ORNL-4191.txt

 

LOCKHEED MARTIN ENERGY RESEARCH LIBRARIES

| qt |
i
i

\ :
;
! 4
[ h

:
i b

:
| R
\ i
‘

i
! 4
| :

B
B
i
"
B
i
iy
4
4
4
v
i

1

3 445k 05155905

 
 

 

 

 

 

 

 

 

 
ORNL-4191
UC-80 — Reactor Technology

Contract No. W-7405-eng-26

MOL TEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT
For Period Ending August 31, 1967

M. W. Rosenthal, Program Director

R. B. Briggs, Associate Director
P. R. Kasten, Associate Director

DECEMBER 1967

OAK RIDGE NATIONAL LABORATORY
Osk Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the

U. 5. ATOMIC ENERGY COMMISSION
TIN ENERGY AESEARGH LIBRARIES

[T

3 4454 0515905 7

i
i

 

 

 

 

 

 

 

 

 

 

 

 

 

§
i
!

i
'

 

 

 

 

b
i
This report is one of a series of periodic reports in which we describe briefly the progress of

the program. Other reports issued in this series are listed below. ORNL-3708 is an especially

useful report, because it gives a thorough review of the design and construction and supporting

development work for the MSRE. It also describes much of the general technology for molten-salt

reactor systems,

ORNL.-2474
ORNL.-2626
ORNL-2684
ORNL-2723
ORNL-2799
ORNL-2890
ORNL-2973
ORNL-3014
ORNL-3122
ORNL.-3215
ORNL-3282
ORNL.-3369
ORNL-3419
ORNL-3529
ORNL-3626
ORNL.-3708
ORNL-3812
ORNL-3872
ORNL-3936
ORNL-4037
ORNL-4119

Period Ending January 31, 1958

Period
Period
Period
Period

Ending October 31, 1958
Ending January 31, 1959
Ending April 30, 1959
Frnding July 31, 1959

Period Ending October 31, 1959

Periods Ending Jonuary 31 and April 30,

Period

Period

Period
Period
Period
Period
Period

Ending July 31, 1960
Ending February 28, 1961
Ending August 31, 1961
Ending February 28, 1962
Ending August 31, 1962
Ending January 31, 1963
Ending July 31, 1963

Period Ending January 31, 1964

Period
Period
Period
Period
Period
Period

Ending July 31, 1964
Ending February 28, 1965
Ending August 31, 1965
Ending February 28, 1966
Ending August 31, 1966
Ending February 28, 1967

1960
Contents

TN R O DU T N et e ettt et e s s e e be e bt et e oo s e e e oo e e ea e et e e e ae e et s 1

Sl M M A R Y oottt et ettt et et 2ttt nE o e e A A< e et et s e e e e e e e e e e e ae s e e et e e e e e a e e e e nen e e 3

PART 1. MOLTEN-SALT REACTOR EXPERIMENT

L. MO RE O E R A T IO S o oo e oo oo e oo oo e e e et 4 e e e ea et s es et 4o e n s £ bbb e 1t e e aeae e e 13
1.1 Chronological Account of Operations and Maintenance ...l e 13
1.2 Reactivity BalanCe ... 19

Balance s At P oWt e e 19
Balances at Zeto PowWer e 21
1.3  Thermal Effects of Operation ... ..ot et 22
Radiation Heating . oot ettt e e e e e ettt e e ettt e e e e 22
Thermal Cycle HISTOIY .o e et e e 22
Temperature MeaSurement. ... .. .o e e e 22
Fuel Salt Afterheat e e 24
1.4 Equipment PerformamOe et e et e e e 25
Heat T rans  er o e e 25
MR B O OIS i ittt et e ettt et et e e 26
Radiator EnC oS Ure e e e e e e 26
O GaS Sy S MS i i e ettt e ettt a e 27
Cooling Water SySteMS ..o ettt et e e e 29
Component Cooling SyStRIM ... e e e e n e 29
Salt Pump O11 SySTemMS oo i e e e e e 30
Electrical System............... e e e e 31
HEATOIS ittt et e e e e e e e e e e e e e e e et e e et e e e e et e ea e 31
Control Rods and Drives ... e 31
Salt Samplers ... et ee et Meeeeeeeteiessueseteeetateessatetenehe et eeentn Lt et bt s e e st e e .31
oM A It L. i ettt et 33

2. COMPONENT DEVE L O PMEN T e et e et ettt 36
2.1 Off-Gas SamM P T e e 36
2.2 Remote Maint@nanee ... e e e 36

Preparations for Shutdown After Run 11, 36
Evaluation of Remote Maintenance After Run 11 e 37
Repair of Sampler-Enricher and Recovery of Latch ..., 37
2.3 Decontamination SIS . o e e e 40

111
2.4  Development of a Scanning Device for Measuring the Radiation I.evel of

2.5

3. INSTRUMENTS AND CONTROLS

Remote Sources

Experiment with a S-curie '37Cs Gamma SOUICE ..ot
Gamma Scan of the MSRE Heat Exchanger .. ...
Gamma Energy Spectrum Scan of the MSRE Heat Exchanger.............................................
BT DS L
Mark-2 Fuel Pump ...l UUTTI e

Spare Rotaty Elements for MSRE Fuel and Coolant Salt Pumps
Stress Tests of Pump Tank Discharge Nozzle Attachment

MSRE Oil PUMPS o U

0il Pump Endurance Test

3.1 MSRE Operating ERPeriCnCe ... e e
Control System Components ........................oco. e

N AT I S T I O S oo e e e e

Safety SyStem .

3.2 Contiol System Design ...

4. MSRE REACT OR AN ALY SIS e e
B It oAU O IO o e

4.2 Neutron Energy Spectra in MSRE and MSBR ... .

4.3  Other Neutronic Characteristics of MSRE with 233U Fuel .. o

4.4  MSRE Dynamics with 233U Fuel

5. DESIGN

5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8

6. REACTOR PHYSICS

6.1

Critical Loading, Rod Worth, and Reactivity Coefficients ...
Fission Rate and Thermal Flux Spatial Distributions ... ...

Reactor-Average Fluxes and Reaction Cross Sections

Effect of Circulation on Delayed-Neutron PrecurSors ...
Samarium Poisoning Effeets ...

Cell Arrangement

Reactor

Fuel Heat Exchanger
Blanket Heat Exchanger
Fuel Drain Tanks

MSBR Physics Analysis
Optimization of Reactor Parameters
Useful Life of Moderator Graphite

Flux Flattening

Temperature Coefficients of Reactivity

PART 2. MSBR DESIGN AND DEVELOPMENT

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

41
43
43

45
45
45
46
46
46
47
47
47
47
48

48

50
50
50

54
54
55
55
58
60

61

63
63

71
76

79
79
81

82

82
82
84
37
7. SYSTEMS AND COMPONENTS DEVELOPMENT ... oo e 90
7.1 Noble-Gas Behavior in the MSBR .. e 90
7.2 MSBR Fuel Cell Operation with Molten Salt ...t e 95
7.3  Sodium Fluoroborate Circulating Loop Test ... e et e e e 95
T4 MSBR PUIIDS oottt cie et ettt ee e et e ettt et e et e e e ettt e er et s s et e e et e s en e s e e e 96

Survey of Pump Experience Circulating Liquid Metals and Molten Salts ... 96
Introduction of MSBR Pump Program. ... ..ot e et e e 96
Fuel and Blanket Salt Pumps oo ettt et ee e e e ens 97
Coolant Salt PUMDPS ..ot e ettt et e e et £t ane s 99
Water Pump Test Facility . ... e 99
Molten-Salt Bearing TeSES ..ottt s s ra e e e e e e eeneae s s 100
Rotor-Dynamics Feasibility Investigation ... 100
Other Molten-5alt PUIPS ... ettt et s e ae s st et e e oo e ers e e e e e e e st ee s e s mnmeeiaee e s 101

PART 3. CHEMISTRY

8. CHEMISTRY OF THE MORE L. ittt e e s e ee e et s e aene e e ce e crien e 102
8.1 MSRE Salt Composition and PUrity .. ...t e ee e e e 102

L GBIt i e et e e b e e et e et aenene e 103

Co01ant Sall Lo e e e e e e taa e e e e e e e et e en e naaae o 103

ISR Sat oo et ettt ba ettt e 2 e st et nnae e enea 103
Implications of Current Experience in Future Operations ... .. 108

8.2 MSRE Fuel Circuit Corrosion ChemiStry ...ttt 110

8.3 Adjustment of the UF, Concentration of the Fuel Salt ... 110

9. FISSION PRODUCT BEHAVIOR IN THE MSRE ... s 116
9.1 Fission Products in MSRE Cover Gas .. ... 116

9.2 Fission Products in MSRE Fuel ..o e 119

9.3 Examination of MSRE Surveillance Specimens After 24,000 Mwhr ... 121
Examination of Graphite ... e 121
Examination of Hastelloy N ..o e e s 124

Fission Product Distribution in MSRE . et e 125

9.4 Deposition of Fission Products from MSRE Gas Stream on Metal Specimens................. 128

9.5 Deposition of Fission Products on Graphites in MSRE Pump Bowl ... 131

10. STUDIES WITH LiF-BeF , MELTS ..o 136
10.1 Oxide Chemistry of ThFd—UF‘4 MBS et e e et e 136
10.2 Containment of Molten Fluorides in SiliCaA. ...t st e 137
CREMIESEIY .ottt ettt e e et e s e e s ek et 137
Spectrophotometric Measurements with Silica Cells ... 139

10.3 Electrical Conductivity of Molten Fluorides and Fluoroborates ... 140
11. BEHAVIOR OF MOLYBDENUM FLUORIDES ... ... e 142
11.1 Synthesis of Molybdenum Fluorides ... 142
11.2 Reaction of Molybdenum Fluorides with Molten LiF-BeF , Mixtures ... 143

11.3 Mass Spectrometry of Molybdenum Fluorides ... 144
vi

12. SEPARATION OF FISSION PRODUCTS AND OF PROTACTINIUM FROM
MOLTEN FLUORIDES

12.1 Extraction of Protactinium from Molten Fluorides into Molten Metals... ... ... . . .

12.2  Stability of Protactinium-Bismuth Solutions Contained in Graphite

12.3  Attempted Electrolytic Deposition of Protactinium

12.4 Protactinium Studies in the High-Alpha Molten-Salt Laboratory
Reduction of Iron Dissolved in Molten LikF'- ThF
Thorium Reduction in the Presence of an Iron Surface (Brillo Process)
Thorium Reduction Followed by Filtration
Conclusion

12.5 MSBR Fuel Reprocessing by Reductive Extraction into Molten Bismuth .............................

12.6 Reductive Extraction of Cerium from LiF-BeF, (66-34 Mole %) into Pb-Bi
Eutectic Mixture

13. BEMNAVIOR OF BF, AND FLUOROBORATE MIXTURES

13.1 Phase Relations in Fluoroborate Systems

13.2 Dissociation Vapor Pressures in the NaBF -Nal" System

13.3 Reactions of Fluoroborates with Chromium and Other Hastelloy N Constituents ...
Apparent Mass Transfer of Nickel

13.4 Reaction of BF3 with Chromium Metal at 6507 C e e e,
13.5 Compatibility of BF  with Gulfspin-35 Pump Oil at 150°F

14. DEVELOPMENT AND EVALUATION OF ANALYTICAL METHODS FOR
MOL TEN-SALT REAC T ORS e e e

14.1 Determination of Oxide in MSRE Salts ...
14.2 Determination of U3* in Radioactive Fuel by a Hydrogen Reduction Method ........................
14.3  In-Line Test Facility . ..., U OPR P
14.4 Electroreduction of Uranium(IV) in Molten LiF-Bel' -ZrF, at Fast Scan Rates

and Short TransSition TImMES Lo e e e e e e e bt nan e
14.5 Spectrophotometric Studies of Molten Fluoride Salts ...
14.6  Analysis of Off-Gas from Compatibility Tests of MSRE Pump Oil with BF , ...
14.7 Development of a Gas Chromatograph for the MSRE Blanket Gas ...

PART 4. MOLTEN-SALT IRRADIATION EXPERIMENTS

15. MOLTEN-SALT CONVECTION LOOP IN THE ORR ... e
15.1 Loop Description .. ... BT ST ST T U OO UP PP PRSPPI
15, O PEIA IO oo oo e
15.3  Operating TemMPEIALUIES .......cooiitiiiiiiie it e et e e e et
15.4  Salt Circulation by ConveTtIon i e e aeieee et e e e s s s a s a2 aaaassareree e aenaeans
15.5 Nuclear Heat, Neutron Flux, and Salt Power Density ...

LS. OTTOSIOM . oo

148

167
16.

17.

18.

19,

vii

15.7  Oxygen AnalySiS oo e 182
15.8 Crack in the Core Qutlet Pipe ... e e e 182
15.9 Cutup of Loop and Preparation of Samples ... 184
15.10 Metallographic EXaminalion .. ..o ik e 185
15.11 Isotope Activity Calculation from Flux and Inventory History ... 187
15.12 Isotope Activity BalanCe ... s 190
15.13 Uranium-235 Observed in Graphite Samples.. .. 190
15.14 Penetration of Fission Products into Graphite and Deposition onto Surfaces ... 192
15.15 Gamma Irradiation of Fuel Salt in the Solid Phase ... e 193

PART 5. MATERIALS DEVELOPMENT

MSRE SURVEILLANCE PROGRAM L ottt s e e en e e e e e e samn e 196
16.1 General Description of the Surveillance Facility and Observations on
Samples Removed ................oein e eeeeaseiutetete e e et eeaaaeante e e et ee 196
16.2 Mechanical Properties of the MSRE Hastelloy N Surveillance Specimens ... 200
GRA P H I TE ST U S et e et e e et e e e et eh et e ee it e bt ee e s et e e s an e eamsearneeaeeeean 208
17.1 Materials Procurement and Property Evaluation ... 208
17.2  Graphite Surface Sealing with MetalS ...t 211
17.3  Gas Impregnation of MSBR Graphites ... e 212
17.4  Trradiation of Graplit@ .. ... ettt 215
HA S T E L LY N ST U S . ittt e et e e ettt e s er e e e e 2 et eab ot e e et a2 etenesebre s s e baeaesaneenns 217
18.1 Improving the Resistance of Hastelloy N to Radiation Damage by
Composition MOifiC@tIONS . . oo ittt et e et e et et et ae e e e e s eearasaeneaeeeranees 217
18.2 Aging Studies on Titanium-Modified Hastelloy N ... ..ot et 217
18.3 Phase Identification Studies in Hastelloy N . e ee e eee e ne s 2195
18.4 Hot-Ductility Studies of Zirconium-Bearing Modified Hastelloy N...........ooinn . 221
18.5 Residual Stress Measurements in Hastelloy N WeldsS..........ooo e 223
18.6  COrtOSION SEUAIES ..o oottt e et et e et 226
FE] SLES it ittt e e ettt e e s et e e e n s e e e 226
00 Mt SaIES it ee e e et et e et e et eaeet s ettt e e e et e eataeeae e etan e e e aee e e nnenne 227
Equipment Modifications ... e e e e nans 230
18.7 Titanium Diffusion in Hastelloy N .. e e 230
18.8 Hastelloy N—Tellurium Compatibility ... e 233
GRAPHITE-TO-METAL JOINING oottt et s e e oo 236
19.1 Brazing of Graphite to Hastelloy N et e 236
JOIOE DIESIEIL ..ot et ettt e e e a e e e e st 236
Brazing Development. . ... e et et e e e e, 237

19.2 Compatibility of Graphite-Molybdenum Brazed Joints with Molten Fluoride Salts............... 237
20.
21.
22.
23.

24.
25.

26.

viil

PART 6. MOLTEN-SALT PROCESSING AND PREPARATION

VAPOR-LIQUID EQUILIRRIUM DATA IN MOLTEN-SALT MIXTURES . 239
RELATIVE VOLATILITY MEASUREMENT BY THE TRANSPIRATION METHOD . _..................... 242
DISTILILATION OF MSRE FUEL CARRIER SAL T o e 243
STEADY-STATE FISSION PRODUCT CONCENTRATIONS AND HEAT GENERATION

IN AN MSBR AND PROCE SSING P L AN o e e e e e 245
Heat Generation in a Molten-Salt Still ... TR 247
REDUCTIVE EXTRACTION OF RARE EARTHS FROM FUEL SAL T o 248
MODIFICATIONS TO MSRE FUEL PROCESSING FACILITY FOR SHORT DECAY CYCLE ......... 251
PREPARATION OF 233UF4.7L1F FUEL CONCENTRATE FOR THE MSRE .. ... 252
IO G oo 252
Equipment and Operatlons ... e 252
Introduction

The objective of the Molten-Salt Reactor Program
is the development of nuclear reactors which use
fluid fuels that are solutions of fissile and fertile
materials in suitable carrier salts. The program is
an outgrowth of the effort begun 17 years ago in
the Aircraft Nuclear Propulsion (ANP) program to
make a molten-salt reactor power plant for aircraft.
A molten-salf reactor — the Aircraft Reactor Ex-
periment ~ was operated at ORNL in 1954 as part
of the ANP program.

Our major goal now is to achieve a thermal
breeder reactor that will produce power at low cost
while simultaneously conserving and extending the
nation’s fuel resources. Fuel for this type of re-
actor would be ?33UF, or 235UrF'4 dissolved in a
salt of composition near 2LiF-BeF . The blanket
would be ThF4 dissolved in a carcrier of similar
composition. The technology being developed for
the breeder is also applicable to advanced con-
verter reactors.

Our major effort at present is being applied to
the operation of the Molten-Salt Reactor Experi-
ment {MSRE). This reactor was built to test the
types of fuels and materials that would be used in
thermal breeder and converter reactors and to pro-
vide experience with the operation and maintenance
of a molten-salt reactor. The experiment is demon-
strating on a small scale the attractive features
and the technical feasibility of these systems for
large civilian power reactors. The MSRE operates
at 1200°F and at atmospheric pressure and pro-
duces about 7.5 Mw of heat. Initially, the fuel
contains 0.9 mole % UF4, 5 mole % ZrF ,, 29 mole
% Ber, and 65 mole % LiF, and the uranium is
about 33% ?3°U. The melting point is 840°F. In
later operation we expect to use ?3°U in the lower
concentration typical of the fuel for a breeder.

The fuel circulates through a reactor vessel and
an external pump and heat exchange system. All
this equipment is constructed of Hastelloy N, a

nickel-molybdenum-chromium alloy with exceptional
resistance to corrosion by molten fluorides and
with high strength at high temperature. The re-
actor core contains an assembly of graphite moder-
ator bars that are in direct contact with the fuel.
The graphite is new material of high density and
small pore size. The fuel salt does not wet the
graphite and therefore does not enter the pores,
even at pressures well above the operating pres-
sure.

Heat produced in the reactor is transferred to a
coolant salt in the heat exchanger, and the coolant
salt is pumped through a radiator to dissipate the
heat to the atmosphere. A small facility installed
in the MSRE building will be used for processing
the fuel by treatment with gaseous HF and F,.

Design of the MSRE started early in the summer
of 1960, and fabrication of equipment began early
in 1962. The essential installations were com-
pleted and prenuclear testing was begun in August
of 1964. Following prenuclear testing and some
modifications, the reactor was taken critical on
June 1, 1965, and zero-power experiments were
completed early in July. After additional modifi-
cations, maintenance, and sealing of the contain-
ment, operation at a power of 1 Mw began in
January 1966.

At the 1-Mw power level, trouble was experi-
enced with plugging of small ports in control
valves in the off-gas system by heavy liquid and
varnish-like organic materials. These materials
are believed to be produced from a very small
amount of oil that leaks through a gasketed seal
and into the salt in the tank of the fuel circulating
pump. The oil vaporizes and accompanies the
gaseous fission products and helium cover gas
purge into the off-gas system. There the intense
beta radiation from the krypton and xenon poly-
merizes some of the hydrocarbons, and the products
plug small openings. This difficulty was largely
overcoimie by installing a specially designed filter
in the off-gas line.

Full power — about 7.5 Mw — was reached in
May. The plant was operated until the middle of
July to the eguivalent of about six weeks at full
power, when one of the radiator cooling bloweis —
which were left over {rom the ANP program — broke
up from mechanical stress. While new blowers
were being procured, an airay of graphite and metal
surveillance specimens was taken from the core
and examined.

Power operation was resumed in October with
one blower; then in November the second blower
was installed, and full power was again attained.
After a shutdown to remove salt that had acciden-
tally gotten into an off-gas line, the MSRE was
operated in December and January at full power for
30 days without interruption. A fourth power run
was begun later in January and was continued for
102 days until terminated to remove a second set
of graphite and metal specimens. The end of that
run came almost a year after full power was first
attained. In spite of the time required to replace
the blowers, the load factor for that year was 50%.
An additional operating period of 46 days during
the summer was interrupted for maintenance work
on the sampler-enricher when the cable drive mech-
anism jammed.

The reactor has performed very well in most re-
spects: the fuel has been completely stable, the
fuel and coolant salts have not corroded the Has-
telloy N container material, and there has been no
detectable reaction between the fuel salt and the
graphite in the core of the reactor. Mechanical
difficulties with equipment have been largely con-
fined to peripheral systems and auxiliaries. Ex-
cept for the small leakage of oil into the pump
bowl, the salt pumps have run flawlessly for over
14,000 hr. The reactor has been refueled twice,
both times while operating at full power.

Because the MSRE is of a new and advanced
type, substantial research and development effort
is provided in support of the operation. Included
are engineering development and testing of reactor
components and systems, metallurgical develop-
ment of materials, and studies of the chemistry of
the salts and their compatibility with graphite and
metals both in-pile and out-of-pile.

Conceptial design studies and evaluations are
being made of large power breeder reactors that
use the molten-salt technology. An increasing
amount of research and development is being di-
rected specifically to the requirements of two-
region breeders, including work on materials, on
the chemistry of fuel and blanket salts, and on
processing methods.
Summary

PART 1. MOLTEN-SALT REACTOR
EXPERIMENT ’

1. MSRE Operations

There were two long runs at full power during this
report period. The first, run 11, began in January
and lasted into May. After 102 consecutive days
of nuclear operation {over 90% of the time at full
power), the reactor was shut down to retrieve and
replace part of the graphite and metal specimens
in the core. The six-week shutdown also included
scheduled maintenance and annual tests of con-
tainment, instruments, and controls. Run 12 in-
cluded 42 days in which the reactor was at full
power continuously except for two brief periods
after spurious scrams. The run ended when the
fuel sampler-enricher drive mechanism jammed,
making it inoperative. The reactor was then shut
down, the drive was removed, and the sampler
latch, which had accidentally been severed from
the cable, was retrieved from the fuel pump bowl.

During the long runs at high power, interest
focused primarily on reactivity behavior and on
fuel chemistry. Slow changes in reactivity due to
fission product ingrowth and uranium burnup fol-
lowed expectations, and no anomalous effect was
observed outside the very narrow limits of pre-
cision of measurement (£0.02% 5k/k). Over 2 kg
of #3°U was added to the fuel during full-power
operation. The operation, using the sampler-
enricher, demonstrated quick but smooth melting
and mixing into the circulating fuel. Six additions
of beryllium metal were made to the fuel during
operation to maintain reducing conditions in the
salt. Corrosion in the salt systems was practically
nil, as evidenced by chromium analyses and exami-
nation of the core specimens. Studies of the be-
havior of certain fission products continued.

Component performance, on the whole, was very
good. There was no deterioration of heat transfer

capability or evidence of unusual heat generation
in the reactor vessel. Six thermocouples in the
reactor cell began giving anomalous readings during
mn 11, but all other themocouples showed no
tendency to become less accurate. The new off-
gas filter showed no increase in pressure drop and
apparently remained quite efficient. Restrictions
that built up slowly at the main charcoal bed in-
lets were effectively cleared by the use of built-in
heaters. While the reactor was down in May for
sample removal, two conditions that had existed
for some time were remedied: an inoperative posi-
tion indicator on a control rod drive and a leaking
space cooler in the reactor cell were replaced.
Until the sampler failure at the end of run 12, the
only delays in the experimental program due to
equipment difficulties were brief ones caused by
the main blowers and a component cooling pump.
A main blower bearing was replaced in run 11, and
shortly after the start of run 12 a main blower
motor mount was stiffened to alleviate a resonance
condition. Also at the start of run 12, low oil
pressure made a component coolant pump inopera-
tive until the relief valve was replaced. Secondary
containment leakape remained well within pre-
scribed limits, and there was no leakage from pri-
mary systems during operation. During the six-
month period, the reactor was critical 2925 hr
(66% of the time), and the integrated power in-
creased by 2597 to a total of 5557 equivalent full-
power hours.

2. Component Development

Extensive preparations were made for remote
maintenance in the May-June shutdown, including
training of 30 craftsmen and foremen. Work pro-
ceeded during the shutdown on two shifts. Pro-
cedures and tools prepared in advance worked well
in replacing cote specimens, repairing a control-rod
drive, replacing a reactor cell space cooler, and
inspecting equipment in the reactor cell.

When the sampler became inoperative, prepara-
tions were first made for shielding and containment
during replacement of the mechanism and retrieval
of the latch. The mechanism was then removed,
and a maintenance shield was set up for the latch
retrieval. Various long, flexible tools were de-
signed and tested in a mockup before use in the
sampler tube. The latch was grasped readily, but
difficulties were encountered in bringing it up until
a tool was designed that enclosed the upper end of
the latch. Tools removed from the sampler tube
were heavily contaminated, and a shielded carrier
with disposable liner was devised to handle them.
The sample capsule had broken loose from the
latch and cable and was left in the pump bowl after
an effort to retrieve it with a magnet failed. The
sampler repair and capsule retrieval were accom-
plished without spread of contamination and with
very moderate radiation exposures.

A sampler manipulator was successfully decon-
taminated for reuse in a test of decontamination
methods.

A scheme for mapping and identifying fission
product sources remotely was tested in the reactor
cell during the May shutdown. A lead-tube colli-
mator and an ionization chamber mounted in the
movable maintenance shield were vsed to map
gamma-ray sources in the heat exchanger and ad-
jacent piping; then a collimator and a gamma energy
spectrometer were used to characterize the souice
at various points. Results were promising.

Installation of the off-gas sampler was delayed
when the valve manifold had to be rebuilt because
of imperfect Monel—stainless steel welds.

Stress tests on a Mark-1 pump tank nozzle were
completed. Results compared favorably with cal-
culated stresses, and the design was judged ade-
quate. The Mark-2 replacement fuel pump tank for
the MSRE was completed, and preparations for a
test with salt proceeded.

Oil pumps removed from the MSRE were repaired
and tested. A replacement rotary clement for the
coolant salt pump was modified by seal welding a
mechanical seal that might have become a path for
oil leakage to the pump bowl.

3. Instruments and Controls

During the May shutdown a complete functional
check of instrumentation and control systems was

made. Preventive maintenance at that time included

modifying 139 relays and replacing capacitors in
33 electronic control modules. The type of com-
ponent failures that occurred did not compromise
safety or cause excessive inconvenience. Four of
the eight neutron chambers were replaced, one
because of a short and three because of moisture
inleakage.

Separate power supplies were installed for each
safety channel tc improve continuity of operation
and preclude a single compromising failure.
Various other modifications to circnits or com-
ponents were made to provide more infomation, to
improve performance, or to increase protection.

4. MSRE Reactor Analysis

As part of planning for future operation of the
MSRE, computational studies were made of the
neutronic properties of the reactor with 233U in
the fuel salt instead of the present %3°U (33%
enriched). The neutron energy spectrum was coni-
puted and compared in detail with that for a core-
lattice design being considered for a molten-salt
breeder reactor. The strong similarities indicate
that the results of the MSRE experiment will be
useful in evaluating design methods for the MSBR.
Other computations were made, with the following
results. The critical loading will be 33 kg of 233U,
compared with 70 kg of 2*°U in the first critical
experiment. Control rod worth will be higher by
a factor of about 1.3. The important reactivity
coefficients will also be considerably larger than
with 2°°U fuel. The themmal-neutron flux will be
up by more than a factor of 2, and the steady-state
samarivm concentmations will consequently be
Since more samarium will be left in the salt
from 23°U operation, it will act at first as a burn-
able poison, causing the reactivity to rise for
several weeks despite 233U burmup. Fission power
densities and importance functions will be similar
to those for 235U fuel. The effective delayed-
neutron fraction in the static system will be 0.0026,
decreasing to 0.0017 when fuel circulation starts.
(Corresponding fractions for 23°U are 0.0067 and
0.0046.)

The dynamic behavior with 233U was also ana-
lyzed from the standpoint of the inherent stability
of the system. Because of the small delayed-
neutron fraction, the neutron level responds more
sensitively to changes in reactivity, but the re-

lower.

sponse of the total system is such that the maigins
of inherent stability are greater with 233U fuel.
PART 2. MSBR DESIGN AND DEVELOPMENT

5. Design

The conceptual design wotk on molten-salt
breeder reactors during the past six months has
been concemed largely with a general advance in
the design of cells, containment, piping, and com-
ponents, and with stress analysis. In addition,
major effort has been devoted to preparation and
evaluation of a reactor design in which the average
core power density is reduced to 20 kw/liter from
the 40 kw/liter we were using during the previous
reporting period. At this lower power density the
core life before replacement is required would be
adequate even if the graphite behavior under irra-
diation is no better than that which has been
achieved to date. The performance at the lower
power density is more nearly representative of
current technology, and better perforimance should
be achievable as better graphite is developed.
Going from 40 to 20 kw/liter increases the capital
cost by $6/kwhr (electrical). No new design work
was performed on the steam system, but all salt
systems (fuel, blanket, and coolant) have been in-
vestigated more thoroughly than has been done
heretofore. Afterheat removal and thermal shield
cooling have been evaluated.

6. Reactor Physics

Parametric studies have been carried out which
reveal the dependence of MSBR performance on
such key design features as the average core power
density. They indicate that the power density may
be reduced from 80 w/cm? to 20 w/em?® with a
penalty not greater than 2%/year in annual fuel
yield or 0.1 mill/kwhr (electrical} in power cost.
At 20 w/cm? the life of the graphite will be in
excess of ten years.

Studies of power flattening in the MSBR core
show that a maximum-to-average power density
ratio of 2 or less can be achieved with no loss of
performance.

Calculations of tempetaiure ceefficients of re-
activity show that the lamge negative component
due to fuel expansion is dominant, and yield an
overall temperature coefficient of —4.3 x 107 %/°C.

7. Systems and Components Development

An analytical model was developed to compute
the steady-state migration of noble gases to the

graphite and other sinks in the MSBR. Work done
to date indicates that the mass tmansfer coefficient
from the circulating salt to the graphite is more
important than the diffusion ceefficient of xenon in
graphite in minimizing the poisoning due to xenon
migration to the graphite. In addition, the work
has shown that removal of xenon from molten-salt
fuels is strongly controlled by the mass transfer
coefficient to entrained gas bubbles as well as by
the surface area of those bubbles. Studies indicate
that the xenon poison fraction in the MSBR is
greater than 0.5% with the parameter values con-
sidered previously and that the poison fraction may
be about 1% with those parameters. The xenon
poisoning can be decreased slightly by increasing
the surface area of once-through bubbles, decreased
significantly by increasing the surface area of re-
circulating bubbles, decreased significantly by
increasing the mass transfer coefficient to circu-
lating bubbles, decreased proportionately by re-
ducing the graphite surface area exposed to salt,
and decreased significantly if the diffusion co-
efficient of xenon into graphite can be decreased to
1077 ft2/hr or less. Similar studies of the after-
heat in the graphite from the disintegration of the
radioactive noble gases and their decay products
show that the afterheat is affected by a variation of
the parameters in very much the same manuner as
the xzenon poisoning.

An experimental program was started to provide
an early demonstration of the compatibility of a
full-sized graphite fuel cell with a flowing salt
stream. The cell will include the graphite-to-
graphite and the graphite-to-metal joints.

Ag part of a program to qualify sodium fluoro-
borate (NaBF4) for use as a coolant for the MSBR,
an existing MSRE-scale loop is being prepared to
accept NaBF | as the circulating medium under
isothermal conditions. The principal alterations
are to the cover gas system, to include the equip-
ment necessary for handling and controlling the
required overpressure of BF ;. The objective will
be to uncover any problems associated with the
circulation of NaBF , and to devise and test suit-
able solutions or corrective measures.

A report was issued of a survey of experience
with liquid-metal and molten-salt pumps. An ap-
proach to producing the breeder salt pumps, which
invites the strong participation of U.S. indastry,
was evolved. The dynamic response and critical
speeds for preliminary layouts of the MSBR fuel
salt pump are being calculated, and a survey of
fabrication methods applicable to the pump is being
made. Preliminary layouts are being made of
molten-salt bearing and water pump test facilities
for the MSBR fuel salt pump. The pump with the
molten-salt bearing was fitted with a new salt
bearing and a modified gimbals suppost and was
satisfactorily tested with oil.

PART 3. CHEMISTRY
8. Chemistry of the MSRE

Results of regular chemical analyses of MSRE
fuel, coolant, and flush salts showed that after
40,000 Mwhr of power operation generalized corro-
sion in the fuel and coolant circuits is practically
absent and that the salts are currently as pure as
when charged into the reactor. Although statisti-
cally satisfactory, fuel composition analyses are
much less sensitive to variations in uranium con-
centration than is the reactivity balance, and im-
proved methods will be required for future MSR
fuels whose uranium concentrations need to be
only 0.25 that of the MSRE fuel salt.

A program for adjusting the relative concentra-
tion of U3*/XU to approximately 1.5% by addition
of small amounts of beryllium metal to the MSRE
fuel was completed. Specimens of fuel salt taken
from the pump bowl during this program showed
occasional temporary perturbation in the chromium
concentration, giving evidence that the identity
and concentrations of the phases present at the
salt-gas interface of the pump bowl are not neces-
sarily typical of the salt in the fuel circuit.

9. Fission Product Behavior in the MSRE

A second set of graphite and Hastelloy N long-
terrm surveillance specimens, exposed to fissioning
molten salt in the MSRE core for 24,000 Mwhr, was
examined and analyzed. As for the first set, ex-
posed for 7800 Mwhr, examination revealed no
evidence of chemical damage to the graphite and
metal. Very similar fission product behavior was
observed, with heavy deposition of the noble-
metal fission products — *"Mo, 132Te, '?%Ruy,
106Ru, 2°Nb, and !'1Ag — on both metal and
graphite specimens. A refined method of sampling
of the graphite surfaces showed that about 99% of
the ?°Mo, ?°Nb, '%3Ru, and '9%Ru was deposited
within the outer 2 mils of the surface. By con-
trast, appreciable fractions of the !32Te, ?5Zr,

140B4 and %Sy penetrated 50 mils or farther into
the graphite.

Ten additional exposures of metal specimens in
the MSRE pump bowl and five additional samplings
of pump bowl cover gas were cariied out. The re-
sults from tests under normal operating conditions
were similar to those of previous tests; they
showed heavy depositions of noble mctals on
specimens exposed to the cover gas and the fuel
phase. Of special interest were the observations
under unusual operating conditions: nearly as
much deposition occurred after reactor shutdown
with the fuel pump stopped and with the reactor
drained as occurred under nommal conditions.

Analysis of the time dependence of fission
product deposition on Hastelloy N indicated that
there was a short-temm rapid process that reached
saturation in about 1 min and a long-term process
that proceeded at slow constant rate for over
3000 hr. Results from only three exposures of
graphite specimens indicated that deposition rate
decreased with exposure time for long exposures.

10. Studies with LiF-BreF2 Melts

Equilibrium data have been obtained for the
reaction

U*HE) + Th*™ (o) == Th*™ (1) + U*¥0) ;

(XTh)f(XU)o

where (f) indicates that the species is dissolved in
molten 2LiF . BeF  and (o) indicates that the
species is in the sparingly soluble oxide solid
solution (U, Th)Oz. These data show that, over
the interval 0.2 to 0.9 for mole fraction uranium in
the oxide phase and 0.01 to 0.07 for mole fraction
Th** in the molten fluoride, the equilibrium con-
stant is in excess of 1000. Uranium is strongly
extracted from the fluoride phase to the oxide solid
solution. It seems very likely that protactinium is
even more sirongly extracted. If so, equilibration
of an LLiF-BeF Z—ThF4-UF 4»PaF4 melt with the prop-
er (stable) (U,Th)O2 solid solution should remove
protactinium, Recovery of 2?3Pa from a one-region
breeder fuel would, accordingly, be possible.
Vitreous silica (510 ,) has been shown to be a
feasible container material for Lii*fi—BeF2 melts,
especially when the system is stabilized by a
small overpressure of SiF . Preliminary measure-
ments have shown that the solubility of SiF

ZLiF - BeF ) is moderately low (about 0.035 mole
of 5K per kilogram of melt per atmosphere of

SiE ) at 5504C and at least threefold less at 700°C.

No ev1dence for silicon oxyfluoudes has been ob-
served. [t appears that, at least for temperatures
near 500°C and lor short times, an electrically in-
sulating and optically transparent coatainer for
LiF-BeF , solutions is available.

Optical cells of transparent Si0, have been used
toy establish, with a Cary model 14M spectrophotom-
aeter, that solutions of UF , in 2LiF . Ber under
400 mm of SiF were stable for 48 hr at tempera-
tures up to TOOOC These ztudies have led to a
considerably more precise definition of molar ab-
sorptivity of U*" as a function of temperature and
inc ident wavelength than had previously been
possible with windowless optical cells. In similar
spectrophotometric studies with silica cells, the
solubility at 550°C of Cr®” in 2LiF . BeF | was
shown to be at least 0.43 mole %. ‘

Silica apparatus has also been shown to be
feasible for studies of electrical conductivity of
2L1F-BE.-F2, of the I_Ji.F‘fI'}‘ll:“4 eutectic mixture, and
of NaBF4. Preliminary values obtained in this
study are to be refined in the near future by use of
an improved cell design which will provide a much
longer current path length through the melts.

11. Behavior of Molybdenum Fluorides

Molybdenum hexafluoride, the only commercially
available fluoride of this element, has been used
as raw material for preparation of MoF | and MoF ..
Direct reduction of Mo¥F . by molybdenum metal in
glass appatatus at 30 to 100°C yields, as shown by
other investigators, MoF _ of good quality. Dis-
proportionation of Mo under vacuum at 2000C
yields pure MoF , as the solid residue; we have
prapared several samples of the material by this
method, which seems not to have been described
before. The MoF | reacts on heating with Lil to
form at least two binary compounds; the optical
and x-ray characteristics of these materials have
been determined, but their stoichiome try has not
vet been established.

Molybdenum hexafluoride has been shown to re-
act rapidly with UF _ in Lil"-BeF | solution and
with nickel in conta(.t with such qolutmnf-, Molyb-
denum trifluoride has been shown to be relatively

stable when heated to 700°C under its vapor in
sealed capsules of nickel or copper. However,
when such heating is done in the presence of

2LiF . BeF , the MoF | reacts readily with nickel,
yielding l\IiF‘2 and Mo; the reaction is less marked
if the capsule is of copper. Molybdenum trifluoride
has been shown to react completely at 500°C with
UF3 in LiF . Ber mixtures; the products are UF4
and Mo.

Vaporization behavior of MoF | has been shown,
by examination with a time-of-flight mass spec-
trometer, to be complex and temperature dependent.
The behavior observed may suggest that the free
energies of formation (per fluorine atom) of these
intemediate molybdenum fluorides are so nearly
equal that the descriptive chemistry of these sub-
stances is dominated by kinetic factors.

12. Separation of Fission Products and of
Protactinivm from Molten Fiuorides

Very dilute solutions of %%3Pa in bismuth have
been shown to be stable for extended periods in
graphite containers, but the protactinium appears
to be strongly adsorbed upon any added metal or
any precipitated phase. More than 90% of the con-
tained #?3Pa has been successfully transferred
from LiF-BeF ThF blanket mixtures through a
molten Bi-Sn metdl phase and recovered in an
LiF-NaF-K¥F salt mixture by adding Th reductant
to the blanket mixture and oxidant HF to the re-
covery salt; successful operation of this experi-
mental assembly suggests that a redox transfer
process for Pa should be feasible. More concen-
trated solutions of ?3'Pa plus 2°*Pa in realistic
blanket mixtures continue to be successfully re-
duced to insoluble solid material by the addition
of thorium metal. Passage of such reduced mix-
tures through sintered nickel filters produces a
virtually protactinium-free filtrate but fails to
localize the Pa in a readily manageable form.

Preliminary attempts to reduce 27 'Pa plus
233pa solutions in simulated blanket mixtures to
inscluble materials by electrochemical means were
unsuccessful; such reduction certainly seems
feasible, and the experiments will continue.

Rate-earth fluorides in uraninm-free LiF-Bel
solutions ate readily reduced to the metallic state
and are transferred to the molten bismuth upon
contact with a molten alloy of lithium in bismuth.
Preliminary evidence suggests that separations of
uranium from the rare earths and, perhaps, of
uranium from zirconium may be possible by this
reductive extraction technique. Material balances
on the reductant are poor in experiments to date;
this problem will receive additional attention in
future experiments. Use of a Pb-Bi alloy with

51 at. % Bi as a substitute for pure bismuth in
similar extractions gave generally unsatisfactory
results.

13. Behavior of BF; and Fluoroborate Mixtures
Recrystallization of NaBF , and of KBF | from
dilute (usually 0.5 M) aqueous hydrofluoric acid
solutions yields preparations which melt at higher
temperatures and which are almost certainly more
pure than those reported by previous investigators.
These preparations, and our standard differential
thermal analysis and quenching techniques, have
been used to examine the binary systems Nal'-
NaBF  and KF-KBF , and the NaBF -KBF , and

Nal-KBF  joins in the temary system NaF-KF-BF ..

The NaF—NaBF4 and the KF-KBF |, systems show
single simple eutectics; phase diagrams which we
consider to be correct, but which are at variance
with data from other laboratories, are presented in
this report.

Pressures of BF | in equilibrium with NaF-NaBF
mixtures over the composition interval 65 to 100
mole % NaBF, have been measured at temperatures
of interest to the MSRE. Introduction of chromium
metal chips into the system with the NaF-NaBF |
eutectic (92 mole % NaBF ) led to perceptible re-
action. After the sample bad been above 500°C
for 26 hr, the BF | pressure observed was twice that
from the melt without added chromium. Subsequent
examination of the materials revealed NaCrF _ as
one of the reaction products with an additional un-
identified black material also present. Other ex-
periments with Hastelloy N, iron, and molybdenum
showed little or no visual evidence of attack;
these tests (for which dissociation pressure was
not monitored) did show perceptible weight losses
for both the Hastelloy N and iron specimens. In
addition, nickel vessels used in the routine de-
composition pressure measurements showed shiny
interior surfaces, which suggest that some mass
transfer had occurred.

Boron trifluoride gas has been showin to react at
650°C with essentially pure metallic chromiam in
the foim of thin flakes. Weight gain of the chro-
mium sample increased linearly with square root of
time; x-ray diffraction techniques have revealed the

mixed fluoride Cr¥ o CrF , as a reaction product,

Gulfspin-35 pump oil (the type used in MSRE) has
been exposed for 600 hr at 150°F to helium gas
containing 0.1 vol % BF‘E. In these tests the gas
mixture was bubbled at 1 liter/min through 1.5
liters of the lubricating oil. Some discoloration of
the oil was noted, but there was no distinguishable
increase in viscosity.

14. Development and Evaluatien of Analytical
Methods for Mclten-Salt Reucters

The determination of oxide in highly radioactive
MSRE fuel samples was continued. The replace-
ment of the moisture monitor cell was the first
major maintenance performed since the oxide equip-
ment was installed in the hot cell.

The U*" concentrations in the fuel samples run
to date by the transpiration technique do not re-
flect the beryllium additions which have been made
to reduce the reactor fuel. This may be accounted
for by an interference stemming from the radiolytic
generation of fluorine in the fuel samples. This
problem will receive further investigation. Experi-
mental work is also being carried out to develop a
method for the remote measurement of ppm concen-
trations of HF in helium or hydrogen gas streams.

Design work was continued on the experimental
molten-salt test loop which will be used to evaluate
electrometric, spectrophotometric, and transpiration
methods for the analysis of flowing molten-salt
streams.

Controlled-potential voltammetric and chrono-
potentiometric studies were carried out on the re-
duction of U(IV) in molten fluoride salts using a
new cyclic voltammeter. It was concluded that
the U(IV) - U(III) reduction in molten LiF-Bek -
ZrF4 is a reversible one-electron process but that
adsorption phenomena must be taken into account
for voltammetric measurements at fast scan rates
or for chronopotentiometric measurements at short
fransition times.

An investigation of the spectrum of U(VI) in
molten fluoride salts has been initiated. It was
found that the spectrum of Na,UF  dissolved in
LiF-Bel", in an SiO, cell with SiF, overpressure
was identical to the spectrum of U0 2F , dissolved
under identical conditions. It appears that the
equilibrium concentration of 02~ may be sufficient
to react with the components of the melt. An at-
tempt to use the 510 ,-5iF | system in the spectro-
photometiic investigation of electrochemically
generated species in molten fluorides also met
with difficulties. The SiF, overpressure interferes
with cathodic voltammetric studies by causing
very high cathodic currents.

It is planned to install a spectrophotometric
facility with an extended optical path adjacent to
a high-radiation-level hot cell to permit the obser-
vation of absorption spectra of highly radioactive
materials. The basic spectrophotometer and asso-
ciated equipment have been ordered.

Measurements were made of increases in hydro-
carbon concentrations of an He-BF | gas stream
after contact with MSRE pump oil. A thermal con-
ductivity detector was used to monitor the BF
concentration in the test gas stream.

Development studies are being made on the de-
sign of a gas chromatograph to be used for the
continuous determination of sub-ppm, low-ppm, and
high concentrations of permanent gas impurities and

water in the helium blanket gas of the MSRE. This .

problem of analyzing radioactive gas samples
prompted the design and construction of an all-
metal six-way pneumatically actuated diaphragm
valve, A helium breakdown voltage detector with
a glass body was designed and constructed to per-
mit the observation of the helium discharge. Under
optimum conditions, this detector has exhibited a
minimum detectable limit below 1 ppb of impurity.
It appears to be possible that the detector will
also operate in the less-sensitive mode necessary
for the determination of high-level concentrations
of impurities in the blanket gas.

PART 4. MOLTEN-SALT IRRADIATION
EXPERIMENTS

15. Molten-Salt Convection Loop in the ORR

Irradiation of the second molten-salt convection
loop in beam hole HN-1 of the Oak Ridge Research
Reactor was terminated after the development of
8.2 x 108 fissions/cc in the 7LiF-BeF2rZrF4«UF4
(65.3-28.2-4.8-1.7 mole %) fuel. Average {uel
power densgities up to 150 w per cubic centimeter
of salt were attained in the fuel channels of the
core of MSRE-grade graphite.

The experiment was terminated after radioactivity
was detected in the secondary containment systems
as a result of gageous fission product leakage from
a crack in the core outlet tube. Salt samples were
removed routinely during irradiation, and the fuel

salt was drained from the loop before removal from
the reactor beam hole.

Me tallurgical examination revealed a nonductile
crack in the Hastelloy N core outlet pipe. The
loop was made from unmodified material, and we
believe that the failure was caused by loss of
strength and ductility under operating conditions
of high temperature (™~730°C) and irradiation
(™5 x 101° avi).

The distribution of various fission products in
the system was obtained by the examination of sam-
ples of core graphite and loop metal. Some ad-
herence of fuel salt to the graphite and entry into
cracks in the graphite were found. Molybdenum and
tellurium (and probably ruthenium) were largely
deposited on graphite and metal surfaces. Other
isotopes, including 131y 89¢, 14%p, and 2°Nb,
which could have been transported as gases, were
found to have penetrated the graphite.

Solid MSR fuel salt (LiF-BeF -Z¢F -UF , about
£65-28-5-2 mole %) was subjected to very high-
intensity gamma irradiation in a spent HFIR fuel
element at a temperature of 320°C to determine
possible radiation effects on the salt and its com-
patibility with graphite and Hastelloy N. Post-
irradiation examination did not reveal any signifi-
cant effects.

PART 5. MATERIALS DEVELOPMENT
16. MSRE Surveillance Program

The materials surveillance program for following
the changes in the properties of the two major
MSRE structural materials — graphite and Hastel-
loy N — has been maintained. Graphite and metal
specimens were removed for examination on July 28,
1966 (7820 Mwhr), and on May 9, 1967 (32,450
Mwhr). We plan to run various physical and me-
chanical property tests an the graphite, but we
have not considered thig an urgent item since the
doses are quite low (approximately 1 x 10%! neu-
trons/cm?, E > 0.18 Mev). Extensive mechanical
property tests have been run on the Hastelloy N.
Its high-temperature creep-rupture life and rupture
ductility were reduced, but these changes are
quite comparable with what we have observed for
Hastelloy N irradiated in the ORR. There was a
slight reduction in the low-temperature ductility,
which we attribute to the irradiation-induced pre-
cipitation of intergranular M _C.
A set of Hastelloy N specimens located outside
the reactor core was removed on May 9, 1967, after
about 11,000 hr of exposure to the cell environ-
ment. There was some surface oxidation, about
0.003 in., but no evidence of nitriding.

The surveillance program has been expanded to
include some heats of modified Hastelloy N, and
specimens that contained 0.5% Ti and 0.4% Zr were
removed from the core on May 9, 1967. The me-
chanical testing has not been completed, but metal-
lographic studies revealed no significant corrosion.

17. Graphite Studies

Much of our materials program is directed toward
finding suitable materials for future molten-salt
reactors. In our present concept of a molten-salt
breeder reactor, graphite tubes will be the struc-
tural element that separates the fuel and fertile
salts. This will require a graphite with very
special properties, particularly with respect to a
small pore spectrum, low gas permeability, and
dimensional stability under high neutron doses.
We are looking closely at many grades of graphite
that are available from commercial vendors. Sev-
eral grades look promising, but none completely
satifies our requirements.

L.ow gas permeability in graphite seems very
hard to obtain, and we feel that producing mono-
lithic graphite bodies with helium permeabilities of
<10~ % em?/sec will be quite difficult. However,
we may be able to satisfy this requiremeat by
surface-sealing techniques. Our initial efforts
with pyrocarbon and molybdenum sealants look
very promising. The proof test will be to demon-
strate that graphite sealed in this manner retains
its low permeability after neutron exposure.

The dimensional instability of graphite continues
to be a major problem. We are analyzing very
critically all the data obtained to date in an effort
to determine what types of graphite appear most
stable. We have started our own experiments in
the HF[RR, where we can obtain doses of 4 x 1022
nvt (E > 0.18 Mev) in one year.

18. Hastelloy N Studies
Although the Hastelloy N will not be in the core,

it will be located in peripheral arcas where it will
receive rather high doses. We have found that the

10

properties of this basic alloy can be improved sig-
nificantly by slight modifications in the composi-
Reducing the molybdenum from 16 to 12%
suppresses the formation of M C, and small
amounts (approximately 0.5%) of either Ti, Zr, or
Hf improve the resistance to radiation damage.
The titanium-modified alloy looks very good, and
we are proceeding further with its development.
Experiments are being run to determine the sta-
bility of this alloy at elevated temperatures, and
specimens aged at 1200 and 1400°F actually show
some improvement in ductility. Our electron micros-
copy studies show that TiC and Ti O precipitates
‘“solution annealed’ condition.

tion.

are present in the
The changes in distribution and quantity of these
precipitates in the aged specimens will be de-
terminad.,

Since titanium can be leached from Hastelloy N
by fluoride salts in a manner analogous to chromium,
we must consider the corrosion resistance of the
titanium-modified alloy. The process is likely
controlled by the diffusion rate of titanium in
Hastelloy N, and measurements we have made in-
dicate that titanium diffuses at a rate comparable
with that of chromium at 2000°F. Thus, our small
titanium addition will probably not adversely affect
the corrosion resistance of the alloy.

Our welding studies have shown that Ti and Hf
additions to Hastelloy N do not affect the welda-
bility adversely, but that Zr is quite detrimental.
However, the postirradiation ductility of the Zr-
modified alloy is quite high, and we have tried to
find a suitable technique for joining this alloy.
Since residual stresses from welding can cause
dimensional changes and even cracking, we have
developed a technique for measuring these stresses.
We now can adjust welding parameters and post-
weld heat treatments to minimize the magnitude of
the residual stresses.

We have two thermal convection loops munning
which contain an I,,iF-BeFQ-ZrF4—UP‘4-ThF4 fuel
salt. One loop is constructed of Hastelloy N and
has operated satisfactorily at 1300°F for 47,440 hr.
The second loop is constructed of type 3041 stain-
less steel with removable hot-leg specimens of the
same material. The loop has operated at 1250°F
for 36,160 hr, and the removable specimens have
indicated a corrosion rate of 2 mils/year. Two
loops have also been run using NaF—KF-BFg, which
is a possible coolant salt. One loop, constructed
of Croloy 9M, plugged in 1440 hr because of mass
transfer and the deposition of iron crystals in the
cold leg. The second loop, of Hastelloy N, was
terminated as scheduled after 8765 hr ot operation,
but it was partially plugged and considerable cor-
rosion had occurred. Effort is being concentrated
on the compatibility of Hastelloy N and the fluoro-
borate salts.

Of the various fission products that will be pro-
duced in the MSBR, tellurium appears to be the
only one that may not be compatible with Hastelloy
N. We have coated specimens with tellurium and
annealed them for long periods of time. There is a
very slight penetration of tellurium into the metal,
but the mechanical properties are not affected
adversely for the conditions investigated.

19. Graphite-to-Metal Joining

We are investigating several joint designs for
brazing graphite to Hastelloy N. One approach has
proven successful, but we are trying to develop a
cheaper and simpler type of joint. One promising
braze is the 60 Pd-35 Ni--5 Cr alloy, and we have
n corrosion tests that confirm its compatibility
with molten salts.

PART 6. MOLTEN-SALT PROCESSING
AND PREPARATION

The concept of processing the fuel salt con-
tinuously by fluorination and distillation persists
esgentially in its initial form. The critical opera-
fion in this flowsheet is the distillation of the
carrier salt, and most of the effort in this period
has been concentrated here.

20. Vapor-Liquid Equilibrium Data in
Molten-Salt Mixtures

The relative volatilities of ZrF4, NdF‘S, CeF ,
BaFZ, YFE,, LaFS, and Ser in the ternary system
REF -LiF-Bel have been measured using an
equilibrium still at 1000°C. Most values are in
close agreement with those predicted by Raoult’s

law.

11

21. Relative Yolatility Measurement by

the Transpiration Method

Results from initial experiments using the trans-
piration method for measuring the vapor pressure
of LiF-BelF  over the range 920 to 1055°C con-
formed to the correlation of log vapor pressure vs
1/T. These data are also in good agreement with
data obtained from equilibrium still measurements.

22. Distillation of MSRE Fuel Carrier Salit

Equipment for demonstration of vacuum distilla-
tion using MSRE fuel salt has been built and as-
sembled in its supporting framework. It is being
installed in a test facility to perform nonradioactive
experiments. 'FThis unit has been subjected to ex-
tensive examination, and numerous dimensional
measurements have been taken to afford a reference
for postoperational examination. Only if the unit
appears to be in good condition after nonradio-
active tests will it be installed at the MSRE for
carrier salt distillation demonstration.

23. Steady-State Fission Product Concen-
trations and Heat Generation in an

MSBR and Processing Plant

A computer code that considers individual fission
products has been prepared to provide information
on fission product heat generation in the various
components of an MSR processing plant. This pro-
gram allows [or the generation and removal of fis-
sion products by several different processes which
can differ according to their chemical nature. It
has been used to compute heat-generation curves
for a fuel processing still, and the results com-
pare favorably with other programs based on gross
fission product heat data.

24. Reductive Extraction of Rare Earths
from Fuel Salt

One alternative to the distillation process for
decontaminating MSBR fuel salt uses the reductive
extraction of the rare earths from the salt after
uranium has been recovered by fluorination. Ex-
periments have been performed using lithium dis-
solved in molten bismuth as a reductant. Although
the results are complicated by an unexplained loss
of metallic lithium to the salt phase, the distribu-
tion of rare earths between the salt and metal
phases can be correlated with the lithium metal
concentration in the metal phase.

25. Modifications to MSRE Fuel Processing
Facility for a Short Decay Cycle

Provisions are being made for processing the
MSRE fuel salt for uranium recovery on the shortest
possible cycle after shutdown of the MSRE in early
1968. The flush salt will be processed first, and
then the fuel salt will be treated with H -HF to
establish the oxygen concentration. Allowing time
for these operations, the fuel salt may be fluorinat-
ed after 35 days (initial plans called for a 90-day
cooling time). This shorter cooling time requires
some modification of the processing facility at the
MSRE. The higher concentration of iodine requires
improvement of the off-gas system, and the pres-

12

ence of molybdenum requires increased shielding
around the UF . product absorbers.

26 Preporotion of 233UF4-7LiF Foel
Concentrate for the MSRE

Refueling and operating the MSRE with 233U
fuel early in 1968 is planned; this will require
approximately 40 kg of ?°°U as *¥3UF -TLif (27
and 73 mole %) eutectic salt. This fuel concen-
trate will be prepared in a cell in the TURF build-
ing because of the radiation from the 32U daugh-
ters in the 33U, The uranium will arrive as an
oxide in cans, which will be opened and dumped
into a reaction vessel. Lithium fluoride will be
added, and the mixture will be treated with hydro-
gen and finally HF to produce the entectic melt.
Three 12-kg 23°U batches will be prepared for the
major additions to the barren MSRE salt and one
7-kg *#3U batch will be loaded into 60 enriching
capsules. The engineering design is almost com-
plete, and most of the equipment has been fabri-
cated.
Part 1.

Molten-Salt Reactor Experiment

P. N. Haubenreich

1u the six-month period reported here the promise
of the MSRE as a practical and reliable reactor
was, in a large measure, realized. From the be-
ginning, operation of the reactor had strengthened
our confidence in the basic technical feasibility
of molten-salt reactors. At first, however, me-~
chanical problems with the peripheral equipment
did not allow the practical virtues of the molten-
salt system to be emphasized by a long period of
sustained operation at high power. But the delays
were not excessive, and within a year after the
first operation at full power, the reactor did com-
plete a very satisfactory demonstration of sus-
tained operation, Between Jjanuary and May 1967,
there were 102 consecutive days of auclear opera-

1.

tion with remarkably few difficulties; operation
was terminated only because of scheduled re-
moval of specimens from the core,

The first part of this report details the experience
with operation and maintenance of the MSRE. Then
it covers development efforts directly related to
the reactor. - Finally there is a section relating to
a future experiment, namely, the predicted nuclear
characteristics of the MSRE with 223U fuel. We
plan to strip the present uranium from the fuel
salt and replace it with *??U in the spring of
1968 in an experiment that promises to lend worth-
while support to design calculations for *3*U-
fueled breeder reactors.

MSRE Operations

P. N. Haubenreich

1.1 CHRONGLOGICAL ACCOUNT OF
OPERATIONS AND MAINTENANCE

Robert Blumberg C. K. McGlothlan
J- L. Crowley R. R. Minue

R. H. Guymon M. Richardson
P. H. Harley H. C. Rolier

T. L. Hudson R. C. Steffy

A. 1. Krakoviak B. H. Webster

Run 11 began in January and continued into May
for 102 consecutive days of nuclear operation (see
Fig. 1.1). Between February 1 and May §, the
reactor was at full power (7.3 Mw) 93% of the time.
The longest interruption in full-power operation
was four days, initiated because of excessive
vibration in a bearing on a main blower. On this

13

occasion the reactor operated at 5.9 Mw on one
blower for a day; then the power was lowered to
10 kw for the beating replacement and was held
there for three days to allow the xenon to strip
out for a special reactivity measuremeni. Once
the power was reduced to 10 kw for 7 hr to permit
replacement of the coolant off-gas filter, and
once one blower was off for 9 hr after unusual
cold (97F) caused bearing vibrations, Twice,
spurious scrams caused by false signals produced
brief interruptions (1 to 2 hr). Three times the
power was lowered for periods from 6 to 38 hr for
experimental purposes.

In addition to demonstrating the capability of the
MSRE for sustained operation, the lengthy period
at high power in run 11 afforded usefu! information
on long-term reactivity changes due to samarium
 

POWER (Mw)

 

 

 

 

 

 

————

SALT CIRCULATING { FUEL

 

Fig. 1.1.

There were no signif-
Fuel chemistry, in

and other fission products.
icant reactivity anomalies.
particular the behavior of volatile fission products,
was investigated throughout the run by taking aa
average of three samples per week from the fuel
pump. Twice, a few grams of beryllium was added
to the fuel to counteract the tendency of the fis-
sjon process to make the salt chemically less re-
ducing. An adequate margin against corrosive,
oxidizing conditions was maintained, and chromium
analyses showed practically no corrosion,

Run 11 ended with a scheduled shutdown to re-
move core samples. After the fuel system was
flushed and cooled, the array of metal and graphite
specimens was removed to a hot-cell facility,
There the array was disassembled, new samples
were substituted for one of the three stringers,
and the array was reassembled. Meanwhile,
maintenance and inspection were cairied ont on
the reactor. Several maintenance and inspection
jobs were performed in the reactor cell with semi-
remote techniques (see p. 37 on development and
evaluation of procedures and tools). One control
rod drive was removed for replacement of a posi-
tion-indicating device and inspection of the
grease, A set of metallurgical specimens adjaceat
to the reactor vessel was replaced, and a new
americium-curium-beryllium neutron source was
installed in the source tube in the thermal shield,
One of the two space coolers in the reactor cell
was replaced after it proved to be leaking as
suspected. The maintenance shield was then set
up over the salt heat exchanger to test an ex-
perimental device for mapping radiation sources.

    

 

14

ORNL-DWGC 57-11785

 

 

 

 

Outline of MSRE Power Operntion from January to August 1967,

Equipment in the reactor cell was viewed in an
attempt to determiiie the cause of some anomalous
thermocouple readings, and four new thermocouples
to read ambient temperature were installed, White
dust observed in the reactor cell was analyzed
and found to be aluminum oxide, presumably from
thermal insulation in the cell, but the source
could not be located, After the core samples were
reinstalled and the inspection of the cells com-
pleted, the reactor and drain-tank cells were
sealed on June 9.

Other maintenance work at the same time in-
cluded overhaul and repair of the radiator door
brakes and enclosure, overhaul of the main blower
motors, inspection of the main blowers, preventive
maintenance on component coolant pump 2, re-
placement of a differential pressure element on the
fuel off-gas system, and planned modifications
and improvements in the instrumentation and con-
trol systems.

During the shutdown the annual tests of instru-
mentation and control systems and secondary con-
tainment were conducted. The latter included
leak-testing all containment valves and measuring
the reactor cell leak rate at 20 psig.

The shutdown work was completed ahead of
schedule, and nuclear operation in run 12 began
on June 19, 39 days after the reactor was taken
subcritical at the end of run 11.

Run 12 was another period of extended opera-
tion at full power. The first week of nuclear
operation was marked by difficulties with some of
the equipment. These were remedied, however,
and there followed 42 days in which the reactor
was at full power continuously except for two
brief periods following scrams — one accidental
and one from loss of normal power due to lightning,.

Part of the delay in the lirst week was caused
by vibration of a main blower motor. After over-
haul the motor had a slight imbalance which would
have been acceptable, except that the resonant
frequency of the motor mount was very neart the
operating speed. Stiffening the mount by welding
on reinforcing plates solved this problem. During
the first weekend, a component coolant pump lost
oil pressure, so it was necessary to switch to the
standby. A few hours later the reactor scrammed
when lightning knocked out the main power supply
and damaged a period safety amplifier. Full-
power operation was suspended for two days for
modifying the blower motor mount, repaiting the
oil system on the component coolant pump, and
restoring the safety amplifiers to service. Then
began the seven weeks at full power.

During the weeks at full power, there were no
other eguipment problems that threatened con-
tinuity of operation, and interest focused pri-
marily on the studies of the fuel salt. Four ,
additions of beryllium, ranging from 8 to 12 g each,
were made in the first three weeks. After the
fourth addition, there was an anomalous, temporary
rise in chromium concentration in the salt samples,
and over the next week ten fuel salt samples were
taken to follow the behavior as the chromium con-
centration returned to normal. Nex! came a series
of uranium additions: 18 capsules in seven days.
This brought the 23°U inventory up enough for
=ix months of power operation without further
additions. Operating for a period of considerable.
burnup without refueling will make it possible to
determine the capture-to-fission ratio for 225U in
the MSRE neutron spectrum from the changes in
uranium isotopic ratios. -

Run 12 was brought to an end because of dif-
ficulties with the fuel sampler-enricher. During
an attempt to take a routine 10-g fuel sample on
August 5, the cable latch became hung as the
capsule was being lowered. There was no external
sign of trouble; however, as the cable unreeled,
it coiled up in the drive unit housing. Then, as
it was being rewound, it tangled in the gears. The
exact situation could not be diagnosed, and when
the isolation valves between the sampler and the
pump bowl were closed, the drive cable was
severed just above the latch (see discussion on
p. 32). After two days of low-power operation

15

to obtain reactivity data in the absence of xenon,
the fuel was drained, and the loop was flushed
and cooled down to permit replacement of the
sampler mechanism and retrieval of the latch.

A temporary containment enclosuwe was erected
around the sampler, and a filtered exhaust system
was connected to the sampler housing to minimize
contamination problems. After the sampler mecha-
nism was removed in a shielded carrier to the
equipment storage cell, a steel work shield was
set up on top of the sampler to permit insertion
of retrieval tools down the sampler tube. By this
time several long, flexible retrieval {ools had been
designed and tested in a mockup (see p- 38). A
noose-type tool was used first, but broke because
the Jatch was stuck at the latch stop. After an
effort to dislodge the tatch, it was enpaged with
another noose tool. The latch was still stuck,
and 1t was necessary to heat up the pump bowl to
loosen it. (Apparently salt mist on the latch stop
had frozen the latch in place.) The latch was
lifted until it became hung in the tube, and the
noose again broke, The latch was picked up again
with a corkscrew-type tool, but it pulled loose
at the first bend in the sampler tube. Then another
tool was designed to slip down over the latch and
clutch it with a knob on the end of a cable. Fig-
ure 1.2 shows workers atop the shield, inside the
enclosure, manipulating this tool onto the latch
20 ft below. The latch was retrieved successfully
this time, but as shown in Fig. 1.3, the capsule
was missing.

After the latch was removed, a go gage was to
be inserted to determine whether the tube was
clear. It had already been concluded that leaving
the capsule in the sample cage in the pump bowl
would cause no harm, but it was simple to medify
the go gage to house a retractable magnet that
could pick up the capsule. When the tool was in-
serted, the tube was clear, but the capsule was
not found in the cage. Figure 1.4 shows details
of the sampler installation in the pump bowl.
There is enough cleatance for the capsule to
slip out under the ring at the boftom of the cage,
but the capsule is then confined by the baffle.
The capsule, with a copper body and nickel-
plated steel cap, should not deteriorate in the
salt, nor are the salt currents strong enough to
cause movement and erosion., Therefore, no more
efforts were made to remove the capsule,

Startup for run 13 then began while a new sampler
mechanism was being installed and checked out.
16

PHOTO 88992

 

Fig. 1.2. Team Fishing for MSRE Sampler Latch.
 

Fig. 1.3. Sampler Latch, Key, and Cable After Retrieval.

Analysis and details of operations and main-

17

tenance are given in the sections which follow.

Table 1.1,

Summary of Some MSRE Operating Statistics

IPHOTO 89000

Table 1.1 summarizes some operating statistics.

 

March—August 1967

Total Through Aug. 31, 1967

 

Critical time, hr
Integrated power, Mwhr
Equivalent full power hours

Salt circulation
Fuel loop, hr
Coolant loop, hr

2025 (66%)
18,795

2597 (59%)

3024 (68%
3113 (71%

7018
40,307

357

wn

10,361
12,059

 
18

ORNL-DWG 67-10766

 

   
 
 
 
 

 

 

 

 

— T ]
g LATCH ASSEMBLY CABLE SHEARED OFF
1 |1 APPROXIMATELY AT THIS LOCATION
A
11
1
W
1
%
L
!
g
12 in i
LATCH ASSEMBLY NORMAL PQOSITION
LATCH
STOP. &
TOP OF FUEL PUMP
7

 

 

—==-BAFFLE

   

 

 

 

NS

POSSIBLE MOVEMENT
OF CAPSULE\fi

SAMPLE CAPSULE
/ CABLE P
L

SAMPLE CAPSULE

    

 

N
5
N

NORMAL OPERATING _ /& Ameorienpdmron
SALT LEVEL//f“'

1A
CAGE FOR
L~

N
F——

 

SAMPLE CAPSULE

 

 

 

 

    
 

 

 

 

CAPSULF CAGE"" o 74 aLor
BAFFLE 7z’
& } SECTION A-A
/ ||
A ‘_\\h 2
1HE | ~SAMPLE CAPSULE

 

 

LOST SAMPLE CAPSULE

 

 

 

 

SALT INTAKE SLOT

 

 

E3/|6 in. MAX
/ 1 0 1 2 3 4

 

INCHES

Location of Latch and Sample Capsule in Fuel Pump Bowl.

Fig. 1.4. Location of Sampler Latch and Capsules in Fuel Pump Bowl.
1.2 REACTIVITY BALANCE
J. R. Engel

The extended periods of full-power reactor
operation in runs 11 and 12 have provided the
most severe tests to date of the on-line reactivity
balance calculation. Runs 11 and 12 increased
the integrated power by 16,200 and 7650 Mwhr,
respectively, to a total of 40,307 Mwhr. In addi-
tion to the usual calculations of power- and time-
dependent factors, calculations were required in
each of these runs to compensate for 233U addi-
tions that were made with the reactor at full
power, The overall performance of the calcula-
tion was highly satisfactory, and no anomalous
reactor behavior was indicated at any time. How-
ever, some additional calculation modifications
were required to eliminate errors that developed.

19

Balances at Power

Figures 1.5 and 1.6 summarize the results of
the on-line calculations during this report period.
These results are reproduced exactly as they
were generated, with no corrections for computer-
induced errors, For legibility, only about 2% of
the data points are shown, but each plotted point
ig the result of an individual calculation. Thus
the scatfer in the plotted points is an indication
of the precision of the calculation. The points
at which changes were made to correct errors are
indicated by notes.

Except for one negative excursion caused by
circulating voids immediately after a power shut-

down in run 11 (Fig. 1.5), all the calculated values

of residual reactivity were between - 0,03 and
+.0.10% 8k/k. An apparent gradual decline in

ORNL-DWS 67-4044R2

 

 

 
 

Pedanal

 

   

I
|
REACTIVITY

Y 8k/k

 

 

 

T T
.L..u S ot AReL e, N T e g.._u
? . |

 

 

 

 

 

el S esp ™, nlatne,

 

 

 

 

 

 

 

 

 

 

Mw

 

 

 

 

 

0

‘ 1

0.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25
* COMPUTER QUT OF SERVICE

 

G5 - L P i b el |

Yo 3K/ K

- |
-005 --1- Lo
] REACTIVITY
.......... L]

 

 

 

r CORARFCTED KTEMP, KSAM MORFED KXE

. o L . o TR
e +‘f B et e e e et e et ‘ "

 

 

 

 

 

 

Mw

 

 

 

 

 

 

 

 

 

15 i7
MARCH, 1967

 

 

 

 

 

 

 

 

% Sk/k

 

 

 

 

 

 

 

 

 

13 15 17

APRIL 1967

Fig. 1.5. Residual Reactivity During

 

 

 

21 23

 

 

 

 

 

), i s |

SeaTnas weluns -~ L }

P AT e aa ! e ! SSnatre, Y g et St -
: ? . .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7

29 1

24 23

MAY, 1367

MSRE Run 11,
015
ciC

0.05 psommai™ag e, o

 

% Bk /k

oft——m—n e
REACTOR SUBCRITICAL -

- . ; o o A8
-
 E

20

~ ORNL-DWG &7-9191

 

 

REACTIVITY

. . . *
o8¢ $ 2 e '.fla  atat et g n e .
P

 

 

 

 

 

 

 

 

 

-0.05
-0.0 - - e el
8 {#fi . T T T
s ‘
i POWER
Z 4 ‘
2
0| L . b . o i Lo
19 21 23 25 27 29 1 3 5 7 2 15
JUNE, 1967 JULY, 1987
o5 —— - e e s _
REACTIVITY
0.10 : s et e e : SHIFT IN COMPUTER.
b tanee o et - READINGS
% 005 ‘ o R PP .
z O - CHANGE CONTROL? RGD ,,;fwv . -:T-.'.'f"_'j.’f:.."-_-_" "'.‘?“’.’Mfi%{%‘fi{’: ,‘
&2 CONFIGURATION
-0.05 . .
-o40lL e - : TR
8  — T T T o L i T
POWER H‘
= - : .
= 4 .
2 . . . . ‘.
ol el S _
16 18 20 22 24 26 28 30 1 3 5 7 2
JULY, 1967 AUG 1967

Fig. 1.6. Residual Recctivity During MSRE Run 12,

residual reactivity occurred during the first
several weeks of run 11. Detailed analysis of the
individual terms revealed two sources of error.
One was a gradual downward drift in the tempera-
ture indicated by two of the four thermocouples
used to calculate the average reactor outlet tem-
perature. These two thermocouples were elimi-
nated and replaced by one other that had not
drifted. The second error was caused by loss of
significance in the calculation of the '#?Sm con-
centration. In the program, only the change in
samarium concentration is computed, and that
change is added to the last value to obtain the
current value. As the '*?Sm concentration ap-
proached 85% of its equilibrium valiue, the incre-
mental concentration change computed for the 5-
min time step between routine reactivity balances
was outside the five-decimal-digit precision of the
computer. As a result, these increments were
lost when the concentration was updated. To
avoid using double-precision arithmetic, the pro-
gram was modified to only update the *#%Sm and
the !°!Sm concentrations every 4 hr while the
reactor is at steady power. Summary calcula-
tions made off line were used to verify the ade-
quacy of this change.

When these corrections were introduced on
March 17, the apparent downward drift in re-
activity disappeared. At the same time, minor
changes were made in some of the !35Xe stripping
parameters to make the calculated steady-state
xernon poisoning agree more closely with the ob-
served value.

Other small reactivity variations were observed
in run 11, for example, from March 29 to April 9.
These changes are directly related to changes in
the helium overpressure on the fuel loop; a 1-psi
pressure increase leads to a reversible reactivity
decrease of slightly less than 0.01% 8k/k. The
mechanism through which pressure and reactivity
are coupled has not yet been established. The
direct reactivity effect of the change in circulat-
ing voids caused by a change in absolute pressure
is at least a factor of 10 smaller than the observed
effect of pressure on reactivity. The time con-
stant of the pressure-reactivity effect is relatively
long, suggesting a possible connection through the
Xenon poisoning.

Fuel additions were made for the first time in
run 11 with the reactor at full power. Nine cap-
sules containing a total of 761 g of 23°U were
added between April 18 and 21. The reactivity-
balance results during this time show good agree-
ment between the calculated and observed effects
of the additions. The transient effects of the
actual fuel additions were very mild. Figure 1.7
shows an on-line plot of the position of the
regulating control rod made during a typical fuel
Con-
trol rod movement to compensate for the additional
uranium in the core started about 30 sec after
the fuel capsule reached the pump bowl, and the
entire transient was complete about 2 min later.
This indicates rapid melting of the enriching salt
and quick, even dispersion in the circulating fuel.
The weights of the emptied fuel capsules indi-
cated that essentially all their contained *33U
was transferred to the fuel loop.

The reactivity-balance results in run 12 (Fig.
1.6) were essentially the same as those in the

addition with the reactor on servo control.

preceding run. Minor variations, associated with
pressure and power changes, were again observed.
Another series of fuel additions at full power
was made in this run between July 19 and 26,
This series consisted of 18 capsules containing
1527 g of 232U, The purpose of this large addi-
tion was to provide sufficient excess uranium so
that a large amount of integrated power could be
produced without intermediate fuel additions. We
plan to perform a detailed evaluation of the uranium
isotopic-change effects associated with power
operation, and substantial burnup is required to
make the analyses of isotopic composition useful.
A secondary result of this large fuel addition
(0.5% 6k/k) was a drastic change in the control
rod configuration. At the end of the additions
the separation between the tips of the shim rods

ORNL-DWG 710132

 

 

ol |

37 - —
’ \“‘MJJ\,\““;‘"‘*/‘ l\fi.f-f‘.f\...‘_r-‘-t'-\/\.\,

{inches withdrown)

 

 

REGULATING RCD POSITION

 

 

 

 

 

G 1 2 3 4 5 s 7
TIME {min)
O=4150 hr APRIL 20,1967

Fig. 1.7. Regulating Control Rod Position During
Fuel Addition.

and that of the regulating rod was 15.5 in., whereas
the normal separation has been 4 to 8 in. The
variation in apparent residual reactivity as a
function of control rod configuration was reexamined,
and we observed a decrease of 0.02% 6k/k when the
more usual configuration was established. This

was consistent with an earlier evaluation (May

1966) of the accuracy of the analytic expression
used in the computer to calculate control rod
poisoning as a function of rod configuration.

On August 3 a computer failure occurred which
required recalibration of the analog-signal ampli-
fiers after service was restored. As a result of
this recalibration, there were small shifts in the
values of several of the variables used in the re-
activity balance. Errors in reactor-outlet tem-
perature and regulating-rod pesition caused a down-
ward shift of 0.03% dk/k in the residual reactivity.

Balances at Zero Power

Figure 1.8 shows the long-term variation in
residual reactivity since the start of power opera-
tion (December 1965). The values shown are
average results at zero power with no xenon
present, Corrections have also been applied for
computer-induced errors such as those at the end
of run 12. The results are plotted to show their
relationship to the reactor operating limits at
+0.5% Ok/k. The discovery of a 0.5-in, shift in
the absolute position of rod 1 at the end of run 12
(see p. 31) adds some uncertainty to the last point
in this figure. This shift represents a reactivity
effect of 4-0.02% 8k/k, which would have been de-
tected if it had occurred during a run. However,
the dilution corrections which must be applied

ORNI-DWE 67-3803RA

o

 

 

 

 

ACTIVITY % 8k/k?
o
W :

. .
-
os |
|

v
.

o

 

e

RESIDUAL RE

 

 

|

 

 

400 T | B

16 20 24 28 32 36 40 (x100)
INTEGRATED POWER (Mwnr)

Fig. 1.8. Long-Term Drift in Residual Reactivity of
the MSRE at Zero Power.
between runs contain enough uncertainty that an
error of this magnitude could be lost. Thus the
shift in rod position cannot be assigned to either
the beginning or end of run 12. Even with this
uncertainty in residual reactivity, the zero-power
results fall within a very narrow band, which
demonstrates the continuing good performance of
both the reactor system and the reactivity-balance
calculation.

1.3 THERMAL EFFECTS OF OPERATION

C. H. Gabbard

Radiation Heuting

Reactor Yessel. — The temperature differences
between certain thermocouples on the reactor
vessel and the reactor inlet temperature are moni-
tored by the computer to determine whether there
is any evidence of a sedimentation buildup in the
lower head or on the core support {lange. In the
previous semiannual report,’ it was stated that
these temperature differences had increased. Fulil-
power data were reviewed from runs 6 thiough 12,
and it now appears that the increase reported is
within the data scatter. The average temperature
differences for run 6 were 2.11 and 1.54°F/Mw
for the core support flange and the lower head,
respectively, and were 2.205 and 1.55°F/Mw for
run 12,

Fuel Pump Tonk. — An unexplained downward
shift in the temperature of the upper pump-tank
surface was mentioned in the previous semiannual
report.? Past data for the pump-tank temperature
and for the heat removal by the oil system were
reviewed to determine if a better thermal coupling
could have developed between the pump tank and
the shield-plug oil cooler. No evidence of in-
creased heat removal by the oil system was found.

The temperature distribution remained essentially
the same throughout run 11, with pump operation
continuing without cooling air. When the reactor
was taken to power in run 12, the full-power tem-
perature distribution had shifted downward another
15 to 30°F, and the pump tank continued through

 

IMSR Program Semiann. Progr. Rept. Feb. 28, 1967,
ORNL-4119, p. 19.

Ibid., p. 18.

22

the run at the lower temperatures. The tempera-
tures at zero power were consistent with the run
11 zero-power data. This would seem to indicate
that less fission product activity was being re-
leased in the pump tank. The lower temperatures
are not detrimental to the operation or to the life
of the pump tank.

Thermel Cvele History

The accumulated thermal cycle history of the
various components sensitive to thermal cycle
damage is shown in Table 1.2, Approximately 63%
of the design thermal cycle life of the fuel system
freeze flanges has been used to date; 54% had
been used at the time of the previous semiannual
report.3

Temperature Mecsurement

Salt Systems. - Approximately 330 thermocouples
are used to measure the temperature at various
locations on the fuel and coolant circulating salt
systems. Only two thermocouple wells are pro-
vided, one each in the coolant radiator inlet and
outlet pipes. The remaining thermocouples are
attached to the pipe or vessel walls., The thermo-
couples on the radiator tubes are insulated to pro-
tect them from the effects of the high-velocity air
that flows over them during power operation; the
others are not insulated and thus are subject to
error because of exposure to heater shine and to
thermal convection flow of the cell atmosphere
within the heater insulation. In March 1965, with
the fuel and coolant systems circulating salt at
isothermal conditions, a complete set of readings
was taken from all the thermocouples that should
A
similar set of data was taken in June 1967 at the
start of run 12. The results of the two sets of
measurements are shown in Table 1.3. Compari-
son of the standard deviations for the radiator

read the temperature of the circulating salt.

thermocouples with those for the other thermo-
couples shows the effect of insulation on reducing
the scatter. Comparison of the sets of data taken
over two years apart shows very little change,
certainly no greater scatter, Figure 1.9 shows
that the statistical distribution of the deviations

 

SIbid., p. 20.
23

Table 1.2, MSRE Cumulative Thermal Cycle History Through August 1967

 

 

 

 

 

 

 

 

 

-20 0 20 4G
DEVIATION FROM AVERAGE (°F)

-60 —-40

Fig. 1.9. Comparison of MSRE Thermocouple Daita

from March 1965 and June 1947.

 

 

 

 

 

 

 

Thaw
Component Heat /Cool - Fill/Drain Power On/Off Thaw and
Transfer
Fuel system 8 37 57
Coolant system 6 11 53
Fuel pump 9 32 57 428
Coelant pump 7 12 53 113
Freeze flanges 100, 101, 102 3 33 57
Freeze flanges 200, 201 7 11 53
Penetrations 200, 201 7 11 53
Freeze valve
103 6 29 35
104 14 9 25
105 16 18 43
106 18 26 38
107 10 11 18
108 9 17 14
109 9 20 18
110 2 2 3
111 5 4
112 2 1 2
204 8 15 26
20 8 L 13 24
100 ‘ --------------------- proe d&Q.OR:L.DWG:—??WB:; Tabie 1.3, Comparison of Readings of Thermocouples
! { ‘ DO;"‘ | of Salt Piping and Yessels Taken with
o ) - | o | the Salt Isothermal
L 0
;20 | o Lol Thermocouple Indicated Temperature ("F)
'E § [ ( ‘ o ! Location March 1965 June 1967
e b S
w Radiator 1102.6 £ 6.7 1208.5 +3.3
'; || ® DATA TAKEN 1 ) ' 777777777 tubes
£ 60 MARCH 1965 4 Sy
= © DATA TAKEN . Other 1102.1 £13.0 1206.7 £12.3
@ JUNE 1967 | | ‘
3 80 e — S R All 1102.3 £10.6 1207.4 £9.8
Z 40— — — \ B e |
é ’ \‘ ! of individual thermocouples from the mean also
Boapl L i & changed little in the two years.
'E ‘ l The scatter in the various thermocouple read-
g 50 J , B o o ir.tgs is reduced to 31.1 acce[{table level by using
i | ’ biases to correct each reading to the overall
ol \ 7777777777777777777777777 average measured while both fuel and coolant
systems are circulating salt at isothermal condi-
5 L,.J—_o‘¥gi‘,f’o§{"§i... 0 tions. These biases are entered into the computer

and are automatically applied to the thermocouple
readings. The biases ate revised at the beginning
of each run and are checked when isothermal con-
ditions exist during the run, Generally the biased
 

PHOTO 87924

Fig. 1.10. East Side of Heat Exchanger Showing Heater Box HX-1 and the Cocked Spacer on the Right.

thermocouple readings have been reliable, but
there have been a relatively few cases when there
have been shifts in thermocouple readings that
have resulted in calculation errors.

Temperature Disturbance in Reactor Cell. — Dur
ing run 11, a shift upward of a reactor cell ambient
thermocouple was noted. This upward shift,
which occurred on only one of ten ambient couples,
took place the day after reaching maximum power.
A rather extensive investigation revealed that
several other thermocouples were affected at the
same time, all in the area between the fuel pump
and the heat exchanger.

Many tests were performed to determine the
cause of this temperature disturbance, but none
gave any conclusive answers. This area was
viewed with closed-circuit television during the
run 11 shutdown in May and June. The only ab-

normality noted which might have caused this in-
crease was a cocked heater spacer between
heaters HX-1 and HX-2 on the heat exchanger.
This spacer is shown in Fig. 1.10, a photograph
of the television screen. It was concluded that
this cocked spacer, which was viewed only after
the fuel system was drained and cooled, was an
indication of an even larger opening which existed
during operation.

Fuel Salt Afterheat

At the conclusion of run 11 power operation, an
experiment was run to determine the amount of
fission product afterheat in the fuel salt. Power
operation of run 11 was terminated by a rod and
load scram from full power, and the temperature
100 -y T

 

  
 

 

1787
e T T T T | | L"H;T
5 - —~ CALDRON CAL(,ULAnom WITH Kr AND Xe STRI ppwe L I
z —4/__'__‘#_—_‘1__"1 PP e e o e f 41
9 5o [ CALDRON CALCULATIONS WITHOUT STHIPF‘INb Tl
= I i T |
& 11 ] mere Measurements N0r2 ]
ui ) /M/:.i IZED 1O CALDRON cAL- | |
|
20

 

FISSION PRODUA

 

 

20

O':iNL DWG 87~147

 
   
 
  

4

" TORAIN TANK - - IR
ETTL

.)O 100 200 500 1000

  
 
   
 

 
 
    

     
 

 

TIME AFTER SHIEITOOWAL (ke

Fig. 1.11.

transient that followed was recorded by the com-
puter. The net heat input fo the system was
evaluated several times from the combined effects
of the temperature slope and thermal capacity of
the system, the power to the electric heaters, and
the 10 kw of nuclear power when the reactor was

critical. The fuel was drained shortly after the
final heat input data were taken in the fuel loop
55.5 hr after the scram. Two additional sets of
heat input data were taken in the fuel drain tank
at times of 168.5 and 745 hr, but there was no ex-
perimental method to cortelate the drain-tank data
with the fuel-loop data because of the difference
in heat losses. The analysis of the experimental
data gave the change in afterheat between the
55.5-hr data and the various other sets of data
taken in the fue!l loop and between the two sets of
data in the fuel drain tank.

