ORNL-2626.txt

MS LIBRA

Iil HA'tnl]iTI"jiiIilvnsis'Timl Il ulmil

3 445k 03b132kL 7

CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

!
'{ LIBRARY LOAN COPY
‘ DO NOT TRANSFER TO ANOTHER PERSON

If you wish Someone else to see this
document, send in name with document
] and the librory will arrange a loan,

ORNL-2626
Reactors=Power

TID-4500 (14th ed.)

Contract No. W-7405-eng-26

MOL TEN-SALT REACTOR PROGRAM
QUARTERLY PROGRESS REPORT

For Period Ending October 31, 1958

H. G. MacPherson, Program Director

DATE ISSUED

SANG 1959

OAK RIDGE NATIONAL LABORATORY
Ock Ridge, Tennessee
operated by
| UNION CARBIDE CORPORATION
| for the
JE 1111

i

3 4456 03kL32L 7

MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT

SUMMARY

PART 1. REACTOR DESIGN STUDIES

1.1. Conceptual Design Studies
and Nuclear Calcvlations

Further studies of the Interim Design Reactor
have resulted in an improved fill-and-drain system
and alternate ways of fransferring heat from the re-
actor fuel to steam. The revised fill-and-drain
system retains the previously described after-heat
removal scheme, but the components have been re-
designed to make remote repair and removal feasible.
As now designed the entire heat removal system may
be removed or inserted from overhead without dis-
turbing the fuel circuitry.

A simplified fuel-to-steam heat transfer system is
being considered that eliminates intermediate heat
transfer fluids. The direct fuel-to-steam heat trans-
fer system used in conjunction with a Loeffler
boiler cycle allows control of the reactor and load
relationship with conventional steam cycle com-
ponents. Preliminary cost studies based on a 640-
Mw (heat) two-region homogeneous reactor were
made for several heat transfer cycles, including the
direct fuel-to-steam cycle, and it was found that
the direct cycle gave the lowest first cost. A
helium-cooled cycle gave the highest first cost, but
optimization of the reactor system would lower the
cost appreciably.

Nuclear calculations relative to a 30-Mw experi-
mental reactor with a 6-ft-dia core have indicated
that the critical mass of the system at 1180°F
would be 49 kg of U235 and the total inventory for a
system volume of 150 f#2 would be 65 kg of u23s,
After a year of operation, the critical mass would
have increased to 61 kg, with a net integrated
burnup of 14.4 kg.

Calculations of the initial states of
reactors were completed. Data were obtained for
critical mass, inventory, regeneration ratio, etc. for
core diameters ranging from 4 to 12 ft and thorium
fluoride concentrations in the core of 2, 4, and 7
mole % for 600-Mw (thermal) reactors. Additional
calculations covering a 20-year operating period
of the 8-ft-dia core with 7 mole % thorium showed
that, after five years, the reactor is self-sustaining.

Calculations were also made of the fissions in the
blanket of the Interim Design Reactor. To date,
calculations covering only the first five years of

U233 fyeled

operation have been completed, but the fraction of
fissions in the blanket appears to have stabilized
at about 0.02. Methods for calculating the nuclear
characteristics of graphite-moderated, hetero-
geneous, molten-fluoride-salt reactors are being
developed.

Gamma-ray energy absorption coefficients were
computed as a function of energy for 24 elements:
hydrogen, carbon, beryllium, nitrogen, oxygen,
magnesium, sodium, aluminum, silicon, sulfur,
phosphorus, argon, potassium, iron, calcium, copper,
molybdenum, iodine, tin, tungsten, platinum, lead,
thallium, and uranium. The calculated values have
been assembled on paper tape for use with the
gamma-ray heating program Ghimsr being written for
the Oracle.

1.2. Component Development and Testing

Development tests of salt-lubricated pump bear-
ings were continued. Satisfactory molten-salt
hydrodynamic bearing films were established in
two tests of an INOR-8 bearing and an INCR-8
journal. Slight rubbing marks found after extended
service are being investigated. Equipment for test-
ing hydrostatic bearings with stationary pockets was
completed, and data on bearing loads and flows were
obtained in preliminary tests with water. Air en-
trainment encountered in preliminary dynamic tests
of a rotating-pocket hydrostatic bearing was cor-
rected, and the pressure distribution in the pockets
is being investigated.

A conventional aluminum bearing and an Inconel
journal that had been lubricated with Dowtherm
““A’ (a eutectic mixture of diphenyl and diphenyl-
oxide) during 3200 hr of operation in a pump that
was circulating NaF-ZrF -UF , at 1200°F was ex-
amined. The clearance between the bearing and
journal had not changed and there were only slight
indications of wear.

The oil-lubricated centrifugal-pump rotary element
assembly being operated in a gamma-irradiation
facility at the MTR had accumulated a total of
5312 hr of operation and a gamma-ray dose rate to
the lower seal region of 9.3 x 107 r by the end of
the quarter. Samples of the bulk oil and the seal
leakage oil show that the viscosity has increased
and that the bromine number has increased. The

increases have not been sufficient to affect oper-
ation, however. The acidity number of the oil in-
dicates that little oxidation of the oil has occurred.
The test will be terminated when the dose to the
lower seal region has reached 1070 .

Materials required for a motor suitable for long-
term operation at 1250°F in a radiation field are
being investigated. Six coil assemblies that in-
corporate electrical insulation developed for use
at high temperatures were received from the Louis-
Allis Company for testing. Suitable magnetic core
materials and electrical conductors are being sought,

Fuel pump design studies that were contfracted to
the Allis-Chalmers Manufacturing Company and the
Westinghouse Electric Company were completed.
The results of both studies emphasized the need
for the development of salt-lubricated bearings.
Various layouts of centrifugal pumps, both sump
type and submerged, were suggested.

A small, submerged, centrifugal pump with a
frozen-lead seal is being operated in evaluation
tests. During continuous isothermal operation of
this pump for 2500 hr at 1200°F, with @ molten salt
as the pumped fluid, there has been no leakage of
lead from the seal. A similar lead-sealed pump
with a 3]/4-in.-dia shaft is being fabricated for test-
ing.

Freeze-flange and indented-seal-flange joints that
had sealed successfully in high-temperature molten-
salt lines were tested with sodium. Both joints
operated successfully; there was no indication of
sodium leakage. Two large freeze-flange joints
were tested in a 4-in. line carrying a high-temper-
ature molten salt, There was no salt [eakage, but
gas leakage of the flanges was slightly in excess
of the allowable 107 cm? of helium per second.
Modifications are now being made to improve the
gas tightness.

Heater-and-insulation units designed for use on a
4-in. line were tested along with the large freeze
flange. The thermal loss from two units was found
to be about four times the heat loss from an equi-
valent length of ''Hy-Temp’' pipe insulation 3 in.
thick.

Three commercially available expansion joints
were tested and found to be unsatisfactory for use
in molten salt or sodium lines. Since the bellows
that failed in two of the tests were weakened by the
attachment of thermocouples, similar units will be
tested without thermocouples.

Data were obtained from which a plot of the sub-
limation temperature vs vapor pressure was pre-
pared, and the information was used in the design
of a thermal-convection loop for evaluating the heat .
transfer performance of aluminum chioride gas.
Exposure of Inconel and INOR-8 specimens to
aluminum chloride vapor for 1000 hr showed that
either material would be satisfactory for con-
struction of the thermal-convection loop.

Construction and operation of forced-circulation
corrosion testiag loops was continued. Of the
thirteen facilities with operating loops, six are of
the new construction that gives maximum protection
against freezing of the salt in the event of power
failures, and the remaining seven loops have been
improved so far as possible without changing the
loop piping. Assembly of the first in-pile loop was
almost completed, and a second loop is being
assembled.

1.3. Engineering Research

The enthalpies, heat capacities, and heats of
fusion were determined for the mixtures NaF-BeF _-
UF , (53-46-1 mole %), LiF-BeF -UF , (53-46-1 mole
%); NaCl-CaCl , (49-51 mole %), LiCI-KCI (70-30
mole %), LiCI-KCl {60-40 mole %), and LiCl-KCI
(50-50 mole %). Preliminary values were obtained
for the surface tension of LiF-BeF -UF , (62-37-]
mole %) over the temperature range of 460 to 750°C.
Apparatus for measurements of the coefficient of
thermal expansion, thermal conductivity, viscosity,
and electrical conductivity of beryllium-containing
fluoride salt mixtures are being assembled.

The cause of a major discrepancy in heat balance
in the study of the heat-transfer coefficients of LiF-
BeF ,-UF, (53-46-1 mole %) flowing in a small-
diameter Inconel tube is being investigated. A
pump system for investigating such long-range
effects as film formation by comparison of heat-
transfer coefficients is being constructed.

1.4. Instrumentation and Controls

The results of a series of tests of Inconel re- .
sistance-type fuel-level-indicating elements have
not been reproducible, as yet, but it has been
established that fuel 130 has sufficient resistance -
to provide a useful milliwatt output from the probe.
Polarization and surface tension effects are being
investigated as possible sources of the incon-
sistencies.

1.5. Advanced Reactor Design Studies

Graphitesmoderated molten-salt reactors fueled
with low-enrichment material were studied, and the
results of preliminary calculations indicated that
initial enrichments as low as 1.25% U233 might be
used in a circulating-fuel system. Highly enriched
uranium would be added as makeup fuel, and such
reactors could probably be operated to burnups as
high as 60,000 Mwd/ton.

A simplified natural-convection molten-salt re-
actor for operation at 576 Mw (thermal) was studied
to determine the approximate size of components
and the fuel volume. Natural-convection power re-
actors of this size are characterized by high fuel
volumes and large numbers of heat exchanger tubes.
Therefore it is expected that the initial cost of a
natural-convection reactor would be higher than that
for a forced-circulation system, and the large number
of heat exchanger tubes casts some doubt as to
whether the natural-convection system would
actually be more reliable than the forced-circulation
system.

PART 2. MATERIALS STUDIES

2,1. Metallurgy

Two Inconel and four INOR-8 thermal-convection
loops, which circulated various fused-fluoride-salt
mixtures, were examined. An Inconel loop which
operated 1000 hr at 1350°F with the mixture LiF-
BeF,-UF , (62-37-1 mole %) was attacked to a depth
of 3 mils. A loop operated previously with this
mixture at 1250°F showed approximately 1 mil of
attack over a similar test period. Operation of an
Inconel loop at 1050°F for 4360 hr with the mixture
NaF-LiF-KF (11.5-46.5-42 mole %) resulted in the
formation of subsurface voids to a depth of 2 mils.
No attack was evident in any of the INOR-8 loops
even after operation, in one case, for 3114 hr.

Test results were also obtained for three Inconel
forced-circulation loops operated with fluoride
mixtures. A loop which circulated the mixture
NoF-Zer-UF4 (57-42-1 mole %) and also con-
tained a secondary sodium circuit showed maximum
attack to a depth of 1.5 mils in the salt circuit and
only slight surface roughening in the sodium cir-
cuits. No cold-leg deposits were present in either
circuit. Mixtures of LiF, BeF, and UF produced
attacks of 5 and 8 mils respectively in two other

loops operated for slightly more than 3000 hr at a
maximum hot-leg temperature of 1300°F.

INOR-8 samples removed from a forced-circulation
loop that was shut down after 3370 hr of operation
for repair of a leak at a heater lug were found to have
no evidence of surface attack. Etching rates showed
slight differences between as-received and as-tested
specimens, Operation of the loop has been resumed.
The mixture LiF-Ber--UF4 (53-46-1 mole %) is
being circulated in this loop. The maximum walil
temperature is 1300°F. The failure occurred at a
weld of the tubing to a heating lug adapter, and it
was found that the wrong weld material had been
used,

Routine inspection of the INOR-8 material now
being received has revealed that the quality is as
good as that of Inconel and of stainless steels made
to ASTM standards. The rejection rate for fully in-
spected welds is now about 10%,

A sodium-graphite system was used to prepare
Inconel and INOR-8 tensile test specimens for a
study of the effect of carburization on the mechanical
properties, The carburization at 1500°F caused in-
creases in both the yield and tensile strengths of
Inconel and decreased the ductility at room temper-
ature. For the INOR-8 specimens, the yield strength
was increased, the tensile strength was slightly
decreased, and the ductility was greatly reduced.
The changes were dependent on the amount of car-
burization. In similar tests at 1250°F the carburized
specimens again showed a trend to fower ductility,
but some specimens were more ductile than control
specimens.

In comparative tests of the carburization of INOR-8
by fue! 130-graphite systems, 3-mil surface cuts of
specimens exposed only to fuel analyzed 640 ppm
carbon, while similar cuts of specimens exposed to
fuel containing a graphite rod analyzed 940 ppm
carbon. Mechanical property tests showed the same
trends as those found for the specimens carburized
in sodium-graphite systems.

For a test of the penetration of graphite by fuel
130, a graphite crucible 4% in. long with a tapered
hole 0.43 in. ID and 3‘/2 in. deep was loaded with the
fuel and tested in a vacuum at 1300°F. Radio-
graphs were made at 500-hr intervals to study the
penetration of the graphite by the fuel. At the end
of the first 500 hr test of a CCN graphite crucible,
there had been no penetration of the graphite, but a
disk of UO, had formed at the bottom of the tapered
hole that was 0.1 in. thick. Subsequent analyses
showed that it contained 28% of the vranium
originally present in the fuel. Further tests with
TSF in contact with fuel 130 have shown similar
precipitation of UQ, but to a lesser extent. Since
the precipitation was least in the TSF graphite,
which is purer than the CCN grade, it is thought that
the precipitation may be caused by impurities in the
graphite.

The precipitation that occurred in these tests was
not observed in the carburization tests described
above. Further it is in contrast to the lack of pre-
cipitation in similar tests in which fuel 30 (NaF-
ZrF ,-UF ,, 50-46-4 mole %) was used. Metallographic
examinations of the crucible showed no attack and
no penetration of salt into the pores.

Corrosion tests of pure silver and silver-base
brazing alloys in static fuel 130 showed silver and
its alloys to have very limited corrosion resistance.
On the other hand, gold-bearing alloys have shown
excellent corrosion resistance in fuel 130.

A time-temperature-stress relationship was formu-
lated for the creep strength of INOR-8 to aid in
predicting long-time creep strengths. Heatsto-heat
variation studies indicated a 0,14%-carbon-content
heat to be appreciably stronger than two other INOR-8
heats containing 0.02 and 0.06% carbon.

Eleven alloys with compositions bracketing the
maximum amount of each major alioying element of
INOR-8 have been vacuum-induction melted and are
presently being aged at temperatures in the range of
900 to 1800°F. These alloys will be tested in order
to determine the effect of composition on structural
stability.

Specimens of commercially fabricated INOR-8 were
aged for 5000 hr at test temperatures of from 1000
to 1400°F. These specimens exhibited no in-
stabilities in tensile tests that would be detrimental
during long-time service.

Room- and elevated-temperature mechanical
property tests of INOR-8 weld metal both in the as-
welded and in the welded and aged conditions were
made and the results were compared with similar
data for Hastelloy B and Hastelloy W. Aging
seriously reduces the room-temperature ductility of
Hastelloy B and Hastelloy W weld metal, while the
ductility of INOR-8 weld metal is improved. The
influence of melting practice and carbide spheroid-
ization on the elevated-temperature ductility of
INOR-8 is being studied.

The brazing of thick tube sheets is being studied.
Promising results were obtained in a test braze of

vi

a ¥%-in,-OD, 0.065-in.-wall Inconel tube into a 5-in.-
thick Inconel tube sheet with Coast Metals No. 52
alloy. The brazing alloy was preplaced in annular
trepans in the tube sheet and fed to the joint through .
three small feeder holes. Methods of joining
molybdenum and nickel-base materials are also being
studied for pump application.

2,2. Radiation Damage

The electrically heated mockup of the INOR-8
in-pile thermal-convection loop was operated under
simulated in-pile conditions. This experiment
served as a performance test for the various com-
ponents of the loop and indicated that modifications
of the cooling=air injection system were required.
The in-pile model of the loop is being assembled.

Two fluoride-fuel-filled graphite capsules were
examined that had been irradiated in the MTR for
1610 and 1492 hr at 1250°F. The graphite was un-
damaged.

2.3. Chemistry

Phase equilibrium studies of fluoride-salt systems
containing UF , and/or ThF , were continued. Data
are being obtained on the NaF-Ber-ThF4-UF4
system to provide a comparative basis for evaluation
of lithium-base quaternary systems.

Studies have revealed the cause of low concen-
tration of uranium in 50-lb batches of LiF-BeF -
UF4 (62-37-1 mole %) that were transferred from
250-1b batches. The first liquid to form when the -
mixture is melted is the eutectic composition con-
taining 3.1 wt % uranium rather than the nominal
6.5 wt % uranium in the 250-1b batch. Thus the
composition of any partially molten batch must be
intermediate between that of the lowest melting
eutectic and the nominal composition of the liquid.
Complete melting of the mixture is necessary to
maintain composition homogene ity.

Extensive gradient-quenching experiments were
completed on the LiF-BeF ,-ThF, system. Asa
result of these experiments several modifications
were made in the preliminary phase diagram issued -
previously. Three-dimensional models of this
system and of the previously studied system LiF-
ThF -UF , were constructed.

The effects of divalent and trivalent fission-
product ions on the solubility of PuF, in LiF-BeF,
mixtures was studied, The effect of the trivalent
ions was found to be sufficient to indicate that they
should not be allowed to build up in the fuel, but
the effect of the divalent ions was insignificant.
Data were obtained that provide a basis for the
use of trivalent ions in a scheme for reprocessing
fused-salt mixtures containing plutonium.

Solubility studies were made on several systems,
including argon in LiF-BeF, (64-36 mole %); HF in
LiF-BeF , mixtures containing 10 to 50 mole %
Ber; Cer in Nc:F-LiF--BeF2 solvents of various
compositions; CeF3, LoF3, and SmF3 in LiF-Ber-
UF, (62.8-36.4-0.8 mole %); CeF, and LaF, simul-
taneously in LiF-BeF -UF, (62.8-36.4-0.8 mole %);
and CeF3 and SmF ; simultaneously in LiF-BeF ,-
UF, (62.8-36.4-0.8 mole %). An elementary model
was studied with which estimates can probably be
made of the solubilities of the noble gases in media
of high surface tension to within an order of mag-
nitude, which is sufficiently accurate for reactor
applications.

Two methods of separating uranium from un-
desirable fission products in fluoride-salt fuel
solvents are being studied. Experiments are in
progress in which a chromatographic type of
separation of uranium and fission-product rare
earths is effected by using beryllium oxide as the
column packing material. In other experiments
uranium precipitation upon the addition of water in
the influent gas stream is being investigated.

Tentative values were obtained for the activities
and the activity coefficients of nickel metal in
nickel-molybdenum alloys. The values were calcu-
lated from measurements of the electromotive force
of cells containing pure nickel and nickel-molybdenum
alloy electrodes in a bath of NaC|-KCl eutectic con-
taining 5 wt % NiCI2.

Apparatus was set up for measuring the surface
tensions of BeF , mixtures. Trial runs gave a value
of 195 dynes/cm at 425 to 450°C for LiF-BeF ,-UF
(53-46-1 mole %) and 196 dynes/cm at about 480°C
for LiF-BeF2 (63-37 mole %).

In the study of the chemical equilibria involved in
the corrosion of structural metals, the effect of
solvent composition on the equilibria

f‘)Cer,?:"‘?CrF3 + Cr®

and
3FeF2—\‘\__ 2FeF , + Fe®
was investigated. The solvent LiF-NaF-ZrF  was

used because it permitted a continuous variation
from a basic to an acidic melt without interference

from precipitation. Increasing disproportionation ot
CrF, was found with decreasing ZrF , content.
Similar tests in solvents containing BeF ,, which

is considered to be more basic than ZrF ,, showed
that in an LiF-BeF, mixture containing 52 mole %
BeF, the disproportionation of CrF, was roughly
comparable to that in a solvent containing 35 mole
% LrF . These results confirmed the supposition
that BeF, mixtures are more basic than correspond-
ing ZrF , mixtures and that the extent of dispro-
portionation of (:rl:2 is greater in the more basic
solvents. Experiments carried out with FeF, in the
same solvents showed no disproportionation.

Further measurements of the activity coefficients
of CrF, dissolved in the molten mixture NaF-ZrF
(53-47 mole %) were determined and equilibrium
quotients were calculated. The results of the
measurements of CrF_ and the corresponding in-
vestigations of NiF, and Fer, along with the
values for the activities of chromium in alloys, can
now be used in the study of corrosion of Inconel
and INOR-8 alloys. [t is thought that the rate of
corrosion in these alloys is diffusion controlled.

Calculations were made of the amounts of
chromium metal which should be added to pre-
equilibrate a salt prior to a test in an Inconel loop.
The corrosion by a pre-equilibrated salt should de-
pend only on the hot-to-cold-zone transfer mechanism,
and the results should be indicative of long-term
cofrosion rates in reactors,

Studies of chromium diffusion in chromium-con-
taining nickel-base alloys were continued. It is
hoped that the results of these experiments will
provide a means for predicting the void penetration
distances to be expected for a given set of cor-
rosion conditions.

An experimental study was made of the corrosion
of Inconel by aluminum chloride vapor in an all-
metal system. The results indicated that the amount
of attack of Inconel can be expected to be pro-
portional to the pressure of the aluminum chloride
vapor and that at 5 atm the corrosion after 300 hr
would be about 0.7 mil. INOR-8, in which the
activity of chromium is much lower than in Inconel,
should prove quite resistant to aluminum chloride
vapor. Other alloys selected on the basis of
chemical and structural considerations are to be
tested,

The containment of aluminum chloride is of in-
terest because the gas has unique properties which
make it a heat transfer medium of potential utility

vii

at high temperatures. First, because of the large
number of atoms per molecule, the heat capacity of
the gas is high. Further, because at high temper-
atures the major species is AICl, and at lower
temperatures the major species is Al Cl, cooling
the gas in a heat exchanger will remove, in addition
to the heat obtained from the temperature change,
the heat from the dissociation reaction, Estimates
of the physical properties of the gas were made to
provide a basis for the design of a thermal-con-
vection loop in which to test the heat transfer
characteristics and the corrosiveness of the gas.

In tests of the penetration of graphite by molten
salts, it was found that NaF-LiF-KF penetrated to
greater extent than LiF-MgF ,, as indicated by
weight gains. The NaF-LiF-KF mixture has a lower
density and a lower viscosity than the LiF-MgF,
mixture. In both cases there was even penetration
to the center of the graphite rod.

A new method of preparing chromous fluoride was
explored in which the reaction is

SnF, + Cr—CrF, + Sn

The chromium-containing portion obtained in the

viii

experiment had the crystallographic properties of
pure chromous fluoride. Improved methods for pre-
paring vanadium trifluoride and ferrous fluoride were
also developed.

2,4, Fuel Processing

Sufficient laboratory work has been done to con-
firm that fluorination of the fuel salts LiF-BeF ,-
UF , or LiF-BeF -UF ,-ThF , results in good
uranium recovery. Therefore the fluoride volatili-
zation process, which was developed for hetero-
geneous reactor fuel processing and was used
successfully for recovery of the uranium from the
fuel mixture (NaF-ZrF4-UF4) circulated in the Air-
craft Reactor Experiment, appears to be applicable
to processing of fuels of the type now being con-
sidered for the molten-salt reactor.

Developmental work has been initiated on the
processing of the solvent salt so that it can be re-
cycled. The recovery process is based on the
solubility of LiF-BeF, in highly concentrated
aqueous HF and the insolubility of rare earth and
other polyvalent-element flueorides.

CONTENTS

SUMMARY .ottt ettt e e ets b e tee b e et e s ae st e b et beeae s o4eshb e bebe e st e sabete 4 ebmenbeentebe e seenseanserneseseennenneesenneene iii
PART 1. REACTOR DESIGN STUDIES

. 1.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS ....cccoeiiiiiiiinincniic 3

Interim Design Reactor ...ttt s 3

Fuel Fill-and-Drain Sy stem ... s 3

Fuel-to-Steam Heat Transfer ...ttt et s 3

Comparative Costs of Several Systems ... 4

Experimental Reactor Calaulations .......ccieciiiecienierei et s 6

NUCTEAr CalEUlHOoNS c.oo ottt e e et eee e easeae e st s e s eaneeeebe kb e e aas e bbb bt iasera e e s 6
Nuclear Characteristics of Homogeneous, Two-Region, Molten-Fluoride-Salt

Reactors Fueled with UZ33 ettt es st b b st e 6

Blanket Fissions in the Reference Design Reactor ... 9

Comparison of Ocusol and Cornpone Caleulations ... 10

Heterogeneous ReaCIOrs ....o.coiviiieiiiiiiii ettt 11

Gamma-Ray Heating Calculations ... 11

1.2. COMPONENT DEVELOPMENT AND TESTING oottt 19

Fuel PUmp Development ...ttt bttt e bbbt 19

Development Tests of Salt-Lubricated Bearings......coocoveooiiiiiic 19

Development Tests of Conventional Bearings.......ccoiiniimiiiin i, 20

MEChANTCAI SEAIS ettt ettt et et esr e ettt er e ae s b rbs b s e ab s R e e e e et e e e et 21

Radiation-Resistant Electric Motors for Use at High Temperatures ..o 22

Design Studies of Fuel PUMPS .. et 22

. Frozen-Lead Pump Seal ..ot 23

Development of Techniques for Remote Maintenance of the Reactor System ... 25

Mechanical Joint Development ... i s s 25

i Evaluation of Expansion Joints for Molten-Salt Reactor Systems ..o, 32

Remote Maintenance Demonstration Facility....cociiiii e, 35

Heat Exchanger Development ... ..ot 37

ALCl, Thermal-Convection LOoop ... rineissiinis it e 37

Moflten-Solf Heat Transfer Coefficient Measurement.....o.coiiiriiiiiiiii e 38

Design, Construction, and Operation of Materials Testing Loops .o 39

Forced-Circulation LOOPS . oiiieieireeie ettt b e e e 39

[oPi 1@ LLOOP'S «eveeareuietiseesree st ceetsi e eeeeme bbb s h bbb bbb SRS 42

1.3. ENGINEERING RESEARGCH ...ooctocee ettt ettt ieteetec st b e st st en bbb st 44

Physical Property Measurements ... ..o i 44

- Enthalpy and Heat Capacity et st 44

S UPFGCE T NS TON otitvtereresssseeeeee e e eeeets et sasbasssaas samseanseeases b et b e s b e e e b e e R e s e s na e Ls e b4 e ab e e e e e be s sa e eas b s b e s ne b0 45

Apparatus Fabrication and Calibration ... 46

Molten-Salt Heat-Transfer SHUIES ..o ireierert et ettt s s eb e b et et s s s 46

1.4, INSTRUMENTATION AND CONTROLS oottt e 47

Resistance-Type Continuous Fuel Level Indicator ..o 47

1.5. ADVANCED REACTOR DESIGN STUDIES oottt et srae s ssaessssenssesmnesenesevesanssnnessnn 48
Low-Enrichment Graphite-Moderated Molten-Salt Reactors ......cccvemieiiviiniiniei i 48
A Natural-Convection Molten-Salt ReQCtor ....ciiiciii ittt et besses e s st s samae s s 48

2.1.

2.2,

2.3.

MET ALLURGY ottt ettt ta et et e sttt e st s evtaeee s eteenbaeebeaasasseenssenseaseeseesseassebaeereensansbennnenten 53
Dynamic Corrosion StUAIies ..ttt et as b e ers e snrereesre et e snnanbes 53
Inconel Thermal-Convection Loop Tests ..ottt eer e e 53
INOR-8 Thermal-Convection Loop Tests ..t cere et e s re s sas et e annas 53
Inconel Forced-Circulation Loop T @SS oot cciicccctrievctes e ceres e veevsrsneesstessnneessreessneessnnnssnsres 53
Results of Examination of Samples Removed from INOR-8 Forced-Circulation
00D D354 T ettt st ets e et s b e ra e e e eatee e abeaearbbeareaenba e reneenteeabeaeanbebeas 56
MOtErIAl [NSPECTION eiiieiiiiieeicieie s et e et e rers e e s s nr e res e e s s ane e s e te e e s resaenbeseea ssnanaesareessernenerranteens 57
General Corrosion StUdI@s.......ooiiici it ce e s e e v e s e sbaea s er s e st e et e e e e e e teebsenraennrn 57
Carburization of Inconel and INOR-8 by Sodium-Graphite Systems ......ccccooeeiinvccrviinnnicineenne, 57
Carburization of INOR-8 by Fuel 130—Graphite Systems......coooiiiiiiiiiiiee e 59
UO, Precipitation in Fused Fluoride Salts in Contact with Graphite ........ccooniviiiiiinincn, 61
Precious-Metal-Base Brazing Alloys in Fuel 130 .o 64
Mechanical Properties of INOR-=8. ...t et et et eesser e s are s ne e 64
Influence of Composition on Properties of INOR-8 ..o 65
High-Temperature Stability of INOR-8......c..cveieiiicie e 67
Welding and Brazing Studies. ...ttt 68
INOR-8 Weldability ...ciciieieereierireieicee sttt s sim e e st bttt s b s ma s sbs s ba st 68
Brazing of Thick Tube Sheets for Heat Exchangers. ..., 71
Fabrication of Pump Components .......c.ueiiiier ittt sa bbb s 73
RADIATION DAMAGE ..ottt et e se st se e e st ree e e et e s e st es e et e e ete et e st e be e es s ese e s e e b e aabesn e e s n st a s s sanis 76
INOR-8 Thermal-Convection Loop for Operation in the LITR .. 76
Graphite Capsules Irradiated in the MTR ... e 76
CHEMI S T R Y oottt ettt e te sttt e b e eabeeteabs £ ert e aas e et e easear et e ane e sre s e s e e eamn s s ne st ene e e e st e st b e st b s sane b s nanass 78
Phase Equilibrium SHudies ..ooooiiiiiiiiiiiii s e s 78
Systems Containing UF4 and/or ThF4 ............................................................................................ 78
Solubility of PuF, in LiF-BeF, Mixtures ... s 80
FiSSiON-Product BehaVior ..ot ettt et e e ettt e eebe et s bbb s bbb er s e b ases pene 83
Solubility of Noble Gases in Molten Fluoride Mixtures ... 83
Solubility of HF in LiF-BeF, Mixtures ..., 85
Solubilities of Fission-Product Fluorides in Molten Alkali Fluoride—Beryllium
FIUOEIE SOIVENES oot eti e et e et s ete b en st e rae e s e et e bt ereesaesas e st s s e s ebe ot e s bn bt e beere peas 87
A Simple Method for the Estimation of Noble Gas Solubilities in Molten
FlUOTIAE MIXPUTES ooneeeeeeeiieieeiiieseei st e st e v e st e e e taeas e s e teesse sebe s eas s eatesaresamaan b e s reeabas s ssatsrnssrnnesanssnesns %0
Chemical Reactions of Oxides with Fluorides in LiF-BeF ..o 92
Chemistry of the Corrosion Process ... s 93
Activities in Metal Aoy s .o e 93
Surface Tensions of BeF , Mixtures ..., 94
2.4.

Disproportionation of Chromous FIUoride..........uuviiiiiieiiieioiiie e e 94
Effect of Fuel Composition on the Equilibria 3CrF2\=‘ 2CrF, + Cr° and

3FeF2: 2FeF 5 4+ Fel s 95
Activity Coefficients of CrF, in NaF-ZrF 4o, 96
Equilibrium Amounts of CrF, in Molten Salts Containing UF; Which are in

Contact With INConel ... ettt et eeene e 98
Chromium Diffusion in ATIoys ..ot et ettt s b eraas s b eas s e 99

Corrosion of Metals by Aluminum Chloride........occooviiiiiiiiiiiie et 100
Corrosion of Inconel by Aluminum Chloride Yapor in a Fused Silica Container ........ccoe...... 100
Corrosion of Inconel by Aluminum Chloride Vapor in an All-Metal System.......cccccvvveecivvvnnnnene, 100
Significance of Experimental Results ..o 101
Theoretical Considerations of Aluminum Chloride Vapor Corrosion .......ccccoveeeieerivicecniecrennnen, 101
Structural Metal Considerations ... s e s s e e e s e snerenes 102

Gaseous Aluminum Chloride as a Heat Exchange Medium .....ccooecviicriiniiiiiceeec e 103

Permeability of Graphite by Molten Fluoride Salts ..o 107

Preparation of Purified Materials........c.cciiiiiiiiiiiiiie st se e sresree s 107
Preparation of Pure Fluoride Compounds.......coceieiieriiiiiniiiccnnneneesrcnce e e e 107
Production-Scale Operations .......cciieciioriioiiiciiereiesctesse s ese s e se e s sar e s e ensseacsaessesennes 108
Experimental-Scale Operations ...ttt e et 108

FUEL PROCESSING ..ottt e stebs e ettt er et sb e et ets e b st enenb s esteseen e nae e e nseenesseenebeesns 110

Flowsheet for Fluoride Volatility and HF Dissolution Processing of

Molten-Salt Reactor Fluids .ocoiviieriieeee e ettt e s aa s a e er et 110
Experimental Studies of Volatility Step ...ovvieiiiiiiineciec e 110
Solubilities of LiF-BeF, Salts in Aqueous HF ..oy 112

x1i

Part 1
REACTOR DESIGN STUDIES

1.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS

H. G. MacPherson

Reactor Projects Division

INTERIM DESIGN REACTOR

A study has been made of improvements and
variations of the Interim Reactor Design described
in ORNL-2634 and, briefly, in the previous quar-
terly report.! Specifically, as described below,
an improved fill-and-drain system has been de-
signed, and alternate ways of transferring heat
from the reactor fuel to steam have been con-

sidered.

Fuel Fill-and-Drain System
G. D. Whitman

Design studies of the fuel fill-and-drain system
for the interim design reactor were described pre-
viously, "2 and further study of possible configu-
rations has resulted in a system that eliminates
maintenance difficulties inherent in the initial
design. The original concept of pipes and water
walls resulted in a relatively low first cost sys-
tem; however, the components were not readily
accessible for repair. In addition the entire
system was contained in a gastight thermally in-
sulated volume that complicated primary entry for
maintenance operations.

The new concept retains the after-heat removal
scheme, that is, radiant heat transfer to a boiling
water system, but the components have been re-
designed so that overhead accessibility makes
remote repair and removal more practical. The
fused salt is contained in a cylindrical tank into
which a number of vertical bayonet tubes are in-
serted. These tubes are capped at the bottom and
welded into the top of the vessel, which forms a
tube sheet. These tubes serve as the primary
after-heat removal radiating surface and contribute
substantial nuclear poison to the geometry.

The water boiler, which consists of a number of
mating double-pipe bayonet tubes projecting down-
ward from a combination water-and-steam drum,
rests on the top of the drain vessel. The boiler
tubes fit inside the tubes in the fuel vessel, and

1H. G. MacPherson, Molten-Salt Reactor Program
Status Report, ORNL-2634 (Nov. 12, 1958), and MSR
Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 3.

26, D. Whitman, MSR Quar. Prog. Rep. Jan. 31, 1958,
ORNL-2474, p 15.

the combination forms a radiant heat exchange
system. This concept retains the double con-
tingency protection ogainst fluid leakage from
either system,

Electric heaters and insulation are installed
around the fuel vessel for preheating. The boiler
tubes are flooded with water for after-heat removal,
and the rate of heat removal may be controlled by
the boiler ‘‘water rate’” or by dividing the boiler
into sections which may be used in combinations
to control the rate and locations of heat removal
in the fuel vessel.

The entire boiler system may be removed or
inserted from overhead without disturbing the fuel
circuitry. A failed fuel tube could plug at the
tube sheet and only failure of the vessel walls
would necessitate complete removal of the system.

The drain system required for a 600-Mw (thermal)
molten-salt reactor system was considered, and it
was found that four vessels 5 ft in diameter and
12 ft high would be needed to contain the 600 ft3
of salt from the reactor. Each of the vessels
would be capable of removing 1.8 Mw of after heat
within the practical temperature limitations of the
container materials.

Criticality calculations were made at off-design
conditions. The nuclear poison contribution of the
removable boiler tubes was neglected and the
bayonets projecting into the fuel tanks were con-
sidered to be flooded with water. It was de-
termined that approximately 60 lb of 1.5% boron
steel would have to be placed in the system to
give multiplication constants below 0.5 with the
probable U239 concentrations in the fuel salt.

Fuel-to-Steam Heat Transfer
M. E. Lackey

A direct fuel-to-steam heat transfer system has
been studied for application in a molten-salt
power reactor. The complete heat removal sys-
tem consists of four circuits in parallel to remove
the heat from the fuel and a single circuit to re-
move the heat from the blanket. Each of the cir-
cuits is separate and independent up to the point
where the superheated steam paths join ahead of
the turbine. Each circuit has a salt-to-steam heat

MOLTEN-SALT REACTOR PROGRESS REPORT

exchanger located in the reactor cell and an evap-
orator and steam compressor located adjacent to
the reactor cell.

A simplified flow diagram of the heat transfer
system is shown in Fig. 1.1.1. This system is
entirely free of sodium and its associated prob-
lems. The fuel-to-steam heat couple in conjunc-
tion with the Loeffler boiler cycle provides a
differential modulating thermal block between the
fuel and the water which allows complete and
conventional control of the reactor and load re-
lationship.

Comparative Costs of Several Systems
G. D. Whitman

Preliminary cost studies have been made for a
number of heat transfer cycles for a molten-salt
power reactor, and the capital costs for seven
heat transfer systems are summarized in Table
1.1.1. In all cases the system was assumed to
include a fixed-size, two-region, homogeneous
reactor in which 640 Mw of heat was generated
and then transferred to a reheat-regenerative
steam cycle, The steam conditions were set at

1800 psia and 1000°F reheat.

In all cases four parallel heat transfer loops
and four steam generating systems were coupled
to the fuel circuit, and a single heat transfer loop
and steam generator were used in the blanket
circuit. Reheat was supplied by the fuel system
exclusively in all the systems.

The reactor plant portion of the cost summary
included the shielding, containment vessel, in-
strumentation, remote maintenance equipment,
auxiliary fluid systems, original fluid inventories,
and the chemical plant equipment, in addition to
the reactor and heat transfer circuitry. The con-
ventional plant costs included the costs of land,
structures, steam system (not including boilers,
superheaters, or reheaters), turbine-generator,
accessory electrical equipment, and miscellaneous
power plant equipment.

The general expense item listed in Table 1.1.1
was assigned to cover the design and indirect
costs incurred during the construction and startup
phases of operation. It was set at approximately
16.5% of the reactor and conventional plant sub-
total.

The heat transfer linkage used in each system is
indicated, and in order to compare the costs more

Table 1.1.1. Molten-Salt Recctor Costs

Net electrical output: 260 Mw

0.80

Plant load factor:

Interest rate on capital investment:

14% per annum

Reactor with

Reactor with

Reference Interim Reactor with Natural-
Design Reactor Design Reactor Loeffler System Loefter System Loetfler System Gas-Cooled Convection Reactor
(Na-Steam) (Fuel-Steam) Reactar
{Na-Na-Steam) (No-Steam) (No-Steam) {Steam-Steam (Steam-Steam {He at 300 psia) (Na-Na-Steam)
{Na-Na-H,0) (Na-Na-H,0) (Na-5team Reheat} Reheat) Reheat) (Na-Na-H ,0)
Reactor plant $20,232,000 $19,600,000 $17,432,000 $17,100,000 $15,688,000 $26,234,000 $24,500,000
Contingency (40%) 8,093,000 7,840,000 6,973,000 6,840,000 6,275,000 10,493,000 9,800,000
Subtotal 28,325,000 27,440,000 24,405,000 23,940,000 21,963,000 36,727,000 34,300,000
Conventional plant 29,350,000 29,350,000 31,200,000 31,850,000 32,400,000 29,350,000 28,250,000
Contingency (7,5%) 2,201,000 2,201,000 2,340,000 2,389,000 2,430,000 2,201,000 2,119,000
Subtotal 31,551,000 31,551,000 33,540,000 34,239,000 34,830,000 31,551,000 30,369,000
Reactor and conventional 59,876,000 58,991,000 57,945,000 58,179,000 56,793,000 68,278,000 64,669,000
plant subtotal
General expense 9,950,000 9,800,000 9,625,000 9,650,000 9, 440,000 11,350,000 10,750,000
(including design)
Total cost $69,826,000 $58,791,000 $67,570,000 $68, 629,000 $645, 233,000 $79,628,000 $75,419,000
Fuel volume, ft? 575 555 589 589 777 766 1765
Fixed cost contribution 5.37 529 5,20 529 5.09 6.13 5.80

ta power cast, mills/kwhr

BLANKET
@—»

1210°F
23 psia

REACTOR

BLANKET

UNCLASSIFIED
ORNL-L.LR-DWG 34504

408,100 Ib/hr

Fig. 1.1.1. Flow Diogrom of Fuel-to-Steam Heat Transfer System For Interim Design Reactor.

