ORNL-TM-2578.txt

 

ORNL-TM-2578

Contract No. W-7405-eng-26

Chemical Technology Division

PROCESSING OF THE MSRE FLUSH AND FUEL SALTS

R. B. Lindauer

AUGUST 1969

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

i1ii

CONTENTS

1. Introduction

2, Summary .

3. Process Description .

3.1
3.2

4, Equipment Description and Performance .

.

O o0 N O 1 s W N

kb et
N = O

A B A B B DDA D B DDA B
[
W

.14

Fluorination .

Reduction

Plant Layout .
Shielding

Fuel Storage Tank
NaF Trap .
Caustic Scrubber .
Mist Filter
Soda-Lime Trap .
Charcoal Traps
Off-Gas Filter .
Gas Supply System
UF6 Absorbers
Salt Sampler .
Salt Filter

Special Instrumentation

5. Operating History .

.1

Ui g1 1o n»

2
3
.4
5
6

Fluorine Reactor Tests

Caustic Scrubber Tests
Fluorination of the Flush Salt
Reduction of the Flush Salt
Fluorination of the Fuel Salt
Reduction of the Fuel Salt

Page

o

o oo o Oy

13
15
17
17
17
20
21
22
22
25
27

30
30
39
41
43
44
47
iv

Discussion of Data .

.1

s N O8O OO OO
O oo 9 O U1 W N

Fluorine Utilization
Corrosion .

Molybdenum

Plutonium .

Niobium .

Iodine and Tellurium
Uranium Absorption
Uranium Product Purity

Reduction of Metal Fluorides

Acknowledgments

References

Page

50
50
53
56
57
57
62
64
66
70

72

74
PROCESSING OF THE MSRE FLUSH AND FUEL. SALTS

 

R. B. Lindauer

ABSTRACT

The MSRE Fuel Processing Plant, shakedown tests of equip-
ment and procedures, and the uranium recovery operation are
described. The MSRE flush and fuel salt batches were fluori-
nated to recover 6.5 and 216 kg of uranium, respectively.
Known losses during processing were less than 0.1%. Gross
beta and gamma decontamination factors of 1.2 x 109 and 8.6 x
108 were obtained. Corrosion averaged about. 0.1 mil/hr. The
corrosion product fluorides were reduced and filtered to pro-
vide a carrier salt having a lower concentration of metallic
contaminants than the original carrier salt,

1. INTRODUCTION

The Molten Salt Reactor Experiment (MSRE) is an 8-Mw circulating
liquid fuel reactor operating at 1200°F. The original fuel contained
65 mole % LiF,, 30 mole % BeF,, 5 mole % ZrF, and 0.9 mole % 238723°UF,.
The reactor first went critical on June 1, 1965, and began full-power
operation in May 1966. Prior to shutdown on March 26, 1968, the reactor
had operated for slightly more than.one equivalent full-power year, or
72,400 Mwhr.

The MSRE Fuel Processing Facility was constructed in a small cell in
the reactor building for two purposes: (1) to remove any accumulated
oxides in the fuel or flush salt by H2€HF sparging, and (2) to recover
the original uranium charge from the salt to permit the addition of the
233y fuel charge for the second phase of reactor operation. The plant
had been operated once beforé, in 1965, to remove oxide (115 ppm} from
the flush salt before the reactor went.critical.1 Since that time the
oxide contents of both the flush salt and the fuel salt have remained
constant; therefore, it appears unlikely that such an operation will be
required in the future.

A design and operations report describing the MSRE Fuel Processing

2 . . : . .
Plant™ was written before the initial test run; it was revised 1in
December 1967 to reflect the modifications being made to permit handling
of short-decayed fuel salt. The present report covers the modifications
made to the plant during the final plant testing operations and describes

the results of these tests and of the flush and fuel salt processing.

2. SUMMARY

The uranium recovery process consists of fluorine sparging to vola-
tilize the uranium, followed by decontamination of the gas stream with a
750°F NaF bed and absorption of the UFg on the 200°F NaF beds. The excess
fluorine is removed by an aqueous scrubber. The corrosion product fluo-
rides are reduced to the metals, which are filtered from the salt before
the salt is returned to the reactor system.

After four months of testing and modifications, during which the
fluorine disposal system was changed from the gas-phase reaction with SO,
to the aqueous KI-KOH scrubber, processing of the MSRE flush salt (66 mole
% LiF - 34 mole % BeF,) began. The flush salt was fluorinated to recover
6.5 kg of uranium. About 216 kg of uranium was. recovered from the MSRE
fuel salt. Corrosion of the fuel processing tank during fluorination of
the fuel salt averaged about 0.1 mil/hr. The most accurate calculations
were based on the corrosion of nickel (0.04 mil/hr) and chromium (0.14
mil/hr). Fluorine utilization averaged 7.7% during fluorination of the
flush salt @nd 39% during fluorination of the flush salt) and 39% during
fluorination of the fuel salt. Reduction and filtration produced carrier
salt containing less impurities than the original salt. The recovered
uranium was decontaminated from fission products by gross gamma and gross
beta decontamination factors of 8.6 x 10% and 1.2 x 10?, respectively.

Identifiable uranium losses were less than 0.1%.

5. PROCESS DESCRIPTION

3.1 Fluorination

The. flowsheet used is shown in Fig. 1. The.molten salt was forced

from the drain tank under pressure, through a freeze valve in the drain
ORNL—-DWG 68-—-8994

200°F NaoF ABSORBERS IN CUBICLE

 

 

 

 

 

  
   
 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SALT
SALT
CHARGING H'EF';'E%AY”
o Fv AFI-{I% FUEL
[ Yev PROCESSING W ACTIVATED CHARCOAL
& CELL o TRAP
SALT TO OR : )i
FROM DRAIN ; | L 0 VENT
: 1
?E,SKFS"USH FUEL | . MIST SYSTEM
TORAGE B FILTER L]
TANK | ¢ ] SODA ABSOLUTE
| LIME FILTER
L‘% : TRAP \
WASTE 750°F CAUSTIC CHARCOAL TRAP
SALT NaF BED NEUTRALIZER
TANK

Fig. 1. MSRE Fuel Processing System.
tank cell, through a metallic filter (back-flow) and another freeze valve
in the processing cell to the fuel storage tank. The transfer was made at
1000 to 1i00°F, a temperature that is well above the freezing point of the
sait but not hot enough to reduce the strength of the metallic filter ele-
ment below a safe limit. The transfer was made with the lowest possible
pressure in order to minimize stress on the filter element as the gas blew
through at the end of the transfer and also to minimize entrainment of
salt or activity from the fuel storage tank.

The salt was then cooled to within 50°F of the liquidus temperature to
minimize corrosion and fission product volatilization during fluorination.
The sample line was purged with helium to prevent condensation of UFg at
the cold upper end. The salt was sparged with either pure fluorine or a
fluorine-helium mixture at a relatively high flow rate (40 liters/min) to
cowvert the UF, to UFg. Fluorine utilization was high during this period.
When all the UF, had been converted to UFg, UFg began to form and vola-
tilize, as indicated by a temperature rise in the first absorber and by
a rise on the inlet mass flowmeter {see Sect. 4.14). When this occurred,
the fluorine flow was reduced to 15 or 25 liters/min to increase the
absorber residence time for more efficient absorption and to increase the
fluorine utilizatilon.

The gas leaving the fuel storage tank consisted of UFg, excess fluorine,
helium, MoF¢ and some CrF, or CrFc¢ from corrosion, IFy, and the fluorides
of some other fission products such as tellurium, niobium, ruthenium, and
antimony. The gas passed through a 750°F sodium fluoride trap where the
chromium fluoride and most of the volatilized fission products, except
iodine and tellurium, were retained. A small fluorine flow was introduced
upstream of this bed to ensure an excess of fluorine and to prevent any
nonvolatile UFg from forming and remaining on the heated NaF in the event
that the fluorine utilization was near 100%. Although the fluorine utili-
zation was not high enough to require this additional fluorine stream, a
small flow was maintained to prevent diffusion and condensation of UFg in
the line.

The gas stream, which now consisted of UFgz, fluorine, helium, MoFg,

IF-

;2

TeFy, and a trace of other fission products, left the shielded cell
and passed through five NaF absorbers in a sealed cubicle in the opera-
ting area. These absorbers were heated to 200 to 250°F to increase the
reaction rate and to minimize MoFg absorption. The UFg piping was heated
to at least 140°F to prevent condensation of UFg. As the UFg began to
load on a given absorber and the temperature started to rise, the cooling
air was turned on that particular absorber to limit the temperature to a
maximum of 350°F. High temperature reduced. the uranium loading by promot-
ing surface absorption and reducing penetration of the UFg to the inside
of the pellets. The final absorber was operated below 250°F, where the
partial pressure of UFg over the UFg-2 NaF complex allowed only a neglig-
ible amount of uranium to reach the caustic scrubber.

If the gas stream leaving the absorbers contained greater than 50%
fluorine, it was diluted with helium before reaching the caustic scrubber.
This allowed a smooth reaction in the scrubber without the production of
a flame or pressure oscillations. The caustic scrubber was charged with
1300 liters of 2 M KOH -- 0.33 M KI containing 0.2 M K;B,07, which was
added as a soluble neutron poison. The reaction that occurred in the

scrubber is:
6 F, + 12 KOH + KI » 12 XF + 6 Hy0 + 3/2 0, + KIOj.

The scrubber solution was replaced before one-half of the KOH had been con-
sumed, as determined by fluorine flow and calculated utilization, because
dip tube corrosion was increased when the OH concentration was less than

1 M. Besides fluorine, most of the molybdenum and iodine were removed in
the scrubber.

A high-surface-area filter located downstream of the scrubber removed
any particulate matter from the scrubber. During test runs, hydrated
oxides of molybdenum collected at sharp bends in the line from the scrubber.
A soda-lime trap (a mixture of sodium and calcium hydrates) provided a
means for detecting fluorine and a means for removing traces of fluorine
from the scrubber off-gas before it reached the charcoal absorbers. The
trap was operated at room temperature, and the temperatures were monitored.

Activated impregnated charcoal traps sorbed any iodine that was not
removed in the caustic scrubber. The gas leaving the charcoal traps con-

sisted of only helium and oxygen that was produced in the caustic
scrubber; the oxygen amounted to about 25% of the fluorine flow to the
scrubber. This gas flowed through a flame arrester and then through an
absolute filter before being routed to the cell exhaust. It was moni-
tored for gamma activity and iodine before being mixed with the rest of
the building exhaust gas, which passed through additional filters and was
finally discharged from a 100-ft-tall stack.

When there was no longer any evidence of absorber heating, the fluo-
rine flow was stopped and the salt was sampled and analyzed. The uranium
concentration in the salt was expected to be less than 50 ppm at this

time.

3.2 Reduction

Before the salt could be returned to the reactor system, the NiFj,
FeF,, and CrF, produced by corrosion of the Hastelloy-N fuel storage tank
had to be removed from the salt (MoFg is volatile). Because the concentra-
tion of nickel in Hastelloy-N is higher than that of chromium or iron,
the concentration of NiF, was higher than that of CrF, or FeF, in the fuel
salt. Since nickel is more noble than iron or chromium, it was reduced
by hydrogen sparging at 1225°F. After the reduction of NiF, was complete,
as indicated by filtered salt samples, pressed zirconium metal shavings
were added to the salt and hydrogen sparging was continued to reduce the
FeF, and CrF;. In this operation, the ZrF, concentration of the salt was
increased by a negligible amount (<1%). The reduced metals were removed
by a fibrous metal filter, and the efficiency of the filtration was veri-

fied by sampling the filtered salt.

4. EQUIPMENT DESCRIPTION AND PERFORMANCE
This section describes the performance of equipment not included in
the discussion in Sect. 6.
4.1 Plant Layout

Most of the processing equipment 1s located in the fuel processing
cell (Fig. 2), which is 13 x 13 x 17 ft deep and is situated just north

of the reactor drain tank cell. This processing cell contains the fuel
68576

PHOTO

Photograph of Fuel Processing Cell.

 

Fig. 2.

 

 

 
storage tank (fluorinator), the 750°F NaF trap, the caustic scrubber,
two remotely-operated air valves, three salt freeze valves, and the
blower for the absorber cubicle.

The cell to the east ("'spare cell') is also 13 x 13 x 17 ft and

contains the off-gas equipment -- mist filter, soda-lime trap, charcoal
traps, and off-gas filter -- which is located downstream of the caustic
scrubber.

The plant is operated from the high-bay area over these two cells.
This area contains the absorber cubicle, instrument cubicle, instrument
panelboard, sampler and sampler panelboard, hydrogen and oxygen monitors,
and radiation detection instruments.

The gas supply station, which contains the fluorine manifold (where
two 15,000-1liter fluorine tanks mounted on trailers can be connected),
the hydrogen manifold, and the pressure and flow instrumentation associ-
ated with these two gases, is situated outside the building, to the south-
west of the building. The helium for purging and sparging comes from the

reactor system supply.

4.2 Shielding

The shielding around the fuel processing cell was designed to limit
the radiation level to 5 mrad/hr when a fully irradiated, 3-day-decayed
fuel batch was in the fuel storage tank. To meet this requirement, new
roof plugs were installed, and the east and west walls of the cell were
backed up by an additional 2 ft of stacked barytes block.

Additional shielding was designed for 30-day-decayed fuel, but was
not installed because the plant testing required more time than expected.

This shielding consisted of:

(1) charcoal beds in the spare cell — to permit limited entry for main-
tenance;
(2) UFg absorbers in the operating area — to reduce the background activ-

ity from iodine and tellurium in the gas stream;
(3) an NaF trap in the fuel processing cell — to shield the electrical

insulation and valve diaphragms for absorbed 3°Nb.

