ORNL-5132.txt

ORNL-5132

3 4y45L 0445255 3 6/‘/5

Molten-Salt

Reactor ProgSram

Semiannual Cproqreoo %eport
jor Deriod Snding February 29, 1976

OOOOOOO E NATIONAL LABORATORY
CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION
LIBRARY LOAN COPY
DO NOT TRANSFER TO ANOTHER PERSON

one else to see this
document, sen d in name witl h document

and the library will arrange a loan

OAK RIDGE NATIONAL LABORATORY

OPERATED BY UNION CARBIDE CORPORATION FOR THE ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION

Printed in the United States of America. Available from
National Technical Information Service
U.S. Department of Commerce
5285 Port Royal Road, Springfield, Virginia 22161
Price: Printed Copy $7.75; Microfiche $2.25

This report was prepared as an account of work sponsored by the United States
Government. Neither the United States nor the Energy Research and Development
Administration/United States Nuclear Regulatory Commission, nor any of their
employees, nor any of their contractors, subcontractors, or their employees, makes
any warranty, express or implied, or assumes any legal li
the accuracy, completeness or usefulness of any information, apparatus, product or
process disclosed, or represents that its use would not infringe privately owned rights.

ORNL-5132
Dist. Category UC-76

Contract No. W-7405-eng-26

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT
FOR PERIOD ENDING FEBRUARY 29, 1976

L. E. McNeese
Program Director

AUGUST 1976

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee 37830
operated by '
UNION CARBIDE CORPORATION
- for the
ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION
MARTIN ENERGY RESEARCH LIBRARIES

LOCKHEED MAR |

3 4456 0445255 3 l

This report is one of a series of periodic reports that describe the progress of the program. Other reports issued in
this series are listed below. ‘

ORNL-2474
ORNL-2626
ORNL-2684
ORNL-2723
ORNL-2799
ORNL-2890
ORNL-2973
ORNL-3014
ORNL-3122
ORNL-3215
ORNL-3282
ORNL-3369
ORNL-3419
ORNL-3529
ORNL-3626
ORNL-3708
ORNL-3812
ORNL-3872
ORNL-3936
ORNL-4037
ORNL4119
ORNL4191
ORNL4254
ORNL-4344
ORNL-4396
ORNL-4449
ORNL 4548
ORNL4622
ORNL4676
ORNL4728
ORNL 4782
ORNL 4832
ORNL-5011
ORNL-5047
ORNL-5078

Period Ending January 31, 1958
Period Ending October 31, 1958
Period Ending January 31, 1959
Period Ending April 30, 1959
Period Ending July 31, 1959
Period Ending October 31, 1959

Periods Ending January 31 and April 30, 1960

Period Ending July 31, 1960
Period Ending February 28, 1961
Period Ending August 31, 1961
Period Ending February 28, 1962
Period Ending August 31, 1962
Period Ending January 31, 1963
Period Ending July 31, 1963
Period Ending January 31, 1964
Period Ending July 31, 1964
Period Ending February 28, 1965
Period Ending August 31, 1965
Period Ending February 28, 1966
Period Ending August 31, 1966
Period Ending February 28, 1967
Period Ending August 31, 1967
Period Ending February 29, 1968
Period Ending August 31, 1968
Period Ending February 28, 1969
Period Ending August 31, 1969
Period Ending February 28, 1970
Period Ending August 31, 1970
Period Ending February 28, 1971
Period Ending August 31, 1971
Period Ending February 29, 1972
Period Ending August 31, 1972
Period Ending August 31, 1974
Period Ending February 28, 1975
Period Ending August 31, 1975
Contents
INTRODUCTION . . oo e, e, e L vii

SUMM A R Y L e e e e e e ix

PART 1. MSBR DESIGN AND DEVELOPMENT

I. SYSTEMS AND AN ALY SIS L. i e e et e e et et et e e e 2
1.1 TRITIUM BEHAVIOR IN THE COOLANT-SALT TECHNOLOGY FACILITY ................. 2
1.1.1 Analysisof Experiments Tl and T2 .. ... ... ... . . . i i, 2
1.1.2 Analysisof Experiment T3 .............. e e e e 3
1.1.3 Analysisof Experiment T4 . . ... ... . i i it r e e 5
1.2 NEUTRONIC ANALYSIS . ...ttt [ 7
1.2.1 Cross Section Processing for MSBR Calculations ............... .. ... .. .. ........... 7
1.2.2 MSBR Performance Calculations .......... ... .. ... . ittt i 9
1.2.3 Helium Production in the MSBR Vessel ..................... e 9
1.2.4 Production of 232U in the MSBR ... ... ..\ttt et 10
1.2.5 Neutronic Analyses of TeGen Experiments .................. e 12
1.3 HIGH-TEMPERATURE DESIGN METHODS . .. ... ... . i, .. 12
1.3.1 Circular Cylindrical Shells . ........ .. .. ... [P 13
1.3.2 Nozzle-to-Spherical Shell Attachment ........ ... ... . i, 13
1.3.3 Nozzle-to-Cylinder Intersection .................. @ et e e e 13

1.3.4 Inelastic Analyses of MSR Piping Subjected to Internal Pressure and
Transient Temperature Cycles ............. ittt it inan P 13
2. SYSTEMS AND COMPONENTS DEVELOPMENT . ..... ... .ottt e 15
2.1 GAS-SYSTEMS TECHNOLOGY FACILITY ................... e e e e 15
2.2 COOLANT SALT TECHNOLOGY FACILITY ...t i, e 16
221 Loop Operation . ... ...ttt it e e e e e 16
222 Tritium Tests ... .o e e 16
2.2.3 Equipment Modifications .. ............... ... ... ..., e e 18
2.3 FORCED CONVECTIONLOOPS ........................ e 19
2.3.1 Operation of MSR-FCL-2b . . ... ... . e et 19
2.3.2 Heat Transfer Studies in MSR-FCL-2b . .. ... ... . i e e, 21
2.3.3 Design and Construction of FCL-3and FCL-4 .................. e 22

PART 2. CHEMISTRY

3. FUEL-SALT CHEMIST RY . . e e e e e e et e 24
3.1 Solubility of Lithium Tellurides in Fluoride Melts .. ......... ...ttt i, 24

3.2 Spectroscopy of Telluride Speciesin Molten Salts ......... ... ... ... ... . i, 27

iii
iv

3.3 Decomposition Pressure of LiTes . . ... vtn ittt ittt et e et e e e e e 28
3.4 Porous Electrode Studies in Molten Salts — Electrochemistry of Tellurium
in the LiCl-KCl Eutectic System . ... ... . e e et e 29
3.5 The Uranium Tetrafluoride-Hydrogen Equilibrium in Molten Fluoride Solutions ................ 29
. COOLANT-SALT CHEMISTRY . ... et 32
4.1 Hydrolytic Behavior of NasB3FgO3 ... .o . i e 32
4.2 Vapor Density Studies in the System BF3-Ho O . ... .o 34
DEVELOPMENT AND EVALUATION OF ANALYTICALMETHODS ..., 36
5.1 In-line Analysis Of MSBR FuUel:. . .. ... .. iuiet it we... 36
5.2 Tritium Addition Experiments in the Coolant-Salt Technology Facility ....................... 36
5.3 Electrochemical Studies of Tellurium in Molten LiF-BeF 2-ThF4 (72-16-12mole %) .............. 38
5.4 Electrochemical Studies of Oxygenated Species in Molten Fluorides .. .. .. e .39

PART 3. MATERIALS DEVELOPMENT

. DEVELOPMENT OF MODIFIED HASTELLOY N e e e e e e e e 42
6.1 Procurement and Fabrication of Experimental Alloys ............ ...t iirinininnnnnn.. 42
6.1.1 Production Heats of 2% Titanium-Modified Hastelloy N. . .............. PP 42

6.1.2 Semiproduction Heats of 2% Titanium-Modified Hastelloy N That Contain Niobium ....... 45

.6.2  Stability of Various Modified Hastelloy N Alloys in the Unirradiated Condition ................. 45
6.3 Mechanical Properties of Modified Hastelloy N Alloys in the Unirradiated Condition ............. 47
6.4 Postirradiation Creep Properties of Modified Hastelloy N ........................... DR 52
6.5 Microstructural Analysis of Modified Hastelloy N . ... .. [ e 59
6.5.1 Production of 2% Titanium-Hastelloy N Alloys with Uniform Carbide Distribution ........ 59

6.5.2 Carbon Behavior in Ni-Mo-Cr-Ti AllOyS .. .. ... vt e e e 70

6.6 Salt Corrosion StUBIES . ... ... ...\ttt 75
6.6.1 Thermal Convection Loop Results ................. e e e i e e, 76

6.6.2 Forced Circulation Loop Results .. ........... oo, 77

- 6.7 Corrosion of Hastelloy N and Other Alloys in Steam ................... P 80
6.8 Vapor Pressure Measurements of Metal Tellurides ................... e e 85
6.9 Operation of Metal-Tellurium-Salt Systems ............ ... .. ... .. .. . .. iuiiuun... e 86
6.9.1 ~Tellurium Experimental Pot No. 1 ... ... . ... . .. .. . . . . i 86

6.9.2 Chromium Telluride Solubility Tests .. ...ttt e, 87

6.9.3 Tellurium Screening Test ........ e e e e e e e e e e e e e 87

6.9.4 NizTe, Capsule Test . ... ... e e e 87

6.9.5 Chromium-Tellurium-Uranium Interaction Experiment .............................. 87

6.10 Examination of a Hastelloy N Foil Sample Embrittled in the Molten-Salt Reactor Experiment . . . . .. 88
6.10.1 Sample History . ... .. ... 88

6.10.2 Sample Fracture and Analysis Techniques .. ........... ... .o uimneennnnn. ... 88

6.10.3 Observations................. e e e e e e e e 89

_ 6.10.4 Discussion of Observations . .............. . it 93
6.11 High-Resolution Fractography of Hastelloy N Alloys Exposed to Tellurium .................... 95
6.12 Metallographic Examination of Samples Exposed to Tellurium-Cpntaining Environments .......... 100

6.13 Salt-Tellurium Creep StUAIes . ... ... .ovuounreeen e e e e e e e 121
6:14 Salt Preparatlon and Fuel Pin Flllmg for MSR Program Capsule Irradiation Experiment TeGen-2 . 127
6.14.1 Salt Preparation . ... ... i it e e 129
6.14.2 Hydrofluorination .. ......... .. ... it e e 130
6.14.3 UM/U Ratio AdJUSEIMENT . . ...ttt ettt e e e e e e e e e e e e 132
6.14.4 Preparation of Salt Fill Vessel ....... ... ...t 132
6.14.5 Salt Transfer . .. ... .. e 132

6.15 Salt Preparation and Filling of TeGen-3and-4................... et 135

6.16 Operation of TeGen-2and -3 . ... ... ... .. . . .o i it e e 136
6.16.1 Operating History of TeGen-2and -3 ... ... . . .. . . . . .. .. 139
6.16.2 Data Analysis for TeGen-2and -3 ..... e e e e e e e e 139
6.16.3 Preliminary Results of TeGen-2 ............... P 141
6.16.4 Future Irradiations . ... .. ... i i i i i et e 142

6.17 Examination of TeGen-2 ... .. .. . . 143

7. FUEL PROCESSING MATERIALS DEVELOPMENT . ...... ... . e, 163

7.1 Summary of Compatibility Studies . .......... .. ... .. . e 163

7.1.1 Ta—10% W Thermal Convection Loop Test . .................. ... vvn... e 163

7.1.2  Dissimilar Material Tests . . . . . ..o e e, 163

- PART 4. FUEL PROCESSING FOR MOLTEN-SALT REACTORS

8. CHEMISTRY OF FLUORINATION AND FUEL RECONSTITUTION . ............couuieeni... 170
9. ENGINEERING DEVELOPMENT OF PROCESSING OPERATIONS ... ....... ..o, 172
9.1 Metal Transfer Process Development ... ... .. .. .. . . . . it 172
9.1.1 Entrainment Studies in Experiment MTE-3B ... ....... .. .. .. .. ... . ... 172

9.1.2 Removal of LiCl and Bi-Li Phases and Addition of Purified Solutions ......... e 173

9.1.3 Experiments Nd-3andNd-4 . ... ... . . . .. .. 173

O L4 ReSUILS ..o e 174

9.2 Mass-Transfer Studies Using Water-Mercury Contactors . ...........c.ouvrininon e, 177
9.2.1 Experimental Equipment and Procedure ........... ... ... ... . .. ... ... 178

9.2.2 Experimental Results .......... ... . i i 179

9.3 Continuous Fluorinator Development ............ .. .. ... ... ........... e ... 184
9.3.1 Autoresistance Heating Test AHT-4 . . ... ... ... . .. 185

9.3.2 Frozen-Salt Corrosion Protection Demonstration (F SCPD) .. i 186

9.4 Fuel Reconstitution Engineering Development .............ciitiitiiiiiiiitinnnnnnnnn. 189
9.4.1 Hydrodynamic Operation . . ... ... ...ttt ittt ettt ettt eaenn. 191

9.4.2 Calibration of the UFg Metering System . ....... ...ttt it 192

9.4.3 Calibration of Ar-UF¢ Gas Density Detector ......... ... ... .. .. 0000t 192

9.4.4 Calibration of the HF-H, GasDensity Cell . . .. ... .. ... .. . . . . . ... 193

PART 5. SALT PRODUCTION
10. PRODUCTION OF FLUORIDE SALT MIXTURES FOR RESEARCH AND DEVELOPMENT ......... 197
Introduction

The objective of the Molten-Salt Reactor (MSR)
Program is the development of nuclear reactors that
use fluid fuels that are solutions of fissile and fertile
materials in suitable carrier salts. The program is an
outgrowth of the effort begun over 20 yearsago in the
Aircraft Nuclear Propulsion (ANP) Program to
make a molten-salt reactor power plant for aircraft.
A molten-salt reactor, the Aircraft Reactor Experi-
ment, was operated at Oak Ridge National Labora-
tory in 1954 as part of the ANP Program.

The major goal now is to achieve a thermal breeder
reactor that will produce power at low cost while
simultaneously conserving and extending the na-
tion’s fuel resources. Fuel for this type of reactor
would be ?*UF, dissolved in a mixture of LiF and
BeF,, but ***U or plutonium could be -used for
startup. The fertile material would be ThF4dissolved
in the same salt ot in a separate blanket salt of similar
composition. The technology being developed for the
breeder is also applicable to high-performance
converter reactors.

A major program activity through 1969 was the
operation of the Molten-Salt Reactor Experiment
(MSRE). This reactor was built to test the types of
fuels and materials that would be used in thermal
breeder and converter reactors; it also provided
operation and maintenance experience. The MSRE
operated at 650°C and produced 7.3 MW of heat.
The initial fuel contained 0.9 mole % UF., 5% ZrF.,
299% BeF,, and 65% 'LiF; the uranium was about 33%
23U, The fuel circulated through a reactor vessel and
an external pump and heat exchange system. Heat
produced in the reactor was transferred to a coolant
salt, and the coolant salt was pumped through a
radiator to dissipate the heat to the atmosphere. All
this equipment was constructed of Hastelloy N, a
nickel-molybednum-iron-chromium alloy. The reac-
tor contained an assembly of graphite moderator
bars in direct contact with the fuel.

Design of the MSRE was started in 1960,
fabrication of equipment began in 1962, and the
reactor became critical on June 1, 1965. Operation at

vii

low power began in January 1966, and sustained
power operation was begun in December 1966. One
run continued for six months, until stopped on
schedule in March 1968.

Completion of this six-month run ended the first
phase of MSRE operation, in which the objective was
to show, on a small scale, the attractive features and
technical feasibility of these systems for commercial
power reactors. The conclusion was that this
objective had been achieved and that the MSRE had
shown that molten-fluoride reactors can be operated
at 650° C without corrosive attack oneither the metal
or graphite parts of the system; also the fuel is
stable; the reactor equipment can operate satisfactor-
ily at these conditions; xenon can be removed rapidly
from molten salts; and when necessary, the radioac-
tive equipment can be repaired or replaced.

The second phase of MSRE operation began in
August 1968, when a small facility in the MSRE
building was used to remove the original uranium
charge from the fuel salt by treatment with gaseous
F.. In six days of fluorination, 221 kg of uranium was
removed from the molten salt and loaded into
absorbers filled with sodium fluoride pellets. The
decontamination and recovery of the uranium were
very good.

After the fuel was processed, a charge of 23U was
added to the original carrier salt, and in October 1968
the MSRE became the world’s first reactor to operate
on *¥U. The nuclear characteristics of the MSRE
with the ***U were close to the predictions, and the
reactor was quite stable. In September 1969, small
amounts of PuF; were added to the fuel to obtain
some experience with plutonium in a molten-salt
reactor. The MSRE was shut down permanently
December 12, 1969, so that the funds supporting its
operation could be used elsewhere in the research and
development program.

Because of limitations on the chemical-processing
methods available in the past, most of our work on
breeder reactors was aimed at two-fluid systems in
which graphite tubes would be used to separate
uranium-bearing fuel salts from thorium-bearing
fertile salts. However, in late 1967 a one-fluid breeder
became feasible with the development of processes
that use liquid bismuth to isolate protactinium and

remove rare earths from a salt that also contains,
thorium. Our studies showed that a one-fluid breeder-

based on these processes can have fuel-utilization
characteristics approaching those of our two-fluid
design concepts. Since the graphite serves only as
moderator, the one-fluid reactor is more nearly a
scale-up of the MSRE. These advantages caused a
change in the emphasis of the program from the two-
fluid to the one-fluid breeder; most of the design and
development effort is now directed to the one-fluid
system. -

In the congressiondl authorization report on the
AEC’s programs for FY 1973, the Joint Committee
on Atomic Energy recommended that the molten-salt
reactor be appraised so that a decision could be made
about its continuation and the level of funding
approprate for it. Consequently, a thorough review
of molten-salt technology was undertaken to provide

information for an appraisal. A significant result of .

. the review was the preparation of ORNL-4812, The
Development Status of Molten-Salt Breeder Reac-
tors. A subsequent decision was made by the AEC to
terminate work on molten-salt reactors for budgetary
reasons; in January 1973 ORNL was directed to
conclude MSR development work. }

In January 1974, the AEC program for molten-salt
reactor development was reinstated. A considerable
effort during 1974 was concerned with assembling a
program staff, making operational a number of
development facilities used previously, and replacing

viii

a number of key developmental facilities that had
been reassigned to other reactor programs. A
significant undertaking was the formulation of
detailed plans for the development of molten-salt
breeder reactors and the preparation of ORNL-5018,
Program Plan for Development of Molten-Salt
Breeder Reactors. :
During 1974 and 1975, work in the Molten-Salt

- Reactor Program was devoted to the technology

needed for molten-salt reactors. The work included
conceptual design studies and work on materials, the
chemistry of fuel and coolant salts, fission-product
behavior, processing methods, and the development
of systems and components. The most important
single aspect of the program was work on the
development and demonstration of an alloy that is
suitable for the primary circuit of an MSBR and has
adequate resistance to tellurium-induced shallow
intergranular cracking, which was first observed in
MSRE surveillance specimens. A second important
area consisted of studies of the chemical interaction
of tritium with the MSBR secondary coolant, both in
laboratory chemistry studies and in a large engineer-
ing facility (Coolant-Salt Technology Facility).
These studies culminated in the demonstration of an
adequate basis for management of tritium in a 1000-
Mw(e) MSBR.

In February 1976, ORNL was directed by ERDA
to again terminate the Moltén-Salt Reactor Program
for budgetary reasons. Work during the remainder of
FY 1976 was directed toward completion of short-
term work in the Program, reporting of associated
information, and the assigniment of the MSRP staff

and experimental facilities to other ORNL programs.
Summary

PART 1. MSBR DESIGN AND DEVELOPMENT
J. R. Engel

1. Systems and Analysis

The investigation of tritium behavior in the MSBR
reference coolant salt, NaBF.-NaF eutectic, is
continuing in the Coolant-Salt Technology Facility
(CSTF). Results have been evaluated for three short-
term tritium-addition tests, and preliminary results
have been obtained for a steady-state test that is still
in progress.

The buildup rate of tritium in the salt during the
short-term (10-hr) additions indicated that 50—60%
of the added tritium was being trapped in the saltina
chemically combined form. However, subsequent
evaluations of the total tritium flow through the
CSTF off-gas system indicated that additional
tritium was being accumulated somewhere in the
system and released to the salt after the additions
were stopped. The extra material may have been
* temporarily accumulated in the loop walls as
elemental tritium. The concentrations of elemental
tritium in the salt immediately after each of the short-
term additions were estimated from the concentra-
tions of elemental tritium in the off-gas samples and
an assumed value for stripping efficiency in the pump
bowl. The results, in conjunction with the measured
concentrations of combined tritium in the salt,
indicated that the ratio of combined-to-elemental
tritium in the salt was about 300 to 500.

" Before the fourth (steady-state) tritium addition
test, equipment was added to the CSTF to permit
direct measurement of the partial pressure of
hydrogen in the salt and to provide data on the rate of
hydrogen permeation through the loop walls. These
additions were designed to provide alldata necessary
for an overall hydrogen (and tritium) material
balance. '

The attainment of steady-state conditions required
considerably more time than had been projected
from the short-term tests. (Accumulation of elemen-

ix

tal hydrogen in the loop walls is thought to be the
rate-limiting process.) After about four weeks of
tritium addition, the sample results indicated that 96
to 99% of the tritium leaving the CSTF was being
removed in the loop off-gas, and only'] to 4% was
permeating through the loop walls. The overall
material balance on tritium flow rate appeared to be
between 0.89 and 0.97. Measurements of the partial
pressure of hydrogen in the salt (0.43 Pa) were
reasonably consistent with the apparent partial
pressure of elemental hydrogen in the pump-bowl gas.
space, as indicated by the observed concentration of
elemental tritium in off-gas samples. Use of the
measured partial pressure with the measured tritium
concentration in the salt gives a value of 4700 for the
ratio of combined-to-elemental tritium in the salt. If
this concentration ratio were attained in the
reference-design MSBR, only ~3.5 Ci/d of tritium
would be released to the steam system and, hence, to
the environment.

The development of a 123-energy-group neutron
cross-section library, based on ENDF/ B, Version 1V,
was completed. The library includes cross sections at
four temperatures from 300 K to 1200 K for several
fuel-moderator ratios that are appropriate to the
reference-design MSBR. Further processing to
produce few-group cross sections for multidimen-
sional diffusion-theory calculations of the reactor
performance were delayed by difficulties with the
televant computer codes. However, estimates were
obtained for the flux spectrum and the flux density in
both- the reactor and the reactor vessel of the
reference-design MSBR. While the work required to
complete’ the cross-section processing and the
performance calculations has been identified, it will
not ‘be done because the MSR program will be
terminated at the end of FY 1976.

The estimated helium concentration in the Hastel-
loy N pressure vessel of an MSBR after 30 full-power
years, due principally to two-step *°Ni(n,v) *’Ni(n,a)
reactions, is 38 ppm at the inner surface. The con-
centration. would decrease by about 500% for
each centimeter of penetration into the vessel wall,
which is expected to have a thickness of about 5 cm.

The calculated steady-state concentration of ***U
in the uranium inan MSBR is 20 ppm if protactinium
is removed from the salt on a 10-day cycle and if no
»*Th is present. (About 15 years would be required to
approach equilibrium.) Approximately 27% of the
232U results from (n,2n) and (7,n) reactions in **’U,
and the remainder is due to similar reactions in ***Th.,
A concentration of 23 ppm “*Th in thorium would
double the steady-state concentration of ***U.

Neutron-flux monitors were provided for two
additional in-pile irradiation capsules, TeGen-2 and
TeGen-3. The irradiated wires were recovered from
TeGen-2 and are being analyzed, and TeGen-3 is still
being irradiated.

The development of simplified high-temperature
structural design methods, which could be applicable

to MSBR design, is continuing, principally for other

reactor programs. Some cases have been identified in
which the simplified, elastic-analysis rules are
' nonconservative in comparison with more accurate
(and more expensive) inelastic analyses. Potential
sources of the lack of conservatism are being
investigated. Some of the data (thermal transients
and material properties) have been obtained to
estimate the magnitude of thermal ratchetting and
creep-fatigue damage that might occur in an MSBR.
These estimates will indicate the applicability of the
simplified design procedures to Hastelloy N.

2. Systems and Components Development

A larger diameter impeller was installed in the salt
pump of the Gas-Systems Technology Facility, and
additional water tests were made. The amplitude of
the pump shaft oscillations was acceptable; however,
the hydraulic imbalance between the volute and
impeller, caused by operation far from the pump
design point, resulted in unacceptable deflections of
the shaft. Extensive pump modifications would be
required to meet the loop-design requirements. It
appears that the back vanes installed on the larger
impeller overcompensated for the shaft seal leakage
and reversed the direction of the flow. Because of the
termination of the MSR Program at the end of FY
1976, the loop has been drained and put in standby
condition.

Another transient -tritium addition test was

completed on the Coolant-Salt Technology Facility,
and a long-term steady-state test is in progress.
Several changes were made in the sampling systems

for the loop off-gas and for the enclosure ventilation

air to improve the reliability of the data and to permit
the collection of data that will allow the evaluation of
tritium material balances for the system. Operation
of the facility has been highly reliable.
The forced-convection loop, MSR-FCL-2b, has
accumulated 4000 hr of operation with MSBR
reference fuel salt at design conditions (566°C
minimum, 704°C maximum; 2.5 and 5.0 m/s salt
velocities), with the expected low corrosion rates in
standard Hastelloy N. Additions of NiF; were made
in preparation for determining the corrosion rates for
this alloy at higher U*"/U"" ratios. Increases in pump
power, which previously had accompanied increases
in the oxidation potential of the salt, were again
observed at each addition. These tests were inter-
rupted by two successive loop piping ruptures that
resulted from salt freezing in the loop during off-
normal conditions. Modifications were made In
conjunction with the loop repairs to reduce the
possibility of future similar failures. Because of the
decision to terminate the MSR Program at the end of
FY 1976, the corrosion studies with standard
Hastelloy N were discontinued; tests are now in
progress to examine the corrosion of Hastelloy N
modified by the addition of 1% niobium.
Construction of two additional forced-convection
loops, MSR-FCL-3 and 4, is being stopped because
of cancellation of the MSR Program.

PART 2. CHEMISTRY
3. Fuel-Salt Chemistry

Measurements were made of the solubility of Li,Te
in molten Li2BeF, over the range from 500 to 700° C.
Li;Te was prepared using a mixture of stable
tellurium and '**"Te tracer. Salt that had been
equilibrated with the solid Li;Te was sampled, and
the tracer was counted using a lithium-drifted
germanium detector. The most recent data indicatea
solubility below 10~° mole fraction at 700°C.

Spectrophotometric studies of lithium telluridesin
molten LiF-BeF; mixtures were continued. This
work has progressed to the point where three
tentative conclusions can be made: (1) Li;Te is quite
insoluble in LiF-BeF; melts; (2) the light-absorbing
species in LiF-BeF, melts apparently can be
represented as Te,"; (3) although the results are not
conclusive, it is reasonable to expect the Te; ion to
exist in molten-salt solutions.

In other spectral studies, the pressures of tellurium
due to the equilibrium 2LiTe; = Li;Te + °/» Tex(g)
were determined over the range from 500 to 750°C.
Previously determined vapor pressures for pure
tellurium were used to calibrate the system. The
measured equilibrium pressures of Te; over LiTe;
were only about a factor of 1.5 lower than the vapor
pressures of Te. and can be represented by the
equation In P(mm Hg) = 15938 - 12,720/ T(K).
Evidence for a slightly volatile lithium telluride was
found at temperatures around 1000°C.

Electrochemical studies were initiated on the
lithium tellurides Li,Te and LiTe; in LiCl-KCl
eutectic at 400°C. Both of these tellurides were
insoluble in the melt at this temperature. Cathodiza-
tion of a tellurium electrode did show that a one-
electron process was occurring. 1t was also observed
that the presence of moisture in these systems led to
the formation of colored soluble species.

Studies of the equilibrium UF.(d) + /> Ha(g) =
UF;(d) + HF(g) in molten LiF-BeF, mixtures were
continued using a spectroscopic method. Initially,
the molar extinction coefficient for UF; was
determined in LiF-BeF; (66-34 and 48-52 mole %).
Values of the quotient for the above equilibrium were
then determined in the two solvents over the range
from 500 to 800° C. The quotients in the 66-34 mole %
solvent were essentially the same as those measured
previously by Long and Blankenship; however, those
in the 48-52 mole % solvent were about 10 times
greater than those determined by Long and Blanken-
ship. The change in the present values with change in
solvent composition is consistent with expectations
based on prior studies of similar equilibria.

4. Coolant-Salt Chemistry

Raman and "’F NMR spectroscopy were used to
study the hydrolytic behavior of Na3B;F¢Os since this
compound appears to be the stable oxygen-
containing species in NaF-NaBF, (8-92 mole %)
when the total oxygen concentration 1s low. The
results showed that Na;B;F«O; and NaBF;OH
reversibly interconvert in the presence of water.
Assuming that this is also true in molten NaF-NaBF,,
a possible mechanism for trapping of tritium in the
melt is postulated.

Studies of the vapor density in the system BF;-H,O
were continued. At temperatures above 200°C,
BF;-2H,0 is completely dissociated. Below 200°C,
association in the vapor phase becomes pronounced,
and, with sufficient BF;:2H0 in the system, a stable
liquid phase is formed. The vapor pressure of the
liquid reaches 1 atm at about 200°C. Attempts to
determine equilibrium constants for the vapor-phase
reactions are in progress.

Xi

5. Development and Evaluation
of Analytical Methods

The monitoring of U*'/ U’ ratios, which reflect the
oxidation potential of the fuel salt, was continued
during this period for one forced-convection loop,
four thermal-convection loops, and eight creep-test
machines. Forced-convection loop FCL-2b, after a
shutdown period, was recharged with new salt and is
back in operation. The U*"/U"" ratio at startup was
about 5.3 X 107. Thermal-convection loops NCL-
21A and -23 continue to operate at stabilized redox
conditions. Thermal convection loops 18C and 24
have shown a gradual decline in the U*"/U" ratio,
which is presently about 1.7 X 10° and 80 respectively.
The U*"/ U™ ratios for the eight creep-test machines
are presented in tabular form. Generally, the melts
have tended to become more reducing with time.

The results from the third tritium injection
experiment at the Coolant-Salt Technology Facility
are similar to those from the first two experiments.
Most of the tritium occurs in a water-soluble or
combined form. Very little tritium in the off-gas was
in the elemental form. A fourth tritium injection
experiment is now under way.

Voltammetric measurements were made in
molten LiF-BeF.-ThF, following additions of LiTe;
and Cr;Tes compounds in an effort to identify soluble
electroactive tellurium species. No voltammetric

evidence of such compounds was obtained. Electro-

chemical studies were carried out on the tellurium
species generated in situ in molten LiF-BeF»-ThF,
when a tellurium electrode is cathodized. The results
indicated that the species generated is of the type
Ten (M = 1) and appears to be unstable under the
existing experimental conditions.

Voltammetric studies were initiated on two
anodic waves that are observed at a gold electrode in
molten LiF-BeF,-ThF, and also in melten LiF-BeF,-
ZrF,. Although the results are tentative, it is believed
that these waves are associated with oxygenated
species in the melts. The first wave possibly conforms
to the oxide — peroxide electrode reaction, and the
second wave represents the continued oxidation of
peroxide species ultimately to oxygen gas. Noise on
the diffusion current plateau indicates gas-bubble
formation at the electrode surface.

PART 3. MATERIALS DEVELOPMENT

6. Dévelopment of Modified Hastelloy N

Tubing of 2% titanium-modified Hastelloy N was
produced in a pilot run using the fabrication schedule
for austenitic. stainless steels. One commercial 2500~
1b melt of Hastelloy N modified with 2% titanium and
1% niobium was fabricated successfully into several
bar configurations. Eight small commercial alloys
containing 2% titanium and various amounts of
niobium were melted and fabricated into '/>-in.-thick
plate. Laboratory alloys containing up to 4%
niobium were prepared and converted to '/s-in.-diam
rod for evaluation.

Various types of tests were run in which specimens
were exposed to tellurium-containing environments.
The source of tellurium that is most representative of
tellurium in an MSBR appears to be a mixture of
CrTe* plus CrsTet. Examination of specimens
exposed in these various screening tests indicated
that alloys containing from 0.5 to 2% niobium are
most resistant to intergranular cracking by tellurium.
Mechanical property tests showed that these alloys
have slightly lower creep strength than 2% titanium-
modified Hastelloy N, but higher strength than
standard Hastelloy N. Postirradiation creep tests
showed that the niobium-modified alloys have
excellent properties after irradiation at 650°C,
acceptable properties after irradiation at 704°C, and
poor properties after irradiation at 760°C.

A fueled capsule containing pins of 2% titanium-
modified Hastelloy N, 2% titanium plus rare-earth-
element-modified Hastelloy N, and Inconel 600
revealed that all three materials were embrittled
intergranularly by exposure to tue fission-product
containing salt. Two other fuel capsules (six
materials) are in various stages of assembly and
irradiation.

7. Fuel Processing Materials Development

A thermal-convection loop constructed of Ta—10%
W is being operated to evaluate the compatibility of
this alloy with fuel salt. Several graphite capsules
containing various bismuth-lithium solutions and
either molybdenum or Ta—-109% W specimens were
heated for 1000 hr at 600 or 700°C. All capsules
demonstrated excellent compatibility although some
important differences were noted between the
various capsules. '

8. Chemistry of Fluorination and
Fuel Reconstitution

Studies of the chemistry of fuel reconstitution were
resumed. A test of the effectiveness of smooth
platinum for catalyzing the hydrogen reduction of
U** to U™ in small gold equipment has shown that

xi1

smooth platinum sheet of limited surface area would
provide appreciable catalytic activity in the hydrogen
reduction column of the Fuel Reconstitution
Engineering Experiment. Niobium is an important
fission product with volatile fluorides and would be
carried from the fluorinator to the fuel reconstitution
step. Studies of .the hydrogen reduction of NbF,
showed that in the absence of granular platinum, the
NbF, was reduced slowly to Nb°. In the presence of
granular platinum, the rate of NbF, was rapid for the
first 2 hr and decreased to a ‘value similar to that
experienced in the uncatalyzed reaction. The reason
for this behavior is being sought, since, if it is due to
poisoning of the platinum, it has significant implica-
tions for the use of platinum catalysts in a reactor
processing plant.

9. Engineering Development
of Processing Operations

Two additional runs were made in Metal Transfer
Experiment MTE-3B. These runs were made using
agitator speeds of 4.17 and 1.67 rps to determine the
effect of agitation on the transfer rate of neodymium
from the fluoride fuel salt to the bismuth-lithium
stripper solution. Prior to these runs, it was
determined that the previously observed entrainment
of the fluoride salt into the LiCl resulted from
operation of the agitators at 5 rps. This was
unexpected since no entrainment was seen in
experiment MTE-3 under similar conditions. Tests
showed that no entrainment occurred at agitator
speeds up to 4.58 rps. Before the two additional (and
final) runs were made, the LiCl and bismuth-lithium
solutions, contaminated with fluoride salt, were
removed from the process vessels and replaced with
fresh LiCl and bismuth-lithium. Results of the two
runs show that the rate of transfer of neodymium was
increased by 300 to 400% when the agitator speed
increased from 1.67 to 4.17 rps. However, overall
mass-transfer coefficients for neodymium were lower
than predicted by literature correlations, particularly
at the LiCl-bismuth interfaces.