EMERGENCY
RELIEF
LEAK-DETECTION
SYSTEM
[}
—100 psia
FUEL FROM FUEL
[ ' AND BLANKET TURBINE STOP
RN ’ B HIGH -PRESSURE
|_83 osia ATTEMPERATOR 408,1001b/hr - TURBINE
. 1805 psia 260psia
78 psio— 333,000 1b/hr E‘
T iINTERMEDIATE-
p Q22°F AND LOW-PRESSURE
t < 240 psia TURBINES
SUPERHEATER , REHEATER =
| = %
1
COOLING
: WATER
621°F 881°F RSETT’TC'NG CONDENSER
40 psia—1 — 1854 psia 1800 psia — ION
1,492,000 Ib/hr 1,083,900 Ib/hr DESUPERHEATER CONDENSATE
PUMP
\j N d
Lt J
1050°F
13.5 cfs
D ——-
STEAM PUMP M
ENTRAINMENT SEPARATOR
- FEEDWATER
~ 17\ —- HEATERS
IN SERIES
[ [
BOILER }
1
P F— DEAERATOR
(OF 3 PER FUEL CIRCUIT i BLANKET
FUEL FEEDWATER
.l PUMP
475°F FEEDWATER

HEATERS IN SERIES

8561 ‘L€ ¥390L20 9NIGNI dOl¥3d
MOL TEN-SALT REACTOR PROGRESS REPORT

easily certain design features were fixed, as pre-
viously described. More nearly optimized systems
would probably result in lower costs; however,
with the exception of the gas-cooled cycle, it was
not evident that optimization studies would sig-
nificantly alter the relative costs.

The Loeffler direct fuel-to-steam system gave
the lowest first cost. In this case the elimination
of a sodium circuit resulted in a net saving even
though the fuel-to-steam heat exchanger was more
expensive than the heat exchangers in the sodium-
coupled system.

The helium-cooled cycle resulted in the highest
first cost, This was due primarily to the relatively
large amount of heat transfer surface required in
the fuel-to-helium and helium-to-water-to-steam
heat exchangers. This particular system would
probably be the most sensitive to optimization
studies. In particular, increasing the coolant gas
pressure and using gases other than helium could
reduce the size of the heat transfer equipment.

Differences in fuel volume will also be reflected
in fuel cycle costs, which are not included in
the capital costs shown in Table 1.1.1. A first
approximation, based on changes in fuel inventory
charges alone, indicate that each additional 100
ft3 of fuel volume would add about 0.06 mill/kwhr
to the cost. On the basis that each plant had the
same net electrical output, the maximum deviation
in the power cost contributed by the capital charge
was found to be approximately 1 mill/kwhr.

EXPERIMENTAL REACTOR CALCULATIONS
J. W. Miller

A 30-Mw experimental reactor is being con-
sidered that would have a 6-ft-dia core enclosed
in a ]/Z-in.-fhick INOR-8 pressure vessel. Sur-
rounding the core there would be a 6-in.-thick
layer of insulation, a 6-in.-thick air space, and
five 2«in,-thick boron steel thermal shields cooled
by 4-in.-thick layers of diphenyl.

Nuclear calculations indicate that the critical
mass at 1180°F would be 49 kg of U233, and the
total inventory for a system volume of 150 43
would be 65 kg of U235, After a year of operation,
the critical mass would have increased to 61 kg,
with a net integrated burnup of 14.4 kg.

Approximately 7% of the neutrons would be
captured in the core vessel. This compares with
5% for the 600-Mw reference design reactor.
Weighting these numbers by the reactor power

levels, indicates that the heat-generation rate in

the experimental reactor would be approximately

/% of the heat-generation rate in the reference

design reactor. )
Approximately 31% of the neutrons would escape

from the core vessel, and these would ail be cap-

tured by the time they reached the fourth boron

steel thermal shield.

NUCLEAR CALCULATIONS
L. G. Alexander

Nuclear Characteristics of Homogeneous,
Two-Region, Molten-Fluoride-Salt
Reactors Fueled with U233

The calculations entailed in the investigation of
initial states of U233.fueled reactors were com-
pleted, and it was discovered that serious dis-
crepancies existed between results obtained from
the Oracle program Cornpone and the UNIVAC
program QOcusol, especially when the results were
used as input for the Oracle burnout code, Sorghum.
The difficulty was traced to fluctuating end-point
fluxes generated by the Cornpone program, Ac-
cordingly, the Cornpone cases were discarded,
and the series of calculations was completed with
the Ocusol program. Meanwhile, the defect in the
Cornpone program was remedied by the author,

W. E. Kinney, and the two programs are now in
fair agreement, as discussed below.

The results of the calculations are given in -
Table 1.1.2, where critical mass, inventory, re-
generation ratio, etc., are given for a series of
cases covering core diameters ranging from 4 to
12 ft and thorium fluoride concentrations in the
core of 2, 4, and 7 mole %. The regeneration ratio
is plotted against the critical inventory for a
600-Mw (thermal) system in Fig. 1.1.2. The dotted
line in Fig. 1.1.2 is the envelope of the curves
shown and is the locus of optimum conversion for
a given inventory. As may be seen, the 8-ft-dia
cores are nearly optimum at all thorium concen-
trations, and initial regeneration ratios of up to .
1.08 can be obtained with inventories ranging up
to 1300 kg of U233,

The subsequent performance of the 8-ft-dia core
with 7 mole % thorium was studied by means of
the Oracle code Sorghum. Extracts from the cal-
culated results are given in Table 1.1.3, where
inventories, absorption ratios, regeneration ratios,
etc., are given for periods of operation of up to

PERIOD ENDING OCTOBER 31, 1958

Table 1.1.2, Initial Nuclear Characteristics of Two-Region, Homogeneous, Molten-Fluoride-Salt Reactors Fueled with U233

Fuel salt: 37 mole % BeF, + 63 mole % LiF + UF,+ ThF,
Blanket salt: 13 mole % ThF, + 16 mole % BeF, + 71 mole % LiF
Total power: 600 Mw (heat)

External fuel volume; 339 3

Case No, 51 52 53 54 55 56 57 58 59 60 61 62 63 64 85
Core diameter, ft 4 4 4 6 6 6 8 8 8 10 10 10 12 12 12
ThF , in fuel salt, mole % 2 4 7 2 4 7 2 4 7 2 4 7 2 4 7
U233 in fuel salt, mole % 0.619 0.856 1.247 0.236 0,450 0.762 0.152 0.316 0.603 0.121 0.262 0.528 0.101 0.222 0.477

U233 51om density, atoms/em>  19.8%10'7 27.4x 10'7 39.9% 10" 7.55x 10"7 14.4x10'7 24.4% 10" 4.88 10" 10.1x10"7 19.3x 10" 3.8 % 10" 8.39%1017 16.9x10'"7 3.24x10"7 7.39x 10" 1525% 10'7
Critical mass, kg of U233 72.0 100.5 146.5 94.2 177.8 301 143 299 566 221 481 970 320.5 733 1510
Critical inventory, kg of U233 810 1118 1630 374 7N 1204 325 677 1283 364 793 1600 441 1007 2078

Neutron absorption ratios*

U233 (fissions) 0.871 0.874 0.881 0.864 0.868 0.876 0.867 0.865 0.873 0.87 0.864 0.871 0.873 0.864 0.870
U233 (n,) 0.129 0.126 0.119 0.136 0.132 0.124 0.133 0.135 0.127 0,130 0.136 0.129 0.127 0.136 0.130
Be-LiF in fuel salt 0.070 0,066 0,069 0.093 0,075 0.076 0.120 0.082 0.078 0.142 0.088 0.081 0.164 0.093 0.083

Core vessel

Be-Li-F in blanket salt 0.129 0.111 0.095 0.101 0.080 0.062 0.086 0.059 0.043 0.062 0.044 0.031 0.049 0.032 0.020
Leakage
Th in fuel salt 0.343 0.426 0.517 0.581 0.650 0.740 0.716 0.785 0.865 0.800 0.865 0.938 0.872 0.922 0.998
Th in blanket salt 0.653 0.600 0.538 0.403 0.382 0.330 0.264 0.254 0.213 0.189 0.180 0.146 0.115 0.130 0.092
Neutron yield, 7 2.20 2.20 2.22 2.18 2,19 2,21 2,19 2.18 2,20 2.19 2.18 2.20 2.20 2.18 2.19
Regeneration ratio 0.996 1.026 1.055 0.984 1.032 1.070 0.980 1.039 1.078 0.989 1.045 1.084 0.987 1,052 1.090
*Neutrons absorbed per neutron absorbed in U233.

|
4

Fuel: U233

Core diameter: 8 ft
ThF4 in core: 7 mole %
Power: 600 Mw (thermal)

Lood factor: 0.8
External volume: 339 503

Chemical processing rate: 1.66 ft3/day for core

1.97 ¥ /day for blanket

Table 1.1.3. Long-Term Nuclear Performance of o Two-Region, Homogeneous, Spherical, Molten-Fluoride-Salt Reactor

Initial State After 1 Yeor Afrer 2 Years After 5 Years After 10 Yeors Aftrer 20 Yeors
Inventory Absorption Inventory Absorption Inventory Absorption Inventory Absorption Inventory Absorption laventory Absorption
(kg) Ratio (ka) Ratio (kq) Ratio (kg) Ratio (kg) Ratie (kg) Ratio
Core elements
Th232 14,527.1 0.8652 14,527 0.8416 14,527 0.8343 14,527 0.8237 14,527 0.8086 14,527 0.7892
Pq?33 0 0 18.8 0,00692 18.43 0,00671 18.38 0.00642 18.11 0.00605 17.67 0,00562
Y233 1,283.28 1,0000 1,363 0.9998 1,388.3 0.9994 1,420.7 0,996 1,453.03 0.9876 1,500.02 0.9721
Fissions 0.8730 0.8740 0,8741 0.8721 0.8648 0.8519
n,y 0.1270 0.1258 0.1253 0.1245 0.1228 0.1202
234 0 0 27.32 0,00348 53,91 0,00672 129.3 0.015 242.05 0,0305 428.7 0,0472
y3s 0 0 0.353 1.73 « 104 1.32  6.37x107* 717 0.00339 23.17 0.01235 63.07 0.0279
Fissions 0 123« 10~* 4,53 x 10~* 0.00241 0,00880 0.0199
n,y 0 0.50 » 104 1.84 » 10-4 0,00098 0.00355 0,0080
Y23 0 0 0.00347 6.39 < 10~7 0.0252 4.50 = 10~* 0.343 5.89%10"% 2.206 0,000458 11.73 0.00178
Np237 0 0 274105 9.76x 10" 3.23410~* 113x10"7 0.00725 2.46 x 10-% 0.58 2.40 = 10~3 0,321 9.94x10°3
TEAL 0 0 412107 101100 9.44,10~% 225.107 565x10°* 1,30~ 1077 0,00922 2.863 x 10~¢ 0.104 2.711x10"%
py23? 0 0 3.46 <107 1.98x10-'? 1.49%10°7 B37x10""" 2.29-10"% 1.25«10% 0.000732 5.952 = 10~7 0.0148 7,36 x 10~
Fissions 0 1.22% 1072 515« 10="" 0.77 » 108 3,687 » 10”7 4.58 x 10~¢
",y 0 0.76 » 10~'2 3.22x 107" 0.48 » 10~® 2,255 < 10~7 2.78 x 10~8
Fission products 0 0 115,2 0,0165 158.2 0.0221 180,5 0,0244 181.5 0.0234 181.5 0.0221
Li 3,699.12 0.00522 36,991 0.00488 3,699 0,00478 3,699 0.00455 3,699.1 0.00449 3,699.1 0.,00429
Be 2,739.23 0.00907 2,739 0.00906 2,739 0.00906 2,739 0,00905 2,739.2 0.00903 2,739,2 0.00900
F 26,330.9  0.06433 26,331 0.0643 26,331 0.0642 26,331 0,0642 26,331 0.0640 26,331 0,0638
Blanket
Th232 30,821.6 0,2132 30,822 0,2132 30,822 0.2131 30,822 0.2128 30,822 0.212 30,822 0.211
Pg23 0 4,796 4,79 4.78 4,771 4.752
y233 0 26.26 38.15 44.36 44.53 44,35
Other 0.0428 0.0428 0,0428 0.0428 0.0426 0,0423
Total inventory, kg of 1,283.28 1,389,6 1,427.8 1,472.2 1,520.73 1,607.5
fissionable isotopes
Fission yield, 5 2.20 2.203 2.204 2.204 2,202 2.196
Fuel feed rate, ka/year 118 27.4 6.0 -0.1 -0.8
Net integrated fuel 0 11.46 2.5 -6.42 -75.87 -167.36
bumup, kg
Regeneration ratio 1.079 1.051 1,047 1.046 1.045 1.042

LY0dI Y SS3IYO0AJ 4OLOVIY LTVS"NILTIOW

UNCLASSIFIED
ORNL-LR-DWG 34505

140

7 mole%. Th,

8 ’/T
10 4 mole% ThF,

4 4

/

/
20412 4

|
2T 2 mole% ThE,
° |

REGENERATION RATIO
?;

!
" FUEL SALT: 37 mole%s BeF, + 63 mole7s LiF +
, UF, + ThE,

EXTERNAL FUELVOLUME: 339 fTS

|

! |

e ;

0.90 g §'mole ThE, IN FUEL SALT ' ‘ |
NUMBERS ON CURVES ARE

I
8
‘ CORE DIAMETERS IN FEET
I L H 1 1
2

~

4 6 8 10 12 44 16 18 20 (X109
TOTAL INVENTORY OF FISSIONABLE ISOTOPES {kg)

0

Fig. 1.1.2. Initial Regeneration Ratio in Two-Region,
Homogeneous, Molten-Fluoride-Salt Reactors Fueled

with U233.

20 years. The regeneration ratio, total inventory
(including U233 in the blanket), annual feed rate,
and net integrated burnup are plotted in Fig. 1.1.3.
It may be seen that, although the regeneration ratio
remains above 1.04 and therefore there is a net
breeding gain, it is, nevertheless, necessary to
add fresh U233 during the first five years to com-
pensate for the ingrowth of fission products and
nonfissionable isotopes. The annual additions
amount to 118, 27.4, 10.8, 6.0, and 1.6 kg, re-
spectively, but less than 1.0 kg per year is
required after five years, and the reactor is then
self-sustaining. By the end of the second year
the net integrated burnup (total inventory less
total fuel purchased) becomes negative.

Blanket Fissions in the Reference Design Reactor

The burnout code, Sorghum, used in this study
does not take into account fissions occurring in
blanket. In order to estimate these, a series of
sequential Cornpone and Sorghum calculations was
performed, in which the concentrations of the 16
elements in the reactor predicted by one-year
Sorghum calculations were submitted periodically
to a criticality test by the Cornpone program. The
outputs, consisting of fraction of fissions in the

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
ORNL -LR-DWG 34506

1.08
o
q
< 106
=
O
= 1.04
Ll
o
4
Z 102
(U]
&
£.00
1600 TOTAL INVENTORY [
- L INY I S
w E _.——-'._—-—-——'_'-_-—
7= 1400 !
[%2]
< 7 /
% o
=0 1200
0o
e TOTAL POWER: 600 Mw ( HEAT)

1000 '— (0AD FACTOR: 0.8
EXTERNAL FUEL VOLUME: 339 1t3
CORE AND BLANKET PROCESSING RATE:ONCE PER YEAR
150 — CORE DIAMETER: 8 f1

[ O S O
—~ 100 I T i 1 :
g =1 AVERAGE FEED RATE, kg OF UZ33/year
@ 50
a
o 1
5 0
% T~
J -50 ~
m \
; >\
© 100 |— NET INTEGRATED ~ .
& BURNUP, kg — | S~
Y 450 ~
200

0 2 4 6 8 {0 12 44 & 48 20
TIME OF OPERATION (years)

Fig. 1.1.3. Nuclear Performance of Two-Region,
Homogeneous, Molten-Fluoride-Salt Reactor Fueled

with U233.

blanket, together with space-averaged group fluxes
and leakage probabilities, were then used as in-
puts for further one-year Sorghum calculations.
The neutron source in Sorghum was also reduced
by the fraction of fissions occurring in the blanket.
In addition to providing information about the
fissions in the blanket, this series of calculations
also provided a test of the usual method of operat-
ing the code. The case selected for study was
the Interim Design Reactor with a core 8 ft in
diameter, a fuel salt consisting of 36.5 mole %
BeF,, 62.5 mole % LiF, and 1 mole % ThF ,, to-
gether with sufficient UF , to make the system
critical (about 0.25 mole %). The case was run
as Cornpone calculation 01045 and the resulting
data were used as input for Sorghum calculation
01065. In the integration of the time-dependent
difference equations, a time increment of 5 days

MOLTEN-SALT REACTOR PROGRESS REPORT

was used until the time variable corresponded to
one year of operation of the reactor; at this time,
the increment was increased to 30 days for the
next 20 years of reactor time. This is customarily
done to shorten the time required for the compu-
tation and is justified by the fact that, after the
first year, concentrations change siowly.

A second series of calculations was performed
as described above, except that the time incre-
ment was held constant at 5 days. After one year
of reactor operation, the concentrations predicted
by the Sorghum code were found to render the
reactor slightly subcritical (£ = 0.9982 on Corn-
pone 02045), even though the U233 in the blanket
was faken into account. The fraction of fissions
in the blanket was estimated to be 0.013. This
information, together with flux spectrum and leak-
age probabilities, was used as input for further
Sorghum calculations, etc., with the results shown
in Fig. 1.1.4, where the fraction of blanket fissions
is shown as a function of reactor operating time.
Also shown is a comparison of two key parameters,

[}

g UNCLASSIFIED
= ORNL—LR—DWG 34507
D o002 , :

v W | l ! !
b= ! ‘ i
O o ; |

= !

S @ l

—

(51) Z L !

08

[V

o7 | | { | [
CORNPONE BASE, SECOND SERIES E
[ ! ‘ [

/CORNPONE BASE, FIRST SERIES
- R [ B [
{ ey

= =— 0CUSOL BASE | }

| ! \ \ f

0.6

REGENERATION
RATIO

DIAMETER OF CORE: 8ft

0.5
1250 ThF, IN FUEL" {1 mole Do
EXTERNAL FUEL VOLUME: 338 ft3
uLI)J TOTAL POWER: 600 Mw (THERMAL)
% LOAD FACTOR: 0.8
> = !
O !
55 ] _
Z L, 1000 l —
& OCUSOL BASE _—F
; 2’ A== / —//
22 /
S & /
] | A
=0 CORNPONE BASE, FIRST SERIES
S rso 1\l L bbb —t
© ’ CORNPONE BASE, SECOND SERIES
g ]
|
500 ‘

Q 2 4 6 8 10 12 14 16 18 20
TIME {years)

Fig. 1.1.4. Blanket Fissions in Interim Design

Reactor.

10

critical inventory and regeneration ratio, com-
puted in the two series of calculations. The cal-
culations to date cover only the first five years of
operation; however, the fraction of fissions in the
blanket appears to have stabilized at about 0.02.
A comparison between the two lower curves for
critical inventory shows satisfactory agreement
between the two methods of calculation, especially
since the top points of the discontinuous curve
represent the critical condition. Similarly, the
low points of the discontinuous curve for regener-
ation ratio agree well with the smooth curve. How-
ever, it is not presently understood why the slopes
of the discontinuous curves differ so much from
the smooth curves, and it may be that the agree-
ment of the points is fortuitous.

Comparison of Ocusol and Cornpone Calculations

As mentioned above, a mathematical defect in
the Cornpone program was discovered, and hence
all prior Cornpone calculations were discarded.
After the error had been corrected, the program
was used to recompute an Ocusol case. A com-
parison of the results obtained from the two pro-
grams with identical input conditions is shown in
Table 1.1.4. it may be seen that, while there is
good agreement between the regeneration ratios,
there is a substantial difference between the
critical inventories and that the Cornpone program
predicts the lower values.

The discrepancies have been resolved in favor
of Cornpone, which embodies a fourth-order approx-
imation to the difference equations, whereas
Ocusol uses only second-order approximations that
result in an under-estimate of the absorptions in
the core. In order to achieve a neutron balance
in Ocusol, the currents at the boundary of the core
are adjusted arbitrarily and are not computed from
the gradients of the flux. As a result, the leakages
are overestimated. In Cornpone, the fourth-order
solutions satisfy the neutron balances to less than
one part in a thousand, and the boundary currents
are correctly computed from the gradients of the
flux.

The difference in the behavior of this reactor as
predicted by Sorghum calculations based on the
Ocusol and on the Cornpone solutions are shown
in Fig. 1.1.4. It may be seen that the differences
in regeneration ratio are negligible, but the cal-
culation based on Cornpone predicts a critical
inventory about 50 kg lower than that based on
Ocusol.
PERIOD ENDING OCTOBER 31, 1958

Table 1.1,4, Comparison of Ocusol and Cornpone Calculations for the Interim Design Reactor

Fuel: U235

Core diametar: 8 ft

ThF4 in core: 1 mole %
Power: 600 Mw (thermal)
Load factor: 0.8

External fuel volume: 339 £

Ocusol Cornpone
Inventory Absorption Inventory Absorption
(kg) Ratio (kg) Ratio
Core elements
Th232 2,121 0.364 2,121 0.378
u233 616 565
Fissions 0.729 0.734
ny 0.271 0.266
y238 45.9 0.039 46.7 0.042
Li 3,969 0.033 3,969 0.037
Be 3,044 0.010 3,044 0.010
F 24,291 0.058 24,291 0.047
Blanket
TH232 30,822 0.232 30,822 0.227
Other 0.063 0.071
Fission yield, 7 1.800 1.812
Regeneration ratio 0.635 0.647

Heterogeneous Reactors

As mentioned in previous reports, graphite-
moderated, heterogeneous, molten-fluoride-salt
reactors show promise of achieving higher regener-
ation ratios for a given investment in fuel than the
homogeneous reactors. As a result, attention has
been given to developing a reasonably reliable
method of calculation. The multigroup Cornpone
program for Oracle was modified by W. E. Kinney
to provide a boundary condition of zero net cur-
rent in each diffusion group. Thus it is now pos-
sible to treat a heterogeneous unit cell in an
infinite lattice and to compute group disadvantage
factors which may then be applied to the group
cross sections in a homogeneized version of the
finite reactor. A program modifying the Cornpone
program to apply these factors has been written,

Gamma-Ray Heating Calculations

Work on a program for computing the gamma-ray
heating in the core vessel and blanket of homo-
geneous, molten-salt reactors was resumed.
Gamma-ray energy absorption coefficients as a
function of energy for 24 elements were computed
from the x-ray attenuation coefficients listed by
Grodstein, 3 according to a procedure developed
previously.® The defining equation is:

Mo =tpe * [ppltpy + (T + [ duc o

3G. W. Grodstein, Natl. Bur. Standards (US) Circ. 583
(1957).

4L. G. Alexander, The Gamma Energy-Absorption
Coefficient, ORNL CF-56-8-219 (August 8, 1956).

11

MOLTEN-SALT REACTOR PROGRESS REPORT

where
i, = gamma-ray energy absorption coefficient,
Hpe = photoelectric absorption coefficient,
.= Compton scattering coefficient,

f. = fraction of energy of photon deposited
locally in a pair-production collision,

f_ = fraction of energy of photon deposited
locally in a Compton collision,

fo= fraction of energy of photon carried away
from Compton collision by photons having
energies less than E , the Heut-off”’
energy.

In these calculations /., was taken to be equal to
1.0; that is, the annihilation radiation was in-
cluded in the local heating effect. This is not
strictly consistent with the use of a low-energy

5L. G. Alexander, The Integral Spectrum Method for
Gamma-Heating Calculations in Nuclear Reactors, ORNL

CF-56-11-82 (Nov. 4, 1956}.

12

cut-off limit, E ), of 0.1 Mev for the Compton
process. The energy-absorption coefficients com-
puted here will be used, however, in an ‘‘integral
spectrum’’ calculation of the type described in
ORNL CF-56-11-82 (ref 5) in which the coefficient
is averaged over the photon spectrum which is not
known with sufficient accuracy to warrant taking
into account the annihilation radiation.

For a cut-off energy of 0.1 Mev, /_ is equal to
zero with negligible error for all energies down to
and including 0.15 Mev. The energy-deposition
coefficient for Compton scattering, f_, was taken
from the values listed in ref 4,

The calculated values are listed in Table 1.1.5,
and they have been assembled on paper tape for
use with the gamma-ray heating program Ghimsr
being written for the Oracle. A closed, three-
address subroutine has been written for the com-
putation of Storm’s attenuation function for spheri-
cally symmetric systems.®

SM. L. Storm et al., Gamma-Ray Absorption Distribu-
tions ..., KAPL-783 (July 24, 1952),

PERIOD ENDING OCTOBER 31, 1958

Table 1.1.5. Gamma-Ray Energy-Absorption Coefficients, i,

e . o -
E., (Mev) barns /atom cm?/g E., (Mev) barns /atom cm?/g
Hydrogen (1;2) Beryllium (4;9.02)
0.1 0.0620 0.0375 0.1 0.251 0.0167
0.2 0.0856 0.0514 0.2 0.342 0.0228
0.4 0.0982 0.0589 0.4 0.396 0.0264
0.6 0.0991 0.0595 0.6 0.396 0.0264
0.8 0.0970 0.0581 0.8 0.388 0.0258
1.0 0.0927 0.0556 1.0 0.372 0.0248
1.5 0.0850 0.0510 1.5 0.328 0.0219
2.0 0.0815 0.0589 2.0 0.328 0.0219
3.0 0.0669 0.0401 3.0 0.273 0.0182
4.0 0.0593 0.0356 4.0 0.247 0.0165
6.0 0.0491 0.02%4 6.0 0.213 0.0142
8.0 0.0425 0.0255 8.0 0.191 0.0128
10.0 0.0376 0.0226 10.0 0.177 0.0118
Carbon (6;12) Nitrogen (7;14)

0.1 0.391 0.0199 0.1 0.470 0.0202
0.2 0.512 0.0257 0.2 0.599 0.0258
0.4 0.589 0.0296 0.4 0.689 0.0296
0.6 0.594 0.0298 0.6 0.6%94 0.0298
0.8 0.582 0.0292 0.8 0.679 0.0292
1.0 0.557 0.0279 1.0 0.650 0.0280
1.5 0.512 0.0257 1.5 0.597 0.0257
2.0 0.493 0.0247 2.0 0.579 0.0249
3.0 0.416 0.0209 3.0 0.490 0.0211
4.0 0.381 0.0191 4.0 0.451 0.0194
6.0 0.334 0.0167 6.0 0.400 0.0172
8.0 0.306 0.0153 8.0 0.370 0.0159
10.0 0.390 0.0145 10.0 0.251 0.0151

13

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 1.1.5 (continued)

14

- . o .

E., (Mev) barns /atom em?/g E., (Mev) barns/atom em?/g
Oxygen (8;16) Sodium (11;23)
0.1 0.531 0.0200 0.1 1.04 0.0273
0.2 0.702 0.0264 0.2 1.031 0.0270
0.4 0.790 0.0297 0.4 1.119 0.0293
0.6 0.784 0.0294 0.6 1.098 0.0288
0.8 0.776 0.0292 0.8 1.067 0.8279
1.0 0.745 0.0280 1.0 1.021 0.0268
1.5 0.683 0.0257 1.5 0.940 0.0246
2.0 0.662 0.0249 2.0 0.895 0.0234
3.0 0.563 0.0212 3.0 0.791 0.0207
4.0 0.522 0.0196 4.0 0.745 0.0195
6.0 0.452 0.0170 6.0 0.658 0.0172
8.0 0.410 0.0154 8.0 0.608 0.0159
10.0 0.394 0.0148 10.0 0.599 0.0157
Magnesium (12;24.32) Aluminum (13;26.96)

0.1 1.280 0.0316 0.1 1.595 0.0356
0.2 1.085 0.0268 0.2 1.189 0.0265
0.4 1.185 0.0293 0.4 1.285 0.0287
0.6 1.188 0.0294 0.6 1.288 0.0287
0.8 1.162 0.0288 0.8 1.264 0.0282
1.0 IREN 0.0276 1.0 1.210 0.0270
1.5 1.025 0.0254 1.5 LN 0.0248
2.0 0.074 0.0248 2.0 1.088 0.0242
3.0 0.869 0.0215 3.0 1.031 0.0230
4.0 0.822 0.0203 4.0 0.901 0.0202
6.0 0.767 0.0190 6.0 0.848 0.0189
8.0 0.739 0.0182 8.0 0.822 0.0184
10.0 0.727 0.7180 10.0 0.714 0.0160

PERIOD ENDING OCTOBER 31, 1958

Table 1.1,5 (continued)

= . e .
E., (Mev) barns /atom em?/g E., (Mev) barns/atom em?/g
Silicon (14;28.06) Phosphorus (15;31,02)

0.1 1.986 0.0426 0.1 2.49 0.0485
0.2 1.209 0.0259 0.2 1.45 0.0282
0.4 1.392 0.0300 0.4 1.49 0.0290
0.6 1.383 0.0297 0.6 1.48 0.0288
0.8 1.360 0.0292 0.8 1.45 0.0282
1.0 1.303 0.0280 1.0 1.39 0.0270
1.5 1.198 0.0257 1.5 1.28 0.0249
2.0 1.173 0.0252 2.0 1.26 0.0275
3.0 1.030 0.0221 3.0 1.11 0.0216
4.0 0.984 0.0211 4.0 1,068 0.0208
6.0 0.933 0.0200 6.0 1.017 0.0198

8.0 0.912 0.0196 8.0 (1.007)
10.0 0.906 0.0195 10.0 1.022 0.0195

Sulfur (16;32.06) Argon (18;39.91)

0.1 3.10 0.0593 0.1 4.7 0.0709
0.2 1.60 0.0300 0.2 1.95 0.0294
0.4 1.60 0.0300 0.4 1.82 0.0274
0.6 1.60 0.0300 0.6 1.80 0.0272
0.8 1.56 0.0296 0.8 1.75 0.0264
1.0 1.49 0.0280 1.0 1.67 0.0252
1.5 1.37 0.0257 1.5 1.55 0.0234
2.0 1.34 0.0252 2.0 1.52 0.0229
3.0 1.19 0.0224 3.0 1.36 0.0205
4.0 1.15 0.0211 4.0 1.32 0.0199
6.0 1.10 0.0206 6.0 1.25 0.0189
8.0 1.10 0.0206 8.0 1.29 0.0294
10.0 1.10 0.0206 10.0 1.30 0.0196

15
MOL TEN-SALT REACTOR PROGRESS REPORT

Table 1.1.5 (continued)

e 2 o 2
E., (Mev) barns/atom em?/g E., (Mev) barns /atom em?/g -
‘ Potassium (19;39.1) Calcium (20;40.07) .
0.1 5.8 0.0895 0.1 7.26 0.1092
0.2 2.14 0.0330 0.2 2.38 0.0358
0.4 1.94 0.0299 0.4 2.05 0.0309
0.6 1.90 0.0293 0.6 2.01 0.0325
0.8 1.85 0.0285 0.8 1.95 0.0294
1.0 1.76 0.0272 1.0 .85 0.0278
1.5 1.63 0.0251 1.5 1.72 0.0259
2.0 1.61 0.0248 2.0 1.69 0.0255
3.0 1.44 0.0222 3.0 1.52 0.0229
4.0 1.41 0.0218 4.0 1.50 0.0226
6.0 1.39 0.0214 6.0 1.49 0.0224
8.0 1.41 0.0218 8.0 1.50 0.0226
10.0 1.43 0.0220 10.0 1.53 0.0230
Iron (26;55.84) Copper (29;63.57)
0.1 20.7 0.223 0.1 32.5 0.308 .
0.2 4.45 0.0479 0.2 6.2 0.0287
0.4 2.84 0.0206 0.4 3.33 0.0316
0.6 2.67 0.0288 0.6 3.03 0.0287
0.8 2.57 0.0277 0.8 2.90 0.0275
1.0 2.44 0.0263 1.0 2.74 0.0260
1.5 2.24 0.0242 1.5 2.50 0.0237
2.0 2.23 0.0240 2.0 2.52 0.0239 -
3.0 2.07 0.0223 3.0 2.45 0.0232
4.0 2.08 0.0224 4.0 2.40 0.0228 .
6.0 2.15 0.0232 6.0 2.52 0.0239
8.0 2.22 0.0240 8.0 2.63 0.0250
10.0 2.40 0.0258 10.0 2.73 0.0259

16

PERIOD ENDING OCTOBER 31, 1958

Table 1.1.5 (continued)

Photon Photon
Energy fe Energy He
E,, (Mev) barns/atom em?/g E., (Mev) barns /atom em?/g
Molybdenum (42;96.0) Tin (50;118.7)
0.1 147 0.930 0.1 289 1.46
0.2 22.3 0.140 0.2 43.6 0.221
0.4 6.9 0.0433 0.4 10.5 0.0534
0.6 5.04 0.0316 0.6 6.90 0.0350
0.8 4.53 0.0284 0.8 5.86 0.0297
1.0 4.20 0.0264 1.0 5.29 0.0268
1.5 3.81 0.0239 1.5 4.71 0.0239
2.0 3.86 0.0242 2.0 4.77 0.0242
3.0 3.86 0.0242 3.0 4.79 0.0242
4.0 3.99 0.0250 4.0 5.12 0.0260
6.0 4.39 0.0275 6.0 5.77 0.0292
8.0 4.71 0.0296 8.0  6.27 0.0318
10.0 5.03 0.0315 10.0 6.74 0.0341
lodine (53;126.93) Tungsten (74;184.0)
0.1 363 1.72 0.1 1254 4N
0.2 54 0.266 0.2 192 0.630
0.4 12.4 0.0589 0.4 37.0 0.121
0.6 7.7 0.0365 0.6 18.8 0.0616
0.8 6.4 0.0303 0.8 13.1 0.0430
1.0 5.75 0.0272 1.0 1.2 0.0368
1.5 5.08 0.0240 1.5 9.5 0.0311
2.0 5.15 0.0244 2.0 8.5 0.0278
3.0 5.21 0.0247 3.0 8.75 0.0286
4.0 5.60 0.0265 4.0 9.51 0.0312
6.0 6.31 0.0299 6.0 10.86 0.0356
8.0 6.90 0.0327 8.0 11.82 0.0388
10.0 7.44 0.0352 10.0 13.0 0.0426

17
MOLTEN-SALT REACTOR PROGRESS REPORT

Table 1,1,5 (continued)

Photon Photon
Energy Fe Energy Fe
E., (Mev) barns/atom em?/g E., (Mev) barns/atom em?/g
Platinum (78;195.2) Thallium (81;204.4)
0.1 1505 4.65 0.1 1715 5.04
0.2 232 0.716 0.2 268 0.789
0.4 44.8 0.139 0.4 51.5 0.152
0.6 21.6 0.0667 0.6 24.4 0.0719
0.8 15.2 0.0470 0.8 16.9 0.0497
1.0 12.1 0.0374 1.0 13.3 0.0392
1.5 9.5 0.0294 1.5 10.2 0.0301
2.0 9.3 0.0288 2.0 10.2 0.0301
3.0 9.6 0.0297 3.0 10.3 0.0304
4.0 10.4 0.0321 4.0 1.0 0.0324
6.0 10.8 0.0333 6.0 12.6 0.0371
8.0 12.9 0.0398 8.0 13.8 0.0406
10.0 14.2 0.0439 10.0 15.1 0.0445
Lead (82;207.20) Uranium (92;238)
0.1 1785 5.18 0.1 3801 0.962
0.2 282 0.819 0.2 533 1.35
0.4 53.7 0.156 0.4 82.2 0.208
0.6 25.4 0.0738 0.6 38.3 0.0970
0.8 17.4 0.0506 0.8 24.9 0.0630
1.0 13.8 0.0401 1.0 19.1 0.0483
1.5 10.5 0.0304 1.5 13.7 0.0346
2.0 10.4 0.0302 2.0 13.1 0.0331
3.0 10.4 0.0302 3.0 13.1 0.0331
4.0 1.4 0.0332 4.0 14.0 0.0354
6.0 12.9 0.0375 6.0 16.6 0.0419
8.0 13.2 0.0384 8.0 17.1 0.0432
10.0 15.5 0.0451 10.0 18.8 0.0475

18

PERIOD ENDING OCTOBER 31, 1958

1.2, COMPONENT DEVELOPMENT AND TESTING

H. W. Savage

W. B. McDonald

Reactor Projects Division

FUEL PUMP DEVELOPMENT
W. F. Boudreau A. G. Grindell

Development Tests of Salt-Lubricated Bearings

P. G. Smith W. E. Thomas
H. E. Gilkey

Hydrodynamic Bearings. — The results of the first
attempt ! to establish a molten salt (LiF-BeF ,-UF ,
62-37-1 mole %) hydrodynamic bearing film have been
assessed and three additional tests with the same
combination of materials, an INCR-8 bearing and an
INOR-8 journal, have been made. A review of the
bearing installation procedure used for the first test
revealed that the bearing had been placed into the
test apparatus upside down. This error accounts
for the difficulty experienced in maintaining a
hydrodynamic film during bearing operation.

Testing of a second bearing and journal of INCOR-
8 was terminated on schedule at the end of a 500-hr
period of satisfactory operation. Examination of
this bearing and journal revealed a very small
amount of wear and a nearly unchanged surface
finish. Every indication points to the establishment
and maintenance of a hydrodynamic film during
operation of this bearing and journal. Fifty-seven
stops and starts were made during a 54-hr period
of the test. The remainder of the test was con-
ducted at constant conditions, that is, a shaft
speed of 1200 rpm and an applied radial load of
200 lb. A report? covering the first and second
tests was written.

The same bearing and journal were then assembled
for a third test to investigate the effect of repeated
operation. The third test covered a period of 284
hr during which 37 stop-start tests were performed.
The last 100 hr of operation were at constant con-
ditions: shaft speed, 1200 rpm; applied radial load,
300 |b. Examination of the bearing and journal re-
vealed slight rubbing marks at each end of the
bearing.

P, G. Smith, W. E. Thomas, and L. V. Wilson, MSR
Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 20.

2p, G. Smith, Salt-Lubricated Hydrodynamic Journal
Bearing Tests Nos. 1 and 2, ORNL CF-58-8-10 (Aug. 7,
1958).

A fourth test with a new (third) bearing and
journal of INOR-8 in molten salt at 1200°F is under
way to trace the cause of rubbing marks to either
stop and start tests or to steady-load operation. [t
is planned to continue the testing of the second
bearing and journal at a later date.

Hydrostatic Bearings. — Fabrication of the hydro-
static bearing tester® was completed and tests were
run on two hydrostatic bearings with water as the
lubricant. During the first test, in which a radial
clearance of 0.003 in. was used, the bearing and
journal experienced considerable wear. It was found
that the supply pressures for the four hydrostatic
pockets differed from each other, instead of being
constant. The radial distribution of these pressures
was very similar to the radial distribution of the
volute pressures, which give rise to the load on the
bearing. As a consequence, the bearing force tended
to be least in the direction of the greatest bearing
load. This difficulty was traced to four partitions
which had been instalied in the manifold supplying
the hydrostatic pockets. The partitions were re-
moved for the second test, which was conducted
with a second hydrostatic bearing having a radial
clearance of 0.0075 in. For this test, additional
instrumentation was installed to permit indication
of the eccentricity of the journal. This instru-
mentation consists of nozzles through which air
flows radially inward toward the cylindrical shaft

- surface; the back pressure on the nozzle is a

function of the distance between the nozzle outlet
and the cylindrical surface. Further refinements in
this technique of measurement appear to be de-
sirable.

In the second test, data were obtained at shaft
speeds from 600 to 3000 rpm at various pump cir-
cuit resistances in an attempt to determine tne
effect of varying the bearing load. Data on bear-
ing loads and flows were obtained by measuring the
pressure in each bearing pocket and the pressure
differential across each orifice. The second bear-
ing and journal experienced slight wear of the type
that might be expected to occur during stopping and

3L. Y. Wilson, MSR Quar. Prog. Rep. June 30, 1958,
ORNL.-2551, p 19.

