ORNL-3369.txt

ORNL=3369
UC-80 — Reactor Technology

MO LTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT

FOR PERIOD ENDING AUGUST 31, 1962

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the

U.S. ATOMIC ENERGY COMMISSION

DISCLAIMER

This report was prepared as an account of work sponsored by an
agency of the United States Government. Neither the United States
Government nor any agency Thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or
usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product,
process, or service by trade name, trademark, manufacturer, or
otherwise does not necessarily constitute or imply its endorsement,
recommendation, or favoring by the United States Government or any
agency thereof. The views and opinions of authors expressed herein
do not necessarily state or reflect those of the United States
Government or any agency thereof.
DISCLAIMER

Portions of this document may be illegible In
electronic image products. Images are produced
from the best available original document.
Printed in USA. Price: $2.75 Available from the
Office of Technical Services
U. S. Department of Commerce
Washington 25, D. C,

LEGAL NOTICE

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

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

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

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

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

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

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

or his employment with such contractor.

| B A

ORNL-3369

Contracf No. W-7405-eng-26

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT

For Period Ending August 31, 1962

R. B. Briggs, Program Director

Date Issued

DEC 12 1962

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

e
- THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
iii

SUMMARY

Part 1. MSRE Design, Engineering Analysis,
and Component Development

1. MSRE Design

No significant changes in design concept or in detail of any component
or system were made. Design work was essentially completed, and a design
report giving all engineering calculations and analyses of the systems is
being compiled. The remaining design work involves primarily the incor-
poration of results of development work into the final drawings.

Work on the major building modifications is proceeding as scheduled,
- and the estimated completion date is October 15, 1962. A contract was
- (w awarded for construction work required outside the building, and this work
too is to be completed soon. It consists of construction of the exhaust
filter house, the cooling towers, the underground piping, and the inlet
filter house.

Materials procurement is on schedule with the exception of the core
graphite, delivery of which is now expected November 1, 1962. Fabrication
of components is approximately 50% complete and on schedule.

The basic instrument criteria and the control philosophy were estab-
lished. The layout of the instrument and controls system was established,
and the design of control-panel interconnection facilities is under way.

Designs for 30 of the 42 instrument panel sections required in the
MSRE system are complete. Fabrication of these panels by the ORNL shops
is under way.

) A freeze-valve test facility was completed and placed in operation.
./' .
A The selection of detectors for the process and personnel radiation
monitoring system is approximately 80% complete.

Procurement of a data-handling system was initiated. Specifications
were prepared and procurement initiated for 80% of the components required
in the process instrument system.

2. MSRE Reactor Analysis

Conceivable reactivity accidents were analyzed to permit evaluation
of reactor safety. Most of the calculations were done using Murgatroyd,
a digital kinetics program, but some analog analyses were also used. None
of the accidents analyzed led to-catastrophic failure of the reactor.
Undesirably high temperatures were predicted, however, for circumstances
associated with extreme cold-slug accidents, premature criticality during
filling, and uncontrolled rod withdrawal.
iv

Criticality and flux calculations were revised to correspond to the
latest fuel composition and the detailed design of the core. Calculations
were done for a core model consisting of 19 regions with different volume
fractions of fuel, graphite, and INOR-8. The clean critical concentration
for the reactor fueled with salt containing no thorium was calculated to
be 0.15 mole % uranium (93% UZ3%),

High-energy fluxes were also calculated. At a reactor power of 10
Mw, the maximum flux was about 1.3 X 1013 neutrons/cmZ.sec for neutrons
having energies greater than 1 Mev; for neutrons having energies greater
than 0.1 Mev, the maximum flux was about 3.7 x 102 neutrons/cmZ-sec.

_ Fluxes and power densities from the Equipoise calculations were used,
together with the detailed flow distribution in the MSRE core, to predict
the gross temperature distributions in the fuel and graphite. At 10 Mw,
with fuel entering the core at 1175°F, fuel in the hottest channels left
the reactor at 1261°F. Under the same conditions the peak graphite tem-
perature was 1296°F, A low-temperature region existed through the center
of the core where the fuel velocity was above average.

The nuclear average temperature, defined as that uniform temperature
which results in the same reactivity as the actual distribution, was cal-
culated for the fuel and for the graphite. Perturbation theory was em-
ployed, using coefficients from the Equipoise calculations. At 10 Mw,
with the fuel inlet at 1175°F and the mixed-fuel outlet temperature at
1225°F, the nuclear average temperatures for the fuel and graphite were
1213 and 1257°F, respectively.

Importance-averaged temperature coefficients of reactivity were cal-
culated for the case of fuel containing no thorium. The results were
A hs x 1075 and -T7.27 x 1072 (8k/k)/°F for the Tuel and the graphite,
regpeetively.

Power coefficients of reactivity were calculated from the nuclear
average temperatures and reactivity coefficients. This quantity, defined
as the reactivity adjustment required as the power is raised, depends upon
which temperature is used as the control parameter. If the mean of the
fuel inlet and outlet temperatures were held constant, the power coef-
ficient of reactivity would be -L4.7 x 10™* (8k/k)/Mw. If the fuel outlet
temperature were held constant, the power coefficient would be only
-1.8 x 107% (dk/k)/Mw.

Effective delayed-neutron fractions were calculated, taking into
account the delayed-neutron energies and the spatial source distributions
in a one-region model of the core. For a total yield of 0.0064 delayed
neutron per fission neutron, the effective fraction in the static core
was 0.0067. With fuel circulating at 1200 gpm, the effective steady-state
fraction was 0.0036. '

Preliminary tests were made of a new IBM-7T090 program for the analysis
of simulated power transients, based on a simplified space-dependent ki-
netics model. This program calculates fuel and graphite temperature

Iz i

7 A\

LY

&\
[ A

e
I“ &

L

distributions during transients and uses the nuclear-average-temperature
concept to relate temperature changes to the reactivity.

3. Component Development

Measurements were made on INOR-8 freeze flanges for 5-in.-diam sched-
4LO pipe under several different steady-state and transient conditions with
both an octagonal and an oval nickel ring gasket. The following conclusions
were drawn from these measurements: the flange joint design appears to
be satisfactory from an accumulated thermal stress standpoint after 43
cycles from room temperature to 1300°F and higher; the thermal distortion
of the flanges is small in the operating range and is not of a nature to
cause the flanges to separate at the seal; thermal cycling does not cause
the gas leakage rate to increase above an acceptable level; the flange
Joint will not form a frozen plug in the pipe under no-flow conditions if
the pipe heaters are kept on; the thermal and distortion properties of the
INOR-8 flanges are superior to those of the Inconel flanges tested previ-
ously; and thermal cycling of one flange did not affect its ability to
seal with a new flange,

A full-scale conceptual model of the control rod and drive unit was
assembled and tested under operating conditions for over 24,000 cycles.
Difficulty with & bushing was solved by replacement with bail bearings.
Over-all operation was satisfactory.

Evaluation of a removable 5-in. pipe heater insulated with hardboard
and mineral wool indicated excellent performance except for the tendency
to produce dust after high-temperature operation. Nuclear activation
analyses of insulating materials indicated that the dust would present a
problem of air-borne activity during maintenance. Another unit, con-
structed of foamed fused-silica, was better with respect to dusting but
had an excessive heat loss. An improved fused-silica pipe heater, incor-
porating some canned hardboard is being fabricated. An all-metal reflec-
tive type of unit was procured for evaluation.

The prototype cooling bayonets for removing fuel afterheat from the
drain tanks were thermally shock tested 384 times without failure. Reactor-
grade thermocouples were still intact after 6 cycles, whereas other types
of thermocouples failed in fewer cycles.

Detail drawings of essentially all the components and instrumentation
of the sampling and enriching system were finished, and construction of
a mockup was started. The sampling-capsule access-chamber mockup passed
the hydrostatic test requirements, and hot cell removal of a sample from
the capsule was demonstrated.

The full-scale MSRE core model was operated at 85°F with water to
make preliminary measurements of the velocity distribution, heat transfer
coefficients, and solids-handling characteristics. The velocity distri-
bution in the lower core wall-cooling annulus was found to be uniform
around the circumference, indicating that the flow into the lower head
was also uniform. The heat transfer coefficients in the lower head agreed
Vi

with the predicted values, except near the axis, where the measurements
were irregular because of the random flow. Preliminary measurements of
the solids distribution in the lower head indicated that solids which do
remain in the core tend to collect near the drain line at the center. It
appeared, however, that they would not plug the drain line.

The engineering test loop was revised to include a graphite container
and INOR-8 pipe and was placed in operation with a newly made zirconium-
free coolant salt. Information was obtained of the operating character-
istics of the frozen seal in the graphite container. It was found that a
salt mist existed in the pump-bowl gas space that led to some difficulties
with frozen salt deposits, but no such mist was found in the drain-tank
gas space. After 1500 hr of operation with the empty container, graphite /
was installed, vacuum pretreated at 1200°F, and prepared for flushing with
salt that had been treated with HF to lower its oxide content.

Design and construction of equipment for opening and closing freeze-
f'lange joints was completed, and procedure testing was started. A peri-
scope was purchased for use with this equipment. The design of a portahle
shield facility for semidirect maintenance operations is 60% complete.

Tt is a motorized unit with the appropriate openings and lighting for
specific operations.

Six prototype brazed joints were fabricated for quality evaluation
"using the equipment and general techniques required for remote operation.
A similar joint was removed from a loop for metallurgical examination
after 6780 hr at 1250°F and exposure to fuel salt for 123 hr.

The design of the overflow-line disconnect to the fuel pump was com-
pleted, and a carbon-steel prototype was built and successfully rough-
tested as a demonstration of the concept. An INOR-8 model was constructed:
for finel testing.

The design drawings for the coolant pump, the drive motors of the
fuel and coolant pumps, and the lubrication stands were approved. The
assembly of the rotary element and the fabrication and assembly of the
hot test stand for the prototype fuel pump were completed and testing was
started. Acceptable INOR-8 castings of impellers and volutes for the fuel
and coolant pumps and dished heads for the pump tanks were obtained. Fab-
rication of all components required for the reactor pumps was initiated.

A prototype model of the two-level conductivity-type liquid-level
probe being developed for use in measurement of level in the MSRE storage
tanks was constructed and is being tested.

Developmental testing of an element for continuous measurements of
the molten-salt level in the reactor pumps was continued. The maximum
variation recorded during a four-month test was 0.1 in., and there has
been no significant degradation in performance characteristics to date.

Investigations of single-channel thermocouple alarm switches for use
in freeze-flange and freeze-valve monitoring continued. A commerical
system is undergoing field .tests.

0
vii

Development of the mercury-jet-commitator thermocouple-scanning
system is continuing. Modifications of the switch were made to reduce
spurious noise generation. An alarm discriminator was designed and con-
structed. A complete system was assembled and tested.

Testing of mechanical attachments for use on radiator tubes in the
MSRE was resumed after modifying the test apparatus. The differential
between indicated inner and outer wall temperatures under simulated opera-
ting conditions was reduced to 9°F when the thermocouple was insulated
with Fiberfrax paper and Thermom cement was applied between the thermo-
couple and the tube wall.

Six Inconel-sheathed MgO-insulated Chromel-Alumel thermocouples are
being tested in 1200 to 1250°F air. The observed drift in all units
after 5000 hr of operation was less than *2°F,

Ten MSRE-prototype wall-mounted thermocouples were exposed to fast
temperature transients on the drain-tank bayonet-cooler testing facility.
One thermocouple failed immediately. The remaining nine units have with-
stood repeated raplid temperature changes.

Results of tests of thermocouples installed on an MSRE-prototype
freeze valve indicate that either 1/8-in.-OD duplex or 1/16-in.-0D single-
conductor thermocouples will measure wall temperatures on the freeze valve
within the required accuracy and will withstand the temperature transients
inherent in this application. No significant difference in the perform-
ance or durability of the units was noted.

A six-circuit radiation-resistant thermocouple-disconnect assembly
was fabricated and tested in a remote-maintenance facility. The discon-
nect is simpler in construction than the assembly described previously
and is more compatible with MSRE requirements.

Part 2. Materials Studies

4., Metallurgy

Tests were completed for determining the corrosive effect of CFa
vapor on INOR-8 immersed in fuel salt at 1112 and 1292°F. Results of
these indicate that CF4 vapor is effectively nonreactive toward INOR-8
at 1112°F but that minor attack may be promoted by CF; in the presence
of fluoride salts at 1292°F.

A final design and a welding procedure were established for the tube-
to-tube sheet joints of the MSRE heat exchanger. The final design pro-
vides for ultrasonic inspection. In order to demonstrate the welding and
inspection procedures, a 96-joint test assembly was successfully fabri-
cated and inspected by radiographic and ultrasonic Lamb-wave techniques.
Brazing procedures for remotely preparing INOR-8 pipe joints were demon-
strated by the fabrication of six joints. Only minor unbonded areas
were found in these by ultrasonic inspection.

L
viii

Additional mechanical properties data were accumulated for INOR-8,
including thermal fatigue data at 1250 to 1600°F. Stress relieving was
found to improve the elevated-temperature stress-rupture properties of
weldments. Evaluation studies of the mechanical properties of heats of
INOR-8 procured for MSRE construction were started.

Grade TS-281 graphite, considered representative of MSRE material,
was found to be more porous than experimentally made samples of similar
graphite; however, it met design requirements. Specimens of TS-281 were
permeated to 0.2% of the bulk volume in the standard permeation tests.
Oxygen contamination was purged from R-0025-grade graphite with the de-
composition products of NH4F.HF at temperatures as low as 392°F. Exposure
of INOR-8 specimens to this environment at temperatures varying from 392
to 1300°F resulted in minimum attack at 752°F.

Sintering studies were begun to develop procedures for fabricating
Gdz03 and Gdz03-Alz0sz cylinders for MSRE control rods. The shrinkage
characteristics and bulk density changes as a function of green density
were determined for Gdz0s pellets that were fired at 1750°C in hydrogen.

5. In-Pile Tests

Two molten-salt-fueled capsule assemblies, ORNL-MTR-LT-L and 47-5,
were built to study the formation of CF4 in the.gas over fissioning MSRE
fuel containing submerged graphite. In assembly 47-4, four capsules con-
taining a graphite core submerged in fuel and two withvgraphite crucibles -
containing fuel were irradiated.

Assembly 47-4 was irradiated from March 15 to June 4, and it is now
undergoing postirradiation examination at ORNL. The maximum measured tem-
perature of the capsules was 1400 # 25°F during approximately 1500 hr of-
full-power steady-state reactor operation.. Sixty. of the 121 recorded tem-
perature changes of ~30°C or more included decreases to the solidus tem-
perature of the fuel salt.

Two capsules containing submerged graphite cores were installed in
assenbly L47-5 to permit gas sampling during irradiation. Additional sealed
capsules designed to provide a range of oxidation-reduction levels by
altering the accessibility of chromium metal to the fuel were included.

The range of the ratio of INOR-8 surface area to graphite surface area in
contact with the fuel is O to 46:1., Assembly L47-5 is scheduled to be ir-
radiated from September 17 to December 10, 1962.

Postirradiation examination of assembly L47-4 revealed that elemental
fluorine at pressures as high as 35 atmospheres was generated in some of
the rapidly frozen salt used in tests of the effect of radiation on evo-
lution of CF4 from fissioning fuel and graphite. The fluorine, undoubtedly
of radiolytic origin, appears in the frozen salt after shutdown and is
possibly the main causative agent of the CF4 in the capsules that contained
large quantities of CF4. Since capsules of different configuration in the
same lrradiated assembly showed widely divergent behavior, the undesirable
products may be strongly dependent on some incidental effect, such as the
ix

freezing history of the capsule. Present indications are that the fluo-
rine problem will not constitute a severe threat to the reactor operation.
Tests are in progress (ORNL-MTR-47-5) to obtain direct confirmation of
the evidence for a negligible concentration of fluorine in the molten
fuel.

6. Chemistry

A detailed examination of the U50°C contour in the LiF-BeFs-ZrF, ter-
nary system suggests that 67 mole % LiF, 29 mole % BeFz, and 4 mole % ZrF4
should be a suitable solvent for UFs in a simplified fuel that contains
no ThFy. A fuel based on this solvent, LiF-BeFp-ZrF.-UF, (66.85-29-4-0.15
mole %), has an estimated melting point of 445°C. This composition has
a density of 2.15 g/cm® at the operating temperature. _

The eutectic 1n the system LiF-UF4 appears to be useful as a concen-
trate (27 mole % UF4; mp, 527°C) for fuel makeup. The amount and nature
- of segregation on freezing of MSRE-type fuels are tolerable.

Experiments to demonstrate the reaction of CF, with MSRE systems gave
positive indications, mainly at 800°C. Presumably, they would provide
more conclusive results at higher temperatures than the 600 to 800°C range.

Studies of the cleanup of graphite and of fuel continued and efforts
to improve analytical aspects of MSRE fuel technology were directed toward
oxygen determinations in fuel and in graphite and toward homogenization
of radioactive samples of fuel.

T. Fuel Processing

Work on the detailed design of the MSRE fuel-processing system was
started. The system provides for oxide removal from the salt and uranium
recovery. A simplified flowsheet was prepared.

8. Engineering Research on Thermophysical Properties of Salt Mixtures

The liquid-state enthalpy and the viscosity of a LizsC0a-NasCOz-KCOa
(30-38-32 wt %) mixture proposed for use in out-of-pile studies relating
to the MSRE were experimentally obtained. Unusual scatter in the data
warranted only a linear fit of the enthalpy data. The derived heat ca-
pacity over the temperature span 475 to T715°C is therefore given by the
constant value 0.413 cal/g-°C. The kinematic viscosity was found to vary
from 16.5 centistokes at 460°C to 3.1 centistokes at T15°C. An estimate
of the mixture density as a function of the temperature allowed prediction
of the absolute viscosity.
THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
xi

CONTENT'S

------------------

PART 1. MSRE DESIGN, ENGINEERING ANALYSIS,
AND COMPONENT DEVELOPMENT

1. MERE DESIGN .....ciitiiieiniiierennnennnnns

Design Status ....... . i,
Reactor Procurement and Installation ......
Major Modifications to Building 7503 ...
Construction QOutside Building 7503 .....
Procurement of Materials ...............
Fabrication of Components ..............
Reactor Instrumentation and Control Systems

------------------

------------------

------------------

------------------

Layout of Instrumentation and Control System .............

Design Status ........ ... i,
Procurement Status ........ i v

2. MSRE REACTOR ANALYSIS .....cvvivinnenennnn.

Nuclear Accident Analyses .........ccieiuunn
Calculational Procedures .........oeveee,

------------------

------------------

------------------

------------------

Results of Fuel Pump Power Failure Analysis ..............

Effects of Cold-Slug Accidents .........
Results for Filling Accidents ..........

------------------

Effect of Uncontrolled Withdrawal of Rods ................

Effect of Graphite Movement ............
Results of Concentrated Fuel Addition ..
Reactor Statics Calculations ..............
Criticality Calculations ............ ...
Flux Distributions ............ ...
Power Density .......... ... i,
Steady-State Fuel and Graphite Temperatures
Temperature Distributions ..............
Average Temperatures .............c.....
Reactivity Coefficients ...................
Temperature Coefficients ...............
Power Coefficient .......... ... ... ... ...
Effective Delayed-Neutron Fractions .......
Reactor Kinetics Code Development .........

3. COMPONENT DEVELOPMENT ...........c.oieuinine...

Freeze-Flange Development .................
Freeze-Flange-Seal Test ................
Thermal-Cycle Test Loop ................
Freeze Test in Thermal-Cycle Test Loop .

------------------

oooooooooooooooooo

------------------

------------------

------------------

------------------

------------------

------------------

------------------

------------------

------------------

------------------

------------------

------------------

Test of a Freeze Flange in the Prototype-Fuel-Pump

Testing Facility ....... ..o,
MSRE Control Rod ...ttt nnnennnan

------------------

------------------

iii
Heater Tests .......
Pipe Heaters ....

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

-----------------------------------------

Activation Analysis of Insulating Materials ..............

Drain-Tank Coolers .

-----------------------------------------

Sampling and Enriching System ....... .. it

Detail Design ...

-----------------------------------------

Mockup of Sampling and Enriching System ..................

Core Development ...
Model Operations

-----------------------------------------

-----------------------------------------

Velocity Measurements .........ciiiiiiiireninrinninnnnas
Measurement of Heat Transfer Coefficients ................
Experiments on the Behavior of Solids ........... ... ...
MSRE Engineering Test Loop (ETL) . vviiiiiteerinennnnneenns
Frozen-Salt-Sealed Graphite-Container Access Joint .......

Salt Composition
HF Treatment ....

-----------------------------------------

-----------------------------------------

Cold-Finger Collections in Cover Gas ...... . ..ot iiivnnnn
ETT, Graphite Treatment ...ttt ittt tiontianeennnros

MSRTE Maintenance ...

-----------------------------------------

Placement and Removal of Freeze-Flange Clamps ............
Miscellaneous Flange-Servicing Tools ..........coiviiivenn

Remote Viewing ..

-----------------------------------------

Portable Maintenance Shield ........c.ii ittt inntnnenenns
Brazed-Joint Development .. ...ttt i i
Brazed-Jdoint Fabrication .....c. it ittt ineteineneeoonns

Brazed-Joint Salt

Y= v

Mechanical-Joint Development ......... ... .. i,

Pump Development ...

-----------------------------------------

Fuel Pump Design and Fabrication .......... ... .
Coolant Pump Design and Fabrication ............ ... v
Design and Fabrication of Lubrication Stands for Pumps ...
Lubrication-Fump IEndurance Test .............. ... v
Drive Motor Deoign and Fabrication i..:i:iivisssisisssivessas
Test Pump Fabrication ..........ciiiiiiiiiiiiiiiiennns
Test Facility Construction ......... ... BT
Hydraulic Performance Tests ........cc i,
PKP Pump Hot Endurance Test ... i,
Test Pump With One Molten-Salt-Lubricated Bearing ........
Instrument Development ......... ... ittt innnnanns
Single-Point Liquid-Level Indicator ............ . ciiuie.n..
Pump-Bowl Liquid-Level Indication .............ciiiie..n.
Single-Point Temperature Alarm System ....................
Tenperaturc DeamIeT . ...ttt it it i i ettt i e e
Thermocouple Development and Testing ........... ... . ...,

PART 2. MATERIALS STUDIES

METALLURGY .........

Corrosion Effects of

-----------------------------------------

CF L o e s

Welding and Brazing Studies ........iiiii ittt
Heat Exchanger Fabrication ............ it nans

Remote Brazing ..
Welding of INOR-8

-----------------------------------------

----------------------------------------
-

46
46
47
48
49
49
49
49
49
50
51
52
53
53
54
56
56
58
59
59
60
60
60
61
61
62
62
62
62
62
63
63
63
63
63
65
66
66
66
66
68
68
ran
V%

87

87
88
88
92
92
xiii

Mechanical Properties of INOR-8 ... .... . ciiiiiinnen..
Properties of Reactor-Quality INOR-8 ...............
Thermal Fatigue of INOR-8 .......cciiiiiiiiinnnnnnn

Graphite Studies ... ..t i i it i e i
Evaluation of Grade TS-281 Graphite ................
Effects of Thermal Decomposition Products of

NH,F-HF on Graphite-INOR-8 Systems ................

Fabrication of Gd;03 and Gds03—Al1,03 Pellets ..........

IN-PILE TR S .. ittt et ettt et et e v s ssaansaenonnsensnns

ORNL-MIR-47 Molten-Salt Irradiation Experiments .......
Experiment ORNL-MIR-47-4 ... ... iuiiiinnnnennennnnnn.
Experiment ORNL-MIR-47-5 ... ... ..t iniiiniinnnan.

Postirradiation Examination of Experimental Assembly
ORNL-MIR =4 7=3 .. ittt ittt it tins e iaane e,
Chemical Analyses of Irradiated Fuel for Corrosion

Products . ... . i e
Distribution of Radicactivity and Salt in Irradiated
Graphite ... .. i e
Postirradiation Examination of Experimental Assembly
ORNL-MIR -4 74 i ittt et ettt et et s e
Description of Capsule ..........ciiiiin e,
Fabrication and Loading of Capsules ........... ...
Fuel o e i i e e e
Graphite ... . i i i i e e e e e
Decay Bnergy ... e e e e
Irradiation and Temperature Data ...................
Production of Xenon and Krypton ....................
GasS ANa Ly Se S vt ittt e i et e e e e
Discussion of Results ........ ... .,

Phase Equilibrium Studies .......c.cuiii ittt nnnnnens
The System LiF-BeFo-ZrFs; . ... i iiiiiiaen..
The System LiF-BeFp-ZyF 4-ThF4-UFs ..o iivn...
Li¥-UF, Eutectic as a Concentrated Solution of UF,

for Fuel MaKkeup ..... vt ittt it nnnaeeennns
Fractionation of LiF-BeF,-ZrF,-ThF,~-UF, on Freezing
Crystal Chemistry ........ ... i,

Oxide Behavior in Molten Fluorides — Dehydration of LiF

Physical Properties of Molten Salts ...................
Density of LiF-BeFo-2rF4-UFys ...ttt
Estimation of Densities of Molten Fluorides ........

Graphite Compatibility ....... ... i .
Removal of Oxide from Graphite Moderator Elements ..
Behavior of CF, in Molten Fluorides ................

Analytical Chemistry . ... ittt ten it an
Determination of Oxygen in MSRE Fuel Mixtures ......
Determination of Oxygen in Graphite ................
Homogenization of Radiocactive MSRE Fuel Samples ....

oooooo

------

......

oooooo

------

------

------

------

------

------

------

------

------

oooooo

------

------

oooooo

------

95
95
95
96
96

97
98
100

100
100
102

105
105
106
106
108

108
10%

.109

110
110
110
112
115

117

117
117
118

119
121
121

122

123
123
123
125
125
128
130
130
131
131
FUEL PROCESSING

Xiv

---------------------------------------------

ENGINEERING RESEARCH ON THERMOPHYSICAL PROPERTIES OF
SALT MIXTURES .« v otvvee ettt et e e e e e e e e e e e e

Enthalpy
Viscosity

----------------------------------------------------

---------------------------------------------------

133

135

135
136
PART 1. MESRE DESIGN, ENGINEERING ANALYSIS,
AND COMPONENT DEVELOPMENT
~ THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
1. MSRE DESIGN

Design Status

Design work on the MSRE was essentially completed. There remains
only final revising of system layout and design drawings and completion
of the electrical design. The design report, in which all engineering
calculations and analyses of the systems are presented, is presently be-
ing compiled. Much of the design work remaining to be completed involves
incorporation of development work into final design drawings. An example
of such a design is the removable salt-pipe heaters and insulation. Final
test evaluations of two different concepts of such removable insulated
heater sections are nearing completion. Design drawings of these units
must be completed after these tests indicate which type will be used in
the system.

Another item that awaits final test evaluation before the design can
be completed is the control rod drive. A model incorporating all the
principles has been operated successfully, and testing of an actual pro-
totype unit will permit the design to be finished. Such a prototype 1is
being ordered.

All layouts for the three major areas, reactor cell, drain tank cell,
and coolant cell, are complete, The variocus penetrations into these cells
have also been designed. The layout of interconnecting gas, water, elec-
trical, and instrument lines between these cells is nearing completion.

Considerable effort was devoted to the design of the containment ves-
sel, and all parts of this design were completed. Several methods of ef-
fecting a closure with the seal membrane that welds to the steel shell
were analyzed before an acceptable concept could be detailed. The pe-
riphery of the containment vessel just below the closure flange was re-
inforced with a circumferential steel ring to bring the stresses within
limits at design-point pressure. This pressure was dictated by the maxi-
mum credible accident. The vapor-condensing system, which ensures toler-
able pressures within the cell in the event of the maximum credible acci-
dent, was designed.

Since the molten-salt system employs electrical heat to bring all
components to the operating temperature, the electrical distribution for
this heater network is rather elaborate. This electrical system is about
40% designed. The heater control centers are complete. Conduit and
cable-tray layouts are 6U% complete. 'I'ne design ot moditications to
existing auxiliary diesel power controls are 60% complete. All electri-
cal designs are scheduled to be coumpleled Ly March 1, 19G3.

Design drawings for the charcoal-adsorber system for the offgas are
finished. Layout drawings of offgas piping in the reactor cell, drain tank
cell, and special equipment room are 95% complete. - Very little remains
to be done to bring the design of the entire gas system to completion.
4

The design of the jigs and fixtures for the assembly of the initial
and replacement components for the test cell was completed. The compo-
nents, piping, and flanges are supported by adjustable supports that du-
plicate or resemble the test cell structure and are positioned precisely
by optical tooling. Design work on the fixture for the assembly of the
drain tanks and flush tank is in progress.

Completion of the items mentioned above will complete the design of
all major equipment and facilities, including the maintenance control
room, the remote crane control, the television viewing devices, and the
remotely operable lifting tongs. The design of the work shield, protec-
tive environment enclosure (standpipe), tools, and operating procedure
for replacing the 7/8 in, -0D graphite sampler is complete. The graphite
sampler is identical in principal to the sampler illustrated previously,1
which was modified to accommodate the control rods and to raise the work
shield 3 ft above the lower shield beams for working convenience.

There were no changes in design concept or significant changes in
detail of any component or system of the MSRE within this report pericd.
A plan drawing of the major systems showing arrangement of vurlous com-
ponents in the MSRE building is shown in Fig. 1.1.

Reactor Procurement and Installation

Major Modifications to Building 7503

The Kaminer Construction Corporation, which has the prime contract
for the major modifications to Building 7503, has completed approximately
90% of its contract. The estimated completion date is October 15, 1962,
as scheduled. The status of the work in progress is outlined in the fol-
lowing sections,

Reactor Cell, Stress relieving of the reactor cell vessel is com-
plete, and installation of the sleeves between the inner and outer ves-
sels is progressing. This work will be completed when the expansion
joints (bellows) are received.

The support beams for the lower roof plugs were concrete-filled and
put in place. The frames for the lower roof plugs were fabricated and
are being fitted to the cell. These plugs will be poured soon. The up-
per roof plugs were poured but were not delivered because of lack of
storage space. The seal membrane is on-site.

Drain Tank Cell. All the concrete was poured in the drain tank cell,
including the upper and lower roof plugs and the support shield beams.
Plug welding of the stainless steel liner was completed, as shown in
Fig. 1.2. The lower roof plugs should be installed and ready to weld
in the seal membrane when it is delivered.

1"MSRP Prog. Rep. March 1 to Aug. 31, 1961," ORNL-3215, p 17,
Figo lo 126 :
5
-~ — UNCLASSIFIED
i3 EXI7 S % ORNL-LR-OWG. 74970
LUBE OIL
L ! g INSTRUMENT
9 i BUILDING NORTH PANEL
(o°REE) i
| ‘—¢——4 l EUEL PUMP LUBE OIL
{ /.‘ LUBE oIt LUBE QIL INSTRUMENT .
! / POWER PANELS PANEL
| F1REWA / B =
| i N
‘ ‘ | cooLa Mo
‘ | i LUBE O
2 . ‘ PACKAGE\
Sl £ |
, : e 7 £7 A 2
‘ £LEV. 2400 £LEY 8400 o 5 ,
3 : 7 |
LRI E: h
= § N
-, o A EXIST. SPECTROMETER
/ 3 W3S FEACTH TUNNEL FLOOR £L.833-3" . EXIST. ACCESS
/ ! 3 HATCH.
N | FiLey
% )| ExisT oucT BasKSIM ew
g E!fl:: pE!ZL 0 BE waLL o F/L?f»i/7
pucr RUN* &
P : ! N
R SLEEVES-R, IR NeyrroN. |, )
e FOR_INSTRUMENTATION \ SHELL INSTRIMENT /] Ar N 5
AUX, INSTRUMENTE) s sl 1
ELECTRICAL SERVICE 2 £ 5
AREA-FLOOR EL.84070) ’ REACTOR -~
QE§AEL A 30" BUTTERFLY
— 7 o YALVES
e 75 i
| TRANSMITTER | | ALt & . e
I ! (= 7 sAulE > 4800
b Y Joner MR e i L NEWIO" 2 SLEEVE
2 (BELO i L 5 7
b s P (p— = —3+—— | ——n "
i 3 &
/ 4 ( ; o steEves Ry, 2 R
A —} 3 A ConTAck FOR HEATER LEADS .
MAINTIENANCE| |, i - L A S i e 30"EXHAUST Ducr
PLAQTICE i | ey = A
S i | K A
N\ W STE d| ! eouipMENT al =~ IS A j f
TREATMENT STORAGE i [ | fecmsaow ||| N / . i RAD/ATOR
ELEV, 8310, /A
joé Tl R & DUCTWORK
| | JUNET, HATCH, | | K387
REYIRYD N EXCHINGE < e S s o
A ¢ roR NEW 247D
SLEEVE Re- LEAK DETECTION. CELL =3 s 1
00L i
/4 \} DRAIN TAKK
ety
DRAIN TANK CELL S REACTOR 3
FLo 81420 i oo | = N XK= -
=== — 7 &9 2, 5 A 1
i =i £ !
i . ox 'r-eml. s q z 0 N
¥ DECONTAMINATION AUEL by S ¢ ?t\v‘"’@
L4l CE p srorace | | 14 /, 2
. ' ! FLOOR £1.832°C - // [~ %K 748 SPARES| Spuzi Mz “ r
“ : S 1 { ? =S
) (R . N \HEW
v \ Negacroz 7
f I N ‘)’ - N 8 SLEEVE Q) S 7 n
| K, { . DRAIN ¥ DA FLUS, RN, =0 Mié;%
e s = . TANK et | TANK K62 | Tk 7 M Bk 7
" - - - - b \, N Np? / \\‘; / %
A N, L ! & SO - EXIS]. VENT
By 2TACKE, £LocK 4 b, 7o o HOUSE
/ S S T o DRAIN LINE oe'% Y/ r,‘_‘;
%) ; 3 TRANSFORMER 2N e
= e = = 2 0
_____ ‘ e
CONDENS, i - ! | PENrAOUSES |
3 73 / WALL (ABOVE) | s
< a5 y i
S o ST SR 2
SWITCH _HOUSE ; wa = i Wy
2oLET S e 1 0 % w0 U
_FIN FLR ELR40°0" XPAN3, i s B ADJUSTABLE muih
rr;;)y(e::wu Al F:Z 3 ?: ORY STACKED BLOC % E‘F’; | .
SIORAGE TANK = - | hitvd| If,
. COOLING : i
5 & t . / WATER I
T /] PUMBPS > S AL FLow
FAN-10000 CFM
i - v
i 7 ¢ A
7 ! // ) —1 =) N
ELOWFAH | T == |
# f 10,000 CEM E".fi’g CAHIHe )
Q Ex157 co"? _— >
¥ AXIVANE EANS N
= < y / 5
il £L00R EL.B38O°
| iy EX/ST BLOWER
HOoUSE

Fig.

