thorium both have a value of 1.3
(Pauling scale) while uranium is 1.7. The Pu intermetallic
will therefore probably exhibit comparable stability.

If plutonium comes out before or with uranium in the
reduction extraction process, it can be concluded that the
utilization of PuF; in molten salt reactors would have little
or no €ffect on fuel reprocessing costs. It would be necessary
to separate the uranium and plutonium, probably by fluorina-
tion, but this step should not increase overall fuel process-
ing costs appreciably.

If it is found that the separation coefficients for
plutonium in reductive extraction processes are unfavorable,
alternative methods for its removal could be devised. One
possible method would involve fluorination of the fuel first
at 550°C to remove uranium as UF,, then at higher temperatures,
2700°C to remove plutonium as PuF,, leaving any undecayed
protactinium with the carrier salt. This procedure would
utilize the increase in stability of PuF, with increasing
'temperaturel5 and the fact that protactinium does not form
volatile fluorides. Such a method would, of course, not be
applicable during operation of a reactor with PuF; fuel, but
rather is a possible means of final removal of plutonium.

The chemical feasibility of incorporating PuF, in molten
salt reactor fuels, as demonstrated by operating the MSRE
with a PuF; fuel, would not be impaired by the incomplete

development of a chemical process for its separation from
24

ORNL—DWG 68-7227 =

 

 

 

2.6
O £, EXPERIMENTAL
2.4 ® £, CALCULATED BY MOULTON
)
£, RQUEOUS, FROM REF. 9
2.2 O —

 

 

o\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

@«
e \-!—‘
Ly
Ec'; IN LiF-BeF, (66—34 mole %) /

.6 "””/////;

{.4 ,d

*,2 .--__-‘

.0

Ac Th Pa U Np Py Am

Fig. 5. Standard Reduction Potentials for Some Actinides in Aqueous,
Molten Salt, and Metallic Solvents.
25
protactinium. It should be inferred therefore that while a
flowsheet for its: separation cannot now be devised, .the
research and development efforts are readily identifiable

and are experimentally tractable.

FISSION PRODUCTS

In one of the important continuing investigations with-
in the MSRP, we are attempting to establish experimentally
the chemical identities and modes by which a number of the
fission products, notably those of the near noble metals, are
distributed, partly as a means for predicting the behavior
of spent fuel in the chemical reprocessing plant and partly
to establish its corrosion potential accurately with increas-
ing burnup. No significant differences are believed to exist
in the yields or chemistry of the principal species of fission
products which would result from incorporation of PuF; in
MSR fuels. The feasibility of using PuF; in startup opera-
tions of an MSR does not therefore appear to require a
separate research program relative to fission products from
plutonium.

With 235UF, fuel, the fission reaction is mildly oxi-
dizing, resulting in the oxidation of ~0.8 equivalent4 of
UF, per gram atom of fissioned uranium. The oxidation
potential results from the anion-cation imbalance which
develops as the fission products reach thermodynamic equili-

brium. With UF,; as the fissile solute, a slight excess of
26

fluoride ions develops. Use of a trifluoride solute, how- i
ever, should result in a cation excess, and should cause the n
fuel solution to generate a mild reducing potential.
The yield of !35Xe from plutonium fission is somewhat
creater than from 233U, and in turn, is less than from %33U.
The relative poison fraction of 135Xe in the fuel would be
at a minimum at initiation of power operations with PuF,
fuel and would increase as ?3%U was generated within the

system.

USE OF THE MSRE TO DEMONSTRATE FEASIBILITY OF OPERATION OF
MSR'S WITH PuF,

Consideration of several developments in molten salt
reactor chemistry within recent years suggests that the most
appropriate and earliest demonstration of the applicability
of PuF; in molten salt reactors would come from its use in
the MSRE. Sufficient basic chemical information exists to
conclude that it is neither necessary nor important to
demonstrate chemical compatibility with metal alloys in
engineering laboratory scale tests. Laboratory scale tests
with plutonium should be restricted to the minimum number
necessary to establish stabilities because of the inhala-
tion hazard of plutonium-239. Plutonium-239 is, in fact,
regarded as one of the most toxic substances known to the
experimentalist.

The fact that plutonium-beryllium mixtures are neutron
27

sources also complicates laboratory and engineering scale
experiments in which "LiF-BeF,-PuF,; mixtures are used. Some
typical valuesl()_l8 of the neutron energies produced from
actinide-beryllium o¢,n reactions are listed in Table 3.

Tate and Coffinberry19 have computed theoretical neutron
yields of pilutonium-beryllium alloys employing calculations
which include a term from the exXperimental stopping power
of the matrix elements for alpha particles. The available
data suggest that in a dilute Be? "t solution, such as in a
MSRE fuel mixture, e.g., "LiF-BeF,-ZrF,-ThF,-PuF; (64-28-5-
3-0.2 mole %) the neutron yield would not be so great as to
require special shielding of salt lines, drain tanks, or
fuel sample transport containers.

