ORNL-2440.txt

 

 

RGY S g A

KRR DoCUMEN BRARY _fs¢
ORNL-2440

Jd 445k 0251023 & C-84 ~ Reactors-Special Features !&grfi Reactors P‘
o

 

 

"""Jk-" 4 4 \_r'l‘ i

 

 

 

 

-AEC RESEARCH AND DEVELOPMENT REPORT
el |
Rt '
e~ ‘

¥. 1 "
Laatmn ! i

" !

B osn.-

Crss
Cr2
bt W s o4

3415 & | TN
T e

 

{.%5)

19 .

-
AL e s o o i,

- —

... 2.E63.

- EAse s e

-G

- -
-

AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

.
S AL
-

FOR PERIOD ENDING DECEMBER 31, 1957

— e}
-

   
  

..M_'_t;“ GRS o
CLassrrication CAANGED To:

By Au:rluyfi
- i

By

 

 

CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

e LIBRARY LOAN COPY
B RGN N LR R LA RGN

If you wish someone else to see this
document, send in name with document

 

‘e T o : and the library will arrange a loan.
- OAK RIDGE NATIONAL LABORATORY 8
operated by
UNION CARBIDE CORPORATION
for the

3 U.S. ATOMIC ENERGY COMMISSION @

% camr

N

   
 

 

 
ORNL-2440

C-84 — Reactors-Special Features of Aircraft Reactors

This document consists of 324 pages.

Copy/'z\-s—of 273 copies. Series A.

Contract No. W7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Period Ending December 31, 1957

A. J. Miller, Project Coordinator

DATE ISSUED

APR 101958

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORA ENERGY SYSTEMS W7

. v mome teeconnesos (| TTHIARTRLER

- 3 yy5k 0251023 b

 

 

 
 

 

 

 
 

 

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,
 

 

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART 1. AIRCRAFT REACTOR ENGINEERING

1.1. Reactor and Facility Construction

A summary of the status of the ART-ETU
program at the time it was cancelled is presented.
Reference is made to comprehensive reports that
have been prepared on the various aspects of the
development work. Much of the operations planning
had been done, and reference is made to documents
containing detailed operating procedures.  All
outside contract work on the facility in which the
ART would have been operated had been com-
pleted. A pictorial summary of that work is
presented, as well as cost summaries.

1.2. Component Development and Testing

Tests of actual and prototype ETU and ART
components and systems have been terminated or
are being completed before being terminated. The
experiments designed for establishing a detailed
pump sparging procedure were finished. The
reactor fuel pump rotary element that was being
operated in a new gamma-irradiation facility at
the MTR was shut down for repair of an ion
chamber, and during further inspection it was
found that other repairs were needed. This test
will be resumed, since this experimental study
of the effect of irradiation on bearings, seals, and
lubricants is of general interest.

An ART-type fuel pump that is operating at
service temperatures with NaF-ZrF4-UF4 (50-46-
4 mole %, fuel 30) has accumulated 3350 hr of
continuous operation. This pump is to be operated
for an additional 3000 hr under induced cavitation
in order to determine the effect on the impeller.
An identical pump is being operated with NaK in
order to study the effect of the type of fluid being
pumped.

Operation of a fuel pump in an endurance test
was terminated after a total operating time of
4410 hr. Replacement of the lower seal would
have permitted further operation, but this test is
not to be resumed.

Operating data on electromagnetic flowmeters
were obtained in a primary NaK pump loop, and a
split-purge arrangement for the shaft annulus is

being tested as a means of obtaining long-time
trouble-free operation of the lower seal. A Fulton-
Sylphon bellows lower seal on an auxiliary NaK
pump was found to have operated successfully for
936 hr with an average leakage of 16 cm3/day.
Approximately 8700 hr of high-temperature oper-
ation of primary NaK pumps with NaK has been
accumulated, and the accumulated operating time
for auxiliary NaK pumps with NaK is 5100 hr.

A sodium pump equipped with the thermal barrier
designed to protect the bearing-housing region from
thermal and radiation damage and also equipped
with a simulated ETU type of lubricating and
cooling oil system was found to operate quite
successfully in a high-temperature test. The
system was operated a total of 3800 hr, of which
2600 hr were without interruption,

The heat exchanger and radiator development
testing program is being concluded with a thermal
cycling test of a 20-tube semicircular heat ex-
changer designed to simulate the ART main heat
exchanger tube geometries. The heat exchanger
has completed 155 cycles of the proposed 300-
cycle program. A tube in ART test radiator No. 2
failed after a total of 181 thermal cycles, which
was six times the number of thermal cycles
proposed for the ART radiators.

Measurements made on the second quarter-scale
outer core shell model, which was subjected to
339 thermal cycles, are being analyzed in order
to determine whether dimensional changes oc-
curred. The shell is being prepared for a creep
buckling test at 1500°F., A welded shell is to
be tested in order to determine the reliability of
that type of construction.

Zirconium fluoride vapor traps designed for use
in the ETU were found to be satisfactory with
minor modifications. The additional changes that
would be required for satisfactory ART operation
were established.

1.3. Instrumentation and Controls Development

The ART control rod drive system, which was
successfully operated under simulated ART con-
ditions for 3000 hr, was disassembled and ex-

omined. There were no indications of wear and

- v
  

no sodium deposits on rod drive parts located
above the surface of the molten sodium.

In metering tests made in the aluminum north
head mockup just prior to termination of the ART
program, it was demonstrated that the purged-dip-
tube continuous level indicator could detect
changes in level in the fuel expansion tank with
an accuracy of 5% at design speed and flow
conditions. Turbulence prevented absolute level
Considerable work remained to be
done in determining the dynamic characteristics
of the measuring system under ART conditions.
A resistance type of fuel level indicator was
designed as a possible alternate to the purged
dip tube, and it is to be tested further. In pre-
liminary experiments the performance of the indi-
cator was in reasonable agreement with the
calculated performance,

Construction work on the resistance-type level
transducers that were to be used in the NaK pump
bowls of the ART and ETU was terminated with
the completion of a total of 22 wunits. Design
work was completed on all the continuous level
probes required for the ETU and the ART in the
sodium expansion tank, the sodium auxiliary tank,
and the NaK-furnace-circuit drain tank. Tests to
determine the error introduced by an unwetted level
probe were continued,

On-off level probes with brazed, rather than
welded, copper-Inconel junctions at the tip were
installed in a NaK system. After a sufficiently
long operating period these probes will be ex-
amined in order to evaluate the use of the braze
method of construction. Sixteen on-off probes
installed in the gas pots of two NaK systems have
each operated at high temperatures for 3000 hr
without a failure.

measurements,

The program of tests of 2- and 3]/2-in. magnetic
flowmeters was cancelled. The data obtained in
the operation of 20 such flowmeters are being
processed ond analyzed. Indications are that the
accuracy of these flowmeters was well within the
specified ART and ETU requirements. The fifteen
% -in, magnetic flowmeters that had been delivered
for installation on ETU cold traps, as well as
parts for 15 more units, will be packaged and
stored.

Examination of experimental turbine-type flow-
meters indicated the need for further modifications.
The bearing radial clearance was increased, as
was the clearance between the downstream end of

the rotor body and the positioner, In addition, the
over-all blade diameter was reduced to give greater
clearance between the blades and the housing.
Units that incorporate these modifications have
operated in excess of 350 hr at temperatures
ranging from 1200 to about 1600°F.

In further tests, six strain-gage pressure trans-
mitters of 0- to 100-psig range indicated an
average output-input deviation of 0.4% of full
scale over their operating pressure range from
ambient temperature to 1400°F, This accuracy
compares favorably with the accuracies of other
pressure transmitters,

Drift testing of thermocouples was continued,
and reports are being prepared that describe the
design and testing of ART thermocouples. One
report will cover fabrication techniques and the
results of aging tests in fuel, sodium, and NaK,
and the other will give the results of aging tests
in air.

1.4. Engineering Design Studies

Work on the applied mechanics and stress
analysis problems of the ART was discontinued.
Reports have been prepared that describe the
techniques used and summarize the results., The
results that had been checked by component tests
were found to be valid.

1.5. Design Physics

The design of the ART was based on extensive
reactor physics studies that were largely com-
pleted. The problems studied are described, and
reference is made to reports that give the details
of the work.

1.6. Materials and Components Inspection

A summary of the quantities and types of
materials and components inspected is presented.
The wvarious inspection methods used include
visual, dye penetrant, radiographic, ultrasonic,
eddy current, and others,

1.7. Heat Transfer Studies

Several additional tests were completed in the
high-frequency thermal-cycling loop. In this pulse-
pump apparatus the fuel mixture NaF-Zer-UF4
(56-39-5 mole %) is used as the circulating fluid
and helium is the pulsing and cover gas, For
these recent tests, the Inconel test sections were
made up of welded pieces of tubing and pipe that

 

 
 

 

were approximately 0.5 in, in inside diameter but
varied in wall thickness from 0.029 to 0.250 in.
A cursory examination of the data indicated that
the severity of cracking as a result of thermal
cycling increased with increasing wall thickness
and with proximity to a weld. This study is being
continued.  Similar experimental investigations
with the use of a pressurized system were dis-
continued.

Tests of a full-scale plastic ART core model
that was divided into three flow regions by in-
stalling two annular shell liners were also dis-
continued.  The data obtained, which consist
primarily of Pitot traverse measurements and dye-
injection movies, are being assembled. Summary
reports were prepared on previous experimental
studies of flow through quarter-scale models of
the ART core, including the screen-packed core
and header arrangement, and on the phosphorescent-

particle technique for velocity-profile visuali-
zation.

Forced-convection heat transfer studies with
KCI-LiCl flowing in a heated tube were dis-

continved after some tests of stainless steel test
sections. A summary report has been prepared on
the inconclusive results obtained in these tests.

An experimental system is being designed for
the study of forced-convection heat transfer with
molten lithium flowing in metal tubes. The test
section will be constructed of stainless steel and
will be instrumented for measurements of wall
temperatures, fluid flow rates, and mixed-mean
temperatures.

A study of the effect of screens in a diverging
annular channel on the thermal structure of a
volume-heated fluid is being continued. Also, a
system is being constructed for an experimental
study of heat transfer from an internally heated
liquid metal (mercury, initially).  An empirical
relation for forced-convection volume-heat-source
heat transfer with liquid metals was devised.

Preliminary data on burnout heat fluxes for water
in nucleate boiling have confirmed the concept
that source-vortex flow would delay the burnout
condition to a higher heat flux. The higher burnout
values obtained were not counterbalanced by
excessively large power expenditures.

Equipment is being designed and assembled for
a study of flow distribution in the coolant annuli

of the Tower Shielding Reactor Il (TSR-II). The

 

dota obtained will be combined with heat-gener-
ation rates in order to determine the redistribution
of flow required to approximate a flat exit-water
temperature profile.

The thermal conductivity of fuel 30 was re-
determined, and measurements were made of the
enthalpies ond heat capacities of the eutectic

mixtures LiCI-SrCI2 and LiCl-BaClz.

PART 2. CHEMISTRY
2.1. Phase Equilibrium Studies

Extensive studies of phase equilibria in fused-
salt mixtures have been made over a period of
years in order to provide information needed for
the selection of fuel and coolant mixtures for use
in aircraft reactor systems. Reference is made
to detailed reports on these studies.

A comparison of the Russian and ORNL investi-
gations of the NaF-ThF, system is presented. It
has been found that unclassified Russian reports
presently represent the predominant number of
contributions to the literature of phase equilibrium
studies of fused-salt mixtures. Since 1946 more
than 200 Russian reports have been issued con-
cerning the phase equilibria of oxide, silicate,
chloride, nitrate, and fluoride systems.

A disproportionation reaction was found to inter-
fere with investigations of phase equilibria in the
NaF-CrF, system. Various efforts to produce
pure NaF-CrF, samples that are free of Cettt
have been unsuccessful.

The upper operating temperature for the thermal-
gradient quenching apparatus has, through equip-
ment modifications, been raised from 850 to
1300°C. In the revised apparatus a single,
traveling, recording thermocouple was substituted
for 18 monitoring thermocouples; the furnace was
redesigned to permit the sample tubes to be
protected by a static helium atmosphere during
brazing; and a roller crimping machine was adapted
for use in sample preparation in order to minimize
sample tube expansion during heating.

Results of the use of the petrographic micro-
scope for the routine inspection of solidified fuel
mixtures are described. In such examinations,
oxygen-containing compounds are detected and
variations from a nominal composition are revealed,

An optical method is described for the determi-
nation of equilibrium paths and tie lines in ternary

vii

 
 

fused-salt systems. Application of the method to
the LiF-ThF ;-UF, system is demonstrated.

2,2. Chemical Reactions in Molten Salts

The results of a study of the disproportionation
of UF3 in fused-salt fuel mixtures are presented.
Nonacidic fuel solvents that contain a high con-
centration of fluoride donor constituents, such as
alkali fluorides, were found to increase the extent
of disproportionation by complexing the acidic UF
(fluoride acceptor). The ease of alloy formation
between the uranium formed and the container
metal also noticeably influences the extent of the
disproportionation,  Less disproportionation oc-
curred in copper vessels than in nickel vessels,
and tantalum and molybdenum were found to be
even more resistant than copper to alloying with
the uranium formed. It was also found that UF,
was oxidized by the alkali fluorides.

In comparison with UF ,-containing fuels, the
UF ;-containing fuels exhibit a marked improvement
in initial corrosion attack at high temperatures.
Unfortunately, however, the difficulties arising
from the deposition of uranium by disproportion-
ation prevented the easy exploitation of UF, as
a solution to corrosion problems.

Analyses of the results of determinations of the
activity coefficients of FeF, and of NiF, in
molten Nc:F-ZrF4 (53-47 mole %) are presented.
The numerical values of the activity coefficients
of FeF, are, at 800, 700, and 600°C, 1.46, 2.20,
and 3.28, with respect to the pure crystalline
solid; 0.64, 0.66, and 0.66, with respect to liquid
FeF,; and unity, with respect to the standard state
of reference at infinite dilution,

The results of similar studies of NiF, gave very
large discrepancies between the calculated and
experimental equilibrium constants and very high
values of the activity coefficients. It is clear
from the values of the activity coefficients that
when NiF, is dissolved in NaF-ZrF4 (53-47
mole %) it behaves in a manner vastly different
from that predicted by the free energies of for-
mation for solid NiF,. It is postulated that the
values available for the free energies of formation
of solid and liquid NiF , are not accurate.

Solubilities and activity coefficients of NiF
as functions of solvent composition in NaF-ZrF
mixtures are being measured in an effort to locate
the positions of the maximums of the activity
coefficient and the solubility. The present results

viii

indicate that the effective Z/R average of the
solvent may correspond closely to that predicted
by the previously proposed acid-base scale. An
observation that the value of the activity co-
efficient of the NiF2 solute is not dependent on
the solubility has verified a previous conclusion
that the precipitating phase is not NiF,.

A summary of the results of studies of the
reduction of UF, by chromium in various fluoride
reaction mediums is presented, as well as data
for the reaction of UF, with various other metals,
including Inconel, in which the activity of chromium
is much less than unity. In all solvents studied
the chromium concentration decreased as the UF
concentration increased. The implications of the
studies are analyzed in terms of the corrosion
mechanism. Molybdenum was the only metal found
to have stability in the Ncll:-LiF-KF-UF4 mixture,
and accordingly nickel-molybdenum alloys would
be expected to resist corrosion by this fused salt.
In the ZrF4-based melts, molybdenum, niobium,
and tungsten appear to be stable toward UF .
Tantalum and vanadium appear to offer no ad-
vantage over chromium for either type of melt.

Extensive study of the reduction of FeF, by
chromium in molten fluorides has indicated that
the behavior is far from ideal. An analysis of
Cr**to-Fe™  ratios obtained from equilibrium
quotients determined by the filtration method and
a comparison with activity coefficients obtained
from the standard free energies of formation in-
dicates that equilibrium was not attained in these
experiments., |t is apparent that the filtration
method is not capable of establishing the Cr**-
to-Fe** ratio in these reaction mediums at 600°C.

The results of experimental studies of the
temperature coefficient of the equilibrium

2UF4 + M= .'ZUF3 + MF2

when molten-salt mixtures are used as the reaction
mediums, were used as a basis for a thermo-
dynamic interpretation of equilibrium data on the
corrosion of iron and chromium by molten-salt
fuels. The discussion covers the tendency toward
mass transfer. !t was expected from the similarity
of the compounds CrF, and FeF, that their
solution behaviors would be similar and that any
difference would be governed by the difference of
their AH® values, Experimental data proved this
supposition to be invalid, and it must be concluded
that some of the data are incorrect or that CrF,
and FeF , have entirely different solution behavior.

 

 
 

 

A summary of the results obtained thus far in
the determination of the solubilities of various
noble gases in a number of different solvents is
presented. In all the cases investigated, Henry's
law was found to apply, and the solubilities were
found to increase with temperature. For the same
solvent, the solubilities of the gases decreased
with increasing solute gas size. Variations in
solvents showed corresponding gas solubilities
to be approximately the same in two related ZrF -
bearing solvents, but the solubility of helium in
NaF-KF-LiF was found to be less than half that
in the ZrF ,-bearing solvents.

The solubility of HF in mixtures in the NaF-
Zrl:4 system has been determined as a function
of temperature, pressure, and melt composition,
and a summary of the results is presented. In all
cases the solubility decreases with increasing
temperature.  Also, the Henry's law constants
for the HF solubility illustrated a constant heat
of solution over the experimental temperature range
used for each composition.

Experimental studies of the
fission-product fluorides in molten NaF-ZrF , mix-
tures were made. It was found that varying the
ratio of sodium to zirconium in the solvent had
a considerable effect on the solubility of CeF ..
The solubility of YF, in NaF-ZrF , (53-47 mole %)
was found to follow a pattern similar to the
patterns obtained for the rare earths. The solu-
bility values for CeF, and SmI:3 in combination
in NuF-ZrF4 (53-47 mole %) were found to be
substantially the same as those found previously
with  NaF-ZrF -UF (56-40-4 mole %) as the
solvent.

Measurements were made of the solubility of
BaO in KF-LiF (50-50 mole %). The results of
three experiments unexpectedly showed the solu-
bility to be quite high. This study is related to
a systematic investigation of the possibilities of
removal of fission-product fluorides from NaF-
KF-LiF base mixtures. A comparison study of
the reaction of CeF ; and UF4 with CaO in KF-LiF
The precipitation of

solubilities of

eutectic was also made.
cerium as Ce203 suggests the possible separation
of tetravalent metals, as oxides, from the trivalent
Alternatives would be to pre-
leave the fission

fission products.
cipitate uranium oxide and
products in solution or to precipitate the rare
earths as oxides and leave the uranium in the
solvent,

 

The results of a determination of chromium
diffusion coefficients in Inconel-molten salt
systems are described. This investigation was
undertaken because experiments carried out over
the past few years on the migration of metal as
a result of temperature gradients in a molten-salt
environment suggested that the over-all migration
rate is controlled by the diffusion rates within
the metal.

An analysis of steady-state chromium migration
loops is described. In-
formation available from the many thermali-con-
vection loops operated to date is correlated in
order to develop a method whereby the amount of

in thermal-convection

chromium transported can be estimated as a
function of loop operating time, and the extent
of corrosion to be expected at any time can be
predicted.

Measurements of the activities of the metallic
components of container materials for fused-salt
fuels have been made in order to study the effects
of composition and temperature on the activities
and thus on the corrosion resistance of container
alloys. The activity of chromium in nickel-
chromium alloys (Inconel) has been studied, and
the results of studies of the nickel-molybdenum
system are presented. In future work the chromium-
molybdenum system is to be studied, as well as
the ternary system nickel-molybdenum-chromium,
which includes the INOR alloys developed by the
Metallurgy Division,

The possibility of dissolving stainless steel fuel
elements in molten fluorides, with the resultant
advantage of adaptability to the present volatility
method for fuel reprocessing, prompted some ex-
ploratory experiments. In general it appears that
a practical process could be evolved for removing
the uranium from the steel and leaving a constant
nonaccumulative heel in the reactor vessel. This
heel could be flushed out periodically with a
dummy salt mixture.

In experiments for studying the oxidation of
sodium and potassium mixtures it was found that
the variables could not be controlled sufficiently
to establish any quantitative effect of original
composition on the Na-to-K ratio in the mixed
The data show, however, a significant
in the ratio of Na to X in the mixed

oxides.
decrease

oxides upon an increase in the potassium content
of the original alkali metal mixture from 49 to
77 wt %.
 

2.3. Physical Properties of Molten Materials
Methods

molten salts are described, and the results ob-
tained in these studies are summarized. Reference
is made to a concurrent effort to measure vapor
pressures that was carried out at Battelle Memorial
Institute.

Future vapor pressure investigations of fused
fluorides of reactor interest will involve systems
containing BeF,. Difficulty will be encountered

used for vapor pressure studies of

at BeF2 concentrations of less than 75 mole %
because vapor-phase compounds will contribute
appreciably to the total pressure. It will be
necessary to obtain partial pressures. Methods
for these measurements are being developed.

A mass spectrometric study of salt vapors is
also being planned in order to obtain an under-
standing of the nature of dimeric and trimeric
molecules in the vapor of alkali halides. Polymer
formation can cause components of salt mixtures
to be even more volatile as part of an associated
species than the pure component is, in some
cases. A preliminary study is described for which
potassium chloride was chosen as a suitable
material with which to develop the method.

The results of surface tension determinations on
molten salts by three different methods are
compared. The methods used were the sessile-
drop technique, the maximum-bubble-pressure tech-
nique, and a tensiometer technique. During the
course of these experiments it has become apparent
that further information is needed on the wetting
characteristics of fluoride salts,

2.4, Production of Purified Mixtures

A summary of é€xperience in the synthesis for
use in ANP chemistry studies of compounds that
were not available commercially is presented.
Of particular interest were fluorides of the metals
being considered as container materials for various
fuels and the products of their reactions with
constituents of proposed fuels. Corrosion products
found in fuels were identified by comparison of
their x-ray diffraction patterns and optical crystal-
lographic properties with those of synthesized
compounds. Fluorides of the fission products were
prepared for use in the study of their effects on
the fuel mixtures of interest.

Summaries are presented of the experience in
the preparation of fuel for the Aircraft Reactor

Experiment and of the establishment and operation
of pilot-scale and production-scale purification
apparatus.

2.5. Analytical Chemistry

Work was terminated on the development of an
instrument which would serve to monitor the
exhaust gases from the ART by detecting traces
of NaK in air. The instrument was based on the
measurement of the absorption of sodium resonance
radiation by sodium atoms. It is thought to be
applicable to the detection of a concentration of
sodium in air that would be equivalent to the
introduction of a total volume of 6 em® of NaK
per hour to the exhaust gases. This corresponds
roughly to an increment of 10 ppb of NaK across
the NaK-to-air radiators. The test model showed
that the response of the instrument to small
increments in the concentration of sodium in air
samples was almost instantaneous, and the model
was satisfactorily stable and reliable in all
respects. Other methods of detection which were
evaluated during feasibility studies included
photometric smoke detectors, emission spectrog-
raphy, flame photometry, and surface ionization
detectors. These methods were not considered to
be as promising as the resonance radiation ab-
sorption technique because they either lacked
adequate sensitivity or presented technological
complications that would make their application
to the ART extremely difficult.

The apparatus for the determination of oxygen
in fluoride salts with potassium bromotetrafluoride
was put into operation. The gas-measuring system
of the apparatus was calibrated to a precision of
about 2% with oxygen. The method was also
tested on samples of yttrium oxide and is being
used routinely for the determination of oxygen in
yttrium fluoride. The precision of the method with
respect to yttrium fluoride appears to be less than

5%.

PART 3. METALLURGY

3.1. Nickel-Molybdenum Alloy Devel opment
Studies

The developmental work on nickel-molybdenum-
base alloys for use as reactor structural materials
in contact with the fused salt fuel NaF-KF-LiF-
UF  (11.2-41-45.3-2.5 mole %, fuel 107) at normal

 
_

 

operating temperatures of 1650°F and above is
summarized. The various alloys studied fell
into three general categories: solution-hardening,
precipitation-hardening, and dispersion-hardening
alloys. The alloys were evaluated in terms of
fabricability, mechanical properties, corrosion
resistance, weldability, and aging characteristics.
The results of the evaluations showed that the
solid solution-hardening alloys most nearly met
the specifications. Extensive work has been
done in order to establish the final composition
of the alloy selected on this basis (INOR-8)
and to produce commercial quantities of the alloy.

Since the chromium content of INOR-8 was
expected to limit the peak metal-salt interface
temperature to the range 1650 to 1700°F, an alloy
designated INOR-9 was being investigated. In
this alloy the chromium is replaced with niobium,
which does not corrode in fuel 107. Preliminary
data on the properties of this alloy are presented.

3.2. Mechanical Properties Studies

properties of the various
materials to be used in circulating-fuel aircraft
reactors have been investigated during the past
several years. Fairly complete design data were
obtained for Inconel, the structural alloy that
was being used for the ART, for beryllium, the
moderator material, and for various shielding
materials.  Basic information on the behavior
of metals was also obtained in order to be able
to predict the safe operating life of various com-
ponents under a variety of operating conditions.

In the investigation of Inconel it was found
that, in the fused-salt environment, the fine-
grained material is stronger than the coarse-
grained material at 1300°F but weaker at 1500°F
and above. It was also found that, in order to
improve the strength of Inconel without impairing
its fabricating characteristics, it was feasible
to add carbon through carburization of finished
tubing.

It was found in the beryllium investigations
that data obtained on a laboratory scale could
be reliably applied to large-scale pressings.

Methods for interpolating and extrapolating the
available data properties have
been explored, and several relationships have
been found that have merit. Empirical equations

The mechanical

on mechanical

based on specific data are available for Inconel.
Tests of both rod and tube specimens of Inconel
were made in order to evaluate the effect of

strains that would be produced by thermal cycling

during reactor operation. The results obtained
in tests in which the strain was mechanically
induced compared favorably with those of a
supporting investigation at the University of
Alabama in which thermally induced strains are
being studied. The effects of temperature, grain
size, environment, specimen geometry, and cycle
frequency are summarized.

The shielding materials that have been inves-
tigated included an 80% Mg-20% Li alloy, an
alloy of very pure lead with 0.06% copper, and
a lead-lithium alloy.

3.3. Welding and Brazing Studies

A production facility for the manufacture of
brazing-alloy rings according to techniques and
procedures developed by the Metallurgy Division
was set up by the Engineering and Mechanical
Division. Approximately 750,000 rings were
produced by the graphite-mold technique, and
a production rate of 60,000 to 70,000 rings per
day was achieved. Production of these rings at
ORNL became necessary because the fabricator
of the ART radiators could not obtain satisfactory
brazing-alloy rings.

A procedure was developed for fabricating
thermocouples for high-temperature service that in-
volves brazing an end cap to the shield and
thermocouple wires. This procedure replaces
a welding technique used previously. Closures
made by the welding technique were frequently
rejected because of cracks and pores in the
Closures that are satisfactory as to
both integrity and consistency in the distance
between the end of the seal and the thermocouple
junction are made by the use of the new brazing
procedure.

An investigation of optimum methods for fab-
ricating the tube-to-tube-sheet joints of the ART
fill-and-drain tank was completed. A wunique
method for brazing the unit was developed that
may be applicable to tube-to-tube-sheet joints of
other types of heat exchangers.

Investigations were started of methods for
welding niobium. Fusion welds were tested in
simple bend tests and found to be ductile. Ex-
periments are being carried out in order to deter-
mine differences in the weldability of arc-cast
and powder-metallurgy niobium. Filler wires for
welding niobium are also being tested. Both
butt welds and tubing saddle welds have been

weld zone.

 
 

fabricated in an argon-atmosphere dry box by
using the tungsten-arc process and zirconium-
niobium filler wire. Impurities in the welding
atmosphere have been found to have an embrit-
tling effect.

Brazing tests were made on carburized Inconel
tubes. The desired fine-grained structure was
retained.  E xtensive grain growth would have
occurred in conventional Inconel subjected to
the same brazing cycle.

3.4, Corrosion and Mass Transfer Studies

In continuing screening tests of new brazing
alloys, Handy and Harman brazing alloys No. 91
(91.25% Ni-4.5% Si-2.9% B) and No. 93 (93.25%
Ni-3.5% Si-~1.9% B) were found to have good
corrosion resistance to NcF-ZrF4-UF4 (53.5-40-
6.5 mole %, fuel 44) when tested in seesaw
apparatus at a hot-zone temperature of 1500°F
for 100 hr.

In similar screening tests, iron-base brazing
alloys are being exposed to lithium as they
became available. (High-nickel-content alloys have
very limited corrosion resistance to lithium.)
An 86% Fe-5% Si~5% Cu~-4% B brazing alloy
showed good corrosion resistance under static
conditions but only fair resistance in seesaw-
furnace tests at 1500°F for 100 hr.

Ihconel welds made with nitrogen as the shield
gas were tested for comparison with welds made
with helium as the shield gas. The corrosion
resistance of the welds made in nitrogen was
about the same as that of Heliarc welds.

Several cermet-metal joints formed by fusion
or brazing of TiC-Ni, WC-Co, and WTiC,-Co
cermets to metals were found to be corrosion
resistant to molten sodium and to NoF-ZrF -UF,
(50-46-4 mole %, fuel 30) in 100-hr exposures at
1500°F. These cermet-metal joints were developed
to permit these types of cermets to be joined
to Inconel by brazing.

An examination was made of the rare-earth-oxide
ceramic cylinders that constituted a prototype
ART control rod and which were exposed to
molten sodium for 3000 hr in o rigorous test of
the control rod drive system. Some of the ceramic
pieces were broken as a result of rather large
thermal changes. There was no detectable cor-
rosion of the ceramic cylinders by the sodium.
Despite the breakage of some of the cylinders,
the rod would have been satisfactory for ART

operation because there was no change in the
location of the cylinders.

Extensive grain-boundary penetration was found in
the hot-leg areas of niobium thermal-convection
loops in which lithium was circulated. These
tests were conducted in a vacuum chamber.
This penetration is thought to be due to attack
on impurity phases, such as carbides, in the
grain boundaries. Experiments are planned in
which various grades of niobium will be tested
in order to determine what factors control grain-
boundary penetration.

A forced-circulation loop is being designed for
studying lithium-niobium systems. The niobium
tubing of the loop is to be protected from air
oxidation by a continuous outer sheath of type
310 stainless steel.

A summary of the results of the extensive cor-
rosion tests that were made of Inconel exposed
to fused fluorides and to sodium and NaK is
presented.

3.5. Materials Fabrication Research

An investigation of the recrystallization of
niobium and the effects of recrystallization on
recovery and on tensile properties was made.
Except for the large grain size of recrystallized
arc-cast material, no significant differences were
found between arc-cast and powder-metallurgy
material. Work has been initiated on fabrication
of material for use in forced-circulation corrosion-
testing loops. Attempts to fabricate tubing by
extrusion of compacted pellets were unsuccessful
because of failures in the tube walls during tube
reducing. Billets of arc-cast material are being
secured and will be extruded to produce tube
blanks for redrawing. Methods for the protection
of pump parts and other odd-shaped sections of
a niobium loop from oxidation during testing are
being investigated. Sprayed coatings of brazing
alloy appear to offer a possible solution to the
problem.

Further development of equipment has provided
means for producing yttrium sponge of increasing
purity,  Analysis of a recent batch of sponge
detected only 1000 ppm of oxygen, and a sample
of this sponge after arc-melting was reduced 37%
in area by cold-swaging.

Equipment is being prepared to make possible
the direct transfer of the molten fluoride charge

 

 
 

into the reaction retort in order to eliminate con-
tamination from handling. Molten-state transfer of
the lithium reductant is now being practiced.

The work on the neutron shield and the shield
plug for the ART fuel pumps was completed, and
a report describing the development work was
prepared.

The tensile strength and ductility of core material
for tubular control rods were determined from
extruded specimens. These specimens were pre-
pared from especially processed Lindsay oxide
and had somewhat greater ductility than previously
tested material.

Developmental work was initiated at the request
of GE-ANP on high-density BeO bodies for use
as moderating material. Very fine particles of
BeO are being obtained by the controlled cal-
cination of BeSO,.

Rods of MgO and of BeO were extruded to
provide rod insulators with four longitudinal
holes. Rods of this type are required for thermo-
couple assemblies.

Methods are being developed for hydriding dense
yttrium metal when it becomes available. Reactor-
grade zirconium, which is available and which
responds to hydrogen in much the same manner
as yttrium does, is being used in developing the
hydriding process.

3,6. Metallographic Examinations of Engineering
Test Components After Service

The results of examinations of the ART pro-
totype test radiator No. 1, which failed immediately
after reaching full power in its first controlled
thermal cycle, are presented. Data on fin-to-tube
joint braze integrity and the degree of oxidation
of the fin collar are compared with similar data
on other radiators. Incomplete melting of the
braze alloy rings was found in several instances.
Corrosion and mass transfer were negligible, as
was expected in view of the short operating period.
It is concluded that the tube failure which per-
mitted the NaK leak could have been the result
of at least three things: reaction with an in-
sulating material, damage to the tube during
fit-up and brazing, and the presence of stringers
across the tube wall.

Test radiator No. 2, which failed during the
182nd thermal cycle, was also disassembled and
sectioned. Forty tubes from the damaged area
were examined, and only one was found that

showed signs of incipient failure. The evidence
of incipient failure was similar to that found
previously in creep test specimens.

3.7. Nondestructive Testing

The inherent accuracy of the new eddy-current
thickness gage was increased by the replacement
of the previously used variable-frequency signal
generator with drift-free transistor oscillators. Small
cup cores that will produce symmetrical electro-
magnetic fields and will decrease the size of
the field are being constructed to improve eddy-
current cladding-thickness determinations.

Preliminary use of a closed-circuit television
system for direct x-ray imaging has given en-
couraging results. The technique for use of this
system is similar to that used in conventional
radiography. The greatest area of usefulness of
the system will be in the inspection of thin
sections through which high-intensity radiation can
reach the photoconductive target.

Routine inspection of tubing, pipe, and plate
was performed by using immersed-ultrasound and
eddy-current techniques. A special lot of Hastelloy
B was found to be among the best tested to date.

PART 4. RADIATION DAMAGE
4.1. Radiation Damage

Examinations of the four forced-circulation loops
that were operated in the MTR with fused fluoride
fuels as the circulated fluids have been completed.
The final items examined were the impeller
housings of the pumps used in loops 4, 5, and
6. No evidences of erosion were found on the
impellers.

Examinations were continued of the moderator
materials that were irradiated in the second ex-
posure in the MTR. No cracking or flaking of
the graphite slug that was encased in nickel was
evident, and the dimensions of the slug after
irradiation were essentially the same as those
prior to irradiation. Metallographic examination re-
vealed no reaction between the nickel and the
graphite. When the zirconium hydride slug that
was encased in molybdenum was being sectioned,
the casing separated from the slug, and thus it
appears that there was no bonding. The molyb-
denum was very brittle. Metallographic studies of
these specimens are continuing in an investigation
of significant’ differences between irradiated and
control capsules.

xiii
The out-of-pile tests that correspond to the MTR
tube-burst tests were completed. The times to
rupture were two to five times longer than in the
MTR tests, but were of the order of a factor of
4 shorter than in tests conducted by the Metallurgy
Division under similar conditions. Further in-
vestigation has revealed that the fused-quartz
insulation used in these in-pile and out-of-pile
tests caused changes in the calibration of the
Chromel-Alumel thermocouples that were operating
in a helium atmosphere. These changes gave
higher temperatures than those indicated. The
higher temperatures partially account for the
shorter times to rupture, but they do not entirely
account for the loss of ductility being greater
in the in-pile tests than in the out-of-pile tests.
Results of extensive tests in the LITR are also
reported.  The tube-burst specimens have ex-
hibited the low ductility found in the MTR test
specimens.

Disassembly of the vertical in-pile loop which
circulated fused salt fuel in the LITR for 235 hr
was completed, and various parts are being
examined and analyzed. The lubricant was found
to have decomposed and become waxlike.

An Inconel capsule and two Hastelloy B
capsules were examined that had been irradiated
in the MTR in contact with fused salt fuel. The
Inconel was attacked 1 to 2 mils in 304 hr at
1500°F; the power density was 6000 w/cmS,
There were thin films on the surfaces of the
Hastelloy B capsules, but the corrosion was
negligible. Attempts are being made to identify
the film and to determine the mechanism of de-
position.

In an investigation of irradiation effects in photo-
conductive detectors, both neutron and gamma-
irradiation have been found to decrease the photo-
response of PbS cells, and the decrease is pro-
portional to the total integrated flux. There is
threshold damage that seems to depend upon the
dimensions of the sample, which, if exceeded,
completely destroys the cell. Below threshold
the cell recovers in a first-order process. The
photoresponse at 24,000 A and above fails before
that at shorter wavelengths. In addition to a
decrease in response, the noise level increases
with irradiation.

A state-of-the-art survey of irradiation effects
in semiconductor barriers is presented to unify
the material presented in previous progress reports

Xiv

and to establish a reference of current concepts.
With this survey as a basis, subsequent reports
will present a coherent picture of the progress
being made in these investigations,

PART 5. SHIELDING

5.1, Shielding Theory

For most calculations of neutron scattering in
air, isotropic scattering in the center-of-mass
system is assumed, although it does not represent
the actual case. For purposes of comparison, a
calculation was performed in which anisotropic
scattering was assumed. The resuits from the
calculations indicate that the isotropic scattering
method of calculation is conservative.

Another calculation was performed to determine
the spectra of bremsstrahlung radiation resulting
from the degradation of electrons in sodium iodide
and in air. The assumptions made, the equations
used, and the results obtained are presented.

A third calculation was prompted by the need of
aircraft shield designers to know the production
of secondary gamma rays by neutron interactions
in air. The problem was idealized to that of
determining the dose rate and angular distribution
of gamma rays at a detector located an arbitrary
distance from a neutron source. The source and
the detector were considered to be embedded in an
infinite volume of air. The source was taken to be
either a point isotropic source of monoenergetic
neutrons or a point source which emitted a mono-
energetic line beam of neutrons at an angle 6
with respect to the source-detector axis. For a
point isotropic source ond a 50-ft separation
distance, it was found that the dose rate varies

with the neutron source energy approximately as
-1/3
E .

0

When curves for various monoenergetic
line-beam sources and a 65-ft separation distance
were plotted as a function of 6, they were rela-
tively flat over the region from 30 to 180 deg, and
they became flatter as the source energy increased.
Since aircraft reactor shields are usually de-
signed so that most of the neutron leakage is
from the rear, the results of the line-beam calcu-
lations indicate that a knowledge of the exact
surface distribution of low-energy neutrons is of
less importance than a knowledge of the energy
spectrum and total neutron leakage.

 

 
5.2, Lid Tank Shielding Facility

A second group of tests of advanced shielding
materials was performed in conjunction with the
Pratt & Whitney ANP shield design effort. In
this group of tests particular emphasis was placed
on a study of the production of secondary gamma
rays in a gamma-ray shield of depleted uranium.
In most cases, slabs of nickel and stainless steel
were placed between the fission source and the
uranium. Borated water was always used as the
neutron shield behind the uranium.

Another series of tests with advanced shielding
materials was performed at the request of General
Electric. These tests, again, were designed for
a study of the production of secondary gamma
rays in gamma-ray shields. The materials used in
this series of tests were Hevimet (largely tungsten),
stainless steel, and depleted uranium. Lithium
hydride was also included in most of the con-
figurations tested, and, since it is highly reactive
with water, oil rather than water was used as
the neutron shield.

5.3, Tower Shielding Reactor |l

The mechanical design of the Tower Shielding
Reactor |l (TSR-1I) has been modified slightly
to provide more clearance for the thermal ex-
pansion that will occur at the highest operating
powers, A second mockup of the control mechanism
has been constructed and is undergoing testing.
Additional nuclear calculations have been per-
formed, from which it has been found that the
control plates could be worth 6% in k. The
critical mass, which was found to be 8.2 kg of
U235, will be checked with further calculations
and in a mockup experiment with dummy fuel

  

elements and uranium solutions. The design of
the water cooling system is approximately 95%
complete.

5.4. Tower Shielding Facility

Gamma-ray measurements in various crew-shield
mockups at the Tower Shielding Facility have
indicated an intensity which is greater than that
This
prompted a study of the production of secondary
gamma rays in air, both those resulting from
thermal-neutron capture in nitrogen and those
resulting from inelastically scattered fast neutrons,
For this study an experiment was performed in
which a beam of neutrons from the reactor irradiated
a cone of air, and the gamma rays resulting from
neutron interactions within the cone were observed
by a gamma-ray spectrometer several feet away.
The ratio of the nitrogen capture gamma rays
observed by the detector to other gamma rays
varied by covering the reactor

expected from scattered gamma rays alone,

observed was
neutron beam with boron-impregnated Plexiglas for
some of the measurements. The difference between
the measurements was then used to separate the
pulse-height spectrum associated with nitrogen
capture gamma rays from the spectrum associated
with other sources. In this report the spectrum
which remained after the subtraction of the capture-
gamma-ray spectrum was presumed to be due to
inelastically scattered fast neutrons; however,
contributions from other sources, such as air-
scattered gamma rays, must be considered. The
results of the experiment are to be compared with
theoretical calculations of the pulse-height spectra
to be expected from neutron interactions with air,
as well as from scattered gamma rays.

 
 

 

 
 

CONTENTS

PART 1. AIRCRAFT REACTOR ENGINEERING
. 1.1. REACTOR AND FACILITY CONSTRUCTION ..ottt 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 ... S
Summary of ART Facility COnStrUCHION ....cooviiiiieniiiii s 6
1.2. COMPONENT DEVELOPMENT AND TESTING ...coorii i 24
PUMp Development TeSTS ..ot s 24
Bearing, Seal, and Lubricant Tests ..o 24
Aluminum North Head Water Tests ....ooiiiiiieeeicirrtiice e sin st s 24
Fuel Pump High-Temperature-Performance Tests ..o, 24
Fuel Pump Endurance Tests ..ot o 25
Primary and Auxiliary NaK Pump Development ..., 25
Sodium Pump Endurance TeSt ..ottt s 25
Reactor Component Development Tests ... 27
Heat Exchanger and Radiator Development Tests ..o, 27
- Outer Core Shell Thermal Stability Tests ..o, 27
Zirconium Fluoride Vapor TraPS ..ot 28
. In-Pile EXPEFiMENTAION ..o ieimts it rses et bbb 28
1.3. INSTRUMENTATION AND CONTROLS DEVELOPMENT ..ot 30
ART Contro] Rod Drive TeSt coooeieeiiiiiieeteeeeee et erecimmiasas e smess e st e e bbb 30
ART Fuel-Expansion-Tank Continuous Level Indicator ... 30
Resistance-Type Fuel Level Indicator ... 34
Liquid-Metal-Level Transducers ...t s 38
ON-OFf Leve! ProbDeS . ooovieeeeeieceeeeee et etes ettt e b bbbt S 39
MAGNEHIC FIOWMETEIS ...t iees s 41
TUrDINE FlOWMETEIS c.voveeeeeeeietiv it esee b et e e e SRR 42
PresSUrE TrOMSMITTEIS cooveeieeeeieieeeotestessaeecesee et eeeb bbb sesaeas s e e eE s e a S eh SRR IR e bR 45
ART THEIMOCOUPIES oeierrveiiteteteciceeiececeat e ras s e 45
Sheathed TherMOCOUPIES .oo.iviiiieiesire ettt bbb b 45
Well ThermMoCOUPLES .vuiviuisieeseieieeee et cr e es b s 46
' 1.4. ENGINEERING DESIGN STUDIES ..ottt b e 49
Applied Mechanics and Stress AnalySis ..ot 49
| 15, DESIGN PHYSICS. oottt ot o e 50
. 1.6. MATERIALS AND COMPONENTS INSPECTION ..ot 52
MOtErials INSPECTION .ouiiivieiseerer e eir s 52
We ld TRSPECHION wevereriveeereseieeescesesieseeans s e oo b LSS 52

 

 
1.7.

2.1.

2.2,

xviii

 

HEAT TRANSFER STUDIES ...ttt ettt

Thermal-Cycling Research ..ottt
PUlse-Pump System. .ot e et eb e e e e e es
Pressurized System .. ... et st ee e sbe et e

ART Hydrodynamics ... it et et st se s et se s eat b ee b sbb et s s e eeeebeeeeeeeneene e
Fus@d Salt Heat T rans er oot e e et eee e e e eeatue e eeensseeesseseesesessesesaaee s e s e eee e e e e e e e e

Volume-Heated SyStems .......ccooiiiiiiiiieiiiiicece ettt e e st e n e st et sre e st rerae s saeresaesbesante s
ART-Type Core with SEre@ns .......o..o.oooviiiiir ettt r e s ere et a e
Liquid-Metal EXperiment ............ccooiiiiiiiiiiiicec st e e st ssaestneebe et e s

Heat Transfer with Vortex Flow oottt eeeeee e e e e e et s et e e s e e e ees e ee s s e e es e e e e e e

Tower Shielding Reactor !l (TSR-I1) Hydrodynamics .......ccccoooviimiiciececcee e

Physical Properties . ..ccoiiiiics ottt sttt st er e e e rennans
Thermal ConduUCtiVIty ..ottt esre s e e e et e st e st s st e st te e enee et eerne e e eene
Enthalpy and Heat Capacity ..ottt et e et reeeeeeenean

PHASE EQUILIBRIUM STUDIES ......ooouitiitiitiet ettt eeee et ee s e
Summary of Salt System INVestigations .....cc.cv...oieieeceeeeeieeeeeeee e eeeeeese e s s eesceseeeseseesees e oo
Comparison of Russian and ORNL Investigations of the NaF-ThF, System ..o,
The System NaF-CrF o et st
Apparatus for High-Temperature Thermal-Gradient Quenching Experiments ..........o.ocoovovoveveeinn .
Petrographic Analyses of Fuel Mixtures ...........cccoooooiiiiuiiioeee oot veee e e,

An Optical Method for the Determination of Equilibrium Paths and Tie Lines in
Ternary Fused-Salt Systems ...t v et
Systems Without Solid SolUtion ... e oo
Systems With Solid Solution ... e,
Index of Refraction vs Solid Solution Composition .......oo.eeceeeieieeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
Composition as @ Function of Temperature ..............cc.ocoovoeeviivoveeiirieeeseeee et eeee e eeee e
Constructing Tie Lines and Equilibrium Paths .........ccocooiiiiie oo ee e,
The System LiF-ThF j-UF , (oot
CHEMICAL REACTIONS IN MOLTEN SALTS ..o ee e et ee e s s et s,
The Preparation of UF, and Its Stability in Molten Fluoride Mixtures .......ccccoovercercrrcccieinnn,
Activity Coefficients of FeF , and of NiF, in Molten NaF-ZrF , s

Solubilities and Activity Coefficients of NiF , as Functions of Solvent Composition
in NGF-ZrF o MixtUres ....ccoooooovi et ettt

Reduction of UF , by Chromium in Molten Fluorides ..........c.cococoovinvniiiinniiiice s
Reaction of UF, with Various Structural Metals in Molten Fluorides ........ccccocvvieiniinienccn,
Reduction of FeF, by Chromium in Molten Fluorides ..o,

Thermodynamic Interpretation of Equilibrium Data on the Corrosion of Iron and of
Chromium by Molten-Salt Fuels ... e

Solubilities of Noble Gases in Molten Fluoride Mixtures ......ooooovoooeeooeooeee e

 

 
 

2.3.

2.4.

2.5.

 

Solubility of HF in NaF-ZrF 93
Solubilities of Fission-Product Fluorides in Molten NaF-ZrF , Mixtures ..o 95
CeF, in NGF-ZrF o s 96
YF, in NGF -ZEF | b 96
CeF, plus SmF, in NGF-ZEF 4 s 97
Solubility of BaO in KF-LiF .t 97
Reaction of CeF , and UF, with Ca0 in KF-LiF et e b 99
Determination of Chromium Diffusion Coefficients in Inconel-Molten Salt Systems.............c.cc...c. 100
D ifFUSTION CORFFICIEIES oooeeeeiieeeiceeieceier et e e et e e e e sbe sttt e s e e b et smeere b s ke sar st e s pan st bbb 100
Measurements Based on Radioactive Tracers ......ccccooicciiieiiiiiiiiiiiiinisers e 100
Battelle Diffusion EXpPeriments .. ...ttt s 102
Steady-State Chromium Migration in Thermal-Convection Loops ..., 104
. Chromium Migration Mechanism ... 107
Steady-State MIGration ... e s 107
Unsteady-State Migration ...t et s 108
Estimating Equilibrium Concentrations for Steady-State Loop Operation.........ccocooiiiiiiinciens 108
Mass-Transport Equations for Steady-State Migration ... 109
Salt and Wall Temperature Relationship ... 110
ACHIVIHIES 1M ATTOYS .ooitetiere ettt s 113
Dissolution of Stainless Steel Fuel Elements in Fluoride Melts ..o et eeans 115
Oxidation of Mixtures Composed of Sodium and Potassium ... i 118
PHYSICAL PROPERTIES OF MOLTEN MATERIALS ..o, 120
Summary of ARE Fuel Vapor-Pressure Studies ... 120
Vapor Pressures of Molten Fluoride Mixtures ... 122
A Mass Spectrometric Study of Salt Yapors ... 124
Surface Tensions of Molten Salts ..o 127
PRODUCTION OF PURIFIED MIXTURES ..ottt et 129
Preparation of Fluoride Compounds .........ocooeiiiiiii s 129
ALK i-Meta] FlUOCRIOMAEES .. ooeeeeeiiiiceieeie e eteete et e st e et sre e s nbn e aast e sae e st et e 129
Other Double FIUOFIAES .oovoiireieeire i eeterre e eees e eia b r e s e b st 129
Simple Inorganic FIUorides ...t 129
CRromous FHUOIIAE oottt ettt ettt e btttk bbb RS 130
Summary of Experience in the Preparation of Fuel for the Aircraft Reactor Experiment ............. 131
Purification of Fuel SOIVERT oottt b b 131
Purification of Fuel Concentrate ..ot s s 132
Storage and Ultimate Transfer. ..o 132
Pilot-Scale Purification OPerations ... s 133
Production-Scale OPErations ..........imimiimmmses sttt e o s 133
SErVICE OPEFAHIONS .. ..cvvereeerereieiareuaimie s csintss e s e e s 134
ANALY TICAL CHEMISTRY oceeeeeeeee ettt et s s e a s aras s e bbb bt s 135
Detection of Traces of NaK in AQr ..o 135
Determination of Oxygen in Fluoride Salts ... 137
3.1.

3.2.

3.3.

3.4.

XX

 

PART 3. METALLURGY

NICKEL-MOLYBDENUM ALLOY DEVELOPMENT STUDIES ......c.ooiieeeeeeeeeee e 141
Material Development ... ettt e ettt eete et e erans 141
Status of Development ...ttt 141
Properties of INOR-8 ...ttt et et et e e es 143
Properties of INOR-G ettt 143
Phase Diagram Studies .......c.ccciviiiiieicee et b s e sas e e bttt st et eeeeeneeeaes 143
Stress-Rupture Testing by New England Materials Laboratory .......occoovvvvicivevierieeieee, 143
Status of Production Heats ...ttt st 143
Composite Tubing and Plate ..., 145
Mechanical Properties Investigations ........ccooviiiii it et ettt 145
COrTOSTON SHULIES .oviiiiiiic ettt e s e et b ettt st et e st eemte et et aeene et e et e e eneeeee e 146
Forced-Circulation Loop Tests of Hastelloys ..o 147
Forced-Circulation Loop Tests of Experimental Nickel-Molybdenum Alloys.........ccccoevenennne. 147
Thermal-Convection Loop Tests of Experimental Nickel-Molybdenum Alloys............c............. 149
MECHANICAL PROPERTIES STUDIES ...t e 151
Design Data oo e et ettt ettt e et n et e s 151
IRCONET oot et ettt a et e r ettt ettt et e ettt e e 151
Recrystallized Inconel ..., et er et et e ereneas 152
Bery llium oottt n et et see et enes 152
Extrapolation of Test Datd ..ot e 152
Thermal Stress INVestigations ........ccoociiiiiiiiictc ettt et 153
REIAXAHION ittt ettt ettt ee et e et e e e e et e e et et et et e e ee e e e sre et et eneaes 154
QU ottt et e ettt e be et et et e eaeeseereteer et e eeette e e eeeeaneteenenee e 154
Biaxial Cre@P . cciei it r et e be et et b b e e et n e e 154
Tests of Shielding Materials ... 155
WELDING AND BRAZING STUDIES ..., e e et e ees 156
Production of Cast Rings of Brazing Aoy ..ot 156
Fabrication of Thermocouples for High-Temperature Service ...........ooooeiiioimiiie e 156
Fabrication of ART Heat Exchangers..........cocooviiiiiiiii it 157
ART Fill-and-Drain Tank Fabrication ..........ccocoooioiiiieeeee e v 159
NTOBTUM Welding ...ocoiiiicc ettt bt e et e s e et e s 160
Carburization of Inconel Tubing ..ottt 166
Welding and Brazing Procedures ...........cooovoioiiiiiiicc et sersn e 166
CORROSION AND MASS TRANSFER STUDIES ...ttt st 169
General Corrosion Investigations ..ottt 169
Brazing Alloys in Fuel Mixtures ..ot 169
Brazing Alloys in Lithium ... et st v 169
Inconel Welded with Nitrogen Shield Gas ..ot et ene s 170
Cermet-Metal Joints in NaF-ZrF ,-UF , and in Sodium .........cocoommmiiii 170
Postoperative Examination of Prototype ART Control Rod ..o 172

 
 

3.5.

3.6.

3.7.

4.1.

 

Corrosion in Dynamic Systems . ..o 173

Lithium in Niobium Thermal-Convection Loops ..c.ccoviiiiiciiioiiiiiriiie ettt 173
Lithium in Niobium Forced-Circulation Loops ...ttt cnree e sncesac s 175
Developmental Studies of Inconel Systems ... 175
Incone! Exposed to Fused Fluorides ...........ccccooiioiiiiviiii e, 175
Inconel Exposed to Sodium and to NaK ... 176
MATERIALS FABRICATION RESEARCH ..ot e, 177
NTODTUM FaBriC @ ION ceoeeeeeeeeeeeeeeeeeeee et ee e vt ers e es e e e st b et e s bbb sa e b e sa bbb bbb et an e a s sees 177
Recrystallization of Niobium ..c.ccoiiiiiii 177
Fabrication of Niobium Loops ...ttt bbb st s 177
Consolidation of Niobium Pellets ..o 177
Y HriUm MetAl ProdUCHiON ..ooeeeeii oot ete e sttt ea et et s sa e a st saeas aeasasnna s e en et naneanenaan 178
Shielding Materials Development ............cooiiiiii 179
Tubular Control Rod ProduUction .......ccciiiiiiii ittt st s n b s 180
Ceramic INVESHIGAIONS ...iiviiiciiictt st bbb s 180
Development of High-Density BeO Bodies ... 180
Extrusion of Magnesia and Beryllia Insulators ... 180
Production of Metal Hydrides ..........ocooiiiieiiiiiii e e 180
METALLOGRAPHIC EXAMINATIONS OF ENGINEERING TEST COMPONENTS
AFTER SERVICE oot et et s e e a e e e ettt ei o b aeas s arae s e e e s e ea s aeen s naneas s ane s 182
ART Prototype Test Radiator No. 1 .. 182
ART Prototype Test Radiator No. 2 ... 187
NONDESTRUCTIVE TESTING oot et s 192
Eddy-Current Thickness Measurements ... 192
Remote X-Ray VIEWINg .ot s 192
Materials TNSPECHON oot e 193
TUBING ..ttt ettt 193
P i ettt ettt et e d e b RS SRR 194
P IO oottt e e ettt ettt it te eyt et aeatae s ee et e b ehbrer e et st e e e e e et s are e e b et Ee e s 194
Hastelloy B Tubing ..o 194

RADIATION DAMAGE . oot et e et et e et be s b bbb a s s s st m et e b 199
Examination of lrradiated Components and Materials ... 199
MTR IN-Pile LOOPS ..oviivieeireieiee oottt ce e ees e ee et s 199
MO EratOr MAtErTALS .o oee oot ete e et e e re e oo e e b er e s ae s b s b e s e g e e s bt e e 200
Creep and Stress-Rupture Tests on Inconel ... s 207
MTR EXPEriMENtS oovovuiteeieeeeteteteeeeeessiesecesssmas s srs s b b8t 207
LITR EXPEFiMENTS ..oovoivoeeeieieiees st eoesieeesaes bbb b 208
ORR EXPEIriMENTS ..oooeovietetceiiceeiseecses oo inis s tins s 208
LITR Vertical Fused-Salt Loop .. ..ot s 208
MTR Statie COrroSion TSt overieeeeeeeiieee e stesre e steteseeaease et s ar bbb e s e e s e E s m et e b ea s et s a etk b s 209
Irradiation of Moderator Materials for Use at High Temperatures..........cooonincnn 209

XX
o il AP

 

 

 
5.1.

5.2.

5.3.

5.4.

XXii

I “ |- .
o .

Irradiation Effects in Photoconductive DeteCtors ..o iieeiee i ieeeeeieeee et eeee e ee e e e e ee st ee e e e eas oo 210
Irradiation Effects in Semiconductor Barriers ........................ e eerreeeeieeeaseete e st e teet e arreeteere e be s tin 218
SEmMICONAUETOr Material s oo e e 218
S M CONAUCTOr D BVICES oottt et e e et e e e e e e e e saaeeeeeserseaaereesereeeaseseee e e ee e e e 221

SHIELDING THEORY oottt ettt ettt et et et e es et e e e ee e 229
Effect of Anisotropic Scattering on the Neutron Flux in Air........coooiiiiiie, 229
Bremsstrahlung Spectra in Sodium lodide and in Air ..o, 231
A Calculation of Gamma-Ray Dose Rates from Neutron Capture in Air.......cccooovvvvviieeciiieiiiieee. 233
Calculational Method ... ettt neene 234
Results for Point |sotropic SoUurces ..o 235
Results for Line-Beam Sources .......o..coiviiiiiiin ettt st st 236
LID TANK SHIELDING FACILITY oottt st s sae st sv et ene e 238
Pratt & Whitney Series of Advanced Shielding Materials Studies ... 238
General Electric Company Series of Advanced Shielding Materials Studies......c..ccoocvveviivecciirennn. 261
TOWER SHIELDING REACTOR Il ettt eaee e et e e 275
Mechanical Design ... e et en e et eee ettt 275
Control Mechanism .. ... ettt aee e ere e e s ee 275
Nuclear Calculations ... ... et e e et s eeene s e enn 275
Water Cooling and Purification System ..ottt e 281
Extension of Gamma-Ray-Heating Calculation for the TSR-11....cooiiiiee 281
TOWER SHIELDING FACILITY oottt vttt et st st st e 285
Study of Gamma Rays Produced by Neutron Interactions in Air ..o.ocoovieiiiieeeeeeeeeeet e 285
Experimental Configuration .......cccoooiiiiiiiiii i 285
Determination of Detector ReSPoOnSe ...occicoiiiiiiioi et 286
Experimental Results and Discussion ... 287
Thermal-Neutron Flux Measurements ..........cccccooooiiiiiiiiiiii e, 292

 
 

Part 1
AIRCRAFT REACTOR ENGINEERING
e 1,]1. REACTOR AND FACILITY CONSTRUCTION

W. F. Boudreau

A. P. Fraas

H. W. Savage

STATUS OF THE ART-ETU PROGRAM
AT TIME OF TERMINATION

A, P, Fraas

The Aircraft Reactor Test was designed as a
60-Mw  circulating-fuel reactor having a major
portion of the features essential to an aircraft pro-
pulsion reactor. The details of the design have been
reported, as they were developed, in previous
reports in this series, and they were summarized
in a design report.! Diagrams of the reactor and
the cell in which it was to operate are presented in
Figs. 1.1.1, 1.1.2, and 1.1.3. At the time of
termination of the program, a full-scale, non-
nuclear engineering test unit (ETU) that essentially
duplicated the ART was being fabricated and
assembled. The ETU was to demonstrate the
feasibility of fabrication and assembly, and, in
high-temperature operation, to test the effects
of the complex thermal and pressure stresses
expected during nuclear operation. The com-
plex geometry, the wide variety of stresses,
and the difficulty of predicting the magnitude and
direction of thermal warping and distortion had
indicated the need for a comprehensive test on a
non-nuclear assembly. Upon completion of the
planned tests, the ETU was to have been dis-
assembled and thoroughly examined before operation
of the ART.

At the time of cancellation, the design, the shop
drawings, and much of the component testing and
fabrication had been completed. All outside
contract work on the facility in which the ART
was to operate had been completed, and work was
to have started on the installation of the process
piping.

The analysis of the design of the ART was com-
pleted except for a number of thermal stress analy-
sis problems intimately related to test work which
was in progress. Sufficient analytical and test
work had been completed to indicate that the
design was adequate to carry the basic pressure
stresses and a modest number of thermal cycles.
The design objective was a nominal life of 300
thermal cycles, whereas the operating program

 

VA. P. Fraas and A. W. Savoltinen, Design Report on
the Aircraft Reactor Test, ORNL-2095 (Dec. 7, 1956).

anticipated only 30 full thermal cycles; thus a
factor-of-10 margin was provided to ensure a high
degree of reliability. The results of the analytical
design studies agreed surprisingly well, in most
instances, with the results of tests of components,
for example, heat exchangers. |t was expected
that the programmed tests would have confirmed the
results of analytical studies of the designs of
other components.

Much of the extensive program of component
testing had been completed. The performance
testing of the pumps was finished, together with
much of the extensive endurance test program.
The fuel-to-NaK heat exchanger component test
work was largely completed, except for some tests
on circularearc tube bundles designed to simulate
in considerable detail the geometry of the full-
scale reactor.

Difficulties were experienced with leaks caused
by thermal stresses in the prototype NaK-to-air
radiators that were tested, but design modifications
were made which, it was believed, would give an
acceptable arrangement. Tests of the modified
radiators were to have been started shortly after
the program was cancelled, The reflector cooling
circuit was completely worked out and checked in
flow tests. The core hydrodynamics had posed
many problems, but tests were made in the fall of
1957 on a combination of screens which was
found to be effective in suppressing local temper-
ature fluctuations. The control-rod drive mechanism
had been thoroughly tested and found to be satis-
factory in extensive endurance tests, The principal
area in which component test work remained to be
done was in structural tests of the thin shells,
This work had been delayed by difficulties in
fabricating these shells, but analytical and pre-
liminary test work on models indicated that the
shell designs were satisfactory.

Much of the fabrication of the first set of parts
was completed. The pumps were available; the
basic parts of the north head were finished, in-
cluding the sodium-to-NaK heat exchangers, and
the assembly work was 70% completed. The
island beryllium and core shell assembly was
completed, and the beryllium for the reflector was
machined and ready for assembly, as were the
ANP PROJECT PROGRESS REPORT

     
   
  
 

ORNL-LR-DWG {6011

.- FUEL PUMP
CONTROL ROD._

Na EXPANSION TANK

NoK
FUEL EXPANSION TANK..

Na—-TO—-Nak
HEAT EXCHANGER e

 
 
  

     
   
     
 
   
    
 

  

     
   
    

 

    
   

  

 

     
 
  
   
 
 
     

     

-
5
=
3 - HEAT EXCHANGER
2 - ASSEMBLY
_~PRESSURE
AN SHELL
REFLECTOR + s .
ASSEMBLY . FUEL TUR
L J
\\‘\ S \\\‘\\‘\ . '.\ \.." I ‘ o ...\‘ . “I‘ :
THERMOCOUPLE t[*"0 R N
INSTALLATION —- ‘
S SNy
. N RN .
- > SN ISLAND-.
TILE LAYER— NN ISLAND:.
~
BERYLLIUM /AL~
REFLECTOR— /17"~
MODERATOR. -] |
. FUEL-TO-NgK
. A HEAT EXCHANGER
s - R o
SPHERICAL B,C 7 : ,

TILE LAYER

THERMOCOUPLE SLEEVE

.

      

“SFILLER PLATES
Na

“DRAIN

FUEL DRAIN (HOT}

Fig. 1.1.1. Vertical Section Through Reactor Assembly.

 
PERIOD ENDING DECEMBER 31, 1957

ORNlm 16457

CONTROL ROD

18-in~THICK BORATED WATER SHIELD

ALUMINUM TANK

     

TYPICAL LiCH GAN

!

w
S
=
o
=
e
=
>
<L
o
o
-
X
w
>
o
o
o
=
=2
a
—
w
2
w

 

T
e
3
T
e
=
a
w
-
=
w
0
3
a
a
z
o
o
q
x
i
I

 

PARAFFIN

      

  

ALUMINUM TANK

   

 

e
mT
w2
Sae
WUL
o Souw
w 2T
= =Wy
29 98
Sw Eay
I xa
x
Sv mmw
Tx 00
=w nd=Z
T = Swd
cg nCo
£%
1 W W <g
o T
al/ Uk

 

 

 

//////// /n////,//ux/ S
R GO R o
RS T R

h“\!‘.‘.‘.

N§§/

SN2

 

 

REACTCR SUPPORT PADS

ALUMINUM TANK

HANGER RODS

SUPPORT FOR WATER SHIELD
SUPPORTING LEAD SHIELD

 

&§§%%%\§§\%%7

‘\‘DO R

4.3-in.— THICK LEAD
GAMMA-RAY SHIELD

33-in-THIGK BORATED
WATER SHIELD

oo 0 .

V,—in.~ THIGK
AIR GAP

 

  

REACTOR
PRESSURE SHELL

ASSEMBLY

&\\%&/
%%%§C>%7§§§%%w
N AN AN ]

NNV b SN NN iy

GCOOLING WATER PASSAGES,
¥, -in-0OD TUBES ON

4-in. CENTERS

S %//z\\\\\|¢///////\\t|\/////\\\\\ 7

\Q\%%%

%97\\

/Z7§§§%y
§§§%%§§§

N\

 

Va=in.-THICK
THERMAL INSULATION

3-in~-THICK LEAD GAMMA-RAY SHIELD

   

HEXAGONAL CANS FILLED WITH LiOH

Ya~in-THICK THERMAL INSULATION

FUEL DRAIN LINE

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

ANP PROJECT PROGRESS REPORT

A
ORNL-LR-DWG 15056

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 1.1.3. Reactor Assembly Cell.

 
 

boron-bearing parts for the boron curtain. A proto-
type main fuel-to-NaK heat exchanger tube bundle
had been completed and was found to meet specifi-
cations. The remaining tube bundles were ex-
pected at the rate of two per month. The forging
for the pressure shell was being machined.

The ART design presented some difficult prob-
lems in the fabrication of thin-walled Inconel
shells, particularly in connection with the core
shells and the thin, hemispherical shells for jack-
eting the boron. Various techniques, including
hydrospinning, pressing, and machining from forgings
and pressings, were investigated. Hydrospinning
was found to be unsatisfactory, but the core shells
were machined readily to the desired tolerances from
pressings and were ready for assembly. It was
found that the large-diameter thin-walled hemi-
spherical shells could not be fabricated by hydro-
spinning, but similar parts were made (with dies
built for a missile application) which showed that
these parts could be fabricated by hydraulic
pressing. New dies for the ART parts were sched-
uled for completion in November 1957.

Construction work on the facility for the ETU was
under way in the experimental engineering building.
The ART facility in which the second reactor
assembly was to be tested was, as stated above,
completed and ready for the installation of process
piping. A summary description of the construction
of the facility is given in a subsequent section.
Plans had been completed for training operating
crews, and operating procedures were being pre-
pared. These items are also described in a sub-
sequent section.

CANCELLATION OF FABRICATION AND
CONSTRUCTION ACTIVITIES

W. F. Boudreau

Arrangements were made, immediately after
notification was received to cancel work on the
ETU and the ART, to halt construction and fabri-
cation activities in the Oak Ridge area and at the
various vendors’ plants. All work was terminated
in a manner that preserved tooling, materials, and
partially completed components in a condition
suitable either for the future resumption of the
work or for adaptation to other uses. The ETU
and the ART facilities were put into standby con-

dition pending a determination of their usability

for other purposes.

PERIOD ENDING DECEMBER 31, 1957

Termination reports have been prepared which
summarize the reactor and facility construction
activities. The first of these, which describes
the construction and procurement work for the ETU
and ART reactors and the ETU facility, has been
essentially completed except for material which is
to be prepared by the vendors of heat exchangers,
radiators, and other equipment. In addition to
descriptions of the various problems encountered
in construction, assembly, and procurement, and
the solutions that were found for them, this report
will contain a brief analysis and summary of
expenditures for the various phases of the work.

A summary report on the planning for removal
and disassembly of the ART has also been pre-
pared.2 This document contains a great deal of
information that can be expected to be of value to
other projects, in particular, sections concerning
methods for making accurate dimensional measure-
ments on highly radicactive parts and methods for
dissecting a relatively large and complex object
by remote means.

SUMMARY OF ART AND ETU
OPERATIONS PLANNING

D. B. Trauger

W. B. Cottrell S. M, DeCamp
C. W. Cunningham R. E. MacPherson

The ETU-ART Operations Group was established
in the spring of 1956 to review the reactor flow
sheets and general design from an operational
standpoint, to prepare preliminary procedures for
reactor operation, and, eventually, to ftrain per-
sonnel and to operate the reactors. To augment
the few people who could be made available for
this work and to utilize experience and information
from the wvarious groups within the project, a
committee was formed to assist in the preparation
of procedures. The Operating Committee studied
ART and ETU operational problems and conditions,
and preliminary drafts of ART operating procedures
were prepared. The initial procedures constituted,
essentially, a basic philosophy of operation. With
the pattern thus established, the activities of the
Committee were completed, and the Operations
Group went forward with the task of preparing de-
tailed procedures.

 

2A. A. Abbatiello and F. R. McQuilkin, ART Removal
and Disassembly, ORNL-2464 (in preparation).

 
ANP PROJECT PROGRESS REPORT

The reactor flow sheets and control circuitry
were used as the bases for the detailed operating
procedures. At the time of cancellation of the pro-
gram, most of the flow sheets were complete and
up to date. Systematic tabulations of various
components, such as thermocouples and valves,
were made to clarify the requirements and to check
the flow sheet nomenclature. An instrument tabu-
lation was prepared which specified the service and
requirements that every instrument was to fulfill.
of procedures served as a final
check on the over-all design. The procedures
prepared first included reactor disassembly,
maintenance, operation of certain auxiliary equip-
ment, pump startup, and filling and draining of the
reactor, The studies of procedures led to numerous
design changes; for example, the review of dis-
assembly requirements alone revealed 45 necessary
or desirable changes, while the fill-and-drain
system studies resulted in the elimination of un-
necessary equipment. Particular attention was
given to radiation levels and other conditions to
be expected where maintenance was likely to be re-
quired during operation. At the time of termination

Preparation

of the project, consideration had been given to
procedures for most of the normal operating situ-
ations of both the ETU and ART.3~% In addition
to the procedures pertaining specifically to sched-

 

3ETU Operations: Status Report, ORNL CF.57-10-7
(in press).

4ART Operations: Status Report, ORNL CF-57-10-9
(in press).

SETU Operating Procedures, ORNL CF-57-10-8 (in
press). :

SART Operating Procedures, ORNL CF-57-10-10 (in
press).

vled reactor operation, separate procedures were
to be prepared for servicing and maintaining in-
struments and other equipment and for complex
maintenance, such as the replacement of a pump or
a radiator,

Procedures for an operation such as the ART
could not be developed in such detail that an
operator unacquainted with the device could be
expected to perform a safe and satisfactory oper-
ation. Therefore, a program of operator training
was planned to provide the necessary background.

SUMMARY OF ART FACILITY CONSTRUCTION
F. R. McQuilkin

Lump sum contract work on the ART facility
was completed. This work was contracted in four
stages, as determined by the design program.
The first stage, package 1, included major building
alterations, additions to existing buildings, cell
installation, installation of main air duct, instal-
lation of a 1500-kva substation, and installation
of 480-v main switch-gear. The second stage,

package A, included installation of auxiliary
service piping. The third stage, package 2, in-
cluded installation of diesel-generators and

facility, electrical motor control centers, spec-
trometer room electrical system, and the spectrom-
eter air-conditioning system. The fourth stage,
package 3A, included installation of an electrical
system for supplying power to the pipe and equip-
ment heaters, a dry-air plant and facility, NaK
pump motor controllers, a lube-oil fill-and-waste
system, and a hydraulic system for louver operation.
Tables 1.1.1, 1.1.2, and 1.1.3 present pertinent
information concerning all ART lump sum contract
work performed at the 7503 building.

Table 1,1.1. Contract Cost Summary

 

Total Lump Sum Contract Cost

Contract AT-(40-1)-1952
Package 1, phase |
Package 1, phase ||
Package A

Contract AT-(40-1)-1954
Package 2
Package 3A

$1,096,390.42

881,852.25
$607,526.68
274,325.67
50,351.62
214,538.17
71,557.67
142,980.50

 

 

 
PERIOD ENDING DECEMBER 31, 1957

Table 1,1,2. Controct AT«(40-1)«1952 Summary

 

Total Contract Cost $881,852.25
Phase | — package 1 $557,175.06
Phase Il — package 1 274,325.57
Package A 50,351.62
Package 1
Initial total contract cost 765,835.00
Phase |, construction of 501,462.00

additions to Building 7503

Phase |Il, construction of 264,373.00
7503 cell (operational)

Final total contract cost 831,500.63
Phase | 557,175.06
Phase 11 274,325.57

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 | $100/day
Phase Il $300/day

Liquidated damages charged None

Reason liquidoted damages Contract deviations plus equipment
charge not invoked delivery delays resulting from

Westinghouse strike
Prime contractor VY. 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
Final contract price
Official starting date

Official scheduled completion

date
Actual completion date
Prime contractor

Major subcontractors
Electrical

Mechanical
Designed by

Approximate number of drawings

$50,351.62

4-10-56

6-29-56

9-26-56

V. L. Nicholson Co.

Davis-LLongsworth Electrical Corp.

John F. Humphrey Co.
ORNL Engineering Department

18

$ 50,351.62

 

Table 1,1.3.

Contract AT«(40-1)-1954 Summary

 

Total Contract Cost
Package 2

Package 3A

Package 2
initial contract price
Final contract price
Total number of contract deviations
Official starting date

Official scheduled completion dates
Part 1
Part 2

Actual completion dates
Part 1
Part 2

Liquidated damage rate

l.iquidated demage charged

$ 72,557.67

142,980.50

58,400.00

8

1-2556

6-22-56
1-27-57

6-22-56
2422-56

$100/day

None

$214,538.17

71,557.67
PERIOD ENDING DECEMBER 31, 1957

Table 1.1.3 (continued)

 

Reasons liquidated damages not

invoked

Contractor

Major subcontractors
Electrical

Mechanical
Designed by

Approximate number of drawings

Package 3A

Initial contract price

Final contract price
Capital expenditures

Operational expenditures
Total number of deviations
Official starting date

Official scheduled completion dates
Initial
Final
Actual completion date
Liquidated damage rate
Liquidated damage charged
Contractor

Major subcontractors
Electrical

Mechanical
Designed by

Approximate number of drawings

Government-furnished diesel-generators
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,

Rentenbach Engineering Co,

Broodway Electrical Service Co.

Hickselngie Co.
ORNL Engineering Department

15

$114,434.00

118,394.71
24,585.79

6

12-7-56

4-25-57
8-21-57

8-20-57
$100/day
None

Rentenbach Engineering Co.

Broadway Electrical Service Co.

Hicks-Ingle Co.
ORNL Engineering Department

34

$142,980.50

 
ANP PROJECT PROGRESS REPORT

A pictorial resume of the facility construction
activity is presented with the following series of
pictures which were taken at various stages of
the construction program: Figure 1.1.4 shows a
portion of building 7503 as it existed at the start
of construction of the ART facility, When this
picture was taken in September, some remodeling
had been started. Figure 1.1.5 is an early view
looking into the south side of the building. In the
foreground can be seen the concrete reinforcing
steel for the absorber pit and the cell foundation,
In the background can be seen the forms required for
the control-tunnel wall, Figure 1.1.6, which was
taken in January 1956, shows the structural steel
required for additions to the building. On the
left can be seen the switchhouse. The foreground
shows the stack foundation and the top of the
absorber pit. Figure 1,1.7, which was made in
January 1956, shows the early stages of cell and
water vessel fabrication. Figure 1,1.8, also made
in January 1956, shows the cell foundation and the
first portion of the water tank that was installed.
Figure 1.1.9 shows the status of the cell instal-
lation during the early part of March 1956. Figure
1.1.10 is a close-up view of the water vessel and
cell which was taken in the latter part of March
1956. In the foreground can be seen the cell and a
portion of the water vessel. In the background is
the top section of the water vessel. Figure 1,1,11
was photographed during vacuum leak-testing of the
cell in June 1956. It shows the cell water vessel
in place as it would have been during ART oper-
ation. Figure 1,1,12 is a view looking into the
main air duct. It was photographed in June 1956.
Figure 1,1.13, which was also taken in June 1956,
shows the 480-v switchgear. Figure 1.1.14, which
was taken in October 1956, shows the cell after
the head had been removed and placed in a storage
area. This view represents the status of the cell
after contract work was completed. Figure 1.1.15
was photographed in May 1956. |t shows a portion
of package A service piping in the auxiliary equip-
ment room. Figure 1,1,16 was photographed in
April 1956 and is a view looking at the west side
of the building. In the right foreground can be seen
the foundation for the diesel-generator building,
which was installed under the package 2 contract.
Figure 1.1.17 is a view of one of the two sections
of the motor control center installed by the package
2 contractor, This was photographed in June 1956.
Figure 1.1.18 is a view showing the five diesel
generators which were installed by the package 2

10

contractor. This was photographed in July 1956.
Figure 1,1.19, which was photographed in April
1957, shows portions of the electrical heater dis-
tribution panels which were installed by the package
3A contractor, Figure 1.1.20, which was also
photographed in April 1957, shows the dry air
station which was installed by the package 3A
contractor. Figure 1,1.21 was also photographed
in April 1957 and shows the regulator pit in the
north basement and some of the induction voltage
regulators installed by the package 3A contractor.
Figure 1.1.22 shows the NaK pump motor controllers
which were installed by the package 3A contractor.
Figure 1.1.23 was photographed in August 1957.
It shows a portion of the louver hydraulic operating
system. Figure 1.1.24, which was photographed
after construction activity ceased in August 1957,
is a view looking south toward the building. In
the foreground can be seen the diesel-fuel storage
tank which was installed by the package 1 con-
tractor.  The transformer in the center of this
view was also installed under the package 1 con-
tract. On the right can be seen the '‘Braden’’ type
building which houses the five diesel-generator
It was constructed by the package 2 con-
tractor. On the left can be seen the process water
tank which was part of the existing ARE facility.
Figure 1.1.25, which also was photographed in
August 1957 after construction had ceased, is a
view looking toward the southwest corner of the
building. In the left background can be seen the
diesel-generator house and in the left foreground is
the compressor house. In the center is the blower
house. In the right foreground is the vent house,
stack, and top of the absorber pit. The generator
house was installed by the package 2 contractor;
the compressor house was installed by the package
3A contractor; and the blower house, vent house,
absorber pit, and stack were installed by the
package 1 contractor.

units.

ORNL forces had painted the spectrometer tubes,
installed tracks for the collimator dolly and the
neutron spectrometer dolly, fabricated one set of
heat barrier doors, vacuum leak-tested the lube-oil
waste system, installed the unit heater in the
switchhouse, and painted miscellaneous piping
and structural steel,

A termination report on the construction of the
ART facility has been prepared.’

 

7F. R. McQuilkin et al,, Termination Report for Con-
struction of the ART Facility, ORNL-2465 (in press).

 
PERIOD ENDING DECEMBER 31, 1957

T ¥ UNCLASSIETED
PHOTO 15413

UNCLASEIE|ED
PHOTO 15844

 

Fig. 1.1.5. Early Work on Absorber Pit and Cell Foundation.

11
ANP PROJECT PROGRESS REPORT

12

# T
§ -

S

"

e

AT

Fig.- 1.1.7. Cell Fabrication. January 1956.

 

 

Ly . Y
R ik e

 

URCLASSIF|ED
FHOTD 16189

 
 

    

: *"‘E:} e
L . s -;-_ _'
o } T fi !
ST G

  
 

i
e

    

HCLASSIFIED
PHOTD 16314

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
PHOTO 14317

 

Fig. 1.1.8. Cell Foundation and Portion of Water Tank. January 1956.

13

 
 

ANP PROJECT PROGRESS REPORT

e UNC L ASS I IED
F | PHOTO 14504
>

T
-s

Ff

 

ORCLASSFIED
PHOTO 17254

   

s 9

— "% | _-; |
ATSY ke N

Fig. 1.1.10. Status of Cell Installation in Late March 1956.

 

14

 

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIF IED
FHOTO 17980

 

 

- UNCLASSIFIED
= PHOTO 17764

 

 

 

Fige 1.1.12. Status of Main Air Duct in June 1956.

15
 

 

 

UNCLASTIFIED
PHOTO 17988 |

ANP PROJECT PROGRESS REPORT

 

== [ aafe o N Moaf s

q“.__ -l il n.%

I.ll-..-. - .llt.- #ll-_..-' \

 
          

 

 

 

 

 

 

 

 

 
   

-

 

Fg
Lomell e el l

i i =]
g (g T ] L

 

 

 

-
r————
it F

o ¥

- i

LS

LIHCL ASSIFIED

PHOTC 10872

Inside of Cell. QOctober 1956.

Fig. 1.1.13. The 480-v Switchgear Installation in June 1956.
Fia 1114

 

16
PERIOD ENDING DECEMBER 31, 1957

LMCLASMIFIED
PHOTO 17725

 

- = R . .r'_ N : 3 T -"--.'.' ™ ._.‘,. 5 - _I-Ah-lma
o« SO - ey e . s R : T
. 5 5

Fig. 1.1.16. View of Building Showing Foundction of Diesel-Generator Building in Right Foreground. April 1956.

17

 
 

  
   
 
  
   

ANP PROJECT PROGRESS REPORT

S UNCL ASSIFIED
PHOTO 17948

   
 

 

 

 

 

Fig.- 1.1.17. A Section of the Motor Control Center. June 1956.

UNCLASLIFIED -
® PHOTO 18154

Fig. 1.1.18. Diesel Generators. July 1956.

18

 
PERIOD ENDING DECEMBER 31, 1957

- — EURCLASSIFIED
- | . - ke - = H B PHOTO £0710

 
 
 

 

P il || |||| fi
§ H iy -
2y
Fig. 1.1.19. Electrical Heater Distribution Panels. April 1957.
- T

 

Fig. 1.1.20. Dry-Air Plant. April 1957.

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
PHOTD 40712
k|

 

Fig. 1.1.21. Remotely Controlled Induction Voltage Regulators. April 1957.

 
PERIOD ENDING DECEMBER 31, 1957

PHOTO 40711

2
L
w
5
vy
<
=3
LU
=
-

April 1957.

Fig. 1.1.22. NaK Pump Motor Controllers.

 

 

21

 
ANP PROJECT PROGRESS REPORT

 

 

22

Fig. 1.1.23. Portion of Louver Hydroulic Operating System. August 1957.

 

 

 

 

 

- -,
St
ol a4
T
A
e
= o -
o -—
B
\ i -

L —
——

i _.
o o

UNCLASSIFIED
PHOTO 41476
PERIOD ENDING DECEMBER 31, 1957

§ UNCLASSIFIED

4 7]

 

Fig. 1.1.24. Stotus of Facility as of August 1957. View locking south.

 

F——-—; I AT fi——‘-— e e

 

 

Fig. 1.1.25. Status of Facility as of August 1957. View looking toward southwest corner of building.

23

 
ANP PROJECT PROGRESS REPORT

1.2. COMPONENT DEVELOPMENT AND TESTING
H. W. Savage

The experimental engineering activities as-
sociated with the development and testing of
components for the ART and ETU systems are
being terminated gradually with the objective of
obtaining the maximum information that 'will be
of value in further aircraft reactor developments.
Equipment and test stands for which no more
essential tests were planned or from which the
data to be obtained were applicable only to the
ART or ETU have been disassembied. Parts that
could be used elsewhere were salvaged and
stored. All outstanding orders for equipment in
this category were cancelled.

Test components that had been completed in
the shop, or almost completed, from which in-
formation of general interest can be obtained
are being tested as test facilities become avail-
able. Similarly, tests that were under way at
the time of cancellation of the ART-ETU program
will be completed if they provide data of value
to other ANP programs. Summary reports on the
major test programs are being prepared as the
work is completed.

A forced-circulation niobium-lithium loop is
being designed in which to evaluate system
components and metallurgical developments, Con-
struction of the loop will be initiated when the
niobium becomes available., Thermal-convection
loops will also be operated. A development
program will be carried out on valves, bearings,
pump seals, and the other components that will
be required in an operating system. Secondary
programs will include the testing of equipment
for the continuous processing of lithium to
achieve and maintain the required purity in an
operating system.

PUMP DEVELOPMENT TESTS
W. F. Boudreau  A. G. Grindell
Bearing, Seal, and Lubricant Tests

D. L. Gray

The tests described previously' that were being
conducted in order to establish a detailed pump
sparging procedure for use during ETU and ART

 

p. L. Gray and W, K. Stair, ANP Quar. Prog., Rep.
Sept. 30, 1957, ORNL-2387, p 32.

24

operation were completed. Detailed reports
covering these tests and previous bearing and
seal developments are being prepared.

As mentioned in the previous reporf,] the re-
actor fuel pump rotary element that was being
operated in a new gamma-irradiation facility at
the MTR was shut down for repair of a defective
ion chamber. The chamber was found to be in-
operative because of the deterioration of the
Teflon used to insulate the high-voltage element.
The system has now been modified so that dose
rates can be determined from flux readings taken
external to the test capsule. A subsequent attempt
to resume operation of the rotary element was
halted by ruptures of two pressure-measuring
devices. Inspection has revealed that the hy-
draulic drive system has failed somewhere inside
the sealed unit, but it is not yet known whether
the hydraulic line is plugged or the motor is
defective. The unit will be examined and repaired.
Another attempt will be made to resume the ex-
periment when the repairs have been made.

Aluminum North Head Water Tests

J. W. Cooke?

Terminal reports3+4 are being prepared that

will give details of the water tests of the ART
fuel and sodium systems. These tests were
performed with the use of an aluminum mockup

of the ART north head.

Fuel Pump High-Temperature-Performance Tests
P. G. Smith  W. E. Thomas

The loop in which high-temperature performance
tests are being run of an ART-type fuel pump
with Nc‘1[=-ZrF4-U|:4 (50-46-4 mole %, fuel 30) as
the circulated fluid operated throughout the
quarter and has accumulated a total of 3350 hr
of continuous operation. The loop is presently

 

2On assignment from Pratt & Whitney Aircraft,

3J. W. Cooke, Terminal Report of the Aluminum North
Head Twin MF Pump Water Test Stand, ORNL CF-57-12-4

(in press).

