ORNL-3872.txt

ORNL-3872
UC-80 — Reactor Technology
TID-4500 (46th ed.)

 

 

" MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT

FOR PERIOD ENDING AUGUST 31, 1965

 

 

 

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
| for the

U.S. ATOMIC ENERGY COMMISSION
 

Printed in USA. Price $5.00. Available from the Clearinghouse for Federal
Scientific and Technical Information, National Bureau of Standards,

U.S. Department of Commerce, Springfield, Virginia

 

 

 

 

 

LEGAL NOTICE

This report was prepared as an account of Government sponsored work. Neither the United States,

nor the Commission, nor any person acting on behalf of the Commission:

A. Makes any warranty or representation, expressed or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
any information, apparatus, method, or process disclosed in this report may not infringe
privately 6wned_righfs; or

B. Assumes any liabilities with respect to the use. of, or for damages resulting from the use of
any information, apparatus, method, or process disclosed in this report.

As used in the above, ‘‘person acting on behalf of the Commission'’ includes any employee or

contractor of the Commission, or employee of such contractor, to the extent that such employee

or contractor of the Commission, or employee of such contractor prepares, disseminates, or
provides access to, any information pursuant to his employment or contract with the Commission,

or his employment with such contractor.

 

 

 
Contract No. W-7405-eng-26

/‘
{
. MOLTEN-SALT REACTOR PROGRAM

R

SEMIANNUAL PROGRESS REPORT
For Period Ending August 31, 1965

R. B. Briggs, Program Director

DECEMBER 1965

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

$
_L’{\; 7
"IN

s C) 0 3/
1ii

Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING
ANALYSIS, AND COMPONENT DEVELOPMENT

1. MSRE OPERATIONS..... s e ecesessecesececesene s ssce s e eeseo s e s 7
Chronological AcCoUnt..coeeeeeeess cecsescesssenns ceeeeesencens 7
Component and System Performance.....ccceeeececsscocsocccscsns 10

Control RodS.eeeeeeeoosns s s eesceccccceresns s s s seses s e s 10
Sampler-Enricher...ceceeeeeeces ceoscoescesesess s et e an e 11
Freeze Valve S .. eeeeerssooscesscososcossocosssosososossssscsossos 12
Freeze FlangeSeeeeeo oo 6t e ocecsecceevsesecessessssas s e e s s e s 13
Fuel and Coolant System Pressure Control.....cceeeeeecesss 13
Heaters and Insulation..eeeeeceeeccsccsoocess ceeseeseneene 15
Reliable Tnstrument Power SyStelm....eoceeessceosccosooosas 15
Thermal-Shield Cooling Waler..oseeeoeeeececocrssocescccnossoe 16
Fuel-Pump Overflow TanK...eeeeoeeoeceocssssssossesscccsnse 16
Analysis of Operation...cceeeeececesescscescsssoscccsossosccacscse 16
Basic Nuclear CharacteristiCS.eeeeececcsssooccscososccsesss 16
Dynamic TeStS.useeeeeoseeeeoeeeeososasasosssscssssossnosssss 24
Undissolved Gas in Fuel LOOPeeeceecsssssoosoosooocoossass coe 30
Salt Density and Inventory EXperience......ceeececesccesss 30
External NeUbron SOUILCE .. eeeessscesoccecooossossossoscsssssssse 31
Instrumentation and Controls Design and Installation.......... 32
CEIIET A L e e s 0o s ovoooooeoccsssossscsssesosasososssscssssssssessss 32
Design, Installation, and Checkout........ooeeveeencaeennn 32
Tnstrumentation and Controls System Performance........... 40
Revisions and ModificationsS.ceeeesceeesoccccosccosccsacsanss 43
DOCUMENLEEI0N . s s e e vesesoeecsesssscssssssssscssescaassccsses 46

2. COMPONENT DEVELOPMENT. . cocavcoscesossoascsasossssosssscocnssseocs 48

AT e T OE S e s e s s evoeosososoesasessesesseosssssssssonsescsssasssos 48
FrEEZE VOLV e e oo eseesoeosocessssssosassssssosessossosasconsss 48
Pipe Healer e iveeeeoseooonsocsocssassssassssscsasscses eeaes 48
Drain-Tank Heaber.. v e eeeeeesaosossssscssssossscccososssss 48

Checkout and Startup of ComponentS...eeeeeeeeeecescccoacasnocss 48
CONETOL ROAS . e s esecsosesesecsesscssessssssssscsscsscsscsnssse 48
Control Rod Drive Unibts..eeeeeeeesocosesoosssoscosccscncssss 49
Freeze ValveS..... cecoessssssess ceoeees ceecoseecsecssasesasene 51
Freeze Valves 204 and 206...eeeesseoesocccscssosssssossssnss 51
Freeze Valve 103, . ceeeecesoeeoooscsssosssocsssossssssssscascesaes 51
Freeze Valves 104, 105, and 106...cceeeesseccsccssancosons 53
Noble-Gas DynamicCS.eeeseeceosocscosscososss cecesereacacnos 55

Sampler-Fnricher..ceeeeeecescoscocsscsss ceseceensno e cesece 56
iv

Removal Valve and S€a8l...ceecocssseossossoscsocassocsosnossss . 56
MAnIPULAEO e s e e veesrosossocososossssssssscsessocssasooscnsse 56
Access Porteeeeeesoseoes cecoeeaceos ceeeeseessssesss s e e e 57
Operational Valve...oeeeoeese ceceon e teceeseccescssssens e 57
Sampling CapsUle. e eeesseoeosooccsoesoseees ceeseeseceseene 58
Fuel-Processing SampPler.ceceessssscosscossossssssosocscassos 58
Off-Gas SySteMeoeesssseasssoseoosns Gt e e seeereseeseessasanns 58
Remobte MainbenanCe.eeeeeseoccsoossososssosesossoosssscssesesssscoseco 60
Pump Development.ecessseosoocoeocssesosssosscsocsscsssocassonsos 61
MSRE PUIDS e e e o sessosesesssscsososssossesssssososossossssss o1
Measurement of the Concentration of Undissolved Helium
in Circulating Molten Sall..eeceeeecccoscessnscsoccacsnccne 62
Other Molten-Salt PUMPSeececscoceccccssossoosossssocsssocssss 65
Tnstrument Developmenteeeeeessesrsooossesesrscssssoossssessccscs 66
Ultrasonic, Single-Point Molten-Salt Level FProbe..... conee 66
High-Temperature NaK-Filled Differentlal-Pressure
T aNSIMIi T e e e v sesveceseossccesosssssssoosossssscsssocssssss 70
Float-Type Molten-Salt Level Transmitter.....ccocceeeeecses 71
Conductivity-Type, Single-Point Molten-Salt Level Probe... 71
gingle-Point Temperature Alarm Switches.....coceeeeeeasees 72
Helium Control Valve Trim Replacement,..ecoeseesoesocoscocss 72
Thermocouple Development and Testinge.ceseeeseerecosseeses 73

MSRE REACTOR ANALYSTIS:eeoeeoeecsssosooscoscocns cessecees e e e 75

Theory of Period Measurements Made with the MSRE During
Fuel Circulation..e.eeeeoseosoossssosescssoososssssssossssosssocesss 75

Part 2. MATERIALS STUDIES
B/[ETA—LL.[JRGY....0..0...0..000.0......9.....0.0....‘..O’O...'0.0. 8]—

Dynamic Corrosion StudieS..veeeeeseressecssooaesccsocscsescans 81
Corrosion Studies Using Lead as a Coolanf...eecevesoncssse 82
Revised Thermal Convection Loop DesSigleceoceecsoscocssonss 85

MSRE Maberials Surveillance TeStiNgeeeeececsecscesssssessssees  O7
TNOR-8 Surveillance SpPeCimMeNnS.eceeseososscccscsssssssssssos 88
Graphite Surveillance SPECIMENS.ceeecoesssosossssosoocsons 88
Assembly of Survelllance SpPECIMENS.cseecesvessoocasacnscss 20

Fvaluation of Radiation-Damage Problems of Graphite for

Advanced Molten-Salt ReaCtorS.eeeesescessscoossosssososscoscssce 93

Mechanical Properties of Irradiated INOR-8...ceeecevoscssccson 94

Mechanical Properties of Unirradiated INOR-8 Used in

the Reactor VessSel..veeeeessoososososooososessssssososssos V.
Postirradiation Stress-Rupture Properties of INOR-8....... 102
In-Reactor Creep TesTS..ieeereserncosrceeccancnnscsocccncen 104

RADIATTION CHEMISTRY ¢ v s evevescesosoncscsossssensossseassseassses 106

IntrOduCtiOIl.O..O..OOO..0.00‘0'0'..0.00.0..O...Q.OO’.O..O..... 106
Experimental Objectives and Design ConsiderationS..eeeeese 106
Experiment Design and Mockup Test Operation........ ceeeees 107
CHEMISTRY ¢ o s oecseoeesssosonocooncss ceceens ceecesensenen s e ees 111
Chemistry of the MSRE..eeseesss ceeev oo cessssses e cereserssen e 111
MSRE Fuel LOBA1NEeoeersoesossosososasossssssssocssssassassonsaes 111
MSRE Salt Chemistry During Precritical and Zero-Power
Experiments.eeee.. c e s e s e e e e ees e s e e s et et es e et s e e e s ceees 112
Examination of Salts After the Zero-Power Experiment...... 117
Density of MSRE Fuel and Coolant SallsS.eieeeesescesaccenes 119
Chemistry of LiF-BeF, Systems.. ceereacns ceeecescacpencanes 123
Solubility of HF and DF in Molten LiF-Bel's
(66-34 Mole B)eereensanss ceveeeessssesassseesss s ceeesas 123
VaPOr PreSSUlCessecessscescesocsasessscesss ceeeseseseens 123
The Quest for quuld Ligquid Tmmiscibility in
LiF-BeF, Melts., 6t e e s e s cecesseese s s s e s s e s e ns ees 126
Removal of Iodlde from.LlF -Bek, MElts by HF-H,
SPATrEING . ceeservocoses s s esseeresecesees ceseecessenees e s 127
Salt Compositions for Use in Advanced Reactor Systems........ . 133
Blanket Salt Mixtures for Molten-Salt Breeder Reactors.... 133
Coolants for the Molten-Salt Breeder Reactor...eeseeesss.. 135
Viscosity Of NaBF,eeveoececococcscossosoosssconsssocsocscss 136

Fuel and Blanket Materials for the Proposed MOSEL Reactor..... 137
Recovery of Protactinium from Fluoride Breeder Blanket

MixXbUreSeseesoeess 6t oo s e e eeec e et s e e s e e st es s e s e e es s e ns 137
Introduction.seeeeeeos. ce e e e s s s e st e e s e e s et tee s s s e eeo 137
Facility DescriptioN..ceecee. cesessecsese st s ee s ee oo s s 137
Oxide Precipitation of Protactinium.......eceeceeeee cesess 140

Development and Evaluation of Methods for the Analy51s

of MSRE Fuel..... creceseeessesscee s cecesecen cesessecenses eo. 140
Determination of Oxide in MSRE Fuel....... cesens ceeeseness 140
Electrochemical AN8lySiS..eseceeeesessesescsosssoscossasas 143
Spectrophotometric Studies of Molten-Salt Reactor Fuels... 145
Analysis of MSRE Blanket Gas. ceeeseseeresecsensnseenns 147

Development and Evaluation of Equlpment for Analy21ng

Radioactive MSRE Fuel SampleS...... ceese s cereaes cecsesecesss 148
Sample Preparation..eeeecoececocesass ce s esesacesecesoseonees 148
Sample Analyses.. ceeereecenoaes cesecsveco s ceesescanee 148
Quality Control Program ........ cevenas ceeresetreseanas eee. 148

FUEL PROCESSING....... cessecssscevvsessreenes e veoecconon eesse 152
.-

 
vii

SUMMARY

 

Part 1. MSRE Operations and Construction, Engineering Analysis,
and Component Development

 

1. MSRE Operations

 

The first run in which salt was circulated ended on March 4, after
the planned prenuclear tests were completed. In this run flush salt was
circulated for 1000 hr in the fuel loop and coolant salt for 1200 hr
in the coolant loop; equipment performance was generally satisfactory.

With the reactor shut down, the operators recelved advanced tralin-
ing. They were then examined and, if they qualified, were certified as
nuclear reactor operators. Meanwhile the system was prepared for low-
power nuclear operation by installation of the nuclear instruments, the
fuel sampler-enricher, and one layer of concrete blocks over the reactor
cell.

Fuel carrier salt (lacking enriched uranium) was charged in late
April. As a final checkout and to provide base-line data, this salt was
circulated for ten days before the addition of enriched uranium began on
May 24. Criticality was first attained on June 1, with circulation
stopped, control rods almost fully withdrawn, and a 2357 concentration
very close to predictions. June was spent in adding more 235U} while
one control rod was calibrated over its full travel, reactivity coeffi-
cients were measured, and dynamics tests were conducted. The nuclear
power was held to a few watts except during planned transients, in which
it was allowed to rise to a few kilowatts. Nuclear characteristics were
in good agreement with predicted values.

Performance of the mechanical components of the system was gratifying.
Only minor difficulties were encountered, interfering in no way with the
experimental program.

The fuel system was drained and flushed on July 4—5. Radiation
levels were low enough to permit maintenance and installation work to
begin immediately in all areas in preparation for high-power operation.
The remainder of the period was spent in this work.

Except for a small amount of work on the fuel-processing system
sampler, completion of documentation, and a few minor additions and re-
visions, the design, installation, and checkout of the MSRE Instrumen-
tation and Controls System are now complete. Since our last semiannual
report the design, installation, and checkout of all instrumentation and
controls systems required for power operation of the reactor were com-
pleted, the computer—data-logger system was installed and checked out,
several revisions and modifications were made to improve performance or
correct errors in design, and some safety instrumentation and associated
control circuitry were added. Documentation of design and as-built changes
viii

is nearing completion. In general, performance of the instrumentation has
been very good. Although some failures and malfunctions have occurred,
once the causes of the failures were determined, they were easlly cor-
rected, and no major changes in instrumentation or in design philosophy
have been necessary.

2. Component Development

 

A thermal cycling test was completed on the prototype of a freeze
valve after 1800 cycles through the temperature range from 600 to 1200°F.
No cracking or other physical damage was evident.

Prototypes of the removable heater for 5-in. pipe and the drain-tank
heater completed over 12,000 hr of satisfactory test operation.

The modified control rods were operated in the reactor during the
criticality test and performed satisfactorily. A satisfactory procedure
for attaching the control rods to the drive unit by use of remote methods
was demonstrated.

The limit switch actuators, which operate at the limit of the con-
trol rod stroke, were modified to improve the bearing surfaces. A pro-
totype of the modified switch was operated through 14,760 switch opera-
tions without difficulty. The buffer stroke of the shock absorber had
changed by as much as 50% during the criticality test and had to be re-
adjusted before reinstallation. Temperature alarm switches were installed
in the lower end of each control rod drive to indicate gross changes in
the cooling air flow. The wire cables for the electrical disconnects were
lengthened to facilitate remote removal and installation.

The cooling air flows to the freeze valves were increased to about
2/ sefm. Freeze valves Tor the coolant drain tank were modified to re-
duce the thaw time under power failure conditions. This time is now about
13 min. We continued to experiment with the reactor drain valve in an
effort to improve the procedures for operating the valve under the dual
set of operating conditions. Also there was some difficulty in operating
this valve when the reactor temperature was below about 1100°F. A 4if-
ferential controller was installed in the cooling air stream, and the valve
performed satisfactorily during the critical experiment. We also had some
difficulty with the freeze valves in the distribution lines to the drain
tanks, and methods were developed to ensure proper filling of the valves
with salt before the valve was frozen. In addition, the problem of me-
chanical failure and inadequate capacity for the heaters in these valves
was solved by installing new heaters.

Analysis of the experiment on the behavior of noble gas in the re-
actor was continued. Preliminary results of one experimental run yielded
five rate constants which control the transient behavior of 85Ky, A rea-
sonable agreement was found between these constants and the equivalent
physical constants measured elsewhere.
1X

A total of 54 samples were isolated and 87 capsules of enriching
salt were added by means of the sampler-enricher during the zero-power
runs. The equipment was tested for the first time and operators were
trained during the same period.

Maintenance was sufficient to improve the performance of the sample-
removal valve and the operational valve. The manipulator boot failed
three times, but the causes have been eliminated by modifying the pro-
tecting interlock system which controls the pressure and by altering a
protrusion in the normal path of the manipulator. A sample capsule was
retrieved from the top of the operational valve after it had been acci-
dentally dropped. The design of the fuel-processing sampler was started.

A holdup test using 85Kr as an indicator was made to check the per-
formance of the MSRE charcoal beds. The beds performed as predicted.

Difficulties with the operation of the valves in the fuel and coolant
off-gas system were traced to an accumulation of glassy spheroids of the
salt and a carbonaceous material which could be a normal residual of the
manufacturing process. The filter element upstream of the valve is being
changed from one with a 25-j pore size to one with a 1-K pore size, and
no further trouble 1s expected. Design was started on an off-gas sample
unit which is to incorporate an on-line indicator of the total contami-
nants as well as provide the means of concentrating and collecting batch
samples of the off-gas.

Practice with the remote-maintenance tools and procedures was con-
tinued. FPhotographs were taken of all the installed equipment in order
to provide a final record of the as-installed condition.

The spare rotary assembly for the fuel pump was operated for 2644
hr circulating salt. The radiation densitometer used to determine the
concentration of undissolved gas in the circulating salt during this
test was used again at the MSRE to make a similar measurement for the
circulating fuel salt. Also, back-diffusion tests using 85Ky were made
during the same test.

The spare rotary assembly for the MSRE coolant pump was completed.
Tests on the mockup of the MSRE lubrication system were completed. The
priming problem with the standby lube pumps was resolved by modifying
the jet pump in the oil return circuit to the reservoir. The PKP molten-
salt pump was placed in operation at 1200°F for an endurance run.

Fabrication and installation of an ultrasonic level probe system in
the MSRE fuel storage tank were completed. The probe performed very well
during the initial filling of the fuel storage tank but did not operate
when the tank was drained. This malfunction was determined to have been
caused by frequency drift of the excitation oscillator. Performance
studies revealed other characteristics which could limit the usefulness
of the instrument for long-term service under field conditions. Modifi-
cations are being made or are under consideration which would eliminate
the unwanted characteristics.
The coolant-salt-system flow transmitter that falled in service at
the MSRE was replaced by a spare transmitter. A new transmitter has been
ordered for use as a spare. Tests are being performed on the defective
transmitter to determine the cause of failure and to determine whether it
can be repaired.

Inspection of the core tube in a prototype ball-float-type molten-
salt level transmitter showed that the buildup of vapor-deposited salts
in critical areas has not been sufficient to affect the performance of
the instrument.

The fuel flush tank level probe was modified and repaired. Consid-
erable difficulty was experienced with sulfur embrittlement of nickel wire
in the replacement assembly. Performance of the repaired probe and of
other probes installed in MSRE drain tanks was satisfactory during critical
and low-power operations.

Modifications of the temperature alarm switches to eliminate spurious
set-point shifts were completed. 1t i1s not known at this time whether the
modifications effectively eliminated the set-point shifts.

Results of investigations indicate that the failure of four helium
control valves in MSRE service was probably caused by misalignment and
complete lack of lubrication rather than incompatibility of plug and seat
materials.

Calibration drift of eight thermocouples made of materials selected
from MSRE stock remained within the limits previously reported.

Ten MSRE prototype, surface-mounted thermocouples installed on the
prototype pump test loop continued to perform satisfactorily throughout
the test.

Revisions were made in the MSRE coolant-salt radiator AT measurements

system to eliminate long-term drifts and noise found to be present in the
MORE installation.

3. MSRE Reactor Analysis

 

As an aid in interpretation of the zero-power kinetics experiments
with the MSRE, the theory of period measurements while the fuel is in
circulation has been developed from the general reactor kinetic equa-
tions. The resulting inhour-type equation was evaluated numerically by
machine computation, and results are presented relating the reactivity
and the asymptotic period measured during circulation. By means of this
analysis, the measured and calculated reactivity differences between the
noncirculating and circulating critical conditions were found to be in
close agreement.
X1

Part 2. Materials Studies

 

4. Metallurgy

 

Thermal convection loops made of INOR-8 and type 304 stainless steel
have circulated LiF-BeF,-ZrF,-UF,-ThF, (70-23-5-1-1 mole %) fuel for 29,688
and 18,312 hr respectively. A maximum attack of 0.002 in. was found
on specimens removed from the type 304 stainless steel loop.

Thermal convection loops were run to evaluate the compatibility of
lead with Croloy 2—1/4 steel, low-alloy steel, type 410 stainless steel,
and. Cb—1% Zr at 1100 to 1400°F meximum temperatures. The steel loops
tended to plug in the cold regions and have general surface corrosion in
the hot region. The Cb—1% Zr was found to have no measurable attack at
1400°F in 5280 hr. A new loop design that allows improved temperature
control of the cold region was tested for use in coolant evaluation
studilies.

MSRE surveillance specimens were assembled and placed in the re-
actor core, the control test rig, and the area adjacent to the reactor
vessel. An assembly of graphite specimen, INOR-8 tensile bars, and flux
monitor wires will be exposed to fluxes at various points of the reactor
to anticipate and match the effects of radiation on the materials of the
reactor core and vessel.

The radiation-damage problems were evaluated for graphite in ad-
vanced molten-salt reactors, considering growth rate, creep coefficient,
flux gradient, and geometric restraint as important factors. The stress
developed because of differential growth in an isotropic graphite should
not exceed the fracture strength of the graphite and cause failures. Evi-
dence exists that %raphite should have the ability to withstand damage to
doses up to 4 X 10 2 Data are needed for greater dose levels.

A study of INOR-8 specimens made from heats of material used in the
MSRE reactor vessel indicates that (1) creep strength is comparable to
those reported previously, (2) properties of the alloy are very sensitive
to mechanical and thermal treatment, and (3) welding of air-melted heats
without subsequent heat treatment causes large reductions in rupture life
and ductility, whereas welding of vacuum-melted heats causes small effects
on these properties. A microstructure study using the electron microprobe
indicates that the large precipitates present are nickel-molybdenum inter-
metallics with high silicon content. Several modified alloys are being
studied that have potentially improved properties.

Postirradiation stress-rupture properties of heat Ni-5065 and heat
2477 at 650°C were determined at doses varying from 5 X 1016 to 5 x 1020
nvt. These data show that ductilities at the higher levels vary from 1
to ~5%, the lower-boron-bearing heat 2477 having the higher ductility.
otress-rupture life was reduced as dose level was raised. In-pile creep
data were developed for two heats of material.
Xi1

5. Radiation Chemistry

 

The in-pile irradiation program is being changed from an MSRE-oriented
program to one that will provide an understanding of both short-term and
long-term effects of irradiation and fissioning on molten-salt reactor
fuels and materials. Ixperimental objectives of the program are: (1)
200'w/bm3 fuel fission power, (2) meximum fission product production,
and (3) long~-term in-pile operation (up to one year).

The i1rradiation tests are to be conducted in beam hole HN-1 of the
ORR with an autoclave (capsule type) experiment. Design features in-
clude: (1) circulation of the salt by means of thermally induced flow,
(2) sampling and replacement of fuel salt while operating in-pile, (3)
cover gas sampling, and (4) keeping fuel molten at all times.

oeveral prototype models of the in-pile molten-salt autoclave ex-
periment have been constructed and operated in a mockup facility. Some
4000 hr of operation with a salt mixture similar to the MSRE fuel salt
have been accumulated. Results of these mockup tests indicate that the
presently designed autoclave is suitable for in-pile experiments with
molten-salt fuel.

6. Chemistry

A1l fluoride mixtures — coolant, flush, and fuel — for the operation
of the MSRE were prepared and loaded into the reactor facility by the Re-
actor Chemistry Division. Capsules of fuel concentrate, containing about
85 g of 2357 each, were provided for use in reaching criticality and for
criticality maintenance during nuclear operation.

Chemical analyses of the MSRE fuel were carried out during the pre-
critical, zero-power, and postcritical stages for the purpose of estab-
lishing analytical base lines for use in the full-power operating period.
Chemical composition, contaminant levels, and isotopic analyses were ob-
tained regularly on samples obtained daily throughout the zero-power ex-
periments. Judging from the concentration of Cr<™, which is the primary
corrosion product, essentially no corrosion occurred during the 1100-hr
precritical and zero-power test period. From the standpoint of chemical
evidence, the MORE salts were maintained in an excellent state of purity
during all transfer, fill, and circulation operations. Unless dilution
of the fuel by flush salt is postulated, uranium analyses for samples
obtained in both precritical and zero-power experiments were about l%
below book values.

New experimental values for the densities of the fuel and the coolant
agreed well with results on the weights and volumes of salts in the drain
tanks at the MSRE site.

The solubilities of HF and DF in the molten mixture LiF-BeF, (66-34
mole %) were measured over the range 500 to 700°C at pressures of 1 to 2
X111

atm. The solubilities were of the order of 2 X 10™% mole of HF per mole
of melt, and DF solubilities were lower than HF solubilities by about 10%.

Vapor pressures were measured for the LiF-BeF, system over the entire
composition range. The vapor above liquid compositions containing 70%
or more BelFp was virtually pure BeFy. Vapor pressures of importance in
recovering the MSRE fuel by distillation were also determined.

The fea81b111ty of removing 1357 from the fuel, as a way of reducing
the amount of *3 Xe, was examined in greater detail. Half the iodide con-
tent could be removed by using 388 cc of gaseous HF in an Ho-HF mixture
per kilogram of melt.

A Turther search for liquid-liquid immiscibility in the LiF-BeF,
system at high Bel, concentrations was made; no immiscibility region
was found.

Considerations of phase behavior in ternary fluoride systems con-
taining ThF, have led to the selection of suitable blanket systems for
breeder reactors. A search for suitable coolants, however, continues.
At present, interest is centered on the potentialities of fluorides and
fluoborates, possibly in combination with Bs03. Reports in the Russian
literature of a low-melting eutectic of NaF-NaBF, could not be confirmed.

The new facility was designed for laboratory-scale studies of the
removal of protactinium from fluoride breeder-blanket mixtures and for
supporting research work. Glove boxes will permit use of the *21Pg
isotope to give concentrations in the expected operating range of 50 to
100 ppm. Hot cells would be required for work with equivalent concen-
trations of <° Pa but millicurie amounts of this gamma-active isotope
will be mixed'w1th 231pg in order to minimize the need of alpha anal-
yses.,

A preliminary experiment was performed to test the equipment and to
confirm the previously reported precipitation of protactinium by addition
of oxides. Protactinium at tracer concentration K1 ppb) was completely
pre01p1tated'by addition of thorium oxide to molten LiF-BeF,-ThF, (73-2-
25 mole %), and treatment of the melt with a dry mixture of HF and H,
redissolved the protactinium.

A prototype apparatus was constructed and tested for the determina-
tion of oxides in the MoRE fuel using the hydrofluorination principle.
The water produced from the reaction of HF with oxides i1s measured auto-
matically by means of an electrolytic moisture monitor. The entire ap-
paratus 1is being assembled for insertion into a hot cell in order to an-
alyze the fuel after the reactor has gone to power.

otudies were continued on adapting electrochemical methods to in-
line analysis of impurities in the fuel. When the simulated fuel is
subjected to controlled-potential electrolysis, gas evolution, primarily
X1V

of COz, CO, and Oz, is observed at the indicating electrode. This indi-
cates removal of oxide from the melt by electroreduction. This discovery
holds promise for a possible in-line determination of oxide. Absorbance
spectrophotometric studies are also under way which are designed to de-
termine trivalent uranium and tetravalent uranium in the fuel by their
characteristic absorbance peaks.

A process gas chromatograph equipped with s helium breakdown-voltage
detector is under construction for the continuous analysis of the helium
cover gas 1n the reactor. A metal diaphragm sampling valve has been de-
signed specially to withstand the temperature and radiation effects that
nullify the use of conventional sampling valves.

camples from the MSRE precritical and zero-power experiment were
analyzed in the HRLAL hot cells. The results, using the specially de-
veloped equipment and analytical methods, were satisfactory with the
exception of those for uranium and beryllium. Statistical evaluation of
the control data indicated a negative bias of ~0.8% for uranium and none
for beryllium.

7. Fuel Processing

 

Construction of the MSRE fuel-processing system was completed, the
system was tested, and the flush salt was processed for oxide removal.
Operation of the plant was generally satisfactory, and about 115 ppm of
oxide was removed from the salt in reducing the concentration to about
50 ppm.
INTRODUCTION

 

The Molten-5Salt Reactor Program is concerned with research and de-
velopment for nuclear reactors that use mobile fuels, which are solu-
tions of fissile and fertile materials in suitable carrier salts. The
program is an outgrowth of the ANP efforts to make a molten-salt reactor
power plant for aircraft and is extending the technology originated there
to the development of reactors for producing low-cost power for civilian
uses.

The major goal of the program is to develog a thermal breeder re-
actor. Fuel for this type of reactor would be 33UF4 or 235UF4 dissolved
in a salt of composition near ZLiF-Bel';. The blanket would be ThF, dis-
solved in a carrier of simillar composition. The technology being devel-
oped for the breeder is applicable to, and could be exploited sooner in,
advanced converter reactors or in burners of fissionable uranium and plu-
tonium that also use fluoride fuels. osolutions of uranium, plutonium,
and. thorium salts in chloride and fluoride carrier salts offer attractive
possibilities for mobile fuels for intermediate and fast breeder reactors.
The fast reactors are of interest too but are not a significant part of
the program.

Our major effort is being applied to the development, construction,
and operation of a Molten-Salt Reactor Experiment. The purpose of this
Experiment is to test the types of fuels and materials that would be used
in the thermal breeder and the converter reactors and to obtain several
years of experience with the operation and maintenance of a small molten-
salt power reactor. A successful experiment will demonstrate on a small
scale the attractive features and the technical feasibility of these sys-
tems for large civilian power reactors. The MSRE will operate at 1200°F
and atmospheric pressure and will generate 10 Mw of heat. Initially, the
fuel will contain 0.9 mole % UFy, 5 mole % ZrF,, 29.1 mole % BeF,, and
65 mole % LiF, and the uranium will contain about 30% 235U. The melting
point will be 840°F. In later operation, highly enriched uranium will
be used 1n lower concentration, and a fuel containing ThF, will also be
tested. In each case the composition of the solvent can be adjusted to
retain about the same liquidus temperature.

The fuel will circulate through a reactor vessel and an external
pump and heat exchange system. All this equipment is constructed of
INOR-8,1 a new nickel-molybdenum-chromium alloy with exceptional re-
sistance to corrosion by molten fluorides and with high strength at
high temperature. The reactor core contains an assembly of graphite
moderator bars that are in direct contact with the fuel. The graphite
1s a new material® of high density and small pore size. The fuel salt
does not wet the graphite and therefore should not enter the pores, even
at pressures well above the operating pressure.

 

1S0ld commercially as Hastelloy N and Inco No. 806.
2Grade CGB, produced by the Carbon Products Division of Union
Carbide Corp.
Heat produced in the reactor will be transferred to a coolant fuel
1n the heat exchanger, and the coolant salt will be pumped through a
radiator to dissipate the heat to the atmosphere. A small facility is
being installed in the MSRE building for occasionally processing the fuel
by treatment with gaseous HF and F».

Design of the MORE was begun early in the summer of 1960. Orders
for special materials were placed in the spring of 1961. Major modifi-
cations to Building 7503 at ORNL, in which the reactor is installed,
were started in the fall of 1961 and were completed by January 1963.

Fabrication of the reactor equipment was begun early in 1962. Some
difficulties were experienced in obtaining materials and in making and
installing the equipment, but the essential installations were completed
so that prenuclear testing could begin in August of 1964. The prenuclear
testing was completed with only minor difficulties in March of 1965. Some
modifications were made before beginning the critical experiments in May,
and the reactor was first critical on June 1, 1965. The zero-power ex-
periments were completed early in July. Additional modifications, main-
Tenance, and sealing and testing of the containment are required before
the reactor begins to operate at appreciable power. This work should be
completed in October, and the reactor should be at full power before the
end of the year.

Because the MoRE is of a new and advanced type, substantial research
and development effort is provided in support of the design and construc-
tion. Included are engineering development and testing of reactor com-
ponents and systems, metallurgical development of materials, and studies
of the chemistry of the salts and their compatibility with graphite and
metals both in-pile and out-of-pile. Work is algso being done on methods
for purifying the fuel salts and in preparing purified mixtures for the
reactor and for the research and development studies.

This report is one of a series of periodic reports in which we de-
scribe briefly the progress of the program. ORNL-3708 is an especially
useful report because it gives a thorough review of the design and con-
struction and supporting development work for the MSRE. It also describes
much of the general technology for molten-salt reactor systems. Other re-
ports issued in this series are:

ORNL-2474 Period Ending January 31, 1958

ORNL-2626 Period Ending October 31, 1958

ORNL-2684 Period Ending January 31, 1959

ORNL-2723 Period Ending April 30, 1959

ORNL-2799 Period Ending July 31, 1959

ORNL-28°90 Period Ending October 31, 1959

ORNL-2973 Periods Ending January 31 and April 30, 1960

ORNL-3014% Period Ending July 31, 1960
ORNL-3122
ORNL-3215
ORNL-3282
ORNL-3369
ORNL-3419
ORNL-3529
ORNL-3626
ORNL-3708
ORNL-3812

Period Ending
Period Ending
Period Ending
Period Ending
Period Ending
Period Ending
Period Ending
Period Ending
Period Ending

February 28, 1961
August 31, 1961
February 28, 1962
Avgust 31, 1962
January 31, 1963
July 31, 1963
January 31, 1964
July 31, 1964
February 28, 1965
Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING
ANATYSIS, AND COMPONENT DEVELOPMENT
1. MSRE OPERATIONS

Chronological Account

 

The principal accomplishments for the period from March through Au-
gust 1965 were the preparation for, and completion of, the initial criti-
cal and associated "zero-power' experiments. Initial criticality was
achieved on June 1, and the experiments were concluded on July 3. The
remainder of the period was used in preparing the system for operation
at power.

The initial, precritical operation of the reactor system (run PC-l),
with flush salt in the fuel loop, was concluded on March 4, after the ex-
periments on noble-gas behavior were completed. (The analysis of the re-
sults of these experiments is discussed in Chap. 2, Component Development.)
In this run, salt was circulated at high temperature for 1000 hr in the
fuel loop and 1200 hr in the coolant loop. After the flush salt was
drained from the fuel loop, it was transferred to the fuel storage tank,
there to await processing to remove oxides. (The processing is described
on page 152.)

The next five weeks (until midaApril) were spent in advanced class-
room training of the reactor operators and supervisors and the administra-
tion of qualifying examinations. Before the nuclear experiments started,
there were at least one engineer and one technician on each crew who had
been qualified and certified. Others were certified later as they com-
pleted individual oral examinations.

While the operator training was in progress, final physical prepara-
tions were made for zero-power nuclear operation. These included.:
1. 1installation and checkout of the fuel-salt sampler-enricher,
2. fTinal installation and checkout of the control rods and drives,
3. completion and checkout of the nuclear instrumentation and controls,
2

installation of a gamma-ray densitometer on the fuel-salt inlet line
to the reactor,

5. installation of the lower layer of shield plugs on top of the reactor
cell,

6. miscellaneous minor maintenance Jobs.

The fuel "carrier" salt (a mixture of LiF, BeF,, and ZrF,) was
charged into fuel drain tank NO.ZB(FD-Z), starting April 21. The con-
tents of 35 shipping containers (4560 kg of salt) were melted and trans-
ferred to the drain tank in six days. To this was added 236 kg of LiF-
UF, eutectic containing 147 kg of 238y (depleted in 235U).

The first operation with this barren salt was to obtain neutron
counting rates with the salt at various levels in the core. Then, as a
Tinal check on the operation of the equipment and to establish base lines
for chemical analyses of the fuel salt, the carrier salt was circulated
for ten days in run PC-2. Eighteen samples, taken through the newly in-
stalled sampler-enricher, showed that the salt composition was as ex-
pected. (See page 113.) This, coupled with satisfactory operation of
all equipment, indicated that at last all was ready for the initial crit-
ical experiment.

The addition of #3°U was started on May 24, and initial criticality
was achieved at 6:00 PM on June 1, 1965. The 255U’was added as the IiF-
U, eutectic with highly enriched uranium (93%). The bulk of this ma-
terial, containing 69 kg of 235U} was loaded in four charging operations
to FD-2. After each addition the salt was transferred to the second drain
tank (FD-1) and back again to ensure thorough mixing. The mixed salt was
loaded into the reactor system after each charging operation, and count-
rate data were taken at several salt levels in the core and with the re-
actor vessel full. These data were compared with the barren-salt dats

to monitor the neutron multiplication and to establish the size of the
next addition. Extrapolation of inverse-count-rate plots with the re-
actor vessel full showed that the loading after the fourth addition was
within 0.8 kg 2327 of the critical loading when circulation was stopped
and the control rods were withdrawn to their upper limits. The remainder
of the #3°U was added directly to the circulating loop with enriching
capsules. These were inserted into the fuel-pump bowl via the sampler-
enricher to increase the loading 85 g at a time. Count rates were meas-
ured after each capsule with the fuel pump off and the control rods with-
drawn. The reactor became critical after the eighth capsule with the
pump off, two rods fully withdrawn, and one poisoning 0.03 of its avail-
able worth.

After the initial critical condition was established, additional en-
riching capsules were added to increase the uranium loading to the op-
erating level. Enough excess reactivity was added in this way so that
one control rod could be calibrated over its entire range of travel. The
various zero-power experiments were performed during this phase of the
operation. These included, in addition to the rod-calibration experi-
ments, measurements of

temperature coefficient of reactivity,
uranium-concentration coefficient of reactivity,
effects of fuel circulation on reactivity,

effects of system overpressure on reactivity,

bW

dynamic characteristics.

Use was made of the on-line digital computer for collecting data for some
of these experiments even though the equipment was not completely checked
out and in normal service.

kxperiments specifically aimed at rod worth were stable period meas-
urements and rod drop experiments. These were done with the fuel static
and with it circulating. The results will give, as accurately as possible,
the total and differential worths of the regulating rod (rod 1) over its
entire travel with the other two rods fully withdrawn. In addition,
worth values will be obtained for each of the three rods with the other
two withdrawn and at intermediate positions. These will lead to evalu-
ations of rod shadowing and "ganged" rod worth.