LI

Plan View of MSRE.

UNCLASSIFIED
PHOTO 58774

Fig. 1.2. Fuel Drain Tank Cell.

Maintenance Control Room. The grade beams for the malntenance con-
Lrol room were poured. The [orms [or pourlng up Lo elevallon 862 'L are
being installed, and the concrete will be poured soon.

Radiator Cell. The sheet metal work in the radiator cell is approxi-
mately 50% complete. It is expected that this work will be completed on
schedule.

Sheet-Metal Liner for Crane Bay. All framing for the crane-bay liner
was fabricated and 40% of the framing was installed. The sheet metal liner
was fabricated and is ready to install.

Building Ventilation Duct Work. The duct work is 90% complete. The
remaining work consists of the installation of dampers, which are sched-
1led to be delivered soon.

Reactor Cell Exhaust Duct. The 30-in. exhaust duct is installed to
the end of the Kaminer portion, including the butterfly valves.

Hot Waste Storage Cell. The storage tank is installed and 95% of the
piping complete.

Hot Storage, Fuel Transfer, Decontamination, and Spent Fuel Storage
Cells. The stainless steel liner is being installed in the hot storage
cell. Work in the other three cells is complete, except for miscella-
neous concrete patching and stacking of shielding blocks in the spent fuel
storage cell.

Electrical System. Electrical conduit installation in the crane bay
is 90% complete. Wiring for the air compressors and controls is complete,
except for some miscellaneous control wiring, which is awaiting delivery
of control switches.

Air and Water Piping. The compressed air and water piping is approxi-
mately 98% complete.

Auxiliary Line Penetrations into Reactor Cell, O0Oil, water, and other
service lines were fabricated, and installation is approximately 50% com-
plete,

Electrical Service Area, All concrete was poured, and ladders and
hatchway covers were installed. Work in this cell is 80% complete.

Construction Outside Building 7503

The R. S. Hixson Construction Company of Oak Ridge was awarded the con-
tract for the construction work outside Building 7503. Work was started ap-
proximately May 1 and is expected to be completed approximately September 15.

The exhaust filter house is complete except for the installation of
filters, which are expected to be delivered approximately September 15.
The offgas exhaust stack is installed. The exhaust blowers, duct, and
dampers will be delivered by September 1 and installed by September 15.

The cooling towers were installed, and the piping was completed, ex-
cept for connections to the pumps. The pumps are to be delivered soon.
The underground piping is complete.

The inlet filter house is complete, and the steam coils, dampers,
duct, and filters will be delivered soon.

Procurement of Materials

With the exception of the graphite for the MSRE core, materials pro-
curement is approximately on schedule. The graphite is now promised for
November 1, 1962, and efforts are being made to improve this delivery
date. The status of procurement is outlined in the following sections.

INOR-8. Delivery of all plate, rod, weld rod, and pipe is complete,
except for some miscellaneous pieces. Partial delivery was made of pipe
fittings and tubing; delivery is approximately 75% complete. Delivery
of forgings, that is, freeze flanges, heat-exchanger tube-sheet blank
flanges, O-ring gaskets, and other special forgings, is approximately
80% complecte,
Graphite. The National Carbon Company has formed, baked, and graphi-
tized all the base-stock graphite bars. The individual lots are at vari-
ous stages of the multiple impregnation treatment. All bars received the
first impregnation. There are about 80 graphite bars completely through
all the required impregnations; however, the fabrication of sound base-
stock graphite bars is the critical part of the fabrication. The delivery
of the completely machined graphite bars was tentatively rescheduled from
September 1 to November 1, 1962.

Auxiliary Heat Exchangers. A contract was awarded Berlin Chapman
Company for the fabrication of three stainless steel auxiliary heat ex-
changers.

Stainless Steel Piping. Contracts were awarded for the manufacture
of type 304 stainless steel pipe, tubing, and fittings for all auxiliary
piping systems. Most of this material has been manufactured and is be-
ing inspected at the vendors' mills. Contracts have not been awarded
t'or gspecial f'langes and gaskets.

Reactor Cell Support Steel. Steel for the reactor cell equipment
supporting structures is being furnished by O'Neal Steel Company. Special
pipe spring supports are being manufactured by Bergen Pipe Support Company.

Fabrication of Components

The fabrication of MSRE components in the Y-12 Machine Shops is es-
timated to be approximately 50% complete and on schedule. Delivery of
the first component is expected approximately October 1, and all compo-
nents are expected to be delivered by January l. The status of component
fabrication is described in the following sections.

Reactor Vessel. The reactor vessel (Figs. 1.3 and 1.4) is being fab-
ricated and is approximately 45% complete. Delivery is expected approxi-
mately January 1, 1963.

The stainless steel thermal shield for the reactor vessel is being
fabricated in the Union Carbide Nuclear Company Paducah machine shop and
is approximately 35% complete. Delivery is expected approximately
October 1, 1962. Procurement was initiated for approximately 75 tons of
steel balls that will be poured into the thermal shield for gamma shield-
ing.

Ilecnt Exchanger. The INOR=C primary heat exchanger is approximately
65% complete (Fig. 1.5), and delivery is expected approximately January 1,
1963. A vendor, Wall Colmonoy, is under contract to furnace-braze the
tubes to the tube sheet with a gold-nickel alloy.

Radialor. The salt-to-alr radiator is approximately 12% complete.
The headers for this component are being fabricated in the Union Carbide
Nuclear Company shops at X-10 and Paducah and are approximately 95% com-
plete. Delivery is expected approximately January 1, 1963.
UNCLASSIFIED
PHOTO 38994

Fig. 1.3. 1Inside of Reactor Vessel Showing Flow Straightening
Vanes.

Radiator Enclosure. The enclosure for the radiator (Fig. 1.6) is
approximately 65% complete, and delivery is expected approximately
January 1, 1963. The door-lifting mechanism was fabricated in a local
Oak Ridge Machine Shop and is complete. Delivery is complete on all
motors, clutches, brakes, and electric heaters for this component.

Salt Storage Tanks (Total of 4). The several INOR-8 fuel and cool-
ant storage tanks (Fig. 1.7) are approximately 45% complete. Delivery
of these components will begin approximately October 1, 1962, and be com-
Pleted by October 1, 1963.

Fuel and Coolant Pumps. Fabrication of the fuel and coolant pumps
(Fig. 1.8) is approximately 75% complete. Delivery of the first unit is
expected approximately January 1, 1963.

Miscellaneous Fabrication. The ORNL and Paducah machine shops are
fabricating miscellaneous parts for the major reactor components. In
addition, certain equipment supports, assembly jigs, and panel boards are
being fabricated in the K-25 shops. The sampler enricher, charcoal beds,

10

UNCLASSIFIED
PHOTO 38932

TR A

Fig. 1.4. Flow Distribution Nozzle [or Reactor Vessel.

UNCL ASSIFIED
PHOTO 38530

Fig. 1.5. Some Completed Components of INOR-8 Primary Heat Ex-
changer,.
i

UNCLASSIFIED
PHOTO 38995

Fig. 1.6. Partially Completed Radiator Enclosure.
12

UNCLASSIFIED
PHOTO 38680

() “0‘
®,

0’.
LX)

.0

Fig. 1.7. Coolant Salt Drain Tank.
13

UMCLASSIFIED
PHOTO 38989

Fig. 1.8. Volute and Upper Half of Fuel Pump Bowl.
14

salt piping, and other miscellaneous parts are being fabricated in the
ORNL machine shop. A contract was awarded to Taylor Engineering Company
for the fabrication of seven special equipment supports and four special
vessels.

Reactor Instrumentation and Control Systems

Layout of Instrumentation and Control System

The layout of the instrumentation and controls system proceeded
closely in accordance with the scheme outlined in the previous report.2
Drawings showing the exact location of all control panels were completed,
including drilling details for floor penetrations in the main and aux-
iliary control areas. The relay cabinet was moved to a new position, and
three panels for control of the diesel generators were added to the
auxiliary-board panel arrangement. Requirements for wire ways between
the main control area and field-mounted equipment were determined, and
detail design is now under way. 'These wire ways consist of three sepa-
rate sets of 3- by 24-in. open trays. Thermocouples, signal cables, and
control circuit wiring will be run in separate trays. Cable from isolated
signal transmitters will be run in conduit to the nearest tray. Details
for wire terminal boxes between field-mounted electrical equipment were
also determined.

Design Status

Instrument Application Diagrams. Instrument application flow dia-
grams for the fuel sampler-enricher system, the fluorination system, and
the containment system were completed and approved for construction. Nine
diagrams previously approved were revised in accordance with recent de-
sign changes. A tabulation of all instruments shown on the flow diagrams,
giving identifying numbers, location, function in process, and a brief
description, was completed. These will be revised periodically to in-
corporate design changes.

Control Panels. The design of 8 of the 12 modular instrument panel
sections required for the main board was completed and fabrication of
these panels is presently under way in the ORNL panel shops. The three
remaining panels, as well as the control console, will be completed when
all the nuclear instrument requirements are known. A complete design
for one panel section includes instrument-mounting-hole cutout drawings,
pneumatic tubing diagrams, and electrical wiring diagrams. <Twenty-five
drawings for the main-board panel sections were approved.

Twenty-four drawings covering the design of 11 auxiliary-board panels
located in the transmitter room and auxiliary room were completed. 1In
addition, the design of the thermocouple-patch-panel cabinet was com-
pleted.

2"MSRP Prog. Rep. Feb. 28, 1962," ORNL-3282, pp 10-19.
15

Designs of three sampler-enricher control panels and two cover-gas-
treatment control panels were also completed and approved. Eighteen
drawings were approved and issued for construction. These panels are
presently being fabricated in the ORNL shops and when completed will be
used to control mockups of these systems.

The designs of six other auxiliary instrument panels located in the
field are complete. These panels serve the fuel and coolant pump lube
oil systems, cooling water system, and the containment air system. 5ix
drawings were completed. Fabrication of the containment vessel air sys-
tem instrument panel was completed, and the panel was delivered to the
contractor.

Forty-two 2-ft modular-panel sections are required in the MSRE in-
strument system. Designs are complete for 30 of these panels. Design
of the remaining panel sections is approximately 50% complete.

Field Installations. Information necessary for the development of
the electrical power system criteria was prepared. This included the
safety control circuit 48-v power distribution system, 115-v ac instru-
ment power requirements, and remote control requirements for switchgear
and motor-control centers. The design of electrical control circuitry
for two 40-hp instrument air compressors was completed. '

The requirements for the instruments and controls system reactor cell
penetrations for electrical signal cables, thermocouple lead wire cables,
and pneumatic lines were determined in detail.

A mockup of a typical MSRE freeze valve, with reactor-type thermo-
couples, solid-state switch modules, coolant-air system, and control cir-
cuits was designed and constructed. Initial test operations indicate that
all control-circuit components will perform satisfactorily.

Data Handling. A requisition was issued for a data system to be ac-
quired on a three-year lease periocd with an option-to-buy clause. Pre-
liminary proposals from three manufacturers were received the week of
July 9. The three proposals were evaluated in terms of cost and equip-
ment in accordance with the specification. On the basis of the prelimi-
nary proposals, a system was tentatively selected. This selection was
based on preliminary information and is subject to change when the final
detailed proposals are received,

Process and Personnel Radiation Monitoring. A review of the process-
monitoring requirements was completed in May. The selection of the neces-
sary dctectors and associated electronic equipment is approximately 80%
complete., Preliminary detector locations and mounting requirements were
discussed. Preliminary investigations were started for the design of the
two-stack monitoring systems. The instrument application diagrams now
show all the proposed process monitors, with the exception of the high-
level gamma monitors for the reactor and drain tank cells.

16

The resulting recommendations were reviewed by the Laboratory's Fixed
Instrument Advisory Committee and by ORNL Radiation Safety Control per-
sonnel. The recommendations for personnel-monitoring instrumentation
are summarized below: -

Instrument Quantity

Continuous air monitor, Q-2240

Monitron, Q-1154

Radiation and contamination monitor, Q-2091
Hand and foot monitor, Q-1939

Alpha sample counter, Q-2345

Beta sample counter, Q-2344

Central control and annunciator panel

H e300~

The design of this system was initiated and procurement of the required
instruments is under way.

Procurement Status

Specifications were prepared and procurement initiated for approxi-
mately 80% of the components required in the process instrument system.
All panel-mounted instruments are either on order or on hand. One hun-
dred and twenty-one specifications for commercially available standard
items were prepared, and procurement was initiated.

Thirteen detailed job specifications for some components requiring
special development or procurement effort were completed, and procurement
was initiated. These include such items as weld-sealed transmitters,
valves for radioactive gas service, thermocouple material, a data-handling
system, a coolant-salt-system venturi flow element, and high-temperature
NaK-filled traenemittcro.
17

2. MSRE REACTOR ANALYSIS

Nuclear Accident Analyses

Analyses of reactor behavior in potentially hazardous situations were
carried out, and the results were reported in the second addendum to the
MSRE Preliminary Hazards Report.! The incidents that were analyzed in-
cluded fuel pump failure at high power, cold-slug accidents, premature
criticality during core filling, breakage of a graphite stringer, passage
of a slug of concentrated fuel through the core, and runaway rod withdrawal.

Calculational Procedures

The case of stoppage of the fuel pump with the reactor at power wvas
analyzed with an analog computer.Z All other situations were analyzed
using Murgatroyd, a digital program for MSRE kinetics analysis.3

In Murgatroyd the fuel and graphite in the core are treated as sep-
arate bodies, each with its own mean temperature and temperature coeffi-
cient of reactivity. The reactor is treated as a point, however, in that
the mean fuel temperature is assumed to be the average of the inlet and
outlet temperatures. Heat transter between fuel and graphite is assumed-
to be proportional to the difference between the fuel and the graphite mean
temperatures. Reactivity disturbances can be represented in Murgatroyd as
a series of steps and ramps, but the inlet temperature is regarded as being

constant.

The reactor properties used in the analysis were appropriate for the
MSRE with 1 mole % thorium in the fuel. (Elimination of the thorium would
increase the temperature coefficients of reactivity and the prompt-neutron
lifetime.)

Results of Fuel Pump Power Failure Analysis

The results of a simulated fuel pump failure at a reactor power of
10 Mw, with no corrective action, are shown in Fig. 2.1. The decrease in
fuel-salt flow led to a decrease in the core cooling rate and to an in-
crease in the effective delayed-neutron fraction, which resulted in an
initial rise in the core temperatures. Also the loss in flow decreased
the rate of heat transfer to the secondary coolant, and with no corrective
action the coolant in the radiator reached the freezing temperature in
less than 2 min. Heat transfer from the radiator can, however, be de-
creased quickly by closing the radiator doors. Results for the case of

1'"Molten-Salt Reactor Experiment Preliminary Hazards Report," Adden-
dum No. 2, May 8, 1962, unpublished internal report.

2p. N. Haubenreich and J. R. Engel, "Safety Calculations for MSRE,"
ORNL-TM-251 (May 15, 1962).

3C. W. Nestor, Jr., "Murgatroyd - An IBM 7090 Program for the Analysis
of the Kinetics of the MSRE," ORNT.-TM-203 (April 6, 1962).
18

UNCLASSIFIED
ORNL-LR-DWG 67578

1400 — ——]
CORE OUTLET
/ <
———~———. CORE FUEL MEAN
/// B
.//',/, [ ———t""1 GRAPHITE MEAN|
.~ 1200 ~
3
- [ ———
@ 1100 ——]
a '\\\ RADIATOR INLET
= \
e —
1000 \\‘\
\
.\-____--
\ \
\ \
\ --\-.
900 — ——
SALT LEAVING RADIATOR e
"-_____ —_—
= e ]
800
12
/\
10 N
\\SSWER
— 8
3
s \\\\\‘
N
= & \ .
= \
S -
\\
2 -_—'fl—..‘
0 >
0 20 40 60 . 80 100 120
TIME (sec)
Fig. 2.1. Power and Temperatures Following Fuel Pump Failure, with

No Corrective Action.
Ls

19

control rod action and closing of the radiator doors following the fuel
pump power failure are given in Fig. 2.2; the control rods were considered
to decrease reactivity at the rate of 0.07% per second 1 sec after power
failure, and heat removal from the radiator was stopped after about 30 sec.
For these circumstances the fuel salt temperature rise was small, and the
secondary coolant temperature did not become too low.

Effects of Cold-Slug Accidents

The consequences of cold-slug accidents were estimated by calcu-
lating the reactivity transient associated with passage of a cold fuel
slug through the core and introducing this transient into a Murgatroyd
calculation by considering a series of reactivity ramps. The temperature
rise calculated by Murgatroyd for a constant inlet temperature was then
superimposed on the temperatures that would be produced by the cold slug
without nuclear heating. Results are shown in Fig. 2.3 for two extreme

UNCLASSIFIED
ORNL-LR-DWG 67579

1300
CORE OUTLET
. _— GRAPHITE MEAN
u — ] I
0 1 —
° 200 CORE FUEL MEAN
(VS
g CORE INLET
'_
<I
v f
W
o 1100
= i\\\\FADIATORINLET
'_ \
\"- -__——_—_‘-———_—
——T" RADIATOR
1000 RADIATOR OUTLET
10 1
\ FISSION POWER

° \
=
s \
z 6
« \
s \
B 4
a

2 AN

N
\‘.
0
0 20 40 60 80 100 120

TIME (sec)

Fig. 2.2. Pump and Temperatures Following Fuel Pump Power Failure.
Radiator doors closed and control rods driven in at 0.4 in./sec after
failure.
20

UNCLASSIFIED
ORNL-LR-DWG 67580R

1300 |
MEAN TEMPERATURES A | T~ T =
[~ L7 A “'-..___.
9 / // e =
> 1200 GRAPHITE A /
g \ __'_--_,__ / ./_--_——_—_‘___--_-
= N 7 i
ac \ /"——"—-__/
w
a \ _ /a
= 1100
e \ / — 20 ft3 COLD SLUG
EL —==30 f

PRESSURE RISE {psi)
o
\

300

\ POWER

ZOuU : \ '
100 \Q
3

POWEFR (Mw)

TIME (sec)

Fig. 2.3. Bystem Behavior Associated with PaSSagé of Cold Slugs at
900°F and a 1200-gpm Flow Rate with the Core Initially at 1200°F.

cases: 20 or 30 ££t3 of fuel at 900°F was assumed to be pumped at 1200 gpm
through the core; the reactor was initially critical at 1200°F and at low
power with no circulation prior to introduction of the cold slug. Peak
temperatures in the core were estimated to reach about 1600°F. The core
pressure rise was not appreciable in either case.
21

Results for Filling Accidents

Criticality could be reached prematurely while the core was being
filled if the fuel were abnormally concentrated, or if the fuel tempera-
ture were abnormally low, or if the control rods were withdrawn too far.

The power excursions that would result from the continued addition of re-
activity associated with such filling accidents were computed by Murgatroyd,
assuming no heat removal from the fuel and using a fuel-graphite heat trans-
fer coefficient appropriate for a static core.

In the first case it was postulated that the fuel was concentrated by
selective freezing in the drain tanks of phases containing no uranium (or
thorium). With 39% of the fuel frozen,4 filling the core with the remain-
ing liquid at 1200°F at a rate of 1 ft3/min would result in a 55-Mw power
excursion about 60 sec after initial criticality. No significant pressure
surge would result.

Filling the core with salt having the normal fuel concentration and a
temperature of 1200°F would produce an excursion similar to the one de-
scribed above if all rods were fully withdrawn. Filling at 1000°F with
the fuel concentration and rod positions corresponding to criticality at
1200°F would produce a much smaller excursion.

Rod insertion would be quite effective in preventing the power ex- |
cursions in these accidents because the rates of reactivity increase would
be very low. If no corrective action were taken (e.g., insertion of rods
or draining of core), temperatures in the core would exceed 1800°F, in the
worst case, by the time the core became completely full.

Effect of Uncontrolled Withdrawal of Rods

Simultaneous withdrawal of all three rods could produce a large power
excursion and undesirably high temperatures if these rods were withdrawn
rapidly with the reactor initially subcritical or at very low power. In
the worst case examined, the rate of reactivity addition was assumed to be
0.08% 6k/k per second and the initial power was taken to be 1 watt. A
pover excursion to 730 Mw and a pressure surge of 19 psil resuited. More
serious, the calculated mean fuel temperature rose from the initial 1200°F

to a high of 1600°F, and local temperatures in excess of 2000°F were esti-
mated. .

Effect of Graphite Movement

If a graphite stringer were to break in two near the center of the
core and the upper half could float up, the reactivity would increase
0.004% 6k/k for each inch of stringer replaced by fuel. If the entire
central stringer were replaced with fuel, the reactivity increase would
be only 0.13% 6k/k, No power or temperature excursion of any consequence
would result, even in the latter event.

“This is the highest fraction that could be frozen and still leave
enough liquid to fill the core.
22

Results of Concentrated Fuel Addition

An upper limit on the disturbances which could result from a fuel
addition of 120 g of U®35 yas calculated. For this case it was postulated
that the added fuel passed through the core as a flat pancake-shaped ele-
ment under normal flow-rate conditions. If this were possible and the
reactor were initially at 10 Mw, the power would rise to a peak of 120 Mw,
and the fuel mean temperature would rise 90°F. The pressure surge would
be about 2 psi. Smaller disturbances would result at lower initial powers.

Reactor Statics Calculations

The reactor statics calculations reported here are based on the cur-
rent reactor design. In this connection, a two-dimensional, 19-region
nuclear model of the reactor was utilized to represent the various reactor
regions. Because of the limitation to two dimensions, the control rod
thimbles were considered as a single INOR-8 cylindrical annulus concentric
with the core center line. A vertical half-section of the reactor model
showing the relative sizes and positions of the various regions is shown
in Fig. 2.4. The boundaries of the regionc and the volume fractions of
fuel, graphite, and INOR-8 in each are summarized in Table 2.1.

A fuel salt mixture containing x mole % UF4, 70 mole % LiF, (25 - x)
mole % BeFs, and 5 mole 9% 7rF4 was considered in the calculations. The
isotopic compositions of lithium and uranium were based on material com-
positions that might reasonably be available for use in the reactor. The

Table 2.1. Nineteen-Region Core Model Used in Equipoise Calculations for MSRE
(See Fig. 2.4 for graphical location of rcgions)

Rodiun (in.) Height (in.) Composition (val &) ,
Region Region Represented
Inner Outer Bottom Top Fuel, (raphite  INOR-8
A 0. 29.56 Th.92  T76.04 0 0 100 Vessel top
B 29.00 29.56 -9.1h4 Th .92 0 o) 100 Vessel sides
C 0 29.56 -10.26 -9.1k 0 0 100 Vessel bottom
D 3.00 29.00 67 b7 T4 .92 100 ) 0 Upper head
E 3.00 28.00 66.22 67 .47 93.7 3.5 2.8
F 28.00 29.00 0 67 .47 100 0 0 Downcomer
G 3.00 28.00 65.53 66.22 94 .6 5.4 0
H 3.00 27.75 64.59 65.53 63.3 36.5 0.2
I 27,75 28.00 N A5 53 0 0 100 Core can
J 3.00 27.75 5.50 64,59 22.5 7.5 0 Core
K 2.94 3.00 5.50 T4 .92 O 0 100 Similated thimbles
L 0 2.94 2.00 64,59 25.6 T4 .L 0 Central region
M 2.94 27.75 2.00 5.50 22.5 TT+5 ) Core
N 0 27.75 0 2.00 23.7 76.3 o) Horizontal stringers
0 0 29.00 -1.k1 0 £A.Q 15.3 1.7.8
P 0 29.00  -9.1hk -1.k1 90.8 0 9.2 Bottom head
Q 0 2.94 66.22 T4 .92 100 0 0
R 0 2.94 65.53 66.22 89.9 10.1 0
S 0 2.6h  64.59  65.53 43.8 56 .2 o

23

UNCLASSIFIED
ORNL-LR-DWG 736140

Y I P PO B |
0O 2 4 6 8 10
INCHES

Fig. 2.4. Nineteen-Region Core Model for Equipoise-3A Calculations.
See Table 2.1 for explanation of letters.
calculations described below were for an isothermal system at 1200°F with
no reactivity controlled by the control rods, no fission-product poisons,
and no chemical impurities in the fuel.

Criticality Calculations

The initial criticality calculations were made with Modric, a 34-group,
multiregion, one-~dimensional, neutron-diffusion program for the IBM-7090.
Because this program is limited to one dimension, two types of geometry
calculations were made. One was based on slab geometry, with region and
material specifications corresponding to an axial traverse through the main
part of the core; a radial buckling term was used in this calculation.

The other was a radial calculation, with region and material specifications
corresponding to a radial traverse through the reactor midplane; an axial
buckling term was used in this calculation. The two calculations agreed
within 1% in predicting a critical uranium concentration of 0.1% mole %.

In addition, these calculations provided the two-group nuclear constants
for each of the regions as required for two-dimensional, two-group calcu-
lations.

Flux Distributions

The spatial distributions of the fast and thermal fluxes and adjoint
fluxes were calculated for the 19-region, 2-dimensional model using the
Equipoise-3A program. The results for a reactor pover of 10 Mw are pre-
sented in Figs. 2.5 and 2.6. The radial distributions (Fig. 2.5) are for
an axial position that corresponds to the maximum in the thermal flux,
which is at a position very close to the core midpiane. The axial dis~
tributions® (Fig. 2.6) are given for a position 7 in. from the core center
line; this radius corresponds to the maximum value of the thermal flux.

Information about the space-energy distribution of the fast flux was
obtained by combining the Equipoise-3A results with the results of the two
Modric calculations. Figure 2.7 shows the radial distribution, near the
core midplane, of the neutron fluxes for energies greater than 0.1 and
1.0 Mev, and Fig. 2.8 shows the axial distribution of the same fluxes 3 in.
from the core center line. This radial position corresponds to that of
the control-rod thimbles and the small grapnlte test specimeris. The Tluxes
are normalized to a power level of 10 Mw.

Power Density

'he spatial distribution of the power density is determined by the
total fission rate. About 87% of the total fissions in the MSRE are caused
by thermal neutrons. As a result, the spatial distribution of the fuel
power density is essentially the same as that of the thermal-neutron flux.
The maximum fuel power density at 10 Mw with no fuel permeation of the
graphite is 33.6 w/cm®. The corresponding maximum graphite power density
is 0.83 w/cm®. '

SThe reference plane for axial measurements was the bottom of the
horizontal graphite bars on which the vertical stringers are supported.

~y
UNCLASSIFIED
ORNL-LR-DWG 73611

(x10'3)

,4 ~C

UNCL ASSIFIED
ORNL-LR-DWG 73612
FAST FLUX | (x10'5) I

o
Q
{.')
I y _
P : ~ FAST FLUX
£ \ 3
g Co \
3 S 4
= P N
| f‘* / N\
'E’: 8 -~ SLOW FLUX N E 1o
N e :
o =~ ™~ 2 SLOW FLUX
e SLOW ADJOIN'I\\ L. 8 o
iR ) T — 5 TN
2 Bt W 7 SLOW ADVOINT S
] T e 3 d "\
o« S, ':33. 6 N
= FAST ADJOINT S« > —_feeees <L N\ \
Qo 5 Pl S
I: 4 5 -~ FAST ADJOINT b, \
iy <L
2
@ g 8
X
2 E
3 2
i
Y o
0 5 10 5 -1C 0 10 20 30 40
RADIUS (ir.) . AXIAL POSITION (in.)
Fig. 2.5. Redial Distribution of Fluxes and Fig. 2.6. Axisl Distribution of Fluxes and

Ad joint Fluxes in the Plane of Maximum Slow Flux. Adjoint Fluxes at a Position 7 in. from Core Center
Line. :

Ge
UNCLASS'FIED
ORNL-LR-DWG 733513
(x10'3) .

4.0

3.5 -\ . UNCLASSIFIED

ORNL -LR-0WG 736144

{x 1013}
g 3.0 \ : is //’—“\\
E ’ #(£ > 0.1 Mev)
N -
5 3
2 25 ' G (£ > 0.1 Mev! N 3.0
e . E
z >
= c
o g
S 2 as
r 2.0 A z
] . =
2
[+ 4 lCL) 2.0 \ /\
= g v
S s z
g \ &
0.4 . - . 4 14
t ol P LE > Mev) Q
3 &  $lE>iMew)
w 1.0 A\ o / \

2 /
x 1.2 / Vs ™,
0.5 4
\ 0.5 // \
5 | . ,////
0 5 10 5 20 25 30 -0 0 1c . 20 30 40 S0 80 70 80
RADIUS (in.) : AXIAL POSITION (in.)

Fig. 2.7. Radial Prcfiles of High-Energy Fig. 2.3. High-Energy Neutron Fluxes 3 in.
Neutron Fluxes in Plane of Maximum Flux Based on from Core Center Line Based on Modric Calculations.
Modric Calculations. :

- ¢ 1 k!l

o¢
27

- Steady-State Fuel and Graphite Temperatures

Fluxes and power densities obtained from Equipoise calculations were
used to estimate fuel and graphite temperature distributions and average
temperatures for steady-state reactor operation at 10 Mw.®

Temperature Distributions

The shape of the temperature distribution in the reactor is deter-
mined by the shape of the power-density distribution and the fuel flow
. pattern. In the main part of the core, where the fuel flows in discrete
channels among the graphite stringers, local temperature variations across
individual fuel channels and graphite stringers are superimposed on the
gross distributions. The gross distributions, which have the greatest
effect on the reactor average values, are described below. They are based
on 10-Mw reactor operation and a total flow rate of 1200 gpm.

In describing the gross temperature distributions, only that part of
the core which generates the major fraction of nuclear energy needs to be
considered, that is, the central core associated with the regions designated
J, L, M, and N in Fig. 2.4t and Table 2.1. The fuel temperature in any chan-
nel at a particular elevation is a function of the radial position of the
channel, the inlet fuel temperature, the flow rate in the channel, and the
total amount of heat produced in or transferred to the fuel as it passes
from the inlet to the elevation in question. Figure 2.9 shows the axial
variation of fuel temperature in the hottest channel for the power dis-
tribution calculated for the MSRE. The channel inlet temperature is some-
what higher than the reactor inlet temperature because of the heat produced
in the entrance regions through which the fuel must pass in reaching the
main part of the core. Figure 2.10 shows the radial variation in fuel
temperature at the midplane of the core. The relatively lower temperatures
in the central and outer regions of the core are caused by the higher fuel
velocities in these regions. The fuel velocity in the central channels is
expected to be about three times that in most of the remaining channels.
Less extreme variations are expected in the outer channels.

The over-all temperature distribution in the graphite depends on the
fuel temperature, the graphite power density, and the resistance to trans-
fer of heat produced in the graphite to the fuel in the adjacent channels.
The internal resistance of the graphite to heat transfer produces some
elevation of the graphite temperature above that of the fuel, but most of
the temperature difference results from the Poppendiek effect.” Figure 2.9
shows the axial distribution of the mean temperature in a graphite stringer
immediately adjacent to the hottest fuel channels. The calculated axial

8J. R. Engel and P. N. Haubenreich, "Temperatures in the MSRE Core
During Steady-State Power Operation,' ORNL-TM-378 (in preparation).

7H. F. Poppendiek and L. D. Palmer, "Forced Convection Heat Transfer
in Pipes with Volume Heat Sources Within the Fluids," ORNL-1395-C, Nov. 11,

1953
UNCLASSIFIED
CRNL—LR— JWG 73615
1300

UNCLASSIFIED

o ORNL-LR-DWG 73616
/’_\ 1290
/ \ /
1280 / : 1280

GRAPHITE

1270

GRAPHITE

1260 / ‘ - 1260 E N
/ / | \
|

. 4250
1240 / yd

, 1240 \
1220 - / 1230 \

' 1220 _ - \
y

1200 / ,/ | \JEL \
/ - 1210 i \
/ | 1200 i ;‘
100 ™ | : \ \
190 by \
5

TEMPERATURE (°F)

TEMPERATURE (°F)

1160 1180
0 10 20 30 40 52 60 70 0 5 10 15 20 2 30
DISTANCE FROM BOTTOM OF CORE fin.) RADIUS (in)
Fig. 2.9. Axiel Temperature Profiles in Hottest Fig. 2.10. Radial Temperature

Channel of MSRE Core (7 in. from Core Center Line). Profiles in MSRE Core Near Midplane.