19

MOLTEN-SALT REACTOR PROGRESS REPORT

starting. At low speeds, during stopping and start-
ing, no supply pressure is available to the hydro-
static pockets; thus contact between the journal
and bearing may be expected.

Rotating-Pocket Hydrostatic Bearings. — Static
tests were completed on the rotating-pocket hydro-
static bearing,4 and dynamic tests are in progress.
In the initial dynamic test it was found that there
was entrainment of air in the process water. This
problem has been solved by the use of an external
circuit that carries flow from the upper side of the
impe ller-bearing region to the impeller suction
through a stilling well. Another problem that is
being investigated is that of the pressure dis-
tribution in the pockets. The pressure distribution
obtained thus far in dynamic tests does not con-
form to that found in the static tests and pre-
dicted by the analytical studies.

Bearing Mountings. -~ Two types of bearing mounts
are being studied. One of these is a column type
of sleeve mount designed at ORNL that takes into
account the thermal expansion differences of the
materials. Plans have been made to test this mount
statically prior to applying it to an actual bearing
in a molten-salt system. Fabrication of a mount for
testing awaits a satisfactory solution to the problem
of brazing molybdenum to Inconel and to INOR-8.
The other type of mount being studied is a patented
fixture supplied by Westinghouse. Tests of this
fixture, designated ““Thermoflex,'’ are under way.

Development Tests of Conventional Bearings
D. L. Gray W. E. Thomas

Organic-Liquid-L ubricated Bearings. — The high-
temperature sump pump being used to conduct a
demonstration of the behavior of Dowtherm **A’’
(eutectic mixture of diphenyl and diphenyloxide)
as a bearing lubricant operated satisfactorily for a
period of 3206 hr. The pump circulated fuel 30
(Na F-ZrF4-UF4, 50-46-4 mole %) at a temperature
of 1200°F; the shaft speed was about 2600 rpm; the
temperature of the Dowtherm ‘‘A’’ supplied to the
bearings was maintained at 180°F, with approxi-
mately a 5°F rise through the test bearings. The
lubricant flow rate was 2 gpm.

Upon completion of 3200 hr of operation, the
pump was disassembled, and the test bearings and
seals were inspected. The double-row angular-

. v, Wilson, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 17.

20

contact ball bearings (MRC-5309), which were sub-
merged in the lubricant during operation, showed
no detectable wear or imperfections. The diametral
clearance between the aluminum bearing and the
Inconel journal was found to be essentially the
same (0.0044 in.) as it was at the time of the
previous 100-hr inspection. Seal nose heights
were measured at the beginning and the end of a
2008-hr test period, and wear indications of 0.0016
in. on the upper seal and 0.0004 in. on the lower
seal were noted. Lower seal leakage averaged

2 cm®/day during the operating period, but the
upper seal leakage was essentially zero.

The pump was reassembled and reinstalled in the
test stand and operation at the above-mentioned
steady-state conditions was resumed. After 6 hr of
operation, that is, an over-all total of 3206 hr of
operation, a building power interruption caused the
lube oil pump to stop. Within about 30 sec, it was
noticed that the drive equipment was heavily over-
loaded and that the pump speed had been reduced
to zero. Upon disassembly of the pump it was
found that the pump shaft had seized in the aluminum
bearing. It was further observed that the three oil-
supply grooves in the bearing were not covered by
the journal for approximately T/4 in. Thus a path
existed during the entire test for direct escape of
the pressure-fed lubricant to the region external to
the load-carrying film of the bearing. However, it
is believed that this bearing would have continued
to operate for a longer period of time had the lube
oil pump remained in operation. The demonstration
was terminated after this incident and a final report
is being written,

Bearing and Seal Gamma Irradiation. — The
centrifugal-pump rotary element assembly being
operated without an impeller in a gamma-irradiation
facility at the MTR had accumulated a total of
5312 hr of operation by the end of the quarter; during
4426 hr of the operating period, the assembly has
been exposed to gamma irradiation. The accumu-
lated gamma-ray dose to the lower seal region has
reached approximately 9.3 x 10% r. The bulk oil
has accumulated an approximate gamma-ray dose of
2.25 x 108 r.

On June 12, 1958, a momentary power failure
stopped shaft rotation and caused seal leakage of
53 ¢m? in the following 24 hr. During the succeed-
ing 24 hr the leakage was 12 cm?, after which the
leakage leveled off to the normal 6 to 7 cm®/day.
On June 23, 1958, an attempt was made to ac-
celerate the bulk oil radiation damage by reducing

the lube oil dilution factor. A total of 5430 cm®

of oil had been removed from the lube oil reservoir
when the low lube oil flow alarm halted rotation of
the shaft. Sufficient oil was then replaced to
permit stable operation, but subsequently operation
was stopped twice because of a low lube oil level.
The leakage of the lower seal increased during this
quarter from approximately 6 ml/day to approxi-
mately 60 ml/day. It is believed that the stoppages
may have contributed to the increased rate of leak-
age.

[t is planned to terminate the test when the lower
seal region has received a dose of 10'% r. This
should be achieved within three additional weeks,
The test unit will then be returned to ORNL for
examination.

Table 1.2.1 indicates the effect of the radiation
on samples of the lubricant, Gulfspin 34, which
were taken at various times and tested by the ORNL
Analytical Chemistry Division.

The viscosity of both the bulk and the seal-
leakage oil has increased, but is not yet deleterious
to the operation of the test, The increase in the
bromine number indicates that the gamma irradiation
has displaced some hydrogen atoms from the hydro-
carbon molecules of the lubricant. The acidity

PERIOD ENDING OCTOBER 31, 1958

number indicates that the lubricant system has
suffered very little oxidation during the test.

Mechanical Seals
D. L. Gray

Labyrinth and Split-Purge Arrangement. — Load-
deflection data were obtained for new seal springs
and for the seal springs used in the NaK pump that
was operated to test a labyrinth and split-purge
arrangement, as described previously.® The data
are being analyzed to determine the effect of the
test conditions on the seal-face loading created by
the springs. The load-deflection tests were made
both with and without the elastomeric O-ring which
seals between the stationary and floating members
of the seal in order to determine the effect of the
resistance offered by the O-ring to the axial move-
ment of the flexibly mounted seal ring element of
the seal assembly. The results of these tests and
the operation of the pump are being correlated.

Bellows-Mounted Seal. —~ The modified Fulton-
Sylphon bellows-mounted seal being subjected to an
endurance test in a NaK pump has accumulated a

D, L. Gray, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 21.

Table 1,2,1, Effect of Gamma Irradiation on Lubricating Properties of Gulfcrest 34

in a CentrifugalsPump Rotary Element Assembly

Bulk Circulating Qil

Accumulated Oil Leakage

from Lower Seal

Control Sample Taken 3-13-58 5.24-58  7-29-58
Sample  3.77.58 6-5-58 7-31-58 to 4-28-58 to to
3-27-58 5.21-58  8.8-58
Approximate gamma- 0 0.47 x10% 1.3x10% 1.82x108
ray dose (r)
Viscosity (SUS at 90.2 89.4* 104.9* 118.5 108.0 112.6 140.2 168.2
25°C)
Bromine No. {mg of Br 0.87 2.8 4.2 5.6 3.4 4,1 7.3 6.8
per 100 g of oil)
Acidity (mg of KOH 0 0 0 0.013 0 0 0 *

per g of oil)

*These quantities were checked within 5.5% by an independent analysis performed by the Special Samples

Laboratory of the Y-12 Technical Division.

**Result not yet available.

21

MOLTEN-SALT REACTOR PROGRESS REPORT

total of 8110 hr of operation. The pump has been
operating at a temperature of 1200°F and a speed of
2500 rpm. The observed seal leakage has continued
to decrease to the point that it has become difficult
to measure and can be considered to be negligible.
Two stoppages occurred: (1) a power failure caused
the pump to be stopped for 5 min, and (2) the pump
was stopped to replace the motor brushes. The

test has been scheduled for termination at the end
of 8760 hr, at which time the seal will be examined.

Radiation-Resistant Electric Motors for Use
at High Temperatures

S. M. DeCamp

The investigation of the materials that would be
required for radiation-resistant motors for use at
high temperatures has included studies of electrical
steels and high-strength low-resistance electrical
conductors, in addition to the studies of electrical
insulation previously reported.® This work is
directed toward the development of a motor suitable
for long-term operation at 1250°F in a radiation
field.

Six coil assemblies that incorporate the electrical
insulation system developed by the Louis-Allis
Company for use at high temperatures were re-
ceived early in September. These coil assemblies
(called *‘motorettes’’) are arranged on a steel frame
in such a fashion that the three separate aspects of
the insulation problem normally encountered in an
electric motor are properly simulated and can be
separately measured. These aspects of the problem

' eoil-to-

are usually referred to as *‘turn-to-turn,
coil,’’ and “‘coil-to-ground’’ resistance.

During the course of development work by the
Louis-Allis Company it was found that a new wire
product (Silotex-N) made by Anaconda Copper
Company was superior to ceramic-coated wire at
elevated temperatures. Silotex-N wire is a nickel-
plated copper conductor covered with a glass serv-
ing. The wire is also coated with a silicone
varnish that is stable up to about 350°C. At this
temperature the silicone is driven off, and SiO, is
left as a residue. The glass serving appears to be
a good turn-to-turn insulation at high temperatures
based on tests run by the Louis-Allis Company.
No tests have been run by ORNL to date.

The coil-to-coil and coil-to-ground insulation
supplied with the motorettes consists of mica that

65, M. DeCamp, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 22.

22

was coated and impregnated with a modified ceramic
cement by the High Temperature Instrument Com-
pany. Again, preliminary tests by the Louis-Allis
Company at 1200°F indicate good results. After
the coils are wound, they are placed in the “‘slots”’
of the motorette with the mica as a slot liner {(coil-
to-ground insulation) and as a coil separator (coil-
to-coil insulation). The whole assembly is then
impregnated with a ceramic compounded by the
Louis-Allis Company. All materials used in the
motorettes appear to offer good insulating properties
at a tamperature of 1200°F. Radiation resistance
has not been checked by test, as yet, but this re-
quirement was considered in the selection of ma-
terials. Testing of the motorettes by ORNL will
include assessing the effects of time at elevated
temperatures, thermal cycling operation during nu-
clear irradiation, and operation during nuclear irradi-
ation at elevated temperatures.

A preliminary investigation of magnetic core ma-
terials for use at high temperatures indicates that
little progress has been made in this field. More
work on obtaining satisfactory magnetic steel is
planned.

Operation of an electric motor at 1250°F will
require electrical conductors with good mechanical
strength properties and low electrical resistivity.
Work is presently under way to determine types of
materials which might be considered for such an
application.

Design Studies of Fuel Pumps

Engineering study contracts for the evaluation of
various types of molten-fluoride-salt pumps and for
the preparation of preliminary design studies were
awarded to two companies late in the previous
quarter, and work on both contracts (Allis-Chalmers
Manufacturing Company and Westinghouse Electric
Company) was completed during this quarter,

The work by Allis-Chalmers tock the form of a
feasibility study. Nineteen different types of pump
and drive combinations were evaluated, including
centrifugal pumps of the canned-motor type, similar
pumps with remote drives, positive displacement
pumps, jet pumps, and electromagnetic pumps. The
study resulted in a recommendation that two types
of centrifugal pumps be selected for further study:
(1) a pump with a vertical shaft and a free gas-to-
salt interface and a totally enclosed fan-cooled
motor drive; and (2) a turbo-pump energized by a
secondary fluid, possibly a barren-salt mixture that
is compatible with the reactor salt and the container
material.

Three preliminary centrifugal pump layouts were
selected by Westinghouse for inclusion in their
report (a number of other, less applicable, layouts
were shown to ORNL engineers during the course of
the work). Two of the layouts included in the re-
port showed pumps of the sump type (having a free
liquid-to-gas interface in the body of the pump).
Both of these layouts made use of a salt-lubricated
lower bearing, while the upper bearing was lubri-
cated by a low-melting-point salt in a container
physically separated from the process salt. Both
pumps made use of totally enclosed motors.

The third Westinghouse layout showed a sub-
merged pump completely filled with salt that made
use of a motor designed for operation at high
temperatures. Both the journal and the thrust
bearings were lubricated by the molten salt. The
motor windings were to be made with a specially
formed mica insulation system being developed by
Westinghouse for another reactor project. A sub-
merged pump of this type would have advantages in
a molten-salt reactor in that (1) the heating of the
pump parts by fission gases would be eliminated and
(2) the pump could be made considerably smaller
than a sump type of pump.

One of the primary purposes of these design
studies was to determine which particular elements
of a fuel pump should be selected for inclusion in
the development program. In this respect, both
vendors emphasized the need for salt-lubricated
bearings.

Frozen-Lead Pump Secl

W. B. McDonald E. Storto
J. L. Crowley

A small, submerged, centrifugal pump with a
frozen-lead seal was designed and fabricated and
is being operated in evaluation tests. The pump
volute and impeller were patterned after a standard
Eastern Industries, Inc., Model D-11 oil pump and
were fabricated of Inconel. The pump barrel, which
is the container for the lead, contains the seal
gland. The gland incorporates a long tapered

PERIOD ENDING OCTOBER 31, 1958

annulus which narrows to a short annulus of con-
stant diameter around which the cooling coil is
wrapped (see Fig. 1,2,1).

The pump shaft is a directly coupled extension
of the motor shaft and depends upon the motor for
support and guidance. There is a 0.015-in. diam-
etral clearance between the pump shaft and the
pump body at both the seal gland and the volute.
The drive motor is a fractional horsepower (]/8) in-
duction motor rated to draw 2.0 amp under a full
load at 3460 rpm. During operation with the
frozen-lead seal this motor utilizes 1.5 amp at

3580 rpm.

The test loop in which the pump is operating con-
sists of 3/3-in. sched-40 Inconel pipe with a small
surge tank at the top. Heat is applied by Calred
heaters controlled by Variacs, and temperatures are
measured and recorded by thermocouples and a
recorder. The pump, motor, and loop assembly is
shown in Fig. 1.2.2 without the heaters and thermo-
couples.

In preparation for filling, the loop piping was
heated to 1000°F. With the pump shaft rotating,
1100 g of pure lead was loaded into the pump
barrel; the chamber was about two-thirds full. The
remainder of the loop and the pump barrel were then
filled with fused salt mixture 30 (NaF-ZrFA-UF“)
and flow was started. |sothermal operation was
established at 1200°F for the main loop, and the
temperatures of the pump barrel were regulated to
establish a lead-to-fused salt interface temperature
of 1200°F and a seal-gland temperature of 820°F,

Since the start of operation on June 13, 1958, the
pump and seal have operated continuously for 2600
hr. During the first 100 hr of operation, a slight
leakage of lead was observed. This was stopped
by adjusting the coolant flow, and no leakage has
occurred since. A closeup view of the seal gland
while in operation is shown in Fig. 1.2.3.

In order to further evaluate the possibility of a
lead-sealed pump, a larger model is being pre-
pared for testing. A 3]{‘-in.-d ia shaft, which will
operate at 1200 rpm, will be used in a seal gland of
similar configuration. The peripheral speed of the
shaft will be 1020 fpm compared with 190 fpm for

the smaller model.

23

MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL— LR—DWG 34508

/ PUMPHEAD

PUMP VOLUTE IMPELLER

TANK BARREL

N

PUMP SHAFT

7
| |
| 7 N
/ ZPUMP FOOT

TANK FILL LINE

9

COOLING COIL

LA U

! 2
1 I

; 1

r— O

INCHES

Fig. 1.2,1, Sectional View of Lead-Sealed Pump.

24

UNCLASSIFIED
PHOTO 31626

Fig. 1.2.2, Frozen-Lead Sealed Pump and Test Loop

Without Heaters and Thermocouples.

PERIOD ENDING OCTOBER 31, 1958

Fig. 1.2.3. Frozen-Lead Seal Gland During Operation.

DEVELOPMENT OF TECHNIQUES FOR REMOTE
MAINTENANCE OF THE REACTOR SYSTEM

W. B. McDonald

Mechanical Joint Development
A. S. Olson

The necessity for the development of a reliable
mechanical joint was explained previously,” and
detailed descriptions of the three types of joints
under consideration were presented. Screening
tests of the three joints were conducted under
thermal-cycling conditions in a high-temperature
loop that circulated fuel 30 (NcF-Zer-UFA) at
temperatures up to 1500°F. 87

During this quarter, the freeze-flange and in-
dented-seal-flange joints were tested with sodium,
since sodium may be used as the primary coolant
for a molten-salt power reactor. The two joints
are shown installed in the loop in Fig. 1.2.4. The
testing procedure consisted of the following: (1)
assemble joint and leak check (cold) on helium leak
detector; leakage specification, =1 x 107 cm? of
helium per second; (2) weld joint into loop and
operate through 50 thermal cycles between 1100
and 1300°F; (3) drain loop and remove joint intact;

A S, Olson, MSR Quar. Prog. Rep. Jan., 31, 1958,
ORNL-2474, p 20.

8A. . Olson, MSR Quar. Prog, Rep. June 30, 1958,
ORNL-2551, p 24.

9W. B. McDonald, E. Storto, A. S. Olson, Screening
Tests of Mechanical Pipe Joints for a Fused Salt Re-

actor System, ORNL CF-5 -8-33 (Aug. 13, 1958).

25
MOLTEN-SALT REACTOR PROGRESS REPORT

CL AJQ 'EO
flOTO 31627

Fig. 1.2.4. Freeze-Flange and Indented Seal-Flange Joints Installed in Loop for Testing with Molten Sodium.

(4) leak check joint on leak detector; and (5)
separate joint, remake joint, and leak check.

The leakage rate for the freeze flange joint,
which cgntained a seal ring made of annealed
copper, was 2 x 10=8 cm3/sec before the test and
1 x 107 ecm3/sec after the test. The flangese
were then separated and reassembled with a new
seal ring of the same material, and the leakage
rate dropped to 1.3 x 10~8 cm?/sec.

The leakage rate for the indented-seal-flange
joint, which contained a seal ring made of nickel-
plated Armco iron, was 5 x 10~8 cm3/sec before
the test and 3 x 10=8 cm®/sec after the test. The
flanges were then separated and reassembled with

a new seal ring of the same material, and the [eak-

age rate dropped to 1 x 10~8 cm3/sec.
Temperatures were measured during the test at
the points indicated in Fig. 1.2.5. The molten

26

THERMOCQUPLE

3| ‘ EY
Ll 1 L N
T —4 =
l [ SODIUM
“u | FLOW
32 I
L
33| |
. J! - ]
[ C3—fe
— 4

FREEZE-FLANGE JOINT
TESTNO.5

FLOW

| o8
SODIUM ) ‘ [ /

UNCLASSIFIED

ORNL~LR-DWG 34509
THERMOCOUPLE
21
> Ry —
Fe { ;J(\ ‘
231 »] \
|
ks /J
\\ /
|
24 o (
e A VK - -
o~ °

(
| &%

NDENTED-SEAL-FLANGE JOINT

TEST NO. 3

Fig. 1.2.5. Diagram of Locations of Thermocouples on
Mechanical Joints Tested in a Loop That Circulated

Sodium.

sodium was circulated at a flow rate of 2 gpm, and
the average temperature cycle time for the 50-cycle
test was 60 min. Representative temperatures
measured during the test are listed in Table 1.2.2.
Both joints operated successfully; there was no
indication of sodium leakage. The separated joints
are shown in Figs. 1.2.6 and 1.2.7 as they appeared
after the test.

The annealed copper seal ring used in the freeze-
flange joint was identical to that used in the pre-
vious tests with molten salt. The gasket used in
the indented-seal-flange joint was made of annealed
Armco iron that was nickel plated to prevent oxi-
dation of the iron. This gasket material was also
successfully tested with molten salt.

A detailed description of the large freeze-flange
foint for use in a 4-in.-dia line was presented
previously.® Tests of two of these joints were con-
ducted in a high-temperature loop that circulated

PERIOD ENDING OCTOBER 31, 1958

fused-salt fuel 30 at temperatures up to 1300°F
under thermal-cycling conditions. The flanges are
shown installed in the loop, prior to testing, in

Fig. 1.2.8, and the test loop is shown schematically
in Fig. 1.2.9.

A heater-insulation unit, designed for use on a
4-in.-dia line, was installed on the loop and tested
along with the large freeze-flange joints. The
heater-insulation unit is shown in Fig. 1.2.10. A
spacer is shown in Fig. 1.2,11, Two of the heater-
insulation units and one spacer, located between
the units, are shown installed on the test loop in
Fig. 1.2.8. It was found that the thermal loss from
the two units was about four times the heat loss
from an equivalent length of ‘‘Hy-temp'’ pipe in-
sulation 3 in. thick on a 4-in. pipe.

The cold leakage rates for the two large freeze-
flange joints, before installation in the loop, were
6.7 x 10~ cm? of helium per second for flange No.

Table 1,2,2, Temperatures Measured During Thermal Cycling of

Flanged Joints in a Sodium-Filled Test Loop

Thermocouple Minimum Cycle Maximum Cycle
Nomber* Temperature Normber Temperature Nomber
(°F) (°F)
Freeze-Flange Joint
27 110 10-20-28-29 133 9
28 170 21 215 9-41
29 905 29-31-32-34 1095 1-2
30 1075 31-32 1295 8
31 1045 31-32-39 1255 Numerous
32 890 38 1085 2
33 170 20 212 8-9
34 108 10 130 8-9
Indented-Seal-Flange Joint
21 908 37 1025 35
22 920 4 1035 18-28-41
23 925 3.4 1050 35
24 965 Numerous 1110 35
25 1060 32-38 1055 Numerous
26 1078 21-33 1288 8-41

*See Fig. 1.2.5 for location of thermocouples.

27

MOLTEN-SALT REACTOR PROGRESS REPORT

Fig. 1.2.6. Indented-Seal-Flange Joint After Removal from Molten Sodium Test Loop.

Fig. 1.2.7. Freeze-Flange Joint After Removal from Molten Sodium Test Loop.

28

UNCLASSIFIED
PHOTO 319R4

UNCLASSIFIED
PHOTO 31985

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
PHOTO 31950

Fig. 1.2.8. Large Freeze-Flange Joints and Heater-Insulation Units Installed in Test Loop.

1 and 1.6 x 10~8 cm?/sec for flange No. 2. During
the loop test, temperatures were measured at the
points indicated in Fig. 1.2.12. Molten salt was
circulated at a flow rate of ~40 gpm. The salt
temperature was cycled 50 times between 1100 and
1300°F; each cycle was of 2 hr duration. Repre-
sentative temperatures measured during the test
are listed in Table 1.2.3.

There was no indication of salt leakage during
the test, but gas leakage tests made on the flanges
while they were installed in the loop indicated
leakage rates in excess of the maximum allowable
leakage rate of 1077 cm® of helium per second.
The bolts on the special clamps around each joint
were tightened from original loads of 180 ft-Ib each
to approximately 400 ft-lb each without appreciable
improvement of the leakage rates.

The clamps were therefore removed from the
flanges, and additional bracing was welded to them
to permit higher loads. In addition, oxide was re-
moved from the clamps, and they were given an
oxidation-resistant coating. A special lubricant
containing graphite and molybdenum disulfide in a
resin base was also coated over the surfaces of
the clamp and bolts to reduce friction between
mating surfaces and thus provide a clamp of greater
efficiency. Two seal rings are being fabricated
from 2S aluminum. [t is believed that the use of
these seal rings and a higher bolt loading will
provide a gas-tight joint. The thermal-cycle testing
will be repeated following these modifications.

Both of the joints were separated, after removal
of the clamps, and the inside surfaces of the
flanges were examined for signs of cracking caused

29

MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL-LR~DWG 31380R

,

PRESSURE-MEASURING DEVICES g

! Ya=in PIPE

poV

Lo

4-in PIPE

FREEZE FLANGES

FILL LINE AND FREEZ\E VALVE
)

Fig. 1.2.9. Freeze-Flange Mechanical Joint Development Test Loop.

30

PERIOD ENDING OCTOBER 31,

UNCLASSIFIED
PHOTO 31558

Fig. 1.2.10, Heater-Insulation Shown in Open Position.

UNCLASSIFIED
PHOTO 31557

Fig. 1.2.11. Spacer for Heater-Insulation Unit.

1958

31

MOL TEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL--LR—DWG 3451

() FLANGE NO.1 THERMOCQUPLES

| ] FLANGE NO.2 THERMOCOUPLES
DOUBLE LINES INDICATE WELLS

Fig. 1.2,12, Diagram of Location of Thermocouples on
Joints Tested in Mechanical Joint Development Test

Loop.

by thermal stress. No such stress cracks were ob-
served in dye checks.

The joints were easily disassembled in about 20
min. The frozen salt in each joint was firmly
attached to the stainless steel screen that had been
placed there for that purpose, as shown in Fig.
1.2,13. A flange is shown with the screen and
salt seal removed in Fig. 1.2.14. Note that the
flange face is sufficiently clean for reassembly of
the joint.

Evaluation of Expansion Joints for Molten-Salt
Reactor Systems

J. C. Amos

Three commercially available expansion joints
have been tested in the test facility shown in Fig.
1.2,15. The joints were cycled over their maximum
allowable traverse at rated conditions of 1300°F
and 75 psig in an environment of molten salt or

Table 1.2.3. Temperatures Measured During Thermal Cycling of Freeze-Flange Joints

in Mechanical Joint Development Test Loop

Thermocouple Minimum Cycle Maximum Cyele
. Temperature Nomber Temperature Number
(°F) (°F)
Flange No. 1
] 440 17 550 3
2 827 17 1030 6-10
3 962 17-18 1140 10
4 965 17-18-29 1135 10
5 875 10-17-29 1030 10
6 353 5 435 30
7 265 17 305 3
8 247 17 288 11-32
Flange No, 2
9 458 28 598 32
10 882 28-29 1045 10
11 970 17-18 1140 Numerous
12 932 18 1098 26
13 898 18 1063 8
14 413 17 510 2-3-31-49
15 242 17-18-41 307 49
16 330 10-18 378 49

*See Fig. 1.2.12 for location of thermocouples.

32
PERIOD ENDING OCTOBER 31, 1958

‘ UNCLASSIFIED

PHOTO 32544

UNCLASSIFIED
32545

’
S

Fig. 1.2.14, Disassembled Freeze-Flange Joint After Removal of Frozen Salt Seal and Screen.

33

MOL TEN-SALT REACTOR PROGRESS REPORT

Qo ~ oY '}. ; —.'
PNEUMATIC OPERATOR
i .

L

N\ _
VAPOR TRAP

FURNACE

Fig. 1.2.15. Expansion Joint Test Facility.

sodium. The joints and the results of the tests are

described in Table 1.2.4.

The failure of the Zallea joints cannot be con-
strued as a reflection of the integrity of the joints,
since the bellows were evidently weakened by the
attachment of thermocouples. The Cook Electric
Company joint will be examined metallurgically in
an effort to determine whether the failure was due
to a weld defect or to excessive stresses at rated
operating conditions. Three more expansion joints,
shown in Fig. 1.2.16, are currently being installed
in the test facility. Thermocouples will not be
attached to these units.

PERIOD ENDING OCTOBER 31, 1958

Remote Maintenance Demonstration Facility
C. K. McGlothlan

The contractor’s portion: of the field construction
work for the remote maintenance demonstration
facility (Fig. 1.2,17) in Experimental Engineering
Building 9201-3 has been completed. This work
consisted of fabricating and erecting the salt safety
spill tray and enclosure and the railway for the
manipulator. In addition, the contractor installed
structures for the dummy reactor, piping, and dummy
heat exchangers (Fig. 1.2.18).

In Fig. 1.2.17 the fuel dump tank and two dummy
heat exchangers are also shown in their approximate

Table 1,2,4. Descriptions of Expansion Joints and Results of Tests

Vendor Cook Electric Co. Zallea Bros. Zallea Bros.
Material Stainless steel Inconel Stainless steel
Test fluid Sodium Molten salt Sodium
Maximum traverse of joint 2]/4 in. 25/8 in. 25/8 in.

Number of cycles to failure 71 49 60

Location of failure Bellows seam weld Thermocouple attachment Thermocouple attachment

to bellows to bellows

COOK ELECTRIC CU TG s

UNCLASSIFIED
PHOTO 32544

FLEXCONICS CORP

JIOINTS .

JOINT

Fig. 1.2.16. Expansion Joints To Be Tested in the Expansion-Joint Test Facility.

35

e e

MOL TEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
PHOTO 32443

Fig. 1.2.17. Remote Maintenance Demonstration Facility.

position. Although this facility contains a dummy
reactor and two dummy heat exchangers, it will con-
tain a loop with a pump for circulating molten salt
at 1250°F through 3%-in. and 6-in. Inconel pipe.
It will be demonstrated in this facility that any

or all components of this quarter-scale simulated
reactor system can be maintained remotely by a
General Mills manipulator as viewed through a
stereo-type closed-circuit television system. The
procedures developed can be adapted for use on a
full-size reactor system.

The manipulator will operate on a 40-ft-long track
with a 22-ft span. The vertical travel of the mani-
pulator arm will be 20 ft. This manipulator is now
being constructed at General Mills Company.

The first of three shipments of equipment for the
two closed-circuit television systems has been

36

made by General Precision Laboratory. The dummy
reactor has been fabricated and is on the job site.

Fabrication in a local job shop of the 3‘/2-in. and
6-in. loop piping subassemblies is in progress.
Bids are being obtained for the fabrication of 20
sets of freeze flanges and the removable electric
heater and insulation units for the 3%-in. and 6-in.
piping. In addition, bids are being obtained to
modify an existing molten-salt pump for use in
this system.

Approximately two-thirds of the engineering work
for this job has been completed. The remaining
portion of the engineering work consists mostly of
electrical, instrumentation, and remote tooling de-
sign.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
PHOTO 12442

Fig. 1.2.18. Heat Exchanger, Piping and Reactor Support Structure of Remote Maintenance Demonstration Facility.

HEAT EXCHANGER DEVELOPMENT
J. C. Amos

Al 2Cl6 Thermal-Convection Loop

A thermal-convection loop has been designed for
evaluating the heat transfer performance of aluminum
chloride gas. This gas has been suggested as a
possible low-pressure gas coolant which would
eliminate many of the containment problems in-
herent in high-pressure gas systems. Aluminum
chloride gas is of interest for the molten-salt re-
actor concept because of the almost complete dis-
sociation reaction (Al ,Cl ,—=2AICl,) that occurs
between 900 and 1200°F. This dissociation pro-
cess is reversible and results in a large increase
in the heat capacity of the gas over the temperature
range presently considered for the molten-salt re-
actor fuel (see Chap. 2.3, ‘‘Chemistry,”” for dis-
cussion of ““Gaseous Aluminum Chloride as a Heat

Exchange Medium'’). Operation of a thermal-
convection loop should provide an initial qualitative
evaluation of the effect of dissociation on the heat
transfer characteristics of this gas.

The small-scale apparatus shown schematically
in Fig. 1.2.19 was assembled to obtain a subli-
mation temperature vs pressure curve for aluminum
chloride and to provide preliminary compatibility
data for the gas in the container materials of
principal interest. A plot of the temperature vs
vapor pressure data is presented in Fig. 1.2.20.
Inconel and INOR-8 corrosion samples installed in
the test apparatus were exposed to aluminum
chloride vapors for 1000 hr. Visual and metallo-
graphic examination disclosed intergranular attack
to a maximum depth of 2 mils at the high-temperature
(1200°F) end of the Inconel specimen and grain-
boundary penetration to a depth of 0.5 mil at the
high-temperature end of the INOR-8 specimen. The

37

MOL TEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 34512

PILOT
VALVE
MOORE NULL-
BALANCE PRESSURE- O O INSTRUMENT
MEASURING DEVICE — AIR SUPPLY
— :
BELLOWS F M \PRESSURE GAGE
. Pl
CALROD HEATER—2 O HERMOCOUPLE

~—CLAMSHELL HEATER

—

— [—‘M_—\\J

CREINTIONLNE o= SHEDULE 0 P
/>< w——CORROSION SPECIMENS
THERM PLE <<
CRMOMIUPLE S % THERMOCOUPLE
O (O—=——CALROD HEATER
7::- E—
Ny __—>THERMOCOUPLE
THEwocouHE\_,g 9;
4 —S0LID AICI4

CONTAINER

Fig. 1,2.19. Diagram of Apparatus for Determining
Aluminum Chloride Sublimation Temperature vs Pressure

Data and Preliminary Compatibility Data.

UNCLASSIFIED
ORNL-LR-OWG 34513

130

110 |

100

PRESSURE ( psig)

/La—OATA BY FRIEDEL AND CRAFTS, FROM _
C.A THOMAS, " ANHYDROUS ALUMINUM
7 CHLORIDE IN ORGANIC CHEMISTRY," REINHOLD J

~

300 400 500 600 700
TEMPERATURE (°F)

Fige 1,2.20. Aluminum Chloride Sublimation Tempera-

ture vs Pressure Data Plot.

38

cold end (750°F) of both specimens revealed a
light, metallic-appearing deposit of a maximum
thickness of 2 mils on the INOR-8 specimen and 1
mil on the Inconel specimen. Spectrographic
analysis showed this deposit to be predominantly
aluminum. Representative photomicrographs of
these specimens are shown in Figs. 1.2.21 through
1.2.24. These preliminary data indicate that either
of the materials tested would be satisfactory for
use in the thermal-convection loop and suggest that
there may be no serious compatibility problem
between aluminum chloride gas and INOR-8, the
container material proposed for a molten-salt power
reactor system.

Molten-Salt Heat Transfer
Coefficient Measurement

Modifications to the molten-salt heat transfer co-
efficient test facility required for a salt-to-salt heat
transfer test are essentially complete. Data runs
will be started early in the next quarter. A de-
scription of this test facility and the results of a
salt-to-liquid metal test were reported pre-
viously. 1%

10, . Amos, R. E. MacPherson, and R. L. Senn,
MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 27.

1), C. Amas, R. E. MocPherson, oand R L. Senn,
MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 31,

UNCLASSIFIED
T-15799

o

Fig. 1.2.21. Top of Inconel Specimen Exposed to
Aluminum Chloride Gas at 1200°F for 1000 hr. Etchant:
glyceria regia. 250X,

I ) UNCLASSIFIED
SAMPLE 1 -

3

? T-15798 LT
(& ]
e z

SAMPLE 2 & ok

Fig. 1.2,22. Bottom of Inconel Specimen Exposed to
Aluminum Chloride Gas at 750°F for 1000 hr.
glyceria regia. 250X.

Etchant:

Fig. 1.2.23. Top of INOR-8 Specimen Exposed to
Aluminum Chloride Gas at 1200°F for 1000 hr. Etchant:
aqua regia. 250X,

PERIOD ENDING OCTOBER 31, 1958

Fig. 1.2.24. Bottom of INOR-8 Specimen Exposed to
Aluminum Chloride Gas at 750°F for 1000 hr. As
polished. 250X,

DESIGN, CONSTRUCTION, AND OPERATION OF
MATERIALS TESTING LOOPS

Forced-Circulation Loops
J. L. Crowley

Construction and operation of forced-circulation
corrosion testing loops was continued. Operation
of seven new loops was started during this period,
and three loop tests were terminated. By the end
of September, 13 loops were in operation and one
loop was ready to be filled. Nine of the operating
loops are constructed of INOR-8 and four of Inconel.
A summary of operations as of September 30, 1958,
is presented in Table 1.2.5.

Of the three terminated loops, one was an In-
conel loop (9377-1), one an INOR-8 loop (9354-2),
and the other an Inconel bifluid loop (designated
CPR). The decision was made to terminate loops
9377-1 and 9354-2 after the power failure on April
6 caused their failure.'? The CPR loop was termi-
nated after completion of a year of operation to

]2J. L. Crowley, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 33.

39
oy

Table 1.2.5. Forced«Circulation Loop Operations Summary as of September 30, 1958

Composition Aporoximate Approximate Maximum Minimum Maximum Hours of
Loap Loop Material Number of pp Reynolds Wall Fluid Fluid Operation at
. . . ) Flow Rate L Comments
Designation and Size Circulated (gpm) Number at Temperature  Temperature  Temperature Conditions
Fluid* 9 1100°F (°F) (°F) (°F) Given
CPR Inconel, 3/fl in. sched 40, 122 1 5,000 1250 1095 1190 9148 Operation terminated to provide cperating
1 in. OD x 0,035 in. wall, Sodium 7 97.700 1085 1135 space for an additional loop
% in. OD x 0,035 in. wall '
9344.1 Inconel, ]/2 in. 0D, 123 2 3,250 1300 1105 1230 7930 Fluid dumped on July 3 for changing of
0.045 in. wall pump motor
9354-3 INOR-8 84 2.75 1200 1100 1145 6289 Operation stopped tempororily on June 16
- for removal of section of loop piping for
Hot feg, /g In- sched 40 4,500 metollurgical examinotion; restarted
Cold leg, ¥ in. OD, 5,400 with new batch of fluid after pump re-
8.045esi’n./3vclxrlll placed and emergency heaters added
9344-2 Inconel, ]/2 in. OD, 12 2.5 8,200 1200 1000 1090 5861 Fluid froze on July 2 when faulty relay
0.045 in. wall caused loss of pump drive; fluid
partially froze again on Aug. 5 when a
building power failure caused failure
of the pump drive motor; operation re-
sumed after both incidents
9377-3 Inconel, !/2 in. OD, 131 2 3,400 1300 1100 1240 4774 Fluid dumped on July 23 for repair of
0.045 in. wail leaking oil rotory wnion
9354-1 Inconel, I/2 in. OD, 126 25 2,000 1300 1100 1190 4308 Fluid dumped on June 20 for instal-
0.045 in. woll lation af emergency heaters; fluid
dumped on July 29 when a lug
adapter made of SP-16 material de-
veloped a crack; adapter replaced
and operation resumed
9377-1 Inconel, I/2 in. OD, 126 2 1,600 1300 1100 1175 3413 Operation terminated after rupture de-
0.045 in. wall scribed previously
9354.5 INOR-8, :"/8 in. OD, 130 1 2,200 1300 1095 1230 3383 Fiuid partially froze on Aug. 5 when
0.035 in. wall building power failure caused pump

drive motor failure; loop contains
graphite samples

LY0d3 Y SSIHYI0¥d JyOLIVI Y LTVS*NILTIOW
Table 1,2,5 (continued)

Composition Approximate Approximate Maximum Minimum Maximum Hours of
Loop Loop Material Number of PP Reynolds Wall Fluid Fluid Operation at
- . . . Fiow Rate o Comments
Designation and Size Circvlated (gpm) Number at Temperature Temperature Temperature  Conditions
Flyid* s 1100°F (°F) (°F) (°F) Given
9354-4 INOR-8 130 2.5 1300 1130 1200 1646 Operation started July 23; loop con-
. toins three INOR-8 sample inserts
Hot leg, 7% in. sched 40 3,000
Cold leg, % in. OD, 3,500
0.045 in. wall
MSRP-7 INOR-8, I/2 in. OD, 133 Not Not 1300 1100 1210 1302 Operation started on Aug. 7
0.045 in. wall available available
9377-4 Inconel, ‘/2 in. OD, 130 1.75 2,600 1300 1100 1185 1241 First loop with cooler coil and con-
0.045 in. wall trols of new design; operation started
July 24; new protective devices pre-
vented freezing of flvid on two
occasions of pump failures; pumps re-
placed and operation resumed
9354-2 INOR-8, '/2 in. OD, 12 2 6,500 1200 1050 1140 1052 Operation terminated after rupture de-
0.045 in. wall scribed previously
MSRP-6 INOR-8, ]/2 in, OD, 134 Not Not 1300 1100 1190 822 Pump bound on first startup attempt;
0.045 in. wall available available during preheating after pump change,
loop tubing parted between cooler and
pump before fluid was introduced;
tubing being examined by Metallurgy
Division for cause; loop repaired and
operation storted on Aug. 27
MSRP-8 INOR.8, ‘/2 in. OD, 124 2.0 4,000 1300 1100 1210 605 Operotion started on Sept. 4
0.045 in. wall
MSRP-9 INOR-8, 1/2 in. OD, 134 Not Not 1300 1m0 1225 471 Operation started on Sept. 10
0.045 in. wall available available
MSRP-10  INOR-8, ‘/2 in. OD, 135 Not Not 1300 1100 1210 264 Operation started on Sept. 19
0.045 in. wall available available
*Composition 12: NaF-LiF-KF (11,5-46,5-42 mole %) Composition 130: LiF-BeF ,-UF (62-37-1 mole %)
Composition 84; Na F-LiF-BeF2 (27-35-38 mole %) Composition 131: LiF-BeF.‘,-UF4 (60-36-4 mole %)
Composition 122: NaF-ZrF4-UF4 (57-42-1 mole %) Composition 133: LiF-BeF2-UF (71-16-13 mole %)
Composition 123: NaF-BeF_.UF, (53-46-1 mole %) Composition 134: LiF-BeF,-ThF UF  (62-36.5-1:0.5 mole %)
Composition 126: LiF-BeF,-UF, {53-46-1 mole %) Composition 135;: Na F-BeF.‘,-ThF“-UF‘1 (53-45.5-1-0.5 mole %)

8%

8561 ‘I ¥3IFOLD0 ONIANT @Ooi¥d3d
MOL TEN-SALT REACTOR PROGRESS REPORT

provide space for an additional loop (see Chap. 2.1,
*Metallurgy,’’ for results of metallurgical exami-
nations).