4J. W. Cooke, Terminal Report of the Aluminum North
Head Twin MN (Moderator Coolant) Pump Water Test
Stand, ORNL CF-57-12-5 (in press).

 
running with the pump suction pressure reduced
sufficiently to induce cavitation so that the effect
on the impeller can be determined. The pump
has operated at 1200°F for 1200 hr at the cavi-
tation conditions (2.5 psig pump surge pressure,
645 gpm, and 2700 rpm), and the test is scheduled
for termination at the end of 3000 hr under these
conditions.

An identical fuel pump was placed in operation
with NaK (56-44%) in order to obtain information
on the effect of the type of fluid being pumped.
Preliminary data indicate that the head is 2 to
4 ft higher than in water and fuel tests. A total
of 350 hr of continuous operation of this pump
has been accumulated. Investigations are under
way to determine the cause of the difference in
head, and, in further tests, cavitation effects
will be studied.

Fuel Pump Endurance Tests
P. G. Smith

Operation of the fuel pump endurance test was
resumed, and 860 hr of additional operation was
accumulated to bring the total operating time to
4410 hr. The test was terminated when leakage
of oil from the lower seal exceeded prescribed
limits. Examination of the pump indicated that
replacement of the lower seal would have per-
mitted further operation; however, the tests of
this pump are not to be resumed. A report on the
details of the test operation of the pump has
been issued.’

Primary and Auxiliary NaK Pump Development
‘ D. L. Gray

Operation of primary NaK pump loop No. 1
at elevated temperatures was resumed in order
to observe the effects on the lower seal area
of operating without a purge flow down the shaft
annulus.  Shortly after startup, the NaK in the
loop had to be drained so that a leaking drain
valve could be replaced. The loop was then
restarted and operated for approximately 1172 hr
prior to a scheduled shutdown on November 7,
1957.  Operation of this loop will be resumed
for testing preheating procedures and for tests
at temperatures above 1500°F.

Data on electromagnetic flowmeters were ob-
tained in primary NaK pump loop No. 2, and at

 

Sp. G. Smith, MF Pump Endurance Test — 860 Hours,
ORNL CF-57-123 (Dec. 31, 1957).

PERIOD ENDING DECEMBER 31, 1957

the same time, a test was started on a split-purge
arrangement for the shaft annulus in an attempt
to obtain long-time trouble-free operation of the
lower seal. The helium purge supply is brought
directly to the pump shaft annulus and a portion
of the gas flows down and through the pump tank
and into the off-gas system. The remainder of
the gas flows upward into the catch basin to
perform the function of sparging the lower seal
oil leakage. A scheduled shutdown to remove
the flowmeters halted the split-purge test after
953 hr; the rotary element is to be moved to the
auxiliary NaK pump test stand No. 1 and the
test will be resumed. Approximately 8700 hr
of hot testing with NaK FLave been accumulated
in the two primary NaK purip test stands.

On August 16, 1957, an auxiliary NaK pump
with a Fulton-Sylphon bellows type of lower
seal was installed in auxiliary NaK pump loop
No. 2 for testing. When the test was stopped
for a scheduled shutdown on November 5, approxi-
mately 936 continuous hours of operation had been
accumulated. The seal had operated successfully
with an average leakage of 16 cm3/day. Testing
of this pump will not be resumed.

Performance tests were completed on the
auxiliary NaK pump operating in loop No. 1.
The pump in this loop had operated approximately
2048 hr with an estimated 200 hr ot 1500°F. The
pump has been removed from this test stand and
is being replaced with the pump from the primary
NaK pump loop No. 2, except that the original
auxiliary NaK pump impeller will continue to be
used. A total of 5100 hr of operation at elevated
temperatures with NaK has been accumulated on
the two auxiliary NaK pump test stands.

Sodium Pump Endurance Tests
W. F. Boudreau  P. G. Smith

The Inconel thermal barrier that was designed
to protect the bearing-housing region of the ART
sodium pumps from thermal and radiation damage
was subjected to tests in the sodium pump en-
durance test stand. The test was performed with
sodium at 1250°F and with the pump rotating
at 3500 rpm. For this test, the expansion tank
cover and the instrumented section of the pump
barrel were insulated.

A typical set of temperature data obtained
from the pump barrel and from the barrel and
barrier arrangement is presented in Fig. 1.2.1.

25
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TC's 72-64

THERMOFRAX

INSULATION (3, -in. THERMOGLASS THERMOCOUPLE NO. | TEMPERATURE (°F) )
o BACKED WITH ALUMINUM FOIL) T 1252
o 14 770 -
= 15 938
o Yy in. —= 16 430 .
Yig in. —=] 17 240
56 275
A 57 295
PUMP BARREL (INCONEL) . x | =5 i
% | I 59 365
! THERMAL BARRIER //x | 60 450
5 61 475
° ) | A 1 TC l?l < TC'S 60-56 62 525
| N 1 -in. SPACING 63 560
| 64 640
/>< | 65 652
INCONEL / -TC 16| 66 672
SLUG l"éx | 2; ;‘23‘23
| > *
< \ L : & TC'S 64-60 fg :50
@ ] | = Ya-in. SPACING - 65
o« x 812
TC 14| ¢ 72 a5
| f -
¥

INSULATION

K
|

 

l Ya—in. SPACING

 

2 in. OF SUPEREX tNSULATION

R

 

 

SR

5/64 in, ——

 

 

 

 

 

 

N\

SODIUM LEVEL\

N

EXPANSION TANK COVER (INCONEL)

 

 

 

 

X TC {§

 

 

 

 

T

NOTE :
NOT DRAWN TO SCALE

Fig. 1.2.1, Sodium Pump Thermal Barrier Test.

26

 

 
 

The temperature data are plotted in Fig. 1.2.2
to give the temperature profiles along the outside
and the inside of the pump barrel.

The test system was operated for 400 hr, and
the test was terminated because of an apparent

high leakage rate from the lower seal. The test
was not resumed because of termination of the
ART program.

This test stand was equipped with a lubricating
and cooling oil system which simulated that to
be used in the ETU. The results of tests on this
system are described in a separate report.® This
system was tested for a total of 3800 hr of which
2600 hr were without interruption.

UNCLASSIFIED
ORNL-LR-0WG 27004

o
)
p

 

®,  _—INSIDE OF BARREL

P

 

 

 

 

 

 

 

 

 

N
CUTSIDE OF BARREL— "

1 N

* N

‘e

‘e
°
‘e
.,
0 -
0 200 400 600 800 1000

DISTANCE ABOVE EXPANSION TANK COVER (in.)

TEMPERATURE (°F)

Fig. 1.2.2. Sodium Pump Barrel Temperature Profile.

REACTOR COMPONENT DEVELOPMENT TESTS
Heat Exchanger and Radiator Development Tests

R. E. MacPherson J. C. Amos
R. L. Senn D. R. Ward

The heat exchanger and radiator development
testing program is currently being concluded
with a thermal cycling test of a 20-tube semi-
circular heat exchanger designed to simulate
the ART main heat exchanger tube geometries.
The heat exchanger has completed 155 cycles
of the proposed 300-cycle program. A report is
being prepared that will summarize the results
of the over-all ANP heat exchanger development

program. 7

 

¢p, G. Smith, Simulated ETU Reactor Pump Lube and
Cooling Qil System Test, ORNL CF-57-10-60 (Oct. 10,
1957).

7J. C. Amos et ul., ANP fieut [Ixchanger Development,
1955 Through 1957, ORNL-2439 (in preparation).

PERIOD ENDING DECEMBER 31, 1957

The radiator development program was concluded
with the termination of ART test radiator No. 2
because of failure of a radiator tube after a
total of 181 thermal cycles. The number of cycles
imposed on this radiator was six times the
maximum number of thermal cycles proposed for

the ART radiators.

Details for testing of this radiator and ART
test radiator No. | are presented in a spearate
report.8  Details of the development testing of
smaller 500-kw units of the same basic design
have also been summarized in a separate report.

Outer Core Shell Thermal Stability Tests

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

52 psi.

A third quarter-scale core shell is being fab-
ricated by welding together oversize sections
of a complete shell and bringing the shell to
proper dimensions by final machining. This shell
will be subjected to thermal-cycling tests similar
to those described previously in order to obtain
data on the structural reliability of the welded
shell This method of shell con-
struction was one of those proposed as a means
of meeting the rigid dimensional tolerances re-
quired for the ART core shells.

construction.

Fabrication of a shell-buckling rig for testing
full-scale shells is 95% complete. This apparatus
will be shipped to Pratt & Whitney Aircraft, where
tests will be made on ART core shells.

 

8). C. Amos and D. R. Ward, Performance and En-
durance Tests of ART NaK-to-Air Test Radiators Nos, 1
and 2, ORNL CF-58-1-2 (in press).

?R. E. MacPherson and R. J. Gray, ART Prototype
Radiators — Development Program and Metallurgical

Results, ORNL-2436 (in preparation).

27
ANP PROJECT PROGRESS REPORT

Zirconium Fluoride Vapor Traps

D. B. Trauger J. A. Conlin
A. G. Smith'®

Developmental work on Zer vapor traps for use
in the ART system has been terminated. The
trap, shown in Fig. 1.2.3 and described pre-
viously,'! was tested and found to be satisfactory
for ETU operation with the addition of a conical
inlet to reduce heat losses by convection and
radiation to the cooling tubes and the inclusion
of a retainer web to prevent filter material from
being blown out of the trap.

Additional modifications were to be made to
assure that the unit would be satisfactory for

 

105, assignment from Pratt & Whitney Aircraft.

”M. H. Cooper, ANP Quar. Prog. Rep. March 31,
1957, ORNL-2274, p 61.

)

1}

/
|

1
"lll

——
I
o

—

—
——
e

—
——
—
-
—
——

INLET COIL FOR
FIRST SECTION

ART operation. Further, it was essential to
final design considerations that the quantity,
density, and conductivity of the expected deposits
be determined precisely. The quantity and density
would affect the trap size, ond the conductivity
would aoffect the size of the tubes in the second
section. A detailed report on the vapor trap
developmental work is being prepared.'?

IN-PILE EXPERIMENTATION
D. B. Trauger  J. A. Conlin

An in-pile loop in which ART operating con-
ditions could be simulated was designed, and
three such loops were operated as supporting
research for the ART program. The loops provided

information needed for final choice of the

 

 

12M. H. Cooper et al., Zirconium Fluoride Vapor
Traps, ORNL-2245 (in preparation).

UNCLASSIFIED
PHOTO 28575

HEAT EXCHANGER
FOR SECOND SECTION

Fig. 1.2.3. Component Parts of ZrF, Vapor Trap.

28

 
materials of construction and fuel compositions
proposed for the ART. The three loops were
operated in the MTR with varying degrees of
success. The results indicated no measurable
effect of radiation on the corrosion resistance
of the structural material or on the stability of
the fuel. Some work in support of an advanced

PERIOD ENDING DECEMBER 31, 1957

ART reactor design was being considered at
the time of the termination of the ART program.
This work will be described in a forthcoming
repor\‘.13

 

]3D. B. Trauger et al., Circulating Fluoride Fuel
In-Pile LLoop, ORNL-2444 (in preparation),

29
ANP PROJECT PROGRESS REPORT

1.3. INSTRUMENTATION AND CONTROLS DEVELOPMENT

E. R. Mann

ART CONTROL ROD DRIVE TEST
C. S. Walker

The test of the ART control rod drive system,
as stated previously,! was terminated after 3000 hr
of successful operation with the control rod thimble
at or above the design temperature of 1100°F.
The drive mechanism operated through 1780 full
stroke cycles.

Before the termination of the test on September 17,
1957, electrical leakage resistance measurements
were taken with the system temperature maintained
as in the test in order to determine whether sodium
deposition had reduced insulation resistances.
The leakage resistance to ground and the leakage
resistances between isolated conductors which
penetrated the tank that enclosed the rod drive
mechanism were measured. All these leakage
resistances were above 4 x 106 ohms and were
acceptable in each case. The lowest resistance
measured between the open contacts of the limit
switches exceeded 107 ohms.

The sodium was drained from the control rod
thimble, and the drive mechanism was completely
dismantled and carefully examined. = There were
no indications of wear; the backlash in the gears
had not increased beyond that in the original
assembly, No sodium deposition could be found
on the rod drive parts that were located above the
surface of the molten sodium. The speed of
operation of the rod had not changed during the
test, and thus there was apparently no binding or
increase in friction between moving parts.

The heat exchanger which surrounded the thimble
at the upper surface of the sodium was found to
have a deposit of scale on all areas that were
in contact with water. This deposit was 0.018 in.
thick, and chemical analysis indicated it was
practically all calcium, with small quantities
of magnesium and zinc. A porous ceramic water
filter in the effluent water line was found to
contain a sizable quantity of rust.

A chemical analysis of the oil in the gear box
showed no significant amount of sodium or rare
earths. This further indicates that no deposition

 

TE. R. Mann and C. S. Walker, ANP Quar, Prog. Rep.
Sept, 30, 1957, ORNL-2387, p 54.

30

C. S. Walker

R. G. Affel

of sodium occurred on the cool portions of the drive
machinery.

The only troubles encountered during the test
stemmed from equipment external to the rod drive
itself; namely, difficulty with regulation of the
water flow through the sodium-to-water heat ex-
changer, and faults in the temperature recording
apparatus. The rod drive mechanism successfully
completed the test.

Samples of the sodium were taken for analysis
and the control rod was disassembled for ex-
amination. Details of the results of the sodium
analysis and rod examination are presented in
Chap. 3.4 of this report.

ART FUEL-EXPANSION-TANK CONTINUOUS
LEVEL INDICATOR

R. F. Hyland

In the early stages of ART design it became
apparent that a method would have to be developed
for continuous measurement of the level in the re-
actor fuel expansion tank. The level indicator was
required to function continuously in a fuel
30-helium environment in an Inconel container in
the temperature range of 1000 to 1500°F and a gas
pressure of 15 psigs The range of level indication
was 0 to 5 in. The indication was to be accurate
to 12% (percentage of maximum level) and
reproducible to +1%. The minimum life of the
device was to be 1000 hr of continuous operation.

Any portion of the device that was above the
fuel level in the expansion tank was to be cooled
because there would be severe beta heating during
power operation of the reactor. The device was to
be operable under the extremely turbulent fuel
flow conditions that would exist during power
operation. Turbulence of the type to be expected
is illustrated in Fig. 1.3.1, which presents two
pictures taken during water tests of the aluminum
north head mockup. Penetrations through the top
of the north head were to be kept to a minimum
because of the complexity of the system in that
area and the number of layers that would have to be
penetrated to reach the fuel expansion tank, as

 
UNCLASSIFIED
PHOTO 42767

—ve’ 2

 

 

UNCLASSIFIED
PHOTO 42788

    

Fig. 1.3.1.
Aluminum North Head Mockup During Simulated Power
Operation with Water as the Circulated Fluid.
with coption)

Two Views of Fuel Expansion Tank of

(Secret

shown in Fig. 1.3.2. The level sensor and any
seals or stuffing boxes used were to be leaktight,
as determined with @ mass spectrometer, to
preclude leckage of radicactive gases.

An appraisal was made of known level-measuring
techniques, and it was apparent that, because of
its basic simplicity, the conventional purged-
dip-tube level-measuring device offered a possible
solution to the problems. Penetrations through the
top of the reactor north head could be avoided by
entering the side of the fuel expansion tank, as
shown in Fig. 1.3.3, and the beta heating problem
could be avoided by entering the tank below the

PERIOD ENDING DECEMBER 31, 1957

fuel level. The use of @ multiple-pass Heliarc-
welded dip tube penetration would remove the seal
problem, and the use of materials that were not
subject to radiation damage would assure the
lifetime required. On the basis of this study,
static tests with fuel and dynamic tests with
water and with fuel were initiated. These tests
were not completed because of the termination of
the ART experiment; however, a summary of the
the test results obtained is presented below.

Continuous measurement of nonturbulent fuel
levels in the temperature range 1150 to 1500°F
and a level range of 0 to 5 in. with an accuracy of
+2%, a reproducibility of 1.5%, and a sensitivity
of 0.03 in. of fuel was demonstrated. Typical
results of these tests are presented graphically
in Fig. 1.3.4. Dip tube life averaged approximately
1500 hr before plugging occurred. The plugs were
of two general types: zirconium dioxide and other
metallic oxide deposits in the bottom of the dip
tube or fuel vapor deposits in the region of the
dip tube that was below the melting temperature
of the fuel. Several of the zirconium dioxide plugs
were traced to contamination of the helium purge
gas with oxygen and moisture, and it is felt that
this type of plug could be eliminated by controlling
the oxygen and moisture content of the gas. The
fuel vapor deposits were readily melted out of the
dip tubes and could undoubtedly be eliminated
completely by preheating the purge gas. The
all-welded Taylor Instrument Companies model
X333RD differential-pressure level sensor which
was being tested appeared to be satisfactory for
reactor service, contingent upon individual units
passing mass speztrometer leak checks.

Metering tests conducted with the use of the
aluminum north head mockup demonstrated that
with both pumps ot design speed (2700 rpm), the
system at design flow (645 gpm), and the dip tubes
properly placed, changes in level could be detected
with an accuracy of +5%. Determination of an
absolute level under these simulated reactor
operating conditions proved to be impossible. As
shown in Fig. 1.3.1, the water in the expansion
tank was extremely turbulent, and, of course, the
turbulence caused appreciable level oscillations.
Attempts to establish an average level reference by
using high-speed photography were inconclusive
because of the difficulty of photographing the
expansion tank through the small viewing port
provided.

31
ANP PROJECT PROGRESS REPORT

SODIUM FILLING PORT

  

 

HELIUM INLET FOR PRESSURIZING
Na PUMP

SODIUM EXPANSION TANK

UPPER OUTER

SHELL \(

-
y
//

   

-

VENT LINE

    
  
  

LOWER DECK

S
ORNL-LR-DWG 27019
CONTROL ROD SLEEVE

OFF GAS
FUEL PUMP

  
  

UPPER DECK =~
- s

    

e
z

   

Ng PUMP

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

Since the level-measuring system depends upon
the sensing of the fuel pressure, the
indication is dependent on the fluid density. The
magnitude of the density effect is shown in
Fig. 1.3.5. If the reactor operating temperature
of 1150°F is taken as the point of zero error, a
level error range of +2.5 to ~5.7% exists in the
temperature range 1000 to 1500°F. This error
could be eliminated by installing a separate
temperature- (and thus density-) measuring system.
The signal from this system would be multiplied
by the signal from the level-measuring system to
produce a signal corresponding to the true level.

32

level

Considerable work remained to be done in de-
termining the dynamic characteristics of the
measuring system under ART conditions. The
planned test program included installation and
testing in water in the aluminum north head test
rig of shielded dip tubes of the type to be used in
the ART and similar testing.in fuel in the ETU.
Details of the tests described here are available in
a separate repon‘.2

 

2R F. Hyland, ART Fuel Expansion Tank Continuous
Leve! Indicator, ORNL CF-58-2-1 (in press).

 
-

TO ALL WELDED DIP CELL4———\

MEASURED LEVEL ERROR (in. OF FUEL)

PERIOD ENDING DECEMBER 31, 1957

LI ]

 

4
ORNL— LR—- DWG 27020

 

- — FUEL LEVEL

SIDE OF FUEL EXPANSION TANK

 

LEVEL SENSOR TUBE

 

  

 

 

 

 

LEVEL SENSOR SHIELD

 

 

 

 

 

 

 

 

 

 

 

HEAT EXCHANGER

\ —=———— FUEL EXPANSION TANK FLOOR

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

<4
ORNL-LR-DWG 2702t

 

 

0.20

—— LEVEL INDICATOR S1
————— LEVEL INDICATOR S2

 

0.15

 

o
o

 

0.05

 

 

——— DECREASING LEVEL

 

 

 

 

 

 

—=—— |[NCREASING LEVEL

 

MEASURED LEVEL ERROR {% OF MAXIMUM LEVEL)

 

 

 

 

 

1 2 3 4 5 4 3

REFERENCE LEVEL (in. ON CALIBRATING PROBE)

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

33
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a ORNL-L!-DNG 27022
o
w
2 T~
Z o
5, e
- T
"~
S -4 T~
2 ~—
-6
1000 1100 1200 {300 1400 1500

FUEL TEMPERATURE (°F)

Fig. 1.3.5. Effect of Fuel Density, as Indicated by
Fuel Temperature, on Accuracy of Purged-Dip-Tube
Level Measurements.

RESISTANCE-TYPE FUEL LEVEL INDICATOR
T. V. Blalock R. F. Hyland

A resistance-type fuel level indicator was
designed as a possible alternate to the purged
dip tube described in the previous section. The
operation of this level element depends on a
resistance change that is a function of fuel level,
A schematic electrical diagram of the system is
shown in Fig. 1.3.6.

In Fig. 1.3.6, R is the current adjusting po-
tentiometer, R, is the lead resistance, R_ is a
temperature-compensation resistance, R, is the
fuel resistance between the Inconel rod and the
11/2 -in. Inconnglpe (Fig. 1.3.7), R, is the Inconel
tube resistance, r, is a resistance representing
the shorting-out effect of the fluid on the Inconel
tube, and R is the total input resistance of the
parallel section of the level element.

In reference to Fig. 1.3.6, if R >> 2R, « R,
the current I will depend primarily on R and
therefore the level element can be supplied with
a constant current. The output voltage must be
measured with an instrument of high input im-
pedance, such as a vacuum-tube voltmeter or a
Brown recorder. The current flows from the source
through the compensating resistance R _ to junction
B. From this junction the current has three
possible parallel paths. The first path is through
the lnconel rod, the fuel, the ]/ -in. pipe, and
the ! /-in. plate to junction C; the second path is
through the Inconel tube and the /-ln. plate to
junction C; and the third path is through the fuel,
the Inconel tube, and the ]4 -in. plate to junction C.

34

-
ORML-LR-DWG 27023

 

 

 

 

 

 

 

 

 

 

A A A

AAA —AAAA 2 O

yYyy YYYY
15y 3.5y 7
60 cycles o

c
OUTPUT
g {mv)
: 2 =
Ar ,é/ R 2 :F A
AAAR { O

 

 

VVVY Bl

A

Fig. 1.3.6. Electrical Diagram of Resistance-Type
Fuel Level Probe.

As the fuel level rises, the current flow through
the fuel between the Inconel rod and the ll/ -in.
pipe will become less inhibited because of the
increase in the flow-path cross-sectional area.
The resistance of the rod and pipe to this current
flow will be negligible compared with the fuel
resistance. At zero level the element output will
be given by the product of the supply current and
the Inconel tube resistance, since the resistance
R, will be infinite for nonimmersion. The maximum
output will occur at zero level, and the output will
decrease as the level increases. The resistance
R _ is a function of temperature only, and R, and

c
r, are functions of both temperature and level.

The slope of the resistivity curve of the fused
salt mixture is negative; that is, the resistivity
decreases with increasing temperature. The
resistivity of Inconel is almost constant over the
temperature range under consideration (1100 to
1600°F), and therefore R, can be assumed to
be constant. It follows then that R, will be a
function of R and T and will decrease with
increasing temperature.

An increase in fuel temperature will have the
same effect on the output as an increase in level
if the system is not temperature compensated.
One method of temperature compensation is
described below. The element is to be compensated
at a fluid height, b, so that a temperature excursion
from T, to T, will produce no change in output.
If the compenscmng resistance (R_) is placed in
_senes with R the temperature-compensation

 
INPUT e

3.45v, 1amp

GAS SUPPLY
AND VENT LINE —

 

LRI }\\\\\%

 

 

FLUID TRANSFER LINE —s=—

 

 

PERIOD ENDING DECEMBER 31, 1957

L v
ORNL—LR—-DWG 27024

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TEST VESSEL\\

 

 

-

L

— — — — CURRENT LEADS
——————— VOLTAGE LEADS
TEMPERATURE COMPENSATION
RESISTANCE (&)

1 FUEL RESISTANCE (A))

 

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

 

I ¢ OQUTPUT
' {mv)
________ _I I
I I
| I
Y/a-in. PLATE ¢ i
~ TS 3
N 1 /.
P i VENT N
~ VENT HOLES 1| HOLES -
N ; N
INCONEL ROD 1l ?
s |
\ . l ||
N
N | a-in-1PSN
N | SCHED-40N
N INCONEL | INCONEL
- TUBE ——=| § PIPE —=}
N \ |
\ I
N a8 4 /——'“"J
G _ N
\ R, \ | N
N /
B [\

35
ANP PROJECT PROGRESS REPORT

criterion will be satisfied if

 

 

kS |/ v - V
pr—, oy =Py,
(P ) e

T, - T, ) e A
or

AV AV

Rc RT

(2) AT + X 0 for AT=T2-—-T1 .

As stated previously, AV, /AT is negative, and
therefore AV, /AT must The positive and the

resistivity curve of the compensating resistor must
have a positive slope.

A short length of A-nickel wire can be used for
R_and its dimensions can be arrived at as follows:

 

 

Pe
vV = IR_=1 [
R c c !
% c
VRT— IRT G
!E
VR =] T P, at temperature T' ,
1 = 1
;C
v = | at temperature T., ,
Rcz A PCZ P 2
c
‘VRT = IRT1 at temperature T, ,
1
INCONEL PIPE

 

INCONEL TUBE

 

   
      
 

INCONEL ROD

at temperature T2 ;

 

 

R &
'3'12 2
I e i (I = TR )
Po. =l P = T T
T . ] 2
Ec RT] - RTE
(3) T Wy
c Pc2 i Pc,l
where
P, = specific resistance of compensating wire,
[. = length of compensating wire,
A_ = area of compensating wire.
Two resistance-type level elements without

temperature compensation were fabricated (Figs.
1.3.8 and 1.3.9) and are being used to experi-
mentally determine the resistance-level character-
istics of the element and to observe polarization
and other effects not included in the design calcu-
lations. The elements are of a rugged all-welded
design with no seals or stuffing boxes. They are
designed for a maximum fuel temperature of 1600°F
and a maximum allowable working pressure of
5 psig. Calculated performance curves for these
elements, both temperature compensated and
uncompensated, are shown in Fig. 1.3.10. These
curves are based upon a constant current of 1 amp
through the measuring element.

UNCLASSIFIED
PHOTO 29725

INTERNAL STRUCTURE

Fig. 1.3.8. Component Parts of Resistance-Type Fuel Level Probe. (Secret with caption)

36

 

 
Fig. 1.3.9. Assembled Resistance-Type Fuel Level

Probe.

 

(Secret with caption)

UNCLASSIFIED
PHOTO 29726

 

32

28

  
   
   
    

PERIOD ENDING DECEMBER 31, 1957

E——
ORNL-LR-DWG 27025

 

 

! 1

WITH TEMPERATURE COMPENSATION

|
_1600°F
A400°F
_4200°F

 

 

   

 

 

     
   

 

 

 

 

 

24
HOO0°F
CURRENT: 1amp
20 .
|
-
E |
- !
5 16 — —
& WITHOUT TEMPERATURE COMPENSATION
> | '
o |
12 m—_ {200°F
1400°F
1600°F —
I
8 r
| =
| |
3 S 1
|
: |
0 2 4 6 B

FUEL LEVEL (in.)

Fig. 1.3.10. Calculated Performance Characteristics
of Resistance-Type Fuel Level Indicator.

A test rig designed to test two level elements
simultaneously was fabricated in accordance
with the flow diagram shown in Fig. 1.3.11. The
liquid level was varied by a helium-forced ex-
change of fuel between the two vessels. The
initial readout of the level signal was accomplished
with a Hewlett-Packard model 400H vacuum-tube
voltmeter. A sliding probe which completed a lamp
circuit when contacted by the fuel was used for
calibration.

37
ANP PROJECT PROGRESS REPORT

     

Q-160 psig

100-psig
HELIUM

FILTER

 

ORNL-L!-EWG 27026

SCOLENOID VALVE SWITCH MAY BE
REPLACED BY A REPEAT CYCLE
TIMER FOR AUTOMATIC CYCLING.

KEY

 

 

 

P& THROTTLING VALVE

THREE WAY VALVE

\/
N\ DARKENED PORT CLOSED -
DESIGNATES COMMON PORT
VENT v ¢ ° ©
ULIUELECTRIC HEATERS
\/

“A @ 0-2psig
e 0-5 psig

 

 

 

 

ZrFy

- - VAPCR SWITCH "OFF"
\ TRAP 5 POSITION
\\\
TO VACUUM  _ -~ A
TUBE VOLTMETER L v
. CALIBRATION

 

 

 

 

 

VALVES SHOWN IN

 

!

ZrFy

VAPOR .
TRAP E

i S TO VACUUM
: " TUBE VOLTMETER
CALIBRATION

PROBE S
E GAS ’ RESISTANCE PROBE

 

 

 

 

 

 

 

   

TANCE PROB o PROBE
RESISTANCE PROBE -\ EGAS
o — FLUID
g SPARE
o Q
DRAIN  THERMOWELL

VESSEL NO. 1

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

Preliminary experimentally determined per-
formance characteristics of an uncompensated
element are shown in Fig. 1.3.12. The actual
performance of the probe was in reasonable
agreement with the calculated performance. A
family of curves suitable for designing temperature
compensation into the element is now being
obtained.

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-
perienced earlier in the fabrication of these

 

3On loan from Radio Corp. of America.

38

FILL SPARE

     

DRAIN THERMOWELL -

VESSEL NO. 2

transducers that had resulted in internal electrical
shorts and a high rejection rate.

Design work was completed on all the continuous
level probes required for the ETU and ART for use

A
ORNL-LR-DWG 27027

 

>
£
'_
2
a
'-_
2
o
L
m -
o
o
a
.
FUEL LEVEL (in.) -

Fig. 1.3.12. Preliminary Experimental Performance

Characteristics of Resistance-Type Fuel Level Probe.
in the sodium expansion tank, the sodium auxiliary
tank, and the NaK-furnace-circuit drain tank. All
work was cancelled for construction and testing of
these level probes. During this quarter, two level-
element testing rigs were rebuilt that had been
damaged because of NaK leaks in the system
(Fig. 1.3.13), and the test program was resumed in
order, primarily, to determine the error introduced
by an unwetted level probe. Before this type of
level probe is wetted by NaK there is contact
resistance between the surface of the probe and
the NaK that results in the indicated level being
lower than the actual level.

As part of the wetting investigation, three newly
fabricated NaK-pump-bow! tranducers were cleaned
with emery cloth and rinsed with alcohol just prior
to installation in the test facility. The NaK was
heated and then raised to approximately 4 in. on
one of the probes with the NaK at 300°F and on
two of the probes with the NaK at 400°F. The
results of the first level change of approximately
4 in. on each level probe are given in Table 1.3.1.

After the level changes above, the probes were
submerged and calibrated at the test temperature.
With the NaK at 300°F, the indicated level for
the 18 points measured was between (.15 and
0.28 in. below the actual level. For the test with
NaK at 400°F the indicated level for the 18 points
was between 0 and 0.1 in. below the actual level.
Further wetting tests are being made at other
temperatures, and the results will be included in a
final report.

Four level transducers were installed in a life
testing facility. Two of these transducers were
previously operated for 3480 hr and the other two
were operated previously for 2440 hr. These units

Table 1.3.1.

PERIOD ENDING DECEMBER 31, 1957

are shown in Fig. 1.3.14 before installation in the
life testing facility. It is planned to continue the
life tests indefinitely.

ON-OFF LEVEL PROBES
G. H. Burger R. E. Pidgeon, Jr.

Plans for construction and testing of the ETU
and ART on-off level probes described previously?
were cancelled, but similar probes were con-
structed for instrumenting test rigs. These probes
differed from those built previously in that the
copper-and-Inconel junction at the tip of the probe
was brazed instead of being welded in an attempt
to eliminate weld failures. These new probes
were installed in the NaK systems, but they have
not yet operated long enough for an evaluation of
them to be made.

Sixteen on-off probes installed in the gas pots of
two NaK system test loops have each operated for
3000 hr without a probe failure. These probes
have usually operated with the NaK temperature
below 800°F. The control circuit associated with
the on-off probes installed in one test loop failed
because of in-gassing of the pump. Apparently,
during operation of the loop, the NaK level reached
the high-level probe in the throat of the gas pot,
and the control circuit acted to lower the level.
When the level had dropped to below the high-level
probe, the sensitive relay across the probe failed
to energize and allowed helium to be forced into
the loop. The relay failed to operate because the
“pull-in'’ voltage of the relay had increased and
the level probe current was adjusted too low. The

 

4G. H. Burger and R. E. Pidgeon, Jr., ANP Quar,
Prog. Rep. Sept. 30, 1957, ORNL-2387, p 58.

Results of Tests of Especiolly Cleaned Level Prabes

 

NaK Temperature

Time After Initial Level

Level Error Indicated Accuracy

 

(°F) Change (br) {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 Q6.6
21 0.13 9646

 

39

 
 

ANP PROJECT PROGRESS REPORT

 

Fig. 1.3.13. A Pot of o Level-Element Testing Rig Showing Oxide Deposits Resulting from o NaK Leak.

40

 
 

Fig. 1.3.14.
erated for 3480 or 2440 hr.

on-off control circuit of the other test loop has
operated without failure and without maintenance
or readjustment. These level probes and their
circuitry will be discussed in o detailed final
report.

MAGNETIC FLOWMETERS
G. H. Burger C. L. Pearce, Jr3

The 2- and 31é-in. magnetic flowmeter calibration
and test program, as described in the previous
report,® was terminated. The data obtained from
the operation of 20 such flowmeters in pump test
loops are being processed and analyzed, and the
findings will be discussed in a final report.

The loop tests were terminated on November 5,
1957, and the instrument panels have been dis-
mantled. The flowmeters are being removed from

 

0n loan from Radio Corp. of America.

%G, H. Burger and C. L. Pearce, Jr., ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 38.

PERIOD ENDING DECEMBER 31, 1957

UMCL ASSIFIED
PHOTO-29995

Liquid-Metal-Level Transducers Thot Were Installed in a Life-Testing Facility After Having Op-

the loops and will be cleaned and crated for
storage. The total operating times accrued on the
2-in. units are given below:

Acecumuloted Operating

Neo. of Units Time for Each Unit
(hr)
4 1038
4 1916
2 2935

For the 31/5-in. units the operating times were as
follows:

Accumulated Operating

No. of Units Time for Each Unit
(hr)
4 770
4 2368
2 3138

4]

 
ANP PROJECT PROGRESS REPORT

There was no obvious deterioration in performance
of any of the flowmeters during the periods
indicated. |

As mentioned in the previous report,an apparent
drop in sensitivity of the second lot of units
installed in the test loop was observed in con-
junction with plugging of the loop flow-control
valve at low loop temperatures. An attempt to
relate the apparent nonlinearity of the units in
this lot with a high oxide level was unsuccessful,
since the effect remained after the oxide level of
the loop was reduced. It was subsequently found
that a slightly bent pen on the venturi differential-
pressure flow recorder, which employs a square-
root differential-pressure chart, was responsible
for the apparent shift in sensitivity. The bent pen
caused a fixed error in the differential pressure,
which appeared as an error in the square root of
the differential pressure that increased as the
flow decreased. The lower flows, as read from the
chart, showed the effect of the slight shift in pen
position more than the higher flows. The effect
of this error appeared as a shift in sensitivity in
the caleculations. It appears therefore that the
linearity of the units will be better than 1%, as
stated in a previous report. The data which have
been analyzed indicate that all the units that
were calibrated may be expected to provide flow
accuracy well within the specified ART and ETU
requirements over the entire flow and temperature
range prescribed in the flowmeter specifications.
The indications are that, upon completion of data
processing and analysis, an expected accuracy
may be assigned that will be somewhat higher than
that specified.

A total of fifteen 343-in. magnetic flowmeters for
ETU cold traps had been assembled and delivered
for installation in the ETU. The units, as noted
in a previous report, were delivered without the
magnet attached to permit easier handling for
installation in the systems. In addition to these
fifteen units, parts for 15 more units for the ART
were completed but not assembled. All com-
pleted and incomplete units will be packaged and
stored for future disposition.

As noted in a previous report, it was not planned
to calibrate the %-in. units. However, temperature
tests were made on the units to determine magnet
flux changes with magnet temperature and to de-
termine temperatures . of the pipe in the air gap in
order to check for loop cold spots which could

42

cause plugging during operation. It was found that
the flux changed by approximately 3.9% when the
magnet was heated from room temperature to 365°F.
It was also found that, with the insulated section
of pipe at 610 to 1722°F, the section of pipe
through the air gap was at 262 to 680°F when it
was insulated with approximately 2 in. of Fiberfrax
and the flow pipe was held in each mounting clamp
by a series of four corrugated clamps.

Based upon tests on the large flowmeters and
the flux data obtained during the 3/B-in. flowmeter
tests, it was expected that the accuracy of the
units could be predicted to well within the required
+15% specified for ETU and ART operation,
without the need for magnet flux and temperature
monitoring. It was also expected that, with proper
insulation of the magnet air gap, the problem of
plugging because of cold spots could be eliminated
and yet the magnet would remain cool enough for
proper operation at the specified fluid temper-
atures.

TURBINE FLOWMETERS
G. H. Burger

The 1-in. turbine flowmeter designated No.
2-Rev. 3, which was described previously,” was,
as stated previously,® installed in a forced-
circulation gas-fired Inconel loop containing
Nc:F-ZrF4-UF4 (50-46-4 mole %, fuel 30) for
calibration and for an endurance test of 3000 hr.
Operation of the loop was started on January 8§,
1957, and the turbine operated satisfactorily for
about 340 hr. An investigation then showed that
the unit was not operating, but it was not removed
from the loop until the loop was drained. At that
time the loop had operated for 5538 hr at temper-
atures ranging from 1200 to 1300°F. !

An examination of the unit upon its removal from
the loop showed that the bearing shaft on the end
of the large turbine had broken and allowed the
unit to drop and bend; the turbine therefore ceased
to rotate. No oxidation other than that found after
previous operation in NaK was apparent. No
reason for the broken shaft was apparent, but it
is thought that the break was probably an indirect

 

7G. H. Burger, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 25

8. H. Burger, ANP Quar. Prog. Rep, March 31, 1957,

 
result of the small diometer of the shaft and the
brittleness of the material.

Two similar 1-in. flowmeters, designated
No. 4-Rev 0 and No. 5-Rev 0, were calibrated
with water and were installed in a forced-circu-
lation gas-fired Inconel loop containing NaK that
was being operated in order to test pressure
transducers. This test loop is similar to those
used previously except that it can operate with
much higher flow rates. Unit No. 5-Rev 0, which
is shown in Fig. 1.3.15, differed from Unit
No. 4-Rev 0 in that its turbine had 10 blades,
whereas the turbine of Unit No. 4 had five blades.
The results of the water calibration of the units
are presented in Fig. 1.3.16.

The two turbines were installed in series, and
the loop was filled October 2, 1957, at 200°F.
The loop operation was started at approximately
room temperature, and was held at room temperature
for approximately 120hr before the temperature was
raised to 1000°F at a slow rate. Unit No. 5-Rev 0
ceased to operate properly after 162 hr, but Unit
No. 4-Rev 0 continued to operate. Unit No. 5
could be started momentarily by varying the pump

PERIOD ENDING DECEMBER 31, 1957

speed but would cease to operate after a very
short operating period. After the loop had been
operating for a total of 500 hr, the temperature was
again raised with the intention of reaching 1200°F.
At 1150°F, Unit No. 4 also stopped, but the temper-
ature was quickly dropped to 1125°F and the unit

UNCLASSIFIED
ORNL-LR—DWG 27028

| ' i | i I
o UNIT NO.5-REV 0,40 BLADES :
« UNIT NO.4-REVD, 5 BLADES

Az

 

B
O

LA
T

 

Y]
o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OUTPUT FREQUENCY OF FLOWMETER (counts /sec)

80 .
60 ////
% /,-f"i 7 |
"
20 frr‘ !
bt L]
0 { 2 3 4 5 5 T B8 g {0

FLOW RATE (gpm)

Fig. 1.3.16. Calibration Daota Obtained with Water for
Turbine Flowmeters No. 4-Rev 0 and No. 5-Rev 0.

 
 
 

UHC LAHIF!EDE
PHOTO-40854

 

   

UPSTREAM
POSITIONER

2 3 4 5 6

 

 

COWNSTREAM
POSITIONER

    
 

BEARING SHAFT

    

7 8 9 {0

 

 

Fig. 1.3.15. Turbine Flowmeter No. 5-Rev 0.
ANP PROJECT PROGRESS REPORT

again started operating and continued operating
for approximately 100 hr at 1100°F. The loop
temperafure was then again raised, very slowly,
to 1200°F. This time the unit operated in a normal
manner and continued operating until the loop was
shut down after 644 hr for removal of a defective
pressure transducer that was being tested.

During the operating period, the flow rate was
varied from 1.5 to 8.9 gpm. The turbine rotational
speed varied from 300 to 3300 rpm, with 300 rpm
being the lowest running speed of the turbine
because of bearing friction and blade or rotor drag
in the fluid. It was noted during the test that the
output frequency of the unit became increasingly
nonlinear below 4 gpm. This was not unexpected,
since the unit was operating from below to just
within the design flow range of from 5 to 50 gpm
for 1-in. turbine flowmeters. No particular effort
was made to achieve linearity or accuracy to a
high degree, since the major objective of the
tests was to determine the operating life of the
units at various flow rates and temperatures.,

After the loop was drained for removal of the
pressure transducer, it was decided to remove the
turbine flowmeters for examination and repair or
modification, if necessary., The examinations
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
downstream shaft and thus allow the rotor body
to contact the positioner body and cause the unit
to drag and stop. It is evident that increased
temperatures would aggravate this condition be-
cause of the increased softness of the Inconel.
Unit No. 4 did not show this type of wear to so
great an extent, since it was only a five-blade
unit, and it therefore continued to operate.

Based upon the examination of the units, the
bearing radial clearances were increased, the
clearance between the downstream end of the rotor
body and positioner was increased to 0.030 in.
when the rotor was fully positioned in the down-
stream direction, and the number of blades was
reduced to five on Unit No. 5 in order to reduce
downstream thrust. In addition, the over-all blade
diameter was reduced to give greater clearance
between the blades and turbine housing and thus
reduce the chances of jamming as a result of dirt

44

or oxide plugs in the loop. The units were then
reassembled and reinstalled in the test loop
without water calibration, Since the units were
not recalibrated, the accuracy is not expected to
be as good as before, but the linearity should be
about the same. The flow rates indicated by the
turbine and magnetic flowmeters in the system are
plotted against pump speed in Fig. 1.3.17. It may
be noted that the linearity of both the turbine
flowmeters is quite good when compared with
that of the magnetic flowmeter, which is known to
be linear. It may also be noted that there is
considerable error in the flow rates indicated by
the turbine flowmeters in comparison with the rates
given by the magnetic flowmeter. This was
expected, as explained above, since the units
were not recalibrated and are still operating near
the lower limit of the design operating range.

UNCLASSIFIED
ORNL—LR—DWG 27029

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3600 ‘ ' ‘ ‘
3200 e MAGNETIC FLOWMETER
4 TURBINE FLOWMETER NO. 4-REV {
o TURBINE FLOWMETER NO.5-REV 1
2800
E oA ®
Q
& 2400 ,///./
w s
g /’/./
$ 2000 o =/ 7~
5 L~
o. //c
1600 M,/ e
;/./ -
/ v rt
1200 // /=
o A/.‘
800
1 2 3 a 5 6 7 8

INDICATED FLOW RATE (gpm)

Fig. 1.3.17. Comparison of Flow Rates Indicated by
Two Turbine Flowmeters and a Magnetic Flowmeter in a

Test Loop Circulating NaK at 1200°F.

Since the units were reinstalled in the test loop
they have operated in excess of 350 hr at temper-
atures ranging from 1200 to approximately 1600°F.
It has been demonstrated during these tests that
the units can be stopped at high temperatures and
then restarted. It thus appears that self-welding
of the bearings is not a problem. A detailed
report on turbine flowmeter development and
testing is to be prepared.
PRESSURE TRANSMITTERS
Co Mo Burfon

Additional tests were made of strain-gage
pressure transmitters designed for use at high
temperatures. In these pressure transmitters the
process pressure acts on a diaphragm, which
transmits the pressure to a NaK-filled capillary
tube, which, in turn, transmits the pressure to a
Bourdon spring. The Bourdon spring is linked
to an unbonded strain gage, which is connected
electrically in the form of a Wheatstone bridge.
Similar strain gages have been used extensively in
aircraft instrumentation work.

Primary pressure sensors of the same type but
with pneumatic means for signal transmission were
tested previously and are being used to obtain
important pressure measurements in experimental
test facilities. The possibility of leakage of
radioactive material through pneumatic tubing in
the event of an explosion, however, has directed
attention toward instrumentation with electrical
transmission.  The condition of the electrical
readout equipment can be periodically checked by
switching to a calibrated resistor which should
have a predetermined ‘‘synthetic’’ reading.

The strain-gage bridge is energized with a 5-v d-c
supply, which can be unregulated if the voltage-
ratio type of measuring equipment is used to
determine the millivolt signal. This type of
readout equipment is commercially available as a
slightly modified form of the familiar electrical
self-balancing potentiometer,

Under test conditions, six strain-gage pressure
transmitters of O- to 100-psig range indicated an
average output-input deviation of 0.4% of full
scale over their operating pressure range from
ambient temperature to 1400°F. This accuracy
compares favorably with the accuracies of other
pressure transmitters. Two of these pressure
transmitters are now being tested in a NaK loop
under more severe conditions.

ART THERMOCOUPLES
J. T. DelLorenzo
Sheathed Thermocouples

Inconel-sheathed  thermocouples (0.250-in.-OD
sheath, MgO insulation) with hot-junction closure
welds made by the Heliarc welding process are

PERIOD ENDING DECEMBER 31, 1957

still being tested, with approximately 2000 hr of
additional exposure time having been accumulated
since the previous report. Eighteen Chromel-
Alumel thermocouples, with closure welds that
passed dye-penetrant and x-ray inspection, have
now completed over 8000 hr of operation at 1500°F
in sodium. The operating temperature has been
reduced approximately once a week to 1300 and
to 1100°F for readings at these temperatures.
Since the last report,’ no major shifts in per-
formances have been observed. Three of the seven
assemblies that earlier exhibited a slight downward
shift have started slight upward drifts. Four of
the seven assemblies with slow upward drifts have
finally drifted approximately 2°F out of tolerance
at the 1100°F test temperature. (An accuracy of
iaé of 1% over the range of 530 to 2300°F is
claimed by the vendor of the material being tested.)
One assembly that showed a sharp downward drift
at about 3000 hr has continued to drift downward
and is well out of tolerance. All assemblies
other than the five mentioned above are presently
operating within the vendor's tolerance. The data
for all 18 assemblies are summarized in Table 1.3.2.

Thirteen similar specimens and three platinum,
platinum-=10% rhodium units have now completed
over 6000 hr of operation at 1500°F in NoF-ZrF4-
UF4 (50-46-4 mole %, fuel 30). Readings at
approximately one-week intervals were also taken
at 1300 and 1100°F. No major shifts in per-
formance have occurred over the past 2000 hr of
operation. All thermocouples except one are now
showing upward drift trends but are still within the
vendor’s tolerance. The data for these 16 thermo-
couples are summarized in Table 1.3.3.

The effect of nuclear radiation on the resistance
of the magnesium oxide insulation was checked
by inserting three samples of the sheathed ma-
terial in hole HB-3 of the LITR. Two samples
were maintained ot a constant temperature of
1200°F and the third sample was not heated (gamma
heating raised its temperature to 450°F), The
samples were irradiated for three weeks, and a
preliminary examination has revealed that no
significant immediate effect was produced by
radiation — a 10 to 30% decrease in resistance
was observed in going from zero power to full

 

%5 T. Del.orenzo, ANP Quar. Prog. Rep. Sept. 30,
1957, ORNL.-2387, p 61.

45
ANP PROJECT PROGRESS REPORT

reactor power. Further, no significant drift in
resistance was noted during the three-week interval
of the test. An upward drift that was noted during
the first 48 hr of the test was believed to be a
result of additional drying of the magnesium oxide.

Toble 1.3.2, Results of Drift Tests of Inconel-
Sheathed Chromel-Alumel Thermocouples in Sodium

 

Thermocouple Reading (°F)¢

 

 

Test Period
(hr) Deviation? Spread®
Test Temperature: 1500°F
0 542 6s6 to 3.0
1000 5.1 66 to =~0.2
2000 5.4 7.4 to 0.8
3000 4.3 7.0 to —645
4000 4.4 7.8 to —11.5
5000 2.4 8.5 to —10.0
6000 1.9 11.5 to —17.8
7000 0.8 8.0 to —20.0
8000 0.7 9.9 to —27.5
Test Temperature: 1300°F
0 73 8.8 to 6.0
1000 8.1 10.0 te 5.0
2000 5.7 8.0 to 2.0
3000 4.7 85 to —3.8
4000 1.5 8.0 to —8.5
5000 3.8 9.7 to —6.0
6000 2.7 9.0 to —10.5
7000 3.5 10,1 to —14.4
8000 0.9 9.2 to —23.0
Test Temperature: 1100°F
0 6.8 8.1 to 5.5
1000 6.9 8.2 to 4.8
2000 6.2 8.0 to 3.4
3000 6.9 8.5 to 1.5
4000 4.2 9.5 to ~14.0
5000 4.9 10.0 to —4.0
6000 4.6 9.5 to —5.0
7000 4.2 106 to = 13.0
8000 4.2 105 to ~3.1

 

a. .

Nineteen assemblies were tested (one was removed
after 5000 hr)}; each assembly contained two thermocouples
and was aged 24 hr at 1350°F inheliumprior to testing.

Deviation of the average of the thermocouple readings
from the test temperature.

“Maximum spread of the thirty-eight readings obtained.

46

Tests have revealed the extreme susceptibility
of magnesium oxide to water absorption. Con-
siderable baking was necessary to insure dryness
of the material prior to the tests,

Well Thermocouples

Drift tests on well thermocouples designed for
temperature measurement of high-velocity, high-
temperature (1400 gpm, 1100 to 1500°F) liquid
metals in Inconel piping were continued; over
2000 hr of testing has been completed. The ten
thermocouples (two per well) employed in these
tests were fabricated from 20 AWG, solid, Chromel-
Alumel wire (accuracy, +% of 1% over range of
530 to 2300°F) and were welded, with Heliarc
welding techniques, to the bottom of the well.
The test environment is static liquid NaK at
1500°F. The temperature of the system is reduced
to 1300 and 1100°F weekly for readings at these
temperatures.

No major shifts in drift have occurred since the
start of the tests. All the thermocouples appear
to be operating within the tolerance given above.
The very close agreement between the two thermo-
couples in any given well is surprising, and the
maximum deviation for all five units is 1°F. The
variation in readings between wells can possibly
be explained by the high temperature differences
that exist in the region of the wells. The data
taken to date are summarized in Table 1.3.4.

All the work done on the design and testing of
ART thermocouples will be summarized in two
ORNL reports to be issued soon. One report will
cover the fabrication techniques used and the
results of aging tests on Chromel P—Alumel and
platinum, platinum=10% rhodium systems in fuel,
sodium, and NaK, in addition to the results of
rapid heating tests. The second report will
describe all aging tests in air on Chromel P—Alumel
and platinum, platinum—=10% rhodium systems; this
report will also include some recent aging infor-
mation on thermocouple systems made of iron-
constantan, British-made Chromel P-Alumel,
Kanthal alloys, Driver-Harris 242-33 alloys, and
Chromel P-Alumel systems with a titanium oxygen
getter.
PERIOD ENDING DECEMBER 31, 1957

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

 

Chromel-Alumel

Chromel-Alumel

Pt, Pt—-10% Rh

 

 

 

Test Poriod Assemblies? Aged 24 hr Assemblies® Aged 200 hr Assemblies®
(hr) Deviation? Spread® Deviation? Spread® Deviation? Spread®
(°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 y 6e5 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 Bs5 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 18 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 Temperature: 1300°F
0 7.0 8.0 to 5.0 649 7.0 to 6.9 ~0.1 1.8 to —3.8
1000 4.3 7.6 to ~0.5 649 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 12 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 36 8.4 to —B8.5 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 3.5 604 6:8 to 6.0 —0.3 0:4 to —0.9
2000 2.8 8:2 to —5.0 7.7 82 to 7.2 —0.5 12 to —1.5
3000 y 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 86 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 6el 10:2 to —4.5 10.6 11.4 to 9.7 3.1 5.4 to 3.3

 

aEighf assemblies were tested; each assembly contained two thermocouples and wos aged 24 hr at 1350°F in

helium prior to testing.

bTwo assemblies were tested; each assembly contained two thermocouples and was aged 200 hr at 1350°F in

helium prior to testing.

“Three assemblies were tested; each assembly contained two thermocouples.

 

dDeviafion 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

 

 

 

Test Well No. 1 Well No. 2 Well No. 3 Well No. 4 Well No. 5§
Period  peyiation®  Spread?  Deviation®  Spread?  Deviation®  Spread®?  Deviation®  Spread?  Deviation®  Spread®
(he) (°F) (°F) (°F) (°F) (°F) (°F) (°F) (°F) (°F) (°F)
Test Temperature: 1500°F

0 2‘2 008 _212 0-2 ] 00 0.7 —4-3 ] -4 _0n4 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 ]-0 --].5 0-2 3- 7 0.5 -003 005 3-4 000

Test Temperature: 1300°F

0 3-] 0-8 0-0 0-] _305 0-] "'5-6 0-6 -405 I'O
1000 2.7 . 0.8 ~1.5 0.1 2.9 0.0 ~2.3 0.1 2.4 0.0
2m0 4-0 0-7 _0-6 0-2 306 0-6 "206 009 219 0-]

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

 

9Deviation of the average of the two thermocouple readings from the test temperature.
Spread between the two thermocouple readings.

48
1.4. ENGINEERING DESIGN STUDIES

A. P. Fraas

APPLIED MECHANICS AND STRESS ANALYSIS
R. V. Meghreblian

Summaries and final reports on the applied
mechanics and stress analysis problems of the
ART have been prepared. The reports that have
been prepared and issued include an analysis of the
north-head pressure-stress distribution,! a des-
cription of a general relaxation method for deter-
mining two-dimensional temperature distributions on
an IBM-650,2 an analysis of a thermal cycling
experiment performed on beryllium by using
volume heat sources,?® the results of a study of
the temperature distribution at the junction of
shell V (lower) and shell VI,4 the results of
the stress analysis of controlled structural tests

 

's. E. Moore, ART North Head Analysis, ORNL CF-57-
10-114 (in press).

2p, M. Miller, Relaxation Calculations by Machine for
Two-Dimensional Heat Conduction Problems, ORNL

CF-57-11-4 (Nov. 1, 1957).

3s. E. Moore, Canned Cylindrical Beryllium Specimen
t{/qitb)lntemal Heating, ORNL CF-57-10-109 (Oct. 30,
57).

4D. M. Miller, Temperature Profile of Shell V Lower—
Shell VI Connection, ORNL CF-57-11-9 (Nov. 1, 1957).

 

PERIOD ENDING DECEMBER 31, 1957
|
|

performed on the IHE-8B heat exchangers, and

a summary description of the core shell-liner
experimenf.6

Reports were also completed on the temperature
distribution in the ART reflector-moderator as-
sembly,” on the stress analysis of the radiators,®
and on the thermal stress analysis of the ART
heat exchanger channels and header pipes.’
Two memoranda were also prepared on general
methods for analyzing plates and shells,10:!]

 

3p. L. Platus, Summary of IHE-8B Heat Exchanger
Stress Analysis, ORNL CF 57-12-7 (in press).

p. H. Platus, Summary of Core-Shell-Liner Experi-
ment, ORNL CF-57-1-134, Addendum (Dec. 11, 1957).

7D, L. Platus, D. M. Miller, and R. V. Meghreblian,
ART Re{lector Temperature Distribution, ORNL-2425
(Nov. 2, 1957).

8p. H. Platus, ORNL Tube-Fin Radiator Stress Analy-
sis, ORNL-2441 (in press).

°p. L. Platus, Thermal Stress Analysis of the ART
Heat Exchanger Channels and Header Pipes, ORNL-2442

(in press).

106, L. Greenstreet and D. M. Miller, An Extension of
the Esslinger Shell Relationships to Include Effects of
Axial Load and Temperature, ORNL CF-57-11-65 (Nov.
19, 1957).

e, 4, Stanek, Stress Analysis of Flat Circular Plates,
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
the design of the ART were the same as those that
arise in the design of any reactor, but there were
major shifts in relative importance. Typical
problems that were studied included critical core
size, or critical mass of fuel; reactivity effects of
temperature, poisons, control rods, and so forth;
radiation heating of various parts of the reactor;
activation or neutron-induced changes in the compo-
sition of various materials; and shielding.

For a fused-salt reactor, the problem of critical
mass may be viewed in a somewhat different
perspective than for a fixed-fuel reactor. Since
the concentration of fissionable material can be
readily odjusted to achieve criticality or the
desired excess reactivity, the urgent problem is
not so much to know in advance exactly how much
fuel the reactor will require, as to be assured that
the melting point of the fuel, which depends upon
the uranium concentration, will remain within
acceptable limits for any foreseeable additions of
UF, to the fuel.

The criticality problem for the ART was initially
one of finding a suitably compact core geometry,
while at the same time keeping the UF, concen-
tration within acceptable limits. The work of the
Laboratory on this problem'=3 was augmented by
a contract with the Curtiss-Wright Corporation for
a parametric survey? of fuel concentration and core
power distribution, upon which the physical dimen-
sions of the ART core and reflector were finalily
based. The composition and thickness of the core
shell were shown to have a decisive influence on
critical fuel concentration;4® this was confirmed

 

A, P. Fraas and C. B. Mills, A Reflector-Moderated
Circulating Fuel Reactor for an Aircraft Power Plant,
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).

3M. E. LaVerne and C. S. Burtnette, ANP Quar. Prog.
Rep. June 10, 1954, ORNL-1729, p 32.

4y, Reese, Jr., S. Strauch, and J. T. Mihalczo, Geom-
etry Study for an ANP Circulating Fuel Reactor, WAD-
1901 (Sept. 1, 1954); C. B. Mills and H. Reese, Jr.,
Design Study og an ANP Circulating Fuel Reactor,
WAD-1930 (Nov. 30, 1954).

Sw. K. Ergen, ANP Quar. Prog. Rep. June 10, 1955,
ORNL-1896, p 20.

50

by the room-temperature critical experiments.®
While the analysis indicated a negative temperature
coefficient of reactivity, the marked dependence
of critical concentration on the core shell thick-
ness caused some concern that the reactor might
in fact have a positive moderator temperature
coefficient, and led to the performance of a high-
temperature critical experimenf7 which proved the
moderator coefficient to be negative. The experi-
ment was based on an exact mockup of the design
core. Nevertheless, there were a number of dif-
ferences, such as the absence of reflector cooling
and deviations in the outer boundary of the reflector,
whose cumulative effect on critical concentration
proved to be quite significant.® The required
concentration of UF, in the ART fuel, including
allowance for reactivity changes during the core
life, was expected to be about 5 to 5.5 mole %, in
contrast to the critical concentration in the high-
temperature critical experiment of 2.84 mole %.

A major problem in the circulating-fuel reactor
is activation of the secondary coolant. While the
thermal leaving the reflector can be
effectively absorbed in a boron-containing layer
between the reflector and the heat exchangers,
some activation remains because of higher energy
neutrons absorbed in the sodium resonance at
2.9 kev, or moderated in the heat exchanger, and
because of delayed neutrons emitted in the heat
exchanger. The problems of determining the level
of sodium activity and of finding a satisfactory
means of keeping the activity as low as possible
have been examined extensively, both analytically?®
and experimentally.1®

The total Na?? activity expected in the ART
secondary coolant loop was 1000 curies. It was
evident, however, that the level of activity could

neutrons

 

SA. D. Callihan et al., ANP Quar. Prog. Rep. March
10, 1955, ORNL.-1864, p 43.

7. D. Callihan et al., ANP Quar. Prog. Rep. Sept.
10, 1955, ORNL-1947, p 58.

BA. M. Perry, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL.-1947, p 33.

H. W. Bertini, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 28.

19, B. Dee et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, pp 204, 209.

 

 
1
be very substantially reduced — perhaps by a factor
of 3 or more ~ in an aircraft reactor.

Closely related to the NaK activity is the activity
of material activated in the heat exchangers and
carried to the radiators by mass transfer in the
NaK stream. In the case of the ART, the activity
in the radiators after shutdown and removal of the
NaK was expected to be about 1 curie, and the
dose rate 1 ft from the radiator surface was expected
to be about 1 r/hr, due mostly to Co%0 (ref 11).

A major area of investigation in reactors with
high power density is the heating of structural
components by gamma radiation and neutrons.
Calculations showed that, in the ART, the heating
of the reflector, the core shells, the north head
structures, and many other components would be
substantial, especially from gamma radiation, and
special provisions were provided for cooling them.
A complete summary of the radiation heating calcu-
lations may be found in two topical reports;'2/13
several aspects of the problem have been discussed
in progress reports, that is, heating in the vicinity
of the pressure vessel,'4 the north hec:d,]5 the
off-gas system,'® and in the beryllium and core

shells. 1718

 

”A. M. Perry, ANP Quar. Prog. Rep. March 10, 1956,
ORNL.-2061, p 30.

124 W. Bertini et al., Basic Gamma-Ray Data for
ART)Heat Deposition Calculations, ORNL 2113 (Oct. 3,
1956).

13H. W, Bertini et al., Radiation Heating in the ART,
ORNL-2429 (to be published).

4y w. Bertini, ANP Quar. Prog. Rep. June 10, 1956,
ORNL -2106, p 28.

15H. W. Bertini and D. L. Platus, ANP Quar. Prog.
Rep. June 10, 1956, ORNL-2106, p 34.

 

PERIOD ENDING DECEMBER 31, 1957

The control rod in the ART illustrates the usual
considerations of depletion of the active material
and of internal heating to a marked degree.!? |t
was found?%2! that a 13/8-in.-0D annular rod,
l/4 in. thick, would be adequate if composed of
rare-earth oxide at l/3 to ]/2 the normal density,
with the voids occupied by metal or sodium to
achieve high thermal conductivity. The reactivity
value of the control rod was determined entirely
by critical experiments.’

Most of the considerations affecting the shielding
requirements for the ART are described in the
Report of the 1953 Summer Shielding Session,?2
Additional studies relating particularly to the
shielding of various components included the NaK
pipes, 2 fuel drain line,?® fuel overflow line,?3
vapor traps,?3 and the fill-and-drain tank.24

 

16p. B. Stevenson, ANP Quar. Prog. Rep. June 10,
1956, ORNL-2106, p 35.

7r. B. Stevenson, ANP Quar. Prog. Rep. Sept. 10,
1956, ORNL.2157, p 31.

laC. M. Copenhaver, ANP %uar. Prog. Rep. Dec, 31,
1956, ORNL-2221, p 15; R, B, Stevenson, ANP Quar.
Prog. Rep. Dec. 31, 1956, ORNL-2221, p 16.

wW. K. Ergen and H. W, Bertini, ANP Quar. Prog.
Rep. March 10, 1955, ORNL-1864, p 26.

20W. Fader, ANP Quar. Prog. Rep. June 10, 1955,
ORNL -1896, p 22.

2IW. Fader and A. M. Perry, ANP Quar. Prog. Rep.
Sept. 10, 1955, ORNL-1947, p 33.

22E. P. Blizard and H. Goldstein (eds.), Report of
%ge ]1995543) Summer Shielding Session, ORNL-1575 (July

23, W. Bertini, C. M. Copenhaver, and A. M. Perry,
ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 71.

24y, W. Bertini, ANP Quar. Prog. Rep. Sept. 30,
1957, ORNL-2387, p 80.

51
ANP PROJECT PROGRESS REPORT

1.6. MATERIALS AND COMPONENTS INSPECTION
A. Taboada

MATERIALS INSPECTION

Approximately 4000 requests were received for
the inspection of materials for the ART-ETU
program. The total quantities inspected and the
rejection rates are given in Table 1.6.1. At the
time of suspension of the program 61,243 ft of
CX-900 Inconel tubing was being inspected or

WELD INSPECTION

inspected were made almost
inert-gas shielded-arc method
and were predominantly of Inconel material.
Visual, dye penetrant, and radiographic inspection
techniques were used for the inspection of all
thin-walled tubing and other critical weldments

The weldments
entirely by the

 

 

 

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 f%) 15.5
Regular inconel 54,000 ft 54,000 #{?) 100.0
CX-900 Inconel 73,000 f+ 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 #'d) 59.0
small pipe)
INOR-8 128 ft 5 f(d) 4.0
Pipe Regular Inconel 12,300 1,329 (@) 11.0
CX-900 Inconel 1,470 f+ 70 ft 5.0
2 oe)
Plate and sheet CX-900 Inconel 1,574 1+ 198 ft 12.5
Other 7,350 ft2 257 f12 3.5
Red and bar CX-900 Inconel 50 # 0 ft 0.0
Other 2,381 ft 139 fr 6.0
Fittings Machined adapters, reducers, 600 ft 43 #t 7.0
ells, pipe caps, etc.
(a)Used for test loop applications; rejected for defects greater than 5% of the wall thickness or cracks.
b Rejected on the basis of a random 10% sampling inspection.
C)Recctor 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
welds are designated C in the summary presented
in Table 1.6.2. Weldments which did not require
radiography, either because of inaccessibility or
because the weldment did not require complete re-
liability, are designated CN in Table 1.6.2.
Heavy section weldments on test loop pipes were
also designated CN.

The ANP welding performance record from

PERIOD ENDING DECEMBER 31, 1957

December 1954 to September 1957 is summarized
in Table 1.6.2. Prior to December 1954, non-
qualified welders had been used, and rejection
rates approached 40%. The culmination of this
work was the ETU assembly welding. Two
hundred and fifty-six weldments were made, of
which 40, or 15.5%, were rejected. The ETU
weld rejection rate at termination was approxi-
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 Quontity Quantity Rejection Quantity Quantity Rejection
Inspected Rejected Rate (%) Inspected Rejected Rate (%)
1954 to Oct. 1955 1,358 275 2 276 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 25

 

53
ANP PROJECT PROGRESS REPORT

1.7. HEAT TRANSFER STUDIES
H. W. Hoffman

THERMAL-CYCLING RESEARCH
Pulse-Pump System

J. J. Keyes A. l. Krakoviak
J. E. Mott!

Several additional tests were completed in
the high-frequency thermal-cycling loop? with the
fuel mixture NaF-ZrF ,-UF, (56-39-5 mole %) as
the circulating fluid and helium as the pulsing
and cover gas. A Milton Roy pump with a ll/z-in.-
dia piston and a 4-in. stroke was used to generate
approximately sinusoidal pressure oscillations at
frequencies of 0.4 and 1.0 cps. These pulsations
were transmitted to the fuel in the pulse chamber
so as to force alternate hot and cold slugs of
liquid through the test section. Outside wall
temperatures and, in one case, fluid stream tem-
peratures were measured instantaneously; the in-
side wall temperature amplitude was estimated by
using data from previous water tests. The test
sections were constructed of Inconel tubing and
pipe that was approximately 0.5-in. in inside
diameter and varied in wall thickness.

The test conditions and results for runs 3
through 5 are given in Tables 1.7.1 and 1.7.2,

 

]On assignment from Pratt & Whitney Aircraft.

2A complete description of this experimental system
may be found in the following ANP Quarterly Progress
Reports: ORNL.2157, p 56, ORNL-2221, p 54, and
ORNL-2274, p 53.

and the geometries of the test pieces, as well
as the direction of flow and the location of the
samples taken for metallographic study, are
shown in Fig. 1.7.1. The photographs of Fig.
1.7.2 show the results of examination of the
test piece from run 5. In Fig. 1.7.22 some
surface cracks are visible in the transition
section near the entrance of the heavy-walled
test piece (sample e). An approximately 0.070-in.
crack in the wall of the same test piece at a
position 3 in. downstream of the entrance
(sample b) is shown in Fig. 1.7.2b. The total
failure of the Inconel pipe in the vicinity of a
weldment (sample c) may be seen in Fig. 1.7.2c.

A cursory examination of the data indicates
that the severity of cracking as a result of thermal
cycling increases with increasing pipe wall thick-
ness and with proximity to a weld. This study
is being continued on a limited basis.

The basic investigation of thermal transients
at a solid-fluid interface was continued, with
an experimental study in which the conditions
of an adiabatic infinite wall were approximated.
The results, obtained for three different film
heat-transfer coefficients ranging between 2500
and 5000 Btu/hr-f12.°F, have validated the theo-
retical analysis. A memorandum giving a detailed
description of the experimental system and a
discussion of the results has been issued.?

 

3J. E. Mott, An Investigation of Thermal Transients
c]zéa Solid-Fluid Interface, ORNL CF-57-10-122 (Oct. 25,
57).

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

 

Maximum Temperature
Average Fluid

Frequency

Total Total Fluid

 

Test Difference Between of Temperature . Total
No.* Tempoerafure Hot and Cold Streams Oscillations T':e Cycles Flow Rate
(°F) (OF) (cps) (hr) (gpm)
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.

***E ctimated time at failure.

54

 
oS

 

 

1 a - s * . ‘ . . . .
Table 1.7.2. Results of High-Frequency Thermal-Cycling Tests of Inconel Pipes
Wall Meosured Temperature Estimated Temperature ) o
Test Sample Thickness Amplitude on Outside Amplitude on Inside Results of Metallographical Examination
Ne. Designation (in.) Pipe Wall (°F) Pipe Wall (°F)* of Test Piece and/or Adjacent Piping

3 a 0.091 164 168, 195 Moderate to heavy void formation to a depth of 2 mils;
some intergranular cracks or attack

3 b 0.029 164 168, 195 Very light and scattered void formation to @ maximum
depth of 2 mils

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

4 b 0.091 Not measured 1125, £150 Heavy intergranular cracks to a depth of 20 to 80 mils
found immediately above and below welds

4 c 0.065 Not measured 1125, £150 Moderate intergranular cracks which appeared to pro-
duce spalling

4 d 0.029 185 193, +125 VYery few general and intergranular voids to a depth
of 3 mils

5 a 0.133 120 +150, +200 Moderate intergranuiar cracks, especially near weld

5 b 0.250 17 150, +200 Heavy intergranular cracks to a depth of 72 mils

5 c 0.091 168 1100, 175 Moderate intergranular cracks with heavy cracks in
heat-affected zone of weld where failure occurred

5 d** 0.133 15 1100, £175 Light intergranular cracks to a depth of 2 to 3 mils

 

*First value calculated from outside surface temperature; second value calculated from fluid temperature amplitude by using the 7 ratio,

**Subjected to conditions of tests 3 through 5.

L5601 ‘l€ ¥3GW3I3Aa IONIAN3 @0i¥3d
ANP PROJECT PROGRESS REPORT

     

SAMPLE g

SAMPLE b

TEST PIECE 3

 
   
 
 
   

SAMPLE a

SAMPLE b

-~ SAMPLE c¢

- SAMPLE 4d

TEST PIECE 4

UNCLASSIFIED
ORNL-LR-DWG 27030

FLOW

    
   
  

] }SAMF’LE a

 

’ }SAMPLE e

SAMPLE b

SHEATH-TYPE STREAM

WELD — THERMOCOUPLE

WELDS
SAMPLE ¢ { et

 

SAMPLE d
TEST PIECE 5

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

Pressurized System

H. W. Hoffman  D. P. Gregory!

The experimental investigation of the effect
of thermal-stress cycling on Inconel tubes filled
with flowing NaF-ZrF ,-UF, (50-46-4 mole %) with
the use of a pressurized flow system was dis-
continued. A report summarizing the results of
this experimental study is being prepared.?

ART HYDRODYNAMICS

G. L. Muller’  F.E. Lynch
W. J. Stelzman

The investigations of the mechanics of the ART
core fluid flow for which the full-scale plastic
core model was being used have been terminated.
The tests of the core that was divided into three
flow regions by installing two annular shell
liners®> were not completed. A discussion of
the design has been prepared by Platus.® The

mass of data obtained in the course of these

 

4H. W. Hoffman and D. P. Gregory, An Experimental
Study into the Effects of Surface Thermal Cycling on
Inconel in a Fluoride Fuel Environment, ORNL-2435
(to be published).

5W. J. Stelzman, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 96.

56

studies is being assembled and annotated for
future publication. The data consist primarily
of Pitot traverse measurements and dye-injection
movies.

No further work has been done on the experi-
mental studies of flow through quarter-scale
models of the ART core. A complete report on
this work has been prepared.” Work done on the
screen-packed core and header combination while
simulating single and dual pump operation will
be reported in a later memorandum.

A report describing the phosphorescent-particle
technique for velocity-profile visualization is also
being prepared.®

FUSED SALT HEAT TRANSFER
S. |. Cohen H. W. Hoffman

Forced-convection heat transfer studies with
KCI-LiCl (41.2-58-8 mole %) flowing in a heated

tube were continued. The results obtained are

 

6D. H. Platus, Summary of Core-Shell-Liner Exgeri-
ment, ORNL CF-57-1-134, Addendum (Dec. 11, 1957).

76 L. Muiler, J. O. Bradfute, and F. E. Lynch, Investi-
gation of Fluid Flow in the ART and Other Reflector-
Moderated Reactor Cores, ORNL-2199 (in press).

8F. E. Lynch, L. D. Palmer, and G. M. Winn, Instan-
taneous Velocity Profile Measurement by Photography,
ORNL-2257 (to be published).

 

 
UNCLASSIFIED UNCL ASSIFIED
T-14311 T-14312

UNCLASSIFIED
T-14310

 

 

   

 

{
=3

 
 

&

1
o Ut Ay
-.n'.,r_,_r

INCHES

e
!
¢
-
g F x
e
=
T |
;
i

 

3.
¥
e
=
J-
2
¥

 

 

\...'l -~

: Y T 0.04

- X £ e 01—— ,

. ) AL il

;{ AT e SR E‘Tl
C A T TS N :‘e -"_-/:-. 0.08

M Ly "'y" PJ'S s EI-I
S e IR e 1 l0.08

. ¥, y o
Y
4= ’
=

 

 

 

 

/
’.
3
22X
—
<3
':rf\.l'n\l
£ 0N
h 4 V-\ .. I \
.-'-fif;_‘
g_‘ T > *y
-_"l.
.
\
""--_,__1

W “'\"’ e.07

 

Fig. 1.7.2. Effect of High-Frequency Thermal Cycling on Inconel in an NaF-ZrF ,-UF (56-39-5 Mole %) Environment — Test Run 5. (a) Inlet transi-
tion section; sample location e, 4X. (b) Section from sample location b; crack extends about 0.070 in. into 0.250-in. wall. 50X. (c) Exit region; sample

location c; %fi-in. sched-40 pipe. 4X, (Secret with caption)

LS

LS61l ‘1€ ¥39W3D30 ONIAN3I dOI¥3d
ANP PROJECT PROGRESS REPORT

presented graphically in Fig. 1.7.3 in terms of
the heat transfer parameter NNu/Ng.l'4 The data,
which cover the Reynolds modulus range between

6,000 and 18,000, were obtained for both type

347 stainless steel and Inconel test sections.

UNCLASSIFIED

CRNL~-LR-DWG 270314
200 T T T 1T

ASERIES GA, TYPE 347 STAINLESS STEEL
® SERIES GB, TYPE 347 STAINLESS STEEL
0 SERIES GC, 'NCONEL

 

o
o

 

 

 

 

 

 

 

 

 

 

[ M o -
vo =0.023 Mg, «\7/
v
/ . ..o OQJ
20 7 ®0
/ ¢ 0

HEAT TRANSFER PARAMETER, ¥y, /V5:*
3

 

 

| 0
) o Y

:p1]
10 2

2X10° 5 108 o
REYNOLDS MODULUS Mg,

 

 

 

 

 

 

 

 

 

 

 

4x104

Fig. 1.7.3. Heat Transfer Characteristics of KCl-LiCl
(41.2-58.8 Mole %).

The data of the first experiment (series GA)
for a type 347 stainless steel tube fell approxi-
mately 55% below the general heat transfer
correlations for standard fluids. These results
were corroborated in succeeding tests with a
type 347 stainless steel test section
(series GB) and with an Inconel test section
(series GC). An analysis of the system indicated
that the observed deviations might result from
the use of incorrect thermal conductivity values
in evaluating the data and/or the existence of
high thermal resistance films at the salt-metal
interface.

new

The thermal conductivity value used for the
salt in the analyses of the data was obtained by
averaging the preliminary value of 0.8 Btu/hr.ft:°F
experimentally determined by Powers® and the
value of 1.2 Btu/hr-ft-°F predicted on the basis
of correlations by Gambill.'® No attempt was

 

w. D. Powers, unpublished data.

lOW’. R. Gambill, Prediction of the Thermal Conduc-
tivity of Fused Salts, ORNL CF-56-8-61 (Aug. 10, 1956).

58

made to weight the data from either of these
sources. o completely explain the observed
discrepancy in the heat transfer parameter, a
decrease in the thermal conductivity of the salt
to a value of 0.3 Btu/hrft:°F would be needed.
On the basis of previous experimental measure-
ments of the thermal conductivity of fused salts,
a change of this magnitude (0.8 to 0.3 Btu/hr-ft.°F)
between preliminary and final values is not
anticipated. Thus, incorrect thermal conductivity
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 systems'! that
a reaction occurred between the salt and the
metal to form a film of high thermal resistance
at the salt-metal interface. Visual observation
of the inside surface of the test section from
series GA did disclose a very thin reddish-brown
surface deposit. Spectroscopic and x-ray analyses
of this film material have been inconclusive.
Thus, the question as to whether this film is
due to a surface reaction or to deposition from
an impure salt mixture remains unresolved. Work
on this program has been discontinued, and the
results are summarized in a series of reports. ! ' =14

MOLTEN LITHIUM HEAT TRANSFER
N. D. Greene

Preliminary designs are being prepared of an
experimental system for the study of forced-
convection heat transfer with molten lithium
flowing in metal tubes. The loop proposed will
consist of o gas-sealed pump, a magnetic flow-

meter, a test section, and an gir-cooled heat
sink. In order to mitigate the difficulties and
hazards encountered in earlier studies with

lithium, an effort will be made to keep the volume

 

My, W. Hoffman and J. Lones, Fused Salt Heat Trans-
fer. Part 1l: Forced Convection Heat Transfer in
Circular  Tubes Containing NaF-KF-LiF Eutectic,
ORNL-1777 (Feb. 1, 1955).

12H. W. Hoffman, Turbulent Forced Convection Heat
Transfer in Circular Tubes Containing Molten Sodium

Hydroxide, ORNL-1370 (Oct. 3, 1952).

l3}"|. W. Hoffman and S. |. Cohen, Fused Salt Heat
Trans fer, Part lll: Forced-Convection Heat Trans fer in
Circular Tubes Containing the Salt Mixture NaNOQ,-
NaNOs-KNO3, ORNL-2433 (in preparation).

T4y, w, Hoffman, S. |. Cohen, and D. P. Gregory, Ex-
perimental Determination of Forced-Convection Heat
Trans fer with Several Fuel and Coolant Mixtures, ORNL
CF-58-1-1 (in preparation).

 

 
of liquid metal to a minimum. The lithium will
be maintained at approximately 700°F, and the
Reynolds modulus in the test section will be

at least 50,000.

The test section will be constructed of type
347 stainless steel tubing having an inside
diameter of approximately ?'/] in. Heat will be
generated electrically in the tube wall by the
passage of a high amperage current. The heat
generation in the lithium can be reduced to
approximately 7.6% of the total for a tube wall
thickness of ”/32 in. Test section wall tem-
peratures, fluid flow rates, and mixed-mean
temperatures will be measured.

VOLUME-HEATED SYSTEMS
ART-Type Core With Screens
N. D. Greene

The study of the effect of screens in a diverging
annular channel on the thermal structure of «
fluid under volume-heat-source conditions is
being continued on a limited scale. Four screens
of varying radial solidity have been fabricated
and are currently being installed in the test
model. These screens were designed to increase
the resistance in the high-velocity regions of
the flow so as to force the fluid into the lower
velocity region near the outer wall.

Liquid-Metal Experiment
G. L. Muller

Construction of the system described pre-
viously'® for the experimental investigation of
heat transfer in a liquid metal (mercury) with
internal  heat generation has been essentially
completed. The test-section electrical-power leads
and instrumentation - are now being installed.
The three flow-measuring orifices have been
calibrated with mercury by using a weigh tank.
Differential pressure cells and a recorder were
used to measure the orifice pressure drop. A
few minor leaks around valve stems and pump
shaft seals have been corrected.

An attempt has been made to obtain a universal
relation for forced-convection volume-heat-source

 

156, L. Muller, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 106,

PERIOD ENDING DECEMBER 31, 1957

heat transfer with liquid metals. The empirical

equation
k(tw - tm) ]
(1) - = T o e ——,
0.92
wr, 79 + 0.226 Np*.
where
t, = wall temperature,
t, = mean fluid temperature,
k = thermal conductivity,
W = power density,
r, = tube radius,
Np, = Peclet modulus (the Reynolds-Prandtl

product),

was found to describe the analytical results pre-
viously reported'® to within +15% over the limits
0 < Np, 0.1 and Npo 2 10,000. The entire
turbulent range (N, > 5000) can be represented
to a lesser degree of accuracy (approximately
+30% at low N, _) by the equation

- t) 1
2 0.915
Wrw 65 + 0.24 NPo

The results obtained previously are compared in
Fig. 1.7.4 with results of analyses made by
Poppendiek and Palmer.'” It may be seen that
for values of the Peclet modulus above 1000
the agreement is excellent.

HEAT TRANSFER WITH YORTEX FLOW
W. R. Gambill N. D. Greene

Previous experiences with source-vortex flow of
fluids in forced-convection heating (no phase
change) indicated that such a flow field might
be of particular value in a boiling system. It
was felt that, in comparison with other types
of flow fields, the gravitational acceleration pro-
duced by the rotating fluid would keep a concave
heated surface freer of bubbles at a given heat
flux by providing an increased buoyant driving
force for inward radial bubble migration. Delay
of the burnout condition to a higher heat flux

 

166, L. Muller, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 102.

74, F. Poppendiek and L. D, Palmer, Forced Con-
vection Heat Transfer in Pipes with Volume Heat Sources

Within the Fluids, ORNL-1395 (Dec. 17, 1952).

59
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27032

65 +0.24 N5

1

0.92

79 +0.226 Np,

1 2 5 10 2 5 108

 

Nge = 5000
me= 10,000

N, = 1,000,000

N, =100,000

5 10° 2 5 10t 2 5 10°

PECLET MODULUS, N,

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

value was foreseen as the primary advantage
of this phenomenon.

These early concepts have been justified to
a considerable extent by preliminary data taken
for the burnout heat fluxes for water in nucleate
boiling, both subcooled throughout and with small
net steam generation. These data are presented
in Table 1.7.3. Comparisons of these burnout
heat flux values with literature data for straight-
through linear water flow at the same conditions
of axial fluid velocity, L/D, exit pressure level,
and exit subcooling (defined as saturation tem-
perature minus actual liquid temperature, at
exit), indicate approximately a 500% superiority
for the vortex case. A typical comparison is
shown in Table 1.7.4, where test No. 5 by
Gunther '8 is compared with vortex flow run No. 4,
for which the pertinent experimental conditions
were almost the same.

 

18e ¢, Gunther, Trans. Am. Soc. Mech, Engrs, 73,
115~123 (1951).

60

 

If all Gunther’s conditions were identical with
those of ORNL run No. 4, it is estimated that
his flux value would be 2.3 x 10% Btu/hr-ft2, at
the most; thus the advantage for the vortex case
would be 378%. Gunther’'s values were chosen
for comparison here because they are consistently
among the highest reported in the literature and
should give a conservative evaluation. Gunther's
data were obtained for a rectangular flow passage
within which was suspended a thin (0.004-in.)
electrically heated metal strip or wafer that
was cooled on both sides. - This geometry gave
considerably higher burnout values for linear
water flow than those obtained in similar tests
with straight, round tubes.

The high burnout values reported are not
counterbalanced by an inordinately large power
expenditure, as is indicated by the ratio desig-
nated P//Pq 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
to the heat absorption rate of the water in the
same power units, As may be seen the heat

 
 

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 5 /p Burnout
Run  velocity  of Subcooling  Pressure Inside 9”“5'“ Heated . Quality "% Heat Flux
No. (fps) at Exit (°F) (psig) Diameter, D, Diameter Length, L L/Di Material () (%y** (Btu/hr-ft2)
(in.) (in.) (in.)

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

1.8 0 94 0.275 0.300 3.187 11.60 Copper 1.24 0.19 8,240,000

12.35 0 99 0.275 0.2875 (inlet) 2.875 10.44 Copper 1.49 0.20 9,690,000
0.300 (exit)

13.61 72 100 0.250 0.300 1.20 4.80 Inconel 0 0.43 10,980,000

13.18 135 248 0.300 0.375 3.05 10.16  Capper 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  Incanel 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 acrass test section to heat absorption rate af the fluid in the same

power units.

Table 1.7.4, Camparisan of Heat Transfer Data
for Yortex Flaw and Straight-Through Linear Flow

 

 

Vortex Straight-Through
Condition Flow* Flow**
Axial water velocity, 13.6 12.3
fps
Exit pressure, psia 15.0 14.4
Exit subcoeling, °F 72 71
L/Di 4.8 6
Burnout heat flux, 11.0 x 108 2.0 x 10°

Btu/hr-h2

 

*Data obtained by Gambill and Greene (see Table
1.7.3).
**Data obtained by Gunther (ref 18).

absorption rate ranges from 99 to 526 times the
test-section flow power.

An interesting observation was made recently
in connection with runs 6 and 7. In run §, the
exit pressure level was held at 100 psig (rather
than atmospheric as in all the other runs}), but
little change was noted in the burnout heat flux.
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
attained indicates that velocity effects are more

important than pressure effects. These studies

have been reported in greater detail in a recently
issued ORNL memorandum, 1?

TOWER SHIELDING REACTOR Il (TSR-Il)
HYDRODYNAMICS

W. R. Gambill

Detailed design work is nearly completed on
an enclosure assembly for a full-scale model
of a quarter segment of the TSR-Il to be used
in hydrodynamic studies. The data on the flow
distribution in the coolant annuli will be com-
bined with calculated heat-generation rates to
determine the magnitude of redistribution of flow
necessary to approximate a flat exit-water tem-
perature profile. In these tests the exit annulus
gap and the radial location of the control assembly
at the core center will both be varied. The
necessary pumps and instrumentation are being
obtained concurrently with design work on the
enclosure.

PHYSICAL PROPERTIES
W. D. Powers
Thermal Conductivity

The variable-gap th ermal-conductivity device
modified to include a heat meter of larger area

 

%W, R. Gambill and N. D. Greene, A Study of Burnout
Hea: Fluxes Associated with Forced-Convectzon, Sub-
cooled, and Bulk Nucleate Boiling of Water in Source-

Vortex Flow, ORNL CF 57-10-118 (ch. 29, 1957).

