ORNL-TM-3258.txt

 

-

WY S KT2 s

o ----R»-II-."E_-:-‘
AL 1
|

O/ l” '._lli.i||-ll|lll'I|||Il-||||.i.;||||’.l I ||||| l'“' |||'IM \L LABORATORY

by
445k 0566291 2
2 ON  NUCLEAR DIVISION

U.5. ATOMIC ENERGY COMMISSION
ORNL- TM- 3258

ABORAT

 

f"‘b{.: ,::"_f T
M@%?fi& Co . o

ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT
BREEDER REACTOR PROCESSING NO, 8

L. E. McNeese

NOTICE This document contains information of a preliminary nature
and waos prepared primorily for internol use ot the Ook Ridge MNotional
Laoboratery. It is subject to revision or correction and therefore does

not represent o final report.
ORNL-TM-3258

Contract No. W-TL405-eng-26

CHEMICAL TECHNOLOGY DIVISION

ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT
BREEDER REACTOR PROCESSING NO. 8

L. E. McNeese

MAY 1972

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee 37830
operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION
 

ii

Reports previously issued in this series are as follows:

ORNL-4366

ORNL-TM-3053
ORNL-TM-3137
ORNL-TM-3138
ORNIL-TM-3139
ORNL~-TM~3140
ORNL~-TM=-31L41
ORNL-TM-3257

Period
Period
Period
Period
Period
Period
Period

Period

ending
ending
ending
ending
ending
ending
ending

ending

September 1968
December 1968
March 1969
June 1969
September 1969
December 1969
March 1970
June 1970
iii

CONTENTS
SUMMARTES &« ¢ o v v v v 0 v 6 s o 6 4 4 o o e e e e e e e e e e e e
1. INTRODUCTION. « & o s o o o o o o o o o o s s o o o o o o o o o
2. ANALYSIS OF THE FLUORINATION--REDUCTIVE EXTRACTION AND METAL
TRANSFER FLOWSHEET. . ¢ o ¢ & o & o ¢ o o s o o o o s o o o » o o
2.1 Isolation of Protactinium in a Secondary Salt, Using Fluori-
nation and Reductive Extraction . . . « « « « v v « « +« . .
2.2 Mathematical Analysis of Flowsheet in Which Protactinium Is
Isolated in a Secondary Salt, Using Fluorination and Reduc-
tive Extraction . « « ¢« & « v o 0 v 0 e 0 0 e e e e e e e
2.3 Calculated Results on Operation of the Protactinium Isolation
System o ¢ v v v e e e e e e e e e e e e e e e e e e e
2.4 Combination of Discard Streams from the Fluorination--Reduc-
tive Extraction-Metal Transfer Flowsheet. . . . . . . . . .
3. ANALYSIS OF URANIUM REMOVAL FROM FUEL SALT BY OXIDE PRECIPITATION
3.1 Mathematical Analysis of a Uranium Oxide Precipitator . . .
3.2 Calculated Results « ¢ ¢ ¢ o o o o o o o o o o o o o o o
4, DEVELOPMENT OF A FROZEN-WALL FLUORINATOR: INSTALLATION OF A SIM-
ULATED FLUORINATOR FOR STUDYING INDUCTION HEATING « « o ¢ o + & &
4.1 Experimental Equipment. . . . . « « +« « « o o . . . .
L.2 Status of Equipment Installation. . . . « « « &« o « & « + .
5.

MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS AND GAS HOLDUP IN
OPEN BUBBLE COLUMNS . e e e e e e e

5.1 Previous Studies on Axial Dispersion. . « « « « + o o + &

5.2 Mathematical Analysis of Unsteady-State Axial Dispersion in
a Bubble CoOlUmMN « « ¢ « o o o o o o o o o o o s o o o o o

5.3 EBquipment . . « « ¢« ¢ ¢ 0 0 4 e e e e e e e e e e
5.4  Experimental Procedure. . « « « ¢« + ¢« ¢ ¢+ o 0 e 0 0 0 .
5 . 5 Results . ° . * . . . . [] L] . o o ° . . e > . . . . . o ° .

5.6 Discussion of Data on Axial Dispersion. . . « « « + « . .

[
-
.
-
.
-
s
.
.

5.7 Discussion of Data on Gas Holdup. . . .

5.8 Conclusions « « « o o+ 4 0 0 e 4 e e e e e 4 e e 4 e s

12

19
21

21
2l

2k

26
29

31

31

33
38
Ly
L5
L5
54
56
iv

CONTENTS (continued)

Page
6. DEVELOPMENT OF THE METAL TRANSFER PROCESS + + « « « « = +« « « « « . 58
6.1 Equipment for Experiment MTE-2 . « « o« « « o « « o« « o o o « 59
6.2 Materials Used in Experiment MTE-2 . « « « 4« « « & o + « « + 63
6.3 Status of Experiment MTE-2 . « v v « ¢ « ¢« ¢« « o o o « o « o 63
7. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL
FACILITY . * . ® ® * - o » . » . . . » » » * . * -« . . . . . . . . 6h
T.1 Preparation for Mass Transfer Run UTR=3 . . «. + « « « « « o« . 65
7.2  Mass Transfer Run UTR=3 &« « v « « v « o o« o« o« o o« « o o o « o« 67
7.3 Preparation for Mass Transfer Run UTR-4 . . . . . . . . . . . 70
7.)4 MaSS TI‘anSfer Run UTR_)'I' . . . ° . . . . . . . . . . . . . . . 71

7.5 Mathematical Analysis of Mass Transfer with Primary Resist-
ance to Transfer in the Salt Phase. . ¢« + ¢« ¢ ¢« « « &« o o« « o« [5

7.6 Discussion of Mass Transfer Data from Runs UTR-3 and UTR-L. . 78

7.7 Preparation for Zirconium Mass Transfer Experiments; Run

UTR=5 + 4+ « o « o o« o o o+ o o o o e o o o o s o o v o« .80
7.8 Summary of Hydrodynamic Data with Present Column. . . . . . . 8L
T.9 Examination of 304 Stainless Steel Corrosion Specimens from
Salt-Metal Treatment Vessel o« + « 4 « ¢ o o o o o o o o o « o« 87
7.10 Maintenance of Equipment. . . « « ¢ ¢ o ¢« o o o o o o « « o o 87
8. PREVENTION OF AXIAL DISPERSION IN PACKED COLUMNS. + « « « &« s o « « 90

9. ELECTROLYTIC REDUCTION OF LiCl USING A BISMUTH CATHODE AND A

GRAPHITE ANODE . v v & ¢« o o o o o s o o o o o o o o o o o o o+ o « 92
9.1 Equipment and Materials Used. . + « « & & o o o« o o « « o« o« « Ok
9.2 Operating Conditions and ResultS. . « o « ¢« &+ & o« o o o « & o+ 9k
9.3 Postoperational Examination of Equipment . . . . . . « . . . 96
10. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS . . . . . . 97
10.1 Iron Fluoride Reduction Runs R-1l and R-2. .+ . « « + « « « « « 97

10.2 Mathematical Analysis of the Rate of Iron Fluoride Reduction;
Calculated Mass Transfer Coefficients for Runs R-1 and R-2. . 99

10.3 Equipment Modifications and Maintenance . . . . . « . « . . .101
11.

12.

CONTENTS (continued)

ELECTROLYTIC OXIDATION OF Pah+ to Pa5+ IN MSBR FUEL SALT .

11.1

action Between Protactinium and Uranium.

lll2

Between Protactinium and Uranium .

1103

REFERENCES

Calculated Results and Discussion.

-

.

.

Estimated Anode Current Density for the Case of No Inter-

Estimated Anode Current Density for the Case of Equilibrium

107

110
113

116
vii

SUMMARIES

ANALYSIS OF THE FLUORINATION-~REDUCTIVE EXTRACTION
AND METAL TRANSFER FLOWSHEET

An improved flowsheet was developed in which protactinium is isolated
from the fuel salt of an MSBR and held for decay in a secondary salt stream
that is physically and chemically isolated from the reactor. A processing
plant based on this flowsheet should be much easier to control than one
based on the earlier flowsheet in which the protactinium was isolated in
bismuth at a point intermediate in the protactinium extraction column. Use
of the new flowsheet should result in a considerable saving in capital equip-
ment cost. A mathematical analysis showed that essentially complete ex-
traction of the protactinium on a 10-day cycle could be obtained with five
equilibrium stages and a reductant addition rate of about 200 equiv/day
for a uranium removal efficiency of 99% in the primary fluorinator, or an
addition rate of about 300 equiv/day for a uranium removal efficiency of
95%. The number of equilibrium stages could be reduced to as few as three
without increasing the protactinium removal time appreciably if the reductant
addition rate were increased to 371 equiv/day for a primary fluorinator ef-
ficiency of 95%, or to 257 equiv/day for a fluorinator efficiency of 99%.

A method was developed for combining and fluorinating the various waste
streams produced by the flowsheet. Use of this method will eliminate

several potential routes for loss of fissile material from the system.
ANALYSIS OF URANIUM REMOVAL FROM FUEL SALT BY OXIDE PRECIPITATION

Calculations were made to investigate the operation of an oxide pre-
cipitator for removing uranium from fuel salt that contains no protactinium.
The results indicate that greater than 99% of the uranium can be removed
with about three equilibrium stages in a countercurrent system and that the

UO,~ThO,. stream produced will have a UQ, concentration of greater than 90%.

2 2 2
Less than 1% of the thorium fed to the system would be precipitated with
the uranium. No significant effect on precipitator performance was ob-
served when the amount of salt remaining with the oxide during the transfer

of salt between stages was varied from 2 to 10 moles per mole of oxide.
 

viii

DEVELOPMENT OF A FROZEN-WALL FLUORINATOR: INSTALLATION OF A SIMULATED
FLUORINATOR FOR STUDYING INDUCTION HEATING

An experiment to demonstrate protection against corrosion by the use
of layers of frozen salt in a continuous fluorinator requires a corrosion-
resistant heat source to be present in the molten salt. High-frequency
induction heating has been proposed as the source of heat. There are un-
certainties 1in determining the effect of bubbles in the molten salt with
this heating method and in estimating the amount of heat that will be
generated in the metal walls of the fluorinator. Equipment is being in-
stalled for studying heat generation in a simulated frozen-wall fluorinator
containing provisions for induction heating. A 31 wt % HNO3 solution will
be used to simulate molten salt in the system. The simulated fluorinator
consists of a 5-in.-OD by 5-ft-long section of 8-in. 304 stainless steel
sched 40 pipe. The system also contains a pump for circulating the acid

through the column and a heat exchanger for removing heat that is generated

in the acid.

MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS AND
GAS HOLDUP IN OPEN BUBBLE COLUMNS

An improved experimental technique was developed for measuring axial
dispersion coefficients in open bubble columns. This technique consists of
injecting a small amount of KCl tracer solution at the top of a column and
using a conductivity probe to measure the rate at which the tracer is dis-
persed throughout the column. Twenty-nine runs were made in order to
measure the axial dispersion coefficient in open bubble columns having
diameters of 1.5, 2, and 3 in.; 59 runs were made to determine gas holdup.
The new technique appears to be superior to the earlier steady-state tech-
nique in that (1) less scatter is observed in the dispersion coefficient
data, and (2) data can usually be obtained in less than 10% of the time
required for the steady-state technique. The axial dispersion coefficients
obtained with the new technique are in agreement with those measured pre-
viously; however, data were obtained over a wider range of superficial gas

velocities in the present study. At low gas flow rates (where bubble flow
ix

occurs), gas holdup was found to be proportional to the superficial gas
velocity and independent of the column diameter. At higher gas flow rates,
the gas holdup was found to be inversely proportional to the column diameter

and to increase, in a gradual manner, as the gas flow rate increased.
DEVELOPMENT OF THE METAL TRANSFER PROCESS

A second engineering experiment (MTE-2) is presently in progress. The
experiment was designed to demonstrate all phases of an improved rare-earth
removal method known as the metal transfer process. The main objectives of
the experiment are: (1) to demonstrate the selective removal of rare earths
from fluoride salt containing thorium fluoride, (2) to collect the rare
earths in a lithium-bismuth solution, and (3) to verify previous distribution
coefficient data. Experiment MTE-2 is being performed at 660°C in a 6-in.-
diam carbon steel vessel that is divided into two compartments interconnected
at the bottom by a pool of thorium-saturated molten bismuth. One compartment
contains fluoride salt (72-16-12 mole % LiF-BeF,~ThF)) to whichT uCi of 1h4Tyg

and sufficient LaF., to produce a concentration of 0.3 mole % has been added.

The second compartient contains LiCl, a 35 at. % Li-Bi solution (in a cup),
and a pump for circulating the LiCl through the cup at a flow rate of about
25 cm3/min. The pump is constructed of carbon steel and uses molten bismuth
as check valves. Thus far, the experiment appears to be proceeding satis-

factorily; however, no data are available at this time.

SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS
IN A MILD-STEEL FACILITY

We continued to operate a mild-steel system which can be used to carry
out semicontinuous reductive extraction experiments. We are presently study-
ing the mass transfer performance of an 0.82-in.-ID, 2L4-in.-long column

packed with 1/4-in. molybdenum Raschig rings.

Following the second uranium mass transfer run (UTR-2), the salt and
bismuth were returned to the treatment vessel and 122.5 g of thorium metal

was added to the graphite crucible through a 3/b-in.-diam tube. The thorium
dissolved slowly in the bismuth, at about the same rate observed after
thorium was added to the bismuth feed tank prior to run UTR-2. A second
charge of thorium metal (119 g) was lowered into the bismuth phase in a
perforated steel basket to accelerate the reaction between the thorium

and the bismuth. After this addition, the thorium dissolution rate in-
creased to about ten times the rate observed earlier. Subsequent samples
of the bismuth revealed a nonuniform thorium concentration in the bismuth
pool, apparently the result of poor mixing during the period in which the
bismuth and salt approached chemical equilibrium following the reductant
addition. It was also found that thorium concentrates in the lower part

of a sample during the slow freezing of the bismuth because of the tendency
of the more dense thorium bismuthide to settle to the bottom. The previous
sample preparation method, which involves discarding the lower 15% of the
sample, was revised in order to avoid substantial errors in the reported

concentration of thorium in the bismuth.

Run UTR-3 was carried out with metal-to-salt flow rate ratios of 2.05,
1.22, and 0.91. Seven pairs of bismuth and salt samples were removed from
the salt and bismuth streams leaving the column. Analyses of these samples
showed that, as the bismuth-to-salt flow rate ratio was decreased from 2.05
to 0.91, the fraction of the uranium removed from the salt decreased from

0.91 to 0.73.

Following run UTR-3 the salt and bismuth phases were returned to the
treatment vessel. Sufficient thorium metal was then charged to the vessel
to produce a thorium concentration of about 1000 ppm, which is about 10%
greater than the solubility of thorium and bismuth at the bismuth feed tank
operating temperature (540°C). Observed variations in thorium concentration
in the bismuth appeared toc be due to insufficient mixing of the bismuth
phase. In run UTR-4, the fraction of uranium extracted from the salt in-
creased from 0.63 to 0.7L as the flow rate ratio was increased from 0.75 to

l.o.

In correlating the uranium extraction data from runs UTR-3 and -4, it
was assumed that the rate at which uranium transfers to the bismuth is con-—

trolled by the diffusive resistance in the salt film when the extraction
xi

factor is high and when the salt film is composed largely of nontrans-
ferring ions. It was found that the data could be correlated in terms
of the height of an overall transfer unit based on the salt phase. The
HTU value increased from O0.77 ft to 2.1 ft as the bismuth-to-salt flow

rate ratio decreased from 2.05 to 0.75.

In order to measure mass transfer rates in the column under more
closely controlled conditions and under conditions where the controlling
resistance is not necessarily in the salt phase, preparations were begun
for experiments in which the rate of exchange of zirconium isotopes will
be measured between salt and bismuth phases otherwise at chemical equilib-
rium. The salt and bismuth were transferred to the treatment vessel and
were contacted with a T0-30 mole % H2-HF mixture for 20 hr in order to
remove reductant from the bismuth phase. After being sparged successively
with hydrogen and argon, the salt and bismuth were then transferred to their
respective feed tanks. The reported uranium concentrations in salt samples
removed from the column effluent varied by i_35%, which was surprising
since no extraction of uranium from the salt occurred during the run. The
hydrodynamic data obtained during countercurrent flow of salt and bismuth
in the present column were found to be in good agreement with the predicted

column throughput at flooding, which is based on studies with mercury and

aqueous solutions.

PREVENTION OF AXTAL DISPERSION IN PACKED COLUMNS

Packed columns are being considered for use in countercurrently
contacting molten salt and bismuth in MSBR fuel processing systems. We
have previously made axial dispersion measurements in packed columns
during the countercurrent flow of mercury and aqueous solutions and have
shown that axial dispersion can significantly reduce column performance
under some operating conditions of interest. As part of our contactor
development program, we are evaluating column modifications that will
reduce the effect of axial dispersion to .an acceptable level. We have
devised and tested an improved axial dispersion preventer, which consists

of an inverted bubble cap having a single 3/8-in.-0D tube at the upper
xii

surface of the bubble cap. During operation, the metal phase accumulates
around the bottom of the bubble cap and the resulting seal forces the
continuous phase to flow through the tube in the upper surface of the

cap. This design allows the depth of the metal phase to increase to the
point where a sufficiently high head of liquid metal is produced to force
the metal through the dispersion preventer at a high throughput. The ex-
tent of axial dispersion observed with this design is probably sufficient-
ly low for most applications of interest. However, it is believed that
the present design allows for salt and bismuth throughputs that can be as

high as the column throughputs at flooding.

ELECTROLYTIC REDUCTION OF LiCl USING A BISMUTH CATHODE
AND A GRAPHITE ANODE

One method for providing the Li-Bi solutions required by the metal
transfer process for rare-earth removal would consist of electrolytically
reducing LiCl produced in the processing system. We have performed an
experiment in order to determine the general operating characteristics of
an electrolytic cell having a bismuth cathode and a graphite anode. This
experiment was carried out in a b-in.-diam quartz cell vessel in which
a 3.5-in.-diam molybdenum cup containing the bismuth cathode was placed.
The anode consisted of a l-in.-diam graphite rod. The cell was operated
at 670°C at a maximum anode current density of 8.6 A/cmg. Apparently,
there was no limiting anode current density under the conditions used in
this experiment, and disengagement of the chlorine gas produced at the anode
proceeded smoothly and without difficulty. During operation, the LiCl be-
came red in color, probably because the cathode surface became saturated
with Li3Bi (which partially dissolved in the LiCl). Measurement of the
cell current efficiency was not possible because of the reaction of the
chlorine gas produced in the cell with iron components in the upper part
of the cell vessel. The experiment confirms our expectation that elec-
trolytic reduction of LiCl using a bismuth cathode and a graphite anode
should proceed readily and that the attack on the graphite anode should

be minimal, if it occurs at all.
xiii

STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS

We have continued stuiles of the purification of salt by counter-
current contact with hydrogen in a 1.25~in.-diam, T-ft-long column packed
with 1/b4-in. nickel Raschig rings. During this report period, a sufficient

quantity of FeF. was added to the salt (66-34 mole % LiF-BeF,) to increase

the iron conceniration from 20 ppm to 425 ppm. Two iron fluiride reduction
runs (R-1 and R-2) were carried out. A salt flow rate of 100 cm3/min was
used in each; the hydrogen flow rates were about 20 liters/min and 13.5
liters/min respectively. Accurate control of the flow rates of salt and
gas to the column proved to be difficult, and erratic results were obtained.
The c;lculated mass transfer coefficients for the two runs were 2.4 and 5.k
x 10

correct the difficulties noted during these runs.

moles/sec~cm3; Several equipment modifications were made in order to

+
> IN MSBR FUEL SALT

ELECTROLYTIC OXIDATION OF Pah+ TO Pa

A protactinium isolation method based on the selective precipitation
of Pa205 from MSBR fuel saifi has recently beenSEroposed. The proce;i re—
quires oxidation of the Pa in the salt to Pa (and most of the U~ +to
Uh+) prior to precipitation of the Pa205. It was suggested that oxidation
of the PaLL+ be carried out electrolytically. Estimates of the current
densities, the size of the proposed electrolytic oxidation system, and the
feasibility of such a reduction step were made. It was concluded that a
low current density would be observed at the anode because of the low con-
centration of PalL+ in the salt, and that a current density limitation was
not likely at the cathode since the concentration of the material to be
reduced is higher, by about two orders of magnitude, than the total con-
centration of materials to be oxidized at the anode. ZElectrolytic oxida-
tion of Pah+ prior to precipitation of Pa205 is not considered attractive
because of the large anode surface area required. It is believed that

5

+
by the use of HF—HEO—H2 mixtures is more

promising and that additional data should be obtained to allow evaluation

+
oxidation of the Pal‘L to Pa

of this step.
1. INTRODUCTION

A molten-salt breeder reactor (MSBR) will be fueled with a molten

fluoride mixture that will circulate through the blanket and core regions

of the reactor and through the primary heat exchangers. We are developing

processing methods for use in a close-coupled facility for removing fission

products, corrosion products, and fissile materials from the molten fluoride

mixture.

Several operations associated with MSBR processing are under study.

