ORNL-3282.txt

RUN MARTETTA £ gvs

T

3 445k 03L4525 O

 
 

 

 

 

 

 

 

 

 
ORNL-3282
UC-80 - Reactor Technology
TID-4500 (17th ed.)

Contract No. W-TLO5-eng-26

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT

FOR PERIOD ENDING FEBRUARY 28, 1962

R. B. Briggs, Program Director

Date Issued

 

OAK RTDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION

for the
U.5. ATOMIC ENERGY COMMISSION

VR

3 yysk 03k4525 0
SUMMARY

PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS

1. MSRE Design

Introduction.~-There was no significant change in the MSRE design
concept during the past six months; nearly all the design ideas described
in the preceding semiannual report remain unchanged. This report period
wag concerned with detailing the many parts of the system, the kind of
activity that precludes lengthy description here.

Ma jor Components.--Design of all major components (reactor, pump,
heat exchanger, radiator, and drain tanks) was completed, and shop draw~-
ings were issued. A model of the control-rod drive was built, and tests
are just beginning on the operation of the mechanism. Since some ques-
tions arose regarding the stability of unclad graphite in the core, a
design study was made of a core that incorporated clad graphite.

 

Secondary Components.--The designs of the thermal shield, radiator-
door drive mechanism, heat exchanger support, and pump support were com-
pleted in detail. Revisions were made to the layout of electrical,
instrument, and auxiliary gas and water connections to the reactor cell.
Pump-~lubrication package drawings were completed.

 

Pregssure-Suppression System for the Reactor Cell.--Further study of
the hazard associated with a salt spill in the reactor cell indicated the
desirability of providing a pressure-suppression system in order to ensure
that no accident could raise the cell pressure to an unsafe level. A pre-
liminary design was made for such a system.

Electrical Distribution System.--The system was thoroughly investi-
gated, and the complete electrical supply including emergency and stand-
by power was dlagrammed. Detailing of the electrical heater circuits for
bringing the system up to operating temperature is 50% complete.

 

Instrumentation and Controls Design.--Wide-range servo-positioned
figsion chambers will be used in the system in order to provide continuous
coverage of reactor operations from startings to full power. Compensated
ion charmbers will be used to provide an accurate, linear monitor of reactor
power and to provide a flux signal for the servo controller. The concep-
tual design of a servo controller, which will be used to control reactor
temperatures at low power level, was tested on the analog computer and
performed satisfactorily.

 
The preparation of instrument-application drawings is nearing com-
pletion. Nine drawings have been approved. The tabulation is almost 75%
complete. The design of electrical control circuitry is continuing. The
layout of the main control area was revised to provide more space in the
data room. Auxiliary instrument panels will be located in a transmitter
room adjacent to the reactor, in the service room at the north end of the
service tunnel, and in the cooling-water room. The main control-board
layout was approved after several revisions, and a l/h«scale model was
constructed.

Specifications are being prepared for the procurement of a digital
data system that has the capability of performing on-line computations,
detecting and alarming abnormal conditions, and logging process data auto-
matically and on demand. The MSRE instrumentation and control system is
being designed to ensure that failure of the data system will not compro-
mise the safety or operability of the reactor.

Requirements for process and personnel radiation monitors are being
investigated.

Specifications have been completed for five types of instrument com-
ponents, four of which have been approved. Preparation of seven specifi-
cations is nearing completion.

Instrumentation components are being procured. OSome equipment will
be obtained from Homogeneous Reactor Experiment No. 2.

2., Component. Development

Freeze-Flange Development.--Two sets of MSRE prototype flanges were
fabricated of INOR-8, and testing was started. After completion of thermal
distortion and seal leakage measurements, the flanges will be installed in
the thermal-cycle test facility and in the prototype pump test loop.

 

Experiments were conducted to determine the temperature distribution
and distortion characteristics of a 6~in. Inconel freeze flange, using
electrical-resistance heating in the bore to simulate operation with salt.
The freeze position of the molten salt was 4.8 in. from the bore when the
bore was at 1400°F, whereas the predicted position was 2.7 in. from the
bore. The maximum observed distortion varied from 0.024 in. at 1000°F to
0.063 in. at 1400°F. No permanent distortion was noted after 40O cycles to
1300°F.

Control-Rod Development.--A simple chain-driven control-rod mechanism,
using gas as a coolant, is being developed. Complete unit replacement is
planned for maintenance.

 

A 28-in.-long, U200-w section of removable heaters for 5 in. pipe
was built and tested. A heat loss of 450 w per linear foot of heater was
measured at operating conditions.
Heater Tests.~-~The reactor vessel heaters have operated satisfacto-
rily for 3750 hr, with TOO hr at 1425°F. However, the heater supports
and Inconel reflector buckled due to an inadequate support design. Cor-
rections will be made and the test continued.

The prototype cooling bayonets for removing fuel after-heat in the
drain tanks have operated for 3860 hr and 40 thermal shock cycles, with-
out significant damage. Each bayonet can remove 6.3 kw from 1300°F salt.

Sampler-Enricher System.--A few simplifying changes were incorpo-
rated in the design of the sampler-enricher system in order to reduce
maintenance problems and increase system reliability. A double-sealing
gate valve was opened and closed 100 times by hand, with leak rates of
less than 0.3 cc of helium per minute at operating conditions. A double-
sealing, buffered, sliding seal for oxygen exclusion during insertion and
removal of the sample transport container was mechanically cycled 1400
times, with a leak rate <1 cc/min with 15-psig helium buffer pressure.

 

MSRE Core Development.--The pressure vessel for the full-scale MSRE
core model was installed in its test loop, and measurements were made of
the flow distribution in the entrance volute. The results agreed with
previous measurements made on the l/5-scale model to within the precision
of the scaling factor.

 

Helium Purification.-~-Construction of a full-scale oxygen-removal
unit is about 50% complete, and an electrolytic type trace oxygen analyzer
was purchased for testing the oxygen-removal unit.

 

MSRE Engineering Test loop.--Interpretation of the effectiveness of
the oxide flushing runs in the Engineering Test Loop was complicated by
the presence of zirconium fluoride, which had been inadvertently included
in the salt mix. The slow-drain difficulity was the result of an addition
of BeO pellets which migrated to the drain line and formed a viscous plug,
later removed by raising the line temperature 60°F.

 

Oxide sludge was manually removed from the pump bowl after efforts
to dissolve it had failed. The treatment of the salt in the drain tank
with a mixture of hydrogen and HF removed the equivalent of 1025 ppm of
oxyzen without excessive corrosion of the container.

The 8-in,.-diam graphite container was fabricated from INOR-8 and
installed in the loop. A dry-box for loading and unloading graphite
samples was fabricated and is undergoing test.

MSRE Maintenance Development.--A program to produce an inventory of
tools and procedures to cover MSRE maintenance problems was initiated.
A hydraulic actuator system was successfully demonstrated for tightening
the clamps on the freeze flange. The problem of removal and reinstalla-
tion of flange ring gaskets was resolved, and the design of tools was
started. Tools were tested for aligning flanges, jacking pipe, and re-
moving the pump. A l~5/8~in.wdiam wide-angle periscope was superior to
similar smaller-dilameter devices previously tested.

 
Vi

Brazed-Joint Development.--The tapered braze Joint for l-l/Q-in,
sched-40 pipe was modified to use a 0.005~in.-thick sheet braze preform
which reduces to approximately 0.001 in. after the braze is formed.
Ultrasonic and metallographic inspection of completed prototype joints
indicated 81 to 86% bonding. A representative braze joint was held at
1250°F and intermittently exposed to reactor salt for T4 hr.

 

Mechanical-Joint Development.--Tools for remotely cutting, tapering,
and brazing the 1-1/2-in. pipe joint were received and testing started.

 

A mechanical joint, using trapped gas to keep salt out of the gasket
area during intermittent use, is being designed for use in crowded areas.

A moisture separator for use in a re-entrant-tube steam generator-
superheater was tested and found to carry over approximately W% water.
Further work is postponed.

Pump Development.--The design drawings for the MSRE fuel pump were
approved, and the thermal analysis of the fuel and coolant pump tanks was
completed. Water testing of the model of the cooling pump was completed.
Fabrication of the rotary element for the prototype fuel pump was com-
pleted, and fabrication of the pump tank was nearly completed.

 

Design drawings for the lubrication stands and drive motors were
submitted for review,

Additional INOR-8 castings of impellers and volutes for the fuel and
coolant pumps are being made, and dished heads for the pump tanks are
being inspected.

MBRE Instrument Development.--A series of tests 1s being made to
evaluate methods of attaching thermocouples. A test rig was assembled
for developmental testing of mechanical thermocouple attachments for use
on the radiator tubes in the MSRE.

 

Development of a thermocouple scanning system, using a mercury-jet
commitator, is continuing. A noise problem was eliminated by the use of
make~-before~break switch action.

Investigation of methods of economically obtaining signals which re-
liably indicate the operating status of freeze flanges and freeze valves
is continuing. A monitoring system, manufactured by the Electra Systems
Corporation, and a control relay, manufactured by Daystrom Incorporated,
are being evaluated.

Development of a continuous-level element for use in measurement of
molten salt levels is continuing. Several high-temperature differential
transformer designs have been investigated, two alternate level-element
designs were developed, and a level test facility incorporating the two
level element designs was fabricated. Testing of the prototype level
elements is underway.
vii

Testing of a prototype single-point level indicator was continued.
A prototype of the single-point level indicator to be used in the MSRE
is belng fabricated for test.

3. Reactor Engineering Analysis

Reactor Physics.--Further analysis of the MSRE temperature coeffi-
clents was done to obtain estimates of the effect of retaining fission-
product xenon and samarium in the core graphite, and of inserting a non-
1/v absorber such as rhodium. Since the principal contribution to the
temperature coefficient is the increase in leakage with increasing temper-
ature, and since the major effect of poison insertion is a change in the
temperature variation of thermal utilization, the calculated temperature
coefficients are not appreciably changed by poisoning in the core.

 

Results of simplified reactor-kinetics calculations indicate that a
peak pressure rige of 6l psi will result from a 0.3 step reactivity in-
sertion; a 1% step produces a 210-psi peak in the pressure rise, and a
0.6% step produces a 13-psi peak in the pressure rise. In the large step
additions the principal removal of reactivity results from heating the
fuel salt.

An IBM TO90 program, 2DGH, for the calculation of gamma-ray heat-
deposition rates was checked out and put into service. Results of a
heating survey in the top head of the MSRE vessel show a maximum heat
generation of 0.12 W'/cm3 at the lower end of the outlet pipe.

PART TI. MATERTALS STUDIES

4. Metallurgy

Dynamic Corrosion Studies.--Tests are in progress to determine the
effect of oxidizing impurities on the corrosion behavior of fused salts.
A loop containing NaF-ZrF, and contaminated with HF had uniform attack to
a depth of 1/2 mil after approximately 200 hr of operation. Examination
" an INOR-8 thermal convection loop containing molybdenum and graphite

 

of
inserts was completed and corrosion data are reported. The previously
reported embrittlement of molybdenum specimens appears to be associated
with surface contamination.

A series of tests were initiated to evaluate the effects of CFq4 on
MSRE core materials. No significant reaction was noted between CF, and
INOR-8 at the temperatures studied. An analysis was made of data obbtained
from corrosion tests containing inserts; the analysis indicates that
chromium, although a small fraction of the total metal loss, plays a major
role in the corrosion process.

Welding and Brazing Studies.--Improvements were made in the design of
the tube-to-tube-sheet joints for the MSRE heat exchanger. Weld "roll-
over" was minimized and the braze trepan was deepened in order to permit

 
viii

the preplacement of an adequate amount of brazing alloy. Small mock-up
test sections were brazed, and it appears that satisfactory brazes can be
obtained over a range of temperature rise from 75°C/hr to 225°C/hr. A
sleeve-type braze Jjoint was developed for MSRE use, and suitable brazing
conditions were determined which will permit good bonding along the Jjoint
length. Inspection methods are being developed for this Joint.

Initially, difficulties were experienced in qualifying welders for
INOR-8 work, but refinements in welding procedures have resulted in the
production of satisfactory welds. The room- and elevated-temperature
mechanical properties of dissimilar-metal welds containing INOR-8 were
determined. In general, these welds exhibited satisfactory integrity,
and failures occurred in the stainless steel, nickel, or Inconel, or at
the interface between those metals and the weld metal.

Mechanical Properties of INOR-8.--Mechanical properties of cast INOR-8
are being determined. Cast INOR-3 had shorter rupture life and higher
minimum creep rates than wrought INOR-8.

 

Evaluation of MSRE Graphite.--A sample of CGB-X graphite similar to
graphite to be used for the MSRE was evaluated, using MSRE evaluation tests.
Salt permeation tends to be restricted to shallow (less than 0.1 in.)
penetration below the surfaces at a pressure of 150 psi, about three times
that expected in the MSRE. There appears to be a slight heterogeneity in
the accessible pore spaces.

Tests indicated that a simple mercury impregnation test at room tem-
perature can ve a suitable quality-control test for relating standard
molten fluoride permeation into graphite.

Twenty-hour-long purges with the thermal decomposition products of
NH.F.HF in the temperature range from 1300 to 930°F removed oxygen con-
tamination from high- and moderately low-permeability grades of graphite
(AGOT and R-0025) to such an extent that there was no detectable uranium
oxide precipitation from a molten LiF-BeF--UF, mixture when it was exposed
to the purged graphites for 4000 hr at 1300°F. Lower purging temperatures
and smaller quantities of NH.F-HF also appeared promising for the removal
of oxygen from these grades of graphite.

A precursory test showed that refractory monoclinic ZrOs can be con-
verted to a less refractory fluoride compound by exposing it to NH4F.HF
at 1300°F.

5. In~Pile Tests

Interaction of Fissioning Fuel with Graphite: Test No. ORNL-MIR-
47-3.~-An experiment designed to determine whether fissioning fuel in con-
tact with graphite exhibits interfacial characteristics different from the
nonwetting behavior shown out of pile was operated for 1580 hr at a fuel
power density of 200 w/cc. The fuel, LiF-BeFo-7rF4-ThF4-UFy (69.5-23-5-
1-1.5 mole %), contained in encapsulated graphite boats, reached maximum
temperatures of about 900°C in undergoing 8.5% burnup of U?2% and still
remained nonwetting toward graphite. The graphite was virtually undamaged.

 
Supplemental observations, some of which are still in progress, re-
vealed that CF4 was produced, that the frozen fuel appeared black because
of discoloration by beta radiation, and that several other unusual or
puzzling phenomena had occurred. The persistence of CF4 was contrary to
thermodynamic equilibrium and may have been favored by the experimental
arrangement. The escape of CF4 in the offgas is potentially a serious
problem, mainly because the effect is the same as if a strong reducing
agent were acting on the fuel.

In-Pile Testing.--A description of the apparatus and the test con-
ditions for in-pile test No. ORNL-MTR-4T7-4 is given. The test is de-
signed to tell whether CF4 can exist in the cover gas over the fuel, when
the graphite 1s submerged in such a way that the CFs formed at the fuel-
graphite interface must pass through the fuel before escaping into the
cover gas. Also, the experiment was planned to further demonstrate the
compatibility of the fuel-graphite-INOR-8 system under thermal conditions
at least as severe as those expected during MSRE operation.

 

6. Chemistry

Phase-Equilibrium Studies.--Progress was made in the elucidation of
phase equillibrium relations for the system LiF-BeFp-ZrF4, which provides
an analogue of initial freezing behavior of the MSRE fuel. A partial
phase diagram for the system and a number of invariant points were estab-
lished. The ternary compound, 6LiF.BeFn+ZrF,, the first phase to separate
(at 441°C) on cooling the MSRE fuel, was subjected to crystallographic
study. Equilibrium behavior in an important composition section from
the five-component system which contains the MSRE fuel was studied; the
section shows the effects of diluting the fuel with the coolant and of
removing 2LiF.BeF, by distillation. References are given to other phase
equilibrium and crystallographic studies of fluoride systems and com-
pounds of interest.

Oxide Behavior in Fuels.~-~The purification of fluoride melts was
demonstrated by the removal of an estimated 1200 ppm of oxide contami-
nation from a charge of LiF-BeFo-ZrF4 that had been studied in the Engl-
neering Test Loop. The oxide was removed as water by treating the melt
at 565°C for 70 hr with gaseous HF containing 20% hydrogen. ILittle cor-
rosion was observed as a result of this treatment. ILaboratory studies
of the behavior of sulfate, a common contaminant of fuel raw materials,
showed that sulfate ion is not stable in fluoride melts at temperatures
between 500 and 800°C. The behavior of various inorganic sulfates in
LiF-BeFs melts was explored at 600°C.

 

Physical and Chemical Properties of Molten Salts.--The estimation
of the densities of molten fluorides to within 2% of reported values
was achieved by refinements of a method based on the additivity of molar
volumes. Densitles of solid, complex, metal fluorides were estimated to
within about 5% by the use of analogous assumptions; such estimates have
facilitated the cholce of the number of molecules per unit cell for com-
plex fluorides under study.

 
Cryoscopic and calorimetric studies of fluoride systems were extend-
ed to include freezing-point depressions in sodium fluoride containing
uni- and trivalent solutes and in mixtures of NaF-LiF, and enthalpy
changes from 8TL.0°C to 0°C for KF, LiF, and various mixtures in between.

Refined thermodynamic values for the equilibriwm, NiFo + Ho == Z2HF + Ni,
at elevated temperatures, were obtained and interpreted. The theory of
molten~-salt behavior was extended by studies using molten nitrate systems.

Graphite Compatibility.--The behavior of CF4 in contact with normal
and partially reduced MSRE fuel at 600°C was studied in static and in gas-
recirculation systems. FEvidence was obtained that CF4 can react with oxide
contaminants to produce COp. Based on indirect evidence, the solubility
of CF4 in MSRE fuel at 600°C cannot be greater than 1 x 1078 moles of CFgq
per cubic centimeter of melt per atmosphere.

 

Chemical Aspects of MSRE Safety.~--Chemical aspects of MSRE safety were
studied. The sudden injection of molten fuel into water gave no substan-
tial hydrolysis of the fuel, according to petrographic or x-ray diffraction
examination of the products of the reaction, but titration of the offgas
indicated a 2% yield of HF. The solubility of MORE fuel salt components in
water was studied from 25 to 90°C; the rate of uranium solubility and the
amount of uranium dissolved implied that neutron poisons should be provided
in any water which might come in contact with the fuel. To permit safety
calculations on criticality hnazards arising from segregation by partial
freezing of the fuel, a hypothetical crystallization path was defined, and
density of the concentrated fuel was estimated.

 

Fluoride Salt Production.--The time required for the purification of
fluoride salts in the Fluoride Production Facility was shortened by the
adoption of a combined HF-Hs treatment. Current modifications in the size
of salt-transfer containers will give a 50% increase in the production
rate. The addition of premelting furnaces is being studied as a possible
means to a further increase of 6L4% in the production rate.

 

Analytical Chemistry.--Analytical studies included (1) evaluation of
methods for the determination of oxygen in the nonradioactive MSRE fuel
and (2) a survey of methods applicable to the complete analysis of fuel
samples from the reactor during critical operation.

T. Fuel Processing

MSRE Flowsheet,~~Spent MSRE fuel will be fluorinated to recover
uranium. Methods for disposal of the excess fluorine are being investi-
gated.

Fluorine Disposal Tests, Using Charcoal.--Fluorine disposal by ve-
acting fluorine with charcoal was shown to be nonexplosive and greater
than 99.99% effective in seven runs. Solid and 1iguid products were
trapped from the offgas, which contained 52,6 mole % CFy4.
xi

CONTENTS

MMARY S 0 0 0 0 ¢ 000 02 OO O OO OO OO P O OO OO S E OO 0P ONE S eSS e SO eSO OO GG iii
PART I. MOSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS

1. MORE DESIGN 4t eececesoescoasovoseocssssossssscosossscssoosoasoss
INtroduction ccececesesescacecanscscosscsscsoscssonssnasese
Reactor Core and VesSsSel ..eeeecresscooscosccocnscococsocs
Primary Heal EXChanger ...ecececescccocococscsoscscsssone
Radiator coeseseecseescoersersesesceosasacsssaccescscacacss
Fuel-Salt Drain Tanks ..eeeeecscsccscccsscscoosscososons
Equipment Layoul cceeecececsecccesessscscccsscsssssasoas
Cover-Gas OYStemM seecsececssosesosesossssrsscscscsasossssne
Oystem Healers teeeeceesecsessscscsscssesosccosccosss cevene
Maintenance Design, Assenmbly Jigs, and Fixtures ........
Reactor Procurement and Installation ....ccceeccocccceces
1.10.1 Major Modifications to Building 7503 .veeecocens
1.10.2 Construction Outside Building T503 ceieeveosones
1.10.3 Planning and Scheduling of the MSRE

*

o
H\0 0= O\ =W O

O

-

OTIIT0O O DWW

Installation ..... ® 0 O 0 @& O ¢ O OO B PO O 0P8BS P9 O C 0O 0 S s 9
1.10.k Procurement of Materials ....ececececoscssceas .o 2
1.10.5 Procurement of Components eeeeceesecess cecssceocss 10

1.11 Reactor Instrumentation and Controls System .eececececess 10
1.11.1 Nuclear Control System .eeeeeeececsssesescsosccnas 10
Instrument-Application Diagrams ...eeeeecccceses 14
Electrical Control CircultYy .eeceececccecossnccos 15
Layout of Instrumentation and
Controls SYystem .eeeveeeoeecocosoesssoseacsscass 15
5 Control Panels seeeeesescecscsscsssoscscassoanes 17
6 Data Handling ceeeeeececeoeeccooscesccsscssoonss 18
.T Process and Personnel Radiation Monitoring ..... 18
8 Procurement StALUS .eeeeeeeeerreccosceccosccnsos 19

i
=l
Sy
=w

2, COMPONENT DEVELOPMENT +ecececocscessaccesoesosascanssosscscsscoaces 21

2.1  Freeze-Flange Development .cevieeeceerecccccsocnccsscnccs 21
2.1.1 MSRE 5~in. FlangesS .eeeeccececsaccessccass cecocse 21
2,1.2 Freeze-Flange-Seal Test Facility ..... sessccncse 2L
Control-Rod Development c.eeeecssescoscscssscsssssssssns Ry

N
w O

Heater TeSES ceeseeeoccceeasosesscssscesssssocsssscssesscs 26
2.3.1 Pipe Heaters .eceeeeeserscescsscccosscssscssscscans 26
2.3.2 Reactor Vessel Heabers ,...eeeeescsscccccccessos 29
2.1 Drain~Tank COOLETS eeceescecccerssensccsooccsssoscsescosss 29
2.5 Sampler-Enricher SySTeM ceceeesscessssseccsscossssonocoas 29

2,.5.1 General Concept of Sampler-Enricher .eeeeceeceses 29
285.2 Operationalvalve Q2 6 0 0 O 0O DD S O O8O O P 9SO e SN e O ’3—]‘
3.

s

<11l
.12

.13

oo

Xil

2.5.3 Removal Seal for Sample Container ......

2.5.4  Detail Design of Sampler-Enricher .......... ce
MSRE Core Development ...eeceeceeoccosecns ceceeesesoaans
2.6.1 Full-Scale Core Model ....eeeeeeveeecocoennnns ..
2.6.2 Core-Inlet Flow Distribubion ..eeveeeceeeeens .
Helium Purificalion ....eceereecerseossocccnseoscssansns
MSRE Engineering Test Loop (ETL) ........
2.8.1 Toop Operalions ..eeeveeeereeenoeeees R ..
2.8.2  Opecration of Freeze Valve .eeeeeeeeeeoocccconns
2.8.3 Oxide Removal v..eeeeeeeen. Ce et ceeee
2.8 0 HEF TreatmMent veeeeeeeeeeeseoeeeesananceeeaoneses
2.8.5 FTL Graphile Facility voeeeeeeeeeeeereeceecenns
MSRE Maintenance Development ....ceeevee ceeeccesrcsaan s
2.9.1 Placement and Removal of Freeze-~Flange Clamp ...
2.9.2 Flange Alignment and Pipe-Jacking Tools ........
2.9.3 Gasket-Replacement Procedures .....eeeeeeseen .
2.9.h  Miscellaneous Disconnects ..eeeeseces.. ceeen
2.9.5 Remote Viewing ..eeeeceeeceees .
2.9.6 Component Removal .......... Ceeecesseeseeeeees .
Brazed-Joint Development ....eveveeveeosoereoooncsns oo
2.10.1 Joint Desigll seeeeeseescsosoesosasesaanocsnasas
2.10.2 Brazed-Joint Testing ..eeeeeveeeceneconosas ceosoe
2.10.3 Remote Fabrication of Braze Joint ...ceeeeeeeen.
Mechanical-Joint Development ....ceeeeeecosss e esen ceee
Steam Generator ...ceeeeseess cesens S e essenussssasencases
Pump Development ..eeeeeeresersrscncessescsssssosscoanosns
2.13.] MORE Fuel Pump +..veveverececrereroscsoooonaosss
2.13.2 MSRE Coolant PulMD .eveveeeccecoosoceccooononccses
2.13.3 Advanced Molten-Salt PUmMpPS ceeeeeeoccenss .
MSRE Instrument Development ....eeeeeeeseesocosccccocnas
2.1.1 Thermocouple Attachments ..v.eeeeeeeeesoonenss ..
2.14.2 Temperature SCANNET «evveevereveeeeroreeones
2.14.3 Single-Point Temperature-Alarm System ..........
2.14.4 Pump-Bowl-Level Indicator ....... et eeacnann
2.14.5 Single-Point Level Indicator .e.eeo... N

REACTOR ENGINEERING ANALYSTS ..eveeieveescrescossvosonne

3.1 Reactor PhySicCs teeeeeeescccscesssesssessaosssnscsancssnns
3.1.1 Analysis of MSRE Temperature Coefficient .......
3.1.2 Reactor-Kinetics Studies ....eeeeecorercconcenes
2.1.3 Gamma-Heating SUTVEY ceeeeeersesveesosescoconcas
PART TIT. MATERTALS STUDIES
METALLURGY 4t eeoeoeeossaceecaosenaccnansnns ceerecans cecensnennn
401l Dynamic~Corrosion SLUALES weeeeeeeeeeeenorsenessanananns
4.1.1  Fluoride-Salt Contamination Studies ..... . .
4ol.2 Molybdenum~-Graphite Compatibility Tests ...... .
4.1.3 Corrosion Effects of Carbon Tetrafluoride ......
ol 4 Bxamination of Corrosion Inserts from TNOR-8&
Forced~Convection LOOPS veeeeseon. ceesesacaanne
4.2  Welding and Brazing Studies ve.eeeeeees. cheenen ceveseas
4.2.1  Heat Exchanger Fabricabion ...eeececee. ceraene .
4.2.2 RemMObLe BraZiNg tveeeeeeeeeeeeeeanneennaanns oo

o & o 0 0 0 0 0

31
32
32
52
32
33
34
35
35

37
37
41
41
4l
41
43
43
4
45

45
46
51
51
51
52
55
56
58
58
58
60
6l
66

68
68
68
68
70

2
72
72
4

7l

79
81
g1
8/,
xiii

4.2.3 Welding of INOR=8 sceveevevscoscoscssoscesacscas 86
L.2.4 Mechanical Properties of INOR-8-
Dissimilar-Metal WeldS seeeeescecescccsscoccns 88
4.3 Mechanical Properties of INOR-8 ..ceeeeeecoesconscnsaess 89
4.h Evaluation of MSRE Type Graphite .eieecececesscecsssccas 89
h.4.1 Comparison of the Permeation of Graphite
by Mercury and Molten Fluorides ....cceeceecee 92
h.h.2 Removal of Oxygen Contamination from
Graphite with Thermal Decomposition
Products of NH4AF 'HF .eieeecececcecscosooscnsnscs 94
b,k.3 Reaction of ZrOs with Thermal Decompo-
sition Products of NH4AFHF ...eeveeeeonces ceaes 94

5. IN"PILE TESTS S 8 6 0 0 6 0 6 6 6 0 8 6 O 8 00 P O 0 OO OO E OO SO L S OO e PO 0N 00N e 9’7
5.1 Interaction of Fissioning Fuel with
Graphite: Test No. ORNL-MIR-UT-3 .iveiveeeoerecesaeas 97

5.1.1 Description of Experiment c..eeeeecescsccesvence o7
5.1.2 Dismantling of In-Pile Assembly ....... cesesens . 99
5.1.3 Temperatures ....ce.. cevecersocs cesessesesesasss 100
5.1.4 Gas Analyses .e.ceeeeecees Ceesesecscececaseas ... 100
5.1.5 Test Effects on Graphite ...ieeveseeecescnss ceees 105
5.1.6 Analyses of Graphite .ceeeseeeeccsecosesscecoses 106
5.1.7 Test Effects on COUPONS +evvevsescasssssssscsess 106
5.1.8 Test Effects on FUELl vvveeeeereeceaannnns cesess. 106
5.1.9 ConclusSions .eeeeecceessessccsssssoscsscsssasssas 110
5.2 MSRE In-Pile Testing eveececceocssssscscessosssessssecess 110
6. CHEMISTRY +vvvcevocnsoncescacsossosaasssososcassssansos cerasees 114
6.1 Phase-Equilibrium Studies ....ccoeeveveeeee ceececenneeee. 114
6.1.1 The System LiF-BeFo-Z1F4 sevececcesceococenseess 114
6.1.2 The System LiF-BeFo-ZrFs-ThF4-UFs .vvvvveeanee.s 116
6.1.3 Phase Equilibrium Studies in Fluoride

OYSTEMS sveeeescossescscososesesoasscsssnsasans 110
6.2 Oxide Behavior in FUELS .eeveeececcssossccsosssssssssseas 117
6.2.1 Removal of Oxide from a Flush Salt ...eesveesess 117
6.2.2 The Behavior of Sulfates in Molten
FIUOTIAES +oveeoscecscosnsooscesovasosansenens 118
6.3 Physical and Chemical Properties of Moplten Salts ....... 121
6.k  Graphite Compatibility seeeceescecoerscacsssscsnsasseses 122
6.%.1  The Behavior of Carbon Tetrafluoride
in Molten FluoridesS ceevececscesosccoessnesess 122

6.5 Chemical Aspects of MSRE Safeby veeeeececcercccceneeasss 124
6.5.1 Physical Effects of Mixing Molten
Fuel and Water ..eecesececceecscoassossscaesoas 124
6.5.2 Solubility of Fuel-Salt Components
in Water s.vseececeeecrssscccosoosscsscceancaes L27
6.5.3 Solubility of MSRE Coolant in Water .eeceeeeeees 129
6.5.4 Partial Freezing of MSRE Fuel .veevevereen ceeees 130
6.6 Fluoride-S5alt Production ..eeceeecereeceecoeracesaeeaess 130
6.6.1 Production RateS .eveeeeeeeoceosscscocossoncosees 130
6.7 Analytical Chemistry coeeeececcessoessscoscoscssseacesss 131
6.7.1 Oxygen in Nonradioactive MSRE Fuel ...eeveeeeee.. 131
7.

Xiv

6.7.2 Adaptation of Analytical Methods
to Radioactive Fuel ...i.iceveeeencens

e ¢ ¢ &6 & 0 0 0 0

mL}?ROCESSD\]G @ & 5 & 6 0 0 6 0 0 00 & b SO SO S O 0L S e 0SB s 0000 » 8 6 06 ¢ 0 6 4 0 0 0

7.1
T.2

MSI{EFlO.‘JSheet ® & & & & & 0 8 0 O 5 00 " 8 P8OOSOt e e e0e e

Fluorine Disposal Tests, Using Charcoal .........

® & o & 0 0 0

132

134
134
134
PART I. MSRE DESIGN, COMPONENT DEVELOPMENT,
AND ENGINEERING ANALYSIS

1. MSRE DESIGN

1.1 INTRODUCTION

Accomplishments in the past six months were mainly in the field of
working drawings, and there were very few design studies since no change
was made 1in the system concept. A little more work than was anticipated
resulted from moving the pump off the top of the reactor. Design of a
satisfactory pump mount, which provided sufficient degrees of freedom for
the pump, took more effort than expected. This mount has now been detailed
and will be tested on a hot-pump~test stand.

Since no lump-sum bidder could be found for component fabrication,
these items were scheduled for local fabrication. This necesgitated the
making of shop drawings for components showing weld details, etc., which
otherwise would have been left to a vendor.

While access to the top of the reactor is relatively unencumbered,

the control-rod thimbles must bend slightly in order to clear the graphite-
sampling-port flange. This bend made 1t necessary to have some flexibil-
ity in the control-rod drive mechanism. Since cell atmosphere is used to
cool the control rod, some means of containment of fuel salt in case of a
control ~-rod~thimble rupture imposed certain restraints in the control rod
and its drive mechanism. It was decided to build and test a model before
issuing formal design drawings for the drive mechanism and associated cool-
ing and containment piping. The model has Jjust been completed, and testing
has started.

The auxiliary reactor-cell work is nearing completion. The thermal-
shield design is finished. Most of the support steel within the cell has
been detailed. Analyses of stresses in hold-down bolts and containment-
cell skirt cylinder were made. Pipe-support detalls were designed, but
the drawings are not yet finished. Penetrations, disconnect locations,
and assembly tool accessories have been established.

Circuit layouts for the electrical system are approximately 75% com-
plete, and no problem remains other than preparing the necessary drawings.
Layouts have been made for the drain-tank heaters. Power and control
points for the heater circuits have been designated on electrical one-

line diagrams.
Drain-tank-cell piping has been layed out and slightly simplified.
Drain-tank support structure with weigh cells has been completed. Work
has not yet started on the detailed layout of piping and support in the
coolant cell. However, the radiator with its enclosure and heater cir-
cults is completed.