The shielding used is summarized in Table 1.
Table 1. Shielding for the Fuel Processing System

 

 

Thickness
Location (in.) Material
Fuel processing cell
East wall 18 Normal concrete
24 High-D concrete
South wall 36 Normal concrete
12 FEigh-D concrete
West wall 12 Normal concrete
24 iigh-D concrete
North wall 18 Normal concrete
Roof 48 High-D concrete
Salt sampler 4 Lead

 

4,3 Fuel Storage Tank

The construction of the fuel storage tank, or fluorinator (see Fig. 3)
is similar to that of the reactor drain tanks, except that the fluorinator
is 30 in. taller to provide about 38% freeboard above the normal liquid
level. This extra height minimized salt carryover during gas sparging.
The tank is. suspended inside the heater and insulation assembly to mini-
mize the weigh cell tare weight. There is no provision for cooling since
the heat loss limits the temperature to less than 1200°F after two weeks'
decay (Figs. 4 and 5).

The fuel storage tank is heated by four sets of heaters, which provide
5.8 kw on the bottom, 11.6 kw on the lower sides, 5.8 kw on the upper
sides, and 2 kw on the top. Each group of heaters is controlled by a
separate powerstat. During the period in which UF4 was converted to UFg
(when the fluorine utilization was high), the heat of reaction, plus the
afterheat, provided sufficient heat to maintain the salt at 850°F. About
12 kw of electrical heat was required during the reduction operation at

1225°F,
10

ORNL-DWG 65-2509R

  
  

 

   

 

 

 

 

 

 

SANELER WEIGH
ULTRASONIC SALT ,/ CELL
LEVEL PROBE INLET '
TO
TRANSDUCER ~=—
SUPPORT
RING k
[
{ T
; L BAFFLE SUPPORT
/° i; ;I COLUMN
i
I |
b IF |
/// / |
i I
///D i “
6 in, / || |
INSULATION——-../° | |
7 |
/° I I
’/// I I LIQUID
/» }; I LEVEL
| oy
| | 7in,

A4

— HEATERS

v

*
! 62‘in.

\
N_i-in AIR

SPACE

g

O

Q

 

o
n_
u=

I
|
I
|
|
z
I
|
i
|
|
|
]
|
|
|
|
o I
| ] * 116 in.
l ]
|
[
t
|
|
f
|
|
|
l
|
|
|
|
|
i
|

 

 

R .

 

 

 

 

STABILIZER

 

 

fe—————— 50 in, —

Fig. 3. Fuel Storage Tank.
AFTER HEAT (kW)

11

ORNL—DWG 69—-877%9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

S0 z T T T T 7 1 1T T

z' | «w=59« MW | (DECAY TIME 0%~ (IRRAD+ DECAY TIME 702 ]
\ .~ .. CALCULATED FCR 8MW - —}—-} bt i
| i : i ! | | !
i | Lol ' ‘ ‘
) _-_\\R\[ o .
4 \ i ‘
| ! “:\\ ]

o - - | ~E oL T
: S T i T T
e Lo oIS ek
e e - Si 20C days |IRRADIATION -} &

‘ Lo . ; .

e A hE e (o
i L. {00 days IRRAD!ATION _ .

| ' o S

. do ] 1
. : t i I ; ‘ :

; i ‘ | . :
Ii | b \‘

o ! \ ’ | | 1 ’ ‘E i |
1 2 5 10 20 50 100 200 500

DECAY TIME {days)

Fig. 4.

MSRE After Heat.
HEAT LOSS (kW)

10

300

Fig. 5.

12

ORNL— DWG 6&95-6780

|

FLUSH SALT

 

 
 
  
 
  

BEFORE UFg
BREAKTHROUGH e FUEL SALT
: CoOL DOWN
® RUN NO.3
¢ RUN NO.4

 

e -—

500 700 900 1100 1300
FST TEMPERATURE (°F)

Fuel Storage Tank Heat Loss.
13

The gas inlet line for the fuel storage tank has a normal submergence
of 64 in., and the differential pressure between this line and the gas
space provides an indication of liquid level. The tank is equipped with
an ultrasonic probe (Sect. 4.14) to check the weigh cell calibration dur-
ing the filling and emptying operations.

Several safety interlocks are provided on the fuel storage tank:

1. The gas sparge line is automatically vented if the pressure in this
line decreases to that of the gas space. This prevents accidental
backup of molten salt in the cold gas inlet line.

2. An automatic valve will open in the tank off-gas line if the tank
pressure reaches 50 psig. An alarm sounds at 5 psig, which is a
pressure higher than that normally found during operation.

3. Fluorine flow to the tank is stopped if the helium purge flow down
the sampler line stops. This minimizes the amount of fluorine, UFg,
or radiocactive gases which could diffuse into lines leaving the cell
in the event that the helium supply is lost.

4. The fluorine supply valve is designed so that it must be closed before

the salt sampler valve can be opened.

4.4 NaF Trap

The NaF trap, shown in Fig. 6 is operated at 750 to 800°F. In this
temperature range, volatile ruthenium, niobium, antimony, and chromium
fluorides are absorbed, while uranium and molybdenum hexafluorides pass
through. The slightly oversized line from the fuel storage tank is also
heated to 750°F to prevent the absorption of uranium at the trap inlet.
The trap has two sets of heaters that are separately controlled, 2.25 kw
in the center pipe and 9 kw on the sides. There are three thermocouples
located in wells at the inlet, center, and outlet of the trap, respec-
tively, which are maintained at temperatures within 10°F of each other.
With 4 in. of insulation, the heat loss is about 1270 w at 775°F.

Before we had determined that most of the niobium left the salt
during reactor operation, we were concerned that the 3 x 10° curies of
95Nb in the fuel salt would cause overheating of the trap. With the

extended decay time, we calculated that 1 x 10° curies of 2°Nb would be
14

ORNL-DWG 65-2511R

      
  

LIFTING BAIL

 

3 LIFTING
COOLING
AIR FRAME

OOUTLET

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

INLET
T THERMOWELLS
THERMOWELL
[ OUTLET
A | | //l '
|| | I by
i || Iy | A
| | | | Iy ||
| || | LJ [
| o ,
Ly I |
| b | | |
I || I | |
|
| I
| | |
| e | I
18 in. | i | HEATING
| ¥ PIPE OPEN| |
| i1 AT TOP | |
| N I ONLY | I
l L] | | |
I | | I
| : | I
> BAFFLE I | !
| |
| |
I |
y L |
12in, 20in. »
I__czzzzééj

 

Fig. 6. Sodium Fluoride Trap.
15

in the salt at the time of processing. Apparently, however, most of the
niobium either plated out or was carried out in the reactor off-gas stream
since no detectable heating of the trap was detected during fluorination
and the center pipe air cooling system was not required,

We were also concerned that the trap might become plugged with vola-
tilized chromium fluorides. For this reason, the trap was designed to be
replaceable. The piping associated with the trap is flanged, with pro-
vision for remote leak detection, and the heater and thermcouple leads are
provided with disconnects. Because of the anticipated high radiation
level, a defective trap would be disconnected and moved to an empty space
within the cell instead of being removed from the cell. However, we did
not detect any increase in pressure drop between the fuel storage tank and
the absorber inlet. Chromium fluoride volatilization was expected as the
uranium concentration in the salt decreased; but such volatilization was
apparently almost insignificant until the concentration became as low as

20 ppm.

4.5 Caustic Scrubber

The caustic scrubber shown in Fig. 7 was used for disposal of excess
fluorine and of the HF produced during NiF, reduction. The Monel spray
ring distributor originally installed in the tank was replaced with two
dip lines, each with three 3/8-in. holes, as shown in the figure. The
dip lines have shutoff valves with extension handles to permit alternating
the lines when plugging occurs. The new dip lines are made of Incomnel,
which has a higher corrosion resistance (than Monel) to the KOH-KI-F, re-
action. The scrubber and other internal piping were originally constructed
of Inconel. The cooling water to the coil is not metered; however, only
a fraction of the full flow was required to control the temperature during
fluorination. On one occasion, in which the temperature was allowed to
increase from less than 80°F to greater than 100°F, no decrease in plugging
was noted.

The caustic scrubber was equipped with a contact microphone similar
to the one used for the fuel storage tank. It was used intermittently as
a check on the gas flow rate through the dip lines. During a 2-hr period

in run 6, unusually loud, rapid noises, accompanied by rapid but small
ORNL-DWG 65-2513R

THERMOWELL

16

 

 

< <
-
. T O
% o =
o
o s,
UVB
QEB_.HLV
=45
-~ m

 

 

CAUSTIC
CHARGING
LINE

 

O 0N SO0 O U O~ 0~ o5 ‘ H_

 

         

Y e e e o e m——— . — e e e e —— o e — e — e e

 

 

    

 

 

 

 

 

 

 

   

 

o T T TR T T
w7 771 e ST
5 = g~ ——— ] e R
w llllllll —— it — — —— ——— Tt — o ——r - — —— — —_— —_———— — — —— — — s i
o
q
95 -\ Y S o T T T T T .- - T IIT|ZT
© z
3 S S O O g SN NN Y [ 4 LS
= : : ° \O
O
T3 o « o o
Zaxh s > o+ D
|E£ a = wu
I ul Jp=
= wZ =
SF> Td2 g5z
CWO .mlmc QO

Caustic Scrubber.

Fig. 7.
17

pressure fluctuations in the line to the scrubber, were heard. Switching

dip tubes restored the normal bubble flow sound.

4.6 Mist Filter

The mist filter is a high-efficiency Fiberglas air filter consisting
of fifty-five 9-1/2-in.-diam by 28-in.-long bags hanging vertically in a
30-in.-diam by 4-ft-high stainless steel vessel. The total filter area 1s
about 50 ft?. The filter, which is mounted in an opening in a horizontal
support plate near the top of the vessel, was caulked to prevent bypassing.
The tank is heated with steam coils that are wrapped around the tank to
prevent condensation of the moisture that is carried over from the caustic
scrubber. A thermowell installed in the gas space above the filter was
used to adjust the amount of insulation in order to prevent overheating

and damage to the filter binder.

4.7 Soda-Lime Trap

The soda-lime trap, shown in Fig. 8, was charged with activated
alumina when it was used as the fluorine disposal backup trap for the SO,
system. When this disposal method was abandoned in favor of the aqueous
scrubber, the trap was recharged with soda-lime, which is very effective
for the removal of fluorine at room temperature but cannot be used with
SO,. To prevent excessive pressure drop with a gas having a high fluorine
content, the soda-lime was mixed with an equal volume of 4- to 8-mesh
alumina, which has a very slow reaction rate at low temperatures. To pre-
vent condensation of water vapor in the gas from the scrubber, the soda-
lime trap was insulated and traced with low-temperature steam tubing.

Thermocouples were installed near the inlet and near the outlet.
During the processing of the flush and fuel salts, no temperature rise
was observed. Thus it appears that no unreacted fluorine escaped from the

caustic scrubber.

4.8 Charcoal Traps

The charcoal traps shown in Fig. 9 were considerably overdesigned to
permit the handling of the large amount of iodine that is present in 30-day-

decayed fuel, and to allow for the uncertainties associated with a gas
18

ORNL-DWG 68-2730A

 

 

 

 

 

 

INLET -z-c-215 OQUTLET
e
CHARGING
FLANGE
pa il s THERMOWELLS
30in,
6-in. \‘\\‘Q
INCONEL 4
PIPE O
3
Y

 

Fig. 8. Soda-Lime Trap.
19

ORNL-DWG 68-13777

GAS
INLET

]

 

 

 

  
      
  

CHARCOAL-
BARNEBY-CHENEY
NO. 727

 

GAS
OUTLET

 

NICKEL

WOOL\\\

 

NICKEL
WOOL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

~ 24 in,

Fig. 9. Charcoal Traps.
20

stream containing SO, and SO,F,. The original trap (inlet) was initially
charged with replaceable cannisters, each of which had a total bed depth
of 1-1/2 in. The final capacities and residence times are shown in Table

2.

Table 2. Charcoal Traps

 

Residence Time at Average Flow

 

Diameter Length Volume of 16 std liters/min
Trap (in.) (in.) (liters) (sec)
Inlet 6 12 5.84 22
Backup 6 22 3.17 12
Total 34 9.01 34

 

The traps are steam-traced in the same manner as the soda-lime trap.
Surface thermocouples (TE) are located at the inlet of the first trap and
at the inlet and the outlet of the backup trap. Temperatures were main-
tained between 175 and 205°F during processing, and no additional heating
was detected at any time.

Four ion chambers, located at the inlet and the outlet of each trap
were used to detect the accumulation of iodine; however, because the fuel
had been allowed to decay for such a lengthy period, none of the readings
increased during processing. Five chambers (CT-A-E) as shown in Fig. 9,

were used during the fluorine reactor tests described in Sect. 5.1.

4.9 0£ff-Gas Filter

The fuel processing cell ventilation air and the gas stream from the
charcoal traps pass through a 2-in.-deep, 24- by 24-in. Fiberglas pre-
filter and a 11-1/2-in.-deep, 24- by 24-in. Fiberglas absolute filter
before passing through the reactor system containment filters and the
containment stack. There are three 2-in.-diam butterfly valves for iso-

lation and for bypassing the filters during replacement of the latter.
21

Since these valves do not have soft seats, it was necessary to blank the
bypass valve in order to obtain a satisfactory DOP test.

A locally mounted differential-pressure transmitter indicates the
pressure drop across the filters on the fuel-processing system panelboard.
During the processing described in this report, the differential pressure
varied between 1/2 and 1 in. H,0. This variation was probably the result

of variable flow rates rather than a buildup of solids on the filter.

