ORNL-4577.txt

 

 

ORFL-L5TT

Contract No. W-7L0S5-eng-26

CHEMICAL TECHNOLOGY DIVISION
UNIT OPERATIONS SECTION

LOW-PRESSURE DISTILLATION OF A PORTION OF THE FUEL CARRIER SALT
FROM THE MOLTEN SALT REACTOR EXPERIMENT

J. R. Hightower, Jr.
L. E. McNeese

B. A. Hannaford

H. D. Cochran, Jr.

 

This report was prepared as an account of work |
sponsored by the United States Government, Neither '
the United States nor the United States Atomic Energy
Commission, nor any of their employees, nor any of
their contractors, subcontractors, or their employees,
makes any warranty, express or implied, or assumes any
legal liability or responsibility for the accuracy, com-
pleteness or usefuiness of any information, apparatus,
product or process disclosed, or represents that iis use

would not infringe privately owned rights.
AUGUST 1971

 

 

 

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

s ey T
e renmomm e AT o Tl d il e

g R T T TN : S . Loeew @AW
Sule bl ; A e e e

 
 

 

 
iii

CONTENTS

ABSTRACT . . ¢ ¢ v v 4 o & &« « o o &
INTRODUCTION .« + & ¢ o o v o o o« o «
DESCRIPTION OF EQUIPMENT . . . ¢ ¢ + « ¢ o o« « o « o

2.1 Process Equipment . . . ¢« + ¢« ¢« ¢ & o « o
2.2 Instrumentation .

2.2.1 Measurement and Control of Temperature

2.2.2 Measurement and Control of Pressure
2.2.3 Measurement and Control of Liguid Level
2.2.4 Radiation Instrumentation . . . . .

2.2, Instrument Panel .
2.3 Condensate Sampler . .

2.4 Location of Equipment at the MSRE . . . . . . . .
DESCRIPTION COF DISTILLATION OPERATICN . . . . . &
EXPERIMENTAL RESULTS . . . ¢« « « o ¢ &

L.l Summary of Experimental Data . . . . . . . . .
L.2 Material Balance Calculations . . . . . « . .

1.3 Results of Relative Volatility Calculations .

4.4 Possible Explanations of Calculated Results .
h.4.1 Entrazimment of Droplets of Still-Pot Liguid

4,h.2 Concentration Polarization .
h.h.3 Contamination of Samples
L,h.k TInaccurate Analyses .« « ¢« « o o & « o

CONCLUSIONS & & & & ¢ ¢ o o o s o o o o o « o s + o
ACKINOWLEDGMENTS . . . . . v ¢« ¢« ¢« ¢ ¢« ¢ o o o« « &

REFERENCES . « & o o« o ¢ o ¢ o s o o o o =

10

10
12
13
15
15
17

20

20

8. APPENDIX: ANALYSES OF SAMPLES FROM THE MSRE DISTILLATION
EXPERIMENT . . ¢ & & ¢ ¢ & ¢ o o o s o o o o

...........
 

 

 
LOW-PRESSURE DISTILLATION OF A PORTION OF THE FUEL CARRIER SALT

 

 

el FROM THE MOLTEN SALT REACTQOR EXPERIMENT
J. R. Hightower, Jr.
L. E. McNeese
B. A. EHannaford
H. D. Cochran, dJr.

ABSTRACT

An experiment to demonstrate the high-temperature
low=pressure digtillation of irradisted Molten Salt Reactor
Experiment (MSRE) fuel carrier salt has been successfully
completed. total of 12 liters of MSRE fuel carrier salt
was distilled in 23 hr of trouble~free opersation with gtiil-
pot temperatures in the range 900-980°C and condenser pres-
sures in the range 0.1-0.8 torr. ZEleven condensate samples
were taken during the course of the run at intervals of
approximately 90 min and were subsequently analyzed for Li,
Be. 7r, 13T¢s, 957y, 1lbbce, 14Tpy, 155my, 91y, 90sy, and
89gr. Effective relative volatilities, with respect to
Li¥, for Be and Zr were in good agreement with wvalues
measured previously in the laboratory. Effective rela-
tive volatilities for the slightly volatile materials
lthe, 9ly, 908r, and 898r were found to be much higher
than values measured in the laboratory. The high values
are believed to be the result of contamination from other
MSRE salt samples, although concentration polarization may
have also been a contributor. The effective relative vol-
atility for 13Tcs was found to be only 20%, cr less, of
the value measured in the laboratory: no explanation of
this discrepancy is available.

Although the effective relative volatilities for the
lanthanides were found tc be higher than anticipated, the
values observed would still allow adequate recovery of
TLiF from waste salt streams by distillation.

1. INTRODUCTION

Low-pressure distillation may be required in order tc recover

valuable carrier salt components from waste salt streams coming from

the fuel processing plant of a mclten-salt breeder reactor (MSBR).
Typically, TLiF would be vaporized and recovered, leaving & ligquid heel .
more concentrated in the less volatile lanthanide fission products (as

fluorides). This heel would then be discarded. The final stage of g

three-phase experimental program to study and demonstrate the feasibility

of distillation for decontaminating carrier salt components of lanthanide

fission products is described in this report. The experimental program

included measurements of relative volatilities of several lanthanide

and alkaline-earth fluorides in mixtures of LiF and Bng, Ls2 the opera~

tion and testing of a large single-stage still using fuel carrier salt from

the Molten Salt Reactor Experiment (MSRE) with simulated fission prod-
ucts ", and, as described here, a demonstration of the distillation
process using irradiated fuel carrier salt from the MSRE.

The operation of the distillation equipment with unirradiated salt
had the following objectives: (1) obtaining operating experience with
large, low-pressure, high-temperature distillation equipment; (2) in-
vestigating entrainment rates and separation inefficiencies due to con-
centration gradients in the still pot; (3) measuring distillation rates
under a variety of conditions; and (4) uncovering unexpected difficulties.
The objectives of the demonstration distillation of the irradisted MSRE
fuel carrier salt were: (1) to provide MSBR fuel processing technology
with a process tested with fuel salt from an operating resactor, (2) to
provide information (not available from the laboratory investigations)
on relative volatilities of fission preducts, and to give a general

confirmation of predicted fission product behavior, and (3) to uncover

unexpected difficulties associated with radioactive operation.
In the nonradicactive tests, six 48-liter batches of salt that had

the composition of the MSRE fuel carrier salt and contained NdF3 were
distilled at condenser pressures below 0.1 torr¥ and at a still-pot
temperature of 1000°C. Whereas these tests indicated areas in which
further engineering development was required, they also indicated that
decontsmination from lanthanide fluorides by distillation was feasible.
The still that was used in the nonradicactive tests was also used to

distill the radiocactive salt (containing no uraniuvm) from the MSRE.

This operation and its results are described in the secticns that follow.

2. DESCRIPTION OF EQUIPMENT
2.1 Process Eguipment

The equipment used in the MSRE Distillation Experiment included a
18-1iter feed tank containing s salt charge from the MSRE to be distilled,
a 12-liter still from which the salt was vaporized, a 1l0-in.-diam by
51-in.-long condenser, and a 48-liter condensate receiver. This equip-
ment is only briefly described here; a more complete description is given
elsewhere.3

The feed tank, shown in Fig. 1, was a 1/2-in.-diam by 26-in.-tall
right circular cylinder made from 1/L-in.-thick Hastelloy N. It was
designed to withstand an external pressure of 15 psi at 600°C. The
condensate receiver, shown in Fig. 2, was a 16-in.-diam by 16-1/2-in.-
tall right circular cylinder having sides of 1/b-in.~thick Hastelloy N
and a bottom of 3/8-in.-thick Hastelloy N. It was designed to withstand

an external pressure of 15 psi at 600°C.