Instrumentation and control design effort was concentrated on the
final phases of conceptual design and on specification writing, tabula-
tions, and similar procedures required to producc detailed design drawings.
Conceptual design was virtually completed, and, in some areas, component
procurement was initiated.

1.2 REACTOR CORE AND VESSEL

The reactor core and vessel underwent no design changes. For infor-
mation, the cutaway drawing shown in the last report is repeated in
Fig. 1.1.

UNCLASSIFIED
ORNL-LR-DWG 510978

FLEXIBLE CONDUIT TO

CONTROL ROD DRIV
SAMPLE ACCESS PORT" i ES

  
    
  
 
    
 

——

2 ?///v"COOLING AIR LINES

CORE
CENTERING GRID

FLOW DISTRIBUTOR

%
GRAPHITE-MODERATOR
STRINGER —— |

FUEL INLET
REACTGR CORE CAN

REACTOR VESSEL—

   

 

 

   
  

e
i i B

il

 

'v‘\
I
I

 

 

ANTI-SWIRL VANES— <o i \(/
S MODERATOR

g

VESSEL DRAIN LINE e SUPPORT GRID

Fig. 1.1. Cutaway Drawing of MSRE Core and Core Vessel.
Although no changes were made in the reactor, much design effort was
spent in detailing welds, Jjoints, etc., and in defining shop procedures,
because of the necessity for local fabrication.

The drawing (Fig. 1.1) does not show the method of supporting the
reactor vessel. OSince this is the anchor point in the fuel system, the
vessel 1s statlionary and is hung from the top of the thermal shield.
Hanging is effected by twelve 1-1/Lk-in.-diam rods which engage lugs by
means of a cylindrical pin passing through the ears of the lugs and the
suspension bolt. The bolts are of the turnbuckle type, which permits
leveling of the reactor vessel. The lugs are welded to the vessel just
above the fuel-inlet volute.

All design drawings on the core are complete, and fabrication has
begun in the Y-12 shops.

Although fabrication of the reactor vessel is proceeding, results of
a recent in-pile test introduced a new uncertainty into operation of a
reactor with an unclad~-graphite core. Some CFgq was found in the gas space
in capsules which had been irradiated with MSRE fuel in contact with
graphite at temperatures and power densities considerably greater than
those of the MSRE. This finding is not yet fully understood or evaluated,
and 1t will be several months before data can be obtained from tests under
MSRE conditions. On the unlikely chance that additional information would
show the MoRE to be inoperable with unclad graphite, a conceptual design
was prepared for a graphite core which could be installed in the present
design of the reactor vessel. A calandria type of core was designed. This
core 1s made with hexagonal graphite blocks stacked around TNOR-8 tubes
3 in. in diameter, 0.065-in. wall. The tubes are welded to a stationary
header on the bottom and to a flexible header on top, both made of INOR-O.
A 10-in.-dia tube provides an open volume in the center of the core. The
present control-rod thimbles and clad~-graphite or unclad-graphite samples
of the same design as used in the present core can be inserted in this
opening. With this alternative design, the fuel is in turbulent flow
in the cylindrical fuel passages in the core, and the temperatures of
the fuel and graphite are slightly higher than with the unclad core but
are still well within an acceptable range. The critical mass is about
twice that of the unclad core. Enough calculations have been made to
ensure that the alternative design will be acceptable. Detailed calcu-
lations are in progress.

1.5 PRIMARY HEAT EXCHANGER

Shop drawings were released for fabrication of the heat exchanger.
A detailed stress analysis of the exchanger was completed.

This exchanger is of coanventional tube-and-shell design, with the
fuel flowing in the shell side. The inside diameter of the cylindrical
shell is 15-3/k in. The shell has an overall length of 8 ft, and the
active heat transfer length is 6 ft. The tubes are 1.2 in. in outer
diameter, with a 0.042-in.-thick wall. There are 163 tubes arranged in
a U-bend configuration, each end terminating in opposite halves of a
tube sheet.

Fuel 1s directed in the shell side by six baffle plates. The volume
of the holdup in the heat exchanger is 6 ft>. The volume of the coolant
holdup in the heat exchanger is 5.7 ft°. 'The heat exchanger when installed
and full of fuel and coolant weighs 3500 1b.

Heat exchanger mounting was designed to allow for movement of the
exchanger with load and with thermal expansion of the system.

1.4 RADIATOR

Work on the radiator, =2xcept for thermocouple attachments, has been
finished for several months. The radiastor enclosure and door drive mech-
anism required additional detailing. Attachment of heaters inside the
radiator enclosure required many drawings. This was true also of the
door drive mechanism. The entire radiator package is now complete, and
some fabrication on the support structure has already taken place.

Thermocouples were added to each of the 120 radiator tubes near the
cold end of the radiator in order to monitor as accurately as possible the
low salt temperature in order to prevent freezing in the radiator tubes.

1.5 FPUEL-SALT DRAIN 'TANKS

Some changes were made in the drain tanks. The volume of the tanks
was too small to allow ample margin in the freeboard in case unforseen
excessive temperatures should obtain. Therefore the diameter of the tanks
was increased from 48 to 50 in. to provide ample margin of safety in free-
board volume for every reasonable eventuality.

One other change in the drain-tank design involved the addition
of a penetration in the tank dome for the insertion of ligquid-~level
probes. With these changes, the shop drawings were completed and issued
for construction.

The support structure for the drain tanks, including the weigh-
cell -mount detalls, was completed.

1.6 EQUIPMENT LAYOUT

The concept of employing a batching operation in the Fluoride Vola-
tility Pilot Plant for fuel processing was abandoned in favor of eventual
installation of facilities for on-site fluorination of the fuel in the
spent-fuel storage tank. This will result in the elimination of the "fuel
transfer cell" from the layout plans. A waste tank will be required for
storing carrier salt after removal of uranium.
Layout within the reactor cell is complete but will be undergoing
constant modification to a slight degree as more complete information on
maintenance tools and practices becomes available. An overflow pipe was
added from the pump bowl to the drain line downstream of the freeze valve
to eliminate any possible danger of overfilling the pump gas space because
of some inadvertent or unforseen rise in fuel temperature.

Layout in the drain-tank cell is in the same status as that within
the reactor cell. The salt piping here has been completely redone,
because the first layout was thought to be intolerably crowded and conplex.

All auxiliary systems have now received some attention. The punp-
lubrication layout 1s complete. The component cooling system, water
cooling system, and electrical service to the cell are better than 90%
complete.,

Work has barely started on support details for the coolant piping.

A vapor-suppression system was added as part of the containment
(Fig 1.2). The large reactor-cell ventilation line has a tee which takes
off from the cell side of the two large valves. This line goes to a
water-filled tank so that discharge gases from the line pass through a
large volume of water. The water-filled tank vents into another closed
volume in the form of a shielded tank for containing noncondensable gas
which escapes from the first tank. Temperatures, pressures, required

UNCLASSIFIED
ORNL-L.R-DWG 67462R

40 ft FROM REACTOR BLDG.

 

 

 

 

 

. ELEV. 858 ft
2-in-DIA VENT LINE -
TO FILTERS AND STACK /
SPECIAL EQUIPMENT ROOM RN
IN REACTOR BLDG.-
FLOOR ELEV. GROUND ELEV, FILL
852 fi 851 f1-6in.

 

 

A/

A

 

 

 

 

 

 

 

 

 

 

 

 

 

  
  

    

 

 

  

ELEV. 848 1 1] 5 S
\10—in;~DIA RELIEF LINE 10-1 DIA x 54—t LONG
3900 13
BURSTING DISK
R f foem ~ AOO ]
4-in. LINE
ELEV. 836 fi
30-in-DIA REACTOR CELL VENTILATION DUCT
| T I ELEV, 830 ft
10~ft DIA x 23-ft HIGH N |~
1800 ft3-1200 13 OF WATER%; ) 1l
i ELEV. 824 fi

Fig. 1.2. Diagram of MSRE Pressure-Suppression System.
holdup capacity, and shielding requirements were calculated, and a pre-
limipnary layout was made showing arrangement, pipe and tank sizes and
locations, and shielding.

1.7 COVER-GAS SYSTEM

Work was continued on detailing of system piping and components.
Design memorands were issued on the oxygen-removal units, the helium
surge tank, and the secondary containment and shielding for the off-gas
line and charcoal adsorber pit. The cover-gas system is about 00% com-
plete. Routing of pipe between major components constitutes the major
remaining design.

1.8 SYSTEM HEATERS

Heaters for the coolant circuit were layed out for the entire pliping
system. These will be conventional clamshell-type ceramic heaters. They
are not required to be remotely removable; so conventional-type insulation
will be installed over the heaters on the pipe.

For pipes requiring remote maintenance, that is, all the salt piping
in the reactor and drain-tank cells, some kind of removable heating and
insulating "package'" must be made. Design has taken the empirical approach.
A full report on pipe-heater design is given in the Chap. 2. Design of
the actual pipe-heater module to be used in the MSRE has Jjust started.
Drawings will be based entirely on the mockup unit.

Heaters for the reactor vessel also followed the empirical approach,
and design drawings are being made of a thoroughly tested prototype heater.

Heaters have been decsigned for the dralin tanks. These consist of
standard heater units arranged in sectors which, when combined around a
tank, effectively place the tank within an electrically heated furnace.

Since the pump, with its rather long inlet pipe, constitutes an
oddly shaped unit, a different heating method has been employed. Here an
insulated furnace is constructed around the pump, enclosing the inlet
pipe. Into the top of this enclosure and around the periphery of the
pump bowl, arrangements have been made to insert commercial-type bayonet
heaters. This makes the pump-heating plan similar to the one used for
the reactor vessel, except that the actual heater units will be of com-
nercial design and will be purchased items.

All heating designs for the salt systems have now been determined;
drawings are in various stages of completion, but none are yet ready for
issue.
1.9 MAINTENANCE DESTGN, ASSEMBLY JIGS, AND FIXTURES

The design of the Jjigs and fixtures Tor the assembly of the initial
and replacement components was halted, awaiting firmer system design. It
has been restarted and 1s proceeding on the basis of using optical tooling
methods for the precision location of mating parts. Optical tooling will
be used to make in-cell measurements of mating points at the time that
component replacement becomes necesgary. This procedure will reveal any
deformation that may have occurred from operation.

1.10 REACTOR PROCUREMENT AND INSTALLATTON

1.10.1 Major Modifications to Building 7503%

The prime contract to make major modifications to Building 7503
wa.s awarded to the Kaminer Construction Corporation of Atlanta, Georgia.
The major portions of this work include modifications to the 24-ft-diam
containment vessel, modification of the drain-tank cell, modification
to the coolant-system cell, installation of the building ventilation
system, and erection of the secondary contalinment walls inside the high-
bay area.

Construction work started in November 1961 and is scheduled for com-
pletion about October 15, 1962. To date, the following work has been
accomplished by the contractor:

1. Excavation of the drain-tank cell was completed. The structural
steel hold-down beams, both vertical and horizontal, were welded in place,
and a concrete leveling pad was poured on the floor of the drain tank.
The reinforcing steel is being set preparatory to pouring the concrete
floor of the tank.

2. Demolition of the penthouse wall, above the coolant-system cell,
wa.s completed; the structural steel was set in place and the forms were
erected Tor pouring the new concrete wall for the penthouse.

5. Off-gite Tabrication of the stainless steel grid and liner for the
drain-tank cell is almost complete. This will be installed in March.

4. Off-site fabrication of the T-ft extension of the 24-ft-diam
containment vessel is in progress, as are the stainless steel diaphragm
and the shielding support beams. On-site work on this containment vessel
is scheduled to start in March.

5. 0Off~site fabrication of the carbon-steel structure and walls for
the secondary containment is in progress.

Figure 1.5 shows construction progress in the drain-tank cell.
UNCLASSIFIED
PHOTO 56727

i

 

 

 

 

 

 

 

in the Drain-Tank Cell

Down Beams

lold-

ion of |

Installat

3.

1.

g

F
1.10.2 Construction Outside Building T750%

Design drawings and specifications for the construction work exterior
to Building 7505 are completed, and the work will be advertised for bids
in March. The major portions of this work are the off-gas filter pit, the
erection of a 100-ft-high carbon-steel stack, and the installation of a
cooling tower. The work is expected to be done concurrently with the
existing work under contract with Kaminer Construction Corporation and is
expected to be completed at about the same time.

1.10.5 Planning and Scheduling of the MSRE Installation

A detalled review has been made of the activities required for pro-
curing and fabricating components for the MSRE and for installing the
reactor system in Building 7505. The information was used to develop a
critical~path diagram and schedule for the MSRE work. The information
was coded for the IBM 7090 computer, and preliminary results indicate a
completion date of July 1963%. A more detailed review is being made of
the critical and near-critical activities to verify this estimated com-
pletion date and to determine if an earlier date is possible.

The schedule is based on the following assumptions:

1. That the lump-sum contract work will be completed and the cells will
be available for equipment installation by November 1, 1962.

2. That about 80% of the reactor and drain-tank process and service
piping will be prefabricated between July 1, 1962, and December 30, 1962.
5. That about 50% of the process piping, service pliping, power and
thermocouple cable will be installed in the reactor and drain-tank cells
prior to the installation of the process equipment.

The reactor components are to be preassembled on a jig outside the
reactor cell. In order to minimize the effects on the schedule of this
preassembly, as well as the effect of any unanticipated delay in the
completion of the building-modification work, the process and service
lines will be prefabricated as indicated.

1.10.4 Procurement of Materials

Delivery of INOR-8 plate, rod, and weld rod started in November 1961;
by February 15, 1962, about 80% of the material had been received. Approx-
imately 111,000 1b of this material was manufactured by Haynes Stellite
Company at an average price of $%.38 per pound.

Contracts were awarded in September 1961 to International Nickel
Company, Michigan Seamless Pipe and Tube Company, and Wall Tube Company
for manufacture of 18,000 linear feet of INOR-8 seamless pipe and tubing.
This material weighs approximately 18,000 1b and is being manufactured
for an average price ol about $18 per pound. The pipe and tubing are
scheduled for delivery in May 1962.
10

Taylor Forge Company was awarded a contract in September 1961 for
the manufacture of 420 INOR-8 pipe and tubing fittings for an average
price of $260 each. Delivery of these fittings is scheduled to begin
in March 1962, with completion by May 1962.

The National Carbon Company was awarded a contract in September
1961 for the manufacture of approximately 9000 1b of moderator graphite,
completely machined to specifications and tolerances, for an average
price of $21 per pound. It is scheduled for delivery in July 1962.

Contracts have also been awarded for the manufacture of 150 flexible
Inconel metal hoses for the fuel drain tanks.

1.10.5 Procurement of Components

Since efforts to obtain lump-sum bids for fabricating MSRE compo-
nents were unsuccessful, because of industry's lack of familiarity with
TNOR-8, the fabrication of the major MSRE components (reactor, radiator
and enclosure, heat exchanger, fill and drain tanks, and pump parts) is
being performed in the Y-12 Machine Shops, where experience has been

years.

Fabrication 1s started in the shop as material becomes available
and as fabrication drawings, prepared for outside vendors, are modified
for use in the Y-12 Machine Shop. Work was started in January on the
radiator enclosure and in February on the reactor. By February 15, about
18 men were working on these two jobs. This manpower load will increase
to a peak of about 125. Completion of the fabrication of all the major
reacltor components is scheduled for October 1902.

Subcontracts for fabricating INOR-8 have been awarded to the Taylor
Forge Company for forging the heat-exchanger tube sheet; to the UCNC
Paducah Machine Shop for forming the reactor inlet volute, radiator
headers, and dished heads for the heat exchanger, radiator, and reactor
vessel ; and to the Lukens Steel Company and Phoenix Steel Company to form
dished heads for the various fill and drain tanks and pumps.

The dished heads for the radiator, heat exchanger, and pumps have

been delilvered. Scheduled delivery of the other TNOR-8 fabricated items
is for March and April 1902.

1.11  REACTOR INSTRUMENTATION AND CONTROLS SYSTEM

1.11.1 Nuclear Control System

The nuclear instrumentation and controls system is block-diagrammed
in Fig. 1.k, The wide-range servo-positioned fission chamber channels?
are expected to provide continuous coverage of the full range of reactor
 

11

 

    

 

 

 

 

 

 

     

 

       
     

 

   

 

 

     
   

 

 

 

 

 

 

 

 
 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

8 UNCLASSIFIED
{ —_— Afow  Afreactor ORNL-LR-DWG 67585R
| RANGE i FLOW "5 ¢ 7 8
i CHANGE : % ?1 2 o i
¥
COMPENSATED 1 \
{ON CHAMBER ! POWER SUPPLY \LOG» LOG: TWO PEN
Q-1045 | a99s / RECORDER
MICRO-MICRO : LEVEL | RANGE ¢ SERVO POWER
( ™ AMMETER B CONTROL AMP
MEAN COMPUTER
TEMPERATURE
L05; SET POINT
E 17
LINEAR Z
POWER
RANGE S F Y S
| CHANGE > ‘
COMPENSATER ALTERNATE CONNECTIONS SO THAT [ FLUX SET POINT
1ON CHAMBER SUFF EITHER CHANNEL MAY BE USED TO ‘
&-1045 POW;}';%U LY ; PROVIDE SERVO INPUT
{ MICRO-MICRO i . LIiIPNTIE'R
| AMMETER ’
L ~2
LOWER
1 'd LIMIT
| | |
LIMIT ! |
SW.
UPPER LOWER o —_— LOG
LIMIT LIMIT COUNT RATE | POWER
| SET POINT | RECORDER
[ R POWER SERVO | LOG POWER
_— O — | T
[~ AMP, AMP, . | - oo}
i : . ONEAR PERIOD COMPUTER|  \'J
| i manuaL | COUNT RATE ! PERIOD
{ POSITIONING | METER CORDER
-t - ! LINEAAR LOG
1 PREAMP | PULiED MP. COUNT RATE
0-2263 | METER
FISSION CHAMBER ] DISCRIMINATOR ‘
Q-2217 PREAMP - UPPER
POWER SUPPLY SCALER CimIT
Q-2263
WIDE
RANGE FISSION CHAMBER AND PREAMP. Q
SERVO < FLEXIBLE ASSEMBLY RC 2-12-3
FISSION LOWER
CHAMBERS r 7 LIMIT
? l
POSITION ' :
I ’
UPPER i 1 COUNT RATE :
LIMIT { i f SET POINT ™
i POWER SERVO LOG POWER
| AMP, amp. | 1 LOGE
i LINEAR PERIOD COMPUTER
i COUNT RATE
METER
- * LINEAR
i} | PULSE amp l METER
§ A
FISSION CHAMBER PREAME { DISCRIMINATOR
Qe POWER SUPPLY SCALER
N Q-2263
.
ANNUNCIATOR l +
T \ | .~ ALTERNATE CONNECTIONS
{— ‘ ! LOG] DATA LOGGING POINT
— i /
NEUTRON ——
SENSITIVE AMPLIFIER

LOG

 

.

—

CONTINUITY
MONITOR

\Efi%fi\

—

AMP.
FS

ION
CHAMBERS ’\ ANN.
%;/

REACTOR QUTLET 1 2aN

TEMPERATURE ———t=di

 

FUEL SALT FLOW RATE

OTHER INPUTS J
{FLOW, RADIATION,
LEVEL, ETC.)

TRIPS

LOGIC

 

 

 

   

 

 

 

 

 

 

Fig. 1.4, Block Diagram

 

 

 

e t SHIM ROD
? U NO.3

 

MANUAL
SHIM
CONTROL
RELEASE
ALL
RODS

of MSRE Nuclear instrumentation and Control System.

SHIM ROD
NO. 2

&)

TORQUE-SYNCHRO POSITION RECE!VER

AMPLIFIER WITH POWER SUPPLY

INDICATING LAMP

| TORQUE-SYNCHRO POSITION TRANSMITTER

CLUTCH (ELECTRICALLY OPERATED}

ELECTRICAL CONNECTION
————— ALTERNATE ELECTRICAL CONNECTIONS
————  MECHANICAL POWER CONNECTIONS
12

operation, from startup to full power. The linear channels, with com-
pensated chambers as the signal source, are the most accurate nuclear
information channels and are used to monitor reactor power and o provide
the flux signal for the servo controller. These channels have a span of
four decades; and, by combining range switching and chanmber relocation,
they are useful over a wide range. 'The limit channels will be used to
provide a "reverse' mode flux signal which will call for all the rods +to
be Inserted at an average rate of 0.040% Ak/k per second.

Figure 1.5 diagrams the reactor power ranges of the above nuclear
instrument. During initial critical periods, two temporary counting

UNCLASSIFIED
ORNL-LR-DWG 68575

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

FISSION-CHAMBER
CHANNELS
SCALER
LOG CR
LINEAR NEUTRON LOG POWER
FLOW CHANNELS INVERSE POWER LIMIT CHANNELS
A} (SEE NOTE 1) (SEE NOTE 2) (SEE NOTE 5)
10° S - S
SEE NOTE 4
ol %-—m—-<w~4—~w~ ———
5 _
10" -0 Mw -+ —fitw4~ ——— e
1008 b —aMw - — L Lo |
. MBER
10— LT CATION - —
E 404 e \l, [ !¥ o -
5
2 03 |
= m>~1w~\ T S —
z
© 1t — L e e e ]
101 | — e o 4
mofi4w\\{r ———
10‘1\\;i; e ] L - b S U
40”2._.‘fi [ (O S A
10_3L~—1mwf‘]\4-~—-' e

 

 

1. Ceonsists of compensated chamber to micromicroammeter. Used for servo flux input signal as well as
recorder input. Lower threshold, possibly one decade below that shown, if all conditions optimized.
Range switching not shown,

2. These are servo-operated fission channels and can be used to provide all readouts shown.

3. Under ideal conditions {low elecirical noise and relatively clean fuel) the servo will control in the
1C-Mw region.

4. Range of servo-operated fission channel dependent on movement in attenuating medium (H 20), pro-
vided that no gross flux distortion exists in medium traversed by chamber.

5. The two-decade span may be located over a wide range,

Fig. 1.5. Ranges of MSRE Power and Nuclear Instruments.
13

channels with BFs chambers located in the composite shield will be used.
Except for the BFy counters, all nuclear instruments will be located in
the large 36-in.-diam water-filled penetration, inclined approximately

L5° (ref 2).

The servo control system will be derived from a conceptual design
(Fig. 1.6) proposed by E. R. Mann. Servo control of the M3RE is required,
because its extremely conservative operation yields a very high value for
the ratio of heat capacity to power density. Inherent self-control by
temperature coefficient is therefore sluggish and productive of power
oscillations which, while not dangerous, are not acceptable from an oper-
ational standpoint. The need for servo control will diminish rapidly if
reactor power is allowed to exceed the present 10-Mw limitation.

UNCLASSIFIED
ORNL - LR-DWG 68576

FLOW

 

   
 

 
   

POWER
REMOVED-

 

 

 

 

 

 

 

 

 

 

 

e e o e o e e e e e e e e — — —— —— - RE ACTOR
POWER-
¢
T T T T T T T T T T MEAN SERVO
6; — "] | TEMPERAT URE MOTOR
| \ lERROR SIGNAL cowen
| €
' "+ B, AMPLIFIER
S e e 1y |
‘ |
l | "
l | ~ Ll
| REGULATING | |
| SET POINT, 8gp | ROD } 'I
| !
| |

€=4 [qb~ F(6,—6)~¢ (/)] -8 [BSP - 2]

Fig. 1.6. Block Diagram of MSRE Reactor Controller,

Functionally, the servo controller simply augments the temperature
coefficient of the reactor; in the power range it is not used to control
reactor power on operator demand.

The conceptual design was tested on the analog computer® and per-
formed satisfactorily.

The translation of the controller from conceptual design to hard-
ware is in process and consists of substituting and packaging suitable
industrial ~grade components for computer-operational amplifiers and the
14

addition of a power amplifier to match the shim-rod motor. The control
package will be thoroughly tested and will be used in conjunction with
the analog simulator for operator training.

1.11.2 Instrument-Application Diagrams

During the past report period, the 15 instrument-application flow
diagrams listed in Table 1.l were revised and issued for comment. In
general, the result of the revisions was a reduction of the number of
instruments and thermocouples required. Nine drawings were approved
for construction. It is estimated that this basic part of the instru-
mentation effort is 90% complete.

Table 1.1. Instrument-Application Flow Diagrams

 

 

No. of Flow Diagram Application

P-AA-B-40500" Fuel-salt clreuit
40501 Coolant-salt system
Lo502™ Fuel, flush, and drain-tank fault system
MO50§a Cover-gas system
Lo50L" fuel-salt-pump lubricating-oil system
MO505a Fuel sampler-enricher system
40506" Liquid-waste system
40508 Coolant-salt-pump lubricating-oil system
MOSOQa Water system
MO5ZLOa Off-gas system
40513 Fuel, fill and transfer system
Los1 kL Instrument~air and service-air systems
h0515 Containment air

 

aProjectmapproved drawings.

A tabulation of all instruments shown on the flow diagrams, giving
identifying numbers, location, function in process, and a brief descrip-
tion, was also revised in accordance with the latest revision of the
application diagrams. This tabulation is nearly 75% complete.
15

Although 11 thermocouple-location drawings were approved for con-
struction, they were revised because of the reduction in the number of
thermocouples required., A list of all thermocouples, showing the type
and location of the readout device for each one, was issued for comment.

1.11.5 Electrical Control Circuitry

Control-circuit design is continuing. Elementary schematic drawings
showing basic electrical control and the safety circuits were nade for the
sampler enricher, fuel- and coolant-pump lubricating-oil system, water sys-
tem, contalnment-air system, and the coolant-salt radiator,

1.11.4% ILayout of Instrumentation and Controls System

The layout of the instrumentation and controls system i1s continuing
according to the general scheme outlined in the last report.? The lLayout
will permit all routine operations to be performed in the main control-
roonm area. An effort is being made to centralize instrumentation, and
Tfield panels will be used only where the nature of the operation dictates
that controls and instruments be located in the field. The arrangement
of the main control area was revised as shown in Fig. 1.7. The data-
room area is larger by 225 square feet, allowing more space for the data-
handling system. The main control and auxiliary control area were shifted
to the south. The instrument shop was moved from the location shown pre-
viously® to the southeast corner of the building at the same (852 ft)
level.

A second major control area is the transmitter room, which 1is
located on the 84O-ft level adjacent to the reactor. Instruments for
the leak-detector system, weligh system, and the gas-purge systems for
the fuel and coolant circuits will be mounted on auxiliary panels in
this area. Solenoid valves through which air is supplied to salt-system
gas~control valves are to be located in this area, along with ampliflers
and power supplies for the process-~variable transmitters.

Two instrument panels are located in the service room at the north
end. of the service tunnel. These panels house instrumentation and con-
trols equipment associated with the lubricating-oil systems for the pumps
that circulate the fuel and coolant. Both pumps are located in the serv-
ice tunnel. Radiation-monitoring instruments serving the service tunnel
will also be mounted on these panels.

One more instrument panel will be located in the water room near
the radiator blowers at the southwest corner of Building 750%. Instru-
ments connected to the cooling-water system are to be located on these
panels.
UNCL ASSIFIED
ORNL-LR-UWG 68577

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OPERATIONAL AREA PLAN VIEW

ELEVATION: 852 ft

Fig. 1.7. Operctional Area of Main Contro! Room and Auxiliary Control

Room.

 

 

 

f 2 3 4 5 6 7 (8 9
- I { R Y — 21 ft — o 24t e e 24§t AT f1 3N, e {7 ft Gin - AT £ B
! ' ! (T
e b e I *‘f'f”ii‘?—lr’)y“ c = — - — -~
‘ ; "~
1 ; EemE e Em T s *‘r'frfffm‘
| | | T | iINSTRUMENT SHOP
= i | | AND STORAGE
= o }L ‘1 ~ /‘/’ ] 1 : . »
1 s = — — | ‘ | AREA: 441 fi
1 = | | | \
i “GLASS WALL L ‘ |
| ; | |
r Y - X { I
2 RMOCOUPLE | :r ’ I ! I I
- CA. B0 12 THERMOCOUPLE | :
; AREA: 1160 ft Chemer | —
‘ ‘ DATA AND SUPERVISOR ROOM |- ~ ‘ : ; : =
- K b e e e | : \‘
i S AT IXILIARY CONTROL ROOM i \ " ;
2 AREA: 475 #12 MAIN CONTROL ROOM *\/\ NALCX'E : PCOE OL ROO 1\ . | :
AN UCLEAR PANELS | |, - j
| N \/ ,,,,,, e 0 e e -
| Cppry i IS LT
ot L RELAY CABINET--~ PRl = arel -
1 bk i
CONTROL POWER DISTRIBUTION CENTERS
W -
) = FUTURE PANELS

91
17

1.11.5 Control Panels

The main-control-board layout was approved after several revisions.
A 1/h-scale model of this board as revised is shown in Fig. 1.8. The
design of the control console was issued for comment and is shown in
Fig. 1.9. Detail drawings of the control-circult relay and the ther-
mocouple cabinets located in the auxiliary control area were also approved.,
Detailed fabrication drawings showing instrument layouts, panel-cutout
detalls, and wiring and piping diagrams are almost 60% complete, These
drawings will not be completed until exact dimensions of purchased equip-
ment are known.

SUNMCLASSIFIED
O PHOTO 37959

 

Fig. 1.8. One-Fourth-Scale Mode! of MSRE Control Panel,

UNGLASSIFIED
ORNL~ LR~ DWG 68578

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 1.9. Conceptual Design of MSRE Console.
18

1.11.6 Data Handling

A study of M3RE data-handling requirementss was completed and
issued Tor comment. The study indicated that the MSRE could advanta-
geously use a digital data-collecting and -handling system to process
the data from the reactor. The data system would be used in conjunction
with the conventional instrument and control system. The study report
was reviewed by MSR project personnel, and a decision was made toO procurc
a digital data system with the capability of performing on-line compu-
tations, detecting and alarming off-normal conditions, and logging the
process data automatically and on demand.

The MSRE instrumentation and control system is being designed Lo
ensure that failure of the data system will not compromise the safety
or operability of the reactor. All data necessary for the operation and
safety of the reactor will be available from the conventional instrumen-
tation system. The function of the data-handling facility will be o
implement the conventional system by recording data needed for experiment
analysis in a more convenient form. Provisions have been made in the
design of the instrumentation system to permit expansion of the conven-
tional data display and recording system should the need arise.

The study of the MSRE data-handling requirements was extended and
examined in more detail in order to prepare specifications for the data
system. A rough draft of the specifications was completed and distri-
buted for review and comment. The reactor operation and experimental
analysis groups are reviewing their respective data collection and display
requirements. These requirements and the comments on the preliminary
specifications will be incorporated in the final specifications.

At this time it appears that approximately 2L2 process variables
will be transmitted to the data system.

1.11.7 Process and Personnel Radiation Monitoring

An initial study of the MSRE radiation instrumentation requirements
was made, and a preliminary tabulation of the detectors” was issued for
comment.

The memorandum was reviewed by project personnel, and the comments
were incorporated into the instrument application diagrams.

The process monitors are those detectors used to monitor the reactor-
system process lines, some reactor components, the two stacks, the reactor
cell, the drain-tank cell, and the radiator pit. The personnel monitors
are those normally required for personnel protection throughout the inhab-
ited areas of the building. These monitors would normally be such ltems
19

as hand and foot counters, door monitors, continuous alr monitors, etc.,
as required by health physics practices.

The requirements for both types of monitoring is again being reviewed
in order to determine the specific detector and associated electronic
equipment required for each location. The possible detector locations
are being studied in order to determine the ambient operating conditions
so that adequate shielding, cooling, and cholce of materials and compo-
nents can be made.

1.11.8 Procurement Status

Specifications were written for the following eguipment during the
report period.:

. Weigh transmitters.

Thermocouple patch panels.

Valves for radiocactive-gas service.

Pressure and differential -pressure transmitters, weld sealed.
Venturi flow element for the coolant-salt system.
Mineral-insulated Inconel-sheathed thermocouples.

v s O OH

The specifications for items 1, 2, 3, M, were approved, and a
purchase order was placed for item 2. Purchase requisitions were
written for items 1 and 3. Ttems 5 and 6 were issued for comments.

Instrument cabinet requirements were determined and orders for cab-
inet frames, panels, and associated hardware were issued to the ORNL
stores department. (These are ORNL stores catalog items.)

Writing of specifications for the following items is continuing,
and they will be issued for approval early in the next report period.

l. Standard valves.

. Standard pressure and differential-pressure transmitter.
. Variable-area flowmeters.

Pressure regulators.

Recorders, indicators, and controllers.

Pressure switches.

Differential-pressure transmitter, NaK filled for high-
temperature (1500°F) service on the venturi flow element.

—J O\ WO

Equipment that can be economically salvaged from HRT facilities and
that meets MSRE specifications will be used.
‘N

20

REFERENCES

J. L. Anderson and R. E. Wintenburg, Instrumentation and Controls

 

Division, Ann. Progr. Rept. July 1, 1960, ORNL-2787, pp 145-150.

 

MSRP Progr. Rept. Aug. 51, l96l, ORNL-%215, p 9, Fig. 1.5.

 

O. W. Burke, MSRE Analog Computer Simulation of the System with a
Servo Controller, ORNL CF-062-12-50.

 

 

MSRP Progr. Rept. Aug. 3L, 1961, ORNL-5215, p 21.

 

Ibid., Fig. 1.15, p 22.