The remaining parts of this report discuss:

(1)

(2)

an improved flowsheet for processing MSBR fuel salt by fluori-

nation--reductive extraction and the metal transfer process,

analysis of methods for removing uranium from fuel salt by

precipitation as UOE-ThO2 solid solutions,

installation of a simulated continuous fluorinator for studying

induction heating in molten salt,

measurement of gas holdup and axial dispersion coefficients in
open bubble columns, using a recently devised transient tech-
nique,

installation of experiment MTE-2 for demonstration of the metal
transfer process for removing rare earths from MSBR fuel carrier

salt,

experiments conducted in a mild-steel reductive extraction
facility, to increase our understanding of the rate at which
uranium is extracted from molten salt into bismuth in a

packed column,

the design and testing of devices for preventing axial dis-
persion in packed columns during the countercurrent flow of

molten salt and bismuth,

operation of a static electrolytic cell for reduction of LiCl

using a bismuth cathode and a graphite anode,
(9) studies of the purification of salt by continuous methods,
and
(10) analysis of the feasibility of the electrolytic oxidation
of tetravalent protactinium to pentavalent protactinium
prior to the precipitation of Pa205‘
This work was carried out in the Chemical Technology Division during the

period July through September 1970.
2. ANALYSIS OF THE FLUORINATION--REDUCTIVE EXTRACTION
AND METAL TRANSFER FLOWSHEET

M. J. Bell L. E. McNeese

A flowsheet in which fluorination is used for removing uranium and
reductive extraction is used for isolating protactinium from MSBR fuel
salt has been described.l’2 However, we have found that a considerable
simplification in the flowsheet and a significant reduction in partial
fuel cycle cost can be achieved by holding the isolated protactinium in
a secondary salt phase, rather than in bismuth, as previously considered.
A flowsheet that employs this improved mode of operation has been developed,
and the partial fuel cycle costs corresponding to several sets of operating
conditions have been calculated. We have also observed that the waste
streams from the protactinium isolation and the rare-earth removal portions
of the new flowsheet can be conveniently combined for uranium recovery prior
to disposal, thus decreasing the probability of loss of fissile material

T

and also eliminating the need for adding LiF to the protactinium decay tank
in order to obtain an acceptably low liquidus temperature. A method for

combining the waste streams is described.

2.1 Isolation of Protactinium in a Secondary Salt, Using
Fluorination and Reductive Extraction

Analysis of the fluorination--reductive extraction flowsheet for iso-
lating protactinium from MSBR fuel salt has revealed severai undesirable
features and has suggested an improved method for removing fission product
zirconium and for retaining 233Pa during its decay to 233U. In the flow-
sheet described previously, zirconium was extracted into the bismuth stream
exiting from the lower column of the protactinium isolation system; it was
removed from this stream by hydrofluorinating a small fraction of the bis-
muth in the presence of salt that was withdrawn from the system. ©Since the
bismuth stream also contained protactinium and uranium, the portion of the
stream that was hydrofluorinated represented a compromise between (1) main-

taining an acceptably low zirconium concentration in the bismuth in the
lower part of the column and (2) transferring acceptably small amounts of
protactinium and uranium to the waste salt from which these materials must
be recovered. The remaining bismuth was hydrofluorinated in the presence
of salt that was then recycled to a point ahead of the fluorinator in
order to remove uranium as UF6. This operation also resulted in the re-~
cycle of zirconium, which, under operating conditions of interest, was
oxidized and reduced severél times before its removal. Such recycling
caused the quantity of reductant required for isolating the protactinium

to be increased significantly.

We have observed that the flowsheet can be simplified by hydrofluori-
nating the entire bismuth stream in the presence of a secondary salt stream,
as shown in Fig. 1. Salt is withdrawn from the reactor on a 10-day cycle
and is fed to a fluorinator, where 95 to 99% of the uranium is removed.
The salt from the fluorinator is then sent to an extraction column where
protactinium, zirconium, and the remaining uranium are extracted into a
bismuth stream containing reductant. Finally, the bismuth stream is
hydrofluorinated in the presence of the secondary salt stream, which re-
sults in transfer of the extracted materials to the salt. Reductant is
added to the recovered bismuth, and the metal stream thus produced is re-
cycled to the extraction column as in the previous flowsheet. The sec-
ondary salt stream is circulated successively through a hydrofluorinator,
a fluorinator, and a protactinium decay tank. The fluorinator is used to
maintain an acceptably low uranium concentration in the protactinium decay
tank. Periodically, salt is withdrawn from the decay tank to remove zir-
conium and other fission products that have accumulated. The salt is held

for a sufficient period before final discard to allow 233Pa to decay to

233U, which is recovered from the salt by batch fluorination.

These flowsheet modifications offer the following advantages over

the earlier flowsheet:

1. The bismuth inventory in the system is greatly reduced, thereby

avoiding a significant inventory charge.

2. The protactinium decay tank can be fabricated from a nickel-base
alloy rather than molybdenum, which will result in a considerable

saving in the installed equipment cost.
 

 

 

 

 

 

BeF,
Thiy

SALT DISCARD

 

 

 

 

 

 

 

 

 

 

 
  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL-DWG 70-2811RAR

SALT UFg PROCESSED SALT

PURIFICATION RFDUCTION ~ 1

t | 1

| Hz extracTor)

1 I

| I Bi

Jufs SALT CONTAINING RARE EARTHS 1 |
| | | -
| I I
I ! 1
1 EXTRACTOR] i
RFACTOR : [_'_I ! I
EXTRACTOR _! |
I » I

icl mmmmB-]
: L (0.5 MOLE FRACTION Li) |
|7 1
I U EXTRACTOR 1
6 I
I 4 Bi-Li !
HYDROFLUORINATOR FLUORINATOR Pa DECAYJ b mp+ DIVALENT RARE I
I ‘ * ' r—-—- EARTHS i
emmpgi
F2 | HE Fa Po DECAY I (0.05 MOLE FRACTION !
| F o UFg Li) i
o FLUORINATOR exTracTorl I
I Bi-Li 1
| SALT TO l= = o+ TRIVALENT RARE I
WASTE Ll EARTHS
L———‘ REDUCTANT ————-———————————————»l
ADDITION
Li

~y

Fig. 1.

J

URANIUM REMOVAL

v
PROTACTINIUM REMOVAL

Reductive Extraction and the Metal Transfer Process.

—
RARE EARTH REMOVAL

Flowsheet for Processing a Single-Fluid MSBR by Fluorination--
3. Control of the protactinium isolation system is greatly

simplified.

L. Zirconium will not be recycled in the lower part of the prot-
actinium isolation system, thus reducing the consumption of

reductant.

5. The isolated protactinium is retained in such a manner that
maloperation of the extraction column cannot return large

quantities of protactinium to the reactor.

6. The flowsheet is simplified; recycle of fuel salt containing
uranium, zirconium, and protactinium to the primary fluorinator
is avoided. Operation of the secondary salt circuit is restricted

only by heat removal and uranium inventory considerations.

T. Very efficient hydrofluorination of the bismuth stream would
permit the initial salt inventory in the protactinium decay
tank to contain natural lithium rather than 7Li.

2.2 Mathematical Analysis of Flowsheet in Which Protactinium Is Isolated
in a Secondary Salt, Using Fluorination and Reductive Extraction

A mathematical analysis was carried out for the protactinium isolation
system described in the previous section. Of main interest were (1) the ef-
fects of the primary fluorinator uranium removal efficiency and the reductant
addition rate on the performance of the protactinium extraction column, and
(2) the rate at which LiF must be added to the secondary salt in the prot-
actinium decay system in order to maintain an acceptably low liquidus temp-
erature. Also of interest was the uranium inventory in the protactinium
decay tank for various values of the secondary fluorinator efficiency and

the flow rate of the secondary salt through the fluorinator.

In making the analysis, we assumed that salt of a designated composi-
tion was withdrawn from the reactor at a known flow rate and fed to the
primary fluorinator, where a specified fraction of the uranium was removed
as UF6. The extraction column was assumed to consist of a specified number

of theoretical stages. With this approach, calculations proceed from one
end of the column, where flow rates and concentrations are known or as-
sumed, to the opposite end of the column, where an appropriate check is
made on the calculated values (if the starting concentrations and flow
rates were assumed values). The calculations involve the successive
application of equilibrium and material balance relations for each of
the theoretical stages. Data relative to the equilibrium distribution
of materials of interest have been obtained by Ferris and co-workers.
The distribution coefficient for material A, between salt and bismuth
containing a reductant, is given by the expression:

1
log D, = n + log K, , (1)

A A log D

Li
where

DA = distribution coefficient for A,

= Xy /%gpo

XMA = concentration of element A in metal phase, mole fraction,

XSA = concentration of fluoride of element A in salt phase, mole
fraction,
n, = valence of element A in salt phase,
DLi = distribution coefficient of lithium,
KA = modified equilibrium constant.

1
The variation of the modified equilibrium constant, K,, with temperature

is given by the relation:

1
log K, = A + B/T, (2)
where
A, B = constants,

T = temperature, °K.

In setting up the equilibrium relations for a system containing N + 1
components that distribute between the molten salt and bismuth phases, one
component is conveniently chosen as the reference component. The relative
concentrations of materials in the two phases are then related by the fol-

lowing set of expressions that can be derived from Eq. (1):
X ni/nr n
X . =X A | ' i ..
ui T Vst | X exp [log X, - —~logK 1,1i=1...., (3)
r

where
XSi’XMi = mole fraction of component i in salt and metal, respectively,

XSr’XMr = mole fraction of reference component in salt and metal,
respectively,
N, L0, = valence of component i and reference component, respectively,
in salt.
The final expression required for calculating equilibrium concentrations

in the two phases is given by the following relation:

N+1

E ;X4
=1
N+1

where

XMR = equivalents of transferrable components per mole of bismuth.

This relation requires that the number of equivalents of transferrable

metals per mole of bismuth remain constant.

A material balance around stage J of a column in which the stages

are numbered from the top ylelds the relation:

+ X = X.,. . + . .
FSjXSi,j FMj Mi,J FSj+l Si,J+1 FMJ—lXMl,j—l >

where
FSj = flow rate of salt leaving stage j, moles/day,
FMj = flow rate of metal leaving stage j, moles/day,
XSi,j = mole fraction of component i in salt leaving stage Jj,

XMi 3 = mole fraction of component i in metal leaving stage Jj.
3
The notation used in analyzing the remaining parts of the protactinium

isolation system (see Fig. 2) is as follows:

FDS = flow rate of discarded salt, moles/day,
FH = flow rate of salt entering hydrofluorinator, moles/day,
FM = flow rate of metal entering hydrofluorinator, moles/day,
FLIF = rate of addition of LiF to Pa decay tank, moles/day,
FS1 = flow rate of salt leaving hydrofluorinator, moles/day,
H = uranium removal efficiency in fluorinator,
XM = mole fraction of a component in metal phase,
XS = mole fraction of a component in salt phase,
VDK = volume of protactinium decay tank, moles,

-1
A = radioactive decay constant for 233Pa, day .

In addition, various suffixes were added to the flow rates and concentra-
tions to indicate the location of a stream in the flow diagram, and the
following suffixes were appended to the concentrations to denote the

element being considered:

L = lithium,

P = protactinium,
T = thorium,

U = uranium,

Z = zirconium.

Thus, the symbol XS2P refers to the mole fraction of protactinium in salt
at point 2 in the flowsheet. In the model, the protactinium decay tank
was treated as a completely mixed vessel and the hydrofluorinator was
assumed to remove all transferrable metals from the bismuth. For these

conditions, one can write the material balance relations given below.

For 233Pa:

FM-XMP + FSL-XShP = FS2-XS82P, (6a)
FS2.XS2P = FS3+XS3P, (6b)

(FS3 - FDS)-XS3P = (A-VDK + Fsk)-XSLP. (6c)
ORNL DWG. 71-13585R|

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o ____|REDUCTANT .
v ADDITION SLiF
| K
! t
Y !
E !
X :
T |
R 1
A |
C I
T :
| :
0 | FS4
N : FH XS4
H i
FS UFg : I (FS4-FH) FLIF
XS ! :
l T | I UFg LiF
. ! $
: FSI| _ Pa
FLUORINATION FMlbHYDROFLUOR!NATION XSIFSZ FLUORINATION —F—Sg—-—— DECAY
? XM f* XS2 f XS3
Fy HF Fo SALT TO WASTE

FDS

Fig. 2. Schematic Flowsheet for Isolation of Protactinium Using
Fluorination and Reductive Extraction.

0T
For

For

For

For

233

232

Zr:

Li:

Th:

11

FM+XMU + FSk-XSLhU = FS2.XS2U,
(FS3 - FDS)+XS3U + A-VDK.XS4P = FSh.Xshu,

FS3:XS3U = FS2.X82U0.(1 - H).

FM+XML + FShXSWI = FS2-XS2L,
FS2.XS2L = FS3-XS3L,

(FS3 - FDS)-XS3L + FLIF = FSL.XSLL.

FM+XMT + FSh.XSLhT = FS2-:XS2T,
FS2.XS2T = FS3-XS3T,

(FS3 - FDS).XS3T = FSL.XSLT.

FM-XMZ + FSL-Xskz = FS2-XS2Z,
FS2:XS2Z = FS3-XS3Z,

(FS3 - FDS)+XS3Z = Fsh.Xskz.

From an overall material balance:

FM: (XMP + XMU + XML + XMT + XMZ) + FSh = FS2,

FS2 = FS3 + FS2-XS2U-H,

FS3 + FLIF = FDS + FSk.

(8a)
(8b)
(8c)

(10a)
(10b)

(10c)

(11a)
(11b)

(11c)

The system is also under the constraint that LiF must be added to the

protactinium decay tank at a sufficient rate (FLIF) to obtain a suitable

salt liquidus temperature. An LiF—ThFh mixture containing 71 mole % LiF

has a liquidus temperature of 568°C, which is acceptably low.

assumed that the presence of zirconium, protactinium, and uranium fluorides
12

at low concentrations would not increase the salt liquidus temperature

to an unacceptable level; and, in the calculations, the value of FLIF was
determined so that XSLL was equal to 0.71l. Each of the sets of equations
[i.e., Egs. (6) = (10)] consists of three equations relating three concen-
trations. However, the equations are not linear because of the dependence
of the flow rates on the uranium concentration, as indicated by Eq. (11b).
The quantities FM, XMP, XMU, XMT, XML, XMZ, VDK, H, XSLL, and FSh are known;
the remaining variables are to be determined. In the algorithm adopted for
solving the above system of equations, the equations were linearized by
assuming a value for the salt discard rate, which, with Egs. (8a) - (8c)

and (1la) - (1llc) fixes all of the salt flow rates. The remaining equations
can be solved for the unknown concentrations, and Egs. (1l1b) and (1llc) can
be used to obtain an improved estimate of the salt discard rate. This
algorithm converges in only a few iterations since the equations are not

strongly nonlinear.

2.3 Calculated Results on Operation of the Protactinium
Isolation System

The mathematical model described in the previous section was used to
compute concentrations and flow rates throughout the secondary salt system
and to evaluate the performance of the protactinium isolation system for
an assumed 10-day processing cycle. The variation of the protactinium re-
moval time with the primary fluorinator uranium removal efficiency and with
the rate at which reductant is added to the extraction column is shown in
Fig. 3. As can be observed, essentially complete extraction of the prot-
actinium is obtained if the reductant addition rate is adequate. A re-
ductant addition rate of about 200 equiv/day is required for a uranium
removal efficiency of 99%, and an addition rate of about 300 equiv/day is
required for a uranium removal efficiency of 95%. The rate at which LiF
must be added to the secondary salt in order to maintain a LiF concentra-
tion of 0.71 mole fraction depends on both the uranium removal efficiency
in the primary fluorinator and on the reductant addition rate. As shown

in Fig. 4, a LiF addition rate of about 40 moles/day is required in order
13

ORNL DWG. 71-13586

 

 

 

 

 

 

 

 

 

 

T T T Y T
REDUCTANT FEED RATE
60 - (EQUIVALENTS PER DAY)
o _
S ~ B 8 = g K~
N N & o ® m o™
50 - -

»n

>

<

o

a0l -

w

=

-

-

<

> 30} -

O

=

w

@

=

_D_ 20 -

<

.—

O

<

-

O

@

a 10 -~
PROCESSING CYCLE TIME 10 DAYS
TEMPERATURE 640°C
STAGES IN Pa EXTRACTOR 5

0 i ] 1 i i
100 98 96 94 92 90

PRIMARY FLUORINATOR EFFICIENCY (%)

Fig. 3. Effects of Primary Fluorinator Efficiency and Reductant Feed
Rate to Protactinium Extraction Column on Protactinium Removal Time.
1k

ORNL DWG. 71-13587

 

   
 
 
   

 

 

 

1 1T 1 1
REDUCTANT FEED RATE
(EQUIVALENTS PER DAY)
606 © -
N~
on w oI YR
N Ao O PROCESSING CYCLE
TIME IO DAYS
TEMPERATURE 640°C
50 STAGES IN Pa i
EXTRACTOR 5
>=
<
o
S
» 40 B
w
-
@]
s
LIJ30 =
-
<
a
<
2,
F 2 i
o
o
- ¢
L.
10 .
0 | 1 |
100 98 S 6 94 92 90

PRIMARY FLUORINATOR EFFICIENCY (%)

Fig. 4. Effects of Primary Fluorinator Efficiency and Reductant Feed
Rate to Protactinium Extraction Column on LiF Addition Rate.
15

to achieve a uranium removal efficiency of 99% when the reductant ad-
dition rate is 200 equiv/day. The same LiF addition rate is also re-
quired to achieve a uranium removal efficiency of 95% when the reductant
addition rate is 371 equiv/day. The variation of protactinium removal
time with number of stages in the protactinium extraction column is
shown in Fig. 5. As can be observed, little benefit is obtained from
use of more than five stages, and as few as three stages could be used
at the higher reductant feed rate without increasing the protactinium
removal time appreciably. The variation of the uranium inventory in
the protactinium decay tank with the uranium removal time from the tank
and with the uranium removal efficiency in the primary fluorinator is
shown in Fig. 6. As can be observed, a uranium inventory that is 0.1%
or less of the reactor uranium inventory can be obtained over a wide

range of operating conditions.

Both the quantity of reductant required and the rate at which fuel
carrier salt must be removed from the reactor to compensate for the LiF
added by the protactinium isolation system depend on the fraction of the
uranium that is removed from the fuel salt by the primary fluorinator.

For a uranium removal efficiency of 95% and a reductant addition rate of

371 equiv/day, fuel carrier salt must be withdrawn at the rate of 0.3
ft3/day; for a removal efficiency of 99% and a reductant addition rate

of 200 equiv/day, fuel salt must be removed at the rate of 0.16 ftB/day.
Table 1 gives a comparison of the components of the partial fuel cycle

cost for the present protactinium isolation system in which the protactinium
is isolated in salt and those for the previous protactinium isolation system
in which the protactinium was isolated in bismuth. For the present system,
the chemical and inventory charges amount to 0.047 mill/kWhr for a uranium
removal efficiency of 95% and a reductant addition rate of 371 moles/day.
These charges are reduced to 0.034 mill/kWhr for a uranium removal effi-
ciency of 99% and a reductant addition rate of 200 equiv/day. The salt
inventory in the protactinium decay tank is 150 ft3 in each case. The

uranium inventory in the tank is about 0.1% of the reactor uranium in-

ventory.
16

ORNL DWG 7!-13588RI
|4 T T T I 1 T
TEMPERATURE 640 °C
Pa CYCLE TIME |10 DAYS

FLUORINATOR EFFICIENCY 95 % —=~~-
99 % ——

 

   
 
  

ol
|

N
l

REDUCTANT FEED RATE
(EQUIVALENTS/DAY)
200

T
/

  

 

o
]
|
|
|
i

PROTACTINIUM REMOVAL TIME (DAYS)

 

 

9 1 | L L 1 L
0 l 2 3 4 5 6 7
NUMBER OF STAGES IN PROTACTINIUM EXTRACTOR

 

Fig. 5. Effect of Number of Stages in Protactinium Extraction Column
on Protactinium Removal Time for Primary Fluorinator Efficiencies of 95%

and 99%.
IN DECAY TANK

INVENTORY )

INVENTORY

URANIUM

OF FISSILE

(%

ORNL DWG. 71-13589

 

    

 

 

 

 

2.0 T ‘
TEMPERATURE 640°C ‘
PROCESSING CYCLE TIME 10 DAYS
STAGES IN Po EXTRACTOR 5
1.6 | -
PRIMARY FLUORINATOR EFFICIENCY =95%
REDUCTANT FEED RATE =371 EQUIVALENTS/DA
12 | -
08| >
0.4} -
Q\{PRIMARY FLUORINATOR EFFICIENCY = 99 %
REDUCTANT FEED RATE=200 EQUIVALENTS/DAY
0 1 1 1
0 | 2 3 4

URANIUM REMOVAL TIME (DAYS )

Fig. 6. Effect of Uranium Removal Time in Secondary Salt Circuit on
Uranium Inventory in Protactinium Decay Tank for Primary Fluorinator Ef-
ficiencies of 95% and 99%.

LT
18

Table 1. Partial Fuel Cycle Costs for Present and Past
Protactinium Isolation Systems

 

 

Pa Isolation Pa Isolation
in Salt in Bi
Reductant addition rate, 371 200 k29
moles/day
Fluorinator efficiency, % 95 99 95
Resulting components of cost,
mill/kWhr
Reductant 0.0131 0.0071 0.0151
Fluorine 0.0115 0.0115 0.0115
Salt replacement 0.0206 0.0134 0.0163
Uranium inventory 0.000k 0.000k 0.0030
Loss in breeding ratio 0.0000 0.0006 0.0013
HF and H2 0.0010 0.0007 0.0010
Bi inventory 0.0097

Total 0.0L6T 0.0336 0.0579

 
19

2.4 Combination of Discard Streams from the Fluorination-—-~
Reductive Extraction— Metal Transfer Flowsheet

The flowsheet shown in Fig. 1 requires that about 40 moles of LiF be
added daily to the protactinium decay tank in order to obtain a suitable
liquidus temperature. Lithium fluoride purchased for this addition would
increase the fuel cycle cost by only 0.001L mill/kWhr; however, we have
observed that an acceptable liquidus temperature can also be obtained by
hydrofluorinating the Li-Bi stream from the divalent rare-earth stripper
in the presence of the salt from the decay tank. This operation adds
50 moles of LiF and about 1.1 moles of rare-earth fluorides to the decay
tank per day. With this addition, the composition of the salt in the
decay tank is 72-24-3 mole % LiF-ThFh—Zth, 1 mole % divalent rare-earth
fluorides, and 360 ppm of trivalent rare-earth fluorides. (A primary
fluorinator efficiency of 99% and a reductant feed rate of 200 equivalents
per day ére assumed.) The salt has a liquidus temperature of 570°C; at
this temperature, the rare-earth fluoride concentration is well within

the rare-earth solubility.