After the initial critical experiment, another eight capsules were
required before the reactor could be made critical at 1200°F with the
fuel pump running (a consequence of the loss of delayed neutrons during
circulation). Thereafter, we measured the critical rod position, with
the pump running, after each capsule. At intervals of four capsules, we
made period measurements with the pump running; then we turned it of'f,
determined the new critical rod position, and made more period measure-
ments. This continued until 87 capsules had been added. Three times
during this experiment (after 30, 65, and 87 capsules), we observed rod
drop effects.

Period measurements were usually made in pairs. The rod on which
the sensitivity was to be measured was adjusted to make the reactor Just
critical at approximately 10 w; then it was pulled a prescribed distance
and held there until the power increased about 2 decades. The rod was
then quickly inserted to bring the power back to 10 w, and the measure-
ment was repeated at a shorter stable period. Periods were generally in
the ranges from 40 to 50 sec and from 70 to 120 sec. The available re-
sults of these and the other zero-power experiments are discussed in the
section Analysis of Operation.

Most of the zero-power experimental program was carried out with the
coolant system empty. However, some of the dynamic tests required cir-
culation of the coolant salt. This loop was filled on June 20, and salt
was circulated for 118 hr while the tests were in progress. The coolant
loop was drained on July 1. The zero-power experiments were concluded,
and the fuel loop was drained on July 4 after 764 hr of circulation in
this run. The loop was then filled with flush salt, which was circulated
1.3 hr, sampled, and drained to prepare the system for maintenance.

With the reactor shut down, the final preparations were started for
operation at significant power. The principal jobs to be accomplished
during this shutdown are:

1. modification of the coolant-radiator door assembly,

2. modification of the coolant-salt penetrations of the reactor contain-
ment cell,

3. dnstallation of a new graphite-sample assembly in the reactor vessel,

4. removal and replacement of the fuel-pump rotary element for remote
maintenance practice,

closure and leak testing of the reactor containment,

installation of stacked-block shielding.
10

Component and System Performance

 

In general, the performance of the many mechanical components and
auxiliary systems during operation was highly satisfactory. This is par-
ticularly true in view of the fact that some of the items were being in-
tegrated into the system operation for the first time. Some difficulties
were encountered which caused temporary inconvenience, but no program de-
lays resulted and no extensive modifications will be required to improve
future performance. This section deals with the difficulties that were
experienced, their actual and potential effects, and the changes which
they incurred. In addition, some routine experience with selected com-
ponents is discussed.

Control Rods

 

Two of the control rods and drives were installed during run PC-1
and were used in simulator training. Before PC-2 all three rods were
installed and subjected to a test consisting of 100 cycles of full with-
drawal and scram. The rods operated freely and never failed to scram,
but occasionally the lower limit switches failled to clear properly as
the rods were withdrawn. We found the cause to be galling in the cam
actuator for the switch. After we installed Stellite bearing surfaces
to remedy this problem, each rod was successfully raised and scrammed
30 times without any malfunction. (The lower limit switch on rod 2 at
first stuck as before, and we found that a shim had been left out of the
switch-actuator assembly. After the shim was replaced, there was no fur-
ther trouble.) Operation continued throughout run 3 without trouble.

Rod drop times were measured in the tests in PC-2 and in a geries
of 40 scrams at the end of run 3. The results (Table 1.1) show that
the drop times became slightly shorter and more uniform. This is con-
sistent with development experience in breaking in new flexible rods.

Table 1.1. Observed Drop Times of Control Rods

 

. a
Number of Timed Drops T eDrop Times éfseg) ——
Rod 1 Rod 2 Rod 3 verag >tandar eviation

Rod 1 Rod 2 Rod 3 Rod 1 Rod 2 Rod 3

 

Test Series

 

Original 45 64 46 820 818 870 17 25 30
At end of 4l 4. 41 793 775 792 4.0 .1 3.7
run 3

 

aTim.e from actuation of the scram switch (with rods at 5l—in.'withdrawal) until
actuation of lower limit switch (0 in.).

Measurements of indicated rod positions with the lower end of the
poison at the fiducial zero position were made after installation and at
intervals during run 3. The data did not indicate any stretching of the
rods.
11

Inspection of the control rods after run 3 showed that two were in
very good condition, but there was severe damage at one point on the flex-
ible extension tube of rod 3. (This extension, connecting the poison
section to the drive assenmbly, consists of a flexible stainless steel
tube covered with a braided stainless steel wire sheath.) A hole was
worn in the sheath, and the tube convolutions were abraded at the point
which was 1n contact with the lower roller when the rods were fully in-
serted. The roller assembly was cut out of the thimble, and the roller
was found to be worn and rough on one side, indicating that it had been
stuck. Presumably, the roller was jammed when the housing was distorted,
while the assembly was being welded in. The extension tube was replaced,
and a new roller assembly will be installed.

After run 3 the rod drives were also inspected. Modified limit
switch actuators were installed, and the worm gears were replaced with
new fully hardened, lapped gears which had a much smoother finish than
the original gears.

campler-Enricher

 

The sampling of the circulating fuel-salt system during run PC-1 was
done with a temporary samplerl which provided an inert atmosphere for the
sample and prevented air from entering the fuel system. A total of 12
flush salt samples was taken with this sampler.

The installation of the fuel-system sampler-enricher was completed
during the shutdown period before run PC-2. Since that time, 53 samples
were withdrawn, and 87 enriching capsules were added. Although several
minor problems occurred during the reactor operation, none of these pre-
vented the sampler from being operational. In some cases delays were in-
curred during sampling or enriching, but the overall schedule was not
significantly affected. ©GSpecific problems that were encountered are as
follows:

1. Both the operational and the malntenance gate valves developed leaks
through one of the two seats of each valve.

2. The removal valve leaked and required a greater-than-normal helium
flow for buffering.

3. A solenoid valve on the removal-valve actuator failed.

4. The removal valve occasionally failed to close completely and had to
be closed manually.

5. The access port periodically failed to close properly because of faulty
operation of the clamps on the clamp actuators.

6. One sample capsule was accidentally dropped down the sample tube to
the operational valve gate.

7. The drive motor stopped once during the removal of an empty capsule.
The capsule was inserted about 12 in. and was then successfully with-
drawn. The reason for this stoppage is unknown.
12

8. There were three boot failures on the manipulator, one of these in-
volving both boots.

9. The manipulator arm and fingers were bent, causing some difficulty in
gripping the latch cable and moving the manipulator arm.

Most of the troubles occurred because this period was one of testing
of equipment and training of operators. A mockup had been thoroughly
tested, but the installation of the sampler-enricher on the reactor was
completed just before the critical experiments were begun. The operators
are now well trained, and some changes are being made to improve the re-
liability of the equipment and the safety of the operation. The device
1s expected to perform satisfactorily during power operation.

Freeze Valves

 

The freeze valve problems of insufficient cooling air which were re-
ported previously were corrected.?® In addition, the coolant system drain
valves were modified to the same basic design as the fuel-system valves.
The revised coolant valves now have sufficient heat capacity to thaw on
loss of power.

All the freeze valves except the system drain valves were relieved
of the fast-thaw requirement by requiring that the valves to both fuel
drain tanks be thawed while fuel is in the reactor. Previously one valve
was to be normally frozen and was to thaw during an emergency drain.

A proportional controller was added to the FV103 cooling-air supply
after run PC-1 to maintain the freeze valve at a preselected temperature
that would result in a suitable thaw time. This temperature and the
valve temperature distribution were controlled satisfactorily at any
steady fuel-system temperature level, but it was necessary to change the
controller set points whenever the fuel-system temperature was changed.
appreclably. The following thaw times were recorded for the drains after
runs PC-2 and 3.

 

Run Drain Thaw Time (min)
PC-2 Carrier salt l6-l/2a

3 Fuel salt 10

3 Flush salt 18

 

®System cooled to 1100°F.

These thaw times compare with the 33-1/2 min which was required at the
end of run PC-1.

An incident occurred during run 3 which could have delayed an emer-
gency drain of the fuel system had it been required. Electrical power to
13

the freeze-valve control modules was lost for a period of about 10 min be-
cause the terminals of a temporary recorder were accidentally shorted. The
loss of module power turned blast air onto the freeze valves. Normal op-
eration of the valves was restored after module power was regained. Con-
Ttrol circult revisions are being made to prevent the recurrence of this
problem.

Although the performance of the freeze valves was acceptable, it was
still difficult to maintain the proper temperature profiles across the
valves because one controller was used to supply heat to both shoulders
of the wvalve. The heater control systems were revised after run 3 to
provide separate controllers for each shoulder heater on the freeze valves
that serve the fuel and coolant drain tanks and the fuel flush tank. Pro-
portional controllers similar to that on FV1O03 were added to the cooling-
alr supplies for the fuel and coolant drain-tank freeze valves (105, 106,
204, and 206).

Freeze Flanges

 

The five freeze flanges on the main fuel- and coolant-salt piping
continued to perform satisfactorily. Leakage of the buffer (leak de-
tector) gas was not excessive at any time. The buffer gas leakage rates
measured at various times during the reactor operation are listed in
Table 1.2.

Table 1.2. Observed Leak Rates of Buffer Gas from Freeze Flanges

 

Leakage Rate (stad cm3/sec)

 

 

oystem
freeze Initial Circulating Hot Drained System Cold System Cold
After After
Flange Heatup oalt After
Run PC-1 Run 3
Run PC-1
X 1077 X 10™2 X 1073 X 1073 X 1077
100 2.0 0.57 Q.7 1.5 2.0
101 1.3 0.4 0.25 2 .20 1.2
102 0.5 0.3 0.30 2.33 1.0
200 1.0 0.21 0.41 1.48 0.3
201 0.6 0.22 0.21 1.23 1.8

 

The freeze flange leakages are normally monitored as a group,; leak-
age of individual flanges would be measured if the group leakage were
higher than normal.

Fuel- and Coolant-System Pressure Control

Difficulties were experienced in controlling the fuel- and coolant-
system pressures within close limits because of accumulation of solids
in the off-gas throttling valve used for fuel-system pressure control and
in the filter just upstream of the coolant-system pressure-control valve.
14

During run PC-1 (Januvary-March 1965), when the coolant salt was
circulated for 1200 hr, the coolant off-gas filter plugged and was re-
placed twice. When the filters plugged, pressure was controlled at
5 * 2 psig by manual venting through a larger bypass valve. Inspection
showed that the filter was covered with amorphous carbon containing traces
of the constituents of the coolant salt and INOR-8. Before run PC-2 the
filter was replaced with one having 35 times the surface area. Coolant
salt was not circulated again until near the end of run 3 and then for
only 118 hr. During this time the pressure control again became erratic,
indicating obstruction of either the filter or the valve. Both were re-
moved for inspection, and although there was no deposit on the filter,
the valve was partially obstructed by a black, granular material. Rinsing
with acetone restored the original flow characteristics of the valve, and
1t was reinstalled.

The fuel-system pressure control became erratic near the end of run
PC-1, and at the conclusion of the run the off-gas filter was removed.
It was clean; so the pressure control valve was removed and found to be
partially plugged. The obstruction was blown out with gas, and the valve
was washed out with acetone. The valve then performed normally and was
reinstalled. The acetone rinse was darkened and contained small (1—5 u)
beads of a glassy substance.

Fuel-system pressure control was satisfactory at the beginning of
run PC-2, but within a week the valve began sticking again. This time
it was replaced with one having a large CV (0.077 instead of 0.02). The
original valve was cut open for inspection, and a black deposit was found
partially covering the tapered stem. The deposit was about 20% amorphous
carbon, and the remainder was the 1- to 5-p glassy beads, which proved
to have the composition of the flush salt.

The larger replacement valve in the fuel off-gas line gave adequate
pressure control for the first four days of salt circulation in run 3.
However, there were four occasions when it appeared to be sticking. When
this happened, the pressure built up slowly to about 6 psig before the
valve opened to drop it back to the normal 5 psig. For the next 20 days
Tthe small pressure variations predominated, suggesting either that the
valve was not functioning properly or that intermittent partial plugging
was occurring. This condition cleared up abruptly, and during the last
ten days of salt circulation the loop pressure remained completely stable.
Since no corrective action had been taken and the valve was functioning
properly when the system was shut down, no explanation for the earlier
erratic. behavior could be established.

The cause of the solids in the off-gas lines is not yet known. There
is reason to believe that some carbon may have been introduced into the
reactor with the salt, accumulated on the surface in the pump bowl, and
carried into the off-gas line as a dust. O0il contamination of the salt
system has also been suggested. The glassy salt beads in the fuel off-
gas line are probably frozen droplets of mist caused by the stripper
spray, but we do not know whether these were carried into the line con-
Tinuously during operation or were swept out of the pump bowl by sudden
venting.
15

Filters that are capable of removing 1l-u particles are to be installed
in both the fuel and coolant systems. (The pore size of the original fil-
ters was about 25 K.) Presumably, this will protect the valves from fur-
ther accumulations. In any event, the coolant off-gas filter and pressure-
control valve can be maintained directly after power operation; those on
the fuel off-gas are designed for remote maintenance.

Heaters and Insulation

 

The two heater elements which failed during the initial precritical
operation3 and their spares were replaced during the shutdown prior to
run PC-2. In both these cases operation had been continued by using the
installed spare elements. ©Six additional heater-element failures and an
electrical ground at a disconnect were discovered during the checkout for
run PC-2. 1In four of the heaters, the failure occurred at the junction
of the lead-in wire and the heating wire inside the ceramic element. These
elements and the electrical ground were repaired before startup, and the
other two elements were left out of service for runs PC-2 and 3.

During subsequent operation, one in-cell heater failure was noted,
and one failure occurred at a heater power supply. (The latter was re-
paired immediately.) 1t may be noted that a number of heater failures
could occur without significantly affecting the operation of the reactor
system. Unless a heater 1s in a particularly sensitive location, its
failure may not be discovered until individual-element checks are made
during a shutdown.

Three more heater-element failures were found during the shutdown
after run 3. In addition, a number of minor defects (grounds, improper
resistances, damaged connectors) were found. All the defects, including
those left from earlier operations, will be corrected before the next
reactor startup.

An important cause of heater-element failure has been separation
of the lead-in from the heater wire. This is apparently due to a com-
bination of a design weakness and excessive flexing during installation
of the elements. The Jjoint was redesigned, and new elements which in-
corporate the change are being installed where such failures occurred.

Reliable Instrument Power System

 

Motor-generator sets 1 and 4 and the 250-v battery system normally
supply power to the fuel and coolant oil pumps and to the instrumentation
system. MG-1 is an ac-to-dc set which supplies 250-v dc power to drive
MG-4 and to charge the 250-v emergency batteries. MG-4 operates from
either MG-1 or the 250-v batteries and normally supplies 120-v ac power
to the instruments and to the fuel and coolant oil pumps.

The initial operation of these MG sets was unreliable because of
failures in electrical control components. The MG sets are used ones
that were installed in the building for earlier experiments, and most
of the failures resulted from aging of the control components.
16

The MG sets were repalred prior to run PC-2 and were placed back in
service. Minor problems occurred, which were corrected by adjustments
to the voltage controls.

Both MG sets were placed in service under normal load at the begin-
ning of run 3, and both performed satisfactorily for the duration of the
run. MG-4 was shut down after the completion of run 3 because of the
failure of an internal cooling blower. The blower motor was replaced to
make the set operational.

Thermal-Shield Cooling Water

 

The thermal-shield water piping was modified to provide an adequate
flow through the three removable segments. The three segments were con-
nected in series, and a line was connected directly to the cooling water
supply. Pressure regulators and pressure-relief valves were installed
in the supply lines to both the main thermal shield and The removable
slides to avold overpressuring the system.

The above changes provided flow rates of 65.8 and 4.6 gpm to the
main shield and to the removable segments respectively. An inspection
of the thermal shield while the fuel system was heated indicated that
the removable segments were adequately cooled by these modifications.

Fuel-Pump Overflow Tank

 

A continuous, but very slow, accumulation of salt in the fuel-pump
overflow tank was observed throughout the operation of the fuel loop.
In 1000 hr of circulation in runs PC-2 and 3, 39 kg of salt was collected
with the pump-bowl level 3 in. below the overflow point. This salt was
recovered Jjust before the run 3 shutdown. Accumulation at this rate will
not significantly affect the operation of the reactor.

Analysis of Operation

 

Bagic Nuclear Characteristics

 

The nuclear characteristics measured during the zero-power tests
generally agreed quite well with the predicted values. Analysis of the
data is still in progress, particularly with regard to control-rod worth
and dynamic characteristics. The results that have been obtained so far
are described below.

Critical Concentration. Predicted and observed <3°U requirements
for criticality are compared most logically on the basis of volumetric
concentration. The required volumetric concentration is nearly invariant
with regard to the fuel-salt density (unlike the mass concentration, which
varies inversely with density) and depends not at all on system volume or
total inventory.

 
17

The observed *3°U concentrations are on a welght basis, obtained
from either inventory records or from chemical analyses. These weight
concentrations must be converted to volumetric concentration by multi-
plying by the fuel-salt density. The amounts of ?2°U and salt weighed
into the system gave a *3°U weight fraction of 1.42% at the time of the
initial criticality. The chemical analyses during the precritical op-
eration and the zero-power experiments gave uranium concentrations which
were 0.985 of the "book" concentrations. (Part of this discrepancy,
about half we believe, is due to dilution of the fuel with flush salt
left in freeze valves and drain-tank heels when the fuel salt was charged.)
Applying this bias to the book concentration at criticality gives an "ana-
lytical"™ #7°U weight fraction of 1.40%. We now believe that the density
of the fuel salt at 1200°F is about 145.5 1b/ft3. This is the preliminary
result of recent laboratory measurements of density, and it agrees with
measurements made in the reactor using the two-point level probes in the
drain tanks. DHRarlier measurements in the reactor, using the drain-tank
welght indications and the volume of salt believed to have been trans-
ferred into the fuel loop, gave 136.6 lb/ftB.

Corrections must be applied because the initial critical conditions
were not exactly the same as those assumed in the predictions. The core
temperature was 1181°F instead of 1200°F, and the control rods were poi-
soning 0.118% 8k/k instead of none. (Two rods were at meximum withdrawal,
51 in., and one was at 46.6 in.) The predicted 23°U concentration for
criticality at the reference condition was 32.87 g/liter; corrected to
the actual conditions, 1t is 33.06. This predicted value is compared
with observed concentrations in Table 1.3.

Table 1.5. Comparison of Critical 2327 Concentrations
(1181°F, pump off, 0.118% dk/k rod poisoning)

 

 

235U 235U

. Fuel Density .

Concentration (lb/ft3) Concentration

(wt %) (g/1iter)
Predicted 33.06
Book 1.42 145.5 33.10
136.6 31.07
Analytical 1.40 145.5 32.60
136.6 30.60

 

1f, as we estimate, the true concentration was about halfway be-
tween the book and the analytical and the density is about 145.5 lb/ft3,
the actual concentration was extremely close to the prediction.

Laboratory measurements now in progress should confirm the value of
the fuel-salt density. The uncertainty between book and analytical con-
centrations will be reduced as a result of the next startup, when analysis
of the fuel after another flushing, draining, and refilling will help us
evaluate the dilution effect.
18

Control Rod Worth. The only data relative to control rod worth that
we have finished analyzing so far are the rod-bump, period measurements
on rod 1 with the fuel static and the other two rods at their upper limits.
These data were used to produce the curve of rod sensitivity as a function
of position shown in Fig. 1.1. Because rod worth is affected by the 235y
concentration in the core, it was necessary to apply theoretical correc-
tions to the measured sensitivities to put them all on the basis of one
concentration. The points in Fig. 1.1l were corrected to the initial crit-
ical concentration, where the sensitivity is the highest. The correction
factors which were applied increase linearly with 237 concentration to
a maximum of 1.087 at the final concentration (the points between 1 and
2 in.). Had the points been corrected to the final concentration, the
curve would have been lower by 8.7%.

 

Figure 1.2 shows a curve of rod effect vs position at the initial
concentration which is the integral of the differential-worth curve in
Fig. 1.1. The curve for the final concentration is simply the first
curve reduced by a factor of 1.087. The predicted worth of this rod at
the initial critical concentration was 2.2%.

We are working on the analysis of the period measurements with the
fuel circulating. As will be discussed later, a new mathematical treat-
ment of delayed-neutron effects in the MSRE has been developed; this will
be used to relate period to reactivity. The sensitivities determined in
this way should, of course, coincide with the results of the static-fuel
measurements. Because of the difficulty of accurately treating the com-
plex pattern of precursor distribution in the circulating fuel, the values
obtained with the fuel static should be more reliable.

ORNL- DWG 65-8033
0.07 :

 

006 1= T e T

0.05 | - . . 4
e \
o .7 ! N | - - \. .
/ .
003 | — ¢ e e Y,

N
./

o
O
2
N\

 

DIFFERENTIAL WORTH  [(% 84/4)/in]

N
N
|

o
=
|
|
|
\

\

\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

|
O 4 8 12 16 20 24 28 32 36 40 44 48 52
ROD POSITION (in.)

O

 

Fig. 1.1. Differential Worth of Control Rod No. 1, Adjusted to
Initial Critical Loading.
19

ORNL-DWG 65-8034

 

e T T

2.0 — AR i
|
|
\
\

1.8 1 . | I I ’ o

el . CRITICAL LOADING

| (61.52 kg 235U IN LOOP) 2/ | e
| | | | | |
! | | } |
- | » | | ST FINAL LOADING N

 

 

 

x

> 1

0

& 235

~ (67.98 kg U IN LOOP)
T ? |

- i

o

O

=

> T N
=

=

}_. -

@)

<

(N R

(et

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 24 28 32 36 40 44 48 52
ROD POSITION (in.)

Fig. 1.2. Integrated Worth of Control Rod No. 1.

The rod-drop experiments are also being analyzed. They will give
independent values of rod worth and may show the effect of 23°U concen-
tration on rod worth. This latter effect was also the subject of multi-
group calculations at the concentrations observed in the experiment. The
correction factor used here is the result of these calculations.

2327 Concentration Coefficient of Reactivity. The 232U concentration
coefficient of reactivity is §iven'by the ratio of the change in reactivity
to the fractional change in 235U concentration (or circulating mass) as a
result of a small addition. The effect of each capsule addition (after
initial criticality) on the critical position of the control rod was de-
termined with the pump running. The critical position with the pump off
was measured after every fourth addition. Rod positions were converted
to reactivity, using Fig. 1.2 and correcting for the concentration effect
on rod worth. Results are shown in Fig. 1.3.

 

The slope of a curve in Fig. 1.3 at any concentration multiplied by
that concentration 1s the desired concentration coefficient of reactivity,
(8k/k)/(8m/m). The coefficient obtained in this way is 0.226, independent
of concentration. The coefficlient predicted from multigroup criticality
searches about the minimum critical concentration was 0.248 (ref. 4).
20

ORNL-DWG 65-803¢

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.4
e
2 e ——— o S e
37 08’
FUEL NOT CIRCULATING ¢ | _&*°

16 T . /;'-.? i;;.?ag e ]
= e
2 ./ <"
~X- 1 2 - - ’;77’{“ ® R - - T
E ./ ,o"..
o ~" " FUEL CIRCULATING

/, .‘.,oo
0.8 LA - - —
/. ...
//. ,o”
0.4 7 ./.‘.‘.; - S I
. /: ....
o8 ..3
O ./ W .'..’
61 62 63 64 65 66 67 68 69

MASS OF 235U IN FUEL LOOP (kq)

Fig. 1.3. Effect of *2°U Mass on oystem Reactivity.

Reactivity Effect of Circulation. The reactivity effect of circula-
tion, given by the difference of the two curves in Fig. 1.3, is —(0.212 =
0.004)% 6k/k. The effect of changes in delayed-neutron precursor distri-
butions with circulation had been predicted to be —0.30% 8k/k.%»© Another
~0.2% dk/k was expected because of entrained bubbles of helium in the cir-
culating salt. As will be discussed later, the evidence shows that there
were practically no circulating gas bubbles except for a brief period when
the fuel level in the pump bowl was lowered. Therefore the gas effect
attending circulation was practically nil.

The difference between the predicted and observed delayed-neutron
losses was apparently due to inadequate accounting for delayed neutrons
emitted just outside the graphite region of the core in the upper and
lower heads. A more realistic model has been developed to account for
these effects. The program computes precursor distributions under both
steady-state and transient conditions, taking into account mixing, ve-
locities, volume fractions, and flux distributions in each of the prin-
cipal regions. The welghted contributions of the delayed neutrons from
each group are computed, taking into account the initial energies of the
neutrons and the nuclear importance of each region. The result is s
"circulating-fuel inhour equation" for the MSRE whose uses will include
the analysis of the circulating-fuel period measurements and, as a spe-
cial case, the steady-state effect of circulation. Application of this
equation to the steady-state condition gave a reactivity effect due to
circulation of —0.22% 8k/k.

Temperature Coefficients of Reactivity. We measured the effect of
temperature on reactivity by adjusting the electric heaters to change the
system temperature slowly (about l5°F/hr) while we observed the critical
2l

position of the regulating rod. This experiment gave the overall tempera-
ture coefficient, that is, the sum of the fuel and graphite coefficients.
We alsc attempted to separate the fuel (rapid) and graphite (Sluggish)
coefficients by an experiment in which the coolant system was used to in-
crease the fuel-salt temperature rather abruptly.

Figure 1.4 shows results of the three experiments involving slow
changes in temperature. The first experiment, with 68 kg of %3°U in cir-
culation, gave a line whose slope ranges from —(6.6 to 8.3) x 10™° (°F)~L.
At 70 kg the experiment gave a straight line with a slope of —7.24 X 107°
(°F)~t. In the last experiment, at 72 kg, the slope of the curve above
sbout 1140°F is —7.3 X 1072 (°F)~t. A value of =7 X 10™° (°F)~1 nad been
predicted: —3.3 X 1072 (°F)~! for the fuel and —3.7 X 10~° (°F)~! for the
graphite. The fuel-salt coefficient of thermal expansion used in this
calculation was obtained by an empirical correlation of density and com-
position of salts other than our fuel salt. Recent measurements of fuel-
salt density gave a higher thermal-expansion coefficient, leading to a
calculated fuel temperature coefficilent of reactivity of —5.6 X 10™°
(°F)"1. The new value for overall temperature coefficient is —9.6 X
107° (°F)~t. The observed coefficient is in better agreement with the
earlier prediction (on which the safety analysis of the MSRE was based) .
The density measurements are belng reviewed and will be checked by an
independent method of density determination.

The experiment at 72 kg 232U shows a lower slope below about 1140°F.
We do not believe that the temperature coefficient is lower in this range;
we believe that another phenomenon became significant during this part of
the experiment. This phenomenon was the appearance of an increasing amount

ORNL-DWG 65-8032R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16 .- . b L [ - i,,,,, fi,,:’ip - ]
e
67.86 kg 23°U IN LOOP ,.¢°
1.4 - - - e e .
/o‘
CD/O

$o b — A
£ 10 |- -~ —t -
N / ..J
3 - al
S 0.8 — — e b e et o ]

235 ° N
71.71kg U IN LOOP e o
7 &7
06 | 4 — e — -
- -~ -~
° o® 235
oa | T weasskg 2PU N Loop
./
o/
"
0.2 e —_— - - e B e
0
1000 1050 1100 1150 1200 1250

FUEL TEMPERATURE (°F)

Fig. 1.4. Effect of Fuel Temperature on Reactivity.
22

of helium bubbles in the circulating salt as the temperature was lowered.
The evidence for this is discussed in the next section. The effect, so

far as the temperature experiment is concerned, was that the bubbles tended
to reduce the amount of fuel salt in the core, compensating to some extent
for the increase in density of the salt itself as the temperature was low-
ered. Thus the slope of the lower part of the curve cannot be interpreted
as a temperature coefficient of reactivity in the usual sense.

The hot-slug transient was done by stopping the fuel pump, raising
the temperature of the circulating coolant salt and the stagnant fuel in
the heat exchanger, and then restarting the fuel pump to pass the hotter
fuel salt through the core. The output of a thermocouple in the reactor-
vessel outlet, logged digitally at 1/4-sec intervals, showed a brief in-
crease of 5 to 6°F as the hot salt first passed. It then leveled at about
3.5°F for a few loop transit times before decreasing gradually. The noise
in the analog-to-digital conversion (£1°F) limited the accuracy of the
measurement, but by taking an average of 50 points during the level period
after mixing and before the graphite temperature had time to change sig-
nificantly, a value was obtained for the step in fuel temperature. Re-
activity change was obtained from the change in rod position, corrected
for the decrease due to circulation, and ascribed to the fuel temperature
increase. The result was a change of —=(4.9 + 2.3) x 10~° (°F)~Lt. Pre-
dicted values of the fuel temperature coefficient lie in this range. We
propose to repeat this test with the thermocouple signals biased and am-
plified to reduce the effect of noise in the analog-to-digital conversion.
More precise results can be expected in this case.

Effect of Pressure on Reactivity. We performed three tests to ex-
plore the effect on reactivity of changing system overpressure. Theo-
retical considerations had indicated that for slow changes a very small,
possibly negative, pressure coefficient of reactivity could be expected,
but for rapilid changes the coefficient would be positive. The existence
of any pressure coefficient was based on the assumption that undissolved
helium would be entrained in the circulating fuel. In each of the three
tests the loop overpressure was slowly increased from the normal 5 psig
to 10 to 15 psig and then quickly relieved, through a bypass valve, to
a drain tank that had been previously vented to atmospheric pressure.

 

The first two tests were carried out at normal system temperature
with the normal operating level of salt in the fuel-pump bowl. No change
in control rod position was required to maintain criticality, and no
significant change in pump-bowl level was observed during either of the
tests. These indicated that the pressure coefficient was negligibly
small and that essentially no helium bubbles were circulating with the
salt. IFurther evidence of the lack of circulating voids was obtained
from a gamma-ray densitometer on the reactor inlet line; this instrument
showed no change in mean salt density during the tests.

The third test was performed at an abnormally low pump-bowl level,
which was obtained by lowering the operating temperature to 1050°F,. Fig-
ure 1.5 shows the pressure transient and the responses of the regulating
control rod, densitometer, and fuel-pump level during the rapid pressure
23

ORNL-DWG 65-8074R

FUEL PUMP LEVEL (%)

 

OVERPRESSURE (psig)

DENSITOMETER

 

CONTROL ROD POSITION (in.)

O 4 8 12 16 20 24 28
TIME (min)

Fig. 1.5. Conditions During Ra@id Pressure Release While Cir-
culating Helium Bubbles.

release. Independent evaluations of the void fraction from these three
parameters all gave values between 2 and 3% by volume. The frequency re-
sponse characteristics of the effects of pressure on reactivity were cal-
culated from the pressure and rod motion and are shown in Fig. 1.6, along
with the predicted high-frequency response for a void fraction of 1.2%.
Extrapolation to the observed curve gives a void fraction of 2 to 2-1/2%.
The low- and high-frequency pressure coefficients were +0.0003 and +0.014
(% 6k/k)/@si, respectively, for this particular condition.
24

ORNL-DWG 65—-8904A
1073

® FREQUENCY VALUE
o PREDICTED FOR 1.2 %

VOID MN

 

1076
0.0001 0.004 0.01 0.1 1 10
w, FREQUENCY (radians/min)

Fig. 1.6. Reactivity-Pressure Transfer Function with 2 to 3%

Void Volume in Fuel. Calculated from pressure release experiment
using Samulons method with 0.2-min sampling interval.

Dynamic Tests

 

We performed a variety of dynamic tests during the operation at zero
power. These tests were the start of an extensive program to evaluate
experimentally the inherent nuclear stability of the MSRE at all power
levels. The reactor system has been analyzed on a theoretical basis,
and the tests are designed not only to characterize the present system
but also to evaluate the techniques and mathematical models used in the
theoretical analysis. Preliminary results from the analysis of the zero-
power tests are presented below.

Frequency Response Measurements. A series of tests was run to de-
termine the frequency response of neutron level to reactivity perturba-
tions. These experiments included pulse tests, pseudorandom binary re-
activity-perturbation tests, and measurements of the inherent noise in
the flux signal. Tests were run with the fuel pump on and with it off.
Noise measurements were also made during the special run with low pump-
bowl level, when there were entrained bubbles in the core.

 

The frequency response is a convenient measure of the dynamic char-
acteristics of a reactor system. Classically, the frequency response
is obtained by disturbing the reactor with a sinusoidal reactivity per-
turbation and observing the resulting sinusoidal neutron-level variations.
The magnitude ratio i1s defined as the ratio of the amplitude of the output
sinusold to that of the input sinusoid. The phase angle is defined as
the phase difference between the output sinusoid and the input sinusoid.
Other procedures, such as those described in this report, can be used
to yleld the same results as the classical method with less experimental
effort.
25

The zero-power frequency response tests serve to check the theoret-
ical zero-power frequency response predictions, but they do not furnish
direct information on the stability of the power-producing reactor. The
zero-power tests, however, do serve as an indirect, partial check on the
at-power predictions because the dynamic behavior at power is simply the
zero-power case with the addition of feedback from the system. Thus, ver-
ification of the zero-power kinetics predictions lends some support to the
predictions regarding power operation.

In the pulse tests a control rod was withdrawn 1/2 in., held there
for 3-1/2 or 7 sec, and then returned to its original position. The rod
was placed so that this rod motion caused a change in kgope of about 0.03%.
The rod-position signal and the flux signal were recorded digitally at 0.25-
sec 1ntervals using the MSRE on-line digital computer (referred to in Figs.
1.13 and 1.14 as "logger"). The frequency response Was obtained by numer-
ical Fourier transformation of the input and output signals.

The pseudorandom binary tests consisted of a gpecially selected series
of positive and negative pulses. The series contained 19 bits each with
a duration of 7 sec and was repeated several times so that the initial
transients could die out. The rod motion about the average position cor-
responded to a Mkgrp of about *£0.015%. The rod-position signal and the
flux-level signal were recorded digitally at 0.25-sec intervals using the
MSRE computer. The frequency response was obtained by two different
methods. The direct method used a digital filtering technique to obtain
the spectral density of the input and the cross spectral density of the
input and the output. The frequency response of the system is the ratio
of the cross to input power spectral densities. The indirect method in-
volved calculation of correlation functions and subsequent numerical Fou-
rier transformation. Both methods gave essentially the same results.

The nolse measurements and analyses used direct analog filtering of
an analog tape record of the inherent noise in the flux-level signal. The
tests required 1 hr of data recording to give statistically accurate re-
sults. These tests were hampered by the unfavorable location of the de-
tector and the resulting low flux-signal level at low power QYlO'w).

Preliminary results of the frequency response measurements are shown
in Figs. 1.7 through 1.11, along with theoretical predictions for the runs
with the fuel pump on and with the pump off. As shown on the legend in
the figures, several different procedures were used to obtain thesgse re-
sults, but this represents only a partial analysis of the available data.
All the experimental points except for the noise-analysis results are in
absolute units (fractional change in neutron flux per change in keff)'
The noise-analysis results are based on the assumption of a white-noige
input and contain an unobtainable proportionality constant. Thus, the
noise-analysis data were arbitrarily multiplied by the factor required
to give agreement with theoretical results at a frequency of 9 radians/
sec.

Figure 1.12 shows the noise-analysis results for operation at a re-
duced pump-bowl level, which caused increasgsed bubble entrainment in the
20

fuel salt. Comparison of the nolise spectrum obtained with bubbles cir-
culating (Fig, 1.12) and the noise spectrum obtained with no bubbles cir-
culating (Fig. 1.7) indicates that the bubbles increased the amplitude

of the power spectral density significantly in the 1- to lO—radian/sec
region. Previous experiments with the MSRE-core hydraulic mockup indi-
cated that random, hydraulically induced pressure fluctuations in the

core would probably cause a significant modulation of the core void frac-
tion, thus causing reactivity fluctuations. Hence, additional flux noise
in this frequency range was expected, although it was not possible to pre-
dict the "shape' or characteristics of this spectrum.

ORNL—DWG 65—8905A

104

SN/,
Sk/ kg

 

 

® BINARY TEST — INDIRECT ANALYSIS
A BINARY TEST — DIRECT ANALYSIS
o NOISE ANALYSIS

 

10
0.004 0.01 0.1 4 10 100
w, FREQUENCY (rodicns/sec)

Fig. 1.7. Frequency Response — Magnitude Ratio of (SN/NO)/
(6k/ko) with Fuel Circulating. Results of binary and noise tests.

ORNL-DWG 65-8906A
104

® TEST NO.1
© TEST NO. 2

103

 

‘SAUM@
S/ kg

 

102

 

10"
0.001 0.0 0.1 1 10 100
w, FREQUENCY (radians/sec)

Fig. 1.8. Frequency Response — Magnitude Ratio of (BN/NO)/
(8k/ko) with Fuel Circulating. Results of two 7-sec-pulse tests.
27

The information presented in Figs. 1.4 through 1.8 was obtained from
the data by stralghtforward analysis, using basic techniques. The pulse
test data were filtered prior to analysis, but the binary test data were
used in unmodified form. A more sophisticated analysis, including smooth-
ing and trend removal, is being made, but little change in the results
is expected. The noise test results will be corrected to remove the in-
fluence of the tape-recorder characteristics, which may be significant at
frequencies below 6 radians/sec.

ORNL—DWG 65—-8907A

10 o 7—sec PULSE, TEST NO.1
A 7—sec PULSE,TEST NO.2
o BINARY-INDIRECT ANALYSIS
-10 s BINARY-DIRECT ANALYSIS

PHASE (deg)

—-80
—90

 

—-100
0.001 0.0/ 01 1 10

w, FREQUENCY (radians/sec)

Fig. 1.9. Frequency Response — Phase of (5N/NO)/(6k/ko) with
Fuel Circulating. Results of pulse and binary tests.

