ORNL-2440

C-84 —Reactors-Special Features of Aircraft Reactors

This document consists of 324 pages.

Copy/*4i of 273 copies. Series A.

Contract No. W-7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

For Period Ending December 31,1957

A. J. Miller, Project Coordinator

DATE ISSUED

APR 151958

OAK RIDGE NATIONAL LABORATORY

Oak Ridge, Tennessee
operated by i MARIETTA ENERGY SYSTEMS LIBRARIES
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

3 MMSfc. 0251023 t.
 FOREWORD

This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL
summarizes the progress that was made in the development of circulating-fuel reactors
and records the status of work now under way on other ANP programs at the Laboratory
under its contract W-7405-eng-26. The report is divided into five major parts: 1. Air
craft Reactor Engineering, 2. Chemistry, 3. Metallurgy, 4. Radiation Damage, and
5. Shielding.
With the cancellation of work on circulating-fuel reactors as of October 1957,
program emphasis has shifted to research in support of the work of other organizations
participating in the national ANP effort. The major fields of the ORNL effort are now
shielding, materials research and development, and investigations of components of
reactors and of systems designed for the nuclear propulsion of aircraft.

in
 ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART 1. AIRCRAFT REACTOR ENGINEERING being tested as a means of obtaining long-time

1.1. Reactor and Facility Construction
trouble-free operation of the lower seal. A Fulton-
Sylphon bellows lower seal on an auxiliary NaK
A summary of the status of the ART-ETU pump was found to have operated successfully for
program at the time it was cancelled is presented. 936 hr with an average leakage of 16 cm /day.
Reference is made to comprehensive reports that Approximately 8700 hr of high-temperature oper
have been prepared on the various aspects of the ation of primary NaK pumps with NaK has been
development work. Much of the operations planning accumulated, and the accumulated operating time
had been done, and reference is made to documents for auxiliary NaK pumps with NaK is 5100 hr.
containing detailed operating procedures. All A sodium pump equipped with the thermal barrier
outside contract work on the facility in which the designed to protect the bearing-housing region from
ART would have been operated had been com thermal and radiation damage and also equipped
pleted. A pictorial summary of that work is with a simulated ETU type of lubricating and
presented, as well as cost summaries. cooling oil system was found to operate quite
successfully in a high-temperature test. The
1.2. Component Development and Testing system was operated a total of 3800 hr, of which
Tests of actual and prototype ETU and ART 2600 hr were without interruption.
components and systems have been terminated or The heat exchanger and radiator development
are being completed before being terminated. The testing program is being concluded with a thermal
experiments designed for establishing a detailed cycling test of a 20-tube semicircular heat ex
pump sparging procedure were finished. The changer designed to simulate the ART main heat
reactor fuel pump rotary element that was being exchanger tube geometries. The heat exchanger
operated in a new gamma-irradiation facility at has completed 155 cycles of the proposed 300-
the MTR was shut down for repair of an ion cycle program. A tube in ART test radiator No. 2
chamber, and during further inspection it was failed after a total of 181 thermal cycles, which
found that other repairs were needed. This test was six times the number of thermal cycles
will be resumed, since this experimental study proposed for the ART radiators.
of the effect of irradiation on bearings, seals, and Measurements made on the second quarter-scale
lubricants is of general interest. outer core shell model, which was subjected to
An ART-type fuel pump that is operating at 339 thermal cycles, are being analyzed in order
service temperatures with NaF-ZrF4-UF4 (50-46- to determine whether dimensional changes oc
4 mole %, fuel 30) has accumulated 3350 hr of curred. The shell is being prepared for a creep
continuous operation. This pump is to be operated buckling test at 1500°F. A welded shell is to
for an additional 3000 hr under induced cavitation be tested in order to determine the reliability of
in order to determine the effect on the impeller. that type of construction.
An identical pump is being operated with NaK in Zirconium fluoride vapor traps designed for use
order to study the effect of the type of fluid being in the ETU were found to be satisfactory with
pumped. minor modifications. The additional changes that
Operation of a fuel pump in an, endurance test would be required for satisfactory ART operation
was terminated after a total operating time of were established.
4410 hr. Replacement of the lower seal would
1.3. Instrumentation and Controls Development
have permitted further operation, but this test is
not to be resumed. The ART control rod drive system, which was
Operating data on electromagnetic flowmeters successfully operated under simulated ART con
were obtained in a primary NaK pump loop, and a ditions for 3000 hr, was disassembled and ex
split-purge arrangement for the shaft annulus is amined. There were no indications of wear and
no sodium deposits on rod drive parts located the rotor body and the positioner. In addition, the
above the surface of the molten sodium. over-all blade diameter was reduced to give greater
In metering tests made in the aluminum north clearance between the blades and the housing.
head mockup just prior to termination of the ART Units that incorporate these modifications have
program, it was demonstrated that the purged-dip- operated in excess of 350 hr at temperatures
tube continuous level indicator could detect ranging from 1200 to about 1600°F.
changes in level in the fuel expansion tank with In further tests, six strain-gage pressure trans
an accuracy of ±5% at design speed and flow mitters of 0- to 100-psig range indicated an
conditions. Turbulence prevented absolute level average output-input deviation of 0.4% of full
measurements. Considerable work remained to be scale over their operating pressure range from
done in determining the dynamic characteristics ambient temperature to 1400°F. This accuracy
of the measuring system under ART conditions. compares favorably with the accuracies of other
A resistance type of fuel level indicator was pressure transmitters.
designed as a possible alternate to the purged Drift testing of thermocouples was continued,
dip tube, and it is to be tested further. In pre and reports are being prepared that describe the
liminary experiments the performance of the indi design and testing of ART thermocouples. One
cator was in reasonable agreement with the report will cover fabrication techniques and the
calculated performance. results of aging tests in fuel, sodium, and NaK,
Construction work on the resistance-type level and the other will give the results of aging tests
transducers that were to be used in the NaK pump in air.
bowls of the ART and ETU was terminated with
the completion of a total of 22 units. Design 1.4. Engineering Design Studies
work was completed on all the continuous level Work on the applied mechanics and stress
probes required for the ETU and the ART in the analysis problems of the ART was discontinued.
sodium expansion tank, the sodium auxiliary tank, Reports have been prepared that describe the
and the NaK-furnace-circuit drain tank. Tests to techniques used and summarize the results. The
determine the error introduced by an unwetted level results that had been checked by component tests
probe were continued. were found to be valid.
On-off level probes with brazed, rather than
welded, copper-lnconel junctions at the tip were 1.5. Design Physics
installed in a NaK system. After a sufficiently The design of the ART was based on extensive
long operating period these probes will be ex reactor physics studies that were largely com
amined in order to evaluate the use of the braze pleted. The problems studied are described, and
method of construction. Sixteen on-off probes reference is made to reports that give the details
installed in the gas pots of two NaK systems have of the work.
each operated at high temperatures for 3000 hr
without a failure.
1.6. Materials and Components Inspection
The program of tests of 2- and 3/,-in. magnetic A summary of the quantities and types of
flowmeters was cancelled. The data obtained in materials and components inspected is presented.
the operation of 20 such flowmeters are being The various inspection methods used include
processed and analyzed. Indications are that the visual, dye penetrant, radiographic, ultrasonic,
accuracy of these flowmeters was well within the eddy current, and others.
specified ART and ETU requirements. The fifteen
1.7. Heat Transfer Studies
/--in. magnetic flowmeters that had been delivered
for installation on ETU cold traps, as well as Several additional tests were completed in the
parts for 15 more units, will be packaged and high-frequency thermal-cycling loop. In this pulse-
stored. pump apparatus the fuel mixture NaF-ZrF.-UF
Examination of experimental turbine-type flow (56-39-5 mole %) is used as the circulating fluid
meters indicated the need for further modifications. and helium is the pulsing and cover gas. For
The bearing radial clearance was increased, as these recent tests, the Inconel test sections were
was the clearance between the downstream end of made up of welded pieces of tubing and pipe that
were approximately 0.5 in. in inside diameter but data obtained will be combined with heat-gener
varied in wall thickness from 0.029 to 0.250 in. ation rates in order to determine the redistribution
A cursory examination of the data indicated that of flow required to approximate a flat exit-water
the severity of cracking as a result of thermal temperature profile.
cycling increased with increasing wall thickness The thermal conductivity of fuel 30 was re
and with proximity to a weld. This study is being determined, and measurements were made of the
continued. Similar experimental investigations enthalpies and heat capacities of the eutectic
with the use of a pressurized system were dis mixtures LiCI-SrCI2 and LiCI-BaCI2.
continued.
Tests of a full-scale plastic ART core model PART 2. CHEMISTRY
that was divided into three flow regions by in
2.1. Phase Equilibrium Studies
stalling two annular shell liners were also dis
continued. The data obtained, which consist Extensive studies of phase equilibria in fused-
primarily of Pitot traverse measurements and dye- salt mixtures have been made over a period of
injection movies, are being assembled. Summary years in order to provide information needed for
reports were prepared on previous experimental the selection of fuel and coolant mixtures for use
studies of flow through quarter-scale models of in aircraft reactor systems. Reference is made
the ART core, including the screen-packed core to detailed reports on these studies.
and header arrangement, and on the phosphorescent- A comparison of the Russian and ORNL investi
particle technique for velocity-profile visuali gations of the NaF-ThF. system is presented. It
zation. has been found that unclassified Russian reports
Forced-convection heat transfer studies with presently represent the predominant number of
KCI-LiCI flowing in a heated tube were dis contributions to the literature of phase equilibrium
continued after some tests of stainless steel test studies of fused-salt mixtures. Since 1946 more
sections. A summary report has been prepared on than 200 Russian reports have been issued con
the inconclusive results obtained in these tests. cerning the phase equilibria of oxide, silicate,
An experimental system is being designed for chloride, nitrate, and fluoride systems.
the study of forced-convection heat transfer with A disproportionation reaction was found to inter
molten lithium flowing in metal tubes. The test fere with investigations of phase equilibria in the
section will be constructed of stainless steel and NaF-CrF, system. Various efforts to produce
will be instrumented for measurements of wall pure NaF-CrF, samples that are free of Cr
temperatures, fluid flow rates, and mixed-mean have been unsuccessful.
temperatures. The upper operating temperature for the thermal-
A study of the effect of screens in a diverging gradient quenching apparatus has, through equip
annular channel on the thermal structure of a ment modifications, been raised from 850 to
volume-heated fluid is being continued. Also, a 1300°C. In the revised apparatus a single,
system is being constructed for an experimental traveling, recording thermocouple was substituted
study of heat transfer from an internally heated for 18 monitoring thermocouples; the furnace was
liquid metal (mercury, initially). An empirical redesigned to permit the sample tubes to be
relation for forced-convection volume-heat-source protected by a static helium atmosphere during
heat transfer with liquid metals was devised. brazing; and a roller crimping machine was adapted
Preliminary data on burnout heat fluxes for water for use in sample preparation in order to minimize
in nucleate boiling have confirmed the concept sample tube expansion during heating.
that source-vortex flow would delay the burnout Results of the use of the petrographic micro
condition to a higher heat flux. The higher burnout scope for the routine inspection of solidified fuel
values obtained were not counterbalanced by mixtures are described. In such examinations,
excessively large power expenditures. oxygen-containing compounds are detected and
Equipment is being designed and assembled for variations from a nominal composition are revealed.
a study of flow distribution in the coolant annuli An optical method is described for the determi
of the Tower Shielding Reactor II (TSR-II). The nation of equilibrium paths and tie lines in ternary

VII
fused-salt systems. Application of the method to indicate that the effective Z/R average of the
the LiF-ThF .-UF . system is demonstrated. solvent may correspond closely to that predicted
by the previously proposed acid-base scale. An
observation that the value of the activity co
2.2. Chemical Reactions in Molten Salts
efficient of the NiF_ solute is not dependent on
The results of a study of the disproportionation the solubility has verified a previous conclusion
of UF3 in fused-salt fuel mixtures are presented. that the precipitating phase is not NiF..
Nonacidic fuel solvents that contain a high con A summary of the results of studies of the
centration of fluoride donor constituents, such as reduction of UF . by chromium in various fluoride
alkali fluorides, were found to increase the extent reaction mediums is presented, as well as data
of disproportionation by complexing the acidic UF. for the reaction of UF4 with various other metals,
(fluoride acceptor). The ease of alloy formation including Inconel, in which the activity of chromium
between the uranium formed and the container is much less than unity. In all solvents studied
metal also noticeably influences the extent of the the chromium concentration decreased as the UF,
disproportionation. Less disproportionation oc concentration increased. The implications of the
curred in copper vessels than in nickel vessels, studies are analyzed in terms of the corrosion
and tantalum and molybdenum were found to be mechanism. Molybdenum was the only metal found
even more resistant than copper to alloying with to have stability in the NaF-LiF-KF-UF mixture,
the uranium formed. It was also found that UF, and accordingly nickel-molybdenum alloys would
was oxidized by the alkali fluorides. be expected to resist corrosion by this fused salt.
In comparison with UF .-containing fuels, the In the ZrF4-based melts, molybdenum, niobium,
UF3-containing fuels exhibit a marked improvement and tungsten appear to be stable toward UF .
in initial corrosion attack at high temperatures. Tantalum and vanadium appear to offer no ad
Unfortunately, however, the difficulties arising vantage over chromium for either type of melt.
from the deposition of uranium by disproportion Extensive study of the reduction of FeF. by
ation prevented the easy exploitation of UF, as chromium in molten fluorides has indicated that
a solution to corrosion problems. the behavior is far from ideal. An analysis of
Analyses of the results of determinations of the Cr -to-Fe ratios obtained from equilibrium
activity coefficients of FeF- and of NiF_ in quotients determined by the filtration method and
molten NaF-ZrF4 (53-47 mole %) are presented. a comparison with activity coefficients obtained
The numerical values of the activity coefficients from the standard free energies of formation in
of FeF2 are, at 800, 700, and 600°C, 1.46, 2.20, dicates that equilibrium was not attained in these
and 3.28, with respect to the pure crystalline experiments. It is apparent that the filtration
solid; 0.64, 0.66, and 0.66, with respect to liquid method is not capable of establishing the Cr++-
FeF2; and unity, with respect to the standard state to-Fe ratio in these reaction mediums at 600°C.
of reference at infinite dilution. The results of experimental studies of the
The results of similar studies of NiF_ gave very temperature coefficient of the equilibrium
large discrepancies between the calculated and
2UF4 + M^ 2UF3 + MF2
experimental equilibrium constants and very high
values of the activity coefficients. It is clear when molten-salt mixtures are used as the reaction
from the values of the activity coefficients that mediums, were used as a basis for a thermo
when NiF2 is dissolved in NaF-ZrF. (53-47 dynamic interpretation of equilibrium data on the
mole %) it behaves in a manner vastly different corrosion of iron and chromium by molten-salt
from that predicted by the free energies of for fuels. The discussion covers the tendency toward
mation for solid NiF_. It is postulated that the mass transfer. It was expected from the similarity
values available for the free energies of formation of the compounds CrF_ and FeF. that their
of solid and liquid NiF. are not accurate. solution behaviors would be similar and that any
Solubilities and activity coefficients of NiF_ difference would be governed by the difference of
as functions of solvent composition in NaF-ZrF. their AH° values. Experimental data proved this
mixtures are being measured in an effort to locate supposition to be invalid, and it must be concluded
the positions of the maximums of the activity that some of the data are incorrect or that CrF„
coefficient and the solubility. The present results and FeF„ have entirely different solution behavior.
 A summary of the results obtained thus far in The results of a determination of chromium
the determination of the solubilities of various diffusion coefficients in Inconel-molten salt
noble gases in a number of different solvents is systems are described. This investigation was
presented. In all the cases investigated, Henry's undertaken because experiments carried out over
law was found to apply, and the solubilities were the past few years on the migration of metal as
found to increase with temperature. For the same a result of temperature gradients in a molten-salt
solvent, the solubilities of the gases decreased environment suggested that the over-all migration
with increasing solute gas size. Variations in rate is controlled by the diffusion rates within
solvents showed corresponding gas solubilities the metal.
to be approximately the same in two related ZrF4- An analysis of steady-state chromium migration
bearing solvents, but the solubility of helium in in thermal-convection loops is described. In
NaF-KF-LiF was found to be less than half that formation available from the many thermal-con
in the ZrF -bearing solvents. vection loops operated to date is correlated in
The solubility of HF in mixtures in the NaF- order to develop a method whereby the amount of
ZrF4 system has been determined as a function chromium transported can be estimated as a
of temperature, pressure, and melt composition, function of loop operating time, and the extent
and a summary of the results is presented. In all of corrosion to be expected at any time can be
cases the solubility decreases with increasing predicted.
temperature. Also, the Henry's law constants Measurements of the activities of the metallic
for the HF solubility illustrated a constant heat components of container materials for fused-salt
of solution over the experimental temperature range fuels have been made in order to study the effects
used for each composition. of composition and temperature on the activities
Experimental studies of the solubilities of and thus on the corrosion resistance of container
fission-product fluorides in molten NaF-ZrF4 mix alloys. The activity of chromium in nickel-
tures were made. It was found that varying the chromium alloys (Inconel) has been studied, and
ratio of sodium to zirconium in the solvent had the results of studies of the nickel-molybdenum
a considerable effect on the solubility of CeFj. system are presented. In future work the chromium-
The solubility of YF3 in NaF-ZrF4 (53-47 mole %) molybdenum system is to be studied, as well as
was found to follow a pattern similar to the the ternary system nickel-molybdenum-chromium,
patterns obtained for the rare earths. The solu which includes the INOR alloys developed by the
bility values for CeF, and SmF3 in combination Metallurgy Division.
in NaF-ZrF4 (53-47 mole %) were found to be The possibility of dissolving stainless steel fuel
substantially the same as those found previously elements in molten fluorides, with the resultant
with NaF-ZrF.-UF, (56-40-4 mole %) as the advantage of adaptability to the present volatility
4 4
solvent. method for fuel reprocessing, prompted some ex
Measurements were made of the solubility of ploratory experiments. In general it appears that
BaO in KF-LiF (50-50 mole %). The results of a practical process could be evolved for removing
three experiments unexpectedly showed the solu the uranium from the steel and leaving a constant
bility to be quite high. This study is related to nonaccumulative heel in the reactor vessel. This
a systematic investigation of the possibilities of heel could be flushed out periodically with a
removal of fission-product fluorides from NaF- dummy salt mixture.
KF-LiF base mixtures. A comparison study of In experiments for studying the oxidation of
the reaction of CeF, and UF . with CaO in KF-LiF sodium and potassium mixtures it was found that
eutectic was also made. The precipitation of the variables could not be controlled sufficiently
cerium as Ce_0 suggests the possible separation to establish any quantitative effect of original
of tetravalent metals, as oxides, from the trivalent composition on the Na-to-K ratio in the mixed
fission products. Alternatives would be to pre oxides. The data show, however, a significant
cipitate uranium oxide and leave the fission decrease in the ratio of Na to K in the mixed
products in solution or to precipitate the rare oxides upon an increase in the potassium content
earths as oxides and leave the uranium in the of the original alkali metal mixture from 49 to
solvent. 77 wt %.
 2.3. Physical Properties of Molten Materials Experiment and of the establishment and operation
Methods used for vapor pressure studies of of pilot-scale and production-scale purification
apparatus.
molten salts are described, and the results ob
tained in these studies are summarized. Reference 2.5. Analytical Chemistry
is made to a concurrent effort to measure vapor
pressures that was carried out at Battelle Memorial Work was terminated on the development of an
instrument which would serve to monitor the
Institute.
Future vapor pressure investigations of fused exhaust gases from the ART by detecting traces
of NaK in air. The instrument was based on the
fluorides of reactor interest will involve systems
containing BeF_. Difficulty will be encountered measurement of the absorption of sodium resonance
at BeF, concentrations of less than 75 mole % radiation by sodium atoms. It is thought to be
because vapor-phase compounds will contribute applicable to the detection of a concentration of
appreciably to the total pressure. It will be sodium in air that would be equivalent to the
necessary to obtain partial pressures. Methods introduction of a total volume of 6 cm of NaK

for these measurements are being developed. per hour to the exhaust gases. This corresponds
roughly to an increment of 10 ppb of NaK across
A mass spectrometric study of salt vapors is
the NaK-to-air radiators. The test model showed
also being planned in order to obtain an under
that the response of the instrument to small
standing of the nature of dimeric and trimeric
increments in the concentration of sodium in air
molecules in the vapor of alkali halides. Polymer
samples was almost instantaneous, and the model
formation can cause components of salt mixtures
was satisfactorily stable and reliable in all
to be even more volatile as part of an associated
species than the pure component is, in some respects. Other methods of detection which were
evaluated during feasibility studies included
cases. A preliminary study is described for which
potassium chloride was chosen as a suitable
photometric smoke detectors, emission spectrog-
raphy, flame photometry, and surface ionization
material with which to develop the method.
detectors. These methods were not considered to
The results of surface tension determinations on
be as promising as the resonance radiation ab
molten salts by three different methods are
sorption technique because they either lacked
compared. The methods used were the sessile-
adequate sensitivity or presented technological
drop technique, the maximum-bubble-pressure tech
complications that would make their application
nique, and a tensiometer technique. During the
course of these experiments it has become apparent
to the ART extremely difficult.
that further information is needed on the wetting The apparatus for the determination of oxygen
characteristics of fluoride salts.
in fluoride salts with potassium bromotetrafluoride
was put into operation. The gas-measuring system
2.4. Production of Purified Mixtures of the apparatus was calibrated to a precision of
A summary of Experience in the synthesis for about 2% with oxygen. The method was also
use in ANP chemistry studies of compounds that tested on samples of yttrium oxide and is being
were not available commercially is presented. used routinely for the determination of oxygen in
Of particular interest were fluorides of the metals
yttrium fluoride. The precision of the method with
being considered as container materials for various respect to yttrium fluoride appears to be less than
5%„
fuels and the products of their reactions with
constituents of proposed fuels. Corrosion products
found in fuels were identified by comparison of
PART 3. METALLURGY
their x-ray diffraction patterns and optical crystal-
lographic properties with those of synthesized 3.1. Nickel-Molybdenum Alloy Development
compounds. Fluorides of the fission products were Studies
prepared for use in the study of their effects on The developmental work on nickel-molybdenum-
the fuel mixtures of interest. base alloys for use as reactor structural materials
Summaries are presented of the experience in in contact with the fused salt fuel NaF-KF-LiF-
the preparation of fuel for the Aircraft Reactor UF (11.2-41-45.3-2.5 mole %, fuel 107) at normal
operating temperatures of 1650°F and above is strains that would be produced by thermal cycling
summarized. The various alloys studied fell during reactor operation. The results obtained
into three general categories: solution-hardening, in tests in which the strain was mechanically
precipitation-hardening, and dispersion-hardening induced compared favorably with those of a
alloys. The alloys were evaluated in terms of supporting investigation at the University of
fabricability, mechanical properties, corrosion Alabama in which thermally induced strains are
resistance, weldability, and aging characteristics. being studied. The effects of temperature, grain
The results of the evaluations showed that the size, environment, specimen geometry, and cycle
solid solution-hardening alloys most nearly met frequency are summarized.
the specifications. Extensive work has been The shielding materials that have been inves
done in order to establish the final composition tigated included an 80% Mg-20% Li alloy, an
of the alloy selected on this basis (INOR-8) alloy of very pure lead with 0.06% copper, and
and to produce commercial quantities of the alloy. a lead-lithium alloy.
Since the chromium content of INOR-8 was
expected to limit the peak metal-salt interface 3.3. Welding and Brazing Studies
temperature to the range 1650 to 1700°F, an alloy A production facility for the manufacture of
designated INOR-9 was being investigated. In brazing-alloy rings according to techniques and
this alloy the chromium is replaced with niobium, procedures developed by the Metallurgy Division
which does not corrode in fuel 107. Preliminary was set up by the Engineering and Mechanical
data on the properties of this alloy are presented. Division. Approximately 750,000 rings were
produced by the graphite-mold technique, and
3.2. Mechanical Properties Studies a production rate of 60,000 to 70,000 rings per
The mechanical properties of the various day was achieved. Production of these rings at
materials to be used in circulating-fuel aircraft ORNL became necessary because the fabricator
reactors have been investigated during the past of the ART radiators could not obtain satisfactory
several years. Fairly complete design data were brazing-alloy rings.
obtained for Inconel, the structural alloy that A procedure was developed for fabricating
was being used for the ART, for beryllium, the thermocouples for high-temperature service that in
moderator material, and for various shielding volves brazing an end cap to the shield and
materials. Basic information on the behavior thermocouple wires. This procedure replaces
of metals was also obtained in order to be able a welding technique used previously. Closures
to predict the safe operating life of various com made by the welding technique were frequently
ponents under a variety of operating conditions. rejected because of cracks and pores in the
In the investigation of Inconel it was found weld zone. Closures that are satisfactory as to
that, in the fused-salt environment, the fine both integrity and consistency in the distance
grained material is stronger than the coarse between the end of the seal and the thermocouple
grained material at 1300°F but weaker at 1500°F junction are made by the use of the new brazing
and above. It was also found that, in order to procedure.
improve the strength of Inconel without impairing An investigation of optimum methods for fab
its fabricating characteristics, it was feasible ricating the tube-to-tube-sheet joints of the ART
to add carbon through carburization of finished fill-and-drain tank was completed. A unique
tubing. method for brazing the unit was developed that
It was found in the beryllium investigations may be applicable to tube-to-tybe-sheet joints of
that data obtained on a laboratory scale could other types of heat exchangers.
be reliably applied to large-scale pressings. Investigations were started of methods for
Methods for interpolating and extrapolating the welding niobium. Fusion welds were tested in
available data on mechanical properties have simple bend tests and found to be ductile. Ex
been explored, and several relationships have periments are being carried out in order to deter
been found that have merit. Empirical equations mine differences in the weldability of arc-cast
based on specific data are available for Inconel. and powder-metallurgy niobium. Filler wires for
Tests of both rod and tube specimens of Inconel welding niobium are also being tested. Both
were made in order to evaluate the effect of butt welds and tubing saddle welds have been

XI
fabricated in an argon-atmosphere dry box by operation because there was no change in the
using the tungsten-arc process and zirconium- location of the cylinders.
niobium filler wire. Impurities in the welding Extensive grain-boundary penetration was found in
atmosphere have been found to have an embrit the hot-leg areas of niobium thermal-convection
tling effect. loops in which lithium was circulated. These
Brazing tests were made on carburized Inconel tests were conducted in a vacuum chamber.
tubes. The desired fine-grained structure was This penetration is thought to be due to attack
retained. Extensive grain growth would have on impurity phases, such as carbides, in the
occurred in conventional Inconel subjected to grain boundaries. Experiments are planned in
the same brazing cycle. which various grades of niobium will be tested
in order to determine what factors control grain-
3.4. Corrosion and Mass Transfer Studies boundary penetration.
In continuing screening tests of new brazing A forced-circulation loop is being designed for
alloys, Handy and Harman brazing alloys No. 91 studying lithium-niobium systems. The niobium
(91.25% Ni-4.5% Si-2.9% B) and No. 93 (93.25% tubing of the loop is to be protected from air
Ni—3.5% Si—1.9% B) were found to have good oxidation by a continuous outer sheath of type
corrosion resistance to NaF-ZrF.-UF4 (53.5-40- 310 stainless steel.
6.5 mole %, fuel 44) when tested in seesaw A summary of the results of the extensive cor
apparatus at a hot-zone temperature of 1500°F rosion tests that were made of Inconel exposed
for 100 hr. to fused fluorides and to sodium and NaK is
In similar screening tests, iron-base brazing presented.
alloys are being exposed to lithium as they
became available. (High-nickel-content alloys have
very limited corrosion resistance to lithium.) 3.5. Materials Fabrication Research
An 86% Fe-5% Si—5% Cu-4% B brazing alloy An investigation of the recrystallization of
showed good corrosion resistance under static niobium and the effects of recrystallization on
conditions but only fair resistance in seesaw- recovery and on tensile properties was made.
furnace tests at 1500°F for 100 hr. Except for the large grain size of recrystallized
Inconel welds made with nitrogen as the shield arc-cast material, no significant differences were
gas were tested for comparison with welds made found between arc-cast and powder-metallurgy
with helium as the shield gas. The corrosion material. Work has been initiated on fabrication
resistance of the welds made in nitrogen was of material for use in forced-circulation corrosion-
about the same as that of Heliarc welds. testing loops. Attempts to fabricate tubing by
Several cermet-metal joints formed by fusion extrusion of compacted pellets were unsuccessful
or brazing of TiC-Ni, WC-Co, and WTiC2-Co because of failures in the tube walls during tube
cermets to metals were found to be corrosion reducing. Billets of arc-cast material are being
resistant to molten sodium and to NaF-ZrF.-UF. secured and will be extruded to produce tube
4 4
(50-46-4 mole %, fuel 30) in 100-hr exposures at blanks for redrawing. Methods for the protection
1500°F. These cermet-metal joints were developed of pump parts and other odd-shaped sections of
to permit these types of cermets to be joined a niobium loop from oxidation during testing are
to Inconel by brazing. being investigated. Sprayed coatings of brazing
An examination was made of the rare-earth-oxide alloy appear to offer a possible solution to the
ceramic cylinders that constituted a prototype problem.
ART control rod and which were exposed to Further development of equipment has provided
molten sodium for 3000 hr in a rigorous test of means for producing yttrium sponge of increasing
the control rod drive system. Some of the ceramic purity. Analysis of a recent batch of sponge
pieces were broken as a result of rather large detected only 1000 ppm of oxygen, and a sample
thermal changes. There was no detectable cor of this sponge after arc-melting was reduced 37%
rosion of the ceramic cylinders by the sodium. in area by cold-swaging.
Despite the breakage of some of the cylinders, Equipment is being prepared to make possible
the rod would have been satisfactory for ART the direct transfer of the molten fluoride charge
into the reaction retort in order to eliminate con showed signs of incipient failure. The evidence
tamination from handling. Molten-state transfer of of incipient failure was similar to that found
the lithium reductant is now being practiced. previously in creep test specimens.
The work on the neutron shield and the shield
3.7. Nondestructive Testing
plug for the ART fuel pumps was completed, and
a report describing the development work was The inherent accuracy of the new eddy-current
prepared. thickness gage was increased by the replacement
The tensile strength and ductility of core material of the previously used variable-frequency signal
for tubular control rods were determined from generator with drift-free transistor oscillators. Small
extruded specimens. These specimens were pre cup cores that will produce symmetrical electro
pared from especially processed Lindsay oxide magnetic fields and will decrease the size of
and had somewhat greater ductility than previously the field are being constructed to improve eddy-
tested material. current cladding-thickness determinations.
Developmental work was initiated at the request Preliminary use of a closed-circuit television
of GE-ANP on high-density BeO bodies for use system for direct x-ray imaging has given en
as moderating material. Very fine particles of couraging results. The technique for use of this
BeO are being obtained by the controlled cal system is similar to that used in conventional
cination of BeS04. radiography. The greatest area of usefulness of
Rods of MgO and of BeO were extruded to the system will be in the inspection of thin
provide rod insulators with four longitudinal sections through which high-intensity radiation can
holes. Rods of this type are required for thermo reach the photoconductive target.
couple assemblies. Routine inspection of tubing, pipe, and plate
Methods are being developed for hydriding dense was performed by using immersed-ultrasound and
yttrium metal when it becomes available. Reactor- eddy-current techniques. A special lot of Hastelloy
grade zirconium, which is available and which B was found to be among the best tested to date.
responds to hydrogen in much the same manner
PART 4. RADIATION DAMAGE
as yttrium does, is being used in developing the
hydriding process. 4.1. Radiation Damage
Examinations of the four forced-circulation loops
3.6. Metallographic Examinations of Engineering that were operated in the MTR with fused fluoride
Test Components After Service fuels as the circulated fluids have been completed.
The results of examinations of the ART pro The final items examined were the impeller
totype test radiator No. 1, which failed immediately housings of the pumps used in loops 4, 5, and
after reaching full power in its first controlled 6. No evidences of erosion were found on the
thermal cycle, are presented. Data on fin-to-tube impellers.
joint braze integrity and the degree of oxidation Examinations were continued of the moderator
of the fin collar are compared with similar data materials that were irradiated in the second ex
on other radiators. Incomplete melting of the posure in the MTR. No cracking or flaking of
braze alloy rings was found in several instances. the graphite slug that was encased in nickel was
Corrosion and mass transfer were negligible, as evident, and the dimensions of the slug after
was expected in view of the short operating period. irradiation were essentially the same as those
It is concluded that the tube failure which per prior to irradiation. Metallographic examination re
mitted the NaK leak could have been the result vealed no reaction between the nickel and the
of at least three things: reaction with an in graphite. When the zirconium hydride slug that
sulating material, damage to the tube during was encased in molybdenum was being sectioned,
fit-up and brazing, and the presence of stringers the casing separated from the slug, and thus it
across the tube wall. appears that there was no bonding. The molyb
Test radiator No. 2, which failed during the denum was very brittle. Metallographic studies of
182nd thermal cycle, was also disassembled and these specimens are continuing in an investigation
sectioned. Forty tubes from the damaged area of significant' differences between irradiated and
were examined, and only one was found that control capsules.
 The out-of-pile tests that correspond to the MTR and to establish a reference of current concepts.
tube-burst tests were completed. The times to With this survey as a basis, subsequent reports
rupture were two to five times longer than in the will present a coherent picture of the progress
MTR tests, but were of the order of a factor of being made in these investigations.
4 shorter than in tests conducted by the Metallurgy
Division under similar conditions. Further in PART 5. SHIELDING
vestigation has revealed that the fused-quartz
insulation used in these in-pile and out-of-pile
5.1. Shielding Theory
tests caused changes in the calibration of the
Chromel-Alumel thermocouples that were operating For most calculations of neutron scattering in
in a helium atmosphere. These changes gave air, isotropic scattering in the center-of-mass
higher temperatures than those indicated. The system is assumed, although it does not represent
higher temperatures partially account for the the actual case. For purposes of comparison, a
shorter times to rupture, but they do not entirely calculation was performed in which anisotropic
account for the loss of ductility being greater scattering was assumed. The results from the
in the in-pi le tests than in the out-of-pile tests. calculations indicate that the isotropic scattering
Results of extensive tests in the LITR are also method of calculation is conservative.
reported. The tube-burst specimens have ex
Another calculation was performed to determine
hibited the low ductility found in the MTR test
the spectra of bremsstrahlung radiation resulting
specimens.
from the degradation of electrons in sodium iodide
Disassembly of the vertical in-pi le loop which and in air. The assumptions made, the equations
circulated fused salt fuel in the LITR for 235 hr used, and the results obtained are presented.
was completed, and various parts are being
examined and analyzed. The lubricant was found A third calculation was prompted by the need of
to have decomposed and become waxlike. aircraft shield designers to know the production
An Inconel capsule and two Hastelloy B of secondary gamma rays by neutron interactions
capsules were examined that had been irradiated in air. The problem was idealized to that of
in the MTR in contact with fused salt fuel. The determining the dose rate and angular distribution
Inconel was attacked 1 to 2 mils in 304 hr at of gamma rays at a detector located an arbitrary
1500°F; the power density was 6000 w/cm3. distance from a neutron source. The source and
There were thin films on the surfaces of the the detector were considered to be embedded in an
Hastelloy B capsules, but the corrosion was infinite volume of air. The source was taken to be
negligible. Attempts are being made to identify either a point isotropic source of monoenergetic
the film and to determine the mechanism of de neutrons or a point source which emitted a mono-
position. energetic line beam of neutrons at an angle QQ
In an investigation of irradiation effects in photo- with respect to the source-detector axis. For a
conductive detectors, both neutron and gamma - point isotropic source and a 50-ft separation
irradiation have been found to decrease the photo- distance, it was found that the dose rate varies
response of PbS cells, and the decrease is pro with the neutron source energy approximately as
portional to the total integrated flux. There is E~ . When curves for various monoenergetic
threshold damage that seems to depend upon the line-beam sources and a 65-ft separation distance
dimensions of the sample, which, if exceeded, were plotted as a function of #., they were rela
completely destroys the cell. Below threshold tively flat over the region from 30 to 180 deg, and
the cell recovers in a first-order process. The they became flatter as the source energy increased.
photoresponse at 24,000 A and above fails before Since aircraft reactor shields are usually de
that at shorter wavelengths. In addition to a signed so that most of the neutron leakage is
decrease in response, the noise level increases from the rear, the results of the line-beam calcu
with irradiation. lations indicate that a knowledge of the exact
A state-of-the-art survey of irradiation effects surface distribution of low-energy neutrons is of
in semiconductor barriers is presented to unify less importance than a knowledge of the energy
the material presented in previous progress reports spectrum and total neutron leakage.
 5.2. Lid Tank Shielding Facility elements and uranium solutions. The design of
the water cooling system is approximately 95%
A second group of tests of advanced shielding
complete.
materials was performed in conjunction with the
Pratt & Whitney ANP shield design effort. In 5.4. Tower Shielding Facility
this group of tests particular emphasis was placed
Gamma-ray measurements in various crew-shield
on a study of the production of secondary gamma
mockups at the Tower Shielding Facility have
rays in a gamma-ray shield of depleted uranium.
indicated an intensity which is greater than that
In most cases, slabs of nickel and stainless steel
expected from scattered gamma rays alone. This
were placed between the fission source and the
prompted a study of the production of secondary
uranium. Borated water was always used as the
gamma rays in air, both those resulting from
neutron shield behind the uranium.
thermal-neutron capture in nitrogen and those
Another series of tests with advanced shielding
resulting from inelastically scattered fast neutrons.
materials was performed at the request of General
For this study an experiment was performed in
Electric. These tests, again, were designed for
which a beam of neutrons from the reactor irradiated
a study of the production of secondary gamma
a cone of air, and the gamma rays resulting from
rays in gamma-ray shields. The materials used in
neutron interactions within the cone were observed
this series of tests were Hevimet (largely tungsten),
by a gamma-ray spectrometer several feet away.
stainless steel, and depleted uranium. Lithium
The ratio of the nitrogen capture gamma rays
hydride was also included in most of the con
observed by the detector to other gamma rays
figurations tested, and, since it is highly reactive
observed was varied by covering the reactor
with water, oil rather than water was used as
neutron beam with boron-impregnated Plexiglas for
the neutron shield.
some of the measurements. The difference between
5.3. Tower Shielding Reactor II the measurements was then used to separate the
The mechanical design of the Tower Shielding pulse-height spectrum associated with nitrogen
Reactor II (TSR-II) has been modified slightly capture gamma rays from the spectrum associated
to provide more clearance for the thermal ex with other sources. In this report the spectrum
pansion that will occur at the highest operating which remained after the subtraction of the capture-
powers. A second mockup of the control mechanism gamma-ray spectrum was presumed to be due to
has been constructed and is undergoing testing. inelastically scattered fast neutrons; however,
Additional nuclear calculations have been per contributions from other sources, such as air-
formed, from which it has been found that the scattered gamma rays, must be considered. The
control plates could be worth 6% in k. The results of the experiment are to be compared with
critical mass, which was found to be 8.2 kg of theoretical calculations of the pulse-height spectra
U235, will be checked with further calculations to be expected from neutron interactions with air,
and in a mockup experiment with dummy fuel as well as from scattered gamma rays.
 CONTENTS

PART 1. AIRCRAFT REACTOR ENGINEERING

1.1. REACTOR AND FACILITY CONSTRUCTION 1

Status of the ART-ETU Program at Time of Termination 1
Cancellation of Fabrication and Construction Activities 5
Summary of ART and ETU Operations Planning 5
Summary of ART Facility Construction °
1.2. COMPONENT DEVELOPMENT AND TESTING 24
OA
Pump Development Tests
Bearing, Seal, and Lubricant Tests ^4
Aluminum North Head Water Tests 24
Fuel Pump High-Temperature-Performance Tests 24
Fuel Pump Endurance Tests •"
Primary and Auxiliary NaK Pump Development 25
Sodium Pump Endurance Test •"
Reactor Component Development Tests *•'
Heat Exchanger and Radiator Development Tests 27
Outer Core Shell Thermal Stability Tests 27
Zirconium Fluoride Vapor Traps 28
28
In-Pi le Experimentation

1.3. INSTRUMENTATION AND CONTROLS DEVELOPMENT 30

ART Control Rod Drive Test 30
ART Fuel-Expansion-Tank Continuous Level Indicator 30
Resistance-Type Fuel Level Indicator 34
Liquid-Metal-Level Transducers 38
On-Off Level Probes 39
Magnetic Flowmeters
AO
Turbine Flowmeters
Pressure Transmitters

ART Thermocouples
Sheathed Thermocouples
Well Thermocouples

1.4. ENGINEERING DESIGN STUDIES 49

Applied Mechanics and Stress Analysis 49
50
1.5. DESIGN PHYSICS

1.6. MATERIALS AND COMPONENTS INSPECTION 52

52
Materials Inspection
52
Weld Inspection
1.7. HEAT TRANSFER STUDIES 54
Thermal-Cycling Research 54
Pulse-Pump System 54
Pressurized System 56
ART Hydrodynamics 56
Fused Salt Heat Transfer 56
Molten Lithium Heat Transfer 58
Volume-Heated Systems 59
ART-Type Core with Screens 59
Liquid-Metal Experiment 59
Heat Transfer with Vortex Flow 59
Tower Shielding Reactor II (TSR-II) Hydrodynamics 61
Physical Properties 61
Thermal Conductivity 61
Enthalpy and Heat Capacity 62

PART 2. CHEMISTRY

2.1. PHASE EQUILIBRIUM STUDIES 67

Summary of Salt System Investigations 67
Comparison of Russian and ORNL Investigations of the NaF-ThF4 System 67
The System NaF-CrF_ 69
Apparatus for High-Temperature Thermal-Gradient Quenching Experiments 70
Petrographic Analyses of Fuel Mixtures 70
An Optical Method for the Determination of Equilibrium Paths and Tie Lines in
Ternary Fused-Salt Systems 72
Systems Without Solid Solution 72
Systems With Solid Solution 72
Index of Refraction vs Solid Solution Composition 73
Composition as a Function of Temperature 73
Constructing Tie Lines and Equilibrium Paths 73
The System LiF-ThF4-UF4 74
2.2. CHEMICAL REACTIONS IN MOLTEN SALTS 75
The Preparation of UF3 and Its Stability in Molten Fluoride Mixtures 75
Activity Coefficients of FeF2 and of NiF2 in Molten NaF-ZrF4 76
Solubilities and Activity Coefficients of NiF2 as Functions of Solvent Composition
in NaF-ZrF4 Mixtures 77
Reduction of UF4 by Chromium in Molten Fluorides 78
Reaction of UF4 with Various Structural Metals in Molten Fluorides 82
Reduction of FeF2 by Chromium in Molten Fluorides 83
Thermodynamic Interpretation of Equilibrium Data on the Corrosion of Iron and of
Chromium by Molten-Salt Fuels 87
Solubilities of Noble Gases in Molten Fluoride Mixtures 92
 Solubility of HF in NaF-ZrF4 93
Solubilities of Fission-Product Fluorides in Molten NaF-ZrF4 Mixtures 95
CeF3 in NaF-ZrF4 96
YF3 in NaF-ZrF4 96
CeF3 plus SmF3 in NaF-ZrF4 97
Solubility of BaO in KF-LiF 97
Reaction of CeF3 and UF4 with CaO in KF-LiF 99
Determination of Chromium Diffusion Coefficients in Inconel-Molten Salt Systems 100
Diffusion Coefficients 10°
Measurements Based on Radioactive Tracers 100
Battel le Diffusion Experiments '02
Steady-State Chromium Migration in Thermal-Convection Loops 104
. Chromium Migration Mechanism 107
Steady-State Migration '07
Unsteady-State Migration 108
Estimating Equilibrium Concentrations for Steady-State Loop Operation 108
Mass-Transport Equations for Steady-State Migration 109
Salt and Wall Temperature Relationship HO
Activities in Alloys H3
Dissolution of Stainless Steel Fuel Elements in Fluoride Melts 115
Oxidation of Mixtures Composed of Sodium and Potassium 118
2.3. PHYSICAL PROPERTIES OF MOLTEN MATERIALS 120
Summary of ARE Fuel Vapor-Pressure Studies 120
Vapor Pressures of Molten Fluoride Mixtures 122
AMass Spectrometric Study of Salt Vapors 124
Surface Tensions of Molten Salts '27
2.4. PRODUCTION OF PURIFIED MIXTURES 129
Preparation of Fluoride Compounds '*"
Alkali-Metal Fluochromates '29
Other Double Fluorides I29
Simple Inorganic Fluorides
Chromous Fluoride
Summary of Experience in the Preparation of Fuel for the Aircraft Reactor Experiment 131
Purification of Fuel Solvent
Purification of Fuel Concentrate
Storage and Ultimate Transfer
Pilot-Scale Purification Operations '33
Production-Scale Operations
1 *XA
Service Operations

2.5. ANALYTICAL CHEMISTRY 135

Detection of Traces of NaK in Air
Determination of Oxygen in Fluoride Salts
 PART 3. METALLURGY

3.1. NICKEL-MOLYBDENUM ALLOY DEVELOPMENT STUDIES 141

Material Development 141
Status of Development 141
Properties of INOR-8 143
Properties of INOR-9 143
Phase Diagram Studies 143
Stress-Rupture Testing by New England Materials Laboratory 143
Status of Production Heats 143
Composite Tubing and Plate 145
Mechanical Properties Investigations 145
Corrosion Studies 146
Forced-Circulation Loop Tests of Hastelloys 147
Forced-Circulation Loop Tests of Experimental Nickel-Molybdenum Alloys 147
Thermal-Convection Loop Tests of Experimental Nickel-Molybdenum Alloys 149
3.2. MECHANICAL PROPERTIES STUDIES 151
Design Data 151
Inconel 151
Recrystallized Inconel 152
Beryllium 152
Extrapolation of Test Data 152
Thermal Stress Investigations 153
Relaxation 154
Fatigue 154
Biaxial Creep 154
Tests of Shielding Materials 155

3.3. WELDING AND BRAZING STUDIES 156

Production of Cast Rings of Brazing Alloy 156
Fabrication of Thermocouples for High-Temperature Service 156
Fabrication of ART Heat Exchangers 157
ART Fill-and-Drain Tank Fabrication 159
Niobium Welding 160
Carburization of Inconel Tubing 166
Welding and Brazing Procedures 166

3.4. CORROSION AND MASS TRANSFER STUDIES 169

General Corrosion Investigations 169
Brazing Alloys in Fuel Mixtures 169
Brazing Alloys in Lithium 169
Inconel Welded with Nitrogen Shield Gas 170
Cermet-Metal Joints in NaF-ZrF4-UF4 and in Sodium 170
Postoperative Examination of Prototype ART Control Rod 172
 Corrosion in Dynamic Systems 173
Lithium in Niobium Thermal-Convection Loops 173
Lithium in Niobium Forced-Circulation Loops 175
Developmental Studies of Inconel Systems 175
Inconel Exposed to Fused Fluorides 175
Inconel Exposed to Sodium and to NaK 176
3.5. MATERIALS FABRICATION RESEARCH 177
Niobium Fabrication >''
Recrystallization of Niobium 177
Fabrication of Niobium Loops 177
Consolidation of Niobium Pellets 177
Yttrium Metal Production 178
Shielding Materials Development 179
Tubular Control Rod Production 180
Ceramic Investigations 1°0
Development of High-Density BeO Bodies 180
Extrusion of Magnesia and Beryllia Insulators 180
Production of Metal Hydrides 180
3.6. METALLOGRAPHIC EXAMINATIONS OF ENGINEERING TEST COMPONENTS
AFTER SERVICE 182
ART Prototype Test Radiator No. 1 182
ART Prototype Test Radiator No. 2 187
3.7. NONDESTRUCTIVE TESTING 192
Eddy-Current Thickness Measurements 192
Remote X-Ray Viewing 192
Materials Inspection '93
Tubing !93
Pipe 194
Plate 194
Hastelloy B Tubing 194

PART 4. RADIATION DAMAGE

4.1. RADIATION DAMAGE 199

Examination of Irradiated Components and Materials 199
MTR In-Pile Loops 199
Moderator Materials 200
Creep and Stress-Rupture Tests on Inconel 207
MTR Experiments 207
LITR Experiments 208
ORR Experiments 208
LITR Vertical Fused-Salt Loop 208
MTR Static Corrosion Tests 209
Irradiation of Moderator Materials for Use at High Temperatures 209
 Irradiation Effects in Photoconductive Detectors 210

Irradiation Effects in Semiconductor Barriers 218

Semiconductor Materials 218
Semiconductor Devices 221

PART 5. SHIELDING

5.1. SHIELDING THEORY 229

Effect of Anisotropic Scattering on the Neutron Flux in Air 229

Bremsstrahlung Spectra in Sodium Iodide and in Air 231
A Calculation of Gamma-Ray Dose Rates from Neutron Capture in Air 233
Calculational Method 234
Results for Point Isotropic Sources 235
Results for Line-Beam Sources 236

5.2. LID TANK SHIELDING FACILITY 238

Pratt & Whitney Series of Advanced Shielding Materials Studies 238

General Electric Company Series of Advanced Shielding Materials Studies 261
5.3. TOWER SHIELDING REACTOR II 275
Mechanical Design 275
Control Mechanism 275
Nuclear Calculations 275
Water Cooling and Purification System 281
Extension of Gamma-Ray-Heating Calculation for the TSR-II 281
5.4. TOWER SHIELDING FACILITY 285
Study of Gamma Rays Produced by Neutron Interactions in Air 285
Experimental Configuration 285
Determination of Detector Response 286
Experimental Results and Discussion 287
Thermal-Neutron Flux Measurements 292

XXII
 Part 1

AIRCRAFT REACTOR ENGINEERING

 1.1. REACTOR AND FACILITY CONSTRUCTION
W. F. Boudreau A. P. Fraas
H. W. Savage

STATUS OF THE ART-ETU PROGRAM anticipated only 30 full thermal cycles; thus a
AT TIME OF TERMINATION factor-of-10 margin was provided to ensure a high
A. P. Fraas
degree of reliability. The results of the analytical
design studies agreed surprisingly well, in most
The Aircraft Reactor Test was designed as a instances, with the results of tests of components,
60-Mw circulating-fuel reactor having a major for example, heat exchangers. It was expected
portion of the features essential to an aircraft pro that the programmed tests would have confirmed the
pulsion reactor. The details of the design have been results of analytical studies of the designs of
reported, as they were developed, in previous other components.
reports in this series, and they were summarized Much of the extensive program of component
in a design report. Diagrams of the reactor and testing had been completed. The performance
the cell in which it was to operate are presented in testing of the pumps was finished, together with
Figs. 1.1.1, 1.1.2, and 1.1.3. At the time of much of the extensive endurance test program.
termination of the program, a full-scale, non- The fuel-to-NaK heat exchanger component test
nuclear engineering test unit (ETU) that essentially work was largely completed, except for some tests
duplicated the ART was being fabricated and on circular-arc tube bundles designed to simulate
assembled. The ETU was to demonstrate the in considerable detail the geometry of the full-
feasibility of fabrication and assembly, and, in scale reactor.
high-temperature operation, to test the effects Difficulties were experienced with leaks caused
of the complex thermal and pressure stresses by thermal stresses in the prototype NaK-to-air
expected during nuclear operation. The com radiators that were tested, but design modifications
plex geometry, the wide variety of stresses, were made which, it was believed, would give an
and the difficulty of predicting the magnitude and acceptable arrangement. Tests of the modified
direction of thermal warping and distortion had radiators were to have been started shortly after
indicated the need for a comprehensive test on a the program was cancelled. The reflector cooling
non-nuclear assembly. Upon completion of the circuit was completely worked out and checked in
planned tests, the ETU was to have been dis flow tests. The core hydrodynamics had posed
assembled and thoroughly examined before operation many problems, but tests were made in the fall of
of the ART. 1957 on a combination of screens which was
At the time of cancellation, the design, the shop found to be effective in suppressing local temper
drawings, and much of the component testing and ature fluctuations. The control-rod drive mechanism
fabrication had been completed. All outside had been thoroughly tested and found to be satis
contract work on the facility in which the ART factory in extensive endurance tests. The principal
was to operate had been completed, and work was area in which component test work remained to be
to have started on the installation of the process done was in structural tests of the thin shells.
piping. This work had been delayed by difficulties in
The analysis of the design of the ART was com fabricating these shells, but analytical and pre
pleted except for a number of thermal stress analy liminary test work on models indicated that the
sis problems intimately related to test work which shell designs were satisfactory.
was in progress. Sufficient analytical and test Much of the fabrication of the first set of parts
work had been completed to indicate that the was completed. The pumps were available; the
design was adequate to carry the basic pressure basic parts of the north head were finished, in
stresses and a modest number of thermal cycles. cluding the sodium-to-NaK heat exchangers, and
The design objective was a nominal life of 300 the assembly work was 70% completed. The
thermal cycles, whereas the operating program island beryllium and core shell assembly was
completed, and the beryllium for the reflector was
A. P. Fraas and A. W. Savolfcinen, Design Report on
the Aircraft Reactor Test, ORNL-2095 (Dec. 7, 1956). machined and ready for assembly, as were the
ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG (6011

FUEL PUMP

Na EXPANSION TANK

FUEL EXPANSION TANK

Na-TO-NaK
HEAT EXCHANGER

HEAT EXCHANGER
ASSEMBLY

REFLECTOR
ASSEMBLY

TILE LAYER-

FUEL-TO-NaK
HEAT EXCHANGER

SPHERICAL B4C
TILE LAYER —

THERMOCOUPLE SLEEVE

FUEL DRAIN

Fig. 1.1.1. Vertical Section TKrough Reactor Assembly.

 PERIOD ENDING DECEMBER 31, J957

CONTROL ROD

-ALUMINUM TANK
TYPICAL LiOH CAN
- 18-in-THICK BORATED WATER SHIELD

- FUEL PUMP DRIVE HYDRAULIC MOTORS

HEXAGONAL CANS FILLED WITH LiOH

REACTOR SUPPORT PADS -

ALUMINUM TANK-
-10-in-THICK BORATED
WATER SHIELD
SUPPORT FOR WATER SHI ELD -

HANGER RODS - STEEL SUPPORT STRUCTURE FOR

SUPPORTING LEAD SHIELD- THE REACTOR ASSEMBLY AND THE
LEAD AND WATER SHIELDS

4.3-in.-THICK LEAD
GAMMA-RAY SHIELD -
- ALUMINUM TANK

33-in-THICK BORATED
WATER SHIELD

- COOLING WATER PASSAGES,

V2-in.-THICK
3/„-in.-0D TUBES ON
THERMAL INSULATION -
4-in. CENTERS

HEXAGONAL CANS FILLED WITH LiOH 3-in-THICK LEAD GAMMA-RAY SHIELD

'/2-in-THICK THERMAL INSULATION

FUEL DRAIN LINE

Fig. 1.1.2. Vertical Section Through Lead-Water Shield.

ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG (5056
jpSjfe,

Fig. 1.1.3. Reactor Assembly Cell.

 PERIOD ENDING DECEMBER 31, 7957

boron-bearing parts for the boron curtain. A proto Termination reports have been prepared which
type main fuel-to-NaK heat exchanger tube bundle summarize the reactor and facility construction
had been completed and was found to meet specifi activities. The first of these, which describes
cations. The remaining tube bundles were ex the construction and procurement work for the ETU
pected at the rate of two per month. The forging and ART reactors and the ETU facility, has been
for the pressure shell was being machined. essentially completed except for material which is
The ART design presented some difficult prob to be prepared by the vendors of heat exchangers,
lems in the fabrication of thin-walled Inconel radiators, and other equipment. In addition to
shells, particularly in connection with the core descriptions of the various problems encountered
shells and the thin, hemispherical shells for jack in construction, assembly, and procurement, and
eting the boron. Various techniques, including the solutions that were found for them, this report
hydrospinning, pressing, and machining from forgings will contain a brief analysis and summary of
and pressings, were investigated. Hydrospinning expenditures for the various phases of the work.
was found to be unsatisfactory, but the core shells A summary report on the planning for removal
were machined readily to the desired tolerances from and disassembly of the ART has also been pre
pressings and were ready for assembly. It was pared.2 This document contains a great deal of
found that the large-diameter thin-walled hemi information that can be expected to be of value to
spherical shells could not be fabricated by hydro- other projects, in particular, sections concerning
spinning, but similar parts were made (with dies methods for making accurate dimensional measure
built for a missile application) which showed that ments on highly radioactive parts and methods for
these parts could be fabricated by hydraulic dissecting a relatively large and complex object
pressing. New dies for the ART parts were sched by remote means.
uled for completion in November 1957.
Construction work on the facility for the ETU was SUMMARY OF ART AND ETU
under way in the experimental engineering building. OPERATIONS PLANNING
The ART facility in which the second reactor
D. B. Trauger
assembly was to be tested was, as stated above,
completed and ready for the installation of process W. B. Cottrell S. M. DeCamp
piping. A summary description of the construction C. W. Cunningham R. E. MacPherson
of the facility is given in a subsequent section. The ETU-ART Operations Group was established
Plans had been completed for training operating in the spring of 1956 to review the reactor flow
crews, and operating procedures were being pre sheets and general design from an operational
pared. These items are also described in a sub standpoint, to prepare preliminary procedures for
sequent section. reactor operation, and, eventually, to train per
sonnel and to operate the reactors. To augment
the few people who could be made available for
CANCELLATION OF FABRICATION AND this work and to utilize experience and information
CONSTRUCTION ACTIVITIES
from the various groups within the project, a
W. F. Boudreau committee was formed to assist in the preparation
Arrangements were made, immediately after of procedures. The Operating Committee studied
notification was received to cancel work on the ART and ETU operational problems and conditions,
ETU and the ART, to halt construction and fabri and preliminary drafts of ART operating procedures
cation activities in the Oak Ridge area and at the were prepared. The initial procedures constituted,
various vendors' plants. All work was terminated essentially, a basic philosophy of operation. With
in a manner that preserved tooling, materials, and the pattern thus established, the activities of the
partially completed components in a condition Committee were completed, and the Operations
suitable either for the future resumption of the Group went forward with the task of preparing de
work or for adaptation to other uses. The ETU tailed procedures.
and the ART facilities were put into standby con
dition pending a determination of their usability 2A. A. Abbatiello and F. R. McQuil)<in, ART Removal
for other purposes. and Disassembly, ORNL-2464 (in preparation).
 ANP PROJECT PROGRESS REPORT

The reactor flow sheets and control circuitry uled reactor operation, separate procedures were
were used as the bases for the detailed operating to be prepared for servicing and maintaining in
procedures. At the time of cancellation of the pro struments and other equipment and for complex
gram, most of the flow sheets were complete and maintenance, such as the replacement of a pump or
up to date. Systematic tabulations of various a radiator.
components, such as thermocouples and valves, Procedures for an operation such as the ART
were made to clarify the requirements and to check could not be developed in such detail that an
the flow sheet nomenclature. An instrument tabu operator unacquainted with the device could be
lation was prepared which specified the service and expected to perform a safe and satisfactory oper
requirements that every instrument was to fulfill. ation. Therefore, a program of operator training
Preparation of procedures served as a final was planned to provide the necessary background.
check on the over-all design. The procedures
prepared first included reactor disassembly, SUMMARY OF ART FACILITY CONSTRUCTION
maintenance, operation of certain auxiliary equip F. R. McQuilkin
ment, pump startup, and filling and draining of the
reactor. The studies of procedures led to numerous Lump sum contract work on the ART facility
design changes; for example, the review of dis was completed. This work was contracted in four
assembly requirements alone revealed 45 necessary stages, as determined by the design program.
or desirable changes, while the fill-and-drain The first stage, package 1, included major building
system studies resulted in the elimination of un alterations, additions to existing buildings, cell
necessary equipment. Particular attention was installation, installation of main air duct, instal
given to radiation levels and other conditions to lation of a 1500-kva substation, and installation
be expected where maintenance was likely to be re of 480-v main switch-gear. The second stage,
quired during operation. At the time of termination package A, included installation of auxiliary
of the project, consideration had been given to service piping. The third stage, package 2, in
procedures for most of the normal operating situ cluded installation of diesel-generators and
ations of both the ETU and ART.3-6 In addition facility, electrical motor control centers, spec
to the procedures pertaining specifically to sched- trometer room electrical system, and the spectrom
eter air-conditioning system. The fourth stage,
package 3A, included installation of an electrical
3ETU Operations: Status Report, ORNL CF-57-10-7 system for supplying power to the pipe and equip
(in press). ment heaters, a dry-air plant and facility, NaK
4ART Operations: Status Report, ORNL CF-57-10-9 pump motor controllers, a lube-oil fill-and-waste
(in press).
system, and a hydraulic system for louver operation.
5ETU Operating Procedures, ORNL CF-57-10-8 (in Tables 1.1.1, 1.1.2, and 1.1.3 present pertinent
press).
6ART Operating Procedures, ORNL CF-57-10-10 (in information concerning all ART lump sum contract
press). work performed at the 7503 building.

Table 1.1.1. Contract Cost Summary

Total Lump Sum Contract Cost $1,096,390.42

Contract AT-(40-l)-1952 881,852.25

Package 1, phase I $607,526.68

Package 1, phase II 274,325.67

Package A 50,351.62

Contract AT-(40-l)-1954 214,538.17

Package 2 71,557.67
Package 3A 142,980.50
 PERIOD ENDING DECEMBER 31, 7957

Table 1.1.2. Contract AT-(40-l)-1952 Summary

Total Contract Cost $881,852.25

Phase I — package 1 $557,175.06

Phase II — package 1 274,325.57

Package A 50,351.62

Package 1

Initial total contract cost 765,835.00

Phase I, construction of 501,462.00

additions to Building 7503

Phase II, construction of 264,373.00

7503 cell (operational)

Final total contract cost 831,500.63

Phase I 557,175.06

Phase II 274,325.57

Total number of contract 45

deviations

Official starting date 8-23-55

Official scheduled completion 8-1-56

date

Actual completion date 10-9-56

Contract liquidated damages rate

Phase I $100/day

Phase II $300/day

Liquidated damages charged None

Reason liquidated damages Contract deviations plus equipment

charge not invoked delivery delays resulting from
Westinghouse strike

Prime contractor V. L. Nicholson Co.

Major subcontractors

Electrical Davis-Longsworth Electrical Corp.

Mechanical John F. Humphrey Co.

Cell installation Chicago Bridge & Iron Co.

Designed by K-25 Plant Engineering Dept.

Approximate number of drawings 118

ANP PROJECT PROGRESS REPORT

Table 1.1.2 (continued)

Package A

Initial contract price $50,351.62

Final contract price $ 50.351.62

Official starting date 4-10-56

Official scheduled completion 6-29-56

date

Actual completion date 9-26-56

Prime contractor V. L. Nicholson Co.

Major subcontractors

Electrical Davis-Longsworth Electrical Corp.

Mechanical John F. Humphrey Co.

Designed by ORNL Engineering Department

Approximate number of drawings 18

Table 1.1.3. Contract AT-(40-l)-1954 Summary

Total Contract Cost $214,538.17

Package 2 $ 72,557.67

Package 3A 142,980.50

Package 2

Initial contract price 58,400.00

Final contract price 71,557.67

Total number of contract deviations 8

Official starting date 1-25-56

Official scheduled completion dates

Part 1 6-22-56

Part 2 1-27-57

Actual comp et on dates

Part 1 6-22-56

Part 2 2-22-56

Liquidated damage rate $100/day

Liquidated damage charged None

 PERIOD ENDING DECEMBER 31, 7957

Table 1.1.3 (continued)

Reasons liquidated damages not Government-furnished diesel-generators

invoked were damaged in shipment and caused
extensive equipment delivery delay.
Contract extension was granted for ex
cessive amount of time beyond the con
tractor's control required for perform
ance of the specified diesel tests.

Contractor Rentenbach Engineering Co.

Major subcontractors

Electrical Broadway Electrical Service Co.

Mechanical Hicks-Ingle Co.

Designed by ORNL Engineering Department

Approximate number of drawings 15

Package 3A

Initial contract price $114,434.00

Final contract price $142,980.50

Capital expenditures 118,394.71

Operational expenditures 24,585.79

Total number of deviations 6

Official starting date 12-7-56

Official scheduled completion dates

Initial 4-25-57

Final 8-21-57

Actual completion date 8-20-57

Liquidated damage rate $100/day

Liquidated damage charged None

Contractor Rentenbach Engineering Co.

Major subcontractors

Electrical Broadway Electrical Service Co.

Mechanical Hicks-Ingle Co.

Designed by ORNL Engineering Department

Approximate number of drawings 34

ANP PROJECT PROGRESS REPORT

A pictorial resume of the facility construction contractor. This was photographed in July 1956.
activity is presented with the following series of Figure 1.1.19, which was photographed in April
pictures which were taken at various stages of 1957, shows portions of the electrical heater dis
the construction program: Figure 1.1.4 shows a tribution panels which were installed by the package
portion of building 7503 as it existed at the start 3A contractor. Figure 1.1.20, which was also
of construction of the ART facility. When this photographed in April 1957, shows the dry air
picture was taken in September, some remodeling station which was installed by the package 3A
had been started. Figure 1.1.5 is an early view contractor. Figure 1.1.21 was also photographed
looking into the south side of the building. In the in April 1957 and shows the regulator pit in the
foreground can be seen the concrete reinforcing north basement and some of the induction voltage
steel for the absorber pit and the cell foundation. regulators installed by the package 3A contractor.
In the background can be seen the forms required for Figure 1.1.22 shows the NaK pump motor controllers
the control-tunnel wall. Figure 1.1.6, which was which were installed by the package 3A contractor.
taken in January 1956, shows the structural steel Figure 1.1.23 was photographed in August 1957.
required for additions to the building. On the It shows a portion of the louver hydraulic operating
left can be seen the switchhouse. The foreground system. Figure 1.1.24, which was photographed
shows the stack foundation and the top of the after construction activity ceased in August 1957,
absorber pit. Figure 1.1.7, which was made in is a view looking south toward the building. In
January 1956, shows the early stages of cell and the foreground can be seen the diesel-fuel storage
water vessel fabrication. Figure 1.1.8, also made tank which was installed by the package 1 con
in January 1956, shows the cell foundation and the tractor. The transformer in the center of this
first portion of the water tank that was installed. view was also installed under the package 1 con
Figure 1.1.9 shows the status of the cell instal tract. On the right can be seen the "Braden" type
lation during the early part of March 1956. Figure building which houses the five diesel-generator
1.1.10 is a close-up view of the water vessel and units. It was constructed by the package 2 con
cell which was taken in the latter part of March tractor. On the left can be seen the process water
1956. In the foreground can be seen the cell and a tank which was part of the existing ARE facility.
portion of the water vessel. In the background is Figure 1.1.25, which also was photographed in
the top section of the water vessel. Figure 1.1.11 August 1957 after construction had ceased, is a
was photographed during vacuum leak-testing of the view looking toward the southwest corner of the
cell in June 1956. It shows the cell water vessel building. In the left background can be seen the
in place as it would have been during ART oper diesel-generator house and in the left foreground is
ation. Figure 1.1.12 is a view looking into the the compressor house. In the center is the blower
main air duct. It was photographed in June 1956. house. In the right foreground is the vent house,
Figure 1.1.13, which was also taken in June 1956, stack, and top of the absorber pit. The generator
shows the 480-v switchgear. Figure 1.1.14, which house was installed by the package 2 contractor;
was taken in October 1956, shows the cell after the compressor house was installed by the package
the head had been removed and placed in a storage 3A contractor; and the blower house, vent house,
area. This view represents the status of the cell absorber pit, and stack were installed by the
after contract work was completed. Figure 1.1.15 package 1 contractor.
was photographed in May 1956. It shows a portion ORNL forces had painted the spectrometer tubes,
of package A service piping in the auxiliary equip installed tracks for the collimator dolly and the
ment room. Figure 1.1.16 was photographed in neutron spectrometer dolly, fabricated one set of
April 1956 and is a view looking at the west side heat barrier doors, vacuum leak-tested the lube-oil
of the building. In the right foreground can be seen waste system, installed the unit heater in the
the foundation for the diesel-generator building, switchhouse, and painted miscellaneous piping
which was installed under the package 2 contract. and structural steel.
Figure 1.1.17 is a view of one of the two sections A termination report on the construction of the
of the motor control center installed by the package ART facility has been prepared.
2 contractor. This was photographed in June 1956.
Figure 1.1.18 is a view showing the five diesel
F. R. McQuilkin et al,, Termination Report for Con
generators which were installed by the package 2 struction of the ART Facility, ORNL-2465 (in press).

10
ANP PROJECT PROGRESS REPORT

1.2. COMPONENT DEVELOPMENT AND TESTING

H.W. Savage

The experimental engineering activities as operation were completed. Detailed reports

sociated with the development and testing of covering these tests and previous bearing and
components for the ART and ETU systems are seal developments are being prepared.
being terminated gradually with the objective of As mentioned in the previous report, the re
obtaining the maximum information that'will be actor fuel pump rotary element that was being
of value in further aircraft reactor developments. operated in a new gamma-irradiation facility at
Equipment and test stands for which no more the MTR was shut down for repair of a defective
essential tests were planned or from which the ion chamber. The chamber was found to be in
data to be obtained were applicable only to the operative because of the deterioration of the
ART or ETU have been disassembled. Parts that Teflon used to insulate the high-voltage element.
could be used elsewhere were salvaged and The system has now been modified so that dose
stored. All outstanding orders for equipment in rates can be determined from flux readings taken
this category were cancelled. external to the test capsule. A subsequent attempt
Test components that had been completed in to resume operation of the rotary element was
the shop, or almost completed, from which in halted by ruptures of two pressure-measuring
formation of general interest can be obtained devices. Inspection has revealed that the hy
are being tested as test facilities become avail draulic drive system has failed somewhere inside
able. Similarly, tests that were under way at the sealed unit, but it is not yet known whether
the time of cancellation of the ART-ETU program the hydraulic line is plugged or the motor is
will be completed if they provide data of value defective. The unit will be examined and repaired.
to other ANP programs. Summary reports on the Another attempt will be made to resume the ex
major test programs are being prepared as the periment when the repairs have been made.
work is completed.
Aluminum North Head Water Tests
A forced-circulation niobium-lithium loop is
being designed in which to evaluate system J. W. Cooke2
components and metallurgical developments. Con Terminal reports * are being prepared that
struction of the loop will be initiated when the will give details of the water tests of the ART
niobium becomes available. Thermal-convection fuel and sodium systems. These tests were
loops will also be operated. A development performed with the use of an aluminum mockup
program will be carried out on valves, bearings, of the ART north head.
pump seals, and the other components that will
be required in an operating system. Secondary
Fuel Pump High-Temperature-Performance Tests
programs will include the testing of equipment
for the continuous processing of lithium to P. G. Smith W. E. Thomas
achieve and maintain the required purity in an The loop in which high-temperature performance
operating system. tests are being run of an ART-type fuel pump
with NaF-ZrF4-UF4 (50-46-4 mole %, fuel 30) as
PUMP DEVELOPMENT TESTS the circulated fluid operated throughout the
W. F. Boudreau A. G. Grindell quarter and has accumulated a total of 3350 hr
of continuous operation. The loop is presently
Bearing, Seal, and Lubricant Tests
D. L. Gray
The tests described previously that were being On assignment from Pratt & Whitney Aircraft.
3
conducted in order to establish a detailed pump J. W. Cooke, Terminal Report of the Aluminum North
sparging procedure for use during ETU and ART Head Twin MF Pump Water Test Stand, ORNL CF-57-12-4
(in press).
J. W. Cooke, Terminal Report of the Aluminum North
1D. L. Gray and W. K. Stair, ANP Quar. Prog. Rep. Head Twin MN (Moderator Coolant) Pump Water Test
Sept. 30, 1957, ORNL-2387, p 32. Stand, ORNL CF-57-12-5 (in press).

24
 PERIOD ENDING DECEMBER 31, 7957

running with the pump suction pressure reduced the same time, a test was started on a split-purge
sufficiently to induce cavitation so that the effect arrangement for the shaft annulus in an attempt
on the impeller can be determined. The pump to obtain long-time trouble-free operation of the
has operated at 1200°F for 1200 hr at the cavi lower seal. The helium purge supply is brought
tation conditions (2.5 psig pump surge pressure, directly to the pump shaft annulus and a portion
645 gpm, and 2700 rpm), and the test is scheduled of the gas flows down and through the pump tank
for termination at the end of 3000 hr under these and into the off-gas system. The remainder of
conditions. the gas flows upward into the catch basin to
An identical fuel pump was placed in operation perform the function of sparging the lower seal
with NaK (56-44%) in order to obtain information oil leakage. A scheduled shutdown to remove
on the effect of the type of fluid being pumped. the flowmeters halted the split-purge test after
Preliminary data indicate that the head is 2 to 953 hr; the rotary element is to be moved to the
4 ft higher than in water and fuel tests. A total auxiliary NaK pump test stand No. 1 and the
of 350 hr of continuous operation of this pump test will be resumed. Approximately 8700 hr
has been accumulated. Investigations are under of hot testing with NaK have been accumulated
way to determine the cause of the difference in in the two primary NaK pump test stands.
head, and, in further tests, cavitation effects On August 16, 1957, an auxiliary NaK pump
will be studied. with a Fulton-Sylphon bellows type of lower
seal was installed in auxiliary NaK pump loop
Fuel Pump Endurance Tests No. 2 for testing. When the test was stopped
P. G. Smith for a scheduled shutdown on November 5, approxi
Operation of the fuel pump endurance test was mately 936 continuous hours of operation had been
resumed, and 860 hr of additional operation was accumulated. The seal had operated successfully
accumulated to bring the total operating time to with an average leakage of 16 cmvday. Testing
of this pump will not be resumed.
4410 hr. The test was terminated when leakage
of oil from the lower seal exceeded prescribed Performance tests were completed on the
limits. Examination of the pump indicated that auxiliary NaK pump operating in loop No. 1.
replacement of the lower seal would have per The pump in this loop had operated approximately
mitted further operation; however, the tests of 2048 hr with an estimated 200 hr at 1500°F. The
this pump are not to be resumed. A report on the pump has been removed from this test stand and
details of the test operation of the pump has is being replaced with the pump from the primary
been issued.
NaK pump loop No. 2, except that the original
auxiliary NaK pump impeller will continue to be
Primary and Auxiliary NaK Pump Development used. A total of 5100 hr of operation at elevated
D. L. Gray temperatures with NaK has been accumulated on
the two auxiliary NaK pump test stands.
Operation of primary NaK pump loop No. 1
at elevated temperatures was resumed in order
Sodium Pump Endurance Tests
to observe the effects on the lower seal area
of operating without a purge flow down the shaft W. F. Boudreau P. G. Smith
annulus. Shortly after startup, the NaK in the The Inconel thermal barrier that was designed
loop had to be drained so that a leaking drain to protect the bearing-housing region of the ART
valve could be replaced. The loop was then sodium pumps from thermal and radiation damage
restarted and operated for approximately 1172 hr was subjected to tests in the sodium pump en
prior to a scheduled shutdown on November 7, durance test stand. The test was performed with
1957. Operation of this loop will be resumed sodium at 1250°F and with the pump rotating
for testing preheating procedures and for tests at 3500 rpm. For this test, the expansion tank
at temperatures above 1500°F. cover and the instrumented section of the pump
Data on electromagnetic flowmeters were ob barrel were insulated.
tained in primary NaK pump loop No. 2, and at A typical set of temperature data obtained
P. G. Smith, MF Pump Endurance Test - 860 Hours,
from the pump barrel and from the barrel and
ORNL CF-57-12-3 (Dec. 31. 1957). barrier arrangement is presented in Fig. 1.2.1.

25
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR- DWG 27000

INSULATION (3/4-in. THERMOGLASS THERMOCOUPLE NO. TEMPERATURE(°F)

BACKED WITH ALUMINUM FOIL)
n 1252
a. 14 770
5
15 938
3

16 430
17 240
56 275
57 295
PUMP BARREL (INCONEL)
58 317
59 365
THERMAL BARRIER- 60 450
61 475
62 525
63 560
64 640
65 652
66 672
67 708
68 722
69 750
70 765
71 812
72 818

2 in. OF SUPEREX INSULATION

EXPANSION TANK COVER (INCON EL)

6.190 in.

SODIUM LEVEL-

X TC 11

NOTE :
NOT DRAWN TO SCALE

Fig. 1.2.1. Sodium Pump Thermal Barrier Test.