We have also observed that it is possible to combine all waste streams
from the metal transfer system and the protactinium isolation system for
uranium recovery prior to disposal, as shown in Fig. 7. 1In this operation,
waste salt from the protactinium decay tank would be combined with the
fuel carrier salt discard stream. The Li-Bi stream from the trivalent
rare-earth stripper would be hydrofluorinated in the presence of the re-
sulting salt, and the combined stream would be held for protactinium decay.
The protactinium concentration in the combined streams would be only 500
ppm initially, and it would decrease thereafter. The specific heat gen-
eration rate would be acceptably low. The salt in the waste holdup tank
would be fluorinated before discard to remove uranium. The composition
of the discarded salt would be Th.7T-13.5-9.5-0.8 mole % LiF-ThF)-BeF -
ZrF) , 1.2 mole % trivalent rare-earth fluorides, and 0.3 mole % divalent
rare-earth fluorides. Although the liquidus temperature of the salt is
near 500°C, the salt temperature would have to be maintained at about

600°C to prevent precipitation of the trivalent rare-earth fluorides.
ORNL DWG 71-2858

Li-Bi +DIVALENT

RARE EARTHS
50 MOLES Li/DAY

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UFg
| }
Bi CONTAINING HYDRO— Pa
Pa, U, Li, Th, Zr — - FLUORINATOR t#{ DECAY
FROM Pa ISOLATION | FLUORINATOR TANK
, , Li-Bi + TRIVALENT
‘ f RARE EARTHS
11O MOLES 50 MOLES Li/DAY
Bi TO HF-H, Fp SALT/DAY}
RECYCLE 0.089 FT¥DAY 4
FUEL CARRIER WASTE
SALT DISCARD ___| SALT HYDRO-
240 MOLES/DAY HOLDUP FLUORINATOR
0.16 FT3/pDAY TANK
[ I l Bi TO
‘ UEG RECYCLE
SALT
{L—& BATCH | o 10 waSTE

 

FLUORINATOR

!

F2

0.27 FT¥DAY

 

 

 

Fig. 7. Method for Combining Waste Streams from Protactinium Isolation
and Rare-Earth Removal Processes. Flow rates are shown for a uranium removal

efficiency in the primary fluorinator of 99% and a reductant addition rate of
200 equivalents/day.

0c
21

This processing scheme would require that salt be discarded at the rate

of 60 £t every 220 days.

3. ANALYSIS OF URANTUM REMOVAL FROM FUEL SALT BY OXIDE PRECIPITATION

M. J. Bell L. E. McNeese

Oxide precipitation is being considered as an alternative method for
selectively removing protactinium from MSBR fuel salt and for subsequently
removing uranium from the fuel salt prior to the removal of rare earths.
We have made calculations that describe the performance of a multistage
countercurrent precipitator for removing uranium from fuel salt which is
free of protactinium. The results indicate that 99% of the uranium can
be removed from the salt by using about three equilibrium stages and that
the UO.-ThO,. solid solution which is precipitated will contain greater

2 2
than 90% UOQ.

3.1 Mathematical Analysis of a Uranium Oxide Precipitatoer

A mathematical analysis of a multistage countercurrent equilibrium
precipitator was carried out to determine the feasibility of precipitating
most of the uranium from MSBR fuel salt without the attendant precipitation
of large quantities of Th02. The data of Bamberger and BaesLL for the
equilibrium concentration of UO2—ThO2 solid solutions in contact with
molten LiF—Bng—ThFh—UFh salts were employed in the analysis. The equi-

librium quotient, Q, defined as

Q= XUo2 XThFh/(XThO Xy, s (12)

for the reaction

= +
UFLL + ThO2 Uo2 ThFh (13)
22

is given by the expression

log Q = (2101 + 550 XUO )/T, (1)
2

where T is the absolute temperature (°K).

In the analysis, each stage was treated as an equilibrium contactor
and 1t was assumed that the solid leaving a stage would have a specified
quantity of salt associated with it. The salt/solid mole ratio in this
stream was assumed to be constant for each of the stages. Material
balances around stage N (see Fig. 8) yield the following relations. For

uranium and thorium in salt:

(FS + FR) [XSU(N) + XST(N)]
= FR [XSU(N + 1) + XST(N + 1)] + FS [XSU(N - 1) + XST(N - 1)]; (15)

and for uranium:

(FS + FR)-XSU(N) + FO-X0U(N)
= FR-XSU(N + 1) + FS*XSU(N - 1) + FO-XOU(N + 1), (16)

where
FO = flow rate of oxide, moles/day,
FR = flow rate of salt accompanying oxide, moles/day,
FS = flow rate of salt flowing countercurrent to oxide, moles/day,
X0U = mole fraction UO2 in oxide,
XST = mole fraction ThFh in salt,
XSU = mole fraction UFM in salt.

Equation (15) states that the rate at which UFh and ThFh leave stage N must
equal the rate at which these materials enter in salt from stage (N - 1)
and in salt associated with solids from stage (N + 1), since it is assumed
that the flow rate of oxide from stage to stage is fixed. Equation (16)
states that the amount of uranium leaving stage N in the salt and the oxide

must be equal to the amount of uranium entering in the salt from stage
23

ORNL DWG. 71-13590RiI

FO; XOU(N+1)

 

 

 

 

FS3XSU(N),XST(N) FR; XSU(N+1), XST(N +1)
STAGE
N
FS; XSU(N-1), XST(N=-I) FO, XOU(N)

FR; XSU(N), XST(N)

Fig. 8. Nomenclature Used in Mathematical Analysis of a Multistage
Countercurrent Uranium Oxide Precipitator.
ol

(N - 1) plus that entering in the solids and associated salt from stage
(N + 1). In solving the set of equations represented by Egs. (15) and
(16), one takes account of two facts: the salt flow rate and the com-
position of the salt are known at the point where salt is fed to the
precipitator, and no salt accompanies tfie oxide fed to the opposite end
of the precipitator. Other data required for evaluating precipitator
performance include the number of stages in the precipitator, the temp-
erature, the moles of salt per mole of oxide in the oxide stream leaving
a stage, and the composition of the UO2--ThO2 solid solution produced in a

given stage, which is given by Eq. (12).

3.2 Calculated Results

Typical results showing the effects of temperature and number of
stages on precipitator performance are shown in Fig. 9. These results
indicate that greater than 99% of the uranium can be removed from MSBER
fuel salt with three or more stages and that the oxide stream produced
will have g UO2 concentration of greater than 90%. Over a wide range
of conditions, less than 1% of the thorium fed to the system would be
precipitated with the uranium. A decrease in performance is observed
as the temperature increases. No significant effect on precipitator
performance was observed when the amount of salt remaining with the oxide

during the transfer of salt between stages was varied from 2 to 10 moles

per mole of oxide.

L. DEVELOPMENT OF A FROZEN-WALL FLUORINATOR: INSTALLATION OF A
SIMULATED FLUORINATOR FOR STUDYING INDUCTION HEATING

J. R. Hightower, Jr. C. P. Tung
L. E. McNeese

An experiment to demonstrate protection against corrosion by the use
of layers of frozen salt in a continuous fluorinator requires a corrosion-
resistant heat source to be present in the molten salt. High-frequency

induction heating has been proposed as the source of heat, and the estimated
25

ORNL DWG 70-8995

100.0

 

 

 

99.0

98.0

 

 

 

 

aQ 970
(0 o8

)

>

S

O 96.0

o

2

2

2 95.0

J

(0

oD

4

® 94,0

/ 580° C
93.0 |- 620°C ———-—
MOLES SALT RECYCLED/ MOLE OXIDE = 2
92.0 l |
1.0 11 1.2 1.3

MOLES OF OXIDE REMOVED PER MOLE OF UF4 FED

Fig. 9. Effects of Temperature, Number of Stages, and Oxide Removal
Rate on Percentage Uranium Recovered in an Oxide Precipitation.
26

5

performance” of a frozen-wall fluorinator having an induction coil em-
bedded in the frozen salt near the fluorinator wall has indicated that
this may be an acceptable heating method. There are uncertainties as-
sociated with determining the effect of bubbles in the molten salt and
in estimating the amount of heat that will be generated in the metal
walls of the fluorinator. Equipment is béing assembled for studying
heat generation in a simulated frozen-wall fluorinator containing an

embedded induction coil. In this experiment, a 31 wt % HNO_ solution,

3
which has electrical properties similar to those of molten salts, will
be used to simulate molten salt in the fluorinator vessel. The equip-

ment for the experiment is described in the remainder of this section.

4.1 Experimental Equipment

The induction-heated simulated fluorinator consists of a 5-in.-0D,
S5-ft-long glass tube inserted in an induction coil and placed inside a
5-ft-long section of 8-in. 30L4 stainless steel sched 40 pipe. The nitric
acid inside the glass tube represents the molten salt in the fluorinator,
the space between the glass tube and the pipe wall (which contains the
induction coil) represents the frozen salt layer, and the pipe represents
the fluorinator vessel wall. As shown by the flow diagram in Fig. 10,
the acid is pumped, first, through the glass column (where it is heated
by the induction coil) and, then, through a heat exchanger (where the
heat is removed). The rate at which heat is generated in the acid will
be determined from a heat balance on the acid passing through the column.
The pipe representing the fluorinator wall is equipped with a jacket
through which cooling water passes. The heat generated in the pipe will
be calculated from the change intemperature of the water flowing through
the jacket. A drain tank is provided for holding the nitric acid during
system maintenance. Air, at rates up to 2 cfm, can be countercurrently

contacted with the downflowing acid stream in the column.

The glass column and induction coil are shown in Fig. 11. The coil
is made up of 17 sections of smaller coils, each of which is 5.6 in. in

inside diameter and 3 in. long. The coil sections are made of 6.5 turns
27

ORNL DWG 70-896¢

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AIR
GLASS ;
COLUMN fi
1
i
e COOLING
g °(“’fi4b { WATER
1 g ! E
N 4 y TE
] b
H Q L]
N b N
1 9 N
/ M b N
’ 1 qQ N
1 | P}
1 H 9 \
1 H q ° N (::>‘—“
4 b N JACKETED
] R 9 b N PIPE
79 R
1 0 q \ HEAT
7 p R EXCHANGER
A N
9 b
/B I~ \
] R b
1 N q \
1 N p R
1 H 9 N
10 P N
1 0 d \
o p 1 COOLING
v 1 SN WATER
1 R 9 \
! \ INDUCTION
1 COoIL
COOLING
WATER ? NITRIC ACID
PUMP
v
§
{ DRAIN

 

 

 

Fig. 10. Flow Diagram of Fluorinator Simulation and Nitric Acid Recir-
culation System.
PHOTO 100288

 

Fig. 11. Photograph of Induction Coil with the Glass Column Thermally
Insulated and Inserted.

8c
29

of 0.25-in.-0D Monel tubing, and adjacent coils are wound in opposite
directions. The total length of the coil assembly is 5 ft. The coils
are connected electrically in parallel between headers made from 5/8-
in.-diam Monel tubing; all coils wound in the same direction are con-
nected to one pair of headers, and the two pairs of headers are, in

turn, connected electrically in parallel. The inductance of the work
coil (when installed inside the 8-in.-diam pipe ) was found to be 5.8
uH. The effective resistance at a frequency of 1000 Hz was found to

be 0.135 Q.

Figure 12 shows the acid recirculation system before thermal insula-
tion was added to the jacketed pipe that simulates the fluorinator vessel
(shown at the right of the photograph). The acid recirculation system is
constructed of stainless steel tubing; the parts of the pump that contact
the acid are made of stainless steel and Teflon; and the gaskets in the
glass-to-metal Jjoints at each end of the glass column are fabricated from
Viton-A. The recirculation system is located behind a splash shield (not
shown in Fig. 12) and can be operated entirely from in front of the splash

shield.

The heat exchanger is an American Standard stainless steel shell
and tube exchanger which has one shell side and one tube side pass. The
heat-transfer area is 9 ftg. The pump is a Crane Company Chempump made
of 316 stainless steel with Teflon gaskets and is designed to operate at
a temperature of 150°F. At a flow rate of 1 gpm (approximately the max-
imum acid flow), the pump develops a total head of T4 ft of fluid.

The rf generator is a 25-kWTher-monic model 1400 oscillator operating
at a frequency of LOO kHz. The rf transmission lines are flexible water-

cooled coaxial cables made by the L. C. Miller Company.

L.2 Status of Equipment Installation

The nitric acid recirculation system and the rf generator have been
installed. We are presently matching the impedances of the generator and

work coil in order to deliver sufficient power to the work coil.
30

- PHOTO 100521

 

    

 

 

"k\%
% R
. :
. %
3
#
- !

 

 

 

      

 

 

 

 

 

 

 

 

 

§
X
X
i
g :
i

    
 

Fig. 12. 1Installed Fluorinator Simulation and Nitric Acid Recircula-
tion System.
31

5. MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS AND
GAS HOLDUP IN OPEN BUBBLE COLUMNS

M. S. Bautistsa J. S. Watson
L. E. McNeese

Axial dispersion is an important consideration in the design and
performance of continuous fluorinators. Since molten salt saturated with
fluorine is corrosive, fluorinators will be simple open vessels having a
protective layer of frozen salt on all exposed metal surfaces. In such
systems, the rising gas bubbles may cause appreciable axial dispersion in
the salt. For the past two years, we have been involved in a program for
measuring axial dispersion during the countercurrent flow of air and water
in open bubble columns. The objectives of this program are to evaluate
the effect of axial dispersion on fluorinator performance and to account

for this effect in the design of fluorinators.

5.1 Previous Studies on Axial Dispersion

Initial investigations on axial dispersion in open bubble columns were
made by Bautista and McNeese,6 who studied the countercurrent flow of air
and water in a 2-in.-ID, T72-in.-long column. Two regions of operation were
observed. The first of these consisted of a "bubbly'" region (at low gas
flow rates) in which the air moved up the column as individual bubbles and
coalescence was minimal. The second region consisted of a "slugging'" region
(at higher gas flow rates) in which the air coalesced rapidly into bubbles
having diameters equal to the column diameter. A plot of the logarithm of
the dispersion coefficient vs the logarithm of the gas flow rate was linear
in both regions. However, the slope of the line representing data in the
slugging region was higher than that for data in the bubbly region. The

transition between the two regions was well defined.

The same column and associated equipment were used by A. M. Sheikh and

T

J. D. Dearth of the MIT Practice School for investigating the effects of

the viscosity and surface tension of the liquid. The dispersion coefficient
32

was found to decrease in the bubbly region as the viscosity of the liquid
was increased from 1 to 15 cP by the addition of glycerol to the water;
little effect was noted in the slugging region. An increase in the dis-
persion coefficient was observed as the surface tension of the liquid was

decreased by addition of n-butanol to the water.

The equipment was also used by A. A. Jeje and C. R. Bozzuto,8 of the
MIT Practice School, who investigated the effects of gas inlet diameter
and column diameter on axial dispersion and obtained data on gas holdup
in bubble columns. In the slugging region, the dispersion coefficient
appeared to be proportional to the square root of the volumetric gas flow
and was independent of column diameter. In the bubbly region, the disper-
sion coefficient was dependent only on the volumetric gas flow rate for
columns having diameters of 2 in. or larger. Dispersion coefficient data
obtained with a 1.5-in.-diam column deviated from this condition. At low
gas flow rates, gas holdup was linearly dependent on the superficial gas
velocity and was independent of column diameter. At superficial velocities
above the transition from bubbly to slug flow, the gas holdup data for the
various column diameters diverged; the holdup was greatest for the smallest

column diameter.

All of the dispersion coefficient data obtained thus far result from
measurements of the steady-state axial distribution of a cupric nitrate
tracer that is continuously injected into the bottom of the column near
the water exit. The tracer concentration was measured at 20 positions,
located 3.5 in. apart, along the column. Water was withdrawn at a low
rate (about 1 cm3/min) by a small magnetically driven centrifugal pump
at each of the positions, and the water stream was circulated through a
photocell and then returned to the opposite side of the column at the
same elevation. This experimental technique operated satisfactorily for
a range of water and air flow rates. It was found that the axial disper-
sion coefficient was independent of both axial position in the column and
water superficial velocity in the range of interest. This experimental
technique had the following two principal disadvantages: (1) the measure-

ments were time consuming since about 2 hr was required for the column to
33

reach steady state, and (2) at sufficiently high gas rates, air that was
entrained with the water circulating through the photocells accumulated

in the centrifugal pumps and prevented satisfactory operation. For these
reasons, an attempt was made to develop an alternative experimental tech-
nique that would circumvent these problems. The remainder of this section
describes a transient technique that has a number of advantages over the

steady-state technique.

5.2 Mathematical Analysis of Unsteady-State
Axial Dispersion in a Bubble Column

A transient technique, which has been used previously by Ohki and
Inoue,9 was examined to determine its applicability under conditions of
interest. In this technique, there is no net flow of water through the
column; however, data obtained by use of the steady-state technique in-
dicated that the water flow rate does not affect the axial dispersion
coefficient in the range of flow rates of interest. A small amount of
electrolyte tracer is rapidly injected into the top of the column, and
the concentration of the tracer is measured continuously at a point near
the bottom of the column by use of a conductivity probe. It was assumed
that the rate of movement of tracer down the column would be described

by the one-dimensional diffusion equation:

2
%§-= De'a%%’ (17)
0Z
where

C = tracer concentration in column at point Z and time t,

t = time,

Z = axial position in column measured from the top of the column,

De = gxial dispersion coefficient.

Since the tracer cannot diffuse across the upper and lower boundaries of
the column, one requires that the following boundary condition be met at

Z =0 and Z2 = L:
3L

gg-= 0. ' (18)
The instantaneous injection at the top of the column of a specified quan-
tity of solution containing a specified tracer concentration is represented

by the following relations at t = O:

© (19)

where
C = concentration of tracer in solution in a thin layer at top of
column at t = O,
A = depth of solution having concentration CO at t =0,

L = column length.

Solution of Eq. (17) with the boundary conditions specified by Egs. (18)
and (19) yields the following relation for the relative tracer concentra-

tion at position Z and time t:

— D n°ret\]
£ = 1+ L ;-sin I-l-TI?-\--cos anz, ex -
C_ A n L L P 12 ’

n=1

where
C_ = final tracer concentration in column after tracer is uniformly

dispersed.

For the condition where A < < L, this solution reduces to the following
relation for the relative tracer concentration in the column at position

Z and time t:
35

The relative tracer concentration is thus a function of the normalized
position in the column, Z/L, and of the dimensionless time, B, which is
equal to the quantity (fl/L)ZDet. In the present case, the relative tracer
concentration is dependent only on the dimensionless time B and was meas-
ured at only one point in the column. In principle, a value for the axial
dispersion coefficient could be obtained by fitting Eg. (20) to the ex-
perimentally determined relative-tracer-concentration-vs-time curve at
only one point (C/C_ = 0.5, for example). Ohki and Inoue point out,
however, that greater accuracy is usually achieved if this curve is fit

to Eq. (20) at two points (other than t = 0) since the result is less
sensitive to errors in the injection time and in the duration of the in-
jection period. In the present case, the curves were fit to Eq. (20) at
relative tracer concentration values of 0.3 and 0.7. Here, the axial

dispersion coefficient is given by the relation:

B0.7 - B0.3 L) 2
De = t -t F' 2 (21)
0.7 0.3

where
80‘3,50‘7 = dimensionless time at which C/Cco has values of 0.3 and
0.7, respectively,
t0,3’t0.7 = time at which C/COo has values of 0.3 and 0.7, respectively.

The error in the measured axial dispersion coefficient resulting from
an error in the relative position of the conductivity probe in the column
can be estimated from Eq. (20) by calculating the variation of the quantity

Bo.1 ™ Fou3

change of the quantity BO 7 BO 3 with a change in relative probe position

at a Z/L value of 0.7 is 1.8, which indicates that an error in probe posi-

with relative probe position, as shown in Fig. 13. The rate of

tion of 2 em would produce an error in the measured axial dispersion co-
efficient of 2.3%. The relative probe position was measured to within 1

cm in the present work.
ORNL DWG 71-13976RI
1

 

.02 I I I T T T T T

98—
94

90—

3
|

SLOPE =18

_ Bor—Bo.3
~ ~ : '
® O 3 a B
| [

o
N

.58

  
 
 
   

 

 

. 1 | | 1 | | | | |

60 62 64 66 68 70 T2 74 76 78
RELATIVE PROBE POSITION (Z/L)

Fig. 13. Variation of BO 7" BO 3 with Relative Probe Position (Z/L).

 

.80

9¢
37

In the previous analysis, it was assumed that the tracer solution is
injected into the top of the column instantaneously. ©Since this is not
actually the case, a new analysis was made in order to estimate the error
in the measured axial dispersion coefficient caused by a nonzero injection
time. In the latter analysis, it was assumed that tracer was added at the
top of the column at a constant rate for a specified time period, and that
this resulted in a nonuniform tracer concentration profile in the column
at the end of the injection period. After this time, the tracer was as-
sumed to diffuse throughout the column with the boundary conditions rep-
resented by Eq. (18). The concentration of tracer in the column during
the period in which tracer is added to the top of the column at a constant

rate is represented by Eq. (17) and the following initial and boundary

conditions:
C=0 , forall Zatt =0 ,
De %%-= Fo , atzZ=0 |,
%%-= 0O , atZ=1 ,
where

FO = flux of tracer at top of column.

The concentration of tracer at the end of the tracer injection period is

given by the following relation:

! o0
FLIPet 3 -2)° - 15 2 1 nn?
C = + - — == cO0S = exp |-
D 2 2 2 2 L
e L 6L T . ©

t
D n2fl2t
e

L2

 

(22)
where

1
t = length of injection period.

The concentration of tracer in the column during the period following tracer
injection is given by solution of Eq. (17) with the boundary conditions rep-

resented by Eq. (18) and the initial condition represented by Eq. (22).
38

The relative tracer concentration in the column at time t and position Z

is given by the following relation:

 

 

o 2 2
‘ Dnrnt 2,
£ oo 1+2 Z cos Eflz-exp - = L - 1 -e® B . (23)
C L 2 2
0 _ L nB
n.—
where ,
Fot
Coo = T = tracer concentration in column after a uniform concentra-
tion is reached,
Deflzt'
1
B 2 .
L

For the column length used in this work and the range of dispersion coef-
ficients observed, an injection time of 2 sec (t') would result in an error
in the measured axial dispersion coefficient of less than 5%. The actual

injection time was always less than 2 sec and usually less than 1 sec.