ORNL—DWG 65—8908A

O 7-sec PULSE, TEST NO.1
A 7—sec PULSE, TEST NO.?2

a 3-1/p—sec PULSE, TEST NO.3
® 2—'/>5—sec PULSE, TEST NO.4
O NOISE ANALYSIS

 

0.001 0.01 0.1 1 10 100
w, FREQUENCY (rodions/sec)

Fig. 1.10. Frequency Response — Magnitude Ratio of (5N/NO)/
(6k/ko) with Fuel btatic. Results of pulse and noise tests.
28

ORNL-DWG 65-8909A
20

-20

-40

PHASE (deg)

e 7sec PULSE, TEST NO. 1
60 4 Tsec PULSE, TEST NO.?2

5 3% sec PULSE, TEST NO.3

_80 ° 3% sec PULSE, TEST NO. 4

 

-100
0.001 0.01 0.1 1.0 10

w, FREQUENCY (radians /sec)

Fig. 1.11. Frequency Response — Phase of (6N/NO)/(SK/RO) with
Fuel Static. Results of pulse tests.

ORNL-DWG 65-8910A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10° T T T T i T T T T ]
R S SN/ Ny |
THE FREQUENCY RESPONSE S_k/T IS PROPORTIONAL
5 _
TO A/PSD IF THE DISTURBANCE IN 84 IS WHITE NOISE |
51—y —— -
® . _ |
O
N
g -
[ |
5 _
[ ]
® 9
o N °
102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 2 5 10 20 50 100
w, FREQUENCY (radians/sec)

Fig. 1.12. Frequency Response — Square Root of Power Spectral
Density (PSD) of Flux oignal — Fuel Circulating with Bubbles. Re-
sults of noise analysis; normalization same as for noise data in Fig.
1.7.

The experience with these "zero-power" dynamics tests has led o
plans for more tests with some improvements in data recording.

Transient Flow-Rate Tests. The purposes of the transient flow-rate
tests were: (1) to obtain startup and coastdown characteristics for fuel-
and coolant-pump speeds and for coolant-salt flowrate; (2) to infer fuel-
salt flow-rate transients from the results of (l); and (3) to determine
the transient effects of flow changes on reactivity and void fraction.
29

Figures 1.13 and 1.14 show the fuel-pump speed, coolant-pump speed,
and coolant-salt flow rate vs time for pump startup and coastdown. Data

were taken

with the computer and with a Sanborn oscillograph. The output

of a differential-pressure cell across the coolant-salt Venturi was re-
corded directly on the oscillograph, and the square root of that signal
was taken as flow rate. The lag in the response of the computer flow

signal is due to the response characteristics of the emf-to-current con-

verter and

the square-root converter between the differential-pressure

transmitter and the computer input.

PERCENT OF FULL SCALE

]

ORNL—DWG 65—8911A
120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 —o-a= 5 O—O0—O0—
FUEL e qp—‘—.—.-fi 8—0—-0‘ 8-g—9-8 e §==o
PUMP /COOLANT ’O,o-"’)/),ozo’(
W SPEED PUMP o®’ O/O/
z[l 80 e 1.0 SPEED !‘v’.,
| @] PYd
3 7/ / ¥ /
— o
> /° a4
L ® [
60 f—f /-
5 s T
- /g> ./ />'COOLANT FLOW RATE
m © e /
E—:) 40 / o (
i O
) / /
/ ’ /
20 —e45 / o LOGGER
/° /" ® OSCILLOGRAPH
¢ / | \ ’
0 §—o—G%-
0 1 2 3 4 5 6 7 8 9 10
TIME (sec)

Fig. 1.13. Pump Speed and Flow Startup Transients.

ORNL-DWG 65-8942A

 

 

 

 

 

100 ¢mgz— J } .
NN | | | |
\ COOLANT SALT FLOW RATE

o LOGGER

k\\ %\:E\\\ e OSCILLOGRAPH

COOLANT PUMP SPEED

;;;fif

 

 

 

 

D
O

 

 
 
 
 
  

 

 

 

 

 

 

 

 

 

 

 

0 i 1 O g ‘ 3 L -
O 2 4 o 8 10 12 14 16 18 20
TIME (sec)

 

Fig. 1.14. Pump Speed and Flow Coastdown Transients.
30

It was hoped that the coclant-pump speed and coolant flow rate would
coast down in unison so that the fuel flow-rate coastdown could be inferred
directly from the fuel-pump speed coastdown curve. This was not the case,
however. Other methods of analysis will be attempted later.

Reactivity effects of fuel flow-rate transients were measured by
letting the flux servo controller hold the reactor critical during the
transients; the reactivity added by the rod is then equal (and opposite)
to the reactivity change due to the flow perturbations. The data for the
pump startup were taken on the computer but were inadvertently erased; the
reactivity transient for the pump coastdown was recorded. Since there
were no voids in the fuel loop during normal operation, this transient
1s due entirely to flow effects on delayed neutrons. A further analysis
of this curve will be made to try to determine the flow coastdown tran-
sient and to check on the model used to represent the circulating delayed-
neutron precursors.

Conclusions from Dynamic Tests. The two most significant conclusions
to be obtained from the dynamic tests are: (1) the information obtained
gives no indication of the existence of inherent characteristics that
might lead to operating difficulties in the low-power runs, and (2) the
selected tests were, on the whole, quite satisfactory. These tests gave
results which show good agreement with theoretical predictions, giving
increased confidence in the theoretical model and in the predictions for
stable power operation. ©Since the zero-power tests of this type are al-
ways more difficult than power-level tests, very good results are expected
from later tests.

 

Undissolved Gas in Fuel Loop

 

Although the nuclear experiments showed that there is no undissolved
gas clrculating in the fuel loop when the salt is at the normal level in
the fuel-pump bowl, the behavior of the pump-bowl level on stopping and
starting the pump suggests the presence of a small compressible volume
somewhere in the loop. Under normal conditions the level change corre-
sponds to 0.4 ft° of gas uniformly distributed around the loop or 0.23
ft> located at the point of maximum pressure. The only known trapped
gas is about 0.14 ft2 in the annuli at the reactor access nozzles. At-
tempts to resolve this anomaly have been inconclusive.

oalt Density and Inventory Experience

 

The weigh cells on all the salt drain tanks were calibrated with
lead weights shortly after the equipment was installed and before the
Tanks and connected piping were heated. Additional data were obtained
with the tanks hot during various salt-charging and transfer operations.
These data served both to recalibrate the weighing systems and to give
a measure of the density of the salts at operating temperature. Through-
out the operation, salt inventories were computed from weigh-cell readings,
using scale factors and tare corrections obtained from the calibration
tests.
31

Cold and hot calibration of the coolant drain-tank weigh cells gave
scale factors differing by less than 0.5%. Fuel drain tank 2 (FD-2) was
calibrated hot twice, with flush salt and with fuel carrier salt; scale
factors were within 0.2% of each other but were about 4% higher than the
original, cold calibration. The reason for this discrepancy has not been
established.

The coolant-salt density was measured in the reactor by three dif-
ferent methods; values ranged from 121.3 to 122.3 lb/ft3 at 1200°F, with
an average of 121.9 1b/ft>. When flush salt, which is identical with
coolant salt, was charged into ¥FD-2, the amount of salt added between two
level probes indicated a density of 124.5 1b/ft2. A density of 120.9
lb/ft3 was computed from the pressure required to 1ift salt from the
drain tank to the fuel loop. Weigh-cell indications of flush-galt den-
sity, using the "hot" calibration factor, ranged from 123.2 to 131.5
1b/ft> at 1200°F; the "cold" calibration factor would have given 118.4
to 126.3 lb/ft3. The data on coolant- and flush-salt densities thus tend
to support the "cold" calibration factor for FD-2.

The density of the fuel carrier salt (65 LiF—30 BeFo—5 ZrF,) was
measured as the salt was being charged to FD-2. This measured density,
computed from externally measured weights and the volume between the level
probes in FD-2, was 140.6 1b/ft> at 1200°F. Addition of all the uranium
added through run 3 would be expected to increase the density by about
5.3 1b/ft>. Four measurements were made after the uranium was added,
using- the weigh cells and the level probes. Densities based on the "hot"
calibration of the wei§h cells ranged from 149.9 to 152.2 1b/ft>, with an
average of 151.0 1b/ft> at 1200°F. With the "cold" scale factor the av-
erage was 145.1 1b/ft°, very close to the expected density.

oalt densities were computed on several occasions from the change in
welgh-cell readings as the fuel loop was filled. In every case the com-
puted density was less than given by other means, suggesting that a full
loop volume may not have been transferred.

The temperature coefficients of density for the salts were computed
from the change in salt level with loop temperature. Measured values
of (Ap/p)/bm were: for the coolant salt, —1.06 X 1074 (°F)—1L (average
of three measurements); for the flush salt, —1.15 X 10—% (°F)“l; and for
the fuel salt, —1.09 X 107% and —1.15 X 107% (°F)~! (two measurements).

The bulk of the inventory data accumulated to date is on the flush
salt, because more transfer and fill-and-drain operations have been done
with this salt. Calculated inventories (using "ot" scale factors) have
ranged from 1.7% below to 2.6% above the nominal or "book" inventory for
no ascribable reason.

Ixternal Neutron Source

 

The MSRE fuel contains an internal neutron source (from the inter-
action of 224U alphas with beryllium and fluorine) that provides about
4 X 10° neutrons/gec in the core when the reactor vessel is full. This
32

source fulfills all the safety requirements imposed on a reactor neutron
source. However, it is highly desirable to monitor closely the operation
of filling the reactor to ensure an orderly sequence of operations. This
requires an external neutron source that will produce significant counting
rates on the nuclear instruments before the filling operation is started.
The geometry of the MSRE is such that an extremely strong neutron source
is needed to satisfy this requirement. (The source tube is in the thermal
shield on the opposite side of the reactor from the neutron detectors.)

The Am-Cm-Be source that was procured was expected to produce about
10 counts/sec on the fission chambers with the reactor vessel empty of
salt. This would have given the source a useful life of one year, with-
out regeneration, for monitoring reactor fills. (A minimum of 2 counts/
sec has been specified as a criterion for starting a fill.) However,
the source, as delivered, produced only 2.5 counts/sec,'which made 1t
adequate for the initial operation but inadequate for future fills.
oince an even stronger source is impractical, because of handling dif-
ficulties and cost, it is expected that some instrument changes will be
made to permit adequate monitoring of the early stages of reactor fills.

Instrumentation and Controls Design and Installation

 

General

Ixcept for a small amount of work on the Fuel Processing oystem
sampler, completion of documentation, and a few minor additions and re-
visions, the design, installation, and checkout of the MSRE Instrumenta-
tion and Controls System are now complete. During the period since the
last report7 the design, installation, and checkout of all instrumenta-
tion and controls systems required for power operation of the reactor were
completed, the data-logging and processing system, referred to as the
data-logger—computer system, was installed and checked out, several re-
visions and modifications were made to improve performance or correct
errors in design, and some safety instrumentation and associated control
circultry were added. Documentation of design and as-built changes is
nearing completion. In general, performance of the instrumentation has
been very good. Some failures and malfunctions have occurred; however,
once the causes of the failures were determined, these troubles were
easlly corrected, and no major changes in instrumentation or in design
philosophy have been necessary.

Design, Installation, and Checkout

Vapor-Condensing System. Design and procurement of instrumentation
for the vapor-condensing system were completed. Instrumentation was pro-
vided to monitor the water level, pressure, and temperature in the tanks.
Commercial devices were used for all applications in this system; how-
ever, since the vapor-condensing system is part of the secondary contain-
ment, some special design and development were required to maintain the
integrity of the containment. Water level is monitored by a four-position

 
33

float-type switch. DPressure is monitored by a local gage and pressure
switch. Temperature is monitored by thermocouples. Abnormal conditions
produce alarms at the vapor-condensing system and in the main control room.
A dip tube was installed for use in continuous measurement of water level
during shutdown. Installation of this system is in progress and will be
complete before the start of power operation.

Cell Air Oxygen Analyzer. Design, procurement, and preliminary
checkout of a cell air oxygen analyzer system were completed. This sys-
tem, which will be used to measure the amount of air leakage into-the
nitrogen-filled secondary containment system, is required to detect a
change of 0.02% (by volume) of oxygen over the range of 4.9 to 4.1%. The
instrument supplied is capable of measuring oxygen content with an accuracy
of 1% of full scale. Two ranges are provided (0-10% and 0-25%). When the
O~25% range is selected, the oxygen content of the cell air can be moni-
tored over the range from normal air (approximately 22%) to, and below,
the normal 5% operating level. The requested sensitivity (0.0Z%) is ob-
tainable over the full range of the instrument when a potentiometer-type
device is used to read the signal. A continuous sample of the cell air
will be obtained by connecting the analyzer across a restriction in a
component cooling air system line. The component cooling air system ob-
tains its air from the reactor cell and is part of secondary containment.
oince the construction of the analyzer would not satisfy the requirements
of the containment system, and since a possibility existed that the sampled
alr could be contaminated, a safety-grade block-valve system was installed
to maintain the integrity of the containment and to protect the operator.
Component parts of the analyzer system were obtained from commercial
sources, and the system was assembled and checked out at ORNL.

 

Figure 1.15 shows the analyzer cabinet. Figure 1.16 shows the block-
valve header assembly. Installation of this system in the MSRE vent house
1s nearing completion. Final checkout and operational testing will be per-
formed before the start of power operation of the reactor.

Fuel-Processing oystem. Fabrication, installation, and checkout of
instrumentation and controls for the fuel-processing system were completed.
Design efforts on this system during the past report period involved the
completion of design, fabrication, and calibration of special flow ele-
ments used in measurement of Fp, HF, and He gas flow and revision of in-
strumentation required to improve performance and to satisfy safety re-
quirements.

 

Controlled Test Rig for Surveillance Program. The instrumentation
and controls design for the surveillance program controlled test rig8 wa.s
completed. The rig will be used to expose samples to conditions which,
except for nuclear radiation, approximate the conditions experienced by
similar samples installed in the MSRE core. Since it is desired that the
out-of-pile samples have the same temperature history as the in-pile
samples, and since 1t 1s not practical to measure the temperature of the
in-pile samples directly, on-line computer control of the test rig heaters

 
3

PHOTO 72554

 

Fig. 1.15. Reactor Cell Oxygen Analyzer Cabinet — Front View.
35

PHOTO 72556A

HAND SOLENOID HAND SOLENOID
SHUT—-OFF BLOCK SHUT-OFF BLOCK

VALVE == VALVES = VALVE VALVES »om
] LEAK TEST LEAK TEST ; LEAK TEST /| LEAK TEST

BLOCK = BLOCK =~ BLOCK / BLOCK

   

  

OUTLET

Fig. 1.16. Reactor Cell Oxygen Analyzer — Block Valve Header
Assembly.

will be used. ©OSet points, error signals, and control functions (propor—
tional, reset, and derivative actions) will be developed in the MSRE data-
logger—computer.

oet points (simulated sample temperatures) will be computed using ex-
isting temperature and power input information to the computer and reactor
temperature profile equations. Error signals will be obtained by computing
the difference between the computed set polnts and computer inputs obtained
from measured temperatures in the test rig. The 10- to 50-ma control
signal supplied by the computer (using the existing analog output capa-
bilities of the data—logger—computer) will be transduced to a pneumatic
signal which will drive the Variacs which control the test rig tempera-
tures.

Three zones of heat control are required on the test rig; therefore,
three control channels are used. To prevent complete loss of control in
the event of computer fallure and to permit completely manual operation
of the test rig, the computer control was superimposed on a manually ad-
justable base heat control, and the amount of control allotted to the
computer was limited to that required to vary the test rig over the nor-
mal range of reactor sample temperature variations.

Due to the inherent flexibility and capability of the MSRE data-
logger—computer, the only difference in equipment requirements (other
than the computer) between the computer-controlled system and a strictly
manually controlled system was three pneumatically operated Variacs and
three current-to-pneumatic converters.

Fabrication of the control panels has been completed, and installa-
tion of equipment in the MSRE high-bay area is in progress.

Closed-Circult Television for Remote-Maintenance Operations. Design
of the television system installation and design and fabrication of the
console assembly were completed. Figure 1.17 shows the console. Although

 
36

PHOTO 72443

 

Fig. 1.17. Remote Maintenance Television Control Console.
37

PHOTO 72445

 

Fig. 1.18. Radiation-Resistant Camera and Accessories for Remote
Maintenance Television System.

only two monitors are used, three camera systems are installed. A video
switching system permits the operator to display the signal from any of
the three cameras on either or both monitors. The "Joy Stick" controls
mounted on the front of the console table enable the operator to control
pan, tilt, focus, and zoom operations on two of the three cameras with
wrist and finger motion. Other, less frequently used controls and ad-
Justments are located on the sloping panel in front of the operator.
opace was provided on the table for addition of crane controls. The de-
sign of this system was strongly influenced by the limited space avail-
able in the maintenance control room. Figure 1.18 shows one of the ra-
diation-resistant cameras with a radiation-resistant zoom lens and a pan
and tilt unit. The complete system was tested after assembly to demon-
strate performance and reliability. A number of minor failures occurred
during the first few weeks of operation; however, since these troubles
were corrected, the performance and reliability have been excellent.

Fuel Sampler-Enricher. Installation and checkout of instrumentation
and controls for the fuel sampler-enricher system were completed.

 
38

Nuclear Instrumentation. Installation and checkout of the nuclear
instrumentation were completed.

Data System. Fabrication of the data-logger—computer was completed
at the vendor's plant in early February and released to the programmers
for program loading and checkout. Starting about the middle of February
and continuing through March and April, the system was checked out at the
manufacturer's plant by both ORNL and the wvendor!'s (Bunker-Ramo) personnel
The checkout consisted in debugging and running all the system programs
written by ORNL and Bunker-Ramo. The programs and system operation were
verified as meeting the preliminary acceptance requirements except for
the analog input system, which could not be effectively checked because
of the lack of live input signals, and for the computer power failure
circuitry, which was not installed.

*

The system was shipped to ORNL and arrived gbout May 1. The instal-
lation of the system, including the connection of the input signals, was
completed in three weeks. All necessary building modifications and in-
stallations of power distribution wiring, signal input transducers, and
signal input wiring were completed prior to delivery of the system. The
installed system is shown in Figs. 1.19 and 1.20. In the front row of

PHOTO 80019

 

Fig. 1.19. MSRE Computer—Data-Logger System Console, Typewriters,
and Tape Units.
39

 

Fig. 1.20. Computer, Typewriters, and Tape Units of the MSRE
Computer—Data-Logger System.

cabinets the first three contain the computer, and the last contains the
power supplies. Five cabinets are located in a row behind the cabinets
shown in Fig. 1.20; they contain the input relays, amplifiers, analog-to-
digital converter, and the associated control circultry.

After the installation was complete, the programs were loaded and
The system was operated with live input signals to test hardware, pro-
grams, and input signals. These tests were continued from the middle
of May through most of June. Numerous hardware and some program dif-
ficulties were corrected. Also, some difficulties with the input signals
were found and corrected. Most of the system problems during this period
were caused by the analog input equipment, which includes the input relays,
amplifiers, analog-to-digital converter, and the associated control cir-
cultry. These problems seemed to result from inadequate design and test-
ing of the input system, which is unique and which was built special for
the MORE installation.

On June 24, after many corrections and modifications of the hard-
ware, the system was initially accepted with the provision that it op-
erate continuously for one week without an error, excluding errors caused
40

by typewriters or magnetic tapes. The system ran for six days before fail-
ing. From then until the middle of August the test was attempted many
times but usuvally failed after a day or two. Many problems were found in
the hardware, usually the analog input equipment. Attempts to run the

test continued until August 16, when the system was shut down at the ven-
dor's request to install extensive modifications intended to improve re-

liability. These modifications are scheduled for completion about Septem-
ber 1.

During the reactor critical experiments in June and July, the data
logger and computer were used to collect and process some of the data
from the reactor dynamics, rod drop, and pressure coefficient of reac-
tivity experiments. The data system was not completely effective during
these experiments due to hardware failures and poor signals from the re-
actor instruments. However, some benefit was obtained from its use.

MSRE Training Simulators. Two reactor kinetics simulators were de-
veloped for the purpose of training the MSRE operators in nuclear start-
up and power level operation procedures. (The startup, or zero power,
simulator was run in February 1965; the power level simulator will be run
in September 1965.) Both were designed to be tied in to the MSRE reactor
control and instrumentation system. The "on-site" simulator has several
advantages over one set up in a computing facility: (1) the operators
get used to the actual instrumentation and controls system; and (2) much
of the actual hardware, such as control rods, is used rather than simu-
lated. The main disadvantage of the on-site simulation is that less com-
puting equipment is gvailable, so the simulation cannot be as accurate as
with the off-site simulator.

 

The startup simulator used the control rod position signals as in-
puts, and provided outputs of log count rate, period, log power, and
linear power. The reactor's period interlocks, flux control system, and
linear flux range selector were also operational.

The power level simulator will include the effects of flux on tem-
peratures. Radiator door position and cooling ailr pressure drop signals,
as well as control rod positions, will be used as inputs, and the usual
nuclear information plus key system temperature outputs will be read out
on the reactor instrumentation. The reactor's servo controller and ra-
diator load control systems will be used.

Both simulators are set up on general-purpose, portable EAI TR-10
analog computers.

Instrumentation and Controls System Performance

 

Performance of the MoRE Instrumentation and Controls System has con-
tinued to be very good. As systems become operational and as operating
time was accumulated, additional minor troubles with instrumentation oc-
curred; however, once the causes were determined, the troubles were easily
corrected and no major changes in instrumentation or in design philosophy
have been necessary.
41

Weigh Systems. In addition to the calibration drift previously re-
ported,? difficulty was experienced with the multiposition pneumatic se-
lector switches, and a failure occurred in one of the weigh cells. The
difficulty with the selector switches, which are used to select signals
from any one of the ten weigh cells for precision readout on mercury ma-
nometers, was determined to be mechanical in nature and was eliminated
by a redesign of the switching mechanism. The weigh cell failure was
determined to be due to pitting of the baffle and nozzle in the cell.
This pitting was apparently caused by amalgamation of mercury with the
plating on the baffle and nozzle. How the mercury got into the welgh
cells has not been determined; however, it is presently believed to have
come from the manometers and to have been precipitated on the baffle by
expansion cooling of the air leaving the nozzle. Appreciable quantities
of mercury were also found in the tare pressure regulators on the con-
trol panel; however, no mercury was found in the interconnecting tubing
or in other portions of the system. Several methods of preventing this
problem from reoccurring in the future (including replacement of the ma-
nometers with precision gages) are under consideration.

 

Temperature Scanner. The scanners continued to operate satisfacto-
rily until March, when four differential amplifiers failed. This failure
was caused by overheating due to failure of the fans in the amplifier
cabinet. The amplifiers were completely rebuilt, since many of the com-
ponents were burned up. The two fans were repaired, and the system was
restarted. However, the ambient temperature in the cabinet was still
around 140°F, partly due to the high ambient temperature in the heater
panel control area. To remedy this, a flexible hose from the control
room was brought down to the input of the amplifier cabinet so that cool
alr from the control room was pulled through the amplifiers by the two
cabinet exhaust fans. This seemed to improve scanner operation consid-
erably, and no further amplifier failures have occurred. A better ar-
rangement of this cooling system is nearing completion.

 

The thermocouple scanner reference voltage supply for the two ra-
diator scanners was installed and checked out.

Before reactor power operation begins, all the mercury switches will
be cleaned and checked and the complete scanner system checked and cali-
brated. An additional 17-in. oscilloscope is being modified for use as a
spare.

Drain-Tank Level Probe Power Supply. Failures due to overheating
also occurred in the drain-tank level probe power supply. In this case
the damage was limited to vacuum tube failures. This problem was elimi-
nated by reducing the load on the supply and by providing better ventila-
tion. It should be noted that the level probe power supplies and the
scanner amplifiers are vacuum tube equipment. No failures due to over-
heating have occurred in any of the solid-state equipment used in the
MSRE .
42

Thermocouples. Performance of the 1033 thermocouples in the MSRE
system continues to be excellent. There were no known thermocouple
failures during critical and low-power operations. An apparent sensi-
Tivity of several thermocouples, located inside the containment, to cell
ambient temperature was determined to have been caused by a double re-
versal of the Chromel and Alumel extension wire leads at the disconnects
and at the out-of-cell Junction box.

 

Manually Operated Helium Throttling Valves. Considerable difficulty
was experienced by the MoORE operators in setting the bellows-sealed hand
valves, which control helium flow to the fuel and coolant pump bowl level
systems, to maintain a constant flow of 366 cm’®/min. The valves were dis-
mantled during the precritical shutdown and found to have been contami-
nated with oil and metal particles. It was also found that the trim had
"on-off" instead of throttling characteristics. The valves were cleaned,
and new trim with throttling characteristics was fabricated and installed
in the valves. Performance of these valves was satisfactory during crit-
ical and low-power operation. The source of the contamination has not
been determined.

 

Component Cooling Air System Control Valves. As previously reported,lo

cooling air flow to the freeze valves was found to be inadequate during
precritical operations. Air flow to the control rods and reactor access
nozzle was also found to be inadequate, and flow to the pump bowl shroud
was fTound to be excessive.

 

Measurements made during the precritical shutdown showed that the
sizing of all of the freeze valve cooling air control valves wasg adequate.
In most cases the low flow was due to restrictions in the lines or freeze
valve shrouds. The low flow to the coolant-salt system freeze valve was
found to be due to improper adjustment of the cooling air control valves.
The sizing of the cooling air control valves for the fuel drain freeze
valve (FV103) was found to be adequate to supply the required blast air
flow; however, the control during the "hold" mode of operation was very
poor because the air flow needed was so low that the valve was forced
to control on the seat. This problem was successfully corrected by in-
stalling specially designed trim in the valve.

The low flow to the reactor access nozzle and control rods was found
to be due, in part, to the equal-percentage characteristics of the valve
together with insufficient actuator stroke. This problem was corrected
by reshaping the valve trim to obtain the full rated capacity of the wvalve
with the available operator stroke.

Reduced area trim was installed in the fuel pump bowl cooling air
control valve to obtain satisfactory control at the new (reduced) air
flow requirement.

oome trouble was experienced with hysteresis in the specially de-
veloped valve actuator motion multipliers.ll This problem was eliminated
by increasing some clearances in the multiplier assenbly and by adding
43

a floating coupling between the valve stem and the multiplier to provide
greater freedom of motion.

Helium Control Valve. Except for some trouble with plugging of the
fuel and coolant system letdown valves, performance of the helium con-
trol valves has been satisfactory. There have been no failures of these
valves due to galling since the start of precritical operations. The
cause of plugging in the letdown valves was determined to be deposition
of particles of carbon and salt on the valve trim. The sizes of these
valves were increased to reduce the effect of the deposits.

 

Revisions and Modifications

 

Nuclear Safety Instrumentation. The nuclear safety system was modi-
fied by the addition of period safety amplifiers in each of the three
channels; see Fig. 1.21. The monitor and test unit was redesigned to
include in-service testing of these.

 

A second modification to the safety system lowers the flux scram
level by a factor of 1000 when fuel salt is not being circulated by the
pump. A measurement of the three-phase current to the pump provides
the input information to accomplish this.

Initial criticality tests disclosed that the neutron flux attenua-
tion in the instrument penetration departed from the ideal exponential
curve. A flat, or nearly flat, region in the attenuation curves prej-
udiced operation of the wide range counting instrumentation. Addi-
tional shielding is being designed for the penetration to obtain the
desired characteristics.

Control and Alarm Circuits. ©Safety-grade circuits were added to
lower the flux scram level when fuel salt is not circulating, to pre-
vent the occurrence of excessively low pressures in the containment
which might damage the reactor containment vessel, and to block the
sample lines that comnect the reactor cell oxygen analyzer to the sec-
ondary contalnment system in the event of high reactor cell pressure or
air activity.

 

The center heaters and the high (1250°F) center temperature inter-
locks and alarm were removed from all freeze valve circuits. The siphon
break and permissive-to-thaw circults were revised to make clear the
causes for alarms. On freeze valves 103, 104, 105, 106, 204, and 206,
the "Freeze Valve Frozen' logic circulitry was changed from one-of-two
to two-out-of-three to prevent false operation of control interlocks
when normal control actions occurred.

A number of revisions were made 1n control-grade circuits for the
purpose of correction minor errors in the design, improving performance
of the system, or providing additional alarms. Ixamples of these changes
are: removal of drain-tank pressure interlocks from the prefill mode cir-
cuits, installation of additional jumpers on the jumper board, revision
4ty

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

"WBJISBI(Q 00T WBIDG POY TOJIJUO) HMSW °TZ°T "ITd
¢S
(¢401no2) [ 5 ,
INIWIT3 ¢ ON
XIHLVW AV 13 ft— 35S —— ¢ STINNYHO
30N30IONI0D < €S ALIIVS =—P— g gy |
\ ><l_mm A..Ilm—\/_l ..N _\/_OW_IJ_
NYY Y
N - .
(¢40 1n0 2) [*=° iNawaTa Tt mdmzmmzo oLime 43 L MAANOD |
XIYIVN AVI3Y le—2S ¢ —— 2S5 ALIAVS ja—p— ] LN3FHYND —~
39NIAIONIOD ~ $ ANV L OINOY10373 OL 4IN3
.« 1S — AVIIE L 2| 2 woyd 37dNOOOWY IHL
WYYV
H
HOLVHYHINOD AL34VS
digl aoIy3d
-GS I~
€ 40 1N0 2] 2S Fz.“_smdAl v | _
XIY1VN AVI3Y [ 2
JON3JIONIOD 1S —{ AL3AVS fe— Vb moflwmwzoo “ " v J
-t— | S
AV 13 | | 43IGWYHD NOI
| _ JAILISN3S
SIHOLNTD | | NOYLN3IN
IAIYA QO , JOLINOW Alddns | |
W JOVLI0N | MO _
43GNYHD _ |
NV Y L

V6.9 -9 OMA-INLO

 

 

STINNVYHO
INdNI AL3IVS
;- J9NLVY3IdINTL

NV
XN713 IVOIdAL

 
45

of the radiator-backflow damper circuits to permit the natural convection
circulation of air through the radiator required for low-power operation,
and addition of alarms to annunciate high temperatures in the control rod
housings.

Freeze Valve Hold Air Controllers. Instrumentation was added to pro-
vide automatic (proportional plus reset) control of shoulder temperatures
on five freeze valves (103, 105, 106, 204, and 206) located in the fuel
and coolant drain system.

 

Thermal Shield Overpressure Protection. A self-regulating flow con-
trol valve was installed, and revisions were made in the control circuits
to prevent the reactor thermal shield cooling water pressure from increas-
ing to the burst pressure rating of the rupture disk in the event of op-
eration of the containment block valves.

 

Temperature Alarm Switches. The temperature alarm switch modifica-
tions discussed in the last report12 were completed. It is not known
at this time whether these modifications have effectively eliminated the
drift problem, because many of the modules were reset prior to and during
critical and low-power operations and because module settings could not
be checked during reactor operations. Observation of the performance of
these devices is continuing.

 

Drain-Tank Level Probe. Modification and repair of the fuel flush
tank level probe excitation cable, which failed in service dve to oxi-
dation and embrittlement,13 was completed during the precritical shut-
down in March. Performance of the probe during subsequent operations
was satisfactory. ©Some difficulties were experienced in the fabrication
and installation of the replacement probe assembly. These difficulties
are discussed in more detail in the development section of this report.

 

Fuel-Processing System. ©OSeveral modifications and revisions were
made in the fuel-processing system instrumentation and controls to im-
prove performance and safety of operation. An interlock was added to
prevent the discharge of hydrogen into the containment air exhaust duct
unless there was flow in the duct. A glass rotameter and a manually op-
erated control valve were replaced with an automatic flow control system
consisting of a weld-sealed orifice and differential-pressure assembly,
a pneumatic controller, and a pneumatically operated bellows-sealed valve.
A motor-operated damper in the containment air system was removed after
1t was determined that the valve was not suitable and was no longer re-
quired.

 

Fuel-Pump-Bowl Cooling Air Flowmeter. Instrumentation, consisting
of an orifice-type flowmeter, a differential-pressure gage, and an indi-
cator, was provided to measure the flow of cooling air to the fuel-pump-
bowl shroud. Because of the configuration of cooling air piping, the

 
46

flow element was installed inside the reactor secondary containment ves-
sel. The transmitter was located outside the containment vessel and
bilological shielding. The transmitted signal is indicated on a gage in
the transmitter room. The design of the system will permit future con-
version to automatic control, if required.

Blectronic Transmission System Signal Modifiers. During checkout
of the computer—data-logger system, it was discovered that most of the
input channels which received their signal from 10- to 50-ma (Foxboro ECI)
instrument loops were inoperative due to amplifier overload. Further in-
vestigation revealed that the amplifiers could not tolerate more than
+10 v common mode voltage and that common mode voltage as high as 59 v
was present at the amplifier input terminals. The cause of the high
common mode voltages was determined to be the electrical location of the
resistance-type (current-to-voltage) modifiers in the 10- to 50-ma trans-
mission loops. This problem was eliminated by revising the 10- to 50-ma
signal interconnection wiring to place the electrical position of all
the resistors at or near a common ground point.

 

Documentation

 

Drawings, tabulations, and specifications were revised to incorporate
additions, revisions, and as-bulilt corrections. Work on the instrumenta-
tion section of the design regort has begun and will continue. A process
instrument switch tabulation®? was completed and issued.

References

 

1. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 32.
<. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 8.
3. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 7.

4. B. E. Prince, MSR Program Semiann. Progr. Rept. Jan. 31, 1964, ORNL-
3626, pp. 53-5%.

5. P. N. Haubenreich et al., MSRE Design and Operations Report. Part
ITI. DNuclear Analysis, ORNL TM-730 (Feb. 3, 1964).

 

6. P. N. Haubenreich, Prediction of Effective Yields of Delayed Neu-
trons in the MSRE, ORNL-TM-380 (Oct. 13, 1962).

 

7. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 17-23.

8. Tbid., pp. 84—86.
10.
11.
12.

13.

14.

4'7

Tpbid., p. 10.

 

Thid., p. 28.

 

Thid., p. 49.

 

Thid., pp. 4647.

 

Thid., pp. 41—42.

Tnternal memorandum (June 1965).
48

2. COMPONENT DEVELOPMENT

Most of the development effort was spent in checking out components,
assisting at the reactor in the operation of various components, and
making changes to the components which will improve the reliability of
the MSRE during the upcoming period of operation at power. We have also
continued selected life tests on components to provide advance informa-
tion of performance. The following is a description of this work.

TLife Tests

 

Freeze Valve

 

A thermal cycling test was conducted on the prototype of FV1i03. The
test consisted in raising and lowering the temperature at the center of
the valve through the range of 600 to 1200°F. The test was conducted
with no salt in the system through 633 cycles, after which the system was
shut down, visually examined, and pressure checked for evidence of crack-
ing or other changes. No changes were observed. The test was terminated
after 1167 additional thermal cycles without evidence of change. The
valve has not been examined metallographically.

Pipe Heater

 

The prototype of the removable heater for 5-in. pipe manufactured
by Mirror Insulation Company has operated for over 13,000 hr with an in-
ternal temperature of 1200 to 1400°F without difficulty. It has operated
continuously since July 28, 1964.

Drain-Tank Heater

 

The prototype drain-tank heater has operated continuously for 12,000
hr at 1200°F without difficulty.

Checkout and Startup of Components

 

Control Rods

 

The modified control rodsl were installed in the reactor prior tO
the criticality tests and operated without incident. The rod lengths
remained within 0.10 in. of the installed lengths. The cooling air flow
To the rods was increased from 2.5 to 4.0 scfm by decreasing the resist-
ance in the supply circuit. The average times required for the rods to
fall through 51 in. of travel were measured before and after the zero-
power test; the results are given below.
49

 

Rod Fall Time (sec)

 

 

No. 1 No. 2 No. 3
Initial 0.828 0.834 0.884
After zero-power tests 0.792 0.775 0.792

 

Connection of the control rods to the drive unit by remote means was
accomplished with minor difficulties which were overcome by an improved
handling technique. The flange attachment at the top of the rod tended
to misalign with the rod drive flange due to the manner of pickup of the
rod when it was resting on the thimble flange.

Control Rod Drive Units

 

The switch actuators which operate at the limits of the control rod
stroke were modified to increase the hardness and area of the bearing
surfaces between the switch retainer and the sliding actuator. The change
was indicated when galling occurred in this area. The bearing surface
hardness was increased from R. 28 to Re 58 on those surfaces in contact
with the actuator. A prototype of the modified switch was scrammed 960
times and operated through 14,760 switch operations without difficulty.
All the drives are now equipped with this type of actuator assembly.

The initial buffer stroke of the shock absorber was set at 3.0 in.
At the end of the criticality test the buffer strokes were: No. 1, 2.4
in.; No. 2, 4.75 in.; and No. 3, 3.1 in. The stroke lengths have been
again adjusted to ~3.0 in. prior to reinstallation.

The APL grease used in the gear head and bearings was in good condi-
tion; however, the reactor cell temperature during the test was less than
the normal 150°F. The temperature measured at the lower end of the con-
trol rod drive can remain at 104°F with ~2.0 scfm of cooling air flowing
down through the rod casings. Temperature alarm switches set at 200°F
are being installed at the lower end of each rod drive. These elements
will serve to indicate cooling air flow difficulties which could cause
excessive temperature increases in the gear head and possible damage to
the gear lubrication.

The wire cables for the electrical disconnects are being lengthened
to facilitate the remote installation and removal operation. One cable
will have sufficient length to permit adjustment of the position indi-
cators on the drive units with the units outside the cell.

Figure 2.1 shows an operational rod drive mechanism removed from its
case.
50

PHOTO 68258

    
      

. GEAR
BOX

     

LIMIT
SWITCHES

 

SHOCK
ABSORBER -

AlR
TUBE

Fig. 2.1. MSRE Control Rod Drive Unit.
51

The tool steel worm and worm gear pair in each drive were out of
tolerance as to surface finish when received but were installed until re-
placements could be obtained. Examination of these pieces at the end of
the criticality test shows evidence of minor galling on the wear surfaces
of the worm and of metal particles distributed into the grease. In gen-
eral, the gears are in fair condition except for the surface roughness.
Acceptable gears of 440-C fully hardened material will be installed be-
fore the drives are returned to the reactor. Prototype gears of this type
have been tested through >60,000 cycles and remain in good condition.

Freeze Valves

 

oceveral problems were found in the operation of the freeze valves
during the zero-power experiments.2 Changes were made in the electrical
and alr supply to the valves to provide more margin for operation during
abnormal conditions. The results of the changes are described below.

