NAT_MSBRrecycle.txt

ENGINEERING DEVELOPMENT OF

THE MSBR FUEL RECYCLE

M. E. WHATLE,Y, L. E. McNEESE, W. L.. CARTER,
L. M. FERRIS, and E. L. NICHOLSON Chemical Technology
Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830

Received August 4, 1969
Revised October 13, 1969

 

The wmolten-sall breeder reactor being de-
veloped at Oak Ridge National Laboratory (ORNL)
requives continuous chemical processing of the
fuel salt, "LiF-BeF;-ThF, (72-16-12 mole%) con-

taining ~ 0.3 mole% *3UF,. The wvreactor and
the processing plant are planned as an integral
system. The wmain functions of the processing
plant will be to isolate ?°Pa from the neutvon flux
and to remove the rare-eavih fission products.
The processing method being developed involves
the selective chemical reduction of the wvarious
components into liquid bismuith solutions at
~600°C, utilizing multistage countev-cuvrrent ex-
traction. Protactinium, which is easily separated
from uvanitum, thovium, and the rare eavihs, would
be trapped in the salt phase in a storage tank
located between two extraction contactors and
allowed to decay to #3U. Rare earths would be
separated from thorvium by a similar reductive
extraction method; however, this operation will
not be as simple as the protactinium isolation step
because the rarve-earth-thovium separation factors
are only 1.3 to 3.5. The proposed process would
employ electrolytic cells to simultaneously intro-
duce reductant into the bismuth phase al the
cathode and to veturn extrvacted materials to the
salt phase at the anode. The practicability of the
reductive extraction process depends on the suc-
cessful development of salt-wmetal contactors,
electrolytic cells, and suitable materials of con-
struction.

 

INTRODUCTION

Oak Ridge National Laboratory is engaged in
the development of a molten-salt breeder reactor
that would operate on the ?**Th-?3%U fuel cycle.

170 | NUCLEAR APPLICATIONS & TECHNOLOGY

 

KEYWORDS: molten-salt re-
actors, fused salt fuel, repro-
cessing, chemical reactions,
reduction, bismuth, liquid
metals, fission products, fuel
cycle, MSBR, separation pro-
cesses, extraction columns

The reference reactor' is a single-fluid, two-
region machine containing ~1500 ft® of carrier
salt having the composition 71.7 mole% ‘LiF, 16
mole% BeF;, 12 mole% ThF,, and ~0.3 mole%
*%UF4. The reactor system would be fabricated of
Hastelloy-N, and would use graphite as a modera-
tor; corrosion of the Hastelloy is very low when
~1% of the uranium in the salt is present as UFs.
Calculations have shown that single-fluid molten-
salt reactors designed to operate economically at
reasonable power densities and fuel inventories
will not breed unless neutron absorbers such as
fission products (mainly xenon and rare earths)
and ***Pa (which is formed from ?**Th and decays
to ***U) are continually removed from the salt.
Protactinium-233, which has a neutron-capture
cross section of ~43 b, must be removed from the
neutron flux on a short time cycle (3 to 5 days).
Rare earths should be removed on a cycle of 30 to
60 days. The chemical processing system for ef-
fecting these separations must be close-coupled to
the reactor to minimize fuel inventory.

The salt from the reactor, even after allowing
1 h for decay of very short-lived nuclides, has a
specific heat generation rate of ~10 kW/ft°. At
various places in the processing plant the protac-
tinium and fission products will be concentrated,
giving rise to heat generation rates that are 3 to 5
times this value. The protactinium isolated in the
processing plant will generate ~5 MW of decay
heat. Thus, the chemical processing system must
be designed to handle much higher levels of radia-
tion and heat generation than are encountered in
the existing aqueous processes for water-cooled
reactor fuels. The separations process being
evaluated involves the selective reduction and ex-
traction at ~600°C of the various species from the
salt into liquid bismuth that contains lithium and
thorium as the reductants. Progress in the devel-
opment of this process is the subject of this
paper.

VOL. 8 FEBRUARY 1970
CHEMISTRY OF THE REDUCTIVE
EXTRACTION PROCESS

Bismuth is a noble metal that does not react
with the components of the fuel salt, but will dis-
solve metallic lithium, uranium, thorium, and the
rare earths to a reasonable extent. (Beryllium,
on the other hand, is almost insoluble in bismuth.)
Bismuth has a low melting point (271°C) and a
high boiling point (1477°C); its vapor pressure is
negligible in the temperature range of interest,
500 to 700°C. These properties, and the fact that
it is almost completely immiscible with a variety
of molten-fluoride salts, made it the first choice
for the metal phase in the reductive extraction
process.