Students from the MIT School of Chemical
Engineering Practice have completed measuring
water-side mass-transfer coefficients in three stirred,
nondispersing, water-mercury contactors. A wide
range of agitator diameters and speeds was covered in
these measurements. These measurements have
provided a great deal of data covering a wide range of

physical parameters which will be useful in develop-

ing correlations to be used for estimating mass-
transfer rates in large-scale nondispersing stirred
contactors required in the MSBR reductive extrac-
tion processes.

A fifth run was made with autoresistance heating
test AHT-4 using a different cooling procedure. This
run demonstrated that the main problem is the
plugging in the unheated end of the salt inlet tube
(electrode). A new electrode has been designed to
alleviate this problem. Eight cooling tests were made
with the Frozen Salt Corrosion Protection Demon-
stration equipment prior to the introduction of
fluorine. The purpose was to define the conditions
under which a satisfactory frozen salt film could be
formed. The fluorine inlet (inner) tube plugged
before the outer wall of the tube was cold enough to
form a satisfactory film. During the sixth test, air
oxidation resulted in a leak in the cooled tube. A
second smaller tube was fabricated with a separate
fluorine inlet tube, but a satisfactory film was not
formed in the first two tests using argon coolant.

During this report period a preliminary hydrody-
namic test of the experimental equipment for the fuel
reconstitution engineering experiment (FREE) was
successfully completed in which salt flow through the
system was maintained under simulated experimen-
tal conditions. A calibration of the UFs metering

xiii

system was completed; a gas density cell used for
measuring concentrations of UFs in argon was
calibrated; and apparatus for producing known
concentrations of HF in hydrogen was developed and
was used to calibrate the gas density cell for
measuring concentrations of HF in hydrogen.

PART 5. SALT PRODUCTION

10. Production of Fluoride Salt Mixtures
for Research and Development

Three 150-kg batches of fuel-carrier salt were
produced in a new copper-lined treatment vessel and
vessel head. The first two of these batches were of
significantly improved purity because the copper
linings reduced vessel corrosion products.

A total of 1975 kg of salts (of various composi-
tions) were produced since activation of the facility in
1974. Of this, 678 kg are stored for possible future
use.

Since the program has now been ended, all
production areas are decommissioned and decon-
taminated. All materials and equipment are appro-
priately disposed of.
Part 1. MSBR Design and Development

J. R. Engel

The overall objective of MSBR désign and

development activities is to evolve a conceptual -

design for an MSBR with adequately demonstrated
performance, safety, and economic characteristics
that will make it attractive for commercial power
generation and to develop the associated reactor and
safety technology required for the detailed design,
construction, and operation of such a system. Since it
is Jjkely that commercial systems will be preceded by
one or more intermediate-scale test and demonstra-
tion reactors, these activities include the conceptual
design and technology development associated with
the intermediate systems. '

Although no system design work 1s currently in
progress, the ORNL reference conceptual design' is
being used as a basis to further evaluate the technical
characteristics and performance of large molten-salt
systems. A major effort in this regard is the
evaluation of tritium behavior in the Coolant-Salt
Technology Facility and the extrapolation of those
results to the MSBR. Early results indicate that the
reference coolant salt, NaBF,-NaF eutectic, will
permit limitation of the tritium release to a few curies
per day.

Additional core neutronics calculations are being
made for the reference MSBR, using nuclear data
from ENDF/B, Version 1V. These calculations will

provide updated estimates of the nuclear perfor- -

mance, as well as additional information on core
characteristics. Analogous methods and data are
employed to provide support for inpile irradiation
work.
" Analytic studies are in progress to assess the
potential significance of thermal ratchetting and
creep-fatigue considerations in the design of Hastel-
loy N components for operation at high tempera-
tures.

The Gas-Systems Technology Facility is an
engineering-scale loop built for use in the develop-
ment of gas-injection and gas-stripping technology
for molten-salt systems and for the study of xenon
and tritium behavior and heat transfer in MSBR fuel
salt. The facility was operated briefly with water to
measure loop and pump characteristics that would be
required for the performance and analysis of
developmental tests with fuel salt. This facility is now

" shut down and in a standby condition.

The Coolant-Salt Technology Facility is bemg
operated routinely to study processes involving the
MSBR reference-design coolant salt, NaBF;-NaF
eutectic. Tests are currently in progress to evaluate
the distribution and behavior of tritium in this
system.

Candidate MSBR structural materials are exposed
to fuel salt at reference design temperatures and
temperature differences (704°C maximum and
139°C AT) and at representative salt velocities in
forced-convection loops to evaluate corrosion effects
under various chemical conditions. These opera-
tions, which are principally in support of the
materials development effort, also provide experi-
ence in the operation of molten-salt systems and data
on the physical and chemical characteristics of the
salt. At present one loop, MSR-FCL-2b, which 1s
made of standard Hastelloy N, is in routine operation
as part of a study of the corrosion of Hastelloy N
containing 1% niobium. Construction of a major
fraction of two other loops was completed before
work on them was discontinued because of the
decision to terminate the MSR Programat the end of
FY 1976.
1. Systems and Analysis

J. R. Engel

1.1 TRITIUM BEHAVIOR
IN THE COOLANT-SALT
TECHNOLOGY FACILITY

J. R. Engel G. T. Mays

Experiments to elucidate the behavior of hydrogen
(tritium) in MSBR coolant salt are continuing in the
Coolant-Salt Technology Facility (CSTF). To date,
three relatively short transient tests have been
performed, and a longer steady-state test is currently
in progress. In all these tests, a substantial fraction of
the tritiated hydrogen that was added to the salt has
been retained in the salt in a chemically combined
form that was slowly removed from the system by
transfer to the loop cover gas. This retention
significantly reduced the amount of elemental
hydrogen that was available for transport through
the loop walls. '

Preliminary results of the first two transient tests
were presented in the previous semiannual report.”
Additional results for these tests and for the third
transient test are presented in the next two sections
(1.1.1-1.1.2). Preliminary results are also presented
for the steady-state test that is in progress. Extrapola-
tion of these results to the reference-design 1000-
MW(e) MSBR indicates that tritium migration to the
steam system (and hence to the environment) would
be less than 10 Ci/day (370 GBq/d).

1.1.1 Analysis of Experiments
T1 and T2

In experiment T1 it was previously reported’ that
the apparent half-life for removal of water-soluble
(combined) tritium from the salt was 9.2 hr. This
value, when incorporated into the buildup of
combined tritium in the salt during the addition
phase of the experiment, implies a trapping efficiency
of 85%. After the results from experiments T2and T3
were analyzed and after the apparent trapping
efficiencies of 50% and 55%, respectively, were
calculated, the trapping efficiency of experiment T1
was reexamined. Depending on the inventory at the
end of the addition period and on how the data for
the subsequent salt samples are extrapolated back to
the end of the addition period, a half-life of 12 hr for
removal of tritium from the salt could be determined.
This would lead to a trapping efficiency of 60%,

which is more nearly consistent with the trapping
efficiencies for experiments T2 and T3.

In test T2, because of a greater number of sample
measurements and less scatter in the salt data during
the removal portion of the experiment immediately
after the addition period, a more reliable estimate of
the half-life for removal of tritium from the salt could
be made. For experiment T2, a 12-hr half-life also
was determined. Even though the data from the
second experiment suggested the presence of one or
more other mechanisms for removal of tritium from
the salt with significantly longer time constants, the
extraction of these time constants was not attempted
because of the scatter in the data at longer times.
Consequently, a single first~-order process with a 12-
hr half-life was used to describe both the removal of
tritium from the salt and the initial buildup. If the
longer time constant or constants could be extracted,
the half-life for the short-term process would be less
than 12 hr. However, neglecting the longer time
constant and using only the 12-hr half-life provide a
reasonable basis for calculating the expected rate of
tritium buildup in the salt. The comparison of
calculated and observed inventories (Fig. 1.1) as
functions of time during the addition period illustrate
reasonable agreement. Improved agreement would
be expected if additional time constants were
factored into the calculation. '

Based on the off-gas flow rate and the measured
concentrations of water-soluble tritium in the off-gas
during experiment T2, approximately 65 mCi (2.4
GBq) of water-soluble tritium had flowed through
the off-gas system during the first 240 hr after the
start of the experiment. However, only about 50 mCi
(1.8 GBq) could be accounted for in terms of water-
soluble trittum released from the salt during the
addition and the inventory in the salt at the end of the
addition. Part of this discrepancy can be explained by
a few apparently high off-gas sample results that were
used in evaluating the integrated tritium off-gas flow.
However, the magnitude of the discrepancy and the
long time constant for tritium removal from the salt
suggest that the water-soluble form of tritium
continued to appear in the salt (and to be stripped out
in the off-gas) after the deliberate tritium addition
was stopped. It is suggested that some of the added
tritium temporarily accumulated in a reservoir other
than the salt (possibly as elemental tritium in the pipe
ORNL-DWG 76—52549

35 ‘ }

20 CALCULATED |° ' //
= ® OBSERVED - 7
£ - /

5 25 A
2 . p
Zz /
= 20 -
E [ 3
r
rd

o
& /
E 10 9
z /
J
>
z /

5 17

0

0 ! 2 3 a4 5 - 6 7 8 9 10

TIME AFTER START OF ADDITION (hr)

Fig. 1.1. Buildup of tritium in salt of CSTF during the addition period of T2.

walls) and then transferred back into the salt when
the salt concentration had decayed to a sufficiently
low level.

1.1.2 Analysis of Experiment T3

In experiment T3 approximately 49 mCi (1.8 GBq)
of tritium were added in 30.5 cm’ of gas during the
9.87-hr addition period. Immediately prior to the
start of the third test, single samples of the salt and
off-gas indicated the following tritium concentra-
tions:

In salt 1.6 nCi/g (59 kBq/kg)
In off-gas, elemental 0.42 pCi/em®  (15.5 kBq/m?)
In off-gas, combined 38 pCi/cm? (1.4 MBq/m3)

These concentrations are comparable to those
obtained at the end of both the first and second
experiments. The sample results for experiment T3
(Fig. 1.2) are presented without corrections for any
baseline concentrations. '

The tritium concentration in the salt peaked at 40
nCi/g (1.5 MBq/ kg) at the end of the addition period
and decreased in a manner similar to that observed in
experiments T1 and T2. If two first-order processes
for removal of tritium from the salt are assumed, as
suggested by the data, the apparent half-lives for the
processes are 8.7 and ~108 hr. These processes lead

to an apparent trapping efficiency of 55% for the salt. .
Recalling that the apparent efficiencies in experi-

ments T1 and T2 were 60% and 50%, respectively, t_he

trapping efficiencies for the three experiments are
reasonably consistent.

Using the 8.7 and 108-hr values for the half-lives, a
comparison was made between the calculated and
observed inventories as functions of time during the
tritium addition (Fig. 1.3). The two processes
associated with these half-lives provide a reasonable
description of both the buildup and decay of tritium
concentration in the salt.

Elemental tritium in the off-gas reached a
maximum concentration of 40 pCi/em’ (1.5
MBq/m"), which was significantly higher in relation
to the tritium concentration in the salt than in test T2.
The concentration of elemental tritium in the salt,
inferred from the elemental tritium concentration in
the off-gas, was 0.1 nCi/ g (3.9 kBq/kg), which led to
a ratio of 380 for the combined-to-elemental
hydrogen in the salt.

The concentration of water-soluble tritium in the
loop off-gas at the end of the addition period was
10,200 pCi/cm’ (~380 MBg/ m’). This result wasalso
significantly higher than the tritium concentration in
the salt at the comparable time during experiment
T2. Two first-order time constants were extracted
from the time variation of the concentration of water-
soluble tritium in the off-gas. The half-life of the
more rapid process ranged from 7 to 15 hr, while the
half-life for the longer process was about 70 hr.

About 150 hr after the start of the third addition,
37 mCi (1.3 GBq) of tritium in the water-soluble form
had passed through the off-gas system while only 25
mCi could be accounted for from salt-sample results.
Again, the indication is that of a tritium inventory
that is not revealed by salt samples.

Measurements of the tritium concentration in the
exhaust air from the loop enclosure of the CSTF were
made shortly after the end of the addition period in

an attempt to determine the rate of tritium permea-
tion through the loop walls. These measurements
were invalidated by simultaneous operations to
dispose of excess tritium from the addition station by
venting to the loop enclosure at the end of the third
addition. When the loop sample results had reached
previous baseline values, the loop sampling fre-

ORNL-OWG 76-5256

104 | =
=LY o SALT (nCi/g) 3
af e OFF-GAS, WATER SOLUBLE -
ri *, _ (pCi/cm3) . ]

~ N & OFF-GAS, ELEMENTAL (pCi/em’)
E 0% EN 5
S Ef\ o % =
O B\ 3
S RN “ e .
S —% _
o BN L] —
'3 S “le %!
£ 10% N TRITIUM ADDITION ~
:\,/ i 3
z =N ®le =
o HN 3
= i o . * . —
s fi .
[ = —
S ibl®
=z 100 B * —
H a =
8 B\« 3
g R L. =
= —§ d cfi 9o _
= BN Qo] 9 o °
x g% Y N
[ ol 0 S aa ol ° % o

{0 E% o B ] r e L Py =
2\ ° 2| ° e
_4§ & P’ &
_§ p—
5N )

\ _
oot LN
16 8 20 22. 24 26 28 30 2 4 6 8
SEPTEMBER : OCTOBER
DATE, 1975

Fig. 1.2. Observed tritium concentrations in CSTF, Test 3.

ORNL—DWG 76—5255

S 25
£
'—
g‘ 20 -
—— CALCULATED L

=
£ ® OBSERVED —]
2 45
- —s
: —
‘5 10 "/
> /
o
'C_) 5 [ ]
P4
w
>
=z

0

0 g 2 3 4 5 6 7 8 9 10

TIME AFTER START OF ADDITION (hr)

Fig. 1.3. Buildup of tritium in salt of CSTF during the addition period of T3.
quency was reduced and several tests were made to
develop a method for determining the rate of
permeation of tritium through the loop walls.
Tritium was added to the loop enclosure and to one
of the cooling shrouds on the loop piping while
samples of the ventilation air were analyzed for
tritium content. These tests led to the use of an air-
sample tap in the discharge line from one of the
coolers and to modification of the cooler inlets to
supply fresh air (instead of air from within the
enclosure) to that cooler. The results indicated that
the tritium concentration in the air leaving the cooler
would be readily and accurately measurable if
anticipated loop-wall permeation rates were experi-
enced.

Immediately after the end of the tritium addition
phase of test T3, some preliminary data were
obtained to determine if the addition tube (or a
similar device immersed in the main salt stream)
could be used to measure the partial pressure of
elemental hydrogen in the circulating salt. The
addition tube was evacuated several times and then
“isolated”; it was expected that any pressure rise
within the isolated tube would be due to hydrogen
partial pressure in the salt, but the pressure did not
reach a steady value. It was subsequently determined
that stray inleakage (presumably of air) was
occurring and that the rate of inleakage was not
sufficiently consistent to permit the extraction of
meaningful data about the hydrogen partial pressure.
However, it appeared that a sufficiently leaktight
device could be used to measure hydrogen partial
pressure. Therefore, a partial-pressure probe was
constructed and installed in.the CSTF for experi-
ment T4.

1.1.3 Analysis of Experiment T4

The fourth tritium experiment—a steady-state
experiment—in the CSTF is in progress. This
experiment was started on February 3, 1976. Data
are still being accumulated, and results obtained thus
far are preliminary. The operational and experimen-
tal conditions for this test are described in Sect. 2.2.

An attempt was made to start this experiment on
January 27, 1976, with tritiated hydrogen being
added to the loop at 6 ¢m’(STP)/ hr. This attempt
was discontinued when the bypass rate of tritiated
hydrogen into the vacuum guard chamber surround-
ing the addition probe was found to be significantly
greater than had been observed during previous
trittum tests. This bypass rate was about 14 times the
rate at which tritium was reaching the circulating salt.

Flow tests to measure the leakage as a function of
addition-tube pressure indicated that, for an
addition-tube pressure of 35 kPa (5 psia), the bypass
rate would be 1.5cm’(STP)/ hr,approximately equal
to the addition rate to the salt, which was considered
tolerable.

Prior to the attempted addition on January 27,
samples from the CSTF showed the following tritium
concentrations:

Tritium in the salt: <l to 8 nCi/g (<40 to 300
kBq/kg)

Water-soluble (combined) tritium in off-gas: 20 to 50
pCi/cm’ (0.7 to 2 MBq/m’)

Elemental tritium in off-gas: 5 to 17 pCi/cm’ (0.2 to
0.6 MBq/m’)

Elemental tritium in cooling-duct air: ~Back-

ground*

Those pretest samples taken on February 3, prior to
the actual start of the test, were biased by the tritium
added on the January 27 attempt.

The tritium concentration in the salt (Fig. 1.4) rose
in a manner similar to that observed in previous tests
to 500 nCi/ g (~20 MBq/kg) about 2.5 days after the
start of the addition. The tritium concentration
subsequently rose to a value of 700 nCi/g (24.5
MBq/kg) (+10%).

The concentration of water-soluble tritium in the
off-gas increased to a value of 150,000 pCi/cm’ (5.5
GBq/m’) about four days after the start of the
addition. It then gradually increased to 180,000 to
190,000 pCi/cm® (6.7 to 7.0 GBq/m’). This corre-
sponds to a flow rate of 22 to 23 mCi/ hr (0.22t00.23
GBq/s) of tritium or 1.5 to 1.6 cm */hr of total
hydrogen (at 5600 ppm tritium in hydrogen) in the
off-gas in the water-soluble form.

The elemental tritium concentration in the off-gas
showed irregular behavior and scatter for about one
day following the start of the addition before
increasing in an orderly fashion to a value of 22,000
pCi/ cm’ (~0.8 GBq/m’). This irregular behavior was
not entirely unexpected and may have been due to
conditioning of new piping in the off-gas system that
had been installed prior to ‘this experiment. A
somewhat similar phenomenon occurred during the
first tritium addition experiment. The concentration
of elemental tritium in the off-gas subsequently
decreased to a value between 10,000 and 11,000
pCi/cm’ (0.37 GBg/m’ and 0.41 GBq/m’). There-

*No detectable radioactivity by comparison with blank sample.
ORNL DWG 76 - 316

O S S e S e B B e e o S B L s ma B e e K
i OFF -GAS, WATER-SOLUBLE ,{p Ci/cm?} e }
L s & _
. o ..-l--- L LR L LA N T Tt L e t : Y s " *
L -
5 ‘.. s
0 E - -
r d ]
[ o ]
E - OFF-GAS,ELEMENTAL , ( pCiZem™) .
= | X xy ' 4
- x & § ) X
E K x kX x l‘x‘ll{llxl“l XX
el oy ll x x % ot X i u‘ I X X % _]
g L x x . x % R
g L o x x 3
Zz i .
z b x i
[ - ~
< | !:“ .
é - % SALT (n Ci/q) ° -
: ) y
g i03 - %o ° @ °° o -
o [] L] ° ° ° 4 o .
§ : ° e° o © 0 oo a e °°°¢ °°‘ 0% o ° "t o % o0 " ,“' ° o o ‘oo % o ° J
B eas °° i
- ° fi° _
- BJ .
- o -
08. a ! a
ot L COOLING DUCT AIR,ELEMENTAL, (p Ci /cm®) x 100 s
E o’ 4 a 4 E
| a a A -
[ .A ‘4 4 a 4 4 . 4 ]
[~ o a a a a a a7
ah
- & aa s a : 4 x ad L) s L. a . . 4 a ~1
. 4 . P a4,  AMuoa .
10 I N S N OO WS NN (NS HNN SN N SN NN SURN NN SN N RO N R S N N R S T
3 5 7 9 T 3 15 17 19 21 23 25 27 29 |
F EBRUARY 1976 MARCH

Fig. 1.4. Tritium concentrations at various points in the Coolant-Salt Technology Facility during experiment T4,

fore, elemental tritium is appearing in the off-gas
system at a rate of 1.2 to 1.32 mCi/hr (~12.5to 13.6
MBgq/s) or 0.08 to 0.09 cm’/hr total hydrogen.

Measurable concentrations of tritium were ob-
served in the first sample of air taken from the loop
cooling duct 2 hr after the start of the experiment.
Since then, the concentration has varied between 0.2
and 0.8 pCi/cm’ (7 to 30 kBg/ m’). These concentra-
tions correspond to a net flow of 0.2 to 0.8 mCi/hr
(2.1to0 2.2 MBq/s) of tritium or0.013 t0 0.054 cm’/ hr
of total hydrogen through the walls of the loop piping
and pump bowl. Thus the total flow rate of tritiated
hydrogen out of the CSTF is from 1.6 to 1.75
cm’(STP)/ hr via the off-gas system and permeation
of the metal walls, with 96 to 999% of the total being
removed through the off-gas system.

Measurement of the bypass rate of hydrogen
around the addition probe indicated an average rate
of ~1.2 cm’/hr. Tritiated hydrogen is being con-
sumed (addition rate to loop plus bypass rate) at ~3.0

cm’/hr. Thus, with 1.6 to 1.75 cm*(STP)/hr being
removed from the CSTF, 89 to 979% of the tritium
that is fed to the addition probe can be accounted for.

The measured concentration of elemental tritium
in the off-gas corresponds to a partial pressure of 0.2
Pa (1.5 um Hg) for elemental tritiated hydrogen in
the pump bowl gas space. This value compares
favorably with the measured hydrogen partial
pressure in the main circulating salt stream, 0.43 Pa
(3.2 um Hg). If the liquid in the pump bowl is
assumed to be in-equilibrium with the gas phase, the
change in the hydrogen partial pressure between the
salt in the loop and that in the pump bowl may be
used to estimate the solubility of elemental hydrogen
in the NaBFs;-NaF mixture. For the nominal flow
rate of salt through the pump bowl, 15 gpm (0.95 X
107 m’/s), the solubility coefficient for elemental
hydrogen would be 1.7 X 107"* atm (atom/cm®)’,
compared to an earlier estimate of 4 X 107"
Conversely, if the earlier estimate of solubility is
used, the salt flow rate through the pump bowl is 36
gpm (2.3 X 10~ m’/s). Since both values are within
reasonable limits, the solubility of hydrogen cannot
be defined more precisely from these data.

A similar analysis of the data for combined tritium

-(hydrogen) in the off-gas indicates a hydrogen partial
pressure of 8.3 Pa (62 um Hg) in the gas phase,
assuming ideal-gas behavior and one atom of tritiurn
or hydrogen per molecule of compound. If this value
represents equilibrium between the liquid and gas
phases, the solubility coefficient for the water-soluble
hydrogen form is 1.7 X 107° atm (molecule/cm’
salt)”. This is close to the value of 1.1 X 107
estimated for the solubility of HF, which is one
possible form of the water-soluble compound.

The apparent rate of hydrogen permeation
through the loop walls, based on the observed tritium
concentration in the loop cooling duct, 1s between 0.7
and 2.7 X 10° cm?*STP)/m’-s. If the hydrogen
partial pressure in the salt 1s0.43 Pa, this rate requires
the effective permeability of the loop piping to be
between 10 and 50 times lower than that of bare
metal. A reduction of this magnitude would be
consistent with the presence of an oxide film on the
exterior metal surfaces.

Based on the measured concentration of water-
soluble tritium in the salt (700 nCi/g) and the
measured partial presure of elemental hydrogen (0.43
Pa), the ratio of the concentration of total hydrogen
in the water-soluble (combined) form to that in the
elemental form is ~4700. If this concentration ratio
were established in the coolant salt of an MSBR and
for the following conditions:

1. removal of combined tritium from a 10% salt
bypass stream (8000 gpm) with 809% efficiency;

2. no sorption of HT or TF on the core graphite;

3. reference value of 10 for the U*'/ U’ ratio in the
primary salt;

4. bare-metal permeability for the steam tubes (no
oxide film);

about 3.5 Ci/day (1.5 MBq/ s) of tritium would enter
the steam system and be released to the environment.
This release rate of tritium to the environment would
be within the range of release rates from pressurized-
water reactors (PWRs) currently operating. The
release rate of tritium for the MSBR includes the
entire fuel cycle, while that for PWR s would be about
80 Ci/day if releases associated with fuel reprocessing
were included.

1.2 NEUTRONIC ANALYSIS
H. T. Kerr E. J. Allen D. L. Reed

The neutronic analysis efforts during this reporting
period have focused primarily on the processing of
cross section data for MSBR performance evalua-
tions and for investigation of specific related
problems. Some work has been done in support of in-
reactor irradiation experiments.

1.2.1 Cross Section Processing
for MSBR Calculations

Neutron cross sections from the ENDF/B-1V data
files* have been processed to form a 123-energy-
group library with thermal scattering matrices for
tempe-atures ranging from 300 K to 1200 K. This
libraty includes all nuclides of interest to the MSR
program and will serve as a data base for all
subsequent neutronics analyses. In addition to the
ENDF/B-IV data, graphite cross sections obtained
from General Atomics Company have been in-
cluded.’ The cross section data available in this base
library must be further processed-for each unique
apphcation. Significant changes in the fuel salt
composition or in the fuel-moderator cell design can
strongly influence the effective cross sections,
particularly for strong absorbers such as ***Th.

The core of the reference design MSBR is divided
into two major zones.' Zone I is characterized by
approximately 13 vol 9% fuel salt, and Zone Il is
characterized by approximately 37 vol % fuel salt.
Each of these two major zones contains two unique
fuel-moderator cell designs (Fig. 1.5). The cylindrical
equivalent of each cell design is also shown.
Neutronic calculations based on one-dimensional
neutron transport theory were done for each
cylindrical cell to produce cell-averaged 123-energy-
group neutron cross sections for the salt constituents
in each fuel channel and for the graphite. Spectrum-
averaged capture and fission cross sections for some’
isotopes are summarized (Table 1.1); the strong
dependence on fuel channel dimensions is apparent.
However, volume averaging the cross sections for
both fuel channels in cells having the same total fuel
salt fraction gives total cell-averaged cross sections
that are not very sensitive to fuel channel dimensions.
Therefore, cell-averaged cross sections for a nominal
fuel salt composition are primarily dependent upon
the volume fraction of fuel salt in the cell.

*ENDF/B-1V is the Evaluated Nuclear Data File, Version 1V,
and is the national reference set of evaluated cross-section data.
peog———

AAAAAAAAAAAAAAAAAAAA
EEEEEEEEEEEEE
0,302—== g
y
0.25R

i |
/A

A

é‘:

04— |=—

/ _L 3500 d; 4.400
0.25 R fl////%////l : 3 | d3=4.513

Rsui

%, =4.000
, ‘ 1,=0.375
ZONE 2-B

Fig. 1.5. Sketches of MSBR core picces and equivalent cylindrical cells.

NNNNNNNNNNNNNNNNNN
Table 1.1.

Cell-averaged neutron cross sections for the four cell designs in an MSBR core

Cell-averaged cross section (b)?

. Core zone 1-A Cotre zone I-B Core zone I1-A Core zone
Reaction LB
Central Quter Central Quter Central QOuter
hole annulus hole annulus hole annulus Gap

For each fuel channel

232Th(n,v) 2.418 2.202 2.044 2902 1.458 2.751 1.905

233y(n,y) 13.03 12.10 12.54 14.25 6.825 8.342 8.805

233y(n,fn 110.5 102.1 108.2 123.3 50.96 62.30 69.44

238y(n,7) 9.721 9.101 8.458 9.220 9.863 11.54 9.774
For total cell

2327h(n,y) 2.225 2.352 1.627 1.905

233y(n,y) 12.20 13.15 7.022 8.805

233y(n.N 103.0 113.6 52.4 69.44

238y(n,v) 9.168 8.731 10.08 9.774

a1 b =108 m?.

As originally conceived, the cell-weighted cross
sections previously described were to be used in a
123-group one-dimensional transport theory calcula-
tion of an MSBR. Neutron energy spectra produced
in this calculation were to be used to collapse the
isotopic cross sections in each core zone into a few
groups (nominally 9) of a form suitable for
multidimensional performance calculations. How-
ever, considerable delays have been encountered with
the transport calculation in the AMPX code." First,
several program errors were discovered which
necessitated re-running all the cell calculations. Then
the observation was made that the calculated radial
flux profiles in the one-dimensional reactor model,
particularly in the reflector and reactor-vessel
regions, were strongly dependent on the initial flux
guess used to start the iterative procedure. This
unacceptable situation results from the use of
convergence criteria in AMPX which are not suitable
for one-dimensional reactor flux calculations in
regions having very low neutron source densities.
Program modifications are being made to incorpo-
rate suitable convergence criteria, and alternate
computational schemes are being considered.

1.2.2 MSBR Performance Calculations

Multidimensional performance calculations for
the reference design MSBR require collapsed, few-
group cross-section data for each zone of the reactor.
However, for the reasons previously given (Sect.
1.2.1), the appropriate cross sections are not

available. This fact and the planned termination of
the MSR Program at the end of FY 1976 will
preclude completion of the performance calcula-
tions.

1.2.3 Helium Production in
the MSBR Vessel

The mechanisms for helium generation from nickel
in Hastelloy N (ref. 5) are: (1) thermal-neutron
capture in **Ni followed by thermal (n,a) reactions in
the product *Ni and (2) direct, high-energy (n,0)
reactions in **Ni. At moderate neutron fluences

- (below about 10?* neutrons/cm’ with the nominal

neutron energy spectrum expected in the MSBR
vessel), the concentration of helium in Hastelloy N
due to the two-step, thermal process is adequately
represented by

N
Hey()) =—52 (60° ,*

where N is the atom fraction of *®Ni in the alloy, o is
the effective, spectrum-averaged cross section for
¥Ni(n,v)’°Ni reaction, o, is the effective spectrum-
averaged cross section for *’Ni(n,a)**Fe reaction, ¢ is
the total neutron flux, and ¢ is time.

*At very high neutron fluences, the helium concentration is
lower than this equation indicates because burnup of **Ni causesa
reduction in the rate of production of *Ni and the *Ni
concentration approaches an equilibrium value.
The dependence on the square of the fluence results
from the fact that the intermediate nuclide, *’Ni, must
be generated by neutron captures. The comparable
approximation for helium concentration due to the
high-energy, (n,a) process is

He2(t) =N03 (f)f s

where o3 is the effective, spectrum-averaged cross
section **Ni(n,a)>’Fe reaction.

The ratio of these two concentrations is a linear
function of fluence. Thus the high-energy process is
the dominant source of helium at low fluences, and
the two-step thermal process becomes dominant at
higher fluences. These relations also indicate that
changing the neutron flux spectrum (e.g., reduction
of the thermal component) would change the relative
importance of the two processes, as well as the total

helium concentration that would be generated in a.

given time period.

Cross sections for the “*Ni(n,7) reaction are fairly
well known, and available information for the
*Ni(n,a) cross sections indicates that a value of 13 b
(13 X 107** m? at 2200 m/sec is reasonable. The
neutron fluxes in an MSBR have been calculated
with several different computational models and are
generally in good agreement for the fueled regions of
the reactor. However, calculating neutron fluxes in
the reactor vessel is a particularly difficult problem
because the codes available for the source-type
problems are not designed for neutron flux conver-
gence in zones of low nuclear importance. A 20-group
S» transport calculation with special convergence
criteria was done. '

By combining the available cross-section data for
nickel with neutron fluxes from this calculation
estimates were made for the amount of helium
generated in the reactor vessel during the reactor
lifetime. The spectrum-weighted cross sections and
flux information used to estimate the helium
concentration at the inside wall are summarized
(Table 1.2). Using these values, calculated maximum
helium concentration in the vessel is 38 ppm.* About
90% of the higher concentration isassociated with the
two-step thermal process discussed above.

Although some uncertainty exists in the helium
concentration calculated at the inner surface of the
vessel, the calculated flux has a steep radial gradient

*Neutron captures in boron, which may be present at low ppm
levels as a trace contaminant in the alloy, would add a small
amount to these concentrations.

10

Table 1.2. Spectrum-weighted neutron cross
sections? and fluxes® for calculation of helium
production in the MSBR vessel

Reaction OT(b)C
S8Ni(n,v) 1.29
58Ni(n,a) 1.20x 1073
59Ni(n,a) 3.789
59Ni(n,abs) 26.8

. _Jo 0B o) dE
Op= po=
Jo o®(E)dE

where
O = spectrum-weighted cross section,
O(F) = energy-dependent cross section,
¢(E} = energy-dependent neutron flux, and

E = neutron energy.

bThe maximum calculated total flux at the inside
wall of the reactor vessel at the horizontal mid-
plane is 5.73 X 10?2 reactions/cm?-sec.

°1b=10"%% m?.

9Normalized to 13 b at 2200 m/sec.

in the vessel. This gradient causes a reduction in
helium concentration by a factor of about five for
each centimeter of penetration into the vessel wall.
Thus the concentration of helium in the Hastelloy N
vessel wall appears to be acceptably low.

1.2.4 Production of ***U in the MSBR

Uranium-232 has a decay chain that approaches
equilibrium over a several-year period; the nuclides
in this chain emit high-energy~alpha particles and
gamma rays. The alpha particles, in turn, may react
with other nuclides to produce neutrons. As a result,
shielding and handling costs of reactor fuels increase
as the *’U content increases. Increasing the content
of U in fuels of conventional reactors for which
reprocessing and refabrication are required raises the
cost by which these processes are performed and
lowers the value of the discharged fuel. In the absence
of light nuclides which can undergo («,n) reactions,
the principal radiations from the standpoint of
shielding are the 2.6-MeV gamma from **Tl and the
2.2-MeV gamma from *'*Bi. However, in the presence
of light nuclides (e.g., F, Be, and Li), as may be the
case for MSBR fuel, a significant neutron source
would also be present.

The ***U production in the reference design MSBR
was examined to determine the concentration of **U
that would be present in the MSBR. Decay of **?U is
not a direct problem for fuel used in the MSBR
{because no fabrication is required), but the content
of this isotope in excess bred fuel could affect
handling and shipping costs. The five processes by
which U can be produced are

230 "'75 2317 fi_é 231p, "'7; 232p, fl-a 232

232Th n,2n> 231Th 6_9 231P3i’% 232Pa fl-> 232U

L

For the “’Th:*’U ratios (~56) expected in the
MSBR, the (n,2n) reaction in ***Th is more important
than the (,2n) reaction in **U. However, removal of
the protactinium isotopes by fuel processing on a 10-
day cycle greatly reduces the production of ***U from
**Threactions. Asa consequence, the net production
rates of 2*U from the two (n,2n) processes may be
comparable in an MSBR.