Other than minor occurrences involving individual
loops, there were only two incidents during the
quarter that affected all loops. The first was a
building power failure on August 5 of less than 1
min duration that affected seven operating loops.
At this time, while an extensive emergency elec-
trical system was being installed, all loops were
equipped so that momentary power failure would
automatically place them on isothermal operation.
Five of the loops did resume operation isothermally
after the power was reapplied. On two loops,
0354-5 and 9344-2, the pump drive motors tripped
out on overload when required to restart under the
load of the magnetic clutch and the pump. These
two loops were partially frozen when circulation
was not maintained but were both successfully
thawed and placed back into operation. The over-
load breakers on all motors, although sized accord-
ing to proper electrical code, were increased to
carry a larger startup current.

The other incident that affected all loops was a
changehouse fire across the street from where the
forced-circulation corrosion loops are located.

The fire destroyed the poles supplying an alternate
power source to the building and also threatened to
spread to the operating area. All operating loops on
that side of the building were placed on isothermal
operation in preparation for dumping. It was not
necessary to dump, however, and the loops were
returned to normal operating conditions after the
danger was past.

Crders were placed for the fabrication of three
INOR-8 loops and two Inconel loops. One INOR-8
loop will be used as a spare and two Inconel loops
will replace two presently operating loops sched-
uled to be terminated.

Of the thirteen facilities with operating loops,
six are of the new design, which gives maximum
protection, as described previously. 2 The re-
maining seven are in various stages of improved
protection that could be provided without disturbing
the loop piping.

In-Pile Loops
D. B. Trauger
J. A, Conlin P. A. Gnadt

Assembly of the first MSR in-pile loop will be
completed by November 1, and the loop is sched-
uled for insertion in the MTR December 1, 1958

42

(MTR cycle 113). A second loop is being as-
sembled, and parts for a third loop are being fabri-
cated. The second loop will provide additional
data and serve as a backup loop in the event of a
failure of the first loop. The instrumentation and
controls required for loop operation are being pre-
pared at the MTR.

The schedule for operation of the second loop
will be determined by the success of the first loop
and the outcome of pump tests presently under way,
as described previously.'® In these tests, two
methods have been tested for starting the loop in
the reactor. The first method consisted of filling
the loop during assembly, freezing the salt, and
then thawing it out after insertion in the reactor.

It was demonstrated in four successive tests with
the first prototype pump that progressive meltout
from the pump to the nose coil could be achieved
without damage to the loop. During meltout, how-
ever, pump speed perturbations occurred, and
several hours after the fourth meltout the pump
seized. Investigation showed a large accumulation
of salt in the intermediate region between the sump
and bearing housing and salt in a granular form
mixed with oil in the copper rotor region. Although
x-ray examination indicated that salt had collected
in the intermediate region without the concurrent
meltout procedure, melting out of the loop was
probably a contributing factor.

A second method of starting the loop in the re-
actor was found to be satisfactory in a second
prototype pump test. The original filling system,
which consisted of an integral fill tank that was
heated and drained into the pump after installation
in the reactor was modified by changing the in-
ternal passages of the pump to accommodate the
high surface tension of the salt. This system was
incorporated in the first loop.

The appearance of salt in the region between the
sump and bearing housing in both the first and
second prototype pump tests presents a further
difficulty. The melting and thawing of the loop in
the first test was thought to have been a con-
tributory cause; however, the salt deposit appeared

again in the second pump test but to a lesser degree.

This salt transfer is not understood, but is be-

lieved to be, in part, a function of the suspected
high surface tension of the fluid. At present the
second prototype pump, similar to the pump used

]3D. B. Trauger, J. A. Conlin, and P, A, Gnadt,
MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 34.
PERIOD ENDING OCTOBER 31, 1958

in the first loop, is being tested to determine the the pump of the second loop. Since construction of
effect of this phenomenon on long-term operation. the first ioop is beyond the point that would permit
A third test of a modified pump is being set up in pump modification, the period of in-pile operation
- an attempt to correct the situation. Modifications will be based on the duration of the pump life test
indicated by these tests will be incorporated in presently in progress.
43

eSS

MOL TEN-SALT REACTOR PROGRESS REPORT

1.3. ENGINEERING RESEARCH
H. W. Hoffman

Reactor Projects Division

PHYSICAL PROPERTY MEASUREMENTS equation,
W. D. Powers R. H. Nimmo Ho = Hygoo = a+ bT + T2 (cal/g) |
Enthalpy and Heat Capacity where the temperature T is expressed in °C. The

coefficients & and ¢ of Table 1.3.1 may be used to

The enthalpies, heat capacities, and heats of

| fusion were determined for six additional salt mix- obtain the heat capacity, “pt from the following

i tures. Of these, two were fuels of the alkali metal- ~ €dquation:

| berylllu.m sy.stem,' and the other four were chlorides ¢ = b+ 2T (cal /g-°C)
of possible interest as secondary coolants. The p
results are given in Table 1.3.1, in which the con- For comparison and completeness, previously de-
stants «, b, c, are the coefficients in the enthalpy termined values of several related mixtures are

Table 1.3.1. Enthalpy Equation Coefficients und Heots of Fusion of Several Fused Salt Mixtures

Enthalpy Equation Heat of Fusion (cal/g)
Temperature Coefficients*
Salt Mixture Phase R o) Temperature
ange (" C) H, - HS
a b c (°C)
x 10~
NaF-BeF,-UF Solid 100-300 ~7.8  +0.251  +11.83
(53-46-1 mole %)  Liquid 400—800 ~5.7  +0.312  +10.87 365 23
LiF-BeF -UF, Solid 100-300 -10.9  +0.293  +19.33 -
(53-46-1 mole %)  Liquid 400-800 -12.8  +0.539 ~1.90 400 63
LiF-BeF -UF ,** Solid 100300 ~9.6  +0.293 +2.18
(62-37-1 mole %)  Liquid 470-790 +24.2  +0.488 +2.52 430 122 .
NaCl-CaCl, Solid 100~500 4.5  +0.158 +7.60
(49-51 mole %) Liquid 550_850 —18.4  +0.358 ~7.70 500 48
LiCl-CaCt,** Solid 100425 ~3.2  +0.167  +T11.31
(62-38 mole %) Liquid 520-895 +25.8  +0.289 496 61
LiCl-SrCl*+ Solid 100—495 3.6  +0.140 +4.25
(52-48 mole %) Liquid 550850 -39.8  +0.356 ~8.16 525 43
LiCl-BaCl,** Solid 100—500 _4.5  +0.146 +2.50
(70-30 mole %) Liquid 550850 —6.6  +0.278 ~3.93 31 45
LiCl-KCl Liquid 450—800 +15.3  +0.361 ~3.12
(70-30 mole %) i
LiCl-KCl Liquid 400—800 +21.2  +0.315 —0.18
(60-40 mole %) .
LiCl-KCl Liquid 500800 +38.5  +0.239 +4.62

(50-50 mole %)

*Heaot copacity may be evoluated from table os cp = b+ 2cT; enthalpy given as HT - H300C =a+bT+ C‘T2.

44

**Previously reported results repeated for completeness.
repeated in Table 1.3.1. Measurements of the
solid state have not been completed for the three
LiCl-KCl mixtures studied in the liquid state.

The results of initial measurements of the mix-
ture LiF-BeF ,-UF, (53-46-1 mole %) were some-
what erratic. Specifically, the enthalpy values
were observed to be a function of the heating-
cooling cycle history of the sample. Thus, for
samples heated to 450°C and suddenly quenched
in the calorimeter, an average enthalpy of 225 cal/g
was obtained. Data for four samples measured at
500°C resulted in an average enthalpy of 220 cal/g,
a value less than that for the original measurements
at 450°C. Since BeF , and mixtures containing
large percentages of BeF, are known to form
glasses, it was suspected that the anomalous be-
havior was associated with this phenomenon. Ac-
cordingly, the samples were remeasured with slower
cooling in the calorimeter (100 min to equilibrium
in comparison with the normal 20 to 30 min). The
slower cooling rate was achieved by inserting an
insulating liner within the cavity of the copper
block. Under these conditions, the enthalpy values
were temperature-consistent and of higher magni-
tude. Since the observed enthalpy is the energy
difference of the salt between the two given tem-
perature levels, the higher values indicate a lower
final energy state and hence a more stable form.
Evidence obtained by varying the heating rate and
pattern in attaining the initial sample temperature
suggests that the upper energy state was identical
for all the samples. The results presented in
Table 1.3.1 for BeF ,-containing mixtures are
based on the slow-cooling data.

The enthalpies and heat capacities of the LiCl-
KCI mixtures are compared in Table 1.3.2 on a
gram-atom basis with results obtained by Douglas
and Dever' at the National Bureau of Standards
for the eutectic mixture LiCl-KCl (59-41 mole %).

It is of interest to note the change in the tempera-
ture dependence of the heat capacity with decreas-
ing LiCl content in the mixture. Between 500 and
800°C the heat capacity of the 70-30 mixture de-
creases by 5.7%, that of the 60-40 mixture by only
0.3%, and that of the 59-41 NBS mixture by 3.5%,
while that of the 50-50 mixture increases by 9.7%.
Additional data will be required for verification of
this indicated effect of composition.

]T. B. Douglas and J. L. Dever, Nationa!l Bureau of
Standards Report 2303, Feb. 20, 1953.

PERIOD ENDING OCTOBER 31, 1958

Surface Tension
Preliminary determinations of the surface tension
of the mixture LiF-BeF ,-UF, (62-37-1 mole %) were
made with the use of the previously described
maximum-bubble-pressure method.? The results,
presented in Fig. 1.3.1, can be represented by the
equation,

o (dyne/cm) = 235,5 - 0.09 T (°C) ,

to within 5% over the temperature range of 460 to

2W. D. Powers, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 38.

Table 1.3.2. Enthalpies and Heat Capacities of
LiCl-K Cl Mixtures of Various Compositions

LiCl-KC! Composition (Mole %)

Temperoture
(°C) 70-30 60-40 50-50  59-41*
Enthalpy (cal/g+atom)
Hp — Hagec Hy - Hgoe
500 4890 4920 4950 5160
600 5740 5790 5800 6030
700 6580 6650 6670 6890
800 7390 7510 7570 7730
Heat Capacity (cal/g-otom+°C)
500 8.58 8.65 8.33 8.76
600 8.42 8.64 8.60 8.63
700 8.26 8.63 8.87 8.53
800 8.09 8.62 9.14 8.45

*Results of Douglas and Dever at NBS (ref. 1).

UNCLASSIFIED

500 ORNL-LR -DWG 34514

—_ 1
. —=—Jo ] I |
g | e S
g ; : : A R —
£ I . \ | |
= {170 ) PR S — —— e ,l e |
> | TUBE NO.  TIP 1D {in.) : |
2 o 4 00700 3 :
& a2 0.0520 | B
H 100 — ' S S P O —
" ° 3 00400 |
<1 4 .
L ‘ [ ‘
fid : !
A 50 L | i i i
450 500 550 600 650 700 750 800

TEMPERATURE (°C)

Fig. 1.3.1. Surface Tension of LiF-Ber-UF4 Mix-
ture (62-37-1 mole %).

45
MOLTEN-SALT REACTOR PROGRESS REPORT

750°C. Three tubes that differed in inside diam-
eter were used. |t is felt that a major source of
the observed scatter of the data is the uncertainty
in the tube-tip inside diameter as a result of pro-
gressive plugging during a test series. Future
studies will include measurement of other salts of
known surface tension and noncorrosive nature in
order to isolate the plugging effect and to estab-
lish accuracy limits for this technique.

Apparatus Fabrication and Calibration

The apparatus? for direct measurements of the
coefficients of thermal expansion of beryllium salt
mixtures has been fabricated and is being as-
sembled. The equipment? for determining the
thermal conductivity of the beryllium salts is being
fabricated. The thermal gradient, and hence the
heat flux, at the ends of the cylinders is to be de-
termined from temperature measurements made at a
large number of positions within the cylinders.
Since it was not found to be feasible to locate the
thermocouples coaxially, interpretation of the ther-
mal profile under conditions of heat exchange at
the cylinder walls becomes difficult. [t is planned
to analyze statistically the experimental tempera-
ture data from the initial measurements so as to
ascertain the thermocouple locations that yield
maximum heat flux precision.

Efflux-cup viscometers are being calibrated that
have larger volumes than the viscometers used
previously.? It is believed that effects such as
““surface-humping’’ will be sufficiently reduced to
allow more precise measurement of the viscosity
of the beryllium salts., Apparatus for the determina-
tion of the electrical conductivity of the beryllium
salts is being assembled.

46

MOLTEN-SALT HEAT-TRANSFER STUDIES
F. E. Lynch

Heat balance difficulties have been experienced
in the evaluation of experimental data on the heat-
transfer coefficient of LiF-BeF2-UF4 (53-46-1
mole %) flowing through an electrical-resistance-
heated smalil-diameter Inconel tube. Specifically,
the measured electrical heat input has been ob-
served to be only one-half the convective heat
gained by the salt, as calculated from temperature
A check made with

recalibrated instruments and with additional instru-

and flow rate measurements.

ments indicated that the power measurements were
not the cause of the discrepancy. New, precali-
brated thermocouples have been installed, and the
weigh-beam flow-measuring device has been re-
checked. A second series of runs will be initiated
shortly under conditions that closely approximate
those of the earlier experiments.

The molten-salt pumping system is being as-
sembled. Since it is the purpose of this study to
determine whether effects such as surface film
formation occur after long exposure of the metal to
the beryllium salt, it is sufficient to obtain rela-
It is
now planned, however, to include an experimental
turbine-type flowmeter in the loop and also to de-

tive values of the heat-transfer coefficient.

termine the flow from the pump calibration, so that
if the flow values obtained by these two methods
consistently agree, an estimate of the heat-transfer
coefficient can be made. The system will contain
an INOR-8 test section, if available, in addition to
the Inconel test section. The two test sections
will be located in adjacent legs of the loop, and a
three-electrode heating scheme will be used. It
will also be necessary to include a heat exchanger
to control the system temperature.
PERIOD ENDING OCTOBER 31, 1958

1.4. INSTRUMENTATION AND CONTROLS
H. J. Metz

Instrumentation and Controls Division

RESISTANCE-TYPE CONTINUOUS FUEL has sufficient resistance to provide a useful milli-
LEVEL INDICATOR volt output from the probe, the data for a given set
R. F. Hyland of conditions could not be reproduced. One pos-

sible source of these inconsistencies is polariza-
tion, a phenomenon commonly encountered in con-
ductivity measurements. This and possible surface
tension effects will be investigated in future tests.

A series of tests of the Inconel resistance-type
fuel-level-indicating elements described previ-
ously] was completed, and typical data are sum-
marized in Table 1.4.1. The results of these tests 'R. F. Hyland, MSR Quar. Prog. Rep. June 30, 1958,
were somewhat disappointing. Although fuel 130 ORNL-2551, p 48.

Table 1.4.1. Summary of Data from Tests of Resistance-Type Fuel Level Probes

Probe supply voltage (regulated): 3.4 v ac

Probe supply current (constant): 1.0 amp

Frequency: 60 cps

Readout device: Hewlett-Packard Model 400H VTVM
Fuel 130: LiF-BeF,-UF, (62-37-1 mole %)

Test Level Range Probe Output Zero to Maximum

Run Date Temperature (in. of fuel at Zero Level Level Probe Qutput Probe Qutput
NO- ]958 (OF) ]30) (mv) Range (mv) Spcn (mv)

1 7/29 1000 0-6 15.7 15.4-12.0 3.4

2 9/2 1000 0-6 15.7 14.9--8.0 6.9

1 8/5 1100 0-6 15.7 15.2-11.5 3.7

2 8/22 1100 0-6 15.7 14.2-6.0 8.2

1 8/11 1200 0-6 15.7 15.5-12.2 3.3

2 8/22 1200 0-6 15.7 14.6-7.0 7.6

1 8/20 1300 0~-6 15.7 14.8-7.4 7.4

2 9/9 1300 0-6 15.7 14.7=7.1 7.6

47
MOLTEN-SALT REACTOR PROGRESS REPORT

1.5. ADVANCED REACTOR DESIGN STUDIES

H. G. MacPherson

Reactor Projects Division

LOW-ENRICHMENT GRAPHITE-MODERATED
MOLTEN-SALT REACTORS

H. G. MacPherson

A rough survey of the nuclear characteristics of
graphite-moderated reactors utilizing an initial com-
plement of low-enrichment uranium fuel has been
made.! The pertinent characteristics of a number
of such reactors are presented in Table 1.5,1. |t
may be seen that reactors could be constructed
that would operate with fuel enriched initially with
as little as 1,25% U235; initial conversion ratios
as high as 0.8 could be obtained with a fuel en-
richment of less than 2%. Highly enriched uranium
would be added as make-up fuel, and such reactors
could probably be operated te burnups as high as
60,000 Mwd/ton. The total fuel cycle cost would
be approximately 1.3 mills/kwhr, of which 1,0 mill
would be the cost of the enriched U233 that would
be burned up.

H. G. MacPherson, Survey of Lou-Enrichment Molten-
Salt Reactors, ORNL CF-58-10-60 (Oct. 17, 1958).

A NATURAL-CONVECTION MOLTEN-SALT
REACTOR

A study was made of a large natural-convection
power reactor to determine the approximate size of
the components and the fuel volume for comparison
with other molten-salt reactors in which the fuel is
circulated by a pump.2 A reactor with a power
rating of 576 Mw (thermal) was chosen for study so
that the heat generated in the core would be com-
parable to that generated in the core of the interim

molten-salt reactor.

The rather simple configuration of reactor, heat
exchanger, and piping shown in Fig. 1.5.1 was
used in the study. All calculations were done on
the basis of one heat exchanger, one riser, and
one downcomer, Since the frictional losses in the
piping are determined primarily by the expansion
and contraction losses and are insensitive to wall

2J. Zasler, 576 Mwt Natural Convection Molten Salt
Reactor Study, ORNL CF-58-8-85 (Aug. 25, 1958).

Table 1.5.1. Choracteristics of Low-Enrichment Grophite-Moderated Molten-Salt Reactors

Uranium Volume of  Uranium  Critical Mass Initial
Case  Yolume Fraction  Multiplication g ;100 Fuelin Core  in Core of U235 Conversion

of Fuel in Core Constaont (%) (ft3) (kq) (kg) Ratio

1 0.05 1.05 1.30 395 22,100 298 0.546

2 1.10 1.45 143 8,000 16 0.492

3 0.075 1.05 1.25 427 23,900 298 0.635

4 1.10 1.39 167 9,350 130 0.600

5 0.10 1.05 1.275 445 24,900 318 0.707

6 1.10 1.46 179 10,000 146 0.668
7 0.15 1.05 1.525 474 26,600 405 0.796

8 1.10 1.80 197 11,000 198 0.780

9 0.20 1.05 2.24 520 29,100 652 0.865
10 1.10 2.88 206 11,540 332 0.865
1 0.25 1.05 4.36 575 32,200 1400 0.900
12 1.10 7.05 240 13,430 952 0.865

48
friction, the single riser can be replaced by a num-
ber of risers having the same height and total cross
sections, Likewise, the heat exchanger can be
replaced by a number of heat exchangers having
the same length of tubes and total number of tubes.

The principal results of these calculations-are
given in Fig. 1.5.2. The diameter of the riser
needed to provide natural convection, the length
and number of heat exchanger tubes, and the total
fuel volume are given as functions of riser height

PERIOD ENDING OCTOBER 31, 1958

and the temperature drop of the fuel in going through
the heat exchanger. Natural-convection power re-
actors of this size are characterized by high fuel
volumes and large numbers of heat exchanger tubes.
Therefore it is expected that the initial cost of a
natural-convection reactor would be higher than
that for a forced-circulation system, and the larger
number of heat exchanger tubes cast some doubt

as to whether the natural-convection system would
actually be more reliable than the forced-circula-
tion system.

UNCLASSIFIED
ORNL-LR-DWG 32561

L —=

HEAT
EXCHANGER

. /S-ft DIA

REACTOR

/

Fig. 1.5.1.
Molten-Salt Reactor.

Schematic Diagram of a Natural-Convection

49

NUMBER OF HEAT VOLUME OF FUEL

LENGTH OF HEAT

RISER

50

UNCLASSIFIED
ORNL-LR-DWG 32565

(x 10%)
/‘)'-\'7—7= H‘SOF
22 / }
o
= _— AT = 200°F
o g e
J— // AT = 250°F
16
(X 403) ‘
Q \%\ AT =475°%F
R 40 < ‘< A7 = 260°F
; s
S 30— AT = 225% = "k\_‘x
: o7 =225 = S e
§ AT = 250°F J |
\ ——
L 20 [
AT = 250°F
E 15 //
; / /AT = 225°%F
b;i P | ———"TAT = 200°F
a _ o
g 10 ///i’// AT =1T5F
T
g '/_/ //
/
5
6
: '_'-l—n_____
cr 5 ——‘.——_“_; R
= 4 e —— AT = 200°F —
) AT = 225°%F
AT = 25C°F
3 i
10 15 20 25 30 35 40 45

RISER HEIGHT (ft)

Fig. 1.5.2. Design Characteristics of a 576 Mw (thermal) Natural-Convection Molten-Salt Reactor.

Part 2
MATERIALS STUDIES

2.1. METALLURGY
W. D. Manly

D. A. Douglas

A. Taboada

Metallurgy Division

DYNAMIC CORROSION STUDIES

J. H. DeVan
J. R. DiStefano R. C. Schulze

The three-phase corrosion testing program, previ-
ously described, for which Inconel and INOR-8
thermal-convection and forced-circulation loops are
utilized, was continued, During the quarter, cor-
rosion studies were completed of two Inconel and
four INOR-8 thermal-convection loops. Examina-
tions were also completed of three Inconel forced-
circulation loops, which were the first loops oper-
ated in the third phase of the corrosion testing
program. Samples removed from an INOR-8 forced-
circulation loop that was temporarily shut down for
repair of a leak were also examined. A number of
new forced-circulation loop tests were started; the
present status of all such loop tests now in oper-
ation is given in Chap. 1.2 of this report.

1), H. DeVan, J. R. DiStefano, and R. S. Crouse, MSR
Quar. Prog. Rep. Oct. 31, 1957, ORNL-2431, p 23,

UNCLASSIFIED
T-15734

L0l0

\Q\ / .01 ]

013

Ql4

o R .018

: | sl 5]
_ | |1z
o~

» L
’ o~

Fig. 2.1.1. Hot-Leg Sample from Inconel Loop 1222
Which Circulated Salt 130 for 1000 hr ot 1350°F.

Inconel Thermal«Convection Loop Tests

One of the two Inconel thermal-convection loops
examined was operated as a supplement to the first
phase of corrosion testing of Inconel in contact
with fused-fluoride-salt mixtures. Loop 1222 circu-
lated salt mixture 130 (LiF-BeF2-UF4, 62-37-1
mole %) for 1000 hr at a maximum hot-leg tempera-
ture of 1350°F, and it was attacked to a depth of
3 mils in the form of intergranular voids (Fig. 2.1.1).
Previously,? loop 1178 had circulated the same salt
mixture for 1000 hr at 1250°F and was attacked to a
depth of 1 mil. Comparison of the results from the
two loop tests emphasizes the effect of operating
temperature on depth of corrosion in Inconel sys-
tems. The 100°F increase in the maximum operating
temperature caused an increase in the depth of at-
tack from 1 to 3 mils in the 1000-hr period.

The second Inconel thermal-convection loop (1207)
that was examined was the first that had operated
for an extended period of time. It circulated salt
mixture 12 (NaF-LiF-KF, 11.5-46.5-42 mole %) for
4360 hr at o maximum temperature of 1050°F. The
attack in loop 1207 was in the form of voids to a
depth of 2 mils.

INOR«8 Thermal«Convection Loop Tests

The four INOR-8 thermal-convection loop tests
which were completed concluded the first phase of
the corrosion testing program for INOR-8, The op-
erating conditions and chemical analyses of the
salts circulated in the loops are presented in Table
2.1.1. None of the loops showed measurable attack,
even though one loop (1213) operated for 3114 hr, A
hot-leg section of loop 1203 after exposure to salt
mixture 125 for 1000 hr at 1250°F is shown in Fig.
2.1.2. Comparable samples from the other INOR-8
loops were similar in appearance.

Inconel Forced«Circulation Loop Tests

The operating conditions and results of three
Inconel forced-circulation loop tests are presented
in Table 2.1.2, These three loops were the first

2), H. DeVan, J. R. DiStefano, MSR Quar. Prog. Rep.
June 30, 1958, ORNL-2551, p 57.

53

MOL TEN-SALT REACTOR PROGRESS REPORT

Table 2.1.1. Test Operoting Conditions and Chemical Analyses of Salts Circulated in INOR-8
Thermal-Convection Loops

Maximum loop temperature: 1250°F

Operatin Major Constituents Minor Constituents
Loop Salt e Salt Composition (wt %) (ppm)
No. No. Time (mole %) Sample Taken
(hr) ) Be Th Ni Cr Fe
1203 125 1000 53 NaF =46 Bel:2 -0.5 Before test 1,73 8.59 2.82 150 72 222
UF,~0.5 ThF, After test 1.82 10,5  2.90 68 235 195
1204 131 1000 60 LiF-36 Be F2-4 UF“ Before test 20.4 6.97 128 30 275
After test 20,6 7.15 <10 79 195
1213 128 3114 71 LiF=29 ’l'l'\F:4 Before test* 62 20 25 115
1221 128 1000 71 LiF=29 ThFA Before test 62.3 35 37 103
After test 61.3 28 70 245

*Results of after-test analyses not yet available.

operated in phase three of the corrosion testing

UNCTLSgg'sTEDI program. The loop designated CPR was the first
forced-circulation loop operated under molten-salt
reactor program conditions. |ts design provided for
the operation of separate sodium and salt circuits
through a U-bend heat exchanger. The loop com-

; .007 pleted 9148 hr of operation with salt mixture 122

™ . (NoF-ZrF4-UF4, 57-42-1 mole %) at a maximum
’ temperature of 1260°F and with the sodium at a

. 009 maximum temperature of 1120°F. Operation was

terminated because of a pump failure.* Metal-

010
lographic examination revealed moderate surface
; j - roughening and slight intergranular attack in the
. e heated leg of the molten-salt circuit. The
‘ . . maximum intergranular penetration was less than
; ' ' ' — 1.5 mils, as shown in Fig. 2.1.3. In the sodium
- 014 circuit, attack was limited to light su.face rough-
ening. No evidences of cold-leg deposits were
25 018 found in either the salt or sodium circuits.
- o The second and third loops described in Table )
a 2 N 2.1.2 also operated for extended time periods.
Fig. 2.1.2. Hot-Leg Sample from INOR-8 Loop 1203 3). L. Crowley, MSR Quar. Prog. Rep. June 30, 1958, 3

Which Circulated Salt 125 for 1000 hr at 1250°F, ORNL-2551, p 33,

After circulating salt mixture 126 (LiF-BeF ,-

UF4, 53-46-1 mole %) for 3073 hr at a maximum hot-
leg temperature of 1300°F, loop 9377-1 showed
moderate to heavy surface roughening and heavy
intergranular void formation to a depth of 5 mils,

as shown in Fig. 2.1.4. The third loop, 9377-2,

UNCLASSIFIED
T-15185

Fig. 2.1.3. Hot-Leg Sample from Salt Circuit of
Inconel Forced-Circulation Loop CPR Which Circulated
Salt 122 for 9148 hr at 1300°F.

PERIOD ENDING OCTOBER 31, 1958

revealed moderate intergranular void formation io
a maximum depth of 8 mils and an average depth
of 4 mils, as shown in Fig. 2.1.5. The loop op-
erated 3064 hr and circulated the salt mixture
130 (LiF-Ber-UF4, 62-37-1 mole %).

UNCLASSIFIED
T-15011

o
-
8

Fig. 2.1.4. Hot-Leg Sample from Inconel Forced-
Circulation Loop 9377-1 Which Circulated Salt 126 for
3073 hr ot 1300°F.

Table 2.1.2. Operating Conditions ond Results of Inconel Forced-Circulation Loop Tests

Maximum salt-to-metal interface temperature: 1300°F

Minimum bulk salt temperature: 1100°F

Temperature difference ocross loop: 200°F*

Salt flow rate: 2 gpm

Loop Fluid Composition Operating Time  Reynolds
Results
No., (mole %) (hr) Number
s CPR (primary circuit) 57 NaF-42 Zer—] UFA 9148 5,000 Intergranular penetration to a
depth of <1.5 mils
CPR (secondary circuit) Sedium 9148 97,700 Light surface roughening; no
cold-leg deposits
9377-1 53 LiF =46 Ber—l UF4 3073 1,600 Voids to a depth of 5 mils
9377-2 62 LiF=37 BeF2—] UF, 3064 3,000 Intergranular voids to a maxi-

mum depth of 8mils; average

depth, 4 mils

*The respective temperature differences across the fused-salt and sodium circuits of the loop were 100°F,

55

MOL TEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
T-14663

\ "" 2w " ol I/' 2 010

2 F /' - Ol
- » S

'~ / 258 b 2 0i2

C v - 013

014

. s

‘ L e 2184
>~ ¥ 2 x

. .‘ f - - ; -' "8"‘
e o

Fig. 2.1.5. Hot-Leg Sample from Inconel Forced-
Circulation Loop 9377-2 Which Circulated Salt 130 for
3064 hr at 1300°F.

Results of Examination of Samples Removed
from INOR+8 Forceds«Circulation
Loop 93541

On July 29, 1958, after successful completion of
3370 hr of operation, loop 9354-1 was temporarily
shut down to permit the repair of a small leak in a
lug adapter near the hottest section in the loop.
The cause of the leak was traced to a probable de-
fect in the material used in fabricating the loop
(see subsequent section of this chapter on ‘“Welding
and Brazing Studies’’). The loop was constructed
of ]/é-in.-OD, 0.045-in.-wall INOR-8 tubing and it
circulated salt mixture 126. The maximum loop
wall temperature was 1300°F and the minimum fluid
temperature was 1180°F. The salt flow rate was
2.5 gpm and the Reynolds number was 2000.

The damaged section, which was about 1 ft long
and which is shown in Fig.2.1.6, was removed from
the loop, and 2-in. samples were cut from the tubing
on both sides of the lug adapter. Metallographic
examination of these samples did not show evidence
of surface attack. However, the electrolytic etch
used to prepare the samples for metallographic ex-
amination produced a greater darkening of the grain

UNCLASSIFIED
T-15512

Fig. 2.1.6. Damaged Section of INOR-8 Forced-Circulation Loop 9354-1 Showing Location of Leak,
boundaries near the exposed surface than in the
interior, as shown in Fig. 2.1.7. This effect indi-
cates a faster etching rate of the boundaries near
the surface and is presumably due to a slight dif-
ference in the chemical composition of the region
near the surface. As may be seen in Fig. 2.1.8,
this surface effect is not apparent in as-received
samples of the tubing used for the loop. The loop
was repaired by replacing the damaged section,
and it has again been placed in operation.

Material Inspection
G. M. Tolson J. H. DeVan

Routine inspection of the INOR-8 material now
being received has revealed that the quality is as
good as that of Inconel and of stainless steels
made to ASTM standards. The high rejection rate
previously found for INOR-8 welded and drawn tub-
ing is felt to have been due to poor welding quality
of the heat of INOR-8 from which the tubing was
made. The rejection of seamless INOR-8 tubing, with
defects greater than 5% of the wall thickness used
as criteria for rejection, is running well below 10%.

UNCLASSIFIED | [
T-15580 | | w
I —
O
=

Fig. 2.1.7. Somple Taken from Damaged Section of
INOR-8 Forced-Circulation Loop 9354-1 Showing the
Effect of Electrolytic Etch on the Grain Boundaries
Near the Exposed Surface.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED | [, |
T-15578 v

Fig. 2.1.8. Sample of As-Received Tubing Used for
Loop 9354-1 Showing Absence of Effect of the Electro-
lytic Etch on the Grain Boundaries.

Fully inspected welds made on INOR-8 material
show a rejection rate of about 10%, which compares
favorably with the rejection rate of welds made on
Inconel and stainless steels. Evaluation tests were
made of thermocouple spot welds to INOR-8 tubing.
Although nondestructive testing did not reveal any
rejectable spot welds, metallographic examinations
showed microcracks in some spot welds.

GENERAL CORROSION STUDIES

E. E. Hoffman
W. H. Cook D. H. Jansen

Carburization of Inconel and INOR-8
by SodiumsGraphite Systems

D. H. Jansen

The sodium-graphite system, which is an effec-
tive carburizing medium for nickel-base alloys,*
was used to carburize Inconel and INOR-8 samples
for additional tests to determine whether the me-
chanical properties of these materials are detri-
mentally affected by various degrees of carburi-
zation. The carburizing system and the test pro-

4E. E. Hoffman, W. H., Cook, and D. H. Jansen, MSR
Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 54.

57
MOLTEN-SALT REACTOR PROGRESS REPORT

cedures used were described previously.® Results
of the tests completed thus far are presented in
Table 2.1.3. As may be seen, the data indicate
that the carburization treatment increases both

the yield and tensile strengths of Inconel and de-
creases the ductility at room temperature, Similar
room-temperature mechanical property data for the
carburized INOR-8 specimens showed that the yield
strength was increased, the tensile strength was
slightly decreased, and the ductility was greatly

SE. E. Hoffman, W. H. Cook, and D. H. Jansen, MSR
Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 59.

reduced. The change in each mechanical property,
especially the ductility, was dependent on the
amount of carburization that had occurred. There
was close agreement of all duplicate and triplicate
mechanical test values, that is, 2 x 102 psi vari-
ation for strength values and £1.5% for elongation
values. Inconel and INOR-8 specimens that were
exposed in the sodium-graphite system for 40 hr at
1600°F are shown in Fig. 2.1.9. Heavy carburi-
zation to adepth of 2mils was found on the Inconel,
and even heavier carburization was found on the
INOR-8.

Creep tests have also been conducted in argon at
1200°F on the Inconel and INOR-8 specimens that

Table 2.1.3. Results of Mechanical Tests in Room-Temperature Air of Inconel and INOR-8

Specimens Carburized in Sodium-Graphite Systems

Mechanical Properties

Exposure Exposure
. Specimen Yield Tensile Elongation in 2-in.
Temperature Period
(°F) (hr) Treatment Strength Strength Gage
(psi) (psi) (%)
Inconel
x 103 x 103
1200 40 Control? 241 80.8 40,1
Carburized 27.2 84.2 38.0
1400 40% Control 23.9 81.3 38.8
Carburized 27.0 83.9 36.8
400° Control 24.3 81.7 37.9
Carburized 30.9 92.9 27.0
1600 40€¢ Control 20.9 75.9 47.3
Carburized 28.5 91.5 25,7
INOR-8
1200 40 Control? 53.3 120.3 43.6
Carburized 53.1 119.6 43.8
1400 40% Control 53.7 122.0 42,6
Carburized 53.9 120.5 36.9
400° Control 53.2 120.7 43.0
Carburized 57.4 103.2 8.8
1600 40¢ Control 50.5 116.3 50.0
Carburized 52.8 98.5 7.8

%Control specimens exposed to argon at the conditions given,

bTesf values given for these specimens are averages of two tests,

c . .
Test values given for these specimens are averages of three tests,

58

¢

~ . “INCONEL

Etchant: copper regia. 500X.

were carburized for 40 hr at 1600°F in the sodium-
graphite system, and the results are presented in
Table 2.1.4. The results of tensile tests of similar
specimens and control specimens in air at 1250°F
are presented in Table 2.1.5. In general, the carbu-
rized specimens show lower ductility thon that of
the control specimens. Exceptions to this trend
were found, as indicated in Table 2.1.5, for some
of the carburized INOR-8 material and for a heavily
carburized Inconel specimen that had been exposed
for 40 hr at 1600°F in the sodium-graphite system;
the carburized pieces were more ductile than the
control specimens. These results were verified by
duplicate tests, and therefore the results of the
duplicate tests are also included in Table 2,1.5.

Carburization of INOR-8 by Fuel-130-Graphite
Systems

A comprehensive test in which INOR-8 (nominal
composition: 70% Ni-16% Mo-7% Cr-5% Fe-2%
other alloying additions) specimens were exposed
to (1) fuel 130 (LiF-Ber-UFa, 62-37-1 mole %)
containing a graphite rod, (2) fuel 130 without
added graphite, and (3) argon, all in the same test
container, at a temperature of 1300°F for 2000 hr,

PERIOD ENDING OCTOBER 31, 1958

lUNCLASSIFIED!
Y-26201

f:fi ,
«m_; ( 4

Fig. 2.1.9. Inconel and INOR-8 Specimens After Exposure to Sodium-Graphite System for 40 hr at 1600°F.

b
INOR-8

has been completed. The 40-mil-thick tensile-test
specimens used were then subjected to tensile
tests in order to determine whether the various
treatments had affected the mechanical properties.

Table 2.1.4, Results of Creep Tests of INOR-8 and
Inconel Control Specimens and Specimens Corburized
in a Sodium-Graphite System for 40 hr at 1600° F
1250°F
argon

18,500 psi

Test temperature:
Test environment:

Test stress:

Elongation in 2-in.
Time to Rupture

Specimen Goge
) (%)
INOR-8
Control 1734 23.8
Carburized 4031* 9.0*
Inconel
Control 38 21.2
Carburized 160 47.5

*Still in test,

59

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.1.5. Results of Mechanical Tests in 1250°F Air of Inconel and INOR-8

Specimens Carburized in Sodium-Grophite Systems

Mechanical Properties

Exposure Exposure
. Specimen Yield Tensile Elongation in 2«in.
Temperature Period
(°F) (hr) Treatment Strength Strength Gage
(psi) (psi) (%)
Inconel
x 103 x 103
1200 40 Control* 50.3 38.5
Carburized 48,6 33.5
1400 40 Control * 46.7 36.5
Carburized 49.9 23.0
400 Control* 16.4 45.2 26.5
Carburized 22.0 50.7 21.0
1600 40 Control* 11.3 42.4 16.5
1.4 18.5
Carburized 20.5 53,9 24.8
51.1 28.5
INOR-8
1200 40 Control* 70.2 17.0
Carburized 74.4 19.0
1400 40 Control* 74.6 18.5
Carburized 76.8 19.5
400 Control* 36.8 70.3 16.5
74.5 20.5
Carburized 42.4 82,9 18.0
85.2 22,0
1600 40 Control* 33.8 68.9 18.5
70.6 6.5
Carburized 40.5 82.8 13.0

*Control specimens exposed to argon ot the conditions indicated,

The specimens were also examined metallographi- system contained 980 ppm carbon. These data and
cally, and, as shown in Fig. 2.1.10, no carburiza- data obtained from carbon analyses of the fuel are
tion was found on the surface of the INOR-8 speci- given in Table 2.1.6.
men that was exposed to the fuel-graphite system. Room-temperature mechanical tests showed the
Chemical analyses and mechanical tests, how- INOR-8 specimen exposed to the fuel-graphite sys-
ever, indicated that carburization did occur. Carbon  tem to have higher yield and tensile strengths and
analyses of 3-mil surface cuts showed that the lower elongation than the INOR-8 specimen that
INOR-8 specimen exposed to fuel 130 withoutadded  was exposed only to argon and given the same heat
graphite contained 640 ppm carbon, while a similar treatment. The same trends were apparent in me-

cut of an INOR-8 piece exposed to the fuel-graphite  chanical tests at 1250°F. Data obtained in the

60

PERIOD ENDING OCTOBER 31, 1958

? - \J
\ 3 <
Q\ -
\ B ¢
\
¥ ° . I
- -
- e—ea- '.'- ' TN F - 2 il
| i -
1 .002 ’
' \
] . - /
. »’
-} .
| . '
t ’
003
5 » SRt R
w - . e / i
/.
o [ 5 -
- P @ 2 " -]
- = . 'O Y.
(0) e UNCLASSIFIED | © (b )" - . UNCLASSIFIED
Y-26647 ’ | 004 " 2 Y-26646

Fig. 2.1.10, INOR-8 Specimens After 2000-hr Exposures at 1300°F to (a) Fuel 130 Containing a Graphite Rod and
(b) to Fuel 130 Without Added Graphite, Etchant: glyceria regia. 1000X, Reduced 9%.