G. H. Burger, MSRE Data Collecting and Handling Requirements, A
Study Report, MSR-61-112.

 

G. H. Burger, MSRE Process and Personnel Radiation Monitoring,
MSR-61-108.

 
21

2. COMPONENT DEVELOPMENT

2.1 FREEZE-FLANGE DEVELOPMENT
2.1.1 MSRE 5-in. Flanges®

Two sets of INOR-8 flange forgings and four ring-joint gaskets were
purchased for production and testing of prototype MSRE flanges. Some
difficulty was encountered in machining the surfaces of the ring grooves
to precise dimensions. Successive cuts of less than 0.0l5-in. resulted
in rapid tool wear, and frequent sharpening was required.

A pair of the flanges is Dbeing installed in the seal~test facility2
for thermal-distortion and gas-leakage measurements. After completion
of these tests, the flanges will be installed in the thermal-cycle test
loop 3 and the prototype-fuel-pump test facility.?

A set of gages was designed to assist in obtaining a consistent over-
all thickness of the flange-gasket assembly. This thickness control is
necessary because oOf the load-deflection characteristics of the spring
clamps used to load the freeze flange. These gages were Tabricated and
used in the fabrication and inspection of the prototype flange parts.

2.1.2 Freeze-Flange-Seal Test Facility5

Testing of the 3—1/2~in. freeze flange, using various ring gasketls
and a resilient clamp, was discontinued after it was found that insuffi-
cient strength in the flange body permitted the ring-groove dimensions to
be changed by repeated thermal cycling.

Experiments were conducted to determine the temperature distribution
and the distortion characteristics of the 6-in. Inconel freeze flange.
The flange and clamp were similar to the 5-in. MSRE freeze-flange assembly
previously described, except for the following: pipe size (6-in. sched-LO
for 5-in. sched-40), material (Inconel for INOR-8), and the absence of
spring action in the clamp. Gap width between flange Taces was maintained
by a 20—3/8—in. diam ring of 1/16-in. Inconel wire.

Two separate heat sources were necessary to obtain even heating of
flanges and pipe extensions. Twelve silicon carbide heating elements
were connected in series and mounted in the bore, centered at the f{lange
faces. Clam-shell resistance heaters were aligned along the exterior of
the pipe stubs beginning 2—1/& in. from the flange face, with the insula-
tion beveled at a 45° angle away from the flange faces. Various combina-
tions of Calrods, clam-shell heaters, and induction heaters were tried in
22

arriving at this combination but they failed either to reach temperature
goals or to maintain them for extended periods.

As shown in Fig. 2.1, the static temperature distribution was meas~-
ured for bore temperatures of 850, 1000, 1100, 1200, 1300 and 1L400°F.
The freeze position of the molten salt was 4.8 in. radially outward from
the flange bore when the bore was at 1400°F. The average increase in
radial position of the freezing temperature was 7/8 in. for each 100KC
increase in bore temperature.

UNCLASSIFIED
ORNL-LR-DWG 68579

|
1400 S - Lo |

NO COOLING AIR CIRCULATED

 

 

1200 |

 

1000

 

850°F
| FREEZING POINT

  
 

800

 

600

 

TEMPERATURE (°F)

400 e

 

200 e

 

 

 

 

 

 

 

 

0 2 4 6

@
o

DISTANCE FROM THE CENTER LINE OF PIPE (in.)

Fig. 2.1. Temperature Distributions in the Central Plane of the 6-in. Freeze Flange

for Yarious Bore Temperatures.

At 1L00°F, the results were checked against the Sturm-Krouse® analyt-
ical predictions. Experimentally the freeze point was 7.8 in. from the
pipe center line, whereas the analytical prediction was 5.7 in. The as-
sumptions of continuity between flange and clamp and of radial temperature
variation are sources Ior the discrepancy.
23

With a bore temperature of 1300°F and of 1400°F, all power to the
assembly was cut off to simulate a salt dump and to note the effect on
temperature distribution. The bore cooled to the salt freezing point of
850°F in 27 min and in 34.5 min after shutdown, respectively.

When the flanges were at eclevated temperatures, there was a dimen-
sional distortion of the plane faces of the flanges, so that they tended
to assume a more conical shape. This is illustrated in Fig. 2.2, which
shows the distortion when the bore temperature was 1300°F. The maximum
distortion varied from 0.024 in. at a 1000°F bore temperature to 0.063 in.
at 1400°F; this is relative to an initial spacing of 0.062 in. between the
flange faces. No permanent distortion was noticed, although the flanges
were cycled 40 times from 200°F to 1300°F.

UNCLASSIFIED
ORNL -~-LR-DWG 68580

 

MALE SECTION FEMALE SECTION

 

. ‘ | ‘

 

RING POSITION

 

i O NO AR THROUGH

 

THE CLAMPS

 

READING

® 15 cfm THROUGH
EACH CLAMP HALF

i
£

 

 

   

RADIAL DISTANCE FROM PIPE CENTER LINE {in.)

 

 

 

 

 

 

 

 

 

 

 

 

-
i
Q
g1 o
]
5 \
B \
Q ®
./\'-~ »»»»»
_____ —AL-850°F: FREEZE POINT
1 \ WITHOUT AIR
®
e "."’850°F: FREEZE POINT
\ WITH AIR
®
5
120 80 40 0 40 80 120

GAP (thousandths of an inch)

Fig. 2.2. Distortion of 6-in, Freeze Flange Operating at a Bore Temperature of 1300°F,

The transient distortion characteristics were checked during a shut-
dovn from 1300°F. Fifteen minutes after power shutdown, the maximum
flange distortion had decreased from 0.061 in. to 0.034 in., while the
position of the freezing-point temperature moved in 1 in. The signifi-
cance of the rapid decrease in distortion, compared with the slow change
in salt freeze position, is that the amount of salt lodged between the
flange faces is reduced during cooling, and the tendency for permanent
distortion of the flange gap is lessened.
24

Lastly, the effects of cooling air were noted by passing 15 cfm of
instrument air through each clamp half (bore temperature of 1300°F).
(See Figs. 2.2 and 2.3.) Whereas the distance from the pipe center line
to the position of the freezing point temperature decreased from 6.5 in.
to 5.8 in., the distortion increased from a maximum of 60 mils (without
air) to 107 mils (with air). It was concluded that the temperature
gradient was the dominant factor leading to the distortion of the flange.

A development program similar to the one just described will be con-
ducted on the INOR-& 5-in. freeze flange. Also, the sealing properties °
of two nickel ring gaskets (one oval and the other octagonal) will be
determined when used with the spring clamp.

UNCL.ASSIFIED
ORNL-LR-DWG 88581

 

 

 

 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

1400 ‘ \
BOGRE TEMPERATURE
1200 +— Y§ - e —
AIR
1000 | — —
FREEZE POINT, 850°F\\\\k
- b LN
O BOO | e e e e e e e e fen
L
o
2 |
g
r {5 cfm BLOWING THROUGH /
o
S 600 | EACH CLAMP HALF ”fif' <
L
= ® \
@
400 } : : —
200 {- --—-
| |
o | |
0 2 a 6 8 10

DISTANCE FROM CENTER LINE OF PIPE (in.)

Fig. 2.3. Cooling Effects on 6-in. Freeze Flange.

2.2 CONTROL-ROD DEVELOPMENT

A simple chain-driven control-rod mechanism is being developed for
use in the MSRE. Figure 2.4 shows the system which is being built for
testing. The weight of the flexible metal hose and the poison elements
25

UNCLASSIFIED
ORNL~-{ R—-DWG 68582

 
  
 
 
  
  
 
 
  
  

REVERSIBLE-DRIVE MOTOR

MECHANICAL CLUTCH
GEAR ARM ... —

 

 

 

 

 

POSITION INDICATOR
SYNCRO TRANSMITTER
FIXED DRIVE
SUPPORT AND 3-in.
CONTAINMENT
TUBE o=
£
v
pR
w“N
\
6!

 

 

COOLANT EXHAUST -~

 

2-in. CONTAINMENT THIMBLE

 

 

 

Fig. 2.4, MSRE Control-Rod-and-Drive Assembly.

combined is 6 to 8 1b. The drive unit is overdesgigned, with the capabil-
ity of exerting a 20-1b downward thrust and a 25-1b upward pull. The
total travel of the poison elements is 66 in., at a rate of 1 in./sec.
26

The rod position is indicated by a calibrated digital voltmeter con-
nected to a synchro-operated linear resistor coupled to the chain-drive
sprocket shaft. A 12-v dc supply is furnished to the resistor. The rod
position is read out as a voltage, Tor convenience in data logging.

For maintenance, it is planned that the entire unit will be replaced
and direct maintenance performed in a "hot" shop. The entire control-rod
assembly can be removed by unbolting the cooling-gas container and drive-
unit base from the fixed support structure. Manually operated electrical
and air disconnects will permit withdrawal of the unit, leaving the
thimble and chain container in position.

Reactor cell air will be pumped to the gas container in order to cool
the drive motor and the control rod. Part of this stream is diverted into
the flexible-metal drive hose by means of a loose-fitting tube mounted
concentrically in the hose. All the gas leaves through a common exhaust
at the top of the containment thimble.

2.3 HEATER TESTS

2.3.1 Pipe Heaters

A full-scale section of removable heaters for 5-in. pipe was built
and tested. The test section is shown in Fig. 2.5. Kach heater unit is

1 BT " % : o JNCLASSIFIED

PHOTO 37519

 

Fig. 2.5. Pipe-Heater Test.
27

28 in. long and contains six flat-plate 700-w heating elements, which can
be seen in Fig. 2.6. A 6-in. yoke piece which is lapped to fit the
heater-box ends 1s inserted between heater sections to make a snug closure.
EFach heater unit can be replaced by first withdrawing the yoke at each end
of the heater, and then simply lifting the box. The heater base is shown
in Fig. 2.7. 1t contains no heaters and is a permanent part of the
support structure.

_UNCLASSIFIED
PHOTO 38320

 

 

Fig. 2.6. Pipe-Heater Section.
1600

1400

{200

1000
i
L
x
2

£ 800
he
w
o
>
|i¥]
’_

600

400

200

0

Fig. 2.8. Temperature Distribution in 5-in. Pipe Heater and Insulation.

28

Fig. 2.7. Pipe Shown Mounted on Heater Base.

UNCLASSIFIED

ORNL-LR-DWG 68583

 

    

 

 

 

 

 

 

 

 

 

   
    
  

 

  
 

 

 

 

 

 

\ T f
_PIPE-WALL AVERAGE TEMPERATURE
P
/;/REVERSE SIDE OF HEATER
¢ ;
!
|
: POWER INPUT (watts per linear foot)
i
o | o 52
? ® 437
........... A 345
| A 270
i
N
a4 UNION ASBESTOS & RUBBER CO.
(1200°F RATING)
® N g
UNARCOBOARD |
\ |
INSULATION ‘
OUTSIDE SKIN
. TEMPERATURE
i
| 0.005-in. ALUMINUM FOIL\\JV }
SUPERTEMP INSULATION ‘ l
- (1900°F RATING) fl_‘ L N\e
i (EAGLE - PICHER CO.)
| o
L L ‘ ‘ l
0 1 2 3 4

DISTANCE FROM PIPE WALL (in.)

UNCLASSIFIED
PHOTO 37975

 
29

Tests indicate a maximum 10°F variation in temperature around the
pipe surface, due to the unheated base section. Figure 2.8 is a plot of
temperature distribution through the insulation at various power inputs.
There is a heat loss of about 450 w per linear foot of pipe at operating
temperature.

2.3+2 Reactor Vessel Heaters”’

The reactor vessel heaters have operated a total of 3750 hr. The
heater-surface was held at 14250F for 700 hr to maintain 1325°F on the
heat~-sink face. The system was then opened for visual inspection, and
the general condition was the same as before. However, the upper guide-
tube rack was inadequately supported to carry the load of the guide tubes
at the higher temperature, and it had buckled. The buckling had not
affected the heater operation but had caused misalignment between the
guide tubes, the heater pins, and the thermal-shield penetrations. This
condition could create serious difficulties for remote malntenance and
possibly cause burnout of the heaters. After the rack is realigned, addi-
tional supports will be installed and the test continued.

The 30-gage Inconel reflector had buckled to some extent in spite of
precautions taken to avoid it. Hanging the reflector material and banding
it in place will be tried in order to reduce this distortion.

2.4  DRAIN-TANK COOLERS

The prototype cooling bayonets® for removing fuel afterheat in the
MSRE drain tanks have operated for a total of 3860 hr. Fach l~l/2~in.
cooling thimble has the capability of removing 6.3 kw when inserted into
1300°F molten salt to a 60-in. depth. The bayonets have been thermally
shock tested through 40 cycles, without apparent damage to the cooling
tubes except for minor warping of the l-in. Inconel boiler pipe. Thermal
shocking was accomplished by drying the steam system out completely,
allowing the bayonet temperature to approach the salt temperature, and
then adding water to the steam drum. The thermocouples for measuring pipe
temperatures were shielded in stainless steel tubes and strapped to the
pipe surfaces. Attempts to use bare-wire thermocouples arc-welded to the
pipe were not successful.

The test system 1s being modified to permit automatic thermal shock-
ing as described above for life testing of the bayonets.

2.5 SAMPLER-ENRICHER SYSTEM
2.5.1 General Concept
The present concept of the MSRE sampler-enricher system is shown in

Fig. 2.9. The prinecipal changes made to the system previously reported®
are:

1. A buffered, double-sealing gate valve replaced the solder freeze
valve as the primary containmment valve during sampler maintenance.
     
  
 

 

    

 

  
   

   
 
   
   
   

   
  

     
   

   
 
 

   

 

 

30

TRANSPORT CASK

  
 
 
    
  
  
       

UNCLASSIFIED
ORNL~LR— DWG 63318R2

 

 

 

 
 

 

   
    
   
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AREA 4
CELL ” SAMPLE TRANSPORT CONTAINER
ILLUMINATOR / REMOVAL VALVE
VIEWING —EXHAUST
SYSTEM s o) Lol by
R SR ZRIERS
: m\g PR CABLE DRIVE
RS X TR " MECHANISM
PCXOKS
SO, N
":’s‘z’s’ ‘KHELIUM
;sts\’s‘ PURGE
T0 b
VACUUM LI
PUMP
MANIPULATOR . 70
URANIUM VACUUM
SHIELDING P PUMP
OPERATIONAL j r“\
VALVE R S
LEAD -
SHIELDING |
AREA 2B
N TIPSR
At
AREA 2A
-OUTER STEEL SHELL
DISCONNECT P ////
FLANGES
®-
EXPANSION
JOINT

CONTAINMENT'J////

LATCH STOP

PUMP BOWL

 

VESSEL SHELL

 

Fig. 2.9. MSRE Sampler-Enricher System.
31

2. A flanged seal isolates the upper manipulator area (3) from the
area of the valves (2B) in place of a sliding seal.

3. The seal between the contaimment vessel (area 2A) and area 2B
was flanged.

k. The periscope was changed to a mirror system external to the
outer compartmwent, with a quartz window for viewing.

The maintenance valve was changed because of the unreliability of the
solder freeze valve tested earlier and the complications produced by the
gas controls necessary to make a double solder seal.® A double-sealing
gate valve similar to the operational valve (sec 2.5.2) will be used. The
other changes were made to simplify maintenance procedures and to increase
general rellability.

2.5.2 Operational Valve

A bellows-sealed, 2-in., Crane double-sealing gate valve considered
for use as the operational and maintenance valves was received and tested.
As received, the Stellite-faced gate did not seat properly against the
Stellite-faced seats and had to be lapped to fit. After the valve was
reworked, 1t was opened and closed by hand 100 times. A torque wrench
wag used to apply a reproducible torque of 7) ft-l1b on closing, considered
to be an acceptable closing force. Periodic checks of buffer-gas leakage
to atmosphere through the seals were made. With a buffer pressure of
L0 psig of helium, the total leakage was 0.3 cc/min. After 100 cycles the
leak rate did not appear to be increasing.

A motor operator was obtained for the valve, and leak rates will be
determined for motorized operation.

2.5.3 Removal Seal for Sample Container

A buffered, double-seal is being tested for use as the sample-
container removal seal. The seal, located belween the sample-transport-
container removal valve and the transport cask, prevents oxygen contami-
nation of the outer compartment while the capsule is being inserted or
removed from the outer compartment. The seal consistsg of two rubber
O-rings, lubricated with a light film of silicone grease and buffered with
15-psig helium.

For the seal test, the sample transport container or rod is moved
through the seal by a hydraulic cylinder; it pauses for 1 min and is then
withdrawn from the seal. Total leakage of buffer gas through the O-rings
is measured periodically while the rod is at rest. After 1400 cycles,
total leakage was <l.cc/min with a 15-psig helium buffer pressure. Sev-
eral scratches from previous tests did not appear to affect the cealing.
The rod is coated with a light film of silicone lubricant.
32 .

2.5.4 Detail Design of Sampler-Enricher

Detail design of the various components of the campler-cnricher
system 1s continuing. Preliminary design of the transfer tube, operation-
al and malntenance valves, valve enclosure, inner compartment, manipulator,
periscope, sample-transport-contaliner removal scal and transport cask are
complete. Detalling of the outer compartment and other components is in
progress. A layout is being made of the shielding and vacuum pump area.
Revisions to the instrument flowsheet arc in progress.

2.6 MSRE CORE DEVELOPMENT
2.6.1 TFull-Scale Core Model

The pressure vessel for the full-scale MSRE core model was installed
in its test loop. Because of difficulties in extruding the aluminum mockup
of the graphite core blocks to tolerance, delivery of the core internals
(including the core shell and support structurec) was delayed five months.
They have just been received and are being installed in the model.

2.6.2 Core-~Inlet Flow Distribution

Without the core internals, the only hydraulic test that could be
performed was the measurement of the flow distribution in the entrance
volute. These data were obtained at 100, 50, and 25% of the reactor flow
rate of 1225 gpm and appear in Fig. 2.10.

UNCLASSIFIED
ORNL-LR-DWG 68584
T o T

 

 

 

 
  
 
   
  

 

  

 

 
 

 

 

 

 

 

 

28 T [ 1 z | ;
| | | ‘ |
| ! 1 i ‘ % i |
2wl e | , :
o | | | |
QO : !
/s |
e 1 /
20 {- 7 | e — L -
e ; WATER USED AS j
3 1 3 TEST FLUID /
v i !
- ‘ 100% DESIGNED FLOW RATE (1225 gpm) |
— 16 _ i — hege, s R £ e
> o | | } wT /
£ | ! -
3 i | |
Ll | ‘
> 12 b— S— ] -~/ -
o ~ | | / /
- :
3 L/“/ |  50% DESIGNED FLOW RATE (612.5 gpm) / //
L { ( [
8 : ‘ } i Ly
O TTe— v
'25% DESIGNED FLOW RATE (305 gpm) | }N@\O /
- ! ! G : ‘ s
| . i O'N"B"'“ T .| P
‘ ‘ ! | (O *"\fll//
| 1
. Lo ‘
0 45 90 135 180 225 270 315 360

8, ANGLE FROM INLET TANGENT (deg)

Fig. 2.10. Center-l.ine Velocity Distribution in Volute of MSRE Full-Scale Core Model.
33

bince water was used for this test, the MSRE Reynolds number was not
reproduced exactly. However, the Reynolds numbers in the volute are well
into the turbulent range, and since the various coefficients involved are
weak functions of the Reynolds number at these high values, the error in
velocity distribution is expected to be small. The velocities reported
here are about 10% higher than those predicted from runs made in the
l/5~scale model.t This difference is attributed to inexact scaling of
the model. The uniform decrease of velocily in the volute with tangential
position indicates the desired uniform supply of fluid to the vessel.

2.7 HELIUM PURIFICATION

Work on the construction of a full-scale oxygen~removal unit is about
50% complete. The design, illustrated on Fig. 2.11, provides for a
titanium getter tube encased in a cylindrical resistance heater which is
in turn surrounded by an outer pressure vessel. ‘The outer wall is pro-
tected from excessive temperatures by an annular layer of high-temperature
insulation. The pressure vessel is L4-in. sched-40 pipe and is about 30 in.
long. The unit will operate at 250 psig, lZOOOF, and 10 liters/min of
helium flow, that is, at design conditions for the reactor gas system.

UNCLASSIFIED
ORNL~ LR—-DWG 68585

THERMOCOUPLE

 

 

 

 

 

 

(1) GETTER TUBE

(2) HEATER, 1000 w

(3) HIGH-TEMPERATURE INSULATION
(3) PIPE, 4-in. SCHED-40 SS

(5) FLANGES, 4-in. 1500-1b SS,

 

 

g in. WELDING NECK, RING JOINT
@ r ®) INSULATION
2tin. @— (7) REFLECTOR

THERMOCOUPLE -1 1JI1 [} |— THERMOCOUPLE

THERMOCOUPLE — | |~ THERMOCOUPLE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

? N0
N
@0\
@}W“ . I
@»-»-—
LTHERMOCOUPLE
THERMOCOUPLE

Fig. 2.11. Oxygen-Removal Unit, MSRE Cover-Gas System.
34

An electrolytic~-type trace-oxygen analyzer was purchased for use in
monitoring helium samplcs from the oxygen-removal unit. Tests with helium
of known oxygen concentration showed the analyzer to work satisfactorily
in the range of O to 10 ppm of oxygen, provided that fluctuations in
sample gas temperature are small.

2.6 MSRE ENGINEERING TEST LOOP (ETL)

The first run of the ETL (Fig. 2.12) was intended to evaluate the
effectiveness of the [lushing operation by following the oxide content of
the salt during extended operation.12 This objective was compromised by
the discovery, after 3150 hr of operation, that some zirconium fluoride

UNCLASSIFIED
ORNL -LR-DWG 54492A

 
  
  
 
 
   

~——SAMPLER-ENRICHER MOCK UP
_—SAMPLER-ENRICHER TRANSFER TUBE

—DANA PUMP
7

Il _.. - - GRAPHITE HANDLING DRY BOX

MANUAL
SAMPLER

   

GRAPHITE ACCESS
( FREEZE SEAL) —_

     

— \

  
 
  

 

  

/ I BOS
S
=y U
= ,/// ;
FUEL TANK ;LVfé?““\FLUSH TANK

MANUAL SAMPLER

 

FREEZE VALVES

Fig. 2.12. MSRE Engineering Test l.oop, The original loop did not contain the sampler-enricher

mockup, the graphite container, and the dry box.

had inadvertently been included in the salt mix, and ZrOs had precipitated.
The balance of the run was devoted to an investigation of various means of
removing ZrOs from the system and the accomplishment of the removal.
35

2.8.1 Loop Operations

Operation with the ETL included (1) tests to verify the cause of the
initial slow drain,*® (2) attempts to dissolve precipitated ZrOs by tem-
perature cycling of the salt and by increasing the ZrF, content, and

(3) removal of ZrOs manually from the pump bowl and chemically (with HF)
from the drain tank.

After the treatment of the salt in the drain tank with HF, the loop
was shut down for major alterations, including the installation of the
INOR-8 graphite container.

A summary of the entire ETL operation is given in Table 2.1. The
elapsed time begins with the startup of the loop on April 20, 1961.

2.8.2 Operation of Freeze Valve

In the previous period it was reported that the initial draining
operation required an excessive amount of time.t3 Tests were performed
to separate two possible causes.

The first test ran for a long period, during which drain-line tem-
peratures were kKept at or below previous settings to see whether the line
would again plug as the result of normal operation. The loop operated
continuously for 560 hr, after which the loop was drained in a normal time
with no difficulty.

The second test duplicated the BeO addition made previously. An
addition of 30 g of BeO in pellet form was made to the pump bowl as before,
requiring over 200 hr. At 3150 hr, or 320 hr after the BeO addition, dif-
ficulty was again experienced in draining the loop. Extra heat applied to
the valve did not help drain the loop. However, when extra heat was
applied to the l/Z—in. drain line to raise its temperature from the range
920 to 1130°F to the range 980 to 1190°F, the loop drained. (The freezing
point of the salt was 850°F.)

From this and other evidence, it was concluded that the difficulty was
due directly to the BeO addition. Portions of the pellets (original size,
O.4 in. diam x 0.5 in. long) left the capsule, entered the pump suction
and settled in the static salt of the l/2~in. drain line, there forming a
sludge due to segregation and reaction with the Zrf, in the system.

Examination®** of the sludge formed directly beneath the addition port
revealed the presence of Zr0s, 2Li¥-*Befs, and 2ZLiF-Zrt,. The 2L1iF-Bels
phase melts at 850, and the 2LiF*Zrf, phase melts at 1105°F. The presence
of the higher-melting 2Z2LiF-ZrF, phase, along with the ZrO0s, formed enough
of a plug so that the loop would not drain. By raising the temperature of
the drain line 6OOF, the sludge became sufficiently less viscous to allow
the loop to drain. In total, the loop was drained and refilled Th times,
without difficulty except for the two above-mentioned occasions.
Table 2.1.

36

Operating History of the ETL

 

 

Time
(hr) Remarks
0 Beginning of the operation, April 20, 1961
940-1140 First addition of BeO pellets
1280 Loop drained to begin freeze valve tests.
Excessive time required to open valve®
1280-1690 Operational tests performed on freeze valves

1690-2250

2250-2830

2830-3050
3159

3150

3200~3500

3820

4180

4180-4380

4380-4680

l'700-5200

5200-5860

Operation for 560 hr to attempt duplication
of the previous slow drain

Additional operational tests performed on
freeze valve

becond addition of BeO pellets

Salt sample ETL-69 revealed prescnce of
Zr0s crystals

Drain time again excessive after BeO
addition

Temperature cycling of salt Trom 1050°F in
drain tank to 1200°F in the loop, for oxide-
transfer attempt

Additions of “Zri, were begun in order to in-
crease oxide solubility

Third addition of Zrl', was made, bringing
salt composition to LiF-BeFs-ZrF,, 062-3L-4
mole ¢

Circulation for 200 hr at temperaturcs up Lo
1200°F

Loop drained, cooled, and pump removed for
examination

salt treated in drain tank with mixture of Ho
+ HF

Drain tank kept hot for additional salt
samples; final sample (E1L-133) extracted

 

g5ee ref 13.

and Loop operation terminated for alterations
37

2.8.3 Oxide Removal

Sludge had been seen and sampled through the lwl/Zuin. sampler connec-
tion to the pump-bowl lid while the system was at temperature (llOOOF).
Two methods of redissolving the sludge were tried, and visual inspection
was used as the method of estimating their effectiveness. First a
"cold trap" type of removal was attempted by circulating salt over the
sludge at 1200°F, draining and cooling in the drain tank to 1050°F. This
cycle was repeated three times during the interval 3200 to 3500 hr (see
Table 2.1). Second, ZrF, was added in the form of ZrF,-LiF (49-51 mole %)
on three occasions to increase the expected oxide solubility. According
to the visual observations, the sludge deposit was not affected appreciably
by either of these attempts. The loop was drained into the flush tank for
cooldown and manual removal of the sludge from the pump.

The results of the oxide analysis of samples taken between the time
that ZrF, was discovered (3150 hr) and the pump was removed (4400 hr)
averaged 563 ppm oxygen, as shown in Fig. 2.13.

2.8.4 HF Treatment

Prior to the complete shutdown of the loop for alteration, apparatus
was set up and operated by the Reactor Chemistry Division for the treat-
ment of the salt in order to remove the precipitated oxide believed to be
in the drain tank. The treatment, consisting of passing a mixture of
hydrogen and HF through the salt, is described in Sec 6 of this report.
The treatment took place over a period of 500 hr and removed the equiva-
lent of 1025 ppm of oxygen from the salt. Figure 2.13 shows the results
of the oxide analysis of salt taken during the treatment period (4700 to
5200 hr) and afterwards (until 5900 hr), while the drain tank was kept
hot. The change in oxide concentration and the amount of oxide removed
by treatment indicates that a considerable amount was precipitated in the
drain tank.

The amount of concomitant corrosion in the Inconel drain tank is in-
dicated in Fig. 2.14 by the increase in the chromium content of the
samples taken between 4700 and 5200 hr. This increase is equivalent to
the removal of all the chromium to a depth of 0.0005 in. over the entire
of drain-tank surface; although the partial depletion of chromium extends
deeper, the attack is not excessive.

Examination of five Inconel dip tubes used for bubbling the HI and Hp
through the salt revealed only very light to moderate surface roughening
and pitting.

2.8.5 ETL Graphite Facility

The &-in.-diam graphite container,® with a longitudinal frozen-salt
seal access, wags made of INOR-8. The container was installed into the
loop as indicated schematically in Fig. 2.12. The upper portion of the
container is shown in Fig. 2.15. At the same time, most of the loop
piping was changed from Inconel to INOR-O.
1300

1200

1100

1000

900

800

700

OXIDE {ppm)

600

500

400

300

200 :-

100

Fig. 2.13. Resuits of Oxide Analysis — Engineering Test Loop, Oxide content of salt vs number of hours

CIRCULATED SAMPLES == -+ DRAIN-TANK SAMPLES
i HF + H, TREATMENT

ZrF, ADDITIONS

UNCLASSIFIED
ORNL-LR-DWG 68586

TERMINATED 12-20-64

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— T T T T — T T
Y Y 1 T e  wivy dyn 99 ‘ 1
D i i 1
- - — ~ , - . - — ]
© O OPEN (UNFILTERED) SAMPLE
I O B ® FILTERED SAMPLE [ 77 |
— S i &5 .
. O i
1 O o
| |
¥ O !
L e i _ | 5
AVERAGE OF 37 SAMPLES o ®
7 ——— FROM 2150 TO 4400 hr o ]
e e
2 :
- - T O O ©
o o ;
o o8 © o
- D - 6\
o ®
— O o . ! O
3 ® ; ;
o o 9 { AVERAGE @ |
T © | e e T e
oc% , e e |
S, - : [ i O ® i 1 i -
SEPTEMBER : OCTOBER NOVEMBER } o ! DECEMBER
! : i ! ; i ,
i | i | | | i | | | 1
30 32 34 36 38 50 42 44 46 48 50 52 54 56 58 (x10%)
TiME {(hr)

ZrF4 additions and HF treatment (four months),

since beginning of test, including

8¢
39

UNCLASSIFIED
ORNL~LR~DWG 68587

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

600
S
A® eee;
- e wp |
500 :”. - Tis®
.
E ole® s o .5"‘.‘&‘
2 * ® . > ®
~ 400 ® e ®
2 : I
E ° 0.,&1.
z ¢ . HF TREATMENT
S 300 o el , IN DRAIN TANK - o]
© a op
oS
z s
= g
Q200 5
S ®
e
100
0
0 1000 2000 3000 4000 5000 6000

CHROMIUM CONTENT (ppm in flush salt) vs HOURS SINCE BEGINNING OF ETL OPERATION

Fig. 2.14. Results of Chromium Analysis ~ Engineering Test Loop. Chromium content (ppmin

flush salt) vs hours since beginning of test.

 

£ UNCL ASSIFLED

i G‘:\’AP;HIYTE ' s . ,‘ . ::’ S S | : : . 7 PHOTO 37826
(65 InLONG) -~ ‘g SR e ; N 1 « :
S GASKET MAKEUP . i ; o 5

TO DRYBOX

~OVAL RING GASKET |
AND FLANGES <

SALT OUTLE

o EXTERNAL
"~ COOLING
COILS

GRAPHITE HOLDDOWN

GRAPHITE SUPPORT PLATE

INTERNAL COOLING BAFFLE

Fig. 2.15. Engineering-Test-Loop Graphite Container, Showing Freeze Joint,
40

The dry box has been fabricated and has recelved preliminary testing.
Figure 2.16 shows the loop and graphite container during construction,
and the dry box in the unmounted position. After operution of the loop,
the dry box would be mounted over the container for access to the graphite
through the flange.

 

UNCL ASSIFIED
PHOTO 37919

i DANA PUMP
AND SUPPORT

GRAPHITE CONTAINER

 

Fig. 2.16. Engineering-Test-l.oop Graphite Facility in Construction.
41

2.9 MBRE MAINTENANCE DEVELOPMENT

The objective of the maintenance development program is to ensure the
capability of replacing any item of equipment which fails after the reactor
has become radiocactive. Reacltor-cell eqguipment problems are being studied
in the maintenance mockup.'® The product of this program will be (1) an
inventory of tools or tool designs which have been appropriately tested
and demonstrated and (2) procedures and techniques for performing the
MSRE maintenance within the capabilities of the available handling equip-
ment and the known operating requirements. The following sections (2.9.1
to 2.9.6) summarize progress made on specific operations.

2.9.1 Placement and Removal of Freeze-Flange Clamp

The first concept for putting clamps on freeze flanges called for
using 1—1/2~in. bolts and trunnion nuts. This method was abandoned because
of repeated severe thread galling.

A second method, utilizing hydraulic cylinders (see Fig. 2.17), vas
successful. The clamps were driven on and off the flanges many times,
with no trouble. A further advantage is that this method does not require
the closely held parallelism between the upper and lower clamps that the
screw-thread system does. The equipment required to operate the clamps
remotely is being designed and fabricated.

2.9.2 Flange Alignment and Pipe-Jacking Tools

These are used to provide force to move the flanges, as shown in
Fig. 2.18. The jack moves axially to open or make up the joint and to
hold the joint closed while the flange clamps are driven on, and the align-
ment tool provides force vertically and laterally, to overcome distortion
and preloading of the spring supports. Both tools were tested and were
able to move the pipe flanges of the simulated MSRE line in the prescribed
directions and amounts.