4.10 Gas Supply System

The fluorine and hydrogen manifolds and associated instrumentation
are located outside the building. Space and piping are available to
connect two 15,000-1iter fluorine tanks mounted on trailers simultaneously.
Each tank holds sufficient fluorine to volatilize the uranium that would
normally be loaded on one group of absorbers; piping is provided for
purging the connecting piping before and after a tank is used. The tanks
can be initially loaded with fluorine under a pressure of 70 psig. This
pressure is reduced to 18 psig by a throttling valve that is controlled
automatically from the operating panelboard. The fluorine then passes
through a NaF trap for removal of HF before being metered through an
orifice. Another throttling valve regulates the flow rate of fluorine
to the fuel storage tank. The fluorine flow downstream of the fuel
storage tank is regulated by a manual throttling valve and measured by
an integral-orifice flowmeter.

The hydrogen manifold consists of two banks of two cylinders each.
The manifold has a two-stage pressure regulator containing an interstage
pressure relief valve. This pressure regulator is set to provide a con-
stant 15-psig pressure upstream of the flow control valve (which is con-
trolled from the panelboard). The flow is measured by an orifice meter.
High-purity hydrogen was used to prevent oxide contamination of the salt
during the reduction step. Spark-proof tools were used for changing the
cylinders, and joints were leak-tested after each change.

Helium is obtained from the reactor system helium supply, which
consists of a 39,000-ft3 tank mounted on a trailer. The 2400 psig
pressure is reduced to 20 psig at the fuel processing system panelboard.

The helium is analyzed for moisture and oxygen before and during usage.
22

4.11 UF6 Absorbers

Five absorbers (see Fig. 10), connected in series, are located in a
9 ft 4 in. x 2 ft 4 in. x 2 ft 2 in. sealed cubicle located in the opera-
ting area above the processing cell. The cubicle is vented to the cell,
where a blower is located for providing additional negative pressure
(below the -0.3 in. Hy0 in the cell) when air is being used to cool the
absorbers. This blower failed during fuel salt run 1 (Sect. 5.5). When
loaded to within 1/2 in. of the top, each absorber holds about 25 kg of
NaF. The minimum cross-sectional area (outlet side) is 0.46 ft?. The
inlet and the outlet Hastings mass flowmeters (Sect. 4.14) are also loca-
ted in the cubicle.

Each absorber is mounted in an insulated can with a removable in-
sulated cover. An air cooling coil and an electric heater are positioned
below each absorber. A separate thermocouple on the bottom plate of each
heater indicates overheating or burnout of a heating element. Each
heater was used to heat each absorber to about 200°F before the start of
UFg absorption. When absorption began, the heat was turned off and air
was turned on.

All absorber piping located upstream of the final absorber is heated
and insulated to prevent condensation of UFg. The absorbers are connected
with removable jumpers having ring-joint flanges. The flanges are leak-
tested before the absorber cubicle is sealed. Copper rings are used, and
a torque wrench is employed to ensure that the pressure on the rings is

uniform.

4,12 Salt Sampler

The fuel storage tank is provided with a salt sampler similar to the
sampler-enricher used with the reactor fuel pump (see Fig. 11). The
sampler, which is mounted on a semipermanent roof plug near the instrument
panelboard, is connected to the tank by a vertical 1-1/2-in. pipe. A
separate instrument panelboard for the sampler is located in the same
area. During processing, the sampler line was purged by a minimum of 5
std liters of helium per minute, mainly to prevent the condensation of

UFg on the cold surfaces. Two thermocouples on the sampler line showed
23

ORNL-DWG 65-2514R

 
 
  
  

THERMOWELL

MATERIAL:
CARBON STEEL

AIR COOLING PIPE
OPEN AT TOP & BOTTOM

INLET OUTLET
— _— / REMOVABLE

COVER

 

 

 

 

 

LA WAL Wiy

T B
;:

A

 

 

 

4in,

 

|
|
|
I
|
| 12in.
l
|
|
|

“
I
|
|
|
|
|

INSULATIQN _

 

 

BAFFLE

 

 

i

l

|
o
| |
| |
|
o
|
| |
| |
| |
| l
| |
| |
| |

\
O

— ) . = COOLING

AAAAAANAVMNAAANNAN AIR

14 in.

 

 

Y

J

 

 

 

ooy

 

- 7 AN
ELECTRIC HEATER/ \THERMOCOUPLE

Fig. 10, UF6 Absorbers.
24

ORNL-DWG 63-5848R

_—REMOVAL VALVE AND

- ’,'-‘,. 7 E,/\)}\ .
i @K} SHAFT SEAL
S 17“ PERISCOPE

v~ LIGHT

  
    
  
  
 
  
 
   

CAPSULE DRIVE UNIT —i.:

   

 

 

LATCH..|" 7 _CASTLE JOINT (SHIELDED

i WITH DEPLETED URANIUM)

 

 

ACCESS PORT~-1l /71

AREA IC s
(PRIMARY CONTAINMENT) -1

 

 

 

SAMPLE CAPSULE -~ T~

CONTAINMENT)

SAMPLE TRANSPORT
CONTAINER

  
 
   
 
  
   
 

 

 

OPERATIONAL AND £
MAINTENANCE VALVES 7545

SPRING CLAMP

p AREA 2B (SECONDARY
DISCONNECT -~

CONTAINMENT)
852" ELEVATION

 

 

 

CRITICAL CLOSURES
REQUIRING A BUFFERED
SEAL

 

SAMPLER-ENRICHER SCHEMATIC

    
   
  

 

 

 

TRANSFER TUBE
{PRIMARY CONTAINMENT)

LATCH STOP

Y - MIST SHIELD
CAPSULE GLIDE

Fig. 11. MSRE Sampler-Enricher.
25

temperatures of 550°F and 130°F, respectively, at positions 3 ft above
the tank and 6 ft above the tank (near the bottom of the roof plugs).
Under the conditions prevalent during the processing of the fuel salt,
condensation of UFg would not occur above a temperature of 115°F, which
would probably be near the temperature of the isolation valve in the
operating area.

The sampler was used for 33 sampling operations and for charging
zirconium capsules for the structural metal fluoride reduction. The only
difficulties experienced involved the filter capsules; the standard "dip"
samples were taken reliably. On some occasions, the vacuum in the filter
capsules was lost due to premature melting of the frozen salt plugs which
were designed to retain the vacuum. When the temperature differential
between the salt and the freezing point was too low, the capsules failed
to provide sufficient filtered salt.

A radiation detector was mounted on the sample line near the isola-
tion valve below the sampler. It was set to alarm when the radiation

level reached 100 mr/hr.

4,13 Salt Filter

The salt filter is installed in the fuel processing cell between the
cell wall and the freeze valve for the fuel storage tank. The replaceable
filter element, which rests on a spherical seal, is mounted in a heated,
insulated 6-in.-diam, 7-ft-long pipe. The 4-ft-long filter consists of
two concentric fibrous metal cylinders with a total filter area of 8 65
ft?. The salt enters the side of the filter housing, just above the
filter element and leaves through a bottom outlet (see Fig. 12). The
upper 3 ft of the filter housing is normally unheated and provides a
transition zone from the molten salt to the cool ring joint seal flange,
which contains a continuously pressured leak monitor. A helium purge
connection at the top is normally used to purge the filter of air; however,
it can also be used for blowing out restrictions in the connecting piping
if necessary. Baffles are included in the gas space to prevent salt from
splattering on the ring joint seal.

The center of the filter element is a sealed can to limit the volume

of salt runback that is unavailable for processing (0.56 ft3) on transfer
26

ORNL-DWG 68-13779

 

LEAK DETECTOR RING JOINT
PRESSURE~\\\ FLANGES
o

  

 

 

 

 

L TE-8

BAFFLE
\ bTE-7

UPPER

HEATED
ZONE PTE-6

SPACER b TE-5
(3 FINS EQUALLY SPACED) ——__
LIFTING BAIL FOR CTE-4

REPLACEABLE FILTER ELEMENT

 

pTE-3
SALT FROM FUEL

STORAGE TANK

 

SPHERICAL SEAT—

FELTMETAL
FILTER ELEMENTS

 

LOWER
HEATER
ZONE

pTE-1

 

 

 

 

 

 

 

 

 

 

SALT TO
DRAIN TANK

Fig. 12. Salt Filter.
27

to the fuel storage tank. This volume also reduces the weight of the
element when it is submerged in salt and prevents the filter from bursting
when salt 1s transferred from the drain tank through a restricted filter
element. The element will 1ift off the seat when a salt pressure of about
1 psi (depending on salt density) is applied under the seat. The outer
element has a burst strength of 25 psi at 1200°F. A possible filter col-
lapse when the full gas pressure is applied to the filter at the end of a
transfer is prevented by a perforated Inconel backup plate placed on the
inside of each filter cylinder. This plate provides a collapse strength
of 131 psi, which is well above the maximum transfer pressure of about

30 psi. More details on the filter design can be found in ref. 3.

During filtration of the flush and fuel salts, no pressure drop across
the filter was detected, even though salts containing about 10 kg of re-
duced metals were filtered. Filtration time was about 2 hr for each batch.
Later, 2 ft3 of fuel salt was returned to the processing tank to be used
for a salt distillation experiment. A sample of this salt, after sparging,

revealed that much of the reduced metals had remained in the tank.

4.14 Special Instrumentation

In addition to the standard orifice flowmeters and temperature,
pressure, liquid level, and radiation instruments, several special instru-
ments were used during the processing of the flush and fuel salt. These
instruments are described below.

Hastings Mass Flowmeters. — These instruments, which indicate the

 

product of heat capacity and mass flow rate of the gas stream, were in-
stalled upstream and downstream of the absorber chain in order to provide
a sensitive indication of the UFg (which has a high heat capacity) concen-
tration in the gas stream. Two detectors (0 to 2 and 0 to 10 equivalent
cfm of air) were required upstream to obtain both range and sensitivity,
while one 0- to 2-cfm detector was required downstream. To prevent plug-
ging of the 0.03-in. capillaries by NaF particles, the primary elements
are protected by 36-u Feltmetal filters. The two inlet meters are heated
by a heating tape (that is controlled separately from the rest of the gas
line heaters) to prevent the temperature on the electrical insulation from

exceeding 200°F. The readings are not affected by temperature changes.
28

These flow meters agreed within +4% of the helium rotameters when
only helium was present. However, unaccountably high readings were
obtained in the last three runs, especially on the exit meter from the
absorbers. At the end of run 6, the exit meter indicated a MoFg (which
also has a high heat capacity) rate corresponding to a molybdenum corrosion
rate of almost 1 mil/hr, while the inlet meter throughout the run indi-
cated a corrosion rate only slightly greater than the 0.2-mil/hr rate
assumed for molybdenum. Soon after the fluorine flow was stopped, both
meters agreed closely with the helium purge flow.

These meters also provided a sensitive indication.of incipient plug-
ging in the caustic scrubber dip lines.

The UFg flow at a given time could be calculated from the following
formula, assuming a MoFg flow rate equivalent to a molybdenum corrosion

rate of 0.2 mil/hr:

UFg flow x A = Absorber Inlet Meter Reading - (F, feed rate x B —
1/2 UFg flow)

 

 

or

_ Inlet Meter Reading — 0.0362 x F, feed rate,
UFg flow = 0.1525
where A = —>eter reading = 0.1706,

std liters of UFg/min

_ __Meter reading
std liters of Fo/min

 

0.0362, and the

inlet meter is corrected for helium, nitrogen, MoFg, and fluorine intro-
duced downstream of the fluorinator. Fluorine utilizations calculated
from these UFg flows near the end of each run were higher than those cal-
culated from absorber weight increase because of the buildup of dense

UFg in the fuel storage tank gas space.

Hydrogen Monitor. — To verify that the system has been completely

 

purged after hydrogen reduction, a hydrogen monitor was installed to
analyze a sample of the gas downstream of the charcoal beds. A commercial
instrument, which determines the partial pressure of hydrogen by measuring
the pressure inside a palladium-silver alloy membrane, was used. The

sample line consists of about 50 ft of 1/4-in. tubing, which provides a
29

flow of about 100 cc/min through the instrument with the differential
pressure available from the charcoal traps to the ventilation header in
the operating area. Because of the long line and low flow rate, the over-
all response time 1is several minutes.

Oxygen Monitor. — The oxygen monitor,. a commercial instrument based on
the paramagnetism of oxygen, was valved into the same sample line on which
the hydrogen monitor was located. This instrument was installed, prima-
rily to determine that oxygen was completely purged from the system before
the salt was treated with H,-HF to effect oxide removal. In the flush and
fuel salt processing operations described in this report, the hydrogen
reduction step immediately followed the fluorination step; therefore there
was little danger of forming an explosive gas mixture.

Since oxygen was produced in the fluorine disposal reaction (equiva-
lent to about one-fourth of the reacted fluorine), it was hoped that the
oxygen monitor could provide an indirect measure of fluorine utilization.
However, during the fluorination of the flush salt, the relatively low
fluorine utilization caused the oxygen that was produced to exceed the
range of the instrument (10%). A malfunction of the instrument prevented
its use during fuel processing.

Ultrasonic Level Probe. — The fuel storage tank contains an ultrasonic

 

level probe to provide a single point check on the weigh cell calibration.
This probe has a 1/2-in.-diam sensing element, which is better suited to
the corrosive conditions during fluorination than the thin-walled conduc-
tivity-type probes used in the reactor drain tanks. The ultrasonic energy
is transmitted through two force-insensitive mounts, one on the tank and
another at the cell wall penetration, to the ultrasonic generator, trans-
mitter, and receiver located just west of the fuel processing cell. The
generator produces a signal in resonance with the system. When the salt
contacted the probe, it caused a decrease in the reflected signal, thus
operating a relay switch.

Difficulty was experienced with the drifting of the signal frequency
away from the resonant peak; this. resulted in a signal even though contact
with salt had not been made. A similar difficulty had been experienced in

1965 during the flush salt treatment; in that instance, a satisfactory
30

signal was obtained when the salt was transferred to the tank but not

when it left the tank. In 1966, the stability of the generator was im-
proved by rewiring the unit with high-stability components. During the
recent processing opeation, a signal was again obtained when the flush
salt was transferred to the fuel storage tank, but not when the salt left
the tank. No signals were obtained for the fuel salt. For proper opera-
tion, the generator would probably require a constant temperature environ-
ment or a much broader signal output.