The relative extractabilities of the important
actinide and lanthanide elements were determined
by measuring equilibrium distribution coefficients
in the two-phase system. The extraction of a
metal fluoride, MF, , from the salt into a liquid
bismuth solution can be expressed as the equilib-
rium reaction |

MF, (salo + # Li(si) = M(pij) + 7 LiF(salr) ,

in which #z is the valence of the metal in the salt
phase. An equilibrium constant for this reaction
can be written as

n
XMYM QLiF
n n
XMmF, YMF, XLi YLi

n
K= M aLfliF _
aMFnaLi

 

(1)

in which a denotes activity, X is mole fraction,
and y is an activity coefficient. Under the experi-
mental conditions used, g ;pand the individual ac-
tivity coefficients were essentially constant; thus,
Eq. (1) reduces to

XM

K'=—"M
Pl
XMFn XLi

(2)

The distribution coefficient for component M is
defined by

_mole fraction of M in bismuth phase _ XM

 

" mole fraction of MF, in salt phase XuE
"(3)
Combination of Eqgs. (2) and (3) gives
D=Xp; K", (4)
or, in logarithmic form,
log D=nlog X1; + log K" . (5)

Thus, a plot of log D vs the logarithm of the lith-
jum concentration in the metal phase (mole frac-
tion or at.%) should be linear with a slope equal to
n. The ease with which one component can be
separated from another is indicated by the ratio of

NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8

Whatley et al. MSBR FUEL RECYCLE
their respective distribution coefficients, i.e., by
the separation factor a. If the separation factor
for two components designated A and B (a =
Da/Dg) is 1, no separation is possible; the greater
the deviation of o from 1, the easier the separa-
tion. |

Data®™® obtained at 600°C using LiF-BeF:-ThF,
(72-16-12 mole%) as the salt phase are summa-
rized in Fig. 1 as plots of log D vs log Cr; . The
slopes of the lines show that, under the conditions
used, zirconium, thorium, and protactinium exist
as tetravalent species in the salt; uranium, pluto-
nium, and rare earths other than europium are
trivalent; and only europium is reduced to the
divalent state prior to extraction. Uranium and
zirconium are the most easily reduced of the spe-
cies shown. In fact, except for the difference in
valence, their behavior is almost identical. Thus,
zirconium, which is a fission product of high
yield, will coextract with uranium in the reductive
extraction process. Uranium and protactinium
should be easily separated, and, under the proper
conditions, a Pu-Pa separation is possible. Under
one expected operating condition, where Dy is ~1,
the corresponding U-Pa and Pa-Th separation
factors are ~100 and 3000, respectively. These
separation factors comprise the basis for the pro-
tactinium isolation flow sheet. As indicated in
Fig. 1, and as shown on an enlarged scale in Fig.
2, the rare-earth-thorium separation factors are
only in the range of ~1.3 to 3.5 under the desired
operating conditions ( Cp;> 0.1 at.%). Thus, re-
moval of the rare earths by the reductive extrac-
tion method will be much more difficult than the
isolation of protactinium.

As noted above, most of the components of in-
terest are adequately soluble in bismuth. Thorium
is the least soluble of the extractable components;
its solubility at 600°C is ~1800 wt ppm.® Uranium
and plutonium (which could be used as the fissile
material for starting up an MSBR) are at least
five times more soluble than thorium. Previously,
no information was available on the solubility of
protactinium in bismuth. However, recent work®
indicated its solubility to be ~1200 wt ppm at
500°C and >2100 ppm at 600°C. By assuming that
the effect of temperature on the solubility between
500 and 700°C is about the same as it is for the
other actinide metals, the solubility of protac-
tinium at 600°C has been estimated to be 4500 wt
ppm.* This concentration is more than adequate
to satisfy the process requirements.

Mutual solubilities of most of the major com-
ponents in bismuth appear to be high enough for
process application. Nickel is the only component
encountered so far that causes a marked effect.
The presence of as little as 100 wt ppm nickel in a
bismuth solution that is nearly saturated with

FEBRUARY 1970 171
Whatley et al. MSBR FUEL RECYCLE

 

  
   
   

 

 

 

2
10 I llllllll T T IIITITI
Zr (1V) U ()
Pu/(l11)
SALT: LiF-BeF,-ThF,
- (72-16-12 MOLE%)
TEMP : 600 C°
10 |- —_
Pa (IV)
-
E
§ 1.0 —
T
™
W
O
O
Z
o
-
>
o,
- ]
.Z 10~} |
o
Nd (1)
10~2}- —
Eu (I1)
Th (1V)
0-3 L1 1 a1l L1t
0.001 0.01 0.1 0.2

LITHIUM CONCENTRATION IN METAL PHASE (at.%)

Fig. 1. Distribution coefficients of major components

between a bismuth phase and a single-fluid
MSBR salt.

thorium can result in the precipitation of an insol-
uble nickel- and thorium-containing intermetallic
compound.®® One method for removing nickel is
described below.