26
 PERIOD ENDING DECEMBER 3 7, 7957

The temperature data are plotted in Fig. 1.2.2 The radiator development program was concluded
to give the temperature profiles along the outside with the termination of ART test radiator No. 2
and the inside of the pump barrel. because of failure of a radiator tube after a
The test system was operated for 400 hr, and total of 181 thermal cycles. The number of cycles
the test was terminated because of an apparent imposed on this radiator was six times the
high leakage rate from the lower seal. The test maximum number of thermal cycles proposed for
was not resumed because of termination of the the ART radiators.
ART program.
Details for testing of this radiator and ART
This test stand was equipped with a lubricating
test radiator No. 1 are presented in a spearate
and cooling oil system which simulated that to
report. Details of the development testing of
be used in the ETU. The results of tests on this
smaller 500-kw units of the same basic design
system are described in a separate report. This
have also been summarized in a separate report.
system was tested for a total of 3800 hr of which
2600 hr were without interruption.

UNCLASSIFIED
Outer Core Shell Thermal Stability Tests
ORNL-LR-DWG 27001

W. B. McDonald E. R. Storto
J. C. Amos R. L. Senn

The second quarter-scale outer core shell model,

which was subjected to 339 thermal cycles, has
been measured, and the measurements are being
analyzed to determine whether dimensional changes
occurred during testing. The test shell is currently
being readied for an extended creep buckling test
at 1500°F with an external helium gas pressure of
S o
52 psi.
5 0 200 100 600 600 1000
TEMPERATURE (°F)

A third quarter-scale core shell is being fab

Fig. 1.2.2. Sodium Pump Barrel Temperature Profile. ricated by welding together oversize sections
of a complete shell and bringing the shell to
proper dimensions by final machining. This shell
REACTOR COMPONENT DEVELOPMENT TESTS
will be subjected to thermal-cycling tests similar
Heat Exchanger and Radiator Development Tests to those described previously in order to obtain
R. E. MacPherson J. C. Amos data on the structural reliability of the welded
R. L. Senn D. R. Ward shell construction. This method of shell con
struction was one of those proposed as a means
The heat exchanger and radiator development
of meeting the rigid dimensional tolerances re
testing program is currently being concluded quired for the ART core shells.
with a thermal cycling test of a 20-tube semi
circular heat exchanger designed to simulate Fabrication of a shell-buckling rig for testing
the ART main heat exchanger tube geometries. full-scale shells is 95% complete. This apparatus
The heat exchanger has completed 155 cycles will be shipped to Pratt & Whitney Aircraft, where
of the proposed 300-cycle program. A report is tests will be made on ART core shells.
being prepared that will summarize the results
of the over-all ANP heat exchanger development
program.
o

J. C. Amos and D. R. Ward, Performance and En

P. G. Smith, Simulated ETU Reactor Pump Lube and durance Tests of ART NaK-to-Air Test Radiators Nos. 1
Cooling Oil System Test, ORNL CF-57-10-60 (Oct. 10, and 2, ORNL CF-58-1-2 (in press).
1957). 9R. E. MacPherson and R. J. Gray, ART Prototype
J. C. Amos et al., ANP Heat i.xcbanger Development, Radiators — Development Program and Metallurgical
1955 Through 1957, ORNL-2439 (in preparation). Results, ORNL-2436 (in preparation).

27
 PERIOD ENDING DECEMBER 31, 7957

materials of construction and fuel compositions ART reactor design was being considered at
proposed for the ART. The three loops were the time of the termination of the ART program.
operated in the MTR with varying degrees of This work will be described in a forthcoming
success. The results indicated no measurable report.13
effect of radiation on the corrosion resistance
of the structural material or on the stability of 13P
D. B. Trauger et al., Circulating Fluoride Fuel
the fuel. Some work in support of an advanced In-Pile Loop, ORNL-2444 (in preparation).

29
ANP PROJECT PROGRESS REPORT

1.3. INSTRUMENTATION AND CONTROLS DEVELOPMENT

E.R.Mann C.S.Walker R. G. Affel

ART CONTROL ROD DRIVE TEST of sodium occurred on the cool portions of the drive
C. S. Walker machinery.

The test of the ART control rod drive system, The only troubles encountered during the test
as stated previously,1 was terminated after 3000 hr stemmed from equipment external to the rod drive
of successful operation with the control rod thimble itself; namely, difficulty with regulation of the
at or above the design temperature of 1100°F. water flow through the sodium-to-water heat ex
The drive mechanism operated through 1780 full changer, and faults in the temperature recording
stroke cycles. apparatus. The rod drive mechanism successfully
Before the termination of the test on September 17, completed the test.
1957, electrical leakage resistance measurements Samples of the sodium were taken for analysis
were taken with the system temperature maintained and the control rod was disassembled for ex
as in the test in order to determine whether sodium amination. Details of the results of the sodium
deposition had reduced insulation resistances. analysis and rod examination are presented in
The leakage resistance to ground and the leakage Chap. 3.4 of this report.
resistances between isolated conductors which
penetrated the tank that enclosed the rod drive
mechanism were measured. All these leakage
resistances were above 4 x 10' ohms and were ART FUEL-EXPANSION-TANK CONTINUOUS

acceptable in each case. The lowest resistance LEVEL INDICATOR

measured between the open contacts of the limit R. F. Hyland

switches exceeded 10' ohms.
The sodium was drained from the control rod In the early stages of ART design it became
thimble, and the drive mechanism was completely apparent that a method would have to be developed
dismantled and carefully examined. There were for continuous measurement of the level in the re
no indications of wear; the backlash in the gears actor fuel expansion tank. The level indicator was
had not increased beyond that in the original required to function continuously in a fuel
assembly. No sodium deposition could be found 30—helium environment in an Inconel container in
on the rod drive parts that were located above the the temperature range of 1000 to 1500°F and a gas
surface of the molten sodium. The speed of pressure of 15 psig. The range of level indication
operation of the rod had not changed during the was 0 to 5 in. The indication was to be accurate
test, and thus there was apparently no binding or to ±2% (percentage of maximum level) and
increase in friction between moving parts. reproducible to ±1%. The minimum life of the
The heat exchanger which surrounded the thimble device was to be 1000 hr of continuous operation.
at the upper surface of the sodium was found to
Any portion of the device that was above the
have a deposit of scale on all areas that were
fuel level in the expansion tank was to be cooled
in contact with water. This deposit was 0.018 in.
because there would be severe beta heating during
thick, and chemical analysis indicated it was
power operation of the reactor. The device was to
practically all calcium, with small quantities
be operable under the extremely turbulent fuel
of magnesium and zinc. A porous ceramic water
filter in the effluent water line was found to
flow conditions that would exist during power
operation. Turbulence of the type to be expected
contain a sizable quantity of rust.
A chemical analysis of the oil in the gear box
is illustrated in Fig. 1.3.1, which presents two
pictures taken during water tests of the aluminum
showed no significant amount of sodium or rare
north head mockup. Penetrations through the top
earths. This further indicates that no deposition
of the north head were to be kept to a minimum
because of the complexity of the system in that
E. R. Mann and C. S. Walker, ANP Quar. Prog. Rep. area and the number of layers that would have to be
Sept. 30. 1957. ORNL-2387, p 54. penetrated to reach the fuel expansion tank, as

30
ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG 27019

SODIUM FILLING PORT -j /— CONTROL ROD SLEEVE

rVENT LINE,
HELIUM INLET FOR PRESSURIZING —x J I r OFF GAS
No PUMP -x \ fl I Nil rf r COOLANT FUEL PUMP

SODIUM EXPANSION TANK FUEL EXPANSION TANK TOP

LIQUID LEVEL GAS INLET

OUTER
RING

Na PUMP

Fig. 1.3.2. ART North Head Showing Fuel Expansion Tank Region.

Since the level-measuring system depends upon Considerable work remained to be done in de
the sensing of the fuel pressure, the level termining the dynamic characteristics of the
indication is dependent on the fluid density. The measuring system under ART conditions. The
magnitude of the density effect is shown in planned test program included installation and
Fig. 1.3.5. If the reactor operating temperature testing in water in the aluminum north head test
of 1150°F is taken as the point of zero error, a rig of shielded dip tubes of the type to be used in
level error range of +2.5 to -5.7% exists in the the ART and similar testing, in fuel in the ETU.
temperature range 1000 to 1500°F. This error Details of the tests described here are available in
could be eliminated by installing a separate a separate report.2
temperature- (and thus density-) measuring system.
The signal from this system would be multiplied
by the signal from the level-measuring system to R. F. Hyland, ART Fuel Expansion Tank Continuous
produce a signal corresponding to the true level. Level Indicator, ORNL CF-58-2-1 (in press).

32
0(SELURiVnOD. POen
PERIOD ENDING DECEMBER 31, 7957

ORNL-LR-DWG 27020

TO ALL WELDED DIP CELL

REACTOR PRESSURE SHELL

FUEL EXPANSION TANK ROOF

FUEL LEVEL

SIDE OF FUEL EXPANSION TANK

LEVEL SENSOR SHIELD

FUEL EXPANSION TANK FLOOR

- HEAT EXCHANGER

Fig. 1.3.3. Dip Tube Installation in Reactor North Head.

ORNL-LR-DWG 27021
0.25

d 0.20

- LEVEL INCJICATOR S1
- LEVEL INC)ICAT0R S2

"*"""-••-**•«• ff=r=^==r^s ^*5:^

'^-j^^
<
w 0.05 1 £
SING LEVEL

*^
0
2 3 4 5 4 3 2
REFERENCE LEVEL (in. ON CALIBRATING PROBE)

Fig. 1.3.4. Results of Level-Measuring Tests in Nonturbulently Flowing Fuel.

33
ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG 27023
ORNL -LR-DWG 27022
4

^
Z 2
o
z
<

5 o
UJ
>

a
UJ

§-4
o
z

-6
1200 1300
FUEL TEMPERATURE CF)

Fig. 1.3.5. Effect of Fuel Density, as Indicated by

Fuel Temperature, on Accuracy of Purged-Dip-Tube
Fig. 1.3.6. Electrical Diagram of Resistance-Type
Level Measurements.
Fuel Level Probe.

RESISTANCE-TYPE FUEL LEVEL INDICATOR As the fuel level rises, the current flow through
T. V. Blalock R. F. Hyland the fuel between the Inconel rod and the 1 k-in.
pipe will become less inhibited because of the
A resistance-type fuel level indicator was increase in the flow-path cross-sectional area.
designed as a possible alternate to the purged The resistance of the rod and pipe to this current
dip tube described in the previous section. The flow will be negligible compared with the fuel
operation of this level element depends on a resistance. At zero level the element output will
resistance change that is a function of fuel level. be given by the product of the supply current and
A schematic electrical diagram of the system is the Inconel tube resistance, since the resistance
shown in Fig. 1.3.6*
Ri will be infinite for nonimmersion. The maximum
In Fig. 1.3.6, R is the current adjusting po output will occur at zero level, and the output will
tentiometer, RL is the lead resistance, Rc is a decrease as the level increases. The resistance
temperature-compensation resistance, Rf is the Rc is a function of temperature only, and R, and
fuel resistance between the Inconel rod and the
r, are functions of both temperature and level.
1!£-in. Incpne^pipe (Fig. 1.3.7), Rt is the Inconel
tube resistance, r, is a resistance representing The slope of the resistivity curve of the fused
the shorting-out effect of the fluid on the Inconel salt mixture is negative; that is, the resistivity
tube, and RT is the total input resistance of the decreases with increasing temperature. The
parallel section of the level element. resistivity of Inconel is almost constant over the
In reference to Fig. 1.3.6, if R » 2RL + RT, temperature range under consideration (1100 to
the current / will depend primarily on R, and 1600°F), and therefore Rt can be assumed to
therefore the level element can be supplied with be constant. It follows then that RT will be a
a constant current. The output voltage must be function of Rf and r, and will decrease with
measured with an instrument of high input im increasing temperature.
pedance, such as a vacuum-tube voltmeter or a
Brown recorder. The current flows from the source An increase in fuel temperature will have the
through the compensating resistance R to junction same effect on the output as an increase in level
B. From this junction the current has three if the system is not temperature compensated.
possible parallel paths. The first path is through One method of temperature compensation is
the Inconel rod, the fuel, the 1k-in. pipe, and described below. The element is to be compensated
the ^-jn. plate to junction C; the second path is at a fluid height, h, so that a temperature excursion
through the Inconel tube and the k-in. plate to from T, to T2 will produce no change in output.
junction C; and the third path is through the fuel, If the compensating resistance (R ) is placed in
the Inconel tube, and the l^-in. plate to junction C. series with Rj, the temperature-compensation

34
 PERIOD ENDING DECEMBER 31, 7957

ORNL-LR-DWG 27024

INPUT •
3.15 v, 1 amp
-• OUTPUT
(mv)

%-in. PLATE I
S LWWWWW
GAS SUPPLY
AND VENT LINE
W
VENT
—- VENT HOLES HOLES —-

INCONEL ROD

lV2-in.-IPS
SCHED-40
INCONEL INCONEL
TUBE PIPE —
N

R,
(>>>
N

X
B A'

r > rr > >rr / ! / / ! r ! >/ r'l

TEST VESSEL-
FLUID TRANSFER LINE

CURRENT LEADS
VOLTAGE LEADS
TEMPERATURE COMPENSATION
RESISTANCE (/?.)
';$$&i££$: EUEL RESISTANCE (Rf)

Fig. 1.3.7. Fabrication Diagram of Resistance-Type Fuel Level Probe.

35
ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG 27026
'*• <MMt

fP.Mo-160pslg
100-psig _ SOLENOID VALVE SWITCH MAY BE
FILTER
HELIUM REPLACED BY A REPEAT CYCLE
TIMER FOR AUTOMATIC CYCLING.

^ THROTTLING VALVE

fc<] THREE WAY VALVE

DARKENED PORT CLOSED
V DESIGNATES COMMON PORT

UTJTJ ELECTRIC HEATERS

TO VACUUM TO VACUUM
TUBE VOLTMETER TUBE VOLTMETER

RESISTANCE PROBE ^—Jy RESISTANCE PROBE

FILL \ ysPARE/

VESSEL NO. I VESSEL NO. 2

Fig. 1.3.11. Diagram of System for Testing Resistance-Type Fuel Level Probes.

Preliminary experimentally determined per transducers that had resulted in internal electrical
formance characteristics of an uncompensated shorts and a high rejection rate.
element are shown in Fig. 1.3*12. The actual Design work was completed on all the continuous
performance of the probe was in reasonable level probes required for the ETU and ART for use
agreement with the calculated performance. A
family of curves suitable for designing temperature
compensation into the element is now being
qbtained.

LIQUID-METAL-LEVEL TRANSDUCERS

G. H. Burger R. E. Pidgeon, Jr.3

At the time of suspension of the ETU-ART
program, a total of 22 resistance-type level
transducers for the NaK pump bowls had been
constructed in the ORNL machine shop. An
acceptable, reasonably simple construction tech
nique had been devised for the production of
reliable units. Considerable difficulty was ex 2 3 4
perienced earlier in the fabrication of these FUEL LEVEL (in.)

Fig. 1.3.12. Preliminary Experimental Performance

On loan from Radio Corp. of America. Characteristics of Resistance-Type Fuel Level Probe.

38
 PERIOD ENDING DECEMBER 31, 7957

in the sodium expansion tank, the sodium auxiliary are shown in Fig. 1.3.14 before installation in the
tank, and the NaK-furnace-circuit drain tank. All life testing facility. It is planned to continue the
work was cancelled for construction and testing of life tests indefinitely.
these level probes. During this quarter, two level-
element testing rigs were rebuilt that had been ON-OFF LEVEL PROBES

damaged because of NaK leaks in the system G. H. Burger R. E. Pidgeon, Jr.

(Fig. 1.3.13), and the test program was resumed in
order, primarily, to determine the error introduced Plans for construction and testing of the ETU
by an unwetted level probe. Before this type of and ART on-off level probes described previously4
level probe is wetted by NaK there is contact were cancelled, but similar probes were con
resistance between the surface of the probe and structed for instrumenting test rigs. These probes
the NaK that results in the indicated level being differed from those built previously in that the
lower than the actual level. copper-and-lnconel junction at the tip of the probe
As part of the wetting investigation, three newly was brazed instead of being welded in an attempt
fabricated NaK-pump-bowl tranducers were cleaned to eliminate weld failures. These new probes
with emery cloth and rinsed with alcohol just prior were installed in the NaK systems, but they have
to installation in the test facility. The NaK was not yet operated long enough for an evaluation of
heated and then raised to approximately 4 in. on them to be made.
one of the probes with the NaK at 300°F and on Sixteen on-off probes installed in the gas pots of
two of the probes with the NaK at 400°F. The two NaK system test loops have each operated for
results of the first level change of approximately 3000 hr without a probe failure. These probes
4 in. on each level probe are given in Table 1.3.1. have usually operated with the NaK temperature
After the level changes above, the probes were below 800°F. The control circuit associated with
submerged and calibrated at the test temperature. the on-off probes installed in one test loop failed
With the NaK at 300°F, the indicated level for because of in-gassing of the pump. Apparently,
the 18 points measured was between 0.15 and during operation of the loop, the NaK level reached
0.28 in. below the actual level. For the test with the high-level probe in the throat of the gas pot,
NaK at 400°F the indicated level for the 18 points and the control circuit acted to lower the level.
was between 0 and 0.1 in. below the actual level. When the level had dropped to below the high-level
Further wetting tests are being made at other probe, the sensitive relay across the probe failed
temperatures, and the results will be included in a to energize and allowed helium to be forced into
final report. the loop. The relay failed to operate because the
Four level transducers were installed in a life "pull-in" voltage of the relay had increased and
testing facility. Two of these transducers were the level probe current was adjusted too low. The
previously operated for 3480 hr and the other two 4G. H. Burger and R. E. Pidgeon, Jr., ANP Quar.
were operated previously for 2440 hr. These units Prog. Rep. Sept. 30, 1957, ORNL-2387, p 58.

Table 1.3.1. Results of Tests of Especially Cleaned Level Probes

NaK Temperature T me After In tial Level L svel Error Indicated Accuracy

(°F) Ch ange (hr) (in.) (% of actual level)

300 0 0.31 92.9

7 0.28 93.5

400 0 0.17 95.6

21 0.16 96

400 0 0.13 96.6

21 0.13 96.6

39
ANP PROJECT PROGRESS REPORT

There was no obvious deterioration in performance cause plugging during operation. It was found that
of any of the flowmeters during the periods the flux changed by approximately 3.9% when the
indicated. magnet was heated from room temperature to 365°F.
As mentioned in the previous report,°an apparent It was also found that, with the insulated section
drop in sensitivity of the second lot of units of pipe at 610 to 1722°F, the section of pipe
installed in the test loop was observed in con through the air gap was at 262 to 680°F when it
junction with plugging of the loop flow-control was insulated with approximately 2 in. of Fiberfrax
valve at low loop temperatures. An attempt to and the flow pipe was held in each mounting clamp
relate the apparent nonlinearity of the units in by a series of four corrugated clamps.
this lot with a high oxide level was unsuccessful, Based upon tests on the large flowmeters and
since the effect remained after the oxide level of the flux data obtained during the %,-in. flowmeter
the loop was reduced. It was subsequently found tests, it was expected that the accuracy of the
that a slightly bent pen on the venturi differential- units could be predicted to well within the required
pressure flow recorder, which employs a square- ±15% specified for ETU and ART operation,
root differential-pressure chart, was responsible without the need for magnet flux and temperature
for the apparent shift in sensitivity. The bent pen monitoring. It was also expected that, wi.th proper
caused a fixed error in the differential pressure, insulation of the magnet air gap, the problem of
which appeared as an error in the square root of plugging because of cold spots could be eliminated
the differential pressure that increased as the and yet the magnet would remain cool enough for
flow decreased. The lower flows, as read from the proper operation at the specified fluid temper
chart, showed the effect of the slight shift in pen atures.

position more than the higher flows. The effect

of this error appeared as a shift in sensitivity in TURBINE FLOWMETERS

the calculations. It appears therefore that the G. H. Burger

linearity of the units will be better than 1%, as
stated in a previous report. The data which have The 1-in. turbine flowmeter designated No.
been analyzed indicate that all the units that 2-Rev. 3, which was described previously,7 was,
were calibrated may be expected to provide flow as stated previously,8 installed in a forced-
accuracy well within the specified ART and ETU circulation gas-fired Inconel loop containing
requirements over the entire flow and temperature NaF-ZrF4-UF4 (50-46-4 mole %, fuel 30) for
range prescribed in the flowmeter specifications. calibration and for an endurance test of 3000 hr.
The indications are that, upon completion of data Operation of the loop was started on January 8(
processing and analysis, an expected accuracy 1957, and the turbine operated satisfactorily for
may be assigned that will be somewhat higher than about 340 hr. An investigation then showed that
that specified. the unit was not operating, but it was not removed
A total of fifteen 3/,-in. magnetic flowmeters for from the loop until the loop was drained. At that
ETU cold traps had been assembled and delivered time the loop had operated for 5538 hr at temper
for installation in the ETU. The units, as noted atures ranging from 1200 to 1300°F.
in a previous report, were delivered without the An examination of the unit upon its removal from
magnet attached to permit easier handling for the loop showed that the bearing shaft on the end
installation in the systems. In addition to these of the large turbine had broken and allowed the
fifteen units, parts for 15 more units for the ART unit to drop and bend; the turbine therefore ceased
were completed but not assembled. All com to rotate. No oxidation other than that found after
pleted and incomplete units will be packaged and previous operation in NaK was apparent. No
stored for future disposition. reason for the broken shaft was apparent, but it
As noted in a previous report, it was not planned is thought that the break was probably an indirect
to calibrate the \-\n. units. However, temperature
tests were made on the units to determine magnet
' G. H. Surger, ANP Quar. Prog. Rep. Dec. 31, 1956,
flux changes with magnet temperature and to de ORNL-2221, p 25.
termine temperatures of the pipe in the air gap in 8G. H. Burger, ANP Quar. Prog. Rep. March 31, 1957.
order to check for loop cold spots which could ORNL-2274. p 26.

42
ANP PROJECT PROGRESS REPORT

again started operating and continued operating or oxide plugs in the loop. The units were then
for approximately 100 hr at 1100°F. The loop reassembled and reinstalled in the test loop
temperature was then again raised, very slowly, without water calibration. Since the units were
to 1200°F. This time the unit operated in a normal not recalibrated, the accuracy is not expected to
manner and continued operating until the loop was be as good as before, but the linearity should be
shut down after 644 hr for removal of a defective about the same. The flow rates indicated by the
pressure transducer that was being tested. turbine and magnetic flowmeters in the system are
plotted against pump speed in Fig. 1.3.17. It may
During the operating period, the flow rate was be noted that the linearity of both the turbine
varied from 1.5 to 8.9 gpm. The turbine rotational
flowmeters is quite good when compared with
speed varied from 300 to 3300 rpm, with 300 rpm
that of the magnetic flowmeter, which is known to
being the lowest running speed of the turbine
be linear. It may also be noted that there is
because of bearing friction and blade or rotor drag
considerable error in the flow rates indicated by
in the fluid. It was noted during the test that the
the turbine flowmeters in comparison with the rates
output frequency of the unit became increasingly
given by the magnetic flowmeter. This was
nonlinear below 4 gpm. This was not unexpected,
expected, as explained above, since the units
since the unit was operating from below to just
were not recalibrated and are still operating near
within the design flow range of from 5 to 50 gpm
the lower limit of the design operating range.
for 1-in. turbine flowmeters. No particular effort
was made to achieve linearity or accuracy to a
UNCLASSIFIED
high degree, since the major objective of the 0RNL-LR-DWG 27029
3600
tests was to determine the operating life of the
units at various flow rates and temperatures. • MAGNETIC FLOWMETER
3200
A TURBINE FLOWMETER NO. 4-REV 1
After the loop was drained for removal of the 0 TURBINE FLOWMETER NO. 5-REV i
pressure transducer, it was decided to remove the 2800
turbine flowmeters for examination and repair or
*°s* v
modification, if necessary. The examinations
Q 2400
showed that the bearings were in excellent con
•*
dition, with no signs of oxidation, and the turbines
rotated freely. Further examination disclosed,
however, that the downstream thrust of Unit No. 5
was sufficient to wear slightly the end of the
2000

1600
^
downstream shaft and thus allow the rotor body j#
to contact the positioner body and cause the unit 1200

to drag and stop. It is evident that increased

temperatures would aggravate this condition be
^ /"

cause of the increased softness of the Inconel. 1 2 3 4 5 6 7 8

Unit No. 4 did not show this type of wear to so INDICATED FLOW RATE (gpm)

great an extent, since it was only a five-blade

unit, and it therefore continued to operate. Fig. 1.3.17. Comparison of Flow Rates Indicated by
Two Turbine Flowmeters and a Magnetic Flowmeter in a
Based upon the examination of the units, the Test Loop Circulating NaK at 1200° F.
bearing radial clearances were increased, the
clearance between the downstream end of the rotor Since the units were reinstalled in the test loop
body and positioner was increased to 0.030 in. they have operated in excess of 350 hr at temper
when the rotor was fully positioned in the down atures ranging from 1200 to approximately 1600°F.
stream direction, and the number of blades was It has been demonstrated during these tests that
reduced to five on Unit No. 5 in order to reduce the units can be stopped at high temperatures and
downstream thrust. In addition, the over-all blade then restarted. It thus appears that self-welding
diameter was reduced to give greater clearance of the bearings is not a problem. A detailed
between the blades and turbine housing and thus report on turbine flowmeter development and
reduce the chances of jamming as a result of dirt testing is to be prepared.

44
 PERIOD ENDING DECEMBER 31, 7957

PRESSURE TRANSMITTERS still being tested, with approximately 2000 hr of

C. M. Burton
additional exposure time having been accumulated
since the previous report. Eighteen Chrome I-
Additional tests were made of strain-gage Alumel thermocouples, with closure welds that
pressure transmitters designed for use at high passed dye-penetrant and x-ray inspection, have
temperatures. In these pressure transmitters the now completed over 8000 hr of operation at 1500°F
process pressure acts on a diaphragm, which in sodium. The operating temperature has been
transmits the pressure to a NaK-filled capillary reduced approximately once a week to 1300 and
tube, which, in turn, transmits the pressure to a to 1100°F for readings at these temperatures.
Bourdon spring. The Bourdon spring is linked Since the last report,' no major shifts in per
to an unbonded strain gage, which is connected formances have been observed. Three of the seven
electrically in the form of a Wheatstone bridge. assemblies that earlier exhibited a slight downward
Similar strain gages have been used extensively in shift have started slight upward drifts. Four of
aircraft instrumentation work. the seven assemblies with slow upward drifts have
Primary pressure sensors of the same type but finally drifted approximately 2°F out of tolerance
with pneumatic means for signal transmission were at the 1100°F test temperature. (An accuracy of
tested previously and are being used to obtain ±\ of 1% over the range of 530 to 2300°F is
important pressure measurements in experimental claimed by the vendor of the material being tested.)
test facilities. The possibility of leakage of One assembly that showed a sharp downward drift
radioactive material through pneumatic tubing in at about 3000 hr has continued to drift downward
the event of an explosion, however, has directed and is well out of tolerance. All assemblies
attention toward instrumentation with electrical other than the five mentioned above are presently
transmission. The condition of the electrical operating within the vendor's tolerance. The data
readout equipment can be periodically checked by for all 18 assemblies are summarized in Table 1.3.2.
switching to a calibrated resistor which should Thirteen similar specimens and three platinum,
have a predetermined "synthetic" reading. platinum—10% rhodium units have now completed
The strain-gage bridge is energized with a 5-v d-c over 6000 hr of operation at 1500°F in NaF-ZrF4-
supply, which can be unregulated if the voltage- UF4 (50-46-4 mole %, fuel 30). Readings at
ratio type of measuring equipment is used to approximately one-week intervals were also taken
determine the millivolt signal. This type of at 1300 and 1100°F. No major shifts in per
readout equipment is commercially available as a formance have occurred over the past 2000 hr of
slightly modified form of the familiar electrical operation. All thermocouples except one are now
self-balancing potentiometer. showing upward drift trends but are still within the
Under test conditions, six strain-gage pressure vendor's tolerance. The data for these 16 thermo
transmitters of 0- to 100-psig range indicated an couples are summarized in Table 1.3.3.
average output-input deviation of 0.4% of full
The effect of nuclear radiation on the resistance
scale over their operating pressure range from
of the magnesium oxide insulation was checked
ambient temperature to 1400°F. This accuracy
by inserting three samples of the sheathed ma
compares favorably with the accuracies of other
terial in hole HB-3 of the LITR. Two samples
pressure transmitters. Two of these pressure
were maintained at a constant temperature of
transmitters are now being tested in a NaK loop 1200°F and the third sample was not heated (gamma
under more severe conditions.
heating raised its temperature to 450°F). The
samples were irradiated for three weeks, and a
preliminary examination has revealed that no
ART THERMOCOUPLES
significant immediate effect was produced by
J. T. DeLorenzo radiation - a 10 to 30% decrease in resistance
Sheathed Thermocouples was observed in going from zero power to full
Inconel-sheathed thermocouples (0.250-in.-OD
sheath, MgO insulation) with hot-junction closure J. T. DeLorenzo, ANP Quar. Prog. Rep. Sept. 30,
welds made by the Heliarc welding process are 1957, ORNL-2387, p 61.

45
ANP PROJECT PROGRESS REPORT

reactor power. Further, no significant drift in Tests have revealed the extreme susceptibility
resistance was noted during the three-week interval of magnesium oxide to water absorption. Con
of the test. An upward drift that was noted during siderable baking was necessary to insure dryness
the first 48 hr of the test was believed to be a of the material prior to the tests.
result of additional drying of the magnesium oxide.

Table 1.3.2. Results of Drift Tests of Inconel-

Sheathed Chromel-Alumel Thermocouples in Sodium
Well Thermocouples

Test Period Thermocoup>le Reading (:°f)° Drift tests on well thermocouples designed for
temperature measurement of high-velocity, high-
(hr) Deviation Sprealdc temperature (1400 gpm, 1100 to 1500°F) liquid
metals in Inconel piping were continued; over
Test Temperature: 1500°F 2000 hr of testing has been completed. The ten
0 5.2 6.6 to 3.0 thermocouples (two per well) employed in these
1000 5.1 6.6 to -0.2 tests were fabricated from 20 AWG, solid, Chromel-
2000 5.4 7.4 to 0.8 Alumel wire (accuracy, ±3/g of 1% over range of
3000 4.3 7.0 to -6.5 530 to 2300°F) and were welded, with Heliarc
4000 4.4 7.8 to -11.5 welding techniques, to the bottom of the well.
5000 2.4 8.5 to -10.0 The test environment is static liquid NaK at
6000 1.9 11.5 to -17.8 1500°F. The temperature of the system is reduced
7000 0.8 8.0 to -20.0 to 1300 and 1100°F weekly for readings at these
8000 0.7 9.9 to -27.5 temperatures.

Test Temperature: 1300°F

No major shifts in drift have occurred since the
0 7.3 8.8 to 6.0 start of the tests. All the thermocouples appear
1000 8.1 10.0 to 5.0 to be operating within the tolerance given above.
2000 5.7 8.0 to 2.0 The very close agreement between the two thermo
3000 4.7 8.5 to -3.8 couples in any given well is surprising, and the
4000 1.5 8.0 to -8.5 maximum deviation for all five units is 1°F. The
5000 3.8 9.7 to -6.0 variation in readings between wells can possibly
6000 2.7 9.0 to -10.5 be explained by the high temperature differences
7000 3.5 10.1 to -14.4 that exist in the region of the wells. The data
8000 0.9 9.2 to -23.0 taken to date are summarized in Table 1.3.4.

Test Temperature: 1100°F

All the work done on the design and testing of
0 6.8 8.1 to 5.5
ART thermocouples will be summarized in two
1000 6.9 8.2 to 4.8
ORNL reports to be issued soon. One report will
2000 6.2 8.0 to 3.4
cover the fabrication techniques used and the
3000 6.9 8.5 to 1.5
results of aging tests on Chromel P-Alumel and
4000 4.2 9.5 to -14.0
platinum, platinum—10% rhodium systems in fuel,
5000 4-9 10.0 to -4.0
sodium, and NaK, in addition to the results of
6000 4.6 9.5 to -5.0
rapid heating tests. The second report will
7000 4.2 10.6 to -13.0
describe all aging tests in air on Chromel P—Alumel
8000 4.2 10.5 to -3.1
and platinum, platinum—10% rhodium systems; this
report will also include some recent aging infor
Nineteen assemblies were tested (one was removed mation on thermocouple systems made of iron-
after 5000 hr); each assembly contained two thermocouples
and was aged 24 hrat 1350 F in helium prior to testing. constantan, British-made Chromel P—Alumel,
Deviation of the average of the thermocouple readings Kanthal alloys, Driver-Harris 242-33 alloys, and
from the test temperature. Chromel P—Alumel systems with a titanium oxygen
Maximum spread of the thirty-eight readings obtained. getter.

46
 PERIOD ENDING DECEMBER 31, 7957

Table 1.3.3* Results of Drift Test of Sheathed Thermocouples in Fuel 30

Chromel-Alumel Chromel-Alumel Pt, Pt-10% Rh

Test Period
Assemblies'1 Aged 24 hr Assemblies Aged 200 hr Assemblies0

(hr) Deviation Spread6 Deviation Spreade Deviation Spreade

<°F) (°F) <°F) (°F) (°F) (°F)

Test Temperature: 1500°F

0 1.9 4.0 to -0.5 2.9 3.0 to 2.8 -1.0 2.5 to -3.0

1000 2.7 6.0 to 0.0 5.1 5.5 to 4.8 0.4 1.0 to 0.0

2000 2.4 6.5 to -10.5 7.3 8.0 to 7.5 -0.6 1.0 to -3.0

3000 0.2 6.5 to -17.0 8.4 8.8 to 8.0 -1.4 1.0 to -4.2

4000 0.7 7.5 to -14.5 8.0 8.5 to 7.5 0.7 2.5 to -5.0

5000 2.0 7.6 to -10.5 8.4 9.2 to 7.6 0.4 1.8 to -1.1

6000 1.4 6.0 to -10.0 9.5 9.7 to 9.2 1.0 1.2 to -3.0

Test Te mperature: 1300°F

0 7.0 8.0 to 5.0 6.9 7.0 to 6.9 -0.1 1.8 to -3.8

1000 4.3 7.6 to -0.5 6.9 7.4 to 6-5 -1.2 0.0 to -2.1

2000 2.5 8.0 to -6.2 7.8 8.0 to 7.6 -0.6 1.2 to -3.0

3000 1.9 8.5 to -12.0 8.8 9.2 to 8.5 -0.8 1.2 to -3.0

4000 3.8 8.2 to -5.5 8.8 9.4 to 8.2 -0.6 1.8 to -3.5

5000 3.6 8.4 to -8.5 9.1 9.6 to 8.5 0.3 2.1 to -1.5

6000 2.3 7.3 to -9.0 8.4 8.8 to 8.0 0.1 2.5 to -1.2

Test Temperature: 1100°F

0 4.2 5.4 to 3.0 4.6 4.8 to 4.5 -2.1 -0.2 to -4.2

1000 5.2 7.8 to 3.5 6.4 6.8 to 6.0 -0.3 0.4 to -0.9

2000 2.8 8.2 to -5.0 7.7 8.2 to 7.2 -0.5 1.2 to -1.5

3000 2.4 8.0 to -10.5 8.3 8.6 to 8.0 -0.3 1.2 to -2.0

4000 2.7 7.8 to -9.5 8.1 8.6 to 7.5 1.0 2.8 to -1.0

5000 4.7 8.5 to -6.5 8.7 10.0 to 7.5 1.5 2.9 to 0.2

6000 6.1 10.2 to -4.5 10.6 11.4 to 9.7 3.1 5.4 to 3.3

aEight assemblies were tested; each assembly contained two thermocouples and was aged 24 hr at 1350 F in
helium prior to testing.
Two assemblies were tested; each assembly contained two thermocouples and was aged 200 hr at 1350 F in
helium prior to testing.
cThree assemblies were tested; each assembly contained two thermocouples.
Deviation of the average of the thermocouple readings from the test temperature.
Maximum spread of the thermocouple readings.

47
ANP PROJECT PROGRESS REPORT

Table 1.3.4. Results of Drift Tests of Well Thermocouples in NaK

Well No . 1 Well No. 2 Well No. 3 Well No. 4 Well No .5

Test
Period Deviation" Spread6 Deviation" Spread Deviation" Spread* Deviation" Spread Deviation" Spread
(hr) (°F) (°F) (°F) (°F) (°F) (°F) (°F) (°F)
(°F) (°F)

Test Temperature: 1500°F

0 2.2 0.8 -2.2 0.2 1.0 0.7 -4.3 1.4 -0.4 0.2

1000 1.5 1.0 -3.2 0.4 0.9 0.2 -4.2 1.2 -0.9 0.1

2000 4.5 1.0 -1.5 0.2 3.7 0.5 -0.3 0.5 3.4 0.0

Test Temperature: 1300°F

0 3.1 0.8 0.0 0.1 -3.5 0.1 -5.6 0.6 -4.5 1.0

1000 2.7. 0.8 -1.5 0.1 2.9 0.0 -2.3 0.1 2.4 0.0

2000 4.0 0.7 -0.6 0.2 3.6 0.6 -2.6 0.9 2.9 0.1

Test Temperature: 1100°F

0 -1.4 1.1 -4.1 0.4 -1.8 0.4 -1.7 1.4

1000 2.5 0.5 0.7 0.5 3.0 0.7 -3.7 0.4 -1.3 0.1

2000 4.8 0.4 1.3 0.2 3.4 0.4 -1.5 0.5 2.6 0.4

deviation of the average of the two thermocouple readings from the test temperature.
Spread between the two thermocouple readings.

48
 PERIOD ENDING DECEMBER 31, 7957

1.4. ENGINEERING DESIGN STUDIES

A. P. Fraas

APPLIED MECHANICS AND STRESS ANALYSIS performed on the IHE-8B heat exchangers,5 and
R. V. Meghreblian a summary description of the core shell-liner
experiment.
Summaries and final reports on the applied Reports were also completed on the temperature
mechanics and stress analysis problems of the distribution in the ART reflector-moderator as
ART have been prepared. The reports that have
sembly, on the stress analysis of the radiators,8
been prepared and issued include an analysis of the and on the thermal stress analysis of the ART
north-head pressure-stress distribution, a des heat exchanger channels and header pipes.9
cription of a general relaxation method for deter Two memoranda were also prepared on general
mining two-dimensional temperature distributions on methods for analyzing plates and shells.10'11
an IBM-650, an analysis of a thermal cycling
experiment performed on beryllium by using
volume heat sources, the results of a study of D. L. Platus, Summary of IHE-8B Heat Exchanger
Stress Analysis. ORNL CF 57-12-7 (in p/ess).
the temperature distribution at the junction of
D. H. Platus, Summary of Core-Shell-Liner Experi
shell V (lower) and shell VI,4 the results of ment, ORNL CF-57-1-134, Addendum (Dec. 11, 1957).
the stress analysis of controlled structural tests 7D. L. Platus, D. M. Miller, and R. V. Meghreblian,
ART Reflector Temperature Distribution, ORNL-2425
(Nov. 2, 1957).

'S. E. Moore, ART North Head Analysis, ORNL CF-57-

8D. H. Platus, ORNL Tube-Fin Radiator Stress Analy
sis, ORNL-2441 (in press).
10-114 (in press). g
D. L. Platus, Thermal Stress Analysis of the ART
D. M. Miller, Relaxation Calculations by Machine for Heat Exchanger Channels and Header Pipes, ORNL-2442
Two-Dimensional Heat Conduction Problems, ORNL (in press).
CF-57-11-4 (Nov. 1, 1957).
3 B. L. Greenstreet and D. M. Miller, An Extension of
S. E. Moore, Canned Cylindrical Beryllium Specimen the Esslinger Shell Relationships to Include Effects of
with Internal Heating, ORNL CF-57-10-109 (Oct. 30, Axial Load and Temperature, ORNL CF-57-11-65 (Nov.
1957). 19, 1957).
D. M. Miller, Temperature Profile of Shell V Lower- F. J. Stanek, Stress Analysis of Flat Circular Plates,
Shell VI Connection, ORNL CF-57-11-9 (Nov. 1, 1957). ORNL CF-57-11-111 (Nov. 25, 1957).

49
ANP PROJECT PROGRESS REPORT

1.5. DESIGN PHYSICS

A. M. Perry

The reactor physics problems associated with by the room-temperature critical experiments.
the design of the ART were the same as those that While the analysis indicated a negative temperature
arise in the design of any reactor, but there were coefficient of reactivity, the marked dependence
major shifts in relative importance. Typical of critical concentration on the core shell thick
problems that were studied included critical core ness caused some concern that the reactor might
size, or critical mass of fuel; reactivity effects of in fact have a positive moderator temperature
temperature, poisons, control rods, and so forth; coefficient, and led to the performance of a high-
radiation heating of various parts of the reactor; temperature critical experiment which proved the
activation or neutron-induced changes in the compo moderator coefficient to be negative. The experi
sition of various materials; and shielding. ment was based on an exact mockup of the design
For a fused-salt reactor, the problem of critical core. Nevertheless, there were a number of dif
mass may be viewed in a somewhat different ferences, such as the absence of reflector cooling
perspective than for a fixed-fuel reactor. Since and deviations in the outer boundary of the reflector,
the concentration of fissionable material can be whose cumulative effect on critical concentration
readily adjusted to achieve criticality or the proved to be quite significant. The required
desired excess reactivity, the urgent problem is concentration of UF. in the ART fuel, including
not so much to know in advance exactly how much allowance for reactivity changes during the core
fuel the reactor will require, as to be assured that life, was expected to be about 5 to 5.5 mole %, in
the melting point of the fuel, which depends upon contrast to the critical concentration in the high-
the uranium concentration, will remain within temperature critical experiment of 2.84 mole %.
acceptable limits for any foreseeable additions of A major problem in the circulating-fuel reactor
UF4 to the fuel. is activation of the secondary coolant. While the
The criticality problem for the ART was initially thermal neutrons leaving the reflector can be
one of finding a suitably compact core geometry, effectively absorbed in a boron-containing layer
while at the same time keeping the UF. concen between the reflector and the heat exchangers,
tration within acceptable limits. The work of the some activation remains because of higher energy
Laboratory on this problem was augmented by neutrons absorbed in the sodium resonance at
a contract with the Curtiss-Wright Corporation for 2.9 kev, or moderated in the heat exchanger, and
a parametric survey of fuel concentration and core because of delayed neutrons emitted in the heat
power distribution, upon which the physical dimen exchanger. The problems of determining the level
sions of the ART core and reflector were finally of sodium activity and of finding a satisfactory
based. The composition and thickness of the core means of keeping the activity as low as possible
shell were shown to have a decisive influence on have been examined extensively, both analytically
critical fuel concentration;' this was confirmed and experimentally.
The total Na activity expected in the ART
'A. P. Fraas and C. B. Mills, A Reflector-Moderated secondary coolant loop was 1000 curies. It was
Circulating Fuel Reactor for an Aircraft Power Plant, evident, however, that the level of activity could
ORNL CF-53-3-210 (March 27, 1953).
2C. S. Burtnette, M. E. LaVerne, and C. B. Mills,
Reflector-Moderated-Reactor Design Parameter Study,
ORNL CF-54-7-5 (Nov. 8, 1954). 6A. D. Callihan et al., ANP Quar. Prog. Rep. March
10, 1955, ORNL-1864, p 43.
3M. E. LaVerne and C. S. Burtnette, ANP Quar. Prog.
Rep. June 10, 1954, ORNL-1729, p 32. 7A. D. Callihan et al., ANP Quar. Prog. Rep. Sept.
10. 1955. ORNL-1947, p 58.
H. Reese, Jr., S. Strauch, and J. T. Mihalczo, Geom
etry Study for an ANP Circulating Fuel Reactor, WAD- 8A. M. Perry, ANP Quar. Prog. Rep. Sept. 10, 1955,
1901 (Sept. 1, 1954); C. B. Milts and H. Reese, Jr., ORNL-1947, p 33.
Design Study of an ANP Circulating Fuel Reactor, 9H. W. Bert in i, ANP Quar. Prog. Rep. March 10, 1956.
WAD-1930 (Nov. 30, 1954). ORNL-2061, P28.
5W. K. Ergen, ANP Quar. Prog. Rep. June 10. 1955, 10J. B. Dee et al, ANP Quar. Prog. Rep. Dec. 10,
ORNL-1896, P20. 1955, ORNL-2012, pp 204, 209.

50
 PERIOD ENDING DECEMBER 31. 7957

be very substantially reduced — perhaps by a factor The control rod in the ART illustrates the usual
of 5 or more — in an aircraft reactor. considerations of depletion of the active material
Closely related to the NaK activity is the activity and of internal heating to a marked degree. It
of material activated in the heat exchangers and was found20'21 that a l^-in.-OD annular rod,
carried to the radiators by mass transfer in the 4 in. thick, would be adequate if composed of
NaK stream. In the case of the ART, the activity rare-earth oxide at /3 to L the normal density,
in the radiators after shutdown and removal of the with the voids occupied by metal or sodium to
NaK was expected to be about 1 curie, and the achieve high thermal conductivity. The reactivity
dose rate 1 ft from the radiator surface was expected value of the control rod was determined entirely
to be about 1 r/hr, due mostly to Co (ref 11). by critical experiments.
A major area of investigation in reactors with Most of the considerations affecting the shielding
high power density is the heating of structural requirements for the ART are described in the
components by gamma radiation and neutrons. Report of the 1953 Summer Shielding Session.
Calculations showed that, in the ART, the heating Additional studies relating particularly to the
of the reflector, the core shells, the north head shielding of various components included the NaK
structures, and many other components would be pipes, 23 fuel drain line,23 fuel overflow line,23
24
substantial, especially from gamma radiation, and vapor traps, and the fill-and-drain tank
special provisions were provided for cooling them.
A complete summary of the radiation heating calcu
lations may be found in two topical reports; ' 16R. B. Stevenson, ANP Quar. Prog. Rep. June 10,
several aspects of the problem have been discussed 1956, ORNL-2106, p 35.
in progress reports, that is, heating in the vicinity 17R. B. Stevenson, ANP Quar. Prog. Rep. Sept. 10,
of the pressure vessel, the north head, the 1956, ORNL-2157, p 31.
18
off-gas system, and in the beryllium and core C. M. Copenhaver, ANP Quar. Prog. Rep. Dec. 31,
1956, ORNL-2221, p 15; R. B. Stevenson, ANP Quar.
shells.17'18 Prog. Rep. Dec. 31, 1956, ORNL-2221, p 16.
19W. K. Ergen and H. W. Bertini, ANP Quar. Prog.
Rep. March 10. 1955, ORNL-1864, p 26.
11 A. M. Perry, ANP Quar. Prog. Rep. March 10, 1956, 20W. Fader, ANP Quar. Prog. Rep. June 10. 1955,
ORNL-2061, p 30. ORNL-1896, p 22.
12
H. W. Bertini et al., Basic Gamma-Ray Data for
21W. Fader and A. M. Perry, ANP Quar. Prog. Rep.
Sept. 10, 1955, ORNL-1947, p 33.
ART Heat Deposition Calculations, ORNL 2113 (Oct. 3,
1956). 22
E. P. Blizard and H. Goldstein (eds.), Report of
13 the 1953 Summer Shielding Session, ORNL-1575 (July
H. W. Bertini et al., Radiation Heating in the ART,
ORNL-2429 (to be published). 12, 1954).
23
14H. W. Bertini, ANP Quar. Prog. Rep. June 10, 1956, H. W. Bertini, C. M. Copenhaver, and A. M. Perry,
ORNL-2106, p 28. ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 71.
15H. W. Bertini and D. L. Platus, ANP Quar. Prog. 24H. W. Bertini, ANP Quar. Prog. Rep. Sept. 30,
Rep. June 10, 1956, ORNL-2106, p 34. 1957, ORNL-2387, p 80.

51
ANP PROJECT PROGRESS REPORT

1.6. MATERIALS AND COMPONENTS INSPECTION

A. Taboada

MATERIALS INSPECTION WELD INSPECTION

Approximately 4000 requests were received for The weldments inspected were made almost
the inspection of materials for the ART-ETU entirely by the inert-gas shielded-arc method
program. The total quantities inspected and the and were predominantly of Inconel material.
rejection rates are given in Table 1.6.1. At the Visual, dye penetrant, and radiographic inspection
time of suspension of the program 61,243 ft of techniques were used for the inspection of all
CX-900 Inconel tubing was being inspected or thin-walled tubing and other critical weldments
was awaiting inspection. where complete reliability was essential. Such

Table 1.6.1. Total Quantities of Materials Inspected for the ART-ETU Program and the Rejection Rates

Quantity Quantity Rejection

Item Type Inspected Rejected Rate (%)

Tubing Regular Inconel 32,000 ft 5,000 ft(a) 15.5

Regular Inconel 54,000 ft 54,000 ft(fe) 100.0

CX-900 Inconel 73,000 ft 18,225 ft(c) 25.0

Stainless steel of various 813 ft 266 ft 33.0

types

Incoloy 70 ft 0 ft 0.0

Hastelloy (tubing and 1,390 ft 816 ft(rf) 59.0

small pipe)

INOR-8 128 ft SftW 4.0

Pipe Regular Inconel 12,300 ft 1,329 ft(a) 11.0

CX-900 Inconel 1,470 ft 70 ft 5.0

2(e)
Plate and sheet CX-900 Inconel 1,574 ft2 198 ft2 12.5

Other 7,350 ft2 257 ft2 3.5

Rod and bar CX-900 Inconel 50 ft 0 ft 0.0

Other 2,381 ft 139 ft 6.0

Fittings Machined adapters, reducers, 600 ft 43 ft 7.0

ells, pipe caps, etc.

(a) Used for test loop applications; rejected for defects greater than 5% of the wall thickness or cracks.
lb) Rejected on the basis of a random 10% sampling inspection.
MrReactor heat exchanger and radiator tubing rejected for defects greater than 2% of wall thickness. Of the re
jected tubing 8700 ft was returnable to vendor and 6150 ft was salvageable.
Rejected for gross defects only.
*e'High rejection rate resulted from poor surface condition.

52
 PERIOD ENDING DECEMBER 31, 7957

welds are designated C in the summary presented December 1954 to September 1957 is summarized
in Table 1.6.2. Weldments which did not require in Table 1.6.2. Prior to December 1954, non
radiography, either because of inaccessibility or qualified welders had been used, and rejection
because the weldment did not require complete re rates approached 40%. The culmination of this
liability, are designated CN in Table 1.6.2. work was the ETU assembly welding. Two
Heavy section weldments on test loop pipes were hundred and fifty-six weldments were made, of
also designated CN. which 40, or 15.5%, were rejected. The ETU
weld rejection rate at termination was approxi
The ANP welding performance record from mately 8%.

Table 1.6.2. Total Quantities of Weldments Inspected for the ART-ETU Program and the Rejection Rates

C-Type Weldments CN-Type Weldments

Period Quantity Quantity Rejection Quantity Quantity Rejection

Inspected Rejected Rate (%) Inspected Re jected Rate (%)

1954 to Oct. 1955 1,358 275 20 276 44 14.5

Oct. 1955 to Oct. 1956 11,067 1,318 12 1,163 31 2.5

Oct. 1956 to Sept. 1957 3,244 573 18 11,292 130 1.0

15,669 2,166 12,731 205

53
ANP PROJECT PROGRESS REPORT

1.7. HEAT TRANSFER STUDIES

H. W. Hoffman

THERMAL-CYCLING RESEARCH and the geometries of the test pieces, as well

Pulse-Pump System as the direction of flow and the location of the
samples taken for metallographic study, are
J. J. Keyes A. I. Krakoviak shown in Fig. 1.7.1. The photographs of Fig.
J. E. Mott1 1.7.2 show the results of examination of the
Several additional tests were completed in test piece from run 5. In Fig. 1.7.2a some
the high-frequency thermal-cycling loop with the surface cracks are visible in the transition
fuel mixture NaF-ZrF4-UF4 (56-39-5 mole %) as section near the entrance of the heavy-walled
the circulating fluid and helium as the pulsing test piece (sample e). An approximately 0.070-in.
and cover gas. A Milton Roy pump with a 1k-in.- crack in the wall of the same test piece at a
dia piston and a 4-in. stroke was used to generate position 3 in. downstream of the entrance
approximately sinusoidal pressure oscillations at (sample b) is shown in Fig. ].7.2b. The total
frequencies of 0.4 and 1.0 cps. These pulsations failure of the Inconel pipe in the vicinity of a
were transmitted to the fuel in the pulse chamber weldment (sample c) may be seen in Fig. 1.7.2c.
so as to force alternate hot and cold slugs of A cursory examination of the data indicates
liquid through the test section. Outside wall that the severity of cracking as a result of thermal
temperatures and, in one case, fluid stream tem cycling increases with increasing pipe wall thick
peratures were measured instantaneously; the in ness and with proximity to a weld. This study
side wall temperature amplitude was estimated by is being continued on a limited basis.
using data from previous water tests. The test The basic investigation of thermal transients
sections were constructed of Inconel tubing and at a solid-fluid interface was continued, with
pipe that was approximately 0.5-in. in inside an experimental study in which the conditions
diameter and varied in wall thickness. of an adiabatic infinite wall were approximated.
The test conditions and results for runs 3 The results, obtained for three different film
through 5 are given in Tables 1.7.1 and 1.7.2, heat-transfer coefficients ranging between 2500
and 5000 Btu/hr-ft2.°F, have validated the theo
retical analysis. A memorandum giving a detailed
description of the experimental system and a
On assignment from Pratt & Whitney Aircraft. discussion of the results has been issued.
A complete description of this experimental system
may be found in the following ANP Quarterly Progress J. E. Mott, An Investigation of Thermal Transients
Reports: ORNL-2157, p 56, ORNL-2221, p 54, and at a Solid-Fluid Interface, ORNL CF-57-10-122 (Oct. 25,
ORNL-2274, p 53. 1957).

Table 1.7.1. Conditions of High-Frequency Thermal Cycling Tests of Inconel Pipe

Maximum Temperature Fi•equency

Average Fluid Total Total Fluid
Test Diflierence Between of T emperature Total
Temperature Time** Flow Rate
No.* Hot and Cold Streams Oscillations Cycles
(°F) (hr) (gpm)
(°F) (cps)

3 1360 288 1.0 99.5 360,000

4 1387 454 1.0 250.0 900,000

5 1412 493 0.4 135.0*** 195,000 5.8

*Tests 1 and 2 were considered cleaning and "shakedown" runs and therefore inconclusive.
**Neglecting isothermal time.
***Estimated time at failure.

54
 Table 1.7.2. Results of High-Frequency Thermal-Cycling Tests of Inconel Pipes

Wall Measured Temperature Estimated Temperature

Test Sample Results of Metallographical Examination
Thickness Amplitude on Outside Amplitude on Inside
No. Designation of Test Piece and/or Adjacent Piping
(in.) Pipe Wall (°F) Pipe Wall (°F)*

0.091 +64 ±68, ±95 Moderate to heavy void formation to a depth of 2 mils;
some intergranular cracks or attack

0.029 ±64 ±68, ±95 Very light and scattered void formation to a maximum
depth of 2 mils

0.133 Not measured ±90, ±112 Heavy subsurface void formation to a depth of 3 mils

0.091 Not measured ±125, ±150 Heavy intergranular cracks to a depth of 20 to 80 mils
found immediately above and below welds

0.065 Not measured ±125, ±150 Moderate intergranular cracks which appeared to pro
duce spoiling

0.029 ±85 ±93, ±125 Very few general and intergranular voids to a depth
of 3 mils
-0
m
0.133 ±20 ±150, ±200 Moderate intergranular cracks, especially near weld 70

5
0.250 ±7 ±150, ±200 Heavy intergranular cracks to a depth of 72 mils O
m

0.091 ±68 ±100, ±175 Moderate intergranular cracks with heavy cracks in o
heat-affected zone of weld where failure occurred
o

o
0.133 ±15 ±100, ±175 Light intergranular cracks to a depth of 2 to 3 mils m
o
m
*First value calculated from outside surface temperature; second value calculated from fluid temperature amplitude by using the 77 ratio. £
CD
k*Subjected to conditions of tests 3 through 5. m
70

(Jl
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27030

FLOW FLOW

FLOW

WELD
MPLE a
WELD
SAMPLE a
SAMPLE a
WELDS SAMPLE e

WELL

SAMPLE b

SHEATH-TYPE STREAM
WELD
THERMOCOUPLE

WELD
WELDS
SAMPLE c
WELD

SAMPLE d-

TEST PIECE 3 TEST PIECE 4 TEST PIECE 5

Fig. 1.7.1. Schematic Diagram of Test Pieces and Sample Locations

Pressurized System studies is being assembled and annotated for

future publication. The data consist primarily
H. W. Hoffman D. P. Gregory1
of Pitot traverse measurements and dye-injection
The experimental investigation of the effect movies.
of thermal-stress cycling on Inconel tubes filled No further work has been done on the experi
with flowing NaF-ZrF4-UF4 (50-46-4 mole %) with mental studies of flow through quarter-scale
the use of a pressurized flow system was dis models of the ART core. A complete report on
continued. A report summarizing the results of this work has been prepared. Work done on the
this experimental study is being prepared. screen-packed core and header combination while
simulating single and dual pump operation will
ART HYDRODYNAMICS be reported in a later memorandum.
G. L. Muller1 F. E. Lynch A report describing the phosphorescent-particle
W. J. Stelzman technique for velocity-profile visualization is also
being prepared.
The investigations of the mechanics of the ART
core fluid flow for which the full-scale plastic FUSED SALT HEAT TRANSFER
core model was being used have been terminated.
S. I. Cohen H. W. Hoffman
The tests of the core that was divided into three
flow regions by installing two annular shell Forced-convection heat transfer studies with
liners were not completed. A discussion of KCI-LiCI (41.2-58-8 mole %) flowing in a heated
the design has been prepared by Platus. The tube were continued. The results obtained are
mass of data obtained in the course of these
D. H. Platus, Summary of Core-Shell-Liner Experi
ment, ORNL CF-57-1-134, Addendum (Dec. 11, 1957).
H. W. Hoffman and D. P. Gregory, An Experimental G. L. Muller, J. 0. Bradfute, and F. E. Lynch, Investi
Study into the Effects of Surface Thermal Cycling on gation of Fluid Flow in the ART and Other Reflector-
Inconel in a Fluoride Fuel Environment, ORNL-2435 Moderated Reactor Cores, ORNL-2199 (in press),
(to be published). a

F. E. Lynch, L. D. Palmer, and G. M. Winn, Instan

5W. J. Stelzman, ANP Quar. Prog. Rep. Sept. 30, 1957, taneous Velocity Profile Measurement by Photography,
ORNL-2387, p 96. ORNL-2257 (to be published).

56
 ANP PROJECT PROGRESS REPORT

presented graphically in Fig. 1.7.3 in terms of made to weight the data from either of these
the heat transfer parameter NN(j/Np'4 The data, sources. To completely explain the observed
which cover the Reynolds modulus range between discrepancy in the heat transfer parameter, a
6,000 and 18,000, were obtained for both type decrease in the thermal conductivity of the salt
347 stainless steel and Inconel test sections. to a value of 0.3 Btu/hr-ft«°F would be needed.
On the basis of previous experimental measure
UNCLASSIFIED
ORNL-LR-DWG E703I
ments of the thermal conductivity of fused salts,
200
_
a change of this magnitude (0.8 to 0.3 Btu/hr«ft'°F)
ASE =?IES GA T-irPE 54 7 STAINLESS STEEL
between preliminary and final values is not
• SE RIES GB T^ fPi 54 7 STAINLESS STEEL

oSE RIES GC IN CC N EL
anticipated. Thus, incorrect thermal conductivity
100
values can provide only a partial explanation of
the experimental results.
It was found in earlier work with NaF-KF-LiF
(11.5-42-46.5 mole %) in Inconel systems1' that
< 50 a reaction occurred between the salt and the
AW 0.8
metal to form a film of high thermal resistance
= 0.0 23 /<
Re at the salt-metal interface. Visual observation
of the inside surface of the test section from
• o
series GA did disclose a very thin reddish-brown
20
I
surface deposit. Spectroscopic and x-ray analyses
A
•A
of this film material have been inconclusive.
ft
Thus, the question as to whether this film is
10
A
A
'I due to a surface reaction or to deposition from
2X10 5 10" 4XI04 an impure salt mixture remains unresolved. Work
REYNOLDS MODULUS N„
on this program has been discontinued, and the
results are summarized in a series of reports.'' ~' *
Fig. 1.7.3. Heat Transfer Characteristics of KCI-LiCI
(41.2-58.8 Mole %). MOLTEN LITHIUM HEAT TRANSFER

The data of the first experiment (series GA) N. D. Greene

for a type 347 stainless steel tube fell approxi Preliminary designs are being prepared of an
mately 55% below the general heat transfer experimental system for the study of forced-
correlations for standard fluids. These results convection heat transfer with molten lithium
were corroborated in succeeding tests with a flowing in metal tubes. The loop proposed will
new type 347 stainless steel test section consist of a gas-sealed pump, a magnetic flow
(series GB) and with an Inconel test section meter, a test section, and an air-cooled heat
(series GC). An analysis of the system indicated sink. In order to mitigate the difficulties and
that the observed deviations might result from hazards encountered in earlier studies with
the use of incorrect thermal conductivity values lithium, an effort will be made to keep the volume
in evaluating the data and/or the existence of
high thermal resistance films at the salt-metal
H. W. Hoffman and J. Lones, Fused Salt Heat Trans
interface. fer, Part II: Forced Convection Heat Transfer in
The thermal conductivity value used for the Circular Tubes Containing NaF-KF-LiF Eutectic.
ORNL-1777 (Feb. 1, 1955).
salt in the analyses of the data was obtained by 12
H. W. Hoffman, Turbulent Forced Convection Heat
averaging the preliminary value of 0.8 Btu/hr«ft'°F Transfer in Circular Tubes Containing Molten Sodium
Hydroxide, ORNL-1370 (Oct. 3, 1952).
experimentally determined by Powers9 and the
value of 1.2 Btu/hr-ft-T predicted on the basis H. W. Hoffman and S. I. Cohen, Fused Salt Heat
Transfer. Part III: Forced-Convection Heat Transfer in
of correlations by Gambill.10 No attempt was Circular Tubes Containing the Salt Mixture NaNO^-
NaNOyKN03, ORNL-2433 (in preparation). Z
o
W. D. Powers, unpublished data. H. W. Hoffman, S. I. Cohen, and D. P. Gregory, Ex
perimental Determination of Forced-Convection Heat
W. R. Gambill, Prediction of the Thermal Conduc Transfer with Several Fuel and Coolant Mixtures, ORNL
tivity of Fused Salts, ORNL CF-56-8-61 (Aug. 10, 1956). CF-58-1-1 (in preparation).

58
 PERIOD ENDING DECEMBER 31, 1957

of liquid metal to a minimum. The lithium will heat transfer with liquid metals. The empirical
be maintained at approximately 700°F, and the equation
Reynolds modulus in the test section will be k(t - t m')
at least 50,000. v w 1
(1) = T
The test section will be constructed of type Wr2 79 + 0.226 N°;92
347 stainless steel tubing having an inside
diameter of approximately 3/,, in. Heat will be
t = wall temperature,
generated electrically in the tube wall by the
passage of a high amperage current. The heat t = mean fluid temperature,
generation in the lithium can be reduced to k = thermal conductivity,
approximately 7.6% of the total for a tube wall
W = power density,
thickness of /32 in. Test section wall tem
peratures, fluid flow rates, and mixed-mean r = tube radius,
temperatures will be measured. Np = Peclet modulus (the Reynolds-Prandtl
product),
VOLUME-HEATED SYSTEMS was found to describe the analytical results pre
viously reported to within ±15% over the limits
ART-Type Core With Screens
0 < NPf £ 0.1 and NRe ^ 10,000. The entire
N. D. Greene turbulent range (NRe ^ 5000) can be represented
The study of the effect of screens in a diverging to a lesser degree of accuracy (approximately
annular channel on the thermal structure of a ±30% at low NR ) by the equation
fluid under volume-heat-source conditions is
k(t - t r.
being continued on a limited scale. Four screens ^ w 1
(2) = T =
of varying radial solidity have been fabricated
and are currently being installed in the test
Wr' 65 + 0.24/Vp;915
model. These screens were designed to increase The results obtained previously are compared in
the resistance in the high-velocity regions of Fig. 1.7.4 with results of analyses made by
the flow so as to force the fluid into the lower Poppendiek and Palmer. It may be seen that
velocity region near the outer wall. for values of the Peclet modulus above 1000
the agreement is excellent.
Liquid-Metal Experiment
HEAT TRANSFER WITH VORTEX FLOW
G. L. Muller
W. R. Gambill N. D. Greene
Construction of the system described pre
viously for the experimental investigation of Previous experiences with source-vortex flow of
heat transfer in a liquid metal (mercury) with fluids in forced-convection heating (no phase
internal heat generation has been essentially change) indicated that such a flow field might
completed. The test-section electrical-power leads be of particular value in a boiling system. It
and instrumentation are now being installed. was felt that, in comparison with other types
The three flow-measuring orifices have been of flow fields, the gravitational acceleration pro
calibrated with mercury by using a weigh tank. duced by the rotating fluid would keep a concave
Differential pressure cells and a recorder were heated surface freer of bubbles at a given heat
used to measure the orifice pressure drop. A flux by providing an increased buoyant driving
few minor leaks around valve stems and pump force for inward radial bubble migration. Delay
shaft seals have been corrected. of the burnout condition to a higher heat flux
An attempt has been made to obtain a universal
relation for forced-convection volume-heat-source
16G. L. Muller, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 102.
H. F. Poppendiek and L. D. Palmer, Forced Con
15
G. L. Muller, ANP Quar. Prog. Rep. Sept. 30, 1957, vection Heat Transfer in Pipes with Volume Heat Sources
ORNL-2387, p 106. Within the Fluids, ORNL-1395 (Dec. 17, 1952).

59
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
-1 ORNL-LR-DWG 27032
10 \

0 91*
65 + 0.24 N
Pe

A— —
—-/VRe = 5000
10
2
7 ..

IN" ., ~^
1 -—-/vRe- 1U,UUU
T --
-,r, rs oo^- .,0.92 ^iSS> 1
-/y = 1,000,000
y
s^*
/VRe-100,000

3
10

4
10
10 10 2 5 10° 2 103
PECLET MODULUS, /VPe

Fig. 1.7.4. Correlation of Liquid-Metal Volume-Heat-Source Forced-Convection Heat Transfer Data.

value was foreseen as the primary advantage If all Gunther's conditions were identical with
of this phenomenon. those of ORNL run No. 4, it is estimated that
These early concepts have been justified to his flux value would be 2.3 x 106 Btu/hr«ft2, at
a considerable extent by preliminary data taken the most; thus the advantage for the vortex case
for the burnout heat fluxes for water in nucleate would be 378%. Gunther's values were chosen
boiling, both subcooled throughout and with small for comparison here because they are consistently
net steam generation. These data are presented among the highest reported in the literature and
in Table 1.7.3. Comparisons of these burnout should give a conservative evaluation. Gunther's
heat flux values with literature data for straight- data were obtained for a rectangular flow passage
through linear water flow at the same conditions within which was suspended a thin (0.004-in.)
of axial fluid velocity, L/D, exit pressure level, electrically heated metal strip or wafer that
and exit subcooling (defined as saturation tem was cooled on both sides. This geometry gave
perature minus actual liquid temperature, at considerably higher burnout values for linear
exit), indicate approximately a 500% superiority water flow than those obtained in similar tests
for the vortex case. A typical comparison is with straight, round tubes.
shown in Table 1.7.4, where test No. 5 by The high burnout values reported are not
Gunther is compared with vortex flow run No. 4, counterbalanced by an inordinately large power
for which the pertinent experimental conditions expenditure, as is indicated by the ratio desig
were almost the same. nated Pt/P in Table 1.7.3. The ratio values
given, in per cent, are the ratios of power input
required for fluid flow across the test section
18
F. C. Gunther, Trans. Am. Soc. Mech. Engrs. 73, to the heat absorption rate of the water in the
115-123 (1951). same power units. As may be seen the heat

60
 PERIOD ENDING DECEMBER 31, 1957

Table 1.7.3. Results of Investigation of Burnout Heat Fluxes for Water in Nucleate Boiling

Test Section
Axial Fluid Degree Inlet Static Exit Burnout
Outside Heated
P,/p
Run inside Quality f i
Velocity of Subcooting Pressure Heat Flux
No.
(fps) at Exit (°F) (psig) Diameter, D. Diameter Length, L l/d{ Material
(%)*
(%)** (Btu/hr-ft2)
(in.) (in.) (in.)

1 12.1 36 93 0.275 0.300 3.063 11.14 Copper 0 0.26 5,940,000

2 11.8 0 94 0.275 0.300 3.187 11.60 Copper 1.24 0.19 8,240,000

3 12.35 0 99 0.275 0.2875 (inlet) 2.875 10.44 Copper 1.49 0.20 9,690,000
0.300 (exit)
4 13.61 72 100 0.250 0.300 1.20 4.80 Inconel 0 0.43 10,980,000
5 13.18 135 248 0.300 0.375 3.05 10.16 Copper 0 0.37 8,630,000

6 16.3 150 248 0.250 0.270 (inlet) 2.25 9.00 Inconel 0 0.39 10,810,000
0.281 (exit)
7 24.0 49 285 0.250 0.270 (inlet) 1.625 6.50 Inconel 0 1.01 17,250,000
0.278 (exit)

*Percentage of water in steam at exit.

"Ratio (expressed in percentage) of power input required for fluid flow across test section to heat absorption rate of the fluid in the same
power units.

Table 1.7.4. Comparison of Heat Transfer Data have been reported in greater detail in a recently
for Vortex Flow and Straight-Through Linear Flow issued ORNL memorandum.

Vortex Straight-Through TOWER SHIELDING REACTOR II (TSR-II)

Condition HYDRODYNAMICS
Flow* Flow**

W. R. Gambill
Axial water velocity, 13.6 12.3
Detailed design work is nearly completed on
fps
an enclosure assembly for a full-scale model
Exit pressure, psia 15.0 14.4 of a quarter segment of the TSR-II to be used
Exit subcooling, F 72 71 in hydrodynamic studies. The data on the flow
distribution in the coolant annuli will be com
L/D. 4.8 6
bined with calculated heat-generation rates to
Burnout heat flux, 11.0 x 106 2.0 x 106 determine the magnitude of redistribution of flow
Btu/hr-ft2 necessary to approximate a flat exit-water tem
perature profile. In these tests the exit annulus
*Data obtained by Gambill and Greene (see Table gap and the radial location of the control assembly
1.7.3). at the core center will both be varied. The
**Data obtained by Gunther (ref 18). necessary pumps and instrumentation are being
obtained concurrently with design work on the
absorption rate ranges from 99 to 526 times the enclosure.
test-section flow power.
PHYSICAL PROPERTIES
An interesting observation was made recently
in connection with runs 6 and 7. In run 6, the W. D. Powers
exit pressure level was held at 100 psig (rather Thermal Conductivity
than atmospheric as in all the other runs), but The variable-gap thermal-conductivity device
little change was noted in the burnout heat flux. modified to include a heat meter of larger area
In run 7, however, a maximum inlet pressure was
generated for the maximum flow rate and gravita
tional field, and the very high burnout heat flux "w. R. Gambill and N. D. Greene, AStudy of Bumout
Heat Fluxes Associated with Forced-Convection, Sub-
attained indicates that velocity effects are more cooled, and Bulk Nucleate Boiling of Water in Source-
important than pressure effects. These studies Vortex Flow, ORNL CF 57-10-118 (Oct. 29, 1957).

61
ANP PROJECT PROGRESS REPORT

has been used to redetermine the thermal con LiCI-BaCI2 (70-30 mole %)
ductivity of the fuel mixture NaF-ZrF.-UF. (50-
46-4 mole %, fuel 30). A preliminary value of Solid (100 to 502°C)
1.3 Btu/hr-ft-0F has been obtained that is to HT - W30oc = -4.57 + 0.146T +
be compared with the previous values that range
-5\ t-2
from 0.8 to 1.6 Btu/hr.ft«°F. This study will + (2.50 x 10-3) T
be completed after replacement of several de
fective components in the circuit that supplies c„P = 0.146 + (5.00 x 10-5) T
power to the apparatus. Liquid (551 to850°C)
HT - tf30oc = -6.67 + 0.278T
Enthalpy and Heat Capacity
- (3.93 x 10~5) T2
Final results have been obtained for the en
thalpies and heat capacities of the eutectic cp = 0.278 + (7.85 x lO-5) r
mixtures LiCI-SrCI2 (52-48 mole %) and LiCI-
BaCL (70-30 mole %). The wide discrepancies Fusion (511°C)
in enthalpy between duplicate samples reported
previously were resolved by making improve
HL - Hs = 49
ments in the technique for mixing the two dry
salts. The equations describing the enthalpies In the expressions above
and heat capacities of these two mixtures are
H = enthalpy in cal/g,
as follows:
c = heat capacity in cal/g»°C,
T = temperature in °C.
LiCI-SrCI2 (52-48 mole %)
To date the enthalpies, heat capacities, and
Solid (99 to 493°C)
heats of fusion have been determined for five
HT - fV30oc = -3.52 + 0.140T + chloride mixtures c ontaining lithium chloride.
In a manner analogous to that previously used
5\ -r2
+ (4.25 x 10-5) T for fluoride-containing mixtures, it has been
cp = 0.140 + (8.50 x 10-5) T possible to correlate these three properties for
the chloride salts, and the results for the five
Liquid (550 to 852° C) mixtures are presented on a gram-atom basis in
Table 1.7.5. It may be seen that the maximum
W30°C = ~39'8 + °-35^r deviation of any salt mixture from the average
-5\ t-2 for the five salts is 9%. The values per gram
- (8.16 x 10-5) T
are obtained by multiplying the_l^sted per-gram-
cp = 0.356 - (16.3 x 10-5) T atom quantities by the factor N/M, where N =
2^,-N^ and M- 2 x.M., x, being the mole fraction
Fusion (525°C estimated) of the zth component; N •, the number of atoms
in the zth component; and M., the molecular
HL ~ "S = 43 weight of the zth component.

2 1
W. D. Powers and G. C. Slalock, Enthalpies and
20W. D. Powers, ANP Quar. Prog. Rep. Sept. 30, 1957, Heat Capacities of Solid and Molten Fluoride Mixtures,
ORNL-2387, p 108. ORNL-1956 (Jan. 11, 1956).