Table 2,1.6. Carbon Analyses of INOR-8 Specimens and Fuel 130 Mixtures Used in Tests
of the Carburization of INOR-8 in a Fuel-Graphite System at 1300° F for 2000 hr

Specimen Analyzed

Carbon Content (ppm)

Before Test After Test
Fuel 130 exposed to graphite and INOR-8 220 190
Fuel 130 exposed only to INOR-8 180 150
INOR-8 exposed to graphite and fuel 130 160-170" 980
INOR-8 exposed only to fuel 160-170* 640

*Carbon analysis of as-received INOR-8 from heat SP-16.

room-temperature and high-temperature mechanical
tests are presented in Table 2.1.7. As in the me-
chanical tests of specimens carburized in a sodium-
graphite system, the decrease in elongation is an
indication that carburization occurred, The fact
that the tensile strength of the INOR-8 specimen
exposed to fuel 130 was greater than that of the
specimen heat treated in argon can be explained
by the decrease in carbon concentration of the fuel
exposed to the INOR-8. The INOR-8 in this case
probably acted as a sump and picked up carbon
from the fuel, This would account for the decrease

in the carbon content of the fuel and the increase
in mechanical strength of the INOR-8 specimen ex-
posed to it.

UO, Precipitation in Fused Fluoride Salts
in Contact with Graphite

W. H, Cook

A test of the compatibility of fuel 130 and graph-
ite at 1300°F in @ vacuum (<0.1 i) was terminated
at the end of a 500-hr period because of UO, pre-

cipitation from the fuel. In preparation for the test,

61

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.1.7.

Results of Mechanical Tests of INOR-8 Specimens Exposed to Fuel 130 With and Without

Graphite and to Argon at 1300°F for 2000 hr

Yield Strength
(psi)

Specimen Treatment

Tensile Strength Elongation in 2-in, Gage

(psi) (%)

Room-Temperatute Tests*

X 103

51.5
51.5

Exposed to fuel 130 and graphite

40.0
39.9

Exposed only to fuel 130

39.0
31.3

Exposed to argon

Tests at 1250°F

Exposed to fuel 130 and graphite *x

Exposed only to fuel 130 b

Exposed to argon **

% 103
116.8 47.0
116.4 46.5
107.3 56.0
106.5 56.0
104.2 55.0
103.4 56.0
73.0 19.5
66.5 30.0
66.6 33.8

*Values obtained in each of duplicate tests are listed.

**Yield strengths were not obtained because the results might have been obscured by a notch effect of the

extensometer,

a machined CCN graphite crucible, nominally, 2 in.
oD, 0.43 in. ID, 31/2 in. deep, and 4]/2 in. long, with
an average bulk density of 1.90 g/cm 3, was de-
gassed at 2370°F for 5 hr in a vacuum of 1 to 3 p,
cooled to room temperature, and blanketed in argon.
The crucible was exposed to room atmosphere for
approximately 3 min in order to weigh it. During
all other preparations for the test, the crucible

was either sealed under an argon atmosphere or
was under a vacuum of <0,1 . Sufficient fuel 130,
13.70 g, to fill the crucible to a depth of 3 in. at
1300°F was cast into the graphite. The loaded
crucible was tested in an inert-arc welded Inconel
container.

Radiographs were used to monitor the location of
the fuel in the graphite so that the test could be
terminated if the fuel threatened to penetrate the
crucible walls. Such a radiograph revealed at the
end of 500 hr a Q.1-in.-high disk in the bottom of
the graphite crucible that was relatively opaque to
the x-rays {see Fig. 2.1.11). The remainder of the
fuel was less opaque to the x-rays than it had been
at the beginning of the test. Subsequent examina-
tion of the disk revealed that it contained U02; the

62

disk contained ~40 wt % U as compared with 7.03
wt % U in the as-received fuel. The uranium con-
tent in the disk amounted to 28% of the total amount
of uranium originally in the fuel 130 charge; stated
in another way, the uranium concentration of the
fuel above the disk had decreased to 59% of its
original value.

Petrographic examinations of the fuel used for
this test did not reveal oxide contamination. The
fuel remaining in the batch from which the test
quantity was taken is being analyzed for oxygen.
This same source of fuel was used in a 2000-hr
static test at 1300°F, described in the preceding
section, in which the fuel was exposed only to
INOR-8. A radiographic examination of the fuel at
the end of that test did not reveal any precipitation
of urantum.

Different grades of graphite and different fuels
are being used in further tests of the compatibility
of fuels and graphite. Radiographic examinations
of systems now being tested, in which TSF graphite
(average bulk density, 1.67 g/cm3) is used to con-
tain the fuel 130, have also revealed uranium pre-
cipitation, but the quantities precipitated are less
in TSF graphite than in CCN graphite. Radiographic

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
PHOTO 45742

B ———VACUUM (<04 p) — &

(@)

3 2. <cALED EVACUATION TUBE BBE—"1

pa——"THERMOCOUPLE WELL

| 4PROTECTIVE CONTAINER =¥
(INCONEL)

CCN GRAPHITE CRUCIBLE—g

FUEL 130 [7.03 wt % U]

UO,-FUEL 130 DISK
[~40 wt PoU]

i‘——1 in.—-—‘

(0)

Fig. 2.1.11. Radiographs of CCN Graphite Crucible Containing Fuel 130 (o) Before Test and (b) After 500 hr at

1300°F.

examinations of a CCN graphite crucible which con-
tains NaF-ZrF4-UF4 (50-46-4 mole %, fuel 30) and
which is also being tested under the conditions de-
scribed above, have shown no uranium precipitation
in more than 1500 hr. Thus it appears that fuel 130
is more sensitive to contamination than fuel 30 is.
It is thought that the brief exposure of the graph-
ite crucibles toroom atmosphere during the weighing
operation may have contaminated them with water
vapor and oxygen and that the subsequent reaction
of these contaminants with the fuel caused the UO2

On the basis of the degassing and
handling techniques used, however, such rapid con-

precipitation.

tamination appears to be in contradiction to the ob-
servations of others.® The greater quantity of
uranium precipitation in the CCN graphite crucible
than in the TSF graphite crucible may have been

SFirst Nuclear Engineering and Science Congress,
Cleveland, Ohio, 1955; selected papers edited by D. J.
Hughes, S. McLain, and C. Williams, Problems in Nuclear
{:’nginf’pring, Vol. 1, p 142, Pergamon Press, New York,

957.

63

MOLTEN-SALT REACTOR PROGRESS REPORT

the result of impurities in the graphite. The CCN
grade is less pure than the TSF grade, which is a
gas-purified product.

Metallographic examinations of the graphite
crucibles used in these tests have shown no attack
and no penetration of salt into the pore spaces in
the graphite. Experiments are now in progress to
determine (1) the extent to which salts can be
forced, by pressure, into the pore spaces of various
grades of graphites and (2) the physical effects on
the graphite of repeated cycles of freezing and
melting of salts trapped in the pore spaces.

PrecioussMetalsBase Brazing Alloys in Fuel 130
D. H. Jansen

Five silver-base brazing materials ranging in com-
position from pure silver to alloys with 42 wt % Ag
were tested in static fuel 130 for 500 hr. It was ex-
pected on the basis of thermodynamic considerations
that silver would show good corrosion resistance to
beryllium and lithium fluorides, but, as shown in
Table 2.1.8, deep subsurface voids (as deep as 50
mils in the case of pure silver) were observed in all
the silver-containing alloys.

A gold-base alloy (75% Au-20% Cu-5% Ag) was
also tested in fuel 130 under the same conditions as
those used in the tests of the silver alloys. No cor-
rosion was found on this alloy after the test, Three
other gold-base alloys are being prepared for testing

in fuel 130,
MECHANICAL PROPERTIES OF INOR-8B

R. W. Swindeman D. A. Douglas

Short time (less than 10,000 hr) creep data for
INOR-8 obtained through tests in air conducted by
the Haynes-Stellite Company and tests in fuel 107

(NaF-LiF-KF-UF,, 11.2-45.3-41-2.5 mole %) con-
ducted at ORNL are presented in Fig. 2.1.12., Speci-
mens tested by Haynes-Stellite were annealed 8 min
at 2150°F; ORNL specimens were annealed for 1 hr -
at 2000°F. Cross plots of log t vs 1/T, where ¢t is
the time to 1% strain and 1/7T is the reciprocal of
the absolute temperature, yielded the straight lines
shown in Fig. 2.1.13. These parallel straight lines
indicate that the equation proposed by Dorn” would
best represent the time-temperature-stress relation-
ship of this material. The slopes of these lines
yield an activation energy for creep of 76,700
cal/mole:°K for INOR-8 sheet in the fused-salt
environment. A plot of various stress functions
against In [ (1/7) e78:700/RT] rayealed that the

]
stress dependence is of the form ¢B9 ", where o
is the stress and B is a constant, A plot of the
data based on the equation
(/1) = M 2B’ /3 o~76,700/RT

to represent the time-temperature-stress relation-
ship, where M is a constant, is shown in Fig.2.1.14.

7). E. Dorn, J. Mech, and Phys. Solids 3, 85 (1954).

UNCLASSIFIED

105 ORNL-LR-DWG 34515
— e e e e e e e e e S AT
e = ! e e =i
0 A O A R A
B = - - o~ MOO°F |
@ L. b 1! i 447 Tt 21T Py |13 N
a N i — - \w. I L
; , o 1 \\D"-;-..__i_‘ : ; \qh.::‘lf 3000 39“0 F
B [ M SR S e e S TP B -\F -
W L TR S T = L EREN
N R = - 5000F n
L[ ® ORNL(FUEL107) *~' oo TiH— |~ | )
b © HAYNES STELLITE (AIR) =10 Lol | {1
[———EXTRAPOLATED ‘ 1700 fi 1 1
IR H ' [ : S i
10° | PP - Ll S eI 2
(OB 1.0 10 100 1000 10,000

¢, TIME TO 1% STRAIN hr)

Fig. 2.1.12. Time to 1%Strain vs Stress for INOR-8
Tested at Various Temperatures.

Table 2,1,8, Results of Corrosion Tests of Silver- and Gold-Base Brazing Alloys Exposed
to Static Fuel 130 for 500 hr at 1300°F

Braze Material

Metallographic Results

Pure silver

90% Ag-10% Cu
50% Ag-33.3% Au=146.7% Cu

42% Au-40% Ag—18% Cu—0.6% Zn

75% Au-20% Cu~-5% Ag

Rather uniform, heavy attack to a depth of 16 mils

Spotty, heavy attack; stringers to a depth of 50 mils
in some places

Spotty attack to @ maximum depth of 5 mils

Uniform ottack to a depth of 15 mils

No attack

0*

5

2

_10®
£
=

I 5
‘_
wn
53
o

=2
i
=
fual

= 10°

5

2

UNCLASSIFIED
ORNL-LR—DWG 34546

10
085 090 0.95

100 1.05
10%/T (°K)

110 145

Fig. 2.1.13. Time to 1% Strain vs Reciprocal of
Absclute Temperature at Constant Stress for INOR-8
Tested in Fuel 107,

38

36

(5]
b

n [“/,.)eTS.TOO/RT]
W
[\

L
Q

28

1.20

UNCLASSIFIED
ORNL- LR—-DWG 34547

t1100°F
1200°F
1300°F

rpOSD

/

1500°F

10,000 hr AT 1200°F

o]

25

30

Fig. 2.1.14, Effect of Stress, g, Vuriations on

35

Parameter Representing Temperature, T, and Time, ¢,
to 1% Strain for INOR-8 Sheet Tested in Fuel 107,

PERIOD ENDING OCTOBER 31, 1958

A horizontal line, representing the value of
In [(1/1) eDAH/RT] yhen T is 1200°F and ¢ is
10,000 hr is included on Fig. 2.1.14. The stress
corresponding to 1% strain as given by this curve
is approximately 9600 psi. Extrapolation of the log
curve for 1200°F in Fig. 2.1.12 to 10,000 hr yields
about 11,000 psi. Thus it would appear that a stress
value of 9600 psi can be used as a conservative
estimate of the metals creep strength for the 10,000-
hr life. The results of longer tests may permit this
value to be revised towards the 11,000-psi figure.
An investigation is also being made of the vari-
ation in creep properties of specimens subjected to
identical conditions of stress and temperature,
Creep curves obtained in this investigation are
presented in Fig. 2.1.15. The data was obtained
in air at 1250°F in duplicate tests for two different
stresses, 20,000 and 25,000 psi. The greatest dif-
ferences appear to be in the rupture life and elonga-
tion rather than the creep rate. Data for all tests
completed thus far on INOR-8 in air are compiled
in Fig. 2.1.16. The variation of the

In [(1/1) e76:700/RT)

term at each stress level is an indication of the
reproducibility of the results. The Haynes-Stellite
data were obtained in short-time tests (times of

less than 100 hr), but the ORNL data include values
obtained at times up to 5500 hr.

The variations in the properties of samples taken
from different heats of the alloy are also of interest.
Creep curves obtained at 1300°F and 20,000 psi for
three different heats of INOR-8 (SP-16, SP-19, and
M-1) are shown in Fig. 2.1.17. The specimen from
heat M-1 was appreciably stronger than those from
the other two heats, probably because of the higher
carbon content (0.14%) of heat M-1. The strengths
of heats SP-16 and SP-19, with 0.02 and 0.06%

carbon, respectively, were similar.

INFLUENCE OF COMPOSITION ON PROPERTIES
OF INOR-8

T. K. Roche

The precipitation of intermetallic compounds in
nickel-molybdenum base alloys, such as Hastelloy
B, has been shown to be deleterious to the tensile
ductility of these materials.® Inasmuch as the upper

aR. E. Clausing, P. Patriarca, and W, D. Manly, Aging
]C;Jar;zcteristics of Hastelloy B, ORNL-2314 (July 30,
57}.

65
MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL—LR—DWG 34548

10
8
/*
E . //
Z /
o4 e
L
Q QS\ /
2 ¥ SPECIMEN RUPTURED
o]
O 200 400 600 800 1000 1200 1400 1600 1800 2000 2200

TIME (hr)

Fig. 2.1.15, Creep Curves for INOR-8 Tested in Air at 1250°F,

UNCLASSIFIED
ORNL-LR-DWG 34549
a8 °

3 /

¢ (ORNL DATA
o HAYNES STELLITE DATA q
34

In [( I/,)em,?oo/fir]
18]
n

30 2

28 z

26

Fig. 2.1.16, Effect of Stress, 0, Variations on
Parameter Representing Temperature, T, and Time, ¢,

to 1% Strain for INOR-8 Sheet Tested in Air.

limit of molybdenum in INOR-8 has presently been
set at 18%, which approaches the solubility limit of
molybdenum in nickel, it was considered to be im-
portant that the structural stability of the alloy be
determined with respect to the maximum amount of
each major alloying element called for by the speci-

fication: 18% Mo—-8% Cr-5% Fe—balance Ni and

66

UNCLASSIFIED

12 ORNL-LR-DWG 34520
¥

. L

8 /.
2 ~
3 /
<1
® & 2, -
[ c® © -
’ /41} /

4 - N\}—/

2 ]

_// % SPECIMEN RUPTURED
. A
o} 100 200 300 400 500 600 700 800 300 4000

TIME {hr)

Fig. 2.1.17. Creep Curves for Three Heats of INOR-8
Tested in Fuel 107 at 1300°F and 20,000 psi.

small quantities of other elements. It was also de-
sired to determine the approximate locations of
phase boundaries that might exist near the speci-
fied composition of INOR-8 so that the suscepti-
bility of the material to intermetallic compound for-
mation could be established.

An investigation of the stability of alloys with
compositions near the specified composition has
been initiated. Eleven alloys with the compositions
given in Table 2.1.9 were prepared by vacuum in-
duction melting. The alloys were then rolled to

0.020-in.-thick strip and treated at 2200°F in a hy-
drogen atmosphere to eliminate as much residual
carbon as possible in order to reduce the potential
for carbide precipitation. The alloys were subse-
quently remelted and rolled, and specimens are
presently undergoing 100- and 1000-hr aging treat-
ments over a temperature range of 900 to 1800°F.
Metallographic examinations of the specimens will

Table 2,1.9. Nominal Compositions of Nickel-Molybdenum
Alloys Prepared for Structural Stability Studies

Nominal Composition (wt %)

PERIOD ENDING OCTOBER 31, 1958

be carried out to determine the presence or absence
of precipitates in the microstructures.

HIGH-TEMPERATURE STABILITY OF INOR-8
H. Inouye

Embrittiement studies of two commercial heats of
INOR-8 are being conducted. Previously, aging
heat treatments for times up to 2000 hr in the tem-
perature range 1000 to 1400°F were shown to cause
no significant changes in the tensile properties of
the material.” The two heats being studied contain
different amounts of carbon, however, and they have
significantly different properties, as shown in

Alloy Tables 2.1.10 and 2,1.11.
Designation Mo Cr Fe Ni The effect of a 5000-hr heat treatment was deter-
mined during the quarter, and the results for speci-
VT-143 18 5 bal mens from the two heats are compared with similar
VT-144 18 10 bal data for annealed specimens from the two heats in
VT-145 18 4 7.5 bal Tables 2.1.10 and 2,1.11, As may be seen, aging
VT-146 18 5 10 bal for 5000 hr did not significantly change the tensile
VT-147 18 6 2 bal properties. |t is therefore concluded on the basis
VT-148 18 6 5 bal of these tensile tests that INOR-8 does not exhibit
VT-149 18 8 5 bal instabilities that would be detrimental in service
VT-150 18 8 8 bal for extended times in the temperature range 1000
VT-151 18 10 bal to 1400°F.
VT-152 18 10 4 bal
VT-153 18 10 10 bal °H. Inouye, MSR Quar, Prog. Rep. June 30, 1958,
ORNL.-2551, p 67.
Table 2,1,10, Tensile Properties of INOR-B Specimens of Heat M-1
Compositien: 16.2% Mo—-7.47% Cr—6.1% Fe—-0.14% C—bal Ni
Specimen Annealed and Aged
Specimen Annealed 1 hr at 2100°F
Test 5000 hr at Test Temperature
Temperature Ultimate Tensile Yield Strength at Elongation Ultimate Tensile
o Strength 0.2% Offset Elongation
("F) ) i (%) Strength
(psi) (psi) (psi) (%)
Room 117,100 51,900 39
1000 101,000 34,200 40 101,800 33
1100 102,600 37,600 37 88,000 19
1200 83,100 38,400 16 78,600 13
1300 70,500 38,400 18 74,800 14
1400 63,900 38,500 14 63,300 15

67

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.1.11. Tensile Properties of INOR-8 Specimens of Heat SP-19
Composition: 16.6% Mo=~7.43% Cr-4.83% Fe—-0.06% C—bal Ni

Specimen Annealed and Aged

. o
Specimen Annealed 1 hr at 2100 F 5000 hr at Test Temperature

Test
Temperature Ultimate Tensile Yield Strength at Elongation Ultimate Tensile ‘

° Strength 0.2% Qffset Elongation
(VF) ) . (%) Strength

(psi) (psi) (psi) (%)
Room 114,400 44,700 50
1000 93,000 28,300 46 93,200 51
1100 93,000 28,900 50 85,000 23
1200 82,400 27,500 37 81,600 33
1300 69,900 78,000 24 73,800 25
1400 61,800 26,200 21 64,300 20
WELDING AND BRAZING STUDIES

P. Patriarca
oRNL LR DWG 34521
INOR-8 Weldability 100 T
80 DUCTILITY .

G. M. Slaughter

Mechanical Properties of Weld Deposits. —Studies
of the mechanical properties of INOR-8 weld de-
posits were continued. A vacuum-melted experi-

mental heat of this material (ORNL heat 30-38)

was found to have room- and elevated-temperature 140,000

60 ‘ /]f/

\LJINCONEL il
40 B HASTELLOY B

HASTELLOY W

ELONGATION IN {in. (%)

properties comparable to those of commercial heats

fabricated by Haynes-Stellite and Westinghouse, !0 o=

and the ductility at 1300°F and above was still ‘ZO’OOO:: STRENGTH

marginal. Thus conventional vacuum-melting did - —
not overcome the ductility problem. In order to de- 100,000

termine the influence of carbon on the mechanical o= ::\ ]

HASTELLOY W |

R
N
N\ HASTELLOMQ\“

properties of INOR-8 weld metal, weld test plates
were constructed with the use of a high-carbon

(0,16% C) filler metal. The tensile strengths of \T\\ ;\
these welds were significantly higher than those of 60,000 T — ™~
conventional INOR-8 welds, but the room-tempera- o ‘T\ M
ture ductilities decreased considerably. The weld |NcIONEL—:£\
metal properties of both Hastelloy B and Hastelloy 70:000 [

80,000

TENSILE STRENGTH {psi)

W were also determined in order to obtain compara-

tive data for other nickel-molybdenum alloys. 20,000 |—
Plots of the as-welded tensile strengths and duc- . |
tilities of INOR-8, Hastelloy B, Hastelloy W, and " i

0 l | i ] | |

Inconel weld metal are shown in Fig. 2.1.18. |t ROOM  #00 {200 1300 1400 1500 1600
may be seen that the tensile strength of INOR-8 TEST TEMPERATURE (°F)

106 M. Slaughter, MSR Quar. Prog. Rep. June 30, Fige 2.1.18, Mechanical Properties of Weld Metal in
1958, ORNL -2551, p 69. As-Welded Condition.

68
weld metal is appreciably lower than that of either
Hastelloy B or Hastelloy W weld metal at ali tem-
peratures studied. The comparative data for Inconel
reveal that INOR-8 is superior in strength, particu-
larly at 1500°F.

Aging studies on INOR-8 weld deposits were also
conducted to determine the influence of time at the
service temperature on the mechanical properties.
The age-hardenable alloys, Hastelloy B and
Hastelloy W, were tested to provide additional com-
parative information. Samples were aged at 1200
and 1500°F for periods of 200 and 500 hr and the
room-temperature mechanical properties after aging
are shown in Fig. 2.1.19. Hastelloy B, as would
be expected, is subject to severe impairment in

UNCLASSIFIED
ORNL- LR—DWG 34522

100
=
= 80 — - ——
= BUCTILITY
z 60 | -
S
= W0
>
2
= =
_J =
VU] i g1 N
180,000
STRENGTH AS-WELDED
AGED AT {200°F
166,000 =1 FOR 200 hr
' = 7] AGED AT 1200°F
= FOR 500 hr
' AGED AT 1500°F
140,000 FOR 200 hr
AGED AT 1500°F
FOR 500 hr
120,000
s
(o
T
5 100,000
p=4
Ll
o
=
w
% 80,000
wn)
=
L
-
60,000 |-
40,000 |-
20,000
EN x ’ e

HASTELLOY B HASTELLOY W INOR-8 INCONEL
Fig. 2.1.19. Room-Temperature Mechenical Properties

of All-Weld-Metal Specimens.

PERIOD ENDING OCTOBER 31, 1958

ductility upon aging, particularly at 1200°F (duc-
tility reduced from 24 to 3% after aging for 500 hr).
Hastelloy W weld metal also exhibits a marked re-
duction in room-temperature ductility after aging.
INOR-8, on the other hand, maintains its good duc-
tility,

Mechanical property studies of aged specimens
were also conducted at the aging temperature in an
effort to more completely determine the behavior of
these materials at elevated temperatures. The data
obtained at 1200°F are shown in Fig. 2.1.20, and
the data obtained at 1500°F are shown in Fig.
2.1.21. Aging at 1200°F definitely reduces the
ductility of Hastelloy B and Hastelloy W at 1200°F,
whereas the ductility of INOR-8 is slightly im-
proved. At 1500°F, a significant ductility increase

UNCLASSIFIED
ORNL —LR-DOWG 34523
100

80 J— R [ —_— — . p— . — _ J— - J—
DUCTILITY

ELONGATION IN tin. (%)
o
o
T

160,000
STRENGTH

AS-WELDED

AGED AT 4200°F

FOR 200 hr .
AGED AT {200°F

FOR 500 br

140,000 |~

120,000 -

L]

100,000 r

80,000 r

60,000 —

TENSILE STRENGTH {psi)

40,000 |-

20,000 [ -

HASTELLOY W

|

HASTELLOY B INOR-8 [INCONEL

Fig. 2.1.20. Mechanical Properties of All-Weld-Metal
Specimens at 1200°F,

69

MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-0WG 34524

.. 100

8

s 80 o oDucTILITY _

2 60

=

=

e 40 |— — —— AR —

<1

2

Z 20 p— —— —]

- BN

o ‘
80,000

AS-WELDED
AGED AT 1500°F

% 600 ] FOR 200 hr

2 60,000 AGED AT 1500°F

he FOR 500 hr

&) S

=4 SN

& o

x 40,000 f—

W

wl

]

& 20,000 | —F

[ s

?:: N

HASTELLOY B HASTELLOY W INCR-8 [INCONEL

Fig. 2.1.21. Mechanical Properties of All-Weld-Metal
Specimens at 1500°F.

is evident for INOR-8, probably as a result of partial
spheroidization of the eutectic carbides.

The influence of carbide spheroidization is being
studied further, and high-temperature stabilization
treatments of weld metal are under investigation.
Mechanical property studies of INOR-8 welded
joints, both in the as-welded and aged conditions,
are also being conducted,

Filler Metal Modification Studies. ~ Since the
ductility exhibited by INOR-8 weld metal at tem-
peratures of 1300°F and higher is marginal, a pro-
gram has been initiated to determine the factors
that influence ductility and to develop procedures
for improving it. Three heats of INOR-8weld metal
have all shown approximately the same ductility,
that is, 10% or less in a 1-in. gage length in the
1300 to 1500°F temperature range. Hastelloy W,
however, has a minimum elongation of 20% in this
range.

Personnel of the International Nickel Company
have indicated that they think the trouble probably
results from trace elements in the 0,009 to 0.0001
wt % range, The use of proper addition agents
during the melting of the ingot and the proper timing
of the additions are extremely important in removing
or neutralizing these trace elements. Accordingly,

70

melts containing recommended quantities of manga-
ese, silicon, aluminum, titanium, magnesium, and
boron have been made. The ingots are being fabri-
cated into filler wire for the preparation of all-weld-
metal specimens for room- and elevated-temperature
testing. Heats of filler metal have also been made
that contain additions of a commercially available
aluminum-titanium-nickel master alloy.

One experiment was conducted which verifies the
premise that minor modifications in filler metal
composition can result in major changes in elevated-
temperature mechanical properties. A special ingot
of INOR-8 filler metal was prepared in which most
of the nickel and all the chromium were added as
Inconel inert-arc welding wire. This Inconel wire,
designated Inco No. 62 (efongation at 1500°F of
70%), also contains 2% columbium to prevent hot
cracking in the weld deposit., The mechanical
properties of this modified INOR-8 weld metal in
the 1200 to 1500°F temperature range were superior
to those exhibited by conventional INOR-8, as is
shown in Table 2.1.12.

As was previously mentioned, the improvement in
ductility of INOR-8 weld metal upon aging appeared
to be associated with the spheroidization of eutectic
carbides. Therefore special heats of INOR-8 having
no carbon and 0.03% carbon have been prepared and
will be fabricated into filler wire as a means of de-
termining the welding characteristics and mechani-
cal properties of lower-carbon-content weld deposits.

Results of Examination of INOR«8 Forceds«Circus
lation Loop Weld Failure, — An INOR-8 heater lug
section of forced-circulation loop 9354-1 failed
during service. The component and the location of
the failure are shown in Fig. 2.1.22. Metallographic
examinations of the failed area and adjacent areas
revealed profuse cracks in the INOR-8 adapter near
the lug-to-adapter weld, with the cracks resembling
those frequently observed in the heat-affected zones
of the crack-sensitive heat SP-16 material.!!

Chemical analyses of the adapter material re-
vealed it to be a high-boron, low-carbon heat of
INOR-8 and confirmed the suspicion that heat
SP-16 material had been inadvertently substituted
for the recommended heats of INOR-8. It is thought
that the cracks initiated during welding and propa-
gated to failure during service.

Welding of INOR-8 to Other Metals. — An evalu-

ation of Inco-Weld *A’’ wire was made to determine

”P. Patriarca and G. M. Slaughter, MSR Quar. Prog.
Rep. Oct. 31, 1957, ORNL-2431, p 18.

PERIOD ENDING OCTOBER 31, 1958

Toble 2,1.12, Comparison of Mechanical Properties of INOR-8 and Modified INOR-8 in As-Welded Condition

Test Temperature
|
\
\
|

' Tensile Strength Elongation in 1-in,
i (OF) Fl“ef Meful (Psi) (%)
Room INOR-8 116,000 38
- Modified INOR-8 113,000 44
1200 INOR-8 73,000 18
Modified INOR-8 77,000 30
1300 INOR-8 64,000 10
Modified INOR-8 72,000 22
1500 INOR-8 53,000 6
| Modified INOR-8 59,000 15

UNCLASSIFIED
Y-26847

Fig. 2.1.22. INOR-8 Heater Lug Section of Forced-Circulation Loop 9354-1 That Failed During Service. Arrow

indicates point of failure.

its suitability as a filler material for weldments
with INOR-8. The study'? indicated that the as-
welded ductility of the metal in short-time tests is
satisfactory up to 1300°F but becomes marginal at
: 1500°F. Significant impairment of the room- and
elevated-temperature ductility was found after aging

at 1200°F for 500 hr.

125, M. Slaughter, Mechanical Property Studies on
INCO-Weld **A’* Wire Filler Metal, ORNL CF-58-10-30
(Nov, 5, 1958).

Brazing of Thick Tube Sheets for Heat Exchangers
G. M. Slaughter

The construction of heat exchangers for molten-
salt power reactor applications poses a variety of
problems. For example, it is probable that the tem-
perature and pressure conditions in some of the heat
exchangers will necessitate the use of relatively
thick tube sheets (thicknesses of 4 to 6 in. or
greater). Present practice and fabrication proce-
dures for constructing high-integrity tube-to-tube

71

MOLTEN-SALT REACTOR PROGRESS REPORT

sheet joints in sections of this magnitude may not Inconel tube sheet with Coast Metals No. 52 alloy.
be adequate for a molten-salt environment. The un- Brazing was performed at 1920°F with a 1 hr hold
avoidable notch obtained in the conventional welded  at temperature. A 600°F per hr rate of temperature
tube-to-header joint is undesirable in several ways.  rise was used in this experiment, and the sample -
First, it may act as a starting point for crack propa-  was furnace-cooled to room temperature. The weld ‘
gation in the welds during cyclic service. Also, the side of the tube sheet was also trepanned to mini- |
notch may act as a crevice and thereby increase the  mize stresses during welding, As may be seen in "
possibility of stress-corrosion cracking. It is evi- Fig. 2.1.24, a photograph of the brazed specimen,
dent that the technique of back-brazing of the tube-  excellent filleting was obtained.
to-header joints would remedy these conditions by
eliminating the notch. Brazing would also improve
heat transfer and decrease the possibility of a leak
through a defective tube-to-header weld, since the
leak-path would be significantly increased.
Although considerable experience exists inbrozing
small diameter tubes into relatively thin tube sheets,
little is known of the problems associated with
adapting these techniques to heavy sections. Ac-
cordingly, a preliminary experiment has been con-
ducted to determine the feasibility of back-brazing
joints in thick tube sheets. A brazing procedure
was utilized which had proved useful in fabricating
simulated components containing ]]/Q-in.-thick tube

UNCLASSIFIED
Y-26383 |

sheets.

As may be seen in Fig. 2.1.23, the brazing alloy
is preplaced in annular trepans in the tube sheet
and is fed to the joint through three small feeder
holes. Melting of the brazing alloy therefore does
not take place until the joint has attained the tem-
perature where wetting can occur,

In an effort to extend this technique, a speci-
men was prepared that consisted of a 5/B-in.-OD,
0.065-in.-wall Inconel tube brazed into a 5-in.-thick

‘ UNCLASSIFIED
ORNL-LR-DWG 27066R

_—~WELD SIDE
— R e

> TUBE SHEET

E
|
E»
I ',
|
:

3 HOLES, 0.062-in DIA,

f
}
}
AT 120 DEG—>__. . 1/
A ZN%)
i
PREPLACED BRAZING ALLOY” [ = TUBING ’
7
Fig. 2,1.23, Alloy Preplacement Procedure for Fig. 2.1.24. Brazed Tube-to-Header Specimen Showing
Brazing Thick Tube Sheets. Excellent Filleting.

72

Flow of the alloy along the joint was evaluated
from metallographic examination of numerous sec-
tions of the sample. A photomicrograph of the fillet
on the braze side of the tube sheet is shown in
Fig. 2.1.25, while the joint on the weld side of the
tube sheet is shown in Fig. 2.1.26. On the basis
of these observations it can be concluded that flow
was probably continuous along the entire 5 in. of
the joint and around the entire 360 deg of the cir-
cumference. Also no cracking of the brazing alloy
was noted.

In view of the promising results of this experi-
ment, a more comprehensive evaluation program is
planned. Two 6-in.-dia, 5-in.-thick tube sheets
containing five tubes are presently being machined
to determine the influence of section size on braze-
metal preplacement procedure, |t is probable that
these experiments may suggest modifications in
joint design and brazing procedure. In addition,
other factors must also be studied if these experi-
ments show promise. For example, the minimum
rate-of-rise to brazing temperature to achieve satis-
factory flow during brazing of heavy tube sheets
will be affected by the choice of materials and the
brazing atmosphere. This rate should be determined
for the base material and brazing alloy compositions

Fig. 2.1.25.

Brazed Side of Tube Sheet Showing Fillet.

PERIOD ENDING OCTOBER 31, 1958

of interest, Further, the influence of time and tem-
perature on the diffusion of microconstituents from
the brazing alloy into the tube walls and to a lesser
extent into the tube sheet must be determined. Since
it is reasonable to expect that the base metal me-
chanical properties will be affected, the nature of
this effect must be known if back-brazing is to be
utilized,

Fabrication of Pump Components
G. M. Slaughter

The problems associated with joining a mo-
lybdenum extension to an Inconel shaft for a pump
application are being investigated. The widely
different coefficients of thermal expansion (that is,
molybdenum, 2.7 x 10~¢ in./in./°F; Inconel,

6.4 x 10~% in./in./°F) present a condition of very
high stresses during service. One method of at-
tachment proposed by the Pump Development Group
was to utilize molybdenum fingers brazed to the
Inconel shaft around the shaft periphery. The
amount of stress built up in each brazed lap joint
would thus be influenced markedly by the size of
the molybdenum fingers.

In order to test this proposed method, several mo-
lybdenum-to-Inconel brazed joints were fabricated

UNCLASSIFIED
Y-26634

vid w4 . K 4
LA R et g 4 B I ey

P BT s e S Wy e P

Etchant: glyceria regia, 20X.

73

MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
Y-26633

Fig. 2.1.26. Weld Side of Tube Sheet Showing Complete Flow to the Root of the Weld. Etchant: glycerio regia. 20X.

and are being examined metallographically for
cracking after brazing and after thermal cycling.

The samples are ]8 <Y x1in, % xY% %1 in, and

. %1 %1 in. pieces of molybdenum brazed to Y-
in.~thick Inconel plate with copper, Au-Ni, and
Coast Metals No. 52 brazing materials. The three
specimen sizes are illustrated in Fig. 2.1.27. These
three brazing alloys represent, respectively, a
weak, ductile alloy; a moderately strong, moderately
ductile alloy; and a very strong, brittle alloy.

Although the evaluation is not yet complete, it
appears that the Coast Metals No. 52 alloy is the
most suitable of the three for brazing the ‘,é x 1 x l-in,
fingers. Severe cracks were sometimes noted in
joints brazed with the other alloys. The cracking
tendencies are decreased as the size of the mo-
lybdenum fingers is decreased.

Thermal expansion difficulties are also encountered
in attaching a molybdenum sleeve-bearing around an
INOR-8 or Inconel pump journal. A metal-sprayed

74

UNCLASSIFIED
Y-26537

Fig. 2.1.27. Molybdenum-to-Inconel Brazed Specimens.
molybdenum surface has spalled during thermal
cycling, and consequently high-temperature brazing
has been suggested for several alternate designs.
- One proposed design (exploded view shown in Fig.
2.1.28) incorporates molybdenum pins which are
brazed to an INOR-8 or Inconel shaft, as shown in

UNCLASSIFIED
Y-26597

ms.

ONE INCH

Fig. 2.1.28. Exploded View of Simulated Pump
Journal Bearing Subassembly. The molybdenum pins
are to be joined to the INOR-8 or Inconel shaft by

brazing.

PERIOD ENDING OCTOBER 31, 1958

Fig. 2.1.29. A molybdenum sleeve would be brazed
to the molybdenum pins. This design minimizes
thermal expansion stresses, since the molybdenum-
to-Inconel joints are located very near the center of
the shaft, Studies are under way to determine the
optimum brazing alloy and joint design for this
application.

UNCLASSIFIED
Y-26730

ONE INCH

Fig. 2.1.29. Pump Journal Bearing Subassembly After

Brazing.

75

MOL TEN-SALT REACTOR PROGRESS REPORT

2.2,

RADIATION DAMAGE

G. W. Keilholtz

INOR-8 THERMAL-CONVECTION LOOP
FOR OPERATION IN THE LITR

W. E. Browning, Jr.
W. H. Montgomery R. P. Shields

The electrically heated mockup of the INOR-8 in-
pile thermal-convection loop, previously described,
was operated under conditions simulating those ex-
pected in the LITR. Operation of the mockup dem-
onstrated that the ports for the injection of cooling
air would have to be more widely distributed around
the loop to prevent cold spots that would be at tem-
peratures below the melting point of the fuel. The
fuel-filled mockup was operated for 771 hr at tem-
peratures ranging from 450 to 750°C, with axial
temperature differences as high as 250°C. The
following situations that will be encountered during
in-pile operation were simulated: melting the fuel,
reactor shutdown, reactor startup, reactor operation
up to one-half power, emergency shutdown required
by stoppage of fuel flow, adjusting the differential
temperature to desired value by altering distribution
of cooling air. Operation of the mockup served as
a performance test for the various components of
the loop, The experiment was terminated by a leak
in the fuel tube near an electrical contact which
conducted electrical current to the loop to simulate
fission heat. This contact, which is not a part of
the in-pile loop, had produced higher than normal
temperatures, The mockup loop is being examined
for evidence of corrosion and for further information
as to the cause of the leak.

The in-pile model of the loop is being assembled,
and the required modifications in the cooling air
supply have been made. The loop was fabricated
of INOR-8 and is to be filled with the fuel mixture
LiF-BeF,-UF, (62-37-1 mole %). Operation of this
loop will test the compatibility of this combination
of fuel and container material under radiation con-
ditions,

GRAPHITE CAPSULES IRRADIATED
IN THE MTR
H. L. Hemphill

The two fuel-filled graphite capsules that were ir-

radiated in the MTR at 1250°F for 1610 and 1492

W. E. Browning, Jr.

'w, E. Browning et al., MSR Quar. Prog. Rep. Oct. 31,
1957, ORNL-2431, p 30.

76

hr, respectively, at integrated power densities of -
1520 and 1375 kwhr/cm? are being examined. These
graphite capsules, described previously,? were en-
closed in Inconel containers and were filled with

the fuel mixture LiF-Ber-UF4 (62-37-1 mole %).
The preliminary examinations have indicated that
the graphite was unaffected by irradiation while ex-
posed to the fuel mixture. The graphite was wetted
by fuel in the irradiated capsules, but was not wetted
in similar unirradiated control capsules. A section
through one of the irradiated capsules is shown in
Fig. 2.2.1, and Fig. 2.2.2 shows a similar view of
an unirradiated control capsule that had been held
for 3389 hr at 1250°F. The irradiated fuel dissolved
more rapidly in water used in cleaning than did the
unirradiated fuel.

These preliminary examinations have indicated
that a fuel-cooled graphite moderator would not be
limited by damage to the graphite by fluoride fuel.
Further detailed examinations of the capsules and
fuel mixtures are under way.