The salt weights obtained at probe contact are compared in Table 3.

Table 3. Performance of the Ultrasonic Level Probe

 

Salt Weight at Probe Contact (1lb)

 

 

Year Weigh Cell Bubbler Instrument
1965 7880 Readings not taken
1968 6550 7950

 

“The weigh cell readings taken in 1968 were unreliable. They were con-
sistently 5 to 10% lower than the bubbler instrument readings and also
the comparable readings in the reactor drain tanks.

5., OPERATING HISTORY
5.1 Fluorine Reactor Tests

In April 1968 (about two weeks after the reactor shutdown), the
fluorine supply and excess fluorine disposal systems, which had not pre-
viously been operated, were tested. According to the original design of
the disposal system, the excess fluorine was expected to react with sul-
fur dioxide in a fluorine reactor (Fig. 13) at temperatures above 400°F
to form sulfuryl fluoride, SO,F,, a relatively inert gas that could be
safely passed through Fiberglas filters and discharged to the atmosphere.
A backup alumina bed, originally unheated, had been installed between the
fluorine reactor and. the charcoal traps to detect and to remove traces of

unreacted fluorine.
31

ORNL-DWG 65-2545R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SO, F,
OUTLET
5-in. &
MONEL PIPE
N _—2-in. PIPE
30in. o D
22
THERMOWELLS o 17 HEAT
' ==:]1 p TRANSFER 18in.
| = RODS
00 in. STEAM s =13
OUTLET |3 ":‘ <>
Wr‘:% =
20in. /%
2=y ¥ f
44 in. ?
THERMOWELLS
SO, AND F,
1.:.-:’/ PREHEATERS
Hln
STEAM
INLET EIGHT %g-in. DISTRIBUTOR
HOLES FOR F,
' Fp INLET
SO, INLET
F, REACTOR

Fig. 13. Fluorine Reactor and Preheaters.
32

Results of four fluorine disposal test runs are summarized in Table 4
and Fig. 14. The first run was started on April 2, but was interrupted
after about 1-1/2 hr by fluorine ignition of two Teflon plug valves. A
total of 900 liters of fluorine had been used; however, the flow rate was
very erratic because of the difficulty encountered in blending a low fluo-
rine flow rate (v7.5 std liters/min) with a high helium flow rate (85 std
liters/min) to obtain dilute fluorine for conditioning the system. During
the time that the fluorine was flowing, there was no evidence of a reaction
in the fluorine reactor although the alumina bed was heated higher than
250°F. We believe that SO, had condensed in the supply line because of
excessive back pressure from check valves. After the flow of SO, and
fluorine was stopped, the temperature of the fluorine reactor rose to
520°F. This temperature rise was probably caused by the evaporation of
SO,, and its subsequent flow through the leaky check valves and reaction
with fluorine being purged from the fuel storage tank.

The test was resumed three days later, after the plug valves had been
replaced. When the helium purge was started, the temperature of the
alumina trap again rose as the result of residual fluorine in the system.
A reactlion was also observed in the fluorine reactor, probably from a
small amount of SO, remaining in the system. When the SO, and fluorine
flows were started, the reactor temperature reading went off scale
(>1000°F) because of insufficient cooling from low-pressure steam. The
alumina trap temperatures decreased, probably because the bed was ex-
hausted and because some of the fluroine had reacted with the charcoal
traps, causing an increase in pressure drop in the system.

The first test indicated the need for the following:

(1) Finer-mesh alumina and a heated alumina bed to increase the capacity
and the reaction rate; additional thermocouples to permit the con-
sumption of fluorine in the bed to be followed.

(2) An SO, flow until all fluorine has been purged from the system.

(3) A greater excess of SO, to compensate for poor flow control

(4) Increased steam pressure on the fluorine reactor cooling coil for
greater cooling capacity.

(5) Check valves with lower pressure differential in the S0, supply line.
 

 

 

 

 

Table 4. Summary of Fluorine Reactor Test Runs
Fluorine Flow Fluorine
Aver. Reactor Alumina Trap
Liters Max. Steam Max. Bed
Time per Temp. Pressure Temp . Pene- Remarks and
Run Date (min) Liters (min) (°F) {psi) (°F) tration Material Modifications Made After Run
1A 4/2 100 900 9 520 10 >250 ~n1/2 1/4-in. balls Run interrupted by ignition of two Teflon
plug valves,
1B 4/5 210 5100 24 >1000 10 >250 To CTa 1/4-in. balls AATb reaction too slow; replaced AATb with
finer AA and installed heaters and more
thermocouples. Increased steam pressure for
better FLR cooling. Improved insulation in-
stalled on SO, and F, preheaters and lines.
Replaced 50, check valves,
2 5/1 130 2650 20 ~1000 S0 >1000 nZf3 Inlet 2/3, System performed sufficiently well to permit
35 8-14 mesh; testing of iodine retention on charcoal in
40 outlet 1/3; next run. Improved S0, Cylinder heating.
50 60-100 mesh.
3 5/3 160 3900 24 >1000 35 or >1000 To CT® Inlet 2/3, Iodine retention satisfactory. AATb heating
50 8-14 mesh; major problem; preheaters added to AATP and
outlet 1/3, FLRC inlets. FLR® steam coil bonding dam-
60-100 mesh, aged; added Thermond bypassed FST for next
run,
4A 5/20 102 7 ? >1000 50-30 >1500 n1/2 Inlet 2/3, Run suspended when AAT temperature exceeded
-10 8-14 mesh; 1500°F, near end of fluorine conditioning,
outlet 1/3, to allow cooling and readjustment of all
60~100 mesh. temperatures.
4B 5/21 202 4400 20 1440 0-25 1480 To CT? Inlet 2/3, Fluorine reactor operation good, but alumi-
8-14 mesh; na trap heating continued. Piping to be
outlet 1/3, temporary modified for caustic scrubber
60-100 mesh, test,
4%CT = charcoal traps. “FLR = fluorine reactor.
bAAT = activated alumina trap. dHeat transfer cement.

e
TEMPERATURE (°F)

34

QRNL-DWG 69-5399

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 
  

 

 

 

 

1600 ' : } OFF SCALE
1 | TN\
1200 -} I+ L orr A 1 l_j ‘ /) 7
. j R 5
800 |— - e S S 4 \ 13
g * 5. , |
<rolFF OFF I |
400 I~ cene | [ scate T ] j '
2 N ! 2 ‘
o [ ! J ?\T , L]
1600 I : | | : :
| : ! i ;
: , i ’ f
1200 1 bed—t = oFF — | I T —F> - OFF -t \\1\!\:’!
% 1 . SCALE3 42 Lo *! SCALE \2 -T
f o I\ : N
el L N e A N
i AN T
i [ i ! i 1‘4\
300 — 3 4 p— | 1—+——

3T [T 2
| \

 

n-—1-—

F. | F

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

|
lFy: F
eyl | ||—T—"’-—| e

 

- . T i
0
15 46 17 1814 15 16 17 18 19 20 2123 24 04 020316 17 8 19 20 2113 14 15 16 17 %4 15 16 17 18 19 20
4-2-68 4-5—-68 5-1-68 5-3-68 5-20-68 5-24-68
RUN 1A B 2 3 4A 48

Fig. 14. Temperatures During Fluorine Reactor Tests.
35

The above equipment and procedural changes were effected, and the
charcoal cannisters (in the inlet charcoal bed), which had reacted with
fluorine, were replaced. Test run 2 was then started on May 1, 1968.
About 1/2 hr after the start of fluorine flow, the reactor inlet tempera-
ture started to rise, indicating a reaction; however, temperature points
16 in. and 32 in. downstream but still in the reaction zone)} showed prac-
tically no heating. It is believed that the increased flow of steam
overcooled the reaction zone, thereby causing the reaction rate to de-
crease and allowing unreacted fluorine to reach the alumina trap. The
alumina trap was being heated at the rate of about 80°F/hr, but the
temperature at a point 8 in, down the bed rose sharply, from 400°F to
>1000°F in less than 15 min, about 2 hr after the start of fluorine flow.
The fluorine flow was then stopped, but the SO, flow was continued while
the system was being purged with helium. Two hours later, the temperature
measured at points farther down the alumina bed rose from about 800°F to
>1000°F. Since the fluorine flow rate to the reactor should have been
low at this time, it appears that the second temperature rise was due to
an SO,F, reaction. (Later, such a reaction was shown to occur above 800°F.)
However, because at this time we were unaware of the SO,F,-alumina reac-
tion, efforts were directed to improving SO, and F, flow control.

One of the main objectives of the fluorine disposal tests was to
demonstrate that iodine could be retained on charcoal when SO; and SO;F,
were flowing through the beds. In the event that this could not be demon-
strated, an alternative fluorine disposal method would have to be found.
Therefore, although operation of the system was still not satisfactory, as
shown by the tests described above, it was felt that the alumina trap had
sufficient residual capacity to remove any unreacted fluorine and that
iodine would be retained on the charcoal traps.

In the third run, a satisfactory heat of reaction was obtained in the
fluorine reactor; this indicated the production of SO,F,. Since heating
was observed at the center of the alumina bed, But.not at the exit, it
appeared that no fluorine was leaving the bed. About 100 min after the
start of fluorine flow, 30 mc of CH313II was 1njected into the gas stream

upstream of the charcoal traps. The loading of the iodide was followed
36

by determination of the radiation levels with radiation monitors (see
Fig. 15). Although fluorine flow was stopped 1 hr after the start of
iodide injection, the fluorine reactor temperature remained above 400°F
for 4 additional hours (from purged fluorine). After the first 72 min

of the 2—hrkiodide injection period, the summation of the levels indi-
cated by the five radiation monitors on the charcoal traps reached a maxi-
mum and leveled off, indicating all the iodide had been loaded. During
the third and fourth hour, there was a very slight increase in radiation
at points CT-D and CT-E (see Fig. 9) and a corresponding reduction at
CT-C. During this time, the temperature at the outlet of the alumina
exceeded 1000°F, and there was some heating of the charcoal, indicating

a reaction with fluorine (probably from SO,F, decomposition}. The slight
shift in the location of the iodide could have been caused by fluorine
rather than SO,F,. The almost immediate penetration of some of the
iodide through the first 12 in. of charcoal was not unexpected, as this
type of behavior has been noted previously for organic iodides.

After the third run, the insulation was removed from the fluorine
reactor to allow the cooling coils to be examined. [The copper coils
were bonded to the Monel reactor with E-Z flow solder (flow point, 1328°F)
containing cadmium. At temperatures above 1300°F, alloying, along with
the formation of cracks and weakening of the metal, can occur.] In runs
1 and 3, the thermowell temperatures exceeded 1000°F (maximum recorder
temperature) for 3-1/2 hr and for 40 min, respectively. Examination re-
vealed that the solder had melted, leaving the middle third of the 44-in.
coil unbonded. Dye-checking revealed some cracks in the Monel. However,
a pressure test showed that the reactor still had adequate strength to
permit the remaining test runs to be made. A high-temperature heat
transfer cement was applied around the coils, and the reactor was rein-
sulated.

Run 4A was started on May 20, 1968. A preheater had been installed
on the fluorine reactor inlet to provide inlet temperatures above 400°F
to ensure a reaction at low fluorine flows as well as at the start of
fluorine flow. A preheater had also been installed at the alumina trap
inlet to obtain more satisfactory temperature distribution and to improve

the efficiency at the bed inlet (in run 3 no reaction occurred in the
10

10

mR/hr

10

10

Fig.

37

ORNL-DWG 68-5494

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

~ 30mc MeI ADDITION
T\ CT-A
[r\ CT-B
f
CT-C
1
{
|
/
[I
{ CT-D
/ CT-E
/
/
/
/
D
2 4 6 8 10 12
TIME (hr)

15. Charcoal Trap Radiation Levels During Addition of CH

3

131

14

I.
38

inlet half of the bed). The ranges on the two temperature recorders

were increased from 0 - 1000 to 0 - 1500 and 0 - 2000°F. A temporary
bypass line was installed around the fuel storage tank to eliminate its
120-ft3 volume between the fluorine supply and the fluorine reactor.

This line had to be conditioned with dilute fluorine, which involved
manual flow control., Control was erratic; however, the fluorine reaction
started immediately and, soon thereafter, the alumina trap inlet began to
heat up. The heated zone proceeded to the second thermowell, and the
fluorine flow was stopped when an off-scale reading of 1500°F was obtained
at this point. We decided to allow the alumina bed to cool and to resume
the run the next day. The system was conditioned so that automatic flow
control of 100% fluorine could then be used.

The fourth test was resumed on May 21; our efforts were centered on
trying to determine the cause of the alumina trap heating. To ensure a
complete reaction of the fluorine with SO,, the fluorine reactor inlet
was preheated so that the temperature of the reaction zone was about
600°F; steam cooling was delayed until the reactor temperature reached
1000°F. Then the steam pressure was carefully adjusted to maintain all
temperatures above 400°F. A maximum fluorine flow of 25 std liter/min
was used, with a minimum of 50% excess SO,. The flow control was good.
In spite of satisfactory operation in the reactor, the alumina trap
temperatures again increased as soon as the reaction started. We con-
cluded that these increases could only have been caused by SO;F, decom-
position. When the exit of the alumina trap reached about 1000°F,
iodide began to move down the charcoal bed. Most of the iodine left
the first 12 in. of bed and accumulated at the inlet of the backup trap.
Fluorine flow was stopped 20 min after this migration began, and thus
no iodine left the trap. About 5 mc of iodide still remained at this
time.