Cooling air flow rates to the freeze valves have been increased by
reducing the line pressure drop. Maximum air flows (scfm) to the valves
with an 8-psi air supply are now: FV204, 22.5; FV206, 25.0; FV104, 26.6;
Fv105, 27.7; and FV106, 24.8.

Freeze Valves 204 and 206

 

The coolant-salt freeze valves 204 and 206 have been modified by re-
moving the center heater and enclosing the 2-in. valve plate in a metal
shroud to contain the cooling air. These valves are now similar to all
the other valves in the system with the exception of FV103. The heat ca-
pacity in the furnace area surrounding the valves was increased by instal-
ling a ceramic liner around the valve heaters. Freeze valve 204 now melts
in 13 min on power failure and 206 melts in 12 min. The control of the
temperature distribution along these valves will be improved by installa-
tion of separate control circuits to the valve shoulder heaters. These
valves formerly used a single Variac to control the heat on both sides
of the frozen wvalve. Independent control of the shoulder heaters should
alleviate the temperature gradient shown in Fig. 2.2.

A differential flow controller is being added to the cooling air
supply system to each of these valves and will be operated from a single
selected thermocouple. The function of the controller is to maintain the
valve temperatures within the module control temperature boundaries by
automatically varying the cooling ailr flow to maintain a constant temper-
ature at the valve. The purpose of this change is to reduce the amount
of temperature cycling, the frequency of the alarms that occur when the
module set points are exceeded, and the attention required by the reactor
operators.

Freeze Valve 103

 

The operation of FV103 is complicated by a dual set of operating
conditions: (1) After the reactor system is filled with salt, FV103 is
52

ORNL—-DWG 65—11795

58

e C Y
T v X T e N\

A4 B4 B4 A4

 

 

B

 

 

b

 

 

 

 

1200 l

 

 

 

 

 

 

TEMPERATURE (°F)

 

 

 

 

 

 

 

 

 

 

 

 

X\ / X X% /X
800 < \
X
X/
13-min POWER t2-min POWER
FAILURE MELT FAILURE MELT
400 \
X V/
| 1
FV-204 FV-204
: 1 ||
A4 fiB 2B 3B B4 5B B4 38 2B 1B A4
TE NUMBER

Fig. 2.2. Freeze Valves 204 and 206 Temperature Distribution —
Hold Freeze Condition.

frozen, and salt is retained in the 103 line for some period. At this
time there is salt on both sides of the valve. (2) After determining
that the salt level in the pump bowl is correct, the 103 line is vented
via line 519. Freeze valve 103 then has hot salt on the reactor side of
the valve and an empty pipe on the drain-tank side. During condition 1,
the temperatures are symmetrical across the frozen valve. During condi-
tion 2, a temperature gradient of ~400°F exists between the reactor side
and the drain-tank side of the wvalwve, probably due to the difference in
the heat conductance of the full and the empty line. The module controls
were set to control over this wide range.

The freeze time for FV103 is 15 to 30 min if the salt is steady in
The system. Thaw time average has been 11 min with the reactor tempera-
ture at 1150 to 1200°F.

A differential controller was installed on FV103 prior to the criti-
cality tests and operated smoothly throughout the run. The controller
53

ORNL-DWG 65-11796

 

 

 

 

 

 

 

 

3A_ 38, 2A 28 1A 1B
N\ T/ |
( . . Nul g/ 0C - ~ REACTOR
U ] A\ A\ e )0
\ \

B A1B

FV-103

1400

1000 o”/////j\\ 4””’t:::::::::::;

 

 

<

 

 

—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

%’L ‘/)]( \ X/
& 800 X \\\~
g \\ /47 X CONTROLLER ON FV103-{B-—960°F SET
o O CONTROLLER ON FV103-3B-540°F SET _|
600
S Nx_ /
L}E I\X
400 i
\[ VALVE IN HOLD FREEZE CONDITION
200 |
0
B AlB 38 2B 1B R27-30 AVG
TE NUMBER

Fig. 2.3. Temperature Distribution on FV103 with Controller.

was operated initially from the drain-tank-side thermocouple (FVlOB-BB),
and 1t was found that this couple was relatively insensitive to changes
in reactor temperature. The control was then shifted to the thermocouple
on the reactor side of the wvalve (FVlOB-lB) and remained there for the
balance of the run. Over the range of 1150 to 1200°F no changes in con-
troller set point were necessary. The reactor temperature was lowered to
1050 to 1100°F at one stage of the operation, and it was necessary to
lower the control point by 50°F. Operation of the controller from the
valve center temperature (FV103-2B) will be tried next to see if an op-
timum mode of operation can be found. Figure 2.3 shows how the tempera-
ture distribution across FV103 changes with the location of the control
thermocouple when the valve is in the hold-freeze condition.

Freeze Valves 104, 105, and 106

 

Operation of these wvalves has been difficult for a number of reasons:
(l) pressure differences in the drain system which affect the salt dis-
tribution in the valve tree (FV105 and 106), (2) poor heat control and
resultant temperature distribution across the valves, (3) inadequate heat
at specific locations, (4) loss of heater elements for mechanical reasons,
and (5) difficulties with the temperature control modules.
54

Methods are being‘developed with operational experience which ensure
adequate Tilling of the valves with salt.

Separate shoulder heater controls were installed and tested on freewze

valves 105 and 106 which relieve the temperature distribution difficul-
ties.

Power to the shoulder heaters was increased from 15 w/in.2 to 30
w/in.®. The effect of these changes is shown in Fig. 2.4.

The loss of some of the heaters from mechanical difficulties was due
to improper installation; warping of the heater boxes to which these heat-
ing elements are attached, with resultant breaking of the ceramic elements;
and abuse of the heater element lead wires during installation. New heat-
ing elements were installed in all the freeze-valve heating units. These

ORNL—-DWG 65-11797

LINE 103
FUEL DRAIN TANK _— FUEL DRAIN TANK

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7 J/KM1Q13
1200
I ]] ]
¢
1000 5 3
/
| 7
{/ )
_ /
< 800 3
w C | X
o
D
— |
<
Q: b
a |1 A
= 600 |i
- {l O BEFORE MODIFICATION |
‘| X AFTER MODIFICATION |l
400 ,
200 l !
1 2 3?5 6 7 8 910, 12 13 14
¢

— -

Fig. 2.4. Effect of Freeze-Valve Heater and Control Modifications
on Temperature Distributions for FV105 and FV106.
55

elements were installed to fit loosely to permit some freedom of movement.
The lead wire size was reduced from No. 12 to No. 1l4. The method of at-
taching the lead wire to the element wire was modified to prevent twisting
and breaking the leads at the welded joint. This was done by adding a
short length of wire at 90° to and extending outward from the welded joint
and embedding the cross and joint in the ceramic.

1t was noted that in a number of units with the straight connection,
the weld had broken; but the lead and resistance wires maintained conti-
nulty due to the twisting together of these wires prior to welding. How-
ever, breakage at this weld point can create an area of high resistance
and possible burnout; so the change in attachment method was made.

The module difficulties are being corrected by the instrument group.
These were multiple alarm points for a single set point and drifting off

set points in some instances.

Differential cooling air controllers are being installed on freegze
valves 105 and 106.

Noble-Gas Dynamics

 

Analysis of the experiment® on the noble-gas dynamics was continued.
The experiment was designed to evaluate the constants needed to describe
the xenon migration in the MSRE and consisted in adding 85Ky to the gas
space of the fuel pump bowl for an interval of time and then observing the
rate of removal during an interval of stripping. Run 3, which consisted
of an 11-1/2—day addition phase and a 6-day stripping phase, yielded five
exponentials. The half-lives associated with these exponentials and their
physical interpretation are given in Table 2.1. Exponentials 3 and 4
are still subJject to interpretation. Although listed as two graphite

Table 2.1. Results of ®9Kr Experiment, Run 3

 

 

Fxponential Half-Life Process Rate Constant Rate Constant
*P (hr) Involved Value
1 0.119 Purging pump bowl Purging 74%
efficiency
2 1.04 Stripping salt Stripping 129
efficiency
3 4e52 Probably stripping Mass transfer 2.0 ft/hr
graphite coefficient
4 15.5 Probably stripping Mass transfer 0.59 ft/hr
graphite coefficient
5 198 Stripping bulk Mass transfer 0.046 ft/nr

graphite coefficient

 
56

regions, they may actually be one graphite region in which krypton diffu-
sion in the graphite must be considered. The estimated stripping effi-
ciency, from work done at the University of Tennessee with a CO,-water
system, is about 17%. The estimated mass transfer coefficient for the
bulk graphite, assuming turbulent flow, is about 0.1 ft/hr. It is rea-
sonable that the measured value is lower than this because the flow is
not turbulent through the entire channel, as was assumed in the estimate,
but only through a part of it.

scampler-Enricher

 

Installation of the sampler-enricher system for the fuel pump was
completed just before the start of the precritical run (PC-2). Therefore,
shakedown testing and training of operators were done concurrently with
sampling and enriching of the fuel loop. The system was used to isolate
54 samples and add 87 capsules of enriching salt during PC-2 and the zero-
power run (run 3). Twenty operators were trained in the use of the sampler
during this period. Little or no contamination of the work area occurred
during operations or maintenance. A maximum radiation field of 5 mr/hr
was measured at the outside of the unshielded transport container which
contained samples isolated either during an extended 10-w operation or
immediately after a high power peak.

During this initial operating period some maintenance was required,
and some changes were indicated which were delayed until the completion
of run 3. BSeveral components were removed from the system for maintenance
and were decontaminated easily by washing.

The mechanical troubles encountered during this initial period are

discussed below. After the indicated changes are completed and after the
Operators gain additional experience, less maintenance should be needed.

Removal Valve and Seal

 

The transport container and removal tool assembly did not slide
through the removal valve and seal unit freely. Examination revealed
that the valve and seal were not aligned properly with area 3A (see Fig.
2.5). After clearances were increased, no further binding occurred.

During the entire test the ball of the removal valve failed to seal
properly even though both the ball and the seals were replaced. There-
fore, the entire valve assembly will be replaced with a modified design
which should improve the sealing characteristics and access for future
maintenance work. During the run, one of the three-way solenoid valveg
which controls the air supply to the removal valve operator failed and
was replaced.

Manipulator

 

The boots used to seal the manipulator arm to area 3A were replaced
three times. On one occasion area 3A was inadvertently evacuated without
57

ORNL-DWG 63-5848R

 

REMOVAL VALVE AND
SHAFT SEAL

PERISCOPE
LIGHT

   
  
 
 
  
 
  

 

 

N
N

/]
L

xrr?

 

CASTLE JOINT (SHIELDED
WITH DEPLETED URANIUM)

 

CAPSULE DRIVE UNIT— e
LATCH\W F

ACCESS PORT—

&
N

| i B
! N
| ] MANIPULATOR

%\AREA 3A (SECONDARY

CONTAINMENT )

SAMPLE TRANSPORT
CONTAINER

| EAD SHIELDING

"D

— ——. .7 o 2 2 )

AREA 1C
(PRIMARY CONTAINMENT) —

 

 

 

SAMPLE CAPSULE ]

 

 

 

   
 
 
  
 
   
  
  
  

 

 

 

 

 

OPERATIONAL AND
MAINTENANCE VALVES

SPRING CLAMP \
DISCONNECT

| ——=AREA 2B (SECONDARY
CONTAINMENT)

 

 

 

 

 

TRANSFER TUBE
(PRIMARY CONTAINMENT )~/

LATCH STOP —+
CRITICAL CLOSURES

REQUIRING A BUFFERED
SEAL

 

 

 

MIST SHIELD ’ O ! 2
CAPSULE GLIDE FEET

Fig. 2.5. ©Schematic Representation of Sampler-Enricher Dry Box.

the manipulator cover, and the resultant pressure gradient across the boot
ruptured it. Modifications to the interlock system are planned to prevent
recurrence of this type of accident. On another occasion, while using

the manipulator to release a sticking access port operator, the boot was
snagged on the bottom piece of the transport container. The height of
this bottom pilece was reduced, and the access port operators were read-
Justed. The third failure resulted from pinching the boot between the
manipulator arm and the housing.

Access Port

 

On several occasions one or two of the six access port operators
failed to open when they should have. The manipulator was required to
release the sticking operator. The tension on the operators was read-
Justed, and the gas-supply tubing to the pistons was realigned to relieve
& restraining stress. To increase the ease of emergency operation, a
knob was added tc the pin which must be moved with the manipulator to
manually release the operator.

Operational Valve

 

The leak rate of buffer gas through the seals on both the operational
and maintenance valves exceeded 5 cm® of helium per minute. In both cases
58

the upper gate seal had the greater leak rate. The operational valve was
removed and examined. A thin black ring, which was easily removed, had
formed. at the upper sealing surface of the valve gate. When the valve
stem and gate were lubricated, the valve sealed almost completely. When
the lubrication was then removed from the gate, the leak rate of buffer
gas (helium) increased to about 2 cm®/min through the upper gate seal and
remained at zero through the lower seal. The maintenance valve will be
removed and cleaned, and the stem will be lubricated.

A small quantity of salt spheres (estimated at <1 g) had collected
between the seats of the gate valve. The analysis of the uranium in the
spheres showed the concentration to be less than enriching salt and more
than fuel salt. Therefore, the spheres must have come from the used en-
riching capsules as they were retrieved from the pump bowl. A possible
explanation is that the bottom hole in the enriching capsule was not
drilled completely through the nickel wall; a small droplet of salt could
bridge the hole and then could be dislodged from the capsule during han-
dling with the manipulator while removing it from area 1C.

campling Capsule

 

On one occasion a sampling capsule was knocked into area 1C before
the latch key was completely engaged in the latch, and the capsule as-
sembly dropped down to the operational valve. The capsule was retrieved
by removing the manipulator assembly from area 3A, opening the access
port, and retrieving with a hook on the end of a stiff wire.

Fuel-Processing Sampler

 

Design of the sampling equipment and instrumentation for the fuel-
processing system has been started. The sampler-enricher mockup will be
modified for this use.,

Off-Gas System

 

Charcoal Beds. A gas-retention test was made to check the perform-
ance of the MSRE charcoal beds. With the bed temperature at 85°F and
helium flowing at a constant rate, a pulse of 85K was injected at the
bed inlet. The time required for passage of the krypton through the bed
was determined by monitoring the effluent gas with a G-M tube.

 

The holdup time for krypton at the design flow rate of 4.2 liters of
helium per minute was 5~l/2 days. Adsorption data published by Ackley
and. Browning4 indicate the equivalent xenon holdup time to be a minimum
of 88 days. By adjusting the holdup times downward to allow for tempera-
ture effects, an estimate was made of the atmospheric concentrations of
krypton and xenon which will be produced by the charcoal bed effluent
during 10-Mw operation. The estimate was based on the following assump-
tions:
59

Minimum holdup time (t) — 4-3/4 days krypton and 75 days xenon
Minimum stack flow — 7 X 10° cm?/sec (15,000 cfm)
Atmospheric dilution factor — 1500 (ref. 5)

Maximum atmospheric concentration will be:

(Ri) (g ) (e™MY) (108)
C; = = 2.6 X 10712 Rjhie
("7 x 10°) (1500) (3.7 x 1010)

=N\t

J

 

where

I

Ry flow rate at pump bowl outlet, atoms/sec,

Ci

atmospheric concentration, pc/cm?,

For the indicated holdup times, 8°Kr, 121MXe, and 133Xe are the only iso-
topes which yileld concentrations of significance. The concentrations for

These are given in Table 2.2 together with the maximum permissible con-
centration (MPC).

Table 2.2. Atmospheric Concentrations of Radiocactive Gases
from the MSRE Off-Gas System

 

MPC (uc/cmB)

 

-7\1't 3
* Ry Ay © Cy (he/en?) 168-hr Week®
85Ky 9.4 %X 1014 2,14 x 10°° 1 5.2 x 1072 3 x 1076
13y, 9,2 x 1012 6.7 x 1007 1.3 x 1072 1.4 x 107° 4 X 1070
133x%e 2.0 x 1016 1.5 x107%® 6.0 x 1075 4.7 x 107° 3 x 10°6

 

SNBS Handbook 69.

 

Solids Entrainment. Difficulties were encountered with operation
of the back-pressure control valves (PCV 522 and 528) in the fuel and
coolant off-gas systems during the precritical and critical test periods.
The trouble was due, at least in part, to an accumulation of solids at
the valve seats. Visual examination indicated that the solids were a
mixture of glassy spheroids, about 1 p in diameter, and carbon or a car-
bonaceous material. Optical measurements indicated that the chemical
composition of the spheroids was the same as that of the circulating
salts, and it is assumed that they are due to carryover of the salt mist
in the pump-bowl gas spaces. The carbon is assumed to be a residual con-
taminant from the manufacturing process. The problem is being approached
in two ways: (1) a study is being made to determine whether the solids

 
60

entrainment can be reduced or eliminated, and (2) filter elements immedi-
ately upstream of the wvalves are being changed from a 25-u pore size to
l-u size.

cample Unit. Design work was started on a sample unit for the re-
actor off-gas stream. The unit will include:

 

1. @a thermal conductivity cell for on-line measurement of total contami-
nants,

& gas chromatograph unit for measurement of gaseous constituents,

a refrigerated molecular sieve trap for batchwise collection of gas-
eous constituents; the trap contents will be transferred to a shielded
sample bottle and then to a hot cell for analysis.

Supply to the sample unit will be a small side stream (<100 cmB/min) taken
from the off-gas line immediately downstream of PCV 522. Initial design
effort is being concentrated on areas which must be completed prior to
operation of the reactor at power (e.g., piping connections to existing
equipment and shielding revisions).

Remote Maintenance

 

The previously established programs for ensuring the maintainability
of the MSRE were continued. Broadly stated, these programs consist of
monitoring new designs and changes, inspecting the installation of the
reactor equipment, and trying out the maintenance procedures where the
possibility of future trouble could be determined.

One control rod and its drive were removed using the work shield.
This Jjob required handling five 3/8 -in. socket head bolts with long-han-
dled tools. Vision and maneuverability were hampered because of close
clearances and because the bolts are not quite accessible from directly
above. Previously developed lights and viewing techniques were not effec-
tive for all the bolts. However, a viewer with a collimated light source
mounted above the work shield was used with success. Because it takes
longer to set up this type of light source, it will be used only on those
bolts where the internal lights are not good enough. In addition, align-
ment guides were added to both the adapter flange and the housing to help
in the installation and removal of the drive housing.

All three space coolers and many of the freeze-valve tree replaceable
heaters were removed from their installed positions for alterations. This
provided the opportunity to check out several remote-maintenance require-
ments. The space coolers were balanced for handling, flange guides were
added, and the effectiveness of blowing out the lines was observed. All
the space coolers were observed to lose water when the flanges were broken,
indicating the need for improving the procedures for blowing out the lines.
The replaceability of some of the drain cell heaters has been adversely
affected by thermal distortions. This has the effect of producing a tight
fit between adjacent heaters. The proposed solution is to alter the heater
units if they have to be removed for maintenance.
61

All the maintenance procedures are affected by revisions, new in-
stallations, and unanticipated requirements; so a continuing effort is
being made to keep track of changes. Toward this end, a program of pho-
tographing installed equipment and marking identifying numbers on thermo-
couple boxes, heaters, and electrical leads is under way.

Revised freeze-~flange clamp operators were assembled and tried for
fit. Long-handled tools were designed and built for the installation of
the revised graphite samples. Preparations are being made for handling
the pump rotary element and the graphite samples, and for inspecting the
core through the 2-1/2—in. access flange. Procedures were worked out for
handling the newly designed control rod shielding. Preliminary procedures
were prepared for maintaining some of the elements of the sampler-enricher.

Pump Development

 

MSRE Pumps

 

Molten-Salt Pump Operation in the Prototype Pump Test Facility. The
spare rotary assembly for the fuel pump® circulated the salt LiF-BeF,-
ZrF,-UF,; (66.4-27.3=4.7-0.9-0.7 mole %) for 2644 hr at 1200 to 1400°F.
During this time, measurements were made to determine the concentration
of undissolved helium in the circulating salt and the effectiveness of
the down-the-shaft helium purge against the upward diffusion of 85K+ to
the catch basin for the lower shaft seal.

Operation of the pump was terminated when strident noises and an un-
wanted increase in power to the pump drive motor were noticed. Inspection
revealed that three vanes, each approximately 2 X 24 X 1/8 in., had be-
come detached from the flow-straightener section in the salt loop. One
of the vanes was carried by the circulating salt to the impeller inlet,
where 1t became lodged and rubbed against the impeller. The only apparent
damage to the pump was the removal of metal from the inlet portion of the
impeller vanes. The failure of the vanes apparently resulted from metal
fatigue caused by vibration induced by the flowing salt. This unit will
be reassembled with a new impeller, shaft bearings, and seals and given
a cold shakedown test to prepare it for future reactor service.

The spare rotary assembly for the coolant pump6 was prepared for re-
actor service and will be held in standby.

The prototype pump test loop is being modified. The Venturi flow-
meter is being relocated upstream of the orifice flow restricter, and a
new flow-straightener section is being installed upstream of the Venturi.

Lubrication Systems. Tests were conducted in the transparent mockup
of the lubrication reservoir to determine the source of gas entrainment
in the circulating lubricant that was noted during shakedown tests of the
MSRE lubrication systems.7 The entraimment made it difficult to prime a
standby pump during emergency operation. Sources of entrainment were

 
62

found to be the action of the jet pump used to scavenge oil from the bear-
ing housing and the agitation of the oil caused by the rotation of the
shaft and bearings. The scavenging Jjet pump was modified to reduce the
entrainment of gas, and the standby pump primed satisfactorily. A modi-
fied jet pump will be installed in the lubrication systems for the fuel
and coolant pumps at the MSRE and tested prior to power operation of the
reactor.

The lubrication pump, which is circulating oil in an endurance® test,
has operated without incident for 18,000 hr. The thermal cycling tests
(1000 cycles, 1 hr running and 1 hr off) were completed. The endurance
test is continuing at 160°F oil temperature and 70 gpm.

Back Diffusion of Fission Gases in the Pump Shaft Annulus. Addi-
tional tests were made to investigate the diffusion of $2Kr in the shaft
annulus against a flow of helium purge gas. In these tests the radial
clearance in the shaft annulus was 0.005 in. compared. to 0.0025 in. used
in the previous tests.’ The following data were obtained with the 0.005-
in. annulus:

 

85Kr Concentration

 

 

Shaft Purge Cat;firBisln (curies/cm?) Dilution

(1iters/day) (literi/da ) Factor
Y/ In Pump Tank In Catch Basin

212 173 2.66 X 107¢ <1.5 x 10710  >317 700

232 164 5.50 X 1007¢  <1.5 x 10719  >36,700

72 105 be2e X 1076 8.17 x 10~? 520

 

It was difficult to measure the dilution factor (ratio of concentration
of fission gas in the pump tank to that in the catch basin) except for
very low purge flow rates of little interest to the project. The diffi-
culty stems from our inability to measure the concentration of fission
gas below 10710 curie/cm®. An additional factor was the relatively low
concentration of fission gas permitted in the pump tank because of hazard
to personnel, the quantity of the gas available, and the problems of han-
dling it. However, the data do indicate that purge flow rates in excess
of 1000 liters/day should protect leakage 01l in the catch basin from
polymerization by fission gas.

Measurement of the Concentration of Undissolved Helium in Circulating
Molten Salt

 

Circulating helium in the MSRE fuel will reduce the fuel density
and, therefore, the reactivity of the reactor. Measurements of both the
void fraction and the void-fraction reactivity coefficient were required.
63

A radiation densitometerl® (ORNL Dwg. 433-7.0 R 41482) was used to
obtain a measure of the helium void fraction. The void-fraction coeffi-
cient was determined from transient measurements with the densitometer
in conjunction with other reactor parameters. The detector signal current
was fed to a suppression circult that indicated only variations in de-
tector output. The circuilt output was recorded on Visicorder tape. The
densitometer sensed only mass variations; these variations were inter-
preted in terms of void fractions and temperature effects in the following
presentation.

Investigations were made on two different experimental installations.
The first was the prototype pump test loop. Here, the densitometer was
evaluated, and a standardization technique was developed. The second in-
stallation was on the fuel inlet line to the MSRE reactor vessel.

Prototype Pump Test Loop. Helium concentration measurements were
made with a radiation densitometer at two pump bowl levels, upper and
normal. The salt was circulated through a 6-in.-diam system at 1615 gpm;
an additional 85 gpm flowed through the pump bowl spray ring.

 

The densitometer axis was horizontally positioned normal to and in-
tersecting the pipe axis. The scintillation detector was maintained at
70°F by water cooling. The detector voltage was maintained at 1700 v,
and the voltage divider-network drain was 9 ma.

The measurement technique involved the following. First, after a
steady-state salt flow and temperature condition was established, refer-
ence measurements were obtained. A Tused-silica beam filter was used as
a mass sensitivity standard. The mass sensitivity was checked frequently
throughout all the measurements. ©Second, measurements of void behavior
were obtained during salt-flow stoppage and subsequent reestablishment.
The third technique involved varying the salt temperature. Flow and no-
flow measurements were obtained at several temperatures.

The two Visicorder traces shown in Fig. 2.6 give a measure of helium
concentration in the loop at two pump bowl levels. The traces are plotted
as 1T normalized to the flow density in order to illustrate the transient
behavior of both the helium purge and buildup when the pump is first
turned off and then turned on. Trace I relates to the normal pump bowl
fuel level, which is 3-3/8 in. above the center line of the volute (ORNL
Dwg. F-RD-9830). The circulating void fraction was approximately 4.6
vol %. Trace II was obtained when the fuel level was 4-7/16 in. above
the center line of the volute; the fuel covered the spray ring discharge
ports. The void fraction was approximately 1.7 vol %. The volume sensi-
tivity of the densitometer was 0.64 vol % per division in each case.

The volume sensitivity and the validity of using the fused-silica
beam-filter standard are illustrated in Fig. 2.7. The densitometer volume
sensitivity is calculated from the following:

100 M
S

>S5
s C
64

where
Mg = mass of standard,
DS = recorder divisions deflection due to standard,
ME = salt mass subtended by the collimator.

The numerical value of S is controlled by adjustment of the electronic
system gain. The density at temperature T is obtained from:

Mr S S
pT=V;-+Pr1—O—O-(Dt-Dr)=Pr[l+-l—o—o-(Dt—-Dr)} ,
where

Mr = salt mass subtended by the collimater at the reference tempera-
ture (in this case, 1200°F),

Pr = salt density at reference temperature,

Dr = deflection at reference temperature,

Dt = deflection at temperature in question,

Vé = volume subtended by the collimator.

The measured densities are in very close agreement with those obtained by
independent measurements.tt

ORNL-DWG 65-11798

SILICA
BEAM PUMP ON

FILTER

I 4.6%

DENSITY INCREASING

I NORMAL PUMP
BOWL LEVEL

L
L
O

II UPPER PUMP
BOWL LEVEL

II .7%

DEFLECTION (div)

SALT TEMPERATURE : 1185°F
VOLUME SENSITIVITY: 0.64% /div

 

O 20 40 o0 80 100 120 140 160 180 200
TIME (sec)

Pig. 2.6. Transient Characteristics of the Helium Concentration
in the Prototype Pump Test Loop.
65

ORNL-DWG 65—-11799

 

e g N

— 3
C—2.286 g/cm

P
16 o =2.29 g/cm3 SENSITIVITY = 0.275 VOL.%/div
M p. = CALCULATED DENSITY

14 ~\\\\\ FM==MEASURED DENSITY

N
2 <

 

 

 

 

 

 

 

\; 259 g/cm?

/%=2
\
N

 

DEFLECTION (div)
®

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\ p.=2.203 g/cm3 _|
4 ¢ _ 3
\%2.20 g/cmx
2 \\
0 \\\j
1100 1150 1200 1250 1300 1350 1400

TEMPERATURE (°F)

Fig. 2.7. Temperature Characteristics of Loop Salt — No Flow.

Molten-5alt Reactor Fuel Circuit. The densitometer was horizontally
mounted on the reactor input fuel line; it was operated in the same fashion
as on the prototype pump test loop. The fuel line was a 5-in.-diam pipe
instead of a 6-in. one. The system flow was 1150 gpm, with a 50 gpm spray
ring flow,

The first measurements were obtained from carrier salt only. The
measurement sensitivity was 0.076 vol %. The data indicated that there
was no measurable void fraction at 1190°F. The flow was stopped and then
started again numerous times, and the quiescent period was varied from 5
to 10 min. The only density changes observed were those caused by temper-
ature changes (approximately 0.0095 mass % per °F).

The second set of measurements were obtained after fuel was added to
the carrier salt. At 1200°F there was no measurable void fraction, either
with pump excursions or with 15 1b overpressure. Overpressure measure-
ments made at 1050°F indicated a void structure. The low-temperature
densitometer data, in conjunction with other measurements, were used in
determining some of the nuclear parameters discussed previously in the
section Analysis of Operation.

Other Molten-Salt Pumps

 

PKP Fuel Pump High-Temperature Endurance Test., This pump12 was oper-
ated for 384 hr circulating the salt LiF-BeF,-ThF,-UF, (65-30-4-1 mole %)
66

at 1200°F and 1550 rpm. The test was halted temporarily due to failure
of the lower bearing in the drive motor, apparently due to insufficient
lubrication. The bearing was replaced, and the pump has since been oper-
ated at 1200°F, 1650 rpm, and 800 rpm for 800 hr.

MK-2 Fuel Pump. Water testing 1%°13 with the pump tank models (full
scale and 4-in. section) was concluded. There was no noticeable reduction
of gas content attalned with the various baffle arrangements tested in
these models. The salt pump tank design12 is near completion, and the
internal baffling and flow passages will duplicate that used on the full-
scale pump tank model.

 

Instrument Development

 

Ultrasonic, Single-Point Molten-Salt Level Probe

 

Fabrication and installation of an ultrasonic level probe system for
the MSRE fuel storage tank (developed by Aeroprojects, Inc., under con-
tract to the AEC with ORNL assistance) were completed prior to the start
of MSRE chemical-processing system operations in May. Figure 2.8 shows
the probe assembly before installation in the fuel storage tank. Figure
2.9 shows the force-insensitive mount assembly that permits ultrasonic
energy 1o be transmitted through the tank wall. Note the rugged all-
welded construction of this device. A final seal weld was made at the
periphery of the flanged section when the probe was installed in the fuel
storage tank. A Tunctional diagram of the system is shown in Fig. 2.10.
Except for the frequency of operation, the principle of operation of the
MSRE probe system is the same as that of the prototype probe system pre-
viously described.t® A lower (25-kc) frequency was required on the MSRE
probe because of the heavier construction required to provide adequate
allowance for the high corrosion rates expected during operation of the
chemical-processing system.

Performance of the probe was satisfactory during initial operation
of the chemical-processing system; however, some difficulties were experi-
enced during subsequent operations. The probe operated very well when
the tank was filled but did not operate when the tank was drained. A
check of the instrument made at this time revealed that the oscillator
frequency had drifted 40 cps from the original setting. Correction of

PHOTO 72098

 

Fig. 2.8. MSRE Fuel Storage Tank Ultrasonic Level Probe Assembly.
o7

  

PHOTO 72099

 

Fig. 2.9. MSRE Fuel Storage Tank Ultrasonic Level Probe, Force-
Insensitive Mount, and Excitation Rod.
68

ORNL-DWG 65-11800

2] e FUEL STORAGE CELL SHIELDING
= CONCRETE WAL L

iaf"flméfi PIEZO ELECTRIC
: L CRYSTALS
LEVEL LIGHTS

    
  

 

 

    

MAGNETO-
STRICTIVE
TRANSDUCER

 

DIFFERENTIAL
AMPLIFIER

 

 

 

 

 

 

)

 

 

\\\\*ELECTRONKIPACKAGE

  

    

  

 

 

 

 

 

 

 

 

 

 

 

7 % FORCE INSENSITIVE MOUNTX
‘. )
% //
17 g
1V #
el /
07 v

g 7 EXCITATION ROD
% 4 f 7
7 7
] //
% L

/////// *PATENT NO. 2891180 AEROPROJECTS INC.

Fig. 2.10. MSRE Fuel Storage Tank Ultrasonic Level Indicator
oystem.

the frequency restored the instrument to an operative condition. Further
checks revealed that the frequency wvaried as much as 300 cps over a period
of a few days. Since the probe is basically a sharply tuned (high-Q)
resonant system, small shifts in oscillator frequency from the resonant
point caused the probe to become inoperative.

The problem of frequency drift was further complicated by the pres-
ence of a number of resonant peaks within the range of oscillator fre-
quency adjustment (some of which were not responsive to level changes)
69

and by the difficulty of checking instrument performance in the field
withoug interfering with operations (salt level must be varied to check
probe performance).

The difficulties experienced with the MSRE probe showed that some
improvements were needed if the device was to be useful for long-term
operation under field conditions. To gain a better understanding of the
problems involved, studies were made of the frequency response and per-
formance characteristics of the probe gystem. Because of the need to
minimize interference with MSRE operations, these studies were made using
the prototype probe system installed in the MSRP level test facility.
Results of tests performed showed that a number of resonant peaks existed
(see Fig. 2.11) and that, while several of the peaks were level sensitive,
only one was usable. When the molten-salt level was ralsed, one peak
disappeared when the salt touched the sensing plate. This peak was not
the one with the highest amplitude. Other peaks disappeared as the salt
level rogse and covered more of the excitation rod. Some peaks were not
affected by level.

The MSRE probe will be checked the next time that salt is transferred
to the fuel storage tank, using information gained and procedures devel-
oped from the studies discussed above. An energy-absorbing (Q-reducing)
slug will be installed in the excitation rod to broaden the bandwidth of
the resonant peaks and to reduce the effects of frequency drifts. The
oscillator will be stabilized by component changes and/or circuit revi-
sions. Installation of a notch filter in the amplifier to eliminate un-
wanted resonant peaks from the signal is being considered. As previously

10 [ ORNL-DWG 65-11801
|

| —— MOLTEN SALT BELOW SENSING PLATE H
-——— MOLTEN SALT TOUCHING SENSING PLATE |

--------- SENSING PLATE 0.4 in BELOW SURFACE fl
8 f--- —t—— OF MOLTEN SALT —f ... .

 

 

 

 

7 -

 

=

 

 

 

 

O

 
 

 

 

——

e e —

e
— |

|
]

et
i
T et —
oo
2 g 0 0 0§

 

 

 

LEVEL DETECTOR OUTPUT (mv dc)

el
oo 08B

 

 

 

 

 

 

 

 

 

 

 

i —
| o S
e ——

. . ) , {\
o4 d A
. .//J \\ : : II ‘\
r o
. \ . » A
e o\e 1, . ":.z\\ . fl°0 \\ o .. ~~ A

-----

O —
Lfim“b—-
|
e
&

3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

49,000 50,000 51,000 52,000
ULTRASONIC GENERATOR FREQUENCY (cps)

Fig. 2.11. Ultrasonic Level Detector Resonance Peaks.
70

reported,l4 replacement of the oscillator with one that would automatically
adjust to desired frequency is also being considered.

The manufacturer (Aeroprojects, Inc.) is studying the effects that
dimensional changes of the sensing plate will have on the resonant fre-
quency. This information will be of wvalue in predicting the effect of
corrosion on probe performance and might be of some value in determining
the rate and extent of corrosion during chemical-processing system opera-
tions.

The work reported above was done with the assistance and cooperation
of the probe manufacturer (Aeroprojects, Inc.).

High-Temperature NaK-Filled Differential-Pressure Transmitter

 

The coolant-salt system flow transmitter that failed in service at
the MSREL® was replaced with a spare transmitter after further field tests
and observations failed to reveal the reason for the shifts in zero and
span calibration.

An attempt to return the defective transmitter to the manufacturer
for repair was abandoned because the instrument was contaminated with
beryllium and the manufacturer did not have the facilities for handling
contaminated materials. The possibility of decontaminating the instrument
at ORNL was discounted when it was determined that the cost of decontami-
nation without further damage to the instrument would probably exceed the
cost of a new instrument.

A new instrument has been ordered for use as a spare. Two high-
temperature, INOR-8, diaphragm seal assemblies, fabricated and inspected
on the original purchase order, will be used in the fabrication of the
replacement instrument.

The defective tTransmitter is being tested at ORNL, under controlled
conditions, to determine the cause of the zero and span shifts and to de-
termine whether repair of the transmitter, at ORNL, is feasible. Results
of these tests to date are inconclusive; however, it has been determined
that the shifts are predominantly temperature-induced zero shifts which
are mechanical in nature. A 10°F change in ambient temperature around
the body of the transmitter caused the output to vary approximately 50%
of full scale. ©Static pressure changes applied equally to both high and
low inputs caused no appreciable shifts in output. The possibility that
the shifts originated in the strain-gage transducers or associated elec-
tronic circuitry was eliminated by disconnecting the strain gage and ob-
serving drift under various load and temperature conditions.

The results of the controlled tests confirm the conclusions reached
from field test and observationl® but have added very little to our under-
standing of the causes of the malfunction. An attempt will be made to
eliminate the shifts by refilling the transmitter body with silicone oil.
If this attempt is not successful, the transmitter will be disassembled
and visually inspected.
71

Float-Type Molten-5Salt Level Transmitter

 

During a recent shutdown of the MSRE prototype pump test loop, the
developmental ball-float-type level transmitterl® was dismantled and in-
spected to determine whether any buildup of vapor-deposited salts was
occurring which would affect future performance of the instrument. The
transformer was removed, and the core chamber was cut off at the top of
the float chamber. No significant deposits were found. A thin deposit
of salt was found on the part of the core tube that was normally in the
float chamber, but the amount was not enough to restrict motion of the
core. A sample of this material was taken for analysis.

Performance of the ball-float-type transmitter installed on the MSRE
coolant-salt pump continues to be satisfactory. The errors in calibra-
tion previously reported still exist; however, the information required
To correct the calibration has been obtained, and corrections will be
made when the system is filled.

Design of the ball-float transmitter installation in the MK-2 MSRE
fuel-circulating pump is in progress. All information necessary for this
installation has been transmitted to the pump designers. Due to consid-
erations of piping layout and pump configuration, it will not be possible
to design this installation to permit remote removal and replacement of
the differential transformer as originally planned. However, since no
failures have occurred on any of the transformers on prototype or field
test installations, it is expected that this compromise will not affect
the overall reliability of the transmitter.