THE CONCEPTUAL PROCESS FLOW SHEET

The principal engineering features of the con-
ceptual process®”’ are combined in a simplified

172 NUCLEAR APPLICATIONS & TECHNOLOGY

 

 

 

 

 

0.05 T | T
0.04 |— SALT: LiF-BeF,-ThF, | Nd (I1t) —
(72-16-12 MOLE%)
TEMP.: 600 C°
0.03 | — —
La (H)
- 0.02— Eu (1) ]
w
O
L
L
L
O
O
Z
3 Th (IV)
-
D
o
o
% 0.01 —
Q
0.008 |— —
0.006 |— —
0.004 1o |
0.05 0.07 0.1 0.2
LITHIUM CONCENTRATION IN METAL PHASE (at.%)
Fig. 2. Distribution of thorium and selected rare earths

between a single-fluid MSBR salt and a bismuth
phase,

flow sheet shown in Fig. 3. In this process fuel
salt from the reactor, after <1 h of cooling, en-
ters the bottom of the protactinium isolation sys-
tem at a rate of ~2.5 gal/min. This system
consists of two 7-in.-diam extractors, each having
six stages. The extractors are separated by a
200-ft° decay tank. Uranium is extracted from the
fuel salt into the bismuth, and the protactinium is
concentrated and trapped in the decay tank, re-
sulting in its removal from the reactor on a 3- to
5-day cycle. The decay tank is actually a heat-
exchanger in which protactinium decay heat is re-
moved.

The bismuth is continuously circulated through
the protactinium isolation system contactors and
an electrolytic cell. At the anode of the cell, ura-
nium present in the bismuth is oxidized to U*T,

VOL. 8 FEBRUARY 1970
MSBR FUEL RECYCLE

Whatley et al.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4OLIVINOD
39VIS €

 

 

 

 

dN IANYW 1VS

‘JGSIN © Sursseooxd J0J 199YS MOTJ UOTJOBIIXS SATIONPSI renydoouo) ‘g *S1d
p-Obx =17 cH/MM 67LE
,0rx02=ul _oix vg'="40d
SNOILOVYS 3TOW NI N3AID 38V SNOILVYLINIONOD 17V g O1x9°'g =0d 5 ¢oo0="n
. 91000=n (31040 Aop €)
wdb 6)°0 : .
69000 =34 . 18 wdb €S ouuuzwhq._a - wdo €se
371260 ONISS3008d3Y 17vs 13n4
e N.._ 1 Aop/sajow 2°9 b/MY B bE
910 = %48 31SUM OL [ ixge
2.°0=1 %ii7- 41 HOLOTHLX c-Obx €€ =0d (370AD Aop 09)
— — ————— o ——— o~ _ —-d. ‘C= wdb 924¢°
r 6900°0=3Y4 Tovis o o 01X Le=N 9210
| 31SvMm OL 1g wdb G
_ Aop /1TVS ¢4 66°0 . c /M 82 5 5
_ 5 4 °H-3H 1000 =0d fl fl
' | wdb ® mlOvXN =0
_ i s ¥ w6800 24 002
_ 4ON14 HON TS | | ANvL VAW (ed yONS
[ 370403y S | OYQAH |—e AVO30Q od 4
| 2 dOLOVYLX3
_ 39V1S 2} Y Y
| , _. Kop/sajow b g°1¢ H AH
| NOLOV3IN 5 ¥O1OVHLX3 200t
_ _ an 39VIS 9 15 4°N
2 2 9 3710403y
H | 3 ol1773N _u 2
_ | 0L%
t , , 4
TWAOW3Y | [0y 40N 4ON dvyl 1| 3130 4ON dvil
°5 2600 [+ ] 2o00% 0109 : 94N 20004 [™] 3700
91000°0 =3y —_—
| b i (37040 kop 0g) | -OMX = '
4H %n wdb G20 91000 =ul| 9N
$S30X3 :
, 300410313/ 4 u 9C’I= 48
(SNONNILNOD - IN3S) _.j 1LA108.10313
Aop;e 14 2 ! 32 912 ¥O10V3Y g 190
Pt 22 ! 9 ¥010V3Y
¥0LOv¥1X3 | 5.00t 0L%3Nn
L 3oviS 2t | “5oN | ) o
T )
1 1
- = = HO1OVYLX3
cHO'BE  gl96E  chI6E CHOBE o4 96E A9V1S € A
e ————————
NOILONA3Y
AV230 °d ANV NOILYINWNOOY HL¥Y3 3yvy L UI_L Ny YAOW3Y
_.qu rqu .|fil_ rlql_ 9100°0 = wdb Hp'0 NOILdH0S8VY ‘9
r||||.—-.|||llln—-|.||||_'llllnl 91000=ul wh_D
shop 1} YO4 AVQ/LTVS ¢HO'E 900000 =3Y
wdbg2o
YOLOVINOD -2
39VIS € dH-"H
_.:._ zro="4ul
910 =428
3008103713 /54 1 2,0=
7730 211A10810373 kop /.44 64°0