8¢
29

distribution is somewhat steeper than may be expected in the MSRE because
axial heat transfer in the graphite was neglected in the calculation.
Figure 2.10 shows the radial variation of the graphite temperature at the
core midplane. The distribution in the plane of maximum graphite tempera-
tufes would have the same general shape but would be displaced upward about
12°F,

The graphite temperature distributions in Figs. 2.9 and 2.10 were
calculated on the basis that 6% of the reactor power is produced in the
graphite by gamma and neutron heating; no other heat sources in the graph-
ite were considered. The graphite-fuel temperature differences will be
greater than the indicated values if fuel salt soaks into the graphite and
produces a source of fission heat in the stringers. If 2% of the graphite
volume were uniformly filled with fuel salt, the maximum difference between
the mean temperature in a graphite stringer and that in the adjacent fuel
channel would increase from 62.5 to T4.5°F.

In addition to the above results for the main core regions, the mean
fuel temperatures and temperature rise were also calculated for a number
of peripheral regions, based on 10-Mw reactor operation and an average
fluid temperature rise of 50°F in passing through the core. The increase
in fuel temperature as the fuel passes through a region and the associated
mean temperatures for the various regions are summarized in 'lable 2.2.

Average Temperaturcs

A useful concept in relating the actual fuel and graphite temperature
distributions to the total reactivity is the nuclear average temperature.
This quantity is defined as follows: At low power, reactor criticality is
assumed to be established at isothermal temperature conditions in fuel and
graphite. Then, as the power is raised, the fuel nuclear average tempera-
ture, T,, is defined as the equivalent uniform fuel temperature that gives
the sam® total reactivily level as that associated with the actual fuel
temperature distribution, with the graphite temperaturg held constant.
Similarly, the graphite nuclear average temperature, T , is defined as the
uniform graphite temperature that gives the same react%vity level as that
given by the actual graphite temperature profile, with the fuel temperature
held constant. A reactor nuclear average temperature could be defined in
an anglogous manner; however, it appears more convenient to consider Tf
and T individually in the manner indicated below.

Using first-order, two-group perturbation theory, the relation de-
rived between the nuclear average temperatures and the actual distributions
was8

8B. E. Prince and J. R. Engel, "Temperature and Reactivity Coefficient
Averaging in the MSRE," ORNL-TM-379 (in preparation).
where

J

Reactor

30

T(r,z) G(r,z)

dv

G(r,z) av

Reactor

* »* e 3 >*
G(r,z) = G 8.8 + G B8 +G,P.0 + 008, -

Table 2.2.

Fuel Temperature Rise and Average
Temperature Asdoelated with Varivus Reaclor

Regions at a Power Level of 10 Mw

Average Temperature

Rise in Passing

Average Temperature

Region” Through Region in Region
(°F) (°F)
D 2.4 1224 .7
B 0.8 1223.1
F 1.3 1175.6
G 0.5 12224
H 0.6 1221.9
J 43.0 (b)
L 22.0 (b)
M 1.0 (v)
N 0.3 (v)
0 0.3 1177.1
P 0.7 1176.6
4] 1.7 1200.9
R 0.3 1199.9
S 0.2 1199.7

aSee Table 2.1 and Fig. 2.4 for location and com-
rosition of regions; only regions which contain fuel are

considered here.

bActual temperature distribution was computed in

this region; see Figs. 2.9 and 2.10.

(1)
31

In Eq. (1), T is either Tf* or T*, and the integration takes place_over
the reactor volume. In thé defin¥tion of G(r,z), B, o, ¢i, and @, are
the fast and slow fluxes, and the fast and slow adjoint fluxes, respective-
ly; the G; 4 are constant coefficients for each region of the reactor in
which the composition is uniform and are related to the macroscopic two-
group nuclear constants.

The values for ¢ and ¢* are the unperturbed fluxes and adjoint fluxes
calculated at isothermal conditions in the reactor. In Eq. (1), the tem-
perature derivations associated with the function G are taken with respect
to Ty in obtaining the fuel-temperature-weighting function, and with re-
spect to Tg in obtaining the graphite function.

The temperature-weighting functions were evaluated from the MSRE,
using the Equipoise two-group fluxes, adjoint fluxes, and nuclear param-
eters. The results are shown in Figs. 2.11, 2.12, and 2.13. These corre-
spond to radial and axial traverses, which pass through the thermal flux
maximum in region J of Fig. 2.4. Discontinuities in the weighting func-
tions occur as the effective fuel volume fraction changes from region to
region, and thus reflect the change in local neutron multiplication, as
well as the variation in nuclear importance from region to region.

UNCLASSIFIED
ORNL-LR-DWG 73617

0.9

| S

0.8

N\
\ FUEL
AN

0.7 N
GRAPHITE\ \
0.6 \

0.4 \
%\
0.3 \

c.2 \

RELATIVE NUCLEAR IMPOR™ANCE
-

o 5 10 15 20 25 30
RADIUS (in.)

Fig. 2.11. Relative Nuclear Importance of Fuel and Graphite Tem-
perature Changes as a Function of Radial Position in Plane of Maximum
Thermal Flux.
32

UNCLASSIFIED
ORNL-LR-DWG 73618

1.0
[ I |

I N
EQUIPOISE 3A RESULT / \
08 sin2 ["/79.4 (2 + a1 +) /

Y8 os : / \ I
=z . Il
b |
E I
2 i
= o4 By
- / \ 3
< / \ | 1]
d 0.2 \ l ” \
2 / W
= \ \
Lé' 0 Jl - \"--._
:j ‘ | je———— MAIN PORTION OF CORE -
(i} |
L ) I

-o'2 |

-0.4

-06 :

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

z, AXIAL POSITION (in.)

Fig. 2.12. Relative Nuclear Importance of Fuel Temperature Changes
as a Function of Position on an Axis Located 7 in. from Core Center Line.:

UNCLASSIFIED
ORNL-LR-DWG 73619

1.0 I T I Py

—— EQUIPOISE 3A RESUL/. \
na L — —5in2 ["/79.4 (z+ 4.71} ] \

8 | / \
s
Z 06
|_
x
(o]
a
Z 54 .
N T \
3 /| \
> 02 N\
w \
v
|-—MAIN PORTION OF CORE m-mmrmormmeserm o m
—0.2
—04
-10 0 10 20 30 40 50 60 70 80

2,AXIAL POSITION (in.)

Fig. 2.13. Relative Nuclear Importance of Graphite Temperature
Changes as a Function of Position on an Axis Located 7 in. from Core
Center Line.
33

Use was made of these temperature-weighting functions to compute
nuclear averages of Tf and T¥ for the conditions of a 10-Mw power level,
with reactor inlet and outle% temperatures of 1175 and 1225°F, respectively.
The results obtained are given in Table 2.3, considering three levels of
fuel permeation of the graphite. Both the bulk average and the nuclear
average temperatures are given.

Table 2.3. Average Temperatures in Reactor Vessel at 10 Mw
as a FMunction of Fuel Permeation of Graphite

Fuel Temperatures Graphite Temperatures
Fuel Permeation » (°F) (°F)
(% of graphite
volume ) Nuclear Bulk Nuclear Bulk
Average Average Average Avcrage
0 1213.1 1198.0 1256.6 1226.3
0.5 1258.4 1227 .4
2.0 1263.7 1230.7

Since axial heat transfer in the graphite was neglected, any heat
produced in the graphite was considered transferred to the fuel immedi-
ately adjacent to it. As a result, the gross fuel temperature distri-
bution was unaffected by fuel permeation. It was assumed that the shapes
of the temperature-weighting functions were also unaffected so that the
different degrees of fuel permeation did not produce any change in the
fuel nuclear average temperature. The changes in the graphite nuclear
average temperature resulted from changes in the temperature level of the
graphite.

Reactivity Coefficients

Temperature Coefficients

Consistent with the definition of the nuclear average temperatures,
importance-averaged temperature coefficients of reactivity for graphite
and fuel were calculated. The equations for the coefficients were based
on first-order perturbation theory and involved the function G(r,z) given
previously.

Results of calculations for fuel containing no thorium, based on per-
turbation theory are compared below with those based on a one-region
homogeneous-reactor model employed in previous studies.® The temperature
coefficients of reactivity, in units of (6k/k)/°F, were:

S'MSRP Prog. Rep. March 1 to August 31, 1961," ORNL-3215, p. 83.
34

For Fuel For Graphite
Perturbation theory | -4 .46 x 107° -7.27 x 107°
Homogeneous reactor model -4.13 x 10*S -6.92 x 10-5

The differences in the results from the two methods are small and
well within the probable error associated with the value of the temperature
coefficient of expansion of the salt. The coefficients from the homogeneous
model do depend strongly, however, on the correct choice of effective re-
actor size. These calculations also account for the effect of the salt
temperature on fast-neutron leakage; taking this into account has the net
result of increasing the fuel coefficient slightly relative to the graphite
coefficient.

TPowel Cocfficient

The power coefficient of reactivity is defined here as the amount of
reactivity reswlting from nuclear heating that must be compensated by con-
trol rod motion per unit increase in power. 'This may be calculated from
the relations between nuclear average temperatures, fuel inlet and outlet
temperatures, and reactivity coefficients. At a 10-Mw thermal power,
these relations are

T = °
o Tp; + 50°F (2)
T = T, - 11.9°F (3)
f = “fo ?
1 = T+ 31.6°F | (4)
. g fo ?
-F ¥ ‘ - * .
Ak = w(4.45x 10 5) AT, = (7.27 x 10 5) aTg , (5)
where
Ak = reactivity change associated with changes in
nuclear average temperatures,
Tfi = fuel inlet temperature,
Tfo =- fuel outlet temperature,
i
Tf = fuel nuclear average temperature,
¥ .
Tg = graphite nuclear average temperature.

In obtaining power coefficients of reactivity, three cases were con-
sidered: (1) the fuel outlet temperature was held constant as the power
was raised to 10 Mw; (2) the mean of the inlet and outlet temperature,
T,., was held constant; and (3) the outlet and inlet temperatures were

lowed to vary to maintain criticality with no control rod motion. The
35

initial temperature at zero power was assumed to be 1200°F; the reactivity
change associated with increasing the power to 10 Mw was calculated for
the three cases considered. The results are presented in Table 2.k4.

Table 2.4. Temperatures at 10 Mw and Associated Power
Coefficients of Reactivity

m ¥ *
lIIfo Tfi Tf Tf Tg POYe?
Condition : Coefficient
(°r)  (°F)  (°F)  (°F)  (°F)  [(% ok/ik)/Mw]
Toy = constant 1200 1150 1175 1188.1 1231.6 -0.018
T’f = constant 1225 1175 1200 1213.1 1256.6 -0.0LT
Teys Tpy varied 118%.9 1134.9 1159.9 1173.0 1216.5

Effective Delayed-Neutron Fractions

Fuel circulating at 1200 gpm spends 9.4 sec in the core and 16.4 sec
in the remainder of the loop. Conseguently, a large fraction of the longer
lived delayed neutrons will be emitted outside the core and effectively
lost. Furthermore, the distortion of the source distributions that will
result from the transport of the precursors will increase the leakage
probability for the neutrons emitted in the core. The reduction in the
contribution of B;, the delayed neutrons of group i, to the chain reaction
would have a strong effect on the reactor kinetic behavior. The computer
programs that have been developed for MSRE kinetics calculations (Murgatroyd
and Zorch) do not treat the delayed neutrons rigorously. Instead, they use
the neutron and precursor equations for a noncirculating system, with values
of B adjusted to allow for losses resulting from circulation. Thus use of
these programs requires previous evaluation of effective B values.

Some calculations were made to determine the effect of the energy and
spatial distributions of the delayed neutrons on the effective values of
B: in steady-state operations.lo For each group, the spatial distribution
of the delayed-neutron source was calculated from the precursor production
distribution (which was assumed to follow the fundamental mode of the neu-
tron flux), taking into account the transport of the precursors prior to
their decay. The core was represented as a single region, with uniform
fuel volume fraction and velocity equal to the averages for the actual MSRE
core. Figures 2.14 and 2.15 show calculated source distributions along a
channel through the core. (The densities are per unit volume of fuel salt,
normalized to one fission neutron, and apply to a channel where the flux

1%p | . Haubenreich, "Effective Delayed Neutron Fractions in MSRE, "
ORNL-TM~380 (in preparatlon)
(x107%)
2.8

T 2.4

E

=

>-

=

W

z 2.0

w

o

w

(&)

@

3 -

3 {.i3

=

o

(144

}_

=

w {.2

=

[&]

w

>

a

m

L 0.8

V3]

2

=

<

o 0.4

w .
0

Fig. 2.14. Axial Distribution of Delayed-
Neutron Source Densities in an MSRE Fuel Channel.

"JNCLASSIFIED
ORNL-LR-DOWG 7:362)

I
GROUP 74, (sec)

4 56

2
4 2.3
5
6

(SEE FIG 2.i5 FOR GROJP
4 DISTRIBUTION)

.

S

!,
//' R
v _

/ 6

7

/

0.4 0.6 0.8 1.0
RELATIVE HEIGHT

Fuel circulating at 1200 gpm.

UNCLASSIFIED
ORNL-LR-DWG 72621

(10793
6
2 2
5 / \
> / \
g) /'S"ATIC \
Wl
S 4 / A
¢ / 7\
3 / / \
& 3 / ‘ \
3 / /. \
uzJ / / CIR(?ULATING \
o 2 / \
3 [/ \
2 |/ / \
[a]
¢ LT \
= | A \
2|/ \
0
0 0.2 0.49 - 0.6 0.8 1.0
RELATIVE HEIGHT
Fig. 2.15. Effect of Fuel Circulation on

Axial Distribution of Source Density of Group 4

Delayed Neutrons.

Fuel circulating at 1200 gpn.

9¢
37

equals the radial average.) The ratio of the nuclear importance of a de-
layed neutron to the importance of a prompt neutron was taken to be the
ratio of the appropriate nonleakage probabilities. The nonleakage proba-
bility for neutrons of a particular group was calculated by representing
the source distribution by the series

S,2) = ) ) Sigy JoUp/R) sin (am/m) (6)

m=]l n=l

and assigning to the m,n term a nonleakage probability given by

P, = < (7)

and
£ (2F (1)

where S.(r,z) is the source spacial distribution for delayed neutrons
of ith group, S;,, orc cxponsion coefficients, j, is the mth root of
the zero order Bessel function, Jg5, R is the radius of the reactor,

H is the height of the reactor, L is the thermal diffusion length, and
T. is the Fermi age for delayed neutrons of the ith group. The age, Tio
wds reduced from that for prompt neutrons to allow for the lower initial
energy of the delayed neutrons. (When the fuel is not circulating, this
effect enhances the importance of the delayed neutrons by reducing the
nonleakage probability.) Results of the calculations are summarized in
Table 2.5.

Reactor Kinetics Code Development

Preliminary results were obtained from a new reactor kinetics program,
Zorch, in which approximate account is taken of the space dependence of
fuel and graphite temperatures.l! This is an improved version of the
Murgatroyd program used in previous studies. In the Zorch calculation,
the reactivity was related more realistically to the temperature profiles
by using the concept of nuclear average temperatures. During an excursion,
the axial pover-density shape was assumed to be that of the fundamental
mode, that is, sin nz/H. The axial temperature profile was recalculated
at each time step, taking into account energy production, transfer, and
transport. In the radial direction, the magnitudes of the temperatures

11c. W. Nestor, Jr., "Zorch, An IBM-7090 Program for Analysis of
Simulated MSRE Power Transients with a Simplified Space-Dependent Kinetics
Model," ORNL-TM-345 (in preparation).
Tablz 2.5, Deiayed Neutrons in MSEE
Group -
Total
1 2 3 L = 6
Half-life, sec 55.7 22,7 .22 2.30 0.61 0.23
B, (see foot- 2,11 x 107%  14.02 x 127*  12.5% x 107%* 25.28 x 1¢"*  7.40 x 107%*  2.70 x 10°* 6k x 107*
note a) _
¥*
B;, static (see 2.23 x 107%  14.57 x 19™* 13.07 x 107* 26.28 x 1¢™* 7.65 x 10”* 2.80 x 10”* 67 x 10°%
footnote ) - -
6;,'circulating 0.52 x 107% 3.73 x 10~* k.99 x 107% 16.98 x 10™* 7.18 x 107* 2.77 x 107* 36 x 107*

®precarsor yield per fission neutron.

bEffective fraction of nesutrons in a given delay group.

8¢
UNCLASSIFIED

UNCLASSIFIED ORNL—LR—DWG 73623

ORNL-LR-DWG 73544 1500
1450 //“\
/ \ TIME
e /\\\ 3.0sec
§ T / N 2.0sec
= 1300
g / \ {.0sec
s Osec
///(/’—_
12€0
/
1200 % //
% INITIAL POWER =10Mw
1150
0] 0.2 04 0.6 08 1.0
[o0) Q1 Q2 Q3 oy Qs Qa6 Q7 Qa8 Q97Z/H 1.0 FRACTION OF DISTANCE ALONG CHANNEL
Fig. 2.16. Output of Calcomp Plotter: Fuel and Graphite Fig. 2.17. Fuel Temperature in
Temperature Distributions in Hottest Channel with Reactor at Hettest Channel Following One Dollar

10 Mw. Step Insertion of Reactivity.

6¢
40

and power density varied with time; however, the power density shape was
assumed to remain the same as initially.

Results of calculations for the MSRE with fuel containing 1% thorium
were obtained for a test case in which one dollar of reactivity was added
instantaneously (Beff was assumed to be 0.0034). The initial power level
was 10 Mw. The maximum rise in the fuel nuclear average temperature was
96°F. (This was about 4°F higher than the maximum rise in the fuel mean
temperature computed in the Murgatroyd program.) Also the rate of rise
of the nuclear average temperature was higher than the rate of rise of the
mean reactor temperature. The maximum power level calculated with the
Zorch program was T8 Mw, as compared with 107 Mw from the Murgatroyd
calculation.

As an option in the Zorch program, the temperature profiles in fuel
and graphite at specified time steps may be plotted using the Calcomp
digital plotter. The temperature profiles at t = O, obtained for the test
case, are shown in Fig. 2.16. The change in the fuel temperature with
time following the reactivity insertion is indicated in lig. 2.17.
41

3. COMPONENT DEVELOPMENT

Freeze-IFlange Development

A pair of INOR-8 freeze flanges for 5-in.-diam sched-40 pipe was
installed in the freeze-flange-seal testing facility® for measurement of
thermal distortion and gas leakage during high-temperature operation.
Following the successful completion of this series of tests, each of two
flanges was mated with new INOR-8 flanges and placed into actual salt
service. ‘lhe several tests are described below.

Freeze-Flange-Seal Test

The flanges mounted in the seal-testing facility were thermally
cycled in air, and precise measurements were made of flange dimensions
and of helium leakage at the ring-joint seal. This preliminary testing
consisted of 10 cycles from a cold condition to a hot eguilibrium con-
dition. The equilibrium temperatures were varied at 1O0CF intervals from
800 to 1400°F. Heat was provided by 12 silicon carbide heating elements
in the bore of the flange and by removable pipe-heaters2 along the ex-
terior of the flange adaptor, beginning 2 l/h in. from the flange faces.

At a bore temperature of 1400CF, the radial position of the freeze
point was 7.l in., which is believed to be satisfactorily far from the
ring-gasket radius of 10.4 in. The average decrease of the freeze-point
radius was ll/16 in. for cach 100°F decrease in bore temperature.

The external thermal distortion of the 5-in. INOR-8 flanges, both
at temperature equilibrium and during temperature transients, was found
to be much less than the thermal distortion of the 6-in. Inconel flanges
tested previously.® (The flanges differed only in material and bore
diameter.) The INOR-8 flanges operated cooler than the Inconel flanges,
as shown in Fig. 3.1. The distortions under equilibrium conditions are
given in Table 3.1. Starting with equilibrium bore temperatures of 1300
and 1400°F, all power to the silicon carbide heating elements in the
INOR-8 flanges was cut off to produce a transient simuwlating a salt dump.
When the bore temperature reached 850°F, the maximum external distortion
wa.s 6 mils, indicating that little distortion is induced during tempera-
ture transients. This limits the amount of frozen salt that could be
trapped between the flange faces during a rapid cooldown and which
could lead to permanent flange distortion. No permanent distortion was
noted at the outside of the flanges. Upon opening the flange joint, how-
ever, the inner faces were found to be slightly convex; the axial dis-
placement of the center of the bore was 1k mils.

1"MSRP Prog. Rep. March 1 to Aug. 31, 1961," ORNL-3215, p 20.
2"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p 26.
3Tbid., p 32.
42

UNCLASSIFIED
ORNL-LR-DWG 73982

1400 S
\\
1200 §}
\
\
\\
1000 °-\ S
™ \\ N
& [ ___850°%F NN N N ]
W oo [LFREEZING POINT | N\ e [N
D \ 4
g :F ‘\:\\w
@ . )
600 Se 0N
i &
\\‘D'\o
400 | FLANGES:
o INCONEL
e [NOR-8
200
0
0 2 4 6 8 10

DISTANCE FROM PIPE CENTER LINE {in.)

Fig. 3.1. Temperatures of Inconel and INOR-8 Freeze Flanges.at Com-
parable Distances from:Center Line of Pipe. Flanges tested in air.

Table 3.1. Maximum External Thermal
Distortion of Freeze Flanges at
‘Various Bore Tcmperatures

Equilibrium Bore Distortion (mils)

Tempgrature 5-in. INOR-8 6-in. Inconel
(°F) F]-angea Flangeb
1400 22 63
1300 18 6.

1200 12 pal
1000 8 2k

®Distortion measured 4.50 in. from center line.

bDis‘cortion measured 5.33 in. from center line.
43

Two nickel ring-joint gaskets were tested. For the first five
cycles, an oval ring gasket was used; for the second five cycles, an
octagonal ring gasket was installed. Indicated leakage, with the buffer
space evacuated, was very low at all times. There was a tendency for the
leak rate to increase as the number of cycles increased and to decrease
at higher temperatures. Leak rates at the outside of the buffer groove
were consistently higher than the leak rates to the bore through the inner
seal of this buffer zone.

Specifically, for the oval ring the cold leak rate steadily increased
from 3.9 x 10-8 to 1.4 x 10-4 cm3/sec after five cycles. There was
initially no detectable leakage at high temperatures, but after five
cycles the total leak rate was 1.5 x 10-2 cm3/sec at 1200°F. For the
octagonal ring the initial cold leak rate was 8.5 x 10-7 cm3/sec, and after
five cycles, it was 4.6 x 10-6 cm3/sec. At 1200°F there was no initial
detectable leakage, but after five cycles the leak rate had increased to
3.0 x 10-7 cm3/sec at 1300°F. For this brief period of testing, the per-
formance of the octagonal ring was superior to that of the oval ring, but
both performed far better than required.

Thermal-Cycle Test Loop

Immediately following the preliminary testing in the secal test
facility, one of the tested flanges was paired with a new flange, and the
pair was installed in the thermal-cycling facility.? The freeze-flange
thermal-cycling facility permits testing of the flange joint under drain-
and-fill conditions, steady-state conditions at various operating tem-
peratures, and temperature cycling between room temperature and a selected
operating temperature. The flange testing consisted of thermal cycling
and of attempts to form a frozen salt plug at the flange bore.

Forty-three cycles have been completed to date. A cycle starts with
the flange bore at ambient temperature and the sump and pipe temperatures
held constant. The freeze flange is then preheated until temperature
equilibrium is reached (~ 8 hr) at a bore temperature of 650°F; molten
salt at the desired high temperature of the cycle is introduced and
oscillated between the two sumps until the flange again reaches tempera-
ture equilibrium (~ 5 hr); and, finally, the flange is allowed to cool
until the bore is at ambient temperature (~ 10 hr). The total cycle
requires approximately 22 hr to complete. The first 15 cycles were made
with an average flange bore temperature of 1020°F and an ambient tem-
perature of 85°F. For the following 28 cycles, the flange bore tem-
perature was increased to 1250°F and the ambient temperature to 150°F.

The size of the salt ring was predicted on the basis of the flange
temperature distribution, shown in Fig. 3.2, as obtained during cycle
No. 35, and the expected 20°F temperature difference between the external
and internal flange faces, which was noted in earlier tests. The salt

4"MSRP Quar. Prog. Rep. July 31, 1960," ORNL-301Lk, pp 24-25.
UNCLASSIFIED
. ONNL-LR-DWG 73983
1400

1200 \

' A
1000 O

SALT FREEZE

800

600 NS

TEMPERATURE {°F)
VY

400 I o TEATEG WITH SALT IN | HERMAL-UYULE TEST LODP ~

& TESTED WITHOUT SALT IN SEAL TESTING FACILITY

200 — NOTE: TEMPERATURES RECORDED AT THE OUTSIDE

FLANGE FACE

L

0 2 4 6 8 10 2
DISTANCE FRUM PIPE CENTER LINE (in}

Fig. 3.2. Temperature Distribution of a 5-in. Freeze Flange in the
Thermal-Cycle Test Loop Compared with the Distribution in Air.

freezing point of 8LOCF was reached at an external face temperature of
0z0”F, whilch occurred at a diameter of 10.8 in. in the current series

of tests. Examination of the salt screen following cycle No. 36 (Fig. 3.3)
showed the salt-ring diameter to vary between 10.5 and 11 in. At the
greatest diameter the salt was still 4 in. from the ring.

Two mechanical dial indicators mounted on & small lathe bed were
employed to measure thermal distortion of the flanges. With the flange
bore at equilibrium at 1250°F, the maximum combined thermal expansion and
distortion was 25 mils. The calculated thermal expansion was 21 mils,
and thus the maximum thermal distortion was 4 mils. No permanent distor-
tion of the outer flange faces was detected in the cold flanges after
30 cycles.

There was a small increase in the interior face warpage of the female
flange from the initial out-of-flatness of 14 mils to 18 mils after
36 cycles. DNo warpage was detected on the male flange face. After 36
cycles all bore dimensions were checked and found to be undersize. The
female bore decreased by 10 mils, the male bore by 30 mils, and the male
lip by 8 mils.

For the first 36 cycles, an octagonal seal ring was used. The initial
cold seal leak rate for the ring was between 10-% and 10-3 cm3/sec. After
the flange joint had reached its initial hot equilibrium temperature
(1020°F), the hot leak rate was less than 5 x 10-9 cm3/sec. When the
45

UNCLASSIFIED
PHOTO 38664

S O0CTAGONAL
RING

INCHES
F 1R R T B R T e e 11%
}111 | i!M l]in%;Ll:iyhlzhIzislsh!!l

Fig. 3.3. Freeze Flange Salt Ring After 36 Cycles.

flange again reached ambient temperature, the cold leak rate had decreased
to 2.6 x 106 cm3/sec. After 36 cycles the hot lesk rate had increased to
3.9 x 10-7 cm3/sec and the cold leak rate was 1.8 x 10-4 cm3/sec; both
these leak rates are considered to be acceptably low.

An oval ring was tested in the next 7 cycles. The precycling cold
leak rate of 3.7 x 10~7 cm3/sec had increased to 1.7 x 10-2 cm3/sec at the
begimning of cycle No. 37. The final hot leak rate at the end of cycle
No. 43 was 9.3 x 10-3 cm3/sec. oince this rate was considered to be only
marginally acceptable, the flanges were opened for inspection. It was
found that foreign material had interfered with proper seating of the
ring. A secoud oval ring was inserted and is presently being tested.
46

Freeze Test in Thermal-Cycle Test Loop

In order to simulate a pump failure, the salt was held in a stagnant
condition between the two sumps to determine whether a freeze plug would
form at the flange bore. In the most severe test, the piping was kept
heated at 1000°F. Although the flange bore temperature was reduced to
350°F under these conditions, only a partial plug formed. It was con-
cluded that the MSRE 5-in.-pipe flanges would not plug so long as the pipe
heaters were operative.

Test of a Freeze Flange in the Prototype-Fuel-Pump Testing Facility

The second INOR-8 flanged joint was installed in the pump-testing
facility® (described later in this chapter), with an oval ring-joint
gasket. ''o date the facility has been operated approximately 300 hr
with three complete cycles from startlup to shutdown. The tempcrature dis-
tribution of the flanged joint has closely approximated the distributions
measured in both the seal-testing facility and the thermal-cycle Llesl
loop. No distortion has been noted, and no leak rates greater than
h x 1o=4 cm3/sec have been detected with the buffer seal system of this
joint. The buffer type of seal system used is similar to the reactor
leak-detection system. Although the test results are encouraging, addi-
tional experience is needed before the current freeze-flange design is
considered acceptable.

MSRE Control Rod

A full-scale model® of one MSRE control rod and drive was assembled
and operated over a range of temperatures from 500 to 1300%F. The drive
is shown in Fig. 3.4. The unit has operated through 24,473 cycles
(56 3/4 in. full travel in each direction). The rod has been dropped
from various heights 862 limes.

While over-all operation has been satisfactory, the drive unit failed
when one of the lower idler-sprocket bushings galled as the result of a
foreign particle. This bushing was replaced with a shrouded ball bearing,
and operation was resumed. Tests indicate that the rod position is known
within #1/8 in. 94% of the time and t1/4 in. 100% of the timc.

Heater Tests

Pipe Heaters

Testing of removable hardboard-insulated heaters for 5-in. pipe was
completed.” The units operated a total of 6136 hr with a pipe temperature

S"MSRP Semiann. Prog. Rep. Feb. 28, 1961," ORNL-3122, p 51.
6"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, pp 24-26.
“Ibid., pp 26-29.
47

e

[-CONTAINMENT POSITION i~SPROCKET MAGNETIC \GEAR BRAKE \-COOLING- |-MOTOF\’
VESSEL TRANSMITTER AND CLUTCH REDUCER AIR LINE
CHAIN

Fig. 3.4. Prototype Control Rod Drive Viewed from Above.

under the heaters of 14LO0°F. The average power input was 520 watts per
linear foot of heated pipe. The over-all functional operation of the
heaters was excellent, but the hardboard exterior tended to crack and dust,
the soft mineral wool lining dusted after heating, and the metal liner
warped. Three alternate designs being considered to overcome these
objections consist of units made with blown fused silica (Glasrock Mfg. Co.),
reflective metal insulation (Mirror Mfg. Co.), and combinations of these
materials. The materials for use in the MSRE will be selected after tests
of the various combinations. :

Testing of two prototype heater boxes using the Glasrock material
indicated an excessively high heat loss of 700 watts/ft. The surface tem-
perature of the heater boxes was 2LO°F when the heaters were operated at
lMOOOF; this would lead to excessive heat losses in the MSRE cell. Further
the material is unable to withstand physical abuse without cracking or
chipping. On the other hand it has the lowest coefficient of expansion
of any of the materials tested to date and thus has very high resistance
to thermal shock. There was no evidence of warping at the joint closures
at the completion of the test.

All-metal prototype heater boxes made of reflective insulation have
been purchased from Mirror Mfg. Co. for testing, and a test unit made of
Glasrock, board insulation, and metal reflective sheets is being fab-
ricated.

Activation Analysis of Insulating Materials

Most high-temperature insulation materials will dust to some extent,
and the dust will present a problem in reactor maintenance if it is highly
activated. The insulating materials listed in Table 3.2 were therefore
tested for activation following irradiation for 16 hr at a neutron flux
48

Table 3.2. Results of Activation Analysis of Insulating Materials

Observed Radioactivity Stable Element

Concentration
Sample Material
: : Activity Amount
Radionuclide (dis/sec-g)a Element (wt %)
Super Temp 260-d Zn®S 2.13 x 102 Znb4 7.8 (t0.2)
(Eagle-Picher Co.) 60-d Sbla4 1.91 x 10 Sbl23 0.0058 (£0.0005)
40-h Lal40 8.64 x 1073 1al3®  0.0019
5.27-y Co®©  1.99 x 102 Co5® 0.028 (*0.001)
85-d Sc46 1.85 x 103 Sc45 0.00014
45-d4 FeS® 5.60 x 104 Fe58 20.3 (%0.1)
Cerafelt LO-h Lal40 3.63 x 103 Lal3®  0.00008
(Johns-Manville) 85-d Sc4° 9.25 x 103 Sc45 0.0007
45.9 Fe5° 1.27 x 103 FeS8 0.46 (£0.02)
15-l Na? 1.10 x 108 123 U.060 (LU.003)
Unarco Board 60-d Sbl24 6.57 x 102 Sbl23  0.002
(Union Asbestos and 5.27=y Co8° 2.11 x 104 Co>®° 0.0003
Rubber Co.) 4o-n Lal40 2.26 x 103 Lal32 0.00005
85-d Gc46 1.98 x 103 Sc#5 0.00015
45-d4 FeS52 8.84 x 10° re>8 3.2 (£0.1)
15-h Na24 6.82 x 10”2 Na23 0.035 (+0.002)
Carey Temp LO-h Lal4® 3.06 x 103 Lal3® 0.00085
(Philip Carey Manufacturing 85-4 Sc46 1.12 x 10t Sc45 0.0008
Boe ) 45-4 FcS° 5.02 x 103 Fe58 1.82 (%0.02)
15-h Na24 8.57 x 107 Na23 4.40 (*0.05)
Fiberfrax LO-h Lal40 2.72 x 102 Lal3® 0.000006
(Carborundum Co. ) 15-h Na24 1.82 x 107 Na23 0.96 (%0.01)
Glasrock” 4O-h La4© 9.10 x 103 Lal3®  0,0002
(Glasrock Manufacturing Co.) 85-d Sc4® 3.96 x 102 Sc45 0.00003
15-h Na2% 7.50 x 10% Na23 0.0038
Glasrock Cement LO-h Lal40 1.36 x 10% 1al3®  0.0003
(Glasrock Manufacturing Co.) 15-h Na24 6.24 x 102 Na23 0.033

#Measured approximately 24 hr after reactor discharge.

bDetermined from the count rate observed for each radionuclide.