Conductivity-Type, Single-Point Molten-Salt Level Probe

 

Modification and repair of the fuel flush tank level probe excitation
cable, which failed in service due to oxidation and embrittlement,17 was
completed during the precritical shutdown in March. Repairs were accom-
plished by replacing the copper-clad, mineral-insulated copper wire exci-
tation and signal cables and portions of the probe head assembly with a
stainless-sheathed, ceramic-beaded nickel-wire cable assembly. Consider-
able difficulty was experienced with breakage of the nickel wire in this
assembly during fabrication and installation. Investigations showed that
Tthe breakage was due to embrittlement produced by sulfur which originated
in potting materials (Thermostix and Ceramicite 200) used to seal the
cable against leakage of gas from the probe head. Although the amount of
sulfur found in these materials was very small and the materials had been
successfully used with other material (such as copper, stainless steel,
Inconel, Chromel, and Alumel), the quantities of sulfur present were ap-
parently sufficient to produce severe embrittlement of the nickel wire.
Since the potted seals were being installed only as a precautionary measure
and were not necessary for either contaimment or performance of the in-
strument, the seals were eliminated.

Performance of the repaired probe and of other probes installed in
the MSRE was satisfactory during subsequent critical and low-power opera-
tions of the reactor.
72

Inspection of the cable assemblies on probes installed in other fuel
and coolant system drain tanks showed that no damage had occurred in those
installations and that no further probe modifications were needed.

oingle-Point Temperature Alarm Switches

 

Modification of the temperature alarm switch modules to eliminate
the spurious set-point shifts experienced during precritical operationsl8
was completed. Printed circuit-board contacts were gold plated to reduce
contact resistance, the trim pots used for hysteresis adjustment were re-
placed with fixed resistors, and resistor values in modules having ambig-
uous (dual) set points were changed to restore the proper bias levels.

It is not known at this time whether these modifications have completely
eliminated the drift problems. The necessity of resetting many of the
modules during critical and low-power operation and of minimizing inter-
ference with reactor operations prevented the accumulation of the data
required to determine whether further shifts had occurred.

There was some evidence of set-point shifts in two of the freeze-
valve control modules; however, since the module could not be checked
during operation, it was not established whether the shifts were due to
module malfunction or poor connections in the thermocouple circuits.
Checks made during the prepower shutdown showed that additional modules
had developed dual set points; however, there was no evidence of recur-
rence of dual set points in modules which had previously been modified
to eliminate this problem. The occurrence of dual set points is presently
believed to be due to component aging and is expected to diminish as com-
ponents stabilize.

Module set points are being rechecked, using improved procedures,
as settings of various channels become firm. Associated thermocouple in-
put circuits are being checked to determine whether poor or intermittent
connections exist which could cause set-point shifts. Observation of the
performance of the temperature alarm switches will continue during power
operation of the reactor.

Helium Control Valve Trim Replacement

 

Results of investigation of the cause of failure of four helium con-
trol valves during MSRE precritical 0perationsl5 indicate that the severe
galling between the 17-4 PH plug and Stellite 6 seat was probably caused
by misalignment and the complete lack of lubricant rather than by incom-
patibility of the plug and seat materials. Inspection of the defective
valves revealed no serious misaligmment in the body, bonnet, or stem as-
semblies. The original trim was damaged too badly to determine whether
misalignments had existed; however, inspection of replacement trim re-
vealed that (in some cases) the hole in the seat was not perpendicular to
the seat within the tolerance required to prevent excessive side forces.
oeveral alternative trim material combinations were tested to determine
whether they would be less susceptible to galling under the conditions
of extreme cleanliness (no lubricant) and dry helium atmosphere. The com-
binations tested were Stellite to Stellite, chrome-plated 17-4 PH to
73

Stellite, and 440C to Stellite. None could be cycled more than 100 times
without evidence of incipient damage even when perfect alignment was main-
tained. However, the 440C to Stellite and the 17-4 PH to Stellite combi-
nations withstood 5000 additional cycles without evidence of further dam-
age after the plug was lubricated with a minute amount of light-grade
machine oil., There was some evidence that the properties of the 440C to
Stellite combination may be superior to those of 17-4 PH to Stellite.

Thermocouple Development and Testing

Drift Tests. Calibration drift of eight metal-sheathed, mineral-
insulated Chromel-Alumel thermocouples fabricated from material randomly
selected from MSRE stock has remained within the limits previously re-
ported.lg Figure 2.12 shows the drift of these thermocouples since the
start of the test.

ORNL-DWG 65-8089A

 

 

 

 

 

 

 

 

/fi‘\~f I ‘/f:_f AN

 

 

 

8 Tol I, %’
=1 /i/f\'# fi \}fi\i\f”?

o — | — ‘_—“‘—'—_—_—__'_-_—"—"—

 

 

 

ERROR (°F)

 

 

 

 

 

 

 

3/8 % OF 1250°F

TEST TEMPERATURE : 1250°F
> TEST ATMOSPHERE : AIR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O 50 100 150 200 250 300 350 400 450 500 550 600 620 700
TEST TIME (days)

Fig. 2.12. Average Drift of Eight MSRE-Type Thermocouples.

Thermocouples on the Prototype Pump Test Loop. Observation of the
performance of ten MSRE prototype surface-mounted thermocouples installed
on the prototype pump test loop was terminated in July, when the section
of pipe on which the thermocouples were installed was removed from the

loop. Performance of these thermocouples continued to be satisfactory to
the end of the test.

Coolant-5alt Radiator AT Thermocouples. Investigation of the effects
of mismatch of thermocouple and thermocouple lead-wire materials on the
accuracy of MSRE coolant-salt radiator AT measurement?® was continued.
Tests performed at the MSRE showed that the effects observed in the lab-
oratorygo were present in the MSRE thermocouple lead-wire installation
and that excessive noise was present on the signal. The existing lead
wire was replaced with a continuous run of higher-quality shielded lead
wire., Tests performed after the shielded lead wire was installed indi-
cated that the long-term drift previously observed had been eliminated
T4

but that excessive intermittent noise was still present. The thermocouple
lead wire was then insulated to eliminate ground loops. Further tests
are being performed to determine whether the noise has been eliminated.

References

 

1. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 29.

 

2. Ibid., p. 8.
3. Ibid., p. 12.

4s R. Do Ackley and W. E. Browning, Jr., Equilibrium Adsorption of Kryp-

 

ton and Xenon on Activated Carbon on Linde Molecular Sieves, ORNL-
CF-61-2-32 (Feb. 14, 1961).

 

5. MSRE Design and Operations Report. Part V, Reactor Safety and Anal-
ysis Report, ORNL-TM-732, p. 269,

 

6. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 50,

 

7' Ibido) po 500
8. Ibid., p. 51.

9. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp. 155—
60.

 

10. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 51—
52.

 

11l. MSR Program Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, pp. 123—
24

12. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 52.

 

13. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 166,

 

14. MSR Program Semiann. Progr. Rept. Feb., 28, 1965, ORNL-3812, pp. 38—
40,

 

15. Ibid., p. 48.
16. Ibid., p. 4l.
i7. Ibid., pp. 4l—42.
18. Ibid., pp. 46—47.
19. Ibid., p. 43.

20. TIbid., p. 44.
75

3. MSRE REACTOR ANALYSIS

Theory of Period Measurements Made with the MSRE
During Fuel Circulation

 

 

Zero-power kinetics experiments performed with the MSRE while the
fuel circulation is stopped can be analyzed by means of the conventional
inhour equation, which relates the measured asymptotic time dependence
of the neutron flux to the reactivity.l When the fuel is circulating,
however, the conventional analysis becomes inadequate due to the trans-
port of delayed neutron precursors by fluld motion and subsequent emis-
sion of the delayed neutrons in regions where they either do not contribute
at all to the chain reaction (i.e., in the external loop) or contribube
with substantially different weight than the prompt fission neutrons (snift
in the spatial distribution of emission within the reactor core). These
effects have long been recognized, and various approximations have been used
for their representation in reactor kinetics analysis.2 The particular
effect of the shift in the spatial distribution of delayed neutron pre-
cursors relative to the prompt fission neutrons has been the subject of
recent studies by'Wolfe3 and Haubenreich.# Wolfe employs a perturbation
approach and obtains an inhour equation for a slab reactor, through which
fuel circulates in the direction of axial variation of the neutron flux.
In an independent approach, Haubenreich considers an explicit analytical
model representing the MSRE in the circulating, just-critical condition
and, by means of a modal analysis, obtains effective values of delayed
neutron fractions for this condition. In the latter analysis, a bare-
cylinder homogenized approximation was used to represent the MSRE core,
with boundaries corresponding physically to the channeled region of the
actual core.

We have extended the preceding analyses to include the contribution
of delayed neutrons emitted while the fuel is in the upper and lower ple-
nums and to include the case in which the flux 1s changing according to
a stable asymptotic period. The result of this analysis is an inhour-type
equation, relating the measured period to the static reactivity of an
equivalent state in which the fuel is not circulating. The static re-
activity, p_, is defined by the relation”

vV — v

. C
ps vV J

in which VvV is the physical energy-averaged number of neutrons emitted
per fission and V., 1s the fictitious value for which the reactor would
be "virtually'" critical with the fuel stationary. Since the static re-
activity is the quantity normally obtained in reactor calculation pro-
grams, it 1s most convenient to relate this quantity directly to the
asymptotic period. This analysis is summarized below. Only the essen-
tlial steps are indicated in obtaining the inhour relation. Specific
approximations sultable for MSRE analysis, together with preliminary nu-
merical results, are given following the discussion of the general prob-
lem. Further details of the analysis will be included in a subsequent
report.
76

When the flux and precursor densities are behaving in time as ewt
Tthe general reactor equations, written to include the transport of de-
layed neutron precursors by fluid motion in the axial direction, are:

J

6

— - = -1
Do + (1 BT)fPP¢ + L A f vrwe (1)

. diCi

1

d o
B;Pe — AC, — 5= (vci) = uC, i=1,2, oo, 6 . (2)

The symbols ¢ and C represent the neutron flux and delayed precursor
densities, which in general are functions of position, energy, and angle
variables. The operators D and P, assumed time independent, represent
neutron destruction (leakage, absorption, energy transfer by scattering)
and fission production processes, respectively. Their explicit repre-
sentation depends on the model used in analysis. In Egs. (1) anda (2),

1 — Bp is the fraction of all neutrons from fission which are prompt,
and PB; and %i are the fractional production and decay rate for the ith
precursor group. The quantities fp and f5 are energy spectrum operators
which multiply the total volumetric production rates of prompt and de-
layed neutrons to obtain the net production of neutrons of a specific
energy. The symbols v and V represent the neutron velocity and the fluid
velocity, respectively.

As defined above, the static reactivity is the algebraically largest
eigenvalue of the equation:

~Do_ + (1 - ps)fP¢S =0, (3)
where
{5;\
T = (1 —-fiT)fp + fii B.f . - (4)

In order to relate the reactivity, Pgs» to the stable period, wrl} use 1is
made of the static adjoint flux, ¢§, the solution of the adjoint equation
corresponding to Eq. (3), as a welghting function in the integration of
Bgs. (1) and (2). The purpose of this procedure is to convert the re-
actor equations from a form which involves only local reactor properties
to a relation which utilizes global, or integral, quantities. This anal-
ysis 1s similar in principle to that for stationary fuel reactor systems,
details of which have been presented in several sources (see, e.g., ref,
5). By forming the inner product of of with Eq. (1), that is, by multi-
plying Eq. (1) by ¢& and integrating over the position, energy, and angle
variables, one obtains the relation

6 6
+
(o1, vo) .Z; Bi(¢g, £4;P¢) —-.Z; A (o, £4:C;)
0 = — . 1=1 ~ 1=1 5 (5)
® (of, TPo) (¢, TPo)

 

 
77

where the symbol (x, y) denotes the inner product of the two functions
x and y. In the special case when w = 0, this equation gives the static
reactivity difference between the stationary-fuel and circulating-fuel
critical states. In the general case, the dependence of the precursor
densities on w is given implicitly by Eg. (2). Several approximations
sultable for MoORE analysis have been made in performing the integrations
of Egs. (2) anda (5):

(a) The shape of the asymptotic flux distribution, ¢, is assumed to
be sufficiently well approximated by the static flux, o5, corresponding to
the stationary fuel state. This is essentially the perturbation approxi-
mation.

(b) The correction for the difference in the energy spectra for
emission of prompt and delayed neutrons appearing in Eqg. (5) can be cal-
culated approximately as a separate step. This is done by reducing the
age for the ith group from that of the prompt neutrons, and using the
relative nonleakage probability factors appropriate to a bare reactor.
This is equivalent to modifying the values of the static fraction, B .
The net correction for the differing "energy effectiveness" of the de-
layed neutrons is small compared with the spatial transport effects under
consideration.

(c) The principal difference in the spatial distributions of prompt
and delayed neutron emission 1s in the direction of fuel salt flow. The
veloclty profile was assumed to be flat in the radial direction across
the core. Also, for this initial study, radial averaging of the neutron
%r§duction rates was neglected in calculating the inner products in Eq.

5).

ORNL-DWG 6512048
0.8

 

 

 

 

 

 

 

 

0.7 - /
. //
os /
Y/
v
2 05 /
- /|
C /
= 0.4 ~ /|
2 yd
Ll L
* 03 7
S T
—“‘/
0.2 =T

 

 

O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10?5 2 1021072 o 5 1072 2 5 10! 2 5 100 2 5 10!
w, FREQUENCY (sec™!)

Fig. 3.1. Static Reactivity Difference Between otationary-Fuel
Critical Condition and Circulating-Fuel Transient Condition vs Stable
Inverse Period Observed During Circulation.
78

Equation (2) was integrated numerically, using a three-region approxi-
mation for the MSRE core, representing the channeled region and the upper
and lower plenums. The fluid axial velocities in each region corresponded
to the average fuel residence times for that region. The axial distribu-
Ttlions of the static neutron flux and adjoint flux corresponded to the case
of all three control rods fully withdrawn from the reactor core. Lffective
fuel residence times in the core regions and the external loop were ob-
tained from ref. 6. The numerical results obtained from the analysis are
summarized in Fig. 3.1, in which the static reactivity is plotted as a
function of the inverse asymptotic period, w. Only the range of values
of w important to experimental measurements of the stable period were con-
sidered in this initial study. The calculated reactivity difference be-
tween the static-fuel and circulating-fuel critical conditions (0.22%,
corresponding to w = 0 in Fig. 3.1) agrees favorably with the value of
0.21%, measured by excess addition of 23°U during the zero-power experi-
ments. This analysis is now being applied to rod calibration measure-
ments made during fuel circulation, and the results will be included in a
future report.

References

 

1. 5. Glasstone and M. C. Edlund, The Elements of Nuclear Reactor Theory,
chap. X, Van Nostrand, New York, 1952.

 

2. dJ. A. Fleck, Jr., Theory of Low Power Kinetics of Circulating Fuel
Reactors with Several Groups of Delayed Neutrons, BNL-334 (April 1955).

 

 

3. B. Wolfe, Nucl. Sci. Eng. 13, 80-90 (1962).

 

4. P. N. Haubenreich, Prediction of Effective Yields of Delayed Neutrons
in the MSRE, ORNL-TM-380 (October 1962).

 

5. T. Gozani, "The Concept of Reactivity and Its Application to Kinetic
Measurements, " NUcleonik!g(Z), 55 (1963).

6. R. B. Lindauer, internal memorandum (June 1964) , p. 8.
Part 2. MATERIALS STUDIES
81

4o METALLURGY
Dynamic Corrosion Studies

A test program is in progress to study the compatibility of struc-
tural materials with fuels and coolants of interest to the Molten-Salt
Program. Thermal convection loops described previouslyl are used as the
standard test in this program. Presently, two long-time tests are in
operation with fuel salts at conditions reported in Table 4.1.

Table 4.1. Operating Conditions for Thermal Convection Loops Containing
LiF-BeF-ZrF 4-UF 4-ThF, (70-23-5-1-1 Mole %)

 

 

Loop 1255 Loop 1258
Loop material INOR-8 Type 304 SS
Maximum salt temperature 1300°F 1250°F
AT 160°F 180°F
Operating time as of 29,688 hr 18,312 hr
Auvg. 31, 1965
Insert specimens in INOR-8-2% Cb Type 304 SS

hot leg

 

The INOR-8 loop contains insert specimens made of INOR-8 modified
with 2% Cb to improve weldability and mechanical properties. In addition
to studying the corrosion resistance of this modified alloy, the loop
serves to demonstrate the long-time compatibility of INOR-8 with an MSRE-
type fuel salt.

The type 304 stainless steel loop is being operated to investigate
the potential of this lower-cost alloy for molten-salt applications. This
material 1s being reconsidered at present in view of the lower tempera-
tures of interest for coolants and the improvements in salt processing
that result in lower corrosion rates.

Specilmens were removed from the stainless steel loop after 15,000 hr
of operation and examined metallographically. A maximum attack of 0,002
in. was observed that was generally intergranular in nature (see Fig.
4.1). This loop is continuing to operate.
82

 

.035 INCHES
N 100X

[

 

 

 

 

\

Fig. 4.1. Appearance of Specimen Removed from Region of Maximum
Temperature (1250°F) in Thermal Convection Loop 1258 Made of Type 304
otainless Steel. The loop had operated for 15,000 hr at this time.

Corrosion Studies Using Lead as a Coolant

 

Six uninhibited thermal convection loops were started at ORNI to ex-
plore the compatibility of structural materials with lead at conditions
expected for molten-salt reactor coolants. The operating conditions of
these loops are summarized in Table 4.2, along with results of a 2-1/4
Croloy steel loop run at Brookhaven National Laboratory2 in which lead
was circulated at 1022°F for several years and in which no measurable cor-
rosion was Observed.

Two of the ORNL loops (one of type 410 stainless steel and of 2-1/4
Croloy steel) had Cb—1% alloy liners in the surge tanks within which MSRE
fuel salt floated on the lead surface, as shown in Fig. 4.2. These tests
also contained graphite specimens in the surge tank which were exposed
to the salt, to the lead-salt interface, and to the lead. A hot-leg tem-
perature of 1200°F was maintained in the loops with a 300°F AT. Two other
loops that contained no salt, Cb—1% Zr, or graphite, were operated at
1100°F with a 200°F AT. One of these loops was constructed of 2-1/%
Croloy steel, the other of low-carbon steel. A fifth loop was operated
to test the compatibility of columbium alloys with lead. This loop was
made from Cb—1% Zr clad 446 stainless steel and contained no samples or
inhibitors.

Of the two loops that operated at 1200°F and contained salt, the
2-1/4 Croloy steel plugged after 288 hr, and the type 410 stainless steel
plugged after 1346 hr. The plugs were made up of dendritic crystals,
83

Table 4.2. Operating Conditions of Thermal Convection Loops Circulating Lead

 

 

Maximum AT Operating
Loop Material Temperature ° Time Comments
o (°F)
(°F) (hr)
Croloy 2-1/4% 1022 227 27,765 25 ppm Mg, Zr inhibitor added
Croloy 2-1/4 1210 300 266  Graphite, fuel salt, Cb—l% Zr
placed in surge tank
ATST type 410 SS 1215 300 1,346 Graphite, fuel salt, Cb—1% Zr
placed in surge tank
Croloy 2-1/4 1100 200 5,156
Low-Carbon Steel 1100 200 5,064
Co—1% Zr (clad with SS) 1400 400 5,280
Croloy 2—1/4b 1200 230 1,848 Zr specimen in hot leg

 

aBrookhaven National Laboratory loop.
bPrototype test of new loop design.

which were determined to be iron and chromium by x-ray diffraction and
chemical analysis. The maximum depth of attack in the hot leg of the 410
stainless steel was 0.002 in. and on the 2-1/4 Croloy loop was 0.001 in.
Typical attack is shown in Fig. 4.3. No corrosion was observed on Cb—1%
Zr liners.,

The 2—1/4‘ Croloy steel loop that operated with a hot leg at 1100°F
showed signs of plugging after 648 hr of operation. A radiographic ex-
amination of the dumped loop at this time indicated several areas where
lead had wet the metal, indicating selective attack. The lead from the
loop contained crystals of iron and chromium. The loop was refilled with
new lead and operated for a total of 5156 hr. Metallographic examination
of the loop revealed large amounts of dendritic crystals in the cold leg
and also revealed hot-leg attack to a depth that varied from 0.004 to
0.008 in. The crystals in the cold leg were similar in composition to
the original alloy.

The carbon-steel loop was constructed of a larger-diameter pipe and
operated for 5000 hr before shutdown. However, prior to being shut down,
it did show some signs of restricted flow. Metallographic examination
showed a maximum attack of 0,010 in., similar in nature to that found in
the 2-1/4 loops. The Cb—1% Zr loop operated for 5280 hr with a hot-leg
temperature of 1400°F and a 300°F AT. Metallographic examination showed
no hot-leg attack or crystals in the cold leg.

Results on the tests made to date indicate that uninhibited lead will
cause excessive corrosion in low-alloy steel and stainless steel systems
at temperatures above 1100°F., However, Cb—1% Zr appears to be a promising
container. Inhibitors should be investigated in steel systems in view
of the BNL results.
84

     
   

| . PHOTO 74545A
VENT LINE . . PROBE LINE TO DETERMINE
~ LIQUID LEVEL

FILL LINE
GAS LINE

STOP TO HOLD LINER
IN PLACE

Cb—4 Zr LINER

LEAD SALT INTERFACE

 

TOP OF HOT LEG

Fig. 4.2. Section Through the Surge Tank Used on the Type 410
octainless Steel Loop. Notice that the salt is floating on the lead
and that the only metal which contacts the salt is the Cb—1% Zr liner.
The graphite specimens were suspended from the small wire running into
the salt and cannot be seen in the picture.
85

 

 

Ton

0.007 IN HES
> 500X

 

 

Fig. 4.3. Metallographic Section of a 2-1/4 Croloy Loop Run in
Lead at 1200°F for 266 hr. Unetched.

Revised Thermal Convection Loop Design

 

The thermal convection loop was redesigned for studying coolants and
now includes the following features:

1. removable hot-leg samples (for weight gain data and metallographic
analysis),

2. 7portable thermocouple control in the hot leg to plot temperature pro-
files,

3. surge tank with Cb—1% Zr liner and graphite to allow the presence of
fuel salt,

4., smaller size to allow radiographing the complete loop at once,
5. sampling devices for both lead and salt,

6. a method by which the lead may be drained and refilled without af-
fecting the salt so that the hot leg can be radiographed to find se-
lective attack and so that mass transfer may be removed by gravity
separation,

7. better control of the cold-leg temperature.
86

Figure 4.4 shows the revised thermal convection loop design.

In order to obtain better control over the cold-leg temperature, a
heater was placed on the cold leg upstream from the coldest part of the
loop. This heater 1s controlled by the cold-leg thermocouple. A shut-
down device to eliminate burnout was also incorporated which turns the
power off when the temperature of the cold leg is equal to the temperature
of the hot leg. With this design, the degree of plugging may be deter-
mined by the power consumption of the cold-leg heater.

A prototype loop of the new design was run at conditions listed in
Table 4.2 to test the new features of the loop, the temperature distribu-
tion, and the sample removal and cleaning techniques. The cold-leg tem-
perature control worked well, as did the sample-removal and dumping fea-
tures. The samples were cleaned by amalgamating the lead with mercury

ORNL—DWG 65—-4744

/,-in. CONNAX FITTING
Hrkp FILL FILL
: AND SAMPLE~_ “AND SAMPLE

EVACUATION LINE

   
   
 
 
  
 

 

 
 
   
 
    

WCORCHESTER
{-in. BALL VALVE

SURGE TANK

HOT LEG CONTROL
THERMOCOUPLE

GAS

 

V4-in. BUTT WELD
SWAGELOCK

 

34 in. OD x.0.089 in. WALL
THICKNESS

 

    
 

SPECIMENS
SUSPENDED
FROM END 4-in, - o
CLAMSHELL -in.
HEATER—— SCHED-40 PIPE
| 4Y/p-in. LONG

6-in. CLAMSHELL :
HEATERS {

 

 

{“_—— T T —_——_-_-_—.]I i
C—C— a3 e )~
e ]
Lo e
\J
=

 
  

 
  

 

 

 

 

  

 

 

 

 

 

E - GRAPHITE
71| 1 COLDLEG -1 Zr
REMOVABLE ' CONTROL LINER
SAMPLES . THERMOCOUPLE
i ////) SLOTS IN GRAPHITE
e ENSURE LEAKAGE SALT
4-in. | LT ‘ R
CLAMSHELL HEATER . . LEAD
SLOTS IN
DUMP TANK—~] Cb-1 Zr LINER
ENSURE LEAKAGE LOOP PIPING
THERMAL CONVECTION LOOP SECTION THROUGH SURGE TANK
USED FOR LEAD CORROSION STUDIES OF THERMAL CONVECTION LOOP

USED FOR LEAD-SALT STUDIES

Fig. 4.4. Sketch of Prototype Loops Used to Study Lead-Salt Low-
Alloy Steel Systems.
87

and then dissolving the amalgam with concentrated nitric acid. A re-
movable sample made of zirconium was included in the test and did not
change weight. The loop plugged after 1848 hr of operation and showed
a maximum attack of 4.5 mils. This high rate of attack indicates that
the zirconium metal specimen did not inhibit corrosion in the loop.

MSRE Materials Surveillance Testing

 

The MSRE surveillance specimens® of INOR-8 and grade CGB graphite
have been assembled and placed in the reactor and control test systems.
These specimens will be used to make periodic surveys of the effects of
the reactor operations on the materials from which the reactor and moder-
ator were constructed.

Three stringers of graphite specimens, each with a pair of INOR-8
Tensile specimen assemblies and flux monitors, were placed in the central
position of the MSRKE core and extend axially the full height. Matching
sets of graphite and INOR-8 specimens were installed in the control test
rig, which will subject these control specimens to molten fuel salt under
approximately the same temperature profile and major temperature fluctua-
tions experienced by the reactor specimens.4 Three other sets of INOR-8
Tensile specimens and flux monitors without graphite specimens have been
installed outside the reactor vessel adjacent to the reactor flow dis-
tributor.

The specimens in the reactor core will be subjected to a range of
temperatures and neutron fluxes (shown in Fig. 4.5) that bracket the tem-
peratures and fluxes of the reactor core materials. Thus results from

ORNL-DWG 65-12049

//\QST FLUX
1220

(x10'3)

 

 

o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T - \ GRAPHITE TEMPERATURE
o >\ (APPROX)
n —— T —
. ~ |
- // | 1210
5 / L | ——==—FUEL
. P _ TEMPERATURE _
- 7 ~ L
3 P
2 6 — g 1200 W
£ _ _ 5
x v ~ <
5 - > 5
L 4 -~ . \ oo &
z / // // ~ U§J
o

& s - \ =
5 7~ |

2 //,‘ ' THERMAL FLUX— 3 1180

-
0 170
0 10 20 30 40 50 60 70 80

DISTANCE ABOVE BOTTOM OF HORIZONTAL GRAPHITE GRID (in.)

Fig. 4.5. Temperatures and Neutron Fluxes Along Graphite-Sample
Assembly. Control rods at upper limits.
88

these specimens will anticipate the condition of the reactor vessel as
well as match the condition of both the thimbles and the reactor vessel.

The specimens adjacent to the reactor vessel will be exposed to neu-
tron fluxes that more nearly approach those at the reactor vessel wall.
Also, the sgpecimens will not be exposed to the salt environment, facili-
tating any comparisons of test results with data developed in the Irradi-
ation Test Program.

INOR-8 Surveillance Specimens

 

The INOR-8 specimen assembly consists of 27 tensile bars approxi-
mately 2 in. long and with a 0.125-in. gage diameter welded end to end.
The first and the 27th specimen are machined from weldments. The lengths
of the sections between tensile bars have been adjusted to hold the graph-
ite specimens firmly in the final assembly.

The specimens in the core and control test were made from heats Ni-
5085 and Ni-5081 material. Ni-5085 is the heat from which most of the
reactor vessel shell is made; Ni-5081 is one of the original heats used
to develop much of the mechanical properties data for INOR-8. Both heats
have been investigated in the INOR-8 radiation damage program. The weld
specimen was made with heat Ni-5055 weld rod and heat Ni-5085 base metal.

The specimen assembly rods outside the reactor vessel and the flux
monitors are 82 in. long and parallel the vertical axis of the reactor.
These specimens are similar in shape to the reactor core specimens but
are made from heats Ni-5085 and Ni-5065. Ni-5065 is the heat used to make
the reactor top and bottom heads and is also being tested in the radia-
tion-damage program.

The analyses of the INOR-8 specimens will include: (1) metallo-
graphic examination for structural changes, corrosion effects, and pos-
sible layer formations; (2) tensile and creep properties, with emphasis
on creep ductility; (3) a general check for material integrity and di-
mensional changes; and (4) chemical analyses for compositional changes
and fission product deposition.

Graphite Survelllance Specimens

 

The graphite specimens were machined perpendicular to and parallel
with the extrusion (grain) directions of the graphite bars listed in Table
4.3. The stringer bars are the vertical bars with grooved fuel channels
that constitute about 98% of the moderator volume. The lattice bars are
the horizontal criss-crossed bars that support the stringers. The lattice
bars were fabricated with a slightly higher permeability than the stringer
bars in order to secure maximum structural integrity.
89

Table 4.3. Typical Data on Grade CGB Graphite Used to
Fabricate MSRE Surveillance Specimens

 

Electrical Resistivity

 

 

bulk (microhm-cm)
Bar No. Lot No. Bar Type Density - -
(g/ch) Parallel Perpendicular
635 8 Stringer 1.853 640 1210
1229 11 Stringer 1.848 690 1400
1559 13A Lattice 1.868 635
1564 13A Lattice 1.881 630

 

“The orientation is with respect to the extrusion, the grain, or ag
direction.

There appears to be somewhat more turbostratic graphite in some of
the stringer bars. Since this less-graphitic material might show more
irradiation-damage effects than the more crystalline graphite, specimens
containing more than the normal quantity of turbostratic graphite were
included with the surveillance specimens. Specimens from bars Nos. 635
and 1229 in Table 4.3, respectively, represent the stringer bars and the
slightly-less-graphitic graphite. The higher electrical resistivity of
bar No., 1229 reflects the latter.

Typical graphite specimens used in a set of surveillance specimens
are shown in Fig. 4.6. All these can be utilized in the (1) metallo-
graphic examination for structural changes and material deposition; (2)
radiographic and autoradiographic examination for salt permeation and
possible wetting effects; (3) dimensional checks for shrinkage effects;
and. (4) measurement of physical properties such as electrical conduc-
Tivity, thermal conductivity, and Hall coefficient. Chemical analyses
for salt and fission product deposition will be made on the larger pieces,
A and B in Fig. 4.6. These have cross sections of 0.470 X 0.660 in. The
lower specimen in Fig. 4.6A is from the lattice bar and will be located
at the bottom of the moderator core. The specimen at the top of Fig. 4.6A
and. the specimen B in Fig. 4.6, respectively, are located at the top and
center of the moderator core and are from stringer bars.

The specimens shown in Fig. 4.6C are essentially multiples of a basic
0.11 X 0,47 X 2,25 in. chape, the smallest pieces. These are to be used
to make flexural strength and modulus of elasticity measurements. The
small specimens can be tested directly, while the long specimens will have
to be cut to the proper lengths. The objective was to preclude machining
of irradiated graphite in hot-cell facilities. Such machining would be
difficult and could modify some of the property changes in the graphite.
These specimens, as shown in Fig. 4.6C, when placed together create a
cross-section dimension that matches those of the larger pieces discussed
above,
20

Y-64822

 

C

 

A B

Fig. 4.6. Typical Graphite Shapes Used in a Stringer of Surveil-
lance Specimens.

Assembly of Surveillance Specimens

 

A subassembly of graphite and INOR-8 surveillance specimens bound.
with INOR-8 straps is shown in Fig. 4.7. For the reactor, three sets were
bound together and placed into a perforated tube as shown in Fig. 4.8.

The specimens were sealed in the container by the ball lock assembly (Fig.
4.8b) and placed into the reactor (Fig. 4.8c).

Similarly, the matching control specimens were assembled in three
sets; however, instead of being bound together and placed into a common
container, each set was placed into a separate, sealed perforated con-
taingr as shown in Fig. 4.9. These were placed in the controlled test
rig.

The flux monitors included in the assemblies placed in the core and
just outside the reactor vessel are 0.020-in.-diam wires of pure iron,
pure nickel, and a type 302 stainless steel. The type 302 stainless wire
91

PHOTO 80761

 

Fig. 4.7. Three Stringers (S, R, and L) of CGB Graphite and
INOR-8 Surveillance Specimens.

PHOTO 81671
SPACER AND GUIDE PIN

UPPER GUIDE ———s= BASKET LOCK\ ASSEMBLY

    
   
  
 
  
 
  

   
  
   
 
   

INOR-8 ROD OF TENSILE SPECIMENS
GRAPHITE (CGB) SPECIMENS
BINDING STRAP

BASKE T—_

 

FLUX MONITORS TUBE

BASKET
AN

\ BALL LOCK ASSEMBLY

REACTOR ¢

  
  
 
 

CONTROL ROD

GUIDE TUBE
TYPICAL GUIDE BAR
FUEL CHANNEL
I REACTOR ¢
o \
Oc

0.400 in.
() SURVEILLANCE SPECIMENS

2in. TYPICAL

(c) R=REMOVABLE STRINGER
Fig. 4.8. 1INOR-8 and Grade CGB Graphite Surveillance Specimens
and Container Basket. (a) opecimens partly inserted into the container.
(b) Container and its lock assemblies. (c) Location of surveillance

specimens in the MGRE.

was selected for 1its uniform cobalt content. The selection of materials

for flux monitors was limited by the relatively high temperature, 1225°F,
of the reactor.
R

PHOTO 81672

Y-66055

   
 

LIFTING TOOL ~ “

     

SPECIMENS

  

T~ BASKET

(a)

Fig. 4.9. One Stringer of Control Specimens of INOR-8 and Grade
CGB Graphite for the MSRE Surveillance Specimens Shown with Its 73-1/2-
in.-long Container. (a) Bottom view. (b) Top view.

The flux monitor wires installed outside the reactor vessel are bare,
while those placed near the central part of the moderator are protected
from exposure to the salt. These have been sealed into an evacuated,
1/8-in.-0D by 0.20-in.-wall INOR-8 tube by tungsten—inert-gas arc welding.

At an appropriate time determined by the radiation-damage program,
examinations as described above will be made on a set of specimens from
each of the three test sites: (1) the central part of the reactor moder-
ator, (2) the controlled test rig (the control test for those from the
reactor), and (3) outside but adjacent to the reactor vessel. A set of
specimens constitutes one-third of the total at each test site. As these
are removed, sets of specimens of like or of advanced materials will be
placed into the empty positions. The sampling and replacing of specimens
will be sequential for each one-third set at each test site.
93

kEvaluation of Radiation-Damage Problems of Graphite
for Advanced Molten-Salt Reactors

 

Recent progress in developing and testing nuclear grades of graphite
provides a basis for specifying the properties and estimating the life
of a graphite for advanced molten-salt reactors such as the MSBR. There
are, of course, several uncertainties which limit the ability to state
categorically that any graphite will withstand the MSBR environment. The
graphite in the regions of highest flux in an MSBR will be irradiated to
doses of 2 to 5 X 10%% nv (>0.18 Mev) per year, and there is no evidence
to demonstrate that any graphite can withstand massive doses of >10%2 nvt
and still retain its integrity. Isotropic graphite has demonstrated the
greatest potential for accommodating this condition. The tubular, thin-
walled design that is proposed for the MSBR is one of the better config-
urations for reducing stresses resulting from differential growth of the
graphite. The tubular shape also is easier to fabricate and to test non-
destructively to ensure maximum integrity.

The pertinent properties of the graphite grade that could be devel-
oped for the MSBR can be projected from those of available grades with
some degree of certainty. The main questions to be answered relate to
the growth and strength of graphite under irradiation. Values for growth
and strength can be estimated fairly well to dose levels where experi-
mental evidence exists; however, it would be presumptuous to extrapolate
these growth and strength values beyond dose levels of 3 to 4 X 102 nvt.

The magnitude of the differential growth problems depends on the
growth rate, creep coefficient, flux gradient, and geometric restraint
in the particular structural component. With a tube 4 in. in outside
diameter, a flux drop of approximately 10% could be expected across a
l/Z-in. wall. The growth rate for an isotropic graphite at 700°C is con-
servatively estimated to be about half®’? that for grade AGOT or about
2.4 %X 107%% (in./in.) (nvt)™L. The creep coefficient® should be about
that for grades AGOT and CGB or about 7 X 107%° (in./in.) (psi)~! (nvt)~ 1.
The geometric restraint for a tubular configuration of this size 1is very
close to 1/2.

The problem is solved by obtaining the maximum strain rate caused
by the differential growth, which is:

s (é?i) (6)(R) = #1.2 X 10725 (in./in.) (avt)~! ,

¢
where
¢ = mechanical strain rate,
4%12 flux drop, 0.10,
G = growth rate, —2.4 X 10724 (in./in.) (nvt )™,
R = restraint factor, *1/2.
9

This strain rate is maximum on the inside surface and is minimum on the
outside surface as in a tube with a thermal gradient induced by heating
from the outside. The maximum stress is then obtained from the creep co-
efficient.

—25
- L2 X 10 = 1.7 x 10° psi ,

0.7 x 10~<8

 

Mo

where o 1s maximum fiber stress, and K is creep coefficient. The bend
fracture stress of isotropic graphite is expected to be >5000 psi; so the
stresses induced by differential shrinkage should not cause the tubes to
fail, at least to doses of 4 X 10°% nvt. Also, there is some evidence”?
that the shrinkage rate of isotropic graphite diminishes after 1 to 2 X
102 nvt. In effect, this will reduce the stress level and introduce more
conservatism into the extrapolation to higher dose levels.

The ability of the graphite to absorb the creep strain regardless
of the stress intensity has been demonstrated. ’? 8 Although a strain lim-
itation to fracture has not been demonstrated, one probably exists. This
limitation, however, is greater than the 2% tensile strain observed by
Perks and Simmons.,’ Thus, to obtain a 2% maximum fiber strain, it would
require a 1.6 X 10°2 nvt dose if the shrinkage rate remains constant.