The possible criticality problems associated with
storage of PuF,-bearing fuel salt have been considered
qualitatively and do not seem to be ominous. Whereas fission
multiplication factors hold for %33V into the epithermal
neutron range, they do not do so in the case of plutonium.
Further, more energetic @,n reactions will accompany #%3U
operation of the MSRE20 than are likely with PuF;, primarily
because of the presence of 232U in the charge which is to
be used. Accordingly, the potential radiation problems
associated with a,n reactions in fuel salt will have been

faced before PuF; is used in the MSRE. Although radiation

from fuel-fertile salt in storage tanks does not seem to

be serious, detailed scrutiny of the possible problems which
T

Table 3.

Typical Values of Neutron Energies Produced

From Actinide-Beryllium ¢,n Reactions

 

@ Source t3 Q(Mev) gig:;;n Nfig;ggzeéifii?éigé
(Mev) -

21 0py 138.4d 5.3 $11, av. 4 80

2z22Rp 3.83d 5.48 $11 460

226Ry 1.62x103y 5.65 $13 460

2337; 1.63x10°y 4.82

235y 7.07x108%y 4.80

239py 2.43x10* 5.15 210, av. 4

8¢
29

= might arise from a¢,n reactions would be necessary before the
PuF; were to be used in the MSRE.

Except for a few data on solubility of the fluorides
and oxides, which are obtainable in laboratory measurements,
chemical compatibility in reactor containment systems is
reasonably assured. If the solubilities of Pu,0,; and PuF,
in LiF-BeF,-ZrF, melts are found to be in excess of 300-400
ppm and 0.2 mole %, respectively, a test in the MSRE would
be virtually assured of success with respect to the chemical

behavior of the plutpnium-bearing salt.

CHEMICAL DEVELOPMENT REQUIREMENTS

It appears that the chemical feasibility of employing
PuF, in molten salt reactors will be assured if two general
properties, solubility of the oxides and fluorides in LiF-
BeF,; -Zr¥,~-ThF, solvents are suitably high, and the extract-
ability of Pu metal from fluoride melts into bismuth amal-
gams 1is sufficiently discrete to be economic. As noted
above, only the absence of solubility data obviates the
conclusion that PuF; could be incorporated in the MSRE fuel
salt at our earliest convenience.

In order to establish that it is chemically feasible
to fuel molten salt reactors with PuF,;, a program of chemi-

cal development should include the following items:
30

a. Determination of the solubility of PuF; in LiF-BeF,-ZrF, s
and LiF-BeF, -ZrF,-ThF, solvents. It should be adeguate &
to carry out most of the necessary measurements with
CeF; as a proxy for PuF,. Thereafter, only a few experi-
ments with PuF; would be required to confirm the conclu-
sions based on CeF; solubilities.

b. Determination of the solubility of Pu,;0; in LiF-BeF,-ZrF,-
ThF,-UF, solvents. The lanthanide oxides do not serve
adequately as proxies for estimation of actinide oxide
solubilities. It will be necessary therefore to determine
the oxide tolerance of fuel salts directly with plutonium
oxide in alpha-laboratory facilities.

¢c. Establishment of the standard reduction potentials and
separation coefficients for plutonium in Bi-Th alloys.

d. Solubility of Pu in Bi-Th alloys.

Items c¢. and d. should become a part of the existing
programs in chemical and chemical engineering development.

It may be unnecessary to initiate experimental work in this

part of the program until it is first demonstrated that PuF,;

fuels perform satisfactorily in a molten salt reactor.

It is likely that some 15 to 20 years will pass before
plutonium trifluoride is incorporated in a full scale power
reactor. If a demonstration that molten salt reactors are
operable with PuF, fuels is regarded as desirable, it can
probably be realized with the MSRE. Since molten salt fuel

processing technology will require a period of years to evolve,
31

f”"- the ambiguous fate of plutonium in fuel reprocessing should
not at this point be considered a deterrent to a continuing
evaluation of the chemical feasibility of employing PuF; as

an MSR fuel.

SUMMARY

A definite economic advantage is associated with startup
of molten salt breeder reactors with PuF;-based fuel. If the
solubilities of plutonium oxide and plutonium trifluoride are
confirmed as exceeding a few hundred ppm and ~0.2 mocl %,
respectively, the chemical feasibility of fueling molten salt
reactors with PuF; will be eésentially assured. Separation
of protactinium and plutonium during operation of a PuF,-
fueled reactor, and removal of ?%!Pu and %4%Pu after two
yvears of operation; as would be desirable in a large breeder
reactor from an economic standpoint, has not yet been demon-
strated, although conceptual designs of processes for effecting

such separations are available for development.
10.