 

%¥1 torr is 1/760 of a standard atmosphere.
ORNL DWG. 66-10983

     
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Np INLET —
3 O.DaI9GA. TUBE !f
M ; /
; ; FEED INLET
; TI7 - 3
]fir ! I T i) i 70.0.x15 GA. TUBE
| il 1 |
| i I I | |
! i |- |
i ! 1t
I H I | i
| 10y 111 |
I I 1 '
| 11 s i |
' it i1 '
I I 1l !
| Hi 1 !
! f.l | | :
i i Hi | |
) |
! 153 0D i 1! ' . o
i3 1.D. 1l | | 72* TYR '
g g | o i ‘
N o 2 { nlfl ! E H
| Hy I [
I Iy 1] I
it |
! i {
i i | [i] :
[ I L
| i I I
i | i1y |
| H 1 |
i . 4 y
b \ I \
it f £
M i \
=3l | :
E“‘\%;—:h‘_fl_ { 'II'H-ERMOCOUPLE WELL ;
= i 3 0:D.x19 GA. TUBE J
; FEED OUTLET —-
: 20.0.118 6A. TUBE
i PLAN VIEW
r fl

 

 

 

 

 

 

 

 

Fig. 1. Molten Salt Distillation Experiment. Schematic diagram of
feed tank.

 

ST
 

ORNL DWG. 66-10984

& VACUUM LINE
(" SCH. 40 PIPE

   
  
   

THERMUCOUPLE WELL
% ©.0.x.042 WALL TUBE

@— THERMOCOUPLE WELL

30.0.2.042 WALL TUBE

30.0.x.042 WALL TUBE

THERMOCOUPLE VIELLA@ >

—

FUEL DIP LINE & Ny ADD'N,
20.0.x.072 WALL TUBE
I3 SCH. 40 PIPE

l&'mloo'&

LEVEL PROBE TUBE

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

P
. . - . _
>~ / -
/i . N 7 I,:/";.A
/ WO gl -
7 N ‘\\‘l“. ,’/,i i P
) T _ ’ \;‘L"- R ]!.: il N
| /1 iyl R |
il Byoa Heov i
| \w/ | 0 o
1 : ol | !
i ! I
. ) 30° - L P l
‘ | N o |
| ! i il |
| [ o |
: 1! I | |
T ol I | K
: b | 1 i |
\ 1 I I
: L ! ] kR !
e —la i |11 b !
“ = i | i |
e 16 Ie.o. L o +
sgLo. || P
! v : 1 |
! L : 11 |
| ‘ | S 3
: | al v "
; : Vol I £ E'[ ;
i [ bho H l
! I I |
. )l Lis

 

 

Fig. 2. Molten Salt Distillation Experiment. ©Schematic diagram of

the condensate receiver.

 
The still snd the condenser are shown in Fig. 3. The still pet
consisted of an annular volume between the vapor line and the outlet
wall, and had a working volume of about 10 liters. Both the still and
the condenser were made of 3/8-in.-thick Hastelloy N and were designed
for pressures as low as 0.05 to 1.5 torr. The design temperature for
beth the still pot and the condenser was 982°C.

The feed tank., the still pot, the condenser, and the receiver were
mounted in an angle iron frame to facilitate their transfer between
Bldg. 3541, where the nonradiocactive tests were carried out, and the
MSRE site. ©Since the equipment was to be installed in a cell only
slightly larger than the equipment frame, the thermocouples, the heaters,
the insulation, and most of the piping were added before the equipment
was placed in the cell. TFigure 4 is a photograph of this equipment
(without the insulation). A stainless steel pan to catch melten salt
in the event of a vessel rupture was placed around the bottom of the
frame.

Because large gquantities of iron and nickel particles were expected
tc be present in the fuel storage tank (FST) at the MSRE (as a result
of the chemical processing of the fuel salt), a porous metal filfier was
installed in the feed tank fill line downstream from the freeze valve
in line 112 (see Fig. 5). The Inconel filter medium consisted of ap-
proximately 28 in.2 of Huyck Feltmetal FM 284 having a mean pore size
of 45 u.

To prevent particulates from reaching the vacuum pump, Flanders
High Purity filters were installed in the vacuum lines from the feed

tank and the receiver. These filters were tested and demonstrated to
 

   
  

 

 

 

ORNL DOWG, 66-10983

HERMOCOUPLE WELL
14270.0. TUBE %19 GA.--
¢ O
LEVEL PROBE - v—— FEED LINE

13 5CH. 40 PIPE : , 34*0.0.1.072 WALL

  
  
 

 
 
 
 

 

 

  

. - HANGER
! BRAGKET

 

 

 

 

  

LEVEL PROBE

 

 
  

I35CH. 40 PIPE

~— DRAIN & SPECIMEN HOLDER

THERMOCQUPLE WELL. 172°0.0. TUBE %19 GA.

172°0.0. TUBExXIS BA,

VIEW A=}

SAMPLE TUBE
k5 SCH. 40 PIPE

 

 

Fig. 3.
the vacuum still and condenser.

 

 

 

 

 

Molten Balt Distillation Experiment. Schematic diagram of

 
PHOTO 93564

 

 

 

 

 

Fig. 4. Molten-Salt Distillation Equipment Before Installation in
Spare Cell at MSRE.

 
ST
iy

 

 

Al o ke o VW — —

ARGON Iii

 

ORNL DWG 70-6280

 

 

 

 

 

 

SUPPLY
FREEZE
vaLve. 1 |
"z
MOLTEN
SALT
FILTER
FUEL
STORAGE
TANK "=} FEED
—1| TANK

 

 

 

     
  

ARGON
SUPPLY
PRESSURE
MEASUREMENT ™ _r -
AND CONTROL
HCV -2 E
- &
- VACUUM
HCV—9 ACUU
-3
I
i
L LEVEL FILTER
[ ] MEASUREMENT
— AND CONTROL
!
I -
! SAMPLER
a -
i
i

 

CONDENSER 4

| CONDENSATE
| RECEIVER

Fig. 5. Simplified Flow Diagram of MSRE Distillation Experiment.
10

effectively remove 99.997% of 0.3-u particles. The hcusings for these
filters can be seen in Fig. 4 on two of the larger lines at the right-
hand side.

A1l valves and piping that did not ccntact the molten salt were made
of stalnless steel and were housed in a sealed steel cubicle containing
pressure transmitters and two vacuum pumps— one to evacuate the reference
side of g differential pressure transmitter, and the other to evacuate
fhe distillation process vessels. The valve box, with its front and rear
cover plates removed, is shown in Fig. 6. This box completed the second-
ary containment around piping and instrumentation when its lower plates
were bolted and sealed in place. With the box under a pressure of 15 in.
B O, the leak rate was 0.1 cfh. During operation, the pressure in the box

2

never exceeded 0.5 in. H. 0, and the lesk rate was negligible.

2

2.2 Instrumentaticn

2.2.1 Measurement and Control of Temperature

Temperatures were measured and controlled over two ranges: 500-600°C
for the feed tank and the condensate receiver, and 800-1000°C for the
still and the condenser. Platinum vs platinum - 10% rhodium thermocouples
were used for the high~temperature measurements, whereas less expensive
Chromel-Alumel thermocouples were used on the feed tanks, condensate re-
ceiver, and salt transfer lines. FEach of the thermocouples {total, 60}
was enclosed in & 1/8-in.-diam stainless steel sheath; insulated junctions
were used. Five 12-point recorders were available for readout: twc for
the Pt vs Pt - 10% Rh thermocouples, and three for the Chromel-Alumel

thermocouples.
 

MSRE

Fig. 6. Containment Box for Instruments
Distillation Experiment.

and Valves Used in the

PHOTO 92843

 

 

T
12

There were a total of nine individually heated zones on the feed -
tank, the still, the condenser, and the receiver. The heaters for each
of these zones were independently controlled by a Pyrovane "on-off" con-
troller; the voltage to the heaters was controlled by Variacs. Heaters
on the various salt transfer and argon lines were manually controlled

by "on-off" switches and Variacs.

2.2.2 Measurement and Control of Pressure

Pressure measurements over three ranges were required: 0 to 15 psia
for monitoring system pumpdown at the start of the run, for monitoring
system repressurization at the end of the run, and for contrclling salt
transfer from the fuel storage tank; 0 to 10 torr for suppressing vapori-
zation while the salt was held at operating temperature in the still; and
C to 1 torr during distillation.

Absclute-pressure transducers(Foxboro D/P cells with one leg evacuated)
covering the 0- to 15-psia range were used to measure the pressure in the
feed tank and in the still-condenser-receiver complex. An MKS Baratron
pressure measuring device with ranges of 0-0.003, 0-0.01, 0-0.03, 0-0.1,
0-0.3, 0-1, 0-3, and 0~10 torr was used to measure very low pressures in
the condensate receiver.