The computer program CALDRON was used to
check the experimental results and to provide
reference points at decay times of 55.5 and 765
hr. The results of the CALDRON calculations
and the afterheat measurements are shown in Fig.
1.11. Two sets of CALDRON calculations are
shown, one set without krypton or xenon stripping
and the other with krypton and xenon stripping at
a rate equivalent to the removal from the MSRE
fuel salt. The MSRE experimental data were
normalized to the 55.5- and 745-hr CALDRON cal-
culations that included stripping. (The heat
losses required to make the observations agree
with the calculation at these points were assumed
to exist at all other times in the same system.)

We had hoped to obtain useful data within about
10 min after the scram. However, there was ap-

Results of MSRE Afterhect Measurement.

parently an air leak through the radiator enclosure
which healed itself in about 1.5 to 2 hr. Since

the calculation procedure required that the heat
losses {rom the reactor system be nearly con-
stant, the first 2 hr of data could not be used.
Actually the calculations made 2 hr after shutdown
also appear to be somewhat low, as s=en in Fig.
1.11.

The thermal capacity of the fuel and coolant
systems for the afterheat calculation was cali-
brated during the run 12 startup. The temperature
transient was recorded following a step increase
in nuclear power of 148 kw. The thermal capacity
of the system was found to be 16.22 Mw-sec/°F.

1.4 EQUIPMENT PERFORMANCE

Heat Transtfer

C. H. Gabbard

The monitoring of the heat transfer performance
of the salt-to-salt heat exchanger continued, both
by periodic measurement of the heat transfer coef-
ficient and by practically continuous observation
of the ‘‘heat transfer index,”” (The heat transfer
index is defined as the ratio of reactor power to
the tempetature difference between the fuel leav-
ing the core and the coolant leaving the radiator.)?
Six sets of data were taken for evaluation of the
heat transter coefficient during runs 11 and 12.

 

Ibid., p. 21.
 

 

CRNL-DWG 37-11404

 

 

 

;

5" 700 — =, 007

o

- 800 } o | DoB ¥
£ | J g;
2500 ¢ - l‘ i — 005
.- ‘ 5
o 400 J - 004 =
2 ‘ | b
Lu'] 300 -t 003 &
S | 2
E 200 — - o HEAT TRANSFER COEFFICIENT (REVISED CALCULAIION) [ : 0.02 f
9 t ® HEAT TRANSFER COFFFICIENT [ORIGINAL CALCULATION){ : <
& 00— L 1 " = HEAT TRANSFER INDEX : ; J 0ot T
ool Loy Ty

& A M J J A S O N S S .

12€6

1

|
-+ {267 -

Fig. 1.12. QObserved Performance of MSRE Main Heat Exchanger.

Coefficients were computed from these data by a
procedure used since the beginning of power
operation and by a revised procedure whose
principal difference is that it uses only the most
reliable thermocouples. Coefficients computed
both ways are shown in Fig. 1.12 along with the
heat transfer index. The coefficients and the
index indicate that the performance of the heat
exchanger has remained practically unchanged.
(The downward shift in the heat transfer index in
March 1967 is the result of revising the tempera-
ture biases in the computer.)

Main Blowers

C. . Gabbard

The rebuilt main blowers, MI3-1 and MB-3, have
now accumulated 4640 and 4220 hr of operation,
respectively, since they were installed in October
and November 1966. The main bearing on MB-3
was replaced in early March after 1800 hr of
operation, when the vibration amplitude started
increasing. The balls and races of the bearing
were severely scored and pitted. The replace-
ment bearing also gave an indication of trouble
and was scheduled for replacement during the run
11 shutdown. However, the problem turned out to
be the result of a loose vibration pickup.

A complete inspection of the blowers and drive
motors was made after the run 11 shutdown. Both
blowers were again in excellent condition after
3585 and 3162 hr of operation, with no indication
of cracking in the blades or hubs. The slip rings
and brushes on the drive motors had become scored,
and the motors were removed for repair. The re-

pairs included refinishing the slip rings, replacing
the brushes and bearings, and balancing the
rotors. Vibration pickups were added at each
motor bearing, and filters were installed to pro-
tect the slip rings from dirt and grit. When the
blowers were test run, there were excessive vi-
brations on the drive motor of main blower 3. The
motor vibration had been satisfactorily low when
the motor was loosened on its mount, indicating
that the motor was not badly unbalanced. The
rotation speed of the motors was found to be very
near the natural frequency of the motor mount.
The vibration amplitude was reduced to an accept-
able level (below 1 mil) by stiffening the mount,
Insulation Dust in the Reactor Cell. ~- During
observation in the reactor cell between runs 11
and 12, a nonuniform coating of white material was
seen on most of the horizontal surfaces of the re-
actor cell (see Fig. 1.13). Samples of the white
coating were obtained with long-handled tools
and identified as being mostly AlZO3 (insulation).
Attempts to further identify it as one of the two
specific types of insulation known to be in the
cell were unsuccessful. The possible sources are
the insulation covering the fuel pump and over-
flow tank, the reactor vessel, the fuel drain line,
or the fuel line under the heat exchanger, The
drain-tank cell was also viewed, but no covering
of insulation dust was noted.

Radiator Enclosure

M. Richardson

The brake shoes in the brakes of the radiator
door lifting mechanism were found to be worn and
 

PHOTO 87750

Fig. 1.13. Motor of Reactor Cell Cooler No. 1 Showing Dust Accumulation.

were replaced after run 11. This reduced the
coastdown after the doors were partially lowered
to 3 in. It was also necessary to replace part

of the outlet door soft seal gasket material which
had burned and blown loose. Operation of the
doors has been without incident, and the radiator
seals have been adequate for operation.

Off-gas Systems

. B. Engel

Operational difficulties with the off-gas sys-
tems were greatly reduced during this period of
operation. One 7-hr power reduction was required
to replace a filter in the coolant off-gas line.
Otherwise, only minor inconvenience, which had
no effect on power operation, was experienced.

Particle Trap. — The new off-gas filter® (particle
trap) installed in the fuel off-gas line before the
start of run 11 continued to function satisfactorily
with no evidence of increasing pressure drop.

The pressure drop across this unit, with one sec-
tion valved out, remained below 0.1 psi through-
out the operation. The temperatures near the
various filter media depend to some extent on
operating conditions other than power. Increased
pump tank pressure or reduced purge-gas flow
increases the transport time for fission products
from the pump tank to the particle trap. This per-
mits more radioactive decay en route and results
in lower temperatures at the particle trap. Never-
theless, under similar conditions the steady-state
temperature near the coarse filtering material

 

*Ibid., p. 42.
(Yorkmesh) was ~ 275°F when the particle trap
was first used and ~ 380°F near the end of run 12,
The temperatures decrease rapidly when the re-
actor power is reduced, however, and the zero-
power steady-state temperatures are essentially
unchanged. These effects indicate the accumula-
tion of some material, presumably organic, on the
filtering media that enhances the retention of
short-lived fission products.

Main Charcoal Beds. — The performance of the
charcoal beds in holding up noble-gas fission
products has continued to be satisfactory. The
gradual development of restrictions at the inlet
ends of the beds has also continued, but this has
not limited the reactor operation in any way, since
effective measures can be taken to reduce the
restriction when necessary. _

Run 11 was started in January 1967 with the
charcoal bed sections 1A and 1B in service with
an initial pressure drop of 2.5 psi at normal off-
gas flow. The pressure drop increased very slowly,
reaching 7 psi on March 29, two months after the
start of the run. At that time the standby beds,
2A and 2B, were put in service, and the restricted
sections were valved out. The pressure drop
across these sections built up from 2.6 to 9 psi
in only ten days. We then cleared the restrictions
from all four sections by forcing clean helium
through sections 2A and 2B in the normal flow
direction and heating the inlet ends of sections
1A and 1B with previously installed® electric
heaters. These operations did not require a re-
actor shutdown but only a temporary lowering of
the water level in the charcoal bed pit to allow
the heaters to function.

After the restrictions in both sets of beds had
been cleared, sections 1A and 1B were put back
in service. In the ensuing three weeks the pres-
sure drop increased from 2.4 to 3.5 psi. At that
time, we decided to increase the helium purge
flow by 1 liter/min to see if the xenon poisoning
would be affected by lower concentrations in the
fuel pump gas space. To accommodate the higher
gas flow without an increase in fuel pump pres-
sure, sections 1A and 1B were valved out and 2A
and 2B were put in service. The next day section
1A had to be reopened to keep the fuel pump pres-
sure at 5 psig. The normal purge flow was re-
stored after three days, and, just before the power
shutdown at the end of run 11, the partial restric-

 

®Ibid., pp. 30-31.

28

tions in all four beds were again removed by heat-
ing sections 1A and 1B and forward blowing sec-
tions 2A and 2B.

Owing to the success of the heaters in clearing
the restrictions from sections 1A and 1B, we in-
stalled similar heaters at the inlets of sections
2A and 2B during the shutdown between runs 11
and 12. The differential pressure transmitter
that senses charcoal bed pressure drop directly
was also replaced. This instrument had failed
earlier, possibly because of the pressure differ-
ences imposed during blowouts of the charcoal
beds. However, all these pressure differences
were within the specified overrange capability of
the instrument.

Power operation in run 12 was started with sec-
tions 1A and 1B in service. The gradual increase
in pressure drop made it necessary to change to
sections 2A and 2B after about three weeks. The
pressure drop across the second sections reached
an unsatisfactory level after only six days. Then
the restrictions were cleared from all four sections
by heating the inlet ends. The remainder of run 12
was completed with sections 1A and 1B in service.

The development of flow restrictions at the char-
coal beds appears to be related to the accumula-
tion of volatile organic matter on the steel wool
packing at the bed inlets. Physical variations in
this packing probably account for the different
times required to plug various individual sections.
The experience in runs 11 and 12 indicates that
the restrictions can be effectively removed by
electrically heating the inlet ends of the beds.
Presumably, this heating drives the volatile matter
off the steel wool packing in the inlets and moves
it farther downstream where the flow areas are
larger. There is no evidence from the charcoal
temperatures that this material has reduced the
fission product retention capability of the charcoal.
Since the heating operations do not affect reactor
performance, there are no plans at present to make
further modifications at the charcoal beds.

Coolant Off-gas System. — Very slow plugging
of the coolant off-gas system at the filter that
precedes the coolant-loop pressure control valve
has been encountered throughout the reactor opera-
tion. The originally installed filter was replaced
in February 1965, during the preoperational check-
out of the system. Subsequent replacements were
made in March and September 1966 and on March
1, 1967, The replacement on March 1, 1967, re-
quired a reactor power reduction for 7 hr to permit
personnel access to the area where the filter is
located. By the end of run 11 (May 1967) the
filter was plugged again, and periodic venting of
the coolant system through an auxiliary line
(1.-536) was required to keep the loop overpressure
below 10 psig.

During the shutdown between runs 11 and 12,
the filter was replaced again, and a minor piping
modification was made to permit the coolant off-
gas activity monitoring to monitor gas vented
through line 536. Before this change, any activity
release would have been detected and stopped by
another monitor on the combined fuel and coolant
off-gas, but identification of the source of the
activity would have been more difficult. No
activity has ever been detected in the coolant
off-gas,

Cooling Water Systems

A. L. Krakoviak

The cooling water systems performed satis-
factorily during this report period. The systems
functioned relatively trouble free except for a
few leaks. In July a 15-gpd leak from the treated
water system was detected and was traced to a
faulty pressure-relief valve in the line leading to
one of the reactor cell coolers. Replacement of
the faulty relief valve restored the system to
normal operation.

Space Coolers, — As reported previously,” leaks
have occurted in the reactor cell space coolers
at the brazed joints on the brass tubing headers,
During the scheduled shutdown at the end of run
11, both space coolers were leak-tested. One
cooler (RCC-1) leaked less than 125 cm?®/day
and was not replaced; however, the other (RCC-2),
which leaked at the rate of 9 liters/day, was re-
placed with a cooler whose headers and nipples
were fabricated of copper. Copper weldments
were used on the new cooler instead of the brazed
joints.

Radiation levels around the removed unit were
sufficiently low that it could be disassembled
directly. The radiator was the most radioactive
component, with readings up to 1000 millirems /hr
at contact. (The radiation was very soft and
caused no contamination problem.) This unit was
discarded. However, the fan motor was retained

 

"Ibid., p. 32.

29

for possible future use, and the new fan, motor,
and radiatot were mounted on the original frame
for installation in the cell.

Reactor Cell Annulus. — Sometime prior to or
during run 11, the fill line to the biological shield
plugged, and water additions to the reactor cell
annulus were made through the level measuring
line. Since the plug in the fill line could not be
cieared, the overflow pipe from the cell annulus
was modified to also serve as a fill line.

Steam Dome Feedwater Tanks. — Water is dumped
automatically from a feedwater tank (FWT) to a
steam dome if cooling of a fuel drain tank (FD)
is required after a fuel drain. - During run 12, small
amounts of water from FWT-1 had randomly ap-
peared in the steam dome of FI-1, causing a tem-
perature decrease in the fuel drain tank. To en-
sure that this drain tank remained available for a
possible emergency drain, the water was removed
from the feedwater tank, which is now in normal
service after having a faulty temperature switch
replaced.

Component Cooling System

P. H. Harley

Although some difficulties were encountered,
the component cooling system operated satis-
factorily during this report period. The two main
blowers (CCP-1 and CCP-2) operated 1536 and
2375 hr respectively; CCP-2 has operated for a
total of 3340 hr without a failure.

The discharge check valve on CCP-2 was re-
placed and the belt drive was tightened as part
of the preventive maintenance program. There was
no indication of any significant aging of the sili-
cone rubber in the removed check valve.

Trouble was encountered in the CCP-1 oil cir-
culating system. First, a loose tubing connection
caused the loss of ~ 2 gal of il during run 12;
this irregularity was easily repaired. Then, fol-
lowing the run 12 shutdown, intermittent low-oi}-
pressure alarms again occurred. An investigation
indicated no significant loss of cil, but a slow oil-
pressure response was observed when the blower
was started. The suspected oil pump and pressure-
relief valve on CCP-1 were replaced with spare
units to cormrect the trouble. The removed pressure-
relief valve was found to be relieving before the
normal oil pregsure developed. In spite of these
difficulties in the system, there was sufficient
lubrication, and no noticeable damage was ob-
served,

Although the temporary strainer in the CCP dis-
charge line worked satisfactorily, a more efficient
strainer, which had been on order for a year, was
received and installed in the line. Over an eight-
month period the temporary strainer accuniulated
~ 30 to 50 g of black, dry powdery material that
appeared to be dust from abrasion of the drive
belts. The new strainer, however, has a 100-mesh
screen and a 0.2-psi pressure drop as compared
with a 1/“,)-in. pore size and a 0.5-psi pressure
drop in the old strainer.

The improved performance of the blower belt
drives, which have not needed replacing during the
past nine months, is attributed to less frequent
starting and stopping of the blowers as well as to
the reinforced motor support. During early opera-
tion, the blowers were alternated twice a month;
now one blower is operated continuously during a
run.,

The stainless steel strainer which was removed
had been in contact with condensate containing
dilute HNO, while in service (see ‘‘Containment,”’
p. 33). After being decontaminated, the strainer
was examined and was found to be in very good
condition. The surface was slightly etched, but
no more than would be caused by the decontamina-
tion process.

Blower CCP-3, which coaols out-of-containment
freeze valves, failed on April 19 after more than
5000 hr of operation, A bearing galled and damaged
the drive shaft. Operation continued without inter-
ruption by using air from the service air compres-
sor. Blower CCP-3 has heen repaired and can now
be used when required.

Salt Pump Oil Systems
A. 1. Krakoviak

The lubricating oil systems for both salt pumps
have been in continuous service except during the
planned oil change during the shutdown after run
11. At this time the oil was sampled, drained, and
replaced with new oil. The oil, which had been
in service since August 1966, showed no signifi-
cant change in its physical or chemical properties.

During the steady full-power operation in run 11,
very good balances were obtained on the oil sys-
tem inventory changes, indicating little or no loss

30

by leakage into the pump bowl. The measured
amounts removed for analysis and accumulated in
the catch tanks actually slightly exceeded the ob-
served decreases in supply reservoir contents.

In March and April the difference was 65 cm?® in
the fuel pump system and 210 cm® in the coolant
pump 8ystein,

The oil leakage through the lower seal of the
fuel-pump shaft had previously accumulated at the
rate of 5 cm®/day; it has now decreased to ~ 1
cm?®/day. The leakage past the lower seal of the
coolant salt pump averaged 17 cm?®/day during run
11 and 30 cm?®/day during run 12; the present ac-
cumulation rate is 15,

Automatic siphons were originally installed on
the oil collection tanks from both pumps to meas-
ure and dispose of seal leakage without manual
draining to keep the level in the sensitive (reduced
cross-sectional area) range of the level-measuring
leg of the collection tanks. These siphons have
failed to function properly at the low oil leakage
rates that actually occurred. The oil simply flows
over the high point of the siphon tube, much like a
liquid flowing over a weir, without bridging the
tube to form a siphon. The overflow points for the
coolant pump and fuel pump oil collection tanks
were reached in March and April respectively.

For the remainder of run 11 the only indicators of
leakage rate were the supply reservoirs, which
are much less sensitive than the leakage collec-
tion tanks. After run 11 the collection tanks were
drained, and the average leakages for the latter
part of the run were determined by measuring the
total accumulated oil leakage. The collection
tanks were drained again after run 12, and periodic
drainings are planned to keep the oil level below
the siphon (overflow) level. Because there is a
high radiation field at the collection tanks, during
power operation the draining operations must be
performed only when the nuclear power is low. If
high leakage rates (500 to 1000 cm®/day) develop,
the auntomatic siphons are expected to function as
designed,

Although the oil from the coolant pump seal
when sampled had a dark appearance, spectro-
graphic and infrared spectrophotometric analyses
showed no significant difference between the seal
leakage oil and a sample of unused oil. The dark
appearance and the somewhat increased leakage
rate past the seal could be an indication of ab-
normal wear at the Graphitar—stainless steel
rotary seal of the salt pump. The somewhat lower
accumulation rate at present indicates that the
seal may have reseated itself.

Electrical System

T, L. Hudson

Power to the MSRE electrical system is supplied
from the ORNIL. substation by either of two 13.8~
kv TVA power lines, a preferred line or an alter-
nate. During this report period, while operating
on the preferred feeder, there were two unscheduled
electrical interruptions when the reactor was at
power, During a thunderstorm on June 25, the
reactor operation was interrupted by the loss of
hoth feeders. Two amplifiers were damaged on the
reactor period safely system. Approximately 32
hr later, the reactor was returned to critical opera-
tion after the period safety amplifiers had been
tepaired. On July 12, the other interruption was
caused by the loss of the preferred feeder during
another thunderstorm, Emergency power from the
diesel generators was in gervice within 2 min, and
low-power nuclear operation was resumed in
about 13 min. After repairs had been completed
on the preferred feeder, full-power operation was
resumed in approximately 6 hr.

Difficulty was experienced in run 12 when re-
starting main blower 1. The breaker tripped off
the blower during the starting sequence for several
attempted starts, On a later occasion, the breaker
tripped several times before main blower 1 was
started. This erratic behavior was explained when
a loose gasket was found in one of the time-
delay orifices when the breaker was checked in
August. Tests have been made that indicate the
total time of the starting sequence is too long
when compared with maximum time delay of the
breaker overload element. Therefore, the total
time of the start sequence will be reduced from
25 to 12 sec.

Heaters

T. L. Hudson

The last of six heating elements in heater HX-1
failed on October 28, 1966. Satisfactory heat ex-
changer temperatures were maintained without this
heater, even with the fuel loop empty. However,
continuous circulation of the coolant salt was
maintained until after run 11, so the full effect of
the heater failure could not be determined. Tests

were performed with both the fuel and coolant
loops empty after run 11 to determine the need for
this heater in preheating the system from a cold
condition. When helium circulation was stopped
in the fuel system, the temperature distribution
was satisfactory for a fill without heater HX-1
operating. However, with helinum circulation in
the fuel system, ‘‘cold’” helium was introduced -
possibly from the fuel pump -~ and one temperature
decreased to below 800°F, Since satisfactory
temperatures could be achieved without it, the
failed heater was not replaced.

Radiator heater CR6-48 failed on March 30, 1967,
and was replaced with a spare heater (CR4-3C).
During the shutdown after run 11 the electrical
lead to heater CR6-4B was repaired, and the heater
was placed back in service. Several broken
ceramic bushings were alsc replaced at this time.

Contro! Rods and Drives

M. Richardson

Performance of the control rods and drives has
been within the operating limits this period. No
mechanical failures of the rods or drives have
occurred. The fine-position synchro of rod 2,
which had failed during the previous period, was
replaced prior to the beginning of run 12, The
drive unit for this rod was inspected at the same
time and found to be in excellent condition. The
grease in the drive unit appeared unchanged after
a total radiation dose of about 10? rads and there-
fore was not replaced,

After run 12, routine checks of absolute rod
position using the single-point indicators in the
rod thimbles revealed an apparent upward shift
of 0.5 in. for rod 1. Since there wag an equjvalent
shift in the upper and lower limit switches, it is
believed that this shift was caused by slippage
of the drive chain on one of the sprockets after a
rod scram. The exact time of the shift is not
known because, between the absolute position
measurements, there were control rod scrams be-
fore and after run 12.

Salt Samplers
R. B. Gallaher

Fuel Sampler-Enricher. — The fuel sampler-
enricher was used intensively during this report
period with only a few minor difficulties until the
failure that brought run 12 to an end. Between
March 1 and August 5 there were 111 operations,
as follows:

Salt samples 71
Freszze-valve samples of cover gas 6
Exposure of graphite to salt and cover gas

Beryllium additions 6
Uranium additions 27

Of the salt samples, 14 were 50-g samples and 2
were taken in special three-compartment capsules.
The special samples are described in the section
‘‘Reactor Chemistry.”” These operations brought
the total, since the sampler-enricher was installed
in March 1965, to 114 uranium enrichments and
279 samples and special exposures.

The uranium additions, the first made with the
reactor critical, provided information on mixing
between salt in the sample enclosure and the
main circulating stream. Response of the re-
activity (Fig. 1.7) revealed a time constant close
to that for mixing betweesnt the pump bowl and the
main stream, indicating rapid melting of the en-
riching salt and good circulation through the
sample enclosure.

Near the end of run 11 the manipulator arm and
boot assembly was replaced after a small leak
appeared in one ply of the boot. The arm was
quite contaminated (300 r/hr at 3 in.), but it was
successfully decontaminated and saved for pos-
sible future use (see p. 40). At the start of
run 12, one ply of the boot was ruptured when ex-
cessive differential pressure was inadvertently
applied, and again the assembly was replaced,
This time a slightly different boot was used. The
new boot was thicker and more durable (at the
expense of ease of manipulation), and the outside
was coated with a white plastic spray which ef-
fectively improved viewing in the sampler by de-
creasing light absorption.

One of the beryllium addition capsules was
dropped while it was being removed from the 1-C
area with the manipulator. The capsule fell onto
the gate of the operational valve, where it was
retrieved by a magnet lowered on a cable.

Enriching capsules occasionally jammed in the
transport container until the difficulty was elimi-
nated by increasing the length of the cavity in the
disposable portion to give more room for the long
capsule and attached key.

32

The neoprene seals on the 1-C access port be-
gan to show increased leakage, possibly due to
radiafion damage. Fission products, mostly the
species found in cover-gas samples, produced
radiation levels in area 3A of several hundred
rads per hour. Before run 12, the radiation in this

area was reduced a factor of 10 by using the manip-
ulator to wipe down surfaces with damp sponges.

In run 12, increased leakage from the buffer zone
between the seals on the 1-C access port was

met by increasing the size of the helium supply
flow restrictor so that a satisfactory buffer pressure
could be maintained. Operation of the pneumatic
clamps on the 1-C access door occasionally re-
sulted in gaseous fission products being vented
through the operator vent line. A small charcoal
filter was added to prevent this activity from
reaching the stack.

As described on p. 15, a situation developed on
August 5 that led to a shutdown of the reactor and
replacement of the 1-C assembly. It now appears
that the trouble started when the latch hung at
the maintenance valve. Indications were that hoth
isolation valves weie fully open and that the
capsule key was hanging properly in the latch at
the outset; so the exact cause of the hangup is
not known. But there is convincing evidence that
the drive cable was severed by the operational-
valve gate while the latch was at the maintenance
valve. The length of the latch and remaining drive
cable (Fig. 1.3) equals the distance from the
maintenance valve up to the gate of the operational
valve, and the cable end appeared to have been
cut. It was believed that the operational valve
was practically closed before substantial resist-
ance was felt, but it would have been easy to
mistake the torque and motion involved in shear-
ing the cable and seating the valve for simply
seating the valve, (The handwheel torque re-
quired to shear the cable was subsequently cal-
culated to be less than 100 in.-1b.)

Once before, in December 1965, the latch ap-
parently hung up at one of the isolatioa valves,
causing the drive cable to coil up in the 1-C area
and in the drive unit. Following that occasion,
the travel of the valve gates on opening was in-
creased slightly, and when the 1-C assembly was
removed after the recent trouble, the valve gates
were observed to completely clear the sample tube,
The valves were also obseived to function reli-
ably, so no changes were made,
The retrieval of the sample latch would have
been facilitated had it been magnetic. Therefore,
the replacement latch was made of magnetic type
430 stainless steel instead of type 304 stainless
steel. Another change in the replacement unit
was the addition of a sleeve to bridge the 2-in.
gap between the cable drive reel and the floor
of the drive compartment. This will prevent the
cable from escaping into this compartment if it
meets resistance while being unreeled.

As explained on p. 15, the detached capsule is
positively confined by the batfle so that it cannot
escape. Swirl velocities observed in the bowl of
the prototype salt pump were less than 0.1 fps,
indicating that continual movement of the loose
capsule is very unlikely. Nor should the capsule
corrode. The body of the capsule is copper, and
the cap is mild steel with a thin nickel plating
that probably leaves some ferrous metal exposed.
On exposure to the molten salt, the copper-iron-
nickel assemblage in contact with the chromjum-
containing Hastelloy of the baffle and Tower head
will temporarily constitute an electrochemical
cell, in which Cr? from the Hastelloy surfaces
will be oxidized to Cr? " and reduced once again
to Cr® metal on the capsule surface. This re-
action slows down as the chromium activity in the
capsule surface approaches that of the Hastelloy
surfaces, Eventually the transfer diminishes to
the rate at which the chromium can diffuse into
the capsule metal. The amount of chromium that
can be transferred in this way will have negligible
effect on nearby Hastelloy surfaces., The con-
clusion is, thetefore, that no ill effects will re-
sult from the presence of the capsule in the pump
bowl.

Coolant Sampler. — During this report period,
seven 10-g samples were isolated from the coolant
salt pump, bringing the total with this sampler to
60. There were no operating difficulties and no
maintenance was required,

Fuel Processing Sampler. — The shielding
around the fuel processing sampler was finished,
completing the installation of this sampler. The
nylon guide in the removal area and the manipula-
tor arm were subsequently removed for use in the
fuel sampler. Other spares are now on hand but
will not be installed until near time for use of the
processing facility.

33

Containment

P. H. Harley R. C. Steffy

Secondary Containment. — During run 11, the
containment cell inleakage was measured to be
~ 10 scf/day with the cell at —2 psig. This de-
termination was made by monitoring the cell pres-
sure with the Hook gage (a sophisticated water
manometer) and by measuring all known purges
into and out of the containment cell. The cell
pressure and system purges remained virtually
constant, with small pressure oscillations occur-
ring as the outside temperature fluctuated, These
oscillations were particularly noticeable during
the colder months.

Between May 16 and June 13, an extensive
containment check was made. Thitteen block
valves out of a total of 160 were found to be
leaking and were repaired. Five were instrument
air block valves, three were in the cover-gas
(helium) system, and five were in the treated-
water system, The most common cause of leak-
age was aging elastomer O-rings; a few, however,
had scale, metal filings, or dirt on the seating
surfaces. None of the valves leaked excessively,
and seven of them are backed by a closed system.
Two others ate in use <0.1% of the time and are
normally backed by closed hand valves.

At the beginning of run 12, the secondary con-
tainment vessel leakage was checked with the
containment cell at 20 psig. At this pressure the
cell leaked only ~ 28 scf/day. The cell was then
evacuated to — 2 psig, and reactor filling opera-
tions were started.

As during earlier runs, there was a water leak
in the cell during run 12 (this leak is discussed
later in the section). As a result of the water
leak the first determinations placed the cell
leak rate at ~ 70 scf/day. However, as soon as
the cell air became saturated, the indicated leak
rate dropped to <20 scf/day. This took only
about seven days. For the major part of run 12,
the indicated cell leak rate remained essentially
constant at ~ 14 scf/day.

For both runs 11 and 12, the cell leak rate
values were well below the 85 scf/day permissible
(extrapolated from accident conditions at a 39-
psig cell pressure).
Although the Hook gage has been the primary
means of determining the cell leak rate, an on-
line oxygen analyzer (Beckman model F3) was
installed at the MSRE with the expectation that
it could be used as a megans of calculating the
cell leak rate as well as keeping track of the cell
oxygen content. Cell oxygen content is held > 3%
to prevent nitriding of the Hastelloy N and <5%
to eliminate the possibility of combustion in the
cell in case of an oil leak.

To obtain consistent data from the analyzer,
we have found it necessary to calibrate the in-
strument every 8 hr. In spite of this calibration
frequency the indicated oxygen content often
changes 0.2 or 0.3% between readings taken every
4 hr, Since a 0.1% change in oxygen coiresponds
to ~ 60 scf of air, it is evident that only data
taken on a long-term basis are meaningful.

During both runs 11 and 12, data from the oxygen
analyzer indicated a negative cell leak rate (the
cell was leaking out instead of in). A possible
explanation for this negative cell leak is that
something is combining with the oxygen in the
cell. If it is assumed that the Hook gage is cor-
rect, then ~ 3.7 scf/day of oxygen must be re-
moved from the cell atmosphere to explain the
oxygen results. Oxidation of a small amount of
oil on hot surfaces could easily consuine this
amount of oxygen. Additional investigation will
be required to resolve the discrepancy between
the two methods of cell leak-rate measurement.

Activity Releases. — The total activity release
to the atmosphere during this repoit period con-
sisted of 3.99 mc of iodine and <0.23 mc of
paiticulate matter. The largest single release was
1 mc of iodine, released while removing graphite
samples from the reactor core on May 15. Other
measurable releases included 0.7 mc of iodine
activity while preparing to remove the broken
sample-cable latch and 0.05 mc of iodine released
when the Fuel Storage Tank was vented to the
stack prior to modification of the fuel processing
piping.

Ventilation. — No difficulties were encountered

34

with the ventilation fans during the past six months.

The ventilation discharge filters showed an in-
crease in AP from 1.13 to 1.85 in. H, O across the
roughing filter section. There was no increase
across the absolute filter section. The annual
filtering efficiency test of the absolute filters in
the ventilation discharge was performed on June

5, 1967. Results of the standard DOP (dioctyl
phthalate) test for the three banks of filters were
99,994, 99,998 and 99.979%; the minimum ac-
ceptable efficiency is 99.95%.

Contamination experience in the vent house
during maintenance operations® and the prepara-
tions for the off-gas sampler led to several re-
visions in the ventilation piping in that area. A
new 0-in. line was installed between the main
ventilation line and the instrument and valve box
of the reactor off-gas system in the immediate
vicinity of the particle traps. The existing vent
line from the valve box was increased from 2 in.
to 4 in. in diameter, and a 4-in. branch was in-
stalled to vent the off-gas sampler box. These
lines will improve the ventilation of the affected
areas and reduce the release of contamination
during maintenance when the areas are open to
the atmosphere.

Moisture in Reactor Cell Atmosphere. — In both
runs 11 and 12, it appeared that the atmosphere
in the reactor and drain-tank cells increased in
humidity for the first few days after the contain-
ment was sealed, until finally moisture began to
condentse in cooler points in the component cool-
ing system. Condensate was drained daily from
the inlets of the component cooling blowers and
the shell of the air cooler at the discharge of the
blowers. During runs 11 and 12 the total conden-
sate collected was 100 and 50 gal, respectively,
averaging about 0.9 gpd.

At least part of the inleakage in run 11 was
from a space cooler (RCC-2) in the reactor cell
which was definitely leaking before it was re-
placed at the end of that run. The source of the
leakage during run 12 was not located, not even
as to whether it was in the reactor or drain-tank
cell, since no water appeared in the sump in
either cell in run 11 or 12,

To determine whether the leak might be in the
nuclear instrument penetration, the water for
shielding and cooling this penetration was spiked
with 6 kg of D, O during run 12. Analysis of the
condensate showed no detectable increase in
deuterium content above the normal water concen-
tration. The sensitivity of the deuterium analyses
was such that a leak of 0.01 gpd into the con-
tainment from the instrument penetration would
have been detectable.

 

81pid., p. 38.
The moisture condensed from the cell atmos-
phere was mildly acidic and contained consider-
able tritium, Results of analyses during each run
were as follows:

 

Date  Sample 3 {dis min~ ] ml“l) pH NO,™ (ppm)
2.21-67 LW-11-1 2.8 % 10° 2.7 221
7-27-67 LW-12-1 2.7 x 10° 2.5 3721

The condensate collected in runs 11 and 12
contained a total of about 700 curies of tritium,
which was stored in the MSRE waste tank before
being transferred to the Melton Valley waste
system. The source of the tritium is presumably
neutron reactions with °Li in the thermal insula-
tion around the reactor vessel. A spectrographic
analysis of Careytemp insulation like that used
in the thermal shield shows the lithium content to
be (.1%. Calculations show that with approxi-
mately 1200 Ib of insulation containing 0.1%

4>

35

natural lithium, exposed to a thermal-peutron flux
of about 10'? neuntrons cm™? sec™ !, the amount
of tritium observed is entirely reasonable.

The nitrate ion in the condensate, which makes
it acidic, is presumably caused by ionization of
nitrogen in the cell atmosphere. (The nitrogen
content is kept at 95 to 97% by addition of nitrogen
through the cell sump bubblers.) The acid con-
centration in the condensate is relatively low, and
little corrosion is evident. A general spectro-
graphic analysis of a condensate sample taken in
August gave the following results:

Element Element

 

ppm ppm
Al 0.03 Mg 0.062
B 0.20 Mn 0.04
Ca 0.80 Ni 1.60
Cr 0.02 Pb 1.00
Cu 0.44 Zn 0.20
Fe 2.20
2. Component Development

Dunlap Scott

2.1 DFF-GAS SAMPLER

R. B. Gallaher A. N. Smith

The system developed for sampling and analyz-
ing the fuel off-gas was described in the last
progress report.' During this report period, in-
stallation of all peripheral equipment was com-
pleted, but the installation of the actual sampling
mechanism was delayed. (Peripherals include
shielding, external tubing and connection points,
power and instrument wiring, panels, and instru-
ments.)

During final inspection and checkout of the
sampling mechanism, a failure was detected in a
Monel-to-stainless-steel weld. Fuither detailed
radiographic inspection revealed some other
welds, mostly Monel to stainless steel, that were
judged to be marginal or unacceptable for primary
reactor containment. Therefore, the mechanism
was modified to eliminate many of the dissimilar-
metal welds (by substituting stainless steel valves
for Mone!l in the valve manifold) and to reduce
mechanical stresses at the welds. The valve
substitution eliminated brazed joints at the valve
bodies, and other minor modifications were made
to improve the containment integrity of the assembly
and the sample transport bottle. The sampler was
reassembled, with all welds fully certified, but
installation was deferred because of other work
during the shutdown following run 12,

2.2 REMOTE MAINTENANCE

Robert Blumberg

During this report period the trend for more re-
mote maintenance to be handled routinely was ac-

 

1 .
MSE Program Semiann. Progr. Rept. Feb. 28, 1967,
ORNL.-4119, pp. 41—42.

36

celerated, but development efforts remained con-
siderable and varied.

The shutdown in May and June was planned, and
considerable effort was spent in advance on de-
velopment of procedures, preparation of tools and
equipment, and training of maintenance persounnel.
As a result, most of the remote maintenance work
during this shutdown was handled by the normal
maintenance forces, Development personnel also
paiticipated and afterward reevaluated procedures
and tools. The next shutdown was unscheduled.
Although some equipment and general techniques
already developed for other jobs proved to be
adaptable, intensive work was required to develop
tools and procedures for retrieving the sample
latch.

Preparations for Shutdown After Run 11

At the beginning of MSKE operation, when tools
and procedures for all remote maintenance jobs
were still being tested and revised, development
personnel provided direct gunidance and much of
the actual work. But the work planned for the
shutdown after run 11 consisted mainly of jobs for
which tools and procedures had alteady been
proved. Therefore, the contribution of the develop-
ment group shifted more toward training main-
tenance personnel, helping plan the operations,
and putting the maintenance equipment in readiness.

The training program consisted of a series of
lectures, demonstrations, and practice sessions
conducted by members of the maintenance develop-
ment group and attended by some 30 people, in-
cluding 5 craft skills and their supervision. The
lectures covered health physics, safety considera-
tions, general requirements of all remote main-
tenance, remote maintenance sirategy used at the
MSRE, details of some of the equipment, and de-
scriptions of specific problems. The demonstra-
tions involved the viewing equipment, the remote
cranes, the portable maintenance shield, and some
of the long-handled tools.
core sample removal and replacement served both to
train personnel and to shake down the equipment.

Practice sessions on

The planning part of the preparation involved
the writing or updating of step-by-step procedures
for each job. These procedures were used to draw
up lists of required tools, equipment, and material,
and to estimate time and manpower requirements.
Possible problem areas were recognized for prior
testing and mockup practice. Before the shutdown
began, the procedures were used to provide informa-
tion to the foremen and the working crews.

A large effort was required to put &ll the physical
equipment into readiness. This meant fabrication
of new equipment; cleaning, maintepance, and
repair of existing equipment; and procurement of
some special items. Numerous long-handled tools,
containers for hot pulls, materials for contamina-
tion control, handling equipment, and devices for
shielding, lighting, and viewing were prepared.

£valuation of Remote Maintenance After Run 11

The maintenance work during this shutdown was
done on a two-crew, two-shift-per-day basis. Al-
though thete were drawbacks, such as spreading
more thinly the knowledgeable personnel and lost
motion at shift change, the work proceeded safely
and smoothly. Radiation levels were about the
same as at the last shutdown and did not require
any changes in procedure. Some bothersome de-
lays were caused by breakdowns of the cranes and
maintenance shield, but the schedule was not
seriously affected. Comments on individual jobs
follow.

Removing and replacing the sample array in the
core was the most difficult task of the shutdown,
involving as it did an extremely intense source
of radiation and contamination, The standpipe
above the reactor access flange contained the
assembly and maintained an inert atmosphere on
it. Before the operation a new charcoal filter
was installed io the vent line from the standpipe
to prevent iodine releases, The procedure used
previcusly was changed to minimize the time of
exposure for the uncontained sample within the
standpipe and to provide complete containment

37

and shielding during the transfer of the sample
from the standpipe to the shielded carrier. Two
unanticipated situations arose. The bushing which
supports the sample assembly and which is sup-
posed to stay locked in position in the vessel
outlet strainer actually came out on the sample
basket. After visually ascertaining that the
strainer was not damaged, a new bushing was
inserted. When difficulty was encountered later
in obtaining a satisfactory seal while replacing
the reactor access flange, four bolts were re-
placed, and all the threads on the nuts were
cleaned by tapping. This permitted an increase
in gasket loading which eliminated the leak.

Removal, repair, inspection, and replacement of
a control rod drive went smoonthly.

The leaking east space cooler was disconnected
through the portable maintenance shield and then
was removed by using the crane, television, and
remotely operated maintenance-shield slide from
the maintenance control room. A new unit, reas-
sembled on the old frame, was reinstalled the same
way without excessive difficulty.

The metallurgical specimens which hang in the
reactor vessel fumace were replaced without dif-
ficulty. The carrier for the core samples was
used to transport the specimens to the hot cell.
The fresh neutron source was transferred from a
carrier and placed in the source tube (on top of
the original source) without incident.

Visual inspection in the cell was by use of a
periscope, binoculars, and remote television.
Thermocouple junction boxes were tested, and
four thermocouples were plugged into a spare dis-
connect box by using long-handled tools through
the maintenance shield,

As a result of the experience during this shut-
down, procedures were reviewed and revised
where desirable, and some tools and equipment
were revised or overhauled. The television camera
mounts were changed for greater flexibility, and
revisions were made to the track and to the roller
and guide bearings on the maintenance shield.

Repair of Sampler-Enricher and Recovery of Latch

After the failure of the sampler-enricher drive
unit, a lead carrier was built, and detailed pro-
cedures for the removal of the defective unit were
prepared and reviewed. The actual removal of
this component was without incident.
Meanwhile, retrieval tools for the latch were
designed and tested in a mockup of the sampler
tube. All the tool shafts had flexible sections
16 ft long to negotiate the two bends in the 11/2-
in. pipe. One noose, one corkscrew, and two
gripper tools were tested originally. All the tools
proved capable of grasping a dummy latch, but the
noose tool (Fig. 2. 1) gave the most secure grip
and a positive indication when the latch was
snared. After the latch was found to be stuck in
place (see p, 15), the dislodging tool shown in
Fig. 2.1 was developed and proved capable of
greater lateral force and considerable impact.

ff/g—in CABLE

{/32-in, CABLE

‘/2 ~in. FLEX
CABLE SHEATH

C/

 

CABLE BEING NOOSE READY
PUSHED OUT FOR FISHING
NOCSE TOOL

'/1g —in. CABLE
/ 16

72 in. SCHED 40 PIPE

 

LATCH RETRIEVAL
TOOL

38

3 in. DIAM x 3% in. |

Later, when the latch became hung in the tube as
it was being lifted with the noose and again while
it was being lifted with the corkscrew, the con-
clusion was that the latch stem was being forced
against the tube wall, causing it to hang. A tool
was then devised that slipped down over the
latch to hold it securely and keep it from hanging
on the way up the tube. This is the third tool in
tig. 2.1 and the one that eventually brought up
the latch. The tool that was developed to check
the sampler tube for obstructions above the latch
stop and to retrieve the capsule is shown last in
fig. 2.1

ORNL-DWG &67-11788

{/4-in. CABLE
.::/

    

—{Y/a-in. DIAM

 

DISLODGING TCOOL

STEEL TAPE— =116 —in. CABLE

s — 1 in. 1D
COPPER

7
L e

 

MAGNET

—~—{. 375 in. DIAM
BRASS

 

S LSS
L /7?‘7;\-.\\\ <

 

 

GO - GAGE AND
CAPSULE RETRIEVAL
TOOL

Fig. 2.1. Tools Developed for Retrieval of Latch and Capsule from Sampler-Enricher.
39

PHOTO 88999

 

Fig. 2.2. General View of Latch Recovery Operation with Tool Retrieval Shield in Place.
Some delays were encountered in the retrieval
operation because of the necessity of building
shielded carriers into which to pull the contami-
nated tools. As a result a very convenient ar-
rangement was developed. A hollow lead cylinder,
24 ft long with 1- to 11/2-in. walls, strapped to a
steel I-beam formed the body of the carrier. A
24-ft length of 11/2-ir1. pipe, with a gate valve at
the lower end, fitted into the shield and could
easily be dropped out at the burial ground with
the contaminated tool safely contained inside.
Figure 2.2 shows the containment enclosure
around the restricted work area, with the shielded
carrier suspended from the crane bridge. The
workers on the bridge are in position to pull a
tool out of the sampler tube with a cable dropped
through a seal on the upper end of the pipe liner.
This system was used for five hot pulls, with no
spread of contamination or excessive radiation.

When the replacement sampler drive unit was
installed, some difficulties were encountered in
fitting up pipe and tubing connections, but con-
tainment joints were made up leak-tight.

The procedures and tools developed for this
shutdown proved effective. The tool retrieval
shield, in particular, was a useful development.

Measures for control of contamination and radia-
tion prevented spread of activity outside the
restricted work zone, and the highest quarterly
dose for any worker was only 300 millirems. In
view of the intense sources involved, this record
attests to the effectiveness of the controls,

2.3 DECONTAMINATION STUDIES

T. H. Mauney

The maintenance scheme for molten-salt breeder
reactors proposes that components which can be
easily decontaminated will be repaired and reused
as spare parts. As an aid in evaluating this
proposal, a study was started to determine the
effectiveness of decontamination procedures in
reducing the activity of contaminated parts from
the MSRE.

The first item used in the study was the manip-
ulator hand which had been used in the sampler-
enricher system during months of power operation,
until it was replaced because of a leak in the
boot. When removed from the sampler-enricher,
the unit was contaminated by mixed fission prod-
ucts to a level of approximately 300 r/hr at 3 in,

40

It was necessary to first remove the two-ply

plastic boot which had provided fission product
containment for the manipulator while in use.

Since the ORNL facility used for routine decontami-
nation can handle only up to 50 r/hr of gamma
radiation, it was necessary to use a facility at

the High-Radiation-Level Analytical Laboratory.
This facility was equipped with remote handling
equipment and high-pressure sprays which were
used in the procedure.

The decontamination began with spraying the
manipulator hand with a 500-psi jet of detergent.
The unit was then soaked in several solutions.
After the radiation level was reduced to less than
3 t/hr, the manipulator was removed to a laboratory ’
hood and hand scrubbed with a wire brush and
another detergent. The radiation level was finally y
reduced from the initial reading of ~ 300 r/hr to a
final level of 300 mr/hr, which was low enough to
permit controlled direct contact for repair. The
detailed results of the various treatments in re-
ducing the radiation level are shown in Table 2.1,

One of the complicating factors affecting the
decontamination of the manipulator hand was the
presence of many crevices in the linkages and
pins which could not be reached by the jet spray.

Extending the cleaning time per step might have
resulted in greater decontamination factors. Fur-
ther studies will be made of this effect as other
components become available,

The decontamination of the manipulator hand
served a dual purpose. It not only helped de-
termine the effectiveness of decontamination
procedures in reducing the activity of MSRE con-
taminated components, but also made available at
a decontamination cost of $300 a spare part that
would have cost about $2000 to duplicate.

2.4 DEVELOPMENT OF A SCANNING DEVICE
FOR MEASURING THE RADIATION
LEVEL OF REMOTE SOURCES

Robert Blumberg T. H. Mauney
Duniap Scott

A method for locating and evaluating concentra-
tions of radioactive materials in areas having
high background radiation would be useful in fol-
lowing the deposition of fission products in com-
ponents of circulating fuel reactors and in chemi-
cal process plants. Location of unusual deposits
would aid in understanding the operation of the
41

Table 2.1, Readiation Levels Following Steps in Decontamination of an MSRE Sampler-Enricher Manipulator Hand

 

 

Cleaning Radiation Level
Cleaning Treatment Time After Treatment
(min) (r/hr)
None (as received) 300
Detergent (Duz) jet spray 5 100
Rinse (water)
Detergent {(Duz) jet spray ‘ 5 75
Formula 50 10 25
Dilute HNOS 10 7.5
Bolt freeing solvent with acetone ringe 5 5.0
Concentrated HNO, 10 3.0
Scrub with scouring powder (BaB-0Q)7 5 0.5—-1.0
Scrub with scouring powder (BaB-0) 5 0.2-1.0
Formula 50 5 No reduction
Concentrated HN(}3 5 No reduction
Scrub with scouring powder (BaB-0) 5 No reduction
Clean with concentrated HN03 5 0.2—-0.3
Acetone rinse 5 Ne reduction

Smearsb

800—13,000°

 

fRemoved to laboratory hood.
bzirconium and ruthenium.

®In disintegrations per minute.

system and in planning for maintenance of the
components. On the basis of these needs, a study
of possibly useful devices was started.

The concept of the pinhole camera using radia-
tion shielding and gamma-ray-sensitive film was
evaluated along with a system using a portable
collimator and a small gamma-sensitive dosimeter.
It was determined that while the camera method
was feasible, the results to date did not provide
very good resolution of position, intensity, and
size, The problem of calibration of the film for
use in radiation fields of unknown energy appeared
difficult. However, the time required to survey an
area would be short, approximately equal to the
required exposure time for the film. The use of
the collimator method would take longer to com-
plete a survey, but possibly could yield better
resolution. This method aiso offered the potential
of giving information on the gamma-ray energy
spectrum of the source. A series of experiments
with the collimator was conducted to evaluate the
method, with the results described below,

Experiment with a 5-curie '37Cs Gamma Source

A collimator was constructed by casting lead in
a 4-in.-diam steel pipe 48 in. long and providing a
l-in.-diam hole the length of the tube. The col-
limator tube was placed in a suitable hole in the
portable shield which was positioned over a pit
containing a S5-curie 12®7Cs source.

Radiation measurements were made with the
equipment and source arranged as shown in Fig.
2.3. The radiation source could be raised and
lowered, and the collimator tube could be moved
horizontally over the source. Measurements were
made with the source 13 ft and 15 ft 6 in. below
the top of the portable shield. The source was
about 1/4 in. in diameter and 2 in. long and was
enclosed in a 1-in.-diam tube 5 in. deep in the
carrier,

The results of the radiation measurements are
shown in Fig. 2.4. It will be noted that as the
coliimated ion chamber was moved horizontally
over the source, the radiation readings changed
PORTABLE SHIELD —.__

 

42

CORNL-DWG 67— 11789

- GAMMA CHAMBER {'%4 in. 00)

s

  

 

 

 

 

 

N

<

 

 

 

 

 

imr/ hr)

:
1

RADIATION LEVEL

30

25

20

 

4-in. LEAD COLLIMATOR .~ ‘ ~—4-in. COLLIMATOR HOLE
48 in. LONG - |
|
COLLIMATOR
MOVEMEMNT
= PR | e R
i
1"
CENTERLINE -~ :
|
| _~CESIUM

-7 GAMMA SOURCE

e

Fig. 2.3. Collimator-Source Geometry for Point-Source Test.

ORNL-—-DWG €7—11790

 

 

 

 

 

 

““““ ; e = 1
| | | |
i ' | :
! ‘ i ; ; f ?
e DISTANCE 43.0ft VERTICALT[™ 17 77—
(SOURCE OUT OF CARRIER) ;
e DISTANCE 455t VERTICAL |
{SOURCE OUT OF CARRIER) ‘
. DISTANCE 155ft VERTICAL | %
(SOURCE IN CARRIER) 3
5 X !
; Z
LJ
©
Ll
O
o
2
| o]
} v
|
‘i
i
|
" L J -
5 4 3 2 1 o 1 2 3 4 5

 

 

  

 

 

 

COLLIMATED SOURCE POSITION TO LEFT OR RIGHT (in)

Fig. 2.4. Collimated Radiation from 5-curie 137¢. source.
significantly, indicating clearly the collimating
effect. The flat top at the peak was of the same
width as the collimator hole and represents the
resolution of the collimator when used with a point
source,

Gamma Scon of the MSRE Heat Exchanger

A gamma radiation scan was made on May 19,
1967, over the area in the reactor cell containing
the fuel-to-coolant-salt heat exchanger and fuel
salt line 102,

To accomplish the radiation scan, the portable
maintenance shield was located over an opening
in the reactor cell above the heat exchanger, and
the collimator and gamma ion chamber were in-
stalled in a hole in the portable shield. Measure-
ments were made by moving the portable shield
until the ion chamber was directly above the de-

ORNL-DWG &7-11721

PORTABLE
l SHIELD.,

.7” Z

ION CHAMBER

Y

1it-71in.

 

  
 

1 ft-2in.

 

 

 

 

 

 

RO

 

 

O —,

 

/I
48-in. COLLIMATOR TUBE
(4% in. OD) 18 £1-9 in,
13 fi=2 in.
~—— = |NCREASING REACT_OR— -
> VESSEL

HEAT EXCHANGER

(_(N* -

 

 

 

 

 

 

5 ft—3in,
~ | LINE 102
YT T T/ - Y
— U )L
FF 102

Fig. 2.5. Source-Detector Geometry for Gamma Scan
of MSRE Heat Exchanger.

sired points in the cell. Figure 2.5 shows the
vertical distances from the tip of the ion chamber
to items in the cell.

The orientation of the various ion chamber
traverses of the cell along with lines of equal
radiation levels are shown superimposed over the
heat exchanger and line 102 in Fig. 2.6. It will
be noted that the maximum radiation levels were
3.5 t/hr above the heat exchanger center line
and 300 mr/hr above line 102, When the collimator
was removed, the radiation level at the top of the
hole in the portable shield was 100 r/hr.

These results show that the collimated gamma
ion chamber measures radiation levels with very
good resolution of position, even in the presence
of high background radiation levels. Therefore,
this method of radiation measurement should be
useful in locating accumulations of fission products
in reactor components and in planning maintenance
operations.

Gamma Energy Spectrum Scan of the

MSRE Heat Exchanger

On May 20, 1967, a gamma energy spectrometer
was set up over the same area in the reactor cell,
and measurements were made at several points
over the fuel-to-coclant-salt heat exchanger. This
exploratory experiment was made primarily to
evaluate the method.

The measurements were made by using a 3- by
3-in, Nal crystal and photomultiplier tube mounted
in a lead shield which had a collimating hole
1/32 in. in diameter and 7 in. long under the crystal.
The array was mounted on the hole in the portable
shield in the same manner as the collimator for
the gamma scan described earlier.

There were strong, well-defined peaks at about
0.48 and 0.78 Mev corresponding to '93Ru and
93 Zr-9°Nb. The resolution of the Nal crystal was
not good enough to determine whether the second
peak was from °°Nb only or from the °°Zr decay
chain. The energy resolution could be improved
by using a different crystal, since the collimator
did not appear to degrade the peaks. The problem
of assigning absolute disintegration rates to these
isotopes is complicated by the geometrical ar-
rangement of the heat exchanger and heaters, the
effect of gamma energy on absorption by the col-
limator, the source distribution in the heat ex-
changer, and the counter efficiency. It is be-
lieved, however, that some measure of the relative
44

ORNL-DWG 6711742

«  x X <
s < w 0 M~ - -
2 < ! v ol a Q o
1l il Il il I I i I Il
> > 3 > S S S

|

 

 

 

 

 

   

HEAT
EXCHANGER

 

 

- X =99,

 

 

 

"REEZE FLANGE 101~
| |

;,.g,/" FREEZE FLANGE 102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LINES OF EQUAL RADIATION LEVEL

Fig. 2.6. Results of Gowmma Scon of the MSRE Heat Exchanger.

disintegration rates can be obtained by comparing is believed that these differences in the ratios
the ratio of the areas under the counting peaks represent a difference in the deposition of the two
for different isotopes at one position with the isotopes at the two positions. Much additional
ratio found at other positions. For the measure- work would he necessary to establish the method
ments made at point 4 on Fig. 2.6, the ratio of for quantitative measurement under a particular
Ru to Zr-Nb was 0.66, while at point B the ratio set of conditions, but it might be the only non-
was 1.20, destructive method for making such deposition

Although not clearly understood at this time, it measurements.
45

2.5 PUMPS needed at the MSRE. The pump tank is being in-
stalled in the pump test facility, and the facility
P. G. Smith A. G. Grindell is being prepared for tests with molten salt to
investigate the adequacy of the running clearances
Mark- 2 Fuel Pump and the performance of the xenon stripper.
As reported previously,? the Mark-2 (Mk-2) fuel § v Koty Elewants fee MERE Epel

pump was designed to give more salt expansion
space in the pump bowl, thus eliminating the need
for an overflow tank. As a consequence of the
deeper bowl, the Mk-2 rotary element has a longer
unsupported length of shaft than the Mk-1. The
device for removing gaseous fission products from
the salt is also different in the Mk-2 bowl.
Fabrication of the Mk-2 pump tank was com-

pleted, and the rotary element was modified so
that the joint between the bearing housing and 2Ibid., p. 64.
shield plug can be seal-welded should the pump be 3tbid., p. 65.

and Coolant Salt Pumps

The spare rotary element for the coolant salt
pump® was assembled, tested, and fitted for
service at the MSRE. The joint between the
bearing housing and the shield plug was seal-
welded to eliminate the possibility of an oil leak-
age path from the lower seal catch basin to the

 

PHOTO 74199

TENSION OR
COMPRESSION

. TSB : ; NOZZLE |
ey AT TACHMENT
[MOMENT } :

 

Fig. 2.7. Experimental Apparatus for Nozzle Stress Tests.
pump tank, It is now available, along with a
spare fuel pump rotary element, for service in the
MSRE.

Stress Tests of Pump Tank Discharge
Nozzle Attachment

The stress tests of the discharge nozzle on the
Mk-1 prototype pump tank® were completed, and
the tank assembly was transferred to the Metals
and Ceramics Division for examination. A photo-
graph of the experimental apparatus is shown in
Fig. 2.7. Strain data were obtained under a pres-
sure of 50 psi in the tank and a moment of 48,000
in,-1b applied to the nozzle. The measured
strains were converted into stresses by use of
conventional relationships. For comparison, the
stresses were calculated by means of the CERL-
IT code, which is written for a nozzle attached

radially to a spherical shell. The actual configura-

tion is a nonradial nozzle attached to a cylindri-
cal shell near the junction of the shell to a formed
head. The computer code underestimated the ex-
perimentally measured pressure stresses by a
factor of about 4.6 and overestimated the moment
stresses by about 3.5 times. When the stresses
are combined, the code-calculated stresses are
higher than the measured stresses for the test

46

conditions. Based on our present knowledge, the
nozzle attachment appears to be adequate for the
intended service.

MSRE Oil Pumps

Two oil pumps were removed from service at the
MSRE, one because of excessive vibration and
the other because of an electrical short in the
motor winding. The one with excessive vibration
had one of two balancing disks loose on the shaft.
The loose disk was reattached, and the rotor, in-
cluding shaft assembly and impeller, was dy-
namically balanced. This pump has been reas-
sembled, has circulated oil up to 145°F, and is
ready for service at the MSRE. The pump with the
shorted motor winding has been rewound, reas-
sembled, and is circulating oil.

Oil Pump Endurance Test

The oil pump endurance test* was continued,
and the pump has now run for 35,766 hr, circulat-
ing oil at 160°F and 70 gpm.

 

*Ibid., p. 66.
3.

Instruments and Controls

L.. C. Qakes

3.1 MSRE OPERATING EXPERIENCE

C. E. Mathews J. L. Redford
R. W. Tucker

Instrumentation and control systems continued to
perform their intended functions. Component fail-
ures did not compromise safety nor cause excessive
inconvenience in the reactor operation. A moderate
amount of maintenance was required, as described
in the sections that follow.

Control System Components

The 48-v dc relays continued to suffer damage
because of heat developed in the resistors built
into the operating coil circuits. Because of this
problem, redesigned relays had been procured and
some had been installed.! After several months,
some of these also showed signs of detenoration.
Therefore, following run 11, all 139 relays operat-
ing from the 48-v dc system were modified by re-
placing the built-in resistors with exterally
mounted resistors. This was done with the relays
in place, without disconnecting control circuit
wiring, thus avoiding the possibility of wiring
mistakes inherent in relay replacement. No trouble
was experienced after the modification.

Component failures were as follows. The coil in
a solenoid valve for the main blower 3 backflow
damper developed an open circuit. The fine-position
synchro on rod drive 2 developed an open circuit in
the stator winding and was replaced at the shutdown
in May. The fuel overflow tank level transmitter
was replaced after a ground developed in the feed-
back motor. The ground was subsequently found

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1967,
ORNL-4119, p. 71.

47

and corrected. The span and zero settings of the
differential pressure transmitter (PdT-556) across
the main charcoal beds shifted drastically. Al-
though the pressure capability had never been ex-
ceeded, the diaphragm had apparently suffered
damage. Thus, when a new transmitter was in-
stalled in June, hand valves were installed in the
lines connecting to the off-gas system to be used
in protecting against unusual pressures during
system operations. The 1-kw 48-v-dc—to~120-v-ac
inverter, which supplies ac power to one of the
three safety channels, failed during switching of
the 48-v dc supply. It was repaired by replacement
of two power transistors. Only one themmocouple
(TE-FP-10B) failed due to an open circuit. (See p.
22 for discussion of thermocouple petformance.)
Water was admitted to the steam dome on fuel drain
tank 1 on several occasions before the cause was
found to he an intermittent fault in a control module.
A freeze-flange temperature control module failed
because of excessive condenser leakage. At the
next shutdown, all similar condensers were checked,
and 33 were replaced because of leakage.

Nuclear Instruments

Four of the eight neutron chambers had to be
replaced. The fission chamber for wide-range
counting channel 1 was replaced because of a short
circuit from the high-voltage lead to ground. The
fission chamber for wide-range counting channel 2
was replaced after moisture penetrated the protec-
tive cover on the cable. Moisture leakage into the
cable to the BF , chamber required that the cable
and chamber be replaced. The output of the ion
chamber in safety channel 2 decreased drastically
during a nonoperating period, and the chamber was
replaced prior to resumption of operations. The
48

trouble proved to be a failure in a glass seal that
allowed water to enter the magnesia insulation in
the cable, which is an integral part of the chamber.

Safety System

A period safety amplifier failed when lightning
struck the power line to the reactor site, and a
replacement amplifier failed as it was being in-
stalled. The field-effect transistor in this type of
amplifier is susceptible to damage by transient
voltages, and it was found that under some condi-
tions, damaging transients could be produced when
the amplifier is removed from or inserted into the
system. A protective circuit was designed, tested,
and installed on the spare unit; in the interim, a
different and more stable field-effect transistor was
installed. The module replacement procedure has
been modified to reduce the possibility of damage
incurred on installation of the module.

Experience at other sites with the type of flux
safety amplifiers used in the MSRE had shown that
the input transistor could be damaged, causing
erroneous readings, if the input signal became too
large too rapidly. Therefore, a protective network
of a resistor and two diodes was added to the input
circuit of each of these amplifiers.

Two relays in the safety relay matrices failed,
both with open coil circuits. A chattering contact
on the fuel pump motor current relay caused safety
channel 2 to trip several times before the problem
was overcome by paralleling two contacts on the
same relay. A defective switch on the core outlet
temperature also caused several channel trips and
one reactor scram before the trouble was identified
and the switch was replaced

Four rod scrams, all spurious, occurred during
operation in the report period. Two were caused
by general power failures during electrical stoms.
Ancther occurred during a routine test of the safety
system when an operator accidentally failed to resef
a tripped channel before tripping a second channel.
The other scram, also during a routine test, came
when a spurious signal from the switch on the core
outlet temperature tripped a channel while another
channe! was tripped by the test.

3.2 CONTROL SYSTEM DESIGN

P. G. Hemdon

As experience showed the need or desirability of
more information for the operators, improved per-

formance, or increased protection, the instiumenta-
tion and controls systems were modified or added to.
During the repoit period there were 25 design
change requests directly involving instruments or
Six of these required only changes in
Fourteen

coitrols.
process switch operating set points.
requests resulted in changes in instruments or
controls, one was canceled, and the remaining four
were not completed. The more important changes
are described below.

The single £32-v dc power supply formerly serv-
ing the nuclear safety and controls instrumentation
was replaced with three independent power supplies,
one for each safety channel. Each obtains #32-v dc
from 110-v ac, but the ac power sources are dif-
ferent. One is the nomal building ac system.
Another is 110-v ac from a 50-kw static inverter
powered by the 250-v dc system. The third comes
from a 1-kw inverter operating on the 48-v dc sys-
tem. Thus continuity of control circuit operation is
ensured in the event of a single power supply or
power source failure. It also increased the re-
liability of the protection afforded by the safety
system by ruling out the possibility that malfunc-
tion of a single voltage-regulating circuit could
compiomise more than one channel. (Before the
change, on one occasion, a failure in part of the
regulating circuit caused its voltage output to in-
crease from 32 to 50 v, and only a second regulator
in series prevented this increase from being im-
posed on all the safety circuits.)

To prevent the reactor from dropping out of the
“‘Run’’ control mode when a single nuclear safety
channel is de-energized, the ‘““nuclear sag bypass’’
interlocks were changed from three series-connected
contacts to a two-of-three matrix.

Circuits were installed to annunciate loss of
power to the control rod diive circuits. This re-
minds the operator that he cannot immediately re-
tum to power simply by manually withdrawing the
rods after the controls have dropped out of the
“Run’’ mode.

A wiring error in a safety circuit was discovered
and corrected. Interlocks had recently been added
in the ‘‘load scram’’ channels to dwop the load when
the control rods scram. A wiring design error re-
sulted in these interlocks being bypassed by a
safety jumper. Although the circuits were wired
this way for a time before being discovered, the
scram interlocks were always operative during
power operation, since the reactor cannot go into
the ‘‘Operate’’ mode when any safety jumper is
inserted.
The rate-of-fill interlocks, including the pneu-
matic instrument components in the drain-tank
weighing system, were removed from the control

circuits for the fuel drain tank helium supply valve.

Experience had shown that these interlocks, in-
tended to limit the rate at which fuel could be
transferred to the core, were not required, since
physical restrictions in the helium supply posi-
tively prevent filling as fast as the design set
point on the jinterlock.

Four new photoengraved graphic panels incorpo-
rating recent circuit changes were installed on the
control and safety circuit jumper board. The main
control board graphics were also revised to show
changes in the fuel off-gas system.

Test circuits were originally provided to check
the operability of electronic safety instruments on
fuel loop pressure and overflow tank level. These
were revised to make it possible to observe the
value of the test signal current at which the re-
lays operate.

The measuring range of the matrix-type flow ele-
ment on the upper seal gas flow from the coolant

49

pump was increased from 5 to 12 liters/hr. This
was to keep the instrument on scale at pump bowl
pressures higher than originally anticipated. A new
capillary flow element was designed to replace the
clement that measures the reactor cell evacuation
rate, raising the range from 1 to 4 liters/min.

To avoid reducing the reactor system helium
supply pressure below the minimum limit when
large quantities of helium are required for fuel
processing, the fuel processing system helium
supply line was removed from the 40-psig supply
header and reconnected upstream to the 250-psig
main supply header. Pressure-regulating instm-
ments were installed.

To provide more information to the operator, 20
additional pairs of themocouple lead wires were
installed between the vent house and the data
logger. Connections at the patch panel were made
available by removing nnused lead wires from some
radiator thermocouples. A new conduit for these
lead wites was also installed between the vent
house and the existing cable trough in the coolant
drain cell.
4. MSRE Reactor Analysis

 

 

4.7 INTRODUCTION Table 4.1, lsotopic Composition of 233y
Available for MSRE®
I?. N. Haubenreich B. E. Prince o
Fraction

The experimental program planned for the MSRE Uranium Isatope (at. %)
includes operation for over half a year with 233U
as the fissile material. The fuel salt presently in 230 0.022
the reactor will be fluorinated in the MSRE process- 233 91.5
ing facility to remove the uranium (33% 233U). 234 7.6
Then 233U will be added to the stripped carrier 235 0.7
salt as the LiF-UF , eutectic (73-27 mole %). 236 0.05

During this report period, we calculated most of 938 0.14

the important neutronic properties of the reactor,
operating with 233U fuel. These results will be -

. . . . ORNL communication, J. M. Chandler to J. R. Engel,
used in a safety analysis and in planning the start- May 10967,
up experiments. We also made some computations
of neutron spectra to use in evaluating the relevance
of MSRE 233U experiments to molten-salt breeder
design calculations. The techniques of analysis

for the new fuel (Table 4.1) is the actual composi-
tion of the uranium available for use.
were similar to those used previously for the MSRE
Wlth.235U fl:lel. The results are discussed 'm the 4.2 NEUTRON ENERGY SPECTRA IN
sections which follow, and references are given

) . i .. MSRE AND MSBR
which provide more detailed descriptions of the
methods of analysis.

The fluorination process removes uranium and
some fission products, but leaves plutonium and
most fission products in the carrier salt. It was
necessary, therefore, to make some assumptions
regarding these concentrations at the time of the
startup with ?33U. It was assumed that all the
uranium is removed and that all the plutonium and
samarium remain in the salt. We assumed that the
changeover would be made at an integrated power
of 60,000 Mwhr and computed the plutonium and
samarium concentrations from the time-integrated
production and removal rates for these nuclides.
Fission products other than samarium were not
c?nsidere‘d explicitly in the base-line ?a‘lcul-ations, ORNLA4119, . 70,
s1r?ce their net effect c?n th.e core_ react1v1t3.z ¥s 2SR Program Semiann. Progr. Rept. Feb. 28, 1967,
gquite small. The uranium isotopic composition ORNI.-4119, p. 193.

B. E. Priince

In a recent report, we described the results of
computational studies of MSRE neutron spectra for
the present 23°U fuel salt.! We have extended
these studies to the spectra with the 233U fuel and
have compared the results with corresponding
spectra for a typical MSBR core lattice currently
under design study.?

The characteristics of the MSBR core design used
for these calculations were taken from ref. 2. The
graphite-moderated portion of the core was 10 ft
long and 8 ft in diameter and contained fuel and

 

'MS& Program Semiann. Progr. Rept. Feb. 28, 1967,

50
51

ORNL-DWG 67-14793

LETHARGY

8 7 6 5 4 3

™
e~

0

 

 

4
i

 

 

 

 

[ T 1 T [ 1| BT T
: T4
; bl‘il'f\ »

-
| Jd

 

 

 

 

\

 

 

 

 

 

 

MSRE- 27U FUEL

W
3
‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l‘-—-“"‘- —6
|
-
;

 

 

R UNIT LETHARGY (normalized to one fission source neutron

£

 

 

 

 

 

FLUX P
M

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O b C
10!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,,,,,,

 

108

ENERGY (ev)

Fig. 4.1. Calculated Epithermal Neutron Spectra in the 2331)_Fyeled MSRE and in the MSBR.

fertile salt volume fractions of 16.48 and 5.85%
respectively. The fuel salt contained approxi-
mately 2.84 mcle % UF , (~64% of total U is 233y
and ~5.9% is 225U), and the fertile salt contained
27 mole % ThF . By compatson, the model of the
MSRE core used for this analysis was a cylinder
with effective dimensions of 58 in. in diameter by
78 in. in height. The fuel salt volume fraction is
22.5%, and there is no fertile salt.
content of the fuel salt was approximately 0.125
mole % UF (™ 01.4% 2331)), based on calculations
summarized in the following section.

As described in ref. 1, the epithermal and themal
neutron spectra calculations were made with the
GAM and THERMOS programs respectively. In the
choice of an effective ‘“‘thermal cutoff’’ energy
separating these spectra, there is a considerable
degree of latitude, with the single restriction that
the cutoff energy be high enough that important
calculated neutronic properties at operating tem-
perature are not strongly influenced by neglect of

The uranium

upscattering corrections above the cutoff energy.

In previous studies of the MSRE, '3 we used

0.876 ev {(for which the present energy group
structures for the GAM and THERMOS programs
coincide). This value was found to be sufficiently
high to satisfy the preceding criterion.?® However,
because current design calculations for the MSBR
are based on the higher value of 1.86 ev, to make a
valid comparison we calculated the MSRE thermal
spectra for this higher cutoff energy.

The results of GAM-II calculations of the epi-
thermal neutron spectra in the two reactors are
shown in Fig. 4.1. As in ref. 1, use is made of the
lethargy variable, proportional to the logarithm of
the neutron energy, in plotting the results. This
relation is

lethargy = — In (energy/10 Mev) .

For a unit fission sourtce the calculated fluxes per
unit lethargy for each GAM-II neutron group are

 

3P. N. Haubenreich et al., MSRE Design and Opera-
tions Repaort. Part IIl. Nuclear Analysis, ORNL-TM-
730, pp. 38--40.
(normalized)

R UNIT LETHARGY

,_
-

FLUX P

0024

0.020

  

52

16

 

  
 
  

 

 

 

 

 

 

 

 

 

 

 

    
 

 

 

 

 

 

 

LETHARGY
17
I
0024
3
Q
N
= 0020 -
£
&
£
> 0016 |
&
o
<I
I
= ool
)
=
=
o
z 0008
il
o
>
3
2 0004
0.0 0.2 05
ENERGY (ev)

Fig. 4.2, Calculated Thermal Neutron Spectra in the MSRE with

20 19

o

MSRE - 233 ) FUEL THERMAL

 

LETHARGY
17

;

ORNL-DWG €7 -11724
T 1117
FLUX SPECTRA AT THE
CENTER OF THE FUEL
CHANNEL ARE PLOTTED.
FLUXES AREL NORMALIZED
TO EQUAL. MAGNITUDES OF
TOTAL NEUTRON FLUX i L
BELOW 1{.86 ev, AVERAGED
OVER THE LATTICE CELL.

 

 

 

 

MSRE—-23%U FUEL

 

  

 

10

235

U and 233U Fuel Salts.

ORNL--DWG 6714795

14

 

 

 

FLUX SPECTRUM AT CENTER
OF FUEL CHANNEL

 
 
  

GRAPHITE

 

  
  

 

0.046

0.042

0o.008

c.004

 

 

POSITIONS FOR THERMAL FLUX PLOTS
A *B oC
- —

l !
2 4 & 8

 

 
  

 

 

  
 
 
 

 

 

RADIAL POSITION (cm)

MSBR CELL GEOMETRY

MSRE AND MSBER FLUX PLOTS ARE
NCRMALIZED TC EQUAL MAGNITUDES

OF TOTAL NEUTRON FLUX BELOW 1.3&ev,
AVERAGED OVER THE LATTICE CELL

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 4.3. Calculated Thermal Neutron Specitra in the

 

ENERGY (ev)

23

 

 

 

20

3U-Fueled MSRE and in the MSBR.

FUEL / FERTILE
- FUEL,%«\#}‘ GRAPHIT E’He

 

 
53

shown as solid points in Fig. 4.1 (points shown
only for the MSRE 233U spectra). The calculated
points are connected by straight line segments. As
might be expected from the gross similarities be-
tween the two reactor systems (temperatures, mod-
erating materials, total salt volume fractions),
there is a matked similarity between the two
spectra. Above energies of about 1 kev, the MSBR
flux spectra lie above that of the MSRE, due to the
smaller neutron leakage from the MSBR core. Below
about 1 kev the spectrum in the MSBR becomes pro-
gressively reduced by resonance neutron abscrption
in the fertile material, and the two flux spectra
tend to approach one another in magnitude.

Results of THERMOS themal spectra calculations
are shown in Figs. 4.2 and 4.3. Figure 4.2 com-
pares the MSRE spectrum for the 237U {uel sait
with that for the current 235y fuel loading. The
spectra at the center of the fuel channel are chosen
as a basis of comparison. In Fig. 4.3, the MSRE
thermal spectrum with the ?3°U fuel salt is com-
pared with the cortesponding spectra at several
points in the MSBR lattice. Note that curve 4 has
the same relative position in the MSBR lattice
(center of the larger fuel salt channel) as that of
the MSRE. For both these reactors the flux spectra
of Figs. 4.2 and 4.3 are nomalized to give the
same total integrated neutron flux below 1.86 ev,
averaged over the total volume of salt and graphite.
On this basis, relative comparisons can be made of
the enerpy (or lethargy) distributions and position
dependence of the spectra in each reactor. One
should note, however, that the actual magnitudes

of the total thermal flux will be quite different in
the systems, depending on the relative fission
densities and the fissile materdal concentrations.
In Fig. 4.2, the slight “energy hardening,”” or
preferential removal of low-energy neutrons for the
2350 fuel loading, reflects the larger uranium con-
centration for this case (V0.9 mole % total UF,,
33% eariched in 235U). Again, in Fig. 4.3 the
higher concentration and absorptions in fissile
uranium in the MSBR and also the absorptions in
the fertile salt produce a hardening of the energy
spectrum as compared with the MSRE. Just as in
the case of the epithermal spectra, however, the
similar temperatures and gross material compo-
sitions result in a marked similarty in the spectra.
An altemate means of comparing the neutron
spectra in the two systems is shown in Figs. 4.4
and 4.5. Here the calculated group fluxes have
been multiplied by the corresponding group micro-
scopic fission cross section for 222U, summed
over the energy groups above each energy point,
and nommalized to one fission event in 233U
occurring in each of the thermal (£ < 1.86 ev) and
epithermal (£ > 1.86 ev) energy regions. In the
latter case, most of the fission reactions occur
between the lower cutoff energy and about 30 ev,
where the flux spectra are quite close to one
another (Fig. 4.1). As a result, there is a close
cortespondence in the normalized distribution of
epithermal fission events in the two systems.
Larger differences are encountered in the thermal
fissions, as shown in Fig. 4.5. Note, however,
that the spectra at the center of the fuel channels

ORNL--DWG 6714796

 

TPITHERMAL {>1.86 ev) FISSIONS

z'
ABCVE ENERGY £

 

 

 

 

 

 

RACTION OF

—_
F

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1l
Ll

 

 

 

100 2 5 10t

ENERGY (ev)

Fig. 4.4. Normalized Distributions of Epithermal Fission Events in the 233U-Fue|ed MSRE ond in the MSBR.
FRACTICN OF THERMAL («188ev) FISSIONS

CRNL -DWG 67-1797

 

 

  
   
    

 

 

 

e e
FLUX SPECTRA TAKEN AT
0B . THE CENTER OF THE
| FUEL CHANNEL
W ‘
&
06 NN msas
& \
L
= | o
o o4 MSRE -233)) FuEL
<I
02
0 L) - - . : H’“‘_s
001 002 005 o4 02 05 f 2
ENERGY (ev]

are chosen as a basis for this comparison, so that
the curves in Fig. 4.5 should represent an upper
limit of the difference between the energy distribu-
tion of thermal fissions in these systems. The
overall ratio of epithermal to thermal fissions cal-
culated for the MSRE lattice was 0.15; this is also
very close to that obtained for the MSBR lattice.

The general conclusion obtained from these
studies 1s that there is sufficient similarity in the
neutron spectra in these two reactors that infer-
ences drawn from physics experiments with the
2337-fueled MSRE (i.e., critical experiments,
measurements of the effective capture-to-fission
ratio in the reactor spectra, etc.) should bear a
strong relation to the nuclear design of the MSBR.
These inferences include the adequacy of the
library of ??3U cross sections and the computa-
tional techniques used to evaluate the experiments.
The present studies, however, do not consider the
equally important questions of control of error and
wtcertainty in such measurements, and the sensi-
tivity of the measurement analysis to uncertainties
in cross-section data and theoretical modeling
techniques. Further studies along these lines are
planned, in order to better detemmine the relevance
of MSRE experiments with 233U fuel to the design
of the MSBR.

43 OTHER NEUTRONIC CHARACTERISTICS
OF MSRE WiTH 233U FUEL

B. E. Prince

Critical Loading, Red Worth, and
Reactivity Coeflicients

A summary of the important parameters calculated
for the 233U fuel loading is given in Table 4.2.

54

Fig. 4.5. Mormalized Distributions of Thermal Fission

Events in the 233U-Fueled MSRE and in the MSBR.

Table 4.2, Neutronic Characteristics of MSRE
with 233 Fuel Salt at 1200°F

 

. L . . a
Minimum critical uranium loading

Concentration, grams of U per liter 15.82
of salt
UF,, mole % 0.125
Total uranium inventory, kgb 32.8
Control rod worth at minimum critical
loading, % Ok/k -
One rod —2.75
Three rods ~7.01
Increase in uranium equivalent to in- 8.78
sertion of one control rod,¢ %
Prompt neutron generation time, sec 4.0 x 10_4
Reactivity coefficientsd
Fuel salt temperature, (OF)m1 -6.13 x 10777
Graphite temperature, (OF)"l —3.23 x 107°
Fuel salt density +0.447
Graphite density +0.444
Uranium concentration® +2.386
Individual nuclide concentrations
in salt
°Li —0.0313
1 ~4.58 x 107 *
°Be +0.0259
g +0.0706
0%, —0.00813 .
1495m ~0.0132 -
15igm —0.00163
233y +0.4100 i
234y ~0.0132 -
235y +0.00216
236y —2.58 x 1077
238y 4,93 % 107>
239 py +0.0122

 

“Fuel not circularing, contrel rods withdrawn to upper
limits.
3
PBased on 73.2 ft” of fuel salt at 1200°F in circu-
lating system aud drain tanks.

c . - . . -

Increase in uranium concentration required to main-
tain criticality with one rod inserted to lower limit of
travel, fuel not circulating.

c"At initial critical concentration. Where units are
shown, coefficients for variable x are of the form
51(/1(8)(; otherwise, coefficients are of the form xak/kfix.

®Isotopic compositions of uranium as given in
Table 4.1,
These are the results of two-dimensional, four-
group diffusion calculations, with the group cross
sections averaged over the neutron spectra shown
in the preceding section. Both RZ and R geo-
metric approximations of the reactor core were con-
sidered, in the manner used previously to analyze
the physics experiments with the *2°U fuel.* One
innovation was the use of the two-dimensional
multigroup EXTERMIN ATOR-2 program,® a new
version which includes criticality search options
and a convenient perturbation and reactivity coef-
ficient calculation.

The properties listed in Table 4.2 differ in
several important respects from the correspending
properties of the present ?35U fuel salt. The
{issile material concentration is only about one-
kalfl that for the present fuel, an effect which re-
sults mainly from the absence of a significant
quantity of #38U in the new loading. This same
effect produces an increase in the thermal diffusion
length in the core, thereby increasing the thermal-
nentron leakage (as reflected in the larger magni-
tudes of the fuel temperature and density coeffi-
cients of reactivity). The increased diffusion
length also results in an increase in control rod
effectiveness (2.75% 8k/k calculated for one rod,
as compared with 2.11% calculated and 2.26%
measured worth of one rod in the present *°U-
fueled reacter).® On the percentape basis given in
Table 4.2, the uranium concentration coefficient
is higher for the #3%U fuel by a factor of about 1.8,
owing to the absence of the 238U and also to the
greater neutron productivity of the *?%U. On the
basis of 1 g of excess uranium added to the fuel
circulating system (70.5 ft3), the reactivity change
is +0.00123% S&/k. This is about 3.8 times the
corresponding reactivity change when 1 g of 23°U
is added to the present fuel salt.

Fission Rate and Thermal Flux
Spatial Distributions

Spatial distributions of the fission density in the
fuel salt, obtained from the EXTERMINATOR dif-
fusion calculations described in the preceding
section, are shown in Figs. 4.6 and 4.7. Figure

 

4}3. E. Prince ef al., MSRE Zero Power Physics Ex-
periments (ORNL report in preparation).

>). B. Fowler ot al., EXTERMINATOR-2: A Fortran
IV Code for Solving Multigroup Neutron Diffusion Equa-
tions in Two Dimensions, ORNL-4078 (April 1967}.

55

4.6 shows the axial distribution of the fission
density at a radial position of about 7.3 in. from
the core center line, which is the approximate
radial position of maximum flux with all control
rods withdrawn. The fission density in the salt is
nomalized to 10°® fission events occurring in the
entire system. Also shown in Fig. 4.6 is the axial
distribution of neutron importance for the fast
group, which is used in the calculation of the
effects of circulation on the delayed precursors
(see later section).

Figure 4.7 shows the comrresponding radial dis-
tributions of fission density in the salt, both with
all rods withdrawn and with one and three rods
fully inserted. These curves are based on calcu-
lations with an R0 model of the core geometry.

The approximate geometry used for the control
element and sample-holder configuration is indicated
schematically in the figure, and the angular position
of the flux traverse along the reactor diameter is
also shown. In the position where the traverse
intersects the control element (Fig. 4.7), the curves
for the center portion of the salt-graphite lattice are
connected with those for the major part of the lat-
tice by dashed straight lines extending through the
control region.

The cortesponding distributions of themal flux
are very similar in shape to the fission density
distributions. These are given in Figs. 4.8 and
4.9. The magnitudes of these fluxes are based on
1 Mw of fission energy deposited in the system,
using the conversion factor of 3.17 x 10'% fis-
sions/Mwsec. The flux magnitudes are also
nommalized to the minimum critical uranium loading,
s0 that renormalization of these values would be
necessary to apply them to the exact operafing
concentration. For this purpose the flux can be
assumed to be inversely proportional to the 2270
concentration. Note alse that, in these calcula-
tions, the thermal neutron {lux is defined as the
integral flux below a cutoff energy of 0.876 ev
(see the discussion in Sect. 4.32).

Reactor-Average Fluxes and Reaction
Cross Sections

Cross sections for important nuclide reactions
in the MSRE 23U spectrum are listed in Table 4.3.
These values of the thermal and epithermal aver-
ages are based on a cutoff energy of 0.876 ev. To
obtain the final column, we have used the average
56

ORNL-DWGC 67-441738
4.8

 

 

 

 

 

 

 

   
 
 
 
 

 

 

AXIAL CENTER LINE OF CORE

 

 

,,,,,,,,,,,, I
v

x —
5 T | z
2 - S - e ~{a0 ¢
© ™~ I >
O \ o o
= < l <
a Pl \\ g | é
" ; — 5
5 81 """ 132,
2 = =
2 S 2
v Q. o
= o g
- - =
A28 4= N e 24 =
& | o
Z RADIAL POSITICN, 7.28 in. FROWM E
> o
._
@
P
wJ
O
=z
Q
o
o
L

 

 

 

 

 

 

 

 

 

 

 

 

Lef
=
1.6 5
O
e
o
}._
wy
<
L
8 ] - 08 -
N N
. *
-

ol -der L L L . 0

20 30 40 50 80

AXIAL DISTANCE (in.)
Fig. 4.6. Axial Distributions of Fission Density and Fast Group tmportance in the MSRE with 233 Fyel Loading.

ORNI-DWG §7-14799
T 10 : R
{27
FISSION DENSITY s FISSION DENSITY
TRAVERSE = fi) TRAVERSE
e ~

 

 

3-5 } ------ | i | T T

 
  
 
 
 

30

   

THREE RODS
WITHDRAWN

    
   
 
 
   
 
 

1’77“\

J

S

CONTROL ROD AND SAMPLE
HCLOER CONFIGURATION

25

§

ROD NO.2 INSERTED - -

\
= -THREE RODS INSERTED - \\
~

  
 

-ROD NO.Z2 INSERTED !

THREE RODS INSERTED

    

FISSION DENSITY IN SALT (fissions/cm> per 08 fissions)

}-- SAMPLE HOLDER REGION
\
\
!

 

INORMALIZED TO AXIAL POSITION OF MAXIMUM FISSION DEMSITY)

 

15 10 5 C 5 10 15
RADIAL DISTANCE (in)

Fig. 4.7. Radial Distributions of Fission Density in the MSRE with 2331) Fyel Loading.

epithermal-to-thermal flux ratios, determined by magnitude of the thermal fluxes, give the approxi-
averaging the EXTERMINATOR group fluxes over mate total reaction rates per atom for neutrons of
the volume of the fuel circulating system. These all energies in the spectrum. Table 4.3 also gives

eftective cross sections, when multiplied by the the calculated magnitudes of thermal and epithermal

«
 

 

'; A2y —— S - PO
= (107 L

.

o

a | .
T

0

3 8 |

5 -

3 g W

£ oo

© Z W

= 2

o =
z jsed
,,,,,,, O e

= G T i

o u

L 5

; o

= 5 o.l

= qal..E 2 I

o 'C_) b]

z 0

& |

‘r..