61
ANP PROJECT PROGRESS REPORT

has been used to redetermine the thermal con-
ductivity of the fuel mixture NaF-ZrF -UF  (50-
46-4 mole %, fuel 30). A preliminary valuve of
1.3 Btu/hr«ft:=°F has been obtained that is to
be compared with the previous values that range
from 0.8 to 1.6 Btu/hr«ft.°F. This study will
be completed after replacement of several de-
fective components in the circuit that supplies
power to the apparatus.

Enthalpy and Heat Capacity

Final results have been obtained for the en-
thalpies and heat capacities of the eutectic
mixtures LiCl-5rCl, (52-48 mole %) and LiCl-
BaCl, (70-30 mole %). The wide discrepancies
in enthalpy between duplicate samples reported
previously?? were resolved by making improve-
ments in the technique for mixing the two dry
salts. The equations describing the enthalpies
and heat capacities of these two mixtures are
as follows:

LiCI-$rCl,, (52-48 mole %)
Solid (99 to 493°C)
~3.52 + 0.140T +
+ (4.25 x 1073) 12
c, = 0.140 + (8.50 x 10-%) T
Liquid (550 to §52°C)

Hy = Higoc =

HT - H300C = '—39-8 + O.SSST -
- (8.15 x 10~5) 12
c, = 0.35% - (16.3 x 10-3) T
Fusion (525°C estimated)

H - H¢ =43

 

20y, p, Powers, ANP Quar. Prog. Rep, Sept. 30, 1957,
ORNL-2387, p 108.

62

LiCl-BaCl,, (70-30 mole %)
Solid (100 to 502°C)
~4.57 + 0.146T +
+ (2,50 x 10=%) 12

Hyp ~ Higoe =

0.146 + (5.00 x 10=3) T

It

C

P
Liquid (551 to 850°C)

Hp = Hyqoe = —6.67 + 0.278T -
— (3.93 x 10~%) 12
c, = 0.278 + (7.85 x 10-% T
Fusion (511°C)
H, ~ H, =49

In the expressions above

H = enthalpy in cal/g,

c heat capacity in cal/g-°C,

b
T

temperature in °C.

To date the enthalpies, heat capacities, and
heats of fusion have been determined for five
chloride mixtures containing lithium chloride.
In a manner analogous to that previously used
for fluoride-containing mixtures,?! it has been
possible to correlate these three properties for
the chloride salts, and the results for the five
mixtures are presented on a gram-atom basis in
Table 1.7.5. It may be seen that the maximum
deviation of any salt mixture from the average
for the five salts is 9%. The values per gram
are obtained by multiplying the listed per-gram-
atom quantities by the factor N/M, where N =
Tx;N, and M =F x.M,, x, being the mole fraction
of the ith component; N, the number of atoms
in the ith component; and M., the molecular
weight of the ith component.

 

2y, D. Powers and G. C. Blalock, Enthalpies and
Heat Capacities of Solid and Molten Fluoride Mixtures,

ORNL-1956 (Jan. 11, 1956).
PERIOD ENDING DECEMBER 31, 1957

Table 1.7.5. Properties of Binary Eutectics with Lithium Chloride

 

 

 

Enthalpy Heat Capacity
i Second LiCl y N N per gram-atom per gram-atom Heat of Fusion
Component (mole %) Solid at Liquid ot  Solid at Liquid at per gram-atom
. 300°C 700°C 300°C 700°C
. NaCl 72 46.89 2,00 23.45 678 8.23 1750 6760 2020
KCl 59 55.59 2.00 27.79  6.56 9.03 1860 6960 1830
CaCl, 62 68.46 2.38 28.77  6.76 8.31 1650 6560 1750
SrCl, 52 98.15 2.48 39.58  6.56 9.55 1670 6690 1700
BaCl, 70 92,16 2.30 40.07  6.45 8.94 1660 6770 1960
Average  6.62 8.81 1720 6750 1850
Maximum deviation from average 3% 8% 8% 3% 9%

 

63

 

 
Part 2
CHEMISTRY
W. R. Grimes

 
2.1. PHASE EQUILIBRIUM STUDIES'

C. J. Barton R. E. Moore

SUMMARY OF SALT SYSTEM INVESTIGATIONS
R. E. Thoma

Comprehensive studies of phase equilibria in
fused-salt mixtures have been made over a period
of years. These studies have been relevant to
(1) selections of fused-salt fuels, solvents, and
fuel-enriching mixtures, (2) development of fuel
reprocessing techniques, (3) considerations of
molten-salt secondary coolants, (4) structural
metal corrosion, (5) crystal structure relationships,
and (6) basic investigations of the phenomena
involved in fused-salt phase equilibria.

A nearly current report of the status of fused-salt
phase equilibrium diagrams was published re-
cently,2 and an extensive discussion of the
equilibrium relationships in some of the more
complex systems which have been investigated
will be available soon.3 Continved phase
equilibrium studies have provided, during the
quarter, sufficient data for essentially complete
diagrams of the following systems: KF-UF ,
RbF-UF,,  NaF-HfF, LiF-ThF,, NaF-ThF,,
NoF-LiF-Ber. For the most part, reliable liquidus
lines or surfaces are well known for the systems
that have been reported,? but for many systems
only tentative conclusions can be drawn regarding
the subliquidus phase relationships.

The search for improved fuel mixtures for
aircraft reactor use has been directed to areas of
low-melting-point regions of fluoride salt diagrams.
Characteristic features of most fluoride salt
diagrams that have strongly influenced the
selections of fuel systems have been the high
melting temperatures (800 to 1000°C) of 3MF.XF
and 2MF-XF, compounds, where MF is an alkali
fluoride and XF, is ZrF,, UF,, or ThF,. These
high-melting-point c0mpounds ond relafed high-
temperature ||qu1dus compositions in ternary

 

]The petrographic examinations reported here were
performed by T. N. McVay, Consultant, and R. A
Strehlow, Chemistry Division. The x-ray examinations
were performed by R. E. Thoma, Chemistry Division.

2R. E. Thoma and W. R. Grimes, Phase Equilibrium
Diagrams for Fused Salt Systems, ORNL-2295 (June 24,
1957).

3. E. Ricci, Guide to the Phase Dz'afirams of the
Fluoride Systems, ORNL-2396 (to be published).

R. E. Thoma

systems have frequently precluded otherwise
useful mixtures for fuels.

COMPARISON OF RUSSIAN AND ORNL
INVESTIGATIONS OF THE Nc:F-TI-nF4 SYSTEM

R. E. Thoma

A search for low-melting fuels requires «a
knowledge of binary phase diagrams for alkali
fluorides and quadrivalent fluorides. An example
which affords a comparison of published Russian
fused-salt data with the type of information de-
veloped for ANP fuels is the system NaF-ThF .
The Russmn phase diagram of the NaF- ThF
system4 is presented in Fig. 2.1.1, and the resuhs
of ORNL studies are given in Fig. 2.1.2. In the
ORNL investigations, because of the need for a
complete understanding of fuel behavior, all the
phase equilibria, including those below the
solidus, received attention. In Fig. 2.1.1,
Emelyanov and Eustyukhin have interpreted
thermal analysis data to indicate that in the
vicinity of 20 mole % ThF , the incongruent melting
temperature of 4NOF'Th& is 648°C, that the
4N<:|F-ThF4-2chF-T|-|F4 eutectic temperature is
620°C, that 4NaF-ThF , decomposes into NaF and
2N0F ThF , at 606°C, and that the a-2NaF.ThF

B-ZNQF ‘ThF , inversion occurs ot 568°C. At
ORNL results of x-ray diffraction and optical
examination of samples from thermal-gradient
quenching experiments have shown that, for the
above cases, no transition occurs at 648°C, that
the 4NaF.ThF -2NaF.ThF, eutectic temperature
is 606°C, and fhot 4NaF 4ThF4 decomposes into
NaF and 2NaF.ThF , at 568°C.

Emelyanov and Eustyukhin have assigned equi-
librium polymorphism to the compound 2NaF-ThF ,,
apparently wholly on the basis of the Zachariasen?
indication that two crystal modifications of this
compound exist. ORNL data show that there are
two forms of 2NaF.ThF ,; however, the hexagonal
O form is either metastable or hos equilibrium
existence below 400°C.

 

4v. s. Emelyanov and A. |. Eustyukhin, Atomnaya
Energ. 1, No. 5, 80 (1956).

W, H. Zachariasen, J. Am. Chem. Soc. 70, 2147 -51
(1948).

67
ANP PROJECT PROGRESS REPORT

68

UNCLASSIFIED
ORNL—LR—DWG 27033

HCCJEJOBAIIHE CHCTEM PACNIABIAEHHEX COMEHl HA OCHUOBE OTUPHIIA TOPHA

ThF,. 310 ykasmsaer, uto npu 606° (0 co-

eangesne  Na,ThF, pacnazacrea, ofpasyvw
NaF n cocnmrense Na,ThF,. Mpn H68°C
HPOMCXOAHT  TNONAMOpQHOEe  iIpeBpalleHMce
a-—B-Na,ThF,.  Cocamuenne Na,ThF, (u

touke ¢ 33,3 mon.% ThF,) naamurca ¢ or-
KPHTHM MAKCHMYMOM TIpH Temmepartype 7087 €.
JpreKTudeckan cmech cocaunenuit Na Th¥, u
Na,ThF, nnasurca npn 620°C u copepmur

K30° €. mpOHCXOMMT TIEPATEKTHYIECKAR peak-
UMA  PASTOMCHHA XHMNYECKOTO COeANHEHNA
NaTh,F, (66,7 mon.% ThF,), coctan xoroporo
HOJITBCPIKABETCA, B TACTROCTR, TEM, 9TO Tel-
soroi addert npr 8I0° maeer MakcEMamBEYI
BeNMAMAY B Todke okono 67 wmox. % ThF,
u BOTHIM HTOr0 cocTaBa mewe3daer e unan-
JAeHHA IBTeKTHKE npr 700° (.

Taxam ofipasow, B cmcreme NaF . ThF,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

24 mon.% ThF, HMCCTCA  9eTHPE  XHMHICCKAX  COOMMNONNA:
/ 065°
/ o0°
992"\ /(_
o N\ A K- ThE a0’
/ 830°
3m° XeNaF - - ] ° dm'
%0° X +WaTh A
Na,ThF, . 756‘; T ayse
o’ o th\ B W eWaThF, 70"
suge”, oW e cel NaThFThE,
J NaF*MaThF, oA ThEaNa,TAF, | .
500" 4 — 800
Na, ThEset MY, Lo
501 i hE e\ NaThEeNaTh,F,
ol NaThF; | . {NaTh,F, o’
NaF~ANa,ThF, |
|
| man B ThF,
0 10 20 30 40 50 60 70 80 9% 007%
Puc. 1. Cuetena NaF- ThF,.

Jlanee JuHMA NWKBHAYCA TOCTCNCHHO 110-
HUKAETCH JI0 CACAYWIICH IBTCKTAYCCKOW TOMKM
npr 700°C w 36 -37 mon.% ThF,. Hagnnan
¢ 36,5 mon.% ThF, w nanee, tennosoir ad-
q’CHT aaaBJACHAA IPTCKTHKHN }'M(‘Hh["flffl‘(‘fl
n ucgeaact okono M mom.% ThF,.

o COBOKYNHOCTH JAHHHX TepMHAYCCKOI'D
M MHKPOCTPYKTYPHOTO AHAJM3A CACAYCT, 9TO
B obnactH okoso U mon.% ThF, eyviuecrayver
XHMHYECKNe COCIAHHCHEE NPEANOIOMKHUTCILHOIO
cocrasa NaThF, umewomee sechma nogorai
MAaKCEMYM, . |HHAA JRKBAAVCA OT COPMHEHHA
NaThF, vesnaumTennHO nmouumaerca Q0 3BTEK-
tkr mpu 7K° Cn 53,5 moa.% ThF, u naace
PE3KO UOQHMMACTCA BBEPX, TMPCTEPIEBAA TPH
830° C r 62 mo1.% ThF, maiom, nocne koto-
poro oHa maaBHO noaamMaetrcs po 1065° C
(remnepatrypa naasiaenma wucroro Th¥,). Ilpu

i «ATOMHaA sHeprudAs N\ &

AT

Na,Th¥, Na,ThF, NaThF; w NaTh,F,. Co-
¢1nHcHue Na 'l‘hF. € VIHeCTBYET B ABYX MoMw(E-
xanaAax: «-Na,ThF, B-Na,ThF,, =, wosmna-
MOMY, B pACIJIABJIEHHOM COCTOHHEEN; COSKE-
nengs Na,ThF, u NaTh,F, pacnagmaorca mo
NePUTCKTHYCCKHM  PCaKUAAM. OTH  JeEEMe
NOATBEPKARKT  CTPYKTYPHHE MCCAeAOBAaHER
Jdaxapuaacra [1]. Coegunenwe wpeamONOXN-
tenrHoro coctapa NaThF, umeer B ofnacra
or 44 o S0 wmon.% ThF, owens monormk
MAKCAMYM Ha JHHHH JTHKBHAYCA, MO KOTOPOMY
TPY/AHO YCTAHOBHTL TOYHWH COCTAB.

Hecaenopanme cumoremm KF - ThHF,

Buao maroroneno w mcenemosano (5], [EJ
cenme 40 conasos yepes arrepsan 2-—3 mon.

ThF,. Ha ocuomanmu aTex mccaenoranmit Guta
NOCTPOCHA JAATPEMMA COCTOAHHA  CHCTOME

Fig. 2.1.1. The System NaF-ThF4 According to Emelyanov and Eustyukhin,

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL—-LR-DWG 18964

 

 

 

{100 / /
{000 //

 

 

80O

 

TEMPERATURE (°C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

700 v
600 /
u_(f
= <
500 = L A £
5 - - o
g 5 i i
3 2 2
400
NaF 10 20 30 40 S0 60 70 80 90 ThF,

ThF, {(mole %)

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

The Russian diagram shows a congruently melting
compound NaF+ThF ,. ORNL data show that the
compound NaF:ThF, melts incongruently to
NaF 2ThF , and liquid.

Unciassi?ied Russian reports presently represent
the predominant number of contributions to the
literature of phase equilibrium studies of fused
salt mixtures. Since 1946 more than 200 Russian
reports have been issued® concerning the phase
equilibria of oxide, silicate, chloride, nitrate, and
fluoride systems.

Most Russian phase work is done with thermal
analysis and differential thermal analysis tech-
niques. Phase identification is accomplished for
the most part with x-ray diffraction. No reports

 

SA.G. Bergman, K. A. Evdokimova, and O. F. Bogush,
lzvest. Sektora Fiz.-Khim. Anal., Inst. Obshchei Neorg.
Khim., Akad. Nauk S.S5.S.R. 27, 419-56 (1956).

indicate the use of the thermal-gradient quenching
techniques or of routine petrographic examinations.

THE SYSTEM NuF-CrF2
H. A. Friedman C. F. Weaver

Phase equilibria in the NaF-structural metal
fluoride systems NaF-CrF,, NaF-FeF,, and
NaF-NiF, have been studied in an effort to
elucidate the corrosion process as it is related to
these materials. Phase diagrams for the systems
NaF-FeF, and NaF-NiF , have been reported.”’

No fused-salt melts from thermal analysis experi-
ments or from thermal-gradient quenching experi-
ments have produced pure NaF-CrF, samples that
are free of Cr***. A variety of means has been

 

’H. A. Friedman and B. S. Landau, ANP Quar. Prog.
Rep. June 30, 1957, ORNL-2340, p 128.

69
ANP PROJECT PROGRESS REPORT

employed to minimize the effects of the dispro-
portionation reaction

3CrF2;\‘_2CrF3 + Cr°

in these phase equilibrium studies. Even the
inclusion of 20 to 30 wt % chromium metal powder
dispersed through samples for thermal-gradient
quenching has not suppressed the forward reaction
sufficiently for NaF-CrF, phase equilibrium data
to be derived.

APPARATUS FOR HIGH-TEMPERATURE
THERMAL-GRADIENT QUENCHING
EXPERIMENTS

H. A. Friedman

A design modification of the apparatus used
in thermal-gradient quenching experiments has
permitted the feasible upper operating temperature
limit to be raised from 850 to 1300°C. Previous
difficulties encountered with quenching apparatus
when operated routinely at 900 to 1000°C included
the frequent need for replacement of the 18 thermo-
couples which monitored the thermal gradient, the
rupture of the sample tube because of oxidation
corrosion, uncertainties as to temperature constancy
of the sample because of discontinuous monitoring,
and sample tube expansion during heating.

In the revised apparatus a single, traveling,
recording thermocouple was substituted for the
18 monitoring thermocouples. The furnace was
redesigned to permit the sample tubes to be
protected by a static helium atmosphere during
heating. The use of a roller crimping machine
contributes to minimizing sample tube expansion.

The apparatus, shown schematically in Fig.2.1.3,
now consists of a round nickel block 10 in. in
length and 2 in. in diameter, which is supported in
a 2]/5-in. tube furnace that has a platinum wound
core. Interchangeable gears allow a choice of
excursion times for the traveling thermocouple.
A complete thermocouple travel cycle of 2 hr is
usually employed in using this apparatus.

The roller and crimping machine (Fig. 2.1.4)
is designed so that a é-in.-long, 0.10-in.-OD,
0.010-in.-wall nickel tube containing a sample

8,9

 

BC. J. Barton er al., ANP Quar. Prog. Rep. Dec. 10,
1954, ORNL-1816, p 57.

9C. J. Barton et al., ‘"Phase Equilibria in the Systems
LiF-UI’:4 and NoF-UF4," J. Am. Ceram, Soc. (in press

for issuance in Feb. 1958).

70

to be quenched is flattened and crimped at V-in.
intervals. The sample tubes are sealed by welding
after they have been rolled and crimped. Sample
tubes prepared in this manner have better heat
transfer characteristics than those of the tubes
used previously. Much less expansion of the tubes
during heating occurs because the void space is
also reduced in the crimping process.

PETROGRAPHIC ANALYSES OF FUEL
MIXTURES

R. A. Strehlow

The petrographic microscope has been used on a
routine basis for the inspection of solidified fuel
mixtures.19 This type of examination is performed
to detect the presence of oxygen-containing com-
pounds and, for some samples, to determine any
variance from the nominal composition. Mixtures
of NaF-ZrF,, NaF-Z¢F ,-UF,, LiF-NoF-KF-UF,,
and related compositions have formed the bulk of
samples submitted for inspection. An illustration
of the results obtained with the technique used in
the inspection of these samples is given here for
NaF-ZrF , mixtures containing about 50% ZrF ,.

Mixtures in the NaF-ZrF, system which contain
between 46.2 and 57.1 mole % ZrF , will contain
some 3NaF:4ZrF, which will, in general, occur
in the slowly cooled melt and will be dispersed in
eutectic aggregates that also contain 7NaF-6ZrF .
The aggregate frequently will exhibit a brown cast,
which has been attributed to hydration of
3NaF-4ZrF ,. Better crystallized 3NaF«4ZrF ; has
been observed in samples containing about
56 mole % ZrF,. The material present as the
matrix in the eutectic aggregates is the compound
with the reported formula of 7NaF-6ZrF4.

This compound occurs as the primary phase in
mixtures containing from 46.2 to 49 mole % ZrF ,.
It is characterized as a crystalline uniaxial
substance that is optically negative with N_ = 1.502
and N_, = 1.508, where N_ and N_ refer to the
ordinary and extraordinary indices of refraction of
uniaxial crystals,

Compositions containing about 50 mole % ZrF,
which have been cooled at intermediate rates
occasionally contain NaF«ZrF , as a nonequilibrium
phase that is uniaxial positive with N_ = 1.417 and

 

101, techniques described were developed over a
period of several years by T. N. McVay, G. D. White,
H. Insley, and B. 5. Landau.

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 25133

\\\\\\\\

\\\\\\\\\

FURNACE INSULATION

.
LSS ST S S A SIS S S

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ -

 

o

-

 
  

   

       

   

  

 

Free e NG
: T \\\\\ .

I e s e
0 0 "

 

HANDLE TO

 

TO REGORDER

SAMPLE TUBE

 

 

 

 

 

 

 

 

 

 

 

 

TO CONTROLLER

e
——T

 

 

   

 

 

 

AL .

  

 

 

N N N T N I S R NN T R A RN TR RN R NN N

<

 

     

      

AU G
: H[J N e
4 o :

G

I

 

 

 

     

    

S
/////

\
/

 

 

 

  

 

CONTROLLING THERMOCOUPLE

Fig. 2.1.3. Thermal-Gradient Quenching Apparatus,

71
ANP PROJECT PROGRESS REPORT

UMCLASSIFIED
PHOTD 20159

— —

-
—

 

Fig. 2.1.4. Rolling and Crimping Machine for Pre-

paring Thermal-Gradient Quenching Samples.

N, = 1.446. This is rare, however, in slowly
cooled fuel mixtures.

At higher NaF concentrations than 53.8 mole %,
crystals of the solid solution between 7NaF.6ZrF
and 3NaF-2ZrF, are observed which are, in
general, poorly crystallized and which may be
distinguished from 7NaF«6ZrF, crystals by their
lowered indices of refraction, their crystalline
character, and their poor extinction when examined
with crossed polarizing disks.

Oxygen will, in general, be observed as the
oxyfluoride, which is wunioxial negative with
indices of refraction greater than 1.54 to 1.7 and
which frequently crystallizes as hexagonal plates.
Zirconium oxide has been observed, and it is
easily seen in samples of this type because of its
pronounced angular cleavages, frequent high
birefringence, and very high refractive indices

(higher than 2.13).

72

The addition of UF, may introduce, as a new
material, a UF4-ZrF solid solution which occurs
in mixtures containing less than 50 mole % NaF
ond is biaxial negative (2V = 50 deg)'! with
marked birefringence and refractive indices of
around 1.6. If UO, is present it is optically
isotropic and garnet red in color. Partial reduction
of UF, to UF, results in a yellow-brown cast to
the otherwise green crystals.

AN OPTICAL METHOD FOR THE
DETERMINATION OF EQUILIBRIUM PATHS
AND TIE LINES IN TERNARY
FUSED-SALT SYSTEMS

C. F. Weaver
Systems Without Solid Solution

If the liquidus temperatures and the boundary
curves in a ternary phase equilibrium system are
known, and if there is no solid solution, all the tie
lines for any over-all composition and temperature
can be drawn. The tie lines will then describe the
equilibrium path (a straight line) and indicate the
composition of the liquid in equilibrium with a
primary phase as a function of temperature and
over-all composition.

Systems With Solid Solution

If there is solid solution, knowledge of the
liqguidus temperatures and the boundary curves is
insufficient to allow the tie lines to be drawn.
The difference between the case without solid
solution and the case with solid solution lies in
properties of the primary phase. In the case of
no solid solution, the composition of the primary
phase is given on the phase equilibrium diagram,
and it is independent of temperature. In the case
of solid solution, the composition of the primary
phase is not specified on the phase equilibrium
diagram. In addition the composition of the primary
phase is, in general, a function of temperature.

Under certain conditions the composition of the
primary phase as a function of temperature can be
determined experimentally for each over-all compo-
sition. This information together with the liquidus
temperatures can be used to draw the tie lines.
As in the case of no solid solution, the tie lines
will describe the equilibrium path (in general, a
curved line) and indicate the composition of the

 

Mn this expression, 2V refers to the acute angle
between the axes in biaxial erystals.

 
liquid in equilibrium with a primary phase as o
function of temperature and over-all composition.

Index of Refraction vs Solid Solution Composition

If a series of fused-salt solid solutions of known
composition are prepared, their indices of refraction
can be measured by means of a microscope and
svitable immersion oils. A hypothetical plot of
index of refraction vs composition for a continuous
fused-salt solid solution between A and B is given

in Figo 20]05.

UNCLASSIFIED
ORNL-LR-DWG 25265

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N
S // Y
-
L | "]
E /( emem e e N,
w A, T
e

fi //
a N —
E a

A COMPOSITION B
Fig. 2.1.5. Hypothetical Plot of Index of Refraction

vs Composition for a Continuous Fused-Salt Solid Solu-
tion Between A and B.

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
Fige 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
of AB is plotted vs temperature, as in Fig. 2.1.6,

PERIOD ENDING DECEMBER 31, 1957

and the curve is valid, of course, only at point P.
Hence a series of such curves must be constructed
to suitably cover the primary phase area of
interest.

Constructing Tie Lines ond Equilibrium Paths

if the over-all composition is considered to be
that shown in Fig. 2.1.7 and it is assumed that
Fig. 2.1.6 gives the solid solution composition

UNCLASSIFIED
ORNL-LR-DWG 25264

 

 

o

 

 

 

COMPOSITION
o

 

o

 

LIQUIDUS AT P

 

 

 

 

BOUNDARY CURVE

 

-
N o ——————— o

3
TEMPERATURE

Fig. 2.1.6.
Poinf Pt

Composition of AB vs Temperature for

UNCLASSIFIED
ORNL-LR-DWG 25263

   

LR
Qfififiuv‘vfie
(RTINSO
V?mm.wmmuvfi;

R
NN/
KPR
AWA%“&WV?G@“#
N/ VAVAWA\AVAV#V# 5" VA
AAVAAVA{AVAVA(AVA

   

§

    
    
 
     
 
      

  

73
ANP PROJECT PROGRESS REPORT

vs temperature plot for P, at temperature 1 the
composition of AB solid solution is 4, the compo-
sition of the liquid is P, and the tie line is Pa.
Likewise for temperatures 2, 3, and 4, the compo-
sition of AB solid solution is b, ¢, and 4, the
composition of the liquid is 4% ¢’ and d’, and
the tie lines are b°b, c’c, and d'd, respectively.
The locus of all points “such as P, b° ¢/, and d°
is the equilibrium path.

The System LiF-ThF,-UF,

The system LiF-ThF -UF, appears to contain
three continuous solid solutions: UF4-ThF4,

74

LiF-4UF ,-LiF:4ThF ,, and 7LiF-6UF ,-7LiF-6ThF .
In the UF ,-ThF, solid solution the plot of index
of refraction vs composition has a small slope.
As previously mentioned, tie lines cannot be
accurately determined when the slope of the plot
is small. In the remaining two solid solutions,
however, the slopes are sufficient for the de-
termination of tie lines. These solid solutions
offer ample opportunity for a study of the tie lines
and equilibrium paths, and such studies have been
initiated.
 

PERIOD ENDING DECEMBER 31, 1957

2.2. CHEMICAL REACTIONS IN MOLTEN SALTS

F. F. Blankenship
R. F. Newton

THE PREPARATION OF UF, AND ITS STABILITY
IN MOLTEN FLUORIDE MIXTURES

F. F. Blankenship

The use of UF, in fuel mixtures was found to be
impractical because disproportionation occurred
and uranium metal deposited on the container
surfaces. The work that led to this conclusion and
the results of investigations of methods for elimi-
nating or
marized here. The mechanism for disproportionation
has been ascertained and means for controlling it
are suggested. Petrographic and x-ray diffraction
examinations were used in this study to characterize
phases in solidified melts, and liquid compositions
were, in general, defined by synthesis or by fil-
tration and chemical analysis.?+3

avoiding disproportionation are sum-

The UF, compound may be prepared in solution
by various redox reactions or by electrolysis, but,
when large quantities were needed for corrosion

studies, solid UF, was used as the starting
material. The reaction
(1) 3UF, + U= 4UF,

was used to produce routinely kilogram batches
of 99% UF,. Essentially complete reaction of
stoichiometric mixtures of UF , and U° was obtained
in 48 hr at 900°C.4

The equilibrium pressure of UF, above UF,
as a result of disproportionation by the reverse
reaction in Egq. 1 was measured at temperatures
between 1270 and 1400°C in tantalum-lined vapor

pressure apparatus.  The results obtained are
represented by the expression®
4187
log p (mm Hg) = - — + 3.945 ,

 

. Iinsley 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.
March 10, 1955, ORNL-1864, p 53.

3M. B. Panish, ANP Quar. Prog. Rep. June 10, 1955,
ORNL-1896, p 69.

4B, J. Sturm and E. E. Ketchen, ANP Quar. Prog. Rep.
March 10, 1955, ORNL-1864, p 66.

3s, Langer, ANP Quar. Prog. Rep, June 10, 1955,
ORNL-1896, p 64.

L. G. Overholser
G. M. Watson

where T is in degrees Kelvin. The equilibrium
thus obtained indicated that crystalline
UF, is more stable than published estimates of
the free energies of formation would indicate. In

values

contrast, UF, was found to disproportionate in
molten fluoride solvents to a much greater extent
than it did in the solid-crystal form. Fartially
complete phase diagrams were prepared® that
useful in approximating the solubility be-
havior of UF; in binary mixtures of alkali fluorides.
The solubility of UF, in NaF-ZrF, and in NaF-
BeF,, which are mixtures of interest as reactor
fuels, is 1to 2 wt % U at 600°C.7

Nonacidic fuel solvents that contain a high con-
centration of fluoride donor constituents, such as
alkali fluorides, were found to increase the ex-
tent of disproportionation by complexing the acidic
UF, (fluoride acceptor). The ease of alloy for-
mation between the uranium formed and the con-
tainer metal also noticeably influences the extent
of the disproportionation.® Less disproportionation
occurred in copper vessels than in nickel vessels,
and tantalum and molybdenum were found to be
even more resistant than copper to alloying with
the uranium formed.

It was also found that UF, was oxidized by the
alkali fluorides. The reaction was particularly
noticeable with mixtures containing KF:

were

(2) UF, + KF — UF, + K°

For example,? when a solution containing 21.4
wt % U as UF, in the LiF-NaF-KF eutectic (46.5-
11,5-42 mole %) was heated under helium in copper
apparatus, about one-third of the UF, was lost by
disproportionation in 2 hr at either 650 or 750°C;
the amount of UF ; oxidized by KF at the same time
increased from 7 to 18%. Sufficient reduction in
the helium pressure caused a rapid increase in the

 

6¢c. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10,
1954, ORNL-1771, p 56.

7L. M. Bratcher ez al., ANP Quar. Prog. Rep, March 10,
1955, ORNL-1864, p 50.

8¢. J. Barton and B. H. Clampitt, ANP Quar. Prog.
Rep. Sept. 10, 1955, ORNL-1947, p 78.

98. H. Clampitt, ANP Quar. Prog. Rep. March 10,
1956, ORNL-2061, p 97.

75
ANP PROJECT PROGRESS REPORT

U***.  The sharp in-

crease in the rate occurred at a pressure corre-

rate of disappearance of

sponding to the bubble pressure of potassium. The
small apparent temperature coefficient of the dis-
proportionation in such experiments was probably
the result of the process being limited, after an
initial rapid deposition of uranium, by the slow
rate of diffusion of uranium into copper.

In comparison with UF j~containing fuels, the
UF ,-containing fuels exhibit o marked improvement
in initial corrosion attack at high temperatures.'?
Moreover, the use of UF, in conjunction with UF,
could provide reserve reducing capacity to com-
pensate for any tendency of the fuel to become
more oxidizing during burnup. Unfortunately the
difficulties arising from the deposition of uranium
by disproportionation prevent the easy exploitation
of UF; as a solution to corrosion problems. How-
ever, even in the UF, fuels, some UF, always
appears as a product of the oxidizing action of
UF,, and the fate of this UF, is of interest in
connection with long reactor lives.

ACTIVITY COEFFICIENTS OF FeF, AND OF
NiF2 IN MOLTEN NoF-ZrF,

C. M. Blood

The activity coefficients!!+12 of FeF, and of
NiF, in NaF-ZrF, (53-47 mole %) were determined
by comparing the values of the equilibrium quotients
of the reaction

MF ,(d) + H,(g) =M°(s) + 2HF(g)

with the corresponding values of eauilibrium con-
stants calculated from the tabulated!? free energies
of formation of FeF,, NiF,, and HF. Experimental

values for the equilibrium auotients for the re-

actions of FeF, ond of NiF, were shown to be in-
dependent of concentration. For the FeF, reaction
the values of K _are 2.6 t 0.2, 0.63 + 0.04, and
0.097 £ 0.007 atm at 800, 700, and 600°C, re-
spectively. These values are in essential agree-
ment with equilibrium constants calculated from

 

105, H. DeVan and E. A. Kovacevich, ANP Quar,
Prog. Rep. Dec, 10, 1955, ORNL-2012, p 107,

“C. M. Blood and G. M. Watson, ANP Quar. Prog.
Rep. March 10, 1956, ORNL-2061, p 84.

12¢, M. Blood, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387,% 117.

]3L. Brewer et al., Natl. Nuclear Energy Ser. Div. IV
198, pp 65, 110 (1950).

76

the standard free energies of formation of pure
solid and supercooled FeF, at these temperatures.
Accordingly, the activity coefficients for FeF, in
this solvent are near unity, regardless of the choice
of standard state. The numerical values of the
activity coefficients of FeF, are, at 800, 700, and
600°C, 1.46, 2.20, and 3.28 with respect to the
pure crystalline solid, 0.64, 0.66, and 0.66 with
respect to liquid FeF,, and unity with respect to
the standard state of reference at infinite dilution,

Equilibrium quotients for the NiF2 reduction
have been determined to be 21,900 + 2,000,
15,300 + 700, 11,000 £ 600, and 7,600 * 400 atm,
as compared with the values 10.9, 7.4, 4.8, and 3.1
calculated from the free energies of formation of
pure solid NiF, at 625, 600, 575, and 550°C,
respectively. The corresponding constants calcu-
lated for liquid NiF, are 43.8, 33.6, 25.2, and
18.4. The very large discrepancies between the
calculated and experimental constants give rise to
very high values of the activity coefficients of
NiF,. By using the solid as the standard state
the numerical values of the activity coefficients
were found to be 2009, 2076, 2292, and 2468 at
625, 600, 575, and 550°C, respectively. With
supercooled liquid NiF, as the standard state, the
corresponding values of the activity coefficients
are 500, 455, 436, and 413.

It is clear from the values of the activity co-
efficients that when NiF, is dissolved in NaF-
ZrF , (53-47 mole %) it behaves in a manner vastly
different from that predicted by the free energies
of formation for solid NiF,. This is somewhat
surprising in view of the agreement shown to exist
between the calculated and experimental behavior
of FeF, in the same solvent. Accordingly, the
results of these investigations confirm that the
tabulated values of the free energies of formation
of solid FeF,, the tabulated melting point, and
the estimated heat of fusion are adequate to pre-
dict the behavior of FeF, in the solvent NaF-ZrF
(53-47 mole %). On the other handthe corresponding
values for NiF, are not at all suitable to describe
its behavior in the same solvent.

In the case of FeF, the experimentally determined
free energy of formation at 600°C for the standard
state of reference at infinite dilution was -135.4
kcal/mole, as compared with the tabulated values
-137.5 and =134,5 for the solid and liquid states,
respectively.  Thus the experimental value for
FeF, was found to be between the predicted
values for the solid and the liquid states. On the

 
other hand, the experimentally determined free
energy of formation of the standard state of ref-
erence for NiF, was found, at 600°C, to be ~114.6
kcal/mole, while the tabulated value for the solid

state is =127.9.
14

As has been pointed out else-
where,'? at first glance it appears that the value
of the free energy of formation of the solid NiF,
is in error by approximately 13 kcal.
however, that the solubility of NiF, is quite low in
in this solvent,!® 0.17 mole % at 600°C; and, if
the saturating phase is pure NiF,, an activity co-
efficient of 588 would be required to give unit
Since the saturating phase

It is known,

activity at saturation.
is not pure NiF, but rather a complex fluoride, it
is only possible to predict from the solubility an
upper limit of 588 for the activity coefficient for
solid NiF, as the standard state. The upper limit
of 588 corresponds to a value of ~125.7 for the
free energy of formation of NiF,. Conversely, it
may be stated that the upper limit at 600°C of the
free energy of formation for solid NiF, is —125.7
kcal/mole, as given by its solubility, the experi-
mental K , and the tabulated free energy of for-
mation for gaseous HF. The magnitude of a more
reasonable value will now be considered.

In order to obtain an activity coefficient in the
neighborhood of unity with
NiF, as the standard state, the free energy of
formation must be approximately ~114.6 kcal/mole
at 600°C. The free energy of formation of solid
NiF, at 600°C, based on an assumed free energy
of formation of liquid NiF, of -114.6, may be
computed according to the equation

supercooled liquid

o - o
AFNin(S) - AFNiFi,(l') - AFMehing '

If the heat of fusion is assumed to be independent
of temperature, then

O
A‘FNiFl,(s)

_ o
- AFNin(l) - ASM (TM - T) .
where AS, = AH/T,, the entropy of fusion, and
T, is the melting point of solid NiF .

In previous calculations,'? a heat of fusion of

8000 cal/mole and a melting point of 1300°K for

NiF, were assumed, as suggested by Brewer.'3

 

}4M. Biander and F. F, Blankenship, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 121.

15¢, M. Blood, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 99.

PERIOD ENDING DECEMBER 31, 1957

These values give a free energy of formation for
solid NiF, of -117.2 kcal/mole. However, some
recent work indicates that the melting point of
NiF, is greater than 1500°C (ref 16). |f the en-
tropy of fusion for FeF, (8000/1375 = 5.82 eu) is
assumed for NiF,, but a melting point of 1800°K is
used, a value of ~120.0 kcal/mole is obtained as a
reasonable estimate of the free energy of formation

of solid NiF2 at 600°C.

SOLUBILITIES AND ACTIVITY COEFFICIENTS
OF NiF, AS FUNCTIONS OF SOLVENT
COMPOSITION IN NoF-ZrF, MIXTURES

C. M. Blood

It has been postulated!7+12 that the activity co-
efficients of a solute in a fluoride solvent will
reach a maximum at the isoelectric point or when
the charge-to-distance ratio of the solute's positive
ion is equal to the average or effective charge-to-
distance (Z/R) ratio of the positive ions of the
solvent mixture. As the composition of the solvent
is_changed and its charge-to-distance ratio is made
to differ from that of the solute, negative deviations
On this basis a
tentative acid-base scale for predicting isoelectric
points of solutes in NaF-ZrF, melts was pre-
pc:red.18 Since not much experimental evidence is
available on the effect of solvent composition on
solute activity coefficients, the proposed plot is
only tentative, since no proven way exists to ob-
tain the effective Z/R of the solvent as a function
of composition. In an effort to providé quantitative
information, hydrogen-reduction equilibrium quo-
tients of NiF, were measured in various NaF-ZrF,
mixtures. Activity coefficients were calculated,
but, since no satisfactory value of the free energy
of formation of either solid or liquid NiF, is avail-
able, the data are presented (Table 2.2.1) as the
ratio of the activity coefficient of NiF, in the
solvent indicated to the activity coefficient of
NiF, in NaF-ZrF, (53-47 mole %). In addition,

the solubility data are given as a function of

from ideality become greater.

solvent composition.

 

16, A, Friedman, private communication to C. M. Blood.

”R. F. Newton and F. F. Blankenship, ANP Quar,
Prog. Rep. March 31, 1957, ORNL-2274, p 123.

8y G, Hill and F. F. Blankenship, ANP Quar, Prog.
Rep. March 31, 1957, ORNL-2274, p 126.

77
ANP PROJECT PROGRESS REPORT

Table 222.1. Equilibrium Quotients, Activity Coefficient
Ratios, and Solubilities of NiF2 in NaF-ZrF
Mixtures at 600°C

 

Solvent Activity

 

Composition  CaUIlibrivm o ghiciony  Solubility
(mole % ZrF ) Quotient Ratio* (ppm Ni)
x 103
57 4.6 0.30 450
52 9.2 0.60 540
47 15.3 1.00 980
44 18.5 1.21 1380
40 25.3 1.65

 

*Based on activity coefficient of NiF2 in Nr.uF-ZrF4
(53-47 mole %).

The experimental measurements of the 40 mole %
ZrF, mixture have not been completed, and the
data given in Table 2.2,1 for this composition may
be revised. It is proposed to maoke measurements
on the solvent containing 37 mole % ZrF, in an
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
the solvent may correspond closely tothat predicted
by the proposed acid-base scale.'® It may also
be noted that the value of the activity coefficient
of the NiF, solute is not dependent on the solu-
bility; this confirms a previous conclusion that the
precipitating phase is not NiF,,.

REDUCTION OF UF, BY CHROMIUM
IN MOLTEN FLUORIDES

J. D. Redman

The possible use of Inconel or stainless steel as
a material for containing UF ,-bearing fluoride
mixtures raised numerous questions as to the
effect of corrosion attack in the temperature range
600 to 800°C. The most active metal present in
these alloys is chromium, and consequently any
fundamental information relating to the interaction
of chromium with UF, dissolved in representative
fluoride mixtures should be applicable to an under-
standing of the corrosion of the alioys by the
fluoride melts.

A summary of the results of studies of the re-
duction of UF, by chromium in various fluoride

78

reaction mediums is presented here, and data for
the reaction of UF, with various other metals are
given in the next section,. Some data are also given
for the reaction of UI"'4 with Inconel, in which the
activity of chromium is much less than unity.

As might be expected, the reaction of UF, with

chromium is influenced by the reaction medium
employed. The reaction which occurs in the
ZrF ;- and BeF,-bearing melts studied may be
written as

2UF, + Cr°= 2UF, + CrF,

The valency of the chromium compound has been
established through numerous experiments, and
it has also been demonstrated experimentally that
the solubility of the uranium and chromium com-
pounds involved is considerably greater than the
concentrations encountered in these studies. The
equilibrium quotient, K., defined as

where N denotes mole fraction, may be evaluated
for this system by adding a known concentration of
UF , to chromium and determining the resulting con-
centration of Cr**. Side reactions that arise from
impurities (H,0, HF, Fe** Ni*H must be eliminated
or suppressed to such a degree that the chromium
blank is small compared with the total chromium
measured. The blanks in these studies ranged from

150 to 250 ppm Cr.

A filtration method was used in which the melt
was equilibrated under designated conditions and a
sample was withdrawn through a sintered-nickel
filter and analyzed for chromium. Nickel apparatus
was used in most cases, because early experi-
mental work demonstrated that nickel is stable
toward UF, dissolved in fluoride mixtures. Equi-
librium is attained in these systems fairly rapidly,
3 hr being sufficient in virtually all cases, al-
though a longer period of heating was utilized for

some studies.

The data given in Table 2.2.2 demonstrate that
the equilibrium quotient, K , is essentially con-
stant over the range of concentrations of UF, and

UF ; studied with NaF-ZrF, (50-50 mole %) as the
PERIOD ENDING DECEMBER 31, 1957

Table 2.2.2. Equilibrium Quotients for the Reaction 2UF4(d) + Cr(c);_—‘_2UF3(d) + Cer(d)
in NaF-ZrF , (50-50 mole %) at 600 and 800°C

 

Conditions of

 

 

Original Equilibrium Equilibri
Equilibration Concentration Chromium quitibrium
(mole %) Quotient,*
Temperature Time mole 7% Concentration K
(°C) (hr) UF, UF, (ppm) x
x 10~4
600 3or5 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 3ord 4.1 2550 * 210 59 +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 Kx were obtained from the arithmetic mean of two to four measurements,

reaction medium. As anticipated, the chromium con-
centrations decrease as the UF; concentration in-
creases; this suggests that the extent of attack on
chromium or chromium-bearing alloys may be de-
creased by the addition of UF3. The reaction has

Table 2.2,.3. Equilibrium Chromium Concentrations for
the Reaction of Chromium with UF4 in NaF«LiF-KF

(11, 5-46, 5-42 mole %) ot 600 and at 800°C

 

 

a small positive temperature coefficient; an average Conditions of Original Equilibrium
heat of reaction of 6 kcal in the temperature range Equilibration Concentration Concentration*
600 to 800°C is obtained from t.he values for K’E' Tommerators  Time (mole %) 0 .
Values of 70 and 20 were obtained for the equi- o —_—
librium constants at 600 and at 800°C, respectively, ¢ ()  UF, UF3 (vt%)  (ppm)
by using .'rhe available values of the free energies 500 3 2.30 12.0 1130
of formation for CrF,, UF,, and UF,. The corre-
sponding equilibrium concentrations were 8100 and 5 230 1.7 1050
7300 ppm. A comparison of these values with 5 2.00 0.24 11.0 205
those given in'Table 2.2.2 ind'icates fhat' this 12 200 024 1.4 250
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 UF, with Cr° cannot be
p:,ecisely defined. A mixture of Cr** and Cr*** > 2% 1. 2750
results, but the ratio of the two ions has not been 5 2.00 0.24 1.1 1850
accurately determined. Equilibrium quotients 12 2,00 0.24 10.8 1750
cannot, therefore, be evaluated, but, rather, the 5 110 1.20 10.2 140
data are presented in terms of total dissolved

5 0.70 1.60 10.8 40

chromium (Cr*t + 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 UF,
concentration increases. The large difference in
chromium concentrations at 600 and at 800°C in-
dicates that the heat of reaction in this medium is
much greater than that calculated for the reaction
in NaF«ZrF, and suggests that the mass fransfer
of chromium is more severe in the former case.

The chromium concentrations for equilibrium with
Inconel can be calculated by using an activity
valve of 0.1 for chromium in the alloy and the
chromium concentration values given in Table
2.2,2. The values thus obtained are 1050 and
1150 ppm at 600 and at 800°C, respectively, in
the reaction medium NaF-ZrF -UF,. These con-
centrations are much lower than that given for
equilibrium with pure chromium at 600°C and
indicate that it should not be possible to deposit
chromium at 600°C when NaF-ZrF ,-UF, melts of
this general composition are circulated in Inconel
loops with hot- and cold-leg temperatures of 600
and 800°C, respectively. Mass transfer of chromium
would occur in loops of pure chromium under these
conditions, and would occur in an Inconel loop if
the deposited chromium were able to diffuse rapidly
enough into the cold-zone material.

Similarly the data given in Table 2.2.3 may be
used to calculate the equilibrium chromium con-
centration in the NaF-LiF-KF-UF, melt ot 800°C
when Inconel is the container material. Values of
1250 and 1500 ppm were obtained based on Cr**
and Cr***, respectively, and the equilibrium con-
centration is probably intermediate between the
two values. Since these concentrations are greater
than the value of 1100 ppm found for equilibrium
with pure chromium at 600°C, Inconel in contact
with the NaF-LiF-KF-UF, melt will support a
higher concentration of Cr** + Cr*** in equilibrium
at 800°C than pure chromium can support at 600°C.
Thus chromium would be removed from Inconel in
the hot zone and deposited as essentially pure
chromium in the cold zone. Since no diffusion
process is required in the cold zone, the rate of
attack in the hot zone would be controlled by the
rate of diffusion of chromium to the metal-salt
interface. The chromium concentrations that re-
sult when Inconel is contacted with UF , dissolved
in either the NaF-ZrF, or NaF-LiF-KF mixture
are giver: in Table 2.2.4.

It may be noted that the chromium concentrations
given in Table 2.2.4 are, in general, lower than

Table 2,2,4, Chromium Concentrations Resulting from the Reaction of Inconel with UF4 in Molten Fluorides

 

 

 

Conditions of Original Final
Reaction Equilibration Concentrations Chromium
Medium Temperature Time —(LnoleL Concentration
(°C) (hr) UF, UF, (pPm)
Na F--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
those for reaction with pure chromium. If 700 ppm
is taken for the chromium concentration in the
NaF-ZrF ,-UF ; melt at 800°C and 0.1 is taken for
the activity of chromium in Inconel, a valve of
5 x 10~> may be calculated for K.. This value is
about an order of magnitude less than that given
in Table 2.2.2 for the equilibrium quotient ob-
tained for pure chromium. The discrepancy un-
doubtedly results from an apparent equilibrium
state in which the Inconel surface is depleted in
chromium because of the relatively slow diffusion
of chromium to the surface.

Values for the equilibrium quotients for the
reaction of UF , with chromium at 600 and at 800°C
in various reaction mediums are presented in
Table 2.2.5. An examination of the data for the

alkali fluoride—~ZrF, binary mixtures containing
approximately 52 mole % of alkali fluoride shows a

PERIOD ENDING DECEMBER 31, 1957

decrease in K, _ upon changing from lithium to
rubidium. This behavior is attributed to the change
in the cation charge-to-separation ratio, Z/R, and
the resulting decrease in fluoride activity upon
going from LiF to RbF while keeping a constant
mole percentage of alkali fluoride. A similar effect
is noted when the NaF content in the NaF-ZrF
binary is increased from 50 to 59 mole %. The
results are also shown in Fig. 2.2.1, where K_is
plotted against either the cation radius or the
mole percentage of NaF. The change in K_ with
temperature, which is a measure of the heat of
reaction, is probably a good index for the mass
transfer of chromium. Thus, in considering the
effect of the solvent for UF, on the corrosion of
chromium-bearing alloys, the change in K _ must
be considered, as well as the equilibrium chromium
concentrations. In addition to the various fluoride

Table 222.5. Equilibrium Quotients for the Reaction 2UF4(d) + CHe) &= 2UF3(d) + Cer(d) in Molten Fluorides

 

 

Initial Equilibrium
Reaction Temperature U F4 Chromium
Medium (°C) Concentration Concentration Kx
(meole %) (ppm)

LiF-ZrF, 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 1073
NaF-ZrF 600 4.0 1730 + 30 (1.3 +0.1) x 104
(53-47 mole %) 800 4.0 2160 + 20 (2.9 +0.1) x 104
KF-ZrF 600 4.0 1080 + 70 (2.4 0.5 x 10~3
(52-48 mole %) 800 4.0 1160 * 60 (3.2 £ 0.7) x 1073
RbF-ZrF, 600 4.0 69C * 30 (1.0 £ 6.2) x 1077
(52-48 mole %) 800 4.0 760 * 40 (1.5  0.4) x 10~5
NaF-Z(F 600 4.1 2220 * 120 (3.2 £ 0.8) x 10~4
(50-50 moie %) 800 4.1 2550 + 210 (5.9 + 1.8) x 10~4
NaF-ZrF, 600 3.7 1020 * 40 (1.5 £ 0.2) x 10~7
(59-41 mole %) 800 3.7 1050 * 110 (1.7 £ 0.7) x 10~°
LiF-NaF-ZrF 600 2.5 540 + 90 (1.0 £ 0.7) x 10~¢
(55-22-23 mole %) 800 2.5 750 + 60 (3.8 £ 1.1) x 10~6

LiF-BeF, 550 1.5 920 (Z x 10~%)

(48-52 mole %) 650 1.5 1300 + 220 (9 x 10~%

800 1.5 2060 £ 210 (7 x 1079

 

81

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27034

CATION RADIUS (&)

=5
(x$0 ™) 07 0.9 14 1.3 15
1000

500

200

100

 

50
O

20
10
5
2
1

50 52 54 56 58 60 62 64

NaF CONTENT (moie %}
Fig. 2.2.1. Effect of Cation Radius and NaF Content

of the Nch-ZrF4 Reaction Medium on the Equilibrium
Quotient for the Reaction of UF4 with Chromium at
600 and at 800°C. (Secret with caption)

mixtures listed in Table 2.2.5, other solvents,
namely, LiF-BeF,-ThF, (67-23-10 mole %) and
LiF-ThF , (72-28 mole %), have been studied with
UF , concentrations in the range 0.1 to 0.2 mole %.
The extent of reaction with chromium at these
low UF , concentrations is so limited, however,
that no measurable difference was found between
the chromium blanks and the chromium concen-
trations resulting from the reaction of UF, with
chromium,

REACTION OF UF4 WITH YARIOUS STRUCTURAL
METALS IN MOLTEN FLUORIDES

J. D. Redman

The compatibility of UF, dissolved in various
fluoride mixtures with various metals having de-
sirable structural properties, either as the metal
or in alloy form, is of considerabie importance in

82

the study of molten fluoride fuels for reactors.
Many of the available alloys have either a nickel or
an iron base, with chromium added to impart
oxidation resistance and high-temperature strength.
Such alloys include Inconel and the stainless
steels. Nickel is stable toward UF, dissolved in
the fluoride mixtures of interest and consequently
is a svitable base for alloys. Data are presented
here which demonstrate that even though iron re-
acts with UF, the extent of the reaction is suf-
ficiently limited that iron might also serve as a
suitable base for the structural material. Data
given in the previous section, in comparison with
the data presented here, show conclusively that
chromium is more reactive with UF, than is iron
and indicate that the chromium present in Inconel
and the stainless steels is the element most likely
to react with UF, and therefore to play the im-
portant role in the corrosion of these alloys by
UF ,-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 UF, dissolved in either
alkali flueride mixtures or alkali fluoride-ZrF ,
mixtures,

A filtration method was employed in which the
UF ,-bearing fluoride mixture was equilibrated in
nickel equipment with the metal under study and a
sample was removed through a sintered-nickel
filter and analyzed for the desired metal. In the
case of iron in ZrF ,-bearing melts the reaction is
believed to proceed as

2UF, + Fe®== 2UF, + FeF,

and equilibrium quotients may be evaluated from
the expression

The valence states of the oxidation products of the
metals other than iron have not been established,
and consequently only the equilibrium metal con-
centrations are given.

The data presented in Table 2.2.6 show a re-
markable insensitivity of iron concentration to
changes in the composition of the reaction medium.
It may be noted that the reaction exhibits a negative
Table Z22.6. Data for the Reduction of UF, by Iron in
Molten Fluorides at 600 and at 800°C

 

 

Reaction Temperature  Fe K
Medium (°C) (ppm) x
LiF-ZF 600 500 1.4 %1076
(52-48 mole %) 800 370 5% 1077
NaF-ZrF, 600 490  2x1078
(53-47 mole %) 80O 380 6 x10~7
KF-Z/F, 600 540 3 x1076
(52-48 mole %) 800 320 6 x 107
RbF-Z/F, 600 490 4 x107°
(52-48 mole %) 800 140 5x1078
NaF-ZrF, 600 450  1x107¢
(50-50 mole %) 800 310 3x10~7
NoF-ZrF , 600 530 1.5x1076
(59-41 mole %) 800 450 8 x10~7
LiF-NeF-ZrF 600 420 6x10"7
(55-22-23 mole %) 800 625 2 %10~
NaF-LiF-KF 600 690
(11.5-46.5-42 mole %) 800 480

 

temperature coefficient in all mediums except the
LiF-NaF-ZrF, mixture. The results given for the
LiF-NaF-KF mixture, a composition quite different
from the alkali fluoride~ZrF, mixtures, indicate a
negative temperature coefficient and equilibrium iron
concentrations very similar to those found in the
ZrF j-bearing melts. In contrast to these results
for iron, the behavior of chromium, as described
above, follows the expected pattern in the alkali
fluoride~ZrF , melts.

An examination of the data presented for various
metals in Tables 2.2.7 through 2.2.11 reveals some
interesting differences in the behavior of the in-
dividual metals and the effects of different re-
action mediums. The only metal that shows any
real stability in an NaF-LiF-KF-UF, melt is
molybdenum, which undergoes rather limited attack.
The attack on molybdenum by an NaF-ZrF ,-UF

melt is less than that of the alkali flueride mixture,

PERIOD ENDING DECEMBER 31, 1957

as is, in general, the case for most of the metals
studied. Niobium is not stable in the NaF-LiF-KF-
UF, meit, but it is unattacked by either the NaF-
ZrF ~UF , or LiF-NaF-ZrF ,-UF, melt. Tantalum,
which would be expected to behave similarly to
niobium, is not stable in either type of melt.
Vanadium, likewise, is not stable in any of the
melts used, whereas tungsten is stable in the
NaF-ZrF ,-UF , melt but not in the NaF-LiF-KF-UF,
melt. |In general, the extent of attack on the metals
increases with a temperature increase from 600 to
800°C.

It is evident that only molybdenum exhibits any
stability toward UF, in the NaF-LiF-KF mixture,
and accordingly nickel-molybdenum alloys would be
expected to resist corrosion by this type of melt.
In the ZrF,-based melts, molybdenum, niobium,
and tungsten appear to be stable toward UF,, and
thus these metals or their alloys might serve as
suitable container materials for this type of melt.
Tantalum and vanadium appear to offer no advantage
over chromium for either type of melt.

REDUCTION OF Ft=.'F2 BY CHROMIUM
IN MOLTEN FLUORIDES

J. D. Redman

A study has been made of the reaction
FeF, + Cr°=Fe® + CrF,
in an attempt to evaluate the equilibrium quotient

NCrF2
Kx = —
NFe F2

at  600°C in the following reaction mediums:
Nch-ZrI"'4 (53-47 mole %), KF-ZrF , (52-48 mole %),
and RbF-ZrF, (52-48 mole %). Although it was
known that the equilibrium quotient for this reaction
was much greater than 1, it was believed that the
equilibrium concentrations of Cr** and Fe** might
be determined sufficiently accurately by the fil-
tration method to permit an evaluation of the ratio.

If the system behaves ideally, a value of approxi-
mately 6 x 103 for K, is obtained from the standard
free energies of formation for CrF, and FeF, at
600°C, and thus essentially complete reduction of
Fe** by the chromium is indicated. A calculation
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 Found in Filtrate

 

Equilibration

 

 

Reaction Total Total Total
Medium Temperature Time U Mo* Ni

(°C) (hr) (wt %) (ppm) (ppm)

NaF-Zr F4 600 3 8.4 7 155

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

S 8.6 11 85

5 8.6 9 80

5 B.6 9 10

NaF-LiF-KF 600 3 13.3 210 85

(]]-5‘46.5‘42 mole %) 5 ]4.6 200 1]0

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

M° + 2UF, =2 2UF, + MF,

where M is iron or chromium, in the NaF-ZrF ; mix-
ture at 600°C yielded a value of ~60, however,
which suggests that the behavior is far from ideal.
The corresponding values in the KF-ZrF, and
RbF-ZrF , mixtures are ~10 and ~3, respectively.
These values for the Cr**to-Fe** ratios, if
correct, should be determinable by the filtration
method.

The data given in Table 2.2.12 were obtained
from filtration runs in nickel equipment. In most

cases the equilibration time was 5 hr. Several runs,

with an equilibration period of 12 hr gave similar

values and demonstrated that equilibrium was

84

attained within 5 hr, A comparison of the experi-
mentally determined chromium values with the
theoretical values shows that, in the case of the
NaF-ZrF, and KF-ZrF, mixtures, the reaction
occurs as indicated and yields a good balance
between the Fe'' added and the Cr*t found,
except at the higher concentrations. The abnormal
behavior at the higher concentrations will be dis-
cussed below. In the data for the runs with RbF-
ZrF , as the reaction medium, it may be noted that
the calculated chromium values are about 10 to
20% lower than the experimentally determined
values, in most cases. No experimental error could
be found to explain this discrepancy, and, although
there is no evidence that the valence of the iron
and chromium is not 2, some abnormal behavior
must occur,

 
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 Found in Filtrate
’ Reaction Equilibration Total Total Total
Medium Temperature Time u Nb* Ni
° (°C) (hr) (wt %) (ppm) {(ppm)
’ NaF-ZrF 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-ZrF, 600 3 8.2 1 100
(55-22-23 mole %) 3 8.8 i 120
5 8.4 2 120
. 5 8.9 1 45
. 800 3 8.5 1 110
- 3 8.6 1 110
- 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 1.3 225 65
5 11.1 655 1
5 11.2 640 6
800 3 10.8 1090 35
3 10.5 1200 15
5 1.1 1540 6
5 11.2 1580 40
12 11.0 1830
) 12 10.9 1470
. 12 11.0 1930
- 12 1.1 1310
12 1.4 1230
. 12 12.0 1360

 

*Blanks present at 800°C for the various mediums were: for NuF-ZrFA, 15 ppm; for LiF-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

 

Conditions of

Equilibration

Found in Filtrate

 

 

 

Reaction Total Total Total
Medium Temperature Time u Ta* Ni

(°C) (hr) (wt %) (ppm) (ppm)

NaF-ZrF, 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-L.iF-KF 600 3 11.2 1320 140

5 12.1 1150 110

5 11.8 1150 240

800 3 11.2 3220 170

11.9 2800 80

5 11.8 3020 180

 

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

The reduction of Fe*t was essentially complete
in all cases, except at the higher FeF, additions
in the NaF-ZrF, mixture. In fact the Fe**? con-
centrations are so low that the ratio of Cr*t to
Fe*" cannot be evaluated in any of the reaction
mediums. A value of ~60 was calculated, as
stated above, for the ratio in the NaF-ZrF , mixture,
and the experimental data do not deny that it may
be this large. However, for the KF-ZrF, and
RbF-ZrF, mixtures, it is evident that the experi-
mentally determined ratios, while not constant, are
considerably greater than the calculated ratios of
~10 and ~3. This lack of agreement must be
attributed to some uncertainty in the K, valuves
found for the iron and chromium reactions with
UF,.

The limited solubility of CrF, in the NaF-ZrF,
mixture precludes the addition of large amounts of
FeF,, and thus the concentration range is rather

86

limited. It will be noted that for runs made at
600°C with either 1 or 2% additions of Fe*™ the
results are not precise and very incomplete re-
duction of Fe** occurred. This behavior probably
results from the presence of solid phases con-
taining NaF, ZrF ,, CrF,, and/or FeF,. The solid
phase contains more ZrF, than NaF, and conse-
quently the liquid phase is enriched with respect to
NaF. The lack of precision indicates that equi-
librium was not attained in these runs. |t is of
interest to compare these runs with similar runs
made at 700°C in which the solubility of the
CrF ,-bearing compound was not exceeded. Here
essentially complete reduction of Fe** occurred,
and a stoichiometric relationship prevailed between
the Fe** added and the Cr** found. It is apparent
that the method employed is incapable of establishing
the Cr**to-Fe** ratio in these reaction mediums at

600°C.

 
 

PERIOD ENDING DECEMBER 31, 1957

Table 2.2, 10. Data for the Reaction of UF4 with Yanadium in Molten Fluorides at 600 and at 800°C

 

Conditions of

Equilibration

Found in Filtrate

 

 

 

Reaction Total Total Total
Medium Temperature Time u vV* Ni

(°C) (hr) (wt %) (ppm) (ppm)

Nch-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 in the NaF-ZrF4 mixture; 150 ppm present in the NaF-LiF-KF mixture at 800°C.

THERMODYNAMIC INTERPRETATION OF
EQUILIBRIUM DATA ON THE CORROSION
OF IRON AND OF CHROMIUM BY
MOLTEN-SALT FUELS

M. Blander

Considerable attention has been given to the
corrosion and metal deposition caused by the tem-
perature coefficient of the equilibrium”

(n 2UF, + M=2UF,; + MF,

when molten-salt mixtures are used as the reaction

mediums. The equilibrium constant for the reac-
tion is
a a
MF2 UF3
Ka = - ’
a a
u I:4 M

 

Yy, R, Grimes, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 96.

where a is the activity of the designated material,
the salts are dissolved in the fuel melt, and the
metal is in the container wall. Some of the dif-
ferences in the behavior of structural metals are
very important from the standpoint of compatibility
between fuel ond container, The effect of tem-
perafure,zo for a steady-state concentration of
UF,, UF;, and MF, in the melt, can be represented
as

d In XM
(2) RT?2 = —(AH®° + AH )
dT s
dIn K
= —RT? *
dT !

 

20This discussion covers only the tendency toward
mass transfer. The rate is gaverned by this tendency,
or driving force, and by a resistance to this force which
is a function mainly of the diffusion coefficient of the
metol in question through the container alloy being con-
sidered.

87

 
ANP PROJECT PROGRESS REPORT

Table 222,11, Data for the Reaction of UF, with Tungsten in Molten Fluorides at 630 and ot 800°C

 

 

 

 

Conditions of Found in Filtrate
Reaction Equilibration Total Total Total -
Medium Temperature Time U W Ni
(°C) (hr) (wt %) (ppm) {ppm) -
NaF-LiF-KF 600 3 11.2 950 215 *
(11.5-46.5-42 mole %) 3 n.2 920 435
5 1.6 1410 250
5 11.2 1350 230
5 10.5 900
5 11.0 1120
800 3 11.3 980 260
3 1.2 920 205
5 11.2 1480 65
5 10.8 2150 250
5 11.8 1260 165
5 1.6 1210
5 1.8 1600 ’
12 10.5 1880 115 v
12 11.8 1400 -
12 11.3 1430 ’
NaF-ZrF 600 3 9.2 70 110
(53-47 mole %) 9.0 95 45
5 9.1 70 30
800 3 9.2 130 40
3 9.3 100 50
5 8.5 95 50
5 8.9 100 20

 

*Blank of 350 ppm present at 800°C in NaF-LiF-KF; <5 ppm present in NaF-Zer.

where X,, is the equilibrium concentration of N2 N

metal M in the wall, AH® is the heat of change of UF3 'MF, .
the reactants in the equilibrium reaction (Eq. 1) Ky = - .
from their standard state (chosen here as the pure NUF4 Ny

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 AH® + AH _ is positive the metal M

librium quotient: will tend to transfer from the hot to the cold end of

88

 

 
PERIOD ENDING DECEMBER 31, 1957

Table 2.2,12. Data for the Reaction FeF2 + Crofi Fe® + CrF2 in Molten Fluorides

 

 

 

Present in Filtrate Calculated
Reaction Temperature FeF2 Added - Cr*
Medium (°c) (ppm of Fe) Ni Fe Cr _—

(ppm) (ppm) (ppm) (ppm)
NuF-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 53 1,010 1,010

970 90 50 980 1,060

885 110 55 880 980

2,000 110 50 2,000 2,020

2,100 110 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 15 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 Filtrate Calculated
Reaction Temperature FeF, Added - Cr*
Medium (°C) (ppm of Fe) Ni Fe cr

(ppm) (ppm) (ppm) (ppm)
KF-ZrF, 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 30 150 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-ZrF 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
the system and, if negative, from the cold end to
the hot end. The most desirable condition would
be for AHI® + AH _ to be zero so that there would
be no mass transfer by the proposed mechanism,

If the pure liquids involved in the reaction are
chosen as the standard state, the values of AH®
tabulated below can be obtained by using the heat
content values listed in Table 2.2.13:

 

M AH®
Cr +9
Fe +22
Ni +33

Since all these values of AH® are positive, there
would be a tendency for mass transfer to take place
from the hot to the cold end of the system if the
AH° term is not compensated for by the term AH _,
which involves all the heats of solution.

The term AH _is given by

——

(3) AH = 2(HUF3 - HSFS) + (HMF2 - HfiFz)"
- 2(HUF4 - H8F4) —(Hy = HQ) .

where H is the partial molal heat content of the
material in solution and H® is the molal heat con-
tent of the pure substance. For a pure metal the
last term in Eq. 3 is zero. The other terms in this
expression are unknown, but if data are available

Table 2.2.13. Heat Content Data* for Reactants of
the Reaction ZUF4 + M ;ZUF3 + MF2

 

Heat of Formation, Heat of Fusion,

 

Compound
AH,gq (keal) AH[ (keal)
UF, —443 5.7
UF, -357 8.5
CrF, -181 (12y*~
NiF, -159.5 (12)**
FeF, -168 (12)**

PERIOD ENDING DECEMBER 31, 1957

they can be obtained from the equation

@ [dlny =+H—H°’
d(1/T)| R

where y is the activity coefficient based on pure
liquid as the standard state and the subscript N
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
given melt composition that is relatively dilute in
CrF, or FeF,. Since the values for the heats of
solution of UF, are the same for a melt dilute in
CrF, as for a melt dilute in FeF,, and similarly
for UF ,, it follows from Eq. 2 that

din(N~- /N_)
2 Cr Fe o
(5) RT — = _AHCrF,‘, +

+ AHS

o
FeF _(H_H)

+
2 CrF2

RT? 4 In [K_(Fe)/K_(Cr)]

 

dT

The values of K _ for chromium and for iron in
NaF-ZrF, (53-47 mole %) at 600 and at 800°C
(ref 21) are given in Table 2,2,14,

From the similarity of the compounds CrF, and
FeF, it would be expected that their solution be-
haviors would not differ greatly and that any dif-
ference would be governed mainly by the difference
of their AHH° values. As may be seen in Table
2.2.14, these suppositions are not valid,
be concluded either

It must
that some of the data are

 

21, p, Redman, ANP Quar. Prog. Rep. june 30, 1957,
ORNL-2340, p 135.

Table 2-2.]4.
Derived from Temperature Coefficients for Chromium

and lron in Na F-ZrF4 (53-47 Mole %)

Equilibrium Quotients and Heat Changes

 

 

 

*From Selected Values of Chemical Thermodynamic
Properties, NBS Circular 500.
**Estimated by using a method discussed by Blander

and Blankenship in a previous report, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 121.

Chromium lron
K
x -4 -6
At 600°C 1.3 x 10 2 x 10
At 800°C 2.9 x 1074 6 x 1077
AH® + AHS 7.5 keal -11.3 kcal
AH°® 9 kecal 22 kcal

 

91

 
ANP PROJECT PROGRESS REPORT

incorrect or that CrF, and FeF, have entirely
different solution behavior,

Speculations as to the source of the discrepancy
may be made on the basis of available data and
some reasonable estimates for the heats of solu-
tion. From data on the activity coefficients of
FeF, (ref 22) based on the solid as the standard
state and the estimate of the heat of fusion of
FeF, given in Table 2.2.13, the heat of solution
of FeF, was found to be -4 kcal/mole. Since the
heat of solution of UF, is probably similar to that
of ZrF, (ref 23) in the fuel melt, it is assumed to
be between 2.0 and 4.3 kcal/mole. Similarly, it is
likely that the heat of solution of UF, in the fuel
melt lies in the range +5 kcal. [f these heats of
solution are combined with the values of AH given
in Table 2.2.13, a range of values for (AH® + AH )
of +15 to +27 kecal/mole is obtained for the FeF,
equilibrium; the apparent value of =11 kcal/mole
given in Table 2.2,14 seems to be improbable, It
may be considered therefore that the measured tem-
perature coefficient of the iron equilibrium reaction
appears to be unreasonable. The probable range
for (AH® + AHS) for the CrF, equilibrium is +2
to +14 kcal/mole. Within the broad limits of the
method of estimation used here the temperature
coefficient of K _ for the chromium equilibrium is
not unreasonable.

SOLUBILITIES OF NOBLE GASES IN MOLTEN
FLUORIDE MIXTURES

N. V. Smith

The solubilities of noble gases in molten salts
have been investigated because such gases are
produced in fused-salt-fueled reactors. Xenon gas
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
salt mixture used as the fuel. A knowledge of the
temperature dependence of the solubility was needed
in order to predict in what portion of the reactor
system the gas would have the greatest tendency
to come out of solution,

Helium and argon gases are used as protective
blankets and for sparging during reactor operations,
and therefore their solubilities in fuel mixtures

 

22 M. Blood and G. M. Watson, ANP Quar. Prog. Rep.
March 10, 1956, ORNL-2061, p 84.

238. Cantor, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 111.

92

were also studied. It was conceivable that a por-
tion of the reactor fuel could become saturated
with these gases in one section of a reactor and
supersaturated in another and that bubble formation
could result. This condition could lead to un-
desirable perturbations in reactivity., The solu-
bilities of helium and of argon were determined as
functions of pressure and temperature.

Because of the chemical simplicity of the noble
gases, the solubility relationships of these gases
in molten salts may be correlated with some funda-
mental aspects of liquid structure and chemical
composition. The variations of the solubilities,
the heats and entropies of solution, and the other
properties with the atomic size of the gas may
perhaps be sufficiently meaningful to reveal some
of the characteristics of the potential-energy func-
tions that govern the behavior of the ionic and
molecular species in the liquid.

A summary of the results obtained thus far
in the determination of the solubilities of various
noble gases in a number of different solvents is
presented in Table 2.2.15, In all the cases in-
vestigated, Henry's law was found to apply. In
addition,
with increasing temperature. For the same solvent,
the solubilities of the gases decreased with in-
creasing solute gas size. Variations in solvents
showed corresponding gas solubilities tobe approx-
imately the same in the two related ZrF ;-bearing
solvents, but the solubility of helium in NaF-KF-
LiF was found to be less than half that in the
ZrF ,-bearing solvents,

The heats and entropies of solution of the four
gases studied in the NaF-ZrF, (53-47 mole %)

mixture are summarized below:

24-26

the solubilities were found to increase

AH As

Gos (kcal/mole) (eu)
Helium 6.2 -1.3
Neon 7.8 0.4
Argon 8.2 -1.6
Xenon 11.1 -0.7

 

24N. V. Smith, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 140.

254, . Smith, ANP Quar. Prog. Rep. March 31, 1957,
ORNL-2274, p 111

26N. V. Smith, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 130.

 

 
 

PERIOD ENDING DECEMBER 31, 1957

Table 2,2.15. Henry's Law Constants for Various Noble Gases in Molten Fluoride Solvents

 

 

 

K*
Solvent Gas
At 600°C At 700°C At 800°C
x 1078 x 1078 x 10~8
Na F-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-Zr F4-U F4 {50-46-4 mole %) Helium 20 27 41
Xenon 2.0 4.0 6.5
Helium 11.3

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

 

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

It may be observed that the heat of solution in-
creases with increasing atomic size of the gas.
Also it is worth noting that the entropy of solution
corrected to equal concentrations is negative and
in all cases quite small. Because of the uncer-
tainties of the measurements, the heats of solution
are considered to be reliable only to 11 kcal/mole
and the entropy changes are reliable only to 11 eu.
Accordingly it appears that the trend in the heats
of solution is real but that no particular significance
should be attached to the different values of the
entropy change, since as far as the experimental
measurements are concerned the values listed for
the entropy change are not significantly different,

SOLUBILITY OF HF IN NaF-ZrF,
J. H. Shaffer
The solubility of HF in mixtures in the NaF-ZrF

system has been successfully determined as a
function of temperature, pressure, and melt compo-

A summary of the recent experimental
7,28 .
is

sition,
results and those obtained previously?
presented in Table 2,2.16.

It may be noted that the solubility of HF, in all

cases, decreases with increasing temperature.

 

27 §. H. Shatfer, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 141.

28, 4, Shaffer, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 136.

Furthermore, as shown in Fig. 2.2.2, the Henry’s
law constants for the HF solubility illustrate a
constant heat of solution over the experimental
temperature range for each melt composition. The

UNCLASSIFIED

-5
(x10 7 ORNL-LR-DWG 27035

70
60

5.0

40

30

2.0

1.0
09
0.8
0.7

0.6

K [{motes OF HF)/(cm? OF melt)/atm ]

0.5 MELT COMPOSITION
0.4 O 19.5 mole % er"4
® 35 mole % er'—'4

A 40 mole % Zrf,

A 47 mole % Zrfy

0 55 mole % ZrF,

03

 

0.2
10.0 H.0 12.0 13.0

10%/T (°K)

7.0 8.0 2.0

Fig. 2.2.2, Temperature Dependence of Henry's Law
Constants for the Solubility of HF in NaF-ZrF4.
ANP PROJECT PROGRESS REPORT

Table 2.2.16, Solubility of HF in NuF-ZrF4 Mixtures as a Function of Temperature, Pressure, and Melt Composition

 

 

Melt Saturating Solubility ,
N Temperature HE P (moles of HE/ 3 Henry's Law
Composition °C) ressure e c Constant, K*
(mole % ZrF4) (atm) of melt)
x 103 x 1073
19.5 800 2,04 9.12 4,47
2.12 9.26 4,38
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 t+ 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 1 0.02
700 0.93 + 0.02
750 0.86 1 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 1 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 = ¢/p in moles of HF per cubic centimeter of melt per atmosphere.

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

94
values for the heats of solution at each composition
have been calculated from Fig. 2.2.2 and are shown
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
undergoes an inflection point at approximately
44 mole % ZrF,, and it increases with increasing
NaF concentration,

Attempts to correlate the observed changes of HF
solubility with corresponding changes in solvent
composition will be withheld until considerably
more information is available. Experiments to
measure the solubility of HF in NaF-KF-LiF (11.5
42-46.5 mole %) and in LiF-—BeF2 and Nc:F-BeF2

mixtures are now under way.

UNCLASSIFIED
ORNL—LR--DWG 27036

 

10,000

3000 \

-
e B

 

 

 

8000

5 7000

o

E \
~

3 .
I \
<]

| 6000

 

5000 \
4000 \

 

 

 

 

 

 

 

3000

 

20 40 60 80
Z¢F, CONTENT (mole %)

NaF

Fig. 2.2.3. Composition Dependence of the Heat of
Solution of HF in NuF-ZrF4.

PERIOD ENDING DECEMBER 31, 1957

 

 

 

 

-5 UNCLASSIFIED
(x10 7) ORNL-LR -DWG 27037
45
o

0 ¢ eoo°C

o 700°C

4 800°C
35

[
o
[—

 

n
w

 

N
o
o

 

# [{moles OF HF)/(cm® OF meit)/atm]

 

 

 

 

 

 

 

 

 

 

1.5 .
N
0.5 \"
10 20 30 40 50 60

ZrFy CONTENT {mole %)

Fig. 2.2.4. Composition Dependence of Henry’'s Law
Constants for the Solubility of HF in NuF-ZrF4.

SOLUBILITIES OF FISSION-PRODUCT
FLUORIDES IN MOLTEN NaF-ZrF , MIXTURES

w. T. Ward R. A, Strehlow

Experiments have been completed in order to
determine (1) the effect that varying the ratio of
sodium to zirconium in the NaF-ZrF, solvent has
on the solubility of CeF,, (2) the solubility of YF,
in NaF-ZrF , (53-47 mole %), and (3) the solubilities
of CeF3 and SmF, when both are present in NaF-
ZrF , (53-47 mole %). The solubility measurements
were made by using the radiochemical tracer tech-
niques described previously,2%+30

 

29, T. Ward, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 125.

0y, 7. Ward, ANP Quar, Prog. Rep. March 31, 1957,
ORNL-2274, p 105.

95

 
ANP PROJECT PROGRESS REPORT

CeF3 in Nc:F-ZrF4

It has been found that varying the ratio of sodium
to zirconium in the solvent has a considerable
effect on the solubility of CeF,. This effect is
clearly shown in Fig. 2.2.5, in which the solubility
vs temperature curves for five different NaF-ZrF ,
compositions are plotted. In addition to the compo-
sitions indicated on Fig. 2.2.5, the NaF-ZrF,
(80.5-19.5 mole %) solvent was briefly investigated.
It was found that all the 4.60 mole % CeF , that was
added was soluble at temperatures from 793 to

915°C.

UNCLASSIFIED
ORNL-LR-DWG 27038

 

5
NaF-ZrF, SOLVENT
= COMPOSITION
a~
% 42-58 mole %
E
o2 50 - 50 mole %
q
E 53-47 mole %
w
=z
P
@
3
05 59 - 4¢ mole %
63-37 mele %
0.2
10%/7 (°K)
Fig. 2.2.5. Solubility vs Temperature for CeF, in

Various NuF-ZrF4 Mixtures.

One interesting observation is that in the solvent
NaF-ZrF, (42-58 mole %) the nature of the pre-
cipitating phase changes at about 640°C, the in-
flection point on the curve. A plot of CeF, solu-
bility at 600°C as a function of composition is
presented in Fig. 2.2.6, and it may be noted that
the curve is quite similar to that obtained earlier3’
for the solubility of UF, in NaF-ZrF,. Such

correlation would be expected from a consideration

 

316, M. Watson and C. M. Blood, ANP Quar. Prog,
Rep. June 10, 1954, ORNL-1729, p 51.

96

UNCLASSIFIED
ORNL-LR-DWG 27039

 

 

5 &

4 /

 

 

 

CeFy SOLUBILITY {mcle %)
w

 

/

 

 

 

 

 

30 40 50 60
ZrF, CONTENT (mole %)

Fig. 2.2.6. Solubility of CeF, at 600°C in NaoF-ZrF,

Mixtures as a Function of Composition.

of the similarity of the UT™" 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
41 mole % ZrF,, since with higher NaF content
the precipitating phase is not CeF, but rather
NaCeF ,, as determined by petrographic observation.

YF, in NaF-Z+F,
The solubility of YF in NaF-ZrF,, (53-47 mole %)

was determined, and the results are presented in

Table 2.2,17 and Fig. 2.2.7. Although some

Table 2.2.17. YF:3 Solubility in NaF-ZrF
(53-47 Mole %)

 

 

 

Filtration Solubility
Temperature
°C) Inwt%Y In mole % YF3
854 11.37 13.48
817 8.68 10.26
756 6.31 7.39
663 4.10 4.75
560 2.37 2,72

 

 

 
scattered earlier results had indicated that the
plot of the log of the solubility vs the reciprocal
of the temperature for YF; was linear, these results
show a behavior pattern more similar to the patterns
obtained for the rare earths,

Cer plus SmF in NaF-Z(F

The solubilities of CeF, and SmF, in combina-
tion in Nc:F-ZrF4 (53-47 mole %) were measured by

UNCLASSIFIED
ORNL-LR-DWG 27040

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20
15 :
°
\

10 e,

8
3 . Ve,
s N
S 6 N
£ AN
w 5 N
= .
<
£ ~
w
z
u” 3 ~
> N

2

1.5

1

8 9 10 A A2 13
10%/7 (°K)
Fig. 2.2.7. Solubility vs Temperature for YF, in

NuF-ZrF4 (53-47 Mole %).

PERIOD ENDING DECEMBER 31, 1957

using the two-tracer technique described pre-
viously.30 Filtrates were obtained over the tem-
perature range of 550 to 800°C for two cerium-to-
samarium ratios. A summary of the solubility data
is presented in Table 2,2.18, along with the solu-
bility of CeF3 by itself in the same solvent, The
solubility vs temperature relationships are illus-
trated in Fig. 2.2.8, where the logs of the sums of
the solubilities are plotted vs the reciprocal of the
temperature.

The solubility of pure SmF; has not yet been
measured in this solvent, but, by utilizing the
correlation of solubility with the cube of the ionic
radius, estimates were made of the solubilities of
SmF, at these temperatures, The values obtained
are shown in Table 2.2.19,

These values are summarized in the pseudo-
ternary diagram presented as Fig. 2.2.9. The rare-
earth fluoride solubility behavior in this solvent
is seen to be substantially the same as the be-
havior reported previously when the solvent was

NaF-ZrF ,-UF , (50-46-4 mole %).%?

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
from NaF-LiF-KF base mixtures. Three experi-
ments have been performed to obtain solubility
values in which the starting barium compound,

either BaF, or Ba0, was labeled with radicactive
Ba'33,

 

323, A, Strehlow and W. T. Ward, ANP Quar. Prog.
Rep. June 30, 1957, ORNL-2340, p 139.

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

 

Total Additive

 

4,48 Mole % SmF3,

4.76 Mole % SmFa, 6.1 Mole % CeFa,

 

 

 

 

Filtration 1.90 Mole % CeF, 7.14 Mole % CeF, no SmF
Temperature
(°C) Solubilities (mole %) Solubilities (mole %) Solubility of CeF,
SmF,  CeF,  SmFy + CeF, SmF,  CeFy  SmF;+CeF, (mole %)
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
DORNL—LR--DWG 270414

 

 

 

 

 

 

' |
TOTAL RARE-EARTHS PRESENT IN
6 SYSTEM (mole %)
\ SmFy  CeFy |
® NONE 6.1
a 4.7¢6 7.14
5 \ A 4.48 1.90
\ A
N\

 

 

 

 

 

 

TOTAL RARE-EARTHS IN FILTRATE (mole %)
72

 

 

 

 

 

 

 

 

O
2 \“‘
N
..
\h.
1
8 9 10 1" 12 13

109/ 7 (K)

Fig. 2.2.8. Sclubilities of SmF3 and CeF, in NaF-
ZrF4 {53-47 Mole %) as a Function of Temperatute.

Table 2.2.19. Estimates of the Solubility of SmF3 in
Na F-ZrF4 {53-47 Mole %) Based on the
Solubility of CeF3

 

Estimated Sm F3

 

Temperature Solubility of CeF3 Solubil
iiit
(O(:) (H10le 9%) olubliity
(mole %)

800 3.78 4.40

700 2,57 2.99

600 1.90 2.21

550 1.67 1.95

UNCLASSIFIED
ORNL-LR-DWG 27042

SOLVENT

 

 

 

 

 

Fig. 2.2.9. Pseudoternary Diagram of the System
CeF3-SmF3 with the Mixture NuF-ZrF4 (53-47 Mole %)
as the Solvent,

The melting point and heat of fusion of BaO are
listed®3 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
from BaF,. The soluble impurities would cause
apparent solubility values of BaO to be erroneously
high. This preparation was attempted in two dif-
ferent experiments, The first consisted in the
addition to the melt of LiOH containing a known
amount of labeled BaF,. A filtrate sample was
obtained prior to the addition of the LiOH, and
filtrates at 700 and 600°C were obtained after the
addition was completed. Since no change was
noted in the concentration of barium in solution,
it was decided to repeat the experiment with sodium
peroxide as the precipitating agent, The results
of the second experiment also showed no change
in the concentration of barium in the liquid after
addition of 20% and of 140% excess Na,0
lated as Na,0.

9 calcu-

 

 

98

334, Glassner, A Survey of the Free Energies of
Formation of the Fluorides, Chlorides, and Oxides of
the Elements to 2500°K, ANL-5107, p 6 (Aug. 1953).

 

 
In view of the unexpected results obtained in
the first two experiments, a third experiment was
carried out with labeled BaO as the additive. The
labeled BaO was prepared by ignition of Ba(NO,},.
Again the results showed that the solubility of
BaO was quite high., The labeled BaO used was
examined with the petrographic microscope and by
The results of these examina-
tions show that the material used was high-quality
BaO. The numerical results of the three experi-
ments are summarized in Table 2.2.20,

x-ray diffraction.

REACTION OF CeF, AND UF, WITH
CaO IN KF-LiF

J. H. Shaffer

Studies of the precipitation of cerium as Ce 0O,
from the molten KF-LiF eutectic®4 and of the solu-
bility of barium oxide in this solvent (see preceding
section) have provided a basis for-investigating
the possible separation of tetravalent metals as
oxides from the ftrivalent fission products of a
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
uranium oxide and leave the fission products in
solution,

In the present study, 5 wt % of labeled Ce was
added as CeF, to the molten KF-LiF eutectic

 

34; H. Shaffer, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL.-2387, p 139.

PERIOD ENDING DECEMBER 31, 1957

containing 1 mole % U as UF, at 600°C. Calcium
oxide was then added in tared increments, and
filtrate samples were taken approximately 2 hr
after each addition. These samples were counted
for radioactivity to determine the cerium concen-
tration and were chemically analyzed for uranium,
The results are shown graphically in Fig. 2.2.10.

It may be noted that the cerium concentration
increases, initially, because of the removal of the
uranium from solution, until the equivalence point
for the uranium is reached. Upon further addition

UNCLASSIFIED
ORNL-LR-DWG 27043

 

 

 

 

 

 

 

 

 

 

 

 

 

5
,_-—-0'-—-‘ |
|
|
|
4 5.
\ o :
2 | 0 CeFy
5:-9 \ EI . UF4
Z 3 \ g i
',:‘_" . 2| \
q G|
5 \ 21 \
a g
o ol
w 2 1
x w
o x|
L
=
\ ' \
4 +
\
0.0
I 32 0.0085 T
0 .-!".--’ [
0 [oR] 0.2 0.3 0.4 0.5 06

EQUIVALENTS OF CaO ADDED

Fig. 2.2,10. Precipitation of CeFa and UF4 with Ca0
in KF-LiF (50-50 Mole %).