The system pressure was controlled in the 0.l1- to 10-torr range by
feeding argon to the inlet of the vacuum pump. The Baratron unit produced
the signal required for regulating the argon flow.

It was necessary to ensure that an excessive internal pressure did
not develop in the system since, al operating temperature, a pressure in

excess of 2 atm would have Deen unsafe. This was accomplished by using

an absolute-pressure transmitter in the condenser off-gas line to monitor
..........

13

the system pressure. When the pressure exceeded 15 psia, the argon

supply was shut off automatically.

2.2.3 Megsurement and Control of Liguid Level

The difference in the pressure at the outlet of an argon-purged dip
tube extending to the bottom of the vessel and that in the gas space
above the salt was used to measure the depth of the salt in both the feed
tank and the condensate receiver.

Twe conductivity-type level probes were used in the still for measur-
ing and controlling the liquid level. These probes essentizlly measured
the total conductance between the metal probes {that extended into the
molten salt)} and the wall of the still; the total conductance was a func-
tion of the immersed surface area of the pro‘be,5

The conductivity probes (see Fig. 7) were similar to the single-point
level prcbes that were used in the MBSRE drain tanks. Tests have shown
that the range of such an instrument is limited to approximately 30% of
the length of the signal generating section. A 6-in. sensing probe was
used to control the liquic level between points that were 1 in. and 3 in.
below the still-pot overflow; a longer sensing probe wag used to measure
very low levels of liquid in the still pot.

Metal disks were welded to the conductivity probes to aid in their
calibration. These disks provided abrupt changes in the irmersed surface
ares of each probe at known liguid levels. During operation, the signal
from a probe changed abruptly when the salt level reached one of the
disks.

The liguid-level controller for the still pot was a Foxboro Dynalog

circular chart recorder-controllier, which consists of a 1-kHz ac bridge-type
1h

ORNL-DWG 67-4776R1

TOP VIEW

SIGNAL AMPLIFIER AND
LEVEL INDICATOR

EXCITATION
SOURCE

 

A

X\

 

 

 

SIGNAL LEADS

 

 

 

 

 

HEAD COVER

FOLDED EXCITATION
SECTION

 

CONTAINMENT VESSEL

 

 

_ BUOSSOSOOOOONINNY

/ |

_ e N N N N N e N N N R s e N S A R S AN AN A S A A S AR R R A Y A Y

 

DISKS TGO AID
CALIBRATION

 

 

 

Fig. 7. OSimplified Schematic of Conductivity-Type Liquid Level
Probe for Still Pot in MSRE Distillation Experiment.

 
15

_____________ measuring device using variable capacitance for rebalance. The proper
control action {see Sect. 3) was accomplished by having a variable dead
zone imposed on the set-point adjustment mechanism. With the controller
set for the desired average liquid level, the argon supply valve to the

feed tank was opened when the level indicator dropped 3% below the set

point and was closed when the level indicator rose 3% above the set point.

2.2.4 Radiation Instrumentation

Ionization-chamber radiation monitors were mounted on process lines
in three locations: one on the filter in the feed tank vacuumn line, one
on the filter in the receiver vacuum line, and one on the liquid-nitrogen
trap in the receiver vacuum line. The two monitors on the filters were
shielded from the radiation field in the cell by an 18-in.-thick barytes
concrete bleck wall and indicated the level of radiocactivity for each
filter. The monitor on the liguid-nitrogen trap was not shielded from
the radiation field produced by the process vessels and thus registered
the general level of radiation in the cell.

Two Geiger-Muiller tubes, which were attached to the valve box,
monitored the vacuum pumps. They were set to sound an alarm when the

radiation level reached 1 mR/hr at a point about € in. from the pumps.

2.2.5 Instrument Panel

 

The instrument panel, from which the process was controlled, con-
tained the temperature controllers, the pressure and level reccorders and
controcllers, the valve operation switches, the electrical power supply
controls, the temperature and pressure alarms, and four of the five

temperature recorders. This panel is shown in Fig. 8. The fifth

it
 

 

 

 

 

 

 

F

ig.

8.

Instrument Panel

for MSRE Distillation Experiment.

R PHOTO 89322

 

 

91
17

e temperature recorder and the radiation measuring instrumentation were

mounted in other cabinets.

2.3 Condensate Sampler

The condensate sampler was the most important item of equipment for
obtaining information from the distillation experiment. The sampler was

233U

patterned after the equipment that had previously been used to add
to the fuel drain tanks and to take salt samples from the drain tanks.
Modifications of this design were made to allow the sampler to be evac-
uated to about 0.5 torr so that condensate samples could be withdrawn
without disturbing the operation of the still.

Figure 9 shows a cutaway diagram of the sampler. The main components
of the sampler were: (1) the containment vessel in which the samples
were stored; (2) the turntable inside the contaimment vessel, which
allowed the sample capsules to be aligned with the handling tool and
also with the removal tool; (3) the capsule handling tool, with which
empty capsules were sttached to the cable tc be lowered into the sample
reservoir; and (L) the reel assembly, with which capsules were lowered
and raised.

The following sequence was used in collecting a condensate sample.
With valve HV-62 (see Fig. 9) closed and the containment vessel at atmos-
pheric pressure, the sample handiing tool was raised to the highest
position. As shown in Fig. 9, the cable was attached about 20 in. from
the top of the tool so that, when the cable was reeled to the highest
position, the top end of the tool protruded through valve HV-66 and the

samples on the turntable could pass under the lower end of the tool.
18

ORNL DWG 69-4945 s

 

 

 

 

 

 

 

 

 

 

 

 

  

VACUUM
PUMP

 
 

  

FILTER—

Fig. 9. Cutaway Diagram of Condensate Sampler for MSRE Distillation
Experiment.

............
 

19

With the tool at its highest position, an empty capsule was rotated
underneath the lower end of the tool. The tocl was then lowered onto

the stem of the capsule and locked in place by an adjustment at the end
of the tool protruding through HV-66. The tool and the attached capsule
were raised again, the turntable was rotated until the sampling notch

was located underneath the tool, and the tool was lowered below valve
HV-66. Valve HV-66 was then closed. At this time, the vacuum pump was
turned on, and the containment vessel was evacuated to about 0.5 torr.
When the pressure in the containment vessel reached 0.5 torr, valve HV-62
was opened and the sample handling tecol and the empty capsule were low-
ered until the sample capsule rested on the bottom ¢of the sample reservoir
at the end of the condenser. The tool and capsule were then ralised above
valve HV-62, which was subseguently closed. Next, the containment vessel
was pressurized to atmospheric pressure with argon. Valve HV-66 was

then opened, and the sample handling tocl was raised to its highest
position. Finally, the empty samplie holder was rotated underneath the
sample handling tool, and the sample was lowered into its holder and re-
leased from the tool. The process was repeated by raising the sample tool
again, rotating another sample capsule undernesth it, etc. The
turntable was designed to contain 11 sample capsules. After the samples
had been collected, they were stored in the containment vessel. At the
end of the experiment, they were removed for analysis.

A blower, which drew air into the top of the line at the reel assem-
bly to prevent radioactive particles from escaping into the operating
area, was provided for the parts of the operation requiring HV-66 to be
open. The air handled by the blower was filtered and exhausted into the

cell.
20

The sample capsules were standard 10-g MSRE sample capsules, each
of which was fitted with a key for attaching to the sample handling tool.
Figure 10 shows one of these capsules. The three wire ribs on the stem
ensured that the capsule would remain vertical in the capsule holder on
the turntable.

¥igure 11 is a photograph of the sampler during installsation at the
MSRE. 1In this photograph, the unloading tube has been capped off, and
the turntable operating handle is not in place. The ring of lead bricks
around the sampler forms the base for the radiation shield, which is

fitted over the containment wvessel.

2.k ILocation of Equipment at the MSRE

The éistillation unit was installed in the spare cell at the MSRE
site; the sampler., the valve box, and the instrument panel were placed
in the high bay area above the cell. Figure 12 shows a schematic diagram
of the spare cell, while Fig. 13 shows a photograph of the unit in the

cell before the cell cover was put in place.