= N
i | 2330 LOADING
> o | i

_

<

=

n-

1l

 
   
   

RADIAL POSITION, 7.28 in. FROM AXIAL
CENTER LINE OF CORE. FLUX MAGNITUDES
NORMALIZED TO MINIMUM CRITICAL

QRNI-DWG 67-11800

-

 

  

 

 

 

 

 

¢r,TH

30

 

40

AXIAL DISTANCE f{in))

Fig. 4.8. Axial Distribution of Therma! Neutron Flux in the MSRE with 233 Fuel Loading.

 

 

  
 
 

     
 
       

 

 

 

     
   
  
 
    

 

  

 

CRNL -DWG 671301
. T TRy e m e | I T T T tTttTT Tttt
(x 40') | | o | ‘ ‘
9 THREE RODS WITHDRAWN — | ; } | /THHEE RODS  WITHDRAWN |
z | || r \3/
. | | | FLUX N FLUX
Fa & || TRAVERSE (5 | 7g) TRAVERSE
§ N\ L I l . ! -
o NN {1\
'e T \ Y~ ROD NO. 2
c | I INSERTED CONTROL ROD AND SAMPLE HOLDER
= CROD NO. 2. CONFIG
N MEERTED | SONFIGURATION
= |
& |
& _ THREE RODS __ |
x S INSERTED | -\
i’ . ‘ |
é 4 - | | NOTE: FLUX MAGNITUDES  \\
i \ | | | NORMALIZED TG POSITION OF
W \ \ | MAXIMUM  AXIAL FLUX, AND TO
_ 3 \ I | MINIMUM  CRITICAL #22U LOADING
<1
2 |
& L
= b
s CONTRCL | | e SAMPLE HOLOER
5 4 ELEMENT —f o RIEGION
‘ ! |
|| |
o 0 R P R N
20 25 20 15 10 5 0 5 1C 15 20
RADIAL DISTANCE {in)

Fig. 4.9. Radial Distributions of Thermal Neutron Flux in the MSRE with 233y Fyel Loading.

flux, averaged over the volume of the fuel circu-
lating system and normalized to 1 Mw. As dis-
cussed in the preceding section, these magnitudes
are nommalized to the minimum critical uranium
concentration. Note also that these values include
the ‘“flux dilution’’ effect of the time the fuel

spends in the external piping and heat exchanger.
(The corresponding value of the thermal flux,

averaged only over the graphite-moderated region
of the core, is 3.60 < 1012 neutrons cm ™ ? sec™?
Mw™1)
58

Table 4.3. Average Cross Sections for Thermal- and Epithermal-Neutron
Reactions in the MSRE Spectium with 233y Fyel

Cross Section Averaged

over Thermal Spectrum,

over Fpithermal Spectrum,

Cross Section Averaged
Effective Cross Section

 

Nuclide Energy <0.876 ev Energy > 0.876 ev in Thermal Flux
(bamns) (barns) (barns)

bLi® 442.1 17.7 473.6
L4 0.017 0.0007 0.018
'Be 0.0047 0.0040 0.012
Boron 353.1 14.1 378.2
e 0.0019 <107t 0.0020
" 0.0047 0.0023 0.0088
Zirconium” 0.0865 0.0764 0.222
13%%e 1.32 % 10° 83.3 1.32 x 10°
149%m 3.91 x 10° 92.1 3.93 x 10°
151gm 2900.0 128.7 3127.0
233y (abs) 272.1 45.6 353.2
233y (fission) 249.2 (1 = 2.50) 38.4 317.6
234y 43.5 38.1 111.3
2350 (abs) 290.3 22.3 330.0
235y (fission) 244.2 (1 - 2.43) 13.6 268.5
236y 2.8 18.9 36.4
238y 1.3 16.7 31.0
239py (abs) 1402.6 22.5 1442.7
239y (fission) 836.0 (1 - 2.89) 13.9 860.8

System-average® neutron fluxes (neutrons cm 2
Thermal: 1.48 x 10°°
Epithermal; 2.64 x 1012

. . 6_ .
“Cross section for the reaction 141(11,(1)3!-1.
PNatural isotopic composition,

“Rased on 2.00 x 10° ¢m? of circulating fuel.

Effect of Circulation on
Delayed-Neutron Precursors

In previous studies, we described a mathematical
model found to be in close correlation with meas-
ured values of the delayed-neutron loss caused by
circulation of the present 23°U fuel. 46 We have

 

6B. E. Prince, Period Measurements on the Molten
Salt Reactor Experiment During Fuel Circulation:
Theory and Experiment, ORNL-TM-1626 (October 1966).

applied this same computational model to the ?33U-
fueled reactor in order to obtain the reactivity
change due to circulation and also to obtain the
reactivity increments necessary to produce a given
stable period during circulation. In these calcu-
lations, temperature feedback effects are neglected.
Plutonium-239 is present in sufficiently small
amounts to be negligible also.

The net effect of the circulation in changing the
shape of the distribution of delayed neutron emis-
sion within the reactor is shown in Fig. 4.10. This
-

59

ORNL—DWG 711802

 

 

 

 

  

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

6 .....
[ [
| |
5l l_ A IR T T TG L |
e il |
£ 4 @il — | e
0
8% &5l = \
T 5wl a1
zZ5 3 sl ot N _
o7 3 LLI (DI
- oL =
0w to- (L w3
Su_ 8 Ol QI
— =
2F Sl o] 7 DIRECTION OF FUEL CIRCULATION —-w ol | |
a® | ©|
52 | S
L)t
W |
<Ll P —
we AT B T B i
| |
| | |
o o J |
—40 0 10 20 30 a0 50 0 70 80

AXIAL DISTANCE (in)

Fig. 4.10. Axia! Distribution of the Total Source of Delayed Neutrons in the MSRE with 233 Fuel Lsading.

corresponds to the just-critical, circulating condi-
tion. By comparison, the distribution of the prompt -
neutron source, as well as the delayed-neutron
source in the noncirculating condition, is propor-
tional to the fission density curve in Fig. 4.6.

The effective (reduced) values of 8, the delayed
fractions for group 1, in the critical, circulating
condition are listed in column 4 of Table 4.4. To
obtain these values, the components of the total
delayed-neutron source distribution shown in
Fig. 4.10 must be weighted by the fuel salt volume
fraction and also by the distribution of fast-group
importance shown in Fig. 4.6. These calculations
were made by using an IBM 7090 numerical inte-
gration program, based on the theory given in ref. 6.

‘The difference between the sums of the 3, columns

of Table 4.4 for the stationary and circulating con-
ditions is equal to the reactivity loss due to ciccu-
lation. In normal reactor operation, this difference
will be observed as an additional increment of
regulating rod withdrawal, asymptotically required
to maintain criticality when going from noncirculat-
ing to circulating conditions.

Table 4.4. Delayed-Neutron Fractions in the
MSRE with 233U Fuel

 

Delay Decay Constant Delay Fractions (n/n)

 

 

G ( ~1.a
Hroup sec ) Stationary®  Circulating
%107 %
1 0.0126 2.28 1.091
2 0.0337 7.88 3.848
3 0.139 6.64 4.036
4 0.325 7.356 5.9672
5 1.130 1.36 1.320
6 2.500 0.88 0.876
Total 26.40 17.14

 

?G. R. Keepin, Physics of Nuclear Kinetics, p. 90,
Addison-Wesley, Reading, Mass., 1965.

In Fig. 4.11 are curves showing the reactivity
addition required to produce a given positive stable
period, starting from the critical condition (w == 0
at zero point of reactivity). Curve A, for the non-
circulating condition, is calculated with the
ORNL-DWG 6711803

 

 
  
 

       

   

030 DT
B-CALCULATED WITH STATIC
025 INHOUR FORMULA,USING |
s REDUCED VALUES OF DE-
= LAYED NEUTRON FRAC -
o TIONS
8 020 °
— C: CALCULLATED WITH NU--
g MERICAL INTEGRATION
& 015 PROGRAM
= | | fi ‘ o
Soop ||
= o ‘ !
g i L ‘ :
< i Ll | :
g . [
005 — . ‘ ‘ |
0 e T i

 

 

 

 

 

0001 0002 0005 0O 002

- i
005 01 02 05 1 2 5 10

 

 

  
  
   
 
 
   
  
   

. T

\ [ ‘ : :
A FUEL

/7 STATIONARY

o

Y

, B FUEL
¥ C CIRCULATING

|

i
i
i
|
i
|
\
1

 

 

 

 

 

 

 

 

 

 

INVERSE STABLE PERICD (sec )

Fig. 4.11. Reactivity Addition Required to Produce a Given Stable Period in the MSRE with 233y Fuel L.oading.

““‘static’’ inhour equation,”® using the 3, given in

column 3 of Table 4.4. As a first approximation to
the corresponding relation for the circulating con-
dition, curve B is also calculated with the static
inhour equation, using the reduced ,81. given in
column 4 of Table 4.4. Finally, in curve C, we
show the relation obtained from the numerical inte-
gration program, which gives the complete treat-
ment of the precursor production, motion, and
decay, when the reactor is on a stable period. In
the range 0.001 S w 5 1.0, curve C “bows’’ away
from curve B, illustrating the fact that the effec-
tive reductions in the delay fractions are actually
dependent on the reactor period.® As «w becomes
large compared with the precursor decay constants,
A;, it can be shown that curve C approaches curve
B asymptotically. Furthermore, curve B differs
asymptotically from curve A, on the vertical re-
activity scale, by an amount equal to the initial
reactivity loss caused by circulation (0.093% 6k/k,
from Table 4.4).

Samarivm Poisoning Effects

At the end of 60,000 Mwhr of operation with the
current fuel charge, the ' *°Sm will be very near its

equilibrium concentration corresponding to 7.4 Mw.

 

7A. F. Henry, Nucl. Sci. Eng. 3, 52--70 (1958).

8E. E. Gross and J. H. Marable, Nucl. Sci. Eng. 7,
28191 (1960).

The '°1Sm will achieve approximately 30% of its
equilibrium concentration. After shutdown the
149Sm will be further enhanced by about 8%, owing
to decay of '*?Pm. These samarium poison con-
centrations will be present in the salt when the
reactor is brought to power with the 233U fuel
loading. For this new loading, however, the theimal
flux will be approximately 2.2 times that for the
235 fuel. In addition, the yields of the 149 and
151 fission product chains are lower than those for
the 235U (0.77 and 0.35% for 233U, vs 1.13 and
0.44% for 23°U, based on ref. 9). The net result

is that the samarium initially present in the salt
will act as a burnable poison during the first part
of reactor operation with ?**U. Figure 4.12 (top
curve) shows the result of calculations of the
reactivity change corresponding to burnout of the
initial samarium and achievement of the equilibrium
poisoning at the higher flux level. Continuous
operation at a power level of 7.4 Mw is assumed.
The lower curve in Fig. 4.12 shows the correspond-
ing reactivity change caused by burnup of 233U
during this period. The algebraic sum of these two
curves (dashed curve) gives the approximate re-
activity effect which must be compensated for by
motion of the regulating rod. (There will be addi-
tional corrections due to other isotopic changes,
but these will be small compared with the samarium

 

°T. R. England, Time-Dependent Fission Product
Thermal and Resonance Absorption Cross Sections,
WAPD-TM-333, Addendum No. 1 (January 19635).
61

o R e

  

[ P e

 

ACTIVITY % B /%)

 

=
T

 

]

 

 

 

 

APPROXIMATE REACTIVITY C“IANGE e
COMPENSATED 3Y COMTROL RCD MOTION -

o [ ,,,,,,,,,,,,,,, |
T

 

200

 

300
TIME-INTEGRATED POWER (Mwd)

ORNL.-DWG 87-11804

 

NET SAMARIUM CHANGE
(BURNOUT - PRODUCTION)

 

-r.;,* -

e

~. 223 BURNUP

 

 

400 500

Fig. 4.12. Time Dependence of Samarium Poisoning and Approximote Control Rod Reactivity During the First

Period of Operation of MSRE with 233y Fyel lcading.

and 2330 bumup effects.) The resultant curve in
Fig. 4.12 indicates that regulating rod insertion
will be required for about the first 70 days of
operation.

Because the samarium acts as a burnable poison,
the amount of excess uranium required for operation
will be largely governed by the requirements for
calibration of the control rods. There should be
considerable latitude in this choice, detemined in
part by criteria of control rod seasitivity at the
operating point and by the maximum time of opera-
tion desited before refueling.

4.4 MSRE DYNAMICS WITH 233U FUEL

S. J. Ball

The dynamic behavior of the MSRE has been
analyzed for the case of *?%U-bearing fuel salt by
using the MSRE frequency response code MSFR. '°
The only differences in input data for the 2370
and the reference 2°50U calculations are in those
parameters compared in Table 4.5. The most im-
portant difference is the lower delayed-neutron
fraction for the *3°U, which makes the neutron
level more responsive to changes in reactivity.
For a given fast change in rod reactivity, the
immediate flux response would be two to three
times greater with *?°U than with #3°U. This
effect will probably require & minor modification in

 

10
5. J.
Molten Salt Reactor Experiment,
ber 1965).

Ball and T. W. Kerlin, Stability Analysis of the
ORNL-TM-1070 (Decem-

the present MSRE rod control system to compensate
for the higher system gain.

The predicted effect of ?3°U on the MSRE in-
herent stability, as indicated by the phase margin,
and the natural period of oscillation are shown as
a function of power level in Fig. 4.13. The phase
margin '! is a typical measure of system stability,
the smaller phase margins indicating reduced sta-
bility. A general rule of thumb in control practice
is that a phase margin of at least 30” is desirable;
phase margins of 20° or less indicate lightly
damped oscillations and thus poor control. Hence
the predicted inherent stability for the *3?U system
is greater than for *°U for all power levels.

The faster response of the ***U system, as indi-
cated by the smaller natural periods of oscillation,
is due to the higher gain of the neutron kinetics.
Experimentally determined values of period of
oscillation for the ?3°U system are also shown in
Fig. 4.13 for comparison.'? Although direct
measurements of phase margin were not made, the
stability characteristics as indicated by frequency
response measurements were also in good agree-
ment with the predictions.

It is concluded that no serious operational diffi-
culties are to be expected due to the differences in
dynamic behavior resulting from the 223U fuel
loading.

 

“_] E. Gibson, Nonlinear Automatic Control, chap. 1,
MCGraw~H111 New York, 1963.

T. W. Kerlin and 8. J. Ball, Experimental Dynamic
An.alysis of the Molten Salt Reactor Experiment, QRN L.-
TM-1647 (October 1966).
62

Table 4.5, Comparison of MSRE Nuclear Data for 233y and 235U-Bearing Fuel Salts

1

 

 

 

 

 

Static delayed-neutron fractions, £, and precursor decay constants, A (sec )
233 235,
Group
A B A .
I I 1 1
1 0.0126 0.228 x 1073 0.0124 0.223 x 1072
2 0.0337 0.788 x 103 0.0305 1.457 = 107 °
3 0.139 0.664 % 10 ° 0.111 1.307 x 1073
4 0.325 0.736 x 103 0.301 2.628 % 1073
5 1.13 0.136 x 10> 1.14 0.766 x 10 >
6 2.50 0.088 x 1073 3.01 0.280 x 1073 -
Total static 8 0.00264 0.006606
Prompt-neutron generation 4.0 x10* 2.4 x 1074 -
time, sec
Temperature coefficients
of reactivity, (OF)_1
Fuel salt 613 x 107 ° —4.84 + 1077
Graphite ~3.23x10°° —3.70 x 107"

 

100

50

20

Q5

0.05 Ot 0.2

 

05

ORNL-DWG &7-118B05

PHASE MARGIN (deg)
T 238y FUEL
233y FUEL

PERIOD OF OSCILLATION (min)
232 FUEL (e EXPERIMENTAL | | -

PERICQDS}
U FUEL

233

1 2 5 10

POWER LEVEL (Mw)

Fig. 4.13. Comparison of MSRE Fredicted Phase Margins and Natural Periods of Oscillation vs Power Level for

2331. and 2:'Lr’l,l-Beui'itng Fuel Salt Loadings.
Part 2. MSBR Design and Development

R. B. Briggs

The primary objective of the engineering design
and development activities of the MSR program is
to design a molten-salt breeder experiment (MSBE)
and develop the components and systems for that
reactor. The MSBE is proposed to be a model of a
large power breeder reactor and to operate at a
power level of about 150 Mw (thermal).

A reference design of a 1000-Mw (electrical)
molten-salt breeder reactor power plant is to pro-
vide the basis for most of the criteria and for much
of the design of the MSBE. We have been spending
most of our effort on the reference design. Sev-
eral concepts were described in ORNL-4037, our
progress report for the period ending August 1966,
and in ORNL-3996, which was published in August
1966. We selected the modular concept — four 250-
Mw (electrical) reactors per 1000-Mw (electrical)
atation — that has fissile and fertile materials in
separate streams as being the most promising for
irmmediate development and have proceeded with
mote detailed study of a plant based on that con-
cept. Initial results of those studies were re-
ported in ORNL-4119, our progress report for the

5.

period ending February 1967. During this report
period, we continued to examine the details of the
reference plant and its equipment. We also sum-
marized the objectives and program for develop-
ment of the MSBE in ORNL-TM-1851.

To date, the component and systems development
activity has been concerned largely with support
of the design and with planning. The development
program was outlined in ORNL-TM-1855 and in
ORNL-TM-1856. Salt circulation loops and other
test facilities are being modified for tests of the
sodium fluoroborate coolant salt, models of graph-
ite fuel cells for the reactor, and molten-salt bear-
ings and other crucial components of the fuel cir-
culation pump. Essentially no experimental work
has heen done yet.

We plan to complete essential parts of the ref-
erence design during the next report period and to
begin to design the MSBE. Experimeatal work
will begin also but will be limited to a few of the
mote obvious important problems. Design and
development on a large scale are planned to start
early in FY 1969.

Design

E. 5. Bettis

5.1 GENERAL

E. S. Bettis R. C. Robertson

The design program during the past peried has
primarily been directed toward refinement and
modification of the previously reported concepts
for the reactor and other major equipment and re-
visions to the cell layouts and piping.

63

The major effort was redesign of the reactor to
accommodate the new data on radiation effects in
graphite. New approaches wete tried to arrive at
a core design which could compensate for dimen-
sional changes in the graphite, but it was finally
decided to simply double the volume of the core
to lower the power density from 40 to 20 kw/liter
and thereby assure a stable core life of at least
ten years. It was also decided to include pro-
visions for replacing the complete reactor vessel
ORNL-DWG &£7-3081AR

Bl ANKET SALT
HEAT EXCHANGLR

AND PUMP
FUSL SALT
HEAT EXCHANGER -
AND PUMP e COOLANT -
. BLANKET 5401
RZACTOR SALT PP R '

   
 

1
PROCESSING | | 2!

 
 
 
   

 

  
 
 

  
  

 

    
     
   
 

 

  

 
  
 

 

   

 

 

 

|
|
|
' |
| e —— |
: Pl ! ‘j |
FUEL : / : L 3z L o ! |
PROCE SSING : T LT P
4 #¥day lpe 23 1 |
4_—} . R ot el GENCRA
¥ ] o NCRATOR
——— ~ . 4 {
z —f1 | H_Lfi FRTE I 5272 Mwte)
= =l o e :
o it | ] ; i 35 : ;el
N 3 1 {u‘v .
w2 ' i . '
| ‘ o - - L GENERAIOR
= 1 o b : v Y
2 o i ILPT ALPT 35077 Mwle) |
o - GASEQUS L 3 P F.
= 3FISSICN ‘ ‘ I L o
i f gflsorggirs _ !
HSPOSAL For — i
o VG TE N Loy ONDENSING
I SYSTEN: e i ! \
el ?E, FEED WATER
HEF SYSTEMS /
{ T
! — 5
i - BOILER SUPCRHEAT-R REHEATE RS REHEAT STEAM 62
i ‘ Ve ' PREHCAIFRS BCOSTER MIXING
L =5 R o PUVP TEE
FUEL SAIT FLUSH-  BLANKET SALT COCLAN: SALT
DRAIN IANKS — SALT DRAIN TANK IRAIN TANK
DRAIN TANK
FUEL COOLANT STEAM PERFORMANCE DATA
POINT ft/sec psig TEMP (°F) POINT t13/sec psig TEMP (¢F) POINT 10% Ib/hr psta | EMP (°F) 1000 Mwle) GENERATION P_AN]
f 25.0 131300 L 375 129 1125 50 257 3800 1000 NET OUTPUT 1000 Mwie)
2 250 9 1300 31375 10 1125 51 10,07 3600 000 GROSS GENERATION 1034.9 Muw (e)
3 250 146 1300 32 375 2680 125 52 715 3515 1000 BOILER FEED PUMPS 29.49 NMw
4 250 110 5% 5.0 208 125 53 297 3515 {000 BODSTER PUMPS 9.2 Mwle)
5 250 50 1000 34 50 212 850 54 292 3500 866 STATION AUXILIARY P57 Mwlel
& 250 31 1000 35 841 252 1125 55 514 600 552 REACTOR HIAT INPUT 2225 Mwl(t]
7 250 18 1000 36 81 194 350 56 513 570 650 MET HFAT RATE 7601 Biudkw hr
37375 P03 350 57 128 570 550 NET EFFICIENCY 44979
BLANKET 8 375 198 850 58 128 540 000 LEGHND
20 43 125 1250 39 375 164 1111 59 513 540 1000 FUEL —_
21 43 60 1250 40 375 38 114 60 500 172 708 BLANKET T
: COOLANT —_—
22 43 1110 1250 61 430 15inHg 97 ST r A -
23 a3 200 1150 62 716 3500 551 WATER e
24 43 445 1150 63 1007 3475 695 FREEZE VAl vE i
64 1007 3800 700 JATA POINT X
65 257 3800 700

Fig. 5.3, 250 Mw {Efectrical) Module Heat Flow Diagram.

va
€

*

rather than just the core and to provide space to
store the spent reactor within the biological shield-
ing of the reactor cell. While this new design rep-
resents the present status of graphite technology,
we believe that an improved graphite will be de-
veloped within the next few years and that longer
core life or higher power density than used here
can be expected in the future.

Analysis of the thermal expansion stresses in
the various systems indicated that rather extensive
modifications were needed. Changes were there-
tore made in the plant layouts, the equipment sup-
ports, and in the arrangement of the salt piping,
with the result that the calculated stresses in the
vessel attachments, supports, and piping systems
now fall well within the acceptable values.

Calculations of the radiation flux levels at the
reactor cell walls allowed designs to be prepared
for the thermal shielding used to protect the con-
crete from excessive temperatures. Drawings were
also made for the peneral cell wall design, in-
cluding the closure blocks at the top, the cell
penetrations, etc.

Relatively few changes were required in the
MSBR flowsheet during the past period, but for
convenience it is presented again in Fig. 5.1.

The MSBR design study is expected to be com-
plete within the next few months. The description,
cost, and performance data for the steam system
will probably need little change but will be up-
dated as required. The August 1966 design study
report on a 1000-Mw (electrical) molten-salt breeder
reactor power station (ORNL-3996) will be revised
to present firmer cost estimates based on the new
and more detailed designs for the reactor plant.

5.2 CELL ARRANGEMENT

C. E. Bettis W. Terry

H. L. Watts C. W. Collins
H. A. Nelms W. K. Crowley
J. R. Rose H. M. Poly

The reactor is now designed on the basis that
the entire vessel will be of all-welded construction
and arranged for relatively simple replacement
after the useful life of the core has been expended.
Replacement of the reactor would of course require
provisions for disposal of the spent assembly.
Rather than lift this relatively large and radio-
active piece of equipment from the reactor plant

biological shielding, which would require a large
amount of heavily shielded transport equipment,
room has been provided in the reactor cell to store
the spent unit until the second reactor vessel is
replaced. In the intervening period the activity
level of the first spent unit would have decayed

to a more manageable value. This is the basis of
the present design, but, of course, the maintenance
procedures will continue to receive careful study
and review,

Storage of the spent reactor vessel in the cell
will require an enlargement of the cell volume by
about 40%. Since simplification and reatrangement
of the interconnecting salt piping were also re-
quired to satisfy stress requitements, the layout
of the entire reactor plant was rearranged. 'The
new layout is shown in Fig. 5.2.

The plant consists of four modules, each having
a reactor with a thermal output of about 556 Mw.
Steam is generated at 3500 psi—1000°F and re-
heated to 1000°F in each of the modules, but the
flows are combined to serve a single 1000-Mw
electrical turbine-generator unit.

As shown in Fig. 5.3, each module consists of a
reactor cell, a steam-generating and steam reheat
cell, and a drain-tank cell, the last being shared
with one other module. A drain-tank cell thus con-
tains four fuel drain tanks, two for each reactor.
The blanket and coolant salt drain tanks are also
contained in the two drain-tank cells,

The reactor and steam cell elevations ate shown
in Fig. 5.4, The equipment in these cells is
mounted on rigid supports, with thermal expansion
loops being provided in the connecting piping to
maintain the stresses within the allowable limits.
The reactor vessel is shown mounted on a single
column that is pinned at the bottom to allow some
lateral movement of the reactor. This arrangement
appatrently presents few major design problems and
reduces the stresses within certain elements of
the system; but, since the stresses are not prohib-
itive even without use of the pinned column sup-
port, this arrangement may receive further review.

In previously reported MSBR concepts the re-
actor was connected to the heat exchangers through
concentric pipes in an arrangement designed to
minimize the salt volumes and to simplify the ther-
mal expansion problems. As will be explained sub-
sequently, however, the doubling of the reactor
core volume made minimization of the fuel sait
volume in the piping less important, and this piping
change had relatively little effect on fuel cycle
66

3fr

3t

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

  

3 1
| Y
i A !
| ,;'_ o
}, — - |
T
| j_ -]
454+ I
] li;Y,, e 1
it ‘
[ e oo
| ¥
oy S
! 41
i
i
148 ft
56 4
|
; !
i e A -
i - ! ; :
i L I P
; . \ b ‘
; [ P :
: 40 ft CHEMICAL | 3 Pl ;
; PROCESSING L= = bor e
| c — -
| T D
g PROCESS ROOM L Ll b [
1 T N
= 7311
e 178 ft

Fig. 5.2. Reactor Salt System Layout. Plan

costs, The concentric piping was also found to
present stress problems at certain temperatures of
operation. For these reasons the piping was
changed to provide separate reactor inlet and out-
let lines for the fuel and blanket salt streams.
The separate piping will also simplify the main-
tenance procedures with remotely operated tooling.

The coolant, or secondary, salt piping in the
steam cell was also modified to include thermal
expansion loops and to include fixed supports for
the equipment. The location of the coolant salt
circulating pump was changed in order tc gain the
necessary flexibility in the piping.

Radiation shielding, formed of 3-in.-thick steel
plates, lines the reactor cell to protect the con-

23 Gins ‘r-«ZO‘T e 23t Gin, -

 
         
 
       
 
       

ORNL-DWG 87-10840A -

3f1 3ft

- 4t

COOLANT DUMP TANKS

COOLANT PUMP
- REHEATERS3

-STEAM GENERATORS
BLANKET DUMP TANKS
FUEL DUMP TANKS

BLANKET
HEAT EXCHANGER

 

 

PRIMARY
HEAT EXCHANGER -

REACTOR

 

FUsH SALT TANK

 

 

OFF-GAS
PRCCESS ROOM

  
  
  

INSTRUMENTATION

HOT CELLS

|
o
and piping for ™~ 1000 Mw (electrical) unit.

crete from excessive temperatures due to absorbed .
This thermal shield was designed to
attenuate a flux of 1 x 101? 1-Mev gammas cimn ™
sec™ ! and assures a concrete temperature of less
than 200°F. As shown in Fig. 5.5, the mounting
columns for the components are fitted with double
bellows seals where they penetrate the thermal
shield. It may also be noted that the columns are
supported on shock mounts to provide the requi-
site protection against seismic disturbances.

The walls of the reactor cell present the most
serious design problem. They are exposed to
gamma flux heating and to an ambient temperature
as high as 1150°F inside the cell, they must be
leak-tight to maintain containment integrity, and

radiation.
2
67

ORNL-DWG 57-10639A

SUPPORT FOR REACTOR

STORAGE \

/ STEAM CELL

   
  

  

 

i)

L&
{

- k) . v, 2 / - : 5 - s e . . . .
. 7 3 IS - ' - . " e e - - - -
r R i : i L Ct| e
Y ’fl g S ] y ": ‘\‘\\ 95:-'1 3 ' ! ’:_‘/" ':’-,' AR et |
: VO« L) T R = - L > S |
LY A / 4 . -
CNEEE ST : - Ry E Fezrdl A
- ey T RN S . STEAN PIPING
A e oo L - ? - ) ; o fi)'_ . g SO g s, v X e
: | @’flw % A

   
  
   

 

 

OFF -GAS 2 X ey
\ PROCESS ROCM oy SYEYEHEY

 

 

 

      

el 1 L Tt e O / A i e W\,__L”
REACTOR CELL / STEAM CELL N CUMP AND STORAGE

TANKS CELL

Fig. 5.3. Reactor and Steam Generator Cells ~ Plan. 250 Mw (electrical) module.

ORNL~-DW(G 67-10645A0

FUEL AND BLANKET
PUMP DRIVE MOTORS —r

N

 
  
  
  
 
 
 
  

STEAM L COOLANT SALT PUMP
REACTOR GEMERATORS

ST TN O
4 g

 

 

 

 

 

 

 

 

STEAM
/e ING

 

 

- REHEATERS

 

62 Tt 2in.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[0 1)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.:-5)}(2—3&1\: e iz
REACTOR SUPPORT FQR STORAGE S RERIMARY HEAT EXCHANGER AND PUMP

Fig. 5.4. Reactor and Steam Generator Cells — Elevation. 250 Mw (electrical) module, section A-A.
 
 

{

   

 

—55ft gin.———-

 

 

 

ISOLATOR —

T REACTOR
|
I

VIBRATION p

68

ORNL-DWG 67-1064848
— - B&ftOin—- - . R —

 

      
   

PRIMARY
HEAT
EXCHANGER

(\fl

! |
J (//l L\\W BLANKET
- | :

HEAT
EXCHANGER

|
l

 

/
/

  

  

 

 

 

 

Fiy. 5.5. Reactor Cell Thermal Shield ond Component Supports.

they must also provide the necessary biological
shielding for the reactor.

As shown in Figs. 5.6 and 5.7, the reactor cell
biological shielding consists of 7-ft-thick rein-
forced ordinary concrete around all sides below
grade and an 8-ft thickness above grade. The top
consists of removable concrete roof plugs having

a total thickness of 8 ft. The method of construc-
tion would be to first pour the 2-ft-thick reinforced
floot pad. A 3-in.-thick carbon steel floor plate
would be laid over this, and the 3-in.-steel vertical
wall plates would be welded to it, the latter serv-
ing as forms for pouring the side walls. A second
3-in.-thick steel plate would be erected inside the
COOLING CHANNEL

7‘

   
  
   
     

 

----- HEAT REFLECTOR

  
   

— DOUBLE CONTAINMENT

69

ORNL~-DWG 67-40644A

 
 
    

 

L

R e i
S

 

 

 

 

 

 

 

 

 

 

 

 

THERMALSHELD///

 
      
  
 

0O 24 6 810
Ll L L]
INCHES

Fig. 5.6. Reactor Cell Construction

first with a 3-in. air gap between the two. A sim-
ilar second plate and air gap would be provided
for the floor. During reactor operation, cooling
air would be circulated through the gap at a ve-
locity of about 50 fps to remove the heat penet-
ated by gamma absorptions in the wall. Cooling
air is also used in the removable roof plugs, the
air duct connections being flanged to facilitate
removal. The total heat losses from the reactor
cell are estimated to be a maximum of 800,000

 

 

 

VIBRATION
ISOLATOR

 

 

 

 

 

 

 

 

 

 

 

 

— Component Support Penetration.

Btu/hr. The thermal shield could tolerate loss of
cooling air for up to 1 hr without an excessive
temperature rise.

A 3/16--in.—thick carbon steel membrane is in-
stalled inside the inner 3-in.-thick steel wall
mentioned above. This membrane is hermetically
tight and would satisfy the containment leak-rate
requirements. The membrane is continuous around
the top plugs and also at the penetrations of the
cell, as shown in Fig. 5.7. The space between it

EQUIPMENT SUPPORT
   

 

ORNL-DWG 67-106374
COOLING SUPPLY

 

 

 

 
 

 

 

 

 

 

 

 

 

,,,,,, oo = REMOVABL.E
. ROCF SLABS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COOLING CHANNEL" N
DOUBLE CONTAINMENT

tjaqw-CELL HEATER THIMBLE

 

 

0 2 4 6 810
i
INCHES

Fig. 5.7. Reactor Cell Construction — Maintenance and Heater Access Details.

and the 3-in. plate is exhausted by a vacuum pump
which discharges into the cell atmosphere; this
“pumped-back’’ system thus provides opportunity
for continuous monitoring of the double contain-
ment system.

The inside surface of the sealing membrane is
covered with a 6-in.-thick blanket-type thermal
insulation protected on the inside by a 1/1 cin.
skin of stainless steel. This skin also serves

as a radiant heat reflector to reduce heat flow
into the wall,

The fuel, blanket, and coolant salts will be
maintained above their liquidus temperatures by
operating the interior of the cells at temperatures
up to 1150°F. During normal operation the tem-
perature is self-sustaining, and, as mentioned
above, the problem is one of heat removal from
the wall. For warmup and low-power operation,
4

however, electric heaters are provided in thimbles
around the inner walls of the cells, as shown in
Fig. 5.7. The electric leads for the heaters are
brought out through sealed bushings in the thimble
caps.

The off-gas cells and the drain-tank cells have
double containment but do not need the thermal
shield to protect against the radiation flux. The
steam cells require only the thermal insulation
and cooling air, since the radiation levels will be
relatively low and double containment is not re-
quired in these spaces.

5.3 REACTOR
G. H. Llewellyn W. C. George
W. C. Stoddaxt W. Terry
H. L. Watts H. M. Poly

During the past teport period, the new data dis-
cussed in Sect. 6.1 became available on the di-
mensional changes that occur in graphite as a
result of neutron irradiation. Because of this
experimental evidence, we decided to redesign
the reactor even though there is optimism that a
more stable graphite will be developed within the
next few years. The MSBR cost and performance
characteristics continue to be attractive even
though penalized by designing on the basis of
the immediate technology. We also decided that -
the reactor should be designed in such a way that
major redesipgn or modification would not be re-
quired, to take advantage of a more stable graphite
when it becomes available.

Several new approaches were tried for the core
design, one of which was to put the fertile salt
in the flow passages through the core graphite
and to allow the fuel salt to move through the in-

71

terstices. This so-called “‘inside-out’’ design
could probably accommodate the dimensional
changes in the graphite, assuming that suitable
adjustments were also made in the fuel enrich-
ment. A major disadvantage, however, is that the
fuel salt would also penetrate into the interstices
of the radial blanket, a position in which it is ex-
posed to relatively low neutron flux and thus pro-
duces relatively litile power. Since the flow in
this area would also be somewhat indeterminant,
this design of the reactor was not pursued further.

Attempts to design a removable graphite core
for the reactor led to the conclusion that such an
arrangement would probably be impractical. One
major problem would be containment of the highly
radioactive fission products associated with re-
moval of a bare reactor core. There would also
be the problem of assuring leak-tightness of a
large-diameter flanged opening which must be
sealed only by use of remotely operated tooling.
As previously mentioned, it wasg decided to re-
place the entire reactor vessel.

Selection of a ten-year life for the reactor, or
about 5 x 1027 nvt (greater than 50 kev) total max-
imum neutron dose for any point in the core, meant
that the power density would be reduced from the
40 kw/liter used in previous concepts to 20
kw/litex. This involved doubling the core volume
from 503 ft® to about 1040 ft3, and also required
that the reactor vessel size be increased cor-
respondingly. The factors entering into selection
of these conditions are given in Table 5.1, which
shows the effect of power density on the perform-
ance factors for the plant. At 20 kw/liter, it may
be noted that the fuel cycle cost is 0.5 mill/kwhr
and the yield is 4%/year. This appears to be the
most practical design point, although it is without
benefit of improved graphite. Tt should be pointed
out that the differences in capital costs shown in

Table 5.1. Performance Factors of MSBR as Function of Average Core Fower Density

 

Fisaile Inventory

 

Power Densitly Core Size {ft) Yield Fuel Cycle Cost Capital Cost Lide [
j et et et o : . ki for 1000 Mw

{kw/liter) . . (7 /year) (mills /kwhr) [$/kw (electrlcal)] (years) T )
Diameter Height : (electrical)l

80 6.3 8 5.6 0.44 117 2.5 880

40 8 10 5.0 0.46 119 5 1040

20 10 13.2 4.1 0.52 125 i0 1260

16 12 i8 2.7 0.62 132 20 1650

 
72

OHRNL-DWG 67-10643A -

e : 14§t 0in. 0D -

- 10 ft O in. DIAM CORE et

FUEL CELLS
420 AT 5% in. PITCH
53 in. HEX
x 13t 3in. LONG

BUANKET CELLS
252 AT 53 DIAM
(~15 in. TOTAL )

REFLECTOR
GRAPHITE
{~6in. TOTAL)

COOLING SKIRT

BLANKET PLENUM

  
 
 
 
 
 

PRESSURE
VESSEL

CONTROL
ROD

HOLD-DOWN PLATE

FUEL SPACER SHEETS

BLANKET CELL SPACER TIES

Fig. 5.8. 250 Mw (Electrical) Reactor — Plan. 420 SS;é-in. cells.

Table 5.1 do not take into account the cost of re-
placing the graphite moderator at intervals deter-
mined by radiation damage to the graphite. The
frequency of replacement and the contribution to
the power cost are greater at the higher power
densities, and this effect essentially offsets the
difference in initial cost.

Plan and elevation views of the reactor are
shown in Figs. 5.8 and 5.9. The design param-
eters are given in Table 5.2,

There are 420 two-pass fuel flow channels in
the reactor core graphite. As shown in Fig. 5.10,

each of these channels is formed from two graphite
extrusions. The outer, and longer, of the pieces
is hexagonally shaped, 53/3 in. across the flats,
and has a 2 23/3 ,-in.-diam hole in the center. The
total length of the extrusion is about 14% ft. The
inner piece is a cylinder, 2% in. OD x 1% in. 1D,
and fits inside the hole of the hexagonal extrusion.
The fuel salt enters the reactor through the inlet
plenum at the bottom of the core and flows upward
through the annulus between the inner and outer
graphite extrusions to the top of the reactor, where
it turns and flows through six - by 17%-in. slots
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

73
DRNL-DWG 67-10538A
- e {4 Oin DIAM
(D_ /CONTROL ROD
TS Bt e - s e
T A e
,r»-‘c’:j’.j_’jf - / Lo -
e TRy
| A7 L CORE 10 [t Oin. DIAM x 43 3 in. HIGH \*qu‘
HOLE-DOW, 1 / 420 FUEL CELLS AT 5¥8 HEX N’
- TO
PU‘TE\H S : . BLANKET HEAT
| — EXCHANGER -~
]_._ - oy /
T 3
19 ft Bin,
REACIOR : ’ |
ALIOR ‘ 13t 2in.
VESSEL - enm| ‘ | CORE
D |
| i
1
REFLECTOR ke
GRAPHITE :}\
niE R -
]
BLANKET %<
CELLS — {r»_;. 2 ft 2in.

 

 

 

 

 

 

 

 

 

   
 

 

 

 

FUEL SALT
PLENUMS—"

-~ SUPPQRT LEGS
{8)

Fig. 5.9. 250 Mw (Electrical) Reactor — Elevation, 420 5?78-in_ cells.
3in. 3in. DIAM e

3Y5in. DIAM

1ft 3in.

e T,

e ———— o

 

13 £ 3in.
REACTOR !

5.3750n. 5

223/321&![) ‘fi \
N

 

10in.

 

 

[

3in. MIN

 

 

 

 

Fig. 5.10. Fue! Cell Eiement.

ORNL-DWG 87-10646A

Y5 in. 1IDx 2% in. 0D

SiX Yo —x1¥-in. SLOTS

2 ¥in. DIAM

1/2“’1.

GRAPHITE TO
METAL BRAZE

17gin. ODx 43/4in. 1D

1Y in. 0DxtYgin 1D

FUEL INLET PLENUM

FUEL QUTILLET PLENUM
-

r .

Tahle 5.2. Reactor Specifications

 

Average core power density, kw/liter

Power, Mw

Number required for 1000 Mw (electrical)

Vessel diameter, ft
Vessel height, ft
Core diameter, ft
Core height, ft
Core volume, £t 3
Fraction of fuel in core
Fraction of blanket in core
FFraction of graphite in core
Blanket thickness, ft
Fraction of salt in blanket
Breeding ratio
Fuel yield, %/year
Fuel cycle cost, mills /kwhr
Fissile inventory, kg
Fertile inventory, kg
Specific power, Mw (thermal)/kg
Number of core zlements
Velocity of fuel in core, fps
Average flux >(0.82 Mev
Fuel Volume,‘ft3

Reactor core

Plenums

Entrance nipples

Heat exchangers and piping

Processing

Total

Peak/average flux ratio

20
556

4

14

22

10
13.2
1040
0.134
0.064
0.802
1.25
0.58
1.06
4.1
0.52
314
54,000
1.8
420
4.8
3.33 % 1013

135
37
13
160
6

355
™2

 

to the inside of the cylindrical graphite extrusion
and returns to the bottom of the reactor and the
outlet plenum. The average velocity of the fuel
in the core is about 4.8 fps as it is heated from
1000°F to 1300°F. The effective height of the
core is approximately 137 ft, and the total length
of the two-pass flow channel, plenum to plenum,

is about 27 ft.

The hexagonal graphite pieces have a cylindrical
portion about 12 in. long turned at the bottom to
which a Hastelloy N nipple, 17/8 in. OD by 1/16 in.
in wall thickness, is brazed. The other ends of
these nipples are welded to discharge openings
in the upper plate of the inlet fuel plenum. Inner

Hastelloy N nipples, 1Y

a4

in. OD by ?

in. in wall

thickness, are welded to the outlet plenum tube
sheet and have an enlarged upper end which fits

snugly, but is not brazed, into the inner 1 1/2—in.—ID
hole in the inner graphite cylinder of the fuel ele-
ment. The lower end of the core assembly is thus
fixed in place by attachment to the plenums, but
the upper end is free to expand or contract in the
vertical direction.

The graphite fuel pieces extend 15 in. above the
end of the fuel flow passages in order to serve as
the top axial reflector for the core. The top 3 in.
of each fuel element is turned to a smaller diam-
eter to establish a shoulder, as shown in Fig.
5.10. Triangular stampings of Hastelloy N sheet
are slipped down to this shoulder and engage
three of the elements, as shown in Fig. 5.8. These
stampings are interleaved to maintain the radial
spacing of the fuel channels, yvet eliminate the
need for a large-diameter upper diaphragm drilled
to close tolerances. The center six fuel channels
engage a ring which is attached through six ribs
to the vessel itself, thus stabilizing the entire
assembly.

Immediately outside the core region of the reactor
are graphite tubes around and through which the
fertile salt of the radial blanket is circulated.
These tubes displace the more expensive fertile
salt and also, by scattering the neutrons, promote
more effective capture by the thorium atoms in the
blanket. The ratio of fertile fraction to graphite
is about 58% in this region, as determined by the
code used for optimization of the reactor design.
The graphite tubes are slipped over short nipples
extending from a mounting plate at the bottom of
the reactor, as shown in Fig. 5.9. The tubes are
radially positioned at the top by overlapping con-
nectors in much the same manner as the fuel ele-
ments,

Solid cylinders of graphite, 5 in. in diameter,
are arranged on the outer circumference of the re-
actor to serve as a reflector. A can of 1/‘#-in. wall
thickness surrounds the reflector graphite and
serves to direct the entering fertile salt down the
inside wall of the vessel and to the bottom of the
core. The fertile salt stream then divides; part
of it moves upward through the interstices between
the fuel elements, while the major portion flows
through the graphite tubes in the blanket region.

It may be noted that the fuel channels them-
selves provide sufficient graphite for moderation
of the reactor and for the top reflector without use
of any special shapes or pieces, as was required
in earlier MSBR concepts. All the graphite con-
sists of extrusions which require little in the way
of machining or close tolerances. These design
improvements were made possible by relaxing the
restriction that the wall be thin enough to permit
reduction of the permeability to the 10~ 7-cm?/sec
range by impregnation. This requirement was
eliminated because, as discussed in Part 5, there
appears to be more hope for obtaining very low
permeability through a surface sealing technique,
which would not be limited by wall thickness, than
through impregnation, which requires thin sections.

The fuel enters and leaves the reactor through
separate pipes rather than by the concentric pipe
arrangement employed in the previous concepts.
The new design eliminates the need for the inner
slip joint and also relieves the thermal stress

problems that existed in certain temperature ranges.

The thermal expansion loops now shown in the
salt piping add to the fuel salt volume in the sys-
tem, but this is of less significance thau it would
have been in the 40-w/cm? reactor, which had only
one-half the core volume.

5.4 FUEL HEAT EXCHANGER

T. W. Pickel W. Terry

The heat exchanger for transfering heat from the
fuel salt to the coolant salt remains essentially
unchanged since the last report. The exchanger
is shown in Fig. 5.11, and the design data are
given in Table 5.3. As mentioned previously,
there has been some change in the salt headers.
Design of the gas sparging system has not been
completed.

5.5 BLANKET HEAT EXCHANGER

T. W. Pickel W. Terry

No major changes have been made in the heat
exchanger which transfers heat from the fertile,
or blanket, salt to the coolant salt. The exchanger
is shown in Fig. 5.12, and the pertinent data are
listed in Table 5.4. The blanket heat exchanger
has been lowered in the cell so that the pump
drive shaft will be the same length as the fuel
pump drive shaft and thus allow interchangeability
of parts. The coolant salt piping has also been
modified to replace the concentric piping formerly
used.

76

Table 5.3. Fuel Heat Exchanger Specifications

 

Number required per reactor module
Rate of heat transfer, Mw
Rate of heat transfer, Btu/hr
Shell side
Hot fluid or cold fluid

Entrance temperature, °F

Exit temperature, °F

Entrance pressure, psi

Ex1it pressure, psi

AP across exchanger, psi

Mass flow rate, 1b/hr
Tube side

Hot fluid or cold fluid

Entrance temperature, °F
Exit temperature, Oy
Entrance pressure, psi
Exit pressure, psi
AP across exchanger, psi
Mass flow rate, lb/hr

Tube material

Tube OD, in.

Tube thickness, in.

Tube
tube sheet, ft

Inner annulus

length, tube sheet to

Outer annulus
Shell material
Shell thickness, in.
Shell ID, in.
Tube sheet material
Tube shect thickness, in.
Top outer annulus
Top inner annulus
Floating head
Number of tubes
Inner annulus
Outer annulus
Ditch of inner ananulus tubes, in.
Radial
Circumferential

Pitch of outer annulus tubes, in.

Type of baffle
Number of baffles
Inner annulus

Outer annulus

1
520
1.80 x 10°

Cold
(coolant salt)
850
1110
198
164
34
1.68 x 107

Hot
(fuel salt)
1300
1000
146
50
96
1.09 % 107
Hastelloy N
0.375
0.035

15.3

16.7
Hastelloy N
1

67
Hastelloy N

1.5
2.5 _
3.5 ]

4347
3794

0.600

0.673

0.625,
triangular

Doughnut

 
CRNL:-DWG 67 -10642A

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

= — &
i Il =
= > =
L8k
w - &
~ = Z 5 o w8 o o
W 3 < 4 7z = a4 m G C
@ = = £ = 5 O n - .
= = = 5 c »ng
..Am - O o = M = 1 M ] = o=
= . -
_ =z o o D ox g © b o 9L @Ff o
W 2 = C Weoow O » m 2 3 o
) 7 £ 2 & 4 29z o4 H o= :
= o = & = 8 6w b =z : r < < <
2 b C == L O bl ooy Lo ~
= w = — r I 2 - oo il = Ll
= o n © ~ I JoLw Z ¥ < om £« o
_ f ' : i : < | ' :
| W 1 ,,‘hmi\ll, . "
: i e i _ i
: ! i[\\ifl(}.n fi ! :
! N
i "
0.
: Z
_ <
¢ =
e 35 o
./!l - e / m\.\»w/ :
““““ =TT
Vi =
L !
) e 7 a
/ ” fl Kl ooy
.,‘T " ..«\VM. .fl =
W m_ : . 7 ,
b i : / A
N | L
W v 0 / P
) , al / Lo
W\ Z| T
//nfis = < g 4 -
: : & A
; : (&
i : [
j : o
: : %
” : & : ] . fi =
| _ N < : = = s £ L
; - ” 0 : 5 ) ; = = e
! * — - —— —r — - - ./._| —— fi +— - - -
| e : - - - = =
| o ” b - 0 o -
-

 

 

 

 

23 1 7Yin
]
29ft 3'%in

DRAIN LINE

2-in.

Fuel Heat Exchanger and Pump. 250 Mw (electrical) unit.

Fig. 5.11,
30ft 6¥4in.
|

 

|
GAS CONN |
i
|
|
|
|
|

19ft Qin.

GAS SKIRT-

4t 8in. OD |

COOLANT 20in. NPS

 

Fig. 5.12. Blonket Heat Exchanger and Pump. 250 Mw (electrical)

\
|

4f+ Qin.

8in. NPS
' }
i
|

2ftCin.

14§t 1%in.

|
211 0in. D|f-\M--n~- . ]

78

N

 

R . o ""I T
— _
. '

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2Tt 9% in.

‘:;1 i
I
A
Lo
0l :
I i
] ! 1“\; it
- |\ [
. H
\ ] il
B : { r+fl;1‘
w al
0L i
...... L -
; S !

 

 

o
e T
g oS

e s e

 

 

 

R by

V(7 rain NeS L
! \

S . i

ORNL-DWG 67-406294A

- MOTOR FLANGE

SPCOL

CPERATING LEVEL

MOLTEN SALT BEARING

SLANKET PROCESSING
-DRAIN LINE

BLANKET TO
REACTOR 8in. NPS

FLANGE

IMPELLER

8ft 4in.

-- QUTER TUBES

822 AT ¥gin. OD

INNER TUBES
834 AT ¥gin. OD

unit.
Table 5.4. Blanket Heat Exchanger Specifications

 

Number regquired
Rate of heat transfer, Mw
Rate of heat transfer, Btu/hr
Shell side

Hot fluid or cold fluid

Entrance temperature, °F

Exit temperature, °F

Entrance pressure,® psi

Exit pressure,? psi

AP across exchanger,b psi

Mass flow rate, ib/hr
Tuhbe =side

Hot fluid or cold fluid

Entrance temperature, °F
Exit temperature, °p
Entrance pressure,? psi
Exit pressure,?® psi
AP across exchanger,b psi
Mass flow rate, lb/hr
Velocity, fps

Tube material

Tube OD, in.

Tube thickness, in.

Tube length, tube sheet to
tube sheet, ft

Shell material

Shell thickness, in.

Shell ID, in.

Tube sheet material

Tube sheet thickness, in.

Number of tubes
Inner annulus
Outer annulus J

Pitch of tubes, in.

Total heat transfer area, £t ?
Basis for area calculation
Type of baffle

Number of baffles

Baffle spacing, in.

Disk OD, in.

Doughnut ID, in.

Cverall heat transfer coef-
ficient, U, Btu el a2

4
27.8
9.47 % 107

Cold
(coolant salt)
1110
1125
138
129
15
1.68 x 107

Hot
(blanket salt)

1250

1150

111

20

91

4.3 % 10°

10.5

Hastelloy N

0.375

0.035

8.3

Hastelloy N
0.50

55

Hastelloy N
1

834

822

0.8125,
triangular

1318

Tube OD

Disk and
doughnut

4

19.8

33.0

31.8

1030

 

“Includes pressure due to gravity head.

PPressure loss due to friction only.

79

5.6 FUEL DRAIN TANKS

H. L. Watts T. W. Pickel

Two 60-in.-diam by 20-ft-high dump tanks are
required for draining the fuel salt from each of the
reactor systems. These are shown in Fig. 5.13.
One drain tank is connected to the 5-in.-diam
overflow connection on the sump tank of the fuel
circulating pump and receives all the salt from
the reactor when it is drained. The drain time
is estimated to be short, probably less than 10
sec, and this method could be used for emergency
shutdown if necessary. The other drain tank is
connected to the fuel heat exchanger and fills
through a siphon action due to the differences in
elevation.

The maximum heat to be removed in each fuel
drain tank is 12 Mw (thermal). The two tanks are
cooled by a total of 422,000 1b/hr of steam taken
from the high-pressure turbine-generator exhaust
at about 550 psia. The steam is heated from 650°F
to about 1000°F in vertical thimbles spaced on 3-
in. centers in each dump tank. There are 271 of

these thimbles per tank, extending to within 1 in.
of the bottom. The reentrant-type steam thimbles
are inserted in 2-in.-OD INOR-8 thimbles which
contact the fuel salt, the 0.025-in. radial clearance
between the two thimbles being filled with stag-
nant inert fuel salt to setve as a heat transfer
medium. The double thimble arrangement provides
the necessary double barrier between the steam
and the enriched fuel salt. Steam leaves the
thimbles at about 72 fps and enters the steam
chest at the top of the dump tank.

5.7 BLANKET AND COOLANT SALT
DRAIN TANKS

H. L. Watts W. C. George

The drain tanks for the blanket salt and for the
secondary coolant salt are essentially simple
storage reservoirs. Cooling systems such as
those used in the fuel salt tanks are not required.
The details of the designs for these tanks have
not been completed. The location of the tanks
in the cells is shown in Fig. 5.3.
80

ORNL-DWG 67 - 10644 A

.- STEAM {*8-in. PIPE)
e 650 °F - 570 psig

   
  
 
    

271 TUBES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2% A PITCH
4 . -11/2 in. i
44 DISH R TYP
ALL HEADS
2 in
L
; |
’ !
1
41 in.
_BUFFER
"]
. GAS
0072 in o _FUEL
A15-in, PIPE) ‘
o ey
W 24 f1-9'% in.
| ‘
Ff_fij,..cvj
AN ] TTUPLENUM
\}\ ‘z;
R T RISER SKIRT
N
.
~~
HEAT SHIELD
19 -6
- TUBE 6in
SUPPORT
¢.042 n. - =
- --jo.O?z in
A1
'y
. , _ e |
\5(_’_ | S R - 1 = i
| e /, b= / ) !
N A TO FREEZE kfi_fl/ Y
| A VALVE
[t
b
\ ; T~ FUEL DRAIN
|

;
2 in. 0D x Q035 in.

L 1 % in. OD x 0.048in.

1'% 1. OD x 0.025in

Fig. 5.13. Fue! Drain Tank with Decay-Heat Removal System Details.
81

58 STEAM GENERATOR-SUPERHEATER The initial concepts have changed little, however,
AND REHEATER and the design work has consisted primarily of
detailing tube sheets, supports, headers, etc.
C. E. Bettis T. W. Pickel Details of this equipment will be covered in sub-
Some work was continued on the design of the sequent reports,

steam generator-superheater and on the reheater.
6. Reactor Physics

A. M. Perry

6.1 MSBR PHYSICS ANALYSIS

O. L. Smith

Optimization of Reactor Parameters

H. T. Kerr

The two principal indices by which the perfor-
mance of the Molten-Salt Breeder Reactor is cus-
tomarily evaluated are the cost of power and the
annual fuel yield, that is, the annual fractional in-
crease in the inventory of fissionable material.
These are used as figures of merit in assessing the
influence of various design parameters or the effect
of design changes that may be contemplated, and,
in fact, we customarily combine them into a com-
posite figure of merit,

F-y+100(C 1+ X)77,

in which y is the annual fuel yield in percent per
year, C is that part of the power cost which de-
pends on any of the parameters considered, and X
is an adjustable constant, having no physical sig-
nificance, whose value merely determines the rela-
tive sensitivity of F to y and C. Since a large num-
ber of reactor parameters are usually involved, we
make use of an automatic search procedure, carried
out on an IBM 7090 computer, which finds that com-
bination of the variable design parameters that max-
imizes the figure of merit, F, subject to whatever
constraints may be imposed by the fixed values of
other design parameters not allowed to vary. This
procedure, called OPTIMERC, incorporates a mul-
tigroup diffusion calculation (synthesizing a two-
space-dimensional description of the flux by al-
ternating one-dimensional flux calculations), a
determination of the fissile, fertile, and fission
product concentrations consistent with the proc-

82

essing rates of the fuel and fertile salt streams,

and a method of steepest gradients for optimizing -
the values of the variabies. By choosing different
values for the constant X in the figure of merit, F,
we can generate a curve showing the minimum cost
associated with any attainable value of the fuel
yield, and by carrying out the optimization proce-
dure for different, successive fixed values of se-
lected design parameters, we can generate families
of such curves of C vs y. (In OPTIMERC any of
some 20 parameters may either be assigned fixed
values or be allowed to vary within specified limits
subject to the optimization procedure.)}

One of the design parameters which has a sig-
nificant influence on both yield and power cost is
the power density in the core. (Actually, the core
dimensions for a given total power are the param-
eters used.) The performance of the reactor is
better at high power densities. At the same time,
the useful life of the graphite moderator, which is
dependent on the total exposure to fast neutrons,
is inversely proportional to the power density (see
next section). It is necessary, therefore, to de-
termine the effect of power density on performance
with considerable care.

In Fig. 6.1 the fuel-cycle cost is used because
it contains, in fact, most of the variation of power
cost with the parameters being varied. It may be
seen from Fig. 6.1 that a reduction in power density
from 80 to 20 w/cm? involves a fuel-cycle cost
penalty of about 0.1 mill/kwhr (electrical) and a
reduction in annual fuel yield of perhaps 1.5%.
There is, of course, also an increase in capital
cost (cf. Chap. 5), but this is essentially offset by
a reduction in the cost of replacing the graphite
and reactor vessel at intervals determined by ra-
diation damage to the graphite. The combined
penalty for having to replace the graphite, com-
pared with a high-power-density core not requiring
replacement, is about 0.2 mill/kwhr (electrical),
1.00

o
o
[

20ST [ milis/kwhr(eiec‘rricali]

~
u

L CYCLE

FUE

075 o

025 [~

 

83

ORNL--DWG 67—448B06

 

 

 

 

 

 

 

0 2.5 50

 

 

 

 

 

 

 

 

 

 

 

 

whether this comprises higher capital cost plus
lower teplacement cost at low power densities, or
lower capital cost plus higher replacement cost at
higher power densities. Figures 6.2 and 6.3

show the variation of other selected parameters
both with power density and with the adjustable
constant X.

It is apparent from these results that the useful
life of the graphite is not increased by reducing
core power density without some sacrifice in other
aspects of reactor performance. The reduction in
yield and the increase in cost are quite modest for
a reduction of power density from 80 to 40 w/cm?,
but they become increasingly more significant for
each further factor-of-two reduction in power den-
sity. Nonetheless, it appears that with an average
power density as low as 20 w/cm>, the MSBR can
still achieve an annual fuel yield of 3.5 to 4% and
a fuel-cycle cost of less than 0.5 mill/kwhr (elec-
trical).

Fig. 6.1. MSBR Fuel-Cycle Cost vs Annual Fuel

 

 

 

 

 

 

 

 

 

 

) 7.5
FUEL YIELD (% per year) Yield.
ORNL-DWG 67141807
0.20 } I [ [
SALT VOLUME FRACTIONS 3
|
i
Q4B Joooe
042
2/
O
0.08 L
10m ke __‘_._EEETLLE
6 = T ee— e 0
21 TE R0
o~ 0
0.04
3 _

 

 

 

SPECIFIC POWER [Mw({thermal}/kg ]

 

 

 

 

 

 

o 20 40 60
POWER DENSITY (w/cm3 )

 

 

/U RATIO (X107}

 

 

 

 

 

 

 

 

0 20 40 50 80
POWER DENSITY (w/cm3)

Fig. 6.2. Variation of MSBR Parameters with Average Core Power Density. Numbers attached to curves are

values of the adjustable constant X.
ORNL-DWG 67-141808

RADIAL BLANKET THICKNESS

THICKNESS (ft)

 

 

 

 

 

0.08 [ e e e
NEUTRON I OSSES TO Li Bef
e e o [10
0.06 4 40
0.08

 

| T r
NEUTRON LOSSES TO MODERATOR

0.06

0.04 +—

NEUTRONS ABSORBED /7 SCURCE NEUTRONS

 

0.02

 

 

2.25

 

 

 

 

 

2.24
w
=
2.23
2.22
400 e b
“.
200 TR e, ‘—J
ERNAL |

 

™

 

 

 

 

 

20
POWER DENSITY (w/cm3)

40 60 80
Fig. 6.3. Variation of MSBR Parameters with Average
Core Power Density. Numbers attached to curves are

values of the adjustable constant X.

84

Useful Life of Moderator Graphite

A. M. Peny

Information currently used in the MSR project re-
garding the dependence of graphite dimensional
changes on fast neutron dose is derived primarily
from experiments carried out in the Dounreay Fast
Reactor (DFR).

A curve of volume change vs fast neutron dose
for a nearly isotropic graphite at temperatures in
the range 550 to 600°C is shown in Fig. 6.4, which
is taken from the paper of Henson, Perks, and Sim-

mons.! (The neutron dose in Fig. 6.4 is expressed

R W. Henson, A. J. Perks, and J. H. W. Simmons,
Lattice Paramefer and Dzmensronal Changes in Graphite
Irradiated Between 300 and 1350°C, AERE-R5489, to be
published in the proceedings of the Eighth Carbon Con-
ference.

ORNL
44 _fiv.,.....................”‘\‘........................,....... T

s

r ’
400-440°C B
550-600°C

8 , [

DWG 6? 61809

+
10

 

0o

{Ye)

CHANGE

,.
-
—

 

VOLUM

&

 

 

 

3
neuhrons,/cm2 )

0 1
EQUIVALENT PLUTO DOSE

HO‘ZZ

Fig. 6.4. Yolume Changes of Near-lsotropic Graphite
Resuvlting from Meutron lrradiation. See text for dose in

terms of MSBR flux.
85

ORNL OWG 67-11810

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

’O Y - L et b
jg’wgmrg)dg
RAEG = wggm—
[ eI dE
ED
el
2
.‘é
S i = Hy,0 SPECTRUM
< B 2=0p,0 specTRUM | LT L 1 L v T
g,'f 3 = CARBON SPECTRUM //
2 4 = FAST SPECTRUM e /
w3 A
7 L .V o ~23% ;/.//
" l S /
u A
<ol e LT {/ ,,,,,,,,,,, (o
Q& ~3% /"/
/ L~ >
‘ %r/“"f 2
1
A @
= et T2 S
B | it - e S e e O b gl b
o
L—-= ®
.«-'/ O
|-~ T
4
10? 2 5 10° 2 5 10°
EO (eV)

Fig. 6.5. Fost Flux as ¢ Measure of Radiation Damaoge.

in terms of an equivalent Pluto dose; the total DFR
dose, that is,

frfmqsuz, f) dE dt |
0 0

is 2.16 times the equivalent Pluto dose.) From an
inspection of all the available data, we have con-
cluded that a dose of about 2.5 x 10?2 neutrons/
cm? (equivalent Pluto dose) could be sustained
without any significant deterioration of the physi-
cal properties of the graphite, and this has been
adopted as an allowable dose, pending further de-
tailed consideration of mechanical design problems
that might be associated with dimensional changes
in the graphite.

In order to interpret these experiments to obtain
predictions of graphite damage vs time in the
MSBR, it is necessary to take into account the dif-
ference in neutron spectra in the two reactors.
This, in turn, requires assumptions regarding the

effectiveness of neuntrons of different energies for
producing the observable effects with which one

is concerned. At present the best approach avail-
able is to base one’s estimates of neutron damage
effectiveness on the theoretical calculations of
graphite lattice displacements vs carbon recoil
energy carried out by Thompson and Wright.? The
Thompson and Wright ““damage function’’ is inte-
grated over the distribution of carbon recoil en-
ergies resulting from the scattering of a neutron

of a given energy, and the result is then multiplied
by the energy-dependent scattering cross section
and integrated over the neutron spectrum in the re-
actor. Tests of the model have been made by
Thompson and Wright by calculating the rate of
electrical resistivity change in graphite relative to
the 5®Ni(n,p)°8Co reaction, in different reactor

 

’M. W. Thompson and S. B. Wright, J. Nucl. Mater. 16,
146.-54 (19635).
86

DRNL-OWG 67-{1814

NEUTRCON ENERGY (MeV}

 

5108 6 4 3 2 1.5 1.0
; f \ [ [ [ 1 ’ I
!

1= H,0 SPECTRUM

2=D,0 SPECTRUM

3=CARBON SPECTRUM

4=FAST SPECTRUM ({50% Na, 50% U-METAL, 20% U, 80% ->°U)
1.0 S S

¢ (U} Larbitrary units)

 

 

0.8 0.6 G0 0.08

I ] ! \‘ f

 

 

 

 

 

   

 

 

 

 

 

LETHARGY U

Fig. 6.6. Meutron Flux Per Unit Lethargy vs Lethargy. Normalized for equal damage in graphite.

spectra, and comparing it with experimental de-
terminations of the same quantities. The results
indicate that the model is useful at least for pre-
dicting relative damage rates in different spectra.

A useful simplification arises from the observa-
tion that the damage per unit time is closely pro-
portional to the total neutron flux above some
energy £, where Eo has the same value for widely
different reactor spectra. We have reconfirmed this
observation, to our own satisfaction, by comparing
the (calculated) damage per unit flux above energy
E  as a function of E  for spectra appropriate to
three different moderators (HQO, D,0, and C) and
for a ““typical’’ fast reactor spectrum. The results
plotted in FFig. 6.5, show that the flux above about
50 kev is a reliable indication of the relative dam-
age rate in graphite for quite different spectra.

7

Figure 6.0 shows the spectra for which these re-
sults were derived. The equivalence between
MSBR and DFR experiments is simply found by
equating the doses due to neutrons above 50 kev
in the two reactors. We have not yet calculated
the DFR spectrum explicitly, but we expect it to

be similar to the ““fast reactor’’ spectrum of Fig.

6.6, in which 94% of the total flux lies above 50
kev. Since the damage flux in the MSBR is es-
sentially proportional to the local power density,
we postulate that the useful life of the graphite is
governed by the maximum power density rather than -
by the average, and thus depends on the degree of
power flattening that can be achieved (see next -
section). In the MSBR the average flux above 50 -
kev is about 0.94 x 10'* neutrons cm™? sec™! at
a power density of 20 w/cm?. From the DFR ir-
radiations it has been concluded that a dose of
5.1 x 1022 nvt (> 50 kev) can be tolerated. The
lifetime of the graphite is then easily computed;
this useful life is shown in Table 6.1 for an as-
sumed plant factor of 0.8 and for various com-
binations of average power density and peak-to-
average power-density ratio.
It must be acknowledged that in applying the
results of DFR experiments to the MSBR, there
remain some uncertainties, including the pos-
sibility of an appreciable dependence of the dam-
age on the rate at which the dose is accumulated
4

Table 6.1. Useful Life of MSBR Graphite

 

Average Power

 

Density r /P ) Life
max av
(w/cm®) (years)
40 2.0 5.4
40 1.5 7.2
20 2.0 10.8
20 1.5 14.4

 

as well as on the total dose. The dose rate in the
DFR was approximately ten times greater than that
expected in the MSBR, and if thete is a significant
dose-tate effect, the life of the graphite in an MSBR
might be rather longer than shown in Table 6.1.

Flux Flattening

O. L. Smith H. T. Kerr

Because the useful life of the graphite moderator
in the MSBR depends on the maximum value of the
damage flux rather than on its average value in the
core, there is obviously an incentive to reduce the
maximum-to-average flux ratio as much as possible,
provided that this can be accomplished without se-
rious penalty to other aspects of the reactor per-
formance. In addition, there is an incentive to
make the temperature rise in parallel fuel passages
through the core as nearly uniform as possible, or
at least to minimize the maximum deviation of fuel
outlet temperature from the average value. Since
the damage flux (in effect, the total neutron flux
above 50 kev) is essentially proportional to the
fission density per unit of core volume, the first
incentive requires an attempt to flatten the power
density per unit core volume throughout the core,
that is, in both radial and axial directions in a
cylindrical cote. Since the fuel moves through the
core only in the axial direction, the second in-
centive requires an attempt to flatten, in the radial
direction, the power density per unit volume of
fuel. Both objectives can be accomplished by
maintaining a uniform volume fraction of fuel salt
throughout the core and by flattening the power
density distribution in both directions to the
greatest extent possible.

87

The general approach taken to flattening the
power distribution is the classical one of pro-
viding a central core zone with k = 1, that is,
one which is neither a net producer nor a net ab-
sotber of neutrons, surrounded by a ““buckled”’
zone whose surplus neutron production just com-
pensates for the neutron leakage through the core
boundary. Since the fuel salt volume fraction is
to be kept uniform throughout the core, and since
the concentrations of both the fuel and the fertile
salt streams are uniform throughont their respec-
tive circuits, the principal remaining parameter
that can he varied with position in the core to
achieve the desired effects is the fertile salt
volume fraction. The problem then reduces to
finding the value of the fertile salt volume frac-
tion that gives k, = 1 for the central, flattened
zone, with fixed values of the other parameters,
and finding the volume fraction of the fertile salt
in the buckled zone that makes the reactor critical
for different sizes of the flattened zone. As the
fraction of the core volume occupied by the flat-
tened zone is increased, the fertile salt fraction
in the buckled zone must be decreased, and the
peak-to-average power density ratio decreases
toward unity. The largest flattened zone and the
smallest power density ratio are achieved when
the fertile material is removed entirely from the
outer core zone. Increasing the fuel salt concen-
tration or its volume fraction (with an appropriate
adjustment of the fertile salt volume fraction in the
flattened zone) would permit a still larger flattened
zone and smaller Pmax/Pavg, but would be ex-
pected to compromise the reactor performance by
increasing the fuel inventory, at least if carried
too far,

There are obviously many possible combina-
tions of parameters to consider, It is not a priori
obvious, for example, whether the flattened zone
should have the same height-to-diameter ratio as
the entire core, or whether the axial buckled zones
should have the same composition as the radial
buckled zone. While we have by no means carried
the investigation of these questions as far as we
need to, we have gone far enough to recognize
several important aspects of the problem.

First, by flattening the power to various degrees
in the radial direction only and performing fuel-
cycle and economic calculations for each of these
cases, we find that the tadial power distribution
can be markedly flattened with very little effect
either on fuel cost or on annual fuel yield, our
chief indicators of performance. That is, the
radial peak-to-average power density ratio, which
is about 2.0 for the uniform core (which is sur-
rounded, of course, by a heavily absorbing blanket
region and hence behaves essentially as if it were
unreflected), can be reduced to 1.25 or less with
changes in fuel cost and yield of less than 0.02
mill/kwhr (electrical) and 0.2% per year respec-
tively. The enhanced neutron leakage from the
core, which results from the power flattening, is
taken up by the fertile blanket and does not rep-
resent a loss in breeding performance.

Second, attempts at power flattening in two
dimensions have shown that the power distribu-
tion is very sensitive to details of composition
and placement of the flattened zone. Small dif-
ferences in upper and lower blanket composition,
which are of no consequence in the case of the
uniform core, produce pronounced axial asymmetry
of the power distribution if too much axial flat-
tening 1s attempted.
radial buckled zones may interact through the

In addition, the axial and

flattened zone, to some extent, giving a distri-
bution that is concave upward in one direction
and concave downward in the other, even though
the integrated neutron current over the entire
boundary of the central zone vanishes. In view
of these tendencies, it may be anticipated that a
flattened power distribution would be difficult to
maintain if graphite dimensional changes, result-
ing from exposure to fast neutrons, were allowed

to influence the salt volume fractions very strongly.

Consequently, some revisions in the details of the
core design are under consideration as a means of
reducing the sensitivity of the power distribution
to graphite dimensional changes.

Temperature Coefficients of Reactivity

O. L. Smith C. O. Thomas

In analyzing power transients in the Molten-Salt
Breeder Reactor -~ as indeed for most reactors —
one must be able to determine the reactivity ef-
fects of temperature changes in the individual com-
poneints of the core, for example, the fuel salt, the
fertile salt, and the graphite moderator. Since the
fuel is also the coolant, and since only small frac-
tions of the total heat are generated in the fertile
salt and in the moderator, one expects very much
smaller temperature changes in the latter compo-

83

ORNL-DOWG 67-11812

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 e ——— .
MODERATOR
A
e i
x i
&
|
(o) (b)
-—q | Ao
8 |
FUEL SALT OVERALL
4 AN f [
-
Q
X ol N L N _
x
60 -
74 L — u — _
(¢) (¢) !
- 8 e L e s —
800 300 1000 800 300 1000
7 {°K) 7 (°K)

Fig. 6.7. MSBR Multiplication Factor vs Temperature.

nents than in the fuel during a power transient.
Expansion of the fuel salt, which removes fuel
from the active core, is thus expected to be the
principal inherent mechanism for compensating
any reactivity additions to the MSBR.

We have accordingly calculated the magnitudes
of the temperature coefficients of reactivity sep-
arately for the fuel salt, the fertile salt, and the
graphite over the range of temperatures from 800
to 1000°K. The results of these calculations are
shiown in Fig. 6.7.

In Fig. 6.7a we show a curve of change in mul-
tiplication factor vs moderator temperature (with
6k arbitrarily set equal to zero at 900°K). Similar
curves of &k vs temperature for fuel and fertile
salts are shown in Figs. 6.7b and 6.7c, and the
combined effects are shown in Fig. 6.7d. These
curves are all nearly linear, the slopes being the
ternperature coefficients of reactivity. The mag-
nitudes of the coefficients at 900°K are shown in
Table 6.2.

The moderator coefficient comes almost entirely
from changes in the spectrum-averaged cross sec-
Table 6.2, Temperature Coefficients of Reactivity

 

 

Coefficient
1 dk
Component I (OK)——I
k dT
Moderator +1.66 X 1077
Fertile salt +2.05 x 1073
Fue! salt —~8.05 x 1077
Overall —~4.34 x 1077

 

tions. It is particularly worthy of note that the
moderator coefficient appears to be quite insen-
sitive to uncertaiaties in the energy dependence
of the 233U cross sections in the energy range
below 1 ev; that is, reasonable choices of cross

389

sections based on available experimental data
yield essentially the same coefficient.

The fertile-salt reactivity coefficient comprises
a strong positive component due to salt expansion
(and hence reduction in the number of fertile atoms
per unit core volume) and an appreciable negative
component due to temperature dependence of the
effective resonance-absorption cross sections, so
that the overall coefficient, though positive, is
less than half as large as that due to salt ex-
pansion alone.

The fuel salt coefficient comes mainly from ex-
pansion of the salt, which of course reduces the
average density of fuel in the core. Even if all
core components should undergo equal temperature
changes, the fuel-salt coefficient dominates; and
in transients in which the fuel temperature change
is far larger than that of the other components, the
fuel coefficient is even more controlling.
7. Systems and Components Development

Dunlap Scott

Work related to the Molten-Salt Breeder Reactor
was initiated during this period. Studies were
made of the problems related to the removal of
the noble gases from the circulating salt to help
identify the equipment and systems required to
keep the '?3Xe poison fraction and the fission
product afterheat to an acceptable level. Prepara-
tions were begun for operation of a small out-of-
pile loop in which a molten salt will be circulated
through a graphite fuel cell and for operation of an
isothermal MSRE-scale loop with sodium fluoro-
borate.

The major effort of the program at this time is
to help establish the feasibility of improved
concepts and to define problem areas. Since the
production of suitable and reliable salt pumps is
one of the longest lead-time items for molten-salt
reactors, a major emphasis is being placed on this
program,

Some of the work related to problems of the
MSBR but actually performed on the MSRE is
discussed in the MSRE section of this report.

The other work is described below.

7.1 NOBLE-GAS BEHAVIOR IN THE MSBR

R. J. Ked!

In the MSBR conceptual designs, the graphite
in the reactor core is unclad and in intimate
contact with fuel salt. Noble gases generated by
fission and any gaseous compounds can diffuse
from the salt into the porous structure of the
graphite, where they will serve as heat sources
and nuclear poisons.

A steady-state analytical model was developed
to compute the migration of noble gases to the

90

graphite and other sinks in the MSBR. The sink
terms considered are:

1. Decay.

2. Burnup — takes place in the graphite and in the
fuel salt passing through the core.

3. Migration to graphite — these gases ultimately
decay or are burned up.

4. Migration to circulating bubbles -- these gases
are stripped from the fuel loop to go to the
off-gas system.

Two source termms are considered: genertation
directly from fission, which is assumed to occur
only in the core, and generation from decay of the
precursor, which occurs throughout the fuel loop.
The model is based on conventional mass transfer
concepts. Some degree of success has been ex-
perienced with similar models developed for the
MSRE, for example:

1. Xenon-135 poison fraction calculations. The
steady-state model is developed in ORNI.-4069.
Results of the time-dependent form of the model
are summarized in ORNIL-TM-1796.

2. A model was developed to compute the con-
centrations of daughters of very short-lived
noble gases in graphite (ORNL-TM-1810).
Computed concentrations check very well with
measured values.

With this model, steady-state '?5Xe poisoning
calculations have been made for the MSBR [556
Mw (thermal) fueled with 233U and moderated with
unclad graphite] to show the influence of various
design parameters involved. The reactor con-
sidered here is essentially that described in
ORNL-3996 [P. R. Kasten e/ al., Design Studies
of 1000-Mw (e) Molten-Salt Breeder Reactors], with
specific design parameters as listed in Table 7.1.
re

91

Table 7.1. Design Parameters

 

 

Reactor power, Mw (thermal) 556
Fuel 233y
Fuel salt flow rate, ft3/-sec 25.0
Core diameter, ft 2
Core height, ft 10
Volume fuel salt in core, £t3 83
Volume fuel salt in heat exchanger, £t 2 83
Volume fuel salt in piping between core and heat exchanger, £t3 64
Fuel cell cross section
DOWNFLOW CHANNEL
3—7/3 in. HOLES
UPFLOW CHANNEL

Total graphite surface area exposed to salt, £t? 3630
Mass transfer coefficient to graphite — upflow, ft/hr 0.72
Mass transfer coefficient to graphite — downstream, ft/hr 0.66
Mean thermal flux, neutrons see ¥ em ™2 5.0 x 1014
Mean fast fiux, neutrons sec_1 cm ™2 7.6 x 1014
Thermal neutron cross section for 233U, bhams 253
Fast neutron cross section for 233U, bams 36.5
Total core volume — graphite and salt, it 503
233U concentration in core — homogenized, atoms barn ~ 1 em ™! 1.11 x 1073
Graphite void available to xenon, % 10
135%¢ parameters

Decay constant, 1/hr 7.53 % 1072

Generation direct from fission, % 0,32°

Generation from iodine decay, % 6.38°

Cross section for MSBR neutron spectrum, bams 9.94 x 16°

 

“The values for the yield of 135

Xe from the fission of 2330 are from an old

source and were used in the screening calculations. Recent values of 1.11% for
generation direct from fission and 6.16% for generation from iodine decay as reported
by C. B. Bigham et al. in Trans. Am. Nucl. Soc. 8(1), June 1665, will be used in the

future.

The xenon stripping mechanism consists in cir-
culating helium bubbles with the fuel salt and then
removing them from the system. These bubbles
are injected at the inlet to the heat exchanger in
the region of the pump. Xenon-135 migrates to the
bubbles by conventional mass transfer, and the
mass transfer coefficient controls the rate of
migration, The circulating bubbles are then
stripped from the salt by a pipeline gas separator
at the heat exchanger outlet. The heat exchanger,
then, is the xenon stripper region of the fuel loop.

The principal parameters to be discussed here
will be:

1. Diffusion coefficient of xenon in graphite.
2. Parameters associated with circulating bubbles.

(2) Mass transfer coefficient to the bubbles.
(b) The surface area of the bubbles.

3. The surface area of graphite exposed to salt in
the core.

In the plots that follow, the diffusion coefficient
of xenon in graphite at 1200°F with units of ft2/hr
92

ORNL-DOWG 6711813

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    
 

 

 

 

G e : e -
: i L
‘ L
=
L —
o o=
- - i - = z O
8 8 < g
Q g5 x
z2 o
S 5
S o &
7 L _ v E Ll &
- W
| o Wl a
- Wl o
o € g
E° of
b E
6 — s — . _ B — 0
62 £ %z ou
; wn T Lj @
” wl v B m a
e , . e 2 W @ 3
[ ‘ < = -
0 NO CIRCUL ATING BUBBLES D T 0w
c o C:D) w :LD S -
: S
- =
@ SO a
2l 4 L _ 4 . £ &5 33 '
n et [ 2 (:)) < L)
< o L& x = .
s L C 5
- L 2o g
- o S o =
L S 2 L. O
s o L le 82 o%Z
| oIy, <,
» WA w7
z @= T =
i Z aZ 5
ot Ll o W o
—— e = & 8 O L|1_J
ol e . . T%““"““'- LW E o E =z
n oo r S
< 20 20
b e e = " O [ ]
— 1.0 3000 0
VL - 10 3000 300
e 20 3000 300
40 3000 300
0oL e e SO, _
1072 1073 1074 105 10776 107
DIFFUSION COEFFICIENT OF Xe IN GRAPHITE AT 4200°F {Ha/hr)
~PERMEABILITY OF He IN GRAPHITE AT ROOM TEMPERATURE (cm2/sec) -
Fig. 7.1. Effect of Diffusion Coefficient in Graphite on 135%¢ Poison Fraction.

is used as a parameter. Numerically, this is ap-
proximately equa!l to the permeability of helium

in graphite at room temperature with units of
cm?/sec, if Knudsen flow prevails. Knudsen flow
should be the dominant flow character for perme-
abilities less than 107° cm?/sec. For perme-
abilities greater than 10~° cm?/sec, viscous flow
becomes important and this direct relationship
does not exist.

The gas circulated through the system is handled
in these calculations as two groups of bubbles.
The first group, refetred to as the
bubbles, is injected at the bubble generator and
removed with 100% efficiency by the gas separator.

once-through”’

“recirculated’’

The second group, referred to as the
bubbles, is injected at the bubble generator, com-
pletely bypasses the gas separator, and recircu-
lates through the system until the bubbles are
removed with 100% efficiency on their second pass
through the gas separator. In the accompanying
plots the bubble surface aiea is the quoted param-
eter. For proper orientation, note that 3000 ft? of
bubble surface area corresponds to an average void
fraction of 1% in the stripper region of the fuel loop
with bubbles 0.020 in. in diameter, and also cor-
responds to a gas flow rate of about 40 scim.
Figure 7.1 shows the '*%Xe poison fraction as a
function of the diffusion coefficient in graphite.
 

ORNL-DWG 6711814

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9 ................ e —— e pameamsemeemmmsem—————
%
&
B feees s s s s esesi A
w
—
CIFFUSION COFFFICIENT OF Xe IN GRAPHITE=10"5 ffz/hr éuj
O
o=
7 —_— Z
= Z
g2
& 1o
2 L
o &%
; e
6l - S — g EE
o n 9
o
3 x Q4
o 0 .
§‘ a. {'E -
= Q: v w
O o zu
s ——— = T X
© o~
e z ‘&
ia B
w - i
z E
D = o
o wr -
o 1Y m 2
3 _t D a
> m m H-
m =
S L0
o z __‘
= }— o
2 iz
o
I o
- :13 o
.l
wl =
U3
o 8 i
= o
£ e
L. L
[ o 7
< <T _t
wh u! <L
o oo
< .
s
w! w
O Q .4.
" w
x o ‘~2?
s BO
3000 0O
3000 300
3000 600
o Ll I..
O { 2 3 4 5 &

MASS TRANSFER COEFFICIENT TO CIRCULATING BUBBLE (ft/hr)

Fig. 7.2. Effect of Bubble Mass Transfer Coefficient on ]35Xe Poison Fractien.

The top line is for no xenon removal through cir-

culating bubbles, and the poison fraction approaches

that for a solid-fueled reactor. The other lines are
for various circulating bubble parameters as indi-
cated. From this figure it can be seen that the

poison fraction is not a strong function of the dif-

fusion coefficient over the range from 10~2 to 1078,

This is because the mass transfer coefficient from
salt to graphite is the controlling resistance for
migration of 1¥3Xe into the graphite. Since 135%e
in the graphite is the greatest contributor to the
total poison fraction, the parameters that control
its migration will, in turn, control the poison frac-
tion. For permeabilities <10~ ° the resistance of

the graphite starts becoming significant. The
mass transfer coefficients to graphite were com-
puted from the Dittus-Boelter equation as modified
by the heat-mass transfer analogy.

Figure 7.2 shows the effect of the mass transfer
coefficient to the bubble on the poison fraction.
This mass transfer coefficient is one of the least
well known and most significant of the parameters
involved. Available information indicates its
extreme values to be 0.7 and approximately 6
ft/hr. Values of 0.7 to 0.8 ft/hr were estimated
assuming that the bubbles behave as solid spheres
and have a fluid dynamic boundary layer. Values
near 3.5 ft/hr were estimated assuming that as the
 

ORNL-DWG &67-11815

-
11 3
& 10
T
o
<
o
U
Z
9]
}.__
<
x
=z 2.
s 10
“-<E
n
ul
=
L
O
G
bl
I
10? . .
1 10 100 1000 10000 -
TIME AFTER REACTOR SHUTDOWN (min)
Fig. 7.3. Afterheat in MSBR [556 Mw (thermal)l Graphite from Noble Gases and Their Daughters After 10 Years .

of Power Operation.

bubble rises, its interface is continually being
replaced by fresh fluid (penetration theory). Both
of these cases are for a bubble rising at its terminal
velocity in a stagnant fluid. There is very little
information in the literature concerning the effect
of fluid turbulence on the bubble mass transfer
coefficient. Nevertheless, from turbulence theory
it has been estimated that mass transfer coef-
ficients as high as 6 ft/hr could be realized under
MSBR conditions. The analyses that lead to this
number are generally optimistic in their assump-
tions.

The target '*5Xe poison fraction for the MSBR
is 0.5%. From Fig. 7.2 it can be seen that this
goal will be easy to attain if the mass transfer
coefficient 1s over 4.0 ft/hr. It is still attainable
if the mass transfer coefficient is between 2.0
and 4.0, but with more difficulty. From this figure
it is apparent that a small amount of recirculating
bubbles is as effective as a large amount of once-
through bubbles. One reason is that the contact
time for recirculating bubbles is about four times
that for the once-through bubbles.

Another variable that will strongly affect the
poison fraction is the graphite surface area in the
core. Calculations indicate that if the graphite
surface area is doubled, all other parameters re-
maining constant, the poison fraction will increase
by 50 to 70%.

This model has also been used to compute the
noble gas contribution to afterheat of the unclad
graphite. Xenon and krypton are involved in over
30 fission product decay chains. The model was
used to compute the flux of each xenon and krypton
isotope into the graphite, assuming that this flux
1s constant for the entire time the reactor is at
power. From this we computed the concentration
of each noble gas and all its daughters in the
graphite as a function of time that the reactor is
maintained at power. Results of calculations for
the reactor after ten years at full power are shown
in Fig. 7.3. The reactor parameters are the same -
as used in the '?5Xe poisoning calculations. Two
curves are shown in the figure. Rather than listing
all the circulating bubble parameters involved
(e.g., void fraction, mass transfer coefficient, etc.),
it 1s sufficient to list the equivalent !'35Xe poison
fraction. The afterheat is proportional to this
value,

Work is under way in two areas. First, we are
considering ways to introduce circulating bubbles
of uniform size and about 0.020 in. in diameter.