32

REFERENCES
P. R. Kasten, Reactor Division - ORNL, personal communi-
cation, 1968.
P. R. Kasten, J. A. Lane, and L, L, Bennett, "Fuel Value
Studies of Plutonium and U-233," unpublished work, 1962,
D. B. Grimes, Molten Salt Reactor Program Quarterly
Progress Report for Period Ending June 30, 1958, ORNL-
2551, p. 13,

J. A. Lane, H. G. MacPherson, and F. Maslan, Fluid Fuel

Reactors, (Addison-Wesley Publishing Co., Inc., Reading,

Mass.), p. 656.

0. L. Smith, Reactor Division - ORNL, personal communi-
cation.

C. J. Barton, J. Phys. Chem., 64, 306 (1960).

J. H. DeVan and R. B. Evans, III, '"Corrosion Behavior of
Reactor Materials in Fluoride Salt Mixtures,'" ORNL-TM-328,
September 19, 1962.

W. M. Olson afid N. R. Mulford, Proceedings of Symposium on
Thermodynamics of Nuclear Materials, IAEA, September 1967,
Paper No. SM-98-40.

W. R. Grimes, "Radiation Chemistry of the MSRE System,"
ORNL- TM-500, March 31, 1963,

C. F. Baes, Jr., Reactor Chemistry Division - ORNL,

personal communication, 1968.
11,

12.

13.

14.

15.

16.

17.

18.

19.

20.

33

B. B. Cunningham and J. C. Wallman, in The Chemistry of

 

the Actinide Elements, J. J. Katz and G. T. Seaborg, ed.,

 

(John Wiley, New York, 1957), p. 412.

D. M. Moulton, Reactor Chemistry Division -~ ORNL, personal
communication, 1968.

R. H. Wiswall and J. J. Egan, Proceedings of Symposium on
Thermodynamics of Nuclear Materials, IAEA, May 1962,
Vienna, p. 345.

L. C. Tien et‘al., "Thermodynamics,” Vol. 1, Internaticnal
Atomic Energy Agency, Vienna, 1966, pp. 501-514.

B. Weinstock, Rec. Chem. Prog., 23, 23 (1962).

National Bureau of Standards Handbook 72, "Measurement

of Neutron Flux and Spectra for Physical and Biological
Applications," 1960.

G. E. Darwin and J. H. Buddery, Beryllium, (Butterworths,
London, 1960), p. 362.

G. Friedlander and J. W. Kennedy, Nuclear and Radiochemi-

 

stry, (John Wiley, New York, 1955),

R. E. Tate and A. S. Coffinberry, Proceedsings of Second
United Nations Conference on the Peaceful Uses of Atomic
Energy, Geneva, 1958, Paper 700.

R. C. Steffy, "Inherent Neutron Source in MSRE with Clean
2337 Fuel Salt Mixture," ORNL-Technical Memorandum (In

preparation).
 

 