3. DESCRIPTION OF DISTILLATION OPERATION

The transfer of salt from the fuel storage tank (FST) to the feed
tank was initiated by evacuating the feed tank, which contained about
2 liters of unirradiated salt, tc gbout 1.5 psia. Prior to the transfer,
freeze valve FV-112, located between the FST and the feed tank for the
still (see Fig. 5), was heated and opened. (Salt had previously been

frozen in the feed line to the still pot in order to isolate the feed
 

Fig. 10.
Experiment.

~ PHOTO 95933

Condensate Sample Capsule Used in the MSRE Distillation

 

 

Tc
22

PHOTO 94943

 

Fig. 11. Condensate Sampler for MSRE Distillation Experiment.

 
23

ORNL DWG 68-{30R1

SAMPLER

   

 

 

 

 

 

HIGH BAY AREA

 

 

o °° o ollP I:, o o
3'_0'::, : EESCF; : ioo ° o :&:\oo.
l 0% o ° o:o °°°co'--F'EXED
e 0% % ollgpe °.%2p e SUPPORT

 

    
   

°l PLUG
*eNs-BARYTES b o
¥ CONCRETE ®

 

: N BLOCKS

> — ol {'-0"
FROM FUEL LINE 4
STORAGE & 0

 

TANK

 

h fi gg&%rEN ° o
° ELEV. 840'-0"
<7 FILTER o f

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

e Yo o
I 00,00
— [~ 6 © ° 4
STILL &g Pleec
'-:5_',6? :’°° o e ¢
YA e ° ) Do <
S i
.()-( ° : 04 °°
‘.‘q g © o 0
35 o ° LAy 0:
ol fees
CONDENSER — Q s
eo] /g ELEV. 83¢'-0"
5 X

 

 

L W
v v v L) = :
o8l © 2% S0%ns 00 %0 » CONDENSATE RECEIVER
‘0‘,(.’0"o °°°o°°:°:°:°:°°o°°o "0c °°°°°o go
"000 °°°o°°°.°o-°°.°°°°°°°: °o°:°° 0% o ¢

Fig. 12. South Elevation of MSRE Spare Cell, Showing Distilliation
Unit and Sampler Locations.
24

QRNL PHOTO 94840R
FiL’FER fificz:, VER

  
   
  
  
  
  
 
 
 
  
   
 

TG HCV 4 fi& _
VALVE BQK

FILTER IN
FEED TANK -
VACUUM LINE

EXPANSION ok Bt - .
JG NT : TR B amEBER. 0000 e RS S
- - SAMPLER

- VACUUM
PUMP’ -

~SA PL,E .

3 v_%;{:uum LINE
BLOWER £ FILTE
FOR SAMPLER -
VENTILATION

 

 

 

LiQ N, TRAP

C@ND&NS»&TE
REEE VER

| DUITS FROM
.  PROCESSING

STILL POT __ :

DRAIN LINE ™ §

 

JUNCTION BOXES | a@x;%‘ FOR
THERMOCOUPLE
POWER LEA

FILTER IN
FEED TANK
FilLk LINE
(LINE 442)

L LINE M2 PENETRATION | i
 (FROM FUEL STORAGE TANK) - ® N

Fig. 13. Molten Salt S5till Installed in Spare Cell at MSRE.

 
25

S tank from the cther process vessels.) After 12 liters of salt had been
transferred from the FST, the pressure of the feed tank increased from
about 4 psia tc atmospheric pressure over a period of about 2 minj; this
indicated that gas was being transferred from the feed tank through the
salt charging line. More salt could not be transferred from the FST even
though the bubbler in the FST indicated that additional salt was present.
Only one further attempt at salt fransfer was made because of the potential
danger of blowing the trapped salt out cf the freeze valve; this would
have made it impossible to obtain a seal between the feed {tank and the
fuel storage tank. After we had established that & tight seal could be
made, we decided to proceed with the experiment using the salt already
transferred, even though its volume was much less than the anticipated
volume (48 liters).

To start the distillation experiment, the still-pot feed line was
thawed, all process vessels were evacuated to 5 torr, and the still pot
was heated to 900°C. The valve between the feed tank and the vacuum pump
was then closed, and argon was introduced into the feed tank to increase
the pressure to about 0.5 atm; this forced the salt to flow from the feed
tank into the still pot. When T liters of salt had been transferred to
the still pot, the condenser pressure was reduced to 0.2 torr to start
the distillation of salt. At this point, control of the liquid level in
the still pot wes switched to the automatic mode. In this mode, salt was
fed to the still pot at a rate slightly greater than the vaporization rate.
The argon feed valve to the feed tank remained open {forcing more salt
into the still pot) until the liquid level in the still rose to a given

set point; the valve then closed and remained closed until the liquid
26

level decreased to a second set point. In this manner, the salt volume
in the still pot was maintained near T liters.

As the salt vapor passed through the condenser, heat was removed
from it by conduction through the condenser walls and the insulaticn and
by convection to the air in the cell. The condenser was divided into
three heated zones, the temperatures of which could be contrclled sepa-
rately when condensation was nct occurring. A sharp increase in tempera-
ture above the set points near the condenser entrance, as well as a
gradual temperature rise near the end of the condenser, accompanied the
beginning of distillation. Operation of heaters to keep the temperature
of the condenser abové the liguidus temperature of the condensate was not
necessary during vapcr condensation. In this part of the run, the still-
pot temperature was slowly increasing; and, since the concentrations of
volatile BeF2 and Zth were still fairly high, the vaporization rate was
alsc increasing. An abnormally high temperature at the end of the con-
denser indicated that the capacity of the condenser would be exceeded.

By raising the condenser pressure to 0.8 torr, the distillation rate was
reduced sufficiently to maintain the condenser temperature near T00°C, an
acceptably low temperature.

When the contents of the feed tank (7 liters) had been depleted, the
salt in the still feed line was frozen; then & total of 4 of the 7 liters
of salt in the still pot was distilled by batch distillation. At this
point, the still-pot temperature was 980°C. As the more volatile mate-
rials were vaporized from the still pot, the condenser pressure was re-
duced from 0.8 torr to 0.1 torr in order to maintain a fairly high

distillation rate. When the condenser pressure could not be decreased

..........
27

o further, the distillation operation was terminated by increasing the
pressure in all the process vessels to atmospheric pressure and turning
off the power to all the heaters.

The semicontinuous distillation phase lasted for 8.3 hr; the average
distillation rate during this pericd was 0.57 liter/hr. The duration of
the batch distillation period was 13.8 hr; the average distillation rate
for this pericd was 0.36 liter/hr. Eleven condensate samples were taken
during the run at approximately 90-min intervals. At the end of the
experiment, radiation readings of these samples ranged from L R/hr at
contact (the first sample) to 500 mR/hr at contact (the last sample).

After the still had been allowed to ccol down, a2ll electrical and
thermocouple leads to the still were cut and all pneumatic instrument
lines from the valve box were disconnected. The sampler was dismantled
and sent to the burial ground; the projecting end of the sample line was

. flanged. Valve handle extensions were cut flush with the floor of the
high bey ares. The process vessels and the valve box were allowed to

remain in place.

4. EXPERIMENTAL RESULTS

4.1 Summary of Experimental Dats

In order to determine the capability of the distiliation equipment
for separating fission products from the carrier salt, we measured the
following quantities in the course of this experiment: the concentration
of 811 the major and most of the minor components in the feed salt to the

still, the concentration of each component in the condensate as it left
28

the still pot, the volume of liquid fed to the still pot, and the vclume i
of ligquid collected in the receiver. An estimateT was made of the con-
centration of each fission product whose concentration in the feed salt
was not measured. The estimate was based on the production rate of each
fission product in the MSRE and assumed that all the precursors of a
given fission product remained in the salt. This seemed to be a fairly
gcod assumption for isotopes having no long-lived gaseocus precursors.
All the concentraticns were converted to mole fractions; the mole
fractions of radicactive materials were caiculsted as of May 7T, 1969.
The volumes were determined by measuring the weight of liquid over the
end of & bubbler, dividing by & salt density of 2.2 g/cm3 to obtain the
depth of liquid, and multiplying this value by the cross-sectioconal area
of the particular vessel to obtain the velume. We assumed that the mass
density of the liquid was independent of composition and that the wvolume
of liquid in the still pot could be calculated by subtracting the volume .
of condensate collected from the total volume of salt fed to the still
from the feed tank. Assuming that molar volumes can be added (which is
e fairly good assumption for fluoride salts), we calculated mass densities
for liquids in the concentration range seen in this experiment and found
only & 5% varistion; hence, our assumption of constant density appears

to be acceptable. The analyses of all the condensate samples are reported

in the Appendix.