A small model of a mechanical bubble generator
that operates somewhat like a mixer has been built
for testing with air and water. No quantitative
results are yet available. Second, a closer look
is belng taken at the bubble mass transfer coef-
ficients. An experiment is being considered that
will vield a measured value to this parameter.
95

7.2 MSBR FUEL CELL OPERATION pump was completely renovated, and the procure-

Wi

Dunlap Scott P. G. Smith

TH MOLTEN SALT ment of some loop materials was started. The
procurement of a representative graphite for the
fuel cell is expected to be a critical item, and,
therefore, the size of the cell in the initial tests

We have started an experimental program designed  wil] he controlled by the available graphite.
to give an early demonstration of the compatibility

of a full-sized graphite fuel cell with a flowing salt
stream. The cell will include graphite-to-graphite 7.3 SODIUM FLUOROBORATE CIRCULATING
and the graphite-to-metal joints. An existing LOOP TEST

facility, the Engineering Test Loop from the MSRE
development program, is being reactivated for this

work.

The loop will be operated with a single cell over
a range of conditions expected in the MSBR. These

- are!

 

P. G. Smith A. N. Smith

An existing MSRE-scale forced convection salt
loop is being altered to accept sodium fluoroborate
NaBE ) as the circulating medium. This test is

 

 

 

 

 

 

 

 

 

 

 

 

 

part of a program to qualify NaBF , for use as a
Flow rate 18 to 35 gpm coolant for the MSBER.
. Temperature 1000 to 1300°F The salt loop is part of the Fuel Pump High-
Helium overpressure 5 to 20 psig Temper.ature Endurftnce Test Fa'c111ty and normally
uses Li-Be-U fluoride salts, which have very low
Design of the alterations needed for the loop to vapor pressures at the temperatures of interest,
accept the fuel cell was begun, the circulating Since the NaBF _ exerts a BF | partial pressure of
ORNL~DWG 6711816
750 ¢m>/min
PUMP SHAF T PURGE
o TOTAL PRESSURE, 25 psig
HEL I PUMP TANK { BF, PARTIAL PRESSURE, 18 psi T STAGK
U
FLOW
CONTROL —FO0 /
{ 100 Cm3/m|n
T
, - Tfiéw A ™~ HELIUM AND BF
BF g T - o ‘} 1456 em>/min 3
SUPPLY -
SU FLOW
CONTROL e — 1 e FILTER
700 cm¥/min T
" NaBFg4
- 850 gpm
1200°F
i
VARIABLE ORIFICE ~ e
. i _,.:—_;—:‘:_’_:i’,‘\\ j:

 
 

T FLOW ELEMENT

FREEZE

VALVE

.

Fig. 7.4. Cover Gas Control System for the Fluoroborate Test.
about 500 mm Hg at the test operating temperature
(1200°F), it will be necessary to maintain an over-
pressure of BF . in the pump bowl vapor space to
avoid a loss of BE | from the salt with the resultant
increase in salt liquidus temperature. The cover
gas system is being revised to include the neces-
sary equipment for handling and controlling the
BF ,. This system, shown schematically in Fig.
7.4, includes some of the features of the cover

gas system used with the MSRE coolant salt, and
the information developed in the test will be useful
in planning the revisions which will be necessary
to prepare for testing of NaBF | in the MSRE
coolant system.

The operating conditions for the loop are:

Temperature 1200°F
Flow rate 800 gpm
Pump head 120 ft
Pump speed 1800 rpm

The tendency of BF | to induce polymerization of
organic materials could cause problems in the
pump lubrication system and in the off-gas line,
The helium purge to the shaft annulus will be
adjusted to minimize diffusion of BF | into the
pump bearing chamber, and filters are provided

in the off-gas line to protect the pump tank pres-
sure control valve., The BF | flow required will be
dictated by the total pump bowl pressure, the
desired BF3 partial pressure, and the required
helium purge flow.

It is planned to operate the loop isothermally
for a period of about six months. The objective
will be to uncover any problems associated with
the circulation of NaBF , and to devise and test
suitable solutions or corrective measures.

7.4 MSBR PUMPS

A. G. Grindell
P. ;. Smith L. V. Wilson

Survey of Pump Experience Circulating Liquid
Metals and Molten Salts

A survey of experience with pumps for liquid
metals and molten salt was made, and a report?

Ip. . Smith, Expetrience with High-Temperature
Centrifugal Pumps in Nuclear Reactors and Their
Application to Molten-Salt Thermal Breeder Reaciors,
ORNL.-TM-1993 (September 1967).

96

was issued relating pump descriptions, operating
hours, and the problems encountered during oper-
ation.

Introduction of MSBR Pump Program

The objectives of the salt pump program for the
MSBR include the production of suitable and reli-
able pumps for the fuel, blanket, and coolant salt
circuits of the Molten-Salt Breeder Experiment
(MSBE) and its noanuclear prototype, the Engi-
neering Test Unit (ETU). Table 7.2 presents the
pump requirements as they are presently envi-
sioned. A single conditional objective requires
that the pumps developed for the MSHE should be
capable of flow capacity scale-up by a factor of -
approximately 4 to the 550-Mw (thermal) MSBR
with little or no additional development work.

Our approach is to invite the strong participa- -
tion of U.S. pump industry in the design, develop-
ment, and production of these pumps. We will
prepare pump specifications along with a prelimi-
nary pump assembly drawing, pertinent rotor-
dynamic and heat transfer analyses, and the re-
sults of a survey of fabrication methods for sub-
mission to pump manufacturers. The pump manu-
facturers would be asked to make an independent
analysis of the pump specifications and support-
ing material and to define all the changes and im-
provements they helieve necessary. Parentheti-
cally, it may prove necessary to pay for several
independent analyses. The pump manufacturers
would then be asked to bid on the production of
the detailed pump design and shop drawings, the
fabrication and assembly for shop inspection of -
the required quantities of salt pumps, and the
shipment of disassembled pumps to ORNL for
further testing. )

Two important implicit requirements are pro- )
vided in this approach. The individual pump
configurations are matched to the varions MSBE
systems requirements by and at ORNL, and the
responsibility for approval of the final pump de-
sign and the detailed drawings rests with ORNL.

The principal pump components requiring de-
velopment effort are the molten-salt bearing, if
used, the shaft seal, and a full-scale rotor-dynamic
simulator, if supercritical operation of a salt pump
is required, that is, operation of the pump at
speeds above the first critical shaft speed. The
pump manufacturers would be invited to partici-
pate in this and other development work they may
Table 7.2. Pumps for Breeder Reactors

 

 

Fuel Blanket Coolant
2225 Mw (thermal) MSBR
Number required 42 42 47
Design temperature, °F 1300 1300 1300
Capacity, gpm 11,000 2000 16,000
Head, ft 150 26 150
Speed, rpm 1160 1160 1160
Specific speed, Ns 2830 2150 3400
Net positive suction head required, ft 25 8 32
Impeller input power, hp 990 250 1440
150 Mw (thermal) MSBE
Number required 1 1 1
Design temperature, °F 1300 1300 13060
Capacity, gpm 4500 540 4300
Head, ft 150 80 150
Speed, rpm 1750 1750 1750
Specific speed, Ns 2730 1520 2670
Net positive suction head required, ft 27 5 26
Impeller input power, hp

 

reference design or modular design.

deem necessary. However, it would appear more
economical to perform the molten-salt bearing de-
velopment work at ORNL, where the fuel produc-
tion facilities and the handling techniques are
already available. Proof testing of completed
pumps in molten salt prior to operation in either
the ETU or the MSBE will be conducted at ORNL.
Endurance testing of prototype pumps in molten
salt will also be conducted at ORNL in the proof-
testing facilities.

Because experience indicates that production of
suitable and reliable salt pumps is one of the
longest lead-time items for molten-salt reactor ex-
periments, it is important to get an early start in
the pump program.

If study indicates that the MSBE salt pumps can
be operated subcritically but that the MSBR pumps
must be operated supercritically, then the con-
ditional objective may require operation of a salt
pump at supercritical speeds during the course of
the MSBE program to build confidence in the
reliability of a pump with such a long shaft.

Fuel and Blanket Salt Pumps

Preliminary layouts have been made for the fuel
salt pump, blanket salt pump, and coolant salt
pump. The concepts of the fuel salt pumps, shown

410 61 350

in Fig. 7.5, and the blanket salt pump are similar,
having a shaft of the order of 34 ft long, supported
by an oil-lubricated radial and thrust bearing at
the upper end and a molten-salt-lubricated journal
bearing near the impeller at the lower end. The
main differences in the two pumps lie in (1) the
size of the fluid flow passages to, through, and
from the impeller, (2) the absence of a pump tank
on the blanket salt pump, and (3} the sizes of the
drive motors. The similarity of the pumps which
is derived from their common environment and
placement within the cell results in common
analytical and developmental efforts in the areas
of bearings, seals, rotor dynamics, motor contain-
ment, general layout, ditect and remote mainte-
nance, ancillary systems, fabrication and assembly,
and nuclear heating.

The fuel salt and blanket salt pumps are de-
signed so that the rotary element which contains
all the moving parts can be replaced by remote
maintenance without having to cut any of the salt
lines to or from the pump. Direct maintenance can
be performed on the drive motor and the bearing-
seal assembly at the upper end of the pump. A
static seal can be brought into play to separate
and protect the maintenance area from the radio-
activity in the pump when the bearing-seal assem-
bly is to be removed.
98

ORNL-DWG S7-11817

MCTOR

- COUPLING

MBLY

- UPPER BEARING AND
SEAJL ASSE

 

CITRIVE

UL HE

BELLOWS

-- CONCRETE SHICLDING

PUMP TANK

"MOLTEN SALT
BEARING

-IMPELLER

 

 

 

 

 

 

o Ha -

L) ME2
e UIBLE HiE

B

 

-

- HEAT EXCHANGER

 

Fig. 7.5. Preliminary Layout of the MSBR Fuel Salt Pump.
Nuclear heating of the pump tank and the sup-
port structure within the pump tank is removed by
circulating a portion of the fuel salt from the main
salt stream over the heated surfaces. To remove
the nuclear heat from the shaft, a small amount of
salt is bled up the center of the shaft and fed into
an annulus between the shaft and a cooling tube
that extends the length of the pump tank. A
[abyrinth seal at the lower end of the tube forces
most of the salt to flow to the upper end of the
tube, where it spills over into the pump tank. An
added benefit is the increased damping and
stiffness provided to the shaft by the salt in the
annulus.

Analyses are being made of the nuclear heating
in that portion of the pump casings and shaft for
which no cooling is provided. If a problem is
found, we can provide cooling or shielding and
thermal insulation where needed to reduce the
heat generation in the pump structure to an ac-
ceptable level.

The seal arrangement at the upper end of the
shaft is similar to that used in the MSRE salt
pumps. It consists of a face-type seal (Graphitar
against tool steel) with the 1ubricati_ng oil on one
side and the helium in the shaft annulus on the
other. Helium is brought into the annulus to serve
as a split purge between the salt and gaseous
fission products at the lower end of the shaft and
any lubricating oil that leaks through the face seal
Part of the helium passes
down the shaft through a close-fitting labyrinth,
where the increased gas velocity reduces the
upward diffusion of molten-salt vapor and gaseous
fission products. Concurrently, that portion of the
helium passing upward through the labyrinth seal

into a leak-off line.

prevents the downward movement of lubricating oil
vapors and also serves to scavenge oil leakage
and vapors overboard from the pump.

Coolant Salt Pumps

Two preliminary layouts of the MSBR coolant
salt pump have been prepared. One layout utilizes
a pump with a short overhung shaft mounted on two
oil-lubricated rolling element beatings, and the
other is a long shaft with an oil-lubricated bearing
at the top end of the shaft and a molten-salt
bearing located just above the impeller. One
criterion for the pump requires variable-speed op-
eration over the range 300 to 1200 rpm. The dif-

99

ficulty with the short-shaft pump is that to have
the pump operate below the first critical speed,
the shaft diameter would have to be greater than

8 in., which would present a formidable seal de-
sign and development problem. If it were designed
to operate above the first critical and below the
second critical speed, the shaft diameter would be
approximately 3 in., which is inadequate to trans-
mit the torque. For the long-shaft pump configu-
ration, however, a shaft with a diameter selected
on the basis of torque requirements would have a
first critical speed well above the maximum oper-
ating speed. The long-shaft pump would also use
the same upper bearing and seal configuration
planned for the fuel and blanket salt pumps. Hence
the long-shaft concept appears to be preferable for
the coolant pumps.

The coolant salt pump will have the impeller and
volute mounted in a pump tank of sufficient volume
to accommodate the thermal expansion of the cool-
ant salt for the most adverse thermal condition that
might arise during reactor operation. A double vo-
lute pump casing has been selected to reduce ra-
dial loads on the impeller and the resultant loads
on the molten-salt bearing, particularly when op-~
erating at off-design conditions, and to reduce the
diameter of the bridge tube, which provides a
flexible connection from the volute to the pump
tank nozzle.

We believe that the coolant pump drive motor,
although having a greater horsepower, can be de-
signed to fit the same containment vessel as that
for the fuel pump drive motor.

Water Pump Test Facility

Preliminary layouts have been prepared of a
facility for testing the fuel pump with water. The
configuration does not incorporate the long shaft
of the high-temperature pump but only mocks up
those portions which affect the fluid flow. The
layout also includes a mockup of the inlet to the
heat exchanger tube sheet with sufficient instru-
mentation to monitor the flow distribution in the
heat exchanger tubes. The distribution of the gas
injected to remove the xenon will be monitored
also.

The configuration has been designed to permit
water testing of the blanket pump in the same
facility. The purpose of the water test facility in
the pump development program is (1) to determine
head and flow characteristics of the impeller-
diffuser design, (2) to measure radial hydraulic
forces acting on the impeller (needed for designing
the molten-salt bearing), (3) to measure and re-
duce to an acceptable level the axial forces acting
on the impeller, and to determine the relationship
between the axial clearance at the bottom end of
the impeller and the axial force, (4) to observe

the fluid behavior in the pump tank, and to make
the necessary changes to reduce gas entrainment
to an acceptable level, (5) to assure that the
mocked-up molten-salt bearing will run submerged
under all operating conditions, and (6) to check
the point of cavitation inception and the required
net positive suction head of the impeller.

Molten-Salt Bearing Tests

The present layouts of the MSBE and MSBR
salt pumps require a molten-salt journal bearing
near the impeller. A molten-salt bearing presents
three important considerations: (1) the hydro-
dynamic design of the bearing to provide the
requisite lubricating film, (2) the selection of the
kind and form of the bearing materials, and (3)
the design of a bearing mounting arrangement
which will preserve the lubricating film despite
thermal distortions between pump shaft and
casings.

We are studying the use of hard, wear-resistant
coatings on the journal and bearing surfaces.
Such coatings present advantages over the
sintered, solid-body journal and bearing inserts
most often used in high-temperature process fluid
lubrication. The hard coatings are convenient to
apply and hopefully eliminate the differential
thermal expansion problems. Mechanical Tech-
nology, Inc., of L.atham, New York, has been
engaged to produce Hastelloy N specimens with
each of four different hard coatings: (1) cobalt
(6 to 8%) bonded tungsten carbide, (2) nickel
(7%) bonded tungsten carbide and mixed tungsten-
chromium carbides, (3) nickel-chromium (15%)
bonded chromium carbide, and (4) molybdenum
(7%) bonded tungsten carbide.

These coatings will be subjected to corrosion
and thermal cycling tests in molten salt at ORNL.
A test in molten salt will be made with a 3 x 3 in.
bearing using one of these coatings, if one should
prove satisfactory.

A layout is being made of a tester to accom-
modate a full-scale molten-salt bearing for the

100

MSBE fuel salt pump. The tester will be capable
of subjecting the hearing and its mounting arrange-
ment to start-stop wear tests and thermal cycling
and endurance tests in molten salt.

Rotor-Dynamics Feasibility lnvestigation

Mechanical Technology, Inc., is performing an
analysis (Reactor Division subcontract No. 2942)
of the rotor dynamics of the preliminary layout of
the MSBR fuel salt pump to determine its flexural
and torsional critical speeds and flexural response
to a dynamic unbalance. Interim results? of the
analysis show that the pump will operate between
the fourth and fifth flexural critical speeds of the
pump system, which includes the pump shaft, inner
and outer pump casings, and the drive motor.

The third and fifth system criticals are essentially
the first and second simply supported beam criti-
cals of the shaft. The critical-speed results also
show that the pump-system criticals are relatively
independent of the bearing stiffness over a range
representative of practical bearing designs. The
stiffness characteristics of the drive motor
coupling also have little effect on system ciiti-
cals.

The synchronous response amplitudes resulting
from a ‘“‘bowed-shaft’’
been calculated over the complete range of pump
speeds. The response results show only one sys-
tem critical to be significant from a bearing load
standpoint -- namely, the ‘‘first shaft critical®’

unbalance condition have

which occurs at about 700 rpm.

In addition to passing through one shaft critical
speed, three additional system criticals must be
traversed as the pump accelerates to design speed.
Thesec three criticals are basically cantilever
resonances of the outer casing. The first two
cantilever modes occur at quite low speeds and
hence should not be a problem from a steady-state
standpoint. However, if the pump system should
be transiently excited during normal operation,
these cantilever beam modes would be the primary
contributors to the resulting transient vibration
response of the pump system.

The third cantilever beam mode of the outer
casing also excites a simply supported resonance
of the inner casing. This mode occurs between

 

2P. W. Curwen, Rotor-Dynamic Feasibility Study of
Molten Salt Pumps for MSBR Power Plants, MTI-67TR48,
Mechanical Technology, Inc., August 6, 19867,
800 and 930 rpm and makes it appear advisable to
separate the inner and outer casing frequencies by
suitable changes in the wall thicknesses and
diameters of the two casings.

A preliminary undamped torsional critical-speed
analysis has been made for the pump system, and
the results indicate that the two torsional critical
speeds that might affect pump operation can be
strongly dependent upon the electromagnetic
torsional stiffness of the drive motor. We believe
that by changing some of the component dimensions
and accounting for inherent system damping, the
pump will operate satisfactorily between the first
and second torsional critical speeds.

Fabrication Methods Survey. — Based on the
preliminary layouf of the MSBR {uel salt pump, the
shaft and inner and outer casings were detailed,
and a survey is being made of potential fabrication
methods to identify fabrication problems. For the
shaft, it is important to determine the straightness
and concentricity tolerances that can be supplied.
These tolerances have considerable effect on the
provisions that must be made for the dynamic
balancing of the shaft, which must be done rather
precisely when the pump is to operate above the
first shaft critical speed. A manufacturer was
found who could fabricate shafts in the range 7 to
10 in. outside diameter with a wall thickness as
small as 1/2 in., and who would guarantee the
straightness from end to end to 0.005 total indi-
cator reading (TIR) and the outside diameter—
inside diameter concentricity to 0.005 in., but at
considerable expense. As these two tolerances
are relaxed, more manufacturing capability is
available, and the fabrication costs are reduced;
however, dynamic balancing of the shaft becomes
a more important portion of the fabrication process.

101

An investigation is under way to determine the
relationship between shaft precision and total
shaft cost (fabrication plus dynamic balancing) as
well as its effect on pump design.

Several manufacturers have been found who are
capable of fabricating the inner and outer casings
to the tolerances shown on the preliminary layout,
but also at considerable expense. The effects on
pump design of relaxing the preliminary values of
the tolerances are being studied.

Other Molten-Salt Pumps

Fuel Pump High-Temperature Endurance Test
Facility. — The facility,? incloding the salt pump,
gas systems, instrumentation, and handling equip-
ment, is being prepared for operation with sodium
fluotoborate (NaBF 4) . The new drive motor, rated
200 hp at 1800 rpm, was delivered, and the existing
motor support was modified to suit the new motot.
Also, the pump rotary element was removed
from the facility, disassembled, cleaned, and
reassembled.

Molten-Salt Bearing Pump Endurance Test
Facility. — A new salt bearing constructed of
Hastelloy N was installed on this pump.® The
gimbals support for the bearing was modified to
reduce the possibility of the support becoming
disassembled during operation. The bearing and
gimbals arrangement was satisfactorily tested
with oil as the pumped fluid at room temperature.

 

SMSR Program Semiann. Proge. Rept. Feb., 28, 1967,
ORNL-4119, p. 66.

*MSR Program Semiann. Progr. Rept. Aug. 31, 1966,
ORNIL.-4037, p. 82,
Part 3. Chemistry

W. R. Giimes

The chemical research and development effort
in close support of the MSBR program includes, as
described in this chapter, a variety of studies.

A major share of this effort 1s still devoted to the
immediate and anticipated problems of the operat-
ing Molten-Salt Reactor Experiment.

Sampling of the MSRE fuel and coolant salts and
interpretation of the analyses for major and minor
constituents of the melt, and examination of metal
and graphite surveillance specimens from the core
and of specimens exposed to the pump bowl gases,
continue as routine, though obviously necessary,
portions of the total effort. Minor fractivns of
several fission products continue to appear in the
pump-bowl gas space. The possibility that these
may occur as volatile compounds has prompted
examination of the chemistry and the vaporization
behavior of the litile-known intemmediate valence
fluorides of molybdenum.

Oxide-fluoride equilibria in the LiF-BeF , sys-
tem and its more complex counterparts with added
UF , and ThF , are under study, since such
equilibria may well lead to separation processes
of value and seem to have shown us container

materials that will greatly aid our experimental

program.

The plans to substitute a lower melting and more
economical coolant for the LiF-BelF , mixture in
the MSRE have required examination of phase be-
havior among the alkali fluoroborates and of
ancillary questions of the decomposition pressuze
of these materials and of possible undesirable
interactions of BF ; with metals and lubricants to
which it would be exposed.

While recovery of uranium by fluorination (both
from fuel and blanket) and recovery of fuel salt
by vacuum distillation remain as the design re-
processing methods, the recovery of protactinium
from the blanket and the removal of fission prod-
ucts from the fuel by reductive extraction iato
molten metals continue to show promise and are
actively pursued.

Development studies in analytical chemistry
have been directed primarily to improvement in
analysis of radioactive samples of fuel for oxide
and uranium trifluoride and for impurities in the
helium gas from the MSRE.

8. Chemistry of the MSRE

R.

8.1 MSRE SALT COMPOSITION AND FURITY

Molten fuel, flush, and coolanat salts have been
intemmittently circulated and stored in the MSRE
for approximately two years. In use, these salts
have been subjected to chemical analysis on a

regular basis.! The results of these analyses

102

E. Thoma

signify that generalized corrosion in the fuel and
coolant circuits is practically absent, and that

lChemical analyses performed under the supervision
of C. E. Lamb, Analytical Chemistry Division; spec-
trochemical data were obtained by W. R, Musick,
Analytical Chemistry Division.
the salts are currently in essentially as pure
condition as when charged into the reactor.

Fuel Salt

MSRE runs 11 and 12 were completed within the
current report period. During this period, small
amounts of beryllium metal were dissolved into
the fuel salt to adjust the oxidation-reduction
potential of the salt. The total concentration of
uranium in the fuel was also increased by addition

of "LiF-?3°UF, to the circulating salt. Curmently,

the uranium concentration of the fuel salt is
approximately 4.590 wt %, of which the U?" frac-
tion of the total uranium is 1.5%. The effects of
the beryllium and 7LiF—”‘qUF4 additions are
evident in the results of the chemical analyses of
the fuel salt shown in Table 8.1. A refinement of
analytical procedure was introduced during run 11;
preliminary values for the detemination of ura-
nium concentrations were confirmed by a second
group of analysts before final values were re-
ported. Continuous control methods were em-
ployed by both groups. This innovation in pro-
cedure resulted in a significant improvement in
precision. Average scatter was reduced from
10.5% to +0.4% of the value, corresponding to
+0.03 and 10.02 wt % uranium.

MSRE fuel salt is analyzed by HF-H , purge
methods for evidence of oxide contamination. The
results of analyses obtained during runs 11 and 12
indicated that the fuel salt does not contain more
than 50 to 60 ppm of oxide; that is, it is currently
as free of oxide as when it was originally charged
into the MSRE.

Coolant Salt

When run 12 was terminated in August 1967, the
coolant salt had circulated in the MSRE for a
perod of 12,047 hr. Coolant salt specimens were
submitted at one-month intervals during 1967.
Results of those analyses show the composition
and purity of the salt to be

Li Be F Fe Cr Ni 0

 

{wt %) (ppm)

 

13.04 9.47 76.30 63 27 12 ™200

103

as compared with the composition and purity it
was known to possess a year ago:

Li Be F Fe Cr Ni 0

(wt %)

 

(ppm)

 

13.03 9.46 76.16 58 36 16 ™200

The two compositions are not differentiable with-
in the precision of the analytic methods. The
constancy of the trace concentrations of the
impurities attests to the fact that the cover gas,
which is supplied to both the fuel and coolant
circuits from a common source, has been main-
tained in a state of high purity throughout the
entire operational period.

Flush Sait

Whenever the MSRE fuel circuit is flushed with
flush salt, there is cross mixing of fuel and flush
salts by residues which remain in the reactor
after each is drained. We need to know the
amounts of material transferred between the fuel
and flush salts because they enter into the calcu-
lation of the book-value concentration of uranium
in the fuel salt. Sufficient analytical data are
now available to enable us to deduce the average
mass of these residues.

The concentration of uranium in the flush salt
appears to change in nearly equal increments
duting each flush operation, as shown in Table
8.2. These data indicate that fuel salt which
remains in the fuel circuit after drainage of the
fuel increases the uranium concentration of the
flush salt by 200 ppm each time the drained re-
actor is cleaned with flush salt. An increase of
200 ppm of uranium corresponds to the addition of
approximately 850 g of uranium to the flush salt.
This is the amount of uranium in 19.20 + 0. 10 kg
of fuel salt in which the uranium concentration is
between 4.570 to 4.622 wt % U, the range for the
MSRE during the period considered.

On filling the MSRE with fuel salt, "LiF-BeF,
(66-34 mole %) flush salt residue is incorporated
in the fuel salt, diluting its concentration of UF |
and ZrF | slightly. This dilution is reflected in
the uranium and zirconium analyses shown in Fig.
8.1. The decrease in zirconinm concentration of
the fuel salt from a mean value of 11.33 wt % to
10.85 wt % corresponds to dilution of the fuel by
12.7 kg of salt on each drain-flush-fill cycle.
 

 

Table 8.1, Summary of MSRE Fyel Sait Analyses, Runs 11 and 12
Sample Li Be Zr :na;yt Ubook F Fe Cr Ni Total
No. (wt %) {(wt %) wt %) {wt %) {wt %) (wt %) {ppm) (ppm) (ppm) (wt %)

¥FDP11-01 11.18 6.28 10.90 4,603 4,576 67.46 131 66 54 100.44
FP11-02 11,10 6.27 10,65 4,599 4.576 69,16 112 75 63 101.81
FP11-03 10.42 6.21 11.08 4,604 4,575 565,72 150 61 64 99,06
FP11-04 11,190 6.33 11.19 4,562 4.575 67.87 143 62 22 101.01
FP11-05 wt/SU - 0.37%
FP11-06 11.25 6.31 10,97 4,555 4,574 67.44 131 67 33 100.54
FP11-07 11.38 6.70 11.27 4.558 4,574 69.92 172 62 50 103.86
FP11-08 i1.50 6.62 10,81 4.569 4,573 68.89 312 54 107 102.43
FPi1-09 Gas samplie
¥Pii-10 Beo addition, 11.66 g
Frii-11 10.57 6.57 10.87 4.551 4,572 69.94 165 73 43 102.62
¥FP11-12 10.93 6.27 10.88 4.567 4,572 69.96 76 75 34 102.62
FP11-13 U T2U = 0.42%
FP11-14 10.98 6.35 11.26 4.539 4.571 67.83 98 78 75 100.98
FP11-15 1.11 6.49 18,74 4,552 4,571 68.89 i 62 37 101,80
FP11-i96 Gas sample
FP1i-i7 11,33 6.47 11.21 4,553 4.571 70.25 67 58 47 193.85
FP11-18 10.73 6.67 11.17 4.579 4,570 68.68 120 56 43 101,85
FP11-i9 10.47 6.57 11,11 4.589 4,570 67,12 117 68 42 90.38
FP11-20 10,51 6.72 10.98 4,561 4,570 67.60 122 59 49 100.39
FP11i-21 10.50 6,36 10.89 4.576 4,570 66,823 168 63 46 99.19
FP1:-22 10.55 6.57 10,92 4.572 4.559 66.05 104 62 63 98.68
FP11i-235 10.53 5,49 10.95 4.583 4,569 69,70 136 63 55 102.28
FP1i-24 1G.55 6.51 10,99 4,547 4.5569 58.50 i73 67 63 101,04
FP11i-25 Sampie for oxide analysis; analysis unsuccessful
FP1i-29 10.43 .49 10,91 4,57C 4.568 67.17 118 52 53 99.59
FP11-27 10,52 .85 10.85 4,577 4.568 66.04 72 63 &80 90,84
FP11-28 Sample for oxide analysis; oxide concentration, 58 ppm
FP11-29 10.48 45 10.92 4.584 4.567 64.52 126 64 56 97.08
FPil-30 10.48 .39 11,02 4.597 4.597 64.51 i15 56 73 97.01
FP11-31 1G.53 58 11.06 4,559 4,566 67.06 80 64 50 95,81
FP11-32 U320 - 0.34%
FP11.33 10.53 65.43 11.14 4,567 4,564 66,38 142 72 72 99,07
FP11-34 10.55 5.33 11.37 4.582 4.5646 68.99 146 64 64 101,85
FPii-35 10.53 6.35 1i.12 4.566 4.565 67.24 194 73 64 99,84

1
Table 8.1. (continued)

 

 

Sample Li Be Zr Zn‘alyt Ubook F Fe Cr Ni Total
No. (wt %) (wt %) {(wt %) (wt %) (wt %) (wt %) . {ppm) {ppm) {ppm) {wt 7o)

FP11-35 Gas sample
FP11-37 10.55 6.33 10.75 4,541 4,565 65.93 79 80 49 96.12
FP-11-38 Sample for us +/S.U analysis; analysis unsuccessful
FP11-39 11.57 6.44 10.92 4.536 4,565 66.55 182 69 52 100,05
FP11-40 Be? addition, 8.40 g
FP11.4% 10.42 6,37 10.77 4.579 4,564 68.5 135 56 58 100.74
FP11-42 Gas sample
FP11-43 ud +/):U — no analysis performed
FP11-44 10.50 6.60 11.01 4.561 4,564 69,88 140 59 44 102,57
FP11-45 10.58 6.50 10.65 4,548 4,563 67.13 88 54 41 09,43
Average 10.80 * 0.35 6.46 £ 0.15 10.97 £0.18 4,570 £ (.018 67.81 £1.46 131 +48 64 L6 54 £ 6
FP11-46 Gas sample
FP11-47 10.95 6.48 10,96 4,604 4.582 66,65 169 71 75 949.67
FP11-48 10.45 6.52 10.85 4,578 4.581 67.23 210 49 58 99,66
FP11-49 U +/EU — no analysis performed
FP11-50 Gas sample
FP11.51 11.20 6.45 10.97 4,571 4,580 69.61 114 61 61 102,83
FP11-52 11.33 6.45 10.79 4,566 4,580 69,27 80 60 25 102.43
FP11-53 Gas samnmple
FP11-54 10.93 6.63 10.95 4,551 4.579 £69.91 158 61 63 103.00
FP11-35 Special 50-g sample
FP11-56 Sample for oxide analysis; oxide concentration, 50 to 100 ppm
FP11-57 No sample obtained
FP11-58 10.48 6.60 11.27 4.607 4.578 70.95 131 81 42 104.00
Average 10.82 £0.35 6.47 +0.15 10.97 £0.17 4,571 £0.019 57.09 £1.33 133 £47 64 17 54 T 16
FP11-59 13.43 7.57 0.345 0.0292 76.18 222 68 40 97.51
FP11-60 13.40 9.36 8,260 0.0268 79.23 110 74 34 102,30
FP12-01 13.40 8.64 <0.20 0.6778 80.22 108 66 26 102.56
FP12-02 13.70 0,590 <0.20 8.07%3 75.40 104 7 23 88,90
FP12-03 13.60 9,38 <0.20 0.0826 77.4Q a3 63 24 100,68
FP12-04 Sample for oxide analysis
FP12-05 11.20 6.74 10.94 4.550 4.548 66.32 123 52 60 99.77
FP12-06 udt/Zu = 0.37%
FP12-07 Gas sample
FP12-08 Be addition, 7,93 g
FP12-09 Be additien, 9.840 g

SOT
Table 8.1,

(continued)

 

 

Sample Li Be Zr :naiyt Ubook F Fe Cr Ni Total
Ne. (wt 70) {wt %) (Wt %) (wt %) {wt %) (wt 7o) {ppm) {ppm) {ppm) (wt %)
FP12-10 11.60 6.91 10.57 4.525 4,547 67.27 i34 71 72 141.90
FPi2-11 U2 = 1%
FP12-12 11.60 6.54 10.91 4,545 4.547 66.66 113 64 62 100,78
FPI2-13 Be addition, 8.33 g
FP12-14 11.50 6.50 11.22 4.557 4.546 67.95 145 82 47 101.76
FP12.15 Be addition, 11,68 g
FPi12-16 11.40 6.40 10,62 4.567 4.546 68.27 269 110 08 101,31
FP12-17 11.30 6.40 10,656 4.532 4,545 66.92 215 144 53 99.85
¥P12-18 Sample for oxide analysis; oxide concentration, 57 ppm
FP12-19 11.50 6.19 11.00 4,522 4.545 65.05 100 102 62 98.29
FP12.20 10.60 6.36 10,76 4,557 4.545 65.76 &1 654 44 G8.06
FP12-21 U3 /ZU - 0.5%
FP12.22 10,50 6.52 10.43 4.566 4.544 86.18 247 ad 50 GR.22
FP12-23 10.60 5,68 10,58 4,526 4,544 H5.26 154 78 76 a7.67
FP12-24 11.38 5.36 10.67 4.567 4.544 66.46 176 av 56 99.47
¥FP12-25 10.70 5.46 1G.53 4.496 4.544 66,10 208 68 62 98.32
FP12-26 Gas sample
FP12-27 10.70 6.61 10.78 4,550 4.544 67.50 195 75 72 100.18
FP12-28 10,70 6.44 10.656 4.569 4,544 69,00 150 68 52 101.40
FP12.29 10.70 6.41 10.87 4,520 4.543 67.11 177 &4 78 99.65
FP12-30-35 "LiF-?37UF , additions
FP12-36 10.50 6.58 10.96 4,554 4,555 66.40 156 84 359 99.15
FP12-37-40 "LiF-?*°UF additions
FP12-41 10.60 6.52 10.68 4.562 4,565 68.00 110 64 192 100,40
FP12-42-46 "LiF-??UF, additions
FPi12-47 10.60 6.50 10.50 4,586 4.576 65.34 i20 72 70 98.56
FP12-468-50 "LiF-??PUF | additions
FP12-51 i1,22 6.60 10.32 4.588 4,576 66,32 94 72 39 99,07
FP12-52 10.70 6.47 10.71 4.5084 4.576 65.98 119 72 932 98.48
FP12-53 10.50 6.39 10.66 4,503 4.575 65.48 132 72 60 97.56
FP12-54 Sample for isotopic dilution analysis
FP12-55 10.75 6,44 11.12 4.577 4,574 65.72 156 72 300 98.66
10.80 6.42 10.95 4,575 4.574 54,57 136 58 720 97.42

901
Table 8.1. (continved)

 

 

Sample Li Be Zr :nalyt Ubook F Fe Cr Ni Total
No. (wt %) (st %) (wt %) (Wt %) (wt Te) {(wt %o) (ppm) (ppm) {(ppm) (wt To}
FP12-56 Be addition, 9.72 g
FP12-37 11.33 6.59 11.08 4.587 4,574 65,14 156 34 170 98.77
11.30 6,74 11.28 4,549 4,574 66.53 160 654 424 100.46
FP12-58 11.00 6.58 10.77 4,600 4,574 66.62 138 74 606 99.60
FP12-56 Sample ladle remained in pump bowl
Average 10.93 6.25 10.78 66,50
0,40 +3.15 10.24 10,03

 

a - . . fak
Corrected to compensate for isotopic composition,

LOT
108

Table 8.2, Chemical Analyses of MSRE Flush Salt Specimens

 

Average Uranium

Average Increase in
Number of Samples

 

Run No. Found Uranium for Flush

(ppm) Analyzed (ppm)
FP-3 (final) 195 1 195
P-4 (initial) 218 6 218
FP-8 (initial) 460 3 230
FP«8 (final) 616 1 205
FP-9 (final) 840 1 210
¥P-11 (final) 930 1 186
FP-12 (initial 799 3 160
FP-13 (initial) 1186 3 197

e

Overall average = 200 ppm

 

Implications of Current Experience
in Future Operations

Currently, the MSRE is entering its final period
of operation with ?3°U fuel salt. It is planned
that the MSRE will operate with ?33U fuel in
1968, % as will the MSBE later, and that the con-
centration of uranium tetrafluoride in those fuels

will be only one-fourth that employed in the MSRE.

Several inferences may be drawn from the ex-
perience developed during previous "operation
which have significant implications regarding
operation of the MSR[Z when it is charged with
2337 fuel, as well as for larger molten-salt re-
actors. In general, we must conclude that if
chemical analyses are to function as operational
controls, appreciably greater precision than is
now available must characterize the methods for
determining the concentration of uranium as well
as the U®" fraction in the total uranium.

The overall composition of the present fuel salt
may be determined in routine chemical analysis
with a precision of 0.2 to 0.3 wt % (Fig. 8.1).
Precision in the determination of the uranium con-
centration is considerably better, £0.02 wt % on a
statistical basis (Fig. 8 2). Such precision in the
detemination of the uranium concentration is,
however, only one-tenth that which is obtained in
outine computations of reactivity balance. The

 

Zp. N, Haubenreich et al. to R. B. Briggs, private
communication, Dec. 19, 1966,

high sensitivity in the reactivity balance to varia-
tions in uranium concentration vitiates applica-
tion of periodic batch analysis of fuel as a sig-
nificant control parameter in reactor operations;
such analyses have come to function primarily as
an independent basis for cross-checking burnup
and inventory computations. It is anticipated that
when the reactor is fueled with 233U, the pre-
cision of the reactivity balance will be improved
by a factor of at least 4,° while the precision of
the chemical assay of uranium will fall in pro-
portion to the uranium concentration as it is re-
duced from 0.83 to 0.20 mole %. The limitations
on the present methods of analyzing the MSRE
fuel indicate, therefore, that it will be necessary
to develop improved methods for detemining fuel -
composition. In reactor systems in which fre-
quent or nearly continuous chemical reprocessing
is carried out, composition of the fuel and blanket -
systems will undergo constant change. Un-
questionably, composition determination will then
be necessary by way of on-line techniques sup-
plemented by methods which have high intrinsic
accuracy.

The intiinsic corrosion potential of the fuel
salt is proportional to the UF | concentration,
which, to date, has been determined directly only
by an intricate and difficult method which is
probably near its limit of capability with salt of

 

3J. R. Engel to R. E. Thoma, private coiununica-
tion, April 28, 1967.
LITHIUM (wt %)

BERYLLIUM {(wi %)

ZIRCONIUM (wi 9%

(W‘.L (yo}

URANIUM

CHROMIUM (oprn)

-~
~

IRON {ppm)

NICKEL {ppr}

12.00

14,50 -
14,00 b

§0.50 ot

10.00

7.50 -

7.00

4.600 -

4,550 +—

4.500
100

80

&0

40 |-

20
200

150

160

ORNL.-DWG 67 -14818

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50 |-

Fig.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TrTY fu
T e
,4}"’____ {I L L |

T T LI
________ Ldoe topord g1
T
TTTTT T
{ 4>}{17§ } \}i ,,,,,,, *
e } 1 L
. LH}E %4} ,,,,,,, } i
AmE ”M_;w “““ -
[ T I ] .
T
L g
o ® { - fl
_______ S __d B
""" T T T
Tl e b
1T 1 | -
T e
" T T 7T ]
| L |L
i *""“filf T

109

ORNL-OWG 67T-11353
QB0 (e
?{CHEM{CAL ASSAY

2600 | % - R —

  
 
 

A5G0 &

4540 |

URENILM (wt %)

4520

 

 

 

 

4500 |1
J RUN A1---—

 

 

 

 

2 24

™}

30
VEGAWATT HOURS (x10%)

&8 a2

Fig. 8.2. Urenium Conceniration in MSRE Fuel Seit.

4 While this

method has been used with moderate success with
the MSRE fuel salt, the low total concentration of
uranium which is anticipated in future fuel salts
makes it improbable that this method can have
continued application. It will be of considerable
importance in the near future to employ direct
spectrophotometric metheds for the determination
of U?" concentration in the fuel salt. Results of
recent laboratory experiments indicate that this
approach is feasible. 5

In the future, the MSRE fuel salt as well as the
fuel salts in the large reactor plants will be sub-
jected to fluorination and to HF-H , purge streams
during chemical reprocessing. Salt streams in
those reactors may be expected to contain even
lower concentrations of contaminant oxides than
currently exist in the MSRE and should therefore
not require oxide analysis.

the present uranium concentration.

Results of the chemical analysis for chromium
have shown sufficient precision (::10%) that the
method has come to serve as an excellent and
reliable measure of generalized corrosion within
the MSRE. The utility of this analysis as an
indicator results from the fact that at present
the total concentration of chromium in the fuel
salt is low (~70 ppm). Relatively minor changes
in cotrosion are, therefore, reflected in significant
changes in chromium concentration. In future
operation it is possible that the total concentra-
tion of chromium in the fuel circuit will increase

 

4A. 8. Meyer, Jr., to R. E. Thoma, private commumni-
cation, May 12, 1967.

SJ. P. Young, MSR Program Semiann. Progr. Rept.
Feb. 28, 1967, ORNL-4119, p. 163.
tenfold or more as a consequence of chemical
reprocessing. Unless either the precision of
chromium analysis is improved or low concentra-
tion of chromium is maintained in the salt which
is returned to the reactor fuel circuit, much of the
capability for immediate detection of corrosion
will have been lost.

It is anticipated that gas chromatographic
methods for analysis of gas streams will be tested
soon at the MSRE. If application of such methods
to cover gas analysis succeeds in providing a
sensitive means forthe quantitative detemmination
of volatile fluoride, hydrogen, and oxygen-bearing
phases, a major advance toward on-line analysis
of salt purity will have been achieved.

8.2 MSRE FUEL CIRCUIT
CORROSION CHEMISTRY

Corrosion on salt-metal interfaces in the MSRE
is signaled by an increase of chromium concen-
tration in salt specimens. An increase of 10 ppm
corresponds to the removal of approximately 40 g
of chromium from the Hastelloy N surfaces. Cur-
rently, the chromium concentration of the fuel
salt is 72 =7 ppm; this concentration represents
an increase of only 34 ppm and removal of about
170 g of chromium from the Hastelloy N container
since operation of the MSRE began in 1965. If
the total amount of chromium represented by this
increase were leached uniformly from the fuel
circuit, it would correspond to removal of chromium
from a depth of 0.22 mil. Recent evidence indi-
cates, however, that only half the chromium in-
crease observed in the fuel salt may be attributed
to corrosion in the fuel circuit.

On temnination of run 7, graphite and metal
surveillance specimens were removed from the
core of the reactor and were replaced with speci-
mens contained in a new perforated metal basket.
Fuel specimens taken throughout the next run,
No. 8, were found to contain a chromium concen-
tration of 62 ppm, rather than 48 ppm, the average
concentration which had persisted almost from the
beginning of power operations. Since the only
known environmental alteration was the installa-
tion of the new surveillance specimen assemblage,
we speculated that the container basket and the
Hastelloy N specimens had sustained most of the
corrosion responsible for the observed increase
in chromium, and that corrosion might be evident

110

to a depth of 10 mils. However, recent inspection
of specimens from the basket (see Part 5, this
report) did not disclose that the anticipated cor-
rosion of the metal had occurred.

On several previous occasions, salt was re-
turned to the fuel circuit after storage in the drain
tanks without developing evidence of an increase.
Nevertheless, we are forced to conclude that the
increase of chromium in the fuel salt took place
while it was stored in the drain tank during the
ten-week interval between runs 7 and 8.

Although it is not evident how the drain tank
may have become contaminated, its surface seems
to be the source of the additional chromium in the
fuel salt. If all the chromium was leached uni-
formly, corrosion in the drain tank will have
reached a depth of 0.7 mil. If the increase of
48 to 62 ppm is attributed to the drain tank, the
total increase of chromium resulting from fuel-
circuit corrosion is only 20 ppm throughout the
entire operation of the MSRE, and corresponds to
~ 100 g of chromium, or 0.13 mil of generalized
corrosion in the fuel circuit.

8.3 ADJUSTMENT OF THE UF, CONCEN.
TRATION OF THE FUEL SALT

The fuel salt, free of moisture and HF, should
remove chromium from Hastelloy N only by the
equilibrium reaction

L 0 . o
,'2CI' + UF4 —_— /2ch 2

+UF3
(d)

(d} {d)

When the above corrosion equilibrium was first

established in MSRE power operations, the UF, .
produced in this reaction, together with that -
originally added to the fuel concentrate, should

have totaled 1500 g, with the result that as much

as 0.65% of the uranium of the system could have

been trivalent soon after the beginning of power

operation. The UF, content of the MSRE fuel

was determined after approximately 11,000 Mwhr

of operation to be no greater than 0.05%. The

fuel salt was considered to be far more oxidizing

than was necessary and certain to become more

so as additional power was produced unless ad-

justment was made in the UF, concentration. A

program was initiated early in 1967 to reduce 1 to

1.5% of the uranium inventory to the trivalent

state. The U®"* concentration has been increased
111

since that time by addition of 84 g of beryllium
metal. The method of addition was described
previously. ®

The low corrosion sustained by the MSRE fuel
circuit, which is in general accord with the re-
sults from a wide variety of out-of-pile corrosion
tests, might have been expected to be greater
during the first 10,000 Mwhr of operations be-
cause the UF, concentration of the fuel was
markedly less than was intended. According to
Grimes,’ ““The lack of corrosion in the MSRE
melts which appear to be more oxidizing than
intended can be rationalized by the assumption
(1) that the Hastelloy N has been depleted in Cr
(and Fe) at the surface so that only Mo and Ni
are exposed to attack, with Cr (and Fe) reacting
only at the slow rate at which it is fumished to
the surface by diffusion, or (2) that the noble-
metal fission products are forming an adherent
and protective plate on the reactor metal.”’

If it is assumed that corrosion of the fuel pro-
duced 3.3 equivalents of UF , and that fission has
resulted in the oxidation of 0.8 equivalent of UF,

 

®R. E. Thoma and W. R. Grimes, MSR Program
Semiann. Progr. Rept. Feb. 28, 1967, ORNL.-4119,
p. 123,

7W. R. Grimes, Chemical Research and Develop-
ment for Molten-Salt Breeder Reactors, ORNL-1853,
p- 70 (June 6, 1967).

per gram-atom of fissioned uranium,® the addi-
tions of beryllium metal to the fuel have been fol-
lowed by the concentrations listed in Table 8.3.
The calculated values ate for the most part higher
than the measured values. The cause of this
disparity is not currently understood, but is under
investigation.

All dissolutions of beryllium metal into the fuel
sall proceeded smoothly; the bar stock which was
withdrawn after exposure to the fuel was observed
to be smooth and of symmetrically reduced shape.

No significant effects on reactivity were ob-
served during or following the beryllium additions,
nor were chemical analyses indicative that such
additions were made until after run 12 was begun.
The additions preceding run 12 had increased the
U3" concentration in the total uranium by ap-
proximately 0.6%. Four exposures of beryllium
were made at close intervals during the early
part of that run. Samples taken shortly after the
last of these four exposures (FP12-16 ef seq.,
Table 8.1) began to show an unprecedented in-
crease in the concentration of chromium in the
specimens, followed by a similar decrease during
the subsequent sampling period.

Previous laburatory experience has not dis-
closed comparable behavior, and no well-defined

 

81bid., p. 65.

Table 8.3, Concentration of UF3 in the MSRE Fuel Salt?

 

 

Uranium  Uranium yl * Net ul +/EU u? +/EU
Sample Mwhr Burned Burned  Oxidized Be Added Be Added Equivalents Calculated Analysis
No. (@)  (moles) (moles) )  (equivalents)  poguced %) ()
9-4 10,978 554 2.34 1.87 0 0 3.13 0.31 0.1
10-25 16,450 829 3.50 2.80 16.28 3.61 5.81 0.58 0.5
11-5 17,743 953 4.02 3.20 16.28 3.61 5.21 0,53 0.37
11-13 20,386 1029 4.34 3.46 27.94 6.20 8.74 0.88 0.42
11-32 25,510 1287 5.43 4.34 27.94 6.20 6.86 0.69 0.34
11-38 27,065 1365 5,76 4.60 27.94 6.20 6.60 Q.66
11-49 30,000 1514 6.39 5.10 36.34 8.06 7.96 0.80
12-6 32,450 1637 6.91 5.50 36.34 8.06 7.76 0.77 0.37
12-11 33,095 1670 7.05 5.60 54.11 12.01 11.40 1.14 1.2
12-21 35,649 1798 7.59 6.10 74.12 16.45 15.35 1.5 0.5

 

“These numbers assume that the salt originally was 0.16% reduced; that the increase in Cr from 38 to 65 ppm was

real, occurred before 11-14-66, and resulted in reduction of U

+
atom of U3+; and that there have been no other losses of U3

+ + . .
4 to U3 ; that each fission results in oxidation of 0.8
mechanism is available which satisfactorily
A possible
cause, which is partially supported by experi-

accounts for the observed behavior.

mental data, is described as follows.

On the two occasions when the most rapid rates
of dissolution of the beryllium rods were observed,
chromium values for the next several fuel samples,
FP11-10 et seq., and FP12-16 et seq., rose tempo-
rarily above the lo level and subsequently re-
turned to normal. That the increase in chromium
levels in samples FP12-16 to -19 was temporary
indicates that the high chromium concentration
of fuel samples removed from the pump bowl was
atypical of the salt in the fuel circuit and implies
that surface-active solids were in suspension at
the salt-gas interfaces in the pump bowl.

That atypical distribution of species in this
location does indeed take place was demonstrated
earlier by the analysis of sample capsule support
wires that were (1) submerged below the pump-
bowl salt surface, (2) exposed to the salt-gas
interface, and (3) exposed to the pump-bowl cover
gas. The results showed that the noble-metal
fission products, Mo, Nb, and Ru, were deposited
in abnormally high concentrations at the salt-gas
interface. Such behavior suggests that the high
chromium concentrations in the fuel specimens
were caused by the occurrence of chromium in the

 

Fig. 8.3. Surface Appearances of Fuel Salt Specimens Taken Before and After Beryllium Addition.

(b) FP12-57.

112

R-39267

pump bowl in nonwetting, surface-active phases
in which its activity was low. A possible mecha-
nism which would cause such a phenomenon is

the reduction of Cr?" by Be® with the concurrent
reaction of Cr? with graphite present on the salt
surface to form one or more of the chromium car-

bides, for example, Cr3C2 (AHOf: —21 keal at

298°K). Such phases possess relatively low
stability and could be expected to decompose,
once dispersed in the fuel-circuit salt.

The possibility that surface-active solids were
formed as a consequence of the Be? additions was
tested late in run 12 by obtaining salt specimens
at the salt-gas interface as well as below the sur-
face. First, specimens were obtained in a three-
compartment sample capsule that was immersed
so that the center hole was expected to be at the
interface. Next, a beryllium metal rod was ex-
posed to the fuel salt for 8 hr with the result that
9.71 g of beryllium metal was introduced into the
fuel salt. Twelve hours later a second three-
compartment capsule was immersed in the pump
bowl. Chemical analyses of the fuel salt speci-
mens FP12-55 and -57 (Table 8.3) do not show
significant differences in chromium; however, the
salt-gas interface in FP12-57 is blackened as
as compared with FP12-55 (see Fig. 8.3).

R-39266

 

(a) FP12-55,
113

 

Fig. 8.4. Appearance of Upper Part of Three-Compartment Sample Capsule FP12-57 After Use.
An additional purpose of sampling with the
three-compartment capsule was to detemine
whether foamlike material was present in the
sampler area and would be collected in the upper
compartment. However, analysis of the upper
compartment has not yet been performed. Glob-
ules were noted on the upper part of FP12-57
(Fig. 8.4), indicating that conditions in the pump
bowl were different when samples FP12-57 and
-55 were obtained.

Examination of the metal basket which con-
tained the beryllium rod while it was exposed to
the fuel showed the presence of dendritic crystals
along with a small amount of salt residue (Fig.
8.5). Spectrochemical analysis of material re-
moved from the basket (Fig. 8.6) indicated that
the material contained 7.8 wt % chromium and
less than 10 ppm of iron and nickel.

The evidence obtained to date does not permit
inference as to the identity of the phases which
have formed within the pump bowl as a conse-
quence of the beryllium additions. It does strong-
ly imply that nonwetted flotsam can be formed and
accumulated temporarily in the MSRE pump bowl.

R-39264

 

114

The overflow tank has not been mentioned as a
possible factor in the behavior of chromium fol-
lowing a beryllium addition. It could, however,
bear some responsibility for the persistence of
chromium for several days following an exposure
of the salt to beryllium. Fuel salt accumulates
in the overflow tank steadily during operation and
remains there in relative isolation from the fuel
stream. At intervals of about one day, part of the
salt (60 1b) is retumed to the fuel stream. Recog-
nizing that chromium might be injected into the
pump bowl as the salt retums, we performed an
experiment in which salt samples were obtained
from the pump bow! within an hour after fuel was
retumed from the overflow tank to the pump bowl.
The purpose of the experiment was to determine
whether material from the overflow tank con-
tributed appreciably to the perturbations in the
chromium concentration. The results were nega-
tive, possibly, in part, because sampling and
salt transfer operations were not performed con-
currently for safety reasons.

Fig. 8.5. Dendritic Crystals and Salt on Basket of
Beryllium Addition Capsule FP12-56.
115

R-39265

 

Fig. 8.6. Residue from Beryllium Addition Capsule FP12-56.
9. Fission Product Behavior in the MSRE

S. S. Kirslis

Results of previous tests in-pile and in the
MSRE have been very reassuring regarding most
aspects of the chemical compatibility of graphite
and Hastelloy N with molten fissioning sait. The
aspect of chemical behavior currently causing
some practical concem is the observed tendency
of noble-metal fission products (Mo, Ru, T¢, Te,
and Nb) to deposit on graphite surfaces exposed
to fissioning fuel in the MSRE. Some of the iso-
topes of these elements have neutron cross sec-
tions high enough to affect significantly the
neutron economy of an MSBR after several years
of operation if a large fraction of these isotopes
deposited in the graphite core. Recent work in
the MSRE has been directed mainly toward eluci-
dating the behavior of these noble-metal fission
products.

Information derived from several kinds of tests
is reported in some detail in the following sec-
tions. These include (1) quantitative measure-
ments of the concentration of fission product
species in samples of the MSRE pump bow] cover
gas captured during a number of diverse reactor
operating conditions, (2) analysis of fuel samples
for fission products, (3) examinations and analyses
of graphite and Hastelloy N specimens exposed to
molten fissioning fuel salt in the MSRE for 24,000
Mwhr of power operation, and (4) qualitative meas-
urements of fission product deposition on metal
and on one set of graphite samples exposed to the
cover gas and fuel phases in the MSRE pump bowl
under a variety of reactor operating conditions.

9.1 FISSION PRODUCTS IN MSRE COVER GAS

Sampling of the cover gas in the MSRE pump
bowl has been continued in an effort to define the
nature and the quantity of the gas-borne species.

F.

116

F. Blankenship

Five gas samples were obtained and analyzed
during the period covered by this report. In these
studies we attempted to determine how the nature
and amount of gas-borne species are affected by
(1) addition of beryllium to the fuel, (2) stopping .
the MSRE fuel pump, (3) a long reactor shutdown,
and (4) increasing the volume of helium bubbles

in the fuel.

All samples were taken in evacuated 20-cc
capsules sealed by fusible plugs of 2LiF - BeF ;
these plugs melted and permitted the capsules to
fill with the cover gas upon insertion into the
heated pump bowl. Samplers, sampling procedures,
and analytical operations were, in each case, very
similar to those previously described.!

Since it was not possible to instrument the
sampling assemblies, no definitive information
concerning the temperature of the gas at time of
sampling is available. Temperatures wete cer-
tainly higher than the melting point of the fusible
plug (~450°C). It seems likely that the tempera-
ture is near 600°C and that the capsule actually
drew 20 cc of gas at this temperature and 5 psig;
the sampled volume of gas at STP was, therefore,
about 8.5 cc.

The results obtained from these five samples
are shown in Table 9.1. All values (except those
for uranium) are given in disintegrations per
minute for the total sample, and all have been
corrected to correspond to the time of sampling
if the reactor was operating, or to time of shut-
down for the two cases where the reactor was
not at power. The uranium values are in micro-
grams of uranium per sample. Conditions of re-
actor operation and special features of each test
are indicated in Table 9.1.

 

1s. s, Kirslis, MSR Program Semiann. Progr. Rept.
Aug. 31, 1966, ORNL-4037, p. 165.
117

Table 9.1. Fission Products in MSRE Pump Bow! Gas as Determined from Freeze Valve Capsules

 

 

 

Experiment No. FP11-42 FP11-46 FP11-53 FP12-7 FP12-26
Sampling date 4/11/67, 02:49  4/18/67, 02:28 5/2/67, 10:43  6/2/67, 06:50 7/17/67, 06:03
Operating time, days On 65, off 1.5 hr  Off 14, on 72 Off 14, on 86 On 92.3, off 42.5 Off 46, on 23
Nominal power, Mw 0 7.2 7.2 0 7.2
Be addition After 8.40 ¢ No No No After 37.8 g
Features Pump off 1.2 hr Regular Helium bubbles Power off 42.5 days Regular
Accumulated Mwhr 27,000 29,100 31,700 32,650 36,500
Isotope Yield Disintegrations per Minute in Total Sample
%10 6.06 1.05 x 10! 2.31 x 10! 1.57 x 1011 2.74 % 1011
1030, 3.0 2.51 % 107 4.64 % 10° 1.12 x 10" 4 % 10°?
106p, 0.38 8.2 % 107 9.49 x% 107 4.03 % 108 1.7 x 10°®
132 ~4.7 1.15 x 10! 3.35 % 101! 1.88 % 101 4.17 % 107 3.16 x 10'7?
12%pg 0.35 7.98 x 10% 3.51 x 107 2.17 x 10 6.6 x 103
5NL 6.2 6.45 % 108 1.3 x 107 1.05 % 161° 2.26 % 1092 3.52 x 10°
957, 6.2 <4.4 =% 107 ~2 % 107 1.8 % 108 8.64 % 107 2.98 % 10”7
1405, 6.35 6.16 x 10°
1317 ~3.1 5.7 % 10° 9.81 x 108 8.63 % 10° 1.67 x 101'°
895y 4.79 2.13 x 108 4.08 % 107 3.72 % 10° 8.35 x 105 3.71 x 10°
111 8 . 8
Ag 0.019 5.5 x 10 1.33 < 10
1410, ~6.0 4.79 x 10°
14404 ~6.0 4.17 x 107
2350 (ue 59 9.2 23 25

 

The considerable scatter in the data for ?°Zr
probably reflects scatter in the quantity of salt
mist in the sampling region and trapped in the
gas samples. The largest value (1.8 x 103 dis/
min of ?3Zr) was found under conditions in which
helium bubbles in the fuel were at a maximum.

If this ?°Zr were present as fuel mist, it would
represent about 1.3 x 1072 g of salt per sample
or about 1.6 x 10™* g of salt per cubic centimeter
of helium in the sampling region. In all other
cases studied the ?*Zr, and presumably the mist,
was two- to tenfold more dilute,

The values for 50.5-day ®°Sr, which is born
from 3.2-min 8%Kr, afford an opportunity to check
whether the mist shield which encloses the sam-
pling station permits equilibrium mixing with the
rest of the gas in the pump bowl. The values for
8951 are nearly an order of magnitude higher in
every case than are those for *3Z1. A small

fraction of the ®°Sr probably arises from salt
mist; it seems vittually certain that most of it
arises in the vapor phase through decay of the
89Kr. In the three gas samples taken while
MSRE was at power, the 8%Sr activity in the
samplers averaged 3.8 x 10° dis/min.

The total number of #7Sr atoms per standard
cubic centimeter of gas deduced from calculating
the 29Sr counting rate back to the time of sam-
pling was, therefore, 4,7 x 1013, If the sample
as taken is assumed to represent a uniform mix-
ture of the pump bowl gas containiag **Kr and
its daughters ®°Rb and #°Sr formed since the
B9Kr left the salt, then the ®9Kr was being
stripped at 2 x 1017 atoms/min or at some 31%
of its production rate. If, on the other hand, it
is assumed that all the ®*°Rb and ®°Sr are washed
back to the salt phase by the spray from the spray
ring, then the quantity of ®*°Kr that must have
been stripped is more nearly 50% of the production
rate,

With the simplifying assumptions that (1) no
89Kr is lost to the moderator graphite, (2) no
89Kr is lost through burnup, and (3) the fuel
which flows through the pump bowl is stripped
of 89Kr with 100% efficiency, it may be simply
shown that

89Kt stripped/min

F

89Kr produced/min_ A+ F

 

where F is that fraction of the fuel volume which
passes the stripper (pump bowl) per minute and A
is the decay constant (0.693/3.2 min) for 8?Kr.
For MSRE, with the pump bowl flow rate at 4% of
total flow, this expression yields the value 29%
for the 89Kr lost to the stripper gas. It is un-
likely that the stripping efficiency in the pump
bow! is 100%; moreover, it is certain that some
fraction of the 8?Kr is lost by penetration into
the graphite. This rate of loss of 8°Kr to the
moderator — which is probably less than 30% of
the production rate — will in effect introduce a
third tern to the denominator of the equation and
lower the fraction lost to the off-gas. From these
data, therefore, it seems possible that the gas

at the sampling station may be more concentrated
in 29Sr than is the average gas in the pump bowl,
but the concentration factor is not large.

The absolute amounts of ?°Mo, corrected back
to sampling time or time of previous shutdown,
show rather minor variations with reactor op-
erating conditions. The lowest value is for run
F'P11-42, in which the sample was taken 1.5 hr
after shutdown and 1.2 hr after stopping the fuel
pump. The normal runs showed the higher ?°Mo
concentrations, although the highest value might
have been expected for FP11-53, in which the off-
gas pressure was suddenly lowered just before
sampling to ensure a release of helium bubbles
from the pressurized graphite bars into the cir-
culating fuel. From the amount of ?°Mo found in
these samples (mean value 1.9 x 10'! dis/min or
2.2 x 1019 dis min~! ce~! or 1.3 x 10'* atoms
?9Mo/cc), the effective partial pressure of the
volatile species is calculated to be 4 x 10~ °
atm. The total ??Mo lost at 4.2 liters/min of
helium flow is 7.8 x 1029 atoms/day. This is
ahout 18% of the total inventory of 4.4 x 102!
atoms, or about 70% of the daily production rate.
If it is assumed that the material is lost as MoFfi,

118

this corresponds to a loss of 4.6 x 102! fluorine
atoms/day, or an equivalent of F~ per 150 days.
The very considerable quantities of ®?Mo found
on the graphite and the Hastelloy surveillance
specimens as well as in the fuel stream do not
seem consistent with such large losses to the
gas system.

The 193Ru, 1%°Ru, 129Te, and '3?Te concen-
trations generally showed parallel behavior over
all the runs, with particularly good correspondence
between the 23Ry and '9%Ry values. The ratios
for the two isotopic pairs agreed satisfactorily
with the ratios of fission yields divided by half-
lives.
stoppage was minor on this set of isotopes. The
highest values for '93Ruy, '9%Ru, and '?%Te were
obtained during the pressure release run, FP11-53.
However, 132Te, like ??Mo, did not show a high
value during this particular test.

The effect of short shutdown and pump

The appreciable concentrations of #°Sr found in
samples FP11-42 and FP12-7 are puzzling. In
each case the reactor had been shut down long
enough to permit the complete decay of 89Kr be-
fore the sample was taken. The ?°Zr analyses
indicated fuel mist concentrations too low to
account for the 3°Sr values. It may be that other
activities interfered with the beta counting re-
quired for 3%Sr analysis. If these values are
accepted as real (and further sampling must be
done to confirm them) it would appear that, at
least when the pump is off or the reactor is not
at power, the sampling volume is poorly flushed
by the cover gas system.

The ?°Nb values showed a fivefold increase of
magnitude from the first run to the fifth, with a
marked peak at the pressure release run, FP11-
53. Not all of the overall increase could be as-
cribed to the increase of °Nb inventory in the
reactor system with continued reactor operation
(cf. moderate increases in !°3Ru and '?°Ru).

The analyses for ?3°U from four of these runs
yielded values higher by at least a factor of 10
than those observed in earlier tests.! These
relatively high values for 233U are, at best, dif-
ficult to explain. If, as noted above, the values
for °°Zr are taken to indicate the existence, and
quantity, of salt mist in the sample, then sample
FP11-53 might have been expected to have about
20 pg of 35U in such mist. However, in no other
case of those shown in Table 9.1 can such an ex-
planation account for more than 20% of the observed
uranium.
The losses indicated by these samples are ap-
preciable. The mean of four samples shows nearly
30 pg/sample or about 3.5 pg per cubic centimeter
of helium. This would correspond to somewhat
less than 20 g/day of 275U or nearly 60 g/day of
the uranium in the fuel. (The figure remains at
nearly one-half this level if the highest figure is
rejected,) It seems absolutely certain that the
reactivity balances on the reactor through 250
days of full-power operation would have speedily
disclosed losses of far smallet magnitude than
these. It is difficult to conceive of a mechanism
other than volatile UF  to account for losses of
uranium to the vapor. It is, however, extremely
difficult to see how losses of this sort could be
reconciled with the nuclear behavior (and the
chemical analyses) of MSRE. IFurther samples
of the gas system will be made to help resolve
this apparent dilemma.

The effect of short shutdown and pump stoppage
(FP11-42) on noble-metal volatilization was in-
significant. This suggests that the volatilization
process does not depend on fuel nor bubble cir-
culation nor on the fissioning process itself, but
is probably a local phenomenon taking place at the
fuel-salt—gas interface. Beryllium additions to the
fuel had no discernible effect on the volatilization
of noble-metal fission products. This fact is rein-
forced in some detai! (see subsequent sections of
this chapter) by data obtained by insertion of
getler wires into the pump bow! vapor space.

The effect of reactor drain and long shutdown
(FP12-7) was consistently to lower observed
noble-metal volatilization (observed activities
were, of course, calculated back to the time of
previous shutdown). This observation is con-
sistent with all previous data on the effect of
long shutdowns. Presumably an appreciable
fraction of the noble metals is deposited on the
walls of the drain tank and reactor system during
a long shutdown. The minor effect of short shut-
down and the incomplete deposition during long
shutdown seem to indicate a slow deposition
process,

The pressure release experiment (FP11-53) re-
sulted in an appreciable increase of volatilization
of most fission products but not of %Mo or 32 Te,
Niobium-95 showed the most distinct increase. It
is particularly puzzling that the '?°Te concentra-
tion increased markedly while the '3?Te concen-
tration fell slightly. Clearly, further experiments

119

of this kind would be required to permit definite
interpretations.

It cannot be claimed that the mechanism for
volatilization of many of these wmaterials is known.
As has been shown elsewhere! ? the volatile flu-
orides of many of these materials (notably Mo, Ru,
Te, and, probably, Nb) are too unstable to exist
as equilibrium components of a fuel system con-
taining appreciable quantities of UF . It is true
that, for example, Mo formed at about 5 x 10%?
atoms/day probably exceeds greatly the solubility
limit for this material in the salt. Since it is born
in the salt in an atomic dispersion, it may, in ef-
fect, behave as a supersaturated solution; that is,
the activity of molybdenum may, in fact, be mark-
edly greater than unity. It seems most unlikely
that it can have activities greater than 1019,
which would be required if MoF . were to be stable.
Such findings as the presence of 111Ag, for which
it is difficult to imagine volatile fluorides, seem
to render the volatile {luoride mechanism more un-
likely. It still seems more likely that an ex-
planation based upon the finely divided metallic
state will prove true, though such explanations
also have their drawbacks.

9.2 FISSION PRODUCTS IN MSRE FUEL

The program of sampling and analysis of samples
of MSRE tuel for fission product species has been
continued by use of apparatus and techniques de-
scribed previously.' =3 Table 9.2 shows data for
16 fission product isotopes and for 23°Np obtained
from a series of four samples from a long stable
run of MSRE, from a sample obtained shortly after
shutdown in this run, a sample after the salt had
been cooled for 42.5 days, and a sample from a
relatively early stage in the subsequent power run.

Captions in this table show the operating history
of MSRE and the special conditions prevailing in
the reactor at the time the sample was taken.
Table 9.2 also includes estimates of the fraction
of the isotope represented by these analyses in
the 4860 kg of circulating fuel,

 

23, 8. Kirslis and F. F, Blankenship, Reactor Chem.
Div. Ann. Progr. Rept. Dec. 31, 1966, ORNL-4078, p.
48,

33. 8. Kirslis and F. F. Blankenship, MSR Program
Semiann. Progr. Rept. Feh. 28, 1967, ORNL-4[19, p.
124,
 

 

Table 9.2. Fission Products in Fuel Samples
Sample No. FP11-45 FP11-51 FP11-52 FP11-34 FP11-58 FP12-6 FP12-27
Accumuiated Mwhr 29,000 30,800 31,250 32,000 32,650 32,650 36,550
Mominal power, Mw 7.2 7.2 72 7.2 0 0 7.2
Operating time, days® Oif 14, an 71 Off 14, on 82.5 Off 14, on 85 Off 14, on 89 On 92.3, off 9.1 0n 92.3, off 42.5 Off 46, on 23.5
Exposure time, min 1 1 1 2 i P 6

 

 

 

 

Features No purge,” after 8.4 g Minimum leve! Regular Regular Pump off 2 hr, Before startup. After 38 g of
of Be added in pump bowi power off 2.3 hr after shutdown Be udded
Fission o Percent S Pe;;rI{ Percent Percent Percent Per.cent Percent
Isotope Yield dis min~' g~ ! of dis min™! g7 ! of dis min~! g ! of dis min—!' g ! of dis min=! g~ ! of dis min' ! g7! of dis min~! ¢! of
(%) Total® Total® Total€© Total® Total® Total® Total®
67-hr *Mo 6.06 1.53 - 194! 108 7.22 . 1010 5% 8.17 . 1910 58 3.50 < 1010 25 3.19 . 10'° 23 Decaved 1.24 . 104! 88
39,7-day '%?Ru 3.0 1.40 » 1010 23 3.8 .10° 7.1 8.86 . 1010 165 1.82 . 10° 3 9.5 . 10° i7 6.9 . 10° 12 5.55 . 10° 23
1.02-year '?°Ru 0.38 ~4 . 108 ~3 . 108 2.38 . 108 7.5t . 107 3.6 « 10® 3.9 . 108 2.43 . 10%
78-hr 132Te ~ 4,7 3.50 . 10!° 32 1.68 « 101° 15 1.56 - 10° 14 1.07 . 10!° 10 8.39 . 10¢ 8 Decayed 1.27 . 1910 12
37-day '¥7Te 0.35 5.20 . 10% 3 5,00 - 10° 8 5.45 - 10% 8 2.06 -« 10° 3.1 .108 i0
35-day *SNb 6.2 2.61 . 101° 4.0 . 10° 1.91 ~ 10° 311 1010 2.15 . 1010 8.04 . 101° 1.91 . 10'°
65-day *%Zr 6.2 2.97 . 10192 1.20 . 10%! 1.34 . 101 .34 . 101! 1.54 . 101! 1.46 . 101! 1.00 . 107¢
12.8-day '*°Ba 5.35 1.67 . 10! 115 1.82 . 10" 125 2,17 .0t 149 1.74 . 101! 119
50.5-day *°Sr 4.79 8.6 . 1010 123 8.02 . 19'° 106 8.11 . 101° 106 1.02 101! 128 8.89 . 10%° 111 6.07 - 101 116
9.7-hr °'Sc 5.81 1.29 . 10"! 95
33-hr '*2Ce 5.7 1.14 . 101! 86 7.67 . 10'° 53 1.78 . 101! i34
33-day *!Ce ~6,0 1.48 « 10'! 129 1.4 . 101! 142 1.71 « 10! 143 1.37 . 101! 115
260-day '*4Ce ~6.0 6.36 « 10'° 6.05 « 101? 6.24 « 10'° 6.01 . 10'0
7.6-day '''Ag 0.019 5.1 x 107 12 6.5 - 107 15 4.9 . 107 11 1.63 - 107 4
3hr P 2Ag 0,019 8.0 « 107 34 8.9 . 107 38
8.05-cay '?'1 -3 9.17 - 1017 128 8.20 - 10'° 114 8,22 . 12 ° 115 6.97 - 10'° 97 1.15 « 10%° 16 7.12 . 10t° 114
5.6 - 107 ! 3.29 « 1017

2.33-day **°Np

“Duration of previous shutdown and of continuous operating time just before sampie was taken; vice versa

5.36 -

 

bUsually a 200-cc/min gas purge was passed down the sample line,

¢Percent of calculated total amount in reactor based on fission yields and power history.

for samples taker after shutdown.

0c1
121

As in previous analyses of the reactor fuel, it
is clear that a very high fraction of isotopes (such
as '4%Ba, 89Sr, 1%1Ce, 1*%Ce, and °!Sr) that form
very stable and soluble fluorides is present in the
circulating fuel. As judged by the behavior of
13171 the isotopes of iodine also seem to be present
in a high fraction.

On the other hand the more noble elements, such
as Mo, Nb, Te, and Ru, are generally present in
much lower fractions. It is worthy of note that
these materials seem to show a much wider scatter
than do the ‘‘salt seekers’’; this may indicate that
they are not present in true solution. It must also
be noted that these noble metals are present in
considerable quantity in the gas stream and that
they adhere tenaciously to metal specimens ex-
posed to them. It is possible, therefore, that a
considerable fraction of the material found in the
“‘salt’’ is due to the fact that the sampler is low-
ered and then raised through this gas phase. It is
possible that the relatively high values obtained
for %Mo, 19%Ru, and 132Te in FP11-45, where
no gas was used, may be due to this phenomenon.
This possibility will be verified, if possible, ex-
perimentally in the near future.

6.3 EXAMINATION OF MSRE SURVEILLANCE
SPECIMENS AFTER 24,000 Mwhr

A second set of long-term surveillance speci-
mens of MSRE graphite and Hastelloy N was ex-
amined after exposure to citculating molten-salt
fuel in an axial core position of the MSRE for
24,000 Mwhr of power operation. During the last
92 days of this exposure, the MSRE operated at a
steady power of 7.2 Mw essentially without in-
terruption, whereas power operation was frequently
interrupted during the previous 7800-Mwhr expo-
sure. Examinations of the specimens gave results
very much like those previously reported after a
7800-Mwhr exposure.! ™3

Examination of Graphite

As before, three rectangular graphite bars were
available for examination: a 9.5 x 0.47 x 0.66 in.
bar from the middle of the core, and 4.5 x 0,47 «
(.66 in. bars from the bottom (inlet} and top (out-
let) of the core. Visually, the graphite appeared
undamaged except for occasional bruises incurred
during the dismantling. The two 4.5-in.-long bars

gained 0.002 t 0.002 in. in length and 13 mg (0.03%)
in weight. Metallographic examination showed no
radiation or chemical alteration of the graphite
structure and no evidence of surface films.

X radiography of thin transverse slices showed
occasional salt penetration into previously exist-
ing cracks that happened to extend to the surface
of the sample. Similar penetration was observed
in control samples that were exposed to molten
salt but were not irradiated. X-ray diffraction by
the graphite surface exposed to fuel gave the
normal graphite lines except for a very slightly
expanded lattice spacing. A few weak foreign
lines that were probably due to fuel salt were
observed. Autoradiography of the samples showed
a high concentration of activity within 10 mils of
the surface, with diffuse irregular penetration to
the center of the cross section. These observa-
tions confirmed previous demonstration of the
satisfactory compatibility of graphite with fis-
sioning molten salt as far as damage and per-
meation by fuel are concerned.

Concentration profiles for fission products in
the graphite were determined by milling off layers,
which were dissolved and analyzed radiochemically.
Near the surface, layets as thin as 1 mil were ob-
tained; farther in, layers as thick as 10 mils were
collected until a total depth of about 50 mils was
reached. The predominant activities found were
molybdenum, tellurium, ruthenium, and nicobium.
These elements can be classed as noble ele-
ments since their fluorides (which are generally
volatile) are relatively unstable, The practical
concern, because of long-term neutron economy,
is with Mo, *"Mo, °9Tc, and '°'Ru, but these
particular isotopes are either stable or long-lived
and thus were not amenable to direct analysis.
However, we considered as sound the assumption
that their deposition behavior was indicated by
either that of a radicactive isotope of the same
element or that of a radioactive noble-metal pre-
cursor with a suitable haif-life.

Over 99% of the noble-metal activities were en-
countered within the first 2 or 3 mils of the
graphite surface; the same was true for °Zr. This
is illustrated in Figs. 9.1-9.4. The values for
blanks shown in the figures were obtained from
samples of about 0.2-g (1 mil) size that were
milled from fresh unexposed graphite in the hot
cell and with the equipment used for the irradiated
specimens. The blanks show that the apparent
tails on the curves for Mo, ?3Nb, and '°°Ru are
 

1044

 

109 4

 

 

-

Q
-
n

 

disintegrafions per minute per grum

 

 

2
1010
5
2
10° L
0 10 20 30 40 50
DISTANCE FROM SURFACE (mils)
Fig. 9.1. Distribution Profile of %0 in Grephite
from MSRE Core.
meaningless. Figure 9.4 indicates that there
probably was measurable penetration of ?5Zr to
a depth of 50 mils.
Figure 9.5 shows that 132Te concentrates, as

do the noble metals, at the surface; however, a
measurable concentration of this element pene-
trates to a depth of 50 mils. The behavior of
fission products with rare-gas precursors is
shown in Figs. 9.6 (1*%Ba) and 9.7 (3°Sr).

Because of the concentration at the surface,
the amounts present in the several layers near
the surface were representative of the total

122

 

  
  

 

 

 

 

 

'3 ORNL--DWE 67—12232
10 e — } - . T
Lo ]
5
|
2 ]
ol ]
5
5 2
o
2 o't _;
o e
5 R
c
E °r '"f,'j
@ _
o
E 2
5
£10° —
£
5 )
5 .......
2
10°
5
2 L
108 L ]
0 10 20 20 40 50
DISTANCE FROM SURFACE (mils)

Fig. 9.2. Distribution Profile of 95Nb in Grophite
from MSKRE Core.

amount of many isotopes in the graphite. Table
9.3 shows the quantity of several materials (in
disintegrations per minute per square centimeter)
on graphite specimens as functions of position
and of exposed face. In each case, face No. 3
was in contact with other graphite and was not
exposed to salt. In general, this face contained
about as much activity as those exposed to salt;
this finding supports the view that the deposited
activities were gas boine.

In Table 9.4 the values (averaged for all faces
of the graphites) obtained in the specimens ex-
posed for 24,000 Mwhr are compared with the
123

" ORNL—DWG §7--12233

 

 

 

 

 

 

 

1010 H— o —

* BLANK

5 4~ —_— —_ .

 

 

 

 

 

 

o
w N

w

 

 

n

 

o
®

disintegrotions per mingie per gram

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50
DISTANCE FROM SURFACE (mils)

Fig. 9.3, Distribution Profile of 106Ru in Graphite
from MSRE Core.

values previously obtained after 7800 Mwhr. It
is clear from this table that the data for °?Mo,
132Te, and '°?Ru are generally similar, while
the values obtained for °*Nb and °°Zr are ap-
preciably higher for the longer exposures. It
should be noted that the data for 3Nb are ob-
tained from gamma scans without chemical sep-
arations and are probably less certain than the
others.

The graphite samples accumulated by milling
these bars were analyzed for uranium by neutron
activation. Typical data from one of these bars

disintegrations per minute per gram

10

——— -y

12 ORNIL--DWG 6712234

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 20 40 50
DISTANCE FROM SURFACE (mils)

Fig. 9.4. Distribution Profiie of 951r in Graphite
from MSRE Core,

are shown as Table 9.5, It is clear from this
table that the 23°U is present in all faces of

the graphite and that the concentration diminishes
with the depth of milling. The small concentra-
tion in the blank (as in blanks from other speci-
mens) indicates strongly that these uranium con-
centrations are real.

The absolute quantity of 233U or of uranium in
the graphite moderator is apparently very small.
If these analyses are, in fact, typical of those
in the 2 x 2 in. moderator bars, then the average
concentration of ?3°U in and through the bars
124

ORNL—-DOWG ©67-12235

 

1013

 

disintegrations per minute per gram

 

 

 

 

20 30 40
DISTANCE FROM SURFACE (mils)

 

50

Fig. 9.5. Distribution Profile of 132Te in Graphite
from MSRE Core.

is much less than 1 ppm. Since the MSRE con-
tains about 70 ft® or slightly less than 4000 kg

of graphite, the total 2?°U in the moderator stack
is almost certainly less than 4 g. Since the MSRE
fuel is slightly more than 30% enriched in 23°U,
the total uranium in the moderator should be, at
most, 10 to 15 g.

Examination of Hastelloy N

The graphite specimens described above were,
as was the case for the specimens examined after
7800 Mwhr, contained in a cylindrical, perforated
holder of Hastelloy N. Samples were cut from this
cylinder as before! =3 and analyzed by standard
techniques for several fission products and for
2357y,

10"2 _

 

 

N

O‘
o

w

disintzgrations per minute per grom

n

102 |-

 

 

 

 

DISTANCE FROM SURFACE (mils)

Fig. 9.6. Distribution Profile of 1404, in Graphite
from MSRE Ceore.

Data are shown in Table 9.6 for seven fission
product isotopes and for uranium at five positions
from the top to the bottom of this Hastelloy N as-
sembly, Data are also shown for the comparable
samples examined after 7800 Mwhr of operation
in mid-1966.

It is clear from these data that the behavior of
?9Mo was similar in the two sets of specimens,
with the 24,000-hr exposures generally showing
somewhat more deposited activity. Behavior of
the !*2Te seems generally similar in the two ex-
posures. Somewhat more '°3Ru was found on the
24,000-Mwhr specimens than on those previously
examined, while ?°Zr (present in small amounts,
in any case) seems less concentrated on the
specimens exposed to the higher dosage. lodine-
131, which is present in appreciable concentra-
tions, shows some scatter with location of the
1o"

N

O-h-
o

disintegrations per minute per gram

DISTANCE FROM SURFACE (mils)

ORNL—DWG &7--12237

 

Fig. 9.7. Diswribution Profile of B¢ in Graphite

from MSRE Core.

125

specimen but about the same deposited activity
in the two series of tests.

The uranium values correspond in every case to
negligible quantities. It appears that, if these
analyses are typical of all metal surfaces in the
system, about 1 g of 233U may be deposited on or
in the Hastelloy of the MSRE.

Fission Product Distribution in MSRE

The computer code for calculation of total in-
ventory of many fission product nuclides in the
operating MSRE has not yet been completed. Ac-
cordingly, no detailed information as to expected
quantities of several isotopes is available. How-
ever, from the data given in previous sections of
this chapter approximate values can be given for
the distribution of several important fisston prod-
ucts among the salt, graphite, and metal phases
of the reactor system. These are presented in
Table 9.7. Such approximate values must, of
course, assume that the fission products found
on the relatively small samples from the surveil-
lance specimens in midcore are those to be found
on all graphite and metal in MSRE,

Since there is, as yet, no test of this assumption
we regard the values in Table 9.7 as tentative,

Table 9.3. Fission Product Deposition on MSRE Graphite Faces After 24,000 Mwhr

 

Fission Product (dis min~! cm"g)

 

 

Graphite
Location Face 66-hr 78-hr 37.6-day 1.02-year 35:day 65-day
99M0 132Te IDSRH IOGRU QaNb gSZr
x 1010 x 1010 x 10° x 108 x 10'° x 108
Top 1 2.16 1.45 4.54 2.26 5.00 2.26
2 3.59 1.75 8.86 3.58 6.15 3.58
3 2.76 1.74 6.97 2.77 4.77 2.77
4 6.30 3.52 15.9 3.08 g9.16 3.08
Middle 1 2.28 2.51 6.92 3.51 2.00 3.51
2 5.32 3.47 12.9 6.33 17.8 6.33
3 6.16 3.31 7.94 7.94
4 5.41 4.68 15.9 6.98 15.4 6.98
Bottom 1 3.24 1.79 6.95 3.11 11.6 3.11
2 8.05 4.50 18.6 8.04 30.6 8.04
3 2.85 2.22 9,72 3.64 14.2 3.64
4 3.43 2.14 8.65 3.74 13.3 3.74

 
126

Table 9.4. Fission Product Deposition on MSRE Graphite After 7800 and 24,000 Mwhr

 

 

 

 

 

Fission Product (dis min " ? cm“'z)a
Graphite 1.02-year
. 67-hr Mo 78-hr **?Te 39.7-day '?3Ry 106 35-day °°Nb 65-day 252Zr
Location _ _—~ 7 7 o T ) Rn ™™™ = ="
b ]
A B A B A B A B A R A B
x 101 x10'% x10'% x10'? x10!'0 %1010 « 1019 w10'%  x10'% x10'% x10!'®
Top 3.97 3.69 3.22 2.12 0.83 0.91 0.003 0.46 6.2 0.0004 0.001
Middle 5.14 4.79 3.26 3.82 0.75 1.19 0.006 2.28 1.17  0.003  0.007
Bottom 3.42 4.39 2.78 2.66 0.48 1.10 0.005 2.40 17.4 0.002  0.02

 

IMean values for all four faces.

b

A: value obtained after 7800 Mwhr, B:

value obtained after 24,000 Mwhr.