®Density: 25 1b/ft3.

of about 7 x 101t neutrons/cmz-sec. The data obtained (see Table 3.2)
indicate that Glasrock, which is nearly pure silica, may be the only
ceramic insulation that can be utilized uncanned in the MSRE.

Drain-Tank Coolers

The prototype cooling bayonets for removing afterheat from fuel in
the MSRE drain tanks were thermally shock tested 384 times without
failure. The test system was then shut down for installation of proto-
type thermocouples (see later section of this chapter) on the l1-in. cooling
bayonets that slide into the 1 l/Z-in. thimbles in the salt pool. 1In
previous operations, thermocouples installed in this area came loose in
a few cycles. The reactor-grade thermocouples now being tested are intact
after 6 cycles and 48 hr of exposure at 1L4OQCF.
49

A carbonate salt is being used in place of the MSRE fuel salt in
these tests to avoid beryllium-handling problems. The carbonate consists
of LipCO3, NapCO3, and KpCO3 (38-40-32 wt %). Its physical properties at
1200°F are given below, along with those of the fuel salt, for comparison:

Fuel Salt Carbonate Salt

Density, 1b/ft> 154.5 138.5
Average specific heat, Btu/1b*°F 0.455 0.391
Viscosity, cp 7.64 7.8

Sampling and Enriching System

Detail Design

Drawings of nearly all the components of the sampling and enriching
system® were issued for comment. Detail drawings of the transfer tube
and the spool piece between the disconnect flanges, with its positioning
tool, were approved for construction. Instrumentation flowsheets and
panel layouts were also approved for construction.

Mockup of Sampling and Enriching System

A full-scale mockup of the sample transfer box (area 3), the drive
box (area lc), and the operational valve is being fabricated for installa-
tion on the Engineering Test Loop. The transfer tube, which had to be
shortened to fit into the space available, has bends and slopes that are
the same as those to be used in the reactor system.

The mockup will be used to demonstrate that each component operates
satisfactorily and to develop adequate sampling and enriching procedures.
The sampling-capsule access-chamber mockup was successfully hydrostatically
tested at 90 psig to check the tightness and strength of the access port.
The copper sampling capsule was modified to simplify handling, and remote
removal of a sample from the revised capsule was demonstrated.

Core Development

Model Operations

The full-scale MSRE core model was operated at 85°F with water to
make preliminary measurements of the velocity distribution, wall heat
transfer coefficients, and solids-handling characteristics. Since the
viscosity of water is lower than that of the fuel salt, the model is not
exactly similar to the MSRE with respect to fluid dynamics. The pre-
liminary tests will be usetul, however, in planning later tests with a
more viscous liquid.

®Ibid., p 30.
50

Velocity Measurements

The peak resultant velocities at the bottom of the core-wall cooling
annulus were measured at various flow rates as a function of the angle
around the core. The data obtained are presented in Fig. 3.5. As may
be seen, the velocity profile is flat, indicating that the flow entering
the lower head from the annulus is distributed uniformly. Based on the
observed resultant velocity of k.55 ft/sec and the known axial velocity
component of 2.15 ft/sec at 1200 gpm, the average tangential velocity
component is k.0 ft/sec at the bottom of the core annulus.

A discrepancy was observed between the flow in the one-fifth .scale
model and the full-scale model. In the one-fifth scale model, the flow
in the lower head was observed to have a large rotational component,
whereas, in the full-scale model, the flow was observed to be almost
purely radial. On examination of the one-fifth scale model, it was noted
that the swirl killers were not resting on the lower head but, rather,
wcre celevated 1/8 to l/h in. off the head. This allowed water to flow
underneath the swirl killers and maintain its rotational component. New
swirl killers were made for the one-fifth scale model, and the flow then
behaved the same as that in the full-scale model. With this change the
measured heat transfer coefficients at the center line of the one-fifth

UNCLASSIFIED
ORNL-LR-DWG 73984

L, . gl . l o 's
0 o I o
1200 gpm
4
°
Qv
v
pd
.>__ .
G 3
O
s
w ,
> a’
2
L Q >
it % o) 2 -0 1o o} Q
4 2 600 gpm
>
7
w
o
=
)
E ¢ — 3 <
A Fa 4
hd Y 300 gpm -
o
6] 90 {80 270 360

ANGLE AROUND CORE (deg)

Fig. 3.5. Resultant Velocity Distribution Around Bottofi of Core-Wall
Cooling Annulus at Various Flow Rates. -

Lt
51

scale model were higher than those measured previously, so the change was
inferred to be beneficial.

Measurement of Heat Transfer Coefficients

Heat transfer coefficients are indices both of wall cooling and of
wall shearing, which tends to sweep out sediment. Coefficients measured
in the lower head at a radius of 17 in. were found to be proportional to
the 0.8 power of the flow rate, with a value of 522 Btu/hr.ft2-CF at
1200 gpm. Extrapolating the measurements to reactor conditions, the coef-
ficient for the reactor at 1200 gpm was estimated to be 1500 Btu/hr-fta-OF.

As a check on the water heat transfer coefficient, the fluid velocity
profile next to the wall was measured at the same point the coefficient was
measured. The velocity profile shown in Fig. 3.6 was found. A heat trans-
fer coefficient can be calculated from the data of Fig. 3.6 by assuming
flat-plate geometry and that the distance between plates is twice the
distance from the wall to the point of peak velocity. The calculated heat
transfer coefficient for this case is 522 Btu/hr-ftz-OF, which is in
fairly good agreement with the measured value of 667 Btu/hr-ft2.CF.

Attempts wcre made to infer the heat transfer coefficient at a
position about 4 in. from the core center line from data obtained with a

UNCLASSIFIED
ORNL-LR-DWG 73985
3.0

n
O
h7¢

»
/

FLUID VELQCITY (ft/sec)
o
/

o NORTH POSITION N N
e EAST POSITION
0.5 | & SOUTH POSITION
A NORTH WEST POSITION
0
0 0.5 1.0 1.5 2.0 2.5

DISTANCE FROM WALL (in.)

Fig. 3.6. Velocity Profile of Fluid Near Wall of Lower Vessel Head
2L Rudius of 17 in.
52

heat meter. These attempts were unsuccessful because there was no pre-
dominant direction of flow at that position; rather, the flow was "gusty."
The flow was first in one direction and then another. It was determined
that the heat transfer coefficient would be lower than at a radius of

17 in., but the exact value was not obtained. Further attempts to measure
the heat transfer coefficient in this region will be made.

Experiments on the Behavior of Solids

Experiments were run to obtain data for use in predicting the
behavior of solid particulate matter in the fuel salt, especially its
tendency to settle out in the reactor vessel lower head. The experimental
procedure consisted of injecting 3 1b (1362 g) of sized iron filings into
the model inlet during operation. After the solids were added, circula-
tion of water was continued for various periods nf time (0 %o 8 hr). The
core was then drained, and the quantity of solids that remained in the
bottom of the reactor vessel was determined. As far as the solids werc
concerned, the loop was considered as a once-through system. The sedimen-
tation data obtained are summarized in Table 3.3.

Table 3.3. Collection of Sediment in Lower Head of Core Model
Following Addition of 1362 g of Iron Filings

Nominal Time Pump Running Weight of Iron

Run Mesh Screen Required Time After Filings Re-
Nd’ Size Aperture to Add Solids covered from
) of () Solids Added Lower Hcad

Solids (min) (hr) (g)
€ 150 1oL 35 0 127
2 150 104 35 8 0#
3 60 248 20 0 823
L 60 2L8 20 8 750
5 150 10k 30 71/2 02

%A11 solids swept out with water.

In the case of the fine solids, there appeared to be some tendency
for sediment to form during their first pass through the core. Additional

running time caused them to be picked up, however, and swept out.. 1In the
case of the coarse material, a larger fraction settled out on the first
pass through the core and very little was removed with additional pumping
time.

The 150-mesh solids that remained in the lower head appeared in a
concentrated pile near the drain line. On the other hand, the 60-mesh
solids, as shown in Table 3.3, left larger amounts of sediment that tended
to form a torus around the drain. These larger piles were 2 to 3 in. wide
and from 1/2 to 1 in. thick, and they were against the drain line on one
side. The results of these tests indicate that the MSRE drain line will
53

not plug as the result of sedimentation of solids. Although no sedimen-
tary materials are expected in the reactor, these experiments provided
confidence in the design.

MSRE Engineering Test Loop (ETL)

The engineering test loop® was modified to include a graphite con-
tainer in the salt-circulating loop, and most of the Inconel pipe was
replaced with INOR-8 pipe. The loop was then placed into operation with-
out graphite in the graphite container. Most of the previous zirconium-
containing flush salt was removed, and the drain tank was flushed with
newly made coolant salt. Information was obtained during 1500 hr of
operation on the operational characteristics of the graphite-container
access joint, on the chemical analysis of the salt being circulated, and
on the formation of salt deposits above the liquid level.

After receipt, inspection and cleaning of the graphite to be tested,
the loop was shut down for installation of the graphite. Also, as part
of the proposed graphite-treatment procedure,lo the flush salt was treated
with HF and Hp2 in the drain tank to lower its oxide content before it
contacted the graphite. Details of these tests are discussed in the
following sections.

Frozen-Salt-Sealed Graphite-Container Access Joint

The flush salt was circulated at 50 to 70 gpm and 1200°F through the
loop and the graphiie container while awaiting delivery of the ETL
graphite from the vendor. The graphite container, which is an 8-in.-diam
INOR-8 pipe positioned vertically in the loop,9 has a flanged closure that
incorporates a frozen-salt seal to prevent contact of molten salt with
the conventional metal oval-ring seal. This closure is similar to a
slightly larger one that will be used in the MSRE. A seal of this type
between graphite and meta) was tested previously.tt

The ETL graphite-container closure is an extension of the 8-in.-diam
graphite container with a 20-in.-long tapered plug inside to form an
annutus varying from l/h—in. radial clearance at the lower end to l/lé-in.
radial clearance at the upper end. The MSRE closure is 10 in. in diameter
and 28 in. long, with gaps of l/h and 1/8 in., respectively. The tapered
gap prevents axial movement of frozen salt as the result of surges in
salt pressure.

The seal was made and broken three times, and it was leak-tight
during 1500 hr of operation. Since gas is trapped in the annulus during

B S

®Ibid., pp 34-4O.

107, L. Crowley, "Preliminary Operating Seguence for Graphite
Testing in the ETL, " March 19, 1962, unpublished internal report.

117. L. Crowley and W. B. McDonald, "A Metal to Graphite Joint for
a Molten Salt System," ORNL CTF-60-8-82, August 17, 1960.
54

filling, either initial evacuation or latcr ovcr-pressurc or venting must
be used to force salt into the cooled region. On the ETL closure the gas
was partially vented after the loop was filled.

An interesting observation made while determining the minimum cooling
air required for the seal was that when the cooling air flow was reduced
below the critical value, the salt plug moved upward; however, instead
of moving uniformly around its circumference, in which case gas compression
would have limited its travel, the salt extended upward in one semi-molten
sector that was contained within frozen salt at the edges. This character-
istic precludes the use of trapped-gas compression for preventing salt
from reaching the metal oval ring unless the entire-joint is held at tem-
peratures above the salt melting temperature. The temperature distribution
and a diagram of the closure are shown in Fig. 3.7.

Salt Composition

In order to remove as much of the ZrF) from the system as possible,
the original flush salt was removed, the freeze valves were purged, and
new salt was introduced into the unused drain tank (previously designated
the fuel drain tank). Salt samples were removed from the pump while in

UNCLASSIFIED
ORNL-LR—-0OWG 73986

TN,
PI
N
N
gl
| / THER%&?OUPLE | . | | : : ' : | | I
Ak ‘ 18 [-—— 0 COOLING AIR FLOW RATE: 23.8 s¢fm -
] EXTERNAL COOLING —— {14 1 INTFRNAI + 97 FXTFRNAI }; 344kw RFMOVFN —
Yl ON 8-in.<DIAM 16 _ -
CONTAINER ur ® CONLING AIR FLOW RATE: 36.3 scfm
- THC T g | R (214 INTERNAL +14.9 EXTERNAL) ; 4.14 kw REMOVED
INTERNAL COOLING i/ £1 o —
ON TAPERED PLUG -+ — — — —————- 8E o .
11 N
| 4 o 12 \
A = \\\\
| TE- I ~eFo~ |
y: APPROX . N it S —
SALT RN o i e i i A I s Y
PSRRI LEVEL g 8 <3~
'y 42 .90 S L
) LA A NN :, O b \
: s hY
KX ; [ . \
INSULATION < INSULATION r 4 A
: 8CS 5|
RS S  v vvs ~~ 8¢ . O N N
EZE NaYaYa¥a¥a¥ala)) <
= 0 N . s
= 100 300 500 700 900 1100 1300
_¢ TEMPERATURE (°F)
- 0O 2 4
— ——=—=]
INCHES

Fig. 3.7. Temperature Profile of ETL Frozen-Salt-Sealed Graphite-
Container Access Joint During Operation at 1200°F with a Salt Flow Rate
ol 50 to 70 gpmn.
OXYGEN (ppm )

55

operation and from the drain tanks when the loop was drained. Chemical
analyses of the salt samples indicated that the zirconium content was
reduced from about 90,000 ppm at the end of the previous run to 500 to
1600 ppm. In comparison, dilution of 1 wt % MSRE fuel salt into the flush
salt would result in 945 ppm zirconium.

An investigation was made of the oxygen pickup in the salt as a result
of contact with loop-metal oxides. To simulate the expected condition of
reactor components after heat treatment, the INOR-8 loop and container
were heated to 1300°F and held for 12 hr with an oxygen-contaminated cover
gas. The loop was then filled with flush salt, and samples were taken for
oxide analysis. No increase in salt oxygen concentration was detected.
This is consistent with the expected increase of less than 1 ppm based on
weight gain of INOR samples that had received similar treatment.

The results of oxygen analyses during loop operation are presented in
Fig. 3.8. There is considerable scatter.of the data, but no definite
increase of oxide content is detectable. The loop was opened between
operating periods at 800 to 1100 hr, but it was purged and held at a low
bake-out temperature (220°F) before refilling.

UNCLASSITIED
ORNL-LR~-DWG 73987

O DRAIN TANK SAMPLES

® PUMP TANK SAMPLES HF TREATMENT)/—-—__\

I-Loop OPERATION— l-LOOP OPERATION——I ’-—I
700
600 Q : 2]

o)
®
® &
500 . Q o5
d 00
LN
400 & o o O O O~
.
S ¢ %P © ¢ o
300 a - )
o
200 g
APRIL 12, 1962 : 0
| o

100

e APRIL —ele—— MAY - JUNE - JULY le AUGUST -+

o |
0 500 1000 1500 2000 2500 3000
TIME { hr}

Fig. 3.8. Results of Oxide Analyses of Engineering Test Loop Salt
Samples Removed from Pump and Drain Tank During Operation and HF Treat-
ment.
56

HF Treatment

Because the small amount of zirconium in the flush salt is believed
to have lowered the oxide solubility to an unknown limit, the salt was
treated with HF to assure oxygen nonsaturation before operation with the
graphite. The first treatment (July 1962) of T4 hr with HF removed 42 g
of oxygen,1l2 equivalent to 175 ppm in the salt. Samples taken several
days later, however, indicated that the oxygen content had increased to
the original level. While awaiting further samples and to cover the
possibility of the oxide having precipitated and been relatively inacces-
sible to the HF, the treatment was repeated in August, as indicated in
Fig. 3.8. The second HF treatment lasted 38 1/2 hr and removed only
9.5 g, or L0 ppm, of oxygen from the salt. Although this quantity of
oxygen was higher than expected, the actual rate of oxygen removal at the
end of the treatment period was approximately l/lO of the maximum rate
near the beginning of the period. If there were oxide solids left in the
bottom of the tank, it could mean that the dissolving rate was very low
during the 400 hr between treatments. Another source of contaminetion is
a small unknown amount of oxygen that was admitted to the tank during the
introduction and removal of the dip tube used in the HF treatment. Further
study will be made of the effect of this source.

The rate of corrosion of the container walls by the HF was apparently
greater than during the treatment performed in the other tanks in
November 1961.13 There was a negligible increase in the salt chromium
content during operation of the largely INOR-8 system, which is to be com-
pared with the chromium content increase from 100 to 200 ppm during HF
treatment of salt in the Inconel drain tank. Considering the amount. of
tank surface exposed to the salt, the weight of salt treatcd, and the
number of hours of treatment, the removal of chromium is equivalent to
0.00084 in. total or 0.011 mil per hour ol HF treatment. The INOR-8
tanks of the MSRE are expected to be more corrosion-resistant.

Cold-I'inger Colleclions in Cover Gas

Some difficulty was experienced during operation of the loop with
plugging of the pump offgas line. Examination after the first 788 hr of
operation revealed a colliection of salt at the junction of the unbaffled
l/z—in.-pipe offgas connection with the DANA pump-bowl 1lid. Around the
baft'led gas-inlet line, there was a small but uncbjectionable ridge of
salt. A salt deposit that was of concern appeared 6 in. above the ligquid
level in the capsule stop area of the sampler-enricher guide tube.#*

A buildup there would interfere with the lowering of the sample capsule
during testing of the sampler-enricher mockup. Cold fingers were therefore

127, H. Shaffer, "Hydrofluorination of LiF-BeF, (66-34 mole %) in
the Sump Tank of the Engineering Test Loop," MSR-62-58, unpublished
internal report.

13 o ,
"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p 37.
14"MSRP Semiann. Prog. Rep. Feb. 28, 1961," ORNL-3122, p 59.
27

inserted into the gas space above the salt level in the pump and the drain
tank to determine whether these deposits were due to mist caused by
agitation in the pump or by the vapor pressure of the salt.

A l/M-in.-diam copper tube sealed at one end was inserted into the
pump bowl gas space and positioned 1 in. above the liquid level. The
tube was air-cooled internally for a period of 89 hr while the pump was
turning at 900 rpm in salt at 1100°F. The extent of the salt deposit,
which was distributed uniformly around the circumference, is shown in
Fig. 3.9. The deposit was examined petrographicallyt® and found to be

UNCLASSIFIED
PHOTO 57914

Fig. 3.9. Collection of Salt Particles on a 1/4-in.-diam Tube Cold
Finger Inserted in Gas Space Above DANA Pump of ETL.

2LiF-BeF2, the composition of the salt mixture being circulated. A similar
cold finger was inserted into the quiescent atmosphere of the drain tank

on two occasions for periods of 4l and 14k hr at 1200 and 1220°F,
respectively. Both tubes had a thin deposit, which was scraped off,
examined by x-ray diffraction,®® and delermined Lo be 90 to 95% BeFo.

The difference in chemical composition and weight of the deposits found

15R. E. Thoma, June-July 1962, private communication.
58

in thc pump bowl and the drain tank led to the conclusion that the pump
bowl difficulties were caused by an aerosol-type dispersion of salt
particles.

A cylindrical shield of sheet Inconel was then inserted into the
sampler guide tube so that it extended below the liquid level, and a
1/16—in.-wide longitudinal slit was positioned away from the pump impeller
to prevent liquid salt from being forced up into the shield. Examination
of the sampler guide tube again after 740 hr of operation revealed no salt
deposit in the capsule stop area.

ETL Graphite Treatment

The ETL graphite, which is the same shape as MSRE graphite except
that the pieces are only 50 in. long, was loaded into the ETL graphite
container July 13, 1962. The loading, as shown in Fig. 3.10, consisted

UNCLASSIFIED
PHOTO 38700

Fig. 3.10. ETL Graphite Container Showing Graphite Pieces.
59

of five stringers, complete with INOR-8 lifting knobs, plus 16 other
partial elements forming 16 flow channels. Several miscellaneous graphite
and metal samples were attached to the graphite holddown section of the
freeze-joint plug. The graphite (approximately 60 kg total) was dusted
with a vacuum brush and weighed before installation.

The container, pump, and loop were then sealed and evacuated at
room temperature to 130 p Hg through liquid-nitrogen-cooled traps for
24 hr. The total weight gain of the cold traps was 10 g. The system
was then pressurized with helium and heated. The container and graphite
were held at 1300°F for 9 hr while helium was being circulated by the
pump and also purged through a liquid-nitrogen-cooled trap. The weight
gain of the trap during this operation was 4.5 g.

The dry box, described previously,l® was positioned over the graphite
container following the cooldown. After the dry box was evacuated, leak
tested, and purged with dried argon, the container was opened and graphite
samples were removed for various types of oxygen analysis. The analytical
results, as well as the results of analyses of as-received material, will
be reported at a future date.

MSRE Maintenance

Maintenance of the MSRE will be accomplished, as discussed previously,®

primarily with semidirect methods. Remote methods will be used for opera-
tions which require large openings in the shield and produce radiation
levels so high as to preclude occupancy of unshicldcd areas near the
reactor cell. Almost all the in-cell operations will be performed semi-
directly with long-handled tools operated through bushed holes in the
portable maintenance shield. Viewing will be provided by special lights,
windows, and periscopes. Moving shield blocks and large pieces of
equipment into or out of the cell will be done remotely from inside the
shielded maintenance control room, using the overhead crane system with
direct and television monitoring. Design and development of the in-cell
equipment, the portable shield, and the maintenance tools and development
of the techniques for their use have proceeded simultaneously in order

to produce a coordinated system. A procedure that outlines the steps
necessary to accomplish a representative maintenance task was prepared.l®

Placement and Removal of Freeze-Flange Clamps

The freeze-flange clamps will be driven on and off the flanges by
hydraulic cylinders mounted in frames on the ends of tubular masts. A
framework consisting of an upper and lower beam, two vertical push rods,

16"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p LO.
17"MSRP Prog. Rep. March 1 to Aug. 31, 1961, " ORML-3215, p 13.°

18E. C. Hise, "MSRE Maintenance Memo A-1, Maintenance Philosophy, "
April 27, 1962.

-
60

and two ears welded on the horizontal center line of the flangel® will
remain in the reactor cell attached to the clamps. This framework will
serve to center the clamps, to eliminate tilting during disassembly, and
to simplify the maintenance tools. When the clamps are driven off they
will be stored on an installed bracket by a hook-type handling tool.

Miscellaneous Flange-Servicing Tools

Flange alignment will be effected by a system of jacks, hooks, and
brackets installed in the cell, and each flange will require custom
installation. The equipment has been designed, and a prototype system
will be built and tested. The long-handled tools for adjusting the system
were fabricated. :

A tool to handle replacement gaskets was designed and constructed,
and the complete gasket-installation procedure was successfully tested.
The covers to be placed on open flanges during maintenance were designed
and constructed, and their handling was demonstrated.

Remote Viewing

A 1.625-in.-diam periscope suitable for reactor use was purchased
from the Lerma Company. It will be used in mockup demonstrations. Appro-
priate sheaths, windows, mirrors, and lights are being designed for use
with the periscope.

Portable Maintenance Shield

Design work is nearing completion on a portable maintenance shield to
provide the MSRE with a facility for shielded semidirect maintenance
operations. The shield is designed to temporarily replace any Two
adjacent lower roof plugs over the reactor or the drain tank cells. A
series of built-in tool access ports provide space for multiple mainte-
nance operations.

The portable shield assembly consists of four basic components: slide,
modulcs, frame, and track. The slidc dimensions arc 8 1/2 or 13 ft
long by 5 £t wide and consist of three layers of 4-in.-thick steel plate
welded together to a total thickness of 12 in. of shielding. The slide
includes a circular cutout that accommodates a 35~-in.-diam plug. The
plug, in turn, includes a series of small cutouts that provide access for
various lights, windows, and tools. Another cutout, at one end of the
slide, borders the eccentric plug of an adjacent module attachment, as
well as a removable crane access module.

The modules are additional short sections of slide that can be added
to the main slide to allow positioning of tool holes above the desired
work area. The module adjacent to the slide contains a duplicate of the

19Dwg. No. E-GG-C-L0610, Freeze Flange and Clamps Assembly.
61

35-in.-diam slide plug mounted eccentrically within a 50-in.-diam plug.
This plug-within-plug arrangement provides necessary reach to permit tool
access over any point within the block opening. The hole layout within
the inner plug permits multiple access to the reactor freeze flange, as
specifically required for flange maintenance and for optical positioning
and alignment. The plugs are manually gear-driven to allow precise
positioning of their multiple openings. Pneumatic socket drivers may also
be employed for driving the gears. The modules and plugs are fabricated
from 12-in.-thick carbon steel.

The frame, constructed of heavy steel plate, serves as the support
for the slide. After the frame is installed over a lower roof plug, the
slide is moved to an end track to permit the removal of the concrete plug;
an opening 52 by 127 in. is provided. The slide is then returned to
position over the cell opening to provide complete shielding for personnel.
A 3/h-hp motor and gear-box combination mounted on the frame moves the
slide through a rack-and-pinion arrangement. The rate of travel is about
2 in./sec. Movable end stops are provided to limit travel as desired.
Motor controls are connected to the maintenance control room operating
station with a cable.

The track is an extension to the frame that can be mounted at either
end of the frame and on which the slide and module rollers move. -I1t con-
sists of I beams with channel and angle bracing. '

A shielding thickness of 12 in. of steel is provided throughout
(slide, modules, and frame sides and ends). Sufficient shielding is pro-
vided to limit radiation levels 24 hr following reactor shutdown to about
10 mr/hr. '

As many as eleven 150-w lighting fixtures can be mounted in special
plugs provided in the slide and the modules for special illumination of
the work area. All work holes are designed with stepped split inserts to
accommodate tools of various diameters. Some inserts include a drilled
lead ball through which tools may be "swung' into positions other than
directly underneath the work hole. Three 8-in.-diam lead-glass windovs
may be mounted simultaneously for direct viewing. Periscope and televi-
sion camera openings are also available.

Brazed-Joint Development

Brazed-Joint Fabrication

Six prototype brazed joints in 1 1/2-in. sched-40 INOR-8 pipe were

fabricated using tools and equipment developed for this purpose. One

joint was made in a mockup ot the drain cell with all the work accomplished
remotely, as it will be done in the MSRE. Another was made on a pipe in
which molten salt had been circulated and in which some salt remained. All
six joints were subjected to a hydrostatic test, a helium leak test, and
ultrasonic inspection. None evidenced any leakage, and all indicated
adeguate bonding. The development work on the joint and methods for its

JE g 1
62

remote fabrication is considered to be complete. Development work on
techniques and equipment for remote inspection is still under way.

Brazed-Joint Salt Test

A prototype 1 l/2-in. brazed joint was tested in service that simu-
lated a drain line. The joint was held at 1250°F for a total of 6780 hr,
including 123 hr in contact with salt and 402 simulated drains. There was
no apparent damage to the joint throughout the test. The joint was removed
from the loop for metallurgical examination.

Mechanical-Joint Development

The design of the overflow-line disconnect at the fuel pump bowl20
was completed, and a carbon-steel prototype was built tor testing. 'Lhe
disconnect is essentially a ring-joint flange mounted in a vertical line
with a cylindrical baffle extending downward into the pipe to trap gas
next to the ring gasket. )

A series of tests of the prototype at 1000°F and 70 thermal cyc¢les
with heat applied both internally and externally indicated favorable tem-
perature distribution and excellent leak tightness. Therefore an
improved INOR-8 prototype flanged joint was built and installed in the
freeze-valve test loop. Testing with salt is scheduled to start in the
near future. '

Pump Development

Fuel Pump Design and Fabrication

The design of the fuel pump was approved, as previously reported, =1
and. fabrication of the pump tank is approximately 30% complete. An INOR-8
volute casting was weld repaired to the required quality for reactor
service, and a spare volute casting is being weld repaired to the same
gquality. Fabrication of two rotary elements for the fuel pump is approxi-
mately 80% complete.

Coolant Pump Design and Fabrication

The design of the coolant pump®l was completed and approved. The
coolant and fuel pumps are similar in design, except that the hydraulic
characteristics are different and the coolant pump tank does not contain
the xenon-removal spray and the cooling shroud for the upper shell of the
tank used in the fuel pump.

The fabrication of the pump tank is approximately 30% complete. An
INOR-8 volute casting was weld repaired to the quality required for
reactor service, and a spare volute casting is being weld repaired to the

20"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p 51.
211pid., pp 51-57-
63

same quality. The fabrication of two rotary elements for the coolant pump
is approximately 80% complete.

Design and Fabrication of Lubrication Stands for Pumps

The design of the lubrication stands was completed and approved.
Two stands will be required, one for the fuel pump and one for the coolant
pump. The lubrication systems are designed to lubricate and cool the
pump-shaft bearings and seals and to transport heat from the shield plug
and the upper tank structure. The two stands will be interconnected to
provide time for orderly shutdown of the reactor in the event one of the
systems should become inoperative.

Each stand provides a water-cooled reservoir, two centrifugal pumps
in parallel, a full-flow oil filter, and the necessary valves and instru-
mentation to ensure adequate lubrication and cooling of the reactor pumps.
The fabrication of the stands is approximately 40% complete.

Lubrication-Pump Endurance Test

The test lubrication pump has circulated a turbine type of oil for
5413 hr at 160°F, 70 gpm, and 3500 rpm. Upon resumption of operation
after the initial test®! of 1428 hr, the motor failed after operating an
additional 913 hr, when one phase of the stator winding became grounded.
The stator was replaced and tecsting was rcscumed.

Drive Motor Design and Fabrication

The design of the drive motors for the reactor pumps was completed
and approved. The basic design consists of a totally enclosed, water-
cooled, explosion-proof, NEMA design "B," special-purpose, squirrel-cage
induction motor. The motor for the fuel pump is rated 75 hp at 1200 rpm,
and the motor for the coolanl pump is rated 75 hp at 1800 rpm.

The specification for the drive-motor contaimment vessels requires
that the vessels pass a mass-spectrometer leak test with a helium leak
rate of less than 1 x 10-8 cm3 (STP)/sec. The vessel fabricator has
experienced difficulty acquiring acceptable plate material; a second order
of the material failed to pass the impact-test requirements and is being
heat.-treated prior to retesting.

Test Pump Fabrication

Fabrication of the prototype pump tank was completed, and the tank
was installed in the test facility. Bench testing of the rotary element
was completed, and it was installed in the pump tank for elevated-
temperature testing.

Test Facility Construction

The construction of the test facility (Fig. 3.11) was completed. In
addition to the prototype pump and loop piping, it includes (Fig. Fu:12 )
PUMP
TANK
COOLING
SYSTEM

RIVE MCTOF

T

PUM-
LU3RICATION
SYSTEM

Prototyre Pump Test Fazility.

UNCLASSIFIES
g PHOTZ 33757

79
65

UNCLASSIFIED
ORNL-LR-DWG 72414

AIR COOLER
F

THERMAL
, WELL
( »” -.
DISCHARGE L
PRESSURE—] MSRE
l PIPE HEATER
FREEZE £
8-in. PIPE FI ANGF
VENTURI

METER ~6-in. PIPE

| "
—I— INLET  FREEZE VALVE l MSRE FREEZE
T PRESSURE \ FLANGE

THROAT I i
PRESSURC HAND VALVE

DUMP
TANK

Fig. 3.12. Prototype Pump Test System.

two freeze flanges, one of which is of the MSRE design, a section of MSRE
pipe heater, the cooling system for the upper shell of the fuel pump tank,
the nozzle for the sampler-enricher (not shown), the bubble-type liquid-
level devices (not shown), and a thermocouple installed in a well for
comparison with thermocouples attached to the external surface of the pipe.

Hydraulic Performance Tests

Hydraulic performance data were oblained with the 13-in.-diam impeller
circulating the salt LiF-BeF2-ZrF)-ThF)-UF), (70-23-5-1-1 mole %) at
1200°F on three different system resistance lines at speeds ranging from
600 to 1150 rpm. The hydraulic performance of the pump with molten salt
compares well with the performance with water. While data were being
taken on the last and steepest resistance line, the pump shaft seized
along the lower 4 in. of the shield plug.