4.2 Material Balsnce Calculatiocns

One of the mcst concise ways to express the separation performance of

 

the distillstion equipment is to convert the condensate analyses tc effective
29

S relative volatilities with respect to LiF. The effective relative

=
volatility of component i with respect to LiF is defined as:
. - v,/ X, -
. . = s
1-LiF = ypip/¥p;p
where y = mole fraction in the condensate,
x = mole fraction iIn the still pot.

Although the composition of each component in the still pot was not meas-
ured directly during the run, the data allowed the composition of the
still pot to be estimated from a material balance for each component.
In this section, we derive the material balance eguations and outline the
calculational procedure.

For the semicontinuous mode of operation, a differential mole balance
for component i gives:

£,dI - y,dO = dM,, (2)
where I = total moles fed to the still pot,
0 = total moles removed from the still poct,
Mi = moles of component i in the still pot,
fi = mole fraction of component i in the feed,

v, mole fraction of component i in the condensate.
Since volume was the measured quantity, we can make the following
substitutions:

&Vin
al = N ’ (32)
ot f Ve
30

40 = {(3p)

where Vin = volume of salt fed to the still pot, liters,

out = volume of condensate collected, liters,

partial molar volume® of component j, liters/mole.

v,
J

When we substitute the quantities in Eq. (3) into Eg. (2), we obtain:

f Vi
AV, - V=AM (k)

j=lfjvj ) b.v

Integration of Eq. (L) yields

Vout

fl Vi
| in N dVout =M (5)

f.v, Y.V,

j=1 J J g;l Jd d

C
where the conditions V, =V =M, = 0 at the start of the experiment
in out i

vere used. The compositlion in the still pot can be determined by solving
for Mi for each component. This eguation is valid up to the point where
Vout = 5.07 liters (the end of the semicontinuous distillation).

If Mi moles of component i are present in the still pot and dMi

moles of component i are vaporized during the batch distillation, the

mole fraction of ccmponent i in the vapor is given by

dM.
i

 

¥pssumed to be independent of the composition of the liquid.
31

similarly, for component j (i # 3),

My
v v, = (1)
<7
o aM
k=1 1k
From Hgs. (6) and (7), we obtain the expression
dMi de
. (8)
i J
or
Y5
dM. = dM.o (9)
1 Y. J
J
Assuming that the partial molar volumes v, are independent of the composi-
tion of the liquid (as before), we multiply both sides of Eg. (9) by vy
and sum both sides over all 1 to obtain the following equation:
n T Loa
do Vi T Wopsny 50 o yvy) (10)
i= i___}zl . J
where Vstill = volume of salt in the stiil pct, liters.
Sclving Eq. (10) for de vields:
73
de = Tm" dvstill' (11)
PY.LV,
e
Integrating Eq. {11) yields:
Vstinn
’ Y,
M, =M, + / i
Oy N Vsti11 12)
vV <Y
i=1
32

where N%O = moles of component jJ present in the still pot at the start E%#;

of the batch distillation,
VO = volume of salt in still pot at start of batch distillation.

in Eq. (12) is calculated by evaluating Eq. (5), using

The quantity MjO

the values of Vin and Vou that ccrrespond to the end of the semicontinuous

t
distillation. (At the end of semicontinuous distillation, Vin = 13.8
liters and Vout = 5,07 liters.) The composition of the still pot is

determined by evaluating Eq. (12} for each component that is present.

The integrals in Egs. (5) and (12) were evaluated by plotting

Y.

N
§lyivi

1

vs the volume of condensate collected in the receiver during the semi-

continuous distillation and

vs the still-pot volume (calculated as explained previously) during the
bateh distillation, fitting the resulting curves to simple empirical
functions of the appropriate volume, and then integrating these equaticns.
For any component, a fairly wide range of equation parameters fit the
scattered data equally well; however, the values of the integrals were
not significantly sensitive to these variations in the parameters. The
sums ggyivi and gtffivi were adequately represented by considering only
i=1 i=1

the major sslt components LiF, BeF2, and ZrFu since the mole fractions

 
33

g of the fission products were negligible as compared with the mole
fractions of these components.
Using Egs. (5) and (12), we calculated the number of moles of each
component present in the still pot at the time each condensalte sample was
taken. The mole fraction of each component in the still pot at that time

wag calculsted using the following equation:

 

X, = . (13)

The measured values of N and the calculated values of X, were then used
in Eq. (1) to calculate the relative volatility of each component, with

respect to LiF, for each sample.

k.3 Results of Relative Volatility Calculations

The effective relative volatilities, with respect tc LiF, of the

95

major salt components, BeF_ and ZrFu, and of the fission products ir.,

2

1
luhCe, L}TPm, 155Eu, 91Y, 9OSr, and ls?Cs were calculated using the methods

outlined shove. The variations in relative volatilities during the ccurse
of the distillation operation, as cobtained from the calculations, are
given in Figs. 1hk-1T7. 1In these figures, the most self-consistent values
resulted when the carrier salt composition in the feed was assumed to be
65-30-5 molie % LiF—BeFZ—Zth rather than the slightly different composi-
tion indicated by the analysis of the selt from the fuel storage tank.

The difference between the analytical values and the values actually

used can probasbly be attributed to zirconium metal, which was present in
RELATIVE VOLATILITY OF BeF,

WITH RESPECT TO LiF

ORNL-DWG 70 - 4516

 

 

 

 

 

 

 

 

L ! [ 1 v | ! | 3 [ i 1 ¥ I ' I
_ g O B
A 4
10 |- A -
8t =
6 —
@
5H o . ® ® ® o S -
| 2 _
) & ®
3 = —
® @
2r -
® BefF,
A ZrF,
O "7y
g L | | ! i | | i | i ] ] i I |
0 1 2 3 4 5 € T 8
LITERS CONDENSATE COLLECTED
Fig. 14. FEffective Relative Volatilities of BeF,, ZrF), and 9521‘.

10

& RN O

RELATIVE VOLATILITY OF ZrF, & %2

S

WITH RESPECT TO LIiF

n€
35

ORNL- DWG 70 - 4545

 

 

 

 

 

 

 

 

i il T T T T T T
@
@ 1940q
10-% &
& @ @
z ® :
if B
7 ¢ & ]
= & & m
: . s
g A A 147pm_|
5 &
= |
<
!
il
&
{03 & ]
@® 94Ce A
;‘ & %7pm i
- O 1885, .
G
. O Q O -
e}
o o O QSSEU -
O
<i.4 x10°° 0.86 x 107¢
10-% {: | 1 ] ) j ; | '[1/ | ; | 5 | i
O 1 2 3 4 O 5 S 7 g8

LITERS CONDENSATE COLLECTED

Fig. 15. Effective Relative Volatilities of lthe, 1h7

 
RELATIVE VOLATILITY WITH RESPECT TO LIiF

0~

ot

{0°%

i0~*

36

ORNL DWG T70-45t4

 

 

 

 

 

 

i v 1 E ¥ ] ¥ I I E ¥ I 1 I
®
: ® i
@
—- ® 5
8
® Y
C ® ~
: . &
1 i, 4 e A i
A QOSr
_ A i
A
: .QIY :
. A %5y i
] ] ] ; ] | 1 i ] j ] ] i
Q 2 3 4 5 6 7 g
LITERS CONDENSATE COLLECTED
. . . e e s 91 Q0
Fig. 16. Effective Relative Volatilities of 7Y and ~ Sr.

 
37

g

ORNL DWG T70-4513
100 T T T T T | T T T T T Y I T

 

L. ‘ A -
f=4.02 x10°% mole fraction 4 A

 

0.0

Lol

L

o

ruln]

RELATIVE VOLATILITY WITH RESPECT TO LiF

v

 

 

 

] | | i ! 1 ] ] f ] | i
0 { 2 3 q 5 6 7 8
LITERS CONDENSATE COLLECTED

 

OJ 1 E ]

137

Fig. 17. Effective Relative Volatility of Cs, Bhowing Effect of

Varying the Assumed Feed Concentration of 137¢s.