62
 PERIOD ENDING DECEMBER 31, 1957

Table 1.7.5. Properties of Binary Eutectics with Lithium Chloride

Enth alpy Heat C opacity

per gram-atom per gra m-atom

Second LiCI - -

M N M/N
Component (mole %) Solid at Liquid at Solid at Liquid at per gram-atom
300°C 700°C 300°C 700°C

NaCI 72 46.89 2.00 23.45 6.78 8.23 1750 6760 2020

KCI 59 55.59 2.00 27.79 6.56 9.03 1860 6960 1830

CaCI2 62 68.46 2.38 28.77 6.76 8.31 1650 6560 1750

SrCI2 52 98.15 2.48 39.58 6.56 9.55 1670 6690 1700

BaCI2 70 92.16 2.30 40.07 6.45 8.94 T660 6770 1960

Average 6.62 8.81 1720 6750 1850

Maximum deviat on from average 3% 8% 8% 3% 9%

63
 Part 2

CHEMISTRY

W. R. Grimes
 2.1. PHASE EQUILIBRIUM STUDIES1
C. J. Barton R. E. Moore R. E. Thoma

SUMMARY OF SALT SYSTEM INVESTIGATIONS systems have frequently precluded otherwise

R. E. Thoma useful mixtures for fuels.

Comprehensive studies of phase equilibria in

fused-salt mixtures have been made over a period COMPARISON OF RUSSIAN AND ORNL
of years. These studies have been relevant to INVESTIGATIONS OF THE NaF-ThF4 SYSTEM
(1) selections of fused-salt fuels, solvents, and R. E. Thoma
fuel-enriching mixtures, (2) development of fuel
reprocessing techniques, (3) considerations of A search for low-melting fuels requires a
molten-salt secondary coolants, (4) structural knowledge of binary phase diagrams for alkali
metal corrosion, (5) crystal structure relationships, fluorides and quadrivalent fluorides. An example
and (6) basic investigations of the phenomena which affords a comparison of published Russian
involved in fused-salt phase equilibria. fused-salt data with the type of information de
A nearly current report of the status of fused-salt veloped for ANP fuels is the system NaF-ThF4.
phase equilibrium diagrams was published re
The Russian phase diagram of the NaF-ThF4
cently,2 and an extensive discussion of the system4 is presented in Fig. 2.1.1, and the results
equilibrium relationships in some of the more
of ORNL studies are given in Fig. 2.1.2. In the
ORNL investigations, because of the need for a
complex systems which have been investigated
will be available soon. Continued phase complete understanding of fuel behavior, all the
equilibrium studies have provided, during the
phase equilibria, including those below the
quarter, sufficient data for essentially complete solidus, received attention. In Fig. 2.1.1,
diagrams of the following systems: KF-UF , Emelyanov and Eustyukhin have interpreted
thermal analysis data to indicate that in the
RbF-UF4, NaF-HfF4, LiF-ThF4, NaF-ThF4,
NaF-LiF-BeF2. For the most part, reliable liquidus vicinity of 20 mole %ThF4 the incongruent melting
lines or surfaces are well known for the systems
temperature of 4NaF-ThF4 is 648°C, that the
that have been reported,2 but for many systems 4NaF>ThF4-2NaF-ThF4 eutectic temperature is
only tentative conclusions can be drawn regarding 620°C, that 4NaF»ThF4 decomposes into NaF and
the subliquidus phase relationships. 2NaF-ThF4 at 606°C, and that the a-2NaF-ThF4
The search for improved fuel mixtures for to (8-2NaF-ThF4 inversion occurs at 568°C. At
aircraft reactor use has been directed to areas of
ORNL, results of x-ray diffraction and optical
low-melting-point regions of fluoride salt diagrams. examination of samples from thermal-gradient
quenching experiments have shown that, for the
Characteristic features of most fluoride salt
above cases, no transition occurs at 648°C, that
diagrams that have strongly influenced the
selections of fuel systems have been the high the 4NaF-ThF4-2NaF-ThF4 eutectic temperature
melting temperatures (800 to 1000°C) of 3MF-XF4 is 606°C, and that 4NaF-ThF4 decomposes into
and 2MF«XF. compounds, where MF is an alkali NaF and 2NaF-ThF4 at 568°C.
fluoride and XF . is ZrF., UF., or ThF.. These Emelyanov and Eustyukhin have assigned equi
high-melting-point compounds and related high- librium polymorphism to the compound 2NaF-ThF4,
temperature liquidus compositions in ternary apparently wholly on the basis of the Zachariasen5
indication that two crystal modifications of this
compound exist. ORNL data show that there are
The petrographic examinations reported here were two forms of 2NaF«ThF4; however, the hexagonal
performed by T. N. McVay, Consultant, and R. A.
Strehlow, Chemistry Division. The x-ray examinations 8 form is either metastable or has equilibrium
were performed by R. E. Thoma, Chemistry Division. existence below 400°C.
2
R. E. Thoma and W. R. Grimes, Phase Equilibrium
Diagrams for Fused Salt Systems, ORNL-2295 (June 24, V. S. Emelyanov and A. I. Eustyukhin, Alomnaya
1957). Energ. 1, No. 5, 80 (1956).
3
J. E. Ricci, Guide to the Phase Diagrams of the 5W. H. Zachariasen, ;. Am. Chem. Soc. 70, 2147-51
Fluoride Systems, ORNL-2396 (to be published). (1948).

67
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR-DWG 27033

HCCJIE30BAHHB IHCTKM PU'IIJIABilElIIIblX C0JIF.I1 HA OCHOBE *TOPHJIA TOPHH

ThF4. 3to VKanhiBaeT, mto npH 606° (; co- 830° (', npoHfxonjir iit'pnTeKTn<je( kbh peaK-
eAHHt'HBc Na,ThF, pacnaAacTCH, o6pa:ivn HHfl p83Jio>kphhh xHMaiecKoro coenaHCHBfl
NaF h coeflBHeHiip Na.ThF,. nP« 568° C NaTh,F, (66,7 moji. % ThF4), coct«b Koroporo
npoMcxoflHT nojiRMopdpHoe iipoBpamcHMc iioATBcpwflaeTCft, b lacTHoCTii, reM, «rro Ten-
a—-^-Na.ThF,. CotanHPHiie .Na.ThF, (b .ioboh a<Jx|>pKT npH 830° hmppt MaKCRMajn»Hy»
TOiKe c 33,3 moji. % ThF4) njiaBHTCH e ot bpjihihhv b toikp okojio 67 moji.% ThF4
KphlTWM MaKCHM vmom npH TPMnepaType 708° ( '.. H Bt)JIH3H 3T0K0 COCTaBB RC*P3aeT S^dppKT miaR-
•)BTeKTH4ecKaw rMPCb cocahhc'hhh Na4ThF, II JldlHH 3BTPKTHKH npH 750° C.
Na,ThF, naaBHTCH npn 620c C h ro,iep>«HT TaKHM ofipaaoM, b chctpmp NaF ThF4
24 moji.% ThF,. MMPPTCH ICThlpP XHMH<tPCKHX COejtBMWM:

piaJhFA NaThF,*HaThtF,
NaThF, J NaThtF,
500 mw
NaF*3Ha,ThFt
moa %ThFt
10 20 30 10 50 SO 70 80 90 100%
Put. 1. Cn(Teiin NaF-Tliry

flaJiee jihhhh JiHKBHflyra hoctpiu-hho uo- Na4ThK„ Na,lhF4, NaThF, h NaTh,Ft. Co-
HHJKaCTCH BO C^PAVIomeH rtBTPKTHHPCKOH TOHKH e.iHHpHMP Na.ThF, cymecTBypT b flByx MOfunfa-
npw 700° C h 36--37 moji.% ThF4. HaqnH«H khuhhx: a,-Na,ThF,, 3-Na,ThF,, •, *ob*ah-
c 36,5 moji.% ThF4 h abjipp, tpiijioboh adp- momv, b pacnyiaBJiPHHOM cottohhhb; coe^M-
<J>ckt nJiaBJiPHHfl sbtckthkh YMPHbinapTCH HeHHH Na4ThF, h NaTh,F, pacnaaaiOTi'a no
h ncie3aPT okojio 50 moji.% ThF4. nPpHTPKTM4PPKHM pCaKUHHM. 3tM UHIIK
Ilo COBOKynHOCTH ABHHNX TepMHMPCKOro nOflTBPpH<flaH>T CTpVKTypHUe RCCJleAOBaRHR
H MHTKpOCTpyKTypHOrO aR8JIH3a CJIPflyPT, IT" HaxapHa3PHa [1]. (xx-aHHpHHe npeAHonoMCB-
B 06jI8CTH OKOJIO 50 MOJI.% ThF4 rylllpCTByPT TpjihHoro cocTaBa NaThF6 HMeer b ofj/iacnt
XHMHHeCKOP rOCAHHCHHe IipOanOJIO>KHTPJII,H()r() ot 44 ;io 50 moji.% ThF4 ohphi. noaorwi
rocTaBa NaThFj, HMewmpp Beci>Ma iioaoriiii MahTHMVM H8 JIHHHH JIHKBHAVI'a, no KOTOpOMJ
MaKCBMVM. . IhHHH JIHKBHflVOa OT OOPAHHeHHH TpVAHO VCTaHOBHTb TOIHUH COCTBB.
NaThF, HesHaiHTpjibHO noHnwaPTCH flo 3btpk-
ikh npn 750° C h 53,5 moji. % ThF4 h ^aaec HccjieaoBaHBe chct«mu Kf —ThF4
pc3KO noAHHxaPTCH BBepx, npcTppiieBan npn Bwjio R3roTOBaPHo h HiT/ienoBaHO [5], [6]
830° C r 62 moji.% ThF4 h3hom, nocjie koto- cBtime 40 cmiaBOB Mppes HRrepBaJi 2—3moji.%
poro 0Ha naaBHO nooHRMacTOi no 1065° C ThF4. Ha ocHOBaHHH amx HCcjienoBaHBB riujta
(TpMnepaTypa iuiaBJiPHHH HHCToro ThF4). Hpn nocTpopHa ARarpaMMa poctorhrh chctpmu

'i .ATOMUaH dH^prHH. V ."i ^1

Fig. 2.1.1. The System NaF-ThF. According to Emelyanov and Eustyukhin.

68
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 18964
HOO

IOOO

900

£ 800

700

600

/
/
11?
.c
,,*•
500
.c 1-
1-
o \- OJ
Ll_
o Li_ Li.
o O
2
z z
CO

400
NaF 10 20 30 40 50 60 70 80 90 ThF„

Fig. 2.1.2. ORNL Data on the NaF-ThF. System.

The Russian diagram shows a congruently melting indicate the use of the thermal-gradient quenching
compound NaF«ThF4. ORNL data show that the techniques or of routine petrographic examinations.
compound NaF-Thr4 melts incongruently to
NaF«2ThF . and liquid.
THE SYSTEM NaF-CrF2
Unclassified Russian reports presently represent H. A. Friedman C. F. Weaver
the predominant number of contributions to the Phase equilibria in the NaF—structural metal
literature of phase equilibrium studies of fused fluoride systems NaF-CrF2, NaF-FeF2, and
salt mixtures. Since 1946 more than 200 Russian NaF-NiF2 have been studied in an effort to
reports have been issued' concerning the phase elucidate the corrosion process as it is related to
equilibria of oxide, silicate, chloride, nitrate, and these materials. Phase diagrams for the systems
fluoride systems. NaF-FeF2 and NaF-NiF2 have been reported.7
Most Russian phase work is done with thermal No fused-salt melts from thermal analysis experi
analysis and differential thermal analysis tech ments or from thermal-gradient quenching experi
niques. Phase identification is accomplished for ments have produced pure NaF-CrF2 samples that
the most part with x-ray diffraction. No reports are free of Cr+++. A variety of means has been

A. G. Bergman, K. A. Evdokimova, and 0. F. Bogush,

Izvesl. Sektora Fiz.-Khim. Anal., Inst. Obshchei Neorg. H. A. Friedman and B. S. Landau, ANP Quar. Prog.
Kbim., Akad. Nauk S.S.S.R. 27, 419-56 (1956). Rep. June 30, 1957, ORNL-2340, p 128.

69
ANP PROJECT PROGRESS REPORT

employed to minimize the effects of the dispro to be quenched is flattened and crimped at k-in.
portionation reaction intervals. The sample tubes are sealed by welding
after they have been rolled and crimped. Sample
3CrF 2^2CrF3 + Cr° tubes prepared in this manner have better heat
transfer characteristics than those of the tubes
in these phase equilibrium studies. Even the
inclusion of 20 to 30 wt % chromium metal powder used previously. Much less expansion of the tubes
dispersed through samples for thermal-gradient during heating occurs because the void space is
quenching has not suppressed the forward reaction also reduced in the crimping process.
sufficiently for NaF-CrF. phase equilibrium data PETROGRAPHIC ANALYSES OF FUEL
to be derived. MIXTURES

R. A. Strehlow
APPARATUS FOR HIGH-TEMPERATURE
THERMAL-GRADIENT QUENCHING
The petrographic microscope has been used on a
EXPERIMENTS
routine basis for the inspection of solidified fuel
mixtures.'0 This type of examination is performed
H. A. Friedman to detect the presence of oxygen-containing com
A design modification of the apparatus8,9 used pounds and, for some samples, to determine any
in thermal-gradient quenching experiments has variance from the nominal composition. Mixtures
permitted the feasible upper operating temperature of NaF-ZrF4, NaF-ZrF4-UF4, LiF-NaF-KF-UF4,
limit to be raised from 850 to 1300°C. Previous and re fated compositions have formed the bulk of
difficulties encountered with quenching apparatus samples submitted for inspection. An illustration
when operated routinely at 900 to 1000°C included of the results obtained with the technique used in
the frequent need for replacement of the 18 thermo the inspection of these samples is given here for
couples which monitored the thermal gradient, the NaF-ZrF4 mixtures containing about 50% ZrF4.
rupture of the sample tube because of oxidation Mixtures in the NaF-ZrF4 system which contain
corrosion, uncertainties as to temperature constancy between 46.2 and 57.1 mole % ZrF4 will contain
of the sample because of discontinuous monitoring, some 3NaF-4ZrF4 which will, in general, occur
and sample tube expansion during heating. in the slowly cooled melt and will be dispersed in
In the revised apparatus a single, traveling, eutectic aggregates that also contain 7NaF»6ZrF4.
recording thermocouple was substituted for the The aggregate frequently will exhibit a brown cast,
18 monitoring thermocouples. The furnace was which has been attributed to hydration of
redesigned to permit the sample tubes to be 3NaF'4ZrF4. Better crystallized 3NaF-4ZrF4 has
protected by a static helium atmosphere during been observed in samples containing about
heating. The use of a roller crimping machine 56 mole % ZrF4. The material present as the
contributes to minimizing sample tube expansion. matrix in the eutectic aggregates is the compound
The apparatus, shown schematically in Fig. 2.1.3, with the reported formula of 7NaF«6ZrF4.
now consists of a round nickel block 10 in. in This compound occurs as the primary phase in
length and 2 in. in diameter, which is supported in mixtures containing from 46.2 to 49 mole % ZrF4.
a 2 k-in. tube furnace that has a platinum wound It is characterized as a crystalline uniaxial
core. Interchangeable gears allow a choice of substance that is optically negative with NQ = 1.502
excursion times for the traveling thermocouple. and Ng = 1.508, where NQ and Ne refer to the
A complete thermocouple travel cycle of 2 hr is ordinary and extraordinary indices of refraction of
usually employed in using this apparatus. uniaxial crystals.
The roller and crimping machine (Fig. 2.1.4) Compositions containing about 50 mole %ZrF4
is designed so that a 6-in.-long, O.lO-in.-OD, which have been cooled at intermediate rates
0.010-in.-wall nickel tube containing a sample occasionally contain NaF-ZrF4 as a nonequilibrium
phase that is uniaxial positive with No = 1.417 and
C. J. Barton et al., ANP Quar. Prog. Rep. Dec. 10,
1954. ORNL-1816, p 57.
Q
C. J. Barton et al., "Phase Equilibria in the Systems
The techniques described were developed over a
LiF-UF4 and NaF-UF.," J. Am. Ceram. Soc. (in press period of several years by T. N. McVay, G. D. White,
for issuance in Feb. 1958), H. Insley, and B. S. Landau.

70
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 25153

-TO CONTROLLER
BLOCK SUPPORT

He LINE

TO RECORDER

Fig. 2.1.3. Thermal-Gradient Quenching Apparatus.

71
 PERIOD ENDING DECEMBER 31, 7957

liquid in equilibrium with a primary phase as a and the curve is valid, of course, only at point P.
function of temperature and over-all composition. Hence a series of such curves must be constructed
to suitably cover the primary phase area of
interest.
Index of Refraction vs Solid Solution Composition
If a series of fused-salt solid solutions of known Constructing Tie Lines and Equilibrium Paths
composition are prepared, their indices of refraction If the over-all composition is considered to be
can be measured by means of a microscope and that shown in Fig. 2.1.7 and it is assumed that
suitable immersion oils. A hypothetical plot of Fig. 2.1.6 gives the solid solution composition
index of refraction vs composition for a continuous
fused-salt solid solution between A and B is given
in Fig. 2.1.5. UNCLASSIFIED
ORNL-LR-DWG 25264

UNCLASSIFIED
ORNL-LR-DWG 25265

O
<
~e~r.
LJ
r£ UJ
>
a.
U_ M _«-=• -""T or
r-
o "r U <
>-
X
K
< a
a
Z>
5 z
a
o
_ . _ .
m
—

3 2

TEMPERATURE
t\ C OMPO SITIO g E

Fig. 2.1.6. Composition of AB vs Temperature for

Fig. 2.1.5. Hypothetical Plot of Index of Refraction Point P.
vs Composition for a Continuous Fused-Salt Solid Solu
tion Between A and B.
UNCLASSIFIED
ORNL-LR-DWG 25263

In general, when the solid solution AB ap

pears in the system of interest its composition
is not immediately known. It is readily apparent
that the composition can be determined from
Fig. 2.1.5, if the indices of refraction for the
solid solution AB are measured. In order to
avoid ambiguity, plots such as Fig. 2.1.5 must
contain no maximums or minimums. Further it
should be noted that the accuracy of the de
termination increases with the slope of the curve
of index of refraction vs composition of the solid
solution.

Composition as a Function of Temperature

The composition of the solid solution AB must
be determined, as described above, for a series of
temperatures at a given over-all composition P in
order to obtain the information needed for con
structing the equilibrium paths. The composition Fig. 2.1.7. Over-All Composition of Ternary System
of AB is plotted vs temperature, as in Fig. 2.1.6; ABC.

73
ANP PROJECT PROGRESS REPORT

vs temperature plot for P, at temperature 1 the LiF-4UF4-LiF.4ThF4,and7LiF.6UF4-7LiF-6ThF4.

composition of AB solid solution is a, the compo In the UF4-ThF4 sojid solution the plot of index
sition of the liquid is P, and the tie line is Pa. of refraction vs composition has a small slope.
Likewise for temperatures 2, 3, and 4, the compo As previously mentioned, tie lines cannot be
sition of AB solid solution is b, c, and d, the accurately determined when the slope of the plot
composition of the liquid is b', c' and d', and is small. In the remaining two solid solutions,
the tie lines are b'b, c'c, and d'd, respectively. however, the slopes are sufficient for the de
The locus of all points "such as P, b', c' and d' termination of tie lines. These solid solutions
is the equilibrium path. offer ample opportunity for a study of the tie lines
and equilibrium paths, and such studies have been
The System LiF-ThF4-UF4 initiated.

The system LiF-ThF4-UF4 appears to contain

three continuous solid solutions: UF4-ThF4,

74
 PERIOD ENDING DECEMBER 31, 7957

2.2. CHEMICAL REACTIONS IN MOLTEN SALTS

F. F. Blankenship L. G. Overholser
R. F. Newton G. M. Watson

THE PREPARATION OF UFj AND ITS STABILITY where T is in degrees Kelvin. The equilibrium
IN MOLTEN FLUORIDE MIXTURES values thus obtained indicated that crystalline
F. F. Blankenship UF, is more stable than published estimates of
the free energies of formation would indicate. In
The use of UF, in fuel mixtures was found to be contrast, UF, was found to disproportionate in
impractical because disproportionation occurred molten fluoride solvents to a much greater extent
and uranium metal deposited on the container than it did in the solid-crystal form. Partially
surfaces. The work that led to this conclusion and complete phase diagrams were prepared6 that
the results of investigations of methods for elimi were useful in approximating the solubility be
nating or avoiding disproportionation are sum havior of UF, in binary mixtures of alkali fluorides.
marized here. The mechanism for disproportionation The solubility of UF, in NaF-ZrF4 and in NaF-
has been ascertained and means for controlling it BeF2, which are mixtures of interest as reactor
are suggested. Petrographic and x-ray diffraction fuels, is 1 to 2 wt %U at 600°C.7
examinations were used in this study to characterize Nonacidic fuel solvents that contain a high con
phases in solidified melts, and liquid compositions centration of fluoride donor constituents, such as
were, in general, defined by synthesis or by fil alkali fluorides, were found to increase the ex
tration and chemical analysis. ' tent of disproportionation by complexing the acidic
The UF, compound may be prepared in solution UF4 (fluoride acceptor). The ease of alloy for
by various redox reactions or by electrolysis, but, mation between the uranium formed and the con
when large quantities were needed for corrosion tainer metal also noticeably influences the extent
studies, solid UF4 was used as the starting of the disproportionation. Less disproportionation
material. The reaction occurred in copper vessels than in nickel vessels,
and tantalum and molybdenum were found to be
0) 3UF4 + U°; 4UF, even more resistant than copper to alloying with
the uranium formed.
was used to produce routinely kilogram batches It was also found that UF, was oxidized by the
of 99% UF3. Essentially complete reaction of alkali fluorides. The reaction was particularly
stoichiometric mixtures of UF4 and U°was obtained noticeable with mixtures containing KF:
in 48 hr at 900°C.4
The equilibrium pressure of UF4 above UF3 (2) UF3 + KF UF4 + Kc
as a result of disproportionation by the reverse
reaction in Eq. 1 was measured at temperatures -or exam pie, when a solution containing 21.4
between 1270 and 1400°C in tantalum-lined vapor wt % U as UF3 in the LiF-NaF-KF eutectic (46.5-
pressure apparatus. The results obtained are 11.5-42 mole %) was heated under helium in copper
represented by the expression apparatus, about one-third of the UF, was iost by
disproportionation in 2 hr at either 650 or 750°C;
4187 the amount of UF, oxidized by KF at the same time
log p (mm Hg) = + 3.945 ,
increased from 7 to 18%. Sufficient reduction in
the helium pressure caused a rapid increase in the
H. Ins ley et al., Optical Properties and X-Ray Dif
fraction Data for Some Inorganic Fluoride and Chloride
Compounds, ORNL-2192 (Oct. 23, 1956).
2R. J. Sheil and B. H. Clampitt, ANP Quar. Prog. Rep. 6C. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10,
March 10, 1955, ORNL-1864, p 53. 1954, ORNL-1771, p 56.
3M. B. Panish, ANP Quar. Prog. Rep. June 10, 1955, 7L. M. Bratcher et al.. ANP Quar. Prog. Rep. March 10,
1955, ORNL-1864, p 50.
ORNL-1896, p 69.
4B. J. Sturm and E. E. Ketchen, ANP Quar. Prog. Rep. 8C. J. Barton and B. H. Clampitt, ANP Quar. Prog.
March 10. 1955, ORNL-1864, p 66. Rep. Sept. 10, 1955, ORNL-1947, p 78.
5S. Longer, ANP Quar. Prog. Rep. June 10, 1955, 9B. H. Clampitt, ANP Quar. Prog. Rep. March 10,
ORNL-1896, p64. 1956, ORNL-2061, p 97.

75
ANP PROJECT PROGRESS REPORT

rate of disappearance of U . The sharp in the standard free energies of formation of pure
crease in the rate occurred at a pressure corre solid and supercooled FeF2 at these temperatures.
sponding to the bubble pressure of potassium. The Accordingly, the activity coefficients for FeF2 in
small apparent temperature coefficient of the dis this solvent are near unity, regardless of the choice
proportionation in such experiments was probably of standard state. The numerical values of the
the result of the process being limited, after an activity coefficients of FeF2 are, at 800, 700, and
initial rapid deposition of uranium, by the slow 600°C, 1.46, 2.20, and 3.28 with respect to the
rate of diffusion of uranium into copper. pure crystalline solid, 0.64, 0.66, and 0.66 with
In comparison with UF4-containing fuels, the respect to liquid FeF2, and unity with respect to
UF3-containing fuels exhibit a marked improvement the standard state of reference at infinite dilution.
in initial corrosion attack at high temperatures. Equilibrium quotients for the Ni F2 reduction
Moreover, the use of UF3 in conjunction with UF4 have been determined to be 21,900 ± 2,000,
could provide reserve reducing capacity to com 15,300 ± 700, 11,000 + 600, and 7,600 ±400 atm,
pensate for any tendency of the fuel to become as compared with the values 10.9, 7.4, 4.8, and 3.1
more oxidizing during burnup. Unfortunately the calculated from the free energies of formation of
difficulties arising from the deposition of uranium pure solid NiF2 at 625, 600, 575, and 550°C,
by disproportionation prevent the easy exploitation respectively. The corresponding constants calcu
of UF3 as a solution to corrosion problems. How lated for liquid NiF- are 43.8, 33.6, 25.2, and
ever, even in the UF4 fuels, some UF3 always 18.4. The very large discrepancies between the
appears as a product of the oxidizing action of calculated and experimental constants give rise to
UF4, and the fate of this UF3 is of interest in very high values of the activity coefficients of
connection with long reactor lives. NiF2. By using the solid as the standard state
the numerical values of the activity coefficients
ACTIVITY COEFFICIENTS OF FeF2 AND OF were found to be 2009, 2076, 2292, and 2468 at
NiF2 IN MOLTEN NaF-ZrF4 625, 600, 575, and 550°C, respectively. With
C. M. Blood supercooled liquid NiF2 as the standard state, the
The activity coefficients ' of FeF, and of corresponding values of the activity coefficients
are 500, 455, 436, and 413.
NiF2 in NaF-ZrF4 (53-47 mole %) were determined
It is clear from the values of the activity co
by comparing the values of the equilibrium quotients
of the reaction
efficients that when NiF2 is dissolved in NaF-
ZrF4 (53-47 mole %) it behaves in a manner vastly
MF2(d) + H2(g) —M°(s) + 2HF(g) different from that predicted by the free energies
of formation for solid NiF2. This is somewhat
with the corresponding values of equilibrium con surprising in view of the agreement shown to exist
stants calculated from the tabulated free energies between the calculated and experimental behavior
of formation of FeF2, NiF2, and HF. Experimental of FeF2 in the same solvent. Accordingly, the
values for the equilibrium auotients for the re results of these investigations confirm that the
actions of FeF2 and of NiF2 were shown to be in tabulated values of the free energies of formation
dependent of concentration. For the FeF2 reaction of solid FeF2, the tabulated melting point, and
the values of Kx are 2.6 ± 0.2, 0.63 ± 0.04, and the estimated heat of fusion are adequate to pre
0.097 + 0.007 atm at 800, 700, and 600°C, re dict the behavior of FeF2 in the solvent NaF-ZrF4
spectively. These values are in essential agree (53-47 mole %). On the other hand the corresponding
ment with equilibrium constants calculated from values for NiF2 are not at all suitable to describe
its behavior in the same solvent.
In the case of FeF2 the experimentally determined
J. H. DeVan and E. A. Kovacevich, ANP Quar. free energy of formation at 600°C for the standard
Prog. Rep. Dec. 10. 1955. ORNL-2012, p 107. state of reference at infinite dilution was —135.4
1]C. M. Blood and G. M. Watson, ANP Quar. Prog. kcal/mole, as compared with the tabulated values
Rep. March 10. 1956, ORNL-2061, p 84.
-137.5 and -134.5 for the solid and liquid states,
12C. M. Blood, ANP Quar. Prog. Rep. Sept. 30. 1957. respectively. Thus the experimental value for
ORNL-2387,-p 117.
13
L. Brewer et al.. Natl. Nuclear Energy Ser. Div. IV FeF2 was found to be between the predicted
19B, pp 65, 110 (1950). values for the solid and the liquid states. On the

76
 PERIOD ENDING DECEMBER 31, 7957

other hand, the experimentally determined free These values give a free energy of formation for
energy of formation of the standard state of ref solid NiF2 of —117.2 kcal/mole. However, some
erence for NiF2 was found, at 600°C, to be -114.6 recent work indicates that the melting point of
kcal/mole, while the tabulated value for the solid NiF2 is greater than 1500°C (ref 16). If the en
state is -127.9. As has been pointed out else tropy of fusion for FeF2 (8000/1375 = 5.82 eu) is
where, at first glance it appears that the value assumed for NiF,, but a melting point of 1800°K is
of the free energy of formation of the solid NiF2 used, a value of -120.0 kcal/mole is obtained as a
is in error by approximately 13 kcal. It is known, reasonable estimate of the free energy of formation
however, that the solubility of NiF2 is quite low in of solid NiF2 at600°C.
in this solvent,15 0.17 mole % at 600°C; and, if
the saturating phase is pure NiF2, an activity co
SOLUBILITIES AND ACTIVITY COEFFICIENTS
efficient of 588 would be required to give unit
activity at saturation. Since the saturating phase OF NiF2 AS FUNCTIONS OF SOLVENT
COMPOSITION IN NaF-ZrF4 MIXTURES
is not pure NiF2 but rather a complex fluoride, it
is only possible to predict from the solubility an C. M. Blood
upper limit of 588 for the activity coefficient for It has been postulated '' that the activity co
solid NiF2 as the standard state. The upper limit efficients of a solute in a fluoride solvent will
of 588 corresponds to a value of -125.7 for the reach a maximum at the isoelectric point or when
free energy of formation of NiFj. Conversely, it the charge-to-distance ratio of the solute's positive
may be stated that the upper limit at 600°C of the ion is equal to the average or effective charge-to-
free energy of formation for solid NiF2 is -125.7 distance (Z/R) ratio of the positive ions of the
kcal/mole, as given by its solubility, the experi solvent mixture. As the composition of the solvent
mental Kx, and the tabulated free energy of for is. changed and its charge-to-distance ratio is made
mation for gaseous HF. The magnitude of a more to differ from that of the solute, negative deviations
reasonable value will now be considered. from ideality become greater. On this basis a
In order to obtain an activity coefficient in the tentative acid-base scale for predicting isoelectric
neighborhood of unity with supercooled liquid points of solutes in NaF-ZrF"4 melts was pre
NiF2 as the standard state, the free energy of pared. Since not much experimental evidence is
formation must be approximately —114.6 kcal/mole available on the effect of solvent composition on
at 600°C. The free energy of formation of solid solute activity coefficients, the proposed plot is
NiF2 at 600°C, based on an assumed free energy only tentative, since no proven way exists to ob
of formation of liquid NiF2 of -114.6, may be tain the effective Z/R of the solvent as a function
computed according to the equation of composition. In an effort to provide quantitative
information, hydrogen-reduction equilibrium quo
AFNiF2(s) = AFNiF2(/) _ AFMelting ' tients of NiF- were measured in various NaF-ZrF4
mixtures. Activity coefficients were calculated,
If the heat of fusion is assumed to be independent but, since no satisfactory value of the free energy
of temperature, then of formation of either solid or liquid NiF2 is avail
able, the data are presented (Table 2.2.1) as the
AFNiF2(,) = AF°NiF2(/)- ^m(TM ~ T) ' ratio of the activity coefficient of NiF- in the
solvent indicated to the activity coefficient of
where ASM = AW JTM, the entropy of fusion, and NiF2 in NaF-ZrF4 (53-47 mole %). In addition,
TM is the melting point of solid NiF2. the solubility data are given as a function of
In previous calculations, a heat of fusion of solvent composition.
8000 cal/mole and a melting point of 1300°K for
NiF, were assumed, as suggested by Brewer.
H. A. Friedman, private communication to C. M. Blood.
14M. Blander and F. F. Blankenship, ANP Quar. Prog. 17R. F. Newton and F. F. Blankenship, ANP Quar.
Rep. Sept. 30. 1957, ORNL-2387, p 121. Prog. Rep. March 31. 1957. ORNL-2274, p 123.
15C. M. Blood, ANP Quar. Prog. Rep. June 10, 1956, 18D. G. Hill and F. F. Blankenship, ANP Quar. Prog.
ORNL-2106, p 99. Rep. March 31. 1957, ORNL-2274, p 126.

77
ANP PROJECT PROGRESS REPORT

Table 2.2.1. Equilibrium Quotients, Activity Coefficient reaction mediums is presented here, and data for
Ratios, and Solubilities of NiF, in NaF-ZrF4 the reaction of UF4 with various other metals are
Mixtures at 600°C given in the next section. Some data are also given
for the reaction of UF4 with Inconel, in which the
Solvent Activity activity of chromium is much less than unity.
Equil ibrium Solubility
Composition Coefficient
Quotient (ppm Ni) As might be expected, the reaction of UF4 with
(mole % ZrF4) Ratio*
chromium is influenced by the reaction medium
employed. The reaction which occurs in the
xlO3 ZrF4- and BeF2-bearing melts studied may be
57 4.6 0.30 450 written as

52 9.2 0.60 540

2UF4 + Cr°; 2UF3 + CrF2
47 15.3 1.00 980

44 18.5 1.21 1380

The valency of the chromium compound has been
40 25.3 1.65
established through numerous experiments, and
it has also been demonstrated experimentally that
*Based on activity coefficient of NiF, in NaF-ZrF4 the solubility of the uranium and chromium com
(53-47 mole %). pounds involved is considerably greater than the
concentrations encountered in these studies. The
The experimental measurements of the 40 mole % equilibrium quotient, K , defined as
ZrF4 mixture have not been completed, and the
data given in Table 2.2.1 for this composition may N
UF.
-N
CrF,
be revised. It is proposed to make measurements K =
on the solvent containing 37 mole % ZrF4 in an N-
UF.
effort to locate the positions of the maximums of
the activity coefficient and solubility. The present
results indicate that the effective Z/R average of where N denotes mole fraction, may be evaluated
the solvent may correspond closely to that predicted for this system by adding a known concentration of
by the proposed acid-base scale. It may also UF4 to chromium and determining the resulting con
be noted that the value of the activity coefficient centration of Cr . Side reactions that arise from
of the NiF2 solute is not dependent on the solu impurities (H20, HF, Fe++, Ni +'t) must be eliminated
bility; this confirms a previous conclusion that the or suppressed to such a degree that the chromium
precipitating phase is not NiF2. blank is small compared with the total chromium
measured. The blanks in these studies ranged from
REDUCTION OF UF4 BY CHROMIUM 150 to 250 ppm Cr.
IN MOLTEN FLUORIDES
A filtration method was useg1 in which the melt
J. D. Redman was equilibrated under designated conditions and a
The possible use of Inconel or stainless steel as sample was withdrawn through a sintered-nickel
a material for containing UF4-bearing fluoride filter and analyzed for chromium. Nickel apparatus
mixtures raised numerous questions as to the was used in most cases, because early experi
effect of corrosion attack in the temperature range mental work demonstrated that nickel is stable
600 to 800°C. The most active metal present in toward UF4 dissolved in fluoride mixtures. Equi
these alloys is chromium, and consequently any librium is attained in these systems fairly rapidly,
fundamental information relating to the interaction 3 hr being sufficient in virtually all cases, al
of chromium with UF4 dissolved in representative though a longer period of heating was utilized for
fluoride mixtures should be applicable to an under some studies.
standing of the corrosion of the alloys by the The data given in Table 2.2.2 demonstrate that
fluoride melts. the equilibrium quotient, K , is essentially con
A summary of the results of studies of the re stant over the range of concentrations of UF. and
duction of UF4 by chromium in various fluoride UF3 studied with NaF-ZrF4 (50-50 mole %) as the

78
 PERIOD ENDING DECEMBER 31. 7957

Table 2.Z2. Equilibrium Quotients for the Reaction 2UF4(rf) + Cr(c) ^±2UF3{d) + CrFj(rf)
in NaF-ZrF4 (50-50 mole %) at 600 and 800°C
Cond itions of Original Equilibrium
Equilibrium
Equi libration C sncentration Chromium
(mole %) Concentration
Temperature Time"
K
X
(°C) (hr) (ppm)
UF4 UF3

xlO-4
600 3 or 5 4.1 2220 + 120 3.2 + 0.8

5 3.0 0.82 950 4

5 2.0 1.7 155 3

5 0.89 2.5 30 5

800 3 or 5 4.1 2550 ± 210 5.9 ± 1.8

5 3.2 0.92 1120 6

5 2.2 1.9 320 6

5 1.3 2.8 60 6

5 2.0 1450 4

♦Standard deviations were obtained from 10 measurements at 600°C and 13 measurements at 800 C. The other
values for K were obtained from the arithmetic mean of two to four measurements.
x

reaction medium. As anticipated, the chromium con Table 2.2.3. Equilibrium Chromium Concentrations for
centrations decrease as the UF3 concentration in the Reaction of Chromium with UF4 in NaF-LiF-KF
creases; this suggests that the extent of attack on (11.5-46.5-42 mole %) at 600 and at 800°C
chromium or chromium-bearing alloys may be de
creased by the addition of UF,. The reaction has
a small positive temperature coefficient; an average Conditions of Oris inal Equi ibrium
heat of reaction of 6 kcal in the temperature range Equilibrat on Concentration Conce ntration*
600 to 800°C is obtained from the values for Kx. Temperature Time
(mole %) U Cr
Values of 70 and 20 were obtained for the equi
(°C) (hr) UF4 UF3
(wt %) (ppm)
librium constants at 600 and at 800°C, respectively,
by using the available values of the free energies 12.0 1130
600 3 2.30
of formation for CrF2, UF., and UF3. The corre
5 2.30 11.7 1050
sponding equilibrium concentrations were 8100 and
7300 ppm. A comparison of these values with 5 2.00 0.24 11.0 205
those given in Table 2.2.2 indicates that this 12 2.00 0.24 11.4 250
system behaves in a manner that is far from ideal.
When dealing with melts in the NaF-LiF-KF 800 3 2.30 11.2 2650

system, the reaction of UF4 with Cr° cannot be 5 2.30 11.5 2750
precisely defined. A mixture of Cr and Cr
5 2.00 0.24 11.1 1890
results, but the ratio of the two ions has not been
accurately determined. Equilibrium quotients 12 2.00 0.24 10.8 1750
cannot, therefore, be evaluated, but, rather, the 5 1.10 1.20 10.2 140
data are presented in terms of total dissolved
5 0.70 1.60 10.8 40
chromium (Cr++ + Cr+++). The data obtained with
NaF-LiF-KF as the reaction medium are summarized
in Table 2.2.3. Again it may be observed that the ♦Arithmetic mean of two to six measurements.

79
ANP PROJECT PROGRESS REPORT

chromium concentrations decrease as the UF3 Similarly the data given in Table 2.2.3 may be
concentration increases. The large difference in used to calculate the equilibrium chromium con
chromium concentrations at 600 and at 800°C in centration in the NaF-LiF-KF-UF4 melt at 800°C
dicates that the heat of reaction in this medium is when Inconel is the container material. Values of
much greater than that calculated for the reaction 1250 and 1500 ppm were obtained based on Cr++
in NaF-ZrF4 and suggests that the mass transfer and Cr , respectively, and the equilibrium con
of chromium is more severe in the former case. centration is probably intermediate between the
The chromium concentrations for equilibrium with two values. Since these concentrations are greater
Inconel can be calculated by using an activity than the value of 1100 ppm found for equilibrium
value of 0.1 for chromium in the alloy and the with pure chromium at 600°C, Inconel in contact
chromium concentration values given in Table with the NaF-LiF-KF-UF4 melt will support a
2.2.2. The values thus obtained are 1050 and higher concentration of Cr++ + Cr+++ in equilibrium
1150 ppm at 600 and at 800°C, respectively, in at 800°C than pure chromium can support at 600°C.
the reaction medium NaF-ZrF4-UF4. These con Thus chromium would be removed from Inconel in
centrations are much lower than that given for the hot zone and deposited as essentially pure
equilibrium with pure chromium at 600°C and chromium in the cold zone. Since no diffusion
indicate that it should not be possible to deposit process is required in the cold zone, the rate of
chromium at 600°C when NaF-ZrF.-UF. melts of attack in the hot zone would be controlled by the
this general composition are circulated in Inconel rate of diffusion of chromium to the metal-salt
loops with hot- and cold-leg temperatures of 600 interface. The chromium concentrations that re
and 800°C, respectively. Mass transfer of chromium sult when Inconel is contacted with UF. dissolved
ft 4
would occur in loops of pure chromium under these in either the NaF-ZrF4 or NaF-LiF-KF mixture
conditions, and would occur in an Inconel loop if are given in Table 2.2.4.
the deposited chromium were able to diffuse rapidly It may be noted that the chromium concentrations
enough into the cold-zone material. given in Table 2.2.4 are, in general, lower than
Table 2.2.4. Chromium Concentrations Resulting'from the Reaction of Inconel with UF. in Molten Fluorides

Conditions of Original Final

Reaction
Equilibration Concentrations
Chromium
Medium Temperature Time
(mole %) Concentration

(°C) (hr) UF, UF, (ppm)

NaF-ZrF4 600 12 4.1 135

(50-50 mole %)
5 3.1 1.0 35

800 3 4.1 360

5 4.1 610

12 4.1 840

5 3.1 1.0 210

NaF-LiF-KF 600 5 2.3 350

(11.5-46.5-42 mole %)
12 2.3 310

5 2.0 0.24 30

12 2.0 0.24 30

800 5 2.3 680

12 2.3 900

24 2.3 820

5 2.0 0.24 480

12 2.0 0.24 700

80
 PERIOD ENDING DECEMBER 31, 1957

those for reaction with pure chromium. If 700 ppm decrease in K upon changing from lithium to
is taken for the chromium concentration in the rubidium. This behavior is attributed to the change
NaF-ZrF4-UF4 melt at 800°C and 0,1 is taken for in the cation charge-to-separation ratio, Z/R, and
the activity of chromium in Inconel, a value of the resulting decrease in fluoride activity upon
5 x 10-5 may be calculated for K . This value is going from LiF to RbF while keeping a constant
about an order of magnitude less than that given mole percentage of alkali fluoride. A similar effect
in Table 2.2.2 for the equilibrium quotient ob is noted when the NaF content in the NaF-ZrF4
tained for pure chromium. The discrepancy un binary is increased from 50 to 59 mole %. The
doubtedly results from an apparent equilibrium results are also shown in Fig. 2.2.1, where Kx is
state in which the Inconel surface is depleted in plotted against either the cation radius or the
chromium because of the relatively slow diffusion mole percentage of NaF. The change in Kx with
of chromium to the surface. temperature, which is a measure of the heat of
Values for the equilibrium quotients for the reaction, is probably a good index for the mass
reaction of UF4 with chromium at 600 and at 800°C transfer of chromium. Thus, in considering the
in various reaction mediums are presented in effect of the solvent for UF4 on the corrosion of
Table 2.2.5. An examination of the data for the chromium-bearing alloys, the change in K must
alkali fluoride—ZrF4 binary mixtures containing be considered, as well as the equilibrium chromium
approximately 52 mole % of alkali fluoride shows a concentrations. In addition to the various fluoride

Table 2.2.5. Equilibrium Quotients for the Reaction 2l}F4(d) + Cr{c) ^± 2UF3(<i) + CrFAd) in Molten Fluorides

Initial Equil ibrium

Reaction Temperature UF4 Chromium

Medium (°C) Concentration Conce itrotion Kx

(mole %) (ppm)

LiF-ZrF4 600 4.0 2930 + 170 (7.5 + 2.0) x 10~4

(52-48 mole %) 800 4.0 3860 ± 140 (2.4 + 0.4) x 10-3

NaF-ZrF4 600 4.0 1730 ± 30 (1.3 ± 0.1) x 10~4

(53-47 mole %) 800 4.0 2160 + 20 (2.9 ± 0.1) x 10-4

KF-ZrF4 600 4.0 1080 ±70 (2.4 ± 0.5) x 10"5

(52-48 mole %) 800 4.0 1160 ± 60 (3.2 ± 0.7) x 10-5

RbF-ZrF4 600 4.0 69C ± 30 (1.0 ± 0.2) x 10"5

(52-48 mole %) 800 4.0 760 ± 40 (1.5 + 0.4) x 10-5

NaF-ZrF4 600 4.1 2220 ± 120 (3.2 ± 0.8) x 10~4

(50-50 mole %) 800 4.1 2550 + 210 (5.9 ± 1.8) x 10-4

NoF-ZrF4 600 3.7 1020 ± 40 (1.5 + 0.2) x 10"5

(59-41 mole %) 800 3.7 1050 + 110 (1.7 + 0.7) x 10-5

LiF-NaF-ZrF4 600 2.5 540 ± 90 (1.0 + 0.7) x 10-6

(55-22-23 mole %) 800 2.5 750 ± 60 (3.8 + 1.1) x 10-6

LiF-BeF2 550 1.5 920 (2 x 1C -6)

(48-52 mole %) 650 1.5 1300 + 220 (9 x 1C -6)

800 1.5 2060 ± 210 (7 x 1C -5)

81
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27034
the study of molten fluoride fuels for reactors.
Many of the available alloys have either a nickel or
CATION RADIUS (2)
UIO5) 0.7 0.9 1.1 1.5
an iron base, with chromium added to impart
1000 II II I 1 1
1 oxidation resistance and high-temperature strength.
Li No K Rt
Such alloys include Inconel and the stainless
500 steels. Nickel is stable toward UF4 dissolved in
the fluoride mixtures of interest and consequently
J1
is a suitable base for alloys. Data are presented
200 here which demonstrate that even though iron re
acts with UF4 the extent of the reaction is suf
100
ficiently limited that iron might also serve as a
/ ^ suitable base for the structural material. Data
\ V given in the previous section, in comparison with
\|
50 \*o
\Oo the data presented here, show conclusively that
V
ST
^\T
chromium is more reactive with UF4 than is iron
\ and indicate that the chromium present in Inconel
20 &
and the stainless steels is the element most likely
^

to react with UF4 and therefore to play the im

10
portant role in the corrosion of these alloys by
UF4-bearing fluoride mixtures. The possibility
of replacing chromium in alloys by other elements,
such as molybdenum, niobium, tantalum, vanadium,
and tungsten, led to a study of the compatibility
of these materials with UF4 dissolved in either
\\a alkali fluoride mixtures or alkali fluoride—ZrF.4
mixtures.
50 52 54 56 58 60 62 64
NaF CONTENT (mole %)
A filtration method was employed in which the
UF4-bearing fluoride mixture was equilibrated in
Fig. 2.2.1. Effect of Cation Radius and NaF Content nickel equipment with the metal under study and a
of the NaF-ZrF. Reaction Medium on the Equilibrium sample was removed through a sintered-nickel
Quotient for the Reaction of UF4 with Chromium at filter and analyzed for the desired metal. In the
600 and at 800°C. (Secret with caption) case of iron in ZrF4-bearing melts the reaction is
believed to proceed as
mixtures listed in Table 2.2.5, other solvents,
namely, LiF-BeF2-ThF4 (67-23-10 mole %) and 2UF4 + Fe°^2UF3 + FeF2
LiF-ThF4 (72-28 mole %), have been studied with
UF4 concentrations in the range 0.1 to 0.2 mole %. and equilibrium quotients may be evaluated from
The extent of reaction with chromium at these
the expression
low UF4 concentrations is so limited, however,
that no measurable difference was found between N •N
UF. FeF,

the chromium blanks and the chromium concen K =

trations resulting from the reaction of UF4 with N
UF.
chromium.

REACTION OF UF4 WITH VARIOUS STRUCTURAL The valence states of the oxidation products of the
METALS IN MOLTEN FLUORIDES
metals other than iron have not been established,
and consequently only the equilibrium metal con
J. D. Redman centrations are given.
The compatibility of UF4 dissolved in various The data presented in Table 2.2.6 show a re
fluoride mixtures with various metals having de markable insensitivity of iron concentration to
sirable structural properties, either as the metal changes in the composition of the reaction medium.
or in alloy form, is of considerable importance in Itmay be noted that the reaction exhibits a negative

82
 PERIOD ENDING DECEMBER 31, 7957

Table 2.2.6. Data for the Reduction of UF4 by Iron in as is, in general, the case for most of the metals
Molten Fluorides at 600 and at 800°C studied. Niobium is not stable in the NaF-LiF-KF-
UF4 melt, but it is unattacked by either the NaF-
Reaction Temperature Fe
K
ZrF4-UF4 or LiF-NaF-ZrF4-UF4 melt. Tantalum,
Medium (°C) (ppm) X
which would be expected to behave similarly to
niobium, is not stable in either type of melt.
LiF-ZrF4 600 500 1.4 x 10-6 Vanadium, likewise, is not stable in any of the
(52-48 mole %) 800 370 5 xlO-7 melts used, whereas tungsten is stable in the
NaF-ZrF4-UF4 melt but not in the NaF-LiF-KF-UF4
NaF-ZrF4 600 490 2 xlO-6 melt. In general, the extent of attack on the metals
(53-47 mole %) increases with a temperature increase from 600 to
800 380 6 x 10"7
800°C.
KF-ZrF4 600 540 3 xlO-6 It is evident that only molybdenum exhibits any
(52-48 mole %) 800 320 6 xlO-7 stability toward UF4 in the NaF-LiF-KF mixture,
and accordingly nickel-molybdenum alloys would be
RbF-ZrF4 600 490 4 xlO-6 expected to resist corrosion by this type of melt.
(52-48 mole %) 800 140 5 x 10-8 In the ZrF4-based melts, molybdenum, niobium,
and tungsten appear to be stable toward UF4, and
NaF-ZrF4 600 450 1 xlO-6 thus these metals or their alloys might serve as
(50-50 mole %)
800 310 3 x 10-7 suitable container materials for this type of melt.
Tantalum and vanadium appear to offer no advantage
NaF-ZrF4 600 530 1.5 xlO-6 over chromium for either type of melt.
(59-41 mole %)
800 450 8 xlO-7

LiF-NaF-ZrF4 600 420 6 xlO-7 REDUCTION OF FeF2 BY CHROMIUM

(55-22-23 mole %) IN MOLTEN FLUORIDES
800 625 2 xlO-6
J. D. Redman
NaF-LiF-KF 600 690
A study has been made of the reaction
(11.5-46.5-42 mote %) 800 480
FeF2 + Cr-
-°^Fe° + CrF„

temperature coefficient in all mediums except the in an attempt to evaluate the equilibrium quotient
LiF-NaF-ZrF4 mixture. The results given for the
LiF-NaF-KF mixture, a composition quite different "CrF.

from the alkali fluoride—ZrF4 mixtures, indicate a K = —

* N
negative temperature coefficient and equilibrium iron FeF,

concentrations very similar to those found in the

ZrF.-bearing melts. In contrast to these results at 600°C in the following reaction mediums:
for iron, the behavior of chromium, as described NaF-ZrF4 (53-47 mole %), KF-ZrF4 (52-48 mole %),
above, follows the expected pattern in the alkali and RbF-ZrF4 (52-48 mole %). Although it was
fluoride-ZrF4 melts. known that the equilibrium quotient for this reaction
was much greater than 1, it was believed that the
An examination of the data presented for various equilibrium concentrations of Cr and Fe might
metals in Tables 2.2.7 through 2.2.11 reveals some be determined sufficiently accurately by the fil
interesting differences in the behavior of the in tration method to permit an evaluation of the ratio.
dividual metals and the effects of different re If the system behaves ideally, a value of approxi
action mediums. The only metal that shows any mately 6 x 10 for Ka is obtained from the standard
real stability in an NaF-LiF-KF-UF4 melt is free energies of formation for CrF2 and FeF2 at
molybdenum, which undergoes rather limited attack. 600°C, and thus essentially complete reduction of
The attack on molybdenum by an NaF-ZrF4-UF4 Fe by the chromium is indicated. A calculation
melt is less than that of the alkali fluoride mixture, of the Cr -to-Fe ratio from values found for K„

83
ANP PROJECT PROGRESS REPORT

Table 2.2.7. Data for the Reaction of UF4 with Molybdenum in Molten Fluorides at 600 and at 800°C
Conditions of Foun d in Filtrate

Equilibration Total Total Total

Reaction

Medium Temperature Time U Mo* Ni

<°C) (hr) (wt %) (ppm) (ppm)

NaF-ZrF4 600 3 8.4 7 155

(50-50 mole %) 3 8.6 7 135

5 8.5 7 230

5 8.6 9 215

800 3 8.5 8 90

3 8.4 8 85

5 8.5 9 30

5 8.6 11 85

5 8.6 9 80

5 8.6 9 10

NaF-LiF-KF 600 3 13.3 210 85

(11.5-46.5-42 mole %) 5 14.6 200 110

5 11.8 325 170

800 3 13.5 130 170

3 13.9 105 85

5 14.4 55 205

5 13.9 65 145

*Blanks of 20 ppm were present at 800°C in both reaction mediums and have been subtracted from the amount
found in the filtrate.

for the reaction attained within 5 hr. A comparison of the experi

mentally determined chromium values with the
M° + 2UF4^2UF3 + MF2 theoretical values shows that, in the case of the
NaF-ZrF4 and KF-ZrF4 mixtures, the reaction
where M° is iron or chromium, in the NaF-ZrF4 mix occurs as indicated and yields a good balance
ture at 600°C yielded a value of ~60, however, between the Fe added and the Cr found,
which suggests that the behavior is far from ideal. except at the higher concentrations. The abnormal
The corresponding values in the KF-ZrF4 and behavior at the higher concentrations will be dis
RbF-ZrF4 mixtures are ~10 and ~3, respectively. cussed below. In the data for the runs with RbF-
These values for the Cr++-to-Fe++ ratios, if ZrF4 as the reaction medium, it may be noted that
correct, should be determinable by the filtration the calculated chromium values are about 10 to
method. 20% lower than the experimentally determined
The data given in Table 2.2.12 were obtained values, in most cases. No experimental error could
from filtration runs in nickel equipment. In most be found to explain this discrepancy, and, although
cases the equilibration time was 5 hr. Several runs, there is no evidence that the valence of the iron
with an equilibration period of 12 hr gave similar and chromium is not 2, some abnormal behavior
values and demonstrated that equilibrium was must occur.

84
 PERIOD ENDING DECEMBER 31, 1957

Table 2.2.8. Data for the Reaction of UF4 with Niobium in Molten Fluorides at 600 and at 800°C
Conditions of F ound in Filtrate

Equilibration Total Total Total

Reaction

Medium Temperature Time U Nb* Ni

(°C) (hr) (wt %) (ppm) (ppm)

NaF-ZrF4 600 3 8.2 25 55

(50-50 mole %) 3 7.9 20 60

5 8.9 25 55

5 8.7 25 80

800 3 8.8 20 40

3 8.8 20 70

5 8.8 20 80

5 8.6 20 85

LiF-NaF-ZrF4 600 3 8.2 100

(55-22-23 mole %) 3 8.8 120

5 8.4 120

5 8.9 45

800 3 8.5 110

3 8.6 110

5 8.9 5 15

5 8.3 4 40

NaF-LiF-KF 600 3 11.2 675 20

(11.5-46.5-42 mole %) 3 11.3 725 65

5 11.1 655 1

5 11.2 640 6

800 3 10.8 1090 35

3 10.5 1200 15

5 11.1 1540 6

5 11.2 1580 40

12 11.0 1830

12 10.9 1470

12 11.0 1930

12 11.1 1310

12 11.4 1230

12 12.0 1360

♦Blanks present at 800°C for the various mediums were: for NaF-ZrF4# 15 ppm; for Li F-NaF-ZrF4, 15 ppm; for
NaF-LiF-KF, 315 ppm.

85
ANP PROJECT PROGRESS REPORT

Table 2.2.9. Data for the Reaction of UF4 with Tantalum in Molten Fluorides at 600 and at 800 C

Cond tions of Found in Filtrate

Reaction
Equi ibration Total Total Total
Medium Temperature Time U Ta* Ni

(°C) (hr) (wt %) (ppm) (ppm)

NaF-ZrF4 600 5 9.4 1710 175

(53-47 mole %) 5 9.4 1630 145

800 5 9.0 2420 315

5 9.2 2820 270

NaF-LiF-KF 600 3 11.2 1320 140

(11.5-46.5-42 mole %) 3 11.6 1220 90

5 12.1 1150 110

5 11.8 1150 240

800 3 11.2 3220 170

5 11.9 2800 80

5 11.8 3020 180

*Blankof 130 ppm present at800°C in the NaF-ZrF4 mixture; 650 ppm present inthe NaF-LiF-KF mixture at 800°C.

The reduction of Fe was essentially complete limited. It will be noted that for runs made at
in all cases, except at the higher FeF2 additions 600°C with either 1 or 2% additions of Fe++ the
in the NaF-ZrF.4 mixture. In fact the Fe++ . con-
,
results are not precise and very incomplete re
centrations are so low that the ratio of Cr to duction of Fe occurred. This behavior probably
Fe cannot be evaluated in any of the reaction results from the presence of solid phases con
mediums. A value of ~60 was calculated, as taining NaF, ZrF4, CrF2, and/or FeF2. The solid
stated above, for the ratio in the NaF-ZrF4 mixture, phase contains more ZrF4 than NaF, and conse
and the experimental data do not deny that it may quently the liquid phase is enriched with respect to
be this large. However, for the KF-ZrF4 and NaF. The lack of precision indicates that equi
RbF-ZrF4 mixtures, it is evident that the experi librium was not attained in these runs. It is of
mentally determined ratios, while not constant, are interest to compare these runs with similar runs
considerably greater than the calculated ratios of made at 700°C in which the solubility of the
~10 and ~3. This lack of agreement must be CrF2-bearing compound was not exceeded. Here
attributed to some uncertainty in the K values essentially complete reduction of Fe occurred,
found for the iron and chromium reactions with and a stoichiometric relationship prevailed between
UF4. the Fe added and the Cr++ found. It is apparent
The limited solubility of CrF2 in the NaF-ZrF4 that the method employed is incapable of establishing
mixture precludes the addition of large amounts of the Cr -to-Fe ratio in these reaction mediums at
FeF2, and thus the concentration range is rather 600°C.

86
 PERIOD ENDING DECEMBER 31, 1957

Table 2.2.10. Data for the Reaction of UF. with Vanadium in Molten Fluorides at 600 and at 800°C

Cond itions of Found in Filtrate

Equi libration Total Total Total
Reaction

Med i um Temperature Time U V* Ni

(°C) (hr) (wt %) (ppm) (ppm)

NaF-ZrF4 600 3 7.4 460 120

(53-47 mole %) 3 8.1 500 25

5 8.8 420 25

5 10.3 390 35

800 3 8.3 2200 70

3 8.3 2150 30

5 8.3 1850 30

5 8.3 1600 70

NaF-LiF-KF 600 3 11.4 550 25

(11.5-46.5-42 mole %) 3 11.4 750 45

5 11.8 700 50

800 3 11.0 1450 25

3 10.6 1350 20

8 12.0 1800 50

8 11.9 1900 5

•Blank of 180 ppm present at 800°C inthe NaF-ZrF4 mixture; 150 ppm present in the NaF-LiF-KF mixture at 800°C.

THERMODYNAMIC INTERPRETATION OF where a is the activity of the designated material,

EQUILIBRIUM DATA ON THE CORROSION the salts are dissolved in the fuel melt, and the
OF IRON AND OF CHROMIUM BY metal is in the container wall. Some of the dif
MOLTEN-SALT FUELS ferences in the behavior of structural metals are
M. Blander
very important from the standpoint of compatibility
between fuel and container. The effect of tem
Considerable attention has been given to the perature, for a steady-state concentration of
corrosion and metal deposition caused by the tem UF4, UF3, and MF2 in the melt, can be represented
perature coefficient of the equilibrium as

(1) 2UF4 + M; (2UF3 + MF2

d In X
M
when molten-salt mixtures are used as the reaction (2) RTJ -(AH° + AHs)
dT
mediums. The equilibrium constant for the reac
d In K
tion is
-RT'
dT
*MF2aUF3
Ka = 20-
'This discussion covers only the tendency toward
aUF4flM mass transfer. The rate is governed by this tendency,
or driving force, and by a resistance to this force which
is a function mainly of the diffusion coefficient of the
19W. R. Grimes, ANP Quar. Prog. Rep. June 10. 1956. metal in question through the container alloy being con
ORNL-2106, p 96. sidered.

87
ANP PROJECT PROGRESS REPORT

Table 2.2.11. Data for the Reaction of UF4 with Tungsten in Molten Fluorides at 600 and at 800°C

Conditions of Found in Filtrate

Reaction Equilibration
Total Total Total
Medium Temperature Time U W* Ni

(°C) (hr) (wt %) (pom) (ppm)

NaF-LiF-KF 600 3 11.2 950 215

(11.5-46.5-42 mole %)
3 11.2 920 435

5 11.6 1410 250

5 11.2 1350 230

5 10.5 900

5 11.0 1120

800 3 11.3 980 260

3 11.2 920 205

5 11.2 1480 65

5 10.8 2150 250

5 11.8 1260 165

5 11.6 1210

5 11.8 1600

12 10.5 1880 115

12 11.8 1400

12 11.3 1430

NaF-ZrF4 600 9.2 70 110

(53-47 mole %)
9.0 95 45

9.1 70 30

800 9.2 130 40

9.3 100 50

8.5 95 50

8.9 100 20

'Blank of 350 ppm present at 800"°C in NaF-LiF-KF; <5 ppm present in NaF-ZrF,.
4

where XM is the equilibrium concentration of N N

UF.
metal M in the wall, \H° is the heat of change of MF,

the reactants in the equilibrium reaction (Eq. 1) K =

x

from their standard state (chosen here as the pure N

UF
liquids for the salts and pure solid for the metal),
AH is the difference between the heats of solution where N is the mole fraction of the designated
of the reactants and products, and K is the equi material. If AW0 + /S.Hs is positive the metal M
librium quotient: will tend to transfer from the hot to the cold end of
 PERIOD ENDING DECEMBER 31, 1957

Table 2.2.12. Data for the Reaction FeFj + Cr° ^=i Fe° + CrFj in Molten Fluorides
Present in F Itrate Calculated
Reaction Temperature FeF2 Added Cr*
Ni Fe Cr
Medium (°C) (ppm of Fe)
(ppm) (ppm) (ppm) (ppm)

NaF-ZrF4 600 0 55 70 165

(53-47 mole %) 0 35 60 165

0 85 65 155

0 85 65 160

870 60 85 960 970

915 70 55 1,010 1,010

970 90 50 980 1,060

885 110 55 880 980

2,000 110 50 2,000 2,020

2,100 no 45 2,130 2,120

2,000 195 75 1,960 2,020

2,050 65 50 1,690 2,070

4,020 20 140 4,100 3,910

4,050 70 120 3,730 3,940

4,100 55 150 4,170 3,980

3,940 30 160 4,160 3,840

6,220 20 195 5,740 5,960

6,240 35 160 4,360 5,980

6,150 30 125 6,380 5,900

6,150 25 115 5,940 5,900

10,000 50 1,320 4,460 9,400

10,000 55 1,500 2,700 9,400

10,000 90 2,010 2,210 9,400

10,000 60 1,670 2,560 9,400

20,000 55 2,140 4,620 18,700

20,000 70 3,280 1,290 18,700

700 6,000 40 15 5,840 5,740

6,000 35 15 5,660 5,740

10,000 40 20 7,700 9,400

10,000 40 25 9,500 9,400

20,000 75 70 18,400 18,700

20,000 110 80 18,400 18,700

89
ANP PROJECT PROGRESS REPORT

Table 2.2. 12 (continued)

Present in Filtr ate Calculated

Reaction Temperature FeF2 Added Cr*
Ni Fe Cr
Medium (°C) (ppm of Fe)
(ppm) (ppm) (ppm) (ppm)

KF-ZrF4 600 0 90 4 420

(52-48 mole %)
0 60 15 200

0 60 8 395

0 85 5 220

2,330 75 115 2,770 2,470

2,330 40 120 2,170 2,470

2,270 30 ISO 2,180 2,410

6,540 45 150 6,240 6,390

6,560 30 310 6,230 6,410

20,000 55 35 18,400 18,700

20,000 80 140 18,400 18,700

RbF-ZrF4 600 0 30 90 315

(52-48 mole %)
0 40 100 215

0 40 85 235

0 35 100 280

990 55 100 1,300 1,180

840 20 95 1,080 1,040

890 20 90 1,050 1,090

850 35 95 1,120 1,050

1,920 110 135 2,330 2,050

1,860 35 95 2,300 1,970

1,900 20 90 2,360 2,030

1,950 30 105 2,400 2,070

4,050 35 145 4,620 4,030

3,890 35 85 4,610 3,870

3,900 25 90 4,670 3,880

3,970 30 100 4,880 3,960

6,030 50 105 7,360 5,880

6,100 80 140 7,100 5,940

6,100 75 95 7,320 5,940

6,050 30 90 7,610 5.890

*Complete reduction of FeF, assumed.

90
 PERIOD ENDING DECEMBER 31, 7957

the system and, if negative, from the cold end to they can be obtained from the equation
the hot end. The most desirable condition would
d In y H - H'
be for AW° + AW to be zero so that there would (4) = +•

be no mass transfer by the proposed mechanism. d (l/r) R

If the pure liquids involved in the reaction are

chosen as the standard state, the values of AW° where y is the activity coefficient based on pure
tabulated below can be obtained by using the heat liquid as the standard state and the subscript N
content values listed in Table 2.2.13: denotes that the mole fractions are held constant.
A comparison can be made of the behavior of two
different pure metals, such as Cr and Fe, in any
M_ AH0
given melt composition that is relatively dilute in
Cr + 9 CrF2 or FeF2. Since the values for the heats of
Fe + 22 solution of UF3 are the same for a melt dilute in
Ni +33 CrF2 as for a melt dilute in FeF2, and similarly
for UF4, it follows from Eq. 2 that
Since all these values of AW° are positive, there d\n(Nc/NFm)
would be a tendency for mass transfer to take place (5) RT' -AW
CrF,
from the hot to the cold end of the system if the dT

AW° term is not compensated for by the term AW5,

which involves all the heats of solution. + AW° _
FeF,
(w - w °) CrP2 +(«-«°)FeP2
The term AW is given by RT2 d In [Kx(Fe)/Kx(Cr)l
(3) AWs =2(WUF3 -HSF3) +(ffMF2 -W°F2)- ~ dT

The values of K for chromium and for iron in

-2{HU -W°F )-(«M -"S) - NaF-ZrF4 (53-47 "mole %) at 600 and at 800°C
4 4
(ref 21) are given in Table 2.2.14.
where W is the partial molal heat content of the From the similarity of the compounds CrF2 and
material in solution and W° is the molal heat con FeF, it would be expected that their solution be
tent of the pure substance. For a pure metal the haviors would not differ greatly and that any dif
last term in Eq. 3 is zero. The other terms in this ference would be governed mainly by the difference
expression are unknown, but if data are available of their AW° values. As may be seen in Table
2.2.14, these suppositions are not valid. It must
be concluded either that some of the data are
Table 2.2.13. Heat Content Data* for Reactants of
the Reaction 2UF4 + M; s2UF3 + MF2
21 J. D. Redman, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 135.
Heat of Formation, Heat of Fusion,
Compound
AH298 (kcal) AH, (kcal)
Table 2.2.14. Equilibrium Quotients and Heat Changes
Derived from Temperature Coefficients for Chromium
-443 5.7
UF4 and Iron in NaF-ZrF4 (53-47 Mole %)
-357 8.5
UF3
Chromium Iron
CrF2 -181 (12)**

NiF2 -159.5 (12)**

FeF„ -168 (12)** At 600°C 1.3 x 10~4 2 x 10~°

At 800°C 2.9 x 10~4 6 x 10~7
*From Selected Values of Chemical Thermodynamic
Properties, N3S Circular 500. Aw° + Ah5 7.5 kcal -11.3 kcal
**Estimated by using a method discussed by Blander
Ah° 9 kcal 22 kcal
and Blankenship in a previous report, ANP Quar. Prog.
Rep. Sept. 30. 1957, ORNL-2387, p 121.

91
ANP PROJECT PROGRESS REPORT

incorrect or that CrF2 and FeF2 have entirely were also studied. It was conceivable that a por
different solution behavior. tion of the reactor fuel could become saturated
Speculations as to the source of the discrepancy with these gases in one section of a reactor and
may be made on the basis of available data and supersaturated in another and that bubble formation
some reasonable estimates for the heats of solu could result. This condition could lead to un
tion. From data on the activity coefficients of desirable perturbations in reactivity. The solu
FeF2 (ref 22) based on the solid as the standard bilities of helium and of argon were determined as
state and the estimate of the heat of fusion of functions of pressure and temperature.
FeF2 given in Table 2.2.13, the heat of solution Because of the chemical simplicity of the noble
of FeF2 was found to be -4 kcal/mole. Since the gases, the solubility relationships of these gases
heat of solution of UF4 is probably similar to that in molten salts may be correlated with some funda
of ZrF4 (ref 23) in the fuel melt, it is assumed to mental aspects of liquid structure and chemical
be between 2.0 and 4.3 kcal/mole. Similarly, it is composition. The variations of the solubilities,
likely that the heat of solution of UF3 in the fuel the heats and entropies of solution, and the other
melt lies in the range ±5 kcal. If these heats of properties with the atomic size of the gas may
solution are combined with the values of AW given perhaps be sufficiently meaningful to reveal some
in Table 2.2.13, a range of values for (AW° + AW,.) of the characteristics of the potential-energy func
of +15 to +27 kcal/mole is obtained for the FeF2 tions that govern the behavior of the ionic and
equilibrium; the apparent value of —11 kcal/mole molecular species in the liquid.
given in Table 2.2.14 seems to be improbable. It A summary of the results obtained thus far
may be considered therefore that the measured tem in the determination of the solubilities of various
perature coefficient of the iron equilibrium reaction noble gases in a number of different solvents is
appears to be unreasonable. The probable range presented in Table 2.2.15. In all the cases in
for (AW° + AW ) for the CrF2 equilibrium is +2 vestigated, Henry's law was found to apply. In
to +14 kcal/mole. Within the broad limits of the addition, the solubilities were found to increase
method of estimation used here the temperature with increasing temperature. For the same solvent,
coefficient of K for the chromium equilibrium is the solubilities of the gases decreased with in
not unreasonable. creasing solute gas size. Variations in solvents
showed corresponding gas solubilities to be approx
SOLUBILITIES OF NOBLE GASES IN MOLTEN imately the same in the two related ZrF4-bearing
FLUORIDE MIXTURES solvents, but the solubility of helium in NaF-KF-
N. V. Smith LiF was found to be less than half that in the
ZrF4-bearing solvents.
The solubilities of noble gases in molten salts
The heats and entropies of solution of the four
have been investigated because such gases are
gases studied in the NaF-ZrF4 (53-47 mole %)
produced in fused-salt-fueled reactors. Xenon gas
mixture are summarized below:
is an important reactor poison, and the ease of its
removal from the reactor core depends primarily
on its absolute solubility in the particular fused Gas
(kcal/mole) (eu)
salt mixture used as the fuel. A knowledge of the
temperature dependence of the solubility was needed Helium 6.2 -1.3

in order to predict in what portion of the reactor Neon 7.8 -0.4

system the gas would have the greatest tendency Argon 8.2 -1.6

to come out of solution. Xenon 11.1 -0.7

Helium and argon gases are used as protective

blankets and for sparging during reactor operations,
and therefore their solubilities in fuel mixtures
24N. V. Smith, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 140.
22C. M. Blood and G. M. Watson, ANP Quar. Prog. Rep. 25N. V. Smith, ANP Quar. Prog. Rep. March 31, 1957.
March 10. 1956, ORNL-2061, p 84. ORNL-2274, p 111.
23S. Cantor, ANP Quar. Prog. Rep. June 10. 1956, 26N. V. Smith, ANP Quar. Prog. Rep. Dec. 31. 1956,
ORNL-2106, p 111. ORNL-2221, p 130.

92
 PERIOD ENDING DECEMBER 31, 1957

Table 2.2.15. Henry's Law Constants for Various Noble Gases in Molten Fluoride Solvents

Solvent
At 600°C At 700°C At 800°C

X 10"8 x 10~8 x 10~8

NaF-ZrF4 (53-47 mole %) Helium 21.6 29.2 42.0

Neon 11.3 18.4 24.7

Argon 5.06 8.07 12.0

Xenon 1.9 3.6 6.3

NaF-ZrF4-UF4 (50-46-4 mole %) Helium 20 27 41

Xenon 2.0 4.0 6.5

NaF-KF-LiF (11.5-42-46.5 mole %) Helium 11.3

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

It may be observed that the heat of solution in Furthermore, as shown in Fig. 2.2.2, the Henry's
creases with increasing atomic size of the gas. law constants for the HF solubility illustrate a
Also it is worth noting that the entropy of solution constant heat of solution over the experimental
corrected to equal concentrations is negative and temperature range for each melt composition. The
in all cases quite small. Because of the uncer
UNCLASSIFIED
tainties of the measurements, the heats of solution ORNL-LR-DWG 27035
are considered to be reliable only to +1 kcal/mole
and the entropy changes are reliable only to +1 eu. 6.0

Accordingly it appears that the trend in the heats 5.0

of solution is real but that no particular significance

should be attached to the different values of the
3.0
entropy change, since as far as the experimental
measurements are concerned the values listed for
the entropy change are not significantly different. » 2.0

1.5 •
'^
SOLUBILITY OF HF IN NaF-ZrF4 'A

J. H. Shaffer U- 10
J*

o.y
A^
The solubility of HF in mixtures in the NaF-ZrF4 li
o o.a JS&
<y
system has been successfully determined as a <u 0.7
o

function of temperature, pressure, and melt compo fc

",
OK

sition. A summary of the recent experimental * 0.5

M ELT COMPOSITION

results and those obtained previously ' is 0.4

o 19.5 mole % ZrF. _

presented in Table 2.2.16. • 35 mole % ZrF4

A 40 mole % ZrF4
It may be noted that the solubility of HF, in all A 47 mole % ZrF4
cases, decreases with increasing temperature. 0 55 mole % ZrF„

0.2
7.0 8.0 9.0 10.0 11.0
104/M°K)
27J. H. Shaffer, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 141.
Fig. 2.2.2. Temperature Dependence of Henry's Law
28J. H. Shaffer, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 136. Constants for the Solubility of HF in NaF-ZrF4>

93
ANP PROJECT PROGRESS REPORT

Table 2.2.16. Solubility of HF in NaF-ZrF^ Mixtures as a Function of Temperature, Pressure, and Melt Composition

Melt Saturating Solubility

Temperature 3 Henry's Law
Composition HF Pressure (moles of HF/cm
(°C) Constant, K*
(mole % ZrF4) (atm) of melt)

X 10~5 X 10-5

19.5 800 2.04 9.12 4.47

2.12 9.26 4.38

Av 4.43 + 0.05

850 1.54 5.94 3.85

2.06 6.85 3.33
2.07 7.33 3.55

Av 3.58 ± 0.18

900 1.48 4.28 2.90

2.07 6.39 3.09

Av 3.00 + 0.09

35** 700 1.46 ± 0.03

800 1.06

40** 600 1.52 ± 0.03

700 1.03

800 0.81 ± 0.01

47** 550 1.38 ± 0.02

600 1.23 ± 0.04

650 1.03 ± 0.02

700 0.93 ± 0.02

750 0.86 ± 0.02

800 0.73 + 0.01

55 600 1.14 0.96 0.83

1.40 1.09 0.78
1.44 1.12 0.78
1.87 1.33 0.71
1.92 1.53 0.80

Av 0.78 + 0.03

700 1.24 0.82 0.66

2.04 1.27 0.62
2.36 1.60 0.68

Av 0.65 ± 0.02

800 1.46 0.74 0.51

1.51 0.79 0.52
2.46 1.22 0.49

Av 0.51 ± 0.01

*K — c/p in moles of HF per cubic centimeter of melt per atmosphere.

"Summarized results of previously reported data are listed for K.

94
 PERIOD ENDING DECEMBER 31, 1957

values for the heats of solution at each composition UNCLASSIFIED

ORNL-LR-DWG 27037
have been calculated from Fig. 2.2.2 and are shown 4.5

graphically in Fig. 2.2.3. The composition de

pendence of the solubility of HF is illustrated in .

Fig. 2.2.4. As may be noted, the solubility of HF 4.0

0 600°
0 700°
undergoes an inflection point at approximately A 800°
44 mole %ZrF4, and it increases with increasing
NaF concentration.
3.5

Attempts to correlate the observed changes of HF

solubility with corresponding changes in solvent E 3.0

composition will be withheld until considerably L-

o

more information is available. Experiments to E

t>

measure the solubility of HF in NaF-KF-LiF (11.5- >2.5

42-46.5 mole %) and in LiF-BeF2 and NaF-BeF2
mixtures are now under way.

2.0

UNCLASSIFIED
ORNL—LR-DWG 27036
10,000

\ 1.5
<
\

9000
1.0

0.5
8000
10 20 30 40 50 60
ZrF4 CONTENT (mole 7,1

Fig. 2.2.4. Composition Dependence of Henry's Law

-Z 7000 Constants for the Solubility of HF in NaF-ZrF4.

<
6000
\\ SOLUBILITIES OF FISSION-PRODUCT
FLUORIDES IN MOLTEN NaF-ZrF4 MIXTURES
1 1
W. T. Ward R. A. Strehlow

5000 \ Experiments have been completed in order to

determine (1) the effect that varying the ratio of

\ sodium to zirconium in the NaF-ZrF4 solvent has

on the solubility of CeF3, (2) the solubility of YF3

4000 \ in NaF-ZrF4 (53-47 mole %), and (3) the solubilities

of CeF3 and SmF3 when both are present in NaF-
ZrF4 (53-47 mole %). The solubility measurements
were made by using the radiochemical tracer tech
niques described previously. '
3000
20 40 60 80

29W. T. Ward, ANP Quar. Prog. Rep. Dec. 31. 1956.

ORNL-2221, p 125.
Fig. 2.2.3. Composition Dependence of the Heat of
30W. T. Ward, ANP Quar. Prog. Rep. March 31. 1957,
Solution of HF in NaF-ZrF.. ORNL-2274, p 105.

95
ANP PROJECT PROGRESS REPORT

CeF3 in NaF-ZrF4 UNCLASSIFIED

ORNL-LR-DWG 27039

It has been found that varying the ratio of sodium

to zirconium in the solvent has a considerable
effect on the solubility of CeF3< This effect is •
clearly shown in Fig. 2.2.5, in which the solubility
vs temperature curves for five different NaF-ZrF4
compositions are plotted. In addition to the compo
sitions indicated on Fig. 2.2.5, the NaF-ZrF4
(80.5-19.5 mole %) solvent was briefly investigated.
It was found that all the 4.60 mole %CeF3 that was -i 3

added was soluble at temperatures from 793 to

915°C.

UNCLASSIFIED
ORNL-LR-DWG 27038
10 ^
'N^A
>A -s.

i 30 40 50 60
NaF-2 rF4 SOLVENT
COM POSITION

\ 42- 58 mole % Fig. 2.2.6. Solubility of CeF3 at 600°C in NaF-ZrF4

Mixtures as a Function of Composition.