The seizure, which was of the friction-weld type, was caused by
rubbing between the outer surface of the shaft and the surface of the bore
in the shield plug. Deflection of the shaft by radial hydraulic force on
the impeller had reduced the running clearance to zero just prior to
seizure. The pump was operating at an off-design condition well removed
from its balance line when the seizure occurred. The pump had operated
66

satisfactorily for 335 hr at 1200°F. Another rotary element is being
assembled in which the running clearances are increased to accommodate
off-design operation, and the test will be resumed. |

PKP Pump Hot Endurance Test

The PKP test pump continued to operate and had accumulated, at the
end of this report period, a total of 5445 hr at 12250F, 1950 rpm, and
510 gpm, circulating the salt LiF-BeFp-ThF),-UF), (65-30-4-1 mole %). The
differential pressure across the lower shaft seal has been maintained at
1 psi. The seal leakage across the upper seal has been approximately
5 cm3/day, and across the lower seal, it has been too small to measure
conveniently.

''est Pump With One Molten-Salt-Lubricated Bearing

The test pump with a molten-salt-lubricated bearing was equipped with
a new journal and bearing fabricated of LNOR-8. During shakedown ot the
test facility, it was noted that the hot, wetted end of a thermocouple
well (3/8—in.—diam tube) had corroded away. Operation will be resumed as
soon as an analysis of the fuel supply and a metallurgical examination of
the remainder of the thermocouple tube have been made.

Instrument Development

Single-Point Liquid-Level Indicator

A prototype model of the two-level conductivity-type liquid-level
probe being developed for use in single-point measurement of liquid level
in the MSRE storage tanks was constructed. The completed probe assembly
is shown in Fig. 3.13. The basic principle of operation of the probe was
reported previously;®* however, the structural design of Lhe probe differs
tfrom the conceptual design described previously in that the control-
sampling access hole and the mounting flange have been eliminated. The
changes were necessitated by changes in the design of the storage tanks
that precluded the use of the central access hole for level-probe instal-
lation. In the revised installation, the probe will be welded into a pipe
nipple and wall not be removable.

The probe assembly is being tested in molten salt under simulated
operating conditions. Initial results of the tests indicate that the
design concept is correct. Signals of 100-mv amplitude were received when
the sall was 1n coulacl wilhh Lhe probe. The signal level was less than
20 mv when the salt level was below the probe. It is expected that this
latter signal level can be reduced by balancing capacities and by use of
a phase-sensitive demodulating circuit in the receiving equipment. Testing
of the prototype probe will continue in order to determine the long-term
reliability of the device and to obtain data for use in the design of the

22'MSRP Prog. Rep. March 1 to Aug. 31, 1961," ORNL-3215, pp 72-76.
Pig. 3ul3,

Prototype Model of Two-Level Conductivity Probe.

UNCLASSIFIED
PHOTO 38613

L9
68

MSRE indicator. Detailed designs of probe assemblies for the MSRE are
being prepared.

Pump-Bowl Liquid-Level Indication

Developmental testing of a continuous liquid-level-indicating
element®> for use in measurement of molten-salt levels in the MSRE fuel
and coolant pump bowls was continued. A four-month initial test at teme
peratures between 900 and 13OOOF was terminated June 4. During the last
three months of this period the temperature was maintained at 1200°F, and
the molten salt was held at a constant level. The maximum variation
recorded during this period was O.l in.; the span of the instrument is
5 in. Tests made at the end of the four-month test period revealed no
evidence of calibration shift, hysteresis, or other degradation of the
performance characteristics of the instrument in excess of 2% of full
scale (0.1 in.). Calibration curves made three months apart are shown in
Figs. 3.14 and 3.15. There has been no indication of buildup of solids
as a result of vapor deposition in the core tube of the instrument; how-
ever, solid deposits were found on a spark-plug level probe removed from
the test stand at the conclusion of the four-month test. After completion
of the four-month test, the instrument was drained and the temperature was
cycled three times between 500 and 1000°F. Subsequent checks revealed no
evidence that the instrument was damaged by this repeated heating and
cooling. The instrument was returned to operation at 1200°F, and long-
term testing is continuing.

A third instrumcent is being constructed and will be inslalled on
the MSRE pump testing facility to obtain experience under operational
conditions approximating those which will be encountered in the MSRE.
This instrument will use an INOR-8 metal-ball float, since graphite will
not float in the barren salt used in this facility. The transformer will
be mounted above the float because this type of assembly is preferred for
the MSRE installation and because tests of a similar unit, described
above, revealed no objectionable operational characteristics.

Single-Point Temperature Alarm System

Investigations of single-channel thermocouplc alarm switches for use
in freeze-flange and freeze-valve monitoring and control systems con-
tinued. Several components of the Electra Systems Corporation=% moni-
toring system were received in late March. The system included a dual-
limit alarm module (ET-4210), two single-limit alarm modules (ET-4200),

a control module (ET-MBOO), and the power supply and module enclosure.

The units were tested in the laboratory. As previously reported, &4
the single-limit alarm modules appear to be suitable for freeze-flange

23"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, pp 61-66.
247pid.; p 60.
69

UNCLASSIFIED
ORNL-LR-DWG 73988

O CALIBRATION 6-6-62
90 |—
® CALIBRATION 3-16-62 {
TRANSFORMER [N 1225°F FURNACE
80 //
®
70 /
S 60
3
<<
Ll
o
=
n
T B0 °
o f
O
(®]
=i
<
=
& 40
30 o
20
10
o
0 »
4 5 6 7 8 9 10

TANK NO. 2 SCALE READING SHOWING MOLTEN-SALT LEVEL CHANGE (in))

Fig. 3.14. Calibration Curve of Recorder of Pump-Bowl Liquid-Level
Indicator No. 1 Before and After Three Months Operation at 1200°F.
70

UNCLASSIFIED
ORNL-LR-DWG 73989

100 ’ ’ ®
O CALIBRATION 6-6-62
90 |—
® CALIBRATION 3-16-62
TRANSFORMER ABOVE FURNACE
AND NOT HEATED
80 /#
°
{/
g 60
(&)
<
(0]
(0
[—.
2
< 50 /T
(&)
O} &
(&)
|
—<’ /
> 40 5
30
20
10
U
4 5 6 7 8 9 10

TANK NO. 2 SCALE READING SHOWING MOLTEN-SALT LEVEL CHANGE (in.)

Fig. 3.15. Calibration Cirve of Recorder of Dump-Bowl Liguid-Level
Indicator No. 2 Before and After Three Months Operation at 1200°F.
b

monitoring. Recent tests indicate that control module ET-4300 is
suitable for the valve-monitoring system. The evaluation thus far is
based only on laboratory tests.

The Electra System was placed in field operation on a freeze-valve
testing facility in mid-August. It is now undergoing full field tests
utilizing the control system to be used for the reactor. The Electra
System is composed of modular units, as previously described.2® The
internal wiring of one module card containing two single-alarm switches
is shown in Fig. 3.16, and the entire system is shown in Fig. 3.17. The
top unit is the power supply and the bottom unit the module enclosure.
The module enclosure will handle 20 single-limit alarm switches or 10
control modules.

Tests of the Daystrom Magsense Control Relay, Model A-82, were
completed. As previously reported, &2 this unit could be utilized for the
flange- and valve-monitoring systems; however, the individual units
would have to be integrated into a system. This would require adding
alarm-1limit lights, relays, and enclosure, as well as a common power
supply. Further, the input resistance of the units is so low (350 ohms)
that it could cause a problem with long thermocouple leads.

Temperature Scanner

Development work on the mercury-jet-commutator thermocouple-scanning=e
system is continuing. The switch was modified to reduce spurious noise
generation. An alarm discriminator was designed and constructed, and the
reference-thermocouple isolation system was improved and packaged with
the alarm discriminator. A differential dc amplifier with variable band-
width was received, and an existing 17-in. oscilloscope was repaired to
display the thermocouple signals.

The complete system was assembled and tested in the laboratory in
early May. A diagram of the system as it now exists is shown in
Fig. 3.18. The laboratory tests with simulated thermocouple signals
were successful, and the system was installed on a liquid-level test
facility in early August. One hundred thermocouples from the test
facility were used as input signals to the scanner. At present, the
system continues to operate satisfactorily after more than 700 hr of
operation. Operation will be continued to determine the operating life
of the mercury switch.

The diagram of the system (Fig. 3.18) shows the principle of opera-
tion. One hundred thermocouples are switched in sequence by the mercury-
jet switch at a rate of 2000 points per second. The output of the switch
is fed to an integrating system consisting of a capacitor and the re-
sistance of the dc amplifier. The signal level is held during the point-

25"MSRP Prog. Rep. March 1 to Aug. 31, 1961," ORNL-3215, p 25-27.
26"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p 59.
Fig. 3.16.

Dual-Limit Alarm Module.

UNCLASSIFIED
PHOTO 38688

cL
SERIES 4000
ypERAFINS MNITOR SYETS

vSTEMS COR
ELICT::L&YONV CALIFORNIA

AL roAtpn
powWER SUPFLY 5

£
pese -, G o B v 8
K - <

%

- R
¥

Pig. 317

1200-rpm
MOTOR

73

o, gr4008
fr-a00 WODEL "/

UNCLASSIFIED
PHOTO 38689

UNCLASSIFIED
ORNL-LR—DWG 58910 R3A

THERMOCOUPLES
ON REACTOR PIPING
. TWO DOURIL F-POI F
DOUBLE- THROW CHOPPERS
OPERATING SYNCHRONOUSLY n
‘ ’ [ ONE COUPLE ABOVE xREFERENCE
REFERENCE
je——1 { | | J”L' 5| THERMOCOUPLE
— |
L TN T
K< > R = O o e
< - :5:; | 2 , 00000
|

~N— P i ~

| ... - ~17-in. OSCILLOSCOPE
l«—REACTOR *REFERENCE APPROXIMATELY

PIPE EQUAL TO SCANNED
THERMOCOUPLE SIGNAL
| = o B
DISCRIMINATOR
ALARM ANNUNCIATOR
MERCURY-JET SWITCH SETPOINT
(100 CONTACTS) + 50°T Q
K=" = N
o REFERENCE
THERMOCOUPLE

ON REACTOR PIPE

fig. 3.18. Block

Diagram of Temperature Scan

§G 1 i
T4

to-point switching time. The reference thermocouple signal is fed to two
double-pole double-~throw choppers with a capacitor between the poles of
each chopper. The reference thermocouple is connected in opposition to
the output of the mercury switch. The choppers store the charge from each
side of the reference thermocouple during one half of the operating cycle
and then switch it across the large integrating capacitors in each out-
put line of the mercury switch. The difference signal is fed to the dc
amplifier and amplified to 1 volt. The output of the amplifier is fed to
the oscilloscope and to the alarm discriminator. The alarm discriminator
can be adjusted to produce an alarm when the difference signal exceeds an
adjustable range of 50 to t300°F. If the reference signal and the
commutated signal are equal, a straight line is seen on the oscilloscope.
If a signal from one thermocouple is higher than the reference, it is fed
through the system as an ac signal and appears as a positive pulse on the
oscilloscope. By using a synchronizing pulse from the switch, the 100
thermocouple signals are displayed on the oscilloscope in a fixed position
so that each signal is identified with a particular thermocouple.

The switch modification consistcd of installing shorting straps
between the coalescer coll and the common mercury pool, as shown in
Fig. 3.19. This reduced the static charges which were generated between
the coalescer coil and the mercury pool and resulted in a significant re-
duction in the spurious noise generated in the switch.

A block diagram of the alarm discriminator is shown in Fig. 3.20.
This circuit is designed to produce an alarm when pulse signals having an
amplitude of 1 volt or greater and a repetition rate of 20 pulses per
second are applied to the input. With proper adjustment of the differen-
tial dc amplifier, the discriminator may be adjusted to produce an alarm
with signals as small as t50°F. A pulse-integrating (count-rate) circuit
prevents spurious alarms from random noise pulses. Since the mercury
scanning switch samples each point 20 times per second, input pulses
resulting from a true alarm condition will have a repetition rate of at
least 20 pulses per second and an alarm will occur. The discriminator-
alarm circuitry is completely transistorized and is packaged in a
3 l/Z-in.-wide, T l/Z-in.—high, panel-mounted housing. Also packaged in
the housing are the reference-thermocouple isolation system, a sync-pulse
amplifier for driving the oscilloscope trigger, and the necessary power-
supply circuits.

Thermocouple Development and Testing

Radiwlor Thermocouple Tect. Teoting of mecchanical thermocouple
attachments for use on radiator tubes®’ in the MSRE was resumed after the
test apparatus was modified to eliminate errors in measuring the inner
wall temperature. Provisions for purging with inert gas were added; the
reference thermocouple was attached to the inner wall of the INOR-8 tube
by spot welding; and thermal insulation was placed between the thermo-
couples and the copper slug (Fig. 3.21).

=T Thid., p 98
UNCLASSIFIED
ORNL—LR—-DWG 72559

SIGNAL PIN ROTOR
ASSEMBLY HARNESS
GASKET,NOZZLE /ASSEMBLY
( .

NOZZLE

COUNTERWEIGHT

A II é 'Ill’ ‘
S
¢ Taks §v—=—, g% o1 L.
\\-

CLAMP ls ’\ Q /
ASSEMBLY ” .,‘ " // > COALESCER
"0 RING
\\\\\

SHORTING STRAP
\
~MERCURY POOL

|

1@'@ COO 00 1O
i

—_
e

. —— MOTOR ASSEMBLY

J\ Qg COOO00O A

Fig. 3.19. Model 210 Delta Switch.
76

UNCLASSIFIED
ORNL-LR-DWG 72558A

" | DISCRIMINATOR .y AT
INPUT SIGNAL TRISGER LSE
1 volt >+ 50°F SHAPER CIRCUIT
0
)
o
O]
I
O EMITTER

REI AY

RIVER
gé URIVE FOLLOWER

Fig. 3.20. Block Diagram of Alarm Discriminator.

UNCLASSIFIED
ORNL-LR-DWG 73990

REFERFNCE THERMQCOUPLE
SPOT WELDED TO WALL

THERMAL HEATER THERMOCOUPLE

INSULATION

INOR-8 TUBE INCONEL-SHEATHED TEST THERMOCOUPLE

Z22
m : 3
SOSNN N AN SN N S SO N SOOSNNN SN SNSRI
AT A T A AT 3 WS Lt Lo L ) 4
> 2 L7 7 Z Z

= ; P AT 3
? | NNNNN\N| 1%
7 e
% / . I . : _ASSESSXY
< aasSsSSsSs S SS OO IS VAR A R I I I T T T T T T T T I T T I T T I T M LTI T L LI LI LLL LALLM LS LL LU L R R R R,
PURGE-GAS
“ PURGE-GAS WEDGE-FITTED LARIRIDGE=- “INCONEL CLAMPING BAND INLET )
FXHAIIST COPPER SLUG TYPE HEATER ATTACHFN TN TEST THERMOCOUNLL
WITH GOLD-NICKEL BRAZING ALLOY
HEATER
AIR STREAM POWLR LEADS

Fig. 3.21. ©Sectional View of Radiator Thermocouple Test Apparatus.

Several test thermocouples were preparcd with one section of a Lwo-
piece clamp made of 3/16-in.—wide, 0.020-in.-thick Inconel attached to
the sheath near the end seal with gold-nickel brazing alloy. The thermo-
couples were then attached to the tube by placing the npper section over
the tube, engaging the lower section, and crimping both, as shown in
Fig. 3.82.

Several tests were run with Lhe thermocouples attached as described
above and with thermal insulation consisting of different thickncss of
Fiberfrax board or paper placed over the junction end of the thermocouple.
The thermocouple shown at the far right in Fig. 3.22 is insulated with
Fiberfrax paper, and the thermocouple second from right is insulated with
Fiberfrax board. The accuracy of the wall-temperature measurements with
- e eecem

UNCLASSIFIED
PHOTO 38840

L

Fig. Z.22. Radiator Tube Thermocouple Attachments.
78

the test thermocouple was improved considerably with both forms of ,
Fiberfrax insulation. The flexible paper form of Fiberfrax is preferred
over the rigid board, however, because it requires no machining or pre-
forming prior to installation. Typical results obtained in recent tests
of thermocouples attached and insulated as described are listed in

Table 3.4. The test thermocouple used in run 5 was covered with Thermom,
a heat-conductive cement rated to 1250°F, and 3 layers of l/8-in.-thick
Fiberfrax paper. It is not known yet whether Thermom will be acceptable
for this application. A test specimen of a typical thermocouple attach-
ment covered with Thermom is now undergoing a corrosion test.

Engineering Test Loop Installation. Eight MSRE-prototype surface-
mounted thermocouples were installed in the ETL facility to determine,
under simulated operating conditions, the reliability of attachments and
the accuracy of wall temperature measurements taken with thermocouples
located on the walls of pipes and components adjacent to heaters.

Sheathed, l/l6-in.-OD, single-conductor two-wire, and l/8-in.-OD, sheathed,
duplex thermocouples were mounted in pairs adjacent to 30-ga bare-wire
reference thermocouples at three locations. A similar pair of thermo-
couples was located adjacent to a reference thermocouple installed in a
well at a fourth location. A typical installation is shown in Fig. 3.23.

A1l the thermocouples have been in service since April 16, 1962, and
are still performing satisfactorily; however, they have accumulated only
1500 hr of operation time at 1040 to 1200°F during this period because of
loop downtime. With salt circulating at the above temperatures, differ-
ences in readings between the test thermocouples and the respective

Table 3.4. Results of Tests of Radiator Tube Thermocouples

Approximate Inner Wall Tect Thermocouplc Tcmpcraturc

fig?& Air Flnw Tpmpgrflturc Temperature Diffgrcncc
(£ps) (°r) (°F) (°F)
1 0 996 985 -1l
90 990 95k -36
% 0 e uslk -1k
70 95 905 -30
g 0 1009 996 -13
50 1009 983 -26
L 0 1009 996 -13
50 1000 972 -28
5 0 1007 1002 -D

50 1010 1001 -9

“The test thermocouples used in runs 1, 2, 3 and 4 were insulated
with Fiberfrax board; the run 5 test thermocouple was covered with
Thermom and insulated with Fiberfrax paper; all readings were taken with
the test thermocouple downstream of the air flow.
Fig. 3.23.

Thermocouple Installation on Engineering Test Loop Piping.

UNCLASSIFIED
PHOTO 38195

6L
80

reference thermocouples, as noted periodically, were inconsistent. The
limits of variation in temperature difference between an individual test
thermocouple and its respective reference thermocouple are listed in
Table 3.5.

Drift Test. Six Inconel-sheathed MgO-insulated Chromel-Alumel
thermocouples are being tested to determine the stability of calibration
at MSRE temperatures. Over 5000 hr of soaking time in 1200 to 1250°F air
has been accumulated to date. None of the thermocouples have drifted more
than f2°F equivalent in emf output.

Bayonet Thermocouples. Ten MSRE-prototype wall-mounted thermocouples
were installed in the drain-tank bayonet-cooler test facility to determine
the effects of fast temperature transients on the life of these thermo-
couples and to determine how fast they would respond to transients. The
thermocouples tested consisted of both l/l6-in.-OD, sheathed, single-
conductor, two-wire and l/8-in.-OD, sheathed, duplex thermocouplecs with
Junctions grounded to walls and end seals. One l/8—in.—OD unit failed
immediately after startup of the test. Its early failure is thought to
have been due to faulty construction, since the other nine units have
been subjected to rapid temperature changes between 1200 and 600°F for
over a week without failing. The thermocouples that endure the thermal-
cycling test will be checked for speed of response.

Freeze-Valve Thermocouple Tests. Six MSRE~prototype wall-mounted
thermocouples were installed on a freeze valve in a test conducted to
determine the durability of these units in this type of service and to
determine the accuracy of the measurement of wall temperature at the
cooled and heated area of the valve. Sheathed, l/l6-in.-OD, single=-
conductor and l/8-in.—OD duplex thermocouples were mounted in pairs adja-
cent to a 30-ga bare-wire reference thermocouple at three locations on
the valve. The locations and methods of attachment are shown in
I'ig. 3.24. The test thermocouples agreed with the reference thermocouples,

Table 3.5. Variations in Readings of Thermocouples
Installed on the Engineering Test Loop

Outside Diameter Limites of Varia-

Type of
Re%grezce Station of Test tion ig Tempersa.-
Thermocouple Ther@ocouple ture Differcnce
(1in.) (°F)
No. 30 AWG bare wire 1 1/8 +1 to +6
1/16 £l to +
E 1/8 -2 t6 +4
1/16 -2 tao 0
- 1/8 -3 to +5
1/16 =2 to +5
1/8-in.-0D, sheathed, 4 1/8 =5 to =6

in well 1/16 =3 to +6

Fig. 3.24.

Thermocouple Installation on the MSRE Freeze Valve.

UNCLASSIFIED
PHOTO 38800

8
82

except that during rapid heating and cooling, differences of 15 to 20°F in
temperature readings were noted. All the thermocouples were still func-
tioning after two weeks of intermittent operation of the valve. No sig-
nificant difference in the performance or durability of the l/l6-in.
single-conductor and l/8-in. duplex thermocouples has been noted.

Thermocouple Disconnects. A six-circuit radiation-resistant thermo-
couple-disconnect assembly that was fabricated from commercially available
components is shown in Fig. 3.25. This disconnect is simpler in con=-
struction than the assembly previously described®® and is more compatible
with MSRE requirements. The plug and jack panels were supplied by the
Thermo Electric Company and are similar to their standard plug and jack
panel assemblies, except for the insulating material, which is Electro-
bestos. The housings are modified FS-type conduit boxes. The jack
housing (top half) has a removable back plate to facilitate wiring connec-
tions. Swaged-type tube fittings are utilized to support and restrain the
individual metal-sheathed thermocouples. A disconnect of this type was
tested 1n the remote-maintenance test facility and operated satisfactorily.
Alignment of the pins during remote maintenance operations will be
accomplished by means of a guide incorporated in the handling tool.

28'"MSRP Semiann. Prog. Rep. Feb. 28, 1961, " ORNL-3122, pp 61-62.
neect.

Fig. 3.25.

83

UNCLASSIFIED
PHOTO 38838

g

o ol
(LT
\
A

Six~Circuit Radiation-Resistant Thermocouple Discon-
PAGES 85 to 86
WERE INTENTIONALLY
LEFT BLANK

87

4, METALLURGY

Corrosion Effects of CF4

INOR-8 specimens, as reported previously,l sustained only superfi-
cial attack by CF4, vapor in 500-hr tests at 1112 to 14T72°F. The presence
of both oxide and fluoride reaction products were detected, however, on
specimen surfaces examined under the electron microscope and diffraction
camera. In order to determine whether these films had effected passiva-
tion of the INOR-8 surfaces, two additional experiments were run with
specimens immersed in a fuel salt of the composition LiF-BeFs-ZrF,-ThF,-
UFs (70-23-5-1-1 mole %) and with CF4 vapor passed continuously through
or over the fluoride salt.

The two experiments were performed in 2-in.-diam isothermal INOR-8
pots at temperatures of 1112 and 1292°F, respectively. Although both
were scheduled for 1000 hr of operation, depletion of the CF4 supply at
the higher temperature required termination of this experiment after TO0O
hr. Specimens being tested at 1112°F were removed from the vessel at
100-hr intervals. The specimens exposed to the salt underwent negligible
weight changes and showed no consistent pattern of weight loss or gain;
however, specimens tested in the vapor phase exhibited a slight weight
increase, in accordance with previous test results. Specimens immersed
in salt at 1292°F gained substantial weight (1 to 3 mg/cm?), and in some
cases a metallic-appearing coating was noted. Analyses of the coating
showed it to consist principally of nickel, with relatively large amounts
(>5%) of molybdenum. Specimens in the vapor phase showed small, although
consistent, weight losses and were covered with a greenish reaction
product.

Metallographic examinations of the specimens tested at 1112°F
revealed no evidence of surface changes. Specimens tested at 1292°T
were lightly pitted in areas not covered by the metallic coating, as
shown in Fig. 4.l. Areas covered by the coating were unattacked. Some
evidence of carburization could be detected metallographically; however,
chemical analyses of the specimens after the test showed no changes in
carbon content compared with that of as-received samples.

The results of these and earlier experiments indicate that CF4
vapor is effectively nonreactive toward INOR-8 at 1112°F but that minor
attack may be promoted by CF4 in the presence of fluoride salt at 1292°F.
At least part of the attack at 1292°F appears to be attributable, however,
to hydrolysis of the CF4 by traces of moisture in the system. This con-
clusion is based on the detection of small amounts of CO and COs in the
gas leaving the test vessel.

1"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, pp TT7-T9.
88

UNCLASSIFIED
Y-46810

Fig. 4.1. Corrosion Effects of CF,; on INOR-8 at 1292°F.

Welding and Brazing Studies

Heat Exchanger Fabrication

Tube-Joint Design and Fabrication. A final design and a welding
procedure were established for the tube-to-tube sheet joints of the
MSRE primary heat exchanger that incorporated modifications of the
previous design® to provide tf'or inspection of the back-brazed region.

The joint, shown in Fig. 4.2, was changed by flaring the weld
end sufficiently so that weld rollover would not interfere with the
insertion of an ultrasonic inspection probe into the tube. The welding
conditions established to assure a minimum weld penetration of 42 mils
are listed below:

2Tbid., pp 81-8k4.
89

UNCLASSIFIED
ORNL-LR-DWG 65682R2

TUBE

TREPAN GROOVE WITH
/BRAZING ALLOY RING

WELD SIDE/ \TREPAN

(a) BEFORE BRAZING

77
////////////
7

(L) AFTER BRAZING

Fig. 4.2. Tube-to-Tube Sheet Joint Design for MSRE Heat Exchanger.

Welding current 39 to 40 amp

Welding speed 4o TOM

Electrode diameter 3/32 in.

Electrode-to-work distance 0.035 in.

Electrode position 0.005 in. outside Jjoint interface
Inert gas Argon

A test assembly, shown in Figs. 4.3 and 4.4, was fabricated to
demonstrate suitable fabrication procedures and welding techniques
and to study the effect ot size on the back-brazing operation. The
assembly contained 96 welded and back-brazed joints. Welding was
completed without difficulty, and good flow of braze metal, indicated
by filleting, occurred in all but one of the joinls. Metallographic
investigation of this brazed joint revealed that contamination in the
vieinity of the feeder holes had interfered with flow of the alloy, as
may be seen in Fig. 4.5. To avoid a similar situation in the MSRE
heat exchanger, precautionary inspection measures have been included
in the fabrication procedure.

Tube-Joint Inspection. Methods of inspecting tube joints were
developed that are applicable to the MSRE heat exchanger. These include
radiography of the weldments and an ultrasonic Lamb-wave technique for
inspecting the brazed joints.

Test welds were examined with both x-ray and iridium sources under
a variety of conditions and with curved strips of film placed in the
20

UNCLASSIFIED
PHOTO 58284

Fig. 4.3. Weld Side of Test Assembly with MSRE Heat Exchanger Type
of Joint.

tubes at the welds. Of those tested, the optimum radiographic conditions
found for the heat exchanger weld were:

X-ray energy 160 kv
Film to focal distance 48 in,
Tncident angle 3035 deg
Time 2.5 min
Film type M

Four to six exposures were required to completely inspect each weld;
however, the tube arrangement on the MSRE heat exchanger allows for
several tube welds to be radiographed at one time, and the operation
will be more economical than for the test welds. The defects observed
were somewhat distorted because of the method of film placement, but,
despite this, pores several mils in diameter were detected.

The ultrasonic Lamb-wave technique developed for the evaluation of
bonding in brazed tube Jjoints utilizes a small probe that fits inside
the l/2—in.-diam tube. The probe contains two piezoelectric crystals
a9l

UNCLASSIFIED
PHOTO 58287

2

N 3
RATORY

QOAK RIDGE NATIONAL LABO
U

Fig. 4.4. Braze Side of Test Assermbly with MSRE Heat Exchanger Type
of Joint.

that serve as ultrasonic generator and receiver, respectively. The
proper ultrasonic frequency and beam incident angles were determined,
and a custom probe was designed and fabricated. Approximately 20 brazed
joints with various degrees of bonding were evaluated using this system,
and metallographic examinations are in progress to provide correlation
of the test results.

A mechanical device was designed that provides for uniform scanning
of the probe in the tube and thus assures complete inspection coverage.
This device should permit more rapid inspection and may also allow
antomatic recording ol uubonded arcas.
92

UNCLASSIFIED |
Y-46952

TUBE WALL

Fig. 4.5. Defective Braze Caused by Metal Chips in Feeder Holes.

Remote Brazing

The brazing procedure and equipment described previously® for
remotely fabricating INOR-8 pipe were used in the preparation of sgix
prototype joints. Evaluation of these by ultrasonic techniques developed®
previously indicated only minor nonbonding in five of the brazed areas.
The sixth joint contained a number of unbonded areas in the upper portion
of the braze. The bonding in a typical remotely brazed joint is described
in Fig., 4.6.

Welding of INOR-8

A mechanical properties investigation is being made of INOR-8 weld
material from heats to be used in the MSRE. As part of thc preliminary
studies, data were obtained on the effect of stress-relieving on stress-
rupture properties. These initial tests were made using transverse weld

SIbid., pp 84-86.

“K. V. Coock and R. W. McClung, "Development of Ultrasonic
Techniques for the Evaluation of Brazed Joints," ORNL-TM-356 (in
press) .
93

UNCLASSIFIED
/OUTER SURFACE ORNL-LR-DWG 69026R

NOMINAL LENGTH
END EFFECT OF JOINT (1in) . END EFFECT
FROMFILLET .77 ), 7] I FRoM FILLET
1 o
= [ . - :
| - ST
! |
| } L I
I
I I
| |
1 | - -
| . |
: i I
fin— |
|
| |
Tt || BONDED AREA —
e
= = . || UNBONDED AREA =
- o
T e - |
e =i
- |
|
y |
2in. _’ }
|
|
I
|
|
|
|
|
|
I
|
3in — i
1
r |
|
|
|
= 2 I
=) -e |
= |
5 ¢
3} |
= {
= I
< 4in. |
= . i
4
w |
i |
]
e |
= I
3 I
[}
© I
S I
< l
S |
. I
/ |
Sin. |
* |
I
|
|
i
|
|
|
S I
I
[
i N !
| ’ l
6in.—D| :
' - 1
i .
j = !
I
' |
|
S g I
| I
i s
4 | . |
{ |
1
7in.—| I |
I
! s i
2 i
! I
! |
| - |

Fig. 4.6. Results of Ultrasonic Inspection of a Remotely Brazed Pipe
Joint.
9%

specimens cut from a weld-containing base plate (heat No. N1-5090) and
weld wire. Two samples were tested at 1300°F in the as-welded condition
and two in the stress-relieved condition (2 hr at 1600°F). The data
obtained are presented in Table 4,1, It is evident that the stress-
relieving treatment improved the over-all properties of the weldment,
probably because of a redistribution of microconstituents, such as
grain-boundary carbides.

Table 4.1. Elevated-Temperature Stress-Rupture Test
Data on Transverse INOR-8 Weld Specimens in the
As-Welded and Stress-Relieved Conditions

Time to Rupture Total bStrain

Sample Condition (hr) (%)
As welded 113,9 %445
79.1 6.6
Stress relieved 234.,2 125
233,9 1269

Tests were also performed on welds made with the short-arc semi-
automatic welding process. Tensile tests at 1300°F provided the data
presented in Table 4.2. The low ductilities were probably due to
defects in the weld. It is expected that improvements in the process
can produce welds having properties comparable with those of welds
produced by the manual, tungsten-arc process.

Table 4.2. Tensile Tests at 1300°T of Transverse INOR-8
Short-Arc Weld Specimens in the As-Welded Condition

Sample Tenesile Strength Yield Strength Elongation
No. (psi) (psi) (%)
1 58,300 47,500 Ted
2 60, 800 4'7, 500 965
(2) 67,300 47,200 15,5

aTypica] 1300°F tensile data for manual, tungsten-arc,
INOR-8 welds.
95

Mechanical Properties of INOR-8

Properties of Reactor-Quality INOR-8

The mechanical properties of typical INOR-8 material procured for
MSRE construction are being determined in order to evaluate the effects
of large-gquantity production and improved-quality requirements. Speci-
mens for tensile and stress-rupture testing were prepared from
representative INOR-8 plate, including heats N1-5055, N1-5075, and
N1-5081. Tensile properties will be tested from room temperature to
1800°F, and stress-rupture properties will be determined in air at
1100, 1300, and 1500°F.

Thermal Fatigue of INOR-8

The thermal-fatigue properties of TNOR-8 are being determined
with methods described by Carden.> Fifteen tests were completed at
maximum test temperatures in the range 1250 to 1600°F and cycle periods
of 1 and 10 min. Results of these tests are presented in Fig. 4.7, which

UNCLASSIFIED
ORNL-LR-DWG 74157

E
£ > .
= . INCONEL (13C0°F MEAN)
w
g .
l% 2 N \\
>~ ™
a 10-2 — \\ N
- ——TYPE 304 STAINLESS -
o ——  (1300°F MEAN) S ~
=z N S
E:I 5 """-%,_ \\
= .-'0"'--\. = \\\
g q’ T TR
E 2 | i ‘--._‘-_‘
? OPEN SYMBOLS # | min /cycle ~Pr- ~
[ J
Q 453 |- SOLID SYMBOLS 10 min/cycle AN |
2 — o 1250-1350°F T
Z 5[ a 1400-1450°F
- —— n 1500 =1550°F | MAX
4 —— { 1600°F
2
104 ' - -
0 2 5 10° 2 5 10° 2 5 1% 2 5 10

M, CYCLES TO FAILURE

Fig. 4.7. Thermal-Fatigue Data for INOR-8 (Heat NL-5050) in Com-
parison with Inconel 'and Type 304 Stainless Steel.

gives the plastic strain range (maximum internal width of stress-strain
loop) versus the number of cycles to failure for LNOR-8, Inconel, and

5A. E. Carden, "Thermal Fatigue — An Analysis of the Experimental
Method," ORNT.-TM-LOS (4in preparation).
96

type 304 stainless steel. The data on Inconel and stainless steel are
included for comparison.