5.3 Equipment

The equipment used in the study (shown schematically in Fig. 14) con-
sisted of an open bubble column, a means for injecting KCl tracer solution
at the top of the column, a conductivity probe located at an intermediate
axial point along the column for determining the KC1 concentration in the
aqueous solution at that point, an electronics system and a recorder for
recording the output from the conductivity probe, an air supply and meter-
ing system that allowed air to be fed at a known flow rate to a gas dis-
perser located in the base of the column, and a manometer for obtaining
data on gas holdup in the column. These parts of the system are discussed

in detail in the remainder of this section.

5.3.1 Column

The open bubble columns used in this study consisted of Lucite tubes,

8 ft long and 1.5, 2, and 3 in. in inside diameter, that were mounted in a
39

ORNL DWG 71—13966

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

KClI
SOLUTION
INJECT ION — T 1— AR g——‘
WET TEST
METER
=
- A
] MANOMETER
BUBBLE—*
COLUMN °
O
O
O
o b—/
O
o| CONDUCTIVITY
] PROBE
o :~/—— CONgg,CDTG'g'TY AMPLIFIER
O
1
© RECORDER
GAS 0
DISPERSER
ROTAMETER
AIR SUPPLY

Fig. 1bk. Schematic Flow Diagram of Equipment Used for Study of Axial
Dispersion in an Open Bubble Column by a Transient Technique.
Lo

vertical position. Provision was made for installing a conductivity probe
at a point T0.5 cm from the bottom of the column. Provision was also made
for attaching a manometer for obtaining gas holdup data at any of 20 points
located 3.5 in. apart along the axis of a column. A gas disperser was

located in the bottom of the column.

5.3.2 Gas Disperser

The gas disperser consisted of a 0.04-in.-ID opening in the bottom of
the column. Provision was made for use of gas dispersers having a diameter
as large as 1/4 in. However, in the present study, only one disperser size

was used.

5.3.3 Air Supply and Metering System

Compressed air (90 psig) from the building supply was passed through
a demister trap and was reduced in pressure to 15 psig. The air was then
metered through either of two rotameters having maximum flow capacities
of 0.216 liter/min and 0.840 liter/min and éubsequently passed into the
gas disperser in the bubble column. The rate at which air flowed through
the column was monitored with a wet-test meter as the air exited from the

column.

5.3.4 KCl Tracer Solution Injection System

A 5-ml syringe was connected to a sparger through which a known quan-
tity of concentrated KCl tracer solution could be injected. The total time
required for injecting the tracer was less than 2 sec. The tracer solution
was distributed throughout the column by use of four sections of 1/16-in.-

diam tubing.

5.3.5 Conductivity Probe

The conductivity probes were prepared by use of 1/4-in.-diam bolts,

as shown in Fig. 15. A 0.094-in.-diam hole was drilled along the axis of
L1

ORNL DWG 71—-13967RI

 

Pt

 

 

 

 

 

WIRES SILVER
SOLDERED TO

LUCITE COLUMN
Pt ELECTRODES e od

  

 

 

 

Fig. 15. Method for Installing Conductivity Probe in Column.
L2

the bolt, and two insulated wires were silver-soldered to separate square
platinum electrodes that measured about 0.5 cm on a side. The electrodes
were positioned 1 to 2 mm apart and were attached to the head of the bolt
by use of epoxy resin. The hole in the bolt through which the wires ex-
tended was also sealed with epoxy, and all surfaces other than those of

the platinum electrodes were coated with epoxy. The separation distance
between the electrodes was adjusted in order that water would be held be-
tween the electrodes by surface tension. The surfaces of the electrodes
were platinized to reduce polarization by inserting the probe in a platinum
chloride solution and imposing a dc potential of about 3 V across the probe
for a period of about 1 min. The direction of current flow was then re-
versed for 1 min. The output from the cohductivity probe and electronics
system was shown to be linearly dependent on the concentration of KCl in an
aqueous solution in which the probe was placed (see Fig. 16). When the
probe is used in a column, relatively large fluctuations in output signal
are observed during the first part of an experiment. Initially, these
fluctuations were believed to be the result of air passing between the
electrodes. later, it was observed that they become minimal after the

KC1l tracer has been dispersed uniformly throughout the column. Apparently,
then, the fluctuations are produced by velocity gradients in the column in

the presence of a significant KC1l concentration gradient.

A probe was installed (see Fig. 15) so that the head of the bolt was
located inside the column; the threaded section of the bolt extended through
a rubber gasket and a hole in the column. A nut was tightened on the bolt
from outside the column to compress the gasket and thereby secure the probe

and seal the hole through which the bolt was passed.

5.3.6 Electronics System and Recorder

The conductivity probe was connected to a Leeds and Northrup Model
4988 conductivity monitor, which imposed a 60-Hz signal across the elec-
trodes of the probe and produced an output voltage proportional to the
current passing through the probe. It was not possible to vary the fre-

quency of the signal impressed across the probe; however, there was no
L3

ORNL DWG 71-13969
! l l I I

6l Imm GAP SPACE

 

) H

PROBE OUTPUT (millivolts)
N

 

 

 

o | | | | |
0 I 2 3 4 5

KCl CONCENTRATION (meq/liter)

 

Fig. 16. Variation of Output from Conductivity Probe and Electronics
System with Concentration of KCl in an Aqueous Solution.
Ly

evidence of polarization. The signal from the conductivity monitor passed
through a high-impedance amplifier, having a gain of unity, in order

not to overload the conductivity monitor. The output signal from the am-
plifier was damped slightly to reduce the effect of high-frequency fluc-
tuations. Several damping periods in the range 0.1 to 0.5 sec were tested;
no effect on the quantity tO.T - t0.3 could be detected. A damping period
of 0.5 sec was used in succeeding runs. The output from the electronics
system was recorded on a Hewlett-Packard Model T100B recorder having a

chart speed of 0.1 in./sec, or 2 in./min, and a chart width of 12 in.

5.3.7T Manometer for Gas Holdup Measurements

The relative pressure and the relative height of the air-water mixture
were measured at several points along the axis of the column in order to
determine the average gas holdup in the portion of the column above the
measuring point. A manometer made from 0.25-in.-diam glass tubing was used
for the pressure determinations. The lower end of the manometer was con-
nected to a point on the column, and the upper end waé‘connected to a point
on the column off-gas line near the top of the column. The variation of
gas holdup with axial position along the column could be determined by ob-

taining readings at several positions along the column.

5.4 Experimental Procedure

Prior to a run, the column was filled with distilled water to a level
that would produce an air-water mixture height of approximately 230 cm at
the gas flow rate to be used in the subsequent run. The gas holdup in each
column had previously been determined as a function of gas flow rate. The
air flow rate was then set at the approximate value by use of a rotameter,
and air exiting from the column was routed through a wet-test meter to ac-
curately measure the air flow rate. The air was allowed to flow through
the meter for a sufficient length of time to obtain an accurate measurement;
at low gas flow rates, this period was considerably longer than the time

required for a dispersion coefficient measurement.

The tracer injection syringe was then filled with the desired volume
of tracer solution, and the conductivity monitor and recorder were turned

on. The tracer was injected as quickly as possible, and the injection
45

time was marked on the recorder chart. As the tracer was dispersed
throughout the column, the recorder indicated the increase in conductivity
of the solution at the probe position. The run was continued until no
further increase in recorder output was observed, and the times required
for the recorder reading to increase to 30% and 70% of its total deflection
during the run were determined and recorded. The sensitivity of the re-
corder and the volume of tracer injected were selected so that the total

recorder deflection observed during a run was 70 to 90% of the chart width.

During the run, the height of the air-water mixture in the column was
determined to within 1 cm. A more accurate determination would have been
difficult because of variation in the position of the air-water mixture at

the top of the column.

5.5 Results

Twenty-nine runs were made in order to measure axial dispersion coef-
ficient values in open bubble columns having diameters of 1.5, 2, and 3 in.
These data are summarized in Tables 2-4. Fifty nine runs were made to
determine gas holdup in the same columns. In each run, values for the pres-
sure at a particular location and thé height of the air-water mixture above
this location were determined at three to seven points along the column axis.
The resulting data were fit by the method of least squares to a straight line
on a plot of pressure vs distance down the column. The resulting data showed
that gas holdup does not depend on axial position along the column, since
the standard deviation values were less than 2.5% of the holdup values in
most cases. For this reason, experimentally determined values for the pres-
sure and height of the air-water mixture at the various points along the
column axis are not given. Values for the gas holdup (and the associated

standard deviation) for each of the runs are summarized in Tables 5-7.

5.6 Discussion of Data on Axial Dispersion

The variation of the axial dispersion coefficient with gas superficial

velocity is shown in Fig. 17 for columns having diameters of 1.5 2, and 3 in.
Table 2. Summary of Data on Axial Dispersion in a 1.5-in.-ID Open Bubble Column

. . 3
Tracer injection volume: 72 cnm

 

 

Superficial
Gas Flow Gas Relative & Dispersion

Run Rate? Velocity Probe 0.3 0.7 Coefficient
No. (cm3/sec) (cm/sec) PositionP (sec) (sec) (cme/sec)
1 7.19 0.631 0.697 181.8 411.0 18.3

2 89.8 7.88 0.700 51.0 108.6 Th.2

3 35.4 3.11 0.701 87.6 196.2 39.4

Y 288.9 25.3 0.700 18.5 39.5 194

5 236 20.7 0.691 23.5 49.5 154

6 152.1 13.3 0.696 31.8 73.2 100.3

T 128.1 11.2 0.683 34,2 75.6 97 .6

 

®Measured under conditions at top of column.

bRatio of distance of probe from surface of gas-water
water mixture.

mixture to total height of gas-

91
L7

Table 3. Summary of Data on Axial Dispersion in a 2-in.-ID Open Bubble Column

3

Tracer injection volume: ~ 2 cm

 

 

Superficial
Gas Flow Gas Relative L Dispersion

Run Rate? Veloeity Probe 0.3 0.7 Coefficient
No. (cm3/sec) (cm/sec) PositionP (sec) (sec) (cm?/sec)

1 206 10.2 0.700 30.6 69.0 113.2

2 30k 15.0 0.687 27.6 54.0 148

3 375 18.5 0.683 17.4 37.5 139

4 LT3 23.3 0.680 16.7 38.0 176

5 L25 21.0 0.679 15.7 34.8 145

6 495 2k . kL 0.680 13.8 27.1 282

T 515 25.4 0.680 16.0 35.7 192

8 T6h 37.7 0.705 11.7 23.0 265

9 60k 29.8 0.67hL 12.3 26.3 258
10 T.79 0.384 0.682 102 326 21.L
11 57.3 2.83 0.690 T1.1 158.1 L6.0
12 32.3 1.59 0.703 100.8 219.6 33.1
13 T.71 0.380 0.79L 103.5 234.9 31.0

 

*Measured under conditions at top of column.

bRatio of distance of probe from surface of gas-water mixture to total height
of gas-water mixture.
L8

Table 4. Summary of Data on Axial Dispersion in a 3-in.-ID Open Bubble Column
3

Tracer injection volume: ~ 2 cm

 

 

Superficial
Gas Flow Gas Relative ¢ : Dispersion

Run Rated Velocity Probe 0.3 0.7 Coefficient
No. (em3/sec) (cm/sec) PositionP (sec) (sec) (em2/sec)
1 L88.0 10.7 0.701 27.0 61.8 123.6

2 832.8 18.3 0.696 19.6 45.0 164.0

3 882.1 19.3 0.692 18.0 37.6 207.1

L 168.6 3.70 0.700 45.0 99.0 Th.O

5 334.0 7.32 0.709 39.0 87.0 oL.2

6 31.9 0.700 0.684 67.2 153 k3.0

T 218.0 L.78 0.700 Lh.1 99.0 78.9

8 528.3 11.6 0.692 25.9 56.5 131.8

9 505.6 11.1 0.695 25.5 60.0 119.4

 

aMeasured under conditions at top of column.

b Ratio of distance of probe from surface of gas-water mixture to total height of
gas-water mixture.
L9

Table 5. Summary of Data on Gas Holdup in a 1.5-in.-ID Open Bubble Column

 

 

Superficial Relative

Run Gas Velocity® Number of Gas Standard Deviation
No. (cm/sec) ~ Points Holdup Deviation (%)
1 5.26 5 0.140 0.00387 2.76
2 9.7L 5 0.228 0.00537 2.35
3 15.2 5 0.320 0.00838 2.62
L 21.6 5 0.397 0.00852 - 2.15
5 26.4 5 0.460 0.00892 1.9k
6 31.1 5 0.469 0.01017 2.17
7 3.03 5 0.0905 0.00291 3.22
8 L.28 5 0.119 0.002k4T 2.08
9 T.65 5 0.191 0.00406 2.13
10 12.5 5 0.278 0.00743 2.67
11 24.8 5 0.433 0.01269 2.93
12 18.4 5 0.371 0.011L46 3.09
13 32.6 5 0.485 0.01500 3.09
1k 27.2 5 0.k4b4s 0.01265 2.84
15 28,4 5 0.L467 0.01326 2.84

 

aMeasured under conditions at top of column.
50

Table 6. Summary of Data on Gas Holdup in a 2-in.-ID Open Bubble Column

 

 

Superficial Relative
Run Gas Velocity@ Number of Gas Standard Deviation
No. (cm/sec) Points Holdup Deviation (%)
16 0.260 5 0.009k 0.000Lk46 0.47
17 1.90 L 0.0638 0.00268 L.20
18 3.88 L 0.109 0.00124 1.1k
19 7.80 L 0.191 0.00336 1.76
20 13.4 L 0.277 0.00592 2.1h
21 16.3 L 0.320 0.00857 2.68
22 19.5 L 0.356 0.00873 2.45
23 26.9 L 0.412 0.00996 2.2
2l 1k4.3 L 0.294 0.00717 2.4k
25 9.13 L 0.197 0.00L450 2.28
26 9.32 3 0.197 0.00239 1.21
27 20.7 L 0.341 0.00851 2.50
28 6.02 5 0.151 0.00261 1.73
29 6.37 5 0.152 0.00222 1.46
30 5.92 5 0.1k45 0.001kk 0.99

 

Measured under conditions at top of column.
ol

Table 7. Summary of Data on Gas Holdup in a 3-in.-ID Open Bubble Column

 

 

Superficial Relative
Run Gas Velocity Number of Gas Standard Deviation
No. (cm/sec ) Points Holdup  Deviation (%)
31 12.6 T 0.2k9 0.00312 1.25
32 11.9 5 0.230 0.00k455 1.98
33 11.0 5 0.217 0.00482 2,00
3L 10.2 5. 0.20k 0.00360 1.76
35 9.32 5 0.189 0.00345 1.83
36 8.4o 5 0.177 0.00290 1.64
37 7.61 6 0.165 0.00288 1.75
38 5.94 6 0.137 0.00259 1.89
39 9.03 5 0.154 0.00259 1.68
Lo 1.75 > 0.0481 0.00101 2.10
L1 13.1 5 0.234 0.00299 1.28
L2 T.57 5 0.146 0.000562 0.38
43 8.93 5 0.182 0.00108 0.59
L 11.0 5 0.209 0.00207 0.99
L5 5.71 5 0.13k% 0.00203 1.51
L6 5.08 5 0.134 0.00133 0.99
L7 3.55 5 0.102 0.00125 1.23
L8 2.65 5 0.0813 0.00161 1.98
L9 1.86 6 0.0598 0.000613 1.03
50 0.988 5 0.0360 0.00112 3.11
51 14.0 5 0.237 0.00509 2.15
52 2L .0 5 0.319 0.00508 1.59
53 15.2 5 0.2L8 0.00485 1.96
54 20.2 5 0.268 0.00476 1.78
55 7.50 5 0.152 0.00266 1.75
56 2L.8 5 0.320 0.00662 2.07
57 33.0 5 0.378 0.0088L 2.3kL
58 29.6 5 0.3b42 0.00554 1.62
59 21.6 5 0.291 0.00L57 1.57

 

*Measured under conditions at top of column.
ORNL DWG 71-13968

 

 

 

 

 

1000 I I | I l 1 I I
500 _
©
®
"
’I
S 8 A~y
~ 7
- o i
z /ޤ:
s 5~
o
w
w
O
O
z
©
n
x S5O0 —
w 7
a ///o/ © 15-in.-1.D. COLUMN
o ——A T T, A 20-in.-1.D. COLUMN
- —_—— e 0 3.0-in-1.D. COLUMN
g L
> ~
<< A _/‘
»/O/
v/
'/
/
0 1 1 1 | 1 1 | 1
O.1 03 05 1.0 3 5 10 30 50 100

GAS SUPERFICIAL VELOCITY (cm/sec)

Fig. 17. Variation of Axial Dispersion Coefficient with Gas Superficial
Velocity and Column Diameter in Open Bubble Columns Having Diameters of 1.5,
2, and 3 in.

2§
53

The line shown for a column diameter of 2 in. is the same as that re-
ported earlier, based on data obtained by use of the steady-state tech-
nique. In the present study, however, data have been obtained at higher
gas flow rates than were previously used; these data show minimal scatter,
and thus the curves for the various column diameters are more clearly
defined. On the basis of these data, the transient technique appears to
be far superior to the earlier steady—state technique in that (1) less
scatter is observed in the dispersion coefficient data, and (2) data can
be obtained much more rapidly with the transient technique. A typical
run using the transient technique requires less than 10 min, whereas more

than 2 hr is required for a run using the steady-state technique.

The data on axial dispersion obtained with the 1.5-in.-diam column
are quite similar to those reported earlier. However, the high quality
of the present data shows that, at high gas flow rates, the dispersion
coefficient values obtained with the 1.5-in. column are slightly lower
than values obtained with a 2-in.-diam column at the same superficial
gas velocity. When the gas flow rate is increased sufficiently to produce
a slugging condition, the dispersion coefficient data for the two column
diameters result in parallel lines that are separated by about 15%. At
lower gas flow rates where bubbly flow occurs, the curves for the two
column diameters have widely different slopes (as had been reported
earlier). The slope of the curve for the l1.5-in.-diam column appears

to change only slightly in moving from bubbly to slug flow.

Thus far, the principal advantage of using the transient technique
has been improved axial dispersion coefficient data for the 3-in.-diam
column. Although the scatter in the present data increases as the column
diameter is increased, the data for the 3-in.-diam column appear to ad-
equately define the dependence of axial dispersion coefficient on super-
ficial gas velocity. There appear to be three distinct regions of opera-
tion, and the data from each region can be fit by a straight line as in-
dicated in Fig. 17. In the first region, bubbly flow is observed for
superficial gas velocities up to about 3 cm/sec. The second mode of

operation consists of a transition region covering superficial gas veloc-
54

ities from about 3 to about 25 cm/sec. The third mode of operation con-
sists of slug flow, which occurs at superficial gas velocities greater
than 25 cm/sec. The axial dispersion coefficient values measured with
the 2- and 3-in.-diam columns in the slug flow region show little dif-

ference.

5.7 Discussion of Data on Gas Holdup

The variation of gas holdup with superficial gas velocity aund column
diameter is shown in Fig. 18. At low gas flow rates (in which bubbly flow
occurs), gas holdup is proportional to the superficial gas velocity and
is independent of the column diameter. At higher gas flow rates, gas hold-
up data from the different column diameters begin to diverge and holdup
increases, but in a more gradual manner, as the gas superficial velocity
increases; for a given gas superficial velocity, holdup increases as the
column diameter is decreased.

Davies and Taylorlo report that the rate of rise of a single gas "slug"

in a column filled with stationary liquid is given by the relation:

Vl = 0.35yad , (2k)
where
Vl = rate of rise of gas slug,
g = gravitational acceleration,
d = column diameter.

In an open bubble column operating at steady state, a number of gas slugs
will be essentially evenly spaced throughout the column. In this case, the
relative velocity of the liquid between the gas slugs will be equal to the
superficial gas velocity. If the slug velocity as defined by Davies and
Taylor is assumed to be the relative velocity between liquid and gas in

an open bubble column, the gas holdup would be given by the relation:
GAS HOLDUP

 

 

 

 

0.6 ORNL DWG 71-13970
' I | | | | | 1 | | l | 1 | 1 | 1 T
051 _
A A
A
0.4} A -
O o
0.3 s —
o
®
o
0.2\ A-11/2 in. COLUMN -
®
e O-2 in. COLUMN
: -3 in. COLUMN
O.1 _
o
0 oy
0 2 4 6 8

10 12 14 16 I8 20 22 24 26 28 30 32 34 36
SUPERFICIAL GAS VELOCITY (cm/sec)

Fig. 18. Variation of Gas Holdup with Superficial Gas Velocity and
Column Diameter in Open Bubble Columns.

qS
56

Vv
= £
¢—V + V (25)
g 1
)
—_—

where
¢ = gas holdup,
Vg = superficial gas velocity.

A plot of the gas holdup data in the form suggested by Eq. (25) is shown
in Fig. 19. It should be noted that the holdup data from the 1.5-, 2-,
and 3-in.-diam columns are brought together on a single line which passes
through the origin. However, the slope of the line is approximately 0.78,
rather than the expected value of 1.0. Thus, while Eq. (25) does not
produce an exact fit of the gas holdup data, it appears to be useful in
suggesting a form for correlating the data. Because the slope of the
resulting line is not unity, the theoretical basis for the correlation

still must be established.