2.9.3 Gasket-Replacement Procedures

The time interval in the maintenance procedure while the flanges are
open is critical because of the possibility of damage to the gasket and
groove, because of the danger of spreading particulate contamination from
the reactor internals to the cell atmosphere, and because of the danger of
contaminating the reactor system with materials in the cell atmosphere.

It is believed that four tools will be required for gasket replacement:
(1) some form of dust catcher must be utilized until the openings can be
covered, (2) a tool capable of freeing a stuck ring gasket must be used
to remove the existing gasket, (3) two separate covers must be placed to
seal the two pipe openings, and (h) a handling tool is needed to position
and lock the new gasket in place. The flange covers were received and are
being improved with respect to handling, space requirements, and sealing
ability. The stuck-ring tool is being fabricated. The remainder of the
tools will be designed in the near future.
42

UNCLASSIFIED
PHOTO 37988

R

C
R

HYDRAUL |
OPERATO

|

 

 

ic Clamp Operators.

Hydraul

7.

2

ig.

F
 

43

UNCL ASSIFIED
PHOTO 37993

FLANGE ~ALIGNING. |
TOOL

PIPE JACK o

Fig. 2.18. Flange Alignment and Pipe-Jacking Tool.

2.9.4 Miscellaneous Disconnects

Several components are directly accessible in the plan view, and
their removal constitutes unit remote operations that do not require co-
ordination with other operations. These are: power, thermocouple and
electronic disconnects, and pipe line and reactor heaters. Remote han-
dling of one of these components generally consists in engaging a hook of
some sort, imparting either a lifting or twisting motion to free the unit
and then transporting it to the storage position. Several of these hook
tools have been built and tested. A test of an electric-power disconnect
indicated the need for refining the viewing and handling techniques, with
special emphasis on protecting the insulated leads.

2.9.5 Remote Viewing

In some maintenance work, remote viewing will be required. A
1.625-in.-dianm wide-angle periscope was borrowed for testing disconnect
operations. This device was superior to an 0.875-in.-diam periscope used
earlier, especially in its ability to "see" in dimly 1lit areas. Prelimi-
nary quotations have been received on a similar modular periscope capable
of working in the MSRE radiation environment.
44

2.9.6 Component Removal

The only work done on this phase has been to practice the direct-
vision handling of the circulating pump, using the special 1ift sling and
the overhead crane (see Fig. 2.19). All auxiliary piping will be added
to the pump mockup as final details become known, and a full replacement
procedure will be tested.

#® UNCL ASSIFIED
PHOTO 37995

 

Fig. 2.19. Simulated MSRE Pump with Lifting Sling.
45

2.10 BRAZED-JOINT DEVELOPMENT7
2.10.1 Joint Design

The tapered braze Jjoint for l-l/z-in. sched-40 pipe was further modi-
fled (Fig. 2.20) to use 0.005-in.-thick sheet-braze preform that places
braze metal throughout the joint. Axial pressure is maintained on the
Joint members during brazing so that when the braze metal melts, the joint
thickness is reduced to 0.001l to 0.002 in.

 
  
  
 
 

UNCLASSIFIED
PHOTO 37965

  

COMPLETED LAPERED JOINT

005 INCH THICK SHEET BRAZE METAL
WAS PREPLACED IN THE JOINT. AXIAL
PRESSURE WAS MAINTAINED ON THE ASSEMELY.
WHILE IT WAS HEATED TO 1900°F, HELD for

ONE MINUTE, AND PERMITTED TO COOL.

     

 

 

 

Fig. 2.20. Brazed Pipe Joint, Showing the Sectioned, Completed Joint (Above) and Starting Materials
(Below).

2.10.2 Braze-Joint Testing

Ultrasonic and metallographic inspection of completed joints made
with prototype equipment on the bench indicated 81 to 86% bonding. One
completed l-l/2~in. Jjoint was subjected to tensile loading, with the
following results: at 1300°F it held a 28,000-1b load, held a 10,000-1b
load at 1500°F for 5 min, an 8000-1b load at 1600°F for 5 min, and failed
at 1700°F under an 8000-1b load. Visual inspection of the separated joint
indicated complete or nearly complete wetting of the mating INOR-8 sur-
faces by the braze metal.
46

A representative braze joint was held at 1250°F and exposed to
reactor salt for a total of 74 hr of intermittent exposure, simulating
drain-line conditions. At the completion of the test, the joint will be
examined metallographically.

2.10.3 Remote Fabrication of Braze Joint

Tools for the various mechanical operations required to fabricate a
braze joint remotely were received and satisfactorily operated. The func-
tions of the tools and the sequence of operations are as follows:

1. The traveling vise is lowered over a continuous run of
pipe and bolted to the base plate (Fig. 2.21). The vise
jaws are then closed on the pipe.

2. The pipe cutter is lowered over the pipe onto its sup-
port on the traveling vise (Fig. 2.22).

3. The pipe-cutter drive is operated, carrying the cutter
around the pipe. The knife is automatically fed by the
star wheel and pin device (Fig. 2.23) about 0.009 in.
each revolution. After the cut is completed, the cutter
is opened and removed.

i,  After the component is removed, the tapering tool is
placed on the base plate. The traveling vise feeds
the pipe into the rotating shaped cutter (Fig. 2.24),
which machines a 6° included-angle taper to a depth of
1 in. When the cut is finished, the pipe is withdrawn
and the tapering tool removed.

5. One portion of the furnace can now be placed over the ta-
pered male-pipe stub.

6. The new component with the female-joint half, contalning
the braze preform, and the remainder of the furnace are
installed.

7. The fixed vise is installed over the female-joint half to
hold and align it (Fig. 2.25).

8. 'The sliding vise then inserts the male into the female
pipe joints, the furnace is assembled over the joint,
and the thermocouple leads are connected to the record-
ing instruments. After an adequate inert-gas purge,
the joint is inductively heated to 1850°F while axial
force is maintained; it is held at that temperature for
a minute and then permitted to cool.
R

R

S

UNCLASSIFIED
PHOTO 37963

%

P

47

i

3

 

ipe.

inuous P

ing Vise Being Lowered over Cont

21. Slid

@

Fig. 2
48

o
LJ
L
%21
%)
<
J
O
=
o

PHOTO 37964

R

s
St

 

ipe,

ise over Continuous P

ing V

ipe Cutter Being Mounted on Slid

P

Fig. 2.22.
49

UMCL ASSIFIED
FHOTO 37971

   
   

e REMOTE DRIVE SHAFT

Fig. 2.23. Pipe Cutter. Drive shaft rotates cutter around pipe. Star wheel and pin feed the knife with

each revolution,
50

UNCL ASSIFIED
PHOTO 37968

  
   
   
 

 

REMOTE
DRIVE SHAFT

Fig. 2.24, Pipe-Tapering Machine,

UNCLASSIFIED

PURGE LINE =~ »- , A t g THERMOCOUPLE PHOTO 37967
FEED SCREW f ’. LEADS
ORIVE VISE-JAW | ) } a

'''' ! : e V| SE - JAW
DRIVE

     
   
 

- {NDUCTION -
HEATING

   

VISE §

o

  

Fig. 2.25. Joint and Furnace Ready for Assembly.
51

2.11 MECHANICAL-JOINT DEVELOPMENT

The overflow line to the fuel-pump bowl described in the design
section of this report is routed through an area too crowded to permit
making a braze joint after component replacement. Therefore, a mechanical
jJoint is being developed for this service. Since the Jjoint will be ex-
posed to salt for only short periods during reactor-fill operations and
during infrequent overflow from the pump, a trapped-gas pocket will be
used to keep salt out of the region of the gasket. Initially, the over-
flow line will be continuous with joint halves installed, and preparations
will be made for cutting at the appropriate points for installation of the
mechanical Joint. A prototype will be tested.

Z2.1l2 STEAM GENERATOR

A steam separator for a bayonet-tube steam generator-superheater was
constructed and tested in the separator test chamber previously reported_.l8
The flow areas within the test separator were as follows:

"Boiling" annulus — 0.312 in.<
Swirling-separator annulus — 0.0638 in.
Steam-outlet nozzle — 0.0283 in.=2
Water-return tube — 0.1452 in.?®

2

Swirl vanes l-7/8—in. long and having a 300 angle with the separator axis
were used.

Alr and water were circulated through the separator at various flow
rates and ratios of flow rates. 1In all cases, the water carryover was
very high. As an example, with an air flow of 3 ¢fm and a water flow of
Os5 gpm, the water carryover was approximately 0.2 gpm.

Further work was postponed due to a shortage of manpower.

.13 PUMP DEVELOPMENT

The design drawings for the MBRE fuel pump were approved, and the
thermal analysis of the fuel and coolant pump tanks was completed. Water
testing of the coolant pump model was completed. Fabrication of the
rotary element for the prototype fuel pump was completed, and fabrication
of the pump tank was nearly completed. Design drawings for the lubrica-
tion stands and the drive motors were submitted for review. Additional
ITNOR-8 castings of impellers and volutes for the fuel and coolant pumps
are being made, and dished heads for the pump tanks are being inspected.
52

Z.13.1 MOSRE Fuel Pump

PK-P Pump Hot Test

 

This test pum;pl9 was placed back in operation to obtain more testing
of the resistance of the lower shaft seal and of the impeller to cavi-
tation damage and to obtain additional observations on lubricant inven-
tory. The pump circulated LiF-BeFo-ThF,-UF, (65-30-4-1 mole %) at 1225°F,
510 gpm, and 1950 rpm, and has operated for (88 hr. The differential
pressure across the lower shaft seal was maintained at 3/M psi, and no
measurable oll leakage was noted.

Prototype Fuel Pump and Hot-Test Facility

 

Fabrication of the rotary element®® was completed, and dimensional
checking, assembly, and bench testing were started. Fabrication of the
pump tank is about 80% complete.

Modifications were made to the support structure of the test facil-~
itygl to accommodate the flexible mount designed for the fuel pump. Sev-
eral other modifications were completed to accommodate testis of various
items including (1) the bubble liquid-level device, (2) the freeze flange,
(3) a section of pipe heater, (4) operation of the sampler-enricher device
in the pump tank, and (5) the comparison between temperature readings for
thermocouples installed in a thermowell and attached to the external sur-
face of the piping.

Thermal Analysis of MSRE Fuel and Coolant Pumps

 

The thermal stress and strain Tatigue ana1y51522 of these two pumps
has been completed. The thermal stresses in the fuel pump during reactor
operation at 10 Mw which were previously reported have been revised to
include the effects of the meridional heat flow along the surface of the
pump tank. The maximum principal stresses are shown in Figs. 2.20 and
2.27, and in Figs. 2.28 and 2.29 for the fuel and coolant pumps, respec-
tively.

Calculations were completed, showing that a constant cooling— air
flow of 200 cfm on the upper surface of the fuel pump tank will provide
adequate life; therefore, the previously proposed automatic air-flow-
control system will not be used. A forced-convection air-cooling systen
is not required for the coolant-salt pump. The predicted total usage
factor of the fuel and coolant pumps are 0.30 and 0.57, respectively, com-
pared with a safe design value of 0.08.

. 2 . . . . .
A detuiled report®® covering the analysis was written and is being
reviewed.

MSHR ruel Pump
The design of the rotary element®? was completed and approved. Four

each of the rotary elements were placed Ior fabrication by off-area fabri-
cators. The design of the pump tank was completed and approved, and
53

UNCLASSIFIED
ORNL-LR~DWG 64496

 

30,000

VOLUTE SPHERE

Ti

20,000 /- JUNCTION
-—/7\ 10 Mw OPERATION

10,000 f-or-rrm ffod 200 acfm COOLING AIR

 

 

 

 

INTERNAL CYLINDER EXTERNAL CYLINDER

Iy
-10000 —A—— 1 ) \“/Q\\

-------------- O- POWER OPERATION
200 acfm COOLING AIR

 

 

STRESS (psi)
o
N

 

 

  

 

—

-20,000 OP FLANGE

 

 

 

 

 

 

 

 

 

 

-30,000
-6 -4 -2 0 2 4 6 8
AXIAL POSITION (in.)

Fig. 2.26. Principal Thermal Stresses in Cylinder of Fuel Pump.

UNCL.ASSIFIED
ORNL. ~LR—-DWG 64497

 

4000

2000
\ O POWER, 200 acfm COOLING AIR

CYLINDER T

,_——_—-——‘—"M
JUNCTION \4
-2000 / fi

\M

~4000 \
/ 10 Mw, 200 acfm GOOLING AIR

-6000

~8000 I///,

~10,000

~12,000

 

 

 

 

STRESS { psi}

 

 

 

 

 

 

 

 

 

 

 

0 | 2 3 4 5 6
MERIDIONAL  POSITION (in.)

Fig. 2.27. Principal Thermal Stresses in Sphere of Fue! Pump.

partial fabrication will precede the delivery of the INOR-3 volute castings
and the requisite INOR-8 pipe and tubing.

The dished heads for the pump tank were received and are being in=-
spected. Three pairs of impeller and volute castings of INOR-8 are being
poured by the founder of the coolant salt pump castings. Design drawings
of the drive motors were reviewed and returned to the manufacturer for
revision.
54

UNCLASSIFIED

ORNL~LR-DWG 64498

 

 

30,000
VOLUTE

T

 

 

20,000 -

 

 

O POWER NO EXTERNAL
COOLING

 

 

  

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10,000 |-
Q
o 10 Mw NO EXTERNAL COOLING J
w
ul O | |
o
% 10 Mw NO EXTERNAL COOLING

240,000 fboreoreeeer e e e

O POWER NO EXTERNAL
SPHERE JUNCTION COOLING
-20,000 \\
TOP FLANGE
-30,000 I
-6 -4 -2 0 2 4 6 8

AXIAL POSITION (in.)

Fig. 2.28. Principal Thermal Stressesin Cylinder of Coolant Pump.

ORNL-LR-DWG 64499

UNCLASSIFIED

 

60

 
  
 
 

 

10 Mw NO EXTERNAL COOLING

O-POWER NO EXTERNAL COOLING

 

 

 

 

 

 

 

 

@
Q
O
Q
Q
; 20 e e i
wn
&
- CYLINDER
n JUNCTION
0
|
|
!
-20 {
0 1

2

3
MERIDIONAL POSITION (in.)

4

Fig. 2.29. Principal Thermal Stresses in Sphere of Coolant Pump.

MSRE Iubrication Stands

 

The design of the Jlubrication stands for the MSRE fucl and coolant
and valves have been pro-

punips has been completed.
cured for the two stands.

The pumps,

filters,

 
55

A stand was constructed for proof testing the canned-rotor-type pumps
to be used in the lubrication stands, and one pump was operated for
1428 nr, circulating a turbine-type oil at 160°F, 70 gpm, and 3500 rpm.
Discoloration of the oil was noted, and examination of the pump revealed
a varnish~like coating on most of the surface of the rotor, indicating an
insufficient flow of cooling oil through the motor cavity. The pump
support bearings were modified by adding three axial grooves spaced at
120° to to reduce resistance to the flow of cooling oil through the motor
cavity. The flow was increased several fold and the temperature of the
external surface of the motor stator was reduced from 270 to 195OF. An
endurance test will be made.

2.13.2 MSRE Coolant Pump

Coolant-Pump Water Tests

 

The MSRE coolant-pump water tests®> have been completed, and the
impeller diameter has been tentatively set at 10.3 in. to provide a Tlow
rate of 850 to 920 gpm, assuming the actual head loss in the coolant-salt
system is within t10% of the design value of 78 ft. The hydraulic char-
acteristics of the pump are shown in Fig. 2.30, and the efficiency of the
pump at operating conditions is approximately 78%.

UNCLASSIFIED
ORNL~LR-DWG 68588

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1O
HEAD
—mh\
100 /— \ ]
T PREDICTED FLOW. RANGE
-\\\\\\\ WITH SYSTEM RESISTANCE

90 ~ — AT B850 gpm =78 ft 10%

80 ™
2 70 \\\\ )
2 N
¥ 60 N
= \\\\
£
- 50 [ ]
é /
x SHAFT Hp AT SG =1,92‘,,,/f”“’
n 40

320 //

(,//f””/ IMPELLER 0.D: 10.27 in.
20 frr CALCULATED FROM HYDRAULIC TEST DATA ~ ]
WITH IMPELLER 0.D: 9.90 in.
{0
0
0 200 400 600 800 1000 1200 1400

CAPACITY. (gpm)

Fig. 2.30. Hydraulic Characteristics of MSRE Coolant Pump.
56

Reactor Coolant Pump

 

The design drawings are being checked prior to review. One satisfac-
tory volute and two satisfactory impeller castings of INOR-0O were received.
Weld repairs are being made to one impeller and volute to raise the
casting quality to Class I level. The other impeller is Class I quality,
without repair. An additional volute casting of INOR-8 was poured by the
founder, who reports that radiographic and dye-penetrant inspections
indicate an acceptable casting.

2.13.3 Advanced Molten-Salt Pumps

Pump with One Molten Salt Lubricated Bearing

 

The test of this pump®® was terminated at 12,500 hr of operation
during which LiF-Bels-UF, (62-37-1 mole %) was circulated at 1225°F,
(5 gpm, and 1200 rpm; 92 start-stop operations were sustained. For the
last two start-stop operutions, rubbing could be detected while the pump
shoft was rotated by hand; however, no evidence of the rubbing could be
found in the recorded trace of the power input to pump-drive motor.

Examination of the Journal and bearing surfaces indicated slight
rubbing. The bearing assembly is shown in Fig. 2.31. Each of the two
sets of gimbals supporting the hearing were satisfactorily operable at
the conclusion of the test. During disassembly, difficulty was experi-
enced with the removal of the INOR-8 journal sleeve from the Inconel pump
shaft. The difficulty resulted from self-welding of the INOR-8 sleeve
and the Inconel shaft, and is shown in Fig. 2.32. The bearing and journal
will be replaced for further endurance testing.

UNCL ASSIFIED
PHOTO 37393

 

 

TkT’/’l'l",'lk‘l'ri""f‘l"‘"l"}f"f T
. 1 2 3
[ INCHES \

 

Fig. 2.31. Gimbals-Mounted, Molten-Salt-Lubricated Journal Bearing.
57

UNCL ASSIFIED
PHOTO 37896

AR
R

 

Welding Between Inconel Shaft and INOR-8.

0320 Self-

Figo 2
58

2.1% MSRE INSTRUMENT DEVELOPMENT
2.14.1 Thermocouple Attachments

Welded thermocouple attachments made with Heliarc welded, TNOR-8
adaptor lugs of different weights and shapes27 are being ecvaluated in a
series of tests. Initial tests were made in order to determine the
bonding strength and effects, if any, on the calibration. Attachments
made with side lugs welded along the edge parallel to the axlis of the
sheath®® withstood the greatest pull, bend, and prying force before sepa-~
ration occurred.

Seven thermocouples were attached to a section of INOR-G pipe by the
methods tested above and then checked for calibration accuracy under
static conditions between 900 and 1400°F. The greatest error noted for
any couple tested was +6°F, which is within the 3/4% tolerance specificd
for the Chromel-Alumel material used in this test. These thermocouples,
which are now being soaked at 1200°F, will be rechecked in several weeks.

Additional thermocouples are being prepared for testing under simu-
lated operating conditions. They will be tested on the Engineering Test
Loop, freeze-valve test, and a pipe heater section in the pump-test loop.

A test rig was assembled for developmental testing of mechanical
attachments for use on the radiator tubes in the MSRE. 'The procedure con-
sists of heating a L in. section of 7/8-in. 0D x 0.065-in.-wall INOR-8
tube with a cartridge~type heater which was inserted in a silver-plated
copper slug located inside the tube. Reference thermocouples are located
between heater and slug, slug and tube wall, and on the outside wall of
the tube. The latter is a 30-gage bare-wire thermocouple, spol welded to
the tube wall. The results of a preliminary test made with a l/8uin.wOD
Inconel-sheathed, MgO-insulated thermocouple are as follows:

Outer-Wall Temp.(°F)

 

 

 

Air Flow Inner-Wall Temp. (°F) Reference Couple Test Couple
0 1230 1150 1060
80-90 fps 1230 1040 850

As indicated by the results, the inner-wall thermocouple was not
measuring true wall temperature. This will be corrected before further
tests are conducted. Also, the heated section will be purged with inert
gas to prevent cracking of the plating on the copper slug. Test thermo-
couples have been made with the junction ground to the sheath's wall, which
will be placed next to the heated surface.

2.14.2 Temperature Scanner

Development of a thermocouple scanning :system,,g9

commutator, is continuing.

using a mercury-jet
59

During the early development of the system, the method of switching
created considerable noise. The switch was then being operated in a
break-~before-make switching mode. Operation in this manner resulted in
the generation of 0.25 v and greater noise pulses, with pulse durations
about 1 to 2 pusec. Further examination of the switch output signal re-
vealed that the noise pulses always occurred during the break portion of
the switch action. It was postulated that the pulses were caused by
static charges generated by the mercury jet in the insulated region
between the signal pins. The noise pulses did not appear when the jet
was in contact with the signal pin because the source impedance (thermo-
couple and lead-wire resistance) was about 100 ohms or less. However,
when the Jjet moved away from the signal pin, the input impedance became
very large, and the nolse pulses appeared in that edge of the output
signal.

The switch manufacturer verified the cause and existence of the noise
pulses.

The problem was eliminated by changing the switch operation to make-
before-break, resulting in a low input impedance at all times. The output
signal from this type of operation is different from the break-before-make
action because, during the make period, the mercury jet is in contact with
two signal pins simultaneously for about 10% of the outpubt-signal width.
During this time, the output signal is the average of two adjacent input
signals. However, in the break portion, the jet is again on a single
signal pin, and the output signal is equal to the emf of a single thermo-
couple. This type of output signal is acceptable as an input to the pro-
posed alarm discriminator and does not materially alter the display and
identification of signals on the oscilloscope.

The development has progressed to the point that a preliminary design
has been completed. The system design appears to be workable, and
detailed circuit design is in progress. An alarm-detector circuit is
being designed and is expected to be completed by April. One complete
100-point scanning system is being built for testing.

It appears that at this point in instrument-system design, approxi-
mately 400 thermocouples will be required to be scanned by the scanning
system.

The radiator-temperature monitoring system will require 120 points.
These will utilize two scanners of 60 points each, with provisions for
detecting and alarming on low radiator-tube temperature.

The remaining 280 points will utilize three scanners. One scanner
each, with provisions for producing high or low alarms will be used on the
fuel system, the coolant system, and the drain tank system.

The final arrangement, operation, and integration of the five scan-
ners into a system will await testing of the first prototype unit, now
being built. It should be completed in April.
60

2.14.3 Single-Point Temperature-ilarm System

Investigation of methods of economically obtaining signals which
reliably indicate the operating status of freeze flanges and freeze valves
is continuing.

Results indicate that the Electra Systems Corporation monitoring
system, described previously, 2° will be suitable for monitoring freeze-
flange temperatures but will not be sultable for the freeze-valve opera-
tions because of the lock-in (seal) feature inherent in this system. 1In
the freeze-flange monitoring system, only high-temperature-alarm monitor-
ing is required, and since all temperatures would normally be below the
alarm point, an alarm would occur if any frecze-flange temperature rises
above the alarm point. However, in the frecze-valve-monitoring opecra-
tions, both high- and low-temperature signals arce required to indicate
whether the valve 1s open or closed, and the monitor is required to indi-
cate immediately when the operating state of the valve is changed. The
lock-in feature would prevent the change of state of the alarm monitor
until the monitor is manually or automatically reset. This reset action
is considered undesirable as it would produce spurious signals which could
interfere with the operation of safely interlocks and which would be
annoying to the operator.

Other methods of monitoring the freeze valves arc being investigated.
One device which appears promising is a magnetic relay, Daystrom Magsense
Control Relay, Model A-82, manufactured by Daystrom, Lncorporated,
Ia Jolla, California.

The unit is a complete solid~state device consisting of a magnetic
amplifier with an isolated winding for the input signal and a silicon-
controlled rectifier to furnish the output switching action.

The relay operates from a 25 to 30 v dc supply. The signal input
range is O to 100 pa dec, standard. The input resistance is a nominal
360 ohms. The relay has a response time of 200 msec, standard. The alarm
setpoint is adjustable with tl% of' the full rated input range by a poten-
tiometer or by external means. The output is analogous to a single-pole
double-~throw relay rated at 1 amp at 25 to 30 v de. This can be considered
as two outputs, one energized above the set point and one below set point.
The repeatibility of the alarm set point is *1 pa or +1% of the input
range. 'The hysteresis is less than 2 pa and can be increased or decreased
by use of an external resistance. The operating temperature range is 35
to 100°F at full output rating and at stated operating specifications.

An Klectra Systems Corporation monitoring system has been placed on
order. It will be tested to determine whether it mects {the operational
and reliability requirements of the MSRE.

One A-82 control relay is on hand and being tested. The results to
date indicate that it performs to specifications and would be sujtable for
use 1in monitoring the freeze valves.
61

The freeze-valve-monitoring operation will require three relays for
each valve. Integration of three A-82 relays into a monitoring system
would require additional design to supply indicator lamps, power supply,
and a calibration method. However, at this time i1t appears that this type
of system offers the greatest reliability and best operating characteris-
tics at the least cost.

2.14.4 Pump-Bowl-Level Indicator

Development of a continuous-level 31 eclement for use in measurement
of molten salt levels in the MSRKRE fuel and coolant salt pump bowls has
continued. During the past report period several high-temperature differ-
ential transformer designs were investigated, two level-element designs
were developed, and fabrication of a level test facility incorporating the
two level-element designs was completed. Testing of the prototype level
elements 1s underway.

The major effort in the program during this report period was devoted
to the development of a differential transformer which would operate
reliably (without excessive shifts in characteristics) in the temperature
range from 850 to 1300°F. Several variations in the design of the
transformers were investigated, and three designs were selected for fur-
ther testing. These included the nickel-wound lava-insulated transformer
previously reported,sl'a transformer similar to the nickel -wound
transformer but with Inconel windings, and a guartz-insulated transformer
with Hastelloy C windings.

Although previous tests of the nickel-~wound transformers 3% were prom-
ising, the known and tested durability and low temperature coefficient of
resistivity of Inconel, the lower resistivity (compared with Inconel) of
Hastelloy. C, and the negligible temperature coefficient of resistivity of
Hastelloy C indicated that these materials should be tried. Neither
Inconel nor Hastelloy C were as satisTactory as nickel. The Inconel
transformer was durable, but the temperature effects on the output signal
of this transformer were greater than on the nickel-wound transformer
(see Fig. 2.33). The Hastelloy-C transformer also exhibited excessive
temperature effects on the output signal. "The wire in this transformer
became very brittle and broke during a Tirst inspection made after one
heat cycle to 1LOOCF.

Two unforseen difficulties were encountered while fabricating and
testing the transformers. One was caused by the use of unfired Fiberfrax
insulating material, the other by the magnetic characteristics of the tube
furnace used in the test. All transformers tested were encased in an
Inconel coil form which gave mechanical protection to the coils and insu-
lation. Fiberfrax paper was placed between the transformer and the sur-
rounding case to absorb mechanical shock. When the nickel transformer
assembly was heated, the organic binder in the Fiberfrax insulation evap-
orated, condensed on the lava forms and on the ceramic insulating beads
on the lead wires, permeated the insulation, and carbonized. The result-
ing carbon deposits effectively shorted the transformer; measured resist-
ance to ground from either primary or secondary was less than 10 ohms.
62

UNCLASSIFIED
ORNL-LR-DWG 68589

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50
40 |
uw INCONEL |
2 }
Q
w
i T —
5 !
- ’LM._ p
© T ? w»sszzr.':i,:::m»m@
f
S.[)
Z NICKEL | .
or
]
r<:I: ‘ i
z CONSTANT VOLTAGE EXCITATION
BOTH TRANSFORMERS ADJUSTED AT 1050°F
1 ,,,,,,,, - - ]
0 CORE MOVED WHILE TRANSFORMER WAS AT 1050°F SO THAT RECORDER
INDICATED 25% OF SCALE ’ ‘
\ &
o | | |
600 700 800 900 1000 1100 1200 1300

TEMPERATURE (°F)

Fig. 2.33. Temperature Characteristics of High-Temperature Differential Transformers Wound with
Nickel and Incone!l Wire. Constant voltage excitation; both transformers adjusted to 1050°F; core moved

while transformer was at 1050°F so that recorder indicated 25% of scale.

The nickel winding was att.cked and, for all practical purposes, destroyed.
The trouble was corrected by completely dismantling the transformer, baking
the lava forms at lSOOOE and then reassembling, using new insulating beads,
nickel wirce, and Fiberfrax paper which had been prefired in air to drive
out the organic binder.

During initial testing of the trunsformers considerable difficulty
was encountered in controlling the furnace temperature and in obtaining
useable data. 1t was noted that, when the furnace was on, the Honeywell
temperature recorder-controller indicated temperature variations of about
200°F but did not control the furnace. It was also noted that the Dynalog
recorder, used to indicate the differential transformer output, becamc
insensitive and locked in one position when the heaters were on and that
the temperature of the transformer core was 30 to 50°F higher than the
outer transformer temperature.

Investigation showed that the heaters in the Marshall tube furnace
were spirally wound. The 60-cycle magnetic field produced by this spiral
configuration induced voltages in the secondary of the transformer and in
the thermocouples sufficient to saturate the amplifiers of the Honeywell
and Dynalog recorders.

The difficulty was avoided in subsequent tests by heating the trans-
former above the temperature required, turning off the heaters, and taking
63

data as the transformer cooled. (Care will be taken in the design of
reactor-system heaters to ensure that this difficulty does not occur in
the reactor level-element installation.)

During testing of the transformers, an interesting phenomenon was
observed. When the transformers were heated above the curie point of the
i1ron core and then allowed to cool so that the temperature passed back
through the curie point, the signal output from the differential trans-
former increased very rapidly to a high peak value at the curie point of
the iron and then decreased rapidly to a much lower value, thus producing
a gpike on the recorded trace at the curie point of the core material.
The spikes occurred repeatedly at the same temperature (see Fig. 2.3L4).
There 1s a possibility that this phenomencn could be used to obtain an
accurate temperature calibration check point or as the basis for a simple
high-temperature-alarm device. The lemperature at which the spike occurs
could be varied by the use of core materials of different curie points.

UNCLASSIFIED
ORNL-LR-DWG 68590

 

60

 

50

 

 

40

 

 

30

DYNALOG CHART READING

 

20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1300 1400 1300 1400
TEMPERATURE (°F)

Fig. 2.34. Temperature Effect on OQutput Signal of Differential Transformer

as lron Core Passes Through lts Curie Point,

While the transformers were being tested, the design of the level
test facility was modified to permit testing of an additional level-
indicating system. A new tank and head assembly was added, and provisions
64

were made for the detection of molten-salt level with a spark plug probe.
Provisions were also made for future addition of a bubbler-type level
device. As shown in Iig. 2.35, the new head assembly is designed to mount
a differential transformer above the tank. The core for this transformer
projects from the top of the float through the gas space above the float

UNCLASSIFIED
ORNL -LR-DWG 68591

0 e )

 

SPARK PLUG
//

PROBE PRESSURE

_# INDICATOR

LEVEL
RECORDER f———--
NO. 2

DIFFERENTIAL
TRANSFORMER NO. 2~

. ~CORE NO. 2 Jfi

 

 

Prirenn

TEMPERATURE ~
RECORDERJ ] ]_HELIUM

 

 

X
S

 

 

 

 

- FLOAT
NO. 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o/

LEVEL
RECORDER
NO. 4

- DIFFERENTIA:
TRANSFORMER NO. 4

 

 

| |
1 |
| |
| l
| |
| |
| |
| | (
l o / e CORE NO. 4 }
FURNACE = ;
| |
| S
| I
l |
| l
| I
| |

 

[J CHEATERS- - -
. - ’, - A
FREEZE VALVE d FREEZE VALVE -
NO. 3 - NO. 2 = —— | f--
o8

 

 

 

 

 

 

 

  

 

Fig. 2.35. Simplified Diagram of MSRE Level Test Facility.
-

65

and into the core tube inside the differential transformer. This differ-
ential transformer is insulated but at present 1s not heated. The end of
the transformer next to the furnace is operating between 980 and 1000°F,
the other end between U480 and 500°F. While unnecessarily severe, this
operating condition does provide every opportunity for the level system
to fail. If there are to be any deposits of solids from the vapors above
the molten salt in the temperature range 1200 to 500°F, the proper condi-
tion exists. If there are deposits, they presumably will accumulate on
the core and core tube to the extent that the core will no longer move.

Construction of the level test facility and fabrication of the two
prototype level-element assemblies was completed, and testing of the level
elements is proceeding. A photograph of the completed test facility, made
prior to installation of insulation, is shown in Fig. 2.36. In this photo-
graph, the differential transformers can be seen above the tank on the
left, and below the tank on the right.

| UNCLASSIFIED
{ PHOTO 37897

 

Fig. 2.36. Molten-Salt Level Test Facility.
66

The system has operated three weeks atl temperature. 5Salt was added
after the first weck of operation at temperature. Preliminary results are
very cncouraging. They indicate that the temperature effects on span and
zero, over the range 800 to 1200°F, are acceptably low and that there is
no pressure effect and negligible hysteresis. There has been no evidcnce
of salt absorption by the graphite float. The tests are being continued
to determine the effects of continuous high-temperature operation on the
characteristics of the transmitters.

2.14.5 Single-Point Level Indicator

The test of the single-point level indicator previously reported32
was terminated after two months operation. At the end of this period there
was an 11% reduction in the amplitude of the signal. Due to the on-off
characteristics of the device this was not objectionable. When removed
from the salt pot for inspection, the portion of the probe which was
inside the pot but above salt level had an even deposit of green material
adhering to it. The portion which was below salt level was clean and
shiny. The probe was damaged during the inspection, and teslts were dis-
continued.