Existing data show, then, that irradiation effects should not produce
failures in one year in the graphite in the highest flux regions of a
large MSBR. The data can be used to infer a life of several years. The
actual life of the graphite will, however, remain uncertain until its
ability to withstand irradiation damage beyond 1023 nvt has been demon-
strated.

Mechanical Properties of Irradiated INOR-8

Mechanical Properties of Unirradiated INOR-8 Used in the Reactor
Vessel

A program has been initiated to evaluate the properties of the heats
of INOR-8 used in constructing the MSEE reactor vessel. Although mechan-
ical property tests were conducted previously on several MSRE heats,g the
radiation effects were not superimposed. The present studies include
further mechanical property tests on the heats of material comprising the
reactor vessel in both the wrought and welded conditions to evaluate their
properties under service conditions. The obJjectives of this study are
to estimate the safe operating life of the MSRE and to determine whether
any reasonable steps can be taken to increase this 1life,

oignificant experiments and conclusions in this program include the
following:
95

1. The creep strengths of the wvarious heats of material appear to
be comparable with those reported previously.9 However, the creep-rupture
data appear to be more accurately represented by an equation of the form

e’ = AtB than the more conventional form of o = AtB. Hence, the extrapo-

lated values are altered slightly. This is illustrated by the data shown
in Figs. 4.10 and 4.11 for heat Ni-5065.

2. The properties of the alloy at 650°C are very sensitive to me-
chanical and thermal treatments. The data in Table 4.4 show that, in
general, both annealing at temperatures below 1600°F and cold working im-
prove the rupture life and ductility. If working and annealing are con-
tinued sufficiently to cause recrystallization, the strength is reduced.
Annealing at 2300°F caused a significant reduction in rupture life.

3. Welding the air-melted heats of material caused a large reduction
in rupture life and ductility. This is also illustrated by the data in
Table 4.4. The weld between the top head and the flow distribution ring
of the MSRE, which was not stress relieved, was reproduced as nearly as
possible. ©Since this weld was found to have very poor properties in the
as-welded condition, the rate of recovery of strength and ductility was
measured at several temperatures. The results of these tests are given
in Figs. 4.12 and 4.13. Properties similar to those of the base metal
can be obtained by stress relieving for 8 hr at 1600°F or for 50 to 100
hr at 1400°F,

4o Welding of a vacuum-melted heat of INOR-8 was found to produce
very small changes in properties. This is illustrated in Fig. 4.14.

5. Microprobe and autoradiographic studies have shown that the large
precipitates in INOR-8 are intermetallics rather than carbides. The
structure is typified by the microstructure shown in Fig. 4.15. When the
alloy is heated to about 2500°F, the microstructure reverts to that shown
in Fig. 4.16. A grain-boundary lamellar phase results, with some local-
ized melting. The transition from one phase to another is illustrated by
the photomicrograph of a fusion line shown in Fig. 4.17. Table 4.5 pre-
sents the results of microprobe analyses on both types of precipitate.
The precipitates are basically nickel-molybdenum intermetallic phases
with large amounts of silicon also present. The spherical phase is ap-
proximately Ni-Mo (8), and the lamellar phase is Thought to be NisMo (7),
since iron and chromium are both known to suppress B-phase formation in
nickel-molybdenum alloys. The autoradiograph shown in Fig. 4.18 was made
on a heat of INOR-8 containing the isotope 140, The light band shows
where the emulsion has been scraped from the surface to reveal the la-
mellar phase., The film is present on the dark area, and the darkening
is produced by beta-ray emissions from the 140, The lamellar precipitate
is actually depleted in carbon rather than enriched.

6. Several experimental alloys are being studied. The amount of
molybdenum is being reduced to about 12 wt % to produce a solid solution,
and the silicon content is being reduced to reduce the "hot-short" char-
acteristics of the alloy. Various alloy additions, such as zirconium,
titanium, and niobium, are being investigated as a means of improving the
resistance of the alloy to neutron irradiation.
96

ORNL-DWG 65-10806

100
80

60

40

20

STRESS (1000 psi)

 

1 10 100 1000 10,000 100,000
RUPTURE LIFE (hr)

Fig. 4.10. Stress-Rupture Properties of Hastelloy N (Heat 5065)
at 650°C.

ORNL-DWG 65-10805R

 

S T

A e AS RECEIVED

 

o/

 

 

70

 

®

60

 

50 S

 

40 J

 

STRESS (1000 psi)

30 \\\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\
N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 10 100 1000 10,000 100,000
RUPTURE LIFE (hr)

Fig. 4.11. Creep-Rupture Properties of Hastelloy N (Heat 5065)
at 650°C.
97

Table 4.4. Influence of Cold Working and Heat Treatment on the
Creep Properties of Hastelloy N (Heat 5065)

o = 40,000 psi, T = 650°C

 

 

 

Heat Trestment Rupt%ii)Life Elon%%gion iieiggzi?%)

As received 312.3 16.6 15.6
2 hr at 871°C 502.8 48.4 40.8
8 hr at 871°C 652.3 54.7 30.3
200 hr at 760°C 566.3 46 .9 35.3
Cold worked (C.W.) 0% 179.7 9.4 8.1
C.W. 10%, 2 hr at 871°C 729.1 57.6 31.7
C.W. 10%, 8 hr at 871°C 373.5 37.5 21.8
C.W. 20% 316.1 28.1 8.5
C.W. 20%, 2 hr at 871°C 414 .3 29.7 34.1
C.W. 20%, 8 hr at 871°C 156.3 25.0 24.0
1 hr at 1260°C 34.0 A 13.2
As welded 18.7 1 0.77

. ORNL-DWG 65-10797

O

 

2
<
Ll
o 2
<
=
=Z 1
o
i_
O
D
8 0.5 o = 40,000 ps
o 7 =650°C
871°C
0.2 760°C
650°C
O.1

0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 {000
TIME AT TEMPERATURE (hr)

Fig. 4.12. Influence of Stress Relieving on the Rupture Ductility
of Weld 1.
(hr)

RUPTURE LIFE

98

ORNL-DWG 65-10798

 

1000
40,000 ps
650°C
500 871°C
760°C
650°C
200
100
50
20
10 '
041 02 0.5 1 > 5 10 20 50 100 200 500 1000 2000

TIME AT TEMPERATURE (hr)

Fig. 4.13. Influence of Stress Relieving on the Rupture ILife of
Weld 1.

ORNL-DWG 65-10808
70

 

 

 

 

 

 

 

 

 

 

 

 

 

N | 2477 BASE METAL

 

50

 

40

 

 

 

 

 

STRESS (1000 ps1)

DN
30 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10

 

1 10 100 1000 10,000
RUPTURE LIFE (hr)

Fig. 4.14. Influence of Welding on the Creep-Rupture Properties
of Hastelloy N (Heat 2477) at 650°C.
99

 

 

 

 

 

 

| 0
| o "
o
S3IHONI o ° XOOl :
1 o o L o
»* A % Kossorcsenaroe: e « o - e e
w i X\x&\\a .w,« ) %\ !.Sez( ékfi\%«w’ W - ARSI ; N g “ . »
0 * ,\&w{ o ) » rs&, . T - J.\.t. »< m. y
o e s l&,\\} > - . | i,.{z.,\‘\,,,v.. . . . §
~ : e s * \p\
1 - \w\ . & j: ~~ o~ 0 B . <
Y M . ! ,\« l» fi . 'X\ o..'«‘ Wrflw@
e w ° ..H» * * - ’ .
/?((J..Mz . S * 9u o‘ s o g& \.Q % W &, Z .
n}, f» &V A&o’-l & P ) .&M Ya s a‘ Ginflufl %;LA& t%.w« * °° e » ¢ ¢A, flw«
2 AT , M
\M -/
\« \\ x - »
\\«\
\
\N -
\\‘»\. -
N\ * P
\\\ - ;-
/
o EEy e

QWWP&%¥PA%W?

 

OO»«/«}»‘

©

#*

 

Typical Microstructure of As-Received INOR-8, Showing

Fig. 4.15.
Large Intermetallic Prec

ingers.

tates in Stri

ipi
100

e

Fig. 4.16.
Boundary Lamellar Phase and Localized Melting.

Y-49700
\

é:""‘:fl. .

 

INCHES

0.02

 

0.03

 

 

{0.05

Structure of INOR-8 Heated to 2500°F, Showing Grain

 
101

 

 

 

 

 

Table 4.5. Microprobe Analysis of Hastelloy N
(Heat 5075)
Element Bulk opherical Precipitates Lamellar Precipitates
Composition Matrix Precipitate Matrix Precipitate
Ni Bal 71.0 30.0 71.8 59.5
Mo 15.95 11.5 49 . 4 11.5 20.4
Cr 6.87 6.8 4.3 6.8 6.8
F'e 3.84 4.0 0.7 3.7 245D
Sl 0.62 0.3 24 0.3 1.0
Mn 0.50 0.6 0.01-0.10 0.6 0.6
Mo _ Mo _
T 1.65 T 0.34
Mo\ _ Mo\ _ Mo\ _
(=) com (=) -om (=) -1

Fig. 4.17.
Normal Structure to Structure with Lamellar Phase.

   

 

T

0.007 INCHES
1> 500X

[on

[

 

 

Photomicrograph of INOR-8, Showing Transition from
102

l

Tro

 

0.005 INCHE S
750X

T

[

 

 

 

Fig. 4.18. Autoradiograph of INOR-8, Showing Carbon Depletion
in Lamellar Phase.

Postirradiation Stress-Rupture Properties of INOR-8

The bulk of the data available on the effects of irradiation on
INOR-8 are postirradiation tensile data.lo In order to better assess
the lifetime of reactor components that were designed on the basis of un-
irradiated creep data, a program was started to ascertain the in-reactor
creep behavior and the postirradiation creep behavior of INOR-8. The dose
dependence of the ductility and rupture life of two heats of INOR-8 (Ni-
5065 and 2477) are being determined at several stress levels. Heat Ni-5065
is a reactor vessel heat, and heat 2477 is a low-boron heat. These uni-
axial creep tests at 650°C show the influence of irradiation on creep
rate, time to rupture, and ductility.

The postirradiation stress-rupture properties of heat Ni-5065 and
heat 2477 at 650°C are given in Tables 4.6 and 4.7 for various thermal-
neutron doses. These data show that ductilities for heat Ni-5065 at the
higher dose levels are less than 1% and for the low-boron heat are ~5%.
With both heats the time to rupture for a given stress level decreased
with increasing dose. An attempt is being made to correlate these data
To predict the rupture time for stresses below 10,000 psi and to determine
103

a correlation between in-reactor and postirradiation tests. With these

correlations one would be able to predict allowable stress limits and
rupture lives for reactor structural application.

 

 

 

Table 4.6. Stress-Rupture Times and Ductilities® for INOR-8
(Teat Ni-5065) Tested at 650°C
ot Times (hr) and Ductilities (%) for Specified
(pgifis Thermal-Neutron Dose (nvt)
0 5% 106 1.3 x10% 9 x 108 5 x 102 5 x 102°
X 107
52,57 52 21.2 11.9 5.1 0.43 0.9
(10) (2.6) (2.0) (3.4) (0.85) (0.64)
39,80 ~270 424 .7 147.0 80.2 23.5 5.3
(9.2) (2.8) (2.4) (0.2) (0.3)
32.35 ~800 686.7 647 .0 232 5.2 8.0
(6.7) (6.3) (2.5) (1.8) (0.42)
21.5% ~2500 Test in 1603 521.8 312
progress (2.7) (1.5) (0.79)

 

a C e .
Ductilities are in parentheses.

 

 

 

Table 4.7. Stress-Rupture Times and Ductilities”™ of Irradiated
INOR-8 (Heat 2477) Tested at 650°C
Times (hr) and Ductilities (%) for Specified
otress Thermal-Neutron Dose (nvt)
(psi) 5 x 10%° 1.3 X 10%8 9 x 10%8 5 x 10%° 5 x 1020
X 103
52.6 70.6 57 .4 30.4 21.1 19.9
(2.8) (3.0) (4.6) (4.3) (4.7)
32 .4 1070 1525.7 1425 849.6 515
(37) (6.3) (5.6) (5.1) (3.4)

 

a'Ductill_:'Lt:'Les are in parentheses.
104

In-Reactor Creep Tests

 

ceveral tests have been run in which the specimens were irradiated
and stressed simultaneously. The testing technique was described previ-
ously.!! The data obtained on heat Ni-5065 at 650°C and a nominal thermal
flux of 6 X 10'° nv are compared with unirradiated data in Fig. 4.19. A
similar comparison is made in Fig. 4.20 for another experiment on heat Ni-
5085. In both experiments the specimen at the lower stress did not fail,
and its behavior suggests that some experimental problem may exist. The
reduction in rupture life is greater for heat 5085 than for heat 5065.
This behavior seems consistent with the fact that the boron content of
heat 5085 is about double that of heat 5065.

ORNL-DWG 65-12050
70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6O\\\\\\\\
N
~\\
\
50
\UNIRRADIATED

2 40 NG
O \
Q N
O N
= \\
n \\
wn
g 30 ® \
n ° \\\

20 ? o—» \\

N
N
® IN-REACTOR DATA FOR SPECIMENS EXPOSED TO 650°C
AND TO A THERMAL FLUX OF 6X41013 nv
{0
0
10 20 50 100 200 500 1000 2000 5000 10,000

RUPTURE LIFE (hr)

Fig. 4.19. Influence of Irradiation on Creep-Rupture Properties
of INOR-8 (Heat 5065).
105

ORNL-DWG 65-12051

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60

50 \\

N
N
\\
\\
\\

40 \\ n
—~ UNIRRADIATED
2 \\k
3
o ™
= 30 S
o \
& o
= N
wn \\

20 ° ™

o— N
10 | ® SPECIMENS EXPOSED TO THERMAL FLUX OF 6X10'3 nv
AND 650°C IN REACTOR
0
10 20 50 100 200 500 1000 2000 5000 10,000

RUPTURE LIFE (hr)

Fig. 4.20. Influence of Irradiation on Creep-Rupture Properties

of INOR-8 (Heat 5085).

References

 

G. M. Adamson, Jr., et al., Interim Report on Corrosion by Zirconium
Base Fluorides, ORNL-2338 (Jan. 3, 1961).

 

 

A. J. Romano, C. J. Klamart, and D. H. Gurmski, The Investigation of
Container Materials for Bi and Pb Alloys, Part 1, BNL-811 (T-313)
(July 1963).

 

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 83.

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 84—86.

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 84—86.

 

J. W. Helm, Radiation Effects in Graphite at High Temperature, Bat-
telle Northwest, BNSA-67.

 

A. J. Perks and J. H. W. Simmons, 'Dimensional Changes and Irradia-
tion Creep of Graphite at Very High Doses," paper No. 188, Seventh
Biannual Conference on Carbon, June 21-25, 1965, held at Case In-
stitute of Technology, Cleveland, Ohio.

C. R. Kennedy, "Irradiation Creep of Graphite," Metals and Ceramics
Ann. Progr. Rept. June 30, 1965, ORNL-3870.

 

J. T. Venard, Tensile and Creep Properties of INOR-8& for the MSRE,
ORNL-TM-1017 (February 1965).

 

10. W. R. Martin and J. R. Weir, Nucl. Appl. 1, 160 (April 1965).

11.

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 82-83.

 
106

5. RADIATION CHEMISTRY

Introduction

 

A number of irradiation experiments with molten-salt reactor fuel
and construction materials have been carried out in irradiation facili-
ties in the Materials Testlng Reactor (MTR). Results from these experi-
ments have been reported and indicate that operation of the MSRE should
not be adversely affected by irradiation effects on fuel salt and ma-
terials. However, a continuing irradiation program is needed to pro-
vide supporting information for an understanding of both short-term and
long-term effects of irradiation and fissioning on fuels and materials,
especially for reactors that will follow the MSRE.

Design and development of an in-pile molten-salt experiment have been
in progress during the past year. A description of the proposed in-pile
experlment and preliminary results of mockup tests have been reported else-
where?® but are briefly summarized here along with progress made through
this report period.

Experimental ObJjectives and Design Considerations

 

oome objectives of the proposed irradiation (in—pile) program with
an MSR fuel salt are tabulated below:

Irradiation objectives: ZOO'w/me fuel fission power
Maximum fission product production
One year of in-pile operation

Information objectives: Graphite—INOR-8-salt compatibility
Salt (fuel) stability
Fission product chemistry

To meet these objectives, the following experimental requirements
were established:
oalt circulation through hot and cold regions
oalt sampling during irradiation
Cover gas sampling
Ability to drain fuel salt
valt addition to replace that removed in sampling
Keeping fuel molten at all times
The above objectives and requirements indicate the need for daily
attention to the experiment operation and data, rapid analysis of gas
and fuel samples, and possible concessions in reactor operating sched-
ules. These requirements and the availability of hot-cell and analytical

facilities especially equipped to handle radicactive fuel dictate that
such experiments be carried out in the Oak Ridge Research Reactor (ORR).
107

Therefore, the experimental equipment is designed to make use of horizon-

tal beam hole HN-1 in the ORR. Previous neutron flux mapping of beam hole
HN-1 indicates that a thermal flux of ~5 X 10%*3 would be avallable. It

is estimated that a thermal flux of 5 X 10%3 will produce a fission power

density of l9O'w/cm3 in a fuel salt containing 0.6 mole % 235U and a den-

sity of 320 w/em® in a 1 mole % 23°U salt.

Experiment Design and Mockup Test Operation

In order to avoid costly, complex pump loop experiments, an auto-
clave (capsule type) experiment with thermally induced flow was designed.
Initial tests with prototype autoclaves showed that flow rates of 2—10
cmB/min could be achieved in an autoclave of this type. These flow rates
are considered adequate to study transport and deposition of fission and
corrosion products and to provide mixing for fuel sampling and enrichment.
Figure 5.1 is a schematic diagram of the autoclave assembly, and a photo-
graph of the partially assembled experiment is shown in Fig. 5.2.

ORNL-DWG 65-12052
SALT RESERVOIR
PRESSURE MONITOR LINE

SALT FLOW PASSAGES
GRAPHITE CORE

THERMOCOUPLE WELL

S SAMPLE LINE

  
 
  
  

 

 
 

SALT LEVEL

 
    

   

THERMOCOUPLE <f;§fl:!%€
WELL LA - —=

  

ALROD HEATER

   

   
    
 
   
  

i COOLING

COIL

CALROD HEATER

OOLING COIL
GRAPHITE
JACKET
COLD LEG

AUTOCLAVE

  

\
N

.
\~
.

 

  

SALT SAMPLE LINE

CALROD HEATERS
'6. 7

SPRAYED NICKEL

Fig. 5.1. ©Schematic Diagram of In-Pile Molten-Salt Autoclave.
10

PHOTO 71141A

     

SALT FLOW PASSAGES

SALT OUTLET

COOLING COIL

= GRAPHITE JACKET

CALROD HEATER

Fig. 5.2. Partially Assembled In-Pile Molten-Salt Autoclave.
109

The main autoclave body is constructed of 2-in. sched-40 INOR-8 pipe
with a graphite core. Longitudinal holes through the graphite core serve
as fuel passages. The fuel volume in the core, where the maximum thermal
neutron flux will exist, is ~%42 cm®. A reservoir tank serves as a salt
reservoir and provides for a liquid-vapor interface. A '"cold leg" return
line completes the circwit for salt circulation. The total salt volume
in the experiment is ~85 cm>.

harlier models of the autoclave experiment had a horizontal line
connecting the "cold leg" with the main autoclave. However, the design
incorporating this horizontal return line was changed to the triangular
configuration shown in Fig. 5.1, after test operation indicated flow block-
age by deposition of bubbles of cover gas in the return line. Gas depo-
sition and flow blockage were observed after ~16 hr of operation with
helium cover gas and ~135 hr with argon cover gas. Flow invariably re-
covered after evacuation and repressurization of the capsule, but con-
tinuous salt flow in the experiment 1s desirable, so the horizontal return
line was eliminated to facilitate gravity release of the bubbles. The
bubble formation appears to be consistent with a mechanism based on mass
transport due to the temperature dependence of gas solubllity, but ap-
parently is under diffusion control in the return line. Successful elimi-
nation of the loss of flow by removal of the horizontal return line was
evidenced by the fact that the redesigned test model (Fig. 5.1) has op-
erated for 550 hr under helium cover gas with a continuous flow rate of
913 cm®/min.

Calrod heaters and cooling coils, in which air or air-water mixtures
are used as coolant, provide temperature control and maintain the thermal
gradients necessary to induce flow. Heaters and cooling coils are em-
bedded in a graphite Jjacket around the autoclave body. Tubes of capillary
dimensions connect the vapor space of the reservoir tank with remotely
located pressure monitoring and gas sampling equipment. The salt sample
line (1/8-in. OD X 0.075-in. ID) is ~12 ft long and is traced with Calrod
heaters. The sample line is connected to a sample station at the reactor
shield face, where samples of molten salt can be routed either to a dump
tank for storage or a sample capsule for subsequent chemical analysis.
The sample station also contains necessary tanks, valves, and lines so
that salt can be added to the autoclave experiment during in-pile opera-
tion.

To date some 4000 hr of operation has been accumulated with several
prototype models of the in-pile molten-salt experiment. Test operation
has been conducted in a mockup facility which duplicates beam hole HN-1
of the ORR. Test operation is at a nominal salt temperature of ~1200°F
with a salt mixture whose composition is LiF-BeF,-ZrF,-UF, (65-29-5-1
mole %, liquidus temperature ~840°F). Salt circulation rates of ~L10
cm /min are obtained by maintaining a median temperature gradient of
50-100°C between the main autoclave and the '"cold leg." The flow rate
is monitored by heat balance measurements around the "cold leg."

 

The adequacy of the salt sampling and addition system was demonstrated
by removing 15 samples (10 cm® each) of the salt while operating at high
110

temperature. After each sample removal, an equivalent amount of salt was
added in order to maintain the original salt inventory.

A test was made of the ability of the autoclave heater-cooler unit
to remove the predicted 12 kw of fission and gamma heat (8 kw of fission
heat) during in-pile operation. Some 9 kw of heat, the maximum that
could be generated in the test equipment, was removed with the autoclave
operating at 1000°F; the heat removal capacity is accordingly considered
adequate for in-pile operation, taking into account the higher operating
temperature of 1200°F. Also, the cooler is capable of a larger through-
put of air-water coolant than was used in the above tests.

Results of mockup tests indicate that the presently designed auto-
clave is suitable for use in a fuel salt irradiation program with the
obJjectives previously outlined. Additional design and development work
to provide necessary facilities in beam hole HN-1 of the ORR is under
way. Included in these facilities are a beam hole liner, a beam hole
shield plug, salt and gas sampling equipment, revisions to an existing
instrument panel, and shielded carriers needed to remove the experiment
after in-plle operation.

References

 

1. Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp.
36—45 (April 1965).

 

2. Ibid., pp. 45-48.
111

6., CHEMISTRY

Chemistry of the MSRE

 

MSRE Fuel Loading

 

The initial fuel loading of the MSRE required approximately 75 ft>
of fused fluorides having the composition 7LiF-BeF2-ZrF4-UF4 (65.0-29.1-
5.0-0,9 mole %); fissionable <3°U comprised about one-third of the uranium
inventory. To provide for an orderly approach to critical operation of
the reactor and to facilitate fuel preparation, the fuel was produced as
Three component mixtures. The enriched fuel concentrate mixture, in which
all ?3°U was combined with ‘LiF as UF, 93% enriched in 23°U to form the
binary eutectic mixture (27 mole % UF,), was prepared in six small batches
(15 kg 2357 each) for nuclear safety and for planned incremental additions
to the reactor fuel system. The balance of the uranium required for the
fuel was provided as a chemically identical mixture with UF, depleted of
235U, The third component mixture, the barren fuel solvent, consisted
of the remaining constituents of the reactor fuel and had the chemical
composition of ‘LiF-BeFy-ZrF, (64.7-30.1-5.2 mole %). The preparation
of these mixtures from simple fluoride salts has been previously re-
ported.*t

Reactor fueling operations began on April 20, 1965, with the loading
of the barren fuel solvent and the depleted fuel concentrate mixture.
Approximately 10,050 1b of solvent from 35 batch containers and 520 1b
of depleted fuel concentrate from two batch containers were added directly
into the fuel drain tank. Batch containers were heated above the liquidus
temperature of the salt mixture in auxiliary furnaces and connected by a
small-diameter Inconel tube to the drain tank. The connecting tube ex-
tended to the bottom of the batch container so that the salt mixture could
be transferred as a liquild by controlling the differential gas pressure
in the two containers. All fill operations were accomplished in a routine
manner without causing detectable beryllium contamination to the reactor
facilities.,

The addition of the enriched fuel concentrate mixture to the MSRE
to within 1 kg #2°U of criticality was accomplished during the latter part
of May 1965. This operation was coordinated in accordance with planned
zero-power experiments of the reactor system.2 The first major addition
of enriched fuel concentrate consisted of the transfer of about 44.17 kg
232U from three containers directly into the fuel drain tank. Three sub-
sequent additions of 235U to the reactor drain tank increased its 23°U
inventory to 59.35, 64.42, and finally 68.76 kg. The transfer of less
than batch-size quantities of 437U was made by inserting the salt transfer
line to a predetermined depth in the batch container.

All fluoride mixtures — coolant, flush, and fuel — required for the
initial operation of the MSRE were prepared and loaded into the reactor
facility by the Reactor Chemistry Division. Enough fuel-enriching cap-
112

sules, each of which contained about 85 g 227U each, to bring the MSRE to
its critical operation and to maintain nuclear operation of the reactor
during its scheduled tests were also provided.

MSRE Salt Chemistry During Precritical and Zero-Power Experiments

 

Chemical analyses were conducted during the MSRE precritical and
zero-power experiments for the purpose of establishing analytical base
lines for use in the full-power operating period. Results of these tests
provided not only the desired base lines, they also indicated that, with
respect to preventing contamination of the fuel salt, current reactor op-
erating procedures have afforded excellent protection of the fuel salt.
Fuel samples were removed daily from the MSRE pump bowl throughout the
zero-power experiments. Chemical composition, contaminant levels, and
isotoplic analyses were determined on a regular basis. The salient con-
clusion derived from the results of these analyses i1s that no corrosion
was sustained by the fuel containment circuitry during the approximately
1100-hr precritical and zero-power test period. At the end of the zero-
power experiments, the fuel salt was drained from the fuel circuit and
replaced briefly by the flush salt. The amount of uranium subsequently
detected in the flush salt (0.0195 wt %) is sufficiently small to indicate
that no appreciable holdup of fuel salt occurred in the drain operation.

The MSRE fuel was constituted within the reactor in two steps. The
first was addition of "LiF-?28UF, to the "LiF-BeF,-7ZrF, (64.75-30.09-5.16
mole %) carrier salt. The mixture produced in this manner was circulated
through the fuel circuit for some 250 hr during the PC-2 precritical test.
The reactor fuel for the zero-power experiments was produced subsequently
by adding small increments of LiF-235UF4 into this carrier salt in the
drain tanks and finally into the pump bowl. The final composition of the
salt was controlled by the amount of 235UF4 required for criticality to
be sustained with one control rod completely inserted. It was anticipated
that the composition of the fuel at this point would be 7LiF—BeF2—ZrF4-UF4
(65.0-29.17-5.0-0.83 mole %); the composition calculated from the weights
of the carrier and enriching salts added to the reactor was 7LiF-BeF2-
ZrF4-238UF4—235UF4 (64.85-29.28-5.04-0.55-0,28 mole %). The composition
of the fuel salt was changed steadily throughout the zero-power experi-
ments as capsules of the enriching salt were added to the pump bowl. Com-
positional analysis during this period served, therefore, to permit eval-
uation of the fuel composition dynamically rather than as a statistical
base for reference during the power run. Final composition established
analytically on the basis of the last four samples was 7LiF-BeF2-ZrF4—
238yF,-23°UF, (62.31-31.68-5.19-0.54-0.28 mole %).

Complete results of all analyses performed with MSRE salts during
the precritical and zero-power experiments were reported earlier.? The
data are summarized in Table 6.1.

Uranium Assay. During the initial fill operations, ‘LiF-BeF, flush
salt was admitted to all parts of the fuel system. On draining the re-
actor, flush salt remained in the freeze valves as well as residue in the
drain tanks. At this point, some 80 1b of salt was unaccounted for by

 
*quawtaadxe T-0d JO PpUS 4V

*quawTIadXs T=Dd JO SUTUUTISQ 4V

" (%'3eP qBT-30U) quel "H D

*(®B3EP QBT-PTOO) UUSNeBA *Jd "M

* SUSWTOads JNOJ UO Pasey BABP fAQUSUTISIXD JoMOd-0J9Z JO PUS 1V

113

 

 

 

 

 

 

0L6 ¥ GOTT GLOT oCE7 O
6L ¥ 8% e ¢c LT ol i TN
8 F LE LE 61l cc oG 13 <D
GG F YGT coT 86 G6 oG 56 o4
oAE@@V UAEQQV Qd&@@v_ AEQQV d&@@v AEQQV
L8 89 ¢ L9 06" /L9 20°89 76° 69 6869 80" L 169/ A
909°% L9 70° ¢ 80°'¢ O O 0 O nx
09¢° 1 796G T O O O O O 0 Negz
o¥0°¢ 901" ¢ 70° ¢ g0'¢ O 0 O O Ngegz
TT°TT cO°"TT 9¢"TT Te* 1T G611 c6°TT O O L7z
T4 9 ce'9 649 cG'o Ge* /L 789 89°6 9c 6 od
Gc 0T L8 0T 70T cO'TEL 98°01 Ge 1T cr'¢l £8°¢ct I,
(% 3) (% %) (% ) (% 3) (% ) (% ) (% ) (% )
»STSsATeuy UITS9( sTsATeUY UITSoQ sTsATeUVY USTS9( STSATRUY uSTsoq e
IS Tondg qummmnmfigu + 318G JI9TIIE) ITBS JISTJLIB) ITeS USNTH

 

sosATeUY 4TBS MMSW JO AIsummg *T*'9 STABRT
114

the weigh-cell data and is presumed to have remained at these locations.
As fuel was constituted at the beginning of the PC-2 test, 7LiF—238UF4
and carrier salt were blended and circulated through the same system cir-
cuitry. On terminating the zero-power experiment, the fuel was drained
into the storage tank via a single freeze valve. At this stage, there-
fore, the fuel salt was diluted with a significant amount of flush salt.
Its effect is evident in the disparity between nominal and analytical
values for uranium concentrations in PC-2 and in the differences between
nominal and measured 238U enrichment values during the zero-power experi-
ments. It may be demonstrated that if the MSRE fuel salt were diluted
with no more than 140 1b of 7LiF-BeF, flush salt, the bias between nominal
and analytical values for the uranium concentration in both the precrit-
ical test, PC-2, and the zero-power experiment would be reduced to zero.
We infer, therefore, that approximately 140 1b of ‘LiF-BeF, flush salt
diluent remained in the fuel circuit at the end of the PC-1l test and is
responsible for the apparent blas in nominal and analytical values for
uranium concentration and *2°U enrichment fractions.

An interesting implication of fuel dilution by 140 1b of flush salt
is that a revised figure for nominal uranium concentration at criticality
becomes available. This figure now becomes 4.453 wt % uranium (of which
1.39 wt % is <2°U), lower by 1.26% than the previously accepted value of
4e51 wt % uranium.

During the zero-power run, approximately 70 1b of salt gradually ac-
cumulated in the fuel-pump overflow tank. This salt, of unknown composi-
tion, was not removed until after the last fuel-salt sample was obtained.
Removal of salt from the circulating system would be expected to have had
only negligible effect on values for the uranium concentration but may
have contributed slightly to error in estimates of 235U enrichment frac-
tions.

otructural-Metal Impurities. Normalized values for Fe, Cr, and Ni
(see Table 6.1) indicate that no appreciable change in the concentrations
of these elements occurred in the approximately 1100-hr test period. This
result is in interesting contrast with the changes in the PC-1 test, in
which the concentration of chromium in the fuel and coolant circuilits in-
creased by 25 and 10 ppm, while iron in the same circuits was decreased
by 52 and 50 ppm respectively. The concentration of nickel remained at
approximately 5 ppm throughout the test. The free energies of formation
of NiF,, FeF,, and CrF, at 1000°K are —58, —66.5, and —74 kcal (°F)"1
atom™1; that is, the fluorides of NiFs and Fel, are unstable with respect
to Cr® in the Hastelloy N vessel walls. The results of the zero-power
tests as well as those from the first prenuclear test* indicate that
nickel is present in the MSRE salt only as a metallic phase and that iron
is also present predominantly, if not entirely, in metallic form. We con-
cluded that iron and chromium concentrations varied inversely with time
in the PC-1 test but not in the zero-power experiment primarily because
the "LiF-UF, enriching salt contained some 2% of the uranium in the tri-
valent state; that is, U3* in the enriching fluid served to allow the re-
action

 
115

FeF, + 2UF; = Fe0 + 2UF,
to proceed rather than the reaction
cr® + Fel's, — CrFy + Fel

Unquestionably, iron and nickel were introduced into the reactor fuel
circuit from a quiescent salt reservoir in the fuel storage tanks and
continued to be circulated throughout the zero-power-level experiments.

We must conclude that these suspended metallic impurities were introduced
into the fuel circuit from salt which was well mixed by thermal-convection
stirring in the drain tanks. That thermal mixing occurs in these tanks
was substantiated by the first composition analysis in PC-2. As evidenced
by the composition of a sample which was obtained from No. 2 drain tank
shortly after ‘LiF-?38UF, was added to the carrier salt, virtually com-
plete mixing of 7LiF-UF4 and carrier salt was achieved before intertank
transfers were made to ensure homogenization. Comparably efficient mixing
1s also probably adequate to prevent settling of the very fine (below the
limit of microscopic detection) metallic particles of iron and nickel in
the drain tanks.

Coolant salt was circulated through the reactor for 117 hr near the
end of the zero-power experiments. During this period, three samples were
obtained from the pump bowl and analyzed for structural metal and oxide.
The average concentration of iron, chromium, nickel, and oxygen was found
to be 108, 38, 23, and 548 ppm, respectively, impurity concentrations
which indicate no appreciable change in the coolant salt since its use in
the PC-1 experiment.

Oxide Contamination. Analytical wvalues for oxide concentration of
MSRE specimens obtained during the PC-2 and zero-power experiments were,
for the most part, incredibly high. If real, the higher wvalues would
represent the presence, in fact, of more than 0.5% ZrO, in the specimen.
Each of the 52 specimens removed from the reactor was subjected to careful
microscopic examination. Neither crystalline ZrO, nor oxyfluorides were
detected in any specimen. Sensitivity of detecting well-formed crystals
of ZrOp, in such salts by the petrographic method is considered to be well
below 100 ppm. Although sampling was performed under extremely dry at-
mospheres, subsequent handling operations occasionally exposed the salts
to moisture-laden atmospheres. Beryllium fluoride is formed on freezing
all MSRE salts. It i1s a hygroscopic material which, once moistened, can-
not be redried effectively by room-temperature desiccants. The anoma-
lously high values for oxide are therefore, to a large extent, a measure
of the exposure of the salt to moisture-ladened atmospheres. Despite the
unrewarding results, we consider that performing oxide analyses was a
worthwhile effort for several reasons. The sporadic values of oxide con-
centration are anomalous in contrast with the microscopic data and with
the unchanging value of chromium in the salt; their magnitude in both PC-1
and PC-2 tests also appears to be unrealistically high in contrast with
the amount of oxide (115 ppm) recovered by R. B. Lindauer in reprocessing
the flush salt.” These facts indicate that the results of such analyses
cannot in themselves serve as an alarm signal in reactor operation, but
rather that techniques for application of the method on-site should be
considered.

 
116

Mean values for the concentration of HF in the helium cover gas were
obtained during the PC-2 and zero-power experiments by adaptation of a
continuous internal electrolysis analyzer for gaseous fluorides to this
monitoring method. Through the cooperation of ORGDP personnel, data were
obtained regularly indicating that HF was evolved from the salt at a max-
imum mean value of 150 to 200 ug/liter.6 Some question exists with re-
gard to the possible contribution of particulate fluorides to this value.
Nevertheless, as a maximum possible concentration, the HF value corre-
sponds to the introduction of no more than 1 ppm of oxide into the salt
per day through the reaction

2H-O + Zr¥, — ZrOy + 4HF ,
an amount which would escape detection by other methods.

Spectrochemical Data. MSRE carrier and fuel salts were examined for
trace quantities of heavy-metal cations which would go undetected in gen-
eral chemical analyses. The following amounts of impurities were typical
of the results found (in ppm): Al, 1 to 10; Ca, 1 to 10; Cu, 0.1 to 1.0;
Mg, 0,01 to O.1; Mn, 0,001 to 0.01.

 

Appraisal of Results. From a chemical standpoint, the operation of
the MSRE during the precritical and zero-power experiments was an unqual-
ified success. The salts were maintained in an excellent state of purity
during all transfer, fill, and circulation operations. The results.of
chemical analyses reinforce the long-standing conclusion that pure molten
fluorides are completely compatible with nickel-based alloys.

 

ceveral aspects of the MSRE are currently unsatisfying and must re-
ceive Tfurther attention.

1. Unless dilution of the fuel by flush salt is postulated, uranium
analyses for specimens obtained in both the precritical and zero-power
experiments exhibit a bias of approximately 1% below book values. It may
be anticipated that recurrence of this disparity will take place during
the full-power run after any shutdown which requires circulation of the
flush salt through the fuel circuit. Weigh-cell data, while very useful,
are probably not sufficlently accurate to permit bias in uranium analyses
to be established to within *0.5%. Application of additional methods for
corroborating uranium assay would therefore be very useful but are not now
planned.

2. No rapid, absolute method exists for establishing the level of
contamination of the salts by small amounts of oxide ion. This assay was
performed in a partly satisfactory manner during the precritical and zero-
power experiments, but the method has not yet been adapted for application
in the power run.
117

3. Analytical values for lithium have continued to appear anoma-
lously low and beryllium values anomalously high during the most recent
experiments, much as they were in the first precritical experiments. Mod-
ifications of analytical procedures for measuring the concentrations of
these components will be necessary for the requisite confidence in compo-
sitional analyses to be established in the future.