5.8 Conclusions

An improved experimental method for measuring axial dispersion coef-
ficients in open bubble columns has been developed. This method, an un-
steady-state technique, has been shown to be more accurate and to be ap-
plicable over a wider range of superficial gas velocities than the steady-
state technique employed previously. In addition, use of the transient
technique allows completion of an experiment in less than 10% of the time
required for the steady-state technique. Axial dispersion coefficilent
data obtained in columns having diameters of 1.5, 2, and 3 in. are in
agreement with previously obtained data. A theoretical basis for cor-
relating these data, as well as those on gas holdup in bubble columns,
must be developed before such results can be used to predict gas holdup

and the effect of axial dispersion in continuous fluorinators.
o7

ORNL DWG 71-1397]
I | i | I | T T T

 

0.60 T T

0.52|— —
0 48— —
044 o -—
040 —]
0.36|— A —
0.32— o0 —

A
0 28— O —
°

GAS HOLDUP

0.20+— AN e —~3 in. Diam. COLUMN —
. A-2in. Diam. COLUMN

0.16 — . O —15in Diam. COLUMN

0.08— . —

0.04— o

 

 

 

Coor e L
OO .04 08 12 16 20 .24 28 32 36 40 44 48 52 56 60 64 68 .72 76

Vg /{Vg + 0.35 V/gd)

Fig. 19. Variation of Gas Holdup in 1.5-, 2-, and 3-in.-diam Open
Bubble Columns with the Quantity Vg/(Vg + 0.359gd).
58

The unsteady-state technique for measuring axial dispersion coef-
ficients does not require water to be fed through the open bubble column,
and it may prove especially useful in studying effects on axial dispersion
resulting from changes in the physical properties of the continuous (water)
phase. Since at least one measurement (and possibly several measurements)
of the axial dispersion coefficient can be made with one column volume of
the liquid phase, it should be possible to work with materials that would

have been considered too expensive with the steady-state technique.

6. DEVELOPMENT OF THE METAL TRANSFER PROCESS
E. L. Youngblood L. E. McNeese

It has been found that rare earths distribute selectively into molten
lithium chloride from bismuth solutions containing rare earths and thorium,
and an improved rare-earth removal process based on this observation has
been devised. We are currently engaged in a demonstration of all phases
of the improved rare-earth removal method, which is known as the metal
transfer process. In a previous engineering experiment (MTE-l),ll we
studied the removal of rare earths from single-fluid MSBR fuel salt
by this process. During this experiment, approximately 50% of the
lanthanum and 25% of the neodymium originally present in the fluoride
salt were removed at approximately the predicted rate. However, the
lanthanum and neodymium that were removed from the fluoride salt did
not accumulate in the lithium-bismuth solution used for removing these
materials from lithium chloride as expected. Reaction of impurities
in the system with the rare earths is believed to have caused this un-
expected behavior. A second engineering experiment (MTE-2) is currently
under way. The main objectives of this experiment are: (1) to demonstrate
the selective removal of rare earths from fluoride salt containing thorium
fluoride, (2) to collect the rare earths in a Li-Bi solution, and (3) to

verify previous distribution coefficient data.
29

6.1 Equipment for Experiment MTE-2

Experiment MTE-2 is being carried out in a vessel made of 6-in.-diam
carbon-steel pipe (see Fig. 20). The vessel is divided into two compart-
ments by a partition that extends to within 1/2 in. of the bottom of the
vessel. The compartments are interconnected by a 2-in.-deep pool of
thorium~saturated molten bismuth, which forms a seal between the two com-
partments. One compartment contains MSBR fuel carrier salt (72-16-12 mole
% LiF—BeFQ-ThFh) to which tracer quantities of lhTNd and sufficient LaF3
to produce a lanthanum concentration of 0.3 mole % were added. The other
compartment contains LiCl, Li-Bi solution (in a cup), and a pump for cir-
culating the LiCl through the cup at a flow rate of about 25 cm3/min. The
concentration of reductant (35 at. % Li) in the Li-Bi solution is suf-
ficiently high to permit essentially all of the lanthanum and neodymium to

be extracted from LiCl in equilibrium with the Li-Bi solution.

In order to obtain mixing in the main bismuth pool, about 10% of the
metal volume is forced to flow back and forth every T min through the 1/2-
in. slot located below the partition between the fluoride and chloride
compartments. This flow is effected by reducing the pressure of the argon
cover gas in the fluoride compartment relative to that in the chloride
compartment. Gas-lift sparge tubes are used to disperse droplets of bis-~
muth in the salt phase to improve contact between the salt and bismuth
phases in each compartment. Lumps of thorium metal (a total of 149 g),
measuring approximately 0.5 in. on a side, were placed in the bottom of
the vessel in order to ensure that the bismuth phase is saturated with
thorium. The amount of thorium added in this manner is about five times
the amount that could dissolve in the bismuth. The lower section of the
vessel is heated to the operating temperature (650 to 660°C) by an 8-kW
furnace. The flange and upper 6 in. of the vessel are wrapped with
cooling coils through which water flows in order to maintain the flange
at approximately 100°C. This provision prevents damage to the neoprene
gasket used for sealing the flange to the vessel. The exterior of the
vessel was spray coated with 20 mils of nickel aluminide to protect

against air oxidation of the carbon steel.
60

ORNL DWG 70-12503
LEVEL
ELECTRODES

ARGON INLET
AND VENT .

 

 

L ]
!
| —CARBON STEEL PUMP
WITH MOLTEN Bi
CHECK VALVES

6- -
CARBON STEEL — T STEEL Pipgt

ITION \ STEEL PIP

 

 

 

 

    

 

 

 

 

 

 

 

 

24 in
E‘. LiCl
72-16-12 MOLE %~ = 7- |
FUEL CARRIER SALT = =
> 7 N —— 11—
\l /22 Li-Bi
Y

 

Fig. 20. Carbon Steel Vessel Used for Metal Transfer Experiment MTE-2.
61

The pump used for circulating the LiCl through the cup containing
the Li-Bi solution is constructed of carbon steel and uses molten bismuth
as check valves (see Fig. 21). The quantity of LiCl pumped during one
cycle of operation is controlled by level electrodes (in the pump chamber)
that actuate a solenoid valve in a gas supply line to the pump chamber.

A pressure of about 18 in. H,O is maintained in the LiCl compartment of

the main vessel by the use oi a mercury bubbler in the vessel off-gas.

When the pressure in the pump chamber is decreased to near atmospheric
pressure, LiCl flows into the pump chamber until the upper electrode is
contacted by salt. At this point, the solenoid valve opens and argon is
admitted into the pump chamber to discharge the LiCl. When the salt

level falls below the lower electrode, the solenoid valve is closed and
the pump chamber is again vented. The pumping rate is determined by the
distance between the two electrodes and the frequency of the pumping cycle.

An electrically operated counter is attached to the solenoid valve in order

to record the number of pumping cycles that have occurred.

Provisions were made for obtaining filtered samples of the salt and
‘bismuth phases in the experiment. The samplers, which are constructed of
stainless steel, consist of 1/L4-in.-0D, l-in.-long stainless steel capsules
that have a length of 1/16-in.-0D capillary tubing attached to one end and
a porous metal filter (mean pore size, 20 u) attached to the other. 1In
preparation for obtaining a sample, the capillary tubing is inserted through
a hole in a Teflon plug located in a chamber above a ball valve leading
into the system. The chamber is purged with argon before the ball valve
is opened in order to prevent air from entering the system. The capillary
tube is used to push the sample to the proper location within the vessel.
An argon purge is maintained through the capillary tubing to prevent ma-
terial from entering the sampler before the sampler reaches the level at
which a sample is to be taken. When the sampler is in position, a sample
is obtained by applying vacuum to the capillary tubing. The sample is
then removed and prepared for analysis by removing the filter and capillary
tubing and cleaning the outside surface of the sampler with emery cloth.
The sample is analyzed for neodymium and radium by direct gamma counting;
the lanthanum concentration is determined by a neutron activation technique

after the sample has been dissolved.
ORNL DWG 70 -898I

 

 

 

 

 

 

 

 

 

 

LA TO VENT AND
N \QQ WARGON SUPPLY
TEFLON \
PLUG \
N
i)
N\
ELECTRODES 172" CARBON STEEL
FOR LIQUID TUBING 28" LONG
LEVEL %
MESUREMENT N
N
.
| )
N
N I/4"CARBON STEEL
( s TUBING
N
N
DISCHARGE
L I — MOLTEN BISMUTH
|1 ///
N “
; /|
[
N
N
S
A\

 

 

 

 

 

 

 

INTAKE

Fig. 21. Carbon Steel Pump Having Molten Bismuth Check Valves Used to
Circulate LiCl in Metal Transfer Experiment MTE-2.
63

6.2 Materials Used in Experiment MTE-2

The quantities of materials used in experiment MIE-2 are given in
Table 8. The LiCl was purified prior to use by contact with thorium-
saturated bismuth at 650°C. Both the carbon-steel vessel and the bismuth
were treated with hydrogen at 650°C to remove oxides. The fluoride salt
used in the experiment was purified by the Reactor Chemistry Division.
The argon used as cover gas for the experiment was purified by passage
through a bed of uranium turnings held at 600°C and a bed of molecular

sieves.

Table 8. Materials Used in Metal Transfer Experiment MTE-2

 

 

 

Quantity
Material em3 g-moles

Fluoride salt 789 Lo.7

(LiF-BeFp~ThF),-LaF3,

72-15.7-12-0.3 mole %,

7 mci 14TnaF3)
Bismuth saturated with thorium 799 36.9
LiCl 10Lk2 - 36.6
Li-Bi (35 at. % lithium) 164 9.5

 

6.3 Status of Experiment MTE-2

To date, experiment MTE-2 appears to be operating satisfactorily;

however, no data are available at this time.
6L

7. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS
IN A MILD-STEEL FACILITY

B. A. Hannaford C. W. Kee
L. E. McNeese

We have continued operation of a facility in which a system constructed
of mild steel can be used to carry out semicontinuous reductive extraction
experiments.12 Initial work with the facility was directed toward obtain-
ing data on the hydrodynamics of countercurrent flow of molten salt and
bismuth in an 0.82-in.-ID, 2hk-in.-long column packed with 1/4-in. molybdenum
Raschig rings. We have been able to show that flooding data obtained with
this column are in agreement with predictions from a correlation based on
studies of the countercurrent flow of mercury and aqueous solutions in
packed columns.13 We have recently undertaken experiments for determining
the mass transfer performance of the packed column. In the initial experi-
ments, a salt stream containing UFh was countercurrently contacted, over a
wide range of operating conditions, with bismuth containing reductant. The
first uranium mass transfer experiment (UTR-1) was very successful hydro-
dynamically and demonstrated that simultaneous samples of the bismuth and
salt streams leaving the extraction column can be taken easily. However,

the run failed to provide mass transfer data because of difficulties that

14

prevented the addition of reductant to the bismuth prior to the experiment.
During the second uranium mass transfer run (UTR-2), bismuth containing
reductant was fed to the column at the rate of 24T cm3/min, and salt (72-
16~12 mole % LiF—Bng—ThFh) containing 3000 ppm of uranium as UF) was fed
to the column at the rate of 52 cm3/min. The flow rates of bismuth and
salt were both steady over a period of about 40 min, and 95% of the uranium
was extracted from the salt.lLL The experiment represented the first known
demonstration of the continuous extraction of uranium from molten salt into
bismuth containing reductant and was encouraging in that it suggested that

high uranium removal efficiencies can be obtained in a packed column having

a reasonable length.
65

T.1 Preparation for Mass Transfer Run UTR-3

Following run UTR-2, the salt and bismuth were returned to the treat-
ment vessel and 122.5 g of thorium metal was added to the graphite crucible
through a 3/4-in.-diam tube. Periodically, bismuth samples were taken
with graphite ladles. Analyses of these samples showed that the thorium
concentration in the bismuth increased at the rate of about 2 ppm/hr,
which is the same rate of increase observed following the earlier addition
of thorium to the bismuth feed tank prior to run UTR-2. The thorium con-
centration in the bismuth was only 800 ppm after 250 hr, which is far below
its solubility at 660°C (i.e., 3500 ppm). The rate of dissolution appeared
to be limited by poor contact of the bismuth with the 0.5-in. cubes of
thorium. In order to improve the extent of contact between thorium and
bismuth, a second charge of thorium metal (119 g) was loaded into a l-in.-
diam, 6~in.-long perforated steel basket and was lowered into the bismuth
phase. After this addition, the observed thorium dissolution rate increased
to about ten times the earlier rate (20 ppm/hr based on graphite ladle
samples, and 25 ppm/hr based on change in weight of the addition basket).
About 85% of the thorium dissolved during a 22-hr period. After approxi-
mately 24 additional hours, the thorium concentration in a ladled bismuth
sample was 1665 ppm, which is 95% of the value expected from complete

dissolution.

Surprisingly, bismuth samples taken 50, 100, and 350 hr after com-
pletion of the thorium addition showed a sharp decrease in thorium con-
centration (to 1000 ppm), which was followed by a slow increase in thorium
concentration (to 1200 ppm). In order to check for a concentration gradient
within the pool of bismuth, samples were taken with a ladle at a point about
1 in. from the bottom of the vessel and at a point close to the bismuth-
salt interface (about 6 in. from the bottom of the vessel). Two pairs of
samples showed that the thorium concentration was higher near the bottom
of the vessel (17% higher in one case and 88% in the other). This behavior
can be explained as follows. After thorium had been added to the bismuth,
the composition of the bismuth phase changed slowly as the bismuth and salt
66

approached a state of chemical equilibrium. During this period, part

of the thorium was oxidized from the bismuth phase by reaction with LiF
from the salt and was replaced with the same number of equivalents of
lithium. In the absence of mixing in the bismuth phase, a sharp density
gradient would have been produced in which the least-dense material would
occur at the top of the bismuth pool and the most-dense material would
occur at the bottom of the bismuth pool. It was concluded that the mixing

within the bismuth phase was inadequate to maintain a uniform composition.

Filtered samples of the bismuth phase were usually taken, using the
stainless steel samplers, at the same time that samples were obtained with
a graphite ladle. The thorium content of samples obtained in the stainless
steel samplers was consistently lower than that of samples obtained with
the graphite samplers. In order to check the possibility that the method
for preparing the filtered samples for analysis might be at fault, two
stainless steel samplers>were cut into two or three segments and analyzed
separately. The results showed that the thorium concentration in the bot-
tom third of the sample capsule was much higher (about 1900 ppm) than in
the upper third (about 300 ppm). This clearly confirmed our suspicion
that thorium was concentrating in the bottom of the sample capsule during
the slow freezing of the sample. Presumably, the observed concentration
of thorium resulted from the fact that the density of thorium bismuthide
(11.L4 g/cmB) 1s greater than the density of bismuth (9.66 g/cm3). Thus,
the method for preparing samples taken in a stainless steel sampler,
which involved discarding the bottom of the capsule (containing the porous
metal filter and about 15% of the bismuth), resulted in substantial errors
in the measured thorium concentration. The sample preparation method was
revised so that the outside surface of the porous metal frit was mechanical-
ly cleaned; however, the frit was left attached to the capsule. We have
continued to rely primarily on samples obtained with a graphite ladle for
analyzing for thorium in bismuth because the metal sample can easily be
removed from the ladle and is not subject to difficulties peculiar to the

leaching operation required for the filtered samples.
67

7.2 Mass Transfer Run UTR-3

Immediately prior to run UTR-3, about 5 liters of salt was transferred
from the treatment vessel to the salt feed tank. On completion of this
operation, a total of 15 liters of salt, in which the uranium concentration
was about 2000 ppm, was present in the feed tank. About 15 liters of bis-
muth was then transferred from the treatment vessel to the bismuth feed tank.
The experiment was initiated by setting the salt flow rate through the column
at approximately the desired value; subsequently, the flow of bismuth through
the column was initiated as shown in Fig. 22. During the initial part of
the run, bismuth and salt were countercurrently contacted in the packed col-
umn at flow rates of 205 and 100 cm3/min,Arespectively. Five sets of samples
were obtained from the salt and bismuth streams leaving the column at the
times indicated in Fig. 22. The salt flow rate was then increased to 160
cm3/min, and a single set of salt and bismuth samples was taken after a
salt volume equivalent to three column volumes had passed through the column.
The salt flow rate was then increased to 220 cm3/min, and a final set of
salt and bismuth samples was taken after three additional column volumes
of salt had passed through the column. The temperature along the extrac-
tion column during the run varied from 627°C at the bottom of the column

to 640°C at the top.

Data obtained during run UTR-3 are summarized in Table 9. The high
uranium concentration in the first salt sample is probably the result of
dilution in the salt effluent piping which initially contained salt with a
uranium concentration equal to that in the feed tank. For this reason,
the point corresponding to this sample was excluded from consideration in
calculating the fraction of uranium that was transferred. As the bismuth-
to-salt flow rate ratio was decreased, the fraction of the uranium removed
from the salt decreased, as expected, from an average value of 0.91 during
the first period of operation to 0.73 during the final period of operation.
As in the previous run, the uranium concentration in the bismuth in the re-
ceiver tank (528 ppm) was significantly higher than that which would be
calculated (about 44O ppm) by integrating the product of the uranium con-
FEED REMAINING (liters)

ORNL DWG 71-13982R|

 

 

    
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 T T T T T T T T T
+! +2 +3 +4 +5 +6 +7 SAMPLE NUMBERS
I8+ —
——BISMUTH
c ———SALT
£ €
I6F=~<_ € -~ —
~== SALT, 100 ¢m3/min - <t
~——— -
- < E "o
n° o
14 o & -
12+ —
10 \----,_\\ -
\\\
8 BI,205 cm3/min 7]
61— _
4|- —
2'_ —
Bi, 195 cm3/min ]
0 | ll | I I ] 1 T] |
0 10 20 30 40 50 60 70 80 . 90 100
TIME (min)

Fig. 22. Volumes of Bismuth and Salt Remaining in Feed Tanks vs Run
Time, Run UTR-3. Volumetric flow rate (ml/min) for each indicated interval
was inferred from the slope.

89
Table 9. Summary of Mass Transfer Data Obtained During Run UTR-3

 

 

 

 

Uranium
Conc. in Fraction of
Salt Uranium Volumetric Bismuth Phase
Phase Remaining Flow Ratio, U Comnc. Zr Conc. Th Conc. L.i Conc. Sum
Sample (ppm) in Salt Bismuth/Salt ppm  meq/liter ppm meqg/liter ppm meg/liter ppm meq/liter meg/liter
Feed tanks 2100 200 32.% L0 6.6 1200 199 307 b1k 260
Flowing stream
1 461 0.220 2.05 LTS 77.1 57 9.4 25k ho.3 3k L7.3 176
2 103 0.0k4g 2.05 405 65.8 37 6.1 420 70.0 37 51.5 193
3 160 0.076 2.05 Lok 65.6 L7 7.7 300 50.0 Lo 55.7 i79
L 158 0.075 2.05 362 58.8 43 7.1 208 3L.6 36 50.1 151
5 326 0.155 2.05 L2o 68.2 54 8.9 282 L7.0 35 L8.7 173
6 LLs 0.217 1.22 458 Th kL 55 9.0 192 32.0 27 37.6 153
7 S61 0.267 0.91 502 81.5 53 8.7 226 37.€ 27 37.6 165
Ave. 170
b
Receiver tanks 272 528 85.7 Lot 67.8 34 LT.3 209
U balance, flowing stream samples = U entering B1 _ 129 millimoles U _ 0.78

 

U leaving salt 160 millimoles U

69

 

aEquilibrium lithium concentration calculated from thorium analysis.

bAssume zirconium concentration = 8.1 meg/liter.
70

centration in the flowing bismuth samples and the bismuth flow rate. A
similar difference is also noted in the measured and calculated values for
the total number of equivalents of transferrable materials (U, Th, Zr, and
Li) per liter of bismuth. This quahtity should be invariant during an
experiment, except for changes resulting from the fact that the uranium

in the exit salt has a valence of 3+ rather than 4+ as in the entering salt.
However, this effect accounts for only about 5 meq of reductant per liter,
or less than 2% of the total milliequivalents of transferrable materials
per liter of bismuth. Therefore, the observed variation in the number of
milliequivalents of transferrable materials per liter of bismuth must be
ascribed to errors in the sampling procedure or in the method of analysis,
or to the presence of oxidants in the system. The uranium material balance

during the run was T78%, based on analysis of the flowing stream samples.

7.3 Preparation for Mass Transfer Run UTR-k4

Following mass transfer run UTR-3, both the salt phase and the bismuth
phase were transferred to the treatment vessel for a 16-hr equilibration
period. After both phases were sampled, most of the salt was transferred
to the salt feed tank. At this point, the treatment vessel contained about
15 liters of bismuth and about 2 liters of salt. Sufficient thorium metal
was charged to the treatment vessel to produce a thorium concentration of
about 1000 ppm, which is about 10% greater than the solubility of thorium
in bismuth at the bismuth feed tank operating temperature (900 ppm at 540°C).
A filtered sample of the bismuth taken 44 hr after addition of the thorium
indicated a thorium concentration of 980 ppm, which is near the expected
value. The rate of dissolution of the thorium, based on the observed thorium
concentration, was about that observed prior to run UTR-3. However, visual
inspection of the thorium addition basket showed that less than half of the
thorium had dissolved in the bismuth. The basket was returned to the bis-
muth phase, and the temperature of the treatment vessel was increased from
600°C to 660°C. After a total period of 115 hr, the addition basket was
removed and weighed. It was found that about 90% of the thorium had dis-

solved. The bismuth was then sampled and transferred to the bismuth feed
71

tank. The reported thorium contents of two samples removed from the
treatment vessel were 828 ppm and 710 ppm. Because of the discrepancies
in these concentrations, three additional samples of the bismuth were
taken from the bismuth feed tank and were analyzed by both the colori-
metric method and a spectroscopic technique. Results of these analyses
are shown in Table 10. Although the values reported for the two different
analytical techniques showed appreciable scatter, the bismuth in the feed
tank appeared to have a nonuniform composition. The lower concentration
of thorium near the bottom of the tank might have been due to a lower
temperature at the bottom of the tank and a correspondingly lower thorium

solubility.

Table 10. Thorium Concentrations Reported for Bismuth Samples Removed
from the Bismuth Feed Tank Prior to Run UTR-L

 

 

 

Location
at Which (Thorium Concentration (ppm)
- Sample Colorimetric Spectroscopic
Was Taken Type of Sampler Analysis Analysis
Bottom of tank Graphite ladle T1h 830
15 in. from bottom Graphite ladle 1080 1200
Middle of bismuth pool Stainless steel, 1189 920

filtered

 

7.4 Mass Transfer Run UTR-L

Immediately prior to run UTR-4, about 15 liters of bismuth and about
15 liters of salt were transferred from the treatment vessel to the salt
and bismuth feed tanks. The uranium concentration in the salt was 1681 ppm.
As shown in Fig. 23, the run was initiated by starting a salt flow at a low

rate through the column. After the salt metering system was determined to
ORNL DWG 71—13983RI
T T i

 

 

 

 

 

 
  

 

  
 
    
  

 

  
   

 

 

 

20 T 1 I T T I ¥ T 1 T
+ + + + + + + + + + + 4+ + +
| 2 34 5 6 7 8 910 ni213 14 SAMPLE NUMBERS
18}~ ———BISMUTH
BISMUTH 140 e¢m¥min | ===-- SALT
16 _
= b
~14} 5E ]
v 2
® 9%
= ®
2k . = 4
o ~
2 R BISMUTH
> . 157 cm¥min
< |10 -
=
Wl
[0
a 8 -
w
w
"%
°r ]
T |
SALT _
2r 18 cm¥min
0 1 1 l I { 1 1 ! 1 1
0 10 20 30 40 50 60 70 80 90 1o 120 130 140
TIME (min)

Fig. 23. Volumes of
Time, Run UTR-L.