-*- 53--47 mole la

>
of the similarity of the U and the Ce ionic
radii, that of U being 1.03 A and that of Ce
being 1.02 A. It is quite likely that an inflection
exists in the solubility curve at a composition near
X 41 mole % ZrF4, since with higher NaF content
\
0.5
«- *9- % the precipitating phase is not CeF3 but rather
NaCeF4, as determined by petrographic observation.
YF3 in NaF-ZrF4
- 63--37 mole 10

The solubility of YF3 in NaF-ZrF4 (53-47 mole %)

10 11 12 13 15 was determined, and the results are presented in
10V7" (°K) Table 2.2.17 and Fig. 2.2.7. Although some

Fig. 2.2.5. Solubility vs Temperature for CeF-, in Table 2.2.17. YF3 Solubility in NaF-ZrF4
Various NaF-ZrF4 Mixtures. (53-47 Mole %)

One interesting observation is that in the solvent Filtration

Solubil ty
NaF-ZrF4 (42-58 mole %) the nature of the pre Temperature
cipitating phase changes at about 640°C, the in (°C)
In wt % Y Ir mole %YF3
flection point on the curve. A plot of CeF3 solu
bility at 600°C as a function of composition is 854 11.37 13.68
presented in Fig. 2.2.6, and it may be noted that
817 8.68 10.26
the curve is quite similar to that obtained earlier
for the solubility of UF3 in NaF-ZrF4. Such 756 6.31 7.39

correlation would be expected from a consideration 663 4.10 4.75

31 560
G. M. Watson and C. M. Blood, ANP Quar. Prog. 2.37 2.72
Rep. June 10, 1954, ORNL-1729, p 51.

96
 PERIOD ENDING DECEMBER 31, 1957

scattered earlier results had indicated that the using the two-tracer technique described pre
plot of the log of the solubility vs the reciprocal viously. Filtrates were obtained over the tem
of the temperature for YF3 was linear, these results perature range of 550 to 800°C for two cerium-to-
show a behavior pattern more similar to the patterns samarium ratios. A summary of the solubility data
obtained for the rare earths. is presented in Table 2.2.18, along with the solu
bility of CeF3 by itself in the same solvent. The
CeF, plus SmF3 in NaF-ZrF4 solubility vs temperature relationships are illus
The solubilities
ubilities of CeF3 and SmF3 iin combina-
CeF3 and trated in Fig. 2.2.8, where the logs of the sums of
tion in NaF-ZrF4 (53-47 mole %) were measured by the solubilities are plotted vs the reciprocal of the
temperature.
UNCLASSIFIED
ORNL-LR-DWG 27040 The solubility of pure SmF3 has not yet been
20
measured in this solvent, but, by utilizing the
correlation of solubility with the cube of the ionic
15
\
• radius, estimates were made of the solubilities of
\ SmF, at these temperatures. The values obtained
to
• are shown in Table 2.2.19.
9
\ These values are summarized in the pseudo-
8 \ ternary diagram presented as Fig. 2.2.9. The rare-
\
7
earth fluoride solubility behavior in this solvent
6 is seen to be substantially the same as the be
5 havior reported previously when the solvent was
•
<
or NaF-ZrF4-UF4 (50-46-4 mole %).32
SOLUBILITY OF BaO IN KF-LiF

J. H. Shaffer
\
The solubility of BaO in KF-LiF (50-50 mole %)
is of interest in the systematic investigation of
possibilities of removal of fission-product fluorides
1.5 from NaF-LiF-KF base mixtures. Three experi
ments have been performed to obtain solubility
values in which the starting barium compound,
10 11 12 13
either BaF, or BaO, was labeled with radioactive
io4/r(°K) Ba133.

Fig. 2.2.7. Solubility vs Temperature for YF3 in 32

R. A. Strehlow and W. T. Ward, ANP Quar. Prog.
NaF-ZrF4 (53-47 Mole %). Rep. June 30, 1957, ORNL-2340, p 139.

Table 2.2.18. Solubilities of CeF3 and SmF3 in NaF-ZrF4 (53-47 Mole %)

Tota 1 Additive

4.48 Mole %SmF3, 4.76 Mole %SmF3, 6. 1 Mole %CeF3,

Filtration %CeF3
1.90 Mole %CeF3 7.14 Mole no SmF,
Temperature
Solubil i ties (mole %) Solubilities (mole %) So lubil ity of CeF,
(°C)
(mole %)
SmF3 CeF3 SmF3 + CeF3 SmF3 CeF, SmF, + CeF.

800 3.23 1.48 4.71 1.56 2.48 4.04 3.78

700 2.17 0.95 3.12 1.13 1.66 2.79 2.57

600 1.62 0.62 2.24 0.86 1.16 2.02 1.90

550 1.45 0.50 1.95 0.76 0.97 1.73 1.67

97
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED UNCLASSIFIED
ORNL —LR-DWG 27041 ORNL-LR-DWG 27042

Fig. 2.2.9. Pseudoternary Diagram of the System

CeF3-SmF3 with the Mixture NaF-ZrF4 (53-47 Mole %)
as the Solvent.

The melting point and heat of fusion of BaO are

listed33 as 1923°C and 13.8 kcal/mole, respec
tively. Accordingly the estimated solubility of
BaO at 600°C would be quite low (less than 0.5
mole %). Since it was considered to be difficult
to prepare labeled BaO free of traces of soluble
impurities, such as carbonates, hydroxides, and
chlorides, it was decided to prepare the BaO in situ
1Cr/7"(°K)
from BaF2. The soluble impurities would cause
apparent solubility values of BaO to be erroneously
Fig. 2.2.8. Solubilities of SmF, and CeF, in NaF-
high. This preparation was attempted in two dif
ZrF. (53-47 Mole %) as a Function of Temperature.
ferent experiments. The first consisted in the
addition to the melt of LiOH containing a known
Table 2.2.19. Estimates of the Solubility of SmF, in
amount of labeled BaF2. A filtrate sample was
obtained prior to the addition of the LiOH, and
NaF-ZrF4, (53-47 Mole %) Bas
Based on the
filtrates at 700 and 600°C were obtained after the
Solubility of CeF,
addition was completed. Since no change was
noted in the concentration of barium in solution,
_. - , , ... , _ r- Estimated SmF-
Temperature Solubility ot Cer, •» it was decided to repeat the experiment with sodium
,o^ , i „,* Solubility
°C) (mole %) ' peroxide as the precipitating agent. The results
(mole %)
of the second experiment also showed no change
in the concentration of barium in the liquid after
800 3.78 4.40
addition of 20% and of 140% excess Na202, calcu
700 2.57 2.99 lated as Na20.
600 1.90 2.21
3 3
A. Glassner, A Survey of the Free Energies of
550 1.67 1.95 Formation of the Fluorides, Chlorides, and Oxides of
the Elements to 2500°K, ANL-5107, p 6 (Aug. 1953).

98
 PERIOD ENDING DECEMBER 31, 1957

In view of the unexpected results obtained in containing 1 mole % U as UF4 at 600°C. Calcium
the first two experiments, a third experiment was oxide was then added in tared increments, and
carried out with labeled BaO as the additive. The filtrate samples were taken approximately 2 hr
labeled BaO was prepared by ignition of Ba(N03)2. after each addition. These samples were counted
Again the results showed that the solubility of for radioactivity to determine the cerium concen
BaO was quite high. The labeled BaO used was tration and were chemically analyzed for uranium.
examined with the petrographic microscope and by The results are shown graphically in Fig. 2.2.10.
x-ray diffraction. The results of these examina It may be noted that the cerium concentration
tions show that the material used was high-quality increases, initially, because of the removal of the
BaO. The numerical results of the three experi uranium from solution, until the equivalence point
ments are summarized in Table 2.2.20. for the uranium is reached. Upon further addition

UNCLASSIFIED

REACTION OF CeF3 AND UF4 WITH ORNL-LR-DWG 27043

1
CaO IN KF-LiF

J. H. Shaffer
i5i \
Studies of the precipitation of cerium as Ce203 a:

£ o CeF3
from the molten KF-LiF eutectic and of the solu • UF4
Z 1
bility of barium oxide in this solvent (see preceding O 1

section) have provided a basis for-investigating •

o 1
Z 1
UJ |
the possible separation of tetravalent metals as
oxides from the trivalent fission products of a \ <!
u 1

\
UJ !
fused-salt reactor. Alternatives to this separation
are to precipitate the rare earths as oxides and
leave the uranium in the solvent or to precipitate \8I
uranium oxide and leave the fission products in
solution.
In the present study, 5 wt % of labeled Ce was
added as CeF3 to the molten KF-LiF eutectic
0.2
\l .0.032

0.3
EQUIVALENTS OF CoO ADDED
0.4
0.0085-/\ ?
0.5 0.6

34J. H. Shaffer, ANP Quar. Prog. Rep. Sept. 30, 1957, Fig. 2.2.10. Precipitation of CeF3 and UF. with CaO
ORNL-2387, p 139. in KF-LiF (50-50 Mole %).

Table 2.2.20. Concentration of Barium in Liquid KF-LiF After Addition of BaF2-LiOH, BaFj-NajOj, and BaO

Initial Barium Concentration Final B arium Concentration

Temperature (wt % Ba)
(wt % Ba)
Additive
(°C)
Ca Iculated Experimental Calculated Exp erimental

BaF2-LiOH 600 11.4 9.85 9.93 10.45

700 9.93 10.95

BaF2-Na202 (20% excess) 600 8.28 8.1 7.82 7.6

700 7.82 7.6

800 7.82 7.6

BaO 600 6.94 6.6

700 6.94 6.6

800 6.94 6.6

99
ANP PROJECT PROGRESS REPORT

of CaO the soluble cerium concentration decreases AC - concentration change of diffusing ma

linearly to zero. The uranium concentration de terial across L, C _. —C ., g/cm ,
creases to a value of approximately 0.1 wt % at the D = diffusion coefficient, cm /sec.
equivalence point. This experiment is now being
repeated with the exception that the initial CeF3 The diffusion coefficient is a function of the
concentration has been decreased to 0.5 wt %. temperature, the diffusing material, and the ma
Another experiment is proposed in which ZrF4with terial through which diffusion takes place. It does
a labeled hafnium tracer is to be substituted for not depend on the geometry of the flow system.
the UF4. It is entirely independent of the molten salt in
contact with the surface of the metal, although
DETERMINATION OF CHROMIUM DIFFUSION
the salt has a mqrked effect on AC and on the rate
COEFFICIENTS IN INCON EL-MOLTEN
equations. Therefore it is not necessary to employ
SALT SYSTEMS
molten salts in experiments designed to determine
R. B. Evans diffusion coefficients, even though the ultimate
The results of experiments carried out over the application involves molten-salt systems.
past few years on the migration of metal as a re The steady-state equation (Eq. 2) is useful for
sult of temperature gradients in a molten-salt en discussion and forms the basis for the determina
vironment have strongly suggested that the over tion of flow constants in analogous systems (flow
all migration rate is controlled by the diffusion of heat, flow of electricity, and fluid flow in porous
rates within the metal. Accordingly, a thorough mediums); however, its use in solid-state diffusion
knowledge of solid-state diffusion phenomena is studies has been discouraged because of the ex
considered to be an important phase of corrosion tremely low values of AM/At and D involved. The
studies. Several investigations of these phenomena only alternative is utilization of experiments and
have been made, but only a small portion of the equations based on unsteady-state flow. The basic
work covered cases in which molten salts were linear flow equation for this case is
involved. Therefore an investigation was initiated
dC d / dC
of the over-all mass transfer rate within isothermal (3) —Id —
systems in which Inconel is the container material dt dx \ dx
and NaF-ZrF4 is the salt. The diffusing metal in
Inconel in this system is chromium. If self-diffusion measurements are to be made, for
example, diffusion of radioactive silver into pure
Diffusion Coefficients silver, and/or if the values of dC/dx involved
The diffusion coefficient is a flow-resistance
are not high, Eq. 3 may be written as37
parameter used in diffusion rate—time relationships. dC d2C
It is defined by the linear flow equation (4)
dx2
1 dM dC
(1) Equation 4 is known as Fick's diffusion law.
A dt dx
Measurements Based on Radioactive Tracers
which, if steady-state conditions are established
and D is assumed to be independent of concentra Tracer-Layer Method. — One method for experi
tion, may be written as mentally determining diffusion coefficients for a
given metal diffusing through a bar composed of
1 AM AC the same or a different metal is based on the use
(2) D
A ~aT of radioactive isotopes of the given metal. A very
thin layer of solid labeled metal is placed, by any
where of various methods, on the face of a solid specimen
AM/At constant rate of diffusion, g/sec, 3 5
It will be shown, however, that utilization of molten
A area through which diffusion takes salts leads to a unique and simple method of indirectly
place, cm , determining diffusion coefficients for metals in which
the salt rather than the metal is analyzed.
L = length, cm, of system in x direction 36J. C. Fischer, J. Appl. Phy. 22, 74 (1951).
(L is zero at the surface and increases 37
A. C. Wahl and N. Bonner (eds.), Radioactivity
with penetration of the metal wall), Applied to Chemistry, Wiley, New York, 1951.

100
 PERIOD ENDING DECEMBER 31, 1957

of unlabeled metal. Uniform placement ensures Fig. 2.2.11 that a has been corrected for time
linear flow, and the layer is never replenished after decay. The diffusion coefficient determined for
initial placement. The specimen is then subjected the labeled material is essentially the diffusion
to a constant temperature (T > 600°C) for a speci coefficient for the unlabeled material.
fied period of time during which the labeled metal Excellent results have been reported for
diffuses into the bar. Even though the specimens this type of experiment. In some instances, two
may be of small dimensions, they may be con types of diffusion occurred simultaneously. The
sidered as being infinite in length with respect to two types are (1) diffusion through the grains and
the direction and distance of labeled-metal in (2) diffusion along the grain boundaries; the latter
vasion, since the rates of diffusion involved are controls the over-all diffusion rate at lower tem
very low. peratures, and a combination of the two controls
A solution of Eq. 4 under boundary conditions ' the rate at higher temperatures. In instances in
which closely approximate the described experi which both types of diffusion occur, log a vs x2
mental condition is plots tend to bend. ' Such results leave con
2*0 siderable doubt as to the value of the over-all
(5) C* = -x2/4Dt diffusion coefficient.
1/2
(nDt) Labeled-Molten-Salt Method. — If consideration is
given to a semi-infinite isothermal system wherein
where all quantities other than Q* have been pre
an alloy containing metal M° is in contact (at one
viously defined, and the starred variables refer to
the labeled metal. A physical interpretation of face) with a molten salt containing a fixed amount
Qt, g/cm , is somewhat difficult for the case of of MX2, it may be seen that at equilibrium the net
rate of transfer of M between the metal and the salt
diffusional flow; it is analogous to an "impulse of
energy" for heat or electrical flow systems. It will be zero, although a random dynamic exchange
may be thought of as being the product of the layer of M will take place. If a small amount of the MX-
were removed and replaced by an equal amount of
thickness and the layer density, the thickness
being zero. The layer is put on at an infinite labeled M*X2, the random dynamic exchange could
be studied by following the exchange rate of the
rate over a zero time interval.
A sketch of C* vs x at various times / is shown
labeled M*X2. The expression which relates the
in Fig. 2.2.11. The quantity dC/dx at x = 0 is quantities of interest in terms of the surface ex
zero, and thus dM/dt at x - 0 is zero for all / > 0. change reaction or "wall reaction,"
Also shown in Fig. 2.2.11 is a sketch of log C vs x M° + M*X, ;MX, + M°* , AF° = 0,
at various times. It is apparent from Eq. 5 that the
slopes of the lines of this plot are equivalent to
39
-(4Dt)~ . Thus D may be determined from a log W. A. Johnson, Trans. Am. Inst. Mining Met. Engrs.,
Inst. Metals Div. 143, 107 (1941).
a vs x plot. It is assumed in the sketches of
P. L. Gruzin and G. B. Fedorov, Doklady Akad.
Nauk S.S.S.R. 105, 264 (1955).

38
S. T. Kishkin and S. Z. Bokstein, Proc. Intern.
R. V. Churchill, Modern Operational Mathematics in Conf. Peaceful Uses Atomic Energy, Geneva, 1955 15,
Engineering, p 117, McGraw-Hill, New York, 1944. 87 (1955).

UNCLASSIFIED
ORNL-LR-DWG 27044

TIME:
log C
OR OR
a, ACTIVITY log a

x (cm) xz (cm2)

Fig. 2.2.11. Concentration of Labeled Metal at Various Times

101
ANP PROJECT PROGRESS REPORT

is be directly determined from the slopes of AM* vs

t plots. These slopes are equivalent to the
[M°*][MX2] quantity 2 CM°JX=0 (D/tt) / . The conditions set
(6a) forth for the equations would be difficult to dupli
[M°] [M*XJ
cate experimentally, however, since the necessity
The equilibrium of Eq. 6a is established within a of using small sample containers and of excluding
short period of time after the M*X2 is added. Since air suggests the use of sealed capsules. Con
tinuous addition of M*X2 could not be accomplished,
K
a
= 1 = K
>
and Eq. 8 would be applicable only at early flow
times. In view of the low diffusion rates involved,
the concentration of M°* at the surface of ihe alloy
it is not important that the conditions cannot be
at any time is given by the equation duplicated experimentally. Equation 4 can be
[M*X,] solved in such a way as to take into account de
(6b) [M°]. pletion behavior. A rigorous solution is
[MO*U [MX2] AM*
= 1 - ea~l erfc at]/2
The bracketed concentration expressions are taken M*
'f=0
to be weight fractions.
The M°* will diffuse back into the metal as time where a is proportional to D . It can be shown
progresses and tend to deplete both the M°* at the that Eq. 9 is a first approximation to the solution
surface and the M*X2 in the salt. Depletion will for small values of t.
occur unless M*X2 is added to the system at a Diffusion coefficients obtained by the labeled-
rate equal to the diffusion rate. If the latter molten-salt method are of value because they are
manipulation is carried out, or if it can be assumed over-all diffusion coefficients, that is, coefficients
that M*X2 remains constant, tM°*]x=0 is constant that are independent of microscopic behavior within
with time. The solution of Fick's law for this the alloy. Further, the experimental procedure is
case is simple: the initial value of AM* is taken to be the
early-time depletion of M*X2 in the melt, and the
(7) CMo, = CMo,Jx=0 erfc melt is assayed at various times.
2(Dt) 1/2
Battelle Diffusion Experiments
whereby Isothermal capsule diffusion experiments have
1/2
recently been conducted at Battelle Memorial
1 dM*
Institute on Inconel—fused fluoride salt systems.4
(8)
~~A~dT x=0
:m°*L (^ Radioactive chromous fluoride was added to the
salt in order to study the combined effect of the
and wall reaction and the diffusion coefficient on the
chromium transfer rate.
AM* ' (dm* Briefly, the experimental procedure consisted of
(9)
41 dt
placing a series of identical Inconel capsules,
each containing 2 g of salt and known amounts of
1/2 labeled CrF2, in a furnace. The capsules were
2 CMo, removed individually at various time intervals.
x=0
The salts in the capsules removed were analyzed
In Eq. 9, AM* is the cumulative amount of M°* for the amount of CrF2 and Cr*F2 present. In
(in grams) which has passed through A (cm2) of the addition, the walls of one capsule from each series
metal surface at any time t (sec). The quantity of isothermal tests were sectioned by an electro-
C..o*] x—u
_n is the concentration of , metallic labeled polishing technique to obtain the wall distribution
M at the surface of the alloy (g/cm ).
42
If AM* is measured as a function of time under R. Price et al., A Tracer Study of the Transport of
Chromium in Fluoride Fuel Systems, BMI-1194 (June 18.
the conditions stated, the diffusion coefficient can 1957).

102
 PERIOD ENDING DECEMBER 31, 1957

of labeled chromium. The data obtained are used Eqs. 6 and 9. The expression is
below in a critical analysis of the proposed method
for the determination of diffusion coefficients. % Depletion
(10)
Data Based on Molten-Salt Analyses. —A portion [* (hr)] 1/2
of the results of the Battelle experiments is repro 1/2
A [Cr°] D
duced in Fig. 2.2.12. This figure indicates that (60 (200)— pMeta,
(1) at low temperatures and (2) at early flow times N, [CrF2]
for high temperatures the amount of labeled chro
mium which diffuses from the salt into the metal where PMeta) is the density of Inconel (g/cm3),
varies directly with the square root of the flow NT >s the amount of salt present (g), [CrFj] is the
time. It may also be noted that as time goes on weight fraction of CrF2 in the salt, reported as
the experimental points at higher temperatures fall Cr , and m is the slope given on Fig. 2.2.12.
away from the straight-line relationship. This This method was used to estimate values of D
was to be expected because the quantities C_ oj _. from the results of several capsule experiments.
and [Cr*FJ were not held constant during the The experiments were divided into two groups,
experiment. with the base salt for one group being NaF-ZrF4
(50-50 mole %) and for the other NaF-ZrF4-UF4
A reasonable approximation of the diffusion (50-46-4 mole %). It was necessary to assume that
coefficient can be obtained, however, from the equilibrium amounts of UF3, UF4, and CrF2 were
slopes of the straight lines in Fig. 2.2.12. The present in the base salt of the second group,
expression relating the slope of these lines to the since data on the [UF3]/[UF4] ratios and the
diffusion coefficient can be obtained by combining [CrF2] variations as functions of time were not

UNCLASSIFIED
ORNL-LR-DWG 27045

SQUARE ROOT OF EXPOSURE TIME (hr c)

Fig. 2.2.12. Net Loss of Cr*F2 Activity from NaF-ZrF4-UF4 (50-46-4 Mole %) Because of Cr* Diffusion into
Inconel Capsule Wall. (Secret with caption)

103
ANP PROJECT PROGRESS REPORT

available. The estimated values are presented in In order to evaluate the discrepancies, the value
Fig. 2.2.15, which is discussed in a subsequent of D from the distribution data was used to con
section of this report. struct a depletion curve. This in turn was com
Data Based on Distribution of Labeled Metal in pared with the actual depletion data obtained for
Wall. - The Cr°* wall distribution data formed the the same experiment. The results are shown in
basis for the values of the diffusion coefficients Fig. 2.2.14.
reported by the Battelle investigators. Equation 5 Additional comparisons with chromium diffusion
was utilized for correlating the results, although coefficients available in the literature are pre
the boundary condition sented in Fig. 2.2.15. These values are also
tabulated in Table 2.2.21 along with the corre
dM sponding activation energies for diffusion. Although
= 0
~dt J Eq. 7 did not fit the distribution data, it was con
x=0
cluded on the basis of Table 2.2.21 that Eq. 9
was obviously violated by the experimental condi was suitable for estimating values of the over-all
tions. A plot of one set of the Battelle distribution diffusion coefficient.
data is shown in Fig. 2.2.13. The "volume" The magnitudes of the variations encountered in
diffusion curve reported by the Battelle investi terms of amount of material transferred are not
gators, as well as the distribution predicted by immediately apparent from plots such as Fig. 2.2rl5;
Eq. 7, was superimposed over the data, and both hence, calculated rate values are presented in
curves failed to fit the data. The "average D" Table 2.2.22 to give a "feel for the numbers" in
calculated by both equations was 10" cm /sec. terms of rate of diffusion and amount transferred as
The fact that the curve of Eq. 7 did not fit casts a function of time.
some doubt as to the validity of the related Eq. 9. The values presented in Table 2.2.22 were cal
culated by using Eq. 2 for the amount of steady-
UNCLASSIFIED
state chromium diffusion which would occur over

4000
ORNL-LR-DWG 27046
a one-year interval through a thin Inconel sheet,
one face of which was backed up by a thick chro
mium plate, and the other face of which was con
tinuously exposed to a fresh molten NaF-ZrF4-UF4
salt mixture at 900°C. The sheet was assumed to
be 3 mils thick, with a 130-cm cross-sectional
area (equivalent to the area exposed in a 16-in.
loop of 0.5-in. Inconel tubing). A comparison of
these values suggests that the diffusion coeffi
cients reported by the Battelle investigators yield
diffusion rates which are somewhat pessimistic
with respect to Inconel corrosion properties.

STEADY-STATE CHROMIUM MIGRATION IN

THERMAL-CONVECTION LOOPS

R. B. Evans

The results of many investigations of the

corrosion of polythermal Inconel loops have shown
that chromium is transported from hot to cold
zones by circulating NaF-ZrF4-UF salts. Mass
transfer of chromium also occurs with other salt
4 8 12 16 20

PENETRATION DEPTH (<i)

W. D. Manly et al., Metallurgical Aspects of Molten
Fluoride Reactors, ORNL-2349 (Sept. 17, 1957).
Fig. 2.2.13. Cr51 Distribution in Inconel at 700°C 44W. R. Grimes, ANP Quar. Prog. Rep. June 10, 1956,
After 739 hr of Exposure. ORNL-2106, p 96.

104
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
:-.:->» ORNL-LR-DWG 27047
12 •

/
/
1

(2) . _^-
1
/ <>
___^-

(3)
Q
UJ
or
or
\ ^^ ^~~
•

UJ
u. 1
</)
z
<
1
1
or

O 4 1
1
yy
s^
is DIFFUSION COEFFICIENT BASED ON
1 ,--0* fMOTCOMTI^M ,„ ,.,», ,

1 • y? (2) Cr*F2 DEPLETION RATE

(3) DATA FROM GRUZIN AND FEDER 3V7
r [TYDirPIMFMT/U pniMTQ 1IQFn TCI CONSTRUC T CURVE (2 )

•
1 X •

200 400 600 800 1000

DIFFUSION TIME (hr)

Fig. 2.2.14. Effect of Diffusion Coefficients as Determined by Various Methods on Predicted Diffusion Rates
of Labeled Chromium in Inconel.

Table 2.2.21. Diffusion Coefficients for Chromium at 700 C in Various Nickel-Base Alloys from Various Sources

Activation Energy*
Chromium Diffusion Coefficient Data
Type of Metal for Diffusion Values Reported by
(wt %) (cm /sec) Reference
(kcal/g-at.)

-15
x 10

Ni-Cr alloy 3 20.4 3.2' 66 Gruzin and Fedorov 40

Inconel 16 3.7 Present investigation

Pure nickel 0 71** 48 Gruzin and Fedorov 40

Inconel 16 100 41 Price et al. 42

-B JRT
act
*D=D0e
'Extrapolated values.

105
ANP PROJECT PROGRESS REPORT

ORNL-LR-DWG 27048

TEMPERATURE CO

1200 1100 1000 900 800 700 600

17
III II 1 I I 1 1 1 1
DATA
SYMBOL CONTAINER SALT REFERENCE
A NICKEL NONE 40
16 •
A Ni-CrALL0Y2 NONE 40
o Ni-CrALL0Y3 NONE 40 s
s
• Ni-CrALL0Y4 NONE 40 s
s

15 —
— X INCONEL* NaF-ZrF4-UF4 s

• INCONEL* NaF-ZrF4 s

• INCONEL** NONE 42
* INCONEL** NONE 42 j>
' D ♦
14 s

•
a s

s
x--'
*DATA FROM PRESENT INVESTIGATION s
xy S*
^

o 13 —
BAS ED ON C *F2 DEP LETION
**DAT A FROM F EPORT B Y BATTEL LE42
r*>>*''
*'
BAS ED ON C r°* WALL DISTRIB JTION y/1
12

10

10 12

io4/r (°k)

Fig. 2.2.15. Comparisons of Various Diffusion Coefficients.

Table 2.2.22. Calculated Steady-State Diffusion Rates for Chromium Through Thin Alloy Sheets at 900°C

Diffusion
Diffusion Rate Diffusion Rate Calculations Based on Data
Sheet Material Coefficient
(g/sec) (g/year) Coefficients Reported By Reference
(cm /sec)

-13* -8
Inconel 8x 10 1.4x 10 4.5 Present investigation

-13
Ni-Cr alloy 3 8x 10 1.8 x 10" 5.7 Gruzin and Fedorov 40

-12
Inconel** 3x 10 5.2 X 10"° 17 Price et al. 42

-11
Inconel*** 2x 10 3.5 x 10~7 110 Price et al. 42

* Extrapolated value.
**By volume diffusion alone.
***By boundary diffusion alone.

106
 PERIOD ENDING DECEMBER 31, 7957

systems and metals; however, this report deals and is carried as CrF, to the cold zone via the
only with the chromium migration peculiar to the circulating salt. At the cold-zone surface, the
NaF-ZrF4-UF salt system contained in Inconel. reaction is reversed, and the chromium reverts
In all experiments the loops have been loaded to the metallic state. The deposition process
with pure salt mixtures, that is, mixtures which results in a chromium concentration gradient
did not initially contain corrosion products. along the cold-zone surface which causes the
Therefore the over-all transport vs time relation chromium metal to diffuse into the tube walls.
ships observed could be broken down into three The diffusion into or out of the walls occurs in
consecutive flow situations or regions. The a direction which is essentially normal to the
study of each region required a separate analytical surface and is governed by semi-infinite, linear
approach, since the first two regions involve flow relationships.
unsteady-state solutions and the transport phe The CrF, concentration in the salt builds up
nomena corresponding to the third and final region initially with respect to time and then remains
are governed by a quasi-steady-state situation. essentially constant. Thus a quasi-steady-state
The analysis of the third region is relatively situation appears to exist whereby chromium is
simple compared with the analyses of the un transported at very low rates and under conditions
steady-state regions. The final steady-state of a fixed chromium surface concentration vs
situation is of considerable interest because low loop position relationship. This idea is supported
migration rates occur in this region. by the fact that corrosion continues even though
One objective of the present study is to develop the amounts of CrF,, UF , and UF, in the salt
a method whereby the amount of chromium trans remain constant.
ported can be estimated as a function of loop
Steady-State Migration
operating time, and thus to be able to predict
the extent of corrosion to be expected at any Based on experimental evidence alone, there is
time. This method will be applicable only for still some doubt as to whether the loops operated
the case in which loop operation is initiated with thus far have actually attained the quasi-steady
the equilibrium concentration of the corrosion state. Although the CrF, concentration remains
products in the salt mixture. Chromium migration constant, within the accuracy of the measurements,
behavior for this case approximates the behavior it has not been possible to make a direct, quan
in the quasi-steady-state region, with the ex titative measurement of the total over-all amount
ception that the distribution of chromium in the of chromium transported. An irrefutable statement
wall does not depend on what might have happened about the existence of a quasi-steady state cannot
during the unsteady-state period. The finding be made without such measurements.
of a relatively simple method of estimating the Another approach to a study of the migration is
equilibrium concentrations is a further objective to assume that the equilibrium amounts of CrF,
of this study. and UF, are present in the salt before the loop
and salt are subjected to the operating conditions.
With this and the migration mechanism in mind,
Chromium Migration Mechanism the following conditions may be specified for the
The chromium migration process in thermal- resulting migration:
convection loops involves two chemical changes 1. At any given time the cumulative amount of
chromium which leaves the hot zone must equal
and two transport mechanisms. Over the entire
hot zone, the metallic chromium slowly diffuses the cumulative amount deposited in the cold zone.
Furthermore, the over-all instantaneous rate of
from the matrix of the Inconel tubing to the salt-
transfer out of the hot zone must equal the in
exposed surface. The chromium concentration
stantaneous rate of transfer into the cold zone.
gradient in the metal is the driving force for the
diffusion process. The transported chromium is 45
The equilibrium constants for the reaction
then oxidized at the surface according to the Ni° + 2UF, 2UF3 + NiF2
reaction
are so low compared with those for the chromium reaction
that the transport of nickel by the mechanism described
(1) Cr° + 2UF, 2UF3 + CrF„ here may be ignored.

107
ANP PROJECT PROGRESS REPORT

2. The over-all rate of transfer is diffusion the CrF, concentration in the circulating salt
controlled, since the diffusion rates are very much ceases to vary with time. The first two regions
lower than the other rates involved. mentioned in the introduction are separated by
3. There must be at least two points within the the point Z. The last region starts when the
loop at which chromium is neither deposited nor balance points become stationary.
depleted. In subsequent discussions, these points A very important observation can be made on
will be referred to as the balance points. the basis of the above discussion; that is, the
4. At the balance points (a) no diffusion takes over-all amount of chromium transferred during
place within the walls; (b) the chromium concen the unsteady-state time interval would be higher
tration at the surface is the same as the concen than that transferred in any identical pre-equili-
tration in the original Inconel; (c) the circulating brated loop which started at the same time. This
UF,, UF4, and CrF2, the concentrations of which observation is valid, since the depletion areas
do not vary with loop position or time, must be in and driving forces for the unsteady-state case
equilibrium with the original concentration of are greater (CrF concentrations much lower)
chromium in Inconel,46 and thus the concentrations than those for the steady-state case. In addition,
of the circulating salts are fixed by the equilibrium the degree of chromium depletion (corrosion)
constants (and temperatures) at these points; and per unit area in the maximum-temperature portion
(d) the temperatures controlling the equilibrium of the loop is highest in the unsteady-state case.
constants must be the same. In other words, pre-equilibration is good operating
5. The loop position vs temperature relation procedure with respect to suppressing chromium
ships are constant with respect to time. Thus migration within loops. The remainder of this
the balance points must remain at fixed loop discussion will deal with loops which have been
positions. pre-equilibrated at conditions close to the op
6. The concentration of chromium at any point erating balance-point conditions.
on the tubing surface varies with loop position
but not with time.
Estimating Equilibrium Concentrations for
Steady-State Loop Operation
Unsteady-State Migration The equilibrium (balance-point) concentrations
As previously pointed out, unsteady-state cor of UF , UF , and CrF, can be estimated for
rosion arises in loops in which the salt initially steady-state loop operation by applying the six
contains UF4 but no UF, or CrF2. At zero time, conditions previously outlined to a given set
tn, the entire loop area will be a depletion zone of loop parameters and operating conditions.
(zero deposition zone), and it will remain so until Basic material data needed to carry out numerical
time t z is reached. At t z', the CrF.Z concentration calculations are (1) diffusion coefficients for
of the circulating salt will have increased to chromium in Inconel as a function of temperature,
an amount Z, where Z is the concentration of and (2) equilibrium ratios for the redox equation
CrF, in equilibrium with the chromium concen (Eq. 1) as a function of temperature. '
tration of the wall surface at the position which Tentative values of these variables are given
has the lowest temperature of the loop. The in Fig. 2„2„16. The diffusion coefficients are
balance points are then a single point at this based on extrapolations of published data and
position. As the concentration of CrF, in the the equilibrium ratios are based on an inter
circulating salt increases to values greater polation of experimental values for two NaF-
than Z, the initial balance point will develop ZrF .-UF. mixtures in contact with chromium.
into a growing deposition area bounded by two In order to present a numerical example of the
balance points. Steady state will be attained calculations involved, the following parameters
when the balance points become stationary and
P. L. Gruzin and G. B. Fedorov, Doklady Akad.
Nauk S.S.S.R. 105, 264 (1955).
This statement is based on the important assumption 48J. D. Redman, ANP Quar. Prog. Rep. June 30, 1957,
that chemical equilibrium with respect to Eq. 1 is ORNL-2340, p 132.
assumed to exist at every surface point within the
loop. Concentration gradients within the salt normal to L. G. Overholser, private communication to R. 6.
the circulation path are ignored. Evans.

108
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED be written in terms of the diffusional flow which

ORNL-LR-DWG 27049
takes place in the metal; that is, it is assumed
TEMPERATURE (°K)
1040 1000 960 that the circulating rate of the salt does not
I I
1
1 contribute to the transport equation.
(XlO

15
Condition 6 states that, for a unit area at any
loop position, L, the chromium concentration at
the surface remains constant with time. If it is
—
assumed that all chromium diffusion through a
unit area, A, into the wall is always normal to A,
the following equation may be written to relate
the chromium concentrations, C, time, t, the
diffusion coefficient, D, and penetration distance,
\d''* x:

/£--. (2) C(x,t), C.Inconel, +

+ (C surface
, - C.Inconel.) erfc
2(Dt) 1/2
and

(3) C(x,t).hot, zone

= Csurface
t
9.2 9.6 10.0 10.4 10.8
<oVr (°K) + (C,Inconel, - Csurface
. ) erfc
2(Dt)]/2
Fig. 2.2.16. Effect of Temperature on Parameters The related rate expression for both the cold and
Which Control Rate of Chromium Migration in Inconel- hot zones is
Salt Systems.
dM dC
and operating conditions, which are those of a (4) — = - DA
dt dx
surfa
typical experimental loop, were assumed: s urface

1/2
D
Salt NaF-ZrF4-UF4 (50-46-4 mole %) = A AC
Ttt
Flow rate 1 fpm

Loop Inconel tubing, 0.5 in. OD, 16 in. where AC = Csurface

. - C.Incpnel.. The signs
3
of
long, containing 16 wt % Cr dM/dt, which is the weight of chromium diffusing
Maximum temperature 1060.8 K
per unit time (in grams per second), depend on
AC; thus, a positive dM/dt means that chromium
Minimum temperature 933.3°K is diffusing into the wall. The cumulative amount
In addition it was also necessary to make some which will have migrated across any unit area of
simplifying assumptions regarding both con surface from time zero to any time t is given by
centration and temperature effects at certain
loop positions and time. The reasons for this (Dty/2 dM
are brought out in the following section, in which (5) AM = 24 AC — =2/ — .
\ 77 / dt
the diffusional flow equations for applicable
steady-state migration in loops are discussed. The chromium concentration at the tubing surface,
that is,
Mass-Transport Equations for Steady-State C surface
, = [Cr°] surface
, p . . ',
rmetal
Migration
Condition 2, given above, for steady-state mi where C and p are in density units and [Cr°] is
gration implies that the over-all migration rate can expressed in terms of weight fraction of chromium

109
ANP PROJECT PROGRESS REPORT

in the metal, is governed by the reaction available as a function of L in order to obtain

over-all values of dM/dt.
(6) Cr° + 2UF4 ^ 2UF3 + CrF2
Salt and Wall Temperature Relationship
Thus,
When a loop is externally heated, the temperature
[UF3]2[CrF2] of the circulating molten salt, T , will lag the
(7) K[Cr°] = loop wall temperature, T , with the temperature
[UFJ2 gradient being a function of the heat transfer
coefficient and heat capacity of the salt. A heat
The right-hand side of Eq. 7 is constant, according balance on the simplest model of such a system
to steady-state condition 4(c). Therefore is given by the equation
KBp[Cr°]Bp = KjCr°]L , dT
(ID B(T - T ) = 0
where the subscript BP refers to balance point dt
and L refers to a unit area at any loop position, or where

K 2b
BP
B =
(8) AC = "metal
p . . [Cr°], .
v ' Jlnconel
K, Pcpr
The rate of diffusion as a function of time at any the quantities h, C , and p are the heat transfer
position L may then be obtained by combining coefficient, specific heat, and density of the salt,
Eqs. 4 and 8 to give and r is the inner radius of the tubing.
An equation relating T , tubing length, L,
dM
(9) = Ap [Cr velocity of circulating salt, u, and time, t, must
Inconel
~d~t metal
be obtained in order to solve Eq. 11 under steady-
state conditions. Actually all that must be
1/2
BP assumed is the shape of the plot of T vs loop
- 1 position. It is not at all unreasonable to assume
K,
that the equation of the plot is
In order to obtain the over-all rate at which the (12) Tw = T mean + a sin cot ',
chromium is being transferred, it is necessary to
where a is one-half the loop AT, that is, 0.5
integrate the instantaneous rates over an entire
(1060.8 - 933.3) °K, and co is 2n times the flow
hot or cold zone, as follows:
velocity divided by the length of the loop.
dM
1/2 The salt temperature, which may be obtained by
(10) 2rpr metal
, . [Cr°] \ -. combining Eqs. 11 and 12, is
dt
Ba
T = T sin (cot — 4>)
K 2x1/2
BP (Bz + co1)
/ - 1 D]/2dL ,
where

co

where r is the radius of the Inconel tube. If cf> = tan"

B
TL < TBp, KL < KBp, and dM/dt > 1, chromium
is diffusing into the wall in the cold zone; the A rplot of Tw and T s is shown in Fiq.
3
2.2.17.
converse is true for the hot zone. The physical properties of the salt, NaF-ZrF -UF
The integration indicated in Eq. 10 cannot be (50-46-4 mole %), give a value for B that is very
performed until the K, and D vs L relation nearly equal to the value of co for the loop de
ships are known. Two important parameters are scribed above. Therefore <j> was chosen to be
involved: (1) the wall temperature, which controls 45 deg. A rotating vector representation of the
D , and ^2) the metal-salt interface temperature, Tw and T s relationships
r
is also shown in Fiq.
3
which controls K, . These temperatures must be 2.2.17. After the graph was prepared, L was

110
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 27050

/. = 12 in.

UJ
Q
co
Q L = 16 in. 37T
UJ
_J Z. = 0 4
o
O
o

oj/= 0

L = 4 in.

0 2 4 6 8 10 12 14 16
LOOP LENGTH (in.)

7T 37T
77 2TT
2 2
0J/ (radians)

Fig. 2.2.17. Temperature-Time Relationships for Idealized Thermal-Convection Loop.

Ill
ANP PROJECT PROGRESS REPORT

redefined or, rather, moved back one-fourth the

loop length to give a T
metrical about L/2.
curve which was sym (Area)2 =JrBPl Dj" ^—
fKBP
- lj\ dL,
1
The equilibrium ratios, K, are applicable only
when the wall and salt are at thermal equilibrium, cm2/sec'/2
that is, T = T , and it may be seen in Fig. 2.2.17
that the wall and salt are at thermal equilibrium K
BP
- 1 ) dL,
only at the two points which correspond to the
maximum and minimum salt temperatures. To
(Area),
•/ BP2 ""' \ L
circumvent this difficulty, three assumptions may
be made regarding the surface temperature, which cm2/sec,/2
governs the value of K: If L =0 is taken at the center of the cooled
1. At x = 0, a thin layer of metal is at thermal section, as shown in Fig. 2.2.17, then
and chemical equilibrium with the bulk of the
salt. The equilibrium temperature is T . 1/2
2. At x = 0, a thin layer of salt is at thermal
and chemical equilibrium with the bulk of the
t\, ' hot zone
-^°\t (Area).

metal. The equilibrium temperature is T . Likewise,

3. The equilibrium temperature is at some value
between T and T . 1/2
s w dM
In line with the above discussion, it is in 2rpJG°\\-
~di
teresting to note the effect of increasing the cold zone

circulation rate, u, on Ts while holding Tw x [(Area), + (Area)3] .

constant. The two variables h and co will in
crease, with the rate of increase of co being the Since the CrF, content of the salt remains fixed,
greater. This means that the ratio co/B will (Area), must equal (Area). + (Area),. If the areas
increase, which, in turn, means that cf> will in are not equal, new balance points must be chosen
crease and the maximum variation of T will and the entire procedure must be repeated until
decrease. Thus the chemical equilibrium ratios this condition is satisfied.
are most applicable at low flow rates, since the The results of calculations performed as out
differences between T 5 and T w increase as the lined above are shown on Fig. 2.2.18, where the
flow rate increases. solid curve represents the results based on
All the information required to obtain an ap assumption 1 and the dotted curve represents the
proximate estimate of the rate and amount of net results based on assumption 2. Values of the
chromium transport within a particular loop is over-all transfer rate and cumulative amount trans
given on Figs. 2.2.16 and 2.2.17. The mechanics ferred as a function of time are shown on Fig.
of the calculations are given below:
UNCLASSIFIED
1. Choose one of the three assumptions given (X10 ") ORNL-LR-DWG 27051
+10
above regarding the surface temperature.
2. Arbitrarily choose a balance point. o +5
3. Tabulate the factors [(KBp/KL) - 1] and /

D for various values of L.

E
\
\
\ \l+# (INTO W U_L) /
4. Integrate, by graphical means, the area under !-#< DUT OF W ALL) \ /
/

\ / /
a curve representing the product of the above
I ^allJJ?-0
™ 'salt
factors as a function of L. =Sk \
i

\ \ i
The results will be values of the following \
\ \—IS — 'salt]X-0" Twall]
X—- a

\
expressions: /

2 4 6 8 10 12 44 16

(Area).
-/.""•Hir 1 dL,

Fig. 2.2.18.
LOOP POSITION, L (in.)

Product of Diffusion-Controlling Paran

cm /sec.1/2 eters as a Function of Loop Position.

112
 PERIOD ENDING DECEMBER 31, 1957

2.2.19. The curves on Fig. 2.2.19 are tentative; as the ternary system nickel-molybdenum-chromium,
however, they strongly suggest that the amount which includes the INOR alloys developed by
of steady-state chromium migration within actual the ORNL Metallurgy Division.
loops (similar to the type considered) will be very Measurements of the activity of nickel in Ni-Mo
small. alloys are being made by using the emf method
described in the previous report. The cell
ACTIVITIES IN ALLOYS
being used for the measurements is designed
S. Longer as shown in Fig. 2.2.20 to permit the tight gas-
keting required to maintain an inert atmosphere
Measurements of the activities of the metallic
components of container materials for fused-salt over the liquid metal bath in which the cell is
fuels have been made in order to study the effects immersed. The liquid metal bath was found to
be necessary to prevent a large temperature
of composition and temperature on the activities
and thus on the corrosion resistance of container
gradient along the electrodes. A tin bath was
alloys. The activity of chromium in nickel- unsatisfactory because a hard, sintered mass of
chromium alloys (Inconel) has been studied, oxide formed on the surface and frequently cracked
and data are presented here for the nickel-molyb the cells. Liquid bismuth baths are quite sat
isfactory, and the small amount of oxide that
denum system. In future work the chromium-
forms is not deleterious to the cell.
molybdenum system is to be studied, as well

50M. 3. Panish, ANP Quar: Prog. Rep. Sept. 10, 1956, 5,S. Longer, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2157, p 100. ORNL-2387, p 130.

ORNL-LR-DWG 27052

6 8
LOOP OPERATING TIME (months)

Fig. 2.2.19. Chromium Migration Characteristics for a Pre-equilibrated Thermal-Convection Loop.

113
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27053

THERMOCOUPLE WELL ^ f

J TO HELIUM SUPPLY
TUNGSTEN-GLASS SEAL

SILVER-SOLDERED TUNGSTEN-NICKEL JOINT

CROSS SECTION OF CELL

ELECTRODE -

;U
GROUND FUSED SILICA PACKING "

Fig. 2.2.20. EMF Cell for Measuring Alloy Activities.

114
 PERIOD ENDING DECEMBER 31, 1957

Only preliminary data on two alloy compositions process. The choice of a solvent salt was based
are available thus far. The activity of nickel in an on the ability of the salt to dissolve or dis
alloy containing 90 mole % Ni is 0.70 at 800°C. integrate a set minimum quantity of stainless
This corresponds to an activity coefficient of steel fuel elements to yield a concentration of
0.76. The behavior of this alloy in the cell uranium in the molten salt which would be
appears to be quite reproducible. An alloy con economically feasible for adaptation in the
taining 80 mole % Ni has a nickel activity of volatility process. From previous experience
about 0.32 at 800°C The activity coefficient in handling various fluoride mixtures, the alkali
in this case is about 0.40. The data on this fluoride compositions containing NaF, LiF, and
alloy have not yet been reproduced to within KF seemed to be most promising. The composition
the desired precision. chosen was NaF-LiF-KF (11.5-46.5-42 mole %),
The preliminary data seem to indicate large which combined relatively high solubilities of
negative deviations from Raoult s law. From the structural metals nickel, chromium, and iron
inspection of the nickel-molybdenum phase dia with a fairly low melting point (460°C).
gram,52 this is to be expected. The 90 mole %Ni For the first experiment, which was run to
alloy is in the solid-solution region, and the rather gain knowledge of the general characteristics of
large negative deviation should not be surprising the dissolution process, a 2-kg batch of the
in view of the proximity of the compound MoNi4. chosen fluoride mixture was purified according
The 80 mole % Ni alloy might have been expected to normal procedures in a copper-lined reactor
to exhibit even greater deviation from ideality, vessel. The reactor vessel was equipped with
since it is within the region of stability of the copper thermocouple wells and dip lines, so
compound at 800°C. that copper was the only metal that contacted
It is expected that the precision and accuracy the molten salt. After the purification and
of the preliminary data will be improved by meas sampling steps, a type 347 stainless steel plate
urements now being made. Measurements are 4 in. square and 0.0625 in. thick, which weighed
to be made of the activities of alloys containing approximately 71 g, was suspended in the melt
as high a molybdenum content as is metallurgically attached to a copper rod. The temperature was
feasible over a temperature range between 700 and maintained at 650°C, and anhydrous HF was
1000°C. Measurements over this temperature range bubbled through the molten salt at approximately
will permit the determination of the temperature- 1 liter/min. Hydrofluorination was continued for
dependent thermodynamic properties, and the data a period of 24 hr, and salt samples were taken
obtained at temperatures above 900°C will have every 4 hr. A sample of the stainless steel
additional interest, since the 900 to 1000°C was submitted for analysis after the 24-hr period,
temperature range is above the region of stability and the following report of the analysis was
of both the compound MoNi4 and the higher- received:
molybdenum compound MoNij. In this region it
would be expected that the activity of nickel Component Amount (wt %)
would be more nearly ideal. Ni 10.3

Cr 17.6
DISSIOLUTION OF STAINLESS STEEL FUEL
Fe 69.8
ELEMENTS IN FLUORIDE MELTS
Nb 1.72
J. E. Eorgan G. J. Nessle
The possibility of dissolving stainless steel The results obtained from analyses of the seven
fuel elements in molten fluorides, with the re salt samples are compared in Table 2.2.23 with
sultant advantage of adaptability to the present the theoretical concentrations in the salt if the
volatility process, made it desirable to do some structural metals in the type 347 stainless steel
exploratory experiments to determine the feasi plate had completely dissolved.
bility and general characteristics of such a At the conclusion of the experiment the molten
salt was transferred from the reactor vessel so
American Society for Metals, Metals Handbook, that the copper liner could be examined. It was
1948 ed., p 1230, American Society for Metals,
Cleveland, 1948. found by examination of the cooled melt and the

115
ANP PROJECT PROGRESS REPORT

Table 2.2.23. Results of First Experimental Study of the Dissolution of Type 347 Stainless Steel
in NaF-KF-LiF Eutectic by Hydrofluorination

Metal Concentrations in Salt Samples

Period of
Sample Hydrofluorination Ni Cr Fe
No.
(hr) ppm ppm ppm

Type 347 stain ess 7.3 12.5 49.6

steel*

1 0 40 30 280

2 4 2955 5.9 720 1.4 9,620 19.2

3 8 3465 6.9 1280 2.5 10,795 21.6

4 12 3840 7.7 1850 3.6 12,375 24.7

5 16 3900 7.8 1410 2.8 9,620 19.2

6 20 3730 7.5 960 1.9 7,830 15.6

7 24 4005 8.0 1490 3.0 9,720 19.4

'Concentrations given are calculated values based on complete dissolution of sample in salt.

reactor vessel that the steel plate had completely through a sintered nickel filter with 0.0015-in.
dissolved or disintegrated and that some odd pores.
riodule-type salt formations had remained in
the reactor, attached to the support rod and the The nodule type of formation occurred again on
dip leg at about the liquid level. These formations the steel plate support rod and the dip leg. A
appeared to be water soluble, and no metal rough material balance is given below:
particles were visible. From the results given
Amount (g)
in Table 2.2.23, it seemed that the plate was
completely dissolved at the end of 12 hr, since Type 347 stainless steel plate 71
metal concentrations did not increase after this Total salt charge 2000
point. All the nickel was accounted for in the Transferable salt 1657
bulk of the transferable melt, but only half the Filter-stick samples (3) 67

iron and less than one-fourth the chromium could Nontransferable salt left 347 (by difference)
in reactor
be found.
The first experiment established that dis
Samples of each item were submitted for chemical
solution of the steel plate was possible, and
analysis to determine the distribution of the
therefore a second experiment was run to in
nickel, chromium, and iron, and the results are
vestigate further the disposition of the structural
presented in Table 2.2.24.
metals in the molten mixture. The conditions of
the first experiment were duplicated as closely Accurate weights of the salt sections that
as possible, but only three filter-stick salt made up sample 4 were difficult to obtain, be
samples were taken, on the assumption that cause these nodules had to be broken off the
dissolution of the steel plate would occur as support rods and the sides of the copper liner.
it had in the first experiment. An attempt was The weights were sufficiently accurate, however,
made to obtain a closer material balance in order that by making some liberal assumptions a rough
to give a better idea as to the disposition of material balance was obtained. Samples 2 and
the dissolved metal. At the completion of the 3 were assumed to be the most homogeneous
experiment the salt batch was transferred from samples as far as nickel concentration was con
the reaction vessel into a tared receiver can cerned, and the analyses were in good agreement

116
 PERIOD ENDING DECEMBER 31, 7957

Table 2.2.24. Results of Second Experimental Study of the Dissolution of Type 347 Stainless Steel
in NaF-KF-LiF Eutectic by Hydrofluorinption

Metal Concentrations in Salt Samples

Sample
Sample Description Ni Cr Fe
No.
ppm

Filter stick samples (67 g)

1 Taken before hydrofluorination 2300* 4.6 0.025 0.5 0.058 1.1

2 Taken after 12 hr of hydrofluorination 3700 7.4 0.18 3.6 2.1* 41.8

3 Taken after 24 hr of hydrofluorination 3600 7.2 0.26 5.2 1.36 27.2

4 Nondule-type formations on support rod 2300 0.7 3.25 9.0 9.6 26.4
and dip leg (~347 g)

5 Transferable material collected in re 9100* 15.0 0.42 6.9 1.70 28.2

ceiver can (1657 g)

* Analyses were questionable but were not rechecked.

with those of similar samples from the first ex Of greatest concern in the dissolution process
periment. A rough material balance of the struc was the possibility of concentrating the uranium
tural metals in the fluoride salt is given in Table in the nontransferable, that is, high-melting,
2.2.25, and fairly close agreement with theoretical phase of the salt bath during dissolution of the
values may be seen. steel fuel element. Previous experiments had
In order to determine the effect of uranium on shown that such phases could occur.
the dissolution of type 347 stainless steel fuel In the first of the two experiments, a steel
elements in an alkali metal fluoride salt such
plate 4 in. square and '/16 in. thick, which weighed
as NaF-LiF-KF (11.5-46.5-42 mole %), two ex 71 g, was dissolved in 2 kg of molten salt at
periments were run in which the uranium-bearing
1200°F by bubbling HF gas through the bath at
salt NaF-LiF-KF-UF4 (10.9-44.5-43.5-1.1 mole %) approximately 1 liter/min. The dissolution process
was used as the solvent. In a practical ap
was complete at the end of 12 hr, according to
plication the uranium would come from the fuel analytical results. At the end of 24 hr the melt
elements.
was transferred through a sintered nickel filter
Table 2.2.25. Material Balance of Structural Metals
into a storage can. Approximately 300 g of
from the Dissolution of Type 347 Stainless Steel Plate material remained in the reactor vessel as the
in NaF-LiF-KF Eutectic
nontransferable high-melting phase. The uranium
analysis of the salt bath prior to dissolution of
Weight the stainless steel was 6.3%. The analysis of
Metals Found
Sample of Material
the transferable salt showed a uranium content
in Salt (g)
No. Analyzed of 6.5%, and the nontransferable salt contained
Ni Cr Fe 2.4% U. The chromium and iron tended to con
(g)
centrate in the nontransferable phase. In general,
therefore, the presence of uranium seemed to
3 1724 6.4 4.5 23.4
have little effect, if any, on the dissolution
4 347 0.8 11.3 33.3 process, and the uranium showed no tendency
Totals 2071 7.2 15.8 56.7
to concentrate in the nontransferable phase.

Theoretical val ues for 71 7.3 12.5 49.5

In the second experiment the conditions de
type 347 stain less scribed above were duplicated, except that a
steel
piece of stainless steel 4 x 12 x V, in., which

117
ANP PROJECT PROGRESS REPORT

weighed 213 g, was used. In this case dis For the first experiments, which were performed
solution again seemed to be complete at the at 25 to 30°C, the equipment consisted of a
end of 12 hr, according to chemical analysis. glass capsule 6 in. long and 1 in. in diameter,
Attempts to transfer the salt bath through a with a sintered glass disk installed midway in
sintered nickel filter were fruitless, however, the capsule. A similar capsule that was con
even though the salt temperature was raised to structed of nickel was used at elevated tem
1400°C in 50-deg increments. The filter was peratures. The oxidation was accomplished by
finally removed, and the salt was transferred the introduction of measured volumes of dry
through tubing. Approximately 740 g of salt air. The oxides were separated from the bulk
remained in the reactor vessel. This nontrans of the liquid metals by filtration, and the re
ferable phase appeared to be crystalline after maining liquid metals were removed by amal
freezing and could be described as mushy at high gamation.
temperature. The nickel again appeared to remain The early runs, in which Na-K (50-50 mole %)
completely in solution, there being little difference was used, gave a value of 25 for the Na-to-K
in nickel concentration between the transferable ratio in the mixed oxides formed at 30 to 35°C.
and nontransferable phases. Small filter samples No significant difference was found at 200°C
taken at 6-hr intervals prior to transfer of the if the oxides were removed immediately, but,
bath gave nickel concentrations close to those if they were formed at 200°C and equilibrated
found in the bulk phases. The iron and chromium for 40 hr at 400°C, the ratio was found to be
again concentrated in the nontransferable phase less than 10.
and were present to a fair extent as metal crystals.
The early experiments were not precise, and
The uranium analysis showed 6.5% in the original
bath, 6.7% in the transferable phase, and 3.3% in
therefore a more thorough study of the oxidation
the nontransferable phase. was made at 30 to 35°C in an attempt to find
In general it appears that the stainless steel conditions that would yield more precise results.
concentration used in this experiment was too The glass apparatus was modified to provide
high, although the dissolution or disintegration a larger surface for oxide formation, -and all
process was not appreciably affected. If less operations were carried out in a vacuum dry
stainless steel were used, a practical process box in order to reduce contamination by water
could be evolved for removing the uranium from vapor. The oxidation of two compositions of
the steel and leaving a constant and nonaccu- the Na-K mixture was studied under the modified

mulative heel in the reactor vessel. This heel conditions, and the results are presented in
could be flushed out periodically with a dummy Table 2.2.26.

salt mixture if desired. It is evident from the lack of precision in the

data that a close control of the variables involved
OXIDATION OF MIXTURES COMPOSED is absolutely essential. It is appafent, however,
OF SODIUM AND POTASSIUM that the necessary control is extremely difficult
E. E. Ketchen to accomplish. Traces of moisture, which is
the most likely contaminant, should tend to
A study of the oxidation of sodium and potas reduce the Na-to-K ratio, and therefore the larger
sium mixtures under carefully controlled conditions values are probably closest to the correct ratio.
has been made in an attempt to determine the The lack of precision in these experiments makes
effect of certain variables on the final composition it difficult to establish any quantitative effect
of the mixed oxides. The variables studied in of composition of the alkali metal mixture on
clude (1) the original composition of sodium- the Na-to-K ratio in the mixed oxides. The data,
potassium mixture, (2) the temperature of oxidation, however, show a significant decrease in the ratio
and (3) the temperature and time of equilibration upon an increase in the K content from 49 to
following the oxidation. 77 wt %.

118
 PERIOD ENDING DECEMBER 31, 1957

Table 2.2.26. Composition of Oxides Formed by Oxidation of Sodium-Potassium Mixtures at 30 to 35°C

Com Position of Oxides Con- position of Oxides

Composition of Composition of
Na-K Mixtures Na K Na-to-K Na-K Mixtures No K Na-to-K
(wt %) (mg) (mg) Ratio (wt %) (mg) (mg) Ratio

51-49 13.8 0.17 82 23-77 8.1 0.27 30

6.7 0.089 75 7.4 0.23 32

5.3 0.058 91 8.6 0.31 28

4.1 0.085 48 4.7 0.21 22

4.3 0.049 88 5.9 0.30 20

11.6 0.13 89 4.4 0.31 14

10.4 0.21 50 8.5 0.31 27

8.5 0.15 56 8.1 0.31 26

7.7 0.11 70 6.0 0.23 26

4.0 0.24 17 5.4 0.19 28

6.6 0.14 47 5.6 0.30 19

21.0 0.37 57 4.9 0.36 14

7.5 0.22 34
Av 64 ± 22
18.0 0.48 37

Av 25 ± 7

119
ANP PROJECT PROGRESS REPORT

2.3. PHYSICAL PROPERTIES OF MOLTEN MATERIALS

F. F. Blankenship

SUMMARY OF ARE FUEL VAPOR-PRESSURE The ZrF4 used for vapor pressure measurements
STUDIES and for the preparation of purified mixtures was
material which chemical analysis showed to be
R. E. Moore
98 to 99% pure. Uranium tetrafluoride was ob
The vapor pressures of a number of fused salt tained from the Mallinckrodt Chemical Works.
mixtures in the NaF-ZrF4 and NaF-ZrF4-UF4 The NaF used was reagent-grade material. The
systems were measured to obtain information NaF-ZrF4 and NaF-ZrF4-UF4 mixtures were
needed for the selection of a fuel for the Aircraft prepared by fusion under an inert atmosphere,
Reactor Experiment. A quasi-static method of followed in every case, except that of the compo
vapor pressure measurement, which originated sition 75 mole %ZrF4-25 mole %NaF, by treatment
with Rodebush and Dixon1 and was later used by with HF, H-, and inert gas, in the order stated.
Fiock and Rodebush,2 was chosen for this work A summary of vapor pressure data for pure ZrF4
because it appeared to be the most reliable method and mixtures of ZrF4 qnd NaF is given in Table
for use at the temperatures and pressure range 2.3.1, and a summary of the data for UF4-bearing
(3-200 mm Hg) considered. The apparatus and mixtures related to the ARE fuel or fuel concen
technique have been described by Beusman.3 trates is given in Table 2.3.2. References in the
tables indicate the ANP progress report in which
the data are reported. In some cases, as indicated,
]W. H. Rodebush and A. L. Dixon, Phys. Rev. 26, the data were refined or extended later.
851 (1925).
2E. F. Fiock and W. H. Rodebush, /. Am. Chem. Soc. Petrographic examinations4 of samples distilled
48. 2522(1926). from binary mixtures containing more than 43
3
C. Beusman, Activities in the KCl-FeCl-^ and LiCl-
FeCl2 Systems, ORNL-2323 (May 15, 1957). 'Made by T. N. McVay, Consultant, ORNL.

Table 2.3.1. Vapor Pressures of NaF-ZrF4 Mixtures and of ZrF4

Constants for the

Vapor Pressure
ZrF Content Temperature Range
Equation* Reference
(mole %) (°C)
A B

43 868-1034 7,289 7.34 ORNL-1609, p 106

47 795-994 7,206 7.62 ORNL-1649, p 99**

50 812-990 7,675 8.14 ORNL-1649, p 99**

57.8 732-906 8,250 9.00 ORNL-1609, p 106

75 754-844 9,369 10.57 ORNL-1771, p 129

100 713-839 10,936 12.11 ORNL-1375, P 147**

*The constants, A and B, are for the general equation

•og10MmmHg) =~J^ +B'

**The data given were revised and extended from those given in the ANP quarterly progress report cited.

120
 PERIOD ENDING DECEMBER 31. 1957

Table 2.3.2. Vapor Pressures of NaF-ZrF4-UF4 Mixtures

Constants for the

Compos ition (mo e %) Temperature
Mixture Vapor Pressure Equation*
Range Reference
Number NaF
ZrF4 UF4 (°C) A B

27 40 50 4 807-968 7779 8.281 ORNL-1375, p 148**

30 50 46 4 807-979 7551 7.888 ORNL-1375, p 148**

33 50 25 25 848-1096 6906 6.844 ORNL-1515, P 159

39 65 15 20 1166-1266 6944 5.86 ORNL-1609, p 106

40 53 43 4 790-1012 7106 7.37 ORNL-1556, p 88**

*The constants, A and B, are for the general equation

A
log1Q p (mm Hg) = - " 0 ' + B •
**The data given were revised and extended from those given in the ANP quarterly progress report cited.

more %ZrF4 show them to be virtually pure ZrF4«

ORNL-LR-DWG 27054
Thus the total pressures are approximately 300

equivalent to the partial pressures of ZrF4> As

may be seen in Fig. 2.3.1, which is a plot of
pressure vs composition at 912°C obtained from 700

the data in Table 2.3.1, large negative deviations

from ideal behavior occur. Thus it appears that
complexing exists in the solution. 600

The actual ARE fuel composition, 53.09 mole %

^ 500
NaF-40.73 mole % ZrF4-6.18 mole % UF4, was E
E
determined by titration to criticality, and its vapor
pressure was not measured. The vapor pressures v/
400
of compositions in the neighborhood of the ARE $
fuel composition were low enough, however, that
operational difficulties traceable to ZrF4 vapor
were neither expected nor encountered in the ARE.

Estimated vapor pressures for a number of other 200

ZrF4-bearing mixtures containing UF4 were given

in a recent report.5 The composition given that
is nearest to that of the actual ARE fuel contains 100

6.5 mole % UF4_40 mole % ZrF4- Its vapor

pressure at the fuel outlet temperature (1522°F)
is about 8 mm Hg. 10 20 30 40 50 60 70 80 90 100
ZrF. (mole 7o)

Fig. 2.3.1. Vapor Pressures of NaF-ZrF. Mixtures at

5S. Cantor, ANP Quar. Prog. Rep. June 30. 1957.
ORNL-2340, p 158. 912°C.

121
ANP PROJECT PROGRESS REPORT

VAPOR PRESSURES OF MOLTEN FLUORIDE for which the transpiration or carrier-gas technique
MIXTURES was used. The results of the BMI research are

S. Cantor discussed here in relation to the work done at

ORNL.
Vapor pressure studies have provided, as part The binary systems studied in both laboratories,
of the search for optimum fuels, directly de along with the references which most fully describe
terminable physical property data, as well as a the results, are given in Table 2.3.3. In most
measure of the thermodynamic activities of com cases it was necessary to measure the vapor
ponents in fuel mixtures. The main objective of pressure of the pure components in order to be
the studies was to obtain the vapor pressures of able to calculate the activities. The references
binary fuel solvents, such as the mixtures of for these measurements are given in Table 2.3.4.
ZrF4 with alkali fluorides. It could be assumed For the ZrF4 systems, it was usually found that
that the addition of small amounts of fuel (2—4 sharp decreases in ZrF4 vapor pressures occurred
mole % UF ) to the solution would not markedly as the alkali fluoride concentrations were in
alter the solution structure and that the vapor creased. Also, there was a correlation between
pressure would change in accordance with Raoult's decreasing ZrF4 vapor pressure and alkali cation
law.
size. The purpose of studying the 7NaF-6ZrF4-
As stated in the preceding section, a quasi-static 7NaF'6UF4 system was to investigate the effect
method developed by Rodebush and Dixon was of partiaj substitution of UF4 for' ZrF4 while
used to measure vapor pressures. The compo maintaining a constant NaF concentration. The
sitions to be studied were prepared by weighing results clearly indicated that this substitution
and mixing the purified components in a dry box. decreased the vapor pressure, but to a far less
The o4WLte*4luorides were purified by fusing, degree than increasing the amount of alkali
slowly cooling the melt, and selecting clear crystal fluoride.
fragments. The commercially obtained ZrF4 was For the compositions studied at ORNL, it was
further purified by vacuum sublimation of the best at first assumed that the measured pressures were
hafnium-free material. All other pure components due to ZrF4 alone. Above about 40 mole %ZrF4,
and mixtures used were produced at the Laboratory. this assumption, tested by sampling of the vapor,
It is important to note here that there was a proved to be quite valid. At lower liquid-phase
a concurrent effort to measure vapor pressures concentrations of ZrF4, the probable formation of
being carried out at Battelle Memorial Institute vapor-phase compounds made the assumption

Table 2.3.3. Binary Solvent Systems Whose Vapor Pressures Were Measured as Part of the ANP Program

System Reference

LiF-ZrF, Vapor Pressures of the Rubidium Fluoride—Zirconium Fluoride and

Lithium Fluoride—Zirconium Fluoride Systems, BMI-1199 (June
27, 1957)

NaF-ZrF, Vapor Pressures of the Sodium Fluoride —Zirconium Fluoride

System and Derived Information, BMI-1064 (Jan. 9, 1956)

KF-ZrF. ANP Quar. Prog. Rep. June 10. 1956, ORNL-2106, P 111

RbF-ZrF. Vapor Pressures of the Rubidium Fluoride—Zirconium Fluoride

and Lithium Fluoride—Zirconium Fluoride Systems, BMI-1199
(June 27, 1957)

7NaF-6ZrF4-7NaF-6UF4 ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL-2221, p 146

NaF-BeF^ Vapor Pressure and Equilibrium Studies of the Sodium Fluoride-

Beryllium Fluoride Systems, BMI-1186 (May 27, 1957)

122
 PERIOD ENDING DECEMBER 31, 7957

Table 2.3.4. References for Vapor Pressures of Pure Components

Compound Reference

ZrF, Vapor Pressures of the Sodium Fluoride—Zirconium Fluoride

System and Derived Information, BMI-1064, P 7 (Jan. 9, 1956)

UF4 ANP Quar. Prog. Rep. June 10, 1955, ORNL-1896, p 65

BeF„ Vapor Pressure and Equilibrium Studies of the Sodium Fluoride-

Beryllium Fluoride Systems, BMI-1186, p 3 (May 27, 1957)

LiF, NaF, KF, RbF, CsF ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 128

NaF Vapor Pressures of the Sodium Fluoride—Zirconium Fluoride

System and Derived Information, BMI-1064, p 7 (Jan. 9, 1956)

LiF, RbF Vapor Pressures of the Rubidium Fluoride—Zirconium Fluoride

and Lithium Fluoride—Zirconium Fluoride Systems, BMI-1199,
p 3 (June 27, 1957)

poorer. The vapor compounds were first postulated Equation 2, in turn, was used to calculate alkali
by Sense0 in order to explain apparent NaF partial fluoride activities.7,8 The alkali fluoride activities
pressures obtained in the NaF-ZrF4 system. could be considered to be only approximate,
Vapor compounds also had to be postulated in the because it was necessary to extrapolate Eq. 1 to
LiF-ZrF4, RbF-ZrF4, and NaF-BeF2 systems if iV_ = 1 in order to solve the Gibbs-Duhem equation.
the thermodynamic stability of these systems was The results of these calculations showed that the
to be preserved. Because the vapor above the alkali fluoride activity decreased sharply with
molten salt solutions under study at ORNL increasing ZrF4 concentration.
(RbF-ZrF4, KF-ZrF4, NaF-ZrF4) was virtually all Once a large body of vapor pressure data was
ZrF„ it was still possible to calculate ZrF4 accumulated for molten mixtures of ZrF4 and
activities7'8 directly from total pressure measure alkali fluorides, estimates of vapor pressures
ments. These activities were then used to form were made for a number of unmeasured compo
the analytical expression sitions of ZrF4-base fuels.9 In these estimates,
the vapor pressures of ZrF4 were assumed to be
(1) log y, = AN2 + BN^ , linear functions of composition over the appropriate
extrapolation or interpolation range. The vapor
where y. is the activity coefficient of ZrF4, N2
is the mole fraction of alkali fluoride, and A and B pressures of fuel mixtures composed of UF4 and
alkali fluorides were too small to measure by the
are constants. This expression was used in the
methods used, but, with the vapor pressure known
Gibbs-Duhem equation, in the form
for each pure fluoride used in the melt, estimates
(2) /V1 d log y, + N2 d log y2 = 0 , of the upper limits of vapor pressure were compiled'
by. assuming that Raoult's law was obeyed by the
where N} is the mole fraction of ZrF. and y2 is alkali fluorides and that Henry's law was obeyed
the activity coefficient of the alkali fluoride. by UF4.
Any continuation of vapor pressure investi
gations of fused fluorides of reactor significance
K. A. Sense et al., Vapor Pressures of the Sodium should involve systems containing BeF_. One
Fluoride—Zirconium Fluoride System and Derived In
formation. BMI-1064 (Jan. 9, 1956).
7S. Cantor and S. D. Christian, ANP Quar. Prog. Rep.
Sept. 10, 1956. ORNL-2157, p 108.
8S. Cantor, ANP Quar. Prog. Rep. Dec. 31, 1956, VS. Cantor, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2221, p 145. ORNL-2340, p 158.

123
ANP PROJECT PROGRESS REPORT

such important system that is amenable to measure A molecular-beam mass spectrometric study,
ment by the methods now in use or under develop which continued the work of Rosenstock et a/.,13
ment is LiF-BeF2. The determination of activities was undertaken in order to develop a method of
in BeF„ systems presents a greater problem than exploring the vapor-phase associations so that
for the corresponding ZrF4 systems. The difficulty the occurrence of the phenomenon and its effects
that arises is the inability to equate BeF_ vapor could be predicted. The mass spectrometer
pressures to total vapor pressures except at higher provides a measure of the ionized progeny of
BeF2 concentrations (75 mole % or more). In individual associated vapor species. With this
compositions below these higher concentrations, technique, the individual vapor species which
vapor compounds10 start contributing sizably to were the precusors of the observed ions can be
the total pressure. Therefore, it will be necessary inferred, and by control and measurement of the
to obtain partial pressures of each component in temperature of the vapor source the internal energy
order to calculate activities. At present, the of vaporization and of association of the detectable
carrier-gas technique is being developed to aid species can be measured. The mass spectrometer
in measuring the partial pressures. The transpi is uniquely suited to measuring" these energies of
ration vapor pressures are not sufficient to be association.
useful in obtaining the activities of the alkali Potassium chloride was chosen as a suitable
fluorides in these mixtures. Recent evidence material with which to develop the method, since
convincingly shows that there are appreciable prior studies of that compound had been re
quantities of dimeric and even trimeric molecules ported.3*11'13 It was incompletely studied by
in the vapor of most alkali halides.11 Since the Rosenstock by using a hot filament as a source of
vapor pressures obtained in the transpiration ions in the mass spectrometer. This method yields
method are dependent on the molecular weight of only relative energies of vaporization of the various
the vapor, total pressures will still have to be vapor species. The ion species observed were
measured in order to help obtain the partial K+ and KCI+, which are derived from monomeric
pressures of the species involved. KCI, primarily, and K Cl+, which is derived from
a dimeric species, K2Q2. The existence of these
A MASS SPECTROMETRIC STUDY OF
species agrees, in general, with the findings of
SALT VAPORS
Miller and Kusch11 and of Beusman.3
L. 0. Gilpatrick M. Blander The mass spectrometer used in this study was a
The vapors of some fused salts have been 6-in.-radius 60-deg Nier-type machine with a metal
shown, as stated above (and in ref 12), to contain envelope; it was evacuated by a mercury diffusion
double and triple molecular species. An under pump.14 The accelerating voltage, 2000 v, was
standing of the fundamental nature of these supplied by a well-regulated source capable of
associated species in the vapor is necessary for delivering a current of 2 ma. The ion-beam
the interpretation of the vapor pressure data of intensity was measured by a vibrating-reed
fuel mixtures in terms of the activities of the electrometer. The output from the electrometer was
volatile components. It is also necessary for the recorded on a 0 to 10 mv strip-chart recorder,
estimation of the volatility of components of fused which produced a permanent record of the ion-beam
salt mixtures. Polymer formation can cause intensities.

components of salt mixtures to be even more The effusion cell used to produce the molecular
volatile as part of an associated species than the beam of salt molecules, as finally modified, is
pure component is, in some cases. shown in Fig. 2.3.2. The cell was heated from
the 110-v main by means of a voltage regulator
a'nd a variable autotransformer. This arrangement
10K. A. Sense, R. W. Stone, and R. B. Filbert, Jr.,
Vapor Pressure and Equilibrium Studies of the Sodium 13
Fluoride-Beryllium Fluoride System, BMI-1186 (May 27, H. Rosenstock et al.. Mass Spectra Studies of Silver
1957). Halides and Alkyl Halides, paper No. 34 presented at
the 4th Annual Meeting of the American Society for
R. C. Miller and P. Kusch, /. Chem. Phys. 25. Testing Materials, Committee E-14on Mass Spectrometry,
860 (1956). Cincinnati, May 28, 1956.
2S. A. Ochs, R. E. Cote, and P. Kusch, J. Chem. G. F. Wells, Stable Isotope Research and Production
Phys. 21,459 (1953). Semiann. Prog. Rep. May 20, 1954. ORNL-1732, p 27.