2y, E, Browning and H. L. Hemphill, MSR Quar. Prog.
Rep. June 30, 1958, ORNL-2551, p 81.

1UNCLASSIFIED
RMG-2391 -

FLOURIDE
ELE

GRAPHITE

INCONEL

0.0251N.

B .

Fig. 2.2.1. Fluoride-Fuel-Filled Graphite Capsule
Sectioned After Irradiation in the MTR for 1610 hr at
1250“F. A low-permeahility grophite was used and the
fue |l mixture was LiF-BeFQ-UF4 (62-37-1 mole %). 7X.

PERIOD ENDING OCTOBER 31,1958

B UNCL ASSIFIED
¢ RMG-2370

FLUORIDE FUEL
GRAPHITE

, g o
.:CONEL »

. .

g 0.025 IN.

Fig. 2.2,2. Fuel-Filled Graphite Control Capsule
Tested Out-of-Pile at 1250°F for 3389 hr. 7X.

77

MOL TEN-SALT REACTOR PROGRESS REPORT

2.3. CHEMISTRY

W. R. Grimes
Chemistry Division

PHASE EQUILIBRIUM STUDIES
Systems Containing UF , and/or ThF

R. E. Thoma H. A. Friedman
H. Insley C. F. Weaver

The System NaF-BeF -ThF ;. — Phase studies of
sodium-based beryllium-containing fluoride salt mix-
tures were initiated in order to obtain data for com-
parisons of the properties of the system NaF- Ber-
ThF ,-UF, with those of the system LiF-BeF,-
ThF UF Such comparisons will be of volue in
apprclsmg the usefulness of the lithium-base qua-
ternary system. The completed phase equilibrium
diagram of the system NaF-ThF ,, shown in Fig.
2.3.1, was described in a recen'r report. |
nary investigations of the system NaF-BeF,-ThF,

Prelimi-

with the use of thermal-analysis and thermal-
gradient quenching techniques have shown that the
system is remarkably similar in several respects to
the system Lif:-BeF;,_-UF“.2 Recent results of
phase studies have also shown that the two com-
pounds 2NaF+ThF , and 2NaF.BeF, form the simple
binary system illustrated in Fig. 2.3.2. Since the
compound 4NaF+ThF ; does not form a solid solution

R. E. Thoma et al., Phase Equilibria in the Systems
L,iF-TbF'4 and NaF-TbF4, Paper No. 39, ACS Meeting,

Chicago, Illinois, Sept. 8, 1958,

2¢, J. Barton et al., MSR Quar. Prog. Rep. Oct. 31,
1957, ORNL-2431, p 32.

UNCLASSIFIED
ORNL LR- DWG 28528AR

1150 . : T T
H | '
R
1050 \
950 N | ‘
&850 L. A y ]
2 ‘ \ {
< ! \
& 750 ) S ! -
a \ \
z \ ] '
= 650 i Lous ]
o= o
NE o
— 1 W — .
290 FNoF They T 2 % 2 { i
: 2NuF~ThF4J-— m : ] ‘
450 R L b i | |
NoF 10 20 30 40 50 60 70 B8O 90 Thf
Thfy (mole %)

Fig. 2.3.1. The System NaF-ThF .

78

containing BeF_ and, in the binary system, has a
lower limit of stability (565°C) that is higher than
the lowest thermal break in the composition triangle
2NoF-BeF2-2NcF-ThF4-NoF (512°C), this triangle
necessarily becomes the single compatibility tri-
angle at higher NaF compositions than those which
occur on the alkemade line 2NaF:ThF ,-2NaF:BeF,.
The single eutectic in this compatibility triangle
has the approximate composition 70 mole % NaF -
25 mole % BeF ,—5 mole % ThF, and melts at
512°C.

Thermal-analysis data lead to the inference that
the maximum solubility of ThF , in NaF-BeF, mix-
tures is approximately 7 mole % at temperatures
below 550°C, except in quite small composition
regions adjacent to boundary curves separating
NaF-ThF , primary phase fields. Thermal-gradient
quenching experiments have been completed which
outline roughly the liquidus at 7 mole % ThF . The
results of both thermal-analysis and thermal-gradient
quenching experiments are shown in Fig. 2.3.3.

Urenium Phase Separation in LiF-BeF -UF , Mix-
tures. — Fifty-pound portions of LiF-BeF -UF  (62-
37-1 mole %) transferred from a 250-1b batch have
been found to contain an average of 3.5 wt % vra-
nium (nominal concentration of uranium in this com-
position is 6.5 wt %), although previous samples

UNCLASSIFIED
ORNL-LR-DWG 34530

TOOT
! :
| |
1 !
S 600 [——— i A
b 2N0F ThF
L + o
= LIQUID /
5 1 —
iy i ‘ ;
i 1
% ‘ 2NaF * Th, @-2NaF - By
500 | + -
‘ a-2NaF - BeF, '—l'QU'D |
i |
| , |
| @ THERMAL BREAKS '
0 PETROGRAPHIC DATA ?
400 | 3 ‘5 %
0 5 10 15 20 25 30
2NaF - ThFy BeF, (male %) a-2NaF- Bef,

Fig. 2.3.2. The System 2Na F'ThF4-2Nu F-Be F2.

UNCLASSIFIED
ORNL-LR-DWG 34534

800 1
LIQUID
~ 700 :
(&)
< : o THERMAL BREAKS
§ \ o PETROGRAPHIC RESULTS 7
= 600 Y | //
i AN i f
b= NoF+ "\ g/2NoF-ThE+ o 0™ "\ UNIDENTIFIED
9 500 [LIQUID LIQUID PHASE +
LIQUID
400 I
20 30 40 50

BeF, (mole %)

Fig. 2.3.3. Liquidus Diagram for Mixtures Containing
7 Mole % ThF4 in the System NoF-Ber-ThF4.

considered to be representative of the composition
of the 250-1b batch contained approximately the
nominal uranium concentration. At the time of
transfer of the 50-1b batches, it was known that
some solid remained unmelted in the reservoir con-
tainer. Studies have shown that the abnormally low
uranium concentration of the transferred batch is
related, as shown below, to the quantity of solid
remaining in the reservoir container.

The lowest melting eutectic in the system LiF-
BeF,-UF ; has the composition 47.5 mole % LiF-52
mole % BeF,~0.5 mole % UF ,, according to a phase
diagram derived at Mound Laboratory. The first
liquid to form when LiF-BeF ,-UF , (62-37-1 mole %)
is melted must have the eutectic composition, and
thus it contains 3.1 wt % uranium. The liquid com-
position of any partially molten mixture of LiF-

BeF ,-UF, must be intermediate in composition be-
tween that of the lowest melting eutectic and the
nominal composition of the mixture, although it need
not be located linearly between these two compo-
sition points. Thus, complete melting of the mixture
is necessary in order to maintain composition homo-
geneity.

The System LiF-BeF ,-ThF ,. — As was noted pre-
viously,? LiF-ThF, solid phases take a BeF ,-con-
taining phase into solid solution to a limited extent.
Recent, extensive, gradient-quenching experiments
have delimited the range of solid solutions for the
various LiF-ThF4 compounds. |t has been found

3R. E. Thoma ez al., MSR Quar. Prog. Rep. June 30,
1958, ORNL-2551, p 83.

PERIOD ENDING OCTOBER 31, 1958

that LiF-4ThF4 and 7LiF-6ThF4 exist as practi-
cally pure compounds in the ternary system. The
compound LiF-QThF4, however, can take as much

as > mole % BeF, into solid solution and 3LiF+«ThF
can take as much as 11 mole % BeF,. Moreover, the
introduction of BeF, into the structure of 3LiF+«ThF,
apparently causes enough distortion of the lattice
for some 7LiF-6ThF4 also to be taken into solid so-
[ution. Thus a roughly triangular area of homo-
geneous solid solution exists with one apex located
on the LiF-ThF, boundary at the composition
3LiF«ThF, and the other two apices at about the
compositions 67 mole % LiF-11 mole % BeF ,-22
mole % ThF, and 60 mole % LiF-13 mole % BeF,—
27 mole % ThF ,. Detailed quenching tests have
established that there is practically no solid so-
lution of 7LiF«6ThF , in 3LiF«ThF  in the binary
system LiF-ThF4. The nature of the solid solutions
in the ternary system is not known, but obviously
such solid solutions are not of the substitutional
type. With the increases of 7LiF:6ThF , in
3LiF+ThF, in the ternary system, the refractive
indices of 3LiF-ThF4 are raised somewhat and the
birefringence is lowered. The lines of the x-ray
powder-diffraction pattern also shift with increasing
solid solution. The compound 2LiF:BeF, appar-
ently exists only as the pure compound in the
ternary system, and there is no evidence that it
enters into solid solution with any of the LiF-ThF,
compounds.

Further gradient-quenching studies have required
the following modifications in the preliminary dia-
gram of the LiF-Ber-ThF4 system. (1) The
primary phase 7LiF-6ThF4 does not have a common
boundary curve with the 2LiF<BeF, phase. (2)
There is a peritectic at the composition point 64
mole % LiF-28 mole % BeF2—8 mole % ThF4, which
is ot a temperature of 450°C. The solid phases
3LiF-ThF4 (solid solution), 7LiF-6ThF4, and
LiF+2ThF , (solid solution) are in equilibrium with
liquid at the peritectic point. (3) There is a peri-
tectic at the composition point 64 mole % LiF-31
mole % BeF,-5 mole % ThF4, which is at a temper-
ature of 434°C. The solid phases 3LiF+ThF  (solid
solution), 2LiF«BeF,, and LiF:2ThF, (solid so-
lution) are in equilibrium with liquid at this peri-
tectic point. (4) The peritectic point at which LiF,
3LiF+ThF, (solid solution), and 2LiF:BeF, are in
equilibrium with liquid has the composition 66
mole % LiF-30 mole % BeF ,~4 mole % ThF ,, and
the temperature is 450°C. (5) The eutectic of

79

MOLTEN-SALT REACTOR PROGRESS REPORT

2LiF.BeF,, BeF,, and LiF+2ThF , (solid solution)
has been established at a composition of approxi-
mately 48 mole % LiF-51 mole % BeF ,—1 mole %
ThF,, and it has a temperature of 355°C.

Models of the Systems LiF-ThF ,-UF, and LiF-
BeF ,-ThF ;. — A three-dimensional model of the
system LiF-ThF ,-UF , has been constructed on the
basis of the terminal phase diagram presented pre-
viously? and is shown in Fig. 2.3.4. The three-
dimensional model® of the system LiF-BeF,-ThF,
presented in Fig. 2.3.5 was based on phase equi-
librium studies of the system LiF-BeF,-ThF,,
which are complete, as described above, with re-
spect to measurements of the liquidus surface and
determinations of temperatures and compositions of
invariant points.

Solubility of PuF, in LiF-BeF , Mixtures
C. J. Barton R. A. Strehlow

Data were obtained on the effect of CeF; and
BaF, on the solubility of PuF, in LiF-BeF, (63-37
mole %), and the solubility of pure PuF, in the
solvent was determined at two temperatures. The
effect of CeF, on the solubility of PuF, is of
interest as an example of the possible effect of tri-
valent elements resulting from the fissioning of plu-
tonium in a plutonium-fueled fused-salt reactor. The
possibility of precipitating PUF3 in the form of a
CeF ,-PuF ; solid solution, along with trivalent
fission-product fluorides of limited solubility, is
also of interest as one step in a reprocessing
scheme.

The effect of BaF, on PuF, solubility is of in-
terest as an example of the possible effect of di-
valent fission-product ions. The data are shown
graphically in Fig. 2.3.6, along with previously re-
ported data for ThF ,-containing mixtures.® The
theoretical curves shown in Fig. 2.3.6 for mixtures
containing CeF ; and PuF, were calculated with the
use of the equation

_ O
(]) NPuF3(d)—SPuF3NPuF3(ss) !

adapted from the solubility relation developed by W.

4Ibld, e5p. Figo 2.3.3-

>The construction of these models was a part of the
assignment to C. S, Johnson, MIT summer participant,

1958,

éC. J. Barton and R. A, Strehlow, MSR Quar, Prog.
Rep. June 30, 1958, ORNL-2551, p 84.

80

T. Ward et al.,” in which 5B uF
vhs

(solubility) of F’uF3 in the solvent at a specified
temperature, as shown by the top line in Fig. 2.3.6

(labeled *‘PuF; only™’), and Nqua(ss) is the mole

fraction of PuF, in the solute mixture. Two of the
four data points obtained with a 1:1 molar ratio of
CeF:3 to PUF3 added to the LiF-BeF2 (63-37 mole %)
mixture agree quite well with the theoretical curve,
while four of the five points obtained with a 5:1
molar ratio of CeF ; to PuF, lie reasonably close to
the theoretical curve. The reason for the divergence
of the results in the other three experiments is not
clear.

Similar plots of CeF, solubility values based on
preliminary analytical data show even poorer agree-
ment with theoretical curves obtained with the use
of data reported by W. T. Ward® for the solubility of
Cel:3 in LiF-BeF2 (62.4-37.6 mole %). Analytical
methods for determining cerium in the presence of
plutonium in mixtures of this type are being studied
and it is hoped that a better correlation between
theoretical and experimental values will be obtained
with later analytical results. It seems reasonably
safe to assume that Eq. 1 holds well enough for
mixtures of this type.

Data in Fig. 2.3.6 show that the solubility of
PuF, in LiF-BeF, (63-37 mole %) can be redyced
from 0.15 to 0.026 mole % at 500°C by dissolving
0.75 mole % CeF , in the mixture at a higher temper-
ature before cooling to 500°C. The calculated con-
centration of CeF ; remaining in solution at 500°C
is 0.14 mole %. By going to a lower temperature
(the liquidus temperature for the solvent is about
458°C) or a higher cerium-to-plutonium ratio, the
recovery of plutonium from used fuel could undoubt-
edly be improved, but it would be difficult to recover
more than 90% of the plutonium by a single operation
of this type. By reheating the filtrate and adding
more CeF, to the mixture, followed by a second
cooling and filtration operation, further improvement
in the efficiency of plutonium recovery could be
cbtained.

The data in Fig. 2.3.6 seem therefore to provide
the basis for a suitable reprocessing scheme for
fused-salt mixtures containing plutonium. Concen-
tration of the plutonium by precipitation would

is the mole fraction

"W, T.Ward et al., Solubility Relations Among Some
Fission Product Fluorides in NaF'-ZrF'4- UF'4 (50-46-4
mole %), ORNL-2421 (Jan. 15, 1958).

8w, 1. Ward, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 90.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
PHOTO 32550

ThF,
1111°C

+ LIQUID

LiF-4ThF4—LiF-4UF4 SOLID
SOLUTION + LIQUID

PR Y S
P

LiF-2ThF4 + LIQUIDE

3 LiF-ThF, + LIQUID —
e

! g

PR .6 ThF,—7LiF-6UF, ‘
SOLID SOLUTION + LIQUID .

|

+
400 TUF-6ug,

u'+7[jf.‘u“ .5! 7[5‘-‘"; §'
= A :
300 - + =
S UF-4ug, UF-4UE, + ug,
N
~_

Laanabo bl Vbl s, cul i __,_‘.J

Fig. 2.3.4. Model of the System LiF-ThFa-UF“.

81

MOLTEN-SALT REACTOR PROGRESS REPORT

LiF; 7LiF-6ThF,
845°C + LIQUID

B 3 LiF ThF,

. + LIQUID
70
600

<
N

g

= g4

ThFg + LIQUID

LiF-4ThF,
+ LIQUID

3\
s <
LiF-2ThF, B
+ LIQUID §

Fig. 2.3.5. Model of the System LiF-Ber-Th FA'

82

ThFg,
1111°C

UNCLASSIFIED
PHOTO 32551

UNCLASSIFIED
ORNL—-LR-DWG 329414

TEMPERATURE {°C)
650 600 550 500
| I [ I

! T 4:4/\ -
EXPERIMENTAL POINTS N— 0 —— 1]
PuFz ONLY IN SALT . %:

SOLUBILITY OF PuFy IN LiF -BeF, (63-37 mole o)

0.05 [—
e -
® PuFy + ThF, IN SALT \\
& PuFy + BaFp IN SALT 4
4 1:1 MOLAR RATIO OF CeF3 TO \\
PuF3 IN SALT
0.02 I'g 5:4 MOLAR RATIO OF Cefy TO
PuF3 IN SALT } |
0.0 ‘ W : ‘ ‘
100 105 M0 M5 120 125 130 135
10%/7 (°K)

Fig. 2.3.6. Effect of Other Fluorides on Solubility of
PuF, in LiF-BeF, (63-37 mole %).

simplify subsequent operations required to purify
the plutonium, possibly by fluorination. The effect
of trivalent fission products on the solubility of
PuF, is large enough so that they could not be
allowed to build up to high concentrations, and
some of them probably should be removed anyway to
reduce nuclear poisoning.

Comparison of the PuF; solubility data for mix-
tures containing ThF, and BaF, with the values ob-
tained in the absence of additives, as shown in
Fig. 2.3.6, seems to indicate that ThF, may reduce
the solubility of PuF,; slightly at the lower end of
the temperature range investigated, while the solu-
bility in BaF ,-containing mixtures is about the
same as that in the ThF ,-containing solvent at
about 500°C and significantly less at 650°C. In
neither case is there sufficient data to be con-
clusive and additional experiments are planned, but
the very odd behavior of the solubility curve for
3aF ,-containing mixtures may indicate that iso-
therms in the psuedo-ternary system BaF ,-PuF ,-

PERIOD ENDING OCTOBER 31, 1958

solvent are not linear, although they apparently are
linear for the system CeF ;-PuF ;-solvent and for a
number of similar systems investigated by Ward

et al.” Phase behavior in the system BaF -PuF,
has apparently not been investigated, but it is very
likely similar to that in the system B<JF2-UF3.9
According to these investigators, UF ; dissolves in
solid BaF, to the extent of about 50 mole %, while
BaF, dissolves in UF; to the extent of about 20
mole %. The residue from one of the filtration ex-
periments with BaF,-PuF; mixtures was found by
microscopic examination to contain PUF3 that had
optical properties that were altered sufficiently to
indicate that it may have contained as much as 10
mole % BaF,. It should be noted that, although the
concentration of BaF, remained relatively constant
in the filtrates and indicated that the amount added
was soluble in the temperature range investigated,
the molar ratio of PuF; to BaF, present in the un-
filtered mixtures varied considerably (from about 1.0
to 1.9). Since the concentration of barium in these
mixtures is far in excess of the concentration of di-
valent ions that would be produced in a power re-
actor, it seems clear that this type of fission
product would not significantly affect the solubility
of PuF, in the fuel mixture.

FISSION-PRODUCT BEHAVIOR
G. M. Watson F. F. Blankenship

Solubility of Noble Gases in Molten Fluoride
Mixtures

N. V. Smith

The solubility of argon in LiF-BeF, (64-36

mole %) was studied during the quarter at 500, 600,
700, and 800°C over a pressure range from 0 to 2
atm. The results are summarized in Table 2.3.1 and
presented graphically in Fig. 2.3.7; a comparison
with the solubility in NaF-KF-LiF (11.5-42-46-5
mole %) is presented in Fig. 2.3.8. Corresponding
values for the solubilities of other noble gases in a
variety of molten fluoride mixtures were reported
previously. 19~ 13 A correlation of noble gas solu-
bility with atomic radius of the gas and the surface

IR, W. M, D’Eye and F. S. Martin, J. Chem. Soc.
(Londen), 1847 (1957).

wW. R. Grimes, N. V. Smith, and G. M. Watson, [.
Phys, Chem, 62, 862 (1958).

11, H. Shaffer et al., MSR Quar. Prog. Rep, Oct. 31,
1957, ORNL-2431, p 41.

12N, V. Smith, MSR Quar. Prog. Rep. Jan. 31, 1958,
ORNL-2474, p 91,

13N, V. Smith, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 88.

83

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.3.1. Solubility of Argon in LiF-BeF2 (64-36 mole %)

Temperature Saturating Solubility .
(°C) Pressure (atm) (moles of t:lrgon/Cm3 of melt) K
x 10~8 x 1078
500 0.996 0.562 0.56
1.014 0.523 0.52
1.024 0.519 0.51
1.489 0.755 0.51
1.500 0.855 0.57
1.960 1.041 0.53
1.964 1.088 0.55
Av 0,54 0,02
600 0.984 1.001 0.98
0.996 0.949 0.95
0.997 1.001 1.00
1.013 1.046 1.03
1.513 1.466 0.97
1.513 1.417 0.94
1.971 1.927 0.98
1.987 1.919 0.97
Av 0,98 10,02
700 0.989 1.596 1.61
0.996 1.736 1.74
1.008 1.566 1.55
1.025 1.582 1.54
1.249 1.960 1.57
1.479 2.688 1.82
1.513 2.485 1.64
1.514 2,697 1.78
1.934 3.444 1.78
1.954 3.618 1.85
1.967 3.337 1.70
Av 1.69 0.10
800 0.993 2.628 2,65
0.999 2.612 2.62
1.020 2.696 2.64
1.239 3.227 2.60
1.476 3.908 2.65
1.499 3.888 2.59
1.532 4.077 2.66
1.532 4,242 2.77
1.922 5.260 2.74
1.967 5.145 2.62
2,008 5.007 2.50

Av 2.66 £ 0.06

84

*K =c/p in moles of gas per cubic centimeter of melt per atmosphere.

UNCLASSIFIED
ORNL- LR~ DWG 34532

A)oc
o

b

. A)"C
.

N

©
E
-
5
)
E
(o]
~
[ =
o]
o
=
o
w“
]
uw
o
)
[=
)—
=
=
@B
>
0
Q
7]

v 1»/500"(:
/ 4
0
0 0.5 10 1.5 20 25

PRESSURE (atm)

Fig. 2.3.7. Solubility of Argon in Molten LiF-BeF,
(64-36 mole %) at Yarious Temperatures.

tension of the solvent is presented in a following
section of this chapter that is titled ‘‘A Simple
Method for the Estimation of Noble Gas Solubilities
in Molten Fluoride Mixtures.”’

Comparisons of the solubilities of helium'? and of
argon in LiF-BeF, (64-36 mole %) with their solu-
bilities in other solvents previously studied have
indicated a number of similarities and one differ-
ence. The similarities include (1) conformity to
Henry’s law, (2) increase of solubility with in-
creasing temperature, (3) decrease of solubility with
increasing atomic weight of gas, and (4) increase in
the enthalpy of solution with increasing atomic
weight of gas. The difference is the value of the
entropy of solution in the LiF-BeF, mixture. While
all the entropies of solution for helium, neon, argon,
and xenon in Nc:F-ZrF‘1 (53-47 mole %) were quite
small, about —1 eu, the entropy values for helium
and argon are —3.73 and —3.95, respectively, in the
LiF—BeF2 (64-36 mole %) mixture. The significance
of this difference is not immediately apparent, but
the difference seems to be related to the nature of
the solvent.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
ORNL—LR—DWG 34533

TEMPERATURE (°C)

-9
(X'BOO ) 8OO 700 600 500
] I I T
e ,<1,7 -

5 4\7}\

—~ARGON IN NgF-KF~LiF (REF1{3) |
AH=12,400 cal /mole
AS= —0.10eu |

-

’E ‘ .
.2 | ;
2 AN ]
O
E ™, ‘ !
- 20 e ———. ———
8 | |
£ \/ _~ARGON IN LiF -Bef,
B 10 \ _ AH= 8850 ccl/moie _
n ! Oy
£ -+ - — AS=—395eu -
) | A\ ‘ T
3 _ I R
oS . L : \____ I
= [ L
o 5 i _ 5 -
G
E —— — —_t I i —— ]
e i
E NN
X \
2L L « 2.09%
\ | | |
8 9 10 1 12 13 14 15

10%/T (°K)

Fig. 2.3.8. Solubility of Argon in LiF-BeF, (64-36
mole %) and in NaF-KF-LiF (11.5-42-46.5 mole %).

Solubility of HF in LiF-BeF , Mixtures
J. H. Shaffer

The solubility of HF in LiF-BeF, mixtures was
determined as a function of temperature, pressure,
and solvent composition. The composition was
varied from approximately 10 to 50 mole % BeF,, at
pressures from O to 3 atm, and at temperatures from
500 to 950°C, depending on the liquidus temper-
atures of the mixtures.'* Within the precision of
the measurements, the solubility of HF was found to
follow Henry’s law over the pressure range studied.
Accordingly, at a given temperature and solvent
composition, the solubility of HF can be measured
at several saturating pressures and the results can

MComposi're diagram compiled by R. E. Thoma, ORNL,
from D. M. Roy, R, Roy, and E. F. Osborn, J. Am. Ceram.
Soc. 37, 300 (1954); A. V. Novoselova, Y. P. Simanov,
and E. |, Yarembash, Zbur. Fiz. Khim. 26, 1244 (1952).

85

MOLTEN-SALT REACTOR PROGRESS REPORT

be averaged as a Henry’s law constant character-
istic of that temperature and composition. The
average constants obtained experimentally are pre-
sented in Fig. 2.3.9. For each point, at least three
constants were obtained that agreed with each other
to within £5%. The averages of the 800 and 900°C
data at 31 mole % BeF2 are believed to be low,
possibly because of consistently reproducible errors
in the temperature measurement as a result of a de-
fective thermocouple. The values obtained at this
composition will be rechecked before proceeding

-6 UNCLASSIFIED

(Xx10 ) ORNL—LR-DWG 34534
30 | l
———EXTRAPOLATED

o
o 20
-
a ——
(1] — e
=] e
§ 15 e
B ““‘- 6OO°C
E " ——— '\ 70
S 10 ke Ooc ‘xot
[=]
K S — ! \ N
£ . l-._____\ \ 500 T~
[ » °
& . \\\C \.L\
(TR 3
E -\OOOC\D\
5 » \
v
% 5 \.
E \
x 4

3

G 10 20 30 40 5C

LiF BeF, (mole %)

Fig. 2.3.9. Composition Dependence of Henry's Law
Constants for the Solubility of HF in LiF-BeF2 Mixtures.

with a corresponding investigation of another sol-
vent system.

For purposes of comparison, Henry’s law con-
stants expressed at rounded values of composition
are more useful, and therefore values taken from a
large-scale plot of the data are summarized in Table
2.3.2. The enthalpies and entropies of solution of
HF as functions of composition of the solvent are
listed in Table 2.3.3.

It may be seen that in LiF-BeF, solvents, the
solubility of HF does not show a very strong com-
position dependence, such as it did in NaF-ZrF,
mixtures, as indicated previously. 'S The weaker

15, H. Shaffer, MSR Quar. Prog. Rep. Jan. 31, 1958,
ORNL-2474, p 93.

Table 2.3,.3. Enthalpies and Entropies of Solution
of HF in LiF-BeF2 Mixtures

Solvent Enthalpy Entropy

Composition of Solution of Solution*

(mole % LiF) (cal/mole) (ev)
100 -4700 -4.4
90 -4900 -4.8
80 -5100 -5.2
70 -5400 -5.8
60 -5400 -6.2 '
50 —-5400 —6.6

*Entropy of solution calculated at equal concentra-
tions in the gas and liquid phases at 800°C,

Table 2,3.2, Henry’s Law Constants for the Solubility of HF in LiF-BeF, Mixtures

Solvent

K (moles of HF per t:m3

of melt per atmosphere)

Composition

(mole % LiF) At 600°C At 700°C At 800°C At 900°C
x 1078 x 1078 x 1078 x 1070
100 18.2* 13.7+ 1.0+ 9.0
90 17.3%* 130+~ 10.3+* 8.4
80 16,0+ 1.8 9.1 7.6
70 13.9 10.1 7.7 6.3
60 11.2 8.2 6.4 5.2
50 8.6 6.6 5.2 4.2

*Graphical extrapolation to 0 mole % BeF2 (Fig. 2.3.9).

**Extrapolated from measurements at higher temperatures,

86

composition dependence may be a result of much
less stable acid fluorides in the LiF system'® than
in the NaF system. However, this effect should be
more precisely demonstrated when measurements
using NaF-BeF, mixtures are carried out.

Solubilities of Fission-Product Fluorides
in Molten Alkali Fluoride—Beryllium
Fluoride Sol vents

W. T. Ward

Solubility of CeF in NaF-LiF-BeF,. — The solu-
bility of CeF, in NaF-LiF-BeF, solvents of various
compositions was determined over the temperature
range 400 to 750°C. The mixtures investigated con-
sisted of the NaF-LiF eutectic (40-60 mole %) to
which BeF, was added in amounts that varied from
0 to 56 mole %. The results of the experiments are
shown in Fig. 2.3.10, in which the solubility of

wH. J. Emeleus in Fluorine Chemistry, Yol. 1, p 25,
J. H. Simons (ed.), Academic Press Inc,, N. Y. (1950).

UNCLASSIFIED
ORNL-LR-DWG 34535

© |
B — — —_— I__ —
4 I I | s LAY
GREATER THAN 4.5 \ \
| moie 7, AT 650°C \
2 e \ -
\ - 750°C o /
'.
L ]
1 \ o_Ye_700°C
\ A\ ® »
-~ 08 ANEAN l yd
: N ovore a2
%’ 0.6 \\ \ /
E .
E 04 ¥. 6000%
P_: ~—
—
c s
=
o 0.2 o .
w©
i \e” )/
[ ]
0.4 /
0.08 .//
0.06 //
0.04 e
0.02
0 10 20 30 40 50 60
t00% NoF - LiF Bef, IN SOLVENT {mole T}

EUTECTIC

Fig. 2.3.10. Solubility of CeF3 in Nt:F-LiF-BeF2

Mixtures.

PERIOD ENDING OCTOBER 31,1958

CeF, (mole %) is plotted against solvent compo-
sition at several temperatures. A comparison of the
solubility of CeF , in this solvent with the solu-
bilities of CeF, in the LiF-BeF, and NaF-BeF,
binary systems is presented in Fig. 2.3.11 for the
single temperature of 600°C.

UNCLASSIFIED
2 ORNL-LR-DWG 34536

Al \\ o
& .
1 [ VO
2 \ 2
= 08 : WY /4
= — \ - P 4
z \ a4
5 06 NN LFBeR o %
o \\ b b E\/
g —e O
L N 7
[0 04
© NuFuF~BeF2__,\€\_1/
\G'z\NcF—BeFZ
0.2
o 10 20 30 40 50 60

BeF, IN SOLVENT (mole Ta)

Fig. 2.3.11, Comparison of CeF, Solubilities in
Three Solvent Systems at 600°C.

Solubility of CeF, LaF,, and SmF, in LiF-BeF ,-
UF4. ~ The solubilities of CeF,, LoFa, and SmF3
in LiF-BeF ,-UF, (62.8-36.4-0.8 mole %)} were de-
termined over the temperature range 500 to 750°C.
The data are given in Table 2.3.4, and, for com-
parison, are plotted against the reciprocal of the
temperature in Fig. 2.3.12, As may be seen, the
solubility of CeF, in this solvent is a little higher
than in LiF-BeF, (63-37 mole %) over the full tem-
perature range investigated.

Simultaneous Solubilities of CeF, and LaF, in
LiF.BeF -UF ,. — The solubilities of CeF; and
LaF, in the presence of each other in LiF-BeF ,-
UF4 (62.8-36.4-0.8 mole %) were determined with
the use of two radioactive tracers in the manner de-
scribed previously.!” The values obtained are
listed in Table 2.3.5 and plotted in Fig. 2.3.13, in
which the sums of the solubilities may be seen to
be intermediate between the individual solubilities.
The lines represent the solubilities of CeF, and

7W. T. Ward et al., Solubility Relations Among Some
Fissiop Product Fluorides in NaF'-ZrF4
mole %), ORNL-2421 {(Jan. 15, 1958).

-UF ; (50-46-4

MOL TEN-SALT REACTOR PROGRESS REPORT

Table 2.3.4. Solubilities of CeF3, LuF3, and SmF3 in LiF-Beer-UF4 (62.8-36.4-0.8 mole %)

Temperature CeF3 in Filtrate Temperature La F3 in Filtrate Temperature SmF3 in Filtrate
(°C) Wt % Mole % (°C) Wt % Mole % (°C) Wt % Mole %
764 >9,1* >1.8* 741 8.65 1.71 757 >10.22*% >1.93*
661 5.1 0.97 676 >10.27* >1.94*
652 4,5 0.85 655 4.19 0.79
578 2.35 0.44 560 1.73 0.32 587 5.35 0.97
499 1.04 0.19 480 0.79 0.15 495 2.20 0.39
*Solution not saturated,
UNCLASSIFIED UNCLASSIFIED
ORNL-LR-DWG 34537 . ORNL-LR-DWG 34538
2
\ * CALCULATED TOTAL RARE—-EARTH FLUQRIDE PRESENT
0\ . e 182 mole % Cefs, 0.78 mole % LaF3
2 2\ & 1.85 mole % Cefy, 2.27 mole % LaFy
_ \ C 1.83 mole % Cefs;3.30 mole % LaFy
SmF o o ___.___E ]
— \ s
e i
% AR * £
E \L:N N\ &
= 0.8 *\ N E
w AN \ E
g L
£ 06 NN ;
E \N\eer, -\ z I
E L
w =
) \\\ 5
£ 04 =
[} \ .
N\
z =
% w
0 LoF3 W CeF, ONLY
w <
& : ]
o > Y = \
'_
\‘\ I I I N
LY
ot
9 {e) 11 12 13 14 15
(oK ]
9 10 11 12 13 14 104/ T{°K)

10Y/7 (o)

Fig. 2.3.12, Solubilities of CeF,, LaF,, and SmF,
in LiF-Ber-UF4 (62.8-36.4-0.8 mole %).

LaF ; individually (same curves as those shown in
Fig. 2.3.12), while the points represent the sums of
the solubilities obtained when both are present in
the proportions indicated.

It may be observed from the data given in Table
2.3.5 that the solubility of a rare earth fluoride is
decreased by the addition of another rare earth fluo-

ride. This is believed to be the result of the for-

mation of a solid solution by the rare earth fluorides

88

Fig., 2.3.13, Comparison of Solubilities of CeF, and
LoF3 Individually and When Both are Present in LiF-
Be FZ-U F4 (62.8'36.4‘0.8 mo'e %)o

involved. |t was shown previously that the equi-
librium quotient of the reaction

LaF, (d) + CeF, (ss) &= LaF (ss) + CeF, (d)

is given by the expression

SO
CeF3

PERIOD ENDING OCTOBER 31, 1958

where S° is the mole fraction (solubility) of a rare lated constants. These results are also illustrated
earth fluoride in the solvent at the specified tem- in Fig. 2.3.14.

perature and in the absence of a second rare earth

fluoride. A comparison of the calculated and the Simultaneous Solubilities of CeF; and SmF in
experimentally determined equilibrium quotients is LiF-BeF ,-UF,. — One experiment was run in which
given in Table 2.3.6. It may be seen that, regard- both CeF; and SmF, were added to LiF-BeF,-UF ,
less of rather large composition changes, the equi- (62.8-36.4-0.8 mole %). There was evidence that
librium quotients, or partition coefficients, remain this particular batch of SmF; was slower in dis-
fairly constant and in fair agreement with the calcu-  solving than usual, and the individual solubilities

Toble 2.3.5. Solubilities of CeF, and L0F3 Separately and in the Presence
of Each Other in LiF-Ber--UF4 (62.8-36.4-0.8 Mole %)

Constituents of System Rare Earth Flucride in Filtrate
(mole %) Temperature (mole %)
€c)
CeF3 L0F3 Solvent CeF3 L<:|F3 Total
180 0 98.20 700 1.36 0 1.36
600 0.55 0 0.55
500 0.193 0 0.193
1.82 0.78 97.40 700 0.890 0.387 1.28
600 0.375 0.141 0.52
500 0.148 0.043 0.191
1.85 2.27 95.88 700 0.588 0.525 1.1
600 0.271 0.195 0.47
500 0.120 0.067 0.187
1.83 3.30 94.87 700 0.486 0.775 0.126
600 0.220 0.296 0.52
500 0.093 0.102 0.195
0 2.20 97.80 700 0 1.19 1.19
600 0 0.475 0.48
500 0 0.177 0.177

Table 2.3.6. Solid-Solvent Extraction Coefficients for LaFg-Cng in LiF--Ber-UF4 (62.8-36.4-0.8 Mole %)

Experimental KN*

Temperature  Calculated Ky

(OC) © /5° For 1.82 Mole % CeF3 + For 1.85 Mole % CeF3 + For 1.83 Mole % CeF3 +
CeFgy = LaF, 0.78 Mole % LaF , 2.27 Mole % LaF 3.30 Mole % LaF,

700 1.14 0.97 1.55 1.18

600 1.15 1.18 1.82 1.38

500 1.09 1.52 2.28 1.68

NLaFS(ss) NCeF3(d)

CeF 4(s2) NLaFB(d)

89

MOLTEN-SALT REACTOR PROGRESS REPORT

SOLVENT UNCLASSIFIED
ORNL-LR—DWG 34539

& 3 /
o .
é AN\
&m s Nooe
N OVER-ALL
3 COMPOSITION
\\ / / \\ /
\ / \\ ///'
\ ) ’ A .
4q

/ S ovER-ALL S\ /
\\ COMPOSrTlON _ \

/ / \\ // \

. OVER-ALL
COMPOSITION

“

Fig. 2.3.14. Portion of Psvedo-Ternary System Ce F3~
La F3-So|venf Where the Solvent is LiF-Ber-UF4
(62.8'36.4'0.8 mo'e %)-

obtained were considered to be unreliable. How-
ever, the sums of the solubilities were intermediate
between the solubility curves obtained when only
CeF3 and only SmF3 were present, as shown in Fig.

2.3.15.

A Simple Method for the Estimation of Noble Gas
Solubilities in Molten Fluoride Mixtures

M. Blander G. M, Watson

The solubilities of noble gases in liquids may be
interpreted with the use of a model similar to that
described by Uhlig. 817 Although the model is
elementary, it yields an interesting correlation with
experimental data. If the gas does not interact with
the liquid in which it is dissolved, the free-energy
change upon solution is related to the ‘‘surface’”
energy of the hole created by the gas molecule or
atom. The solubility of such a gas in a continuous
fluid medium can be shown to be given by the re-
lation

A

d  -18.08r%y . /RT
mic

K T —— = @
©c C
g

8p 4, Emmett, private communication.

"9H. H. Uhlig, J. Phys, Chem. 41, 1215 (1937).

90

UNCLASSIFIED
ORNL- LR-DWG 34540

3
[ ]
2
g . SmF5 ONLY
B \
[}
E 1
= N\ N\
5 o8 N\ AN
o X
H N\ N\
z \ N\
=
2 06 \ . \
(& ]
T
@}
o
= oa \
z N
2
L \CeF3ONLY
Ll
o
< .
o
|
2002
: N
° \
CALCULATED TOTAL RARE-EARTH
— FLUORIDE PRESENT
. 2.83 mole % C,eF3
2.22 mole % SmFs
0.4
9 10 1 12 13 14

10% 7 (°K)

Fig. 2.3.15. Solubilities of CeF3 and SmF3 Indi-
vidually and When Both are Present in Li |:-Be|:2-U|:4
(62.8‘36-4‘0-8 mole %)l

where K _is the equilibrium ratio of the concen-
trations in the liquid (Cd) and in the gas phase
(Cg), 7 is the radius of the gas atom (in ‘Angstroms),
Y .ic iS the microscopic surface tension of the sol-
vent (in ergs/cmz), R is the gas constant (in
cal/moles®K), and the number 18.08 is the resultant
of the factors (6.02 x 1023) (477 x 10~ 1¢) (2.39 x
10‘8), which are used in the calculation of the sur-
face free energy (in cal/mole of a mole of spherical
holes of radius r (A} in a continuous liquid of sur-
face free energy y (e (ergs/cm?). Although a real
liquid cannot be considered a continuous fluid, it is
interesting to speculate that it behaves as if it were
one and that y
face tension Y.,
The area of the holes is at least as large as the
surface area of the spherical gas molecules, and
because of thermal motions the hole would be ex-
pected to be larger than the gas atom. The radii of

;. is related to the macroscopic sur-

PERIOD ENDING OCTOBER 31, 1958

the noble gas atoms in the solid will be a lower Atomic_Radius
limit of the radius of the hole, and the solubility (A)
calculated on this basis would be expected to be ‘

. high. Values of K_ are listed in Table 2.3.7 that He lium 1.22
were calculated for the noble gases by assuming Neon 1.52
thaty . =y_._and by using the radii of the noble Argon 1.93

- gas atoms in the solid. The numerical values used

Xenon 2,18

for the surface tensions of the solvents are given in
Table 2.3.8, and the radii of the noble gases, as

given by Goldschmidt, 20 are listed below: The comparison in Table 2.3.7 of the calculated

and measured values of K_ for solutions of noble
gases in NaF-KF-LiF and in NaF-ZrF, shows good
agreement in view of the elementary nature of the
20y, M. Goldschmidt, Geochemistry, Oxford University mo.dt'al. -I_-he n-mgnn‘ude and the variation of the sol-
Press (1954). ubility with size of the gas molecule are correctly

Table 2.3.7. Comparison of Calculated and Observed Values of Henry’s Law Constants,
K= Cd/cg’ for Noble Gas Solubilities in Molten Fluoride Mixtures

Sol c Temperature Kc Ratio of Experimental to
olvent as
(°C) Experimental Calculated Calculated K. Values
x 1073 x 1073
NaF-ZrF, Helium 600 15.5 137 0.11
700 23.3 188 0.12
800 37.0 243 0.15
Neon 600 8.09 45,9 0.18
} 700 14.7 74.9 0.20
800 21.7 112 0.20
Argon 600 3.62 7.33 0.49
. 700 6.44 16.0 0.40
800 10.6 30.2 0.35
Xenon 600 1.39 1.77 0.79
700 2,84 4.84 0.59
800 5.56 10.7 0.52
NoF-KF-LiF He lium 600 8.09 (28.3)* (0.29)
700 14.0 (46.8) {0.30)
800 20.3 (70.7) (0.29)
Neon 600 3.12 (3.94) (0.79)
700 6.00 (8.63) (0.70)
3 800 9.84 (16.4) (0.60)
Argon 600 0.645 (0,146) (4.4)
700 1.43 (0.509) (2.8)
) 800 2.99 (1.41) (2.1)
Xenon 600 (0.011)
700 {0.057)
800 (0.212)

*Values enclosed in parentheses are based on estimated surface tensions.