Laboratory tests4 were made by W. S. Pappas* to verify the decompo-
sition of SO,F,. While these tests showed no gaseous decomposition
products leaving the alumina bed, the laboratory bed was not completely
exhausted as was the case at the MSRE. The cause of the iodine movement

in the last run could, therefore, have resulted either from unreacted

 

*
ORGDP.
39

fluorine passing through the exhausted alumina bed or from SO,F; decompo-
sition products leaving the exhausted bed. The movement was probably not
due to SO0,F, since Pappas' work showed only partial trapping of SO,F,
below 1200°F. Also, SO5F, was produced for 2 hr before the alumina bed
temperature reached 1200°F, and no iodine movement was observed during
this time.

With steam cooling of the fluorine reactor, it was not possible to
maintain the reactor walls above 400°F. Thus the complete reaction of
traces of fluorine could not be ensured. If work had been continued on
the S0, system, we would have installed controlled electric heat over the
entire 8 ft of the reactor instead of steam heat over the first 3-2/3 ft.
Future work would also have involved decreasing activated alumina trap
temperatures (to 700 to 750°F) to prevent the SO,F, reaction. Some means
of cooling would probably have been required to prevent a small fluorine
reaction from heating the bed above the SO,F, reaction temperature.

After this last fluorine reactor test, it was thought that the
reactor could be operated at high efficiency; however, this could not be
readily demonstrated. Since no highly efficient backup trap was avail-
able for the detection and removal of traces of fluorine in the presence
of SO, and SO,F,, we decided to make a test run with KOH-KI solution in
the caustic scrubber. Temporary piping changes were, therefore, made to

place the backup alumina trap downstream of the caustic scrubber tank.

5.2 Caustic Scrubber Tests

Two weeks after the last fluorine reactor test was completed, the
piping was modified to test disposal, using the scrubber. The scrubber
was charged with a 2 M KOH -- 0.33 M KI solution, and 30 mc of CHgl3t1
was again loaded on the charcoal traps. The backup alumina trap material
was replaced, and the trap was heated to 700-800°F. Fluorine diluted to
less than 50% with helium was used to condition the scrubber -- an opera-
tion requiring several hours. After about 50 min of conditioning the new
or recently exposed piping with dilute fluorine the fluorine flow rate
was stabilized at 35 std liters/min and diluted with an equal flow of

helium upstream of the scrubber.
40

As soon as the fluorine flow began, the temperature near the exit
of the alumina trap (containing 5-mesh alumina) began to increase slowly.
There was no iodine movement; the other alumina bed temperatures were
steady; and a gas sample downstream of the scrubber showed only a faint
trace of fluorine. The temperature rise, was therefore, attributed to
the increased gas flow due to oxygen formation in the scrubber. This
was verified two days later by injection of 6 std liters of oxygen per
minute into the system, using a helium flow equal to that used during
the test run. The temperature rise was duplicated. Further confirmation
was obtained by repeating the fluorine flow test; no additional tempera-
ture rise was obtained.

Following the successful test at the MSRE, the Fuel Processing
Plant was converted from the SO, system to the scrubber method of fluo-
rine disposal. The fluroine reactor, gas preheaters, and associated
piping were removed from the cell. The Monel inlet distributor ring in
the caustic scrubber was blanked and replaced by an Inconel dip tube.
(Laboratory tests had shown that Inconel had much greater corrosion
resistance than Monel.) Since dip tube plugging (by corrosion products)
had occurred in laboratory tests, a spare dip tube was installed. Ex-
tension handle valves were supplied to permit dip tubes to be switched
from the operating area. A contact microphone was installed on the
scrubber tank to detect any abnormal reaction such as that which occurred
in laboratory tests when a fluorine concentration of greater than 50% was
used.

A second scrubber test was made on July 4, 1968. The fuel storage
tank bypass line was removed, and the tank was heated to 1100°F for a
pressure test. The tank was maintained at this temperature during the
scrubber test. After 70 min of fluorine flow, the pressure upstream of
the scrubber started to increase. The flow was, at first, reduced from
35 to 25 std liters/min and, then, had to be stopped completely 15 min
later to prevent the solution from leaving the scrubber tank. After
the system had been purged overnight with helium at a low flow rate,
the gas line from the scrubber was removed and found to be almost com-
pletely restricted by a white solid containing a high percentage of

molybdenum.
41

Laboratory tests using MoFg and KOH-KI solution produced a heavy
mist that could be removed from the gas stream by a Fiberglas filter. A
50-ft? filter was, therefore, installed downstream of the scrubber to
remove such material. To prevent plugging of the connecting line, the
size of the line was increased from 3/4-in. to 1-1/2 in., and all sharp
bends were eliminated. The alumina backup trap was relocated downstream
of the Fiberglas filter, and the alumina was replaced with soda-lime,
which removes fluorine effectively at low temperatures. Both the filter
and the soda-lime trap were steam heated to prevent condensation of the
saturated gas stream from the scrubber.

Another. equipment change resulting from. the experience of the second
scrubber test was the addition of means for heating the UFg absorbers.
The MoFg was expected to be retained on these absorbers, which were at
room temperature, instead of passing through to the scrubber. Follow-up
laboratory tests showed a much slower reaction rate for both MoFg and UFg
with low-surface-area NaF than with the high-surface-area material used
in previous volatility processing. Heating and insulating the absorbers
provided two advantages: (1) the amount of UFg leaving a particular ab-
sorber before that absorber is fully loaded was reduced, and (2) the

amount of coabsorbed molybdenum was minimized.

5.3 Fluorination of the Flush Salt

On August 1, the flush salt was transferred by gas pressure, from
the fuel flush tank in the drain cell to the fuel storage tank in the
fuel processing cell. The transfer required about 11-1/2 hr. Following
this transfer, the isolation freeze valves were frozen and leak-tested,
and the flush salt was sampled. Analyses showed that fuel salt had been
retained by the flush salt during the seven flushes of the reactor system.
Each of these flushing operations contributed about 200 ppm (or 1 kg) of
uranium to the flush salt, for a total of about 6-1/2 kg of uranium.
Pertinent irradiation and decay data are shown in Table 5.

Fluorination of the flush salt was started on August 2. The salt
was sparged with 25 std liters of fluorine per minute diluted with 20 std
liters of helium per minute. An additional 15 std liters of helium per

minute was used to purge the sample line. The reading on the inlet mass
42

Table 5. Fuel System Flushes

 

 

Flush Decay at Time of Fuel Irradiation at
No. Fluorination (months) Time of Flush (Mwhr)

1 37 <1

2 25 7,800

3 21 9,500

4 20-1/2 10,000

5 14-3/4 33,000

6 12 40,000

7 4-1/4 72,400

 

flowmeter rose rapidly and reached a maximum after 48 min of fluorine
sparging. The peak reading corresponded to a fluorine utilization of
about 15%. The readings of the thermocouples on the No. 1 and No. 2
absorbers rose, but lagged the initial rise in the mass flowmeter read-
ing by 20 to 30 min. The reading on the exit mass flowmeter rose as
expected as the fluorine utilization decreased from its peak, but attained
a much higher reading than expected. When the reading for this meter
went off-scale, we decided to bypass the fluorine around the salt while
we 1lnvestigated the reason for this high reading. There was some concern
that a uranium breakthrough had occurred; however, no temperature rise had
been observed on any of the last three absorbers. It was finally con-
cluded that MoFg was the cause of the high reading; therefore, fluorine
sparging was resumed after 95 min. After a total of 5 hr of fluorine
sparging, the fluorine supply was exhausted. Since the reading on the
inlet flowmeter had almost become constant, we decided to sample the salt
while the fluorine trailer and the caustic scrubber solution were being
replaced. Fluorination was resumed after the caustic solution was re-
plenished; it was then terminated when a uranium analysis of 24 ppm was
obtained for the flush salt, since fluorine sparging had been carried out
for an additional 109 min and the reading of the inlet mass flowmeter had
become constant. Analysis of a second salt sample, taken at this time

indicated a uranium concentration of only 7 ppm (desired <10 ppm).
43

The absorbers were loaded with low-surface-area NaF for the flush
salt processing and were heated to 200 - 250°F to increase the reaction
rate (see Sect. 6.7). Since the amount of uranium in the flush salt was
only about half the capacity of one absorber, very little uranium was ex-
pected to be deposited on the last four absorbers. However, almost 30%
was collected on the second absorber and possibly 3% on the third, al-
though no temperature rise was detected on the last three absorbers. For
the fuel salt runs, it was, therefore, decided to use only high-surface-
area material in the final three absorbers to permit operation with only
one backup absorber. The temperature rise on the first absorber was only
39°F; thus cooling air was not required.

Although some plugging of the caustic scrubber dip lines occurred,

switching to the spare line for several minutes always cleared the line.

5.4 Reduction of the Flush Salt

After the analysis of the first fluorination salt sample (uranium,
24 ppm) was obtained, heatup of the salt for hydrogen reduction was
started. During the heatup, the caustic scrubber dip lines were checked
and found to be free of plugs. The caustic solution was left in the tank
to neutralize the HF formed during NiF, reduction. A period of 43 hr at
maximum wattage was required to raise the salt temperature from 860 to
1220°F. August 5, 1968, hydrogen sparging was started. A mixture of
hydrogen (30 std liters/min) and helium (20 std liters/min) was used in
conjunction with a 10-std-liter/min helium sample line purge. The hydro-
gen monitor in the off-gas line downstream of the charcoal beds was
observed in order to determine the time at which hydrogen consumption by
NiF, ceased. After 2 hr, the rate of increase of the monitor was essen-
tially zero, which indicated a nickel reduction rate of less than 3 ppm/hr.
Hydrogen sparging was continued for a total of 10.8 hr.

Zirconium was then added through the salt sampler to reduce the FeF,
and CrFy, as well as the remainder of the NiF,. Three 1l-in.-diam cylin-
ders of pressed zirconium shavings, weighing 604 g total, were dropped
through the sample line with the manipulator. The contact microphone on
the fuel storage tank provided an audible verification of the passage of

the cylinder down the length of the sample line. Although there was some
44

concern that the cylinders would be held up by the valve or the bellows,
this did not occur. The salt was then sparged with a mixture of hydrogen
(30 std liters/min) and helium (10 std liters/min). An additional flow
of helium (20 std liters/min) was used to purge the sample line. After

9 hr of sparging, the hydrogen flow was stopped; then the system was
purged with (75 std liters/min) for 85 min before the salt was sampled.
Two samples were obtained: a standard dip sample, and a filtered sample,
using a capsule sealed with lithium-beryllium fluoride salt (''freeze-
valve capsule’). Analyses of the samples indicated that, although the
nickel concentration had been reduced to less than 50 ppm, there had only
been a slight reduction in the iron and chromium concentrations. An
additional 470 g of zirconium was added, and the salt was sparged with
the same hydrogen-helium mixture (see above) for an additional 16 hr.
After failure to obtain a sample with the freeze-valve capsule, the salt
was transferred through the filter to the drain tank. In the drain tank,
the salt was successfully sampled by using a new type of capsule, designed
by A. I. Krakoviak,* that relies on the vacuum produced by the cooling
capsule to pull the salt sample through the filter. This eliminates the
danger of losing vacuum by the melting of the frozen salt seal in the
freeze-valve capsule. Analysis of the filtered portion of the salt showed
that reduction was adequate and that the filter had removed the reduced

metals.

5.5 Fluorination of the Fuel Salt

Results obtained during the processing of the flush salt pointed out
the need for several equipment and procedural changes before the much
more radioactive fuel salt was processed. These changes, which required
about two weeks, consisted of the following:

(1) Enough high-surface-area NaF (1400 1b) was procured from ORGDP to
load the last three absorbers in each run, and the first two ab-
sorbers in three of the six runs, to obtain a comparison between
the performance of the two types of NaF.

(2) Piping was installed to permit the helium diluent (required to limit
the fluorine concentration to the caustic scrubber to less than 50%)

to be introduced downstream of the absorbers instead of into the

 

*
Reactor Division.
45

fuel storage tank. This change increased the residence time in the
absorbers, thus permitting more efficient absorption.

(3) The flow rate of the sample line purge gas was reduced from 15 to
10 std liters/min in run 1 and to 5 std liters/min thereafter to
further increase the absorber residence time.

(4) For more efficient absorber purging, a connection was provided at
the absorber inlet to permit purging with pure helium instead of
purging through the fuel storage tank.

(5) A small soda-lime trap was installed in the absorber cubicle to
allow the absorbers to be vented to atmospheric pressure instead of
the scrubber liquid head.

(6) Purge flow rates following fluorination were kept low to allow more
efficient absorption of UFg, which was being purged from the gas

space.

While the above changes were in progress, the flush salt was re-
turned to the reactor system and circulated to test the fuel pump sampler.
The transfer of the fuel salt to the processing cell was started on
August 17, 1968, after a difficulty with a restriction in the line at the
cell wall penetration was resolved. The spare heaters at the penetration
were energized, thereby raising the temperature from 1108°C to 1255°C, but
the restriction persisted. However, when the drain tank was pressurized,
the hot salt evidently thawed the plug, and transfer was started. This
operation required about 11 hr at an average rate of 15 1b/min. After the
transfer, a radiation level of about 2 r/hr was detected at the filter
located upstream of the inlet mass flowmeter in the absorber cubicle., This
radioactivity was caused by carryover of metallic ?°Nb during the gas
blowthrough at the end of the salt transfer. The filter was decontami-
nated to approximately 100 mr/hr to minimize contamination of the uranium
product during fluorination. Analysis of the decontamination solutions
confirmed that the principal activity present was 2°Nb.

The caustic scrubber was emptied, rinsed, and recharged; then several
salt samples were taken. The low-surface-area NaF in the trap for remov-
ing HF from the fluorine supply was replaced with high-surface-area

material since agglomeration of pellets in the first flush salt absorber
46

indicated the presence of considerable HF in the gas stream. This HF
could have resulted from the reaction of fluorine with the Be(OH), re-
maining in the flush salt after the previous Hy-HF treatment.