173

FEBRUARY 1970

VOL. 8

NUCLEAR APPLICATIONS & TECHNOLOGY
Whatley et al. MSBR FUEL RECYCLE

which transfers to the salt stream that is returned
to the reactor. At the cathode of the cell, Th*t and
Li" from the salt are reduced to metals which
dissolve in the bismuth. The resulting Th-Li-Bi
solution flows into the top of the upper extractor.
A small side stream of salt from the lower ex-
tractor is fluorinated to remove uranium as UFs
for control purposes. The fluorination also re-
moves iodine, bromine, and oxygen from the salt.
After treatment to remove traces of fluorine, the
salt is returned to the decay tank. The UF¢ is de-
contaminated from fission products by passage
through hot sodium fluoride beds and is collected
after subsequent passage through a suitable con-
centration monitor. Most of the UF¢ is absorbed
in salt, reduced with hydrogen to UF4, and re-
turned to the reactor. Excess uranium is removed
and sold.

Batch fluorination of molten salt for uranium
recovery and decontamination is well-established
technology. A similar operation was carried out®
at the MSRE when the ***U fuel was replaced with
3%U. Small-scale tests have shown that continuous
fluorination will be feasible,® and that soluble UF,
i1s produced when UFg¢ is sorbed in salt in the
presence of hydrogen.'" In the preceding opera-
tions, container corrosion could be severe; hence,
protection of the wall by a layer of frozen salt is
being considered.

The bismuth stream from the first stage of the
lower extractor carries the uranium to the elec-
trolytic cell in which the uranium is oxidized,
About 1% of this stream is continuously treated
with hydrogen fluoride in the presence of a salt to
remove fission products (Zr, Zn, Ga, Cd, and Sn)
and corrosion products (Fe, Ni, and Cr). After
fluorination to remove the uranium, the salt is
discarded to waste. This operation removes fis-
sion product zirconium on a 200-day cycle. A
shorter cycle time may be necessary if nickel or
other contaminants build up excessively. Hydro-
fluorination of this side stream of bismuth would
provide a method for removing plutonium from the
circuit of a molten-salt breeder reactor that was
started up with plutonium.

The salt stream leaving the protactinium isola-
tion system contains only traces of protactinium
and uranium but contains practically all of the
rare earths. A portion of this salt stream is
withdrawn and sent to a reductive extraction
process’ for removing rare earths. The rare-
earth extraction system differs from the pro-
tactinium isolation system in that the highest
concentration of rare earths occurs at the lower
end of the contactor rather than in the middle.
The salt-feed stream would enter near the middle
of the contactor. Calculations have shown that a
contactor having 24 theoretical extraction stages

174 NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8

with a bismuth-to-salt flow rate ratio of 80 would
result in a discard salt with the rare earths ~60
times as concentrated as they occur in the reactor
salt. The salt discard rate is set so that the rare
earths are effectively removed on a 50-day cycle.
At this discard rate, the neutron loss to rare
earths in the reactor is kept at an acceptably low
level, and the alkali metal and alkaline earth fis-
sion products (which remain in the salt throughout
the process) are removed from the reactor on an
8- to 10-year cycle. The salt that is discarded
would have a heat generation rate of ~17 kW/ft°
and would have to be stored for radioactive decay.

The reconstituted fuel salt will contain a small
but unknown amount of bismuth. Most of this bis-
muth must be removed from the salt to ensure
that its concentration in the salt returning to the
reactor will not be high enough to cause corrosion
of the Hastelloy.