MOLTEN-SALT REACTOR PROGRESS REPORT

Surface Tensions of Molten Fluoride

Table 2.3.8.

Mixtures

Surface Tension (dynes/cm)

Temperoture
(°C) NaF-KF-LiF* NaF-Zr F4**
(11.5-42-46.5 mole %) (53-47 mole %)
600 230 128
700 220 120
800 210 112

*Surface tension values estimated from extrapolated
values of surface tensions of pure components [F. M.
Jager, |. fir anorg. chemie 101 (1917)] assuming addi-
tive surface tension. Surface tension measurements of
this mixture have not as yet been made.

**F, W, Miles and G. M, Watson, Oak Ridge National
Laboratory, unpublished work.

predicted. This order of solubility has not often
been experimentally observed, since most liquids
have surface tensions so low that other solution
effects predominate. The difference between the
calculated and experimental values is in the di-
rection expected if the estimated size of the hole is
too small (except for argon in NaF-KF-LiF).

The last column of Table 2.3.7 shows a trend in
the ratio of the experimental to the calculated
values of the Henry’s law constants. An approxi-
mate calculation indicates that this trend can be
explained if the polarization of the noble gas atom
by the highly ionic salt medium is taken into ac-
count.

It may be concluded from these calculations that
estimates can probably be made of the solubilities
of the noble gases in media of high surface tension
to within an order of magnitude. This accuracy is
satisfactory for reactor applications.

Chemical Reactions of Oxides with Fluorides

in LiF-BeF2
J. H. Shaffer

An effective means for separating uranium from
undesirable fission products is required in order to
develop a suitable reprocessing scheme for fused-
salt nuclear reactor fuels. Experiments described
previously?! demonstrated that uranium can be sepa-
rated from rare earth fluorides, such as CeF3, in the

21 J, H. Shaffer, MSR Quar. Prog. Rep. Jan, 31, 1958,
ORNL-2474, p 99.

92

simple binary solvent KF-LiF. In these experiments
uranium was preferentially precipitated as UO, by
the addition of calcium oxide to the fused salt mix-
ture. In the presence of BeF,, uranium probably
can be separated from cerium and cerium oxides,

but apparently oxides of cerium and beryllium will
coprecipitate after the precipitation of uranium is
essentially complete.

In practical applications the addition of calcium
oxide to precipitate uranium is objectionable. Con-
tamination of the fuel with calcium fluoride might
induce serious complications, such as increasing
the melt liguidus temperature. However, two alter-
nate methods of uranium separation are being in-
vestigated. First, beryllium oxide may be substi-
tuted for calcium oxide to precipitate uranium and
the rare earths. The primary advantage of this sub-
stitution is that no foreign constituent is introduced
into fuels containing beryllium. The results of pre-
liminary experiments with the use of beryllium oxide
have been reported.2?2 A second method would em-
ploy a precipitant which would form a volatile
product. Reagents of this class might be water,
boron oxide, and silicon dioxide.

Precipitations with BeO. — A chromatographic
type of separation of uranium and fission product
rare earths might be effected by using beryllium
oxide as the column packing material. Separations
of this type would yield a purified solvent as the
column effluent and would provide a means of con-
centrating the uranium in the column for subsequent
recovery. Pertinent information for this process in-
cludes knowledge of reaction rates, optimum particle
size of BeO, column recovery techniques, etc.

The rate of reaction of UF , with a onefold excess
of BeO as a function of the particle size is illus-
trated in Fig. 2.3.16. Experimental results indicate
that the reaction of UI:4 with solid BeO is surface
controlled. Further experiments are in progress to
determine additional characteristics of the ex-
traction process. |t is planned, for the next series
of experiments, to use actual columns packed with
BeO and to force the liquid LiF-Ber-UF4 mixture
to flow through the columns to determine the amount
of uranium extracted per pass.

Formation of Yolatile Products. — Results of ex-
perimental precipitations of uranium from UF | in
LiF-BeF2 (63-37 mole %) at 600°C by the addition

of water to the influent gas stream are presented in

22§ H. Shaffer, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 90.
UNCLASSIFIED
ORNL-LR-DWG 34544

|
t)o BeQ PARTICLE SIZE
\ ® 35 TO 60 MESH
g0 fe) A 20 TO 35 MESH
\. A\ 0 12 TO 20 MESH
o]
\ 0\3\
[ ] 2 \
\ \‘ )'\_\
70 \ s X
<
5
< 60
w
z \ A
D \
z .
Z 50 \
s . \
i
o A
: \
S 40 N
2 N
g \‘\
30 \
[ )
20 -
*J
N
0
0 20 40 60 80 100 120

TIME (min)

Fig. 2.3.16. Rote of Reaction of BeO (100% Excess)
With UF4 in LiF-BeF2 (63-37 mole %) at 600°C.

Fig. 2.3.17. As may be seen, the stoichiometry in-
dicates the formation of U02. This assumption has
been confirmed by petrographic examination.
Further experiments are in progress to investigate
the precipitation with water of cerium, as well as
the selective precipitation of uranium in the pres-
ence of cerium, in various fused salt solvents.

CHEMISTRY OF THE CORROSION PROCESS
Activities in Metal Alloys
S. Langer

Studies of the activity of nickel in nickel-molyb-
denum alloys by measurement of the electromotive
force of cells of the type

NiCl2 (5 wt %} in

i |
Ni (metal) NaCl-KCl eutectic

Ni-Mo Alloy

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
ORNL-LR-DWG 34542

[

/3/

URANIUM ME TAL FOUND (w1 %)
(5]
1
THEORETICAL END-POINT FOR UF,

e

g

0 100 200 300 400 500
HF EVOLVED {meq)

Fig. 2.3.17. Precipitation of UO2 from UF, in Li F-
BeF2 (63-37 mole %) With Hzo-

have been continued. The disagreement between
the potentials of identical pure nickel electrodes
which was experienced in the pusm‘23 has been
greatly reduced by changing to an uncompartmented
cell. The potentials of the nickel electrodes now
agree to within £1.0 mv and frequently to better
than £0.5 mv.

A condensation of the results obtained to date is
given in Table 2.3.9. Cell potentials measured in
the temperature range 680 to 820°C were plotted as
a function of temperature, and a straight line was
drawn that best fitted the data points. The data are
not yet sufficiently precise to warrant more refined
methods. The potential of a given alloy at 750°C
was read from the graph, and the activities and ac-
tivity coefficients listed in Table 2.3.9 were cal-
culated from the potential data.

The alloys used as electrodes were prepared by
annealing for 16 hr at 1200°C and quenching to
room temperature; thus the electrodes may not rep-
resent equilibrium samples even after remaining at

23g, Langer, MSR Quar. Prog. Rep. Jan. 31, 1958,
ORNL-2474, p 111.

93
MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.3.9. Tentative Values of Activities and
Activity Coefficients of Nickel Metal in
Nickel-Molybdenum Alloys

Nickel Potential Activi Activity
Atom Fraction, at 750°C, civity, Coefficient,
a.,.
XNi E (mv) Ni Ni
0.937 1.95 0.96 1.02
0.904 2.95 0.94 1.04
0.800 3.45 0.92 1.16

temperature in a cell for a week. In future experi-
ments, alloys which have been annealed at 750 to
850°C for at least 6 weeks will be used.

Surface Tensions of BeF2 Mixtures
B. J. Sturm

Experiences in pumping LiF-BeF ,-UF , melts led
to interest in the surface tensions of BeF, mix-
tures, and preliminary measurements have been
made. For these measurements a Cenco-DuNouy
interfacial tensiometer?4 was used to measure the
force required to pull a platinum ring out of the
melt. Since the measurements were made with the
liquid surface exposed to air, the results are ap-
proximate.

24p oduct of the Cenco Scientific Company.

A trial measurement of LiF-Ber-UF4 (53-46-1
mole %), with poor temperature control, gave a sur-
face tension of 195 dynes/cm at 425 to 450°C.
Measurements of a melt consisting of LiF and BeF
(63-37 mole %) gave 196 dynes/em at about 480°C.

The surface tension of the liquid can affect the
drainage behavior, and the level of a liquid in a
capillary is directly proportional to the surface
tension and inversely proportional to the density.
Therefore a comparison of the rise in a 2-mm-1D
capillary is indicative of the validity of the results.
Such a comparison is presented in Table 2.3.10 for
two melts and for water.

Disproportionation of Chromous Fluoride
B. J. Sturm

Chromous fluoride in solution with fused alkali
fluorides disproportionates into trivalent chromium
and chromium metal (see the following section). In
an attempt to demonstrate this reaction in concen-
trated solutions, and also to prepare compounds of
NaF and CrF,, nickel capsules containing mixtures
of sodium fluoride and chromous fluoride {(NaF-CrF
50-50 and 67-33 mole %) were heated to 1000°C.
For capsules maintained at the elevated temperature
for 2 hr and cooled slowly, x-ray diffraction showed
sodium hexafluochromate (IHI) to be the principal
phase. Capsules heated for only 2 min and
quenched on a copper plate contained only a trace
of sodium hexafluochromate (l1), the principal
phase being one regarded as sodium trifluochromate
(i1). The rapid cooling and quenching did not allow
time for disproportionation; the product was N(:lF-CrF2
instead of 3N0F-CrF3.

!

Table 2.3.10. Comparison of Surface Tensions and Approximate Rise in a 2-mm-ID Capillary of
Two Fused Salts and Water

Temperature Surface Tension Density Capillary Rise
(°C) (dynes/cm) (g/cma) (em)
NaF-ZrF ,-UF, 630 1329 3.34% 0.8
(50-46-4 mole %)
LiF-BeF, 480 196 1.92° 2.1
(63-37 mole %)
Water 25 72¢ 1 1.5

94

. 1. Cohen, W. D. Powers, and N. D, Greene, private communication.
W. D. Powers, private communication, September 1958,

“N. A. Lange (ed.), Handbook of Chemistry, 5th ed. (1944), p 1577.

Effect of Fuel Composition on the Equilibria
3CrF, &= 2CrF, + Cr° and
3FeF2\—-——-——_)2FeF3 + Fe°

J. D. Redman

A study of the chemical equilibria involved in the
corrosion of structural metals by fluoride fuels has
continued, with attention being given to the effect
of solvent composition on the equilibria

3CrF2#2CrF3 + Cr°
and
3FeF2 —4\—- 2FeF2 + Fe® .

Chemical analyses were made of equilibrated
samples contained in nickel filtration apparatus.
The samples were obtained by filtration after an
equilibration period of 5 hr at temperatures chosen
between 525 and 800°C.

Chromium Fluorides. — The extent to which the
reaction

3CrF2 #QCrF3 +Cr°

proceeds to the right was found to decrease with
increasing ZrF , concentration in the solvent LiF-
NaF-ZrF ,, as shown in Fig. 2.3.18. The solvent
LiF-Nch-ZrF4 was useful because it permitted a
continuous variation from a basic to an acidic melt

without interference from precipitation. The data of

Fig. 2.3.18 are based on duplicate experiments in

PERIOD ENDING OCTOBER 31, 1958

which excess chromium metal was equilibrated with
2 wt % Cr*** in the solvents described in Table
2.3.11. For this series of solvent compositions, in
which the mole ratio of NaF to LiF was held con-
stant at 40:60, the effect of Zr*" in competing with

UNCLASSIFIED
CGRNL.-LR-DWG 34543

[o2d
O

o
(@)

"

(&)
o
Z

o
o
y

T~

o
{

PERCENTAGE OF DISSOLVED GHROMIUM PRESENT AS cr ¥+

0 10 20 30 40 50 60
Zrf, IN SOLVENT (male To)

Fig. 2.3.18, Effect of Solvent Composition on the
Reaction 3Cr++‘=‘- 2Cr+++ + Cr° When 2 wt % Cr+++
is Equilibrated with Excess Cr® in the Solvent LiF-NaF
(60-40 mole %) + ZrF , at 800°C.

Table 2.3.11. The 3CrF2;—_-—'\2CrF3 + Cr® Equilibrium in LiF-NaF-ZrF, ot 800°C

(2wt % ottt Equilibrated with Excess Chromium Metal)

ZrF ; added (mole %) to LiF-NaF 0 15 25 35 47 55
(60-40 mole %)
Cr found by analysis (wt %) in
duplicate experiments
cettr ottt 2.59 2.76 2.80 2.79 2.76 2.83
2.62 2.70 2.78 2.77 2.75 2.86
cett 1.38 2.10 2.30 2.30 2.52 2.82
1.25 2.21 2.27 2.34 2.55 2.50
Calculated Concentrations (wt %)
Average Cr' ' 1.32 2.16 2.28 2.32 2.54 2.66
Average Cr'tt 1.29 0.66 0.50 0.46 0.22 0.18
K, = (cha)z/(chz)3 gx 10-1  4x10~2  2x10°2 1x1072 3x107¥  2x7107°

95

MOLTEN-SALT REACTOR PROGRESS REPORT

Cr*** for fluoride ions is apparent from the decreas-
ing stability of Cr**7,

Similar experiments at lower temperatures and
lower dilutions of chromium ion consistently showed
the same trend. Because of minor quantitative dis-
crepancies, which appeared presumably as a result
of analytical difficulties, the numerical results are
mainly of qualitative interest and have been omitted
from this report. The same is true of numerous
trials involving the addition of CrF,, with no chro-
mium metal or Cr*** initially present, to check the
reaction 3CrF, — 2CrF, + Cr° Increasing dispro-
portionation (more CrF,) with decreasing ZrF,
content was quite evident. Since the chromium
metal was deposited on the nickel walls at an ac-
tivity of less than unity, the ratios of CrF, to CrF,
were higher when only CrF, was added than for the
equilibrium values in the presence of excess chro-
mium metal.

Tests were also made in solvents containing
BeF,, which is considered to be more basic, or less
acidic, than ZrF ,, since it is a better fluoride donor
for furnishing fluorides for complexing. An appre-
ciable amount of disproportionation, CrF3/CrF
0.15 for 2 wt % Cr at 800°C, was found in LiF-BeF,
(48-52 mole %). This is roughly comparable to the
amount of disproportionation in the 35 mole % ZrF
solvent. In another BeF , mixture (LiF- BeF, ThF
67-23-10) the effect of ThF (acidic) was more fhan
counterbalanced by the |ncrecsed LiF (basic) con-
tent to give a net increase in the complexing of
CrF,. Here the equilibrium Cr|:3/CrF2 ratio was
about 0.4 at 800°C with 2 wt % Cr; thus the solvent
containing ThF , has about the same complexing
tendency as the 15 mole % ZrF, solvent. These
results confirm the supposition thcf BeF, mixtures
are more basic than corresponding ZrF, mnxtures
and that the extent of dlsproporflonohon is greater
in the more basic solvents. Numerous auxiliary
trials were made to ensure that solubility limits
were not exceeded and to explore the effect of lower
dilutions and temperatures; the results were in good
qualitative agreement with the foregoing conclu-
sions. Quantitative discrepancies, such as a lack
of material balance, were found that appeared to be
due to experimental difficulties, particularly with
respect to analyses.

Ferrous Fluorides. — Experiments similar to those
with chromium fluorides were also carried out with
FeF, and FeF3 in the same solvents. No dispro-
portionation of FeF, was detected in any of the sol-
vent compositions at either 600 or 800°C. In the

96

course of these experiments it was noted that the
solubility of FeF, in LiF-NaF- ZrF, (27-18-55
mole %) is 4.4 wt % Fe at 800°C ond 0.3 wt % at
600°C; in LiF- NaF-ZrF4 (32-21-47 mole %) the
solubility is 0.4 wt % at 600°C. For comparison the
solubility of FeF in NaF- ZrF, (53-47 mole %) at
600°C is 0.4 w'r %, as defermlned by electrometric
measurements, 2 and 0.26 wt %, as determined by
an earlier filtration experlmenf

Activity Coefficients of CrF, in NuF-ZrF4
C. M. Blood

The activity coefficients of CrF, dissolved in the
molten mixture NaF- ZrF (53-47 mole %) at 850°C
were determined and reporfed previously. 27 The re-
sults of experimental measurements at 750°C are
summarized in Table 2.3.12. The tabulation gives
the experimentally determined partial pressures of
HF and mole fractions of CrF,, together with the
resulting equilibrium quotients. In each case the
partial pressure of hydrogen was determined by sub-
tracting the partial pressure of HF from the meas-
ured total pressure, which was kept essentially
constant and very close to 1 atm. The equilibrium
quotients are also shown graphically as functions of
the mole fraction of CrF, in Fig. 2.3.19.

Examination of the results indicates that, within
the experimental precision, the results are inde-
pendent of the mole fraction of CrF, over the range
from 1 to 10 mole %. If the extrapolation of the
values of K, is valid to infinite dilution, as shown
in Fig. 2.3.19, the activity coefficient of Crr:2 is
unity, over the concentration range studied, with
respect to the standard state of reference having
unit activity coefficient at infinite dilution. Com-
parison of the extrapolated equilibrium quotient with
the equilibrium constants calculated from tabulated
values of thermodynamic properties?® for the re-
actions:

CrF2 (s) + H, (g) #Cr (s) +

25

L. E. Topol, private communication.
6 . . .
J. D. Redman, private communication.

e M Blood, MSR Quar, Prog. Rep. Jan. 31, 1958,
ORNL-2474, p 105.

28 L. Brewer et al,, Natl. Nuclear Energy Ser. Div, |V,
198 (1950).

PERIOD ENDING OCTOBER 31, 1958

Table 2.3.12, Equilibrium Quotients at 750°C of the Reaction Cl'F2 (d) + H2 (g) =— Cr {s) + 2HF (g)
in Nt:lF-ZI'F4 (53-47 mole %)

Pyr XCer _ Pur xCrF2 -

{atm) {mole fraction) * {atm) {mole fraction) *
x 1073 x 1072 x 10~4 x 1073 x 102 x 10~4
4,23** 9.14 1.97 2,63*%** 3.42 2.03
4.97*** 9.18 2.71 2.51** 3.38 1.87
4,38*%** 9.23 2.09 2,76%** 3.44 2.22
5,12%*x 9.23 2,85 2,58*** 3.40 1.96
4,78*** 9.25 2.48 2,55%** 3.44 1.90
4,57*** 9.22 2.28 2,87*** 3.44 2.40
4,31** 9.25 2.02 2,66*"* 3.46 2.05
5.33* % 9.42 3.03 2,71+ 3.44 2,14
4,27** 9.44 1.94 2,70%** 3.50 2.09
4,61** 9.44 2,26 2.60** 3.52 1.93
4,96*** 2.50 2.60 3.02*%* 3.59 2,55
4.84*** 9.56 2,46 2,93*** 3.53 2.44
4,41%* 9.48 2,06 2,69*** 3.55 2,04
4.39%* 9.56 2.03 1,31%** 0.953 1.80
5.56*** 9.63 3.23 1.26** 0.943 1.69
4.72%* 9.71 2.31 1.62%* 1.07 2.47
4.86** 9.60 2.47 1.59** 1.09 2,33
5.56%** 9.88 3.15 1.53** 1.05 2,24
2,18** 3.07 1.30 1.54** 1.05 2.27
2,05** 3.07 1.37 1.57** 1.07 2,32
2.03%** 3.09 1.34 1.60** 0.988 2.60
2,18** 3.13 1.52 1.70** 1.03 2.82
1.91** 3.13 1.17 1.55** 1.05 2.30
2,15** 3.13 1.48 T.49* 1.07 2.09
2.08** 3.09 1.40 1.61*** 1.09 2.39
1.98** 3.13 1.25 1.84*** 1.10 3.07
2,09%* 3.13 1.40 1,79*** 1.09 2.96
1.99** 3.09 1.28 1,69** 1.14 2.51
2,28*** 3,13 1.66 1.86%** 1.07 3.26
2.06** 3.13 1.36 L71** 1.10 2,65
2,00** 3.1 1.29 1,89%** 1.16 3.08
1,93** 3.13 1.1¢9 2,06*** 1.22 3.49
1.91** 3.15 1.16 2, 12%** 1.26 3.58
2.05** 3.11 1.35 1.92** 1.26 2.93
2,31*** 3.13 1.71 2, 11 *** 1.28 3.49
2,84**%* 3.32 2.43 2,07*** 1.32 3.26
2.51** 3.42 1.85 2,16** 1.39 3.35
2,28** 3.38 1.54

Av 2,21 10,52

*K, = PE-I F/XCrF Py where X is mole fraction and P is pressure in atmospheres,

**Determined under reducing conditions.

***Determined under oxidizing conditions.

97

MOLTEN-SALT REACTOR PROGRESS REPORT

e DETERMINED UNDER *

-3 UNCLASSIFIED
(:é% ) ORNL-LR—DWG 34544
. { i [
_ -4 o DETERMINED UNDER
K= (224 £ O-f’f’ x 10 REDUCING CONDITIONS
0.80 K, {s) = 4.2 x10
-4
K,(1) = 8.6 x 10 OXIDIZING CONDITIONS
0.60 Y(s) = 0.526 B, ) .
o y(1) = 0.257
0.40 —1-
"‘c?'o _ e ¢ o
I A, A
020 f=z====s Vi it @ffi; G e T
0
0 1.0 20 3.0 40 5.0 6.0 7.0 8.0 3.0

(x1079)

CrFy IN LIQUID (mote fraction)

Fig. 2.3.19. Equilibrium Quotients at 750°C of the Reaction CrF2 (d) + H2 (9) ==Cr (s) + 2HF (g) in Na F-ZrF4

(53-47 mole %).

and
CrF, (1) + H, (g) ==Cr (s) +

+ 2HF (g), K_(I) = 8.6 x 1074,

yields values of activity coefficients of CrF, of
0.53 and 0.26 with respect to the solid and liquid
standard states respectively at 750°C. The corre-
sponding values at 850°C (ref 27) are 0.28 and 0.18.

The results of this investigation and of the corre-
sponding investigations of NiF, and FeF, and of
the activities of chromium in alloys?® can now be
applied to the study of corrosion of Inconel and
INOR alloys, where it is believed that the rate of
corrosion is diffusion controlled. Experiments
might be carried out where corrosion would be in-
duced by the passage of HF-H, mixtures through
salts in contact with the alloys. By controlling
composition of the gas mixture it may be possible
to attack the chromium without significant effects
on iron or nickel corrosion.

Equilibrium Amounts of CrF, in Molten Salts
Containing UF ; Which Are in Contact with Inconel

R. J. Sheil R. B. Evans

The results of two previous invesm‘igc:tionsw'30
have been combined to calculate the equilibrium
amounts of CrF, in two molten salts containing UF
in contact with an infinitely thick Inconel container

for an infinite period of time. The calculated values

29M. B. Panish, R. F. Newton, W. R, Grimes, and ¥, F.
Blankenship, J. Pbhys., Chem., 62, 980 (1958).

30|, G. Overholser and J. D. Redman, private
communicatjon,

98

are presented in Table 2.3.13 and in Fig. 2.3.20 as
functions of temperature.

The following assumptions were made for the cal-
culations:

1. All the CrF, present at equilibrium resulted
from the reaction

Cr° + 2UF ;== 2UF, + GrF,

2. The activity coefficient for chromium in In-
conel is independent of temperature and salt com-
ponents. The value utilized was 0.563.2°

3. The average atom fraction of chromium in In-
conel is 0.1688.

4. The salt contained no UF; or CrF,, initially.

5. The mean temperature which defined the equi-
librium values corresponds to the balance-point3!
temperature for polythermal flow systems.

31
R. B. Evans, MSR Quar. Prog. Rep. . 31, 1938
ORNL-2474, p 107. € 8. Rep- Jan 2%

Table 2.3.13. Equilibrium Amounts of Chromium in
UFA-Containing Salts in Inconel

Equilibrium Cr (ppm)

Temperature
(OF) NflF-ZrF4'UF4

(50-46-4 mole %)

LiF-Ber-UF4
(47.8-51.7-1.5 mole %)

1000 717 242
1100 785 345 .
1200 852 467
1300 920 600
1400 985 734

UNCLASSIFIED
ORNL~-LR-DWG 34545
1400
1000 —
o)
_am®
__ 900 a6
5
Z 800 L
Q 700
= a L7
o ((\0\e
[ a]
=z 1/\'
g 600 g B
1
s} &
= axe
2
2 500 57
i NS
(&)
400
- / // iiiii
200
1000 1100 1200 1300 1400

MEAN SALT TEMPERATURE (°F)

Fig. 2.3.20, Calculated Equilibrium Amount of

Chromium Present in UFA-Confuining Salt in Inconel,

The chromium concentrations shown also repre-
sent the amounts of chromium metal which should be
added to pre-equilibrate a salt prior to loop tests.
Under this condition, the corrosion rates will de-
pend only on the hot-to-cold-zone transfer mechao-
nism. Such results will be highly indicative of the
long-term corrosion rates to be expected in molten-
salt reactors.

Chromium Diffusion in Alloys

R. B. Evans R. J. Sheil
W. M. Johnson

Studies of chromium diffusion in chromium-con-
taining nickel-base alloys were continued along the
lines described previously.32 New equipment is
being developed for the depletion-method experi-
ments that are based on the rate of loss of Cr37 (as
Cr H*) from a noncorrosive salt contained in the
alloy being studied. The pots used in previous
tests were costly, and the experimental procedures
for their use were complicated. They were designed
to assure a constant ratio of the salt-exposed area
of the container to the salt volume (A4/V). Capsules

325, B. Evons and R. J. Sheil, MSR Quar. Prog. Rep.
June 30, 1958, ORNL-2551, p 95.

PERIOD ENDING OCTOBER 31, 1958

are being designed to replace the pots, and equip-
ment for testing several capsules simultaneously is
being developed.

The constant-potential method was utilized further
to prepare specimens for the study of the distri-
bution of Cr®! within the alloy. In these investi-
gations, measurements are made of the gain of Cr5!
by a metal specimen exposed to molten salt con-
taining a dissolved amount of chromium 51 (as
Cr***) that remains constant with time. It has been
found that the results obtained by this method are
extremely sensitive to the manner in which the
specimens are prepared prior to the experiment.

The diffusion coefficients resulting from initial
experiments, which were conducted with specimens
which had been polished but not hydrogen fired,
were one to two orders of magnitude higher than
those obtained with specimens that had been pol-
ished and then hydrogen fired at 1100 to 1200°C for
3 hr. This information is of importance with respect
to the use of tracers in corrosive media, since a
large portion of the dynamic corrosion studies at
ORNL are conducted in nonhydrogen-fired con-
tainers.

The over-all diffusion coefficients obtained to
date, including data obtained by Gruzin and
Fedorov3?3 and by Price et al., 4 are presented as
log D vs 1/T plots in Fig. 2.3.21. Little signifi-
cance is attached to the low values for the hydro-
gen-fired specimens tested at 640 and 675°C, since
the radioactive count of the Cr>! was virtually at
background level for these measurements.

The coefficients indicated at 750, 700, and 640°C
are based on the results of two or more determi-
nations at each temperature. The data from which
they were obtained are plotted on Fig. 2.3.22. The
straight lines indicate that, with the exception of a
small discrepancy around zero time, the square-root
time law utilized for the calculations is valid.

Equipment for sectioning specimens by electro-
polishing techniques is now being developed jointly
with members of the Dynamic Corrosion Group of the
ORNL Metallurgy Division. The technique thus de-
veloped will be used for tracer studies of chromium
diffusion phenomena in thermal-convection loops. It
is felt that a comparison of diffusion coefficients

33p, L, Gruzin and G, B, Fedorov, Doklady Akad,
Nauk. S.5.5.R. 105, 264 (1955).

34R. B. Price et al., A Tracer Study of the Transport
of Chromium in Fluoride Fuel Systems, BMI-1194 (June
18, 1957).

MOLTEN-SALT REACTOR PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 343546

TEMPERATURE (°C)

1000 900 800 . 700 600

ace IINTO NICIKEL (GRUZIN AND FEDoéov, REF 33)
| & CAPSULE EXPERIMENTS (PRICE &/ o/, REF 34) _____|
® POT WITH CONSTANT A/V AND FILLED WITH

NGF —KF=LiF (11.5-46.5-42 mole %)
-9 [ 0 CONSTANT POTENTIAL EXPERIMENTS WITH UNFIRED |
SPECIMENS
_10 |® CONSTANT POTENTIAL EXPERIMENTS WITH HYDROGEN
FIRED SPECIMENS
-1 ”“‘—\

~

log,, D (WHERE D IS IN cm%sec)

’l
?
/
‘I

¢

!

F—r

7.5 8C 85 90 95 100 105 #1.0 15

104/7 (°K)

Fig. 2.3.21. Current Results of Chromium Diffusion

Experiments.

obtained from salt behavior with those obtained from_
distribution curves is essential to a complete under-
standing of the diffusional processes which take
place within the metal. It is hoped that the results
of these experiments will provide a means for pre-
dicting the void penetration distances to be ex-
pected for a given set of corrosion conditions.

CORROSION OF METALS BY ALUMINUM
CHLORIDE

R. E. Moore

Corrosion of Inconel by Aluminum Chloride Vapor
in a Fused Silica Container

Purified aluminum chloride vapor in a container of
fused silica at 730°C was found to have a very de-
structive effect on Inconel. The results of the ex-
periment showed that the aluminum chloride had
reacted with the silica to produce silicon tetra-
chloride, which attacked the metal. In essence,
this experiment was a corrosion test of Inconel ex-
posed to silicon tetrachloride rather than to alu-

minum chloride. Aluminum chloride should be much

100

UNCLASSIFIED
ORNL~LR-DWG 34547

(x10‘3) [
/ |
100 //
90 :
80 A +
. 1(0 | l
s/ /!
70 o 17 /
& y
A S
60 O &

50

o
40 Q/’ O /

ACTIVITY IN SALT PER UNIT WEIGHT

o

£

30 1P

Q}?/ l\% 1
N

—
]
|
|

20 / cg\)q ’
v .

/ & [

10 ) 1
={762 ppm

’ +
/ /.MELI\SUREMENTS ot 6a0°C,Lert ]
0 /b 2 2!

0 1 2 3 4 5 6 ?| 8
SQUARE ROOT OF EXPOSURE TIME (hr'2)

RATIO OF Cr3' ACTIVITY IN METAL PER UNIT AREA TO Cr5!

Fig. 2,3.22. Time Dependence of Chromium Diffusion

Data Obtained in Constant Potential Experiments.

more stable toward Inconel on the basis of free
energies of formation. The three polished Inconel
strips that were tested lost about 1% in weight in
500 hr and were covered with a dull coating, which
was identified by x-ray diffraction examination as
mainly Ni Si. Large amounts of chromous chloride
and ferrous chloride, identified by x-ray and spec-
trographic examinations, were found in the aluminum
chloride in the part of the apparatus that constituted
the reservoir. This aluminum chloride reservoir had
been held at about 180°C during the test in order to
maintain a pressure of aluminum chloride of about

1 atm. A small amount of nickel was also found by
analysis to be present in the aluminum chloride in
the reservoir. A white coating on the fused silica
was identified by x-ray diffraction as an aluminum
silicate.

Corrosion of Inconel by Aluminum Chloride Yapor
in an All-Metal System

A further corrosion experiment was carried out
with an all-metal system in order to avoid the for-
mation of SiCl,. Purified aluminum chloride gas
was maintained in contact with the walls of an In-
conel tube for 300 hr at 720°C. A copper metal

reservoir containing aluminum chloride, which was
attached to the Inconel tube, was held at about
150°C to maintain a pressure of aluminum chloride
of about 125 mm Hg during the test.

Metallographic examination®3 of the metal showed
the complete absence of subsurface voids. A photo-
micrograph of the test specimen (Fig. 2.3.23) shows
very large grains 1o a depth of about 10 mils in com-
parison with the untreated fine-grained Inconel
sample (Fig. 2.3.24). The grain growth, as well as
the irregular nature of the surface, is probably a re-
sult of the heat treatment or hydrogen firing rather
than the corrosive attack. Chemical analysis of the
contents of the reservoir showed 3.0 mg chromium,
0.2 mg iron, and 0.5 mg nickel. As expected, chro-
mium was attacked to a much greater extent than
iron or nickel. The chromium in the reservoir repre-
sents attack on the Inconel of the equivalent of
0.024 mil over the 44-cm? surface area exposed to
attack at 720°C. Considering the worst case and
assuming that the reaction does not become dif-
fusion controlled after the initial attack on the sur-
face, the amount of attack would be proportional to

35E xamination performed by R. S. Crouse of the ORNL
Metallurgy Division.

UNCLASSIFIED
T-15846

Fig. 2.3.23. Tube Wall of Inconel Test Specimen
Which was Exposed to Aluminum Chloride Vapor for
300 hr at 720°C. 250X,

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
T-15845
/ ‘-
~i, oy
- = X {2 ¥ ;.- \
S e .Q',/ e ? N 1 <(i
/) = X a . 3
" 7 [ " .' . ?/ 09
U I. ‘ { e " D ‘C? Y r
= s (/v.‘ = L R 4 1T
] g . . ,‘ ’r -
’\ K 4 - 'y, " k’., P 2 :«2 (“\ )
t \ s — < 0l
o <N ; 2 »
‘2 1 AN Q > _F‘ i - Are A= 1
¢ g o % ) R et ol 4 L.ol2
: (.25 L e o 2
“® \ — ‘/. L‘/ ( S .
S p = . y s t A -v__\ \ o3
T Q:'A // "I'- - | -‘ A :
v.‘f_,— — \ A RS : {_ s ' ek < - 014
3 C . ' ~ . .|
i . v 3 VS < ¥
~ . v ’ “/‘ L (/ jf‘ ols
& L [ <=5 N2
{\\ N &, .;fi’u P B A g
~. - N \ o, okl L9
Jo L% 5 EAU N N 18
b M . L " < - " - .T
2 O 4T AR TR At S

Fig. 2.3.24. Tube Wall of Untreated Fine-Grained
Inconel Test Specimen. 250X.

the pressure of aluminum chloride and at 5 atm the
corrosion after 300 hr would be about 0.7 mil.

Significance of Experimental Results

The existence of a reservoir of aluminum chloride
at a temperature of 150 to 200°C in a corrosion test
is unfavorable because the relatively cold reservoir
allows condensation of NiCl, and FeCl,,. In future
tests the container will be maintained at 650°C on
one end and 450°C on the other end to duplicate
actual conditions in a reactor. Also, the aluminum
chloride pressure will be maintained at about 5 atm.
Even though it is expected that CrCl, will condense
at 450°C from alloys containing chromium, the
slight attack on Inconel found in the all-metal test
described above is very encouraging. INOR-8, in
which the activity of chromium is much lower than
in Inconel, should prove quite resistant to aluminum
chloride vapor. The corrosion of a number of
metals, including nickel, steel, stainless steel,
INOR-8, and Inconel, is under study.

Theoretical Considerations of Aluminum Chloride
Vapor Corrosion

Theoretical calculations of M. Blander and R. F.
Newton (see following section) suggest that alu-
minum chloride vapor would be an excellent coolant

101

MOLTEN-SALT REACTOR PROGRESS REPORT

for a power reactor. Its primary advantoge over
other gas coolants is the large amount of heat which
could be transferred by the reaction

AlLCl, ==2AICI,

in the vapor phase. An important consideration,
however, in the practical use of aluminum chloride
is a choice of a suitable container material so as to
minimize corrosion.

The reactions of importance in corrosion of metals
by aluminum chloride vapor are

(1) 2AICI, + 3M° ==3MCl, + 2AI°

(2) AICI, + 2A1° &==3AICI

(3) AICH, + M° ==AICl + MCI,

In these equations M represents any component of a
structural metal. Equation 3 was obtained by adding
Egs. 1 and 2.

For the case of a reactor operating at 650°C with
the temperature of the heat exchanger at 450°C,
equilibrium constants may be calculated for the
equations given above for various components from
values of free energies of formation given by
Glassner.38 The principal uncertainty in such cal-
culations is the extent of absorption of aluminum
metal into the surface of the container as a solid
solution.

Equation 3 should be the important reaction for
most components of structural metals because the
pressures of MCl, calculated from this equation are
higher than those calculated from Eq. 1. The equi-
librium pressure of AICI in AlCl3 at 5 atm and
650°C in contact with aluminum metal is about 0.28
mm Hg. The equilibrium pressure of AICI calcu-
lated from Eq. 3 for Fe, Cr, and Ni under the same
conditions is very much less than 0.28 mm Hg.
Hence, disproportionation of AICI should occur only
by the process of absorption of aluminum metal into
the surface of the container. When this occurs there
should be a reduction in pressures of MCI,
action with the aluminum in the container.

The equilibrium pressures of AICI and MCI, at
450°C are less than those at 650°C. Thus, transfer
of metal from the reactor core to the heat exchanger
should take place.

by re-

364, Glassner, A Survey of the Free Energies of For-
mation of the Fluorides, Chlorides, and Oxides of the
Elements to 2500°K, ANL-5107 {Aug. 1953).

102

Another important factor is the vapor pressure of
MCl, at 450°C. If the vapor pressure is lower than
the equilibrium pressure from Eq. 3, MCI, should
deposit in the heat exchanger. Therefore, suitable

vapor should have the
(1) a low oxidation-re-

metals for containing AICI,
following characteristics:
duction equilibrium pressure for the metal chloride
at 650°C and (2) a vapor pressure for the metal
chloride at 450°C that is higher than the equilibrium
pressure at 650°C.

Structural Metal Considerations

Considerations of various possible components of
structural metals has led to the following generali-
zations:

Chromium. — A high chromium content may be un-
desirable because the equilibrium pressure of CrCl
at 650°C over an alloy such as Inconel or INOR-8 is
much higher than the vapor pressure of CrCl, at
450°C. The deposition of CrCl, on the cooler walls
of the heat exchanger might interfere with heat
transfer.

Iron. - The equilibrium pressure of FeCl, is
somewhat less than that of CrClz, but the vapor
pressure’’ of FeCl, at 450°C is higher than its
equilibrium pressure at 650°C. lron would therefore
be expected to deposit in the heat exchanger as
metal (Eq. 3) and not seriously interfere with heat
transfer.

Nickel. — The equilibrium pressure of NiCl,
650°C is less than that of FeCl,. The vapor pres-
sure of NiCl, at 450°C is probably high enough to
preclude deposifion of NiCl, in the heat exchanger.

Manganese. — Manganese appears to be an unde-
sirable component because of a very high equilib-
rium pressure of MnCl .

Molybdenum. — The equilibrium pressures of mo-
lybdenum chlorides are very low. Molybdenum
should be a satisfactory component of a container
metal.

These generalizations are based solely on equi-
librium considerations and have no relation to rates
of corrosion. After initial attack on the surface of
an alloy the rate will probably be diffusion con-
trolled, and it could be very slow.