124
 PERIOD ENDING DECEMBER 31, 7957

UNCLASSIFIED
ORNL-LR-DWG 26649

(CELL EXIT SLIT ( 1.8 mil WIDE) RADIATION SHIELD TOP

RADIATION SHIELD CYLINDER

TOP THERMOCOUPLE
Pt, 10% Rd)

NICKEL

SALT SAMPLE
COMPARTMENT

THERMOCOUPLE WELLS

RADIATION SHIELD
-HEATER COILS
FLOOR

CROSS SECTION OF CELL

FACE PLATE

KOVAR-TO-GLASS SEALS

Fig. 2.3.2. Effusion Cell.

proved to be sufficiently stable as a source of by at least 3°C, to prevent the deposition of salt
heater voltage for temperatures to be maintained on the slits. The lowest temperature, that is,
to within a few tenths of a degree for periods of the cell body temperature, was used in all
1 hr or more. Temperatures of the cell were calculations as the temperature most representative
measured by two platinum—platinum-rhodium thermo of the vapor temperature. The effusing stream of
couples, one fixed in the cell body and the other salt vapors was roughly collimated in order,
fixed in one of the slits. These thermocouples primarily, to control the deposition of salt at
were calibrated with U. S. Bureau of Standards unwanted locations in the source. Salt deposits
thermocouples prior to use. The cell body and the upon insulators were an early source of electrical
cell slits were independently heated in the final leakage and are a troublesome factor in this type
equipment, and an attempt was made to keep the of experiment. Only those molecules which are
slits at a higher temperature than that of the cell, intercepted by the ionizing electron stream are

125
 ANP PROJECT PROGRESS REPORT

effective in forming ions, so narrow collimation of directly by any other experimental method. The
the molecular stream is not necessary. values of the absolute partial pressure of the
The effusion cell was charged with reagent-grade species are not obtainable by this method because
potassium chloride which had been fused in the ionization efficiencies for the production of
platinum. Clear crystals of salt were chosen the ionized progeny are not, in general, known.
from the melt and loosely packed into the effusion If a salt species, KCI, forms a dimer, (KCI)2,
cell. After the cell was assembled and the slits the following equilibrium obtains:
were adjusted to 2.5 mils with a feeler gage, it
was mounted in the mass spectrometer and checked (2) (KCI) 2 ^ - ^ 2KCI
for electrical leakage and thermocouple response.
The heaters were then adjusted to give the cell a The equilibrium constant K is given by
temperature of 350 to 400°C, and the outer envelope
of the mass spectrometer was raised to a temper P\
ature of 250°C. This bakeout-degassing operation (3) K =.

lasted at least 16 hr and, in some cases, for a

number of days, until the hydrocarbon background where p} is the partial pressure of the monomer
was reduced to a negligible level at the region of and p2 is that of the dimer. The standard free
the lower mass position for potassium chloride energy, Ap°, is given by
(74).
The data were taken by raising the temperature (4) AF° = -RT In K = AE° + RT - T AS0 .
of the cell in increments of 20 to 30 deg and
making measurements until constant temperatures The quantity AS°, which can be estimated by the
and mass intensities were observed. This time use of statistical mechanics, increases with an
interval varied from more than 1 V hr at the lower increase of mass.15 The energy, AE°, can be
temperatures to less than 15 min at the higher estimated for alkali halides by considering the
temperatures. An attempt was made to take data ions as charged, polarizable spheres. This quantity
from both an ascending and a descending sequence can also be obtained from the experimental data
of temperatures; however, the salt capacity of the by the relation
effusion cell was such that it showed signs of (5) AE° = 2AE, _ AE2 ,
exhaustion before any complete total sequence
could be measured.
where AE, and AE2 can be calculated from the
If the pressure of the salt vapor is low enough measurements by using Eq. 1. These energies are
so that the mean free path of the molecules is one of the two sets of factors that control the
greater than the slit width of the effusion cell, association, and, at present, they can be measured
the number of molecules of a given kind escaping independently only by the method described here.
from the slit should be a function only of the
temperature. Also if the area of the slit is small
Several measurements were made on KCI prior
compared with the inner surface area of the cell,
to the use of heated slits, and none of the experi
ments were completely successful because of the
the pressure of salt vapor in the cell should
approach the equilibrium pressure at a given
formation of deposits to a greater or lesser degree
at the exit slit. The deposits had the effect of
temperature. In this experiment the ratio of cell
surface area to slit area was 800. The measured
reducing the intensity of the beam in an arbitrary
ion current, or mass spectrometer peak height, of
and variable manner. A plot of the log of the
peak heights of the observed KCI+ and K Cl +
any of the ionized progeny «/ of the species i as a
species vs 1/7, as obtained in one of these
function of the temperature of the salt in the
effusion area is given by measurements, is shown in- Fig. 2.3.3. The slopes
of these lines, calculated by using Eq. 1, yield
d In w •' AE, the energy of vaporization of the monomer (AE ),
(1) and the dimer (AE2), and, by using Eq. 5, the heat
d(\/T)

where Ei is the, energy of vaporization of the I5r

species i, a quantity which cannot be obtained R. Fowler and E. Guggenheim, Statistical Thermo
dynamics, chap III, University Press, Cambridge, 1939.

126
 PERIOD ENDING DECEMBER 31. 7957

UNCLASSIFIED
ORNL-LR-DWG 27056
was chosen for this study and has been described
in previous reports in this series. The data ob
tained to date include fairly reliable measurements
of the surface tension of the NaF-ZrF4 mixture
X (53-47 mole %) at 600 and 700°C and preliminary
data at 800°C. A summary of the data obtained by
—\^-
-K,C the sessile-drop technique is presented in Table
2.3.5 and compared with data obtained by Miles19
by the maximum-bubble-pressure technique. A
KCI+
Table 2.3.5. Surface Tension of NaF-Z«F.
(53-47 Mole %) at Various Temperatures

Temperature
Surface Tens on (dynes/cm)
(°C) Data of Miles Autnor's Data

600 127* 127 + 6

700 120 119 ± 4
1.100 1.140 1.180 1.220 1.260 800 138 + 7**
If XlO"3 (K')

*Average of three measurements corrected for more

Fig. 2.3.3. Relative Intensities of Monomer and Dimer
recent values of the density.
Species of KCI as a Function of Temperature. Data
obtained from a mass spectrometric study of an effusing **The large loss of weight of the sessile drop during
beam of KCI.
the course of the measurement renders this value highly
inaccurate for NaF-ZrF4 (53-47 mole %). It would apply
of association (AE°) can be obtained. The values perhaps to a melt containing about 30 mole % ZrF ..
4
found, in kcal/mole, are:
preliminary value of 180 dynes/cm has been ob
AE, = 42.5 , tained for the surface tension of pure molten UF4
AE2 = 49.8 , at 1050°C. This work was initiated in an effort to
determine the effect of small amounts of UO.
AE° = 35.2 . on the wetting properties of UF4. Data obtained
by Cohen and Jones20 for NaF-ZrF4-UF4 (50-46-4
These values differ from the values of Beusman3 mole %) with a tensiometer are presented for
and those of Miller and Kusch. As mentioned comparison in Table 2.3.6. These tables summarize
earlier, this discrepancy may be due to the problem all the surface tension data on fuel mixtures and
of salt deposition at the effusion slit. Such their components that are available thus far.
deposits have not been reported elsewhere, although
they may be the source of the anomalous behavior
observed in similar experiments on oxides.16 The
M. G. Inghram and W. A. Chupka, Chemical Bonds
modified equipment, shown in Fig. 2.3.2, has not by Thermodynamic Methods, paper No. 40 presented at
as yet produced satisfactory data. the 4th Annual Meeting of the American Society for
Testing Materials, Committee E-14on Mass Spectrometry,
Cincinnati, May 28, 1956.
SURFACE TENSIONS OF MOLTEN SALTS
W. K. Stromquist and R. M. Boarts, Effect of Wetting
S. Longer on Heat Transfer Characteristics of Liquid Metals.
Third Quarterly Report, ORO-60 (Jan. 31, 1952).
The surface tension of a confined fluid has been
18J. W. Taylor, /. Nuclear Energy 2, 15 (1955).
shown to be important, in certain cases, in heat
transfer17 and in corrosion processes.18 A study 19F. W. Miles, ANP Quar. Prog. Rep. Sept. 10, 1956.
ORNL-2157, p 118.
of the surface tension of molten fluoride mixtures 20
S. I. Cohen and T. N. Jones, ANP Quar. Prog. Rep.
was therefore initiated. The sessile-drop technique June 10, 1953, ORNL-1556, p 88.

127
ANP PROJECT PROGRESS REPORT

Table 2.3.6* Data Obtained by Cohen and Jones on Much of the knowledge of fused fluoride salts
the Surface Tension of NaF-ZrF4-UF4 (50-46-4 Mole %) gathered in the past four years is still of a
at Various Temperatures qualitative nature. During the course of these
experiments, for example, it has become apparent
Temperature Surfa :e Tension that further information on the wetting behavior of
(°C) (dy nes/cm) fluoride salts is needed. The sessile-drop experi
ments on metal supporting surfaces and the
530 157 maximum-bubble-pressure experiments have shown
that truly clean hydrogen-fired surfaces are com
630 132
pletely wetted by molten NaF-ZrF4 but that, if
730 115 thin oxide films are allowed to form on the metal
surface after hydrogen-firing, the films will prevent
In connection with these experiments, it should wetting until they are dissolved by the molten
be noted that surface tension data are almost salt. These observations are difficult to reconcile
valueless without a knowledge of the density with the lack of evidence of wetting in fuel
of the liquid. Miles19 developed a method by purification vessels, where it has been consistently
which both the density and the surface tension can observed that the purification pots drain cleanly
be obtained from the measurements involved in and that no residual film of salt (other than
the maximum-bubble-pressure technique. While splattered droplets) is left on the walls. Visual
theoretically the density of a sessile drop can be observation studies21 have tended to confirm this
obtained by weighing the drop and calculating its behavior, since the salt meniscus appears to be
volume from its surface profile, in practice the definitely convex, especially in melts containing
method is subject to extremely large errors. BeF2.
Preliminary pycnometric density data on molten
UF4 and dilatometric measurements of solid UF4
have yielded the interesting result that UF4 melts
21
with negligible expansion, if any. R. J. Sheil, private communication to S. Longer.

128
 PERIOD ENDING DECEMBER 31, 1957

2.4. PRODUCTION OF PURIFIED MIXTURES

G. J. Nessle J. P. Blakely

PREPARATION OF FLUORIDE COMPOUNDS containing alloys have always been within this
B. J. Sturm L. G. Overholser
solid-solution area.2 The compound Rb3CrF,
was identified as the corrosion product that results
Compounds not available commercially have from the attack of a chromium-containing alloy by
been synthesized for use in ANP chemistry an RbF-LiF composition.3
studies. Of particular interest were fluorides of
the metals being considered as container materials Other Double Fluorides
for various fuels and the products of their reactions
with constituents of proposed fuels. Many of
A new series of compounds of the type M"ZrF,
these compounds had not been reported in the
were formed by fusing ZrF4 with fluorides
of divalent transition metals. The compounds
chemical literature. Corrosion products found in
fuels have been identified by comparison of their
CrZrF^, MnZrF6, FeZrFfi, CoZrFfi, NiZrF6, and
x-ray diffraction patterns and optical crystal-
ZnZrF^ have been prepared.4 The compounds
lographic properties with those of synthesized CuZrFfi, CdZrFfi, and PbZrF6 probably exist,
but the products of fusions corresponding to these
compounds. Fluorides of the fission products
compositions have not yet been adequately studied.
have been prepared for use in the study of their
The compound NiZrF. was identified as the
effects on the fuel mixtures of interest.
corrosion product that results from the action on
Alkali-Metal Fluochromates a nickel container by the vapors resulting from the
fusion of ammonium bifluoride with a fluozirconate
A series of bright green chromium fluoride
composition.
complexes of the type M'Cr'"F. were prepared
Fusions of alkali-metal fluorides with transition-
in which M stands for an alkali metal or ammonium
metal fluorides give a series of compounds.
ion.1 The purpose of preparing the alkali metal
Except for lithium fluoride, which does not combine,
fluochromates was to obtain their crystallographic
the alakli metal fluorides form 1:1 compounds with
properties for comparison with those of unidentified
manganese, cobalt, and nickel difluorides.5 This
phases.
is probably also true with chromous fluoride, but
Potassium hexafluochromate(lll), K.CrF., was
the results are confused by the contamination of
prepared by slowly heating to 900°C a mixture of
the preparations by chromic-containing phases as
3 moles of potassium bifluoride with one of
a result of disproportionation of some of the
hydrated chromic fluoride. Substitution of sodium
chromous fluoride. The rubidium and cesium
bifluoride for the potassium bifluoride yielded
compounds were unknown before this work was
Na3CrF6. Heating hydrated chromic fluoride with begun, and very little information was available on
an excess of ammonium bifluoride to 250°C formed
the physical properties of the rest of these
(NH4)3CrFi. The compounds Li3CrF6, Rb3CrF6, compounds.
and Cs3CrF. were then obtained by reacting the
alkali metal fluoride with either (NH4) CrF, or a Simple Inorganic Fluorides
mixture of ammonium bifluoride ana hydrated
Many different anhydrous fluorides were prepared.
chromic fluoride.
Moisture was removed from lithium, sodium, and
Fusion of LkCrF, with Na3CrF, gave the com potassium fluorides by heating in vacuum to
pound Li2NaCrF.. The ternary system KjCrF,-
Li3CrF.-Na3CrF, has a solid-solid-solution
area bounded by the compounds KjNaCrF,, B. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. Sept. 10, 1952. ORNL-1375, p 122.
K Na,(CrF.)2, and KjLiCrF,. The compositions
of corrosion products that result from the action of 3B. J. Sturm and E. E. Ketchen, ANP Quar. Prog.
Rep. Sept. 10. 1954, ORNL-1771, p 67.
NaF-KF-LiF-UF4 or NaF-KF-LiF on chromium- 4B. J. Sturm, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, P91.
1B. J. Sturm and L. G. Overholser, ANP Quar. Prog. B. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. Dec. 10, 1952, ORNL-1439, p 115. Rep. June 10, 1953, ORNL-1556, p 48.

129
ANP PROJECT PROGRESS REPORT

around 500°C. Strontium, cadmium, cerium, Chromous Fluoride

praseodymium, neodymium, and samarium fluorides High-purity chromous fluoride was needed for
were precipitated by the addition of hydrofluoric studying the removal of chromium from container
acid to aqueous solutions of their salts.6-8 Oven materials by fluoride fuels. Reaction of powdered
drying at around 200°C yielded essentially chromium metal with hydrogen fluoride gas at
anhydrous material. about 800°C was found to be a poor method of
Yttrium fluoride, precipitated from aqueous preparing chromous fluoride, because a coating
solution, was treated with anhydrous hydrogen of the fluoride formed on the metal and prevented
fluoride at an elevated temperature to ensure further reaction with the hydrogen fluoride.
freedom from moisture. The hydrated ferric, Grinding of the mixture exposed new metal surface
nickelous, cuprous, and zinc fluorides were to react with hydrogen fluoride, but three successive
dehydrated by treating them in a similar manner.7 treatments failed to react all the metal. The
Twenty hours of hydrofluorination followed by 4 hr product was also contaminated with chromic
of flushing with helium reduced the boron content fluoride that resulted from the reaction
of cupric fluoride to below 2 ppb, the spectroscopic
limit of detection. The anhydrous fluorides
prepared by treatment with hydrogen fluoride gas 2CrF2 + 2HF- •^2CrF3 + H2
were usually flushed with purified helium gas to
remove adsorbed hydrogen fluoride. The hydrated
Hydrofluorination of chromous chloride at about
aluminum, chromic, and nickelous fluorides were
700°C yielded chromous fluoride that was con
reacted with ammonium fluoride or bifluoride at
taminated by chromic fluoride, because of oxidation
around 200°C to form the ammonium fluocomplex,
by hydrogen fluoride, and traces of unreacted
which was decomposed at a higher temperature to
chloride.12 The chromic fluoride could be reduced
form the anhydrous fluoride.6
by hydrogen.
Hydrofluorination of anhydrous ferric chloride at
around 300°C and chromic chloride at around 600°C A few kilograms of chromous fluoride were
converted the chlorides to fluorides.4*9 The prepared by hydrogen reduction of chromic fluoride.
hydrated nickelous, ferrous, and cobaltous chlorides The chromic fluoride was usually made by thermally
were slowly heated to around 500°C during the decomposing ammonium fluochromate, which was
passage of hydrogen fluoride to convert them made by fusing ammonium bifluoride with hydrated
to the anhydrous fluorides.7'10 Silver carbonate chromic fluoride. Usually the preparation of
was similarly converted into the fluoride at around chromic fluoride and its reduction by hydrogen were
150°C, and ferrous oxalate was converted at combined into a single step in which ammonium
around 600°C.11 The tetrafluorides of thorium and fluochromate was placed in a nickel reaction
uranium were freed of oxide and moisture by vessel and the temperature was slowly raised to
hydrofluorination at around 700°C.9 Vanadium 700°C while a flow of hydrogen was maintained.
trifluoride was prepared by condensing hydrogen Chromous fluoride prepared in this manner was
fluoride onto vanadium trioxide powder, heating to usually contaminated with some opaque material.
remove water and excess hydrogen fluoride, and This was especially true of the product near the
rehydrofluorinating at 200°C.10 walls of the nickel reaction vessel. The chromous
fluoride was further purified by fusing at 1100°C,
recrystallizing, and manually separating the clear
E. E. Ketchen and 3. J. Sturm, ANP Quar. Prog.
Rep. June 10, 1955. ORNL-1896, p 76. crystals from the opaque residue.
7B. J. Sturm, ANP Quar. Prog. Rep. Dec. 10, 1953, At first it was assumed that the opaque material
ORNL-1649, p 62.
resulted from reduction of some of the chromous
83. J. Sturm, ANP Quar. Prog. Rep. March 10, 1956, fluoride to the free metal. Free metal could not be
ORNL-2061, p 100.
9B. J. Sturm and E. E. Ketchen, ANP Quar. Prog. Rep. found by x-ray diffraction analysis, but gamma
March 10, 1955, ORNL-1864, p 66. chromium nitride was identified in the opaque
,0B. J. Sturm and L. G. Overholser, ANP Quar. Prog. material. The following reactions are proposed
Rep. June 30. 1957. ORNL-2340, p 164.
12
nL. G. Overholser and B. J. Sturm, ANP Quar. Prog. B. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. Sept. 10. 1953, ORNL-1609, p 70. Rep. Dec. 31, 1956, ORNL-2221, p 153.

130
 PERIOD ENDING DECEMBER 31, 7957

to explain the formation of chromium nitride: at elevated temperatures. Commercial NaF con
tains some sulfur, mostly as sulfate, and small
(NH4)3CrF6- 3NH4F + CrF3 amounts of water.

2NH4F- NH4HF2 + NH„

The fuel solvent, NaF-ZrF4, and the fuel concen
trate, 2NaF-UF4, were prepared separately for the
CrF3 + NH3- ->CrN + 3HF ARE, as required by the startup procedure. For
testing of components and for heat transfer and
The formation of chromium nitride by the reaction metallurgical research, however, large quantities of
of ammonia with chromic chloride was reported by the fuel mixture NaF-ZrF4-UF4 were prepared
Schwarzkopf and Kieffer.13 It is reasonable to in various compositions. Such bulk production was
expect chromic fluoride and ammonia to react in carried out on a 250-lb-batch scale, but con
the same way. siderable production of special compositions in
If chromic fluoride was not completely reduced small batches on a 5-, 10-, and 50-lb experimental
to chromous fluoride, a phase was formed which scale was required for many specific test programs.
had a composition intermediate between the compo In general the fluoride mixtures were purified by
sitions of chromic and chromous fluoride. The
the process described below, with only minor
exact composition of this phase has not been modifications for the small batches when specific
established, but its optical properties and x-ray requirements had to be met for specialized
diffraction pattern have been determined. experiments.

SUMMARY OF EXPERIENCE IN THE

Purification of Fuel Solvent
PREPARATION OF FUEL FOR
THE AIRCRAFT REACTOR EXPERIMENT The equimolar NaF-ZrF4 mixture required for
the ARE was purified in 250-lb batches. The
C. R. Croft J. E. Eorgan reaction vessel, in which the chemical processing
F. A. Doss J. Truitt
was accomplished, and the receiver vessel, into
The nuclear poisons, such as boron, cadmium, which the purified mixture was ultimately trans
and certain of the rare earth elements, are not ferred, were 13-in.-dia, vertical cylindrical con
common contaminants of enriched uranium, sodium, tainers, 48 and 24 in. long, respectively, that were
and zirconium fluorides. Hafnium, which invariably fabricated of A-nickel by the Heliarc welding of
contaminates zirconium and its compounds, can /g-in. plate. The top of the reactor vessel was
be reduced in concentration to acceptable limits pierced by a charging port consisting of a 4-in.-dia
(50 to 100 parts of Hf per million parts of Zr) on a pipe which was capped well above the heated zone
commercial scale. Since the initial concentrations by a Teflon-gasketed flange. The tops of both the
of these poisons were low and the parasitic capture receiver and the reaction vessel were pierced by
of neutrons in the core structure would be con short risers which terminated in Swagelok fittings
siderable, purification processes were designed through which gas lines, thermowells, etc., of
to minimize corrosion and to ensure the absence of /£-in. nickel tubing could be introduced. A transfer
slightly soluble oxides and oxyfluorides rather line of /£-in. nickel tubing terminated near the
than to improve neutron economy. bottom of the reactor vessel and near the top of
The best-quality UF4 available commonly con the receiver; entry of this tube was effected through
tains small quantities of U02, UF5, and UOjFj. copper-gasketed flanges on Hn.-dia tubes which
The commercially available ZrF4, even when pierced the tops of both vessels. This transfer
prepared by hydrofluorination of ZrCL and further line contained a filter of Micro Metallic sintered
refined by sublimation at low pressure, contains nickel and a sampler which collected a specimen
up to 2% excess oxides and oxyfluorides, up to of the liquid during transfer. Through one of the
0.2% Cl~, and Fe+++ to the extent of a few risers in the receiver a 3^-in. nickel tube extended
hundred parts per million. Both these tetrafluorides to the receiver bottom; this tube, which was sealed
are readily hydrolyzed to oxides and oxyfluorides outside the vessel, ultimately served as the means
for transfer of the purified melt to the ARE. A
13
P. Schwarzkopf and R. Kieffer, Refractory Hard schematic diagram of the reaction vessel—receiver
Metals, Macmillan, New York, 1953. assembly is shown in Fig. 2.4.1.

131
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED The hydrogen and all subsequent reagent gases

ORNL-LR-DWG 20157
were fed at about 3 liters/min to the reaction
vessel through the receiver and the transfer line
TEFLON GASKETED FLANGE
and, accordingly, bubbled up through the molten
charge. The hydrogen was then replaced by
CHARGING PORT anhydrous HF, which served, during a 2- to 3-hr
period at 800°C, to volatilize H2S and HCI and to
convert oxides and oxyfluorides of uranium and
zirconium to tetrafluorides, but at the expense of
introduction of considerable NiF_ into the melt
through reaction of HF with the container. A final
3/8-in.-NICKEL 24- to 30-hr treatment at 800°C with H2 sufficed
TRANSFER LINE
to reduce the NiF2 and the contained FeF2 to
insoluble metals.
At the conclusion of the purification treatment,
a pressure of helium above the salt in the reactor
FILTER
vessel forced the melt through the transfer line,
filter, and sampler into the receiver. The metallic
iron and nickel were left in the reactor vessel or
SAMPLER on the sintered nickel filter.
The purified melt was permitted to freeze under
an atmosphere of helium in the receiver vessel,
and, by appropriate manipulation of connections
and valves, the cold receiver was detached and
removed to storage while the helium atmosphere
was maintained over the contents.

- REACTION VESSEL
Purification of Fuel Concentrate
RECEIVER VESSEL- The 2NaF-UF mixtures were prepared by a very
similar processing schedule in similar but smaller
equipment. In order to avoid the possibility of
Fig. 2.4.1. Apparatus for Purification of Fluoride criticality accidents, the fuel concentrate was
Mixtures. prepared by using reactors and receivers 4 in. in
This assembly was connected, through traps for diameter to limit batch sizes to about 30 lb.
collecting sublimed ZrF4, to a manifold through
which He, H2, HF, or vacuum could be supplied Storage and Ultimate Transfer
to either vessel. By a combination of large tube
The receiver vessels containing either fuel
furnaces, resistance heaters, and lagging, sections
solvent or fuel concentrate were connected to
of the apparatus could be brought, independently,
manifolds that supplied pure helium at slightly
to controlled temperatures in excess of 800°C.
more than atmospheric pressure and were stored
The raw materials in batches of the proper
in this manner at room temperature until used.
composition were blended and charged into the
reaction vessel. The material was melted and Transfer of the material for ultimate use was
heated to 700°C under an atmosphere of anhydrous accomplished by forcing the remelted charge under
HF to remove H20 with a minimum of hydrolysis. helium pressure through the transfer line built
The HF was then replaced with H2 for a period of into the receiver vessel. In this simple fashion,
1 hr, during which the temperature was raised to virtually quantitative transfer of the molten
800°C, to reduce U5+ and U6+ to U4+, sulfur material could be accomplished. Transfer of up
compounds to S—, and extraneous oxidants to 250 lb of solvent usually left less than 0.5 lb
(Fe+++, for example) to lower valence states. of solvent in the container.

132
 PERIOD ENDING DECEMBER 31. 7957

PILOT-SCALE PURIFICATION OPERATIONS PRODUCTION-SCALE OPERATIONS

C. R. Croft J. Truitt J. E. Eorgan

In the early stages of the ANP program, extensive The ARE and the large-scale component testing
small-scale experimentation was carried out in program, as stated above, created demands for
order to develop a suitable process and a flexible processed fluorides far beyond the capabilities
processing unit for the purification of fluoride of the small pilot-scale units, and therefore the
mixtures. Small-scale units of 5- and 10-lb capacity 250-lb production units were installed to process
were installed, and later, as demands grew, 50-lb fairly standard compositions in large quantities.
units were added. Finally, six small units and The first production operation of this facility was
three 50-lb units were in continuous use. Although the processing of the approximately 3300 lb of
bulk production was taken over by the 250-lb fuel solvent required for the ARE. Analyses of
capacity units, demands for small, special batches the 14 batches of fuel solvent processed for use in
continued to increase. During the active life of the ARE, as can be seen from the results given in
the ANP fluoride production program, 653 batches Table 2.4.1, showed the purity of the finished
totaling approximately 11,500 lb of various fluoride material to be very satisfactory. Since this was
compositions for use in phase equilibrium studies, the first operation of the production-scale units,
physical property measurements, small-scale the performance of the processing units during
corrosion tests, and other experiments were this period served as a basis for future production
processed in these pilot-scale units. schedules.

Table 2.4.1. Results of Analyses of ARE Fuel Solvent

Spec trogra phic

Chemical Analysis
A nalys
Batch Ma or Constituents Impurities (ppm)
(%) (ppm)
No.
Zr F Na Fe Cr Ni
HF B

1 43.3 45.0 9.76 30 <10 <10 240 0.2

2 43.9 45.2 11.0 45 <20 <20 55 0.2

3 43.9 45.5 10.6 60 <10 25 100 0.9

4 42.0 44.7 10.8 55 <20 45 80 0.9

5 44.1 44.9 11.0 90 <20 <20 60 0.6

6 43.6 44.7 10.5 35 <20 30 85 0.5

7 43.6 45.2 10.4 35 <20 30 70 0.5

8 43.4 44.6 10.4 35 <-20 <20 108 0.1

9 43.1 44.6 10.6 35 <20 <20 67 1.0

10 43.8 44-6 10-6 45 <20 35 60 0.3

11 43.7 45.1 10.5 40 <20 35 71 0.4

12 43.7 44.8 10.7 30 <20 <20 39 0.3

13 44.1 44.5 10.3 40 <20 30 42 0.6

14 43.9 44.7 10.3 <20 <20 35 48 0.6

133
ANP PROJECT PROGRESS REPORT

The efficiency, or production loss, of this During the production operation of this facility,
process and facility, with regard to raw material 322 batches totaling 80,000 lb of processed
usage vs quantity of final product, was very good. fluorides were produced. Since production usually
A material balance for the fuel-solvent production exceeded demand, the facility operated on an
operation is given below: average of eight months per year.

SERVICE OPERATIONS
Weight (lb)
F. A. Doss J. Truitt
Raw fluorides used (NaF and ZrF4) 3412
J. E. Eorgan N. V. Smith
Net fluorides produced 3348
In conjunction with the accelerated component
Net samples 5 testing program, and the resulting acceleration in
Net heels left in reactor vessels 56 fluoride production, a facility was established for
the efficient and accurate handling of the large
Total produced 3409
amounts of fluorides and liquid metals used. The
Net loss 3
250-lb batches could, in some cases, be used
directly, but ordinarily these batches had to be
reduced to quantities ranging from 1 to 50 lb to
On the basis of usable product, a production fit specific test programs. Complete analytical
efficiency of 98% was obtained. records and material balances were kept to avoid
The production facility was normally run on a confusion in the interpretation of component test
five-day-week, 24-hr-shift basis during the early results. Responsibility was taken for the loading,
stages of the component testing program. When sampling, and unloading of charge materials from
demands from off-area consumers (such as Pratt & component test equipment, including the high-
Whitney Aircraft) increased, the operation was temperature critical assemblies. During the course
placed on a continuous seven-day-week schedule. of the program, approximately 70,000 lb of
Maximum production on the seven-day-week processed fluorides and 35,000 lb of liquid metals
schedule was reached at 1000 lb per week. were dispensed. In addition, some 200 storage
Production on a five-day-week schedule averaged cans used in dispensing processed fluorides to
about 600 lb per week. local and off-area consumers were maintained.

134
 PERIOD ENDING DECEMBER 31, 7957

2.5. ANALYTICAL CHEMISTRY

J. C. White

DETECTION OF TRACES OF NaK IN AIR Although flame photometric measurements may

be extended to ultimate sensitivities in the desired
A. S. Meyer, Jr. J. P. Young
concentration range, such measurements might
A study of sensitive methods for the detection not be reliable because of the problem of main
of traces of alkali metals in air was carried out taining a constant flame source by remote control.
in order to provide a monitor for the exhaust air Experience with a Beckman flame photometer
from the ART NaK-to-air radiators. The sensitivity during operation of the ARE indicated that this
requirement for a suitable detector was established problem imposes serious limitations to the
on the basis of a tentative limit of a NaK leak application of this instrument. Additional compli
which would correspond to the introduction of a cations can be predicted in the application of this
total volume of 6 cm3 of NaK per hour into the technique to the proposed detector in that it would
exhaust air stream. On the basis of the design be necessary to maintain two flames at identical
flow rate of air in the radiator ducts, the instrument temperatures and to meter samples of the inlet
was required to respond to a concentration of NaK and exhaust gases to these flames at constant
in air of 10 ppb by weight. Since it was also rates.
assumed that the inlet air could be contaminated The detection of alkali metals by means of a
with alkali metal compounds in concentrations surface-ionization detector is based on the
comparable with those introduced by the radiator measurement of the current that is carried by
leaks, it was considered necessary that the positive ions which are formed when atoms of the
detector be designed to monitor both the inlet alkali metals are brought into contact with a
and exhaust streams and to respond to an increment heated filament. This detector has been applied to
of 10 ppb of NaK across the radiators. Additional the detection of sodium and NaK in helium.1
specifications for a suitable detector included a Concentrations of 2 ppm of sodium and 9 ppm of
rapid response time and a high degree of reliability. NaK have been detected in this manner. It was
Methods of detection which were evaluated during suggested2 that the surface-ionization detector
the feasibility studies included chemical analyses, might be used for air samples if an iridium filament
photometric smoke detection, emission spectro- were used. Under these conditions, detection
graphic analysis, flame photometric analysis, limits of 1 to 10 ppm of NaK could be achieved.
surface-ionization detection, and resonance- The calibration curves reported for the measurement
radiation measurement. In addition, some con of alkali metals in helium (current vs saturation
sideration was given to a technique in which a temperature) are linear over only a very limited
material that could be readily detected by existing range of concentration, however, and this method
conventional methods of analysis would be added would therefore not be readily applicable to the
to the NaK. Radiometric techniques were not proposed differential measurement of concentration.
considered. This method was given some further consideration
Existing chemical methods, emission spectro- since it was thought that, because of the small
graphic techniques, and smoke detection do not size of the sensing elements, numerous elements
offer sufficient sensitivity for the detection of could be placed in strategic locations to detect
NaK at the 10-ppb concentration level. Although the larger concentrations of alkali metals which
mass spectrographic techniques offer adequate would be present in the immediate vicinity of the
sensitivity, they do not appear to be applicable leak. The detection requires the measurement of
to the proposed system because of the problems relatively small direct currents, however, and the
concerned in introducing representative, continuous instrumentation for such a multiple installation
samples of the alkali metal oxide suspensions of detectors might become prohibitive.
through capillary leaks into the ionization chambers
of the spectrometer. For the proposed differential 'H. S. McKown, A Detector for Sodium Vapor - A
measurement it would be necessary to maintain Surface Ionization Sodium Detector, K-996 (Feb. 16,
capillary leaks of equal rates for both inlet and 1953).

exhaust air.
J. R. Sites, private communication to A. S. Meyer, Jr.

135
ANP PROJECT PROGRESS REPORT

The proposed resonance-absorption method is which included the substitution of a Polaroid

based on the measurement of the attenuation of a chopper and refinements of the 90-cycle tuned
beam of sodium D light (589 mj*)on passage through amplifier, the limit of detection was lowered to
sodium metal vapor. This method offers extreme 60ppb.3 Additional modifications of the instrument
sensitivity for the detection of sodium. Since the were made to eliminate the continuous radiation of
absorption coefficient for sodium vapor at the the heated cell walls from the photomultiplier
center of the resonance line is approximately tube and to provide means of exact alignment of
10""12 N, where N is the density of sodium in the optical system. Details of these modifications
atoms per square centimeter, it was calculated were given previously.5 Experimental work on the
that a concentration of 0.01 ppm of atomic sodium project was discontinued before the improved
in air would be sufficient to reduce by 50% the instrument could be tested.
intensity of a beam of sodium radiation that had It is believed that the studies which have been
traversed a 50-cm optical path. Inasmuch as carried out have demonstrated the applicability of
compounds of sodium such as Na20 do not absorb this method to the detection of the radiator leaks.
the resonance radiation, it is necessary to main The test model responded almost instantaneously
tain the sample at a temperature sufficient to to small increments in the concentration of the
cause thermal dissociation of Na_0 to atomic sodium in the air samples, and with refinements
sodium in order to obtain a measurable absorption in the optical and electronic components its
of the resonance line. sensitivity could readily be increased to provide
On the basis of thermodynamic constants it was detection at the required concentration level. The
calculated that a concentration of Na20 of 10 ppb double-beam system is known to offer high stability
could be quantitatively dissociated at temperatures and reliability. With minor modifications, the
as low as 800°C. Dissociation of comparable instrument could be adapted to the detection of
concentrations of K20, which can also be detected other alkali metals in air.
by the resonance-absorption method, is complete at On the basis of these experiments, design
approximately 600°C.3 criteria for a working model of the instrument were
A test model of an instrument designed to com prepared.3 Improvements to the test model would
pare the absorption of samples of the inlet and include: a rigidly mounted optical system to
exhaust air from the radiators was constructed. provide a multiple pass of the light beam through
Detailed descriptions, including a photograph of the sample; an electrodeless source lamp excited
the instrument, were given previously.4 In this by radio frequencies to provide extended lamp life
instrument a beam of sodium D light is divided and a sharper line source; multilayer interference
into two pulsed beams of equal intensity by a filters to isolate the resonance line more ef
90-cycle mechanical chopper. After the beams fectively; and a phase-sensitive amplifier to
are passed through heated absorption cells which discriminate completely against frequencies of
contain samples of the inlet and exhaust air, the other than 90 cycles.
transmitted beams are recombined at a photo- Although the feasibility of the proposed detector
multiplier tube. A 90-cycle a-c signal, which is appears to have been demonstrated, much additional
proportional to the difference in absorption between experimental work would be required to establish
the inlet and exhaust air, is generated by the sampling techniques. Questions to be resolved
photomultiplier. include the selection of the best material for the
This instrument was tested by passing, through construction of sampling ports and transfer lines,
one of the tubes, samples of sodium in air which the determination of the optimum temperature and
were prepared by mixing with the air stream helium flow rates to minimize deposition of the alkali
that had been saturated with sodium vapor. A metals on the walls of the tubing, and the required
detection limit of approximately 100 ppb was recovery time of the assembly after exposure to
noted. With minor modifications of the instrument, the large concentrations of alkali metals which
would be introduced by a major leak. A com
parison of the transfer characteristics of K 0 and
A. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog.
Rep. June 30. 1957. ORNL-2340, p 166.
A. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog. A. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog.
Rep. March 31. 1957. ORNL-2274, p 137. Rep. Sepu 30. 1957, ORNL-2387, p 148.

136
 PERIOD ENDING DECEMBER 31, 7957

Na 0 should be undertaken in order to provide DETERMINATION OF OXYGEN IN

criteria for a choice of the element which could be FLUORIDE SALTS
monitored most efficiently. A. S. Meyer, Jr. G. Goldberg
A major test facility would be required in order The gas-measuring system of the apparatus8
to provide a means of simulating a NaK leak in an for the determination of oxygen in fluoride salts
ART radiator. This facility could be used to with KBrF4 was calibrated by measuring the
establish criteria for the selection of sampling displacement of the manometer levels produced by
points which would assure complete mixing of the the addition of measured volumes of oxygen.
air for representative sampling and to determine Discrete additions of oxygen were made by means
the reliability of the instrument in the detection of of the apparatus shown in Fig. 2.5.1. By ap
small increments in the concentration of the propriate rotation of the three-way stopcock (S )
alkali metal in the presence of large background the volume of oxygen in the sealed arm of the
concentrations. stopcock (0.281 cm3) is connected with the
Toepler pump. The oxygen is transferred to the
Although the tests could be carried out with less measuring system by means of the Toepler pump
extensive equipment if the alkali metal could be in a manner analogous to its transfer from the
introduced to the air stream at an infinitesimal reaction system when determinations are performed.
rate as a dilute solution, preliminary experiments6 The coefficient of variation of the calibration
have indicated that the physical properties of the measurements was less than 2%.
samples vary widely with the method of prepa
ration. Accordingly, to carry out definitive tests The accuracy of the calibration was confirmed by
of sampling procedures, the direct addition of NaK determining the oxygen in 10-mg samples of Y„03.
to the heated air stream would be essential. Tests Comparable precision was obtained in these
carried out with water have indicated that for measurements. Typical results of replicate
minimal rates of injection, as a jet, high pressures determinations by this procedure are given in
and minute orifices are required. In order to inject Table 2.5.1.
NaK at a rate consistent with a test facility of
reasonable proportions, pressures in excess of The range of the method is from a minimum of
1000 psi and orifices 0.0001 to 0.0003 in. in 5 jig of oxygen, as measured with the oil manometer,
diameter would be required. Orifices of this to a maximum of 100 mg, as measured by the
diameter can be drilled by anodic dissolution with mercury manometer. For quantities of oxygen in
a virtual electrode.7 Orifices less than 0.0005 in.
excess of 100/zg, the precision of the method would
in diameter have been fabricated in nickel by this
appear to be less than 5%.
technique. Zirconium orifices would be more
satisfactory, but electrolytic drilling of this
material has not been successful. Q

A. S. Meyer, Jr., and G. Goldberg, ANP Quar. Prog.

Rep. Sept. 30, 1957, ORNL-2387, p 150.
These sampling difficulties could be largely
resolved by adding to the NaK a volatile material
Table 2.5.1. Determination of Oxygen in Samples
which could be detected in small concentrations.
of Y203 and YF3
Mercury has been suggested as an ideal additive,
since it could be readily transferred and can be
detected by resonance-absorption methods at Oxygen (mg)
Sample Sc imple Weight
operating temperatures. The addition of mercury Composition (g) Theoretical Determined
would be contingent on the evaluation of its effect
on the compatibility of NaK with structural ma 0.0103 2.19 2.16
terials and its effect on nuclear properties. Y2°3
0.0102 2.17 2.18

6A. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog. YF3 1.00 0.58
Rep. Dec. 31. 1956. ORNL-2221, p 164.
1.00 0.59
7A. Uhlir, Jr., Rev. Sci. Instr. 26, 965 (1955).

137
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG27057

VACUUM GAGE

Jl

VACUUM ? 0 HELIUM

MEASURING
MANOMETER
TOEPLER PUMP
OXYGEN-METERING
S= HIGH-VACUUM STOP COCKS MANOMETER

Fig. 2.5.1. Constant-Volume Manometer Calibrating Apparatus.

138
 Part 3

METALLURGY

W. D. Manly
 3.1. NICKEL-MOLYBDENUM ALLOY DEVELOPMENT STUDIES

MATERIAL DEVELOPMENT and types, it has been determined that a particular

H. Inouye T. K. Roche property of interest can be improved, but usually
J. H. Coobs
at the expense of some other property. Inasmuch
as the property variations are derived largely from
The developmental work on nickel-molybdenum- additions of third, fourth, or fifth alloying elements,
base alloys for use as reactor structural materials the principal properties of importance such as
in contact with the fused salt fuel NaF-KF-LiF-UF4 strength and corrosion resistance can be balanced
(11.2-41-45.3-2.5 mole %, fuel 107) at normal to give an alloy composition which is best suited
operating temperatures of 1650°F and above has for a particular application. As the severity of
been terminated. Much of the technology and ex the service conditions is increased (for example,
perience gained in this investigation will be of higher temperatures and longer operating times), the
value to the Molten Salt Reactor program. freedom of modification becomes narrower, and
Status of Development thus many of the alloys are eliminated for various
reasons. In short, for every temperature or appli
Based on the requirements for an ideal container cation, there appears to be an optimum alloy
material for fuel 107, the experimental nickel- composition.
molybdenum-base alloys have shown exceptional
promise. The various alloys which have been For 1000 hr of operation, the alloys, as presently
studied fall into three general categories: solution- developed, appear to permit peak temperatures of
hardening, precipitation-hardening, and dispersion- 1650 to 1700°F at the metal-salt interface. This
hardening alloys. The solution-hardening alloys limitation has been tentatively set by the rapid
are represented by alloys INOR-l and -8; the pre increase of the corrosion tendency of the alloys
cipitation-hardening alloys by IN0R-3, -4, -6, and containing chromium and the deterioration of
-7; and the dispersion-hardening alloys by INOR-5 strength above these temperatures.
and -9. Preliminary data on the properties of the The alloys that have been studied have been
INOR alloys 1 through 8 were presented previously, evaluated in terms of fabricability, mechanical
along with the composition ranges of each alloy properties, corrosion resistance, weldability, and
studied.1 The data indicate that, of the three aging characteristics. The results of these in
categories of alloys, the solid solution-hardening vestigations are summarized below.
alloys most nearly meet the requirements. On
this basis, the INOR-8 alloy was selected as Fabricability. - Alloys which are solid solutions
having the best combination of properties, and at the forging temperatures should present no un
much of the effort has been directed toward estab usual fabrication problems. Forging difficulties
lishing the final composition and producing com have been encountered with large heats of the
mercial quantities of the alloy. It is felt, however, INOR alloys, as reported previously, ' but such
that the full potential of alloys of the nickel- difficulties have been associated with poor melting
molybdenum system definitely has not yet been and forging techniques. It is anticipated that
developed. For the immediate future, these alloys further fabrication problems will be those normally
appear to be the only structural materials which associated with developing reduction schedules for
show the possibility of bridging the gap between new materials.
Inconel and the refractory metals niobium and A serious limitation in the production of tubing
molybdenum. has been the tendency toward longitudinal splitting
The most attractive feature of the nickel-base of alloys that contain over 0.10% carbon. This
alloys containing 15 to 20% molybdenum as the tendency has not been noted during hot-forging
principal alloying element is their versatility. operations or during cold rolling of plate and sheet
From the studies of numerous alloy combinations products.

'h. Inouye and T. K. Roche, ANP Quar. Prog. Rep. H. Inouye, T. K. Roche, and J. H. Coobs, ANP Quar.
March 31, 1957, ORNL-2274, p 177. Prog. Rep. June 30, 1957, ORNL-2340, p 175.

141
ANP PROJECT PROGRESS REPORT

Mechanical Properties. — The mechanical prop determined to be 6%. Alloys containing less than
erties of the nickel-molybdenum alloys are de 6% chromium show about 4 mils of oxidation in
scribed in detail in a subsequent section of this 170 hr, while those containing more than 6% show
chapter. In general, the available data show that attack of less than 1 mil.
an increase of up to 200°F in operating temperature The extent of oxidation in 1000 hr at temperatures
should be obtained by substituting INOR-8 for as high as 1750°F has been tolerable for all
Inconel. Thus, operation at temperatures between compositions studied. This conclusion is based
1650 and 1700°F with fuel 107 appears to be upon the fact that approximately 75 loops have
possible. been operated without special coatings or pro
A correlation between microstructure and creep- tection. The probable maximum service temper
rupture characteristics showed that the micro- atures in fuel 107, in sodium, and in air for alloys
structures with grain-boundary networks of fine of current interest are given in Table 3.1.1.
carbides had the best strength properties. Carbide Weldability. - The status of the development of
precipitation studies showed that the desired techniques for welding these alloys has been re
microstructure could be obtained by using high ported in detail in reports in this series. It is
annealing temperatures (2300°F). Improvements felt that the weldability will not be a serious
in the time to rupture and in ductility were made by problem, since approximately 40 alloys of com
minor additions of malleableizing elements and positions representing various modifications of the
vacuum remelting of air melts. nickel-molybdenum base material have been welded
Corrosion Resistance. - Within certain limi satisfactorily. Such welds have been in service
tations, the corrosion-resistance characteristics of for periods of up to 1000 hr at temperatures as high
the nickel-molybdenum alloys approach the optimum as 1850°F without evidence of deterioration. The
for service in fused-salt, sodium, and air environ incidence of poor weldability and cracking ten
ments. The corrosion behavior of these alloys in dencies under restraint in various heats of the
fused salts and sodium is described in detail in a alloys has shown good correlation with evidence
subsequent section of this chapter. that those heats were not properly melted. The
The oxidation rates of the nickel-molybdenum results of high-temperature ductility tests have
alloys have been found to vary from excessive to been reported.
slight, depending upon the composition. The
element having the largest effect is chromium, and E. F. Nippes, W. F. Savage, and R. P. Meister,
Hot-Ductility of Nickel-Molybdenum Alloys, Rensselaer
its critical level at a temperature of 1800°F was Polytechnic Institute, Troy, N. Y., 1957.

Table 3.1.1. Maximum Permissible Interface Temperatures of Nickel-Molybdenum

Alloys in Air, in Sodium, and in Fuel 107 for 1000 hr of Operation*

Alloy
Alloy 1 emperat ure ot bnvironmen t n->
Compos ition
Designation Fuel 107 Sodium Air
(wt %)

20 Mo-bal Ni INOR-1 1850** 1650 1300

17 Mo-6 Fe_bal Ni 1800** 1650 1300

17 Mo-3 Nb-5 Fe-bal Ni INOR-9 1800** 1650 1300

24 Mo-5 Cr-5 Fe-bal Ni Hastelloy W 1800** 1650 1400

17 Mo-7 Cr-5 Fe-bal Ni INOR-8 1700** 1650 1800

17 Mo-7 Cr-5 Fe-3 Nb-bal Ni INOR-8 1700** 1650 1800

(modified)

17 Mo-10 Cr-bal Ni 1650 1650 1800

*Based on an allowable 4 mils of corrosion in 1000 hr.

**Estimates based on tests at 1650 or 1750°F.

142
 PERIOD ENDING DECEMBER 31. 7957

Aging Characteristics. — Of the three elements, ature would be permissible for INOR-9 than for
titanium, aluminum, and beryllium, which cause the INOR-8.
basic nickel-molybdenum alloy to respond to an
The heats of INOR-9 that are being investigated
aging heat treatment, beryllium was found to be
have the nominal composition 12-19% Mo, 3-5% Nb,
the most potent. Alloys containing beryllium in
2-5% Fe, 0.06-0.30% C, bal Ni. The results
amounts as low as 0.25% exhibited strong aging
obtained thus far show that (1) the maximum per
tendencies at 1300°F. The aging response of missible niobium content for forgeability is about
alloys containing titanium and aluminum at 1300°F
7.5%, (2) the solid solubility of niobium up to
was significant, while at 1500°F these alloys 1700°F in the nickel-molybdenum base is 5%, (3)
either overaged in a short time or showed no effect. the creep strength is about equal to that of INOR-8,
and (4) the oxidation rates of the INOR-9 alloys are
Properties of INOR-8 slightly higher than those of Hastelloy B.
It appears at this time that the alloy INOR-8,
which is based on the Ni-Mo-Cr-Fe system, offers Phase Diagram Studies
the best choice of properties for fulfilling the An investigation is being conducted at the
requirements of a structural material compatible University of Tennessee (subcontract 582) for
with fused fluorides. The current optimum com the purpose of establishing phase relationships
position range of this alloy is: 15-18% Mo, 6-8% and characteristic transformations in the binary
Cr, 4-5% Fe, 0.04-0.08% C, 0.8% (max) Mn, nickel-molybdenum and ternary nickel-molybdenum-
0.5% (max) Si, bal Ni. The properties of this chromium systems. In the binary system, three
alloy were summarized and discussed previ secondary phases have been identified and their
ously.1'4 transformation temperatures have been determined,
An investigation of the effect of varying com as stated previously.
position on the mechanical properties of alloys of
the INOR-8 type is under way. Eighteen alloys Solubility limits of molybdenum plus chromium in
have been prepared and tested in creep rupture at a nickel are being determined, and identification of
stress of 10,000 psi and a temperature of 1500°F the phases observed is to be undertaken. The
in argon. Results of these tests, although in compositions studied range from 20 to 35% Mo and
complete, show that (1) increasing the molybdenum 3 to 15% Cr. Apparently the addition of chromium
content is the most effective means of improving suppresses the formation ofthe beta phase (Ni4Mo),
the creep properties, (2) increasing the chromium inasmuch as it has not been observed in any of the
content from 5 to 7% without changing the molyb ternary compositions.
denum and iron content produces only a slight im
provement in the creep properties, and (3) there is Stress-Rupture Testing by New England
no obvious correlation of the time to rupture of the Materials Laboratory
material with the total content of alloying elements The production heats of INOR-1 through -6,
in the nickel base. which were prepared by the International Nickel
Company, were stress-rupture tested at 1350,
Properties of INOR-9 1500, and 1650°F in air at the New England
Materials Laboratory. A final report is being pre
Alloys designated INOR-9 have been studied to
pared in accordance with the terms of the con
determine whether they may be suitable as possible
tract. The data obtained on stress-rupture strengths
substitutes for INOR-8. The base composition of
of these alloys were summarized in a previous
these alloys is that of INOR-1 (a nickel-molybdenum
binary alloy) with niobium and carbon added for report.4
strength. Since the alloys of this group do not
Status of Production Heats
contain elements that corrode in fuel 107, it is ex
pected that a higher metal-salt interface temper- A purchase order was placed with the Inter
national Nickel Company for the production of
4800-lb heats of six nickel-molybdenum-base
4H. Inouye, T. K. Roche, and J. H. Coobs, ANP Quar.
Prog. Rep. Sept. 30, 1957, ORNL-2387, p 161. alloys so that experience might be gained with this

143.
ANP PROJECT PROGRESS REPORT

type of material, The alloy compositions are cold rolling. Specimens were prepared for creep-
given below: rupture studies and for density and thermal ex
Alloy Nominal Composition (wt %)
pansion measurements. The density ofthe material,
as determined by weighing a machined specimen,
INOR-1 20 Mo-bal Ni
was found to be 8.795 g/cm . Thermal-expansion
measurements were made from room temperature to
INOR-2 16 Mo-5 Cr-bal Ni
1832°F with the cooperation of the Ceramics
INOR-3 16 Mo-1 AI-1.5 Ti-bal Ni Group of the ORNL Metallurgy Division. The data
INOR-4 16 Mo-2 AI-1.5 Ti-bal Ni
obtained are presented in Fig. 3.1.1. The co
efficients of linear thermal expansion, a, based on
INOR-5 13 Mo-5 Cr-2 W-2 Nb-0.1 C-bal Ni the data of Fig. 3.1.1 are given below:
INOR-6 16 Mo-5 Cr-1 AI-1.5 Ti-bal Ni
Temperature Average a.
In general, billet conversion of these alloys has Range (°F) (in./in..°F)
been only moderately successful. Difficulties were
encountered with cracking during the rolling of the xlO"5
titanium- and aluminum-bearing INOR alloys 4 and
212-752 0.75
6. INOR alloys 1, 2, and 5 were processed with
considerably more success. Strength and weld 752-1112 0.73

ability studies were conducted on the converted 1112-1832 1.02

products received at ORNL. Recently, after some
212-1832 0.87
delay, the tube reducing of the tube shells was
completed, and these shells are presently being
redrawn to small-diameter seamless tubing. Cur
TEMPERATURE (°F)
rently, the International Nickel Company is engaged 752 IH2 1472

in the conversion of the remaining billet stock to

items required for testing.
//
A 10,000-lb air-melted heat of INOR-8 was if
jf
purchased from the Haynes Stellite Company to It

supply the immediate material needs of the various //

//
test programs. Casting and fabrication of the c t RREC" ED /
TRUE EXPANSION ti NtRUE EXPANSION
alloy by hot and cold working to bar, plate, sheet, FOR (-in. LENGTH
V 0.989-in. LENGTH
and strip products was accomplished with little or \y |
no difficulty. Various shapes of this material
have been received, including forged bar, hot- INS TRUMENT
CORRECTION

rolled plates, hot-rolled sheet, hot-rolled bar, six ff

;urvi

production-size extrusion billets, and rolled strip

for the processing of Weldrawn tubing.
Recently, a review was made of the progress to
date on Westinghouse Electric Corporation sub -PLO TTED )IAL

contract 1067, and the first periodic report on FOR 0.9 89-in LENG TH

this developmental work was submitted to ORNL. |

•
1
In summary, three heats of the specified INOR-8 i
i— i MOR-8
composition were induction-melted in air. Diffi I COMPOSITION (wt%): I
/ .. Mo Cr Fe C Mn Si Ni ...
culties were encountered in forging castings of 16.90 6.86 4.21 0.14 0.84 0.23 BAL
£r
these heats because of internal flaws formed during
400 600 800
cooling. Various changes in mold design and TEMPERATURE CO

melting practice were made to overcome this, as

described in detail previously. Fig. 3.1.1. Linear Thermal Expansion for 1-in.
A sound 3-in.-dia forging from the initial high- Length of INOR-8 as Determined with a Vacuum
carbon heat of INOR-8 prepared by Westinghouse Dilatometer Direct-Reading Dial Gage. Sample used
was received for evaluation. The material was was only 0.989 in. long, and data were interpolated to
successfully worked by extrusion and by hot and values for a 1-in. length.

144
 PERIOD ENDING DECEMBER 31, 7957

Creep-rupture tests on the air-melted and vacuum- properties of the Hastelloy alloys was carried out.
remelted material, as shown previously, indicated It was found that these alloys are not single-phase
that definite improvements in time to rupture and in and that they suffer a serious loss of ductility as
elongation were realized through vacuum remelting, a result of the precipitation of a brittle beta phase
but that the creep rate of the air-melted alloy was in the reactor operating temperature range. These
superior to that of the vacuum-remelted material. alloys also presented fabricational difficulties,
particularly in the production of seamless tubing,
Composite Tubing and Plate and therefore it was considered that the best
In order to take advantage of the corrosion re solution would be to change the alloy composition.
sistance of the nickel-molybdenum alloys in fuel Therefore, special alloys designed to retain the
107 at temperatures exceeding 1650°F, it would inherent advantages of the Hastelloys and to
be necessary to isolate the nickel-base alloy from eliminate their objectionable characteristics were
the sodium through the use of composite tubing. studied.
Preliminary studies of INOR-8 and type 316 stain Screening tests were made of the creep-rupture
less steel diffusion couples show that an unknown properties of various experimental alloys exposed
phase or phases precipitate at the interface as a to fuel 107. For comparison, each alloy was tested
result of diffusion. The extent of the interface re with a stress of 8000 psi at 1500°F in two con
action is negligible after diffusion times of 500 hr ditions: (1) annealed at 2000°F for 1 hr and (2)
at 1300°F. As expected, the severity of reaction annealed at 2000°F for 1 hr and aged for 50 hr at
increases rapidly as the temperature is increased 1300°F. Results of the creep-rupture tests, pre
to 1800°F. The precipitate is confined to the viously reported, included data obtained for
nickel-base side of the interface and occurs as a air-melted heats prepared by Battelle Memorial
grain-boundary phase rather than as a continuous Institute, International Nickel Company, Westing
layer. Room-temperature tensile tests of composite house Electric Corporation, Superior Tube Company,
sheet showed that strengthening occurred after and Haynes Stellite Company, as well as for
diffusion heat treatments at 1300 and 1500°F. A vacuum-melted heats produced at ORNL.
weakening effect was observed in those specimens Attempts were made to increase the strength of
heat treated at 1650 and 1800°F. A minimum the nickel-molybdenum alloys by the addition of
fracture ductility of 28% elongation was observed elements that form complex precipitates and by
in the specimens heat treated at 1650°F. the addition of elements that are simple carbide
The feasibility of fabricating composite tubing of formers. Additions of very small amounts (<1%)
Inconel and type 316 stainless steel was described of aluminum and titanium did not seem to increase
previously. It is anticipated that similar pro the strength, but did increase the elongation be
cedures can be applied to the fabrication of com fore fracture. Fabrication difficulties were en
posites of INOR-8 and type 316 stainless steel. countered when additions of aluminum and titanium
exceeded 1.5%. It was possible, however, to
improve the creep characteristics by additions
MECHANICAL PROPERTIES INVESTIGATIONS
of carbide-forming elements, such as niobium
C. R. Kennedy5 J. R. Weir and tungsten, without seriously compromising
R. W. Swindemann J. W. Woods the other good characteristics. The BMI alloys
Commercial nickel-base, molybdenum-containing
alloys, such as Hastelloy B or W, have been con
sidered for reactor applications because of their °C. R. Kennedy, D. A. Douglas, and W. D. Manly,
High Temperature Mechanical Properties of Hastelloy B
excellent resistance to chemical corrosion and and W, ORNL-2402 (in press).
their relatively high strength at elevated temper 7C. R. Kennedy, ANP Quar. Prog. Rep. Dec. 31, 1956,
atures. Because such alloys showed promise as ORNL-2221, p 249.
structural materials for a circulating-fuel reactor, 8J. W. Woods, ANP Quar. Prog. Rep. March 31, 1957,
ORNL-2274, p 220.
an extensive investigation of the mechanical
9H. Inouye, T. K. Roche, and J. H. Coobs, ANP Quar.
Prog. Rep. June 30, 1957, ORNL-2340, p 176.
10J. W. Woods and D. A. Douglas, ANP Quar. Prog.
On assignment from Pratt & Whitney Aircraft. Rep. Sept. 30, 1957, ORNL-2387, p 171.

145
ANP PROJECT PROGRESS REPORT

containing carbon in excess of 0.1% exhibited UNCLASSIFIED

ORNL-LR-DWG 27059
good high-strength properties, but they tended (0

to fracture when subjected to tube-fabricating

_ 8
processes. Additional alloying elements, such as
chromium (7%) and iron (4%), were found to im &
1 6 0/ .^HASTELLOY B RUPTURED
prove the creep properties of the basic nickel- 5 / AT 2750 hr WITH 4.6%
#/
molybdenum binary alloy.

The Haynes Stellite INOR-8 (nominal composition,

-i
/agmsSi en INOK-g
/
-H
/ 1 1 1
1
(Wril L BEING TESTED)
n
1

in wt %: 16 Mo-7 Cr-5 F-0.02 C-bal Ni) appears 1

800 4000
to be the best alloy produced to date. It has good 200 400 600
TIME (hr)
corrosion resistance and fabricating characteristics,
but, as a result of its low carbon content, the Fig. 3.1.3. Creep Properties of Untreated INOR-8 and
strength properties are only fair. It is considered Carburized INOR-8 in Comparison with Those of
to be feasible to substantially increase the strength Hastelloy B Exposed to Fuel 107 at 1650°F at a Stress
of the alloy after the difficult fabrication processes of 5000 psi. (Secret with caption)
are completed by adding more carbon through gas
carburizing. Tests of specimens in which carbon UNCLASSIFIED
ORNL-LR-DWG 27060
was added by this means showed greatly improved 10

strength properties.
_. 8

A comparison of creep properties of untreated

alloys and the same alloys after carburization with 1 6 c >/
5
the properties of Hastelloy B are presented in i 4
a>

0
Figs. 3.1.2, 3.1.3, and 3.1.4. It is evident that -I
5

(N0R-» 1
carburized INOR-8 becomes equivalent to Hastelloy ^ C£rbur££2,
B in creep strength at 1650°F. _ HAS

200 400 600 800 1000

TIME (hr)

UNCLASSIFIED
ORNL-LR-DWG 27058 Fig. 3.1.4. Creep Properties of Untreated INOR-8 and
Carburized INOR-8 in Comparison with Those of
Hastelloy B Exposed to Fuel 107 at 1800°F at a Stress
of 3000 psi. (Secret with caption)
0/
-<&
.< 9X

t
c 3/ *&
d^Z CORROSION STUDIES

Z 6 J. H. DeVan
3
J. R. DiStefano R. S. Crouse

HASTELl OYB
Two related series of tests have been made in
12.00U nr —
order to evaluate the corrosion properties of
nickel-molybdenum alloys in fused fluoride salts
800 1600 2400 4000 and in sodium. In the first series of tests, com
TIME (hr)
mercial alloys, such as Hastelloy B and Hastelloy
W, were studied, while in the second series data
Fig. 3.1.2. Creep Properties of Untreated INOR-8 and were obtained on the corrosion properties of ex
Carburized INOR-8 in Comparison with Those of perimental nickel-molybdenum alloys, which from
Hastelloy B Exposed to Fuel 107 at 1500°F at a Stress the standpoint of heat treatment and chemical
of 8000 psi. (Secret with caption) properties are better adapted to reactor operation.

146
 PERIOD ENDING DECEMBER 31, 7957

Forced-circulation loops were used for most of the fluid are summarized in Table 3.1.3. A comparison
tests of the Hastelloys, but both thermal-convection of these results with the results for other classes
and forced-circulation systems have been used to of alloys tested in sodium is presented in Fig. 3.1.5,
study the experimental alloys. in which the weights of cold-zone deposits are
indicated as a function of temperature. As may be
Forced-Circulation Loop Tests of Hastelloys seen, the Hastelloys and the nickel-base alloys,
The results of forced-circulation loop tests of in general, are considerably more susceptible to
Hastelloy B with fuel 107 as the circulated fluid mass transfer in sodium at temperatures above
are summarized in Table 3.1.2. Further details of 1300°F than are either the austenitic or the ferritic
these tests were presented in previous reports. ' stainless steels.
While defects present in the as-received Hastelloy
B tubing precluded comprehensive analyses of the Forced-Circulation Loop Tests of Experimental
test results, it was demonstrated in these tests Nickel-Molybdenum Alloys
that basic nickel-molybdenum alloy systems are
Two forced-circulation loops constructed of
compatible with alkali-metal fluorides at temper
experimental nickel-molybdenum alloys have been
atures up to 1760°F.
successfully operated for 1000 hr with fuel 107.
The results of forced-circulation tests of Hastelloy
The first of these loops, 7641-9, was discussed
B and Hastelloy W with sodium as the circulated
previously. The loop was fabricated from an
J. H. DeVan, ANP Quar. Prog. Rep. Sept. 10, 1956. alloy containing 17% Mo-6% Fe-bal Ni and was
ORNL-2157, p 136.
12J. H. DeVan and R. S. Crouse, ANP Quar. Prog. 13J. H. DeVan and R. S. Crouse, ANP Quar. Prog.
Rep. March 31, 1957, ORNL-2274, p 147. Rep. Sept. 30, 1957, ORNL-2387, p 176.

UNCLASSIFIED
(DRNL-LR-DWG 27061
24

20 HAb 1 t ILLOY WJ
/

LU i j
/
16 /
o / yiNcc)NEL
o /

12 /
HASTELLC)Y B/
co
o
a.
LU
/ / / s'lNCONE _ X

a
/ >
8 /
/
x
CD
^TYPE 330 STAIh LESS STE EL
/
^ >*Vtype 310 STAINLESS STE EL
4 s
s
—' r"V/TYPE 430 STAINLESS STE EL
/
•

/
X TYPfiS 347, 31 61304-,,",
r-^ TY PE 316 STAINI FSS STEEL-,

0
1200
-'

1300
£-\ ' SI"AINLESS
1400
STEEb.
-^v
1500
-"t
— -

1600
«

1700
MAXIMUM FLUIDTEMPERATURE (°F]

Fig. 3.1.5. Comparison of Results of 1000-hr Forced-Circulation Loop Tests of Various Alloys with Sodii
All loops contained oxide cold traps and were operated with a temperature drop of 300°F.

147
 i Table 3.1.2. Operating Conditions and Results of Forced-Circulation Loop Tests of j>.
Co Hastelloy B with Fuel 107 as the Circulated Fluid ^
—— -o

Maximum Fuel- Maximum Fuel o

Loopr Period of u, .. .Interface
Wail
, d n cFue i
Ru k
t *
Tf>mnf>rntiir«.
Reynolds
«-^=jr.iv..us>
Flow As-Received Size
Number Operation _ __ ,..,, . Number Rate and Condition Metallographic Results o
Temperature Temperature Differential .J
(hr) .Op> .op, ,Op, (fps) of Tubing
— O

7425-13Aa 1000 1555 1505 200 10,000 3.2 Size: 0.5 in. OD, 0.035 in. Small pits found along hot-leg £
wall; as-received speci- surface to a depth of 1.5 mils; [v
mens had surface pits to no cold-leg deposits or
a depth of 1 mil layers "1
"a

7641-1B 1000 1767 1635 280 10,000 2.6 Size: 0.5 in. OD, 0.065 in. Many pits in hot leg to a depth °
wall; inner surfaces of as- of 4 mils; pits to a depth of 7 ~*
received tubing had pits in mils in vicinity of weld in
some areas to a depth of 5 cold leg; no cold-leg deposits
mils or layers

7641-2B 1013 1710 1640 295 5,500 1.2 Size: 0.5 in..OD, 0.065 in. Many pits in hot leg to a depth
wall; tubing from same lot of 4 mils; pits to a depth of 7
as that used for Loop mils in cold leg; deposits of
7641-1B fine metal crystals in small
area of cold leg up to 0.5 mil
thick

7641-3B 407.5fo 1750 1660 300 10,000 2.6 Same as loop 7641-1B Many pits in hot and cold legs
to a depth of 3 mils; no de
posits or layers

7641-4E 180c 1710 1645 293 20,000 5.1 Size: !/-in. sched-40 pipe; Loop was shut down and re-
inner surfaces badly paired four times because of
pitted and cracks present leaks; cleaning fills omitted
on both sides of longitu- on third and fourth runs; a
dinal closure weld silver-colored loosely ad
herent layer found throughout
loop; layer consisted primarily
of fuel, but contained approxi
mately 10% Ni and 5% Mo

Suffix letters designate number of loop shutdowns prior to termination.

Terminated because of pump failure; tubing ruptured upon attempting to restart circulation.
Terminated because of tube failure.
 PERIOD ENDING DECEMBER 31, 7957

Table 3.1.3. Operating Conditions and Results of Forced-Circulation Loop Tests of Hastelloy B
and Hastelloy W with Sodium as the Circulated Fluid

Temperature differential: 300 F

Fluid flow rate: 2.5 gpm

Ma ximum Thickness Total Weight

Maximum Fluid Operating Maximum Hot-
Loop Loop of Cold-Leg of Cold-Leg
Temperature Time Leg Attack
Number Material Deposits Deposits
(°F) (hr) (mils)
(mils) (g)

7641-1 Hastelloy B 1500 1000 2 27 17.3

7641-2 Hastelloy B 1300 1000 2 3 0.45

7641-3 Hastelloy B 1500 2000 2 30 20.6

7641-4 Hastelloy W 1500 1000 1. 50 22 20.6

operated with a maximum fluid-metal interface comparable with those normally found in Inconel —
temperature of 1760°F. Attack occurred along the fuel 30 systems tested at 1500°F. The chromium
hot-leg surfaces in the form of intergranular voids content of the fuel was approximately 685 ppm,
and was limited to a maximum depth of 4 mils. and chromium was the impurity that was present in
the largest quantity.
A second loop constructed of an alloy containing
17% Mo-10% Cr-bal Ni has now completed oper It is obvious on the basis of these test results
ation with a maximum fluid—metal interface temper that chromium additions in amounts of 10% or
ature of 1850°F. Other operating conditions for greater would limit the usefulness of nickel-
the loop, 7641-10, are given below: molybdenum alloys in fuel 107 to temperatures
below 1800°F. Thermal-convection loop test
results appear to have established a satisfactory
Maximum bulk fuel temperature
operating temperature of 1650°F, as discussed
Fuel temperature drop below.

Reynolds number 10,000

Flow rate 1.4 gpm Thermal-Convection Loop Tests of Experimental

Ratio of heated section surface 2.1 in.2/in.3 Nickel-Molybdenum Alloys
area to total loop volume Approximately 50 thermal-convection loops
fabricated from alloys representing numerous
compositional variations of the 83% Ni-17% Mo
Extensive attack was found metallographically
base have been operated with fuel 107 for periods
throughout the hot leg of the loop that reached, at
up to 1000 hr and temperatures up to 1650°F. The
the point of maximum wall temperature, to a depth
primary purpose of these tests was to evaluate
of 23 mils. The attack was predominantly in the
the effect of specific alloying elements on the
form of intergranular voids and was accompanied
corrosion resistance of the base composition.
by removal of second-phase material from the grain
matrix, as well as from the grain boundaries. Most of the loops were operated with a maximum
Cold-leg sections showed little attack, but there bulk fuel temperature of 1500°F and with a corre
was a thin metallic deposit over most of the ex sponding 1650°F metal-fluid interface temperature.
posed surface. The quantities of metal con The results of these tests were summarized in a
taminants in fuel samples taken after the test were previous report. From the test data it has been

149
ANP PROJECT PROGRESS REPORT

possible to stipulate approximate limits of alloy compositions and their approximate temperature
additions which can be permitted from the stand- limitations for operation with fuel 107 were pre-
point of corrosion penetration and resistance to sented in Table 3.1.1. Operation at temperatures
thermal-gradient mass transfer. As would be only 100°F lower than those shown has been found
expected, the limits are quite dependent on oper- to reduce the corrosive attack for the same time
ating temperatures. A list of several promising period to a depth of 1 to 2 mils.

150
 PERIOD ENDING DECEMBER 31, 1957

3.2. MECHANICAL PROPERTIES STUDIES

D. A. Douglas
C. R. Kennedy1 J. R. Weir
R. W. Swindemann J. W. Woods

An extensive investigation has been under way Details of these investigations have been re
during the past several years in which the me ported.
chanical properties of the various materials to Since the strength properties of Inconel make it
be used in circulating-fuel aircraft reactors were only marginally useful at peak aircraft reactor
evaluated under simulated service conditions. Par operating temperatures, some thought has been
ticular attention has been given to obtaining given to a modification which would increase the
design data and to studying the behavior of the strength without impairing the excellent fabricating
materials under the various conditions of stress, characteristics. Carbon is a very potent
temperature, and environment that will be en strengthening agent because it disperses carbides
countered during operation of the reactor. Fairly throughout the matrix and in the grain boundaries,
complete design data were obtained for Inconel, but, unfortunately, such a structure is difficult
the structural alloy chosen for the ART, for to draw into tubing. It was believed to be feasible,
beryllium, the moderator material, and for various however, to fabricate the tubing and then add
shielding materials. Basic information on the additional carbon through carburization. Several
behavior of metals has also been obtained in order creep specimens were therefore pack-carburized
to be able to predict the safe operating life of and tested in fuel 30 and in argon. The results
various components under a variety of operating of these tests, as presented in Figs. 3.2.1, 3.2.2,
conditions. and 3.2.3, show that a very marked improvement in
creep properties was obtained.
DESIGN DATA

Inconel J. R. Weir, D. A. Douglas, and W. D. Manly, Inconel

as a Structural Material for a High-Temperature Fused-
Creep and stress-rupture data were obtained for Salt Reactor, ORNL-2264 (June 4, 1957).
Inconel in the temperature range of 1300 to 1650°F
in argon, in sodium, and in the fuel mixture UNCLASSIFIED
ORNL-LR-DWG 27062
NaF-ZrF,-UF,
4 4
(50-46-4 mole %, fuel 30). The 12 I I
data for an argon atmosphere were used for com i

parison. Variables such as grain size, section

thickness, and alloy consistency were investi
gated. It was found that in the fused-salt en
vironment the fine-grained material is stronger -

than the coarse-grained metal at 1300°F but

$€
weaker at 1500°F and above. However, in argon
J= 6 4
sS
0^
and in sodium the fine-grained Inconel maintains <
QC $&
its superiority over the coarse-grained material at
temperatures up to 1650°F, where metallurgical 9-^ f
instabilities begin to introduce adverse effects. .

I
A comparison of three heats of Inconel showed i
that quite wide variations in the properties of the
alloy may be expected. There is not a large - -

.enough difference in the major alloying elements I

50 100 150 200 250

to account for the lack of uniformity; therefore it
TIME (hr)
appears that the source of the difference is the
variations of the minor additions or of the residual
Fig. 3.2.1. Effect of Carburization on the Creep
elements which are difficult to control in air melts.
Properties of Inconel Tested in Fuel 30 at 1500 F and
On assignment from Pratt & Whitney Aircraft. a Stress of 5000 psi. (Secret with caption)

151
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27063 prior to operation. Therefore a study was under
14
taken in order to establish the temperatures
12
i required for recrystallization after varying amounts
of cold working and to determine the stable grain
10
1 size achieved after high-temperature anneals.
I The testing procedures used and illustrations of
B8 (?,
T the grain sizes produced in recrystallized struc
<

h- 6 tures have been presented elsewhere. It was

CO
1 noted in this study that the recrystallization
/ ^
K<&
fcV
4
/ O*" temperature ranged from 1200 to 1650° F and was
2
/ dependent on the amount of cold working and the
time held at temperature. Fine-grained Inconel
'
0 recrystallized at a temperature about 100°F lower
100 200 300 400 500 600 700 800 900 1000
than the temperature at which the coarse-grained
TIME (hr)
material recrystallized for the same amount of prior
Fig. 3.2.2. Effect of Carburization on the Creep deformation and time at temperature. The grain
Properties of Inconel Tested in Fuel 30 at 1650 F and size that resulted from high-temperature treatments
3000 psi. (Secret with caption) was found to be dependent on the degree of
working prior to recrystallization up to a temper
UNCLASSIFIED ature of 1800°F. However, at 1900°F only very
ORNL-LR-DWG 27064
5 small differences in the final grain sizes could
1 be seen regardless of the prior treatment.
4
\ Beryllium
<
1 Extensive mechanical tests of beryllium have
ce
I-
</> o2
I been made in the temperature range of 800 to
1500°F by the Brush Beryllium Company under
1 s
^c ARE3UR IZE D
subcontract P.O. W34X-75581.5 Also tests were
1
made of samples trepanned from the actual be
0 100 200 300 400 500 600 700 800 900 1000 ryllium moderator blocks prepared for the ART.
TIME (hr)
The mechanical properties of the actual blocks
Fig. 3.2.3. Effect of Carburization on the Creep
were in excellent agreement with those obtained
Properties of Inconel Tested in Fuel 30 at 1800 F and
in the prior test program for material of like
1000 psi. (Secret with caption) composition. Thus, it appears that data obtained
in the test program can be reliably applied to
large-scale production pressings.
Recrystallized Inconel
A stable metallurgical microstructure is required EXTRAPOLATION OF TEST DATA

in order to obtain optimum-strength properties in

an alloy. Tests have demonstrated that metals One of the problems associated with mechanical
which have been cold worked to various degrees testing work is that there is frequent need for data
may recrystal lize under certain conditions of under conditions outside the ranges of temperature
stress and temperature and that deleterious effects or stress that have been explored. Thus, the first
on the creep properties will result. Since many step in being able to anticipate the behavior of
reactor components are fabricated by pressure metals under service conditions is to develop a
forming methods and are assembled by welding reliable method of interpolating and extrapolating
and brazing, portions of the structural metal will
have experienced a variety of thermal treatments D. A. Douglas and J. R. Weir, Recrystallization and
Grain Growth of Inconel, ORNL-2406 (in press).
R. G. 0'Rourke et al., Mechanical Properties of
°J. R. Weir, Jr., ANP Quar. Prog. Rep. Dec. 31, 1956. Reactor-Grade Beryllium at Elevated Temperatures,
ORNL-2221, p 244. COO-312 (Aug. 1956).