Not all of the fission products having signifi-
cant neutron capture cross sections would be
removed by the reductive extraction process.
Xenon, Krypton, and tritium will be removed from
the reactor as gases on about a 1-min cycle by a
helium purge. Experience with the MSRE has
shown that the noble metal fission products (e.g.
Mo, Ru, Tc, Rh, Nb, and Pd) are not present in the
salt as fluorides.” Instead, they apparently exist
in the metallic state because of the reducing con-
dition in the reactor. A portion of these metals is
deposited on the surfaces of the graphite and the
Hastelloy, and the rest is present in the gas phase
in the form of a smoke. It is estimated that at
least half of the noble metal fission products will
also be removed from the MSBR by the helium
purge. Thus, the reactor off-gas system must be
designed to handle a significant amount of gaseous
and particulate fission products.

PROTACTINIUM ISOLATION SYSTEM CALCULATIONS

The distribution coefficient data for the com-
ponents of interest (Fig. 1) provide a firm basis
for calculation of the separations attainable in a
multistage countercurrent extraction system. A
typical set of concentration profiles for this Sys-
tem’ with six theoretical stages below the decay
tank and six theoretical stages above, is shown in
Fig. 4. The points beyond the left margin of the
figure represent the composition of the reactor
fuel salt entering the extraction system. The
maximum concentrations of uranium, protactini-
um, and thorium in the metal phase are limited by
the solubility of thorium in bismuth, which, in
turn, governs the salt-bismuth flow rate ratio.
The protactinium concentrations in both the salt
phase and the metal phase reach maxima in the
vicinity of the decay tank, where the protactinium

FEBRUARY 1970
concentration in the salt is two orders of magni-
tude higher than that in the fuel salt entering
the extractor. Routing the high-protactinium-
concentration salt stream through a 200-ft° decay
tank results in retention of >90% of the protactini-
um (in the reactor and the chemical processing
plant) in the tank, thus obtaining low protactinium
losses by neutron capture in the reactor.

The concentration profile (Fig. 4) represents a
steady state at very nearly optimum conditions.
However, two significant effects are not readily
apparent in this representation. The firstis that
the profile is very sensitive to the amount of re-
ductant entering the system. Thus, a small error
in this amount, caused by either a change in the
reductant concentration in the bismuth or by a
flow-rate-ratio change, could change the location
of the protactinium. The net effect could be the
eventual return of all the protactinium to the re-

 

     
 

 

 

 

 

 

 

-0
oLiF T 1 T T T T T T T T 10
2$§2 Pa DECAY TANK
- INLET CONC. | TANK CONC. —107!
U=369%x10%| 1.75 x10-
Pa =1.326 x10-3| 1.312x 1073
- _.|O-2
Liand Thin METAL
0——0——0 0——0O 5
TIO
C
—1074
Z
=107 ©
<
@
[TH
[FT]
6 O
[ — o _"O 2
<
o Th in METAL o/
- 0 ~—0 ——0 ——
= _q
& —10
< 0 es () cvems O
Ll
O
Z
3 3
O - —10
O
-
O
<
L
o a 10”7
a
\ o
] | | ] ] ] i | | 1\ 1010
| 2 3 &4 5 6 1 8 9 10 I 12
STAGE NUMBER
Fig. 4. Calculated concentration profiles in the pro-

tactinium isolation column,

NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8

Whatley et al. MSBR FUEL RECYCLE
actor. The second effect is the stabilizing influ-
ence on the system of the capacitance that is
provided by the large volume of salt in the decay
tank.

The effect of flow rate on a system at steady
state is shown in Fig. 5. The dotted line repre-
sents operation in the mode described above. The
concentration of protactinium in the reactor de-
creases to an optimum from which, under a sim-
ple steady-state analysis, a flow rate variation as
small as 1% could result in a tenfold increase in
the protactinium concentration in the reactor if
the flow rate were low, or a threefold increase if
the rate were high. Removal of a small amount of
uranium, for example by fluorination of 2% of the
salt entering the decay tank, would provide con-
siderable relief of the control system sensitivity
(Fig. 5). The typical concentration profiles (Fig.
4) show that the uranium concentration changes
rapidly from stage to stage in the vicinity of the
decay tank and would also be sensitive to flow rate
ratios. In fact, the uranium concentration drops
to 0.0005 of its concentration in the reactor when
the bismuth flow rate is slightly above the opti-
mum, and drops by a factor of 400 per 0.1% in-
crease in the bismuth flow rate close to the
optimum. Thus, the uranium concentration in the
salt provides a very sensitive index to operation,
and by controlling it at ~0.007 times the reactor
concentration (within a factor of 5 or 10), the sys-
tem can be held sufficiently close to the optimum
flow conditions. As noted before, uranium re-
moval from the system would be accomplished by
continuously fluorinating a small side stream of
salt entering the decay tank. Monitoring the UFg
concentration in the gas from the fluorinator pro-
vides a sensitive measurement of the uranium
concentration in the decay tank and, consequently,
the location of the protactinium in the contactor.