Tahle 9.5. Uranium in Millings from Top Graphite Bar in MSRE After 24,000 Mwhr

 

 

Cumulative Specimen 235y 235y
Samplle Sample Depth Weight Analysis in Specimen
IL.ocation Number (mils) () (ppm) (1)
Side 1 1 1.56 0.145 9 1.3
(facing salt) 5 4.04 0.224 2.1 0.47
7 5.51 0.132 2.4 0.32
9 7.50 0.172 0.80 0.13
11 10.75 £.2065 0.03° 0.0099
13 13.60 0.256 .72 0.19
17 17.40 0.340 (.48 0.16
21 21.75 .390 0.29 0.11
25 25.64 0.350 0.49 0.17
26 30.57 0.445 0.39 0.017 -
27 36.58 .540 .28 0.15
89 48.18 1.05 0.15 0.15 A
Side 2 5.58 0.360 8. 3.2 :
(facing salt) 8.60 0.100 3.0 0.57
11.84 0.210 3.6 0.76
10 14.03 0.140 2.6 0.36
12 16.62 0.165 1.4 0.23
14 24.44 0.500 1.6 0.80
Blank 0.28
18 31.90 0.480 0.99 0.48
22 40.94 0.580 1.33 0.77
Side 3 3 5.50 9.495 6. 3.3
(facing 15 8.79 0.300 1.2 0.36
graphite) 19 12.98 0.470 0.84 0.39
23 18.56 0.505 0.78 0.39

“Questionable value.
Table 9.6, Fission Product Depo

 
 

Isotope

Extreme Top

Bb

All values except

235¢7 in disintegrations per

B

sition on Hastelloy N in MSRE Core A

 

  

minute per sguare centim

 

Hastelloy N Location®

A

 

Middle

 

B

    
  
 

fter 7800 and After 24,000 Mwhr

eter of Hastelloy

   

   

Extreme Bottom
B

  

67-hr Mo .84 x 101 3.12 (10t 3.94xtott 2.76 x 1017 2.86 x 10! 2.04 x 10%? 4.42 x 1017 0.74 % 10%?
78-hr 102 Te 6.66 x 10%} 5.08 x 101 5.66 x 101* 3.41 x 10 2.06 x 10%} 4.07 x 10*1 402 % 10!t 1.94 x 101!
37.6-day ‘" Ru 0.09 x 1010 3.55X% 1010 10.6 x 1017 9.55 x 100 1.56 X 1010 2.32x 100 602X 101?

35-day J°Nb 1.04 x 101? 2.32 x 101 .32 x 10** 1.66 x 10%* 0.16 x 10*}
65-day °Zr 2.32 x 10° 18.5 x 10° 1.56 x 10° 18.4 x 10° 8.62 x 10° 25.8 x 103 4.10 x 108 0.32 x 103
50.5-day > Sr 1.69 % 16° 10.3 x 108 46.1 x 10° 5.93 x 10° 0.834 x 108
8.05-day ‘"1 1.60 x 10° 8.24 x 10° 3,93 x 10° 3.97 x 107 3,18 x 10° 5.24 % 10° 7.67 x 10°

235y (ug/cm® $1.3 0.92 0.61 0.61

LTt

1.51

 

 

aa11 surveillance specimens in center axis of core.
ba. after 7800 Mwhr, B: after 24,000 Mwhr.
128

 

Table 9.7. Approximate Fission Product Distributions? that had been exposed to the liquid phase, the
in MSRE After 24,000 Mwhr gas phase, or the interface region. In all cases,
e — e the nickel-plated key exposed at the very top of
Fuel Graphite Hastelloy N the gas phase was analyzed, but these results are
Isotope {(total (total (total not included here. Data from seven separate tests
dis/min)}  dis/min) dig/min) are shown as Table 9.8. In addition, the behavior

 

 

..... i of '°®Ru is shown in Fig. 9.8.

In virtually every case the specimens were
5 leached first with an aqueous chelating agent and
1320a 5.3 % 1018 5.9 x 10%° 4.1 x 1017 then with an oxidizing acid solution. Some iodine
and ruthenium may have been lost from the acid

b
%Mo 1.7 x 1017 8.6 x 1016 3.2 x 10%7

b
%Ry 2.9x 10 2.2« 101° 5% 1018
leach solutions, and it is not unlikely that some

of these values are low. Table 9.8 reports, in all
Wz, 7.3 x 10'7 1.8 x 1014 3.6 x 10t? cases, the sum of values obtained with both types

- b
95Nb 9.6 % 1010 1.6 x 1017 1.5 x 1017

of leach solution.

A nuinber of conclusions from the data, of which
those in Table 9.8 are typical, follow.
T T ' The deposition of noble metals on the pump bowl

8%, 4.9 % 1017 1.3 x 1013

131y 3.5 x 1047 4% 101°

 

Selected mean value from fuel analyses; assume con- specimens under normal reactor operating condi-
centrations on graphite and fuel specimens are typical of tions was quantitatively quite similar to that pre-

all such surface in core. ) 3
BMean of considerably scattered values. viously reported.® Added purge flow down the

ORNL-0OWG 67-12238

9.4 DEPOSITION OF FIS3ION PRODUCTS

  
   
  

tinuing series of studies of deposition of gas- SSI
bomne activities on metal specimens. The tests
carried out since the previous report have attempted
to observe the effect of duration of previous reactor
operation, addition of metallic beryllium, level of
liquid in the pump bowl, exposure time of speci-
mens, stopping of the reactor pump, length of shut-
down time of the reactor, and draining of fuel from
the reactor.

Exposure conditions were generally similar to
those used in previous tests of this kind,! ~3 but
in these studies the stainless steel cables which
hold the sampler were used without additional
getter materials. To reduce the increasing fis-
sion product contamination of the sampler-enricher
assembly, a downward flow of some 200 cc of
helium was used with all insertions after FP11-45.
The cables (in all cases except FP11-50, de-
scribed in some detail in the following section)
attached to standard samplers for obtaining spec- =
imens of fuel were exposed under conditions and FP NUMBER
for times shown in Table 9.2, The cables were
then segmented into approximately equal lengths Fig. 9.8. Deposition of 106y in Pump Bowl Tests.

FROM MSRE GAS STREAM ON ol e e

METAL SPECIMENS o sl dpm/g =+ ,,J

T e SSL l

Access to the pump bow! provided by the ~ o SS6 R
sampler-enricher tube has been used in a con- g Y (

toia! disintegrations per minute

107

12-27 ,[

-2z |
11-45 |
1450

5

#4-52 -
1-54
12-6 -
129

Table 9.8. Deposition of Yarious Fission Products on Metal Specimens lnserted in MSRE Pump Bowl

 

 

235
Exposure 0
a 09 131 103 132 95
Run No. Condition Mo I Ru Te Zr (lug/sarnple)
= 1011 < 1010 x 1019 x 1011 x 107
FP11-45 Liquid 0.65 0.46 0.2 0.9 0.035 21
Interface 1.8 1.0 1.4 1.3 10 20
Gas 1.1 1.1 0.2 0.35 0.80 5.2
FP11-50" Liquid 3.5 12 0.7 14 250 35
Interface 26 30 2.8 100 320 190
Gas 0.2 9.2 23 17 60
FP11-51 Liquid 0.3 0.1 0.13 0.14 1 3.5
Interface 0.85 0.85 0.5 0.52 1.5 19
(Gas 0.4 0.6 0.25 0.3 3 6.0
FP1i-54 L.iguid 0.42 0,22 0.25 0.17 b 6.4
Interface Q.72 0.70 1.0 0.61 3.5 5.6
Gas 0.42 0.55 0.40 G.28 .5 2.0
FP11-58 Liquid 2.05 0.11 0.13 0.17 0.3 8.4
Interface 0.15 0.28 0.15 0.07 0.6 3.3
Gas 0.18 0.60 0.23 0.04 0.7 2.5
FPr12-6 Liquid 0.11 0.82 22
Interface 0.17 0.60 46
Gas 0.38 0.30 a0
FPri2.27 Liguid 0.06 0.14 0.05 0.09 0.9 8.2
Interface 0.72 0.24 0.04 0.18 1. 5.6
Gas 0.23 0.37 0.22 0.40 1.3 6.6

 

ASampling conditions are those shown in Table 9.2.

b . . o - . . .
Special sample connecting two sets of graphite immersion samples immersed for 8 hr; experimental details are

in the following section.

sampler-enricher tube during exposures had no
significant effect on deposition on fresh metal
specimens. The coosiderable additional reactor
operating time also apparently had little effect,
except for possibly a slight decrease in °°Mo and
132T¢ deposition for the later runs.

The deposition of noble metals was usually
heavier on the interface sections of the stainless
steel cable than on either the gas-phase or fuel-
immersed sections. This suggests the presence
of a sticky scum or froth containing noble metals
on the surface of the fuel in the pump bowl.

As previously observed?® and as verified by the
results from sampling the MSRE gas stream, the
reduction of the MSRE fuel with beryilium caused
no discernible change in deposition behavior. This
observation is difficult to reconcile with the hy-

pothesis that deposition behavior is related to the
presence of volatile noble-metal fluorides.

The deposition of noble metals on the sections
of stainless steel cable immersed in the fuel
usually paralleled their concentrations in the {uel
(see Table 9.2),

The B-hr exposure (FP11-30) revealed curious
differences in the deposition behavior of the
various fission products. The deposition of 1??Te
and **Zr was about 100 times that observed in 1-
or 10-min exposures. The cotresponding factor
was 30 or less for 14%Ba, 10 or less for ?°Mo, 3
or less for 1°?Ru and '°*Ru, and less than 1 for
¥5Nb. Usually, comparisons between runs have
shown similar factors for all the noble-metal fis-
sion products. This unusual observation indi-
cates that the noble metals do not travel together
{e.g., in metallic colloidal aggregates) but deposit
individually by separate mechanisms.

In all but one of these tests, the exposure times
varied from 1 to 6 min. The amounts of noble
metals deposited during these intervals showed no
increase with exposure time (if anything a de-
crease) and were not significantly different from
the results of previous 10-min exposures. Even in
the case of 132Te, which deposited heavily in the
8-hr exposure, no significant differences were ob-
served for exposure times between 1 and 10 mia.
This and related information will be discussed in
more detail later.

The deposition behaviors of 40-day !Y3Ru and
1.0-year '9%Ru were fairly closely parallel. The
incomplete data for 33-day '2°Te also followed
well the variations among runs of the 77-hr 132 Te
deposition data. At production-decay equilibrium,
the number of disintegrations per minute of each
fission product should be proportional to its fis-
sion yield, For the tellurium isotopes the ratio
of activities deposited was about 20, while the
ratio of fission yields is 13.4. The ratio of ac-
tivities of the two isotopes in the fuel was also
about 20. The discrepancy between 20 and 13.4
may be partly due to the fact that the 12°Te may
not have reached equilibrium activity. In the case
of the ruthenium isotopes the much larger discrep-
ancy between the ratio of activities and the ratio
of fission yields is very probably due to the fact
that the one-year '°°Ru had not nearly reached
equilibrium activity.

Experiment FP11-51 was carried out with the
pump bowl level as low as possible, with the ob-
jective of introducing more than the usual amount
of circulating helium bubbles into the fuel. Nommal
deposition of noble metals was observed. There-
fore, the fraction of circulating helinm bubbles has
no effect on fission product deposition.

Experiment FP11-58 was carried out after 92.3
days of virtually uninterrupted power operation,
2.3 hr after shutdown, and 2 hr after stopping the
fuel pump. Under these conditions, the deposition
of noble metals decreased by a factor of 2 to 9,
most sharply for tellurium and least for ruthenium.
The concentrations in the salt (see Table 9.2) re-
mained nearly constant except for ruthenium, for
which the data show a fivefold increase.

r

I'he rather moderate decreases in noble-metal
deposition with the reactor shut down for 2.3 hr
and particularly with the fuel pump off are dif-
ficult to explain by the previously mentioned

metal colloid theory of deposition. Any suspended
species in the pump bowl cover gas should have
been swept out by the 4-liter/min helium flow
through the pump bowl gas space. The 3-in.-diam
by 6-in. volume inside the mist shield was addi-
tionally swept by the 200-cc/min purge flow

prior to and during the exposure. The result is
one of the most conclusive indications that tiuly
gaseous species originating from the fuel suriace
are responsible for the observed concentrations of
noble-metal fission products in the pump bowl gas
space.

Experiment FP12-6 was carried out after a re-
actor drain and refill after 42.5 days of reactor
shutdown and prior to resumption of power op-
eration. Only slightly lower than normal depo-
sition of noble metals was observed. After long
shutdowns in the past, sharp decreases in depo-
sition occurred. Also contrary to past experience,
the concentrations of noble metals in the fuel
either remained constant (ruthenium) or increased
(tellurium and niobium).

An additional experiment (FP14-1, not shown in
the table) was carried out after the MSRE had
been shut down for 38.1 days, then at full power
for 2.5 days, then shut down and drained for 2.0
days. The fuel pump continued to circulate helium
in the drained reactor with the temperature of the
pump bowl top at 700°F. Specimens were exposed
for 11 min. The only specimens available for
analysis were the gas-exposed stainless steel
cable and the key.

The analyses are not complete, but the avail-
able results are remarkable. Compared with the
previous ““normal’’ run (FP12-27), the activity
(calculated back to the end of the 2.5-day op-
eration) deposited on the cable was slightly -
higher for °°Mo, '9°Ru, and ?°Nb, distinctly
higher for '3!'I and 235U, slightly lower for
132Te and '°3Ru, and distinctly lower for
957r. Most of the activities on the key dropped
by a factor of about 10 from the previous run.
These high depositions are all the more re-
markable when it is considered that there was
a two-day shutdown before sampling and that
even the short-lived fission products could not
have reached equilibrium activities in the 2.5-
day period of power operation following the 38.1-
day shutdown.

It is not clear whether the observed deposits
originated from material previously deposited on
the graphite or metal walls of the reactor system
or whether it came from the 40 1b of fuel salt
which remains in the reactor system after a fuel
drain. Some of this fuel salt is in the form of
shallow puddles in the intemals of the fuel pump
where contact with the circulating helium should
be good. Another possibly pertinent fact is that
gas circulation through the reactor and pump bowl
with the pump operating in the drained reactor is
much faster than whea the reactor is filled with
fuel, Thus the pump bowl specimens may con-
tact a larger volume of gas per unit time. The
fuel may also act as a scrubbing fluid to remove
gaseous or suspended activities from the pump
bowl gas, even while it is the source of these
activities at a different time and place. Sorbed
gaseous activities or loosely deposited solids
from the reactor surfaces may conceivably move
to and stay in the gas phase more readily in the
absence of fuel salt.

A few analytical results are available from test
FP14.-2, which was run one day after test FP14-1
and after the reactor had been refilled with fuel
but before power operation was resumed. For the
gas-phase stainless steel cable specimen, the
deposition of %Mo, 132Te, and 31 was lower
than for FP14-1 by an order of magnitude. The
ruthenium and niobium activities increased
slightly, and that of ?°Zr decreased slightly.
The amount of 233U deposition increased by a
factor of 4. The decreases in activity are dif-
ficult to explain. The curious fact appears to
stand that deposition of noble metals in the
drained reactor is the same within an order of
magnitude as with fuel in the reactor,

Occasional radiochemical analyses were car-
ried cut on the leaches of the metal specimens
for two additional noble-metal fission products,
111A0 and '12Pd., These behaved like the other
noble metals, with heavy deposition on the gas-
phase and submerged specimens. This behavior
is of chemical interest since no highly volatile
fluorides of silver and palladium are known to
exist. This observation favors the gaseous
metal suspension theory of noble-metal volatili-
zation.

Occasional analyses were also made for other
alkaline-earth and rare-earth species such as
91Qy, 141Ce, 143Ce, 144Ce, and '*'Nd, These
species remained in the fuel melt and showed
light deposition on the pump bowl specimens.
Most of the pump bowl specimens were analyzed
for 235U deposition by delayed-neutron counting

131

of their leaches. Unusually high values were ob-
tained for the FP11-50 (8-hr exposure) specimens,
averaging 70 pg of 2*°U on each specimen. An-
other set of high values averaging 35 ug of 235U
per specimen was obtained on run FP12-6 (42.5-
day shutdown). The deposition of ?5Nb was also
unusually high for this run. The values for run
FP11-45 (normal) were also high, averaging about
30 pg of 23°U per specimen. The results for
2357 did not parallel those for any other fission
product, and the reasons for the wide variations
between runs are not known.

9.5 DEPOSITION OF FISSION PRODUCTS
ON GRAPHITES [N MSRE PUMP BOWL

Graphite specimens have been exposed in the
MSRE pump bowl to observe fisgion product depo-
sition under a known set of short-term conditions.
Such an exposure has been carried out for 8 hr
during steady 7.2-Mw operation of MSRE.. Since
this exposure, with the sampler-enricher tube
opened to its separate containment for 8 hr, posed
some problems in reactor operation, the test was
designed to yield a considerable amount of infor-
mation. Three graphite samples, two of CGB grade
and one of pyrolytic graphite, were exposed in the
gas phase, while similar specimens were exposed
to the liquid. A comparison sample of Hastelloy N
was exposed in each phase. Assemblies used in
this experiment consisted of a pair of the holders
shown in Fig. 9.9. These holders, for safety
reasons, eiclosed the praphite and Hastelloy
gpecimens in perforated metal cylinders. They
were connected by a loop of stainless steel
cable of a length such that the bottom holder was
completely immersed in the molten fuel while the
top holder was exposed only to the gas phase.
This assembly was exposed in the pump bowl for
8 hr with 200 ce/min of helium purge except for
the last 10 min. After the exposure, tiny droplets
(~0.2 mm in diameter) of greenish white fuel salt
were seen adhering loosely to the gas-phase
graphite gpecimens. The mechanical disturbance
involved in removing the specimens from the holder
was sufficient to dislodge the droplets from the
graphite surfaces. Nevertheless, two of the
graphite specimens were wiped with small squares
of cloth, which were analyzed radiochemically.
The remaining graphite and metal specimens were
then leached in a neutral mixture of sodium ver-
senate, boric acid, and citric acid, which dis-
 

SAMPLES,
3/gin. DIAM X tY%gin. ——

 

 

CGB GRAPHITE —-

 

132

ORNL-~DWG 67 —12239

MATERIAL: Ni OR NICKEL PLATED

MILD STEEL

| I/35-in. HOLES FOR CABLE
(SPACE BETWEEN SAMPLES)

   
  

———GET SCREW

=———Ni WIRE RING THROUGH HOLE IN
GENTRAL ROD. MAY BE OMITTED

IF SET SCREW CONSIQERED SAFE
AGAINST LOOSENING

INOR ~8

-4-— CGB GRAPHITE

PYROLYTIC GRAPHITE

Fig. 9.9. Graphite Sample Assembly.

solves fuel salt but which should not attack metal.

Finally, the graphite specimens were dissolved in
H2804—HNO3, and the metal specimens were
leached free of activity with 8 N nitric acid. All
leaches were analyzed separately.

Deposition of noble metals on the sections of
stainless steel cable (shown in Table 9.8 as
FP11-50) was as much as two orders of magni-
tude higher than that on the graphite and Hastel-
loy N specimens of similar surface area. Depo-
sition was heaviest on the interface section of
the cable and least on the vapor-phase section.

It is likely that the perforated sample holders
(which surrounded the specimens and which could

not be analyzed) intercepted much of the activities
which would otherwise have deposited on the
graphite and Hastelloy N specimens.

The fraction of activity removed by the aqueous
chelating leach, which preceded the acid leach
or dissolution, varied markedly from one isotope
to the next. As Table 9.9 indicates for several
typical leaches, a sizable fraction of the °Mo
and °*Nb (and often the predominant amount)
was found in the chelating leach. By contrast,
this leach usually removed only about 10% of
the '¥2?Te and ruthenium from both graphite and
metal samples, but it generally removed more
than half of the 1311, 95Zr, 14%8a and 3°Sr.
133

Toble 9.9. Comparison of Aqueous Chelating Agent with HN03 Leach for Removing Leposited

Fission Products from Graphite and Hastelloy

 

Digintegrations per Minute in Total Sample

 

 

E,x§loied Sample® 55 1
Mo 327 1030y 952y
x 101° % 101° x 10° x 167
Gas CGB V 4.6 2.99 0.15 4.3
A 1.75 32.1 1.19 0.89
T 6.35 35.1 1.34 5.23
Liquid CGB V 4.60 2.99 0.154
A 3.24 2.25 3.00 0.13
T 7.84 5.24 3.15 4.47
Gas Hastelloy N V 4.05 5.89 0.12 1.49
A 0.21 70.2 <5.3
T 4,87 76.1 <6.79
Liquid Hastelloy N V 2.24 1.99 0.07 0.35
A 1.51 15.5 0.66 <2.3
T 3.75 17.5 0.73 ~2.7

 

?V represents leaching in aqueous chelating solution; A represents leaching in acid; T represents total.

Since this solvent dissolves salts but not
rnetals, these observations may possibly indi-
cate that Mo and ?°Nb are present as oxidized
species (fluorides?), while tellurium and ru-
thenium are present as metals. Wiping with a
cloth removed larger fractions of °°Mo and °°Nb
than of activities such as 132Te, 131, 957,
14084, and #9Sr which penetrated the graphite.

The deposition of fission products on pyrolytic
graphite (see Table 92.10) in the gas phase was
similar to that on CGB graphite, although '32Te
deposited to a smaller extent on the pyrolytic
graphite. In the liquid phase, ®°Mo and **2Te
deposited slightly more heavily on pyrolytic
graphite than on the CGB.

The deposition of ?5Nb was markedly heavier
on Hastelloy N in the liquid phase than in the gas
phase. Tellurium-132 and 1*%Ba deposited more
heavily on the metal exposed in the gas phase.
Other isotopes deposited similarly on metal in
the two environments.

When metal and graphite were exposed to the
gas phase, somewhat more 132Te, 1311 and
95Zr deposited on the metal. When both were
exposed to the liquid phase, less °°Mo, 14°Ba,
and 2%Sr and more '32Te and '°'] deposited on
the Hastelloy. The deposition per square centi-
meter of a given isotope varied by less than a

factor of 10 between any Hastelloy N or graphite
samples.

When the activity, in disintegrations per minute
per square centimeter, of fission products on the
graphite samples from the 24,000-Mwhr core ex-
posute is compared with that on submerged
graphite specimens from the 8-hr exposure (Tables
9.4 and 9.10) the ratios were 0.9 for °Mo, 1.7
for 132Te, 3.5 for '?3Ru, 14.0 for 19%Ru, 95 for
?5Nb, and 56 for *3Zr. A similar comparison of
the activities on the Hastelloy N specimens from
the two exposures (Tables 9.6 and 9.10) gives
ratios of 20 for ®°Mo, 6 for 132Te, 170 for 193Ruy,
400 for 19°Ru, 75 for 5Nb, and 30 for 93Zr. The
ratio of full-power exposure times is 417; the ratio
of total exposure times is, of course, considerably
greater., Both sets of activity ratios may be bi-
ased high since the sample holder in the 8-hr ex-
posute may have intercepted a sizable fraction of
the depositing activities.

Examination of the ratios given above shows
that only deposition of '°*Ru on Hastelloy N
proceeds at approximately the same rate for 8
as for 3340 full-power hours. However, the
measured deposit of a given radioactive nu-
clide actually reflects the deposition rate only
over the last two or three half-lives of exposure

time for that nuclide. Thus the measured ac-
Table 9.10. Deposition of Fission Products During B-hr Exposure in MSRE Pum

p Bowl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Exposed Sampte Area . Disintegrations per minute in Total Sample” 235y
in — (sz) 99M0 132Te 1311 E}szr 8981‘ 140-Ba 103Ru 106RU QSNb 143Ce (ng)
w1010 x1o!l  c10° %107 x 108 x 107 x 107 % 107 x10%  x10®
Gas CBG No. 1 3.68 9.35 3.51 3.42 5.23 8 1.62 1,34 4.0 4,84 5.14
Gas CBG No. 2G 3.68 5.56 4,44 1.65 1.4 0.47 2.16 6.0 8.8 1.31
Gas PG 2.76 6.13 0.371 1.25 0.91 7.52 1.09 2.3 6.0 5.3 0.71
Liquid CBG No. 1L 3.68 7.8 0.52 3.14 4.47 13.0 1.58 3.15 9.27 4.0 4,76
fLiquid CBG No. 2L 3.68 11.2 0.464 2.34 0.58 0.33 5.29 15.9 4.4 .27 1.55
Liquid PG 2.76 33.4 0.7¢ 2.81 2.4 158 0.59 25.2 8.06 5.4 15.5 1.37
Gas Hastelloy N 2.46 4.87 7.61 15.6 4.8 10.9 1.09 2.3 6.55
Ligquid Hastelloy N 2.46 3.75 1.75 6.77 2.7 1.97 0.063 0.73 2.0 63.5 13.8

 

 

 

 

8p11 activities calculated for time of sampling.

veT
tivities of ?°Mo and !'32Te deposited in a long
exposure are a measure of deposition rate over
only the last week or so of exposure. If we
assume a constant deposition rate, the amount
deposited in time t is given by the familiar
equation:

k
—At
N=—(10 -,
)\(

where k is the constant deposition rate and A
is the decay constant. Using this equation a
theoretical ratio can be calculated for each iso-
tope for exposure times of 3340 and 8 hr:

. M. .o 1 — e—AX3340

L AXE
Ns 1 -¢

 

These ratios turn out to be 12.6 for °?Mo, 14.5
for 132Te, 157 for 123 Ru, 365 for 19%Ry, 142 for
?5Nb, and 218 for ?3Zr. Whether fortuitous or
not, these calculated ratios agree with the ob-
served ratios for deposition on Hastelloy N
very well for the ruthenium isotopes and within
a factor of 2 for molybdenum, tellurium, and
niobium. This agreement suggests that the
deposition of noble metals on Hastelloy N does
indeed proceed at approximately the same con-
stant rate for a 3340-hr exposure as for an 8-ht
exposure,

By contrast, the calculated ratios are much
higher (usually by teafold or more) than the ob-
served ratios for deposition on graphite except
for >5Nb and ?°Zr, whose calculated ratios are
high by factors of only 1.5 and 4 respectively.
This indicates that the deposition rate of noble
metals (except *°Nh) on graphite decreases with
exposure time. Neutron economy in MSRE would
benefit if the deposition rate of noble-metal fis-
sion products on graphite decreases with expo-
sure time while that on Hastelloy N remains
constant., As the data in Table 9.10 indicate,
the 8-hr deposition rates are not greatly different

135

on graphite and metal, and the activities of
%Mo, 1327Te, '93Ry, and 19°Ru were similar
for the 7800- and 24,000-Mwhr exposures. This
suggests no change of rate between 7300 and
24,000 Mwhr for ?*Mo, 32Te, and '°*Ru and a
decreased rate for the long-lived '%8Ru.

The discussion above gives a generally satis-
factory picture of noble-metal deposition for ex-
posures of 8 hr or longer. The picture is greatly
complicated by the inclusion of results from the
short (1 to 10 min) exposures of the stainless
steel cable specimens in the pump bowl fuel (see
Table 9.8). For most isotopes, the depositions
were only slightly greater (a factor less than 5)
for the 8-hr exposure than for 1- to 10-min ex-
posures. Tellurium-132 and ?°Zy deposits in 8
hr were about 100 times larger than for the short
eXpOsSUures.

The calculated ratios from the formula given
above would be very close to the time ratios for
these short exposures, Comparing the 8-hr run
with a 1-min run, the time ratio is 480. The
short-exposure data thus indicate that (except
for 1*?Te and ?°Zr) the original deposition rate
is very fast compared with the rate after 8 hr.

It may be that when fresh samples of metal are
exposed to fuel salt two deposition mechanisms
come into play. The first process rapidly de-
posits noble metals on the stainless steel in
less than a minute and then stops (or reaches
equilibrium). The second (slow) process then
continues to deposit noble metals at a fairly
constant rate for exposure times up to 3340 hr.
It may be speculated that the first process is a
rapid pickup of colloidal noble-metal particles
from the fuel on the metal surface; the second
may be a slow plating of noble metals in highet
valence states present at very low concentra-
tions in the fuel.

It will be interesting to expose graphite
samples in the fuel for short times to see
whether a short-term mechanism also exists for
deposition on graphite.
10. Studies with LiF-BeF, Melts

10.1 OXIDE CHEMISTRY OF ThF¥ ,.UF, MELTS

B. F. Hitch C. E. L.. Bamberger

C. F. Baes, Jr.

The precipitation of protactinium and uranium
from LiF-BeF -ThF , mixtures by addition either
of BeO or ThO, was demonstrated several years
ago by Shaffer et al.?~? as a possible method for
processing an MSBR blanket salt. The purpose
of the present investigation is (1) to verify that
the oxide solid phase formed at equilibrium with
UF -ThF ,-containing melts is the expected solid
solution of U0, and ThO, and (2) to determine the
distribution behavior of Th** and U*™ between the
oxide and the fluoride solutions:

U*H(E) + Th* %) = U*¥o) + Th* XD (1)
(where f and o denote the fluoride and the oxide
phases).

The experimental technique is similar to that
used in a UO,-Z10, phase study.*® The ThO, and
UO, were contacted with 2LiF - BeF , containing
UF , and ThF , within nickel capsules under a
hydrogen atmosphere in a rocking furnace. The
equilibrated oxide solids were allowed to settle
before the samples were frozen. Provided suf-
ficient fluoride phase had been added originally,
good phase separation was thus obtained.

A (U—Th)O2 solid solution has been found in
every sample examined thus far, and, moreover,

 

'y, H. Shaffer et al., Nucl. Sci. Eng. 18(2), 177 (1964).

2_]. H. Shaffer, G. M. Watson, and W. R. Grimes, Ke-
actor Chem. Div. Ann. Progr. Rept. Jan. 31, 1960,
ORNL-2931, pp. 84—90.

3_]. H. Shaffer et al., Reactor Chem. Div. Ann. Progr.
Rept. Jan. 31, 1961, ORNL-3127, pp. 8--11.

4K. A. Romberger, C. ¥. Baes, Jr., and H. H. Stone,
J. Inorg. Nucl. Chem. 29, 1619 (1967).

136

the lattice parameter determined by x-ray diffrac-
tion® was consistent with the composition calcu-
lated for such an oxide phase. The equilibrium
quotient for the above metathetic reaction was
determined by analysis of the fluoride phase for
the small amount of uranium which it contained.
The results obtained thus far give

Xuy X

X

Th{f) 1000 to 2000
U (f) X’[‘h(o)

Q- )

at 600°C. The mole fraction of uranium in the
oxide phase [XU(O)] was varied from 0.2 to 0.9,
while the mole fraction of thorium in the fluoride
phase was in the range 0.01 to 0.07.

It thus appears that relative to Th**, the U*"
present is strongly extracted from the fluoride
phase by the oxide solid solution formed at equi-
librium. There is every reason to expect that
Pa*" will distribute even more strongly to the
oxide phase. This is in agreement with the ef-
fective precipitation of U*" and Pa* ™ by oxide

~

first reported by Shaffer ef al. The formation of

a single oxide solution phase to which Th*7,
U*" and Pa®™ all could distribute offers a much
more flexible and effective oxide separation
method for a single-region breeder reactor fuel
than would be the case if only the separate oxides
(ThO, and UO,) were formed. For example, a
possible processing scheme for an MSBR might
involve some sort of a countercurrent contactor
(Fig. 10.1) in which a (U-'Th)02 solid solution of

 

5These x-ray examinations were performed by G. D.
Brunton and . R. Sears of the Reactor Chemistry Di-
vision. The mole fraction of ThO;, in the oxide solid
solutions was calculated from the lattice parameter,
using the equation given by L. O. Gilpatrick and C. H.
Secoy, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31,
1965, ORNIL.-3789, p. 241.
ORNL-DWG &7-14820

HF, Hs

TO REACTOR l ["‘“‘”“HF’ Hy0, Hp —mr——

FROM REACTOR

Li+, Belt £
Tht UM et

COUNTERCURRENT

CONTACTOR .
FLUORIDE SOLUTICN
OXIDE SOLID SOLUTION [=——(Th, U) 02

 

 

 

ret—ee {Th, U, P0)O3

 

 

=t HOLDUP f——se]  MAKEWP
=L HOLDUR e MAKELE r

e HYDROFLUORINATION

Lit Be2t -
ThAt ya+

 

 

 

 

PRODUCT MAKELP

Fig. 10.1. Suggested Flow Diagroms for Protactinium Removal by Oxide Extraction.

controlled initial composition is equilibrated with
the blanket salt. It is expected that 233Pa will
distribute strongly to the oxide phase, while the
amount of ??3U removed will be controlled by the
uranium contenf of the influent oxide phase. Sub-
sequently, the oxide phase might be stored while
the ?%°*Pa decays and then largely recycled to the
contactor with some adjustment of the composition
and with removal of a portion of the (Th-U)0, solid
solution as product. Before returning the salt to
the reactor, the dissolved oxide could be removed
by HF-H, sparging. This processing concept
might be applicable to a one-region breeder fuel.
It should only be necessary to adjust the uraninm
content of the oxide phase upward in order to be
compatible with the higher uranium content nec-
essary for a fuel salt.

It is planned to continue the present measure-
ment of the disttibution of U??, and perhaps Pa*®,
between fluoride and oxide phases as a function
of composition and temperature. In these meas-
urements, vigorous agitation will be used in the
hope of shortening the equilibration times. If the
results should warrant, a more detailed study of
the rate-controlling factors in the distribution of
U** and Pa** will then be made.

10.2 CONTAINMENT OF MOLTEN
FLUORIDES IN SILICA
Chemistry

C. E. L. Bamberger
J. P. Young

R. B. Allen
C. F. Baes, Jr.

The advantages of silica as a container material
for molten salts in spectrophotometric, emf, and
other measurements include high thermal shock
resistance, good ultraviolet transmittance, high
electric resistivity, low price, and ease of fabri-
cation. For the containment of molten LiF-BeF
mixtures, an obvious disadvantage is possible
chemical attack as a result of the reaction

2BeF (d) + Si0,(s) & 8iF (g) + 2BeO(s) . (3)

However, when the equilibrium partial pressure of
SiF , was calculated for this reaction from avail-
able formation free energies for crystalline §i0,,°
crystalline BeO, ® gaseous SiF ,® and dissolved

 

GJANAF Thermochemical Tables, Clearing House for
Federal Scientific and Technical Information, U.3. Dept.
of Commerce, August 1965,
138

BeF , in 2LiF - BeF ,, 7 this partial pressure was
found to be surprisingly low, that is, 0.03 mm at
700°K to 21 mm at 1000°K. Furthermore, pre-
liminary direct measurements now in piogress con-
firm the low range of these partial pressures. This
suggested that an overpressure of SiF , might pre-
vent, or at least reduce, the attack of the silica
container and prevent the precipitation of BeO.

The compatibility of Si0, with these LiF-Belk',
melts is primarily associated with the relative
stability of BeF, compared with that of BeO; the
standard free energies of formation at 1000°K are
~206 and 120 kcal/mole respectively. Since
many other metallic fluorides show even greater
stability compared with their corresponding oxides,
it seems that S10, should not be ruled out as a
container material for molten fluorides without
first estimating the equilibrinm position of such
reactions as

X
ZSioz(S) + MF (s or H= MOX/Q(S)

LSPGO, @

which involve the pure metal fluoride and metal
oxide, and

S10,(s) + 4F ~ = 207~ + SiF (g) . (5)

The latter reaction is the one of most interest,
since it gives the level of oxide contamination
of the fluoride melt.
be judged if activity coefficients or solubilities
can be assigned to MF_and MO, ,, in the melt
under consideration. In the present case the ac-
tivity of BeF , in 2LiF - BeF , (~0.03) and the
solubility of BeO (~0.01 mole of O? per kg of
salt at 600°C) have been measured.’ From these
and the avaijlable thermochemical data the follow-
ing equilibrium quotient may be estimated for the
above reaction [Eq. (5)] in 2LiF - BeF , at 600°C:

Its equilibrium position may

Q=[0%"]? PSiF4

=4 %10~ 7 atm mole */kg? . 6)

 

7C. F. Baes, Jr., *"The Chemistry and Thermodynamics
of Molten Salt Reactor Fluoride Solutions,?’ SM-66,/60
in Thermodynamics, vol, 1, JAEA, Vienna.

From this, it can in turn be estimated that in the
presence of 1 atm of SiF , the oxide ion concen-
tration at equilibrium with Si0, should be only
6 x 10™? mole/kg.

Another factor to be considered when using sil-
ica is the possible formation of silicates; for
exaniple,

2MF ,s or 1) + 2510 (s)

= M,Si0,(s) + SiF (&) . (7)

In the present case, for example, Be ,510, (phen-
acite) should be formed as the stable reaction
product rather than BeO at low S, parttial pres-
sures; however, phenacite is not very stable rel-
ative to 2BeO + SiO, and should not alter the
above conclusions about the compatibility of sil-
ica with LiF-BeF , melts. In other fluoride sys-
tems, the formation of metallic silicates may be
the controlling factor.

The solubility of SiF', in the molten fluoride
must be considered because possible reactions
such as

SiF (&) + 2F~ = SiF *~ (&)

could produce high solubilities which not only
would alter the salt phase, but also would cause
the reaction with Si0, to proceed farther than
otherwise expected. Accordingly, the solubility
of SiF , in 2LiF - BelF , was estimated by means
of transpiration measurements, The prepurified
salt was saturated with SiF , by bubbling a mix-
ture of 0.1 atm of SiF , in He throngh the salt
until the effluent and influent compositions were
the same. The dissolved gas was then stripped
with helium and trapped in an aqueous NaOH bub-
bler, the amount of SiF , being equivalent to the
amount of NaOH consumed by its hydrolysis:

SiF ,(g) + 40H™ = 4F~ + 2H,0 1 $i0(s) . (9)

The solubility was calculated by means of

 

[SiF,]
K= 5 ; (10)
SiF4
[SiF 1 =[2ng, . H(PSiF4V/RTm)]/w, (11)
}‘:‘nsm == {otal number of moles of SiF,
4 titrated,
PSiF4 = SiF, pressure at equilibrium
with the melt,
E(PS. . V/RT ) = correction term which in-
iF 4 m

cludes the various dead vol-
umes at different tempera-
tures, T,

w = weight of the melt in kg.

Using Egs. (10) and (11) the solubility was found
to be

K ;= 0.032 £ 0.005 mole kg™ Latm—1! at 499°C,
KH = 0.035 +0.005 mole kg~ ' atm— ! at 540°C.

These solubilities are of the same order of magni-
tude as the solubility of HF in 2LiF - BeF , (~0.01
mole kg~ ! atm ™ 1).® At higher temperatures, in
the range 600 to 700°C; the solubility was so low
that by the rather insensitive method of measure-
ment only an upper limit could be set: K _ <0.01
mole kg~ ! atm ™. The data obtained indicate
that at 1000°K, using an SiF , pressute close to
the equilibrium value over 2LiF - BeF , the solu-
bility of SiF | is ~0.003 mole %.

Another undesirable side reaction that should
be considered in these systems is the formation
of silicon oxyfluorides; for example,

SiF ,(g) + Si0 (s) = 2Si0F ,(g) (12)

3SiF ,(8) + Si0,(s) & 251 ,0F (&) . 13)
Such silicon oxyfluorides have been identified by
Novoselova et al.? during the synthesis of phen-
acite from 5i0,, BeO, and Na ,BeF , at tempera-
tures in the range 700 to 800”C. In one meas-
urement they report equilibrium partial pressures
of 0.162 atm of SiF , and 0.07 atm of SiOF , over
the reaction mixture at 813°C. In other experi-
ments wherein equilibrium was not reached, they
were able to identify Si OF .. We performed tests
in which He after passage through molten 2LiF -
Bel , containing powdered BeO and SiO, at tem-
peratures up to 650°C was analyzed by means of
a mass spectrometer. No silicon oxyfluorides
were found to be present.

 

8P, E. Field and J. H. Shaffer, J. Phys. Chem., in
press.

139

Spectrophotometric Measurements

with Silica Cells

C. E. L. Bamberger
C. F. Baes, ]Jr.

J. P. Young

Behavior of U4™,

which an S1F, overpressure would prevent reaction
(5), UF , a colored solute which would produce an
even less soluble oxide than does BeF ,, was
added to LiF-BeF , melts in SiO, containers. Tet-
ravalent uranium thus should act as a sensitive
““internal indicator’ of the reaction of F~ with
Si0,. It has a known absorption spectrum of suit-
able inteusity, *? and hence its concentration can
be followed spectrophotometrically. For the re-
action

- To estimate the extent to

UF4(d) + SiOz(s) = SiF 4(g) + UOZ(S) , (14)
equilibrium partial pressures of SiF , were pre-
dicted to be reasonably low; for example, with
0.002 mole fraction UF‘5 in 2IAF - BeF ,, calcn-
lated SiF , pressures varied from 3 mm at 700°K

to 115 mm at 1000°K. Spectrophotometric mon-
itoring of the concentration of gt provides a very
good means for studying the stability of the sys-
tem.

Melts of 2LAF - BeF , containing variable con-
centrations of UF, (0.0053 to 0.13 mole % or 0.003
to 0.008 mole/liter) were held under SiF (400
mm) at tempetatures ranging from 490 to 700°C in
sealed silica tubes. The tubes were placed di-
rectly in a special furnace'! located in the light
path of a Cary spectrophotometer, model 14M, or
in a heated nickel metal block and later transferred
to the furnace inside the instrument. Temperatures
were controlled to +1°C by means of a Capacitrol
471 controller.

Constancy of the U*" absorbance peaks at 1090
and 640 nm showed the uranium concentration in
the melt to be constant within the precision of the
spectral measurement (about 1%) for at least 48 hr.
It was, therefore, evident that no significant attack
of the silica container or significant contamination
of the melt occurred. During longer periods of time

 

A. V. Novaselova et al., Proc, Acad. Sci. USSR,
Chem. Sect. {(English Transl.), 159, 1370 (1964); A. V.
Novoselova and Yu. V. Azhikina, [norg. Mater. USSR
(English Transl)), I{9), 1375 (1966).

loj. P. Young, Incrg. Chem. 6, 1486 (1967).
17, P, Young, Anal. Chem. 31, 1892 (1959).
140

(up to one week) the transparency of the containers
was noticeably affected, and the base line of the
spectra rose. The net absorbance of the U** peaks,
however, remained reproducible even though in
some experiments the temperature was cycled from
500 to 700°C. No spectral experiment was con-
tinued for longer than a week. However, in other
studies, melts of 2LLiF * BeF | were maintained as
long as one month at S00°C without noticeable
damage except loss of transparency {(presumably
due to some process of devitrification) of the fused
silica.

The precision of these quantitative spectral meas-
wrements of U*" was better than that possible with
windowless cells.}? The molar absorptivity (Am)
of U** had been observed!® to be 14 + 10% and 7 +
10% at 1090 and 640 nm, respectively, in molten
LiF-Be}l , at 550°C contained in windowless cells.
With the melts contained in the present silica cells
the A_of U*'is 14.4 + 0.5 and 7.3 + 0.3 at 1090
and 640 nm, respectively, in molten 2LiF - BeF ,
at 560°C. The effect of temperature on the 4 _ of
several absorption peaks of U*™ in 2LiF - BeF )
was studied. The results are summarized in Table
10.1.

Behavior of Cr® . — The solubility of CrF | in
Li.F-BeF2 is being investigated spectrophoto-
metrically with the melts contained in SiO, cells.

A minimum solubility of 0.43 mole % Cr®* was ob-
tained at a temperature of 550°C. This is not the
solubility limit, but the solutions are too opaque

to observe spectrophotometrically in a 1-cm path
length. Further studies will be carried out with
510, cells of a shorter path length. The spectrum
of Cr3'in LiFF-BeF , at 500°C consists of three
peaks at 302, 442, and 690 nm, all arising from
transitions within the 3d level.
sorptivity of the 690-nm peak was calculated to

The molar ab-

be 6.6. This value compares favorably with an

Table 10.1. Effect of Temperature on the Molar
Absorptivity of Several Peaks of U4+
in Molten 2LiF - BeF,

 

Molar Absorptivity

 

Temperature (°C)

 

1050 nm 640 nm
498 16.4 3.0
560 14.4 7.3
650 13.8 6.2
698 12.9 5.7

 

approximate value of 7 reported from a cursory
study of the system in windowless cells.!® The
molar absorptivity of the two other peaks is some-
what higher, about 10, but fucther work will be
necessary to establish a more precise value.

10.3 ELECTRICAL CONDUCTIVITY
OF MOLTEN FLUORIDES
AND FLUORGBORATES

J. Braunstein K. A. Romberger

Studies have been initiated to measure the elec-
trical conductivity of molten fluorides and fluoro-
borates of interest as fuel, blanket, and coolant
salts for the Molten-Salt Breeder Reactor. Pre-
liminary measurements were made to provide esti-
mates of the electrical conductivities for engi-
neering use in the design of electrical circuits for
instrumentation in the MSBE. Cell desiga has
been investigated, including the use of silica in
view of recent indications that dry fluorides may
be stable to silica in the presence of low concen-
trations of SiF4 (see Sect. 10.2).

Preliminary measurements were made of the con-
ductivity of 2LiF-BeF ,, of LiF-ThF , eutectic
(29 mole % ThF ), and of NaBF , by observing
current-voltage curves at 100 hertz and using plat-
inized platinum wire electrodes in a silica cell at
4000 hertz. !* Potassium nitrate at 390°C was
used to determine the effective cell constant.
Temperature coefficients were estimated from data
on mixtures of alkaline-earth fluorides with alkali
fluorides or with cryolite !® and from the data of
Greene ' ® on FLINAK and NaF-Z:F -UF , mixtures.

The cell used for the current-voltage measure-
ments was similar to that used by Greene!® and
consisted of a pair of platinum wire electrodes of
0.07-cm diameter separated by about 2 cm and im-
mersed to a measured depth in the melt contained

12}. P. Young, Anal. Chem. 36, 390 (1964).

Bpsr Program Semiann. Progr. Rept. Aug. 31, 1966,
ORNL-4037, p. 194.

1 . . . .
4J. Braunstein to R. I.. Moore, private communication,
Aug. 14, 1967.

1%, w. Yim and M. Feialeib, J. Electrochem. Soc.
104, 626 (1957); J. D. Edwards et al., J. Electrochem.
Soc. 99, 527 (1952); M. de Kay Thompson and A. L.
Kaye, Trans. Electrochem. Soc. 67, 169 (1931).

16N. D. Greene, Measurements of the Electrical Con-
ductivity of Molten Fluorides, ORNIL-CF-54-8-64 (Aug.
16, 1954).
141

in a silica tube. (The standard, KNO,, was con- creasing the current path length through the elec-
tained in a Pyrex beaker.) The voltage drop across trolyte. In the new cell, the electrodes will enter
a resistor in seties with the electrodes and elec- the melt through silica tubes of about 0.3 cm ID
trolyte was measured with an oscilloscope whose and extending 3 cm below the surface of the melt.
vertical amplifier was provided with a calibrated The dependence of the measured conductivity on
voltage offset. Bridge measurements also were the temperature and composition of the melts and
made with the same cell, using a General Radio on frequency is under investigation.

rmodel 1650A impedance bridge. In order to im-
prove the sensitivity and reproducibility of the
measurements, a bridge was constructed with a
100-ohm Helipot to provide the ratio arms and with

Table 10.2, Electrical Conductivity of Molten

Fluorides and Fluorobsrates

 

a decade resistance box and variable parallel ca-

 

pacitance in the balancing arm. An oscilloscope Conductivity Temperature 1cmperature
served as the phase balance and null detector. (ohm ™ em™1) (") Coiffidelm
The preliminary results listed in Table 10.2 are [
probably in error by 20 to 50%, the limiting factor . 3
being the small cell constant in the initial ex- 2LAF Bel 1.5 650 1x10
periments. LiF -ThF 2.0 650 3x107°
A new dipping cell has been constructed to pro- NaBF 0.8 420 2% 1073

vide a cell constant of the order of 100 by in-

 
11

The appearance in appreciable concentrations of
several fission product species in the gas space of
the MSRE pump bowl remains the biggest chemical
surprise in MSRE operation. This behavior does
not, of itself, seem to pose serious problems for the
MSRE or for subsequent large molten-tluoride re-
actors, but it is clearly important to establish the
mechanism by which these diverse species, notably
9%Mo, ?°Nb, '?%Ru, '°Ru, 132Te, and even !!'%Ag,
get into the gas stream.

It is possible that at least some of the gas-bome
species listed above have volatilized from the melt
We have, accordingly, initiated studies
of the (less well known) lower fluorides of these
materials both as pure compounds and as their solu-

as fluorides.

tions in molten LiF-BelF , preparations. The pre-
liminary studies performed to date have been de-
voted to the fluorides of molybdenum, but it is
anticipated that, should the results warrant, the
study will be broadened to include other elements
on the list.

11.1 SYNTHESIS OF MOLYBDENUM FLUORIDES

C. F. Weaver H. A. Friedman

Molybdenum is a constituent of Hastelloy N, the
MSRE structural metal, as well as an important
high-yield fission product. Consequently, and
especially since molybdenum is one of the elements
appearing in the MSRE exit gas stream, the be-
havior of molybdenum and its fluorides in molten
fluoride fuels in contact with graphite and Hastel-
loy N is of considerable interest. ™3

A search of the literature on the tluorides and
oxyfluorides has been completed and reported.?
The fluorides MoK _, M0F4, Mo, F,, MoF, and
MoF _ have been shown to exist, but of these only
MoF _ is commercially available. It has been neces-
sary, therefore, to refine methods for preparation,

142

Behavior of Molybdenum Fluorides

and to prepare gram quantities, of the lower-
valence compounds for study.
Preparation of MoF _ by the reaction

SMOF6 + Mo~ EiMDFS

has been described by Edwards, Peacock, and
Small,® who have also established the crystal
structure of MoF _. We have successfully employed
this method in which Mol _ is refluxed over metal-
lic molybdenum at temperatures in the ianterval 35 to
100°C in glass apparatus.

Molybdenum pentafluoride is a yellow, hygro-
scopic material which melts at about 65°C to a
yellow liquid of high viscosity.® Its vapor pressure
at the melting point is about 2 mm,® and it boils at
about 212°C.%-°

Disproportionation behavior of MoF  appears to
be very complex. We have observed that heating of
the material to 100°C while maintaining a good
vacuum leaves the MoF _ unchanged. However, we
have shown that if Mok _ is heated at 200°C and if
the volatile product is removed by pumping, the
reaction is

3M0F5—> 2M0F6 + MOF3 i

 

'W. R. Grimes, Chemical Research and Development
for Molien Salt Breeder Reactors, ORNIL.-TM-1853 pp.
61--81 (June 6, 1567). ’

3. s. Kirslis, F. F. Blankenship, and C. F. Baes, Jr.,
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1967,
ORNI.-4076, pp. 48—53.

3S. 5. Kirslis and F. F. Blankenship, MSR FProgram
Semiann. Progr. Rept. Feb. 28, 1967, ORNL-4119, pp.
124-43.

4C. F. Weaver and H. A. Friedman, A Literature Sur-
vey of the Fluorides and Oxyfluorides of Molybdenuin,
ORNL-TM-1976 (October 1967).

SA. J. Edwards, R. D. Peacock, and R. W. H. Smail,
J. Chem. Soc. 4486—91 (1962).

5D. E. LaValle et al., J. Am. Chem. Soc. 82, 243334
(1960).
The MoF , so obtained has been identified by its
x-ray diffraction pattem established by LaValle

143

¢t al.® on the material synthesized by other methods.

This convenient method of preparation, which seems
not to have been used before, has served to pre-
pare several pure batches of MoF .

When MOF5 is pumped, in this same way at 150°C,
the reaction has been reported® ” to be

2M0F5 —_— MOF4 + MOF6 .

We have established that the solid product is not
Mo¥ | (as produced in the 200°C reaction), but we
have not vet completed our identification of the
material.

The compound MoF | has been shown to react
with LiF to fom at least two binary compounds.
These materials, whose stoichiometry has not yet
been established, are both birefringent. The mean
index of refraction for compound I is at 1.520,
while its major x-ray diffraction peaks (copper
x-tay target; 26 values in degrees) are at 21.4,
19.4, and 27.0; for compound II, the mean refrac-
tive index is 1.480 and the major diffraction peaks
ate at 22.3 and 20.3. The stoichiometry and the
optical and x-ray data will be established as soon
as well-crystallized samples are available.

11.2 REACTION OF MOLYBDENUM FLUORIDES
WITH MOLTEN LiF.BeF, MIXTURES

C. F. Weaver H. A. Friedman

The appearance of ?Mo in the exit gas from the
MSRE ? has suggested the possibility that a vola-
tile fluoride of molybdenum exists in the MSRE. We
have accordingly begun an examination of reactions
of molybdenum {luorides with MSRE fuel and fuel
solvent mixtures.

Molybdenum hexafluoride is a very volatile
material (boiling point, 34°C) and is a very strong
oxidant. The nickel container was rapidly attacked
when Mol diluted to ™ 50% with helium was
passed through the MSRE fuel mixture (in which all
of the uranium was as UF,) at 650°C; after 1 hr the
fuel melt contained 9500 ppm of nickel. A small
guantity of uranium was transported (presumably
as UF ) by the gas, but, as expected, no appre-

 

7G. H. Cody and G. B. Hargreaves, J. Chem. Soc.
156874 (1961).

ciable reaction of the MoF , with the melt was ob-
served. When MoF  was passed through a fuel
mixture to which 0.08 mole % UF | had been added,
the UF | was rapidly oxidized to UF . A small
quantity of uranium was again observed in the exit
gas lines, and violent corrosion of the nickel was
observed. The valence state to which the MoF
was reduced was not detemined in these experi-
ments. It is probable, however, that the molybde-
num was present as either molybdenum metal or
MoF ,, since MoF and MoF , have not been reported
to exist and MoF ,, Mo I, and MoF  seem to be
thermally unstable above 200°C.

The compound MoF , dispropertionates readily®
under a vacuum at temperatures greater than about
600°C to form molybdenum and MoF .. However,
this material has been heated under its own pres-
sure i closed nickel and copper capsules at 500
and 710°C, respectively, for perieds in excess of
ten days without disproportionation. In addition,
it has been shown that MoF , at 500 mm will react
at 560°C with molybdenum to form MoF .. Conse-
quently, an equilibrium

2MoF , T==Mo + MoF

must exist in this temperature range with MoF
pressures of a fraction of an atmosphere. When
MoF . was heated in a nickel capsule to 500°C in
the presence of 2LiF-BeF , (50-50 mole %), the

reaction
2}\/[0173 + 3Nj —> EiNiF2 + Mo

occurred. Apparently the 2LiF.BeF  liquid dis-
solved the protective coating on the nickel wall.
Repeating the experiment with copper indicated
that much less corrosion occurred. In this case
the molybdenum was found both as MoF ; and as
the 1.480 refractive index LiF-MoF , compound of

3
unknown stoichiometry. The reaction

MoF , (1 mole %) + 3UF , (4 mole %)~—> Mo + 3UF,

was demonstrated to proceed essentially to comple-
tion at 500°C in a copper container. The reverse
reaction was not observed, although it is possible
that equilibiium was achieved with very small con-
centrations of MoF , and Ul ,.

Present information suggests that the following
events (equations unbalanced) may be significant
in the kinetics of the reduction of MoF | to molyb-
denum metal by UF,:
144

UF
source ——> MOF6 —_— M0F3

MoF , <= Mo + MoF

\U F, T

This scheme allows the molybdenum to be ‘‘trapped”’
in the trivalent state until the source of MoF _ is
removed. Then the molybdenum is converted to the
metal by the above reactions, which continue to
produce the volatile MoF | at a decreasing rate

until the process is complete. Attention is now
being given to experimentally checking this hy-
pothesis with molybdenum concentrations in the

ppm fange.

11.3 MASS SPECTROMETRY OF
MOLYBDENUM FLUORIDES

R. A. Strehlow J. D. Redman

The volatilization behavior of molybdenum and
other fission product fluorides in the MSRE has led
to a study of molybdenum fluorides. Mass spectro-
metrically derived information is of particular value
in studies involving volatilization, since, at least
in principle, the vaporizing species ate analyzed
with a minimum time lapse. This gives an oppor-
tunity to observe some transient phenomena and to
distinguish among various oxidation states and im-
purities which may be present.

The wotk so far has been concemed with the mass
analysis of vapors from three molybdenum fluoride
samples. The first objectives were to assess

material purity and to establish the mass spectro-
metric cracking pattems for these materials which
have not previously been subjected to mass analy-
sis. The three samples are designated and de-
scribed in Table 11.1.

Sample I, duting an increase of temperature from
400 to 725°C, yielded first M002F2 at the lowest
temperature. As the temperature was increased, the
peaks associated with this species decreased in
magnitude and a family of peaks attributed to
MoOF , appeared. Near the upper limit of the tem-
perature excursion, a mass peak family was ob-
served which is attributed to MoF _ and MoF ; vapor
species. The large amount of volatile oxides indi- ]
cated that an oxidation-hydrolysis had occurred -
and that better, or at least fresher, material was
needed. A somewhat increased amount of mass 96
was observed from this sample, which is attributed .
to orthosilicic acid (H ;Si0O ) rather than to the
molybdenum, since its peak height was not a con-
stant multiple of the other Mo ' peak heights.

Sample II, Mol ,, was prepared by C. F. Weaver
and H. A. Friedman and was heated in the Knudsen
cell inlet system of the Bendix time-of-flight mass
spectrometer. The compound MoO,F  was not ob-
served, but some MoOF | was evident (along with
the usual SiFS,Q,l ions) at temperatures as low as
350°C. Beginning at 275°C, M0F5+, M0F4+,
M0F3+, MoF;, and MoF " were also observed. The
M0F’5+/I\.roF4Jr peak height ratio was about unity,
indicating some Mol as well as MoF ( (or MoF ).
We have insufficient evidence to demonstrate that
MOF4 has been part of our sampled vapor. At tem-
peratures greater than 600°C, only fluoride species
were observed. The spectra for sample I at tem-
peratures of 250, 300, and 725°C are shown in
Fig. 11.1. A photograph of an oscilloscope trace of .

Table 11.1. Mass Analysis of Vapars from Three Molybdenum Fluoride Samples

 

 

Nominal
Sample Composition Source
1 MDF3 Exposed to air for D. EE. LaValle, Analytical Chemistry
several years Division
11 MOF3 Recent synthesis C. F. Weaver and H. A. Friedman,
Reactor Chemisiry Division
111 MOF5 Recent synthesis C. F. Weaver and H. A. Friedman,

 

Reactor Chemistry Division
145

102

ORNL ~-DWG 87 - 103408

 

 

 

 

   

 

 

 

 

 

T i T T
=725 MoF.* MoF,t MoFgt
—4 MoF b :
10 MoF * ¢ J
Mot
31077
: ‘ fi
34
:) '
¢ g0 Lo L AL b d ! LU SR -
& 1072 I T T T e,
4 7 = 300°C MoOF
3
4wt
-
E Ji l
5 402 ffi?‘[—_ T“_'1 """"" .
2 7= 250°C MoOF 5
% . MoF !
3 MOFZ +
*j 0 " MDF+ MOOF+ MOOFZ
0 +
'{—:J) Mot MoO
1o L
o L LAl b _ 4 oo
90 100 10 20 430 140 150 166 470 130 90 200
’ amu

Fig. 11.1.

the MoOF ,*/MoF ¥, MoOF _ "/MoF ,*, and MoF _*
peaks at 375°C is shown in Fig. 11.2. This shows,
for illustrative purposes, the comhbined spectra for
the various isotopes and compounds in the mass
range from 150 to 194 amu. At 725°C the peak
hmght for the Mol , * relative to the height of the
MOF peaks as Lompared with the spectrum at
375D allowed calculation of a tentative cracking
pattem assigned to MoF ,. This is seen only at
elevated temperatures. An impurity at m/e = 167
was assigned the probable formula 5t ,0F . '

Mass analysis of these species is facilitated by
the unique isotopic abundance ratios of molybdenum
isotopes and the difference of 3 amu between oxy-
gen and fluorine. This yields, if one selects ad-
Jac,ent pairs of species from the MoO_F, */MoOF +/
MoF, " family of peaks, three unigue peaks for edch
spec1e{~_., allowing a precise calculation of the peak
height ratios for the ‘‘overlapping” peaks to be

Mass Spectra of Sample 1l (MoF 4) During a Heating Cycle from 250°C 10 725°C,

made. The dimer Mo ¥ has, of course, the 15
peaks which would be expected from the seven
stable molybdenum isotopes. Figure 11.3 shows

the good agreement between the calculated and the

PHGTO 89172

 

155 193

Fig. 11.2. Oscilloscape Trace of Mass Spectrum of

Somple H (M0F3).
observed peak heights for Mo 2F9+. This, with the
observed precision of the analyses of the various
families of peaks, gives a satisfactory level of
confidence in the applicability of mass spectral
information to the molybdenum fluoride studies.

Sample IHl, MoF, yielded the spectrum shown
in Fig. 11.4 at 125°C. The presence of MOZFQ+
is attributed to Mo F , rather than to dimeric
MoF _, since Mo, F is a reported compound which
could be expected to yield a parent ionic species
in the spectrometer. The Knudsen cell with the
sample in it was removed from the vacuum sys-
tem, exposed to air for ¥ hr, remounted, evacu-
ated, and heated. No species was observed until
a temperature greater than 450°C was reached.
A scan at 525°C was made which yielded the spec-
trum shown in Fig. 11.4. The sole parent species
yielding this spectrum appeared to be MoO,F .

Cracking patterns have been derived for the
various species primarily based on scans in which
the dominant or sole species was well character-
ized. Since MoF _ has not yet been studied in our
equipment, we have assumed the applicability of
the reported data (see Table 11.2). Our derived
cracking patterns are shown in this table. The
compound designation for MoOF , is uncertain
(MOO'F3 is possible as far as we know). The com-
pound designated Mok _ may be a mixture of MoF
and perhaps 5 to 10% MoF .

The observation of Mo, 9+ at temperatures of
325°C and of MoF | at temperatures up to 725°C
is noteworthy. The compound Mo k' is reported to
decompose completely at 200°C, and MoF _ is re-
portedly not stable at the temperatures at which
we have observed it. One may, therefore, infer that
the free energies of formation of these fluorides
per fluorine atom are so nearly equal that kinetic
factors are expected to dominate the descriptive
chemistry of these substances.

146

ORNL~DWG 67-10342R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FOOQ rormrmrmsmrssn oo s
M02F9+ ISOTOPIC PATTERN AT 125°C
100 -
10 |—
w
},—
T
&
L
T
7
I { b d i JJ 1 ‘[ }‘
Ll
a {100 ‘ I T T
LéJ CALCULATED ISOTOFPIC PATTERN
2 FOR MopFy!
w
o
{10 — .
-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q.1 [ """" ! ! |
355 360 365 370 375
amy

 

Fig. 11.3. Comparison of Calculated and Observed
Isotopic Abundance for the Dimer M02F9.
147

-3 ORNL-DWG 67-103HR
10 Tt T T T i TR T fm \A,‘"

BEYER Z0min EXPOSURE TO M002F-+
LLABORATORY AIR AT 25°C MoOF S Mo, FF
1g7* | T=525°C '\.400+ MoOF -

 

 

 

 

 

MOF MOF?

5
L 1678 | e e
S JOT2 e e — i T T : w’\/\“““ T 7
wy i
0 r=125° MoF:* .
; wrg
- - oL
g - MoF} MoOFy s o
= i + + 02'9
- MoF MoOF
o +

1075 |- MoG MoFe ]

 

 

| fl Il

r T

1OO HO 1¢O 130 440G 150 160 170 180 490 355 3ba
amu

Fig. 11.4. Mass Spectra of Sample 11l (MoF5) at 125°C (Note Dimer, M02F9+) and at an Elevated Temperature

After Exposure to Air.

Table 11.2. Mass Spectrometric Cracking Patterns for Yarious Molybdenum Compounds

 

 

MoF MoF MoF 7 MoOF () MoO,F
MoF,* <7 MoF," 100 MoF,' 100 MoOF," 100 MoO,F," 100
MoF,' 100 MoF " 15 MoF ,* 30.4 MoF 6 MoO,F 65
MoF, " 56 MoF 10 MoF ,* 17.5 MoOF " 10+ MoOF ,* 6
MoF," 18 Mo 10 MoF )" 13.1 MoF " 11 Mor 2 2
Mo 8 MoF 6.9 MoOF ¥ 10 MoOF * 23
Mo® 8 Mo” 4.5 MoF 9 MoF * 5

Mo 3.5 MoO 7
Mo 5 Mo 16

 

“From J. C. Horton, ORGDP.
12. Separation of Fission Products and of Protactinium

from Molten Fluorides

12.1 EXTRACTION OF PROTACTINIUM FROM
MOLTEN FLUORIDES INTO MOLTENMN METALS

J. H. Shaffer
D. M. Moulton

F. F. Blankeaship
W. R. Grimes

The removal of protactinium from solution in IL.il-
Bel’ -ThE, (73-2-25 mole %) by addition of thorium
or beryllium metal has been demonstrated as a
method for reprocessing the fertile blanket of a two-
region molten-salt breeder reactor.! This investiga-
tion has been directed toward the development of a
liquid-liquid extraction process where reduced *33Pa
would redissolve into a molten metal phase for sub-
sequent back extraction by hydrofluorination into a
second storage salt mixture. Upon #s radiolytic de-
cay, fissionable 23U could be returned to the reactor
fuel via fluoride volatility.

Earlier experiments on static systems where the
simulated blanket salt, in contact with bismuth, was
contained in mild steel further substantiated the re-
duction of protactinium from the salt mixture but
failed to demonstrate its quantitative dissolution into
the metal phase. Since the extraction method utilizes
the molten metal only as a transport medium, the ap-
parent very low solubility of 2*3Pa or that of its
carrier in molten metal is not necessarily detrimental
to the process. A dynamic system was constructed
and was operated by recirculating bismuth containing
dissolved thorium through the simulated blanket mix-
ture.? Recovery of protactinium on a steel wool
column in the bismuth circuit accounted for about
43% of the ?33Pa initially in the salt phase. How-

 

1MSR Program Semiann. Progr. Rept. Feb. 28, 1966,
ORNL-3936, p. 147.

‘MSR Program Semiann. Progr. Rept. Aug. 31, 1966,
ORNL-4037, p. 148,

148

ever, the recovery of 2*3Pa was predominantly by
filtration rather than by absorption. These filtered
solids showed that ?33Pa was associated with both
iron and thorium.

More recent experiments have been designed to
demonstrate the reductive extraction of 233Pa from
the blanket salt mixture into molten metal and its
back extraction by hydrofluorination into 2 second
salt mixture. This apparatus, shown schematically
in Fig. 12.1, contained all liquid phases in graphite
to preclude the indicated absorption of 2?3Pa on
solid metal surfaces. The initial experiment of this
series demonstrated essentially complete removal of
233Pa from the blanket; only 65% of the ?33Pa was
found in the recovery salt. The loss of ?3*Pa in this
experiment and in a later experiment was attributed
to its absorption on metal particles formed by re-
duction of iron impurities in the blanket mixture.

The solubility of iron in bismuth at 650%C is about
90 ppm.> Thus, the addition of a second component
to the metal phase which would increase the solu-
bility of iron might promote the ?33Pa extraction
process. The solubility of iron in tin at temperatures
proposed for the extraction process is in excess of
1000 ppm.* Since tin is completely miscible with
bismuth below its melting point of 271°C, it was
chosen as an additive for bismuth in a recent ex-
periment. In this experiment approximately 1.78 kg
of the simulated hlanket salt, I_,il?"-Ber-ThF4 (73-
2-25 mole %), with 1 mc of 2?3Pa was in contact with
2.98 kg of bismuth which contained 0.41 kg of tin.
Approximately 0.533 kg of LiF-BelF, (60-40 mole %)
was also in contact with the molten metal mixture as

3_]. R. Weeks et al., Proc. U.N. Intern. Conf. Peaceful
Uses At. Energy, Geneva, 1955 9, 341 (1956).

1. A. Kakovskii and N. S, Smirnov, [zv. Akad. Nauk
SSSR, Otd. Tekhn. Nauk 1957(11), pp. 44—51.
149

ORNL.-DWG &67-11821

,_
£
o

SAMPLE AND Th

         

=

ADDITION PORT e SAMPLE PORT
. 0O o
1 He &
TS
oy

oFFeas—9 HT L —— OFFGAS
— o
i
T -
BLANKET & ‘1. RECOVERY
SALT - .l : SALT

 

 

 

AL MATERIAL IN CONTACT WITH SALT AND METAL PHASES
1S GRAPHITE

233

Fig. 12.1. Extraction Yessel for Pa Removal from

Molten Fluorides.

a recovery salt. The results of 233Pa transfer during
reduction by Th? additions to the blanket salt and
hydrofluorination of the recovery salt are shown in
Fig. 12.2. Approximately 90% of the 233Pa ac-
tivity was found in the recovery salt mixture, 2%
wass present in the molten metal phase, and 8% re-
mained in the blanket salt. Thus, this experiment
demonstrates quantitative accountability of ?33Pa
by analyses of samples taken from the three liquid
phases. Subsequent experiments will further eval-
uate the effects of tin, as a metal phase additive,
on the extraction process,

12.2 STABILITY OF PROTACTINIUM-BISMUTH
SOLUTIONS CONTAINED IN GRAPHITE

D. M. Moulton  J. H. Shaffer

Previous experiments on the extraction of #3%Pa
from Li E‘—ThF4-BeF2 into Bi by Th or Li have re-

ORNL--DWG 8714822

 

100

  
  

  
  
  
   
 
   

  

 

-~ 80 ] ]
&~
1 | o BLANKET SALT
T \ & RECOVERY SALT
o
® METAL PHASE
o 60 |— L‘% H 5 |
a = 3 BLANKET: 1.78 kg
~ 2, RECOVERY: 0.533 kg
= o N METAL: 2.98 kg Bi
=
o A0 | L O AND 0.408 kg Sn —
= T fij azz_ .
= v z Pal~ {mc
& T
ooy
g £ o
oy
- 200
N

 

 

 

 

 

10 15 25

THORIUM METAL ADDED (g)

Fig. 12.2. Extraction of 23%Pa from LiF-BeF,-ThF,
(73-2-25 mole %) by Reduction with Thorium Metal into
LiF-BeF, (60-40 mele %) by Hydrofluorination via
Molten Bismuth-Tin Mixture at $50°C,

sulted in an apparent disappearance of Pa from the
system. It seems as though protactinium is stable

as an ion in the salt, and only begins to act peculiarly
when it enters the liquid metal phase. We accordingly
carried out an experiment in which thorium irradiated
to produce 233Pa was dissolved in and kept in bis-
muth in an all-graphite apparatus. Data obtained in
this experiment are shown in Fig. 12.3, upon which
the 233Pa decay curve has been superimposed.

The solution of 1 me of protactinium with carrier
thorium showed an initial ?**Pa count about one-
third of that expected on the basis of 1-mc solutions
made previously in salt. The protactinium stayed
constant (though with wide scatter) for 16 days at
600 and 750°C with mote thorium addition. When the
metal was cooled to 350°C so that ~98% of the tho-
rium should precipitate, the protactinium fell (after
an unexplained one-day lag) to ~12% of its former
value. Reheating to 700°C and then to 725°C {where
all of the thorium should redissolve) brought the prot-
actinium back to only 30% of its initial value. When
salt was added, protactinium did not appear until
enough BiF, had been added to oxidize half the
thorium added; then 82 to 86% of it appeared in the
salt phase. In fact, protactinium should not have
appeared until practically all the thorium was oxi-
dized, so there must bave been about a 50% reductant
loss.
150

5 ORNL-DWG 87-11823

 

ol 7o — S e ]

s00C° \ 750° ‘ 350° ‘700"

 

>

counts min ! (g of B!

 

 

 

 

L [ { - e — e e —
L]
STATIC J

®

— 0 AGITATED
A AVERAGE
A

SALT {(AS Bi EQUIVALENT}

Q A0 20 3G 40 50 60 70
TIME (days)

 

 

Fig. 12,3. Stability of Solutions of Protactinium in Bismuth,

Behavior of this sort cannot be explained by a vigorously, suggestiog that the protactinium was on
simple mechanism. However, slow reversible ad- a dispersed phase rather than on the vessel walls.
sorption (negligible at high temperature) followed by Strips of various metals were immersed for 5 min
irreversible diffusion into a solid (but oxidizable) and 2 hr in the bismuth to see how much protactinium
phase is consistent with the data. This phase does was adsorbed; results are shown in Table 12.1. When
not seem to be the thorium-bismuth intermetallic. the metals were dipped directly into bismuth, niobium
After the cooling step, the metal samples, which were picked up the most protactinium, with beryllium, iron,
taken with an open dipper, tended to show more prot- and carbon picking up less (carbon, much less). Ad-
actinium when the liquid metal was being stirred sorption on iron reached about one-third of its 2-hr
151

 

 

 

Table 12.1. Adherence of Protactinium to Metal Samples at 725°C
Pa on Metal Bi Equivalent to
Metal E'Kplr)sure Are; (counts/min) Pz in ?11 » Counts of Pa (g)
Time (cm®) T T T (counts min™ g~ ) T
Total Per om Total Per om?
x 104 x 10% x 103

Be-1 2 hr 5.52 6.27 1.14 5.9 11 1.9
Nb~1 2 hr 5.22 15.4 2.95 5.8 27 5.1
Fe-1 2 hr 5.52 2.84 0.514 5.2 5.5% 0.99
Nb-2 2 hr 4.34 11.2 2.58 5.1 22° 5.1
Fe-2 5 min 5.52 0.796 0.144 4.9 1.6 0.29
Fe-3 2 hr 5.52 2.59 0.469 4.9 5.37 0.96
C-1 2 hr 11.94 0.170 0.014 4,7 0.36 0.03

After C-1, a salt layer of 72.5 mole % LiF and 27.5 mole % BeF2 was added to the experiment,

Metal samples were held in this salt for about 2 hr to remove surface oxides hefore ex-

posure to the bismuth.
Nb-3 2 hr 4.34 0.292 0.0673 3.4 0.86 Q.20
Nb-4 5 min 4.34 1.07 0.247 3.0 3.6 0.82
Nb-5 2 hr 4.34 0.617 0.142 3.0 2.1 0.47
Fe-4 2 hr 5.52 3.66 0.663 3.3 11 2.0

 

Weight gain (adhering bismuth) less than 1 g.
Pa counted differentially at 305 to 315 kev.

value in 5 min. When the metals had been exposed to
salt to remove oxides, the amounts of adsorption were
similar; but now, niobium picked up less protactinium
than did iron, and the process was much faster, In
all cases the protactinium adsorbed was not a large
amount of the total.

12.3 ATTEMPTED ELECTROLYTIC DEPOSITION
OF PROTACTINIUM

D. G. Hill®>  H. H. Stone

A number of unsuccessful attempts to deposit
protactinium electrolytically from molten fluoride
mixes have been reported.®’” None of these experi-
ments involved the use of a standard reference elec-
trode such as the hydrogen—hydrogen fluoride
couple.® Summarized here are an experiment to

 

*Cons ultant, Duke University.

6C~ 1. Barton, MSR Program Semiann. Progr. Rept.

Aug. 31, 1966, ORNL-4037, p. 156.

c. J. Barton and H. H. Stone, MS5R Program Semiann.
Progr. Rept. Feb. 28, 1967, ORNL-4119, p. 153.

IqGr, Dirian, K. A. Romberger, and C. F. Baes, Jr., Re-
actor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNL.-3789, p. 76.

measure the deposition potential for thorium by using
an Hz-HF anode and a nickel cathode, and an unsuc-
cessful attempt to deposit protactinium with the same
electrolytic cell.

A nickel pot was fabricated for these experiments,
The salt compartment had a diameter of 2.5 in, and
was 8 in. high. It was fitted with two 0.5-in.-ID
chimneys welded to the top of the pot and extending
9 in. above the pot, high enough above the furnace to
permit the use of Teflon insulators for the electrodes
or electrode supports. One of the chimneys also ex-
tended to 1 cm above the bottom of the pot, and it
was closged at the lower end by porous nickel filter
material. Thus anode and cathode compartments wete
isolated in such a manner that bulk diffusion would
be very slight, but ions could migrate between elec-
trodes.

The pot was loaded with 1556 g of purified Lil-
Bel" -ThF, (73-2-25 mole %). We calculated that
the material would have a volume of 275 cm?® at
600°C and would fill the pot to a depth of 9 cm at
this temperature. Initially, % -in. nickel tubing was
inserted in both chimneys so that gas could be
passed through each compartment. A cylinder of
platinum gauze closed at the bottom and extending
1.5 cm below the end of the tube was tightly wired
152

to the end of the nickel anode. A mixture of H
and HF that passed through this electrode estab-
lished the reaction at the anode as

1/2H2+F_:HF+8—.

The cathode reaction was the deposition of any of
the cations present, particularly thorium at above its
deposition potential, and hopefully protactinium after
it was added to the melt.

For the preliminary study of the decomposition po-
tential of ThF4, the vessel was maintained at 650°C
while the cathode compartment was swept for 3 hr
with HF and then overnight with H,. The cathode
coinpartment was capped both at the inlet and the
outlet, and a stream of HF-H_ was introduced
through the anode tube. The gas flow was adjusted
to approximately 100 cm?®/min. The mixture was '
analyzed at the exit tube and was found to contain
3.8 mole % HF,

Electrolysis was performed with dry cells as the
source of power. The voltage was tapped off a
slide-wire voltage divider, and simultanecus read-
ings were taken of the voltage and the current.

Four runs were made, starting each time at zero
voltage and measuring the current at 0.2-v inter-
vals. The plot of current vs voltage given in Fig,.
12.4 shows a decomposition potential in this cell
of —0.86 v.

The mole fractions of each of the reactants are re-
lated by the equation

RT [HF]
E=FE - ln e
0 F [Hg]l/Z [T}lF4]1/4
4.59 {(923)
~-086=K -0 -
o 2.3 «10*
ORNL--DWG 67--118724
AOQ g -~ T
. 75 .
g /}
é 50 /f— -
44
o5 | L _ ;/ T
g
A
0 e 8 e :‘,;__,_ 7“4%_J
0 05 1.0 L5
Ev)

Fig. 12.4. Current-Voltage Curve.

| (0.038) )
0F e

& (0.962)172 (0.25)t 74

from which E0 = —1.09 is calculated for the reaction

2H, 4 ThF, = 4HF ¢ Th.

After completing these measurements the tlow of
HF was cut off, while H, flow continued, After 1 hr,
both electrodes were raised out of the melt, and the
H, tlow was allowed to continue until the pot was
cold.

For the study of protactinium removal, the cathode
compariment was opened with as much care as pos-
sible to avoid the entrance of moisture, and to that
compartment was added 5 g of LiF-ThF  which had
been irradiated in a reactor to give a 233Pa activity
of about 1 mc. The salt also contained approximately
27 mg of ?31Pa. The vessel was closed and installed
in the furnace in a glove box in the High-Alpha Molten-
Salt Laboratory. After HF treatment for 4 hr, followed
by H, overnight, a filtered salt sample was counted for
233pa_

The anode gas mixture was adjusted to be approxi-
mately the same as that used in the previous run ex-
cept that a slightly higher HF concentration, 7.5
mole %, was used. The decomposition potential for
ThF, at this concentration of HI from Eq. (1) should
be --0.92 v, so electrolysis was carried out at - 0.90 v.
The current averaged about 10 ma, although it was less
for the first 20 min, after which it rose rather rapidly
to the 10-ma value expected from the earlier experi-
ment. Assuming 100% Faraday efficiency, the depo-
sition of 27 mg of protactinium would require 75 min
at 10 ma. Electrolysis was carried out for 165 min,
more than twice the minimum time for complete
deposition.

The gas at the end of the experiment contained only
4.1 mole % HF, although the H, rate was unchanged.
There is no indication at what time during the elec-
trolysis the decrease in HF occurred. At this lower
concentration of HF the decomposition potential is
calculated to be —0.87 v, so that if the decrease in
rate took place early in the electrolysis period, the
applied potential was too high.

At the conclusion of the run the cathode was raised
out of the melt, thus interrupting electrolysis. A
sample of melt was then taken, and both cathode and
sample were counted with a multichannel analyzer.

The sample of melt gave essentially the same count
as before electrolysis. A number of gamma peaks were
observed in the cathode spectrum, but none corre-
sponded to the ??3Pa peaks, which were strong in the
melt sample. The location of the peaks and the
limited data on their decay rates indicate that these
activities can be attributed to various mewbers of the
thorium decay series, such as 21%2Pb, ?12Rj, and
20871, although insufficient decay rate data were
obtained for conclusive identification. The lack of
233Pa activity on the cathode, coupled with almost
identical counts on samples of melt taken before
and after electrolysis, indicated that deposition of
a significant amount of protactinium on the cathode
did not take place in this experiment.

The failure to reduce protactinium in this experi-
ment may be explained by its low concentration in
the melt. The amount added, 27 mg in 1556 g of
LiF-BeF -ThF,, is approximately equal to a mole
fraction of 2 x 1073, If the protactinium is four-
valent in the melt, and if it is assumed that its
EG 15 close to that of thorium, that is, 1 v, the
potential required to reduce it at that concentra-
tion and at the HF-H_ ratio used in the electrol-
ysis is 1.007 v. The very small denominator makes
the logarithm term positive, so that the reduction
potential is even more negative than Eo in spite
of the low HIF pressure.

The same argument would deny that protactinium
can be reduced by thorium at this concentration,
which is contrary to fact. The nicke! cathode did
not provide a low-activity alloy in this attempt.
This suggests the possibility of using a cathode
that would form a thorium alloy in which protactinium
dissolves at low enough activity to permit reaction.
For further experiments of this type we recommend
a potential of 1.1 v, at which thorium will be de-
posited, hopefully, along with protactinium, and
stopping the experiment when five equivalents of
thorium have been deposited. This would give a
concentration of about 0.2 for protactinium in the
metal, and thus a low enocugh activity to permit
reduction of essentially all the protactinium.

12.4 PROTACTINIUM STUDIES iN THE HIGH-
ALPHA MOLTEN-SALT LABORATORY

C. J. Barton H. H. Stone

In the previous progress report’ we mentioned the
possibility that iron coprecipitated with protactinium,
when solid thorium was exposed to molten LiF-’_l‘hF’;,

might be carrying the reduced protactinium to the steel

wool surface in the Brillo process. It was also
pointed out that variable iron analyses made it dif-
ficult to determine the role of iron in protactinium re-

duction experiments. We performed one reduction
experiment with 5%Fe tracer dissolved in LiF-ThE,
(73-27 mole %) in the absence of protactinium, and
the results are summarized below. Most of the
protactinium recovery experiments performed during
the current report period were thorium reduction tests.
An unsuccessful effort to deposit protactinium elec-
trolytically is discussed in the preceding section of
this report.

Reduction of Iron Disselved in Molten LiF-ThF4

We studied the reduction of iron dissolved in
molten LiF-ThF  (73-27 mole %) by using hydrogen
and metallic thorium as the teducing agents. The
tracer iron tesults indicated that more than 40 hr
was required to reduce the iron concentration from
330 to 2 ppm with hydrogen at a temperature of
600°C. During a 3-hr exposure to metallic thorium
at the same temperature, the iron concentration in
filtered samples, calculated from tracer counts,
diminished from 550 to 13 ppm.

Colorimetric iron determinations performed by two
different laboratories agreed with the tracer iron data
for about half the samples. In general, agreement
was poorest for samples that gave °?Fe counts in-
dicating an iron concentration less than 0.1 mg/g.
It appears that the colorimetric iron method tends
to give high results with samples having a low
iron concentration. This finding is in agreement
with results of earlier unpublished studies per-
formed by Reactor Chemistry Division personnel.

The results of this experiment will be discussed
in more detail in another report.

Thorium Reduction in the Presence of
an lron Surface (Brille Process)

One Brillo experiment of the type discussed in the
previous report’ was performed during this report
period. A salt solvent, LiF-ThF  (73-27 mole %},
containing initially 17 mg of ?*'Pa and 57 mg of Fe
was exposed to thorium rods for two 1-hr periods in
the presence of 4 g of steel wool (0.068 m? per g of
surface area). The tracers 233Pa and *°Fe were
added to aid in following the behavior of these ele-
ments in the experiment. The first thorium exposure
removed 95% of the protactinium, as determined by
analysis of a filtered sample of salt, and almost all
the iron. The second thotium exposure resulted in
only a slight further decrease in protactinium con-
154

centration (97% removed). The distribution of prot-
72.8% in the steel wool
plus untransferred salt, 4.3% in the unfiltered salt
transferred away from the steel wool, 8.7% in the
steel liner and dip leg, and 12.0% in samples, for
a total of 97 8% recovered. The data obtained in
this experiment confirmed the resulis of previously
reported” experiments that indicate the Brillo
process may warrant further examination.

actinium was as follows:

Therium Reduction Followed by Fiitration

We reported earlier® that a large fraction of the
reduced protactinium that would not pass through a
sintered copper filter was found in samples of un-
filtered salt. This suggested the possibility of col-
lecting the reduced protactinium on a metal filter
from which it could presumably be removed by dis-
solving it in a molten salt after passing HF through
the filter. We performed several experiments to test
the efficiency of protactinium recovery by filtration.
The initial treatment of molten LiF—"l"I'lI*"4 (73-27
mole %) mixed with enough ?3'Pa to give a con-
centration of 20 to 60 ppm was performed in unlined
nickel pots. The molten salt was treated first with
mixed hydrogen and HF, followed by a brief hydmgen
treatment before it was transferred through a nickel
filter into a graphite-lined pot equipped with a
graphite dip leg. Here the reduction of protactinium
was carried out in the usunal fashion, either by ex-
posure to a solid thorium rod or to thorium turnings
suspended in a nickel basket, taking samples of fil-
tered and unfiltered salt after each thorium treatment
of the melt.
back into the nickel pot through the transfer filter.

The reduced melt was then transferred

In four experiments the amount of protactinium
found in the transfer filter varied from 10 to 30% of
the total amount present. This represented 40 to
95% of the amount of protactinium suspended in the
reduced salt (average, 69%). The amount of prot-
actinium in the graphite liner and dip leg varied
from 20 to 57% of the total (average, 33%). The
data show that a filtiation method will not catch
protactinium on the filter, but nevertheless the re-
moval of protactinium from a melt does appear fea-
sible.

One experiment was attempted with a niobium
liner and dip leg. The reduction of protactinium
with thorium proceeded normally (12% of the Pa re-
mained in the filtered salt after 2 hr of thorium
treatment), but transfer of the reduced salt through
the filter could not be effected because of a clogged

filter. A considerable amount of grayish material
was found in the bottom of the pot after it cooled to
room temperature. A sample of this material was
reported to contain only 0.35 mg of Nb per g, but

it is quite possible that this amount of impurity in
the molten salt would have been sufficient to clog
the filter.

The effect of iron on the hehavior of protactinium
in thorium reduction experiments has not been de-
fined urambiguously as yet, but we continue to find
reasonably good coirelation between the distribution
of iron and protactinium in these experiments. Count-
ing of 2?*Pa and *%Fe in both solid samples and
solutions of samples provided a check on the ac-
curacy of 2?!Pa alpha pulse-height analyses and
colorimetric iron determinations.

Conclusion

Thorium metal is an effective agent for reducing
protactinium in molten fluoride breeder blanket
mixtures, but further study will be required to de-
termine the best method of separating the reduced
protactinium from the salt mix.