The equilibrium concentration* of **U in the
MSBR is given by the equation:

1 A2 039 (90;Ngo + R2Ng2)

Ny, = + RN
asdths | u ) (maptag)
where

N22=232Uconcentration,’

Ngo = 23°Th concentration,
Ny, =232Th concentration,
Nas

¢ = neutron flux (average in core),

=233 concentration,

A, = decay constant of 232?Pa,

As = decay constant of 232U,
Ag =removal constant of Pa,

o, =(n,v) cross section of 23°Th,
R, =¢0; t ¢,0,,,

0, =(n,2n) cross section of 232Th,

*About 95% of equilibrium is attained in 15 full-power years.

11

024 = (7,7) cross section of 222 Th,
¢, = gamma flux,
03 =(n,7) cross section of 23! Pa,
Ry =¢o4 + 9,04,
04, = (7,) cross section of 223U,
04 = (n,2n) cross section of 233U, and

0s = (n,abs) cross section of 232U.

The equation assumes that the concentrations of
23U, ¥°Th, and **’Th are constant and that other
isotopic concentrations are in equilibrium.

Based on neutron fluxes obtained from the
diffusion theory model of the MSBR, values were
obtained for the various reaction rates. The reaction
rate values, in reactions per atom per second, are
summarized below.

0,6=1.25X 1078
0,0=1.92X 10713,
0py0, =196 X 1074
030=430X 1078 |
oa9=1.11X 10713
0449y =1.96 X 10714 |
gsp=1.57X 1078,
A, =6.08X 107
As =3.05X 10710,
Ag = 1.16 X 107% (ten-day cycle) .

The »°Th(n,¥) cross section was based on a 2200
m/sec value of 22.7 b (22.7 X 107* m’) and a
resonance integral of 996 b (966 X 107** m®). The (v,n)
cross sections for both *’U and ***Th were taken as
0.05 b (0.05 X 10"** m?) for gamma energies above 6.0
MeV and zero for'gamma energies below 6.0 MeV.
These very conservative (y,n) cross sections are based
on the limited information available for these
reactions.® The gamma flux was approximated by
assuming a volumetric fission source of gammas inan
infinite medium of pure graphite with a density equal
to that of pure graphite in the MSBR. The remaining
cross sections are better known, and the correspond-
ing reaction rates should be more reliable.

Substituting the reaction-rate values into the
equation for the equilibrium concentration of 2y
gives

Nyspyy =2.35X 1072 N30 4399 X 1077 Noyp | +8.19 X 1075 Nyyy
Solution of this expression for the reference design
1000 MW(e) MSBR with a ten-day processing cycle
for protactinium removal and no **°Th leads to an
equilibrium concentration of 20 ppm for U in total
uranium. Of this amount, 27% is produced by
reactions in ***U. The ***U source from **Th would
about equal the combined source from ***Thand *’U
at a concentration of 23 ppm “°Th in thorium.
Thorium-230, which is a decay product of 2*U, is
present to about 17 ppm in natural uranium. Thus, if
the thorium to be used in a reactor is obtained from
mineral deposits that also contain uranium, such
thorium could contain a significant amount of Z**Th.
The Th:U ratio in Monazite sands, a major thorium
resource, 1s typically about 20. Hence, thorium from
such deposits could be expected to contain only
about 1 ppm *°Th—an insignificant concentration
from the standpoint of ***U production. However, in
the absence of a large thorium industry, initial
demands for thorium conceivably could be met by
material recovered from uranium mill tailings, where
the Th: U ratio in the raw ore could have been (.2 or
less. (The uranium ores at Blind Riverin Canada, for
example, have Th:U ratios of about 0.2.) Hence the
2°Th content could exceed 100 ppm, which would
raise the equilibrium ***U concentration to about 107
ppm in total uranium in an MSBR,

1.2.5 Neutronic Analyses
of TeGen Experiments

The neutronic analysis for the TeGen experiments,
as described in the previous semiannual progress
report,” was based primarily on a flux mapping
experiment in the ORR poolside and on flux
monitors included in the TeGen-1 experiment. Flux
monitors have now been loaded into the TeGen-2 and
TeGen-3 capsules and will provide additional data
for characterizing the TeGen experiments.

The flux monitors in both TeGen-2 and TeGen-3
are identical and consist of two naturaliron wiresand
two vanadium—cobalt (0.215 wt %) wires. One wire of
each type is placed on the side of the fuel specimen
toward the reactor core, and the other wire of each
type is placed on the side away from the reactorcore.
The iron wires are 343 mm long and extend along the
entire length of the fuel pin region. The vanadium-
cobalt wires are 76 mm long and are located only
along the middle fuel pin.

The flux monitors have been recovered from the
TeGen-2 capsule and are being analyzed to determine
the induced activities. The irradiation of the TeGen-3
capsule has not been completed.

12

1.3 HIGH-TEMPERATURE
DESIGN METHODS

G. T. Yahr

Simplified analysis methods in ASME Code Case
1592 (ref. 7) and RDT Standard F94T (ref. 8) permit
the assessment of ratchetting and creep-fatigue
damage on the basis of elastic-analysis results,
provided a number of restrictive conditions are met.
Otherwise, detailed inelastic analyses, which are
usually quite expensive for the conditions where they
are currently necessary, are required to show that
code requirements are met. Analytical investigations
to extend the range over which simplified ratchetting
and creep-fatigue rules may be used to show
compliance with code requirements are being made
under the ORNL High-Temperature Structural
Design Program, which is supported in part by the
MSR Program. Modeling procedures for applying
the simplified ratchetting rules to geometries and
loadings prototypic of those encountered in LMFBR
component designs are to be identified. Then the
conservative applicability of these ratchetting rules
and procedures and of elastic creep-fatigue rules will
be demonstrated and placed on a reasonably sound

~ engineering basis. Finally, an assessment will be

made of the applicability of the simplified design
methods to Hastelloy N under MSBR design
conditions, and the importance of thermal ratchet-
ting in an MSBR will be determined.

The detailed plans for achieving the stated
objectives were given in a previous progress report.’
The basic approach is to perform a relatively small
number of carefully planned and coordinated
rigorous elastic-plastic-creep ratchetting-type anal-
yses of several geometries. Each geometry is
subjected to the axial, bending, thermal-transient,
and pressure loadings described in Table 1.7 of ref. 9.
Several two-dimensional cylindrical shell problems
are being analyzed at ORNL using the PLACRE
computer program,'® while three-dimensional
nozzle-to-shell problems are being analyzed by
Atomics International and Combustion Engineering
using the MARC computer program.* Each inelastic
analysis will include a complete code evaluation for
accurnulated strains and creep-fatigue damage. Also
associated with each inelastic analysis are a number
of elastic analyses to provide the input parameters
required to apply the various simplified ratchetting
rules and procedures and elastic creep-fatigue rules.

*MARC-CDC, developed by MARC Analysis Research
Corporation, Providence, RI.
The inelastic analysis results for the nine axisym-
metrical shell cases have been tabulated and
compared with the results of the Code Case 1592-3
elastic creep-fatigue rules. The elastic rules were
found to be conservative in all cases for combined
creep-fatigue, but the creep damage components
were not conservative for the notched cylinder and
the stepped cylinder, compared to the predictions of
the inelastic rules. One-dimensional inelastic anal-
yses simulating variouslocations in the shells are now
being performed.

Both of the three-dimensional inelastic analyses of
nozzle-to-shell configurations are under way again
after various problems, including cost overruns,
temporarily stopped the work. Combustion
Engineering has completed an analysis of three of the
five thermal downshocks to be applied to the nozzle-
to-cylinder configuration, and Atomics International
is approximately halfway through the analysis of the
one thermal downshock to be applied to the nozzle-
to-spherical shell configuration.

1.3.1 Circular Cylindrical Shells*

Results for the nine circular cylindrical shell cases
of this study have now been tabulated, and the results
of Code Case 1592-3 elastic rules have been
compared with the inelastic results. The elastic
fatigue damage rules have been found to be
conservative in all cases studied. Both the elastic
ratchetting rules and the elastic creep-damage rules
have been found to be nonconservative in some cases.
There are two types of geometries in which

nonconservatism hasarisen—ata notch in a pipe wall

and in the thick-walled portion of a shell involving a
stepped-wall thickness. :

Work 1s now under way to carry out one-
dimensional inelastic analyses simulating various
locations inthe nine circularcylindrical shell cases. It
is expected that these simulations will throw further
light on the nonconservatism found.

1.3.2 Nozzle-to-Spherical Shell Attachment™

The project on three-dimensional inelastic analysis
of a nozzle-to-spherical shell attachment was
discontinued for approximately five months because
of funding problems associated with computer costs.
1t was reactivated early in November, but the scope
has been reduced from a five-cycle to a one-cycle

*Work done at ORNL by W. K. Sartory.
TWork done at Atomics International by Y. S. Pan,

13

inelastic analysis. The elastic thermal stress analysis
and the full-scale inelastic analysis are now being
continued.

At this time, 26 thermal load incrementsand 1 zero
load increment have been completed. The tempera-
ture in the entire thin-walled nozzle section of the

-attachment reached 427° C (800° F) at the twenty-first

increment. The maximum temperature in the
structure at the twenty-sixth increment is 576°C
(1036°F). There are approximately six more incre-
ments to complete the thermal down-transient.
Problems associated with the number of increments
which could be submitted at one time, computer
hardware and system problems, and problems
associated with the error messages in the MARC
program have slowed the rate of analysis, but these
problems have been resolved.

1.3.3 Nozzle-to-Cylinder Intersection?

This investigation involves a detailed inelastic
analysis (elastic-plastic and creep analysis) of a
nozzle-to-cylinder intersection. The geometry of the
intersection is that of the primary inlet nozzle of the
intermediate heat exchanger (IHX)'f(')r the fast-flux
test facility (FFTF).

Cyclic thermal and mechanical loading will be
repeated until the strain-stress history shows a
repetitive pattern or steady-state response. A
maximum of five cycles will be done. The inelastic
analysis has been performed for the first three cycles.

1.3.4 Inelastic Analyses of MSR
Piping Subjected to Internal
Pressure and Transient
Temperature Cycles

The temperature histograms for two of the most
severe thermal -transients that might occur in an
MSBR have been obtained to serve as a basis for
evaluating the magnitude of thermal ratchetting and
creep-fatigue damage that might be expected. One of
the transients results from a 60% decrease in load
demand from full power operation in just 3 sec. The
maximum thermal downshock rate during this event
is only 3.3°C/sec (5.9°F/sec), and 120 such events
might occur over a 30-year life. The other event
considered is a control-rod scram; which resultsina
maximum thermal downshock rate of 19.3°C/sec
(34.7°F/ sec). This rather severe transient is expected
to occur no more often than once each year.

¥ Work done at Combustion Engineering by R.S. Barsoum.
Creep-rupture data and creep curves for Hastelloy
N at 593, 704, and 816°C (1100, 1300, and 1500°F)
have also been obtained. When cyclic stress-strain
data are obtained for Hastelloy N, solution of the
sample problems will proceed. Results from the
sample problems will be examined to determine
whether thermal ratchetting and/or creep-fatigue
damage will be significant problems in an MSBR.

REFERENCES

1. Molten-Salt Reactor Program Staff, Conceptual Design
Study of a Single-Fluid Molten-Salt Breeder Reactor, ORNL-
4541 (June 1971).

2. J. R, Engel and G. T. Mays, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 3-8.

3. O.L.Smith, Preparationof 123-Group Master Cross Section
Library for MSR Calculation, ORNL/TM-4066 (March 1973).

14

4. N. M. Greene et al., AMPX: A Modular Code System for
Generating Coupled Multigroup Neutron-Gamma Libraries from
ENDF/B, ORNL/TM-3706 (March 1976).

5. H. T. Kerr, D. L. Reed, and E. J. Allen, MSR Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 9-12.

6. E. G. Fuller, National Bureau of Standards, personal
communication, February 1976, ,

7. Code Case 1592, Interpretations of ASME Boiler and
Pressure Vessel Code, AmericanSociety of Mechanical Engineers,
New York, 1974,

8. RDT Standard F94T, Requirements for Construction of
Nuclear System Components at Elevated Temperatures (Supple-
ment to ASME Code Cases 1592, 1593, 1594, 1595, and 1596),
September 1974. :

9. J. M. Corum and G. T. Yahr, MSR Program Semiannu.
Progr. Rep. Feb. 28, 1975, ORNL-5047, pp. 1522.

10. W. K. Sartory, “Finite Element Program Documentation,”
High-Temperature Structural Design Methods for LMFBR
Components Quart. Progr. Rep. Dec. 31, 1971, ORNL]TM-3736,
p. 66. '
2. Systems and Components Development

R. H. Guymon

2.1 GAS-SYSTEMS TECHNOLOGY FACILITY
R. H. Guymon

The Gas-Systems Technology Facility (GSTF) was
operated for a short time during this period to obtain
more information on the salt pump problems.
Previous tests had indicated that the pump shaft
oscillations might be reduced if the operating speed
was not close to the critical speed.' This could be
accomplished by using a larger impeller and
operating at lower speeds. Thereforea 330-mm-diam
impeller was finish-machined from an existing
casting to replace the old 298-mm-diam impeller.
Back vanes were also installed on the upper surface to
reduce the fountain flow. After final balancing and
installation of the pump rotary element, the loop was
refilled with water and testing was resumed on
October 2, 1975. This impeller developed the desired
loop flow rate at significantly lower speed (Table 2.1).

The amplitudes of the shaft oscillations were
smaller with the larger impeller, but the deflections
were greater. At 1509 rpm with the 330-mm-diam
impeller, the deflections varied between 1.5 and 2.2
mm from run to run and at times showed definite
signs that the rotary element was rubbing against
stationary parts. The reasons for the variations under
apparently identical conditions are not known. The
deflections are high because of the performance
mismatch between the impeller and the volute. This
mismatch resulted from pump design modifications
made in an effort to use an existing pump.’

The operating characteristics of an MSRE coolant
pump would be satisfactory for the present require-
ments of this loop, and the shaft oscillations and
deflections should be low. Therefore, consideration
was given to converting the Mark Il pump into a
coolant pump. This could be accomplished by
installing an available coolant-pump volute in the
Mark II bowl and using a coolant-pump impeller.
The bowl could also be shortened so that an existing
short-shaft rotary element could be used. The cost of
these modifications was estimated at $104,000.

The effectiveness of the impeller back vanes in
reducing the fountain flow was determined.* With no
gas flow to the bubble generator, the rate of off-gas
flow from the bubble separator was about 1 liter/ min

when the water level was below the centerline of the ;

pump volute and a few tenths of a liter/ min at levels

15

Table 2.1. Performance comparison for impellers

in GSTF pump
Parameter Value

Impeller diameter, mm 298 330
Main-loop flow rate ? m>/min 1.93 1.89
Pump speed, rpm 1770 1509
Pump speed, % of critical speed 78 735
Pump-tank gage pressure, MPa 0.10 0.10
Pump discharge gage pressure, MPa 0.41 0.41
Motor power, kW 28 32
Shaft oscillation amplitude, mm 2.4 0.5
Shaft deflection, mm 0.5 >1.5

“Desired nominal value is 1.90 m3/min.

about 100 mm above the centerline of the volute. This
was thought to be due to a reverse fountain flow
injecting gas from the pump bowl into the loop,and it
would indicate that the impeller back vanes are too
long. The water flow rate in the fountain area was
0.38 m’/min at a pump speed of 1500 rpm, based on
gas-flow material balances in the pump tank.
However, these tests do not provide information to.
differentiate between fountain flow and reverse
fountain flow. Operation at lower pump speeds gave
a lower fountain flow rate. This would be expected
for normal fountain flow because of the lower head
developed. However, changing the speed could also
affect the reverse fountain flow by changing the ratio
of the pumping efficiency of the back vanes to that of

" the main impeller vanes. It might be possible to

determine whether the flow is normal fountain flow
or reverse fountain flow by running tests at different
heads but at the same salt-pump speed or at different
liquid levels in the pump tank.

Work restarted on fabrication of components
required for salt operation was stopped because of
the planned termination of the MSR Program at the
end of FY 1976. The loop hasbeendrained and put in
standby condition, and all records and parts have
been stored.

*Fountain flow is the flow of pumped liquid from the volute,
past the pump-shaft labyrinth, to the pump tank. This flow returns
to the loop at the pump suction, which is always submerged in
liquid.' Reverse fountain flow implies a flow from the pump tank
to the loop at the shaft penetration, which may not be fully
submerged; hence, this flow could cause gas ingestion.
2.2 COOLANT SALT TECHNOLOGY
FACILITY

A. N. Smith

A third transient tritium addition test was
completed by the end of October 1975. The loop was
then shut down to (1) install a probe to measure the
partial pressure of hydrogen in the circulating salt, (2)
revise the off-gas sample system, and (3) install a
cooling-air sample system to measure pipe-wall
permeation. Operation of the loop was resumed on
January 16, 1976, and a steady-state tritium addition
test was started on February 3, 1976. Attheend of the
report period, the steady-state test was still under
way, and work on the fabrication of a second
hydrogen addition tube had been started. The first
tube, now in service, developed a leak, which makes it
unsuitable for use in subsequent planned tests. The
data accumulated thus far indicate that improvement
is still needed in both the salt and air sampling
procedures; efforts are being directed toward this
end.

2.2.1 Loop Operation

The loop was in operation at the beginning of the
report period, and operation continued until Novem-
ber 4, 1975, when the system was shut down to make
equipment changes related to the tritium tests (Sect.
2.2.3). Operation was resumed on January 16, 1976,
and continued throughout the remainder of the
report period at 540°C with other basic loop
operating conditions described previously.® The
operating experience with the off-gas cold trap
continues to agree with the previous experience.* The
off-gas system has operated without any sign of
plugging, indicating that the salt mist filter’ is
effective. As of 0800 on February 29, 1976, the loop
had accumulated 5673 hr of salt circulating time since
being reactivated in December 1974.

16

2.2.2 Tritium Tests

A third transient tritium addition test (Test No.
T3), similar to those described previously,® was
started on September 16, 1975. Data for calculation
of the amount of added gas are shown (Table 2.2).
The tritium concentration in the salt declined to
about background levels approximately 300 hr after
the addition was stopped, and the concentrations in
the loop off-gas continued to decline for several more
days. A few additional salt and off-gas samples were
obtained during the next four weeks while experi-
ments were performed to determine the nature of the
planned system modifications. The analyses of these
samples showed no significant changes from earlier
results. The CSTF was then shut down on November
4, 1975, to make the modifications.

On January 27, 1976, an attempt was made to start
a steady-state tritium addition test by pressurizing
the tritium addition tube to an absolute pressure of
550 kPa. The test was discontinued when it was noted
that the bypass into the vacuum annulus surrounding
the tritium addition tube was about 1400 times higher
than the rate which had been observed during the
previous tritium tests. Under these conditions, the
hydrogen bypass rate was about 14 times higher than
the planned rate of input into the salt (84 cm’/hr vs 6
cm’/hr). Tests indicated that the bypass rate
decreased rapidly with addition tube pressure, and at
an absolute pressure of 35 kPa, the leak rate was
about the same magnitude as the permeation rate
into the salt. A bypass rate of this size would be
tolerable; so the steady-state test was started on
February 3, 1976, with a nominal absolute pressure in
the tritium addition tube of 35 kPa.

The basic plan for the steady-state test is to add
tritiated hydrogen to the circulating salt at a
relatively constant rate long enough for the tritium
concentration in the salt and the rate of transfer of
tritium to the two sinks (off-gas and permeation

Table 2.2. Trifium addition data for Test T3

Addition started
Addition time, hr

Initial absolute pressure in the transfer cylinder, Pa
Final absolute pressure in the transfer cylinder, Pa

Total gas transferred, cm> (STP)
Stray leakage, cm?® (STP)

Net gas transferred to salt, cm’ (STP)

Tritium concentration, vol %

Total tritium transferred to salt, mCi

Average tritium addition rate, mCi/hr

0924, 9-16-75
9.87

1.37 x 105
1.08 X 10°
30.8

through the salt pipe walls) to approach steady state.
The absolute pressure in the addition tube is
controlled at an average value of about 35 kPa by
pressurizing to 42 kPa and then repressurizing when
the pressure has dropped to about 32kPa. A 500-cm’
volume was added to the addition-tube piping (Fig.
2.1) to reduce the absolute rate of pressure drop and
thus permit operation for at least 24 hr before having
to repressurize the addition tube. The rate of pressure
drop in the addition tube is continually monitored,
and these readings are used to evaluate the rate at
which tritiated hydrogen is being supplied to the
system. The rate at which tritiated hydrogen is being
bypassed into the vacuum annulus is evaluated by
periodically isolating an appended 1000-cm’ volume
and noting the rate of pressure increase. The net rate
at which tritiated hydrogen is being added to the
circulating salt is then the difference between the total
addition rate and the rate of bypass into the vacuum
annulus (in the absence of other leaks). The net input
of tritium into the system is determined from the
input gas rate and the tritium concentration of the
input gas, which is determined by collecting periodic
samples of the input gas for analysis by mass
spectrometry.

The outflow of tritium in the off-gas stream is
determined, as in the previous transient tests, by
sampling the off-gas stream. To estimate the rate at
which tritium permeates through the salt pipe walls,
samples of the air from the lower loop cooling duct
are taken, and the air flow from the duct 1s measured.
The ratio of the total permeation rate to the
permeation rate into the cooler is assumed to equal

500-cm®
VOLUME

TRITIUM
TRANSFER
CYLINDER

17

the ratio of the respective pipe surface areas. The
inventory of chemically combined tritium in the salt
is determined by sampling the salt, while the
concentration of elemental tritium in the salt is
inferred from readings taken on the hydrogen
pressure probe (Sect. 2.2.3). Independent estimates
of the elemental tritium in the salt are also inferred
from the concentration of elemental tritium in the
off-gas stream and from the rate at which tritium is
permeating the salt pipe walls.

The status of the steady-state test at the end of the
report period, at which time the test was still in
progress, is summarized as follows:

1. As of 0800 on February 29, 1976, the steady-state
test had been in progress for 622 hr.

2. The concentration of the addition gas was about
5700 ppm tritium, the net rate of gas input to the
salt was about 1.8 cm’(STP)/hr, and the rate of
tritium input was about 26 mCi/ hr.

3. During the initial phase of the test, samples were
taken at about 2-hr intervals. After the tritium
concentrations started to level off, the sample
frequency was reduced to three per day.

4. The salt sample results showed considerable
scatter, and work isin progress to determine if this
is caused by the sampling procedure.

5. The tritium level in the air samples from the loop
cooling duct varied randomly between about 0.3
and 0.8 pCi/cm’. This concentration is near the
limit of sensitivity of the analytical method being

ORNL-DWG 75-12616R

PURIFIER

SAMPLE

HYDRQGEN

VACUUM

1000-cm®> VOLUME

b

359C
VACUUM
ANNULUS ~_| |
100 °C
ADDITION
TUBE ‘

?SALT
FLOW ™

Fig. 2.1. Tritium addition system for CSTF.
used, and various schemes are being considered
for increasing the tritium concentration, such as
reducing the cooling-air flow and raising the
concentration of tritium in the addition gas.

6. The data from the hydrogen partial-pressure
probe indicate that the partial pressure of
elemental hydrogen in the salt is quite low (less
than 10 um Hg), which is generally in agreement
with the low concentration of tritium found in the
cooling-air samples and also with the concentra-
tions of elemental tritium found in the off-gas
samples. Plans are being made to verify the
hydrogen pressure probe readings by using a
mass-spectrometer leak detector machine.

Test results are discussed in Sect. 1.1.3.

At the end of the report period, the tritium levels in
the salt, off-gas, and cooling air appeared to be at
steady state. The plan is to continue the test long
enough to examine various schemes for improving or
verifying the data for tritium concentration in the
cooling air and in the salt and to provide better
estimates of the tritium material balance, trapping
efficiency, and .mass-transfer processes. In the
meantime, a new tritium addition tube is being
fabricated for subsequent tests. These will be
designed to show variations in trapping efficiency as
a function of the gas addition rate and the presence of
chemical additives in the salt.

2.2.3 Equipment Modifications

During the loop shutdown beginning November 4,
1975, the test equipment was changed to improve the
quality of the data and to provide additional data.
The following is a summary of the changes:

|. Hydrogen Pressure Probe. A probe was
inserted in the main salt stream at a point about 1.25
m downstream from the salt pump discharge to
measure the partial pressure of elemental hydrogen in
the salt. The probe consisted of a sealed Hastelloy N
tube, 12.7 mm OD X 1.1 mm wall, connected to the
necessary valving and instrumentation to permit
pressure readings inside the probe in the range 0of 0.01
to 100 Pa (0.1 to 1000 um Hg). Since this probe is
completely submerged in salt, its response should
depend only on conditions in the salt.

2. Off-Gas Sample System. Several changes were
made for the purpose of improving the reliability of
the sample system and the quality of the sample data:
(a) A salt mist filter was fabricated from porous
copper sheet (made from pressed and sintered copper

18

powder) with a porosity of 0.23 and.a pore hydraulic
radius of 6.0 um. The filter, housed in a Hastelloy N
body, was installed at the inlet to the gas sample line
with a heater and insulation to permit operation at
540°C and was oriented such that trapped salt could
drain back to the off-gas line. (b) A 3-way Monel
valve was installed in place of the 3-way glass
stopcock immediately upstream of the first tritium
pretrap. The glass unit had some leakage and
condensation of fluid and also was not suited for
operation at the higher temperature now being used.
(c) The temperature of the gas sample line from the
salt mist filter to the inlet to the first tritium pretrap
was increased from 150 to 200°C. This change was to
minimize the condensation of fluids in the gas sample
line. Such fluids, collected in the sample lines, were
probably partly responsible for some earlier erratic
sample results.

3. Exhaust-Air Sample System. Equipment was
installed for collecting samples that can be used to
estimate the rate at which tritium is being lost due to
permeation through the loop pipe walls. A sample
tap was installed in the air line coming from the lower
loop cooling duct (Fig. 2.2); this tap was connected to
a diaphragm pump that circulates sample gasat a rate
of about 1700 ¢cm’/min and vents the gas into the
loop enclosure. At the pump discharge, where the
pressure is controlled at an absolute pressure of 140
kPa, a 100 cm’/min side stream is taken off for
tritium trapping. The sample stream is mixed with 20
cm’/min of humidified helium, and the combined
streams are passed through a CuO furnace to convert
the tritium to tritiated water and then through a
water trap and a cold trap to collect the tritiated
water. The tritium content is determined by scintilla-
tion counting in a manner similar to that used for the
off-gas samples. The air intakes to the lower cooling
duct were extended so that the air is drawn from
outside the loop enclosure, thus avoiding contamina-
tion of the air samples by extraneous tritium which
might be in the enclosure. The flow of air through the
cooling duct is determined by sampling the air stream
while adding a known flow of tritium to the air inlet.

In the original design of the tritium addition
system, a vacuum pump maintained a low pressure
(less than 75 Pa) in the vacuum annulus surrounding
the addition tube. The discharge gas from the
vacuum pump was vented into the loop enclosure and
was eventually discharged to the stack along with the
loop ventilation air. Thus the air exhausted from the
loop enclosure normally contained a small amount of
tritium from the vacuum annulus in addition to the
tritium that permeated through the loop pipe walls.
19

ORNL-DWG 76-5258

AR o LOOP
Al E
‘ AIR NCLOSURE
| IYITIIIEETIEITISITIETTEIN I I RTINS N U AT ITIIIISS zzml r_zzzzzzzzzzzmzzzzzz;
LOWER COOLING DUCT 4
' : APPROX ;
1.3 m /s AIR g : .
; AIR SAMPLE
,
FLOW
CHECK ) |+ B- VENTILATION BLOWER
B“:':LT; ~ APPROX. POINT 7 PP:::T:‘;N '-3'/”5 -
PIPE 0.1 m3/s ? A m>v/s Al
LOWER COOLING DUCT e /
EXHAUST LINE "—a
' AIR
SAMPLE %fi
RECIRCULATING AIR STREAM
1.5 TO 20 liters/min ":E .
FILTER \ Il FI-72
CIRCULATING APPROX. 100 c¢m®/min
PUMP
FLOW

XX S0
KR I
S5 ....._."' SR )

Cu0 FURNACE
_ 600 °C
~ APPROX. 20 cm®/min

RESTRICTOR

HUMIDIFIER

HELIUM ~50 °C

Fig. 2.2. Diagram for air sampling

During the steady-state test, (Sect. 2.2.2), the amount
of addition gas which is bypassed into the vacuum
annulus is about 10’ higher than was the case during
the earlier transient tests. This extra tritium must be
kept out of the loop enclosure because extra tritium
there would interfere with the evaluation of pipe-wall
permeation from the tritium content of the loop
exhaust air. Therefore, shortly after the steady-state
test was started, a new line was installed to route the
vacuum pump exhaust around the loop enclosure
and directly to the SLlCth]'l of the stack fan.

2.3 FORCED CONVECTION LOOPS

W. R. Huntley  R. H. Guymon
M. D. Silvermdn

Operation of forced-convection loopsis part of the
effort to develop a satisfactory structural alloy for
molten-salt reactors. Corrosion loop MSR-FCL-2b

S ..o'o‘

WET TEST
METER

U

TRAP

coLD
TRAP

system CSTF tritium experiment.

is presently operating with reference fuel salt at
typical MSBR velocities and temperature gradients
to evaluate the.corrosion and mass transfer in
Hastelloy N alloys. At this time, the loop has
operated approximately 4000 hr with fuel salt at
MSBR reference design conditions with the expected
low corrosion rates in the standard alloy. A test has
been started to obtain corrosion data on specimens of
Hastelloy N containing 1% niobium. |

Construction of two additional corrosion loop
facilities, designated MSR-FCL-3 and MSR-FCL4,
continued for most of this reporting period, but this
effort has been discontinued because of the decision
to terminate the MSR Program at the end of FY
1976. : :

2.3.1 Operation of MSR-FCL-2b

A 4000-hr corrosion test was completed on
October 2, 1975, in which standard Hastelloy N
specimens were exposed to fuel salt at typical MSBR
operating temperatures. During the test, 18 corrosion
specimens were exposed to salt at velocities up to 5
m/sec at three different temperature levels—565,
635, and 705°C. As expected, corrosion rates were
low throughout the test period and <3 um/year (0.1
mil/ year) during the final 1000-hr increment of the
test. A detailed discussion of the corrosion results is
presented under Salt Corrosion Studies (Sect. 6.6).

On October 6, a series of NiF; additions to the loop
salt was begun to increase the U*"/ U ratio. It was
planned to increase the U*'/ U’" ratio from the earlier
equilibrium value of about 100 to a new level of about
1000 and then to evaluate corrosion rates at the
higher level, which reflects a more oxidizing
condition in the salt. The NiF, was added by
dropping small pellets of NiF,-fuel carrier salt (50-50
wt %) from a sealed air lock, via a ball valve and riser
pipe, to the fuel-salt surface within the auxiliary
pump tank. During the additions, the loop was
operated isothermally at ~695° C with a pump speed
of 4000 rpm to provide a salt flow rate of about 12
liters/ min. Four separate additions of 0.8 g of NiF;
were made between October 6 and October 21,
followed by a final addition of 1.6 g on November 3,
1975 (4.8 g total addition to the 14-kg salt inventory).
The work was interrupted at this time by operational
problems, but the limited observations made are
described below.

The effects of NiF, additions were similar to those
observed in September and October 1974,' when fine
NiF: powder was added to the salt inventory of FCL-

2b. Immediate pump-power increases of 8 to 10%

were again noted with each 0.8-g addition. The salt
level in the auxiliary tank decreased with each NiF;
addition, which suggests that NiF; affected the salt
surface tension and may have thereby reduced helium
ingassing in the pump bowl. During the first four
NiF; additions of 0.8 g each, the salt level dropped
about 1> mm immediately after each addition. A
1'>-mm level change represents an apparent salt
density increase of about 19%; therefore, density
change of the pumped. salt should account for only
about one-tenth of the pump-power increases noted,
and the remaining power increase is believed to be
due to increased drag forces on the pump impeller.
The final addition of 1.6 g of NiF; resulted in a 14%
increase in pump power and a larger-than-expected
level drop of 12 mm at the auxiliary tank.

Each NiF, addition to the salt resulted in an
immediate increase in the U*/U’ ratio, which
reached a maximum in a few hours, followed by
asymptotic decay; e.g., the first addition increased

20

the ratio from ~ 100 to >>1000, but after three days of
isothermal operation at 695°C, the ratio had
decreased to 130. Similarly, the fourth addition
increased the ratio from ~250 to 10,000, but it had
dropped to 200 after eight days. Just prior to the final
1.6-g NiF; addition on November 3, the U*/ U** ratio
was 150. Two hours after the 1.6-g addition, the ratio
was 2.6 X 107, and a day later the ratio had fallen to
4.8 X 10°. No further readings were obtained because
of a salt leak in the piping system which occurred on
November 5, 1975. The plan to run a corrosion test at
a high U”/U" ratio was abandoned at this time.

The loop did not accumulate any significant
operating time between November 5, 1975, and
February 18, 1976, because of salt leaks that occurred
on November 5 and December 10, 1975. Both leaks
resulted from tube ruptures during reheating opera-
tions immediately after transfer from AT to isother-
mal operation of the loop.

Salt freezing had always occurred in the coolers of
FCL-2b after a loop scram* because of the large mass
of air-cooled metal within the cooler housing and the
relatively small mass of hot salt within the cooler
coils. This had not caused significant problems
during the past 20,000 hr of operation, other than
delaying resumption of salt circulation for a few
hours while gradual remelting occurred. However, in
each of the two recent shutdowns where leakage
occurred, salt froze in another portion of the loop not
monitored by a thermocouple, in addition to the
known freezing in the coolers. Since the second
frozen area was not apparent to the operators during
either incident, normal operating and remelting
procedures were followed which resulted in salt
liquid expansion and pipe rupture between the frozen
coolers and the unsuspected salt plug.

The salt leak of November 5, 1975, occurred in a
resistance-heated section of the loop piping (I°R
heater No. 2). The location and cause of the frozen
area which precipitated this failure are not definitely
known, but the cold spot was probably at metallurgi-
cal station No. 2. The salt leak of December 10, 1975,
occurred at metallurgical station No. 3, where an
adapter (for line S08) is saddle-welded to 19-mm-OD
X 1.8-mm-wall tubing. The cause of the frozen area
related to this leak was a 10-cm-long uninsulated
portion of a salt drain line (S-112). Insulating this
portion of line had been overlooked during prior

*Scram is used in this context to indicate automatic corrective
action (initiated by instruments or manually) taken in response to
the development of an off-normal condition to establish minimatl
stable operating conditions.
repair operations because of piping complexity and
inaccessibility.

Several design modifications were made to the
loop to reduce the likelihood of further incidents of
this type. The scram circuits were revised to provide
continuous salt-pump operation at a reduced speed
of 2000 rpm after each scram instead of the previous
procedure, which stopped the pump. This provides
more heat energy to the cooled metal within the
cooler housing via the flowing salt. Also, the cooler
housing was modified by adding internal thermal
insulation and electric heaters to reduce the mass of
cold metal to which the salt-containing cocler coil
can transfer heat. Thirdly, an automatic solenoid
valve was added to turn off auxiliary cooling air to
the resistance-heating lugs on the main heatersaftera
scram or whenever the main heaters are deactivated.