Table 2.2.20, Concentration of Barium in Liquid KF-LiF After Addition of Ban-LiOH, Ba F2-N0202, and Ba0O

 

Initial Barium Concentration

Temperature

Final Barium Concentration

 

 

 

Calculated Experimental Calcuiated Experimental

Ba F2-LiOH 600 11.4 9.85 9.93 10.45

700 9.93 10.95
BOFZ'N°202 (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 Ca0 the soluble cerium concentration decreases
linearly to zero. The uranium concentration de-
creases to a value of approximately 0.1 wt % at the
equivalence point, This experiment is now being
repeated with the exception that the initial CeF,
concentration has been decreased to 0.5 wt %.
Another experiment is proposed in which ZrF , with
a labeled hafnium tracer is to be substituted for

the UF4.

DETERMINATION OF CHROMIUM DIFFUSION
COEFFICIENTS IN INCONEL-MOLTEN
SALT SYSTEMS

R. B. Evans

The results of experiments carried out over the
past few years on the migration of metal as a re-
sult of temperature gradients in a molten-salt en-
vironment have strongly suggested that the over-
all migration rate is controlled by the diffusion
rates within the metal. Accordingly, a thorough
knowledge of solid-state diffusion phenomena is
considered to be an important phase of corrosion
studies, Several investigations of these phenomena
have been made, but only a small portion of the
work covered cases in which molten salts were
involved. Therefore an investigation was initiated
of the over-all mass transfer rate within isothermal
systems in which Inconel is the container material
and NaF-ZrF , is the salt. The diffusing metal in
Inconel in this system is chromium.

Diffusion Coefficients

The diffusion coefficient is a flow-resistance
parameter used in diffusion rate—time relationships.

It is defined by the linear flow equation

I oM ac

1 - . = vV
(M A ot Ox

which, if steady-state conditions are established
and D is assumed to be independent of concentra-
tion, may be written as

1 AM AC
(2) ——=D—
A At L

where
AM/At = constant rate of diffusion, g/sec,

A = area through which diffusion takes
place, cm?,

length, c¢m, of system in x direction
(L is zero at the surface and increases
with penetration of the metal wall),

100

AC = concentration change of diffusing ma-
terial across L, C__, ~ C__ , g/em’,

D = diffusion coefficient, cm?/sec.

The diffusion coefficient is a function of the
temperature, the diffusing material, and the ma-
terial through which diffusion takes place. [t does
not depend on the geometry of the flow system,
it is entirely independent of the molten salt in
contact with the surface of the metal, although
the salt has a mgrked effect on AC and on the rate
equations. Therefore it is not necessary to employ
molten salts in experiments designed to determine
diffusion coefficients, even though the ultimate
application involves molten-salt systems.>”>

The steady-state equation (Eq. 2) is useful for
discussion and forms the basis for the determina-
tion of flow constants in analogous systems (flow
of heat, flow of electricity, and fluid flow in porous
mediums); however, its use in solid-state diffusion
studies has been discouraged because of the ex-
tremely low values of AM/At and D involved. The
only alternative is utilization of experiments and
equations based on unsteady-state flow. The basic
tinear flow equation for this case is

dC d dC
(3) — =—(D .
ot Ox Ox
If self-diffusion measurements are to be made, for

example, diffusion of radioactive silver into pure
silver,3 and/or if the values of 9C/dx involved

 

 

are not high, Eq. 3 may be written as37
ocC d2C

(4) _— = D .
ot Ox 2

Equation 4 is known as Fick’s diffusion law.

Measurements Based on Radioactive Tracers

Tracer-Layer Method, — One method for experi-
mentally determining diffusion coefficients for a
given metal diffusing through a bar composed of
the same or a different metal is based on the use
of radicactive isotopes of the given metal. A very
thin layer of solid labeled metal is placed, by any
of various methods, on the face of a solid specimen

 

35“‘ will be shown, however, that utilization of molten
salts leads to a unique and simple method of indirectly
determining diffusion coefficients for metals in which
the salt rather than the metal is analyzed.

36, C. Fischer, J. Appl. Phy. 22, 74 (1951).

37A. C. Wahl ond N. Bonner (eds.), Radioactivity
Applied to Chemistry, Wiley, New York, 1951.

 

 
 

of unlabeled metal. Uniform placement ensures
linear flow, and the layer is never replenished after
initial placement, The specimen is then subjected
to a constant temperature (T > 600°C) for a speci-
fied period of time during which the labeled metal
diffuses into the bar. Even though the specimens
may be of small dimensions, they may be con-
sidered as being infinite in length with respect to
the direction and distance of labeled-metal in-
vasion, since the rates of diffusion involved are
very low,

A solution of Eq. 4 under boundary conditions37+38
which closely approximate the described experi-
mental condition is

*
(5) o o o -x2/4D1 ’
(nD2)1/2

where all quantities other than QO* have been pre-
viously defined, and the starred variables refer to
the labeled metal. A physical interpretation of
Q;, g/cmz, is somewhat difficult for the case of
diffusional flow; it is analogous to an ‘‘impulse of
energy’’ for heat or electrical flow systems. [t
may be thought of as being the product of the layer
thickness and the layer density, the thickness
being zero. The layer is put on at an infinite
rate over a zero time interval.

A sketch of C* vs x at various times ¢ is shown
in Fig. 2.2.11. The quantity dC/dx at x = O is
zero, and thus dM/d¢ at x = 0 is zero for all ¢ > 0,
lAlso shown in Fig. 2.2.11 is a sketch of log C vs x2

at various times. |t is apparent from Eq. 5 that the
slopes of the lines of this plot are equivalent to
—~(4D1)~'. Thus D may be determined from a log

a vs x% plot. It is assumed in the sketches of

 

38R, V. Churchill, Modem Operational Mathematics in
Engineering, p 117, McGraw-Hill, New York, 1944,

o, ACTIVITY

   

el
| o

x(cm)

PERIOD ENDING DECEMBER 31, 1957

Fig. 2.2.11 that a has been corrected for time
decay. The diffusion coefficient determined for
the labeled material is essentially the diffusion
coefficient for the unlabeled material.

Excellent results3?=4! have been reported for
this type of experiment. In some instances, two
types of diffusion occurred simultaneously.®® The
two types are (1) diffusion through the grains and
(2) diffusion along the grain boundaries; the latter
controls the over-all diffusion rate at lower tem-
peratures, and a combination of the two controls
the rate at higher temperatures. In instances in
which both types of diffusion occur, log a vs x?2
plots tend to bend.?':3% Such results leave con-
siderable doubt as to the value of the over-ail
diffusion coefficient,

Labeled-Molten-Salt Method. — |f consideration is
given to a semi-infinite isothermal system wherein
an alloy containing metal M® is in contact (at one
face) with a molten salt containing a fixed amount
of MX,, it may be seen that at equilibrium the net
rate of transfer of M between the metal and the salt
will be zero, although a random dynamic exchange
of M will take place. If a small amount of the MX,
were removed and replaced by an equal amount of
labeled M*X,, the random dynamic exchange could
be studied by following the exchange rate of the
labeled M*X,. The expression which relates the
quantities of interest in terms of the surface ex-
change reaction or ‘‘wall reaction,"’

MO + M*X, =2 MX, + M°* , AF® = 0,

 

3. A. Johnson, Trans. Am. Inst. Mining Met, Engrs.,
Inst. Metals Div. 143, 107 (1941).

Op. {. Gruzin and G. B, Fedorov, Doklady Akad.
Nauk 5.5.5.R. 105, 264 (19553).

41s, T. Kishkin and S. Z. Bokstein, Proc. Intern.
Conf, Peaceful Uses Atomic Energy, Geneva, 1955 15,
87 (1955).

UMCLASSIFIED
ORNL-LR~DWG 27044

 

 

Fig. 2.2.11. Concentration of Labeled Metal at Various Times,

101
ANP PROJECT PROGRESS REPORT

[Mo*1{MX,]

ba K =c—.
(6a) T T IMX,)

The equilibrium of Eq. 6a is established within a
short period of time after the M*X,, is added. Since
K, = 15K,

a r

the concentration of M®* at the surface of the alloy
at any time is given by the equation

[M*X,]

ox = o
(6b) M*] _, = M ]—[sz]
The bracketed concentration expressions are taken
to be weight fractions.

The M°* will diffuse back into the metal as time
progresses and tend to deplete both the M°* at the
surface and the M*X, in the salt. Depletion will
occur unless M*X, is added to the system at a
rate equal to the diffusion rate. If the latter
manipulation is carried out, or if it can be assumed
that M*X, remains constant, [MO*]x:o is constant
with time. The solution of Fick’s law for this
case is

X

Effc -_—

() Cpou = Cyou

x=0

whereby

Adr |, Moo \mt '

and
AM* T ¢ /dM*
(9) =—f dt
A A 0 dt =0
_ac ] Dt 1/2
- Mo x=0 w :

in Eq. 9, AM* is the cumulative amount of M°*
(in grams) which has passed through A (cm?) of the
metal surface at any time ¢ (sec). The quantity
C,o+1,-o is the concentration of metallic labeled
M at the surface of the alloy (g/cm?).

 

 

 

If AM* is measured as a function of time under
the conditions stated, the diffusion coefficient can

102

be directly determined from the slopes of AM* vs
t172 plots. These slopes are equivalent to the
quantity 2 CMO*]x=0 (D/rr)'/z. The conditions set
forth for the equations would be difficult to dupli-
cate experimentally, however, since the necessity
of using small sample containers and of excluding
air suggests the use of sealed capsules. Con-
tinuous addition of M*X, could not be accomplished,
and Eq. 8 would be applicable only at early flow
times, In view of the low diffusion rates involved,
it is not important that the conditions cannot be
duplicated experimentally, Equation 4 can be
solved in such a way as to take into account de-
pletion behavior. A rigorous solution is

AM* 2
——— =1 - e%terfcat'? |
M*]
t=0
where a is proportional to D2, It can be shown

that Eq. 9 is a first approximation to the solution
for small valves of z.

Diffusion coefficients obtained by the labeled-
molten-salt method are of value because they are
over-all diffusion coefficients, that is, coefficients
that are independent of microscopic behavior within
the alloy, Further, the experimental procedure is
simple: the initial value of AM* is taken to be the
early-time depletion of M*X, in the melt, and the
melt is assayed at various times.

Battelle Diffusion Experiments

Isothermal capsule diffusion experiments have
recently been conducted at Battelle Memorial
Institute on Inconel—fused fluoride salt systems.,*?
Radicactive chromous fluoride was added to the
salt in order to study the combined effect of the
wall reaction and the diffusion coefficient on the
chromium transfer rate.

Briefly, the experimental procedure consisted of
placing a series of identical Inconel capsules,
each containing 2 g of salt and known amounts of
labeled CrF,, in a furnace. The capsules were
removed individually at various time intervals.
The salts in the capsules removed were analyzed
for the amount of GCrF, and Cr*F, present. In
addition, the walls of one capsule from each series
of isothermal tests were sectioned by an electro-
polishing technique to obtain the wall distribution

 

42R. Price et al., A Tracer Study of the Transport of
]C;Jg?;nium in Fluoride Fuel Systems, BMI-1194 (June 18,

 
of labeled chromium. The data obtained are used
below in a critical analysis of the proposed method
for the determination of diffusion coefficients.

Data Based on Molten-Salt Anolyses. — A portion
of the results of the Battelle experiments is repro-
duced in Fig. 2.2,12, This figure indicates that
(1) at low temperatures and (2) at early flow times
for high temperatures the amount of labeled chro-
mium which diffuses from the salt into the metal
varies directly with the square root of the flow
time. |t may also be noted that as time goes on
the experimental points at higher temperatures fall
away from the straight-line relationship, This
was to be expected because the quantities Ccr°*]x =0
and [Cr*Fz] were not held constant during the
experiment,

A reasonable approximation of the diffusion
coefficient can be obtained, however, from the
slopes of the straight lines in Fig. 2.2,12, The
expression relating the slope of these lines to the
diffusion coefficient can be obtained by combining

PERIOD ENDING DECEMBER 31, 1957

Egs. 6 and 9. The expression is

% Depletion

(10) m =
[t (hr)]'/2
A [Cre] /D \/?2
= (60) (200) — _ [ —
( ) ( ) NT pMeful [CrF2]<fl ) I

where p. . is the density of Inconel (g/cm?),
N.. is the amount of salt present (g), [Cer] is the
weight fraction of CrF, in the salt, reported as
Cr**, and m is the slope given on Fig. 2.2.12.
This method was used to estimate values of D
from the results of several capsule experiments,
The experiments were divided into two groups,
with the base salt for one group being NaF-ZrF
(50-50 mole %) and for the other NaF-ZrF ,-UF ,
(50-46-4 mole %). It was necessary to assume that
equilibrium amounts of UF,, UF,, and CrF, were
present in the base salt of the second group,
since data on the [UF3]/[UF4] ratios and the

[Cer] variations as functions of time were not

UNCLASSIFIED
ORNL—LR—DWG 27045

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60
/ - Ooc-'/ ",—-— N %
o
5 / /
y '//. /4/ ’ T
p—— @
3 40 //, ,/.-f
2 | / .
=
o /V ® /
L / /
E.J A - [Cer] SLOPE —
w ( cn % /hr'2)
= ppm Cr (% /hr
. A 1467 0.0652
;.;_ / O 2285 0.3125
~ 20 o ® A 1444 1.800"
© . A ®  NOT AVAILABLE
/ * AVERAGE
®
/ o 7000(:/
e
pummemm
/ EL’_———‘——‘—‘U—’_-—
/
—/ o
0
0 10 20 30 40
4/2

SQUARE ROOT OF EXPOSURE TIME (hr

)

Fig. 2.2.12. Net Loss of Ce*F, Activity from NaF-ZrF ,-UF, (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
Fig. 2.2.15, which is discussed in a subsequent
section of this report,

Data Based on Distribution of Labeled Metal in
Wall. — The Cr°* wall distribution data formed the
basis for the values of the diffusion coefficients
reported by the Battelle investigators. Equation §
was utilized for correlating the results, although
the boundary condition

am

dt x=0
was obviously violated by the experimental condi-
tions. A plot of one set of the Battelle distribution
data is shown in Fig. 2.2.13. The ‘‘volume"”’
diffusion curve reported by the Battelle investi-
gators, as well as the distribution predicted by
Eq. 7, was superimposed over the data, and both
curves failed to fit the data. The ‘‘average D'’
calculated by both equations was 10='3 em?/sec,
The fact that the curve of Eq. 7 did not fit casts
some doubt as to the validity of the related Eq, 9.

UNCLASSIFIED
ORNL—LR—DWG 27048
4000

 

2000 S

 

1000 oo ANEAN FROM EQ 7

FROM EQ5 \
\
400 N

 

 

 

 

SPECIFIC ACTIVITY OF LAYERS {counts/min - mg)

 

 

 

 

 

 

 

 

 

\ DATA FROM
200 \“ REF 42
\\\\\
\\
NN
100 \—y
L\
AN
\‘ \
80 L\ N
v\
40 A\
0 4 8 12 16 20

PENETRATION DEPTH {u)

Fig. 2.2.13. Ce®! Distribution in Incone! at 700°C
After 739 hr of Exposure.

104

In order to evaluate the discrepancies, the value
of D from the distribution data was used to con-
struct a depletion curve, This in turn was com-
pared with the actual depletion data obtained for
the same experiment., The results are shown in
Fig. 2.2.14.

Additional comparisons with chromium diffusion
coefficients available in the literature*® are pre-
sented in Fig. 2.2,15, These values are also
tabulated in Table 2.2.21 along with the corre-
sponding activation energies for diffusion. Although
Eq. 7 did not fit the distribution data, it was con-
cluded on the basis of Table 2,2.21 that Eq. 9
was suitable for estimating values of the over-all
diffusion coefficient.

The magnitudes of the variations encountered in
terms of amount of material transferred are not
immediately apparent from plots such as Fig. 2.2:15;
hence, calculated rate values are presented in
Table 2.2.22 to give a '‘feel for the numbers’’ in
terms of rate of diffusion and amount transferred as
a function of time.

The values presented in Table 2,2,22 were cal-
culated by using Eq. 2 for the amount of steady-
state chromivm diffusion which would occur over
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-ZrF ,-UF ,
salt mixture at 900°C., The sheet was assumed to
be 3 mils thick, with a 130-em? 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-CONYECTION LOOPS

R. B. Evans

investigations?® of the
corrosion of polythermal Inconel loops have shown
that chromium is transported®* from hot to cold
zones by circulating NaF-ZrF4-UF4 salts. Mass

transfer of chromium also occurs with other salt

The results of many

 

4':‘,'W. D. Manly et al., Metallurgical Aspects of Molten
Fluoride Reactors, ORNL-2349 (Sept. 17, 1957).

44y, R, Grimes, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 96.

 
 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 27047

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12 :7.-; 7
. /' 1)
i
. 1 /
. . / _ o —
2 8 = =
— e
a I / — - o
& ! / —_ g
x -
"U'S , - e —
2 ! ////
o o~
— -
3 I y ~ DIFFUSION COEFFICIENT BASED ON
© 4} A~ ’ (1) cr°® DISTRIBUTION IN WALL
, * / (2) cr¥F, DEPLETION RATE
. /| {3) DATA FROM GRUZIN AND FEDEROV7
@ EXPERIMENTAL PQINTS USED TO CONSTRUCT CURVE (2)
®
! °
0
. 0 200 400 600 800 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*

 

 

T ¢ Motal Chromium  Diffusion Coefficient tor Diffosi Val R b Data
Ype © eta (wt %) (cmz/sec) or httusion alues Reported by Reference
(keal /g-at.)
x 10713

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 4] Price et al, 42

-Eflc‘l/RT

*DflDOe .

**Extrapolated values.

105
ANP PROJECT PROGRESS REPORT

Al
ORNL-LR-DOWG 27048

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TEMPERATURE (°C) -
1200 1100 1000 900 800 700 600
17
T | | | l | | l
DATA -
SYMBOL  CONTAINER SALT REFERENCE
6 | 4  NICKEL NONE 40 e .
a Ni—Cr ALLOY 2 NONE 40
0 Ni—Cr ALLOY 3 NONE 40 ’
®  Ni-CrALLOY 4 NONE 40 7
5 |— X  INCONEL¥ NaF —ZrF, —UF, <
D INCONELY  NaF=ZrF, e
a INCONEL NONE 42 oK
¢  INCONEL NONE 42 47 0 '
14 |— =
//
/”
S ¥ DATA FROM PRESENT INVESTIGATION X -7
o3 * 7 i
o BASED ON Cr™ F; DEPLETION 4 ’ \3‘4&;,
T *¥*DATA FROM REPORT BY BATTELLE 42 3 NOZ - -
BASED ON Cr®™ WALL DISTRIBUTION / /2 u V-
12 e ) P
y/f’ od=~
~ ) v
7y a”
/ ,,/ &
1 A 1
5 " -
% I
-
-~ w
10 ~ =~
o -
6 T 8 9 10 " 12
10*/7 (oK)
Fig. 2.2.15. Comparisons of Yarious 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 Dato
Sheet Material Coefficient o
2 (g/sec) {g/year) Coefficients Reported By Reference
(cm*/sec)
*
inconel 8 x 10-13 1.4 x 108 4.5 Present investigation
Ni-Cr alloy 3 8 % 10_]3 1.8 x 10-8 5.7 Gruzin and Fedorov 40
Inconel** 3% 10712 52x 1078 17 Price et al. 42
Inconel*** 2x 10711 3.5x 1077 110 Price et al. 42 .

 

*Extrapolated value.

**By volume diffusion alone.

***By boundary diffusion alone.

106

 

 
systems and metals; however, this report deals
only with the chromium migration peculiar to the
NoF-ZrF“-UF4 salt system contained in Inconel.

In all experiments the loops have been loaded
with pure salt mixtures, that is, mixtures which
did not initially contain corrosion products.
Therefore the over-all transport vs time relation-
ships observed could be broken down into three
consecutive flow situations or regions. The
study of each region required a separate analytical
approach, since the first two regions involve
unsteady-state solutions and the transport phe-
nomena corresponding to the third and final region
are governed by a quasi-steady-state situation.
The analysis of the third region is relatively
simple compared with the analyses of the un-
steady-state regions. The final steady-state
situation is of considerable interest because low
migration rates occur in this region.

One objective of the present study is to develop
a method whereby the amount of chromium trans-
ported can be estimated as a function of loop
operating time, and thus to be oble to predict
the extent of corrosion to be expected at any
time. This method will be applicable only for
the case in which loop operation is initiated with
the equilibrium concentration of the corrosion
products in the salt mixture. Chromium migration
behavior for this case approximates the behavior
in the quasi-steady-state region, with the ex-
ception that the distribution of chromium in the
wall does not depend on what might have happened
during the unsteady-state period. The finding
of a relatively simple method of estimating the
equilibrium concentrations is a further objective
of this study.

Chromium Migration Mechanism

The chromium migration process in thermal-
convection loops involves two chemical changes
and two transport mechanisms, Over the entire
hot zone, the metallic chromium slowly diffuses
from the matrix of the Inconel tubing to the salt-
exposed surface. The chromium concentration
gradient in the metal is the driving force for the
diffusion process. The transported chromium is
then oxidized at the surface according to the
reaction

() C° + 2UF, = 2UF, + GrF,

PERIOD ENDING DECEMBER 31, 1957

and is carried as CrF, to the cold zone via the
circulating salt. At the cold-zone surface, the
reaction is reversed, and the chromium reverts
to the metallic state. The deposition process
results in a chromium concentration gradient
along the cold-zone surface which causes the
chromium metal to diffuse into the tube walls,4?
The diffusion into or out of the walls occurs in
a direction which is essentially normal to the
surface and is governed by semi-infinite, linear
flow relationships.

The CrF2 concentration in the salt builds up
initially with respect to time and then remains
essentially constant. Thus a quasi-steady-state
situation appears to exist whereby chromium is
transported at very low rates and under conditions
of a fixed chromium surface concentration vs
loop position relationship. This idea is supported
by the fact that corrosion continues even though
the amounts of CrF,, UF,, and UF, in the salt
remain constant.

Steady-State Migration

Based on experimental evidence alone, there is
still some doubt as to whether the loops operated
thus far have actually attained the quasi-steady
state.  Although the CrF, concentration remains
constant, within the accuracy of the measurements,
it has not been possible to make a direct, quan-
titative measurement of the total over-all amount
of chromium transported. An irrefutable statement
about the existence of a quasi-steady state cannot
be made without such measurements,

Another approach to a study of the migration is
to assume that the equilibrium amounts of CrF
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,
the following conditions may be specified for the
resulting migration:

I. At any given time the cumulative amount of
chromium which leaves the hot zone must equal
the cumulative amount deposited in the cold zone,
Furthermore, the over-all instantaneous rate of
transfer out of the hot zone must equal the in-
stantaneous rate of transfer into the cold zone.

 

4'ls'l'he equilibrium constants for the reaction
Ni° + 2UF4—,;‘ ZUF3 + NiF2

are so low compared with those for the chromium reaction
that the transport of nickel by the mechanism described
here may be ignored.

107

 
ANP PROJECT PROGRESS REPORT

2. The over-all rate of transfer is diffusion
controlled, since the diffusion rates are very much
lower than the other rates involved.

3. There must be at least two points within the
loop at which chromium is neither deposited nor
depleted. In subsequent discussions, these points
will be referred to as the balance points.

4, At the balance points (a) no diffusion takes
place within the walls; (b) the chromium concen-
tration ot the surface is the same as the concen-
tration in the original Inconel; (c) the circulating
UF3, UF,, and Cer, the concentrations of which
do not vary with loop position or time, must be in
equilibrium with the original concentration of
chromium in Inconel,*® and thus the concentrations
of the circulating salts are fixed by the equilibrium
constants {and temperatures) at these points; and
(d) the temperatures controlling the equilibrium
constants must be the same.

5. The loop position vs temperature relation-
ships are constant with respect to time. Thus
the balance points must remain at fixed loop
positions.

6. The concentration of chromium at any point
on the tubing surface varies with loop position
but not with time.

Unsteady-State Migration

As previously pointed out, unsteady-state cor-
rosion arises in loops in which the salt initially
contains UF'4 but no UF3 or Cer. At zero time,
tor the entire loop area will be a depletion zone
(zero deposition zone), and it will remain so until
time ¢_ is reached. At t_, the CrF_ concentration
of the circulating salt will have increased to
an amount Z, where Z is the concentration of
CrF, in equilibrium with the chromium concen-
tration of the wall surface at the position which
has the lowest temperature of the loop. The
balance points are then a single point at this
position, As the concentration of CrF2 in the
circulating salt increases to values greater
than Z, the initial balance point will develop
into a growing deposition area bounded by two
balance points. Steady state will be attained
when the balance points become stationary and

 

46This statement is based on the important assumption
that chemical equilibrium with respect to Eq. 1 is
assumed to exist at every surface point within the
loop. Concentration gradients within the salt normal to
the circulation path are ignored.

108

the CrF, concentration in the circulating salt
ceases to vary with time. The first two regions
mentioned in the introduction are separated by
the point Z. The last region starts when the
balance points become stationary.

A very important observation can be made on
the basis of the above discussion; that is, the
over-all amount of chromium transferred during
the unsteady-state time interval would be higher
than that transferred in any identical pre-equili-
brated loop which started at the same time. This
observation is valid, since the depletion areas
and driving forces for the unsteady-state case
are greater (CrF:2 concentrations much lower)
than those for the steady-state case. In addition,
the degree of chromium depletion (corrosion)
per unit area in the maximum-temperature portion
of the loop is highest in the unsteady-state case.
In other words, pre-equilibration is good operating
procedure with respect to suppressing chromium
migration within loops. The remainder of this
discussion will deal with loops which have been
pre-equilibrated at conditions close to the op-
erating balance-point conditions.

Estimating Equilibrium Concentrations for
Steady-State Loop Operation

The equilibrium (balance-point) concentrations
of UFA, UFS' and CrF2 can be estimated for
steady-state loop operation by applying the six
conditions previously outlined to a given set
of loop parameters and operating conditions,
Basic material data needed to carry out numerical
calculations are (1) diffusion coefficients for
chromium in Inconel as a function of temperature, 4’
and (2) equilibrium ratios for the redox equation
(Eq. 1) as a function of temperature.48:4?

Tentative values of these variables are given
in Fig. 22.16. The diffusion coefficients are
based on extrapolations of published data and
the equilibrium ratios are based on an inter-
polation of experimental values for two NaF-
ZrF4-UF4 mixtures in contact with chromium.
In order to present a numerical example of the
calculations involved, the following parameters

 

475, L. Gruzin and G. B. Fedorov, Doklady Akad.
Nauk S5.5.5.R 105, 264 (1955).

48, p. Redman, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 132.

49L. G. Overholser, private communication te R. B.
Evans,

 

 
UNCLASSIFIED
ORNL-LR-DWG 27049

TEMPERATURE (°K)

1080 1040 1000 960 920
20
(x10”9)

15

O N ® WO
O’)-JOJLDO

n

SQUARE ROOT OF DIFFUSION COEFFICIENT

UFg)2[CrF,
CORROS!ON REACTION CONSTANT, X, _ R[]
¥ [uR]4[cr]
OF CHROMIUM IN INCONEL [(cm?/sec) 2 ]

-~
o

 

92 9.6 100 10.4 10.8
10%/7 (°K)

Fig. 2.2.16.

Which Control Rate of Chromium Migration in Inconel-

Effect of Temperature on Parameters
Salt Systems.

and operating conditions, which are those of a
typical experimental loop, were assumed:

Salt NOF'ZI'F4'UF4 (50-46-4 mole %)

Flow rate 1 fpm

Loop Inconel tubing, 0.5 in. OD, 16 in.
long, containing 16 wt % Cr

Maximum temperature 1060.8°%K

Minimum temperature 933.3°K

In addition it was also necessary to make some
simplifying assumptions regarding both con-
centration and temperature effects at certain
loop positions and time. The reasons for this
are brought out in the following section, in which
the diffusional flow equations for applicable
steady-state migration in loops are discussed.

Mass- Transport Equations for Steady-State
Migration

Condition 2, given above, for steady-state mi-
gration implies that the over-all migration rate can

PERIOD ENDING DECEMBER 31, 1957

be written in terms of the diffusional flow which
takes place in the metal; that is, it is assumed
that the circulating rate of the salt does not
contribute to the transport equation.

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 A4,
the following equation may be written to relate
the chromium concentrations, C, time, ¢, the
diffusion coefficient, D, and penetration distance,
x:

 

(2) C’(x’t)cmld zone Inconel
X
+ (Csurface Inconel) erfc
and
(3) C(x’t)hof Ione surface
X
+ (Clnconal surfuco) erfc
2(Df)]/2

The related rate expression for both the cold and
hot zones is

M aC
(4 — = — DA —-—}
¢ d% surface
( D )1/2
= AAC| — ,
mt
where AC = Ceuvrface — Clnconelr The signs of

dM/dt, which is the weight of chromium diffusing
per unit time (in grams per second), depend on
AC; thus, a positive dM/dt means that chromium
is diffusing into the wall. The cumulative amount
which will have migrated across any unit area of
surface from time zero to any time ¢ is given by

Dt \ /2 am
(5) AM = 2AAC | — = 2t — .,
" dt

The chromium concentration at the tubing surface,
that is,

C = [Cr°]

surface surface pmefcll ’

where C and p are in density units and [Cr°] is
expressed in terms of weight fraction of chromium

109

 

 

 
ANP PROJECT PROGRESS REPORT

in the metal, is governed by the reaction
(6) Cr + 2UF, = 2UF, + CrF2 .
Thus,

[UF ;12 [CrF )
(7) KICP] = —————
(UF ]2

The right-hand side of Eq. 7 is constant, according
to steady-state condition 4(c). Therefore

KBP[Cr°]BP = KL[CrC’]L .

where the subscript BP refers to balance point
and L refers to a unit area at any loop position, or

pmefal Inconel K

L
The rate of diffusion as a function of time at any
position L may then be obtained by combining
Eqs. 4 and 8 to give

Kpp
(8) AC = [Cr°] — =]

daM

(9) = ) [Cr°]

X

Inconel

KBP p \1/2
x - 1 -
KL Tt

In order to obtain the over-all rate at which the

metal

 

chromium is being transferred, it is necessary to
integrate the instantaneous rates over an entire
hot or cold zone, as follows:

aM m 1/2
(]0) a’_t = 2rpmefu| [CI’ ](*;')

L
2 / K
BP
f —-l> pV?2 4
L K

1 L

where r is the radius of the Inconel tube. If
T, <Tgp, K <Kgp, and dM/dt > 1, chromium
is diffusing into the wall in the cold zone; the
converse is true for the hot zone.

The integration indicated in Eq. 10 cannot be
performed until the K, and DV2 ys L relation-
ships are known. Two important parameters are
involved: (1) the wall temperature, which controls
D2 and {2) the metal-salt interface temperature,
which controls K, . These temperatures must be

110

available as a function of L in order to obtain
over-all values of dM/dt.

Salt and Wall Temperature Relationship

When a loop is externally heated, the temperature
of the circulating molten salt, T_, will lag the
loop wall temperature, T, , with the temperature
gradient being a function of the heat transfer
coefficient and heat capacity of the salt, A heat
balance on the simplest model of such a system
is given by the equation

de
(1n —a’t— + B(Tw - Ts) =0,
where
2h
B = — 7
pCpr

the quantities b, C_, and p are the heat transfer
coefficient, specific heat, and density of the salt,
and r is the inner radius of the tubing.

An equation relating [ tubing length, L,
velocity of circulating salt, », and time, ¢, must
be obtained in order to solve Eq. 11 under steady-
state conditions.  Actually all that must be
assumed is the shape of the plot of T, vs loop
position, |t is not at all unreasonable to assume
that the equation of the plot is

(12) T, =T + asin wt ,

mean

where a is one-half the loop AT, that is, 0.5
(1060.8 - 933.3) °K, and w is 27 times the flow
velocity divided by the |length of the loop.
The salt temperature, which may be obtained by
combining Eqgs. 11 and 12, is
Ba

T =T f— sin(at - @),
S medadn (B2 . 0)2).'/2 gs

w
¢ = tan~! —
B

A plot of T and T_ is shown in Fig. 2.2.17.
The physical properties of the salt, Nc:F-ZrF“-UF4
(50-46-4 mole %), give a value for B that is very
nearly equal to the value of @ for the loop de-
scribed above. Therefore ¢ was chosen to be
45 deg. A rotating vector representation of the
T, and T_ relationships is also shown in Fig.
2.2.17.  After the graph was prepared, L was
F- 7 mean (°K)

COOLED SIDE

80 —

-80

     

 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR-DWG 27050

L=12 in.

 

 

r~r~
i
S
3
-
ol
nlo
fif
P~
1l
m
5
HEATED SIDE

 

 

wao L

 

L =4in.

T mean = 997°K iy

 

 

   

 

 

 

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

|
0 g T 3 21T

 

w/ (radians)

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

11
2.2.19. The curves on Fig. 2.2.19 are tentative;
however, they strongly suggest that the amount
of steady-state chromium migration within actual
loops (similar to the type considered) will be very
small.

ACTIVITIES IN ALLOYS
S. Langer

Measurements of the activities of the metallic
components of container materials for fused-salt
fuels have been made in order to study the effects
of composition and temperature on the activities
and thus on the corrosion resistance of container
alloys. The activity of chromium in nickel-
chromium alloys (Inconel) has been studied,3°
and data are presented here for the nickel-molyb-
denum system. In future work the chromium-
molybdenum system is to be studied, as well

 

M. 3. Panish, ANP Quar: Prog. Rep. Sept. 10, 1956,
ORNL-2157, p 100.

PERIOD ENDING DECEMBER 31, 1957

 

as the ternary system nickel-molybdenum-chromium,
which includes the INOR alloys developed by
the ORNL Metallurgy Division.

Measurements of the activity of nickel in Ni-Mo
alloys are being made by using the emf method
described in the previous report.3!  The cell
being used for the measurements is designed
as shown in Fig. 2.2.20 to permit the tight gas-
keting required to maintain an inert atmosphere
over the liquid metal bath in which the cell is
immersed. The liquid metal bath was found to
be necessary to prevent oa Iqrge temperature
gradient along the electrodes. A tin bath was
unsatisfactory because o hard, sintered mass of
oxide formed on the surface and frequently cracked
the cells. Liquid bismuth baths are quite sat-
isfactory, ond the small amount of oxide that
forms is not deleterious to the cell.

 

5ls, Langer, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 130.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L
ORNL-LR-DWG 27052
1.0 {
-
—
”
-
-
0.8 -~ /T 0.8
' .-v’, aM / '
"/ cr® .
on
— // E
o ~ =
£ - 5
Eo.e /,’ 0.6 E
o e :
nd: ” =
> 7 ; Tware ] = Tsaur T
S ’ ¥=o0 T
— ” _
< 0.4 7 -~ - ToaLT] = TwarL | 0.4%
O v X=0 X —eo 8
= / =
/7 <
/’ CONDITIONS::
/ i6-in. LONG,0.5~in.— OD INCONEL LOOP CIRCULATING
0.2 \v dMcpo ——— NoF - Zng- UF, (50—46—04 mole o) 0.2
% a7 7(-JV=997 K Afloup=127 C
/ \/ FLOW VELOCITY = 1 fpm
Py
i__- -l-_-_- - —
.-_-_—---—_-=_——-—
0 l 0

 

 

 

 

0 2 4 6

8 10 {2 14

LOOP OPERATING TIME (months)

Fig. 2.2.19. Chromium Migration Characteristics for o Pre-equilibrated Thermol-Convection Loop.

113
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 27053

 

THERMQCOUPLE WELL

 
 

TUNGSTEN-GLASS SEAL TO HELIUM SUPPLY

   

SILVER-SOLDERED TUNGSTEN-NICKEL JOINT

w/('r—-———'——\w

 

CRQSS SECTION OF CELL

 

 

 

ELECTRODE—__ |

 

 

 

 

 

 

 

GROUND FUSED SILICA PACKING

Fig. 2.2.20. EMF Cell for Measuring Alloy Activities.

114
\»

Only preliminary data on two alloy compositions
are available thus far. The activity of nickel in an
alloy containing 90 mole % Ni is 0.70 at 800°C.
This corresponds to an activity coefficient of
0.76. The behavior of this alloy in the cell
appears to be quite reproducible. An alloy con-
taining 80 mole % Ni has a nickel activity of
about 0.32 at 800°C. The activity coefficient
in this case is about 0.40. The data on this
alloy have not yet been reproduced to within
the desired precision.

The preliminary data seem to indicate large
negative deviations from Raoult’s law, From
inspection of the nickel-molybdenum phase dia-
gram,52 this is to be expected. The 90 mole % Ni
alloy is in the solid-solution region, and the rather
large negative deviation should not be surprising
in view of the proximity of the compound MoNi,.
The 80 mole % Ni alloy might have been expected
to exhibit even greater deviation from ideality,
since it is within the region of stability of the
compound at 800°C.

It is expected that the precision and accuracy
of the preliminary data will be improved by meas-
urements now being made. Measurements are
to be made of the activities of alloys containing
as high a molybdenum content as is metallurgically
feasible over a temperature range between 700 and
1000°C. Measurements over this temperature range
will permit the determination of the temperature-
dependent thermodynamic properties, and the data
obtained at temperatures above 900°C will have
additional interest, since the 900 to 1000°C
temperature range is above the region of stability
of both the compound MoNi, and the higher-
molybdenum compound MoNij. In this region it
would be expected that the activity of nickel
would be more nearly ideal.

DISSIOLUTION OF STAINLESS STEEL FUEL
ELEMENTS IN FLUORIDE MELTS

J. E. Eorgan  G. J. Nessle

The possibility of dissolving stainless steel
fuel elements in molten fluorides, with the re-
sultant advantage of adaptability to the present
volatility process, made it desirable to do some
exploratory experiments to determine the feasi-
bility and general characteristics of such a

 

32 American Society for Metals, Metals Handbook,
1948 ed., 1230, American Society for Metals,
Cleveland, 1;48.

 

PERIOD ENDING DECEMBER 31, 1957

process. The choice of a solvent salt was based
on the ability of the salt to dissolve or dis-
integrate a set minimum quantity of stainless
steel fuel elements to yield a concentration of
uranium in the molten salt which would be
economically feasible for adaptation in the
volatility process. From previous experience
in handling various fluoride mixtures, the alkali
fluoride compositions containing NaF, LiF, and
KF seemed to be most promising. The composition
chosen was NaF-LiF-KF (11.5-46.5-42 mole %),
which combined relatively high solubilities of
the structural metals nickel, chromium, and iron
with a fairly low melting point (460°C).

For the first experiment, which was run to
gain knowledge of the general characteristics of
the dissolution process, a 2-kg batch of the
chosen fluoride mixture was purified according
to normal procedures in a copper-lined reactor
vessel. The reactor vessel was equipped with
copper thermocouple wells and dip lines, so
that copper was the only metal that contacted
the molten salt. After the purification and
sampling steps, a type 347 stainless steel plate
4 in. square and 0.0625 in. thick, which weighed
approximately 71 g, was suspended in the melt
attached to a copper rod. The temperature was
maintained at 650°C, and anhydrous HF was
bubbled through the molten salt at approximately
1 liter/min. Hydrofluorination was continued for
a period of 24 hr, and salt samples were taken
every 4 hr. A sample of the stainless steel
was submitted for analysis after the 24-hr period,
and the following report of the analysis was
received:

Component Amount (wt %)
Ni 10.3
Cr 17.6
Fe 69.8
Nb 1.72

The results obtained from analyses of the seven
salt samples are compared in Table 2.2.23 with
the theoretical concentrations in the salt if the
structural metals in the type 347 stainless steel
plate had completely dissolved.

At the conclusion of the experiment the molten
salt was transferred from the reactor vessel so
that the copper liner could be examined. It was
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- (hl’)
ppm g PPM g ppm g
Type 347 stainless 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
dissolved or disintegrated and that some odd
nodule-type salt formations had remained in
the reactor, attached to the support rod and the
dip leg at about the liquid level. These formations
appeared to be water soluble, and no metal
particles were visible. From the results given
in Table 2.2.23, it seemed that the plate was
completely dissolved at the end of 12 hr, since
metal concentrations did not increase after this
point. All the nickel was accounted for in the
bulk of the transferable melt, but only half the
iron and less than one-fourth the chromium could
be found.

The first experiment established that dis-
solution of the steel plate was possible, and
therefore a second experiment was run to in-
vestigate further the disposition of the structural
metals in the molten mixture. The conditions of
the first experiment were duplicated as closely
as possible, but only three filter-stick salt
scmples were fclien, on the assumption that
dissolution of the steel plate would occur as
it had in the first experiment. An attempt was
made to obtain a closer material balance in order
to give a better idea as to the disposition of
the dissolved metal. At the completion of the
experiment the salt batch was transferred from
the reaction vessel into a tared receiver can

116

through a sintered nickel filter with 0.0015-in.
pores.

The nodule type of formation occurred again on
the steel plate support rod and the dip leg. A
rough material balance is given below:

Amount (g)
Type 347 stainless steel plate 71
Total salt charge 2000
Transferable salt 1657
Filter-stick samples (3) 67

Nontransferable salt left
in reactor

347 (by difference)

Samples of each item were submitted for chemical
analysis to determine the distribution of the
nickel, chromium, and iron, and the results are

presented in Table 2.2.24.

Accurate weights of the salt sections that
made up sample 4 were difficult to obtain, be-
cause these nodules had to be broken off the
support rods and the sides of the copper liner.
The weights were sufficiently accurate, however,
that by making some liberal assumptions a rough
material balance was obtained. Samples 2 and
3 were assumed to be the most homogeneous
samples as far as nickel concentration was con-
cerned, and the analyses were in good agreement

 

 
PERIOD ENDING DECEMBER 31, 1957

Table 2.2.24, Results of Second Experimental Study of the Dissclution of Type 347 Stainless Steel
in NaF-KF-LiF Eutectic by Hydrofiuorination

 

 

 

Metal Concentrations in Salt Samples

 

 

 

Sc:qmple Sample Description Ni Cr Fe
o.
PpPm 9 % g % g
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 hydroffuorination 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 (V347 g)

5 Transferable material collected in re- ?100* 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-
periment. A rough material balance of the struc-
tural metals in the fluoride salt is given in Table
2.2.25, and fairly close agreement with theoretical
values may be seen.

In order to determine the effect of uranium on
the dissolution of type 347 stainless steel fuel
elements in an alkali metal fluoride salt such
as NaF-LiF-KF (11.5-46.5-42 mole %), two ex-
periments were run in which the uranium-bearing
salt NaF-LiF-KF-UF, (10.9-44.5-43.5-1.1 mole %)
was used as the solvent. In a practical ap-
plication the uranium would come from the fuel
elements.

Table 2.2,25. Material Balance of Structural Metals
from the Dissclution of Type 347 Stainless Steel Plate
in NoF-LiF-KF Eutectic

 

 

Weight Metals Found
Sample of Material in Salt (g)
No. Analyzed
Ni Cr Fe
(a)
3 1724 6.4 4.5 23.4
4 347 0.8 113 33.3
Totals 2071 7.2 15.8 56.7
Theoretical values for 71 7.3 12.5 49,5
type 347 stainless
steel

 

Of greatest concern in the dissolution process
was the possibility of concentrating the uranium
in the nontransferable, that is, high-melting,
phase of the salt bath during dissolution of the
steel fuel element. Previous experiments had
shown that such phases could occur.

In the first of the two experiments, a steel
plate 4 in. square and 1/16 in. thick, which weighed
71 g, was dissolved in 2 kg of molten salt at
1200°F by bubbling HF gas through the bath at
approximately 1 liter/min. The dissolution process
was complete at the end of 12 hr, according to
analytical results. At the end of 24 hr the melt
was transferred through a sintered nickel filter
into a storage can. Approximately 300 g of
material remained in the reactor vessel as the
nontransferable high-melting phase. The uranium
analysis of the salt bath prior to dissolution of
the stainless steel was 6.3%. The analysis of
the transferable salt showed a uranium conteént
of 6.5%, and the nontransferable salt contained
2.4% U. The chromium and iron tended to con-
centrate in the nontransferable phase. In general,
therefore, the presence of uranium seemed to
have little effect, if any, on the dissolution
process, and the uranium showed no tendency
to concentrate in the nontransferable phase.

In the second experiment the conditions de-
scribed above were duplicated, except that a
piece of stainless steel 4 x 12 x l/“5 in., which

117
ANP PROJECT PROGRESS REPORT

weighed 213 g, was used. In this case dis-
solution again seemed to be complete at the
end of 12 hr, according to chemical analysis.
Attempts to transfer the salt bath through a
sintered nickel filter were fruitless, however,
even though the salt temperature was raised to
1400°C in 50-deg increments. The filter was
finally removed, and the salt was transferred
through tubing.  Approximately 740 g of salt
remained in the reactor vessel. This nontrans-
ferable phase appeared to be crystalline after
freezing and could be described as mushy at high
temperature. The nickel again appeared to remain
completely in solution, there being little difference
in nickel concentration between the transferable
and nontransferable phases. Small filter samples
taken at 6-hr intervals prior to transfer of the
bath gave nickel concentrations close to those
found in the bulk phases. The iron and chromium
again concentrated in the nontransferable phase
and were present to a fair extent as metal crystals.
The uranium analysis showed 6.5% in the original
bath, 6.7% in the transferable phase, and 3.3% in
the nontransferable phase.

In general it appears that the stainless steel
&oncentration used in this experiment was too
high, although the dissolution or disintegration
process was not appreciably affected. [f less
stainless steel were used, a practical process
could be evolved for removing the uranium from
the steel and leaving a constant and nonaccu-
mulative heel in the reactor vessel. This heel
could be flushed out periodically with a dummy
salt mixture if desired.

OXIDATION OF MIXTURES COMPOSED
OF SODIUM AND POTASSIUM

E. E. Ketchen

A study of the oxidation of sodium and potas-
sium mixtures under carefully controlled conditions
has been made in an attempt to determine the
effect of certain variables on the final composition
of the mixed oxides. The variables studied in-
clude (1) the original composition of sodium-
potassium mixture, (2) the temperature of oxidation,
and (3) the temperature and time of equilibration
following the oxidation.

118

For the first experiments, which were performed
at 25 to 30°C, the equipment consisted of a
glass capsule 6 in. long and 1 in. in diameter,
with a sintered glass disk installed midway in
the capsule. A similar capsule that was con-
structed of nickel was used at elevated tem-
peratures. The oxidation was accomplished by
the introduction of measured volumes of dry
air. The oxides were separated from the bulk
of the liquid metals by filtration, and the re-
maining liquid metals were removed by amal-
gamation.

The early runs, in which Na-K (50-50 mole %)
was used, gave a value of 25 for the Na-to-K
ratio in the mixed oxides formed at 30 to 35°C.
No significant difference was found at 200°C
if the oxides were removed immediately, but,
if they were formed at 200°C and equilibrated
for 40 hr ot 400°C, the ratio was found to be
less than 10.

The early experiments were not precise, and
therefore a more thorough study of the oxidation
was made at 30 to 35°C in an attempt to find
conditions that would yield more precise results.
The glass apparatus was modified to provide
a larger surface for oxide formation, :and all
operations were carried out in a vacuum dry
box in order to reduce contamination by water
vapor. The oxidation of two compositions of
the Na-K mixture was studied under the modified
conditions, and the results are presented in

Table 2.2.26.

It is evident from the lack of precision in the
data that a close control of the variables involved
is absolutely essential. It is apparent, however,
that the necessary control is extremely difficult
to accomplish. Traces of moisture, which is
the most likely contaminant, should tend to
reduce the Na-to-K ratio, and therefore the larger
values are probably closest to the correct ratio.
The lack of precision in these experiments makes
it difficult to establish any quantitative effect
of composition of the alkali metal mixture on
the Na-to-K ratio in the mixed oxides. The datq,
however, show a significant decrease in the ratio
vpon an increase in the K content from 49 to

77 wt %.

 
PERIOD ENDING DECEMBER 31, 1957

Table 2.2,26. Composition of Oxides Formed by Oxidation of Sodium-Potassium Mixtures at 30 to 35°C

 

 

 

 

Composition of Composition of Oxides Composition of Composition of Oxides
. Na-K Mixtures Na K Na-to-K Na-K Mixtures Na 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
Av 64 122 7.5 0.22 34
18.0 0.48 37
T Av 2517

 

 

 

119
ANP PROJECT PROGRESS REPORT

2.3.

PHYSICAL PROPERTIES OF MOLTEN MATERIALS

F. F. Blankenship

SUMMARY OF ARE FUEL VAPOR-PRESSURE
STUDIES

R. E. Moore

The vapor pressures of a number of fused salt
mixtures in the Nc:F-ZrF4 and NaF-ZrF“-UF4
systems were measured to obtain information
needed for the selection of a fuel for the Aircraft
Reactor Experiment. A quasi-static method of
vapor pressure measurement, which originated
with Rodebush and Dixon! and was later used by
Fiock and Rodebush,? was chosen for this work
because it appeared to be the most reliable method
for use at the temperatures and pressure range
(5-200 mm Hg) considered. The apparatus and
technique have been described by Beusman.?

 

TW’. H. Rodebush and A. L. Dixon, Phys. Rev. 26,
851 (1925).

2E, F. Fiock and W. H. Rodebush, J. Am. Chem. Soc.
48, 2522 (1926).

3C. Beusman, Activities in the KCI-FeCl, and LiCl-
FeC12 Systems, ORNL-2323 {May 15, 1957).

The ZrF, used for vapor pressure measurements
and for 'r11e preparation of purified mixtures was
material which chemical analysis showed to be
98 to 99% pure. Uranium tetrafluoride was ob-
tained from the Mallinckrodt Chemical Works.
The NaF used was reagent-grade material. The
NaF-ZrF, and NaF-ZrF UF, mixtures were
prepared by fusion under an inert atmosphere,
followed in every case, except that of the compo-
sition 75 mole % ZrF ,—25 mole % NaF, by treatment
with HF, H,, and inert gas, in the order stated.

A summary of vapor pressure data for pure ZrF4
and mixtures of ZrF4 and NaF is given in Table
2.3.1, and a summary of the data for UF ,-bearing
mixtures related to the ARE fuel or fuel concen-
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,
the data were refined or extended later.

Petrographic examinations4 of samples distilled
from binary mixtures containing more than 43

 

4Mc:de by T. N. McVYay, Consultant, ORNL.

Table 2.3.1. Vapor Pressures of NuF-ZrF4 Mixtures and of ZrF,

 

ZrF4 Content Temperature Range

Constants for the
Vapor Pressure
Reference

 

 

 

(mole %) (°C) Equation*
A B

43 8681034 7.289 7.34 ORNL-1609, p 106

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

50 812990 7,675 8.14 ORNL-1649, p 99**

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

75 754844 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 *

| A
g np(mmHg) = ———+ 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 Nc:F-ZrI'-'4--UF4 Mixtures

 

Constants for the

 

 

. T t

Mixture Composition (mole %) emperature Vapor Pressure Equation* Ref

Number NeF 7eF UF Range eference
o a Ty 4 (°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
T (°K)

 

Iog]op(mmHg)=—- + B .

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

mole % ZrF, show them to be virtually pure ZrF ,.
Thus the total pressures are approximately 800

SR
ORNL-LR-DWG 27054

 

equivalent to the partial pressures of ZrF,. 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

/7
/

 

NaF -40.73 mole % ZrF4—6.]8 mole % UF,, was
determined by titration to criticality, and its vapor

pressure was not measured. The vapor pressures 200

 

of compositions in the neighborhood of the ARE
fuel composition were low enough, however, that

300

\%‘
it

 

operational difficulties traceable to ZrF, vapor
were neither expected nor encountered in the ARE.

VAPOR PRESSURE (mm Hg)

Estimated vapor pressures for a number of other 200 [/

 

ZrF ,-bearing mixtures containing UF , were given /
in a recent report.® The composition given that /

 

is nearest to that of the actual ARE fuel contains 100 /
6.5 mole % UF4_40 mole % ZrF,. Its vapor
pressure at the fuel outlet temperature (1522°F)

 

 

 

 

 

 

 

 

 

 

 

 

is about 8 mm Hg. O 10 20 30

 

40 50 60 7O 80 90 100
ZrF, (mole %}

5S¢, Cantor, ANP Quar. Prog. Rep. June 30, 1957, Fig. 2.3.1. Vapor Pressures of Nc:F-ZrF4 Mixtures at

ORNL-2340, p 158. 912°C.

121
ANP PROJECT PROGRESS REPORT

VAPOR PRESSURES OF MOLTEN FLUORIDE
MIXTURES

S. Cantor

Vapor pressure studies have provided, as part
of the search for optimum fuels, directly de-
terminable physical property data, as well as a
measure of the thermodynamic activities of com-
ponents in fuel mixtures. The main objective of
the studies was to obtain the vapor pressures of
binary fuel solvents, such as the mixtures of
ZrF , with alkali fluorides. It could be assumed
that the addition of small amounts of fuel (2-4
mole % UF ,) to the solution would not markedly
alter the solution structure and that the vapor
pressure would change in accordance with Raoult’s
law.

As stated in the preceding section, a quasi-static
method developed by Rodebush and Dixon was
used to measure vapor pressures, [|he compo-
sitions to be studied were prepared by weighing
and mixing the purified components in a dry box.
The otealiwhluorides were purified by fusing,
slowly cooling the melt, and selecting clear crystal
fragments. The commercially obtained ZrF, was
further purified by vacuum sublimation of the best
hafnium-free material. All other pure components
and mixtures used were produced at the Laboratory.

It is important to note here that there was a
a concurrent effort to measure vapor pressures
being carried out at Battelle Memorial Institute

for which the transpiration or carrier-gas technique
was used, The results of the BMI research are
discussed here in relation to the work done at
ORNL.

The binary systems studied in both laboratories,
along with the references which most fully describe
the results, are given in Table 2.3.3. In most
cases it was necessary to measure the vapor
pressure of the pure components in order to be
able to calculate the activities. The references
for these measurements are given in Table 2.3.4.

For the ZrF , systems, it was usually found that
sharp decreases in ZrF4 vapor pressures occurred
as the alkali fluoride concentrations were in-
creased. Also, there was a correlation between
decreasing ZrF , vapor pressure and alkali cation
size. The purpose of studying the 7NoF-6ZrF4-
7NaF6UF , system was to investigate the effect
of partial substitution of UF, for ZrF, while
maintaining a constant NaF concentration. The
results clearly indicated that this substitution
decreased the vapor pressure, but to a far less
degree than increasing the amount of alkali
fluoride.

For the compositions studied at ORNL, it was
at first assumed that the measured pressures were
due to ZrF4 alone. Above about 40 mole % ZrF,,
this assumption, tested by sampling of the vapor,
proved to be quite valid. At lower liquid-phase
concentrations of ZrF ,, the probable formation of
vapor-phase compounds made the assumption

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

 

 

System Reference

LiF-ZrF4 Vapor Pressures of the Rubidium Fluoride—Zirconium Fluoride and
Lithium Fluoride—Zirconium Fluoride Systems, BMI-1199 (June
27, 1957)

Na F-ZI'F4 Vapor Pressures of the Sodium Fluoride—~Zirconium Fluoride
System and Derived Information, BMI-1064 (Jan. 9, 1956)

KF-ZrF4 ANP Quar. Prog. Rep. June 10, 1956, ORNL-2106, p 111

RbF-ZrF4 Vapor Pressures of the Rubidium Fluoride—~Zirconium Fluoride
and Lithium Fluoride—Zirconium Fluoride Systems, BMI-1199
(June 27, 1957)

7Nc.|F-6ZrF4-7N4.:F-6UF4 ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL-2221, p 146

NdF-Ber Vapor Pressure and Equilibrium Studies of the Sodium Fluoride—

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

 

122
TCIbIe 2.3.4.

PERIOD ENDING DECEMBER 31, 1957

References for Yapor Pressures of Pure Components

 

 

Compound Reference
ZrF4 Vapor Pressures of the Sodium Fluoride—Zirconium Fluoride
System and Derived Information, BMI=1064, p 7 (Jan. 9, 1956)
UF, ANP Quar. Prog. Rep. June 10, 1955, ORNL-1896, p 65
BeF2 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, BM1-1064, 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
by Sense® in order to explain apparent NaF partial
pressures obtained in the NaF-ZrF, system.
Vapor compounds also had to be postulated in the
LiF-ZrF4, RbF-ZrF,, and NaF-BeF, systems if
the thermodynamic stability of these systems was
to be preserved. Because the vapor above the
molten salt solutions under study at ORNL
(RbF-ZrF4, KF-ZrF,, NaF-ZrF ;) was virtually all
ZrF,, it was still possible to calculate ZrF ,
activities 7+8 directly from total pressure measure-
ments. These activities were then used to form
the analytical expression

(1) log y, = AN3 + BNj ,

where y, is the activity coefficient of ZrF , N
is the mole fraction of alkali fluoride, and A and B
are constants. This expression was used in the
Gibbs-Duhem equation, in the form

(2) Nldlogyl+N2d|ogy2=0,

where N. is the mole fraction of ZrF, and v, is
the activity coefficient of the alkali fluoride.

 

bK. A. Sense et al,, Vapor Pressures of the Sodium
Fluoride ~Zirconium Fluoride System and Derived In-
formation, BMI-1064 (Jan. 9, 1956).

75. Cantor and S. D. Christian, ANP Quar. Prog. Rep.
Sept. 10, 1956, ORNL-2157, p 108.

85. Cantor, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 145.

Equation 2, in turn, was used to calculate alkali
fluoride activities.”«® The alkali fluoride activities
could be considered to be only approximate,
because it was necessary to extrapolate Eq. 1 to
N, =1 in order to solve the Gibbs-Duhem equation.
The results of these calculations showed that the
alkali fluoride activity decreased sharply with
increasing ZrF4 concentration.

Once a large body of vapor pressure data was
accumulated for molten mixtures of ZrF, and
alkali fluorides, estimates of vapor pressures
were made for a number of unmeasured compo-
sitions of ZrF4-base fuels.? In these estimates,
the vapor pressures of ZrF, were assumed to be
linear functions of composition over the appropriate
extrapolation or interpolation range. The vapor
pressures of fuel mixtures composed of UF, and
alkali fluorides were too small to measure by the
methods used, but, with the vapor pressure known
for each pure fluoride used in the melt, estimates
of the upper limits of vapor pressure were compiled?
by. assuming that Raoult’s law was obeyed by the
alkali fluorides and that Henry's law was obeyed
by UF .

Any continuation of vapor pressure investi-
gations of fused fluorides of reactor significance
should involve systems containing BeF,. One

 

95. Cantor, ANP Quar. Prog. Rep. June 30, 1957,
ORNL-2340, p 158.

123

 
ANP PROJECT PROGRESS REPORT

such important system that is amenable to measure-
ment by the methods now in use or under develop-
ment is LiF-Ber. The determination of activities
in BeF, systems presents a greater problem than
for the corresponding ZrF ; systems. The difficulty
that arises is the inability to equate BeF, vapor
pressures to total vapor pressures except at higher
BeF, concentrations (75 mole % or more). In
compositions below these higher concentrations,
vapor compounds !0 start contributing sizably to
the total pressure. Therefore, it will be necessary
to obtain partial pressures of each component in
order to calculate activities. At present, the
carrier-gas technique is being developed to aid
in measuring the partial pressures. The transpi-
ration vapor pressures are not sufficient to be
useful in obtaining the activities of the alkali
fluorides in these mixtures. Recent evidence
convincingly shows that there are appreciable
quantities of dimeric and even trimeric molecules
in the vapor of most alkali halides.!! Since the
;rapor pressures obtained in the transpiration
method are dependent on the molecular weight of
the vapor, total pressures will still have to be
measured in order to help obtain the partial
pressures of the species involved.

A MASS SPECTROMETRIC STUDY OF
SALT VAPORS

L. O. Gilpatrick M. Blander

The vapors of some fused salts have been
shown, as stated above (and in ref 12), to contain
double and triple molecular species. An under-
standing of the fundamental nature of these
associated species in the vapor is necessary for
the interpretation of the vapor pressure data of
fuel mixtures in terms of the activities of the
volatile components. It is also necessary for the
estimation of the volatility of components of fused
salt mixtures.  Polymer formation can cause
components of salt mixtures to be even more
volatile as part of an associated species than the
pure component is, in some cases.

 

]OK. A. Sense, R. W. Stone, and R. B, Filbert, Jr.,
Vapor Pressure and Equilibrium Studies of the Sodium
fguoride—Beryllium Fluoride System, BMI-1186 (May 27,

37).

M
R, C. Miller and P. Kusch, J. Chem. Phys. 2
860 (1956). st J- Chem. Phys. 25,

12, A. Ochs, R. E. Cote, and P. Kusch, J. Chem.
Phys. 21, 459 (1953).

124

A molecular-beam mass spectrometric study,
which continued the work of Rosenstock et al,, 13
was undertaken in order to develop a method of
exploring the vapor-phase associations so that
the occurrence of the phenomenon and its effects
could be predicted. The mass spectrometer
provides a measure of the ionized progeny of
individual associated vapor species. With this
technique, the individual vapor species which
were the precusors of the observed ions can be
inferred, and by control and measurement of the
temperature of the vapor source the internal energy
of vaporization and of association of the detectable
species can be measured. The mass spectrometer
is uniquely suited to measuring these energies of
association,

Potassium chloride was chosen as a suitable
material with which to develop the method, since
prior studies of that compound had been re-
ported.3:11¢13 |t was incompletely studied by
Rosenstock by using a hot filament as a source of
ions in the mass spectrometer. This method yields
only relative energies of vaporization of the various
vapor species. The ion species observed were
K* and KCI*, which are derived from monomeric
KCI, primarily, and K2C|+, which is derived from
a dimeric species, K2C|2. The existence of these
species agrees, in general, with the findings of
Miller and Kusch!! and of Beusman.3

The mass spectrometer used in this study was a
6-in.-radius 60-deg Nier-type machine with a metal
envelope; it was evacuated by a mercury diffusion
pump.14 The accelerating voltage, 2000 v, was
supplied by a well-regulated source capable of
delivering a current of 2 ma. The ion-beam
intensity was measured by a vibrating-reed
electrometer. The output from the electrometer was
recorded on a 0 to 10 mv strip-chart recorder,
which produced a permanent record of the ion-beam
intensities.

The effusion cell used to produce the molecular
beam of salt molecules, as finally modified, is
shown in Fig. 2.3.2. The cell was heated from
the 110-v main by means of a voltage regulator
dnd a variable autotransformer, This arrangement

 

13H. Rosenstock et al., Mass Spectra Studies of Silver
Halides and Alkyl! Halides, paper No. 34 presented at
the 4th Annual Meeting of the American Society for
Testing Materials, Committee E-14 on Mass Spectrometry,
Cincinnati, May 28, 1956.

146, F. Wells, Stable Isotope Research and Production
Semiann. Prog. Rep. May 20, 1954, ORNL-1732, p 27.

 
(CELL EXIT SLIT { 1.8 mil WIDE)

  
   
  
 

TOP THERMOCOUPLE
{P+,10% Rd)

NICKEL

SALT SAMPLE
COMPARTMENT

THERMOCOQUPLE WELLS

HEATER COILS

CROSS SECTION OF CELL

FACE PLATE -

Fig. 2.3.2.

proved to be sufficiently stable as a source of
heater voltage for temperatures to be maintained
to within a few tenths of a degree for periods of
1 hr or more. Temperatures of .the cell were
measured by two platinum—platinum-rhodium thermo-
couples, one fixed in the cell body and the other
fixed in one of the slits. These thermocouples
were calibrated with U, S. Bureau of Standards
thermocouples prior to use. The cell body and the
cell slits were independently heated in the final
equipment, and an attempt was made to keep the
slits at a higher temperature than that of the cell,

RADIATION SHIELD
FLOOR

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
CRNL-LR-DWG 26649

RADIATION SHIELD TOP

 

RADIATION SHIELD CYLINDER

 

CELL SUPPORT PIN

<z

| <
CELL SUPPORT Cod

  
   

KOVAR-TQ-GLASS SEALS
Effusion Cell.

by at least 3°C, to prevent the deposition of salt
on the slits. The lowest temperature, that is,
the cell body temperature, was used in all
calculations as the temperature most representative
of the vapor temperature. The effusing stream of
salt vapors was roughly collimated in order,
primarily, to control the deposition of salt at
unwanted locations in the source. Salt deposits
upon insulators were an tarly source of electrical
leakage and are a troublesome factor in this type
of experiment. Only those molecules which are
intercepted by the ionizing electron stream are

125
ANP PROJECT PROGRESS REPORT

effective in forming ions, so narrow collimation of
the molecular stream is not necessary.

The effusion cell was charged with reagent-grade
potassium chloride which had been fused in
platinum.  Clear crystals of salt were chosen
from the melt and loosely packed into the effusion
cell. After the cell was assembled and the slits
were adjusted to 2.5 mils with a feeler gage, it
was mounted in the mass spectrometer and checked
for electrical leakage and thermocouple response.
The heaters were then adjusted to give the cell a
temperature of 350 to 400°C, and the outer envelope
of the mass spectrometer was raised to a temper-
ature of 250°C. This bakeout-degassing operation
lasted at least 16 hr and, in some cases, for a
number of days, until the hydrocarbon background
was reduced to a negligible level at the region of
the lower mass position for potassium chloride
(74).

The data were taken by raising the temperature
of the cell in increments of 20 to 30 deg and
making measurements until constant temperatures
and mass intensities were observed. This time
interval varied from more than ]'/2 hr at the lower
temperatures to less than 15 min at the higher
temperatures. An attempt was made to take data
from both an ascending and a descending sequence
of temperatures; however, the salt capacity of the
effusion cell was such that it showed signs of
exhaustion before any complete total sequence
could be measured.

If the pressure of the salt vapor is low enough
so that the mean free path of the molecules is
greater than the slit width of the effusion cell,
the number of molecules of a given kind escaping
from the slit should be a function only of the
temperature, Also if the area of the slit is small
compared with the inner surface area of the cell,
the pressure of salt vapor in the cell should
approach the equilibrium pressure at a given
temperature. In this experiment the ratio of cell
surface area to slit area was 800. The measured
fon current, or mass spectrometer peak height, of
any of the ionized progeny n.” of the species i as a
function of the temperature of the salt in the
effusion area is given by

dlnn/ AE,

(]) a(]/T) == R '

where E. is the energy of vaporization of the
species 7, a quantity which cannot be obtained

126

directly by any other experimental method. The
valves of the absolute partial pressure of the
species are not obtainable by this method because
the ionization efficiencies for the production of
the ionized progeny are not, in general, known.

If a salt species, KCl, forms a dimer, (KCl)z,
the following equilibrium obtains:

(2) (KCl), ==2KCli
The equilibrium constant K is given by

b
(3) K =—0 ’
Py

where p. is the partial pressure of the monomer
and p, is that of the dimer. The standard free
energy, AF°, is given by

(4) AF° = —RTIn K< AE® + RT — T As® .

The quantity ASS, which can be estimated by the
use of statistical mechanics, increases with an
increase of mass.’> The energy, AE® can be
estimated for alkali halides by considering the
ions as charged, polarizable spheres. This quantity
can also be obtained from the experimental data
by the relation

(5) AE® = 2AE, - AE, ,

where AE. and AE, can be calculated from the
measurements by using Eq. 1. These energies are
one of the two sets of factors that control the
association, and, at present, they can be measured
independently only by the method described here.
Several measurements were made on KCI prior
to the use of heated slits, and none of the experi-
ments were completely successful because of the
formation of deposits to a greater or lesser degree
at the exit slit. The deposits had the effect of
reducing the intensity of the beam in an arbitrary
and variable manner. A plot of the log of the
peak heights of the observed KCI* and K2CI+
species vs 1/T, as obtained in one of these
measurements, is shown in' Fig. 2.3.3, The slopes
of these lines, calculated by using Eq. 1, yield
the energy of vaporization of the monomer (AEI)'
and the dimer (AE,), and, by using Eq. 5, the heat

 

3R, Fowler and E, Guggenheim, Statistical Thermo-
dynamics, chap Ill, University Press, Cambridge, 1939,

LS

 
»

UNCLASSIFIED
ORNL-LR-DWG 27056

10%

RELATIVE PEAK HEIGHT
o
(&)

1400 1440 1180 1.220 1.260 1300 1340
% x 1073 (K"

Fig. 2.3.3. Relative Intensities of Monomer and Dimer
Species of KCl| as a Function of Temperature. Data
obtained from a mass spectrometric study of an effusing
beam of KCI.

of association (AE® can be obtained. The values
found, in kcal/mole, are:

AE, = 42.5 ,
AE2 = 49-8 7
AE® = 35.2 .

These values differ from the values of Beusman3

and those of Miller and Kusch.1l As mentioned
earlier, this discrepancy may be due to the problem
of salt deposition at the effusion slit. Such
deposits have not been reported elsewhere, although
they may be the source of the anomalous behavior
observed in similar experiments on oxides.16 The
modified equipment, shown in Fig. 2.3.2, has not
as yet produced satisfactory data.

SURFACE TENSIONS OF MOLTEN SALTS
S. Langer

The surface tension of a confined fluid has been
shown to be important, in certain cases, in heat
transfer 7 and in corrosion processes.'® A study
of the surface tension of molten fluoride mixtures
was therefore initiated. The sessile-drop technique

PERIOD ENDING DECEMBER 31, 1957

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 Nan--ZrF4 mixture
(53-47 mole %) at 600 and 700°C and preliminary
data at 800°C. A summary of the data obtained by
the sessile-drop technique is presented in Table
2.3.5 and compared with data obtained by Miles 19
by the maximum-bubble-pressure technique. A

Table 2.3.5. Surface Tension of NuF-ZlF“
(53-47 Mole %) at Yarious Temperatures

 

Surface Tension (dynes/cm)

 

 

 

Temperature
(°C) Data of Miles Author’s Data
600 127* 127 t 6
700 120 119 t 4
800 138 L 7**

 

*Average of three measurements corrected for more
recent values of the density.

**The large loss of weight of the sessile drop during
the course of the measurement renders this value highly
inaccurate for NaF-ZrF4 (53-47 mole %)+ It would epply
perhaps to o melt cantaining about 30 mole % ZrF4.
preliminary value of 180 dynes/em has been ob-
tained for the surface tension of pure molten UF
at 1050°C. This work was initiated in an effort to
determine the effect of small amounts of uo,
on the wetting properties of UF ,. Data obtained
by Cohen and Jones29 for Nc:F-ZrF4-UF4 (50-46-4
mole %) with a tensiometer are presented for
comparison in Table 2.3.6. These tables summarize
all the surface tension data on fuel mixtures and
their components that are available thus far.

 

6. G. Inghram and W. A. Chupka, Chemical Bonds
by Thermodynamic Methods, paper No. 40 presented at
ffi/e 4th Annual Meeting of the American Society for
Testing Materials, Committee E-14 on Mass Spectrometry,
Cincinnati, May 28, 1956.

7w, K. Stromquist and R. M, Boarts, Effect of Wetting
on Heat Transfer Charactleristics of Liquid Metals.
Third Quarterly Report, OR0-60 (Jan. 31, 1952).

]BJ. W. Taylor, J. Nuclear Energy 2, 15 (1955).

1€, W. Miles, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, p 118.

20, |, Cohen and T. N. Jones, ANP Quar. Prog. Rep.
June 10, 1953, ORNL-1556, p 88.

127
ANP PROJECT PROGRESS REPORT

Table 2.3.6. Data Obtained by Cohen and Jones2® on
the Surface Tension of NoF-ZrF4-UF4 (50-46-4 Mole %)

at Various Temperatures

 

 

Temperature Surface Tension
(°C) (dynes/em)
530 157
630 132
730 115

 

In connection with these experiments, it should
be noted that surface tension data are almost
valueless without a knowledge of the density
of the liquid. Miles!? developed a method by
which both the density and the surface tension can
be obtained from the measurements involved in
the maximum-bubble-pressure technique. While
theoretically the density of a sessile drop can be
obtained by weighing the drop and calculating its
volume from its surface profile, in practice the
method is subject to extremely large errors.
Preliminary pycnometric density data on molten
UF, and dilatometric measurements of solid UF
have yielded the interesting result that UF , melts
with negligible expansion, if any.

128

Much of the knowledge of fused fluoride salts
gathered in the past four years is still of a
qualitative nature. During the course of these
experiments, for example, it has become apparent
that further information on the wetting behavior of
fluoride salts is needed. The sessile-drop experi-
ments on metal supporting surfaces and the
maximum-bubble-pressure experiments have shown
that truly clean hydrogen-fired surfaces are com-
pletely wetted by molten NaF-ZrF, but that, if
thin oxide films are allowed to form on the metal
surface after hydrogen-firing, the films will prevent
wetting until they are dissolved by the molten
salt. These observations are difficult to reconcile
with the lack of evidence of wetting in fuel
purification vessels, where it has been consistently
observed that the purification pots drain cleanly
and that no residual film of salt (other than
splattered droplets) is left on the walls. Visual
observation studies2! have tended to confirm this
behavior, since the salt meniscus appears to be
definitely convex, especially in melts containing

Ber.

 

21 1. Sheil, private communication to S. Langer.
 

2.4,
G. J. Nessle

PREPARATION OF FLUORIDE COMPOUNDS
B. J. Sturm L. G, Overholser

Compounds not available commercially have
been synthesized for use in ANP chemistry
studies. Of particular interest were fluorides of
the metals being considered as container materials
for various fuels and the products of their reactions
with constituents of proposed fuels. Many of
these compounds had not been reported in the
chemical literature. Corrosion products found in
fuels have been identified by comparison of their
x-ray diffraction patterns and optical crystal-
lographic properties with those of synthesized
compounds. Fluorides of the fission products
have been prepared for use in the study of their
effects on the fuel mixtures of interest.

AlkalieMetal Fluochromates

A series of bright green chromium fluoride
complexes of the type M:';Cr”'F6 were prepared
in which M! stands for an alkali metal or ammonium
ion.? The purpose of preparing the alkali metal
fluochromates was to obtain their crystallographic
properties for comparison with those of unidentified
phases.

Potassium hexafluochromate(lll), K3CrF6, was
prepared by slowly heating to 900°C a mixture of
3 moles of potassium bifluoride with one of
hydrated chromic fluoride. Substitution of sodium
bifluoride for the potassium bifluoride yielded
NO3CrF6. Heating hydrated chromic fluoride with
an excess of ammonium bifluoride to 250°C formed
(NH,),CrF,. The compounds Li CrF, Rb,CrF,
and Cs CrF, were then obtained by reacting the
alkali metal fluoride with either (NH,),CrF  or a
mixture of ammonium bifluoride unj hygrcfled
chromic fluoride.

Fusion of Li:;(:rl:6 with Na,CrF . gave the com-
pound LizNoCrFé. The ternary system K,CrF -
Li3CrF6-No3CrF6 has a solid—solid-solution
area bounded by the compounds KzNoCrFé,
K3No3(CrF6)2, and K2LiCrF6. The compositions
of corrosion products that result from the action of

NcaF-KF-LiF-UF4 or NaF-KF-LiF on chromiuvm-

 

g, J. Sturm and L. G. Overholiser, ANP Quar. Prog.
Rep. Dec. 10, 1952, ORNL.-1439, p 115.

PERIOD ENDING DECEMBER 31, 1957

PRODUCTION OF PURIFIED MIXTURES

J. P. Blakely

containing alloys have always been within this
solid-solution area.? The compound Rb:,’CrF6
was identified as the corrosion product that results
from the attack of a chromium-containing alloy by
an RbF-LiF composition.3

Other Double Fluorides

A new series of compounds of the type M”ZrF6
were formed by fusing ZrF, with fluorides
of divalent transition metals. The compounds
CrZrFé, ManFé, FeZrF,, CoZrF, NinFé, and
ZnZrF, have been prepared.* The compounds
CquFé, CdZrF,, and PbZrF, probably exist,
but the products of fusions corresponding to these
compositions have not yet been adequately studied.
The compound NinF6 was identified as the
corrosion product that results from the action on
a nickel container by the vapors resulting from the
fusion of ammonium bifluoride with a fluozirconate
composition.

Fusions of alkali-metal fluorides with transition-
metal fluorides give a series of compounds.
Except for lithium fluoride, which does not combine,
the alakli metal fluorides form 1:1 compounds with
manganese, cobalt, and nickel difluorides.5 This
is probably also true with chromous fluoride, but
the results are confused by the contamination of
the preparations by chromic-containing phases as
a result of disproportionation of some of the
chromous fluoride.  The rubidium and cesium
compounds were unknown before this work was
begun, and very little information was available on
the physical properties of the rest of these
compounds.

Simple Inorganic Fluorides

Many different anhydrous fluorides were prepared.
Moisture was removed from lithium, sodium, and
potassium fluorides by heating in vacuum to

 

8. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. Sept, 10, 1952, ORNL-1375, p 122,

38. J. Sturm and E. E. Ketchen, ANP Quar. Prog.
Rep. Sept. 10, 1954, ORNL-1771, p 67.

48, 1. Sturm, ANP Quar, Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 91

3B. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. June 10, 1953, ORNL-1556, p 48.

129
ANP PROJECT PROGRESS REPORT

around 500°C. Strontium, cadmium, cerium,
praseodymium, neodymium, and samarium fluorides
were precipitated by the addition of hydrofluoric
acid to aqueous solutions of their salts.6=8 Oven

drying at around 200°C vyielded essentially
anhydrous material.

Yttrium fluoride, precipitated from aqueous
solution, was treated with anhydrous hydrogen

fluoride at an elevated
freedom from moisture. The hydrated ferric,
nickelous, cuprous, and zinc fluorides were
dehydrated by treating them in a similar manner.”
Twenty hours of hydrofluorination followed by 4 hr
of flushing with helium reduced the boron content
of cupric fluoride to below 2 ppb, the spectroscopic
limit of detection. The anhydrous fluorides
prepared by treatment with hydrogen fluoride gas
were usually flushed with purified helium gas to
remove adsorbed hydrogen fluoride. The hydrated
aluminum, chromic, and nickelous fluorides were
reacted with ammonium fluoride or bifluoride at
around 200°C to form the ammonium fluocomplex,
which was decomposed at a higher temperature to
form the anhydrous fluoride.®

Hydrofluorination of anhydrous ferric chloride at
around 300°C and chromic chloride at around 600°C
converted the chlorides to fluorides.4:? The
hydrated nickelous, ferrous, and cobaltous chlorides
were slowly heated to around 500°C during the
passage of hydrogen fluoride to convert them
to the anhydrous fluorides.7+10 Silver carbonate
was similarly converted into the fluoride at around
150°C, and ferrous oxalate was converted at
around 600°C.!! The tetrafluorides of thorium and
uranium were freed of oxide and moisture by
hydrofluorination at around 700°C.%? Vanadium
trifluoride was prepared by condensing hydrogen
fluoride onto vanadium trioxide powder, heating to

remove water and excess hydrogen fluoride, and
rehydrofluorinating at 200°C. 10

temperature to ensure

 

®E. E. Ketchen and B. J. Sturm, ANP Quar. Prog.
Rep. June 10, 1955, ORNL-1896, p 76.

7B. J. Sturm, ANP Quar. Prog. Rep. Dec. 10, 1953,
ORNL-1649, p 62.

83, J. Sturm, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 100.

8. J. Sturm and E. E. Ketchen, ANP Quar. Prog. Rep.
March 10, 1955, ORNL-1864, p 66.

105, 4. Sturm and L. G. Overholser, ANP Quar, Prog.
Rep. June 30, 1957, ORNL-2340, p 164.

11|, G. Overholser and B, J. Sturm, ANP Quar. Prog.
Rep. Sept. 10, 1953, ORNL-1609, p 70.

130

Chromous Fluoride

High-purity chromous fluoride was needed for
studying the removal of chromium from container
materials by fluoride fuels. Reaction of powdered
chromium metal with hydrogen fluoride gas at
about 800°C was found to be a poor method of
preparing chromous fluoride, because a coating
of the fluoride formed on the metal and prevented
further reaction with the hydrogen fluoride.
Grinding of the mixture exposed new metal surface
to react with hydrogen fluoride, but three successive
treatments failed to react all the metal. The
product was also contaminated with chromic
fluoride that resulted from the reaction

2CrF, + 2HF—>2CrF, + H,

Hydrofluorination of chromous chloride at about
700°C vyielded chromous fluoride that was con-
taminated by chromic fluoride, because of oxidation
by hydrogen fluoride, and traces of unreacted
chloride.'?2 The chromic fluoride could be reduced
by hydrogen.

of chromous fluoride were
prepared by hydrogen reduction of chromic fluoride.
The chromic fluoride was usually made by thermally
decomposing ammonium fluochromate, which was
made by fusing ammonium bifluoride with hydrated
chromic fluoride.  Usually the preparation of
chromic fluoride and its reduction by hydrogen were
combined into a single step in which ammonium
fluochromate was placed in a nickel reaction
vessel and the temperature was slowly raised to
700°C while o flow of hydrogen was maintained.
Chromous fluoride prepared in this manner was
usually contaminated with some opaque material,
This was especially true of the product near the
walls of the nickel reaction vessel. The chromous
fluoride was further purified by fusing at 1100°C,
recrystallizing, and manually separating the clear
crystals from the opaque residue,

A few kilograms

At first it was assumed that the opaque material
resulted from reduction of some of the chromous
fluoride to the free metal. Free metal could not be
found by x-ray diffraction analysis, but gamma
nitride was identified in the opaque
The following reactions are proposed

chromium
material.,

 

]28. J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep, Dec. 31, 1956, ORNL-2221, p 153.

 
 

to explain the formation of chromium nitride:
(NH4)3CrF6—> 3NH, F + CrF,
2NH4F—> NH,HF, + NH,
CrF3 + NH3—-—‘> CeN + 3HF

The formation of chromium nitride by the reaction
of ammonia with chromic chloride was reported by
Schwarzkopf and Kieffer.'3 |t is reasonable to
expect chromic fluoride and ammonia to react in
the same way.

If chromic fluoride was not completely reduced
to chromous fluoride, a phase was formed which
had a composition intermediate between the compo-
sitions of chromic and chromous fluoride. The
exact composition of this phase has not been
established, but its optical properties and x-ray
diffraction pattern have been determined.

SUMMARY OF EXPERIENCE IN THE
PREPARATION OF FUEL FOR
THE AIRCRAFT REACTOR EXPERIMENT

C. R. Croft J. E. Eorgan
F. A. Doss J. Truitt

The nuclear poisons, such as boron, cadmium,
and certain of the rare earth elements, are not
common contaminants of enriched uranium, sodium,
and zirconium fluorides. Hafnium, which invariably
contaminates zirconium and its compounds, can
be reduced in concentration to acceptable limits
(50 to 100 parts of Hf per million parts of Zr) on a
commercial scale. Since the initial concentrations
of these poisons were low and the parasitic capture
of neutrons in the core structure would be con-
siderable, purification processes were designed
to minimize corrosion and to ensure the absence of
slightly soluble oxides and oxyfivorides rather
than to improve neutron economy.

The best-quality UF , available commonly con-
tains small quantities of UO,, UF,, and UO,F..
The commercially available ZrF,, even when
prepared by hydrofluorination of ZrCl, and further
refined by sublimation at low pressure, contains
up to 2% excess oxides and oxyfluorides, up to
0.2% Cl=, and Fe**t to the extent of a few
hundred parts per million. Both these tetrafluorides
are readily hydrolyzed to oxides and oxyfluorides

 

IBP. Schwarzkopf and R. Kieffer, Refractory Hard
Metals, Macmillan, New York, 1953.

PERIOD ENDING DECEMBER 31, 1957

at elevated temperatures. Commercial NaF con-
tains some sulfur, mostly as sulfate, and small
amounts of water,

The fuel solvent, NaF-ZrF4, and the fuel concen-
trate, 2NaF-UF ,, were prepared separately for the
ARE, as required by the startup procedure. For
testing of components and for heat transfer and
metallurgical research, however, large quantities of
the fuel mixture Nc:F-ZrF4-UF4 were prepared
in various compositions. Such bulk production was
carried out on a 250-lb-batch scale, but con-
siderable production of special compositions in
small batches on a 5-, 10-, and 50-1b experimental
scale was required for many specific test programs.
In general the fluoride mixtures were purified by
the process described below, with only minor
modifications for the small batches when specific
requirements had to be met for specialized
experiments.

Purification of Fuel Solvent

The equimolar NuF-ZrF4 mixture required for
the ARE was purified in 250-1b batches. The
reaction vessel, in which the chemical processing
was accomplished, and the receiver vessel, into
which the purified mixture was ultimately trans-
ferred, were 13-in.-dia, vertical cylindrical con-
tainers, 48 and 24 in. long, respectively, that were
fabricated of A-nickel by the Heliarc welding of
3/a-in. plate. The top of the reactor vessel was
pierced by a charging port consisting of a 4-in.-dia
pipe which was capped well above the heated zone
by a Teflon-gasketed flange. The tops of both the
receiver and the reaction vessel were pierced by
short risers which terminated in Swagelok fittings
through which gas lines, thermowells, etc., of
3/a-in. nickel tubing could be introduced. A transfer
line of 3/8-in. nickel tubing terminated near the
bottom of the reactor vessel and near the top of
the receiver; entry of this tube was effected through
copper-gasketed flanges on l-in.-dia tubes which
pierced the tops of both vessels. This transfer
line contained a filter of Micro Metallic sintered
nickel and a sampler which collected a specimen
of the liquid during transfer. Through one of the
risers in the receiver a ¥-in. nickel tube extended
to the receiver bottom; this tube, which was sealed
outside the vessel, ultimately served as the means
for transfer of the purified melt to the ARE., A
schematic diagram of the reaction vessel—receiver
assembly is shown in Fig. 2.4.1.

131

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 20157

TEFLON GASKETED FLANGE

  
   

4-in.~DIA CHARGING PORT

3/g—in~NICKEL
TRANSFER LINE

FILTER

 

SAMPLER

 

 

 

 

 

\

REACTION VESSEL

RECEIVER VESSEL—U

Fig. 2.4.1. Apporatus for Purification of Fluoride

 

 

 

 

Mixtures.

This assembly was connected, through traps for
collecting sublimed ZrF,, to a manifold through
which He, H,, HF, or vacuum could be supplied
to either vessel. By a combination of large tube
furnaces, resistance heaters, and lagging, sections
of the apparatus could be brought, independently,
to controlled temperatures in excess of 800°C,

The raw materials in batches of the proper
composition were blended and charged into the
reaction vessel. The material was melted and
heated to 700°C under an atmosphere of anhydrous
HF to remove H,O with a minimum of hydrolysis.
The HF was then replaced with H, for a period of
1 hr, during which the temperature was raised to
800°C, to reduce U3* and U¢* to U4*, sulfur
compounds to S$—=, and extraneous oxidants
(Fe*** for example) to lower valence states.

132

The hydrogen and all subsequent reagent gases
were fed at about 3 liters/min to the reaction
vessel through the receiver and the transfer line
and, accordingly, bubbled up through the molten
charge, The hydrogen was then replaced by
anhydrous HF, which served, during a 2- to 3-hr
period at 800°C, to volatilize H,$ and HCl 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
24- to 30-hr treatment at 800°C with H, sufficed
to reduce the NiF, and the contained FeF, to
insoluble metals,

At the conclusion of the purification treatment,
a pressure of helium above the salt in the reactor
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
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.