The thermal-fatigue data of Fig. 4.7 indicate that Inconel and
type 304 stainless steel have greater ability to sustain plastic flow
than INOR-8. An analysis of the temperature difference versus cycles
to failure for comparable maximum temperatures and frequencies shows,
however, that INOR-8 will sustain more restrained thermal cycles and
that INOR-8 requires a greater temperature difference than type 30k
stainless steel or Inconel to produce failure in a comparable number-
of cycles. The plastic strain per cycle is less in INOR-8 because of
its lower thermal coefficient of expansion and higher elastic strength.

Considerable plastic flow was observed in INOR-8 during the hold
time in the 10-min per cyc¢le test. This phenvwmenovuy is being studied
for hold times of up to 24 hr.

Graphite Studies

Evaluation of Grade TS-28) Graphite

Specimens of grade TS-281 graphite were evaluated for molten salt
and mercury permeation in standard screening tests, and the specimens
met MSRE design requirements. This graphite was produced for the
Engineering Test Loop in the size required for the MSRE and with
essentially the same characteristics as those of the graphite for
the MSRE. '

Exposure of TS-281 test pileces to molten salt at the conditions
described in Table 4.3 resulted in an average salt permeation of 0.2%
of the bulk volume. The MSRE specification allows a maximum of 0.5%
permeation. Impregnation of similar specimens with mercury at the
conditions described in Table 4.3 resulted in an average weight gain
of 1.2%, as compared with a maximum allowable MSRE specification of

3.5%.
A comparison of the permeation of TS-281 graphite with that of

experimental grades of similar graphite (reported in Table L.lL)

Table 4.3, Test Conditions for Standard
Permeation Screening Tests

Test Molten-Salt® Mercury
Conditions Test Test
Temperature, °T 1300 70
Test period, hr 100 20
Pressure, psig 150 470

41iF-BeF, -ThF,-UF, (67-18.5-14-0.5
mole %),
97

Table 4,4, Salt Permeation for Grade
TS-281 and Similar Experimental
Grades of Graphite

. Size of Cross Bulk Volume
Géaphlte section Permeated
rade .
(in.) (%)
TS-281 2 1/4 x 2 1/4 0.2
CGR-X 4 (diameter) 0.02
CGB-Y 2 1/4 x 2 1/4 0.04%

indicates that full-scale or commercial fabrication imposes the
penalty of greater porosity.

In the testing of the TS-281 specimens, the graphite bars were
found to be slightly more porous and less dense toward their central
axis; however, these regions were well within the requirements specified
for MSRE graphite. Radiographic examinations are being made to determine:
Lhe distribution of salt impregnation in CGB-Y and TS5-281 graphite.

Effects of Thermal Decomposition Products of NH4F-HF on Graphite-
INOR-G Systems '

A determination was made of the effects of temperature on the
purging procedure® developed to remove oxygen from grephite by utilizing
the decomposition products of NH4F+HF. The effects on INOR-8 were
examined concurrently.

Grade R-0025 graphite and INOR-8 specimens were exposed isothermally
for 20 hr in Inconel containers at temperatures varying from 392 to
"1300°F., Both 0.2 and 1.0 g quantities of NH4F-HF were included in the
test system.

The following qualitative results were obtained from this study.

1. Both 0.2 and 1.0 g of NH4F.HF effectively removed oxygen con-
tamination from the graphite, except in one instance. The test made
with 1.0 g at 392°F was inconclusive and is being checked.

2. For purges with 0.2 or 1.0 g of NE,F+HF there was a minimum
reaction of the INOR-8 specimens at 752°F, as determined by weight gain.

%3, In general there was less reaction (weight gain) of INOR-8 at
all purging temperatures for the tests with 0.2 g of NH4F.HF than for
tests with 1.0 g.

S"MSRP Prog. Rep. April 30, 1960," ORNL-2973, p 59.
98

4, The reaction (weight gain) of INOR-8 varied in proportion
to the amount of graphite present and was least in the test with no
graphite. This increased attack in the presence of graphite was
probably caused by oxygen products available in the graphite.

Fabrication of Gds0s; and Gdo03—A1-05 Pellets

A study was begun to develop the fabrication procedure necessary
to make cylindrical shapes of Gd>03 and Gdz03—Alz03 mixtures. These
parts are of interest as control rod elements for the MSRE and are
being designed to withstand a nitrogen atmosphere contaminated with
air and moisture at temperatures up to 1400°F. Al-03 is added to
minimize hydrolysis and subsequent deterioration of the GdzOs.

A preliminary test was made of the sintering characteristics of
the materials under a single set of conditions. Pellets containing
0, 20, and 30 wt % Alo05 were prepared by dry blending, pressing, and
51nter1ng in hydrogen at 1750°F for 3/4 hr. All sintered pieces were
set on molybdenum sheet. Of the compositions tested, only Gdz0s gave
a sound pellet with good density. 'The density and shrinkage data are
listed in Table 4.5. The mixtures of (Gds0s and Al,05 showed evidence
of eutectic . melting or the formation of a low-melting-temperature
compound that indicated too high a sintering temperature. Additional
specimens of Gdp0s—Al505 mixtures with various green densities have
been prepared and will be fired at 1600, 1650, and 1700°C to determine
optimum sintering conditions.

Table 4.5, Characteristics of Gd,03-A1,03 Mixtures Sintered
in Hydrogen at 1750°C for 3/4 hr

. . . Density Shrinkage
Composition (% of theoretical) (%)
Gd203 93 23
80 wb % Gda02—20 wt % Al20a 60 13
70 wt % Gd03—30 Wt % Al,0 Melted

Pellets of Gds05 with green densities ranging from 2.9 to 3.8 g/cm3
were tested for determining the effects of green density on sintering
characteristics. Shrinkage and bulk density data for these pellets
are presented in Figs. 4.8 and 4.9. Using this information, thin wall
cylinders having the dimensions required for MSRE control rods will be
tabricated for proof testing.
UNCLASSIFIED
CRNL-LR-DWG 74158

N
25 >,
DIAMETER
~ 24
B | \\
uj
T LENGTH - UNCLASSIFIED
g 1N N “ ORNL-LR-DWG 74159
z 22 €76
T \Q e
-}) \ >
N\ =
g 20 [z e
z 8
- i / \
o //
w
44
et
16 ' Z 7.0
2.8 3.0 3.2 3.4 3.6 3.8 4.0 ® 2.8 3.0 32 3.4 3.6 3.8 4.0
GREEN BULK DENSITY (g/cm3) GREEN BULK DENSITY (g/cm3)
FPig. 4.8. Sintering Shrinkage of Fig. 4.9. Bulk Density Change of
Gd,04 Soaked at 1750°C in Hydrogen for ‘ Gd,03 Soaked as 1750°C in Hydrogen for

3/4 hr. 3/4 hr.

66
100

5. IN-PILE TESTS

ORNL-MTR-47 Molten-Salt Irradiation Experiments

Experiment ORNL-MTR-47-L

The carbon tetrafluoride found in the cover gas over molten-salt fuel
irradiated in graphite crucibles in assembly ORNL-MTR-47~3 is considered
to have formed at the salt-graphite interfaces and to have passed through
the graphite to the cover gas.l It is postulated that if the CF) had been
forced to pass through molten fuel, it would have been consumed. Two ex-
perimental assemblies, ORNL-MIR-47-4 and 47-5, were designed to test this
hypothesis and to study the formation of CF).

Six cylindrical INOR-8 capsules,g four containing a central graphitc
core submerged in fuel salt and two containing graphite crucibles, were
installed in assembly ORNL-MTR-47-L. The capsules were immersed in molten
sodium, which served as a heat transfer medium. The primary purpose of
the experiment was to determine whether CF), would exist in the cover gas
over MSRE fuel irradiated in contact with completely submerged graphite.
Because the fuel could not be maintained molten during reactor scrams and
shutdowns, the graphite surface was intermittently in direct contact with
the cover gas through cracks in the solid fuel. The crucible capsules,
which contained exposed graphite in contact with fuel, provided a basis
for comparison with the capsules of the 47-3 experiment and an evalualion
of the effectiveness of graphite submersion in preventing the net gener-
ation of CF). Also, experiment -4 provided for @ [further demonstration
of the compatibility of the fuel—graphite—INOR-8 system under thermal
conditions at least as severe as those expected to exist during MSRE
-pperation.

The temperature history of the 47-4 capsule during the irradiation
period of March 15 to June b is summarized in Table 5.1. The four sub-
merged-graphite capsules contained central Chromel-Alumel thermocouples.
The internal temperature distribution (Fig. 5.1), obtained by use of the
computer program TOSS,- indicates that the thermocouple should give, essen-
tially, the graphite-to-salt interface temperature. The temperature history
of the uninstrumented crucible capsules was calculated based on temperature
measurements in the surrounding liquid sodium. The muxlmum.mcasured tem-
perature among the four instrumented capsules was 1400 T 50 °F. The mean
measured temperatures of the other three instrumented capsules were approx-
imately 1380, 1370, and 1310”F, respectively:. The corrccponding caloniated
INOR-8 capsule wall-to-salt interface temperatures were 1130, 1125, 1115,

1"MSRP Prog. Rep. Feb. 28, 1962," ORNL-3282, pp. 97-110.
2Tpid., pp. 110-112. |

SD. Bagwell, "TOSS - An IBM 7090 Code for Computing Transient or
Steady State Temperature Distributions," K-1494, December 1, 1961.
101

Table 5.1.
ORNL-MIR-47-4 Capsules

Temperature History of Fuel Salt in

Time at Temperature (hr)

Temperat?ES)Interval Submerged Graphite

Core
(24, 4A, 54, 6, 6A)P

Graphite@d

Crucible

(4)

Steady-state operation

0-100 390.6
100~700 44,0
700750 1203.7
750800 254,0
800850
850—200

Total 1892.3

Nonsteady-state operation 51.7
Total irradiation 1553.4

390.6
12.1
3.8
17.6
11.4
1456.9

1892.3
51.7
1553. 4

SCalculated temperature history based on temperatures meas-

ured in submerged graphite capsules.

bCapsule identification number.

UNCLASSIFIED
ORNL-LR-DWG 67930

1600 v
1c | P' . '
: /IGRA TI E CORE
1
1400 L,, FUEL SALT
=N |
; "N\ | L~ CAPSLLE waLL
\\ W A
1200 STV
N
AT N S S
A -« #
_ 'R 2~ 'SoDiUM
w =4 —_
¢ tooo : il“‘.__—.——--—-j“*‘SUUIUM TANK WAI I,
w
g
 B0O [— 2 wotts/g GAMMA ENERGY S I
& 74 waits /cc FISSION (FUEL}
o = 38,300 Btu/hr FROM -——VARIABLE GAS GAP
=z SODIUM TANK
= 800

———— VERTICAL TEMPERATURE
DISTRIBUTION ENVELOPE

TEMPERATURE GRAPHITE
400 - MIDPLANE

& EXPERIMENTAL DATA

200 “WATER JACKET
L<\‘ COOLANT H»0 105°F

O ‘A'.
o 0.2 0.9 0.6 Q 0.2 0.4 0.6 0.8
DISTANCE FROM CAPSULE DISTANCE FROM INNER TANK

CENTER LINE (in.} SURFACE (in.)

Fig. 5.1. Radial Temperature Distribution of MSRE

Capsules.
102

and 1110°F. The tcmperature of the surface of the INOR-8 end cap in con-
tact with the fuel salt was calculated to be the same as that of the
graphite-to-salt interface temperature. The area of the end cap wetted by
the salt was about 12% of the total 5.6 in.2 of metal surface in contact
with the salt.

The accumulated time during transient operation includes only times
for temperature changes of approximately 30°C or more. Of the 121 such
temperature changes recorded, 60 included decreases to the solidus tem-
perature of the fuel salt.

The assembly was removed from the MIR, partially disassembled, and
returned to ORNL, where it was completely dismantled. Postirradiation
examination of the capsules is in progress. As an aid to interpretation
of irradiation effects, an out-of-pile experiment was ¥iih that duplicated
the temperature history, including the transients.

Experiment ORNL-MTR~-47-5

The 47-5 experimental assembly was designed and constructed to permit
gas sampling from two capsules during irradiation to study the CF) content,
if any, of the offgas from fissioning MSRE fuel in contact with submerged
graphite. These capsules (Fig. 5.2) are 1 in. in diameter and 2.25 in.
long, and they contain a central CGB graphite core 1/2 in. in diameter and
1 in. long submerged approximately 0.3 in. in about 25 g of fuel. The two
capsules differ only in the uranium concentration in the fuel.

Additional capsules included in the assembly are designed to provide
a range of oxidation-reduction levels that includes what might be both
early and late stages in the life history of a reactor. This is accom-
plished by altering the accessibility of chromium from the container alloy
(INOR~8) to the fuel. These capsules are externally similar to the purged
capsules, but the graphite corés are markedly different. A high metal-to-
graphite wetted-surface-area ratio (46:1) is obtained (Fig. 5.3) by using
a central core of INOR-8, except for a 0.l-in.-thick CGB graphite wafer.
A low metal.-to-graphite surface area ratio (1:4) is obtained (Fig. 5.4)
by using a close-fitting graphite crucible 0.7 in. in internal diameter
and 1.5 in. long, with a central graphite core that is 0.34 in. in diameter
and 1.25 in. long.

An extreme case in the oxidization study is a capsule with no metal,
that is, with graphite prepermeated with fuel so that the only fuel in the
capsule is contained within the pores of an isclated graphite gere. Two
small capsules of this type (Fig. 5.5) are included and will be operated
at different power densities. These are 5/8 in. in diameter by 2.25 in.
long and contain 0.32-in.-diameter by 1 3/h-in.-long centrally suspended
AGOT graphite cores. Irradiation is scheduled to begin the week of
September 17, 1962.
UNCLASSIFIED
ORNL-LR—DWG 71570R

INOR-8 :NOR-8
COVER GAS CCVER GAS
PURGE LH\E\ FURGE LIME~

SN '|
N
INOR-8 CAP —., Yy N v
Y RTINS R A A R
\ b
1
A !
PUNCTURE AREA N\ ' ) ’*—fl—HE‘-'U(“g Eg;’)ER GAS
GAS SAMPLING — \ |
b= |
A=A
INOR-8 2= s
INTERFACE AREA™ k- IS80: -1 M OLTEN-SALT FUEL
3sem?y (LA 7] (25 ql
A~ I
T 1A
INOR-8 CaN—"" [} -~ Rl
- o1
. e
% R
CHROMEL-A_UMEL [/}~ = S CGB GRAPHITE
THERMCCCUPLE —{£==7 S f//_—INTERFACE AREA
/1 LR (125 cm?)
& P

NOR—8 CENTERING PIN

N CKEL POSITIONING LUG

Fig. 5.2. Svbmerged and Purged Graphite—

Molten-Salt Capsule.

&

UNCLASSIFIED
ORNL-LR-DW{ 71568R2

NICKEL POSITIONING LUGS (2)\ f —— INOR~8 THERMOCOQUPLE WELL

/— INOR-8 CAP

N
b\,

/

»
HELIUM COVER GAS —
(3 cm”)

77L& 7

S~ INOR—8 CAN

PUNCTURE AREA

Np SN
bl
FOR GAS SAMPLING” /N /N
\]|—— MOLTEN-SALT FUEL
INOR-8 N R (19 )
ANTERFACE AREA N N
(7 cm") Ny
&Y e e Y
“\ ... fl I §
HELIUM GAS Gap—" NEST ™ EXH <N CHROMEL-ALUMEL
NS . TN THERMOCOUPLE
: TR
GRAPHITE 2
INTERFACE AREA
(27 cm?)
o) ¥ 1 @)~ NICKEL POSITIONING LUG
— e ——— N
INCH O

Fig. 5.3. BSubmerged Graphite-Crucible—
Molten-Salt Capsule.

€0t
UNCLASSIFIED
ORNL-LR-DWG 71867R

NICKEL POSITIONING LUGS (=) _—— INOR-8 THERMOZOUPLE WELL . ) UNCLASSIFIED

ORNL-LR—-DWG 71569
NICKEL POSITIONING LUG—.__
INOR-8 CAP rJ HELIUM COVER GA3
\\ N {3 cm3)
y \V // 7 .fi
b \/ , 2 N -
PUNCTURE AREA /7l 3 ' .i B —FOR ‘GAS SAMPLING
- i - o] g
FOR GAS SAMFLING—=t|  [4 [ \//—INOR 8 CAN . HELIUM COVER GAS— "1 §% \
M (6 em®) J K / \
= \ 3 B
= . = 3] RRASZLN A
=X Y B N
MOLTEN-SALT FUEL—— 17— : > N RIS N
25 g} 7/ .SE INOR-8 5] [RXRIG Y
- — P INTERFACE APEA . R N
- —- 5eme N k) N
\E=7/ 48 em®) < R N
- n [ KK KK K]
' : \_—/é: AGOT GRAPHITE Y REEED N
CGB GRAPHITE N TR IMPREGNATED WITH -~} INOR—8 CAN
INTERFACE AREA—— K= ; MOLTEN-SALT FUEL N \ /_
2 N-- -2 7 -5 (z g !
{1 em?) N[/ WA —p—"NOR-8 cCRE 2 6 i N
pe— 7
Yy, \
N - - : \
CHROMEL-ALUMEL N
THERMOCOUPLE ———?F
FUEL RESERVOIR

0 Vs 1 c %
INCH "INCH NICKEL POSITIONING LUG
Fig. 5.4. Submergei Graphite-Wafer-dMolien- Fig. 5.5. Impregnated-Graphite-Molten-Salt

Salt Capsule. Capsule.

701
105

Postirradiation Examination of Experimental
Assembly. ORNL-MTR-47-3

Examinations of irradiated specimens from assembly ORNL-MTR-47-3, the
assembly in which CF) was first detected,? provided some additional informa-
tion on the corrosion products and on the distribution of radiocactivity and
salt in the graphite. The possibility that much of the CF) could have been
formed from the fluorine released from the frozen salt after shutdown, as
found in test 47-4, placed new light on the interpretation of the results
of test 47-3. The fuel mixture irradiated in this experiment was initially
LiF-BeFy~ZrF) -ThF) -UF), (69.5-23-5-1-1.5 mole %).

Chemical Analyses of Irradiated Fuel for Corrosion Products

Corrosion~product analyses were carried out on an aliquot of the solu-
tion produced during dissolution of the irradiated fuel to obtain gas
samples for reducing-power analyses. The aliquot was evaporated to dryness
three times after additions of methanol to remove the excess boric acid
used to improve the initial dissolution. The dry residue was dissolved in
10 to 15 ml of 9 M HNO, with 2 ml of 3 M hydroxylamine to reduce any hexa-
valent chromium to the trivalent state, and the thorium was removed from
the solution on an anion-exchange column. The solution was then diluted to
5 N HNO,, and the uranium was removed by extraction into 20% TBP in hexane
or Amscd. The iron, chromium, nickel, and molybdenum corrosion products
were determined by the intensity of their spectral lines compared with an
added internal standard. The readings were made with a densitometer on
the photographic plate.

The analyses for corrosion products, as shown in Table 5.2, gave
values somewhat different from those previously reported.® The corrosion-
product concentrations that indicate attack on INOR-8 coupons have been
revised to the order of 100 ppm, rather than several hundred or thousand
ppm, and therefore the data of Table 5.1 for iron, chromium, and nickel
indicate mildly reducing conditions. They present some puzzling aspects,
however, in that differences are not discernible between capsules that pro-
duced widely different amounts of CFj. This can be explained if the larger
amourls of CF), arose from attack of the graphite by fluorine released after
freezing of the fuel.

Another point of interest is the absence of molybdenum (below the
50 ppm limit of detection) in the fuel in spite of obvious attack on some
of the molybdenum coupons. The molybdenum may have volatilized as MoFg,
but its presence in a reducing melt would be difficult to reconcile. It
would be expected, however, in an atmosphere that would produce XeF) or
CFy. DNot finding the molybdenum in the fuel may indicate that it deposited
on the wall of the capsule.

4"MSRP Quar. Prog. Rep. Feb. 28, 1962," ORNL-3282, p. 97.
STbid., p. 110.
106

Table 5.2. Results of Corrosion-Product Analyses of Fuel
Irradiated in Assembly ORNL-MTR-47-3

Corrosion Products (ppm)

Capsule Sample
Cr Fe Mn Mo Ni
3 1 60 310 <02 <40 70
ob 790 230 36 <50 1100
8 1 20 <130 <10 <60 50
2 120 <210 <20 <120 100
15 1 40 220 <20 <80 <230
2 £0 230 20 <100 <310

Values preceded by the less than symbol indicate the
lowest detcctable limit.

bThe unusually high values Tor both chromium and nickel,
as well as the presence of manganese, indicate the possibility
that the sample was contaminated.

The compound XeFu, which was recently discovered at the Argonne
National Laboratory,® is being studied. As mentioned later in connection
with experiment L7-k, there is a close correlation between fluorine or
evidence of fluorination in the form of CF), and a pronounced deficiency of
xenon in the irradiated cover gas. This correlation and an cbserved
increase in the xenon-to-krypton ratio in samples collected over a long
period of time are circumstantial evidence that XeF), was formed in some
capsules. Conversely, the presence of Xenon in other capsules is a possible
index of the absence of fluorine, at least during the time favorable for
reaction with xenon, and this must overlap part of the range in which CF)
is formed. The reaction between xenon and fluorine proceeds at 400°¢.

Distribution of Radioactivity and Salt in Irradiated Graphite

Data on the distribution of radiocactivity and sait were obtained from
cored samples of graphite parts. Patterns of radiocactivity were determined
by gamma spectrometry, and the sall distribution was obtained from
petrographic studies with an optical microscope. The scatter of the data
preclude immediate interpretation, and the data will therefore be held for
examination along with data from future similar experiments.

Postirradiation Examination of
Experimental Assembly ORNL-MTR-47-4

The hypothesis that CF), should react with the melt rather than accu-
mulating in the cover gas in irradiated capsules containing molten fluoride

®H. H. Claasen, H. Selig, and John G. Malm, letter to editor of
J. Am. Chem. Soc. dated Aug. 17, 1962, in press.

107

fuel and submerged graphite was tested further with the irradiation of four
capsules, described previously,? at a flux of 1013 neutrons/cmé-sec for
more than 1500 hr in the MTR. The data sought in the test, namely the
partial pressure of CF) at operating conditions, were obscured in two of
the capsules opened thus far by large amounts of radiolytic fluorine from
the frozen fuel. Of the fluorine content in the gas phase, which was 1%

of the fluoride in the fuel, some 80% was free fluorine and the remaining
20% was CFy, most of which probably was a secondary reaction product
arising from attack on the graphite by the fluorine released by radiolysis
after shutdown.

Two smaller capsules containing molten salt in graphite crucibles
were irradiated in the same assembly (ORNL-MTR-47-4) and received similar
neutron dosages and heating cycles. These capsules were also described
previously.? These two capsules were included in the assembly to provide
conditions favorable for the accumulation of a measurable steady-state con-
centration of CF)y at two power densities; one capsule contained twice as
much UF), as the other capsule. The one capsule opened thus far operated
at a slightly lower power density than the other because of its location.
No fluorine and only the modest amount of CF) for this type of configura-
tion were found.

Iluorine atoms undoubledly are formed in fissioning fuel at an
apprecilable rate as a result of radiolysis. The rate of formation is
probably proportional to the power density. The lack of corrosion of
metal or graphite indicates that rapid reverse reactions in the molten
fuel consume the fluorine atoms and lower the steady-state concentration
of elemental fluorine to negligible levels. All other chemical components
in the operating fuel system, including moderator and container, remain at
virtual thermodynamic equilibrium as long as the fissioning fuel is molten.

As shown by some of the results described below, restricted mobility
in the frozen crystalsuof fuel, along with the pronounced drop in tempera-
tures that prevails in irradiation test capsules after reactor shutdown,
may, on occasion, lead to off-equilibrium conditions in the form of ex-
cessive amounts of unreacted elemental fluorine. L1n such cases, the
reverse reaction mechanisms that consume fluorine are even more drastically
curtailed on shutdown than is the rate of energy release that gives rise
to fluorine atoms. The resulting buildup of fluorine molecules in the
frozen capsule contrasts markediy with the situation at operating conditions.

Also the marked contrast between various frozen samples implies that
small differences in geometry or freezing history may be of controlling
importance. Not all freezing liberates fluorine. Further there does not
seem to have been a significant amount of fluorine formed in the 30
freezings that occurred during the in-pile exposure, probably because of
an insufficient time lapse in some cases. Also there were capsules in
which no fluorine was found even after the final lengthy cooling period.
The gquestion of why some capsules did and some did not contain fluorine
can be answered only conjecturally on the basis of current information.
108

Degcription of Capsules

The capsules were constructed of l/l6-in.-wall INOR-8 tubing. The
larger type contained 25 g of fuel of the nominal composition
LiF-BeFy-ZrF) ~ThF), -UF), (70-23.3-5-1-0.7 mole %), which gave the analysis
shown in Table 5.3; that is, 71.0-22.6-4.7-1.0-0.7 mole %. The capsules
were filled by liquid transfer to a level controlled by blowing excess fuel
back through the dip leg associated with the vent line.* Somewhat less
than the design volume of 3.5 em3 of vapor space at temperature probably
remained in the capsule, perhaps because of capillary effects. Because of
contraction of the fuel, principally on freezing, there was 2.5 cm’ more
vapor space at room temperature than at the operating temperature. The
room~-temperature volumes, as calibrated with known volumes after puncture
and gas collection, were about 4.5 cm3. Initially two concentrations of
UF), were to be used in pairs of the large capsules, but breakage of a
transfer line during filling, attributed to embrittlement by sulfur Lrom an
unknown source, eliminated the possibility of meeting schedules with the
more concentrated fuel. :

Table 5.3. Chemical Analysis of Fuel in Large Capsules in
‘ Assembly ORNL-MIR-47-4

Nominal . . - Corrosion

Uranium Tonic Constituents (wt %) Products (ppm)

Content _ - " - —

(mole %) U4t Li” 7zr4t  Be Th* o Ni Cr  Fe
0.7 3.89 11.6 10.1 4.8 5.8 63.6 25 24 o7
L.47 7.5 10.5 10.4 s 5 5.2

61.5 15 40 123

B A NI KTV

The smaller capsules were loaded with molded ingote of about 10 g of
fuel contained in graphite crucibles or liners, and in one of them the UF)
concentration was increased to 1.4 mole % to provide a power density of
about 160 w/cm3. The cover-gas space had a volume of about 2 cmS. There
were no thermocouples in these capsules, and the fuel should not have been
in contact with any metal, although traces of solvent probably distilled
to the interface between the graphite liner and the INOR-8 capsule wall.

Fabrication and Loading of Capsules

The metal capsulc partc were cleaned by firing in hydrogen gas at
lOOOOC, but a vacuum of questionable quality was inadvertently imposed
while the parts were still at temperature. The resulting contamination by
air caused a greenish oxide film on the INOR-8 parts, but time for another
treatment had elapsed before the error was noted. The film was partially
removed with steel wool prior to assembling and welding the parts in a
helium-filled glove box. The graphite core in capsule 24 was cleaned by
109

heating to 2000°C in an evacuated system. The graphite parts in the other
capsules were cleaned by flushing with a molten fluoride salt.

The larger capsules were then connected with a filling system and
heated. The molten fuel salt was run into each capsule and the excess
blown back with helium pressure. The fuel salt level was monitored with a
television x-ray system. Final closure was made in a glove box where the
external ends of the fill tubes were crimped and welded shut.

Fuel

The results of ionic analyses of the bulk constituents and analyses of
the corrosion products of the two fuels prepared for use in experiment
ORNL-MTR-47-4 are presented above in Table 5.3. The fuel salt was pre-
trecated at 750 to 800°C by sparging for 2 hr with hydrogen gas and then for
8 hr with a 5:1 hydrogen—hydrogen fluoride gas mixture, followed with a
48-hr sparge with hydrogen gas.

Graghite

The CGB graphite used in experiment 47-4 had a surface area of
0.71 mz/g as determined by the BET method. The other properties of this
graphite are listcd below:

Permeability of a 1.5-in.-0D, 6.56 X 10™% cm?® of He (STP)
0.5-in.-ID, 1.5-in.-long specimen per second
Density (Beckman air pycuometer) 2.00 g/cm’
Bulk density 1.838 g/cm’
Bulk volume accessible to air 8.7%
Total void volume as percentage of 19.5%
bulk volume

A semiquantitative spectrographic analysis of the graphite, as shown
in Table 5.4, revealed only the usual low leveles of trace elements, some
of which may have been introduced to the specimens during the various

Table 5.4. Spectrographic Detection of Trace Elements in
Graphite Used in Experimental Assembly ORNL-MI'R-47-4

Element Level Detected Element Level Detected
Detected (ppm) Detected (ppm)

Al 10—100 Fe 10-100

Be 1—-10 Mg 1-10

Ca 1-10 Ni 10-100

Cr 1-10 Pt 10-100

Cu 10100 Si 1—10

110

machining and handling operations. The elements analyzed for but not de-
tected are listed below:

Ag Hge  Sr
Mn

As Ta
Am Mo Te
B Na Ti
Ba - P T1
Bi Po \
Ca Pad Zn
Co Ru - Sr
Ga Sb Cb
Ge Sn

Decay Energy

The [fuel salt froze within 5 min after the reactor shutdown that
terminated the irradiation of assembly Y-k, caleuwlations ol Ue decoy
energy in the large capsules indicated that during the first 1l-hr period
following the initial 5-min freezing period about 10 w-hr of energy was
dissipated, whereas during normal irradiation about 840 w was dissipated.

Irradiation and Temperature Data

Fxposure data for assembly 47-4 are summarized in Tables 5.5 and 5.6.
Differences in flux for the various capsules were due to the flux gradlent
in thc beam hole at the MER. The values for the thcrmal flux were obtained
from the activation of CoP¥ in the stainless steel wires thal were used as
dosimeters, and were confirmed by counting Crdl. They apply to 1553 hr
during the period from March 12 to June 4, 1962. The flux was determined
on both the expused and the shaded side of the capsule, but only averaged
values are tabulated. ''hesé values udre higher than the design figures,
and the gradient is twlce the design value. If these data arc correct,
the burnup in the capsules varied over a threefold range. No explanation
has been found for the discrepancy between these values and the design
estimates, which are more consistent with heat transfer calculations.

The temperatures were controlled from capsule 24, and the changes in
the retractor position required to hold a constant temperature were not
unusual. The thermocouple reading for capsule 45 appeared to drift down-
ward by about 35°C during the 12-week exposure, but this was the only
symptom of deviant temperature recordings.

Production of Xenon and Krypton

Yiclds of long-lived and staple isotopes of krypton and xenon, respec-
tively, were about 4 and 21%, or 0.038 and 0.20 cm3 for 1% burnup of
1 g of U235, Thus the expected yields in capsule 6_were about 0.05 and
0.3 em3, and in capsules 24 and 36, 0.27 and 1.k em3, according to the
burnups listed in Tables 5.5 and 5.6.
Table 5.5.

Exposure Data for Large Capsules in Assembly 47-4

Uraniuwm Contant

Th2rmal-Neutron

Fast-Neutron

Temperature (°C)

Weight of Flux® Based (>3 Mev) Flux Fuel Power Calculated
Capsule Fuel on Co®° Based cn Co°8 Graphite- INOR-8-to- Density Density Burnug
(g) Mole & , Activation Activation to-Sals Salt (g/em®)  (w/em?) (% U?3°2)
-8 (neatrons/cm?-sec)  (neutrons/cm®-sec) Interface Interface
x 1013 x 10*2
3¢ 25.522 0.7 0.933 750
8¢ 26.303 0.7 1.023 750
i2 25.174 0.7 0.979 2.10 2.1 6804 610 2.6134 67 5.5
24 25.3% 0.7 0.987 2.7 3.2 76cd 605 2.5412 83 7.0
36 24,.386 0.7 0.968 2.7 3.3 7.4 595 2.5863 85 7.0
45 25.598 0.7 0.996 3.35 3.8 7od 600 2.5863 117 9.7
aAverage external neutron flux.
bEstimated temperatures.
cUnirradiated-controls.
61Thernocouple readings vprior to termination of final irradiation cycle.
Table 5.6. Exposure Data for Small Capsules in Assembly 47-4
U \ Content Thermal -Neutron Fast-Neutron Tempe rature
Weight of Tanium o Flux® Based (>3 Mev) Flux o g:ntml“ Fuel Pover Calculated
Capsule Fuel on Co%0 Fased on Co’ Pogion Density  Density Burnu?
(g} Mole % Activation Activation ‘(§C) (g/cm?) (w/cn?) (% U23°)
oie g (nevtrons/cm® - sec) (nevtrons/cm?-sec)
x 1013 x 1012
1b 9.381 0.7 0.365 750
3P 6.805 0.7 0.265 750
5 9.€29 1.47 D.737 895
4 10.101 1.47 J2.758 4.79 5.2 895¢ 2.5270 260 1.4
6 9.915 0.7 2.386 1.31 1.3 71L5¢ 2.5803 43 1.3

aAverage external newtrcn Tlux.

bUnirradiated controls.

cThermocouple readings prior to termination of final irradiation cycle.