.............
38

the fuel storage tank at the time the analysis was made but was filtered

out when the salt was transferred to the still feed tank.

952r are )

The effective relative volatilities of BeF Zth, and

5
shown in Fig. 14. The effective relative volatility of BeF2 was essen= _
tially constant during the run, exhibiting an average value of 3.8. This

value.agrees favorably with the value of 3.9 measured by Smith, Ferris,

and Thom.pson2 but is slightly lower than the value of U.T measured by

Hightower and McNeese.l Slight inaccuracies in calculations (especially

the material balance calculations described earlier) and analyses probably

account for such differences.

95

The effective relative volatilities of fission product Zr and of
Zth, the carrier salt constituent, are in agreement with respect to
magnitude and variation during the run. When the analysis of the salt

in the FST was used in the relative volatility calculation, the resulting
relative volatilities of the Zth were about a facteor of 2 lower than the

QSZr.

Figure 14 shows that o decreased from an initial

values for Zth—LiF

value near L4 at the start of the run to about 1 at the end of the run.
These values bracket the value (2.2) measured by Smith g&_g&,g in salt
mixtures having a Zth concentration abtout 2% of that in this system.

The effective relative volatilities of the lanthanide fissicn products

1&&089 147 155

Pm, and Eu are shown in Fig. 15. The effective relative

o 14k -4 . :
volatility of Ce rose sharply from 6.1 x 10 at the time of the first
sample tc about 1.0 x 1O=2, where it remained for nearly all the subse-
quent samples. However, the relative volatility for the fifth sample
was lower than 1.0 x 10—2 by a factor of 3. The low initial value wsas

1.5 to 3.4 times the value measured in an equilibrium still, whereas the
39

high, steady value was 24 to 56 times the value measured in the

S

equilibrium still.

1h7

The effective relative volatility of Pm was based on a computed

feed concentration, which was considered to be a more accurate value
than the measured feed concentration since measured concentrations of

other lanthanide fission products had agreed well with computed concen-

1hh 1hT

trations. As in the case of Ce, the relative volatility of Pm was

low at the time the first sample was taken, that is, less than 7.8 x

-4 3

10 '; it rose sharply to about 3.4 x 10"~ for the remaining samples

except for the fifth sample, which was low. There were no previcusly
measured values for the relative volatility of promethium with which
these results could be compared.

At this point, it is interesting to note that, when the effective

relative volatility of lhTPm was calculated on the basis of the measured

147

concentration of Pm in the FST instead of on an estimated concentra-

L for each sample, except the first, nearly coincided
147 pp=1iF Ll
1

with the respective calculated point for Ce.

tion, the o

155

Eu during the run

k7

The varistion of the relative volatilities of

Pm. Although
’)

closely paralleled the variations observed for 1the and

the value for the first sample was low (less than 1.5 x 10 °), the values

for all subsequent samples, except the fifth, were much higher, that is,
about 2.2 x 10mh. The value for the fifth sample was lower than this by

a factor of 2.6. These calculated effective relative volatilities

155

(which were based on a computed feed concentration of Eu) were lower

than the value of 1.1 X 10-3 meagured in recirculating equilibrium stills.

1 . . .
However, the accuracy of the 55Eu values is questionable since some
Lo

difficulty was experienced in meking the analyses; all of the results

155

for Eu in the condensate samples were reported as approximate (see

the Appendix). On the other hand, it is significant that the variation

of aj . during the run closely paralleled the relative volatilities
2OFu-LiF

of lthe and lkTPm.

Figure 16 shows the effective relative volatilities of le and QOSr.

9l

During the run, the effective relative volatility of Y had an average

value of 1.4 x 10_2, about 410 times the value measured in recirculating
equilibrium stills. The variation of ale - during the run was similar

-Li

to variations of relative volatilities of the lanthanides; the low value
for the fifth sample was most noticeable.

The effective relative volatility of 908r (based on the measured
concentration in the feed) had an average value during the run of 4.1 x
10—39 about 84 times the value measured in recirculating equilibrium

stilis. Although not shown in Fig. 16, the average value of the relative
89

volatility of Sr (based on a computed concentration in the feed) was

0.193, about 3900 times the value measured in equilibrium stills. The

89 G0

ratio of the Sr activity to the Sr activity in the condensate samples

varied from 0.22 tc values greater than 10. It should have been about

0.002 for each sample or, at the least, should have remained constant.

137

Calculations of the effective relative volatility of Cs were

based on an estimated feed concentration since the concentration of

137 137

Cs in the feed salt was not measured. However, Cs has a fairly

137

long-lived gaseous precursor Xe, half-life = 3.9 min) that causes

the actual concentration of 13705 in the salt to be less than that pre-

dicted when the method outlined earlier is used. Also, because the
L1

g sctual relative volatility of CsF is fairly high, the results of

calculations of the effective relative volatility sre sensitive to the

2
assumed feed concentration of lJTCs. Figure 17 shows calculated rela-

137

. tive volatilities:of Cs for two assumed feed concentrations. The

points around the lower curve result from the feed concentration of
lsTCs which would apply if all the precurscrs remained in the salt
during MSRE operation. These pcoints represent lcower limits for the
effective relative volatility in this distillation; the points around
the upper line result from a feed concentration just high enough to keep

137

the computed concentration of Cs in the still-pot liguid from falling

to zero or below, and represent upper limits for the effective relative

137 3Tng

volatility of Cs. The highest effective relative volatility of 1
seen in these calculations was only about 20% of the value measured by

Smith et al. in LiF-BeF  systems.

2
As seen from the previous calculated results, all components except

beryllium and zirconium had effective relative volatilities that differed

(in some cases, drastically) from values predicted from work with equi-

librium systems. Some possible explanaticns for these discrepancies are

discussed in the following section.

4.4 Possible Explanstions of Calculated Results

Possible causes for the discrepancies between the effective relative
veolatilities of some of the fission products in this experiment and values
measured previcusly include: (1) entrainment of droplets of still-pot

liquid in the wvapor, (2) concentration gradients in the still pot,
Lo

(3) contamination of samples during their preparstion for radiocchemical s
analysis, and (4) inaccurate analyses. These possibilities are discussed

below.

L.4.1 Entrainment of Droplets of Still-Pot Licuid
Entrainment rates of only 0.023 mole of liquid per mole of vapor,
or less, would account for the high relative volatilities calculated for

the slightly volatile fission products lthe, lhTPm, 91 90

Y, and Sr.
Rates of this order would not be reflected in the effective relative
volatilities of relatively volatile material (o > 1). The high correla-
tion of the scatter of the calculated effective relative volatilities of
different slightly volatile fission products is consistent with the
hypothesis that entrainment occurred.

Although there was some evidence of entrainment during the non-
radiocactive operation of the still,h the considerably lower rate of
distillation of the MSRE saltmakes entrainment by the same mechanism
less likely. Therefore, some cther reason for entrainment in the radio-
active operation should be sought. Evidence of a salt mist above the
salt in the pump bowl at the MSRE and also sbove .salt samples removed
from the MSRE has been reported;8’9 the studies have indicated that
these mists are present over radicactive salt mixtures but not over non-
radioactive mixtures. According to the reported data, either a mist
concentration (grams of salt per cm3 of gas) or a rate of formation of
mist (grams of salt per second) could be calculated. Entrainment rates

sufficiently large to explain the results of this experiment could only

be obtained by assuming that the gas space above the salt contained salt

...............

mist having the same concentration as that seen in the studies, instead
L3

Y of by assuming equal mist formation rates. However, the concentrations
calculated from the data were scarcely adeguate to explain the entrained
fraction that must have occurred. Furthermore, both the mist formation
rate and the concentration of the mist would be expected to decrease
with decressing decay power density in the liquid. ©Since the salt used
in the distillation experiment had a much lower decay power density than
the salt examined for mist formation {400 days of cocoling for distilla-
tion feed, as compared with less than 30 days of cooling for salt samples
tested for mist formation), it seems unlikely that the mist concentra-
tion would have been high enough to explain the high relative volatili-
ties for the slightly volatile fission products.