Bismuth and Salt Remaining in Feed Tanks vs Run

cL
73

be functioning satisfactorily, a bismuth flow was started through the col-
umn, and the salt and bismuth flow rates were adjusted to 149 and 1k0 cm3/
min, respectively. Seven pairs of salt and bismuth samples were obtained
from the salt and bismuth streams leaving the column. The salt and bismuth
flow rates were then decreased to 118 and 117 cm3/min, respectively, and
three pairs of salt and bismuth samples were withdrawn. The salt and bismuth
flow rates were then adjusted to 210 and 157 cm3/min, respectively. However,
these flow rates were maintained during only about one-half of the remainder
of the run. During this period, four additional pairs of salt and bismuth
samples were taken from the salt and bismuth streams leaving the column.

The temperature along the extraction column during the run varied from

615°C at the bottom of the column to 630°C at the top.

Data obtained during run UTR-4 are summarized in Table 11. A notice-
able change in the fraction of uranium extracted from the salt is observed
as the bismuth-to-salt flow rate ratio varies throughout the range 0.75 to
1.0. The total number of milliequivalents of transferrable materials per
liter of bismuth in the bismuth stream leaving the column was essentially
constant throughout the run, and is in agreement with the value obtained
by analyzing a sample from the bismuth receiver tank. The total number
of milliequivalents of transferrable materials per liter of bismuth, as
indicated by analysis of bismuth from the feed tank, is about 30% higher
than values observed during the run and values obtained by analyzing the
material present in the bismuth receiver tank. The indicated lithium
concentration in the bismuth in the feed tank is higher than expected
and could signify that the bismuth sample was contaminated with salt. A
slight change in the total number of milliequivalents of transferrable ma-
terials per liter of bismuth is expected since the uranium remaining in
the salt has a valence of 3+ rather than 4+ as in the inlet salt. However,
this effect would result in a change of only about 10 meq of transferrable

materials per liter of bismuth.
Table 11. OSummary of Meass Transfer Data Obtained During Run UTR-L

 

 

 

 

 

Uranium
Conc. in Fraction of
Salt Uranium Volumetric Bismuth Phase
Phase Remaining Flow Ratio, U Conc. Zr Conc. Th Conc. Li Conc. Sum
Sample {(ppm) in Salt Bismuth/Salt ppm  meg/liter ppm meq/liter  ppm meqg/liter ppm meq/liter meq/liter
Feed tanks 1681 480 77.8 53 8.7 1097 182 67 92.5 3€1
Flowing stiream
1 LT3 0.281 0.94 753 120 71 11.7 342 57.0 36 50.1 o4
2 549 0.327 0.9k 7hs 121 38 6.3 399 66.5 NG S5 .7 250
3 L3l 0.256 0.04 784 127 LL8 7L.6 13 28,1
Y 636 0.378 0.94 skl 88 32 5.3 332 55.3 29 4o.k 189
5 515 0.306 0.9k 773 126 39 6.4 LET 77.8 31 k3.2 253
6 Los 0.294 0.9k 7ol 118 38 6.2 Log 83.1 34 LT3 258
7 5702 0.340 0.9k 6l 105 Ly 7.2 595 99.1 35 L8 .7 260
Ave. 0.312
8 LL1 0,262 1.0 833 135 32 5.3 508 g8L.,6 33 45,5 271
9 458 0.272 1.0 597 97 50 8.2 395 65.8 3k L7.3 218
10 4518 0.2L9 1.0 ) 105 L2 6.9 391 65.1 34 L7.3 c2k
Ave. 0.261
11 668 0.397 0.75 806 131 Ly 6.7 3Lkg 58.1 31 L3.2 239
1z 720 0.L28 0.75 778 126 3k 5.6 399 66.5 32 Lk .5 243
13 551 0.328 0.75 L5 121 30 L.g 488 81.3 36 50.1 257
ik 567 0.337 762 124 38 6.2 493 82.1 29 40.L 252
Av 0.384%
Receiver tanks 423 803 130.4 53 8.7 586 97.6 32 Ly .5 281

tl

 

a . . . . - R .
Latz from sample 1% was not used in calculating this value because of variation in the salt flow rate.
75

(.5 Mathematical Analysis of Mass Transfer with Primary
Resistance to Transfer in the Salt Phase

15

It has been noted ~ that the uranium extraction data from runs UTR-3
and -4 can be correlated in terms of the height of an overall transfer unit
based on the salt phase (HTU) if several assumptions are made. The chief
assumption is that the rate at which uranium transfers to the bismuth phase
will be controlled by the diffusive resistance in the salt film when the
extraction factor is high and when the salt film is composed largely of non-
transferring ions. In runs UTR-3 and -4, a small amount of uranium was
added to the salt after it had been equilibrated with the bismuth phase,
which contained reductant. Thus, no significant transfer of lithium and
thorium between the salt and bismuth occurred in the column. In this case,
the overall transfer coefficient based on the salt phase is equal to the
individual salt film transfer coefficient. By definition, the HTU and the

number of overall transfer units based on the salt phase developed in the

column are related as:

H = HTU-NTU, (26)

where
H = column length,
HTU = height of an overall transfer unit based on the salt phase,
NTU = number of overall transfer units based on the salt phase.

If it is assumed that uranium is the major component transferring from the
salt and that the controlling resistance to transfer 1s in the salt phase,

the HTU can be defined as follows:

V

HTU=E§-, (27)

where
VS = superficial velocity of salt in the column, cm/sec,
k = overall mass transfer coefficient based on the salt phase, cm/sec,
a = interfacial area between salt and bismuth phases per unit column

3

2
volume, cm /em™.
76

It has been observed previouslyl6 that the dispersed-phase holdup is ap-
proximately proportional to the flow rate of the dispersed phase in a
packed column, except at conditions near flooding. We would, therefore,
expect the interfacial area between the salt and bismuth phases to be

proportional to the bismuth flow rate; thus, one can write the following

relation:
1
ka = k VBi’ (28)
where
k' = a constant, cm_l,
VBi = superficial velocity of the bismuth in the column, cm/sec.

The number of overall transfer units (based on the salt phase) developed

in the column is defined as:

o
NTU = E §;9§j§ , (29)
*in
where
X = uranium concentration in the bulk salt, ppm,
X* = uranium concentration in salt in equilibrium with the bulk bismuth
phase, ppm,
Xi = uranium concentration in the salt fed to the column, ppm,
XO = uranium concentration in the salt leaving the column, ppm.

*
As shown in Table 12, the value of X at the bottom of the column is much
*
smaller than the value of XO; thus, one would expect that the value of X
would be much less than the value of X throughout the column. In this
case, Eg. (29) can be integrated to yield the following expression:
X

NTU = -1n — . (30)
i

=
Table 12. Summary of Mass Transfer Data Obtained During Uranium Mass Transfer Runs UTR-3 and UTR-k

 

Uranium Concentration in Salt (ppm)

 

Fraction of

 

Salt Salt Max imum Metal-to-Salt Fraction Uranium

Feed, Effluent Equilibrigm Flow Rate of Remaining

Run Xi XO Value,? X Ratio Flooding in Salt
UTR-3 2100 159 3.6 2.05 0.87 0.076
Lhs 9.4 1.22 1.04 0.212
561 10.3 0.91 1.23 0.267
UTR-L 1680 52l 9.7 0.94 0.85 0.312
439 7.2 1.0 0.69 0.261
646 11.5 0.75 1.07 0.38L4

 

aCalculated as the concentration that would be in equilibrium with the observed concentrations of
reductant (lithium) and uranium in the bismuth effluent. :

L.
78

Combining Egs. (26), (27), (28), and (30), and rearranging, yields the

relation:
Xo ' VBi
1ln X -k H v (31)
i S

which states that a semilogarithmic plot of the fraction of uranium re-
maining in the salt vs the bismuth-to-salt flow rate ratio should yield
a straight line having a slope of ~k'H. This line should pass through

an ordinal value of 1.0 at a bismuth-to-salt flow rate ratio of zero.

7.6 Discussion of Mass Transfer Data from
Runs UTR-3 and UTR-4

The mass transfer data obtained during runs UTR-3 and -4 are summarized
in Table 12. As shown in Fig. 24, the data are well represented by Eq. (31).
For these data, the constant k' has the value of 0.0208 cm-l. The product
of the overall mass transfer coefficient, based on the salt phase and the

interfacial area, 1s given by the relation:

ka = 0.0208 VBi’ (32)
where
ka = overall rate constant based on the salt phase, sec_l,
VBi = superficial velocity of bismuth in the column, cm/sec.

Values for the overall ratc constant for the present data range from (0.012
to 0.021 sec_l and compare favorably with a preliminary value of 0.0076

sec-l measured:LT for the transfer of uranium from a 96.2-3.6-0.2 wt % Cd-
Mg-U solution to a molten salt (50-30-20 mole % MgClE-NaCl-KCl) at temp-

eratures ranging from 560 to 610°C.

The height of an overall transfer unit based on the salt phase, as
calculated from the data from runs UTR-3 and -4, is given by the expression

ey o Lkl

= o, (33)
VBi/Vs
79

ORNL DWG 71-2865 R2

 

 

 

 

1.0 T T Y ] T T T I T T T
p— -
- O UTR-3 Data 2
—~ OUTR—-4 Daoata —
- O —
}_
-
3) L -
Z
o
Z B .
Z
ot
=
W
T ol
= s e
2 - -
: -
@ - O -
o - —
. - =
o - ]
z | ]
- — ]
Q
s |
vt
W [~ -
- -4
0.0I 91 | L | | | 1 | ] 1L 4
0 0.5 1.0 i.5 2.0 2.5 30

BISMUTH-TO-SALT FLOW RATE RATIO, Wg;/Vs

Fig. 24. Variation of Fraction of Uranium Remaining in Salt with
Bismuth-to-Salt Flow Rate Ratio During the Countercurrent Contact of Salt
and Bismuth in a 0.82-in.-long Packed Column.
80

where
HTU = height of the overall transfer unit based on the salt phase,
cm,-
VBi/Vs = bismuth-to-salt volumetric flow rate ratio.

The lower and upper limits of the HTU values were O0.77 ft and 2.1 ft, which
corresponded to flow rate ratios of 2.05 and 0.75, respectively. The amount
of reductant present in the bismuth apparently had no measurable effect on
the rate of uranium transfer. Uranium mass transfer data from an earlier
experiment (UTR-2) are not in agreement with the data shown in Fig. 2L.
However, in experiment UTR-2, the thorium reductant was added to the bis-
muth in the feed tank and the bismuth and salt phases were not in chemical
equilibrium with respect to thorium and lithium as they entered the column.
Thus, both uranium and lithium were being extracted from the salt phase by
reaction with thorium from the bismuth phase, and it is not surprising that
the resulting uranium mass transfer rate is not in agreement with data ob-

tained during runs UTR-3 and -L.

T.7T Preparation for Zirconium Mass Transfer
Experiments; Run UTR-5

In order to measure mass transfer rates in the column under more close-
ly»controlled conditions and under conditions where the controlling resis-
tance is not necessarily in the salt phase, preparations were begun for
experiments in which the rate of exchange of zirconium isotopes will be
measured between salt and bismuth phases otherwise at chemical equilibrium.
In the present system, there is no means for effecting a separation of zir-
97Zr

conium isotopes; thus, by definition, the separation factor between

(16.8-hr half-life) and natural zirconium will be unity, as shown in Eqg.

(34):

 

=1, (3k4)
81

where
o = 97Zr—Zr separation factor,
D97 = 97Zr distribution coefficient,
2r
DZf = Zr distribution coefficient.
This relation also shows that the distribution coefficient for 972r will

be equal to the distribution coefficient for natural zirconium and hence
can be varied at will by adjusting the concentration of reductant in the
bismuth phase. In the experiments to be carried out, a small amount of
natural zirconium will be added to the system, and prior to each experi-
ment the reductant concentration in the bismuth will be adjusted to ob-
tain the desired zirconium distribution coefficient. Immediately prior

9TZr tracer will be added to the salt or bismuth phase

to an experiment,
in its respective feed tank, and the phases will be countercurrently con-
tacted in the packed column under the desired operating conditions. During

97

an experiment, the ~ Zr tracer will exchange for natural zirconium and the

mass transfer performance of the column can be evaluated by determining the
97Zr activity in the salt and bismuth streams leaving the column. This ex-
perimental technique has several advantages over the technique used thus

far in runs UTR-2, -3, and =4. First, the zirconium distribution coefficient
can be adjusted as desired. However, even for high values of the zirconium
distribution coefficient, no chemical changes other than the exchange of
zirconium isotopes will occur in the salt or bismuth. In this case, the
driving force for the transfer of 97Zr between the salt and bismuth phases
will be accurately known at all points throughout the column. Since the

972r tracer can be added either to the salt or bismuth feed tank, experi-
ments can be carried out in which the primary resistance to mass transfer
is in either the salt or bismuth phase, or in which the resistance to

transfer in both phases is of importance.

Before the zirconium mass transfer runs could be carried out, it was
necessary that (1) most of the reductant be removed from the bismuth, (2)
that undissolved thorium added to the bismuth feed tank prior to run UTR-2
be removed, and (3) that the salt and bismuth phases be brought to chemical
82

equilibrium. It was decided that the best way for effecting these changes
consisted of completely removing reductant from the bismuth in the treat-
ment vessel, after which the salt and bismuth would be transferred to their
respective feed tanks. An adequate period for dissclution of thorium from
the bismuth feed tank would be allowed; then the salt and bismuth would be
countercurrently contacted in the packed column. This operation would pro-
vide additional information on the hydrodynamics of countercurrent flow in
the packed column, and would provide an opportunity to obtain flowing
stream salt and bismuth samples under conditions where essentially no urani-
um mass transfer should occur (and hence essentially no variation should be
observed in the uranium concentration in the salt and bismuth streams

leaving the column).

The salt and bismuth were transferred to the treatment vessel and were
contacted with a nominal T0-30 mole % H2—HF mixture for 20 hr. At the end
of this period, the salt was sparged, first, with hydrogen for 6 hr in order
to reduce FeF2 to metallic iron and, then, with argon at the rate of about
0.13 ft3/hr for 15 hr to ensure complete removal of HF. The salt and bis-
muth were subsequently sampled and were transferred to their respective
feed tanks. After a delay of about 20 hr, the next experiment (UTR-5) was

begun.

In the initial part of the experiment, bismuth and salt were fed to
the column at the rates of 115 and 123 cm3/min, respectively, and seven
pairs of samples were taken of salt and bismuth leaving the column (see
Fig. 25). Throughout the remainder of the experiment, the salt and bis-
muth flow rates were varied in order to operate near expected flooding
conditions. During this period, six salt samples were obtained from the
salt stream leaving the column in order to check for evidence of bismuth
entrainment. After about 116 min of operation, the column flooded and
an alarm indicating a high salt level in the column was noted. At this
point, the bismuth entrainment separator became filled with bismuth. A
restriction in the drain line from the entrainment separator prevented

the accumulated bismuth from draining out of the salt exit line, and, in
20 l ORNL DWG 71-13984RI

FEED REMAINING (liters)

 

  
   
  

 

 

 

    
 

 

 

 

 

 

 

‘ + +]+ U l 8 9 410 | | / +
Fs st bt +8 #9410 +1 saMPLE NUMBERS*!2 I3
FLOWING SALT SAMPLES ONLY FOR Bi
18- ENTRAINMENT ]
La BISMUTH, !5 cm3/min BISMUTH
= [ -—---SALT _
14} -
12+ —
10 © 4
Ne
T
s
N >
8 =5 7
il
©
SALT, 123 cm3¥ min cm3/min
. \
cm3/min N
\\
al- ~ .
\\\ o
2 SALT 7]
50
cm¥Ymin
0 ] ] | 1 : | ! ] | 1 ] 1 1
40 50 60 70 80 90 100 1o 120 130 140 150 160 170
TIME (min)

Fig. 25. Volumes of Bismuth and Salt Remaining in Feed Tanks vs Run

180

Time, Run UTR-5. Volumetric flow rate (ml/min) for each indicated interval

was inferred from the slope.

€8
8

order to overcome the resulting back pressure, an argon overpressure of

21 in. HEO was imposed on the top of the column for the remainder of the
experiment. The salt samples taken during the latter half of the experi-
ment contained large quantities of bismuth (>50 vol %). It is likely

that this was caused by the restriction in the entrainment separator

drain line and the subsequént accumulation of bismuth in the salt stream
sampler. The temperature along the extraction column during the run varied

from 615°C at the bottom of the column to 622°C at the top.

Data obtained during the first half of the run by analyzing salt and
bismuth samples from the treatment vessel, the feed tanks, the salt and
bismuth streams leaving the column, and the salt and bismuth receiver
tanks are shown in Table 13. The concentration of thorium in the bismuth
samples increased, as expected, from about 0.3 ppm for samples from the
treatment vessel to 66 ppm for samples from the bismuth receiver tank. Al-
though the increase was less than expected, it does indicate that the holdup
of thorium in the bismuth feed and receiver tanks was decreased significant-
ly. The uranium concentration in each of the bismuth samples was below the
limit of detection (<1 ppm) except in the case of the receiver tank sample,
for which a uranium concentration of 1.1 ppm was indicated. Thus, no
transfer of uranium to the bismuth phase occurred during the experiment.
Surprisingly, the reported uranium concentrations in the salt samples re-
moved from the column effluent varied by ijS% from the average value, which
was in excellent agreement with the indicated uranium concentrations of the
salt feed and catch tanks. Under almost the same conditions of flow during
run UTR-1, the variation in the indicated uranium concentration in samples
of the salt leaving the column was only about #10%. These results confirmed
that, in any future uranium transfer experiments, we must continue to rely
upon several pairs of samples at a given set of conditions in order to ob-

tain a reliable estimate of the fraction of the uranium that is transferred.

7.8 Summary of Hydrodynamic Data with Present Column

Hydrodynamic data obtained during countercurrent flow of salt and bis-

muth in the present column are summarized in Table 14. The values for the
Table 13. Concentrations of Uranium and Thorium in Salt and Bismuth Samples from Run UTR-5

 

Bismuth Samples

 

 

Thorium Uranium Uranium Conc. in
Conc. Conc. Salt Samples
Sample Source (ppm) (ppm) (ppm)

Treatment vessel 0.32, 10 <1 3030, 3400

Feed tanks 9, 1.2 <1l 3005
Flowing stream

1 3.k <1 2500
2 19. <1l 3000
3 8.3 <1 2650
L 5.1 <1 3240
5 0.83 <1 3340
6 26. <1 4310
T 35. <1 20ko
Ave. 1h. <1 3010

Receiver tanks 66 1.1 3030

 

e
86

 

 

 

Table 14. Summary of Hydrodynamic Data Obtained During Countercurrent Flow of Salt and Bismuth in a
0.82-in.-ID, 2L-in.-long Column Packed with 1/4-in. Molybdenum Raschig Rings
Apparent
Volumetric Flow Bismuth-~ Fraction Bismuth
Interval Rate (ml/min) to-Salt of Holdup
Run (min) Bismuth Salt Flow Ratio Flooding® (%) Comments
HR-9 18 80 85 0.9k 0.48 1k
HR-9 15 115 121 0.95 0.69 1k
HR-9 3 175 17T 0.99 1.03 14
HR-9 2 221 133 1.66 1.02 17
HR-Q b 221 107 2.07 0.93 ~-=
HR-9 15 150 150 1.00 0.88 Lo
HR-10 12 45 68 0.66 0.33 --
HR-10 6 175 68 2.57 0.68 -—
HR-10 7.5 27k T2 3.81 0.92 17
HR-10 3 Lho 20.3 21.7 0.95 32 Holdup increasing; flooding
HR-11 8 210 51 L.12 0.69 17
HR-11 7 330 50 6.60 0.93 17
HR-11 6 406 L7 8.6k 1.07 18-27 Holdup increasing; flooding
HR-11 3.5 228 100 2.28 0.92 --
UTR-1 90 100 100 1.0 0.59 17
UTR~1 25 130 130 1.0 0.76 17
UTR-2 4o 2h7 52 L.75 0.77 17
UTR-3 23.5 205 100 2.05 0.87 6
UTR-3 9 195 100 1.95 0.8k 6
UTR-3 200 160 1.25 1.05 13
UTR-3 5.2 195 220 0.89 1.22 10
UTR-4 46 140 149 0.94 0.85 16
UTR-4 13 117 118 0.99 0.69 16
UTR-k T.5 157 210 0.75 1.07 16
UTR-5 L2 115 123 0.93 0.70 17
UTR-5 8 1L6 161 0.91 0.90 Lo Holdup increasing; flooding
UTR-5 S 276 90 3.07 1.00 29 Holdup increasing; flooding
UTR-5 12 219 50 4,38 0.70 i Holdup increasing; flooding
®Fraction of flooding = [(vsaltl/2 + v .1/2)/\/ l/2]2. The value of V_ (superficial slip velocity) used was 392 ft/hr,

which corresponds to 676 ml/min in thé presen% column.
87

apparent bismuth holdup were calculated from measurements of the pressure
at the base of the column (salt inlet), and the column overpressure. During
most of the periods of steady-state operation, the values for the apparent
bismuth holdup were essentially constant (less than 5% variation during the
period) ; however, in a few periods the variation was considerably larger
(about 33%). A comparison will be made in a later report in this series

of the apparent bismuth holdup with previous measurements of the apparent
holdup of mercury during countercurrent flow of mercury and aqueous solu-
tions in packed columns. In run UTR-5, flooding was observed at 90% of

the calculated flooding condition. This low flooding value could be ex-
plained by the nonwetting condition of the salt following the hydrogen-HF
treatment. A comparison of the predicted flooding relation resulting from
work with mercury and aqueous solutionsl6 with the hydrodynamic data sum-
marized in Table 1lb4 is shown in Fig. 26. As can be observed, the agreement

between the predicted and measured values is quite satisfactory.