Piping replacement and modification were neces-
sary after each of the two incidents of salt leakage.
The entire length of the standard Hastelloy N
resistance-heated section (/°R heater No. 2) was
replaced, as well as metallurgical station No. 3. Drain
lines were replaced because of damage from salt
running down the exterior tube surfaces from the
leakage areas above. All three 6.3-mm-OD drain
lines were replaced with 9.5-mm-OD lines because
the original lines were too small and had hampered
operation and repair functions. A new, larger
standard Hastelloy N fill-and-drain tank was also
installed because the original tank had a marginal
capacity.

The loop repairs and modifications were com-
pleted in February 1976, and 20.25 kg of new fuel salt
mixture was transferred to the new Hastelloy N drain
tank. The salt was sampled after transfer into the
drain tank, and impurity levels were found to be low,
which indicated that the new salt was in good
condition.

The loop piping was filled with salt on February
18. After the salt circulated isothermally for 3 hr, the
oxidation potential of the salt was measured
electrochemically and found to correspond to a
U*/U* ratio of about 4600. The U /U’ ratio
dropped rapidly, as had been the previous experi-
ence, and decreased to a value of about 500 after two
days’ operation at design conditions and full AT.

A scram test from AT operation was made to verify
that the recent design modifications would prevent
salt from freezing in the coolers. The test was
successful and showed that the original scram circuits
functioned and that the newly added automatic
features worked as planned; that is, the pump speed
reduced from 4000 to 2000 rpm, the guard heaters on

21

the coolers were deenergized, and the cooling air on
the resistance-heater lugs was cut off. Because of the
new design modifications, the salt continued to flow
at a reduced rate after the scram and the isothermal
circulating salt temperature fell to only 565° C, which
is considered a safe level above the salt liquidus
temperature of 500°C.

At the completion of the scram tests, 18 new
corrosion specimens were installed in the three
metallurgical sample stations. The corrosion speci-
mens are fabricated from modified Hastelloy N
containing 19 niobium, instead of the standard
Hastelloy N specimens tested previously in FCL-2b.
After the specimens were inserted, the loop was
operated isothermally at about 700°C over the
weekend and was brought to design AT conditions on
February 23. Operation is continuing, and removal
of the corrosion specimens for their first examination
is planned after 500 hr of AT operation.

2.3.2 Heat Transfer Studies in MSR-FCL-2b

The heat transfer performance of the fuel salt
proposed for an MSBR (LiF-BeF-ThF«UF,; 71.7-
16-12-0.3 mole %) was measured in one of the
resistance-heated sections of MSR-FCL-2b. Heat
transfer data were obtained over the following range
of variables: Reynolds moduli, 1540 to 14, 200;
Prandtl moduli, 6.6 to 14.2; heat fluxes, 142 to 630
kW/m? salt velocities, 0.5 to 2.5 m/sec; and. fluid
temperatures, 549 to 765°C.

Film coefficients ranged from 132 to 11.8
kW/m?*K at Nusselt moduli of 11 to 102. There was
satisfactory agreement with the empirical Sieder-
Tate correlation’ in the turbulent region at Reynolds
moduli from ~8000 to 14,200 (Fig. 2.3). Between
Reynolds moduli of ~2100 and 8000, the experimen-
tal data follow a modified Hausen equation,® which is
normally applicable to the transition region between
laminar and turbulent flow. The extended transition
region is probably due to the high viscosity and large
negative temperature coefficient of viscosity of the
fuel salt. Hydrodynamic stability theory predicts the
possibility of such an extension when heat is
transferred from a solid interface to a fluid whose
viscosity decreases with temperature. Insufficient
data were obtained in the laminar flow region to
allow any conclusions to be drawn. The results of
these experiments are similar to those obtained in
FCL-2 with coolant salt’ (NaBF.-NaF; 92-8 mole %)
and indicate that the proposed fuel salt behavesasa
normal heat transfer fluid.
22

ORNL-DWG 76-5259 -

10
— l YA I [T TTTFI
My, = 0.027 A28 Y3 (KB —
Nu Re "Pr lpg —
51— /b.g? —
5 g |
- —
5 —
3 v
: ol /56 044 _|
2/3 1/3[ Hp
o =
E ; My = 0416 (Ng,“—125) N, (H)
4
<« /°
x
- 10! »
R = -
w
T — L/D=152 —
. _/ . _
S 5 9 —
KGR 7
s Ut 3 [ pg |04
":,,i ~ My, =186 [NR.NP,(D/L)]‘ (Ts_) ]
=
L/D =123
2 — , —
© 1.30m FROM INLET CORRESPONDING L/D = 123
® 4.60m FROM INLET CORRESPONDING L/0= 152
0 | I N I! l L 111
10 3 B 4 S
10 2 5 10 2 5 10

Npe, REYNOLDS MODULUS

Fig. 2.3. Heat transfer characteristics of LiF+BeF; *ThF, - UF, (72-16-12-03 mole %) flowing in a 0.416-in. ID tube.

233 Design and Construction of FCL-3
and FCL-4

All components for FCL-3 and FCL-4 are on hand
except the finned-cooler sections, which were to be
fabricated of titanium-modified Hastelloy N tubing.
- Components received during this . report period
include the resistance-heater lugs, loop shielding, two
ALPHA-pump bowls, FCL-4 control panels, and
parts for four ALPHA-pump rotary assemblies.
Assembly of the electrical power supply and auxiliary
piping for FCL-3 progressed as far as practical in the
absence of tubing for the saltloop. Assembly work on
FCL-4 electrical power supplies and auxiliary piping
was continued for most of this reporting period.
However, the FCL-4 assembly work has now been
stopped because of the decision to terminate the
MSR Program at the end of FY 1976.

REFERENCES

1. R. H. Guymon and G. T. Mays, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 16-22.

2. A. G. Grindell et al., MSR Program Semiannu. Progr. Rep.
Aug. 31, 1971, ORNL-4728, p. 35.

3. A.N.Smith, MSR Program Semiannu. Progr. Rep. Aug. 31,
1975, ORNL-5078, pp. 22-23.

4. Ref. 3, p. 23,

5. A.N.Smith, MSR Program Semiannu. Progr. Rep. Feb.28,
1975, ORNL-5047, p. 27.

6. A.N.Smith, MSR Program Semiannu. Progr. Rep. Aug. 31,
1975, ORNL-5078, pp. 24-25. ‘

7. E. N. Sieder and G. E. Tate, Heat Transfer and Pressure
Drops of Liquids in Tubes, Ind. Eng. Chem. 28(12): 1429-35
(1936).

8. H. Hausen, Z. Ver. Deut. Ing. Beih., Verfahrenstechnik (4),
pp. 91-98 (1943).

9. W.R. Huntley, MSR Program Semiannu. Progr. Rep. Aug.
31, 1971, ORNL-4728, p. 152.
Part 2. Chemistry

L. M. Ferris.

Chemical research and development related to the
design of MSBRs have concentrated on fuel- and
coolant-salt systems chemistry and the development
of analytical methods for use in these systems. This
work is presently being phased out along with the rest
of the MSR Program.

The chemistry of tellurium in fuel salt has been
emphasized because this element is responsible for
the shallow intergranular cracking of Hastelloy N.
Preliminary measurements of the solubility of Li;Te
in molten LiBeFs confirmed the expected low

solubility. High-temperature spectroscopy was used

to study the behavior of tellurium species in molten
LiF-BeF; mixtures. The results of these studiesare in
agreement with the measured low solubility of Li;Te,
and they show that the soluble species can probably
be represented as Te," . Spectroscopy was also used
to determine the equilibrium tellurium pressures for
the reaction 2LiTes = Li;Te + */,Te,, but the method
could not be used for similar studies with nickel and
chromium tellurides because of their very low
dissociation pressures. In other spectroscopic work,
equilibrium quotients for the reaction UF.(d) +
'»Ha(g) = UFs(d) + HF(g) were determined in LiF-
BeF, mixtures and in MSBR fuel salt, LiF-BeF:-
ThF: (72-16-12 mole %). The values obtained
generally are in agreement with those obtained
previously by other workers using a different
method. | X

Electrochemical studies of lithium tellurides
confirmed the low solubility of both Li;Te and LiTes
in LiCI-KCI eutectic at 400° C, but they did indicate

23

that an oxidant such as water vapor could generate
more soluble and colored species.

‘Work on several aspects of coolant-salt chemistry
was continued. It was shown that Na;B:F¢Os is the
stable oxygen-containing species in the system NaF-
NaBF.:-B,O: at low oxygen concentrations. This
compound reversibly interconverts with NaBF;OH
in aqueous s.'ystemS, a situation which is probably
also the case in the coolant salt. Based on this
assumption, a plausible mechanism for trapping
tritium from an MSBR in NaF-NaBF, coolant salt
has been postulated. Vapor density studies of the
system BF3;-H>O were also continued. It was shown
that BF3-2H,O is completely dissociated in the vapor
phase at temperatures above about 200°C. Below
200° C, association occurs, and a stable liquid phase
can be formed. .

Monitoring of U*'/U’" ratios in several forced-
convection loops, thermal-convection loops, and
creep-test machines was continued. Data were
obtained during the third tritium injection experi-
ment at the CSTF. The results were similar to those
from the first two experiments in that most of the
tritium was found in a water-soluble or combined
form. Electroanalytical studies indicated that LiTes
and CniTes; are insoluble in MSBR fuel salt.
Cathodization of a tellurium electrode in the
presence of fuel salt apparently produced a species of
the type Te, (n = 1), which was unstable under the
experimental conditions. Some evidence for oxide
and/or peroxide species was obtained in other
electrochemical studies.
3.

Fuel-Salt Chemistry

A. D. Kelmers

3.1 SOLUBILITY OF LITHIUM TELLURIDES

IN FLUORIDE MELTS
D. Y. Valentine A. D. Kelmers

It has been demonstrated that fission-product
tellurium is the agent responsible for the shallow
intergranular cracking of Hastelioy N surfaces which
was observed in the MSRE.' However, the state or
states in which tellurium is actually present in the
LiF-BeF,-ThF4-UF4(71.7-16-12-0.3 mole %) fuel salt
under reactor operating conditions remains to be
determined. The lithium-tellurium system was the
first to be investigated. The two species that are
known to exist in the Li-Te system, Li;Te and LiTes,
have been prepared.’ LiTe; undergoes disproportion-
ation to Li;Te and Te quite easily’ and thus does not
lend itself to dynamic solubility measurements. On
the other hand, Li,Te is quite stable and was used
here to make the first tellurium solubility estimates.
Molten Li;BeF, was used instead of the thorium-
containing fuel salt since the density of Li,Te (3.4
g/cm’) is similar to the density of the fuel salt (3.75
g/cm’). However, Li;BeF, hasa density of 2.2 g/ cm’,
which is probably sufficiently different from L1, Te to
prevent the suspension of Li;Te particulates in the
melt. -

The solubility study was carried out in a closed
nickel vessel completely lined with POCO-AXF-5Q
graphite. The various parts are shown before
assembly (Fig. 3.1). Openings in the graphite liner lid
accommodate a thermocouple (TC) well and a sparge
tube, both Y4-in.-OD, also made from POCO-AXF-
5Q graphite. The TC welland sparge tube were joined
to the top of the nickel pot by brazing a tantalum ring
between the graphite and the nickel pot. Design of the
graphite parts and the graphite-Ta-Ni joint was
contributed by H. C. Cook of the Metals and
Ceramics Division. The brazing techniques were
developed by J. F. King of the Welding and Brazing
Laboratory, Metals and Ceramics Division. This
type of joint is reported to withstand temperatures up
to 1150°C.* It has been in continuous use in this
experiment for many weeks over a temperature range
of 500°C to 800°C without showing signs of failure.

Two stainless steel ball values welded to the top of
the pot allow samples to be taken with filter sticks.
The filter sticks were also made of POCO-AXF-5Q
graphite. A screw cap on the end of each filter stick

24

holds a graphite frit disk in place.* The pore size of
the frit is about 150 w. The filter stick, frit, and screw
cap are shown in the foreground of Fig. 3.1. The filter
stick (s-in.-OD, Yis-in.-1D) was found to be
extremely strong. An enlarged chamber behind the
frit holds the sample. The filter stick is admitted to
the melt through a Teflon sleeve above the ball valves
on the pot top. A Swagelok fitting is used around the
filter stick, allowing it to pass into the pot through a
closely fitting Teflon sleeve. In addition, an over-
pressure of argon is maintained while the fiiter stick 1s
being lowered into the melt, assuring that contamina-
tion of the melt from ingress of air does not occur.

Enriched (95.07%) '**Te was purchased from
lsotope Sales Division, ORNL. Irradiation of this
isotope in the Bulk Shielding Reactor for 48 hr
produced '?"Te having a half-life of 117 days. This
isotope emits a 159-keV gamma ray suitable for
counting with a lithium-drifted germanium detector.
The '*"Te was melted with natural tellurium (purity
99.999+ wt % from Alpha Ventron Products Co.)ina
tungsten crucible. The tracer-labeled tellurium (Te*)
was then used to prépare Li;Te* in the same tungsten
crucible by melting the lithium on the surface of the
Te* and slowly raising the temperature until reaction
occurred. The Li;Te* product was then ground and
sintered to evaporate any excess reactants and
bottled for removal to the counting facility. The
dilution and subsequent preparation of Li;Te* were
carried out in an argon atmosphere vacuum box
equipped with a heater designed especially for use in
vacuum boxes.

After the vessel top was welded into place, the
empty vessel was installed in the vacuum-argon
system and heated under vacuum to 800° C to test the
weld joints. The graphite parts had previously been
outgassed at 1000°C for 24 hr under vacuum. The
evacuated vessel was then transferred to the argon
atmosphere box, where it was loaded through the ball
valves with Li;Te* followed by Li;BeFs. The vessel
was then replaced in the argon-vacuum system and
heated while maintaining a 50 cm’/ min argon sparge
flow. The furnace temperature was controlled by an
Electromax 1l controller to *1°C. The entire
assembly is shown in Fig. 3.2.

*The frit was supplied by J. M. Robbins, Metals and Ceramics
Division.
25

PHOTO 0699-76

Fig. 3.1 Parts of Li;Te solubility pot before assembly: nickel pot, graphite liner with lid, nickel lid with ball valves, graphite
theromocouple well and sparge tube attached. Foreground shows filter stick parts before assembly: tube, frit, screw-on end cap.

Samples were taken at temperatures between
500°C and 700°C. No wetting of the graphite by the
salt occurred, and the filter sticks could be withdrawn
smoothly. After a sample had been taken, the filter
stick end cap was unscrewed, the frit removed, and
the body of the filter stick encapsulated under
vacuum in a quartz tube. The tube was then heated in
a furnace to about 750°C, allowing the sample to
flow by gravity into a quartz cup in the bottom of the

tube. The sample did not adhere to the quartz; so
uniformly shaped sample buttons were produced
which were then weighed and counted.

The 159-keV gamma rays from the '*"Te tracer
were detected with a high-resolution lithium-drifted
germanium detector (10% relative efficiency). A
PDP-11/05-based pulse-height-analysis system was
used to acquire the energy spectra and subsequently

26

PHOTO 062776

Fig. 3.2. Overall view of equipment with Li; Te solubility pot installed in argon-vacuum system.

to perform quantitative isotopic analyses.* The first
few samples were counted at two source-to-detector
distances (0 cm, directly on the detector endcap, and
30.5 cm, a distance at which the photon detection
efficiency was accurately known from previous
calibrations) to establish the efficiency at closer
geometries. Subsequent samples were counted at 0
cm only.

To perform the quantitative analysis, the computer
estimates the baseline under the full energy peak and

*Two identical systems were used for these analyses. The first
samples were counted on the low-background system of F. F.
Dyer, Analytical Chemistry Division. Subsequent samples were
counted on the system of K. H. Valentine, Instrumentation and
Controls Division. One of the samples counted on Dyer's system
was used to obtain a quantitative calibration of Valentine’s system,
thus using Dyer’s careful efficiency calibration work.

then integrates the peak. Data such as half-life and
branching ratio are supplied to the computer in the
form of an isotope table and are used in the activity
calculation (corrected to any arbitrary reference date
and time). The only correction not automatically
applied by the program is that resulting from self-
absorption in the sample. However, the gamma-ray
attenuation length' for Li:BeF. isabout 3.6 cmat 159
keV, and since the samples are small (less than 0.5
cm), this correction does not exceed 10%. In
addition, the production of sample buttons of
consistent size and shape facilitated this correction.

The first samples taken indicated a solubility of
10" to 10~ mole fraction over the range between

¥ Attenuation length is that length by which the intensity is
reduced by a factor of 1/e.

475°C to 700°C. More recent data indicate a
solubility below 107 mole.fraction at 700°C. It is

suspected that particulates of Li;Te may have.

contaminated the first measurements. The melt has
been examined for possible oxidation, and no
evidence of an oxidized tellurium species was seen.
Samples will continue to be taken until reproducibil-
ity can be achieved, which presumably will indicate
that any particulate settling process has been
completed. The melt will also be examined for
suspended colloids. :

3.2 SPECTROSCOPY OF TELLURIDE
- SPECIES IN-MOLTEN SALTS

B. F. Hitch L. M. Toth

Spectrophotometric studies of lithium telluridesin
molten salts have continued. Several experiments
involving the addition of tellurium compounds to
LiF-BeF, (66-34 mole %) mixtures have been
conducted using a diamond-windowed graphite cell.
The primary objective of these experiments is to
identify the anionic tellurium species present in
fluoride melts at 500-700° C which absorb in the 200-
to 2500-nm range.

When Li;Te was equilibrated with LiF-BeF; melts,
the diamond-windowed cell was used in the conven-
tional manner with the atmosphere above the meltin
contact with the atmosphere inside the furnace.
However, when LiTe; was used as a solute, the entire
cell was encapsulated in quartz and the cellcapsule
assembly was positioned in the furnace such that the
entire unit was isothermal to prevent mass transport
of Te; to the cell walls by the decomposition reaction:

2LiTe; — Li;Te + *1Tex(g) . ()

In LiF-BeF; melts, LizTe did not show any
absorbance except in one case where a small (0.1
absorbance unit) broad band appeared at approxi-
mately 470—480 nm. This band disappeared when the
cell was evacuated, indicating that the absorbing
species was probably notdueto Li;Te butinstead toa
species of greater tellurium activity which was
present in Li;Te; that is, the Li;Te was not
stoichiometric and may have contained a slight
excess of tellurium.

Additions of LiTe; to LiF-BeF, mixtures gave a
small band (0.2 absorbance unit) at 450 nm which
remained stable. Previous work in chlorides indi-
cated that maximum absorbance was obtained with
solutions in which the lithium—tellurium stoichiome-
try was approximately one. Thus a Li;Te addition

27

was made to the LiF-BeF, mixture containing LiTe;
which shifted the lithium—tellurium stoichiometry to.
45-55 at. 9%. After the Li;Te addition, a very large
band (2 to 3 absorbance units) at 465 nm was
observed (Fig. 3.3) which remained stable during two
weeks of measurements. The color of the solution
was deep orange. Following these measurements,
enough tellurium metal was added to return the
lithium—tellurium stoichiometry to 25-75 at. %
(LiTes). The absorbance decreased slightly, but not
to the original 0.2 absorbance unit seen for the meltin
contact with LiTes. The anomalous behavior of LiTes
may arise from slow kinetics involved iIn its
equilibration with melts.

These results suggest that the most soluble
tellurium species in LiF-BeF; (66-34 mole %) melts
under isothermal conditions is an ion suchas Te,> or
Te". Our work in chloride melts,” the work of
Bamberger et al.,’ and recent electrochemical
studies” support the interpretation.

Currently, preparations are being made to care-
fully titrate a fluoride melt containing LixTe with
tellurium metal and to locate the lithium—tellurium

ORNL-DWG. 76- 6497
4.0 I Y I — I T

ABSORBANGE

O I 1 I | l I
800 600 - 400.
WAVELENGTH (nm)

200

Fig. 3.3. Absorbance spectrum of Te," ~.

ratio which gives the maximum absorbance. This
information will establish the stoichiometry of the
soluble lithium—tellurium species, and once this
stoichiometry is known; addition of this compound
(or mixture) to LiF-BeF; melts would establish its
solubility and absorption coefficient.

The following tentative conclusions are drawn: (1)
Li;Te does not appear to be soluble in LiF-BekF
melts, or, if it is soluble, the Te’” ion does not absorb
in the region investigated. (2) The light-absorbing
species present in LiF-BeF; solution (as well as in
molten chlorides) appears to be one represented by
Te.” (e.g., Te” or Te;”), based on our titration
measurements. Although solid compounds such as
LiTe or Li;Te; have not been observed in phase
diagram studies,’ the stability of a Te,"” ion in melts
could easily be explained by solvation effects. (3) The
experimental results have not been sufficiently
consistent to establish that the Te; 1on exists in
molten-salt solutions as a light-absorbing species. It
is, nevertheless, reasonable that it should occur as
one of a series of colored ions derived from the Te"
ion in the following manner:

Te + Te; = Tesy

or, in general,

Te,” + mTes = Teprom .

3.3 DECOMPOSITION PRESSURE OF LiTe;
B. F. Hitch L. M. Toth

In spite of frequent reference to LiTe; asa source of
tellurium in MSR experimental work,”™® no
equilibrium data are available for its decomposition
reaction:

2LiTe; — Li,Te + *;Tex(g) . (1)

As a result, Te, pressures over LiTe; (either solid or
liquid) have been a matter of considerable conjecture.
The relatively high Te; pressures (as indicated by the
characteristic yellow color over LiTes at tempera-
tures greater than 600°C) could be determined for
an isothermal LiTe; system by means of absorp-
tion spectroscopy; a Cary 14-H spectrophotometer
measured the Te; pressure over LiTe; by monitoring
the 4130-A band of Te; vapor.

Approximately 155 mg of LiTe; was loaded (under
an inert atmosphere) into a shallow tungsten
crucible, which was, in turn, placed in a 1-cm silica
cell, which was evacuated before being sealed off. The

crucible was necessary to prevent attack of the silica
by Li;Te formed during the decomposition. The
absorbance of Te; vapor over molten tellurium metal
was similarly measured, and the absorbances were
related to known vapor pressure data.'’ This
calibration enabled the conversion of the absorbance
data for Te; over molten LiTe; to pressures and also
permitted determination of the molar extinction
coefficient (e = 3232+ 132 & mole™' cm™) for Te: (Fig.
3.4). The upper curve gives the values for the vapor
pressure of pure tellurium. The equilibrium pressures
of Te; over molten LiTes are given by the lower curve
and can be represented by the equation In P(mm Hg)
= 15.938 — 12,720/ T(°K). As seen, the pressures of
Te, over molten LiTe; are quite high, being only
about a factor of 1.5 lower than the vapor pressures
for pure tellurium. This technique is applicable only
to tellurides. with similarly high Te; decomposition
pressures; therefore, decomposition pressures of
chromium and nickel tellurides cannot be measured
spectrophotometrically.

Earlier measurements by this technique indicated
somewhat lower pressures of Te: over LiTes. The
observed pressures decreased continuously with

ORNL - DWG. 76-6496

TEMPERATURE (°C)
700 600 . 500
| I

100 -

0
H

T T I LI

@
T
E 10'_“"
E Z
N
Q!l’ N
1O
04— : :
0.95 1.0 1.1 1.2 1.3
103 .
— (°K)

Fig. 3.4. Vapor pressure of tellurium over LiTe; and pure
tellurium as a function of temperature.

prolonged exposure at 700° C with the simultaneous
appearance of silica attack in the vapor space above
the crucible. Since Te. does not attack silica, even at
1050°C, probably one of the lithium tellurides has
some slight volatility at these temperatures. The
experiments reported above were performed at lower
temperatures to avoid this attack.

This study will be concluded with a determination
of the Te, pressure over a solid mixture of Li-Te and
LiTe;, which will provide an equilibrium constant for
Eq. (1). With this constant and the above data for
molten LiTe;, the activity coefficients for LixTe
dissolved in the molten LiTe; phase can be deter-
mined.

3.4 POROUS ELECTRODE STUDIES IN
MOLTEN SALTS—ELECTROCHEMISTRY
OF TELLURIUM IN THE LiCI-KCl
EUTECTIC SYSTEM

H. R. Bronstein F. A. Posey

Our previous studies have established the capabil-
ity of the packed-bed electrode of glassy carbon
spheres for monitoring electroactive species in
molten salts by use of voltage-scanning coulometry
techniques.'' It has been generally established that
intergranular attack and subsequent embrittlement
of the structural material of the MSRE was caused by
fission-produced tellurium. It would therefore be
highly desirable to establish the solubility and the
. valence state of tellurium species present-in the
MSBR fuel salt. The complex behavior of tellurium
species derived from the apparent dissolution of
Li;Te and LiTe; in various molten salts has been
amglfiylzdemo'nstrated by spectrophotometric stud-
ies.”

An electrochemical study of these tellurides in the
LiCI-KCl eutectic was initiated to provide additional
information which would help in the elucidation and
interpretation of the results obtained spectrophoto-
metrically. Of almost immediate impact was the
observation that the addition of Li;Te to the highly
purified and dehydrated LiCl-KCl eutectic melt
(400°C) produced no color, as was observed in
previous spectrophotometric studies.”> A volt-
ammetric scan showed no species attributable to the
dissolution of Li;Te. This means that Li;Te is not
soluble in this solvent at 400°C.

Next, LiTe; was added to the melt. No color was
observed and no trace of soluble tellurium species
was detected by volt-ammetry. Again, the conclusion
must be that LiTes also is not soluble in this solvent at

29

400° C. Removal of the carbon addition cup revealed
that the LiTe; had decomposed and had deposited
metallic tellurium on the outer surface surrounding
the holes in the cup. In light of the above results, the
decomposition of LiTe; releasing metallic tellurium
can probably be written as

Teg_ - Te_ + Tez .

Cathodization of an accurately weighed tellurium
electrode produced no color formation around the
electrode. From the coulomb accumulation and
weight loss of the electrode, a sone-electron process
was derived, in agreement with the finding of D. L.
Manning in fluoride melts.® The lack of color on
cathodization is also in agreement with that of
Plambeck." '

Our experiments, along with those of Plambeck "’
have shown that, not only are the lithium tellurides
insoluble in the LiCl-KCl eutectic at 400° C, but also
in highly purified melts the reaction products are not
present which lead to colored solutions.

These colored solutions may be due to interaction
of the tellurides and trace amounts of moisture
present as a contaminant in the melt. To test this
premise, LiCI-KCl eutectic-Li.Te mixtures were
contaminated with .varying quantities of moisture
and melted in sealed quartz tubes. As anticipated, the
tube with the largest contamination of moisture
developed the darkest color, whereas the least
contaminated melt had the lightest color. However, if
true solubility of tellurides should occur in other
melts, colored solutions could arise due to the
presence of polytelluride ions such as those observed
for the polysulfides."

3.5 THE URANIUM TETRAFLUORIDE-
'HYDROGEN EQUILIBRIUM IN MOLTEN
FLUORIDE SOLUTIONS*

L. O. Gilpatrick L. M. Toth

A continuing study" of the redox equilibrium
UF(d) + '2Ha(g) = UFy(d) + HF(g)
has produced values for the equilibrium quotient,

Q = [(UF3)(P,»)] + [(UF) (P )",

*This research in support of the MSR Program was funded by
the ERDA Division of Physical Research.
at a range of temperatures for the solvents LiF-BeF;
(66-34 mole %) and (48-52 mole %). Details of the
procedure have been given previously."

Since the spectroscopically measured UF;and UF,
concentrations are diréctly dependent on their molar
extinction coefficients* (e), it was necessary that
accurate values be known for both the uranium ions
in solution. Although ¢ values have already been
reported ' for UF4, no corresponding values exist for
UF; at 360 nm (which is the wave length used to
monitor the UF3). Therefore, the equilibrium study
was interrupted to obtain these data.

The reduction of a dilute UF; solution with
zirconium metal was found to be the most satisfac-
tory procedure for obtaining UF; excitation coeffi-
cients. By preparing UF,; solutions of known
composition and reducing all the UF4 to UF; with
excess zirconium, an accurate determination of e
at 360 nm was possible. At the same time, the
presence of trace Zr'' ion (<0.1 M) in solution, either
as a result of the reduction or intentionally added,
served to getter any trace oxide ion which would
otherwise precipitate UO; and upset-the accuracy of
the analysis.

Surprisingly, solutions of pure UF; were stable up
to' 700°C with respect to carbide formation in the
graphite spectrophotometric cell; therefore, it was
possible to perform the calibration on pure UF;
solutions. Above 700°C, loss of Ui’; was observed
(with the presumed formation of UC;), and it was
then necessary to extrapolate back to the time where
the instability occurred in order to obtain an estimate
of the absorption coefficients at higher temperatures.
Consequently, the high-temperature calibrationdata
are not as good as those at the lower temperatures.

Values of €30 apparently decrease somewhat with
increasing temperature, but the magnitude is within
the present experimental error. The average values
for the two solvents LiF-BeF, (66-34 mole %) and
(48-52 mole %) were found to be 1030 (for
500~750°C) and 1155 £ mole™ cm™ (for 400~750° C)
respectively. A discrepancy in the calibration
experiments for the latter solvent led to a rather high
(x10%) error. These experiments will be repeated if
time permits at the end of the H; reduction studies.

Equilibrium quotients at a given temperature were
determined by sparging the system with a fixed
HF/H; gas ratio foraday or more until the UF,/ UF,

* Defined by Beer’s Law, A = ec®, where A is the measured
absorbance, € is the molar extinction coefficient (in ¢ mole™'cm™),
c is the concentration (in mole 27"), and 2 is the path length (in cm).

30

ratio became constant. This length of time was also
necessary to obtain consistent HF analyses from the
furnace inlet and outlet gas streams because of the
very low HF/H, ratios (10™ to 10°%) which equili-
brated very slowly with the furnace and gas lines as
well as with the melt.

The values measured for Q are shown in Fig. 3.5
along with standard derivation (¢) bars. Included
also are the Q values calculated from the Long and
Blankenship (L & B) equation,’

logio @ =4.0—9.33 (10°/ T

+3.77 Xyp, + 2.09 (X, — 0.30) ,

in which X denotes mole fraction. The heat of
reaction, AH, for LiF-BeF: (66-34 mole %) deter-
mined from our data is 42 kcal and compares

. favorably with the L & B value 0of42.7 kcal. However,

our individual Q values all fall slightly higher than the
previous L & B values except for those at 800°C.
The equilibrium quotients for the 48-52 mole %
solvent are approximately ten times greater than
those calculated from the L &B equation. These
cannot be due to the 10% uncertainty in the e3so(UF3)
values because the error in the Q values is equivalent
to the error in the absorption coefficient. The 48-52
mole % solvent composition falls outside the range of

ORNL-DWG. 76-3804

107 ]
- N ]
i 48-52 mole %
B This Investigation . -
y i
& - E
E - \ 3
— — 66 34 mole % ]
wl o This ‘ 7]
afa L Investigation |
N o ‘ i
6 B
SleCE . :
wle T E . g
[ N
n - . . N ) {\—
i — 66-34 mole % / \\ ]
Long and Biankenship \
107 — N
a 48-52 male % /\ 3
- Long and Blankenship A
| . .
\ —
800° 700° 600° 500°
' | ! T T T T T
10% 7,

Fig. 3.5. @ values measured in LiF-BeF, solvents (66-34 mole
% and 48-52 mole %) at several temperatures for the reaction
UF4(d) + 4 Hy(®? = UF3(d) + HF(g).
the L & B data, and there is some indication in their
data that the Q values are not linearly related to the
BeF: concentration, as suggested by the above
equation (see Fig. 9 of ref. 17). Therefore, such
deviations due to gross solvent changes do not
necessarily conflict with the L & B data.

The heat of reaction in the 48-52 mole %
composition is 34.0 kcal and is 8.0 kcal lower than in
the 66-34 mole % melt. This is what one would expect
for a change in solution in going from a fluoride-rich
to a fluoride-deficient solvent and is in agreement
with our previous analysis'® for the UF,-UC;
equilibrium with UF; and graphite.

This study will be concluded by measuring Q with
LiF-BeF;-ThF, (72-16-12 mole %) as the solvent.

REFERENCES

I. A. D. Kelmers and D. Y. Valentine, MSR Program
Semiannu. Progr. Rep. Feb. 28, 1975, ORNL-5047, p. 40.

2. D. Y. Valentine and A. D. Kelmers, MSR Program
Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, p. 29.

3. B. F. Hitch and L. M. Toth, Sect. 3.2, this report.

4. J. F. King, personal communication.

31

5. B. F. Hitch and L. M. Toth, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, p. 30.

6. C. E. Bamberger, J. P. Young, and R. G. Ross, J. lnorg. -
Nucl. Chem. 36, 1158 (1974).

7. H. R. Bronstein and F. A. Posey, Sect. 3.4, this report.

8. D. L. Manning and G. Mamantov, Sect. 5.3, this report.

9. P. T. Cunningham, S. A. Johnson, and E. J. Cairns, J.
Electrochem. Soc. 120, 328 (1973).

10. A. A, Kadryartsev and G. P. Ustyugov, Zhur. Neorg. Khim.
6, 1227 (1961).

I1. H. R.Bronsteinand F. A. Posey, MS R Program Semiannu.
Progr. Rep. Aug. 21, 1975, ORNL-5078, p. 32.

12. D. M. Gruen, et al., J. Phys. Chem. 70, 472 (1966).

13. F. G. Bodewig and J. A. Plambeck, J. Electrochem. Soc.
117, 618 (1970); J. A. Plambeck, personal communication, Jan. 14,
1976. :

14. F. G. Bodewig and J. A. Plambeck, J. Electrochem. Soc.
117, 904 (1970). ' ‘

15. L. O. Gilpatrick and L. M. Toth, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, p. 31. '

16. L. M. Toth and L. O. Gilpatrick, The Equilibrium of Dilute
UF; Solutions Contained in Graphite, ORNL-4056, p. 12
(December 1972). .

17. G. Long and F. F. Blankenship, The Stability of Uraniu
Trifluoride, Part ll. Stability in Molten Fluoride Solution,
ORNL/TM-2065, Part 11 (November 1969).

18. L. M. Toth and L. O. Gilpatrick, J. Phys. Chem. 77, 2799
(1973).
4. Coolant-Salt Chemistry

A. D. Kelmers

4.1 HYDROLYTIC BEHAVIOR OF Na3B;F:0;
L. Maya

The hydrolytic behavior of Na3B;Fs0; was studied
by means of "’F nuclear magnetic resonance (NMR)
and Raman spectroscopy to clarify the possible
mechanisms of the conversion of this species into
hydroxide. The compound Na;B;F,O; appears to be

the stable oxygen-containing species at relatively low:

oxide concentrations (3-15 mole % B.03) in the
system NaF-NaBF.-B;O;. These compositions ap-
proach that of the coolant melt, although the latter
contains a lower oxide concentration. This finding is
of significance since tritium trapping by coolant melt

probably involves an equilibrium between oxide and
hydroxide species. A knowledge of the nature of
these species and their chemistry would provide
further understanding of the trapping process.