Purification of Fuel Concentrate

The 2NaF-UF , mixtures were prepared by a very
similar processing schedule in similar but smaller
equipment. In order to avoid the possibility of
criticality accidents, the fuel concentrate was
prepared by using reactors and receivers 4 in. in
diameter to limit batch sizes to about 30 |b.

Storage and Ultimate Transfer

The receiver vessels containing either fuel
solvent or fuel concentrate were connected to
manifolds that supplied pure helium at slightly
more than atmospheric pressure and were stored
in this manner at room temperature until used.

Transfer of the material for ultimate use was
accomplished by forcing the remelted charge under
helium pressure through the transfer line built
into the receiver vessel. In this simple fashion,
virtually quantitative transfer of the molten
material could be accomplished. Transfer of up
to 250 |b of solvent usually left less than 0.5 Ib
of solvent in the container.

 

 
PILOT.SCALE PURIFICATION OPERATIONS
C. R. Croft J. Truitt

In the early stages of the ANP program, extensive
small-scale experimentation was carried out in
order to develop a suitable process and a flexible
processing unit for the purification of fluoride
mixtures. Small-scale units of 5 and 10-lb capacity
were installed, and later, as demands grew, 50-1b
units were added. Finally, six small units and
three 50-1b units were in continuous use. Although
bulk production was taken over by the 250-1b
capacity units, demands for small, special batches
continued to increase. During the active life of
the ANP fluoride production program, 653 batches
totaling approximately 11,500 Ib of various fluoride
compositions for use in phase equilibrium studies,
physical property measurements, smalil-scale
corrosion tests, and other experiments were
processed in these pilot-scale units.

PERIOD ENDING DECEMBER 31, 1957

PRODUCTION-SCALE OPERATIONS
J. E. Eorgan

The ARE and the large-scale component testing
program, as stated above, created demands for
processed fluorides far beyond the capabilities
of the small pilot-scale units, and therefore the
250-Ib production units were installed to process
fairly standard compositions in large quantities.
The first production operation of this facility was
the processing of the approximately 3300 Ib of
fuel solvent required for the ARE. Analyses of
the 14 batches of fuel solvent processed for use in
the ARE, as can be seen from the results given in
Table 2.4.1, showed the purity of the finished
material to be very satisfactory. Since this was
the first operation of the production-scale units,
the performance of the processing units during
this period served as a basis for future production
schedules,

Table 2.4.1. Results of Analyses of ARE Fuel Solvent

 

Chemical Analysis

Spectrographic

 

 

 

 

Batch Major Constituents (%) Impurities (ppm) Anclysis

e _om

Zr F Na Fe Cr Ni HE B

1 43.3 45.0 976 30 <10 <10 240 0.2
2 43.9 452 11.0 45 <20 <20 55 0.2
3 43.9 45.5 10.6 60 <10 25 100 0.9
4 42.0 4447 10.8 55 <20 45 80 0.9
5 44.1 44.9 11.0 90 <20 <20 60 0.6
6 4346 4447 10.5 35 <20 30 85 0.5
7 4346 45.2 10.4 35 <20 30 70 0.5
8 434 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
11 437 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
process and facility, with regard to raw material
usage vs quantity of final product, was very good.
A material balance for the fuel-solvent production
operation is given below:

Weight (Ib)
Raw fluorides used (NaF and ZrF4) 3412
Net fluorides produced 3348
Net samples 5
Net heels left in reactor vessels 56
Total produced 3409
Net loss 3

On the basis of usable product, a production
efficiency of 98% was obtained.

The production facility was normally run on a
five-day-week, 24-hr-shift basis during the early
stages of the component testing program. When
demands from off-area consumers (such as Pratt &
Whitney Aircraft) increased, the operation was
placed on a continuous seven-day-week schedule.
Maximum  production on the seven-day-week
schedule was reached at 1000 Ib per week.
Production on a five-day-week schedule averaged
about 600 Ib per week.

134

During the production operation of this facility,
322 batches totaling 80,000 Ib of processed
fluorides were produced. Since production usually
exceeded demand, the facility operated on an
average of eight months per year.

SERVICE OPERATIONS

F. A. Doss J. Truitt
J. E. Eorgan N. V. Smith

In conjunction with the accelerated component
testing program, and the resulting acceleration in
fluoride production, a facility was established for
the efficient and accurate handling of the large
amounts of fluorides and liquid metals used. The
250-ib batches could, in some cases, be used
directly, but ordinarily these batches had to be
reduced to quantities ranging from 1 to 50 Ib to
fit specific test programs. Complete analytical
records and material balances were kept to avoid
confusion in the interpretation of component test
results. Responsibility was taken for the loading,
sampling, and unloading of charge materials from
component test equipment, including the high-
temperature critical assemblies. During the course
of the program, approximately 70,000 |b of
processed fluorides and 35,000 |b of liquid metals
were dispensed. In addition, some 200 storage
cans used in dispensing processed fluorides to
local and off-area consumers were maintained,
 

DETECTION OF TRACES OF NaK IN AR
A. S. Meyer, Jr. J. P. Young

A study of sensitive methods for the detection
of traces of alkali metals in air was carried out
in order to provide a monitor for the exhaust air
from the ART NaK-to-air radiators. The sensitivity
requirement for a suitable detector was established
on the basis of a tentative limit of a NaK leak
which would correspond to the introduction of a
total volume of 6 cm3 of NaK per hour into the
exhaust air stream. On the basis of the design
flow rate of air in the radiator ducts, the instrument
was required to respond to a concentration of NaK
in air of 10 ppb by weight. Since it was also
assumed that the inlet air could be contaminated
with alkali metal compounds in concentrations
comparable with those introduced by the radiator
leaks, it was considered necessary that the
detector be designed to monitor both the inlet
and exhaust streams and to respond to an increment
of 10 ppb of NaK across the radiators. Additional
specifications for a suitable detector included a
rapid response time and a high degree of reliability.

Methods of detection which were evaluated during
the feasibility studies included chemical analyses,
photometric smoke detection, emission spectro-
graphic analysis, flame photometric analysis,
surface-ionization detection, and resonance-
radiation measurement. |In addition, some con-
sideration was given to a technique in which a
material that could be readily detected by existing
conventional methods of analysis would be added
to the NaK. Radiometric techniques were not
considered.

Existing chemical methods, emission spectro-
graphic techniques, and smoke detection do not
offer sufficient sensitivity for the detection of
NaK at the 10-ppb concentration level. Although
mass spectrographic techniques offer adequate
sensitivity, they do not appear to be applicable
to the proposed system because of the problems
concerned in introducing representative, continuous
samples of the alkali metal oxide suspensions
through capillary leaks into the ionization chambers
of the spectrometer. For the proposed differential
measurement it would be necessary to maintain
capillary leaks of equal rates for both intet and
exhaust air.

PERIOD ENDING DECEMBER 31, 1957

2.5, ANALYTICAL CHEMISTRY
J. C. White

Although flame photometric measurements may
be extended to ultimate sensitivities in the desired
concentration range, such measurements might
not be reliable because of the problem of main-
taining a constant flame source by remote control.
Experience with a Beckman flame photometer
during operation of the ARE indicated that this
problem imposes serious limitations to the
application of this instrument, Additional compli-
cations can be predicted in the application of this
technique to the proposed detector in that it would
be necessary to maintain two flames at identical
temperatures and to meter samples of the inlet
and exhaust gases to these flames at constant
rates.

The detection of alkali metals by means of a
surface-ionization detector is based on the
measurement of the current that is carried by
positive ions which are formed when atoms of the
alkali metals are brought into contact with a
heated filament. This detector has been applied to
the detection of sodium and NaK in helium.!
Concentrations of 2 ppm of sodium and 9 ppm of
NaK have been detected in this manner. |t was
suggested? that the surface-ionization detector
might be used for air samples if an iridium filament
were used. Under these conditions, detection
limits of 1 to 10 ppm of NaK could be achieved.
The calibration curves reported for the measurement
of alkali metals in helium (current vs saturation
temperature) are linear over only a very limited
range of concentration, however, and this method
would therefore not be readily applicable to the
proposed differential measurement of concentration.
This method was given some further consideration
since it was thought that, because of the small
size of the sensing elements, numerous elements
could be placed in strategic locations to detect
the larger concentrations of alkali metals which
would be present in the immediate vicinity of the
leak. The detection requires the measurement of
relatively small direct currents, however, and the
instrumentation for such a multiple installiation
of detectors might become prohibitive.

 

TH. s. McKown, A Detector for Sodium Vapor — A
Surface lonization Sodium Detector, K-996 (Feb, 16,
1953).

2J. R. Sites, private communication to A. S. Meyer, Jr.

135
ANP PROJECT PROGRESS REPORT

The proposed resonance-absorption method is
based on the measurement of the attenuation of a
beam of sodium D light (589 mpu)on passage through
sodium metal vapor. This method offers extreme
sensitivity for the detection of sodium. Since the
absorption coefficient for sodium vapor at the
center of the resonance line is approximately
10-'2 N, where N is the density of sodium in
atoms per square centimeter, it was calculated
that a concentration of 0.01 ppm of atomic sodium
in air would be sufficient to reduce by 50% the
intensity of a beam of sodium radiation that had
traversed a 50-cm optical path. |nasmuch as
compounds of sodium such as Na,O do not absorb
the resonance radiation, it is necessary to main-
tain the sample at a temperature sufficient to
cause thermal dissociation of Na,O to atomic
sodium in order to obtain a mecsurogle absorption
of the resonance line.

On the basis of thermodynamic constants it was
calculated that a concentration of Na,O of 10 ppb
could be quantitatively dissociated at temperatures
as low as 800°C. Dissociation of comparable
concentrations of K, O, which can also be detected
by the resonance-absorption method, is complete at
approximately 600°C.3

A test model of an instrument designed to com-
pare the absorption of samples of the inlet and
exhaust air from the radiators was constructed.
Detailed descriptions, including a photograph of
the instrument, were given previously.4 In this
instrument a beam of sodium D light is divided
into two pulsed beams of equal intensity by a
90-cycle mechanical chopper. After the beams
are passed through heated absorption cells which
contain samples of the inlet and exhaust air, the
transmitted beams are recombined at a photo-
multiplier tube. A 90-cycle a-c signal, which is
proportional to the difference in absorption between
the inlet and exhaust air, is generated by the
photomultiplier.

This instrument was tested by passing, through
one of the tubes, samples of sodium in air which
were prepared by mixing with the air stream helium
that had been saturated with sodium vapor. A
detection limit of approximately 100 ppb was
noted. With minor modifications of the instrument,

 

3A. 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.
Rep. March 31, 1957, ORNL-2274, p 137.

136

which included the substitution of a Polaroid
chopper and refinements of the 90-cycle tuned
amplifier, the limit of detection was lowered to
60ppb.3 Additional modifications of the instrument
were made to eliminate the continuous radiation of
the heated cell walls from the photomuitiplier
tube and to provide means of exact alignment of
the optical system, Details of these modifications
were given previously.? Experimental work on the
project was discontinued before the improved
instrument could be tested.

It is believed that the studies which have been
carried out have demonstrated the applicability of
this method to the detection of the radiator leaks.
The test model responded almost instantaneously
to small increments in the concentration of the
sodium in the air samples, and with refinements
in the optical and electronic components its
sensitivity could readily be increased to provide
detection at the required concentration level. The
double-beam system is known to offer high stability
and reliability.  With minor modifications, the
instrument could be adapted to the detection of
other alkali metals in air.

On the basis of these experiments, design
criteria for a working model of the instrument were
prepared.? Improvements to the test model would
include: @ rigidly mounted optical system to
provide a multiple pass of the light beam through
the sample; an electrodeless source lamp excited
by radio frequencies to provide extended lamp life
and a sharper line source; multilayer interference
filters to isolate the resonance line more ef-
fectively; and a phase-sensitive amplifier to
discriminate completely against frequencies of
other than 90 cycles.

Although the feasibility of the proposed detector
appears to have been demonstrated, much additional
experimental work would be required to establish
sampling techniques. Questions to be resolved
include the selection of the best material for the
construction of sampling ports and transfer lines,
the determination of the optimum temperature and
flow rates to minimize deposition of the alkali
metals on the walls of the tubing, and the required
recovery time of the assembly after exposure to
the large concentrations of alkali metals which
would be introduced by a major leak. A com-
parison of the transfer characteristics of K2O and

 

SA. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog,
Rep. Sept. 30, 1957, ORNL-2387, p 148.

 

 
No,O should be undertaken in order to provide
criteria for a choice of the element which could be
monitored most efficiently.

A major test facility would be required in order
to provide a means of simulating a NaK leak in an
ART radiator. This facility could be used to
establish criteria for the selection of sampling
points which would assure complete mixing of the
air for representative sampling and to determine
the reliability of the instrument in the detection of
small increments in the concentration of the
alkali metal in the presence of large background
concentrations.

Although the tests could be carried out with less
extensive equipment if the alkali metal could be
introduced to the air stream at an infinitesimal
rate as a dilute solution, preliminary experiments®
have indicated that the physical properties of the
samples vary widely with the method of prepa-
ration. Accordingly, to carry out definitive tests
of sampling procedures, the direct addition of NaK
to the heated air stream would be essential, Tests
carried out with water have indicated that for
minimal rates of injection, as a jet, high pressures
and minute orifices are required. Inorder to inject
NaK at a rate consistent with a test facility of
reasonable proportions, pressures in excess of
1000 psi and orifices 0.0001 to 0.0003 in. in
diameter would be required. Orifices of this
diameter can be drilled by anodic dissolution with
a virtual electrode.” Orifices less than 0.0005 in.
in diameter have been fabricated in nickel by this
technique.  Zirconium orifices would be more
satisfactory, but electrolytic drilling of this
material has not been successful.

These sampling difficulties could be largely
resolved by adding to the NaK a volatile material
which could be detected in small concentrations.

Mercury has been suggested as an ideal additive, .

since it could be readily transferred and can be
detected by resonance-absorption methods at
operating temperatures, The addition of mercury
would be contingent on the evaluation of its effect
on the compatibility of NaK with structural ma-
terials and its effect on nuclear properties.

 

6A. S. Meyer, Jr., and J. P. Young, ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 164.

7 A. Uhlir, Jr., Rev, Sci. Instr. 26, 965 (1955).

 

PERIOD ENDING DECEMBER 31, 1957

DETERMINATION OF OXYGEN IN
FLUORIDE SALTS

A. S. Meyer, Jr. G. Goldberg

The gas-measuring system of the apparatus8
for the determination of oxygen in fluoride salts
with KBrF, was calibrated by measuring the
displacement of the manometer levels produced by
the addition of measured volumes of oxygen.
Discrete additions of oxygen were made by means
of the apparatus shown in Fig. 2.5.1. By ap-
propriate rotation of the three-way stopcock (SB)
the volume of oxygen in the sealed arm of the
stopcock (0.281 ecm3) is connected with the
Toepler pump. The oxygen is transferred to the
measuring system by means of the Toepler pump
in a manner analogous to its transfer from the
reaction system when determinations are performed.
The coefficient of variation of the calibration
measurements was less than 2%.

The accuracy of the calibration was confirmed by
determining the oxygen in 10-mg samples of Y203.
Comparable precision was obtained in these
measurements. Typical results of replicate
determinations by this procedure are given in

Table 2.5.1.

The range of the method is from a minimum of
5 pg of oxygen, as measured with the oil manometer,
to a maximum of 100 mg, as measured by the
mercury manometer. For quantities of oxygen in
excess of 100 pg, the precision of the method would
appear to be less than 5%.

 

8A. S. Meyer, Jr., and G. Goldberg, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 150.

Table 2.5.1. Determination of Oxygen in Samples
of Y2 O, and YF,

 

Oxygen (mg)

 

 

Sample Sample Weight
Composition (g) Theoretical Determined
Y203 0.0103 2.19 2.16
0.0102 2.17 2.18
YF, 1.00 0.58
1.00 0.59

 

137
ANP PROJECT PROGRESS REPORT

VACUUM GAGE
n

UNCLASSIFIED

ORNL-LR-DWG 27057

 

  
 
   

|
H oL
VACUUM 54%/3 MANOMETER
ii
|- H e
H g
./' ==

[
LN

&

 

 

DIFFERENTIAL
MANOMETER

-

 

 

Cloofoadiooroed g

‘7/

 

 

 

 

MEASURING
MANOMETER

TOEPLER PUMP

CALIBRATED VOLUME

AL 8

=

D S0
2

   

7.,

    

(=<t TO REACTION APPARATUS
s

M |
> PYREX TRAP

5 = HIGH-VACUUM STOP COCKS

1/1/‘ I l:h

 

 

Ciononayfoo o] |

 

 

 

OXYGEN-METERING
MANOMETER

Fig. 2.5.1. Constant-Volume Manometer Calibrating Apparotus.

138

 

 
Part 3
METALLURGY
W. D. Manly

 
3.1. NICKEL-MOLYBDENUM ALLOY DEVELOPMENT STUDIES

MATERIAL DEVELOPMENT

H. Inouye T. K. Roche
J. H. Coobs

The developmental work on nickel-molybdenum-
base alloys for use as reactor structural materials
in contact with the fused salt fuel NaF-KF-LiF-UF ,
(11.2-41-45.3-2,5 mole %, fuel 107) at normal
operating temperatures of 1650°F and above has
been terminated. Much of the technology and ex-
perience gained in this investigation will be of
value to the Molten Salt Reactor program.

Status of Devel opment

Based on the requirements for an ideal container
material for fuel 107, the experimental nickel-
molybdenum-base alloys have shown exceptional
promise. The various alloys which have been
studied fall into three general categories: solution-
hardening, precipitation-hardening, and dispersion-
hardening alloys. The solution-hardening alloys
are represented by alloys INOR-1 and -8; the pre-
cipitation-hardening alloys by INOR-3, -4, -6, and
-7; and the dispersion-hardening alloys by INOR-5
and -9. Preliminary data on the properties of the
INOR alloys 1 through 8 were presented previously,
along with the composition ranges of each alloy
studied.! The data indicate that, of the three
categories of alloys, the solid solution-hardening
alloys most nearly meet the requirements. On
this basis, the INOR-8 alloy was selected as
having the best combination of properties, and
much of the effort has been directed toward estab-
lishing the final composition and preducing com-
mercial quantities of the alloy. [t is felt, however,
that the full potential of alloys of the nickel-
molybdenum system definitely has not yet been
developed. For the immediate future, these alloys
appear to be the only structural materials which
show the possibility of bridging the gap between
Inconel and the refractory metals niobium and
molybdenum.

The most attractive feature of the nickel-base
alloys containing 15 to 20% molybdenum as the
principal alloying element is their versatility.
From the studies of numerous alloy combinations

 

Th. lnouye and T. K. Roche, ANP Quar. Prog. Rep.
March 31, 1957, ORNL-2274, p 177.

and types, it has been determined that a particular
property of interest can be improved, but usually
at the expense of some other property. Inasmuch
as the property variations are derived largely from
additions of third, fourth, or fifth alioying elements,
the principal properties of importance such as
strength and corrosion resistance can be balanced
to give an alloy composition which is best suited
for a particular application. As the severity of
the service conditions is increased (for example,
higher temperatures and longer operating times), the
freedom of modification becomes narrower, and
thus many of the alloys are eliminated for various
reasons. In short, for every temperature or appli-
cation, there appears to be an optimum alloy
composition.

For 1000 hr of operation, the alloys, as presently
developed, appear to permit peak temperatures of
1650 to 1700°F at the metal-salt interface. This
limitation has been tentatively set by the rapid
increase of the corrosion tendency of the alloys
containing chromium and the deterioration of
strength above these temperatures.

The alloys that have been studied have been
evaluated in terms of fabricability, mechanical
properties, corrosion resistance, weldability, and
aging characteristics. The results of these in-
vestigations are summarized below.

Fabricability. — Alloys which are solid solutions
at the forging temperatures should present no un-
usual fabrication problems. Forging difficulties
have been encountered with large heats of the
INOR alloys, as reported previously,!*? but such
difficulties have been associated with poor melting
and forging techniques. It is anticipated that
further fabrication problems will be those normally
associated with developing reduction schedules for
new materials.

A serious limitation in the production of tubing
has been the tendency toward longitudinal splitting
of alloys that contain over 0.10% carbon. This
tendency has not been noted during hot-forging
operations or during cold rolling of plate and sheet
products.

 

2H. Inouye, T. K. Roche, ond J. H. Coobs, ANP Quar.
Prog. Rep. June 30, 1957, ORNL-2340, p 175.

141
ANP PROJECT PROGRESS REPORT

Mechanical Properties. — The mechanical prop-
erties of the nickel-molybdenum alloys are de-
scribed in detail in a subsequent section of this
chapter. |n general, the available data show that
an increase of up to 200°F in operating temperature
should be obtained by substituting INOR-8 for
Inconel. Thus, operation at temperatures between
1650 and 1700°F with fuel 107 appears to be
possible.

A correlation between microstructure and creep-
rupture characteristics showed that the micro-
structures with grain-boundary networks of fine
carbides had the best strength properties. Carbide
precipitation studies showed that the desired
microstructure could be obtained by using high
annealing temperatures (2300°F). Improvements
in the time to rupture and in ductility were made by
minor additions of malleableizing elements and
vacuum remelting of air melts.

Corrosion Resistance. - Within cértain limi-
tations, the corrosion-resistance characteristics of
the nickel-molybdenum alloys approach the optimum
for service in fused-salt, sodium, and air environ-
ments. The corrosion behavior of these alloys in
fused salts and sodium is described in detail in a
subsequent section of this chapter.

The oxidation rates of the nickel-molybdenum
alloys have been found to vary from excessive to
slight, depending upon the composition. The
element having the largest effect is chromium, and
its critical level at a temperature of 1800°F was

determined to be 6%. Alloys containing less than
6% chromium show about 4 mils of oxidation in
170 hr, while those containing more than 6% show
attack of less than 1 mil.

The extent of oxidation in 1000 hr at temperatures
as high as 1750°F has been tolerable for all
compositions studied. This conclusion is based
upon the fact that approximately 75 loops have
been operated without special coatings or pro-
tection. The probable maximum service temper-
atures in fuel 107, in sodium, and in air for alloys
of current interest are given in Table 3,1.1.

Weldability. — The status of the development of
techniques for welding these alloys has been re-
ported in detail in reports in this series. |t is
felt that the weldability will not be a serious
problem, since approximately 40 alloys of com-
positions representing various modifications of the
nickel-molybdenum base material have been welded
satisfactorily. Such welds have been in service
for periods of up to 1000 hr at temperatures as high
as 1850°F without evidence of deterioration. The
incidence of poor weldability and cracking ten-
dencies under restraint in various heats of the
alloys has shown good correlation with evidence
that those heats were not properly melted. The
results of high-temperature ductility tests have
been reported.?

 

3E, F. Nippes, W. F. Savage, and R. P. Meister,
Hot-Ductility of Nickel-Molybdenum Alloys, Rensselaer
Polytechnic Institute, Troy, N. Y., 1957.

Table 3.1.1. Maximum Permissible interface Temperatures of NickeleMolybdenum
Alloys in Air, in Sodium, and in Fuel 107 for 1000 hr of Operation*

 

Alloy

Composition

Alloy

Temperature of Environment (°F)

 

 

(wt %) Designation Fuel 107 Sodium Air
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

 

 
Aging Characteristics. —~ Of the three elements,
titanium, aluminum, and beryllium, which cause the
basic nickel-molybdenum alloy to respond to an
aging heat treatment, beryllium was found to be
the most potent. Alloys containing beryliium in
amounts as low as 0.25% exhibited strong aging
tendencies at 1300°F. The aging response of
alloys containing titanium and aluminum at 1300°F
was significant, while at 1500°F these alloys
either overaged in a short time or showed no effect.

Properties of INOR-8

It appears at this time that the alloy INOR-8,
which is based on the Ni-Mo-Cr-Fe system, offers
the best choice of properties for fulfilling the
requirements of a structural material compatible
with fused fluorides. The current optimum com-
position range of this alloy is: 15-18% Mo, 6-8%
Cr, 4-5% Fe, 0.04-0.08% C, 0.8% (max) Mn,
0.5% (max) Si, bal Ni. The properties of this
alloy were summarized and discussed previ-
ously.!+4

An investigation of the effect of varying com-
position on the mechanical properties of alloys of
the INOR-8 type is under way. Eighteen alloys
have been prepared and tested in creep rupture at a
stress of 10,000 psi and a temperature of 1500°F
in argon. Results of these tests, although in-
complete, show that (1) increasing the molybdenum
content is the most effective means of improving
the creep properties, (2) increasing the chromium
content from 5 to 7% without changing the molyb-
denum and iron content produces only a slight im-
provement in the creep properties, and (3) there is
no obvious correlation of the time to rupture of the
material with the total content of alloying elements
in the nickel base.

Properties of INOR-9

Alloys designated INOR-9 have been studied to
determine whether they may be suitable as possible
substitutes for INOR-8. The base composition of
these alloys is that of INOR-1 {a nickel-molybdenum
binary alloy) with niobium and carbon added for
strength.  Since the alloys of this group do not
contain elements that corrode in fuel 107, it is ex-
pected that a higher metal-salt interface temper-

 

44. Inouye, T. K. Roche, and J. H. Coobs, ANP Quar.
Prog. Rep. Sept. 30, 1957, ORNL-2387, p 161.

PERIOD ENDING DECEMBER 31, 1957

ature would be permissible for INOR-9 than for
INOR-8.

The heats of INOR-9 that are being investigated
have the nominal composition 12~19% Mo, 3-5% Nb,
2-5% Fe, 0.06-0.30% C, bal Ni. The results
obtained thus far show that (1) the maximum per-
missible niobium content for forgeability is about
7.5%, (2) the solid solubility of niobium up to
1700°F in the nickel-molybdenum base is 5%, (3)
the creep strength is about equal to that of INOR-8,
and (4) the oxidation rates of the INOR-9 alloys are
slightly higher than those of Hastelloy B.

Phase Diagram Studies

An investigation is being conducted at the
University of Tennessee (subcontract 582) for
the purpose of establishing phase relationships
and characteristic transformations in the binary
nickel-molybdenum and ternary nickel-molybdenum-
chromium systems. In the binary system, three
secondary phases have been identified and their
transformation temperatures have been determined,
as stated previously. !

Solubility limits of molybdenum plus chromium in
nickel are being determined, and identification of
the phases observed is to be undertaken. The
compositions studied range from 20 to 35% Mo and
3 to 15% Cr. Apparently the addition of chromium
suppresses the formation of the beta phase (Ni Mo},
inasmuch as it has not been observed in any of the
ternary compositions.

Stress-Rupture Testing by New England
Materials Laboratory

The production heats of INOR-1 through -6,
which were prepared by the International Nickel
Company, were stress-rupture tested at 1350,
1500, and 1650°F in air at the New England
Materials Laboratory. A final report is being pre-
pared in accordance with the terms of the con-
tract. The data obtained on stress-rupture strengths
of these alloys were summarized in a previous
report.4 :

Status of Production Heats

A purchase order was placed with the Inter-
national Nickel Company for the production of
4800-lb heats of six nickel-molybdenum-base
alloys so that experience might be gained with this

143

 
ANP PROJECT PROGRESS REPORT

type of material. The alloy compositions are

given below:

Alloy Nominal Compesition (wt %)
INOR-1 20 Mo—bal Ni

INOR-2 16 Mo-5 Cr-bal Ni

INOR-3 16 Mo—~1 Al-1.5 Ti-bal Ni

INOR-4 16 Mo—-2 Al=1,5 Ti=bal Ni

INOR-5 13 Mo-5 Cre2 W—2 Nb-0.1 C-bal Ni
INOR-6 16 Mo~5 Cr—1 Al=1.5 Ti—bal Ni

In general, billet conversion of these alloys has
been only moderately successful. Difficulties were
encountered with cracking during the rolling of the
titanium- and aluminum-bearing INCR alloys 4 and
6. INOR alloys 1, 2, and 5 were processed with
considerably more success. Strength and weld-
ability studies were conducted on the converted
products received at ORNL. Recently, after some
delay, the tube reducing of the tube shells was
completed, and these shells are presently being
redrawn to small-diameter seamless tubing. Cur-
rently, the International Nickel Company is engaged
in the conversion of the remaining billet stock to
items required for testing.

A 10,000-Ib air-melted heat of INOR-8 was
purchased from the Haynes Stellite Company to
supply the immediate material needs of the various
test programs. Casting and fabrication of the
alloy by hot and cold working to bar, plate, sheet,
and strip products was accomplished with little or
no difficulty. Various shapes of this material
have been received, including forged bar, hot-
rolled plates, hot-rolled sheet, hot-rolled bar, six
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-
contract 1067, and the first periodic report on
this developmental work was submitted to ORNL.
In summary, three heats of the specified INOR-8
composition were induction-melted in air. Diffi-
culties were encountered in forging castings of
these heats because of internal flaws formed during
cooling.  Various changes in mold design and
melting practice were made to overcome this, as
described in detail previously.?

A sound 3-in.-dia forging from the initial high-
carbon heat of INOR-8 prepared by Westinghouse
was received for evaluation. The material was
successfully worked by extrusion and by hot and

144

cold rolling, Specimens were prepared for creep-
rupture studies and for density and thermal ex-
pansion measurements. The density of the material,
as determined by weighing a machined specimen,
was found to be 8.795 g/em3. Thermal-expansion
measurements were made from room temperature to
1832°F with the cooperation of the Ceramics
Group of the ORNL Metallurgy Division. The data
obtained are presented in Fig. 3.1.1. The co-
efficients of linear thermal expansion, @, based on
the data of Fig. 3.1.1 are given below:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Temperature Average a
Range (°F) (ine/in.+:°F)
-5
x 10
212-752 0.75
7521112 0.73
11121832 1.02
212--1832 0.87
UNCLASSIFIED
ORNL—LR—-DWG 21328R
TEMPERATURE (°F)
392 752 12 1472 1832
0.01600
/
7
/
Il
0.01400 ; / e
4
,I
/I
0.01200 CORRECTED 7 o
TRUE EXPANSION /’ TRUE EXPANSION
FOR f=in. LENGTH ~_ L FOR —d
) 0.989-in. LENGTH
= /
£ 0.01000 b - / , /
= /1
z INSTRUMENT
g / CORRECTION /
Z 000800 7 ___CURVE7,
a
>
: J ]
I V
S / // /
’
< 000600 L AL / ,
A 7

 

/ /Qr— PLOTTED DIAL

| A A GAGE READINGS ||
ecos00 L~ FOR 0.989-in. LENGTH
|

000200 |— ——+——t— nor—8 — 1 -+ -
; | COMPOSITION (wt Ta):

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e Pd Mo Cr Fe c Mn Si Ni .
/- 4 16.90 6.86 4.2 044 084 023 BAL
o & L
0 200 400 600 800 1000 1200
TEMPERATURE (°C}
Fig. 3.1.1. Linear Thermal Expansion for 1l-in.
Length of INOR-8 as Determined with a Vacuum

Dilatometer Direct-Reading Dial Gage. Sample used
was only 0.989 in. long, and data were interpolated to

valuves far o 1-in. length,

 
Creep-rupture tests on the air-melted and vacuum-
remelted material, as shown previously,? indicated
that definite improvements in time to rupture and in
elongation were realized through vacuum remelting,
but that the creep rate of the air-melted alloy was
superior to that of the vacuum-remelted material,

Composite Tubing and Plate

In order to take advantage of the corrosion re-
sistance of the nickel-molybdenum alloys in fuel
107 at temperatures exceeding 1650°F, it would
be necessary to isolate the nickel-base alloy from
the sodium through the use of composite tubing.
Preliminary studies of INOR-8 and type 316 stain-
less steel diffusion couples show that an unknown
phase or phases precipitate at the interface as a
result of diffusion. The extent of the interface re-
action is negligible after diffusion times of 500 hr
at 1300°F. As expected, the severity of reaction
increases rapidly as the temperature is increased
to 1800°F. The precipitate is confined to the
nickel-base side of the interface and occurs as a
grain-boundary phase rather than as a continuous
layer. Room-temperature tensile tests of composite
sheet showed that strengthening occurred after
diffusion heat treatments at 1300 and 1500°F. A
weakening effect was observed in those specimens
heat treated at 1650 and 1800°F. A minimum
fracture ductility of 28% elongation was observed
in the specimens heat treated at 1650°F.

The feasibility of fabricating composite tubing of
Inconel and type 316 stainless steel was described
previously.! It is anticipated that similar pro-
cedures can be applied to the fabrication of com-
posites of INOR-8 and type 316 stainless steel.

MECHANICAL PROPERTIES INVESTIGATIONS

C. R. Kennedy? J. R. Weir
R. W. Swindemann J. W. Woods

Commercial nickel-base, molybdenum-containing
alloys, such as Hastelloy B or W, have been con-
sidered for reactor applications because of their
excellent resistance to chemical corrosion and
their relatively high strength at elevated temper-
atures. Because such alloys showed promise as
structural materials for a circulating-fuel reactor,

an extensive investigation of the mechanical

 

50n assignment from Pratt & Whitney Aircraft.

 

PERIOD ENDING DECEMBER 31, 1957

properties of the Hastelloy alloys was carried out.®
It was found that these alloys are not single-phase
and that they suffer a serious loss of ductility as
a result of the precipitation of a brittle beta phase
in the reactor operating temperature range. These
alloys also presented fabricational difficulties,
particularly in the production of seamless tubing,
it was considered that the best
solution would be to change the alloy composition.

and therefore

Therefore, special alloys designed to retain the
inherent advantages of the Hastelloys and to
eliminate their objectionable characteristics were
studied.

Screening tests were made of the creep-rupture
properties of various experimental alloys exposed
to fuel 107. For comparison, each alloy was tested
with a stress of 8000 psi at 1500°F in two con-
ditions: (1) annealed at 2000°F for 1 hr and (2)
annealed at 2000°F for 1 hr and aged for 50 hr at
1300°F. Results of the creep-rupture tests, pre-
viously reported,” =10 included data obtained for
air-melted heats prepared by Battelle Memorial
Institute, International Nickel Company, Westing-
house Electric Corporation, Superior Tube Company,
and Haynes Stellite Company, as well as for
vacuum-melted heats produced at ORNL.

Attempts were made to increase the strength of
the nickel-molybdenum alloys by the addition of
elements that form complex precipitates and by
the addition of elements that are simple carbide
formers. Additions of very small amounts (<1%)
of aluminum and titanium did not seem to increase
the strength, but did increase the elongation be-
fore fracture. Fabrication difficulties were en-
countered when additions of aluminum and fitanium

exceeded 1.5%. It was possible, however, to
improve the creep characteristics by additions
of carbide-forming elements, such as niobium

and tungsten, without seriously compromising
the other good characteristics. The BMI alloys

 

6c. R. Kennedy, D. A. Douglas, and W. D. Manly,
High Temperature Mechanical Properties of Hastelloy B
and W, ORNL-2402 (in press).

7c. R Kennedy, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 249.

8J. W. Woods, ANP Quar. Prog. Rep. March 31, 1957,
ORNL-2274, p 220.

9H. thouye, T. K. Roche, and J, H, Ceobs, ANP Quar.
Prog. Rep. June 30, 1957, ORNL-2340, p 176.

10, W. Woods and D. A. Douglas, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 171.

145
ANP PROJECT PROGRESS REPORT

containing carbon in excess of 0,1% exhibited
good high-strength properties, but they tended
subjected to tube-fabricating
processes. Additional alloying elements, such as
chromium (7%) and iron (4%), were found to im-
prove the creep properties of the basic nickel-
molybdenum binary alloy.

to fracture when

The Haynes Stellite INOR-8 (nominal composition,
in wt %: 16 Mo-7 Cr~5 F-0.02 C~bal Ni} appears
to be the best alloy produced to date. [t has good
corrosion resistance and fabricating characteristics,
but, as a result of its low carbon content, the
strength properties are only fair. It is considered
to be feasible to substantially increase the strength
of the alloy after the difficult fabrication processes
are completed by adding more carbon through gas
carburizing. Tests of specimens in which carbon
was added by this means showed greatly improved
strength properties.

A comparison of creep properties of untreated
alfoys and the same alloys after carburization with
the properties of Hastelloy B are presented in
Figs. 3.1.2, 3.1.3, and 3.1.4. It is evident that
carburized INOR-8 becomes equivalent to Hastelloy
B in creep strength at 1650°F.

UNCLASSIFIED
ORNL-LR-DWG 27058

ELONGATION (%)

12,000 hr

 

0 800 1600 2400 3200 4000
TIME (hr)

Fig. 3.1.2, Creep Properties of Untreated INOR-8 and
Carburized INOR-8 in Comparison with Those of
Hastelloy B Exposed to Fuel 107 at 1500°F at o Stress
of 8000 psi. (Secret with caption)

146

UNCLASSIFIED
ORNL-LR-DWG 27059

TELLOY B RUPTURED

AT 2750 hr WITH 4.6%
ELONGATION

ELONGATION {%)

 

o 200 400 600 800 1000

TIME (hr)

Fig. 3.1.3. Creep Properties of Untreated INOR-8 and
Carburized INOR-8 in Comparison with Those of
Hostelloy B Exposed to Fuel 107 at 1650°F at a Stress
of 5000 psi. (Secret with caption)

UNCLASSIFIED
ORNL-LR-DWG 27060

ELONGATION {%)

 

0 200 400 600 800 1000
TIME (hr)

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)

CORROSION STUDIES

J. H. DeVan
J. R. DiStefano R. S. Crouse

Two related series of tests have been made in
order to evaluate the corrosion properties of
nickel-molybdenum alloys in fused fluoride salts
and in sodium, [n the first series of tests, com-
mercial alloys, such as Hastelloy B and Hastelloy
W, were studied, while in the second series data
were obtained on the corrosion properties of ex-
perimental nickel-molybdenum alloys, which from
the standpoint of heat treatment and chemical
properties are better adapted to reactor operation.
 

Forced-circulation loops were used for most of the
tests of the Hastelloys, but both thermal-convection
and forced-circulation systems have been used to
study the experimental alloys.

Forced-Circulation Loop Tests of Hastelloys

The results of forced-circulation loop tests of
Hastelloy B with fuel 107 as the circulated fluvid
are summarized in Table 3.1.2. Further details of
these tests were presented in previous reports.”'12
While defects present in the as-received Hastelloy
B tubing precluded comprehensive analyses of the
test results, it was demonstrated in these tests
that basic nickel-molybdenum alloy systems are
compatible with alkali-metal fluorides at temper-
atures up to 1760°F.

Theresults of forced-circulation tests of Hastelloy
B and Hastelloy W with sodium as the circulated

 

”J. H. DeVan, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, p 136.

12_,. H. DeVan and R, S. Crouse, ANP Quar. Prog.
Rep. March 31, 1957, ORNL-2274, p 147.

 

PERIOD ENDING DECEMBER 31, 1957

fluid are summarized in Table 3.1.3. A comparison
of these results with the results for other classes
of alloys tested insodiumis presented in Fig. 3.1.5,
in which the weights of cold-zone deposits are
indicated as a function of temperature. As may be
seen, the Hastelioys and the nickel-base alloys,
in general, are considerably more susceptible to
mass transfer in sodium at temperatures above
1300°F than are either the austenitic or the ferritic
stainless steels.

Forced-Circulation l.oop Tests of Experimental

Nickel-Molybdenum Alloys

Two forced-circulation constructed of
experimental nickel-molybdenum alloys have been
successfully operated for 1000 hr with fuel 107.
The first of these loops, 7641-9, was discussed
previously.'®  The loop was fabricated from an

alloy containing 17% Mo-6% Fe-~bal Ni and was

loops

 

13). H. DeVan and R. S. Crouse, ANP Quar, Prog.
Rep. Sept. 30, 1957, ORNL.-2387, p 176.

UNCLASSIFIED
ORNL-LR-DWG 27061

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

24
/
y
. 20 HASTELLOY vyf
2 /
O /
A
" / /
o 16 //
3 / /NCONEL
O /
= /,
o / A{
212
é) HASTELLOY B/ /){|NCONEL ¥
& / /
o 8 £ /
7

—
I / / _-TYPE 330 STAINLESS STEEL
% // ’ 4)L TYPE 340 STAINLESS STEEL

] /’ TYPE 430 STAINLESS STEEL

/ TYPES 347. 316.304 TYPE 346 STAINLESS STEEL~ |
“//fi/ STAINLESS STEEL J____._.-{——"‘"'
O C_——-_-_— | L (= - w—
1200 1300 1400 1500 1600 1700
MAXIMUM FLUID TEMPERATURE (°F)
Fig. 3.1.5. Comparison of Results of 1000-hr Forced-Circulation Loop Tests of Various Alloys with Sodium.

All loops contained oxide cold traps and were operated with a temperature drop af 300°F.

147
Table 3,1.2. Operating Conditions and Results of Forced«Circulation Loop Tests of
Hastelloy B with Fuel 107 as the Circulated Fluid

8!

 

 

Maximum Fuel- Maximuym Fuel
Loop Period of Wall Interface Bulk Fuel Temperature Reynolds Flow As-Received Size
Number Operation Temperature Temperature  Differential Number Rate and Condition Metallographic Results
(hr) ©F) (°F) °F) (fps) of Tubing
7425-13A% 1000 1555 1505 200 10,000 3.2 Size: 0.5in. OD, 0.035 in. Small pits found along hot-leg
wall; as-received speci- surface to a depth of 1.5 mils;
mens had surface pits to no cold-leg deposits or
a depth of 1 mil layers
7641-1B 1000 1767 1635 280 10,000 2.6 Size: 0.5in. OD, 0.065 in. Many pits in hot leg to a depth
i 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,,0D, 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 407.5b 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 180¢ 1710 1645 293 20,000 5.1 Size: lé-in. sched-40 pipe; Loop was shut down and re-

inner surfaces badly
pitted and cracks present
on both sides of longitu-

dinal closure weld

paired four times because of
leaks; cleaning fills omitted
on third and fourth runs; a
silver«colored loosely ad-

herent layer found throughout

loop; layer consisted primarily

of fuel, but contained approxi-

mately 10% Ni and 5% Mo

 

A5uffix letters designate number of loop shutdowns prior to termination.

bTerminated because of pump faiiure; tubing ruptured upon attempting to restart circulation.

“Terminated because of tube failure.

[

LA0d3Y SSIYO0¥d LIIFrodd ANV
 

PERIOD ENDING DECEMBER 31, 1957

Table 3,1,3. Operating Conditions and Results of Forced-Circulu‘tion Loop Tests of Hastelloy B
and Hastelloy W with Sodium as the Circulated Fluid

Temperature differential: 300°F

Fluid flow rate:

2.5 gpm

 

Maximum Fluid Operating

Maximum Thickness Total Weight
Maximum Hot-

 

Loop Loop of Cold-Leg of Cold-Leg
] Temperature Time Leg Attack . .
Number Material op N | Deposits Deposits
(°F) (hr) (mils) i) o
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
temperature of 1760°F. Attack occurred along the
hot-leg surfaces in the form of intergranular voids
and was limited to a maximum depth of 4 mils,

A second loop constructed of an alloy containing
17% Mo~10% Cr-bal Ni has now completed oper-
ation with a maximum fluid—metal interface temper-
ature of 1850°F. Other operating conditions for
the loop, 7641-10, are given below:

Maximum bulk fuel temperature 1750°F
Fuel temperature drop 300°F
Reynolds number 10,000
Flow rate 1.4 gpm

Ratio of heated section surface 2.1 in.2/in.3

area to total loocp volume

Extensive attack was found metallographically
throughout the hot leg of the loop that reached, at
the point of maximum wall temperature, to a depth
of 23 mils. The attack was predominantly in the
form of intergranular voids and was accompanied
by removal of second-phase material from the grain
matrix, as well as from the grain boundaries.
Cold-leg sections showed little attack, but there
was a thin metallic deposit over most of the ex-
posed surface.  The quantities of metal con-
taminants in fuel samples taken after the test were

comparable with those normally found in Inconel —
fuel 30 systems tested at 1500°F. The chromium
content of the fuel was approximately 685 ppm,
and chromium was the impurity that was present in
the largest quantity.

It is obvious on the basis of these test results
that chromium additions in amounts of 10% or
greater would limit the usefulness of nickel-
molybdenum alloys in fuel 107 to temperatures
below 1800°F.  Thermal-convection loop test
results appear to have established a satisfactory
operating temperature of 1650°F, as discussed
below.

Thermal-Convection Loop Tests of Experimental

Nickel-Molybdenum Alloys

Approximately 50 thermal-convection loops
fabricated from alloys representing numerous
compositional variations of the 83% Ni-=-17% Mo
base have been operated with fuel 107 for periods
up to 1000 hr and temperatures up to 1650°F. The
primary purpose of these tests was to evaluate
the effect of specific alloying elements on the
corrosion resistance of the base composition.

Most of the loops were operated with a maximum
bulk fuel temperature of 1500°F and with a corre-
sponding 1650°F metal-fluid interface temperature.
The results of these tests were summarized in a
previous report.'® From the test data it has been

149

 
ANP PROJECT PROGRESS REPORT

possible to stipulate approximate limits of alloy
additions which can be permitted from the stand-
point of corrosion penetration and resistance to
thermal-gradient mass transfer. As would be
expected, the limits are quite dependent on oper-
ating temperatures. A list of several promising

150

compositions and their approximate temperature
limitations for operation with fuel 107 were pre-
sented in Table 3.1.1. Operation at temperatures
only 100°F lower than those shown has been found
to reduce the corrosive attack for the same time
period to a depth of 1 to 2 mils.

 

 
PERIOD ENDING DECEMBER 31, 1957

3.2, MECHANICAL PROPERTIES STUDIES

D. A. Douglas

C. R. Kennedy'

R. W. Swindemann

An extensive investigation has been under way
during the past several years in which the me-
chanical properties of the various materials to
be used in circulating-fuel aircraft reactors were
evaluated under simulated service conditions. Par-
ticular attention has been given to obtaining
design data and to studying the behavior of the
materials under the various conditions of stress,
temperature, and environment that will be en-
countered during operation of the reactor. Fairly
complete design data were obtained for Inconel,
the structural alloy chosen for the ART, for
beryllium, the moderator material, and for various
shielding materials. Basic information on the
behavior of metals has also been obtained in order
to be able to predict the safe operating life of
various components under a variety of operating
conditions,

DESIGN DATA

Inconel

Creep and stress-rupture data were obtained for
Inconel in the temperature range of 1300 to 1650°F
in argon, in sodium, and in the fuel mixture

NaF-Z¢F -UF , (50-46-4 mole %, fuel 30). The

data for an argon atmosphere were used for com-

parison. VYariables such as grain size, section
thickness, and alloy consistency were investi-
gated. It was found that in the fused-salt en-

vironment is stronger
than the coarse-grained metal at 1300°F but
weaker at 1500°F and above. However, in argon
and in sodium the fine-grained Inconel maintains
its superiority over the coarse-grained material at
temperatures up to 1650°F, where metallurgical
instabilities begin to introduce adverse effects.
A comparison of three heats of Inconel showed
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
to account for the lack of uniformity; therefore it
appears that the source of the difference is the
variations of the minor additions or of the residual
elements which are difficult to control in air melts.

the fine-grained material

 

1On assignment from Pratt & Whitney Aircraft.

J. R, Weir
J. W. Woods

Details
ported. 2

Since the strength properties of Inconel make it
only marginally useful at peak aircraft reactor
operating temperatures, some thought has been
given to a modification which would increase the
strength without impairing the excellent fabricating
characteristics, Carbon is a very potent
strengthening agent because it disperses carbides
throughout the matrix and in the grain boundaries,
but, unfortunately, such a structure is difficult
to draw into tubing. It was believed to be feasible,
however, to fabricate the tubing and then add
additional carbon through carburization, Several

of these investigations have been re-

creep specimens were therefore pack-carburized
and tested in fuel 30 and in argon. The results
of these tests, as presented in Figs. 3.2,1, 3.2.2,
and 3,2.3, show that a very marked improvement in
creep properties was obtained.

 

2J. R. Weir, D. A. Douglas, and W. D. Manly, Inconel
as a Structural Material for a High-Temperature Fused-
Salt Reactor, ORNL-2264 (June 4, 1957).

UNCLASSIFIED
ORNL-LR-DWG 27062

 

S
=
<
o
.—
n

0 50 100 150 200 250

TIME (hr)
Fig. 3.2.1. Effect of Carburization on the Creep

Properties of Inconel Tested in Fuel 30 at 1500°F and

a Stress of 5000 psi. (Secret with caption)

151
 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR-DWG 27063

STRAIN {%)

0
e 9.6\3?'\1?‘
<

 

0 100 200 300 400 500 600 700 800 900 4000
TIME (hr)

Fig. 3.2.2.
Properties of Inconel Tested in Fuel 30 at 1650°F and
3000 psi. {Secret with caption)

Effect of Carburization on the Creep

UNCLASSIFIED
ORNL-LR-DWG 27064

 

5
4
S
23
<
o
o2
\ BURIZED
O
0 {00 200 300 400 500 600 700 800 900 (000
TIME (hr)
Fig. 3.2.3. Effect of Carburization on the Creep

Properties of Inconel Tested in Fuel 30 at 1800°F and
1000 psi. {Secret with caption)

Recrystallized Inconel

A stable metallurgical microstructure is required
in order to obtain optimum-strength properties in
an alloy. Tests have demonstrated that metals
which have been cold worked to various degrees
may recrystallize under certain conditions of
stress and temperature and that deleterious effects
on the creep properties will result.® Since many
reactor components are fabricated by pressure
forming methods and are assembled by welding
and brazing, portions of the structural metal will
have experienced a variety of thermal treatments

 

3J. R. Weir, Jr., ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 244.

152

prior to operation. Therefore a study was under-
taken in order to establish the temperatures
required for recrystallization after varying amounts
of cold working and to determine the stable grain
size achieved after high-temperature anneals,

The testing procedures used and illustrations of
the grain sizes produced in recrystallized struc-
tures have been presented elsewhere.® It was
noted in this study that the recrystallization
temperature ranged from 1200 to 1650°F and was
dependent on the amount of cold working and the
time held at temperature. Fine-grained Inconel
recrystallized ot a temperature about 100°F lower
than the temperature at which the coarse-grained
material recrystallized for the same amount of prior
deformation and time at temperature. The grain
size that resulted from high-temperature treatments
was found to be dependent on the degree of
working prior to recrystallization up to a temper-
ature of 1800°F. However, at 1900°F only very
small differences in the final grain sizes could
be seen regardless of the prior treatment.

Beryllium

Extensive mechanical tests of beryllium have
been made in the temperature range of 800 to
1500°F by the Brush Beryllium Company under
subcontract P.O, W34X-75581.% Also tests were
made of samples trepanned from the actual be-
ryllium moderator blocks prepared for the ART,
The mechanical properties of the actual blocks
were in excellent agreement with those obtained
in the prior test program for material of like
composition, Thus, it appears that data obtained
in the test program can be reliably applied to
large-scale production pressings.

EXTRAPOLATION OF TEST DATA

One of the problems associated with mechanical
testing work is that there is frequent need for data
under conditions outside the ranges of temperature
or stress that have been explored. Thus, the first
step in being able to anticipate the behavior of
metals under service conditions is to develop a
reliable method of interpolating and extrapolating

 

4p. A. Douglas and J. R. Weir, Recrystallization and
Grain Growth of Inconel, ORNL-2406 (in press).

>R. G. O'Rourke et al., Mechanical Properties of
Reactor-Grade Beryllium at Elevated Temperatures,

CO00-312 (Aug. 1956).
 

the available data. There has been much work
done in this field, and there are several relation-
ships which appear to have merit. The multitude
of metallurgical variables involved in high-
temperature deformation processes made it ad-
visable, however, to derive empirical equations
based on fairly specific data. Such equations are
available for Inconel, and the method for using
them and the limitations of the results are de-
scribed in a topical report.®

THERMAL STRESS INVESTIGATIONS

The thermal cycles which occur in a reactor
during its operation induce in the structural
members corresponding stress and strain cycles.
Thus for safe operation it is essential to know
the effect of these strain cycles on the endurance
of the structural materials. In order to obtain
design data of this type, tests were conducted to
determine the strain-cycle and stress-relaxation
properties of Inconel, Important factors which
have been investigated in conjunction with these
tests are the effects of temperature, grain size,
environment, specimen geometry, and cycle fre-
quency. Testing devices were constructed for
testing both tube and rod specimens. In order to
gain a clear picture of the true influence of strain
reversals, these rigs were designed to mechanically
strain the specimens; whereas, under true oper-
ating conditions, the strains may be produced by
thermal fluctuations. Supporting investigations in
which thermally induced strain cycles are being
studied are under way at the University of Alabama,

Strain-cycle tests in argon at 1300, 1500, and
1600°F were completed, and the data were pre-
sented previously.”~® Curves have been drawn
which indicate the number of strain cycles to
failure, N, as a function of plastic strain per

cycle, €,. An empirical equation,

“p

 

5c. R. Kennedy and D. A. Douglas, Use of a Rate
Equation for the Extrapolation of Creep Data, ORNL
CF-57-8-119 (Aug. 20, 1957).

’c. R. Kennedy and D. A. Douglas, Plastic Strain
Absorption as a Criterion for High-Temperature Design,
ORNL-2360 (in press).

8J. R. Weir, Jr., ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 246.

°%C. R. Kennedy and D. A. Douglas, ANP Quar. Prog.
Rep. June 30, 1957, ORNL-2340, p 190.

PERIOD ENDING DECEMBER 31, 1957

 

has been formulated to relate the data, which fall
on a straight line when plotted as log €, vs log N.
Similar behavior was observed by Coffin'® in his
work on thermally induced cyclic strains in type
347 stainless steel.

At 1500 and 1600°F, fine-grained Inconel is
superior to coarse-grained material; that is, the
data indicate that fine-grained specimens endure
a greater number of cycles at a given strain-cycle
level before failure. At 1300°F, however, the
difference in cycles to rupture between coarse-
and fine-grained Inconel is negligible.

The influence of geometry is not great. Rod
specimens appear to endure slightly more cycles
before failure than tube specimens. However,
this improvement may be attributed to a less
sensitive method of detecting the initiation of
failure in the case of rod specimens.

The effect of temperature on these properties is
not appreciable over the temperature range studied.
Specimens cycled at 1600°F exhibit only a slightly
greater number of cycles to failure than specimens
tested under comparable conditions at 1300 and
1500°F,

Tests have also been conducted on Inconel tubes
in fuel 30 at 1500 and 1600°F, Results of these
tests were presented in previous reports in this
series.” ! The effect of fuel is to reduce the
number of cycles to failure by 40 to 66% of the
number of cycles in argon, depending on the plastic
strain per cycle and the grain size.

Since the frequency of induced-strain cycles
within a reactor will not be uniform, a program
has been set up for determining the effect of
frequency on the strain-cycle properties. Pre-

tests had been conducted with «a

viously, all
11

2-min/cycle period.
are being conducted with a 30-min/cycle period
on Inconel tube specimens at 1500°F, and there
are indications that frequency may be an important
factor. Specimens cycled at small strains and
30-min/cycle periods do not endure as many cycles
as those cycled with the 2-min/cycle period.
Further investigations will have to be completed
before this effect can be entirely substantiated
and understood.

Present investigat ions”

 

]OL. F. Coffin, Jr., Trans. Am. Soc. Mech. Engrs. 76,
931-50 (1954).

e R Kennedy and D. A. Douglas, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 182,

153
ANP PROJECT PROGRESS REPORT

Results from the tests being made at the
University of Alabama indicate that mechanical
strain-cycle effects at 1300°F correlate well with
thermally induced strain-cycle behavior at the
same mean temperature, 1

Additional tests were made in order to determine
the effect of carburization, weld metal, and braze-
metal coatings on the strain-cycle properties of
Inconel tubes. Tests conducted on carburized
Inconel tubes showed no change in strain-cycle
characteristics as a result of the carburizing
treatment.'' A series of tests on all-weld-metal
tubes was cancelled before completion, but the
preliminary results'' indicated that the welds
were stronger in tension than the base metal at
temperatures from 1400 to 1800°F. Several Inconel
tube specimens coated with braze metal (92% Ni-
3% Si-3% B-2% Fe) were tested in order to
determine whether boron penetration of the base
material would affect the strain-cycle properties.
Rupture data from these tests at 1500°F were in
good agreement with the design curve for coarse-
grained Inconel at 1500°F. Consequently, it may
be assumed that no detectable impairment of
strain-cycle characteristics will be produced by
the presence of a braze-metal coating.

A series of tests for evaluating the effect of
prior strain cycling on the creep properties of
Inconel was also conducted. The procedure and
initial results were described previously.!! In
summary, the available data indicate that the
creep strength and rupture life are significantly
reduced by prior strain cycling and that, with
increasing consumption of strain-cycle life, the
creep-life expectancy is correspondingly reduced.
The inference is that design calculations based
on only one of these conditions would be opti-
mistic for instances where both strain cycling
and creep occurred.

Beryllium is the only other reactor material for
which strain-cycle characteristics have been in-
vestigated. A series of tests was conducted at
1250°F on hot-pressed, reactor-grade beryllium,
and the results were reported previously.'! The
test data revealed that beryllium, although much
inferior to Inconel, nevertheless, does have
ductility and is capable of withstanding cyclic
strains at 1250°F,

Relaxation

Stresses induced by a thermal cycle in a con-
strained structural member usually undergo some

154

degree of decay or relaxation if the temperature
is sufficiently high This relaxation of stress
occurs through a conversion, by creep, of the
elastic strain into plastic strain. In order to
establish the true plastic strain induced by a
thermal loading, the relaxation characteristics of
a material must be known and taken into consider-
ation,

Relaxation tests of Inconel were completed for
coarse- and fine-grained specimens at 1100, 1300,
1500, and 1650°F, and the test procedures,
apparatus, and results have been described.'?
The data obtained in these tests indicate that,
at temperatures greater than 1300°F, relaxation
of stress occurs rapidly. The relaxation rate at
these temperatures varies only when the initial
strain is below the elastic limit.  Moreover,
induced plastic strains do not significontly in-
crease the elastic limit, since recovery appears
to be rapid at temperatures above 1300°F,
Stresses are reduced to low residual values after
a few minutes. Below 1300°F, however, relaxation
occurs less rapidly. At 1100°F high initial
strains produce work hardening, which results in
larger elastic strains and higher stress values;
these strains, in turn, relax slowly. The results
also give indications of a grain-size effect. The
relaxation rates of coarse-grained Inconel at all
temperatures were lower than the rates for fine-
grained material.

Fatigue

Under a subcontract to Battelle Memorial In-
stitute, a program was initiated to investigate the
effect of cyclic stresses imposed on a statically
loaded system. The aim of this program was to
define the stress conditions under which fatigue
is the dominant failure mechanism and the stress
conditions under which creep may be regarded as
the controlling mechanism for failure. Preliminary
tests for checking the performance of the apparatus
and the reproducibility of results were satis-
factorily completed. At present, tests are being
conducted to determine the influence of stress-
cycle frequency on the fatigue properties.

Biaxial Creep

Very little is known concerning the influence of
stress distribution on deformation rates, and the

 

uC. R. Kennedy, D. A. Douglas, and W. D. Manly,
Relaxation Characteristics of Inconel at 1100 1o 1650°F,
ORNL-2407 (in press}).

 
fact that structures are normally subjected to
multiaxial stresses emphasizes the need for ad-
ditional research in this area. Therefore a program
of tests is being conducted in which tubes are
subjected to various ratios of axial and tangential
stress and the resultant creep rates are measured.
A considerable change in rupture life, ranging
from 200 to 2000 hr, and significant variations of
creep rates have been noted.'> A considerable
quantity of data must be accumulated before it
will be possible to determine which of the various
mechanical theories of failure describes the
mechanism of fracture under these test conditions,

TESTS OF SHIELDING MATERIALS

Creep-rupture testing of an 80% Mg—-20% Li alloy
proposed as a neutron shielding material was
completed. Data in addition to those presented
previously'® have been obtained at 200°F for
specimens that were given a chemical surface

 

‘3C. R. Kennedy and D. A. Douglas, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 185.

14, R. Weir and C. W. Dollins, ANP Quar. Prog. Rep.
Sept. 10, 1956, ORNL-2157, p 208.

PERIOD ENDING DECEMBER 31, 1957

treatment ' and are compared in Table 3.2.1 with
data obtained for untreated specimens.

An alloy of very pure lead containing 0.06%
copper has been proposed for use as a gamma-ray
shielding material. Since no mechanical property
data were available at a temperature of 230°F,
the expected maximum operating temperature of
the lead shield, a creep-rupture testing program
was initiated to determine the strength of this
alloy. A design criterion of a creep rate of 0.5%
strain per year for a stress of 100 psi was es-
tablished. The results of tests in air show that
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
year and 0.05% per year, respectively. A lead-
Jithium alloy was also investigated in the hope
that its excellent room-temperature strength in-
dicated good high-temperature strength, However,
it was found that the lead-copper alloy has greater
creep strength at 230°F than the lead-lithium
alloy.'$

 

lsR. E. McDonald and C. F. Leitten, Jr., ANP Quar.
Prog. Rep. June 10, 1956, ORNL-2106, p 173,

]60. H. Jansen, E. E. Hoffman, and D. M. Sheppard,
Lead-Lithium Shielding Alloy ~ Metallurgical Studies,
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 72 121 420 240
800 0.5 1 71

 

155
ANP PROJECT PROGRESS REPORT

3.3. WELDING AND BRAZING STUDIES

P. Patriarca

G. M. Slaughter

PRODUCTION OF CAST RINGS OF BRAZING
ALLOY

G. M. Slaughter

Difficulties were encountered by the York Corp.
in the consistent brazing of radiator test units
with sintered brazing-alloy rings obtained from
Coast Metals, Inc.! Therefore a decision was

P. Patriarca

made to use brazing-alloy rings cast in graphite
molds for the fabrication of future units. Ex-
periments have indicated that the cast rings are
not so sensitive to the rate of temperature rise
as are the sintered rings? and that the required
consistent minimum brazing temperature is some-
what less,

Consequently, a production facility was set up
by the Engineering and Mechanical Division for
manufacturing approximately 3,000,000 rings ac-
cording to the procedures previously developed by
the Welding and Brazing Group of the Metallurgy
Division.® Assistance was given by Metallurgy
Division personnel in the determination and control
of optimum ring dimensions and weights and in the
establishment of systematic weight-control records,
A  production rate of approximately 60,000 to
70,000 rings per day was achieved, and a total
of approximately 750,000 rings was manufactured
prior to termination of production on September 6,
1957. A detailed description of the equipment and
production procedures, including a cost estimate,
is being prepared by the Engineering and Me-
chanical Division.4

FABRICATION OF THERMOCOUPLES FOR
HIGH-TEMPERATURE SERVICE

P. Patriarca

The use of thermocouples in high-temperature
fused-salt systems requires that a reliable thermo-
couple junction be made and that a leaktight seal
be assured. It is also desired that the locations

 

6. M Slavghter, ANP Quar. Prog. Rep. June 30,
1957, ORNL-2340, p 223.

2G. M. Slaughter, ANP Quar. Prog. Rep. March 31,
1957, ORNL-2274, p 205.

156

R, L. Heestand

of the junctions of all thermocouples be at a
consistent and known distance from the end cover.

The problem of sealing the junction end of
four-wire Chromel-Alumel thermocouples which
were housed in Inconel tubing and insulated with
magnesium oxide was first approached by welding.
A small Inconel plug was inserted into a hole
which had been prepared by sand-blasting the
magnesium oxide to the desired depth. The
assembled components were then fused together,
both with and without the addition of filler wire.
The inadequacy of this and other similar pro-
cedures was apparent when radiographic inspection
resulted in a rejection rate of approximately 20%
as a result of cracks and pores in the weld zone.

High-temperature furnace-brazing procedures were
then utilized in developing a suitable fabrication
technique. In initial experiments the magnesium
oxide was excavated for a short distance, the hole
was filled with brazing alloy, and the unit was
brazed. As was expected, this procedure resulted
in excessive porosity and poor wetting of the
2%-aluminum-containing Alumel component of the
thermocouple. Most test specimens leaked during
vacuum testing; thus the seals would have been
unsatisfactory for use in the corrosive fused salt
environment, Further, the distance between the
end of the seal and the thermocouple junction
varied both in seals made by this method and in
those made by the welding technique, and such
variations could give significant errors in temper-
ature measurements during service.

A method was consequently developed which
satisfied the requirements as to both integrity of
the seal and consistency in the distance between
the end of the seal and the thermocouple junction.
A schematic drawing that illustrates the thermo-
couple fabrication procedure is presented in Fig.
3.3.1, and the products of the various steps of
the procedure are shown in Fig. 3,3.2.

 

3p, Patriarca and R. E. Clausing, ANP Quar. Prog.
Rep. Sept. 10, 1956, ORNL-2157, p 200.

4y, s. Hornbaker, Production of Fused Coast Metals
ll\fo.h 3)2 Alloy Brazing Rings, ORNL-2424 (to be pub-
ished).

 

 
 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL-LR—-DWG 27065

 

BRAZING ALLOY RING
OR POWDER (0.04q)

 

 

EXCAVATED ~u g in.

 

MAGNESIUM OXIDE INSULATION

 

-

THERMOCOQUPLE WIRES, 2 CHROMEL
AND 2 ALUMEL,0.030-in.0D —=

A/ANANTINFARNANR

 

 

 

 

 

-«—[INCONEL SHEATH, 0.254-in.0D
0.035-in.WALL,BORED WITH NO.12
DRILL TO TOP OF CAP

 

> INCONEL COVER, 0.025 -in. THICK,
%G—in.DIA, PUNCHED

 

INCONEL CAP,0.025-in. THICK,
1%4—5n.DIA, PUNCHED AND DRILLED

WITH NO.64 DRILL

 

 

 

 

 

 

 

 

 

 

 

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

After the magnesium oxide is excavated to the
desired depth as illustrated, the four-hole Inconel
disk is inserted into the hole, the thermocouple
wires are cut off, a solid Inconel disk is placed
on top, and the assembly is brazed in dry hydrogen
at 1080°C. The brazing alloy can be preplaced
either as powder or as a cast ring. |f o cast ring
is to be used, it is advisable to prime the ring
to the joint with a slight amount of powder,
Radiographic inspection can be used after manu-
facture to assure that high-quality joints are
obtained.

FABRICATION OF ART HEAT EXCHANGERS
G. M. Slaughter

The fabrication of an experimental ART fuel-to-
NaK heat exchanger by Black, Sivalls & Bryson,
Inc,, Tulsa, Oklahoma, was observed closely in
order to assure the adequacy of the high-temper-
ature brazing procedures utilized. The preplace-
ment of Coast Metals brazing alloy No. 52 on the
tube-to-header joints was closely examined before
the unit was placed on the brazing fixture. Cast
brazing-alloy rings, both circular and elliptical,

157
ANP PROJECT PROGRESS REPORT

g -

cuT EXCAVATED

 

UNCLASSIFIED

 

@ o©

COVER CM 52 RING
ASSEMBLED BRAZING ASSEMBLED
RING

Y-24060
CAP BORED COVER
ASSEMBLED
] 1
l?"'}[
OR CM 52 AFTER GROUND
POWER BRAZING TO COVER

Fig. 3.3.2. Products of the Various Steps in the Preparation of Thermocouples for High=Temperature Service.

had been made in graphite molds at Black,
Sivalls & Bryson, but there was considerable mis-
fitting, In order to permit intimate contact between
the rings and headers, many of the rings had to
be removed and replaced by cemented half-rings.
Additional brazing alloy in the form of slurry had
to be applied to a large number of joints to
connect the rings to the capillary joints. In a
few instances, slurry had to be added to joints
where the rings did not intimately contact the
tubes.

The unit was placed on a large fixture, pre-
viously constructed at the York Corporation, that
did not accurotely match the contour of the heat
exchanger; consequently, the extensive use of
shims was required. This fixture was utilized for
expediency, and one machined to greater accuracy
was being constructed for brazing prototype units.
The tube sheets of both headers were in the
vertical position during brazing, with the axes of
the tubes horizontal at the point of entrance to
the header. Various measurements were made for
determining the degree of distortion during brazing.

The unit was sealed in a retort and was brazed
in a Lindberg furnace that had a working space

158

approximately 6 ft wide, 8 ft deep, and 10 ft high.
The maximum power rating of this furnace was
420 kw, and the recommended maximum working
temperature was 2050°F. The hearth of the
furnace was mounted on a flat car so that the retort
could be located on the hearth while outside the
furnace. The furnace temperature could be closely
regulated with ten individual-zone controllers.

The atmosphere purification system for the
furnace consisted of a BAC-150 activated-alumina
Lectrodryer (closed cycle) and a Deoxo model
D-1000-50 palladium catalyst. The inlet dewpoint
to the retort was approximately —100°F throughout
the brazing cycle,

Eight Chromel-Alumel thermocouples were at-
tached to the assembly, job samples, and fixture,
with two attached to each header of the heat
exchanger. An approximate rate of temperature
rise of 300°F/hr was employed, and %—hr temper-
ature stabilization periods were used at approxi-
mately 1400 and 1600°F,

At the stort of the soak period, the header
thermocouples indicated temperatures that varied

from 1900 to 1925°F. At the end of the 30-min

sock period, the temperature variation recorded by

 
these four thermocouples was reduced to 5°F
(1915-1920°F). The maximum temperature re-
corded by these thermocouples was 1925°F,

After removal of the heat exchanger from the
retort, the brazes on both headers were examined.
All visible joints, including those on the two job
samples, had relatively good fillets, and the unit
was, in general, clean and bright. The extent of
runoff from one header was such, however, that
the bottom row of tubes was solidly brazed to the
bottom wedge of the header, More careful control
of the brazing alloy preplacement procedures
should reduce runoff, but it is felt that additional
preventive measures should be studied.

in summary, it appeared that the gas purification
and brazing equipment were adequate and that the
general brazing procedures were well conducted.
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
could prove to be insurmountable distortion
problems, It was recommended, therefore, that
extra time and effort be devoted to the preplace-
ment problem in an attempt to prevent runoff and
to ensure reliable fillets.

ART FILL-AND-DRAIN TANK FABRICATICON
R. L. Heestand

An investigation of optimum methods for fabri-
cating the ART fill-and-drain tank tube-to-tube-
sheet joints was completed. A description of the
techniques used in the study was given pre-
vious!y.s"‘S

The optimum joint design consisted of trepanning
the weld side of the tube sheet, as shown in
Fig. 3.3.3. The trepanning serves to ensure good
heat distribution by providing a uniform weld
geometry and to minimize tube-sheet distortion and
restraint in the vicinity of the weld joint. Test
joints made initially by fusion welding showed
some root cracks, but welds made with Inco No. 62

 

5E. A. Franco-Ferreira and G. M. Slaughter, ANP
Quar. Prog. Rep. March 31, 1957, ORNL.2274, p 200.

SE. A F ranco-Ferreira, ANP Quar. Prog. Rep.
June 30, 1957, ORNL-2340, p 225.

PERIOD ENDING DECEMBER 31, 1957

filler metal added to the joint were found to be
free of root cracks.

In back-brazing the welded tube-to-tube-sheet
joints, a unique heating problem existed in that
the tube-sheet temperature continuously lagged
behind the tube temperature throughout the brazing
cycle because of the much greater thickness of
the tube sheet. When Coast Metals brazing alloy
No. 52 was applied by conventional means, the
alloy would preferentially flow to the hottest
member, and consequently there was poor flow on
the tube sheet, The method of brazing-alloy
preplacement shown in Fig. 3.3.3 and described

UNCLASSIFIED
ORNL-LR- DWG 27066

‘)-—----—0.890 in. —

 

t [—~O?60in.+—4

0430 in. .

0.630in.

‘ | .n“—T % \/WELD SI,E,)FT
\\ \\\\ ‘.\\\_\‘\\.‘ - \\
AN SRR

3 HOLES,0.062~in DIA, NN TON Y

AT {20 DEG \§ . -

     

N
. ~
.
~ ~

. N
. ~, ~
() SO

PREPLACED BRAZING ALLOY "EVB_ in.0D, G.065 -in. WALL TUBING

 

 

 

 

-~

Fig. 3.3.3. Trepanned Joint Detail.

previously® removed the problem by allowing the
brazing alloy to remain at the temperature of the
tube sheet until the brazing temperature was
reacned. The trepanned sump on the braze side
of the tube sheet then acted as an infinite res-
ervoir from which the alloy flowed to fill the
annulus between the tube and the tube sheet and
to form a fillet where the tube met the tube-sheet
surface., A typical joint of this type, cut so as
to show one of the three sump holes, is shown in

Fig. 3.3.4.

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
braze cycle, a 66-tube Inconel sample that simu-
lated a tube-sheet section was prepared. Brazing-
alloy slurry (Coast Metals No. 52) was applied to

159

 

 
ANP PROJECT PROGRESS REPORT

the braze side, and the tubes were then inserted
and were welded with Inco No. 62 weld wire.

 

Fig. 3.3.4. Cross Section of @ Typical Welded and
Brazed Tube-to-Tube-S5heet Joint.

After brazing at 1920°F, no visual distortion was
evident, and the complete fillets at all joints
indicated a high degree of reliability, The weld
side of the completed sample is shown in Fig.
3.3.5, and the braze side is shown in Figs. 3.3.6
and 3.3.7.

The utilization of brazing in the fabrication of
this unit is unique, since high-temperature brazes
ore not normally used on tube sheets of this
thickness.  This method of alloy preplacement
permits a high degree of reliability in both braze
penetration and filleting, however, ond it may well
be applicable to tube-to-tube-sheet joints on other
types of heat exchangers having components of
similar size,

NIOBIUM WELDING

R. L. Heestand

Investigations of methods for welding niobium
were started. Attempts are being made to produce

160

ductile welds by the inert-arc fusion process and
by the addition of filler metals.” Fusion welds
manufactured manually at the Faonsteel Metal-
lurgical Corporation and fusion welds machine-
made ot ORNL have been examined metallo-
graphically.,  Although some differences in the
sizes of the grains in the two welds were evident
(Figs. 3.3.8 and 3.3.9), no other significant
differences were apparent. Both types of fusion
welds were tested in simple bend tests and found
to be ductile. Experiments are also being carried
out to determine differences in weldability of arc-
cast and powder-metallurgy niobium. An inert-
atmosphere dry box and fusion-welding techniques
are being used for these experiments. Pure
niobium wire and wire fabricated from arc-cast
ingots of zirconium-niobium and zirconium-niobium-
molybdenum alloys were prepared as filler metals
for additional welding experiments. Welds fabri-
cated by depositing niobium wire have large
grains, ond they have leaked in high-temperature
service in liquid metals, even though they had
passed room-temperature mass-spectrometer |eak
detection tests prior to operation. A cross section
of a saddle weld made with niobium wire is shown

in Fig. 3.3.10.

In the development of filler wires for welding
niobium, the zirconium-niobium system was in-
vestigated because it was known that the 85% Zr—
15% Nb alloy had satisfactory melting character-
istics, Development studies pertaining to the
transformation kinetics and fabrication of this
alloy had been made by the Physical Metallurgy
Group associated with the Homogeneous Reactor
Project, and a prediction of the behavior of the
alloy could be made. The studies had indicated
that the zirconium-niobium alloy, with ternary
additions to control age-hardening reactions, would
be desirable because of its melting point, melting
range, and transformation kinetics, and becouse
of the possibility of solution of the niobium base
metal in the weld metal.

Both butt welds aond tubing saddle welds have
been fabricated in an argon-atmosphere dry box
(Fig. 3.3.11) by using the tungsten-arc process
and zirconium-niobium filler wire. A typical butt
weld in sheet niobium is shown in Fig. 3.3.12.

 

7R. L. Heestand ond E. E. Hoftman, Welding of
Niobium Progress Report No. 1, ORNL-2423 (to be
published).

 

 
PERIOD ENDING DECEMBER 31, 1957

UMNCLASSIFIED
PHOTO 41927

 

 

 

QAKX RIDGE NATIONAL LABORATORY

 

Fig. 3.3.5. Weld Side of Fill-and-Drain-Tank Sample.

5 , 1 | | - | UNCLASSIFIED |
) ! | | | | PHOTO 41926 |

R

 

e

 

Fig. 3.3.6. Braze Side of Fill-ond-Drain-Tank Sample Showing Filleting of Trepanned Braze Joint.

161
ANP PROJECT PROGRESS REPORT

UMCL ASSIFIED
FHOTO 41008

 

 

 

 

 

 

Fig. 3.3.7. Tube Side of Fill-oand-Drain-Tank Sample.

 

The joint was spaced with a gap approximately
0.030 in. wide prior to welding. As may be seen
in Fig. 3.3.12, the filler metal simulates a brazing
alloy rather than an actual welding filler metal.
The gap was penetrated, and wetting was observed
on the side opposite to the welding area. Welds
of this type have been found in free bend tests
to be ductile when bent through a 90-deg angle.
The major distortion was in the weld.

The more complicated saddle weld shown in
Fig. 3.3.13 and in cross section in Fig. 3.3.14
was also fabricated with zirconium-niobium filler
wire. In this case no joint spacing was used, and
the alloy did not penetrate to the inside of the
joint. It may be noted that the fine grain of the
niobium base metal was retained, A saddle weld
which was sectioned and crushed to demonstrate
the ductility of the joint is shown in Fig. 3.3.15.

Similar butt welds and tubing saddle welds were
fabricated with 80% Zr—15% Nb—5% Mo filler wire,
and the ductility was found upon bending to be
the same as for the 85% Zr—15% Nb wire. Depo-
sition of the 80% Zr-15% Nb-5% Mo alloy is
somewhaf more difficult, however, because of a
higher melting point. The 80% Zr-15% Nb-5% Mo
alloy shows greater strength than the 85% Zr-—
15% Nb alloy. More complete investigations of
alloys of the zirconium-niobium system are planned.

UNCLASSIFIED
Y-24309

Fig. 3.3.8. Fusion Seam Weld Machine-Made in Niocbium Tubing at ORNL. 22X.

162

 
 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
Y-24307

 

Fig. 3.3.9. Fusion Seom Weld Monually Mode in Nicbium Tubing ot Fansteel Metallurgical Corporotion. 22X,

 

Figs. 3.3.10. Cross Section of Niobium Saddle Weld Made with Nicbium Filler Wire. Weld made at ORNL by using
inert-arc method., 22X.

163

 
 

ANP PROJECT PROGRESS REPORT

el UNCLASSIFIED
HEL IUM Y- 10944
; TANK . Lk

HELIARC WELDING Gl
TORCH INLET

ACTIVATED-.CHAR
COAL
NITROGEN-GAS PURJFTE D" )

B

 

Fig. 3.3.11. Argon-Atmosphere Dry Box for Welding Niobium,

UNCLASSIFIED
Y-24312

 

Fig. 3.3.12. Sheet Niobium Butt Weld Made by Using 85% Zr—-15% Nb Filler Wire. 22X.

164

 
 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
Y-24264

 

ONE INCH

 

 

 

 

 

Fig. 3.3.13. Niobium Saddle Weld Made with 85% Zr—15% Nb Filler Wire. 3X.

UNCLASSIFIED
Y-24311

 

Fig. 3.3.14. Cross Section of Weld Shown in Fig. 3.3.13. 22X.

165
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-24267

 

 

 

ONE INCH

 

 

||1l1||

 

Fig. 3.3.15. Weld Shown in Fig. 3.3.13 as Sectioned
and Crushed ot Recom Temperature Te Show Ductility.

It has also been noted that impurities in the
welding atmosphere, nitrogen and oxygen in par-
ticular, have embrittling effects on the weld metal
and the base metal. Therefore the impurity
tolerance of these elements will be investigated.

CARBURIZATION OF INCONEL TUBING
G. M. Slaughter

Means for obtaining fine grain structure in
Inconel heat exchanger tubing after brazing have
been extensively investigated. The influence of
brazing temperature, time at temperature, and prior
cold working were discussed in recent previous
reports in this series.

Experiments have now been conducted with
carburized tubes to study the effect of carburi-
zation on' the grain size. An as-welded joint is
shown in Fig. 3.3.16, and a joint that was brazed
at 1080°C for 30 min is shown in Fig. 3.3.17. It
may be seen that the carbides were distributed
in the as-welded structure and that they went into
solution and reprecipitated during the brazing
cycle. The fine-grained structure evident in the
as-brazed tubing indicates that the carbides were
beneficial in preventing grain growth. Extensive

P. Patriarco

166

INCHES ]

|

3

 

164

 

 

Fig. 3.3.16, As-Welded Tube-to-Heoder Joint Made
with Carburized Inconel Tubing. Hardness numbers are
in DPH units. Etchant: glycerio regia. 75X. Reduced
12.5%.

grain growth would have occurred in conventional
Inconel subjected to the same brazing cycle., No
significant changes in the hordness of the tubing
as a result of the brazing cycle can be noted. The
weld possessed a hardness approaching that of
the uncarburized header material, since con-
siderable dilution of the header material occurred.

Sections taken near brazed fillets on carburized
and uncarburized tubes are compared in Figs.
3.3.18 aond 3.3.19. The stabilization of the grain
size as a result of carburization is evident. The
influence of carburization on mechanical properties,
corrosion, and other fabrication properties has not
yet been determined.

WEL DING AND BRAZING PROCEDURES
G. M. Slaughter

The following welding procedures, which were
developed for internal use, are now being prepared
for AEC publication:

P. Patriarca
PERIOD ENDING DECEMBER 31, 1957

 

 

 

* UNcLaAssIFiED] | .,
Y-23%981 . ——
- > -';'.\- i
5 ¥ w
2 z
- 0 0
a | =
...r,:
. .
" lo.os
- ] o
: ."43_' 0.04
- . "
A
e 0.08
‘\.Ih. -
i
. -
Fig. 3.3.19. Section Token near Brazed Fillet To
£ 7 = Show Large Grain Size of Uncarburized Tubing.
»t 3 E--' g
s ARt 1. Procedure Specification PS-1 for D-C Inert-
- "Tr vl LR Y ; i . A,
L 1 i ety L Arc Welding of Inconel Pipe, Plate, and Fittings.

 

2. Procedure Specification PS-2 for D-C Inert-

ig. e . o =To= int M d - - a .
Fig. 3.3.17. As-Brazed Tube-to-Header Joint Made Arc Welding of Inconel Tubing for High-Corrosion

o with Carburized Inconel Tubing. Hardness numbers are Nor N
" ' : Applications.
in DPH units. Etchant: glyceria regia. 75X. Reduced
11%. 3. Procedure Specification PS-8 for D-C Inert-
Arc Welding of Nickel and Low-Carbon Nickel
UNCLASSIFIED LS Pipe, Plate, and Fittings.

Y-23983

4. Procedure Specification PS-9 for D-C |nert-
Arc Welding of Nickel and Low-Carbon Nickel
Pipe, Plate, and Fittings to Inconel Pipe, Plate,
and Fittings.

5. Procedure Specification PS-12 for Metal Arc-
Welding of Inconel Plate.

6. Procedure Specification PS-14 for D-C Inert-
Arc Welding of Inconel Pipe, Plate, and Fittings
to Types 304, 316, 321, and 347 Stainless Steel
Pipe, Plate, and Fittings.

7. Procedure Specification PS-15 for D-C Inert-
Arc Welding of Hastelloy B and Hastelloy C Pipe,
Plate, and Fittings.

"

 

8. Procedure Specification PS-18 for D-C Inert-
Fig. 3.3.18. Section Token near Brazed Fillet To Arc Welding of Hastelloy B and Hastelloy C
Show Small Grain Size of Carburized Tubing. Tubing for High-Corrosion Applications.

167

 

 
ANP PROJECT PROGRESS REPORT

The corresponding Operator’'s Qualification Test
Specifications have also been developed and are
being prepared for AEC publication.

Summary reports have been published on (1) the
development of high-temperature brazing alloys and
their corrosion and oxidation resistc:nce,3 (2) the

 

8E. E. Hoffman et al., An Evaluation of the Corrosion
and Oxidation Resistance of High-Temperature Brazing
Alloys, ORNL-1934,(Oct. 23, §

168

fabrication of radiators and heat exchangers,g'l0

and (3) the developmental work on the bonding of
cermet valve components to metals. '’

 

%P, Patriarca et al., Fabrication of Heat Exchangers
and Radiators for b Temperature Reactor Ap-
plications, ORNL-1955 (fiune 14, 1956).

100, Patriarca, G. M. Slaughter, and W. D. Manly,
Welding ]. (N.Y.) 36, 1172 (1957).

Hp, Patriarca and G. M. Slaughter, ‘‘Bonding of
Cermet Valve Components to Metals,’' Welding J. (N.Y.)
(in press).

 
PERIOD ENDING DECEMBER 31, 1957

3.4. CORROSION AND MASS TRANSFER STUDIES

GENERAL CORROSION INVESTIGATIONS
E. E. Hoffman

W. H. Coock D. H. Jansen

Brazing Alloys in Fuel Mixtures
D. H. Jansen

The Hondy & Harman brazing alloys Ne. 91
(91.25% Ni—4.5% Si—2.9% B) and No, 93 (93.25% Ni—
3.5% Si—1.9% B) in button form were subjected to
100-hr corrosion tests in the fuel mixture NaF-
ZrF ,-UF , (53.5-40-6.5 mole %, fuel 44) in a see-
saw apporatus at a hot-zone temperature of 1500°F,
Both alloys showed good resistance to the fuel.
Minor constituents of the alloys were leached to a
depth of 1 to 2 mils, as shown in Fig. 3.4.1.
Nickel, because of its comparative inertness to the
fuel, was used for the test containers. Weight

losses of the alloy specimens varied from 0.03 to
0.06%.

UNCLASSIFIED {{ 003
Y-24025

 

 

Brazing Alloys in Lithium
D. H. Jansen

Iron-base brazing alloys, as they become avail-
able, are subjected to corrosion screening ftests
in lithium. The more abundant high-nickel-content
brazing alloys cannot be used in lithium because
of their limited corrosion resistance. Recently,
an 86% Fe-5% Si—-5% Cu—-4% B brazing alloy was
corrosion tested for 100 hr in lithium under dynamic
(seesaw apparatus) and static (capsule) conditions
at 1500°F, The specimens tested were brazed
type 347 stainless steel tube-to-header joints,

Under static conditions, the brazing alloy was
leached of minor constituents to a depth of 4 mils,
and there was some subsurface void formation in
the depleted region, as shown in Fig. 3.4.2, The
attack of the alloy tested in the seesaw apparatus
was similar but of greater magnitude. Subsurface
voids were present to a depth of 10 mils in one
area. The weight losses of the specimens tested

UNCLASSIFIED
¥-24026

Fig. 3.4.1. Hondy & Horman Brazing Alloys Nes. 91 (a) and 93 (&) After Exposure to NoF-ZrF ,~UF , (53.5-40-
6.5 Mole %, Fuel 44) for 100 hr in Seesaw Apparatus ot o Hot-Zone Temperature of 1500°F. Etchant: 10% oxalic

acid. 200X. Reduced 10%. (Secret with caption)

169
ANP PROJECT PROGRESS REPORT

   

8

4
.