Calculations® of the transient behavior of the
protactinium isolation system were made to gain
information about the stability of the system. In
the computer program used, the uranium concen-
tration at the inlet to the decay tank controlled the
bismuth flow rate. The other input information
was similar to that used to generate the con-
centration profile. A random error with a 5%
standard deviation was superimposed upon the
controlled rate which was maintained constant for
1 h; then, a new input generated a different flow
rate with a new random error. Even with this
crude control system, over 85% of the protactini-
um was held outside of the reactor.

THE ELECTROLYTIC CELLS

An ideal electrolytic cell for use in the process
would receive bismuth containing extracted com-

FEBRUARY 1970 175
Whatley et al. MSBR FUEL RECYCLE

 

 

 

 

 

250 T T T T I T I T l
2 P~ <
E \\ REACTOR VOLUME 1042 FT3
g 20 DECAY TANK VOLUME 208 FT3
= 0 SALT FLOW RATE 1.8 GAL/MIN
é . TOWER ABOVE TANK 6 STAGES
S \ TOWER BELOW TANK 6 STAGES
o
x 150 \\ _
5 \
g \'_NO URANIUM
2 REMOVAL \
w 100 |- \
g \
o \
m \
o 50 \ -
o \
Q \
I \
(1 o
o ] 1 1 | | ] ] 1 |
35 3.52 354 3.56 358 36 3.62 3.64 3.66 3.68 3.7

BISMUTH FLOW RATE,GAL/ MIN

Fig. 5. Effect of Bi flow rate in reductive extraction tower on Pa concentration in the reactor,

ponents and would completely oxidize these com-
ponents to fluoride salts which would be carried
out of the cell. The purified bismuth would then
be routed to the cathode where thorium fluoride
from the salt would be reduced to produce a bis-
muth metal phase that is suitable for use as an
extractant. In practice, both lithium and thorium
metals would be produced at the cathode, and the
anode would supply BiF; which would subsequently
react with the uranium or the rare earths in aux-
iliary contactors provided for this reaction to
strip them into the salt phase.

In the rare-earth removal system the cell
would be operated as the center unit of a three
unit complex with salt-bismuth contactors located
above and below it. The cell would operate with
pure bismuth being fed to both the cathode and
anode. The salt in the cell would be practically
free of the rare earths and thorium. In the con-
tactor above the cell, the extracted components
would be removed from the incoming bismuth
stream by oxidizing them with BiF; (produced at
the anode of the cell) and transferring them to the
salt phase. In the lower contactor, the bismuth
leaving the cell, which contains lithium produced
at the cathode, would be contacted with the in-
coming salt, and thorium would be transferred to
the bismuth phase. In the protactinium isolation
system, only the top contactor would be used.

The BiF; that is produced at the anode of the
rare-earth removal system could attain a concen-
tration of up to 20 mole% in the salt. Operation
with high concentrations of this fluoride in the cell
could seriously affect cell efficiency. The BiF;

176 NUCLEAR APPLICATIONS & TECHNOLOGY

concentration in the cell might be controlled by
introducing hydrogen to form hydrogen fluoride,
which would leave the cell promptly; HF is as
effective as BiF; in performing the oxidation.

The cell arrangement described above would
allow the electrodes to be placed in close prox-
imity. This is important because the resistivity
of the salt is fairly high (0.5 @ cm), and a long
current path would cause intolerable electrical
losses and accentuate the heat removal problem.
Corrosion protection of the anode, as well as
electrical insulation, would be provided by freez-
ing a layer of salt on all exposed metal surfaces.
Ample internal heat from radioactive decay and
electrical losses is available to generate the
required thermal gradient for establishing and
maintaining the protective frozen wall.

The theoretical current requirement for the
protactinium isolation system is ~10 000 A. Cur-
rent efficiencies of at least 50% at a current den-
sity of ~5000 A/ft* might be achieved, based on
commercial practice in the aluminum industry.
The protactinium isolation system would then re-
quire ~4 ft* of surface for each electrode, while
the rare-earth removal system would require a
cell that is about three times this size. The elec-
trodes would consist of pools of bismuth in the
flowing bismuth stream.

Alternatively, it is possible that the reducing
agent (either thorium or lithium) and the oxidizing
agent (e.g., hydrogen fluoride) could be supplied
from an external source, eliminating the need for
electrolytic cells. In such a case, the protactini-
um isolation and the rare-earth removal systems

VOL. 8 FEBRUARY 1970
would require ~550 and ~1500 kg of thorium per
day, respectively. However, the cost of this
thorium and the resulting salt discard would be
prohibitive. Thus, it is concluded that electrolytic
cells are required for economical operation of the
processing system.