The Raman spectrum of a 0.66-M solution of
Na3B;F¢O; 1s given (Fig. 4.1). It was concluded that
the lines at 875 and 760 cm ™' were due to H;BO; and
BF;OH resulting from the depolymerization of the
B;FsO:" ion as described by the following sequential
reactions:

Na3;B;Fs0; + 3 H,O = 3 NaBF>(OH):

2 NaBF;(OH),=
NaBF;OH + H;BO; + NaF. (2)

()

ORNL DWG 75-15135R|

/\/

NoBF3OH,H3B80
NaoF SOLUTION
2MINB

SLIT I35 ¢m~!
LOWER TRACE
SLIT97cm~!

900 750

700

650 600 550

cm™!

Fig. 4.1. Raman spectrum of a solution 2 M in boron with an F/B ratio of 2.

32
This situation is not surprising since tetraborate is -

depolymerized in water into H;BO; and B(OH)s .
Hibben' observed such behavior on the basis of
Raman spectra.

The stoichiometry of the depolymerization re-
quired that sodium fluoride, which was not observed
in the Raman spectrum, would be formed. This
suggested that boric acid, sodium fluoride, and
sodium hydroxyfluoroborate that were dissolved in
equivalent proportions should generate the same
spectrum as aqueous NaiB3FsOs. This was found to
be the case; furthermore, evaporation of the solution
yielded pure Na;B;FsOs, as established by elemental
analysis, infrared spectroscopy, alkalimetric titration

33

experiment proved the reversibility of the equations
described above and gave an alternate and simpler
method of synthesizing the complex salt. In spite of
the apparently straightforward chemistry described
above, closer scrutiny of the Raman spectrum
revealed some inconsistencies. Pure NaBF:OH
absorbs at 763 cm ™', and the 760 cm ™' band actually
observed appeared to be asymmetric. A band at 610
cm ' was also detected. The intensity of this band
proved to be temperature dependent, increasing at
lower temperatures. This suggested the presence of
more species than originally assumed. The “FNMR
spectrum (Fig. 4.2) revealed a considerable propor-
tion of species containing an F/B ratio of 2.0. The

and the x-ray powder diffraction pattern. This spectral evidence indicates that the equilibria
ORNL-DWG 75-7869RI
I 1 T | 1
1 | L 1 W
-350 -30.0 -25.0 -200 -150 & ret CgFg
ppmM

Fig. 4.2. 19F NMR spectrum run at 0°C of a solution 2 M in boron with an F/B ratio of 2.
described in Egs. (l) and (2) are not shifted
completely to the right and that there are consider-
able amounts of BF,(OH).;” and B;FsO; ions
present. Analysis of the 760 cm ' band with a curve
resolver revealed two components at 763 and 753
cm ', the first being due to BF;OH', as assigned on
the basis of pure NaBF;OH, and the second due to
BF:(OH), . This gives an additional point to the

correlation of the »; band frequency of partially-

substituted fluoroborates with the degree of substitu-
tion. Thus, BF,~ absorbs at 770 cm™', BF; OH ™ at 763
cm ', BF2(OH),™ at 753 ¢cm’’, and B(OH)s at 745
cm '. Similarly, analysis of the ’F NMR signal at
—29.6 ppm (8 referenced to CsF¢) revealed two
signals, one due to BF;(OH).,” and the other
apparently due to the B;FsO;". The presence of
B3;Fs0;’ is confirmed by the band at 610 cm™ in the
Raman spectrum. This line corresponds to the line at
595 cm™' observed in the spectrum of the solid.

A semiquantitative material balance, based on the

intensities of the ’F NMR and a calibration of the
B(OH): and NaBF;OH lines in the Raman spectrum,
gives the following approximate distribution for a
solution 2 M in boron prepared by dissolving
equimolar amounts of NaBF;OH, NaF, and H;BOs:

BF30H™ BF,(0OH); B3Fg03> F~ H3BO;

Approximate 0.5 0.5 0.2 0.5 0.2
molarity

Free fluoride is not shown in the spectrum given in
Fig. 4.2. It absorbs at about —44 ppm and is not
resolved” from BF(OH); ; thus, among other
uncertainties, it is a potential source of error in the
material balance.

Having established that Na;:B;FsO; and NaBF;0H

interconvert reversibily in the presence of water, it is
possible to assume that a similar mechanism operates
in the melt. This led to the postulation of a number of
plausible reactions which could take place in the melt
involving these species, hydrogen (tritium), and
dissolved metal fluorides.

H, + MF, = 2HF + M’ (3)

2HF + 2NaBF;OH = 2NaBF, + 2H,0  (4)

2H,0 + %sNa;B;Fs0; =
NaBF;OH + NaF + H,;BO; (5)

H;BO; = ,B,0s + ,H,0 | (6)

2B20; + NaF + ',NaBF, = %Na;B;Fs0; (7)

%,NaBF, + *,H,0 = ;NaBF;0H + %,HF (8)

The overall reaction is
2H,; + 2MF; + '5sNa3B3Fs05 =
3HF + 2M° + NaBF;OH ..  (9)
The equilibrium constant for this reaction is

K. = PHF3 (NaBF3OH)
' (Na3B3Fe03)' *(MF;)* Py ?

The-above expression is not entirely representative of
the system since HF and NaBF;OH can react
according to:

HF + NaBF;OH = H,O + NaBF, . - (10)

This equilibrium constant for this reaction is
_ Py, 0
Py p(NaBF3;0H)

K,

NaBF. is not incorporated in the ex pression since 1t 1s
the solvent, Finally, the system could be represented
by combining K, and K; into the following expres-
sion;

) Pyr* (NaBF; OH)?
(N33B3 FGOB)I IB(MF2)2PH22 PH2 O

K3

Evaluation of the constant K; would lead to a
quantitative description of the system. A more
practical qualitative observation based on this
expression is that tritium trapping will be favored by
relatively high concentrations of oxide and/or metal
fluoride in the salt along with a relatively high value
for the partial pressure of water.

4,2 VAPOR DENSITY STUDIES IN
THE SYSTEM BF;-H;0

L. Maya S. N. Russell*

In the last report,’ it was stated that dissociation of
BF;-:2H,0 is complete at 400°C. The present set of
experiments was aimed at establishing the tempera-
ture threshold at which association or condensation
of the vapor species takes place. These experiments
showed that at about 200° C there is a departure from
complete dissociation. This departure becomes

*SCUU student participant.
pronounced at 150°C, and, with increasing amounts
of BF3-2H,0, a stable liquid phase is formed, giving a

saturation vapor pressure which remains constant’

with varying amounts of BF3-:2H,0. This behavior
was unexpected, in view of the report” that BF;-2H,0
decomposes upon distillation; however, that work
was done in glass in an open system. This finding led
to a series of experiments to obtain a set of saturation
vapor pressure data at different temperatures (Table
4.1).

Table 4.1 Saturation vapor pressures of -
BF3+2H; 0 temperature (°C)

167

101 127 150 ~ 166 ° 170
P;, mm Hg? 44 99 212 299 335 362
Py, mm Hgb .42 103 208 324 333 361

ap, = experirfientally determined pressure.
bp, = pressure calculated from a least-squares fit of the
data: log P (mm Hg) = (—2242 + 91)/T + (7.62 + 0.21).

A plot of In P vs T™' for BF;2H:0 gave an ex-
trapolated “boiling point” (actually the temperature
at which dissociation pressure and vapor pressure
reach 1 atm) of 200° C and an enthalpy of vaporiza-
tion (AH,) of 10.6 kcal/ mole, which is similar to the
AH, of a number of etherates of BF;’

Another set of experiments was conducted to study
the dissociation behavior of BF;-2H,0 and BF3-H,O
in the vapor phase. This was done choosing
conditions such that no liquid phase could be present.
Typical results of these runs are presented in Table
4.2. '

35

Table 4.2 Dissociation pressures
(calculated and experimental) of
boron trifluoride hydrates

(mm Hg)
Temperature BF3:2H,0 BF3-H,0
b
) PF? Pcb PFa Pc
150 146 206 158 189
160 166 212 188 214
170 284 341 314 338
181 364 402

"PF = pressure determined experimentally.
ch = calculated pressure assuming complete
dissociation.

Consideration was given to using the data in the
calculation of equilibrium constants and thermody-
namic quantities, but it was felt that because of the
uncertainty in the determination, valid results would
not be obtained. On the other hand, the results clearly
show that there is association between BF; and H,O
in the vapor phase.

REFERENCES

1. J. H. Hibben, Am. J. Sci. 113 (1938).

2. R. E. Mesmer and A. C. Rutenberg, /norg. Chem. 12, 699
(1973).

3. L. Maya, MSR Program Semiannu. Progr. Rep. Aug. 31,
1975, ORNL-5078, p. 33.

4. J. S. McGrath, G. G. Stack, and P. A. McClusker, J. Am.
Chem. Soc. 66, 1263 (1944). : _

5. H. C. Brownand R. M. Adams, J. Am. Chem. Soc. 64,2557
(1942).
36

5. Development and Evaluation
of Analytical Methods

A. S. Meyer*

5.1 IN-LINE ANALYSIS OF
MOLTEN MSBR FUEL

R. F. Apple D. L. Manning

Corrosion test loops described previously' have
continued operation with circulating reference fuel
carrier salt, LiF-BeF,-ThF,-UF, (71.7-16-120.3
mole %). Ratios of U*/U’" in thermal convection
loops NCL-21A and 23 continue to show no unusual
trend in the redox behavior of the melt. A temporary
rise in the ratio in response to the addition of new
corrosion test specimens is still visible as well as a
recovery to the original ratios within a short time.
The U**/ U’ ratios are presently about 5.4 X 10°and 4
for loops 21A and 23 respectively. The U*"/ U’" ratio
in loops 18C and 24, which are operating with
Hastelloy N corrosion specimens, have shown a
gradual decline and are presently about 1.7 X 10*and
80 respectively.

Forced convection loop, FCL-2b, has been
charged with new salt and is back in operation aftera
shutdown period. The U*'/*" ratio is presently of the
order of 5.3 X 10°, which indicates that the melt is
oxidizing. Probably the ratio will decrease rapidly as
the melt comes in contact with new metal surfaces
that have been installed in the loop.

The measurement of U**/ U’ ratios were initiated
in November on eight creep-test machines located in
Building 2011. Since the initial measurements, the
melts have tended to become more reducing. For
comparison, first measurements and present
U*/U’3 ratios are shown in Table 5.1. Tellurium as

Table 5.1 Ratios of U¥*/U in
creep-test machines

Machine U“/U‘“ ratio -
identity Initia] Present
15 4.6 x 10° 1.4 x 10°
16 3.9 x 10? 100
17 2.3x 10° 76
18. 3.5 x 102 90
19 3.5x 10% 35
20 5.3x 10% 63
21 65 22
22 3.1x 10% 64

J. M. Dale

Cr;Tes (~50 mg) was added to machines 19-22. No

detectable change in the redox behavior was
observed. Additional discussions of these tests are
presented elsewhere.’

5.2 TRITIUM ADDITION EXPERIMENTS
IN THE COOLANT-SALT
TECHNOLOGY FACILITY

R. F. Apple

The Coolant-Salt Technology Facility (CSTF) is
being operated for testing NaBFs-NaF (92-8 mole %)
for its suitability as a possible secondary coolant. In
cooperation with loop engineers and technicians, the
Analytical Chemistry Group is trying to determine
the fate and distribution of elemental tritium when it
is added directly to the salt to simulate, at least in
part, the predicted transport of tritium in the coolant
system via diffusion through the primary heat
exchanger. The methodology and the results from the
first two tritium injection experiments were described
previously.' This section presents the results of the
third tritium injection experiment.

About 80 mCi of tritium (diluted about 1:1000 with
protium) was introduced into the salt overa period of
about Il hr beginning at 0900 hours September 16,
1976. Tritium concentrations were measured in the
cover gas and salt from the beginning and for several
days thereafter (Table 5.2).

Preliminary evaluation of the data indicates that
about half of the injected tritium experienced
significant hold up in the salt but was eventually
removed in the off-gas stream. Very little tritium in
the off-gas was in the elemental form. The majority
was in a water-soluble or combined form which is
advantageous from the standpoint of tritium trap-

" ping by the salt. Also, there appears to be evidence

that at low concentrations some of the tritium which
is captured by the salt has a tendency to escape on
standing. A general discussion of the behavior of
tritium in the CSTF and a more complete analysis of
the injection experiments are given elsewhere.’

A fourth tritium injection experiment now
under way uses improved gas addition and sampling
techniques. Tritium will be added in this experiment

*Deceased.
37

Table 5.2. Tritium content in cover-gas samples and salt samples
after the third tritium addition at the CSTF

: Cover gas (pCi tritium/ml) Salt
Date Sample time
Water soluble elemental (nCi/g)
9/16 0810 38 0.42 1.60
0915
1115 785 5.0 11
1315 3082 8.0 18
1545 5770 21 27
1830 9254 23
1847 39
1920 10162 31 38
2115 9498 25 31
2315 8568 28 29
9/17 0824 10639 16
1124 5583 14
1424 4864 11
2024 2750 7.6
9/18 0820 2393 3.8 4.7
0910 1012 3.8
0950 720 3.6
1124 ' 4.3
1424 3.6
2021 895 4.4 3.0
9/19 0825 738 35 3.7
1425 724 2.3 2.6
2117 2.3
9/20 0825 450 1.3 2.6
1420 432 1.3 1.9
2116 3.1
9/21 0845 2.7
1420 1.2
2115 2.4
9/22 0825 1.2
1215 332 1.6
1310 378 1.6
1355 372 1.7
1426
9/23 0824 2.1
0905 191 1.2
1330 159 1.0
1430 1.9
9/24 0828 0.9
0945 143 0.99
1400 144 1.03 0.9
9/25 0824 1.4
0905 150 1.3
1345 125 1.0
1421 1.2
o 1421 1.2
9/26 0836 5.6
0836 - 5.5
0850 150 1.1
1415 209 0.90 1.1
1415 1.1
9/29 0905 100 0.75 0.77
1345 89 0.67 0.95
9/30 0815 86 0.71 2.4
1400 83 0.57 0.47
10/1 0815 73" 0.57 0.23
1400 70 0.59
10/2 0810 57 0.43
10/3 0810 44 0.38

until a steady state is reached. A preliminary
evaluation of the results obtained so far is presented
elsewhere.*

5.3 ELECTROCHEMICAL STUDIES OF
TELLURIUM IN MOLTEN
LiF-BeF.-ThF, (72-16-12 MOLE %)

D. L. Manning G. Mamantov

Tellurium occurs in nuclear reactors as a fission
product and results in shallow intergranular cracking
in structural metals and alloys.” Efforts were
continued to characterize this substrance electro-
chemically and to ascertain the feasibility of in situ
monitoring by electroanalytical means.

Tellurium screening studies were carried out in
cooperation with J. R. Keiser of the Metals and

38

Ceramics Division. As previously reported,® lithium

telluride, Li,Te, was added to molten LiF-BeF-:-
ThF., which was contained in a cell equipped with
viewing ports in addition to the electrode ports.
Subsequent voltammograms did not reveal any
waves that could be attributed to soluble electroac-
tive tellurium species. Chemical analysis indicated
<5 ppm tellurium in the melt. Following cleanup of
the cell and recharging with LiF-BeF,-THF,,
. standard additions of LiTe; were made in the form of
pressed pellets. The pellets disappeared more rapidly
from the melt surface than did the Li;Te pellets. A
greyish film having a metallic appearance formed
on the surface. Subsequent voltammograms did not
show any significant changes over background scans.
The equilibrium potential remained the same, which
was different from the Li>Te additions where the melt
became more reducing. This means that LiTe; is not
stable under our operating conditions. Bamberger et
al.” showed that isothermal conditions weie neces-
sary to hold the characteristic color for any length of
time in experiments on spectral measurements of
what was reported to be LiTes in LiF-BeF,. Under
nonisothermal conditions the color quickly disap-
peared with evidence of tellurium metal formation.
Since our conditions are nonisothermal, the LiTe;
probably decomposed immediately after contacting
the melt. The greyish film having a metallic
appearance that was noted on the melt surface
apparently tended to short out the electrodes after
a few hours because the voltammograms became
“extremely noisy and nonreproducible. However,
there was no change in the equilibrium potential.
Thus, we have been unable to detect stable
electroactive telluride species in molten LiF-BeF;-

ThF,: following standard additions of Li;Te and
LiTe; compounds under these conditions. It appears
that these tellurides are, for the most part, relatively
insoluble and/or thermally unstable under these
experimental conditions.

In the absence of meaningful voltammograms,
experiments were initiated to determine the decom-
position potential of elemental tellurium (mTe + ne”
— Ten,") relative to the half-wave potential of the
U(IV) — U(IID) electrode reaction. This should
provide some insight on the reducing power
[UIV)/ U(I1]) ratio] required to favor the existence
of tellurides (if stable) over elemental tellurium in
MSBR fuel salt. For these experiments, the molten
LiF-BeF;-ThF, was contained in a pyrolytic boron
nitride cup. The holder for the small tellurium pool
electrode (’s:-in.-diam) was fabricated from
spectrographic-grade graphite. Cathodic polariza-
tion curves recorded on the tellurium pool electrode
right after dipping into the melt at ~650°C and for
the next hour or so revealed a decomposition
potential of about +1.15 V vs the melt limit. It was
observed that ~ 100 mg of tellurium volatilized from
the graphite holder in approximately 1 hr. The half-
wave potential for the U(IV) — U(IIl) electrode
reaction relative to the melt limit is ~ +0.75 V. Thus
the decomposition potential of tellurium relative to
Ei, for U(IV) — U(III) reduction is of the order of
+0.40 V, which corresponds to a U(IV)/ U(111) ratio
of ~150at 923 K (650° C). From these measurements,
it appears that a relatively reducing melt is required
to favor the existence of stable telluride species over
elemental tellurium in molten LiF-BeF;-ThFs-U(1V)
at 923 K (650°C).

In a second experimental setup located in J. R.
Keiser’s laboratory and with Keiser’s cooperation, we
are monitoring the U(IV)/ U(11I) ratio in molten LiF-
BeF,-ThF.: following additions of tellurium as
CriTes. With the U(IV)/ U(I11) ratio at a stable value
of ~110at978 K (1300°F), a standard addition of 105
mg of Cr;Te, was made. This addition resulted in no
change either in the U(IV)/U(III) ratio or shape
of the voltammograms over a two-day monitoring

period. A second addition of 1.00 g of CriTes was

then: made, which resulted in an increase in the
U(1V)/ U(II]) ratio from ~110 to 220. This effect has
been observed previously and is believed to be due to
slight moisture contamination during addition
procedures. Within three days the U*/ U’ ratio had
decreased to about 90 and appeared to stabilize at
about this value. No change in the characteristics of
the voltammograms (new waves, etc) was observed
following the larger addition, although the U(IV)/
U(III) ratio appeared to level out at a somewhat
lower value. Metal specimens present in the cell will
be removed and examined for tellurium effects.

In order to obtain additional information on the
formation and stability of tellurides under noniso-
thermal conditions, studies were conducted on the
telluride species produced in situ from cathodizing
elemental tellurium (mTe + ne” — Te,,"). Chronopo-
tentiometric and double potential step experiments
conducted at a tellurium pool electrode contained in
a graphite cup revealed that the telluride species
generated does not appear to be stable at ~650°C.
Instability was indicated from the chronopotentio-
metric experiments by comparing the ratio of the
forward and reverse transition times.® Generation of
a stable but insoluble substance yields 7/ 7,=1; fora
soluble and stable species, 7/ 7, = 3 is predicted. For
an unstable species, on the other hand, the value of
7./ 7. should be greater than three. For these
experiments, the current was reversed ata time ¢ < 7y;
however, the above conclusions remain valid as long
as ! < 7. Potential-time curves recorded at the
tellurium pool electrode produced a value of 74/ 7,23
in all the runs, indicating that the telluride species
generated is not stable, at least within the time frame
of the experiment (seconds).

In the double potential step’ experiments, the
anodic—cathodic current ratio (i,/i;) is plotted vs a
function of time [ £(¢)] during which the potential step
is applied and removed. For a stable system, i,/i. =1
when F(1) is extrapolated to zero. For the generation
of an unstable species, i,/ i. < I; this was observed for
the tellurium experiments. '

Plots of log i vs E from potential step experiments
revealed an n value close to unity. The validity of n
value determinations by this method is discussed by
Armstrong et al.'° and Bacarella and Griess.'' Thus if
n = 1, the telluride generated can be represented as
mTe + ¢ — Te, (m = 1). Bronstein,'” using a
different method, also obtained an n value of unity
from polarization studies of tellurium in molten
chlorides.

A stable telluride of the type Ten should exhibita
color when dissolved in the melt.” In an effort to
observe this effect, we cathodized elemental tellurium
in molten LiF-BeF; at ~480° C. The melt was held in
a quartz tube to permit visual observations. When a
current of ~150 to 300 mA was applied for several
seconds or longer, a dark brown substance was
observed streaming from the tellurium electrode;
however, the material appeared insoluble, and a
colored melt was not produced. After the material

39

diffused away from the electrode a short distance (~1
cm), the color could no longer be seen. These results
also indicate that we are generating an unstable
species that under nonisothermal conditions under-
goes a decomposition reaction. Reasonable reactions
are as follows:

2Te™! — Te,”
Te)” = Tem + Tet (im=1)
2Te, — Te* + 2m—1) Tet (m > 1)

The Te®” does not appear to be soluble in fluoride
melts, at least to the extent that voltammetric
detection is feasible.

5.4 ELECTROCHEMICAL STUDIES OF
OXYGENATED SPECIES IN
MOLTEN FLUORIDES

D. L. Manning G. Mamantov

Voltammetric anodic scans at a gold electrode in
molten LiF-BeF.-ThF,(MSBR fuelsolvent)and also
in molten LiF-BeF,-ZrFs- (MSRE fuel solvent)
revealed two anodic waves at peak potentials ~1.05,
1.3 V and 0.85, 1.2 V vs an iridium quasi-reference
electrode (Ir QRE) respectively. Noise isencountered
on the diffusion current plateau of the second wave,
which is indicative of gas-bubble formation.at the
electrode surface. The waves are irreversible at slow
(0.1 V/sec) scan rates. Although the results are
tentative, these waves are probably associated with
the oxygenated species in the melts.

A calculation of the n value for the first anodic
wave from the ratio of the volt-ammetric i,/ v'/* to the
chronopotentiometric ir"* and also the chronoam-
perometric ir"* revealed a value of approximately
unity in both melts. Scan rate studies of the first wave
in LiF-BeF,-ThF, revealed a decrease in the i v'*
value of ~30% over the range of 0.1 to 5 V/sec. A
possible mechanism that reflects this behavior
involves a charge transfer followed by dimerization."’
If it is assumed that the wave is associated with
“oxide” in the melt, a likely reaction could be

O-l— 0O
200 — 0F
which would account for the one-electron charge and

dimerization. Efforts to carry out similar measure-
ments in LiF-BeF;-ZrF, were hampered by a small
prewave at the foot of the first anodic wave which
became very large at the faster scan rates. The peak
current (i,) of the prewave was roughly proportional
to v (scan rate), which is indicative of adsorption of
product (probably O") at the electrode surface. It is
not clear why this adsorption effect is so pronounced
in LiF-BeF2-ZrF,.

If the first anodic wave represents the oxidation of
oxide to the peroxide, the second anodic wave must
represent the continued oxidation to superoxide and
ultimately oxygen gas. The noise on the diffusion
current plateau of the second wave is indicative of gas
formation at the electrode surface. The reactionsare
complex; however, the overall n value for the second
wave should be 2 (assuming O,” ~ 2e — O).

For chronopotentiometry, the magnitude of the
transition time, 71 and 7, for the stepwise oxidation of
a substance involving » = | and n = 2, respectively,
results’ in a ratio of transition times 71/ 72 = 8. The
experimental data from chronopotentiograms re-
corded on both melts were in reasonable agreement
with the predicted value.

If the second anodic wave is a perioxide oxidation
wave, the addition of peroxide to the melt should
result in a pronounced increase in the i, of the wave.
We added 600 mg of Na;O; to the LiF-BeF,-ZrF,
melt. The first interesting observation was a 200-mV
anodic shift in the equilibrium potential (more
oxidizing) over a period of a few minutesand, then,a
gradual return to about the original value within a
few hours. A pronounced increase was observed for
the second anodic wave height. At faster scan rates
(~5-50 V/sec) the second wave is split, and a small
post-wave can be observed. This led us to believe that
we were seeing the intermediate peroxide — superox-
ide step and that the small post-wave was due to the
superoxide oxidation. To check this further, a
standard addition of sodium superoxide, NaQO,, was
made, and a pronounced increase in the peak height
of the post-wave was observed, which supports our
belief that the wave is due to superoxide oxidation.

It is also worth noting that a new cathodic volt-
ammogram at ~—0.5 V vs Ir QRE was detected
following the Na,O, addition. The wave apparently
involves a one-electron step. Although it is interest-
ing to speculate on various electrode reactions, at
present it can only be said that the wave results from

40

the reduction of peroxy and superoxy species. The
value of the reduction potential [~—400 mV more
negative than that for Ni(II)] for this process is
surprising.

After one day the wave heights of the volt-
ammograms had diminished, and after about four
days the effect of the Na,O, was no longer detected.
The reason that an increase in the first anodic wave
(believed to be due to oxide — peroxide) was not seen
as the peroxide decomposed (207 — Oz +207) is that
the melt may have already been saturated with oxide.
The fact that the second anodic wave increased upon
adding Na;O; supports our belief that this wave,
although complicated, is due to peroxide oxidation.

These observations continue to be encouraging.
The reproducibility of the curves for a given set of
conditions is generally good. It is believed that a
significant step has been taken toward the capability
of in-line monitoring of the oxide level in fluoride
melts under favorable conditions.

REFERENCES

1. A. S. Meyer et al., MSR Program Semiannu. Progr. Rep.
Aug. 31, 1975, ORNL-5078, p. 44.

2. This report, Sects, 6.11-6.13,

3. This report, Sects. 1.1.1 and 1.1.2.

4. This report, Sect. 1.1.3.

5. H. E. McCoy, “Materials for Salt Containing Vessels and
Piping,” in The Development and Status of Molten-Salt Reaciors,
ORNL4812, p. 207 (February 1975).

6. A. S. Meyer et al., MSR Program Semiannu, Progr. Rep.
Aug. 31, 1975, ORNL-5078, p. 48.

7. C. E. Bamberger, J. P. Young, and R. G. Ross, “The
Chemistry of Tellurium in Molten Li;BeF.,” J. inarg. Nucl. Chem.
36, 1158 (1974).

8. A. Weissberger (ed.), Physical Methods of Chemistry, Part
HA: Electrochemical Methods, p. 620, Wiley-Interscience, New
York, 1971,

9. Ry. 8, p. 602.

10. R. D. Armstrong, T. Dickinson, and K. Taylor, “The
Anodic Decomposition of Copper (1) N-Methyl-
hexamethylenetetramine Bromide,” J. FElectroanal. Chem.
64, 155 (1975).

11. A. L. Bacarella and J. C. Griess, Jr., “The Anodic
Dissolution of Copper in Flowing Sodium Chloride Solutions
Between 25° and 175°C,” J. Electrochem. Soc. 12, 459 (1973).

12. This report, Sect. 3.4,

13. W. V. Childs, J. T. Maloy, C. P. Keszthelyi, and A.J. Bard,
“Voltammetric and Coulometric Studies of the Mechanism of
Electrodimerization of Diethyl Fumerate in Dimethylformamide
Solutions,” J. Electrochem. Soc. 118, 874 (1971).
Part 3. Materials Development

H. E. McCoy

The main thrust of the materials program is the
development of a structural material for the MSBR
primary circuit which has adequate resistance to
embrittlement by neutron irradiation and to shallow
intergranular attack by fission product penetration.
A modified Hastelloy N containing 2% titanium has
good resistance to irradiation embrittlement, but
does not have sufficient resistance: to shallow
intergranular cracking by tellurium. It appears
necessary to modify the alloy by the addition of
niobium to impart better resistance to cracking.

Laboratory programs to study Hastelloy N-salt-
tellurium interactipns are being established, includ-
ing the development of methods for exposing test
materials under simulated reactor operating condi-
tions. Surface-analysis capabilities have been im-
proved so that the reaction products in the affected
grain boundaries can be identified.

The procurement of products from two commer-
cial heats (8,000 and 10,000 lb) of 2% titanium-
modified Hastelloy N continued. Some of the
seamless tubing was received and was of good
quality. The products are being used in all phases of
the materials program. One 2500-1b heat of 2%

41

titanium plus 1% niobium modified Hastelloy N was
procured as several bar products. Small, 50-Ib
commercial melts of several niobium-—titanium-
modified alloys were obtained and were included in
all phases of the program.

The work on chemical processing materials is
concentrated on evaluation of graphite for use in
reductive extraction steps involving molten salts and
liquid bismuth. Capsule tests are in progress to study

~ possible chemical interactions between graphite-

metal systems and bismuth-lithium solutions and to
evaluate the mechanical intrusion of these solutions
into the graphite.

Some of the effort during this reporting period was
expended in reestablishing test facilities. Six thermal-
convection loops are in operation in the new loop
facility, which will accommodate at least ten loops.
The mechanical property and general test facility is
partially operational, but numerous test fixtures
remain to be assembled and tests started. An air lock
has been added to the general test facility to make it
more functional, and plans were partially developed
for further expansion of the facility.
42

6. Development of Modified Hastelloy N
H. E. McCoy

The purpose of this program is the development of
metallic structural materials for an MSBR. The
current emphasis is on the development of a material
for the primary circuit, which is the most important
problem at present. Material for the primary circuit
will be exposed to a modest thermal-neutron flux and
to fuel salt- that contains fission products. A
modification of standard Hastelloy N will probably
be a satisfactory material for this application. An
alloy that_'containé 20, titanium appears to ade-
quately resist irradiation embrittlement, but the 2%
titanium does not impart adequate resistance to
intergranular embrittlement by tellurium. Additions
of 0.5 to 2% niobium are effective in reducing
tellurium-induced shallow intergranular cracking,
but the effects of neutron irradiation on alloys
containing -niobium have not been characterized
fully. Obtaining this information has been a major
part of our work during this report period. Increasing
the chromium concentration from the present 7% to
between 12 and 15% may also be beneficial in
preventing tellurium-induced shallow intergranular
attack. Currently, factors associated with production
of the 2% titanium-modified alloy in commercial
quantities are being studied, while smaller heats are
being made of Hastelloy N containing additions of
niobium alone as well as additions of niobium and
rare earths to 2% titanium-modified Hastelloy N.
These materials are being evaluated.

Two large heats, one 10,000 1b and the other 8,000
b, of the 2% titanium-modified alloy have been
melted by a commercial vendor. Product shapes
including plate, bar, and wire have been obtained for
use in the alloy development program. Tubing is
currently being produced by two independent routes.
The various product forms from the two large heats
are being used to fabricate the salt-contacting
portions of two forced-circulation loops. One 2500-1b
heat of an alloy modified with 2% titanium-1%
niobium was produced and converted to bar pro-
ducts.

Laboratory methods for studying Hastelloy N-
salt-tellurium reactions are under development.
Methods must be developed for exposing candidate
structural materials to simulated reactor operating
conditions. Tests are being run in which specimens
are exposed at 700° C to the low partial pressure of
tellurium vapor in equilibrium with tellurium metal
at 300° C. Other tests involve metal tellurides that are

either added to salt or sealed in evacuated quartz vials
to provide a source of tellurium. Several experimen-
tal alloys have been exposed to tellurium, and the
extent of intergranular cracking was evaluated
metallographically. Essential to this program are

"adequate techniques for identifying and characteriz-

ing the reaction products. Several methods for the
analysis of surface layers are under development.

Materials that are found to resist shallow inter-
granular cracking in laboratory tests will be exposed
to fissioning salt in the Oak Ridge Research Reactor
TeGen fueled-capsule series. Three sets of fuel pins
have been filled with fuel salt and are at various stages
of evaluation. These pins are made of alloys modified
with titanium, chromium, niobium, and titanium
plus niobium.

6.1 PROCUREMENT AND FABRICATION
OF EXPERIMENTAL ALLOYS

T. K. Roche B. McNabb J. C. Feltner

- 6.1.1 Production Heats of 2%
Titanium-Modified Hastelloy N

One of the activities in the materials development
program has been the production scale-up of 2%
titanium-modified Hastelloy N, a candidate structur-
al material for the primary circuit of an MSBR. As
reported previously,' two large heats of the alloy, one
10,000 1b and the other 8,000 lb, have been produced
by a commercial vendor. These heats have been used
to establish processing parameters for producing
plate, bar, and wire; and recent emphasis has been
placed on processing seamless tubing. Mill products
from these heats are being used in the general
experimental program, with a principal use being the
construction of two forced circulation loops designed
for studying the compatibility of the alloy with fuel
salt.

Two routes have been taken, one by the vendorand
the other by ORNL, for the fabrication of seamless
tubing, with the starting stock in both cases
originating from the 8,000-1b heat of the alloy (heat
No. 8918-5-7421, generally designated as 75421). The
vendor’s route included forging and. turning six
pieces of bar, each 4.5 in. in diameter by 6 ft long;
production of a tube hollow by trepanning to a 4.5-in.
OD X 0.5-in. wall by a subcontractor; tube reducing
(pilgering) in three steps with intermediate anneals to
2.0-in. OD X 0.083 and 0.095-in. wall;and drawing to
final tubing sizes, 1.0, 0.75, 0.5, and 0.375-in. OD X
0.035 to 0.072-in. wall by a second subcontractor.
Since the problem of stress-induced annealing
cracks or fire cracking had been encountered to some
degree with heat No. 75421 during cold drawing and
annealing of 0.312-in.-diam bar and 0.125-and 0.095-
in.-diam wire, the three tube-reducing operations by
the vendor were initially confined to one tube hollow
to determine if fire cracking would remain a problem.
The intermediate annealing temperature was
1121°C, followed by water quenching. Annealing of
the hollow between each tube-reducing operation
was preceded by the annealing of a sample which was
then liquid-penetrant inspected for evidence of
cracking. With this procedure the hollow was taken
through' the tube-reducing steps and annealing
treatments with no major problems. At this stage,
approximately 24 ft of 2.0-in. OD X 0.187-in. wall
stock was available for further processing. One-half
of this quantity, or 12 ft, was then tube reduced to
1.25-in. OD X 0.095-in. wall X about 36 ft at the
facilities of the subcontractor and was then returned
to the vendor for annealing. The tube was cut into
four lengths, each about 9 ft, and returned to the
subcontractor for the pilot drawing run to final sizes.
The reduction schedule by plug drawing and
sinking for the four lengths of 1.25-in. 01D X0.095-in.
wall tubing closely followed that for stainless steel
(i.e., approximately 25% reduction of area per pass).
The schedule was designed to cover three of the four
finish sizes required. The drawing and sinking
operations were performed on a 15,000 Ib draw
bench. Commercial lubricants were used. The
lubrication practice consisted of applying Houghton
Plastic No. 467 to the outside and inside surfaces,
then drying for about 30 min. During the actual
draw, a lubricant, G. Whitfield Richards EPHA, was
flowed over the outside diameter before the tube
entered the die. The tubing was degreased and
annealed at 1394 K (1121°C) after each draw.
Annealing was carried out in a continuous,
hydrogen-atmosphere furnace without water
quench. The overall annealing cycle was slightly
under 20 min. once the tube entered the furnace, with
time in the hot zone being about 6 min. As with
earlier practice, a 12-in. sample was cut from the
tube, degreased, annealed, pickled, and then visually
and liquid-penetrant inspected before committing

the balance of the tube to these operations. Hardness

measurements were made before and after annealing.
With the procedure outlined above, the quantity of
tubing indicated in Table | was produced from the

43

pilot run with no difficulty. No tendency for cracking
was indicated by the liquid-penetrant inspections.
The hardness before and after annealing was
typically R. 25-35 after cold working and Rg 7881
after annealing.