> W
T2 e
__l_fiu.-.—::?;rr -
g X )
kol

g

Fig. 3.4.2. Brazing Alloy of 86% Fe-5% S5i-5% Cu-4%
After Exposure to Static Lithium at 1500°F for 100 hr.
lographic polishing. As polished. 150X. (Secret with caption)

in the seesaw apparatus were five times as great
as the weight losses that occurred during the
static tests,

Inconel Welded with Nitrogen Shield Gas
Dt H+ Jflnsen

In connection with the initiation of a procedure
for substituting nitrogen for helium as a shield
gas in the welding of Inconel, corrosion tests
were conducted on Inconel and Inconel-type 347
stainless steel welds made by using nitrogen as
the shield gas., It was desired to determine
whether the presence of nitrogen during welding
would produce o weldment with different corrosion
characteristics than those of an ordinary Heliarc
weld on Inconel or on Inconel—type 347 stainless
steel.

The corrosion specimens were exposed to fuel
44 and to NaK (56-44 wt %) for 500 hr in a seesaw
apparatus at a hot-zone temperature of 1500°F,
Upon completion of the tests, metallographic ex-
amination showed no attack on the welds exposed
to NaK. The attack on the welds tested in fuel

170

      
 

   
   
  
   

 

   
 

= UNCLASSIFIED [ &~
‘r 23819 .-S_
"um*' 2
e F=-
~ __h Y ;
b on—Jirs
— -
=2 “!’ fr—
_‘ ,":_"_-. "—- l--"- -:
- —
-\,’- e .
“‘v’f ‘?""-f"_‘ . —;.,fi&_.'!_‘: —
-.__‘:'-fi ..-.(\“_ o ;=-'-‘ Sae 1
‘\ S [0
e —
025
>
‘O—
S

 

B on Type 347 Stainless Steel Tube-to-Header Joint

Specimen nickel plated to preserve the edge during metal-

44 was the same as that found previously on
Heliarc welds: tested under the same conditions.
The only significant attack found on the welds
made in the presence of nitrogen was to a depth of
2 mils on the Inconel-type 347 stainless steel
weld, as shown in Fig. 3.4.3.

Cermet-Metal Joints in Na F-ZrF“-UF‘ and

in Sodium
W. H. Cook

Several methods were developed by the Welding
and Brazing Group for joining cermets to mehls,]
as was required primarily for joining specific
cermet disks and seats to Inconel for the produc-
tion of the fuel dump valves for the ART. In order
to evaluate the joints, the specimens described
in Table 3.4.1 were corrosion tested for 100 hr in
NuF-Zer-UF“ (50-46-4 mole %, fuel 30) in a

seesaw apparatus with a hot-zone temperature of

 

]G M. Slaughter, P. Patriarca, and W. D. Manly,
Bonding of Cermet Valve Camporzem.'v to Metals, ORNL-
2322 (in press).
PERIOD ENDING DECEMBER 31, 1957

UNCLASSI;‘ IED

 

 

 

- Fig. 3.4.3. Inconel-Type 347 Stainless Steel Weld Mode in the Presence of Nitrogen and Tested in NnF-ZrF4-
UF, (53.5-40-6.5 Mole %, Fuel 44) for 500 hr in Seesow Apparatus at a Hot-Zone Temperature of 1500°F. Etchant:
10% oxalic acid. 500X. (Secret with caption)

Table 3.4.1. Cermet-Inconel Joints Fabricated for Testing in Fuel 30

 

 

Joint Ne. Cermet Buffer Type of Join
1 K151A (80% TiC-20% Ni) Nickel High-temperature fusion
2 K162B (70% TiC-25% Ni—-5% Mo) Nickel High-temperature fusion
3 K94 (88% WC-12% Co) Nickel Brozed with copper
4 K94 Nickel Brozed with Coast Metals alloy

50 (93% Ni-3.5% S5i-2.5% B=1% Fe)

5 K138A (B1% TiC-19% Co) Mallery 1000 Brozed with o 60% Pd—-40% Ni alloy
(90% W—6% Ni—4% Cu)

6 KM (89% WTiCz—”% Co) Mallery 1000 Brazed with o 60% Pd-40% Ni alloy

 

171
ANP PROJECT PROGRESS REPORT

1500°F. None of the cermet-metal joints were
attacked, with the possible exception of the joint
which consisted of K94 brazed to nickel with
Coast Metals alloy No. 50. Attack on this speci-
men was doubtful, and it will be tested again.

In these joints, nickel or Mallory 1000 was used
to buffer the stresses developed by differences in
thermal expansion between the cermet and the
structural metal, Inconel. The high-temperature-
fusion joints between cermets and nickel were
made without brazing material, while the nickel
was joined to Inconel by the conventional brazing
techniques used in valve fabrication.

Specimens of joints 1, 2, and 4 were also tested
in sodium for 100 hr at 1500°F in a seesaw appa-
ratus. Results of metallographic examination and
comparisons with untested specimens showed that
these joints were not attacked. Joints 3, 5, and 6
were not tested because it was known that the
brazing materials used were not corrosion resistant
to sodium under these conditions,

Postoperative Examination of Prototype ART
Control Rod

W. H. Cook
The ART prototype control rod used in a 3000-hr

test of the control rod drive system was dis-
assembled and examined. (A discussion of the
test may be “found in Chap 1.3 of this report.)
The rod consisted of a column of porous cylinders
of rare-earth oxides in stagnant, molten sodium in
an Inconel thimbie. The test conditions simulated
those expected during ART operation, except that
heat was applied externally. In the ART the
control rod would be the heat source.

The 23 cylinders of rare-earth oxides (45-49.5%
Sm,0,, 22.5-27% Gd,0;, balance other rare-earth
oxides; apparent porosity, 46.3%) were, nominally,
]'/4 in. OD, 3/4 in, ID, and 1 in, long. They were
contained in a thin-walled Inconel thimble, which
was raised and lowered inside a second thimble
of molten sodium. Holes in the inner thimble
allowed the sodium to contact the ceramic cylin-
ders.?

The ceramic cylinders may be seen in Fig. 3.4.4
as they appeared before the test and in Fig. 3.4.5
as they appeared after the test. The white material
on the as-tested cylinders is residual sodium that

 

25, C. Shuford, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 41.

172

reacted with the room atmosphere after dis-
assembly. The broken pieces (cylinders 12 through
23, as shown in Fig. 3.4.5) were subjected to
1171 thermal cycles from a maximum of 1630 + 20°F
to a minimum of 880°F (+30°F or -20°F) during
the first 1800 hr of the test, This excessive tem-
perature change was inadvertently caused by
faulty temperature measuring and recording equip-
ment, The actual temperatures given above were
found by calibrating the faulty equipment with
propetly operating equipment. During the final
1200 hr of the 3000-hr test, cylinders 12 through
23 were subjected to 609 more cycles from a
maximum of 1140 +30°F to a minimum of 790 £ 80°F;
that is, they were subjected to the simulated ART
conditions. Cylinders 1 through 11 were never
subjected to a temperature change of more than
550°F,

All the thermal cycling was accomplished by
raising and lowering the control rod from the fully
in and fully out positions. Approximately 20 min
was required for temperatures to reach equilibrium
after a change of control-rod position. |t is be-
lieved that cylinders 12 through 23 broke because
of the excessive temperature cycling during the
first 1800 hr of the test when the temperature
changes required to reach equilibrium were about
500 to 800°F.

The results of this test indicate that these rare-
earth cylinders can withstand thermal cycles,
under the conditions described, of 550 to 650°F
from 1650 + 30°F. The most serious fault of these
cylinders appears to be spalling of the type ex-
hibited by cylinders 2, 6, and 9. At the base of
each of the shaillow holes caused by spalling,
there was a small amount of particulate material.
X-ray examinations of the particles showed the
standard pattern of the rare-earth oxides and an
additional, as yet unidentified, pattern,

The sodium that cooled the control rod during
the test was examined spectrographically, and
the quantities of rare earths found ranged from
<0.7 to <0.3 ppm. The samples were taken by
draining the sodium before disassembly of the rod.
Since the analyses indicate that the rare-earth
content of the sodium was negligible, it may be
assumed that the breaking and spalling of the
cylinders occurred without detectable fragmenta-
tion or powdering. The thin-walled Inconel thimble
held the broken pieces in place, and thus the con-
trol rod was operable despite the breakage and,

 
v

 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
PHOTO 26348

Fig. 3.4.4, Rare-Earth-Oxide Compacts for ART Control Rod. (Confidential with coption)

in nuclear operation, would have performed its
control function, because there was no change in
the location of the cylinders.

CORROSION IN DYNAMIC SYSTEMS

Lithium in Niobium Thermal-Convection Loops
E. E. Hoffman L. R. Trotter

The results of previous static tests and recently
conducted seesaw-furnace tests® of niobium in
contact with lithium at elevated temperatures have
indicated that niobium has excellent corrosion
resistance to lithium. In order to study these
materials in a dynamic system, four thermal-
convection loops constructed of sintered niobium
were recently tested. In these tests the operating
periods ranged from 23 to 500 hr, and hot-leg tem-
peratures of 1500 and 1600°F were used. The

 

3. E. Hoffman, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-2387, p 215.

details of these tests are available in a separate
re-pc:lrf.4

In several of the thermal-convection loops, un-
usually heavy grain-boundary penetration of the
niobium occurred, This type of penetration had
not been detected in previous tests and was found
only after the specimens had been polished and
allowed to stand in air for several days. The
presence of lithium was detected by heavy grain-
boundary stains that were the result of the reaction
of the lithium in the grain boundaries with the
atmosphere.  The lithium penetration extended
completely through the hot-leg tube wall (0.020 in.
thick) in three loops that were operated in a
vacuum chamber. The heaviest attack occurred in
the weld areas. Metallographic examination showed
no lithium penetration of cold-leg sections of the
loops. The presence of lithium in the hot-leg tube

 

4E. E. Hoffman and L. R. Trotter, Columbium-Lithium
Thermal Convection Loop Corrosion Tests, ORNL CF-
57-12-2 (te be issued).

173

 

 
 

ANP PROJECT PROGRESS REPORT

  
  

(5)

UNCLASSIFIED
Y-24371

19 20 21 22 23

UNCLASSIFIED
Y-24370

Fig. 3.4.5. Rare-Earth-Oxide Ceramic Cylinders After Exposure for 3000 hr to Molten Sodium in the ART Proto-
type Control Rod Drive System Tests, (a) Cylinders shown in their arrongement in the control rod. (b) Individual

cylinders showing most damaged regions. (Secret with coption)

walls was also confirmed by chemical analyses of
various layers machined from the tube wall. A
lithium concentration gradient was found, with the
concentration of lithium being greatest in layers
near the inner surface., The hot-leg tube walls
that had been penetrated by the lithium cracked
severely when bent, in contrast to cold-leg tube
wall specimens which were quite ductile after the
tests. It is thought at present that the lithium
penetration may be due to grain-boundary impuri-
ties, probably carbides, and experiments are

174

planned in which various grades of niobium will
be tested to determine what factors control grain-
boundary penetration.

Niobium loops are presently being fabricated by
Heliarc welding in an inert-atmosphere chamber,
A high-temperature brazing alloy containing 85%
Zr—15% Nb is used as the filler rod. Static cor-
rosion tests have shown that this brazing alloy
has excellent corrosion resistance to lithium at
1700°F; the static tests were of 100 hr duration.

i
Lithium in Niobium Forced-Circulation Loops
J. H, DeVYan R. S. Crouse

A forced-circulation loop is being designed for
studying lithium-niobium systems under the fol-
lowing conditions:

Lithium flow rate 1-2 gpm
Reynolds number 15,000
Maximum fluid temperature 1500-~1800°F
Temperature differential ' 200 -400°F

The niobium tubing of the loop is to be protected
from air oxidation by a continuous outer sheath or
ciadding of type 310 stainless steei. Either in-
direct resistance heating with Nichrome-wound
clamshell elements or induction heating may be
used with the clad tubing. A preliminary loop de-
sign that includes clamshell heaters is shown in
Fig. 3.4.6. An annulus filled with sodium is
provided around both the hot and cold legs in
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
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,
both clad and unclad, has been initiated, and
delivery is expected within six to eight months.
While awaiting shipment of this tubing, various
mechanical features of the loop design wili be
studied by using an all-stainless-steel prototype.

DEVELOPMENTAL STUDIES OF
INCONEL SYSTEMS

J. H. DeVan R. S. Crouse
Inconel Exposed to Fused Fluorides

From early studies of the compatibility of
structural alloys with fused fluoride salt mixtures,
it was found that nickel-base systems were the
most corrosion resistant of the many types of
alloys investigated. Consideration of the addi-
tional requirements of a structural alioy in a high-
temperature reactor system indicated that Inconel
would be the most suitable of the then available
materials. Following this selection extensive

PERIOD ENDING DECEMBER 31, 1957

<
ORNL~LR-DWG 27067

PUMP CELL

| LITHIUM SURGE TANK
COLD TRAP

INSULATION
//

| TYPE 210 STAINLESS
STEEL CLAD NIOBIUM
TUBING, '-in.ID,
0.040-in. WALL

 

 

[——* SODIUM SURGE

 

- TYPE 310 STAINLESS
STEEL TUBING, Y5-in.
0D, 0.065-in. WALL

——

— SODIUM

T S————
N

L COOLING COILS

\
|

 

T

 

— CLAMSHELL HEATERS

 

 

 

 

 

 

 

 

   

HEATED LENGTH, 7 ft

TOTAL MAIN CIRCUIT
LENGTH,~19ft

   

TYPE 310 STAINLESS
STEEL PIPE, 4-in.IPS,
SCHED 40

 

 

 

Fig. 3.4.6. Niobium-Lithium Forced-Circulation Loop.

corrosion tests were conducted in order to ascer-
tain the specific corrosion rates of Inconel in
fused fluoride salt mixtures under the various
conditions of reactor operation. The results of
such experiments both in- and out-of-pile have
appeared in previous reports in this series, as
well as in several topical reports.”>~7 These

 

5G. M. Adamson et al,, Interim Report on High Tem-
perature Corrosion With Fluoride Mixtures, ORNL-2337
(in publication).

6G. M. Adamson et al., Corrosion by Zirconium-Base
Fluorides, ORNL-2338 (in publication).

7W. D. Manly et al., Aircraft Reactor Experiment —
Metallurgical Aspects, ORNL-2349 (Dec. 20, 1957).

175
ANP PROJECT PROGRESS REPORT

studies”W&e concluded about a year ago, at which
time sufficient information had been obtained to
effectively program the operation of the ETU and
the ART with respect to expected corrosion rates.

Inconel Exposed to Sodium and to NaK

The utilization of Inconel as the primary con-
struction material for the ART and the ARE re-
quired a determination of the corrosion properties
of this alloy in sodium and in NaK over the tem-
perature range from 1000 to 1500°F. Such cor-
rosion data, which were obtained in both thermal-
convection and forced-circulation loop experiments,
have been reported in detail in previous reports
in this series. The primary result of the experi-
ments was an illustration of the very strong tem-
perature dependence of the mass transfer that

176

occurs in Inconel systems exposed to sodium and
to NaK.

The mass transfer that occurred in a 1000-hr
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
formed in the cold-leq regions of forced-circulation
systems. The effects of temperature gradient and
flow rate on mass transfer were also found to be
significant at temperatures akove 1300°F. The
rate of mass transfer showed almost linear corre-
spondence with these variables over the ranges
studied. The level of oxide impurities in sodium
and in NaK, evaluated in Inconel systems, was
also of some importance, but the effect on the rate
of corrosion was not so significant as it has been
found to be at lower temperatures in iron-base
systems.

e
 

 

PERIOD ENDING DECEMBER 31, 1957

o wn 3.5. MATERIALS FABRICATION RESEARCH

NIOBIUM FABRICATION

H. Inouye D. O. Hobson
J. H. Coobs

Recrystallization of Niobium
J. P. Page

Specimens of niobium which had been subjected
to various recrystallization treatments were ana-
lyzed by x-ray and metaliographic techniques, and
the times and temperatures necessary for the
initiation and completion of recrystallization were
determined. In addition a study was made of the
effects of recovery and recrystallization on the
room-temperature properties of niobium. Except
for the larger grain size of the recrystallized
arc-cast material, no significant differences were
found between arc-cast and powder-metaliurgy
niobium. A topical report describing the complete
investigation of the recrystallization of niobium
has been published.’

Fabrication of Niobium Loops

A preliminary design has been prepared, as
described in the previous chapter of this report
(Chap. 3.4), for a forced-circulation loop that is
to be fabricated of niobium tubing and is to be
operated with lithium as the circulated fluid. The
principal problems that will be associated with the
construction and operation of such loops will be
material procurement, oxidation protection of the
niobium, and welding. The development work
under way on the welding of niobium was described
in Chap. 3.3 of this report. Purchase orders have
been initiated for the procurement of 0.500-in.-
OD x 0.035-in.-wall tubing, 1%-in.-OD x 0.035-in.-
wall tubing, and plate, but the finished material
will probably not be received for several months.
Alternative plans for meeting the material require-
ments include the purchase of forged ingots to be
processed into tubing and of arc-cast billets to be
extruded to tube blanks. Since the proposed loops
will be exposed along part of their length to air, it
will be necessary to provide some sort of cladding
that will protect the niobium tubing from exposure.
It will be possible to clad much of the tubing with

 

5. p. Page, The Annealing Behavior of Cold-Rolled
Niobium, ORNL-2372 (Sept. 19, 1957).

stainless steel, but, because of their irregular
shapes, the electromagnetic pump, the surge tank,
and some bends would be difficult to clad by using
ordinary cladding methods. Therefore other means
are being sought for protectively cladding irregular
surfaces and shapes fabricated of niobium. In the
initial studies, several of the high-temperature,
nickel-base brazing alloys are being applied to
niobium specimens, which are then subjected to
oxidation tests. The braze metal is applied by
flame spraying, and it is then sintered in helium at
a temperature just above its melting point. Oxi-
dation tests are presently being made in a hori-
zontal tube furnace at 1700°F in static air, but,
since the specimens must be removed from the
furnace and cooled before they are weighed, long-
time oxidation tests of up to 1000 hr will be run
in vertical-tube oxidation furnaces that do not
require thermal cycling of the specimens during
weighing.

The results of some of the initial oxidation tests
of niobium specimens coated with Wall Coelmonoy
Nicrobraz or Coast Metals alloys No. 52 or 53 are
compared with the results of similar tests of type
310 stainless steel in Fig. 3.5.1. The sharp rise
in the curve for the sample coated with Nicrobraz
is the result of the failure of the cladding at two
corners of the specimen. The niobium, upon ex-
posure to air, oxidized at a very rapid rate, with the
results shown in Fig. 3.5.2. The cladding was
unable to prevent the oxide from spreading and
was rolled back upon itself. A second test of a
Nicrobraz-coated sample, in which the coating did
not fail, showed the Nicrobraz to have oxidation
resistance comparable with that of the Coast
Metals alloys. Tests will also be made of niobium
specimens that have been copper-plated and then
aluminized. Several attempts have been made to
fabricate such specimens, but each attempt has
failed because impure sintering atmospheres pre-
vented bonding between the copper and the aluminum.

Consolidation of Niobium Pellets

An experimental quantity of niobium pellets
produced by the Electro Metallurgical Company
was consolidated into a rod by extrusion. After
being stripped from the stainless steel can in
which it was extruded, the rod was cold-swaged to
a reduction of 97%, without annealing. On the

177
ANP PROJECT PROGRESS REPORT

=
ORNL—LR—DWG 253934

 

 

 

 

 

 

 

 

 

 

 

il T AT T T
0010 = o NiOBIUM COATED WITH COAST METALS ALLOY
NO.52 AND TESTED IN STATIC AIR AT 1700°F
0009 }—— |
.—=—NIOBIUM COATED WITH COAST METALS ALLOY
NO.53 AND TESTED IN STATIC AIR AT 1700°F
0008 ——
—o— NIOBIUM COATED WITH NICROBRAZ AND /
g 0007 TESTED IN STATIC AIR AT 1700°F /
2 —o- TYPE 310 STAINLESS STEEL TESTED /
< 0006 —— IN STATIC AIR AT 1800°F k
<]
1]
= 0005 NICROBRAZ COATING FAILED AFTER~30hr ___
& /
m /
= 0004
-"'l-l___----
D s e |
[ -
_..—-—.—-z:-—.:*::'fl:‘-‘-‘-l'-.l"“ === mon I —_——

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

90 00 MHO 420 430 440 150 160 170
TIME (hr)

Fig. 3.5.1. Oxidation of Niobium Coated with Various Brazing Materials.

UNCLASSIFIED
Y-24413

 

1605 ¢

r|:1:-!:|mr|||||ht|||!||=i:m|';.:.|!||.';1n |

Fig. 3.5.2. Nicrobraz-Coated Nicbium Specimen After
137 hr of Exposure to Static Air at 1700°F. (Secret

with caption)

178

basis of this experiment, five tube-blank ex-
trusions of the pellets were processed. Mild steel
cans were used for compacting and preheating the
pellets. The mild steel was removed after ex-
trusion by using a pickling solution. The extruded
blanks were then vacuum-annealed and sent to the
Superior Tube Company Yor redrawing. However,
the tube blanks failed during reduction because of
severe splitting in the joints between pellets,
Therefore no further attempts will be made to use
this technique for the production of niobium tubing.

YTTRIUM METAL PRODUCTION
T. Hikido? J. H. Coobs
Experimental studies of the production of high-

purity yttrium metal were continued. In the process
being developed,® lithium is used to reduce a

 

20, assignment from the U.S. Air Force.

31, Hikido, R. E. McDonald, and J. H. Coobs, ANP
Quar. Prog. Rep. Sept. 30, 1957, ORNL-2387, p 228.

 

 
 

mixture of YF,-MgF,-LiF to produce an yttrium-
magnesium alloy.

Tweo additional experimental runs were com-
pleted, runs L-5 and -6. In run L-5, 693 g of alloy
was produced for a yield of 67%. Difficulty was
encountered in achieving satisfactory operating
temperatures in run L-6, however, because of in-
adequate insulation, and a very poor yield of alloy
resulted.

Several lots of the Y-Mg alloy obtained in these
and previous runs were vacuum-distilled to remove
the magnesium and the excess lithium. The re-
sulting sponge was then arc-melted under an argen
atmosphere with a tungsten electrode. The results
of chemical analyses of these melts are presented
in Table 3.5.1.

Sample L-5A, which was melted as a finger
casting, was cold-swaged to a cylindrical shape and
then vacuum-annealed at 1000°C for 1,5 hr. After

surface folds were removed by machining, the

Table 3.5.1. Chemical Analysis of Arc=Melted
Yttrium Sponge

 

Semple Impurities Found (ppm)

 

Number Oxygen Nitrogen Lithium Magnesium

 

L-2A 1700 <10 970
L-2B 1400 50 65
L-4A 1900 170 <1 360
L-4B 1800 250 2 69
L-5A 1000 150 <2 480

 

PERIOD ENDING DECEMBER 31, 1957

sample was cold-swaged 37%, as shown in Fig.
3.5.3. The sample was then cold-worked from 0.41
in. to 0.09 in. in diameter, with the use of inter-
mediate vacuum anneals. The amount of cold work
between anneals was over 30%. However, it was
subsequently discovered that some cracks had
developed during the initial 37% cold-swaging
operation.

Equipment for vacuum-distilling larger quantities
of alloy (approximately 4=5 Ib) is presently being
constructed. In the next alloy-reduction experi-
ment, the transfer of molten fluorides into the re-
action retort will be tried. Heretofore the fluoride
charge has been cast into nickel containers and
then crushed to granular particles in an inert-
atmosphere dry box. The charge was loaded into
stainless steel transfer flasks, which were couplead
to the reaction retort through a Jamesbury vacuum
valve system. It is hoped that the transfer of
molten charge material will reduce contamination by
eliminating some of the handling steps.

SHIELDING MATERIALS DEVELOPMENT
J. H. Coobs

A shield plug assembly was devised to protect
the fuel pump motors of the ART from reactor
radiation. The plug was made up of three layers
assembled into an Inconel container. The layers
were a ceanse gamma-ray shield composed of
tungsten carbide bonded with Hastelloy C, a thermal
shield of low-density zirconia, and a neutron
shreld composed of boron carbide dispersed in
copper. The developmental work leading to the
final comronents and the fabrication of a full-size

UNCLASSIFIED
Y-23975

ONE INCH

 

 

Fig. 3.5.3. Cast Yttrium Sponge from Run L-5 After 37% Reduction by Cold-Swoging. (Confidential with

cupfion}

179
ANP PROJECT PROGRESS REPORT

assembly are described in a separate report.* In
addition, a report is being prepared on the de-
velopmental work that led to the selection of the
particular materials used for the neutron shield

of the ART.

TUBULAR CONTROL ROD PRODUCTION
J. H. Coobs

The feasibility study of the extrusion of control
rods was continued. Tensile bars cut from two of
the successful extrusions described previously®
were tested at room temperature and at 1500°F, and
elongations of 2 and 5% were obtained, as com-
pared with elongations of less than 0.1% for the
specimens tested previously.

CERAMIC INVESTIGATIONS
L. M. Doney

Development of High-Density BeO Bodies
R. L. Hamner

Developmental work was initiated at the request
of GE-ANP on high-density BeQO bodies for use as
moderating material. Very fine particles of BeQ
are being obtained by the controlled calcination
of BeSO,. Heating at 950°C for 1 hr in an oxidizing
atmosphere produced a mass of fluffy, white ag-
glomerates, which were broken down in a ball mill
prior to pressing. The surface area of this material
was 20 to 26 m2/g.

Cylindrical specimens were hot-pressedat 200 psi
and temperatures of 1325, 1425, and 1525°C. The
densities obtained were 91, 92, and 96% of theo-
retical, respectively. The high sinterability of
this powder is indicated by the fact that, in order
to obtain comparable densities with commercial
grades of BeO powder, it is necessary to use
temperatures 400°C higher.

Extrysion of Magnesia and Beryllia Insulators
C. E. Curtis

Rods of magnesium oxide and of beryllium oxide
were extruded to provide rod insulators with four

 

4), P. Page and J. H. Coobs, The Shield Plug
Asser)nbly for the ART Fuel Pumps, ORNL-2353 (in
press).

M. R. D'Amore, ANP Quar. Prog. Rep. [une 30,
1957, ORNL-2340, p 242.

180

longitudinal holes as required in thermocouple
assemblies. The magnesium oxide rods were
approximately 0.370 or 0.490 in. in diameter, and
the beryllium oxide rods were approximately 0.242
in. in diameter. Both types of rods were fired at
1400°C for 4 hr. The beryllium oxide rods will be
swaged in the tube assembly.

Production of Metal Hydrides
R. A. Potter

Preparations are being made for hydriding dense
yttrium metal when it becomes available. Reactor-
grade zirconium, which is available and which
responds to hydrogen in much the same manner as
yttrium does, is being used in developing the
hydriding processes.

In the method being tested currently, an air-
evacuated, sealed nickel capsule is used through
which the hydrogen must diffuse before coming
into contact with the metal specimen. The sealed
nickel capsule eliminates the necessity of main-
taining a rather elaborate gas-cleaning train and
also provides a simple method whereby the hy-
drided metal can be kept from contact with the
atmosphere until it is to be clad.

In a typical experiment, a machined, ]/2- by 3-in.,
zirconium metal rod is placed inside a length of
3{4-in.-di0, 20-mil-wall nickel tubing with molyb-
denum spacers at both ends to prevent contact
between the nickel and the sample. Endcaps are
welded in place on the tubing, and the assembly is
evacuated with a vacuum pump connected to a
]{‘-in.-dia nickel tube welded into one end cap.
The small tube is then forged closed, and the
sealed assembly is heated in a hydrogen atmosphere
at temperatures ranging from 1475 to 1850°F and
at pressures from 1 to 2 atm.

Four such encapsulated samples have been
treated under different pressure, temperature, and
time conditions. They were fired in a vacuum to
the peak temperature, hydrogen was allowed to
flow through the retort as the temperature was
dropped to a lower level, and a static pressure
was maintained throughout a soak period. In some
instances, the samples were refired in an attempt
to increase the hydrogen content. The firing
conditions and the results for the four samples
are presented in Table 3.5.2.

a8
PERIOD ENDING DECEMBER 31, 1957

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

 

 

Maximum )
Temperature Minimum Soak T?f?l Time at Hydrogen Hydrogen
- Sample -N‘umber of During Yacuum Temperature Minimum Soak Pressure Content of Sample
Number Firing Cycles Heating °F) Temperature (psig) (wt %)
- (OF) (hr)
1 1 1700 1650 3 3 0.2
2 2 1850 1605 2 3 0.2
3 3 1830 1605 5 15 0.7
4 6 1830 1475 B8 15 1.1

 

sl

181

 
ANP PROJECT PROGRESS REPORT

3.6. METALLOGRAPHIC EXAMINATIONS OF ENGINEERING TEST
COMPONENTS AFTER SERVICE

ART PROTOTYPE TEST RADIATOR NO. 1
R. J. Gray J. E. Van Cleve, Jr.

Test radiator No. 1 failed immediately after
reaching full power in its first controlled thermal
cycle. The NaK outlet temperature for the radiator
remained approximately constant during the tran-
sition, and the inlet temperature was increased
from 1350 to 1500°F at a rate of 50°F/min. The
radiator operated for a total time of 870 hr with
the NaK inlet temperature above 1000°F. The
total number of thermal cycles before the radiator
failed was 9]/2. The examination of the point of
failure was described in the previous reporf,1
and the investigation of fin-to-tube joint braze
integrity, mass transfer, corrosion, and other
items of interest are described here.

 

lR. J. Gray and J. E. Van Cleve, Jr.,, ANP Quar.
Rep. Sept. 30, 1957, ORNL-2387, p 232.

 

 

 

 

 

 

 

 

 

 

 

 

 

SAMPLE LOCATION

 

Fig. 3.6.1. Diagram of Radiator Showing Locations at Which Samples Were Taken for Evaluation of Fineto=Tube

Joint Braze Integrity.

182

Sixteen samples were removed from the radiator
at the positions shown schematically in Fig.
3.6.1 (see Fig. 3.6.10, this report, for photograph
of a radiator with areas numbered as indicated
in Fig. 3.6.1). The samples were sectioned to
expose two opposing joint areas on each tube
and then were mounted in plastic material. The
samples were then ground and polished, and
the joint areas were examined at a magnification
of 50. In Table 3.6.1, the results of the analyses
of fin-to-tube joint braze integrity and the degree
of oxidation of the fin collar are presented and
compared with results for other radiators pre-
viously examined.

Incomplete melting of the braze-alloy rings was
noted in several instances; a typical sample is
shown in Fig. 3.6.2. A partially melted ring
which did not flow over the fin collar to protect
the copper core is shown in Fig. 3.6.3. A crack
in the stainless steel cladding on the fin may

UNCLASSIFIED

 

 

-t {1-2|1-3(1-4|1I-5|{1I-6|1-7|1-8]|1-9

 

 

 

 

 

 

 

 

 

 

 

S| Sl N[ | N

 

'3

 
PERIOD ENDING DECEMBER 31, 1957

Table 3.6.1. Fin-to-Tube Joint Braze Integrity and Degree of Oxidation of Fin Collars

 

Rodiator Designation

 

 

ART
P York York York York PWA
rototype
N Ne. 3 No. 4 MNo. 5 No. 9 Na. 2
o. 1
Mumber of hours of service in the 870 361 1356 1265 1283 1199
temperature range 1000 to 1600°F
Mumber of joint areas examined 2440 2684 2757 3105 2608 3210
Percentage of joint areas having 41.1 90.4 60.7 38.4 44.6 100
75 to 100% adherence
Percentage of joint areas having B.2 59 18.8 9.6 21.2
50 to 74% odherence
Percentage of joint areas having 3.1 1.1 5.5 7.0 9.4
25 to 49% adherence
Percentage of joint areas having 47.8 2.6 15.0 45.0 24.8
0 to 24% adherence
Percentage of joint areas having 10.2 84.5 61.9 29.0 23.5 100
nonoxidized copper fins
Percentage of joing areas having 43.1 ] 19.5 13.6 8.5
slightly oxidized copper fins
Percentage of joint oreas having 46.7 8.0 18.6 57.4 68.0

heavily oxidized copper fins

 

as

also be seen at the point of maximum bending.
A similar crack in the clodding of another fin
is shown in Fig. 3.6.4. The thickness variations
of the stainless steel cladding indicate that the
bending operation may have coused the crack
in the cladding before the radiator was placed
in operation.

UNCL ASSIFIED
Y=23785

The inside and outside diameters and the wall
thicknesses of the tubes were measured at various
intervals between the header and fin matrix during
the polishing process. Variations in the outside
diameter and in the wall thickness were noted
in several instances, but all the differences could
be directly attributed to damage by the fire that
resulted from the NaK leak. There was no vari-
ation of the inside diameter that was sufficient
to indicate the presence of a stress high enough
to produce plastic flow in the tubes.

 

The tube failure which permitted the NaK leak

: _ could have been the result of at least three

e : sh b things: reaction with an insulating material,
_ damage to the tube during fit-up and brazing,
and the presence of stringers across the tube

Fig. 3.6.2. Typical Particlly Melted Braze Ring. wall. Intergranular voids of the type that would
Etchont: 10% oxalic acid. 50X. have indicated deleterious effects of strain

183
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-23786

 

Fig. 3.6.3. Broze Ring That Did Not Melt Sufficiently To Protect the Fin Collar. Note crack in stainless steel
cladding of fin ot the bend. Unetched. 50X.

INCHES

Q.02

 

0.03

 

 

 

Fig. 3.6.4. Stainless-Steel-Clad Copper Fin That Was Found To Have a Crack ot the Point of Maximum Bend-
ing. Unetched. 100X.

184

 

 
 

cycling or creep were not found. Since no signs
of incipient failure were found, the exact cause
of the tube rupture cannot be stated. The fin-to-
tube brazed joints were poor, but they were not
so bad as in some previously operated York
radiators. A job sample that was turned down
because of poor flow of the brazing alloy is
shown in Fig. 3.6.5. As may be seen, most of
the brazing alloy rings are only partially melted,
and, in some instances, there is no evidence
of flow.

Corrosion and mass transfer were negligible be-
cause of the relatively short operating time. The
surfoce of a tube that was exposed to NaK in
the hottest area is shown in Fig. 3.6.6. As may
be seen, the maximum corrosion consists only
of a general roughening of the surface. The area
of maximum mass transfer is shown in Fig.
3.6.7. The tube shown was removed from the
coolest zone of the radiator. The maximum
thickness of the deposited material is 0.004 in.

    

  

"t

1
|
_'
I
|

il < PR

PERIOD ENDING DECEMBER 31, 1957

Short slits, similar to those shown in Fig.
3.6.10, were made through the fins between each
tube at both the top and bottom of the radiator
prior to operation. These slits averaged ten
fins in depth and left a considerable amount of
unprotected copper exposed to the air stream.
Eight samples were therefore removed from the
radiator in order fo investigate the oxidation of
the copper. The positions at which these samples
were taken are shown in Fig. 3.6.8. The maximum
copper oxidation was found near the NaK inlet
header. The maximum depth of oxidation was
0.027 in., and the average in the area of maximum
oxidation was 0.021 in. The average depth of
oxidation found in the cooler outlet region was
0.007 in.

The fin that was most severely oxidized is
shown in Fig. 3.6.9. As may be seen, the stain-
less steel cladding flared out because of the
increase in volume that resulted from oxidation
of the copper.

  

  
    

e UNCL ASSIFIED
Y-23699
'1" .

 

Fig. 3.6.5. Typical Brazed Joint Job Sample. Note partially melted ond unmelted brozing-alloy rings.

185

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-23782

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 3.6.6. Areo of Maximum Corrosion of ART Prototype Raodiator No. 1. Note general roughening of surface.
Etchant: 10% oxalic acid. 200X. (Secret with caption)

INCHES

 

FEREEERELR

ol8

 

kR

ZEIGKI'

 

 

 

 

Fig. 3.6.7. Area of Moximum Moss Transfer in ART Prototype Radiator Ne. 1. Etchant: 10% oxalic acid.
200X. (Secret with caption)

186

 
e

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL—LR—DWG 27077

SAMPLE LOCATION

 

=

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ny

\

 

 

 

 

 

 

 

 

 

 

] ——

 

 

 

 

 

 

 

 

 

Fig. 3.6.8. Diagrom of Radiator Showing Locations ot Which Samples Were Taken for Evaluation of Fin Oxida-

tion.

UNCLASSIFIED
¥-23970

ERE]

S
i
-

 

T 7 -
INC:-IEJS E

L1 [ k]

 

 

B[ TTTRITT]

15?1

|
1

 

 

| [l

Fig. 3.6.9. Stoinless-Steel-Clad Copper Fin with
Unprotected Edge Showing Oxidation That Occurred
During Operation of ART Prototype Radiater MNe. L.
Unetched., 150X. Reduced 24.5%. (Secret with caption)

ART PROTOTYPE TEST RADIATOR NO. 2
R. J. Gray J. E. Van Cleve, Jr.

Test radiator No. 2 failed during the temperature-
differential phase of the 182nd thermal cycle. The
failure and the fire that resulted because of gross
leakage of NaK te air were confined to the tubes
between the NaK inlet header and the fin matrix.
The operating history of the radiator is sum-
marized below:

Total hours ot 1200°F or above 862

Total hours of operation with a 431

temperature differential

Total hours at design temperature 310
differential
Number of successful thermal cycles 181

The radiator was prepared for examination in
the same manner as that used previously for
examination of test radiator No. 1. The cleaned
radiator is shown in Fig. 3.6.10 prior to sectioning
for examination. The failed tubes were found in
sections |-8 and |-7. Each section contains 20
tubes arranged in a four-by-five matrix. Twelve

187

 
ANP PROJECT PROGRESS REPORT

i

5
~
el

'
F—h

ket isfilet i ] b }

I
|

R e e e
T s it b

Sy el ke

:

 

UNCLASSIFIED
T- 1!5“6':

Fig. 3.6.10. ART Prototype Test Radiator No. 2 After It Had Been Cleaned. (mmm» with coption)

of the 40 tubes in the two sections were found
to have perforations.

The two sections that contained the damaged
tubes are shown in Figs. 3.6.11 and 3.6.12.
In the most severely damaged section, |-7, ten
of the 20 tubes were found to be damaged. All
40 of the tubes in the two sections were examined
for evidence of incipient failure. The microstruc-
ture that was typical of 39 of the 40 tubes is
shown in Fig. 3.6.13. As may be seen, there
is no evidence of incipient failure. The micro-
structure of the one tube that showed signs of
incipient failure is shown in Figs. 3.6.14, 3.6.15,
and 3.6.16.

188

Although only one of the 40 tubes examined
showed evidence of incipient failure, it is felt
that this evidence is reliable and that the original
failure occurred in this tube. The other tubes
were probably ruptured by the fire that occurred
because of the leakage of NaK.

Similar incipient failure has also been found
in creep test specimens. The microstructure of
a specimen that was tested in argon ot 1500°F
and a stress of 3000 psi is shown in Fig. 3.6.17;
the tube ruptured after 374 hr of testing. Note
the fissures at the grain boundaries and the
striking similarity to Figs. 3.6.14 and 3.6.15.

ne
PERIOD ENDING DECEMBER 31, 1957

> = B UNCLASSIFIED
Y-24193

 

.
ey
1NN
a
. Fig. 3.6.11. Fire Damage in Section I-7.
2 . UNCLASSIFIED

Y-24194

' L e
Tk < i

s

R

i T

Fig. 3.6.12. Fire Damage in Section I-8.

189
061

 

= : UNCLASSIFIED
: ) Y-24374

 

\
E R

=
T
IN'C|HE5

 

 

 

 

 

 

 

 

 

 

 

" @ P‘J v \ : 0oT
./‘ 2 < _____-—""""

f - - _.£A‘"" k) " #
{ g / Looe
e ,l’“"/ ey 010
o ‘,.-.:—...__‘1\ 1“- 1"

- S I /: = :l'
F J i 02
E 23 } 57 RS i 018
| [ jl/ e \ 014

\

l. - T . - '\ 018
- p £ ]

W ¥ .-'l. » - » -
! QIT
,-"’/- > _\'x__ - i 018

 

 

 

 

 

 

 

 

 

f f
i _x—
[ o R “ e o
~ 9 1 4 -g—
- A0 - 4
\ \' 022
f " i o
i | k ~ l.oz8
\ f o w ¢ |- el = |.o2a
3 ) .
3 B3 | : 028
' - - -

Fig. 3.6.13.
Etchant:

Typical Microstructure Found in 39 of the 40 Tubes Ex-
10% oxalic acid. 200X.

omined,

-

Fig. 3.6,14.
That Ruptured.

-
N
> \
\
- ___A..'" -

UNCL ASSIFIED
Y-24276

 

 

EER]

INCHES

T
]

 

 

 

 

 

 

Microstructure Indicative of Incipient Failure in a Tube

Etchant:

10% oxalic acid. 200X,

1L¥0d3Y S§53¥2908d LO3rodd dNV

 
 

PERIOD ENDING DECEMBER 31. 1957

UNCLASSIFIED UNCL ASSIF IED
Y-24275 ¥-14808

  

BRE

.‘_-

T T
INCHES
Tk

EEEEEEE.

i
i
=

k1

¢

BEE

T
1

T
aox
1

 

 

EEEEI:

 

 

Figs 3.6.15. Another Area of the Tube Shown in Fig. 3.6.17. Microstructure of Inconel Creep-Test
Fig. 3.6.14. Etchant: 10% oxalic acid. 200X. Re- Specimen Showing Incipient Fissures at Grain Bound-
duced 30%. aries. 100X. Reduced 33%.

  
  

  

UNCLASSIFIED § | 007/
Y.24274 ey

 

 

 

 

 

 

 

 

Fig. 3.6.16. Microstructure Closer to the Hole in Tube Shown in Figs. 3.6.14 and 3.6.15. Etchant: 10% oxalic
acid. 200X.

191

 

 
ANP PROJECT PROGRESS REPORT

3.7. NONDESTRUCTIVE TESTING
R. B. Oliver

EDDY-CURRENT THICKNESS MEASUREMENTS

R. B. Oliver
R. A. Nance J. W. Allen

The inherent accuracy of the new eddy-current
thickness gage' was increased by the substitution
of drift-free transistor oscillators for the variable-
frequency signal generator previously employed.
Although transistors are sensitive to thermal
changes, they have a high degree of frequency
stability if a uniform temperature is maintained.

The study of eddy-current cladding-thickness de-
terminations has been continued, with particular
emphasis on alleviating the problems encountered
with the probe coils currently in use. These
coils are contained in commercially procured,
high-permeability, low-loss cup cores, approxi-
mately % in. in diometer. The electromagnetic
field produced by the coil investigates an area
1/2 in. in diameter, which is much larger than
the area of minimum cladding thickness sought.
Since the eddy-current reading is a function
of the area and of the minimum cladding thick-
ness, it is difficult to obtain a true evaluation
of the cladding thickness with this coil. Some
cladding-thickness measurements have been satis-
factory, but the error in a number of cases has
been excessive. Smaller probe coils are being
fabricated to remedy this problem. The accuracy
of the determination will increase as the area
of the field of the coil approaches the area of
minimum cladding thickness. It is also difficult
to locate the exact area of minimum cladding for
metallographic examination because of the asym-
metrical nature of the electromagnetic field pro-
duced by the present probe coils. Therefore the
small cup cores are being constructed to produce
symmetrical fields in addition to decreasing the
size of the field.

A study has been initiated to determine the
feasibility of measuring the thickness of hot-
dipped aluminum coatings on aluminized steel.
Preliminary tests indicate that a nominal coating
thickness of 0.001 in. can be measured to within

0.0002 in.

 

]R. B. Oliver, J. W. Allen, and R. A. Nance, ANP
Quar. Prog. Rep. Sept. 30, 1957, ORNL-2387, p 247.

192

REMOTE X«RAY VIEWING
R. B. Oliver J. W. Allen

A closed-circuit television system (RCA model
ITV-6) for direct x-ray imaging has been installed
in the Nondestructive Test Laboratory, and the
initial adjustment and alignment have been accom-
plished. Preliminary studies have been made on
the system to evaluate it for practical inspection
use.

The system consists of a conventional industrial-
type Vidicon camera and television monitor con-
nected in a closed video circuit. The conven-
tional light-sensitive Vidicon camera tube was
replaced with an x-ray-sensitive tube having a
selenium target (photoconductive receptor) ap-
proximately ten times thicker. The thicker target
allows greater absorption of x rays and thus
greater x-ray sensitivity.

The technique for use of this system is similar
to that used in conventional radiography. The
object to be examined is placed near the photo-
conductive face of the Vidicon tube, and the
source of x radiation is placed at some distance
from the object. The system is, therefore, subject
to the same geometrical considerations as those
of the conventional radiographic method. An
inherent image magnification of approximately 12
diameters is obtained, since the ]/2-by 3/4-in.
roster scanned on the Vidicon target is reproduced
as a 6- by 9-in. roster on the kinescope.

Preliminary studies of the sensitivity and
resolution of the system have been encouraging.
The exposure and contrast sensitivity, although
not so great as those attained with film, are
adequate for many applications. Since the ex-
posure of the photoconductive target is continuous
rather than accumulative, as it is with a photo-
emulsion, it is inherently less sensitive than
film. The greatest area of usefulness of the
system appears to be in the inspection of thin
sections through which high-intensity radiation can
reach the photoconductive target.

A very high degree of resolution is attainable
with this system in comparison with other types
of direct or remote x-ray image-viewing systems;
the resolution is not limited by a mosaic of
fluorescing crystals, as is the case with a

av

 

 
 

fluorescent screen. The system will clearly
resolve the x-ray image of a 500-mesh screen,
and thus it is possible that resolution in excess
of 500 lines per inch may be achieved. The
inherent magnification appears to improve the
resolution of fine detail, as illustrated in Fig.
3.7.1, which is a photograph of the television
screen showing the x-ray image of a small point-
contact semiconductor diode. This image, al-
though trivial, illustrates the desirability of
high-resolution image magnification. The wire of
the diode is less than 0.005 in. in diameter and
terminates on a germanium wafer at the lower
extremity, the assembly being contained in a
glass envelope (light-gray area). In another
test, a fine crack in a Y-in.-dia, 0.025-in.-wall
INOR-8 tube was readily observed; the crack
was less than %-in. long.

It is expected that the detection of small
defects in small-diameter tubing will be improved

 

PERIOD ENDING DECEMBER 31, 1957

by rotating the tube during observation. The
evaluation of this technique and the contrast
sensitivity on small-diameter tubing has been
hampered by the lack of suitable remotely con-
trolled mechanical equipment. Accordingly, an
existing skew-roll feed mechanism is being
modified to accommodate the x-ray Vidicon
system. This mechanism will permit evaluation
of the method for the inspection of tubing to
proceed quickly and effectively.

MATERIALS INSPECTION
R. W. McClung
Tubing

Approximately 13,000 ft of small-diameter CX-900
Inconel tubing was inspected during the last
quarter by the immersed-ultrasound and the

UNCLASSIFIED
PHOTO 42833

Fig. 3.7.1. Radiogroph of a Semiconductor Diode (%-’lé-in.-diu, 5/m--ifl.-lu:mg_:,) Made by Photographing the Menitor

Screen of the X-Ray-Sensitive Television System.

193
ANP PROJECT PROGRESS REPORT

encircling-coil eddy-current techniques; the re-
jection rate was about 5%. The total length
of the tubing inspected for the ANP program was
about 120,000 ft.
Pipe

Approximately 1500 ft of Inconel pipe in sizes
varying from F@-in. sched-40 up to 3%-in. sched-
40 was inspected by the immersed-ultrasound
method; the rejection rate was less than 5%.
About 450 ft of 31’§'i“' sched-40 pipe was also
inspected by the resonance-ultrasound method in
order to accurately determine the wall thickness
throughout each length. The total inspected
length of pipe has been approximately 8250 ft.

Plate

The immersed-ultrasound technique was used
during the last quarter to inspect approximately
375 ft? of Inconel plate in thicknesses varying
from }2 to 1 in. This brought the total inspected
quantity up to about 1500 ft?. The rejection rate
has averaged 1% or less and has varied according
to the location of the defects and the possibility

 

of their removal when the plates were cut to
size.

Hastelloy B Tubing

A special lot of Hastelloy B tubing was fab-
ricated by the Republic Steel Corp., Cleveland,
Ohio, in order to determine their potentiality as
suppliers of this type of material. The tubing
was of three sizes, %-in., % -in., and %-’B-in. 0OD,

4 16
but the wall thickness of each size of tubing
was 0.030 in. It was known that a few gross

flaws were present, and therefore the prime
interest was in the remainder of the tubing.
In general, it was found to be among the best
of the Hastelloy B tubing inspected to date.
The quality of the larger diameter tubing seemed
to be better than that of the ‘;fi-in.-OD tubing.
Among the rejectable conditions noted were
cracks alongside the weld bead of the )-in.-OD
tubing shown in Fig. 3.7.2. The excessive weld

penetration found in the 3.fé-in.-0D' tubing shown
in Fig. 3.7.3 was also considered to be un-
desirable. Pores found on the inner surface

of the 'T'fm-in.-OD tubing are shown in Fig. 3.7.4.

UNCLASSIFIED
Y-23868 6.0l

INCHES

-
o
»

 

0.08

 

Fig. 3.7.2. Weldrawn Hastelloy B Tubing Showing Cracks at the Root of the Weld Bead. Unetched. 50X.

194

 
 

 

 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
Y-23867

 

 

0.08

 

C.08

 

0.07

 

50X

 

 

 

 

Fig. 3.7.3. Weldrawn Hastelloy B Tubing Showing the Results of Excessive Penetration of the Weld Bead.

Unetched. 50X.
E==
¥Y-23876

 

N

 

 

Fig. 3.7.4. Microsection of Weldrown Hastelloy B Tubing Showing Pores on the Inner Surface. Unetched. 500X.

195
Part 4
RADIATION DAMAGE
G. W. Keilholtz
 
 

 

ak

4.1. RADIATION DAMAGE
G. W. Keilholtz

EXAMINATION OF IRRADIATED
COMPONENTS AND MATERIALS

A. E. Richt
N. A. Carter W. B. Parsley
C. Ellis E. D. Sims

E. J. Manthes R. M. Wallace

MTR In-Pile Loops

Examinations of the four forced-circulation loops
that were operated in the MTR with fused fluoride
fuels as the circulated fluids have been completed. '
The impeller housings from loops Nos. 4, 5, and 6,
which were the last items to be examined, were
opened on the remotely operated lathe. The re-
sults of the examination of the impeller from loop
No. 3 were described previcus|y.2 The impellers
could not be removed from the housings because of
a shoulder on the pump shaft, and therefore only
the front faces of the impellers could be examined.
No evidences of fuel were found in the front halves
of the impellers from loops Nos. 4 and 5; however,
there was a fuel deposit in the bottom of the
housing of the impeller from loop No. 6. As may
be seen in Figs. 4.1.1, 4.1.2, and 4.1.3, the

 

TEor descripticns of these loops and the conditions
under which they were operated, see Part 1 of previcus
reports in this series.

2A. E. Richt et al., ANP Quar. Prog. Rep. Dec. 31,
1956, ORNL-2221, p 290.

UNCLASSIFIED
RMG-1945

 

Fig. 4.1.1. Teeth of Impeller from Pump Used in MTR
In-Pile Loop No. 4. 4X. Reduced 11%.

teeth of the impellers from loops Nos. 4, 5, and 6
showed no evidences of erosion. The impeller
housing cover plates and the forward halves of
the impellers from loops Nos. 4, 5, and 6 are
shown in Figs. 4.1.4 through 4.1.9. The fuel de-
posit that may be seen in Fig. 4.1.9 at the bottom
of the housing of the impeller from loop No. 6 is
shown againin Fig.4.1.10 ata higher magnification.
No metallographic specimens were taken from any
of these impellers.

    

o - -,
e —

- - ". —
- S UNCLASSIFIED
E:_ x = “= % RMG-1946

Fig. 4.1.2. Teeth of Impeller from Pump Used in MTR
In-Pile Loop No. 5. 4X. Reduced 11%.

 

Fig. 4.1.3. Teeth of Impeller from Pump Used in MTR
In-Pile Loop No. 6. 4X. Reduced 11%.

199

 
ANP PROJECT PROGRESS REPORT

Moderator Materials

Disassembly and examination of the capsules
irradiated in the second exposure in the MTR of
the high-temperature moderator-materials testing
assembly were completed, except for further
metallographic examination of the molybdenum
capsule that contained zirconium hydride. Three
capsules were irradiated: one was beryllium oxide

UNCLASSIFIED
RMG-1948

 

 

Fig. 4.1.4. Cover Plate of Housing for Impeller of
Pump Used in MTR In-Pile Loop No. 4. /X.

UNCLASSIFIED
RMG-194%

Fig. 4.1.5. Forward Face of Impeller from Pump Used
in MTR In-Pile Loop No. 4. JX.

200

encased in Inconel; another was graphite encased
in nickel; and the third was zirconium hydride en-
cased in molybdenum. The results of the exami-
nation of the BeO-containing capsule were de-
scribed previously;® the examinations of the
other two capsules are described here.

 

3A. E. Richt et al., ANP Quar. Prog. Rep. Sept. 30,
1957, ORNL-2387, p 257.

 

Fig. 4.1.6. Cover Plate of Housing for Impeller of

Pump Used in MTR In-Pile Loop No. 5. !f?x.

 

Fig. 4.1.7. Forward Face of Impeller from Pump Used

in MTR In-Pile Loop Mo. 5. ]’ZX.

e

 
The nickel jacket of the graphite-containing
capsule was removed by milling the jacket longi-
tudinally and then peeling it from the graphite
slug. No cracking or flaking of the graphite slug
was visible after the jacket was removed. The
diameters of the four graphite slugs were found to
be 1.2498, 1.2505, 1.2502, and 1.2502 in., which
compared favorably with the original slug diameter
of 1.251 in. Metallographic examination of the
nickel iacket revealed that no reaction between the

UNCLASSIFIED
: RMG-1952

 

o iR

1 ir
&y fi'-'fi\'.fit&: I‘I-::: SN

&

Y W

Fig. 4.1.8. Cover Plate of Housing for Impeller of

Pump Used in MTR In-Pile Loop No. 6. !‘zx.

 

Fig. 4.1.9. Forward Face of Impeller from Pump Used

in MTR In-Pile Loop No. 6. ) X.

PERIOD ENDING DECEMBER 31, 1957

    

UNCL ASSIFIED
RMG- 1554

 

 

Fig. 4.1.10, Fuel Deposit Indicated by Arrow in
Fig. 4.1.9 at a Higher Magnification. 4X.

graphite and the nickel had occurred, as shown in
Fig. 4.1.11. The outer surface showed void for-
mation to a depth of 8 mils. The small cracks on
the inner surface of the jacket were probably formed
while the sample was being handled during cutting
and mounting. The Inconel jacket from the BeO-
containing capsule examined previously likewise
did not react with the moderator material, as may
be seen in Fig. 4.1.12.

The zirconium hydride capsule was sectioned on
the remote lathe and milling machine. During
sectioning of the capsule, the molybdenum jacket
separated from the zirconium hydride slug; thus, it
appears that there was little or no bonding between
the jacket and the slug. Zirconium hydride chips
were collected during machining and submitted for
chemical analysis.

The zirconium hydride in the brazed end of the
capsule was damaged, as shown in Fig. 4.1.13.
The remainder of the slug was undamaged and
appeared to be structurally intact. A control
capsule showed similar damage of the zirconium
hydride at the brazed end; thus, it appears that
the damage was the result of poor fabrication
techniques rather than irradiation. The molyb-
denum jacket was very brittle and broke easily.
There was no evidence, however, of reaction with
the zirconium hydride on the inner surface of the
molybdenum, as shown in Fig. 4.1.14.

A section of the control capsule with the molyb-
denum jacket and the zirconium hydride slug in-
tact and a section of the irradiated zirconium

201

 
 

ANP PROJECT PROGRESS REPORT

 

. e R o ."‘St__ | UNCLASSIFIED
f J &v. u-\ : EMG".I 955
o & « coup. P __'
: i E x » . 3 "_-'-"‘0'
. . l. h ‘...‘ * R -
. - -
m ,'- H . m
: ; o T I '
E -~ » 5 .:* v 5. .! ; :
e > 22
- = ..' '
\ g y - _F ' *a =
Z \‘/\\,‘ . 3 ? et o P O
s iy G . '. ‘o i ' I‘I m
e :’ ‘.o : - k:
LS e gy T
. S0 S
» ‘;-.
L e
. y s a';’.
o . & .4,
2 ¢ Yo
| > " - -

 

et

J . . }
fr,‘\ ¥/ I

Fig. 4.1.12. Inner Surface of Inconel Jacket Removed from the Be0-Containing Capsule That Was lIrradiated
in the MTR. 250X.

202

 
wo

hydride slugare being examined metallographically.
The irradiated and the control samples appear to be
similar, except that the needlelike phase is
coarser in the control slug than in the irradiated
slug, and, in contrast to those in the irradiated

UNCLASSIFIED
RMG-1957

 

| —

Fig. 4.1.13. Zirconium Hydride Slug Showing Damage
Found ot Brozed End of Copsule Upon Removal of
Molybdenum Jocket After Irradiation in the MTR. This
damage was caoused by poor fabrication techniques
rather than irradiation, 4X. Reduced 17%.

Ry ¥
SR

PERIOD ENDING DECEMBER 31, 1957

slug, the needles in the control slug are in bundles
in the matrix rather than being distributed uniformly.
In both samples there is a third phase interspersed
among the needles, as shown in Figs. 4.1.15
through 4.1.18. The control sample was anodized
to enhance the contrast between the two phases.
Attempts to anodize the irradiated sample have
been only partially successful.

Direct comparisons of the structures of the
irradiated and unirradiated specimens cannot be
made because the unirradioted specimen was
overheated during the brazing cycle. It is not
known at this time whether the structural dif-
ferences evident in Figs. 4.1.15 through 4.1.18
are the result of irradiation or of variations during
fabrication; further, they may possibly be normal
structural variations.

Complex reactions were found to have occurred
at the interface of the zirconium hydride and the
iron layer used between the zirconium hydride and
the molybdenum jacket, and at the interface of the
iron and the molybdenum. The control capsule
exhibited three distinct reaction zones, as may be
seen in Fig. 4.1.19. The irradicted specimen
appeared to have only two reaction zones, as

shown in Fig. 4.1.20. Further metallographic

       
 

P "“‘_ *:'...,'_._'_;.r.{\._- 4#:—:..,,"';" UNCLASSIFIED *

Fig. 4.1.14. Interior Surface of Molybdenum Jocket Removed from on lrradiated Zirconium Hydride—Containing

Caopsule. Specimen nickel-plated for edge preservation. 250X.

203
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
RMG-1958

  

 

 

 
    

 

   

7
i "_, f‘
- I" : ‘I_
ktr')-f e
=r
gy L
e
= ""»I.“ ."
. ' al 20
=8 Iy . = e
¥ = - — - § .
: "'-""‘n.--'-'-'- ~'f Oxh [ i /"' X
- = - & " ~ - r
J.. .k"r' ‘:‘4’ )',; ‘.\‘ ‘r____ /'-. - -
Sty '\ f"/‘ o E S e ’
o - i £
|"." Jf ‘-’l = »I"' e .. -:!' 5 h" o= f I

UNCLASSIFIED
“  RMG-1959 |

 

Fig. 4.1.16. Structure of Unirradiated Zirconium Hydride, as Anodized. 250X.

204

!

ar

 
PERIOD ENDING DECEMBER 31, 1957

’ S - . /f F = UNCLASSIFIED
w, £ / : 7 RMG-1960

 

- Fig. 4.1.17. Structure of Irradiated Zirconium Hydride Shown at a Higher Magnification Than That Used for
5 Fig- 4-'&15- 500}(.
-

UNCLASSIFIED
RMG-1961

 

Fig. 4.1.18. Structure of Unirradiated Zirconium Hydride Shown at o Higher Magnification Than That Used for
Fig. 4.1.16. 500X.

205

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
RMG- 1962

o

  

As polished.

Zirconium Hydride—lron ond Iron-Molybdenum Interfaces of Unirradiated Capsule.

Figl‘ 4- 1-1?.
S00X.

 

B & l,' A ..‘r’\ _E,‘ g / o . s oy,
.’ s - L . f | / ._
"ln /[ {1~ : v ¥ AL . ' o .

Fig. 4.1.20. Zirconium Hydride—lron and lron-Molybdenum Interfaces of Irradiated Capsule. As polished. 500X.

206
 

examinations of both specimens are being carried
out in an attempt to determine the cause of the
structural variations observed in the zirconium
hydride matrix.

CREEP AND STRESS-RUPTURE
TESTS OF INCONEL

J. C. Wilsen
W. E. Brundage N. E. Hinkle
W. W. Davis J. C. Zukas

MTR Experiments

The out-of-pile test was completed that corre-
sponded to the MTR tube-burst tests conducted
previously.  The apparatus for the out-of-pile
test as well as the operating conditions duplicated
those of the in-pile test except for the absence of
radiation. Seven of the eight out-of-pile speci-
mens ruptured before the test was discontinued at
the end of a 304-hr period. The times to rupture,

as shown in Table 4.1.1, were two to five times

 

4], C. Wilson et al,, ANP Quar. Prog. Rep. June 30,
1957, ORNL-2340, p 268.

PERIOD ENDING DECEMBER 31, 1957

longer than in the MTR tests, but the times were
still of the order of a factor of 4 shorter than in
tests conducted by the Metallurgy Division.?

As may be seen from the data in Table 4.1.1, the
strain at fracture in the out-of-pile specimens was
greater than that in the corresponding in-pile
specimens. The in-pile apparatus was disassembled
and the specimens were checked for fracture
location. Metallographic examination of both out-
of-pile and in-pile specimens will be carried out in
order to establish any differences in the mode of
fracture.

The thermocouple wires used in the out-of-pile
test were brittle after the test. In order to de-
termine whether a change in thermocouple charac-
teristics accompanied the loss of ductility, tests
on the constancy of thermocouple calibration were
carried out in helium in an Inconel tube. The
thermocouples were located in air and in helium.

 

5). R. Weir, Jr., D. A, Douglas, and W, D, Manly,
Inconel as a Structural Material {or a High-Temperature
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 Rupture Elongation
{in) Number (psi) Test (hr) (%)
0.030 21 3000 Inepile** 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 17 141
16 4000 In-pile 33 0.75
27 4000 Out-of-pile 78 4e4
31 4000 Out-of-pile 100 5
17 5000 tn-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 Inepile 29 0.96
36 4000 Out-ofepile 105 746
34 4000 Out-of-pile 115 7.6

 

*Original inside diameter = 0.51000 * 0.00025 in.

**Fast flux = 2 X ]0]3 neufrons/cmz-sec (above 1 Mev).

207
ANP PROJECT PROGRESS REPORT

Also a test is being made of some of the fused-
quartz insulation used in the in-pile aund out-of-
pile tests. Preliminary information indicates that
errors of the order of 100°F (below true temperature)
in thermocouple readings may occur after 250 hr
for Chromel-Alumel thermocouples. Iron-constantan
couples are affected to a lesser extent in the
opposite direction. In a test without the fused-
quartz insulation, no disparity between thermo-
couples located in air and in helium was found
after 500 hr at temperatures of 1500°F and above,
and a single tube-burst specimen (at a stress of
4000 psi) in the helium side of the apparatus
ruptured at a period corresponding with the values
obtained in the Metallurgy Division. This suggests
that the fused-quartz insulation supplies a re-
ducing atmosphere that affects the calibration of
the Chromel-Alumel couples so that the specimens
may be at a higher temperature than the thermo-
couples indicate.

Thus it appears that part of the reason for the
reduced times to rupture observed in the in-pile
and out-of-pile tests is the loss of calibration of
the thermocouples. Unless radiation changed the
rate of deterioration of the thermocouples, it is
still not clear why such large ductility changes
were observed.

LITR Experiments

In the most recent LITR test, an Inconel tube
stressed at 2000 psi and exposed to the fused
salt fuel NaF-ZrF ,-UF, (62.5-12.5-25 mole %) has
been in hole HB-3 for over 620 hr without rupture,
in contrast to tubes tested previously which failed
after 100 hr or less. Postirradiation examination
of the tubes previously tested revealed a leak in
one of the Swagelok fittings on the pressure side
of one of the specimens and a defective high-
pressure seal on the thermocouples of another
specimen. In the apparatus being used at present
there are no Swagelok fittings, and all seals were
checked at double the operating pressure. A pure
cobalt foil was inserted inside the specimen tube
to measure directly the flux depression.

A tube-burst specimen (0.010 in. wall thickness)
pressurized to a stress of 3000 psi with helium was
subjected to thermal cycling between 1100 and
1500°F in air, out-of-pile. A joint brazed with
Coast Metals alloy No. 52, near the specimen but
at a temperature approximately 100°F lower than

208

that of the test section, was temperature-cycled
over 4000 times in 1000 hr without failure, but the
test section finally ruptured. The results of this
test indicate that cyclic stresses are not the cause
of the cracks that have been found in brazed joints
in the in-pile rigs.

Three tube-burst specimens tested previously
in hole HB-3 of the LITR at a constant stress of
2000 psi at 1500°F, with both sides in helium,
were examined. One ruptured after 1050 hr, and
showed an increase of 1.1% in inside diameter. A
second ruptured at 1650 hr, and the inside diameter
increased about 0.9%. The third specimen had not
failed when the test was terminated at 1675 hr,
but it showed an increase of about 0.9% in inside
diameter. A fourth specimen, stressed at 1000 psi
at 1500°F for 1675 hr in the reactor, also did not
rupture, but it showed an increase of 0.56% in
inside diameter. The test sections are now being
examined metallographically. The low ductility,
as inthe case of the MTR tests, cannot be accounted
for at this time.

ORR Experiments

The design of a facility for mounting experi-
ments on the pool face of the Oak Ridge Research
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
designed. Apparatus for conducting flux and
gamma-heating surveys was also designed.

LITR YERTICAL FUSED-SALT LOOP

W. E. Browning R. P. Shields

J.E. Lee

Disassembly of the vertical in-pile loop which
circulated fused-salt fuel for 235 hr in the LITR
was completed, and various parts are being ex-
amined and analyzed. One-inch lengths of fuel
tube were cut from each thermocouple location,
and other fuel and metal samples were taken from
the pump region. Particular attention was given
to the pump bearing. As stated in the previous
report,® the lower bearing was locked and the
bearing just below the motor was difficult to turn.
These bearings were separated from the housing
and shaft, and the grease seals were removed.

 

6W. E. Browning et al., ANP Quar. Prog. Rep. Sept. 30,
1957, ORNL-2387, p 270.

 

 
The grease has decomposed and become waxlike,
and it had been pushed to the side away from the
ball bearings. Samples of this damaged lubricant
were taken to be analyzed for fission-gas decay
products and for analysis by the California Re-
search Corporation. An attempt is being made to
degrease and decontaminate the bearings so they
can be cut apart and examined.

An induction furnace {Fig. 4.1.21) has been set
up for melting salt samples from the inconel fuel-
tube specimens. The melt-out chamber is glass,

UNCLASSIFIED
ORNL-LR-DWG 25387

INSTRUMENT

VACUUM

I

He

  

16 in. ALL PYREX MELT-OUT ASSEMBLY

INSTALLED BEHIND SHIELDING

TAPERED QUARTZ SLEEVE

BAAAARA

Vi

TUBING SAMPLE

INDUCTION HEATER

 

PLATINUM CRUCIBLE

 

 

Fig. 4.1.21. Induction Furnace for Removing Fluoride

Fuel from lnconel Tubing.

PERIOD ENDING DECEMBER 31, 1957

and it can be evacuated and filled with helium to
prevent contamination of the salt by the atmosphere.
When the Inconel tube walis are heated quickly, the
salt slides out as a solid pellet. Tests with un-
irradiated samples have shown this system to be
satisfactory. Salt samples from irradiated tubing
will be analyzed chemically, and the tube samples
will be examined metallographically in order to
evaluate the corrosion of the loop that occurred
during the test.

MTR STATIC CORROSION TESTS
W. E. Browning H. L. Hemphill

Three capsules containing fused-salt fuels
which had been irradiated in the MTR were dis-
mantied and examined metallographically. The
results of the examinations are presented in
Table 4.1.2. The Inconel capsule showed the
expected corrosion attack. The two Hastelloy B
capsules were found to have thin films on the
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
capsules was negligible.

Two more Hastelloy B capsules containing
NaF-KF-LiF-UF, fuel were withdrawn from the
MTR. One of these had been irradiated for 2496 hr
at 1250°F and at an initial power density of
6000 w/cm3. There was no indication of failure
when the capsule was withdrawn.

IRRADIATION OF MODERATOR MATERIALS
FOR USE AT HIGH TEMPERATURES

W. E. Browning R. P. Shields

Apparatus for testing beryllium, beryllium oxide,
and yttrium hydride at high gamma-heating densities

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

 

Duration of

 

Capsule Cantainer Fuel Exposure Metallographic Results
Number Material
(hr)
331 Inconel Na F-Zr|=4-UF4 304 1 to 2 mils of penetration
(5345-40-6+5 mole %)
364 Hastelloy B LiF-NqF-l‘(F-UF4 138 Film deposited; penetration negligible
{4443+11:8-3842-5.77 mole %)
335 Hastelloy B LiF-NuF-KF--UF4 119 Film deposited; penetration negligible

(4443-11,8-3842-5.77 mole %)

 

209
ANP PROJECT PROGRESS REPORT

in the ETR is being prepared. Those moderator
materials are to be irradiated with neutrons and
gamma rays in the ETR where the heating rate
will be 25 w/g. A temperature of about 1450°F
will be monitored by a thermocouple on each side
of the capsule and by one inside a well in one
piece of each type of specimen.

All specimens for the test, except the BeO,
have been fabricated and then encapsulated in
Inconel. Capsules for the BeO were completed and
the BeO specimens are complete except for final
checks. Extra beryllium specimens and capsules
were made and encapsulated. A thin molybdenum
foil was wrapped around each pellet, and a disk
was cut to cover the end of each pellet. One set
of the completed beryllium specimens will be
heat cycled by the Metallurgy Division, which
fabricated all the specimens. The heat-cycled
specimens will be examined to determine the com-
patibility of the metals at high temperature.

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-
ment. Details of the gas system for conducting
heat from the specimens to the capsule walls are
being worked out with ETR personnel.

IRRADIATION EFFECTS IN
PHOTOCONDUCTIVE DETECTORS

J. C. Pigg R. A. Zingaro

An initial investigation has been completed of
the effects of neutron and gamma radiation on
chemically deposited thin films of lead sulfide,
which are highly sensitive to infrared radiation
and have a wide variety of commercial and military
applications. The polycrystalline structure of the
films and the relatively high and localized con-
contration of impurities make the behavior of these
films very complex; therefore, the results of this
study are of practical interest. lIrradiation has
been found to cause additional crystalline imper-
fections, and the effects are reflected in changes
in the electrical and photoconducting properties
of this semiconductor. The results of the initial
experiments are described here.

The samples used were chemically deposited
PbS cells, supplied by the Eastman Kodak Com-
pany, and were of the type previously described
by Woods.” One set of contacts, which were pre-
pared by evaporation techniques and which were

 

7J. F. Woods, Phys. Rev. 106, 235 (1957).

210

used in making resistivity measurements, were
located at the center of the sensitive area. The
configuration is shown in Fig. 4,1.22. The thick-
nesses of samples 38 and 40 were 0.86 and 0.85 g,
respectively, and samples 5, 22, and 26 were
assumed to be 0.5 p thick.

UNCLASSIFIED
ORNL-LR-DWG 25452

A,B RESISTIVITY PROBES
C,D HALL PROBES

 

 

 

Fig. 4.1.22. Configuration of Lead Sulfide Sample
Cells.

Measurements were made at a temperature of
24.5 £ 0.5°C. All voltages were amplified by an
Applied Physics Corp. vibrating-reed electrometer,
but the actual readings were taken on a Rubicon
Co. potentiometer. Resistances were computed
from the relation between the direct current and the
voltage. All resistances measured were of the
order of 105 ohms. During all electrical measure-
ments the cell was sealed from light. Thus all
reported values represent dark-current measure-
ments.

The magnetic field for Hall-voltage measure-
ments was furnished by a 9000-gauss permanent
magnet; therefore it was not possible to reverse
the magnetic field. Instead, the sample was rotated
through 180 deg in the field, and readings were
taken with the current flowing in both the forward
and reverse directions during each orientation.
The Hall voltages were caiculated from the dif-
ference of the averages of a number of readings
made with the cell in each orientation. The Hall
measurements could not be made with a precision
of better than 15%. The maximum error in all
other electrical measurements was about 0.5%.
The other significant sources of error in these
experiments were the fluctuations in room temper-
ature (+0.5°C) and in humidity.

Changes in the photosensitivity of the cells were
followed by measuring the signal output when a

[ Y]

 

 
e

cell was subjected to radiation in the 5000- to
25,000-A region. A Cary model 14R spectropho-
tometer supplied the monochromator, light chopping,
and amplification systems. The cell was mounted
at the monochromator exit, and the signal from the
cell was fed directly into the preamplifier input.
This circuitry proved to be very satisfactory, since
the preamplifier system is designed for use with an
identical type of PbS detector, which actually
forms part of the instrument. Absolute wavelength-
response curves could not be obtained with such
an arrangement, since the slit-width drive could
not operate to compensate for the variable energy
output of the light source.

Neutron irradiations were carried out in the
ORNL Graphite Reactor. The samples were placed
in a boron carbide container so that slow-neutron
bombardment would be essentially eliminated. The
fast-neutron flux was taken to be 8 x 10'! neu-
trons/cm?.sec. Gamma irradiations were carried
out in the ORNL ‘‘gamma garden,’’ which utilizes
a cobalt source. The flux density was taken to be
2.65 x 10% r/hr, which is equivalent to about 1072
photons/em?2-sec.

A set of values observed for the four cells on
which electrical measurements were made is given
in Table 4.1.3. The values of the Hall coefficient,
Ry, are in good agreement with those reported by
Woods,” Lawrence,® and Lummis and Petritz.”
By assuming a film thickness of 0.5 p, values of
Ry, from +105 to +245 were calculated for the cells
studied by Lummis and Petritz on the basis of
their data., The resistivity values are, in general,

 

albid.. p 238. Woods' values of R, run from +121
to 63—133, and Lawrence reports values of + 130 and
+ 160.

9F. L. Lummis and R. L. Petritz, Phys. Rev. 105,
502 (1957).

 

PERIOD ENDING DECEMBER 31, 1957

lower than those reported by Woods, who gave
25,9 to 38.3 ohm-cm as the resistivity of his
sample. However, resistivity values calculated
from the data of Lummis and Petritz varied from
15 to 36 ohm-cm. The agreement of these values
within a rather wide range indicates the variance
which is to be expected for thin films of this type.

A plot of the resistivities of PbS cells 26 and
40 as functions of neutron irradiation is presented
in Fig. 4.1.23, A. An initial drop in resistivity
may be seen for both cells that is followed by a
gradual rise as the neutron dose is increased.
The change in resistivity of cell 40 with time,
following neutron irradiation is shown in Fig.
4,1.23, B. There was a rapid decrease in re-
sistivity during the first 24 hr, and it is to be
noted that the resistivity continued to fall with
time until, after the tenth day of annealing, it had
reached an apparent equilibrium value which was
below the resistivity of the cell prior to irradiation.
The five measurements made during the tenth to
twenty-first days gave an average value for the
resistivity of cell 40 of 13.7 £ 0.2 ohm-cm, which
is 17% below the initial resistivity of this film,
16.4 ohm-ecm. The data for cell 40 are plotted on a
larger scale in Fig. 4.1.24 for clarity.

The change in resistivity of the PbS cell after
neutron irradiation followed simple, first-order
kinetic relations, as may be seen in Fig. 4.1.25,
The half-life of the annealing process, determined
graphically, was 11 hr, which gave a value of
0.063 hr=! for the specific rate constant. Thus
the value of the resistivity following neutron
irradiation may be expressed as

M p = pg p—0.063 ¢ ,

where p is the value of the resistivity ¢ hr after
the value p, is measured.

Table 4.1.3. Properties of Chemicolly Deposited Lead Sulfide Cells at Room Temperature in the Dark

 

Hall Coefficient, Resistivity,

Carrier Concentration, Carrier Mobility,

 

Cell Ry (cm3/cou|omb) p (ohmecm) n (cm™3) Ry (cmz/v-sec)
x 106

22 +261 29.4 2.8 8.9

26 +132 12:6 56 10+5

38 +163 14.7 45 11.1

40 +190 1644 3.9 11.6

 

21N
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 25453R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

23 !
T T T T T T T T T ]
,p [~ IRRADIATION - ANNEALING ————————=m .
I ‘ 1 \ S
21— @ CELL 26 ,-/ \
A CELL 40 2 I\
20 /I‘ '\ ;
7 N *
19 ye 3
—_ A N
£ 18 z N,
T hY
£ /'7 ~
£47 A s\
= y ‘( k
l\ ; ~;
" 16
£ ¥ ' N
> n
£ 15 <
= »
L 14 S
o l ---...._____R_A__ A —_—
13
¢
lFLux =8x10'/em?. sec
12
" B
0 i |
I 5 40 15 20 2 4 6 B8 10 12 44 16 418 19 20 24 .
9 TIME (hr) TIME {doys)
0O 10 20 30 40 50 60 48 96 144 192 240 288 336 384 432 -
NEUTRON IRRADIATION TIME (hr)
TIME (min)

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

The percentage changes in the Hall coefficient,
carrier concentration, and carrier mobility of cell
40 are given in Table 4.1.4. Since the Hall
coefficient changed a maximum of less than 5%
during irradiation, which is within the limit of ex-
perimental error, it is concluded that this property,
as well as the carrier concentration, remained un-
changed by irradiation. The carrier mobility, which
is directly dependent on the Hall coefficient and
inversely proportional to the resistivity, showed
an increase as the resistivity decreased initially,
and then decreased when the resistivity increased.
The changes in these properties one to three
weeks after irradiation, as shown
Table 4.1.5, were significant, and all the values
showed definite trends toward limiting values.
The measured values for these properties 18 and

neutron in

212

Table 4.1.4. Changes in Electrical Properties of
Chemically Deposited PbS Cell 40 During

Neutron Irradiation

 

Total

Percentage Changes* in
Fast-Neutron

 

 

Dose Hall Carrier Carrier
(neutron s/cm2) Coefficient Concentration Mobility
2.4 x 104 +1 3 +5
4.8x10'4 +5 5 +4 :
7.2x 10" -1 0 ~4
1.4x10'° -1 0 -n ‘

 

*Based on volues measured prior to irradiation.
or

UNCLASSIFIED

ORNL-LR-DWG 25454R
100

50

20

ANNEALING TIME (hr)

0.5

0.2

24.5

 

0.
22 2 20 9 8 {7 46 15 14 13
RESISTIVITY, p (ohm-cm)

Fig. 4.1.24. Change in Resistivity of Chemically
Deposited PbS Cell 40 Following Neutron Irradiation,

21 days after irradiation were, for the Hall coef-
ficient, 181 and 195 cm3/coulomb, respectively;
for the carrier concentration, 4.1 and 3.7 cm™3;
and for the carrier mobility, 13.4 and 14.0
cm?/vesec. The Hall coefficient and the carrier
concentration reached about the same values after
21 days that they had before irradiation (see
Table 4.1.3), but the carrier mobility showed a
significant increase. Data for the period im-

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL—LR—DWG 25455R

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 i 1 T T
A. CELL 40 AFTER NEUTRON IRRADIATION
TO A TOTAL FLUX OF 3 x 10'° NEUTRONS /
4 em? - sec 1
B. CELL 38 AFTER GAMMA IRRADIATION TO
A TOTAL FLUX OF 3 x 0%/
5
E
T2
E N
o
e ®
Q
Q
v A
zZ
5 \ \
a N\ N\
@
" N N\
S \
A
<1
T 0.5
o 4\ A
N,
\ \
\
0.2 -
0.4
0 10 20 30 40 50

TIME, 7{hr)

Fig. 4.1.25. Change in Resistivity with Time of
Neutron-Irradiated Cell 40 and Gomma Irradiated Cell
38.

mediately following irradiation are not given be-
cause the resistivity was recorded automatically
during that period and the Hall voltage could not
be measured auvtomatically by the method being
used. It is apparent, however, that ot some time
during the first few days of the onnealing period
there was a decrease in the Hall voltage and a
concurrent increase in carrier mobility followed by
a gradual return of these properties to their initial
valves. The annealing mechanism led to a perma-
nent net decrease in resistivity, while only a

213
ANP PROJECT PROGRESS REPORT

Table 4.1.5. Changes in Electrical Properties of
Chemicolly Deposited PbS Cell 40 After Neutron
Irradiation to a Total Dose of
2.9 x 10'% Neutrons/cm?

 

Days After Percentage Changes in

 

Neutron Hall Carrier Carrier

Irradiation  Coefficient 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 -3 +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
curve a shows that fhg cell no longer responded to
radiation of 24,000 A after 30 min of irradiation,
but that it still responded to shorter wavelengths
(curves b, ¢, and d). This was primarily the
result of the light source emitting much more
intense radiation at the shorter wavelengths. An
examination of the data (Table 4.1.6) indicates,
however, that the neutron irradiation effects were
not uniform throughout the wavelength region ex-
amined and that they became more marked at longer
wavelengths. A more quantitative study of a sys-
tem which would provide uniform light intensity
at all wavelengths is necessary before the effect
at longer wavelengths can be considered to be
real. It is of interest to note that restoration of
the initial level of photosensitivity did not occur
even after three weeks of annealing. Indeed, the
signal output after this time was only about one-
half to one-third the original value, and it appeared
to be stabilized at the lower level.

214

UNCLASSIFIED
ORNL-LR-DWG 25456R

NEUTRON IRRADIATION TIME (min)

O 5 10 45 20 25 30 35 40 45 50 55 60
¢

WAVELENGTH OF EXCITING RADIATION
CURVE ¢ 24000 &
02 CURVE 21000 R
CURVE 14500 R
CURVE 10000 A
0.4 CURVE e 7000 A

O

0.3

0.5

0.6

0.7

0.8

09

LOG PHOTOSENSITIVITY
o

1.4 NO SIGNAL AFTE
thr OF IRRADIATION

IRRADIATION (UPPER SCALE)
1.7 —=== ANNEALING (LOWER SCALE)

 

0O 2 4 6 8 10 412 14 16 48 20 22 24
ANNEALING TIME (days)

Fig. 4.1.26. Effect of Neutron Irradiation on Photo-
conductivity of Chemically Deposited PbS Films (Cell
26).

The changes in resistivity, Hall coefficient,
carrier concentration, and carrier mobility produced
by gamma irradiation are plotted as the various
curves of Fig. 4.1.27. The effects during and im-
mediately following irradiation were fairly clean
cut, but it is apparent that the long-term annealing
effects following long periods of irradiation were
very complex.

An increase in resistivity during gamma irradia-
tion (plot A-l1) was clearly demonstrated. Cell 38
showed a threefold increase in resistivity after
receiving a dose of 5 x 10% r and a fourfold in-
crease after a dose of 10 x 10% r. After a dose of

e

 

 
ol

PERIOD ENDING DECEMBER 31, 1957

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

 

 

(Sy - S *
Exciting o
Radiation (A) Atn=2>5 At n =10 At n=135 At n =30 At n =60
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

 

*So 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-
tivity from 29.4 to 352 ohm-ecm. During the time
immediately following irradiation (plot A-li), the
resistivity showed a gradual decrease. The lower
curve, which demonstrates this change following a
2-hr irradiation, suggests that the resistivity
reached a limiting value after about 24 hr; the
actual wvalues of the last four measurements were
41.2, 40.4, 37.8, and 39.7 ohm-cm. The tendency
of the cell to reach an apparent equilibrium state if
the total irradiation did not exceed a certain maxi-
mum was also indicated by the changes in the
photoconducting properties, as will be discussed
later. Nineteen days after a dose of 10 x 10% r,
however, the resistivity was (plot A-llIl) still
undergoing large fluctuations.

If the change in resistivity of sample 28 follow-
ing a total gamma irradiation of 5 x 10° r is plotted
as a function of time as in Fig. 4.1.25, application
of the first-order kinetic law gives the following
relationship for the change in resistivity following
a total gamma irradiation which does not exceed
the ‘‘tolerance’’ for the cell:

(2) P = pg E,--0.041 t
The half-life of the annealing process was 17 hr.
Further examination of Fig. 4.1.27 shows that
there were over-all increases in the Hall coefficient
(plot B-l) and concurrent drops in carrier concentra-
tion and mobility (plot C-I) during gamma irradia-
tion of cells 22 and 38, but that the changes were
quite irregular. For instance, the data for cell 22
show a small decrease from 261 to 247 in R

during the initial 2 hr of irradiation, a threefold
increase during the second 2-hr period of irradia-
tion, and an eightfold increase when the accumu-
lated dose reached 21 x 10° r. Similarly, the data
for cell 38 show a fourfold increase during the
initial 2-hr period of irradiation, but a slight de-
crease during the second 2-hr period. These ir-
regular patterns are, of course, reflected in the
changes in the carrier concentration and mobility,
which depend upon Ry,

The long-range annealing effects (plots A-lll,
B-111, and C-Ill) were so irregular that they will be
mentioned only briefly. It is clear that no equilib-
rium condition had been reached 19 days after
irradiation. There was a trend toward lower values
for the resistivity and Hall coefficient, but it is
not possible to predict, on the basis of the data
obtained thus far, at what time an equilibrium state
would be reached and what these values would be
at equilibrium.  The carrier concentration and
mobility fluctuated so widely that no definite trend
is indicated by the data.

The previously mentioned ability of these cells
to recover from the effects produced by gamma
irradiation is demonstrated in Fig. 4.1.28. As
may be noted, the cells lost their sensitivity
rapidly as they accumulated progressively greater
doses of radiation. Examination of curves a, b,
and ¢ for cell 22 shows that this cell was still
sensitive after receiving anaccumulated irradiation
of 10.4 x 10° r, and after receiving 20.8 x 10% r
it still gave a reasonably good signal at 1.5 and
1.0 u, although no signal was obtained at 1.8 p.

215

 
ANP PROJECT PROGRESS REPORT

(x107'6)

CARRIER CONCENTRATION (cm™3)

HALL COEFFICIENT {cm® coutomb

RESISTIVITY, p (ohm-cm)

216

1650

700

600

500

400

300

200

{00

350

100

20

80

70

80

50

40

30

20

0

UNCLASSIFIED
ORNL-LR—DWG 25457R

DASHEDLINES = CARRIER MOBILITY
SOLID LINES = CARRIER CONCENTRATION

1

0

\,
”
7

MCBILITY (cmZ/ volt » sec)

N

 

IRRADIATION

ANNEALING

@ CELL 38 AFTER2-hr IRRADIATION
O CELL 38 AFTER4-hr IRRADIATION
ACELL 22

   

10 20 30 40 50 &0 8 10
TIME (hr) TIME (days)
2 4 & 8
¥ - IRRADIATION {hr]
FLUX = 26 X40° r/hr

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

it
GAMMA IRRADIATION TIME (hr)
0 2 4 6

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL —LR—DWG 25458R

8

 

<N ]

b guee WS

 

 

 

 

 

 

 

HAR NN

WAVELENGTH OF EXCITING RAD!ATION

CURVE o 18,000 A
CURVE b 15,000 A [ CELL 22
CURVE

N\
\ \EXCESS NOISE CURVE
CURVE e 46,000

CURVE
CURVE g SAME A

| l I I l l l

 

 

10,000 A

!