Because the fluorine supply system had been opened, it seemed advis-
able to condition this equipment with dilute fluorine (this was started
on August 23). Conditioning and a period of steady flow to check the
Hastings mass flowmeter readings were completed, and then salt sparging
was started, After 30 min at a flow rate of 20 std liters/min, the
fluorine flow was increased to 40 std liters/min for the conversion of
UFy to UFs. Fluorine utilization was very high (v70%) during this period.
For almost 6 hr, the absorber inlet and outlet meters read the same, in-
dicating that no absorbable compounds were volatilized. For the next
100 min, the readings on both meters rose (the inlet meter at the faster
rate), but no absorber heating was noted. This was interpreted to indi-
cate MoFg (from Hastelloy-N by corrosion) volatilization and either de-
creasing fluorine utilization or incomplete absorption of the MoFg on the
NaF. After 7-1/2 hr, UFg absorption was indicated by a temperature rise
on the first absorber. Fluorine flow was decreased to increase fluorine
utilization and absorber residence time. Twenty minutes before UFg ab-
sorption began, the pressure in the absorber cubicle rose above the alarm
point (-1 in. H,0). It was found that the blower which normally maintains
a negative pressure in the cubicle was inoperable, apparently because of
a mechanical rather than an electrical malfunction. However, since the
blower was located in the fuel processing cell, it could not be repaired.
By careful adjustment of the cooling air, the pressure in the cubicle was
kept from exceeding atmospheric pressure.

To ensure that essentially ne uranium was lost to the caustic
scrubber, we decided to use the fifth absorber only as backup and not for
collecting uranium. Consequently, the first run was terminated when
uranium broke through to the third absorber. Since no temperature rise
was observed for the fourth absorber during purging, the second run was
continued until uranium broke through to the fourth absorber. The third,
fourth, and fifth runs were terminated after 60, 75, and 95 min, respec-
tively, of loading on the fourth absorber without any temperature rise

being observed on the fifth absorber. The sixth run completed the
47

operation; no uranium was found on the fourth or fifth absorber. Fluorine
flow was stopped when the inlet mass flowmeter showed no further decrease
for an hour (see Fig. 16). End point determination was somewhat difficult
because of the almost continuous buildup of restrictions in the scrubber
dip tubes, accompanied by an increase in system pressure and a decrease
in fluorine flow. The increase in the reading of the exit meter at the
end of the run was probably due to MoFg being desorbed since the reading
was too high to be accounted for by helium and fluorine. The neutron
counter near the fuel storage tank provided another sensitive indication
of the end point. The neutron count rate fell rapidly during the last
few hours of fluorination but had become constant by the time the fluorine
flow was stopped.

After run 6, analysis of a fuel salt sample indicated 26 ppm, or
about 130 g, of uranium remaining in the salt. This is a very acceptable
loss. The salt was, therefore, heated to 1225°C to reduce the structural

metal fluorides.

5.6 Reduction of the Fuel Salt

The temperature of the fuel salt reached 1225°C on August 31, 1968
and a 30-std-liter/min hydrogen sparge was started at this time. A low-
flow hydrogen sparge fiad been used during the previous 15-hr heatup. A
high radiation reading noted a few hours later at the absorber cubicle
appeared to be centered at the metallic filter upstream of the inlet mass
flowmeter. The unshielded radiation level at the cubicle wall was about
80 r/hr. The source of this radiation was later identified as 95Nb (Sect.
6.5).

It is known that niobium exited in the metal form during reactor
operation and that, in this form, most of it plated out or left in the
gas stream. After reactor shutdown, the 2°Nb grew back in from 357r
decay. After a decay period of five months, less than 0.1% of this I5Nb
would be required to produce the observed radiation level. The rest of
the niobium probably plated out on the drain tank or on processing equip-
ment surfaces, or was volatilized during fluorination. About 60% of the
NbFg formed during fluorination would be carried from the salt by 60,000

liters of gas flow.
MASS FLOWMETER READINGS (cfm of air)

48

ORNL-DWG 69-5404

 

2.0

INLET

 

 

1.8

 

EXIT J

 

 

<
L
AL

 

 

 

 

 

1.2

 

FLUORINE
FLOW OFF

 

 

 

— e —

 

0.8

 

 

 

 

0.6
2200

2400 0200 0400
TIME, 8-29-68

Fig. 16. End of Run No. 6.

{x103)

58

50

45

40

35

NEUTRON COUNT RATE (counts/min)

30
49

Because of the higher concentration of corrosion products in the
fuel salt (than in the flush salt), a 24-hr hydrogen sparge was specified;
however, frequent scrubber dip line plugging caused the sparge to be
stopped after 17 hr. At this time, the scrubber solution that had been
used during the last fluorination run was replaced with a KOH solution
containing no KI or K,B,07. Because of frequent plugging of the lines,
the flow rates were reduced below the planned 30 std liters/min for
about 70% of the time. A filtered salt sample was taken and analyzed
while the scrubber solution was being replaced. As expected, there was
no evidence of reduction of chromium or iron; however, the analysis showed
that the nickel concentration had increased from 840 ppm to 1540 ppm.
This increase in nickel concentration was thought to be the result of
contamination by nickel oxide, originating from poorly purged welding
of the nickel filter element in the freeze-valve capsule. After another
17 hr of sparging, the nickel concentration had only decreased to 520 ppm;
again no chromium or 1iron was reduced. A third period of sparging for
17-1/2 hr decreased the nickel concentration further to 180 ppm. The
remaining reduction was accomplished with zirconium metal.

Twenty-seven zirconium slugs, weighing an average of 180 g each,
were charged through the sampler; two others rolled beyond the reach of
the manipulator and could not be charged to the salt at this time.
Approximately 5 kg of zirconium was added, and the salt was sparged for
24 hr with hydrogen at the rate of 10 std liters/min. Analysis of a
filtered sample of salt taken at this time showed the nickel to be
completely reduced and the chromium and iron reduced to about 100 ppm.
The salt had been agitated with helium prior to the sample analysis and
then was sparged for 8 additional hr with hydrogen. Difficulties again
prevented sampling before filtration,

The fuel salt was filtered without difficulty. A dip sample from
the drain tank showed the chromium concentration to be 34 ppm; a second
filtered sample showed only 29 ppm of chromium in solution. More details

on the reduction are given in Sect. 6.
50

6. DISCUSSION OF DATA
6.1 Fluorine Utilization

The efficiency of the fluorination reaction is measured by the per-
centage of the fluorine (after subtracting the fluorine consumed by
corrosion) that is used in the reaction with the uranium. The fluorine
consumed by corrosion is assumed to be constant at 2.5 std liters/min.
Correction is also made for the inert gases present (assumed to be nitro-
gen) in the fluorine supply and the effect of the reduced gas density on
the orifice flowmeter. Fluorine utilization is a function of melt temper-
ature, fluorine flow rate, uranium concentration, and dip tube submergence.
Laboratory tests had shown a decrease in utilization from 8.0 to 3.8% as

> The higher utilization

the temperature was reduced from 930°F to 840°F.
(average, 39%) during the fuel salt processing was probably due both to
lower gas velocities by a factor of 5 and greater dip tube submergence (by
a factor of 15) than used in laboratory tests. Utilization was very high
before the start of UFg volatilization (average, 71%) and, after that,
seemed to be relatively insensitive to uranium concentration until nearly
all the uranium was volatilized (see Figs. 17 and 18). The highest aver-
age utilization during UFg volatilization occurred in the fifth run (see
Table 6).

Until 90 min before the start of UFg volatilization, the readings of
the absorber inlet and outlet mass flowmeters were equal, indicating the
absence of any absorbable constituents in the gas stream. These flow-
meters were used to calculate the fluorine utilization, as shown in Fig.
18, The utilization decreased when the fluorine flow rate was increased
and also when MoFg and UFg began to be vaporized. It is thus assumed that
the fluorine utilization does not change, when the fluorine flow is steady,
until volatilization begins. During the MoF; volatilization period (before
UFg volatilization began), the utilization was calculated by difference,
using the utilizations in the two previous periods and assuming that all
the UF, was converted to UF; prior to uranium volatilization. The fluo-
rine flow rates are adjusted for the fluorine that was consumed by corro-
sion. During the volatilization of UF., the fluorine utilization was

calculated from the increase in absorber weights. The inlet mass
FLUORINE UTILIZATION (%)

FLUORINE SPARGE RATE (sim)

»
O

W
o

N
O

o

H
O O

W
@)

20

10

51

ORNL-DWG €9-5407

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L 1) ’”—"—_L\
l | I l I
I | I ' |
I | | I | \\\\‘_
| I [ I |
.JL—l__“L ~—
—
T I I .
RUN RUN | RUN | RUN | RUN I RUN
T 4 | - 5 ! 6
i L1 |
0 10 20 30 40

FLUORINE SPARGE TIME AFTER UFg BREAKTHROUGH (hr)

Fig. 17. Fluorine Flow Rate vs Utilization.
52

ORNL-DWG 69—-5408

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 ; _
z o
= START OF MoFg VOLATILIZATION
z : ;
S 80 - e e} —— ]
'S
w I
= i !
o : | i ; [
S e 1 . S ] | STARYT OF Uy __ |
> 5 | . ‘ VOLATILIZATION
U ————— SLUCRINE UTILIZATICN
s ~-——=FSLUQORINE FLOW :
< a
% a0 - - + ——— - —_——

——— e e e P gy o e — e e —
& i : ! 1
2 i \ FrT T
= Lo | N |
E ¥y - T A o
LLC\.‘
5 _
c i
C { 2 3 a4 5 6 7 8 9
FLUCRINE SPARGE TIME (hr)

Fig. 18. Fluorine Utilization at the Start of Run 1.
53

Table 6., Fluorine Utilization

 

 

Corrected Percentage
Fo Flow o
(std liters/min) Utilization
Flush salt 19.8 7.7

Fuel salt, overall 18.8 39
Run 1 - Before MoFg Volatilization 17.1 98

- Before MoFg Volatilization 36.0 77

- During MoFg Volatilization 31.5 43

- During UFg Volatilization 26.2 32

Run 2 - During UFg Volatilization 15.3 31
Run 3 - During UFg Volatilization 16.7 29
Run 4 - During UFg Volatilization 16.2 31
Run 5 - During UFg Volatilization 13.5 33
Run 6 - During U¥Fg Volatilization 16.9 13

 

flowmeter indicated a high utilization toward the end of each run because
of the stratification of the dense UFg in the fuel-storage-tank gas space.
The fluorine flow rates calculated from the decrease in pressure in

the fluorire trailers agree reasonably well with the results obtained with
the orifice flowmeter except in runs 2 and 5. In run 2 a fluorine flow

of 90% of the rate indicated by the flowmeter was used since this produces:
(1) utilizations in reasonable agreement with those observed in the other
runs, (2) an integrated UFg flow from the inlet meter that is in agreement
with the absorber weight increase, and (3) a reasonable amount of MoFg
leaving the absorbers as calculated from the outlet meter. In run 5 a

fluorine fiow of 110% of the rate indicated by the flowmeter was used.

6.2 Corrosion

One o7 the major problems encountered during processing was the
corrosion of the fuel storage tank. Coupon testing in the Volatility Pilot

Plant6 and at Battelle Memorial Institute7 had shown Hastelloy-N to be
54

superior, as a material of construction, to nickel (mainly because of

the absence of intergranular corrosion) and as resistant to corrosion as
any of the other materials tested; however, the corrosion rate was still
much higher than desired. The corrosion rates shown in Table 7 were cal-
culated from the increase in Cr, Fe, and Ni contents of the salt and from

the amount of MoFg volatilized.

Table 7. Corrosion During Fluorination

 

Corrosion (mils/hr), Based on:

 

 

Mo Ni Fe Cr Aver,a
Flush Salt 0.10 0.2 0.15 0.11
Fuel Salt 0. 2 0.04 0.2 0.14 0.09

 

aWeighted for concentration in Hastelloy-N.

The corrosion rates calculated during flush salt processing may be
high since they include a 2-hr conditioning period at 1125°F with fluorine.
All other fluorine exposures were at 850 to 860°F. The corrosion rate,
based on molybdenum, was calculated from the rise of the reading of the
inlet mass flowmeter prior to the temperature rise in the first absorber
(Fig. 19). It is believed that this rise was due to the presence of MoFg,
which is more volatile than UFg and has almost as large a heat capacity.
The corrosion due to molybdenum in the flush salt could not be calculated
directly because of the short time interval (20 min) between the two indi-
cations. The equivalent time interval for the fuel salt was 90 min; thus
the calculation in this case should be reasonably accurate. A 15-min time
lag in the temperature response was assumed in the calculations. With the
fuel salt there was a simultaneous, but smaller, rise on the exit flow-
meter reading. This is believed to be caused mainly by the increasing
amount of unreacted fluorine due to reduced utilization. The average exit
meter reading agreed reasonably well with the calculated reading for 43%
fluorine utilization (see Table 6). The higher corrosion rate, based on
molybdenum, is probably due to surface leaching and rapid volatilization

of the highly volatile MoFg.
55

ORNL-DWG 69-5405

 

 

 

 

 

 

* /
5 /
5 / .
£
s
-
=z
1
-
I
2
D
o 2
w
[] /
S INLET METER {
&
=1
V1]
x
@
- d EXIT METER I
Yoy / TEMPERATURE RISE -
ON NO.f ABSORBER
l
i

 

 

 

 

 

 

 

 

FLUORINE OFF FOR
TRAILER CHANGE METER DIFFERENCE
C 6

0 3
TIME (min)

0 90 120 150

Fig. 19. Mass Flowmeter Readings Before Start of Volatilization

of UF6 from Fuel Salt.
56

6.3 Molybdenum

During reactor operation, corrosion product and fission product
molybdenum exists as finely divided metal. In this form, most of it
leaves the salt, either in the gas stream or by deposition on the graphite
and metal surfaces. During fluorination, the molybdenum in the Hastelloy-
N fuel storage tank is attacked and converted to MoFg. Molybdenum is the
only major constituent of the fuel storage tank forming a highly volatile
fluoride. (Chromium fluoride is not volatilized until essentially all of
the UFg has been volatilized.) As seen in Table 8, MoFg is somewhat more
volatile than UFg. The molybdenum complex, MoFg-3 NaF, is much less stable

than the uranium complex, UFg.3 NaF.