A final annealing temperature of 1450 K (1177°C)
has been specified for mill products of the alloy,
including tubing. In the production of cold drawn bar
and wire, the vendor preceded the 1450 K (1177°C)
anneal with a 1394 K (1121° C) anneal since it was felt
that the prior lower-temperature anneal helped to
minimize fire cracking. To determine if there were
any difference between a direct 1450 K (1177°C) and
a combined 1394 K (1121°C) plus 1450 K (1177°C)
final anneal for the tubing, the two lengths of 0.750-
in. OD X 0.072-in. wall tubing were annealed both
ways. No cracking was observed in either case by
liquid-penetrant inspection. These lengths were also
eddy-current inspected at the facilities of the
subcontractor and were compared against an 0.005-
in.-deep transverse notch filed in one of the tubes.
Three indications within a length of about 4 in. were
found in the center of the tube given the direct
1177° C anneal. The “defect” section was cut from the
tube and inspected radiographically against a
penetrameter with an 0.010-in.-diam hole and an
0.010-in.-wide notch. No flaws could be found in the
section. Reinspection with liquid penetrant again
showed nothing. In addition, a Y/-in.-long ring was
cut from each of the annealed tubes and was crushed
flat with no evidence of cracking.

The tubing from the pilot drawing run has been
received at ORNL. Prior to delivery, the various
lengths were hydrostatically tested at 1000 psi; they
passed successfully. Since receipt, additional inspec-
tions have been performed. A 6-in. length of each size
of tubing was cut, split in half longitudinally, and
liquid-penetrant (red-dye) inspected on both OD and
ID; no flaws were found (Fig. 6.1). Metallographic
examination of samples of the 0.750-in. OD X 0.072-
in. wall tubing which were given the two annealing

: b
Table 6.1. Tubing produced of 2%
titanium-modified Hastelloy N
during pilot run

Quantity
Size Number of Total length
pieces (ft)
0.750-in. OD X 0.072-in. wall 2 22
0.500-in. OD X 0.042-in. wall 2 26
0.375-in. OD X 0.035-in. wall 4 44

44

Y-137631

Fig. 6.1. Specimens of 2% titanium-modified Hastelloy N seamless tubing following liquid penetrant inspection. No defects were

observed.

treatments showed recrystallized structures. All
tubing from the pilot run was ultrasonically inspected
and compared against standards of standard Hastel-
loy N with an 0.004-in.-deep notch. No indications
greater than those from the notched standards were
observed.

Following completion of the pilot drawing run, the
remaining tube hollow stock was committed to the
production run. The stock was tube reduced to 2.0-in.
OD X 0.187-in. wall by the vendor, annealed, then

further tube reduced by the subcontractor to the two
previously mentioned sizes, 1.25-in. OD X 0.095 and
0.083-in. wall. (The lighter wall stock is scheduled to
finish at 1.00-in. OD X 0.065-in. wall by drawing.) At
this stage, inspection of both cold worked (as tube
reduced) and annealed specimens showed the ODs to
be free of flaws, but borescopic examination of the
IDs indicated questionable areas. Subsequent metal-

lographic examination of these areas revealed the
presence of cracks generally between 0.001 and 0.008

in. deep with several between 0.012 and 0.018 in.
deep. These cracks were oriented parallel to the tube
axis, and they varied from about 0.125 to 0.5 in. in
length. A total of ten lengths of each of the 0.083-in.
and 0.095-in. wall stock were involved in this
evaluation.

Since the pilot drawing run had been accomplished
without difficulty, the incidence of ID cracking
during the production run cannot be satisfactorily
explained at present. Several optionsare open for the
completion of the production run: (1) hone the 1Ds of
the tube-reduced hollows to remove defects before
proceeding with the drawing operations or (2)
commit a portion of the lot to final drawing with no
ID conditioning. If the flaws do not propagate and
remain a constant percentage of wall thickness, their
depth at finish should be mostly within the inspection
criterion of 3% of the wall thickness or 0.004 in.,
whichever is greater. If the flaws do propagate, the
balance of the lot will be available for 1D condition-
ing prior to drawing. Option 2 is preferred.

As described in the previous progress report,’ the
second route to obtain seamless tubing involved hot
extrusion of tube shells from billets machined froma
forged bar of heat No. 75421, followed by cold
drawing by an outside source. The extrusions were
made with moderate success but required condition-
ing of the IDs to eliminate hot tears. This problem
was felt to be caused by inadequate glass lubrication
during extrusion. To test this assumption, two more
extrusions were made using MoQj3 as the lubricant
rather than glass. The billets were prepared from the
same heat of the alloy with a molybdenum tube lining
the IDs and a plasma-sprayed molybdenum coating
on the ODs. These billets were extruded at 1250° Cto
tube blanks, 1.625-in. OD X 1.0-in. 1D, or an
extrusion ratio of approximately 10 to 1. After
removal of the residual molybdenum, the surfaces of
the extrusions were found to be significantly
improved. These tube blanks were scheduled into the
redraw activities of the outside source. However,
definitive results were not obtained on the redraw
characteristics of these extrusions since the alternate
approach to obtain tubing was terminated at this
point. All tube processing effort was then directed
toward. the routing of the first vendor and his
subcontractor in view of their success with the pilot
drawing run.

6.1.2 Semiproduction Heats of 2%
Titanium-Modified Hastelloy N
That Contain Niobium

Products of nine heats of niobium-titanium-
modified Hastelloy N have been received from a

45

commercial vendor for a more complete characteri-
zation of alloys within this system. Chemical analyses
are given in Table 6.2. Niobium additions to the 2%
titanium-modified Hastelloy N base are being
studied for enhancing resistance to tellurium embrit-
tlement. Eight of the alloys (heats 76025 through
76032) were 50-1b vacuum induction plus vacuumarc
remelted heats with niobium levels between about 0.5
and 2%. Four of these alloys contain different levels
of the residual elements iron, manganese, and sili-
con, which are important because of their beneficial
effects upon oxidation resistance. The eight heats
were successfully converted to '/,-in.-thick hot-rolled
plate, approximately 30 Ib per alloy, which is
presently being prepared for weldability evaluation.
Wire and rod are being processed at ORNL from
strips cut from the plate to provide material for
weldability, salt corrosion, tellurium compatibility,
and mechanical property tests.

The ninth alloy (heat 76902) was a 2500-1b vacuum
induction plus electroslag remelted heat of nominally
1% niobium—2% titanium-modified Hastelloy N.
Product received included 247 Ib of *;¢-in.~diam bar,
538 1b of 4-in.-diam bar, 566 Ib of I'/1¢-in.~diam bar,
and 610 1b of 4-in.-square bar. Several billets have
been cut from the 4-in.-square bar for hot rolling a
small quantity of /-in.-thick plate for weldability
studies. Wire and mechanical property specimens are
being prepared from the ’h6-in.~diam bar.

6.2 STABILITY OF VARIOUS
MODIFIED HASTELLOY N ALLOYS
IN THE UNIRRADIATED CONDITION

T. K. Roche H.E. McCoy J. C. Feltner

The study of the stability of niobium, titanium, and
aluminum containing modified Hastelloy N with
respect to intermetallic precipitation has continued.
The objective of this study is to define the upper limits
of these elements that can be added to Hastelloy N
and still maintain a reasonable degree of stability.
The stabilities of a number of alloys, including
laboratory, semiproduction, and production heats,
containing varying amounts of the elements under
consideration were determined by hardness, tensile
properties, creep-rupture properties, and microstruc-
tural evaluation. From results reported in the
previous progress repa::art2 for titanium—aluminum-
modified Hastelloy N, it was possible to propose
gamma-prime solvus temperature boundaries at 650
and 704°C as a function of titanium and aluminum
concentrations. During the present report period the
locations of these boundaries have been refined, and
Table 6.2. Chemical analyses of semiproduction heats of 2% titanium-modified Hastelloy N containing niobium®

Concentration (%)

Heat Ni Mo Cr Ti Nb C Si Mn Fe Al P ) Co B w Cu Ta
76025 Balance 11.87 740 188 0.51 0.062 0.05 <005 0.06 0.11 0.008 0.001 0.17 0.001 002 <0.05 <0.05
76026  Balance 1207 721 199 095 0.055 .04 <005 0.06 007 0608 0.002 0.08 <0.001 0.01 <0.05 <0.05
76027  Balance 11.77 728 199 137 0.056 05 <005 006 0.08 0.008 0.001 0.09 <0.001 0.01 <005 <0.05
76028 Balance 11.84 7.13 193 1.81 0060 005 <0.05 0.06 007 0.008 0.001 0.09 <0.001 0.01 <0.05 <0.05
76029  Balance 12.06 722 195 099 0060 0.04 <0.05 341 007 0.008 0.001 0.09 <0.001 0.01 <005 <0.05
76030 Balance 1194 729 198 096 0.061 015 <005 0.18 0.10 0.008 0.001 0.09 0.001 001 <0.05 <0.05
76031  Balance 1198 728 2.01 097 0.057 005 <0.05 0.07 007 0.009 0.002 0.08 <0.001 0.01 <0.05 <0.05
76032 Balance 1209 698 198 099 0.062 0.12 0.18 339 0.07 0.008 0.001 0.09 <0.001 0.01 <0.05 <0.05
76902 Balance 1190 7.17 216 095 0.07 0.07 <005 0.81 0.13 0.008 0.005 <005 <0.001 0.01 <005 <0.05

4Vendor analyses; all heats 50 Ib except 76902, which was 2500 Ib.

ot
the 800°C boundary has been proposed. This has
been possible through the accumulation of additional
creep-rupture, hardness, and tensile data on earlier
alloys and the evaluation of new compositions.
Mechanical property data for pertinent alloys are
presented in Tables 6.3 and 6.4, and these data have
been analyzed as a unit in arriving at the proposed
gamma-prime solvus at 650, 704, and 800°C (Fig.
6.2).

The mechanical property data in Tables 6.3 and 6.4
were grouped according to alloy stability at the
various temperatures. At a given test temperature,
unstable alloys show creep rates between one and
three orders of magnitude less than the creep rates of
stable alloys, and the rupture lives of unstable alloys
are significantly longer than those of stable alloys. In
the case of hardness and tensile data, instability or-
age hardening results in an increase in hardness and
strength and a decrease in ductility relative to values
for annealed material. This study has shown that the
29, titanium-modified Hastelloy N alloy can contain

up to about 2.2% titanium with 0.4% aluminum and
still retain reasonable stability.

6.3 MECHANICAL PROPERTIES OF
MODIFIED HASTELLOY N ALLOYS IN
THE UNIRRADIATED CONDITION

T. K. Roche  J.C. Feltner  B. McNabb

The determination of reference mechanical prop-
erty data has continued for the various heats of
modified Hastelloy N alloys being evaluated. These
data include room- and elevated-temperature tensile
properties and creep-rupture properties in air at 650,
704, and 760°C and serve as a reference for
comparison with the properties of standard and other
modified Hastelloy N alloys both in the unirradiated
and irradiated conditions.

Several of the heats of 2% titanium-modified
Hastelloy N which have been factored into the
program include two production heats (74901 and
75421) and six semiproduction heats (74533, 74534,

Table 6.3. Creep-rupture data for various heats of titanium—aluminum-modified Hastelloy N®
at 650°C, 47.0 x 10> psi, and at 704°C, 35.0 X 103 psi

Tested at 47.0 X 103 psi and 650°C

Tested at 35.0 X 10> psi and 704°C

Categoryb Heat .
number Concentration (%) Minimum Rupture Total Minimum Rupture Total
Ti Al C creep rate life elongation creep rate life elongation
' (%/hr) (hr) (%) (%/hr) (hr) (%)

A 74901 1.80 0.10 0.06 0.0200 395.2 26.87 0.0797 193.2 39.47
A 75421 190 0.12 0.07 0.0475 268.4 26.61 0.1525 146.4 39.35
A 74533 2.17 048 0.05 0.0335 464.6 30.55 0.1225 195.5 41.65
A 74534 2.09 0.53 0.08 0.0600 393.8 42.81 0.1490 171.7 46.13
A 427 240 0.18 0.014 0.0300 86.6 21.69 0.1570 82.0 23.49
A 428 247  0.16 0.064 0.0131 852.3 18.43 0.1470 201.8 60.75
B 429 2.40 0.35 0.017 0.0018 957.2 16.30 0.0910 200.6 22.75
B 430 250 034 0.073 0.0016 3456.9 20.63 0.1380 212.4 55.86
B 476 246 040 0.046 0.0012 2563.3 . 13.18 0.1010 271.7 45.15
B 477 255 041  0.052 0.0014 1364.1€ 5.12¢ 0.0950 272.2 42.76
C 444 390 013  0.077 0.00037% 1964.99 1.16% 0.0110 1093.0 29.98
D 431 250  0.74  0.016 0.0006 2309.0 2.19 0.0002 29383 6.40
D 432 235 069 0.057 0.0002 4587 4€ 1.68¢ 0.0003 36115 - 13.72
D 478 250 0.75 0.055 0.0009 1340.2¢ 1.34° 0.00044 1341.0° 2.38°
D 479 245 073  0.058 0.00065 908.6° 0.98° 0.00096 908.0° 1.46°
D 443 1.86 0.053 0.00057 908.9°¢ 0.61° 0.0019 835.9¢ 6.16°

®Base: Ni—12% Mo—7% Cr.

A — Stable at 650, 704, and 800°C.
B — Unstable at 650°C; stable at 704 and 800°C.
C — Unstable at 650 and 704°C; stable at 800°C.
D — Unstable at 650, 704, and 800°C.

“Test discontinued.
Test in progress.
Table 6.4. Hardness and tensile data for various heats of titanium—aluminum-modified Hastelloy N?

in annealed and aged conditions. Underlined data indicate strengthening
relative to the as-annealed condition.

Hardness

Stress (103 psi)

Concentration (%) Test
Categoryb Heat i Condition at room temperature Ultimate Elongation (%)
number Ti Al C temperature cO) . Yield
(RB) tensile
A 74901 1.80 0.10 0.06 Annealed 1 hr, 1177°C 79.4 RT¢ 114.0 44.2 66.7
Aged 100 hr, 650°C 81.8 RT 116 .8 50.0 559
Aged 100 hr, 800°C 85.1 RT 119.0 51.2 53.7
Annealed 1 hr, 1177°C 650 79.3 29.5 47.3
Aged 100 hr, 650°C 650 83.1 334 47.1
Annealed 1 hr, 1177°C 800 549 28.1 34.3
Aged 100 hr, 800°C 800 53.8 33.3 47.7
A . 74533 2.17 0.48 0.05 Annealed 1 hr, 1177°C 81.0 RT 1259 47.5 59.8
Aged 8 hr, 650°C 82.2 RT 127.5 49.2 56.6
Aged 8 hr, 800°C 80.8 RT 125.1 50.1 56.9
Annealed 1 hr, 1177°C 650 90.1 32.2 43.5
Aged 100 hr, 650°C 650 924 32.7 476
- Annealed 1 hr, 1177°C 800 549 33.4 63.5
Aged 100 hr, 800°C 800 521 33.6 614
A 74534 2.09 0.53 0.08 Annealed 1 hr, 1177°C 89.3 RT 131.1 529 53.8
Aged 100 hr, 650°C 87.2 RT 138.3 64.2 45.0
Aged 100 hr, 800°C 90.0 RT 130.4 564 50.6
Annealed 1 hr, 1177°C 650 96.8 44 3 27.1
Aged 100 hr, 650°C 650 99.8 46.8 30.0
Annealed 1 hr, 1177°C 800 52.5 38.9 71.5
' Aged 100 hr, 800°C 800 52.1 37.8 68.3
A 427 2.40 0.18 0.014 Annealed 1 hr, 1177°C 74.7 RT 105.2 37.0 75.9
Aged 100 hr, 650°C 74 .9 RT 106.1 38.7 724
Aged 100 hr, 800°C 77.4 RT 106.9 41.5 59.5
Annealed 1 hr, 1177°C 650 63.5 23.8 35.7
Aged 100 hr, 650°C 650 69.8 24.5 4573
Annealed 1 hr, 1177°C 800 51.2 22.5 30.0
Aged 100 hr, 800°C 800 50.2 25.0 60.9
A 428 247 0:16 0.064 Annealed 1 hr, 1177°C 82.5 RT 124.5 49.0 56.2
Aged 100 hr, 650°C 84.2 RT 125.3 50.5 52.7
Aged 100 hr, 800°C 84.7 RT 123.7 51.1 48.1

8b
Table 6.4. (continued)

Hardness Stress (10° psi
Heat Concentration (%) 4 Test ress (107 psh)
Category” cat Condition at room temperature Ultimate Elongation (%)
number Ti Al C temperature °C . Yield
0 tensile
(Rp)
B 429 2.40 0.35 0.017 Annealed 1 hr, 1177°C 76.9 RT 103.3 440 71.4
Aged 100 hr, 650°C 89.5 RT 126.5 59.5 60.0
Aged 100 hr, 800°C 79.8 RT 1109 440 63.5
Annealed 1 hr, 1177°C 650 71.7 26.0 42.6
Aged 100 hr, 650°C 621 80.4 48.8 264
Annealed 1 hr, 1177°C 800 523 24.6 334
Aged 100 hr, 800°C 800 55.5 28.7 65.2
B 430 2.50 0.34 0.073 Annealed 1 hr, 1177°C 88.6 RT 125.6 50.0 54.1
Aged 100 hr, 650°C 95.7 RT 134.8 634 43.7
Aged 100 hr, 800°C 88.6 RT 126.8 531 50.0
B 476 2.46 0.40 0.046 Annealed 1 hr, 1177°C 84.9 RT 124.3 474 56.7
Aged 8 hr, 650°C 82.5 RT 122.6 58.8 54.7
Aged 100 hr, 650°C 84.5 RT 122.6 48.3 553
Aged 8 hr, 704°C 84.7 RT 122.0 47.6 49.5
Aged 100 hr, 704°C 86.6 RT 121.6 493 62.5
Aged 8 hr, 800°C 86.9 RT 123.0 49.3 59.2
Aged 100 hr, 800°C 86.4 RT 122.6 50.2 61.3
B 477 2.55 041 0.052 Annealed 1 hr, 1177°C 86.3 RT 124.9 475 59.0
Aged 100 hr, 650°C 87.0 RT 124.6 574 61.0
Aged 100 hr, 704°C 87.1 RT 124.6 50.6 594
Aged 100 hr, 800°C 88.0 RT 124.3 55.0 59.0
C 444 3.90 0.13 0.077 Annealed 1 hr, 1177°C 87.8 RT 1274 49.0 62.7
Aged 100 hr, 650°C 95.2 RT 138.3 65.7 504
Aged 100 hr, 704°C 87.2 RT 1274 51.2 61.6
Aged 100 hr, 800°C 87.1 RT 126.6 52.2 58.9
C 447 3.50 0.15 0.043 Annealed 1 hr, 1177°C 86.6 RT 1284 49.6 66.4
Aged 100 hr, 650°C 99.6 RT 155.6 85.1 41.8
Aged 100 hr, 704°C 96.1 RT 142.3 64.3 45.0
Aged 100 hr, 800°C 88.1 RT 126.7 52.5 60.6
b 431 2.50 0.74 0.016 Annealed 1 hr, 1177°C 78.6 RT 1104 40.5 74 .4
Aged 100 hr, 650°C 90.9 RT 128.4 57.1 60.8
Aged 100 hr, 800°C 91.9 RT 129.6 68.3 47.7
D 432 2.35 0.69 0.057 Annealed 1 hr, 1177°C 87.4 RT 124.8 49.2 59.6
Aged 100 hr, 650°C 98.8 RT 148.3 79.3 52.6
Aged 100 hr, 800°C 98.2 RT 142.5 77.1 47.8

6b
Table 6.4. (continued)

3 .
C B Heat Concentration (%) L l-::ri::;;s Test Stress (10” psi) )
ategory number - Al c Condition temperature temperature Ultimate Vield Elongation (%)
(Rp) 0 tensile

D 478 2.50 0.75 0.055 Annealed 1 hr, 1177°C 87.8 RT 126.6 48.7 54.6
: Aged 100 hr, 650°C 95.7 RT 142.8 1.5 4.1
Aged 100 hr, 704°C 978 RT 148.8 795 46.9
Aged 100 hr, 800°C 93.0 RT 132.0 61.8 57.6
D 479 2.45 0.73 0.058 Annealed 1 hr, 1177°C 88.7 RT 127.3 48.6 62.3
' Aged 100 hr, 650°C 96.7 RT 146.2 738 50.7
Aged 100 hr, 704°C 99.3 RT 1504 80.1 46.9
: , Aged 100 hr, 800°C 96.1 RT 133.0 62.1 54.6

D 443 1.86 0.053 Annealed 1 hr, 1177°C 87.5 RT 123.4 48.9 0.
Aged 100 hr, 650°C 98.1 RT 146.9 173 49.6
Aged 100 hr, 704°C 98.7 RT 148.8 80.5 4438
Aged 100 hr, 800°C 95.0 RT 133.0 659 494
D 445 1.2 1.06 0.049 Annealed 1 hr, 1177°C 84.0 RT 121.0 46.3 69.6
Aged 100 hr, 650°C 90.3 RT 138.5 68.3 578
Aged 100 hr, 704°C 95.0 RT 139.5 723 500
Aged 100 hr, 800°C 90.6 RT 124.3 58.8 570
D 474 4.16 0.18 0.050 Annealed 1 hr, 1177°C 904 RT 134.0 539 59.0
Aged 100 hr, 650°C 28.6 R, RT 180.7 99.7 294
Aged 100 hr, 704°C 299 R, RT 181.4 90.6 343
Aged 100 hr, 800°C 931 RT 141.8 60.0 53.0

?Base: Ni—12% Mo—7% Cr.
A — stable at 650, 704, and 800°C.
B — Unstable at 650°C; stable at 704 and 800°C.
C — Unstable at 650 and 704°C; stable at 800°C.
D — Unstable at 650, 704, and 800°C.
°RT is room temperature, approximately 25°C.

0s
51

ORNL —DWG 76-10097

4.5
\o 474
4.0 —X
® 444
35 +—e 447
Y
2
— ® 480
=z
w
-
=
o
(W
s
2
b4
=
= ® 475
i ® 901
650°C
1.5
NO -rr ® 445
1.0
Q.5
443 446 473
0 *o—@ ®
O 0.5 1.0 1.5 2.0 2.5 30

ALUMINUM CONTENT (%)

Fig. 6.2. Proposed boundaries separating stable from unstable alloys of Ni-12% Mo-7% Cr +Al and Ti with respect to gamma prime
precipitation at 650, 704, and 800° C. Alloys above the lines will form gamma prime, and those below will not. (The numbers by the individual

data points refer to alloy numbers.)

74535, 74539, 74557, and 74558) produced by an
outside vendor. Four of the six semiproduction heats
contain small additions of rare earth elements,
lanthanum, cerium, and misch metal. The composi-
tions of these alloys were chosen to study the
effectiveness of rare-earth-element additions for
minimizing the extent of shallow intergranular
cracking. The processing history and chemical
analysis of these alloys have already been reported.3

During this report period, creep-rupture tests were
completed for heat 75421 at three stress levels for
each of three test temperatures. The specimens were
prepared froma Y4-in.-diam bar of the alloy and were

given a pretest anneal at 1177°C. In Table 6.5 a
comparison is made between the rupture lives and
elongations for this heatand the range of these values
determined for the other heats of this series. At 650°C
the rupture behavior of heat 75421 falls on the lower
side of the range. At 704 and 760°C the rupture

“behavior of this heat is more within the established

range. In all cases the 2% titanium-modified alloys
exhibit longer rupture times than standard Hastelloy
N tested under the same conditions.

Tensile properties at room and elevated tempera-
ture for several of the heats in the annealed condition,
1 hr at 1177°C, are presented in Table 6.6. Included
52

Table 6.5. Comparison of stress-tupture data at 650, 704, and
760°C for various heats of 2% titanium-modified Hastelloy N

Heat 75421 Range?
Test Stress Rupture Rupt
temperature 3 Rupture upture Rupture upture
(107 psi) P elongation P elongation
(°C) : g : g
life (hr) (%) hf_e (hr) %)
650 40.0 893 32 1120-1883 23-42
47.0 268 30 302-591 20-43
55.0 68 29 70-166 18-51
704 25.0 1164 48 873-1518 3647
30.0: 361 48 308 -498 37-47
35.0 146 41 172-207 39-63
760 15.0 537 32 544--1555 25-54
20.0 301 47 211-416 42-58
25.0 88 54 79-123 47-57

“Range for 74533, 74534, 74535, 74539, 74557, 74558, and 74901.

with these data are the properties for two earlier heats
(heats 471-114 and 471-583) of 2% titanium-modified
Hastelloy N and typical properties for standard
Hastelloy N. From the tabulation, it can be seen for
any one alloy that increasing temperature within the
temperature range of 650 to 800°C has a greater
effect in reducing the ultimate tensile strength than
the yield strength. In fact, the yield strength of any
one alloy remains relatively constant in this temper-
ature range, but there are heat-to-heat variations
which result in a spread from 40.0 to 29.0 X 10’ psi.
The data are presented in the general order of
decreasing yield strength of the various heats at
elevated temperature and there is an apparent
correlation between yield strength and titanium
content in the range of 1.8 to 2.1%. Note that the two
lanthanum-containing alloys (heats 474-534 and 474-
558) show the highest yield strengths. In making this
correlation, other variables which also influence
tensile properties have not been considered.

All the alloys show a decrease in ductility (rupture
elongation) at elevated temperature relative to the
value at room temperature, a characteristic of nickel-
base alloys. The magnitude of this decrease and the
rate of recovery as temperature increases vary from
heat to heat. As with creep-rupture properties, the
tensile properties of the 2% titanium-modified alloys
are higher than those of the standard alloy.

Niobium-modified Hastelloy N seems to have
good resistance to tellurium embrittlement, and a
new series of laboratory alloys based upon Ni-12%
Mo-7% Cr-0.2% Mn-0.06% C containing between 0
and 1% niobium in 0.2%  increments was prepared for
a closer evaluation of radiation damage resistance.

Creep-rupture tests on unirradiated specimens are
under way in air at 650° C to establish baseline data.
Limited results obtained thus far show that these
alloys have shorter rupture lives at 650°C than 2%
titanium-modified Hastelloy N.

6.4. POSTIRRADIATION CREEP
PROPERTIES OF MODIFIED
HASTELLOY N

H. E. McCoy T. K. Roche

Tests have been made on a number of alloys, but
the main emphasis during this report period was on
alloys modified with niobium. Prior tests on these
alloys measured the properties with the material heat
treated to produce a fine grain size. We later learned
that such anneals did not produce optimum proper-
ties in Hastelloy N, but our work had shown that the
addition of about 2% titanium resulted in excellent
resistance to irradiation embrittlement; so we did not
continue investigation of the niobium-modified
alloys. Recent observations concerning the excellent
resistance of the niobium-modified alloys to embrit-
tlement by tellurium offered added incentive to more
fully investigate the irradiation embrittlement of the
niobium-modified alloys.

Experiment ORR-233 was run for the explicit
purpose of irradiating alloys currently on hand which
contained niobium.* The individual tensile speci-
mens were mounted in clusters of three inside a small
heater (Figs. 6.3 and 6.4).. The temperature was

*Both the equipment assembly and the irradiation were done by
J. W. Woods and C. K. Thomas, Metals and Ceramics Division.
53

Table 6.6. Room- and elevated-temperature tensile properties of various
heats of 2% titanium-modified and standard Hastelloy N*

Test Ultlm.ate Yield . Reduction
b tensile Elongation X
Heat number tempoerature stress st:raess. (%) in area
0 (10% psi) (107 psi) (%)
474-534 25 131.1 52.9 53.8 59.5
650 96.8 44.3 271 32.2
704 87.8 45.5 52.0 46.2
760 68.8 39.3 717.0 64.4
800 525 38.9 71.5 67.2
474-558 25 126.4 521 54.3 58.6
650 87.3 38.8 28.0 30.5
704 81.5 38.3 25.8 30.0
760 65.8 38.7 36.0 32.7
800 531 374 47.8 43.8
474-557 25 1304 51.1 579 60.3
650 87.5 37.3 28.7 28.6
704 83.8 36.7 23.5 27.5
760 72.2 37.1 351 31.6
800 56.9 36.2 46.2 40.4
474-533 25 125.9 47.5 . 59.8 614
650 90.1 . 322 43.5 40.4
704 87.0 3211 38.7 30.8
760 68.6 31.8 54.6 404
800 549 334 63.5 53.2
471-583 25 121.8 459 73.3 60.1
650 859 30.2 49.0 384
760 64.4 294 45.0 47.1
800 52.2 289 47.0 15.0
471-114 25 120.7 44 8 72.6 58.9
' ‘ 650 86.9 29.1 56.2 39.6
704 82.3 28.6 534 40.5
760 64.3 28.7 52.8 504
800 50.6 29.7 51.0 554
674-901 25 114.0 44.2 66.7 554
650 79.3 29.5 47.3 43.4
704 71.9 279 33.5 36.8
760 66.9 29.2 31.2 29.7
800 54.9 28.1 34.3 29.5
Standard 25 114.4 447 50.0
Hastelloy N¢ 650 82.4 27.5 37.0
704 69.9 28.0 24.0
760 61.8 26.2 21.0

%Annealed 1 hr at 1177°C prior to testing. Tested at a strain rate of 0.44 min -

bgee Table 6.21 for detailed chemical compositions.
“Typical data for 0.045-in.-thick sheet annealed at 1150°C.
measured by a thermocouple attached to the center of
the cluster, and the signal from this thermocouple
was also used for temperature control. The assembly
of 102 specimens was welded in an aluminum can for
insertion into the Oak Ridge Research Reactor
(ORR) for irradiation. The specimens were
perpendicular to the face of the reactor where the
peak thermal flux was 1.5 % 10" neutrons cm ™ sec™".
After irradiation for 908 hr to thermal fluences

54

ranging from 2 to 5 X 10” neutrons/cm® and at
temperatures of 650, 704, and 760°C, the assembly
was removed from the ORR. The individual
specimens were recovered in a hot cell and trans-
ferred to another cell for creep testing.

The partial chemical compositions of the alloys in
this experiment are given in Table 6.7. The alloys
contained from 0.45 to 2% niobium and various
combinations of niobiumand titanium. The postirra-

PHOTO 2955-75

Fig. 6.3. ORR-233 irmadiation experiment prior to welding on protective aluminum can. The experiment contains 102 small
creep specimens with thermocouples and heaters for measuring and controlling the temperature. The various heater and
thermocouple lead wires are routed through the access tube to the control room

55

Komand
.

5

AY

—

|
|
|
!
l
I
|
!
f

Fig.
small furna

diation creep properties measured to date are
summarized in Table 6.8. The stress-rupture proper-
ties at 650°C of several of the alloys are shown
graphically in Fig. 6.5. All specimens irradiated at
650° C had rupture lives in excess of those of standard
Hastelloy N in the unirradiated condition. The
fracture strains were quite high and generally fell in
the 10 to 20% range. Irradiation at 704° C caused a
pronounced deterioration of the properties of the

 PHOTO 295775

A closeup of the clusters of creep specimens in ORR-233. The specimens are bound into clusters of three and are centered ina
A thermocouple is attached to the center of the cluster.

alloys modified only with niobium. The alloys
containing 1 to 2% niobium were affected less, and
their properties are still acceptable after irradiation at
this condition. After irradiation at 760°C, only the
alloy containing 2% niobium retained acceptable
properties.

Alloys 303 and 413 (Fig. 6.5), which contained
niobium and titanium additions, had excellent
postirradiation properties after irradiation at 650,
704, and 760° C. However, these alloys do not have
good resistance to embrittlement by tellurium.

The results from three of the niobium-modified
alloys are shown in more detail in Figs. 6.6 and 6.7.
The stress-rupture properties in Fig. 6.6 show thatall
three alloys had good rupture properties after
Jirradiation at 650° C, that alloys containing 1 and 2%
niobium retained acceptable properties after irradia-

Table 6.7. Compositions of alloys

irradiated in ORR-233
Alloy Composition? (wt %)

designation Nb Ti Other

345 0.45 0.22 Si

348 0.62 0.47 Si
21543 0.7

295 0.85

411 1.0°

237 1.07

298 2.04

303 0.84 0.49

304 1.6° 0.8°

305 1.3 0.88
470-786 0.62 0.82

425 0.48 1.9

413 1.0 1.0°

aAlloys nominally are nickel base and contain
12% Mo, 7% Cr, and 0.05% C.
Nominal values.

56

tion at 704° C, and that only the alloy containing 2%
niobium retained good properties after irradiation at
760°C.

The minimum creep rates for these three alloysare
shown in Fig. 6.7. One problem in this comparison is
that the specimens tested in the unirradiated
condition were simply annealed 1 hrat 1177°Cbefore
testing, but the irradiated specimens were annealed 1 .
hr at 1177°C and then held at the indicated
temperature for 908 hr during irradiation before
creep testing. The latter treatment precipitates
carbides and generally increases the creep rate by
removing alloying elements from solution. In the
unirradiated condition, all three alloys have about
equivalent creep rates and are stronger (have lower
creep rates) than standard Hastelloy N. The limited
data on alloy 21543 (0.7% niobium) show that the
creep rate was increased by irradiation and that the
increase was greater as the irradiation temperature
was increased. Alloy 237 (1% niobium) exhibited a
slight increase in creep rate due to irradiation at 650
and 704° C, but a greater increase when irradiated at
760° C. The creep rate of alloy 298 (2% niobium) was
not altered detectably by irradiation at any of the
three temperatures.