 

Q

24,000

I»o o

 

~

 

} CELL 3
d
€

SAME AS
S

 

 

w— |[RRADIATION (UPPER .

 

 

 

LOG PHOTOSENSITIVITY
- O
O @
W
/ A
g
/

//‘
S~—
\

 

 

 

 

 

 

 

£ \d NG S~ ———— SCALE)
1.4 L7 —]
! — —=—— ANNEALING AFTER 2-hr
~_ IRRADIATION (LOWER
1.6 EXCESS —+— NO SIGNAL — SCALE) —
NOISE | AFTER 8-hr
IRRADIATION
1.8
2.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 5 10 45 20 25 30 35

ANNEALING TIME (hr)

40

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

Further examination revealed that after irradiation
to an accumulated dose of 2.6 x 10% r, cell 5
(curves d and e) showed a considerable drop in
photoresponse, but, after annealing at room tem-
perature for 30 hr (curves [ and g), the sensitivity
returned almost to the initial value. After being
subjected to another gamma irradiation of equal
intensity, however, cell 5 had completely lost its
light sensitivity. The photosensitivity of cell 5

‘was measured daily for three weeks following

irradiation, and it had not recovered any of its
valuable photoconducting properties. The loss in
photosensitivity appears to have been the result
of a very large increase in the noise level of this
cell after irradiation. In order to check this, the
cell was scanned on the Cary spectrometer in the
usual manner, except that the light source was
turned off. Instead of the recorder showing no
response under these conditions, as is normally

the case prior to irradiation, a very erratic and
appreciable noise signal was recorded. This
observation is, at best, qualitative, but clearly
indicative.

A consideration of the effects of gamma irradia-
tion on these thin films of chemically deposited
PbS reveals a fundamental property. In brief, these
cells appear to have a fixed tolerance for gamma
irradiation. |f the total dose is maintained below
this tolerance, the physical changes effected in
the matrix can effectively be annealed out, and
the cell will revert to its original state. |f the
total irradiation exceeds this tolerance, either the
cell will be permanently damaged or the annealing
rate will be extremely slow. Which of these latter
possibilities is actually operative has not yet been
definitely established by these studies.

The possibility that fracture of the PbS$ film of
cell 5 may have occurred was considered. The

217
ANP PROJECT PROGRESS REPORT

film was examined under a magnification of 1000,
and no visible change in the smooth continuity of
the layer was found. Electron diffraction studies
should be made, however, before this possibility
is eliminated.

The problem of establishing a gamma tolerance
for these cells is complicated by the rather
individual nature of the cells. Even in this
preliminary investigation a difference of 400% was
observed in the sensitivity of two of these samples
prior to gamma irradiation.

The PbS cells used in this study were supplied
by the Eastman Kodak Company and the Bureau of
Ships, United States Navy.

IRRADIATION EFFECTS IN SEMICONDUCTOR
BARRIERS

Jl Co Pigg
C. C. Robinson 0. E. Schow

The program of study of irradiation effects in
semiconductor barriers is rather specialized and
was initiated at a time when a large, jumbled body
of information had already been accumulated.
Since reports of current progress are necessarily
isolated bits of information that appear out of
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
data. Such a survey introduction to each report
tends, however, to obscure the current achieve-
ments. It has been decided, therefore, to present
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
developments.,

There are many groups engaged in studying
radiation effects in semiconductor components,
and, since the motivations and objectives of the
investigations vary, the data accumulated and
the conclusions drawn seem to be contradictory,
Therefore statements describing motivation, ob-
jectives, and experimental details can frequently
remove contradictions in the data and in the
conclusions. The purpose of the ORNL program is
to contribute to an understanding of the effects of
irradiation on semiconductor barriers in terms of

 

lOW. Shockley, Electrons and Holes in Semiconductors,
with Afplications to Transistor Electronics, Van
Nostrand, New York, 1950.

218

fundamental processes. Such knowledge is appli-
able both to the development of semiconductor
devices with reduced sensitivity to radiation
fields and to a better understanding of the barrier
mechanism.

Semiconductor Materials

A semiconductor is a material whose electrical
conductivity has an energy of activation.'® When
the electrons are all in their lowest energy state,
the valence band is full and the conduction band
is empty (Fig. 4.1.29). In such a condition the
electrons cannot be accelerated by an electric
field, and hence no conduction can occur. If €
is so large that electrons cannot be excited across
the forbidden band at normal temperatures, the
material is an insulator. If € _is zero, the material
is a conductor; if € is such that electrons can be
thermally excited across the forbidden band, the
material is a semiconductor. Such conduction
is a property of the material itself and as such is
called intrinsic semiconduction.

UNCLASSIFIED
ORNL—LR—DWG 25484

b I =

W E ————— (g ———————- FORBIDDEN BAND

: !
///// VALENCE BAND

Fig. 4.1.29. Schematic Band Picture of a Semi-
conductor.

 

When impurities are present in the crystal, such
impurities may contribute energy levels which lie
in the forbidden band. If such impurity atoms have
more valence electrons than are necessary to
satisfy the bonding conditions of the crystal, the
electron may be excited from its ground state E
to the conduction band — such atoms are called
donors. Similarly an impurity atom having less
than the number of electrons required to satisfy
the bonding conditions may accept electrons from
the valence band and leave a hole — such impurity
atoms are called acceptors. For germanium and
silicon, impurities from atoms listed in column V
of the periodic table act as donors, while impurities

 

 
from atoms listed in column |ll act as acceptors.
Conduction by such carriers is called extrinsic
semiconduction. |t can be seen from Fig. 4.1.29
that the activation energy for extrinsic conduction
is smaller than that for intrinsic conduction.

If an excess of donor atoms is present, the
conduction will .occur by electrons (negative
carriers) in the conduction band — such a material
is said to be n-type. If an excess of acceptor
impurities is present, the conduction will occur by
holes (positive carriers) in the valence band — such
material is said to be p-type.

'Il'lae conductivity o of a semiconductor is given
by

(1) O = nep, + pep, ,

n = electron concentration,
= hole concentration,

e = electronic charge,

p, = electron mobility,

By = hole mobility.

If one type of carrier predominates, the expression
for intrinsic conductivity, Eq. 1, is modified by
dropping the term corresponding to the minority
carrier,

The concentration and sign of the majority carrier
can be obtained from Hall measurements. When a
current is established in a sample which has been
placed in a magnetic field, the motion of charged
carriers will be perturbed by the Lorentz force
acting on the moving charges. A voltage will be
established across the sample perpendicular to the
magnetic field and the current. The sign of this
voltage gives the sign of the majority carrier. The
magnitude of this voltage is inversely proportional
to the carrier concentration. The relation of the
quantities, together with appropriate constants
needed to convert the units, is

HI
(2) n = 7.38 x 1018 —— |
(Vv

H = magnetic field, in oersteds,

I = current, in amperes,

t = thickness of sample perpendicular to both
Hand I, in cm,

V = Hall voltage, positive for holes, negative
for electrons, in volts,

n = concentration of n- or p-type carriers.

PERIOD ENDING DECEMBER 31, 1957

If n and o are known, the mobility of an extrinsic
semiconductor may be determined from the extrinsic
expression for conductivity:

O'=?’1€I,L

Since electrons obey Fermi-Dirac statistics they
will be distributed throughout the available energy
states according to the Fermi function:

_ ]
 AE=Ep/T ,
where
T = absolute temperature,

k = the Boltzmann constant,
E = the energy under consideration,
Ef = the Fermi energy,

which can be shown to be the chemical potential
of the carriers. When E, lies at the middle of the
forbidden band, the electrical properties are those
which would be seen if E;, and E_ were absent.
Thus the material exhibits intrinsic behavior when
the Fermi level is at the middle of the forbidden
band. When the Fermi level lies above the middle,
the sample is n-type; when it lies below the middle
of the band, the sample is p-type. If E, lies close
(within o few times iT) to either the valence band
or the conduction band, the sample is said to be
degenerate. |f E, lies several times kT from these
bands, the sample is said to be nondegenerate.
For nondegenerate samples, Bose-Einstein statis-
tics can be used to calculate carrier concentrations.

It can readily be seen that both »- and p-type
conduction occurs simultaneously in intrinsic
material. In extrinsic material, conduction is
normally due to one type of carrier. However,
minority carriers can be produced in significant
quantities by photoelectric effects or by injection
at a barrier, Injection of minority carriers by
forward-biased barriers accounts for the high
forward currents in diodes and for emitter action in
a transistor.

In the case of the transistor, the injected
minority carriers drift across the base of the unit
to the collector. The current amplification, a, of
the transistor is determined by the ratio of the
collector current to the emitter current. The in-
jected carriers constitute a departure from equilib-
rium, and their concentration decays at a rate
1/7, where 7 denotes the lifetime of the minority
carrier. In order for a minority carrier to decay,

219
ANP PROJECT PROGRESS REPORT

there must be a transition between the valence
band and the conduction band.''" 12 These transi-
tions tend to occur via an intermediate step that
involves energy levels or trapping centers which
lie in the forbidden band: for example, E  and E,
as in Fig. 4.1.29. Consequently the lifetime de-
creases as the densities of these energy levels
increase,

When a semiconductor is placed in a radiation
field, there are three major ways in which the
radiation field can induce permanent changes in
the material. Capture of thermal neutrons produces
isotopes which decay to leave substitutional im-
purities in the crystal lattice. Elastic collisions
from energetic neutrons produce disorder in the
form of interstitials and vacancies, as well as
small regions of disorder. Gamma photons produce
Compton electrons, which in turn can produce
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
behavior such as that illustrated in Fig. 4.1.30(a)
will be observed.!'® The conductivity of n-type
material will decrease as the Fermi level falls
until the sample becomes p-type. After the sample
becomes p-type, the Fermi ievel will move toward
the equilibrium position and the conductivity will
saturate. The latter half of the curve also illus-
trates the behavior of high-density p-type material.
Low-density p-type material in which the Fermi
level is below the saturation position will exhibit
a drop in conductivity such as that illustrated in

Fig. 4.1.30(a).

The initial rate at which carriers are rémoved
from n-type germanium is 3.2 carriers per incident
fast neutron,'4 0.02 carrier per incident thermal

 

TR, N. Hall, Phys. Rev. 87, 387 (1952).

‘2w. Shockley and W. T. Reed, Jr,, Pbys. Rev. 87,
835 (1952).

13), C. Pigg and C. C. Robinson, Elec. Mfg. 59(4),
116 (1957).

4, W. Cleland et al.,

Phys. Rev. 83, 312-19
(1951).

220

UNCLASSIFIED
ORNL—LR—DWG 25485R

 

 

 

 

 

 

GERMANIUM SILICON
p (LOW p)
{a) nOR p (d)
. b
b o —p*— — == 2
n # (HIGH p} T T INTRINSIC
EXPOSURE (nvF) {nvt)
BULK MATERIAL
(&) = F (e)
=lF g
S R
z ~
N 3 \_
R T T TINTRINSICT

 

 

 

 

 

 

EXPOSURE (av?)
POINT CONTACT DIODES

{ nvt)

 

 

(¢)

INTRINSIC

I {1 volt)
"
log 7 (4 volt)
|2
“
2

 

 

 

 

 

 

EXPOSURE (avh) (nvt)
JUNCTION DIODES

Fig- 4.]0305

the Behavior of Bulk Germanium and Silicon with the

Comparisons of Effect of Radiation on

Effect on the Behavior of Germanium and Silicon
Diodes.
neutron, '3 0.0015 carrier per incident 1-Mev

photon.'® For p-type material, the rate of carrier
change by fast neutrons depends upon the initial
concentration, '7 but it is less than that for n-type
material.

Both n- and p-type silicon become intrinsic under
fast-neutron bombardment.'® The initial rate of
carrier removal is about 5 carriers per incident

neutron. The behavior of the conductivity of

 

15), w. Cleland, K. Lark-Horovitz, and J. C. Pigg,
Pbggs. Quar. Prog. Rep. March 15, 1950, ORNL-694,
p .

18, W. Cleland, J. H. Crawford, Jr., and D. K.
Holmes, Phys. Rev. 102, 722~24 (1956).

175, W. Cleland, J. H. Crawford, Jr., and J. C. Pigg,
Phys. Rev. 99, 1170=81 (1955).

18, w. Cleland et al., Phys. Solids Institute Quar.
Prog. Rep. April 30, 1951, ORNL-1095, p 35.
W

silicon under bombardment is illustrated by
Fig. 4.1.30(d). 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
thus those imposed by the radiation-induced
changes in the bulk of the semiconductor., The
interim behavior, prior to compiete 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
that results from the bulk resistance of the semi-
conductor, and a barrier which can be assumed to
obey the diode equation

[ = Iy(eaV/RT = 1),
where V is the applied voltage!? and g is the
electronic charge. Both R, and R, (Fig. 4.1.31)
may be functions of voltage or current, but, for
present purposes, they will be considered to be
ohmic.

With the use of the equivalent circuit of
Fig. 4.1.31 and the present knowledge of the
behavior of the electrical properties of semi-
conductors subjected to a radiation field, as
discussed above, the changes in the current-
voltage characteristic of a germanium p-n junction
that can be expected as a result of irradiation may
be described. In the forward direction, the leakage
resistance can be ignored for a practical diode.
The unit can be considered as a barrier in series
with the base resistance. Curve a of Fig. 4.1.32
illustrates the current-voltage relationship which
could be expected to exist for a pure barrier. In
a practical unit, however, the current is limited by
the base resistance, and a characteristic such as

 

Y4, c. Torrey and C. A. Whitmer, Crystal Recitifiers,
chap. 4, McGraw-Hill, New York, 1948,

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL—LR—DWG 25486

 

 

 

"-_QY—*N"_Q_Y——‘{

Fig. 4.1.31. Diode Equivalent Circuit.

curve b is obtained. Upon irradiation the semi-
conducting material is changed.20 Carriers are
removed from the n-type material, and the Fermi
level in that region is lowered. In the p-type
material, the Fermi level moves toward some
equilibrium position which depends upon the
temperature of the unit during exposure. In any
case, the Fermi level does not move so far in the
p-type material as in the n-type material. Hence

the barrier will be lower and there will be a
corresponding increase in I, The resistance of
the n-type base, however, will increase because

of the decrease in carrier concentration and the
decrease in the lifetime of injected minority

 

20 c, Pigg, Solid State Semiann. Prog. Rep. Aug. 31,
1953, ORNL-1606, p 79.

221
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR—DWG 23487

 

DiODE CURRENT

 

 

 

 

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,
a lower current can be expected, The new current-
voltage characteristic will have the appearance
of curve ¢ of Fig., 4.1.32. Under prolonged ex-
posure both sides of the barrier will tend toward
the same ultimate carrier concentration, and hence
the sample will tend toward ohmic behavior.

The base resistance can be neglected in the
consideration of the reverse characteristic curve.
The current through the unit is composed of two
parts, the diode current, [,, and the leakage
current, [, :

1 - e-QV/kt)

and

i vV

L = 5 °

RL

As discussed in the case of the forward charac-
teristic, irradiation will result in an increase in
an increase in I . The behavior of R} depends

222

upon the type of semiconductor, the initial treat-
ment of the surface, and the chemical environment
of the surface. The n-type germanium base units
irradiated to date have exhibited an increase in
R, 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
can be expected.

UNCLASSIFIED
ORNL—LR~DWG 254886R

 

 

2 I
5 T=1 "t
w v
o = 1‘ --__‘-'
g 5 ke
3 e
W ot e e e e i R
Q —_ r=1{1 — o=@
8 g AT
”~
7
/
/
/
/
/
/
/ a
L} L e

 

 

 

APPLIED VOLTAGE

Fig. 4.1.33. Behavior of the Reverse Current-Voltage

Characteristic of a Semiconductor Device.

The expression for I is given by Shockley 19 as:

 

ps ns
where
1/2
I P
ps qu ‘T_ 1
P
1/2
Dn
los = Aap | — :
n

A is the area of the junction, g is the electronic
charge, p and n are the densities of the holes and
electrons, respectively, D is the diffusion constant,
and Tis the lifetime of the indicated carrier.

i ®

 

 
The expression for I has been modified by Pell
to include an additional term to account for charge
generation in the barrier region.2! This added
term varies inversely with lifetime and barrier
height. It has been shown that the minority-carrier
lifetime is extremely sensitive to energetic irradi-
ation; 22 hence, the increase in 10 with irradiation
is expected to be larger than that anticipated on
the basis of the change in carrier concentration
alone. This large expected change is in agreement
with experimental observations.

The forward (F) and reverse (R) currents at lsv
bias, as a function of neutron exposure, were
illustrated in Fig. 4.1.30, parts (&), (¢), (e), and (/)
for germanium and silicon rectifiers, The ger-
manium point contact behaves as illustrated by
part (b).23 The current in the forward direction
drops to a minimum as the bulk conductivity
becomes intrinsic. As the bulk material becomes
p-type and the conductivity increases, the forward
current increases. The reverse current increases
monotonically as the barrier is destroyed, and the
sample becomes ohmic as the Fermi level comes
to the same position in the forbidden band through-
out the sample,

The silicon junction diode compares with bulk
silicon in a similar manner.?4 The forward current
drops as the bulk resistance increases. The
reverse current increases as the barrier is
destroyed. Eventually the sample becomes intrinsic
for both directions of bias.

The characteristic behavior of a silicon point
contact is somewhat different,?? as is shown in
Fig. 4.1.30(e). The forward current behaves in the
same manner as for the other diodes shown in
Fig. 4.1.30. The reverse characteristic, however,
shows no strong effect as a result of a barrier
being destroyed. Such behavior can be understood
if it is postulated that the region around the
contact behaves not as p-type but essentially as
intrinsic material; that is, the barrier would be

 

21E, M. Pell, J. Appl. Phy. 26, 658 (1955).

22J. C. Pigg et al., ANP Quar. Prog. Rep. March 31,
1957, ORNL-2274, p 265.

2:"‘J. H. Crawford, K. Lark-Horovitz, and J. C. Pigg,
Phys. Quar. Prog. Rep. March 15, 1950, ORNL-6%4,
p 64.

24, ¢, 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,
1953, ORNL-1506, p 47.

 

PERIOD ENDING DECEMBER 31, 1957

i-n instead of p-n. Such an i-n barrier would
rectify, but not so well as the p-n barrier, The
barrier would be destroyed as the bulk silicon
became intrinsic. During this time, the reverse
current would also decrease toward the intrinsic
level. Such a barrier would be unipolar, that is,
it would conduct by one type of carrier; hence it
would be more efficient at high frequencies, as
has been observed. The point-contact silicon
diode is the most satisfactory for high-frequency
applications: for example, 3-cm mixer applications.

Since transistor action involves the interaction
of two such barriers, with one barrier biased in the
forward direction and the other in the reverse
direction, all the properties involved in diode
changes apply simultaneously.  The collector
cutoff current will increase as the 1. of the
collector barrier increases. The slope of the
collector characteristic will change as the leakage
resistance across the surface of the base changes.
The emitter current-voltage characteristic will
behave in a manner similar to the behavior of the
forward diode characteristic.

Sah et al.,?% in considering both the diffusion
and charge-generation components of the emitter
current, express the current amplification, a, as

sech (Wb/Lb)

 

],g + ]d’

* J 4tanh (W, /L)

where L, is the minority-carrier diffusion length in
the base, W, is the base width, ] ; is the injected
current density flowing into the base, ] - is the
injected current density flowing into the emitter,
and J,¢ is the current component resulting from
recombination and generation in the emitter barrier.
The recombination-generation current density for
forward bias is shown to be

kT gnl
gF 2'7'0

J = cdV/2kT
rg

where E is the electric field at the junction, V is
the voltage across the emitter barrier, and T, is
the minority-carrier lifetime in the base. While the
diffusion component is inversely proportional to
the square root of the minority-carrier lifetime,

 

26C. T. Sah, R. N. Noyce, and W. Shockley, Proc.
Inst. Radio Engrs. 45, 1228 (1957).

223

 
ANP PROJECT PROGRESS REPORT

the charge-recombination-generation component is
inversely proportional to the first power. Hence
] rg will increase faster than | j as 7 decreases.

At low damage levels, where L, << W,, the
relative importance of the charge-recombination-
generation current with respect to the diffusion
current varies with the applied bias. When the
emitter bias is approximately equal to or greater
than the barrier height, the diffusion current
predominates. At bias voltages that are small
compared with the barrier height, the charge-
generation current predominates.

The diffusion length is connected with the
lifetime by the relation

L = (D'T)]/2

Where L, becomes comparable with W, the factor
sech (¥, /L,) takes control and decreases a
rapidly.

In the range where L, << W,, tanh (W,/L,) is a
rapidly varying function compared with sech
(W, /L,). The change in a with irradiation would

then be expected to be governed by the following
condition:
L, (initial) >> W,

and

]d>> ]rgl ]d' *

In the early stages of irradiation tanh (W, /L,)
changes rapidly and there is an initial increase in
a. However, as the irradiation progresses the
concentration of radiation-induced levels in-
creases, and ]rg increases faster than ] ,. When
L, ~ W, tanh (W,/L,) saturates, but in this range
sech (W,/L,) decreases rapidly. Hence changes
in both J,, and sech (Wb/Lb) tend to decrease the
value of a. Such behavior has been experimentally
observed?7 (Fig. 4.1.34). For units with W, and
L, of the same order of magnitude, such as low-
frequency power devices, this initial increase in
a may not be observed28 (Fig. 4.1.35).

 

27J. C. Pigg and J. W. Cleland, Solid State Quar.
Prog. Rep. Feb. 10, 1953, ORNL-1506, p 48, Fig. 28.

28J. C. Pigg and C. C. Robinson, Solid State Semiann.
Prog. Rep. Feb. 29, 1956, ORNL-2051, p 63, Fig. 54.

UNCLASSIFIED

 

45

  

SSD-B-741
DWG. 18922 AR
INTEGRATED FLUX (nvt)
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 (x10'0)
| | | I | | | | l
_|.

 

40

35

30

25

20

AMPLIFIGATION FACTOR

READ LOWER ABSCISSA l o

 

READ UPPER ARSCISSA

 
  

 

 

 

| | L | ! |
5 l | ’ l l I l l l [ [ i | [
50100 200 300 400 500 600 700 800 900 1000 1200 1400 1600 (x10'0)

INTEGRATED FLUX (nvt)

Fig. 4.1.34. Power Supply for Rectifier Experiment.

224
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
ORNL—LR—DWG 13759

 

 

 

AMPLIFICATION: A-C DROP ACROSS RESISTOR IN COLLECTCOR LEG

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

706
S
=
C
205 \
-
é ¢
5 54 I NO. 4; FLUX 1.86 x 10" oy
a e
\
0.3 v ™~ .
. "‘b \\Q\
9
..11..
0.2 e, o
Tt T ——
..-.' ". & -.--.—-OH-"'--_
0.1 ] S o]
NO.6; FLUX 1.86 x 10" ny "'"'""""T-o-—o-:'_'"?‘—'—-———- e
) L
0 1 2 3 4 5 6 7 8 9 0 1" 2 13 14 15 16 170"

INTEGRATED NEUTRON FLUX (avt,)

Fig. 4.1.35.

Integroted Fost-Neutron Exposure.

In high-frequency units having a very thin base
region, much greater stability should be expected.
When W, /L, <<1,

sech (W, /L,) % 1 and tanh (W, /L)~ — (Wb/Lb) .

In the denominator [ varies with 1/7 and ],
tanh (W, /L) varies as 1/7. The collector current,
as a function of irradiation for various emitter
currents of a surface-barrier transistor, is shown?2?
in Fig. 4.1.36. The unit is changed very little by ir-
radiation to an integrated flux of 1074 neutrons /cmZ2.

The value of I__, the collector cutoff current,
behaves like the value of the reverse characteristic.
The I increases rapidly because of the increasing
saturation diffusion current and the increasing
contribution of the charge-recombination-generation
current. Because of the J  component, Iy does
not saturate, The leakage component can become
the predominant term if surface changes occur.

 

2QJ. C. Pigg and C. C. Robinson, Solid State Semiann.
Prog. Rep. Aug. 30, 1956, ORNL.2188, p 19, Fig. 11.

Changes in Amplification of Two Minneapolis-Honeywell H-2 Power Transistors as a Function of

UNCLASSIFIED

-3
(x10™%) ORNL—-LR-DWG 15063

I = ma

6]

A

TRANSISTOR: PHILCO TYPE L-510C
NO. A-{944

(M

BIAS: 1 volt

ROOM TEMPERATURE

/., COLLECTOR CURRENT {ma)
ro

 

o 14
0 1 2 3 4 5 6 (X10)

nvt, INTEGRATED NEUTRON FLUX (neurrons/cm?-)

Fig. 4.1.36.
Philco L5106 Surface-Barrier Transistor os a Function

Changes in Collector Currents of a

of Irradiation.

225
This will be determined by the type of material
and surface treatment of the unit under con-
sideration.

" The transistor characteristics will change in
the manner illustrated39 in Fig. 4.1.37. Prior to
irradiation there is a family of curves, I_, with
a cutoft I__; subsequent to irradiation, the cutoff
current increases to I’ . 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
I, 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
I”. The curves will continve to collapse unti! 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
be possible to quantitatively describe the behavior,
but techniques for determining the pertinent
parameters for a completed unit have not yet been
developed.

 

30G. L. Keister and H. V. Stewart, Proc. Inst. Radio
Engrs. 45, 931 (1957).

226

f., COLLECTOR CURRENT

UNCL ASSIFIED
ORNL—LR-DWG 25489

 

 

 

 

 

 

 

 

 

¥, COLLECTOR VOLTAGE

Fig. 4.1.37. Collector Voltage vs Collector Current.

 
.t

Part 5
SHIELDING

E. P. Blizard
 

5.1. SHIELDING THEORY

EFFECT OF ANISOTROPIC SCATTERING ON
THE NEUTRON FLUX IN AIR

C. D. Zerby

For many calculations! of fluxes and dose rates
resulting from air-scattered neutrons, isotropic
scattering in the center-of-mass system is assumed.
The question naturally arises as to the validity of
this assumption. In order to obtain some insight
into the problem, the Monte Carlo code used
previously! for calculating the fluxes and dose
rates of multiply scattered neutrons was altered to
include an anisotropic scattering cross section
proportional to (1 + a cos 0) in the center-of-mass

system. Thus, the differential cross section is
given by

do
(1) E=K“ + acos ) ,

The value of the constant K was chosen to make
the integral of the differential cross section equai
to the total cross section.
arbitrary

The value of a is
and constant over the whole energy
range.

The cases investigated were for a point mono-
directional beam of neutrons in an infinite volume
of air. The beam angle with respect to the source
detector axis, 0, and the value of a were the only
parameters varied. The constants of the problem
were:

5 Mev

Source energy, EO

Separation distance, g 50 ft

Air density, p 0.00125 g/crn3

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 &, that is, ®(a), the ratio
of ®(a) to ©(0) was obtained. This ratio then
represents the factorial change which would be
expected in the results of a calculation in which
isotropic scattering was assumed if anisotropy
were to be introduced.

 

TEor example, see C. D. Zerby, A Monte Carlo
Calculation of Air-Scattered Neutrons, ORNL-2277
(April 23, 1957).

The resulting ratios of the total number flux,
energy flux, and dose rate are presented as a
function of beam angle of emission in Figs. 5.1.1,
5.1.2, and 5.1.3. [t may be observed that in each
case the ratio is less than 1 for beam angles
greater than 10 deg, which indicates that the
results obtained by using isotropic scattering in
the center-of-mass system are, in general, con-
servative.

It was found from the calculations in which
isotropic scattering in the center-of-mass system
was used that the equation which fits the neutron

flux data best is?

Ars

 

2 ¢ _
@) 4ng  sin 0

 

2J., Hilgeman, F. L. Keller, and C. D. Zerby, Monte
Carlo Calculations of Fluxes and Dose Rates Re-
sulting from Neutrons Multiply Scattered in Air, ORNL-
2375 (to be published).

UNCL ASSIFIED
2-D1-059-268

@, (a)
Pylo)

 

8, BEAM ANGLE

.
DETECTOR

o
SOURCE PCINT

 

0 30 60 90 {20 150 180
8, BEAM ANGLE (deg)

Fig« 5.1.1. Ratio of the Total Neutron Flux in Air with
Anisotropic Scattering to That Obtained with Isotropic
Scatterings Detector 50 ft from a monodirectional beam of
5-Mev neutrons. Air density = 0.00125 g/cm3.

229
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-059- 269

0.9

0.8

0.
Gelo)
Selo)

 

0.6

04

0.3 8, BEAM ANGLE

o .
0.2 SOURCE POINT DETECTOR

 

0 30 60 20 120 150 180
8, BEAM ANGLE (deg)

Fige 5.1.2. Ratio of the Total Neutron Energy Flux in
Air with Anisotropic Scattering to That Obtoined with
Isotropic Scottering. Detector 50 ft from a mono-
directional beam of 5-Mev neutrons. Air density =

0.00125 g/emS,

UNCLASSIFIED
2-01-059-270

Dpla)
Oyl0)

 

8, BEAM ANGLE

.
SQURCE POINT DETECTOR

 

o 30 60 920 120 50 180
8, BEAM ANGLE {deg}

Fig. 5.1.3. Ratio of the Total Neutron Dose Rate in
Air with Anisotropic Scattering to That Obtoined
with Isotropic Scattering. Detector 50 ft from a
monodirectional beam of 5-Mev neutrons. Air density =

0.00125 g/cm3.

230

 

where
@N(O) = neutron flux per source neutron,
2 (Ey) = macroscopic scattering cross section

at the source energy,
separation distance.
Equation 2 was also found to fit the neutron flux
calculated for singly scattered neutrons on the
assumptions of isotropic scattering in the labo-
ratory system and no attenuation on either leg.

A similar expression fits the calculated total
dose rate, @D(O), per source neutron.  This
equation is

C(Eo) ES(EO) 7 - 0
1.2 4ng  sin 0

 

€) 2,0 =

where C(E,) is the flux-to-dose-rate conversion
factor at the source energy.

Since such a simple approach was so effective
in providinga fit for the data for isotropic scattering
in the center-of-mass system, a similar set of
equations for the cases of anisofropic scattering
was investigated. The assumption that the dif-
ferential scattering cross section given in Eq. 1
was effective in the laboratory system and a
calculation of the singly scattered flux in which
attenuation on both legs was neglected yielded
the following equation:

 

(4) ¢M@=¢Mm6-asmi),

17 ==

where @, (0) is as given in Eq. 2. From Eq. 4 and
a similar equation for ®,(a), it is seen that

(DN(") q)D(a) sin @
(5) _ = = (1 -a > .
‘I’N(O) ®,(0) r—- 0

 

 

Values of these ratios calculated with the use
of Eq. 5 are compared with results obtained in
Monte Carlo calculations in Tables 5.1.1 and
5.1.2. As may be observed, the results of the
Monte Carlo calculation when anisotropic scat-
tering is considered are not fitted by the values
obtained with the use of Eq. 5 as closely as would
be desired. In fact, it is obvious from a qualitative
analysis that Eq. 4 should become less accurate
the more anisotropic the scattering. This is
borne out in the tables.

 
Table 5.1.1. Ratic of the Total Neutron Flux in Air
with Anisotropic Scattering to That Obtained

with |sotropic Scattering

 

 

Error
@ (deg) Formula Monte Carle of Formula
(%)
a=10.3
2 0.997 1.008 - 1.1
30 0.943 0.911 + 3.5
90 0.809 0.806 + 0.3
130 0.737 0.769 - 4,2
180 0.700 0.760 - 7.9
a = 007
2 0.992 1.020 - 2.7
30 0.866 0.796 + 8.8
90 0.554 0.537 + 3.2
130 0.386 0.453 -14.8
180 0.300 0.425 -29.4

PERIOD ENDING DECEMBER 31, 1957

BREMSSTRAHLUNG SPECTRA IN SODIUM
IODIDE AND IN AIRS

C. D. Zerby H. S. Moran

In order to provide some information on the
production of secondary electromagnetic radiation
in materials, a calculation was set up on the
Oracle to determine the spectra of bremsstrahlung
radiation resulting from the degradation of electrons.
The results of calculations for two materials,
sodium iodide and air, are shown in Figs. 5.1.4

and 5.1.5.

 

 

Table 5,1,2. Ratio of the Total Dose Rate in Air
with Anisotropic Scattering to That Obtained

with Isotropic Scattering

 

 

Error
0 (deg) Formula Monte Carlo of Formula
(%)
a=0,3
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
a= 0.7
2 0.992 1.026 - 3.3
30 0.866 0.815 + 6.1
90 0.554 0.545 + 1.7
130 0.386 0.449 -14.,0
180 0.300 0.430 -30.2

 

3Presented in more detail in a report by C. D. Zerby
and H. S. Moran, Bremsstrablung Spectra in Nal and
Air, ORNL-2425 (Feb. 25, 1958).

UNCLASSIFIED
2—01-059-275

1 £, =INITIAL ELECTRON ENERGY (myc?)
k=PHOTON ENERGY (mgc?) :

K
&yt

BREMSSTRAHLUNG PHOTONS PER UNIT(

 

Figs 5.1.4. Bremsstrahlung Spectrum in Sodium lodide.

231
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-041-059-276

£o=INITIAL ELECTRON ENERGY (mye?2)
k=PHOTON ENERGY (mgc?)

BREMSSTRAHLUNG PHOTONS PER UNIT

 

0 oA c.2 0.3 0.4 0.5 0.6 .7 0.8 c.9 1.0
_k
Eg—1

Fige 5.1.5. Bremsstrahlung Spectrum in Air.

In the calculation, the degradation of the elec-
frons was assumed to occur only by inelastic
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
by inelastic collisions to predominate. For the
presented in Figs. 5.1.4 and 5.1.5 the
last assumption is certainly true.?

cases

The energy loss per centimeter of path length
(stopping power) for an electron making inelastic
collisions is given by the equation®

 

4W. Heitler, The Quantum Theory of Radiation,
3d ed., p 252, Clarendon Press, Oxford, 1954.

5H. A. Bethe and J. Ashkin, Experimental Nuclear
Physics (ed. by E. Segre), p 254, Wiley, New York, 1953.

232

a - % =2m%NZ o BUE=D
x g2 212(1 - B2 .

- (2VT=F =14+p)n2+ !
+1 —,82+%(l - @),

 

where
E = total relativistic electron energy (mocz),
I = mean excitation potential (m c?),

B2 = (w/e)? = (E? = 1)/E2,

= classical electron radius (cm),

!

0 =
N = atomic density (cm=3),
Z = charge number.

The Born-approximation cross section for the
production of light (bremsstrahlung radiation) by
an electron in the field of a nucleus without
screening is given by®

 

0%t N | 4 pT + 0?
(2) DUEL) =—— C 5 - 2EE, +
137 & p p%pz .
eEl €]E 1 € EEI
+|— +— = — |+L|Z-— +
p> 3 pp, b
2
EE, + p
k 1
+ — (EzEf + pzp%) + — € —
3,3 2 3
p3p3 ke p
EE, .+ p} 2kEE, ]
- €, + ,
p3 pip?

 

Sw. Heitler, The Quantum Theory of Radiation, 3d ed,,
p 245, Clarendon Press, Oxford, 1954,
where the following relations are used:

.
EE, + ppy = 1
L=2Ih|——mmwuuuu-— ],
k
e=2In(E +p) ,

(3) { €, =2In(E; +p,) ,

)
m
~N

b3

 

\_

As may be seen from Eq. 3, Eq. 2 is simply a
function of E and k&, where E has the same def-
inition as in Eq. 1 and k is the photon energy
expressed in units of mocz. The expression
®%(E,k) dk is the electron cross section (cm=1!)
for producing a photon in the interval £ to & + dk.

When the electron screening of the nucleus is
taken into account and Eq. 2 is reduced to the
relativistic case, E >> mocz, the following
equation’ is obtained:

?'%ZZN 4
137

E - k\2
X 1 + —_— - —=lnZ| -
E 4

S

where y is given by &

4) ®Y(E,k) =

 

—~ X
k

 

 

 

 

5 =100 ——M8M .
) Y E(E - k) z1/3

The functions &,(y) and ¢,(y) are presented
graphically by Bethe and Ashkin;® however, they
are well represented with less than 0.3% error
by the empirical relations

 

"H. A. Bethe and J. Ashkin, Experimental Nuclear
Physics (ed. by E. Segrs), p 260, Wiley, New York, 1953.

81bid., p 259.
%1bid,, p 262.

PERIOD ENDING DECEMBER 31, 1957

5.089¢= 065877 1 15,661¢-0-04817

(6) ()
() $o() = ¢y(y) = 0.595¢=3-047

The quantity y measures the effect of screening.
For y >> 1 the screening can be neglected; how-
ever, for y = 0 the screening is complete. Al-
though Bethe and Ashkin givea series of equations
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
valid for all values of y.

With the use of Eqs. 1, 2, and 4, the total
bremsstrahlung spectra from an electron slowing
down in a compound is given by

Eq E (E,k) dE
j

—-d—dko ,
z<_£.

' dx /.
] j

(8) K(Egkg) dky =

where K(E k) dky is the number of photons pro-
duced by an electron slowing down from energy E,
in the interval k; to k4 + dky. The subscript j
refers to the jth element of the compound. The
quantity ®(E,k) is obtained from Eq. 2 or Eq. 4,
depending on which is appropriate in the course of
integration,

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
number of photons per unit (k/Eo) -~ 1 as a function
of (k/Ej) = 1. The integral under the curve
represents the total number of photons released.

A CALCULATION OF GAMMA-RAY DOSE RATES
FROM NEUTRON CAPTURE IN AIR

F.L. Keller C. D. Zerby
W. W. Dunn!?

Experiments at the TSF and elsewhere have
indicated that the gamma rays which are produced
by the capture of neutrons in air may become a
major contribution to the dose rate inside the crew
compartment of a heavily divided aircraft shield.
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
from these air-capture gamma rays.

 

IOOn assignment from USAF,

233
ANP PROJECT PROGRESS REPORT

The problem may be idealized to that of deter-
mining the air-capture gamma-ray dose rate and
angular distribution at a detector
arbitrary distance from a point source of mono-

located an

energetic neutrons where both the source and
detector are considered to be embedded in an
infinite volume of air. The calculation shouid in-
clude the contributions from captures which take
place during the slowing-down process as well as
contributions from captures which occur during
A portion of the
results of a parameter study of this problem by
the Monte Carlo technique for both point isotropic
and line-beam sources is presented here. The
complete results will be published in a separate
report.”

the diffusion at thermal energy.

Calculational Method

The Monte Carlo machine code which was used
for this problem was a modification of the Monte
Carlo code which was used to calculate neutron
12 This original code was com-
pletely described in the report cited; therefore, the
discussion here will be limited to a very brief
description of the general features of the code as
modified to perform the present problem.

The total cross sections for oxygen and nitrogen
were taken from BNL-325'3 and were considered
to be identical with the scattering cross sections,
except for the absorption cross section of nitrogen.
The absorption cross section of oxygen is ex-
tremely small and was neglected. The absorption
cross sectionofnitrogen is made up of the (#,p) and
(n,y) cross sections, which were taken to be
1.70 barns and 0.08 barn, respectively, at 0.025 ev
and were assumed to have a 1/v dependence, The
scattering probability was taken as isotropic in

scattering in air.

the center-of-mass system.

taken to be either a point
source of monoenergetic neutrons of
energy E_ or a point source which emitted a
monoenergetic line beam of neutrons of energy

The source was
isotropic

E_ at an angle 6, with respect to the source-

n
detector axis. The space about the detector was

 

”W. W. Dunn, F, L., Keller, and C. D. Zerby, A
Calculation of Gamma-Ray Dose Rates from Neutron
Capture in Air, ORNL-2462 (unpublished).

]2C. D. Zerby, A Monte Carlo Calculation of Air-
Scattered Neutrons, ORNL-2277 (April 23, 1957).

]3D. J. Hughes and J. A. Harvey, Neutron Cross
Sections, BNL-325 (July 1, 1955).

234

divided into a number of solid-angle intervals,
and estimates of the angular distribution of the

dose rate at the detector were made,

The neutron histories were generated in exactly
the same manner as described in the report on air
scattering.!? At each neutron collision, the
probability of capture was determined, and a
statistical estimate was
gamma-ray dose-rate confribution at the detector
from that collision. The neutron was then allowed
to scatter randomly and to continue with a reduced
weighting.

made of the capture

The neutrons were degraded in energy
at each scattering by the standard elastic-scattering
relations until the neutron energy reached 0.025 ev.
After the neutron energy was degraded to 0.025 ev,
the neutron was permitted to scatter and diffuse
without further energy loss. A neutron history
was terminated if the neutron weight after the /th
collision was less than about 1/600th of its
original weight or if the collision occurred at a
distance greater than 32 mean free paths from the
detector position. This enabled a 0.025-ev neu-
tron to make approximately 45 collisions before
its history was terminated. For higher source
energies the average number of collisions was, of
course, greater.

The capture gamma rays were assumed to be
emitted isotropically from the collision point, and
an exponential buildup factor was used in the
dose-rate calculation. The spectrum of capture
gamma rays which was used in the dose-rate de-
termination was from the recent work of Bartholomew
and Campion.'4 The doserate contributions were
stored in appropriate solid-angle intervals at the
detector under the assumption that all gamma rays
from a given event entered the detector in the
solid-angle interval which contained the radius
vector from the detector to the collision point.
All the results were normalized to one neutron per
second leaving the source, and all the results
reported here were from calculations which used
500 neutron case histories. It should be pointed
out that because.the method of statistical esti-
mation was employed, the results were relatively
insensitive
used.

to the number of neutron histories

 

]4G. A. Bartholomew and P. J}. Campion, Can. J.
Phys. 35, 1347 (1957).

 

 
The result of the Monte Carlo calculation was
checked against the result obtained from the
neutron flux given by the direct solution to the
transport equation for the case of a point isotropic
source of 0.025-ev neutrons located a distance of
15 m from the detector. The assumptions involved
in both calculations were the same except that the
Monte Carlo calculation took the neutron-scattering
probability to be isotropic in the center-of-mass
system, whereas the ftransport equation treated
the scattering as isotropic in the laboratory system,
This would be expected to lead to a slightly lower
value for the Monte Carlo result. The comparison
showed that the Monte Carlo result was about 10%
smaller than the result from the check calculation.
Hence, the results reported here are believed to
be good to about 10 to 20%. It is interesting to
note that the dose-rate contribution from the flux
given by the nonasymptotic part of the solution to
the transport equation was about 30% of the total
dose rate in the above caose.

Results for Point Isotropic Sources -

A series of calculations were performed for point
isotropic neutron sources with energies of 0.025,
0.1, 1.0, 100, and 10,000 ev and a source-detector
separation distance of about 50 ft. The air density
was taken tobe 0.001205 g/cm?, which corresponds

to approximately sea-level conditions. A plot of
the total gamma-ray dose rate vs the neutron
source energy is shown in Fig. 5.1.6. It may be

)

mr/hr
source neufron/sec

 

2

10—13
001 002

CAPTURE GAMMA -RAY DOSE RATE (

005 04 02 05 0 2 5

PERIOD ENDING DECEMBER 31, 1957

noted that the dose rate varies with the neutron
source energy approximately as E-1/3,

In these calculations, each gamma-ray dose-rate
contribution at the detector was also stored with
respect to the energy, at capture, of the neutron
which gave rise to the contribution. Figures 5.1.7,
5.1.8, and 5.1.9 show histograms of the dose rate
plotted as a function of neutron capture energy
for several of the cases run in this study. It may
be seen that, for relatively low neutron source
energies, a large part of the gamma-ray dose rate
at the detector comes from neutrons which are
captured at energies that are within a factor of a

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(x'!os"O) 2-01-059-266
’ i 1 T T [
w
B 14 —
a
:én w2 NOTE: 48% OF DOSE RATE FROM £, >a&,, _
8 —~ £ = SOURCE ENERGY
> i
a2 _|§wo —
R
AN
i
g Hyos —
2
&~
= 04 — —
Q
o
< 02 - -]
0
0025 00375 0050 00625 0075 00875 0400
£,, NEUTRON ENERGY AT CAPTURE (ev)
Fig. 5.1.7. Air Capture Gomma-Roy Dose Rate 50 ft

from a Point Isotropic Neutron Source as a Function of
EC for Eo = 0-] evs

UNCLASSIFIED
2-01-059-264

 

0 20 50 100 200 500 1000 2000 5000 10,000

£, ,NEUTRON ENERGY (ev)

Fig. 5.1.6. Air Capture Gamma-Ray Dose Rate ot a Detector 50 ft from a Point Isotropic Monoenergetic Neutron Source,

235
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-059-265

 

(x107%2)
5 T | | I

 

)
~
;
|

NOTE: 85% OF DOSE RATE FROM £,>0.;
18% FROM £,>a £y, £5= SOURCE ENERGY

mr-hr_-ev

(

 

N
I
L

 

 

AIR-CAPTURE GAMMA-RAY DOSE RATE
—
50Urce neutron- Sec
o
1
l

 

 

 

 

 

 

 

1 = [ ]
0 ot 02 03 04 05 06 07 08 09 10
£,, NEUTRON ENERGY AT CAPTURE (ev)

 

Fig. 5.1.8. Air-Capture Gamma-Ray Dose Rate 50 ft
from a Point Isotropic Neutron Source as a Function of
EC for EO = 1.0 ev.

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

(x10"' 2-01-059- 267
32
w
E 28 —
@ L -
&
@ 24 | —
O - —
Z |'v20 - NOTE: 40% OF DOSE RATE FROM £, 10 ev —
- @
1 I . —
<'zls
2 s —
2
6 T ; = ]
w o E S 2 —
2 ® — —
B o~—
% 08 |— -
o L _|
o
2 04 |~ —
0 I [ [ 1 ] I
© 0 20 30 40 50 60 70 8 90 100
£., NEUTRON ENERGY AT CAPTURE (ev)
Figs 5.1.9. AireCapture Gamma-Ray Dose Rate 50 ft

from a Point Isotropic Neutron Source as a Function of
E_ for EO = 100 ev,

of the source energy, where a =[{A = 1)/(A + 1)]?
and A is the average mass number for air. This
fact throws considerable doubt on calculations of
the airecapture gamma-ray dose rate which use
the results of the standard age-diffusion equation
for the neutron flux, since these results are known
to break down for neutron energies which are
within a factor of @ of the source energy.

Results for Line«<Beam Sources

Figure 5.1.10 shows the results for a series of
problems which was
sources.

run for line-beam neutron
The gamma-ray dose rate at the detector
is shown plotted as a function of the angle of
emission of the beam for various monoenergetic

236

The source-detector separation distance
was 65 ft, and the air density was again taken
as 0.001205 g/cm®. The curves are relatively
flat over the region from 30 to 180 deg, and they
become flatter as the source energy is increased,
as would be expected.

sources.

UNCLASSIFIED
2-01-059-263

10_‘0

 

2
—
2
w
.
.E§ 10-”
=3
E|E
3
3
-2 5
M
<
o
w
w0
o
e 2
>
«
©
i
3
= (o012
<
®
w
o
2 £,=1000 ev
a 5
O
2
10713

0 30 60 90 120 150 180
85, ANGLE OF EMISSION (deq)

Figo 50]0]00
Detector 65 ft from o Line-Beam Neutron Source Emitted

Air-Capture Gammo-Ray Dose Rate ot a

at Various Angles from the Source-Detector Axis.

Since most aircraft reactor shields are designed
so that most of the neutron leakage is from the rear
of the shield, the curves of Fig. 5.1.10 indicate
that a knowledge of the exact surface angular
distribution of these low-energy neutrons is rela-
tively unimportant. The important thing is to have
a knowledge of the spectrum and total number of
low-energy leakage neutrons. Because of the
e

extreme difficulty in making accurate experi-
mental determinations of low-energy neutron
spectra, it will probably be necessary to generate
this information by Monte Carlo techniques.
Additional results, including angular distributions
of the dose rate and results for other separation

PERIOD ENDING DECEMBER 31, 1957

distances, and a more detailed discussion of
these calculations will be found in the complete
report.”’

 

IsW. W. Dunn, F. L. Keller, and C. D. Zerby, A
Calculation of Gamma.Ray Dose Rates from Neutron

Capture in Air, ORNL+2462 (unpublished).

237

 
ANP PROJECT PROGRESS REPORT

5.2. LID TANK SHIELDING FACILITY
W. Zobel

PRATT & WHITNEY SERIES OF ADVYANCED
SHIELDING MATERIALS STUDIES

E. A. Warman'

A second group of tests of advanced shielding
materials? was performed in the Lid Tank Shielding
Facility (LTSF) in order to obtain information
required in the Pratt & Whitney shield design
effort (LTSF exp. 70). The purpose of these tests
was to determine the feasibility of using various
materials in the shield of a fast-neutron reactor
intended for possible aircraft-propulsion appli-
cation.

The materials tested were lithium hydride, de-
pleted uranium, and borated water, with particular
emphasis being placed on a study of the production
of secondary gamma rays in the uranium. [n most
of the tests, these materials were used in com-
bination with nickel and stainless steel. The
configurations tested are described in Table 3.2.1,
and the physical properties of the materials used
are presented in Table 5.2.2.

The components in configurations 36 through 42
and 90 and 91 were installed dry in the LTSF
configuration tank, which is a steel tank with
5/B-in.-thick walls. The tank has a %-in.-thick
aluminum window on the source side; an un-
avoidable 1.6-cm space between the aluminum
window and the source plate was filled with air
contained in a plastic bag. The borated water
used in configurations 36 through 42 and 90 and 91
was contained in an aluminum tank with I/a-in.-
thick walls. This tank was 31 in. long. Con-
figurations 43 through 49 and 79 through 86 were
submerged in borated water in the configuration
tank.

Gamma-ray dose-rate measurements were taken
beyond each configuration, and fast-neutron and
thermal-neutron measurements were made beyond
all configurations, except configurations 85, 86,
90, and 91. No neutron measurements were made
beyond configuration 86, and only thermal-neutron
measurements were made beyond configurations

 

IOn assignment from Pratt & Whitney Aircraft,

2The first group of tests was described by W. Zobel
et al., ANP Quar. Prog. Re@. June 30, 1957, ORNL-
2340, p 285, and by E. A, Warman, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 300.

238

85, 90, and 91. The fast-neutron and gamma-ray
data represent instrument response measurements
converted to ergs/g-hr-w of tissue dose as cali-
brated with known sources. The thermal-neutron
data are instrument response measurements nor-
malized to cadmium-difference gold-foil fluxes.
The fluxes were corrected for flux depression by
the 2-mil-thick foils and for self-absorption and
self-shielding. The instrumentation used in these
tests was the same as that described previously.?
The measurements made beyond all the con-
figurations are presented here, along with a brief
discussion; the data have not yet been analyzed.

Configurations OB, 43, 45, 48, 81, and 82. -
Gamma-ray dose-rate measurements taken beyond
configuration 81, as shown in Fig. 5.2.1, indicate
the effect of substituting 1 in. of stainless steel
for 1 in. of borated water in configuration OB,
which consisted entirely of borated water. The
dose rate was decreased by about 27%.

In configuration 43, 2 in. of nickel was inserted
in front of the stainless steel of configuration 81.
This resulted in a decrease in the gamma-ray dose
rate of about 55% from that of configuration 81 at
a distance of 40 cm from the source plate and a
decrease of about 42% ot a distance of 160 cm.
The change in slope of the dose-rate curve
possibly indicates a ‘‘harder’’ spectrum of gamma
rays.

In configuration 82, a 1.-in.-thick slab of
depleted uranium was placed behind the stainless
steel of configuration 43. The slope of the
gamma-ray curve was changed, and the curve
itself was lowered considerably. The gamma-ray
dose rate at 40 cm from the source plate was
lowered approximately 70% and at 160 em it was
lowered about 92%. The change in the slope of
the curve indicates a ‘‘softer’’ gamma-ray spectrum.
The measured reduction and the change in slope
also indicate a reduction in the contribution of
primary gamma rays from the source plate.

The effect of inserting 1 ft of borated water in
front of the uranium in configuration 82 is shown
by gamma-ray measurements beyond configuration
48. The dose rate was decreased by 80% at a

 

3D. R. Otis, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 331.

bt

av

 

 
 

PERIOD ENDING DECEMBER 31, 1957

 

Table 5.2.1. Descriptian of Configurations Tested in Pratt & Whitney Series

 

 

® Configuration o ) _ Location of Probe,? Unavoidableb Gap
Number Description of Configuration (cm) Within Configuration
s
* 0B Borated water (B-H20) in configuration tank 2.6
) 36 2in, of Ni + 1 in. of stainless steel + 2 ft of LiH 77.4 3.4 cm of air
+ B-H20 in an Al tank
37 2in. of Ni + 1 in. of stainless steel + 3 in. of 85.6 4.0 e¢m of air
depleted U + 2 ft of LiH + B-H20 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-H20 in an Al tank
39 2in. of Ni + 1 in. of stainless steel + 2 ft of LiH 81.4 3.6 cm of air
+ 11/2 in, of depleted U + B-H20 in an Al tonk
40 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 85.6 4,0 c¢m of air
+ 3 in, of depleted U + 1 ft of LiH + B-H20 in
an Al tank
41 2 in. of Ni + 1 in. of stainless steel + 1 ft+ of LiH 81.4 3.6 c¢m of air
+ 1)y in. of depleted U + 1 ft of LiH + B-H, O
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
- + 11/2 in, of depleted U + 1 4 of LiH+'|]/2 in, of
depleted U + B-H20 in an Al tank
o 43 2in. of Ni + 1 in, of stainless steel + B-HZO 12.8 2.6 cm of B-H20
44 2 in, of Ni + 1 in. of stainless steel + 3 in. of 21.0 3.2 em of B-H20
depleted U + B-H O
45 2in. of Ni + 1 in, of stainless steel + 2 ft of B-H O 75.8 0.8 cm of B-H20
+ 1/ in, of depleted U + B-H O
46 2 in. of Ni + 1 in. of stainless steel + 2 ft of B-H20 81.8 3.0 cm of B-HZO
+ 3 in. of depleted U + B-H20
47 2in. of Ni + 1 in, of stainless steel + 1 {t of B-H20 51.8 3.5 cm of B-H20
+ 3 in. of depleted U + B-H 0
48 2in. of Ni + 1 in. of stainless steel + 1 ft of B+H 0 46.0 1.5 cm of B-H20
+ 1/ in. of depleted U + B-H O
49 2in. of Ni + 1 in, of stainless steel + 1 #t of B-H O 81.0 2.2 cm of B-H2O
+ ]/ in, of depleted U + 1 ft of B-H O + 'I/ in.
. of depleted U+ B-H 0
79 2 in. of Ni + 1 in. of stainless steel + I/2 in. of 23.1 4.0 cm of B-H20
boral + 3 in. of depleted U + B-H O
s 80 2 in. of Ni + 1 in. of stainless steel + / in. of 25.0 4.7 cm of B-HZO
boral + 3 in. of depleted U + / in, of bora|
. + B-H,0
81 1 in. of stainless steel + B-H20 6.6 1.5 cm of B-H20

239
ANP PROJECT PROGRESS REPORT

Table 5.2.1 {continued)

 

 

Configuration L ) . L ocation af Probe,? Unavoidableb Gap »
Description of Configuration
Number z, (cm) Within Configuration
s
82 2 in. of Ni + 1 in, of stainless steel + ]]/2 in, of 16.8 2.8 cm of B-H20 .

depleted U + B-H20

83 2in. of Ni + 1 in, of stainless steel + 3 in. of 27.6 2.1 em of B-H20
B-H20 + 3 in. of depleted U + B-H20

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-H20

85 2 in. of Ni + 1 in. of stainless steel + 30 in. of 97.0 3.0 ecm of B-H20
B-HzO + 3 in, of depleted U + B-HzO

86 2 in. of Ni + 1 in. of stainless steel + 18 in. of 65.6 2.0 cm of B-H20
B-H2O + 3 in. of depleted U + B-H20

90 2 in. of Ni + 1 in. of stainless steel + 2 ft of L.iH 88.0 2.5 c¢cm of air
+ 4% in. of depleted U + B-H,0 in an Al tank

91 2 in. of Ni + 1 in. of stainless steel + 1 ft of LiH 88.2 2.7 em of air
+ 4, in, of depleted U + 1 ft of LiH + B-H,0
in an Al tank

 

2 Actual distance from source plate to back of solid configuration; that is, the borated water at the back of the
configuration is not included.

9

blncludes the l-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 L
Material Description
per Stab (cm)

 

Type 347 stainless 2,54 5ft 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/cm>
Nickel 2.54 Metallic slab, 64 x 66 X 1 in.
Uranium 3.81 Metallic slab, 52 x 55%6 X 1]/2 in., depleted to 0,24% U235
Lithium hydride 32.0 5x 5 x 1-ft slab encased in aluminum cans (]/4-in.-fhick walls); «

composition, 95 wt % LiH=5 wt % unaccounted for (probably
oxygen and hydrogen); density, 0.75 g/cm3

Boral 1.3 48 x 48 x ]/2-in. slab with 0.05-in.-thick Al cladding; composition, >
80 wt % Al-15.7 wt % B—~4.3 wt % C; density, 2.60 g/cm’

Borated water Composition, 1.34 wt % natural boron; density, 1.05 g/cm3 at 20°C

 

240
CONFIGURATION 82:
2in.OF Ni+1in.OF
STAINLESS STEEL
+1¥5 in.OF
U+ B-H,0

GAMMA - RAY TISSUE DOSE RATE (ergs/g-hr-watt)

CONFIGURATION 45:

NFIGURATION 43:
2in. OF Ni+1in. OF STAINLESS STEEL + B~ H,0
FIGURATION OB: |
B-H,0 IN CONFIGURATION TANK

PERIOD ENDING DECEMBER 31, 1957

2-01-057-70 -39

CONFIGURATION 81:
{in. OF STAINLESS STEEL + B-H,0

2in.OF Ni+1in. OF STAINLESS STEEL + 2 ft OF B-H 0

+1!% in. OF DEPLETED U+ B-H,0

CON ATION 48:

2in.OF Ni+1in. OF STAINLESS STEEL +1{ft OF B-H,0+1/2in.OF DEPL u+B

o 10 20 30 40 50 60 70

 

B8O 90 100 1o 120 130 140 150 160

Zy, DISTANCE FROM SOURCE {cm)

Fig. 5.2.1. Gamma-Ray Tissue Dose Rates Beyond Configurations 43, 45, 48, 81, ond 82.

distance of 60 cm from the source plate and by
50% at 160 ecm. The change in the slope of the
curve indicates a harder spectrum. The difference
between the curves for configurations 82 and 48
is believed to be almost entirely due to the
reduction in secondary gamma-ray production in
the uranium. The cause of this reduction is, of
course, the decrease in the thermal- and fast-
neutron fluxes incident on the uranium.

The borated water space of configuration 48 was
increased to 2 ft in configuration 45. The resulting
gamma-ray curve appears to be raised slightly.
This, however, may be attributable to experimental
error.

Corresponding thermal-neutron flux and fast-
neutron dose-rate measurements for the configu-
rations discussed above are shown in Figs. 5.2.2

and 5.2.3.

Configurations 82, 44, 79, and 80. — Configu-
ration 44 represents configuration 82 with a second
Il/z-in.-thick slab of uranium added to it. The
degradation of the higher energy gamma rays in
the added uranium resulted in a softened emergent
gamma-ray spectrum, as is evidenced by the

change in the slope of the curve of the data
plotted in Fig. 5.2.4.

In configuration 79, a !/2-in.-'rhick boral curtain
was placed on the source-plate side of the 3 in.
of uranium in configuration 44. The gamma-ray
dose-rate curve was lowered by approximately 20%.
This decrease is believed to be due almost
entirely to a reduction in the number of secondary
gamma rays resulting from (n,y) captures in the
uranium. The reduction, of course, results from
the reduction in the neutron flux incident on the

241
ANP PROJECT PROGRESS REPORT

THERMAL NEUTRON FLUX (neutrons /cm2-sec-wott)

242

 

2—-04-057-70-392

CONFIGURATION OB: B-H,0 IN CONFIGURATION TANK
GONFIGURATION 81: {in. OF STAINLESS + B-H,0

CONFIGURATION 43: 2in. OF Ni + 1in. OF STAINLESS STEEL + B-H,0

CONFIGURATION 82: 2 in. OF Ni +4in. OF STAINLESS STEEL +1'/5in.
DEPLETED U + B~H,0

CONFIGURATION 48: 2in. OF Ni + {in. OF STAINLESS STEEL + 1 ft OF
B-H,0+ 4¥%,in. OF DEPLETED U + B-H,0

CONFIGURATION 45: 2in. OF Ni + 1in, OF STAINLESS STEEL
+ 2t OF B-H,0 + ¢ Y5 in. OF DEPLETED U + B-H,0

10 20 30 40 50 60 70 80 90 100 140 120
25, DISTANCE FROM SCURGE {cm)

Fig. 5.2.2. Thermal-Neutron Fluxes Beyond Configurations 43, 45, 48, 81, and 82.

 

130

140

&,

 
¥

!

FAST-NEUTRON TISSUE DOSE RATE {ergs/g-hr- watt)

PERIOD ENDING DECEMBER 31, 1957

rugihs gt

L
3 2-04-057-70-393

CONFIGURATION OB: B~H,0 IN CONFIGURATION TANK

CONFIGURATION 43: 2in. OF Ni +4in. OF STAINLESS STEEL + B-H,0

ONFIGURATION 81: tin. OF STAINLESS STEEL +B-H,0

i ! | CONFIGURATION 48: 2 in. OF Ni +1in. OF STAINLESS STEEL +
1t OF B-H,0 +1Y%2 in. OF DEPLETED U+B-H,0

CONFIGURATION 82: 2in. OF Ni +1in. OF STAIN- ——
LESS STEEL +4'2in. OF DEPLETED U+B-H30

CONFIGURATION 45: 2in. OF Ni+4in. OF STAINLESS STEEL *
2 ft OF B-H,;0 + 1Y, in. OF DEPLETED U+B-H,0

 

1973

10 20 30 40 50 60 70 80 90 100 1o 120
Zq, DISTANCE FROM SQURCE {cm)

Fig. 5.2.3 Fast-Neutron Tissue Dose Rates Beyond Configurations 43, 45, 48, 81, and 82.

243

 

 
ANP PROJECT PROGRESS REPORT

OF DEPLETED U + B~H,0

GAMMA-RAY TISSUE DOSE RATE ({ergs /g hr-watt)

20 30 40 50 60 70 80

 

<A
2—-01—057-70-394

CONFIGURATION 82: 2 in. OF Ni+ 1in. OF STAINLESS STEEL + 1% in.

CONFIGURATION 44: 2 in. OF Ni+1in. OF STAINLESS STEEL + 3 in.
OF DEPLETED U +B-H,0

CONFIGURATION 79: 2 in. OF Ni+4in OF STAINLESS STEEL
+ Y, in. OF BORAL + 3 in. OF DEPLETED U+B-H,0

CONFIGURATION 80: 2 in. OF Ni +1in, OF STAINLESS
STEEL + ', in. OF BORAL + 3 in. OF DEPLETED U
+ %, in. OF BORAL + B-H,0

100 110 120 130 140 150 160 170

Zo, DISTANCE FROM SOURCE {cm)

Fig. 5,24, Gamma-Ray Tissue Dose Rotes Beyond Configurations 44, 79, 80, and 82.

uranium, that is, from the absorption of neutrons
in the boral.

In configurotion 80, a second Y -in.-thick boral
curtain was placed behind the 3 in. of uranium in
configuration 79. The gamma-ray dose rate was
almost unchanged at a distance of 34 cm from the
source, but it was lowered approximately 16%
at 160 cm. The change in the slope of the curve
is possibly due to the reduction of secondary
gamma-ray production in the borated water backing.

The neutron measurements beyond configurations
44, 79, 80, and 82 are shown in Figs. 5.2.5 and
5.2.6.

Configurations 43, 44, 46, and 49. — Gamma-ray
dose-rate measurements beyond configuration 44
illustrate the effect of adding 3 in. of uranium to
configuration 43. The gamma-ray dose rate beyond
configuration 44 was lowered about 80% from that
beyond configuration 43 in the forward region and

244

by about a factor of 30 in the region farther from
the source plate, as shown in Fig. 5.2.7. The
radical change in the slope of the curve points
out the much softer energy spectrum of gamma
rays being detected.

Configuration 46 was similar to configuration 44
except that 2 ft of borated water was inserted in
front of the 3 in. of uranium. The resulting
decrease in the gamma-ray dose rate is believed
to be due principally to the reduction in secondary
gamma-ray production in the uranium. If this is
assumed to be true, the difference between dose
rates beyond the two configurations can be taken
to be the approximate secondary gamma-ray con-
tribution to the dose rate beyond configuration 44.

In configuration 49, the borated water spacing
of configuration 46 was broken up into two equal
parts, the second section dividing the uranium in
half. This lamination proved to be a more

4

\‘
  
  

PERIOD ENDING DECEMBER 31, 1957

2—-0{-057-70- 395

   

CONFIGURATION 80: 2 in. OF Ni+ §in. OF STAINLESS STEEL + Y2 in. OF BORAL + 3 in.
OF DEPLETED U + Y% in. OF BORAL + 8-H,0

     
   

CONFIGURATION 79: 2 in. OF Ni+ 1in. OF STAINLESS STEEL + Y% in. OF BORAL + 3 in.
OF DEPLETED U + B-H,0

     
    
  

CONFIGURATION 44: 2 in. OF Ni+ 1in. OF STAINLESS STEEL + 3 in. OF DEPLETED
U+ B-HO
2

   

CONFIGURATION 82: 2 in. OF Ni+ 1in. OF STAINLESS STEEL + 1% in. OF DEPLETED
U+ B—H,0

    
  

102

THERMAL NEUTRON FLUX (neutrons/cm-sec-watt)

1o~

- *O—Z
10 20 30 40 50 60 70 80 90 100 110 120

Z5, DISTANCE FROM SOURCE (cm)

Fig. 5.2,5. Thermal-Neutron Fluxes Beyond Configurations 44, 79, 80, and 82.

245

 
ANP PROJECT PROGRESS REPORT

o oo

]
2-01-057-70-396

103

5
.
CONFIGURATION 44: 2in, OF Ni + {in. OF STAINLESS STEEL + 3in. OF DEPLETED U + B-H;0 .
2
CONFIGURATION 73: 2in. OF Ni +1in. OF STAINLESS STEEL + Y/ in. OF BORAL + 3in. OF .
DEPLETED U+ B-H,0
102
5 »
CONFIGURATION BO: 2 in. OF Ni + tin, OF STAINLESS STEEL + Y, in. OF BCRAL + 3in. OF
DEPLETED U + Y, in. OF BORAL
2
10
5
®
£ 5
o
~
wn
o
2
llf 2
2 CONFIGURATION 82: 2in. OF Ni +1in, OF STAINLESS STEEL + 1'% in.
" OF DEPLETED U + B-H,0
W 4 .
o
o
LIJ -
>
w
w5
|-
g .
x <
[
2
w2
o
w
& 10!

 

1073

20 30 40 50 60 70 80 S0 100 10
Zg, DISTANCE FROM SOURCE (cm)

Fig. 5.2 6. Fast-Neutron Tissue Dose Rates Beyond Configurations 44, 79, 80, and 82.

246
-y

PERIOD ENDING DECEMBER 31, 1957

<
2-01-057-70-397

ONFIGURATICN 43: 2in. OF Ni + {in. OF STAINLESS STEEL + B-H;0

CONFIGURATION 44: 2in. OF Ni +4in. OF STAINLESS STEEL +
3in.OF DEPLETED U +B-H30

GAMMA —-RAY TISSUE DOSE RATE (ergs/g-hr-watt)

CONFIGURATION 46: 2in. OF Ni+1{in. OF STAINLESS
STEEL + 2 ft OF B-H,O+ 3in. OF DEPLETED U +B~H30

CONFIGURATION 49: 2in. OF Ni +1in. OF STAINLESS STEEL +{ft OF B-Hp0 + ™
1% in. OF DEPLETED U + 1 ft OF B=H,0 + 1% in. OF DEPLETED U *B-H,0

30 40 50 60 70 80 90

 

100 110 120 130 140 150 160

25, DISTANCE FRCM SOURCE {cm)

Fig. 5.2.7. Gamma-Ray Tissue Dose Rates Beyond Configurations 43, 44, 46, and 49.

effective gomma-ray shield, the dose rate being
reduced about 14%.

The thermal-neutron and fast-neutron measure-
ments beyond configurations 43, 44, 46, and 49
are presented in Figs. 5.2.8 and 5.2.9.

Configurations 44, 46, 47, 83, 84, 85, and
86. — One series of configurations was tested in
order to investigate the production of secondary
gamma rays in 3 in. of depleted uranium as a
function of the amount of borated water preceding
it. The results are shown in Fig. 5.2.10.

In going from configuration 44, which contained
only the unavoidable amounts of borated water
between the solid sections of the configuration,
to configuration 83, which contained 3 in. of
borated water between the stainless steel and the
vranium, the gamma-ray dose rate was reduced by
about 55%. The slope of the curve was changed
only slightly. A combination of two phenomena
is believed to be the cause of this reduction.
The first, and probably the most important, is that
the production of (n,y) capture gamma rays in the

247
ANP PROJECT PROGRESS REPORT

L
<
2-01-057-70-398

IGURATION 43:
5 2in.OF Ni +{in.OF STAINLESS STEEL + B-H,0

FIGURATION 44:
2in.OF Ni+ 1in.OF STAINLESS STEEL + 3 in. OF DEPLETED

Ou N

ol\)l\) &)

&)

CONFIGURATION 46:
2in.OF Ni +1in. OF STAINLESS
STEEL + 2ft OF B-H,0 +3in.
OF DEPLETED U+ B-H,0

n

S

THERMAL-NEUTRON FLUX (neutrons/cmZ2. sec . watt)
o

CONFIGURATION 49:

2 in. OF Ni +1in. OF STAINLESS STEEL +1 ft OF B-H,0
-1 + 1%, in. OF DEPLETED U+ 1ft OF B-H,0 + 1%, in.OF
DEPLETED U+ B-H,0

 

-2
10
0 10 20 30 40 50 60 70 80 20 100 110 120 130
Z,, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.8. Thermal-Neutron Fluxes Beyond Configurations 43, 44, 46, and 49.

248

 
- watt)

FAST-NEUTRON TISSUE DOSE RATE (ergs /g- hr

PERIOD ENDING DECEMBER 31, 1957

A
2—-01-057-70~399

CONFIGURATION 44: 2in OF Ni+ 4in. OF STAINLESS STEEL
2 + 3in. OF DEPLETED U+ B-H,0

CONFIGURATION 43: 2in. OF Ni+ {in. OF STAINLESS STEEL + B-H,0

10

40_2 CONFIGURATION 46: 2 in. OF Ni + 1in. OF STAINLESS STEEL
+ 2 ft OF B~H,0 + 3in. OF DEPLETED U + B=H,0

CONFIGURATION 49: 2in. OF Ni +1in. OF STAINLESS STEEL
+1ft OF B-H,0 + 1%, in. OF DEPLETED U+ 4 ft OF B-H,0

2 + 1%, in. OF DEPLETED U + B-H,0

 

-3
10
20 30 40 50 60 70 80 30 100 10 120
Zzo, DISTANCE FROM SOURCE (cm)

Fig. 52.9. Fast-Neutron Tissue Dose Rates Beyond Configurations 43, 44, 46, and 49.

249
ANP PROJECT PROGRESS REPORT

103

<4an
2-01-057-70-400

CONFIGURATION 44: 2 in. OF Ni +{in. OF STAINLESS STEEL + 3in. OF DEPLETED U + B-H,0.

102

2 CONFIGURATION B3: 2 in. OF Ni +1in. OF STAINLESS STEEL + 3 in. OF
T B-H,0 + 3in. OF DEPLETED U + B-H,0
o
10
£
<
2 5 URATION 84: 2 in. OF Ni + 1in. OF STAINLESS STEEL + 6 in. OF
3 B-H,0 + 3in. OF DEPLETED U + B-
i
g 2
@
§ 1 CONFIGURATION 47: 2 in. OF Ni +1in. OF STAINLESS STEEL
a +1ft OF B-H,0 + 3 in. OF DEPLETED
" U+ B-H,0
a 5
0
-
=
x 2
&
Z 10!
<[
o

5 CONFIGURATION 86: 2 in. OF

Ni +1in. OF STAINLESS STEEL + 18 in.
, OF B-H,0 + 3 in. OF DEPLETED U + B-H,0
102
CONFIGURATION 85: 2 in. OF Ni + 1in. OF STAINLESS STEEL
5 + 30in. OF B-H,0 + 3 in. OF DEPLETED U + B-H,0

e CONFIGURATION 46: 2in. OF Ni + 1 in. OF STAINLESS STEEL + 2 ft OF B—H,0 + 3 in. OF DEPLETED U + B-H,0

10-3

30 40 50 60 70 80 20

 

100 110 120 130 140 150 160

Z.. DISTANCE FROM SOURCE {(cm)

Fig. 52.10. Gamma-Ray Tissue Dose Rates Beyond Configurations 44, 46, 47, 83, 84, 85, and 86.

uranium was greatly reduced by the reduction of
the incident thermal-neutron flux by the 3 in.
of borated water. The second effect is that
fissions in the uranium slab caused by inter-
mediate and fast neutrons were largely eliminated
by the additional neutron shielding.

In configuration 84, the borated water within the
configuration was increased to 6 in., and the
additional 3 in. lowered the gamma-ray dose rate
by approximately 57% more. This was caused by
a further reduction in the secondary contribution.
Configuration 47 had 1 ft of borated water in front

250

of the wuranium, and a further reduction in the
gamma-ray dose rate of about 55% was effected,
which again is attributable to the reduction of
the number of secondary gamma rays produced in
the uranium.

There was 18 in. of borated water in configu-
ration 86. The gamma-ray dose-rate curve was
lowered about 25% from that beyond configuration
47 and had a slightly different slope, possibly
indicating a harder spectrum. In configuration 46,
in which the amount of borated water was in-
creased to 24 in., the gamma-ray dose-rate curve

 

 
.‘

-

was agairi¥ered slightly in the forward region,
but, owing to the change in slope, it was raised
slightly in the regions farther from the source. It
might be concluded that the curve for configuration
46 represents only gamma rays from the source
plate; that is, secondary gamma rays produced in
the uranium this far out in the borated water are
negligible.

Configuration 85 contained 30 in. of borated
water, which resulted in a further slight reduction
in the gamma-ray dose rate, but this reduction is
sufficiently within experimental error as to be
inconclusive.

Thermal- and fast-neutron measurements beyond
configurations 44, 46, 47, 83, and 84 are shown
in Figs. 5.2.11 and 5.2.12, respectively. Thermal-
neutron measurements beyond configuration 85 are
shown in Fig. 5.2.11, but no fast-neutron measure-
ments were made for this configuration. Neither
thermal- nor fast-neutron measurements were made
for configuration 86.

Configurations 36, 39, 41, and 43. - A com-
parison of configurations 43 and 36 shows the
effect of adding 2 ft of lithium hydride to the
nickel —stainless steel region. There are, how-
ever, two other differences between these configu-
rations. The use of the aluminum-canned lithium
hydride introduced a Y in. thickness of aluminum
in configuration 36 on each side of the lithium
hydride. Also, the solid components in configu-
ration 36 were dry and were backed by borated
water in an aluminum tank, which introduced an
additional ‘/B-in. thickness of aluminum in configu-
ration 36 just behind the lithium hydride. Between
the solid components in configuration 36 there
were distributed air gaps totaling 3.4 cm, while
in configuration 43 there were distributed borated
water gaps totaling 2.6 cm. The use of lithium
hydride increased the gamma-ray dose rate, as
shown in Fig. 5.2.13. The fact that the gaps were
occupied by air rather than borated water probably
accounted for some of the increase because of
the slightly higher thermal-neutron flux incident
on the nickel, stainless steel, and lithium hydride.
The decreased density in the region of the shield
occupied by the lithium hydride (from 1.05 g/cm’
for borated water to 0.75 g/cm3 for lithium hydride)
no doubt accounted for a considerable portion of
the increase.

The thermal-neutron flux beyond configuration 36
was lowered from that beyond configuration 43 by

PERIOD ENDING DECEMBER 31, 1957

about 53% at 83 cm from the source plate and
by about 70% at 110 cm (see Fig. 5.2.14). This
reduction in flux and change in slope shows the
relative attenuation effectiveness of lithium hy-
dride and borated water. This rather large re-
duction in the thermal-neutron flux is accompanied
by a 70% reduction in the fast-neutron dose rate,
The obvious advantages of using lithium hydride
as a shielding material are illustrated by its good
neutron attenuation properties, with only a slight
increase in the gamma-ray dose rate, and relatively
light weight.

In configuration 39, a 11/2-in.-thick slab of
depleted uranium was placed behind the 2 ft of
lithium hydride in configuration 36. The addition
of the uranium resulted in a reduction of the
gamma-ray dose rate by about a factor of 20 (see
Fig. 5.2.13). This is about the attenuation to be
expected for gamma rays of 2 or of 7 Mev, and
is not far from that to be expected for intermediate
energies. Hence it is unlikely that neutron-
induced gamma rays in the uranium are a sig-
nificant contribution to the gamma-ray dose with
the uranium in place.

The slope of the thermal-neutron flux curve for
configuration 39 was drastically changed from that
for configuration 36 as shown in Fig. 5.2.14.
The flux was increased by about a factor of 4
at 88 cm from the source plate but only by about
20% at 100 cm. The curves would probably cross
at greater distances. This change might be
attributed to the fact that the uranium is a strong
inelastic scatterer, and thus it enhances the softer
flux near the slab at the expense of that which
would penetrate to greater distances. The fast-
neutron data for these configurations (Fig. 5.2.15)
indicate that the addition of the uranium behind
the lithium hydride caused an increase in the
fast-neutron dose rate next to the slab, but the
data cannot be accurately extrapolated to indicate
the situation farther back.

In configuration 41, the ll/z-in.-'rhick slab of
uranium was positioned between two 1-ft slabs
of lithium hydride. The gamma-ray dose rates
were approximately the same as those behind
configuration 39 (Fig. 5.2.13), but the thermal-
neutron fluxes were reduced considerably (Fig.
5.2.14).

Configurations 38, 39, 40, 41, 90, and 91. ~ In
configuration 38, a second 1%-in. slab of uranium
was added to configuration 39, with a resulting

251
ANP PROJECT PROGRESS REPORT

2-01-057-70-401

6
10
5 —
-
2 &
5 -
0
5
FIGURATION 44:
5 2in OF Ni +1in. STAINLESS STEEL + 3in.OF DEPLETED U+ B-H,0
CONFIGURATION 83:
> 2in.OF Ni+1in. OF
STAINLESS STEEL
a +3in.0FB-H,0+
10 3in, OF DEPLETED
U+ B-H,0
5
ONFIGURATICON 84:
_ 2in.OF Ni + 1 in. OF STAINLESS STEEL + 6in. OF B-H,0
x + 3in.OF DEPLETED U+ B-H,0
g 2
2
o 3
o 10
E
=
o5
e
5
@
£
Z 2
3 GURATION 47:
T 102 2 in.OF Ni +1in. OF STAINLESS STEEL
z +1ft OF B-H,0 + 3in.OF DEPLETED .
x U+B-H,0 .
S5 5
w
z
I
I
2
2 ~
W
I
-

o

CONFIGURATION 46:
2in.OF Ni +1in.OF
5 STAINLESS STEEL

+2 ft OF B-H,0 +
3in.OF DEPLETED
U+B-

 

2
1g"!
5
FIGURATION 85°
2in. OF Ni + {in. OF STAINLESS STEEL -
+30n. OF B~H,0 + 3in.OF DEPLETED U+ B-
2
1072 .

0 10 20 30 40 50 60 70 80 20 100 1O 120 -
Z,, DISTANCE FROM SOURCE PLATE {cm)

Fig. 52.11, Thermal-Neutron Fluxes Beyond Configurations 44, 46, 47, 83, 84, and 85.

252
e

h

FAST-NEUTRON TISSUE DOSE RATE (ergs /g-hr - watt)

PERIOD ENDING DECEMBER 31, 1957

gyt

S
2-01—057-70-402

CONFIGURATION 44: 2 in. OF Ni+ 1in, OF STAINLESS STEEL
2 + 3in. OF DEPLETED U +B-H,0

CONFIGURATION 83: 2 in. OF Ni +1in, OF STAINLESS STEEL
+ 3in. OF B-H,0 + 3in. OF DEPLETED U+ B-Hp0

10

CONFIGURATION 84: 2in. OF Ni+1in. OF STAINLESS STEEL
+ 6 in. OF B—H,0 + 3in. OF DEPLETED U + B-H,0

5 CONFIGURATION 47: 2in. OF Ni + 1in. OF STAINLESS
STEEL + 4 ft OF B-H,O0 +3in. OF D-"LETED U+B-H,0

> CONFIGURATION 46: 2in. CF Ni +1{in, OF STAINLESS
STEEL + 2 ft CF B-H,0 + 3in. OF DEPLETED U+B-H,0

 

1073
20 30 40 50 60 70 80 90 100 110 120

Z,,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
7o
5 FIGURATION 43: 2 in. OF Ni + fin. OF STAINLESS STEEL + B-H,0 .
-
2 .
\ __—CONFIGURATION 36: 2in. OF Ni +4in. OF STAINLESS
10 STEEL + 2 ft OF LiH + B-H,0 IN AN Al TANK —

—

GAMMA-RAY TISSUE DOSE RATE (ergs/g-hr- watt)

(10] 70 80 90

 

100

=~CONFIGURATION 39: 2in. OF Ni+1in. OF STAINLESS STEEL ——®
+2ft OF LiH + 12 in. OF DEPLETED U + B-H,OIN AN Al TANK

a

< / — .

CONFIGURATION 44: 2in. OF Ni +4in. OF STAINLESS STEEL +
1t OF LiH +1%2 in. OF DEPLETED U+1ft OF LiH + B-H,O IN AN Al TANK

1O 120 130 140 150

zy, DISTANCE FROM SOURCE {cm)

Fig. 52.13, Gamma-Ray Dose Rates Beyond Configurations 36, 39, 41, and 43.

dose rate by about a
This large reduction
secondary gamma-ray
The thermal-neutron
was raised over that

lowering of the gamma-ray
factor of 18 (Fig. 5.2.16).
again indicates negligible
production in the uranium.
curve for configuration 38
for configuration 39 in the forward region (Fig.
5.2.17) because of the replacement of borated
water with the uranium., The fast-neutron dose
rate (Fig. 5.2.18) was lowered about 15%.

The 3 in. of uranium in configuration 38 was
repositioned between the two lithium hydride slabs

254

to form configuration 40. Unlike configuration 41,
in which ]]/2 in. of uranium was positioned be-
tween the lithium hydride layers, the gamma-ray
dose rate in this case was increased by about
80% (Fig. 5.2.16). This increase might be prin-

cipally attributable to the production of secondary

gamma rays in the uranium, .
The depleted uranium gamma-ray shield was
increased to a thickness of 4]/2 in. in configu- .

rations 90 and 91. Configuration 90 represented
the addition of another Il/z-in. slab of uranium

 
THERMAL —-NEUTRON FLUX (neutrons/cm?-sec - watt)

 

80

PERIOD ENDING DECEMBER 31, 1957

Anap
2-01-057~70-404

CONFIGURATION 43: 2in. OF Ni +1in.
OF STAINLESS STEEL + B-H,0

CONFIGURATION 39: 2 in. OF Ni + 4 in.
OF STAINLESS STEEL + 2 ft
OF LiH+ 1'% in. OF DEPLETED
U4+ B—H,0 1t ‘N Al TANK

CONFIGURATION 36: 2 in. OF Ni+ 1in.
OF STAINLESS STEEL + 2 ft OF LiH
+ B—Hy0 IN AN Ai TANK

CONFIGURATION 44: 2 in. OF Ni+ 4in.

QF STAINLESS STEEL + 4 ft OF LiH

+ 1% in. OF DEPLETED U + 4 ft OF |
LiH + B—H,0 IN AN Al TANK

85 90 95 100 105 110 115
Z,, DISTANCE FROM SOURCE (cm)

Fig. 5.2.14. Thermal-Neutron Fluxes Beyond Configurations 36, 39, 41, and 43.

255
ANP PROJECT PROGRESS REPORT

256

e
2—04—057—70—405

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

107"
e iR

5 ‘,\

L CONFIGURATION 43: 2in.OF Ni +1in. OF —]
V STAINLESS STEEL + B—H,0
2 \
9

5
2 -2 \
; 10 \
< N\
o \ CONFIGURATION 36: 2in. OF Ni+ {4 in,
» OF STAINLESS STEEL + 2 ft OF LiH
g \ + B-H,0 IN AN Al TANK
w5 \* \ | CONFIGURATION 39: 2in. OF
< \ \ Ni+1in. OF STAINLESS STEEL
" + 2ft OF LiH + 4%, in. OF
o 0 DEPLETED U+B-H,0 IN AN
o A\ Al TANK
w a
) \
w
L2
|-
s CONFIGURATION 44: 2 in. OF \
o Ni+1in. OF STAINLESS STEEL o
2 +1ft OF LiH+1% in. OF
R DEPLETED U+ 4ft OF LiH
| 1
o
<
w

5

2

1074

60 70 80 90 100 o 120

25, DISTANCE FROM SOURCE {cm)

130

Fig. 5.2.15. Fast-Neutron Tissue Dose Rates Beyond Configurations 36, 39, 41, and 43.