Table 8.. Comparison Between MoF, and UF

 

 

6 6
UFg MoFyg
Boiling point at 760 mm, °F 147 95
Partial pressure over NaF at 200°F, mm 0.001 2

 

At the start of fuel salt fluorination, it is likely that non-vola-
tile MoF3 is formed and is not converted to MoFg until most of the UFy
is converted to UFs. Because of the higher volatility of MoFg, volatili-
zation would be expected to begin before UFg; begins to volatilize. This
was verified by the increase in the reading on the absorber inlet mass
flowmeter about 1-1/2 hr before an absorber temperature rise was noted
(see Fig. 19). During this period of MoFg volatilization, the reading on
the absorber outlet meter indicated that no MoFg was present i1n the exit
gas stream until near the start of UFg absorption.

After UF, began to volatilize, the absorber outlet mass flowmeter
was used to calculate the amount of MoFg leaving the absorbers. The
molybdenum in the absorber outlet stream was also calculated by two other
methods: from the gas flow and the MoFg vapor pressure, and by subtracting

the amount of MoFg retained on the absorbers from the amount of MoFg
57

produced by corrosion. A good correlation of the data, except in run 6,
was obtained, showing that the amount of absorbed MoFg increases as the
amount of UFg passed through the NaF increases. The data, which are only
approximations, are compared in Table 9. This table also shows the amount
of molybdenum found in the caustic scrubber solutions. The reason for
the excessive amount of MoFg leaving the absorbers in run 6 is not clear.
The amount of molybdenum that passed through the scrubber is not known.
There was no accumulation of solids in the line from the scrubber as was
experienced during the second scrubber test run (Sect. 5.2); however, the
mist filter was not examined.

From the random samples of NaF analyzed for molybdenum, the follow-

ing observations have been made concerning the absorption of MoFg:

(1) Absorption is higher on high-surface-area NaF than on low-surface-
area NaF.

(2) Absorption is higher on pellets than powder, probably because MoFg
is trapped.inside the pellets.

(3) Absorption is higher on NaF that has been exposed to UFg, and in-
creases with the UFg concentration in the gas stream. Run 6 appears

to be an exception in this respect.

6.4 Plutonium

Since the MSRE fuel salt contained 66 wt % 238U, a significant
amount of 23%Pu was produced. At the time of reactor shutdown, the plu-
tonium concentration in the fuel salt was calculated to be 112 ppm (560 g)e8
Five salt samples withdrawn after fluorination, after reduction, and after
return of the salt batch to the reactor drain tank showed an average con-
centration of 110 ppm. This confirms the expectation that plutonium would
not be volatilized under the relatively mild conditions required for ura-

nium volatilization.

6.5 Niobium

Figure 20 shows the amounts of fission product fluorides calculated
to be volatilized during fluorination of the fuel salt. The iodine and

tellurium fluorides (Sect. 6.6) pass through both NaF beds to the caustic
Table 9. Disposition of Volatilized Molybdenum

 

Molybdenum (g)
In Absorber Outlet Stream
Amount Produced

 

 

Produced g Absorbed Minus Amount From Flow and From Mass In Caustic

by Corrosion on NaFP Absorbed Vapor Pressure Flowmeter Scrubber
Flush Salt 990 28 962 940 C 260

Fuel Salt

Run 1 1270 130 1140 246 670 107
Run 2 739 210 519 427 364 72
Run 3 868 330 538 257 267 86
Run 4 914 375 539 132 870 24
Run 5 1131 425 706 271 663 17
Run 6 1114 112 1002 399 3120 67
Total Fuel 6036 1582 4444 1732 5954 373
Total 7026 1610 5406 2672 c 633

8S

 

%At a constant corrosion rate of 0.2 mils/hr.
From a correlation of absorbed Mo vs UFg through NaF.

cUnable to calculate.
59

ORNL—DWG 69-5402
106

VAPOR PRESSURE =10%mm:
MoFg, IF7,TeFg

105

104

VAPOR PRESSURE AT 860°F (mm)

103

5 || VAPOR PRESSURE <102mm:
ZrF4, CeF3

 

SnfFa

0 20 40 60 80 100
PERCENT VOLATILIZED DURING FUEL SALT FLUORINATION

102

Fig. 20. Calculated Fission Product Volatilization During Fuel

Salt Fluorination.
60

scrubber and charcoal traps. Because of the five-month decay time, 990
was not present in detectable amounts. Ruthenium, niobium, and antimony
exist as metals during reactor operation and, as such, are continuously
removed by deposition or by the off-gas stream. Since ruthenium is formed
directly by fission and since radioactive antimony has no long-lived pre-
cursors, the amounts of these isotopes in the processed salt were small as
compared with niobium, which grows in from nonvolatile °Zr.

Figure 21 shows the buildup of ?°Nb after reactor shutdown and the
buildup after fluorination, assuming complete volatilization. Also shown
is a comparison between the amount of niobium found in salt samples taken
before and after fluorination and the amount of niobium which should have
been present from zirconium decay (assuming no niobium present at sample
time) when the sample was analyzed several months after processing. Since
the sample prior to fluorination showed less niobium than would have grown
in from zirconium decay, it 1s impossible to say whether there was any
appreciable amount of niobium in the salt before fluorination. At least
60% of the grown-in niobium must have been lost during sample preparation
(before analysis). This loss did not occur in the post-fluorination
samples, as seen in Fig. 21, when the salt was 1n an oxidized state. The
close agreement between the calculated amount and the sample analysis
probably indicates that there was very little niobium left in the salt
after fluorination.

Figure 20 indicates that approximately 60% of the niobium in the fuel
salt volatilized during fluorination. It is not known, however, how much
niobium was present in the salt at that time. From the lack of heat gen-
eration in the Nal trap, the volatilized niobium 1s believed to constitute
less than 10% of the 9 x 10" curies which would have grown in after re-
actor shutdown. (A temperature rise of 20°F would have been observed from
the heat generation from 9000 curies of ?°Nb.) The temperature of the
trap, remained constant; that is, no heater adjustment was required through-
out the six runs. Apparently, most of the niobium was removed from the
salt by deposition in the drain tank or in back-flowing through the salt
filter during salt transfer. At the end of this transfer, the transfer
gas blowthrough carried a small amount of metallic niobium to the absorber

filter (see Sect. 5.5). Altinough this filter was decontaminated prior to
curies

3-25

61

ORNL-DWG 62-5403

95Np AT TIME OF SAMPLE ANALYSIS

O CALCULATED 95Nb FROM %3zr
AT SAMPLING TIME

® 95Nb FOUND IN SAMPLE

SAMPLES AFTER
FLUORINATION

CALCULATED 23Nb BUILDUP
AFTER REACTOR SHUTDOWN

Fig.

21.

 

5-24

&

~

SAMPLE BEFORE
FLUORINATION

E

T T

CALCULATED 95Nb BUILDUP
AFTER FLUORINATION

om0 e T I T 5 Q8 2 3
' ~ t ' o ! I ! f ! |
© ~ @ o Q o - o o
- - - -—

DATE

Buildup of 95Nb from 952r Decay.
62

fluorination, undoubtedly some of the niobium was deposited on the up-
stream piping. During fluorination, this niobium would be converted to
volatile NbFg and collect on the UFg absorbers (any niobium upstream of
the NaF trap should be absorbed on the trap). This is probably the source
of the niobium on the absorbers (Sect. 6.8) rather than niobium that was
volatilized from the salt passing through the NaF trap.

Two days after the end of fluorination, at the beginning of hydrogen
reduction, a large amount of niobium (estimated to be 15 to 20 curies) was
found on the absorber filter. Since 2°Nb should have grown in at the rate
of about 1000 curies per day, apparently less than 1% of the niobium in
the salt was carried to the absorber cubicle. A gamma scan, made of the
filter on March 10, 1969, is shown in Fig. 22. The only definite peak is
that of 9°Nb at 0.77 Mev. Very small amounts of 12°Sb, 193Ru, and 'Y®Rh

could be the cause of the bulge in the curve between 0.4 and 0.55 Mev.

6.6 Iodine and Tellurium

During fluorination, volatile IF; is formed and passes through both
NaF beds; it is removed by the caustic scrubber by reaction with the KOH
and by exchange with the KI. A total of 404 mc of 13!I was calculated to
be in the fuel salt at the start of processing. Analyses of the caustic
scrubber solutions showed that 288 mc was collected during run 1 and 10 mc
was collected during run 2 for a 74% accountability. This is in reason-
able agreement with the results of iodine analyses of fuel salt (during
reactor operation), which showed about 65% of the calculated amount. The
discrepancy between the calculated amount and the analyzed amount is
caused by the loss of metallic 13!Te before decay. There was no indica-
tion that any iodine collected on the charcoal beds downstream of the
scrubber.

Tellurium exists primarily in the metallic state during reactor
operation and, as such, is removed from the salt by deposition and by the
off-gas stream. Freeze-valve samples showed less than 1% remaining in the
salt. Since !2°Te has no long-lived precursor, the activity will not grow
back in after reactor shutdown. Any tellurium remaining in the salt will
be converted to the volatile TeFg during fluorination. Because of the low

MPC for 129MTe (1078 uc/cd, at least 94% of the remaining 1% would have
63

ORNL-DWG 69-5404

 

(x10%)

 

20

o1V

 

 

 

 

95Nb
0.77

counts /minute
o

 

 

 

 

 

 

 

 

 

 

0 0.2 04 0.6 0.8 1.0
GAMMA ENERGY (MeV)

Fig. 22, Gamma Scan of Absorber Cubicle Filter.
64

to be removed from the gas stream before the stream is discharged from
the stack. Tellurium hexafluoride is not absorbed on NaF at any tempera-
ture and removal in the caustic scrubber is not very efficient. However,
since analyses of the scrubber solutions showed tellurium to be below the
limit of detection, it is probable that less than 1% was present in the
salt at reactor shutdown. Any tellurium passing through the scrubber

would have been removed by the activated aluminag in the soda-lime trap.

6.7 Uranium Absorption

As mentioned in Sect. 5.5, two types of NaF were used: a high-surface-
area (HSA) material (1 m?/g), prepared at ORGDP by heating NaHF, pellets,
and a low-surface-area (LSA) material (0.063 mz/g, prepared at the UCNC
Paducah Plant, from NaF powder and water and sintered at 700°C. The LSA
material was originally specified for the MSRE Fuel Processing Plant be-
cause of its higher capacity for both CrFg (on the 750°F NaF trap) and
for UFg. (This high capacity is due to the slower surface reaction and
greater penetration to the inside of the pellets.) When the fluoride
disposal method was changed from the SO, gas-phase reaction to the aqueous
scrubber, a more conservative approach toward possible uranium breakthrough
from the absorbers became necessary. The low MoFg retention observed
during the final fluorine disposal test (Sect. 5.2) suggested that the
reaction rate with UFg might be too slow for complete uranium absorption
under the planned operating conditions. Subsequent laboratory tests10
confirmed the much slower reaction rate with the LSA material. The ab-
sorbers were, therefore, heated to increase the reaction rate, and the LSA
NaF was restricted to the No. 1 and No. 2 positions of the group of five
absorbers.

A comparison of the results obtained with the two types of NaF is
shown in Fig. 23. Three absorbers loaded with each type of material were
used in the first position, and three were used in the second position,
during the six runs. In the first position, the NaF was exposed to a
higher UFg concentration; this resulted in a lower loading than in the
second position'where the slower reaction rate permitted deeper penetra-

tion of the pellets. Except in one case, higher loadings were obtained
URANIUM LOADING (kg)

65

ORNL-DWG 68-11623

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14.5 } G
e NO.1 POSITION
o NO.2POSITION
14 .0
LSA
NQO. 2 POSITION
13.5 4\
\ .
13.0
\ \O
12.5 \ - \
LSA
NO. { POSITION
12.0 S
11.5 o \\
\ HSA
NO.2 POSITION
11.0 \
HSA
NO.1 POSITION .
10.5 . \
100 \
9.5
160 200 240 280

Fig. 23.

AVERAGE TEMPERATURE (°F)

Comparison of LSA and HSA NaF Loading.

320
66

at lower temperatures; in that particular case, the absorber had been
used in a previous run where a small amount of uranium was absorbed.

In spite of the fact that the LSA NaF absorbers were operated at
higher temperatures (to compensate for the slower reaction rate), the
total loading was 13 to 14% higher. The maximum uranium loading obtained
was 0.56 g per g of NaF, as compared with 0.49 g per g of NaF for the HSA
material. Both of these absorbers were in the second position. The three
lowest loadings shown in Fig. 23 for the LSA NaF contained from 11 to 18%
HSA (added later to more fully charge the absorbers), which undoubtedly
reduced the loading on these absorbers.

The time required for a breakthrough of UFg from a given absorber
to the subsequent absorber varies directly with the residence time of the
gas stream in the absorber, as shown in Fig. 24. The first absorber in
run 1 had an unusually short breakthrough time (25 sec), which was poss-
ibly caused by absorption of MoFg that had been velatilized before the
start of UFg volatilization (Sect. 6.3).

The conditions and loadings are summarized in Table 10. These load-

ings are high since no allowance was made for absorbed MoFg.