In addition to the argument against the entrainment hypothesis just
rgivena not all the discrepancies, for example, the variations in the

89Sr/90

Sr activity ratio of and the low wvalue for the effective volatility

137

. of Cs, would be explained by the entrainment hypothesis.

L. 4.2 Concentration Polarization

Concentration polarization would cause the effective relative vola-
tilities of the slightly volatile materials to be higher than the true
relative volatilities. As the more-volatile materials were vaporized
from the liquid surface, the slightly volatile materials would be left
behind at a higher concentration than in the liquid Just below the sur-
face. In turn, the vapor-phase concentration of these slightly volatile

- materials would increase, since further vapcorization would cccur from

& liguid with successively higher concentrations of slightly volatile

materials. Thus, since effective relative volatilities were based on

average concentrsations in the still pot, the vapor concentration would
Ll

be higher than that corresponding to the average liquid concentration, e
and the calculated effective relative volatility woculd be higher than
the true relative volatility, if the concentrations at the surface of
the liquid were higher than average. Similarly, concentration polariza- -
tion would cause the effective relative veolatilities of components with
relative volatilities higher than 1 to be lower than their true relative
volatilities. If mixing or diffusion did not reduce the concentration
gradient in the liquid, then the separation performed by the still would
be adversely affected.

As noted in ref. 4, the extent to which concentration fiolarization
affects the effective relative volatility of a particular component
depends on the dimensionless group D/vL, which qualitatifiely represents
the ratio of the rate of diffusion of & particular component from the
vapor-liquid interface intc the bulk of the still-pot liquid to the rate
at which this material is transferred by convection to the interface by
liquid moving toward the vaporization surface. In this ratio, D is the
effective diffusivity of the component of interest, v is the veloccity of
liquid moving toward the interface, and L is the distance between the
interface and the point where the feed is introduced.

The occurrence of concentration polarization is suggested by the
sharp rise, at the beginning of the run, in the effective relative vola-
tilities of lthe, 1)'L'T‘r"’rnl, 155Eu, and, possibly, 91Y and. 9OSr. This rise
would correspond to the formation of the concentration gradient at the
beginning of the run. The effective diffusivities of NdF3 in the still

pot, calculated from results of the nonradicactive experiments,h ranged -

L

from 1.4 x 10 = to 16 x lO_LL cmg/sec. They form the basis for estimating

 
b5

- the magnitude of the concentration polarization effect in the radioactive
operation. During the semicontinuous operation at the MSRE, the liquid
velocity resulting from veporization averaged 2.2 X 10'h cm/sec and the
depth of liguid above the inlet was approximately 9.4k em. If cne assumes
that the effective diffusivity of the fission products in the still pot
during the MSRE Distillation Experiment was in the same range as that
seen during the nonradiocactive tests, the observed relative volatilities
of the slightly volatile materials would be only 2.0 to 18 times the
actual relative volatilities, and the observed relative volatility of

137Cs would be 0.011 to 0.021 times its true value (in each case, assuming

that the true relative volatilities were those given in refs. 1 and 2).

Although concentration polarization may have been present in the work

with radicactive salt, the effect was less important than that needed to

account for the discrepancies between observed relative volatilities and
what we consider to be the true values. Concentration polarization would

not explain the variation in the ratio of the activities of 89Sr and 9OSr

between samples.

4. L.3 Contamination of Samples

 

The possibility that the condensate samples were contaminated during

preparation for analysis is suggested by the extreme variation in the

ratio of 89Sr and 90

Sr activities between samples. Although routine
precautions against sample contamination were taken in the hot cells
- where the capsules were cut open, no special precautions were taken.

(The same manipulators that are used to handle MSRE salt samples were

employed for opening these condensate samples.) If it is assumed that

the scurce of the contamination was a sample from the MSRE taken just
L6

before the condensate analyses were sent to the hot cells, then the -
amount of radicactive material necessary to result in the observed values

89 90 3

of the ratio of Sr and Sr activities was in the range 10-6 to 10~ g
per gram of sample. Prevention of such a low level of contamingtion is
extremely difficult.

Other observations that csn be explained by assuming that the samples
were contaminated are the high relative volatilities of the slightly vol~-
atile fission products and the high correlation between the variation of
calculated relative volatilities of different fission products. On the

other hand, the low relative volatility for 13TCS is not explained by

this hypothesis.

L. k. Inaccurate Analyses

95 137 1l 90 89

The analyses for ir, Cs, Ce, or, and Sr were made by

proved, reliable methods and are considered toc be accurate within i_B%.

147 91

The analyses for Pm and Y are thought tc be less reliable but,

155Eu was difficult;

nevertheless, accurate to within i_lO%, Analysis for
therefore, the results are only approximate and their accuracy is open
to question.lo
1Lk : .
The analyses for Ce, which were made by gamma scanning and by
radiochemical separation technigues using separate portions of the con-

densate samples, appeared to be accurate. Good agreement was obtained

between the two sets of analyses.
T

5. CONCLUSIONS

The following conclusions were drawn from the results obtained in

the experiment described above.

 

The separation of fuel carrier salt from the lanthanide
fission products was demonstrated by processing 12 liters

of fuel salt from an coperating reactor. Although the
volume of salt that was processed was less than anticipated,
gll the important featureé of the operation were adequately
tested. The operation of the equipment for this run was

smooth and trouble-~free.

The effective relative volatilities for BeFo, and ZrF)
(based on both natural zircenium and fission product 952r)
agreed with previous laboratory measurements.

lSTCS .

The upper limit of the relative volatility of
effective in this run, was found tc be only about 20% of
the value measured in the laboratory. We cannot adequately
explain this discrepancy in terms of the occurrence of
liquid entrainment, concentration polarization in the

still, or condensate sample contamination.

The effective relative volatilities for the lanthanide
figssion products lthe and 1MTPm were unexpectedly high.
The effective relative volatility cof lthe was about 56
times that indicated by previous measurements. No
previous measurements for lhTPm were agvailable; however,
the relative volatility 1s thought to be close to that of
CeF3. It is believed that the discrepancy between
observed values and previous measurements is, in part,

due to sample contamination.

Even if the relative volastilities of the lanthanides are
as high as seen here (v0.0l), adequate recovery of [LiF

from waste salt streams by distillation is still possible.
L8

If the relstive volatilities of the rare-earth fluorides
were as high as 0.01, only 3% of the rare earths would
be volatilized during vaporization of 95% of the TLiF

in a batch distillation.

6. ACKNOWLEDGMENTS

The authors gratefully acknowledge the help of the following pecple
in the installation, operation, and analysis of the MSRE Distillation
Experiment: P. N. Haubenreich, R. H. Guymon, P. H. Harley, A. I.
Krakoviak, and M. Richardson, of the Reactcr Division; R. W. Tucker of
the Instrumentation and Controls Division, R. B. Lindauer of the Chemical
Technology Division; J. H. Moneyhun of the Analytical Chemistry Division;
R. 8. Jackson of the Plant and Equipment Divisicon; and R. 0. Payne,

V. L. Fowler, J. Beams, F. L. Rogers, E. R. Johns, and J. C. Rose,
technicians in the Unit Operations Section of the Chemical Technology

Division.

 
kg

oo 7. REFERENCES

1. J. R. Hightower, Jr., and L. E. McNeese, Measurement of the Relative
Volgtilities of Fluorides of Ce, La, Pr, Nd, Sm, Eu, Ba, Sr, Y. and

Zr in Mixtures of LiF and BeF

P ORNL-TM-2058 (January 1968).

2. F. J. Smith, L. M. Ferris, and C. T. Thompson, Ligquid-Vavor Equilibris

 

in LiF=-BeF,. and LiF-BeF, ~ThF, Systems, ORNL-4L415 (June 1969).
< o i
3. W. L. Carter, R. B. Lindauer, and L. E. McNeese, Design of an

Engineering-Scale Vacuum Distillation Experiment for Molten Salt

 

Reactor Fuel, ORNL-TM-2213 (November 1968).

4. J. R. Hightower, Jr., and L. E. McNeese, Low-Pressure Distillation
of Molten Flucride Mixtures: Nonradioactive Tests for the MSRE
Distillation Experiment, ORNL-LL3Lk (January 1971).