12.5 MSBR FUEL REPROCESSING BY
REDUCTIVE EXTRACTION INTO
MOLTEN BISMUTH

W. R. Grimes
J. H. Shaffer

D. M. Moulton
F. F. Blankenship

An electromotive series for the extraction of
fission products from 2LiF-BeF  into liquid bis-
muth has been constructed in the way described for
the MSBR blanket materials in the preceding report
in this series.® Briefly, standard half-cell reduc-
tion potentials are calculated for each metal, using
as the standard states the ideal solutions in salt
and bismuth extrapolated from infinite dilution to
unit mole fraction. The exception is lithium, for
which the standard state in salt is 2LiF-BeF .
(This standard state is also used for beryllium, but
it is not assigned a standard state in bismuth be-
cause of its very low solubility.) This choice of
standard states is the normal practice when deal-
ing with dilute solutions. These standard poten-

tials are called 590' to distinguish them from 5’:‘0,

IMSR Program Semiann. Progr. Rept. Feb. 28, 1967,
ORNI.-4119, p, 150,
the standard potential for reduction to pure metal.
Thus, 86 is the potential corresponding to the ex-
traction process and does not require the simulta-
neous use of metal activity coefficients that are
far from 1 at infinite dilution.

Ii should perhaps be noted that the electromctive
series is thermodynamically equivalent to other
means of expressing the free energy of the extrac-
tion process, such as equilibrium constants or de-
contamination factors. It is used because it per-
mits ranking the metals in an order which directly
indicates their relative ease of extraction regard-
less of their ionic valence and because it has
proven useful in the experiments where a beryllinm
reference electrode is used to monitor the reduc-
tion process,

From data previously reported!? the series has
been constructed as shown in Table 12.2. The
numbers in parentheses after the rare earths are
the potentials calculated by using the experimental
fractional valences. For beryllium, 80, the poten-
tial for reduction to the pure metal, is shown. All
of these except Be? * are calculated relative to the
Li" value.

Although better measurements may change these
values somewhat, they indicate the relative ease
of extraction. The numbers in parentheses are
probably a better approximation to the true values
despite the peculiar valences. All elements up
through europium can be extracted completely be-
fore metallic beryllium begins to form; this forma-
tion of Be? represents the ultimate limit of the
process. An order of magnitude change in the ratio
[(mole fraction in metal)/(mole fraction in salt)]
corresponds to 0.087 v for divalent and 0.058 v for
trivalent species.

 

Table 12.2. = at 600°C (volts vs Hy—HF = 0.00)

Li 1.02

Bel® 1.81 &)

Bal? 1.79

pu?t 1.62 (1.61)
Na3t 1.58 (1.51)
sm2t 1.58 (1.49)
ce?? 1.87 (1.45)
La3? 1.52 (1.47)

 

155

We have begun a reinvestigation of fission
product extraction. In this study we will measure
fission product distribution as a function both of
[ithium concentration and of temperature with
(hopefully) improved accuracy. Preliminary resulis
of the first of this series, using cerium, are shown
in Fig. 12.5. A failure ended the experiment hefore
pood concentration data could be gotten, but those
we did get indicate a valence close to 3, not 2.3 as
before. The apparent minimum of E‘(; (Ce) at 700°C
is not explained and is rather hard to believe. For
lithium, ‘r.}f(;
lithium concentration and the potential between the
bismuth pool and beryllium metal electrode, for
which the temperature coefficient iz known. "1 The

was calculated from the measured

 

PSR Program Semiann. Progr. Rept. Feb. 28, 1966,
ORNI.-3936, p. 141.

11 .. .. ;
C. ¥. Baes, Jr., Reactor Chem. Div. Ann. Progr.
Rept. PDec. 31, 1965, ORNL-3913, p. 20.

 

 

  

 

 

 

 

 

 

 

 

 

ORNL DWG €7~ 14825
2-1 }
20 |— L“W F S E U A
Li FROM Be
- EQUILIBRIUM
\\.
.
CT >\k ® i
Li HIS WORK - ARGONNE
18— —— D
>
AT —— ] e
g
16 L e e
S o
-~
\ -
P
s -~ -
e T -
400 500 00 T00 800 500
7 (°C)
Fig. 12.5. &’ for Lithium and Cerium.

Q
’

point marked with a cross is the value of 80(1,1)
found earlier by equilibrating metallic beryllium
with Li-Bi amalgams. Its distance from the line in
this experiment may be the result of slightly in-
creased lithium activity at higher concentration

| X -0.16].

LB
E‘,‘.(; for lithium calculated from the data of Foster

The line marked ‘‘Argonne”’ is

and Eppley'? on the excess chemical potential of
Li in Bi. The rather small uncertainty here prob-
ably does not matter much in the separation
process, but it is important, in preparing the elec-
tromotive series, to refer all potentials to the same
The potentials at
600°C measured in this experiment are - 1.89,
-1.81, and —1.46 for Li, Be, and Ce respectively.
The extraction coetficients for the reaction are
given in Table 12.3. This method of presentation
shows more clearly than the potentials the sub-

lithium or beryllium potential.

stantial decrease in extraction with rising temper-
ature.

The difference between this value of Q at 600°C
and that reported earlier by Shaffer!? is mainly due
to his use of 2.3 rather than 3 as the exponent.

The extraction equilibria look very favorable for
using this process in MSBR fuel reprocessing. Re-
covery of ‘Li from the bismuth will be helpful and
should not be difficult. A perplexing problem
which has occurred frequently is the apparent loss
of total reductant species during the course of the
extraction. The extent of this loss is not easy to
measure because of uncertainties in the lithium
analyses, but averages 25 to 50% of the lithium
added. We think we have eliminated the possi-
bility that the loss is due to exiraneous reactions
such as lithium volatilization, dissolution of Li,
Be, or Bi’~ in the salt, or formation of solid beryl-
lium or beryllides. In the Zr-U separation, the
usual experimental technique was changed to min-
imize the possibility of air coming in the access
ports (normally filled with hot helium) during addi-
tions or sampling. The reductant loss was cut
sharply to ~10%, which is within the accuracy of
the measurements. In another experiment we ob-
served a substantial increase in oxide content over
the initial value; because of probable BeO precipi-
tation, the amount of oxygen added could not be

 

12M. S. Foster and R. Eppley. Chem. Engr. Semiann.
Progr. Rept. July—Dec. 1963, ANL-6800, p. 405,

13]. H. Shaffer, MSK Program Semiann. Progr. Rept.
Feb. 28, 1966, ORNL-3936, p. 144,

156

Table 12.

3. Extraction Coefficient for Cerium

Extraction intc Bismuth

 

XCe(m) 4
X =Xy i
Ce(s)

Q T (C)
9.1 % 108 456
4.0 108 545
8.7 % 107 600
2.1 %10 697
2.1 x10° 817

 

determined. A few back oxidations with a meas-
ured amount of BiFs have given more or less 100%
efficiency, indicating that no unknown reduced
species takes part in the reoxidation step. It
seems, therefore, that at least a substantial part of
the reductant loss is due to air entry and can be
eliminated by improved experimental techniques.
Since there is no evidence that any other reduction
process is going on, we do not feel that this reduc-
tant balance anomaly in the laboratory experiments
will have any important bearing on the process
when it is put on a practical scale.

12,6 REDUCTIVE EXTRACTION OF CERIUM
FROM LiF-BeF, (66-24 MOLE %) INTO
Pb-8i EUTECTIC MIXTURE

W. J. Hunt'? W. . Bull'®
J. H. Shaffer

Current investigations on liquid-liquid extraction
as a method for reprocessing the MSBR fuel are di-
rected toward the use of bismuth-lithium mixtures
for removing fission products from the reactor fuel
solvent. A complementary program will survey
other metal phase extractants for possible adapta-
tion in the reprocessing method. This proposed
experimental program will tieat the distribution of
cerium, for reference purposes, between LiF-Bek,
(66-34 mole %) and various molten metal mixtures

14 . .
ORAU summer student from University of Missis-
sippi.

15 . .
Consultant, University of Tennessee.
2

as functions of temperature and metal-phase com-
position. Experimental procedures will essentially
duplicate those developed for the primary program.
The results presented here on the reductive extrac-
tion of cerium from LiF-BeF  (66-34 mole %) into
the Bi-Pb eutectic mixture (56.3 at. % Bi) are the
first of this experimental program.

As in previous related experiments, the distribu-
tion of cerium between the two liquid phases was
followed as a function of the lithium concentration
in the metal phase. However, only small fractions
of lithium added incrementally to the metal phase
remained in solution under equilibrium conditions.
Typical results obtained at 700°C (Fig. 12.6) show
that only 32% of the cerium could be removed from
the salt phase. In addition, analysis of the metal
phase could account for no more than 1 to 2% of
the cerium activity. Variations in temperature from
500 to 700°C had little or no effect on either the
distribution of cerium between the two phases or on
the lithium concentration of the metal phase.

Attempts were made initially to measure the
electrical potential between a beryllium metal
electrode inserted in the salt phase and the molten
metal pool. Despite short contact periods of the
Be® rod with the salt mixture, substantial quanti-
ties of cerium activity accumulated on the beryl-
lium electrode. A large fraction of cerium activity
missing from the two liquid phases collected on
the beryllium electrode during its accidental over-
night exposure to the salt mixture.

157

ORNL~-DWG 67 —-11826

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12 oo e

I [ | i [
i o ANALYSES OF SALT PHASE
2 o ANALYSES OF METAL PHASE
-
o MO e WEIGHT OF SALT PHASE: 2336 kg ]
2 WEIGHT OF METAL PHASE: {526 kg Bi
2 1474 kg Pb
& ooeN —_—
<
£
5
% B e \ _‘;;‘s_._“ ................ 4
8 e O, o o
Z
i_??- 4 — i ————
o3
£
=
2
E 21— . O O SR R
(&)

® ¢
o """ *

O { 2 3 4 5 6
LITHIUM ADDED TO METAL PHASE {(mole fraction xv102)

Fig. 12.6. Reductive Extraction of Cerium from LiF-
BeF2 {(66-34 mole %) into Molten Lead-Bismuth Eutectic
Mixture at 700°C.

Although the results of this experiment are some-
what incoanclusive, they indicate that solid rare-
earth beryllides might be stable in this extraction
system. Because of this rather pronounced effect
of lead, the system Pb-Bi can probably be excluded
from use in the liquid-liquid extraction process for
rare-carth removal from the MSBR fuel solvent.
13. Behavior of BF; and

13.1 PHASE RELATIONS
IN FLUOROBORATE SYSTEMS

C. J. Barton J. A. Bornmann'
I.. 0. Gilpatrick H. Insley?
T. N. McVay? H. H. Stone

Interest in low-melting, low-cost molten salts
for use as the coolant in molten-salt breeder
reactors has prompted reexamination of phase rela-
tions in fluoroborate systems. The system NaBF -
KBP‘4 was studied earlier at this Laboratory, *
while Selivanov and Stender* reported low-melting
eutectics in the systems I\IaF‘—NaBF4 and KF-
KBE . More recently, Pawlenko® investigated the
ternary system KBE -KF-KBF ,OH and its as-
sociated binary systems. Varying melting points
reported in the literature for the compounds NaBF
and KBF  indicated the need for pure compounds
and care in heating them to prevent hydrolysis or
loss of BFs'

We found that recrystallization of commercial
preparations of NaBF , and KBF , from dilute
hydrofluoric acid solutions, usually 0.5 M, gave
compounds with higher melting points than any
previously reported. We find 569°C for KBF
(previous high, 552°)° and 406°C for NaBF
(previous high, 368°).* The differential thermal
analysis (DTA)® melting curve for K]BF4 was much

 

1 ..
Research participant from Lindenwood College, St.
Charles, Mo.

2Censultant.

3R. E. Moore, J. G. Surak, and W. R. Grimes, Phase
Diagrams of Nuclear Reactor Materials, R. E. Thoma
(ed.), ORNL-2548, p. 25 (Nov. 6, 1959).

4V. G. Selivanov and V. V. Stender, J. Inorg. Chem.,
USSR, IK2), 447—-49 (1958).

’S. Pawlenko, Z. Anorg. Aligenm. Chem. 336, 17278
(1965).

®L. 0. Gilpatrick, R. E. Thoma, and S. Cantor, Reactor
Chem. Div. Ann, Progr. Rept. Dec, 31, 1966, ORNL-
4076, pp. 5—6.

158

Fluorcborate Mixtures

sharper than that of NaBK ; this probably indicates

higher purity of the potassium compound, but

repeated crystallizations of NaBI*‘4 from dilute HF

solutions failed to give a preparation having a )
melting point above 406°. We are considering .
alternative methods of purifying this compound.

The principal techniques used in the investiga-
tion of phase relations in fluoroborate systems
were DTA® and quenching.” By use of these
techniques we have examined the binary systems
NaF-NaBF“ and KF-KBF | as well as the compound
joins NaBF -KBF  and NaF-KBF  in the ternary
system NaF-KF-BF .. Diagrams for these systems
are presented in Figs. 13.1 to 13.4. All except
NaBF -KBF  are simple eutectic systems, although
a small thermal effect at 435°C in KBF  -rich NaF-
KBF, mixtures suggests that the compositions
used were slightly off the compound join and that
a small amount of NaF-KF-KBF , eutectic liquid
was formed. Large thermal effects due to the
KBF , phase transition (283 t 2°C) and the
NaBF , transition (243 * 2°C) were noted in the
DTA curves. Efforts to retain the high-temperature
forms of the pure compounds and their mixtures
with NaF or KF by rapid quenching were
unsuccessful.

The phase diagram for the NaF-NaBl system
(Fig. 13.1) indicates that the previously reported
data on this system®* showing a eutectic melting at
304°C and containing 37.5 wt % Nal3F  is grossly
in error. The most recently reported work® on the
KF-KBT | system agrees with our data (Fig. 13.2)
on the location of the eutectic composition, but we
found that the eutectic temperature was 14° higher
than reported. We found no evidence of the com-
pounds KF - KBF, and KF - 2KBF, reported® in this
system. The system NaF-KBF, (Fig. 13.3) has
apparently not been studied by other investigators.
The diagram for the NaBF -KBF, system (¥Fig. 13.4)

 

. J. Barton et al., J. Am. Ceram. Soc. 41(2}, 63
(1958).
ORNL-DWG 67 - 9425A
e e e

MPERATURE {°C)

TE

 

 

 

 

 

 

 

  

NaBF

maie % NaBF4

Fig. 13.1, The System Naf-NaBF ,,

ORNL- DWG 67- 3474

800 | T N }

   

 

- —

 

TEMPERATURE {°C)

 

 

 

 

 

 

 

 

 

 

 

mote % KBF4

Fig, 13.2. The System KF-KBF4_

indicates that the high-temperature modifications
of NaBF | and KBF ,, believed to be cubic, form
a complete series of solid solutions exhibiting a
shallow minimum (392°C) near 15 mole % KBF ,.
Subsolidus relations are not clear at present. A
thermal effect at about 190°C is evident in all
mixtures in the system, while other subsolidus
effects vary with composition. Because of the
previously mentioned difficulty in guenching in
the high-temperature forms of the compounds and
their mixes, resolution of the subsolidus relations
in this system may require detailed investigation
with a high-temperature x-ray diffractometer.

We have started to map the phase boundaries and
isotherms in the region of the ternary system
bounded by the compounds NaF-NaBF -KBF  -KF.

159

 

600 | —— | — .

 

500 |- ..

TEMPERATURE (°C)

 

 

 

 

 

 

 

 

 

 

 

-
ol 5
O j@]
O O
o :
|
|
—————— &
i
|
]

 

L

 

 

 

 

 

 

 

200 1 [ SR .
Nof 20 40 &0 80 KBF,
mole 7% KBF4
Fig. 13.3. The System NaF-KBF .
ORNL-DWG 67-9426
500 T T T T T
> DTA 1.1GUIDUS 1
550 « OTA SOLIDUS
S & QUENCH LIQUIDUS \
T oo i & QUENCH 30LIDUS
L)
&
2 ’ {
g 450 e o \ e : st -
L
: ~I
= 400 | J : fi \ :
F.. ! L
350 ] e Lo Lo
-:!iOO el l.! N [ _J {J AL_L.
KBk, 20 40 O 20 NoBiy

£
maole % MNadh,

Fig. 13.4. The System NaBF ,-K8F ,.

Only a few compositions in addition to the two
compound joins (Figs. 13.3 and 13.4) have been
examined to date. Compositions in the ternary
system containing more than 50 mole % BF
probably exist in equilibrium with very high pres-
sures of BF3-

13.2 DISSOCIATION YAPOR PRESSURES
IN THE NaBF ,-NaF SYSTEM

Stanley Cantor

To determine BF dissociation pressures from
salt mixtures composed of sodium fluoride and
sodium fluoroborate, a static method is being used
in which the melt and its vapor are essentially
enclosed in a metal-glass system. The apparatus
ORNL-DWG &7-11827
VACUUM

=—x}= TO TAYLOR PRESSURE
TRANSMITTER

 

 

s

=——x} GAS SAMPLING
STATION

THERMOCOUPLE
WELL~

|

Hg MANOMETER — .

 

    

 

7

L__I PRECISION ||~
PRESSURE ||~
GAUGE

clebriele el bbb

 

 

 

 

 

 

(it

FURNACE

Fig. 13.5. Apparatus for Measuring Decomposition

Pressure of Fluoroborate Melts,

is shown schematically in Fig. 13.5. The melt,
contained in nickel, exerts its vapor pressure on
a mercury manometer and on diaphragm gages. To
remove adsorbed gases, the vessel, containing a
weighed salt sample, is initially evacuated at 100
to 125°C for at least 15 hr, Then to release gases
encapsulated within the salt crystals the sample is
heated about 25°C above its liquidus temperature.
The sample is subsequently cooled down to room
temperature. If the apparent pressure is greater
than 0.5 mm at room temperature, the system is
reevacuated and then reheated and cooled until
the vapor pressure at room temperature is essen-
tially nil,

After this desoibing procedure, the cbserved
vapor pressures should be caused only by the
dissociation

NaBF (1) - BF ,(¢) + NaF(0) .

Gas samples collected during the vapor pressure
measurements and analyzed mass spectrometrically
confirmed that the vapor was virtually pure BFs.
Usually about 1% SiF4 is observed in the vapor,
this probably originates from the reaction of
desoibed HI" with the glass parts of the apparatus.

After completing vapor pressure measurements,
the containment vessel is cut open and the contents

ORNL DWG 67-11828

 

 

 

 

10,000 - - P S 2
~—e— MEASURED o // 1
T ——o— EXTRAPOLATED P
5000 - » | A gy 4
- T S
L o
— | 0 1'/
| ¢ /O/ //? e o
e _ P
2000 399/-@ //,9 /o/o //
/ — ~
T ! 839?/ s %/ | f
1000 - T - ,;,/
' Q * - ! ‘ .
. = ‘/i 1
¥ 500 !/ ?7/59./: e /
8 , '
£ | | . 7 s
. i
‘U: / ,(OO/' ./. /
. .
x : ! 1
> 200" ‘ a/‘ - i
; / 660., /
A . .
[ ’ : ; ,
" oot | 9/ ot
10 : '/ O ' .
, . . © ; ,
. . . .
/ ; /g o /j
50 . . - O . ! a
® . ; . , S )
‘/l ‘/.
| o4
20 </,' . «
. | . /.
. e
10 - - . e A —

 

 

o
o
—~J
o
~
o
@
O
oo
wn
0]
o
o
o
o
&2

mole % Ng E}F4

Fig. 13.6. Smoathed Vapor Pressures in the NGBF4-
NaF System.

are weighed and chemically analyzed for oxygen
and boron. The oxygen analysis is used as con-
firmation that the system had been leak-tight. The
boron analysis and change in weight are comple-
mentary measurcments of how much BE‘3 was lost
during the course of desorption and vapor pressure
measurement, To obviate composition changes in
the melt because of the inevitable loss of some
BF , a relatively large sample of salt is charged
into the containment vessel.

Measurements have been taken on compositions
between 65 and 100 mole % NaBF .
are depicted in Fig. 13.6. To a fair approximation,
the data may be represented by the equilibrium
quotient

Some results

Qp — (pressure of BF 3)

(inole fraction of Nak')

" (mole fraction of NaBF )
ORNL -DWG 67-11829
TEMPERATURE (°C)
900 850 800 750 700 650 600 550 500
{

 

1000

 

 

 

300

 

200

100

 

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

08 02 1.0 4 12 13
1900 /7 ek

Fig. 13.7. Equilibrium Quotient As a Function of
Temperature for the Reaction Mo BF4(I) = NaF() +
BF,(g).

Values of {  vs temperature are plotied in Fig.
13.7; as might be expected, log (_)p is linear with
1/T(°K).

From the BF | pressure data, thermodynamic
activities of NaBF  and NaF can be evaluated.
This evaluation will be carried forth in the very
near future.

13.32 REACTIONS OF FLUCROBORATES
WiTH CHROMIUM AND OTHER
HASTELLOY N CONSTITUENTS

Stanley Cantor

This investigation seeks to determine the corro-
sion reactions of Hastelloy N and its constituents
with fluorchorate salt melts. Thus far, three
initial experimenis have been carried out; the
conditions and results are summarized in Table
13.1. The nickel containment vessels and the
procedure for desorbing gases other than BFs used
in these experiments were the same as those used
in the vapor pressure measutements (see Sect.
13.2). The salt mixture was composed of 92-8 mole
%o NaBF -NaF; this mixture, which is the only
known eutectic in the NaBF -NaF system, is under
consideration as a substitute coelant for the MSRE.

In the experiment containing chromium mefal
chips (see Table 13.1), the vapor pressures were
measured at several temperatures between 500 and
700°C. After the sample had been af 500°C or
above for 26 hr, BF , vapor pressures were twice
those from the melt alone; that is, the reaction
between chromium metal and salt had yielded
BF | gas.

In the postexperimental examination of the ves-
sel’s contents, the chromium metal chips appeared
blackened and corroded and weighed less than
prior to the experiment. The solidified salt mixture
contained a green solid. Both the chromium metal
chips and the green solid were separated from
adhering fluoroborate by dissolving the water-
soluble fluoroborate. The green solid, when
examined microscopically, showed relatively large
(as much as 50% in volume) black inclusions. From
an x~ray powder pattern the green solid was iden-
tified as predominantly crystalline Na ,CtF . Two
moderately intense lines in the pattern could not
be identified; these, however, cannot be attributed
to chromium metal or to any boride of chromium for
which an x-ray diffraction pattern is available.

Chemical analysis of the green solid showed
27.2% chromium and 0.64% boron. In stoichiometric
NaJCrFB, the chromium content is 22.1%. Presum-
ably, all the boron and the chromium in excess of
22.1% can be assigned to the black inclusions.

An electron microprobe examination of the green
solid, scanning 360 ;?, confirmed the presence

of B, Cr, Na, and F. A limited area scan ditected
at a black surface (of 1 1?) on one crystal yielded
Table 13.1.

162

Summary of Initial Cosrosion Experiments

All experiments were carried out in nickel vessels with 92-8 mole % NaBF“‘—NaF

Total Hours at

Form of Specimen
500°C or Above

Cr Large irregular- 26
shaped chips

Hastelloy N Cylindrical rods 59

Fe Wire® 28

Mo Strips? 28

a .
In the same experiment.

a chromium concentration approximately twice as
large as for the sample as a whole. Boron, a dif-
ficult element to determine by microprobe analysis,
was not detected in this latter scan.

Thus, this first experiment indicates rather
substantial chemical reaction of chromium with
molten NaBF -NaF. Two certain reaction products
are Na_CrF_ and BF ,. The chemical nature of the
black inclusions is not readily apparent. The
presence of boron and the high concentration of
chromium, not as metal, suggest that the black
inclusions (in the green solid) may be a boride of
chromium whose x-ray diffraction pattern is as yet

Maximum Results

Temperature (°C)

 

0
-AGlooo
NaBF | 368 t
NaBCrF’6 581 t
BF3 256 &
Cr B 25 +
X

(Formation)

10

710

Increase of vapor pressure with
time. Formation of green crystals,
largely NaSCrFfi, but containing
black inclusions whose Cr content
is higher than in pure NaSCrFs;
green crystals also contain boron
{(valence unknown). Relatively
extensive coirosion — loss of -

0.56 g of Cr out of 20 g. "

600 Small weight loss (0.017 g out

of 37.8 g of metal).

=

No visible

color in salt,

650 Loss of 209 by weight of wire .~

no change in salt color.

No weight loss or discemible
attack.

 

A chemical reaction that describes the
above observations may be written as follows:

unknowi.

3NaBF4(1) - (1 + x)Cr(s) = NaSCrFfi(s)
+ ZBF3(g) + CrXB(S) ,

where x > 2. For this reaction, AG?OOO, the

standard free energy at 1000°K, is estimated to be
~14 1 20 kcal. The following table gives the free
energies of formation:

Reference or Method of Estimation

From BFS dissociation equilibria
By analogy with formation data of Na3A1F6
JANAYF Thermochemical Tables

Based on estimate of CrzB by K. E. Spear,

Metals and Ceramics Division
163

The two other experiments, one with Hastelloy N
and the other with iron and molybdenum, were not
monitored for vapor pressure. After desorbing the
gaseous impurities from the sample charge (metal
plus salt), the vessels were brought to the tempera-
tures shown in Table 13.1 and were kept there for
the indicated time. After completing these two
experiments the samples were examined; very little
or no visible attack was apparent on the metal
specimens. Also, there were no visible color
changes in the salt mixtures. However, the weight
losses in the Hastelloy and in the iron specimens
(sce Table 13.1) suggest the need for more thorough
investigation of the corrosion reactions of chromium
and iron with fluoroborate salt melts.

Apparent Mass Tronsfer of Nickel

After completing vapor pressure measurements
on each fluoroborate mixture, the nickel vessels
are always cut open for examination of the con-
tents. In almost every case, the top surface of the
solidified salt contains a small amount of black
material. In all cases the inner metal surfaces are
shiny. The top portions of two salt cakes (one
97.5-2.5 mole %, the other 65-35 mole % NaBF -
Nal') when dissolved in water yielded silvery
residues. These residues were ferromagnetic, and,

by z-ray diffraction, were identified as nickel metal.

The reason that nickel metal particles appeared
on the melt is not readily evident. It is unlikely
that metal particles were present in the vessel
prior to loading with salt, nor is it likely that the
stock salts contained nickel metal. The luster of
the walls in contact with salt suggests that nickel
may have been mass transferred. Further investiga-
tion should provide additional information on this
unexpected deposition of nickel metal on the salt.

13.4 REACTION OF BF , WITH CHROMIUM
METAL AT 650°C

J. H. Shatfer H. ¥. McDuffie
Current plans to replace the secondary coolant
of the MSRE with a fluoroborate mixture have
prompted studies of the compatibility of these
materials with structural metals of the reactor
system. Since fluoroborates of interest to this
program exert measurable vapor pressure of 1BF3
at operating temperatures, covering atmospheres

containing equivalent concentrations of BE, must
be maintained in the free volume of the pump bow!
of dynamic systems or used for gas sparge opera-
tions. This experimental program will examine

the reactions of BF | with various structural metals
and alloys that might be applicable to the program.
Preliminary results obtained by contacting BF
with chromium metal at 650°C ate presented here.

The experimental method utilizes a 30-in. length
of 2-in. IPS nickel pipe, mounted horizontally in a
3-in. tube furnace, as the reaction chamber. The
reacting gases are admitted through a penetration
in the end plate that is welded to one end of the
nickel pipe. A sheathed thermocouple also pene-
trated the end plate and extended into the central
region of the heat zone. The other end of the reac-
tion chamber, which extends some 10 in. out of the
tube furnace, is closed by Teflon in a threaded pipe
cap. The gas manifold system provides for the
introduction of helium, BFs, or mixtures of these
gases at known flow rates into the reaction chamber,
The system is sealed from the atmosphere by
bubbling the gas effluent through a fluorocarbon oil.
Concentrations of BF _ in helium can be determined
continuously from the recorded signal of a calibrated
thermal conductivity cell. Metal samples or speci-
mens are carried in nicke! boats inserted through
the threaded access port.

The reaction of BF, with chromium metal was
followed by periodic determination of weight gain
of about 10.88 g of prepared chromium flakes
during a 60-hr reaction period at 650°C. The
chromium metal was prepared by electrolytic deposi-
tion on a copper sheet which was subsequently dis-
solved by an acid leach. The thia film of chromium
metal which remained was broken into smal] flakes
for this investigation. Reaction periods commenced
by introducing BF | at 1 to 2 ce/min after heating
the sample to 650°C in flowing helium. The sample
was also cooled to room temperature under flowing
helium to permit inspection and weight gain deter-
minations,

As shown by Fig. 13.8, the weight gain of the
chrominm sample showed a linear dependence on
the square root of reaction time., The overall reac-
tion period showed that the chromium sample in-
creased its weight by about 4%; there was no
significant difference in the weight or appearance
of the nickel boat during this experiment. Examina-
tion by x-ray diffraction techniques showed that the
chromium sample contained substantial quantities

of Crz()s; minor fractions of the mixed fluoride,
ORNL-DWG &7—14830
044 I | T

INITIAL

 

 

 

0.40

)

 

WEIGHT GAIN OF REACTANT {

 

 

 

 

 

 

 

\

 

8

 

|

4
1
SQUARE ROOT OF REACTION TIME (hr}/2

O 1 2 3 5

Fig. 13.8. Reaction of BF 3 with Chromium Metal at
650°C.

CrF, « CrF , were identified by petrographic tech-
niques, The material is currently being chemically
analyzed for boron.

Although these results are inconclusive with
respect to identification of the oxidation species
in the gas phase, they do illustrate that the direct
use of chemically pure, commercially available
BF3 in high-temperature systems will promote the
oxidation of nearly pure chromium. Further studies
will evaluate the effects of BF _ concentrations on
oxidation rates and will investigate methods for
improving the purity of BF .

13.5 COMPATIBILITY OF BF,
WITH GULFSPIN-35 PUMP OIL AT 150°F

F. A. Doss PP. G. Smith
J. H. Shaffer

The proposed use of a fluoroborate mixture as the
secondary coolant in the MSRE will require that a
covering atmosphere containing BF3 be maintained
above the salt in the pump bowl. Since BF  is

164

known to catalyze the polymerization of certain
organic materials, its effect on the lubricating
properties of the pump oil needs evaluation.
Although these effects will be observed directly
during the planned operation of the PKP-1 loop
with a fluoroborate salt mixture, a preliminary ex-
periment is in progress to determine relative
polymerization rates of Gulfspin-35 oil under con-
ditions which can be related to actual pump opera-
tions. The degree of polymerization should be
indicated by measured changes in oil viscosity
during the experiment.

MSRE-type pumps use a helium purge down the
pump shaft to isolate the lubricated parts of the
pump assembly from the gas environment of the
pump bowl. Previous tests on the prototype pump
loop, using 3 Kr as an indicator, showed that this
isolation technique reduced the concentration of
pump bowl gases at the cil-gas interface to about
1 part in 20,000.% Accordingly, this current study
provides accelerated test conditions by contacting
the pump oil with helium containing about 1000
ppm of BF3 at a maximum operating temperature of
150°F.

Two experimental assemblies have been operated
concurrently to provide comparative data. In each
assembly, helium was bubbled at a rate of about
1 liter/min through 1.5 liters of pump oil. The BF3
was introduced into the helium influent stream to
one experiment at a rate of about 1 cc/min. Sam-
ples of the oil were drained from each experiment
periodically and submitted to the Analytical Chem-
istry Division for viscosity measurements. As
described in a later section, a continuous gas
analysis system was installed and calibrated by
A. S. Meyer, Jr., and C. M. Boyd of the Analytical
Chemistry Division. Concentrations of BF , in the
gas influent and effluent of the experiment are
recorded from the output signal of a thermal con-
ductivity cell. The light hydrocarbon content
(products of o0il polymerization) of the gas effluent
can also be monitored on a semicontinuous basis.

The results obtained through approximately 600
hr of continuous operation of the experiments are
illustrated in Fig. 13.9. Although some discolora-
tion of the oil exposed to BF3 was noted, there is
no distinguishable difference in oil viscosity
between comparative samples. The increase in oil
viscosity (original value, 15.7 centistokes) during

 

8 A. G. Grindell and P. G. Smith, MSR Program Semi-
ann, Progr. Rept, July 31, 1964, ORNL-3708, p. 155.
25 % { (centistoxes )

VISCOSITY GF OIL AT
W

ORNL-DWG &7-11831

 

 

[ ! 1 l l 1
o-QIL CONTACTED WITH 8F 3 (~1000 ppm)
a0 | IN HELIUM.

*-0OIL CONTACTED WITH HELIUM ALONE.

 
 

 

 

 

 

|

VOLUME OF OiL IN EACH EXPERIMENT: 1.5 liters.
GAS FLOW RATE IN EACH EXPERIMENT:1 liter/min.

 

 

 

 

 

 

0 100 200 300 400 500 600 TOOQ
CONTACT TIME (hr)

 

 

 

Fig. 13.9. EHect of BF 3 on Viscosity of MSRE Pump
Oil (Gulfspin-35) at 150°F.

165

the experiment is no more than 0.3 centistoke at
25°C. Concentrations of BF, in the gas effluent
have remained essentially the same as influent
concentrations throughout the experiment. Hydro-
carbon concentrations in the gas effluent were
insignificantly low. On the basis of these results,
the introduction of BF3 as a covering atmosphere
in MSRE-type pump bowls should have negligible
effects on this lubricating propetty of the pump
oil. The experiment will be continued to examine
long-term effects of BF, on the pump oil.
14. Development and Evaluation of Analytical Methods

for Molten-Salt Reactors

J. C. White

The determination of oxide in highly radicactive
MSRE fuel samples was continued. The replacement
of the moisture-monitor cell was the first major
maintenance performed since the oxide equipment
was installed in the hot cell.

The U*" concentrations in the fuel samples run
to date by the transpiration technique do not reflect
the beryllium additions which have been made to
reduce the reactor fuel. This may be accounted for
by an interference stemming from the radiolytic
generation of fluorine in the fuel samples. This
problem will receive further investigation. Ex-
perimental work is also being carried out to develop
a method for the remote measurement of ppm con-
centrations of HF in helium or hydrogen gas
streams.

Design work was continued on the experimental
molten-salt test loop which will be used to evaluate
electrometric, spectrophotometric, and transpiration
methods for the analysis of flowing molten-salt
streams.

Controliled-potential voltammetric and chrono-
potentiometric studies were carried out on the
reduction of U(IV) in molten fluoride salts using
a new cyclic voltammeter. It was concluded that
the U(IV) - U(III) reduction in molten LiF-BeF -
ZrF, is a reversible one-electron process but that
adsorption phenomena must be taken into account
for voltammetric measurements at fast scan rates
or for chronopotentiometric measurements at short
transition times.

An investigation of the spectrum of U(VI) in
molten fluoride salts has been initiated. It was
found that the spectrum of Na UF_ dissolved in
LiF‘-BeF2 in an 3i0, cell with SiF, overpressure
was identical to the spectrum of UO F , dissolved
under identical conditions. It appears that the

166

equilibrium concentration of 02~ may be sufficient
to react with the components of the melt, An
attempt to use the SiOQ—SiF4 system in the spec-
trophotometric investigation of electrochemically
generated species in molten fluorides also met
with difficulties. The Sil", overpressure interferes
with cathodic voltammetric studies by causing very
high cathodic currents.

It is planned to install a spectrophotometric
facility with an extended optical path adjacent to
a high-radiation-level hot cell to permit the observa-
tion of absorption spectra of highly radioactive
materials. The basic spectrophotometer and as-
sociated equipment have been ordered.

The effects of BF | on MSRE pump oil have been
investigated. Measurements were made of increases
in hydrocarbon concentrations of an He-BF | gas
stream after contact with the oil. A thermal con-
ductivity detector was used to monitor the BF 3
concentration in the test gas stream. -

Development studies are being made on the
design of a gas chromatograph to be used for the
continuous determination of sub-, low-, and high- .
ppm concentrations of permanent gas impurities
and water in the helium blanket gas of the MSRE.

This problem of analyzing radioactive gas samples
prompted the design and construction of an all-
metal six-way pneumatically actuated diaphragm
valve. A helium breakdown voltage detector with
a glass body was designed and constructed to
pernnit the observation of the helium discharge.
Under optimum conditions this detector has ex-
hibited a minimuin detectahle limit below 1 ppb of
impurity. It appears to be possible that the detector
will also operate in the less-sensitive mode nec-
esgary for the determination of high-ievel con-
centrations of impurities in the blanket gas.
14.1 DETERMINATION OF OXIDE
IN MSRE SALTS

R. F. Apple J. M. Dale
A. 8. Meyer

During the last week of December the moisture-
monitor cell in the oxide apparaius became inop-
erative. Because of other experiments being
petformed in the same hot cell, the moisture-
monitor cell was not replaced until March. The
insensitive cell showed some supetficial evidence
of radiation damage in that the potting compound
(an RTV preparation which is used to seal the tube
containing the spiral electrodes in a stainless
steel housing) had shrunk and cracked. Flow
checks revealed that substantially all the flow was
still passing through the electrolysis tube, so that
the damage to the potting compound could not have
been responsible for the cell failure, Resistance
measurements indicated that the failure was
caused by either removal of or some alteration to
the PO electrolyte film.

The analyses of eoxide in radioactive salt samples
from the MSRE for this period are summarized in
Table 14.1.

Two samples of radioactive fuel (IPSL-19 and
IPSL-24), submitted from the In-Pile Salt Loop 2,
were found to contain 265 and 240 ppm of oxide
respectively. Sample IPSL-24 was stored under
helium at 200°C for a period of about six months
from the time of sampling uatil the analysis was
made,

14.2 DETERMINATION OF U3*
IN RADIOACTIVE FUEL BY A HYDROGEN
REDUCTION METHOD

J. M. Dale R. F. Apple
A. S. Meyer

A transpiration method is currently being used
to determine the U3" concentration in molten
radicactive MSRE fuel. The molten fuel is sparged
with hydrogen to reduce oxidized species according
to the reaction

n - m

MF 4 —5 H, —> MF_ + (0 — ;) HF ,

in which MF, may be UF _, NiF,, FeF,, CiFF,, or
UF4 in order of their observed reduction potentials.

167

Table 14.1. Oxide Concentrations of MSRE Salt Somples

g Oxide Concentration

 

Sample Date Receive
(ppm)
FP-11.28 (fuel) 3-21-67 58
FP-12~4 (flush) 6-17-67 41
FP=12.18 (fuel) 7=11.67 57

 

The rate of production of HF is a function of the
ratio of oxidized to reduced species in the melt.
The theory of the method has been described pre-
viously.’

The computer program which was under develop-
ment has been completed and permits the calcula-
tion of expected HF yields for any preselected
reduciion steps on any melt composition. Using
the present fuel composition and the experimental
conditions of the transpiration experiment as input
data to the program, HF yields were calculated for
varying initial concentrations of U%*, Sample
concentrations of U3 " were determined from the
comparison of the experimental and calculated HF
yields. Table 14.2 shows the U® " results obtained
from the HF yields of the third and fourth reduction
steps of the analyses and compares them with ex-
pected values calculated by W. R. Grimes.

The calculated results assume that 0.16% of the
uranium in the fuel was originally present as U3,
that the chromium concentration increase from 38
to 65 ppm which occurred before the first sample
was taken resulted in the reduction of U*™ to
U3 +, that each fission event results in the oxida-
tion of 0.8 atom of U3+, and that there have been no
other losses of U3¥,

It will be noted that no analysis results are
listed for samples FP-11-38 and FP-11-49. Although
these samples were run in the normal manner, a
total of over 2000 micromoles of HF was evolved
for the four hydrogen reduction steps for each
sample as compared with about 55 micromoles for
the previous runs. Since this increased HF yield
coincided with an increase in activity in the traps
used to collect the HF, it appeared likely that the
buildup in sample activity during the extended
period of reactor operation might be responsible.

 

11, M. Dale, R. F. Apple, and A. S. Meyer, MSR
Program Semiann. Progr. Repf., Feb. 28, 1967, ORNL-
4119, p. 158,
168

Table 14.2. Concentration of U3+ in MSRE Fuel Salt

 

 

Ay Burnup U3+/Utota1’ U3+/U , Analysis (%)
Sample No. Mwhr Oxidation Added Calculated total -

{(equivalents) (equivalents) (%) Step III Step 1V
FI?.9-4 10,978 1.87 0.31 0 0.1
FP-10-25 16,450 0.93 3.61 0.58 0.35 0.45
FP-11-5 17,743 £2.40 .54 0.37 0.37
FP-11-13 20,386 0.26 2.59 0.77 0.37 0.42
FP-11-32 25,510 0.88 0.69 0.33 0.34
FP-11-38 27,065 0.26 0.66
FP-11-49 30,000 0.50 1.86 0.80
FP-12-6 32,450 0.40 0.76 0.42 0.37
FP-12.11 33,095 0.10 3.95 1.14 0,38 1.2
FP«12-21 35,649 0.50 4.44 1.54 0.39 .50

If the induction period for the radiolytic generation
of fluorine were shortened due to the increased
activity level of the sample, the fluorine evolved
during the loading of the sample could react with
the inner walls of the Monel hydrogenation vessel.
The copper and nickel fluorides formed would be
subsequently reduced during the hydrogenation
steps to produce HF.

The above hypothesis appears to be supported
by the results of the following experiment. One
of the samples which produced the high HF yields
was allowed to stand in the hydrogenator at room
temperature for about a week. The sample was
then subjected to additional hydrogenation steps,
and HF was produced in quantities comparable
with that obtained in the original runs. After
standing several more days at room temperature,
the sample was removed from the hydrogenator.
Smaller but significant quantities of HF were
obtained when the empty hydrogenator was
subjected to the high-temperature hydrogenation
procedure,

The last three samples, FP-12-6, FP-12-11, and
FPP-12-21, were all taken after a relatively brief
period of reactor operation following a lengthy
reactor shutdown period. None of these analyses
produced the excessively high HF yields which
were observed for the previous two samples. This
appears to be further confirmation that the exces-
sive HF yields resulted from a buildup in sample
activity with extended reactor operation.

Since the first addition of beryllium to the fuel,
a1l the determinations of U®" not obviously affected

by radioactivity have fallen in the 0.33 to 0.50%
range (the one result of sample FP-12-11 could be
explained by a leaky valve) and do not reflect the
beryllium additions in the periods between the
samplings. This could be accounted for by the
evolution of fluorine in much smaller quantities
than appeared to be the case in samples FP-11-38
and FP-11-39. If this is the case, the only per-
manent solution would be to maintain the samples
at 200°C during the time of transfer to the hot

cell for analysis. However, an apparatus is now
being designed which will permit the hydrogenation
of synthetic fuel samples under carefully controlled
conditions. It is felt that this experiment will
provide a check of the validity of the transpiration
method and will give further evidence as to whether
or not the fluorine evolution is a real problem.

Experimental work is also being carried out to
develop a method for the remote measurement of
ppm concentrations of HF in helium or hydrogen
gas streams. The technique is primarily for appli-
cation to the U?”™ transpiration experiment but, if
successful, should also be applicable to the deter-
mination of HF in the MSRE off-gas. The method
is based on the collection of HF on a small NaF
trap which is held at 70°C to prevent the adsorp-
tion of water. This is followed by a desorption at
a higher temperature to give a concentrated pulse
of HF that can be measured by thermal conduc-
tivity techniques.

A components testing facility has been set up
which includes a dilution system to produce HF
‘‘standards’’ as low as 20 ppm, a thermostatted
trap with self-resistance heating for fast heatup,
and a themmal conductivity cell with nickel fila-
ments. Initial tests have revealed that the thermal
conductivity cell presently being used is too
sensitive to perturbations in the carrier flow and
that the last traces of HF are desorbed too slowly
from commercial pelletized Nal". The use of
thermal conductivity cells of different geometry
and better flow-control valves is expected to
eliminate the {irst of these problems. The second
problem will require further development studies to
delermine whether the slow desorption rate repre-
sents an inherent property of the NaF or is a result
of impurities in the NaF. Other trapping waterials
will also be investigated.

A special Monel valve has been made and will
be used in the remote HF measuring system. The
valve incorporates two valving systems in the same
metal body and will provide for the simultaneous
adsorption and desorption of HF from alternate
traps. Figure 14.1 shows a schematic of the flow
pattern. This arrangement will eliminate any dead
legs in the HF trapping system and will permit the
measurement of incremental quantities of HF
evolved from one hydrogenation step in the U3"
transpiration experiment.

One additional step is being taken with regard
to the computer program which is being used for
data analysis. Due to equilibrium shifts of
oxidized and reduced species in the molten fuel
salt with temperature changes, the starting U3 *
concentration in the analysis sample at the tem-
perature of the initial hydrogenation steps will
necessarily be different from the U™ concentra-

ORNL-OWG 67-39009

LVENT HF

 

TRAP 1

THERMAL
CONDUCTIVITY
CELL TRAP 2

Fig. 14.1,
System.

 

 

|/ Hy PURGE

Schematic Flow Diagram of HF Trapping

tion in the fuel in the reactor. The computer
program is presently being modified to take this
into account,

14.3 IN-LINE TEST FACILITY

J. M. Dale R. F. Apple
A. S. Meyer

Design work has been continuing with the as-
sistance of J. H. Evans on the experimental
molten-salt test loop which will be used to eval-
uate electrometric, spectrophotometric, and
transpiration methods for the analysis of flowing
molten-salt streams.? The operation of flow
equipment such as capillaries, corifices, and
freeze valves will also be tested. A schematic
tlow diagram of the proposed test loop is shown
in Fig. 14.2. The first draft engineering drawings
have been completed, and it is planned to start
construction of the various components of the
system.

14.4 ELECTROREDUCTION OF URANIUM(1Y)
IN MOLTEN LiF-BeF -ZrF , AT FAST SCAN
RATES AND SHORT TRANSITION TIMES

D. IL.. Manning Gieb Mamantov?

Controlled-potential voltammetric and chrono-
potentiometric studies were carried out on the
reduction of U(IV) in molten LiF-BeF ,-Zik
(65.5-29.4-5.0 mole %). The controlled-potential,
controlled-current cyclic voltammeter was con-
structed in the Instrumentation Group of the
Analytical Chemistry Division at ORNL. In the
controlled-potential mode, scan rates from 0.005
to 500 v/sec are available and cell currents to
100 ma can be measured. In che controlled-current
mode, currents ranging from a few microamperes to
100 ma can be passed through the cell. The built-
in time base allows transition times from 400 sec
to 4 msec to be measured. The instrument can
also be operated in a potential-step mode for
chronoamperometric experiments. Readout of the
curves is accomplished with a Tektronix type 549
storage oscilloscope.

 

2“Ana1ytical Methods for the In-Line Analysis of
Molten Fluoride Salts,” Ansal. Chem. Div, Ann. Progr,
Rept. Oct. 31, 1966, ORNL-4039, p. 18.

3Consu1tant, Department of Chemistry, University of
Tennessee, Knoxville,
   

 

 

ORNL-DWG. G7-3792R

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

...... -
i B e
Q; e CONSTANT ! '
Ui =l HEAD POT (}g $—fi q,;* Jélfi
: | I | |
| TS| -
5
! o @ L0 it g
! 5 i Y,
| Z W l’h ! :u
Z T e
! 5 2 :I | ll>:-|
* -~} o I
] 5 Eu, i 131X
! T OUI i i [ @
! o = Ll
| | e [
I % a || ke |
| .-
: L&__.A\_ ______ J’\.-.L...,quFL..)
| _;l_,r,f\, _______ MM e e — - “-
z ) H,~ HF
8 r““‘T‘Al ,,,,,,,, Ao 5
a8 I HT"\----—;\G---;‘”‘-” . _fl__g;’fx wwwwww T e ANOF >
O I et |
< 1] 'l o Mo T b
N - | r; [‘_ - He
a E’ :‘g‘ || =
w el I HF REMOVAL
4 z mlht
O E - | ||
- s 5 5
&’ = Qo >| :|
wn o EII II.J'\ ______
E ml Il HOLD POT
g alyl -
= I .
o ; j Lo ~~~~T—-——Lfi1”1mmw—<-
|__ fi E | E : He
1 X -
t | N
' r
[ — -7 _Lt"# —— I
T - 7T _| 20 kg SALT
i — = RESERVOIR |
L - :
 — | ——a
| |
- — |
L. GL . r——O_I CIRCULATING
1 PUMP
] riq
|
| 1 |DRIER
r |
L )
b e e e e J

Fig. 14.2. Schematic Flow Diagram of an Experimental Moltzn-Salt Test Loop.

In LiF-BeF ,ZtF , at 500°C, U(IV) is reduced to
U(III) at approximately — 1.2 v vs a platinum quasi-
reference electrode. For voltammetry with linearly
varying potential, the peak current (ip) at 500°C is
given by the Randles-Sevcik equation as

i,=175x 10°a37%2AD Y 2Cvt/?

where v is the rate of voltage scan (v/sec) and the
symbols n, A, D, and C represent electron change,
electrode area (cm?), diffusion coefficient (cm?2/
sec), and concentration of electroactive species
(moles/cm?) respectively. The peak current is
proportional to the concentration of uranium and
also to the square root of the rate of voltage scan
(v'”?) from approximately 0.02 to 1 v/sec. The
diffusion coefficient at 500°C is approximately
2% 1078 cm?/sec. At faster scan rates, however,

the i vs v!/? plots frequently exhibited upward

curva{t)’rure, particularly at platinum and platinum-
rhodium indicator electrodes. It is believed that
this deviation from the Randles-Sevcik equation is
caused in part by adsorption of uranium on the
surface of the electrode. Conway* postulated that
for a charge transfer examined by the voltage sweep
method there are three contributions to the time-
dependent current: (1) a non-Faradaic current

C 4 @v/dt associated with charging or discharging
of the ionic double layer, (2) a pseudo-Faradaic
current C dv/dt associated with the change of
extent of coverage by adsorbed species formed or
removed in the electrochemical step, and (3)
Faradaic current iF associated with any net
reactions which can occur within the range of
potential scans employed. Therefore the net
current 1, that is recorded as a function of time
during a potential sweep is

itsz[dV/dt+CdV/dt-+iF ]

Since dv/dt is the scan rate (v) and i , kvl/2,

it can be written as
1
V1/2

where k = 1.75 x 10%23/24AD'/2C. Corrections for
adsorption plus charging effects, when encountered
at fast scan rates, were made by plotting ip/vl/2
vs v1/2, where the slope reflects adsorption
phenomena and the intercept contains the Faradaic
term. The diffusion coefficient for U(IV) calculated
from the intercept was in good agreement with the
value from the linear i vs v'’/ 2 plots.

In chronopotentiometry, adsorption of reactant
causes the io’fl/2 product, where 1 = current
density (amp/cmz), to increase as 7 (transition
time) decreases; this effect was observed for
uranium at short transition times (<100 msec).

The potential-time traces were not as well defined
as the voltammograms; howevei, reasonably precise
transition times could be established. Approximate
corrections for adsorption effects were made
utilizing a method set forth by Tatwawadi and

 

*B. E. Conway, J. Electroanal. Chem. 8, 486 (1064),

171

Bard.® The diffusion coefficient calculated from
chronopotentiometry was in agreement with the
voltammeiric value.

It is concluded that the U(IV) -s U(II) reduction
in molten LiF-Bel ,~ZyF , is a reversible one-
electron process. However, for the analyses of
current-potential curves at fast rates of voltage
scan or potential-time traces at short transition
times, adsorption phenomena could be pronounced
and must be taken into account.

14.5 SPECTROPHOTOMETRIC STUDIES
OF MOLTEN FLUORIDE SALTS

J. P. Young

Spectrophotometric studies of interest to molten-
salt reactor problems have continued. A major
portion of the work for this period has been con-
cerned with the use of S5i0, as a container for
LiF-BeF , melts. Spectrophotometric techniques
were applied to the evaluation of the compatibility
problems invelved. This work is reported in a
preceding section ot this report. An investigation
of the spectrum of U(VL) in molten fluoride salts
has been initiated. By means of spectral meas-
arements it should be possible to measure the
solubility of U(VI) in these solutions. The spec-
tral measurements can also be nsed analytically
to measure concentrations of U(VI) and to follow
the change in concentration of this species during
various reactions. This work is being done in
cooperation with G. I. Cathers, who has furnished
the solute salt, Na ,UF . The container material
for molten-salt solutions of U(VI) must be inert to
oxidation. It was hoped that Si0 , would satisfy
these requirements. It was found, however, that
the spectrum of Na UF  dissolved in LiF-BeF
in an Si0, cell with SiF, overpressure was
identical to the spectrum of UQ ¥ dissolved
under identical conditions. This would suggest
that the oxide concentration in the melt was suf-
ficient to cause the formation of uranyl ion. The
dissolved species at 550°C exhibited an absorption
peak at 419 nm and the foot of a strong absorption
peak in the ultraviolet with a shoulder at 310 nm.
The absorption peaks generally correspond quite
well to that reported® for UQ ? * in aqueous HCIO o

 

%S, V. Tatwawadi and A. J. Bard, Anal. Chem, 36, 2
(1964).

1. T. Bell and R. E. Biggers, J. Mol. Spectry. 22,
262 (1967); ibid., 22, in press.
This uranium species disappeared slowly when iron
wire was introduced into the melt, The resultant
spectrum was that of tetravalent uranium; thus, it
is demonstrated that an oxidized species of uranium
was originally present. It would appear, then, that
although SiO, containers are stable to oxidation,
the equilibrium 0%~ concentration may react with
components of the melt.

The work on electrochemical generation of solute
species and their spectrophotometric characteriza-
tion is continuing.” This work is carried out in
cooperation with F, L. Whiting® and Gleb
Mamantov.?® Again 510, would offer several
advantages in this study compared with window-
less cell techniques which had been used. Under
the experimental conditions presently required,
however, it was found that the presence of 1 atm
of SiF , gas over the melt under study interferes
with cathodic voltammetric studies by causing
very high cathodic currents. The reasons for this
are not yet understood, but the U(IV)/U(III) reduc-
tion wave is completely lost in the presence of
SiF,. Conversely, SiF , does not affect anodic
voltammograms. Since SiF  is a product of fluoride
salt reaction with SiOz, these results suggest that
Si0, containers may have limited use in voltam-
metric studies.

The development of a recording spectrophoto-
metric system is continuing that will permit spec-
tral measurements of highly radioactive solutions
to be obtained on samples located in a high-
radiation-level cell. The system will make use of
the extended optical path length similar to that
being considered for use in the in-line spectral
measurements of molten-salt reactors. The basic
spectrophotometer and associated eguipment have
been ordered, and the design of the physical and
optical arrangements of the compouents for the
extended path is being considered. The optical
arrangement will be such that both windowed and
windowless cells can be used.

14.6 ANALYSIS OF OFF-GAS
FROM COMPATIBILITY TESTS OF MSRE
PUMP OIL WITH BF,

C. M. Boyd

A total hydrocarbon detector and a thermal
conductivity detector have been used to determine
the hydrocarbons and BF | in the gas from tests
on the effects of BF , on MSRE pump oil (Gulf-

spin 35). These tests are described in an earlier
section. A gas sampling manifold permits the
measurement of increases in hydrocarbon concen-
tration of an He-BF ; stream on contact with the
A parallel stream of helium through a separate
oil reservoir serves as a reference. The hydro-
carbon analyzer was modified by the addition of a
sampling pump which draws in the gas at near
atmospheric pressure, compresses it, and then
passes it through the analyzer. The portions of
the test gas to be analyzed were {irst passed
through a saturated solution of KF to remove the
BY and thereby protect the pump and other
components of the analyzer. This solution has a

oil,

low water vapor pressure, which prevents water
condensation in the analyzer. In the first tests,
with the BF , at the 2000-ppm level and the oil at
150°F, the hydrocarbon level in the off-gas was
less than 50 ppm.

The thermal conductivity detector was used to
monitor the BF | concentration in the test gas,
which was produced by mixing flows of pure BF |
and He. The detector and flow capillary were
calibrated by passing a known volume of the gas
mixture through the detector and then through a
trap of dilute NaOH. The solution was analyzed
for boron and fluoride, and the BF _concentration
of the gas was then calculated. The sensitivity
of this detector under the conditions nsed was
10 ppm of BF ..

14.7 DEVELOPMENT OF A GAS
CHROMATOGRAPH FOR THE MSRE
BLANKET GAS

C. M. Boyd A. S. Meyer
Development studies are being made on the
design of a gas chromatograph to be used for the
continuous determination of permanent gas im-
purities and water in the helium blanket gas of the
MSRE. The chromatograph will require two columns
to separate the components desired. Tests have
shown that a paralle! column system composed of
a 5A molecular sieves column and a Porapak S
column should be practical for analyzing gases
which are not highly radioactive. The H , O,

 

7]. P. Young, MSR Program Semiann. Progr. Rept.
Feb, 28, 1967, ORNL-4119, p. 163.

8University of Tennessee, Knoxville.
Nz, CH,, and CO are separated on the molecular
sieves column, and the H,O and CO, are separated
on the Porapak column. The chromatograph will
tequire separate compartments controlled at dif-
ferent constant temperatures. The sample valve,
columns, and detectors may require different tem-
peratures for optimum operation.

The requirement that determination of impurities
be made at sub-, low-, and also at higher-ppm
levels necessitates the use of two different types
of systems or detectors. A helium breakdown
detector can be operated in either the high- and
low-sensitivity modes or in the high-sensitivity
mode in conjunction with a thermal conductivity
detector.

The chromatograph which is used for analyzing
the reactor off-gas will be subjected to radiation
from samples at the l-curie/cc level. This re-
quires a system free of organic construction
materials. An all-metal sampling valve and a

detector unaffected by this radiation are necessary.

The organic Porapak column cannot be used at
this level of radiation. This eliminated the pos«
sibility of H,0 analysis on these samples, A
silica gel column will be used to resolve CO..

The problem of radioactive samples and the
transmittal of gas samples containing ppm levels
of H,0 from the source to the detector require the
use of a heated all-metal sampling valve. Such
a valve has been designed and constructed. The
valve is similar to the conventional Phillips six-
way pneumatically actuated diaphragm valve (Fig.
14.3). The Teflon diaphragm has been replaced
by a 1-mil-thick Inconel diaphragm which contacts
and seals the redesigned entry orifices (Fig. 14.4).
With this metal diaphragm a spacer is required to
allow gas flow without excessive backpressure.
Spacers made from 2-mil gold were necessary to
give a leak-tight seal between the diaphragm and
the valve faces. The gold was annealed at 1500°F
and, after installation in the valve, subjected to
32,000 psi pressure. Pressure maintained by the
assembly screws was sufficient for retaining this
seal.

The choice of detectors which are sensitive to
sub-ppm levels of permanent gases is limited to
the helium ionization types. A helium breakdown

173

voltage detector was used on the MTR Capsule
Test Facility (test 47-6)° and was not affected

by the radiation present in the gas samples. This
may be explained by the relatively high current
levels (microamperes) used with this type of
detector. A constant-curient power supply is used
with this detector, and a decrease in breakdown
voltage, rather than an increase in ionization cur-
rent, indicates an increase in the electrical con-~
ductivity of the gas. The breakdown voltage of
pure helium is about 500 v and is lowered by about
50 v by 1 ppm of impurity. The minimum detectable
limit is controlled primarily by the noise level
present, but under optimum conditions is below 1
ppb.

In previous attempts to improve the stability of
the detector discharge, various electrode radii
were used with the anode and cathode in a con-
centric arrangement. In this early model of the
detector, which was contained inside a Swagelok
tube fitting, it was impossible to observe the
helium discharge. A test detector was therefore
constructed which has a glass body with Kovar
seal tube connections through which the electrodes
are mounted (Fig. 14.5). These electrodes have
removable tips which allow the testing of various
electrode shapes and spacings. The effects on the
helium discharge can be observed through the
glass. Tests with this detector indicate that a
minimum noise level is obtained with a smooth flow
discharge on the anode probe. Maximum sensitivity
dictates the use of a very pure helium carrier gas,
but this purity level also causes a sparking or
arcing in the helium discharge. The addition of
mercury vapor by the presence of a small source
of the metal in the tip of the anode stabilized the
discharge. The more practical solution of adding
a contaminant by a controllable gas flow is being
tested. This approach may also permit addition of
larger amounts of contaminant to allow the detector
to operate in the less-sensitive mode necessary
for the determination of high levels of impurities
in the blanket gas samples.

 

IMSR Program Semiann. Progr. Rept. July 31, 1964,
ORNL~3708, p. 328.
 

174

ORNL-DWG. 6577

 

 

 

—————> CARRIER
— > SAMPLE

— - =»A  TO SAMPLING LOOP
B——-——> FROM SAMPLING LOOP
> ACTUATING GAS

Fig. 14.3. Conventional **Phillips’* Valve.
.

       

ACTUATING VENT
PRESSURE

CONVENTIONAL "PHILLIPS" VALVE

       

  
    
    

D o
2mit GOl pC BN 22 R

     

!
30 ')’FLJ mil INCONEL

T

ACTUATING VENT
FRESSURE

ALL METAL "PHILLIPS' VALVE

Fig. 14.4. Modifisd All-Metal Sample Valve.

CRNL-LWG. 67-3043

OUTLET

3C4 55 TIPS

Fig. 14.5. Diagram of He!ium Breakdown Detector.
Part 4. Molten-Salt Irradiation Experiments

E. G.

Molten-salt breeder reactors are expected to
operate with a high rate of production of fission
products as a result of fuel salt power densities
in excess of 200 w/cc. The effects of long-term
exposure under such conditions on the stability
of fuel salt, the compatibility of salt with graphite
and metal (Hastelloy N), and the fate of the fis-
sion products are of interest. Undue buildup of
fission product poison on core graphite, for ex-
ample, could reduce the breeding efficiency of the
reactor if not mitigated. Initial loop experiments
are being directed largely at understanding the
fate of important fission products.

The second thermal convection in-pile loop ex-
periment was teminated by the appearance of a

Bohlmann

crack in the core outlet pipe, probably caused by
radiation embrittlement of the alloy and stresses
encountered during a reactor setback. Sufficient
operating time had, however, been achieved to
produce fission product concentration levels
equivalent to those estimated to be present at
processing equilibrium in a breeder; therefore, an
exhaustive evaluation of the experiment is in
progress.

Future loops will be fabricated of Hastelloy N
modified by titanium additions shown to inhibit
the radiation effects. Experiment objectives irni-
clude study of materals compatibility, fission
product behavior, and effects of operation at off-
design conditions.

15. Molten-Salt Convection Loop in the ORR

E. L. Comperse

Resulis of the first in-pile molten-salt convec-
tion loop experiment in this program have been
reported. Irradiation of the second molten-salt
convection loop in beam hole HN-1 of the Oak
Ridge Research Reactor began' January 12, 1967,
and was temminated April 4, 1967, after develop-
ment of 8.2 x 10!® fissions/cc (1.2% 233U burnup)
in the "LiF-Be¥ -ZtF -UF, (65.3-28.2-4.8-1.7
mole %) fuel. Average fuel power densities up to
150 w per cc of salt were attained in the fuel
channels of the core of MSRE-grade graphite.

 

. c. Savage, E. L. Compere et al., MSR Program

Semiann. Progr. Rept. Feb. 28, 1967, ORNI.-4110,
pp. 167--73.

H. C. Savage

J. M. Baker

The experiment was temminated alter radioactivi-
ty was detected in the secondary containment sys-
tems as a result of gaseous iission product leak-
age from a crack in the core outlet tube.

Operation and postirradiation examination of
the second loop are described below.

15.1 LOOP DESCRIPTION

A diagram of the second in-pile molten-salt coi-
vection loop is shown in Fig. 15.1. The core
section of the loop consists of a 2-in.-diam by
6-in.-long cylinder of MSRE graphite. Eight
vertical 1,/4—in~—diam holes for salt flow are bored
through the core in an octagonal pattemn with
177

 

   
 

 

ORNL -DWG 67-11332

Y LTYPICAL SALT LEVEL

~-CORE OUTLET PIPE

 

 

 

 

 

 

 

 

   
  
 
 

 

 

/

I

INCONEL
COOLING COILS

™

TR
RETURN LINE ")
(COLD LEG) }:\

SALT SAMPLE LINE

Fig. 15.1. Diagram of M

.'3,

centers 7 in. from the graphite center line. A
horizontal gas separation tank connects the top of
the core through a retum line to the bottom of the
core, completing the loop circuit. These and the
core shell were fabricated of Hastelloy N, as was
the 12-ft-long sample tube which connects the
loop to the sample station in the equipment cham-
ber at the ORR shield face. The Hastelloy N was
from material in fabrication of the MSRE and which
had not been modified to improve its resistance to
irradiation embrittlement. The electric heating
elements and cooling tubes surrounding the com-
ponent parts of the loop were embedded in sprayed
nickel.

Figure 15.2 is a photograph of the partially
assembled loop showing the configuration of the
salt flow channels in the top of the graphite core.
An identical configuration was used for the bottom
of the core section. Total volume of the loop was
110 cc, with a 43-cc salt volume in the graphite.

 
 

~NICKEL SPRAY‘.._\__?::“.

 
  

 

  
 
 

   
  
 

a4
Uy DL | -
}. it ::.O:::.
0’".‘

- 4~ THERMOCOUPLE WELL
9'0‘:’0:
35‘35& ~HASTELLOY N CORE BODY
(RN
i;fvm'/q -in.~diom FUEL
RO CHANNEL (8)

54 in.

 

 

 

 

 

 

 

 

clten-5alt In-Pile Loop 2.

15.2 OPERATION

After assembly, the loop was tested for leak-
tightness, flushed with argon gas, and vacuum
pumped at 600°C for 20 hr to remove air and
moisture. The loop was flushed with solvent salt,
then recharged with fresh solvent salt and operated
at temperature for 248 hr in the mockup facilities
in Building 9204-1, Y-12.

During the out-of-pile test period, 15 salt sam-
ples were removed from the loop and 12 salt addi-
tions were made, providing a good demonstration
of the salt sample and addition system.

The loop package containing the solvent salt
used in the preirradiation test period was in-
stalled in beam hole HN-1 of the ORR, and in-
pile operation was started on January 12, 1967.
Samples of the solvent salt ("LiF-BeF ,~ZIF 4,
64.8-30.1-5.1 mole %) were taken on January 13
before the start of irradiation and again on
PHOTQ 74244

 

 

   
   
    
 

#
COLD LEG RETURN LINE” %

HASTELLOY N BODY

“ SALT CHANNELS

Fig. 15.2. Photograph of Partially Assembled In-Pile Convection Loop 2 Showing Design of Salt Flow Channels

in Graphite Core.

January 16, 1967, after the reactor was brought to
its full power of 30 Mw. Loop operation was
continued with solvent salt under irradiation at
temperatures ranging between 550 and 650°C.
During this period the equipment and instrumenta-
tion were calibrated and tested, loop performance
was evaluated, and reactor gamma heat was de-
termined as a function of the depth of insertion of
the loop into the beam hole. lLoop operation was
entirely satisfactory during this period, and on
January 30, 1967, 7LiF—UF4 (63-27 mole %) eutec-
tic fuel (93% 23°U) was added, along with addi-
tional solvent salt, resulting in a fuel composi-
tion of "LiF-BeF ,-ZrF -UF, of 65.26-28.7-4.84-
1.73 mole %. The nuclear heat generated in the
loop (fission plus gamma) was again determined as
a function of the depth of insertion, two fuel salt
samples were removed from the loop, and on
February 21, 1967, operation at the fully inserted,
highest flux position was achieved. Operation in
the highest flux position was continued to the end
of ORR cycle No. 71 (March 5, 1967).

The ORR was down from March 5 until March 11,
1967, for the regular between-cycle maintenance
and refueling operations. Loop operation con-
tinued throughout this period. A fuel salt sample
was removed, and, by the addition of eutectic
fuel and solvent salt, the uranium concentration

of the fuel salt in the loop was increased to the
composition of 'LiF-BeF g tF AIF , of 65.4-
27.8-4.8-2.0 mole % in order to attain the experi-
mental objective of 200 w/cc fission-power densi-
ties in the fuel salt. Also a sample of the cover
gas in the loop was taken for analysis.

The ORR was brought to full power of 30 Mw on
March 11, 1967. When the molten-salt loop was
placed in the fully inserted, highest flux position
on March 14, 1967, it was found that the fuel
fission-power density in the graphite core was
150 w/cc instead of the expected 200 w/cc. This
resulted from a rearrangement of the ORR fuel
between cycles 71 and 72 which caused a reduc-
tion in thermal flux in beam hole HN-1 in an
amount sufficient to compensate for the increased
uranium in the loop fuel salt. This reduction in
flux was qualitatively confirmed by other experi-
menters in an adjacent beam hole facility.

When the ORR was started up on March 11,
1967, it was observed that the radiation monitor
on a charcoal trap in the loop container sweep-gas
line read 5 mr/hr, whereas normally this monitor
read zero. We concluded that this increased
activity was caused by radiation from a nearby
ORR primary coolant water line instead of fission
product leakage from the loop into the secondary
containment, and loop operation was continued.
Shortly after full power operation was reached on
March 14, the radiation monitor on the charcoal
trap in the container sweep-gas line increased to
18 mr/hr. Some 8 hr later a further increase to
~3.4 r/hr was noted. No further increase oc-
curred until March 17, when the radiation from the
charcoal trap increased rapidly (over a period of
~3 hr) to ~100 r/hr, indicating a significant
leakage of fission products from the loop. At this
point the loop was retracted out of the high-flux
region to a position at 1 to 2% of the highest flux,
and the fuel salt in the loop was frozen by re-
ducing the loop temperatures to ~400°C in order
to prevent possible salt leakage from the loop.

As a result of these actions, the charcoal trap
activity decreased to ™~ 1 r/hr over a 15-hr period.

From March 17 to March 23, 1967, the loop was
operated in a retracted position at 1 to 2% of full
power, and the fuel salt was kept frozen at a
temperature of 350 to 400°C, except for brief
periods of melting to determine the location of
the leak. It was concluded that the fission prod-
uct leak was in the vicinity of the gas separation
tank and that loop operation could not be con-
tinued. Three fuel salt samples were removed
from the loop during this period.

Beginning on March 27, 1967, the fuel salt was
drained from the loop by sampling to facilitate the
removal of the loop from the reactor and subse-
guent examination in hot-cell facilities. By this
procedure the fuel salt inventory in the loop was

179

reduced from 151.6 g to 2.1 g, requiring ten sam-
ples (12 to 25 g per sample). On April 4, 1967,
the ORR was shut down, and on April 5, 1967, the
loop package was removed from the reactor into a
shielded carrier and transferred into a hot cell
without difficulty.

During the in-pile operating period, fuel salt
containing enriched uranium was exposed to re-
actor irradiation for 1366 hr at average fission-
power densities up to 150 w/cc in the graphite
core fuel channels. For both out-of-pile and in-
pile operating periods, various salt additions and
removals were made, including samples for analy-
sis. During the in-pile period, the reactor power
was altered appreciably 38 times, and the distance
of the loop from the reactor lattice, that is, the
loop position, was changed 122 times. Thus, the
equivalent time at full loop power during fueled
operation was 547 hr, while the ORR was operated
for 937 hr. Table 15.1 summarizes the operating
periods under the various conditions for molten-
salt loop No. 2.

Details of some of the more significant observa-
tions made during operation and results of hot-
cell examinations and analyses are given in
sections which follow.

15.3 OPERATING TEMPERATURES

The salt in the loop was kept molten (> 490°C)
during all in-pile operations until it was frozen on

 

 

 

 

 

 

 

Table 15.1. Summary of Operating Periods for In-Pile Molten-Salt Loop 2
Operating Feriod (hr) Salt Additions and Withdrawals
Full Power
Total Irradiation Dose Equivalent Additions Samples Salt Removal
Out-of-Pile
Flush 77.8 7 1 13
Solvent salt 171.9 5 1 0
In-Pile
Preirradiation 73.7 1 2 0
Solvent salt 343.8 336.5 136.0 1 2 0
Fueled salt 1101.9 G37.4 547.0 2 3 0
Retracted-fuel removal® 435.0 428.3 11.2 0 4 9
Total 2204.1 1705.2 694.2 16 13 22

 

“Maintained at 350 to 400°C (frozen) except during salt-removal operations and fission product leak investiga-

tions.
March 17 following the fission product leak. At
full power, temperatures in the fuel salt ranged
from ~545°C in the cold leg return line up to
~720°C in the core outlet pipe. Since nuclear
heat was removed through the core graphite to the
cooling coils around the outside of the Hastelloy N
core body, the graphite temperature was ™~ 550°C,
or some 70°C below the average fuel salt tempera-
ture in the core. Temperatures of the metal walls
of the component parts of the loop (Hastelloy N)
ranged from a low of 510°C in the core body (where
maximum cooling was used) to ~730°C in the core
outlet pipe, where there were no provisions for
cooling.

The temperature distribution of the salt and loop
components was significantly altered when the
reactor was down. Under this condition, salt
temperatures ranged between ~535°C in the cold-
leg return line to ~670°C in the top of the core
graphite fuel passages. The core outlet pipe was
670°C with no nuclear heat, as compared with
~730°C for full power operation. The core outlet
pipe temperature of 730°C is believed to have
contributed to the outlet pipe failure as discussed
in a following section.

15.4 SALT CIRCULATION BY CONVECTION

In the first in-pile molten-salt convection loop,
the salt circulation rate of 5 to 10 cm?/min was
substantially below the calculated rate of ~45
cm®/min at operating temperature, and frequent
loss of flow occurred. In order to improve salt
circulation rate and reliability in the second loop,
the salt flow channels at the top and bottom of the
graphite core were redesigned to increase the
flow area and improve the flow path (refer to
Fig. 15.2). Further, the top and bottom of the
core section that were horizontally oriented in
the first loop were inclined 5° to minimize trap-
ping of any gas released from the salt, since it is
believed that formation of gas pockets often re-
sulted in flow stoppage.

Salt circulation in the second loop was estimated
to be 30 to 40 cm3/min; this was determined by
making heat-balance measurements arcund the
cold-leg return line and by adding an increment of
heat in a stepwise fashion to one point in the loop
and recording the time required for the heated salt
to traverse a known distance as monitored by
thermocouples around the loop circuit. This flow

180

rate is a fivefold increase over that observed in
the first loop and is attributed to the modification
described. However, occasional loss of flow still
occurred. One possible explanation for this is
that a sufficient temperature difference was not
maintained between the salt in the hot and cold
legs. This is supported by the fact that flow,
when lost, could be restored by adjusting the
temperatures around the loop circuit. Since oc-
casional flow loss did not adversely affect in-pile
operation, this was not considered to be a problem
of any serious consequence.

15.5 NUCLEAR HEAT, NEUTRON FLUX, .
AND SALT POWER DENSITY -

Nuclear heat was detemined at various loop
positions by comparing electric heat input and
cooling rates with the reactor down and at full
power (30 Mw). Figure 15.3 shows the results of
these measurements. Although nuclear heat meas-
urements based on such heat balances are not
precise because of variations in the temperature
distribution around the loop and variations in heat
loss at different loop positions (different power
levels) as well as necessary estimates of the
exact inlet and outlet temperatures of the several
air-water coolant mixtures, they provided a good
basis for determining nuclear heat and resultant
fission-power density during operation. Based on
these determinations, reactor gamma heat with the
loop fully inserted and filled with unfueled salt
was 4200 w. Fission heat was computed by sub-
tracting this value from the total heat generation -
obtained when fuel salt was in the loop.

In the earlier part of the operation with fueled ]
salt (1.73 mole % U), the loop contained 17.84 ¢ -
of uranium, 93% enriched, in a salt volume of
76.2 cc. A fission heat of 8600 w in the fully
inserted position was determined, leading to an
estimate of the average thermal neutron flux of
1.18 x 1013 neutrons cm™ % sec™?, or an average
fuel-salt power density of 113 w/cc. Assuming
from a neutron transport calculation by H. F.
Bauman that the core/average flux ratio was 1.33,
the average power density in the core salt was
150 w/cc at full power. The power density in the
forward core tubes is estimated by using Bauman’s
results to be 180 w/cc.

The average thermal neutron flux in the salt was
also independently determined from flux monitors
 

CRNL-DWG 67-11833

 

»
104 L N e
& A{SSION +GAMMA HEAT— ...
I,\ R L o———

 

 

 

 

 

  
 

AR HEAT {w)
~

NUCL
&
o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we .

23 43 6.3 8.3 103 12.3

LOOP POSITION-DISTANCE FROM REACTOR TANK TO LOOP
CORE CENTER (in.)

Fig. 15.3. Mucliear Heat Generation in Molten-Salt
Loop 2.

recovered during postirradiation disassembly of
the loop and from the fission product activity in
the final salt sainples, as discussed below.
Calculations involving activity of type 304
stainless steel monitor wires or of fission products
in salt samples required taking into account the
relative flux history for the particular isotope as
described in the section on activity calculation.
Thermal neutron flux levels calculated from type
304 stainless steel monitor wires attached to
various regions on the outside of the loop were
(in units of neutrons ecm~ % sec™ 1): core shell,
front, 4.4 to 5.5 x 10!3; coie shell, bottom (rear to
front), 1.7 to 3.8 x 10'3; central well in core
graphite, 2.7 to 3.4 x 10'3; core shell, rear, 1.4
to 1.9 x 10'3; gas separation tank, 1.4 to 2.6 x
103, Applying a calculated attenuation factor
of 0.6 and a fuel blackness factor of 0.8 to a

181

Table 15.2. Mean Nevutron Flux in Salt as Calculated

from Fission Product Activity

 

Flux Estimate

 

 

Isotope (neutrons cm ™2 sec_l)
Sample 23 Sample 26

x 1013 x 1013
30-y 137¢s (6.0%) 0.82 0.90
284-d 1**Ce (5.6%) 0.93 1.05
65-d 2°Zr (6.3%) 0.65 0.96
58.3-d °lv (5.8%) 1.14 0.58
50.4-d %%sr (4.79%) 0.61 0.76
32.8-d 1*!Ce (6.0%) 0.64 0.70
12.8-d 14%Ra (6.4%) 0.27 0.28
11.1-d 1*'Nd (2.6%) 0.59 0.66

 

weighted mean monitor flux of 2.4 x 10!3 results
in an estimated mean flux available to the fuel of
1.15 x 1013,

The flux was estimated on the basis of the
activity of varipus fission products in final salt
samples after accounting for the relative flux
history of the salt. Table 15.2 shows flux esti-
mates hased on the activity of a number of iso-
topes that might be expected to remain in the salt.

As is frequently the case, the flux estimates
based on fission product activity in salt samples
are somewhat lower than the values obtained in
other ways. Generally, chemical separations are
used prior to counting. Because of favorable half-
lives, well-established constants, low neutron
cross sections, good separations, and less inter-
ference from other isotopes in counting, 137Cs,
144Ce, and ?°Zr are regarded as the more reliable
measures of fissions (or flux). The average of the
flux estimates for these three isotopes is 0.88 x
1013,

The results from the different methods of esti-
mating flux are compared in Table 15.3. The
value of 0.88 x 1013 obtained from the fission
product activity is the most direct measure of
fissions in the fuel, and, therefore, it will be used
in estimates of the expected activity of other
fission products produced in the loop.

15.6 CORROSION

Evidence of corrosion was obtained from chemi-
cal analysis of salt samples withdrawn from the
Table 15.3. Comparison of Yalues for Mean Neutron
Flux in Salt Obtained by Various Methods

 

Mean Neutron Flux

 

Method (neutrons cm 2 sec"l}
Power generation rate 1.18 x 10'3
Type 304 stainless steel 1.15% 1013
monitor wires
Activities of 137Cs, '*4Ce, 0.88 x 103

and J S'Zr

 

loop and from metallographic examination of
samples cut from various regions of the loop.

A dissolved-chromium inventory based on analy-
sis of samples indicated that 13 mg of chromium
was dissolved by the flush salt, an additional
35 mg was dissolved during preirradiation opera-
tion, 20 mg more during solvent salt in-pile opera-
tion, and 19 mg more during the ensuing fueled
operation with fissioning, for a total of 87 mg
overall. The flowing salt contacted about 110 cm 2
of loop surface. A uniform 0.5-mil thickness of
metal (7% Cr) from over this area would contain
about 87 mg of chromium.

The largest increases in dissolved chromium
came during the first part of the run; this is con-
sistent with out-of-pile behavior reported by DeVan
and Evans.? There was no indication of any ag-
gravation of corrosion by irradiation.

The regions of the loop contacted by flowing
salt, particularly the core outlet and cold leg,
were seen on metallographic photographs to be
attacked to depths of 1/10 to 1 mil. This agrees
reasonably with the chemical value, which of
necessity was calculated on an overall basis.

The gas separation tank between the core out-
let line and the cold leg showed less attack than
the tubing sections, thereby indicating varying
susceptibility of different items of metal.

Examination of metallographic photographs of
Hastelloy N from the core shell and end pieces
showed darkened areas l/2 to 1 mil deep in re-
gions where the metal was in contact with the
core graphite, indicative of carburization there.

 

2}, H. DeVan and R. B. Evans ITI, ‘‘Corrosion Be-
havior of Reactor Materials in Fluoride Salt Mixtures,'’
pp. 557-79 in Conference on Corrosion of Reacfor
Materials, June 4—8, 1962, vol. 11, International Atomic
Energy Agency, Vienna, 1962.

182

Similar darkening along the core outlet tube
bottom could be either carburization or corrosion.

15.7 OXYGEN ANALYSIS

Oxygen in the salt may come from moisture (or
other oxygen compounds) absorbed by the salt,
either from the atmosphere, the graphite, or else-
where, or from the dissolution of metal oxides
previously formed. Three oxygen deteminations
on salt samples were made. The original solvent
salt contained 115 ppm. Solvent salt withdrawn
from the loop after 353 hr of in-pile circulation
contained 260 ppm. The increase is equivalent to
about 25 mg of oxygen (or 80 mg of chromium oxi-
dized to Cr2"). Fueled salt withdrawn from the
loop after retraction and freezing showed 241 ppm
of oxygen. These values are well below levels
expected to cause precipitation of zirconium or
uranium oxides. However, some or all of the
69-mg increase in chromium content of the salt
noted during the solvent-salt operation could have
been due to corrosion if the oxygen increase in
this period is attributed to moisture, all of which
reacted to dissolve chromium from the metal.

An analysis for the U3*/U*" ratio in a salt
sample was attempted, but a valid determmination
has not been reported.

15.8 CRACK IN THE CORE OUTLET PIPE

Following its removal from beam hole HN-1, the
loop was transferred to hot-cell facilities for
examination. Cutup of the loop is described in a
later section. After the containment vessel was
opened, no evidence of salt leakage from the loop
was seen by visual examination. The loop was
then pressurized to ~ 100 psig with helium, and
L.eak-Tec solution was applied to the extemal
loop surfaces. By this technique a gas leak was
observed in the core outlet pipe adjacent to its
point of attachment to the core body. Subse-
quently, the loop was sectioned for metallo-
graphic examination, and a crack through the wall
of the Hastelloy N outlet pipe (0.406 in. OD x
0.300 in. ID) was found. Figures 15.4 and 15.5
are photomicrographs of the crack, which ex-
tended almost completely around the circumference
of the pipe.

Analysis of the cause of failure in the core out-
let pipe indicates that this failure was probably
e

 

183

i R36583

 

 

 

Fig. 15.4. Inner Surface and Crack in Core Outlet Pipe of In-Pile Loop 2. Top side, near core. 250x.

caused by stresses resulting from differential
thermal expansion of the loop components.

Calculation of the piping stresses in the loop
has been made for two conditions: (1) for the
temperatuce profile around the loop at normal,
full-power in-pile operation, and (2) for the tem-
perature piofile observed during a reactor setback
(change from full power to zero in ™ 11/2 min).

For both conditions (1) and (2) the piping stress
analysis indicates that the maximum stress from
thermal expansion occurs in the core outlet pipe
where the failure occurred. For the nomal operat-
ing condition the bending movement produces a
stress of ~ 10,000 psi in the pipe wall (tension on
the top and compression on the hottom). For the
temperature distribution encountered during a
reactor setback, the direction of the bending move-
ment is reversed, causing a stress of ~ 17,000 psi
in the pipe wall (compression on top and tension
on the bottom).

It appears that two factors could have caused
the failure in the core outlet pipe. First, the

section of pipe where failure occurred was at a
temperature of ~1350°F. Stress-rupture properties
of Hastelloy N at 1350°F (732°C) are below those
at 1200°F (650°C) used for design purposes, and
these properties are further reduced®'* by the
accumulated irradiation dose of ~5 x 10'? nvt.
Under these conditions (1350°F and 5 x 1019

nvt), it is estimated that stresses of about 10,000
psi could produce rupture within moderate times,
possibly of the order of days. Fuither, the duc-
tility of Hastelloy N is reduced such that steains
of 1 to 3% can result in fracture. Thus the thermal
stress of ~ 10,000 psi calculated to exist in the
outlet pipe at full power operation may have been
sufficient to cause failure. A second and more

 

3H. E. McCoy, Jr., and J. R. Weir, Jr., Materials de-
velopment for Molten-Salt Breeder Reactors, ORNL-
TM-1854 (June 16, 1967).

‘n. E. McCoy, Jr., and J. R. Weir, Jr., In- and K x-
Reactor Stress-Rupture Properties of Hastelloy N
Tubing, ORNL-TM-1906 (Septembar 1967).
 
  

  

184

| R36587

W e =

&

 

5

I 250x

 

—— 3044 NCRES
N ] = S o

Fig. 15.5. Inner Surface and Crack in Core Qutlet Pipe of In-Pile Loop 2. Bottom side, near core. 250x,

likely cause of failure is the rapid stress reversal
(+10,000 to —17,500 psi) calculated for the thermal
shock caused by the rapid loss of fission heat
during a reactor setback. Approximately half a
dozen such cycles were encountered during in-pile
operation. In particular, one such cycle occurred
on March 3 after a dose accumulation of ~4 x 10!?
nvt, and it was on March 11 that evidence of fis-
sion product leakage from the loop was first ob-
served. Thus the stress reversals resulting from
such cycles are very likely to have contributed to
failure of the loop, since the rupture occurred at
the point where the stress was a maximum, the
temperature was 1350°F, and a radiation dose suf-
ficient to affect the strength and ductility of
Hastelloy N had been accumulated.

It is evident that for future in-pile loops with
similar configuration and temperatures, a material
superior to the present Hastelloy N in high-
temperature strength under irradiation is required.

15.9 CUTUP OF LOOP AND PREPARATION
OF SAMPLES

Promptness in the examination of the loop was
essential. Major irradiation of the loop had
ceased on March 17, 1967, when the loop was
retracted following the identification of the fission
gas leak. Since such short-lived isotopes as
66-hr *°Mo and 78-hr 13 2Te were of interest, it
was necessary that all such samples should be
counted within less than about eight weeks after
this time.

On April 5, 1967, following the ORR shutdown
on April 4, the loop package was removed from the
beam hole and transferred to the segmenting
facility. The sample and addition system was
stored, and inlet (‘‘cold”’) and outlet {*‘hot’’) gas
tubing sections were obtained from the external
equipment chamber region and submitted for radio-
chemical analysis.