A prototype of the probe assembly that will be used in the MSRE 33
is being fabricated and will be tested.
REFERENCES

1. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 28-32.

 

2. Ibid., p 28.

3. MSRP Quart. Progr. Rept. July 31, 1960, ORNL-301k, p 20,

 

L. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 5l.

 

5. TIbid., p 27.

N

Sturm-Krouse, Inc., Analyses and Design Suggestions for Frecze Ilange
Assemblies for MSRE (Nov. 30, 1960).

 

 

7. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 33.

 

8. 1Ibid., p 37.
9. 1Ibid., p 37-
10. Ibid., p Lo.
11. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 3(-39.

12. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 54-55.
26.

27.
28.

67

lbid., p 5>.

R. E. Thoma, "2240 Service Samples (10/19/61)," private communication.
MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 57.

Ibid., p 58-61.

Ibid., p 61l.

Ibid., p 65.

Tbid., p 46.

Ibid., p 48.

MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 49.

lbid., p 50.

C. H. Gabbard, Thermal Stress and Strain Fatigue Analysis of the
MSRE Fuel and Coolant Pumps, ORNL-TM-78 (to be issued).

MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 50.
Ibid., p 5k4.

Tbid., p 5h.

 

Ibid., p 78.

Tbid., p 79, Fig. 2.41.
Ibid., D 77

Ibid., pp 25-27.

Ibid., pp O7-72.

Tbid., pp 73-T7h.

Ibid., p 76, Fig. 2.39.
68

3. REACTOR ENGINEERING ANMNALYSIS

3.1 REACTOR PHYSICOH

3.1.1 Analysis of MSRE Temperature Coefficient
The previously reported analysisl of the temperature coefficients -
of reactivity in the MSRE was extended to include the effects of retaining
fission-product xenon and samarium in the core graphite and the effect ol
including a small amount of a nonml/v absorber for the purpose of increas- :
ing the size of the temperature coefficient.
Since there are low-lying resonances in ooth Xel®% and Smlég, the av~
erage thermal -absorption cross section Tor these nuclides decreases with
increasing temperature above about 200°C; in certain situations this can
lead to a positive contribution to the temperature coefficient, due to the
reduced poison absorptions relative to fuel absorptions. There are cer-
tain nuclides (e.g., Rh*©3) for which the average thermal cross section
decreases slowly with increasing tewperature; insertion of such an absorp-
er would lead to an increase in the size of the temperature coefficient
due to the change in thermal utilization with temperature.

The principal contributions to the MSRE temperature coeffilcients are
due to the increase in thermal leakage with increasing temperature; alter-
ation of the thermal utilization with temperature is not a large effect.
Therefore the cnanges in temperature coefficients resulting from either -
fission-product buildup or poison insertion are small. Results of calcu-
lations for rhodium insertion are shown in Fig. 3.1; salt, graphite and
total temperature coefficients are plotted vs the ratio of the thermal-
macroscopic~apsorption cross section of tne unpoisoned core. I the as-
sumption is made that all fission-product Xe*3® and Sm»*° are retained in
the core grapnite, then the grapaite tewmperature cocificient 1s changed
from ~6.0 x 1075/°F to -5.2 x 107°/°F; the salt temperature coelficient
is changed from -2.8 x 1075/°F to -3.1 x 1075/°F, so that the total is
changed from ~8.8 x 107°/°F to -8.3 x 1077/°F.

3.1.2 Reactor~Kinetics Studies

Prelininary results were obtained from simplified reactor-kinetics
calculations®’ 3 of the response of tne MSRE to step-reactivity insertions.
A modified version of the PET-I 7O pr@grama was put into operation on the
IBI 7090; additions to the program include revised input and outpult formats
and the treatment of separate fuel-salt and graphite temperatures and cem-
perature coefficients of reactivity. Power, fuel-~salt temperature rise,
and grapnite-temperature rise are plotted as a function of time aflter o
69

% step insertion (starting from 10 Mw) in Fig. 3.2; peak pressures are
plotted as a function of step size in Fig. 3.3.

UNCLASSIFIED
ORNL-IL.LR-DWG 68592

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

{x1075)
-10
p SUM
o —9 W
—— =8
.
o
o~
s
= GRAPHITE
L
o
L
L
L
Q
o
bt
T -4
’_
<L
o
a FUEL SALT
o
k.
= Fig. 3.1. Temperature Coefficients vs Relative
-2 Poison Level,
0
0 0.01 0.02 0.03
2, (RH) /2, (ORIGINAL CORE)
UNCLASSIFIED
ORNL-LR-DWG 68593
5000
— 500
POWER
N\ FUEL
4000 | A 400
_ W
= o
3 3000 300
— to)
R n
Wl @
= L
2 5
}_
2000 200 o
1)
Q.
\ =
i
/ L
1000 e e M/ #400
_—
/ GRAPHITE (X 40)
0 ] 0
0 0.1 0.2 0.3 0.4 05

TIME (sec)

Fig. 3.2, Power, Fuel Temperature Rise, oand Graphite Temperature Rise.
70

UNCLASSIFIED
ORNL-LR-DWG 68594

1000
500
200

100
fg
a

=~ 50

58]

@

a

Ll

o

D

)

w

w

o

0.

« 20

<1

i

Q.

10
5
2

 

0 0.2 0.4 0.6 0.8 1.0
STEP REACTIVITY INSERTION (%)

Fig. 3.3. Peak Pressure Rise vs Step Reactivity Insertion,

3.1.3 Gamnma-leating Survey

An IRM 7090 program, 2DGIH, was developed during the period to per-
form calculations of gamma~ray heat deposition in a reactor using a source
generated by the FQUIPOISE~3 two-group, two-dimensional diffusion theory
program.®  The 2DGH program is a revised and improved version of NIGHTMARE;5
attenuation is computed in the 2DGH program for a two-dimensional distri-
bution of materials within the reactor, while in NIGHTMARE the material
71

distribution must be one-dimensional.. In addition, some modifications
were made in the input and output formats.

Results of gamma~heating calculations with 2DGH for various locations
in the top head of the MSRE vessel are shown in Fig. 3.4. The rectangular
blocks outline the regions used in the reactor model in the EQUIPOISE-3
calculations. FEnergy-deposition rates (in w/cms) are listed for INOR-O
in several locations.

UNCLASSIFIED
ORNL—-LR—-DWG 68595

 

0.063 w/cm3

' 0.119 w/cm3
| __—®005(w/em3— |
T € )

 

 

 

@ 0.045 w/cm® ]

 

 
  
 

—_-
-

% in.

8|
® 0.015 w/cm?

 

 

I
1
M"~~__,___M~,m__Mwfimlq&gtgwé___wwmmmm_«m_ww___JL_fi____~£o792wkms

0.882 w/cm3

Fig. 3.4. Top Head of MSRE Vessel, Showing Results of Gamma-Heating Calculations and the Mathe-
matical Model Used.

REFERENCES

1. MSRP Prog. Rept. Auz. 31, 1961, ORNL-3215.

 

2. 5. Jaye and M. P. Lietzke, Power Response Following Reactivity Addi-
tions to the HRT, ORNL CF-506-12-106 (Dec. 30, 19506).

 

 

3. P. R. Kasten, Homogeneous Reactor Safety, ORNL CPF-55-5-51 (May 9, 1955).

 

ll, T. B. Fowler and M. L. Tobias, HQUIPOISE~3: A Two~Dimensional, Two-
Group, Neutron Diffusion Code for the I1IBM~-7090 Computer, ORNL~3199
(Feb. 7, 1962).

 

5, M. L. Tobias et al., NIGHTMARE - An IBM-7090 Code for the Calculation
of Gamma Heating in Cylindrical Geometry, ORNL-3193 (Feb. 9, 1962 ).

 
72

PART Il. MATERIALS STUDIES
4. METALLURGY

.1 DYNAMIC CORROSION STUDIES
4,1,1 Fluoride-Salt Contamination Studies

A program has been initiated to examine the effects of oxidizing
impurities on the corrosion behavior of fused fluoride mixtures and to
establish tolerances for these impurities in the MSRE cover-gas system.
Tnitial studies have been concerned with the hydrolysis (by water vapor)
of fluoride melts and with the corrosive properties of the hydrolysis by~
products, in particular HF,

As previously reported,l a preliminary INOR-8 thermal-convection-loop
experiment (1252) operating with HF-contaminated salt was terminated by
a flow restriction about 200 hr after startup. The salt was an equimolar
mixture of NaF and ZrFs, almost saturated with HIF'. Operating conditions
for the loop are shown in Table L.1.

Table U4.1. Operating Conditions for INOR-8 Thermal-
Convection Loop 1252

 

Salt mixture, 50-50 mole % NaF-Zr'q
Operating time, hr 223
Maximum salt-metal interface temp., °F 1260
Minimum fluid temp., °F 1205
TLoop AT, °F 35
Total time of HF exposure, hr 193

 

Specimens removed from the hot~ and the cold-leg sections of the
loop were uniformly attacked along exposed surfaces to a depth of l/2 mil.,
as shown in Fig. k.l. Deeper penetrations were evident in some areas but
appeared to follow original flaws in the loop tubing. A series of inserts
73

JUNCLASSIFIED
| T-20522

 

 

 

 

 

 

 

 

 

Fig. 4.1. Corrosion of INOR-8 Tubing Following Exposure to HF-Saturated Molten
Fluoride Salt.

had been incorporated in the hottest section of the loop for the purpose
of weight-loss determinations; however, self-welding prevented their
removal intact. These inserts, whose surfaces had been carefully machined,
exhibited noticeably less attack than surfaces of the loop proper.

Chemical analyses of the salt were obtained during and after test;
results are shown in Table 4.2. The concentration of nickel increased to

Table Lk.2. Chemical Analysis of Salt Circulated in Loop 1252

 

 

 

Sample ?fi?fi NiConstitgint (ppm)Fe
Sstart 0 20 140 55
During test 9% 690 195 105
%During test 196 1800 220 170
Prermination 203 800 170 65

 

aSam;ple obtained with 25-g copper filter stick.

bSample obtained by melting salt from metallographic
specimens.
74

a level of 1800 ppm within about 200 hr of operation. Increases were

also found in the concentrations of the iron and chromium, but to a lesser
extent than the increases for nickel. Analyses of the after-test salt,
determined from samples taken from the metallographic specimens, showed
significantly lower corrosion-product concentrations than did filtered
samples removed shortly before the test was terminated.

Visual examination of the after-test salt revealed evidence of marked
phase segregation. As a result of this observation, a specimen of the
salt was removed from an area near the original plug and submitted Tor
X~ray diffraction studies. Results showed the frozen salt to be composed
of two phases: TNaF«6Zrly and Nal-NiFo.27rF4, in almost equal concen-
trations.

Studies conducted by Blood® on the solubility of NiFs, in the NaF-ZrFg4
salt system showed the saturating phase for NiFs to be NaF-NiFo-27rF4.
It was also found that, as this phase is formed, it immediately precipi-
tates. Thus, 1t appears that the concentration of NiFy in the salt during
the test exceeded the solubility limit under the conditions of operation;
consequently, Nil's precipitated from solution, causing the plug. These
precipitates no doubt also account for the differences in the analytical
results of filtered and after-test sswples.

As a result of this initial test, the program was revised to utilize
static capsules rather than circulating loops. The static-test method will
facilitate the removal of salt and metal samples and will allow better
control of impurity additions. Design of the static pots and auxiliary
equipment is complete, and four units are being assembled.

4.1.2 Molybdenum-Graphite Compatibility Tests

An examination of the second of two INOR-8 thermal-convection loops
operated with molybdenum inserts revealed essentially the same findings
as were derived from the first loop test 1250, reported previously.
Operating conditions for this second loop (1251)9 which were similar to
the first loop, are summarized in Table 4.3. Both loops contained graph-
ite sleeves? adjacent to the molybdenum inserts; however, the volume of
graphite in the second loop was twice that in the first.

Metallography of the INOR-8 specimens removed from the second loop
revealed heavy surface roughening and pitting, dbut, like the first loop,
specimens of the as-received tubing revealed a similar pitted condition.
Examination of the INOR-8 spacers, which held the molybdenum in contact
with the graphite, exhibited attack in the form of surface roughening and
pitting to a maximum depth of 1 mil along the inner surface which had been
exposed to salt. Evidence of scattered carburized areas to a maximum
depth of 5 mils was found on the outer surfaces of the spacers. These
carburized spots corresponded to areas where the INOR-8 was in direct con-
tact with the graphite., A surface layer, similar to that generally present
after long-term INOR-8 tests, was found along the outer surfaces which
w;re in contact with the molybdenum. The maximim depth of this layer was
1/4h mil.
75

Table 4.3. Operating Conditions for Graphite-Molybdenum
Compatibility Test (Loop 1251)

Molybdenum: Universal-Cyclops arc-cast Mo + 0.5 Ti alloy; the
molybdenum was used In the as-rolled condition

 

Graphite: National Carbon Co. Grade R-0Q25
Max. salt-metal interface temp., °F 1300
Max. salt temp., °F 1300
Min. salt temp., °F 1140
Loop AT, °F 160
Operating time, hr 5545
Salt used LiF-BeFo-Z1rF4-ThF 4 ~-UF4
(70-23-5-1~1 mole %)
Salt-to~-graphite volume ratio 1:1

 

Examination of the after-test molybdenum strips revealed no change
when compared with the microstructure of the as-received specimens.
Chemical analyses of the molybdenum after the test, as shown in Table L.k,
likewise compared closely with as-recelved specimens. However, tensile
tests of the specimens, as in the case of the first loop, indicated very
low room-temperature ductilities.

Table L4.4k. Before-and-After Chemical Analyses of Molybdenum Strips
Contained in Loop 1252

 

 

 

. Wt % Component (ppm)
Specimen o T T Gr  Te I P
As received 99,65 0.49 67 40 252 390 5 100
As tested® 99.7 0.k 100 100 200 380 5 100

 

aAverage of four specimens.

Fuarther evaluations of the molybdenum specimens from both loops
traced the cause of apparent ermbrittlement to surface contamination pres-
ent in the as-received molybdenum stock. Evidence of surface contaminants
had been indiscernible in the highly cold-worked microstructures of the
as-received and after-test samples. However, following an annealing
treatment in vacuum at 24LO0°F, which effected recrystallization of uncon-
taminated areas but did not disturb "impurity-stabilized" areas, a contami-
76

nation zone was clearly delineated, as shown in Fig, 4.2. No difference

in the extent of contamination was seen between as-received and after-
test samples.

B UNCL ASSIFIED
_1-20825

 

 

 

 

B UNCLASSIFIED
120832

 

     

 

 

 

 

| . e TN fl./
S AR e o - o ;;Q “{«':,‘w* ,,....m : - i .“QJ »
% /J \ . m - \\M\\‘ ors
N - e T
| i ennren l‘ T f/ oi4
o~ ”‘/P\'F/ . M wwwwww J -\/‘/:
S f—“‘m Pt s p 'C_ - _ o
- ‘5, 3 S i
. \My.‘«ww.«ymw/“‘“m"~--»~~~~-~ I | 5|
(5) e Sy T :
P, - ~ g I .

Fig. 4.2, Comparison of Microstructures of As-Received and Annealed Specimens of

Molybdenum Sheet Used for Loops 1250 and 1251.

(@) As-received microstructure; (&) an-

nealed in vacuum at 2400°F, Etchant: 10 g K3Fe(CN)6, 10 g KOH, 100 cc H2O'
77

Two additional molybdenum-graphite loop experiments have been pro-
gramed to study the effect of various surface preparations and heat treat-
ments on molybdenum compatibility with graphite-~salt systems. The molyb-
denum specimens will also be larger than those used in previous experiments
in order to permit standard mechanical-property evaluations.

4.1.3 Corrosion Effects of Carbon Tetrafluoride

As discussed in another section of this report (see Sec. 5.1), evi-
dence of a reaction between graphite and fluoride jons, leading to the
production of CF4, was indicated in a recent graphite~fuel salt in-pile
capsule experiment conducted at 1600 to 1800°F. If significant, such a
reaction would detrimentally affect both the chemical and nuclear prop=-
erties of the fuel salt. TIts effects might be clrcumvented, however, by
maintaining an overpressure of CF4 on the primary system. Since obvious
corrosion considerations lie behind this expedient, a series of tests
have been initiated to evaluate the effects of CF4 on MSRE core materials.

Three such experiments, incorporating specimens of INOR-8, Inconel,
nickel, molybdenum, and type 304 stainless steel, were recently completed
at 1112, 1292, and 1472°F, respectively. The tests were operated for 500
hr in 2-in.-diam nickel pots containing CF4 vapor at 6 psig. Weight
changes accompanying exposures at each temperature are compared in Table
4,5, None of the material showed significant reaction with CF4 at 1112

or 1292°F, At 1472°F, INOR-8 and nickel still exhibited little attack,
while Inconel, molybdenum, and type 304 stainless steel specimens showed
relatively heavy reaction films. Deep pits were also evident in one of
three molybdenum specimens in this test. (Some seal leakage was encoun-
tered during the 14T72°F test and may have contributed to the attack. )

Table 4.5. Weight Changes of Metal Coupons Exposed
500 Hr in 6 psig CF4 Vapor

 

Weight Gain (mg/cn®)

Type of Coupon IT5°F 1292°%F  1L12°F

 

 

INOR-8 (4 specimens) Maximmm +0.26 +0.28 + 0.59
Minimum +0.00 +0.05 + 0,22
Average +0,12 +0,25 + 0.40
Nickel (2 specimens) Average -0.02 +0.40 + 0.50
Inconel (2 specimens) Average ~0.03 +0.55 + 0.70
Molybdenum (4 specimens) Maximum +0.13 +0.00 + 3.4k
Minimum -0.02 ~0,04 -12.,57
Average +0.065 +0.00 - 0.11
304 85 (1 specimen) +0.0M +1.58 + 1.06%

 

fReaction film was subject to spalling.
78

Metallography confirmed in general the visual and weight-~change
observations., Attack on all specimens was manifested by the formation of
a thin, continuous, surface film, and, in the case of molybdenum and type
304 stainless steel, by surface pitting as well. INOR-8 specimens, which
are shown in Fig. 4.3, exhibited reaction films about 1/10 mil thick at

f UNCLASSIFIED |
i Y-44900 Lx.d
s A e R KRR O

Z

»

 

 

.007

 

 

. * . 1:009

ne

010

- . brrereee s aed

0l

 

 

 

- . - . 018

 

. e ST / // UNCLASSIFIED ®
e 't L. Y-44903 L%
Z

~ To e Cr e f s - T
R \

   

 

 

 

Q09

 

T .010

 

011

 

 

 

 

I s s b

5 . R " .

 

 

Fig. 4.3. Photomicrographs of INOR-8 Specimens Exposed for 500 hr to CF‘1 Vapor
at 1292 and 1472°F. (a) Exposed at 1292°F; (b) exposed at 1472°F. Etchant: 3 parts
HCI, 2 parts H2O, 1 part 10% Cr03.
79

all temperatures. No evidence of carburization of these or other speci-~
mens was detected metallographically (Fig. 4.3). Electron-diffraction
studies of the surfaces of INOR-8 and molybdenum specimens indicated the
presence of both fluoride and oxide reaction products, the latter un-
doubtedly resulting from COs impurity in the test gas.

Equipment is being assembled for a second series of tests in which
the combined effects of CF4 and fluoride salt will be studied., Static
pots containing CFs-saturated salt will be used for these experiments.
INOR-8 disks will be placed in both the liquid and gas-phase regions of
the pots and will be withdrawn through a gas lock after a series of test
intervals.

h,1.4 Examination of Corrosion Inserts from INOR-8
Forced-~Convection Loops

Information is being assembled for a topical report on the results
of INOR-8 forced-convection loop experiments conducted in support of the
MSEE. Data obtained from corrosion inserts located in the hot legs of
four of these loops were analyzed, and the results indicate an interesting
dependence of corrosion rate on time and temperature in these systems.

Table 4.6 summarizes the weight-loss data. (Note that there were
three inserts per loop and that the inserts were withdrawn at a succession
of time intervals ranging from 5000 to 15,000 bhr.) In two test loops
operated at 1300°F, weight losses after 10,000 hr were in the range 2 to
5 mg/cmg. Weight losses at 1400 and at 1500°F were only slightly higher,
being in the range 8 to 10 mg/cm® after 10,000 hr. Thus, when evaluated
from the standpoint of total metal lost from the wall, corrosion rates
did not vary appreciably with temperature (over the range studied). How-
ever, according to the metallographic analyses, the appearance of the
surfaces of the specimens that underwent corrosion losses was noticeably
influenced by the test temperature. Subsurface voids, caused by chromium
depletion, appeared to a depth of L mils in the 1500°F inserts; in con-
trast, inserts from the 1400°F and léOO°F tests displayed only a pitted
surface layer, 1/2 to 2/3 mil thick.

The apparent discrepancy between the metallographic and weight-loss
findings implies that corrosion losses involved not only chromium but
other components of INOR-8 as well. In fact, an analysis of the chromium
losses (based on diffusion studies) showed that their contribution to the
total weight loss in the 1500°F test, assuming selective removal of
chromium, could not have exceeded 1.7 mg/cm® in 10,000 hr.

The involvement of elements other than chromium indicates that the
chemicals leading to corrosion in these tests were relatively strong oxi-
dants, rather than the components of the fuel. It is further inferred
that the losses due to these impurity reactions were only slightly
affected by the test temperature.

The wall-thickness losses calculated on the basis of uniform surface
removal are seen in the Table k.6 to be relatively insignificant. Thus,
Teble 4.6, Results of Meazsurements on Corrosion Inserts Contained in MSRE Forced-Convection Loops

 

 

Test sglt: LiF-BelF-UF4-Thl,
Loop Number and Weight Loss per Weight Loss per Unit Eguivalent Total Loss
Maxirmum Wall Time Unit Area Arez per Unit Time in Wali Thickness
Temperature (br) (mg/cm®) (mg cm © month 1) (mil)
9354-L 5000 1.8 0.25 0.08
(1300°F) 10,000 2.1 0.15 0.09
15,140 1.8 0.08 0.09
MSRP-14 2200 0.69 0.23 .03
O-q‘\
(1300 £ 8LE0 3.8 O.3l+ .17
10,570 5.2 0.33 .23
MSRP-15 8770 11.2 0.93 0.50
IToToNd

(1400°F) 10,880 10.0% 0.67 i
MSRP-16 5250 G.6 1.3 43
(1500°F) 7240 9,0a 0.80 .36

 

a .
Average of two inserts.

08
81

chromium depletion such as occurred in the 1500°F test, in spite of being
a small fraction of the total metal loss, is the single most important con-
sequence of the corrosion process relative to mechanical-property effects.

4.2 WELDING AND BRAZING STUDIES
4.2.1 Heat-Exchanger Fabrication

Modifications were made on the design of the primary heat exchanger
of the MORE in order to improve the integrity and inspectability of the
tube-to-tube~gheet connections. A general discussion of the welded-and-
pack-brazed design proposed for these joints has already been presented.®

The weld-metal portion of the joint shown in Fig. 4.4 now has a con-
tour to minimize weld "rollover" in the tube bore, an undesirable con-
dition that constricts flow and complicates inspection of the brazed
joint. The new weld contour was accomplished by positioning the welding
electrode over the outside edge of the trepan lip instead of over the
faying surface between the lip and the tube., This method of welding
practically eliminates rollover while still giving adequate weld pene-
tration. The welding conditions that will provide the desirable pene-
tration (one tube-wall thickness) are listed in Table 4.7.

Table 4.7. Conditions for Seal Welding
of Tube~to-Tube~Sheet Joints

 

 

Previous Present
Welding current, amp L5 55
Welding speed, in./min 8.5 8.5
Electrode diam, in. 1/16 1/16
Arc length, in. 0.050 0.050
Electrode position Joint center Joint outer

diameter

Inert gas Argon Argon

 

Metallography of welds made using these conditions revealed the de-
fect shown in Fig. 4.5. Since this defect did not intersect the surface
of the weld, it was not found by the standard dye-penetrant inspection.
The photomicrograph illustrates the advantage of brazing between the tube
and the tube sheet to provide a backup seal for a welded Jjoint that can-
not be inspected adequately.
82

UNCLASSIFIED
ORNL-LR-DWG 65682R

    
   

A,

TUBE

DL

TREPAN GROOVE WITH

e
BRAZE SIDE /BRAZING ALLOY RING

  
 
  

)

  

L

A

SIS

.

   
 

VL g2 e 2

  

\TREPAN

(a)

Fig. 4.4, Present Method of Joining. (a) Before
and (&) after brazing and welding.

 

 

Fig. 4.5, A Large Pore in a Seal Weld, |llustrating the Importance of Back Brazing as a Backup Seal.
Etchant: 3 parts HCI, 2 parts HZO' 1 part 10% CrOB, 30X.
83

The braze portion of the joint was also modified to ensure the
presence of sufficient alloy to completely f£ill the annulus and form a
fillet. Calculations showed that the trepan should be deepened to accom-
modate this amount of alloy. Split rings of braze metal were used in
order to simplify assembly, and to ensure that the brazing alloy was re-
tained in the trepan when the tube sheet was inverted to make the seal
welds,

Experiments were conducted to investigate the influence of variations
in rates of temperature rise during brazing, using test assemblies similar
to that shown in Fig. 4.6. Rates of 75, 150, and 225°C/hr were used, with
no apparent variations in brazed-joint quality. The degree of bonding
along the 1—1/2—in. joint length approached 100% in every one of 16 joints
from four different test assemblies sectioned to date.

UNCLASSIFIED
U PHOTO 56641

 

Fig. 4.6. Typical Heat-Exchanger Test Assembly.
84

If a discontinuous fillet should be unintentionally obtained upon
brazing, it would be very desirable to determine the quality of the braze
in the Jjoint. Such a determination might eliminate the need for plugging
the tubes or possibly rebrazing the complete unit. Since previous work
indicated that ultrasonic inspection techniques would be capable of making
such an evaluation, their use is being investigated.

4.2.2 Remote Brazing

The conditions necessary for remotely rejoining MSRE piping compo-
nents are being determined in a cooperative program conducted by the
Welding and Brazing Group of the Metals and Ceramics Division and the
Remote Maintenance Group of the Reactor Division. This method of remotely
brazing INOR-8 pipe connections is proposed as a means for replacing
reactor components if the need should arise.

The sleeve-~type joint which appears Lo be most promising from the
studies conducted to date evolved from the steps shown in Fig. 4.7. De-
sign (b) succeeded the original design (a) because it eliminated problems

UNCLASSIFIED
ORNL-LR-DWG 64750

R

 

 

 

 

 

 

 

 

T < «
{0) STRAIGHT SLEEVE JOINT
WITH WIRE INSERT
zzzzzzz%zz? s
Ao
N NN
(&) TAPERED JOINT WITH
WIRE INSERT
(il /. \\;\SM_Q\
- D,
s .

 

(¢) TAPERED JOINT WITH
BRAZING ALLOY COATING

 

 

NN

 

(¢) TAPERED JOINT WITH
SPOT-WELDED, BRAZING
ALLOY, SHEET INSERT

 

 

T2 MW@ i T S

FFig. 4.7. Evaluation of Joint Designs for Remote Brazing.
85

in joint fit-up.® Design (c) minimized the necessity of obtaining excel-
lent alloy flow because alloy was already everywhere in the joint. The
last design (d) was a simplification of (c), since it required only the
cutting of the braze insert from brazing-alloy sheet and spot welding it
to the joint component.

An immersed pulse-echo ultrasonic technique was developed for the
detection of unbonded areas in the brazed tube-to-sleeve joint. Ultra-
sonic examination of a joint of the latest design (d) revealed various
unbonded areas. Three longitudinal metallographic sections were cut
through the worst of these areas, and the results were compared with the
ultrasonic-examination results. Correlation of the two methods was
extremely good, giving the ultrasonic inspection technique a high degree
of credibility. The measurements on these sections indicated that the
bornding was between 80 and 90% of the joint length. A typical photomicro-
graph of a portion of one of these joints is shown in Fig. 4.8 and illus-
trates the overall adequacy of the remote braze.

Inspection of the jolnt by remote methods will be more difficult. A
contact ultrasonic technique would not be feasible because a frequency of
25 Mc (hence a very thin and fragile transducer) is required to resolve

 

 

 

 

 

 

 

 

 

 

 

Fig. 4.8. A Typical Section of a Remote Joint Brazed with a 5Mil Sheet-Alloy Insert. Bonding on
this portion is about 87%, Etchant: 3 parts HCl, 2 parts H20, 1 part 10% Cr03.
86

reflections from the brazed area. Also, it would not be recommended that
the joint be immersed in a liquid couplant inside the reactor. Promising
results have been obtained, however, with a modified immersion technique
in which a container filled with water was used to couple sound into the
dry test piece through a rubber diaphragm. More development work with
this system 1s necessary before a conclusive application can be made.

The 1300°F shear-strength data presented in a previous reporta showed
that the INOR-8 joints brazed with the 82 Au-18 Ni (wt %) alloy exhibited
an average shear strength of 18,100 psi and a minimum of 12,500. Similar
specimens were tested at room temperature (Table M.B), and the results
indicate average Jjoint strengths of over 70,000 psi.

Table 4.8. Miller-Peaslee Shear-Strength Data
at Room Temperature

 

 

Base metal: INOR-8
Brazing alloy: 82 Au-18 Ni (wt %)
Brazing temperature: 1830°F
Brazing time: 10 min
Brazing atmosphere: Helium
Testing atmosphere: Air
Brazing Gap Number of Shear Strength
(mils) Specimens (psi)
0 1 71,100
2 1 73,000
L 1 72,600
6 3 13,700

 

4.2.3 Welding of TNOR-8

Assistance in the solution of weld-cracking difficulties on initial
welder-qualification tests was provided to the shops that will fabricate
INOR-8 components. The initial welds were definitely unsatisfactory, as
determined by bend tests., Severe cracking was observed in the root-bend
specimens, and, in some cases, they fractured in half. The material being
used for thege qualification tests had satisfactorily met the weldability
requirements specified in MET-WR-L and therefore were considered to possess
adequate weldability for reactor construction.

From a review of the situation it appeared that the major trouble
was associated with overheating of the metal during welding. Modifica-
tion of the Jjoint design and strict adherence to the welding procedure
provided an apparent solution to the problem. Welds are now being pro-
duced which exhibit no cracking in root-bend tests, and at least three
87

welders have demonstrated capability for producing welds of this quality.
Figure 4.9 illustrates the root and side bends of welds made before and
after adjusting the welding technique. The early weld exhibits the root-

bend fissures, while welds in the later bend tests are free from such
indications.

 

UNCLASSIFIED
Y-44825

LBTHJJ

SUNCLASSIFIED
Y-44824

 

 

Fig. 4.9. INOR-8 Qualification Weld Bend Tests.

tests of an unacceptable qualification weld,

(2) INOR-8 root and side-hend

showing root cracking in weld area. (b)
INOR-8 root and side-bend tests of an acceptable qualification weld, showing no cracks
or other detrimental signs.
88

4,2,k Mechanical Properties of INOR-&Dissimilar-Metal Welds

Procedural specifications for INOR-8-to-stainless steel (PS-35),
INOR-8-~to-nickel (PS-34), and INOR-8-to-Inconel (PS-33) were prepared and
qualification tests were completed. Specimens were prepared from the
qualification-test welds for room- and elevated-temperature mechanical
tests. All Joints were prepared by using the commercial nickel-base
filler material, Inconel 82, which is very versatile in the welding of
several dissimilar-metal combinations.® The results of the room~tempera -
ture tensile tests, using transverse specimens, are presented in Table
4,9, The room- and clevated~temperature tensile-~test results for samples
taken from the INOR-8-to-type-34T-stainless-steel Jjoints are presented in
Table 4.10. Tests were performed at room temperature, 1000°F, and 1200°F,
on transverse samples of this joint. All these welds exhibited satis-
factory integrity, and all failures occurred outside the weld-metal area
in the nickel, Inconel, or stainless steel base metal.

Table 4.9. Results of Room-Temperature Tests on
Dissimilar-Metal Welds

Filler metal: 1Tnco 82

 

Tensile Yield

 

Weld Strength  Strength Elongation
(psi) (psi) (% in 2 in.)
INOR-8 to Inconel 98, 200 52,900 29
INOR-8 to nickel 65,200 39,300 29
INOR-8 to type 347 - -
stainless steel 95,500 58,200 22

 

Table 4.10. Room- and Elevated-Temperature Tensile Tests
of INOR-8-to~Type-34T~Stainless~-Steel Dissimilar Welds,
Using Inco-82 Filler Material

 

 

Test Tensile Yield

Temperature Strength Strength Elongation
(°F) (psi) (psi) (% in 1 in.)
Room 95,500 58,200 22
1000 61,700 39,000 9

1200 51,700 35,800 8

 
89

Elevated-temperature creep tests on transverse samples from the
INOR-8-to-type~34T-stainless~-steel welds were tested at 1000°F and 1200°F,
and the results are shown in Table 4.11. Preliminary examination of the
specimens indicates that all failures occurred at the stainless steel-—
Inco-82 interface. Very little strain, as indicated by reduction in area,
was seen in these specimens. Those specimens which ruptured on loading
did so in the type 304 stainless steel-bage material.