6.8 Uranium Product Purity

The 1nlet of the first absorber in each run was analyzed for gross
beta and gamma activity about 12 days after fluorination; the results are
shown in Fig. 25. The activity associated with uranium daughters at the
same time is also shown. At the time of the analyses, the 23°U daughters
had reached equilibrium, while the 23%U daughters had attained only about
30% of equilibrium. The only activities appreciably above background were
the gamma activity in the first run and the gamma activity on the first
absorber of the second run. (The activity in the feed salt before fluori-
nation, counted at the same time, was 2.66 x 10°3 gross gamma and 3,8 x 101!3
gross beta counts per minute per gram of uranium.) The average gross beta
and gamma decontamination factors (DF) for the six absorbers analyzed,
after correction for uranium daughter activity, were 1.2 x 102 and 8.6 x
108 respectively.

The only activity collected in any measurable amount on the NaF ab-

sorbers with the UFg; was ?°Nb. Most of this activity was probably carried
67

ORNL-DWG €9-5400

A

140
/

®
120 o®

100 -
/e

 

 

 

(min)

 

 

 

60

20 // *

 

 

 

ABSORBER BREAKTHROUGH TIME
&

 

 

 

 

 

 

 

 

 

0 5 10 15 20 25 30
ABSORBER RESIDENCE TIME (sec)

Fig. 24. Absorber Residence Time vs UF6 Breakthrough Time.
68

Table 10, Summary of UF, Absorption Data

 

 

 

 

 

 

6
Flow Through Average
Absorber Minus Nj Gas
He and Unreacted F2 Residence
Superficial Time
Aver. Corrected Gas (Excluding Uranium 4
Type Temp. Flow Rate Velocity UFg) Absorbed
Run Absorber NaF (°F) (std liters/min) (fps) {sec) (kg)

Flush 1 LSA 244 20.0 0.038 18.3 4,65
2 LSA 226 20.0 0.037 18.7 1.83

3 LSA 210 20.0 0.036 19.2 0.15

Aver 227 20.0 0.037 18.7 6,63
Fuel-1 1 LSA 261 39.4 0.077 9.1 12.53
2 HSA 258 28.5 0.055 12.6 9.80

3 HSA 258 28.5 0.055 12.6 3.49

2 1 LSA 255 21.2 0.041 17.0 11.98
2 LSA 282 17.9 0.036 19.3 12.86

3 HSA 265 17.0 0.033 20.8 8.36

4 HSA 253 17.0 0.033 21.2 0.61

3 1 HSA 252 18.7 0.036 19.3 10.66
2 LSA 263 18.8 0.037 18.9 13.14

3 HSA 258 18.8 0.037 19.1 11.91

4 HSA 274 18.9 0.038 18.6 5.73

4 1 HSA 193 19.5 0.034 20.2 11,42
2 HSA 218 19.5 0.036 19.5 11.55

3 HSA 217 19.5 0.036 19.5 11.24

4 HSA 220 19.5 0.036 19.4 9.83

5 1 LSA 238 16.0 0.030 23.0 13.51
2 LSA 243 15.1 0.029 24.2 14.45

3 HSA 188 15.1 0.026 26.3 12.37

4 HSA 207 15.1 0.027 25.5 8.94

6 1 HSA 207 38.7 0.070 10.0 10.45
2 HSA 200 38.7 0.069 10.1 12.47

3 HSA 179 38.7 0.067 10.4 0.55

Aver 236 22.7 0.043 18.0 217.85

 

“These weights include absorbed MoFg.

bTotal weights.
69

ORNL—DWG 69— 5406

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-
<2 - — - : -
< |
£
o
o
0
o
3 e GROSS £ IN U PRODUCT
o
/ \ °
I~
1 o t . P —— ‘ — e i e i ]
U DAUGHTER 8 ACTIVITY
UDAUGHTER)fACTHHTYi
|
GROSS ¥ IN U PRODUCT,
#———'___ \
e T
O — -
0 50 100 150 200

URANIUM VOLATILIZED (kq)

Fig. 25.

Gross Activity in Uranium Product.
70

from the salt into the piping and equipment between the fuel storage tank
and the metallic filter upstream of the absorbers at the end of the salt
transfer when the pressure of the transfer gas was released (Sects. 5.5

and 6.5). Because of this contamination, the decontamination factor for

95N ( dpm per gram of U in salt
dpm per gram of U on absorber

niobium volatilized during fluorination and collected on the absorbers.

) was lower than the actual DF from

 

The DF was lower at the start of run 1 when most of the niobium between
the NaF trap and the absorbers was fluorinated and collected on the first
absorber. The DF's varied from 5 x 10® at the start to 1 x 10!0 at the
end of fluorination. Therefore, notwithstanding the piping contamination,
a considerably higher niobium DF was obtained in these runs than in the
Volatility Pilot Plant (5 x 107) for 30-day-decayed uranium,

The activity of the uranium product, in excess of the uranium daugh-
ter activity, would be expected to be negligible after desorption of the
UFg from the NaF. Laboratory-scale runs have shown decontamination fac-
tors of 10 to 103 for this operatione11

The only nonradioactive contaminant of any consequence expected to
be found with the absorbed uranium was molybdenum (from corrosion of the
Hastelloy-N during fluorination). The actual amount on the absorbers
could not be determined until the UFg was desorbed. The behavior of

molybdenum during the processing is discussed in Sect. 6.3.

6.9 Reduction of Metal Fluorides

The concentrations of the structural metal fluorides in the flush

and fuel salts during various stages of processing are shown in Table 11.
Calculations of the hydrogen and zirconium utilizations were only

approximate for the following reasons:

(1) Difficulties in obtaining filtered samples.

(2) Large standard deviation in iron analyses.

(3) Unreliable nickel analyses using filter capsules with nickel filters.
Oxidation of the filter during fabrication caused a high nickel analy-

sis for at least one sample.

Since the flush salt was not sampled after reduction of NiF,, the

hydrogen utilization during reduction could not be calculated. The
71

Table 11. Structural Metal Fluoride Concentrations

 

Concentration (ppm)

 

Cr + 10 Fe + 40 Ni + 15

Flush salt?

In reactor system 76 150 52
Before fluorination 104 —— —
After fluorination 133 210 516
After 10.8 hr of H, sparging No sample taken

After 604 g of Zr and 9 hr of H, sparging 100b 174b SOb
After 1074 g of Zr and 25 hr of H, sparging No sample taken

After filtration 76b 141b 26

Fuel salt?

In reactor system 85 130 60
Before fluorination 170 131 36
After fluorination 420 400 840
After 17.1 hr of H, sparging 420b 430b c
After 33.5 hr of H, sparging 420b 400b SZOb
After 51.1 hr of H, sparging 460b 380b 180b
After 5000 g of Zr and 24 hr of H, sparging 100b llOb <10b
After 5100 g of Zr and 32 hr of H, sparging No sample taken
After filtration 34 110 60

 

aHz sparging times are cumulative.
bFiltered sample.

“Contaminated sample.
72

hydrogen utilization during fuel salt reduction, however, was calculated
to be 1.9% overall, as shown in Table 12. This is somewhat higher than
the utilization experienced in test runs with a 48-liter batch of salt.12

In the test runs, the utilization was less than 1% when the nickel
fluoride concentration in the salt was less than 500 ppm. The higher util-
ization during fuel salt reduction is probably due to greater dip tube
submergency (64 in. vs 26 in.). Apparently, a high hydrogen concentration
is more important for efficient reduction than a high gas sparge rate.

The minimum hydrogen utilization during flush salt reduction was calcu-
lated by assuming complete reaction of the zirconium and an iron reduction
equal to the chromium reduction. The maximum utilization assumes complete
NiF; reduction by the hydrogen. Table 12 summarizes the nickel reduction
step.

Since the NiF; reduction of the fuel salt (and probably also in the
case of the flush salt) was incomplete after hydrogen sparging, the re-
mainder of the NiF, was reduced by addition of zirconium. The reduction
step after zirconium was added to the salt is summarized below in Table
13. A zirconium utilization of only 40% during the flush salt processing

would have been required if all the NiF, had been reduced during the

hydrogen sparge.

7. ACKNOWLEDGMENTS

Much of the credit for the successful processing of the MSRE fuel
belongs to the following: G. I. Cathers for laboratory-scale work in
support of the fluorination, absorption, and fluorine disposal operations;
E. L. Nicholson for assistance in design of final modifications to the
plant; W, H. Carr, G. I. Cathers, and E. L. Nicholson for around-the-
clock technical assistance during the fuel salt fluorination; R. H. Guymon,
T. L. Hudson, and A. I. Krakoviak for supervision of operations; and
W. S. Pappas, ORGDP, for specification of special instrumentation and

assistance during testing and final operation.

J
R4

73

 

 

 

 

 

 

 

 

Table 12. Reduction of Nickel Fluoride
Hydrogen Sparge
Sparge Aver. H, Total Hjp %
Time Flow Rate Flow Hydrogen
Salt (hr) (std liters/min) H2 (liters) Utilization
Flush 10.8 36.2 64 23,500 2.2 - 3.6
Fuel 17.1 21.4 76 21,860 1.1
16.4 34.0 77 33,470
17.6 10.8 92 11,420 5.7
Fuel Aver. 21.8 79 1.9
Table 13. Reduction of Chromium and
Iron Fluorides After Addition of Zirconium
Zirconium Sparge Sparge (Gas
Added Time Flow Rate % Zr
Salt (g) (hr) (std liters/min) % Ho Utilization
Flush 604 9 44 77
470 16 45 78 >40
Aver. 45 78
Fuel 5000 24 11.5 91
—_— 18.5 20 0 (He) 79.5
100 8 13.2 92
Aver. 15 58

 
10.

11.

12.

74

8. REFERENCES

R. B. Lindauer, MSR Program Semiann. Progr. Rept., Aug. 31, 1965,
ORNL-3872, p. 152.

R. B. Lindauer, MSRE Design and Operations Report, Part VII, Fuel
Handling and Processing Plant, ORNL-TM-907 (Revised) (Dec. 28, 1967).

 

R. B. Lindauer and C. K. McGlothlan, Design, Construction and Testing
of a Large Molten Salt Filter, ORNL-TM-2478 (March 1969).

W. S. Pappas, Reaction of Sulfuryl Fluoride with Activated Alumina,
K-L-3062 (Sept. 4, 1968).

ORNL Molten Salt Reactor Program, Monthly Report for February 1968,
MSR-68-42 (Mar. 1, 1968).

A. P. Litman and A. E. Goldman, Corrosion Associated with Fluorination

 

 

 

 

in the ORNL Fluoride Volatility Process, ORNL-2832 (June 5, 1961).
P. D. Miller et al., The Corrosion of an INOR-8 Single-Vessel Hydro-
fluorinator-Fluorinator, BMI-X-234 (May 27, 1963).

 

 

 

R. C. Steffy, Oak Ridge National Laboratory, personal communication.

D. R. Vissers and M. J. Steindler,. Laboratory Investigations in

 

Support of the Fluid-bed Volatility Processes, Part XX. Fission-

 

product Tellurium Off-gas Disposal in the Fluid-bed Fluoride Volatil-
ity Process, ANL-7464 (August 1968}.

G. I. Cathers et al., MSR Program Semiann. Progr. Rept., Aug. 31,
1968, ORNL-4344, p. 324,

G. I. Cathers, et al., Use of Fused Salt-Fluoride Volatility Process

 

 

 

with Irradiated Uranium Decayed 15 - 30 Days, ORNL-3280 (September
1962).

J. H. Shaffer, L. E. McNeese, F. A. Doss, MSR Program Semiann. Progr.
Rept., Feb. 29, 1968, ORNL-4254, p. 155.

 

 

 
0o~ U N

Ez:n >'?321WJPHZEFU*U‘U Arr=Eo0oruppousEsenomuEnmasommmeenEon oS

75

Internal Distribution

L. Anderson 45. R. N. Lyon
. F. Baes 46. H. G. MacPherson
E. Beall 47. R. E. MacPherson
Bender 48. H. E. McCoy
S. Bettis 49, H. C. McCurdy
F. Blankenship 50. C. K. McGlothlan
Blumberg 51. L. E. McNeese
G. Bohlmann 52. J. R. McWherter
E. Brooksbank 53. A. J. Miller
B. Briggs 54. R. L. Moore
H. Carr 55. E. L. Nicholson
I. Cathers 56. W. S. Pappas
W. Collins 57. A. M. Perry
B. Cottrell 58. R. C. Robertson
L. Crowley 59-60. M. W. Rosenthal
L. Culler 61. A, W. Savolainen
J. Ditto 62. Dunlap Scott
P. Eatherly 63. J. H. Shaffer
R. Engel 64. M. J. Skinner
E. Ferguson 65. A. N. Smith
M. Ferris 66. P. G. Smith
K. Franzreb 67. I. Spiewak
P. Fraas 68. D. A. Sundberg
H. Gabbard 69. R. E. Thoma
B. Gallaher 70. D. B. Trauger
R. Grimes 71. A. M. Weinberg
G. Grindell 72. J. R. Weir
H. Guymon 73. M. E. Whatley
H. Harley 74. J. C. White
N. Haubenreich 75. G. D. Whitman
E. Helms 76. L. V. Wilson
W. Hoffman 77. Gale Young
L. Hudson 78-79. Central Research Liabrary (CRL)
R. Kasten 80-81. Document Reference Section (DRS)
J. Kedl 82-84. Laboratory Records (LRD)
S. Kirslis 85. Laboratory Records (LRD-RC)
1. Krakoviak
B. Lindauer
I. Lundin

External Distribution

 

86. C. B. Derring, AEC-ORNL
87-88. T.-W. McIntosh, AEC Washington
89. H. M. Roth, AEC-ORO
90-104. Division of Technical Information Extension
105. Laboratory and University Division, ORO