TTT
112

Gas Analyses

The capsules were punctured at ORNL in the Building 4501 hot cells on
a flat surface previously machined for the purpose. The screw-driven
puncturing tool was sealed by a bellows and, at the flat surface, by a
Neoprene O-ring. Gas escaped into a calibrated collection system that had
been evacuated and leak checked. A Toeppler pump was used to transfer
gases collected at low pressures. The analyses were made by mass spectrom-
etry; those on samples containing fluorine were made at K-25.

Data from Capsules 5 and 6. The volume of helium (from the glove box
atmosphere) and argon (from the welding torch blanket) was only one half
the free volume of irradiated capsule 6, as shown in Table 5.7. This could
have resulted from an effective temperature of 300°C during welding. The
amounts of xenon and krypton are in agreement with the calculated yield.
This may mean that XeF) was absent from the capsule, or that no F, was
present, or, with even less assurance, that the CF) pressure in tfiis cap-
sule was representative of that to be expected ian thc graphite at oper-
ating conditions. If so, the CF) pressure was around 120 mm Hg at 715°C.,

Table 5.7. Analyses of Cover Gas in Irradiated

Capsule 6
Volume of gas removed: 3.2 cm?
Measured gas space in capsule: 2.5 cm?
Corrected volume of capsule gas: 1.7 cm?
Eetimated gas space at operating
temperature : 1.5 cm’
Quantity '
Gas — ' Volure (cm?)
vol % Corrected vol %
CF, 2.2 4.2 0.07
Air 47 0 1.5
He 12 23 0.38 } 1.9
A 27 51 0.86 ’
Kr 2.2 4.2 0.07 } 0.4
Xe 9.9 19 Q.32 )

Kryplou=075 and =zcnon-133 wore the anly radinactive species noted in
the gas. The isotopic distribution of xenon and krypton found by mass
spectrometry of gas from capsule 6 was as follows:

Distribution ' Distribution
Xel36 42 .06 Kr86 51.88
Xel34 31.77 Kr8? 7.51
Xel32 16.82 Kr84 26.62

Xel31 9.3 Kr83 13.99
113

The data of Table 5.8 for control capsule 5 indicate that there was
very little inleakage of air during sampling of capsule 5. The volumes
of helium and argon are in reasonable agreement with the values obtained
for capsule 6. The high temperature of the capsule during welding seems to
be a reasonable explanation for the small volume of those gases.

Table 5.8. Analyses of Cover Gas in Control

Capsule 5

Volume of gas collected: 1.3 cn’

Measured free volume of capsule: 2.7 cm’
Gas Quantity (vol %) Volume (cm?)
He 15 0.2
A 79 1.0 } 1.2
0] 0.7 0.01
CO+N 5.3 0.007

Data from Capsule 24. When irradiated capsule 24 was punctured, 84
cmd (STP) of gas was collected. This was the amount remaining after
some reaction with the apparatus. A correction has been made for the
amount that reacted with the mercury in the Toeppler pump and developed
a black scum that was identified by x-ray as HgpoFo. The gas was later
identified as predominantly Fo. ©Since the collection system was not
designed for a reactive gas, an appreciable amount of Fo must have been lost
to the metallic parts of the system. The remainder reacted with glass to
give S5iF) and O, and with other oxides to give Op. BStudies of the gas were
hampered by the presence of an excessive amount of Tel2d activity. The
free volume of capsule 24 at room temperature was 4.6 cm3, based on the
pressure change when helium was expanded into the capsule from a known
volume.

Based on mass spectrometry ot two O.l-c:m?J samples, the 84 cm3 of gas
was 5% He, 0.1% Xe, 0.4% Kr, 4% (CO + Np), 6% COp, and 17% CFy. The
remainder was Op and SiF) in roughly equal amounts, matching the products
of the reaction

5i0, + 2F, — LlFu pt O2 p o

2 2

Portions of gas were collected under a variety of conditions. The krypton
yield listed above is in fair agreement with the calculated value, but the
xcnon yield is far Loo low. Later samples, obtained after long exposure of
the capsule to an evacuated 3-liter container, had 0.3 em3 of xenon and
less than one-tenth that amount of krypton. This, like other cases of un-
explained holdup of xenon, seems now to be assoclated with the presence of
crystalline Xel).
114

The CF) content, approximately 15 cm3, would have corresponded to a
partial pressure of nearly 20 atmospheres had it been present at operating
temperature; this is higher by a factor of 100 than found in capsules that
yielded no evidence of free fluorine.

Capsule 24 was sawed open to reveal a longitudinal cross section. No
damage or evidence of corrosion was visible under low magnification. Pet-~
rographic samples from random locations were in general too fine grained
(crystallites less than 1 u in size) to be identified. One sample from
near the graphite near the center of the capsule was clearly recognized to
contain the normal phases and to have green U4’ in the TLiF.6(UF,, ThF,)
solid solution. This means that the fluoride deficiency in the salt was
shared randomly, as expected, for loss of fluorine from the frozen state
bhbut not for loss from the liquid state.

Gamma spectrometry of the samples chosen for petrographic examination
revealed a preponderance of ruthenium toward the lower part of the fuel,
vhen compared with the zirconium-niobium activlly Lhal is gencrolly
assumed to be uniformly dispersed with the ZrF) in the fuel. Cerium
activity was fairly evenly distributed with the zirconium-niobium through-
out the fuel, and cesium, along with zirconium-niobium, was found on mctal
surfaces exposed to the gas phase.

Data from Capsule 36. The analyses of the gas from irradiated capsule
36, a duplicate of capsule 24, are more directly representative of the con-
tents of the capsule because an essentially all-metal collection systcm,
comprised largely of prefluorinated nickel, was used to transfer and hold
the samples, and because a surplus of sample was available for mass spec-
trometry. Also, the analysis was made promptly. This is pertinent -
because samples allowed to sit in incompletely conditioned apparatus showed
evidence of reaction of fluorine. The analytical data are presented in
Table 5.9.

The quantity of fluorine was too great to have been present at
operating temperatures. The amount of CFj agrees well with thal found in
duplicate capsule 24. Also, the xenon appears to have been held up, as if
by XeF), while the krypton is present in roughly the expected yield.

The gamma spectrometer was able to detect Kr85 but no other activity
through the nickel walls of the sample containers used for the gas from
capsule 36. Efforts to improve this analysis are under way.

An attempt was made to remove the contents of capsule 36 by melting
and to provide representative ground samples, all under an inert atmosphere,
in a manner suitable for chemical analysis and, in particular, for a deter-
mination of the reducing power that is expected as a conscquence of the
loss of fluorine. An apparatus for accomplishing such a sequence of opera-
tion in a hot cell worked well with capsule 3A, an unirradiated control
sample, and at first appeared to have given satisfactory results with
capsule 36. The melting-out operation yielded, however, only about half
the capsule contents, and the melted-out salt was green. This might be
explained by segregation of the reduced products in the high-melting heel.
115

Table 5.9. Analyses of Cover Gas from Irradiated

Capsule 36
Volume of gas space in capsule at
room temperature: 4.6 cm?
Volume of gas space at operating
temperature: 2.6 cm’
Volume of gas removed from capsule: 188 cm?
Gas Quantity (vol %) Volume (cm®)
Fa 83.5 156
CF, 10.0 19
O, 2.5 4.7
Xe 0 0
Kr 0.3 0.56
He 2.8 5.3
CO+Ne 0.1 0.19
COz 0.7 1.3
A 0
HF (2)

aProbably present in small amounlts from reaction .
with adsorbed water.

Data from Control Capsule 3A. Capsule 3A, which was an unirradiated
control for capsules 24 and 36, was subjected to thermal cycles similar to
those in-pile. It was punctured and 4.5 em3 (STP) of 99.7% helium was
collected. The remainder of the gas was 0.05% (CO + Ne) and 0.02% COo.
The absence of argon and the presence of 1 atm of helium may be explained
by the fact that final closure involved welding crimped capillary tubes.

Discussion of Results

Experimental evidence from previous in-pile tests lends confidence
to the view that high-pressure fluorine is not generated by molten fission-
ing fuel. There is no evidence of severe corrosion on metal exposed to the
gas phase above fissioning fuel or on metal wet by fuel. Likewise, no
severe damage was observed in similarly exposed graphite.

More fluorine was found than should have been produced by the decay
energy liberated during the 5 min required for cooling to essentially room
temperature. When the reactor was shut down, the power produced in
capsules 24 and 36 dropped from 840 w to 14 w in S min, 7 w in X hr, 1.6 w
in 100 hr, and 0.19 w in 95 days (calculated from the Way-Wigner equation).
Assuming that hall of the energy released during a 95-day cooling period
was absorbed in the fuel, a G value of only 0.035 would have been required
to produce fluorine found in capsule 36.

No final choice of a theory explaining why fluorine is not always
observed can be made. A difference in crystallization behavior ic a
116

possibility. The results from capsules 24 and 36 are probably quite anoma-
lous in comparison with anything that has been seen before in earlier ex-
periments. They point to a problem that requires attention but which does
not appear to be a serious threat to the integrity of the MSRE. '
117

6. CHEMISTRY

Phase Equilibrium Studies

The System LiF-Belo-7Zr¥k,

The proposed simplification of the MSRE fuel mixture by the omission
of ThF, led to a more detailed study of the ternary system LiF-BeFs-7ZrFg4,
with a view toward prescribing suitable solvents to contain the 0.15
mole % UF4 now needed for criticality. 'T'he phase diagram of the solvent
csystem did not change significantly,; but new int'ormation on liquidus
contours within the system established the peritectic point involving
the solid phases LiF, 6LiF-<BeFs.Z2rF4, and 2LiF.BeFs at a composition of
LiF-BeFs-ZrFy (67-29.5-3.5 mole %) and at 445°C. Also, better information
on the temperaturcs along the boundary curve separating the primary phase
fields of 6LiF.BeFs.Z2rF, and LiF was obtained.

A minor surprise, in view of the rather unrestricted interchangea-
bility of the quadrivalent cations ZrF4, ThF,, and UF4 in the 10 to 15
mole % MF, concentration range, was found in the fact that 7 mole %
7rF, gives a liquidus 25°C higher than that of the 5-1-1 mole % ZrF4-
ThF,-UF, mixture when the remainder of the composition is 70 mole % LiF
and 23 mole % BeF-, as in the fuel that was nominally LiF-BeFs-ZrF4-
ThF,-UF,4 (70-23-5-1-1 mole %). Such a large change in liquidus tempera-
ture, or more specifically, in the freezing-point depression of LiF,
for a change in composition of only 2 mole ¢ is unusual in [luoride
fuel mixtures and hence provides an interecting point for further
analysis from the standpoint of the thermodynamics of solutions.

One of the specifications for the fuel was a melting point below
450°C so that the fuel would not freeze before the coolant. Since the
0.15 mole % UF, in the fuel was expected to cause only a slight lowering
of the liguidus, the fuel solvent was selected from the 450°C contour
in the LiF-BeFs-ZrF, ternary system. As shown in Fig. 6.1, which depicts
the 450°C contour on a larger scale than that used previously,l the
maximum LiF content occurs, in terms of the nearest whole numbers, near
67 mole % LiF, 29 mole % BeFs, and 4 mole % ZrF,. Since both the vapor
pressure and the viscosity should decrease as the LiF concentration
increases, this composition region is of interest as a fuel solvent.

The mixture LiF-BeFo-ZrF4-UF, (66.85-29-4-0.15 mole %), with an
estimated melting point of LL45°C, was therefore selected for particular
study. The obtuse corner in the U50°C contour marks the change from
LiF as a primary phase to 6LiF-BeFs:ZrF,, and for the fuel cémposition
designated above the primary phase is expected to be 6LiF*BeFo-ZrFy.
Whether significant differences can be detected between liquids such
as LiF-BeFn-7ZrFs (67-29-4 mole %) and LiF-BeF.-ZrF, (67-30-3 mole %)
is conjectural, but a richer ZrF, content may provc advantageous in

1"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, p 11L.
118

UNCLASSIFIED
ORNL-LR-DWG 72906

20 , 20
22

_ PRIMARY PHASES:
24 1 LiF

'6 I 6LIF-BeF,- ZrF,
Il 2LiF -BeF,
‘ IV 2LiF - ZrF,
4
| 7\m

N\
80 78 76 74 72 Tee \ 62 60
2LiR e, P (453) :
LiF (mole %)

Fig. 6.1. The System LiF-BeF,-ZrF,.

increasing the oxide tolerance of the fuel and in giving a primary
phase that would be easier to melt than LiF (mp, 840°C) in case of
accidental plugging by puarllal freezing.

Samples of compositions in the LiF-BeFs-ZrF, system that contain
more than 60% BeF, and hence are too viscous to be of present interest
as fuels were quenched from temperatures above the liquidus. The
quenched samples consisted of two amorphous materials of distinctly
different indices ot refraction, as I two glass phasco of different
composition were simultaneously present. If this implication is borne
out, the first instance of liquid-liquid immiscibility in fluoride
melts will have been encountered.

The System LiF —BEFg ~ZrF4-ThF4-UF4

Studies pertaining to concentrated solutions f'or fuel makeup and
to segregation in the quinary system that includes ThF, in MSRE fuels
were nearing completion at approximately the same time that it was
proposed that the inconvenience of carrylng loken amounta of ThF. could
be avoided. The conclusions from these investigations are applicable,
in a representative fashion at least, to fuels from the quatenary
system that do not contain ThF, and are therefore presented in the next
two sections as examples of the type of behavior that can be expected.
119

LiF-UF4 Eutectic as a Concentrated Solution of UF, for Fuel Makeup

One of the means considered for fueling the MSRE included an
initial nonnuclear operation with a solvent mixture followed by con-
version to the fuel composition LiF-BeFs-ZrF4-ThF,-UF, (70-23-5-1-1
mole %) by titrating to criticality with the liquid LiF-UF, eutectic
mixture (73-27 mole %; mp, 527°C). It was also proposed® that the fuel
makeup salt be added as the frozen LiF-UF, eutectic mixture. A study
of the phase behavior and characteristics of the composition section
that gives all possible combinations of solvent and concentrate shows
(Fig. 6.2 and Table 6.1) that the melting temperature of the LiF-UF4

UNCLASSIFIED
ORNL-LR-DWG 688394

500
<
SOLID PHASES:
480 \ A=4LiF-UR
NS 8=T7LiF-RlIF, OR 7LiF- 6 (U,ThIF, ~]
A+8+L ® g= S
'y = 6LIF - BeFy: ZrF,
5 ~—, £ = 2LiF -BeF,
< ~—
o 460 R,
x N )
2 B+C+L \ L
x
w
S 440 .
B [ J e
= N\ * DL
I Iy £ 21
_ e BIDYETL
420 T o B+C+0+L L —] ®EIhiE
e RIL+DFEFL
B4+C+D+E
400 .
25 20 15 10 5 o}
UF, {mole %)

Fig. 6.2. The Section LiF-UF, (73-27 mole %) — LiF-BeF,-ZrF,-ThF;-UF,
(70-23-5-1-1 mole %). '

mixture is higher than that of any intermediate composition formed
during fueling, that at least four crystalline compounds appear as
primary phases at various intcrmecdiate compositions, and that uranium
and thorium crystallize together in a single solid-solution phase.
From the phase relationships observed in this composition section it
appears that the LiF-UF4 eutectic mixture should serve as a suitable
fueling mixture for the MSRE. At decreasing concentrations of UF4,
uranium-containing phases precipitate later in the crystallizing
sequence. In the fuel itself, the uranium compound TLiF.6(U,Th)F4
precipitates as the tertiary phase, and solids that do not contain U4
are formed during crystallization of the primary and secondary phases.

2p, N. Haubenreich and J. R. Engel, "Safety Calculations for
MSRE," ORNL-TM-251 (May 15, 1962), p 22.

R )
120

Table 6.1. Phase Transition Temperatures in the System LiF-BeF,-ZrF,-ThF,-UF,; and Intermediate Compositions Between LiF-UF,
(73-27 mole %) and LiF-BeF,-ZrF,-ThF,-UF, (70-23-5-1-1 mole %)

Composition (mole %) ‘Transition :
Temperature Phases Observed Above Transition Temperature Phases Observed Below Transition Temperature
LiF BeF, ZrF; Th¥F, UF, . (°C) S =
72.4 4.6 1.0 21.8 0.2 476t 22 LP L + 4LiF.UF4ss® + 7LiF.6(U,Th)F,ss
465 £ 2 L + 4LiF-UF4s8 + 7IiF-6(1,Th)F, a6 L + LiF + 7LiF+6{U,Th)F4co
; 426 + 2 L + LiF + 7LiI"'-6(U,Th)TFsss L + LiF + 7LiF-6(U,Th)F4s8s + 6LiF-BeF, ' ZrF,
421 = 2 L + LiF + 7LiF-6(U,Th)Fzss + 6LiF-BeF, ZrF, L + LiF + 7LiF-6(U,Th)F,; + 6LiF-BeF,-ZrF,; + 2LiF-BeF,
S 417 2 L + LiF + 7LiF-6(U,Th)F,; + 6LiF-BeF,-ZrF, + 2Lil-BeF, LiF + 7LiF-6(U,Th)Fs;ss + OLiF-BeF,:ZrF,; + 2LiF‘BeF,
71.8 9.2 2.0 16.6 0.4 457 % 3 L + LiF + 7LiF-6(U,Th)F,ss
426 * 2 + LiF + 7LiF-6(U,Th)F;ss | L + LiF + 7LiF-6(U,Th)F; + 6LiF-BeF,:ZrF,
422 £ 2 + LiF + 7LiF.6(U,Th)F, + 6LiF'BeF, - ZrF, LiF + 2LiF-BeF, + 7LiF-6(U,Th)F;ss + 6LiF-BeF, -ZrF,
71.2 13.8 3.0 11.4 0.6 . 458 = 2 L L + LiF
439 & 2 L + LiF L + LiF + 7LiF.6(U, Th)F4ss
f 432+ 2 L + LiF + 7Lil'-6(U,Th)F,ss L + LAF + 7LiF-6(U,Th)Fs;ss + 6LiF-BeF,-ZrF,
{ 424 * 2 "L + LiF + 7LiF-6(U,Th)Fsss + 6LiF-BeF, - ZrF, LiF + 2LiT-BeFs + 7LiF-6(U,Th)Fsss + 6LiF BeF, - Z1F,
70.6 18.4 4.0 6.2 0.8 . 436 % 2 L L + 6LiF-BeF,-ZrF,
; 428 + 2 L + 6LiF-BeF,-ZrF, L + 2LiF-BeF, + 6LiF-Bel,-ZrF, + 7LiF-6(U,Th)F,ss
;'424 2 L + 2LiF-.BeF, + 6LiF-BeF,-2rF; + 7LiF-6(U,Th)F,ss 2LiF-BeF, + 6LiF-BeF,-ZrF, ‘+ 7LiF-6(U,Th)F,ss
70.0 23.0 5.0 1.0 1.0 440 =% 2 L L + 6LiF-BeF,-ZrF,
431 £ L + 6LiF:-BeFs3-ZrF, L + 2LiF-BeF,-6LiF -BeF; . ZrF,
429 £ 2 L + 2LiF-BeF, + 6LiF-BeFs-7rF, 2LiF'BeF, + 6LiF-BeF,-ZrF; + 7LiF-6(U,Th)F,ss
*The uncertainty in temperatdres shovn in this column indicates the temperature differences between the quenched samples from which the values were ob-
tained. ]
Prhe symbol "L'" refers to liquid (observed as glass or quench growth).
c 1 (
The

term

ss" means solid solution.
&)

121

Fractionation of LiF-BeFs-ZrF,-ThF,-UF, on Freezing

An assumed sequence of equilibrium fractionations in the freezing
of the five-component MSRE fuel mixture LiF-BeFs-ZrF,-ThF.-UF, (70-23-
5-1-1 mole %) was studied by quenching tests to provide information
regarding (1) the fraction of ZrF, remaining in the liquid state at
the onset of crystallization of phases containing UF4 and whether this
ZrF, concentration was adequate for protection against precipitation
of UOs, (2) the concentration of uranium in the crystalline equilibrium
phases within which it is contained during freezing and the relative
position of these phases in the sequence of crystallization reactions,
and (3) the approximate concentration of BeFs in the liquid remaining
at temperatures just above the solidus.

Although only an exploratory study of the freezing reactions was
made, some qualitative statements regarding the crystallization arc
warranted. During freezing, the ZrF4 concentration in the ligquid
fraction of the mixture is reduced as the primary and secondary phases
(6LiF.BeFo+ZrF, and 2LiF.BeF,) are formed, while the UF. concentration
within the liquid is increased and the zirconium-to-uranium concentration
ratio is decreased. As noted previously,3 the limit of protection can
he exceeded.

The apparent solidus temperature for the five-component mixture
is 429°C. C(Comparison with solidus values in the limiting LiF-BeFso-
MF, systems implies that the composition of the five-component liquid
at the solidus consists of no more than 40 mole % BeF,. The relatively
short temperature interval required for complete freezing is not
conducive to extensive segregation or to a large difference in the
average composition of the initial and final residual liquids. Hence
the products of crystallization do not contain so much free BeF, that
the cooled fuel mixture should be expected to be very hygroscopic. The
last solid phase observed on freezing the fuel mixture is TLiF.6(U,Th)F,,
which contains 1%.% mole % UF4. The proportion of UF, was determined
by two independent methods: (1) by calculation of the material balance
and (2) from measurements of the refractive indices of the solid solution.
Three nonhygroscopic solid phases were found in melts cocled under
equilibrium conditions: 6LiF*BeFs-ZrF,, 2LiF+BeFs,, and TLiF:6(U,Th)F4.
It is to be noted that specimens obtained from a variety of experiments
and experimental engineering tests, in which equilibrium cooling did not
occur, often contain 5LiF-ZrFy, 2LiF.ZrF,, 2LiF.BeFs, and TLiF.6(U,Th)Fg4,
rather than the equilibrium solids.

Crystal Chemistry

Cesium Heptafluorozirconate (CssZrf-). Crystals of the congruently
melting compound 3CsF<ZrF4 were isolated during the investigation of
phase equilibria in the CsF-ZrF, system. X-ray diffraction analysis

S"MSRP Prog. Rep. March 1 to Aug. 31, 1961," ORNL-3215, p 12k,
122

showed that these crystals were isomorphous with KzZrF7 and a;KSUF7.
The Cs and Zr* ion sites were determined, and the F~ ions were shown
to be in a dlsorderea arrangement The face-centered cubic unit cell
has a = 9.70 A%

Indium Trifluoride (InFs). The compound, InFs, which was prepared
for use in molten-salt cryoscopy studies, crystallized with the same
type of structure as FeFg and several other transition-metal trifluorides.
The rhombohedral unit cell has a = 5.73 A and @ = 56°40', and it contains
two InFs formula weights. The crystal density is 4.64 g/cm®.

Complex Alkali Fluorides: LiRbF- and LiCsFo. The crystal
structures of LiRbFz and LiCsFz were found to be isomorphous.>® In
each crystal, the Li ions are surrounded by tetrahedra of fluoride
ions, and these tetrahedra share corners and edgbs in such o manner
as to form continuous sheets. The Rb* and Cs* ions are situated
-between the sheets and bind them together. '

Stabilization of Fluorite Na¥YF, Solid Solution. The equilibrium
iriversion temperature for Lhe pure NaYl'y crystoals inverting from the
high-temperature fluorite form to the low-temperature hexagonal form is
691°C. During continuing investigations of the phase relationships
in the system NaF-YFg5, it was discovered that at nearly saturating
concentrations of YFs in the fluorite cubic solid solution, NaYF4-YFs,
s01id solutions consisting of approximately 65 mole % YFg appear as
equilibrium phases to temperatures below 4T75°C. Hence the cubic form
of NaYF, is stabilized against inversion to the low-temperature form
to the extent of approximately 15°C for each mole percent of dissolved
YF5.

Oxide Behavior in Molten Fluorides — Dehydration of LiF

A marked decrease in the corrosion of grade-A nickel containers
used for the HF dehydration of molten LiF was achieved by using a
hydrogen atmosphere at all times when heat is applied. Previously,
LiF intended for the production of ultrapure crystals was melted under
a heljum atmosphere and dehydrated with a mixture of Ho, and HF. Under
these conditions, both discolored and clear LiF was produced. The
discolored areas contained the impurities found in grade-A nickel,
such as copper, iron, manganese, "and silicon, as well as nickel, in
the 10° ppn range. Material produced using a hydrogen atmosphere
throughout did not contain discolored areas. No copper, iron, silicon,
or nickel was detected by emission spectroscopy. Only 10 ppm Mn was
found, and the total impurity was <300 ppm.

“G. D. Robbins and J. H. Burns, "X-ray Diffraction Study of
CsgZ7rF7," ORNL-TM-310 (August T, 1962). .

SPaper to be presented at the Southeast Regional Meeting of the
American Chemical Society, Gatlinburg, Tennessee, November 1962.
123

Physical Properties of Molten Salts

Density of LiF-BeFs-ZrF4-UF4

The density of a proposed MSRE fuel, LiF-BeFs-ZrF,-UF, (66.85-29-
4-0.15 mole %) was determined from the bouyancy of a platinum plummet
suspended in the melt. The volume of the plummet, corrected to the
high temperatures, was known within *0.1%, and the change in weight of
about 2.5 g could be measured to about 1 mg. The fuel was prepared
in a 200-g batch and purified by treatment with 40 g of NH.HF, to
remove oxide and hydrolysis products. Temperatures were measured
with a calibrated platinum vs rhodium thermocouple in a well that pro-
jected downward into the melt from the side at an angle. The thermo-
couple was calibrated at the National Bureau of Standards.

A stirrer was operated manually to ensure that the salt melt was
originally uniform and was not used further. Protection from the
atmosphere was maintained by flowing argon. The data for LiF-BeFs-
ZrFs-UFs (66.85-29-4-0.15 mole %) fell on the straight line

d =2.478 — (5 x 107%)t
where t is in °C, with a precision of 0.2%.

At 663°C, the fuel temperature at the outlet of the core, the
density is 2.147 g/em®. Unexplained behavior which indicated that the
plummet and stirrer had frozen in place was noted at 542°C, approximately
100°C above the melting point. Points obtained at higher temperatures
both before and after the sticking effect was noted fell on the same
straight line. Subsequent petrographic examination of the fuel disclosed
no deposits of oxide, so the data were tentatively accepted as valid.
Further checks to egtablish this point are in progress.

Estimation of Densities of Molten Fluorides

Densities of many molten fluoride mixtures, especially those of
MSRE importance, can now be estimated to within 3%. The method involves,
first, calculating the molar volumes of the mixture at 600 and 800°C
from the equation

Vnix =Z NV,
1

where N. is the mole fraction of component i and V, is the empirical
molar volume of component. i given in Table 6.2. Tfie average molecular
weight of the mixture is then divided by V . at 600 and 800°C to
obtain the density. By assuming that den51%y is linear with tempera-
ture, within the range of interest, it is a simple matter to obtain
the density equation for the mixture in the form

dt =a+ bt ,
124

Table 6.2. Empirical Molar Volumes of Fluorides

Molar Volume

(em?®/mole)

At 600°C - At 800°C
Lip 14.0 14.7
Bng , . . 23.6 24 .4
ZrFy, . 46 49
THF, ' 47.8 49.0
UF, present up to 8 mole % 58 o 60
UF, present as 8 to 100 mole % 48 - . 50

where d, is the densgity at temperature t (in °C) and a and b are con-
stants. ' |

By way of illustration, the density equation of the tentative MSRE
fuel mixture LiF-BeFs-ZrF,-UF, (66.85-29-4-0.15 mole %) is evaluated
below: '

At 600°C, |
V . = 0.6685(14.0) + 0.29(23.6) + 0.04(46) + 0.0015(58)
miX _ 18.13 cm3;
At 800°C, | |
V. = 0.6685(14.7) + 0.29(24.4) + 0.04(49) + 0,0015(60)
mix 3
= 18.95 cm’; .
Molecular Weight = 0.6685(26) + 0.29(47.01)-+ 0.04(167.22) + 0.0015(311)
= 38.17 g;
00 o0 . 18,13 18.95 —4 3 o
%00 =800 - P~ 500 = — 4.6 x 10™* g/em®.°C .
Therefore

4, = 2.381 — (k.6 x 107%)¢

in units of g/em®. Details of how Table 6.2 and the percentage
reliability were obtained are reported elsewhere.®

65, Cantor, "Reactor Chem. Div. Ann. Prog. Rep. Jan. 31, 1962,"
ORNL-3%262, p 38.
125

Graphite Compatibility

Removal of Oxide from Graphite Moderator Elements

Since the unclad graphite moderator of the MSRE is a possible
source of oxide contamination that could conceivably cause ZrOs to
form in the fuel salt mixture, treatments that favor the removal of
adsorbed gases from the moderator are included in the initial startup
sequence. Some desorption is expected when the reactor core is heated
to about 650°C in an atmosphere of flowing helium, and more can occur
when a barren fluoride salt mixture is circulated for prenuclear testing.
Accordingly, an investigation of the characteristics of gas evolution
from graphite under conditions simulating these startup operations
is in progress.

The experimental program was divided into two parts. In one set
of experiments, gas evolution from graphite blocks placed in a circulating
stream of helium at 650°C is being studied. The second part of this
program is concerned with the removal of adsorbed gases from graphite
immersed in a molten-fluoride mixture.

Gas Evolution During Helium Purging. Initially the reactor core
of the MSRE will be heated to about 650°C at a maximum rate of approxi-
mately 30°C/hr while preheated helium is circulated through the core
at about 75 liters/sec. Accordingly, pieces of graphite (6 to 12 in.
in length) were canned in a manner that provided the same cross-
sectional configuration for flow as an MSRE moderator element, and
gas entry and exil ports were provided at eithcr end. This assembly
was mounted vertically in a tube furnace as part of the circuit of a
gas-circulating loop. The loop was constructed so that portions of
the gas stream emerging from the graphite block could be diverted
through a MEECO Model W Electrolytic Water Analyzer and a gas-sampling
system. The circulated gas stream was passed through a magnesium
perchlorate drying column and a cold trap to remove moisture. Based
on a uniform gas velocity through the MSRE reactor core, a flow rate
of about 2.5 liters/min through the experimental loop was assumed to
approximate gas purge conditions around each moderator element. A
simple programming device was used to control temperatures.

The data on evolution of moisture from a block of AGOT graphite and
from an as-received piece of ETL graphite are presented in Figs. 6.3 and
6.4. The predominant characteristics of the gas evolution were the
removal of physically absorbed moisture at temperatures below 100°C and
the removal of chemisorbed moisture over the temperature interval 200
to 400°C. The additional amount removed between 400 and 600°C should
be relatively inconsequential to reactor operations. In several experi-
ments, AGOT graphite specimens were pretreated with water wvapor by
circulating moist helium through the loop. The data from subsequent
degassing, presented in Fig. 6.5, provide further illustration of the
characteristic bimodal behavior.
126

UNCLASSIFIED
ORNI~1 R-DWG 74358

OFF SCALE | ‘
>1000ppm  GRAPHITE WEIGHT: 6954¢
140 VOLUME : 448 cm3
GAS VOLUME : 2.96 liters (STP)
E GAS FLOW RATE : 2.5 liters /min . ']
8420 : TEMPERATURE RISE : 30°C /hr
=
2
2 M
L
T 400
b4 .
2 I
o
b
2 80
x
= -
w
Q
&
o B0 (3]
S \\;o
a o
‘% 20 o ' \
& h 7 R
E \(
= 4]
[«
0
0 100 200 300 400 500 600 700

TEMPERATURE (°C)

Fig. 6.3. Removal of Moisture from AGOT Graphite by Helium Purging.

UNCLASSIFIED
ORNL-LR-DWG 74359

OFF SCALE
= 1000 ppm GRAPHITE WEIGHT :1.122 kg
140 * VOLUME: 620 cm?3
GAS VOLUME: 2.88 liters {STP)
E FLOW RATE: 2.5 liters /min
o TEMPERATURE RISE : 30%C/hr
- 120 - —_—
=
>
i
o
T
Z 100
= >
=
g
& 80
>
L
o
S
< 60
a
o
% o O
C:: 40 ° AO-M
w . "-—"'Uo
'— |
T
= o
G
20 /
d 5\0/
Q
(4]
o 100 200. 300 400 500 600 700

TEMPERATURE (°C)

Fig. 6.4. Removal of Moisture from ETL Graphite by Helium Purging.

i
127

UNCLASSIFIED
ORNL-LR-DWG 74360
T 1 1 T
' OFF SCALE GRAPHITE WEIGHT: 523 g
>1000 ppm VOLUME : 315 cm3
GAS VOLUME : 2.66 liters (STP)
GAS FLOW RATE : 2.5 liters/min

400 TEMPERATURE RISE: 30°C/hr |

|
|
|

100 3
0 L LOfl{/ Efi\“”*“‘““*°°“‘

) 100 200 300 400 500 600 700
TEMPERATURE (°C}

500

WATER-VAPOR CONCENTRATION IN HELIUM (ppm)

Fig. 6.5. Results of Helium Purging of AGOT Graphite to Remove
Moisture That Had Been Added by Contact with Moist Helium.