In experiments with simple static electrolytic
cells,” current densities in excess of the desired
5000 A/ft> were obtained, and BiF; and lithium
(the salt contained no thorium) were produced at
the anode and cathode, respectively. This experi-
mental series also included the first attempts to
protect the anode from corrosion and provide
electrical insulation with frozen salt. A large-
scale, fully continuous system for electrolytic cell
development is being installed to permit experi-
ments with flowing salt and bismuth.

EXTRACTORS

The required multistage countercurrent ex-
tractors present unique problems due to the phys-
ical properties of the liquids, the unusually high
flow ratios required in the rare-earth removal
system, and the high rate of radioactive-decay
heat generation. The contactor development ef-
fort”’** has, so far, been conducted with mercury
and water to simulate bismuth and salt to permit
estimation of flooding rates, pressure drops, and
holdup with high-density fluids. The density dif-
ference divided by the average density is 2 for the
mercury-water system, 1.0 for the salt-bismuth
system, and ~0.2 for liquid-liquid extraction sys-
tems using organic solvents. It is interesting to
note that the salt-bismuth system may more
closely resemble a distillation system (where this
number may be 2) than a conventional liquid-liquid
system.

The low distribution coefficients available in
the rare earth removal system (Fig. 2) make con-
tactor operation at very high metal-to-salt flow
rate ratios (30 to 80) mandatory. Backmixing
would be expected in a conventional packed column
which would limit performance. Radioactive-decay
heat generation in the salt phase is markedly dif-
ferent in the protactinium isolation and rare earth
removal system. In the protactinium isolation
system, a tenfold change in specific heat genera-
tion occurs between the protactinium decay tank
withdrawal point and the top of the extractor, with
the maximum heat-generation rate being ~34
kW /ft®. Thus, it is apparent that a suitable con-
tactor will require considerable development.

STATUS OF THE DEVELOPMENT WORK

During the past two years, ORNL has been
engaged in studies aimed at developing basic

NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8

Whatley et al. MSBR FUEL RECYCLE
chemical information for the reductive extraction
processing system and in defining the engineering
requirements for this system. During this time,
many of the important problems associated with
the chemistry have been solved. Experiments in
simple equipment that will permit operation in
contactors and electrolytic cells with flowing salt
and bismuth are in progress, while design and
construction of larger, more comprehensive ex-
periments which include electrolytic cell and con-
tactor development are under way.

The selection of a material for construction of
the processing plant requires investigation re-
garding two major areas: (a) surfaces exposed to
oxidizing conditions, and (b) surfaces exposed to
salt and liquid bismuth. As previously mentioned,
a layer of frozen salt will probably serve to pro-
tect surfaces that are exposed to an oxidizing en-
vironment if the layer can be maintained in the
complex equipment. Preliminary experiments'
with a simple vertical column filled with molten
salt and with a centerline heater showed that the
thickness of the frozen film on the wall was pre-
dictable and adhered to the wall. The simulta-
neous containment of bismuth and salt presents a
special problem. The only materials that are
truly resistant to bismuth are refractory metals
and graphite, neither of which is attractive for
fabricating a large and complicated system. To
date, experimental equipment has been built of
low-carbon steel, as an expedient. Development
work to determine if iron-base alloys can be pro-
tected with refractory metal coatings is starting.

Although the development of the chemical pro-
cesses for MSBR’s is at an early stage, we have
attempted to estimate the cost of processing. The
uncertainties of such an estimate are both numer-
ous and serious. Results of previous work™ and a
rough estimate of the complexity of the present
process, suggest that the capital cost of the plant
considered by Perry et al.’® might be ~$8 x 10°
with an operating cost of ~$10°/year, giving a unit
processing cost of ~0.3 mill/kWh, based on an
80% load factor. As additional processing de-
velopment information becomes available, more
meaningful estimates can be made. A more de-
tailed cost study based on the conceptual flow
sheet is under way.

In conclusion, processes for the isolation of
protactinium from the reactor salt and for the re-
moval of the rare-earth fission product poisons
from MSBR’s have been studied through the phase
of establishing chemical feasibility. The pro-
tactinium isolation system is simpler and perhaps
more tractable than the rare-earth removal sys-
tem because of the higher distribution coefficients
and separation factors available in the former
system. Although cost estimates at this time

FEBRUARY 1970 177
Whatley et al. MSBR FUEL RECYCLE

should be considered highly tentative, it appears
that MSBR’s with associated chemical processing
systems hold attractive economic potential in the
evolving nuclear industry.