Alloys 345 and 348 contained additions of niobium
and silicon and the resistance of these alloys to
tellurium embrittlement was excellent. However,
these alloys had very poor postirradiation properties

ORNL-DWG 76-6307

70 : :
[ T TTIHI T I‘NI\!STANDARD HASTELLOY N 1T [ TTTH
<UNIRRADIATED
60
STANDARD HASTELLOY N
50 IRRADIATED
~<{05-1.0) BUBSN, FUB  EUT) GL25)
z o~ l FUISh i) Iz |
" B(4) C(9) D(6) AlIR2IB(IQ) E(16)
» 40 i Fus)f’%-—fl%)
= ! ~ £) 3(3 G(24) ¢ g
@« ):(A(’3) FSIM HeC@) | B30 Xe GU7)
W ALLOY  COMPOSITION % __ iy 1~ T E?’; Ro——C"2
v ; c(a) '8{2) DS
21543 0.7 A ~
20 295 085 B ‘ | =N
414 1 ¢ © §50°C IRRADIATED
237 ) D e 704°C IRRADIATED
10 —— 298 2 E H 760°C IRRADIATED
303 084 049 F () FRACTURE STRAIN
413 { 1 6 .
o I NI N BRI Rl LI
10 10° 10' 10 10° T

RUPTURE LIFE (hr)

Fig. 6:5. Postirradiation creep properties at 650° C of several alloys irradiated at the indicated temperatures. These alloys are nickel base
and contain 12% Mo, 7% Cr, and 0.05% C in addition to the elements listed.

57

Table 6.8. Postirradiation creep properties at 650°C of several

compositions of modified Hastelloy”

Irradiation Minimum . Rupture Loading Total
Test Stress creep . . .
Alloy number temperature a0’ psi) rate life strain strain
cO (%/hi) (hr) (%) (%)
345 R 2047 650 40.0 0.030 39.7 1.72 3.36
704 40.0 10.8 1.96 2.82
R 2068 704 300. 0.0029 75.1 0.20 0.61
760 35.0 2.0 3.15 3.97
348 R 2048 650 40.0 0.021 71.6 1.96 4.12
704 40.0 8.6 1.85 4.08
R 2069 704 30.0 0.0070 74.7 0.84 1.36
760 35.0 12.4 3.00 4.13
21543 R 2033 650 40.0 0.021 292.0 4.85 12.03
R 2064 650 35.0 0.011 558.7 1.84 10.11
R 2039 704 40.0 0.046 24.8 4.05 5.63
704 30.0
760 35.0 7.1 2.30 3.03
760 30.0 34.3 1.13 195 -
295 R 2059 650 47.0 0.069 - 77.2 4.58 13.24
R 2029 650 40.0 0.015 461.3 2.36 9.56
R 2035 704 40.0 0.035 60.5 1.11 3.54
R 2066 704 35.0 0.0063 191.9 0.29 3.36
R 2041 760 35.0 0.0080 419 047 1.91
"R2073 760 30.0 0.0083 110.7 0.29 2.21
411 R 2031 650 40.0 0.018 346.1 4.82 15.42
650 35.0 1101.9 2.99 12.36
R 2037 704 40.0 0.027 105.4 4.81 9.15
R 2070 704 35.0 0.023 55.0 2.49 4.13
R 2043 760 35.0 0.0095 415.2 2.64 7.76
237 R 2028 650 40.0 0.016 206.0 1.24 5.94
R 2058 650 35.0 0.0051 488.1 0.35 4.82
R 2034 . 104 40.0 0.019 65.3 2.02 393
R 2065 704 35.0 0.0027 365.5 1.32 3.16
R 2040 760 350 0.014 37.9 1.77 2.49
R 2072 760 30.0 0.0055 46.0 0.86 1.32
298 R 2060 650 47.0 0.028 413.1 2.56 16.74
R 2030 650 40.0 0.011 1031.7 1.17 15.52
R 2036 704 40.0 0.011 186.6 1.08 4.44
R 2067 704 35.0 990.0 0.25 7.82
R 2042 760 35.0 0.0014 565.6 0.5 n
760 30.0
303 650 47.0 298.6 3.22 233
650 40.0 12849 1.68 23.3
704 47.0 184.1 292 17.6
R 2049 704 40.0 0.019 473.3 1.04 15.4
760 40.0 o '
R 2052 760 35.0 0.0062 316.4 0.30 5.15
304 R 2075 650 47.0 0.12 30.8 2.61 7.06
650 40.0 67.6 0.70 3.46
R 2096 704 40.0 16.6 0.65 1.98
704 35.0
760 35.0 12.8 0.90 1.64
760 30.0
58

Table 6.8. (continued)

T | Irradiation Stress Mlcmézum Rupture Loading Total
Alloy nur:j:er temperature a Oar e;si) rratep life strain strain
€0 (%/hr) (he) ) )
305 R 2093 650 55.0 82.0 4.57 19.2
650 47.0 400.2 1.52 13.8
704 55.0
704 47.0
760 35.0
470786 R 2095 650 47.0 138.3 3.96 20.9
R 2055 650 40.0 0.0083 755.4 1.18 142
704 35.0 348.5 1.00 7.19
R 2056 - 760 35.0 0.0077 446.6 1.77 7.5
425 R 2094 650 55.0 123.9 6.00 17.6
650 47.0 871.3 2.95 16.7
704 47.0
760 40.0
760 350
413 650 47.0 725.3 3.25 ‘249
R 2032 - 650 40.0 0.015 973.2 1.85 24,1
760 35.0 1308.4 1.28 16.8

2All specimens annealed 1 hr at 1177°C prior to irradiation. Irradiated at the indicated temperature for 908 hr to

thermal fluences in the range of 2 to 5 X 1020 neutrons/cm>.

ORNL-DWG 76-6308

70 T T 1T I—I""I\Q T T T T T T T7T T TTTT
STANDARD HASTELLOY N
60 UNIRRADIATED
]
50 < 3
— Y
- ~ N\
2 a0 =~ PV !
(] A
2 : g 3
@ © 21543 (0.7 % Nb) \'f\
W 30 [~ 4237 (1.0% Nb) "“\...\
K D 298 (2.0 % Nb) \\L
OPEN-UNIRRADIATED STANDARD HASTELLOY N .
20 I~ FILLED-IRRADIATED — IRRADIATED —— =
1- IRRADIATED AT 650°C
'O ™ 2- |RRADIATED AT 704°C
3- IRRADIATED AT 760°C
ol vl vl vl e ol
107" 10° 10 102 103

RUPTURE LIFE (hr)

104

Fig. 6.6. Stress-rupture properties at 650°C of three niobium-modified alloys after irradiation at the indicated temperature.
59

ORNL-DWG 76-6309

70 T T TTTIT T TTTTT T TT7TT0 T T T T T T 17T
’/
o L © 21543 (0.7% Nb) © L
4 237 {1.0% Nb) //
O 298 (2.0 % Nb) oo a i
5o | OPEN-UNIRRADIATED el
FILLED- IRRADIATED a & “1
- ///
a 12 1
£ a0 s’y
= 2 3 L oAl
A 37
~
W 30 ~“a
i e
® |~ sTANDARD HASTELLOY N
20 IRRADIATED AND UNIRRADIATED
1- IRRADIATED AT 650°C
{0 | 2-IRRADIATED AT 704°C
3- IRRADIATED AT 760°C
0 L LT [N R R LTIl
107* 107° 1072 o 10° to'

MINIMUM CREEP RATE (%/hr)

Fig. 6.7. Minimum creep rates at 650°C of three niobium-modified alloys after irradiation at the indicated temperature.

(Table 6.8). This result was anticipated because of the
role of silicon in promoting the formation of coarse
M;C-type carbides rather than the finely dispersed
MC-type carbides formed in the absence of silicon. It
is not known whether the good resistance of these
alloys to tellurtum embrittlement was from the
presence of niobium or silicon or both, but these
postirradiation results show clearly that silicon levels
in the range of 0.2 to 0.5 cannot be tolerated because
of the effects on the mechanical properties.

These postirradiation creep data have shown that
the niobium-modified alloys have adequate resist-
ance to embrittlement by neutron irradiation. Some
further work is needed to optimize the niobium level
and to define the maximum allowable service
temperature.

6.5 MICROSTRUCTURAL ANALYSIS OF
MODIFIED HASTELLOY N

D. N. Braski D. J. Bradley
J. M. Leitnaker G. A. Potter

Two largely independent studies are in progress
concerning the microstructure of nickel-base alloys.
One study 1s funded by the Molten-Salt Reactor
Program and involves alloys of direct Program
interest. In the other study, which is beingdone by an
ORAU student, simpler alloys are being investigated
to obtain more basic data. However, both programs

contribute to our understanding of nickel-base
alloys, and research data from both programs are
included in this report.

6.5.1 Production of 2%
Titanium-Hastelloy N Alloys with
Uniform Carbide Distribution

D. N. Braski J. M. Leitnaker G. A. Potter

Introduction. Improved properties, especially
those related to irradiation resistance, have been
achieved by adding about 2 wt % titanium to
Hastelloy N.* However, this new alloy demonstrated
large, and as yet unexplained, variations in postirra-
diation creep properties.” Metallographic examina-
tion showed that the microstructures of the irradiated
samples were quite inhomogeneous, The purpose of
this investigation was to develop metallurgical
techniques that could be used to produce a homo-
geneous microstructure in 2% titanium-Hastelloy N.
Specifically, the goal was to remove the carbide
stringers from the alloy and to redistribute the
carbides uniformly throughout the microstructure.
With this accomplished, it was hoped that both the
mechanical properties and corrosion resistance of the
alloy could be simultaneously optimized. Further-
more, it should be expected that the principles
developed in reaching this goal would be generally
applicable to most structural alloys.
The “memory effect,” that is, how carbides
reprecipitate into stringers after normal solution
annealing and aging treatments, will be discussed
first, after which the results of various experiments
aimed at producing uniform carbide distribution will
be presented and discussed.

The Memory Effect. The microstructure of
titanium-modified Hastelloy N consists of two
phases: (1) a face-centered-cubic, nickel-rich matrix
and (2) face-centered-cubic, MC-type carbides,
where M is a combination of primarily molybdenum,
titanium, chromium, and iron.’ Invariably, the
carbide particles lie in stringers or lines parallel to the
principal working direction, as illustrated in Fig.
6.8a, which shows a polished and etched longitudinal
section of swaged 0.25-in.-diam rod from alloy 452.
The chemical composition of alloy 452 (Table 6.9)
shows that this alloy contains 0.048 wt % carbon—an
amount sufficient to provide good strength proper-
ties to Hastelloy N. When samples of alloy 452 were
solution-annealed for 1 hr at 1177° (a standard
annealing treatment used for Hastelloy N), most of
the carbides were dissolved into the matrix (Fig.
6.8b). Recrystallization and growth of the cold-
worked microstructure have also taken place, as
evidenced by the larger, equi-axed grains. The grain
boundaries were nonlinear—probably caused by the
pining action of remaining carbides during the grain
growth process. When the samples were aged for 160
hr at 760°C, carbides precipitated out in the matrix
(Fig. 6.8¢). Precipitation of carbides occurred
predominately in areas where they existed before the
solution anneal at 1177° C and hence demonstrated
what is called a “memory effect.”

A clearer picture of what happened in the
microstructure of alloy 452 during aging is given in
Fig. 6.9. This transmission electron micrograph was
taken from the same sample shown in Fig. 6.8c and

Table 6.9. Alloy compositions (wt %) -

Alloy number

Element

450 451 452 453
Mo 13.6 13.5 13.5 13.2
Cr 8.1 8.2 7.7 7.2
Ti 2.1 1.9 1.8 2.0
C 0.052 0.017 0.048 0.035
Mn <0.1 <0.1 <0.1 <0.1
Nb <0.1 <0.1 <0.1 <0.1
Fe ~0.3 ~0.3 ~0.3 ~0.3
Ni Balance Balance Balance Balance

60

shows that fine MC-type carbides precipitate outina
stacking fault morphology® along (111) planes. The
large “primary” carbide (one that did not dissolve
during the solution anneal) has a geometric shape
and may have punched out prismatic dislocations
which initiate the growth of stacking fault precipitate
ribbons. However, it will be shown later that primary
carbides are not required to produce stacking fault
ribbons in 2% titanium Hastelloy N. Also, precipita-
tion did not always occur in ribbons. Frequently
carbides precipitated on dislocations lying along the
(111) planes and formed thin, separate platelets
instead of ribbons (Fig. 6.10). Analyses by both
selected-area electron diffractions of thin foils and x-
ray diffraction of electrochemically extracted precip-
itates showed that all the carbides, regardless of
morphology, were of the fcc, MC type.

The main point in defining and trying to under-
stand the memory efféct in titanium-modified
Hastelloy N is that this effect must be overcome in
order to fabricate an alloy without carbide stringers.
As long as the memory effect was occurring, it was
virtually impossible to get rid of carbide stringers.
The general approach to eliminating the memory

- offect was to ensure that the carbides were dissolved

and in solution during the fabrication process. This
could be carried out in two ways: (1) lower the
carbon content of the alloy and continue to use the
same fabrication schedule or (2) maintain a carbon
content of ~0.05 wt % but adjust the fabrication
parameters so that the carbides remain in solution
during fabrication. Both of these approaches were
investigated.

Influence of Carbon Content. The first step in
adjusting the carbon content of the alloy was to
estimate the solubility of carbon in 2% titanium-
modified Hastelloy N at 1177° C. Using precipitation
extraction data from aged samples, the amount of
carbon soluble in the alloy at 1125°C (~50° C below
the desired annealing temperature) was calculated to
be 0.045 wt % carbon. Thus a 2% titanium-modified
Hastelloy-N alloy with less than 0.045 wt % carbon
should be free of carbide particles at 1177°C. Alloys
451 and 453 were prepared with carbon contents of
0.017 and 0.035 wt 9% respectively. They were
fabricated into 's-in.-diam rod according to the
fabrication schedule (Table 6.10). With the lower
carbon content alloys, the solution anneal conducted
at 1177° C after hot swaging was expected to dissolve
the carbides. With the carbides in solution, no
stringers were expected to form during the remaining
cold swaging steps.
Y-134993

Y-136991
= .

19

Fig. 6.8. Optical photomicrographs illustrating the “memory
effect” in alloy 452. (a) As swaged, (b)after | hrat 1177°C, (¢) after
I hr at 1177°C plus 160 hr at 760° C. Electrolytically etched with

oxalic acid.
Fig. 6.9. Transmission electron micrograph showing stacking fault precipitates in alloy 452after annealing 1 hrat 1177°Cand 160 hrat

760°C.
Table 6.10. Fabrication schedule for 2% titanium-Hastelloy N alloys 451, 452, 450
Step Fabrication Diametey  Roduction Alloy 451 Alloy 452 Alloy 450
number process (in) ‘"(;f;"’ 0.017% C) (0.048% C) (0.051% C)
1 Arc-drop cast 1.00
2 Hot swage at 1177°C 15 min reheat 30 min reheat  Can in 316 stainless steel;
Pass 1 0.87 2.3 time between time between 15 min reheat time be-
Pass 2 0.74 21.7 passes passes tween passes
Pass 3 0.64 25.2
Pass 4 0.54 28.8
Pass 5 0.49 17.7
Pass 6 0.43 229
3 Homogenizing anneal 1hrat1177°C,  2hrat1177°C, 1 hrat 1260°C, air cool,
air cool air cool 1 hr at 1300°C, water
quench
4 Cold swage at 25°C 0.34 37.5
s Intermediate anneal 1hrat1177°C 1hrat1177°C 1S minat 1177°C
6 Cold swage at 25°C 0.25 45.9

63

Fig. 6.10. Transmission electron micrograph showing thin MC-type carbide platelets in alloy 452 after annealing | hrat 1177° Cand 160

hr at 760°C.

Unfortunately, both alloys contained carbide
stringers after fabrication was complete, as demon-
strated by the microstructure of alloy 451 (0.017 wt %
carbon) in the as-swaged condition (Fig. 6.11a).
However, when a sample of alloy 451 was givena |-hr
solution anneal at 1177° C, almost all of the carbides
were dissolved (Fig. 6.11h). Moreover, aging for 223
hrat 760° produced no carbide stringers (Fig. 6.11¢).
Instead, tiny carbide particles precipitated out on
grain boundaries, twin boundaries, and preferen-
tially on crystallographic planes. These details were
more apparent with the electron microscope. Figure
6.12 shows MC-type carbides in a grain boundary of
the aged sample of alloy 451. Figure 6.13 shows small
MC platelets located near the center of a grain in the
sample. Trace analysis showed that the plateletsall lie
on (111) planes. Note_that one of the thin MC
platelets also lies in the (111) foil plane. These results

show that the memory effect must have been present
during fabrication of alloy 451 but was eliminated
when small samples of the same heat were solution-
annealed for | hrat 1177°C after swaging. Assuming
that the fabrication furnaces were at the correct
temperatures, the solution annealing time necessary
to dissolve the carbides must be sensitive to the mass
of material being treated. Intuitively, this is reason-
able and has actually been observed repeatedly
throughout the investigation. The results for alloy
453 were similar to those for alloy 451 except thata
final I hrat 1177° Cdid not remove the memory effect
from that alloy. Higher annealing temperatures or
longer times at 1177°C would be required to produce
a uniform carbide distribution in this 0.035-wt %
carbon alloy.

The rate of carbide dissolution was determined
quantitatively for bothalloys452and 451 by solution

X

¥9

20 40 60 MICRONS 120 140
Ly—L 1 —s500%—l; T
0.00¢ INCHES 0.005

Fig. 6.11. Optical photomicrographs of alloy 451. (a) As-
swaged. (b) After annealing | hrat 1177°C. (c) Afterannealing | hr
at 1177°C followed by 223 hrat 760° C. Electrolytically etched with
oxalic acid.

65

YE-11345

Fig. 6.12. Transmission electron micrograph showing fine MC-type carbide in grain boundary of alloy 451 after annealing 1 hr at
1177°C plus 223 hr at 760°C.

YE-11344

Fig. 6.13. Transmission electron micrograph showing platelets of MC lying on (111) planes in alloy 451 after annealing for 1 hr at
1177°C plus 223 hr at 760°C. Plane of foil = (I11).

annealing at 1177°C at different times and then using
the electrochemical extraction techniques to measure
the quantity of undissolved carbide left in the
microstructure (Fig. 6.14). Small samples of both
alloys were annealed at 1177°C for times up to 8 hrin
an argon atmosphere and were then analyzed for the
amounts of carbides remaining. In the case of alloy
451, at least 1 hr was needed to reach what appears to
be an equilibrium amount of precipitate remaining in
the alloy. For times greater than about 1 hr,
essentially all of the carbides were dissolved into the
matrix. However, small amounts of carbide precipi-
tate were extracted at even the longest annealing
time. It is possible that some precipitation (<0.1 wt
%) can occur upon cooling the sample after solution
annealing. There 1s also the possibility that a very
small amount of other insoluble inclusions present in
the alloy were being collected.

Another important point concerning the removal
of the memory effect in 2% titanium Hastelloy N is
that once the carbides dissolve, additional time is
required for the carbide-forming elements to diffuse
away from stringer areas. If the elements are not
given sufficient time to disperse, they will reprecipi-
tate in a stringer morphology. The relation of time
and carbide dissolution is clearly illustrated by the
electrochemical extraction data on annealed samples
of alloys 451 and 452 (Fig. 6.14). The data for alloy
452 show that its higher carbon content resulted in
greater amounts of carbides in the microstructure

ORNL-DWG 76-4734

03
fa
\\
2 02 X
= 1T \&
o AN
o N\
& N
* N 452
R S 4
; 04 ° —l —
. k R
J\. ® &
\ ,/4i51 |
0! [ ] d | I
0 2 4 6 8

TIME AT 1477 °C (hr}

Fig. 6.14. Results of electrochemical extractions on samples of
alloys 451 and 452 annealed for various times at 1177°C,

66

thanforalloy451 for allannealing times investigated.
Its higher carbon content also made carbide
resolution sufficiently sluggish to enable the memory
effect to operate for annealing times up to at least 8
hr. This meant that longer times at 1177° C or higher
temperatures would be needed to eliminate the
memory effect from alloy 452.

Influence of Fabrication Parameters. The previous
section indicated that 2% titanium-Hastelloy N alloys
with less than ~0.035 wt 9% carbon could be
fabricated by the existing shop fabrication schedule
shown in Table 6.10 for alloy 451. Usually, however,
carbon contents from 0.050 to 0.060 wt % have been
used in the alloy development program to obtain

' good mechanical properties. Therefore, experiments

were conducted to determine how to alter the
fabrication process so that the memory effect would
be eliminated when the alloy carbon content was near
0.050 wt %. In fabricating alloy 452, doubling of hot
swaging reheat times as well as the time of the
homogenizing anneal (Table 6.10) was unsuccessful
in removing carbide stringers from the microstruc-
ture (Fig. 6.8a). Therefore, samples of the as-swaged
alloy 452 were solution-annealed for 1 hr at 1177,
1204, 1260, and 1300° C, respectively, in an attempt to
dissolve the MC-type carbides. These samples were
then aged at 760°C for 160 hr to reprecipitate the
carbides. The results of the solution annealing
experiment are shown in optical metallography (Fig.
6.15). Solution annealing at 1177° C (Fig. 6.15¢) and
at 1204°C (Fig. 6.15b) did not eliminate the memory
effect, and many carbides reprecipitated in areas
originally containing stringers. However, the anneals
at 1260°C (Fig. 6.15¢) and at 1300°C (Fig. 6.15d)
successfully removed the carbide stringers. Instead of
stringers, the higher solution anneals produced a fine
Widmanstatten structure of MC-type carbides within
the grains. The habit planes for carbide precipitation
were the matrix (111) planes. The relationship
between MC and matrix was epitaxial, that is,
(11Dmc || (11D manix. MC-type carbides were also
found 1n the grain boundaries. Much of the
precipitates in the sample annealed at 1204°C (Fig.
6.15b) had the Widmanstatten morphology, but
many of the carbides concentrated in prior stringer
areas. This observation strongly suggested that while
most of the carbides were dissolved at 1204°C, a 1-hr

-annealing time at this temperature was not sufficient

to permit adequate homogenization of the carbide-
forming elements throughout the matrix. These
results confirm earlier statements on solution-
annealing which stress the importance of not only
dissolving the carbides but also allowing the carbide-
ey Lk TNl o At . — |v-13a003

60 MICRONS 120 140
L 500X —! L1

T 7
INCHES 0.005

Fig. 6.15. Photomicrographs showing elimination of “memory effect” in alloy 452 (0.048 wt % carbon) by increasing the soluti?n-
annealing temperature. Electrolytically etched with oxalic acid. Annealing times and temperatures: (a) | hrat | 177°C plus 160 hr at 760°C,
(b) 1 hr at 1204°C plus 160 hr at 760°C, (c) 1 hr at 1260°C plus 160 hr at 760°C, and () 1 hr at 1300°C plus 160 hr at 760°C.

4a

Y-134997

"™ v.135004

L9
forming elements to disperse uniformly. The experi-
ments by Bradley and Leitnaker, to be discussed in
the next section of this report (Sect. 6.5.2), indicate
that molybdenum is the controlling element that
must be dispersed. ‘

From the result of the preceding experiments, it
was possible to design.a new fabrication schedule that
produced a microstructure without stringers in an
alloy containing ~0.050 wt 9%, carbon. The pilot run
of this new schedule is given in Table 6.10 for alloy
450 (see Table 6.9 for chemical composition). The arc
cast ingot was 1.0 in. in diameter by ~6 in. long and
was canned in type 316 stainless steel having a 0.120-
in. wall thickness. The purpose of cladding the
material was to minimize decarburization during
processing at high temperatures in air. The as-cast
ingot contained irregular-shaped grains with numer-
ous patches of carbide particles (Fig. 6.16a). Hot
swaging was conducted at 1177°C, as before, but the
reheat times were held to 15 min. After hot swaging,
the microstructure contained a fine grain structure
with carbides in the grain boundaries (Fig. 6.16b).
Areas lacking in carbides experienced more grain
growth, and hence a duplex grain structure was
produced.

The swaged rod of alloy 450 was annealed at
[1260° C for | hr which dissolved most of the carbides
and produced large recrystallized grains (Fig. 6.16¢).
However, enough carbides remained to justify an
additional anneal of 1 hr at 1300°C. The samples
were quenched in water after the 1300° C anneal to
prevent precipitation of carbides during cooling. This
latter anneal was effective in dissolving essentially all
of the carbides (as shown in Fig. 6.16d). The
discrepancy between the results of a 1-hr anneal at
1260° C foralloys 452 and 450 again apparently lies in
the differences in amounts of material annealed.
Thus the annealing practice must be controlled
closely with this alloy to obtain reproducible results.

The final three micrographs in Fig. 6.16 show the
material after cold swaging (Fig. 6.16e), after a
subsequent stress relief anneal for 15 min at 1177°C
(Fig. 6.16f), and after final cold swaging (Fig. 6.16g).
During these final steps the grain size was reduced
through cold work and recrystallization processes.
Some carbides were precipitated in the grain
boundaries, but no stringers were observed. As final
proof that the memory effect had been eliminated, an
as-swaged sample was aged at 760° C for 168 hr. The
microstructure after aging did not contain stririge_rs
but displayed a fine-Widmanstatten structure of thin
MC platelets lying epitaxially on the (111) matrix
planes (Fig. 6.16A).

68

Precipitate extraction data from samples selected
during the fabrication of alloys 452 and 450 are
compared (Fig. 6.17). These data show, in a different
way, how the memory effect was removed from 2%
titanium Hastelloy. Both alloys contained ~0.050 wt
9% carbon, and the cast ingot contained a nominal
amount of carbides. Some precipitation occurred
during hot swaging, and the total for both alloys
increased. Solution annealing was the critical step in
eliminating the memory effect, since substantially
more carbides were dissolved in alloy 450 because of
higher annealing temperatures. The extracted
amounts of carbides remained essentially unchanged
for both alloys after solution annealing. A high-
temperature anneal would effectively remove string-
ers from both alloys after fabrication was completed,
but the material would end up with a very large grain

'size. By performing the high-temperature anneal

early in fabrication, it is possible to produce some
grain refinement during later steps involving cold
work and recrystallization. Alloy 450 is currently

‘being tested to determine its resistance to radiation

damage and also its resistance to embrittlement by
tellurium. _

Conclusions. The following conclusions have been
reached as a result of work discussed above:

I. Carbide stringers form in 2% titanium-
Hastelloy N alloys, as in many other structural
materials, during fabrication from the original
castings. A memory effect has been observed in the
solution-annealed and aged 2% titanium-Hastelloy N
alloys where carbides reprecipitate 1n areas that
originally contained stringers.

2. The memory effect was eliminated, and a
uniform distribution of carbides was produced in
alloy 451 by lowering the carbon content of the alloy
from ~0.050 to 0.017 wt %. The standard 1 hr at
1177° C solution-annealing temperature was used.

3. The memory effect was eliminated in 2%
titanium-Hastelloy N alloys with ~0.050 wt %
carbon by increasing the solution-annealing tem-
perature to ~1300°C.The high-temperature anneal
was given early in fabrication to help maintain
a reasonably small grain size in the final material.

4. To remove the memory effect (carbide string-

ers), a solution-annealing treatment must:

a. dissolve most of the carbides,

b. be of sufficient duration to allow the carbide-
forming elements (especially molybdenum) to
disperse, by diffusion, uniformly throughout
the alloy, and
()

20 40 60 MICRONS 120 1
l—'—uT'—v—L 5uox

INCHES 0,005

Fig. 6.16. Photomicrographs of alloy 450 at various stages of fabrication by a process designed to climinate carbide stringers.
Electrolytically ctched with oxalic acid. (a) As cast, (b) after hot swagingat 1177°C, (¢) solution anncaled I hrat 1260°C, (d) annealed I hrat
1300°C and then water quenched, (¢) 39% reduction in arca by cold swaging, (/) annealed 15 minat 1177°C, (g) 45% reduction in area by cold
swaging (finished '/-in.-diam rod), and (k) aged 168 hr at 760°C.

70

ORNL-OWG 76-4733

0.5
&
_—=""N
N
0.4
he \450
\
\\—
E /‘\ \\
= 03 452
% <>/ :
w
g i B u |
® 1
v 0.2 ;
* \
\
\\
0.1 A ‘
\ _
\ -
________ -
\"__ ——— L
(o]

AS CAST
AT 1477°C

SOLUTION ANNEAL

Fig. 6.17. Extraction data showing the amount of carbides remaining in the microstructures of alloys 450 and 452 after different

fabrication steps.

c¢. take into account the volume of material being
treated—larger quantities of material require
more time. '

5. In 2% titanium-Hastelloy N alloys where the
memory effect was eliminated, the MC-type carbides
formed a fine Wid manstatten structure with thin MC
platelets lying epitaxially on'the (111) matrix planes.
Some small carbide particles were also observed in
the grain boundaries.

6. All the carbides observed in 2% titanium
Hastelloy N, regardless of carbon content or heat
treatment, were identified by both electron and x-ray
diffraction as the face-centered-cubic, MC type.

6.5.2 Carbon Behavior in .
Ni-Mo-Cr-Ti Alloys

D.J. Bradley  J. M. Leitnaker

To understand and predict the properties of
titanium-modified Hastelloy N, a basic study in the
phase relationships has been undertaken. The
primary manpower funding for this portion of the
project is an ORAU Fellowship. The program is
divided into two parts: (1) preparation and character-
ization of specimens and (2) determination of the

phase relationships in the matrix and the precipitate. "

The base alloy has been divided into three suballoys
having the compositions (1) Ni—-149% Mo—2%
Ti-0.19% C, (2) Ni-8% Cr2% Ti0.1% C, and (3)
Ni—2% Ti—0.1% C. A limited amount of work has also
" been done with alloy 472-503, a small commercial

AFTER 2hr  1hr
SWAGING 1:177°C 1260°C

1hr AFTER 1S min  hr AFTER
1300°C coLD n?rec coLD
wQ SWAGING STRESS SWAGING
RELIEF
ANNEAL

heat of 2% titanium-modified Hastelloy N. Three
areas have been investigated as part of the sample

- preparation and characterization portion of this

study. They are decarburization, precipitation rate,
and homogeneity of the material.

Decarburization is thought to be a surface reaction
of the type

O3(absorbed) T 2Csolia solutiom = 2CO .

In principal, this reaction can be easily controlled by
controlling the environment around the sample.
However, experience has shown that this is not a
simple task. Our work with decarburized samples has
shown a considerable vanation of the lattice
parameter of the MC-type precipitate asa function of
.bulk carbon concentration (Fig. 6.18). This can be
explained in terms of a lowering of the carbon
activity, resulting in a carbon-deficient MC lattice.
Similar behavior can be expected from other alloys,
which explains the wide range of lattice parameters of
precipitate found in this system earlier.” The carbon-
deficient MC phase identified in our samples is
probably not -the equilibrium phase for the given
temperature and bulk composition, but a result of the
carbon sink at the surface. Long-term reannealing of
the decarburized material could then result in the
formation of a precipitate with an equilibrium
composition, which would not necessarily be MC, .
Future experiments should show whether these
suppositions are correct and should yield the exact
relationship between the metal-to-carbon ratio and
precipitate lattice parameter.

71

ORNL-DWG 76- 3720 .

432

e

4.30

4.28 /

LATTICE PARAMETER (&)

4.26

200 400

600 800 1000 1200

CARBON CONCENTRATION (ppm)

Fig. 6.18. Lattice parameter of the carbide as a function of carbon concentration in the alloy.

The furnace in which the decarburization took
place has been extensively modified. Decarburiza-
tion can be properly controlled now and possibly
used to advantage. Rather than using O; to
decarburize, a mixture of Ha(g) and CHu(g) will be
flowed through the furnace to prevent decarburiza-
tion.

Homogeneous Alloys. The alloys used in this study
were fabricated in the usual manner with intermedi-
ate heat treatment at 1177°C. Subsequent metallo-
graphic examination revealed MC-type precipitate
stringers parallel to the working direction (Fig. 6.19).
These stringers represent a very inhomogeneous
distribution of the alloying elements; the precipitate
morphology can likely be better controlled if the
alloying elements are distributed more homogene-
ously. To thisend, alloys were annealed at 1260° Cfor
16 hrand then aged for 100 hrat 760° C. Previously, it
had been determined that 4 hr was sufficient to
dissolve the stringers at 1200°C. The samples aged at
760° C were then examined metallographically (Fig.
6.20a—c). As the photomicrographs reveal, the string-
ers are now only present in alloy 1. The matrix of
alloys 2 and 3 contains small, evenly distributed MC-
type precipitates, but in alloy 2 a high density of
carbide stringers is evident. The alloy containing
molybdenum appears to have a “memory effect” even

molybdenum is responsible for the memory effect
and that, once formed, stringers in molybdenum-
containing alloys are extremely difficult to remove.

Precipitation Studies. Another part of the present
characterization studies has centered on the rate of
precipitation. Originally all the samples were cooled
in the cold zone of the furnace. Nonreproducible
solubility data and discrepancies from expected
behavior led to the postulation that precipitation was
occurring on cooling. To verify this postulation, two
sets of samples were annealed at 1200° C (Table 6.11).
One set was cooled by being pulled into the furnace
cold zone, and the other was cooled by dropping into

Table 6.11. Effect of cooling rate on precipitation
at 1200°C in alloy 1 (Ni—13.9% Mo—-1.94% Ti—0.10% C)

Carbon concentration in

Sample |, 50~ Cooling extracted precipitates
number method?

(hr) No.1 No.2 No.3 Average
BA1HS 1 CZ 0.194 0.300 0.247
BAZHS 2 CZ 0.199 0.342 0.270
BA4HS 4 CZ 0.106 0.197 0.151
Ba8HS 8 Cz 0.118 0.209 0.164
BA16HS 16 CczZ 0.130 0.167 0.148
BA1H 1 BQ 0.097 0.084 0.142 0.104
BA2H 2 BQ 0.066 0.084 0.134 0.96

after 16 hr at 1260° C. A high-carbon sample of alloy
472-503 has also been examined, and the memory
effect was found to be still evident after 50 hr at
1170°C (Fig. 6.21). The conclusions are that

@CZ indicates that the sample was cooled by pulling into the
cold zone of the furnace. BQ indicates that the sample was
dropped from the hot zone directly into an H,0-10% NaCl
solution at room temperature.
72

—
. Sy
100 200 MICRONS 600 700
1 1 L 00X—L 1 It
T T T T T
0.005 0.010 INCHES 0.020 0.025

Fig. 6.19. Photomicrographs of alloy 3 in the as-received condition. Note the decarburized zone and the high density of stringers.

Etched with glyceria regia.

an H,0-10% NaCl solution. The quantity of
precipitate would contain approximately 20 to 40%
of the carbon in the alloy. Thus, cooling in the cold
zone is not fast enough for measurement of
solubilities or the rate of solution of precipitates.

Our exploratory work on the solubility of carbon
does not appear to correlate with literature values.
Results reported by Stover and Wulff* on the nickel-
titanium-carbon system indicate that above 800°C,
an alloy containing less than approximately 2.5at. %
titanium cannot precipitate TiC even in the presence
of graphite. Alloy No. 3 contained only 2.3 at. %
titanium; yet TiC was identified in a precipitate
formed at 1170° C. Unfortunately thisalloy appeared
to contain at all temperatures approximately 0.15 wt
Y precipitate, which was amorphous (Table 6.12).
This amorphous material, which is now being
analyzed, has clouded interpretation of the present
results. New alloys are being prepared which should
not contain this material.