Table L.11. Elevated-Temperature Creep Tests of INOR-8-to-Type-347-
Stainless-Steel Dissimilar Welds, Using Inco-82 Filler Material

 

 

Test Stress
Specimen Temperature Level Rupture Time
Number ( °F) (psi) (hr)
1-5 1000 60,000-105,000 Ruptured on loading
1000 55,000 186
1000 50,000 931
1200 39,000 13
9 1200 35,000 14
10 1200 29,000 | 128
11 1200 25,000 222
12 1200 22,000 320

 

4,3 MECHANICAL PROPERTIES OF INOR-8

A cursory examination has been made of the creep properties of cast
INOR-8. The results of the investigation are shown in Table 4.12,

In comparing the test results with data for wrought INOR-8 (ref 10),
it was observed that rupture times were considerably shorter and that
minimum creep rates were appreciably higher for cast material. The data
points are too few to permit a quantitative evaluation of these proper-
ties, but it appears that the difference is roughly equivalent to raising
the test temperature 100°F.

L.t EVAIUATION OF MSRE-TYPE GRAPHITE

A sample of CGB-X graphite, typical of the MSRE material, was evalu-
ated with respect to the requirements of the Molten~Salt Reactor Experiment.
The sample was cut from a treated and fully graphitized b-in.-diam cylinder.

Microscopic examination of a polished section from approximately the
central longitudinal axis of the CGB-X sample indicated that the pore
structure was not uniform but that pores were larger in some zones than
90

Table 4.12. Cast-Metal Creep Data

 

 

Minimum Rupture
Temperature Load Creep Rate Life
(°F) (psi) (%/nr) (hr)
1100 30,000 2.0 x 107° 816.0 RP
1100 20,000 3.3 x 107 2087 D
1200 30,000 2.6 x 10°= 138.0 R
1200 20,000 2,0 x 107° 430.0 R
1200 12,000 5. x 1077 2058 D
1300 20,000 L.8 x 1072 198.0 R
1300 12,000 7.1 x 1072 7h1.0 R
1400 12,000 1.9 x 10”7 198.0 R
1400 8,000 2.5 x 107% 2087
1500 8,000 2.1 x 10°° 162 R
1500 5,000 7.6 x 107° Lol R
1650 3,000 1.k x 1077 119.7
1650 5,000 3.2 x 1074 10.3 R
1800 3,000 5.0 x 10°° 329

 

®Annealed at 2150°F for 1/2 hr in hydrogen prior
to testing.

bR = ruptured.

CD = discontinued.

in others. The general structure of the CGB-X graphite appeared to be
intermediate between that of experimental graphites(grades B-1 and S-4
IB) previously tested.'?

A mercury-impregnation test used in lieu of a molten-salt permeation
test and outlined in the Specification for Graphite Bar for Nuclear Re-
actors, MET-RM-1, was conducted on an evacuated, transverse section from
CGB-X graphite. This specimen gained 0.44% by weight when exposed to
mercury for 20 hr at room temperature under a pressure of 470 psig. The
specified maximum limit is 3.5%.

Since permeation of graphite by molten salt is the preferred method
for determining whether it is suitable for the MSRE, three specimens from
the CGB-X sample were exposed for 100 hr to LiF-BeFo-ThF4-UF, (67-18.5-
14-0.5 mole %) at 1300°F under a pressure of 150 psig. An average of
0.02% of the bulk volume of the graphite specimens was filled with the
salt. The maximum acceptable limit is 0.5%. Some heterogeneity of the
91

sccessible pore spaces was indicated by the different degrees of salt
permeation in individual specimens.

The resistance of CGB-~X graphite to permeation by molten fluoride
was as good as that of best experimental graphite tested (grade B—l),11
despite the low bulk density of 1.83 g/cc. All the previous grades that
met MSRE permeation requirements had bulk densities greater than 1.87 g/cc.
Apparently, CGB-X graphite was fabricated in such a manner as to produce
maximm amounts of (1) small-diameter entrances to accessible pores and
(2) closed pores.

In an attempt to determine salt distribution in the graphite, radio-
graphs were made of thin sections from the salt-permeated specimens of
the standard fluoride salt-screening test reported above.

The thin sections were machined transversely and longitudinally from
specimens having the smallest and greatest amounts of salt permeation.
The specimen with the least permeation had a small, scattered amount of
galt at the original surfaces, with a maximum penetration below the sur-
face of approximately 0.0l in.

The sections from the specimen that had the greater amount of perme-
ation had the same type of salt distribution except for three small
pockets of salt. Two were at the surfaces and penetrated to depths of
0.02 and 0.0k in. The third pocket of salt was located in the interior
of the thin section, with the deepest salt penetration of 0.18 in. With
the exception of the pockets of salt, the salt distribution was similar
to that previously reported12 for grade B-l, the best experimental graph-
ite that has been tested.

This indicates that the heterogeneity of the pores is not severe.
It would be expected that this can be reduced or eliminated in the fabri-
cation of the smaller MSRE graphite bars.

The quantity of salt that may penetrate the accessible pores of
graphite is a function of pore-~size distribution and the pore-entrance
size distribution. The limiting pore-entrance diameter through which
salt can be forced for a given pressure, salt surface tension, and
wetting angle can be calculated by a relation cited by Washburn.*>

To evaluate the susceptibility to salt penetration for CGB-X graphite,
the pore~-entrance-diameter distributions of the accessible pores of two
specimens were determined by using a mercury porosimeter. The data indi-
cated that at approximately 65 psig, the maximum operating pressure ex-
pected in the MSRE, 0.12% of the bulk volume of the graphite should be
permeated by salt. (Under the conditions of a standard fluoride screen-
ing permeation test, 0.16% of the bulk volume of the graphite should be
permeated. )

The pore-entrance-diameter distribution indicated that most of the
accessible pores of CGB-X graphite had entrance diameters less than 0.3h
u. Since theoretically, the molten-salt pressure would have to be 265
92

psig to force salt through a 0.3h4-pu capillary, the pressure required to
permeate the major portion of the accessible pores would have to he
greater than 265 psig.

These are qualitative data because of the limited sampling (only two
specimens) of the graphite, the inherent limitations of the mercury poro-
simeter, the lack of accurate data on the surface tension of the salt,

However, the results should represent the worst conditions, because
the specimens were small and oriented in such a way as to expose the maxi-
mun quantity of accessible pores. The overall volume permeated by molten
fluorides in a full-scale MSRE graphite bar should be less than the quali-
tative, theoretical values above because of (1) better pore-space orien-
tation in the bar and (2) because the shallow permeation by salts as
described above in the radiographic examinations would be "diluted" by
the large salt-free core of the graphite bar.

L.4,1 Comparison of the Permeation of Graphite by
Mercury and Molten Fluorides

The permeation of a particular grade of graphite by a molten salt is
the preferred way to determine the suitability of the graphite to be used
in the special conditions found in a molten-salt reactor. However, this
would be a somewhat cumbersome quality-control test for a graphite vendor.
A mercury-impregnation test was proposed as a method that could be related
to the molten-salt permeation of the graphite. Calculations showed that
a U52-psig pressure on mercury at room temperature should cause it to
permeate graphite to the same extent as does LiF-BeFs-ThF4-UF, (67-18.5-
1%4-0.5 mole %) salt at 150 psig and at 1300°F in the standard salt-perme-
ation screening test. Tests have been made with various grades of graph-
ite to evaluate the permeability relationship between the mercury and salt.

In these tests, the specimens were evacuated to 50 ¢ Hg prior to
being submerged in mercury at room temperature; pressure of U452 psig was
applied and held for 20 hr. Similarly, pieces of the specimens were sub-
jected to the standard fluoride-salt-permeation screening test. The re-
sults are summarized in Table L.13.

Comparison of the bulk volumes of the different grades of graphite
permeated indicated that the mercury-impregnation test was approximately
equivalent to that of the standard screening test. 'The greatest differ-
ence between the percentage of the bulk volume permeated by mercury or
by molten salt was observed in CGB-X graphite. Here mercury took up more
volume than the molten salt. It is believed that this may be due to the
heterogeneity of the pore distribution of this grade and/or that the sur-
face tension assumed for the molten salt is low.

In the Specification for Graphite Bar for Nuclear Reactors, MET-RM-1
(5~10m61), the mercury pressure required is W70 psig, and the maximmun
permissible weight galn in the graphite specimen due to mercury impreg-
nation is 3.5%. On this basis only grades B-~1 and CGB-X would meet the
mercury requirement of the specification.
Table 4.13,

93

Comparison of the Permeations by Mercury and Molten
Fluorides into Various Grades of Graphite

Test Conditions:® Molten SaltP Mercury

 

Temperature, °F: 1300 70
Test period, hr: 100 20
Pressure, psig: 150 L52

 

Bulk Volume
of Graphite

Welght Gain
of Graphite

 

Grade of Dimensions of Specimens  Permeated (%Qd Permeated with
Graphite (in. )¢ Salt Mercury Mercury (%)
CGB-X 1.50 x 0.50 diam 0.02%  0.09° 0.6
CoB-X 0.125 x 1.50 x 1.50 0.01  0.05 0.k
B-1 0.125 x 0.9 diam 0.05 0.06 0.4
Sl -IB~D 0.75 x 0.4 diam 0.4

Sl -LB-D 0.125 x 0.9 diam 0.5 0.6 b.5
CS-112-8 0.125 x 0.9 diam 0.k 0.8 5.9
RHA-1 0.125 x 0.9 diam 0.6 0.7 5.6
S-l-LA-C 0.75 x 0.4 diam 0.8°

Selt-TA-C 0.125 x 0.9 diam 0.8 0.9 6.8
R-0009 1.50 x 0.50 diam 2.0° 2.0° 1.1
R-0009 RG 1.50 x 0.50 diam 2,7° 3,2 22,5
R-0025 1.50 x 0.50 dian 5.7 5.2° 37.2
AcoT 1.50 x 0.50 diam 13.9%7  1h,b3 115.8

 

aSpecimens were evacuated, submerged in the molten salt (or mer-
cury), and then the pressure was applied to the molten salt (or mercury).

Pryis was the TiF-BeFao-ThF,-UF, (67-18.5-14-0.5 mole %) mixture.

C . - -
Nominal dimensions.

dThe superscript number showvn on the figures in the column indi-
cates the nunber of values averaged; the absence of the superscript
denotes that only a single value is involved.
94

h.h.2 Removal of Oxygen Contamination from Graphite with
Thermal Decomposition Products of NH4FHE

Tt was previously r‘@ported,l4¢ that six graphite crucibles were purged
with the thermal decomposition products of NH4F-IF and then were exposed
to "oxygen-sensitive” LiF-BeF--UFs (62-37-L mole %) salt at 1300°F in
order to establish the effectiveness of the purge. (This salt readily
precipitates UOs in the presence of small quantities of oxygen.)

These crucibles have now accumilated L4000 hr of such exposure with-
out any UO, precipitate being detected by radiographic monitoring. This
indicates that the purge with the thermal decomposition products of am-
monium bifluoride was equally effective in removing the oxygen contami-
nation from the graphite crucibles even though the R-0025 graphite is a
moderately low permeability grade and AGOT is a relatively high perme-
ability grade. The tests were terminated for examination of the LiF-BeF o~
UFs mixture and the graphite crucibles.

The success of the preceding tests indicated that lower purging
temperatures might be successful. The lower temperatures would have the
advantage of being easier to attain in large systems and perhaps would
limit the reaction of the decomposition products with TNOR-8, the struc-
tural material for the reactor. Therefore, an R-0025 graphite crucible
was purged simllarly to the preceding tests except that the temperature
was T50°F. The crucible has since held the LiF-BeF--UF, melt at 1300°F
for more than 2000 hr, without any precipitate being detected. A similar
test 18 in progress in which an R-0025 graphite crucible was purged at
390°F.

OCther tests have been made by purging graphite at 1300°F with varying
quantities of NiH4F.HF. Two AGOT graphnite crucibles of the same bulk
volumes as the preceding were purged with 0.2 and 0.4 g of NH,F-HF,
respectively, for 20 hr at 1300°F. The purges appear to have been suc-
cessful in removing the oxygen from the graphite because the crucibles
have held the LiF-BeF»-Ul'y melt for 1000 hr at 1300°F without any detect-
able precipitate forming. Additional exposure time through LO00 hr is
planned.

h.h.3 Reaction of Zr0, with Thermal Decomposition
Products of NH4F.HF

An undesirable refractory sludge found in the MSRE Engineering Test
Toop™™ was tested to establish its composition and to determine a method
for removing it from the loop.

X-ray diffraction analyses indicated that the sludge consisted mostly
of Zy0»o (monoclinie) and Zr¥F4, plus some Belo and traces of other materials
that could have been fluorides. Chemical analysis showed the following:
W11 7r, 2k.6h F, 15.9% 0, 4.85% 14, 3.29% Be, and 0.01% Na.>®

One gram of the sludge was exposed for 20 hr to the thermal decompo-
sition products from 1 g of NH4F.HF crystals at 1300°F to determine

2
95

whether the oxygen of the ZrOp could be replaced by fluorine. This would
reduce the refractoriness of the sludge and make it easier to remove from
the pump bowl,

The particulate sample of the sludge reacted with the NH4F.HF decom-
position products during exposure, became molten, and wetted its Inconel
container. X-ray analysis of this material indicated that it was pri-
marily 2LiF«ZrF4, plus a moderate amount of 3LiF-4ZrF; and a small quantity
of unidentified material. There was no evidence that any monoclinic ZrOs
remained, indicating a maximum of 5 wt % of ZrOo.

The conversion of the refractory oxide sludge to the fluoride form

as described above should permit its removal from the loop by the use of
a molten-fluoride flush salt.

REFERENCES

lo Ro Bo Briggs et alo, I"/ISRP ngro Reptc Augo 31, 1961.’ ORNIJ"'3215,
p 99-100.

2. C. M, Blood, Solubility and Stability of Structural Metal Difluorides
in Molten Salt Mixtures, ORNL CF-61-5-4, p 23 (Sept. 21, 1961).

3. R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215,
p 96-99.

4. R. B. Briggs et al., MSRP Progr. Rept. Feb, 28, 1961, ORNL-3122,
W6
p -6,

 

 

5. R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215,
p 93-96.

6. H. G. MacPherson et al., MSRP Progr. Rept. July 31, 1960, ORNL-301lk,
p 63 ""65 °

 

 

 

T. X. V. Cook and R. W. McClung, Development of Ultrasonic Techniques
for the Inspection of Brazed Joints (to be published),

 

 

80 R. Bo Briggs et 8.1., MSRP PI‘OgI'o Reptn Al)go 31, 1961’ ORNIJ“BE]—-S,
pp 109, 111-12.

 

9. The International Nickel Co., Inconel Welding Electrode "182" and
Inconel Filler Metal "82" (April 1961).

 

 

10. R. W. Swindeman, The Mechanical Properties of INOR-8, ORNL-2780,
P 29-30.

 
11.

12,

13.

96
H. G. MacPherson et al., MSRP Progr. Rept. Apr. 30, 1960, ORNL~-2973,
P 53.
R. B. Briggs et al., MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 93.

H. L. Ritter and L. C. Drake, "Pore-~Size Distribution in Porous
Materials," Industrial and Eng. Chem., Analytical Ed. 17(12), 782
(Dec., 1945). —
R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 11k,
R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 5k,

Private communication from J. L. Crowley.
97

5. IN-PILE TESTS

5.1 INTERACTION OF FISSIONING FUEL WITH GRAPHITE:
TEST NO. ORNL-MTR-47-3

Since MSRE fuel and graphite are chemically inert and thermodynami-
cally compatible, with respect to each other, in the absence of radiation,
relatively extreme exposure conditions were used to accentuate the effects
of fissioning in the irradiation test ORNL-MTR-47-3.1 A power density of
200 w/cc was developed by adjusting the concentration of fully enriched
U235F, to 1.5 mole % instead of the <0.3 mole % to be used in the MSRE.
Consequently the four 3-week cycles in the MTR, accumulating 1580 hr at
power, gave a burnup of 8.5% in comparison with the 6% per year antici-
rated for the MSRE at 10 Mw.

Because so little is known about why some molten salts wet graphite
and some do not, and about changes in interfacial behavior that might
occur in a fissioning fuel, the primary purpose of the experiment was to
tell whether or not permeation of the graphite by fuel was to be expected.
Out-of-pile tests had indicated that it was not and that in clean systems
penetration occurred only in response to pressure in the manner expected
from the pore spectrum of the graphite and the surface tension of the
nonwetting liquid.

The contact angle of the fuel meniscus was chosen as a convenient
and reliable index of possible changes in wetting. A vertical blade of
graphite dipping into a pool of fuel in a graphite boat allowed room in
the same capsule for coupons of INOR-8, pyrolytic graphite, and molybdenunm.
The choice of a boat with a dished inside contour was influenced by the
need for a container which could accomodate freeze-thaw cycles without
stress. Capsules in previous experiments (47-1, 47-2)2 had ruptured from
freeze-thaw cycles. A step along a portion of the length of the bottom
or submerged edge of the blade extended to within 1/16 in. of the floor
of the boat; normal fuel, with a surface tension of almost 200 dynes/cm,
does not penetrate such a small crevice and thus another device for
detecting wetting behavior was provided.

5.1.1 Description of Experiment

As described elsewhere in greater detail,3 each of four sealed INOR-8
capsules, depicted in Fig. 5.1, contained a 3/16-in.-thick R-0025 graphite
blade (7.3 g) dipping into a shallow pool of fuel (11.4 g or 5 cc) held
in a graphite boat 3 in. long and 1-1/4 in. wide. Two of the boats were
of R-0025 graphite (75 g), a relatively impervious grade, and the other
98

UNCLASSIFIED
ORNL-LR-DWG 56754

METAL LOW -PERMEABIILITY

METAL S
SPECIMENS SPECIMENS GRAPHITE

 

 

O J’\\\\.'

 

 

 

 

 

 

 

 

 

 

 

 

 

5
- 77 0N
<

A
7

 

 

 

 

 

 

 

—MOLTEN- !
SALT FUEL GRAPHITE BODY

\THERMOCOUPLE WELL

0 Ya 4/2 3a 1 -
Lo b 0]
INCH

Fig. 5.1. MSRE Graphite-Fuel Capsule Test ORNL.-MTR-47-3,

two were of AGOT graphite (66 g) which had been preimpregnated with 9 g
of fuel. The capsules, horizontally aligned in a vertical diamond array,
were contained in a sodium bath which served as a heat transfer medium.
Thermocouple wells, which were affixed through one end of each capsule,
measured the graphite temperatures at a point roughly midway between the
bottom of the fuel and the capsule wall. Coupons of INOR-8, pyrolytic
graphite, and molybdenum were attached to the blade and were partially
submerged in the fuel.

The fuel was purified by treatment at 800°C with HF and with H,; oxide
was removed from the graphite by degassing in a vacuum at 2000°C. All sub-
sequent manipulations were carried out in a helium-atmosphere glove hox
to avoid contamination by the atmosphere, and several extra capsules were .
assembled at the same time to provide control specimens.

The design neutron flux was 2.3 x 10+3. This figure and 8.5% burnup
have been adopted for purposes of calculation since somewhat discrepant
data from dosimeter wires average to a lower value, while a comparison of
the ratio of uranium isotopes, available from mass spectrometry, indicates
a higher value.

The disassembly and initial observations were carried out in hot cells
at the Battelle Memorial Institute under the supervision of ORNL Reactor
Chemistry personnel, and later the separated pieces were sent to ORNL for
further study.

The wvapor pressure of the fuel increases by approximately a factor
of 10 for each 100° interval between 700 and 10000C, smounting to about
0.2 torr at 1000°C. At MSRE temperatures the vapor pressure is negligible,
and although the control capsules showed no evidence of distillation, a
considerable amount of vapor-phase transport (enhanced by the temperature
gradient toward the cool walls as well as by the high fuel temperature)
99

did occur in the in-pile capsules. The composition of the vapor seems to
be slightly to the BeFo-rich side of the stoichiometry for LisBeF4, and
although the vapor is much poorer in quadrivalent cations than the liquid
fuel, ZrF4 is present perhaps to an extent of a few tenths of a percent.

To a very good approximation, distillation of the fuel corresponds
to the removal of LisBeF4. If an estimated 2.5 g were lost from the lig-
uid fuel in capsules such as No. 15 or 16, leaving 8.9 g of a fuel, the
LiF~-BeF ,-ZrF,-ThF,-UF, proportions would have been altered from 69.5-23-
5-1-1.5 mole % to about 70-17-7.5-1.5-2 mole %. Several effects of the
distillation were noted.

The condensation occurred predominantly at cool metal walls, firmly
cementing the boats to the capsule walls in a manner that required that
the walls be sawed off in strips to open the capsules. This involved
some slight mechanical damage to the contents. FEnough distilled salt
was lost and fuel salt scattered to prevent obtaining a satisfactory mate-
rial balance on either the amount distilled or the amount remaining in
the fuel pool. Some condensation occurred in cooler ports of the graphite
and precluded interpretation of the weight changes of the graphite in terms
of liquid permeation. The concomitant change in fuel composition engen-
dered a crystallization path with LiF as the primary phase.

A considerable amount of the distilled salt was found in the form of
condensed droplets, resembling pearls, that evidently dislodged from the
capsule walls and fell onto or among the broken pieces of frozen fuel.

5.1.2 Dismantling of In-Pile Assembly

After irradiation was campleted, on July 27, 1961, the in-pile assem-
bly was shipped to the Battelle Memorial Institute Hot Cell Facility for
dismantling. The outer water Jjacket was sawed off, and the sodium tank
was cut by a machining device which avoided tumbling the capsule contents.
The sodium was melted under mineral oil, and the capsule assenmbly was
lifted out. The capsules appeared to be in good condition. The individual
capsules were cut loose from the sodium-tank bulkhead and the dosimeter
wires were removed. Micrometer measurements of the outside diameters of
the capsules showed no change from original dimensions. No bulging, bow-
ing, nor distortion was noticeable.

After drilling for gas sampling, each capsule was opened by cutting
through the welds on the recessed end caps with a specially designed cutter
in which the capsule was clamped in a stationary vise and the cutting tool
revolved around the work. The end caps were pried off with the help of
a prying bar and a split collar with a tongue seating in the groove milled
by the cutter.

As mentioned previously, the graphite boats were sealed tightly into
the INOR-8 capsules by salt which had volatilized into the narrow space
between them; therefore, the boats could not be pushed out of the capsule.
A remotely operable cutting machine was designed and built which permitted
the capsule to be clamped horizontally on a milling table and to be moved
100

past a side-milling cutter turning at low speed. Fach cylindrical capsule
wall was cut longitudinally into three sections which then were pried loose,
exposing the graphite boat.

Three identical unirradiated capsules, which had been thermally cycled
in a manner similar to the irradiated capsules, were likewise subjected to
gas sampling and then opened to provide blank samples of salt, graphite,
and metal with which to compare the irradiated materials.

5.1.3 Temperatures

Because of thermal convection in the sodium bath in which the capsules
were immersed, the upper capsule, No. 3, operated at higher temperatures
than its duplicate, No. 8, which was at the bottom of the diamond array.
Capsules 15 and 16 were cooler because of the absence of preimpregnated
fuel in the boat. The operating temperatures are given in Table 5.1. The
temperatures ran 40 to 45°C higher initially than at the end; almost half
the decrease occurred in the first 70 hr, possibly as a consequence of
improved. heat conduction as salt-vapor condensate accumulated in the gas
gap between the graphite and the capsule wall.

5.1.4 Gas Analyses

The nominal volume of the gas space in the sealed capsules, not count-
ing about 4.5 cc of voids in the graphite, was about 16 cc; the initial
amount of capsule gas, though not known, was estimated as about 17 cc
(STP) by assuming an average gas temperature of 50°C at the instant of
sealing.

Gas samples were obtalned by drilling the end of the capsules in an
evacuated chamber which enclosed both the drilling apparatus and the cap~
sule. Before drilling was started, leak rates were reduced to acceptably
low values. The released gas was transferred by a Toepler pump and an
associated manifold into sample bulbs equipped with "break-seals."

A condensed summary of the results from mass spectrometry is given
in Table 5.2. As reflected in amounts shown in the second column of
Table 5.2, some of the capsules gave samples with much-smaller-than-
expected pressures. The difficulty stemmed from a varying combination
of lack of sufficient Toepler-pump cycles, lack of adequate instrumenta-
tion, and lack of information on volumes of various parts of the system.
But, even after these points were at least partially remedied, there was
evidence for a slow rate of gas release from the capsules. A partial
blockage due to the pressure of condensed salt vapor may have contributed
to the difficulty in transferring the expected volumes of gas.

The Generation of Carbon Tetrafluoride

The relatively large amounts of CF4 shown in column 4 of Table 5.2
represent a noneduilibriuwn condition wnich had not been fully anticipated;
usually temperatures as high as the 800 to 900°C prevailing in the capsules
supply sufficient activation energy to prevenl the accumulation of species
Table 5.1 Time-Averaged Operating Temperatures and Identification of Irradiated Capsules

 

Capsule Average Operating Temperature (OC)

Number Graphite? Pretreatment Fuel, Max. Blade-FuelP Boat-Fuell Thermocouple
(estimated) (estimated) (estimated) in Graphite

 

 

15 & 16  R-0025 2000°C in vacuum 835 790 730 710
PN
8 AGOT Preimpregnated with 850 £ 810 750 730
9 g of fuel o
3 AGOT Preimpregnated with g5 00 850 825
9 g of fuel ;

 

8R-0025 graphite is relatively impervious compared to AGOT.

PInterface temperatures; no allowance for films at the fuel interface was made.

101
102

Table 5.2 Off-Gas Compositions in Volume Percent

Averaged values from mass spectrometric results obtained at ORNL and BMI

 

 

 

Standard ccd Components
Capsule No. Sampled AirP He CF, Xe Kr o,  Ar®
15 10 5.84 80.6 09.85 0.012 1.40 0.0k 2.29
16 19 22.75 61.45 8.73 0.003 1.1k4 4,73 1.62
8 7 5.36 76.9 2.32 11.82 1.96 0.11 k.20
3 17 L.57 79.6  0.67 11.43 1.78 0.10 0.96
Contro1d 10 7.38 87  <0.0003 0.17  5.5k4

 

aApparent volume of gas transferred from gas-release chamber.
bMainly inleakage to the evacuated release chamber while drilling the capsules.

CArgon was supplied as a blanket gas for the welding arc used in a helium-atmosphere
glove box to seal the capsules.

d'I‘he control was heated through cycles roughly corresponding to the thermocouple

reading vs time for capsules 15 and 16.

that are thermodynamically unstable toward reaction by several tens of
kilocalories.

The main source of the CF4 was the graphite-fuel interface. The main
sink for CF4 is the fuel. Most of the CF4 from the source was probably
consumed immediately by the fuel, but some fraction of the production dif-
fused into the graphite and thus was bypassed into a reservoir where it
was preserved until it could again come in contact with fuel.

The voids in graphite are interconnected and, of course, commmnicate
to the gas space above the boat. In the voids or the gas space, conditions
were not favorable for access of CFy to a reactive surface. Ehe frequently
encountered kinetic inertness of CF, in many of its reactions™ was also a
contributing factor. Since a higher concentration of CF was found in the
capsules (15 and 16) in which the generation rate was smaller, the consump-
tion rate evidently was controlling as far as the steady-state concentration
in the capsule gas was concerned.

Conditions which may have diminished the consumption rate include the
films of condensed salt vapor, essentially Li.BeF,, which coated the cap~-
sule walls. Hotter regions appeared to have been covered by a film
resembling a carbon or graphite deposit from the pyrolytic or radiolytic
decomposition of CE,. Even in the control capsules a film of black dust,
103

presumably graphite machining dust, accumulated on the surface of the fuel
ingot. In any case the consumption reaction 4id not proceed rapidly enough
to restore equilibrium conditions.

A possible reaction mechanism which accounts for the faster consump-
tion rate in the capsules that contained the prepermeated boats proposes
that CFy reacts most rapidly by means of three-phase contact (gas, graphite,
and fuel) which allows a heterogeneous reaction; this mechanism leans on
the fact that the reduction of CEy even by "unreduced" fuel is thermodynam-
ically favored. Regions of three-phase contact were much more abundant
in the prepermeated boats. The higher temperature of the fuel-graphite
interfaces in the prepermeated cases was also of importance in accelerating
the consumption reaction.

The consumption of CFs by dissolution in the fuel and subsequent homo-
geneous reaction is probably slow, but even for this mechanism, the area
of fuel-gas interface was greater in the prepermested boats.

The point of concern about the CE, generation is the removal of fluo-
ride ions from the fuel-—the reduction of the fuel as CF4 is carried away
in the off-gas. If this removal of fluoride occurs in a system with sub-
merged graphite, like the MSRE, the most readily recognized manifestation
of the reduction would be the conversion of Ufy to UFz. As the concentra-
tion of UF, increases, the disproportionation reaction 4UFs — 3UF, + UPY
leads to the formation of metallic uranium alloys with the container and
also to the formation of uranium carbides by reaction with graphite.

Although the presence of a measurable amount of reducing power in the
fuel from the capsules has not been satisfactorily confirmed, it is instruc-
tive to compare the amounts of CF4 accumulating in the capsules on the
basis of the calculated percentage conversion of UFs to UFs in the fuel.
This was done in Table 5.3; no allowance has been made for the anticipated
reduction due to the fact that the fissioning process produces a total
cation valence requirement greater than can be matched by the four equiv-
alents of fluorides from a gram atom of fissioned uranium.

Xenon

In Table 5.4 the amount of xenon found in each capsule is compared
with the amount expected for 8.5% burnup of the U®®. For the two pre-
impregnated cases, 3 and 8, the proper amount was found, but in the other
two cases only about 0.1% of the expected yield was found.

The missing xenon has not been located, but there is a possibility
that it was somehow selectively absorbed in some L feet of gum rubber
tubing that was used in the gas-collecting system. Also an attempt is
underway to analyze the irradiated graphite for xenon, but since the parts
have been exposed to air several months there is small probability that
the graphite analyses will resolve the question. There were no signifi-
differences in the isotopic distribution of the xenon recovered from the
four capsules.
104

Table 5.3 Calculated Conversion of UF, to UF,,
Based on CF4 Evolution

 

 

Average
Capsule Vol % CF,& Total VolumeP Yield Calculated®
Found (std. cc) (std. cc CFy) % U*" Reduced
3 0.67 23.5 0.157 0.45
8 2.32 21.1 0.49 .41
15 9.8 16.5 1.62 8.3
16 8.73 20.5 1.79 9.1

 

aAverage of BMI and ORNL mass spectrometer analyses.
Ppased on krypton analyses and 8.5% burnup.

CAccording to stoichiometry of MUF4 + C — CFy + UFs, including
impregnated UF4 in calculation.

Table 5.4 Comparison of Results of Xenon Analyses
With Theoretical Yield

 

Capsule Vol % Xe®  Total Volume® Total Xe Theor. Total Xe

 

Found (std. cec) (Std. cc) (std. cc)
3 11.43 23.5 2.69 2.86
8 11.82 21.1 2.52 2.83
15 0.012 16.5 0.0020 1.59
16 0.007 20.5 0.0014 1.61

 

aAverage of BMI and ORNL analyses.

bBased on krypton analyses.

c
Based on 8.5% burnup.
105

Ezzgton

The behavior of the xenon is even more puzzling in view of the fact
that the krypton yields were normal, both in amount and in the proportions
of the isotopes.

5.1.5 Test Effects on Graphite
Dimensional, Weight, and Electrical Resistivity Changes in the Graphite

Dimensionsg of the graphite boats and blades did not change within
the precision with which the measurements in the hot cell were made.
Changes greater than 0.1% should have been detectable.

Weight changes of the graphite parts were also not very meaningful,
since they represented the combined effect of distilled salt condensed in
the pores, broken or lost graphite, and of some sticking salt that was
difficult to remove. The blades of unimpregnated boats gained weight,
probably as a result of salt-vapor condensation. The preimpregnated
boats lost weight by distillation.

The electrical resistance of the graphite parts approximately doubled
as a result of the irradiation exposure. Such changes have been attributed
to the trapping of conductance electrons in defects induced by radiation.

Hardness of Graphite

 

Rockwell hardness measurements on the R-0025 graphite increased by
10% from 9 to 108, as a result of the exposure. The AGOT graphite boats
increased from about 56 to about 80 as a result of preimpregnation with
salt, and the readings after radiation were not significantly higher.
No differences were found between regions under the salt pool and else-
where in the same boat.

Wetting Behavior

 

When the pieces of the fuel ingot, jostled by the disassembly opera-
tions, were fitted into their original positions in the boats, the meniscus
was observed to be the same as in the control specimen and to definitely
show nonwetting behavior toward graphite; the metal coupons were wetted.
Fuel was not found in the 1/16-in. clearance between the blade and boat
but did fill the region under the blade where the clearance was 1/8 in.
The nonwetting behavior is receiving further confirmation from metallog-
raphy and autoradiography of the graphite, along with analyses of semimicro
cores drilled from the graphite. Information from weight changes of the
graphite, though uncertain, was at least indicative of no pronounced per-
meation of the graphite by fuel.
106

5.1.6 Analyses of Graphite

Autoradiography and Gamma-Ray Spectrometry

 

Cut sections of the graphite parts were autoradiographed with a beta-
ray-sensitive film to determine the overall distribution of radioactivity
in the bulk graphite. On the basis of information from autoradiography,
77 selected sites were drilled out, furnishing 1/32-in.-diam core samples,
which were checked for specific activities (gross beta and gamma) and
gamma spectra.