7.9 Examination of 304 Stainless Steel Corrosion Specimens
from Salt-Metal Treatment Vessel

The 304 stainless steel vessel, which contains the graphite crucible
that constitutes the salt-metal treatment vessel, is exposed to an H2—HF
mixture during the infrequent treatment of bismuth and salt to remove
oxides and to transfer dissolved metals from the bismuth to the salt. Cor-
rosion specimens of 304 stainless steel were removed from the treatment ves-
sel after a period of 10 months, which included a total of about 60 hr of
exposure to a 30 mole % HF-H2 mixture. After the specimens were cleaned
with a wire brush, a change of only +0.002 in. in thickness was observed.
The temperature of the specimens was about 550°C as compared with a nominal
vessel temperature of 600°C. On the basis of these data, corrosion of the
stainless steel treatment vessel to the present time is judged to be very

slight.

7.10 Maintenance of Equipment

Three transfer line failures occurred during this report period; all

were due to stresses produced by frozen bismuth. Failures in transfer lines
88

ORNL DWG 71-2840R4

SALT FLOW, ml/ min

 

 

 

 

50 100 200 300 400 500 600 700

20 ! I T T T T T 700
o~ PREDICTED FLOODING CURVE 41600
3 Vei’2 +v,"2 2198, (ft/hr)"?
© 4500
F =4
~ A
< sk A & FLOODED 1400
& e NON FLOODED
- . .
c 4300
Q ¢ A\\ <
o] ° £
~ ° % e ~
s T N\g e d200 E
3 10 . Ne -
< .A o g
(_) ° A ~
w « w
@ * 4100 T
w ° =
a
@ s

150 ©

r °r ¢ =
-
D
S
v
m

0 A 1 ]

o 5 10 IS 20

(SALT SUPERFICIAL VELOCITY,Vs ,ft/hr)"/2

Fig. 26. Summary of Flooding Data with Salt and Bismuth in a 0.82-
in.-diam, 2k-in.-long Column Packed with 1/L-in. Raschig Rings.
89

were noted at two points during heatup of the flow system prior to experi-
ment UTR-3. The first failure occurred on the outside of a bend in the

salt line connecting the salt jackleg to the bottom of the column; the
second failure was located in the bismuth exit line from the column. About
125 ml of bismuth and 500 ml of salt drained from the system as a result of
these failures. The salt leak occurred in a bend that had expanded 5 to 10%
in diameter at a location which is normally filled with bismuth during cool-
down periods. The wall thickness of tubing adjoining the site of failure

in the bismuth exit line was unchanged (0.080 in.), indicating that mass
transfer of iron was not a contributing factor as was the case in some of
the earlier failures. The external surfaces of the tubing examined were
only slightly air oxidized; the extent of oxidation appears to have been
decreased by the oxidation retardant paint used on the lines and by the
practice of heating the transfer lines to operating temperature only when
an experiment is planned. Approximately 15 freeze-thaw cycles had been ac-
cumulated in the failed areas. The implication of these observations was
that future operation should attempt to minimize the number of thermal
cycles at the expense of increasing the length of time during which the
lines are held at elevated temperature. Subsequently, attempts were made

to maintain the lines at temperatures of about 350 to L00°C,which are con-
siderably higher than the melting point of bismuth (271°C). In practice,
this procedure had a serious shortcoming in that freeze valves containing
bismuth would not seal; this, in turn, made the routine transfer of bismuth
and salt between vessels more difficult. We ultimately returned to the
original procedure of cooling the transfer lines to room temperature between

experiments.

During a routine cooldown of the flow system following experiment UTR-5,
a leak occurred in the bismuth drain line from the entrainment separator.
About 50 ml of bismuth was released through a longitudinal crack in the
1/2-in.-diam mild-steel tubing. The failure was similar in appearance to

the failures discussed earlier.
90

8. PREVENTION OF AXTAL DISPERSION IN PACKED COLUMNS

J. S. Watson L. E. McNeese

Packed columns are being considered for use in countercurrently con-
tacting molten salt and bismuth streams in MSBR fuel processing systems.
Previously, we made measurements of axial dispersion for various packing
materials in columns during the countercurrent flow of mercury and agueous
solutions.18 We showed that axial dispersion can significantly reduce the
performance of this type of contactor under some operating conditions of

19

interest. As part of our contactor development program, we are cur-
rently evaluating column modifications that will reduce the effect of

axial dispersion to an acceptable level. The proposed modifications con-
sist of devices to be inserted at points along the column to reduce dis-
persion at these points. If the devices are separated by a column length
equivalent to one theoretical stage (an extent of separation that can be
achieved even if the liquids between the devices are completely mixed),

the stage efficiency of the column segment will be greater than T75% if 15%
or less of the salt flowing through the segment is recycled to the previous
segment. We have previously tested devices consisting of inverted bubble
caps containing a number of small-diameter holes through which the salt flows
at an increased velocity.go It was found that the extent of axial dispersion
could be reduced considerably and, in principle, to very low values by the
use of sufficiently small holes. However, the column capacity, or through-
put at flooding, was reduced by about a factor of 6. During the current
report period, we devised and tested an improved design of axial dispersion
preventer. In this design, shown schematically in Fig. 27, the small-
diameter holes in the upper portion of the inverted bubble cap have been
replaced with a single 3/8-in.-0D tube that is sealed at the top but which
has four 1/L4-in.-diam holes near the top of the tube. This design allows
the metal phase to rise to the bottom of the salt exit holes in the vertical
tube, thereby providing a sufficiently high liquid metal head for forcing
the dispersed phase through the axial dispersion preventer at a high through-

put. Two axial dispersion preventer designs were tested during the counter-
91

ORNL DWG 71-13973

SALT SALT
FLOW FLOW

 

 

o
‘k\~_
y A4

 

 

 

 

  

60 ~—PACKED
COLUMN

 

 

Fig. 27. Schematic Diagram of an Improved Axial Dispersion Preventer.
92

current flow of mercury and water in a 2-in.-diam column packed with 3/8-
in. Raschig rings. In the first design, the water flowed upward through

a 1/4-in.-diam, 1/4-in.-long tube. In the second design, the tube length
was increased to 1/2 in. Data obtained with the second axial dispersion
preventer design is summarized in Fig. 28. The mercury superficial
velocity was varied from about 40 to 150 ft/hr, and the water superficial
velocity was varied from about 2.7 to 14 ft/hr. A marked effect of the
metal flow rate on the extent of axial dispersion was noted; that is, higher
metal superficial velocities led to a greater extent of axial dispersion.
As expected, no difference between the two devices was noted with regard
to axial dispersion. The major effect of using a longer tube is that of
obtaining an increase in the metal throughput. It is believed that this
design would allow for salt and bismuth throughputs that can be as high as
the column throughputs at flooding. Although the extent of axial disper-
sion is still significant with either of the devices, it is probably suf-
ficiently low for most applications of interest. The extent of axial dis-
persion provided with the present device is intermediate between that pro-
vided by bubble caps having one 1/4-in.-diam hole and bubble caps having
four 1/b-in.-diam holes through which the water is allowed to flow. The
performance of the present axial dispersion preventer could probably be
increased by using only a single hole in the side of the vertical tube

instead of the four holes as in the current design.

9. ELECTROLYTIC REDUCTION OF LiCl USING A BISMUTH
CATHODE AND A GRAPHITE ANODE

J. R. Hightower, Jr. C. P. Tung
L. E. McNeese

The metal transfer process for removing rare earths from an MSBR re-
quires the use of Li-Bi solutions for extracting the rare earths from LiCl.
One method for providing these Li-Bi solutions would consist of electro-
lytically reducing LiCl that is produced in the processing system. We have
carried out an experiment in order to determine the general operating char-
acteristics of an electrolytic cell having a bismuth cathode and a graphite

anode. The objectives of the experiment were to determine the maximum anode
93

ORNL DWG. 71-1359I
l I l ] | [

655 ml/min

0.9

 

0.81—

  
   

MERCURY FLOW RATE =

o7l 380 mi/min 1518 ml/min

1020 ml/min

  

o
N
l

BACKMIXING
O
o
l

FRACTION

Ol T

 

 

0 | | | | 1 | |

0 20 40 60 80 100 120 140
WATER RATE , ml/ min

 

Fig. 28. Variation of Fraction of Backmixing with Water and Mercury
Flow Rate for an Axial Dispersion Preventer Having a Single 1/4-in.-ID,
1/2-in.~long Tube Through Which the Water Flowed.
ok

current density achievable, the current efficiency, the extent of attack
on the graphite anode, and operational difficulties associated with forma-

tion of Li-Bi solutions at the cathode.

9.1 Equipment and Materials Used -

The experimental equipment consisted of a L-in.-diam, 18-in.-long
quartz cell vessel; a 3.5-in.-diam, 3.5-in.-high molybdenum cup contain-
ing the bismuth cathode; a l-in.-diam, 7-in.-long graphite anode; a L-in.-
diam, 10-in.-high bismuth purification vessel; and a L4-in.-diam, 15-in.-
high LiCl purification vessel. The assembled electrolytic cell is shown

in Fig. 29.

Provision was made for measuring the volume of chlorine produced at
the anode by allowing the chlorine to displace argon from a 25-ft-long
section of 3/b4-in.-diam copper tubing. A wet-test meter was provided for

measuring the volume of the displaced argon.

Seven kilograms of bismuth was charged to the bismuth treatment vessel,
where it was sparged with hydrogen at 650°C until the water concentration
in the off-gas was less than 1 ppm. The bismuth was then transferred through
a porous molybdenum filter into the electrolytic cell. Eleven hundred grams
of oven-dried LiCl was charged to the salt treatment vessel. The salt was
purified by contact with 2 kg of purified bismuth to which 93 g of thorium
metal had been added. The LiCl was contacted with the thorium-bismuth
solution for a period of about 100 hr at 650°C; then it was transferred
through a porous molybdenum filter into the cell vessel. The porous molyb-
denum filters were made from 0.63-in.-diam, 1/8-in.-thick porous molybdenum
disks having a mean pore size of 30 to 40 u. The filtration rate at 650°C
for the LiCl was 0.8 cm3/sec with a pressure drop of 6.2 psi, and the fil-

tration rate for the bismuth was 0.4 cm3/sec with a pressure drop of 2 psi.

9.2 Operating Conditions and Results

The cell was operated at 6T70°C with the following anode current densi-

ties (based on the projected area of the bottom of the anode, 5.06 cm2):
95

PHOTO 99335

 

Fig. 29. Assembled Electrolytic Cell Vessel Used for LiCl Reduction.
96

4,37 A/cm2 for 7 min, 6.7 A/cm2 for 6 min, and 8.6 A/cm2 for 6 min. There
appeared to be no limiting‘anode current density in this operating range.
Disengagement of the chlorine gas produced at the anode (which had been
rounded slightly to promote gas disengagement) proceeded smoothly and with-
out difficulty. With the initiation of current flow through the cell, the
LiCl became red in color; the color of the salt grew darker as the opera-

tion continued until, finally, the salt became opaque.

Measurement of the chlorine evolution rate (and hence the current ef-
ficiency) was not possible because of reaction of the chlorine with iron

components in the upper part of the cell vessel.

9.3 Postoperational Examination of Equipment

After the cell had cooled to room temperature, examination showed that
the top flange and the anode support, both of which were made of mild steel,
were severely corroded. A mixture of yellow and white powders (FeCl2 and
LiCl containing 1.3 wt % vismuth) covered the flange and nearly filled a
3-ft-long section of the 1/4-in.-diam copper off-gas line. A thin metallic
film, which had covered the LiCl surface, was removed intact when the anode
was raised at the end of the experiment. The metallic portion of the film
consisted of iron and about 3 wt % bismuth. The LiCl appeared to contain a
large amount of suspended black material, which was found to be iron. The
bismuth concentration in the LiCl was relatively high (about 0.63 wt %).
The molybdenum cup, which contained the bismuth cathode, appeared to be

unattacked.

It is believed that the attack of iron components in the cell by gas-
eous chlorine was the only detrimental action which occurred during opera-
tion of the cell. The FeCl_ corrosion product was partially transferred to

3
the LiCl, where it was reduced to metallic iron.

The material that caused the salt to turn red during the experiment

3Bi, which dissolved in the LiCl. It is known that Li3Bi

dissolves to an appreciable extent in an LiCl-LiF eutectic which is in

was probably Li

contact with solid Li.Bi or a bismuth solution that is saturated with

3
971

Li3Bi,21 and that a dark red solution is formed. The cathode probably
polarized rapidly at the cathode current densities used in this experi-
ment, resulting in bismuth saturated with LiBBi at the LiCl-bismuth inter-
face. Polarization of the cathode could be prevented by any of several
methods, including limiting the cathode current density or increasing the
extent of mixing of the bismuth phase near the LiCl-bismuth interface.
The equilibrium concentrations of lithium and bismuth in LiCl are reported
22

to be significantly reduced when the bismuth is not saturated with LiBBi.

Although this experiment pointed out an unexpected complication, it
confirms our expectation that electrolytic reduction of LiCl using a bis-
muth cathode and a graphite anode should proceed readily and that little,

if any, attack should occur on the graphite anode.

10. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS

"R. B. Lindauer L. E. McNeese

We have previously described equipment for study of the purification
of salt by continuous methods.23 Initial work with this system was directed
at measurement of the flooding rates in a 1.25-in.-diam, T-ft-long column
packed with 1/k-in. nickel Raschig rings. Flooding data were obtained dur-
ing the countercurrent flow of molten salt (66-3L4 mole % LiF—BeFQ) and hy-
drogen or argon.gLL The objective of the present work is to study the con-
tinuous reduction of iron fluoride in molten salt by countercurrent contact
of the salt with hydrogen in a packed column. During this report period, a
sufficient quantity of FeF, (34.5 g) was added to the salt (66-3L4 mole %
LiF—Ber) to increase the iron concentration from 20 ppm to 425 ppm. Two
iron fluoride reduction runs (R-1 and -2) were carried out, and several

equipment modifications were made.

10.1 Iron Fluoride Reduction Runs R-1 and R-2

After FeF2 had been added to the 1lh-liter salt charge, two iron fluoride

reduction runs were carried out at a column temperature of TO0°C. During the
98

first run, a salt flow rate of 100 cmB/min and a hydrogen flow rate of
about 20 liters/min were used. A low hydrogen supply pressure resulted

in a fluctuating hydrogen flow rate throughout most of the run. The
variations in the gas flow rate caused the pressure at the top of the col-
umn to fluctuate; in turn, this fluctuation resulted in an irregular salt
feed rate to the column. During the run, the iron fluoride concentration
in the salt was reduced from an inlet value of 425 ppm to 307 ppm. In the
second run (R-2), a salt flow rate of 100 cm3/min and a hydrogen flow rate
of 13.5 liters/min were used. Operation of the column at the desired con-
ditions was difficult in this run, although performance of the system had
been satisfactory during previous flooding runs in which the gas flow rates
were as high as about 25 liters/min. After only 4.0 liters of the 1h-liter
salt batch had been fed through the column, the pressure at the base of the
column resulting from the pressure drop across the column and column off-
gas line was sufficiently high to force the gas-salt interface out of the
salt loop below the column. At this point, the flow of hydrogen was ter-
minated and the remaining salt was transferred to the salt receiver vessel
through the static column of salt. Data for the two runs are summarized

in Table 15.

Table 15. Summary of Data from Iron Fluoride Reduction Runs R-1 and R-2

 

—tr

 

 

Hydrogen Salt Length Analysis of Fraction of
Flow Flow of Filtered Samples® Salt
Run Rate Rate Run (ppm of iron) Contacted
No. (std liters/min) (em3/min) (min) Feed Product with H,
R-1 20.0 100 87 L5 307 0.685
R-2 13.5 100 L1 307 228 0.276

 

aSamples were taken from the combined salt volumes in the salt receiver
vessel.
99

10.2 Mathematical Analysis of the Rate of Iron Fluoride Reduction;
Calculated Mass Transfer Coefficients for Runs R-1l and R-2

The countercurrent contact, in a packed column, of hydrogen with salt
containing FeF2 results in the reaction:

FeF + Hg(g) =Fery+ 2HF(8) . (35)

2(d)

A material balance on FeF2 in the salt contained in a differential length

of the column yields the relation:
-L dx = rA dh, (36)

where
L = salt flow rate, moles/sec,
x = concentration of FeF2 in bulk salt, mole fraction,

r = rate of FeF2 reduction per unit column volume, moles Fng/sec-cmB,
A = cross-sectional area of column, cm2,

h = column height, cm.

The reduction of FeF,. by reaction with hydrogen can be affected by a number

of factors that may Sontribute to setting the overall rate of reaction. one
of the objectives of this work is to determine the rate-controlling steps in-
volved in the reduction of FeF2 and evaluation of the associated rate con-
stants. Two limiting cases for the rate of reduction of FeF2 are of in-
terest: (1) the case in which the rate of reduction is limited by the rate
of transfer of FeF2 to the gas-salt interface from the bulk salt, and (2)

the case in which the rate of reduction is limited by the rate of reaction
between FeF2 and H2 at the gas-salt interface.

10.2.1 Case for Which the Rate of FeFo Reduction Is Limited by the Rate of
Transfer of FeFo to the Gas-Salt Interface

For the case in which the rate of transfer of FeF2 to the gas-salt
interface is rate limiting, the rate of reduction will be given by the

expression:
100

r=%kalx -x), (37)

where
kz = mass transfer coefficient for transfer of FeF2 from the bulk salt
to the salt-gas interface, moles/se0°cm2,

a = gas-salt interfacial area per unit column volume, cmg/cm3,

*

x = concentration of Fer in salt that would be in equilibrium with
the H2—HF mixture adjacent to the salt being considered, mole
fraction.

Combining Egs. (36) and (37) yields the relation:

L dx = —kzaA(x - x*) dh. (38)

*
If the value of x is small in comparison with the value of x, Eq. (38) can

be integrated over the column length to yield the following relation:

L 5
kza = N 1ln z (39)
o
where
H = column height, cm,
X, = concentration of FeF2 in salt fed to column, mole fraction,
x = concentration of FeF, in salt leaving column, mole fraction.

o 2

10.2.2 Case for Which the Rate of FeFo Reduction Is Limited by the Rate
of Reaction of FeFo with Hpo at the Gas-Salt Interface

For the case in which the rate of reaction between FeF2 and H2 at the
gas-salt interface is rate limiting, the rate of reduction will be given by

the expression:

r = kSKfipHQX’ (Lo)
101

where
L reaction rate constant, cm3/sec,
KH = Henry's law constant for hydrogen in salt, moles/cm3‘atm,
Py = partial pressure of hydrogen in gas, atm.
2

Combination of Egs. (36) and (40) yields the relation:
L dx = —kSKHAszx dh. (L1)

If the variation in the value of Py throughout the column is small, this

relation can be integrated over thegcolumn length to yield the expression:

L Xin
kK = ——————— 1pn . (42)

S KH HA pH2 Xout

 

10.2.3 Calculated Reaction Rate Constants for Runs R-1l and R-2

Values for the mass transfer coefficient and the reaction rate constant,
as defined by Egs. (39) and (42) for the two limiting cases of interest, were
calculated. These values and associated information are summarized in Table
16. For the case in which the reduction rate is controlled by mass transfer,

it is seen that the FeF, concentration in salt in equilibrium with gas raving

a composition equal to ihat of gas leaving the column is negligible in com-
parison with the inlet or outlet FeF2 concentrations in the salt. The re-
sulting values for the overall rate constant (kza) are 0.028 and 0.06L
moles/sec'cm3 and are in satisfactory agreement with each other. Since

the partial pressure of hydrogen was the same for both runs, the reaction
rate constant, ks, is proportional to kza. In future runs, the controlling
mechanism will be determined by varying the inlet hydrogen partial pressure.

10.3 Equipment Modifications and Maintenance

Several difficulties were encountered during operation of the system

throughout the current report period. These difficulties were as follows:
Table 16.

Calculated Values for the Mass Transfer Coefficient
and the Reaction Rate Constant for Runs R-1 and R-2

 

Mass Transfer Controlled

Reaction Rate Controlled

 

 

 

 

 

P
FelF~ Concentrations * k HE
: | o x 5 (atm)
o fre i tm
(TOIL raction) x iO (moles/seq-omg) (mole fra tion) (cm3/sec) a
Run To. % o x 105 x 10 x 10-4 Inlet Exit Eycilibrium
R-1 .13 3.70 2.l 0.033% 2.k 1.0 0.9987 1.98873
He D 3,70 - 0.75 5.4 0.198% 5.1 1.0 0.9967 C.9900

 

a . . . -
Calculuted trom composition of exit gas.

Calculated from Fel' | concentration in salt leaving the column.

O

c0T
103

(1) an undesirably high pressure drop in the column off-gas system,
which caused the salt-gas interface in the column to be forced

out of the seal loop below the column;

(2) restrictions in the salt filter inlet and outlet vent lines,
which probably contributed to the erratic operation of the

system during the two iron fluoride reduction runs;

(3) failure of a weld in contact with salt in the sampler in the

salt effluent line from the column;

(4) failure of the liquid level dip line above the salt receiver

tank; and

(5) deposition of BeO (in the column), which reduced the column

throughput.

10.3.1 Restriction in the Off-Gas Line from the Column

In the initial system design, salt leaving the bottom of the column
flowed through a 30-in.-deep loop that formed a gas seal and prevented gas
from flowing out with the salt. The salt used initially in the system had
a specific gravity of 2, and a pressure of 60 in. HEO at the bottom of the
column was sufficient to force salt out of the seal loop. The total pres-
sure at the bottom of the column is the sum of (1) the pressure at the top
of the column, and (2) the pressure drop across the column. The pressure
at the top of the column is usually controlled by a throttling valve; how-
ever, an abnormally high value will result if there is a restrictioh in the
off-gas line leaving the column. The pressure drop across the column in-
creases as the salt and gas flow rates are increased, rising sharply as
the flooding point is approached. During run R-1, the pressure at the base
of the column remained sufficiently low that salt was not forced from the
seal loop. However, during run R-2, the pressure at the top of the column
was higher; and, after about half of the salt (6 liters) had been fed
through the column, the column pressure drop increased to about 35 in. H20.