Examination of Salts After the Zero-Power Experiment

 

At termination of the zero-power experiment, the MSRE fuel was
drained to storage tank FD-2. Flush salt was then circulated through the
fuel circuit for 24 hr. Subsequently, flush and coolant salts were
drained to storage tanks. Four specimens of fuel salt were removed from
the fuel storage tank during the first 300 hr after completion of the
zero-power experiment. The purpose of sampling in this period was to

1. provide assurance that the fuel solution remains at constant composi-
tion during storage,

2. Obtain information on settling of finely divided metal impurities
during storage, and

3. establish an analytical base line for fuel-salt composition for use
in the full-power operating period.

opecimens of fuel salt were submitted for chemical analysis. Pre-
liminary results have been obtained which indicate that the uranium con-
centration in the fuel storage tank is 4.612 wt %. Slight dilution of
uranium actually occurred as the salt in the fuel circuit was blended with
Other less-concentrated salt in the fuel system. The change in concentra-
tion was well within the standard deviation for uranium in the zero-power
experiment analyses, 0.707%. Analytical evidence of the dilution was
therefore not forthcoming.

The thermodynamics of structural-metal corrosion chemistry in molten
fluorides is well established.’ Selective attack upon chromium in INOR-8
must occur in MSRE salts if significant concentrations of FeF, or NiFs
are present in the molten solutions. In a careful study of the reaction

MF, (d) + Ho(g) = M(e) + 2HF(g) ,

where M represents either Cr or Fe, and c, g, and d indicate that the
species are crystalline solid (at unit activity), gaseous, or dissolved
in a molten LiF-Bels; mixture, Blood® showed that the equilibrium constants
for KN.in the expression |

 

2
KN _ EHF
NMFZ X 3H2

are 8 X 1071 and 1.3 x 107% for Fel's and CrFs respectively. Attack of
chromium in INOR-8 by a displacement reaction with FeF, should proceed

until the ratio NCng/NfiéFb ~ 500. In the corresponding case for Ni,
118

available Tfree-energy data and direct experimentation confirm that the
ratio N /N : should be several orders of magnitude greater. The
Cris’ "Nils

constancy of analytical values for the concentrations of Fe and Ni in MSRE
salt during the precritical and zero-power experiments® has therefore been
of some annoyance, for at their low concentrations, ~150 and 50 ppm, re-
spectively, it has not been possible to determine their chemical identity.
It must be concluded that all of the nickel and practically all of the
iron in the salt during the zero-power run were present only in the me-
tallic form. Since all previous transits of the salt were conducive to
retaining suspended metal in the salt, we felt that possibly the finely
divided metal would gradually settle in drain tanks on conclusion of the
zero-power experiment. In anticipation of this possibility, the four
samples obtained from the stored salt were analyzed for structural-metal
impurities. The results of these analyses indicate only that no change
had occurred, which we attribute to thermal-convection stirring. As full
power 1s approached on the MSRE, a AT will be imposed which may, by a
combination of activity change of the component metals in the container
walls and variation with AT of solubility of the iron and nickel metals
in the salt, allow the suspended metal particles to grow in size and de-
posit in the cooler areas of the system. The total quantity of iron and
nickel which could deposit is in the order of 100 g.

PHOTO 81120

©

 

Fig. 6.1. Flush Salt in Rotor Flange Ares.
119

PHOTO 80844

 

Fig. 6.2. Flush Salt Deposit Removed from MSRE Pump Bowl Volute.

On removal of the fuel pump, some 3 1b of flush salt was found +o
have been retained in the rotor flange area (Fig. 6.1) and in the volute
of the pump bowl (Fig. 6.2). Additional small fragments of salt were
found in various adjacent locations, in the grooves of the shield-block
O-ring, and in line No. 903, through which pump-bowl gas was vented o
the HF analyzer. All salt specimens were submitted for microscopic and/or
chemical analysis and were found to be entirely free of oxides or oxyflu-
orides. A minor oil leak allowed oil to enter the shield-block section
of the pump. Thermal decomposition products of the pump oil were in con-
tact with the salt specimens in these areas and were found as partial
films on some of the salt (Fig. 6.2) or dispersed through fragments of
Other specimens. 1In all locations where crystallized salt deposits were
found, it was evident that the molten salt had not adhered to the metal
surfaces. This is particularly evident in the volute deposit (Fig. 6.2),
where the molten specimen cooled rapidly and maintained the surface angle
tangency characteristic of sessile drops which are free of contaminant
oxides.,

Density of MSRE Fuel and Coolant Salts

Relatively few measurements of the densities of liquid-salt mixtures
have been made in development of ORNL molten-salt technology. Some ex-
perimental values, regarded as accurate to within *5%, were made as part
of the ANP program.”’1® Similar measurements, which also employed a sim-
ilar buoyancy principle, were made shortly afterwards on LiF-Bel's -UF,
mixtures.t!? Later measurements were made at ORNL with a proposed fuel
120

containing ThF4.12 Most density wvalues for ORNL project use have been

estimated at first by the method of mixturesl® (regarded as accurate to
within #5%) and, more recently, by an improved method which assumes ad-
ditivity of molar volumesl?’1% and which is considered to be accurate
within *3%., This latter method of estimation requires accurate values

for the molar volume of the pure constituents. For its use, some measure-
ments were made, agaln using the buoyancy principle. These estimates of
densities were sufficiently accurate for the preliminary work on the MSRE
although not accurate enough for current MSRE operations.

In an effort to obtain accurate values for the MSRE fuel and coolant
densities, measurements were obtained of the depth of a melt in a cylin-
drical container of accurately known dimensions which is contained in a
controlled-atmosphere glove box. The depth is measured by a probe at-
tached to a wvernier caliper. Contact of the probe with the melt is indi-
cated by completion of an electrical circuit through the caliper and the
melt. Appropriate corrections for thermal expansion at a particular tem-
perature are applied for the container and the probe. This procedure has
some shortcomings, principally with respect to the reproducibility of de-
tecting the liquid level because of the possibility that small amounts
of liquid may adhere to the probe. The procedure does, however, provide
a direct measurement on liquid mixtures which can be visually observed
during measurement. Density values obtained by this experimental proce-
dure are shown in Tables 6.2 and 6.3; the equation for density, 4, as a
function of temperature is d = a — bt. Variations of salt densities with
temperature are shown in Fig. 6.3. The values are considered to be in
excellent agreement with the values derived from recent measurements of
densities of the salts stored in drain tanks at the MSRE site.?

ORNL—DWG 65-12053

 

1500 \ \\
1400 \ \\

\ \MSRE FUEL

1300
\

\ \

\MSRE COOLANT \

1100 \

 

 

 

 

 

 

 

 

 

 

TEMPERATURE (°F)

1000

900

800

\ \

 

\

\

 

 

 

 

 

\

 

\

 

 

N

 

 

 

 

 

110

115

120

125

Fig. 6.3.

130

135

140

DENSITY (1b/ft3)

145

Density of Melts.

150

155

160
121

"040000 °*0F
"870°0F

*qa10dax sTUY UT UOSTJIeAWOO JOJ OTIASW O POIIDATOD ANg SATUN YSTTIUN UT

iUOTFBTASD DIBPUBLE

tUOTARTASD DPIBPUBILS

q

poaagodaa %HflmQflMfihom

 

 

 

 

CO°9%T  L920°0 L9°8LT g7e*C  069L000°0  g878°C SIoM STUT
m * mfi._” m._um.m . < unpuBIOUW=SUW TBULS29UT JoTadUuagneH
TUSWSINS B
G9°Z%T  9020°0 LE£°L9T YR¢C 2 765000°0 0L9°2  Tv—8¢ *dd ‘Zogc-TN¥O  SsumloA JeTow JO wng
VehCT  €%20°0 9°¢oT GT*2 1666 °0 T9°¢C 2OLT=TNMO SJINIXTU JO POYLSH
91BWTASH
A.,002T q B 0,069 q B
20UDIST Y POYUISN
(¢33/dT) s3TUn USTTIUH (¢wo/8) S3TUN OTJIFSN

 

7qN €8°0 ‘7AIZ 00°G ‘%I LT*62 ‘ATI 0°GO

Tend NS JOo AaTsus *Z°9 STABIL

:(% eTow) uotTaTsoduo)
122

'9020000°0F  fUOTHETASD PILPUBIY
'9TZ0°0F :UOTIRTASD DIEDUBRLS_

"GO6T °dai ‘uoTgedTunumiod o3eATad ‘TsBul Yy °p

 

 

 

 

q
*qJ0dal STUYY UT UOSTIBAdWOOD JOJF OTIASW 09 PIAISAUOD ANg SATUN YSTTIUW UT pajtodad hflfl@qfiwfihom
0°*%CT G9910°'0 cO* 77T €86°T pc87000°0 5962 °¢C fL0M STUL
2086° 1
ceT HC6°T qe ogoxd aanssoaag
6°6TT €82T0°0 AL2°GeT T26°T £,£000°0 8GT 2 980 T~IWIN A10%°2I00ET PUNON
1LUSWDINS B
8°8TT 98¢ TO0 O7°G6¢eT 06°1T 07000°0 09tT*¢ @Efi@thoawfi TeuI2qUT LONEPUT ]
9°8TT  9G€TO'0  06°7€T  868°T  T6£000°0 2qT e Tvr—ee *dd ‘Zozc-TNMO — SumTOoA JBTOW JO UMG
c1T 80c0°0 6°6¢cl G8 T 9000°0 e ¢ cOLT-TNHO SoJANIXTUW JO POYISN
S1BWTASH
4,001 q B D ,069 q B
S0UBIDI Y POV
(¢3F/AT) s3Tun UsSTTIUY (¢wo/3) s3TUn OTIFSN

 

ciesg ¢ ‘4TI 99 (% oTow) uoraTsodwo)
JUBTOO) HMSK JO A3TsuSsQ °*¢°*9 9TIB[
123

Chemigtry of LiF-BeF, Systems

 

Solubility of HF and DF in Molten LiF-BeF, (66-34 Mole %)15

The solubilities of HF and DF in the molten mixture ILiF-BeF, (66-34
mole %) were determined over the temperature range 500 to 700°C and at
gas saturation pressures between 1 and 2 atm. An extrapolation of the
solubility values provides an approximation of the solubility of tritium
fluoride in the molten fluoride mixture. The behavior of tritium fluo-
ride, formed by neutron irradiation of lithium, in the fluoride mixture
would be of interest in molten-salt reactor technology as well as in the
proposed use of a molten-fluoride breeder blanket for a thermonuclear
machine, 10

The experimental procedure has been previously described and employed
for systematic studies of gas solubilities in molten fluoride mixtures.l?
Anhydrous hydrogen fluoride was obtained from a commercial source; anhy -
drous deuterium fluoride was prepared by the Technical Division, ORGDP,
by reaction of elemental deuterium and fluorine.l® Values obtained for
Henry's law constants expressed as moles of HF per mole of melt per at-
mosphere at 500, 600, and 700°C were 3.46 X 107%, 2,15 x 10~%, and 1.39 X
104 respectively. Corresponding values for the solubility of DF were
3.03 x 1074, 1.85 X 1074, and 1.30 X 10™% at 500, 600, and 700°C respec-
tively.

Differences in the solubilities of DF and HF are outside the 95%
confidence level attributed to the experimental data. Evaluation of the
temperature dependence of the Henry's law constants by a least-squares
fit of the data indicated that, within experimental precision, the en-
Thalpies of solution of the two gases are identical and equal to about
—6.5 kcal/mole.

Vapor Pressure

 

LiF-BeF, System. Rodebush-Dixonl? and boiling-point methods were
used to measure total pressure over 16 melts covering the entire composi-
tion range. Each melt was measured over at least a 185° temperature
range; the pressure range usually covered 1 to 100 mm Hg. For all cases
except one, the linear expression

 

log P{mm) = A — B/T(°K)
provided adequate fit to the data (A and B are constants).

The data are summarized in Table 6.4. Isotherms at 1000 and 1100°C
are shown in Fig. 6.4.

With these vapor-pressure data alone, one cannot calculate thermo-
dynamic activities, because the composition of the vapor is not known.
However, the shape of the isotherms shown in Fig. 6.4 suggests that the
predominant vapor species change at a melt composition of about 75 mole %
LiF,
124

Table 6.4. Vapor Pressure in the LiF-BeF, System

 

 

 

 

C o Kquation,

MEl% Cimpo§1tlon Temperature Range log P(mm) =

= mole % = Measured (°C) A — B/T(°K)

2 A B
100 779=114"7 10.487 10,967
84 .99 15.01 826—~1116.5 10.714 11,312
75 .01 24 .99 890~1112.5 10.216 10,707
70.00 30.00 843-1150 10.360 10,950
65.00 35.00 857—1122 9.728 10,194
57 .54 42 46 866—1121 9.790 10,417
50.00 50.00 886—1071 9.130 9,785
45 .04 54,96 932—1155 8.961 9,770
39.95 60.05 930—-1224 8.979 10,040
36.00 64 .00 9501214 9.294 10,718
30.00 70.00 966-1281 8.752 10,277
25.00 75 .00 1020~1272 8.736 10,526
11.00 89.00 891—1239 7.154 8,770
'7.00 ©93.00 76-1234% 7.861 9,910
3.00 97 .00 10201270 10.187 13,360

100 1026—1272 a

 

aValues for this melt composition and temperature range fitted
the equation log P = 3.619 — 15,450/T- 6.039 log T.

To obtain some notion of the vapor species, vapor was collected in
the tubes of the vessel at the conclusion of vapor-pressure measurements
for several compositions. Chemical analysis showed only traces of lithium
ion in condensates where the composition of the melt was 75 mole % or more
BeF,. For melts with 70 mole % or less BeF,, considerable quantities of
lithium were found in the condensates. All that may be concluded from
these analyses is: (1) at greater than 75 mole % BeF,, the vapor is vir-
tually pure BeFs, (2) at less than 70 mole % BeF,, the vapor becomes more
complicated, probably containing the compounds LiBeFs3 and LigBeF4020

LiF=-Bels-Zrl, System. Two melts of importance to the MSR distilla-
tion process were measured by means of the Rodebush-Dixon method. The
data are summarized in Table 6.5 and plotted in Fig. 6.5.

 

The second composition listed in Table 6.5 would be approximately
that of the condensate from a melt in the distillation pot whose composi-
tion corresponded to the first one given in Table 6.5. The second compo-
sition is coincidentally also that of the MSRE carrier salt. Linear ex-
trapolation of the vapor-pressure equation for the carrier salt gives
pressures of 1 X 107° mm at 440°C (the liquidus temperature) and 4 X 10~%
mm at 540°C,
125

ORNL-DWG 65-12054
500

200

100

S0

20

P (mm)

0.5

1000 °C

 

0.2 ,
0 20 40 60 80 100

BeF, LiF (mole %)

Fig. 6.4. YVapor Pressures in the LiF-BeF; Systemn.
126

Table 6.5. Vapor Pressure of LiF-Bel;-ZrF, Melts

 

Composition
(mole %) Temperature Range

Measured (°C) Vapor-Pressure Equation

 

LiF  BeF, ZrF,

 

90.0 7.5 2.5 1121 .5-1283 log P(mm)

i

8.940 — 11,480/T(°K)

64.7 30.1 5.2 960—-116"7 log P(mm) = 8.760 — 9,880/T(°K)

il

 

ORNL-DWG 65-10871R
TEMPERATURE (°C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1250 1150 1050 950
100 | NG T |
80 a<Q%- —t—
, , I
60 - f\iq%f e
I . \\ V)g\ . L
40 ot O 6L - :
N NG Fig. 6.5. Vapor Pressure of
- —\\?@ %, | LiF-BeFp-ZrF, Melts.
S \, Or B ° 5
£ 20— X2 - oo
Q /A\\\.OO \ />)O/,
Gk \io
\Q\<_\ 6\\@ N <//<\
10— ——— A8 e e N
8 e A*¥ /;\ \3;5\ - [ \@@A\ ]
L O)O/ e N \e
6 o - N\, |
_ _ N
4
0.64 0.68 0.72 0.76 0.80
10097 k)

The Quest for Liquid-Liquid Immiscibility in LiF-BeF, Melts

Investigation of the activities of BeF, obtained from HF-H-0 equi-
libria®! suggested that at composition 80 mole % BeF,, 20 mole % LiF,
liquid=-liquid immiscibility might occur above 700°C. In an experiment
carried out in a glove box filled with helium, a nickel crucible contain-
ing a mixture of pure Bels and pure LiF in mole ratio of 4 to 1 was vis-
ually observed during temperature cycling. After heating first to 700°C,
two liquid layers were visible; the more dense layer was very viscous;
the top layer was much thinner and easily stirred. As the temperature
was slowly raised, the viscous layer diminished until at 850°C the mixture
appeared homogeneous. Subsequent coolings and reheatings between room
temperature and 850°C failed to reproduce the liquid-liquid immiscibility.
127

These observations indicate:

1. Liquid-liquid immiscibility was not exhibited by the 80-20 mole %
BeFs-LiF mixture.

2. Insufficient initial mixing of the pure components can give the ap-
pearance of immiscibility; mixing is somewhat difficult because of
the high viscosity of BeFs.

During cooling of this melt, a translucent phase was observed to set
in at approximately 500°C. This temperature coincides with the published22
liquidus temperature for this composition.

Removal of Ilodide from LiF-Bel's Melts by HF-H, Sparging

 

The rapid removal of fission product 1357 from a molten-salt fuel
would be an effective way of reducing the amount of 13°Xe which is formed
in the fuel. Previously reported preliminary studies®? indicated that
the equilibrium

HF(g) + I"(d) = F(4a) + HI(g) (1)

would permit the removal of I~ (the form of iodine expected to be present)
by HF sparging at a rate which was large compared to the 6.7-hr decay of
1357 to 13°Xe. The technique for these measurements has since been im-
proved, and the effect of temperature and melt composition on the equi-
librium has been studied.

Apparatus. A nickel reaction vessel (2 in. in inside diameter by
15 in.é was used in these measurements. Available thermochemical data
indicate that in the temperature range studied (474 to 635°C) no reaction
between HI and nickel should occur in the presence of approximately 1 atm
of hydrogen:

2HI(g) + Ni(s) « NiI,o(s) + Ho(g) .

However, as the gases are cooled in the off-gas line, reaction should
become apprecilable. Consequently, the off-gas line was provided with a
gold liner sealed at the end to a Teflon tube (Fig. 6.6) that carried the
exit gases to the aqueous scrubber used for gas analysis. Tests in which
130T tracer was used indicated that with this arrangement there was no
significant holdup by adsorption of HI on the walls of the off-gas system.

Procedure. After sparging the melt (~275 g) with HF-H, to remove
oxide and to reduce any metallic impurities, the system was flushed with
helium and then opened briefly at the upper end of the off-gas tube to
allow the addition of a weighed pellet of Nal. The system was flushed
again with helium. During this period, small amounts of iodine were
evolved, presumably because traces of oxygen or water had entered the
system during the sodium iodide addition. Hydrogen sparging caused this
evolution to cease after a short time. The HF flow was then begun, at a
partial pressure in Hp, of <0O.1 atm.
ORNL-DWG 65-12055

<-——TEFLON TUBING (Y4-in. DIAM)
TO FIRST BUBBLER

3/4—in. SWAGELOK
FITTING

SWAGELOK NUT

Au TUBE FLARED OVER
Ni TUBE

 

Y/230—in—thick Au TUBE
TEFLON STOPPER
3/4-in~diam NICKEL TUBE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N GAS INLET— ™\
) Au TUBE HELIARC
23 in. ' _— WELDED TO Ni
\
| ——THERMOCOUPLE WELL
l«—2—in—diam NICKEL
VESSEL
15 in
L 4 | ]
/'\\—//-\-‘__—
~ ™
L] N
i ]
Fig. 6.6. Reaction Vessel Used for lodide Removal.

The effluent HIF-HI-H, mixture was analyzed by bubbling it through a
solution containing a known amount of NaOH to neutralize the HI and HF,

a known amount of arsenite to reduce any il1odine present, and a pH indi-

When the indicator showed that all the NaOH had been neutralized,
In each run a succession of

These were analyzed for iodide

cator.
the solution was replaced with a fresh one.

neutralized scrub solutions was obtained.
by potentiometric titration vs standard AgNOs; solution and for Io by back

titration of The arsenite with standard iodine solution. From the re-
sults, the number of millimoles of HF, HI, and I, which had entered a
129

scrub solution could be determined. No significant amounts of I, were
f'ound.

Results. From the equilibrium constant for reaction (l),

K = Bypope/Prpde- (2)
the equilibrium quotient
Q = PHI/PHF [T”] kg/mole (3)

should be a constant at a given melt composition and temperature, pro-
vided the activity coefficient of iodide ion in the melt does not change
with the iodide concentration. If reaction (1) is the only important re-
action taking place, the moles of HI (nHI) and HF (nHF) evolved will be

related to Q and the weight of the melt (w) as follows:

Any g Q
_ San (4)
(n%r = nyp) W HF

1n (ifi.:—:—nfi-I-):—@- (5)

 

HEI w HF °’

where HEI 1s the maximum quantity of HI which can be evolved and should

be equal to the total number of moles of I present when = 0., In terms
of the iodide concentration in the melt, Eq. (5) becomes

In Bty = -2, (6)

where [I7]° is the iodide concentration when Nem = 0.

The results were generally consistent with Eq. (5). Plots of

log [ngp/(ngy = ngp)] ve TgF

were linear over wide ranges (Fig. 6.7) of the fraction of iodide remain-
ing in the melt. The value of Q was found not to be significantly de-
pendent on the HF flow rate (0.35 to 1.57 millimoles min~1 kg™1), on the
partial pressure of HF (0.02 to 0.1 atm), or on the initial iodide con-
centration (0.004 to 0.04 mole/kg). This indicated that reaction (1) was
The only significant reaction occurring, that equilibrium sparging con-
ditions were obtained, and that the activity coefficient of iodide ion
did not vary appreciably over the concentration range of iodide employed.

One anomaly encountered, however, was that the values of nfil or [I~]°

deduced from plots of the data were typically 20% less than the amount of
130

ORNL-DWG 65-9872R
APPROXIMATE TIME (min)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 20 40 60 80 100 120
b I l T 7T  TT 1 I T
- >\.\" ST T L l ) ] 1 B .
0.5 N S na - i
. \\\."* | @=44.7 kg mole
< " Fig. 6.7. Removal of Todide
L 02 e ' S N by HF Sparging.
S .\\
8T Of | — S NG e e ]
< N
c 005 — [1719: ~0.005 mole kg=! |~
8'~ * Hy, FLOW: ~90 ml min~" A
T
<

— Pyg: ~0.047 atm - °
MELT: 275 g 2LiF —BeF,
~ TEMPERATURE:  487°C

 

0.02 +—

 

 

 

 

 

 

o
0.04 1
0 5 10 15 20 (x1073)
MOLES OF HF PASSED

 

iodide added to the salt initially. Runs with 1307 revealed that this
poor material balance definitely was not caused by retention of iodide

in the reaction vessel or on the walls of the off-gas line. Rather, these
tests indicated that a fraction of the HI passing through the aqueous
scrubbers was adsorbed on particulate matter in the gas stream and in this
form was not caught by the NaOH solution.

While this effect will be investigated further, it is felt that it
will not appreciably alter the present estimates of Q (Figs. 6.8 and 6.9).
For 2LiF-Befs, the variation of @ with temperature is given by the fol-
lowing expressions:

log @ = —1.094 + 2.079(10%/1) . (7)
The initial results on the effect of the BeF, mole fraction on Q indicate
a maximum in Q somewhere in the range XBng = 0.33 to 0.50,

From Eq. (6), the number of moles of HF which must be passed per
kilogram of melt in order to remove half the iodide present in the melt
is given simply by 0.693/Q. For example, at 500°C, Q = 40 kg/mole, and
hence, half the iodide present will be removed by passage of 0.0173 mole
of HF (388 std cm3) per kilogram of melt. In order to achieve a removal
half-time of 1 hr, the HF flow rate thus must be 0.0173 mole/hr per kilo-
gram of melt.

In a reactor, 1357 could be removed by passing continuously a frac-
tion F of Tthe fuel per hour through a gas-liquid contactor. If This con-
tactor removed a fraction f of the iodide present in the side stream,
then the half-time in hours for removal of iodide from the entire system
would be

ty/2 = 0.693/F(f) . (8)
131

Fig. 6.8. Variation with Tem-
perature of Equilibrium Quotient for

ORNL—-DWG 65-—9874

TEMPERATURE (°C)

650 550

600

 

 

 

 

 

 

 

 

 

 

 

 

 

 

500

1.314 1.35

T
[
lodide Removal from Z2LiF-BeFs,. 2 AH=-95 keal
<
S
1.07 .1 145 119 1.23 .27
1000/
7 (CK)
ORNL-DWG 65-12056
80
/.
o
70 LN
v N\
s N\
d
s \\
s
s AN
60 ~ .
s
— s N\
T v N
g _~ \
o o N\
@ 50 N
o
£ \
S \\\
\
40 N\
\\
®
S
30 N,
20
0.35 0.40 0.45 0.50
MOLE FRACTION BeF,
Fig. 6.9. Variation with LiF-BeFy Melt Composition of Equilibrium

Quotient for Iodide Removal at 482°C.
132

If a countercurrent gas-liquid contactor equivalent to n ideal stages were
usg%, the fraction of lodide extracted from the side stream would be given
by

<QnHF/ws) -1

n-+1 /

QW) =1

(9)

 

wherein the ratio QHF/WS denotes the moles of HF passed per kilogram of

melt passed through the contactor. From this expression it is clear that
the moles of HF passed per kilogram of salt in the side stream must be
greater than l/Qo 1f it is appreciably greater, the number of stages re-
quired for good iodide removal efficiency 1s 1low.

 

QQHF/WS n (for £ > 0.9)
1.1 7
1.2 5
1.5 Z,
2 3

Hence, the removal half-time is determined primarily by the side-stream
flow rate F. Under these conditions the molegs of HF per hour required to
achieve the desired removal half-time is given by

 

moles HF 0,693 an
e (l*>wT,

hr T1/2 WS

where Wp 1s the total weight of the fuel in kilograms. Since the product
(QHF/WS§Q must be greater than unity, the minimum amount of HF required

would be

moles HF 0.693 W
T
 hr T

 

>
hr tl/g

In future tests attempts will be made to improve the efficiency of
HI trapping. Preliminary results of tests with 1307 indicate that the
trapping efficliency is greatly increased by prefiltering the gas stream,
In addition, the effect of melt composition on Q will be investigated
further.
133

calt Compositions for Use in Advanced Reactor Systems

 

Blanket Salt Mixtures for Molten-Salt Breeder Reactors

 

The blanket salt moest recently proposed for use in molten-salt breeder
reactors:hsen17LiF-BeF2—ThF4 mixture (71-2-27 mole %) whose liquidus tem-
perature is 1040°F (560°C). Other compositions of ‘LiF-BeF,-ThF, may be
attractive as blanket salts if the advantages of lower liquidus tempera-
tures are not offset by the attendant reduction of thorium concentration.
(The concentrations of BeF, are not sufficient to affect the viscosity
adversely.) Mixtures of 7LiF-BeF2—ThF4 whose compositions lie along the
even-reaction boundary curve L = 3LiF«ThF,(ss) + 7LiF-6ThF, appear to
qualifly as the best blanket salts from phase-behavior considerations.

The ThF, concentration of these mixtures varies from 6.5 to 29 mole % as
LiF concentrations change from 63 to 71 mole %. Liquidus temperatures
for this range of compositions vary from 448°C (838°F) to 568°C (1044°F).
In the molten state these mixtures contain thorium concentrations ranging
from 850 to 2868 g/liter at approximately 600°C. These data are summa-
rized in Figs. 6.10 and 6.11 and in Table 6.6. Optimization of thorium

ORNL-LR-DWG 37420AR2

 

 

 

 

Theg, 1444
0
MO TEMPERATURE IN °C
COMPOQSITION IN mole %
1050
LiF-4ThE,
LiF-2ThF, 1000
LiF-2ThE;
950
P 897
2L|F-BeF2
900
7LiF-6ThE,
P62 , 850
P 597
£ 568 \
3LIF-ThE 4., 750
E 565
65
6. 20 06
%?%bQ’C)Qb 20
S 0 00 S50 £ 526
o0/ \L 3
LiF BeF,
845 2LiF- BeF; 500T450 400 400 450 500 548
P 465 £ 370

Fig. 6.10. Phase Diagram of the System LiF-BeF,-ThF,.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL-DWG 65—-12057
(°F) (°C)
1100 593 3400
1050 566 < 3000
LIQUIDUS TEMPERATURE‘/////,

/////////// o
o
& 1000 538 , 2600 3
D / O
5 -
o <
LJ -+ ¢
a Q
2 / 2
W 950 510 2200
~ / o
//////////// Z
N O
T
-
900 482 , 1800
=
/// o
O
=
O
O
/ THORIUM (g/li‘rer AT ~600°C) s
850 454 ——1 1400 S
/ / =
O
I
// ]

// 1000

600

5 10 15 20 25 30
ThF, (mole %)
Fig. 6.11. Thorium Content of LiF-BeF,-ThF', Compositions on the

Even-Reaction Boundary Curve L = 3LiF-ThF, (ss) + 7LiF .6ThE, .

 

 

 

 

Table 6.6. MSBR Blanket Compositions
Thorium Liquidus Temperature Composition
Concentration
(g/liter) °C °F ThF, LiF BeF,
863 450 842 % 63 30

1207 469 876 10 66 24
1828 500 932 16 69 15
2419 550 1022 22.5 71 6.5
2868 568 1054 29 71

 
135

concentration and liquidus temperatures may be made with the aid of Fig.
6.11; if 2000 g/liter of thorium is adequate to achieve good breeding
gain, the composition containing 17.5 mole % ThF, may be used, and the
liquidus temperature of this mixture is 516°C (960°F).

Coolants for the Molten-Salt Breeder Reactor

 

The steam circuit of the Bull Run Plant of the TVA, the type of plant
to which a proposed 1000-Mw(electrical) MSBR would be coupled, operates
with temperature extremes of 700 and 1125°F (371 and 607°C). An essential
requirement for the development of the reactor is the availability of a
cheap, chemically stable secondary-coolant fluid which in the liquid state
has acceptable nuclear, chemical, and physical properties. No salt mix-
ture has as yet been discovered which meets all the criteria imposed.

Preliminary examination of the NaF-NaBF, system, which according to
Russian reports®®’?® forms a eutectic mixture melting at 304°C, disclosed
that the eutectic formed from the pure fluorides actually melts at 373 to
375°C, but that the oxide contaminant B,0; markedly depresses the liquidus
temperatures. Although high solubility of oxide ion in the coolant is
chemically advantageous, By05 exhibits a strong tendency to form glasses
in the molten state, an effect which has already been noted in the vis-
cosity of NaP-NaBF,-B,O; liquids we have examined in the laboratory.
Little information is available from the literature concerning the prop-
erties of these and other liquid mixtures of fluorides and fluoborates.
Evaluation of thelr potential as secondary coolants for the MSBR will
accordingly require considerable additional investigation.

In another set of experiments, phase transition temperatures of two
samples, one consisting of pure NaBF,, the second of a mixture of 39.1
mole % NaBF, and 60.9 mole % NaF, were determined from an examination of
cooling curves. The samples were encapsulated under vacuum in INOR-S8.
Temperatures were read with an NBS-calibrated Pt-Rh thermocouple in an
INOR-8 thermowell which extended about 1 in. into the melt.

Cooling curves for the NaBF, sample were obtained at cooling rates
of from 1.0°C/min down to 0.2°C/min. Every run exhibited supercooling,
ranging from 5 to 15°C, which made it difficult to obtain a very precise
melting temperature. FIFrom these curves, the best estimate for the melting
point of NaBF, is 396 * 2°C.

Cooling curves for the 39.1 mole % NaBF;~60.9 mole % NaF mixture were
obtained at cooling rates from 0.3°C/min to 1.4°C/min. Relatively little
supercooling was encountered at the only phase transition temperature,

375 £ 1°C, observed for this mixture. The sample was cycled through the
temperature interval 255 to 465°C. Again, the 304°C eutectic reported
in the Russian literature?? was not observed.

The capsule containing the NaBF,-NaF mixture was cut open after com-
pleting the cooling curves and was examined; there was no evidence of
corrosive attack.
136

Viscosity of NaBF,

 

In a glove box containing very low concentrations of oxygen and mois-
ture, the viscosity of NaBF, was measured with a Brookfield viscosimeter
and found to be 7 * 2 centipoises at 466°C and 14 * 3 centipoises at 436°C.
The precision is poor because the apparatus was primarily designed to
measure high viscosities.

No mixtures of NaF and NaBF, were measured, but such melts would most
likely have been less viscous than pure NaBF,; the poor precision occur-
ring at Tthese lower viscosities discouraged us from attempting further
measurements.

ORNL-DWG 64-1993

ThE,
114

PRIMARY PHASE AREAS: _
LiF TEMPERATURES IN °C

NaF i} COMPOSITIONS IN mole %
TNaF - 2ThF, . +H-+++++  INDICATES SOLID SOLUTION
4NaF - ThF,

2NaF - ThE, 1050

2LiF-NaF - 2ThF, ~NaF - ThE, LiF-4ThFg—g—- —
SOLID SOLUTION

3LIF- ThF, 1000

7LiF- 6ThE,

LiF-2Thg N\ W /\

LiF- 4ThF, LiF- 2ThF,~ 950 —NaF - 2ThF,
NaF - 2ThF,

3NaF-2ThF, A 900 A
ThE P-897—/— [~ P-83
4

SGlIGICISIOMUIWICIGICIS)

 
 
 
  
   

7LiF- 6ThFg

 

AN 3NaF - 2ThE, 712
"TN—£-690

 

\~7NaF - 2ThF,
N>P-645
4NaF - Th,
8507 o,
//Kjf/// 950
7 NaF
845 E-652 995

Fig. 6.12. The System LiF-NaF-ThF,.
137

Fuel and Blanket Materials for the Proposed MOSEL Reactor

 

A proposal for investigation of the feasibility for developing an
intermediate-energy, molten-fluoride-salt reactor with the fuel salt
cooled by direct contact with liquid lead was described recently.28 In
preliminary discussions, Kernforschungsanlage Julich and ORNI staff mem-
bers have arrived at initial choices for fuel and blanket compositions.
Inspection of the ternary phase diagrams containing thorium fluoride and
mixtures of lithium fluoride, sodium fluoride, beryllium fluoride, or lead
fluoride shows that at concentrations above 10 mole % only the LiF-NaF-
ThF, system (Fig. 6.12) affords mixtures in which the liquidus is as low
as 500°C. Accordingly, the initial choice of the fuel composition is
approximately LiF-NaF-ThF,-UF, (44-32-20-4 mole %). The blanket-salt
composition would be LiF-NaF-ThF, (44-32-24 mole %); alternatively, the
composition LiF-ThF, (73-27 mole %) may be used. Composition of the salt
mixtures used ultimately in the experimental program will be based on
the results of further calculations to be performed in the next several
months.

Recovery of Protactinium from Fluoride Breeder Blanket Mixtures

 

Introduction

 

Demonstration of a process for extracting protactinium from a molten-
fluoride breeder blanket 1s an important part of studies directed toward
establishing the feasibility of molten-fluoride thermal breeder reactors.
An earlier investigation29 showed that protactinium could be removed by
oxide precipitation from one molten mixture, LiF-BeF,-ThF, (67-18-15
mole %), that is suitable for use as a breeder blanket. Other mixtures
are currently of interest, and it is also desirable to develop other,
more efficient, removal processes. The glove box used in the earlier
investigation, when 231pg ywas present in the fluoride mixture in addition
to ?33Pa, is no longer available. The use of ?31Pa permits attainment of
expected operating concentration levels (50 to 100 ppm) without the ex-
cessive gamma activity associated with milligram quantities of 233pg,

This report contains a brief description of a new glove box facility suit-
able for laboratory-scale studies of methods of recovering protactinium
from fluoride mixtures and results of a preliminary experiment on precip-
i1tation of protactinium at the tracer level,

Facility Description

 

The High-Alpha Molten-Salt (HAMS) ILeboratory is located in room 127,
Building 4501, an area formerly occupied by a hot cell. A view of the
laboratory, Fig. 6.13, shows the seven interconnected glove boxes that
are presently installed in the laboratory. The glove box at the right end
of the train is connected to a hood which exhausts into the hot cell off-
gas system. Air is pulled into the glove boxes through a rectangular-
shaped filter on the left side of the top of each box and exhausts through
138

 

Fig. 6.13. View of High-Alpha Molten-Salt Laboratory.

a pair of 8 X 8 in. absolute filters mounted in the back of the box and
through another enclosed absolute filter before reaching the glass-fiber-
reinforced header pipe which discharges into the plant hot off-gas system.
The pressure in the header line is controlled automatically to —0.5 in.
Ho0 in the glove boxes, and the loss of a glove from two boxes in the
glove train results in a negligible increase in pressure in the other
boxes. The room can be maintained at a pressure of —0.5 in. H20 with
respect to the adjoining areas by removal of air through the glove boxes
and hood. The large filters on the wall back of the glove boxes form the
alr intake of a recirculating air-conditioning system.

Most of the molten-salt work will be performed in the stainless steel
box shown in Fig. 6.14. The manifold and gages mounted on the back wall
of the box control application of vacuum or admission of helium, hydrogen,
and anhydrous HEF to a flanged nickel pot in a well below the floor level
of the glove box.20  The well, made of 4-in.-ID stainless steel pipe, is
heated by a 5-in. tube furnace supported by a jack below the box. The
Pyrovane controller at the top of the panel to the right of the box con-
trols the furnace temperature indicated by a thermocouple adjacent to the
heating coil, while the Brown recorder shows the temperature at the junc-
tion of an Inconel-sheathed thermocouple immersed in molten salt about
1/2 in., from the bottom of the nickel liner inside the pot. Other boxes
in the train contain an analytical balance (Mettler, type H-l5), a Zeiss
polarizing microscope, and a quenching furnace for supporting research.
139

PHOTO 80772

 

Fig. 6.14. Stainless Steel Glove Box for High-Temperature Studies
with Molten Salts Containing 231Pa.
140

Oxide Precipitation of Protactinium

 

The precipitation of protactinium from LiF-BeF,-ThF, (73-2-25 mole %)
was studied at tracer level (~1 ppb) using “32Pa in purified salt to test
the equipment and operating procedure.

Filtered samples of fused salt were removed from the melt at 620 or
630°C by inserting a l-l/Z—in. length of 3/8-in. copper tubing with a
sintered-copper filter welded in one end and 18 in. of l/8—in. nickel
tubing brazed to the other end.