37C Beusman, Activities in the KCl- FeCl and LiCl-
FeCl Systems, ORNL-2323 (May 15, ]957)

GASEOUS ALUMINUM CHLORIDE AS A HEAT
EXCHANGE MEDIUM

M. Blander R. F. Newton

Aluminum chloride has faverable and unique prop-
erties which make it, as a gas, a heat transfer
medium of potential utility at high temperatures.
Gases ordinarily transfer heat because of the lower-
ing of their heat content upon cooling from a temper-
ature T, to a temperature T, in a heat exchanger
according to the equation

where H, — H , is the heat content increment and Cp
is the heat capacity of the gas at constant pressure.
Aluminum chloride exists as a dimer Al,Cl, at low
temperatures and as a monomer AICI, at high tem-
peratures. Because of the large number of atoms
per molecule the heat capacity of the gas is high.
Yalues of the heat capacities, of AI2CI6 and of
AICI; were estimated theoretically by statistical
mechanical methods, 38 from the infrared vibrational
frequencies measured or estimated by Klemperer.3?

38J. E. Mayer and M. G. Mayer, Statistical Mechanics,
Wiley (1940), p 440.

3w, Klemperer, J. Chem., Phys. 24, 353 (1956).

o

PERIOD ENDING OCTOBER 31, 1958

The values obtained were 42 cal/moles®C for
AlCl, and 19 cal/moles°C for AICl;. For ease of
calculation the approximation was made that the
heat capacity of AlCl; was half that of A|2C|6,
which was equivalent to assuming that the heat ca-
pacity of AICl is 21 cal/moles°C. Table 2.3.14
presents comparisons of the heat transferred per
mole of Al ,Cl,, CO,, and He for various temper-
ature intervals. On this basis alone, it may be seen
from Table 2.3.14 that the heat transfer capacity of
AlL,Cl, is high. A further mode of heat transfer is
possible, however, since at high enough temper-
atures the major gaseous species is AlCl; and at
lower temperatures the major species is Al Cl,.
Cooling the gas in a heat exchanger will remove, in
addition to the heat obtained from the temperature
change, the heat from the reaction

2AICI, == AlCI,

The heat content as a result of the dissociation re-
action was calculated by using a value of 29.6 kcal
for the heat change for the reaction and 34.6 cal/°C
for the entropy change,*® and the data are presented
in Table 2.3.15. The heat changes per mole of
Al,Cl, from the reaction are given in column 5 of

404, Shepp ond S. H, Baver, J, Am, Chem, Soc. 76,
265 (1954).

Table 2.3.14. Heat Transfer Data for AI2C|6' C02, and He for Several Temperature Intervals

1 2 3 4 5

6 7

Heat Change

Temperature Heat from Sum of Heat Changes, Adjusted Heat Change,
Interval (keal/mole) Association of Columns 2 + 5 Column 6 Divided by
(°C) Al,Cl,  CO, He  AICI4 (kcal/mole) (kcal/mole) (1 + ¥)* (kcal/mole)
4501000 12.8 3.6 1.5 8.8 21.6 18.8
450-1300 19.8 5.7 2.3 25.1 4.9 31.7
750—1300 12.8 3.8 1.5 23.7 36.5 26.0
750-1000 5.8 1.7 0.7 7.4 13.2 1n.2
10002000 23.3 7.3 2.8 20.6 43.9 26.6
1500-2000 11.7 3.8 1.4 1.4 13.1 6.6
1000-1500 11.3 3.6 1.4 19.2 30.9 19.0
12001500 7.0 2.2 0.8 6.7 13.7 7.5

*In this expression, (1 + X} is the number of moles of gas molecules at the average temperature of the interval

stated.

MOLTEN-SALT REACTOR PROGRESS REPORT

Toble 2.3.15. Dissociation of AI2CI6 and Its Contribution to Heat Content

Temperature Temperature Fraction of Gas Dissociation Heat Content Due to Dissociation
(°F) (k) at One Atmosphere (kcal per formula weight of AI2CI6)
450 505 0.001 0.03
750 672 0.05 1.48
1000 811 0.30 §.88
1200 922 0.725 21.46
1300 978 0.85 25.16
1500 1089 0.95 28.12
2000 1366 0.997 29.5

Table 2.3.14 for the stated temperature intervals. In
column 6 of Table 2.3.14 are listed the sums of the
heat changes resulting from cooling and the heat
contribution from the dimerization. Since the number
of moles of gas molecules in a forced-circulation
loop is greater than that calculated if it is assumed
that the gas is all Al Cl, the values of the heat
changes must be adjusted to obtain the heat change
of the gus per mole of gas molecules. A crude ad-
justment can be made by dividing the heat change of
column 6 by the number of moles of gas molecules
at the average temperature of the interval stated.
These values are listed in column 7 of Table 2.3.14.
The large values of the heat transferred by this
cycle of association and dissociation further en-
hance the possible utility of aluminum chloride as a
heat transfer medium.

No measurement of viscosity or thermal condue-
tivity of AICI, or A|2CI6 appears to have been
made. Values for these properties are needed in
design of thermal-convection loops for the equilib-
rium mixture of these gases. Accordingly values for
these properties have been estimated for the gases
at 500 and 1000°K and at 1 atm pressure. These
approximations are not claimed to be accurate and
are presented only to serve as rough working figures
for engineering calculations.

The calculations were performed as suggested by
Hirschfelder, Curtiss, and Byrd,“ and it must be

41 J. 0. Hirschfelder, C, F. Curtiss, and R. B. Byrd,
Molecular Theory of Gases and Liquids, Wiley (1954).

104

emphasized that the theory used may be applied
with real accuracy only to the very simplest of
gases. The crudest model of a gas is an assembly
of monatomic, noninteracting hard spheres. Hirsch-
felder et al. show the equations for viscosity, 7y,
and thermal conductivity, A, where the subscript
HS indicates the hard sphere approximation, to be:

VMT

(]) T]HS = 2.6693 x ]0—5 —2 ' g/cm.sec‘
o
T/M
(2) A, =1.9891x 1074 =
52
15R
——7, cal/em®Cisec,
4 M

where
. -« O
o = molecular diameter in A,
T = temperature in °K,
M = gram molecular weight,
R = gas constant.

The molecular diameter may be estimated crudely
for AI2C|6 from electron diffraction data;4? struc-
tural estimates have been published for AlC[a. The
crude approximations which result from these calcu-
lations are presented in Table 2.3.16.

Application of the hard-sphere model to poly-
atomic molecules yields values for thermal conduc-
tivity which are too small. The transfer of energy

42\ | R, Maxwell, J. Optical Soc. 30, 374 (1940).
PERIOD ENDING OCTOBER 31, 1958

Table 2.3.16. Viscosity and Thermal Conductivity of Al2C|6 and AlCi3 Calculated According
to Hard-Sphere Approximation

—7 Viscosity Thermal Conductivity
(gz (g/cm.sec) (cal/cm:°C.sec)
)
At 500°K At 1000°K At 500°K At 1000°K
x 108 x 10~8 x 10~ x 1076
AI2CI6 65 150 212 4.2 5.9
AI('.:I3 40 172 243 . 9.3 13.6
to or from vibrational and rotational degrees of free- T*, defined by
dom (which are obviously missing in the hard-sphere LT
case) permits greater heat transfer per collision for (6) T* =,
£

polyatomic molecules. The anharmonic vibrations
of AICl, and Al,Cl, and the relatively high temper-
atures for the system of interest make it likely that
all the modes of vibration, rotation, and translation
are near equilibrium at all times. |f such equilib-
rium is attained, the Eucken correlation

4 ¢, 3
(3) Ap=Ays | T 5o

where C_ is heat capacity at constant volume, R is
the gas constant, and A is the corrected thermal
conductivity, is valid. Since C_ is about 40 for
A12C|6 and about 17 for A|C|3, the AE/AHS values
are about 6 and 2.9 for the dimer and monomer, re-
spectively. For Al,Cl  the values for A are 24,3 x
10-¢ and 34.3 x 107¢ cal/em«°C sec at 500 and
1000°K, respectively, and the corresponding values
for AICI3 are 27.1 x 10~% and 38.1 x 105, respec-
tively.

The hard-sphere model must be further corrected,
however, for effects of energetic interactions of the
gas molecules (van der Waals forces, etc). These
corrections take the form

@ MTHs
n= Q ’

(5) A e
-

where Q, for the Lennard-Jones potential, can be
calculated from first principles for potential func-
tions of sufficient simplicity. Hirschfelder et al.
tabulate values of £ as a function of the parameter

where & is the gas constant per molecule and € is
the depth of the Lennard-Jones potential welli:

m wae(7) -]

Unfortunately there is no way to estimate € for
AI2CI6 or AICFB. However, Hirschfelder et al. list
values of €/k for a variety of substances. The
values for halogen-containing substances range from
324 for HI to 1550 for SnCl . If these limits for /k
are used, the values for T* and Q shown in Table
2.3.17 result. These values indicate that 7 and A
are smaller than 5, ¢ and A, by a factor less than
3. A value of 2 for Q was, accordingly, chosen ar-
bitrarily, and the resulting reasonable estimates of
viscosity and thermal conductivity for dimer and
monomer of aluminum chloride are presented in

Table 2.3.18.

Table 2,3.17. Probable Limits for Lennard-Jones

Potential Function

Temperature
€ &
°K) /k T Q
500 324 1.5 1.3
1550 0.32 2.7
1000 324 3.09 1.0
1550 0.65 2.0

105

MOL TEN-SALT REACTOR PROGRESS REPORT

Toble 2.3,18. Estimated Values for Viscosity and
Thermal Conductivity of A|2CI6 and AlCI3

Thermal Conductivity

{cal/em:°Cssec)

Viscosity
{(g/cmesec)

At 500°K At 1000°K At 500°K At 1000°K

the usual manner?® by assuming a constant compo-
sition for a change of temperature, AH is the heat of
dimerization, D , 5 is the interdiffusion coefficient
of the dimer A and monomer B, P is the gas pres-
are the weight fractions of dimer
and monomer, R and R“ are the gas constants in the
different units required, and T is the temperature.

sure, W, and W

% 106 % 10~ % 10-8 % 106 The interdiffusion coeffi4cien'r D ,gP may be esti-
mated from the equation?’
A|2C16 75 106 12,2 17.2 ;
0.0026280 / T3[(M , + M )/2¥ ,Mp]
AICl 86 122 13.6 19.1 B
3 (9 D, zP =
UiB Q-

Substances which dimerize have, at temperatures
at which there is an appreciable fraction of both
monomer and dimer, a higher effective thermal con-
ductivity than a gas whose composition is invariant
with temperature.43~4% This increased thermal
conductivity is a result of the change of the equi-
librium constant for the association with temper-
ature. Theoretical calculations?? for this effect
yielded the equation
8 N AH2 D apP WoWg
® e =" g2 R'T 2

where A _ is the effective thermal conductivity, A
is the "“trozen’’ thermal conductivity calculated in

43k, P. Coffin and C. O’Neal, Jr., NACA Technical
Note 4209 (1957).

44 ). N. Butler and R. S. Brokaw, J. Chem. Phys. 26,
1636 (1957).

45}, 0. Hirschfelder, J. Chem. Phys. 26, 274 (1957).

where M, and M are the molecular weights of the
dimer and monomer, O,p i8S the sum of the collision
radii for dimer and monomer, where 0, = ]/2(0A +
ch), and £ is a correction parameter similar to that
described in Egs. 4 and 5 above and which can be
calculated for simple potential functions. It does
not differ greatly from € (ref 48). The values of the
parameters chosen for the calculations described

below were: AH =29.6 kcal, M; = 133.4, 0, =

8.06A, o, = 6.32A, o, = 7.19A, and Q" = 2. Some
of the calculated values of D, P and DABPAHz/
RR’T3 for several temperatures are given in Table
2.3.19.

Values of the weight fractions of the two gaseous
species can be calculated from the equilibrium con-
stant for the dimerization. These were calculated,

46Hirschfe|der, Curtiss, and Byrd, Op. cit., p 534.

47 .
16id., p 539,
8 1bid., p 534.

Table 2.3.19. Calculated Effective Thermal Conductivities of Aluminum Chloride ()\e) and Some

Parameters Needed in the Calculations

2 Thermal Conductiviti .deg-
Temperature DABP DABP AH WB for WB for N " f Aermc:ACfon c:vmes (CUVCH; d:g sec) W
ok 2 —————— p-4.4at P =10 ot . — A for — A, for veroge*® or or
(k) [(cm?/sec)-atm] RR'T? o omm P= 4.t{atm Pe= 10 atm A/ P =2.4 atm P =e'|0 atm
x 10-3 x 10-3 x 10~6 x 10-% x 10-6 x 10~¢ x 10—%
500 21.4 0.92 5x10"4 3.2x10°4 <0.1 <0.1 12 12 12
600 28.1 0.70 6x10"% 39x10-3 2 1 13 15 14
700 35.4 0.56 3.5x10"2 23 x10"? 10 6 14 24 20
800 40.3 0.42 0.13 0.087 24 15 16 40 31
900 51.6 0.38 0.35 0.24 43 35 17 60 52
1000 60.5 0.33 0.65 0.49 38 4) 18 56 59
1100 69.3 0.28 0.86 0.74 17 27 20 37 47

*A/ is a weighed average of the calculated values for AICl, and Al,Cl .

106

as before, by using a value of 29.6 keal for the heat

of dimerization and 34.6 cal/°C for the entropy of
dimerization in the equation

(10) AF°=AH -~ TAS°=-RT In K

The calculated weight fractions of monomer are

listed in columns 4 and 5 of Table 2.3.19 for total

pressures of 4.4 and 10 atm, respectively. Columns

6 and 7 list the contribution to the thermal con-
ductivity from the dimerization process at each of
the two pressures, 4.4 and 10 atm. These values
are to be added to the ‘‘frozen’’ thermal conduc-
tivity given in column 8. The calculated values
of A_are given in the last two columns of the
table. It is to be noted that the largest values of
the thermal conductivity are for compositions
which exhibit the largest change of composition
with temperature, and hence at temperatures at
which there is the fastest release of the ‘‘chemi-
cal’’ heat of association. Such results are a con-
sequence of the theory and lead to the result that
within the range that the ‘‘chemical’’ heat is
available, the average effective heat conductivity
is about 80% of the maximum value. Thus the

most effective temperature range for the utilization

of AICl, as a heat transfer medium is limited to
temperatures for which the composition of the
vapor still has an appreciable fraction of both
monomer and dimer (that is, at least 5% of either).
The effective thermal conductivities and effective
heat capacities are not very different from the
“‘frozen’’ values outside this range.

These calculations contain many approxima-
tions. Although an attempt was made to be con-
servative in these estimates, they should be used
only as tentative working figures until experi-
mental measurements can be made.

PERMEABILITY OF GRAPHITE BY MOLTEN
FLUORIDE SALTS

G. J. Nessle J. E. Eorgan

Tests of the penetration of graphite by molten
salts were continued. Graphite rod specimens
}4, ]/2, and 1 in. in diameter, prepared as de-
scribed previously,? were outgassed and then
immersed in NaF-KF-LiF (11.5-46.5-42 mole %)
at 1700°F for a specific length of time. The
weight gains of the graphite in this medium are

49(5. J. Nessle and J. E, Eorgan, MSR Quar. Prog.
Rep. June 30, 1958, ORNL-2551, p 99.

PERIOD ENDING OCTOBER 31, 1958

compared in Table 2.3.20 with similar data for
graphite tested in LiF-MgF , (67.5-32.5 mole %). 1
The NaF-LiF-KF mixture has a lower density and |
a lower viscosity than the LiF-MgF , mixture.
It may be seen from the data in Table 2.3.20
that NaF-LiF<-KF penetrated the graphite to a
greater extent than did the LiF-MgF ,, mixture.
Chemical analyses of machined cuttings from the
]/2- and l-in. rods exposed to NaF-LiF-KF indi-
cated uniform penetration of the salt to the center
of the rods, as was found to be the case with the
LiF-MgF , mixture.

Table 2,3.20. Weight Gains of Graphite Exposed to
Fused Salts

Average Net Weight Gain of

Rod Graphite Rod (%)
Diameter
(in.) In NaF-LiF-KF in LiF-Mg F2
(11.5-46.5-42 mole %) (67.5-32.5 mole %)

) 9.1
A . 6.8
A 9.8 7.5

1 1.6 8.4

PREPARATION OF PURIFIED MATERIALS

J. P. Blakely G. J. Nessle
F. F. Blankenship

Preparation of Pure Fluoride Compounds
B. J. Sturm

Further attention was given to the problem of
preparing anhydrous chromous fluoride for use in
research related to corrosion problems with fused
fluorides. Chromous fluoride resulting from the
hydrogen reduction of chromic fluoride is con-
taminated with chromium metal, while that made
from ammonium hexafluochromate (l11) usuaily
contains chromium nitride. A difficult recrystalli-
zation is necessary to purify these products., The
product of the reaction of stannous fluoride with
anhydrous chromic chloride is contaminated with
chloride. 59

A different method of preparing chromous fluo-
ride was therefore explored. Stannous fluoride was

30g, J. Sturm, MSR Quar. Prog. Rep. June 30, 1958,
ORNL-2551, p 101.

107
MOLTEN-SALT REACTOR PROGRESS REPORT

heated with granulated chromium metal to over and/or ThF , component, under an HF atmosphere,
1100°C and the meolten material was allowed to and treating the barren melt, at 1500°F, first with
stand at this temperature for a few hours, The H. for 2 hr and then with HF for 12 hr. The melt
products were chromous fluoride and tin metal in is then cooled to below freezing (about 400°F),
well-defined layers that were easily separated. and the UF , and/or ThF  is added. The melt is
The reaction is then reheated to 1500°F and treated with HF for

an additional 4 hr. The melt is then stripped with

SnF, + Cr — CrF, + Sn
2 2 H, gas until chemical analyses of periodic ‘‘thief"’

The chromium-containing portion had the crystal- samples indicate the batch to be acceptable.

; . . 51
lographic properties of pure chromous fluoride. The 9 batches processed included the six dif-
The chemical analysis of the material was in good

) ferent compositions listed in Table 2.3.21. The
agreement with the theoretical analysis. Micro-

analyses of the final batches are given in Table
scopic observation showed the product to be free 2.3.99

of opaque material, and thus the absence of free Batch 1149, for which the analysis showed 1140
metal is indicated.

In cooperation with the ORNL Isotopes Division,
vanadium trifluoride was prepared by thermal de-

ppm of Fe, is being rechecked and if necessary
will be reprocessed or replaced before use.

Copper-lined stainless steel reaction vessels
and copper thermocouple wells and dip legs were
used during this operation of the production

composition at about 500°C of ammonium hexa-
fluovanadate (l11) made by fusing ammonium bi-
fluoride with vanadium trioxide.”? This method

_ . facility, because the relatively high sulfur content
is more convenient than the usual hydrofluorina-

of the beryllium fluoride had previously caused

tion procedures. frequent failures of nickel thermocouple wells and

Fusion of ammonium bifluoride with hydrated
ferric fluoride formed ammonium hexafluoferrate
(111} of the same crystallographic properties as
those reported for the aqueous preparation.53 The

dip legs. The copper dip leg was replaced with a
nickel dip leg when a batch was completed and
was to be transferred to a storage vessel. No

_ e failures of thermocouple wells or dip legs occurred
crystals were isotropic with a refractive index of

1.444. Thermal decomposition at about 500°C in
a helium atmosphere yielded ferrous fluoride
according to the equation C. R. Croft J. Truitt

during this operation,

Experimental-Scale Operations

6(NH ). FeF ,—> 6FeF. + 16NH .t + 24HF* + N_? Eleven small batches totaling approximately
4/3 6 2 3 2 .
88 kg were processed for use in small-scale cor-
Production-Scale Operations rosion tests and physical property studies. Some
of the small-scale units are being modified to
J. E. Eorgan permit their use in a thorough investigation of the
The production facility was activated for approx- graphite permeability problem.
imately 5 weeks of the quarter to process 9 batches

; . . . Table 2.3.21. Beryllium-Containing Mixtures
of various beryllium compositions to be used in

Prepared for Experimentation

component tests. Since the purification process

used for ZrF , mixtures is not adequate for oxide
Constituents (mole %)

removal from these compositions, the procedure Composition

was modified and extended to allow longer hydro- No. UF, ThF, BeF, LiF NaF

fluorination time. In brief, the procedure consists

of melting the blended salts, without the UF 123 1 46 53
124 7 35 58

STy, Insley et al., Optical Properties and X-Ray Dif-
fraction Data for Some Inorganic Fluoride and Chloride 130 1 37 62
Compounds, ORNL-2192 (Oct, 23, 1956).

52B. J. Sturm and C, W, Sheridan, Preparation of Va-
nadium Trifluoride by the Thermal Decomposition of
Ammonium Hexafluovanadate (111), ORNL CF-58-5-95.

33Data on Chemicals for Ceramic Use, Bulletin of the 135 0.5 1 45.5 53
National Research Council, No, 118, p 6, June 1949,

133 13 16 71
134 0.5 1 36.5 62

108

PERIOD ENDING OCTOBER 31, 1958

Table 2.3.22, Analyses of Production Facility Batches of Beryllium-Containing Mixtures

Batch Composition Major Constituents (wt %) Minor Impurities (ppm)

’ No. No. U Th Be F Li Na Ni Cr Fe S

g 1142 130 5.02 9.22  73.4 12.1 40 220 70 2
1143 134 3.14 5.27 9.34 71.3 1.2 65 250 335 2
1144 134 3.24 5.86 8.82 70.5 1.2 100 150 95 6
1145 133 46.5 1.76 43.9 7.76 45 65 190 2
1146 134 3.50 5.88 70 210 150 2
1147 124 26.8 4.80  47.0 21.6 65 15 65 2
1148 135 2.47 3.76 58.4 27.3 20 60 205 35
1149 123 3.61 8.83 61.2 27.0 100 65 1140 6

109

MOL TEN-SALT REACTOR PROGRESS REPORT

2,4, FUEL PROCESSING

M. R. Bennett
G. L

Chemical processing of the fuel and blanket salts
of a molten-salt reactor will be required in order to
remove the neutron poisons and fransuranic elements
and to prepare the fissionable materials and the
solvents for recycling. The fluoride volatilization
process, which was developed for heterogeneous re-
actor fuel processing and was used successfully
for recovery of the uranium from the fuel mixture
(NcF-ZrF4-UF4) circulated in the Aircraft Reactor
Experiment (ARE), appears to be applicable to
processing fuels of the type now being considered
for the molten-salt reactor.! Sufficient laboratory
work has been done to confirm that fluorination of
the fuel salts LiF-BeF -UF , or Li|=-Bef:2—UF4-'|'|'1F4
results in good recovery of the uranium.

Developmental work has also been initiated on
the processing of the solvent salt to prepare it for
recycling. The salt-recovery process is based on
the solubility of LiF-BeF, in highly concentrated
aqueous HF (70-100%) c:nd the insolubility of rare-
earth and other polyvalent-element fluorides, 2 which
results in the recovered salt being decontaminated
from the important rare-earth neutron poisons.

FLOWSHEET FOR FLUORIDE VOLATILITY
AND HF DISSOLUTION PROCESSING OF
MOLTEN-SALT REACTOR FLUIDS

A tentative flowsheet for application of the fluoride
volatility and HF dissolution processes to molten-
salt reactor fluids has been prepared (Fig. 2.4.1).

In this system the uranium is separated from the salt
as UF before HF dissolution of the salt, although
the reverse system also seems to be feasible. The
LiF-BeF, salt is then dissolved in concenfrated HF
(>90% HF) for separation from the rare-earth neutron
poisons. The salt is reformed in a flash evaporation
step, from which it proceeds to a final makeup and
purification step. This last step would perhaps in-
volve the H. and HF treatment now believed to be
necessary for all salt recycled to a reactor.

1G. 1. Cathers, W. H. Carr, R. B, Lindaver, R. P.
Milford, and M. E. Whatiey, Recovery of Uranium from
Highly Irradiated Reactor Fuel by a Fused Salt-Fluoride
Volatility Process, UN-535.

25, W. Jache and G. W. Cady, J. Phys. Chem. 56,
1106-1109 (1952).

110

D. O. Campbell

Cathers
Chemical Technology Division

Some of the specific features of the process are to
be investigated. For example, a high solubility of
UF,

the voloflllzcn‘lon step, except for final recovery.

in aqueous HF would eliminate the need for

This possibility appears to be unlikely, however, in
view of the low solubility of many tri- and quadri-
valent elements. The solubilities of corrosion-
product fluorides are to be determined; a method for
recovering thorium from the waste salt is needed;
and a study is to be made of the behavior of plu-
tonium, neptunium, and protactinium in the process
steps. Experience in the loborufory and in ARE
volatility pilot plant operations? has shown that the
recycled UF is highly decontaminated from plu-
tonium.

EXPERIMENTAL STUDIES OF
VOLATILITY STEP

A series of small-scale fluorinations was carried
out with a 48 mole % LiF-52 mole % BeF , eutectic
mixture containing 0.8 mole % UF ,. (MSR fuel will
contain from 0.25 to 1.0 mole % UF ,, depending on
the operating time.) The eutectic salt was used in- -
stead of the fuel salt in order to permit operation at
lower temperatures, Fluorinations at 450, 500, and
550°C indicated that the rate of uranium removal in- »
creased with the temperature (Table 2.4,1). It is not
known whether the increased fluorination time ap-
pdrently required for quantitative uranium recovery at
the lower temperature was compensated for by a
decreased corrosion rate.
The thorium-containing blanket salt cannot be pro-
cessed at so low a temperature as that used to
process the fuel salt. The uranium concenfration in
the blanket, however, is very low; it has been esti-
mated that with continuous processing at the rate of
one blanket volume per year, the blanket salt
(LiF-BeF -ThF 71-16-13 mole %) will contain *
opproxmafely 0. 004 mole % UF after one year and
0.014 mole % UF , after 20 years.? Fluorinations of

3Communicmion from W. H. Carr at ORNL Volatility
Pilot Plant,

4Moiten-Solt Reactor Program Status Report, ORNL
CF 58-5-3, p 40.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED
ORNL-LR-DWG 28749A

i
—»-UF
iF - 6
LiF-Bef, SALT | FLUORINATOR
U,Th,FP FROM ~ 450°C
REACTOR
SALT HF, H,0 VAPORS SOLVENT
Th,FP - 7
CONDENSER
Y
DISSOLVER > 30 % HF,< 10% H,0
32°C
~10% SALT SALY SOLUTION
SOLUTION SOLID FLASH
IN HF =H20 LIQUID EVAPORATOR
+Th,FP SOLIDS SEPARATOR 100-400°C
CONTINUOUS
BATCH
TH, FP SOLIDS MOLTEN
SOME SALT AND SALT
SOLVENT
HF,H
M2 U, ThF,
HF
H,0 !

WASTE
EVAPORATCR AND

FUEL MAKEUP

PURIFICATION

Th RECOVERY

LiF - BeF,-Thf,-UF,
TO REACTOR

Fig. 2.4.1. Tentative Flowsheet for Fluoride Volatility and HF Dissolution Processing of Molten-Salt Reactor Fuel.

two such mixtures at 600°C for 90 min gave uranium
concentrations in the salt of 0.0001 to 0.0002 wt %,
which were the lowest uranium concentrations ever
observed in laboratory fluorinations. Over 90% of
the uranium was removed in 15 min. It is concluded
therefore that fluorination of uranium from the
blanket salt offers no problem.

The behavior of protactinium in the blanket salt
during fluorination is of interest, although the pro-
tactinium is not lost, in any case, since the salt is

returned to the reactor. A LiF-Ber-Th F4 (71-16-

13 mole %) mixture containing sufficient irradiated
thorium to give a Pa?33 concentration of 5.5 x 10~°
g per gram of salt was fluorinated for 150 min at
600°C; no measurable decrease in protactinium
activity in the salt was observed. Protactinium
volatilization in the process seems to be unlikely,
but the protactinium concentration in the blanket of
the reference design reactor is ~ 104 g per gram of
salt, which is 2 x 104 times larger than the quantity
used in the experiments. Further tests will there-
fore be required for a definite conclusion regarding
volatility.

111

MOLTEN-SALT REACTOR PROGRESS REPORT

Table 2.4,1. Effect of Fluorination Temperature on the
Fluorination of Uranium from LiF-BeF2 (48-52 Mole %)*

Fluerination Uranium in Salt (wt %)

Time After Treatment

(hr) At 450°C At 500°C At 550°C
0 3.39*+ 5.10 4.91
0.5 1.96 0.20 0.55
1.0 0.39 0.17 0.20
1.5 0.21 0.12 0.06
2.5 0.32 0.11 0.05

*No induction period before uranium eveolution.

**5 wt % added; some of the uranium probably pre-
cipitated as oxide.

SOLUBILITIES OF LiF-BeF2 SALTS
IN AQUEOUS HF

Initial measurements of the solubilities of LiF
and BeF., separately, in aqueous HF solutions indi-
cated that both materials are soluble to an appre-
ciable extent in solutions containing 70 to 90 wt %
HF (Tables 2.4.2 and 2.4.3). In general, LiF is
more soluble than BeF, at higher temperatures, but
the effect of temperature on BeF, solubility was not
definitely established. In fact, it is thought that
some of these preliminary values for beryllium
solubility are too low. The LiF solubility decreases
rapidly as water is added to anhydrous HF, and the
BeF, solubility increases from near zero; the solu-
bilities are roughly the same in 80 wt % HF, being
about 25 to 30 g/liter. The BeF. was quite glassy
in appearance and was slow in dissolving. The
solubility values reported for LiF in Table 2.4.2,
except those at 12°C, were obtained after refluxing
HF over the salt for 3 hr. No further measurements
were made with LiF or BeF2 alone, since the solu-
bilities of the two components together were of
primary importance.

The BeF, solubility was measured in nominal 70,
80, and 90 wt % HF containing a known amount of
LiF that was below the solubility limit. The results
of these measurements (Table 2.4.4) show a marked
increase in BeF, solubility with LiF present, as
compared with the results given in Table 2.4.3.

Values obtained for the solubilities of LiF and
BeF, in solutions saturated with both salts are

112

Table 2,4,2, Solubility of LiF in Aqueous HF Solutions

LiF in Solution {mg/g of solution)

Temperature

°C) HF Concentration in Selvent (wt %)
73.6 82.3 90.8 96.2 100

12 10.7 31.7 58.4 74.8 88.2

62* 20,6

47* 41.0

37 62.8 80.4

32.5* 91.4

*Reflux temperature.

Table 2.4.3. Solubility of BeF2 in Aqueous HF Solution

BeF, in Solution (mg/g of solution)

Temperature - -
°C) HF Concentration in Solvent {wt %)
72.8 78.3 90.8 95.2 100
12 45.8 26.3 9.2 2.8 0.0012

Table 2,4.4. Solubility of Ber in Aqueous HF
Containing LiF

BeF2 in Solution
(mg/g of solution)

Temperature LiF Added P ,
€O mafg et solveny e %)
68.6 79.5 90.0
12 7.0 68.8
12 15.0 42.8 15.6
~60 7.0 65.8
—60 15.0 8.2 16.5

presented in Table 2.4.5. Comparison with the
previous results indicates again that LiF strongly
increases the solubility of BeF ,; a plot of the BeF,
solubilities (Fig. 2.4.2) clearly demonstrates this.
The presence of BeF2 appears to increase the LiF
solubility, but to a smaller extent; this effect may
Table 2.4.5. Solubility of LiF and BeF2 in Aqueous HF
Saturated with Both Salts

{10
T |
Salt in Solution {mg/g of solution) 100 ; 'é';: /
Temperature HF Concentration in Solvent (wt %) i /
(°C) 68.6 79.5 90.0 90 ]
LiF BeF, LiF BeF, LiF BeF2 Q SOLI-_lSBILITY_7
80 N 14
12 12,9 824 29.9 53,8 64.2 48,2 \
. SATURATED
-60 8.5 768 22.6 54.2 46.0 58.2 N

be more apparent in more dilute HF where the LiF
solubility is lower and the BeF,
Because of the increase in BeF2 solubility in the
higher HF concentrations when LiF is present, the
90 to 100 wt % HF range was studied further. [t was
expected that the more nearly anhydrous system
would present fewer complications resulting from
hydrolysis. The solubilities obtained in 90 to 100
wt % HF saturated with both LiF and BeF, are pre-
sented in Table 2.4.6. Measurements were made at
the reflux temperatures indicated, at 12°C, and at
~60°C; in some cases, the measurements at ~60°C
were probably taken without the system being at
equilibrium. The BeF, solubility is appreciable
even in anhydrous HF in the presence of a large
concentration of LiF, in marked contrast to the
solubility in the absence of LiF. Thus solubilities

solubility is higher.

PERIOD ENDING OCTOBER 31, 1958

UNCLASSIFIED

ORNL—-LR—-DWG 34548

AN

60

\

/

50

40

SALT SOLUBILITY (mg /g of solvent}

30

i
X
\\

20 /l/
|
1
l

NO Li

65 70

75 80 85

90 95

HF CONCENTRATION IN SOLVENT {wt %)

Fig. 2.4.2. Effect of LiF on BeF2 Solubility in HF-
H20 Solutions at 12°C.

Table 24,6, Solubility of LiF and BeF2 in Solvents Containing 89,5 to 100 wt % HF

Salt in Solution (mg/g of solution)

Temperature

HF Concentration in Solvent {wt %)

(°C)

LiF

95 98

LiF

60

50

69

78 27
60
87 66

98 46

105 49

*Reflux tem

perature.

113

MOLTEN-SALT REACTOR PROGRESS REPORT

high enough for a workable processing system are
available even in anhydrous HF, provided the LiF
concentration can be maintained at the saturation
value.

The effect of temperature on the solubilities is
complicated by the interaction between the two con-
stituents, LiF is less soluble at lower temperatures.
However, the change in solubility of BeF, with
temperature is less obvious, with an apparent in-
crease in solubility at =60°C compared with 12°C,
especially at the higher HF concentrations.

The fuel salt will contain 48 wt % LLiF and 52 wt %
BeF ,, sothe highest fuel solubilities should be ob-
tainable with the HF concentration for which the
LiF and BeF2 solubilities are about the same, that
is, about 90 wt % HF. In order to check this ob-
servation, solubility measurements were made in
80, 90, 95, and 98 wt % HF solutions for an LiF-
BeF2 (48-52 wt %, 63-37 mole %) salt containing
about 0.1 mole % ZrF4, 0.2 mole % mixed rare-earth
fluorides (Lindsay Code 370), and trace fission
products, The salt was crushed; it was not ground
to a powder. Average particles were flakes about
10 mils thick and 50 to 100 mils across. The so-
lutions were sampled 15 and 60 min after salt
addition to permit an estimate of the rate of disso-
lution. The results (Table 2.4.7) indicate that an
appreciable concentration is reached rapidly, but
that the solutions are not saturated, especially with
respect to BeF,, in less than several days. These

results also demonstrate that the fused salt be-
haves similarly to the two components added in-
dividually.

The aqueous HF solutions of the salt were fil-
tered through sintered nickel and radiochemical
analyses were made in order to determine the
fission-product solubilities. The results, pre-
sented in Table 2.4.8, show that rare earths, as
represented by cerium, were relatively insoluble in
the HF solution and were therefore effectively
separated from the salt. The solubility decreased
as the HF concentration was increased. The total
rare earth {TRE) and trivalent rare earth {except
cerium) analyses do not show this separation be-
cause of the presence of the yttrium daughter of
strontium. No rare earth activity other than cerium
was detected in the HF solutions. The slight
apparent decontamination from strontium is not
understood; strontium is expected to be fairly
soluble in these solutions. Cesium is known to be
soluble, but the solubilities of other neutron poisons
have not yet been investigated.

Thus, the rare earths, as represented by cerium,
are removed from LiF-BeF, salts by dissolution of
the salt in an aqueous HF solution; and strontium
and cesium are not. The rare-earth solubility in
HF saturated with LiF and BeF, increases from
about 10~4 mole % in 98 wt % HF to 0.003 mole %
in 80 wt % HF, based on the amount of LiF + BeF,
dissolved; the reactor fuel will contain rare earths
at a concentration of about 0.05 mole %.

Table 2.4.7. Solubility in Aqueous HF Solution of LiF and BeF, from Salt Mixture
Containing 63 Mole % LiF and 37 Mole % BeF,

Solt in Solution {mg/g of solution)

HF Concentration in Solvent {wt %)

Time Temperature
(°C) 79.5 89.5 95 98
LiF BeF, LiF BeF, LiF BeF, LiF BeF,

15 min 12 28 33 50 44 40 17 28 4
1 hr 29 35 51 50 43 17 22 3
20 hr 31 58 68 74

4 days 32 68* 79 82~ 63 28 70* 22
5 hr -60 21 65 74 102 70 36 71* 22
20 he 12 34 100 76 92 88* 53 60* 22

*Essentially all the component indicated had dissolved and therefore the solubility may be higher than the value

given. Insufficient salt was added to some of the solutions to ensure an excess of both components.

114

PERIOD ENDING OCTOBER 31, 1958

Table 2.4.8. Fission-Product Solubilities in Aqueous HF Solutions

Activity in Original Salt and in Selution (counts/min/g of salt)

P Fission ..
Product* Original HF Concentration in Solvent (wt %)
Salt** 79.5 89.5 95 98
x 104 x 10* x 104 x 10 x 104
Gross 3 745 230 200 225 250
Gross vy 225 293 213 244 282
Cs ¥ 174 251 196 223 258
Sr 8 104 73 67 72 75
TRE 8 650 97 92 94 101
Ce 510 7.9 3.6 1.3 0.28
Trivalent g*** 105 73 66 74 81.5

*Zr and Nb precipitated from molten salt before this experiment and were not present in significant
concentration,
**LiF-BeF2 (63-37 mole %) + ™~ 0.2 mole % rare-earth fluoride + trace fission products between 1 and 2
years old.
***This activity later identified as Y90, daughter of Srqo; activity of rare-earth elements 59-71 not

detectable in HF solutions.

115

XN AN

TAC-TWINEFOQOEPP>PIOMECPE-MPOC-ICETME-DO0OEIOEAIMZIOEO>»PMMOME

. Alexander
Bettis
Billington

. Blankenship

. Blizard

. Boch

. Borkowski

, Boudreau

. Boyd

. Bredig
Breeding

. Briggs

. Browning

. Campbell

. Carr

. |. Cathers

. Center (K-25)
. Charpie

. Coobs

. Culler

. DeVan

. Emlet (K-25)
Ergen

. Estabrook

. Ferguson

. Fraas

. Franco-Ferreira
Frye, Jr.

. Gall

. Gresky

. Gregg

. Grimes

Guth

S. Harrill

W. Hoffman
Hollaender

A4 nIr>rom<xeIrrIrm

. S. Householder

H. Jordan

. W. Keilholtz

P. Keim
T. Kelley

. Kertesz

W. Kinyon

. E. Lackey

. A, Lane

. S. Livingston
. G. MacPherson

INTERNAL DISTRIBUTION

50.
51.
52.
53.
54.
55.
56.
57.
58.
59.

60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.

81.
82.

83.

84.
83.
86.
87.
88.
89.
0.
91-94.

95-134.
135.
136-138.

ORNL-2626
Reactors~Power

TID-4500 (14th ed.)

. Manly

. Mann

. Mann

. McDonald

. McNally
Metz

. Milford

. Miller

. Miller

. Morgan

. Murray (Y-12)

. Nelson

. Nessle

. Osborn

. Patriarca

M. Perry

. Phillips

. Reyling

. Roberts

. Robinson

. Savage

D—rr ONENT-A@m>» OO

. Savolainen
Scott
. Seagren
. Shipley
. Skinner
. Snell
. Storto
A. Swartout
. Taboada
. Taylor
. Thoma
. Trauger
. Vonderlage
. Watson
. Weinberg
. Whatley
. Whitman
. Williams
. Winters
J. Zasler
ORNL - Y-12 Technical Library,
Document Reference Section
Laboratory Records Department
Laboratory Records, ORNL R.C.
Central Research Library

OEEEFPONOAMP>PE-MP>PIMIS->IZS-VOP>VEOQT-A-MATE-FICME
T-omC=x44%

mOooOmTTNOomMmI

117
EXTERNAL DISTRIBUTION

139. F. C. Moesel, AEC, Washington
140. Division of Research and Development, AEC, ORO
141.728. Given distribution as shown in TID-4500 (14th ed.) under Reactors—Power category
(75 copies — OTS)

118