¥
The assembly was segmented into shield plug,
connector, and loop container regions. Sections
of the gas addition and gas sample tubing and
sections of the salt sample line from these re-
gions were obtained for radiochemical analysis.
The loop container region was opened; no evi-
dence of salt leakage from the loop was seen.
Several sections of the gas addition and sample
lines and of the salt sample line were taken. The
loop was then pressurized with dry argon, and
bubbles from a leak-detecting fluid indicated the
crack on the top of the core outlet tubing near the
core.

During these operations, the core region was
kept at 300°C in a fumace when not being worked
on to keep radiolytic fluorine from being generated
in residual salt, although the salt inventory was
only about 2 g. The loop was cut into three
segments — gas separation tank, cold-leg retuin
line, and core — and was transferred to the High-
Radiation-Level Examination L.aboratory for
further cutup and examination. There the Hastel-
loy N core body was removed, and the graphite
was cut into upper and lower sections with thin
sections removed at top, middle, and bottom for
metallographic examination.

Samples of loop metal were taken such that
surfaces representing all regions of the loop were
submitted for both radiochemical and metallo-
graphic examination. In total, some 16 samples
of loop metal, & of the salt sample line, 11 of the
“‘hot’’ gas sample tubes, and 9 of the ‘“‘cold’’ gas
addition tube were submitted for radiochemical
analysis. A dozen specimens of metal from the
loop, some of which contained parts of several
regions of interest, have been subjected to metal-
lographic examination. The results of these
analyses and examinations are described in sec-
tions which follow.

Small amounts of blackened salt were found in
the gas separation tank near the outlet, in the
core bottom flow channels, and in the first few
inches of salt sample line near the core. A drop-
let also clung to the upper thermocouple well in
the fuel channel. The total residual salt in the
loop did not appear to exceed the inventory value
of 2 ¢g.

The penetration profiles of the various fission
products in graphite were determined by collecting
concentric thin shavings of core graphite from
representative fuel tubes for radiochemical analy-
sis. For this purpose a graduated series of

185

broaches or cylindrical shaving tools were de-
signed by S. E. Dismuke. Fourteen broaches per-
mitted sampling of the core graphite fuel channels
(‘/4 in. ID) to a depth of 45 mils, ia steps nomi-
nally ranging from 0.5 mil for the first few mils in
depth up to 10 mils each for the final two cuts.

In some cases, two or three cuts were collected
in the same bottle in order to reduce the number
of samples to be analyzed.

A sample bottle was attached directly below the
hole being sampled, and the broach with shaved
graphite was pushed through the hole into the
bottle from which it was subsequently retrieved
after brushing into the bottle any adhering graphite
particles. The bottle was closed, a new bottle
clipped into place, and the next larger broach used.
For each bottle, all the graphite sample was
weighed and dissolved for radiochemical analysis.

Total recovery from given holes ranged from 94
to 111% of values calculated from the broach di-
ameter and graphite density. The higher values
were almost entirely due to high initial cuts, indi-
cating fuel channels narrower than the nominal
0.250-in.-diam, irregular original holes, or in some
cases, some salt adhering to the surfaces. Since
penetration depth should be measured from the
original surface, actual depths were calculated
from the cumulated weight of material actually
removed. Forward, next-to-forward, next-to-rear,
and rear fuel tubes, in top and bottom sections,
were sampled in this way; a total of 76 such
samples were submitted for analysis.

Graphite was also shaved from the outer surface
of the core cylinder in four samples to a depth of
19 mils. In addition, eight 1/8— by 3/4-in. core
drillings were taken of the interior central part of
the graphite.

1510 METALLGGRAPHIC EXAMINATION

Samples of loop metal taken from the core out-
let, gas separation tank inlet and onutlet ends,
cold leg, and other regions were subjected to
metallographic examination. This examination
defined the nature of the break in the core outlet
line and gave evidence of some carburization
and corrosion of loop metal surfaces.

Bottom and top inner surfaces of the core outlet
tube near the core are shown in Figs. 15.4 and
15.5. These photographs give a metallographic
view of the break in this tube. The break appears
 

186

[ T&

 

o

0.014 INCHES
lo 250x

1o

 

 

Fig. 15.6. Inner Surfuce of Cold-Leg Tubing from In-Pile Loop 2 (Hastelloy N). 250x.

to have been intergranular and without any indi-
cation of ductility. Such behavior at temperatures
in excess of 650°C and at a thermal neutron dose
of 5 x 10'? neutrons/cm ? is consistent with recent
ORNL studies?® of the effect of irradiation on
elevated-temperature properties of Hastelloy N, as
discussed in Sect. 15.8 in connection with the
break in the core outlet pipe.

Some evidence of attack on the inner surface of
the core outlet may be noted in Figs. 15.4 and
15.5. The upper inner surface shows evidence
of corrosion, even though part of the 1-mil, more
finely grained layer had been removed by reaming
prior to the assembly of the loop. The lower
surface does not show corrosive pitting but does
have a datkened bank almost 1 mil deep that
could be carburization.

The bottom surface (not shown) at the core end
of the gas separation tank showed no evidence
of either corrosion or carburzation.

The cold leg of the loop showed substantial
corrosive attack — probably largely intergranular

in the 1-mil, more finely grained inner layer as
shown in Fig. 15.6. An unexposed piece of the
same tubing is shown for comparison in Fig. 15.7.
This tubing was also used for fabrication of the
core outlet pipe.

The depth of the attack on the cold-leg pipe is
of the same magnitude as would be anticipated
from chtomium and oxygen analyses of the salt
reported above. The tubing used to fabricate the
cold leg and core outlet appears to be more sensi-
tive to corrosion than materials used in other
parts of the system.

Carburization to a depth of about 1 mil appears
to have occurred in the inner surface of the core
shell wall in contact with the graphite core, as
shown in Fig. 15.8. Similar carburization was
also noted on core top and bottom pieces (not
shown). Hardness tests were taken at various
depths below the surface of the metal, the nearest
about 1 mil. The test nearest the surface showed
definitely greater hardness, as would be expected
from carburization.
-

187

 

i . L. . g . i
o o LN A A Ty

o ’@—*’fi : e - SRl %
« £5F "

j R—39708

o

 

i

e 260x

[WO

 

 

 

Fig. 15.7. Inner Surface of Unexposed Hastelloy N Tubing Used in Cold Leg and, After Slight Reaming, in Core

Outlet of In-Pile Loop 2. 250x.

15.11 ISOTOPE ACTIVITY CALCULATION
FROM FLUX AND INVENTORY HISTORY

It was useful to estimate how much of a given
isotope (either fission product or activation prod-
uct in flux monitors, etc.) was to be expected in
the system at a patticular time. This may be done
by detailed application of standard equations® to
the individual irradiation and inventory periods with
appropriate adjustment for decay tc a standard
reference time (reactor shutdown, 4-4-67, 0800).
Counting data on the various samples were also
referred to this time. In the in-pile period of the
experiment, 38 changes in ORR power and 122
changes in experiment position altered the neutron
flux, 3 salt additions altered the fuel composition,

and 18 withdrawals, including 9 samples, were made.

 

5_]. M. West, pp. 7—-14, 15 in Nuclear FEngineering
Handbook, ed. by H. Etherington, McGraw-Hill, New
York, 1958.

The equations described below, along with the
detailed flux and inventory history, were pro-
grammed for computer calculation and estimates
of activity of the respective isotopes made as
described below. The equations will be discussed
in terms of fission products but are readily adapted
to activation piroducts. The activities (A-)\] and
B.)\ ) of isotopes that are the first and second
significant members of a decay chain produced
by a given period, ¢, of steady irradiation fol-
lowed by a decay period, t, are givens by
AN =Y-Fel—e e h‘["),

1

~At]
)\2)\1 1'—8 1°i __Alt
B-A -Y.F. (e d)

 

&
S
f
>
>

MAQQ
(1= 7 (e“"z‘d>
A
 

188

[ro =

[~

0.014 INCHES
e 250x%

 

Fig. 15.8. Inner Surface of Core Shell in Contact with Graphite Core of In-Pile Loap 2. 250x.

These equations pemnit the calculation of activity
if flux is given, or of flux if activity has been
measured.

The ¥+ F term is the chain production rate at a
given flux or power density for the quantity of
material under consideration. It was useful to
treat the Y. F term in the case of fissioning asa
product of fission yield ¥, a standard fission
rate F© per gram of uranium (at a given flux), a
uranium inventory for the interval j (from salt
inventory actually under irradiation, Wi times
uranium concentration U,), and the irradiation
intensity factors, p. and q., relative to full re-
actor power and fully inserted position. Thus for
any set of intervals, for example,

Z(AA )]
— e LU Pty

—A. £, —A -t
c(1—e ' e P E))

A similar expression can be written for the
daughter isotope (such as the 35-day ?°Nb daugh-
ter of 65-day °3Zr).

The tem on the right may be computed for any
given isotope from a knowledge of decay rate,
irradiation, and inventory history. The term on
the left represents the count to be expected,
divided by the saturation value from 1 g of ura-
nium at full power.

In practice, the total loop inventory was cor-
rected for the amount of salt not under irradiation,
at the time, in sample lines and the purge tank.
The activity withdrawn with each removal and that
remaining in the loop were calculated. The total
activity (at reference time) of a given isotope
produced during the experiment was thus estimated.

As mentioned in an earlier section, it appears
reasonable to use the activities of certain fission
products in the salt samples, in particular !*’Cs,
144Ce, and ?°Zr, as internal standards to esti-
mate the flux. A mean flux to the salt of 0.88 x
10'% was thus estimated from activities measured
Table 15.4. Comparison of Fission Product Activity Produced with Activity Found in Various Loop Regions

All activities in units of 101° dis/min, referred to ORR shutdown, 4-4-67, 0800

 

147

 

Isotope QQMO 103Ru lOGRu 132Te 129Te QSNba gszr 131I 140Ba 8981‘ 91Y 137CS 141CE 144Ce Nd
Half-life 66h 39.7d 366643 77.7h 334 35d 65 d 8.054 12.8d 50.4d 583d 30y 32.8d4 284d 11.1d
Fission vield, % 6.1 3.0 0.38 4.3 0.34 6.2 6.2 2.93 6.35 1.79 5.8 6.0 6.0 5.6 2.5
Total activity formed” 940 8100 1990 1140 970 7400 12,800 5000 16,500 11,500 12,900 108 17,300 3500 61090
Found in salt samples® 2 6 d 9 d 1530 89840 1520 4800 7550 15,600 94 12,100 3830 4400
20 1290 13,3060 1570 S0O00 9370 7900 103 13,500 4330 23960

Found in residual salt 1110 180 220 902 =
Found in graphite 385 d o 360 57+ 1001 126 24 410 753 d d 277 68 146 ©
Found on loop metal 220 d d 160 55+ 1300 77 30 220 122 d d 77 16 42
Found in salt sample line 93 o d 82 5+ 246 104 57 i50 84 d o il4 23 115
Found in hot-gas sample line 0.2 d d 0.4 <01 <1 <1 5 <1 15 d d <1 0.01 <1
Found in cold-gas inlet line <0.1 4d d <1 <ol 1 <1, 1 <1 1 d d <1 <0.01 <1

®#Daughter of Bz,

EJAs;sn:tming a mean flux to salt of 0.88 x 1013, based on average of values from 952:, 137(’35, and 144Ce in final salt samples.

“Estimated for total salt based on each of two final samples,

d : . , L. .
Determinations not sufficiently complete to be given here.
on final salt samples. This value was used in
calculating total activities of the various iso-
topes produced in the experiment.

15.12 iSOTOPE ACTIVITY BALANCE

The calculations above provide an estimate of
the amount of isotope to be accounted for. We can
estimate how much was actually found in a kind of
‘“isotope activity balance’ by accounting for all
regions of the loop in terms of the measured
activity of samples.

For the various samples of metal, graphite, and
salt obtained from the loop, activity determina-
tions for the 15 isotopes shown in Table 15.4 were
requested. The concentration of ***U was also
determined. Because of the short half-life of such
isotopes as “’Mo and '3?Te, these were mun as
promptly as possible, and others were detemmined
later. Results for some isotopes are not yet
complete. The available data will be examined
in tems of activity balances for the respective
isotopes and the penetration profiles of these
isotopes in graphite.

The ratio of the area of each loop (or fuel
channel) region to the sample representing the
region was determined so that the total loop area
was accounted for in terms of samples. The
cumulative activity of all shavings from a graphite
channel section was used as the sample from that
channel. The sample activities multiplied by the
proper ratios have been totaled for each isotope
under the categories of graphite, loop metal, salt
These values,
plus values for salt based on the final sample
activities, are shown in Table 15.4. Estimated
total activities from the ¢alculations based on

sample lines, gas lines, and salt.

irradiation and inventory history are also shown.

From Tahle 15.4 it may be seen that over half
(and generally less than all) the expected activity
appeared to be accounted for in the cases of
990, 132Te, 95Nb, 95Zr, 89S, 137Cs, 1*1Ce,
144Ce, and '47Nd. A substantial proportion,
although less than half, was accounted for in the
cases of '*%Ba and '3!'I. Inasmuch as iodine
readily volatilizes from all samples, without
doubt especially from the powdered graphite, it
is to be expected that iodine determinations shall
be low. Determinations are not yet complete in
the cases of 1%3Ru, '°%Ru, '2%Te, °'Y, and
1370

190

Molybdenum, tellurium, and ruthenium are almost
entirely departed from the salt, along with sub-
stantial proportions of 2°Sr, 9°Nb, '*°Ba, and
most probably '*'I. Except for 89Sr and possibly
140R4, which favor graphite, these elements show
no strong preference for graphite or metal but
seem to deposit on whatever surface is available.
The alkali-metal and rare-earth isotopes, includ-
ing 'y, '37Cs, '*!Ce, '**Ce, and '*7Nd, and
also ?®Zr, remain almost completely in the salt,
the amounts found in graphite generally being
ascribed to salt contained in the samples, as
discussed later.

15.13 URANIUM-235 OBSERVYED IN
GRAPHITE SAMPLES

Uranium-235 on the various graphite and metal
samples was also detemmined. An activation
technique was used in which delayed neutrons
were counted; it is sensitive to less than 1 pg
of 2?3U. The deteminations served the dual
purpose of measuring small quantities of salt
which could have adhered to surface samples and
of determining the penetration of uranium in one
form or another into the graphite.
In seven of the eight graphite channels from
which samples were taken, the quantity of 25U
ranged from 1.6 to 2.8 mg (per 8.3 to 9.5 cm %) with
over half being found within the first mil and over
80% generally within the {irst 3 mils. However,
some uranium was detected even in the 35- to
45-mil cuts. The first sample from the remaining
channel weighed 290 mg and contained 18.9 mg of .
233y, equivalent to 190 mg of fuel salt. The
samples from deeper cuts contained uranium at _
levels only moderately higher than those from -
other channels. Thus, it appears that this sample
piobably contained a small piece of fuel salt
which had remained on the surface.
X-ray diffraction patterns obtained from the
surface of a specimen of graphite cut from an
exit fuel channel surface showed patterns of
Li 28@5‘4 and Li 2ZrF6,
oxides or uranium compounds. Such a pattern is
characteristic of normally frozen fuel salt, so
these observations indicated that fuel salt had
adhered to the graphite surface even though not
directly visible under ordinary hot-cell viewing
conditions.

with no indication of
1w

191

Table 15.5. Salt Constituents in Graphite at Given Depths

 

Method of Petermination

 

Masg Spectrograph (MS-7)

 

Li Be

Zr

Activation, U
u u

 

(g /me)
{mole %) (mole %) (mole %) {mole %) (pg/meg)
L.cop inventory: 27.8 65.3 4.8 2.0
Graphite — top section next-to-front channel
0.2 to 0.4 mil 22 71 5.3 1.8 39 419
0.4 to 2.05 mils 21 73 3.8 2.2 2.6 3.6
2.05 to 4.4 mils 15 80 3.5 1.7 2.2 0.91
4.4 to 9.8 mils 16 80 4.5 a (<0.34) 0.32

 

“Not determined.

The MS-7 spark mass spectrograph was used to
examine solutions of graphite shaved from a
typical fuel tube for salt constituents. Total
amounts used were at the microgram level. Re-
sults are shown in Table 15.5.

The salt ingredients Li, Be, Zr, and U are in
the proper molar ratio (with the proper total of
"Li and "Be, but some drift in their ratic). Fur-
themore, the absolute quantity of uranium agrees
reasonably with the values from the activation
analysis by delayed neutron counting.

Congequently, the uranium entering the graphite
did so in the form of salt. The difficult question
is whether it pemeated the pores or adhered to
the surface and filled the cracks.

Three-phase contact of graphite, molten LiF-
BeF, or MSRE fuel salt, and slightly moist gas at
elevated temperatures has been shown by Kreyger,
Kirslis, and Blankenship® to result in the forma-
tion of adherent oxide films on the graphite, pre-
sumably BeQ, with wetting of this oxide (and
thereby the graphite) by molten salt. But similar
adherence was reported not to occur under the
zalt where the graphite and molten salt were in
contact before the moisture entered. Our oxygen
analyses of the fuel, reporied above, do not indi-
cate any great uptake of moisture, and such as ~
did enter appears to have been involved in cor-

 

op, J. Krevger, S. S. Kirslis, and F. F. Blankenship,
MSR Program Semiann. Progr. Rept. Jan. 31, 1964,
ORNL-3620, pp- 132--37,

rosion processes and is not necessarily to be
associated with adherence of salt to graphite.

Prior to admission of any salt to the loop, the
graphite was heated at 600°C for 20 hr under
vacuum, whereas evacuation for several minutes
at 400°C has been shown to be sufficient to
remove nomal moisture from graphite. Thus,
internal moisture in the graphite is not believed
to account for adherence of the salt to graphite.

Examinations of autoradiographs indicate that
several fuel channels had one or two cracks ex-
tending from them and that appreciable radio-
activity was located along such cracks. Uranium-
235 was detected by alpha radiography in cracks
as well as on the hole surface. Although some or
all might be a polishing artifact (or from hot-cell
contamination), this does not dispute, and possi-
bly supports, the idea that salt entered cracks.
Thus, we think that the presence of the uranium
in the graphite can be associated with the entry
of salt into cracks and the substantial adherence
of thin films of salt to irregularities in graphite
surfaces.

The adherence to graphite by the salt, which
we will not further expound here, is believed to
have resulted in only coverage of surface (in-
cluding cracks), with no penetration of graphite
potes being indicated. Recognition of this un-
expected entry of uranium-containing salt into
cracks did assist in understanding the behavior
of certain fission products, as discussed below.
15.14 PENETRATION OF FiISSICN PRODUCTS
INTO GRAPHITE AND DEPOSITION
ONTO SURFACES

The average concentrations of individual fission
products in the graphite at different depths from
the surfaces were detemined by radiochemical
analysis of the samples shaved concentrically
from the fuel passages. Results for the individual
elements will be discussed below. It may be said
here that they are generally consistent with de-
terminations on the first and second set of MSRE
surveillance specimens after taking into consider-
ation the adherence of small amounts of uranium-
containing salt to the graphite surfaces in the
loop and its entry into cracks, as discussed in
Sect. 15.13. By also considering the deposition
onto loop metal surfaces, some tentative con-
clusions are drawn below.

The results were consistent from channel to
channel for a given isotope except for variations

192

attributed to the presence of salt and irregulari-
ties in the original surface. Consideration of
such variation is not complete but is not expected
to alter the conclusions below. At this time, we
will therefore consider data obtained from a single
representative channel (bottom graphite section,
front channel). The samples from this channel
contained a total of 2.84 mg of 235U, and the
original channel surface was 8.3 cm®. The activi-
ties of the various fission products that would be
generated in this amount of uranium during the
total irradiation period were calculated and are
compared in Table 15.6 with the activities found
in the channel.

The isotopes shown in Table 15.6 are grouped
into three sections which are related to different
kinds of behavior. The first section includes the
elements Zr, Ce, Nd, and Cs. The isotopes found
in the samples appear to be as expected from the
amount of salt present and do not represent any

interaction with the graphite. The concentration-

Table 15.6. Activities of Isotopes in Grophite Compared with the Activities
That Would Be Generated in the 22U in the Sample®

 

Total Observed

 

Isotope Half-Life Chain Yield Calculated Activity Activity Ratio,
(%) (dis/min) (dis /min) Obsecrved/Calculated
x 109 % 1019
Sz, 65 d 6.2 2.5 2.96 1.2
143ce 284 d 5.6 0.73 0.88 1.2
141 ce 32.8 d 6.0 3.21 4.81 1.5
147Nd 11.1d 2.6 1.04 1.70 1.6
13 cs 30y 6.0 0.0225 b b
131y 8.05 d 2.93 0.86 1.54+ 1.8+
1400, 12.8 d 6.35 2.83 15.4 5.5
895, 50.4 d 4.79 2.22 28.5 13
¥3Nb 35 d° 6.2 1.64 23.7 14
1296 33 d 0.34 0.18 2.04+ 114
1321q 3.25d 4.3 0.36 14.9 41
Mo 2.75 d 6.1 0.44 22.1 50
103p, 39.7 d 3.0 1.53 b b
106p, 366.6 d 0.38 0.039 b b

 

 

“Cumulative values of samples from bottom front fuel passage containing a total of 2.84 mg of 235U; all counts
referred to reactor shutdown (4-4-67, 0800). Calculation based on average loop flux at full power of 0.88 x 1013

neutrons cm™ 4 sec™ L,
PDeterminations not completed.

CDaughter of 65-day 93zr.
18

depth profile of these isotopes (plotted semi-
logarithmically) parallels that of the uranium in
the graphite. Fission product zirconium, of
course, had to remain mostly in the salt because
it necessarily mingled with the 5 mole % of Zr in
the salt itself.

The next group of elements (I, Ba, Sr, and Nb)
shows an appreciably greater activity in the
samples than can be accounted for by the amount
of uranium present. These elements (except
possibly ?°Nb) also show concentration-depth
profiles that do not fall off nearly as rapidly as
does the uranium concentration. Such behavior
has heen described by Kedl” and others (based on
MSRE surveillance specimen data obtained by
Kirslis) as being due to gaseous transport of the
isotope or a short-lived precursor. Niobium-95
could exist as a volatile pentafluoride, 8°Sr has a
3-min krypton precursor, and '*°Ba has a 16-sec
xenon precutsor. lodine-131 is included in this
group, since iodine volatilizes readily from sam-
ples, and the true values should doubtless be
higher than those reported.

The comparatively flat profiles (except for the
first mil or two) imply rapid transport through pores
in the graphite without very strong tendency for
adberence to graphite surfaces. Niobium-95, the
profile of which is not as flat as the others,
appeared to be adsorbed appreciably on the graph-
ite surface, though not as strongly as elements of
the group to follow.

The third group contains elements Te, Mo, and
Ru. (The inclusion of ruthenium is a guess, at
present, based on MSRE surveillance specimen
studies by Kirslis, since our data are incomplete
for the ruthenium isotopes.) This group is charac-
terized by the highest concentrations in the sam-
ples, relative to the uranium present. The con-
centration-depth profiles are even steeper than
for uranium. It appears evident that these ele-
ments are strongly adsorbed by graphite surfaces
and are able to penetrate cracks readily, either by
rapid diffusion in the salt in the crack or by
surface diffusion. They do not tend to penetrate
bulk graphite pores to a comparable degree. The
highest concentration of *’Mo and '*?Te per unit
of surface, either graphite or metal, was reported
for the Hastelloy N thermowell in the front core

 

7R. J. Kedl, A Model for Compuiting the Migrations of

Very Short-Lived Noble Gases into MSRE Graphite,
ORNL-TM-1810 (July 1967).

193

fuel channel. This implies that the elements in
this group tend to deposit on the first surface en-
countered, whether metal or graphite.

15.15 GAMMA IRRADIATION OF FUEL SALT
IN THE SOLID PHASE

The fuel salt in molten-salt in-pile loop 2 was
kept frozen at temperatures generally above 300°C
for a number of days before the salt was re-
moved from the loop. At temperatures below
100°C, fission product or gamma radiolysis of
fluoride salt is known to result in the generation
of fluorine gas to appreciable pressures. Fluorine
could oxidize various species (U, Mo, Ru, Nb, etc.)
to produce volatile fluorides which could then be
transported to other parts of the system. Al-
though a temperature of 300°C was expected to
be more than adequate to suppress such radiolysis
completely, direct data were not available but
appeared readily obtainable. Since the phenome-
non is also of general interest in the MSR program,
the experiment described below was undertaken.

Solid MSR fuel salt (LiF-Bel” -ZrF -UF ,
about 65-28-5-2 mole %) from the stock used to
supply fuel for in-pile loop 2 was subjected to
very-high-intensity gamma irradiation in a spent
HFIR fuel element at a temperature of 320°C to
determine possible radiation effects on the salt
and its compatibilities with graphite and Hastel-
loy N.

For the irradiation experiment, HFIR fuel ele-
ment 9 was placed in a storage rack in the HFIR
pool, and the experiment assembly was placed in
the center of the fuel element as shown in Fig.
15.9. The capsule-type irradiation assembly
consisted of a Hastelloy N capsule, 0.93 in.

OD x 0.78 in. ID x 3.5 in. long, containing two
CGB graphite test specimens, 3.0 x .87 x 0.125
in., placed back to back in the capsule. Twenty-
five grams of fuel salt was added to the capsule,
melted, and then allowed to solidify. A pressure
transducer was connected to the gas space above
the salt, and the capsule assembly was welded
shut. A heater assembly using Nichrome V heater
wire surrounded the capsule, and thermocouples
were located in the capsule wall to monitor tem-
peratures. The experiment assembly was then
placed in an aluminum container to isolate it from
the pool water, as shown in Fig. 15.10.
 

 

194

ORNL-DWG 67-11834

Y 848 ft Gin.

 

 

HFIR EXPERIMENT RCOM
MEIR STORACE POOL N
=1 [DIDEE0)
N e
— — | BRESS] |
OO e
) 3 )
X 4
R VAR 837 ft 8in,
Y N T NSNS
O
N
N
B B 3
=
N
A
W
HFIR FUEL ELEMENT - ~EXPERIMENT
ASSEMBLY
NN
3
G N
N
__________________________________ 828 Al

 

 

 

Fig. 15.9. Equipment Layout for Gamma lrradiation Experiment in an HFIR Fuel Element.

The gamma flux of the spent fuel element used
for the irradiation experiment was measured by
the reduction of a 0.02 M Ce(SO4)2 solution irradi-
ated in the center of the fuel element. Such meas-
urements indicated that the gamma intensity was
8.06 x 107 r/hr before the start of irradiation (6
days decay time) and 2.23 x 107 t/hr after termi-
nation of irradiation (30 days decay time).

The capsule assembly was placed in the center
of HFIR fuel element 9 on July 20, 1967. This
element had been removed from the reactor on
July 7, 1967, after 23 days of operation at 100 Mw.
The salt temperature reached 320°C from gamma
heat alone (calculated to be 0.2 w/g). The as-
sembly was removed on August 14, 1967, after

600 hr (25 days) of irradiation, accumulating an
estimated absorbed dose of 1.5 x 102* ev per g of
salt based on the irradiation intensity measure-
ments described above. The temperature of the
fuel salt was held constant at 320°C throughout
the irradiation period by adding furnace heat as
the fuel element gamma flux decayed. No signifi-
cant pressute change was noted,; a decrease from
~19 psia to ™~ 18 psia appeared to be associated
with the change from internal gamma heating to
extemal electrical heating. Analysis of a sample
of the argon cover gas obtained during postirradi-
ation examination showed no evidence of radiol-
ysis products.
1%

195

 

S VLR
I
L f‘v . 4, 1.)'24

)T el T

ORNL-DWG B67-11B32

o — PRESSUSE CEIL

 

 
 

 

 
  
  

——SEAL WELD

 

 

     

 

 

THERMOCOUPL ES ~r-mewneee. G|

FLUORIDE SALT— o X

 

GRAPHITE PLATES *fi [

 

 

 

 

 

  

/.f——ALUMlNUM CONTAINER

o ——SPACER

e NICKEL PLATED COPPER FURNACE

e HASTELLOY N CAPSULE

Fig. 15.10. Autoclave Assembly for Gamma lrradiation of Fluoride Salt in an HFIR Fuel Element.

The Hastelloy N capsule containing salt and
graphite test specimens was cut up for examina-
tion in an inert-atmosphere dry box. Preliminary
examination did not reveal any gross effects.
Thin salt films adhered to the Hastelloy N con-
tainer with no evidence of noteworthy interaction.
Parts of the salt were darkened to a dull grey-
black, an expected result of irradiation of Li BeF,
and LiZrF  crystals. Zirconium-95 and 1408,
activity in the salt indicated 3 x 1019 total fis-
sions per g of salt (a relatively trivial figure),
presumably from beryllium photoneutrons (the
uranium was 93% enriched). Graphite from below
the salt surface showed slight stains or films.

X-ray diffraction examination of the surface
showed only graphite lines. Samples of graphite
from gas and solid phase regions were analyzed,
showing 1 ug of ?3%U per em ? in the gas phase
and 15 to 35 pg/cm? in the graphite contacting
solid salt. This appeared to be due to traces of
salt adhering to the graphite. The guantity is
much lower than was noted on graphite from fuel
channels of in-pile loop 2, where 350 pg of 2*°U
per cm ? was characteristic.

We find no evidence of radiolytic generation of
fluorine or transport of uranium or other substances
due to irradiation at 320°C.
Part 5. Materials Development

H. E. McCoy

Our materials program has been conceined both
with the operation of the MSRE and with the de-
velopment of materials for an advanced MSBR.

The primary role of the materials program in the
operation of the MSRE has been a surveillance
program in which we follow the property changes of
the MSRE graphite and Hastelloy N as the reactor
operates. We have surveillance facilities in the re-
actor core and on the outside of the core tank.

The successful operation of the MSRE has
strengthened our confidence in the molten-salt re-

J. R. Weir

actor concept, and we have initiated a materials
program in support of an MSBR. In this reactor the
graphite will be a primary structural material, since
it separates the breeder and the fuel salts,
Hastelloy N will be the metallic structural material,
and it will be necessary to develop a joint between
graphite and Hastelloy N. Thus, most of our re-
search effort is concerned with the structural ma-
terials, graphite and Hastelloy N, and the joint be-
tween them.

16. MSRE Surveillance Program

16.1 GENERAL DESCRIPTION OF THE
SURVEILLANCE FACILITY AND
OBSERVATIONS ON SAMPLES REMOVED

W. H. Cook

The effects of the operational environments of the
MSRE on its unclad moderator of grade CGB graph-
ite and its structural alloy of Hastelloy N continue

to be monitored periodically with surveillance speci-

mens of these materials.
the general sampling schedule, and the examination
program have been described previously.!

The second group of reactor core specimens,
stringer RS2, has been removed from the core after
accumulating peak fast- (£ > 0.18 Mev) and thermal-
neutron doses of approximately 1 x 102! and 4 x
1020 neutrons/cm? respectively. These specimens
were subjected to a temperature of 1190 +18°F for
5500 hr. The first group of vessel specimens,

 

sk Program Semiann. Progr. Rept. Aug. 31, 1265,
ORNIL.-3872, pp. 87-92.

The surveillance assembly,

stringer X1, was also removed. The peak thermal-
neutron dose on these is approximately 2.5 x 101°
neutrons/cm?.
these specimens was 900 to 1200°F, and the time
at temperature was approximately 11,000 hr. Initial
results of examination of these groups have not re-
vealed any unexpected or serious effects. These
two groups have been replaced with new sets of
specimens which are being exposed in the MSRE
environments together with other groups that were
not sampled at this time.

The estimated temperature range on

To review briefly, the reactor core surveillance
specimens consist of both graphite and Hastelloy N
mounted approximately 3 in. away from, and parallel
with, the axial center line of the moderator core.
The reactor vessel specimens are Hastelloy N
mounted approximately 41/2 in. outside the reactor
vessel, The reactor core specimens are exposed to
the molten fluoride fuel, and the reactor vessel
specimens are exposed to the nitrogen—2 to 5 vol %
oxygen atmosphere of the reactor containment cell.
The result is that the current and future irradiation

196
Table 16.1. Status of and Future Plans for the MSRE Surveillonce Program®

 

Stringer Pulled

Peak Neutron Dose Stringer Inserted

 

 

 

 

 

Total
Sampling Reacior Approximate Core Vesisel Stringer Vesself Cote Vessel
No. Power Date Designation” Graphite® Hasteiloy N Designation Hastelloy N Fast® Thermal Tremal Designation Graphite Hastelloy N Designation Hastelioy N
(Mwhr) - Heat No.¢ Heat No. Heat No. Heat No.
1 7,820 7-28-68 RS1, RR1, CGB 5081 None None 3.2x 107" 1.3x10%" $.5x 108 RS2 CGB 215458 None None
RL1 5085 21554%
RLZ CGB 5065
5085
RR2 CGB 5065
5085
2 32,450 5-5-67 RS2 CGB 21545 9.9 x 1028 4.1 x 1020 RS3 CGB 67-302% X4 67-502
21554 AXF-5QBG £7-504 67-504
X1 5065 2.7 % 101?27 x 10!° CGB-L!
5085 By
3 50,000  January 1968 RS3 CGR 67-502 7.1 x 1020 2.910%0  4.1x101° RS54 CGB Expti heat 5
AXF-5QBG 67-504 AXFS3QBG Exptl heat 6
RR2 caB 5065 1.7 x 10%1 7.0 1020 RR3 CGB Expti heat 7
5085 Expt! graphite  Expti heat 8
4 70,000 July 19687 RS54 CGB Exptl heat 3 8.1x 102% 3.3 1020 5.8x10'° RS5 CGB Exptl heat 9
AXF-5QBG Exptl heat & Expt! graphite  Exptl heat 10
RR3 CGBE Exptl heat 7 8.1x 1G62% 3.3 % 102¢ RR4 CGB Exptl heat 11
Exptl graphite  Expti heat 8 Exptl graphite  Expil heat 12
5 90,000  March 1969 RL2 cGB 5065 3.3 x 1071 14,1020 7.4x10'°
S085
RS5 CGB Exptl heat 9 8.1x 1020 3.3 1020
Exptl graphite  Exptl heat 10
RR4 CGB Exptl heat 11 8.1x 1020 3.3 1020
Exptl graphite  Exptl heat 12
X2, X3 5065 7.4 % 10t°
5085
¥4 57-502 4.8 % 101°
67-504

 

“Planned and compiled by W. H. Cock, H, £, McCoy, A. Taboada, and others,

YAll these reactor core specimens have control specimens exposed 1o a static fuel szit under MSRE
conditions excep! that there is no neutron radiation.

“Graphite grade designations: Grade CGB is the MSRE moderator praphite which is anisotropic. Grade
AXF-50QBG is an isotropic graphite. Expt! graphite refers to otber experimental grades of isotropic graphite,

“Hastelloy N heat no. refers fo the heat number of 1 standard Hastelloy N composition unless noted other-
wise. Lxptl heat refers to modifications of the basic lastelloy N thai are to be determined later.

¥Based on a calonlated fiux supplied by J. R, Engel for nentrons with E > 0.18 Mev.

£ . sy ¢
Approximate values for both the reactor vessel wa

fExperimental heat 1 which contains an addition of 0.52 wt % Ti.
hExperimental heat 2 which contains an addition of 0,42 wt % 21.

‘Impregnated with graphitized pitch.
JGrade AXF-30BG grughite brazed to Mo with 60 Pd—35 Ni~3 Cr (wt %) brazing alley.

and the reactor vesse! specimens.

*Experimental heat 3 which contains additions of (.5 Ti and 2 W wt %.
’Experimenial heat 4 which contains an addition of 0.5 wt % Hf.
™o be concurrent with salt change to one using 2330,

L6l
198

 

 

(o) S FLUX

MONITORS

Fig. 16.1.

 

ORNL-DWG &67-10918R

STRAP Lo Cod
BAND i

 

{R52) REMOVED

(RS3) INSTALLED

MSRE Reactor Core Surveillance Specimens. (a) Assembled detailed plan view ond {b) Unassembled.

The RI.2 and RR2 stringers were not disassembled for the removal of R52 and the installation of RS3 stringers.

effects on the Hastelloy N can be monitored by the
reactor vessel and reactor core specimens respec-
tively. The radiation dose on the structural
Hastelloy N, because of its location, is less than
that on the Hastelloy N reactor core specimens (see
Table 16.1). This is a desirable condition of the
surveillance specimens of the MSRE, because
Hastelloy N is more strongly affected by the radia-
tion than is graphite under the MSRE conditions.
The previous reactor core specimen assembly for
the first sampling had nonnuclear, mechanical dam-
age that broke some of the graphite and bent some
of the Hastelloy N specimens.? This assembly was
replaced by a slightly modified one, from which the
current sampling was made.? The complete assem-
bly is withdrawn from the reactor in order to remove
any one of the three stringers. This time, the as-
sembly appeared to be in the same condition as it
was prior to its exposure, except that the surfaces
of the Hastelloy N specimens had been dulled. One-
third of the assembly, stringer RS2, was removed,
and a new stringer, RS3, was joined to the other
two stringers (Fig. 16.1). These were returned to

 

‘w. H. Cook, MSR Program Semiann. Progr. Rept. Aug.
31, 1966, ORNL-4037, pp. 97--103,

the reactor in a new containment basket, as is our
standard practice. This completed our second
sampling operation, shown as sampling 2 in Table
16.1.

The controls for the reactor core specimens, which
are exposed to fuel salt under MSRE conditions ex-
cept that the salt is static and there is no radia-
tion, were also removed from the controlled test rig.
The typical appearance of the reactor core speci-
mens and their controls is shown in Fig. 16.2.
Visually, the graphite, with the exception of a few
salt droplets, appeared unchanged; note that
Hastelloy N tensile specimens are reflected in the
bright surfaces of the graphite (Fig. 16.2¢) .

The graphite specimens were pinned at their
tongue-and-groove joints within Hastelloy N straps
with 0.030-in.-diam Hastelloy N wire. The move-
ment of the graphite column of specimens relative
to that of the Hastelloy N tensile specimen rods
was sufficiently difficult to deform the pins in the
reactor core specimens. These pins had to be
drilled out during disassembly. The pins of the
control specimens were not deformed, but the con-
trol assembly is less complex, since space limita-
tions are not as restricted for it as for those in the
reactor. Each control stringer is separate from the
199

PHOTO 89482A

 

Fig. 16.2. MSRE Reactor Core Specimens of Grade CGB Graphite and Hastelloy N after Run 11. (a) Controls,
stringer CS2, exposed to salt under MSRE conditions except the salt was static and there was no radiation. (b) and

(c) Stringer RS2, exposed to the fuel in the reactor core for 27,630 Mwhr operation of the MSRE,
200

others, but all three stringers in the reactor are
bound together. The new stringers for the reactor
core specimens and their controls are all pinned at
the joints with 0.040-in.-diam Hastelloy N in order
to eliminate the pin deformation problem.

Studies of the deposition of fission products on
the graphite and the Hastelloy N are being con-
ducted by the Reactor Chemistry Division. For
their work, they used the graphite specimens from
the bottom, middle, and top of the reactor core
specimen stringer, plus matching controls and
selected Hastelloy N samples from the stringer and
basket. The results of their examinations are re-
ported in Chap. 9.

The remaining graphite samples are being meas-
ured for dimensional changes. Tests of the types
outlined in ref. 3 will be made when the dimen-
sional measurements are completed.

The results of the examination of the Hastelloy N
tensile specimens from the core position and from
outside the reactor vessel are reported in detail in
Sect. 16.2.

We have begun to expose samples in the MSRE
core that are of interest for future molten-salt re-
actors and thus have extended the scope of these
studies beyond surveillance of the MSRE. In the
specimens just removed, the tensile specimen rods
were made of heats of Hastelloy N modified to yield
increased resistance to radiation damage. One of
the rods had an addition of 0.52 wt % Ti, and the
other had 0.42 wt % Z:. Besides evaluating the
radiation resistance of these modified alloys, we
should also be able to obtain data on their corrosion
resistance in the MSRE environment.

The alloy modification study was continued in the
new stringer, RS3, just returned to the MSRE. One
of the tensile specimen rods had an addition of 0.5
wt % Ti and 2 wt % W; the other had an addition of
0.5 wt % Hf.

The graphite samples included the anisotropic
MSRE graphite (grade CGB), graphitized-pitch
impregnated grade CGB graphite, isotropic graphite,
pymolytic graphite, and a joint of isotropic graphite
brazed to molybdenum with a 60 Pd—35 Ni-5 Cr
(wt %) alloy. This test of the brazed joint with
radiation and in flowing salt should supplement the
data on corrosion of such joints in static salt with-
out radiation that are discussed later in this
section. Since the radiation dose received in the

 

IMSR Program Semiann. Progr. Rept. Aug. 31, 1965,
ORNL-3872, p. 89.

MSRE is small compared with what will be en-
countered in an MSBR, the primary purpose for in-
cluding various grades of graphite is to extend the
study of fission product behavior with these grades.

The principal objective of this program is to en-
sure the safe operation of the MSRE, We are,
therefore, continuing to retain 65% or more of the
space in the assembly for exposure of the MSRE
grades of Hastelloy N and graphite (see Table
16.1).

16.2 MECHANICAL PROPERTIES OF THE MSRE
HASTELLOY N SURVEILLANCE SPECIMENS

H. E. McCoy

The results of tensile tests run on the first group
of specimens removed from the MSRE were reported
previously.? These specimens were removed after
7823 Mwhr of reactor operation, during which they
were held at 1190 + 18°F for 4800 hr and accumu-
lated a thermal dose of 1.3 x 102?° neutrons/cm?.
The creep-rupture tests on these specimens have
now been completed, and the results are summarized
in Table 16.2. The times to rupture for the surveil-
lance specimens and the controls are compared in
Fig. 16.3. The rupture life is reduced greatly as a
result of the neutron exposure. However, Fig. 16.4
shows that the minimum creep rate is not appreciably
affected. Although the reduction in the rupture life
is large, it is quite comparable with what we have
observed for specimens irradiated to comparable
doses in the ORR in a helium environment. This
point is illustrated in Fig. 16.5. The superior rup-
ture life of heat 5081 is evident. The fracture
strain is the parameter of greatest concern, since
the rupture life does not appear to be reduced
greatly by irradiation at low stress levels. The
fracture strains for the various heats of irradiated
material are compared in Fig. 16.6. The data indi-
cate a minimum ductility for a rupture life of 1 to
10 hr, with ductility increasing with increasing
rupture life. A lower ductility of heat 5085 after
irradiation in the MSRE is also indicated, although
the data scatter will not permit this as an un-
equivocal conclusion. The superior ductility of
heat 5081 and the least ductility of heat 5085 (heat
used for the top and bottom heads of the MSRE) are
clearly illustrated.

 

*W. H. Cook and H. E. McCoy, MSR Program Semiann,
Progr. Rept. Feb. 28, 1967, ORNL-4119, pp. 95-103.
201

Table 16.2. Postirradiation Creep-Rupture Properties of MSRE Hastelloy N Surveillance
Specimens? {rradiated and Tested at 1202°F

 

Stress Rupture Life Rupture Strain Minimum Creep Rate

Test No. Heat No. (psi) (hr) () (% /1r)

R-230 5085 47,000 0.8 1.45 0.81

R-266 5085 40,000 24.2 1.57 0.031

R-267 5085 32,400 148.2 1.05 0.006

R-250 5085 27,000 25.2 1.86 0.004

R-229 5081 47,000 8.7 2.28 0.20

R-231 5081 40,000 98.7 2.65 0.017

R-226 5081 32,400 474.0 3.78 0.0059

R-233 5081 27,000 2137.1 4.25 0.0009

 

STRESS (1000 psi)

 

 

 

 

 

bl

RUPTURE TIME (hr)

ORNL-DWG &7-7938

“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 16.3. Comparative Creep-Rupture Properties of MS5RE Hastelloy N Surveillance and Control Specimens Ir-
radiated and Tested at 1200°F,

We also did some further work to determine the whereas the amount of precipitate in the control

cause of the reduction in the low-temperature

specimens was much less. Hence, the reduction in

ductility reported previously., Using the extraction low-temperature ductility is probably due to the

replica technique, we found that the irradiated

irradiation-enhanced nucleation on the growth of

specimens had extensive intergranular precipitation, grain-boundary precipitates.
 

202

ORNL-DWG 67-7939

 

 

 

 

 

¥ T
: !
[
o
b
= _A‘
‘B
(=N
o
QO
o
@
[6a] i
L) cph
n- :
F .
w
i ] 5081 5085
| o .. CONTROL o A
P ! ‘ SURVEILLANCE @ a
0 4 | C e T :
‘ 11 1 !
‘ o | ‘ :
O, ________________ ‘ JJ‘. ] i . o | AJ
10°° 107> 10 2 10" 10° 10"

MINIMUM CREEP RATE {%/hr)

Fig. 16.4. Comparative Creep Rates for MSRE Haostelloy N Surveillance and Control Specimens

Tested at 1200°F.

ORNL.-DWG 67-7940

 

 

|
50 e 1] Lo | .
8 40 - . i, . ,‘“.,i__ —_ . Or”" | T,. . ,,.
2 AVERAGE IRRADIATED DATA -
@ | o 3
g 30— per ]y T
5 MSRE  ORR | |
o 5065 | .
20 {— & BOBT -1l 1
i . o 5085 | |
' 5084 |
L ; i
0ol —. %,J,J‘ e L P
. | ;\!“; T L
il
0 o i
107! 10° Te}

 

  

   

 

 

 

 

RUPTURE TIME (hr)

Irradiated and

Fig. 16.5. Comparative Stress-Rupture Properties at 1200°F for Various Heats of Hastelloy N Irradiated in the

MSRE and the ORR at 1200°F.
203

ORML-DWG 67794

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

? et TTTTYTTT ] """"" 1Tt [ | ST ’ T T Tt T
g ]
6 MSRE ORR T e '. cperalb L bl cmree otk
A A 5085 o ’
o 5065 || o 3 ‘
L e b L ceeden bbb o b LD
L0 ¢ o 5081 | ‘ | |
2 B 5067 | J
- f10] | ! :
z 4 ‘l ‘ b | .|
<L ' :
o | | | ’ al [T
= i ‘
’ | | |
! ; Lo - _ do ol
Wz - | ‘ | ;\ 1
D_ 1 ; P e ‘
E } i i i o
2 T I e
- ? ..... ol J . 4‘; W -+ A:_ i Lo
LA
i \
........ . ‘ : | s P ‘ - ‘
1 ‘i ! ;“ l . ‘ | ‘
| ol ! i : a i o | i
|‘ | ‘ © | ‘ ‘
Lo L | n
o — 1 bt e Loo...b L g . ‘ sl
107! 10° 10! 10° 10° 1?

RUPTURE TIME (hr)

Fig. 16.6, Comparative Rupture Strains for Various Heats of Hastelloy N Irradiated and Tested at 1200°F,

We removed a second set of Hastelloy N speci-
mens in June 1967. The specimens removed from
the core had been at temperature for 5500 hr and had
accumulated a peak thermal dose of approximately
4« 10%Y neutrons/cm?. The core specimens were
modified alloys containing approximately 0.5% Ti
(heat 21545) and 0.5% Zr (heat 21554). A stringer
of vessel specimens was also temoved. These
specimens had been exposed to the MSRE cell
environment for about 11,000 hr and had accumulated
a peak thermal dose of about 3 x 10 ° neutrons/cm?.
The vessel specimens were made of the same heats
used in constructing the MSRE.

We have not completed the mechanical testing of
these materials, but preliminary metallographic
studies have been completed. The primary concern
with the vessel specimen was whether they were:
being nitrided by the cell environment. Figure 16.7
shows the surface of heat 5085. There is no evi-
dence of nitriding, and the maximum depth of oxida-
tion is about 0.003 in. The specimen taken from

heat 5065 happened to be a weld made in construct-
ing the stringer. This specimen is shown in Fig.
16.8. The amount of surface oxidation is greater,
but the depth of oxidation is still small, and there
is no evidence of nitriding.

Heat 21545 contained 0.5% Ti, and our main con-
cern was whether the alloy would corrode in the
fused salt environment. Fignre 16.9 shows that
there is some slight surface reaction but no ex-
tensive corrosion. Heat 21554 contained 0.5% Zr,
and we were again concerned with the corrosion re-
sistance. Figure 16.10 shows the surfaces of the
surveillance specimens of this heat and indicates
the lack of significant corrosion.

Thus we are encouraged by the observations that
(1) the MSRE grades of Hastelloy N are not being
oxidized at a high rate, and there is no evidence of
nitriding in the cell environment; and (2} the small
zirconium and titanium additions that we are making
to Hastelloy N do not appear detrimental.
204

 

 

*
- * ¥
. . . . - «
¥ . i
i ‘ . -
¥ - ‘? - o
e * 2
- ® S
¥ »
1 . *
.. . t 2
’ ’ . =, ~ ,«’;
wt3
. T
) >
Z Q@
~| 8
O
O
S8
»
5
i .
;
» ' c
Z
(o) *
17

 

 

 

Tew

0.007 INCHES
500x%

Jon

G

   

“ 5)

 

 

Fig. 16.7. Photomicrographs of Surface of Hastelloy N Heat 5085 Exposed to the MSRE Cell Environment for
11,000 hr. (a) As polished. (b) Etched: glyceria regia. The long thermal treatment has resulted in extensive inter-

granular carbide precipitation in addition to the oxidation near the surface.
o

205

Ihs

 

Q.007 INCHES
500X

Ton

 

 

 

R~39368

 

 

 

 

Fig. 16.8. Photomicrographs of the Surface of o Weld in Hastelloy N Heat 5065 After Exposure to the MSRE Cell
Environment for 11,000 hr. (a) As polished. (») Etched: glyceria regia.
206

 

 

[_..

 

S
T

0.007 INCH
500X

Jen

[

 

 

 

INCHES

Q028

 

 

 

Fig. 16.9. Photomicrographs of the Surface of a Specimen of Titanium-Medified Hastelloy N Exposed to the MSRE
Fuel Salt for 4300 hr, {a) As polished. (b) Etched: glyceria regia.
207

Tra

 

O.007 INCHES

T

 

 

 

 

 

 

 

 

 

Fig. 16.10. Photemicrographs of the Surface of o Specimen of Zirconium-Medified Hastelloy N Exposed to the
MSRE Fuel Salt for 4300 hr. (a) As polished. (b) Etched: glyceriaregio.
17. Graphite Studies

17.1 MATERIALS PROCUREMENT
AND PROPERTY EVALUATION

W. H. Cook

The procurement and evaluation of potential
grades of graphite for the MSBR remain largely
limited to experimental grades of graphite. Gener-
ally, most of these were not fabricated specifi-
cally for MSBR requirements.! Consequently, they
tend to have pore entrance diameters larger than
the specified 1 ;1 and gas permeabilities greater
than the desired 107 ° cm?/sec for helium. Other
properties are reasonably good. The latest graph-
ites, discussed below, fall into these classifica-
tions.

Precursory examinations of five experimental and
two specialty grades of isotropic graphite have
been made using 0.125-in.-diam by 1.000-in.-long
specimens machined from stock and tested in a
mercuty porosimeter. The pore entrance diameter
distributions are summarized in Fig. 17.1. The in-
sets on each plot give the grade designation, the
bulk density, and accessible porosity of the speci-
mens tested. The data indicate trends rather than
absolute values for each grade.

Grades AXF-5QBG and H-315A are the specialty
grades, and the others are experimental grades.
Grade AXF-5QBG is an impregnated version of
grade AXF (previous designation, EP-1924), which
looked promising in previous tests in that the en-
trance diameters of most pores were less than 1 p.

The specimens for grade AXF-5QBG, 1-1, and
16-1 (Fig. 17.1) were taken perpendicular and spec-
imens 1-4 and 16-4 were taken parallel with the
4 x 6 in. plane of two different plates, each 11,/2 x
4 x 6 in. Individual results on these were plotted
to show the variations of properties. The pore

 

'W. H. Cook, MSR Program Semiann. Progr. Rept. Feb.
28, 1967, ORNL-4119, pp. 108—10.

208

sizes of the impregnated materials vary morée than
those of the base stock but are around the 1-u size
sought. It would be desirable for the pore entrance
diameters to be smaller and more uniformly distrib-
uted. The impregnation of the base stock reduced
its accessible porosity approximately 30%.

The experimental grade H17-T174 has a pore
distribution similar to that of grade AXF-5QBG.

Grades H-335, -336, -337, and -338 were small
samples of isotropic graphite that had been impreg-
nated and the impregnate graphitized. Grade
H-315A was the base stock impregnated to make
grade H-335. The similar pore size distributions
of H-315A and H-335 illustrate again the well-
known fact that graphite with pore entrance diam-
eters greater than 1 j¢ is not changed much by con-
ventional impregnation techniques.?'* The plot
for H-315A is an average of measurements on four
specimens machined from different locations of a
41Y -in.-OD x 3%-in.-ID x 11%-in. pipe section;
the uniformity of the spectrum of pore entrance
diameters was good.

The three materials H-336, -337, and -338 appear
to have been made from base stocks with pore
spectra similar to that of H-315A. The accessible
porosities measured for H-337 and -338 are rela-
tively low for the small size of test specimens
used.

Some mechanical properties, specific resistivi-
ties, and gas permeabilities are given in Table
17.1 for H-315A and -335 through -338. The me-
chanical properties and specific resistivities of
these grades are satisfactory for MSBR graphite.
The gas permeabilities are high relative to those
sought but seem to be typical for grades of graphite

 

2W. P. Eatherly et al., Proc. U.N. Intera. Cont.
Peaceful Uses At. Energy, 2nd, Geneva, 1958 7, 389
401 (1958).

3w, Watt, R. L.. Bickerman, and L. W. Graham, Engi-
neering 189, 110—-11 (January 1960).
)

209

 

AXF—5QBG
11 3
1.89g/cm
11.56 %

 

 

ORNL —-DWG 6711836

 

 

AXF--5QBG
14
195 g/cm3
B8.81 %

AXF -5Q86
16---4

192 g/cm3
10.83 %

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AXF—5QBG
16 -4
3
8 1.84 g/cm
b 14.94 %
o [
l.-.
w
=z B
1t
a. HAT-Ti74
il §
3 1.80g/cm3
a 13.42 %
x
ooz o
@ H1?;T4?4
& ~ 4B84g/cm3
— L 9.57 %
=
W
= H-335
a 1.83 g/cm?®
i2.07 %
0 TR TS — e T T T Ko
H-338
1.82 g/cm®
13.24 %
4 — - e ———— e - —
H~337
1,98 g/cm3
7.27 Yo
O o ——— e ieamas
3 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, s
H-~338
1.93 9/cm3
8. %
0 —— e A e yoo=e=y T e ETIT et
3 .............. L D R T
H-315A 1.8% g/em3
170 %
() rrmmereemrm———pmnenngmeneerrreeee] )L _—i—]—'—?—“—f"r——r——-r—gl—wjlrj 1 Tt I sy A
2089 10.0 4330 2.0 40 05 04 Q.05 Q02 001

Fig. 17.1. Comparison of the Distributions of the Pore Entrance Diameters for Various Grades of Graphite.

PORE ENTRANCE DIAMETER (u)
210

Table 17.1. Summary of Some of the Properties of Isntropic Graphite

 

 

Bulk Specific Flexural Fracture Modulus of Permeability
Graphite Orientation® Density Resistance Strength Strain Elasticity to He lium Porosity
- Grade (g/cms) {microhms /em) (psi)b (% b (psi)b (cmz,/sec)c (%)
% 10° x 10 *
H-315A |-L 1.830(29° 985(8) 4790 p.a9) 1.35¢4 8.9° 11.705"
C 8o0(®) 588047 0,49 1.65(4)
H-335 ||-L 1.820(1%) 973(13) 4240 0.48(® 1.29¢8) 6.8 12.07
H-336 L 1.823(1%) gs0(! 4900®?  0.49(%) 1.47(0) 3.6 13.24
H-337 |-L 1.967¢13) 901 (1) 7210¢%7  0.70(® 1.82(%) 7.27
H-338 ||-1. 1.92001% 123513 6320 0.54(6 1.81¢9 6.91

 

a“l

I-L” indicates that the length of the specimen
“‘C’" indicates that the length of the specimen w
bWork performed by C. R. Kennedy.

“Work performed by R. B. Evans III.

d . . . .
The superscript numbers in parentheses indicate

was taken parallel with the length of the stock.

as taken parallel with a chord in the pipe circle.

the number of values averaged. Absence of superscript

numbers indicates that the data are from a single specimen.

“Measured perpendicular to the length of the stock.

Table 17.2. Current Fabrication of Graphite Pipe for MSBR Studies

 

 

Graphite Grade Manufacturer Type @ — e Remarks
oD 1D Total Length
1425-64 A Anisotrapic 3% _ 2% 306° Received 12-23-66
H-337 B Isotropic 5 2%, 36° To be shipped by 10-1-67
or sooner
1 1 b
2 {4 1 /2 768
RBRY 12 C Isotropic 5 223/32 C To be supplied during
FY 1968
1 1,

 

“Random lengths, 38 to 51 in.
PRandom lengths, 10 to 36 in.

“Random lengths, 10 to 48 in.; total lengths not fixed at this time,

that approach MSBR requirements. Reducing the
permeability below these by orders of magnitude
appears to be difficult with conventional tech-
niques. This has prompted the backup work on
sealing graphite with metal or pyrolytically de-
posited graphite that is discussed later in this
section.

We are receiving potential irradiation samples of
anisotropic and isotropic graphites in small quanti-
ties from Carbon Products Division of Union Car-
bide Corporation, the Chemical Engineering Devel-
opment Department of the Y-12 Plant,* Great Lakes

 

*Operated by the Union Carbide Corporation for the
U.5. Atomic Energy Commission.
Carbon Corporation, Poco Graphite, Inc., Stackpole
Carbon Company, and Speer Carbon Company.

The first grades of graphite that have been re-
ceived or will be received in larger quantities are
listed in Table 17.2. Grade 1425-64 is being used
in irradiation studies and graphite-to-metal joint
investigations. Grades H-337 and BY12 will be
used in graphite-to-metal and graphite-to-graphite
joint studies and in an engineering test loop. This
loop will be used to determine the operating char-
acteristics of the proposed MSBR fuel cell, with
special emphasis on gas permeability studies to
aid in evaluation of the fission-gas behavior in the
MSBR.

17.2 GRAPHITE SURFACE SEALING
WiTH METALS

W. C. Robinson, Jr.

Chemical vapor deposition is one of the methods
being investigated to decrease the gas permeability
of the graphite. The two metals presently being
considered for a sealant are molybdenum and ni-
obium. The initial objective will be to obtain a
helium permeability of 1077 em?/sec or less with
a minimum thickness of metal deposit. The depo-
sition parameters will be varied in order to deter-
mine the conditions which produce an optimum
coating.

211

The basic technique involves the deposition of
metal on a heated substrate by hydrogen reduction
of the metal halide. In this particular case a
halide-hydrogen gas mixture is passed over
graphite that is contained in a sealed furnace
chamber. The metal halides being used are MoF
and NbCl .

The initial studies are being carried out on a
nearly isotropic grade of graphite, designated
R-0025, which has a helium permeability of ap-
proximately 107" em?/sec and an accessible pore
spectrum with maxima at 0.7 and 4 1. Molybdenum
is deposited via the reactien

IVIOF6 + 3H2 —> Mo + 6HF .

Nine runs have been completed using the experi-
mental parameters given in Table 17.3.

An assembly was built for estimating the helium
permeability of the coated samples. This as-
sembly, which attaches to a Veeco leak detector,
is shown in Fig. 17.2. Two coated and one un-
coated graphite sample are shown. Qualitative
measurements of the helium permeability have been
performed to demonstrate the utility of the as-
sembly. The leak detector will be calibrated with
known leak sources to make quantitative measure-
ments possible. The present qualitative evidence
indicates that the helium permeability of these
samples can be made less than 107 ° with 0.05 mil
or less of molybdenum.

Table 17.3. Molybdenum Coatings on R-0025 Graphite

 

Gas Flow Rates

 

Run Numbor (cms/mirl) Temlzera‘rure Pressure Tix:ne
S e - o (torrs) (min)
MoF 6 H2
M-Mo-1 50 800 700 5 5
M-Mo-2 50 300 700 5 10
M-Mo-3 50 800 700 10 5
M-Mo-4 50 800 700 5 5
M-Mo-5 50 800 700 10 10
M-Mo-6 50 800 800 5 5
M-Mo-7 50 800 800 5 10
M-Mo-8 50 800 800 10 5
M-Mo-9 | 50 800 800 10 10

 
 

212

PHOTO 88991

Fig. 17.2. Apparatus for Measuring the Helium Permeability of Graphite Cylinders; It Is Used with a Standard

Leak Detector.

17.3 GAS IMPREGNATION OF MSBR GRAPHITES

H. Beutler

We are exploring the feasibility of impregnating
graphites with pyrocarbon to reduce the permeation
of gaseous fission products into graphite compo-
nents of the MSBR core. It has been estimated
that the permeability should be reduced to below
107° cm?/sec (for helium) to prevent diffusion of
fission products effectively. So far, we have
carried out a number of exploratory experiments
which demonstrate that this objective can be at-
tained with a gas-phase impregnation technique.

Gas impregnation of graphite for a similar pur-
pose has been studied by Watt et al. > They found
that the gas permeability could be effectively re-
duced by passing hydrocarbon vapors (mainly ben-
zene) in a nitrogen carrier gas over graphite speci-
mens at 1472 to 1652°F. The process relied
entirely on diffusion to carry reactants into the
pores, and there was no need for a pressure differ-
ential across the specimen. With benzene as a
reactant, temperatures near 1382°F were required
for maximum penetration. Above 1472°F, a defi-
nite concentration of deposit on or near the surface

 

SW. Watt et al., Nucl. Power 4, 86 (1959).
was found. In view of the low reaction tempera-
tures, very long treatment times (up to 800 hr) were
required; however, the permeability of tube speci-
mens (1 in. OD, 0.5 in. ID, 1 in. long) was suc-
cessfully reduced from 8.5 x 1073 to 5.4 x 10~8
cm?/sec. By machining off surface layers it was
found that the thickness of the impervious layer
was relatively thin (less than 100 y). The most
desirable attribute of the starting material to be
impregnated was found to be a uniform pore size
distribution.

In the course of our High-Temperature Gas-
Cooled Reactor coated-particle development, we
gained considerable experience with the deposition
of high-density pyrocarbon outer coatings from pro-
pylene over porous buffer coatings. Extensive in-
filtration of pyrocarbon into the porous buffer
layers was noted, even under fast-deposition rate
conditions, if propylene was decomposed at low
temperatures (2282°F and lower). We reasoned
that the same procedure might effectively reduce
the permeability of graphite.

We have carried out a few exploratory experi-
ments using propylene as an impregnant. The
graphite specimens used so far were made from
NCC graphite type R-0025, with the following di-
mensions and properties: 0.400 in. OD, 0.120 in.
ID, 1.5 in. length, 1.9 g/cm? density, and 1 x 101
cm?/sec.

Table 17.4.

Impregnation Experiment

Condition and Results of Gas

 

Specimen number GI-7

Impregnant C‘?'H6 (partial pressure,
32 torrs)
Temperature of impregnation 2012°F
Treatment time 51/2 hr
Weight increase due to 1.2%
impregnation
Thickness of surface deposit 25

Helium permeability

As received ~1 X 103 cm2/sec

After impregnation® 1.4 x 10~7 cmz/sec

After impregnation and 1.7 x 1077 cm? /sec
subsequent heat freat-

ment?

 

%Determined at room temperature by helium probe gas
technique after 16 hr equilibration.

213

Each specimen was immersed in fluidized coke
particles so that it would be supplied uniformly
with a mixture of propylene decomposition prod-
ucts. The conditions and results of a typical ex-
periment are given in Table 17.4.

For the evaluation of impregnated specimens, we
determined weight and dimensional changes and
measured the helium permeability before and after
subsequent heat treatments up to 5432°F in argon.
We are determining the gas permeability of impreg-
nated specimens using a helium gas probe tech-
nique in apparatus similar to that shown in Fig.
17.2.

For a cylindrical specimen, the permeability co-
efficient (K) is given by:

(72)

F = mass flow rate, moles/sec,

21K
In yo/yi

where

K = permeability coefficient, cm?/sec,
! = length of specimen, cm,
yO

y. = inside radius, cm,
I

= outside radius, cm,

P - pressure, atm,
R = gas constant, cm>-atm (°K)™! mole ™!,

T = temperature, °K.

The principal source of error using this technique
is due to inaccuracy in calibration. More impor-
tant, it suffers the disadvantage that it yields a
lower value for K than the true value if measure-
ments are made before steady-state conditions
have been established. We have therefore equili-
brated our specimens with helium for at least 16 hr
prior to determining helium permeability coeffi-
cients. We have, however, noted that the rubber
gaskets which we employed for sealing the speci-
mens are slightly permeable to helium, and our
quoted permeability coefficients are probably
slightly too high.

The results in Table 17.4 indicate that we suc-
cessfully reduced the permeability coefficient from
1x1072%to 1.4 x 1077 cm?/sec after impregnation
for 51/2 hr. A subsequent heat treatment to 5432°F
did not significantly increase the permeability of
the specimen. We noted an increase in specimen
PYROCARBON
DEPOSIT

 

214

Fig. 17.3. Micrograph of Gas-lmpregnated Graphite Specimen, GI-5. Graphite NCC R0025. Impregnation,

2012°F /6 hr/C3H6. Bright field. 750x.

diameter, which suggested a buildup of a surface
layer. Metallographic examination indeed re-
vealed a surface buildup of approximately 15 p.
We found it difficult to resolve pyrocarbon deposits
inside the graphite pores by optical microscopy;
but in isolated areas (see Fig. 17.3), there was
clear evidence of a buildup of pyrocarbon in in-
ternal pores. We machined approximately 4 mils
from the specimen surface and found that the per-
meability increased (K > 10 —° cm?/sec) rapidly,
which indicates that the low-permeability layer is
very thin. At present we are optimizing our depo-

sition conditions to increase the depth of the im-
pervious layer.

We are also preparing irradiation specimens for a
future HFIR experiment. Although our preliminary
results are encouraging, the feasibility of the tech-
nique can only be assessed by fast-flux irradiation
experiments. Because of the dimensional insta-
bility of graphite when irradiated, we must deter-
mine whether the low permeability of the graphite
will be maintained or whether the pyrocarbon will
change dimensions at a different rate and the per-
meability increase.
SiC TEMPERATURE
SENSOR

ALUMINUM
HOUSING TUBE -~

ORNL-DWG 67—-12716

INSIDE SURFACE
COATED WITH
COLLOIDAL GRAPHITE

TUNGSTEN HEATER
iN END POSITIONS

 

 

 

 

 

 

 
   

 

 

 

 

 

 

 

 

 

 

   

 

 

7
/ DRSS
. gfl N
a7 | N N
e
5 . A ; s 73

 

 

£e
kST.AVZUNLESS \

STEEL TUBE

RADIAL
SPACER

“~GRAPHITE SPECIMEN

SiC TEMPERATURE

0.400 in. IAM SENSOR

Fig. 17.4. Schematic Drawing of the Graphite lrradiation Experiment Inserted in the HFIR.

17.4 IRRADIATION OF GRAPHITE

C. R. Kennedy

Experiments to irradiate graphite in target-rod
positions in the core of the High Flux Isotope Re-
actor have begun. The first two experimental con-
tainers have been installed in HFIR for a one-
cycle (three-week) exposure to determine the
accuracy of heating rate and thermal analysis cal-
culations. After this experiment has been removed
and analyzed and the design parameters have been
altered, the long-term graphite irradiations will be
started. It should be possible to obtain integrated
doses (E > 0.18 Mev) of 4 x 1022 neutrons/cm? in
one year.

The facility shown in Fig. 17.4 is designed to
operate by nuclear heating at 1292 to 1328°F. A
uniform axial heating rate will be maintained by
adding tungsten susceptors along the axis of the
samples to compensate for the axial falloff in
nuclear heating. The HFIR control rod design is
excellent for constancy of heating rate, and the
temperature will vary only about 2% during a re-
actor cycle.

Irradiation temperatures will be determined using
beta SiC located in a center hole in each graphite
specimen. The method of temperature determina-
tion using the dimensional expansion and anneal-
ing characteristics of SiC is that described by
Thorne et al.® This procedure has been verified
by irradiation of three SiC specimens in the ORR

at a controlled temperature of 1400°F. The re-
sults indicate that temperatures can be determined
within 9°F of the operating temperature.

Past graphite irradiations to exposures greater
than 1022 (refs. 7 and 8) have been limited to
graphite grades that are similar, with only slight
variations in the filler material or the coke used in
their manufacture. When irradiated at about 1200°F
all graphites seem to be characterized by an ini-
tial shrinkage and then a very rapid expansion.
This rapid expansion corresponds closely to that
observed in the axial direction for single crystals,
so it appears that the binder has deteriorated and
the axial expansion of the individual crystals is
controlling the growth. There are, however, indi-
cations? obtained from short-term irradiations that
there may be potential modifications of the coke
materials that could extend the exposure required
to cause the binder degradation.

 

®R. P. Thorne, V. C. H. Howard, and B. Hope, Radia-
tion-Induced Changes in Porous Cubic Silicon Carbide,
TRG 1024(c) (November 1965).

7J. W. Helm, ‘"Long Term Radiation Effects on
Graphite,’’ paper MI 77, Eighth Biennial Conference on
Carbon, Buffalo, N.Y., June 1567.

5R. W. Henson, A. S. Perks, and S. H. W. Simmons,
““Lattice Parameter and Dimensional Changes in
Graphite Irradiated Between 300 and 1350°C,” paper MI
66, Eighth Biennial Conference on Carbon, Buffalo,
N.Y., June 1967.

gj. C. Bokros and R. J. Price, ‘‘Dimensional Changes
Induced in Pyrolytic Carbon by High-Temperature Fast-
Neutron Irradiation,’’ paper MI 68, Eighth Biennial Con-
ference on Carbon, Buffalo, N.Y., June 1967.
It is the purpose of our studies to develop grades
of graphite which will circumvent the breakdown of
the binder phase that limits the lifetime expec-
tancy of the graphite core. The tailoring of a
graphite for use in a molten-salt reactor will very
likely require the use of impregnates and/or sur-
face-sealing techniques to obtain the required pore
size and permeability. Therefore, studies of the

216

effects that these treatments will have on the irra-
diation behavior will be made. The studies must
include examinations to demonstrate that the core
lifetime is not reduced either through the loss of
effectiveness of the sealing techniques or through
an enhanced potential of breakdown of the binder
phase.
18. Hastelloy N Studies

18.1. IMPROVING THE RESISTANCE
OF HASTELLOY N TO RADIATION DAMAGE
BY COMPOSITION MODIFICATIONS

H. E. McCoy

Although Hastelloy N has suitable properties for
long-term use at elevated temperatures, we have
found that the properties deteriorate when it is ex-
posed to neutron irradiation. This type of radi-
ation damage manifests itself through a reduction
in the creep-rupture life and the rupture ductility.
This damage is a function of the thermal neutron
dose and is thought to be associated with the
helium that is produced by the '°B(n,a) transmu-
tation. However, the threshold helium countent
required for damage is so low that the property
detericration cannot be prevented by reducing the
108 Jevel in the alloy. We have found that slight
modifications to the composition offer considerable
improvement.

Our studies have shown that the normal massive
precipitate, identified ! as M C, can be eliminated
by reducing the molybdenum level to the 12 to 13%
range. 2 The strength is not reduced significantly,
and the grain size is more uniform and more easily
controlled. The addition of small amounts of ti-
tanium, zirconium, or hafnium reduces the irradi-
ation damage problem significantly. Figure 18.1
illustrates the fact that several alloys have been
developed with postirradiation properties that are
superior to those of unirradiated standard Hastelloy
N. We are beginning work to optimize the compo-
sitions and heat treatments of these alloys. We

 

IR. E. Gehlbach and H. E. McCoy, Metals and Ce-
ramics Div. Ann. Progr. Rept. June 30, 1967, ORNL.-
4170,

’H. E. McCoy and J. R. Weir, Materials Development
for Molten-S5alt RBreeder Reactors, ORNIL.-TM-1854 (June
1957).

217

are also initiating the procurement of 1500-1b com-
mercial melts of some of the more attractive alloys.
The properties of the modified Hastelloy N in the
unirradiated condition seem very attractive.
Strengths are slightly better than standard Has-
telloy N, and fracture ductilities are about double.

18.2. AGING STUDIES ON TiTANIUM-
MODIFIED HASTELLOY N

C. E. Sessions

Although normal Hastelloy N is not known to
suffer from detrimental aging reactions, the ad-
dition of titanium to increase the resistance to
radiation damage introduces another variable
which might possibly affect the aging tendencies.
Thus, aging studies are in progress to evaluate
the creep and tensile properties of small com-
mercial heats of modified Hastelloy N containing
varying amounts of titanium.

Initial tests on heat 66-548, which contains
0.45% Ti and 0.06% C, have been completed. Spec-
imens of this alloy were fabricated by two different
procedures and were then annealed at either 2150
or 2300°F. They were aged for 500 hr at tempera-
tures of 1200 and 1400°F and tested at 1200°F.
The results of these tests are given in Table 18.1.

These results indicate that there are no del-
cterious effects of aging up to 500 hr on the ten-
sile properties of modified Hastelloy N containing
0.5% Ti. There ate significant changes in the
strength; the changes in the yield stress do not
seem to follow a pattern, but the ultimate strength
is consistently increased by aging. The elonga-
tion at fracture is increased by aging in most
cases, remains unchanged in a few cases, but is
never reduced significantly.
218

CRNL-DWG 67-3523R

 

 

 

 

    
   
     
      

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

TG | 1 """ NN T i
i ) | ‘ } Sl [
) ‘M o ’
60 (_‘I T | R MELTS ’ P ‘
i | J ’ | J ‘ “\
‘| \ |‘ ! by ! V ‘\'l
- B | | | 16,05 HE,L) #
— i ,
o0 | wmesw, oz
7 [ ‘ (10,05 Ti ,C) ¢ 46,05 'I'li’ | ’
Saol T L, N
S || \}\\l 2,03%0) l(7505HfL “J |
e : |
o | IRRADIATED AIR MELTS NG L
o HH| ] ! o M /{(905TI,L)'
E 30 — SR O O ‘ Py Pl |
5 ’ | ’ | ‘ ‘ (13, 1T|—L) |
3 ‘ .“ ‘ ’ l ]
20— IRRADIATION CONDITIONS t (12.5)
7 =650°C |

D= 2-5%10%neutrons /ocm?

ol Rl _._\___\___i__}_J..l_..

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 1000 10,000
RUPTURE LIFE (hr)

Fig. 18.1. Comparisen of the Postirradiation Creep Properties of Several Hastelloy N Alloys liradiated and
Tested at 1200°F,

Table 18.1. Effect of Aging at 1400 and 1200°F on Tensile Properties
of Titanium-Modified Hastelloy N (Heat 66-548)

 

Tensile Properties®

 

 

Solution
Annealing G_rai;’ Test Condition Yield Ultimate Total
Temperature® Size Strength Strength Elongation
(°F) (psi) {(psi) (%)
x 103 % 103

2300 Large No age 18.5 60.5 50.0
500 hr at 1200°F 18.3 65.5 61.9

500 hr at 1400°F 23.8 68.1 50.0

Small No age 26.6 64.8 33.7

500 hr at 1200°F 22.5 79.1 63.7

500 hr at 1400°F 27.5 81.0 51.4

2150 Large No age 17.7 52.8 37.0

' 500 hr at 1200°F 17.8 64.0 61.5

500 hr at 1400°F 19.2 62.4 51.0

Small No age 21.7 71.3 53.0

500 hr at 1200°F 21.9 77.7 60.1

500 hr at 1400°F 25.3 77.4 50.5

 

“Anncaled 2 hr at temperature.
bLarge and small grain sizes resulted from the two different fabrication procedures.

“Tests were conducted at 1200°F in air using a strain rate of 0.05 min !
These aging studies will be expanded to include
longer aging times, alloys with varying titanium
and carbon levels, and complete metallographic
examination to ascertain metallurgical changes.

18.3. PHASE IDENTIFICATION STUDIES
IN HASTELLOY N

R. E. Gehlbach

The effects of thermomechanical treatments on
the mechanical properties of Hastelloy N indicated
the need for an electron microscope investigation
to understand the nature of the elevated-tempera-
ture embrittlement problem. It was noticed that
the transition from transgranular to intergranular
failure in the temperature range of the ductility
dectease indicated probable formation of a brittle
grain-boundary product — either a fine precipitate
or segregation of impurity elements to grain bound-
aries.

Preliminary examination of thinned foils by trans-
mission electron microscopy showed precipitation
occurring in the same temperature range as that of
the pronounced ductility change. Thus a detailed
study of precipitation in Hastelloy N was initiated
to identify the various types, morphologies, and
compositions of precipitates and to determine their
relationship to the mechanical properties.

Several complementary approaches are being
used for the purposes of precipitate identification.
Thin-foil transmission microscopy is employed for
observations of fine matrix precipitate, dislocation
structure, and, to a lesser extent, grain-boundary
piecipitate. Due to the heterogeneous nature of
precipitation in Hastelloy N, sampling problems
are encountered with transmission microscopy.
This technique is also limited in that the true
nature of the precipitates is not apparent and that
very thin grain-boundary films are frequently un-
detectable.

Extraction replication techniques overcome the
transmission limitations. Here, conventional
metallographic specimens are lightly etched to
remove the polished surface and to reveal pre-
cipitates and are then coated with a layer of car-
bon. Extraction replicas are obtained by heavy
electrolytic etching through the carbon film, which
dissolves the matrix, leaving the precipitates ad-
hering to the carbon substrate. This permits ob-
servation of precipitates in the same distribution

219

as that existing in the bulk sample with the ex-
ception that grain-boundary precipitates are no
longer supported by the matrix and collapse against
the substrate as shown in Fig. 18.2. This allows
direct observation of the grain-boundary precip-
itates on essentially the surface of the original
grain boundary, providing information on true pre-
cipitate morphology and permitting examination

of thin films.

Identification of precipitates is simplified with
extraction replicas, since interference from the
matrix is eliminated. Selected area electron dif-
fraction is being used for the ideatification of
ciystal structure and the determination of lattice
parameters.

Semiquantitative compositional analysis of in-
dividual particles is being performed using an
electron probe microanalyzer accessory on one
of the electron microscopes. Thus, by coordinating
various approaches available in electron micros-
copy, precipitation processes may be studied
carefully,

Much of our effort has been concentrated on
standard Hastelloy N. The microstracture of this
material is characterized by stiingers of massive
M.C carbides, the metallic constituents being
primarily nicke! and molybdenum with some chro-
mium.® The carbides are very rich in silicon com-
pared with the matrix. On aging or testing in the
temperature range 1112 to 1652°F, a fine grain-
This
precipitate is also M, C with the same lattice pa-
rameter (approximately 11.0 A) as the large blocky
carbides. Work is in progress to attempt to detect
any compositional differences between the two
morphologies of M,C carbides. Initial microprobe
work using extraction replicas indicates that the
grain-boundary carbides are richer in silicon than
the large blocky type. All precipitates found in
Hastelloy N which has not been subjected to tem-
peratures in excess of 2372°F have been M.C
carbides.

When standard Hastelloy N is annealed at tem-
peratures above about 2372°F, the M C begins to
transform to an intergranular lamellar product.
Autoradiography, using '*C as a tracer, shows

boundary precipitate forms (Fig. 18.2).

that this product is not a catbide and that the
carbon seems to be rejected (Fig. 18.3). The
blocky precipitates (FFig. 18.3) are seen to be

 

3H. E. McCoy, MSR Program Semiann. Progr. Rept.
Aug. 31, 1965, ORNL-3872, pp. 94-102.
 

220

Y— 77654

 

 

Fig. 18.2. Extraction Replica from Hastelloy N (Heat 5065) Aged 4 hr at 1600°F. All precipitates are M6C

1000x.

carbides.

carbides. Figure 18.4 shows the correspondence
between the morphologies of the extracted pre-
cipitate and that indicated by the autoradiograph.
A selected area diffraction pattern and an elec-
tron probe microanalyzer trace are also included.
This type of precipitate has not yet been iden-
tified, but the diffraction pattern does not cot-
respond to M .C. A comparison of the microprobe
results with standards shows that the noncarbon
constituents are about 90% Mo and 10% Cr. How-
ever, some M _C carbides are also present after
high-temperature anneals and exist in several
morphologies.

We are concluding our investigation of precip-
itation in standard Hastelloy N and will be eval-
uating the information generated and attempting
to relate our observations to observed changes in
mechanical properties.

Because of the attractive in-reactor properties
of the titanium-modified Hastelloy N, we are ex-
panding our studies to include this material. Op-
tical metallography shows the alloy to be quite
free of precipitate. Initial electron microscopy
studies of two experimental heats revealed that

a considerable amount of very thin, fine precip-
itate was present in the grain boundaries. This
was rarely detected in thin-foil transmission mi-
croscopy but was readily seen by the extraction
replica technique.

Selected area electron diffraction studies on
the fine grain-boundary precipitates have revealed
at least two phases, which exist in many mor-
phologies. Although these precipitates have not
been positively identified, indications are quite
good that Ti O is one phase and that it is present
in conjunction with the second type. The other
phase has a face-centered cubic structure with a
lattice parameter of about 4.27 A and may exist
independently of the former. Considering chemical
composition, TiC seems to be the most likely com-
position of the second phase. However, the re-
ported lattice parameter for TiC is 4.33 A, com-
pared with our measured value of 4.27 A. TiN and
TiB have lattice parameters more closely related
to the measured value; however, chemical analyses
of the heats involved show the boron and nitrogen
contents to be extremely low.
AUTORADICGRAPH

 

AUTOR

 

 

 

 

 

   

ey

AS POLISHED

f-82526
2000X

  

190X

Fig. 18.3. Hastelloy N Containing Te Annealed 1 hr ot 2400°F., Autoradiographs made in situ on tightly etched
surface using Kodak NTE liquid emulsion; 312-hr exposure. Reduced 38%.

The thin film of precipitate in Fig. 18.5a is of
the face-centered cubic type with ““Ti, 0" existing
as the ‘‘spines’’ and polyhedral particles. The
threadlike morphology of the face-centered cubic
precipitate is seen in Fig. 18.55.

The final objectives of this work are to identify
all precipitates present and to correlate their for-
mation with the mechanical behavior of the alloy.

18.4. HOT-DUCTILITY STUDIES
OF ZIRCONIUM-BEARING MODIFIED
HASTELLOY:N

D. A. Canonico

Qur previous work? has shown that zirconium
additions cause weld-metal cracking in Hastelloy

N. However, the zirconium addition appears to
be very desirable from the standpoint of improving
the resistance of the material to embrittlement
by neutron irradiation. For this reason, we are
continuing studies to determine how we can im-
prove the weldability of zirconium-bearing alloys.

One tool that we are using in our study is the
Gleeble. The Gleeble permits us to assess the
hot ductility of the base material as it is being
subjected to a thermal cycle simulating that re-
ceived by the heat-affected zone (HAZ) of a weld-
ment.

Specimens are fractured, on heating, at increas-
ing temperatures until they no longer exhibit any

 

*Metals and Ceramics Div. Ann, Proge, Rept. fune 30,
1967, ORNTL.-4170.
     

. i o o T nu-?=§#$
L f#( B TR TR )l' -
v e A% o .

EXTRACTION REPLICA

 

SELECTED AREA
ELECTRON DIFFRACTION

Y—-8254C

 

 

AUTORADIO GRAPH

 

ELECTRON MICROPROBE OQUTPUT

Fig. 18.4. Hastelloy N Containing 14¢ Annealed 1 hr at 2400°F. Correlation of extraction replication, auto-

radiography, selected area electron diffraction, and microprobe analysis for effective precipitate identification.

degree of ductility (as measured by reduction in
area). This zero ductility temperature (ZDT) is
then set as the maximum temperature, and sub-
sequent specimens are subjected to the ZDT,
cooled at a rate which simulates that experienced
by the HAZ, and fractured at a predetermined tem-
perature. The ductility exhibited at these tem-
peratures is compared with the on-heating duc-
tility at the same temperature. The criteria for
evaluating the weldability of the base metal are
its ZDT and, even more important, its ability to
recover its ductility after being exposed to the
ZDT.

The nominal analyses of the experimental alloys
studied in this program are given in Table 18.2.
For comparative purposes, a hot-ductility study
was conducted on an MSRE grade of Hastelloy N.
Its nominal composition is also given in Table
18.2. The zirconium levels in the experimental

alloys ranged from 0 (No. 168) to 0.7% (No. 172).

The influence of the zirconium on the ZDT is
summarized in Table 18.2. It can be seen that
the presence of 0.3% Zr lowered the ZDT by
225°F. This is a rather significant decrease;
however, as has been pointed out earlier, the
ZDT 1is not the sole criterion for evaluating the
effect of zirconium. The recovery of ductility
upon cooling is equally important. The on-heating
and on-cooling results of the hot-ductility study
are shown in Fig. 18.6. It can be seen that alloy
168, which contained no zirconium, had both a
high ZDT and a satisfactory recovery of its duc-
tility. Alloy 169 had a ZDT identical to 168;
however, its ductility upon cooling is unsatis-
factory. Alloys 170 and 171 both had ZDT’s of
approximately 2150°F and both exhibited good
ductilities on cooling.

Figure 18.7 shows the results obtained from
alloys 168, 171, and 174. The results obtained
from an MSRE grade of Hastelloy N have also
YE~ 9340

 

YE— 9344

 

Fig. 18.5. Hastelloy N (Titanium Modified, Heat 66-
548) Annealed 1 hr at 2150°F, Aged 30 min at About
1200°F, Typical precipitote morphologies determined

by extraction replice technique.

been included. It can be seen in Fig. 18.7a that
the Hastelloy N modified by reducing the molyb-
denum content (No. 168) is superior to the com-
mercial material (MSRE grade Hastelloy N). The
addition of 0.5% Zr to the modified alloy lowers
the ZDT by 225°F. As is shown in Fig. 18.75,

 

Table 18.2, The Zero Ductility Temperature
of the Base Materials Studied

 

 

 

Identification Compositioi“(ft %__)__ ZD7T Fiid::;:;n

»  Number Ni Mo Cr Fe Zr (°F) %)
168 bal 12 7 2350 22
169 bal 12 7 0.1 2350 21
170 bal 12 7 0.3 2125 20
171 bal 12 7 0.5 2125 32
172 bal! 12 7 0.7 2120 5

MSRE grade bal 16 4 2300 3

Hastelloy N

 

Measured at 50°F below ZD'T.

the recovery of all four of these alloys is quite
good. Excellent recovery is exhibited by alloy
174. Its recovery of ductility was nearly instan-
taneous and to a rather high level; however, from
the standpoint of weldability, its low ZDT is not
acceptable.

Further efforts will be made to develop a zir-
conium-bearing alloy with a higher ZDT. If we
can find a combination of alloying elements that
will raise the ZDT and retain the good recovery
characteristics, we shall proceed further with the
development of this alloy system.