5. M. W. Rosenthal, MSR Program Semiann. Progr. Rept. Feb. 28, 1967,
ORNL-L119, p. T6.

6. M. W. Rosenthal, MSR Program Semiann. Progr. Rept. Feb. 28, 1969,
ORNL-4396, p. 2k.

7. E. L. Compere, ORNL, personal communication, July 1, 1969.

8. M. W. Rosenthal, MSR Program Semiann. Progr. Rept. Feb. 29, 1968,
ORNL-L254, p. 100.

9. M. W. Rosenthal, MSR Program Semiann. Progr. Rept. Feb. 28, 1969,
ORNL-L4396, p. 1i5.

10. J. H. Moneyhun, ORNL, personal communication., Feb. 9, 1970.
 

 
51

.........

8. APPENDIX : ANALYSES OF SAMPLES FRCM THE MSRE DISTILLATION EXPERIMENT
 

 
53

Table A-1. Analyses of Samples from the MSRE Distillation Experiment

 

 

 

. Component, Salt Volumes Associated
(w6 %) (v %) (vt %) (ais min"lg™l) (ais min"lgl) (ais min-lgl) (ais min-lg-l) (ais min-lg=1) (dis min-lg-1) (ais min-Ye-1) (aie min-lgl) Fed  Collected

Fuel storage tank-1 — — — 1.79 x 109 3.1k x 1010 3,48 x 109 — — 3.11 x 107 — .
(Date analyzed) (h/2k/69) (k/25/69) (k/24/69) (L/24/69)

Fuel storage tank-2  10.6 5.67 13.39 1.62 x 109 3.21 x 1010 3,75 x 107 — — 3.35 x 107 — —
(Date analyzed) (5/1/69) (5/1/69) (5/1/69) (4/29/69) (5/2/69) (4/30/69) (4/29/69)

Condgnsate samples 3.88  11.1k 15.66 1.53 x 107 6.83 x 10°  <2.6 x 10° <8 x 103 6.5h x 10°  <2.1 x 10° 2.10 x 107 9.4 x 10° 7.9 0.42
-2 Y67 8.48 12.05 1.39 x 107 1.21 x 108 1.53 x 107 N1k x 102 1.18 x 108 <5.2 x 106 5.15 x 107 6.k9 x 108 10.0 1.93
-3 6.86 10.39 10.04 1.14 x 109 2.39 x 108 2,78 x 107 ~3.0 x 107 1.4h x 108 8.54% x 106 1.61 x 107 3.93 x 109 11.2 2.96
- 7.24 9.76 10.09 1.17 x 109 1.97 x 10° 2.23 x 107 A2.1 x 107 1.32 x 106 5.76 x 106 1.3% x 107 3.53 x 109 12.6 3.81
-5 8,15  10.00 10.48 1.24 x 107 8.09 x 107 8.85 x 10° A1.0 x 10° 6.14 x 10°  <2.5 x 10° 2.49 x 107 3.1k x 10% 13.% : 4.35
-6 7.05 9.93 10.49 1.21 x 107 2.89 x 10° 3.20 x 107 n3.L % 107 1.81 x 10° 8.93 x 108 1.08 x 107 .61 x 109 13.6 %.95
- 7.82 9.24 10.53 1.16 x 10° 1.99 % 108 2,16 x 107 n2.5 x 10° 1.15 x 106 8.57 x 100 1.91 x 107 1.66 x 107 13.8 5.80
-8 8.77 8.148 9.83 1.17 x 10° 3.01 x 10° 3.08 x 107 3.3 x 107 1.60 x 10° 8.75 x 106 3.38 x 107 5.91 x 108 13.8 6.3k
-9 9.80 T.66 8.71 9.78 x 108 4.63 x 108 5.98 x 107 n5.5 x 10° 2.29 x 106 1.56 x 107 3.59 x 107 4,06 x 108 13.8 6.89
-10 13.02 5.11 8.26 9.4k x 108 h.75 x 108 6.27 x 107 . 6.6 x 10° 6.01 x 106 2.37 x 107 1.69 x 107 2.63 x 108 13.8 7.%9
-11 13.28 5,20 8.11 9.25 x 100 .91 x 108 6.90 x 107 n8.1 x 10° k.10 x 108 3.04 x 107 .2 x 107 .96 x 108 13.8 7.85
(Date analyzed) (5/21/69) (5/23/69) (5/23/69) (5/21/69) (7/21/69) {7/2/69) (10/8/69) (9/30/69) (7/8/69) (7/8/69) (5/21/69)

 

aDuplicate samples differed by a factor of 4 or more.
 

 
55

UC-80 — Reactor Technology

INTERNAL DISTRIBUTION

1-3. Central Research Library 84-85. J. R. Hightower, Jr.
4-5. MSRP Director's Office 86. H. W. Hoffman
6. ORNL — Y-12 Technical Library 87. R. W. Horton
Document Reference Section 88. W. H. Jordan
7-41. Laboratory Records Department 89. P. R. Kasten
42. Laboratory Records, ORNL R.C. 90. C. W. Kee
43. R. K. Adams 91, M. J. Kelly
44, G. M. Adamson 92. 8. S. Kirslis
45, J. L. Anderson 93. J. W. Koger
46. C. F. Baes 94, R. B. Korsmeyer
47. C. E. Bamberger 95. A. 1. Krakoviak
48, C. J. Barton 96. T. S. Kress
49. H. F. Bauman 97. J. A. Lane
50. S. E. Beall 98. R. B. Lindauer
51. M. J. Bell 99. A. P. Litman
52, E. S. Bettis 100. M. I. Lundin
53. R. E. Blanco 101. H. G. MacPherson
54. F. F. Blankenship 102. 3. C. Mailen
55. J. 0. Blcomeke 103. H. E. McCoy
56. R. Blumberg 104, L. E. McNeese
57. E. G. Bohlmann 105. A. S. Meyer
58. G. E. Boyd 106. R. L. Moore
59. J. Braunstein 137. D. M. Moulton
60. M. A. Bredig 108, J. P. Nichols
6l. R. B. Briggs 109. E. L. Nicholson
62. S. Cantor 110. A. M. Perry
63. W. L. Carter 111. J. L. Redford
64. H. D. Cochran, Jr. 112. M. Richardson
65. E. L. Compere 113. G. D. Robbins
66. W. H. Cook 114. K. A. Romberger
67. B. Cox 115. J. Roth
68. J. L. Crowley 116. W. F. Schaffer
69. F. L. Culler 117. Dunlap Scott
70. J. H. DeVan 118. J. H. Shaffer
71. S. J. Ditto 119. M. J. Skinner
72. W. P. Eatherly 120. A. N. Smith
73. J. R. Engle 121, F. J. Smith
74. D. E. Ferguson 122. D. A. Sundberg
75. L. M. Ferris 123, R. E. Thoma
76. A. P, Fraas 124, D. B. Trauger
77. W. R. Grimes 125. Chia~Pao Tung
78. A. G. Grindell 126, W. E. Unger
79. R. H. Guymon 127. G. M. Watson
80. B. A. Hannaford 128. A. M. Weinberg
8l1. P. H. Harley 129, J. R. Weir
S 82. P. N. Haubenreich 130. M. E. Whatley
83. R. ¥. Hibbs 131. J. C. White
132.
133.
134,
135.

139.
140.
141-142,
143.
144,
145.
146.
147-149.

150-151.

152-375.

56

Gale Young 136. C. H. Tce (consultant)
E. L. Youngblood 137. E. A, Mason {(consultant)
P, H. Emmett (consultant) 138. R. B. Richards (consultant)

J. J. Katz (consultant)
EXTERNAL DISTRIBUTION

A. Giambusso, U.S. Atomic Energy Commission, Washington, D.C.
Kermit Laughon, AEC Site Representative, ORNL

T. W. McIntosh, U.S. Atomic Energy Commission, Washington, D.C.
M. Shaw, U.S. Atomic Energy Commission, Washington, D.C.

J. A. Swartout, Union Carbide Corporation, New York, N.Y. 10017
Laboratory and University Division, AEC, ORO

Patent Office, AEC, ORO

Director, Division of Reactor Licensing, U.S. Atomic Energy
Commission, Washington, D.C.

Director, Division of Reactor Standards, U.S. Atomic Energy
Commission, Washington, D.C.

Given distribution as shown in TID-4500 under Reactor Technology
category (25 copies — NTIS)