Weighed amounts of ThO, were added to the molten mixture through the
same fitting that permitted introduction of the filter stick, and mixing
was accomplished by bubbling helium through the melt for 1/2 to 1 hr.
After precipitation of protactinium was complete, the mixture was allowed
to cool to room temperature under an atmosphere of helium. It was later
remelted and treated with a mixture of H, and HF (10 to 1 by volume) to
demonstrate that the precipitated protactinium could be returned to so-
lution. The results of this experiment confirmed the conclusion of the
earlier study that protactinium can be readily precipitated from a molten
fldoride mixture by addition of thorium oxide and that the precipitate
can be returned to solution by treatment with HF. It also appeared that
precipitation occurred due to inadvertent addition of water. Quantitative
conclusions are not Jjustified because insufficient time was allowed in
the early stages of the experiment for equilibrium to be obtained, and
because a large part of the mixture was found to have solidified on the
liner about 2 in. above the surface of the melt. A large vertical gra-
dient in the furnace well is necessitated by the requirement of a cool
glove box floor. The operating temperature was only 60 to 70° above the
freezing point of tTthe mixture GV560°C), and gas bubbling through the melt
apparently carried portions of it to a point on the liner that was cool
enough to permit freezing to occur.

Operating experience obtained in this experiment indicated that no
special problems should be encountered with similar experiments in the
glove boxes with higher-specific-activity materials such as %31pa., A
speclal furnace and other equipment modifications to minimize the thermal
gradient problem are under consideration.

Development and Iivaluation of Methods for the
Analysis of MSRE Fuel

 

 

Determination of Oxide in MSRE Fuel

 

On the basis of continued study of methods of oxide determination
in MSRE salts, the hydrofluorination method was selected as most adapt-
able to immediate hot-cell operation. While the method is less sensgitive
than the inert—gas-fusion3l or KBrF432 methods, the limits of oxide de-
tection can be lowered by increasing the sample size.
141

d33

The hydrofluorination metho is based on the reaction

0%~ + 20F(g) — Hy0(g) + 27

which occurs when a molten salt sample 1s purged with an Ho-HEF gas
mixture. Either the water evolved or the HF consumed can serve as a
measure of oxide; however, the water offers the more facile measure-
ment .

Preliminary tests have shown that the oxide is readily removed
from LisBeF, salts at 500°C with 0.02 atm of HF in the purge gas.
Adding zirconium to the melt causes a shift in equilibrium of the re-
action, resulting in an incomplete recovery of the oxide at these re-
action conditions. By increasing the melt temperature to 700°C and
enriching the HF in the hydrofluorinating gas to 0.15 atm, the oxide
is removed from a 100-g melt in about 4 hr.

The application of this method to the analysis of radioactive
samples requires the development of (1) a sampling technique which
minimizes atmospheric contamination, (2) the incorporation of a water-
measurement technique which is convenient for hot-cell operations, and
(3) the fabrication of a compact apparatus to conserve space in the
hot cells. It was necessary to adapt the sampling technigques from
methods already developed to obtain samples for wet analyses. By
using a copper enricher ladle, a 50-g sample which can be transported
in the existing transport container is obtained. Atmospheric expo-
sures will be minimized by remelting and hydrofluorinating the entire
sample 1in the sampling ladle. The ladle will be sealed in & nickel-
Monel hydrofluorinator, with a dellvery tube spring loaded against
the surface of the salt. Prior to melting, the apparatus will be
purged with hydrofluorinating gas mixture to remove water on the inner
surface of the hydrofluorinator and, hopefully, any contamination on
the exposed surface of the salt. When the sample melts, the delivery
tube will be driven by spring action to the bottom of the ladle for
efficient purging of the melt.

In all preliminary tests, the water in the effluent gas has been
measured by Karl Fischer titration. While the Karl Fischer reagent
has been shown to be remarkably stable to radiation, the titration would
be difficult to perform in the hot cell. An electrolytic moisture moni-
tor is ideally sulted to remote measurements but is subject to interference
and damage from HF. A sodium fluoride column operated at about 90°C
removes HF from the effluent gas without significant holdup of water.

A schematic flow diagram of the apparatus for hot-cell installa-
tion is shown in Fig. 6.15. The apparatus has been designed and is
now being fabricated during component testing. A modular design has
been selected to facilitate any changes found necessary during com-
ponent testing and to permit necessary repairs in the hot cell. Except
for the hydrofluorination furnace, all hot-cell components are contained
within a 16-in.-square compartment.
142

ORNL-DWG 65—8855A

TO
SODA
LIME
TRAP

|

 

 

  
   

MOISTURE
/ MONITOR

| 5| Pb SHIELD

SPLITTER

 

 

 

 

 

 

 

NN
%
P
I
@D

1
HF |
LHe BACK FLUSH

__——NaF —HF TRAP

PURIFIER

 

 

\

 

 

 

 

 

 

 

 

 

CELL WALL

 

 

 

 

 

 

N\

HYDROFLUORINATOR

 

Fig. ©.15. ©Schematic Flow Diagram of the Hydrofluorination Ap-
paratus for the Determination of Oxide.

Improved results have been obtained with a components test fa-
cility. The apparatus, which has nickel gas lines maintained at a
temperature above 100°C, is designed to minimize void space and dead
legs and to improve melt purging efficiency. While an apparatus for
filling ladles and evaluating the effects of atmospheric contamination
was being assembled, test analyses were run on standard additions of
ZrOz and U0y to a 50-g fuel melt. In each analysis the purge gas flared
at 150 cm”/min. When the salt temperature was held at 700°C with the
HF' concentration =0.1 atm, the oxide was evolved in 1 hr. Recovery
data obtained by the Karl Fischer titration are shown below.

 

 

Sample Oxide Added. (ppm) Oxide Recovered
P Based on 50-g Sample (ppm)
UO2 425 455
355 345
415 390
ZTOQ 405 435
520 490
805 790
425 435
415 420

325 320
143

oeparate tests of a Beckman electrolytic moisture monitor have
yielded 98 * 2% recovery of injected water. These tests also indi-
cated that the cell efficiency is easily decreased by overloading;
therefore, the capillary splitter shown in Fig. 6.15 is required. A
moisture monitor and splitter have been installed on the component test
facility, and recoveries of water from ZrO; addition are being measured.

Electrochemical Analysis

 

Work is continuing toward the goal of adapting electroanalytical
methods to the in-line analysis of impurities (e.g., corrosion products)
and other electroactive species in the molten MSRE fuel.3%: 3% A new
phenomenon, which suggests a potential method for the coulometric de-
termination of oxide, has been observed in both the MSRE fuel and
solvent salts. Current-voltage curves recorded in the positive di-
rection have always shown an anodic peak at the pyrolytic-graphite
indicator electrode (PGE) at about +1.0 v vs the platinum quasi-reference
electrode. It has been concliuded from the shape of the curve that the
electrode was belng passivated. Recently, when the PGE was observed
visually while running current-voltage curves under vacuum conditions,
1t was noted that gas bubbles were coming from the end of the sheathed
PGE at the potential of the anodic wave. Gas formation was also ob-
served at this potential when a platinum wire was used as an indicator
electrode.

The gas lines servicing the electrolytic cells were redesigned to
incorporate a liquid-nitrogen molecular-sieve trap to collect the gases
which are evolved during electrolysis of the melt. The results of the
gas chromatographic analyses of the evolved gases are shown in Table
6.7.

These preliminary results indicate that the anodic wave which has
been observed consistently in these fluoride melts is due to the oxi-
dation of oxide (0%~ — 1/2 Os + 2e). It is believed that the CO and
COp result from the reaction of the oxygen with the graphite cell prior
to being pumped from the electrolytic cell.

Table 6.7. Composition of Gases Obtained from the Electrolysis
of LiF-Bels-Zrl, and LiF-BeF,

Values in mole percent

 

 

0,  CO CO,  CH,  H Anode
LiF-BeF,-Z2rF, 2.3 3.4 83.5 8.1 3.0 Pyrolytic graphite
0.05 5.2  95.6 Graphite cell walls

LiF-BeF» 6.6 16.2 75.6 0.7 <0.3 Platinum

 
144

The presence of Hp and CHy, may be a result of evolution of Hy at
the platinum counterelectrode followed by reaction with the graphite
cell to form CHy,. This is supported by the fact that neither Ho nor
CH, was found to be present in the gases evolved from the second elec-
trolysis experiment in Table 6.7. In this case, due to a high current
density, zirconium was plated on the counterelectrode and probably re-
acted with any Hp to form zirconium hydride. Further support is ob-
tained from the fact that if the potential of the cathodic 1limit of
LiF-Bel's-Zrt, is exceeded (reduction of zirconium) and the scan is re-
versed, gas evolution from the electrode is observed only at that in-
stant when all the zirconium has been stripped from the electrode.

The current efficiency for the reduction of oxide in the first
run of Table 6.7 was about 85%; however, a low current density was
used. Higher current densities were used in the second and third
runs, but the current efficiencies dropped to about 40%.

In order to make feasible the controlled-potential coulometric
analysis of oxide in these melts, it will be necessary to determine
the proper conditions for theoretical current efficiencies. To de-
crease the time of analysis, a better means of stirring the melt under
vacuum is also required. It is planned to study these problems through
further electrolysis experiments.

Electrochemical studies were also made to identify a small re-
duction wave which occurs in the MoRE fuel solvent salt. This wave,
which 1s observed at both platinum and pyrolytic-graphite indicator
electrodes at approximately —1.2 v vs the platinum quasi-reference
electrode, 1s now believed to be due, in part, to the reduction of s
small amount of hydroxide ion impurity in the melt. As previously
pointed ou.t,35 however, the overall electrode reactions appear to be
more complex than first envisioned.

After making standard additions of anhydrous lithium hydroxide to
the melt, it was observed that the overall limiting current increased
and then slowly decreased over a period of several days to a reason-
ably constant value. Further tests revealed that appreciable quantities
of O do not remain as such in the melt bul instead react with the
fluoride ion (OH™ + F~ — HF + 0°7) to form oxide and HF. This was
confirmed by adding LiOH and pumping the off-gas from the electrolysis
cell into a cold trap that contained a known amount of standard sodium
hydroxide and then determining the amount of sodium hydroxide neutral-
ized at 24-hr intervals. In one test where 220 mg of LiOH was added
to the molten LiF-BeF, (43 ml), approximately 80% appeared to have
reacted with the melt to evolve HI' over a period of eight days.

Current-voltage curves recorded with the melt under reduced pres-
sure were not significantly different from those recorded with the melt
in a helium atmosphere; however, a few gas bubbles could be seen evolv-
ing from the indicator electrode at a potential of approximately —1.2 v.
This appears to be visual evidence that the wave is in part due to the
reduction of hydroxide impurity (O + e — 0% + 1/2 Ho) and that hy-
drogen is being discharged at the cathode.
145

opectrophotometric Studies of Molten-Salt Reactor Fuels

The possible use of spectrophotometry is being investigated as an
analytical tool for the in-line determination of U(III) at the parts-
per-million level and estimation of U(IV) in molten-fluoride-salt re-
actors. In past studies, experimental techniques have been developed
for the spectrophotometric study of molten fluoride salts, which are
corrosive to most known window materials. Spectra of both U(III)3°
and U(IV)37 in fluoride salts of interest have been obtained. Tri-
valent uranium in LiF-BeFp-type melts exhibits a very intense absorp-
tion peak at_ 360 mp with a molar absorptivity of approximately 500
liters mole™t em™t. Tetravalent uranium in LiF-Bel's-type melts is
almost transparent in the region of 360 mi, with a molar absorptivity
of less than 2 liters mole™t cm™t. At 1000 mu, U(IV) exhibits a peak
with a molar absorptivity of 15 liters mole ™t cm™t. Both U(IV) and
U(III) are essentially transparent in the region of 700 to 710 mp. It
is therefore feasible to determine U(III) down to a level of ca. 300
Ppm in the presence of up to 1 mole 9 of U(IV) in molten LiF-EEFg—ZrF4.
Likewise, if the concentration of U(III) does not exceed ca. 1000 ppm
it would be possible to at least estimate the concentration of U(IV)
in this molten salt by determining the absorbance of the solution at
1000 mp. For both determinations the optical response of the solvent
could be measured at 710 mp. The results of other spectral studies
show that the corrosion products, which could be dissolved in the salt,
will not interfere with the proposed determination.

An idealized drawing of the spectrophotometric facility is shown
in Fig. 6.16. A modified cylindrical captive-liquid cell3® would be
used to contain the molten salt fuel in a windowless cell for the
spectral determination. The cell would be filled either by continuocusly
dripping the fuel into it or by activating the freeze valve periodically
to flush the cell with fuel salt. In either case, the volume of the
fuel in the cell will rapidly, in less than 6 sec, attain an equilibrium
volume and a relatively constant optical path length. Excess liquids in
either case will flow out of the bottom of the cell, collect on the cone-
shaped bottom, and drip into the return circuit. With liquids that wet
the modified captive-liquid cell, the optical path length is very re-
producible, 0.064 cm * 2%, with either method of filling. With liquids
that do not wet the cell, the path length is less reproducible, but tech-
niques are being developed which will improve this situation. It is ex-
pected that the radiation level of the volume of fuel to be held in the
cell, ca. 100 pl, will not be sufficiently high to cause defect colora-

tion of the windows if the Al,03 is pure.

The model 14H Cary recording spectrophotometer is optically and
electronically designed so that any interference caused by thermal
emission from heated samples i1s essentially eliminated. This optical
design will also eliminate interference from Cerenkov radiation. In
the optical design of a spectrophotometer for in-line use with a nu-
clear reactor, it is necessary to remove the electronic components from
the highly radioactive region. OSeveral advantages can occur as a re-
sult of the necessary lengthening of the optical path. The light beam
146

ORNL-DWG 65—-3975A

FREEZE VALVE

R

 

 

 

 

   

 

Ya-in- diam CELL\\\\\\\\Q\

Al,05 WINDOW

FOCUSED LIGHT BEAM

 

 

  
  

 

 

{/g-in—diam APERTURE

 

 

He —

MATERIAL: INOR-8

Fig. 6.16. Molten-5alt Reactor In-ILine Spectrophotometric Fa-
cility.

of the instrument can be concentrated and imaged specifically for use
with the captive-liquid cell; presently, ca. 90% of the light which is
incident on a captive-liquid cell is lost because of necessary masking.
oecond, the extension of the light beam by the use of mirrors will re-
move any possible problems caused by high-energy gamma rays entering
the spectrophotometer; these rays will, of course, be lost in reflec-
tion. In cooperation with the vendor of the Cary spectrophotometer,

a proposed extended optical design is being considered.

A study of the analytical usefulness of reflectance spectra for
the identification and estimation of U(III) and other ionic species
147

of interest in solidified powdered fluoride salts has been initiated.
This technique should be of value in the analysis of irradiated solid
fluoride salts and as an aid in identification of reduced species if

significant reducing power 1is ever found in solid MSRE salts.

In screening tests, excellent reflectance spectra have been ob-
tained for salts such as UF,, PrFi, NdF3, and halides of structural
metals. These spectra were obtained with a model 14 Cary recording
spectrophotometer equipped with a model 1411 diffuse-reflectance ac-
cessory. A sealable reflectance cell with a quartz window has been
designed and fabricated so that the reflectance spectra of hygroscopic
samples can be obtained without exposure of the samples to air; the
cell would also serve to confine toxic or radioactive dusts in order
to prevent contamination. With this cell it should be possible to
examine the reflectance spectra of solidified LiF-BeFs-type salts which
contain Uk, and/or UF3. This work is continuing.

Analysis of MSRE Blanket Gas

 

otudies have been continued on the helium breakdown-voltage de-
tector for determining permanent gases in helium. The application of
this type of detector to the analysis of helium blanket gas which will
contain radiocactive fission gases seems promising. Tests with x-ray and
gamma-~-ray sources and with the off-gas of the MSRE capsule test No.
ORNL 47-6 indicated that radioactive gases would not be detrimental to
the detector operation.

The detector has been tested in a Beckman model 520 process chro-
matograph for an evaluation of its operation on a continuous analysis
basis. On a series of 35 samples of a standard gas mixture at the 10-
ppm level, taken over a 5-hr period, the relative standard deviation
of the peak heights was l%. This indicates the possibility of analysis
at sub-ppm levels.

The sensitivity of the detector for water in helium is estimated
to be 1 ppb. A major problem in water determinations at the ppm and
sub-ppm levels is the transmittal of the gas sample from the source to
the detector. All lines and valves must be heated to 250-300°C to pre-
vent excessive time delays in detecting concentration changes. At the
present time there 1s no chromatographic sampling valve which will op-
erate satisfactorily at these temperatures. A design has been conceived
utilizing a pneumatically actuated six-way metal diaphragm sampling
valve. Because of sealing problems in the first models of this valve,
a redesigned version is being fabricated. |

oince all components of this valve are metallic, problems of ra-
diation damage are virtually eliminated. This valve will also be ideally
sulted for application to gas analyses in a proposed in-line hydrofluori-
nation method for the determination of oxide content and reducing power
in the fuel of future molten-salt reactors.
148

Development and Evaluation of Equipment for
Analyzing Radioactive MSRE Fuel Samples

 

 

The equipment39 necessary to analyze radioactive MSRE salt samples
for Be, Zr, U, Fe, Cr, Ni, ¥, and Mo was installed in the hot cells of
Building 2026 by April 1965. The reliability of it was checked out by
the Analytical Chemistry Division personnel responsible for its use.
oeveral minor modifications were necessary; however, it was performing
properly at the conclusion of the checkout period.

oample Preparation

 

The salt samples delivered to Building 2026 during the precritical
and zero-power experiments were crushed, welghed, and dissolved in the
hot cells. Two portions of each pulverized salt were weighed and dis-
solved. The preparation of the salt samples for analysis proved to be
satisfactory after a few modifications were made to the equipment and
to the method of dissolution.

Sample Analyses

 

The solutions prepared from the salt samples were analyzed for
U, Cr, Zr, Li, Be, Fe, Ni, and Mo. The crushed salt was analyzed for
fluoride. The results obtained were generally satisfactory except for
uranium and'berylliuma4o The uranium determinations exhibited a nega-
tive bias of approximately l% from book values. The bias was evident
during both the precritical and zero-power experiments. The beryllium
results exhibited a positive bias of approximately 2% from book values.
Both methods are being investigated to determine if the biases actually
exist and, if so, what steps are necessary to correct them.

Quality Control Program

 

A quality control program was initiated prior to precritical sam-
pling. Control samples of synthetic solutions similar to dissolved non-
radioactive fuel-salt samples were analyzed along with the fuel-salt
samples. The program was established to verify the true percent standard
deviation in the individual methods employed in analyzing the MSRE fuel-
salt samples. The accumulated results on the synthetic solutions were
insufficient to determine the true percent standard deviations. How-
ever, the estimated values for U, Be, Zr, Cr, Fe, Ni, and Mo were 1.0,
5.0, 5.0, 15, 15, 15, and 15%, respectively, based on the data to Au-
gust 23.
10.

11.

12.

13.

14.

149

References

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 146.

 

P. N. Hauvbenreich, Preliminary Report on Results of MSRE Zero-
Power Experiments (internal memorandum).

 

R. E. Thoma, 'MSRE Salt Chemistry During Precritical and Zero-Power
Experiments, ' MSR-65-40 (Aug . 2 1965) (internal use only).

R. E. Thoma, 'Chemical Analysis of MSRE Flush and Coolant Salts in
Pren?clear Test Period,"” MSR-65-19 (March 19, 1965) (internal use
only/.

R. B. Lindauver, Preoperational Testing of the MSRE Fuel Reprocess-
ing Facility and Flush Salt Treatment No. 1 (internal memorandum).

Resgults are summarized in memo K-1-2079 from J. G. Million to R. E.
Thoma, Sept. 1, 1965.

MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp.
23841

 

C. M. Blood et al., "Activities of Some Transition Metal Fluorides
in Molten Fluoride Mixtures,' pp. 108-10 in Proceedings of the In-
ternational Conference on Coordination Chemistry, /th, Stockholm
and Uppsala, Jan. 25-29, 1962, Butterworths, London, 1963.

 

 

S. I. Cohen, W. D. Powers, and N. D. Greene, A Physical Property
Sumary for ANP Fluoride Mixtures, ORNL-2150 (Aug. 23, 1956, de-
classified Nov. 24, 1959).

 

 

S. I. Cohen and T. N. Jones, A Summary of Density Measurements on
Molten Fluoride Mixtures and a Correlation for Predicting Densities
of Fluoride Mixtures, ORNL-1702 (July 19, 1954, declassified Nov.
2, 1961).

 

 

B. C. Blanke et al., Density and Viscosity of Fused Mixtures of
Tithium, Beryllium, and Uranium Fluorides, MIM-1086 (December
1956, issued March 23, 1959).

 

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1961, ORNL-3122, pp.
118, 122-25.

P. B. Bien, S. Cantor, and F. F. Blankenship, Reactor Chem. Div.
Ann. Progr. Rept. Jan. 31, 1961, ORNL-3127, pp. 24—25.

 

S. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962,
ORNL-3262, pp. 3841.

 
15.

16.

17.

18.

19.

20.

21.

22 .

23,

2

150

Ixperimental work by P. B. Field, summer participant with Reactor
Chemistry Division, 1965, Assistant Professor of Chemistry, Vir-
ginia Polytechnic Institute.

D. J. Rose and M. Clark, Jr., Plasmas and Controlled Fusion, p.
296, MIT Press, Cambridge, Mass., 1961.

 

J. H. Shaffer, W. R. Grimes, and G. M. Watson, J. Phys. Chem. 03,
1999 (1959). =

 

5. T. Benton, R. L. farrar, Jr., and R. M. McGill, Preparation of
Anhydrous Deuterium Fluoride by Direct Combination of the Elements,
K-1585 (Jan. 29, 1964).

 

 

W. H. Rodebush and A. L. Dixon, Phys. Rev. 26, 851 (1925).

A. Buchler and J. L. Stauffer, IAEA Symposium on Thermodynamics,
July 1965, paper SM-66/26.

 

A. L. Mathews and C. F. Baes, Oxlide Chemistry and Thermodynamics
of Molten Lithium Fluoride—Beryllium Fluoride by Equilibration
with Gaseous Water—Hydrogen Fluoride Mixtures, ORNL-TM-1129, p.
104 (May 7, 1965).

 

 

 

R. E. Thoma, ed., Phase Diagrams of Nuclear Reactor Materials,
ORNL-2548, p.33 (Nov. 6, 1959).

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 137.

 

Equation (9) was obtained by use of the following relation which
applies to a countercurrent extraction system of n stages if the
distribution coefficient D is constant:

fraction extracted = 1 — { fiiifl 1 ;
(br)™ " =1

where r 1s the ratio of the flow rates of the two phases. In the present

case,

and

P__/RT
HI :
D = —=—— = QP __/RT kg/liter
[I...] I_IEw/ g/ J
r = liters of gas per kilogram of melt = Vg/'wS 5
QP .V
o g
br = WRT Ay /W
25.

20.

27 .

28.

29.

30.

31.

32.

33.
34.
35.

36.

37.

38.

39.

40.

Transl.) 3, 279 (1958).

151

V. G. Selivanov and V. V. Stender, Zh. Neorgan. Khim.
(1958).

hw

5 a7

 

Tbid., 4, 934 (1959).

V. G. Selivanov and V. V. Stender, Russ. J. Inorg. Chem. (English

 

W. F. Schilling, Proposed MOSEL Reactor Program at KFA Jiilich
(internal memorandum).

 

J. H. Shaffer et al., Nucl. Sci. Eng. 18, 177 (1964).

 

J. H. Shaffer, Reactor Chemistry Division, designed the manifold
system and supervised its construction.

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 160.

 

G. Goldberg, A. 5. Meyer, Jr., and J. C. White, Anal. Chem. 32,
314 (1960).

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 162.

 

MSR Program Semiann. Progr. Rept. Jan. 31, 1964, ORNL-3626, p. 151.

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 164.

J. P. Young, Anal. Chem. Div. Ann. Progr. Rept. Nov. 15, 1964,
ORNL-3750, p. 6.

 

J. P. Young, Anal. Chem. Div. Ann. Progr. Rept. Dec. 31, 1961,
ORNL-3243, p. 30.

 

J. P. Young, Anal. Chem. 36, 390 (1964).

 

F. K. Heacker et al., MSR Program Semiann. Progr. Rept. Feb. 28,
1965, ORNL-3812, pp. 155-60. |

 

R. E. Thoma, MSRE Salt Chemistry During Precritical and Zero-Power
Experiments (Aug. 2, 1965) (internal use only).

 

 
152

7. FUEL PROCESSING

The MSRE fuel processing system has been described in a previous
report.l Construction was completed, the system was leak tested, and
the tanks were calibrated. A test run using dry nitrogen with a small
metered flow of vaporized water was made to test and calibrate the water
metering equipment, that is, the cold trap and siphon pot for measuring
the total volume of HyO and HF leaving the salt and the water analyzer
developed at ORGDP . ? Operations were satisfactory, and preparations were
then made to process the flush salt for oxide removal.

The MSRE flush salt (66 mole % LiF, 34 mole % BeF,) had been cir-
culated in the reactor fuel system for 1000 hr to remove oxide film and
shake down the reactor system. The salt was tranferred, by gas pressure,
to the fuel storage tank in the fuel processing system and sparged with
a mixture of Hp and HF. Oxide removal was followed mainly by means of
the water analyzer. Difficulty was experienced with the HF flow control
causing occasional overloading of the cold trap and erratic siphon pot
discharges. After 32 hr of operation, the analyzer indicated a removal
of 115 ppm of oxide, and the removal rate had decreased by a factor of
15 to <1 ppm of oxide per hour. By this time the operating pressure in
the system had increased from 2.2 to 6 psig due to an incorrectly in-
stalled charcoal trap, and the increased pressure had caused partial
tranfer of the KOH scrub solution to the liquid waste tank. Sparging
with Hp and HF was therefore terminated. After the system was purged
with helium at a low rate, the defective trap was removed.

The salt was next sparged with Hsy, and 135 liters of HF was evolved
as measured by an HF monitor in the off-gas stream. This indicated that
24 ppm of oxide as Be(OH)g was left in the salt at the end of Hy-HF sparg-
ing. From equilibrium quotients at the salt temperature of 1200°F and
the peak HF concentration of 0.0l atm, an [OH™] to [0°~] ratio of 0.8
was calculated. The total oxide remaining in the salt after processing
was therefore 54 ppm.

Although the Hp:HF ratio was closer to 5:1 than the desired 10:1
ratio, there was no discernible increase in the chromium content of the
salt.

A more detailed description of these operations has been reported.3

References

 

1. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 20L.

 

2. W. 5. Pappas, Continuous Moisture Analyzer Monitors Hydrogen Fluoride
Off-Gases During Oxide Removal at MSRE, K-L-6059A (June 29, 1965).

 

3. R. B. Lindauer, internal memorandum (July 1965).
153

OAK RIDGE NATIONAL LABORATORY

 

JUNE 1, 1965

 

R. B. BRIGGS, DIRECTOR
P. R. KASTEN, DEPUTY DIRECTOR* RD
W. B. McDONALD, ASST. DIRECTOR** RD

D

 

 

MOLTEN-SALT REACTOR PROGRAM

 

 

 

 

MSBR STUDIES

 

 

 

 

ACepomArPOoEO I M

REACTOR DESIGN

BETTIS

BAUMAN
BETTIS*
BRAATZ*
CARTER**
CRISTY*
GRINDELL*
PICKEL*
ROBERTSON
SCOTT*

E. SPALLER*
H. WESTSIK*
V. WILSON*

J. YOST

W. TERRY

C. JONES**

W. KRICK*

DEePprr-mMm ¥

 

 

 

 

 

 

MSRE OPERATIONS

 

 

 

 

 

 

 

 

 

COMPONENT DEVELOPMENT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

INSTRUMENTATION AND CONTROLS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R. L. MOORE j&C
J. R. TALLACKSCN j&C
A. H. ANDERSON 1&C
S. J. BALL* 1&C
G. H. BURGER 1&C
D. G. DAVIS [1&C
E. N. FRAY 1&C
P. G. HERNDON 1&C
J. W. KREWSON 1&C
J. L. REDFORD** 1&C
T. M. CATE 1&C
B. J. JONES* 1&C
P. E. SMITH 1&C
C. E. STEVENSON 1&C
W. WEIS |&C
NUCLEAR ANALYSIS
B. E. PRINCE R
T. W. KERLIN* R
FUEL PROCESSING DEVELOPMENT
M. E. WHATLEY* CT
R. E. BLANCO* CT
H. E. GOELLER* CcT
W. L. CARTER** CT
G. I. CATHERS cT
J. R. HIGHTOWER CT
R. W. HORTON CT
R. B. LINDAUER CT
L. E. McNEESE CT
C. E. SCHILLING CT
C. D. SCOTT~ CT
J. BEAMS CT
C. J. SHIPMAN CT
W. G. SISSON* CT

 

mIm=E
oMM

rEm
= 0=

REACTOR CHEMISTRY

GRIMES*
BLANKENSHIP*
McDUFFIE*
BOHLMANN~*

MSRE ON-SITE CHEMISTRY

THOMA**
S. KIRSLIS
. D. NEUMANN

o em

IRRADIATION PROGRAM

. COMPERE
KELLY
SAVAGE

R. HART
W. MYERS

RC
RC
RC
RC

RC
RC
RC

RC
RC
RC

RC
RC

PHYSICAL AND INORGANIC CHEMISTRY

 

 

 

COENTTAEEP

METALLURGY

TABOADA~

H. COOK™**

W. DAVIS*

G. DONNELLY*
G. GILLILAND*
INOUYE™

R. KENNEDY*
R. MARTIN*

M. TOLSON*

R. WEIR™

J. L. GRIFFITH*
E. J. LAWRENCE*

M&C
M&C
M&C
M&C
M&C
M&C
M&C
M&C
M&C
M&C
M&C
M&C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

*
PAUL R. KASTEN* R P N. HAUBENREICH 2 Dl. zzloETWTAK E
W. B. McDONALD** R .
CAPITAL COST COORDINATION NUCL EAR AND MECHANICAL ANALYSIS FLOW MODELS
R R. C. OLSON” R C. H. GABBARD** R J. R. ENGEL R R. J. KEDL* R
R M. L. MYERS R C. H. GABBARD** R
GE&C J. J. GEIST***
CE&C A. HOUTZEEL***
cT H. B. PIPER R
GE&C J. A WATTS R
R
E
G &(é METALLURGY OPERATION
: O ADA Jec REMOTE MAINTENANCE
GE&C A-TAB R. H. GUYMON R
2 W. H. COOK* M&C J. H. CROWLEY R R. BLUMBERG R
2 C. R. KENNEDY* M&C P. H. HARLEY R CHEMISTRY J.R. SHUGART R
R G. M. TOLSON* M&C V.D.HOLT R
0 T. L. HUDSON R R. E. THOMA** RC
R A. |. KRAKOVIAK R
R H. R. PAYNE R
H. C. ROLLER R
R. STEFFY R
CHEMISTRY W. $ E\;mfi:{z E COMPONENTS AND LOOPS
L ANALYTICAL CHEMISTRY W. H. DUCKWORTH R
J. D. EMCH R MATERIALS A. N. SMITH R
J. C. WHITE* AC E. H. GUINN R R. B. GALL AHER R
T.G. HILL R W. H. COOK™ M&C M. RICHARDSON R
REACTOR CHEMISTRY C. C. HURTT R R. E. CARNES R
J. C. JORDAN R C. A. GIFFORD R
W. R. GRIMES* RC LA
H. F. McDUFFIE* RC L. E. PENTON R C. JONES** R
W. E. RAMSEY R
C. J. BARTON* RC R. SMITH, JR. R
F. F. BLANKENSHIP* RC J. L. STEPP R
S. CANTOR* RC J. L. UNDERWOOD R
S. S. KIRSLIS RC s R WEST 5 CHEMICAL PROCESSING
J. H. SHAFFER* RC
R ps J. E. WOLFE R R. B. LINDAUER** R
PUMP DESIGN AND DEVELOPMENT
AC ANALYTICAL CHEMISTRY DIVISION MAINTENANCE | .
CT CHEMICAL TECHNOLOGY DIVISION ’;' g ;‘:I'?HDELL g
, B. H. WEBSTER** R - G.
D DIRECTOR’S DIVISION N E GILLEN® R L. V. WILSON R
GE&C GENERAL ENGINEERING AND CONSTRUCTION DIVISION L. P. PUGH** R
18C INSTRUMENTATION AND CONTROLS DIVISION
M&C METALS AND CERAMICS DIVISION
P&E PLANT AND EQUIPMENT DIVISION
R REACTOR DIVISION
RC REACTOR CHEMISTRY DIVISION DESIGN LIAISON
* PART TIME ON MSRP
** DUAL CAPACITY C. K. McGLOTHLAN** R
#x* EURATOM EXCHANGE SCIENTIST

 

 

 

 

C. F. BAES, JR. RC
J. H. BURNS RC
S. CANTOR RC
J. H. SHAFFER RC
R. E. THOMA** RC
G. BRUNTON RC

F. A. DOSS RC

H. A. FRIEDMAN RC
G. M. HEBERT RC

K. A. ROMBERGER RC
D. R. SEARS RC
H. H. STONE RC

W. T. WARD RC
W. K. R, FINNELL RC

B. F. HITCH RC

W. P. TEICHERT RC

ANALYTICAL CHEMISTRY

L. T. CORBIN* AC
J. C. WHITE™ AC
R. F. APPLE AC

J. M. DALE AC
F. K. HEACKER AC
C. E. LAMB* AC
w. L. MADDOX* AC

 

 

 
H e
O NWH O

17-26.
27 .
28.
29.
30.
31.
32.
33.
34,
35.
36.
37.
38.
39.
40,
41 .
42 .
43.

45,
46 .
47 .
48.
49,
50.
51.
52.
53.
54.

oo~ o0ut W H

=

GgrEHgOOPHA W HEH OO EAE®

TN EsQgrHdgEudgsoEge R ESgE S STEHZEOOQOEHPPROEHH OEHON QHQ

DO d@m@HdRdoEHepQEHETQHHPEPHEHOPPHEPES QOETW

Adamson

. Alexander
Baes
Beall

. Bettis

. Billington
Blankenship
. Blizard
lumberg
Bauman

. Boch
Bohlmann
Borkowskil
Boyd

. Bredig

. Breeding
. Briggs
Bruce

. Burger
antor

. Cardwell
Carter
Compere
Conlin
Cook
Corbin

. Cristy
Crowley

. Culler

. Davis

. Davis

. DeVan

. Donnelly
Douglas

. Dunwoody
. fIngel
Iipler
hrgen
Fraas
Frye, Jr.
Gabbard
Gall

. Gallaher
. Gilliland
. Grimes

155

INTERNAL DISTRIBUTION

55.
56.
5.
58.
59.
60.
ol.
02 .
63.
64 .
65.
66.
o'7.
68.
69.
70.
71.
72-81.
82 .
83.
84 .
85.
86.
87.
88.
89.
20.
9.
2.
93.
X,
95.
9.
7.
98.
99.
100.
101,
102.
103.
104.
105.
106.
107.
108-110.

ORNL-3872
UC-80 — Reactor Technology
TID-4500 (46th ed.)

2Rz aoHdomnssssm=EaddHdoagrrHaoR =2 0gdg = >=$=Wj<ijFj?fi?fl?j?j?)rd?U?>

Ho=ooxHE o238 n@mDEDnae

GO HOQONPHE Y rOREHdOEOQHOWEepEHEHS HS Wg ey 0H

Grindell
Guymon
Harley
Harrill
Haubenreich
Herndon
Hibbs (Y-12)

. Hill
. Hise
. Hoffman
. Holt

Holz

. Hollaender

. ©. Householder
. L. Hudson

. Inouye

. Jordan

. Kasten

. Kedl

. Kelley

. Kelly

Kennedy

. Kerlin
. Krakoviak

Krewson

. Lamb

. Iarson (K-25)
. Lincoln

. Lindauver

. Livingston
. Lundin

. MacPherson
. Martin

. McDonald

. McDuffie

. McGlothlan
. Miller

. Mills

Mixon

. Moore

Morgan
Moyers
Nelson

. Osborn
. Parker
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134,
135.
136.
137.
138.

200-207.
208.
209,
210.
211.
212.
213.
214,
215.
216.

217-554,

C)fijk1ziw>fij>=C>E:E1tqfiitfl>=Eizizizififlziti?iF:21C7E1w1fi

156

. F. Parsly 139. C. D. Susano
. Patriarca 140. A. Taboada
. R. Payne 141. J. R. Tallackson
. Phillips 142, E. H. Taylor
B. Pike 143. R. E. Thoma
B. Piper 144. G. M. Tolson
E. Prince 145. D. B. Trauger
L. Redford 146. R. W. Tucker
Richardson 147. W. C. Ulrich
C. Robertson 148. D. C. Watkin
C. Roller 149. G. M. Watson
W. Rosenthal 150, B. H. Webster
C. Savage 151. A. M. Weinberg
W. Savage 152. dJ. R. Weir
W. Savolainen 153. J. H. Westsik
Scott 154. G. C. Williams
E. Seagren 155. J. C. White
H. Shaffer 156. L. V. Wilson
D. Shipley 157. K. J. Yost
J. Skinner 158, G. J. Young
M. Slauvghter 159. Bioclogy Library
N. Smith 160-161. Reactor Division Library
G. Smith 162-166. ORNL Y-12 Technical Library,
H. Snell Document Reference Section
F. Spencer 167-169. Central Research Library
opiewak 170-204. Laboratory Records Department
oteffy 205. Laboratory Records, ORNL RC

E. Stevenson
FXTERNATL DISTRIBUTION

. Cope, AEC, ORO
. Deering, AEC, ORO
. Garrison, AEC, Washington
. McNamee, Manager, Research Administration, UCC, New York
. Shaw, AEC, Washington
. E. Slnclalr, AEC, Washington
. L. Smalley, AEC, ORO
. A. Swartout, AEC, Washington
. J. Whitman, AEC, Washington
Division of Research and Development, AEC, ORO
Given distribution as shown in TID-4500 (46th ed.) under Reactor
Technology category (75 copies — CFSTI)

Szibd*lj

Sy EHE00 QO