 

 
¢

PERIOD ENDING DECEMBER 31, 1957

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2-01-0567-70-406
! T |
| I
[ ! =
-0 CONFIGURATION 39: ]
5 2 in.OF Ni + fin, OF STAINLESS STEEL 7
¥2F OF LiH + 1% in. OF DEPLETED — =2
U+B-H,0 IN AN Al TANK ~~
\(\ 9
—_ 2 I | /
© ® CONFIGURATION 4{: \ .
> 2in.OF Ni + {in.OF STAINLESS STEEL + { ft .
£ 10 OF LiH + 1Y%, in.OF DEPLETED U+ 1 ft OF /
o LiH + B-H,0 IN AN Al TANK _3 S
5 —— —7
g —CONFIGURATION 40: : | —Sog
o9 2in.OF Ni +1in. OF STAINLESS STEEL + 1 ft OF
w LiH + 3in.OF DEPLETED U +1 ft OF LiH +B-Hp0 —opm
g | INAN A|_Tm\u<__//|’/_71
T | L
& CONFIGURATION 38: |
S 2 2in.OF Ni+1in, OF STAINLESS STEEL + 2 ft OF —— 29
o Lil—||+3in.OF DEPLETED U+ B~Hp0 IN AN Al TANK A&
? ‘ | \
o)
-2
£ 10" |- CONFIGURATION 91: : r T ~ ~_
> 2in.OF Ni + {in. OF STAINLESS STEEL + { ft OF LiH+ 45 in. —L s | e
< OF DEPLETED U 41 ft OF LiH + B~H,0 IN AN AL TANK——""" e —~
Los e T
§ & \ \
"-
S \\A
2 |—CONFIGURATION 90: -
2in. OF Ni+1in. OF STAINLESS STEEL +2ft OF LiH +4Y in.0F |~ T—~2
DEPLETED U + B=H,0 IN AN Al TANK ——————| ~——_
107 A_
\
Ty
5x107%
40 50 60 70 80 90 100 o) 120 130 140 150 160

Zy, DISTANCE FROM SQURCE PLATE (cm)

Fig. 5.216. Gamma-Ray Tissue Dose Rates Beyond Configurations 38, 39, 40, 41, 90, and 91.

to configuration 38, and the gamma-ray dose rate
beyond configuration 90 was considerably lower
than that beyond configuration 38 (Fig. 5.2.16).
The slope of the curve was changed slightly, the
dose rate for configuration 90 being a factor of 7
lower in the forward region but only a factor of
5 lower in the region farther from the source plate.
The thermal-neutron flux curve (Fig. 5.2.17) was
shifted by the additional uranium.

In configuration 91, the 4]/2 in. of uranium was
positioned between the two I1-ft-thick slabs of
lithium hydride. As was the case when configu-
ration 40 was compared with configuration 38, the
gamma-ray dose rate beyond configuration 90 was
raised by roughly a factor of 4 over that beyond
configuration 91 (see Fig. 5.2.16). This increase
is also believed to be due to an increase in the
production of secondary gamma rays in the uranium
caused by the greater neutron flux incident on

the uranium. The accompanying decrease in the
thermal-neutron flux measured behind the configu-
ration (Fig. 5.2.17) tends to justify at least a part
of this assumption. No fast-neutron data were
obtained for configurations 90 or 91. [t should
be pointed out that the statistics on the data for
this group of configurations were much poorer than
those for other data included in ¢his report.
Configurations 36, 37, 38, 40, and 42. —~ In con-
figuration 37, a 3-in.-thick slab of depleted
uranium was placed between the stainless steel
and the 2-ft+-thick lithium hydride neutron shield.
The gamma-ray dose rate beyond the configuration
(Fig. 5.2.19) was about a factor of 9 below that
beyond configuration 36, in which there was no
uranium. This substantial decrease is believed
to be largely the result of the increased attenu-
ation of primary gamma rays. A comparison of
the data for configurations 37 and 38 indicates

257
ANP PROJECT PROGRESS REPORT

258

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o 4NN
PR, 2-01-057-70-407
10
CONFIGURATION 39: 2in. OF Ni + 4in. OF STAINLESS STEEL + 2 ft OF
. LiH + 1%, in. OF DEPLETED U + B-H,0 IN AN Al TANK
CONFIGURATION 38: 2 in. OF Ni + {in. OF STAINLESS STEEL + 2 ft _|
/ / OF LiH + 3in. OF DEPLETED U + B-H,0 IN AN Al TANK
) | |
| |
CONFIGURATION 41: 2 in. OF Ni+ 1in. OF STAINLESS STEEL
° c( +1ft OF LiH +1%,in. OF DEPLETED U +1 ft OF LiH
) \ P +B-H,0 IN AN Al TANK
N NT TN
NN\ X
TN
= 7\ o N\ |_— CONFIGURATION 90: 2in. OF Ni +4in. OF
g A NI\ N\ STAINLESS STEEL + 2 ft OF LiH + 4%, in, OF
- DEPLETED U + B-H,0 IN AN Al TANK
g N\~ NN \\ ?
e \ \ \ A\
LS
~
£ e 3 X
2
g \I
>
-
& 10~ \ \
X S\
2 ", S\
& \ AN
5 \\ N\,
ul
2 5 \‘E\ \
4 NN
< Ng \\
= \ )
w N
= CONFIGURATION 40: 2in. OF Ni +4in. OF f
2 I STAINLESSSTEEL +1ft OF LiH + 3in. N
OF DEPLETED U +1 ft OF LiH
+ B-H,0 IN AN Al TANK
9
1072 ~
4 u
/£
/
| CONFIGURATION 94: 2 in. OF Ni + {in. OF /.
5 STAINLESS STEEL + 4 ft OF LiH + 4% in. OF /
DEPLETED U +1ft OF LiH + B~H,0 IN AN Al TANK
2
10~3
85 90 95 100 105 110 15 120

Z4, DISTANCE FROM SOURCE PLATE (cm)

Fig. 52.17. Thermal-Neutron Fluxes Beyond Configurations 38, 39, 40, 41, 90, and 91.

 

 
 

PERIOD ENDING DECEMBER 31, 1957

 

 

 

 

 

 

 

 

 

 

 

 

 

Epn- 2-01- 05770~ 408
— 2
10
CONFIGURATION 39: 2 in. OF Ni + {in.
" OF STAINLESS STEEL + 2 ft OF LiH
+1Y%, in. OF DEPLETED U + B~ H,0
IN AN Al TANK
5
o

 

 

 

 

 

+ 1Y% in, OF DEPLETED U + 1 ft OF

\v N CONFIGURATION 44: 2 in. OF Ni + 1 in.
OF STAINLESS STEEL + 1 ft OF LiH
) o

 

 

 

LiH+ B—H,0 IN AN Al TANK

 

 

 

AN N

N\ \\ /—CONFIGURATION 38: 2in. OF Ni+ 1in. ™

OF STAINLESS STEEL +2 ft OF LiH

 

+ 3in. OF DEPLETED U+ B—-H,0 —
IN AN Al TANK

 

 

 

 

 

 

 

 

FAST-NEUTRON TISSUE DOSE RATE (ergs/g-hr - watt)

1
\CONFIGURATION 40: 2 in. OF Ni + 1 in.
OF STAINLESS STEEL + { ft OF LiH
+ 3in. OF DEPLETED U + 1 ft OF

 

LiH+B—-H,0 IN AN Al TANK

 

 

 

 

 

 

 

 

 

 

104

20 95 £00 105 110 115 120 125

ZO‘

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
rate between these configurations is essentially
the contribution of secondary gamma rays from
the uranium to the total measured dose beyond
configuration 37.

Configuration 42 represents a repositioning of
configuration 38 so that both the uranium and the
lithium hydride components are broken up into
two separate layers. The slope of the gamma-ray
dose-rate curve (Fig. 5.2.19) was changed by this

repositioning. The dose rates were identical for
both configurations at 95 cm from the source
plate, but the dose rate for configuration 42 was
lower by about 15% at 140 cm. The change in
the slope of the curve possibly indicates softening
of the spectrum as a result of the increased
degradation of primary gamma rays. A comparison
of configuration 42 with configuration 40, in which
only the lithium hydride is separated into two
layers, shows that the dose rate beyond configu~
ration 42 is about 45% lower. This decrease is

259
ANP PROJECT PROGRESS REPORT

GAMMA RAY DOSE RATE (ergs /g-hr-watt}

2—-01-057-70-409

CONFIGURATION 36: 2in. OF Ni + tin. OF STAINLESS STEEL + 2 ft OF
LiH + B-H,0 IN AN Al TANK

10

CONFIGURATION 37: 2 in. OF Ni + 1in. OF
STAINLESS STEEL + 3 in. OF DEPLETED U
5 + 2 ft OF LiH + B-H,0 IN AN Al TANK

CONFIGURATION 40: 2in. OF Ni+{in. OF
10~! STAINLESS STEEL +1ft OF LiH + 3in. OF
DEPLETED U + 1 ft OF LiH + B-H,0 IN
AN A}l TANK

CONFiGURATION 38; 2in. OF Ni +1in. OF
STAINLESS STEEL + 2 ft OF LiH + 3in,
OF DEPLETED U + B-H,0 IN AN A| TANK

1072
> CONFIGURATION 42: 2in. OF Ni +1{in. OF
STAINLESS STEEL + 1 ft OF LiH + 1% in. OF
DEPLETED U +1 ft OF LiH + 1% in. OF DEPLETED U
+ B-H,0 IN AN Al TANK
2
10~3

80 20 100 10 120 130 140 150
Z,, DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.19. Gamma<Roy Dose Rates Beyond Configurations 36, 37, 38, 40, and 42.

260

 

16C
believ%&be due principally to a reduction in
the production of secondary gamma rays in the
second 1),-in.-thick slab of uranium.
Thermal-neutron and fast-neutron measurements
beyond configurations 37 and 42 are compared with
corresponding measurements beyond configurations

36, 38, and 40 in Figs. 5.2.20 and 5.2.21, re-

spectively.

PERIOD ENDING DECEMBER 31, 1957

GENERAL ELECTRIC COMPANY SERIES
OF ADVANCED SHIELDING MATERIALS
STUDIES

D. W. Cady*? J. M. Miller

In addition to the series of tests described in
the preceding paper, the production of secondary

 

4On assignment from Wright Air Development Center,

2-01-057-70- 440

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10
5
CONFIGURATION 36: 2 in. OF Ni+ { in.
OF STAINLESS STEEL + 2 ft OF LiH
+ B—H,0 IN AN At TANK
2 I
g . | _— CONFIGURATION 38: 2 in. OF Ni+ 4in.
o o \ OF STAINLESS STEEL + 2 ft
S - (/ OF LiH + 3 in. OF DEPLETED
~ N[ N U+ B—H,0 IN AN Al TANK
E ‘-{\
<
@ \ B\
s . N A AT
5 \ I CONFIGURATION 42: 2 in. OF Ni +1in.
g < OF STAINLESS STEEL + 1 ft OF LiH _|
- Q. v, +1%,in. OF DEPLETED U + 1 ft OF
3 N\ LiH + % in. OF DEPLETED
3 Na ’ \ U+ B=H,0 IN AN Al TANK
o
=
2 N& \
o u
W \
|
0| .\\
3 - CONFIGURATION 37: 2 in. OF Ni + 1in. O
E:J — OF STAINLESS STEEL + 3 in. OF / N \ \ N
T —— DEPLETED U + 2 ft OF LiH/ C %\ <N
= —— + B—Hp0 IN AN Al TANK <o N
5 -
\ N
2NN\
CONFIGURATION 40: 2 in. OF Ni+ 1 in. .
OF STAINLESS STEEL + 4 ft OF LiH / 9 \)
+ 3 in. OF DEPLETED U + 4 ft OF Q
5 LiH + B—H20 IN AN Al TANK —] A
n
102 !
80 85 90 95 100 105 10 "5

Z,, DISTANCE FROM SOURCE (cm)

Fig. 5.2.20. Thermal-Neutron Fluxes Beyond Configurations 36, 37, 38, 40, and 42.

261

 
ANP PROJECT PROGRESS REPORT

2-01-057-70-411

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10-2
CONFIGURATION 36: 2in. OF Ni + {1in. OF STAINLESS STEEL + 2 ft OF
3{ LiH + B-H,0 IN AN Al TANK
5 \
N
- ) \ CONFIGURATION 38: 2 in. OF Ni +1in. OF STAINLESS STEEL
£ }-“'” + 2ft OF LiH + 3in. OF DEPLETED U + B~H,0 IN ———
z AN Al TANK
£
I 2
>
3 CONFIGURATION 37: 2 in OF Ni+{in.
- OF STAINLESS STEEL + 3 in. OF
'& DEPLETED U + 2 ft OF LiH + B-H,0
o IN AN Al TANK
=3
§ 0 TN
a NN
w
Z — CONFIGURATION 42: 2in. OF Ni+1in. OF \\A
S 5 STAINLESS STEEL + 1 ft OF LiH + 1'% in. OF N\
Q DEPLETED U +1 ft OF LiH +1Y%, in. OF A
- DEPLETED U + B-H,0 IN AN Al TANK
o
2
i
',_
o
&
CONFIGURATION 40: 2 in. OF Ni +1in. OF STAINLESS
2 STEEL +1ft OF LiH + 3 in. OF DEPLETEDU ——
+1ft OF LiH + B-H,0 IN AN Al TANK
10-4
80 85 90 95 100 105 110 15 120

2y, DISTANCE FROM SQURCE 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
depleted uranium is being studied in a series of
tests designed by the General Electric Company
to support their ANP effort. This experiment is
essentially a continuation of an earlier experiment,
No. 69, which was described previously.® In
general, the configurations used in this series
consist of an oil layer adjacent to the source
plate followed by a 4-in.-thick beryllium layer,
a gamma-ray shield of one of the three materials
being studied, and a 12-in.-thick lithium hydride

 

SR. W. Peelle et al., ANP Quar. Prog. Reps. June 10,
1956, ORNL-2106, p 269; Sep: 10, 1956, ORNL-2157,
p 278; Dec. 31, 1956, ORNL-2221, p 335.

262

neutron shield. All the configurations are sub-
merged in oil, and measurements of the gamma-ray
dose rate are made in the oil at various distances
beyond the lithium hydride. Thermal-neutron flux
measurements are made beyond a few configu-
rations.  The effect of inserting boral within
several of the configurations is also being studied.
The configurations tested thus far are described
in Table 5.2.3, and the physical properties of the
various materials used in the configurations are
presented in Table 5.2.4.

The first group of tests in this series was de-
signed for a determination of the effectiveness of
an oil layer preceding the configuration in reducing
the secondary gamma-ray production within the

 

 
PERIOD ENDING DECEMBER 31, 1957

Table 5.2.3. Description of Configurations Tested in GE Series

 

idable Qil G
Configuration Location of Probe, Unavoidable Oil Gap

 

Description of Configuration Within Configuration
Number Z, {(em)
{cm)

1-A 2.9 em of cil + 4 in. of Be + 2 in. of Hevimet + 12 in, 55.2 3.4
of LiH + oil

1-B 12.9 em of 0il + 4 in. of Be + 2 in. of Hevimet + 12 in, 66.0 3.2
of LiH + oil

1-C 22.9 em of oil + 4 in, of Be + 2 in, of Hevimet + 12 in. 74.6 1.8
of LiH + oil

1-D 32.9 cm of oil + 4 in, of Be + 2 in, of Hevimet + 12 in, 85.0 2.2
of LiH + oil

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 1 cm of oil + 4 in. of Be + Jy in. of boral + 2 in. of 56.2 3.1
Hevimet + 12 in. of LiH + oil

1-G 1 em of oil + 4 in. of Be + Jj in. of boral + 2 in, of 58.0 3.6
Hevimet + ]/2 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 + oil

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 + 4 in. of stainless steel 93.6 0.7
+ 12 in. of LiH + oil

2.F 1.0 cm of oil + 4 in. of Be + J; in. of boral + 4 in. of 59.8 2.9
stainless steel + 12 in. of LiH + oil

26 1.0 cmofoil +4in, of Be + /, in. of boral + 4 in, of 61.0 2.8
stainless steel + 1/2 in. of boral + 12 in, of LiH + oil

2-H 9.7 cm of oil + 4 in. of Be + )y 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
+ ]/2 in. of boral + 12 in. of LiH + oil

2-J 9.7 cm of oil + 4 in. of Be + Jj in. of boral + 4 in. of 69.4 2.5
stainless steel + ]/2 in. of boral + 12 in. of LiH + oil

2.K 27.8 cm of oil + 4 in. of Be + J, in. of boral + 4 in. of 84.6 0.9
stoinless steel + 12 in. of L.iH + oil

2-L 1.0 ¢m of 0il + 4 in, of Be + 8.7 cm of oil + ]/ in. of 68.4 2.8

2
boral + 4 in. of stainless steel + 12 in. of LiH + oil

263

 
ANP PROJECT PROGRESS REPORT

Table 5.2.3 {continued)

 

Configuration

Description of Configuration

Unavoidable Oil Gap

L ti f Prob
ocation of Probe, Within Configuration

 

Number g (cm) (cm)

2-M 1.0 e¢m of oil + 4 in. of Be + ]/2 in. of boral + 8.7 e¢m 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 + ]]/2 in. of depleted U 53.4 1.9
+ 12 in, of LiH + oil

3-B 13.8 cm of oil + 4 in. of Be + 'Il/2 in, of depleted U 62.8 1.7
+ 12 in, of LiH + oil

3-C 24,2 cm of oil + 4 in. of Be + 1), in. of depleted U 73.4 1.9
+ 12 in. of LiH + oil

3-D 33.2 cm of oil + 4 in. of Be + 11/2 in. of depleted U B2.2 1.7
+ 12 in, of LiH+ oil

3-E 40.8 cm of cil + 4 in, of Be + 11/2 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 + 17 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 + 11/2 in, 53.8 1.9
of depleted U + ]/2 in, of boral + 12 in. of LiH + oil

4-A 1.0 cm of oik+ 4 in. of stainless steel + oil 15.8 2.0

4-B 11.0 em of oil + 4 in. of stainless steel + oil 26.4 2.6

4-C 21.0 ¢m of oil + 4 in, of stainless steel + oil 35.2 1.4

4-D 31.0 em of oi!l + 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 e¢m of oil + 4 in, of staintess steel + oil 65.2 1.4

4-G 61.0 cm of il + 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 e¢m 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-
creasing the thickness of the oil layer preceding a
beryllium-Hevimet-lithium hydride shield (con-
figurations 1-A through 1-E) from 2.9 to 12.9 cm
reduced the gamma-ray dose rate beyond the shield
approximately 35%. Increasing the thickness of
the oil layer another 10 ¢cm decreased the gamma-
ray dose rate an additional 16%. Oil layers
greater than 22.9 cm appeared to have no further
effect. This information can be used to determine
the optimum thickness of the oil layer for the

contemplated reactor-shield configuration. The oil

264

layers had approximately the same effect in tests
of similar configurations with stainless steel
(configurations 2-A through 2-E) and with depleted
vranium (configurations 3-A through 3-E), as
shown in Figs. 5.2.23 and 5.2.24.

In all the tests described above an attempt was
made to increase the thickness of the oil layer
in 10-em increments. The fact that the configu-
ration tank used to hold the shield components
had a l-em recess in the tank wall adjacent to
the source plate (see preceding section) made it
impossible to obtain an oil layer smaller than
PERIOD ENDING DECEMBER 31, 1957

Table 5.2.4, Physical Properties of Materials Used for LTSF Mockup Tests
of Advanced Shielding Materials (G-E Series)

Sl

 

. Material

Actual Thickness

per Slab (em)

Description

 

Tronsformer oil

Specific gravity, 0.88 ot 15.6°C; nominal composition,
86.5 wt 2 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 % W6 wt % Ni=4 wt % Cu; density of Hevimet,
16.9 10 17.2 g/cm>
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
. l em, as is evident in configuration 2-A. More-
. o | [ _ _____ zorostrem over, the design specifications set by GE for the
% 5 1 CONFIGURATION -A, 2.9—cm OIL GAP experiment required that the centers of the gamma-
£ 5 [ o CONFIGURATION 1-B, {2.9—cm OIL GAP ray shields in configurations 1-A, 2-A, and 3-A
- = [T 4 CONFIGURATION 1-C, 22.9~cm OIL GAP A )
3, | ° CONFIGURATION 4-0, 329-cmoi GaP | | | be equidistant from the source plate. Since the
‘ = * CONFIGURATION 1-E, 40.6-cm OIL GAP gamma-ray shields used in configurations 1-A and
g 10 ~ CONFIGURATION: 4 in. OF Be + 2 in. OF 3-A were thinner than the gamma-ray shield used
8 s RS HEVIMET »12in. OF LiH » OIL i in configuration 2-A, the oil layers preceding
‘ s L S these configurations were necessarily thicker.
22 i —een A cross plot of the data presented in Figs.
‘ ; , \%b S _ 5.2.22 through 5.2.24 is shown in Fig. 5.2.25,
% ] SIS~ which presents the gamma-ray dose rate at 100 ¢m
z . . from the source plate as a function of the distance
3 ) j o e between the source plate and the center of the
, *-2\";{ gamma-ray shield for each of the three tests.
0 0 80 80 100 o 120 130 140 150 160 This cross plot indicates that the minimum thick-
2, DISTANCE FROM SOURCE PLATE (¢m) ness of the oil layer preceding the beryllium for
the lowest gamma-ray dose is approximately
Fige 52,22, Gamma-Ray Tissue Dose Rates Beyond 20 em.
- Configurations 1-A Through 1-E: Effect of Increasing In several configurations ]/z-in.-fhick sheets of

Thickness of the

Containing Hevimet.

the

Oil

 

Gap Preceding Shields

boral were placed adjacent to or near the gamma-
ray shield. The purpose of these tests was to

265
ANP PROJECT PROGRESS REPORT

GAMMA-RAY TISSUE DOSE RATE (ergs/q-hr - watt)

Figo 5.2.23-

CONFIGURATION
CONFIGURATION
CONFIGURATION
CONFIGURATION

CONFIGURATION

<
2—041—057—71—380

1.0-cm CIL GAP
14.0-cm OIL GAP
24.0-cm OIL GAP
34.0-cm OIL GAP
37.2-cm OIL GAP

/OIL ONLY IN CONFIGURATION TANK

™~

A
\

CONFIGURATION: 4in, OF Be + 4in. OF STAINLESS STEEL
+ 12in. OF LiH + OIL

 

8O S0 100 110 120 130
Z,,DISTANCE FROM SCURCE PLATE {cm)

T~

140 150 160

Gamma-Roy Tissue Dose Roates Beyond Configurations 2-A Through 2-E: Effect of Increasing the

Thickness of the Oil Gop Preceding Shields Contoining Stainless Steel.

266
[y LT S

4x10

PERIOD ENDING DECEMBER 31, 1957

2-01-057 - T1-382

CONFIGURATION 3-A, 4.2-cm OIL GAP

CONFIGURATION 3-8, 13.8-cm OIL GAP

 

CONFIGURATION 3-D, 33.2-cm OIL GAP

GONFIGURATION 3~E, 40.8-cm OIL GAP

10 CONFIGURATION 3-C, 24.2-cm OIL GAP

5
x
=

- 5
S~
o
o
e
o
<L

x 2
w
D
o
O
w
o

a 1
=
S—
<1
a
I

3 5
=
3

CONFIGURATION: 4in. OF Be +1% in. OF DEPLETED
URANIUM + 12 in. OF LiH + OIL
2
1o™!
70 80 90 100 10

120 130 140 150 160

Z,, DISTANCE FROM SOURCE PLATE (cm)

Fige 5.2.24. Gamma-Ray Tissue Dose Rates Beyond Configurations 3-A Through 3-E: Effect of Increasing the
Thickness of the Ojl Gap Preceding Shields Containing Depleted Uranium.

determine the effectiveness of boral in reducing
the secondary gamma-ray production in the gamma-
ray shield. In configurations 1-F, 2-F, and 3-F,
a single sheet of boral was placed immediately
in front of the gamma-ray shield, while in configu-
rations 1-G, 2-G, and 3-G the boral was positioned
both before and after the shield. The measure-
ments (Fig. 5.2.26) showed that placing the boral
in front of the Hevimet reduced the dose rate
beyond the shield approximately 31%. Positioning
a second boral sheet behind the Hevimet produced
a further slight reduction, which may be within
experimental error. Placing the boral in front of
the stainless steel reduced the dose rate beyond
the shield approximately 47% (Fig. 5.2.27). The
second boral sheet behind the stainless steel
again produced a further slight reduction. In

 

configurations with depleted uranium, the addition
of the boral sheet in front of the uranium reduced
the gamma-ray dose rate 51% at a distance 70 cm
from the source but only 39% at a distance of
160 em (Fig. 5.2.28). A second boral sheet placed
behind the gammo-ray shield again reduced the
dose rate only slightly.

Several other tests were performed with vari-
ations of the configurations containing stainless
steel, the results of which are plotted in Fig.
5.2.29. In configurations preceded by an approxi-
mately 9-cm-thick oil layer {configurations 2-H and
2-J), the addition of a second boral sheet behind
the stainless steel reduced the gamma-ray dose
rate about 10%. It is felt that no particular
significance can be attached to the 10% reduction

267
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A am——
2—04—057—T71—383
10
A
° \
] \
;
\CONFIGURATION 2A-E: OIL + 4in. OF Be + 4in. OF
STAINLESS STEEL + 42in. OF LiH +0IL
6 \
A
\
5
5
2 g
E " \
< \ A
o
i \A. A
g
3
W
3
- ‘ \ " CONFIGURATION 3A-E: OIL+4in. OF Be +1Y,in OF
g \ DEPLETED U+12in. OF LiH + 0OIL
«
= "
z \
o \
2
N N\
0\ \ x
' x
CONFIGURATION 4A-E: OIL+ 4in. OF Be
+2in. OF HEVIMET + 42in, OF LiH +OIL \._ o
..
1 |
0 10 20 30 40 50 60 70

Fig. 5.2.25.

DISTANCE FROM SOURCE PLATE TO CENTER OF GAMMA-RAY SHIELD (cm}

Gamma-Ray Tissuve 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
2-01-05;—!*—413

6x10

CONF OiL Be BORAL HEVIMET BORAL LiH
SYMBOL NO. {(cm) {in) (in) {in.) tin)  (in)

. 1-a 29 4 ¢ 2 0 12 +0lL
A t~F 10 4 2 0 12 +0I
¢ -6 10 4 ‘5 2 Yo 12 40l

GAMMA~-RAY TISSUE DOSE RATE {ergs/g-hr-watt)

 

60 80 100 120 140 160 180

Io,DiSTANCE FROM SOURCE PLATE (¢m)

Fige 5.2.26. Gomma-Roy Tissue Dose Rates Beyond
Configurations 1-A, 1-F, ond 1-G:

Boral Within Configurations Containing Hevimet.

Effect of Inserting

in dose rate because the magnitude of experimental
errors may account for this variation.

A comparison of measurements beyond configu-
rations 2-F (from Fig. 5.2.27) oand 2-K and 2-H
(from Fig. 5.2.29), which contain a single sheet
of boral immediately in front of the stainless
steel, shows that increasing the thickness of the
oil layer adjacent to the source plate from 1.0 to
9.7 cm reduced the gamma-ray dose rate approxi-
mately 29%. Increasing the oil layer from 9.7
to 27.8 cm reduced the gamma-ray dose rate only
10%, which, again, may be within experimental
error. This appears to indicate that, for configu-
rations with boral in front of stainless steel,
increasing the thickness of the oil layer preceding
the configuration to more than 10 cm does not
effectively reduce the dose rate beyond the shield.

 

PERIOD ENDING DECEMBER 31, 1957

A
2 2-01-057-N-384

CONF. OIL Be BORAL  STEEL  BORAL LiH
SYMBOL NO. {cm} (in) (in) (in.) tiny  (in)
. 2-A 4+ 4 0 4 0 12 +0L
s e-F {4 4 0 12 +0L
o 2-6 4 Y% 4 Y, 12 4ol

GAMMA -RAY TISSUE DOSE RATE (ergs/g-hr-watt)

 

60 80 100 120 140 160 180
Z,,DISTANCE FROM SOURCE PLATE (cm)

Fige 5.2.27. Gamma-Ray Tissue Dose Rates Beyond
Configurations 2-A, 2-F, and 2-G:
Boral Within Configurations Containing Stainless Steel.

Effect of Inserting

The results of this test are to be compared with
measurements beyond similar configurations in
which no boral was used (configurations 2-B and
2-D in Fig. 5.2.23), in which cases increasing
the thickness of the oil layer preceding the
configurations from 11 to 31 cm reduced the dose
rate approximately 38%.

A comparison of configurations 2-H and 2-I,
based on the assumption that the thicknesses of
the oil layers preceding the two configurations
are essentially the same, shows the effect of
placing a boral sheet in front of the stainless
steel rather than behind it. The dose rate beyond
configuration 2-H is approximately 28% lower than
that behind configuration 2-1. A comparison of

269
ANP PROJECT PROGRESS REPORT

configyegtions 2-1 and 2-J shows the effect of
using boral both in front of the stainless steel
and behind it, rather than just behind it. The
resulting decrease in the dose rate was about 35%.

2—-01-057-T71-386

DEPLETED
ONF. OIL Be BORAL URANIUM BORAL LiH
. fem} {in)  (in) {in.} (in.) (in.)
42 4 0 14 0 12 40
10 4 s 12 0 12 +0L

1.0 4 '/2 f'/z '/2 {2 +0IL

GAMMA-RAY TISSUE DOSE RATE (ergs/q-hr watit)

 

60 BO t00 120 140 160 180
Z,, , DISTANCE FROM SOURCE PLATE (cm)

Fig: 5.2:28. Gamma-Ray Tissue Dose Rates Beyond
Configurations 3-A, 3-F, ond 3-G:
Baral  Within

Uranjum.

Effect of Inserting

Configurations Containing Depleted

270

Covdations 2-L and 2-M were tested to
determine whether an oil layer within the configu-
ration is best located in front of or behind a boral
sheet preceding the stainless steel gammea-ray
shield. The measurements show that the dose
rates beyond the two configurations are the same,
and apparently indicate that most of the neutrons
which cause capture gamma rays in the steel are
thermalized in the beryllium rather than the oil.

In order to determine the effectiveness of oil
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
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
from 17 ergs/g-hr-w for a l-cm-thick oil layer to
approximately 3.5 ergs/g-hr-w for a 3l-cm-thick
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
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.-fhick boral sheets within
the configurations was also studied. From the
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 gomma-ray shield is no more effective in
reducing the flux than placing a single sheet in
front of the shield only.

Other

series.

configurations will be tested in this

 
L

3x102

10

GAMMA-RAY TISSUE DOSE RATE (ergs/g hr watt)

70

Figu 502!29-

CONFIGURATION OIL
SYMBOL 0. {cm)

9.7
1.2
9.0
27.8
1.0
1.0

80 S0 100

Be

 

PERIOD ENDING DECEMBER 31, 1957

2-01-057-71-38

STAINLESS
OlL BORAL OlIL STEEL BORAL
(cm) {in.) (cm) (in.)

110 120 130 40 150 160

Z,, DISTANGE FROM SOURCE PLATE {em)

Gamma-Ray Tissve Dose Rates Beyond Configurations 2-H Through 2-M: Effect of Inserting Boral

and Qil Gaps Within Configurations Containing Stainless Steel.

271

 
ANP PROJECT PROGRESS REPORT

2—01—057—71—-287

O INITIAL RUN
A RERUN

CONFIGURATION:
OIL GAP+ 4in. OF STAINLESS STEEL +OIL

GAMMA-RAY TISSUE DOSE RATE (ergs/g-hr-woit)

 

0 10 20 30 40 50 60 70
THICKNESS OF Ol GAP PRECEDING CONFIGURATION {cm)

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
4-A Through 4-H).

2—04—057—!1—-388

5 X 10°

OiL ONLY IN CONFIGURATION TANK

CONFIGURATION 5-A:
1.0cm OF QIL+4in, OF Be
+12in. OF LiH+OIL

10

GAMMA-RAY TISSUE DOSE RATE {ergs/g-hr-watt)

 

40 60 80 100 120 140 160
Z, , DISTANCE FROM SOURCE PLATE (em)

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

Z—OM—MZ

3x103

CONF. OIL Be BORAL HEVIMET BORAL LiH
SYMBOL NO. (cm) (in) (in.}) (in) (in) (in)

& 1-A 29 4 O 2 0 12 +0IL

10 o i-F 1.0 4 Y 2 0 12 +0IL

. 1-6 1.0 4 /2 2 Y2 12 +0IL

— 2
g
o 102
@»
“
[)
E s
L
n
c
2
3
< 2
X
o
T 10
=z
£
= 5
D
L
=z
{
I 2
=
o
w
I
- {

50 60 70 80 90 100 10 120

2, DISTANCE FROM SOURCE PLATE (cm)

Fige 5.2.32. Thermal-Neutron Fluxes Beyond Con-
figurations 1-A, 1-F, and 1-G: Effect of Inserting

Boral Within Configurations Containing Hevimet.

 
 

-
w»

2-01-057-71-389

STAINLESS
CONF. OIL Be BORAL STEEL  BORAL LiM
SYMBOL NO. (cm) (in) (in) {in.) {iny (in.)
° 2-A 1 4 o] 4 Q {2 +0IL
a 2-F { 4 12 4 0 12 +0IL
o 2-6 t 4 15 4 5 12 +0I1L

ONLY IN
CONFIGURATION TANK

THERMAL - NEUTRON FLUX (neutrons/cmZ2 - sec - watt)

 

60 70 80 90 00 110 120
Zqy . DISTANCE FROM SOURCE PLATE (cm)

Fige 5.2.33. Thermal-Neutron Fluxes Beyond Con-
figurations 2-A, 2-F, ond 2-G:
Boral Within Configurations Contoining Stainless Steel.

Effect of Inserting

 

PERIOD ENDING DECEMBER 31, 1957

<.
2-01~057-71-390

DEPLETED
CONF. OlIL Be BORAL URANIUM BORAL LiH
SYMBOL NO. (in) {in) (in) (in.) tin)  (in)
. 3-A 42 4 0 1 0 12 + OIL
a 3-F 10 4 Yo s 0 12 + 0OIL
o 3-6 1.0 4 Vs 1 Yo 12400

THERMAL -NEUTRON FLUX (neutrons /cm2- sec - watt)

 

50 70 90 110 130 150 170
7y, DISTANCE FROM SQURCE PLATE (¢m)

Fige 5.2.34. Thermal-Neutron Fluxes Beyond Cone
figurations 3-A, 3-F, and 3-G:
Boral Within Configurations Containing Depleted

Effect of Inserting

Uranium.

273

 
L

e

 
5.3. TOWER SHIELDING REACTOR Il

C. E. Clifford

MECHANICAL DESIGN'

The design of the core of the Tower Shielding
Reactor 1l (TSR-I1) has been modified slightly to
provide more clearance for the thermal expansion
that will occur when the reactor is operating at
a power of 5 Mw.
planned to place the lead-boral shield adjacent
to the aluminum pressure tank. In order to ensure
that the tank is not heated above o safe level
by gamma-ray heating in the lead and neutren
heating in the boral, it will be separated from
the shield by a l/4-in.-'rhick layer of water, as
shown in Fig. 5.3.1. The lower portion of the
lead-boral shield will be hemispherical, and it
will be suspended from a ring on the tank wall
at the horizontal midplane of the core. With this
arrangement there will be sufficient room for the
shield to expand downward with heating. Radial
expansion at the midplane will also be permitted.
When the reactor is rotdted about a horizontal
axis, the shield will be restrained from moving
by keys on the hemisphere which will engage
In order to minimize

In earlier layouts? it was

slots in the support ring.
fission gamma-ray leakage through the break in
the midplane, the lead shield will be divided so
that an extension of the interface will not intersect
any of the fuel region.

The annular fuel elements and the lower central
fuel elements will be mounted on the central cyl-
inder.  Each fuel element will be fixed at the
top and guided at the lower end to permit some
growth. These annular fuel elements will be re-
movable remotely from above. The lower central
fuel elements will be bolted to the cylinder so
that they cannot be removed while the annular
elements are in place. This is to prevent the
possibility of lifting the central fuel
elements into the midplane of the core region
with the annular elements in place.

CONTROL MECHANISM

lower

A full-scale mockup of the control mechanism
has been designed, as shown in Fig. 5.3.2. |t

 

¥ The engineering design work for the TSR-H is being
performed by the ORNL Engineering Department,

2C. E. Clifford and L. B. Holland, ANP Quar. Prog.

Rep, March 31, 1957, ORNL-2274, p 294; june 30, 1957,
ORNL-2340, p 323; Sept. 30, 1957, ORNL-2387, p 304.

L. B. Holland

will have a 2.5-in. stroke, and the method of
operation will be the same as that for the pro-
totype, which was described previously.3 Normally
the spring will hold the control plate against the
fuel. Water will flow through the mechanism until
the flow area around the piston is closed by the
traveling nut. The pressure will build up in the
cylinder, and the control plate can then follow
the traveling nut to control the reactor. Inter-
rupting the water flow will scram the mechanism.

NUCLEAR CALCULATIONS?

The preliminary nuclear calculations® were per-

formed for a TSR-II design in which there was a
water reflector exterior to the core, as shown in
Fig. 5.3.3. By a modification of the 3G3R Oracle
code, this reflector was represented as a boundary
condition, so that the internal water reflector
could be divided into two regions for control
studies. A recent further modification of the
3G3R code provides a means of representing
the several layers of the external reflector (see
Fig. 5.3.1) as a set of boundary conditions.

The two outermost regions of the spherical re-
flector can be removed at one time. These two
regions are assumed to be arranged in their exact
geometry with a critical core region inside. The
values of a are then solved for in the following
equations:

by + 2 ay, =¢, -2, ,
(B + 2]y agy + Dy + 2 ) ay, =b, ~ 2, ,

where

ay, = number of fast neutrons incident on the
reflector that are returned as fast neu-
trons,

ay, = number of fast neutrons incident on the
reflector that are returned as thermal
nevtrons,

 

3c. E. Clifford and L. B. Holland, ANP Quar, Prog.
Rep. June 30, 1957, ORNL-2340, p 328.

4The nuclear calculations for the TSR-ll are being
performed by M. E. LaVerne.

5C. E. Clifford and L. B. Holland, ANP Quar. Prog.
Rep. March 31, 1957, ORNL-2274, p 294,

275

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED

  
  
 
  
       
    
     
 
 
 
  

  

2-01-060-37
CENTRAL CYLINDER
{ALUMINUM)
SHIELD
SUPPORT RING
ANNULAR FUEL ELEMENT
- +fi

~WATER FLOW
| RESTRICTION
\

CENTRAL FUEL ELEMENT

 

  

SINGLE PLUG
FUEL ELEMENT

LEAD

BORAL

ALUMINUM

PRESSURE TANK

(ALUMINUM)

WATER

Fig. 5.3.1. Suspension System for TSR-1l Fuel Elements and Lead-Boral Shield Section.

276

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-060-414

DRIVE SHAFT

OPERATING WATER FLOW

CONTROL PLATE

WATER FLOW ACTUATING
SEAT SWITCH

DRIVE NUT

SHOCK ABSORBER
POCKET

SPRING

RTRAVEL DEVICE

 

3 INCHES

Fig. 5.3.2. Experimental Mockup of the Control Mechanism for the TSR-IL.

277

 
ANP PROJECT PROGRESS REPORT

WATER FLOW

   
     
 

 

 

 

 

 

 

 

 

 

DIMENSIONS
4 16,20 OR 25cm
B 16,20 OR 25cm
C 20 em

Fig. 5.3.3. Vertical Section of the Reactor Core Model

a,, = number of thermal neutrons incident on
reflector that are returned,

¢, = fast-neutron flux,

¢, = thermal-neutron flux,

], = fast-neutron current,

], = thermal-neutron current.

The values of a can be obtained by running two
cases, one of which has o black shell placed
between the core and the reflector to eliminate
the thermal current. The values obtained for a
are independent of the ratio ¢,/¢.,.

Two additional terms can be derived, as follows:

 

 

 

2y 1=y, 1

) =]+a11=351'
2], 1= ay 12 b
@, T aysy - (1 + a; )0 + a,,) E

278

UNCLASSIFIED
DWG 2-01-060-7

LOCATION OF CYLINDRICAL
TUBE IN ACTUAL REACTOR

INTERNAL WATER REFLECTOR

MOVABLE BORAL SHELL

CORE REGION (WATER PLUS
URANJUM - ALUMINUM FUEL
PLATES)

 

 

 

 

 

 

 

EXTERNAL WATER REFLECTOR

Used in the Nucleor Parameter Study for the TSR.II.

The £ terms, which are not independent of the
ratio ¢ ,/¢é,, are input numbers of the 3G3R Oracle
code and represent boundary conditions. Thus a
new reactor configuration can be studied which
has three regions and boundary conditions that
represent two additional regions that have been
removed, This can be repeated to remove several
regions.

This procedure was used to strip off four outer
regions of the TOSR-Il ond represent them as
boundary conditions, the four regions being a
20-cm-thick water reflector, a ‘/2-in.-1'hick boral
shell, the 2-in.-thick lead region, and an aluminum-
water region. With these boundary conditions, cal-
culations were performed to obtain the critical
mass ond the amount of reactivity available for
control in a reactor having a 17.5-in.-dia internal
reflector and a 5.5-in.-thick core region which
were separated by a ]{‘-in.-thick shell of aluminum
and water. From these calculations it was found
that a thermaily black control shell which could
move away from the core in the internal reflector

region could control 12% in %. Since the control

 

 
= .

FERIOD ENDING DECEMBER 31, 1957

plates actually will cover only one-half this areq, UNCL ASSIFIED
2-01-060-39

their total worth could be 6% in k.

From the information now available, it is possible
to estimate the total amount of excess reactivity
required to compensate for the following effects:

Xenon bui|dup6 (for 10 hr of operation at 5 Mw) 0.8%

Temperoture rise (estimated from Bulk 0.7%
Shielding Reoctor data)

Shield chonges 1.0%
Control 0.5%
Teotal 3.0%

Two other items should be included in such o
list: (1) burnup of fuel and (2) possible errors
in fuel loading as a result of errors in the cal-
culation of the critical mass or fabrication of
the fuel plates. In order to provide for these
twe factors, calculations of the critical mass
were made for the condition of a critical reactor
with a solid control shell 1.6 ¢cm from the fuel.
Thus, when the control plates are against the
fuel, the reactor is 3% subcritical, and, when
the reactor is just critical with the control plates
1.6 cm from the fuel, an additional 3% in reactivity
can be obtained to compensate for temperature
rise, etc. by moving the plates farther from the
fuel. Fixed boron-loaded plates can be mounted
1.6 cm from the fuel in the areas not covered by
the movable control plates. The fixed control
plates can be adjusted either way to provide a
3% change in reactivity at the time of the initial
critical experiment and subsequently as required
by fuel burnup, depending, of course, on the
amount and direction of the initial adjustment,

The relative fast- and thermal-neutron flux dis-
tributions in the internal reflector ond core region
for the above operating condition are shown in
Fig. 5.3.4 ond Table 5.3.1. The average nor-
malized fluxes and currents are shown in Table
S

Nuclear calculations are being continued to de-
termine the temperature coefficient of reactivity
and the reactivity worth of voids in the core and
internal reflector. This information will be used
to study the kinetic behavior of the reactor and
its control system. A kinetic onalysis of the
water system is presently being set up for study.

 

6Culcu|oted on the basis of data given by H. K. Clark
and J. C, English, Xenon Tables, DP-200 (May 1957).

 

Fig. 5.3.4. Neutron Flux in the TSR-1l with a Thermally
Black Control Shell in the Internal Water Reflector

1.6 cm from the Core-Reflector Interface.

The calculated critical mass for the reactor is
8.2 kg of U235, Since the fuel element devel-
opment was based on o loading of 8 kg, no addi-
tional developmental work will be required for
fuel plate fabrication. Additional checks on the
critical mass calculations will include a cal-
culation with a multigroup, multiregion reactor
code, probably the Cornpone code on the Oracle
computer, and a simple experiment at the Critical
Experiments Facility.

The proposed critical ossembly will not be an
exact mockup of the TSR-ll, but it will utilize
the TSR-1l dummy aluminum fuel elements. The
fuel will be contained in a water solution that
will fill the spaces between the aluminum plates.
Two experiments will be performed, each with a
different fuel concentration, from which the
critical mass can be deduced. In order to check
the validity of the method, o dummy aluminum
element has been fabricated for use in the Bulk
Shielding Reactor. This element is sealed, top
and bottom, at the end of the fuel regions so
that fuel, in solution, can be placed in the element;
the reactivity worth of the solution element can
then be compared with that of a regular element
which contains an equivalent amount of UZ33,

279
ANP PROJECT PROGRESS REPORT

Teable 5.3.1. Calculoted Fast-Neutron and Thermal-Neutron Flux Distributions
in a TSR-ll Configuration with a Control Shell 1.6 cm from the Fuel

 

Distance from Center

 

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
2.76 0.6725 0.0865
20.76-21,40*
21.40 0.6332 0.0685
21,80 0.6661 0.1417
22,20 0.70 11 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
3127 0.9460 0.2095
31.91 0.9215 0.2041
32.55 0.8931 0.1979
33.18 0.8611 0.191
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 1/4-in.-fhick control shell.

** At reflector-core interface.

280
 

 

 

 

 

 

 

Table 5.3.2, Average Normalized Fluxes and Currents
in TSR-Il Configuration with a Control
Shell 1.6 cm from the Fuel

 

Fast Thermal

Region Neutrons Neutrons

 

Average Normalized Flux

Internal reflector, inside 0.4068 0.4505

control shell

Internal reflector, outside 0.7046 0.1759
control shell

Core 0.8674 0.1962

Net Current

Outer core boundary 0.1003 0.0030

 

WATER COOLING AND PURIFICATION SYSTEM

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
of February 1958 and the work is scheduled to
be completed in September. Earlier estimates had
indicated that the work could be completed by
July 1.

A flow diagram of the water system, along with
tentative locations of the pressure- and tem-
perature-measuring elements, is shown in Fig,
3.3.5. The main components of this system
will be:

1. a 60-hp circulating pump capable of delivering
1000 gpm when the reactor is operating at a
200-ft altitude,

2. a venturi nozzle for measuring water flow,

3. the reactor, which will operate up to a 5-Mw
heat output,

4. a two-cell, air-cooled heat exchanger capable

of removing 17,065,000 Btu/hr from water
entering at 160°F and leaving at 125.8°F when
the ambient air temperature is 75°F,

5. a 1500-gal detention tank,

There will also be the following auxiliary units
of the flow system:

I. a main bypass valve for reduced flow when
the reactor is operating in the power range
from 5 kw to 5 Mw;

2. a 5-hp d-c emergency citculating pump which
will operate continuously and which is re-
quired for removing afterheat following any

PERIOD ENDING DECEMBER 31, 1957

emergency shutdown that may occur while the
radiator is operating at 5 Mw and for cir-
culating the water while the reactor is op-
erating at a power below 5 kw;

3. o water-level control system which will consist
of a 250-gal level tank, compressed gas to
remove water from the system, and a fill pump
which will take water from the pool and put
it into the system;

4., a 40-kw heater to prevent water lines from
freezing when the reactor is not operating (the
air cooler will be supplied with louvers and
additional heaters);

5. a mixed-bed demineralizer which will bypass
_the detention tank with a flow of 40 gpm and
which can also be used to keep the reactor
poocl| water clean;

6. a Tracerlab fission break monitor which will
give an alarm if iodine activation, which would
indicate a rupture in a fuel element, is noted
in the reactor outlet water.

EXTENSION OF GAMMA-RAY-HEATING
CALCULATION FOR THE TSR-II

W. W. Dunn?  D. K. Trubey

In the previous computation® of the gamma-ray
heating in the TSR-Il, two assumptions were made
concerning the angular ond energy dependence of
the gamma-ray flux at the core-shield interface.
The first assumption was that the collided energy
spectrum was represented by the NDA differential
energy spectrum after passage fhrough one mean
free path of water,? ahd the second was that the
collided and uncollided fluxes were isotropic at
the core-shield interface. In order to check the
validity of these assumptions, an additional cal-
culation has now been completed by using the
Bowman-Trubey Monte Carlo data on the gamma-
ray penetration of lead shields'® in order to
compute the heating in the first % in. of lead
in the shield, that is, the lead adjacent to the
core. This first lead interval was chosen as

 

70n assignment from U.S. Air Force.

8W. W. Dunn and D. K. Trubey, ANP Quar. Prog. Rep.
Sept. 30, 1957, ORNL-2387, p 307316, esp 312,

9H. Goldstein and J. E. Wilkins, Calculations of the
Penetrations of Gamma Rays. Final Report, NY0-3075
(June 30, 1954).

IOL. A. Bowman and D. K. Trubey, ANP Quar. Prog.
Rep. Sept. 30, 1957, ORNL-2387, p 320.

281
 

ANP PROJECT PROGRESS REPORT

<10 psi

  

WATER TEMPERATURE 126°F 2 1.5°

UNCLASSIFIED
2-01-060-40

 

 

   

VARIABLE PITCH
CONTROL
AIR-COOLED HEAT
EXCHANGER
25hp;
460v—3¢

 

 

VALVE SYMBOLS

 

P NORMALLY CLOSED
<] NORMALLY OPEN
P<] FLOW CONTROL
T~ CHECK
INSTRUMENT SYMBOLS
F — FLOW M -~ MOTOR
P — PRESSURE R — RECORDER
T — TEMPERATURE L — LEVEL
X — SWITCH d — DIFFERENTIAL

_ — CENTRAL PANEL
O LOCAL INDICATOR @ INDICATOR

282

 

MANIFOLD AND TWO-STAGE
REGULATOR, 2000 TO
0-200psig RANGE

VARIABLE PITCH
CONTROL

FISSION BREAK MONITOR

AND PUMF TRACERLAB
MwP -1 SYSTEM

   

NOTE:
INSTRUMENT SYMBOLS ACCORDING TO
INSTRUMENTATION AND CONTROLS DIVISIDN STANDARDS

FLOW <10gpm

 

 

 

 

    
 

4

MIXED -BED

DEMINERALIZER
P

TO RETENTION
POND

  

COMPRESSED AIR
CYLINDERS

 

 

ALARM WHEN VOLUME RESIST-
IVITY <5 x 105 ohm -cm
pH =55

(CAN

(Fa)
_/

¥4—in. VENT LINE

  
 
   
 
 
  
 
  

 

 

 

 

 

 

 

 

 

>340F

 

%

 

HIGH WATER

“‘} FILL PUMP

__) circurt

CONTROL

LOW WATER

250-gal WATER LEVEL
CONTROL TANK

STRAINER

DEMINERALIZER PUMP
CONTRQL CIRCUIT

|t
i I'TO FiLt. PUMP CONTROL CIRCUIT

ol

 

 
    
    
     
 
     
   

150 psig

1500~gol
DETENTION TANK

1000-gpm
MAIN PUMP

 

<]

 

S50 gpm
DEMINERALIZER PUMP

 

 

 

60 hp;
460v-3¢

10 TO 40gpm
EMERGENCY PUMP

5 hp;
120vDC

 

B

  
 

 

rcd

 
 
   
    
  

 

2000 gpm

ACCURACY {7

MAIN BY-PASS VALVE

RANGE: 200 TO 1200gpm
ACCURACY *2 %, AT 1000gpm

HEAT POWER

RANGE: 200 TO

 

   

~d— -

FROM CONTROL ROOM

LIMIT STROKE SPEED PER

REACTOR SAFETY

MANUAL FROM
CONTROL ROOM

- AUTOMATIC FROM

 

 

10 TO 40 gpm
FILL PUMP

CONTROL CIRCUIT

 

 

D<-

 

BLANK FLANGE h——»q————

Shp; 460v -3¢

 

FLOW RANGE 1O TO 50 gpm
ACCURACY 5,

(FAY <tOgpm
./
(FRY
N

 

 

MAX. TEMPERATURE 126 °F
MIN. TEMPERATURE 50°F

TC

 

 

HEATER ; 40kw

 

 

 

 

 

 

 

 

 

 

 

NOTE:

ALL ANNUNCIATOR FIELD CONTACTS OPEN ON TROUBLE.

EXISTING REACTOR POOL

Fig. 5.3.5. Cooling Water Flowsheet for the TSR-Il,

gl
w.

 
the check point, since a large part of the gamma-
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
vtilized.

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(f), the percentage
of initial energy E, deposited in the first 1/8 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
identical to those produced in the previous cal-
culation when [(6), the percentage of initial
energy deposited, was plotted for various depths
of penetration into the shield as a function of
incident angle. It was possible to compare /(6)
and g(f) at 0.05 mean free path of material pen-
etrated, and that comparison is also shown in
Fig. 5.3.6. Since the minimum energy of the
Bowman-Trubey data was 1 Mev, the Monte Carlo
code was used to produce data at 0.5 Mev for
use in this calculation.

The normalized gamma-ray flux, A(EO, shell),
for each E, from each source shell was found
from the following expression, which incorporates
the wusual transformation from a spherical-shell
source to an equivalent infinite plane source
(neglecting the back plane):

1 |+~ }
(1 —2"'[; S(Eu,shell) b(r) th}

m/2
=f A(E g, shell) sin6df ,
0

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-060-32
0.40

0.35

0.30

£(8), ur=0.05

0.25

0.20

¢{8) OR £{8)

0.15

0.10

0.05

 

0 15 30 45 60 75 90
8 (deg)

Fig. 5.3.6. Per Cent of Energy Absorbed in F’b.| as
a Function of the Incident Angle for Various Source
Planes; E = 4 Mev,

where
r = distance from center of core to
spherical source shell, e¢m,
ro = distance from center of core (or
center of internal water reflector)
to core-shield interface, cm,
S(E,, shell) = source term for each shell, number

of gamma rays of energy E  gen-
erated per square centimeter per
fission per second,

thermal-neutron flux in the source
shell at point r,

©-
=
i

Z/ = macroscopic fission cross section
of U233,
t = shell thickness, em.

With g(6) and A(Eo,shell) for each value of E,
and @ in each source shell, the gamma-ray heating
in the first ]/8—in. lead interval was found by

283
ANP PROJECT PROGRESS REPORT

summing over E; and the source shell the nu-
merically integrated expression:

(2) H(E,)

= [, 80}, shell) 4(E, shell) sin 6 cos 649 ,

where the cos & is required for the flux-to-current
Table 5.3.3 presents the result of the
summation of Eq. 2 compared with the results
obtained in steps 1 and 2 of the previous cal-
culations.®  The heating estimated by step 1
was determined by multiplying the heating at the
core-shield interface, as a function of initial
energy group, by the ratio of the lead to core
energy absorption coefficients for that energy

group.

conversion.

Table 5.3.3. Heating in the First I/a-in.-Thick Layer
of Lead in the TSR«ll Shield Calculated
by Various Methods

 

Calculation Heating (wafl's/cmz-Mw)

 

Previous calculation

Step 2 0.445
Present calculation 0.215

 

It may be seen that the result from step 1 of
the previous calculation represents the extreme
limiting case, since it was assumed
that the flux was the same in the !’B-in. lead
interval as in the core. It is expected that step 2
of the previous calculation also gave an over-
estimate of the true heating for several reasons.
First, the NDA buildup factors assumed an infinite
medium and did not take into account the reduction
in the low-energy fluxes at the core-shield inter-
face, which in turn would reduce the heating.
This total flux reduction, which is not reflected
in the tabulated resuits, is of the order of 20%
for the 1.0-Mev fluxes'! and would be much
greater for the lower energies and less for the
higher energies, Second, the collided fluxes
which were obtained from the point kernel and

maximum

 

”Unpublished results of the Bowman-Trubey calcu-
lation.

284

buildup factors at the core-shield interface were
assumed to be in such a direction as to penetrate
the shield, and this would not be entirely true
in the actual case, especially at low energies.
Other effects which would contribute to
an overestimate were: the assumption of nearly
isotropic distribution for the collided fluxes,
which would overestimate the heating in the first

small

lead interval and underestimate the heating in
the succeeding layers if the collided flux were
more peaked in the forward direction, as is
probably the case; and the assumption of the
differential energy spectrum for water to represent
the core, which would overestimate the low-energy

group, since the effect of the aluminum and
vranium would be to reduce the low-energy
spectrum.

The present calculation by necessity required
considerable interpolation to perform the necessary
cross-plotting, and for this reason there is an
uncertainity in the results, The neglect of the
source gamma rays in the source planes directed
away from the shield is probably the one source
of error which lowers the calculated result. While
this calculation avoids many of the errors inherent’
in step 2 of the previous calculation, the use of
the transformation from a spherical shell source
to an infinite plane source does introduce some
error, since the medium is not infinite and homo-
geneous.

In conclusion, it can be stated that the original
assumptions on energy spectra and angular de-
pendence in the previous calculation (step 2)
give a reasonably conservative estimate of the
heating in the first lead interval. Since the heating
in this interval is most dependent on the low-
energy spectrum, it would be expected that the
overestimation caused by the energy-spectra as-
sumption would be decreased in succeeding lead
layers, and hence there is the prospect of im-
provement with penetration into the shield. The
angular-distribution assumption gives an over-
estimation in the initial layer, and, if the collided
flux is more peaked in the forward direction, it
tends to underestimate the heating deep in the
shield.  No further refinements of these cal-
culations are contemplated, since the heating is
considered to be sufficiently defined for purposes
of construction.
 

PERIOD ENDING DECEMBER 31, 1957

5.4. TOWER SHIELDING FACILITY
C. E. Ciifford

STUDY OF GAMMA RAYS PRODUCED BY
NEUTRON INTERACTIONS IN AIR

Y. R, Cain F. J. Muckenthaler

During the past few years a large number of
measurements have been made of gamma-ray dose
rates inside crew shields for several shielding
configurations. Dose-rate measurements of this
type, however, give very little information as to
the original energy of the gamma rays entering
the crew compartment and still less as to their
origin, Gamma-ray dose rate measurements in
the crew shields have indicated an intensity
greater than that expected from the scattering of
gamma rays alone. Therefore, consideration has
been given to the possibility of secondary gamma-
ray production from neutrons leaving the reactor
shield.

inelastic scattering, as well as elastic scattering,

{nteractions of neutrons in air include

Calculations of
gamma-ray intensities from inelastically scattered
neutrons are somewhat difficult because neither
the scattering cross section nor the energy dis-
tribution of gamma rays for this interaction is well
known, The fact that the neutron energy spectrum
leaving a shield is difficult to measure further
complicates the problem, Measurements for de-
termining the inelastic-scattering cross section
for neutrons in air would be very difficult to make
at the TSF., An experiment was performed, how-
ever, in which gamma-ray spectral measurements
were made from which it is hoped that the order of
magnitude of the dose rate to be expected from
inelastic scattering can be deduced.

and thermal-neutron capture.

Experimental Configuration

In order to detect the high-energy gamma rays
resulting from neutron interactions with the air,
it was necessary to eliminate, where possible,
other sources of high-energy gamma rays. Since
the number of gamma rays born in the reactor was
many orders of magnitude larger than the number
born in the air, the reactor had to be well shielded
from the detector., This was accomplished by
submerging the reactor in the handling pool 8 ft
below the surface of the ground and thus inter-
posing approximately 40 ft of earth as shielding.

As shown in Fig. 5.4,1, the experimental con-
figuration included an 8-in.-dia by 10-ft-long air-
filled collimator attached to one face of the
reactor so that a beam of radiation was emitted
into the air at an angle of 60 deg from the hori-
zontal, The axis of the collimator passed through
the center of the face of the reactor. A partially
shielded sodium iodide crystal (3 by 3in.) mounted
on a DuMont 6363 phototube was located approxi-
mately 1 ft above ground level at a horizontal
distance of 52 ft from the reactor. The crystal
was enclosed in a cavity surrounded by an 8-in,
thickness of lead and at least 4 ft of concrete.
On the side of the detector shield facing the re-
actor, a 4-in,-wide collimator slit extended through
the shield to the detector position, The floor of
the slit slanted 10 deg upward from the detector
position so that no direct radiation from the
ground would be admitted. The crystal and photo-
tube were mounted so that the axis of the photo-
tube was at an angle of 45 deg from the horizontal
to ensure maximum sensitivity,

The collimator slit, which was 5 ft long and 5 ft
deep at the deepest point (at the detector position),
was filled with borated water contained in a 4-in,
by 4-ft by 4-ft plastic bag., The borated water pre-
vented neutrons from reaching the detector and
also reduced the number of capture gamma rays
produced in the concrete walls of the collimator,
In addition, a 3’/]6-in.-thick sheet of boron-impreg-
nated Plexiglas (hereafter referred to as boron
Plexiglas) was placed over the outside of the
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
centimeter.) The effectiveness of this neutron
shielding was verified by a series of measurements
with a BF; chamber and a fast-neutron dosimeter
placed in the detector cavity. The measurements
indicated that there was no detectable neutron
flux in the cavity.

The reactor and its collimator could be rotated
the vertical axis of the reactor. The
measure of this rotation is indicated by the
angle 6, which is the angle between a vertical
plane including the reactor center and the detector
center and a vertical plane including the reactor

about

285
ANP PROJECT PROGRESS REPORT

-

8

DETECTOR?% -t /

PLAN VIEW {NCT TO SCALE)

REACTOR

 

 

 

 

 

UNCLASSIFIED
2-01-056-22-D-594

  
   
    
       
 
 
  
  

  
 

LONG - ™~

REACTOR TANK

  

8-in-DIA x 10-ft-
COLLIMATCR

 

TSF REACTOR

 

3-x 3-in. Nal(T1) CRYSTAL
AND PHOTOMULTIPLIER

—g-in,-THlCK LEAD SHIELD ARCUND 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-
ever, most of the measurements were taken with
@ = 0 deg, and for this geometry the infersection
of the detector lines of sight with the cone of
radiation from the reactor is depicted in Fig. 5.4,2,
It can be seen that the detector solid angle easily
includes all the radiation cone near the ground
level, but the fraction of the cone included in the
solid angle decreases as the distance from the
reactor is increased.

Other than the measurements with 6 = 0 deg,
measurements were actually made for only one
other value of 6, that is, 6§ = 90 deg, @ being
measured in the clockwise direction from the
reactor-detector axis. For the measurement taken

286

with this beam angle, additional concrete shielding
was placed above and in front of the detector
collimator to reduce the number of primary gamma
rays scattering from the reactor beam that would
enter the detector collimator and be scattered into
the detector.

Determination of Detector Response

As mentioned above, the detector consisted of
a 3- by 3-in. sodium iodide crystal mounted on a
DuMont 6363 phototube. Pulses from the tube
were sorted by two 20-channel analyzers connected
in series, and the pulse-height output was cali-
brated with respect to energy against the 1,12-Mev
Zn®3 gamma ray.
UNCLASSIFIED
2-01-056-22-D-598

   
   
      
   

CONE OF RADIATION FROM
REACTOR COLLIMATOR

FIELD OF “VISION"
OF DETECTOR T

TSF

 

  

 

 

 

 

 

 

| o REACTOR
-~ @
- / |
-
CYLINDRICAL
REACTOR COLLIMATOR
5 o 10

/' DETECTOR COLLIMATING SLIT

Fig. 5.4.2. Geometric Relationships When @ = 0 deg:

TSF Experiment To Measure the Spectro of Gamma Rays
Resulting from Neutron Interactions in Air.

|

In order to better understand the effect that the

borated water in the collimator might have on the

gamma-ray spectrum, a polonium-beryllium source

was placed outside the collimator and the pulse-

height spectrum of the 4,4-Mev gamma ray from the

Be9(a,n)C]2* reaction was observed. The re-

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

the Van de Graaff accelerator with the N! 5(p,a)C 12*

- reaction when no collimator was used.! It can be

seen that the borated water caused a significant

reduction in the magnitude of the peak correspond-

ing to the 4,4-Mev gamma ray and that the Compton
scattering was increased substantially,

 

p, R. Bell, private communication,

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-056-22-597

PULSE HEIGHT
0 200 400 600 800 1000 1200

3.93 Mev

4.44
3.43

N'S(p,a)C'2% NO COLLIMATOR

counts/sec

3.90 Mev
3.40

445

Be? (a,n) C12* WITH COLLIMATOR

 

0 100 200 300 400 500 600
PULSE HEIGHT

Fig. 5.4.3. Response of Detector to 4.4-Mev cl?

Gamma Rays. Top curve represents response of Nal
crystal to gamma rays produced at Van de Graaff ac-
celerator. Lower curve represents response to Po-Be

gamma rays filtered through a B-H,0-filled collimator.

Experimental Results and Discussion

The top curve in Fig, 5.4.4 represents the total
gamma-ray pulse-height spectrum resulting from
neutron interactions in air measured in the geometry
described above when 6 = 0 deg. This spectrum
presumably consists of gamma rays resulting from
the capture of thermal neutrons in the nitrogen of
the air plus those from other sources. The highest
energy gamma ray from thermal-neutron capture by

287

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-056-22-5395
20.00
BORON CARBIDE ON REACTOR COLLIMATOR
AND DETECTOR COLLIMATOR. 8§ = 9¢° ‘

10.Q0 NO BORON CARBIDE ON REACTOR COLLIMATOR;
BORON CARBIDE ON DETECTOR COLLIMATOR. 8 =0

BORON CARBIDE ON REACTOR COLLIMATOR
500 AND DETECTOR COLLIMATOR. 8 = O

DIFFERENCE (4 MINUS »)

BORON CARBIDE ON REACTOR COLLIMATOR AND

: DETECTCR COLLIMATOR; 6-in. PARAFFIN ON
2.00 DETECTOR COLLIMATOR. 8= 0

1.00

0.500

0.200 107 Mev

0.100

counts /sec
o
o
t
o

0.020

0.010

0.005

0.002

0.001

0.0005

0.0002

 

0.00014
0 100 200 300 400 500 600 700 800 200 1000 1100 1200

PULSE HEIGHT

Fig. 5.4.4. Total Pulse-Height Spectra of Gamma Rays Resulting from Neutron Interactions in Air,

288
N'4 which is known to have been observed? is at
10.8 Mev, and this can be identified in the curve
at a pulse-height setting of 1000. The two peaks
that occur at pulse-height settings of 960 and 920
are typical of high-energy gamma-ray responses in
a 3- by 3-in, sodium iodide crystal, The higher
peaks that correspond to pulse-height settings of
900 and 860 are caused by improperly operating
channels in the analyzer, Some of the high count-
ing rates below a setting of 150 were due to the
detection of Zn®® gamma rays from portions of the
calibrating source which had been left in the
borated of the collimator; low-energy
Compton-scattered gamma rays also contributed.

water

Unfortunately, it was not possible in the ex-
periment to isolate the gamma rays contributed .by
thermal-neutron captures from those contributed by
inelastically scattered fast neutrons. It was
possible, however, to reduce the effect of the
capture-gamma-ray contribution by blacking out
most of the thermal neutrons so that a difference
spectrum could be obtained which could then be
used to determine the approximate contribution
from inelastically scattered neutrons. In order to
obtain measurements with a reduced capture-
gamma-ray contribution, the upper end of the
reactor collimator was covered with boron Plexiglas.
The resulting gamma-ray spectrum is shown in the
third curve from the top in Fig. 5.4.4. It can be
seen that the 10.8-Mev capture-gamma-ray peak was
reduced in intensity by a factor of 40, However,
the portion of the curve that corresponds to lower
pulse-height settings was not reduced by the same
factor, which indicates that this portion of the
spectrum not only represented gamma rays which
originated from thermal-neutron capture but also
represented gamma rays from another source. This
other source is presumably inelastically scattered
fast neutrons that were unaffected by the boron
Plexiglas on the reactor collimator. |f this curve
is subtracted from the top curve in the figure,
which represents the total capture-gamma-ray con-
tribution plus the contribution from inelastically
scattered neutrons, the difference curve would
represent the spectrum of capture gamma rays from
low-energy neutrons which were eliminated by the
boron Plexiglas on the reactor collimator. This
difference curve, which is also plotted in Fig.

 

2G, A Bartholomew and P. J. Campion, Can. J.
Phys. 35, 1347 (1957).

PERIOD ENDING DECEMBER 31, 1957

5.4.4, shows very little fine structure; however,
the energy-level distribution in N' is such that
it is possible for gamma rays of the order of 15
different energies ranging from 3.5 to 10.8 Mev to
be emitted, and these would be smeared by scat-
tering in air and in the borated water and by inter-
actions in the crystal itself. Several energy levels
from 5.2 to 5.6 Mev are responsible for about
one-fourth of the gamma rays from thermal-neutron
capture in nitrogen, and this difference curve does
seem to indicate a prominent peak at about 4.5 to
5 Mev. Determination of the actual capture-gamma-
ray spectrum, however, can only be done by what
is known as the ‘‘subtraction’’ method, and an
analysis of this type is yet to be performed. The
peak observed in this difference curve at a pulse-
height setting of about 100 is the result of a shift
in gain during the foreground and background runs
which brought about improper subtraction of the
Zn®5 gamma rays. (While the reactor collimator
was covered with boron Plexiglas, one series of
measurements was made with the reactor collimator
rotated to ¢ = 90 deg. The results of these
measurements are included in Fig. 5.4.4, but no
attempt has yet been made to interpret the data.)
|f it is assumed that the 10.8-Mev peak obtained
in the measurements in which boron Plexiglas was
used on the reactor collimator is due only to
thermal-neutron capture in nitrogen, this peak can
be normalized with the same peak in the capture-
gomma-ray difference curve. The difference be-
tween these two curves, shown in the lower curve
in Fig. 5.4.5, was thought to be due to gamma rays
resulting from inelastically scattered fast neutrons,
The large number of high-energy gamma rays
(>2.25 Mev), however, indicated the possibility
that a significant contribution to this inelastic-
scattering spectrum might be the result of neutron
interactions in the concrete walls of the collimator.
(For example, silicon capture gamma rays have
very high energies.) In order to resolve this
question, additional measurements were made in
which the intensities of the neutrons and gamma
rays entering the collimator slit were changed
through the use of lead and several thicknesses of
paraffin placed over and in front of the slit.

Data obtained when 4‘/2 in, of paraffin, 9 in. of
paraffin, and/or 4 in. of lead was placed over the
collimator slit and boron Plexiglas covered the
reactor collimator are given in Fig. 5.4.6. For
these measurements the gain of the amplifier was

289
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01—056—22—596
20

2 DIFFERENCE CURVE!
NO BORON PLEXIGLAS ON REACTOR
COLLIMATOR MINUS BORON PLEXIGLAS

ON REACTOR COLLIMATOR

0.5

0.2

oA

COum‘s/sec
o
o
o

INELASTIC-SCATTERING GAMMA RAYS

0.02

0.04

0.005

0.002

0.00!

0.0005

0.0002

 

0.0001
100 200 300 400 500 600 ‘700 800 8900 1000 1400 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—041—-056—22—-670

 

10 T T T T T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 B NO SHIELD (BORATED WATER IN COLLIMATOR SLIT)‘
6 A 4', in. OF PARAFFIN COVERING SLIT
® 9in. OF PARAFFIN COVERING SLIT
O 4in. OF LEAD COVERING SLIT
4 A 4in. OF LEAD + 4 %, in. OF PARAFFIN COVERING SLIT
O 4in. OF LEAD +9 in. OF PARAFFIN COVERING SLIT
> 8 B ) BORON PLEXIGLAS ON REACTOR COLLIMATOR
Py X
o 8 A
° 5 T
* A
, ¢
8 w
=
6
[ 4 A . g
L ® A &%
4 ® A A T u B
A ® o 8 A A " g
H—&f u
0 e e v | B
a ® 9 A u
, & ° f A L
o A
0 : e A .
Q o 0 ® A a
@ oo ®
2 o940 oo g l‘)
—4 o O .
3 8 o —™
o - | J 0 »
a2 a A
[ A j_fl
Ak 84 4
4 O o
o L a ~ A g
59 e U
O A
A A 0 u
> 0O AC o 0 A é ~ ® A
A A% a A 8
o
o o o A o % ™
_ o A aA4g ®
1072 8 &
~ 0
8 A
6 0
U
4 0 O
0
2 ° AQ
A
1072
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 (8 = 0 deg).

291
ANP PROJECT PROGRESS REPORT

reduced so that a pulse-height setting of 100
corresponded to about 2.25 Mev. If a relaxation
length of about 30 cm for gamma rays in paraffin
is *assumed, the reduction in the gamma-ray in-
fensity was that to be expected for the two paraffin
thicknesses (a factor of 1.8 for 4]/2 in, of paraffin)
if the gamma rays were coming from outside the
slit. A comparable reduction was found when
6 in. of paraffin covered the collimator slit (see
Fig. 5.4.4). Had these gamma rays resulted from
neutron interactions in the detector collimator
slit, however, the gamma-ray intensity would have
been reduced much more (about a factor of 7 when
41/2 in. of paraffin was added).

When 4 in. of lead covered the siit, the counting
rate decreased only a factor of 5 below 4 or 5 Mev
and only a factor of 2 above this energy. This
large contribution above 5 Mev appeared to be due
to neutron captures in the lead. This was verified
by placing 4]/2 in. of paraffin on the lead, which
further reduced the counting rate a factor of about
10, A 9-in. thickness of paraffin was also placed
on the lead, as shown in Fig. 5.4.4, but the
additional 4]/2 in, did not effect a further reduction
in the gamma-ray intensity, since the first 4% in.
was sufficient to attenuate the neutrons. |t should
be pointed out that the measurements in which
both lead and paraffin were used were very near
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
rays were the result of interactions in air and not
from thermal-neutron capture in the detector
collimator. It is assumed here that the source
was almost entirely inelastically scattered neu-
trons; however, contributions from other sources,
such as multiply scattered gamma rays, may be
included,  Theoretical calculations of probable
gamma-ray spectra from inelastic scattering of
neutrons in N”, thermal-neutron capture, and
scatftered primary gamma rays are being made,

Thermal-Neutron Flux Measurements

Thermal-neutron flux measurements in the air
around the reactor were made with a bare single-
barreled BF ; counter to facilitate calculations of
the air capture gamma-ray intensities; the calcu-
lations will be made for the same configurations

292

for which the gamma-ray spectra given above were
measured, It should be emphasized that these
measurements give a count rate that is proportional
to the neutron density, The calibration procedure
used essentially multiplies the neutron density by
the thermal-neutron velocity (2200 m/sec), and
the result is the ‘‘effective’’ thermal-neutron flux.
This has only to be multiplied by the capture
cross section at 2200 m/sec to obtain the capture
rate in a 1/v absorber.

For the thermal-neutron flux measurements the
counter was suspended at a fixed point in air and
the reactor and its collimator were rotated about
the reactor vertical centerline, as shown in Fig.
5.4.7. The counter position fell along the center-
line of the beam when the beam was in the vertical

UNCLASSIFIED
2-01-05%6-22-D-678

DETECTOR \ ’, REACTOR AXIS

AXIS OF
COLLIMATED BEAM

   

 

Fig. 5.4.7. Geometry for Measurement of the Thermal-

Neutron Flux in Air,

plane defined by 9 = 0 deg. Measurements were
taken both with and without a boron Plexiglas
cover on the reactor collimator. The resulting
effective thermal-neutron fluxes, with and without
the collimator cover, are plotted as a function of
¢, the angle between the axis of the beam and a
line through the center of the reactor and detector,
in Figs. 5.4.8 and 5.4.9 for several separation
distances, r, measured from the reactor center to
the detector center, The same information is
given as a function of r for & = 0, 15, 30, and
60 deg in Figs. 5.4.10, 5.4.11, 5.4.12, and 5.4.13,
respectively, The curves of Fig. 5.4.9 have been
fitted with an expression of the form e"'/’\/rz, and

 
PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-056-22-6T1

r=216ft
r = 47.2 ft
r = 55.4ft
r=82.4ft
r=855#H
r=96ft

r=126ft
r =194 ft

o
A
o
n
A
®
0
¢

EFFECTIVE THERMAL—NEUTRON FLUX (”Vth /w)

 

60 50 40 30 20 10 0
¢, 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 ¢: No Boron

Plexiglas on Reactor Collimator.

 
ANP PROJECT PROGRESS REPORT

10
8
4
2
1
T 8
~
£
<
~ 4
>
o
1
w
z 2
O
o
}_
o
Y 40!
|
I
g
s 6
(et
T
I 4
Ll
>
—
o 2
LJ
L
L
Ll
1072
8
4
2
10°

60 50

r = 21.6 ft
= 4772 ft
r =554 ft
r = 824 ft
r = 96 ft
r =142.6 ft
r =194 ft

40

30

 

20

UNCLASSIFIED
2—01—056—22— 672

10 0 -

¢, ANGLE BETWEEN BEAM CENTERLINE AND REACTOR-DETECTOR CENTERLINE (deg)

Filg. 5.4.9. Effective Thermal-Neutron Flux at Various Distances r as a Function of the Angle ¢: Boron

Plexiglas on Reactor Collimator.

294
o

UNCLASSIFIED
2—01—056-22-673

o WITHOUT COVER

A WITH COVER

3

   

O

Ny

LA (n=5am)

—
o

o

~ e=* (A=961)
r

EFFECTIVE THERMAL-NEUTRON FLUX (7 1,)

- 0 20 40 60 80 100 120 140
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft}

Effective Thermal-Neutron Flux as a
With and

- Fig. 5.4.10.
Function of the Distance r for ¢ = 0 deg:

Without Boron Plexiglas on Reactor Collimator.

relaxation lengths of 54 and 96 ft, respectively,
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
the reactor with energies higher than thermal. For
this same reason, the curves of Figs. 5.4.11
through 5.4.13 cannot be fitted with a simple
expression.

In addition to the measurements described in the

preceding paragraph, a series of measurements

. was also made along short vertical lines at fixed
’ horizontal distances from the reactor as the reactor
and its collimator were rotated. These vertical

¢ lines were centered above and below the beam
axis defined by @ = 0 deg and served to show that

the neutron beam was symmetrical about its axis.

 

PERIOD ENDING DECEMBER 31, 1957

UNCLASSIFIED
2-01-056-22-674

o WITHOUT COVER
& WITH COVER

EFFECTIVE THERMAL-NEUTRON FLUX (rtvyy, /w)

 

0 30 60 90 100 150 180 210
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Effective Thermal-Neutron Flux as a
With and

Fig. 5.4.11.
Function of the Distance r for ¢ = 15 deg:
Without Boron Plexiglas on Reactor Collimator.

UNCLASSIFIED

2—01—056—-22-675

© WITHOUT COVER
A WITH R

O
i m

(%1

-
Ol
N

EFFECTIVE THERMAL-NEUTRON FLUX (nv*h/w)
w

 

Tod
0 30 60 90 120 150 480 210
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Effective Thermal-Neutron Flux as a
With and

Fi g. 5!4-]2.
Function of the Distance r for ¢ = 30 deg:
Without Boron Plexiglas on Reactor Collimator,

295
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2—0—056—22—676

.

I5

T o WITHOUT COVER

= 2 A WITH COVER

2

T

- 10

&

£ 5

2

W

T

7 2

z

W 1072

I

-

Y 5

|

m

w 2

b

w 3
10

 

0 30 60 90 120 150 180 210
r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Effective Thermal-Neutran Flux as a
With and

Figo 5-4-]30
Function of the Distance r for ¢» = 60 deg:

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
gamma-ray spectrometer.

A measure of the relative importance of thermal
neutrons from the source and higher energy neu-
trons from the source which have been slowed
down in air may be obtained by taking, for a given
position in space, the ratio of the measurements

296

for which a boron Plexiglas cover was used to
those for which no cover was used. This ratio is
plotted in Fig. 5.4.14 as a function of distance
from the reactor for angles of ¢ = 0, 15, 30, and
60 deg. The higher energy neutrons become more
important as the ratio decreases.

UNCLASSIFIED
2-01-056-22-677

100

 

50

3

=

3|2

|5 o

2|E

—

£|5 20

x|~

Y| #

als ¢ =15 deg
10 =30 deg

¢ =60 deg
5
0 30 60 90 120 150 180 210

r, DISTANCE FROM REACTOR CENTER TO DETECTOR (ft)

Fig. 5.4-]4.
Fluxes With and Without Boron Plexiglas on the Reactor

Ratio of the Effective Thermal-Neutron

Collimator as a Function of the Distance r for Various

Angles A

 
1. R. GNAffel

2. J. W, ANen

3. C, J. Bar¥gn

4, M. Bender

5. D. S. Billing

6. F. F. Blankens\ip
7. E. P. Blizard
8. C, J. Borkowski
9. W. F. Boudreau
10. G. E. Boyd

11. M. A, Bredig
12, E. J. Breeding
13. W. E. Browning
14, F. R. Bruce

15. A. D. Callihan
16. D. W, Cardwell
17. C. E. Center (K-25)
18. R. A. Charpie
19. R. L. Clark
20. C. E. Clifford
21, J. H. Coobs
22. W. B, Cottrell
23. R. S. Crouse
24, F. L, Culler
25. D. R. Cuneo
26. J. H. DeVan
27. L. M, Doney
28. D. A. Douglas
29. W. K. Eister
30. L. B. Emlet (K-2
31, D. E. Ferguson /
32. A, P. Fraas
33. J. H. Frye
34. W. T. Furgegfson
35. R, J. Gray
36. A. T. Grgfsky
37. W.R. Gfimes
38. A. G, frindell
39. E. Gyth
40. C. ¥ Harrill
41, MR, Hill
42, E. Hoffman
43, N. H. Hotffman
44, A, Hollaender

. A. S. Householder

o
k‘« «
¥ o
..

C-84 - Reactors-Special Features of Aircraft Reactors

INTERNAL DISTRIBUTION

54,

ORNL-2440

46. J. T. Hffwe

47. W. H. Sordan

48, G, W Keilholtz
49, C.P. Keim

50. Ef L. Keller

51. M. T. Kelley

S F. Kertesz

. J. J. Keyes

5, J. A, Lane
" 56. R. B. Lindauer
57. R. S. Livingston
58. R. N. Lyon

59. H. G. MacPherson
60. R. E. MacPherson
61. F. C. Maienschein
62, W. D. Manly
63. E. R. Mann

64, L. A. Mann
65. W. B. McDonald
66. J. R. McNally
67. F. R. McQuilkin
68. R. V. Meghreblian
69. R. P. Milford

70. A, J. Miller
71. R. E. Moore
72, J. G. Morgan

73. K. Z. Morgan

74, E. J. Murphy
75. J. P, Murray (Y-12)
76. M. L. Nelson
77. G, J. Nessle
78. L. G. Overholser
79. P. Patriarca
80. S. K. Penny

. A. M. Perry

82. D. Phillips

3. J. C. Pigg

84, P. M. Reyling
853 A. E. Richt

86. W T. Robinson
87, W. Savage

88. Savolainen
89. R. R. Schultheiss
90. D. S¥ott

91. J. L.%cott

 
298

92.

D. Shipley

93. AX\Simon

94.
95,
96.

97.

98.

99.
100.
101.
102,
103,
104,

O.

OOOAMEN>E -

Risman

. B. Trauger
. K. Trubey
. M. Watson

105,
106.
107.
108.
109,
110.
111.
112.

113-115,

116-122,

123
124-124

Wiitman ¥

igner (consultant)
C.Nilliams
J. CfWilson
C. F. VWinters
W.fZobel
ORNL -~ Y-12 Technical Library,
Jocument Reference Section
Laboratory Records Department
Laboratory Records, ORNL R.C.
Central Research Library

127-129.

130-131.
132,
133.
134.
135-137.
138,
139-140.
141,
142,
143-144,
145,
146.
147,
148-161.
162.
163165,
166.
167.
168.
169.
170.
171-176.
177.
178.
179-180.
181.
182.
183.
184.
185.
186.
187,
188.

EXTERNAL DISTRIBUTIPN

Air Force Qallistic Missile Dfvision

AFPR, BoeNgg, Seattle

AFPR, Boeiny, Wichita

AFPR, Curtiss\Wright, Clifton

AFPR, Douglas\Long Bghich

AFPR, Douglas, Yanta Monica

AFPR, Lockheed, Qurhfink

AFPR, Lockheed, Myrfetta

AFPR, North Americgh, Canoga Park

AFPR, North Amerigark Downey

Air Force Special Yeapdps Center

Air Materiel Comngtind

Air Research angfDevelopNent Command (RDGN)
Ajr Research anfl Developm{nt Command (RD TAPS)
Air Research gffid Developme}t Command (RDZPSP)
Air Technicalfintelligence Cekter

ANP ProjectOffice, Convair, Nprt Worth
Albuguerqugf Operations Office

Argonne Ngtional Laboratory

Armed Fogtes Special Weapons Prdject, Sandia
Armed Fdrces Special Weapons Projkct, Washington
Assistafft Secretary of the Air Force \R&D
AtomicfEnergy Commission, Washingtdyg

Atomifs International ‘

Battglle Memorial Institute

Betfls Plant (WAPD)

Bugeau of Aeronautics _
Byteau of Aeronautics General Representati

AR, Aerojet-General, Azusa |
BAR, Convair, San Diego

BAR, Glenn L. Martin, Baltimore

BAR, Grumann Aircraft, Bethpage

Bureau of Yards and Docks

Chicago Operations Office

 

 

 
189.
RO.

19

192-195.
196.
197.
198.
199.
200.
201.
202,
203.
204,
205.
206.
207.
208,
209.
210.
211,
212,
213.
214,
215.

216-219.
220,
221,
222,
223.
224,
225,
226,
227.

228-229.

230-247
248-2

Chicago Patent Group
Curtiss-Wright Corporation
Engineer Research and Development Laboratories
General Electric Company (ANPD)
Xeneral Nuclear Engineering Corporatign
rtford Area Office
lddgo Operations Office
Kno\s Atomic Power Laboratory
Lockhgnd Area Office
Los ANmos Scientific Laborator
Marquardt Aircraft Company
Martin Coygpany
National Alvisory Committeeffor Aeronautics, Cleveland
National AdWsory Committef for Aeronautics, Washington
Naval Air De%glopment Cegfter
Naval Air MateNal Center,
Naval Air Turbidg Test Jration
Naval Research [ yborafory
New York OperatioNs Jtfice
Nuclear Developmen\Corporation of America
Nuclear Metals, Inc
Office of Naval Refeargh
Office of the Chigf of Ndyal Operations (OP-361)
Patent Branch, Washingto
Pratt and Whitngy Aircraft Rivision
San Franciscgf Operations O¥ice
Sandia Corpgration
School of gfviation Medicine
Sylvania-forning Nuclear Corpokation
Technighl Research Group
USAF feadquarters
USAP Project RAND
U.§ Naval Radiological Defense LaBgratory
iversity of California Radiation L.aBgratory, Livermore’
Wright Air Development Center (WCOSI-§

. Technical Information Service Extension} Oak Ridge

73.

Division of Research and Development, AxC, ORO

 
 

Reports previously issued in this series are as follows:

ORNL-528
ORNL -629
ORNL -768
ORNL-858
ORNL-919
ANP-60
ANP-65
ORNL-1154
ORNL-1170
ORNL-1227
ORNL-1294
ORNL-1375
ORNL-1439
ORNL-1515
ORNL -1556
ORNL-1609
ORNL-1649
ORNL-1692
ORNL-1729
ORNL-1771
ORNL-1816
ORNL-1864
ORNL-1896
ORNL-1947
ORNL-2012
ORNL -2061
ORNL-2106
ORNL-2157
ORNL-2221
ORNL-2274
ORNL-2340
ORNL-2387

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending December 10, 1950
Period Ending March 10, 1951
Period Ending June 10, 1951
Period Ending September 10, 1951
Period Ending December 10, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952
Period Ending September 10, 1952
Period Ending December 10, 1952
Period Ending March 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953
Period Ending December 10, 1953
Period Ending March 10, 1954
Period Ending June 10, 1954
Period Ending September 10, 1954
Period Ending December 10, 1954
Period Ending March 10, 1955
Period Ending June 10, 1955
Period Ending September 10, 1955
Period Ending December 10, 1955
Period Ending March 10, 1956
Period Ending June 10, 1956
Period Ending September 10, 1956
Period Ending December 31, 1956
Period Ending March 31, 1957
Period Ending June 30, 1957
Period Ending September 30, 1957

!n -
h) v ¥