Data on evolution of other gaseous impurities from AGOT graphitc,
as determined spectroscopically, are presented in Fig. 6.6. Since no
purification train, except for moisture removal, was included in the
helium circuit, these data demonstrate the accumulation and recombination
of impurities in the system and thus are not directly applicable to the
MSRE startup conditions.

Experiments are in progress to study various gas flow rates and
thermal conditions. Also the removal ¢of air and carbonaceous gases
will be followed under conditions that better simulate MSRE operations,
and the results will be compared with data obtained by other experi-
mental techniques.”

Gas Evolution by Immersion in a Molten Salt. When the primary
circuit of the MSRE is filled with the barren fluoride mixture for
prenuclear test operations, additional adeorbed gas may be released
from the unclad graphite moderator. Also, the salt mixture is expected
to remove the oxide film from INOR-8 and to permit this source of fuel
contamination to be flushed from the system. A laboratory demonstration
of these processes was attempted in the following manner.

An experimental flush salt, LiF-BeFs (66-34 mole %), was prepared
in a copper container by conventional techniques. Treatment with HF
and hydrogen was continued until the analyzed oxide content of the melt
was less than 200 ppm. A portion of this melt was transferred into a

L. G. Overholser and J. P. Blakely, "Reactor Chem. Div. Ann.
Prog. Rep. Jan. 31, 1962," ORNL-3127, p 125.
128

UNCLASSIFIED
ORNL-LR-DWG 74361

5 ! : | 25
GRAPHITE WEIGHT : 695¢ .

20

VOLUME : 418 cm3 /
GAS VOLUME : 2.96 liters (STP} . /o'

4 — FLOW RATE: 2.5 liters/min
TEMPERATURE RISE: 30°C/hr /
(
v/

CO, AND O, FOUND IN HELIUM (mole %)
~N W
\
)
N
N \
OO& *
o
MASS 28 (CO+ No) MATERIAL FOUND IN HELIUM {mole %

0 100 200 300 400 500 600 700
TEMPERATURE (°C)

Fig. 6.6. Gas Evolution from AGOT Graphite by Helium Purging.

closed vessel of INOR-8 alloy that contained graphite, as described
below. A connecting gas manifold permitted a known volume of helium
to be circulated over the melt and sampled for spectrographic analysis.
Provisions were made to sample the melt periodically without serious
alteration of the gas-phase constituents.

Simulation of reactor conditions- was based on duplication of the
ratio of salt volume to graphite volume and the ratio of graphite -
surface areas. Right circular cylinders of AGOT graphite, 2, 1, and
1/2 in. in diameter, were held in a positioning rack that was submerged
in the molten-salt mixture. Agitation was achieved by slowly revolving
the graphite holder. ' '

The experimental data obtained thus far in this program are shown
graphically in Fig. 6.7. A possible conclusion from these results
is that the evolution of gas from the graphite should be of little
consequence to MSRE operations. Only a few tenths of a cubic centimeter
of carbon monoxide escapes into gas space per 100 em® of graphite per
100 hr, and there is 1little effect on the oxide content of the salt.
The initial rapid rise in oxide content resembles, qualitatively, the
expected dissolution of oxide scale from the INOR-8, but in general

the values for oxide content are puzzling.

Behavior of CF, in Molten Fluorides

Additional attempts to measure the solubility of CF, in MSRE melts
brought the total to three experiments at 600°C and two experiments at
800°C in which the solubility was apparently below the limit of detection

J
129

UNCLASSIFIED
ORNL-LR-DWG 74362

3.0
CONCENTRATIONS OF Op,CHa, AND H, IN GAS PHASE —1 700
WERE NEGLIGIBLE
2.5 | VOLUME OF SALT: 1371 cm?
VOLUME OF GRAPHITE: 1070 cm3 —| s00
VOLUME OF GAS:™ 4.6 liters (STP)
-
°
2.0 500
®
//
—1 400
P

1.5
A MA%>//,. A/
T 28, ° /////
A -~
l .,/’///’ B
r“
®
\\A A N ””,/’ OXIDE —! 200
Y
A

EVOLVED GAS FOUND tN HELIUM (mole %)

OXIDE FOUND IN MOLTEN - SALT MIXTURE {(ppm)

/
o/ s e
® l.-—l,r
0 L—_—.l | A 0
o] 50 100 150 200 250 300 350
TIME (hr)

Fig. 6.7. Removal of Adsorbed Gas from AGOT Graphite Immersed in
LiF-BeF, (66-34 mole %) at 600°C.

of about 10~ moles of CF, per cm® of melt per atmosphere. The experi-
mental equipment, described previously,® was recalibrated with argon,
and the data were found to be in agreement with eariier measurements
with that gas.® Analyses by mass spectrometry of a mixture of CF. and
helium that was recirculated continuously at 700 or 800°C through
LiF-BeFs-ZrF,-ThF,-UF, (70-23-5-1-1 mole %) to which uranium turnings
had been added indicated a slow loss of CF4 at 700°C that reached =
rate of, perhaps, 4% per day at 800°C. Enough uranium had been added
to convert 20% of the uranium in the melt to UFsz, but, for reasons not
fully understood, analyses for determining the reducing power of the
melt gave consistently negative results, that is, less than 0.1 wt % U3+.
The experiment. was interrupted by trouble with a gas line before the
rate of reaction at 800°C could be established with certainty.

8y. R. Grimes, N. V. Smith, and G. M. Watson, J. Phys. Chem., 62:
862 (1958).

°G. M. Watson, R. B. Evans III, W. R. Grimes, and N. V. Smith,
J. Chem. and Eng. Data, T(2): 285 (1962).

130

Tn static experiments at 700°C involving 15 g of LiF-BeF, (66-3h
mole %) containing 1.45 wt % US" (added as UFs crystals), no evidence
of reaction of CF4 with the melt was found when a one-third excess
of CF4 was used in the cover gas during three trials in a 155-cm®
reaction vessel. The possibility that CF, might react metathetically
with oxides at 600°C was also investigated, but the results were
inconclusive.

In experiments now being prepared, CF4 will be brought into con-
tact with a reduced melt at a graphite interface to see whether the
rate of reaction will increase, and an attempt will be made to measure
the hydrolysis rate of CF,4.

Analytical Chemistry

Determination of Oxygefl in MSRE Fuel Mixtures

Work continued on the determination of oxygen in salt mixturecs
by the inert-gas carbon-reduction method. A graphite crucible tapped
to accept a threaded plug that, in turn, screws onto a graphite spindle
is used to contain the salt. The plugged capsule facilitates the
loading of the sample within the dry box and subsequent rapid weighing.

The quartz tube that is used to provide an envelope for the inert
gas was modified by enlarging the diameter of the central expanded
section and by inserting an inner quartz shell. These modifications
were intended to lengthen the useful life of the quartz tube.

Standard calibration curves were obtained by using metallic tin
standard samples of a known oxygen content that were supplied by the
Laboratory Equipment Corporation. The oxygen content of reagent-grade
LiF was assayed as a test determination. The value obtained was 250 *
40 ppm based on five determinations.

The method was further tested on samples of LiF-BeFs-ZrF,-ThF,-UF,
that were also analyzed by the KBrF, method'® and the "triton" activation
analysis methods [0*©, (H3,n)F'®]. The measured oxygen concentrations
are listed in Teble 6.3.

Table 6.3. Results of Comparative Oxygen Analyses
of MSRE Fuel Mixture

Oxygen Content (ppm)

Sample - '

No. Neutron Activation Inert Gas Fusion
KBrFy Method Method Method

4285 952, 963 804, 793, 706 897, 568, 550

4282 416 - 626, 668, 597 430, 425

4987 330, 386 747, 745, 605 404, 604, 518

(¥

131

Although the comparison is not as good as anticipated, the results
are in general agreement. Rather wide spreads are observed, however,
in the results of a single method. This was noted previously and is
possibly the result of (1) inherent errors in the methods used when
applied to fluoride salts, (2) inhomogeneity of the sample, or (3)
contamination during sampling and handling in the dry, inert-gas box.
The latter has been proved to be a definite cause of error in that the
results vary by a factor of 5 in some cases when the sample is
deliberately exposed to the atmosphere. It is not possible at this
time to eliminate either of the other aforementioned possibilities of
error, but efforts are continuing to ascertain the sources of error.

Determination of Oxygen in Graphite

Several methods were tested for the determination of oxygen in
graphite. The methods used to date include (1) fluorination with
KBrF,,'° (2) inert-gas fusion, (3) vacuum fusion, and (4) neutron
activation with 1l4-Mev neutrons. The graphite examined was high-purity
material, and the samples varied in weight from 50 to 250 mg. In the
initial tests the quantity of oxygen obtained by means of the first
three procedures was essentially equivalent to the representative
blanks for the procedures. Based on the weight of "sample used, the
oxygen concentration in the graphite was less than 50 ppm. Considerably
larger samples and attendant modifications of the apparatus would be
required to establish lower limits of sensitivity. For example, if
l-g samples were used, the oxygen concentrations as low as 10 ppm could
be deterwined.

Currently, efforts are being directed toward use of the neutron- -
activation technique using 1h4-Mev neutrons generated by the Cockroft- -
Walton generator. This method is nondestructive and capable of a high
sensitivity that depends only on the condition of the tritium target. -
Oxygen (0*®) in the graphite undergoes an n,p reaction upon irradiation
with high-energy neutrons to produce N1®, which has a 7.3-s half-life.
The gamma activity of the N'€ is counted using a gamma spectrometer
and a multichannel analyzer. Special pneumatic transfer of the sample
to the counter is necessary to achieve the ultimate sensitivity. Only
cursory results have been obtained thus far. It is planned to extend
this technique to oxygen determinations in other materials.

Homogenization of Radiocactive MSRE Fuel Samples

Equipment was designed to remove and homogenize radioactive MSRE
solid salt samples preparatory to analytical manipulations.'! The

10G. Goldberg, A. S. Meyer, Jr., and J. C. White, Anal. Chem.,
32; 31k (1960).

1M, J. Gaitanis, C. E. Lamb, and L. T. Corbin, "Homogenization
of Molten Salt Reactor Project Fuel Samples," ORNL-TM-291 (August 15,
1962) .

132

equipment includes a copper pulverizer-mixer and a copper sampling
ladle that contains the solidified fuel. The sampling ladle is placed
in the pulverizer-mixer, which is agitated on a mixer mill. The fuel
is fractured out of the ladle, pulverized into a homogeneous powder,

and transferred to a storage bottle. The homogenized fuel sample is
then available for analysis.
133

7. FUEL PROCESSING

Work on the detailed design of the fuel-processing system, which
consists of a fluorination system for uranium recovery, was started.
Equipment in addition to that described previouslyl is now included for
treatment of the salt with a mixture of hydrogen and hydrogen fluoride
for oxide removal. A simplified flowsheet of the system is presented
in Fig. T.1. ‘ :

UNCLASSIFIED
ORNL-LR-DWG 73994

NaoF ABSORBERS AT 1BO°F

50,
CONDENSER
o PREHEATERS
SAMPLER
SAMPLE fl
POT
FUEL OR FLUSHING SALT REACTOR
FROM DRAIN TANKS
Hp-HF MIXTURE
OR FLUORINE S0, F,
L~ TO LIQUID- 1
: WASTE TANK
TO OFFGAS
FUEL ' CAUSTIC FILTER
STORAGE SCRUBBER AND

TANK STACK

, N

NoF

TRAP

AT 750°C| -

Fig. 7.1. MSRE Fuel-Processing System.

In the proposed system, oxide removal is accomplished by sparging
the molten salt with an Hp-HF mixture. The Hp, H20, excess HF, and some
volatile fission products, principally niobium, tellurium, and iodine,

1"MSRP Semiann. Prog. Rep. Feb. 28, 1962," ORNL-3282, pp 13k-5.
134

are evolved. The TS50°F NaF trap removes the bulk of the niobium. The
HF-Hz0 mixture is condensed, sampled, and analyzed in order to follow
the progress of the treatment. The sample pot is emptied, after sparg-
ing, to the caustic tank, where iodine and tellurium are collected.
After sufficient decay, the caustic is sent to the liquid-waste tank
and the ORNL waste system. -

Fluorination of the fuel salt volatilizes some fission products, in
addition to UFg, but an over-all decontamination factor of >107 is ex-
pected. Additional decontamination from ruthenium, zirconium, and nio-
bium is obtained as the UFg passes through the T750°F NaF trap. The
uranium is collected as UFg-+3NaF in the 180°F NaF absorbers. These are
portable units that will be moved to the Volatility Pilot Plant for
desorption and cold trapping. Excess fluorine will be disposed of by
reacting it with SOz at LOO°F, A relatively inert gas SO F. is formed
and can be safely sent through the Fiberglas filters and discharged to
the stack. .

13
135
8. ENGINEERING RESEARCH ON THERMOPHYSICAL PROPERTIES OF SALT MIXTURES

The mixture LisC03-NazC03-K2C0s (30-38-32 wt %) has been proposed
for use in out-of-pile development studies relating to the MSRE on the
basis of similarity in properties to those of the MSRE fuel salt. Use of
this mixture will ease the handling problems because it is essentially
noncorrosive to stainless steel without a protective atmosphere, Meas-
urements of the enthalpy and viscosity of this salt are reported here.

Enthalgl

Experimental data on the enthalpy of the LizC03-NazC03-K=C0s (30-38-
32 wt %) salt in the liquid state are presented in Fig. 8.1. Over the
temperature span 475 to 715°C the data can be represented by the equation

Hy - Hypep = 18.10 + 0.413 &

UNCLASSIFIED
ORNL-LR-DWG 73992

350
5 i
S 300 —&7
a &
2 Ve
= J°
= 1 o
Z 250 | A
5 o] g
o Al o7
< =i )
T 200 ol
= _1}
|
I
150
300 400 500 600 700 800

TEMPERATURE (°C)

Fig. g.1. Enthalpy of a LigCO3—Na2003-—K2003 (30—38—32 wt %) Mix-
ture.

where the enthalpy, H, is in cal/g when the temperature, t, is in °C. A
linear fit was selected in view of the unusually large scatter in the data
(6 = 6.1 cal/g). The derived heat capacity is 0.413 cal/g-°C. The meas-
urements will be repeated with a freshly prepared sample of the salt mix-
ture in the hope of reducing the experimental scatter and establishing the
effect of temperature on the enthalpy and heat capacity.
136

Viscosity

The viscosity of the 1i>C03-Naz2C03-Kz=CO3 (30-38-32.Wt %).mixture was
determined by efflux-cup measurements.l The kinematic viscosity (v = p/p)
in the temperature range 460 to 715°C can be expressed as

v = 0.024 ¢*818/T

2

where v is in centistokes and T is in °K.

The density of the mixture has not been measured; however, assuming
ideal solution, the mixture density (based on individual component data)
is estimated to be

p = 2.212 - 0.00039 t ,
where the density, p, is in g/cms and. the temperature, t, is in °C, The

absolute viséosity'of the salt, calculated using this estimated density,
is shown in Fig. 8.2.

1"MSRP Quar. Prog. Rep. Jan. 31, 1959," ORNL-2684, p 65-67.

*
UNCLASSIFIED
ORNL-LR-DWG 73993

30

T0
&0 \ 25

NN

— 20 ~
Q
— v
- 'g
L 40 N, . £
o .
E. \ —15 8
- >
> 30 - - =
-
@ N a
8 \ 4w 8
l_f,_l -
> 20 |~ -

—~— |

0 i e vl
"800 ST (¢ 1000 100 1200 1300
TEMPERATURE {°F) :

Fig. 8.2. Predicted Viscosity of a Lip,C0O3-NasC03-K5C03 (30-38-42
wt %) Mixture. _ .

f AN
137

Previous reports in this series are:

ORNL-2799
ORNL-2890
ORNL-2973
ORNL-3014
ORNL-31.22
ORNL-3215
ORNL-3282

Period Ending July 31, 1959

Period Ending October 31, 1959

Periods Ending January 31 and April 30, 1960
Period Ending July 31, 1960

Period Ending February 28, 1961

Period March 1 to August 31, 1961

Period Ending February 28, 1962
THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
-

139

OAK RIDGE NATIONAL LABORATORY

MOLTEN SALT REACTOR PROGRAM

AUGUST 1, 1962

R. B. BRIGGS, DIRECTOR ]

MSRE OPERATIONS

MSRE DESIGN MSRE PROCUREMENT AND CONSTRUCTION
S. E. BEALL® R
E. 5. BETTIS 3 R e 8 . B. MCDONALD R
ENGINEERING INSTRUMENTATION AND CONTROLS
P. H. HARLEY R R. L. MOORE 18
:. : ;:1‘17:!: ; S. 1. BALL 1Bc
W. €. ULRICH R 2. R BROWN 8¢ PROCIIREMFNT AND FARRICATION
e ; 0, H. BURGER 15¢
4 HWESTSIX . T.M.CATE 18 C. K. MCGLOTHLAN R
L. V. WILSON D. G. DAVIS 18C F. N. ROUSER R
P. G. HERNDON s 1. M. TEAGUE R
8. J. JONES 18C
J. W. KREWSON e
C. E. STEVENSON 18C
J. R. TALLACKSON 1ac
R. WEIS cvs
MECHANICAL DESIGN =
W. M. BROWN R | CONSTRUCTION
R R 8. M. WEBSTER R
oy N. E. DUNWOODY* E&M
W, C. GEORGE R
¢ ¢ RURTY R J. P. JACKSON E8M
" R. H. JONES R ANALYSIS R. S. JACKSON E&M
A R.KERR " ). P.JARVIS E&M
- R ; |—— P. M. HAUBENREICH R ). W, FREELS FAM
J. 0. MICHOLEOH r )R ENGEL R &b huch im
L. E. PENTON R B. E. PRINCE R o
C. A. ROBERTS R -
W. ROBINSON R
F. L. ROUSER R
. G. STERLING R
J. A WATTS R
SCHEDULING
FUEL PROCESSING
— . W. GOOLSBY EaM
R. B. LINDAUER . T C. H. LAMASTER E&M
SITE PREPARATION
G. B. CHILD E&M
W, L. DERIEUX EaM
N, E. DUNWOODY* E&M
g' & ;'25::3" E:: INSTRUMENTATION AND CONTROLS
Ve SALLEE on INSTALLATION
T. A. WATSON 3] J. L. REDFORD 18C

A
D. sCOTT

COMPONENT DEVELOPMENT
SPIEWAK*

ExY

METALLURGY

A. TABOADA
K. v. COOK*
W. H. CONK
4. H. DEVAN*
R. G. DONNELLY"
R. G. GILLILAND*
J. 7. VENARD
C. H. WODTKE

J. L. GRIFFITH*

M.

. A. REDDEN*

IXTTXTTTEX X

FLOW MODELS

] R.J. KEDL*

B. J. YOUNG

£

REMOTE MAINTENANCE

bt E. €. HISE®
R. BLUMBERG
P.P.HOLZ®

R. F. BENSON
J. R. SHUGART

ERE R R

COMPONENTS

f——oA N. SCnTT

M. RICHARDSOM
. AN, SMITH

H. 5. WEBER

L. SHERSKY

EE R R

LOOPS

EER R ]

PUMP DEVELOPMENT

A. G. GRINDELL*
C. H. GABBARD"

P. G. SMITH

EEE ]

REACTOR CHEMISTRY

ANALYTICAL CHEMISTRY

L. T. CORBIN®
J. C.WHITE®

R. F. APPLE
M. J. GAITANIS
W. L. MADDOX*

CHEMICAL TECHNOLOGY DIVISION

GOMTIMEMTAL TECHMICAL SERVICES
D OIRECTOR'S DIVISION

E&M  ENGINEERING AND MECHANICAL

INSTRUMENTATION AND CONTROLS DIVISION

M METALS AND CERAMICS DIVISION

C  KEAUIUR CHEMISIRY UIVISIUN

R REACTOR DIVISION

* PART TIME ON MSRP

W, R. GRIMES* RC
F. F. BLANKENSHIP RC
S. CANTOR® RC
H, F. MCOUFFIE* RC
J. H. SHAFFER RC
R. E. THOMA® RC
D. R. CUNEO* RC

F. A. DOSS RC

J. E. EORGAN RC
H. A, FRIEDMAN® RC
G. M. HEBERT*" RC

S. 5. KIRSLIS RC

J. E. SAVOLAINEN RC
C. F. WEAVER* RC
W. K. R. FINNELL RC

B. F. HITCH RC

W. JENNINGS RC

W. P. TEICHERT RC

R. G. ROSS* RC

RADIATIAN TFSTING
C. A. BRANDON R
J. A, CONLIN® R
8C  &NAI YTIC AL CHEMISTRY DIVISION

IV.[JDocument enclosed is intended for Oral Presentation.

: Name, Location, Sponsor of Meeting ' : Date
Publication plans ave: :
1. D This paper will be included in the published proceedings of the meeting.
2. D This paper will not be included in published proceedings. After the date indicated above:
a.[ ] DTI is requested to reproduce and make TID-45 00 distribution, including copies to OTS, depository
libraries and announce in NSA.
b. D We (originator) will make TID-4500 distribution, including copies to OTS. (Note: When printed, please
transmit copies to DTI with a new PRF appropriately checked in Section I.) .
c. [:| Paper will be submitted for journal publication. (NOTE When paper is submitted for journal pubhcatlon, please
submit to DTI a new PRF appropriately checked in either Section II or IIL.)

V.[]Document enclosed is an internal or informal report not intended for TID-4500 Standard Distribu-
tion, Journal Publication or Oral Presentation.

AEC Manual Chapter 3202 requires that informal reports generally be given TID-4500 distribution, and that technical infor-
mation contained in internal reports also appear in a distributable document which receives appropriate distribution.

Chapter 3202 does recognize that issuing organizations may wish to recommend distribution limitations for informal reports
and internal reports (subsequently distributed externally) under certain conditions. It also provides for negotiation between
DTI and the originator, or DTI and the cognizant AEC Program Division if distribution limitations specified by the originator
appear questionable to OTI.

Recommendations are: _

1. DTI is|:| is not[__Jto make selected positive distribution to certain AEC contractors as appropriate.

2. DTI isD is notDto_ fill requests for this document from AEC contractors.

3. DTI is[:] is not Dto selectively distribute and fill requests for this document from other Government agencies.

4. DTI isD is not Dto make a single copy available to OTS and announce in NSA.

'IF DISTRIBUTION LIMITATIONS ARE INDICATED ABOVE, LIST JUSTIFICATION OR REASONS AS REQUIRED BY AEC
MANUAL CHAPTER 3202. ‘ ‘

Patent clearance for the document cited in this Publication Release Form has[x] has not{ ] been
obtained.

This release is submitted by/

(name typed) v
ORNL

b
-

. . . Organization

11-15-62

Date

If it is desired that correspondence concerning this document be directed to an individual other than the name above,
please indicate |

USAEC Division of Tachnicel Information Extamion, Ook Ridge, Tanneiies

Form AEC -426(11-61)
AEC Manual Chapter 3202, Appendix 064a UNITED STATES ATOMIC ENERGY COMMISSION

PUBLICATION RELEASE FORM ORNL 3369

Document Number

INSTRUCTIONS: This form should accompany each UNCLASSIFIED document the first time it is submitted to the

USAEC Division of Technical Information Extension, Post Office Box 62, Oak Ridge, Tennessee.
Molten~Salt Reactor Program Semiannual Progress

Document Title Report for Period Ending August 31, 1962 Date of Document

”

Author(s) Contract No. _~ W=7405-eng-26

|[HResearch and Development Report Enclosed is a TID-4500 Standard Distribution Report as defined in
AEC Manual Chapter 3202. No Journal Publication or Oral Presentation is Intended. (Use Section
Il or 1l below if Journal Publication is Intended or Section IV if Oral Presentation is Intended.)

1. [ xlComplete TID-4500 distribution has been made, including copies to the Office of Technical Services, Department of
Commerce (OTS sale price is $2.75 ). The number of copies specified in TID-4500 have been forwarded to the Di-
vision of Technical Information Extension for stock and for further distribution to domestic and foreign depository
libraries, foreign eéxchange organizations, etc., and for announcement in Nuclear Science Abstracts.

2. '_—l Document has been printed but complete TID-4500 distribution has not been made, Copries are being furnished for the
Division of 'I'echnical Intormation Extension to: .
a.DMake complete TID-4500 distribution including copies to OTS (sale price is § )andto depository libraries, etc.
b.[_JMake distribution to OTS and depository libraries, etc. AEC and other Government agency distribution has been .

made in accordance with TID-45U0 (OTS sale price is $ ).
c.[__]Other. Please specify

S.DNO copies have been printed for TID=-4500 and OTS distribution. DTI may reproduce from copy enclosed and make
TID-4500 distribution, including copies to OTS for public sale, depository libraries, etc. Enclosure is:
a.[_]Printed copy b.[ ] Typed copy c.[_JReproducible or multilith plates
(Up to 25 copies will be furnished to authors if desired. Indicate number .)

4.|:|This document, previously distributed as a classified report, has been declassified with[ ] without[ ] deletions.

DTI may reproduce from their master copy and make TID-4500 distribution, including copies to OTS for public sale,
and to depository libraries, etc.
(Up to 25 copies will be furnished to authors if desired. Indicate number .)

Il.L_IDocument Enclosed is a TID-4500 Standard Distribution Report which is also intended for Journal
Publicalion: .

1.[ ] T1-45uu (AEC) distribution has been made. Copies are enclosed for DTI to make single copy distribution to
OTS and to the domestic depository libraries, and for announcement in NSA.

2.[\ ] Copies are being fumished DTI to make TID-4500 (AEC) distribution, and single copy distribution to OTS and to
the domestic depository libraries, and for announcement in NSA.

3.[ ] From the copy enclosed, DTI is requested to reproduce in Microcard form and make TID-4500 (AEC) distribution,
domestic depository library distribution and send 1 full size copy to OTS, and announce in NSA.

Document enclosed has or will be submitted for publication in the following scientific journal:
. N\

(Name of Journal) (Expected date of issuance)

I11.[_]Document enclosed is intended for publication in a journal whose publication policy precludes
advance distribution within the AEC and single copy distribution to OTS and to domestic deposi-
tory libraries.

1. Paper has been or will be submitted for publication in the following scientific journal:

(Name of Journal) (Expected date of Issuance)
(NOTE: DTI will hold this document for internal use and will not announce in NSA. No further distribution will be
made except in unique circumstances when the report is required by present work of another AEC Contractor in
advance of the paper’s appearance in the journal. Such further distribution by DTI will be limited to specific re-
quests for this information.)

{Over)
smp——

7

Rpt. No.

Std. C- [§Z1; mc- [_]; PNT C- C] LTDD pBL_1;BK [1; Und|:|

(J ACCELERATORS
[(J ADMINISTRATIVE
[ AEROSOLS

[(J ANALYTICAL

[0 BIBLIOGRAPHIES
O BIOMED

[0 BLAST EFFECTS
[0 CERAMICS
JZXCHEMISTRY

(] COMPUTERS

[0 CORROSION

O cOsMIC RAYS

O CRITICALITY

[J CRYOGENICS
O DECONTAMINATION

(] DIRECT ENERGY CONVERSION

(0 DOSIMETRY -
0 ECOLOGY

(0 ELEMENTARY PARTICLES

[0 ENGINEERING

[0 EQUIPMENT

J FALL-OUT

] FEED MATERIALS

[0 ¥ISSION PRODUCTS
(] FLUID FLOW

(J GEOLOGY

] HEALTH & SAFETY

] HEAT TRANSFER

O INSTRUMENTATION
[J 1ISOTOPE APPLICATION
[C] ISOTOPE SEPARATION
C] LABORATORIES

|

(6-62)

[J LUBRICANTS

| MAGNETOHYDRODYNAMICS
(] MATERIALS

O MATHEMATICS

[J MEETINGS
%@ETALLURGY

] METEOROLOGY

[J MODERATORS

[] NUCLEAR EXPLOSIONS
(] PARTICLE PHYSICS

(] PHYSICS

E“PLASMA PHYSICS

[0 PLOWSHARE

' 4
qu PROGRESS REPORTS

] PROPULSION SYSTEMS
O RADIOBIOLOGY

(] RADIOACTIVE WASTE
J RADIATION EFFECTS
J RAMJETS

[CJ] RARE EARTHS

'] RAW MATERIALS

[0 REACTOR CONTROL
] REACTOR COOLING
] REACTOR ECONOMICS

@ REACTOR ENGINEERING

[J] REACTOR FUELS

[J REACTOR HAZARDS
[fignEACTOR MATERIATS
(J REACTOR PHYSICS

(O REACTORS

J REACTORS—~AHR

-

] REACTORS-BWR

REACTORS— FFR
] REACTORS—-FSR
[J REACTORS—GCR
[[] REACTORS—HWR
[J REACTORS-LMCR
[ REACTORS—LMFR
[J REACTORS—-OMR
[ REACTORS—PWR
(J REACTORS—SGR
[ REACTORS—AIR
[J REACTORS—CENTRAL
[(J REACTORS-DUAL PURPOSE
[] REACTORS—MARITIME
[ REACTORS—NAVAL
(] REACTORS—PACKAGE
[(CJ REACTORS—-POWER _
0 REACTORS-PRODUCTION
(J REACTORS—PROPULSION
[ REACTORS-RESEARCH
[(J REACTORS—SPACE
[J REACTORS—TEST
[J REFRACTORIES
[J ROCKETS
[ SATELLITES
(] SEPARATION
[C] SHIELDING
] STRUCTURES
(J THERMONUCLEAR
7 TRANSLATIONS
(] WEAPONRY

O

USAEC Divitlon of Technicol Information Exterlon, Ock Kidge, Tenneuse
Ly W
W

HMuoWLwWwLuw
SYoBEIG S

N~ R
Sl

HCEN S LSNP OHNEOONEEEADEOREQ

WEEEQPEPNEONUPEETR S

.

QOEPPHUEUEON UNNNEEQ X

TECHNONPNEENIAYFAECE PP Q

Adamson
Alexander
Beall
Bettis
Bettis
Billington
Blankenship
Blizard
Boch
Bohlmann
Bolt
Borkowski
Boyd.
Breeding
Briggs
Bruce
Burke
Campbell

antor

Cobb
Conlin
Cook
Corbin
Cristy
Crowvley
Culler
DeVan
Donnelly
Dougl.as
Epler
Ergen
Fraas
Frye, Jr,
Gabbard
Gall
Gallaher
Grimes
Grindell
Harrill
Hill
Hise
Hoffman
Holz .

141

ORNL-3369 _
UC-80 — Reactor Technology
TID-4500 (18th ed.)

Internal Distribution

44, A, Hollaender

45. A. S. Householder

46, L. N, Howell

47. W. H. Jordan

48, R. G. Jordan

49, P. R. Kasten

50. R. J. Kedl

51. M. T. Kelley

52. B. W, Kinyon

53. R. W. Knight

54. J. A. Lane

55. C. E. Larson

56. T. A. Lincoln

'57. S. C. Lind

58. R. B. Lindauer

59, R. S. Livingston

60. M. I. Lundin

6l.. H. G. MacPherson

62. W. D. Manly

63. E. R. Mann

64. H. F. McDuffie

65. W. B. McDonald

66. C. K. McGlothlan -

67. E. C. Miller

68. R. L. Moore

69. K. Z. Morgan

70. J. C. Moyers

71. J. P. Murray (K-25)

72. M. L. Nelson

73. C. W. Nestor

74. T. E. Northup

75. W. R, Osborn

76, L. F. Parsly

77. P. Patriarca

78. H. R. Payne

79. D. Phillips

80. W. B. Pike

81l. M. Richardson

82. R. C. Robertson

83, T. K. Roche

84. H. W. Bavage
85~-86, A. W. Savolainen

87. D. Scott
88.
89.
0.
91.
92.
93.
% .
95.
96.
97.
98.
99.
100.
101.
102.
103.

153~

159-

154.
155.
156.
157.
158.
765.

U!fitfl:lc-ibc-qfll.—ibfdbfizm_&q?d

E. Seagren 104.
H. Shaffer 105.
D. Shipley 106.
J. Skinner 107.
M. Slaughter 108.
N. Smith -109.
G. Smith 110.
H. Snell 111.
Spiewak 112.
D. Susano 113-114.
A. Swartout 115-118.
Taboada ‘
R. Tallackson 119-121.
H. Taylor 122-151.
E. Thoma 152.
B. Trauger

D. F. Cope, AEC, ORO

J. Wett, AEC, Washington
R
A. W. Larson, AEC, ORO

142

Ulrich

. Watkin

Watson

Weinberg

Westsik

White

Wilson

. Wodtke

Biology Library

Reactor Division Library
ORNL—~Y-12 Technical Library,
Document Reference Section
Central Research Library
Laboratory Records Department
Laboratory Records, ORNL R. C.

Qg r Qg =
mdQm=E2=2Zaa

Py
3,
R

External Distribution . | .

. W. McNamee, Manager, Research Administration, UCC, New York

Division of Research and Development, AEC, ORO
Given distribution as shown in TID-4500 (18th ed.) under Reactor
Technology category (75 copies — OTS)