18.5. RESIDUAL STRESS MEASUREMENTS
IN HASTELLOY N WELDS

A. G. Cepolina

Welding inherently leads to large residual
stresses. These stresses can lead to dimensional
changes and can even be large enough to cause
cracks in the weldment. For these reasons we
have initiated a study to investigate the stress
distribution in and near welds in Hastelloy N.

A technique was developed which allowed con-
tinuous reading of the stress values, thus making
it possible to know, with sufficient precision, the
stress gradient adjacent to the weld axis. For
this study, a 12-in.-diam, l/z*irl."tllick piece of
Hastelloy N base metal was used. Two circular
bead-on-plate welds were simultaneously made,
one on each side of the plate. The diameter of the
REDUCTION IN AREA (%)

ION IN AREA (%)

BUcCT

RE

224

ORNL—DWG 67--11837
100 | ‘ ey e - s [

| EXP HEAT 168 — 0% Zr EXP HEAT 169 — O.( % Zr
| |
! |

 

 

 

   
  

L. b . b . e e -

60 |- -
40 J\.BZ

20 ] S ‘ :
L\.,__W ' | O,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 ! ! ' ! ] l !
EXP HEAT 474 —1% Zr —0.4% Re

S

 

 

 

 

 

 

 

 

TN !

 

 

 

 

 

 

 

 

|
|
|

 

 

 

1400 1800 2200 2600 1400 1800
TESTING TEMPERATURE (°F)

2600

 

Fig. 18.6. Results of the Hot-Ductility Study Showing the Effect of Zirconium on the Modified Hastelloy N.
Symbols: O.H., tested on heating; 0.C., tested on ccoling.

ORNL—DWG 67--11838

 

 

 

 

 

 

 

  
  

 

   

 

 

 

ALLOY NOMINAL COMPOSITION (wi %)
‘68 B/TIL :\/120 C?r Fe zr Fig. 18.7. The Hot-Ductility Results of Selected
174 BAL 412 7 0.5 Alloys.
MSRE GRADE
HASTELLOY N BAL 16 7 4
100 —- —
TESTED ON HEATING , circular weld beads was 6 in. (see Fig. 18.8). The
8O | welds were made with stationary inert-gas tungsten
60 arc torches while the disk rotated around its center
in a vertical plane. This technique was selected
40 in order to minimize the bending effect associated
e with the transverse shrinkage. For the radius and
thickness ratio of our specimen, it is correct to
o & assume that the stress distribution associated
100 - with the weld shrinkage is planar. The perpen-
80 | dicular shrinkage is negligible; that is, there are
no stresses perpendicular to the plane of the disk.
60 — With these assumptions, the stress distribution
a0 | may be determined by measuring only the tangen-
tial strain on the rim of the disk while machining
20 t a series of concentric sections beginning at the
Sl l . disk center.

1400 1606” 1800 200 2200 2400 We followed essentia Hy the “boring Sachs”’
TEST TEMPERATURE (°F) method which was set up for pipes. Since plane
225

ORNL—DWG 6711832
ORML-DWG 67 - 11840

  

o WELD BEAD(ON BOTH $IDES)

——A=in.— THICK PLATE

Fig. 18.8. Sketch Showing the Location of the Weld

Bead in Relation to the Overall Specimen Geometry.

strain and stress problems can be studied with the
same equations, the relationships for the pipe will
be valid for the disk with the elimination of the

 

 

 

terms due to the longitudinal stresses: Fig. 18.9. Pictorial Explanation of the Symhols Used
in the Stress Distribution Equations.
Ee [ R?
O o —O —
r 2 R2 ’
ORNL -~ DWG &7-11841
de A, +A | [T
o, = E (Ao —A) — — ——p
dA 2A )
(see Fig. 18.9),
where

o, = radial stress,
o, = tangential stress,
E - Young’s modulus for the disk material,

e = total tangential strain measured on the
external rim,

R, = disk radius,

R = internal hole radius,

AO = initial disk area,

A = hole area.

The tangential strain is measured by strain
gages that are mounted on the external rim with
their axes oriented along the middle of the rim
thickness (see Fig. 18.10). We used the gages
recommended for stress measurement, type MM
£A06 500 BH. These self-compensated strain
gages have low transversal sensitivity and are Fig. 18.10. Location of the Strain Gages Used to
stable over a relatively long time period. The Measure the Tangential Strain.

 
epoxy used for bonding was EA500, and the pro-
tective coatings (GageKote Nos. 2 and 5) further
assure their stability. A Budd model P350 strain
measuring device was used to obtain the strain
values.

The metal was removed by a milling cutter.

This machining process eliminated the need to
remove the electrical connections to the strain
gages between readings. Thermal effects were
avoided by submerging the specimen in a cutting
fluid solution that was continuously circulated
during machining. To avoid any errors due to the
holders, eccentric clamping was used in order that
the pressure could be easily relieved before read-
ings were taken.

Currently, we are analyzing the results obtained
from the first welds and checking the reproduci-
bility of results with identically prepared speci-
meis.

We intend to study the residual stress distribu-
tion that results from welds made by various proc-
esses and from varying parameters within a given
welding process. We initiated the program with
bead-on-plate welds. The investigation will be
expanded to include the effect of multipass welds
in a V-groove joint configuration. Fach weld pass
will be deposited under identical welding param-
eters, thus permitting us to study the influence of
joint geometry.

The metal-arc inert-gas process will also be in-
vestigated in order to study the influence of this
mode of filler metal addition on the residual stress
distribution. The effect of postweld heat treatment

on the residual stress level will also be determined.

The completion of this program should allow us to
define the welding process, optimum parameters
within that process, and the correct postweld heat
treatment that will minimize the residual stress
level in Hastelloy N.

18.6. CORROSION STUDIES

A. P. Litman

We are continuing to study the compatibility of
structural materials with fuels and coolants of
interest to the Molten-Salt Reactor Program. Nat-
ural-circulation loops are used as the standard
test in these studies.

226

Two loops are presently in operation, Nos. 1255
and 1258. One loop, No. 10, has recently com-
pleted its scheduled circulation time; one loop,
No. 12, prematurely plugged recently; and four new
loops, Nos. 13-16, will start operation as test
salts become available. The latter loops will con-
tain candidate MSBR fuel, blanket, or coolant
salts. Table 18.3 details the service parameters
of these test units.

Fuel Salts

Loop 1255, constructed of Hastelloy N and con-
taining a simulated MSRE fuel salt plus 1 mole %
ThF ,, continues to operate without difficulty after
more than 5.4 years. Loop 1258, constructed of
type 304L stainless steel and containing the same
salt as loop 1255, has logged 4.1 years’ circulation
time with only minor changes in flow character-
istics. To examine the corrosive behavior of the
relatively old simulated fuel salt in this loop, ten
fresh stainless steel specimens were placed in the
hot leg last January. A plot of the weight change
for the new specimens at the hottest point in the
system and a comparison with earlier data as a
function of time are shown in Fig. 18.11. It is
clear that very rapid attack occurs in the first 50

ORNL-DWG &7 ~11842

 

- DATA FOR SPECIMENS DURING LOOP START-UP 1964
DATA FOR NEW SPECIMENS 1967
ZONE = 4 mil/yr UNIFORM ATTACK
ZONE Z 2 mil/yr UNIFORM ATTACK

i

  
   

WEIGHT CHANGE (mg/cm?)

 

AN

\‘4\

2000 3000 4000
TIME {hr}

20— ——————
0 1000

Fig. 18.11. Weight Change as a Function of Time for
Type 3041 Stainless Steel Specimens Exposed at 1250°F
in Loop 1258 Containing LiF-BeFZ-ZrF4-UF4-ThF4 (70-
23-5-1-1 mole %).
227

Table 18.3. Thermal Convection Loop Operation Through August 31, 1967

 

 

 

Maximum A Time
Loop AT
No Loop Material Hot-Leg Specimens Test Fluid Temperature (°F) Operated
LCF) (hr)
1255 Hastelloy N Hastelloy N + 2% Nb® LiF-BeF Z-ZrF 4—UF4*ThF4 1300 160 47,440
(70-23-5-1-1 mole %)
1258 Type 304L Type 304L stainless LiF-BeF?_-ZrF4-UF4—ThF4 1250 180 36,160
stainless steel steel (70-23-5-1-1 mole %
10 Hastelloy N None NaF-KF-BF | 1125 265 8,765°
(48-3-49 mole %
12 Croloy 9M*? Croloy 9M¥ I‘L’:lF-KF-BI"T3 1125 260 1,440
(48-3-49 mole %) (plugged)
13 Hastelloy N Ti-modified LiF-BeF Q-UF 4 1300 300 (c)
Hastelloy N (65.5-34-0.5 mole %)
14 Hastelloy N Ti-modified LiF—ThF4 (71-29 mole %) 1250 100 ()
Hastelloy N¢
15 Hastelloy N Ti-modified NaF—BF3 (50-50 mole %) 1125 300 (e)
Hastelloy N9
16 Hastelloy N Ti-modified NaF-BF3 (50-50 mole %) 1125 300 (e)
Hastelloy N9
a : d. .
Permanent specimens. Removable specimens.
PScheduled loop shutdown 5-23-67. “Scheduled to start operation 9-15-67.

°Scheduled to start operation 9-30-67.

hr of exposure, and the rate of weight loss sub-
sequently declines with time. While several per-
turbations in the rates are obvious, in general,
the rate loss has remained constant at between 1

PHOTO 74314A

   
  
 

Fe PLUG
and 2 mils per year equivalent uniform attack. -

This is many times the corrosion rate observed
for Hastelloy N under similar conditions.

~ Ainch

Coolant Salts

FLUCROBORATE SALT = L COLD LEG

Loops 10 and 12, both of which have ceased
operation, contained a fluoroborate salt which is
a candidate coolant salt because of its low cost
and low melting point. Loop 12, constructed of
Croloy 9M, ° plugged after 1440 hr circulation due
to mass-transfer and deposition of essentially
pure iron crystals in the coldest portion of the
loop. Similar crystals were found adhering to
specimens in the hot leg. A green deposit with " DRAIN PIPE
a variable composition of 15 to 17% Fe, 11 to 14%
Cr, 2 to 4% B, 1.5% Mn, 10 to 15% Na, and 46% F
was noted in the drain pipe (Fig. 18.12). A best Fig. 18.12. Plugging in Croloy 9M Loop 12 Containing
Nc:F-KF-BF3 (48-3-49 mole %) after 1440 hr at 1125°F,
SNominal analysis 9% Cr—1% Mo—Fe balance. AT = 260°F,

2 NoF-Fe F,-CrF,-BFy DEPOSIT

 
  

PHOTO 74665A

PIPE WALL

3NaF —Crfy PLUG

Fig. 18.13. Plug Formed in Hastelloy N Loop 10 Containing N(JF-KF-BF3 (48-3-49 mole %) After 8765 hr at

1125°F. AT = 265°F. 8x. Reduced 23.50%.

estimate of the stoichiometry of the deposit is
2NaF-FeF ,~CrF ,-BF ;. Metallographic examination
of the hot-leg specimens disclosed only moderate
surface roughening.

Loop 10, fabricated from Hastelloy N, operated
without incident for 8335 hr, at which time the
hot-leg temperature increased about 50°F, ac-
companied by a simultaneous temperature decrease
of the same magnitude in the cold leg. A pertur-
bation of this type is usually an indication of
plugging. The temperature fluctuations ceased
after 1 hr, and no further incidents occurred during
the life of this loop. The loop was shut down
after its 1l-year scheduled operation. Examination

of the loop piping disclosed a partial plug in the
lower portion of the cold leg (Fig. 18.13). The
plug, which closed approximately 75% of the cross-
sectional area of the pipe, was emerald green in
color and analyzed to be essentially single Crys-
tals of 3NaF—CrF3. 6

Analysis of the drain salt cake from loop 10,
Fig. 18.14, revealed extensive increases in the
concentrations of Ni, Mo, Fe, and Cr when com-
pared with the before-test salt. Nickel was par-
ticularly high in the bottom and top portions of

 

6R. E. Thoma, private communication, June 21, 1967.
229

ORNL—DWG 67-10980

Impurity Analysis: NaF-KF-BF; (48—3—49 mole %)

 

 

 

 

Ni Cr
Before test 87 83
After test
Top slag 11.15% 1000
Center layer 90 210
Bottom layer 4,472 1500
Lwt. %.

 

 

Analysis (ppm)

 

 

 

Mo Fe 0 H-0 S
1400 400

7 146 3000 300 2,6
4850
1200

1.352

35 4200 1750 1800 5

3540

160 270 3120 1300 2, <5

7300 1500 gggg 2800 <5,19

 

Fig. 18.14. Drain Salt Cake from Laop 10.

the cake, 4.47 and 11.15 wt % respectively. Mi-
crometer measurements and metallographic exam-
ination of the hot leg of the loop disclosed 1 to 2
mils of metal loss and slight surface roughening.
The crossover line to the cold leg, the cold leg,
and the crossover line to the hot leg all showed
slight increases in wall thickness due to depo-
sition of complex surface layers (Fig. 18.15).
Chemical analysis of the layers disclosed that
they were primarily metallic nickel (60 to 90 wt %)
and molybdenum. Iron was present at locations
close to the base metal, but chromium was absent
in all the corrosion product layers examined. This
is reasonable in view of the complex fluoride plug
and the observation that in the cold leg two com-
plex iron fluorides, 3NaF‘—FeF3 and NaF-FeF ,,
were identified.

Analysis of this loop is still proceeding, but
the mode of attack appears to be dissolution of
the container material in the hot leg, followed by
temperature-gradient mass transfer. Chromium
and iron, the latter to a lesser extent, tend to
form complex fluorides, while nickel and molyb-
denum plate out in metallic form. The findings
to date indicate that corrosion damage was due
to (a) an impure salt containing residual HF or
H,O from processing or (b) the intrinsic corrosive
characteristics of the fluoroborate salt (BF3 is
probably the active agent). In any case, while
the effect on the Hastelloy N is small in terms
of metal loss from the hot section, the formation
of metallic layers and complex fluorides in the
colder sections is disturbing. Blockage of flow
and modification of heat transfer characteristics
ORNL-DWG 67-9342

 

 

32.5in

Q007 inches

   
   
  

 

 

 

LOCP 10
SALT: NaF-KF-BFy (48-3:49 mole %)

TEMPERATURE:H125°F, A7T=285°F, TIME=8760hr

LOCATION QF
3NaF-Cr F3 PLUG-...

0.007 inches

 

 

 

 
 

Fig. 18.15. Surface Laoyers in Loop 10 after 8760 hr Operation.

are, of course, the deleterious results of primary
concern.

Equipment Modifications

During the last six months extensive modifi-
cations were made to the thermal convection loop
area in Building 9201-3 so as to make it more
suitable for advanced studies on the compatibility
of MSBR salts with Hastelloy N and modified
Hastelloy N (0.5% Ti). All instrumentation has
been or is in the process of being upgraded, out-
moded furnaces are being replaced, and special
facilities have been installed to handle toxic BF
gas. Two loops ready to be charged with BF ,-
bearing salts are shown in Fig. 18.16. Our future
program includes operation of natural-circulation
and possibly forced-circulation loops with the
prime candidate fuel, blanket, and coolant salts
for the MSBR (loops 13-16). A study will be made
of the compatibility of graphite—Hastelloy N braze
joints in fuel salt. Capsule tests to determine the
effect of BF , pressure on the compatibility of the
fluoroborate salts with Hastelloy N will also be
conducted.

18.7. TITANIUM DIFFUSION
IN HASTELLOY N

C. E. Sessions T. S. Lundy

Titanium additions to Hastelloy N improve the
resistance of this alloy to damage by neutron ir-
radiation. However, we have some concern about
how the addition of titanium will influence the
cotrosion resistance. Evans et al. 7 have shown
that the primary corrosion mechanism of standard
Hastelloy N in pure fluoride salts is the leaching
of chromium by the reaction

UF, +Cr— UF, + CtF, .

They demonstrated that the process was con-
trolled by the diffusion of chromium in the alloy.
Titanium will tend to undergo a similar reaction,
since the standard free energy of formation of
TiF , is even more negative than that for CrF,

 

7R. B. Evans III, J. H. DeVan, and G. M. Watson,
Self-Diffusion of Chromium in Nickel-Base Alloys,
ORNL-2982 (Jan. 20, 1961).
231

PHOTO 74946

 

 

Fig. 18.16. Hastelloy N Thermal Convection Loops Ready for Charging with BF ;-Bearing Salts.
ORNL-DWG 67-11844

   

m 0.500in. L——

*—-‘ 0.250in.

Fig. 18.17. Geometry of Diffusion Specimen.

(at 1110°F, AF®for TiF, is ~90 kcal per gram-
atom of F and AF®for CrF, is —77 kcal per gram-
atom of F).® Since protective films do not seem
to form in fluoride systems, we would assume that
both the chromium and the titanium will be re-
moved as rapidly as these elements can diffuse.
Our previous studies indicate that the corrosion
rate of the standard alloy is acceptable, and the
question of paramount importance is whether the
addition of titanium will accelerate the corrosion
tate.

To answer this question we are measuring the
rate of titanium diffusion in modified Hastelloy N.
The following technique? is being used. Dif-
fusion samples, 0.625-in.-diam cylinders (Fig.
18.17), were machined from small commercial heats
of modified Hastelloy N (heat 66-548). They were
heat treated 1 hr at 2400°F to establish a stable
grain size and impurity distribution. The radio-
active **Ti isotope was deposited on the polished
face of the sample using a micropipet. The iso-
tope was supplied in an HF-HC1 acid solution;
therefore, additions of NH,OH were made after
depositing the isotope to neutralize the solution.
The sample was then heated in vacuum for 0.5
hr at 932°F to decompose the mixture to leave
a thin layer of **Ti.

Each sample was given a diffusion anneal for
appropriate times in flowing argon at precisely
controlled temperatures. Sections were then
taken on a lathe at 0.001-in. increments, and the
activity of the turnings was measured using a
single-channel gamma spectrometer with an Nal(T1)
scintillation crystal detector. At lower diffusion
anneal temperatures, a hand-grinding technique

 

8A. Glassner, The Thermodynamic Properties of the
Oxides, Fluorides, and Chlorides to 2500°K, ANL-5750.

9_]. F. Murdock, Diffusion of Titanium-44 and Vana-
dium-48 in Titanium, ORNL-3616 (June 1964).

232

ORNL-DWG 67-11845
TEMPERATURE {°C)

 

 

_9 125C 1200 1150 1100 1000
1C . -
5
—_ 2
<«
@
< -0
NE 10
L
I
=z 5
wl
O
m
L
e 2
O
=
o —1
& 1°
o
L
L
o 5
2 .....
10—12
65 70 75 80 85

10,000 /7 s

Fig. 18.18. Diffusivity of Titanium in Modified
Hastelloy N.

was used to obtain smaller increments since the
penetration distances were less. From a plot of
the specific activity of each section against the
square of the distance from the original interface,
a value of titanium diffusivity was obtained for
the diffusion anneal temperature of that specimen.

To date we have determined the diffusivity of
titanium at five temperatures from 2282 to 1922°F.
The results of these measurements are shown in
Fig. 18.18. Although the temperature range over
which we have data is considerably higher than
the proposed reactor operating temperature, we
can compare the rates of titanium diffusion with
that of chromium to get some ideas of the rela-
tive mobilities of the two constituents. At
2012°F the diffusion rate of chromium in an Ni—20%
Cr alloy is reported to be approximately 8 x 10— 1!
cm?/sec.’® From our data at 2012°F the diffu-
sivity of titanium in modified Hastelloy N is 3.9 x
10— ! e¢m?/sec, which is a factor of 2 lower than
for chromium at this temperature.

Thus on a very rough basis there does not ap-
pear to be too much difference in the rate of dif-
fusivity of these constituents at these higher tem-
peratures. However, we cannot conclude at this

 

10P. L. Gruzin and G. B. Federov, Dokl. Akad. Nauk
SSSR 105, 26467 (1955).
time what the effective diffusivities of titanium
wiil be in the alloy at 1100 to 1400°F, because

at these lower temperatures short-circuit diffusion
paths become an important factor in the material |
transport. Since we currently have no estimate of
this latter contribution to the net diffusivity for
titanium, we must extend our diffusion measure-
ments to lower temperatures before concluding
what the expected behavior would be under reactor
opetating conditions.

18.8. HASTELLOY N-TELLURIUM
COMPATIBILITY

C. E. Sessions

The compatibility of Hastelloy N with fission
products in the MSRE is of concern, since the
strength and ductility of the structural material
could be reduced after prolonged exposure at ele-
vated temperatures. Consideration as to which
of the products might be detrimental to the strength
of the alloy revealed that tellurium was a poten-
tially troublesome element. Tellurium is in the
same periodic series as sulfur, a known detrimental
element in nickel-base alloys.

To evaluate the possible effects ot tellurium on
Hastelloey N, several tensile samples were vapor
plated with tellurium and then heat treated in
quartz capsules to allow interdiffusion of the tel-
lurium with the alloy. Also, several samples were

233

vapor coated with tellurium and then coated with
an outer layer of pure nickel in order to reduce

the vaporization of the tellurium during subsequent
heat treatments.

After heat treatment of the coated samples, the
specimens were tensile tested at either room tem-
perature or 1200°F using a strain rate of 0.05
min~!. Table 18.4 lists the test conditions and
results for 12 samples of Hastelloy N. No effect
of the tellurium coating on the ductility of Hastel-
loy N was found at either test temperature. At
12007 F, the ductility following the various treat-
ments ranged from 20 to 34% elongation, which
is within the range normally obtained in the ab-
sence of tellurium. At room temperature the duc-
tility was in the range 52 to 57%, which again is
normal.

Metallographic examination of the specimens
was made after testing to evaluate the interaction
of tellurium with Hastelloy N. A representative
area on the shoulder of one sample is shown in
Fig. 18.19. Irregular surface protuberances are
evident in a localized region of the edge of the
Hastelloy N. The gray phase around the protu-
berances is tellurium metal that remained on the
sample after mechanical testing in air at room
temperature. The roughness of the Hastelloy N
specimens resulted from corrosive interaction of
the vapor or liquid tellurium during the heat treat-
ment. Al higher magnifications it is evident that
a slight amount of grain-boundary penetration of
the tellurium into the Hastelloy N had taken place.

Table 18.4. Tensile Properties of Tetlurium—Hastelloy N Compatibility Studies®

 

Maximum

 

) Annealing Test Yield Total
Sample Coating Annealing Time Temperature Strength Elongation
Technique Temp:rature (hr) (°F) (psi) @)
(CF)

Tellurium in 2012 24 75 49,100 57
capsule with 2012 24 1200 31,300 22
specimen 2012 150 75 49,800 57

2012 150 1260 37,000 20

Vapor coated with 1652 100 75 53,200 52
tellurium and 1652 100 1200 36,700 27
nickel

Vapor coated with 1472 100 75 52,800 54
tellurium 14772 100 1200 40,500 34

 

"Tested at a strain rate of 0.05 min —1
234

Y-79588

 

 

100X

T

 

 

 

Fig. 18.19. Hastelloy N Tensile Specimen Following Exposure to Tellurium Vapor for 150 hr at 1832°F. 100x.

 

Y-79930

 

 

 

 

. 5
& # .0 =
1
e 1
A
1
* e
‘j;_‘ % 2
* *
- -3
' o
.
i * . ZJ é
= T
Ol
. “|S
. o Sls
O
-
-
» o2
Choen
N X - 5
7 o
, *
o
- a.
* -
T =
o *; o, -
T o
. ,__/’? - * -

Fig. 18.20. Hastelloy N Sample Showing Slight Intergranulor Penetration by Metallic Tellurium. 500x.
Figure 18.20 shows an area near the edge of the
sample, where grain boundaries contain a tellurium-
tich phase. Electron microprobe examination of
this region confirmed the presence of tellurium and
indicated that it was preferentially associated with
the large carbide precipitates within grain bound-
aries.

The penetration of the tellurium into the alloy
was only about 0.005 in. following the 1832°F
heat treatment. No penetration was found for
samples heat treated at 1472 or 1652°F.

Elemental scans of the microprobe within the
tellurium phase shown in Fig. 18.19 indicated

235

that manganese and chromium were also present.
These elements were preferentially sepregated
into bands within the tellurium matrix and un-
doubtedly resulted from the leaching of these ele-
ments from the Hastelloy N during the 1832°F
heat treatments.

The results of these tests indicate that liquid
and/or vaporized tellurium metal interacts with
Hastelloy N to a minor extent under the conditions
tested, which included up to 150 hr in contact at
1832°F. No reduction in room-temperature or
elevated-temperature ductility was found. These
tests will be extended to longer exposure times.
19. Graphite-to-Metal Joining

19.1. BRAZING OF GRAPHITE TO

HASTELLOY N

W. J. Werner

Studies were continued to develop methods for
brazing large graphite pipes to Hastelloy N. At the
present time, three different joint designs are being
tested for circumventing the differential thermal ex-
pansion problem associated with joining the two ma-
terials. Concurrently, two different brazing tech-
niques are under development for joining graphite to
graphite and to Hastelloy N. In addition to wetta-
bility and flowability of the brazing alloy on graph-
ite, the brazing technique development takes under
consideration the application-oriented problems of
service temperature, joint strength, compatibility
with the reactor environment, and braze stability
under irradiation in neutron fluxes.

Joint Design

The first joint design, which has heen reported
previously,! is based on the incorporation of a
transition material between the graphite and
Hastelloy N that has an expansion coefficient be-
tween those of the two materials. Molybdenum and
tungsten are two applicable materials, and, in ad-
dition, they possess adequate compatibility with
the reactor system. The transition joint design is
illustrated in Fig. 19.1. The design incorporates
a 10° tapered edge to reduce shear stresses arising
from thermal expansion differences.

The second design is the same as the first except
that the transition portion of the joint is deleted
and the graphite is joined directly to the Hastelloy

 

Lysr Program Semiann. Progr. Rept. Feb, 28, 1966,
ORNIL-2936, p. 140,

236

 

Fig. 19.1. Transition Joint Design with 10-deg
Tapered Edges to Reduce Shear Stresses. From left to
right the material s are graphite, molybdenum, and

Hastelloy N.

ORNL-DWG 67 -1184¢

e e ]

 

 

0000 CLEARANGCE
HASTELLOY N~

GRAPHITE

Fig. 19.2. Schematic Hlustration of a Direct Huastelioy
N—to—Graphite Jaoint,

N. This joint is, of course, more desirable from

the standpoint of simplicity. In addition, the graph-
ite is in compression, which is highly desirable for
a brittle material. The amount of compressive strain
induced in a 3% -in.-0D by ¥,-in.-wall graphite tube
using this design is approximately 1%, which is
postulated to be an acceptable compressive strain
for high-density, low-permeability graphite.
The third design, which is also a direct graphite—
to—Hastelloy N joint, is shown schematically in
Fig. 19.2. Once again we have the graphite in
compression; however, in this case allowance is -
made for keeping the joint in compression even if
the graphite should shrink under irradiation.

Brazing Development

We are continuing wotk on the development of
alloys suitable for joining graphite to graphite and
to structural materials, We are currently looking at
several alloys based on the corrosion-resistant
Cu-Ni, Ni-Pd, Cu-Pd, and Ni-Nb binary systems.
Quatemary compositions were prepared containing
a carbide-forming elemeat plus a melting-point de-
pressant. Preliminary discrimination between the
various alloys was obtained through wettability
tests on high-density graphite. Poor flowability
was obtained with the Ni-Nb and Cu-Pd alloys. In
the Pd-Ni system, the carbide-forming elements and
melting-point depressant were added to the 70-30,
60-40, and 50-50 binary alloys. In the Cu-Ni sys-
tem, the carbide-forming elements and melting-point
depressant were added to the 80-20 and 70-30 binary
alloys. Most of the alloys seemed to wet graphite
well at temperatures ranging from 2102 to 2192°F.

Concurrent with the brazing development work,
we are investigating the radiation stability of the
brazing alloys. We are cutrently irradiating four
batches of Hastelloy N Miller-Peaslee braze speci-
mens in the ORR. The specimens will receive a
dose (thermal) of approximately 2 x 1020
neutrons/cm? at 1400°F,

19.2. COMPATIBILITY OF GRAPHITE-
MOLYBDENUM BRAZED JOINTS WITH MOLTEN
FLUORIDE SALTS

W. H. Cook

The salt-corrosion studies of joints of grade CGB
graphite brazed to molyvbdenum with 60 Pd-35 Ni--5
Cr (wt %) have continued. The specimens are ex-
posed to static LiF-BeF -ZrF ,-ThF -UF  (70-
23.6-5-1-0.4 mole %) at 1300°F in Hastelloy N. We

237

have reported previously?'3 that there was no visible
attack on the braze after a 5000-hr exposure, but
there was a coating of palladium on the braze and
some Cr3C2 on the graphite. A 10,000-hr test has
now been concluded with similar results.

All salt-corrosion tests of this series were sealed
at room temperature with a pressure of approximately
4 % 107 ° torr by TIG welding. A thermal control
for the 10,000-hr salt test was made in which the
test components and test history were the same ex-
cept that no salt was present. The results are
shown in the microstiuctures of the two joints in
Fig. 19.35 and ¢. The diffusion of the palladium
out of the brazing alloy to form a nearly pure
palladium coating on the surfaces of the braze oc-
curred both in the control (the one exposed to a
vacuum) and the one exposed to the salt. The
coating formed in the vacuum may be mote uniform.
Formation of the coating in the vacuum eliminates
the salt as an agent in its formation. The more
probable explanation is that the palladium is
diffusing to the surface of the braze metal. The
thickness of the coating appeared to be a function
of time in the 5000-hr test, but this time depend-
ence does not seem to continue for as long as
10,000 hr.

There is some possibility that the palladium
coating may help prevent corrosion of the brazing
alloy by decreasing ot preventing exposure of the
alloy to the salt.

The chemical analyses of the salts remained
essentially unchanged, as shown in Table 19.1,
with the exception that the chromium content of the
salt in the 10,000-hr test rose sharply relative to
the others. This is higher than one would expect
with Hastelloy N in these types of tests.

This particular test series for this brazing alloy
will be terminated by a 20,000-hr test which is in
progress. Another corrosion test of this brazing
alloy in a similar joint configuration is being made
in the MSRE core surveillance assembly, where the
joint is being exposed to radiation and flowing fuel
salt.

Other potential brazing alloys will be subjected
to similar tests as they are developed. The most
promising alloys will be more rigorously tested in
dynamic salts and irradiation fields.

 

2W. H. Cook, MSR Program Semiann. Progr. Rept.
Aug. 31, 1966, ORNL-4037, pp. 115-—-17.

w. H. Cook, MSR Program Semiann. Progr. Rept.
Feb. 28, 1967, ORNL-4419, pp. 11115,
238

o

Pd COATING =g

 

PHOTO 89483

0.030in, -

Fig. 19.3. Microstructure of Joints of Grade CGB Graphite Brazed to Molybdenum with 60 Pd—35 Mi—5 Cr (wt %).
(a} As brazed, (b) after 10,000-hr exposure at 1300°F to vacuum, and (c) after 10,000-hr exposure at 1300°F to LiF-
BeF2-ZrF4-ThF4-UF4 (70-23.6-5-1-0.4 mole %). The exposed surfaces are te the right.

Etchant:

Table 19.1. A Summary of the Results of the Chemical Analyses of LiF-Ber-ZrF4-ThF4-UF4

 

Hastelloy N to Grade CGB Grophite Joints Brozed with 60 Pd—35 Ni—-5 Cr (wit %) ?

10% oxalic acid.

 

 

 

 

 

100x.,

(70-23.6-5-1-0.4 Mole %) Salts Before and After Various Periods of Exposure in
Test Period

Test No. (hr) Pd Cr Fe

Control 0 <0.005 0.047 0.018

A2361 100 <0.005 0.055 0.015

A2362 1000 <0.005 0.056 0.012

A2363 5000 < 0.00006 0.053 0.004

A2365 13,000 <0.0003 0.4000 0.0040

Ni Mo C
0.013 <0.001 < 0.01 0.133
0.016 <0.003 0.048 0.038
0.006 <0.003 0.049 0.026
0.003 <0.005 0.20 0.22
0.0605 <0.0003 0.0012 0.0748

 

?The salt components remained constant within the tolerances of the chemical analyses.
Part 6. Molten-Salt Processing and Preparation

M. E. Whatley

We have continued development work on the more
critical parts of the fuel processing system for an
MSBR. The basic concept of an on-site plant which
will continuously process 14 £t of fuel salt per day
through a fluorination operation to recover the ura-
nium, a distillation operation to separate the carrier
salt from the less-volatile fission products, and a
recombination of the purified carrier salt and uranium
remains unchanged. The area of work which has re-
ceived most attention during this period is the dis-
tillation operation. Relative volatility values have
been more accurately determined and extended to
additional fission product poisons. The technique
of using a transpirational method for relative vol-
atility measurement is being exploited. We have
fabricated the experimental still which will be used

20. Vapor-Liquid Equilibrium

J. R. Hightower

at the MSRE to demonstrate distillation of radioactive
salt, and it will be tested soon. The alternative
process, removal of rare-earth fission products by
reductive extraction, still appears promising; there-
fore, this work is continuing. The problem of heat
generation in the various components of the process-
ing plant is being elucidated by a new computer code
which will provide concentrations of individual fis-
sion products with adequate accuracy for this purpose.
We now plan to recover the uranium from the MSRE
in early 1968, with a cooling time of only 35 days;
for this procedure some modification of the process-
ing cell will be necessary. The new loading of the
MSRE in 1968 will use ?33U. Plans for the prepara-
tion of this fuel are reported here.

Data in Molten-Salt Mixtures

L.. E. McNeese

Relative volatilities of several rare-earth trifluorides having the desired composition was melted in a

and alkaline-earth fluorides with respect to LiF have

been measured in an equilibrium still at 1000°C.. The

measured relative volatilities are in close agreement
with those predicted by Raoult’s law and are low
enough that a simple distillation system will work
well for removing rare-earth fission products from the
MSRE fuel stream. '

The equilibrium still used for these measurements
is shown in Fig. 20.1. The still consists of a 1Y
in.-diam still pot, a 1-in.-diam condenser, and a trap
through which condensate flows prior to its return to
the still pot. For each experiment, a salt charge

239

graphite crucible. The mixture was then cooled to
room temperature, and the resulting salt ingot was
loaded into the 1% -in.-diam section of the still.
The still was welded shut, the condenser section
was insulated, and the still was suspended in the
furnace. Air was purged from the system by re-
peatedly evacuating the system and filling to at-
mospheric pressure with argon, and the pressure
was set at the desired level. The furnace tempera-
ture was raised to 1000°C and the condenser tem-
perature set at the desired value. During runs with
a given component dissolved in LiF, the operating
ORNL OWG 66-8393

    
  
 
  
   
  

 

 

 

1-in.
NICKEL
PIPE
|
g TO VACUUM
P PUMP
C/> T
q
/p
d
1'% -in. c/
NICKEL PIPE —___ -
d
q
¢
VAPCRIZING
LIQUID
CONDENSED
f }zfli- SAMPLE
o AIR-WATER
MIXTURE
“NICKEL TUBING

 

Fig. 20.1. Molten-Salt Still Used for Relative ¥Yola-

tility Measurements.

Table 20.1. Relative Volatilities of Rare-Earth

Trifluorides, YF3, Ban, Ser, Z.rF4,
Ber at 1000°C with Respect to LiF

and

 

 

Relative Volatility Relative Volatility

 

Compound in LiF—BeF2-REF in LLiF-REF

Mixture Mixtures

CeF, 3.3x107* 4.2 x 1074

LaF 1.4 x 1074 3x107*

NdF <3 x10~" 6.3 x 1074

PrF 1.9x 1073 2.3 x 1074

SmE <3x10"*

YF, 3.4 x 1077

BaF, 1.1 x 1074

SIF,, 5.0 x 107°

ZrF 0.76 to 1.4

BeF, 4.73

 

240

pressure was 0.5 mm Hg and the condenser outlet
temperature was 855 to 875°C; during runs with the
LiF-Ber mixture, the pressure was 1.5 mm Hg and
the condenser outlet temperature was 675 to 700°C.
An experiment was continued for approximately
30 hr, after which the system was cooled to room
temperature. The still was then cut open in order
to remove the salt samples from the still pot and
condensate trap. The samples were analyzed for
all components used in the experiment.
Experimentally determined relative volatilities of
six rare-earth trifluorides, YF3, Ban, Ser, Ber,
and ZrF  with respect to LiF (measured at 1000°C
and 1.5 mm Hg in a ternary system having a molar
ratio of LiF to Bel', of approximately 8.5) are given
in Table 20.1. The mole fraction of the component
of interest varied from 0.01 to 0.05. The relative
volatilities of the fluorides of the rare earths, Y,
Ba, and Sr are lower than 3.3 x 107 *, with the ex-
ception of Pr, which has a rclative volatility of
1.9 x 1073, The relative volatility of Z:F, varied
between 0.76 and 1.4 as the ZrF, concentration
was increased from 0,03 to 1.0 mole %. The average
value of the relative volatility of Bel") was found
to be 4.73, which indicates that vapor having the
MSBR fuel carrier salt composition (66 mole %
LiF~34 mole % BeF,) will be in equilibrium with
liquid having the composition 91.2 mole % LiF -
9.8 mole % BeF .

Relative volatilities are also given for five rare-
earth trifluorides in a binary mixture of a rare-earth
fluoride and LiF. These measurements were made
at 1000°C and 0.5 mm Hg using mixtures having
rare-earth fluoride concentrations of 2 to 5 mole %.
Except for Prk (and possibly Smk',) the relative
volatilities for the rare-earth fluorides are slightly
lower when BeF, is present.

If the molten salt solutions were ideal, the rela-
tive wolatility of the respective fluorides could be
calculated using Raoult’s law and the vapor pres-
sure of the fluorides and L.iF. The relative vola-
tility of a fluoride with respect to I.iF would be
the ratio of the fluoride vapor pressure to that of
LiF. In Table 20.2, experimentally determined rela-
tive volatilities are compared with calculated rela-
tive volatilities for several fluorides for which vapor-
pressure data are available,

Since the experimental errors in measuring rela-
tive volatilities or vapor pressures could be as large
as the discrepancies between the experimental and
the calculated relative volatilities (except in the
case of SrFQ) shown in Table 20.2, one can infer
241

that the behavior of the rare-earth fluorides, YFS, The relative volatilities of the fluorides of the
and BaF2 in both LiF and the LiF-BeF_ mixture rare earths, yttrium, barium, and sfrontium are low
studied can be approximated by assuming Raoult’s enough to allow adequate removal in a still of
law. The deviation of SiF, and, in the binary sys- simple design. Zirconium removal, however, will
tem, LaF, from this pattern is unexplained. be insignificant.

Table 20.2. Comparison of Experimental Relative Yolatilities with Calculated Yalues

 

 

Calculated Measured

 

Component Relative Relative Volatility aexperimental Relative Volatility iexperimentai
Volatility in Ternary System a’calc ulated in Binary System acafcula ted
NdF 3.0x10"* <3x107* <1 6 x107* 2.0
CeF, 2.5 x 107* 3.3x10"4 1.32 4.2 x 107* 1.68
BaF 1.0 x 10~* 1.1 x107* 0.69
YF, 5.9 x 104 3.3 x107° 0.56
LaF, 4.1 %1077 1.4 x 10”4 3.4 3x107? 7.3

 

SrF 6.8 x 1078 5.0 x 1077 7.4
21.

Relative Volatility Measurement

by the Transpiration Method

F. J. Smith

The transpiration method for obtaining liquid-
vapor equilibrium data is being used (1) to cor-
roborate data obtained by the equilibrium still
technique and (2) to determine relative volatilities
for compounds of interest that have not yet been
studied. The apparatus being used closely re-
sembles that described by Cantor.!

The initial experiments were made with LiF-
BeF2 (86-14 mole %) over the temperature range
920 to 1055°C. Plots of the logarithms of the
apparent vapor pressures of LiF and BeF, (based
on the assumption that only I.iF and BeF, were
present in the vapor phase) vs 1/7T were linear,
although some scatter was evident in the data
points for Bel',. The relative volatility of BeF,
with respect to LiF was 4.0 + 0.2 over the tem-
perature range investigated. This is in reason-
able agreement with a value of 3.8 reported by
Cantor’! for LiF-BeF, (88-12 mole %) at 1000°C.

C. T. Thompson

242

L. M. Ferris

Recent experiments with LiF-BeF  (90-10 mole %)
gave a relative volatility of 4.7 (BeF, with re-
spect to LiF), which is in good agreement with
the average value of 4.71 obtained by Hightower
and McNeese? using an equilibrium still. Experi-
ments with LiF-BeF -CeF, (85.5-9.5-5 mole %)
gave vapor samples in which the concentration of
CeF, was below the limit of detection by spectio-
chemical methods. The relative volatility (Cel,
with respect to LiF) was less than 1 x 1073, in
agreement with the reported® value of 3.33 x 1074,

'Reactor Chem. Div. Ann. Progr. Rept. Dec. 31,
1966, ORNL.-4076.

2M. IZ. Whatley to D. E. Ferguson, private communica-
tion, June 12, 1967.

M. E. Whatley to D. E. Ferguson, private communica-
tion, July 20, 1967,
22. Distillation of MSRE Fuel Carrier Salt

L. E. McNeese W. L. Carter J. R. Hightower

Equipment for demonstration of vacuum distil-
lation using MSRE fuel salt has been built and
assembled in its supporting framework. Installa-
tion of the unit is in progress in Building 3541,
where experiments using nonradioactive salt are
to be carried out. During construction, a series
of radiographic and ultrasonic examinations were
made over areas of the vessels that will be sub-
jected to 850 to 1000°C temperatures, and numerous
dimensional measurements were taken between
points on the still, condenser, and condensate
receiver, Similar data will be obtained after non-
radioactive operation and will be compared with

initial data in evaluating vessel integrity for sub-
sequent radioactive operation. The assembled

unit was stress relieved by an B-hr anneal at

1600°F in an argon atmosphere in order to preclude
dimensional changes during operation due to stresses
arising during fabrication.

Specimens of candidate materials of construc-
tion for vacuum distillation equipment were in-
stalled in the still so that each is exposed to
molten-salt vapor and liquid in the region of high-
est temperature. The test specimens are attached
to the drain tube and include Hastelloy N, grade
AXF-50BG isotropic graphite, Mo-TZM, Haynes

{

ORNL-—DWG 67-—-5B80%A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1967 1968
J A $ o N D J F M Al M J
MOUNT  AND
CHECK  INSTRUMENTATION - 3] 4 o
W ) = A
2% a2
W
HEATERS DELIVERED—{oR] w S S|«
r b O
« & w
a . Z T
PREPARE 3544 W = ©
a. >
| : \\;\ NN
FABRICATE STILL @ T ARSI RN N
.‘2 &l = o
= sl g -
z _ z| G| z
- o = - Lt
- i < - 1
= H oo £l o
“ =45 2 -
S| B 1 BN
~ 9 = <3 . « N
g B § 0 - Bty N
2y N oy &
3 Z 3 5 & D
3 5 2| EZ
= < X g W
o v L N
Do d
Bo|a " & o
o0 o ot
= o oo ==
W g
o Q
o o
o

Fig. 22.1. Schedule of Activities Associated w

243

ith Distillation of MSRE Fue!l Carrier Salt.
alloy No. 25, and an experimental alloy (alloy No.
82) having the composition 18 Mo—0.2 Mn--0.05
Cr—bal Ni.

Activities associated with the system during
the past two months and scheduled activities
through completion of radioactive operation are
shown in Fig. 22,1, Still fabrication has been
completed, and the thermal insulation and electric
heaters have been received. Four faulty heaters
out of a total of 41 were rejected, and replace-
ments have been ordered. Instrumentation work
The first of several 48-liter
salt charges for nonradicactive operation is being

is 90% complete,

prepared by Reactor Chemistry Division personnel.

Experiments using nonradioactive salt will begin

244

in early October and will continue through Feb-
ruary 1968. The still will then be inspected for
radioactive operation and installed at the MSRE
site in the spare cell adjacent to the fuel processing
cell.

The equipment will then be used for distilling
48 liters of radioactive MSRE fuel salt (decayed
90 days) from which uranium has been removed by
fluorination. It was determined that the roof plugs
in the spare cell do not provide adequate biological
shielding for 90-day-old salt; the plugs have been
redesigned and will consist of 30-in,-thick barytes
concrete rather than the present 24-in. ordinary
concrete.
23. Steady-State Fission Product Concentrations and

Heat Generation in an MSBR and Processing Plant

 

 

 

J. S. Watson L. E. McNeese W. L. Carter
Concentrations of fission products and the heat V_/V, = core volume/total fuel volume,
generation associated with their decay are re- T, = processing cycle time for isotopes
quired for design of systems for processing the with atomic number Z (sec),
fuel and fertile streams of an MSBR. A method t = time (sec).
for calculating the steady-state concentration of
fission products in an MSBR will be described, The reactor was assumed to be at steady state
and heat-generation rates in the fuel stream will with respect to all fission products. At steady
be given. Changes in heat-generation rate re- state, sz’A/d; - 0; so the differential equation
sulting from removal of fission products in both reduces to
the reactor and the fuel-stream processing system
will be shown. N TKPY +Nz,_.q-1 (’AO-Z,A-I(VC/Vt)JrNZ-I.A )\Z-I,A
Several simplifying assumptions were made in Z.4 A, a0, SV V)T ’
calculating the concentration of fission products
in the fuel stream of an MSBR. The differential which defines the steady-state concentration of
equation used to define the concentration of a each isotope. This equation was solved for each
given isotope was isotope using an 1BM 7090. The calculations used
2357 yields given by Blomeke and Todd;' yields
v, . KPY + N 4 V. for 233U were obtained by distributing ‘‘mass
7 P z,41% 2,40 v yields” reported by England? among the isotopes
f in each mass chain so that each iscotope had the
+ N Y N same fraction of the mass yield as reportied for
z-1,4 2-1.4 Zoa 2,4 235U, Decay chains were simplified and approxi-
. y s implifie PP
v N mated by ‘‘straight’’ chains considering only beta
- c;')c:rz'A 7‘3 NZ,A + f'A , decay. Where isomeric transitions or other
t z complications were noted, personal judgment was
where used to approximate the real process with straight
beta decay.
N = number of atoms of an isotope with

Z,A
atomic number Z and mass number A,

P = power (w),

 

1]. 0. Blomeke and Mary F. Todd, Uraniumn-235
Y = primary yield of isotope of atomic Fission=-Product Production as a Function of Thermal

. Neutron Flux, Irradiation Time, and Decay Time. 1.
number Z and mass number A (atoms/ Atomic Concentrations and Gross Totals, ORNL-2127

fission), (part 1), vol. 1.

P = tlux (neutrons sec™ ! Cm‘z), 2. R England, Time-Dependent Fission-Product
o _ s : I : Themal! andResonance Absorption Cross Sections
¢ A7 cross section of isotope with aton;lc (Data Revisions and Calculational Extensions), WAPD-
number Z and mass number 4 (cm*), TM-333, addendum No. 1.

245
Effective ‘‘one group’’ cross sections were cal-
£ P

culated from results from OPTIMERC (a reactor
analysis code) for the MSBR reference design,
The fuel salt was assumed to be completely mixed,
and capture terms were reduced by a factor equal
to the ratio of the core volume to the total fuel
salt volume.

Heat-generation rates, both at the instant of
removal from the reactor and after various decay
periods, were calculated from the steady-state
fission product concentrations in the fuel using
CALDRON, a fission product heat-generation and
decay code,

Heat-generation rates obtained with 233U vields
were in good agreement with rates calculated by
a modified (to treat a steady-state reactor with
continuous processing) version of PHOBE, a code
written by E. D. Arnold and based on experimental
data for decay of fission products, Agreement with
a reputable code indicates that the simplifications
in the present calculations were reasonable.

In the MSBR, some fission products will be re-
moved from the fuel by the gas purge or by plating
on solid surfaces, and other fission products will
be removed in processing (fluorination) before the

 

246

salt reaches the still. Removal of fission products
by plating and gas stripping was taken into account
by assigning appropriate residence times, T, for
volatile or noble elements. After withdrawal from
the reactor, the fuel salt was assumed to be re-
tained in a hold tank for 12 hr, after which specified
fractions of elements having volatile fluorides were
removed. The remaining fission products were
then allowed to decay for an additional 12 hr be-
fore entering the still.

Figure 23.1 shows fuel salt heat-generation
rates calculated for various times after removal
These calculations were made

2220 Mw (thermal),

2

from the reactor.
with MSBR design conditions:
3.7 % 10** neutrons sec™! cm™?, core volume of
9400 liters, fuel salt volume of 25,400 liters, and
4.5 % 10° sec (52 days) processing cycle. The
uppermost curve represents no fission product re-
moval in the reactor or in the fluorinator. The
next lower solid curve represents Kr and Xe re-
moval from the reactor with a 30-sec residence
time, and the lowest solid curve represents re-
moval of Kr and Xe (30-sec residence) along with
removal of Mo, Tc, Ru, Rh, Pd, Ag, In, Nb, Te,
and Se with a mean residence time of 1.8 x 10°

 

   

 

 

 

  

 

 

 

 

 

 

T 1 [ ] ‘ i T [ oo | \
a ’ | 0 i | ‘ o | { ]
0" | o 4#5“ = NO REMOVAL IN REACTOR VOLATILE - . - |l : l| P i s
1 | r == FLUCRIDES REMOVED IN PROCESSING -1 707 & | } Lo by
i A 1 1 o I - . ERE I . .. tr . '
el A R T”]L,L ’ T ! .} L T
T _ .~ Kr—Xe REMOVED IN REACTOR | } o | | o
5 | N =k » | - o
10 e o I : C N o I
==} 1 0l TSad T Kr-Xe AND PLATING ELEMENTS . |
- e , | gf: REMOVED IN REACTOR Lo L
= I Kr-Xe REMOVED IN REACTOR |71 | = RN o S |
= | VOLATILE FLUCRIDES REMOVED| . | e b
o "7 IN PROCESSING | l | % MO REMOVAL ST T
- 17 B Kr-Xe AND PLATING ELEMENTS REMOVEGC | > i Lol Tl
= L ) IN REACTOR VOLATILE FLUORIDES REMOVED | i [l i Lo
= - ;== IN PROCESSING - ' | S
= { i ' o ‘ i | ! ‘ :
5ot D P - 4 Cobee Lo
© 1 Dl oo [ T b b
Z ; $o | o - ' i :
5 Gl e
£ ol e
" FUEL = 23yF, ' T
. POWER PER REACTOR = 556 Mw (thermal) == = |
-~ FUEL VOLUME (iotal} = 224 #2 L
- PROCESSING CYCLE TIME = 52.08 days o :
P oo L | i
1 ! | | i ; 3 ‘ i i i
N Lo Ly
5 10" 2 5 10° 2

   
  

 

 

 
 

 

      

 

 

 

TIME AFTER REMCVAL FROM REACTOR {days)

Fission Product Heat Generation Rate in M$BR Fuel After Removal from the Reactor.
247

sec (50 hr). These latter residence times are
rough estimates, and as better information becomes
available from figsion product behavior in the
MSRE, more reliable estimates of the residence
times for these elements will be possible. This
lowest solid curve in Fig. 23.1 is shown only to
supggest the general range of conditions under
which the fuel processing plant may operate,
The dashed curves in Fig. 23.1 were obtained
from the same reactor calculations, but some fis-
sion products were assumed to be removed in the
fluorinator. The elements Xe, Kr, Br, I, Mo, Tc,
Te, and Se were considered to be completely re-
moved, and 15% of the Ru, Rh, Nb, and Sb were
removed. After fluorination, however, these
elements may ‘‘grow’’ back into the system.

HEAT GENERATION IN A MOLTEN-SALT STILL

Buildup of heat generation in a molten-salt
distillation system was calculated using the heat-
generation data given in Fig. 23.1. The reactor
and that part of the processing system prior to the
still were assumed to be at steady state, although
the transient associated with buildup of fission
products in the still was considered. TFuel salt
containing fission products remaining after
fluorination was fed to a distillation system, and
complete retention of fission products was as-
sumed. The fuel salt was assumed to have been
held up 24 hr in processing prior to entering the
still (12 hr before fluorination and 12 hr after),

The heat-generation data from Fig. 23.1 was
fitted by the method of least squares to the rela-
tion

15 Kt
Ho=Y 4de 7,

i=1

where

H(t) == heat-generation rate at time ¢ (Btu
hre ! £t d),
A k, = constants,
t - time after salt leaves fluorinator (days).

The rate of heat generation in the still is then
given as

A —k. t
Q(t)r:Ff H(x)dY—Fz (I—-e 1),

0 i=1 i

where

1) = heat generation in still after operating
for a time ¢ (Btu/hr),
F - fuel salt processing rate (ft3/day),
t = length of time still has operated (days).

Calculated heat-generation rates in the still are
shown in Fig. 23.2 for a processing rate of 15
ft*/day for several assumed removal efficiencies
(the same assumptions noted for Fig. 23.1) in
processing steps prior to the still. The still
system will be near steady state in two to three
years, and heat generation rates as high as 3.0
Mw can be expected. Removal of fission products
by gas stripping, plating on metal surfaces, and
formation of velatile fluorides during fluorination
will reduce this rate to about 2.2 Mw,

DRNL~DWG 67-97384

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10" & Er T T T T T ST
P R e P T
L~ REACTOR POWER 2224 Ww (thermal) | |T— AT “:ijiiiiFi,“ T
.- PROCESSING CYCLE 8208 Davs 1]l L ’ :
L Al * 11
| Cbeemmmemen L LD
"""" T 77T '::““ e o e T |
3 | | |
2,0 e '
L b D HO REMOVAL 1N REACTOR SR
= — g I /}4,:,6{ - VOLATILE FLUCRIDES REMOVED IN PROCESSING - 1
@ T JTiT RN N AN
Z o LT T K - Xe REMOVED I REACTOR LT
g ,,,,;fi‘/,;Fi [H - xo» ATHLE FI UOR\D“S RCWOVED IN PROCESSING....\....
P VA AT I | \ b Lt !
o A T Ir(r -Xe AND PL :\rmc H FNFN[: REMOVED IN REACTOR & ‘
& . 7 VOIIATILIFJF\ L(‘RIEF) REMOVED 11 PRO(FS"S NG : |
by J L L Lol 1
i o REMOVALE AT SR ST
- ) ST LT CUTTRITIY
| . I i } ,,,,,,,,,,,,,, 1144 1 e dibo J\ L
\
, L] L] Ll
—— ‘ ‘ 4
: 1 il
10— oL L L v it oLl S
4of ! 13 10" 10°

THIE (days)

Fig. 23.2. Fission Product Decay Heat in MSBR 5till and Accumulation Tank.
24. Reductive Extraction of Rare Earths from Fuel Salt

L. M. Ferris C.

One alternative to the distillation process for
decontaminating MSBR fuel salt involves reductive
extraction of the rare earths (and other fission
products) after the uranium has been recovered
from the salt by fluorination.' ™ The reductant
currently being considered is ‘Li dissolved in
molten metals such as bismuth or lead which are
immiscible with the salt. Studies in the Reactor
Chemistry Division''?'* have defined the equi-
librium distribution of several rare earths between
Li-Bi solutions and LiF-BeF, (66-34 mole %) at
600°C. The results of these studies indicate that
the rare earths can be effectively extracted in
such a system.

The apparent stoichiometry of the extraction
reaction is such that, under the experimental con-
ditions employed, the lithium concentration in the
metal phase should not change detectably, even
if the rare earth were quantitatively extracted.
However, in these, and similar, experiments,
usually 50 to 75% of the lithium originally present
in the Li-Bi solution apparently was consumed.?
The mechanism by which lithium is ‘‘lost’ in
these systems has not yet been elucidated.

In order to properly evaluate the practicality of
the reductive extraction process, the behavior of
lithium in these systems must be clearly defined.
Knowledge of its behavior is especially important
in designing a multistage extraction system, be-
cause the extent to which the rare earths are ex-

 

'rR. B. Briggs, MSIR Program Semiann. Progr. Rept.
Aug. 31, 1966, ORNL-4037.

M. W. Rosenthal, MSR Program Semiann. Progr.
Rept. Feb. 28, 1967, ORNL-4110,

’p. E. Ferguson, Chem. Technol. Div. Ann. Progr.
Rept. May 31, 1966, ORNL.-3945,

4W. R. Grimes, Reactor Chem, Div, Ann, Progr.
Rept. Dec. 31, 1966, ORNL-4076.

i

. Schilling

248

J. F. Land

tracted (at equilibrium) is directly related to the
lithium concentration in the metal phase,!-?*
Consequently, mote experiments have been con-
ducted, with the primary objective being the de-
termination of the cause of the apparent lithium
loss. As discussed below, these experiments did
not produce a definite answer to the question,
and more study will be required. In each experi-
ment a rare-earth fluoride, usually Equ, was
added to the salt to provide a basis for compari-
son with the results of earlier studies and to allow
us to obtain preliminary information on the effect
of temperature on the distribution of europium be-
tween the two phases.

Experimentally, solutions of lithium in bismuth
(usually about 7 at. % Li) and of EuF, in LiF-
BeF (66-34 mole %; mole fraction of EuF |,
about 10=%) were prepared at 600°C in separate
mild steel or graphite crucibles, using pure argon
as a blanket gas. Then, the salt was transferred
to the crucible containing the Li-Bi alloy, and the
system was equilibrated at high temperature in an
argon atmosphere, Filtered samples (stainless
steel samplers) of both phases were removed
periodically for analysis.

No detectable ““loss’’ of lithium occurred during
preparation of the Li-Bi alloys in either mild
steel or graphite crucibles, When the system was
not agitated (by sparging with argon), the time
required to achieve equilibrium was generally
about 24 hr,
in bismuth when the system is quiescent has been
observed by others.! The rate-controlling step
may be the dissolution of Li Bi, a high-melting,
insoluble compound that probably forms rapidly
when mixtures of the two metals are heated.

After about 24 hr at 600°C, the lithium concentra-
tion in the solution, as determined by chemical

The low rate of dissolution of lithium
analysis of filtered samples and thermal analysis
of the system, usually reached the expected
value, When the system was sparged with argon,
equilibrium was reached in 1 to 2 hr or less.

As was the case in other studies,’ a significant
(30 to 70%) decrease in the lithium concentration
in the metal phase occurred during equilibration
of salt and Li-Bi solutions at 500 to 700°C.
Several possible causes of this phenomenon were
considered: (1) dissolution of lithium in the salt,
(2) reduction of BeF  from the salt by lithium with
formation of LiF and berylium metal (which is
insoluble in both the salt and bismuth), and (3)
reaction of the lithium with water that was in-
advertently admitted to the system. Neither of
the first two mechanisms seems likely, based on
the experimental evidence. If either lithium or
beryllium metal were present in the salt in the
amounts expected from the lithium ‘‘loss,”” dis-
solution of the salt in hydrochloric acid should
have resulted in the evolution of more than 10
cc (STP) of hydrogen per gram of salt. However,
it was found that samples of both filtered and un-
filtered salt generally gave less than 0.2 cc of
H, per gram. This finding, along with thermo-

249

dynamic considerations,? appears to rule out
either of these mechanisms.

The inadvertent admittance of water into the
system (as a contaminant in the salt and/or
blanket gas, or during sampling) seems to be a
more reasonable explanation for the consumption
of lithium. Although this hypothesis has not yet
been confirmed, several observations have heen
made which support this mechanism, The salt
after equilibration is invariably permeated with an
ingoluble material which tends to concentrate at
the salt-metal interface. The presence of this
phase does not appear to be related to the presence
of a rare earth, and, on the basis of chemical
analysis, is not the result of corrosion of the
crucible or samplers, High oxygen concentrations
(about 0.5%) have been detected in salt samples,
and, in a few instances, the presence of BeO in
the system has been confirmed, These results,
if caused by the presence of water, are consistent
with a mechanism in which the water reacts first
with BeF , to form insoluble BeO and gaseous HF
which, in turn, reacts with lithium to form LiF
that dissolves in the salt, Accordingly, the LiF
concentration in the salt would increase and the

Table 24.1. Distribution of Europium Between LiF-BeF, (66-34 Mole %) and Lithium-Bismuth Solutions

 

L.ithium

Approximate Amount

 

Concentration in Temperature of Europium in
Experiment Sample Metal Phase o Metal Phase D
(mole %) (7o)
CES75 1 0,85 608 525 1.20
CES66 3 2.6° 602 81° 5.4°
CES66 2 3.4% 583 93P 17.1%
JFL64 3 3.97 500 92°¢ 14.8°
JFL64 1 4.0P 605 84° 6.7°
JFL64 2 4.0° 700 88¢ 9.1¢
JFL64 1 4.84 605 84¢ 6.7
CES66 1 5,00 583 kL 16.8”
JFL64 2 5.19 700 88° 9.1°¢

 

mole fraction of Eu in metal phase
“Defined as D =

 

mole fraction of Eu in salt phase
PEmission spectrographic analyses.
“Neutron activation analyses.

A lame photometric analyses.
260

BeF concentration would decrease. The changes
in salt composition calculated from the amount of
lithium consumed were too small to be detected
by chemical analysis; however, in some instances
x-ray diffraction analyses of salt samples that
had been cooled to room temperature revealed the
presence of LLiF, a phase that would be expected
if the LiF/BeF | mole ratio in the salt were
higher than its original value of about 2. Ob-
viously, more work, conducted under very care-
fully controlled conditions, will be required to
confirm or refute the tentative hypothesis that
water is the cause of the apparent loss of lithium
in reductive extraction experiments.

Despite significant changes in the lithium con-
centration in the metal phase during an experiment,
it was possible to determine the distribution of
europium between the two phases at various tem-
peratures. This distribution is expressed as a
ratio, D, defined?* as

mole fraction of Eu in metal phase

 

mole fraction of Eul", in salt phase '

The values of D obtained in this study are given
in Table 24.1 and are compared in Fig. 24.1 with
those obtained by Shaffer, Moulton, et al.? at
600°C. Agreement between the two sets of data
is reasonably good. It also appears that tempera-
ture has no marked effect on the equilibrium dis-
tribution of europium between the two phases.

An emission-spectrographic or a neutron-activa-
tion method was used to analyze for europium in
both the salt and metal phases. The data reported
are from experiments in which the europium balance
was 90 to 110%. FEach sample of the metal phase

ORNL-DWG 67-11848
20

TEMPERATURE

© (°c)
+ 500

583

c02

5 A 605
608

700

maie fraction of kuin merai phase
moig frection of Eu in sait phase

 

 

05
05 ! 2 5 10

LITHIUM CONCENTRATION IN METAL
PHASE (at %)

Fig. 24.1. Distribution of Europium Between LiF-
BeF2 {66-34 male %) and Li-Bi Solutions. Data points,
this study; solid line, data of Shaffer, Moulton, et al.
(see Reactor Chem. Div. Ann. Progr. Rept. Dec. 31,
1966, ORNL-4076).

was analyzed for lithium by both an emission
spectrographic and a flame photometric method.

In some instances, the values obtained were
markedly different; this is readily apparent in the
data obtained from experiment JFI1.64 (Table 24.1).
25. Modifications to MSRE Fuel Processing Facility

for Short Decay Cycle

R. B.

In order to keep reactor downtime to a minimum
during the next planned shutdown, it has been pro-
posed that the flush and fuel salts be processed
with a minimum decay time.

Problems anticipated because of the short decay
are due to the much larger amount of '3[ and to
the radiation level at the UF _absorbers because
of coabsorbed ??Mo. The processing schedule,
with pertinent radiation information, is given in
Table 25.1. Design of modifications to the MSRE
fuel processing facility is in progress to permit
safe operation at these increased activity levels.
These consist of:

Lindauer

1. A deep-bed backup charcoal trap downstream
of the existing charcoal canisters. These
traps will be tested under simulated operating
conditions to demonstrate at least 99.998% re-
moval of 1*11,

2. An activated alumipa trap downstream of the
SO_ system for excess fluorine disposal.
This will prevent fluorine from releasing iodine
on the charcoal beds in case of a malfunction
of the SO2 system.

3. Shielding and revised handling procedures for
removal of the loaded UF6 absorbers from the
absorber cubicle,

Table 25.1. MSRE Processing Schedule and Pertinent Radiation Levels

 

 

 

ISII
Uranium
Decay Required Removal to Be Radiation Dose Rate
Operation Time Curies in Charcoal Volatilized 2 ft from f\bsorber
(days) in Salt Traps® (k) (millirems /hr)
(%)
Hz—HF treatment 6 600 99,95
of flush salt
Fluorination of 18 210 99.88 7 620
flush salt
HQ-HF treatment 28 23,000 99.990
of fuel salt
Fluorination of 35 12,000 99,997 230 170

fuel salt

 

AFor 0.3 curie release.

251
26. Preparation of 2*3UF,-’LiF Fue! Concentrate

for the MSRE

J. M. Chandler

Plans are now being made for refueling and op-
erating the MSRE with 273U fuel early in 1968,;
approximately 40 kg of ***U as ?**UF -"LiF (27
and 73 mole %) eutectic salt will be required.
This material must be prepared in a shielded fa-
cility because of the high 22?U content (240 ppm)
of the ??3U. The following sections describe the
process, equipment, and status of the fuel prep-
aration process.

PROCESS

The process for preparing eutectic-UF -LiF (27
and 73 mole %) salt for the original MSRE charge
started with UF, and LiF.' Since **°U is avail-
able only as the oxide, the process was modified.
The UO3 is charged to a nickel-lined vessel, and
LiF is added on top of it. The charge is heated,
under helium, to 900°C to melt the LiF, which
wets the oxide and thermally decomposes the UO,
to an average composition of UO, - The charge
is treated with H, to reduce the uo, | to UO2 and
then hydrofluorinated with HZ-HF (approximately
10:1 mole ratio) to convert the UO2 to UF4. The
progress of the reaction of UO2 to form UF, is
followed by the changes in the liquidus tempera-
ture of the binary mixture as the melting point of
the charge decreases from the initial melting point
of 845°C for the LiF to 490°C for the UF -LiF
(27 and 73 mole %) eutectic product. Hydrogen
fluoride utilization and water production are also
indicative of the progress of the reaction. After

 

'MSR Program Semiann. Progr. Rept. Feb. 28, 1965,
ORNL-3812, pp. 149-50.

252

E. L. Nicholson

a final H, sparge to reduce impurities, such as
iron and nickel fluorides, to metals and to sweep
out dissolved HF, the salt is sampled, filtered,
and stored in a product container. This process
has been satisfactorily tested on a small scale,
and the results indicate that the processing time
will be about 15 days/batch.

Tentative plans are to make three 12-kg 233U
batches of salt, which will be used for the major
additions to the barren salt in the MSRE, and one
7-kg 233U batch, which will be loaded into 60 en-
riching capsules.

EQUIPMENT AND OPERATIONS

The schematic equipment flowsheet for salt
preparation is shown in Fig. 26.1. The cans of
UO,, each containing about 500 g of uranium, will
be transported in a shielded cask from the *33U
storage facility in Building 3019 to the TURF
building (7930) and introduced into cell G. All
operations in cell G will be done remotely. By the
use of manipulators, the cans will be loaded into
the alpha-sealed can opener and the dumpet box,
which discharges the oxide into the heated nickel-
lined 8-in.-diam by 3-ft-high reaction vessel and
routes the empty cans to storage for subsequent
removal from the cell. The "LiF is charged to
the reaction vessel via the dumper. Process
gases are supplied from outside the cell, and
instrumentation is provided to control gas flows
and processing temperatures. Waste gases are
filtered to remove entrained 233U and scrubbed
with KOH to remove unreacted HF before dis-
posal to the radioactive off-gas system. After
sampling, the product is transferred via a heated
 

253

ORNL-DWG 67-- 3644 A

 

     

 

 

 

 

GAS
KOk ANALYSIES
e . CAS SAFETY Ta ]
OFF=GAS =1 goqopck TRAR [ OHH v

X ES’EUOS 1l

[ ALUMINUM LiF
e WASTE Y et CANS CHARGING

GAS
ANALYSIS

HYDROGEN
FLUORIDE ™7

HYDROGEN —i>
233 -
UF, - LiF

PRODUCT STORAGE
G-in. DIAM. -----

 

 

 

   

CAN DPENER
AN DUMPER

SMETY CAN
STORAGE

    
  
 

FUHTER }

SALT

  
 
  
 

 

    
 
 

GAS
FLOW --

 

RIEAGCTION VESSEL
8-mn. DIANM.

Fig. 26.1. MSRE 233U Fuel Salt Preparation System.

nickel line through a sintered metal filter into
heated 4-in.-diam by 3-ft-high nickel product
storage vessels. The capsule-filling operation
is not illustrated but will consist in transferring
salt from a product vessel to an array of 60 en-
riching capsules in a furnace. Product vessels
and enriching capsules will be stored in cell G
until needed at the MSRE.

STATUS

Engineering design is estimated to be 95% com-
pleted, and fabrication of the equipment is 85%

completed. The project has been delayed because
the TURF building is not completed; consequently,
the cost-plus-fixed-fee (CPFF) contractor and the
laboratory forces have been delayed in starting the
installation of equipment. Assuming that the build-
ing will be available on September 15, 1967, and
that the CPFF contractor and ORNL forces each
require about one month for their portions of the
job, the full-scale cold run should start December
1, 1967; hot runs should start January 20, 1968,
and be completed in mid-April 1968,
255

OAK RIDGE NATIONAL LABORATORY
MOLTEN-SALT REACTOR PROGRAM

AUGUST 33, 1967

 

M. W. Rasenthal, Director R
R. B. Briggs, Associate Director D
P. R. Kasten, Associate Director R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSBR DESIGN STUDIES COMPONENTS & SYSTEMS DEVELOFMENT‘I PHYSICS MATERIALS FUEL PROCESSING DEVELOPMENT CHEMISTRY MSRE OPERATIONS
E. 5. Bettis R Dunlap Scott R A. M. Perry* R J. R, Weir M&C M. E. Whatley cT W. R. Grimes* RC P. N. Haubenreich R
H. E. McCay M&C E. L. Nicholson T
C. E. Bettis* GE MSRE PHYSICS HASTELLOY N STUDIES M5BR PROCESSING
O, W. Burke™ |1&C
W. L. Carter’ CcT B. E. Prince R D. A, Canenice* MAC J. 0. Blomeke* cT
F.H. Clark* 18C PUMP DEVELOPMENT A G. Cepolina” MAC W. L. Carter €1 REACTOR CHEMISTRY COORDINATION NUCLEAR AND MECHANICAL ANALYSIS
C. W. Collins R MSBR ANALYSIS R. E. Gehlbach* M&C L. M. Ferris cT
W. K. Crowley” £ A. G. Grindeli R O, L. Smith R A P. Litman* MBC 8. A. Hannaford cT H. F. McDuffie* RC C. H. Gobbard** R J. R. Engel R
5. J. Ditte* 1L P. G. Smith R "n T Ken R T. 5. Lundy* MAC J. R. Hightower LT R. E. Thoma** RC €. H. Gabbard™* R
W. P. Eatherly uce L. V. Wilsen R T C. E. Sessions MsC 1. C. Mailen T =. F. Blankenship RC J. A Watts R
A, G. Grindell R D. D. Owens R MSBR EXPERIMENTAL PHYSICS G. M. Slaughter* M&C L. E. M.cNeese T
R. B. Lindaver* cT J. P. Nichols cT MSRE ON-SITE CHEMISTRY
GRAPHITE STUDIES E senilli e
o Liewellyn” e B b g " ey T R. E. Thoma'* RC OPERATION
R. L. Moore* 18C H. Beutler* MLC F. J. Smitl ~ 5 noma
H. A. Netms® CE CONSULTANT W. H, Caok MAC J. 5. Watson g 5. 5. Kirsfis RC R. H. Guymon kR INSTRUMENTATION AND COMTROLS
L. C. Qokes™ 14C COMPONENTS & SYSTEMS . . €. R. Kenned M&C J. Beams
T w Pkl cE T W. Keriin uT ¥, C. Robimeon® MaC V.L. Fowler T PHYSICAL AND INORGANIC CHEMISTRY 2 Crowley R — L. C. Ockes 18C
- - * . . v >
R. C. Robertson® R Dunlap Scott 1. F. Lend cT C.F. Baes, Jrt RC T Huds:n R g K. Adams. l&g
J. R. Rose GE GRAPHITE-METAL JOINING R. D _Fr’ayne ET C: J.- Bar!a’r\“ ’ RC A, I. Krakoviak R C- 2 :u::n ~ ;ic
W. C. Stoddar™ %EC COMPONENTS W. Castleman* ML €. T Thompsan T J. Breunstein® RC R. R. Minue R L R:df::“: &G
d' ? Tallackson . R. B. Gatloher R K. V. Cook* MAC MSRE PROCESSING G. D. Brunton* RC . Richardson R R. W. Tucker JAC
- Terry 8 R J Kedl R 3. P. Hommond* M&C 5. Cantor” RC R. Stefty R L. E. Basler 1%L
E' L W!:Hs . R T. H. Mouney RC W. J. Verner* M C R. B. Lindauer** cr J. H. Shatier* RC T. Amwine R J. Campbell* 1ac
Y. Wilson R AN Smith R e 7 R. A. Strehlow* RC W. H. Duckworth R 5. E. Ellis* 1&C
W. C. George R 2 E Comes R TECHNICAL SUPPORY PREPARATION OF 2*3UF  7LiF FUEL FOR MSRE R. L. Thomo** RC J. D. Emch R e E Rickwood* 18
H. M. Poly § S . . i - E
€. A. Gifford R M. D. Alien* MaC E. L. Nicholson T G e oo Re T o ;
R. W. Cunningham* MaC J. M. Chandler cT - L. Bamberger RC C- C. Hurtt R
REMOTE MAINTENANCE 1. C. Feimer MAC 5. Mann cr FoiQoss RC 1. C. Jor don R
- —~ - . nedmoern - T
R. Blumberg R -Jr #. Goer 1 AL A . Rolm t‘ L. Q. Gilpatrick* RC H. J. Millar R
F. 0. Horvey MAC W. F. Schotfer < 2 CHEMICAL PROCESSING
J. R. Shugart R D. M. Moulton RC L. E. Penton N
. R. Shug, V. G. Lane M&C E. L. Youngblood cT —
E. J. Lawrence MAC J. D. Re¢man'* RC W. E. Ramsey R R. B. Lindeuer** cT
L K. A. Remberger™ RC R. Smith, Jr. R
CONSULTANT E. H. Lee* M&C DR S N RC J. L. Stepp R
- . - JeQrs - .
F. N. Pecbies uT L C- Monley MaC H. H. Stane* RC 5. R. West R
N W. J. Mason* M&C E. W
B. McNabb MEC ¥W. K. R. Finnell RC J. E. Wolfe R
RA P‘Ldge"* VEC 3. F. firch RC S. J. Davis R
N. 0. Pleasant MEC _C E. R°!’°"s EE
R S Pultiam® MAC W. P. Teichert
R. G, Shooster* mé&C
W. H. Smith, Jr.* M&C
L. D Stack* M&C DESIGN LIAISON
G. D. Stohler™ M&C C. K. McGlathlan R
0 R Tinch MaC IRRADIATION EXPERIMENTS R Stoter .
E. M. Thomas M&C E. G. Bohlmann* RC F. J. Underwood GE
1. i. Woodhouse* M&C
E. L. Compere RC
H. £. Savage RC
J. M. Baker RC —
J. R. Hort RC MAINTENANCE
B. L. Jahnsan RC B. H. Webster R o
J. W. Myers RC AE Gill
R. M. Waller RC - F. Gillen R
L. P. Pugh R
ANALY TICAL CHEMISTRY DEVELOPMERT
AC  ANALYTICAL CHEMISTRY DIVISION J. C. White* AC
CT CHEMICAL TECHNOLOGY DIVISION .
D DIRECTORS DIVISION
& DEVELOPMENT
GE GENERAL ENGINEERING DIVISION RESEARCH M
1£C INSTRUMENTATION AND COMTROLSS DIVISION A, S. Meyer AC
M&C METALS AND CERAMICS DIVISION R. F. Apple AC
R REACTOR DIVISION C. M. Boyd AC
RC REACTOR CHEMISTRY DIVISION J. M. Dote AC
UCC UNION CARBIDE CORP H. W. Jenkins uT
UT UNIVERSITY OF TENNESSEE G. Maomantav* uT
* PART TIME ON MSRP D. L. Murlmrlg" fig
** DUAL CAPACITY ; R Maller o
_ L. Whiting
1. P. Young" AC
ANALYSES
L. T. Corbin* AC
G. Goldberg” AC
F. K. Heacker AC
C. E. Lamb* AC
P. F. Thomason* AC

 

 
0 ONO LA LN

ErELMOCOE0VOOONEIAMICONMPAIS-TDOMOITVINN-NATAIOX

W4T >rPr==X

. K. Adams

. M. Adamson
. G. Affel

. G. Alexander
. F. Apple

. F. Baes

M. Baker
J. Ball

. J. Barton

F. Bauman

E. Beadll
Bender

. Bettis
Bettis
Billington

. Blanco

. Blankeriship
. Blomeke

OCmmwem

. Blumberg

L.. Boch

. G. Bohlmann
. J. Borkowski
. E. Boyd

Braunstein
Bredig
Breeding
. Briggs

. Bronstein
. Browning
. Bruce

. Brunton

. Burger
Canonico
Cantor

PITTAAMAUET = >

. W. Cardwell
. L. Carter
. I. Cathers

. Chandler

. Collins

. Compere
Cenlin

. Cook

. Corbin

. Cottrell

INTERNAL DISTRIBUTION

257

59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
39.
90.
91.
92.
93.
94,
95.
96.
97.
98.
99.
160.
101.
102.

G.
S.
. L. Crowley

. L. Culler

. M. Ddle

. G. Davis

. W. Davis

. J. DeBakker

ME-PIMIONDIPRARPETINRNETO-IPIUEMEZEOP ZAL P ETEME

OXATMNOPZLOIPTIOAMOBIITLT>VEIMATAAIT>Y MO

ORNL-4191
UC-80 — Reactor Technology

A. Cristy
J. Cromer (K-25)

H. DeVan

J. Ditto

. Donnelly
. Dunwoody
Dworkin

. Dyslin

. Eatherly

. Edlund (K-25)
Engel

. Epler

. Ergen

. Ferguson
. Ferris

. Fraas

. Friedman
Frye, Jr.

. Gabbard
Gall

. Gallaher

. Gilliland
. Goeller
Grimes

. Grindell

. Guymon

. Hammond
. Hannatord
. Harley

. Harman
Harrill

. Haubenreich
. Heddleson
. Herndon

. Hibbs {Y-12)
. Hightower
. Hitl

. Hise
103.
104.
105.
106.
107.
108.
109,
110.
111-115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131
132.
133.
134.
135.
136.
137.
138.
139.
140.
141,
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.

—
%
———

2OmMP>r0O0;

IqufiénUmmIEZOp%?fifim@%fi&?pmp;fipg;ppgIa>mm<i

258

. W. Hotfman

. D. Holt

. P. Holz

. W. Horton

. S. Householder
. L. Hudson

. Inouye

H. Jordan
R. Kasten
Kedi

. Kelley

. Kelly

. Kennedy
. Kerlin

. Kerr

. Kimbali
Kirslis

. Knowles
Krakoviak
. Krewson
. L.amb

. L.ane

. Larson

. Lawrence
. Lincoln

. Lindauer
. Litman

. Liverman

DT g eemPmMmECLCL AT A

. Livingston

. H. Llewellyn
. I. Lundin

. Lyon

. MacPherson

. MacPherson

. Maienschein

. Manning

Martin

. Martin

. Mathews

. Mauney

cClain

. McClung

. McCoy

. McDuffie

. McGlothlan
McHargue

. McNeese
Meyer

. Miller

. Mills

. Minue

ZTMITOrOmMoOZ

mrrOUmMmS-SATME

158.
159.
160.
161.
162.
163.
164.
165.
166.
167 - 168.
169.
170.
171.
172.
173.
174,
175.
176.
177.
178.
179.
180.
181.
182-316.
317,
318.
319.
320.
321.
322.
323.
324,
325.
326.
327.
328.
329.
330.
331.
332.
333.
334,
335.
336.
337.
338.
339.
340.
341.
342.
343.

ON=E»IZTIEAZTL-OWIAPI I AEIME-L-ARE

C
C
W
C
W
. F. Schaffer
E
al
L
E
E
H

COXTIEORTEPDOORPOEM OIS,

R. Mixon

. L. Moore
. Z. Morgan

C. Moyers

. A. Nelms
. P. Nichols
. L. Nicholson
, C. Oakes
. R. Osbhorn
. B. Parker
. F. Parsly
. Patriarca
. R. Payne
. M. Perry

. W. Pickel
. B. Piper

. E. Prince
. l.. Ragon

L. Redford

. Richardson
. C. Robertson

. Robinson
. Roller

, Rosenthal
. Savage

. Savolainen

. Schilling

. Seagren
. Sessions
. Shatfer
D. Shipley
J. Skinner
M. Slaughter
N. Smith

J. Smith

P. Smith

L. Smith

G. Smith

1. Snell

F. Spencer
Spiewak

C. Steffy

€. Stevenson
C. Stoddart
H. Stone

A. Steehlow
A. Sundberg
R

. Tallackson
413.
414.
415.
416.
417,
418.
419,
420.
421--422.
423.
424,
425.
426,
427.
428.
429.
430,
431,
432.
433.
434.
435.
436.
437.
438.
439.
440,
441.
442443,

259

344, E. H. Taylor 360. W. J. Werner

345, W. Terry 3461, K. W. West

346. R. E. Thoma 362. M. E. Whatley

347, G. M. Tolson 363. J. C. White

348. L. M. Toth 364. G. C. Williams

349. D. B. Trauger 365. L. V. Wilson

350, R. W. Tucker 366. G. J. Young

351. W. C. Ulrich 367. J. P. Young

352, D. C. Watkin 368. F. C. Zapp

353, G. M. Watson 369. Biology Library

354, J. S. Watson 370-373. ORNL. — Y-12 Technical Library
355, H. L. Watts Document Reference Section
356. C. F. Weaver 374-376. Central Research Library

357. B. H. Webster 377~411. Laboratory Records Department
358. A. M. Weinberg 412, Laboratory Records, ORNL R.C.
359. J. R, Weir .

EXTERNAL DISTRIBUTION

W. O. Allen, Atomics International, P.0O. Box 309, Canoga Park, California 91304

J. G. Asquith, Afomics Infernational, P.O. Box 309, Canoga Park, California 91304

J. C. Bowman, Union Carbide Technical Center, 12900 Snow Road, Parma, Ohio 44130

G. D. Brady, Materials Systems Division, UCC, Kokomo, Indiana 46901

J. H. Brannan, Corbon Products Division, 270 Park Avenue, New York, New York 10017
W. S. Butler, Dow Chemical Company, Ludington, Michigan

R. A. Charpie, UCC Electronics Division, 270 Park Avenue, New York, New York 10017
Paul Cohen, Westinghouse Electric Corp., P.O. Box 355, Pittsburgh, Pennsylvania 15230
D. F.
J W
M. W.
C. B.
D. A,
H. L.

Cope, Atomic Energy Commission, RDT Site Office (ORNL)

Crawford, Atomic Energy Commission, Washington 20545

Croft, Babcock and Wilcox Company, P.0. Box 1260, Lynchburg, Virginia 24505

Deering, Atomic Energy Commission, RDT Site Office (ORNL)

Douglas, Materials Systems Division, UCC, Kokomo, Indiana 46901

Falkenberry, Tennessee Yalley Authority, 303 Power Building, Chattanooga, Tenn. 37401
C. W. Fay, Wisconsin Michigan Power Company, 231 W. Michigan Street, Milwaukee, Wisconsin
Gregory Flynn, General Motors, 12 Mile and Mound Roads, Warren, Michigan

A, Giambusso, Atomic Energy Commission, Washington 20545

Gercld Golden, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, lllinois 60439

. Grigorieff, Assistant to the Executive Director, Oak Ridge Associated Universities

. Kehoe, Burns and Roe, Inc., 700 Kinderkamack, Oradell, New Jersey 07649

. Lang, Douglas United Nuclear, 703 Bldg., Richland, Washington 99352

. Larkin, Atomic Energy Commission, ORO

. Lieberman, Atomic Energy Commission, Washington 20545

. Lorenzini, Foster Wheeler, 110 5. Orange, Livingston, N.J. 07039

. Manly, Material Systems Division, UCC, 270 Pork Avenue, New York, New York 10017

. Matthews, Atomic Energy Commission, RDT Site Office (ORNL)

. Mays, Great Lakes Research Corporation, Elizabethton, Tennessee

. McDonald, Battelle-Pacific Northwest Laboratory, Hanford, Washington 99352
. Mcintosh, Atomic Energy Commission, Washington 20542

»
-

a
o
-

HAESOEADA-ED-E
s UrgprPr-=x-=x
444,
445,
446,
447.
448,
449,
450,
451.
459,
453,
454,
455,
456.
457,
458,
459,
460,
461.
462,
463
464,
465,
466,
467.
468.

469470,

471-733,

260

W. J. Mordarski, Nuclear Develop;menf, Combustion Engineering, Windsor, Connecticut

Sidney Parry, Great Lakes Carbon, P.O. Box 667, Niagara Falls, New York 14302

Worth Percival, General Motors, 12 Mile and Mound Roads, Warren, Michigan

G. J. Petretic, Atomic Energy Commission, Washington 20545

M. A. Rosen, Atemic Energy Commission, Washington 20545

H. M. Roth, Atomic Energy Commission, ORO

R. W. Schmitt, General Electric Co., Schenectady, N.Y. 1230]

M. Shaw, Atomic Energy Commission, Washington 20545

Remo Silvestrini, United Nuclear Corporation, 5 New Street, White Plains, N.Y. 10601

E. E. Sinclair, Atomic Energy Commission, Washington 20545

W. L. Smalley, Atomic Energy Commission, ORO

T. M. Snyder, General Electric Co., 175 Curtner Ave., San Jose, California 95103

L. D. Stoughton, UCC, P.0O. Box 500, L.awrenceburg, Tennessee 38464

J. A. Swartout, UCC, 270 Park Avenue, New York, New York 10017

R. IF. Sweek, Atomic Energy Commission, Washington 20545

Richard Tait, Poco Graphite, P.0. Box 1524, Garland, Texas 75040

D. R. Thomas, Commonwealth Assiciates, Inc., 209 E. Washington Ave., Jackson, Michigan 49201
M. Tsou, General Motors, 12 Mile and Mound Roads, Warren, Michigan

J. W, Ullmann, UCC, P.0. Box 278, Tarrytown, New York 10591

C. H. Waugaman, Tennessee Valley Authority, 303 Power Building, Chattanooga, Tenn. 37401
D. B. Weaver, Tennessce Valley Authority, New Sprankle Bldg., Knoxville, Tennessee

G. O. Wessenauver, Tennessee Yalley Authority, Chattanocoga, Tennessee 37401

M. J. Whitman, Atomic Energy Commission, Washington 20545

H. A. Wilber, Power Reactor Development Company, 1911 First Street, Detroit, Michigan
James Wright, Westinghouse Electric, P.O. Box 344, Pittsburgh, Pa. 15230

L.aboratory and University Division (ORQ)

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