152
 PERIOD ENDING DECEMBER 31, 7957

the available data. There has been much work has been formulated to relate the data, which fall
done in this field, and there are several relation on a straight line when plotted as log e vs log N.
ships which appear to have merit. The multitude Similar behavior was observed by Coffin in his
of metallurgical variables involved in high- work on thermally induced cyclic strains in type
temperature deformation processes made it ad 347 stainless steel.
visable, however, to derive empirical equations At 1500 and 1600°F, fine-grained Inconel is
based on fairly specific data. Such equations are superior to coarse-grained material; that is, the
available for Inconel, and the method for using data indicate that fine-grained specimens endure
them and the limitations of the results are de a greater number of cycles at a given strain-cycle
scribed in a topical report. level before failure. At 1300°F, however, the
difference in cycles to rupture between coarse-
THERMAL STRESS INVESTIGATIONS and fine-grained Inconel is negligible.
The influence of geometry is not great. Rod
The thermal cycles which occur in a reactor
specimens appear to endure slightly more cycles
during its operation induce in the structural before failure than tube specimens. However,
members corresponding stress and strain cycles.
this improvement may be attributed to a less
Thus for safe operation it is essential to know
sensitive method of detecting the initiation of
the effect of these strain cycles on the endurance
failure in the case of rod specimens.
of the structural materials. In order to obtain
The effect of temperature on these properties is
design data of this type, tests were conducted to
not appreciable over the temperature range studied.
determine the strain-cycle and stress-relaxation
Specimens cycled at 1600°F exhibit only a slightly
properties of Inconel. Important factors which
greater number of cycles to failure than specimens
have been investigated in conjunction with these
tested under comparable conditions at 1300 and
tests are the effects of temperature, grain size,
1500°F.
environment, specimen geometry, and cycle fre
Tests have also been conducted on Inconel tubes
quency. Testing devices were constructed for
in fuel 30 at 1500 and 1600°F. Results of these
testing both tube and rod specimens. In order to tests were presented in previous reports in this
gain a clear picture of the true influence of strain
series. ' The effect of fuel is to reduce the
reversals, these rigs were designed to mechanically
number of cycles to failure by 40 to 66% of the
strain the specimens; whereas, under true oper
number of cycles in argon, depending on the plastic
ating conditions, the strains may be produced by strain per cycle and the grain size.
thermal fluctuations. Supporting investigations in
which thermally induced strain cycles are being Since the frequency of induced-strain cycles
studied are under way at the University of Alabama. within a reactor will not be uniform, a program
Strain-cycle tests in argon at 1300, 1500, and has been set up for determining the effect of
1600° F were completed, and the data were pre frequency on the strain-cycle properties. Pre
sented previously.7-9 Curves have been drawn viously, all tests had been conducted with a
which indicate the number of strain cycles to
2-min/cycle period. Present investigations '
failure, N, as a function of plastic strain per are being conducted with a 30-min/cycle period
cycle, e . An empirical equation, on Inconel tube specimens at 1500°F, and there
are indications that frequency may be an important
factor. Specimens cycled at small strains and
"% = C • 30-min/cycle periods do not endure as many cycles
as those cycled with the 2-min/cycle period.
C. R. Kennedy and D. A. Douglas, Use of a Rate Further investigations will have to be completed
Equation for the Extrapolation of Creep Data, ORNL
CF-57-8-119 (Aug. 20, 1957). before this effect can be entirely substantiated
and understood.
C. R. Kennedy and D. A. Douglas, Plastic Strain
Absorption as a Criterion for High-Temperature Design,
ORNL-2360 (in press).
L. F. Coffin, Jr., Trans. Am. Soc. Mech. Engrs. 76,
8J. R. Weir, Jr., ANP Quar. Prog. Rep. Dec. 31, 1956, 931-50 (1954).
ORNL-2221, p 246.
9C. R. Kennedy and D. A. Douglas, ANP Quar. Prog. "C. R. Kennedy and D. A. Douglas, ANP Quar. Prog.
Rep. June 30, 1957, ORNL-2340, p 190. Rep. Sept. 30, 1957, ORNL-2387, p 182.

153
ANP PROJECT PROGRESS REPORT

Results from the tests being made at the degree of decay or relaxation if the temperature
University of Alabama indicate that mechanical is sufficiently high This relaxation of stress
strain-cycle effects at 1300°F correlate well with occurs through a conversion, by creep, of the
thermally induced strain-cycle behavior at the elastic strain into plastic strain. In order to
same mean temperature. establish the true plastic strain induced by a
Additional tests were made in order to determine thermal loading, the relaxation characteristics of
the effect of carburization, weld metal, and braze- a material must be known and taken into consider
metal coatings on the strain-cycle properties of ation.
Inconel tubes. Tests conducted on carburized Relaxation tests of Inconel were completed for
Inconel tubes showed no change in strain-cycle coarse- and fine-grained specimens at 1100, 1300,
characteristics as a result of the carburizing 1500, and 1650° F, and the test procedures,
treatment. A series of tests on all-weld-metal apparatus, and results have been described.
tubes was cancelled before completion, but the The data obtained in these tests indicate that,
preliminary results indicated that the welds at temperatures greater than 1300°F, relaxation
were stronger in tension than the base metal at of stress occurs rapidly. The relaxation rate at
temperatures from 1400 to 1800°F. Several Inconel these temperatures varies only when the initial
tube specimens coated with braze metal (92% Ni— strain is below the elastic limit. Moreover,
3% Si-3% B-2% Fe) were tested in order to induced plastic strains do not significantly in
determine whether boron penetration of the base crease the elastic limit, since recovery appears
material would affect the strain-cycle properties. to be rapid at temperatures above 1300°F.
Rupture data from these tests at 1500°F were in Stresses are reduced to low residual values after
good agreement with the design curve for coarse a few minutes. Below 1300°F, however, relaxation
grained Inconel at 1500°F. Consequently, it may occurs less rapidly. At 1100°F high initial
be assumed that no detectable impairment of strains produce work hardening, which results in
strain-cycle characteristics will be produced by larger elastic strains and higher stress values;
the presence of a braze-metal coating. these strains, in turn, relax slowly. The results
A series of tests for evaluating the effect of also give indications of a grain-size effect. The
prior strain cycling on the creep properties of relaxation rates of coarse-grained Inconel at all
Inconel was also conducted. The procedure and temperatures were lower than the rates for fine
initial results were described previously. In grained material.
summary, the available data indicate that the Fatigue
creep strength and rupture life are significantly
reduced by prior strain cycling and that, with Under a subcontract to Battelle Memorial In
increasing consumption of strain-cycle life, the stitute, a program was initiated to investigate the
creep-life expectancy is correspondingly reduced. effect of cyclic stresses imposed on a statically
The inference is that design calculations based loaded system. The aim of this program was to
on only one of these conditions would be opti define the stress conditions under which fatigue
mistic for instances where both strain cycling is the dominant failure mechanism and the stress
and creep occurred. conditions under which creep may be regarded as
Beryllium is the only other reactor material for the controlling mechanism for failure. Preliminary
which strain-cycle characteristics have been in tests for checking the performance of the apparatus
vestigated. A series of tests was conducted at and the reproducibility of results were satis
1250°F on hot-pressed, reactor-grade beryllium, factorily completed. At present, tests are being
and the results were reported previously. The conducted to determine the influence of stress-
test data revealed that beryllium, although much cycle frequency on the fatigue properties.
inferior to Inconel, nevertheless, does have
Biaxial Creep
ductility and is capable of withstanding cyclic
strains at 1250°F. Very little is known concerning the influence of
stress distribution on deformation rates, and the
Relaxation

Stresses induced by a thermal cycle in a con C. R. Kennedy, D. A. Douglas, and W. D. Manly,

Relaxation Characteristics of Inconel at 1100 to 1650°F,
strained structural member usually undergo some ORNL-2407 (in press).

154
 PERIOD ENDING DECEMBER 31, 7957

fact that structures are normally subjected to treatment and are compared in Table 3.2.1 with
multiaxial stresses emphasizes the need for ad data obtained for untreated specimens.
ditional research in this area. Therefore a program An alloy of very pure lead containing 0.06%
of tests is being conducted in which tubes are copper has been proposed for use as a gamma-ray
subjected to various ratios of axial and tangential shielding material. Since no mechanical property
stress and the resultant creep rates are measured. data were available at a temperature of 230°F,
A considerable change in rupture life, ranging the expected maximum operating temperature of
from 200 to 2000 hr, and significant variations of the lead shield, a creep-rupture testing program
creep rates have been noted. A considerable was initiated to determine the strength of this
quantity of data must be accumulated before it alloy. A design criterion of a creep rate of 0.5%
will be possible to determine which of the various strain per year for a stress of 100 psi was es
mechanical theories of failure describes the tablished. The results of tests in air show that
mechanism of fracture under these test conditions. this alloy has a deformation rate of about 2% per
year at 100 psi. Specimens tested at 50 and
20 psi showed deformation rates of 0.08% per
TESTS OF SHIELDING MATERIALS year and 0.05% per year, respectively. A lead-
lithium alloy was also investigated in the hope
Creep-rupture testing of an 80% Mg—20% Li alloy
that its excellent room-temperature strength in
proposed as a neutron shielding material was
dicated good high-temperature strength. However,
completed. Data in addition to those presented
it was found that the lead-copper alloy has greater
previously have been obtained at 200°F for
creep strength at 230°F than the lead-lithium
specimens that were given a chemical surface
alloy.16
13 1SR. E. McDonald and C. F. Leitten, Jr., ANP Quar.
C. R. Kennedy and D. A. Douglas, ANP Quar. Prog. Prog. Rep. June 10, 1956, ORNL-2106, p 173.
Rep. Sept. 30, 1957, ORNL-2387, p 185.
D. H. Jansen, E. E. Hoffman, and D. M. Sheppard,
14
J. R. Weir and C. W. Dollins, ANP Quar. Prog. Rep. Lead-Lithium Shielding Alloy — Metallurgical Studies,
Sept. 10, 1956, ORNL-2157, p 208. ORNL-2404 (in press).

Table 3.2.1. Results of Creep-Rupture Tests at 200°F of an 80% Ma-20% Li Alloy

Time to Indicated Strain (hr)

Specimen Condition Stress (psi)
0.2% 0.5% 1% 2% 5% 10%

Treated 100 3579 6647

200 862 2284 3484

Untreated 600 3.6 25 72 121 420 940

800 0.5 1 71

155
ANP PROJECT PROGRESS REPORT

3.3. WELDING AND BRAZING STUDIES

P. Patriarca
G. M. Slaughter R. L. Heestand

PRODUCTION OF CAST RINGS OF BRAZING of the junctions of all thermocouples be at a

ALLOY consistent and known distance from the end cover.
The problem of sealing the junction end of
P. Patriarca G. M. Slaughter
four-wire Chromel-Alumel thermocouples which
Difficulties were encountered by the York Corp. were housed in Inconel tubing and insulated with
in the consistent brazing of radiator test units magnesium oxide was first approached by welding.
with sintered brazing-alloy rings obtained from A small Inconel plug was inserted into a hole
Coast Metals, Inc. Therefore a decision was which had been prepared by sand-blasting the
made to use brazing-alloy rings cast in graphite magnesium oxide to the desired depth. The
molds for the fabrication of future units. Ex assembled components were then fused together,
periments have indicated that the cast rings are both with and without the addition of filler wire.
not so sensitive to the rate of temperature rise The inadequacy of this and other similar pro
as are the sintered rings and that the required cedures was apparent when radiographic inspection
consistent minimum brazing temperature is some resulted in a rejection rate of approximately 20%
what less. as a result of cracks and pores in the weld zone.
Consequently, a production facility was set up High-temperature furnace-brazing procedures were
by the Engineering and Mechanical Division for then utilized in developing a suitable fabrication
manufacturing approximately 3,000,000 rings ac technique. In initial experiments the magnesium
cording to the procedures previously developed by oxide was excavated for a short distance, the hole
the Welding and Brazing Group of the Metallurgy was filled with brazing alloy, and the unit was
Division. Assistance was given by Metallurgy brazed. As was expected, this procedure resulted
Division personnel in the determination and control in excessive porosity and poor wetting of the
of optimum ring dimensions and weights and in the 2%-aluminum-containing Alumel component of the
establishment of systematic weight-control records. thermocouple. Most test specimens leaked during
A production rate of approximately 60,000 to vacuum testing; thus the seals would have been
70,000 rings per day was achieved, and a total unsatisfactory for use in the corrosive fused salt
of approximately 750,000 rings was manufactured environment. Further, the distance between the
prior to termination of production on September 6, end of the seal and the thermocouple junction
1957. A detailed description of the equipment and varied both in seals made by this method and in
production procedures, including a cost estimate, those made by the welding technique, and such
is being prepared by the Engineering and Me variations could give significant errors in temper
chanical Division. ature measurements during service.
FABRICATION OF THERMOCOUPLES FOR
A method was consequently developed which
HIGH-TEMPERATURE SERVICE
satisfied the requirements as to both integrity of
the seal and consistency in the distance between
P. Patriarca
the end of the seal and the thermocouple junction.
The use of thermocouples in high-temperature A schematic drawing that illustrates the thermo
fused-salt systems requires that a reliable thermo couple fabrication procedure is presented in Fig.
couple junction be made and that a leaktight seal 3.3.1, and the products of the various steps of
be assured. It is also desired that the locations the procedure are shown in Fig. 3.3.2.

JP. Patriarca and R. E. Clausing, ANP Quar. Prog.

'g. M. Slaughter, ANP Quar. Prog. Rep. June 30, Rep. Sept. 10, 1956, ORNL-2157, p 200.
1957. ORNL-2340, p 223.
W. S. Hornbaker, Production of Fused Coast Metals
2G. M. Slaughter, ANP Quar. Prog. Rep. March 31. No. 52 Alloy Brazing Rings, ORNL-2424 (to be pub
1957, ORNL-2274, p 205. lished).

156
 PERIOD ENDING DECEMBER 31, 7957

UNCLASSIFIED
ORNL-LR-DWG 27065

INCONEL SHEATH, 0.254-in. OD

BRAZING ALLOY RING 0.035-in. WALL,BORED WITH N0.I2
OR POWDER (0.04g) DRILL TO TOP OF CAP

EXCAVATED ^% in INCONEL COVER,0.025-in.THICK,

3^6-in.DIA, PUNCHED
MAGNESIUM OXIDE INSULATION

INCONEL CAP,0.025-in.THICK,
n/6„-in.DIA,PUNCHED
^64
AND DRILLED
WITH NO.64 DRILL

THERMOCOUPLE WIRES, 2 CHROMEL

AND 2 ALUMEL, 0.030-in.OD

Fig. 3.3.1. Assembly Diagram of Thermocouple for High-Temperature Service.

After the magnesium oxide is excavated to the FABRICATION OF ART HEAT EXCHANGERS
desired depth as illustrated, the four-hole Inconel
disk is inserted into the hole, the thermocouple G. M. Slaughter
wires are cut off, a solid Inconel disk is placed The fabrication of an experimental ART fuel-to-
on top, and the assembly is brazed in dry hydrogen NaK heat exchanger by Black, Sivalls & Bryson,
at 1080°C. The brazing alloy can be preplaced Inc., Tulsa, Oklahoma, was observed closely in
either as powder or as a cast ring. If a cast ring order to assure the adequacy of the high-temper
is to be used, it is advisable to prime the ring ature brazing procedures utilized. The preplace-
to the joint with a slight amount of powder. ment of Coast Metals brazing alloy No. 52 on the
Radiographic inspection can be used after manu tube-to-header joints was closely examined before
facture to assure that high-quality joints are the unit was placed on the brazing fixture. Cast
obtained. brazing-alloy rings, both circular and elliptical,

157
 PERIOD ENDING DECEMBER 31, 7957

these four thermocouples was reduced to 5°F filler metal added to the joint were found to be
(1915—1920°F). The maximum temperature re free of root cracks.
corded by these thermocouples was 1925°F.
After removal of the heat exchanger from the In back-brazing the welded tube-to-tube-sheet
retort, the brazes on both headers were examined. joints, a unique heating problem existed in that
All visible joints, including those on the two job the tube-sheet temperature continuously lagged
samples, had relatively good fillets, and the unit behind the tube temperature throughout the brazing
was, in general, clean and bright. The extent of cycle because of the much greater thickness of
runoff from one header was such, however, that the tube sheet. When Coast Metals brazing alloy
the bottom row of tubes was solidly brazed to the No. 52 was applied by conventional means, the
bottom wedge of the header. More careful control alloy would preferentially flow to the hottest
of the brazing alloy preplacement procedures member, and consequently there was poor flow on
should reduce runoff, but it is felt that additional the tube sheet. The method of brazing-alloy
preventive measures should be studied. preplacement shown in Fig. 3.3.3 and described
In summary, it appeared that the gas purification
and brazing equipment were adequate and that the UNCLASSIFIED

general brazing procedures were well conducted. ORNL-LR- DWG 27066

Previous experience had indicated that it might

be more advisable from the standpoint of fillet
reliability to have the units brazed with the tube
sheets in the horizontal position (tube axes
vertical). It is recognized, however, that this
could only be accomplished in this case by brazing
one end at a time and thereby introducing what . l/e in.
could prove to be insurmountable distortion 3 HOLES, 0 062
AT 120
problems. It was recommended, therefore, that
extra time and effort be devoted to the preplace
ment problem in an attempt to prevent runoff and PREPLACED BRAZING ALLOY
:—y8-in.0D, 0.065-in WALL TUBING
to ensure reliable fillets.

ART FILL-AND-DRAIN TANK FABRICATION Fig. 3.3.3. Trepanned Joint Detail.

R. L. Heestand

An investigation of optimum methods for fabri previously removed the problem by allowing the
cating the ART fill-and-drain tank tube-to-tube- brazing alloy to remain at the temperature of the
sheet joints was completed. A description of the tube sheet until the brazing temperature was
techniques used in the study was given pre readied. The trepanned sump on the braze side
viously. ' of the tube sheet then acted as an infinite res
The optimum joint design consisted of trepanning ervoir from which the alloy flowed to fill the
the weld side of the tube sheet, as shown in annulus between the tube and the tube sheet and
to form a fillet where the tube met the tube-sheet
Fig. 3.3.3. The trepanning serves to ensure good
heat distribution by providing a uniform weld surface. A typical joint of this type, cut so as
geometry and to minimize tube-sheet distortion and to show one of the three sump holes, is shown in
restraint in the vicinity of the weld joint. Test Fig. 3.3.4.
joints made initially by fusion welding showed
some root cracks, but welds made with Inco No. 62 In order to further prove the techniques described
above and to study the amount of distortion in the
tubes which might be encountered in a typical
E. A. Franco-Ferreira and G. M. Slaughter, ANP
Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 200. braze cycle, a 66-tube Inconel sample that simu
6E. A. Franco-Ferreira, ANP Quar. Prog. Rep. lated a tube-sheet section was prepared. Brazing-
June 30, 1957, ORNL-2340, p 225. alloy slurry (Coast Metals No. 52) was applied to

159
ANP PROJECT PROGRESS REPORT

The corresponding Operator's Qualification Test fabrication of radiators and heat exchangers, '
Specifications have also been developed and are and (3) the developmental work on the bonding of
being prepared for AEC publication. cermet valve components to metals.

Summary reports have been published on (1) the P. Patriarca et al., Fabrication of Heat Exchangers
development of high-temperature brazing alloys and and Radiators for High Temperature Reactor Ap
plications, ORNL-1955 (June 14, 1956).
their corrosion and oxidation resistance,8 (2) the
10P. Patriarca, G. M. Slaughter, and W. D. Manly,
Welding J. (N.Y.) 36, 1172 (1957).
o

E. E. Hoffman et al.. An Evaluation of the Corrosion P. Patriarca and G. M. Slaughter, "Bonding of

and Oxidation Resistance of High-Temperature Brazing Cermet Valve Components to Metals," Welding J. (N.Y.)
Alloys, ORNL-1934,(Oct. 23, 1956). (in press).

168
ANP PROJECT PROGRESS REPORT

1500°F. None of the cermet-metal joints were reacted with the room atmosphere after dis
attacked, with the possible exception of the joint assembly. The broken pieces (cylinders 12 through
which consisted of K94 brazed to nickel with 23, as shown in Fig. 3.4.5) were subjected to
Coast Metals alloy No. 50. Attack on this speci 1171 thermal cycles from a maximum of 1630 ±20°F
men was doubtful, and it will be tested again. to a minimum of 880°F (+30°F or -20°F) during
In these joints, nickel or Mai lory 1000 was used the first 1800 hr of the test. This excessive tem
to buffer the stresses developed by differences in perature change was inadvertently caused by
thermal expansion between the cermet and the faulty temperature measuring and recording equip
structural metal, Inconel. The high-temperature- ment. The actual temperatures given above were
fusion joints between cermets and nickel were found by calibrating the faulty equipment with
made without brazing material, while the nickel properly operating equipment. During the final
was joined to Inconel by the conventional brazing 1200 hr of the 3000-hr test, cylinders 12 through
techniques used in valve fabrication. 23 were subjected to 609 more cycles from a
Specimens of joints 1, 2, and 4 were also tested maximum of 1140 ±30°F to a minimum of 790 ±80°F;
in sodium for 100 hr at 1500°F in a seesaw appa that is, they were subjected to the simulated ART
ratus. Results of metallographic examination and conditions. Cylinders 1 through 11 were never
comparisons with untested specimens showed that subjected to a temperature change of more than
these joints were not attacked. Joints 3, 5, and 6 550° F.
were not tested because it was known that the All the thermal cycling was accomplished by
brazing materials used were not corrosion resistant raising and lowering the control rod from the fully
to sodium under these conditions. in and fully out positions. Approximately 20 min
was required for temperatures to reach equilibrium
Postoperative Examination of Prototype ART after a change of control-rod position. It is be
Control Rod lieved that cylinders 12 through 23 broke because
W. H. Cook of the excessive temperature cycling during the
first 1800 hr of the test when the temperature
The ART prototype control rod used in a 3000-hr
changes required to reach equilibrium were about
test of the control rod drive system was dis
500 to 800° F.
assembled and examined. (A discussion of the
The results of this test indicate that these rare-
test may be "found in Chap 1.3 of this report.)
earth cylinders can withstand thermal cycles,
The rod consisted of a column of porous cylinders
under the conditions described, of 550 to 650°F
of rare-earth oxides in stagnant, molten sodium in
from 1650 + 30°F. The most serious fault of these
an Inconel thimble. The test conditions simulated
cylinders appears to be spoiling of the type ex
those expected during ART operation, except that
hibited by cylinders 2, 6, and 9. At the base of
heat was applied externally. In the ART the
each of the shallow holes caused by spoiling,
control rod would be the heat source.
there was a small amount of particulate material.
The 23 cylinders of rare-earth oxides (45-49.5%
X-ray examinations of the particles showed the
Sm203, 22.5-27% Gd203, balance other rare-earth standard pattern of the rare-earth oxides and an
oxides; apparent porosity, 46.3%) were, nominally,
additional, as yet unidentified, pattern.
l'/4 in. OD, \ in. ID, and 1 in. long. They were The sodium that cooled the control rod during
contained in a thin-walled Inconel thimble, which
the test was examined spectrographically, and
was raised and lowered inside a second thimble
the quantities of rare earths found ranged from
of molten sodium. Holes in the inner thimble
<0.7 to <0.3 ppm. The samples were taken by
allowed the sodium to contact the ceramic cylin-
draining the sodium before disassembly of the rod.
ders.
Since the analyses indicate that the rare-earth
The ceramic cylinders may be seen in Fig. 3.4.4
content of the sodium was negligible, it may be
as they appeared before the test and in Fig. 3.4.5
assumed that the breaking and spoiling of the
as they appeared after the test. The white material
cylinders occurred without detectable fragmenta
on the as-tested cylinders is residual sodium that
tion or powdering. The thin-walled Inconel thimble
held the broken pieces in place, and thus the con
2S. C. Shuford, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 41. trol rod was operable despite the breakage and,

172
 PERIOD ENDING DECEMBER 31, 7957

Lithium in Niobium Forced-Circulation Loops

ORNL-LR-DWG 27067

J. H. DeVan R. S. Crouse
A forced-circulation loop is being designed for
studying lithium-niobium systems under the fol
LITHIUM SURGE TANK
lowing conditions:
COLD TRAP

Lithium flow rate 1—2 gpm

INSULATION
Reynolds number 15,000
TYPE 310 STAINLESS
Maximum fluid temperoture 1500-1800°F STEEL CLAD NIOBIUM
TUBING, V2-in.ID,
0.040-in. WALL
Temperature differential 200-400°F

The niobium tubing of the loop is to be protected

from air oxidation by a continuous outer sheath or
cladding of type 310 stainless steel. Either in
SODIUM SURGE
direct resistance heating with Nichrome-wound
clamshell elements or induction heating may be TYPE 310 STAINLESS
used with the clad tubing. A preliminary loop de STEEL TUBING, IV2-in
OD, 0.065-in. WALL
sign that includes clamshell heaters is shown in
Fig. 3.4.6. An annulus filled with sodium is SODIUM

provided around both the hot and cold legs in

COOLING COILS
order to obtain a much higher power density per
unit length with the available clamshell heaters
than could be obtained with heating units placed
around the relatively small-diameter niobium
tubing. Use of the sodium annulus also alleviates
CLAMSHELL HEATERS
the thermal stress problems associated with bonding
the niobium to the stainless steel cladding.
Although it is presently planned to utilize clad
tubing for all portions of the loop, the portions
external to the annular sections could be protected
with a dry box or other sealed compartment if
necessary. Procurement of the niobium tubing, HEATED LENGTH, 7 ft
both clad and unclad, has been initiated, and "•TYPE 310 STAINLESS
TOTAL MAIN CIRCUIT
LENGTH~19ft
delivery is expected within six to eight months. STEEL PIPE, 4-in.IPS,
SCHED 40
While awaiting shipment of this tubing, various
mechanical features of the loop design will be Fig. 3.4.6. Niobium-Lithium Forced-Circulation Loop.
studied by using an all-stainless-steel prototype.

DEVELOPMENTAL STUDIES OF
corrosion tests were conducted in order to ascer
INCONEL SYSTEMS
tain the specific corrosion rates of Inconel in
J. H. DeVan R. S. Crouse fused fluoride salt mixtures under the various

Inconel Exposed to Fused Fluorides conditions of reactor operation. The results of

such experiments both in- and out-of-pile have
From early studies of the compatibility of appeared in previous reports in this series, as
structural alloys with fused fluoride salt mixtures, well as in several topical reports 5-7 Th e s e
it was found that nickel-base systems were the
most corrosion resistant of the many types of G. M. Adamson et al., Interim Report on High Tem
alloys investigated. Consideration of the addi perature Corrosion With Fluoride Mixtures, ORNL-2337
(in publication).
tional requirements of a structural alloy in a high-
G. M. Adamson et al., Corrosion by Zirconium-Base
temperature reactor system indicated that Inconel Fluorides, ORNL-2338 (in publication).
would be the most suitable of the then available W. D. Manly et al., Aircraft Reactor Experiment —
materials. Following this selection extensive Metallurgical Aspects, ORNL-2349 (Dec. 20, 1957).

175
ANP PROJECT PROGRESS REPORT

studies^^efe concluded about a year ago, at which occurs in Inconel systems exposed to sodium and
time sufficient information had been obtained to to NaK.
effectively program the operation of the ETU and The mass transfer that occurred in a 1000-hr
the ART with respect to expected corrosion rates. period was found to be quite minor at temperatures
below 1300°F, but at higher temperatures noticeable
metallic deposits that were rich in nickel were
Inconel Exposed to Sodium and to NaK
formed in the cold-leg regions of forced-circulation
The utilization of Inconel as the primary con systems. The effects of temperature gradient and
struction material for the ART and the ARE re flow rate on mass transfer were also found to be
quired a determination of the corrosion properties significant at temperatures above 1300°F. The
of this alloy in sodium and in NaK over the tem rate of mass transfer showed almost linear corre
perature range from 1000 to 1500°F. Such cor spondence with these variables over the ranges
rosion data, which were obtained in both thermal- studied. The level of oxide impurities in sodium
convection and forced-circulation loop experiments, and in NaK, evaluated in Inconel systems, was
have been reported in detail in previous reports also of some importance, but the effect on the rate
in this series. The primary result of the experi of corrosion was not so significant as it has been
ments was an illustration of the very strong tem found to be at lower temperatures in iron-base
perature dependence of the mass transfer that systems.

176
 PERIOD ENDING DECEMBER 31, 7957

««W*J»
3.5. MATERIALS FABRICATION RESEARCH

NIOBIUM FABRICATION stainless steel, but, because of their irregular

shapes, the electromagnetic pump, the surge tank,
H. Inouye D. 0. Hobson
and some bends would be difficult to clad by using
J. H. Coobs
ordinary cladding methods. Therefore other means
Recrystallization of Niobium are being sought for protectively cladding irregular
surfaces and shapes fabricated of niobium. In the
J. P. Page
initial studies, several of the high-temperature,
Specimens of niobium which had been subjected nickel-base brazing alloys are being applied to
to various recrystallization treatments were ana niobium specimens, which are then subjected to
lyzed by x-ray and metallographic techniques, and oxidation tests. The braze metal is applied by
the times and temperatures necessary for the flame spraying, and it is then sintered in helium at
initiation and completion of recrystallization were a temperature just above its melting point. Oxi
determined. In addition a study was made of the dation tests are presently being made in a hori
effects of recovery and recrystallization on the zontal tube furnace at 1700°F in static air, but,
room-temperature properties of niobium. Except since the specimens must be removed from the
for the larger grain size of the recrystallized furnace and cooled before they are weighed, long
arc-cast material, no significant differences were time oxidation tests of up to 1000 hr will be run
found between arc-cast and powder-metallurgy in vertical-tube oxidation furnaces that do not
niobium. A topical report describing the complete require thermal cycling of the specimens during
investigation of the recrystallization of niobium weighing.
has been published. The results of some of the initial oxidation tests
of niobium specimens coated with Wall Colmonoy
Fabrication of Niobium Loops
Nicrobraz or Coast Metals alloys No. 52 or 53 are
A preliminary design has been prepared, as compared with the results of similar tests of type
described in the previous chapter of this report 310 stainless steel in Fig. 3.5.1. The sharp rise
(Chap. 3.4), for a forced-circulation loop that is in the curve for the sample coated with Nicrobraz
to be fabricated of niobium tubing and is to be is the result of the failure of the cladding at two
operated with lithium as the circulated fluid. The corners of the specimen. The niobium, upon ex
principal problems that will be associated with the posure to air, oxidized at a very rapid rate, with the
construction and operation of such loops will be results shown in Fig. 3.5.2. The cladding was
material procurement, oxidation protection of the unable to prevent the oxide from spreading and
niobium, and welding. The development work was rolled back upon itself. A second test of a
under way on the welding of niobium was described Nicrobraz-coated sample, in which the coating did
in Chap. 3.3 of this report. Purchase orders have not fail, showed the Nicrobraz to have oxidation
been initiated for the procurement of 0.500-in.- resistance comparable with that of the Coast
OD x 0.035-in.-wall tubing, l^-in.-OD x 0.035-in.- Metals alloys. Tests will also be made of niobium
wall tubing, and plate, but the finished material specimens that have been copper-plated and then
will probably not be received for several months. aluminized. Several attempts have been made to
Alternative plans for meeting the material require fabricate such specimens, but each attempt has
ments include the purchase of forged ingots to be failed because impure sintering atmospheres pre
processed into tubing and of arc-cast billets to be vented bonding between the copper and thealuminum.
extruded to tube blanks. Since the proposed loops
will be exposed along part of their length to air, it
Consolidation of Niobium Pellets
will be necessary to provide some sort of cladding
that will protect the niobium tubing from exposure. An experimental quantity of niobium pellets
It will be possible to clad much of the tubing with produced by the Electro Metallurgical Company
was consolidated into a rod by extrusion. After
being stripped from the stainless steel can in
'j. P. Page, The Annealing Behavior of Cold-Rolled which it was extruded, the rod was cold-swaged to
Niobium, ORNL-2372 (Sept. 19, 1957). a reduction of 97%, without annealing. On the

177
ANP PROJECT PROGRESS REPORT

assembly are described in a separate report. In longitudinal holes as required in thermocouple

addition, a report is being prepared on the de assemblies. The magnesium oxide rods were
velopmental work that led to the selection of the approximately 0.370 or 0.490 in. in diameter, and
particular materials used for the neutron shield the beryllium oxide rods were approximately 0.242
of the ART. in. in diameter. Both types of rods were fired at
1400°C for 4 hr. The beryllium oxide rods will be
TUBULAR CONTROL ROD PRODUCTION swaged in the tube assembly.
J. H. Coobs

The feasibility study of the extrusion of control Production of Metal Hydrides

rods was continued. Tensile bars cut from two of
R. A. Potter
the successful extrusions described previously
were tested at room temperature and at 1500°F, and Preparations are being made for hydriding dense
elongations of 2 and 5% were obtained, as com yttrium metal when it becomes available. Reactor-
pared with elongations of less than 0.1% for the grade zirconium, which is available and which
specimens tested previously. responds to hydrogen in much the same manner as
yttrium does, is being used in developing the
hydriding processes.
CERAMIC INVESTIGATIONS

L. M. Doney In the method being tested currently, an air-

evacuated, sealed nickel capsule is used through
which the hydrogen must diffuse before coming
Development of High-Density BeO Bodies into contact with the metal specimen. The sealed
R. L. Hamner nickel capsule eliminates the necessity of main
taining a rather elaborate gas-cleaning train and
Developmental work was initiated at the request
also provides a simple method whereby the hy-
of GE-ANP on high-density BeO bodies for use as
drided metal can be kept from contact with the
moderating material. Very fine particles of BeO
atmosphere until it is to be clad.
are being obtained by the controlled calcination
of BeSO.. Heating at 950°C for 1 hr in an oxidizing In a typical experiment, a machined, L- by 3-in.,
atmosphere produced a mass of fluffy, white ag zirconium metal rod is placed inside a length of
glomerates, which were broken down in a ball mill ^-in.-dia, 20-mil-wall nickel tubing with molyb
prior to pressing. The surface area of this material denum spacers at both ends to prevent contact
was 20 to 26 m /g. between the nickel and the sample. End caps are
Cylindrical specimens were hot-pressed at 200 psi welded in place on the tubing, and the assembly is
and temperatures of 1325, 1425/ and 1525°C. The evacuated with a vacuum pump connected to a
densities obtained were 91, 92, and 96% of theo /^-in.-dia nickel tube welded into one end cap.
retical, respectively. The high sinterability of The small tube is then forged closed, and the
this powder is indicated by the fact that, in order sealed assembly is heated in a hydrogen atmosphere
to obtain comparable densities with commercial at temperatures ranging from 1475 to 1850°F and
grades of BeO powder, it is necessary to use at pressures from 1 to 2 atm.
temperatures 400°C higher.
Four such encapsulated samples have been
Extrusion of Magnesia and Beryl Iia Insulators treated under different pressure, temperature, and
time conditions. They were fired in a vacuum to
C. E. Curtis the peak temperature, hydrogen was allowed to
Rods of magnesium oxide and of beryllium oxide flow through the retort as the temperature was
were extruded to provide rod insulators with four dropped to a lower level, and a static pressure
was maintained throughout a soak period. In some
4J. P. Page and J. H. Coobs, The Shield Plug instances, the samples were refired in an attempt
Assembly for the ART Fuel Pumps, ORNL-2353 (in
press). to increase the hydrogen content. The firing
5M. R. D'Amore, ANP Quar. Prog. Rep. June 30, conditions and the results for the four samples
1957, ORNL-2340, p 242. are presented in Table 3.5.2.

180
 PERIOD ENDING DECEMBER 31, 7957

Table 3.5.2. Conditions and Results of Hydriding Experiments on Four Zirconium Samples

Maximum
_ ... . - , Total Time at ,_. , Hydrogen
Temperature Minimum Soak nyarogen
SamDle
sample Number ul
"""""" of „
During Vacuum —
Temperature Minimum Soak _Pressure r- . mt. ofi cSample
Content Sc I
Number Firing Cycles H . ,Op, Temperature (psig) (wt %)
m9 (hr)
1700 1650 3 0.2

1850 1605 3 0.2

1830 1605 15 0.7

1830 1475 15 1.1

181
ANP PROJECT PROGRESS REPORT

3.6. METALLOGRAPHIC EXAMINATIONS OF ENGINEERING TEST

COMPONENTS AFTER SERVICE

ART PROTOTYPE TEST RADIATOR NO. 1 Sixteen samples were removed from the radiator
at the positions shown schematically in Fig.
R. J. Gray J. E. Van Cleve, Jr.
3.6.1 (see Fig. 3.6.10, this report, for photograph
Test radiator No. 1 failed immediately after of a radiator with areas numbered as indicated
reaching full power in its first controlled thermal in Fig. 3.6.1). The samples were sectioned to
cycle. The NaK outlet temperature for the radiator expose two opposing joint areas on each tube
remained approximately constant during the tran and then were mounted in plastic material. The
sition, and the inlet temperature was increased samples were then ground and polished, and
from 1350 to 1500°F at a rate of 50°F/min. The the joint areas were examined at a magnification
radiator operated for a total time of 870 hr with of 50. In Table 3.6.1, the results of the analyses
the NaK inlet temperature above 1000°F. The of fin-to-tube joint braze integrity and the degree
total number of thermal cycles before the radiator of oxidation of the fin collar are presented and
failed was 9k. The examination of the point of compared with results for other radiators pre
failure was described in the previous report, viously examined.
and the investigation of fin-to-tube joint braze Incomplete melting of the braze-alloy rings was
integrity, mass transfer, corrosion, and other noted in several instances; a typical sample is
items of interest are described here. shown in Fig. 3.6.2. A partially melted ring
which did not flow over the fin collar to protect
'r. J. Gray and J. E. Van Cleve, Jr., ANP Quar. the copper core is shown in Fig. 3.6.3. A crack
Rep. Sept. 30, 1957, ORNL-2387, p 232. in the stainless steel cladding on the fin may

UNCLASSIFIED
ORNL-LR-DWG 27068
FLOW

0-1 0-2 0-3 0-4 0-5 0-6 0-7 0-8 0-9 1-1 1-2 1-3 1-4 1-5 1-6 1-7 1-8 1-9

SAMPLE LOCATION E^

Fig. 3.6.1. Diagram of Radiator Showing Locations at Which Samples Were Taken for Evaluation of Fin-to-Tube
Joint Braze Integrity.

182
ANP PROJECT PROGRESS REPORT

3.7. NONDESTRUCTIVE TESTING

R. B. Oliver

EDDY-CURRENT THICKNESS MEASUREMENTS REMOTE X-RAY VIEWING

R. B. Oliver R. B. Oliver J. W. Allen

R. A. Nance J. W. Allen
A closed-circuit television system (RCA model
The inherent accuracy of the new eddy-current ITV-6) for direct x-ray imaging has been installed
thickness gage was increased by the substitution in the Nondestructive Test Laboratory, and the
of drift-free transistor oscillators for the variable- initial adjustment and alignment have been accom
frequency signal generator previously employed. plished. Preliminary studies have been made on
Although transistors are sensitive to thermal the system to evaluate it for practical inspection
changes, they have a high degree of frequency use.

stability if a uniform temperature is maintained. The system consists of a conventional industrial-

The study of eddy-current cladding-thickness de type Vidicon camera and television monitor con
terminations has been continued, with particular nected in a closed video circuit. The conven
emphasis on alleviating the problems encountered tional light-sensitive Vidicon camera tube was
with the probe coils currently in use. These replaced with an x-ray-sensitive tube having a
coils are contained in commercially procured, selenium target (photoconductive receptor) ap
high-permeability, low-loss cup cores, approxi proximately ten times thicker. The thicker target
mately /2 in. in diameter. The electromagnetic allows greater absorption of x rays and thus
field produced by the coil investigates an area greater x-ray sensitivity.
/2 in. in diameter, which is much larger than The technique for use of this system is similar
the area of minimum cladding thickness sought. to that used in conventional radiography. The
Since the eddy-current reading is a function object to be examined is placed near the photo-
of the area and of the minimum cladding thick conductive face of the Vidicon tube, and the
ness, it is difficult to obtain a true evaluation source of x radiation is placed at some distance
of the cladding thickness with this coil. Some from the object. The system is, therefore, subject
cladding-thickness measurements have been satis to the same geometrical considerations as those
factory, but the error in a number of cases has of the conventional radiographic method. An
been excessive. Smaller probe coils are being inherent image magnification of approximately 12
fabricated to remedy this problem. The accuracy diameters is obtained, since the \- by 3/-in.
of the determination will increase as the area roster scanned on the Vidicon target is reproduced
of the field of the coil approaches the area of as a 6- by 9-in. roster on the kinescope.
minimum cladding thickness. It is also difficult Preliminary studies of the sensitivity and
to locate the exact area of minimum cladding for resolution of the system have been encouraging.
metallographic examination because of the asym The exposure and contrast sensitivity, although
metrical nature of the electromagnetic field pro not so great as those attained with film, are
duced by the present probe coils. Therefore the adequate for many applications. Since the ex
small cup cores are being constructed to produce posure of the photoconductive target is continuous
symmetrical fields in addition to decreasing the rather than accumulative, as it is with a photo-
size of the field. emulsion, it is inherently less sensitive than
A study has been initiated to determine the film. The greatest area of usefulness of the
feasibility of measuring the thickness of hot- system appears to be in the inspection of thin
dipped aluminum coatings on aluminized steel. sections through which high-intensity radiation can
Preliminary tests indicate that a nominal coating reach the photoconductive target.
thickness of 0.001 in. can be measured to within A very high degree of resolution is attainable
0.0002 in. with this system in comparison with other types
of direct or remote x-ray image-viewing systems;
'R. B. Oliver, J. W. Allen, and R. A. Nance, ANP the resolution is not limited by a mosaic of
Quar. Prog. Rep. Sept. 30, 1957, ORNL-2387, p 247. fluorescing crystals, as is the case with a

192
 Part 4

RADIATION DAMAGE

G. W. Keilholtz
 PERIOD ENDING DECEMBER 31, 7957

examinations of both specimens are being carried longer than in the MTR tests, but the times were
out in an attempt to determine the cause of the still of the order of a factor of 4 shorter than in
structural variations observed in the zirconium tests conducted by the Metallurgy Division.
hydride matrix.
As may be seen from the data in Table 4.1.1, the
CREEP AND STRESS-RUPTURE strain at fracture in the out-of-pile specimens was
TESTS OF INCONEL greater than that in the corresponding in-pile
specimens. The in-pile apparatus was disassembled
J. C. Wilson
and the specimens were checked for fracture
W. E. Brundage N. E. Hinkle location. Metallographic examination of both out-
W. W. Davis J. C. Zukas of-pile and in-pile specimens will be carried out in
order to establish any differences in the mode of
MTR Experiments fracture.
The out-of-pile test was completed that corre
sponded to the MTR tube-burst tests conducted The thermocouple wires used in the out-of-pile
previously. The apparatus for the out-of-pile test were brittle after the test. In order to de
test as well as the operating conditions duplicated termine whether a change in thermocouple charac
those of the in-pile test except for the absence of teristics accompanied the loss of ductility, tests
radiation. Seven of the eight out-of-pile speci on the constancy of thermocouple calibration were
mens ruptured before the test was discontinued at carried out in helium in an Inconel tube. The
the end of a 304-hr period. The times to rupture, thermocouples were located in air and in helium.
as shown in Table 4.1.1, were two to five times
3J. R. Weir, Jr., D. A. Douglas, and W. D. Manly,
4J. C. Wilson et al., ANP Quar. Prog. Rep. June 30, Inconel as a Structural Material for a High-Temperature
1957, ORNL-2340, p 268. Fused-Salt Reactor, ORNL-2264 (June 4, 1957).

Table 4.1.1. Rupture Life and Elongation of Inconel as Measured in In-Pile and Out-of-Pile Tube-Burst Tests

Time to
Tube Wall Thickness* Specimen Stress Type of Elongation
Rupture
(in.) Number (psi) Test (%)
(hr)

0.030 21 3000 In-pile** 78 0.8

22 3000 In-pile 78 0.7
33 3000 Out-of-pile 144 4

29 3000 Out-of-pile 304 3.1

20 4000 In-pile 17 1.1

16 4000 In-pile 33 0.75

27 4000 Out-of-pile 78 4.4

31 4000 Out-of-pile 100 5

17 5000 In-pile 13 1.2

24 5000 In-pile 19 1

30 5000 Out-of-pile 55 5

25 5000 Out-of-pile 103 5.2

0.050 1 4000 In-pile 23 0.85

3 4000 In-pile 29 0.96
36 4000 Out-of-pile 105 7.6
34 4000 Out-of-pile 115 7.6

♦Original inside diameter = 0.51000 ± 0.00025 in.

**Fast flux = 2 X 1013 neutrons/cm2-sec (above 1 Mev).

207
ANP PROJECT PROGRESS REPORT

Also a test is being made of some of the fused- that of the test section, was temperature-cycled
quartz insulation used in the in-pile and out-of- over 4000 times in 1000 hr without failure, but the
pile tests. Preliminary information indicates that test section finally ruptured. The results of this
errors of the order of 100°F (below true temperature) test indicate that cyclic stresses are not the cause
in thermocouple readings may occur after 250 hr of the cracks that have been found in brazed joints
for Chromel-Alumel thermocouples. Iron-constantan in the in-pile rigs.
couples are affected to a lesser extent in the Three tube-burst specimens tested previously
opposite direction. In a test without the fused- in hole HB-3 of the LITR at a constant stress of
quartz insulation, no disparity between thermo 2000 psi at 1500°F, with both sides in helium,
couples located in air and in helium was found were examined. One ruptured after 1050 hr, and
after 500 hr at temperatures of 1500°F and above, showed an increase of 1.1% in inside diameter. A
and a single tube-burst specimen (at a stress of second ruptured at 1650 hr, and the inside diameter
4000 psi) in the helium side of the apparatus increased about 0.9%. The third specimen had not
ruptured at a period corresponding with the values failed when the test was terminated at 1675 hr,
obtained in the Metallurgy Division. This suggests but it showed an increase of about 0.9% in inside
that the fused-quartz insulation supplies a re diameter. A fourth specimen, stressed at 1000 psi
ducing atmosphere that affects the calibration of at 1500°F for 1675 hr in the reactor, also did not
the Chromel-Alumel couples so that the specimens rupture, but it showed an increase of 0.56% in
may be at a higher temperature than the thermo inside diameter. The test sections are now being
couples indicate. examined metallographically. The low ductility,
Thus it appears that part of the reason for the as in the case of the MTR tests, cannot be accounted
reduced times to rupture observed in the in-pile for at this time.
and out-of-pile tests is the loss of calibration of
ORR Experiments
the thermocouples. Unless radiation changed the
rate of deterioration of the thermocouples, it is The design of a facility for mounting experi
still not clear why such large ductility changes ments on the pool face of the Oak Ridge Research
were observed. Reactor was completed. The facility is to be in
stalled before the reactor goes critical. Experi
ment cans and locking devices for the cans were
LITR Experiments designed. Apparatus for conducting flux and
In the most recent LITR test, an Inconel tube gamma-heating surveys was also designed.
stressed at 2000 psi and exposed to the fused
salt fuel NaF-ZrF4-UF4 (62.5-12.5-25 mole %) has LITR VERTICAL FUSED-SALT LOOP
been in hole HB-3 for over 620 hr without rupture, W. E. Browning R. P. Shields
in contrast to tubes tested previously which failed J. E. Lee
after 100 hr or less. Postirradiation examination
of the tubes previously tested revealed a leak in Disassembly of the vertical in-pile loop which
one of the Swagelok fittings on the pressure side circulated fused-salt fuel for 235 hr in the LITR

of one of the specimens and a defective high- was completed, and various parts are being ex
pressure seal on the thermocouples of another amined and analyzed. One-inch lengths of fuel
specimen. In the apparatus being used at present tube were cut from each thermocouple location,
there are no Swagelok fittings, and all seals were and other fuel and metal samples were taken from
checked at double the operating pressure. A pure the pump region. Particular attention was given
cobalt foil was inserted inside the specimen tube to the pump bearing. As stated in the previous
to measure directly the flux depression. report, the lower bearing was locked and the
A tube-burst specimen (0.010 in. wall thickness) bearing just below the motor was difficult to turn.
pressurized to a stress of 3000 psi with helium was These bearings were separated from the housing
subjected to thermal cycling between 1100 and and shaft, and the grease seals were removed.
1500°F in air, out-of-pile. A joint brazed with
Coast Metals alloy No. 52, near the specimen but SW. E. Browning et al.. ANP Quar. Prog. Rep. Sept. 30,
at a temperature approximately 100°F lower than 1957, ORNL-2387, p 270.

208
 PERIOD ENDING DECEMBER 31, 1957

The grease has decomposed and become waxlike, and it can be evacuated and filled with helium to
and it had been pushed to the side away from the prevent contamination of the salt by the atmosphere.
ball bearings. Samples of this damaged lubricant When the Inconel tube walls are heated quickly, the
were taken to be analyzed for fission-gas decay salt slides out as a solid pellet. Tests with un
products and for analysis by the California Re irradiated samples have shown this system to be
search Corporation. An attempt is being made to satisfactory. Salt samples from irradiated tubing
degrease and decontaminate the bearings so they will be analyzed chemically, and the tube samples
can be cut apart and examined. will be examined metallographically in order to
An induction furnace (Fig. 4.1.21) has been set evaluate the corrosion of the loop that occurred
up for melting salt samples from the Inconel fuel- during the test.
tube specimens. The melt-out chamber is glass,
MTR STATIC CORROSION TESTS

W. E. Browning H. L. Hemphill
UNCLASSIFIED
ORNL-LR-DWG 253S7
Three capsules containing fused-salt fuels
which had been irradiated in the MTR were dis
mantled and examined metallographically. The
results of the examinations are presented in
Table 4.1.2. The Inconel capsule showed the
INSTRUMENT
expected corrosion attack. The two Hastelloy B
capsules were found to have thin films on the
•VACUUM
surfaces that were exposed to the fuel. Tests are
under way to identify the film and to study the
mechanism of its deposition. Corrosion of these
ALL PYREX MELT-OUT ASSEMBLY capsules was negligible.
INSTALLED BEHIND SHIELDING
Two more Hastelloy B capsules containing
TAPERED QUARTZ SLEEVE NaF-KF-LiF-UF4 fuel were withdrawn from the
TUBING SAMPLE MTR. One of these had been irradiated for 2496 hr

INDUCTION HEATER
at 1250°F and at an initial power density of
6000 w/cm . There was no indication of failure
when the capsule was withdrawn.
PLATINUM CRUCIBLE
IRRADIATION OF MODERATOR MATERIALS
FOR USE AT HIGH TEMPERATURES

W. E. Browning R. P. Shields
Fig. 4.1.21. Induction Furnace for Removing Fluoride Apparatus for testing beryllium, beryllium oxide,
Fuel from Inconel Tubing. and yttrium hydride at high gamma-heating densities

Table 4.1.2. Results of MTR Static Corrosion Tests at 1500°F and at a Power Density of 6000 w/cm3

Duration of
Capsule Container
Fuel Exposure Metallographic Results
Number Material
(hr)

331 NaF-ZrF4-UF4 304 1 to 2 mils of penetration

(53.5-40-6.5 mole %)

364 Hastelloy B LiF-NaF-KF-UF4 138 Film deposited; penetration negligible

(44.3-11.8-38.2-5.77 mole %)

335 Hastelloy B LiF-NaF-KF-UF4 119 Film deposited; penetration negligible

(44.3-11.8-38.2-5.77 mole %)

209
ANP PROJECT PROGRESS REPORT

in the ETR is being prepared. Those moderator used in making resistivity measurements, were
materials are to be irradiated with neutrons and located at the center of the sensitive area. The
gamma rays in the ETR where the heating rate configuration is shown in Fig. 4.1.22. The thick
will be 25 w/g. A temperature of about 1450°F nesses of samples 38 and 40 were 0.86 and 0.85 /x,
will be monitored by a thermocouple on each side respectively, and samples 5, 22, and 26 were
of the capsule and by one inside a well in one assumed to be 0.5 p. thick.
piece of each type of specimen.
All specimens for the test, except the BeO, UNCLASSIFIED
ORNL-LR-OWG 25452
have been fabricated and then encapsulated in A, B RESISTIVITY PROBES
Inconel. Capsules for the BeO were completed and C, D HALL PROBES
the BeO specimens are complete except for final
checks. Extra beryllium specimens and capsules GOLD

were made and encapsulated. A thin molybdenum

i.25 cm 5B1 PbS
foil was wrapped around each pellet, and a disk
was cut to cover the end of each pellet. One set I I GLASS
of the completed beryllium specimens will be
heat cycled by the Metallurgy Division, which ^ cm —

fabricated all the specimens. The heat-cycled

specimens will be examined to determine the com
patibility of the metals at high temperature. Fig. 4.1.22. Configuration of Lead Sulfide Sample
Cel
A reactor-core filler block has been procured
from the factory which manufactured the parts for
the ETR, and it has been modified for this experi
Measurements were made at a temperature of
ment. Details of the gas system for conducting
24.5 ± 0.5°C. All voltages were amplified by an
heat from the specimens to the capsule walls are
Applied Physics Corp. vibrating-reed electrometer,
being worked out with ETR personnel.
but the actual readings were taken on a Rubicon
IRRADIATION EFFECTS IN Co. potentiometer. Resistances were computed
PHOTOCONDUCTIVE DETECTORS from the relation between the direct current and the
voltage. All resistances measured were of the
J. C. Pigg R. A. Zingaro
order of 10 ohms. During all electrical measure
An initial investigation has been completed of ments the cell was sealed from light. Thus all
the effects of neutron and gamma radiation on reported values represent dark-current measure
chemically deposited thin films of lead sulfide, ments.
which are highly sensitive to infrared radiation The magnetic field for Hall-voltage measure
and have a wide variety of commercial and military ments was furnished by a 9000-gauss permanent
applications. The polycrystalline structure of the magnet; therefore it was not possible to reverse
films and the relatively high and localized con the magnetic field. Instead, the sample was rotated
centration of impurities make the behavior of these through 180 deg in the field, and readings were
films very complex; therefore, the results of this taken with the current flowing in both the forward
study are of practical interest. Irradiation has and reverse directions during each orientation.
been found to cause additional crystalline imper The Hall voltages were calculated from the dif
fections, and the effects are reflected in changes ference of the averages of a number of readings
in the electrical and photoconducting properties made with the cell in each orientation. The Hall
of this semiconductor. The results of the initial measurements could not be made with a precision
experiments are described here. of better than ±5%. The maximum error in all
The samples used were chemically deposited other electrical measurements was about 0.5%.
PbS cells, supplied by the Eastman Kodak Com The other significant sources of error in these
pany, and were of the type previously described experiments were the fluctuations in room temper
by Woods.7 One set of contacts, which were pre ature (±0.5°C) and in humidity.
pared by evaporation techniques and which were Changes in the photosensitivity of the cells were
7J. F. Woods, Phys. Rev. 106, 235 (1957). followed by measuring the signal output when a

210
 PERIOD ENDING DECEMBER 31, 1957

cell waso
subjected
'
to radiation in the 5000- to lower than those reported by Woods, who gave
25,000-A region. A Cary model 14R spectropho 25.9 to 38.3 ohm-cm as the resistivity of his
tometer supplied the monochromator, light chopping, sample. However, resistivity values calculated
and amplification systems. The cell was mounted from the data of Lummis and Petritz varied from
at the monochromator exit, and the signal from the 15 to 36 ohm-cm. The agreement of these values
cell was fed directly into the preamplifier input. within a rather wide range indicates the variance
This circuitry proved to be very satisfactory, since which is to be expected for thin films of this type.
the preamplifier system is designed for use with an A plot of the resistivities of PbS cells 26 and
identical type of PbS detector, which actually 40 as functions of neutron irradiation is presented
forms part of the instrument. Absolute wavelength- in Fig. 4.1.23, A. An initial drop in resistivity
response curves could not be obtained with such may be seen for both cells that is followed by a
an arrangement, since the slit-width drive could gradual rise as the neutron dose is increased.
not operate to compensate for the variable energy The change in resistivity of cell 40 with time,
output of the light source. following neutron irradiation is shown in Fig.
Neutron irradiations were carried out in the 4.1.23, B. There was a rapid decrease in re
ORNL Graphite Reactor. The samples were placed sistivity during the first 24 hr, and it is to be
in a boron carbide container so that slow-neutron noted that the resistivity continued to fall with
bombardment would be essentially eliminated. The time until, after the tenth day of annealing, it had
fast-neutron flux was taken to be 8 x 10 neu reached an apparent equilibrium value which was
trons/cm-sec. Gamma irradiations were carried below the resistivity of the cell prior to irradiation.
out in the ORNL "gamma garden," which utilizes The five measurements made during the tenth to
a cobalt source. The flux density was taken to be twenty-first days gave an average value for the
2.65 x 106 r/hr, which is equivalent to about 1012 resistivity of cell 40 of 13.7 ± 0.2 ohm-cm, which
photons/cm -sec. is 17% below the initial resistivity of this film,
A set of values observed for the four cells on 16.4 ohm-cm. The data for cell 40 are plotted on a
which electrical measurements were made is given larger scale in Fig. 4.1.24 for clarity.
in Table 4.1.3. The values of the Hall coefficient, The change in resistivity of the PbS cell after
R„, are in good agreement with those reported by neutron irradiation followed simple, first-order
Woods,7 Lawrence,8 and Lummis and Petritz. kinetic relations, as may be seen in Fig. 4.1.25.
By assuming a film thickness of 0.5 p., values of The half-life of the annealing process, determined
RH from +105 to +245 were calculated for the cells graphically, was 11 hr, which gave a value of
studied by Lummis and Petritz on the basis of 0.063 hr for the specific rate constant. Thus
their data. The resistivity values are, in general, the value of the resistivity following neutron
irradiation may be expressed as
hbid., p 238. Woods* values of RH run from +121 (1)
• 0.063 t
to + 133, and Lawrence reports values of + 130 and P = Po C
+ 160.
9F. L. Lummis and R. L. Petritz, Phys. Rev. 105, where p is the value of the resistivity / hr after
502 (1957). the value pQ is measured.
Table 4.1.3. Properties of Chemically Deposited Lead Sulfide Cells at Room Temperature in the Dark

Hall Coefficient, Resistivity, Carrier Concentration, Carrier Mobility,

Cell
fi„ (cm /coulomb) p (ohm-cm) n (cm-3) Pfj (cm /vsec)

x 10 16

22 +261 29.4 2.8 8.9

26 + 132 12.6 5.6 10.5

38 + 163 14.7 4.5 11.1

40 + 190 16.4 3.9 11.6

211
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 25453R
23

"" ^
22

/ <
21 — • CELL 26

• CELL 40
/ \

20
/ *
/
19 X
/ \
\
> r \
E
i
' | \ V
V
^S
\J s L
\
H V >
>
<
\
in
^> -!^
— ,i _^ _ ... .
n

A
13 / i
1 FLUX =8xl0,,/cm2.sec
12

11

10
5 10 15 20 2 t ; 5 3 10 12 14 16 48 49 20 24

9 Tl ME(h r) TIME (days)

0 10 20 30 40 50 60 48 96 144 192 240 288 336 384 432
NEUTRON IRRADIATION TIME(hr)
TIME (mini

Fig. 4.1.23. Effect on Resistivity of PbS Cells of Neutron Irradiation Followed by Annealing.

The percentage changes in the Hall coefficient, Table 4.1.4. Changes in Electrical Properties of
carrier concentration, and carrier mobility of cell Chemically Deposited PbS Cell 40 During
40 are given in Table 4.1.4. Since the Hall Neutron Irradiation
coefficient changed a maximum of less than 5%
during irradiation, which is within the limit of ex Total
P erce ntage Changes* in
perimental error, it is concluded that this property, Fast-Neutron
Hall Carrier Carrier
as well as the carrier concentration, remained un Dose
Coeff ici ent Concentration Mobil ity
changed by irradiation. The carrier mobility, which (neutrons/cm )
is directly dependent on the Hall coefficient and
inversely proportional to the resistivity, showed 2.4 x 1014 + 1 3 +5
an increase as the resistivity decreased initially,
and then decreased when the resistivity increased.
4.8 x 1014 + 5 5 + 4

The changes in these properties one to three 7.2 x 10M -1 0 -4

weeks after neutron irradiation, as shown in
Table 4.1.5, were significant, and all the values 1.4 x 1015 -1 0 -11

showed definite trends toward limiting values.

The measured values for these properties 18 and "Based on values measured prior to irradiation.

212
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
UNCLASSIFIED
ORNL-LR-DWG 25454R ORNL-LR-DWG 25455R
100 10

A. CELL 40 AFTER NEUTRON IRRADIATION

TO A TOTAL FLUX OF 3 x 10'" NEUTRONS /
. cm1" • sec
50
a.
^ B. CELL 38 AFTER GAMMA IRRADIATION TO
• \ ATOTAL FLUX OF 3 x106 r
4

7 •

20

10

< i\

i\

_l
<

r
/
z AN.
< I 0.5

/ a\. \ A

/ \a
a

0.5
0.2
s

/
0.2
/ 0.4
10 20 30 40 50

0.1
22
/
21
21.5

20 19 18 17 16 15 14 13
Fig. 4.1.25.
Neutron-Irradiated
TIME, /(hr)

Change in Resistivity with Time of

Cell 40 and Gamma Irradiated Cell
RESISTIVITY, p (ohm-cm)
38.

Fig. 4.1.24. Change in Resistivity of Chemically

Deposited PbS Cell 40 Following Neutron Irradiation. mediately following irradiation are not given be
cause the resistivity was recorded automatically
21 days after irradiation were, for the Hall coef during that period and the Hall voltage could not
ficient, 181 and 195 cmVcoulomb, respectively; be measured automatically by the method being
for the carrier concentration, 4.1 and 3.7 cm- ; used. It is apparent, however, that at some time
and for the carrier mobility, 13.4 and 14.0 during the first few days of the annealing period
cm2/v«sec. The Hall coefficient and the carrier there was a decrease in the Hall voltage and a
concentration reached about the same values after concurrent increase in carrier mobility followed by
21 days that they had before irradiation (see a gradual return of these properties to their initial
Table 4.1.3), but the carrier mobility showed a values. The annealing mechanism led to a perma
significant increase. Data for the period im- nent net decrease in resistivity, while only a

213
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Table 4.1.5. Changes in Electrical Properties of
ORNL-LR-DWG 25456R
Chemically Deposited PbS Cell 40 After Neutron NEUTRON IRRADIATION TIME (min)
Irradiation to a Total Dose of 10 15 20 25 30 35 40 45 50

2.9 X 1015 Neutrons/cm2

Days After Percentage Changes in

Neutron Hall Carrier Carrier

Irradiation C Deff icient Concentration Mobility

7 -21 + 20 + 7.8

11 -18 + 18 -0.9

12 -13 -13 + 34

15 + 9 -8 + 29

18 -5 + 5 + 15

21 + 2 -5 + 20

temporary change took place in the carrier con

centration and in the Hall coefficient.
The effect of irradiation on the photoconduc
tivity of the PbS film (cell 26) was also examined,
and the results are shown in Fig. 4.1.26. Neutron
irradiation brought about a significant decrease in
the photoconducting properties of the material,
and the decrease was proportional to the total
neutron dose. The curves of Fig. 4.1.26 should
not necessarily be interpreted as indicating that
the change in sensitivity was also a function of
the wavelength of the exciting radiation. Thus 6 8 10 12 14 16 16 22 24
curve a shows that the cell no longer responded to ANNEALING TIME (days)
radiation of 24,000 A after 30 min of irradiation,
but that it still responded to shorter wavelengths Fig. 4.1.26. Effect of Neutron Irradiation on Photo
(curves b, c, and d). This was primarily the conductivity of Chemically Deposited PbS Films (Cell
result of the light source emitting much more 26).

intense radiation at the shorter wavelengths. An

examination of the data (Table 4.1.6) indicates, The changes in resistivity, Hall coefficient,
however, that the neutron irradiation effects were carrier concentration, and carrier mobility produced
not uniform throughout the wavelength region ex by gamma irradiation are plotted as the various
amined and that they became more marked at longer curves of Fig. 4.1.27. The effects during and im
wavelengths. A more quantitative study of a sys mediately following irradiation were fairly clean
tem which would provide uniform light intensity cut, but it is apparent that the long-term annealing
at all wavelengths is necessary before the effect effects following long periods of irradiation were
at longer wavelengths can be considered to be very complex.
real. It is of interest to note that restoration of
the initial level of photosensitivity did not occur An increase in resistivity during gamma irradia
even after three weeks of annealing. Indeed, the tion (plot A-l) was clearly demonstrated. Cell 38
signal output after this time was only about one- showed a threefold increase in resistivity after
half to one-third the original value, and it appeared receiving a dose of 5 x 10 r and a fourfold in
to be stabilized at the lower level. crease after a dose of 10 x 10 r. After a dose of

214
 PERIOD ENDING DECEMBER 31, 7957

Table 4.1.6. Absolute Decrease in Signal Output of PbS Detector (Cell 26) at Various Wavelengths
as a Function of Neutron Irradiation

Wavelength of
Exciting <*0 " S"V
o
At n =5 At n = 10 At n = 15 At n = 30 At n = 60
Radiation (A)

24,000 0.25 0.48 0.54 0.73

20,000 0.25 0.44 0.52 0.70 1.02

16,000 0.26 0.43 0.52 0.72

10,000 0.20 0.35 0.45 0.66 0.95

7,000 0.19 0.35 0.46 0.67 0.94

*Sq is the absolute signal output prior to irradiation, and S is the absolute signal output after n min of neutron
irradiation.

21 x 10 r, cell 22 showed an increase in resis during the initial 2 hr of irradiation, a threefold

tivity from 29.4 to 352 ohm-cm. During the time increase during the second 2-hr period of irradia
immediately following irradiation (plot A-ll), the tion, and an eightfold increase when the accumu
resistivity showed a gradual decrease. The lower lated dose reached 21 x 10 r. Similarly, the data
curve, which demonstrates this change following a for cell 38 show a fourfold increase during the
2-hr irradiation, suggests that the resistivity initial 2-hr period of irradiation, but a slight de
reached a limiting value after about 24 hr; the crease during the second 2-hr period. These ir
actual values of the last four measurements were regular patterns are, of course, reflected in the
41.2, 40.4, 37.8, and 39.7 ohm-cm. The tendency changes in the carrier concentration and mobility,
of the cell to reach an apparent equilibrium state if which depend upon R„.
the total irradiation did not exceed a certain maxi The long-range annealing effects (plots A-lll,
mum was also indicated by the changes in the B-lll, and C-lll) were so irregular that they will be
photoconducting properties, as will be discussed mentioned only briefly. It is clear that no equilib
later. Nineteen days after a dose of 10 x 10 r, rium condition had been reached 19 days after
however, the resistivity was (plot A-ll I) still irradiation. There was a trend toward lower values
undergoing large fluctuations. for the resistivity and Hall coefficient, but it is
If the change in resistivity of sample 28 follow not possible to predict, on the basis of the data
ing a total gamma irradiation of 5 x 10 r is plotted obtained thus far, at what time an equilibrium state
as a function of time as in Fig. 4.1.25, application would be reached and what these values would be
of the first-order kinetic law gives the following at equilibrium. The carrier concentration and
relationship for the change in resistivity following mobility fluctuated so widely that no definite trend
a total gamma irradiation which does not exceed is indicated by the data.
the "tolerance" for the cell:
The previously mentioned ability of these cells
(2)
-0.041 / to recover from the effects produced by gamma
Po
irradiation is demonstrated in Fig. 4.1.28. As
The half-life of the annealing process was 17 hr. may be noted, the cells lost their sensitivity
Further examination of Fig. 4.1.27 shows that rapidly as they accumulated progressively greater
there were over-all increases in the Hall coefficient doses of radiation. Examination of curves a, b,
(plot B-I)and concurrent drops in carrier concentra and c for cell 22 shows that this cell was still
tion and mobility (plot C-l) during gamma irradia sensitive after receiving an accumulated irradiation
tion of cells 22 and 38, but that the changes were of 10.4 x 106 r, and after receiving 20.8 x 106 r
quite irregular. For instance, the data for cell 22 it still gave a reasonably good signal at 1.5 and
show a small decrease from 261 to 247 in R„ 1.0 p, although no signal was obtained at 1.8 p.

215
ANP PROJECT PROGRESS REPORT

•1
I 1 1 1 1
DASHED LINES = CARRIER MOBILITY
10

N - <

V< \ i-
A-\
8 «I0
\ / >>

\ o
\ /
' < r
s
A \

/P
4 , o.. 1
' \
\ e \
- 6 /
1 > V \ \ %

5
\
\
\
i\.
\
i 7
o
*; /
/
\ j
\j
/
i
/
/
\
>
1 \ \ o >
^^ d \
/
4
^G —i A \
3 .
\

Lt 1
/4\ At
/r
\

\
•'•n

f\
—<u \

•^ ^ i 1
<%> \
\
22
C-1I
\
C-I c-ni
i
\

(650 i 1
1, Y'Vrf
700
r

600
38/
i l o^
"T
i i
500

1 71 \>^ X \
f y/
400

^.
300 • •-

200 u\
>
B-I B-II B-III
(00

•"— IRRADIATION —•"[

350 1 1 1 Jr
(00
i {s
90

80 ANN EALI

22 / A-I
70

1J i

A-H A-III
1
Vie 1°
N«v
38/
• s
— •
U22
• 1 38 AFTF R?-hr IRRA[JIATIO N
CE
OCELL 38 AFTER4-hr IRRADIATION
ACELL22 | | | |
10 20 30 40 50 60 i i (0 12 t4 16
T MElh r) TIMF (days )
0 2 4 6 8
y-IRRADIATION(hr)
FLUX =2.6 XlO6 r/hr

Fig. 4.1.27. Effect of Gamma Irradiation on Electrical Properties of Chemically Deposited PbS Films.

216
 PERIOD ENDING DECEMBER 37, 7957

UNCLASSIFIED
ORNL-LR-DWG 25458R

GAMMA IRRADIATION TIME (hr)

2 4 6

WAVELENGTH OF EXCITING RADIATION

CURVE a 18,000 A
> CELL 22
CURVE b 15,000 A
CURVE c 10,000 A

CURVE d 24,000 A
> CELL 5
CURVE e 16,000 A

CURVE f SAME AS d
CURVE g SAME AS <?

IRRADIATION (UPPER.
SCALE)
ANNEALING AFTER 2- •hr'
IRRADIATION (LOWER
SCALE)

0 5 10 15 20 25 30 35 40
ANNEALING TIME (hr)

Fig. 4.1.28. Effect of Gamma Irradiation on the Photoconductivity of Chemically Deposited PbS Films.

Further examination revealed that after irradiation the case prior to irradiation, a very erratic and
to an accumulated dose of 2.6 x 10 r, cell 5 appreciable noise signal was recorded. This
(curves d and e) showed a considerable drop in observation is, at best, qualitative, but clearly
photoresponse, but, after annealing at room tem indicative.
perature for 30 hr (curves / and g), the sensitivity A consideration of the effects of gamma irradia
returned almost to the initial value. After being tion on these thin films of chemically deposited
subjected to another gamma irradiation of equal PbS reveals a fundamental property. In brief, these
intensity, however, cell 5 had completely lost its ceils appear to have a fixed tolerance for gamma
light sensitivity. The photosensitivity of cell 5 irradiation. If the total dose is maintained below
was measured daily for three weeks following this tolerance, the physical changes effected in
irradiation, and it had not recovered any of its the matrix can effectively be annealed out, and
valuable photoconducting properties. The loss in the cell will revert to its original state. If the
photosensitivity appears to have been the result total irradiation exceeds this tolerance, either the
of a very large increase in the noise level of this cell will be permanently damaged or the annealing
cell after irradiation. In order to check this, the rate will be extremely slow. Which of these latter
cell was scanned on the Cary spectrometer in the possibilities is actually operative has not yet been
usual manner, except that the light source was definitely established by these studies.
turned off. Instead of the recorder showing no The possibility that fracture of the PbS film of
response under these conditions, as is normally cell 5 may have occurred was considered. The

217
ANP PROJECT PROGRESS REPORT

film was examined under a magnification of 1000, fundamental processes. Such knowledge is appli-
and no visible change in the smooth continuity of able both to the development of semiconductor
the layer was found. Electron diffraction studies devices with reduced sensitivity to radiation
should be made, however, before this possibility fields and to a better understanding of the barrier
is eliminated. mechanism.
The problem of establishing a gamma tolerance
Semiconductor Materials
for these cells is complicated by the rather
individual nature of the cells. Even in this A semiconductor is a material whose electrical
preliminary investigation a difference of 400% was conductivity has an energy of activation.10 When
observed in the sensitivity of two of these samples the electrons are all in their lowest energy state,
prior to gamma irradiation. the valence band is full and the conduction band
The PbS cells used in this study were supplied is empty (Fig. 4.1.29). In such a condition the
by the Eastman Kodak Company and the Bureau of electrons cannot be accelerated by an electric
Ships, United States Navy. field,' and hence no conduction can occur. If e g
is so large that electrons cannot be excited across
the forbidden band at normal temperatures, the
IRRADIATION EFFECTS IN SEMICONDUCTOR material is an insulator. If e is zero, the material
BARRIERS is a conductor; if e is such that electrons can be
J. C. Pigg thermally excited across the forbidden band, the
material is a semiconductor. Such conduction
C. C. Robinson 0. E. Schow
is a property of the material itself and as such is
The program of study of irradiation effects in called intrinsic semiconduction.
semiconductor barriers is rather specialized and
was initiated at a time when a large, jumbled body
of information had already been accumulated. UNCLASSIFIED
ORNL-LR-DWG 25484

Since reports of current progress are necessarily

isolated bits of information that appear out of CONDUCTION BAND
context to the over-all picture of the development,
each report, in order to be clear, must include the
background information and thinking for each bit of Uj E, FORBIDDEN BAND

data. Such a survey introduction to each report

tends, however, to obscure the current achieve
ments. It has been decided, therefore, to present VALENCE BAND

at this time a survey of the pertinent work and to

summarize the concepts on which the current work
is based. Subsequent reports will present program Fig. 4.1.29. Schematic Band Picture of a Semi
developments. conductor.
There are many groups engaged in studying
radiation effects in semiconductor components, When impurities are present in the crystal, such
and, since the motivations and objectives of the impurities may contribute energy levels which lie
investigations vary, the data accumulated and in the forbidden band. If such impurity atoms have
the conclusions drawn seem to be contradictory. more valence electrons than are necessary to
Therefore statements describing motivation, ob satisfy the bonding conditions of the crystal, the
jectives, and experimental details can frequently electron may be excited from its ground state E ,
remove contradictions in the data and in the to the conduction band — such atoms are called
conclusions. The purpose of the ORNL program is donors. Similarly an impurity atom having less
to contribute to an understanding of the effects of than the number of electrons required to satisfy
irradiation on semiconductor barriers in terms of the bonding conditions may accept electrons from
the valence band and leave a hole — such impurity
atoms are called acceptors. For germanium and
W. Shock ley. Electrons and Holes in Semiconductors, silicon, impurities from atoms listed in column V
with Applications to Transistor Electronics, Van
Nostrand, New York, 1950. of the periodic table act as donors, while impurities

218
 PERIOD ENDING DECEMBER 31, 7957

from atoms listed in column III act as acceptors. If n and a are known, the mobility of an extrinsic
Conduction by such carriers is called extrinsic semiconductor maybe determined from the extrinsic
semiconduction. It can be seen from Fig. 4.1.29 expression for conductivity:
that the activation energy for extrinsic conduction
cr = nep .
is smaller than that for intrinsic conduction.
If an excess of donor atoms is present, the Since electrons obey Fermi-Dirac statistics they
conduction will .occur by electrons (negative will be distributed throughout the available energy
carriers) in the conduction band - such a material states according to the Fermi function:
is said to be rc-type. If an excess of acceptor
impurities is present, the conduction will occur by 1
holes (positive carriers) in the valence band - such / =
,(E-Ef)/kT + ]
material is said to be p-type.
The conductivity a of a semiconductor is given
by10
T = absolute temperature,
(1) = neV-n + pep , k = the Boltzmann constant,
E = the energy under consideration,
where
E. = the Fermi energy,
n = electron concentration, which can be shown to be the chemical potential
p = hole concentration, of the carriers. When E, lies at the middle of the
e = electronic charge, forbidden band, the electrical properties are those
p *= electron mobility, which would be seen if E d , and E a were absent.
p = hole mobility. Thus the material exhibits intrinsic behavior when
If one type of carrier predominates, the expression the Fermi level is at the middle of the forbidden
for intrinsic conductivity, Eq. 1, is modified by band. When the Fermi level lies above the middle,
dropping the term corresponding to the minority the sample is rc-type; when it lies below the middle
carrier. of the band, the sample is p-type. If E, lies close
The concentration and sign of the majority carrier (within a few times kT) to either the valence band
can be obtained from Hall measurements. When a or the conduction band, the sample is said to be
current is established in a sample which has been degenerate. If E, lies several times kT from these
placed in a magnetic field, the motion of charged bands, the sample is said to be nondegenerate.
carriers will be perturbed by the Lorentz force For nondegenerate samples, Bose-Einstein statis
acting on the moving charges. A.voltage will be tics can be used to calculate carrier concentrations.
established across the sample perpendicular to the It can readily be seen that both n- and p-type
magnetic field and the current. The sign of this conduction occurs simultaneously in intrinsic
voltage gives the sign of the majority carrier. The material. In extrinsic material, conduction is
magnitude of this voltage is inversely proportional normally due to one type of carrier. However,
to the carrier concentration. The relation of the minority carriers can be produced in significant
quantities, together with appropriate constants quantities by photoelectric effects or by injection
needed to convert the units, is at a barrier. Injection of minority carriers by
forward-biased barriers accounts for the high
HI
forward currents in diodes and for emitter action in
(2) 7.38 10 18
(±V)t ' a transistor.
In the case of the transistor, the injected
minority carriers drift across the base of the unit
H = magnetic field, in oersteds, to the collector. The current amplification, a, of
/ =* current, in amperes, the transistor is determined by the ratio of the
t = thickness of sample perpendicular to both collector current to the emitter current. The in
H and /, in cm, jected carriers constitute a departure from equilib
V = Hall voltage, positive for holes, negative rium, and their concentration decays at a rate
for electrons, in volts, 1/r, where r denotes the lifetime of the minority
n = concentration of n- or p-type carriers. carrier. In order for a minority carrier to decay,

219
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
there must be a transition between the valence ORNL-LR-DWG 25485R

band and the conduction band. ' These transi SILICON

GERMANIUM
tions tend to occur via an intermediate step that
involves energy levels or trapping centers which V p (LOW P>
(0)
lie in the forbidden band: for example, E, and Ea
as in Fig. 4.1.29. Consequently the lifetime de p*- - — ^= =
creases as the densities of these energy levels
increase. \n jfp (HIGH P)

When a semiconductor is placed in a radiation

EXPOSURE (nvt) (nvt)
field, there are three major ways in which the BULK MATERIAL
radiation field can induce permanent changes in
the material. Capture of thermal neutrons produces ~~

isotopes which decay to leave substitutional im o

\ F {e)
>

purities in the crystal lattice. Elastic collisions 3 R\

>^
from energetic neutrons produce disorder in the n>
o

form of interstitials and vacancies, as well as INTRINSIC

small regions of disorder. Gamma photons produce
Compton electrons, which in turn can produce EXPOSURE {nvt) (nvt)
POINT CONTACT DIODES
interstitial-vacancy pairs. The results of all
these interactions is to produce new levels in the
forbidden band.