Unimpregnated Graphite.--The unpreimpregnated graphite seemed to have
not been penetrated by the fuel, although there was considerable radiocac-
tivity present. The principal gamma activity was found to be Zr-Nb°S> .
(frequently 106 dis min-1 mg-1), which appeared to have been dispersed by
distillation of the fuel; relatively high readings in the lower-temperature
portions of the graphite were fairly common. The beta activity, which was -
roughly paralleled by the Rul®®7109¢ getivity, generally diminished gradu-
ally with distance from the fuel interface; there were sharp lines on the
autoradiographs indicative of pronounced activity at interfaces, stronger
for liquid than for vapor exposure.

 

On the bhasis of experience in an earlier tests with fuel that did
not contain 7rF4, Cs*>% and Cs'®7 activity was expected in the graphite
at perhaps 10D dis min~t mg“l. Surprisingly, the cesium,activ}ty appears
to have been swamped; none was found, though 10” dis min-1 mg — for cesium
should have been detectable. The autoradiographs had a grainy appearance,
as though the activity had accumuwlated in finely dispersed but relatively
large pores.

Preimpregnated Graphite.--With the exception of the absence of Cel%%
and of ruthenium activity, there were no recognizable abnormalities in the
gamma. spectrum of the preimpregnated graphite. Both the beta and the Zr-
Nb 2 gamma activity were uniformly present at more than 100 times the
intensity encountered with the unimpregnated graphite. Attempts to under-
stand the failure to find Ce'** or ruthenium have been initiated.

 

5.1.7 Test Effects on Coupons

Metallographic examinations of coupons on INOR-8, pyrolytic graphite,
and molybdenum, which were attached to the blade, have not been completed,
but visual inspection revealed that the molybdenum had been severely
corroded, having lost about half its thickness, while the other specimens
appeared unaffected. INOR-8 wires binding the coupons to the blade had
become brittle.

5.1.8 Test Effects on Fuel

Petrographic and X-ray Examination

 

Except for discoloration the fuel appeared, under the optical micro-
scope, to have no excepltional features when the composition changes were
107

taken into account. Atbtempted x~-ray examination proved unfruitful because
of interference from background gamma radiation. Petrographic observation
established that fuel composition had shifted sufficiently that LiF was

the primary phase and that the main cause of the dark color was a brown
discoloration of LiF. The compound Li-BeF,, as expected, was also distin-
guishable in the irradiated fuel; scme of the crystals of this compound,
ordinarily uncolored, had a faint brownish-purple tint as a result of the
radiation. No oxide, no UFs, and no opaque materials were found. Attempted
comparisons of samples from the exterior and interior of the fuel ingot
revealed no evidence of segregation.

Chemical Analyses

 

Because of the bad effects which a strongly reduced fuel would have
on MSRE operation, the most important question posed for chemical analysis
was whether or not the fuel was reduced to the extent implied by the amocunt
of CF, produced. Although the feasibility of such a determination was
moot, an attempt was made and is still in progress.

To preserve the reducing power of the samples, single chunks or sege
ments representing a complete cross section of the fuel ingot were added
to the dissolver without grinding. Facilities for grinding, homogenizing,
and transferring powder in an inert atmosphere in the hot cell were not
available, but complete dissolution of the sample under an inert atmos-
phere was attainable in about 4 hr. Dissolution in HC1-HBO, solutions
under an atmosphere of helium or argon gave an evolution of hydrogen, the
amount of which was ascertained by mass spectrometry of the cover-gas.

Six dissolutions producing hydrogen at less than 2% concentration had
been carried out in a hastily contrived dissolver and gas colleclting appa-
ratus when results from mass spectrometry on the gas samples indicated
that difficulties had been encountered. Efforts to establish the correc-
tion for radiolytic hydrogen, which accounted for most, if not all, of
the yield led to discrepancies and erratic results on the apparent rate
of production of hydrogen; the trouble was attributed to the design of
the sweep-gas system and to the lack of a means of concentrating the recov-
ered hydrogen. Further dissolutions were postponed until improved
techniques could be developed and tested.

After a two-months development period an assembly was adapted in
which COp was used as the sweep gas, and the hydrogen was collected sbove
KOH solution in a gas burette from which any insoluble gases could be
directly removed for identification by mass spectrometry.

Results from dissolutions in the new apparatus are not yet available;
the earlier results can be tentatively interpreted as indicating that only
in capsule 16 (two samples) was there any evidence of reduced species,
and this is not certain. With the exception of O.l% CH, in the gas from
capsule 15, for which duplication has not yet been attempted, none of
the gas samples gave detectable amounts of hydrocarbons (limit of detec-
tion 0.01%) which were sought as evidence of the presence of carbides.
108

There were several reasons for selecting chunks or radially fractured
segments of ingots for analytical samples. This was the best way of pre-
serving the reducing power. Sampling for purposes other than chemical
analysis had already depleted half the ingot from some of the capsules,
and no complete ingots were available. Only untested grinding techniques
were available, and there was no evidence for, or reason to expect, exten-
sive segregation around the ingot.

The results of the analyses for major constituents, shown in Table
5.5, did not provide a good basis for conclusions regarding the overall
composition of the ingots and suggested that rather extensive segregation
has occurred. Six more samples, giving a total of twelve, are scheduled
for dissolution, but the results are not yet available.

Since the scatter of the results in Table 5.5 was somewhat larger
than expected, a program was initiated to validate the hot~cell procedures
for uranium analysis, using reference samples from a large batch of normal
(unenriched) fuel. A fuel which on the basis of previous routine analyses
on three different dates was believed to contain 5.8 + 0.2% uranium (mean
deviation) was used for reference samples that were reported by routine
facilities to contain 5.57% uranium and by hot-cell procedures to contain
5.8 + 0.1% uranium. The solution which was used in obtaining the 5.57%
value in the routine facilities gave a value of 5.79 when checked at the
hot-cell site. The implied bias, if any, between routine and hot-cell
analyses for uranium was regarded as negligible for ingot samples of pre-
sent interest.

Results of spectrographic analyses for corrosion products are shown
in Table 5.6. The chromium results, if interpreted as Cr®', were compara-
ble with, though slightly higher than, equilibrium values that have been
found for the isothermal corrosion of INOR~8 in the absence of irradiation;
they were at least an order of magnitude higher than expected for a "reduced"
fuel. In the absence of irradiation, INOR-8 corrosion equilibrium concen-
trations for Fe2" any greater than 20% of the Cr?' concentration have been
questioned; the iron concentrations in Table 5.6 were so high that efforts
to establish the possibility of accidental contamination are indicated.

As Tar as the direct interpretation of the analytical data on CFa
and. on corrosion products so far available were concerned, the fissioning
process wag oxlidizing. According to fission ylelds, the fissioning pro-
cess, even if slightly oxidizing, is not sufficiently so (for 10% burnup )
to correspond to the amount of CF, found in capsules 15 and 16.

Analyses of Fuel by Gamma-Ray Spectrometry

 

Attempts to compare the exterior and interior of the fuel ingot by
gamma~ray spectrometry showed that ruthenium was concentrated at the
surface, particularly at the gas interface. Cerium-1kl, barely detectable
in samples from the surface, was prominent in dissolved samples that in-
cluded interior portions. In all cases the predominant activity was the
mass-95 zirconium-niobium decay chain.
109

Table 5.5 Preliminary Results of Chemical
Analysis of Fuel Salt from MSRE Test

ORNL~-MTR-47-3

 

Component (wt %)

 

 

232§cga§£ U h 7r® Be® 1iP
Original batch® 7.19 6.08 10.4 L.57  10.5
Unéiii?iiged 6.004¢ 5.1 11.1 L.8% 9.6
Hypothetical® (8.3) (7.77)  (13.3) (3.3) (9.5)
Capsule 16 5.453 5.17 11.0 4.5 13.3
Capsule 16A 6.340¢  --D 10.8 4.8 9.4
Capsule 15 7.861¢ ..h 1h.5 5.4 1k
Capsule 3 11.703 8.6 11.5 4.1 15.2
Capsule 3A 6.519¢ ..h 11.3 - 3.2-3.5  15.2
Capsule 8 9.82 7.59 11.8 4.8 16.7

 

aSpectrographic determination.

Prlame photometric determination.

CAnalysis obtained from original preparation.

dAnalysis of portion of ingot from unirradiated control sample.

ePolarographic analysis of wranium (the other uranium snalyses
were by the coulometric method).

Tealorimetric determination gave 4.60% Be.

&Calculated approximate composition based on distillation of
2.5 g of LipBeFy from original 11.4 g of fuel salt and 10%
burnup of uranium.

hPrecipitation occurred in original solution before analyses

vere completed.
110

Table 5.6 Spectrographic Analysis of Corrosion Products

 

 

 

Corrosion Wt % of Corrosion Product Found in Capsule
Product No. 3 No. O No. 15 No. 16A
Cr 0.06kL 0.077 0.06 0.05
Fe 0.095 0.12 0.21 0.15
Mn <(0.014)&  <(0.013) <(0.007) <(0.006)
Mo <(0.11) <(0.11) <(0.05) <(0.05)

Ni <(0.1k) <(0.13) <(0.07) <(0.06)

 

a <(-~~) indicates below the detectable limit shown.

5.1.9 Conclusions

Due in part to the difficulties of hot-cell examinations, some aspects
of the postexposure examinations have not been satisfactorily culminated,
but the main objectives of the experiment have been accomplished.

The nonwetting behavior of fissioning fuel toward graphite has been
demonstrated, and the nature of radiation effects which might arise dur-
ing MSRE operation has been disclosed.

Only the evolution of CF,4 promises to be a potentially serious problem,
and this effect may have been greatly accentuated by the boat-and-pool con-
figuration chosen for the experiment. In any case the gas occupying the
voids in graphite exposed to fissioning fuel will contain several percent
of CF4; in reactor operation, however, the graphite is submerged, furnish-
ing a more favorable arrangement for the reaction of CF4 with the fuel,
and the loss of CF4 to the offgas in these circumstances should be consid-
erably less, or perhaps negligible.

5.2 MSRE IN~-PILE TESTING

The third MSRE experiment, ORNL-MIR-47-3 (refs 6 and T) was irradiated
in the MIR reactor from May 5 to July 24, 1961. The graphite used in the
experiment was in contact with fuel but not submerged in it. During the
postirradiation examination by Battelle Memorial Institute during October
and September, CF, was found in the cover gas over the fuel. As a result,
an irradiation experiment, ORNL-MIR-W7-4, has been designed to determine
whether CF, will exist in the cover gas over MSRE fuel when the graphite
is submerged such that CF, formed at the fuel-graphite interface must pass
through the fuel before escaping into the cover gas. Also, the experiment
111

will further demonstrate the compatibility of the fuel-graphite—INOR-G
system under thermal conditions at least as severe as those expected dur-
ing MSRE operation.

The L47-4 experimental assembly contains six capsules. Kach of four
larger capsules (Fig. 5.2) is 1 in. in diameter and 2.25 in. long and con-
tains a CGB graphite core (1/2 in. in diameter and 1 in. long) submerged

UNCLASSIFIED
ORNL~LR-DWG 67714R

Cr—Al THERMOCQUPLE-—-. NICKEL THERMOCOUPLE
WELL

NICKEL. POSITIONING LUGS {2)~—.

NICKEL FILL LINE ~ NICKEL VENT LINE

  

 
  
       

N
\— ~INOR -8 CAP
HELIUM COVER GAS

(3.5cm3) — N ; ‘
PUNCTURE AREA ; 4 A
FOR GAS SAMPLING” K| Jy==

- INOR~8 CAN

 

 

   
 

- TN

      
       
     
      
       
  
  

%
A
XX

 

MOLTEN SALT FUEL
(25 ¢g)

  
 
 

RN
R
XXX

%X
5

 
    
   
  

T
XD
8

  

N - Al
THERMOCOUPLE

7

 

 

 

 

 

Fig. 5.2, Submerged Graphite-Molten-Salt Capsules.,

  
   
  
 
 

 

CGB GRAPHITE

| TS
& &‘0’0
/ 0.0 %

 

~NICKEL POSITIONING L.UG

approximately 0.3 in. in ~25 g of fuel. The fuel will generate from 62
to T6 w/cc, depending upon the capsule location. The expected surface-
averaged temperatures at the graphite-fuel interface and the fuel-INOR-8
vessel interface are respectively 1340 and 1120°F. However, the INOR-8
used to fasten the graphite core is also expected to be in contact with
the fuel at 13400F. The two smaller capsules (Fig. 5.3) are designed to
study the effect of power density and temperature on the formation of CF4.
These capsules contain 1/2-in.-diam cylindrical crucibles which contain
fuel at 55 w/cc and 130 w/cc and are expected to operate at ~1320°F and
~1650°F respectively. These capsules, which contain exposed graphite in
contact with fuel, should also provide a basis for comparison with the
previous experiment to evaluate the effectiveness of graphite submersion
in preventing the net generation of CF4. The experiment is to be irradi-
ated in the MTR from March 12 to June 4, 1962.

Qut-of-pile tests with dummy capsules were utilized in the develop-
ment of a hot-filling method and apparatus to ensure that the salt properly
112

UNCLASSIFIED
ORNL-LR-DWG 67745

r-NICKEL POSITIONING LUG

 

 

re-—-PUNCTURE AREA FOR

 

 

HELIUM COVER GAS _ [ GAS SAMPLING
(24 ¢cm3)
7 >><1 l
SN |~ MOLTEN-SALT FUEL
K] - — (10 g)
A
HELIUM GAS GAP—T ] - Klig—CGB GRAPHITE

3
] CRUCIBLE

“ige 5.3. Crucible-Molten-Salt Capsule,

= INOR~ 8 CAN

 

 

 

 

 

 

 

>
152:’\ SNELITSIIRET I I

% 3
#
Y,

 

~ NICKEL POSITIONING LUG

 

covers the graphite and to minimize contamination of the fuel to be irradi-
ated. Also, the results from three types of thermal cycling tests on
dummy capsules indicate that the in~pile capsules should survive the pres-
sure stresses created by the expansion of melting fuel (caused by the
thermal cycling anticipated from MTR reactor shutdowns and startups dur-
ing irradiation).

Preliminary conceptual design is in progress for the fifth irradiation
experiment, ORNL-MTR-47-5. This experiment will be similar to 47-4 but
will be equipped with a system to purge the capsules to permit measurement
of the formation of CF, during operation.

REFERENCES

1. MSRP Progr. Rept. Aug. 31, 1961, ORNL~3215, p 117.

 

2. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 108.

 

3. F. F. Blankenship et al., Fuel-Graphite Irradiation Test, ORNL~h47-3,
ORNL~TM-118 (in preparation).

 

L. T. J. Brice, pp 432-33 in Fluorine Chemistry, vol. 1, ed. by J. H. Simons,

 

Academic Press, New York, 1950.
113

5. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 108.

 

6. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, pp 101-102.

 

7. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 117-18.

 
114

6. CHEMISTRY

6.1 PHASE-EQUILIBRTIUM STUDIES
6.1.1 The System LiF-BeFs-ZrF,

The ternary phase diagram for LiF-BeFs-ZrFs (ref 1) contains a pri~
mary phase field for LiF which persists to temperatures more than 400°C
below the melting point of ILiF (84L4°C). A valley lying just outside the
LiF phase field can be considered as a prototype of the MSRE fuel region.
The MSRE fuel composition, LiF-BeFo-ZrFu-ThF4-UF,s (70-23.7-5-1-0.3 mole
%; liguidus temperature, 441°C), contains only a little more than 1 mole
% of additional guadrivalent fluoride not found in the LiF-BeFo-ZrF,
ternary system, and on freezing the fuel, no ThFs or UF, appear in equi-
librium solids until perhaps half the liquid has solidified. Thus the
system LiF-BeFo~ZrF4 provides a useful facsimile of the initial freezing
behavior of the MSRE fuel.

New informalion on liquidus contours in the low-melting regions is
included in the expanded, but still incomplete, diagram for the LiF-BeFo-
ZrF4 system shown in Fig. 6.1, and values for invariant points are listed
in Table 6.1.

Of the three ternary systems involving LiF-BeFo, with either Zr¥,,
ThFs (ref 2) or UF4 (ref 2), only LiF-BeFo-ZrFs contains a ternary com-
pound. This compound, 6LiF.BeFs.ZrF,s melts semicongruently to LiF and
liquid at L75°C, and is the first phase to separate (at 441°C) on cooling
the MSRE fuel.

Single~-crystal specimens of this compound, when studied by x-ray dif-
fraction, were found to be body-centered tetragonal with ag = 6.57 A,
cy = 18.62 A, The space group was one that is uniquely determined by the
Laue symmetry and systematic absences, namely, Ihi/amd--D}.

A density of 3.19 g/cc was calculated for a crystalline fluoride of
this stoichiometry, on the assumption that the molar volume of the complex
compound is the sum of the molar volumes of the components. This value
is in good agreement with 3.06 g/cmS, calculated on the basis of the unit
cell dimensions and the presence of four stoichiometric units of LigBeZrFis.
In order to satisfy the space-group requirements, the Be and Zr ions must
lie in planes that are separated by 4.6k A and are perpendicular to the ¢
axis. The locations of the other ions will be the subject of further
studies.
115

UNCLASSIFIED
ORNL~LR— DWG 65852

  

  
 
    
    

 

i;g PRIMARY PHASE AREAS
@) 2LiF - ZeF,
® 3LiF - Z¢F,
© 6LiF - Ber,: ZrF,
/ N © LiF
® 2LiF - BeF,
® 3LiF - azrfy
/- ) @BQFZ
BLIF - BeF, - ZF, ® 2R
3LiF-4ZrF,
P 520
£ 507 AL N
LN
b N
~ ~
\ \
e ~ N, e
\ NN
S ~
% \\\\
2LIF - ZrF, 596
“
£ 570 d ~ "/
3LiF- ZrF, 662 ~ c e
F 2eF, 662 —_ \,(
~
B ’ \\
£ 598 A S
s \\\\
.
A NN\ ™~
O
,_-l\ @ g A
S ® N
0
0 £358°
e \\/ SN s N BeF,
648 K\ 2LiF-BeF -~ P 453 £ 370 548

6LIF - Bef, - ZrF,

Fig. 6.1. The System LiF-‘Ber-ZrF4.

Table 6.1. Invariant Equilibria in the System LiF-BeFo-ZrFg4

 

 

So0lid Phases Present gggposigi§2 (mol;réi gggzzizzt Tem%fg?ture
3LiF«ZrFys, LiF, 2LiF.ZrF, 5 5 20 Peritectic ~4180
LiF, 2LiF«ZrF,, OLiF.BeFoZrF, T 12 1h Peritectic 70
LiF, 2LiF.BeFp, 6LiF.BeFo.ZrFy 67 30 3 Peritectic ~35
2LiF«.BeFyo, 2LiF«ZrF4,

6LAF -BeF-ZrFy 64 32 I Peritectic 428
2LiF.BeFp, BeFp, 2LiF.ZrF4 L7 52 1 Eutectic 358

2LiF.Bel'p, ZrF4, BeFyo ~28 ~U5 2l Eutectic 470

 
116

6.1.2 The System LiF-BeFs-ZrFs-ThF,-UF4

An important composition section from the five-~component system
which contains the MSRE fuel is shown in Fig. 6.2. This section, estab-
lished with quenched samples from high-~temperature equilibrations, in-
cludes 2LiF.BelFs and the nominal fuel composition LiF-BeFo-ZrFg-Thka-UF4
(70-23-5~1~1 mole %). The dilution of the fuel with coolant (approxi-
mately 2LiF-BeF2), as a possible result of a leak in the MSRE heat ex-
changer, gives compositions that lie between the fuel and 2LiF.BeFs, as
shown in Fig. 6.2. The liquidus temperature varies linearly with mole
fraction between the fuel and the coolant. The presence of 1 mole of
coolant in 25 moles of fuel changes the primary phase from HLiF.BeFo.7rF4
to 2LiF.Bel».

UNCLASSIFIED
ORNL-LR-DWG 65684
525 -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TR
i \
g | i U;! —‘
\Q\‘ ) T !
S 500 : ; & |
o . D
J =
@ !
2 ‘
% 475 PRIMARY PHASE
a -LiF-
z L ..
w 6 LIF-BeF,-ZrFy - @
450 \‘ 2 L[F-BeFafi_"J/
‘ o
425 S
0 | UF, ; i . ’_ ) . "Lw_,,w_,
e ‘
ThFg ! 1 ‘
&" ‘1 T
RS {fizzé,,/~/*’// : o Fig. 6.2. The System LiF-BeF »ZrF -UF -ThF .
£
///'* | \ The section containing MSRE fuel and 2LiF-BeF2.
20 L. ‘ _
60
§° e
2 70 . UIF T
£ ST ; |
£ e
-—.«f"«‘w,"’
w“‘"-/fi“m"
80 ,,,,,,,,,,,, e N
0 5 10 15 20 25 30 33l

BeF, (mole %)

Distillation of the fuel, as exemplified in the high-temperature
irradiation described in Chap. 5, closely corresponds to the removal of
2LiF.BeF» from the fuel; the effect of this on phase behavior is also
shown in Fig. 6.2. Relatively little distillation occurs before LiF
becomes the primary phase and the liquidus temperature rises sharply.

6.1.3 Phase Equilibrium Studies in Fluoride Systems

Revisions and additions, discussed in more detail elsewhere,3 were
made to the phase diagrams for KF-ThF,, BeF-ThF,s, NaF-YFs3, NakF-ThF,-UbFg4,
NaF-BeF,-Tht',, Nab-BelFo-ZrFs, LiF-NaF-ThF, and CrF,~CrFa, and crystal
structure studies were carried out on LiSblg, Na¥ly, KsUF7z, KaTh¥y,
LiRbF5, and chromium(II,ITT) fluoride.
117

Correlations based on relative cation sizes lead to the expectation
that RbF.5rFp, CsF.CalFp, and CsF.BaFs could occur as stable compounds.
These expectations were confirmed, and some of the optical character-
istics of the crystals were established.*

6.2 OXIDE BEHAVIOR IN FUELS
6.2.1 Removal of Oxide from a Flush Salt

The use of gaseous HF containing 20% hydrogen to remove oxide from
a flush salt was demonstrated on a charge of LiF-BeFs-ZrFs (62-34-4 mole
%) for the Engineering Test Loop.® In the course of operation of the
loop, treatments of the charge with known amounts of BeO (~40O ppm) had
resulted in saturation with oxide.® Presumably 600 ppm of dissolved oxide
(a figure from chemical analysis) was present at saturation with ZrOs.
When neither temperature cycling nor an increase in the ZrFs concentration
from 1 to 4 mole % was effective in dissolving the oxide, an effort to
remove dissolved oxide from the fuel with HF was initiated. The equation
is 2HF + 0%  — 2F + Hz0?. To minimize corrosion of the Inconel drain
tank which held the salt mixture, 20% hydrogen was used to help maintain
reducing conditions. This concentration of hydrogen was adequate to keep
nickel and iron, but not chromium, reduced to the metallic statei7 ‘the
induced corrosion, Jjudged both by the 50~-ppm increase in the Cr&
concentration in the fuel, and by metallographic examination of dip legs
from the drain tank, was within tolerable limits.®

The amount of oxide removed from the melt at 565°C over a period of
70 hr, as plotted in Fig. 6.3, was monitored by estimating the amount of

UNCL ASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o~ ORNL-LR-DWG 66932
g 200 ]
° e
= HF - Hy RATIO: ~1/5 (vol) e
= 50 | H, FLOW RATE: 2.5 liters/min T
b TOTAL WEIGHT OF FLUORIDE SALT: ./
> 137 kg J
= e
2 100 - AT
7 -
i v
5 50 A~
o o
2 -
[ned
8 o////
Fig. 6.3. Removal of Oxides from Engineering é 0 10 20 30 a0 50 60 20
Test Loop by HF-H, Treatment at 1050°F. HF TREATMENT TIME (hr)

Hs0 in the off-gas. Since there was 176 kg of salt in the tank, the
estimated 215 g of oxide removed corresponded to a decrease in oxygen con-
tent of about 1200 ppm. Although the material balance on oxide 1s poor,
the procedure does seem to have been effective because less than 200 ppm
of oxide was found by chemical analysis of the salt after the HF treatment.
118

6.2.2 The Behavior of Sulfates in Molten Fluorides

The total oxygen in a fluoride melt can be comprised not only of
oxide ions but also of such ions as S04% and C02®". The case of 8047
is of particular interest since 5047 1is the most common form of sulfur
contamination in the starting materials for fuel preparation.

A complete removal of sulfur is essential for avoiding the very
detrimental sulfide embrittlement of nickel and nickel-based alloys, and
of course the oxygen accompanying the sulfur in sulfate must also be re-
moved. In the regular purification treatment, sulfates are reduced by
hydrogen to sulfides, and HF converts the sulfides to fluorides, with the
release of HoS in the effluent gas. The effectiveness of this procedure
has been demonstrated previously.8

To learn more about the behavior of 504% in fluoride-melt experi-
ments, 920 ppm of sulfur as LisS04 was added to a previously purified
mixture of LiF-BeFs (63-37 mole %) at 500°C. After maintaining the melt

UNCLASSIFIED
ORNL-LR-DWG 68597

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

450 é ‘ l
® L AT 500°C FOR ~1 WEEK
2
4001 ]
350 INITIAL SULFUR CONCENTRATION: < 3ppm —
\ SULFUR ADDED AS Li,SO, = 920 ppm
TIME AT EACH TEMPERATURE: ~ 24 hr
SAMPLED ~ 8 hr AFTER EACH TEMPERATURE

~ 300 ELEVATION fi
£
a
&
=
2 250 A
>
J
O
o
Z
e 200 - I SR
D
t
d
2
w

150 f(——— — B D

100

50 —d et e e 4 S S —
<3 ppm\
500 550 600 650 700 750 800 850

TEMPERATURE (°C)

Fig. 6.4. Apparent Thermal Decomposition of Sulfate lon in Molten LiF-
Ber (63-37 mole %) in Copper.
119

at 500°C for about a week under a continuous helium sparge, the tempera-
ture was elevated in 50° increments to 800°C. The welt was maintained at
each temperature level for approximately 24 hr, Eight hours after each
temperature increase, a filtrate sample of the melt was obtalned for
chemical analysis. The decrease in sulfur content, shown in Fig. 6.k,
indicates an apparent thermal instability of sulfate ion in the fluoride
solvent.

In a second experiment LipS504 was added to a purified mixture of LiF-
BeFo (66-34 mole %) containing 2 wt % uranium as UFg in an attempt to
verify the following reaction mechanism:

80,27 —» 0% + 803
UFy + 2027 - UOp + UF~

Following a 48-hr eguilibration at 500°C, the daily temperature in-
crease of 50°C, as in the first experiment, was repeated. Filtered samples
of the melt were analyzed for sulfur and uranium, and the effluent gas
stream was also analyzed by mass spectrometry. The results of chemical
analyses of the salt, illustrated in Fig. 6.5, show the anticipated re-
moval of sulfur and uranium from solution, but somewhat more sulfur was

UNCLASSIFIED
ORML~-L.R-DWG 68598

 

 

6000

5000 -1 2.5

{wt %)

4000 e e e e S R e e e e 2.0

 

3000 Al e 1.5

SULFUR REMAINING IN SOLUTION (ppm)
URANIUM REMAINING iN SOLUTION

1000 - : A e 0.5

 

 

 

 

 

 

 

400 500 600 700 800 200
TEMPERATURE (°C)

Fig. 6.5. Decomposition of L.i2504 and Associated Removal of Uranium
from Solution in LiF-BeF2 (66-34 mole %),
120

lost than prescribed by the stoichiometry of the postulated reaction
mechanism. Analyses of the gas, given in I'ig. 6.6, show the presence of
502 and Ho5 in addition to the anticipated SO0s.

UNCLASSIFIED
ORNL-LR-DWG 68599

 

millimoles)

{

SULFUR EVOLVED FROM SOLUTION

 

 

 

 

400 500 600 700 800 900
TEMPERATURE (°C)

Fig. 6.6. Gaseous Products from Sulfates in Molten LiF-BeF2 (66-34 mole %).

Table 6.2. Behavior of Metal Sulfates in Melts of
Purified LiF-BeF, (66-34 mole %) at G00°C

 

 

Equivalent Sulfur Found

Sulfate Sulfur Added in Solution
Added (ppm) (ppm)
115504 2881 823
NasS04 2232 1900
K504 1818 1810
MgS04 2631 1780
CaS04 2328 1240
PbS04 1045 1130
CuS04 1984 1400
NiS04 2048 1400
FeS04 2085 1650
Cro(504)3 2426 775
A15(S04 )3 2779 1830

Ces(S04)a 1601 936

 
121

In a third experiment the behavior of some metal sulfates in purified
LiF-BeFo (66-34 mole %) was observed visually at 600°C in glass equipment.
Filtered samples were teken from each preparation for chemical analysis.
The results are shown in Table 6.2. Of these mixtures, precipitation was
visually observed after additions of NiSO4, FeSO4, Cro(804)s, and Zr(S04)s.

Because of the relative complexity revealed by these exploratory
experiments, further studies will be required before conclusions can be
drawn.

6.3 PHYSICAL AND CHEMICAL PROPERTIES OF MOLTEN SALTS

A nunber of studies of the physical chemistry of molten salts, recent-
ly reported in some detail, provided the following summary paragraphs
which describe fundamental research of interest as a background for chem-
ical problems rising from the use of molten salts in reactors,”

All the published density data on molten fluoride mixtures were re-
examined in order to develop a more useful method of predicting densities
of related systems. To a good approximation, the molar volumes of these
melts were found to be expressed by an additive function of the components.
By using empirically adjusted values for the molar volumes of the compo-
nents of the melt and assuming additivity, the densities could be calcu-
lated to within 2% of the reported experimental values.

Densities of solid, complex, metal fluorides were calculated with an
average error of 5.5% by assuming that the molar volume of the complex
is the sum of the molar volumes of the simple fluorides that formed the
complex. This method of estimating densities facilitated the choice of
the nunber of molecules per unit cell for complex fluorides whose struc-
tures weres under investigation.

Further studies of freezing-point depressions of sodium fluoride
showed that trivalent fluorides cause negative deviations from ideality
[the smaller the solute cation size, the greater the deviation, and al-
kali fluorides caused mainly positive deviations (attributable to changes
in London dispersion forces)]. Alkali fluorides caused freezing-point
depressions in lithium fluoride somewhat different from those in sodium
fluoride because of differences in coulombic forces. Ixcess partial
molal free energies of mixing, calculated from liguidus temperatures in
the system NaF-LiF, were expressible in terms of concentration and two
constants.

The enthalpy changes from 87L4.0 to 0°C were measured with a Bunsen
ice calorimeter for samples of KF, LiF, and mixtures containing 0.193,
0.385, and 0.747 mole fraction LiF. The calculated molar enthalpies of
mixing when plotted vs composition gave an approximately symmetrical
curve, with a maximm at ~50 mole % LiF,

Previously reported equilibrium constants for the reaction,

NiFs + Hp ==2HF + Ni ,
122

at elevated temperatures provided a basis for refined calculations lead -
ing to an evaluation of /H°z0a.16 and estimation of AF° values at other
temperatures for the reaction with crystalline solid and (hypothetical)
supercooled liquid NiF, as reference states. From the true equilibrium
constant, obtainable directly from these free energy values, and the Ky
values, activity coefficients of NiFy have been obtained. These activity
coefficients and their changes with solvent composition, especially those
with supercooled NiF, as reference state, offer suggestions as to the
nature of such solutions.

The temperature coefficients of the association constants K,, Ko,
and Kjo for the formation of the species AgBr, AgBrgi, and+AggBr+ in
molten KNOs and, K; and Kz for the formation of CdBr , CaT , CdBrs and
CdIls- in alkali nitrate mixtures indicate that the entropies of these
associations are consistent with the "configurational"” entropy calculated
from the quasi-lattice model. The entropy of association of Ag with the
polyatomic ion CN™ is much more positive than for the associations with
monatomic cations. The influence of solvent on the association constants
for the formation of the associated species AgBr or AgCl in the molten
solvents NalNOs and XNOs is consistent with the reciprocal coulomb effect.

 

The perturbation theory of Reiss, Katz, and Kleppa for systems with
a common anion was extended to the third- and fourth-order terms to yield
the equation for the excess free energy of mixing of uni-univalent salt
mixtures.

‘M&“‘E‘%‘ = XaXoPOZ + Kife (X1 - X2)Q0% + (XaXeR + XaXz (X3 - Xp)78)8% + ...

where X; is a mole fraction of component 1, O is related to the ionic
sizes, and P, Q, R, and S are constants. The third- and fourth-order
terms are necessary to rationalize experimental measurements in molten
salts.

An estimate of the contribution of London dispersion energy to the
heats of mixing of mixtures of alkali nitrates with AgNOz or T1NO3 is
consistent with the observed differences between these mixtures and mix-
tures of alkali nitrates. This suggests a method of making estimates of
this effect in mixtures containing other polarizable cations.

6.4 GRAPHITE COMPATIBILITY
6.4.1 The Behavior of Carbon Tetrafluoride in Molten Fluorides

Recent postirradiation examinations of in-pile test capsules (ORNT~
MIR-4T-3)?© disclosed the presence of CF4 in the cover gas above graphite
boats containing MSRE fuel. Carbon tetrafluoride was absent in unirradi-
ated control capsules and represented a wmarked departure from thermodynamic
equilibrium in the irradiated capsules.
123

To study the kinetics of the reaction of CF4 with the fuel, the CF4
pressure over the fuel (LiF-BeFp-ZrF4-ThF4-UF,, 70-23-7-1-1 mole %) in
closed static systems was measured over a period of time. No evidence of
reaction with either normal or reduced fuel was found at 600°C. Since
the conditions for pressure measurements were not closely controlled and
a slow rate of reaction could have escaped observation, more refined
experiments were continued.