At this point, the pressure at the base of the column was sufficient to
104

force salt out of the seal and the run was terminated. Examination of the
1/2-in.-0D off-gas line leaving the column showed that the line was almost
completely restricted with salt. An 8-in.-long, b4-in.-diam heated vessel
packed with a 6-in.-long section of nickel wool (see Fig. 30) was installed
in the off-gas line in order to remove salt from the gas stream leaving the
column. The vessel was positioned in such a manner that salt would drain
from the packing and return to the top of the column through a gas seal

loop.

10.3.2 Restrictions in Vent Lines on Salt Filter

 

Following reduction runs R-1 and -2, it was found that the vent lines
on the inlet and outlet sides of the salt filter had become plugged with
salt. The restriction in the inlet vent line probably resulted from -en-
trainment of salt with the gas that was forced out of the seal loop below
the column during run R-2. The salt in the outlet vent line on the filter
was the result of a cold section in the line (between the filter and the
salt receiver tank) that caused salt to fill the filter vessel. The

restricted portions of the vent lines were removed and replaced.

10.3.3 Failure of a Weld on the Flowing-Stream Salt Sampler

During flooding test 11 (which followed reduction runs R-1 and -2),
the pressure drop across the column suddenly increased. It was found
that failure of a submerged weld in the flowing stream salt sampler had
allowed salt to contact a Calrod heater on the salt exit line between
the sampler and the receiver, thereby causing the heater to fail. Failure
of the heater and, in turn, a decrease in the temperature of the exit line
(to below the salt liquidus temperature) caused an accumulation of salt in
the column and resulted in the observed increase in the column pressure
drop. To circumvent these difficulties, the lower section of the sampler

was redesigned to eliminate welds below the normal salt level.
105

ORNL DWG 72-1384

DIFFERENTIAL PRESSURE
INSTRUMENT LINES

  
  

GAS OUTLET
FROM COLUMN

 

 

 

GAS
T0

VENT ?
SALT

INLET

 

 

 

LIQUID SEAL AND
RETURN TO COLUMN

PACKED
COLUMN

Fig. 30. Salt Entrainment Separator.
106

10.3.4 PFailure of a Liquid Level Dip Line on the Salt Receiver Tank

One of the dip lines in the salt receiver tank was found to be re-
stricted by salt in the vicinity of the top of the vessel. A drill was
used to remove the salt from the line; however, we found that attempts
to transfer salt from the receiver to the feed tank led to a failure in
the liquid level dip line at the point where the salt plug had been re-
moved. We believe that the reason for the failure was improper centering

of the drill and a decrease in the wall thickness of the tubing.

10.3.5 Deposition of BeO in Column

 

After the first two reduction runs (R-1 and -2) had been completed,
we found that the pressure drop across the column with argon flow only (flow
rate, 5 liters/min) had increased from 5.7 in. HQO to 12.0 in. H20. Since
only 7 g of iron had been reduced in the system, it was believed that the
restriction was caused by an accumulation of relatively insoluble BeO,
which may have been formed in the system as the result of low concentra-
tions of impurities in the argon or hydrogen or because of inleakage of
air into the system. Up to this time, our usual procedure has been to
maintain a low argon flow rate through the system, which is held at atmo-
spheric pressure during periods of nonoperation. We have now adopted the
procedure of maintaining a positive pressure of at least 1 psig in the sys-
tem during such periods in order to minimize the possibility of air inleak-
age.

L

11. ELECTROLYTIC OXIDATION OF Pa * TO Pa5+

IN MSBR FUEL SALT
J. S. Watson L. E. McNeese

25

Baes, Bamberger, and Ross have recently proposed a protactinium iso-

lation method based on the selective precipitation of Pa The process

O..
54

+
requires oxidation of the Pah in MSBR fuel salt to Pa (and most of the

+ +
U3 to Uh ) prior to the subsequent oxide precipitation step. A process
107

step was proposed in which oxidation of the protactinium and uranium, as

well as subsequent reduction of about 1% of the Ul#+ to U3+

,» would be car-
ried out electrolytically. The estimated current densities, the size of

the proposed electrolytic oxidation reduction system, and the feasibility
of such a reduction step will be considered in the remainder of this

section.

The proposed system contains three electrodes: a cathode, an anode,
and a third low-current electrode that functions sometimes as a cathode
and at other times as an anode. The current between the first two elec-
trodes is expected to be limited by the rate of diffusion of U3+ and PaLL+
to fhe anode surface, where a low current density is estimated because of
the low concentrations of these materials in the salt. The total concen-
tration of materials to be reduced at the cathode (primarily Uh+) is higher
by about two orders of magnitude than the total concentration of materials

to be oxidized at the anode, and a current density limitation is not ex-

pected at the cathode.

11.1 Estimated Anode Current Density for the Case of No
Interaction Between Protactinium and Uranium

+

The maximum rate at which Pah can transfer to the anode surface is

L+
t

the rate which would be obtained by assuming the concentration of Pa o

L+
be negligible at the anode surface. In this case, the Pa  transfer rate

can be estimated by the following expression:

DCeA
RPaLH. = ——(5 s ().,LB)
where
RPah+ = rate at which Pah+ transfers to the anode surface and is

5+

oxidized to Pa” , moles/sec,

D = diffusivity of Pau+ in MSBR fuel salt under conditions present
in the boundary layer at the anode surface, cm2/sec,

C = effective concentration of Pah+ in the bulk salt adjacent to

the anode surface, moles/cm3,
108

=3
It

anode surface area, cm2,

&
n

thickness of boundary layer adjacent to anode surface, cm.

The effect of the electrical potential gradient within the boundary layer
has been neglected since the potential gradient will be dissipated by the
high concentration of nonreacting ions in the layer that will act as a
supporting electrolyte. The required anode surface area is given, then,

by rearranging Eq. (43) to obtain the following expression:

§ R, L+
A= —2 (b1)

DCe
It should be noted that the rate at which Pa)-L+ transfers to the anode sur-
face is approximately equal to the rate at which protactinium is produced
under desirable operating conditions. It will be useful, subsequently, to
also note the following relations for the fraction of Pah+ that is oxidized

5+

+ +
to Pa and for the rate at which Pah is oxidized to Pa5 in a system op-

erating at steady state:

C L+
£ =1 - agngfflgi , (L45)
Pa"*tin
where
. + . . o o+
f = fraction of Pa that is oxidized to Pa” ,
+
CPah+Out = concentration of Pal‘L in bulk salt leaving anode surface,
moles/cm3,
+
CPah+in = concentration of Pah in salt entering oxidizer from reactor,
moles/cm3;
and
RPah+ = fFCPal“'in’ (46)
where

F = rate at which salt is processed for protactinium removal, cm3/sec.
109

11.1.1 Case for No Axial Mixing Along Anode Surface

If there is no axial mixing along the anode surface, the effective Pa
concentration will be the logarithmic mean average of the inlet and outlet

+

Pah concentrations, which is given by the relation:

_ Coglitin = Cpalitout
Pa**in
Cpal+out

 

1n

Substitution of Eg. (47) into Eg. (kL) yields the following relation for

the required anode surface area:

_OBpgh+ 1
FDC, bty L - T

 

A

11.1.2 Case for Complete Mixing of Salt Along Anode Surface

If the salt along the anode surface is assumed to be perfectly mixed,
+
the effective Pau)'L concentration will be equal to the outletjPah+ concen-

tration, which can be expressed by use of Eq. (45) in the following manner:

(49)

Ce = (1 -1) Cpali+in:

Substitution of Eq. (41) into Bq. (kL) yields the following relation for

the anode surface area in this case:

ORpgli+ 1

= DC 1 - f ° (50)

 

A
Pah+in

The processing cycle time is defined by the following relation:

T, = V/F, (51)
110

where

T processing cycle time, sec,

b
v

volume of fuel salt in reactor, cm3.

The protactinium removal time is then related to the processing cycle time

by the following expression:
T, = Tp/f, (52)

where

g T Pa removal time, sec.

Substitution of Egs. (L46), (51), and (52) into Eq. (50) yields the following
expression for the required anode surface area for the case in which perfect

mixing along the anode surface is assumed:

A = m———, (53)

11.2 Estimated Anode Current Density for the Case of
Equilibrium Between Protactinium and Uranium

Tn the cases discussed thus far, no consideration has been given to
the simultaneous oxidation of U3+ to ULL+ during the oxidation of Pa * or
to the effect of the oxidation of uranium on the rate at which Pah+ is ox-
idized. If we assume that perfect mixing occurs along the anode surface
and that the uranium and protactinium species are at chemical equilibrium,
the concentrations of the latter species are related by the following ex-

pression:

’

3
(CPa5+ Cy3+
\ CPa)-H' \CU)-#-P

 

 

= Q, (54)

where

CPa5+ = concentration of Pa in bulk salt, moles/cm3,
111

. L+
CPah+ = concentration of Pa  in bulk salt, moles/cm3,
CU3+ L+ 3
CUu+ = concentration of U in bulk salt, moles/cm~,

+
concentration of U3 in bulk salt, moles/cmB,

Q = equilibrium constant.

5+

The following equation shows the relationship of the Pa” and Pah+ concen-

L+

trations to the fraction of the Pa that is oxidized:

Pad* _ 1
CPa)_l,+ 1 -7

 

: (55)

+
The concentration of ULL will remain approximately constant during oxidation
L+ + +

of both Pa and U3 . Thus, the concentration of U3 will be given by the
following relation, which is obtained by substituting Eq.. (55) into Egq.
(5k4):

Cy3+ = Qb+ (1 - r)/f. (56)

: L+ 3+ s o :

The total rate at which Pa and U~ are oxidized is given by the following

expression:
R = FCo lay T+ F(C34; = Cy3+)s (57)

where

+ +
R = total rate at which Pah and U3 are oxidized, moles/sec.

3+

+
If it is assumed that the concentrations of Pah and U are negligible at
the anode surface in comparison with their concentrations in the bulk sait,
the total rate at which these materials are oxidized will also be given by

the following expression:

A
_ t
R = E—-(DPau+ CPah+ + DU3+ CU3+)’ (58)
where
. o L+ 3+
At = total anode surface area required for oxidation of Pa and U~ ,

2
cm .
112

If Egs. (57) and (58) are equated and if it is assumed that the diffu-
s L+ +
sivities of Pa and U3 in molten salt are equal, one obtains the fol-

lowing relation for the required anode surface area:

+ -
_F¢ CPah"'in £ CU3+in CU3+out
A% D C o+ C : (59)
Pattin ~ “U3%out

 

where

+ +
D = diffusivity of Pal‘L or U3 in salt.

Substitution of Egs. (57) and (58) yields the following expression for the

required anode surface area:

 

 

 

 

5 -
RPalH' Pah+in CPah”‘in £ . (60)
t DC h.*_ . C )_H. ’
Pa=""Tin U 1 -7
1 - +Q z 7
Pah+in

An expression showing the relative anode surface area required for the
case in which equilibrium is assumed between the protactinium and uranium
species and the case in which no interaction between uranium and protactinium

is assumed is given by dividing Eq. (60) by Eq. (50), which yields:

 

 

- “y3tin 0 Crb+ (1 - f)
At _ CPah+in CPah+in t .
L - . (61)
A CULH-
f + Q,Er————-—
Pah+in

The rate at which protactinium is produced in a 1000-MW(e) MSBR is about 11.1
moles/day, and the concentration of Uh+ in the salt is about 0.0033 mole
fraction. If removal of protactinium in the reactor by neutron capture énd
radiocactive decay of 233Pa is neglected, the relative concentrations of ULL+
and Pa4+ in salt entering the protactinium isolation system will be given by

the following relation:
113

T = T3/7p, (62)

where

T, = Pa removal time, days.

R

Since the ratio of the U3+ concentration to the ULL+ concentration in MSER
fuel salt is about 0.01l, the ratio of the U3+ and Pah+ concentrations in
salt entering the protactinium isolation system is given by the following
relation:

C i3+
8 = 73/, (63)

paltin

Substitution of Egs. (62) and (63) into Eq. (61) yields the following ex-

pression for the relative anode surface area:

A i 1+ Tp[7.h3 - T43Q (1 - rp/rR>]
1 + Th3Q rp

 

t
A

11.3 Calculated Results and Discussion

Processing cycle times of interest (3 to 10 days) correspond to salt
flow rates of about 1 to 3 gpm for a 1000-MW(e) MSBR. The diffusion coef-
ficients for PaL)++ and U3+ in molten salt are estimated to be 10_5 cm2/sec.
Since the salt flow rate is relatively low, recirculation of salt through
the anode compartment will be required to achieve Reynolds numbers suffi-
ciently high to obtain film thicknesses in the desired range of 0.01 to
0.05 ecm. For example, a salt flow rate of 1 gpm through an anode compart-
ment that is 2 in. wide and 78.7 in. (2 m) long corresponds to a Reynolds
number of 16. A film thickness of about 0.03 cm would result from a
Reynolds number of L4000 for the same anode configuration. It is thus ap-

parent that sufficient recirculation of salt will be required in the anode

compartment to give salt which will be essentially perfectly mixed. In
11k

this case, Eq. (50) must be used for estimating the anode surface area
rather than Eq. (48) which applies to the case in which no mixing along

the anode surface 1s assumed.

Values for the required anode surface area for several combinations
of protactinium removal time and processing cycle time are shown in Table
17. The minimum required anode surface area ranges from 16.7 to 167 n° as
the protactinium removal time is decreased from 10 days to 1 day. An anode
area of about 83.5 m2 is required to obtain a 3-day protactinium removal
time with a processing cycle time of 1 day, and an area of 23.9 m2 is re-
quired to obtain a protactinium removal time of 10 days with a processing

cycle time of 3 days.

Table 17. Estimated Values for the Required Anode Surface Area
for the Case of No Interaction Between
Protactinium and Uranium Species

Reactor salt volume = 1700 ft3

Salt film thicknesses = 0.03 cm

 

 

Pa Removal Time Processing Cycle Time Required Anode Area
(days) (days) (m?)
1 0 167
0 55.7
10 0 16.7
3 1 83.5
10 3 23.9

 

Values for the relative anode surface area required for the case in
which equilibrium between the protactinium and uranium species is assumed
are shown in Table 18 for two processing cycle times, four protactinium
removal times, and two equilibrium constznts that represent extremes of

2

the estimated values for this constant. It is seen that the required

area depends strongly on the value of the equilibrium constant and ranges
115

from about 20% of the area required for the case in which no interaction
of protactinium and uranium species is assumed to about 11 times the ares

required for that case.

Table 18. Relative Anode Surface Areas Required for the Case of
Equilibrium Between Protactinium and Uranium Species

 

 

 

Protactinium Relative Anode Ares,
Removal Time Processing Cycle Time A+ /A

(days) (days) Q = 0.0005 @ = 0.01

3 1 5.97 0.h12

6 1 5.92 0.266

10 1 5.90 0.207

3 10.9 0.761

3 10.8 0.521

10 3 10.6 0.330

 

It is concluded that electrolytic oxidation of Pa.Ll+ prior to precipi-
tation of Pa205 is not attractive because of tii largeSinode surface areas
required. We believe that oxidation of the Pa  to Pa” by the use of HF-
HQO—H2 mixtures (as suggested by Baese7) is more promising and recommend

that additional data be obtained for evaluation of this step.
1.

10.

11.

12.

l3o

1k,

15.

16.

17.

116

12. REFERENCES
L. E. McNeese, MSR Program Semiann. Progr. Rept. Feb. 28, 1970, ORNL-
4sh8, pp. 282-88.

M. J. Bell and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug. 31,
1970, ORNL-L4622, pp. 199-202.

MSR Program Semiann. Progr. Rept. Feb. 28, 1970, ORNL-4548, pp. 289-92.
C. E. Bamberger and C. F. Baes, Jr., J. Nucl. Mater. 35, 177 (1970).

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257, pp. 16-29.

 

M. 8. Bautista and L. E. McNeese, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 4, ORNL-TM-3139, pp. 38-83.

 

 

L. BE. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 5, ORNL-TM-3140, pp. 22-30.

 

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 6, ORNL-TM-31L41, pp. 13-23.

 

Y. Ohki and H. Inoue, Chem. Eng. Sci 25, 1 (1970).

R. M. Davies and G. I. Taylor, Proc. Roy. Soc. (London), Ser. A 200,
375 (1950).

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257, pp. 29-L6.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder

 

Reactor Processing No. 1, ORNL-TM-3053, pp. 1-1k.

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 6, ORNL-TM-31L41, pp. 73-79.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257, pp. 52-56.

 

 

J. 5. Watson, ORNL, personal communication, September 1970.

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 5, ORNL-TM-3140, pp. 102-1T.

 

 

T. Johnson et al., Chem. Eng. Div. Semiann. Rept. January-June 1965,
ANL-T055, p. Lk,
18.

19.

20.

21.

22.

23-

ol

26.

270

117

REFERENCES (continued)

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder

 

Reactor Processing No. 7, ORNL-TM-3257, pp. 58-89.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder

 

Reactor Processing No. L4, ORNL-TM-3139, pp. 27-38.

 

L. E. McNeése, Engineering Development Studies for Molten-Salt Breeder

 

Reactor Processing No. 5, ORNL-TM-3140, pp. 97-102.

 

M. S. Foster et al., J. Phys. Chem. 68, 980 (196L).

M. S. Foster, "Laboratory Studies of Intermetallic Cells," p. 1kk in
Regenerative EMF Cells, American Chemical Soclety, Washington, D.C.,

1967,

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 6, ORNL-TM-31hkl, pp. 59-T72.

 

L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder
Reactor Processing No. 7, ORNL-TM-3257, pp. L46-52.

 

C. F. Baes, Jr., C. E. Bamberger, and R. G. Ross, personal communica-
tion, May 8, 1970.

C. F. Baes, Jr., personal communication to J. 5. Watson, May 1970.

C. F. Baes, Jr., personal communication to L. E. Mclleese, June 8, 1970.
=
OO0 O~ o\ W o+

o
o

13.

‘__S
I~

15.
16.
17.
18.
19.
20.
2]1.
22,
23.
ol ,
25.
26.
27
28.
29.
30.
31.
32,
33,
34,
35.

85.
86.

87.
88.

89.
90.

119

INTERNAL DISTRIBUTION

. Anderson
Baes

. Bauman

. Beall

. Bell

. Bennett

S. Bettis
Blanco

. F. Blankenship
. Bohlmann

. Boyd

. Briggs
Brooksbank
. Brown
Carter
Cochran, Jr.
Culler
Distefano
Eatherly

. Ferguson
Ferris

Frye

Grimes
Grindell
Haas
Hannaford
Haubenreich

oA

=

. . . » . . . * .
- - . . » . . » . . -

-
.

Kee

. Keyes
Lindauer
Lundin
MacPherson
Ma.cPherson
McCoy

. . .
. . .

HOEETSQUUEWYE GRS S R TQEE OERRNE QS

HEHOHODGOGO=E Y rprQd DR HEE®

Hightower, Jr.

70.
T1.
72.
73,
Th.
75-T6.
77-78.
79-81.
82.
83.
8h.

. McNeese

Moulton

Nichols
Nicholson
Pashley (K-25)
Perry

Rosenthal

Ryon

Schaffer, Jr.
Dunlap Scott

J. H. Shaffer

M. J. Skinner

F. J. Smith

D. D. Sood

Martha Stewart

0. K. Tallent

Thoma

Trauger

. Unger

Watson

Watson

. Weinberg

Weir

. Whatley

White

. Woods

. Wymer

Youngblood
Central Research Library
Document Reference Section
Laboratory Records
Laboratory Records, RC
Y-12 Document Reference Section
ORNL Patent Office

RGO e
HO==HedE A

.
. -

oo aoaHdHdT 2 noHWH

.

HO-=Swu=E4grpqga=s0w

EXTERNAL DISTRIBUTION

J. A. Accairri, Continental 0il Co., Ponca City, Oklahoma TL601
R. M. Bushong, UCC, Carbon Products Division, 12900 Snow Road,

Parma, Ohio 44130

D. F. Cope, Atomic Energy Commission, RDT Site Office (ORNL)
C. B. Deering, Black & Veatch, P.0O. Box 8405, Kansas City,

6L11k

Missouri

A. R. DeGrazia, USAEC, RDT, Washington, D.C.

20545

Delonde R. deBoisblanc, Ebasco Services, Inc., 2 Rector Street,

New York, N.Y.

10006
91.
92.
93.
ol ,
95-96.

ot .
98.

99.

100.
101.

102.
103.
10L.

105-106.
107

120

D. Elias, RDT, USAEC, Washington, D.C. 20545

Norton Haberman, RDT, USAEC, Washington, D.C. 20545

T. R. Johnson, Argonne National Laboratory, 9700 S. Cass Avenue,
Argonne, Illinois 60439

Kermit Laughon, Atomic Energy Commission, RDT Site Office (ORNL)

T. W. McIntosh, Atomic Energy Commission, Washington, D.C. 20545

E. H. Okrent, Jersey Nuclear Co., Bellevue, Washington 98004

R. D. Pierce, Argonne National Laboratory, 9700 S. Cass Avenue,
Argonne, Illinois 60439

J. Roth, Combustion Engineering Inc., Prospect Hill Road, Windsor,
Cennecticut 06095

M. Shaw, Atomic Energy Commission, Washington, D.C. 20545

N. Srinivasan, Head, Fuel Reprocessing Division, Bhabha Atomic
Research Center, Trombay, Bombay T4, India

C. L. Storrs, Combustion Engineering Inc., Prospect Hill Road,
Windsor, Connecticut 06095

B. L. Tarmy, Esso Research and Engr. Co., P. 0. Box 101, Florham
Park, N.J. 07932

J. R. Trinko, Ebasco Services, Inc., 2 Rector Street, New York,
N.Y. 10006

TIC, Oak Ridge

Laboratory and University Division, ORO