In germanium the center of density of the radia

tion-induced levels lies below the middle of the
forbidden band. The Fermi level will thus shift
toward this center of density, and conductivity EXPOSURE (»r) {nvt)
JUNCTION DIODES
behavior such as that illustrated in Fig. 4.1.30(a)
will be observed. The conductivity of rc-type Fig. 4.1.30. Comparisons of Effect of Radiation on
material will decrease as the Fermi level falls
the Behavior of Bulk Germanium and Silicon with the
until the sample becomes p-type. After the sample Effect on the Behavior of Germanium and Silicon
becomes p-type, the Fermi level will move toward Diodes.
the equilibrium position and the conductivity will
saturate. The latter half of the curve also illus neutron,15 0.0015 carrier per incident 1-Mev
trates the behavior of high-density p-type material. photon.16 For p-type material, the rate of carrier
Low-density p-type material in which the Fermi change by fast neutrons depends upon the initial
level is below the saturation position will exhibit concentration, 7 but it is less than that for ra-type
a drop in conductivity such as that illustrated in material.
Fig. 4.1.30(a). Both n- and p-type silicon become intrinsic under
fast-neutron bombardment.'8 The initial rate of
The initial rate at which carriers are re'moved carrier removal is about 5 carriers per incident
from rz-type germanium is 3.2 carriers per incident neutron. The behavior of the conductivity of
fast neutron, 4 0.02 carrier per incident thermal
15J. W. Cleland, K. Lark-Horovitz, and J. C. Pigg,
Phys. Quar. Prog. Rep. March 15, 1950. ORNL-694,
nR. N. Hall, Phys. Rev. 87, 387 (1952). p 56.

12W. Shockley and W. T. Reed, Jr., Phys. Rev. 87, 14J. W. Cleland, J. H. Crawford, Jr., and D. K.
835 (1952). Holmes, Phys. Rev. 102, 722-24 (1956).
13J. C. Pigg and C. C. Robinson, Elec. Mfg. 59(4), 17J. W. Cleland, J. H. Crawford, Jr., and J. C. Pigg,
116 (1957). Phys. Rev. 99, 1170-81 (1955).
18
14J. W. Cleland et al.. Phys. Rev. 83, 312-19 J. W. Cleland et al., Phys. Solids Institute Quar.
(1951). Prog. Rep. April 30. 1951, ORNL-1095, p 35.

220
 PERIOD ENDING DECEMBER 31, 7957

silicon under bombardment is illustrated by UNCLASSIFIED

ORNL-LR-DWG 25486
Fig. 4.1.30(a'). The acceptor and donor levels
both fall at about the middle of the forbidden band,
and hence the Fermi level moves to this position.

Semiconductor Devices

Irradiation-induced changes in carrier concen

tration, lifetime, mobility, and trap density all
contribute to the change in the operating behavior
of a semiconductor device. Units which depend
upon the existence of a p-n junction will, obvi
ously, be destroyed when all the material becomes
of the same type. The limiting conditions on the
operation of such a device in a radiation field are
r* *-i
thus those imposed by the radiation-induced Vd
changes in the bulk of the semiconductor. The
interim behavior, prior to complete breakdown, can
also be indicated by the change in properties of
the bulk material.
The properties involved in the case of a diode
may be illustrated by means of the equivalent
circuit shown in Fig. 4.1.31. There is a leakage
path over the surface of the unit, a base resistance vt
that results from the bulk resistance of the semi
conductor, and a barrier which can be assumed to
obey the diode equation

/ qV/kT
- Ve 1) ,

where V is the applied voltage19 and q is the

electronic charge. Both Rb and RL (Fig. 4.1.31)
may be functions of voltage or current, but, for
present purposes, they will be considered to be Fig. 4.1.31. Diode Equivalent Circuit.
ohmic.
With the use of the equivalent circuit of
Fig. 4.1.31 and the present knowledge of the curve b is obtained. Upon irradiation the semi
behavior of the electrical properties of semi conducting material is changed.20 Carriers are
conductors subjected to a radiation field, as removed from the ra-type material, and the Fermi
discussed above, the changes in the current- level in that region is lowered. In the p-type
voltage characteristic of a germanium p-n junction material, the Fermi level moves toward some
that can be expected as a result of irradiation may equilibrium position which depends upon the
be described. In the forward direction, the leakage temperature of the unit during exposure. In any
resistance can be ignored for a practical diode. case, the Fermi level does not move so far in the
The unit can be considered as a barrier in series p-type material as in the rc-type material. Hence
with the base resistance. Curve a of Fig. 4.1.32 the barrier will be lower and there will be a
illustrates the current-voltage relationship which corresponding increase in /Q. The resistance of
could be expected to exist for a pure barrier. In the «-type base, however, will increase because
a practical unit, however, the current is limited by of the decrease in carrier concentration and the
the base resistance, and a characteristic such as decrease in the lifetime of injected minority

19 20
H. C. Torrey and C. A. Whitmer, Crystal Recitifiers, J. C. Pigg, Solid State Semiann. Prog. Rep. Aug. 31,
chap. 4, McGraw-Hill, New York, 1948. 1953, ORNL-1606, p 79.

221
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
upon the type of semiconductor, the initial treat
ORNL-LR-DWG 25487 ment of the surface, and the chemical environment
of the surface. The ra-type germanium base units
•• irradiated to date have exhibited an increase in
RL with exposure. Hence a change in the reverse
characteristic from that shown as curve a in
Fig. 4.1.33 to that shown in curve b in Fig. 4.1.33
lb IC
can be expected.

1-
2
\ / UNCLASSIFIED
ORNL-LR-DWG 25488R

UJ
cc
\ /
o

UJ
o
\ /
1 /
o
o
// ••-
II

,*'
/
*-**

]/•'" APPLIED VOLTAGE

Fig. 4.1.32. Behavior of the Forward Current-Voltage

Characteristic of a Semiconductor Device.

carriers. At low voltages, high current can be

expected for a given voltage. At higher voltages,
where the current is limited by the base resistance,
APPLIED VOLTAGE
a lower current can be expected. The new current-
voltage characteristic will have the appearance Fig. 4.1.33. Behavior of the Reverse Current-Voltage
of curve c of Fig. 4.1.32. Under prolonged ex Characteristic of a Semiconductor Device.
posure both sides of the barrier will tend toward
the same ultimate carrier concentration, and hence
the sample will tend toward ohmic behavior. The expression for IQ is given by Shockley10 as:
The base resistance can be neglected in the
consideration of the reverse characteristic curve. '0 - !ps + lns '
The current through the unit is composed of two
inhere
parts, the diode current, /d, and the leakage
current, IL:
dp V/2
-qV/kt 'ts
ps - A1P
'd = '„0 - e rp
and 1/2
D.

'„ =Aqp\-^r
^ *L
A is the area of the junction, a is the electronic
As discussed in the case of the forward charac charge, p and n are the densities of the holes and
teristic, irradiation will result in an increase in electrons, respectively, D is the diffusion constant,
an increase in /.. The behavior of RL depends and ris the lifetime of the indicated carrier.

222
 PERIOD ENDING DECEMBER 31, 7957

The expression for / has been modified by Pell i-n instead of p-n. Such an i-n barrier would
to include an additional term to account for charge rectify, but not so well as the p-n barrier. The
generation in the barrier region. 1 This added barrier would be destroyed as the bulk silicon
term varies inversely with lifetime and barrier became intrinsic. During this time, the reverse
height. It has been shown that the minority-carrier current would also decrease toward the intrinsic
lifetime is extremely sensitive to energetic irradi level. Such a barrier would be unipolar, that is,
ation;22 hence, the increase in /. with irradiation it would conduct by one type of carrier; hence it
is expected to be larger than that anticipated on would be more efficient at high frequencies, as
the basis of the change in carrier concentration has been observed. The point-contact silicon
alone. This large expected change is in agreement diode is the most satisfactory for high-frequency
with experimental observations. applications: for example, 3-cm mixer applications.
The forward (F) and reverse (R) currents at 1-v Since transistor action involves the interaction
bias, as a function of neutron exposure, were of two such barriers, with one barrier biased in the
illustrated in Fig. 4.1.30, parts (b), (c), (e), and (/) forward direction and the other in the reverse
for germanium and silicon rectifiers. The ger direction, all the properties involved in diode
manium point contact behaves as illustrated by changes apply simultaneously. The collector
part (fe).23 The current in the forward direction cutoff current will increase as the /. of the
drops to a minimum as the bulk conductivity collector barrier increases. The slope of the
becomes intrinsic. As the bulk material becomes collector characteristic will change as the leakage
p-type and the conductivity increases, the forward resistance across the surface of the base changes.
current increases. The reverse current increases The emitter current-voltage characteristic will
monotonically as the barrier is destroyed, and the behave in a manner similar to the behavior of the
sample becomes ohmic as the Fermi level comes forward diode characteristic.
to the same position in the forbidden band through Sah et al.,26 in considering both the diffusion
out the sample. and charge-generation components of the emitter
The silicon junction diode compares with bulk current, express the current amplification, a, as
silicon in a similar manner.24 The forward current
drops as the bulk resistance increases. The sech (V/b/Lh)
a =
reverse current increases as the barrier is
destroyed. Eventually the sample becomes intrinsic hs + w
1 +
for both directions of bias. Jd\anh(Wb/Lb)
The characteristic behavior of a silicon point
contact is somewhat different,25 as is shown in where Lfo is the minority-carrier diffusion length in
Fig. 4.1.30(e). The forward current behaves in the the base, Wb is the base width, Jd is the injected
same manner as for the other diodes shown in current density flowing into the base, ] ,t is the
Fig. 4.1.30. The reverse characteristic, however, injected current density flowing into the emitter,
shows no strong effect as a result of a barrier and / is the current component resulting from
being destroyed. Such behavior can be understood recombination and generation in the emitter barrier.
if it is postulated that the region around the The recombination-generation current density for
contact behaves not as p-type but essentially as forward bias is shown to be
intrinsic material; that is, the barrier would be
kT qnl ,qV/2kT
/,.
rg
= 2 —
qE 2rn
21E. M. Pell, J. Appl. Phy. 26, 658 (1955).
22J. C. Pigg et al., ANP Quar. Prog. Rep. March 31, where E is the electric field at the junction, V is
1957, ORNL-2274, p 265. the voltage across the emitter barrier, and T. is
23J. H. Crawford, K. Lark-Horovitz, and J. C. Pigg, the minority-carrier lifetime in the base. While the
Phys. Quar. Prog. Rep. March 15, 1950, ORNL-694, diffusion component is inversely proportional to
p 64.
o a the square root of the minority-carrier lifetime,
J. C. Pigg and C. C. Robinson, Solid State Semiann.
Prog. Rep. Aug. 30, 1956, ORNL-2188, p 19, Fig. 12.
25J. C. Pigg, Solid State Quar. Prog. Rep. Feb. 10, 26C. T. Sah, R. N. Noyce, and W. Shockley, Proc.
1953, ORNL-1506, p 47. Inst. Radio Engrs. 45, 1228 (1957).

223
ANP PROJECT PROGRESS REPORT

the charge-recombination-generation component is then be expected to be governed by the following

inversely proportional to the first power. Hence condition:
Jrg will increase faster than Jd as Tdecreases. Lb (initial) » Wb
At low damage levels, where Lb « Wb, the and
relative importance of the charge-recombination-
generation current with respect to the diffusion h»h8'h' •
current varies with the applied bias. When the In the early stages of irradiation tanh (W,/L.)
emitter bias is approximately equal to or greater changes rapidly and there is an initial increase in
than the barrier height, the diffusion current a. However, as the irradiation progresses the
predominates. At bias voltages that are small concentration of radi at ion-induced levels in
compared with the barrier height, the charge- creases, and / increases faster than Jd. When
generation current predominates. Lb % Wb, tanh (Wb/Lb) saturates, but in this range
The diffusion length is connected with the sech (Wb/L/) decreases rapidly. Hence changes
lifetime by the relation in both / and sech (Wb/Lb) tend to decrease the
value of a. Such behavior has been experimentally
L = (DT^2 . observed27 (Fig. 4.1.34). For units with Wfe and
Lb of the same order of magnitude, such as low-
Where Efe becomes comparable with Wfe, the factor frequency power devices, this initial increase in
sech (\('b/Lb) takes control and decreases a a may not be observed28 (Fig. 4.1.35).
rapidly.
27
J. C. Pigg and J. W. Cleland, Solid State Quar.
In the range where Lb « Wb, tanh (Wb/Lb) is a Prog. Rep. Feb. 10, 1953, ORNL-1506, p 48, Fig. 28.
rapidly varying function compared with sech 28
J. C. Pigg and C. C. Robinson, Solid State Semiann.
(\/Lb) The change in a with irradiation would Prog. Rep. Feb. 29, 1956, ORNL-2051, p 63, Fig. 54.
UNCLASSIFIED
SSD-B-741
DWG. I8922AR
INTEGRATED FLUX (nvt)

30 35 40 45 (Xt0'°)

50 100 200 300 400 500 600 700 800 900 1000 1200 1400 1600 (X<0'°)
INTEGRATED FLUX (nvt)

Fig. 4.1.34. Power Supply for Rectifier Experiment.

224
 PERIOD ENDING DECEMBER 31, 7957

UNCLASSIFIED
ORNL-LR-DWG 13759

0.9

0.8

AMF LIFICAT ON: A- : DROP ACROSS RESISTOR IN COLLECTOR LEG

0.7

; 0.6

1 \
£0.4
--NO. 4 FLUX 1 86 XlO1 V
•

V
V%

0.2
l«>»
-~,»

««..
0.1 -7- •—•—
' >-•-•_
N 0.6; FL JX 1.86 <10,0w ^ —•—•• - • •> »

6 7 8 9 10 II 13 14 (5 16 17UI0'3)
INTEGRATED NEUTRON FLUX (nvt.)

Fig. 4.1.35. Changes in Amplification of Two Minneapolis-Honeywell H-2 Power Transistors as a Function of
Integrated Fast-Neutron Exposure.

In high-frequency units having a very thin base UNCLASSIFIED

region, much greater stability should be expected. (X10~3) ORNL-LR-DWG 15063
•^-
When Wb/Lb « 1, -*i , , , , 4-5.
<
35 ma 1

sech (Wb/Lb) %1 and tanh (Wb/Lb) %- (Wb/Lb) . 4.9 ma^

c 5 r
In the denominator / varies with 1/V and Jd
tanh (Wb/Lb) varies as 1/r. The collector current, —, .- i
1—"
as a function of irradiation for various emitter
currents of a surface-barrier transistor, is shown29 TRANSISTOR: PHILCO TYPE L-510C
NO. A- j^f-f

in Fig. 4.1.36. The unit is changed very little by ir BIAS: 1 volt

radiation to an integrated flux of 1014neutrons/cm2. ROOM TEMPERATURE

The value of / CO', the collector cutoff current, i 1.7 ma

behaves like the value of the reverse characteristic.

L
i**^^_
[ • 1.3 ma

The /. increases rapidly because of the increasing 1 1

0.75 ma
i
y
saturation diffusion current and the increasing
contribution of the charge-recombination-generation
current. Because of the / component, /Q does
not saturate. The leakage component can become nvt, INTEGRATED NEUTRON FLUX (neutrons/cm2)
the predominant term if surface changes occur.
Fig. 4.1.36. Changes in Collector Currents of a
29, Philco L5106 Surface-Barrier Transistor as a Function
J. C. Pigg and C. C. Robinson, Solid State Semiann.
Prog. Rep. Aug. 30, 1956, ORNL.2188, p 19, Fig. 11. of Irradiation.

225
This will be determined by the type of material UNCLASSIFIED
ORNL-LR-DWG 254B9
and surface treatment of the unit under con
sideration.
The transistor characteristics will change in
the manner illustrated30 in Fig. 4.1.37. Prior to
irradiation there is a family of curves, / , with
a cutoff / ; subsequent to irradiation, the cutoff
current increases to / 'Q. If surface damage
occurred, the slope would increase as illustrated.
Because of the decrease in minority-carrier
lifetime, a diminishes, as is indicated by the
decreased separation of the curves. Although the
/. of the barrier increases, a smaller part of the
emitter curve reaches the collector, and the in
creased base resistance, in the case of silicon
or n-type germanium, results in a lower value for
/'. The curves will continue to collapse until the
sample tends toward ohmic behavior.
The current state of the knowledge of the effect
of radiation on semiconductor devices, such as
transistors and diodes, is sufficient for qualitative
understanding of the observed behavior; that is,
the observed changes can be explained in terms of
predictable physical changes in the crystal
structure of the unit. In principle, it should also
V, COLLECTOR VOLTAGE
be possible toquantitatively describe the behavior,
but techniques for determining the pertinent
parameters for a completed unit have not yet been Fig. 4.1.37. Collector Voltage vs Collector Current.
developed.

30G. L. Keister and H. V. Stewart, Proc. Inst. Radio

Engrs. 45, 931 (1957).

226
 Part 5

SHIELDING

E. P. Blizard
 5.1. SHIELDING THEORY

EFFECT OF ANISOTROPIC SCATTERING ON The resulting ratios of the total number flux,
THE NEUTRON FLUX IN AIR energy flux, and dose rate are presented as a
C. D. Zerby function of beam angle of emission in Figs. 5.1.1,
5.1.2, and 5.1.3. It may be observed that in each
For many calculations of fluxes and dose rates case the ratio is less than 1 for beam angles
resulting from air-scattered neutrons, isotropic greater than 10 deg, which indicates that the
scattering in the center-of-mass system is assumed. results obtained by using isotropic scattering in
The question naturally arises as to the validity of the center-of-mass system are, in general, con
this assumption. In order to obtain some insight servative.
into the problem, the Monte Carlo code used
previously for calculating the fluxes and dose It was found from the calculations in which
rates of multiply scattered neutrons was altered to isotropic scattering in the center-of-mass system
include an anisotropic scattering cross section was used that the equation which fits the neutron
proportional to (1 + a cos 0) in the center-of-mass
flux data best is
system. Thus, the differential cross section is
given by
r,(EQ) n - 0
da (2)
(1) —- = K(l + a cos 6) VQ> " 4rrg
dil

The value of the constant K was chosen to make

J, Hilgeman, F. L. Keller, and C. D. Zerby, Monte
the integral of the differential cross section equal Carlo Calculations of Fluxes and Dose Rates Re
to the total cross section. The value of a is sulting from Neutrons Multiply Scattered in Air, ORNL-
2375 (to be published).
arbitrary and constant over the whole energy
range.
The cases investigated were for a point mono- UNCLASSIFIED
2-01-059-268

directional beam of neutrons in an infinite volume

of air. The beam angle with respect to the source
detector axis, 6, and the value of a were the only
parameters varied. The constants of the problem
were: ^ o = 0.3

Source energy, £n 5 Mev

Separation distance, g 50 ft o -0.7

Air density, p 0.00125 g/cm3

Three problems were solved for each value of 6,

corresponding to the values of a of 0, 0.3, and 0.7.
By expressing the value obtained from the calcu
lation as a function of a, that is, <t>(a), the ratio
b7" / \9, BEAM ANGLE
„/
of <£(a) to <t(0) was obtained. This ratio then
represents the factorial change which would be
expected in the results of a calculation in which 60 90 120

isotropic scattering was assumed if anisotrop) 9, BEAM ANGLE (deg)

were to be introduced.
Fig. 5.1.1. Ratio of the Total Neutron Flux in Air with
Anisotropic Scattering to That Obtained with Isotropic
For example, see C. D. Zerby, A Monte Carlo Scattering. Detector 50 ft from a monodirectional beam of
Calculation of Air-Scattered Neutrons, ORNL-2277
(April 23, 1957). 5-Mev neutrons. Air density = 0.00125 g/cm •

229
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-059-269
$N(0) = neutron flux per source neutron,
2s(Eq) = macroscopic scattering cross section
at the source energy,
g = separation distance.
^^0-0.3 Equation 2 was also found to fit the neutron flux
calculated for singly scattered neutrons on the
assumptions of isotropic scattering in the labo
<t>E(o)
\» =0.7 ratory system and no attenuation on either leg.
*E(o)

A similar expression fits the calculated total

dose rate, 5> (0), per source neutron. This
equation is

C(E0) £ (E0) n -
ANGLE
(3) V»
o- 1.2 4ng sin 6

where C(E„) is the flux-to-dose-rate conversion

0 30 60 90 120 150 180 factor at the source energy.
9, BEAM ANGLE (deg)
Since such a simple approach was so effective
Fig. 5.1.2. Ratio of the Total Neutron Energy Flux in in providingafit for the data for isotropic scattering
Air with Anisotropic Scattering to That Obtained with in the center-of-mass system, a similar set of
Isotropic Scattering. Detector 50 ft from a mono- equations for the cases of anisotropic scattering
directional beam of 5-Mev neutrons. Air density = was investigated. The assumption that the dif
0.00125 g/cm3. ferential scattering cross section given in Eq. 1
was effective in the laboratory system and a
UNCLASSIFIED
Z-01-059-270
calculation of the singly scattered flux in which
attenuation on both legs was neglected yielded
the following equation:

sin 6
^»^o = 0.3 (4) $N(a) =*N(0) 1 - a
n - 6

where ®N(0) is as given in Eq. 2. From Eq. 4 and

<t>D(g) a -0.7 a similar equation for <£>D(a), it is seen that
*n(0)

**<*> *D(«> / sintf

(5) ^-77 = —~ - 1 - a
*N(0) 3>D(0) IT - 6

/I
/ \ e, be Values of these ratios calculated with the use
/ of Eq. 5 are compared with results obtained in
0URCE POINT DETECTOR

Monte Carlo calculations in Tables 5.1.1 and

5.1.2. As may be observed, the results of the
60 90 120
Monte Carlo calculation when anisotropic scat
9, BEAM ANGLE (deg)
tering is considered are not fitted by the values
Fig. 5.1.3. Ratio of the Total Neutron Dose Rate in obtained with the use of Eq. 5 as closely as would
Air with Anisotropic Scattering to That Obtained be desired. In fact, it is obvious from a qualitative
with Isotropic Scattering. Detector 50 ft from a analysis that Eq. 4 should become less accurate
monodirectional beam of 5*Mev neutrons. Air density = the more anisotropic the scattering. This is
0.00125 g/cm3. borne out in the tables.

230
 PERIOD ENDING DECEMBER 31, 1957

Table 5.1.1. Ratio of the Total Neutron Flux in Air BREMSSTRAHLUNG SPECTRA IN SODIUM
with Anisotropic Scattering to That Obtained IODIDE AND IN AIR3
with Isotropic Scattering C. D. Zerby H. S. Moran

Error
In order to provide some information on the
of Formula
production of secondary electromagnetic radiation
6 (deg) Formula Monte Carlo
in materials, a calculation was set up on the
(%)
Oracle to determine the spectra of bremsstrahlung
a = 0.3
radiation resulting from the degradation of electrons.
The results of calculations for two materials,
2 0.997 1.008 - 1.1 sodium iodide and air, are shown in Figs. 5.1.4
0.911 + 3.5 and 5.1.5.
30 0.943

90 0.809 0.806 + 0.3 Presented in more detail in a report by C. D. Zerby

and H. S. Moran, Bremsstrahlung Spectra in Nal arid
130 0.737 0.769 - 4.2 Air, ORNL-2425 (Feb. 25, 1958).

180 0.700 0.760 - 7.9

UNCLASSIFIED
2-01-059-275

0.7 —-J -
"•= —

2 0.992 1.020 - 2.7

2 —

30 0.866 0.796 + 8.8

I02 -
.

90 0.554 0.537 + 3.2 I - --1 £"„= INITIAL ELECTRON ENERGY (mc c2)
5

•130 0.386 0.453 -14.8

2 •• —

180 0.300 0.425 -29.4

10 -~^--
— - -

5
" VN \
Table 5,1.2. Ratio of the Total Dose Rate in Air Sr~
£0=2I.O
with Anisotropic Scattering to That Obtained
with Isotropic Scattering
•—• \—
-V-
|s^£0-<6.0
^
Error 2
{/)

#(deg) Formula Monte Carlo of Formula o

10
*
\f0 = 6.0"
o
X
(%) a.
5 N^- f0=2. 3 ^S.

a = 0.3 £ 0=l.l J^
~*N^
2 0.997 1.011 - 1.4

30 0.943 0.920 + 2.5 —

90 0.809 0.810 - 0.2

130 0.737 0.768 - 4.0

180 0.700 0.761 - 8.0

2 . . _ . _

a = 0.7 10"

2 0.992 1.026 - 3.3

2
30 0.866 0.815 + 6.1
10"
90 0.554 0.545 + 1.7 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

130 0.386 0.449 -14.0 fo-<

180 0.300 0.430 -30.2

Fig. 5.1.4* Bremsstrahlung Spectrum in Sodium Iodide.

231
ANP PROJECT PROGRESS REPORT

10
_..| -| !
dE 2nr0NZ f /32(E - 1)
_„.z;=t
(1) - — = 1 In
5 \ I l l
dx o2 o/2/i
2/^(1 -_. tf)
«2x
l\ P'
2 w A= PH0T0N ENERGY (/77„C2)

I
Tmi1 j--..-. * - —
- {ly/\ - /32 - 1 +jS2) In 2+
—--
5

~V
2

10"
^
+1- /32 +1(1 -Vl- yS2)2 I ,
V^'
—

2
Xs s: -^

S5
^fe J>N where
E = total relativistic electron energy (m-c2),
10~: _ -

s^N
^v° s_ —pv ^ I = mean excitation potential (m.c2),
5

:z:-
P2 = (v/c)2 « (E2 - \)/E2,
2 rQ = classical electron radius (cm),
to"
N = atomic density (cm-3),
." —-1-

5
_ _ . „
— -

Z = charge number.

2 - - -
The Born-approximation cross section for the
10" ~.^=.
production of light (bremsstrahlung radiation) by
-

1
an electron in the field of a nucleus without
5
screening is given by6
2
—

10"
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
*2 , .2
roz Nh 4 P] + P
(2) $a(E,k) = •i - 21
137 k p 3
Fig. 5.1.5. Bremsstrahlung Spectrum in Air. py

e,£ e,e\ 8 EE\

In the calculation, the degradation of the elec + L
trons was assumed to occur only by inelastic 3 pp^
P? PP ;
collisions with the bound electrons of the medium
and by radiative collisions with nuclei. It was
also assumed that the initial electron energy
was sufficiently low for the energy degradation ££, + px
by inelastic collisions to predominate. For the E2E2 +P2P2) + e -

P3P* 2pPi
cases presented in Figs. 5.1.4 and 5.1.5 the
last assumption is certainly true.4
The energy loss per centimeter of path length
(stopping power) for an electron making inelastic
collisions is given by the equation5

4W. Heitler, The Quantum Theory of Radiation,

3d ed., p 252, Clarendon Press, Oxford, 1954.
H. A. Bethe and J. Ashkin, Experimental Nuclear W. Heitler, The Quantum Theory of Radiation, 3d ed.,
Physics (ed. by E. Segre), p 254, Wiley, New York, 1953. p 245, Clarendon Press, Oxford, 1954.

232
 PERIOD ENDING DECEMBER 31, 1957

where the following relations are used: (6) 9,(y) - 5.089e-°-6587r + 15.661e-0-0481? ,
fEEy +pP] - 1 (7) 92(y) = 0,(y) - 0.595e-3-04? .
L = 2 In
The quantity y measures the effect of screening.
For y >> 1 the screening can be neglected; how
e = 2 In (E + p) , ever, for y - 0 the screening is complete. Al
though Bethe and Ashkin givea series of equations
(3) -< e, = 2 In (E, +p,) similar to Eq. 4 that are valid for various ranges
of y, the use of Eqs. 6 and 7 in Eq. 4 makes Eq. 4
E, = E - k ,
valid for all values of y.

P2 = E2 - 1 , With the use of Eqs. 1, 2, and 4, the total

bremsstrahlung spectra from an electron slowing
P21 = £1 " 1 • down in a compound is given by

As may be seen from Eq. 3, Eq. 2 is simply a

function of £ and k, where E has the same def '0 I *.(E,A) dE
/»El
inition as in Eq. 1 and k is the photon energy (8) K(EQ,k0)dkQ dk
0 '
expressed in units of m0c2. The expression •'fc+l
$a(E,k) dk is the electron cross section (cm"')
for producing a photon in the interval k to k + dk.
When the electron screening of the nucleus is where K(E0,kQ) dkQ is the number of photons pro
taken into account and Eq. 2 is reduced to the duced by an electron slowing down from energy EQ
relativistic case, E >> m^c , the following in the interval kQ to kQ + dkQ. The subscript /'
equation7 is obtained: refers to the ;th element of the compound. The
quantity <£(£,&) is obtained from Eq. 2 or Eq. 4,
depending on which is appropriate in the course of
r\z2M 4 integration.
(4) <bb{E,k) =
137 ~k Equation 8 was solved on the Oracle to provide
the data presented in Figs. 5.1.4 and 5.1.5. For
convenience, the results are presented as the
x <
number of photons per unit (&/En) - 1 as a function
of (k/E0) — 1. The integral under the curve
represents the total number of photons released.
In Z > , A CALCULATION OF GAMMA-RAY DOSE RATES
3\ E
FROM NEUTRON CAPTURE IN AIR

where y is given by F. L. Keller C. D. Zerby

W. W. Dunn10
(5) y = 100 Experiments at the TSF and elsewhere have
E(E - k)Z 1/3
indicated that the gamma rays which are produced
The functions <pAy) and cf>2(y) are presented by the capture of neutrons in air may become a
graphically by Bethe and Ashkin, however, they major contribution to the dose rate inside the crew
are well represented with less than 0.3% error compartment of a heavily divided aircraft shield.
by the empirical relations Very little quantitative work has been done,
however, on the problem of determining at a point
in air the gamma-ray dose rate which results
H. A. Bethe and J. Ashkin, Experimental Nuclear
Physics (ed. by E. Segre), p 260, Wiley, New York, 1953.
from these air-capture gamma rays.
8lbid.. p 259.
10
9Ibid., p 262. On assignment from USAF.

233
ANP PROJECT PROGRESS REPORT

The problem may be idealized to that of deter divided into a number of solid-angle intervals,
mining the air-capture gamma-ray dose rate and and estimates of the angular distribution of the
angular distribution at a detector located an dose rate at the detector were made.
arbitrary distance from a point source of mono
energetic neutrons where both the source and The neutron histories were generated in exactly
detector are considered to be embedded in an the same manner as described in the report on air
infinite volume of air. The calculation should in scattering. At each neutron collision, the
clude the contributions from captures which take probability of capture was determined, and a
place during the slowing-down process as well as statistical estimate was made of the capture
contributions from captures which occur during gamma-ray dose-rate contribution at the detector
the diffusion at thermal energy. A portion of the from that collision. The neutron was then allowed
results of a parameter study of this problem by to scatter randomly and to continue with a reduced
the Monte Carlo technique for both point isotropic weighting. The neutrons were degraded in energy
and line-beam sources is presented here. The at each scattering by the standard elastic-scattering
complete results will be published in a separate relations until the neutron energy reached 0.025 ev.
11
report, After the neutron energy was degraded to 0.025 ev,
the neutron was permitted to scatter and diffuse
Calculational Method without further energy loss. A neutron history
The Monte Carlo machine code which was used was terminated if the neutron weight after the /th
for this problem was a modification of the Monte collision was less than about l/600th of its
Carlo code which was used to calculate neutron original weight or if the collision occurred at a
scattering in air. This original code was com distance greater than 32 mean free paths from the
pletely described in the report cited; therefore, the detector position. This enabled a 0.025-ev neu
discussion here will be limited to a very brief tron to make approximately 45 collisions before
description of the general features of the code as its history was terminated. For higher source
modified to perform the present problem. energies the average number of collisions was, of
The total cross sections for oxygen and nitrogen course, greater.
were taken from BNL-325 and were considered
to be identical with the scattering cross sections, The capture gamma rays were assumed to be
except for the absorption cross section of nitrogen. emitted isotropically from the collision point, and
The absorption cross section of oxygen is ex an exponential buildup factor was used in the
tremely small and was neglected. The absorption dose-rate calculation. The spectrum of capture
cross section of nitrogen is made up of the (n,p) and gamma rays which was used in the dose-rate de
(n,y) cross sections, which were taken to be termination was from the recent work of Bartholomew
1.70 barns and 0.08 barn, respectively, at 0.025 ev and Campion. 4 The dose-rate contributions were
and were assumed to have a 1/f dependence. The stored in appropriate solid-angle intervals at the
scattering probability was taken as isotropic in detector under the assumption that all gamma rays
the center-of-mass system. from a given event entered the detector in the
The source was taken to be either a point solid-angle interval which contained the radius
isotropic source of monoenergetic neutrons of vector from the detector to the collision point.
energy En or a point source which emitted a All the results were normalized to one neutron per
monoenergetic line beam of neutrons of energy second leaving the source, and all the results
E at an angle 0q with respect to the source- reported here were from calculations which used
detector axis. The space about the detector was 500 neutron case histories. It should be pointed
out that because the method of statistical esti
,]W. W. Dunn, F. L. Keller, and C. D. Zerby, A mation was employed, the results were relatively
Calculation of Gamma-Ray Dose Rates from Neutron insensitive to the number of neutron histories
Capture in Air, ORNL-2462 (unpublished).
12 used.
C. D. Zerby, A Monte Carlo Calculation of Air-
Scattered Neutrons, ORNL-2277 (April 23, 1957).
13
D. J. Hughes and J. A. Harvey, Neutron Cross G. A. Bartholomew and P. J. Campion, Can. J.
Sections. BNL-325 (July 1, 1955). Phys. 35, 1347 (1957).

234
 PERIOD ENDING DECEMBER 31. 7957

The result of the Monte Carlo calculation was noted that the dose rate varies with the neutron
checked against the result obtained from the source energy approximately as E~ .
neutron flux given by the direct solution to the
transport equation for the case of a point isotropic In thes'e calculations, each gamma-ray dose-rate
source of 0.025-ev neutrons located a distance of contribution at the detector was also stored with
15 m from the detector. The assumptions involved respect to the energy, at capture, of the neutron
in both calculations were the same except that the which gave rise to the contribution. Figures 5.1.7,
Monte Carlo calculation took the neutron-scattering 5.1.8, and 5.1.9 show histograms of the dose rate
probability to be isotropic in the center-of-mass plotted as a function of neutron capture energy
system, whereas the transport equation treated for several of the cases run in this study. It may
the scattering as isotropic in the laboratory system. be seen that, for relatively low neutron source
This would be expected to lead to a slightly lower energies, a large part of the gamma-ray dose rate
value for the Monte Carlo result. The comparison at the detector comes from neutrons which are
showed that the Monte Carlo result was about 10% captured at energies that are within a factor of a
smaller than the result from the check calculation.
Hence, the results reported here are believed to UNCLASSIFIED

be good to about 10 to 20%. It is interesting to 2-01-059-266

III!
note that the dose-rate contribution from the flux
1.4
given by the nonasymptotic part of the solution to
NOTE: 48% OF DOSE RATE FROM Ec > a£0,
the transport equation was about 30% of the total 1.2
f0 - SOURCE ENERGY
dose rate in the above case. £ lo -

|0.8 -

Results for Point Isotropic Sources *

0.4
A series of calculations were performed for point
isotropic neutron sources with energies of 0.025, 0.2 -

0.1, 1.0, 100, and 10,000 ev and a source-detector 0.025 0.0375 0.050 0.0625 0.075 0.0875 0.IO0
separation distance of about 50 ft. The air density £c, NEUTRON ENERGY ATCAPTURE (ev)
was taken to be 0.001205 g/cm3, which corresponds
to approximately sea-level conditions. A plot of Fig. 5.1.7. Air Capture Gamma-Ray Dose Rate 50 ft
the total gamma-ray dose rate vs the neutron from a Point Isotropic Neutron Source as a Function of
source energy is shown in Fig. 5.1.6. It may be Ec for E0 ~ u#1 ev*

UNCLASSIFIED
2-01-059-264
10

55 10

? ,0
< 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2 5 10 20 50 100 200 500 1000 2000 5000 10,000
£„, NEUTRON ENERGY (ev)

Fig. 5.1.6. Air Capture Gamma-Ray Dose Rate at a Detector 50 ft from a Point Isotropic Monoenergetic Neutron Source.

235
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Z-01-059-Z65 sources. The source-detector separation distance
was 65 ft, and the air density was again taken
as 0.001205 g/cm3. The curves are relatively
flat over the region from 30 to 180 deg, and they
NOTE: 85% OF DOSE RATE FROM f<.>O.I;
(8% FROM £c>a£0, £0* SOURCE ENERGY
become flatter as the source energy is increased,
as would be expected.

UNCLASSIFIED
2-01-059-263
-10
10"

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
£c, NEUTRON ENERGY AT CAPTURE (ev) 5

Fig. 5.1.8. Air-Capture Gamma-Ray Dose Rate 50 ft

from a Point Isotropic Neutron Source as a Function of
2
Ec for EQ = 1.0 ev.

UNCLASSIFIED
S 10"-II
(xlO""4 ) 2-01-059-267
3.2
I I I I I I I I
tn u.u
2.8
5
"

2.4 -

S 2.0 NOTE: 40% OF DOSE RATE FROM fc>IOev —

2
1 l6 —

1 '2 ^ t = 1.C ev
4

0.8 -
12

7 = 10.0ev
-

0.4
^1
~n= 1000 cv
I \ " "I 1 1 1 1
-

5
0 10 20 30 40 50 60 70 80 90 100
£c, NEUTRON ENERGY AT CAPTURE (ev)
I

Fig. 5.1.9. Air-Capture Gamma-Ray Dose Rate 50 ft

from a Point Isotropic Neutron Source as a Function of
Ec for EQ = 100 ev.
13

30 60 90 120 150 180

of the source energy, where a •= [(A - 1)/(A + l)]2 90, ANGLE OF EMISSION (deg)
and A is the average mass number for air. This
fact throws considerable doubt on calculations of
Fig. 5.1.10. Air-Capture Gamma-Ray Dose Rate at a
the air-capture gamma-ray dose rate which use
Detector 65 ft from a Line-Beam Neutron Source Emitted
the results of the standard age-diffusion equation
at Various Angles from the Source-Detector Axis.
for the neutron flux, since these results are known
to break down for neutron energies which are
within a factor of a of the source energy.
Since most aircraft reactor shields are designed
so that most of the neutron leakage is from the rear
Results for Line-Beam Sources
of the shield, the curves of Fig. 5.1.10 indicate
Figure 5.1.10 shows the results for a series of that a knowledge of the exact surface angular
problems which was run for line-beam neutron distribution of these low-energy neutrons is rela
sources. The gamma-ray dose rate at the detector tively unimportant. The important thing is to have
is shown plotted as a function of the angle of a knowledge of the spectrum and total number of
emission of the beam for various monoenergetic low-energy leakage neutrons. Because of the

236
 PERIOD ENDING DECEMBER 31. 7957

extreme difficulty in making accurate experi- distances, and a more detailed discussion of
mental determinations of low-energy neutron these calculations will be found in the complete
spectra, it will probably be necessary to generate report.15
this information by Monte Carlo techniques. TT
Additional results,, including
..a angular
u..yw.u. distributions
u. ^.i .uu.iu.i^ rCalculation
, W,; ,W> DTV- F' L^ Kel'er'
of Gamma-Ray Dose a"d,C- ?' Zt,by,'
Rates from A
Neutron
of the dose rate and results for other separation Capture in Air, ORNL-2462 (unpublished).

237
ANP PROJECT PROGRESS REPORT

5.2. LID TANK SHIELDING FACILITY

W. Zobel

PRATT & WHITNEY SERIES OF ADVANCED 85, 90, and 91. The fast-neutron and gamma-ray
SHIELDING MATERIALS STUDIES data represent instrument response measurements
E. A. Warman
converted to ergs/g-hr-w of tissue dose as cali
brated with known sources. The thermal-neutron
A second group of tests of advanced shielding data are instrument response measurements nor
materials was performed in the Lid Tank Shielding malized to cadmium-difference gold-foil fluxes.
Facility (LTSF) in order to obtain information The fluxes were corrected for flux depression by
required in the Pratt & Whitney shield design the 2-mil-thick foils and for self-absorption and
effort (LTSF exp. 70). The purpose of these tests self-shielding. The instrumentation used in these
was to determine the feasibility of using various tests was the same as that described previously.
materials in the shield of a fast-neutron reactor The measurements made beyond all the con
intended for possible aircraft-propulsion appli figurations are presented here, along with a brief
cation. discussion; the data have not yet been analyzed.
The materials tested were lithium hydride, de Configurations OB, 43, 45, 48, 81, and 82. -
pleted uranium, and borated water, with particular Gamma-ray dose-rate measurements taken beyond
emphasis being placed on a study of the production configuration 81, as shown in Fig. 5.2.1, indicate
of secondary gamma rays in the uranium. In most the effect of substituting 1 in. of stainless steel
of the tests, these materials were used in com for 1 in. of borated water in configuration OB,
bination with nickel and stainless steel. The which consisted entirely of borated water. The
configurations tested are described in Table 5.2.1, dose rate was decreased by about 27%.
and the physical properties of the materials used In configuration 43, 2 in. of nickel was inserted
are presented in Table 5.2.2. in front of the stainless steel of configuration 81.
The components in configurations 36 through 42 This resulted in a decrease in the gamma-ray dose
and 90 and 91 were installed dry in the LTSF rate of about 55% from that of configuration 81 at
configuration tank, which is a steel tank with a distance of 40 cm from the source plate and a
^.-in.-thick walls. The tank has a ^-in.-thick decrease of about 42% at a distance of 160 cm.
aluminum window on the source side; an un The change in slope of the dose-rate curve
avoidable 1.6-cm space between the aluminum possibly indicates a "harder" spectrum of gamma
window and the source plate was filled with air rays.
contained in a plastic bag. The borated water In configuration 82, a 1/2-in.-thick slab of
used in configurations 36 through 42 and 90 and 91 depleted uranium was placed behind the stainless
was contained in an aluminum tank with /g-'n.- steel of configuration 43. The slope of the
thick walls. This tank was 31 in. long. Con gamma-ray curve was changed, and the curve
figurations 43 through 49 and 79 through 86 were itself was lowered considerably. The gamma-ray
submerged in borated water in the configuration dose rate at 40 cm from the source plate was
tank. lowered approximately 70% and at 160 cm it was
Gamma-ray dose-rate measurements were taken lowered about 92%. The change in the slope of
beyond each configuration, and fast-neutron and the curve indicates a "softer" gamma-ray spectrum.
thermal-neutron measurements were made beyond The measured reduction and the change in slope
all configurations, except configurations 85, 86, also indicate a reduction in the contribution of
90, and 91. No neutron measurements were made primary gamma rays from the source plate.
beyond configuration 86, and only thermal-neutron The effect of inserting 1 ft of borated water in
measurements were made beyond configurations front of the uranium in configuration 82 is shown
by gamma-ray measurements beyond configuration
On assignment from Pratt & Whitney Aircraft.
48. The dose rate was decreased by 80% at a
The first group of tests was described by W. Zobel
et al., ANP Quar. Prog. Rep. June 30, 1957, ORNL-
2340, p 285, and by E. A. Warman, ANP Quar. Prog. D. R. Otis, ANP Quar. Prog. Rep. Dec. 31, 1956,
Rep. Sept. 30, 1957, ORNL-2387, p 300. ORNL-2221, p 331.

238
 PERIOD ENDING DECEMBER 31, 1957

Table 5.2.1. Description of Configurations Tested in Pratt & Whitney Series

Configuration Location of Pre be, Unavoidable Gap

Description of Configuration
Number zQ (cm) Within Configuratio

OB Borated water (B-H~0) in configuration tank 2.6C

36 2 in. of Ni + 1 in. of stainless steel + 2 ft of LiH 77.4 3.4 cm of air

+ B-H-0 in an Al tank

37 2 in. of Ni + 1 in. of stainless steel +3 in. of 85.6 4.0 cm of air

depleted U+ 2 ft of LiH + B-HjO in an Al tank
38 2 in. of Ni + 1 in. of stainless steel + 2 ft of LiH 85.2 3.6 cm of air
+ 3 in. of depleted U+ B-HjO in an Al tank
39 2 in. of Ni + 1 in. of stainless steel + 2 ft of LiH 81.4 3.6 cm of air
+1^2 in. of depleted U+B-HjO in an Al tank
40 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 85.6 4.0 cm of air
+ 3 in. of depleted U + 1 ft of LiH + B-HjO in
an Al tank

41 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 81.4 3.6 cm of air

+1\ in. of depleted U+1ft of LiH +B-Hj 0
in an Al tank

42 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 85.0 3.4 cm of air

+l]^ in. of depleted U+1ft of LiH +1\ in. of
depleted U + B-HjO in an Al tank
43 2 in. of Ni + 1 in. of stainless steel + B-HjO 12.8 2.6 cm of B-HjO
44 2 in. of Ni + 1 in. of stainless steel + 3 in. of 21.0 3.2 cm of B-HjO
depleted U + B-HjO
45 2 in. of Ni + 1 in. of stainless steel +2 ft of B-H20 75.8 0.8 cm of B-HjO
+\\ in. of depleted U+B-HjO
46 2 in. of Ni + 1 in. of stainless steel +2 ft of B-HjO 81.8 3.0 cm of B-HjO
+ 3 in. of depleted U + B-HjO
47 2 in. of Ni + 1 in. of stainless steel + 1 ft of B-HjO 51.8 3.5 cm of B-HjO
+ 3 in. of depleted U + B-HjO
48 2 in. of Ni + 1 in. of stainless steel + 1 ft of B-HjO 46.0 1.5 cm of B-HjO
+1}2 in. of depleted U+B-HjO
49 2 in. of Ni + 1 in. of stainless steel + 1 ft of B-H20 81.0 2.2 cm of B-HjO
+l}2 in. of depleted U-*- 1ft of B-HjO +1^ in.
of depleted U+ B-HjO
79 2 in. of Ni + 1 in. of stainless steel + L in. of 23.1 4.0 cm of B-HjO
boral + 3 in. of depleted U + B-HjO
80 2in. of Ni +1in. of stainless steel +\ in. of 25.0 4.7 cm of B-HjO
boral + 3 in. of depleted U + -^
\ in. of boral
+ B-H20
81 1 in. of stainless steel + B-H,0 6.6 1.5 cm of B-HjO

239
ANP PROJECT PROGRESS REPORT

Table 5.2.1 (continued)

Configuration Location of Probe, Unavoidable Gap

Description of Configuration
Number zQ (cm) Within Configuration

82 2 in. of Ni + 1 in. of stainless steel + 1 /« in. of 16.8 2.8 cm of B-H20

depleted U + B-H20
83 2 in. of Ni + 1 in. of stainless steel + 3 in. of 27.6 2.1 cm of B-H20
B-H20 + 3 in. of depleted U+ B-HjO
84 2 in. of Ni + 1 in. of stainless steel + 6 in. of 35.4 2.4 cm of B-H20
B-H20 + 3 in. of depleted U + B-HjO
85 2 in. of Ni + 1 in. of stainless steel + 30 in. of 97.0 3.0 cm of B-H20
B-HJ3 + 3 in. of depleted U + B-HjO
86 2 in. of Ni + 1 in. of stainless steel I- 18 in. of 65.6 2.0 cm of B-H20
B-H20 + 3 in. of depleted U + B-HjO
90 2 in. of Ni + 1 in. of stainless steel + 2 ft of LiH 88.0 2.5 cm of air

+ 4/2 in. of depleted U + B-HjO in an Al tank

91 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 88.2 2.7 cm of air

+i\ in. of depleted U+• 1ft of LiH +B-HjO

in an Al tank

a Actual distance from source plate to back of solid configuration; that is, the borated water at the back of the
configuration is not included.
Includes the 1-cm recession of the aluminum window in the configuration tank.
Probe inside the configuration tank at the aluminum window.

Table 5.2.2. Physical Properties of Materials Used for LTSF Mockup Tests
of Advanced Shielding Materials (Pratt & Whitney Series)

Actual Thickness
Material Description
per Slab (cm)

Type 347 stainless 2.54 5-ft x 5-ft x 1-in. slab; composition, 58.35 wt % Fe-18.0 wt % Cr-
steel 10.3 wt % Ni-2.0 wt % Mn-1.0 wt % Si-0.08 wt % C-
0.04 wt %P-0.03 wt %S; density, 7.8 g/cm3
Nickel 2.54 Metallic slab, 64 X 66 X 1 in.
•5
Uranium 3.81 Metallic slab, 52 x 557)6 x 1^1 in., depleted to 0.24% U235
Lithium hydride 32.0 5 X 5 X 1-ft slab encased in aluminum cans (/.-in.-thick walls);
composition, 95 wt % LiH—5 wt % unaccounted for (probably
oxygen and hydrogen); density, 0.75 g/cm

Boral 1.3 48 X 48 X /,-in. slab with 0.05-in.-thick Al cladding; composition,

80 wt %AI-15.7 wt %B-4.3 wt %C; density, 2.60 g/cm3
Borated water Composition, 1.34 wt % natural boron; density, 1.05 g/cm at 20 C

240
 PERIOD ENDING DECEMBER 31, 7957

* 'a*--!**
2-01 -057-70-391
10

^v

10

nv -s^
j^^CONFIGIIRATinN 43:
2 n.OF Ni f lin. OF STAINLESS STEEL + B-H,0
L ^^'v- ^v. ^CONFIGURATION OB: | | |
\ L
10 R-H..0 IN CnNFIRIIRiTinW TflWlC

~ "^cr lvs^ IGURATIC

"o-^ ~^<;
'-CONFIGURA TION 82: lin. OF STAINLESS STEEL + B- H20
,

2 n.OFNi + lin.OF
O
Q 10
S rAINLESS STEEL
V2 in. OFDEPLETE
/\
1 u + B-H20
^V^ ' """^''vl ^^ '^s^^**^
\ •"i L
"^,
^^•IL it^^*^i>'*+^***y
^i y

fes^
"""^^l. ^s^
CONFIGURATION 45: i^^^ '^^
2 in. OF Ni +1 in. OF STAINLESS STEEL + 2 f t 0FB-H20
10 ^^

CONFIGU RATION'18:
2in.0F Ni +lin. OF STAINLESS STEEL +1 ft OF B-H.,b +\% in.OF DEPLETED U+B-HyO^fcs^v^ I5^k
^^.^
"^i i

10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.1. Gamma-Ray Tissue Dose Rates Beyond Configurations 43, 45, 48, 81, and 82.

distance of 60 cm from the source plate and by Configurations 82, 44, 79, and 80. - Configu
50% at 160 cm. The change in the slope of the ration 44 represents configuration 82 with a second
curve indicates a harder spectrum. The difference 1/2-in.-thick slab of uranium added to it. The
between the curves for configurations 82 and 48 degradation of the higher energy gamma rays in
is believed to be almost entirely due to the the added uranium resulted in a softened emergent
reduction in secondary gamma-ray production in gamma-ray spectrum, as is evidenced by the
the uranium. The cause of this reduction is, of change in the slope of the curve of the data
course, the decrease in the thermal- and fast- plotted in Fig. 5.2.4.
neutron fluxes incident on the uranium.
The borated water space of configuration 48 was In configuration 79, a 72-in.-thick boral curtain
increased to 2 ft in configuration 45. The resulting was placed on the source-plate side of the 3 in.
gamma-ray curve appears to be raised slightly. of uranium in configuration 44. The gamma-ray
This, however, may be attributable to experimental dose-rate curve was lowered by approximately 20%.
error. This decrease is believed to be due almost
Corresponding thermal-neutron flux and fast- entirely to a reduction in the number of secondary
neutron dose-rate measurements for the configu gamma rays resulting from (n,y) captures in the
rations discussed above are shown in Figs. 5.2.2 uranium. The reduction, of course, results from
and 5.2.3. the reduction in the neutron flux incident on the

241
ANP PROJECT PROGRESS REPORT

«Jp*'<'
2-01-057-70-392
10'
_ —

— CONFIGURATION OB: B-H,0 IN CONFIGURATION TANK

^
,. .*—
1 1 1 1 i 1
^- CONFIGURATION 81: lin. OFSTAINLESS STEEL + B-H20
10°

5 CONFIGURATION 43: 2 in. OF Ni + 1in. OF STAINLESS STEEL + B-H20 -

2
m
-CONFI 3URATI0 N82: 2 n. OFNi + lin. OF STAINLE:ss STEE L + lV2i 1. OF

105 \ DEPL ETEDU 1-B-H2C

\V\
5

\\\
2 \5 ^
^
10"

10-

3 2

102
o
or
'\V

or.

I 10

;V%
^v
2

1
.

1 1 1 1 1 1 1
5 " CONFIGURATION 48- 2 in OF Ni + 1 In. OF STAINLESS STEEL + 1 ft OF
^^. ,_._.,-.,„ . ~ . .z~ !v
2 tax i!l
10-1
V
5
CON FIGURATION 45: 2 in. OF Ni + 1 in. Oh SI AINLtbb 51 ttL ^0
+ 2 ft OF B-H20 + 1V2 in. OF DEPLETED U+ b-m2u ^£
NS \
2

10-2
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.2. Thermal-Neutron Fluxes Beyond Configurations 43, 45, 48, 81, and 82.

242
 PERIOD ENDING DECEMBER 31, 7957

3 2-01 -057-70-393
10

\
V ^ ^ ONFIGURATION OB: B-H20 IN CONFIGURATION TANK

.\
10
2 \ k"
[*
-o
•
' :ONFIGURATION 43: 2 in. OF Ni + 1 in. OF STAINLESS STEEL + B-H, 0

&V ^-CONFIGURAT ION 81: 1 in OF STAINL ESS STEEL 4-B-H20

to

^$s.
\VI
v^
\r
A^'
\\
TN'^\.
vsa
V,
s< ?K\v
vO\
\0\
\\\( >

> k
10
-1
-
CONFIGU NATION 48: 2 in. OFNi + 1 in. OF STA NLESS STE EL +
1ft OF B-H20 + 1/2 in. OF DEPLETED U + B-l1,U —
Sk ^V
XV
^ w

LESS STEEL+ l'/2 in. OF DEPLETED U+ B--H?0 —

^W ^J i

2 V c?^^ ^v
\SS\ \V

V^V.
CONF GURATION 45: 2 in. OF Ni +1 in. OF STAINLESS S TEEL+_^!-2^V
2 ft 0 F B-H20 + 1V2 in. OF DEPLETEO UH HB-H20 —
a\ k
3
10
10 20 30 40 50 60 70 80 90 100 110 120

Fig. 5.2.3. Fast-Neutron Tissue Dose Rates Beyond Configurations 43, 45, 48, 81, and 82.

243
ANP PROJECT PROGRESS REPORT

2-01-057-70-394

10^

102 _- CONFIGURATION 82: 2 in. OF Ni + 1 n. OF STAINLESS STEEL + 1V2 n. . _

2
CONFIGURATION 44: 2 in. OF Ni + 1 in. OF STAINLESS STEEL + 3 in.
10 r*. • f OF DEPLETED U+ B-H20 '

5
^s>.J'i \
\ 1 1 1- A r- ' J
^^>
NSs^ V '

f- V2 in. OF BORAL + 3 in. OF DEPLETED U+ B-H20

O
Q
2
^T\ f^.

1
. _ . . 1.. _ _ _ _ _ —
r^
5 1

<
or

<
2
5
5 + V2 in. OF BORAL + B- H20 ^^^ i^^^
<
1CT4
ID
_^^i I^>^
5
jC^T-^^?^
5^\>.
2

10"°
i

^
20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.4. Gamma-Ray Tissue Dose Rates Beyond Configurations 44, 79, 80, and 82.

uranium, that is, from the absorption of neutrons by about a factor of 30 in the region farther from
in the boral. the source plate, as shown in Fig. 5.2.7. The
In configuration 80, a second k-in.-thick boral radical change in the slope of the curve points
curtain was placed behind the 3 in. of uranium in out the much softer energy spectrum of gamma
configuration 79. The gamma-ray dose rate was rays being detected.
almost unchanged at a distance of 34 cm from the Configuration 46 was similar to configuration 44
source, but it was lowered approximately 16% except that 2 ft of borated water was inserted in
at 160 cm. The change in the slope of the curve front of the 3 in. of uranium. The resulting
is possibly due to the reduction of secondary decrease in the gamma-ray dose rate is believed
gamma-ray production in the borated water backing. to be due principally to the reduction in secondary
The neutron measurements beyond configurations gamma-ray production in the uranium. If this is
44, 79, 80, and 82 are shown in Figs. 5.2.5 and assumed to be true, the difference between dose
5.2.6. rates beyond the two configurations can be taken
Configurations 43, 44, 46, and 49. - Gamma-ray to be the approximate secondary gamma-ray con
dose-rate measurements beyond configuration 44 tribution to the dose rate beyond configuration 44.
illustrate the effect of adding 3 in. of uranium to In configuration 49, the borated water spacing
configuration 43. The gamma-ray dose rate beyond of configuration 46 was broken up into two equal
configuration 44 was lowered about 80% from that parts, the second section dividing the uranium in
beyond configuration 43 in the forward region and half. This lamination proved to be a more

244
 PERIOD ENDING DECEMBER 31, 1957

2-01-057-70- 395

50 60 70 80
DISTANCE FROM SOURCE (cm)

Fig. 5.2.5. Thermal-Neutron Fluxes Beyond Configurations 44, 79, 80, and 82.

245
ANP PROJECT PROGRESS REPORT

2-01-057-70-396

— CONFIGURATION 44: 2 in. OF Ni + 1in. OF STAINLESS STEEL + 3 in. OF DEPLETED U + B-H20

1 1 1 1 '
1 1 1 1
CONFIGURATION 79: 2 in. OF Ni + 1 in. OF STAINLESS STEEL + 72 in. OF BORAL + 3 in. OF
DEPLETED U + B-H20
10'

n.OF

DEPLETED U + 1/2in. OF BORAL

10

- CONFIGURAT ION 82: 2 in. OF Ni + 1 in. OF STAINLESS S 'EEL + lV2in.

OF D EPLETED U + B-H20

n- 10"

^*\

10"
20 30 50 60 70 80 90 110

z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.6. Fast-Neutron Tissue Dose Rates Beyond Configurations 44, 79, 80, and 82.

246
 PERIOD ENDING DECEMBER 31, 1957

2-01-057-70-397
10

( »«^

10'

1 1 1 1
5
~"">^ ^-""'CONFIGURATION 44: 2 in. OF Ni + 1 in. OF STAINLESS STEEL +
fXL 3in.OF DEPLETED U+ B-H20
2 ^<

10

"••Sfc^^
^"""•"•s^
5 ^< i.^

^""""••i '-,
2
^""""•""•H
1 ^""""^Ow, "
—^

10"

CO ^FIGURATION 46: 2 in. OF Ni + 1 in. OF STAINLESS

^^
ST EEL + 2 ft OF B-H20+3in.0F DEPLETED U+B-H2 o-^
10
— 1 """^-^
I
i-^^
CONFIGURATION 49: 2in. OF Ni + 1in. OF STAINLESS STEEL + 1ft OF B-H20 +
11/2 in. OF DEPLETED U+ 1ft OF B-H,0 + iV? in.OF DEPLETED U+ B-H20-^ "^•^

10
30 40 50 60 70 80 90 100 110 120 130 140 150 160
z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.7. Gamma-Ray Tissue Dose Rates Beyond Configurations 43, 44, 46, and 49.

effective gamma-ray shield, the dose rate being In going from configuration 44, which contained
reduced about 14%. only the unavoidable amounts of borated water
The thermal-neutron and fast-neutron measure between the solid sections of the configuration,
ments beyond configurations 43, 44, 46, and 49 to configuration 83, which contained 3 in. of
are presented in Figs. 5.2.8 and 5.2.9. borated water between the stainless steel and the
Configurations 44, 46, 47, 83, 84, 85, and uranium, the gamma-ray dose rate was reduced by
86. — One series of configurations was tested in about 55%. The slope of the curve was changed
order to investigate the production of secondary only slightly. A combination of two phenomena
gamma rays in 3 in. of depleted uranium as a is believed to be the cause of this reduction.
function of the amount of borated water preceding The first, and probably the most important, is that
it. The results are shown in Fig. 5.2.10. the production of (n,y) capture gamma rays in the

247
ANP PROJECT PROGRESS REPORT

2-01-057-70-398
10
l =t =
--*-CON r|fI1 iRATinw ^^-
_ "•

2 in .OF Ni + un.ui- biAiNLtbb bl ttL + _-M,u

^CONFIG URATION 44:

10 \ v 1 2in.0FNi-Mir .OF STAINLESS STEEL + 3 in. OF DEPLETEC)
_-H20 -
\ \ U +
\ \
&
T,
*t
\
\
10

* 2

? 10

§ 10
=^k 1
-z.
o
or
r^ VS. r
Vv 2in.0F Ni +1 in. OF STAINLtbb
LlJ
Z
Ns, -STEEL + 5>« of n-H-n + ^in
OFDEPl.ETED U + B-H2C
< 10 -
i i\ >>
or
UJ
X

_ -*^
\S

=1 V

^A
>V>
Vr^
CONFIGURATION 49: \
2 in .OFNi + 1in. OF STAINLESS STEEL + 1 f t OF E!-H20

10
4-1 % in. OF DEPLETED U+ 1ft OF B- H20 +\% in.OF
fjFPI FTFD ll+R-H,n
*
^
10
10 20 30 40 50 60 70 80 90 100 110 120 130
z0, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.8. Thermal-Neutron Fluxes Beyond Configurations 43, 44, 46, and 49.

248
 PERIOD ENDING DECEMBER 31, 1957

2-01-057-70-399
10J

CONFIGURATION 44: 2 in. OF Ni + 1 in. OF STAINLESS STEEL

—Ju •^ + 3 in. Ul- UtKLL I tU U 1- B -

10' \ i
V
V\
\ _- CONFIGURATION 43: 2 in. OF Ni + 1 in. OF STAINLESS STEEL + E3 H,0

\r
\
>
i l\

10
—N^
yv
v3
\\
^.
vv

III
h- ^
<
or

</> 1
o
Q

UJ

5
in

F-

Z
tl
_ 2
1 l\
10
\ \
\ \
0)
<
\ \
\ S

\\

10
-2
np Mi + i i t of stain _ESS STEE
^
+ 2 ft OF B-H20 + 3 in. OF DEPLETED U + B-H ,0 — Ws
v*\
i i i i i v^\
Tv^. ri
+ \ ft OF B-H20 + 1V2 in. OF DEPLETED U+ 1ft OF B-H20
• " "2"

10
20 30 40 50 60 70 80 90 100 110 120

z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.9. Fast-Neutron Tissue Dose Rates Beyond Configurations 43, 44, 46, and 49.

249
ANP PROJECT PROGRESS REPORT

80 90 100 110 120 130 150 160

.?„, DISTANCE FROM SOURCE (cm)

Fig. 5.2.10. Gamma-Roy Tissue Dose Rates Beyond Configurations 44, 46, 47, 83, 84, 85, and 86.

uranium was greatly reduced by the reduction of of the uranium, and a further reduction in the
the incident thermal-neutron flux by the 3 in. gamma-ray dose rate of about 55% was effected,
of borated water. The second effect is that which again is attributable to the reduction of
fissions in the uranium slab caused by inter the number of secondary gamma rays produced in
mediate and fast neutrons were largely eliminated the uranium.
by the additional neutron shielding. There was 18 in. of borated water in configu
In configuration 84, the borated water within the ration 86. The gamma-ray dose-rate curve was
configuration was increased to 6 in., and the lowered about 25% from that beyond configuration
additional 3 in. lowered the gamma-ray dose rate 47 and had a slightly different slope, possibly
by approximately 57% more. This was caused by indicating a harder spectrum. In configuration 46,
a further reduction in the secondary contribution. in which the amount of borated water was in
Configuration 47 had 1 ft of borated water in front creased to 24 in., the gamma-ray dose-rate curve

250
 PERIOD ENDING DECEMBER 31, 7957

was agaifr^eWered slightly in the forward region, about 53% at 83 cm from the source plate and
but, owing to the change in slope, it was raised by about 70% at 110 cm (see Fig. 5.2.14). This
slightly in the regions farther from the source. It reduction in flux and change in slope shows the
might be concluded that the curve for configuration relative attenuation effectiveness of lithium hy
46 represents only gamma rays from the source dride and borated water. This rather large re
plate; that is, secondary gamma rays produced in duction in the thermal-neutron flux is accompanied
the uranium this far out in the borated water are by a 70% reduction in the fast-neutron dose rate.
negligible. The obvious advantages of using lithium hydride
Configuration 85 contained 30 in. of borated as a shielding material are illustrated by its good
water, which resulted in a further slight reduction neutron attenuation properties, with only a slight
in the gamma-ray dose rate, but this reduction is increase in the gamma-ray dose rate, and relatively
sufficiently within experimental error as to be light weight.
inconclusive. In configuration 39, a 1/,-in.-thick slab of
Thermal- and fast-neutron measurements beyond depleted uranium was placed behind the 2 ft of
configurations 44, 46, 47, 83, and 84 are shown lithium hydride in configuration 36. The addition
in Figs. 5.2.11 and 5.2.12, respectively. Thermal- of the uranium resulted in a reduction of the
neutron measurements beyond configuration 85 are gamma-ray dose rate by about a factor of 20 (see
shown in Fig. 5.2.11, but no fast-neutron measure Fig. 5.2.13). This is about the attenuation to be
ments were made for this configuration. Neither expected for gamma rays of 2 or of 7 Mev, and
thermal- nor fast-neutron measurements were made is not far from that to be expected for intermediate
for configuration 86. energies. Hence it is unlikely that neutron-
Configurations 36, 39, 41, and 43. - A com induced gamma rays in the uranium are a sig
parison of configurations 43 and 36 shows the nificant contribution to the gamma-ray dose with
effect of adding 2 ft of lithium hydride to the the uranium in place.
nickel-stainless steel region. There are, how The slope of the thermal-neutron flux curve for
ever, two other differences between these configu configuration 39 was drastically changed from that
rations. The use of the aluminum-canned lithium for configuration 36 as shown in Fig. 5.2.14.
hydride introduced a /.-in. thickness of aluminum The flux was increased by about a factor of 4
in configuration 36 on each side of the lithium at 88 cm from the source plate but only by about
hydride. Also, the solid components in configu 20% at 100 cm. The curves would probably cross
ration 36 were dry and were backed by borated at greater distances. This change might be
water in an aluminum tank, which introduced an attributed to the fact that the uranium is a strong
additional k-in. thickness of aluminum in configu inelastic scatterer, and thus it enhances the softer
ration 36 just behind the lithium hydride. Between flux near the slab at the expense of that which
the solid components in configuration 36 there would penetrate to greater distances. The fast-
were distributed air gaps totaling 3.4 cm, while neutron data for these configurations (Fig. 5.2.15)
in configuration 43 there were distributed borated indicate that the addition of the uranium behind
water gaps totaling 2.6 cm. The use of lithium the lithium hydride caused an increase in the
hydride increased the gamma-ray dose rate, as fast-neutron dose rate next to the slab, but the
shown in Fig. 5.2.13. The fact that the gaps were data cannot be accurately extrapolated to indicate
occupied by air rather than borated water probably the situation farther back.
accounted for some of the increase because of In configuration 41, the 1/2-in.-thick slab of
the slightly higher thermal-neutron flux incident uranium was positioned between two 1-ft slabs
on the nickel, stainless steel, and lithium hydride. of lithium hydride. The gamma-ray dose rates
The decreased density in the region of the shield were approximately the same as those behind
occupied by the lithium hydride (from 1.05 g/cm configuration 39 (Fig. 5.2.13), but the thermal-
for borated water to 0.75 g/cm for lithium hydride) neutron fluxes were reduced considerably (Fig.
no doubt accounted for a considerable portion of 5.2.14).
the increase. Configurations 38, 39, 40, 41, 90, and 91. - In
The thermal-neutron flux beyond configuration 36 configuration 38, a second 1 A-in. slab of uranium
was lowered from that beyond configuration 43 by was added to configuration 39, with a resulting

251
ANP PROJECT PROGRESS REPORT

2-01 -057-70-401

40 50 60 70 80 120

z,, DISTANCE FROM SOURCE PLATE (cm!

Fig. 5.2.11. Thermal-Neutron Fluxes Beyond Configurations 44, 46, 47, 83, 84, and 85.

252
 PERIOD ENDING DECEMBER 31, 1957

2-01-057-70-402
3
10

\ , V™ GURATION 44: 2 in. OF Ni + 1 in. OF STAINLESS STEEL

+ 3 in OF DEPLETED U+B- H20

10
? \ i
a

,- CONFIGURATION 83: 2 in. OF Ni + 1 in. OF STAINLESS STEEL

+ 3 in. DF B-H20 4-3 in. OF DEPLETED U+B-H20

S 10

\/ + 6 in. OF B-H20 + 3 in. OF DEPLETED U+ B-H20

<
or

\
O
o ^\^
^L
\r

in 5 Ot-« CONFIGURATION 47: 2 in. OF Ni + 1 in. OF STAINLESS

STEEL + 1 ft OF B-H20 + 3 in. OF DELETED U + B-H20

10-1

/^°
FC*

10 -2
Vv\
V^«
x&.
x^.
^^ ,ao
'g-
CONFIGURATION 46: 2 in. OF Ni + 1 in. OF STAINLESS J\
10-3
20 30 40
STEEL +

50
2

60
ft OF B-H; 0

70
+ 3 in. OF DEPLETE!3

80 90
U+ B-H20

100
N \

110 120

z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.12. Fast-Neutron Tissue Dose Rates Beyond Configurations 44, 46, 47, 83, and 84.

253
ANP PROJECT PROGRESS REPORT

2-01-057-70-403
10' 1

l»s*'..*i*V

^1 1 -CONFIGURATION 43: 2 in. OF Ni + 1in. OF STAINLESS STEEL + B-H2 0

^\
^«

<

^_—CONFIGURATION 36: 2 in. OF Ni + lin. OF STAINLESS

10 •
"^^^<!* STEEL+ 2 ft OF LIH + b-H,U IN AN Al IANK
^^^.^ —ex.
^^^^

S 5
^^"-^ ^

^^( ^^D
m 1

f
t/ - -.. ^
<n 5
^*>"V_ +2ftOF LiH +1V2 in. OF DEPLETED U+B-H20 IN AN Al TANK

A^ ^^*_-
\^ /
A

^^, iL

A
10
GURATION 41: 2in. OF Ni +1in. OF STAINLESS STEEL + ' """V.^ /
/ ""'" "
1ft OF LiH +11/2 in. OF DEPLETED U+1ft OF LiH +B-H20 IN AN Al TANK^Lj
*-k
• • - - • - -

10
50 70 80 90 100 110 120 130 140 160
zQ, DISTANCE FROM SOURCE(cm)

Fig. 5.2.13. Gamma-Ray Dose Rates Beyond Configurations 36, 39, 41, and 43.

lowering of the gamma-ray dose rate by about a to form configuration 40. Unlike configuration 41,
factor of 18 (Fig. 5.2.16). This large reduction in which 1/2 in. of uranium was positioned be
again indicates negligible secondary gamma-ray tween the lithium hydride layers, the gamma-ray
production in the uranium. The thermal-neutron dose rate in this case was increased by about
curve for configuration 38 was raised over that 80% (Fig. 5.2.16). This increase might be prin
for configuration 39 in the forward region (Fig. cipally attributable to the production of secondary
5.2.17) because of the replacement of borated gamma rays in the uranium.
water with the uranium. The fast-neutron dose The depleted uranium gamma-ray shield was
rate (Fig. 5.2.18) was lowered about 15%. increased to a thickness of 41/2 in. in configu
The 3 in. of uranium in configuration 38 was rations 90 and 91. Configuration 90 represented
repositioned between the two lithium hydride slabs the addition of another 1k-in. slab of uranium

254
 PERIOD ENDING DECEMBER 31, 1957

'/_***»

2-01-057-70-404

o
ac
\-
Z>
UJ
_•
I
_l
<

tr
UJ
I

z0, DISTANCE FROM SOURCE (cm)

Fig. 5.2.14. Thermal-Neutron Fluxes Beyond Configurations 36, 39, 41, and 43.

255
ANP PROJECT PROGRESS REPORT

2-01-057-70-405
10

^-.*.ft*f!fc-

2 in. OF Ni + 1 in. OF
H,0

10

CONFIGURATION 36: 2 in. OF Ni + 1 in.

OF STAINLESS STEEL + 2 ft OF LiH
+ B-H20 IN AN Al TANK

CONFIGURATION 39: 2 in. OF

< Ni+ 1 in. OF STAINLESS STEEL
or

UJ + 2 ft OF LiH + 11/2 in. OF

in
O DEPLETED U + B-H20 IN AN
o
Al TANK
UJ
_•

10

O CONFIGURATION 41: 2
or
H
Ni + 1 in. OF STAINLESS
Z> + 1 ft OF LiH + 1V2 in. OF
DEPLETED U + 1 ft OF
10
01
<

10
60 70 80 90 100 110 120 130
zQ, DISTANCE FROM SOURCE (cm)

Fig. 5.2.15. Fast-Neutron Tissue Dose Rates Beyond Configurations 36, 39, 41, and 43.