A gas mixture of CF4 and helium was recirculated continuously, bub-
bling through a 6-in. depth of melt at 600°C. Accurately measured volumes
of gas were initially admitted to the system, and gas samples were peri-
odically withdrawn for analysis by mass spectrometry. Based on the CFy-
to-helium ratio (about 40:60) in the analyzed samples, there again was no
discernible reaction in 25 hr at 600° with either reduced or unreduced
fuel,

Two factors make these results tentative rather than conclusive.
There seemed to be discrepancies in the gas analyses, and control samples
of the reduced fuel exhibited none of the characteristic phases, colors,
or reducing power expected on the basis of the amount of zirconium metal
added as the reducing agent.

More definite results were obtained from circulating a CF4 and helium
mixture through a salt mixture containing an indefinite amount of oxide.
Figure 6.7 shows that, on a per-mole-of-helium basis, 0.15 mole of CFy was
consumed, and 0.007 mole of COp was produced. Although the stoichiometry
of the reaction 1is quite puzzling (a black deposit, presumably elemental
carbon, was noted, and the amount of CO has not yet been determined), the
evidence that CF4 reacts with oxide in the fuel is encouraging.

UNGCLASSIFIED
ORNL-LLR-DWG 68600

0.60 0.020

0.56 oote 3|2

He

(&}

CO»o
He

MOLE RATIO —=

0.52 0.012

MOLE RATIO

0.48 0.008

0.44 0.004

 

0.40
0 10 20 30 40 50 60 70

TIME (hr)

Fig. 6.7. Apparent Reoction of CF4 with Oxide Contominants in MSRE Fuel
Salt Contained in Nickel at 600°C.
124

Experiments to determine the solubility of CF4 in the molten fuel
employed apparatus and techniques previously described in connection with
determinations of the solubility of noble gases in molten fluorides.*?t
Two determinations of the CF4 solubility were made by saturating the fuel
mixture at 600°C with CF4 at 1.3 atm for 6 hr. A portion of saturated
melt was transferred to an isolated section of the apparatus and stripped
with a known quantity of dry helium. Spectrographic analyses of the strip
gas failed to indicate the presence of CF4; however, a small measurable
guantity of COz was found. The estimated solubility of CF4 in the fluo-
ride mixture was assumed to be no greater than the value corresponding to
the COz found in the strip gas, that is, 1 x 107® moles of CF4 per cubic
centimeter of melt per atmosphere.

Additional determinations of the CF4 solubllity at higher tempera-
tures have not been completed. The stripping section of the solubility
apparatus has been thoroughly hydrofluorinated in an attempt to eliminate
the postulated reaction of CFa with oxide contaminants.

6.5 CHEMICAL ASPECTS OF MSRE SAFETY
6.5.1 Physical Effects of Mixing Molten Fuel and Water

The overriding physical effect of suddenly mixing molten MSRE fuel
and water would come from the steam pressure generated in the cell. A
study made for the Aircraft Reactor Test system included observations of
the gross effect of dumping or injecting a sizeable amount (~400 1b) of
salt composition No. 12 (NaF-LiF-KF, 11.5-46.5-42.0 mole %) into several
hundred gallons of water contained in an open vessel.’® These tests did
not include pressure measurements or analysis of gaseous products. The
uncertainties in calculating the rate of loss of heat to the MSRE cell
structure led to a study of additional methods of relieving the pressure
in case of an accident.

A small-scale laboratory study of the short-term effects of mixing
fuel (LiF-BeFo~ZrF4-Thf'y-UF,, T70-23-5-1-1 mole %) and water was undertaken
to provide information regarding possible reaction products and the
physical nature of the salt after its interaction with water.

The apparatus shown in Fig. 6.8 was constructed to permit the obser-
vation of the effects of injecting a small amount of molten salt into
water. The hazard involved was considered small since 2 g of salt (0.5
cal g~* °C™*, assumed) at 600°C would convert only 1 g of water to steam.
The salt (5 to 6 g) was loaded into the upper chamber (a 1/2-in.-diam x
3-in.-long, flanged, Hastelloy tube). A small furnace mounted around the
chamber was used to heat the salt to the desired temperature. Helium
pressure was then applied to force the molten salt into water (2 to 5 cc)
contained in an inverted nickel cone in the lower Pyrex-pipe chamber,
which contained an argon atmosphere. By means of a high-speed (3 in./min)
potentiometric recorder and a switching arrangement, pressure bulldup and
temperature variations at selected points could be followed.
125

UNCLASSIFIED
PHOTO 56520

 

Fig. 6.8, Salt Jet Apparatus.

Data from a recent experiment are shown in Fig. 6.9. The pressure
in the water chamber rose some 10 1b above atmospheric and remained there
after cooling. Therefore, the increase must have been due to the slight
126

UNGCLASSIFIED

ORNL-LR-DWG 68601
122

110 «F

THERMOCOUPLE NO. 3
BOTTOM TEMPERATURE

;

86

)

74

 

62

i
1

TEMPERATURE (°C

50 ~THERMOCOUPLE NO.1

WATER - CHAMBER

WALL TEMPERATURE
38 :

o6 o] THERMOCOUPLE NO. 2
VAPOR TEMPERATURE

 

0 2 4 6 8 10 12
TIME FOLLOWING INTRODUCTION OF MOLTEN SALT TO WATER (min)

Fig. 6.9. Temperature Variations in Salt Jet Apparatus Following Introduc-
tion of Salt at 650°C (5.51 g of Salt into 2 cc Water).

excess of helium used to effect salt transfer. The salt transfer occurred
in some 6 to 8 sec and that transferred into the water had an exploded
(popcorn) appearance but was unaltered in color. Salt flowing into the
nickel cone after the water had all been vaporized appeared solid and
dense, with a grayish film. In all experiments, finely divided droplets
of salt were found sprayed on the chamber walls and top. No marked pres-
sure increases were encountered in eight trials.

Petrographic examination of salt from a previous, similar experiment
showed a poorly crystallized material containing eutectic-~type growths
and exhibiting refractive indices typical of the fuel. Oxides were not
detected and if present must have been <1 p in size. The x-ray dif-
fraction pattern also indicated very poorly crystallized material--no
7r0ps, 7r0Fo, ThOo, or UOo. Following one experiment, the gas contained
in the water chamber after introduction of the salt was swept gently
through standard NaOH. Titration of the caustic indicated that some L
millimoles of HF or some substance which reacted with NaOH was formed.
This represents about 2% of the calculated yield for complete reaction;
in none of the experiments was any etching of the glass reaction chamber
noted.

Future experiments will include much larger volumes of water in order
to conserve the heat lost in the finely divided salt seen to spray through-
out the reaction charber, but the prospects for obtaining marked pressure
rise in this apparatus are poor, presumably because of the great diffi-
culty in obtaining sufficiently rapid heat transfer from salt to water.
127

Qualitative and quantitative analysis of the gaseous products is also
planned.

6.5.2 Solubility of Fuel-Salt Components in Water

Following a hypothetical accident in which molten fuel and water were
mixed in the containment cell it might be necessary to wait for weeks or
months before the radiation levels decrease sufficiently to permit the
approach of people for examination, decontamination, or repair purposes.
During this time even a slow rate of uranium leaching from the fuel could,
if the equilibrium solubility were sufficiently high, lead to a criti-
cality accident.

The solubility of uranium tetrafluoride in water is indicated to be
low at 25°C (1.6 x louéAl).la This uranium concentration is below that
necessary for criticality, but several factors in a real situation may
combine to give a higher concentration. The action of the oxygen from
alr or the action of peroxide generated by beta-gamma radiation in water
could oxidize the uranium to the very soluble hexavalent state. The
temperature may well be considerably higher than 25°C for a long time.
The acidity of the agueous phase in contact with the frozen salt may be
such as to increase the uranium solubility. Anions other than fluoride
may be present. Cations other than Uttt are certain to be present. A sol
or slurry of uranium-bearing solid could be produced, and it would create
the effect of a solution. For these reasons measurements of the solu-
bility of MSRE fuel salt in water at a variety of temperatures have been
made available for evaluating the probable course of events in the weeks
following the postulated accident.

To maximize the solubility and the rate of solution, a quantity of
nominal MSRE fuel salt (from the Fluoride Production Facility) was ground
to powder in a glove box so that it would present a large surface area to
the aqueous phase. Portions of it were added to flasks containing water,
and the resulting mixtures were stirred at controlled temperatures;
samples of the supernatant solution were taken at appropriate intervals
for chemical analysis.l4 Figure 6.10 shows that, at 25°C, the uranium
concentration rose, within one day, to approximately 0.002 m and then
slowly climbed to 0.00275 m in another week; the other components simi-
larly rose quickly to values which thereafter changed only slightly.
Figure 6.11 shows the solution concentrations after one-day exposures at
different temperatures, together with one value for the uranium concen-
tration obtained previously.l5 The sequence of temperatures is indicated
by the arrows in the figure; since the solution concentrations were not
reduced when the temperature was lowered, the equilibrium expected for
simple solubility behavior was not re-established on cooling. The tests
reported in Figs. 6.10 and 6.11 were made without particular attempts to
protect the mixtures from contact with the air. Although the flasks were
stoppered during equilibration at a fixed temperature, they were opened
to the alr vwhenever samples were taken. Consequently some oxidation of
uranium to the hexavalent state probably occurred.
128

UNCLASSIFIED
ORNL-LR-DWG 66935

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02 ; — —
Be ‘ : 1
0.1 T P .
0 S S, L e | 1 1 i
0.08 oy e i -
‘ | !
0.06 ¢ ;
S 004 L — |
2 0.006 ; -~
3 ;
|
s f
o, 0004 s
S ‘
S
2 0.002 ;
g ol . , I N L
E 0.45 - T , T e —
Li ‘ 5 i‘___....,
e ) eI Sa $ | . . . .
035~ * ‘ ’ ' T Fig. 6.10. Solution Concentrations of U, Li, Th,
: I | : :
oosd 1 | , P P | Zr, and Be Found upon Mixing MSRE-Fuel Solid
‘ : T - 1 | with HZO; Concentration vs Time at 25°C,
0.004 L{ . ‘ « , , o
® ® ® i B
0002 1 Ll ¢ : : | ;
{

2 3 4 5 6 7 8 9 10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0
TIME (days)
UNCLASSIFIED
ORNL-LR-DWG 66936
0.3 1
0.2 _
[OR ]
0.20
0.16
02
0.08
c | i
2 0.04 ‘ L ‘ |
3 [
5 008 — N ‘ ] e W
o MQWW & T i
S . B | |
o 004**'%?-—”%"“’"_'”#——” | . v ! T‘
g i } i
g O0— — : |
g 006 a—
¢ / J
0.05 ! . /"‘($ ’ ‘
. !
M"”‘—M . ‘
0.04 . % - - .
i : . [ e e \ i ‘v
FFige 6.11. Solution Concentrations of U, Li, 0.03 T . APPLE'S DATA J
Th, Zr, and Be Found upon Mixing MSRE-Fuel 0012 e e . T
. | - + ® ¢ ! >
Solid with H2O; Concentration vs Temperature, ‘ u ‘ ! ‘/‘ j

: i ) )
| APPLE'S DATA [

20 30 40 50 60 70 80 30 100
TEMPERATURE (°C)
129

Because of the appreciable uranium concentrations achieved in these
tests, neutron poisons will be provided in any water which might come in
contact with fuel. Since this provision has been established, solubility
studies in the presence of peroxide or radiation have been postponed.

6.5.3 Solubility of MSRE Coolant in Water

In experiments similar to those reported above for the fuel salt,
the solubility of the coolant salt was ourveybd A gquantity of 2LiF.BeFjs
conmposition (containing approximately 2 wt p exces LiF) was obtained
from the Fluoride Production Facility and ground to powder in a glove box
for use in the solubility experiments. TFigure 6.12 shows the analytical
results at 25°C, and Fig. 6.13 the results of experiments over a range of
temperatures up to 90°C. To provide information relevant to purification
processes for fluorides as well as to the behavior of coolant in contact
with water, a more thorough investigation of phase equilibria in the
system LiF-BeFo-H»-0 has been initiated.

UNCLASSIFIED
ORNL-L.R-DWG 66933

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

&z [FLUORIDE | , . |
25 e
té (—'j o (7} ‘
8% ol | J
§ Q06 T I ] 3 3
Fig. 6.12. Solution Concentrations of Li, Be, §§ 0.4 ,«f/ 5 B[ERL\:l;L:LLm I AAAAAAAAA A |
and F Found upon Mixing Li2BeF4 Solid with %éoz . ‘ x
HzO; Concentration vs Time at 25°C. ég \ l
° 0 5 7 10 15 20 25

TIME (days)

UNCLASSIFIED
ORNL-LR-DWG 66934

 

 

  
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30 [
% FLUORIDE ‘ ‘ ‘ [
- PERFECT, PROC. PENN. ACAD. 5C/.
o ‘ ¢6 54 19"3?) ‘
= T
8 2.0 A ,,,T, _______________ e ~- L
5 j T “‘%"* ~~ A
o A -
o 4 A e 4
8 iol “i"'—”"lf'—m‘ T T AT — ]
< 4 $
@ A
L
2 1 \
o 1
08 T
Be AND Li ’ ‘
= i
Q
I:j— A
3 4 -4_--— R —
2 § g
- s | ] Fig. 6.13. Solution Concentrations of Li, Be,
o e 1
S & Be (PERFECT’S DATA) and F Found upon Mixing Li,~)8e=:F4 Solid with
- o Li | .
Jo2l—aA —g | H,0; Concentration vs Temperature,
2 * 1
O e
20 30 40 50 60 70 80 20 100

TEMPERATURE (°C)
130

6.5.4 Partial Freezing of MSRE Fuel

A hazard could conceivably be produced upon partial freezing of the
reactor fuel while in the drain tank. If the freezing occurs at a suf-
ficiently slow rate, the first solids deposited would not contain uranium,
and the remaining solution, enriched in uranium, could cause the reactor
to be critical when only partly full.

To permit tentative calculations of the potential, though unlikely,
condition, the following hypothetical crystallization path of the MSRE
fuel}6 and estimations of density of the molten liquid remaining along
this crystallization path, have been proposed.

1. The compound 2LiF-.BeF, appears as a secondary phase after 20 wt
% of the fuel has frozen as 6LiF.BeFo+ZrFs, producing a liquid of compo-
sition LiF-BeFp-ZrF4-ThFse-UFs (68.9-26.2-3,3-1.222-0.366 mole %), density
2.246 g/cc at the first appearance of the secondary phase.

2, The compounds 2LiF-<Befs and 6LiF.BeFs.ZrF4 precipitate in a 1:1
molecular ratio until only 50 wt % of the fuel remains as a liquid of
composition LiF-BeFo-ZrFa~ThFs-UFy (67.7-28.2-1.52-1.98-0.595 mole %),
density 2.87 g/cc.

The segregation induced by slow freezing is potentially of great
value as a fuel recovery step since the fission products will be concen-
trated in the last liquid to freeze.

6.6 FLUORIDE SALT PRODUCTION
6.6.1 Production Rates

The use of HF-Hpo mixtures in place of alternating H- and HF treat-
ment in the purification of melts shortened the processing time. Further
gains expected from the use of strong reducing agents such as zirconium
metal have been temporarily thwarted by an unexplained and as yet unex-
plored corrosion of nickel equipment by melts which were treated with
zirconium metal,

Fused fluoride mixtures are currently purified in l.3wft3 batches,
and two batches per week are routinely obtained by simultaneous operation
of two ldentical rigs on a single-shift five~day week in accordance with
the following schedule, which takes advantage of off-shift hours for
intervals requiring minimal attention.
131

Ogeration Time ghrz

Charge and melt down 12
HF-Hs treatment 48
Ho stripping L2
Salt transfer )
Reactor cool-down for recharging 48

154

At the present rate of approximately 2.6 i per week, a minimum of
67 weeks would be required to produce the flush salt, fuel, and coolant
for the MSRE. TFor storing and transporting the purified products, some
50 shipping containers are available. Residual salts from previous usage
of these containers are being removed. OSince the size of a batch was
limited by the volume of the shipping container, each vessel is being
lengthened by 12 in., thereby increasing the production rate to approxi-
mately 4.2 £t2 per week.

To further accelerate the rate of salt production, consideration is
belng given to the addition of a separate furnace asseunbly adjoining each
batch facility. By proper scheduling, this arrangement would compensate
for 60 hr of waiting time (cooling, charging, and melting) per batch,
with an increase in production of about 64%. A preliminary engineering
study to obtain cost estimates Tor this possible modification is in
Progress.

6.7 ANALYTICAL CHEMISTRY
6.7.1 Oxygen in Nonradioactive MSRE Fuel

The method commonly used for the determination of traces of oxy§en
(as oxide) in the MSRE fuel is the fluorination method, using KBrFa. 7
This procedure has been in use for some time and, from the standpoint of
developmental evaluation of the method, has been satisfactory for concen-
trations of oxygen in excess of 400 ppm. (Application of the method to
practical problems in experimental melts has not yet been generally
achieved because of recurring questions regarding the cause of unexplained
variances in the range above 4OO ppm.) At concentrations below 40O ppm,
precision is unsatisfactory. The source of the poor precision is not
known. The lack of standard or representative samples makes definitive
studies extremely difficult. ©Satisfactory precision has been achieved,
however, using standard Uz0g samples.

In order to ascertain the source of possible error, alternative
methods of determining oxygen in the fuel have been explored. No spe-
cific method, other than the KBrF4 procedure, has been thoroughly evalu-
ated. An activation analysis method for oxygen, utilizing the
132

06 (3%,n)F'® reaction, has been under study. This reaction has been used
primarily for the determination of oxygen iIn lithium. Fluorine-18 has a
110~-min half-life, sufficiently long for this application; its positron
emission is measured by means of a scintillation counter.

Samples of MSRE fuel that had been analyzed by the KBrFa method were
analyzed by the activation procedure. The samples had been stored in a
water-free atmosphere to avoid contamination and subsequent hydrolysis of
the beryllium fluoride. The results indicated that no bias existed, and
there was an overall agreement of about 20% at a concentration of approxi-
mately 500 ppm. A chunk of salt (~5 g) was divided into two parts and
portions of each analyzed by the two methods. The agreement in this case
was within 5% at a concentration level of 1250 ppm. Comparative samples
will be analyzed in the future to ascertain orders of agreement at various
concentration levels and under various conditions. An important advantage
of the activation analysis method is that the fluoride salt need be pro-
tected from atmospheric contact only prior to irradiation. (The MSRE fuel,
due to the presence of Bel'», is particularly sensitive to hydrolysis by
contact with the atmosphere.)

Upon installation of a neutron generator, the fast-neutron reaction
0% (n,p)N*® will be examined for the determination of oxygen in the MSRE
fuel.
6.7.2 Adaptation of Analytical Methods to Radioactive Fuel
The feasibility of any of these methods for oxygen determination in
radioactive fuel is being studied. The proposed methods for analyses of

other constituents of the MSRE fuel are as follows:

1, Total Uranium: Controlled-potential coulometry is probably the most
satisfactory method. An alternative method is polarography.

2., Beryllium: FEmission spectroscopy. A colorimetric procedure is under
study as an alternate method.

3. Lithium: Flame photouetry.

Lk, Thorium: Amperometric titration.

5. Zirconium: Amperometric titration. The proposed method will give the
sum of zirconium and thorium. Zirconium must thus be obtained by

difference. FEmission spectroscopy is also available.

6. Corrosion Products—7Fe, Ni, Cr, Mo: Bmission spectroscopy. Colori-
metric methods are available if required.

T. Fission Products: Gamma-ray spectroscopy.
10,

11.
12,
13.

1k,
15.
16.

17.

133

REFERENCES

MSRP Quart. Progr. Rept. Feb. 28, 1961, ORNL-3122, p 111,

 

Phase Diagrams of Nuclear Reactor Materials, R. E. Thoma, ed., ORNL-
2548 (Nov. 6, 1959).

 

Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962, ORNL-3262,
chap. 1.

 

Tbid., last page of chap. 1.

MSRP Quart. Progr. Rept. Aug. 31, 1961, ORNL-3215, p 5hL.

 

MSRP Quart. Progr. Rept. Feb. 28, 1962, ORNL-3282.

 

C. M. Blood, Solubility and Stability of Structure Metal Difluorides
in Molten Fluorides, ORNL CF-6l-5-L (Sept. 21, 1961).

 

MSRP Quart. Progr. Rept. Feb, 28, 1961, ORNL-3122, p 121.

 

Reactor Chem. Div. Ann. Progr. Rept., Jan. 31, 1962, ORNL-3262,
chap. 5.

 

MSRP Quart. Progr. Rept. Feb. 28, 1962, ORNL-32832,

 

W. R. Grimes, N, V. Smith, and G, M. Watson, J. Phys. Chem. 62, 862
(1958). —

L. A, Mann and J. Y. Estabrook, ART Reactor Hazards Tests, ORNL CF-
55-2-100 (Feb. 11, 1955).

H. C. Nikolaef and Yu. A. Luk'yanchef, Atomnaya Energiya 11(1),
67-9 (1961). o

 

Analyses performed by the Analytical Chemistry Division, ORNL.
R. Apple, Analytical Chemistry Division, ORNL.,
Memorandum from H. F. McDuffie to R. B. Briggs, Data on MSEE Fuel

Salt Required for Nuclear Safety Calculations, MSR-61-21 (Feb. 13,
1962).

 

 

((}. Goldberg, A. S. Meyer, Jr., and J. C, White, Anal. Chem. 32, 31k
1960). —
134

7. FUEL PROCESSING

T.1 MSRE FLOWSHEET

Spent MSRE fuel will be fluorinated in Puilding 7503 to recover
uranium as UFg, and the waste salt will be stored in a tank at the reactor
site. After 90 days' decay, the fuel salt will be sparged with fluorine,
while still in the fuel storage tank. The UFg evolved will be passed
through an adsorbing bed of NaF held at 400°C for removal of chromium and
the bulk of the volatile fission-product fluorides, principally niobium.
The UFs will then be adsorbed on NaF at 100°C. The adsorbers will be
transported to the Volatility Pilot Plant for desorption and cold-trapping
of the UFg.

Excess fluorine will pass through the adsorbers and be disposed of
by reacting with charcoal or 80». (If not removed from the exit gas
stream, attack on the Fiberglas filters might result in the release of
radioactivity to the stack.) The testing of methods of fluorine disposal
is in progress. A method other than caustic scrubbing is desirable be-
cause of space limitations.

T.2 Fluorine Disposal Tests, Using Charcoal

oeven runs were made in order to determine the characteristics of
the charcoal-fluorine reaction and its possible use for a fluorine dis-
posal. system. An average of 5.89 g of fluorine was successfully combined
with 1.00 g of charcoal, which is 3% higher than the NASA figure for
adsorption.l The combustion reaction was nonexplosive and reached a
steady-state temperature of ~U50°C, independent of the fluorine flow rate
between 1.00 and 2.25 standard liters per minute of 100% fluorine.

Mass spectrographic analysis of the off-gas at steady state indicated
an average mole-percentage composition of 52.6% CF4; 23.4% No; 12.2% COs;
5.6 COFp; 2.1% CaFg; 1.3% of a mixture of CqFig, CgFi1s, C.Fis and higher
homologues; Ar; and Hp; with Fo always <0.01%.

Hydrogen fluoride, solids, and condensable liguids were trapped from
the off-gas before the mass-spectrographic samples were taken. During the
three longest runs an average of 0.846 g HF was produced per gram of char-
coal consumed; ~0.01 g of solids was deposited in the off-gas line per
gram of charcoal consumed; and ~0.1l g of liquid was trapped per gram of
charcoal consumed. The solids, of a resinous nature, contained 29.3%
wt % carbon and 4T7.9 wt % fluorine and melted over a range of 50 to 105°C
but decomposed before a boiling temperature was reached. The liquid con-
tained 54.0 wt % HIP, I3, 0% H20, and a small amount of high-molecular
weight fluorocarbons; the liquid had a density of ~1.25 g/cc.
135

Although the commercial charcoal used was not homogeneous, an average
composition was reported: T73.16% fixed carbon, 8.7T%h Ho0, 2.13% ash, and
15.92% volatile matter (all weight percentages). To reduce the HF, con-
densable liquids, and resinous solids entrained in the off-gas, two runs
were made with charcoal preheated to 650°C. The amount of HF produced
was reduced by a factor of 2; a negligible amount of liquid was noted,
but enough solid was formed in the off-gas line to plug it.

Studies using a thermogravometric balance indicated that the greatest
rate of weight loss occurred between 40O and 750°C. Heating the charcoal
to 1000°C removed almost all the volatile matter, the Hs0, and some of the
mineral-ash components.

Fluorination and thermal stress cauged the size of the charcoal
pieces to change from an initial ~1.0 in.” to an ash-charcoal residue
size distribution of T7.9% >2 mesh, 6.6/ between 2 and U mesh, 0.69% be-
tween 4 and 8 mesh, 1.23% between 8 and 25 mesh, 0.92% between 25 and 50
mesh, 1.31% between 50 and 100 mesh, and 0.84% <100 mesh (all weight
percentages).

REFERENCE

1. NASA Memorandum 1-27-59E, February 1959, p 8.
137

 

OAK RIDGE NATIONAL LABORATORY
MOLTEN SALT REACTOR PROGRAM

FEBRUARY 4 1962

 

 

R. B. BRIGGS, DIRECTOR

 

 

 

 

 

 

 

 

 

 

 

 

oA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSRE DESIGN : MSRE OPERATIONS
E.S. BETTIS R ! s E.BEALL®
! R.H.GUYMON*
ENGINEERING INSTRUMENTATION AND CONTROLS
P.H. HARLEY R R. L. MOORE 18C
H. R. PAYNE R J. R. BROWN 18C
A. N, SMITH R G. H. BURGER 13C
W. C. ULRICH R 0. W. BURKE* 13C
J. H., WESTSIK* R T. M. CATE 13C
L. V. WILSON R D. G. DAVIS 18C
P. G. HERNDON 13C
B. J. JONES 18C
J. W. KREWSON 18C
C. E. STEVENSON 18C
B. SQUIRES B&R
J. R. TALLACKSON 18C
MECHANICAL DESIGN R. WEIS cTS
W. M. BROWN R
J. 7. DICKIE R
J. K. DUNCAN R
W. C. GEORGE R
C. C. HURTT R
R. H. JONES R ANALYSIS
A. R.KERR R
J. 0. NICHOLSON R J. R, ENGEL R
L. E. PENTON R P.N. HAUBENREICH R
C. A. ROBERTS R C. W.NESTOR R
W. ROBINSON R
F.L.ROUSER R
W. G. STERLING R
1. ALWATTS R
FUEL PROCESSING
R. B. LINDAUER cT
SITE PREPARATION
G. B. CHILD E&M
W, L. DARIEUX E&M
N. E. DUNWOODY E&M
T. E. NORTHUP E&M
G. W. RENFRO E&M
W. E. SALLEE E&M
T. A. WATSON E&M

 

 

 

MSRE PROCUREMENT AND CONSTRUCTION

 

 

 

 

 

 

 

 

 

 

COMPONENT DEVELOPMENT

I SPIEWAK*
D. SCOTT, JR.
£

 

 

 

 

 

 

 

 

 

 

 

 

 

FLOW MODELS

R. J. KEDL™
B. J. YOUNG

)

 

 

 

 

 

 

 

 

W. B. MCDONALD R
PROCUREMENT AND FABRICATION
C. K. MCGLOTHL AN R
J. M. TEAGUE R
CONSTRUCTION
N. E. DUNWOODY* E&M
J. P. JACKSON E&M
B. H. WEBSTER R
L.P. PUGH R
SCHEDULING
J. P. JARVIS E&M
B. N. ROBARDS E&M
J. W. FREELS E&M

 

REMOTE MAINTENANCE

. C. HISE™

. BLUMBERG
P.HOLZ*

R. F. BENSON
4. R. SHUGART

oo m

WD AD

 

 

 

 

 

 

 

 

COMPONENTS

SCOTT, JR.
RICHARDSON
. SMITH
WEBER

. JONES

I»x0

onz

AWV OAD

 

 

 

 

 

 

LOOPS

. CROWLEY

. GALLAHER

R. E. CARNES

W. H. DUCKYORTH

-
w

AN DA

 

 

 

 

PUMP DEVELOPMENT

A. G. GRINDELL
C. H. GABBARD

P.G. SMITH

 

 

T

METALLURGY

ABOGADA

. COOK

. DEVAN™

. DONNELLY
. GILLILAND*

H
H
G
G
J. LEONARD™
K.
T
H
G
J
M

ROCHE*

. VENARD*
. WODTKE

. D. BRADY*

. L. GRIFFITH

. A. REDDEN

TTFTITIZFETFTETIXTX X

 

 

0

BURNS & ROE

CHEMICAL TECHNOLOGY DIVISION
CONTINENTAL TECHNICAL SERVICES

DI

ENGINEERING AND MECHANICAL D'VISION
INSTRUMENTATION AND CONTROLS DIVISION
METALS AND CERAMICS DIVISION

REACTOR CHEMISTRY DIVISION

RECTOR'S DIVISION

REACTOR DIVISION
PART TIME ON MSRP

 

REACTOR CHEMISTRY

 

 

 

 

 

W. R. GRIMES* RC
F. F. BLANKENSHIP RC
S. CANTOR RC
M. J. KELLY* RC
W. L. MARSHALL* RC
H. F. MCDUFFIE* RC
J. H. SHAFFER RC
R. E. THOMA* RC
D. R. CUNEO RC
F. A. DOSS RC

J. E. EORGAN RC
G. M. HEBERT RC

S. S. KIRSLIS RC

J. E. SAVOLAINEN RC
R. SLUSHER* RC
C. F. WEAVER* RC
W. K. R. FINNELL RC

B. F. HITCH RC

W. JENNINGS RC

R. G. ROSS RC

W. P. TEICHERT RC

RADIATION TESTING

C. A, BRANDON R
J. A, CONLIN R

 

 

 

 

 
139

ORNL-~3282
UC-80 — Reactor Technology
TID-U500 (17th ed., Rev.)

INTERNAL DISTRIBUTION

~

1. G. M. Adamson W7, P. R. Kasten
2. L. G. Alexander 48, R. J. Kedl
3. S. E. Beall L9, G. W. Keilholtz
L. C. E. Bettis 50. M. T. Kelley
5. E. S, Bettis 51. B. W. Kinyon
6. D. 3. Billington 52. R. W. Knight
T. F. F. Blankenship 53. dJd. A, Lane
8. E. P. Blizard 5L, C. E. Larson
9. A. L. Boch 55. T. A. Lincoln
10. S. E. Bolt 56. S. C. Lind
11. C. J. Borkowski 57T. R. B. Lindsuer
12, G. E. Boyd 58. R. S. Livingston
13. E. J. Breeding 59. M. I. Lundin
14. R. B. Briggs 60. H. G. MacPherson
15, F. R. Bruce 61, W. D. Manly
16. 0. W. Burke 62, E. R. Mann
17. D. O. Campbell 63. W. B. McDonald
18. W. G. Cobb 6. C. K. McGlothlan
19. J. A. Conlin 65. E. C. Miller
20. W. H. Cook 66. R. L. Moore
2l. G. A. Cristy 67. K. Z. Morgan
22, J. L, Crowley 68. J. C. Moyers
23. F. L. Culler 69. J. P. Murray (K-25)
2, J. H. DeVan 70. M. L, Nelson
25. R. G. Donnelly TL. C. W. Nestor
26, D. A. Douglas 2. T. E. Northup
27, J. L. English T3. W. R. Osborn
23. E. P. Epler . L. F. Parsly
29, W. K. Ergen ™. P. Patriarca
30. A. P. Fraas 76. H. R. Payne
31. dJ. H. Frye, Jr. Tf. D. Phillips
32. C. H. Gabbard 8. W. B. Pike
33. W. R. Gall 9. M. Richardson
34k, R. B. Gallaher 80. R. C. Robertson
35. W. R. Grimes 81. T. K. Roche
36, A. G. Grindell 82. H. W. Savage
37. C. S. Harrill 83. A. W. Savolainen
38, M. R. Hi1l 8k, D. Scott
39. E. C. Hise 85. H. E. Seagren
L0, H. W. Hoffman 86. E. D. Shipley
41, P. P. Holz 87. 0. Sisman
42, A. Hollaender 88. M. J. Skinner
43. A. S. Householder 89. G. M. Slaughter
Wy, 1. N, Howell 90. A. N. Smith
5. W. H. Jordan 91. P. G. Smith
L6, R. G. Jordan 92, A. H. Snell
93.
ol
9%5.

L aH
« & o

-

97.
98.
99.
100.
101.
102,
103.

*

O 9 e

.

163-16k.,
165,
166.
167.
168.
169-771.

Spiewak

D. Susano
A. Swartout
Taboada

R. Tallackson
H. Taylor
E. Thoma

B. Trauvuger
C. Ulrich
C. Watkin
M. Weinberg

D. F. Cope, AEC, ORO

140

10k,
105.
106.
107.
108-109.
110-113.

114-158,

159.
160-162,

J. F. Kaufmann, AEC, Washington
R. W. McNamee, Manager, Research Administration, UCC, New York

F. P. Self, AEC, ORO

J. H. Westsik

L. V. Wilson

C. H. Wodtke

Biology Library

Reactor Division Library

ORNL -~ Y-12 Technical Library,
Document Reference Section
Laboratory Records Department
Laboratory Records, ORNL R.C.
Central Research Library

EXTERNAL DISTRIBUTTION

Division of Research and Development, AEC, ORO
Given distribution as shown in TID-4500 (17th ed., Rev.) under
Reactor Technology category (75 copies - OTS)