ACKNOWLEDGMENTS

The authors thank the many people in the Chemical
Technology Division and Reactor Chemistry Division of
ORNL who contributed to this work, especially D. E,
Ferguson for his encouragement and guidance. We are
also indebted to the groups of W. R. Laing, C. Feldman,
J. H. Cooper, and W. T. Mullins of the Analytical
Chemistry Division who performed the many difficult
analyses and did the necessary analytical development,
This work was sponsored by the U.S. Atomic Energy
Commission under contract with the Union Carbide
Corporation.

REFERENCES

1. E. S. BETTIS and R. C. ROBERTSON, ‘“MSBR De-
sign and Performance Features,” Nucl. Appl. Tech., 8,
190 (1970).

2. D. E. FERGUSON, ‘‘Chemical Technology Division
Annual Progress Report, May 31, 1969, USAEC Report
ORNL-4422, Oak Ridge National Laboratory.

3. L. M. FERRIS, J. C. MAILEN, J. J. LAWRANCE,
and F. J. SMITH, ‘“MSR Program Semiannual Prog-
ress Report, February 28, 1969,”” USAEC Report ORNL-
4396, Oak Ridge National Laboratory (1969).

4. L. M. FERRIS, J. C. MAILEN, F. J. SMITH, E. D.
NOGUEIRA, J. H. SHAFFER, D. M. MOULTON, C. J.
BARTON, and R. G. ROSS, ‘‘Isolation of Protactinium
from Single-Fluid Molten-Salt Breeder Reactor Fuels
by Selective Extraction into Li-Th-Bi Solutions,”’ to be
published in the Proc. Thivd Intern. Protactinium Conf.

5. L. M. FERRIS, J. C. MAILEN, J. J. LAWRANCE,
F. J. SMITH, and C. E. SCHILLING, ‘“MSR Program
Semiannual Progress Report, August 31, 1968, USAEC
Report ORNL-4344, pp. 292-297, Oak Ridge National
Laboratory (1969).

6. M. E. WHATLEY and L. E, McNEESE, ‘“MSR Pro-
gram Semiannual Progress Report, February 29, 1968,’’

178 NUCLEAR APPLICATIONS & TECHNOLOGY

USAEC Report ORNL-4254, pp. 248-251, Oak Ridge
National Laboratory (1968).

7. L. E. McNEESE, ‘“MSR Program Semiannual Prog-
ress Report, February 28, 1969, USAEC Report ORNL-
4396, Oak Ridge National Laboratory (1969).

8. J. M. CHANDLER and R. B. LINDAUER, ‘“Prepara-
tion and Processing of MSRE Fuel,’”’ to be published in
the Proc. AIME Symp. on Reprocessing of Nuclear
Fuels, August 1969.

9. D. E. FERGUSON, ‘““Chemical Technology Division,
Annual Progress Report, May 31, 1966,”> USAEC Report
ORNL-3945, p. 72, Oak Ridge National Laboratory (1966).

10. L. E. McNEESE, “Fuel Reconstitution,’”” in ‘‘Unit
Operations Section Quarterly Progress Report, April-
June 1965,> USAEC Report ORNL-3868, p. 43, Oak
Ridge National Laboratory (1965).

11. L. E. McNEESE and C. D. SCOTT, ‘‘Reconstitution
of MSR Fuel by Reduction of UFg Gas to UF, in Molten
Salt,””> USAEC Report ORNL-TM-1051, Oak Ridge Na-
tional Laboratory (1965).

12, W. R. GRIMES, ‘‘Molten Salt Reactor Chemistry,”’
Nucl. Appl. Tech., 8, 137 (1970),

13. J. S. WATSON and L. E. McNEESE, ‘‘Simulated
Salt-Metal Contactor Studies,’’ in ‘‘Unit Operations
Section Quarterly Progress Report, July-September
1968, USAEC Report ORNL-4366, Oak Ridge National
Laboratory (in press).

14. B. A. HANNAFORD and L. E. McNEESE, “MSR
Program Semiannual Progress Report, August,31, 1968,”’
USAEC Report ORNL-4344, pp. 302-306, Oak Ridge
National Laboratory (1969).

15. C. D. SCOTT and W. L. CARTER, ““Preliminary
Design Study of a Continuous Fluorination Vacuum-
Distillation System for Regenerating Fuel and Fertile
Streams in a Molten Salt Breeder Reactor,”’” USAEC
Report ORNL-3791, Oak Ridge National Laboratory
(1966).

16. A. M. PERRY and H. F, BAUMAN, ‘‘Reactor Physics
and Fuel-Cycle Analyses,”” Nucl. Appl. Tech., 8, 208
(1970). |

VOL. 8 FEBRUARY 1970