256
 PERIOD ENDING DECEMBER 31, 1957

W-.*m*i

2-01-057-70-406

5X10

80 90 100 HO 120
z0, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.16. Gamma-Ray Tissue Dose Rates Beyond Configurations 38, 39, 40, 41, 90, and 91.

to configuration 38, and the gamma-ray dose rate the uranium. The accompanying decrease in the
beyond configuration 90 was considerably lower thermal-neutron flux measured behind the configu
than that beyond configuration 38 (Fig. 5.2.16). ration (Fig. 5.2.17) tends to justify at least a part
The slope of the curve was changed slightly, the of this assumption. No fast-neutron data were
dose rate for configuration 90 being a factor of 7 obtained for configurations 90 or 91. It should
lower in the forward region but only a factor of be pointed out that the statistics on the data for
5 lower in the region farther from the source plate. this group of configurations were much poorer than
The thermal-neutron flux curve (Fig. 5.2.17) was those for other data included in tfhis report.
shifted by the additional uranium. Configurations 36, 37, 38, 40, and 42. - In con
In configuration 91, the AL in. of uranium was figuration 37, a 3-in.-thick slab of depleted
positioned between the two 1-ft-thick slabs of uranium was placed between the stainless steel
lithium hydride. As was the case when configu and the 2-ft-thick lithium hydride neutron shield.
ration 40 was compared with configuration 38, the The gamma-ray dose rate beyond the configuration
gamma-ray dose rate beyond configuration 90 was (Fig. 5.2.19) was about a factor of 9 below that
raised by roughly a factor of 4 over that beyond beyond configuration 36, in which there was no
configuration 91 (see Fig. 5.2.16). This increase uranium. This substantial decrease is believed
is also believed to be due to an increase in the to be largely the result of the increased attenu
production of secondary gamma rays in the uranium ation of primary gamma rays. A comparison of
caused by the greater neutron flux incident on the data for configurations 37 and 38 indicates

257
ANP PROJECT PROGRESS REPORT

2-01-057-70-407
10

CONFIGURATION 39: 2 in. OF Ni + 1 in. OF STAINLESS STEEL + 2 ft OF

LiH + 1V2 in. OF DEPLETED U + B-H20 IN AN Al TANK

95 100 105 110 115 120

z0, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.17. Thermal-Neutron Fluxes Beyond Configurations 38, 39, 40, 41, 90, and 91.

258
 PERIOD ENDING DECEMBER 31, 1957

105 110 125

DISTANCE FROM SOURCE (cm)

Fig. 5.2.18. Fast-Neutron Tissue Dose Rates Beyond Configurations 38, 39, 40, and 41.

that the large difference in the gamma-ray dose repositioning. The dose rates were identical for
rate between these configurations is essentially both configurations at 95 cm from the source
the contribution of secondary gamma rays from plate, but the dose rate for configuration 42 was
the uranium to the total measured dose beyond lower by about 15% at 140 cm. The change in
configuration 37. the slope of the curve possibly indicates softening
of the spectrum as a result of the increased
Configuration 42 represents a repositioning of degradation of primary gamma rays. A comparison
configuration 38 so that both the uranium and the of configuration 42 with configuration 40, in which
lithium hydride components are broken up into only the lithium hydride is separated into two
two separate layers. The slope of the gamma-ray layers, shows that the dose rate beyond configu
dose-rate curve (Fig. 5.2.19) was changed by this ration 42 is about 45% lower. This decrease is

259
ANP PROJECT PROGRESS REPORT

*-.!**»

2-01-057-70-409
10Z

^—- CONFIGURATION 36: 2 in. OF Ni + 1in. OF STAINLESS STEEL + 2 ft OF

•-,' ^ *--^ LiH + B-H20 IN AN Al TANK
10
^ 4 i.

^*^^
" •* ^
^^^
4 i^

^~""o—.

- CONFIGURATION 37: 2 in. OF Ni + 1 in. 01-

STAINLESS STEEL + 3 in. OF DEPLETED U
+ 2 ft OF 1 iH 4- R-H.O IN AW Al TANK

UJ
l-
<
or
UJ
in i
O
o
.CONFIGURAT ION 40: 2 in. DF Ni + lin. 0
•a 10" ' STAINLEISS STEEL +1 ft OF LiH + 3 in .OF
or /
/ DEPLET t u u T i n u r Lin t- _-i-ipU IN
/ AN Al TtMK
/
< i

A^H
_

CONFIGURATION z>a: c\ in. ur- i\ii + i in. ur

• STAINLESS STEEL + 2 ft OF LiH + 3 n.
i. / OF DEPL ETED U + B-h 20 IN AN Al TANK

i ^^^1
L~^^^
1
10"
• ~^*"' ._

—-^_

^"^ , i i

"sS^1 '--~_
CONFIGURAT ION 42:
STAINLEISS
2 in. OF Ni + 1 in. 0
STEEL + 1ft OF LiH + l'/j in. 01-
is^i ~- ^
1
DEPLET ED U +1 ft OF I i'H +1Vo in OF DEPLETED U
-/
+ B-H2C) IN AN Al TAN K

10"
80 90 100 110 120 130 140 150 16C
z0, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.19. Gamma-Roy Dose Rates Beyond Configurations 36, 37, 38, 40, and 42.

260
 PERIOD ENDING DECEMBER 31, 7957

believed to be due principally to a reduction in GENERAL ELECTRIC COMPANY SERIES

the production of secondary gamma rays in the OF ADVANCED SHIELDING MATERIALS
second 1 /--in.-thick slab of uranium. STUDIES
Thermal-neutron and fast-neutron measurements
D. W. Cady4 J. M. Miller
beyond configurations 37 and 42 are compared with
corresponding measurements beyond configurations In addition to the series of tests described in
36, 38, and 40 in Figs. 5.2.20 and 5.2.21, re the preceding paper, the production of secondary
spectively. On assignment from Wright Air Development Center.

2-01-057-70-410

90 95 100 105
zQ, DISTANCE FROM SOURCE (cm)

Fig. 5.2.20. Thermal-Neutron Fluxes Beyond Configurations 36, 37, 38, 40, and 42.

261
ANP PROJECT PROGRESS REPORT

95 100 105 120

Z0, DISTANCE FROM SOURCE PLATE

Fig. 5.2.21. Fast-Neutron Tissue Dose Rates Beyond Configurations 36, 37, 38, 40, and 42.

gamma rays in Hevimet, stainless steel, and neutron shield. All the configurations are sub
depleted uranium is being studied in a series of merged in oil, and measurements of the gamma-ray
tests designed by the General Electric Company dose rate are made in the oil at various distances
to support their ANP effort. This experiment is beyond the lithium hydride. Thermal-neutron flux
essentially a continuation of an earlier experiment, measurements are made beyond a few configu
No. 69, which was described previously. In rations. The effect of inserting boral within
general, the configurations used in this series several of the configurations is also being studied.
consist of an oil layer adjacent to the source The configurations tested thus far are described
plate followed by a 4-in.-thick beryllium layer, in Table 5.2.3, and the physical properties of the
a gamma-ray shield of one of the three materials various materials used in the configurations are
being studied, and a 12-in.-thick lithium hydride presented in Table 5.2.4.
The first group of tests in this series was de
signed for a determination of the effectiveness of
5R. W. Peel Ie et al., ANP Quar. Prog. Reps. June 10, an oil layer preceding the configuration in reducing
1956, ORNL-2106, p 269; Sept. 10, 1956, ORNL-2157,
P 278; Dec. 31, 1956, ORNL-2221, p 335. the secondary gamma-ray production within the

262
 PERIOD ENDING DECEMBER 31, 7957

Table 5.2.3. Description of Configurations Tested in GE Series

Unavoidable Oil Gap

Configuration Location of Probe,
Description of Configuration Within Configuration
Number z~ (cm)
(cm)

1-A 2.9 cm of oil + 4 in. of Be + 2 in. of Hevimet + 12 in. 55.2 3.4

of LiH -Foil

1-B 12.9 cm of oil + 4 in. of Be + 2 in. of Hevimet + 12 in. 66.0 3.2

of LiH + oil

1-C 22.9 cm of oil + 4 in. of Be + 2 in. of Hevimet + 12 in. 74.6 1.8

of LiH -Foil

1-D 32.9 cm of oil + 4 in. of Be + 2 in. of Hevimet + 12 in. 85.0 2.2

of LiH -Foil

1-E 40.6 cm of oil + 4 in. of Be + 2 in. of Hevimet + 12 in. 92.0 0.5

of LiH + oil

1-F 1cm of oil +4in. of Be +\2 in. of boral +2in. of 56.2 3.1
Hevimet + 12 in. of LiH + oil

1-G 1cm of oil +4 in. of Be + \ in. of boral + 2 in. of 58.0 3.6

Hevimet + L in. of boral + 12 in. of LiH + oil

2-A 1.0 cm of oil + 4 in. of Be + 4 in. of stainless steel 58.0 2.3

+ 12 in. of LiH + oil

2-B 11.0 cm of oil + 4 in. of Be + 4 in. of stainless steel 67.6 2.9

+ 12 in. of LiH -Foil

2-C 21.0 cm of oil + 4 in. of Be + 4 in. of stainless steel 77.2 2.5

+ 12 in. of LiH + oil

2-D 31.0 cm of oil + 4 in. of Be + 4 in. of stainless steel 87.2 2.5

+ 12 in. of LiH + oil

2-E 37.2 cm of oil + 4 in. of Be-I-4 in. of stainless steel 93.6 0.7
+ 12 in. of LiH -Foil

2-F 1.0 cm of oil 44in. of Be +'/j in. of boral +4in. of 59.8 2.9
stainless steel -I- 12 in. of LiH + oil

2-G 1.0 cm of oil +4in. of Be + \2 in. of boral +4in. of 61.0 2.8

stainless steel + L in. of boral + 12 in. of LiH + oil

2-H 9.7 cm of oil + 4 in. of Be + }-, in. of boral + 4 in. of 68.4 2.8
stainless steel + 12 in. of LiH + oil

2-1 11.2 cm of oil + 4 in. of Be + 4 in. of stainless steel 69.4 2.3

+ \ in. of boral + 12 in. of LiH +oil
2-J 9.7 cm of oil +4 in. of Be + \ in. of boral +4 in. of 69.4 2.5
stainless steel + L in. of boral + 12 in. of LiH + oil

2-K 27.8 cm of oil +4in. of Be +\ in. of boral +4in. of 84.6 0.9

stainless steel + 12 in. of LiH + oil

2-L 1.0
.0 cm
cm of
of oil
oil +
+ 4
4 in.
in. of
of Be
Be +
+ 8.7
8.7 cm
cm of
of oil + /j in. of 68.4 2.8
boral + 4 in. of stainless steel + 12 in. of LiH + oil

263
ANP PROJECT PROGRESS REPORT

Table 5.2.3 (continued)

Unavoidable Oil Gap

Configuration Location of Probe,
Description of Configuration Within Configuration
Number zQ (cm)
(cm)

2-M 1.0 cm of oil + 4 in. of Be + /j in. of boral + 8.7 cm 68.4 2.8

of oil + 4 in. of stainless steel + 12 in. of LiH + oil

3-A 4.2 cm of oil + 4 in. of Be + 1/j in. of depleted U 53.4 1.9

+ 12 in. of LiH -Foil

3-B 13.8 cm of oil + 4 in. of Be + 1/j in. of depleted U 62.8 1.7

+ 12 in. of LiH -Foil

3-C 24.2 cm of oil + 4 in. of Be + 1/j in. of depleted U 73.4 1.9

+ 12 in. of LiH + oil

3-D 33.2 cm of oil + 4 in. of Be + 1/2 in. of depleted U 82.2 1.7

+ 12 in. of LiH+ oil

3-E 40.8 cm of oil + 4 in. of Be + 1/2 in. of depleted U 89.8 1.7

+ 12 in. of LiH + oil

3-F 1.0 cm of oil +4 in. of Be + /j in. of boral + 1/j in. 51.8 1.2

of depleted U + 12 in. of LiH + oil

3-G 1.0 cm of oil + 4 in. of Be + /2 in. of boral + 1^ in. 53.8 1.9

of depleted U + /o in. of boral + 12 in. of LiH + oil

4-A 1.0 cm of oil"+ 4 in. of stainless steel + oil 15.8 2.0

4-B 11.0 cm of oil + 4 in. of stainless steel + oil 26.4 2.6

4-C 21.0 cm of oil + 4 in. of stainless steel + oil 35.2 1.4

4-D 31.0 cm of oil + 4 in. of stainless steel + oil 44.8 1.0

4-E 41.0 cm of oil + 4 in. of stainless steel + oil 56.0 2.2

4-F 51.0 cm of oil + 4 in. of stainless steel + oil 65.2 1.4

4-G 61.0 cm of oil +4 in. of stainless steel + oil 75.4 1.6

4-H 71.0 cm of oil + 4 in. of stainless steel + oil 86.2 2.4

5-A 1.0 cm of oil +4 in. of Be + 12 in. of LiH + oil 46.0 0.5

gamma-ray shield. Figure 5.2.22 shows that in layers had approximately the same effect in tests
creasing the thickness of the oil layer preceding a of similar configurations with stainless steel
beryllium—Hevimet—lithium hydride shield (con (configurations 2-A through 2-E) and with depleted
figurations 1-A through 1-E) from 2.9 to 12.9 cm uranium (configurations 3-A through 3-E), as
reduced the gamma-ray dose rate beyond the shield shown in Figs. 5.2.23 and 5.2.24.
approximately 35%. Increasing the thickness of In all the tests described above an attempt was
the oil layer another 10 cm decreased the gamma- made to increase the thickness of the oil layer
ray dose rate an additional 16%. Oil layers in 10-cm increments. The fact that the configu
greater than 22.9 cm appeared to have no further ration tank used to hold the shield components
effect. This information can be used to determine had a 1-cm recess in the tank wall adjacent to
the optimum thickness of the oil layer for the the source plate (see preceding section) made it
contemplated reactor-shield configuration. The oil impossible to obtain an oil layer smaller than

264
 PERIOD ENDING DECEMBER 31, 7957

Table 5.2.4. Physical Properties of Materials Used for LTSF Mockup Tests
of Advanced Shielding Materials (G-E Series)

Actual Thickness
Material
per Slab (cm) Description

Transformer oil Specific gravity, 0.88 at 15.6 C; nominal composition,

86.5 wt % C-12.5 wt % H

Beryllium 10.16 48 X 49 X 4-in. metallic slabs

Hevimet 6.35 5-ft X 5-ft X 2-in. slabs made from smaller pieces held
together with 5-ft X 5-ft x /.-in. Al sheet; composition,
90 wt % W-6 wt % Ni-4 wt % Cu; density of Hevimet,
16.9 to 17.2 g/cm3
Type 347 stainless 2.54 See Table 5.2.2
steel

Uranium 3.81 See Table 5.2.2

Lithium hydride 32 See Table 5.2.2

Boral 1.3 See Table 5.2.2

1 cm, as is evident in configuration 2-A. More

2-01-057-71-379 over, the design specifications set by GE for the
-10 =p
— • CONFIGURATION l-A, 2.9-cm OIL GAP :
experiment required that the centers of the gamma-
a CONFIGURATION l-B, 12.9-cm OIL GAP ray shields in configurations 1-A, 2-A, and 3-A
» CONFIGURATION l-C, 22.9-cm OIL GAP
o CONFIGURATION l-D, 32.9-cm OIL GAP be equidistant from the source plate. Since the
• CONFIGURATION l-E, 40.6-cm OIL GAP
gamma-ray shields used in configurations 1-A and
CONFIGURATION: 4 in. OF Be + 2 in. OF 3-A were thinner than the gamma-ray shield used
HEVIMET + 12 in. OF LiH + OIL
in configuration 2-A, the oil layers preceding
these configurations were necessarily thicker.
A cross plot of the data presented in Figs.
5.2.22 through 5.2.24 is shown in Fig. 5.2.25,
which presents the gamma-ray dose rate at 100 cm
from the source plate as a function of the distance
between the source plate and the center of the
gamma-ray shield for each of the three tests.
90 100 110 120 130 140
This cross plot indicates that the minimum thick
,?,,, DISTANCE FROM SOURCE PLATE (cm) ness of the oil layer preceding the beryllium for
the lowest gamma-ray dose is approximately
Fig. 5.2.22. Gamma-Ray Tissue Dose Rates Beyond 20 cm.
Configurations 1-A Through 1-E: Effect of Increasing In several configurations A-in.-thick sheets of
the Thickness of the Oil Gap Preceding Shields boral were placed adjacent to or near the gamma-
Containing Hevimet. ray shield. The purpose of these tests was to

265
ANP PROJECT PROGRESS REPORT

100 110 120 130 160

,?0,DISTANCE FROM SOURCE PLATE (cm)

Fig. 5*2.23. Gamma-Ray Tissue Dose Rates Beyond Configurations 2-A Through 2-E: Effect of Increasing the
Thickness of the Oil Gap Preceding Shields Containing Stainless Steel.

266
 PERIOD ENDING DECEMBER 31, 7957

aw**.***
2-01-057-71-382
4x 10

100 110 120 130 140 150 160

z., DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.24. Gamma-Ray Tissue Dose Rates Beyond Configurations 3-A Through 3-E: Effect of Increasing the
Thickness of the Oil Gap Preceding Shields Containing Depleted Uranium.

determine the effectiveness of boral in reducing configurations with depleted uranium, the addition
the secondary gamma-ray production in the gamma- of the boral sheet in front of the uranium reduced
ray shield. In configurations 1-F, 2-F, and 3-F, the gamma-ray dose rate 51% at a distance 70 cm
a single sheet of boral was placed immediately from the source but only 39% at a distance of
in front of the gamma-ray shield, while in configu 160 cm (Fig. 5.2.28). A second boral sheet placed
rations 1-G, 2-G, and 3-G the boral was positioned behind the gamma-ray shield again reduced the
both before and after the shield. The measure dose rate only slightly.
ments (Fig. 5.2.26) showed that placing the boral
in front of the Hevimet reduced the dose rate Several other tests were performed with vari
beyond the shield approximately 31%. Positioning ations of the configurations containing stainless
a second boral sheet behind the Hevimet produced steel, the results of which are plotted in Fig.
a further slight reduction, which may be within 5.2.29. In configurations preceded by an approxi
experimental error. Placing the boral in front of mately 9-cm-thick oil layer (configurations 2-H and
the stainless steel reduced the dose rate beyond 2-J), the addition of a second boral sheet behind
the shield approximately 47% (Fig. 5.2.27). The the stainless steel reduced the gamma-ray dose
second boral sheet behind the stainless steel rate about 10%. It is felt that no particular
again produced a further slight reduction. In significance can be attached to the 10% reduction

267
ANP PROJECT PROGRESS REPORT

10 20 30 40 50 60

DISTANCE FROM SOURCE PLATE TO CENTER OF GAMMA-RAY SHIELD (cm)

Fig. 5.2.25. Gamma-Ray Tissue Dose Rates 100 cm from the Source Plate as a Function of the Distance Be
tween the Source Plate and the Center of the Gamma-Ray Shield (Configurations 1-A Through 1-E, 2-A Through 2-E,
and 3-A Through 3-E).

268
 PERIOD ENDING DECEMBER 31, 1957

6x10 2-01-057-71-384
5
I 1 1 1 1 1 1 1 1 1 1 1 1 1 1
STAINLESS
c ONF OIL 3e BC)RAL HEVIHilET BORA L LiF
CONF OIL Be BORAL STEEL BORAL LiH
SYME OL NO. cm) ( n.) in.) (in ) (in.) (in )
.SYMBOL NO. (cm) (in.) (in.) (in.) (in.) (in.)
• 1-A 2.9 4 0 2 0 12 + 0 IL -
_• 2-A 1 4 0 4 0 12 + 0IL.
t-F 1.0 4 2 0 12 +0 L
A 2-F 1 4 V2 4 0 12 +0IL
0 l-G 1.0 4 v2 2 v2 12 +0 L _
~ o^, 2-G 1 4 V2 4 V2 12 +0IL

10
\ i

st
\ 1
"v
10
N\ •s
\
S,
\
\
I \
l\
\ s.
?\
XV
L
r
\ 1
Is

80 120 180

10
60 120 140

Fig. 5.2.26. Gamma-Ray Tissue Dose Rates Beyond

Configurations 1-A, 1-F, and 1-G: Effect of Inserting
Boral Within Configurations Containing Hevimet. Fig. 5.2.27. Gamma-Ray Tissue Dose Rates Beyond
Configurations 2-A, 2-F, and 2-G: Effect of Inserting
Boral Within Configurations Containing Stainless Steel.
in dose rate because the magnitude of experimental
errors may account for this variation.
A comparison of measurements beyond configu The results of this test are to be compared with
rations 2-F (from Fig. 5.2.27) and 2-K and 2-H measurements beyond similar configurations in
(from Fig. 5.2.29), which contain a single sheet which no boral was used (configurations 2-B and
of boral immediately in front of the stainless 2-D in Fig. 5.2.23), in which cases increasing
steel, shows that increasing the thickness of the the thickness of the oil layer preceding the
oil layer adjacent to the source plate from 1.0 to configurations from 11 to 31 cm reduced the dose
9.7 cm reduced the gamma-ray dose rate approxi rate approximately 38%.
mately 29%. Increasing the oil layer from 9.7 A comparison of configurations 2-H and 2-1,
to 27.8 cm reduced the gamma-ray dose rate only based on the assumption that the thicknesses of
10%, which, again, may be within experimental the oil layers preceding the two configurations
error. This appears to indicate that, for configu are essentially the same, shows the effect of
rations with boral in front of stainless steel, placing a boral sheet in front of the stainless
increasing the thickness of the oil layer preceding steel rather than behind it. The dose rate beyond
the configuration to more than 10 cm does not configuration 2-H is approximately 28% lower than
effectively reduce the dose rate beyond the shield. that behind configuration 2-1. A comparison of

269
ANP PROJECT PROGRESS REPORT

configurations 2-1 and 2-J shows the effect of C(JM#g*ations 2-L and 2-M were tested to
using boral both in front of the stainless steel determine whether an oil layer within the configu
and behind it, rather than just behind it. The ration is best located in front of or behind a boral
resulting decrease in the dose rate was about 35%. sheet preceding the stainless steel gamma-ray
shield. The measurements show that the dose
rates beyond the two configurations are the same,
10' 1 1 1 1 1 1 1 1
2-01-057-71-386
1 1 I
and apparently indicate that most of the neutrons
DEPLETEO which cause capture gamma rays in the steel are
CONF. OIL Be BORAL URANIUM BORAL LiH
-SYMBOL NO. (cm) (in.) (in.) (in.) (in.) (in.) thermalized in the beryllium rather than the oil.
- • 3-A 4.2 4 0 lV2 0 12 +0IL- In order to determine the effectiveness of oil
_ A 3-F 1.0 4 V2 lV2 0 12 +0IL.
o 3-G 1.0 4 V2 l'/2 V2 12 +0IL in reducing the secondary gamma-ray production
in a gamma-ray shield when no beryllium, boral,
or lithium hydride is present, a group of tests
was run in which a 4-in.-thick stainless steel
slab was moved in the oil from a position close
10 to the source plate to a position 71 cm from the
source plate. As the stainless steel was moved
away from the source, the dose rate at a point
( 100 cm from the source (see Fig. 5.2.30) varied
K
> ,
from 17 ergs/g-hr-w for a 1-cm-thick oil layer to
approximately 3.5 ergs/g-hr-w for a 31-cm-thick
\ \
\ i
oil layer, which represents an 80% reduction.
Greater thicknesses of oil did not reduce the dose
rate further.
A test was also carried out to determine the
attenuation of the gamma-ray dose rate by 4 in.
\
\ of beryllium and 12 in. of lithium hydride. A plot
\
\ \ of the results (Fig. 5.2.31) shows that the be
\ ryllium and lithium hydride reduced the dose rate
)
K '

1^ N f
about 40% at a distance 70 cm from the source
and about 10% at a distance of 160 cm.
The effect on the thermal-neutron flux of in
serting one or two /2-in.-thick boral sheets within
100 120 140 160 the configurations was also studied. From the
DISTANCE FROM SOURCE PLATE (cm)
plots in Figs. 5.2.32, 5.2.33, and 5.2.34 it appears
that placing a boral sheet both before and after
the gamma-ray shield is no more effective in
Fig. 5.2.28. Gamma-Ray Tissue Dose Rates Beyond reducing the flux than placing a single sheet in
Configurations 3-A, 3-F, and 3-G: Effect of Inserting front of the shield only.
Boral Within Configurations Containing Depleted Other configurations will be tested in this
Uranium. series.

270
 PERIOD ENDING DECEMBER 31, 1957

#*»*

Z- 01-057-71-381
3x10'
1
STAINLESS
CO MFIGURATION OIL Be OIL
SYMBOL NO. (cm) (in.) (cm) (in.) (cm) (in.) (in.) (in.)

0 2-H 9.7 4 0 V2 0 4 0 12 +• OIL

10 • • 2-1 11.2 4 0 0 0 4 V2 12 +• OIL

• 2-J 9.0 4 0 h 0 4 V2 12 t OIL

• 2-K 27.8 4 0 Vz 0 4 0 12 + OIL

A 2-L 1.0 4 8.7 V2 0 4 0 12 f OIL

A 2-M 1.0 4 0 '/2 8.7 4 0 12 f OIL

1^
10

.// W
-N-^^-i ^v^
~\-
0./
-C^
o-^^ / ^>^
l^
•y
ItS? >»»-. ^^1 1-
^ >^w ^"1 I-.

o<
^ 1^^

• A — ^ |—
Is"" -• <?s^
"--v^
o^V- ^^
*-v —*—^: *^ ^' 1-^
XT
3<^ ^j<3^^
w°iy
m/
5^
--.
*^5^ ^^Ss-.

!?P2
A<

70 90 100 110 120 130 140 150 160

z„, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.29. Gamma-Rqy Tissue Dose Rates Beyond Configurations 2-H Through 2-M: Effect of Inserting Boral
and Oil Gaps Within Configurations Containing Stainless Steel.

271
ANP PROJECT PROGRESS REPORT

3x10 1— 1 1 1 1 1
CONF. OIL Be BORAL HEVIMET BORAL LiH
SYMBOL NO. (cm) (in.) (in.) (in.) (in.) (in.)
a 1-A 2.9 4 0 2 0 12+OIL
10" -- o l-F 1.0 4 V2 2 0 12+0IL-
- • 1-G 1.0 4 V2 2 '/2 12 FOIL -

0 10 20 30 40 50 60 70
\
THICKNESS OF OIL GAP PRECEDING CONFIGURATION (cm)

^ 10
Fig. 5.2.30* Gamma-Ray Tissue Dose Rates 100 cm
from the Source Plate as a Function of the Thickness of
the Oil Gap Preceding Stainless Steel (Configurations A.

4-A Through 4-H). \

2-01-057^71-388
5X10'
\
\
: ,

\
\
\
Cf"""~

. DILOfgLYIN CON FIGUF ATIO M TAr<K 10

50 60 70 80 90 100 110 120
.?„, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.32. Thermal-Neutron Fluxes Beyond Con

figurations 1-A, l-F, and 1-G: Effect of Inserting
Boral Within Configurations Containing Hevimet.

CONFIGURATION 5-A:
1.0cm OF 0IL + 4in. OF Be
+ 12in. OF LiH-FOIL
10

40 60 80 100 120 140 160

z0 , DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.31. Gamma-Ray Tissue Dose Rates Beyond

Configuration 5-A: Effect of Oil, Beryllium, and Lithium
Hydride Components When Gamma-Ray Shield Is Omitted.

272
 PERIOD ENDING DECEMBER 31, 7957

2-01-057-71-389 2-01-057-71-390
4x10' 3x10
I I 1 1 1 1 1 1 1 1 1 1 1 1 1
STAINLES
DEPLETED
CONF OIL Be BORAL STEEL BORAL LiH
SYMBOL NO. (cm) (in.) (in.) (in.) (in.) (in.) _ CONF OIL Be BORAL URANIUM BORAL LiH
• 2-A 1 4 0 4
^SYMBOL NO. (in.) (in.) (in.) (in.) (in.) (in.) E
0 12 +0IL
A 2-F 1 4 V2 4
Z • 3-A 4.2 4 0 l'/2 0 12 + OIL —
0 12 + OIL
- o 2-G t 4 V2 4 V2 12 +OIL _ ~ A 3-F 1.0 4 V2 l'/2 0 12 + OIL-
o 3-G 1.0 4 V2 lV2 V2 12+OIL

10°
^A
/A

h \
\
\
\
\
3IL ONLY IN
:0NFIGURA HON TANK

^cJV
2^A

\ H

\
\
V -*5--
\ •v
10

T7 \ \
\
\
1
V

70 90 110 130 150

i0, DISTANCE FROM SOURCE PLATE (cm)
\
\
Fig. 5.2.34* Thermal-Neutron Fluxes Beyond Con
1 figurations 3-A, 3-F, and 3-G: Effect of Inserting
70 80 90 100 110
70. DISTANCE FROM SOURCE PLATE (cm)
Boral Within Configurations Containing Depleted
Uranium.

Fig. 5.2.33. Thermal-Neutron Fluxes Beyond Con

figurations 2-A, 2-F, and 2-G: Effect of Inserting
Boral Within Configurations Containing Stainless Steel.

273
*«*»*» «•<»**•
 5.3. TOWER SHIELDING REACTOR II
C. E.Clifford L. B. Holland

MECHANICAL DESIGN will have a 2.5-in. stroke, and the method of

The design of the core of the Tower Shielding operation will be the same as that for the pro
Reactor II (TSR-II) has been modified slightly to totype, which was described previously. Normally
provide more clearance for the thermal expansion the spring will hold the control plate against the
that will occur when the reactor is operating at fuel. Water will flow through the mechanism until
a power of 5 Mw. In earlier layouts it was the flow area around the piston is closed by the
planned to place the lead-boral shield adjacent traveling nut. The pressure will build up in the
to the aluminum pressure tank. In order to ensure cylinder, and the control plate can then follow
that the tank is not heated above a safe level
the traveling nut to control the reactor. Inter
by gamma-ray heating in the lead and neutron rupting the water flow will scram the mechanism.
heating in the boral, it will be separated from NUCLEAR CALCULATIONS4
the shield by a /-in.-thick layer of water, as The preliminary nuclear calculations5 were per
shown in Fig. 5.3.1. The lower portion of the formed for a TSR-II design in which there was a
lead-boral shield will be hemispherical, and it water reflector exterior to the core, as shown in
will be suspended from a ring on the tank wall
Fig. 5.3.3. By a modification of the 3G3R Oracle
at the horizontal midplane of the core. With this code, this reflector was represented as a boundary
arrangement there will be sufficient room for the condition, so that the internal water reflector
shield to expand downward with heating. Radial could be divided into two regions for control
expansion at the midplane will also be permitted. studies. A recent further modification of the
When the reactor is rotated about a horizontal
3G3R code provides a means of representing
axis, the shield will be restrained from moving
the several layers of the external reflector (see
by keys on the hemisphere which will engage Fig. 5.3.1) as a set of boundary conditions.
slots in the support ring. In order to minimize The two outermost regions of the spherical .re
fission gamma-ray leakage through the break in flector can be removed at one time. These two
the midplane, the lead shield will be divided so regions are assumed to be arranged in their exact
that an extension of the interface will not intersect
geometry with a critical core region inside. The
any of the fuel region. values of a are then solved for in the following
The annular fuel elements and the lower central equations:
fuel elements will be mounted on the central cyl
inder. Each fuel element will be fixed at the
(</>, + 2/,)an = 0, - 2/, ,
top and guided at the lower end to permit some
growth. These annular fuel elements will be re (<£i + 2/,) a12 + (<f>2 + 2]2) a22 = <p2 - 2/2 ,
movable remotely from above. The lower central
fuel elements will be bolted to the cylinder so where
that they cannot be removed while the annular a... = number of fast neutrons incident on the
elements are in place. This is to prevent the reflector that are returned as fast neu
possibility of lifting the lower central fuel trons,
elements into the midplane of the core region
a.. = number of fast neutrons incident on the
with the annular elements in place.
reflector that are returned as thermal
CONTROL MECHANISM neutrons,
A full-scale mockup of the control mechanism
has been designed, as shown in Fig. 5.3.2. It
C. E. Clifford and L. B. Holland, ANP Quar. Prog.
Rep. June 30, 1957, ORNL-2340, p 328.
The engineering design work for the TSR-II is being
performed by the ORNL Engineering Department. The nuclear calculations for the TSR-II are being
performed by M. E. LaVerne.
2C. E. Clifford and L. B. Holland, ANP Quar. Prog.
Rep. March 31, 1957, ORNL-2274, p 294; June 30, 1957. 5C. E. Clifford and L. B. Holland, ANP Quar. Prog.
ORNL-2340, p 323; Sept. 30, 1957, ORNL-2387, p 304. Rep. March 31, 1957, ORNL-2274, p 294.

275
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-060-37

SHIELD
SUPPORT RING

WATER

Fig. 5.3.1. Suspension System for TSR-II Fuel Elements and Lead-Boral Shield Section.

276
 PERIOD ENDING DECEMBER 31, 1957

r-i

'it
X

^-
'—•

Fig. 5.3.2. Experimental Mockup of the Control Mechanism for the TSR-II.

277
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
DWG 2-01-060-7

WATER FLOW

LOCATION OF CYLINDRICAL
TUBE IN ACTUAL REACTOR

INTERNAL WATER REFLECTOR

MOVABLE BORAL SHELL

CORE REGION (WATER PLUS

URANIUM-ALUMINUM FUEL
PLATES)

DIMENSIONS

A 16,20 OR 25cm
B 16,20 OR 25 cm
C 20 cm EXTERNAL WATER REFLECTOR

Fig. 5.3.3. Vertical Section of the Reactor Core Model Used in the Nuclear Parameter Study for the TSR-II.

a22 = number of thermal neutrons incident on The p3 terms, which are not independent of the
reflector that are returned, ratio </>./<£2, are input numbers of the 3G3R Oracle
(/>.. = fast-neutron flux, code and represent boundary conditions. Thus a
new reactor configuration can be studied which
e>, = thermal-neutron flux, has three regions and boundary conditions that
/1 = fast-neutron current, represent two additional regions that have been
/- = thermal-neutron current. removed. This can be repeated to remove several
regions.
The values of a can be obtained by running two
cases, one of which has a black shell placed This procedure was used to strip off four outer
between the core and the reflector to eliminate regions of the TSR-II and represent them as
the thermal current. The values obtained for a boundary conditions, the four regions being a
are independent of the ratio 0,/<£2. 20-cm-thick water reflector, a ^-in.-thick boral
Two additional terms can be derived, as follows: shell, the 2-in.-thick lead region, and an aluminum-
water region. With these boundary conditions, cal
1 - a
*h 11 1 culations were performed to obtain the critical
1 + a 0 51 mass and the amount of reactivity available for
11
control in a reactor having a 17.5-in.-dia internal
1 - a.
22 L12 *1 reflector and a 5.5-in.-thick core region which
were separated by a l^-in.-thick shell of aluminum
4>i 1 + a22 (1 + an)(l + a22) c>2 and water. From these calculations it was found
that a thermally black control shell which could
move awoy from the core in the internal reflector
/352 region could control 12% in k. Since the control

278
ANP PROJECT PROGRESS REPORT

Table 5.3.1. Calculated Fast-Neutron and Thermal-Neutron Flux Distributions

in a TSR-II Configuration with a Control Shell 1.6 cm from the Fuel

Distance from Canter

Along a Radius Fast-Neutron Flux Thermal-Neutron Flux
(cm)

1.15 0.1317 0.3302

2.31 0.1344 0.3357

3.46 0.1389 0.3448

4.61 0.1454 0.3575

5.77 0.1540 0.3739

6.92 0.1649 0.3939

8.08 0.1784 0.4173

9.23 0.1949 0.4436

10.38 0.2147 0.4723

11.54 0.2384 0.5020

12.69 0.2666 0.5307

13.84 0.2999 0.5554

14.10 0.3394 0.5711

16.15 0.3860 0.5705

17.30 0.4409 0.5424

18.46 0.5058 0.4705

19.61 0.5823 0.3304

20.76 0.6725 0.0865

20.76-21.40*
21.40 0.6332 0.0685

21.80 0.6661 0.1417

22.20 0.7011 0.1913

22.60 0.7381 0.2177

23.00** 0.7774 0.2209

23.64 0.8318 0.2152

24.27 0.8773 0.2140
24.91 0.9147 0.2150

25.55 0.9448 0.2171

26.18 0.9680 0.2192

26.81 0.9847 0.2210

27.45 0.9952 0.2222

28.09 1.0000 0.2224

28.73 0.9992 0.2218

29.36 0.9931 0.2202

30.00 0.9821 0.2175

30.64 0.9663 0.2140

31.27 0.9460 0.2095

31.91 0.9215 0.2041

32.55 0.8931 0.1979

33.18 0.8611 0.1911

33.82 0.8257 0.1835

34.45 0.7872 0.1755

35.10 0.7458 0.1671
35.73 0.7019 0.1585

36.36 0.6556 0.1499

37.00 0.6071 0.1417

*Position of /-in.-thick control shell

**At reflector-core interface.

280
 PERIOD ENDING DECEMBER 31, 1957

Table 5.3.2. Average Normalized Fluxes and Currents emergency shutdown that may occur while the
in TSR-II Configuration with a Control radiator is operating at 5 Mw and for cir
Shell 1.6 cm from the Fuel culating the water while the reactor is op
erating at a power below 5 kw;
Fast Thermal 3. a water-level control system which will consist
Region
Neutrons Neutrons of a 250-gal level tank, compressed gas to
remove water from the system, and a fill pump
Average Normalized Flux which will take water from the pool and put
Internal reflector, inside 0.4068 0.4505
it into the system;
control shell
4. a 40-kw heater to prevent water lines from
freezing when the reactor is not operating (the
Internal reflector, outside 0.7046 0.1759
air cooler will be supplied with louvers and
control shell
additional heaters);
Core 0.8674 0.1962 5. a mixed-bed demineralizer which will bypass
the detention tank with a flow of 40 gpm and
Net Current
which can also be used to keep the reactor
Outer core boundary 0.1003 0.0030 pool water clean;
6. a Tracerlab fission break monitor which will
give an alarm if iodine activation, which would
WATER COOLING AND PURIFICATION SYSTEM indicate a rupture in a fuel element, is noted
in the reactor outlet water.
The design and specifications for the water
system are 95% complete. The lump sum contract
for the installation is to be let the last week EXTENSION OF GAMMA-RAY-HEATING

of February 1958 and the work is scheduled to CALCULATION FOR THE TSR-II

be completed in September. Earlier estimates had

W. W. Dunn7 D. K. Trubey
indicated that the work could be completed by
July 1. In the previous computation of the gamma-ray
A flow diagram of the water system, along with heating in the TSR-II, two assumptions were made
tentative locations of the pressure- and tem concerning the angular and energy dependence of
perature-measuring elements, is shown in Fig. the gamma-ray flux at the core-shield interface.
5.3.5. The main components of this system The first assumption was that the collided energy
will be: spectrum was represented by the NDA differential
1. a 60-hp circulating pump capable of delivering energy spectrum after passage, through one mean
1000 gpm when the reactor is operating at a free path of water, and the second was that the
200-ft altitude, collided and uncollided fluxes were isotropic at
2. a venturi nozzle for measuring water flow, the core-shield interface. In order to check the
3. the reactor, which will operate up to a 5-Mw validity of these assumptions, an additional cal
heat output, culation has now been completed by using the
4. a two-cell, air-cooled heat exchanger capable Bowman-Trubey Monte Carlo data on the gamma-
of removing 17,065,000 Btu/hr from water ray penetration of lead shields in order to
entering at 160°F and leaving at 125.8°F when compute the heating in the first /. in. of lead
the ambient air temperature is 75°F, in the shield, that is, the lead adjacent to the
5. a 1500-gal detention tank. core. This first lead interval was chosen as
There will also be the following auxiliary units
of the flow system: On assignment from U.S. Air Force.
1. a main bypass valve for reduced flow when 8W. W. Dunn and D. K. Trubey, ANP Quar. Prog. Rep.
the reactor is operating in the power range Sept. 30, 1957, ORNL-2387, p 307-316, esp 312.
o
from 5 kw to 5 Mw; H. Goldstein and J. E. Wilkins, Calculations of the
2. a 5-hp d-c emergency circulating pump which Penetrations of Gamma Rays. Final Report, NYO-3075
(June 30, 1954).
will operate continuously and which is re
L. A. Bowman and D. K. Trubey, ANP Quar. Prog.
quired for removing afterheat following any Rep. Sept. 30, 1957, ORNL-2387, p 320.

281
ANP PROJECT PROGRESS REPORT

VARIABLE PITCH
CONTROL

VALVE SYMBOLS

NORMALLY CLOSED

tX] NORMALLY OPEN

•<! FLOWCONTROL

f\l CHECK

INSTRUMENT SYMBOLS

F - FLOW M - MOTOR

P - PRESSURE R - RECORDER

T - TEMPERATURE L - LEVEL

X - SWITCH d — DIFFERENTIAL
NOTE:
ALL ANNUNCIATOR FIELD CONTACTS OPEN ON TROUBLE.
- CENTRAL PANEL
- LOCAL INDICATOR
INDICATOR

Fig. 5.3.5. Cooling Water Flowsheet for the TSR-II.

282
 PERIOD ENDING DECEMBER 31, 7957

the check point, since a large part of the gamma- UNCLASSIFIED

2-01-060-32
ray heating occurs within it, and, in addition,
it is in the interval in which the Bowman-Trubey
results could be most accurately and readily
utilized.

The Bowman-Trubey results yielded the gamma-

ray heating, in percentage of the initial energy,
E., as a function of the angle of incidence, 6,
on the lead after passage through a varying number
of mean free paths of water. Data were available
for initial energy groups of 1, 3, 6, and 10 Mev
and for incident angles of 0, 60, 70.5, and 75.5
deg. For this calculation the source gamma rays
were considered to be isotropic in the source
planes, and the core source-shell regions and
the source gamma-ray energies were grouped
exactly as was done for the previous calculation.

The Bowman-Trubey results were first cross-

plotted to produce plots of g(6), the percentage
of initial energy EQ deposited in the first /^ in.
of lead, as a function of the angle of incidence
of the gamma rays. This was done for various
mean free paths of water preceding the lead.
A typical plot is shown in Fig. 5.3.6. It is to
be noted that the resulting curves have shapes Fig. 5.3.6. Per Cent of Energy Absorbed in Pb. as
identical to those produced in the previous cal a Function of the Incident Angle for Various Source
culation when f(6), the percentage of initial Planes; E = 4 Mev.
energy deposited, was plotted for various depths
of penetration into the shield as a function of
incident angle. It was possible to compare f(6)
and g(d) at 0.05 mean free path of material pen r = distance from center of core to
etrated, and that comparison is also shown in spherical source shell, cm,
Fig. 5.3.6. Since the minimum energy of the rQ = distance from center of core (or
Bowman-Trubey data was 1 Mev, the Monte Carlo center of internal water reflector)
code was used to produce data at 0.5 Mev for to core-shield interface, cm,
use in this calculation.
S(EQ, shell) = source term for each shell, number
The normalized gamma-ray flux, A(EQ, shell), of gamma rays of energy E. gen
for each E- from each source shell was found erated per square centimeter per
from the following expression, which incorporates fission per second,
the usual transformation from a spherical-shell
= thermal-neutron flux in the source
source to an equivalent infinite plane source
shell at point r,
(neglecting the back plane):
= macroscopic fission cross section
of U235,
1
(1) — S(EQ, shell) <f>(r) 1-t t = shell thickness, cm.
With g(6) and A(EQI shell) for each value of EQ,
77/2 and 6 in each source shell, the gamma-ray heating
f A(EQ, shell) sin ddd , in the first k-in. lead interval was found by
8

283
ANP PROJECT PROGRESS REPORT

summing over EQ and the source shell the nu buildup factor's at the core-shield interface were
merically integrated expression: assumed to be in such a direction as to penetrate
the shield, and this would not be entirely true
(2) H(EQ) in the actual case, especially at low energies.
Other small effects which would contribute to
g(6)(E0, shell) A(EQ, shell) sin d cos Odd , an overestimate were: the assumption of nearly
-/. isotropic distribution for the collided fluxes,
where the cos 6 is required for the flux-to-current which would overestimate the heating in the first
conversion. Table 5.3.3 presents the result of the lead interval and underestimate the heating in
summation of Eq. 2 compared with the results the succeeding layers if the collided flux were
obtained in steps 1 and 2 of the previous cal more peaked in the forward direction, as is
culations. The heating estimated by step 1 probably the case; and the assumption of the
was determined by multiplying the heating at the differential energy spectrum for water to represent
core-shield interface, as a function of initial the core, which would overestimate the low-energy
energy group, by the ratio of the lead to core group, since the effect of the aluminum and
energy absorption coefficients for that energy uranium would be to reduce the low-energy
group.
spectrum.
Table 5.3.3. Heating in the First 4-in.-Thick Layer
of Lead in the TSR-II Shield Calculated The present calculation by necessity required
by Various Methods considerable interpolation to perform the necessary
cross-plotting, and for this reason there is an
Calculation Heatiing (watts/cm >Mw) uncertainity in the results. The neglect of the
source gamma rays in the source planes directed
Previous calculation away from the shield is probably the one source
of error which lowers the calculated result. While
Step 1 0.548
this calculation avoids many of the errors inherent
Step 2 0.445 in step 2 of the previous calculation, the use of
the transformation from a spherical shell source
Present calculation 0.215
to an infinite plane source does introduce some
error, since the medium is not infinite and homo
geneous.
It may be seen that the result from step 1 of
the previous calculation represents the extreme In conclusion, it can be stated that the original
maximum limiting case, since it was assumed assumptions on energy spectra and angular de
that the flux was the same in the J^-in. 'eac' pendence in the previous calculation (step 2)
interval as in the core. It is expected that step 2 give a reasonably conservative estimate of the
of the previous calculation also gave an over heating in the first lead interval. Since the heating
estimate of the true heating for several reasons. in this interval is most dependent on the low-
First, the NDA buildup factors assumed an infinite energy spectrum, it would be expected that the
medium and did not take into account the reduction overestimation caused by the energy-spectra as
in the low-energy fluxes at the core-shield inter sumption would be decreased in succeeding lead
face, which in turn would reduce the heating. layers, and hence there is the prospect of im
This total flux reduction, which is not reflected provement with penetration into the shield. The
in the tabulated results, is of the order of 20% angular-distribution assumption gives an over-
for the 1.0-Mev fluxes and would be much estimation in the initial layer, and, if the collided
greater for the lower energies and less for the flux is more peaked in the forward direction, it
higher energies. Second, the collided fluxes tends to underestimate the heating deep in the
which were obtained from the point kernel and shield. No further refinements of these cal
culations are contemplated, since the heating is
Unpublished results of the Bowman-Trubey calcu considered to be sufficiently defined for purposes
lation. of construction.

284
 PERIOD ENDING DECEMBER 31. 7957

5.4. TOWER SHIELDING FACILITY

C. E. Clifford

STUDY OF GAMMA RAYS PRODUCED BY As shown in Fig. 5.4.1, the experimental con
NEUTRON INTERACTIONS IN AIR figuration included an 8-in.-dia by 10-ft-long air-
V. R. Cain F. J. Muckenthaler filled collimator attached to one face of the
reactor so that a beam of radiation was emitted
During the past few years a large number of
into the air at an angle of 60 deg from the hori
measurements have been made of gamma-ray dose
zontal. The axis of the collimator passed through
rates inside crew shields for several shielding
the center of the face of the reactor. A partially
configurations. Dose-rate measurements of this
shielded sodium iodide crystal (3 by 3 in.) mounted
type, however, give very little information as to
on a DuMont 6363 phototube was located approxi
the original energy of the gamma rays entering
mately 1 ft above ground level at a horizontal
the crew compartment and still less as to their
distance of 52 ft from the reactor. The crystal
origin. Gamma-ray dose rate measurements in
was enclosed in a cavity surrounded by an 8-in.
the crew shields have indicated an intensity
thickness of lead and at least 4 ft of concrete.
greater than that expected from the scattering of
On the side of the detector shield facing the re
gamma rays alone. Therefore, consideration has
actor, a 4-in.-wide collimator slit extended through
been given to the possibility of secondary gamma-
the shield to the detector position. The floor of
ray production from neutrons leaving the reactor
the slit slanted 10 deg upward from the detector
shield. Interactions of neutrons in air include
position so that no direct radiation from the
inelastic scattering, as well as elastic scattering,
ground would be admitted. The crystal and photo
and thermal-neutron capture. Calculations of tube were mounted so that the axis of the photo
gamma-ray intensities from inelastically scattered tube was at an angle of 45 deg from the horizontal
neutrons are somewhat difficult because neither to ensure maximum sensitivity.
the scattering cross section nor the energy dis
tribution of gamma rays for this interaction is well The, collimator slit, which was 5 ft long and 5 ft
known. The fact that the neutron energy spectrum deep at the deepest point (at the detector position),
leaving a shield is difficult to measure further was filled with borated water contained in a 4-in.
complicates the problem. Measurements for de by 4-ft by 4-ft plastic bag. The borated water pre
termining the inelastic-scattering cross section vented neutrons from reaching the detector and
for neutrons in air would be very difficult to make also reduced the number of capture gamma rays
at the TSF. An experiment was performed, how produced in the concrete walls of the collimator.
ever, in which gamma-ray spectral measurements In addition, a /,.-in.-thick sheet of boron-impreg
were made from which it is hoped that the order of nated Plexiglas (hereafter referred to as boron
magnitude of the dose rate to be expected from Plexiglas) was placed over the outside of the
inelastic scattering can be deduced. collimator slit. (The boron Plexiglas used in this
experiment is Plexiglas impregnated with boron
carbide to a density of 0.55 g of boron per cubic
Experimental Configuration centimeter.) The effectiveness of this neutron
In order to detect the high-energy gamma rays shielding was verified by a series of measurements
resulting from neutron interactions with the air, with a BF, chamber and a fast-neutron dosimeter
it was necessary to eliminate, where possible, placed in the detector cavity. The measurements
other sources of high-energy gamma rays. Since indicated that there was no detectable neutron
the number of gamma rays born in the reactor was flux in the cavity.
many orders of magnitude larger than the number The reactor and its collimator could be rotated
born in the air, the reactor had to be well shielded about the vertical axis of the reactor. The
from the detector. This was accomplished by measure of this rotation is indicated by the
submerging the reactor in the handling pool 8 ft angle 6, which is the angle between a vertical
below the surface of the ground and thus inter plane including the reactor center and the detector
posing approximately 40 ft of earth as shielding. center and a vertical plane including the reactor

285
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-056-22-D-594

DETECTOR

-8-in-THICK LEAD SHIELD AROUND CRYSTAL

Fig. 5.4.1. Configuration for the Measurement of the Spectra of Gamma Rays Resulting from Neutron Interactions
in Air.

center and the axis of the reactor collimator. How with this beam angle, additional concrete shielding
ever, most of the measurements were taken with was placed above and in front of the detector
0 = 0 deg, and for this geometry the intersection collimator to reduce the number of primary gamma
of the detector lines of sight with the cone of rays scattering from the reactor beam that would
radiation from the reactor is depicted in Fig. 5.4.2. enter the detector collimator and be scattered into
It can be seen that the detector solid angle easily the detector.
includes all the radiation cone near the ground
Determination of Detector Response
level, but the fraction of the cone included in the
solid angle decreases as the distance from the As mentioned above, the detector consisted of
reactor is increased. a 3- by 3-in. sodium iodide crystal mounted on a
Other than the measurements with d = 0 deg, DuMont 6363 phototube. Pulses from the tube
measurements were actually made for only one were sorted by two 20-channel analyzers connected
other value of 6, that is, 6 = 90 deg, 6 being in series, and the pulse-height output was cali
measured in the clockwise direction from the brated with respect to energy against the 1.12-Mev
•7 65
reactor-detector axis. For the measurement taken Zn gamma ray.

286
 PERIOD ENDING DECEMBER 31, 7957

UNCLASSIFIED
UNCLASSIFIED
2-01-056-22-D-59B
2-01-056-22-597

PULSE HEIGHT
200 400 600 800 1000 1200

-CONE OF RADIATION FROM

REACTOR COLLIMATOR

FIELD OF VISION
OF DETECTOR

Fig. 5.4.2. Geometric Relationships When 9 = 0 deg:

TSF Experiment To Measure the Spectra of Gamma Rays
Resulting from Neutron Interactions in Air. 200 300 400 500 600
PULSE HEIGHT

In order to better understand the effect that the

borated water in the collimator might have on the Fig. 5.4.3. Response of Detector to 4.4-Mev C.12*
gamma-ray spectrum, a polonium-beryllium source Gamma Rays. Top curve represents response of Nal
was placed outside the collimator and the pulse- crystal to gamma rays produced at Van de Graaff ac
height spectrum of the 4.4-Mev gamma ray from the celerator. Lower curve represents response to Po-Be
Be9(a,ra)C12 reaction was observed. The re gamma rays filtered through a B-H«0-filled collimator.
sulting spectrum was compared, as shown in
Fig. 5.4.3, with the spectral response of this
type of crystal to 4.4-Mev gamma rays produced at Experimental Results and Discussion
the Van de Graaff accelerator with the N' 5(jb,a)C' 2 The top curve in Fig. 5.4.4 represents the total
reaction when no collimator was used. It can be gamma-ray pulse-height spectrum resulting from
seen that the borated water caused a significant neutron interactions in air measured in the geometry
reduction in the magnitude of the peak correspond described above when 6=0 deg. This spectrum
ing to the 4.4-Mev gamma ray and that the Compton presumably consists of gamma rays resulting from
scattering was increased substantially. the capture of thermal neutrons in the nitrogen of
the air plus those from other sources. The highest
P. R. Bell, private communication. energy gamma ray from thermal-neutron capture by

287
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-056-22-595

20.00 "r i i i i i i i i i i
1
1

A
• BORON CARBIDE ON REACTOR COLLIMATOR
AND DETECTOR COLLIMATOR. 8 = 90°
•
A
1000 * NO BORON CARBIDE ON REACTOR COLLIMATOR;
• BORON CARBIDE ON DETECTOR COLLIMATO r. y - 0 —
1 •
*
• • BORON CARBIDE ON REACTOR COLLIMATOR
'
AND DETECTOR COLLIMATOR. A = C
5.00 • 1 1

TT
• DIFFERENCE (* MINUS •)
• A
1
**, t BORON CARBIDE ON REACTuk uulliivimiuh nnu
i

'
•
DETECTOR COLLIMATOR; 6-in. PARAFFIN ON
I • *
2.00 Ut 1 t 1. 0
•<
1

• •

1.00 --4
*r~r T. •V
'"• V
A

T
"• •±.
0.500 "•'• 1 A
•
•w
1 ' T
1 1 •I

i •
• A -
•l i
•
-.. • ** * ♦v
0.200 ft*
' \ Mi <0.7 Mev
** «**
'

i
• '1
1 •".
".• 'TT
•

1 -*—
O.tOO _
»
-••-
*T
• a
• -
-f*r • ' T 1
0.050 i
l» \
-I
'..' *
.'..
1

•
*
•
0.020
i A
•
* -
,'• 1
*
• T

0.010 h-M- ,
—•-
T^ t-
• 4
•

I T

0.005 •V ". I I
1
It"' A
•
l
"">'.
1
»'/ I

V
•

<
•••• •

T
T
• T •• .

1 •

0.002
•' A
• • T

T 1
• I I
•
%—
0.001
1
1 *A—
f
•

1 •
1
0.0005 1
•

• •

0.0002 •

fr
0.0001
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
PULSE HEIGHT

Fig. 5.4.4. Total Pulse-Height Spectra of Gamma Rays Resulting from Neutron Interactions in Air.

288
 PERIOD ENDING DECEMBER 31, 7957

N which is known to have been observed2 is at 5.4.4, shows very little fine structure; however,
10.8 Mev, and this can be identified in the curve the energy-level distribution in N is such that
at a pulse-height setting of 1000. The two peaks it is possible for gamma rays of the order of 15
that occur at pulse-height settings of 960 and 920 different energies ranging from 3.5 to 10.8 Mev to
are typical of high-energy gamma-ray responses in be emitted, and these would be smeared by scat
a 3- by 3-in. sodium iodide crystal. The higher tering in air and in the borated water and by inter
peaks that correspond to pulse-height settings of actions in the crystal itself. Several energy levels
900 and 860 are caused by improperly operating from 5.2 to 5.6 Mev are responsible for about
channels in the analyzer. Some of the high count one-fourth of the gamma rays from thermal-neutron
ing rates below a setting of 150 were due to the capture in nitrogen, and this difference curve does
detection of Zn gamma ray.s from portions of the seem to indicate a prominent peak at about 4.5 to
calibrating source which had been left in the 5 Mev. Determination of the actual capture-gamma-
borated water of the collimator; low-energy ray spectrum, however, can only be done by what
Compton-scattered gamma rays also contributed. is known as the "subtraction" method, and an
Unfortunately, it was not possible in the ex analysis of this type is yet to be performed. The
periment to isolate the gamma rays contributed .by peak observed in this difference curve at a pulse-
thermal-neutron captures from those contributed by height setting of about 100 is the result of a shift
inelastically scattered fast neutrons. It was in gain during the foreground and background runs
possible, however, to reduce the effect of the which brought about improper subtraction of the
capture-gamma-ray contribution by blacking out Zn gamma rays. (While the reactor collimator
most of the thermal neutrons so that a difference was covered with boron Plexiglas, one series of
spectrum could be obtained which could then be measurements was made with the reactor collimator

used to determine the approximate contribution rotated to 6 = 90 deg. The results of these
from inelastically scattered neutrons. In order to measurements are included in Fig. 5.4.4, but no
obtain measurements with a reduced capture- attempt has yet been made to interpret the data.)
gamma-ray contribution, the upper end of the If it is assumed that the 10.8-Mev peak obtained
reactor collimator was covered with boron Plexiglas. in the measurements in which boron Plexiglas was
The resulting gamma-ray spectrum is shown in the used on the reactor collimator is due only to
third curve from the top in Fig. 5.4.4. It can be thermal-neutron capture in nitrogen, this peak can
seen that the 10.8-Mev capture-gamma-ray peak was be normalized with the same peak in the capture-
reduced in intensity by a factor of 40. However, gamma-ray difference curve. The difference be
the portion of the curve that corresponds to lower tween these two curves, shown in the lower curve
pulse-height settings was not reduced by the same in Fig. 5.4.5, was thought to be due to gamma rays
factor, which indicates that this portion of the resulting from inelastically scattered fast neutrons.
spectrum not only represented gamma rays which The large number of high-energy gamma rays
originated from thermal-neutron capture but also (>2.25 Mev), however, indicated the possibility
represented gamma rays from another source. This that a significant contribution to this inelastic-
other source is presumably inelastically scattered scattering spectrum might be the result of neutron
fast neutrons that were unaffected by the boron interactions in the concrete walls of the collimator.
Plexiglas on the reactor collimator. If this curve (For example, silicon capture gamma rays have
is subtracted from the top curve in the figure, very high energies.) In order to resolve this
which represents the total capture-gamma-ray con question, additional measurements were made in
tribution plus the contribution from inelastically which the intensities of the neutrons and gamma
scattered neutrons, the difference curve would rays entering the collimator slit were changed
represent the spectrum of capture gamma rays from through the use of lead and several thicknesses of
low-energy neutrons which were eliminated by the paraffin placed over and in front of the slit.
boron Plexiglas on the reactor collimator. This Data obtained when 4k in. of paraffin, 9 in. of
difference curve, which is also plotted in Fig. paraffin, and/or 4 in. of lead was placed over the
collimator slit and boron Plexiglas covered the
G. A. Bartholomew and P. J. Campion, Can. J.
reactor collimator are given in Fig. 5.4.6. For
Phys. 35, 1347 (1957). these measurements the gain of the amplifier was

289
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-056-22-596

20
.

d o

10
0
6
o

•o
(
o

• 0
v^v-
o
. 1 DIFFERENCE CUKVt:
)
NO BORON PLEXIGLAS ON REACTOR
•
0 °oo« c OLLIMATOR MINUS BORON PLEXIGLA s
•
Orffcf,<*>' ON REACTOR COLLIMATOR
•

'•••
* *• o3>
• o
0.5 —«v o <>
i » on]
•- _3a7
< V*
•
• ^ »
o 'cPno<
0.2 "\i o

•
V, "oooo* >o<* 0

fV. o

0.1
•-.

/
0.05
' 1

I MEL *STI c-sc: a t t ERI NG 3AM MA RAYS y •

o

• V
0.02
o
•
o

•
0.01 • f1
>

0.005 * 0
•

•
o

0.002

>

0.001

0.0005
1

0.0002

0.0001
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
PULSE HEIGHT

Fig. 5.4.5. Pulse-Height Spectra of Capture and Inelastic-Scattering Gamma Rays Resulting from Neutron
Interactions in Air*

290
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-056-22-670

10 1 1

MO SHIELD (BORATED WATER IN COLLIMATOR SLIT )

8
A 4V2 in. OF PARAFFIN COVERING SLIT
6
• 9 in. OF PARAFFIN COVERING SLIT
D 4 in. OF LEAD COVERING SLIT
A 4 in. OF LEAD + 4 V2 in. OF PARAFFIN COVERING SL
0 4 in. OF LEAD + 9 in. OF PARAFFIN COVERING SLIT
<1

X.5 A
BORON PLEXIGLAS ON REACTOR COLLIMATOR

• •
• 11 A
<• •
•
•
A •
• A '
A
3
•• \ A. """..
•
* • 1
' . " ^r A A A 1
• •
•
D • • • ' «
D
[p ' * * • A' > I1
D > A •
• A
D 0n ° t •

DDa ° t
u> io
1 ? D ° n D ^ D° ° B J1 A
c 1
=> 8 /* •
LI £
*. n •
I
A A i • R

["
A
• A

• *1 A .
O ° AA
0
0 C o i
i

'***?<!> ° A
A
A
A
A
* A .
1
A

o
0 O ° A •
Ji • D
0
?
0 • AA - •
10" A
• .
8
A

6 a

VJ

A T
A

3
10"
50 100 150 200 250 300 350 400

PULSE HEIGHT SETTING

Fig. 5.4.6. Pulse-Height Spectra of Gamma Rays Resulting from Neutron Interactions in Air: Effect of Various
Shields Covering the Collimator Slit {6 = 0 deg).

291
ANP PROJECT PROGRESS REPORT

reduced so that a pulse-height setting of 100 for which the gamma-ray spectra given above were
corresponded to about 2.25 Mev. If a relaxation measured. It should be emphasized that these
length of about 30 cm for gamma rays in paraffin measurements give a count rate that is proportional
is 'assumed, the reduction in the gamma-ray in- to the neutron density. The calibration procedure
fensity was that to be expected for the two paraffin used essentially multiplies the neutron density by
thicknesses (a factor of 1.8 for 4/2 in. of paraffin) the thermal-neutron velocity (2200 m/sec), and
if the gamma rays were coming frpm outside the the result is the "effective" thermal-neutron flux.
slit. A comparable reduction was found when This has only to be multiplied by the capture
6 in. of paraffin covered the collimator slit (see cross section at 2200 m/sec to obtain the capture
Fig. 5.4.4). Had these gamma rays resulted from rate in a \/v absorber.
neutron interactions in the detector collimator For the thermal-neutron flux measurements the
slit, however, the gamma-ray intensity would have counter was suspended at a fixed point in air and
been reduced much more (about a factor of 7 when the reactor and its collimator were rotated about
4/2 in. of paraffin was added). the reactor vertical centerline, as shown in Fig.
When 4 in. of lead covered the slit, the counting 5.4.7. The counter position fell along the center-
rate decreased only a factor of 5 below 4 or 5 Mev line of the beam when the beam was in the vertical
and only a factor of 2 above this energy. This
large contribution above 5 Mev appeared to be due UNCLASSIFIED

to neutron captures in the lead. This was verified 2-01-056-22-0-678

by placing 4/2 in. of paraffin on the lead, which DETECTOR

REACTOR AXIS

further reduced the counting rate a factor of about AXIS OF

10. A 9-in. thickness of paraffin was also placed COLLIMATED BEAM

on the lead, as shown in Fig. 5.4.4, but the

e = 0 deg
additional 4k in. did not effect a further reduction
in the gamma-ray intensity, since the first 4k in.
was sufficient to attenuate the neutrons. It should
be pointed out that the measurements in which
both lead and paraffin were used were very negr
background.
Since the gamma rays at the detector could be
greatly reduced by using lead and paraffin but
were not reduced by the paraffin alone more than
was expected for gamma rays formed outside the
collimator, it was concluded that these gamma
Fig. 5.4.7. Geometry for Measurement of the Thermal-
rays were the result of interactions in air and not
Neutron Flux in Air.
from thermal-neutron capture in the detector
collimator. It is assumed here that the source
was almost entirely inelastically scattered neu plane defined by 6 = 0 deg. Measurements were
trons; however, contributions from other sources, taken both with and without a boron Plexiglas
such as multiply scattered gamma rays, may be cover on the reactor collimator. The resulting
included. Theoretical calculations of probable effective thermal-neutron fluxes, with and without
gamma-ray spectra from inelastic scattering of the collimator cover, are plotted as a function of
neutrons in N , thermal-neutron capture, and cp, the angle between the axis of the beam and a
scattered primary gamma rays are being made. line through the center of the reactor and detector,
in Figs. 5.4.8 and 5.4.9 for several separation
Thermal-Neutron Flux Measurements distances, r, measured from the reactor center to
Thermal-neutron flux measurements in the air the detector center. The same information is
around the reactor were made with a bare single- given as a function of r for cp = 0, 15, 30, and
barreled BF. counter to facilitate calculations of 60 deg in Figs. 5.4.10, 5.4.11, 5.4.12, and 5.4.13,
the air capture gamma-ray intensities; the calcu respectively. The curves of Fig. 5.4.9 have been
lations will be made for the same configurations fitted with an expression of the form e~r' /r , and

292
 PERIOD ENDING DECEMBER 31, 7957

UNCLASSIFIED
2-01-056-22-671

60 50 40 30 20 10 0

4>, ANGLE BETWEEN BEAM CENTERLINE AND REACTOR-DETECTOR CENTERLINE (deg)

Fig. 5.4.8. Effective Thermal-Neutron Flux at Various Distances r as a Function of the Angle cp: No Boron
Plexiglas on Reactor Collimator.

293
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2 — 01 — 056-22-672

X
3

o
or
H

UJ

or
LU
X

o
LU

60 50 40 30 20 10 0
4>, ANGLE BETWEEN BEAM CENTERLINE AND REACTOR-DETECTOR CENTERLINE (deg)

Fig. 5.4.9. Effective Thermal-Neutron Flux at Various Distances r as a Function of the Angle cp: Boron
Plexiglas on Reactor Collimator.

294
 PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED UNCLASSIFIED
2-01-056-22-673 2-01-056-22-674
10° 10
— — -

0 WITHOUT COVER
.o WITHOUT COVER
A WITH COVER
• WITH COVER
10

o 2
-^e"r/4, (X=54ft) cr
w
10 3 10-'
z
O
a.
kS 5
or
LU

£ 2 —

UJ

UJ 1 .-r/\ ,\ p 10-2
o — — —
5J-W^
UJ
U- c zzzzi:
LlJ u. 5
> UJ

i-
o
UJ o
u. '- 2
••
u_
UJ _
10 io-3
0 30 60 90 100 150 180 210

r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Fig. 5.4.11. Effective Thermal-Neutron Flux as a

10 '
Function of the Distance r for cp = 15 deg: With and
0 20 40 60 80 100 120 140 Without Boron Plexiglas on Reactor Collimator.
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

UNCLASSIFIED
2-01-056-22-675
Fig. 5.4.10. Effective Thermal-Neutron Flux as a 10
- "..
Function of the Distance r for cp = 0 deg: With and - —

- —

Without Boron Plexiglas on Reactor Collimator.

0 WITHOUT COVER
relaxation lengths of 54 and 96 ft, respectively, a WITH COVER

were obtained without and with the cover. These — — — - - -

expressions are probably not valid for distances

greater than 120 ft (for which measurements were
not obtained), because a larger fraction of the
neutrons that contribute to the counts started from 10-

the reactor with energies higher than thermal. For — -

f^.
this same reason, the curves of Figs. 5.4.11
through 5.4.13 cannot be fitted with a simple
expression. 10

In addition to the measurements described in the

preceding paragraph, a series of measurements
was also made along short vertical lines at fixed 10 '
0 30 60 90 120 150 180 210
horizontal distances from the reactor as the reactor r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)
and its collimator were rotated. These vertical
lines were centered above and below the beam Fig. 5.4.12. Effective Thermal-Neutron Flux as a
axis defined by 6 = 0 deg and served to show that Function of the Distance r for cp = 30 deg: With and
the neutron beam was symmetrical about its axis. Without Boron Plexiglas on Reactor Collimator.

295
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED for which a boron Plexiglas cover was used to

2 -01-056-22-676
those for which no cover was used. This ratio is
plotted in Fig. 5.4.14 as a function of distance
\ o Wl THOUT C OVER from the reactor for angles of cp = 0, 15, 30, and
A W TH COVE R 60 deg. The higher energy neutrons become more
-1
important as the ratio decreases.

UNCLASSIFIED
2-01-056-22-677
^s^ > 100

2
uj 10

-3
10
0 30 60 90 120 150 180 210
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Fig. 5.4.13. Effective Thermal-Neutron Flux as a

Function of the Distance r for cp = 60 deg: With and
Without Boron Plexiglas on Reactor Collimator.

It also indicates that the ground did not perturb

the flux appreciably in the air volume seen by the 0 30 60 90 120 150 180 210

gamma-ray spectrometer. a, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

A measure of the relative importance of thermal

neutrons from the source and higher energy neu Fig. 5.4.14. Ratio of the Effective Thermal-Neutron
trons from the source which have been slowed Fluxes With and Without Boron Plexiglas on the Reactor
down in air may be obtained by taking, for a given Collimator as a Function of the Distance r for Various
position in space, the ratio of the measurements Angles cp.

296
 ORNL-2440
C-84 - Reactors-Special Features of Aircraft Reactors

INTERNAL DISTRIBUTION

1. R. GMffel J. T. H*we
2. J. W. A\jen W. H./ordan
3. C. J. BarVon G. yKeilholtz
4. M. Bender\ *. Keim
5. D. S. Billing\n L. Keller
6. F. F. BlankensVp T. Kelley
7. E. P. Blizard \ Kertesz
8. C. J. Borkowski \ J. J. Keyes
9. W. F. Boudreau 54^5. J. A. Lane
10. G. E. Boyd R. B. Lindauer
11. M. A. Bredig R. S. Livingston
12. E. J. Breeding R. N. Lyon
13. W. E. Browning H. G. MacPherson
14. F. R. Bruce R. E.MacPherson
15. A. D. Callihan F. C. Maienschein
16. D. W. Cardwell W. D. Manly
17. C. E. Center (K-25) E. R. Mann
18. R. A. Charpie L. A. Mann
19. R. L. Clark W. B. McDonald
20. C. E. Clifford J. R. McNally
21. J. H. Coobs F. R. McQuilkin
22. W. B. Cottrell R. V. Meghreblian
23. R. S. Crouse R. P. Milford
24. F. L. Culler A.J.Miller
25. D. R. Cuneo R. E. Moore
26. J. H. DeVan J. G. Morgan
27. L. M. Doney K. Z. Morgan
28. D. A. Douglas E. J. Murphy
29. W. K. Eister , J. P. Murray (Y-12)
30. L. B. Emlet (K-25/ M. L. Nelson
31. D. E. Ferguson/ G. J. Nessle
32. A. P. Fraas / L. G. Overholser
33. J. H. Frye / P. Patriarca
34. W. T. Furgojfson S. K. Penny
35. R. J. Gray/ A. M. Perry
36. A. T. Grafeky D. Phillips
37. W. R. G/mes J. C. Pigg
38. A. G. Brindell P. M. Reyling
39. E. Gu/h E. Richt
40. C. Sr Harrill T. Robinson
41. M.A Hill W. Savage
42. a! E. Hoffman \H. Savolainen
43„ •m. H. Hoffman . Schultheiss
44. A. Hollaender SWt
45. A. S. Householder L.\cott

297
 92. \ D. Shipley 105. A. M. Weinberg
93. A\Simon 106. J. C. Whfc
94. 0. \isman 107. G. D. w/tman *
95. J. Sits 108. E. P. Wigner (consultant)
96. M. J.\<inner 109. G. C.JrVilliams
97. A. H. Sntell 110. J. C/Wilson
98. C. D. Sum no 111. C. Jf. Winters
99. J. A. SwaiVjut 112. W.iEobel
100. E. H. Tayl\ 113--115. ORNL —Y-12 Technical Library,
101. R. E. ThomaV document Reference Section
102. D. B. Trauger* 116--122. ^Laboratory Records Department
103. D. K. Trubey 123J r Laboratory Records, ORNL R.C.
104. G. M. Watson 124--126/ Central Research Library

EXJERNAL DISTRIBUTION
Air Force ^Ballistic Missile D/vision
AFPR, BoeVg, Seattle
AFPR, BoeirY Wichita
AFPR, CurtislWright, Cli/ton
AFPR, Douglas\l_ong Beftch
AFPR, Douglas, \inta /Jbnica
AFPR, Lockheed,\urhpnk
AFPR, Lockheed, MWetta
AFPR, North Americife, Canoga Park
AFPR, North AmeriAinVDowney
Air Force Special Weapons Center
Air Materiel Comnfcnd
Air Research anc#Develop\ent Command (RDGN)
Air Research ari Development Command (RDTAPS)
Air Research aJd Development Command (RDZPSP)
Air TechnicalJIntelligence CeVter
ANP ProjectiOffice, Convair, r\prt Worth
Albuquerque/Operations Office
Argonne National Laboratory
Armed Fodces Special Weapons Prelect, Sandia
Armed Fdfces Special Weapons ProjVct, Washington
Assista/t Secretary of the Air Force,wR&D
Atomic/Energy Commission, Washingtc
Atomi/s International
Batteflle Memorial Institute
Betrfs Plant (WAPD)
Bujeau of Aeronautics
B/'eau of Aeronautics General Representati\
\R, Aerojet-General, Azusa
lAR, Convair, San Diego
JAR, Glenn L. Martin, Baltimore
JAR, Grumann Aircraft, Bethpage
Bureau of Yards and Docks
Chicago Operations Office

298
Chicago Patent Group
Curtiss-Wright Corporation
Engineer Research and Development Laboratories
^General Electric Company (ANPD)
general Nuclear Engineering Corporation
krtford Area Office
Iddfco Operations Office
Kno\s Atomic Power Laboratory
LockYind Area Office
Los A^mos Scientific Laboratory
Marquar\t Aircraft Company
Martin CoVpany
National AjUMSory Committee/or Aeronautics, Cleveland
National Ad\sory Committe/for Aeronautics, Washington
Naval Air Development Cemer
Naval Air MateXal Center
Naval Air Turbinte Test Station
Naval Research L\bora|ory
New York Operation Qrffice
Nuclear Developmem^orporation of America
Nuclear Metals, Inc-
Office of Naval Re/ear\h
Office of the Chie/of NcWal Operations (OP-361)
Patent Branch, rcishingtoa
Pratt and Whitrjey Aircraftdivision
San FranciscafOperations 0\ice
Sandia Coronation
School of 4n/iation Medicine
Sylvania^Corning Nuclear Corporation
Technical Research Group
USAF/leadquarters
USA/Project RAND
U.S/Naval Radiological Defense Laatratory
fiversity of California Radiation LaB»ratory, Livermore'
i/right Air Development Center (WCOSI-1
Technical Information Service Extension\Oak Ridge'
Division of Research and Development, AlEC, ORO

299
 Reports previously issued in this series are as follows

ORNL-528 Period Endi ng November 30, 1949

ORNL-629 Period Endi ng February 28, 1950
ORNL-768 Period Endi ng May 31, 1950
ORNL-858 Period Endi ng August 31, 1950
ORNL-919 Period Endi ng December 10, 1950
ANP-60 Period Endi ng March 10, 1951
ANP-65 Period Endi ng June 10, 1951
ORNL-1154 Period Endi ng September 10, 1951
ORNL-1170 Period Endi ng December 10, 1951
ORNL-1227 Period Endi ng March 10, 1952
ORNL-1294 Period Endi ng June 10, 1952
ORNL-1375 Period Endi ng September 10, 1952
ORNL-1439 Period Endi ng December 10, 1952

ORNL-1515 Period Endi ng March 10, 1953

ORNL-1556 Period Endi ng June 10, 1953
ORNL-1609 Period Endi ng September 10, 1953

ORNL-1649 Period Endi ng December 10, 1953

ORNL-1692 Period Endi ng March 10, 1954
ORNL-1729 Period Endi ng June 10, 1954

ORNL-1771 Period End' ng September 10, 1954

ORNL-1816 Period End ng December 10, 1954
ORNL-1864 Period End ng March 10, 1955
ORNL-1896 Period End ng June 10, 1955

ORNL-1947 Period End ng September 10, 1955

ORNL-2012 Period End ng December 10, 1955
ORNL-2061 Period End ng March 10, 1956
ORNL-2106 Period End ng June 10, 1956
ORNL-2157 Period End ng September 10, 1956
ORNL-2221 Period End ng December 31, 1956
ORNL-2274 Period End ng March 31, 1957
ORNL-2340 Period End ng June 30, 1957
ORNL-2387 Period End ng September 30, 1957

300