 
35

ORNL~-TM~ 2256

INTERNAL DISTRIBUTION

1. R. XK. Adams 47. D. R. Cuneo

2. G. M. Adamson 48. J. M. Dale

3. R. G. Affel 49, D. G. Davis

4., L. G. Alexander 50. R. J. DeBakker

5. J. L. Anderson 51. J. H. DeVan

6. R. F. Apple 52. 8. J. Ditto

7. C. F. Baes 53. A. S. Dworkin

8§. J. M. Baker 54. I. T. Dudley

9. S. J. Ball 55. D. A. Dyslin
10. C. E. Bamberger 56. W. P. Eatherly
11, C. J. Barton 57. J. R. Engel

12 H. F. Bauman 58. E. P. Epler

13. S. E. Beall 56. D. E. Ferguson
14, R. L. Beatty 60. L. M. Ferris
15, M. J. Bell 61. A. P. Fraas

16. M. Bender 62. H. A. Friedman
17. C. E. Bettis 63. J. H. Frye, Jr.
18. E. S. Bettis 64. C. H. Gabbard
19. D. 8. Billihgton 65. R. B. Gallaher
20. R. E. Blanco 66. R. E. Gelbach
21 F. F. Blankenship 67. J. H. Gibbons
22 J. O. Blomeke 68, L. O. Gilpatrick
23. R. Blumberg 69. H. E. Goeller
24. E. G. Bohlmann 70. W. R. Grimes
25. C. J. Borkowski 71. A. G. Grindell
26. G. E. Boyd 72. R. W. Gunkel
27. C. A. Brandon 73. E. D. Gupton
28. M. A. Bredig 74. R. H. Guymon
29. R. B. Briggs 75. 4. P. Hammond
30. H. R. Bronstein 76. B. A. Hannaford
31. G. D. Brunton 77. P. H. Harley
32. D. A, Canonico 78. D. G. Harman
33. 8. Cantor 79. W. O. Harms
34, W. L. Carter 80. C. S. Harrill
35. G. I. Cathers 8l. P. N. Haubenreich
36. 0. B. Cavin 82. R. E. Helms
37. A. Cepolino 83. P. G. Herndon
38. W. R. Cobb 84. D. N. Hess
39. C. W. Collins 85. J. R. Hightower
40. E. L. Compere 86. M. R. Hill
41. K. V. Cook 87. H. W. Hoffman
42. W. H. Cook 88. D. K. Holmes
43. L. T. Corbin 89. P, P. Holz
44. B. Cox 90. R. W. Horton
45. J. L. Crowley 91. T. L. Hudson

E 46. F. L. Culler 9Z2. H. Inouye
93.
94.
95.
96.
97.
G8.
99.
100.
101.
102.
103.
104.
105,
106.
107.
108.
109.
110,
111.

112

113.
114,
115,
116.
117.
118,
119,
120.
121.
122,
123.
124.
125.
126,
127.
128.
129.
130.
131.
132.
133.
134.
135,
136.
137.
138.
139.
140.
141.
142.
143,

ORG-S edG OO NSO GO EEQrPAIEGGOGHEHPRPGRIEOEREE0OAITE

36

H. Jordan 144.
R. Kasten 145,
J. Kedl 146,
L Keller 147.
T. Kelley 148,
J. Kelly 149,
R. Kennedy 15¢C.
W. Kerlin 151.
T. Kerr 152.
S. Kirslis 153.
W. Koger 154,
I. Krakoviak 155,
S. Kress 156.
W. Krewson 157.
E. Lamb 158.
A. Lane 159,
J. Lawrence 160.
S. Lin 161.
B. Lindauer 162.
P. Litman 163.
H. Llewellyn 164,
L. Long 165.
I. Lundin 166.
N. Lyon 167.
L. Macklin 168.
G. MacPherson 169.
E. MacPerson 170,
C. Mailen 171.
L. Manning 172,
D. Martin 173,
H. Mauney 174.
McClain 175.
¥W. McClung 176.
E. McCoy 177,
F. McDuffie 178.
K. McGlothlan 179.
J. McHargue 180.
E. McNeese 181-190.
R. McWherter 191,
J. Metz 192.
S. Meyer 193.
L. Moore 194,
M. Moulton 195,
R. Mueller 196.
A, Nelms 197.
P. Nichols 198.
L. Nicholson 199.
D. Nogueira 200,
C. Oakes 201.
Patriarca 202.
M. Perry 203,

cqgméc«sbwmmhmemsmumméw»fluomm:r-masmommuosmmgmmsmmggmwme‘

om&uwzmwr‘mwaujtzflv-am:dmmcommwfiazzummmmrwmmomfia»oootj

Pickel =
Piper -
Prince
Ragan
Redford
1chardson
Robbins
Robertson
Robinson
Roller
Romberger
Rosenthal
Ross
Savage
Schaffer
Shilling
p Scott
Scott
Seagren
Sessions
Shaffer
Sides
Skinner
Slaughter
Smith
Smith
Smith
Smith
Smith
plewak
Steffy
Stoddard
Stone
Strehlow
Tallackson
Taylor
erry
Thoma
Thomason
Toth
Trauger
Watson
Watts
Weaver
Webster
Weinberg
Weir
Werner
West
Whatley
White

SUI:""‘L'"L'IthE
204.
205.
206.
207.
208.
209.
210-211.
212-213,
214-238.
239.

240-241,
242,
243.
244,
245.
246.

247-248.
249.
250.
251.
252.
253.
25%.

255-269.

37

L. V., Wilson

Gale Young

H. C. Young

J. P. Young

E. L. Youngblood

F. C. Zapp

Central Research Library

Document Reference Section

Laboratory Records (LRD)

Laboratory Records - Record Copy (LRD-RC)

EXTERNAL DISTRIBUTION

F. Cope, AEC-CRNL R.D.T. Site Office
W. Crawford, AEC-RDT, Washington

B. Deering, AEC-ORO

Giambusso, AEC-Washington

Larkin, AEC-ORO

Matthews, AEC-0ORO

McIntosh, AEC, Washington

Miller, Jr., AEC, Washington
Resnich, AEC, Washington

. M. Roth, AEC-ORO

lton Shaw, AEC, Washington

W. L. Smalley, AEC-ORO

R. F. Sweek, AEC-Washington

Division of Technical Information Extension (DTIE)

EHE=G

D
J
C
A
W.
C.
T
C
B
H
Mi