Alloy No. 2 also appeared to contain ~0.15 wt %
amorphous material (Table 6.12). X-ray diffraction
of the precipitates yielded only weak TiC peaks. In

one case, no detectable diffraction occurred. Assum-
ing that the nondiffracting material in both alloys 2
and 3 was not TiC, the solubility of carbon in each
alloy is greater than 0.10 wt % at 1260°C. At 760°C
the solubility of carbon in alloy 2 was =0.05 wt % and
in 3 was >0.1 wt %.

The precipitate from alloy No. | gave large, sharp
diffraction peaks, even with very small amounts of
material. If it is assumed that the material was
stoichiometric TiC (Table 6.12), the solubility of
carbon was ~0.08 wt % at 1260°C. The precipitate
formed at 760°C could not be stoichiometric TiC,
since 0.73 wt % precipitate would require a bulk
carbon concentration of 0.14 wt %, and the alloy
contained 0.10 wt % carbon. Previous investigators
estimated that in titanium-modified Hastelloy N the
precipitate was approximately 60 wt % molybde-
num.

A limited amount of work was also done on the
effect of the alloying elements on the matrix lattice
parameter (Table 6.13 and 6.14). It is hoped that
changes in the lattice parameter of the matrix can be
correlated with precipitation. Unfortunately, the

100 200 MICRONS 690 7?0
1 1 1 1 3

T ! e i
0.005 0.010  INCHES 0.020 0.625

Fig. 6.20. Photomicrographs of the three alloys annealed 16 hr
at 1260° C followed by aging at 760° C for 100 hr. (a) Alloy 1 (note
the presence of stringers); (b) alloy 2; and (c) alloy 3. Etched with
glyceria regia.
74

Y-136613

100 200 MICRONS 600 700
L L 400X—L LI
0.005 0.010 INCHES 0.020 0.025

Fig. 6.21. Photomicrographs of Hastelloy N, heat 452-503, aged 50 hr at 1170°C. Note the presence of stringers. As polished.

Table 6.12. Solubility and lattice parameter data on titanium—carbon-type precipitates

Annealing s inimum o
Alloy Sample condifions® Carbnnb Precipitate Mcarbon Lattice
- ———————  content extracted® parameter?
Number Elements number  Temperature  Time Wt %) (Wt %) solubility @)
o (hr) (wt %)
1 Ni,Mo,Ti,C  AS$ 1260 16 0.102 0.0930 0.084¢ 4.288
A10 1260 16 0.092 0.0370 0.085¢ 42991
4.289
AS-A 760 100 0.096 0.7320 0 4313
3 Ni, Ti, C BI3 1260 16 0.083 0.1150 >0.08/
BIS 1260 16 0.080 0.1560 >0.08/
BI5SA 760 100 0.080 0.1410 >0.08/
2 Ni, Cr, Ti, C c7 1260 16 0.098 0.1170 >0.10/
C6 1260 16 0.103 0.1490 >0.10/
C6A 760 100 0.104 0.3590 0.05" 4288

“The samples were annealed together and then cooled in the cold zone on the furance.
'Carbon concentration was determined on a leco apparatus by analytical chemistry. Precision of the determination at this
concentration is approximately +0.0030 wt %.
“The weight percent precipitate was obtained by electrolytic extraction of the samples in a 10% HCl/methanol solution at 1.5 Y-
The normal extraction electrolyzed 1 g of metal. The precision of the weight percent determination is approximately :0.015 wt %.
‘The lattice parameter of the precipitate was determined by x-ray diffraction. The spectrometer normally covered a 26 range of
20 to 120°. An internal standard of TaC was used to calibrate each run. The data were treated by means of a cosé coté extrapolation
function.
®Based on the premise that the carbon content of the precipitate is 20 wt %. This is the case for stoichiometric TiC.
'Calculated assuming that 0.15 wt % ppt was not TiC.

75

Table 6.13. Lattice parameter data on
titanium-modified Hastelloy N matrices

Composition i
Alloy (at. %) Lattlce( ga;iameter
Ni Ti Mo Cr
1 88.55 2.50 8.95 3.56739 + (_).00041
3 97.7 2.30 3.53266 = 0.00042
2 88.65 2.34 9.01 3.54170 + 0.00026

“The lattice parameter was measured with a D.S. camera. The
data were extrapolated to 26 = 180° with Nelson-Riley
extrapolation function.

Table 6.14. Effects of alloying elements
on the nickel lattice

ffe tivelatomic
Aagyfat. % Etfec

Alloy radii®
hi i a - .
This work  Literature value This work Slater radii€

Ti 0.00387 0.0034 1.398 1.40
Mo 0.00398 0.0042 1.402 . 145
Cr 0.00100 0.0011 1.348 1.40

“W. B. Pearson, A Handbook of Lattice Spacings and Struc-
tures of Metals and Alloys, Pergamon Press, Oxford, 1967.

bCalculated assuming close packing along fcc diagonal and the
value of 3.5238 A as the lattice parameter of nickel (see ASTM
card 4-0850).

“Values taken from J. C. Slater, “Atomic Radii in Crystals,” J.
Chem. Phys. 41, 3199 (1964).

previously mentioned problems have not yetallowed
this result, but the data on the decarburized
specimens do agree well with existing data on similar
systems. ,

6.6 SALT CORROSION STUDIES
J. R. Keiser E. J. Lawre_nce

Studies of the corrosion resistance of potential
MSBR containment vessel materials (modifications
of Hastelloy N)have been carried out in both thermal
convection and forced circulation loops. Results
indicated a limited amount of mass transfer of
chromium from the hot to the cold portions of the
Hastelloy N systems, but the overall resistance of
Hastelloy N to molten fluoride salt corrosion is very
high. Unfortunately, Hastelloy N has been found to
be susceptible to irradiation embrittlement and to
grain boundary attack by fission products. Conse-
quently, investigations were begun with different
commercial and experimental alloys containing
compositional modifications to Hastelloy N in order
to find materials that provide sufficient resistance to
the detrimental effects of irradiation and fission
products. Since it is necessary to ensure that these
new or modified alloys are resistant to fluoride salt
corrosion, most of the loops now in operation are
being used to evaluate the corrosion resistance of
these various materials. The status of the operating
thermal convection loops is given in Table 6.15. In

Table 6.15. Status of operating thermal convection loops

Loop Loop Specimen : Maximum AT
. P
number material material Salt type tem(p;eg)lture (9] urpose
21A Hastelloy N Hastelloy N, MSBR fuel salt 704 139 Baseline corrosion data for
1% niobium-modified Hastelloy N
Hastelloy N
Screening test for 1% niobium-modified
Hastelloy N
Tellurium mass-transfer and
corrosion studies
23 Inconel 601 Graphite MSBR fuel salt 671 117 Effect of graphite on stability of
uranium fluorides in molten salt
31 316 stainless 316 stainless LiBeF, 649 156 Baseline corrosion data
steel steel
Effect of reductant additions on
corrosion
24 Hastelloy N Chromium-modified MSBR fuel salt 704 139 . .
Determine the effect on corrosion
Hastelloy N . .
rate of chromium concentration
18C Hastelloy N Chromium-modified MSBR fuel salt 704 139 of Hastelloy N

Hastelloy N

addition to these five loops, three others have been
built and are in varying stages of final preparation.
Because of the anticipated termination of the
program, these three loops will not be put into
operation.

6.6.1 Thermal Convection Loop Results

Thermal convection loops provide a very good
experimental tool for measuring the corrosion rate of
a material exposed to a corrosive medium under an
imposed temperature gradient. Removable speci-
mens permit periodic examination to determine
weight changes, and electrochemical probes permit
on-line measurement of the oxidation potential of the
salt and the concentration of impurities in the salt.

Thermal convection loop 23 is constructed of
Inconel 601, a material that has shown good
resistance to grain boundary attack by tellurium.
However, results have shown that Inconel 601
undergoes extensive corrosion in fuel salt, making ita
material with little prospect for use in an MSBR.
Because of the extensive corrosion which this alloy
has undergone by reaction with fuel salt_the U*"/ U
ratio of the salt has fallen to about 4, the most
reducing level ever achieved in a loop. As reported
previously10 it was decided to use this very reducing

salt to try to reproduce the results of Toth and’

Gilpatrick,'' which predict that under the operating
conditions of this loop and in the presence of
graphite, UC; would be formed by the reaction

4UF; + 2C - 3UF + UG, .

In the first attempt to produce this reaction, pyrolytic
graphite was exposed to the salt for 500 hr, and no
UC; was found. In the second attempt a less dense
graphite was used and a new phase was detected
through x-ray examination. This phase was tenta-
“tively identified as uranium oxide. A third set of
graphite specimens which- has been kept free of
moisture is now undergoing a 2000-hr exposure and
is scheduled to be examined in March 1976.

Hastelloy N thermal convection loop 21 A, which
contains MSBR fuel salt, has been used to obtain
baseline corrosion data. Sixteen Hastelloy N speci-
mens have been exposed a total of 10,009 hr.
According to theory, the weight change of these
specimens due to mass transfer of chromium, AM,
per unit area during time ¢, should be given by

AM < C,\/ Dt ,

76

where C, is the initial concentration of chromium in
the alloy and D is the diffusivity of chromium in the
alloy. Figure 6.22 shows a plot of experimentally
determined weight ‘change, AM, vs the /¢ for a
specimen near the hottest point of loop 21A, The
good fit of the data to a straight line provides further
evidence of a diffusion-controlled corrosion mechan-
ism. A new set of specimens fabricated from 1%
niobium-modified Hastelloy N has been inserted into
the loop and will be used to obtain corrosion data for
that alloy.

Thermal convection loop 31 is constructed of type
316 stainless steel, has 316 stainless steel specimens,
and contains LiF-BeF; (66-34 mole %) salt. For the
first 1000 hr of operation, this loop was used to gather
baseline corrosion data with “as-received” salt.
Subsequently, reductant (beryllium) was added to
the salt at the level of about 0.1 mg Be/cm’ salt. New
specimens were then inserted, and the corrosion rate
has been measured for stainless steel in this reducing
salt. Addition of beryllium made the salt quite
reducing, but once the source of beryllium was
removed, species in the salt were no longer in
equilibrium and the salt became progressively more
oxidizing with increasing time. While the source of
beryllium was in the salt, the corrosion rate was
extremely low. However, after removal of the
beryllium the specimen in the hottest position
showed a pattern of increasing weight loss as a
function of time. The increasing corrosion rate was

ORNL-DWG T7€-4843

AN

oAM (rng/crflz)
[ )

-2.5 ~
N
.
=30
~N
-35 J
0 20 40 60 80 100

A/
(EXPOSURE TIME) 2 {¥hr)

Fig. 6.22. Weight change vs square root of exposure time for
Hastelloy N specimens exposed to MSBR fuel salt at §90°C in
thermal convection loop 21A.
associated with the increasingly oxidizing state of the
salt. Weight change results for specimens in both the
“as-received” and “reducing” salt are shown in Fig.
6.23. Fresh salt and new specimens are to be added to
this loop, and corrosion experiments will be contin-
ued.

Two Hastelloy N loops, NCL 24 and NCL 18C,
have been filled with MSBR fuel salt and put into
operation recently. These two loopsare being used to
measure the effect of chromium concentration on the
corrosion rate of Hastelloy N. Increasing the
chromium content of a nickel-base alloy generally
increases the alloy resistance to grain boundary at-
tack by tellurium. However, increasing the chromi-
um content decreases the alloy’s resistance to mass
transfer. From these tests and tellurium exposure
tests, a possible optional chromium concentration
may be determined. For these corrosion tests, four
modified Hastelloy N alloys containing 7, 10, 12,and
15% chromium have been made. Specimens of the 7%
chromium alloy have been exposed in NCL 24 a total
of 1060 hr. The maximum corrosion rate observed
was —1.9 mg/cm’-year. The 10% chromium speci-
mens have shown a maximum corrosion rate of —2.3

ORNL-DWG 76-3496

30

510°C

20

® "AS RECEIVED" SALT
X AFTER BE ADDITION

* 510°C

WEIGHT CHANGE (mg)

-10 =
650°C

-20

650°C [ |

] 500 1000 1500 2000
EXPOSURE TIME (hr)

Fig. 6.23. Weight changes of type 316 stainless steel speci-
mens exposed to LiF-BeF; (66-34 mole %) salt at the indicated
temperature. Surfaces areas of the specimens were approxi-
mately 3.4 cm?.

77

mg/cm’-year after 550 hr in the salt of NCL 18C.
Exposure will continue until each set of specimens
has been exposed about 1000 hr.

6.6.2 Forced Circulation Loop Results

Hastelloy N pump loop FCL-2b was used during
the first six months of 1975 to obtain baseline
corrosion data under conditions where the U*"/ U**
ratio was about 100. A maximum corrosion rate of
~2.5 mg/cm’-year (—0.11 mil/year) was reported
after a total of about 3200 hr of a planned 4000-hr
exposure to salt. Figure 6.24 shows the effect of
material transfer on specimens from the hottest and
coldest positions. There is evidence of material
removal and deposition. Heat transfer measurements
were scheduled to follow the 3200-hr corrosion test
but had to be delayed because a leak was discovered
in the Hastelloy N tubing. Following repair of the
tubing, the heat transfer measurements were made
and then the corrosion specimens were inserted for
another 1000-hr exposure. Weight change measure-
ments made at the end of this period revealed that
over the total 4309 hr at specimen temperatures of
566, 635, and 704° C the corrosion rates were 0.0,
—0.2, and —2.3 mg/ cm’-year respectively. This rate of
corrosion is well within the limit that could be
tolerated in an MSBR.

It was planned to adjust the U*'/ U’ ratio of the
salt upward to 1000 by means of NiF: additions
followed by a 4000-hr corrosion test in the more
oxidizing salt. Several NiF, additions were made,
and the U*'/ U ratio did increase; but another leak
occurred before the ratio reached 1000. This leak was
repaired, but shortly after operation was resumed, a
third leak occurred. Because of plans to terminate the
MSR Program, it was decided to forego further
testing of standard Hastelloy N and to use the loop to
gather baseline corrosion data on 19 niobium-
modified Hastelloy N, an alloy that shows improved
resistance to tellurium grain boundary attack.
Specimens were inserted at the end of February and
will be removed periodically for examination.

After each of the three leaks, a section of tubing
containing the failure was removed, and new tubing
was welded into place. These failed sections were
examined metallographically. Figure 6.25 shows the
appearance of the '/,-in.-OD tubing at the site of the
first leak. This leak occurred at a 90° bend on the
outermost portion of the tubing. Examination of a
transverse section of the tube (Fig. 6.26) showed
extensive cracking in the vicinity of the leak, and the
¥-134310
.
(@)
v-134308
]
L]
.
(®)
20 40 60 MICRONS 120 140
} U 500X —b— L1
0.001 INCHES 0.005

Fig. 6.24. Photomicrographs of standard Hastelloy N exposed to fuel salt in FCL-2b for 3200 hr. (a) 704°C, (b) 566°C. As
polished.

Fig. 6.25. Photograph of first failure in FCL-2b. 12X

Y-138067

0.056in
1.41mm

Fig. 6.26. Photomicrograph of a transverse section of the failure shown in Fig. 6.25. Inside diameter of tube on the lower side and
outside diameter on the upper side
80

longitudinal section (Fig. 6.27) showed that these longitudinal sections of the failure (Figs. 6.29 and

small cracks ran parallel to the axis of the tube. The 6.30 respectively) show a single crack. The transverse
presence of a shorted heater and a burned-out view shows evidence of necking or reduction of area
thermocouple on the section of tubing where this indicative of a fairly large stress. A high stress could
failure occurred provides additional support for the have occurred during heating of this region if salt
conclusion that this failure resulted from a combina- were trapped there between two frozen salt plugs.
tion of excessive temperature and low stress for an The third failure occurred ata weld joint in one of
extended time. This leak possibly could have been the specimen access positions (Fig. 6.31). Metallo-
avoided if the tubing had been annealed after bending graphic examination of the failure showed one
or if the temperature had been maintained in the crack, which probably occurred when a high stress
desired range. This contention is further supported again resulted from heating trapped salt (Fig. 6.32).
since sections of the failed tubing which had not been As reported by Huntley in Sect. 2.3.1, modifica-
severely stressed during bending did not show tions have been made to the loop to prevent develop-
cracking and since examination of a nearby 90° ment of frozen sections.
elbow that had been operated at the proper
temperature showed no cracking even in the most 6.7 CORROSION OF HASTELLOY N
severely deformed regions. However, the length of AND OTHER ALLOYS IN STEAM
tubing inv9lvcd would have made annealingdiffiqull. B. McNabb H. E. McCoy
and changing the thermocouple would have required
drainage and shutdown of the loop. The corrosion of Hastelloy N and other nickel,
The second failure occurred in a straight section of iron, and cobalt-base alloys is being evaluated in

resistance-heated tubing; the appearance of the failed TVA’s Bull Run Steam Plant at steam conditions of
section is shown in Fig. 6.28. Transverse and 538°C and 3500 psig. Unstressed sheet specimens 2

. Y-134876
D

|50 100 MICRONS 300 350
F—————r—200X—r—————
0.005  INCHES 0.010 0015

Fig. 6.27. Longitudinal section of the failure shown in Fig. 6.25. As polished.

81

Y-135146

Fig. 6.28. Photograph of the second failure in FCL-2b. The tubingis % in. in diameter. 6.

Y-135283

200 400 MICRONS 1200 1400
I i I Vs |~ el 1 h
I T T T T
0.04 INCHES 0.05

Fig. 6.29. Transverse section of the failure shown in Fig. 6.28. As polished.

82

Y-135279

200 400 MICRONS 1200 1400
1 1 1 50X —L I 1
T T T T L
0.01 INCHES 0.05

Fig. 6.30. Longitudinal section of the failure shown in Fig. 6.28. As polished.

PHOTO 4970-75

Fig. 6.31. Photograph of the third failure in FCL-2b.

Y-138354

0.056in
1.41mm

Fig. 6.32. Photomicrograph of the failure shown in Fig. 6.31. As polished

in. long X ' in. wide X 0.035 in. thick have been
exposed to flowing steam (approximately 1000 Ib/ hr)
for 21,000 hr. The various alloys of modified
Hastelloy N have thin, adherent oxides, with no
evidence of spalling. Some of the Croloy type steels
are beginning to develop blisters and some of the
blisters have cracked, but none of the specimens have
spalled at this time. The weight changes had
increased approximately 0.5 mg/cm’ overa period of
4,000 hr, as reported previously,'* but they decreased
during the last 2,000 hr (two weighing periods), until
some of the specimens are approaching the rates
previously established. Some of the specimens have
weight changes still slightly above the previous rates,
but apparently some of the deposited materials are
being removed.

The weight change behavior correlates well with
the steam quality and with observations by Bull Run
engineers. Several leaking condenser tubes were
sealed off during the ten-week shutdown in the latter
part of 1975, and the steam quality improved
considerably. No condenser leaks had been experi-
enced before 1974, but several occurred during 1974

and 1975. During the shutdown the condenser tubes
were ultrasonically inspected, and any tube whose
wall thickness had been reduced by 35% was sealed
off on both ends. The condensers were sized such that
a certain percentage of the tubes could be sealed off
and the condenser still perform satisfactorily. The
condenser tubes are constructed of admiralty brass,
and corrosion occurs near baffles in the regions
where droplets can form and then evaporate. This
leads to an increased concentration of impurities
(such as chlorides) and can cause corrosion even
though the bulk concentrations may be in the low
parts per million or parts per billion range. The
manufacturer of the condensers has made some
recommendations to remedy the problem, such as
water injection to wash the tubes in this area. The
effect of plugging some of the condenser tubes is to
shift the position of the vapor-to-droplet area of high
corrosion to another location in the condenser, thus
exposing areas of the tubing that have had less
corrosion and exposing the corroded areas to a less
corrosive condition. Some experimental condenser
tubes of Cu-10% Ni do not exhibit as much

susceptibility to this type of corrosion as the
admiralty brass. These have been proposed for
replacement tubes and should relieve the problem.
Two heats of standard Hastelloy N tubing (N1
5095 and N2 5101) are being evaluated in the stressed
condition from 28.0 X 10° to 77.0 X 10’ psi. The
specimens of both heats in the annealed condition (1
hr at 1177°C) have shorter rupture times in steam
than in argon. Specimens of heat N1 5095 tested in
the as-received condition and stressed below 50.0 X
10° psi in steam have rupture times equal to those of

84

specimens tested in argon. Figure 6.33 shows the
stress-rupture properties of heat N1 5095. As
reported earlier,”’ flaws in the specimens may have
contributed to the scatter, but at stresses <<50.0 X 10°
psi, where wall thicknesses are heavier, there appears
to be no significant effect of steam on the rupture
time. Two of the specimens stressed at 28.0 and 40.6 X
10° psi have an accumulated 15,000-hr exposure to
steam at 538° C without rupture, and three others at
slightly higher stresses have from 10,000 to 11,000 hr
of exposure. Figure 6.34 shows the stress-rupture

ORNL-0OWG 75-7202R

T Iy T Ty T PO T T rmmp T 1T
. AS- ANNEALED 1 hr
80 o RECEIVED AT {477°C ]
$0 \‘\\ ARGON © o
ool STEAM ® .
70 — 8o \\OT:D—‘OC |
R * . ;\\‘? o
£ 60 - oo —p—
A ¢ d 4 \'\O\
fi ) © 131\0\
'(7) 50 L) 2 .‘\
[
&
40 o .-
30
@
S I A T AV O A
10° To} 102 103 10% 10°

RUPTURE TIME (hr)

Fig. 6.33. Rupture life of Hastelloy N (heat N1 5095). Tubes stressed in argon and steam environments at 538°C.

a0

ORNL-DWG 75-7201R

RN HlHH\ RN
o
o) AS- ANNEALED 1 hr
80 ‘*\\ RECEIVED AT {77°C —f
. S~
s O~ STEam o .
70 L \f“\{ @
= ~o
:’.‘m, n a \\"--..\ ° "
¢ 60 as o— o .
w e
2 s . o~ _e?
n : s e
50 - O ™~
*
40 hahat
0 R e
10° 10! 102 10° 104 10°

RUPTURE TIME (hr)

Fig. 6.34. Rupture life of Hastelloy N (heat N2 5101). Tubes stressed in argon and steam environments at 538°C.
ORNL-DOWG 76-7424

T T TR T T T
60 ‘/:
o‘/0
50 //
//o
Oo "ol
o /
9 /
w /
|9_:30 ’o
v /
HEAT
® N25101
20 o N15095
10
B I I e N i
10°6 1073 10-4 1073

MINIMUM CREEP RATE (% /hr)

| Fig. 6.35. Creep properties of Hastelloy N at 538°C in steam
environment.

properties of heat N2 5101. Specimens in the
annealed condition tested in steam have shorter
rupture times than specimens tested in argon.
Specimens in the as-received condition tested in
steam have rupture times equal to or greater than
specimens tested in argon. The longest rupture time
of this heat to date was 12,000 hr for a specimen
stressed at 48.8 X 10’ psi in steam at 538°C.

The minimum creep rates of both heats of material
were plotted as a function of stress (Fig. 6.35). The
creep rates are calculated from plots of diametral
strain, AD/ D, vs time, which are based on measure-
ments of the internal diameter of the specimens at
1000-hr intervals. There is considerable scatter in the
data because of difficulty in measuring the internal
diameters at the point of maximum strain, but the
figure is useful as an indication of the strain rates
expected in this stress range. Buildup of scale on the
inside diameter exposed to steam also contributes to
the inaccuracy of this method since the apparent
strain 1s reduced as the scale thickens.

6.8 VAPOR PRESSURE MEASUREMENTS

OF METAL TELLURIDES
S. L. Bennett J. Brynestad

Screening  tests were performed to establish
whether standard and modified Hastelloys are

85

embrittled by exposure to tellurium activities defined
by the system Ni;Tez(s) + Ni(s) at 700°Cand 750°C; a
knowledge of the tellurium activity (i.e., vapor
pressure) under these conditions is important.
Arrangements were made to attempt to obtain vapor
pressure measurements over the two-phase system of
nickel plus Ni;Te; at 700° C by mass spectrometry at
Y-12. However, the sensitivity of the instrument was
far from satisfactory; a temperature of 850°C was
required in order to begin to observe the species Te
and Te. The experiment did confirm, however, that
tellurium monomer and dimer are the only gaseous
species present in this system, in accordance with the
observations of Weaver and Redman." |
Since measureable weight losses were observable
with the Mettler Recording Vacuum Thermoana-
lyzer at about 800°C and above, quantitative vapor
pressure measurements are being made by this
technique. The Knudsen effusion equation,
Pp,. =(my, [44.33at)(T/My, )"/, (1)
relates the partial pressure of the species to the mass -
of that species of molecular weight M which effuses in
time ¢ through an orifice of effective area a at
temperature 7. A similar expression can be written in
terms of the Te monomer. The dissociation equilib-
rium '
Teag) = 2Te(g) , (2)
for which the equilibrium constant K, = Pr.’/ Pre,,
must be simultaneously satisfied. Letting m = mr,, +
mr., one can derive the relationship

Pr,, = {[Kp/8+ (m[4433 ar)(T/My, )' 1?1/

~(K,/8)' P} . (3)

That is, the partial pressure of one species can be
expressed in terms of the experimentally measured
total mass loss m and the equilibrium constant. The
equilibrium constant as a function of temperature is
calculated from Eq. (4):

_RTQHKP =Do0 t T{Z[(GTO “Hoo)/T]Te
~((Gr° ~H)T)re }, @)
where the (G7 — Ho%)/ T terms are the free energy

functions" and Dy’ is the enthalpy of dissociation of
Te: at 0°K."*
The samples are contained in tungsten liners inside
a high-density graphite (POCO Grade AXF-5Q)
Knudsen cell. The effective orifice area is the product
of the cross-sectional area, measured by a photomi-
croscopic technique, and the Clausing factor, the
latter being a function of the ratio of the length to the
radius of the orifice.'’

Silver was vaporized over the temperature region
975 to 1320°C, representing a pressure range of 2',
orders of magnitude. The derived enthalpy at 298°K,
AHas', of 67.0 kcal mole™' is in good agreement with
the literature value of 67.9 kcal mole™,"® verifying the
temperature measurements and the general tech-
nique.

So far, three experiments have been done with the
system NiyTe: plus nickel over the temperature
region of 800 to 1000° C with various effective orifice
areas. The logarithm of the calculated Te; partial
pressures vs reciprocal temperature is shown in Fig.
6.36. If the orifice area is too large, then equilibrium
may not be achieved, and the calculated pressures
may be too low (Fig. 6.36). Reducing the effective
orifice area by a factor of 10 resulted in Te: pressures

ORNL-DWG 76-3657

-4.2
—aq %
-4.6 %X\\
[
=50 b8
E ANAY
2 52 \ \Ou.
N \
Q »
o 54 .
(=
- b
-5.6 o (cm?) N
0 29.8 x1074 :\Q \
-5.8 a 291x1074 .Y
e 1.22x1074 \\ \
-6.0 N
6.2 \\
6.4
78 B0 82 84 B6 88 90 92 94

10000/ (0

Fig. 6.36. Clausius-Clapeyron plot of Te, partial pressure

over Ni3Te, + Ni for several Knudsen cell effective orifice areas.

86

approximately twice as great. However, on reducing
the effective orifice area an additional factor of 2.5,
the Te, pressures remained unchanged within
experimental error. Hence it seems that equilibrium
has been achieved in the latter two experiments; some
confirmatory experiments are planned for even
smaller effective orifice areas. If we tentatively accept
the data obtained with an effective orifice area 0f2.91
X 10 cm® as representing equilibrium, then
extrapolation of the data to 700°C gives P 1., = 8 X
10" atm and Pre =2 X 107 atm.

6.9 OPERATION OF
METAL-TELLURIUM-SALT SYSTEMS

J. R. Keiser E. J. Lawrence

The intergranular cracking of Hastelloy N which
was observed in the Molten-Salt Reactor Experiment
has been shown to be caused by the fission product
tellurium. Because the rate of cracking observed In
the MSRE might not be acceptable over the 30-year
design life of an MSBR, it was concluded that the
MSBR primary containment vessel material should
have improved resistance to grain boundary attack
by tellurium. To evaluate various alloys and
modifications of Hastelloy N for their resistance to
tellurium attack, it was necessary to develop tests that
would simulate the appearance of tellurium as a
fission product in the fuel salt of a reactor. Five
experiments are now being conducted to learn more
about the behavior of various tellurium compounds
in salt and to determine the effect of these telluride-
salt mixtures on potential containment vessel materi-
als.

6.9.1 Tellurium Experimental Pot No. 1

This system was built to evaluate the use of lithium
telluride asa means for adding tellurium to salt which
would simulate ‘the production of tellurium by
fission. The pot was constructed to permit periodic
additions of salt pellets containing small, measured
amounts of lithium telluride.

For the first experiment, Li;Te was used in the
pellets that were added to the LiF-BeF;-ThF, salt in
the pot.”” No evidence of a soluble tellurium
specimen was found by either chemical analysis or
electrochemical examination.” A second experiment
was conducted using the reportedly more soluble
lithium telluride, LiTe;. After the pot had been
drained, cleaned, and fresh salt added, three pellets
containing a total of about 0.1 g of LiTes were
inserted. During the following three weeks, electro-
chemical examinations of the salt by D. L. Manning
of the Analytical Chemistry Division revealed no
indication of a soluble tellurium species. Since other
experiments had shown that metal specimens could
be embrittled by tellurium even when tellurium levels
in the salt were below chemical detection limits, a
Hastelloy N tensile specimen was inserted into the
salt and three additional LiTes-salt pellets were
added. After about 500 hr the specimen was removed
from the salt, tensile tested to failure, and examined
metallographically. Intergranular cracking, most
probably due to tellurium, was found, indicating that
‘a mechanism exists for transport of tellurlum or
LiTe; through the salt.

6.9.2 Chromium Telluride Solubility Tests

The original purpose of this test pot was to
measure the solubilities of Cr;Tes and Cr;Tes in LiF-
BeF->-ThF, salt. First, additions of CryTes were made,
and later additions of Cr;Te; were added. Appropri-
ate salt samples were taken at various temperatures,
together with follow-up double checks with filtered
samples; the results showed that, within the limits of
chemical analysis, these compounds were not soluble
in salt. However, Hastelloy N specimens exposed to
these chromium telluride-salt mixtures, when
strained, cracked intergranularly and showed evi-
dence of tellurium on crack surfaces when examined
with Auger electron spectroscopy. Consequently, it
was decided to expose additional specimens for
Auger studies. Eight specially prepared specimens
were inserted in the Cr.Tes-salt mixture, and one was
removed every 24 hr. The examinations of these
specimens are described in Sects. 6.10 and 6.11.

Because of the high rate of attack on specimens in
the salt containing Cr:Te;, a second test was Started
using Cr;Tes, which has a lower tellurium activity.
Several modified Hastelloy N alloys with different
combinations of niobium, titanium, and chromium
were exposed to the salt-CriTes system. Eight
standard Hastelloy N specimens for Auger studies
were exposed to the salt, and one was removed every
48 hr for the purpose of gainifig information about
the rate of tellurium attack. All specimens were

‘submitted for examination, and the results are given
in Sects. 6.10-6.12. '

6.9.3 Tellurium Screening Test

. A large pot has been constructed -for exposing a
large number of specimens to a salt-tellurium system
in order to compare the resistance of various alloys to
tellurium attack. The pot is equipped with a stirring

87

mechanism, electrochemical probes, and five access
ports for insertion of specimens. In late February the
pot was filled with about 44 kg of LiF-BeF:-ThF,

salt, and a mixture of Cr;Tes and CrsTes was added.
Specimens of 19 alloys have been inserted for a 250-
hr exposure, with duplicate sets in the salt and vapor
regions. Further studies of potential containment
materials will be carried out during the next few
months.

6.9.4 Ni;Te; Capsule Test

Results reported by Brynestad”' indicated that a
Hastelloy N specimen exposed to vapor above NisTe
+ nickel for 1000 hr .at 700°C did not show
intergranular cracking. This suggests that the NisTe;
+ nickel system has a tellurium activity at which
Hastelloy N is not attacked. To further evaluate this
potentially significant result, a test was designed. to
expose Hastelloy N tensile specimens to salt

containing NisTe; + nickel. A capsule was built and

filled with LiF-BeF,-ThF, salt. A mixture of 62 at. %
nickel-38 at. % tellurium which had been given a
high-temperature anneal to promote formation of
Ni;Te; was added to the capsule, and four Hastelloy
N specimens were inserted. One specimen was
removed after 1000 hr at 700°C, strained to failure,
and examined metallographically. Contrary to
Brynestad’s earlier results, extensive intergranular
cracking was found. At present, the NisTe: + nickel

mixture is being checked to be certain it contains no

unreacted tellurium which would giVe an anoma-
lously high tellurium activity. If the telluride checks
out satisfactorily, it must be concluded that even the
low tellurium activity associated with nickel+ N13Te2
causes intergranular attack of Hastelloy N.

6.9.5 Chromium-Tellurium-Uranium
Interaction Experiment

The research effort to solve the problem of
tellurium intergranular attack of Hastelloy N has
been almost solely concerned with the development
and testing of new alloys. The four experiments
previously described in this section are part of the
alloy testing program. The possibility of a chemical
change in the salt which could ‘alter the extent of

-attack by tellurium is another manner in which this

problem may be approached. It has been suggested
by Brynestad that since the chromium activitycanbe
controlled within a certain range by the U* /U™
ratio, it might be possible to control the tellurium
activity by the reaction:

CrF; + 2UF; + Cr3Teq ~ 4“CrTe” + 2UF, .
In this equation, “CrTe” is used to denote a telluride
species having activity low enough not to promote
intergranular cracking.

To check. this hypothesis, a salt pot was built,
equipped with electrochemical probes, and filled with
MSBR fuel salt (LiF-BeF;-ThF:-UF4). The oxida-
tion potential and the impurity content of the salt
were monitored.* Additions of CrF; and CrsTes were
made, and a Hastelloy N specimen was exposed for
500 hr to salt with a U*/U” ratio of about 90.
Metallographic examination of the specimen after
deformation showed extensive cracking. To lower
the oxidation potential of the salt, a beryllium rod
was immersed in the salt for an hour, after which time
the U*/ U ratio had decreased to about 7. Another
specimen will be exposed to the salt-telluride
mixture, and a comparative evaluation will be made
of the extent of intergranular attack.

6.10 EXAMINATION OF A HASTELLOY N
FOIL SAMPLE EMBRITTLED IN THE
MOLTEN-SALT REACTOR EXPERIMENT

R. E. Clausing L. Heatherly

Surveillance specimens of Hastelloy N removed
from the MSRE in late 1969 were observed to form
shallow intergranular surface cracks when they were
strained in postirradiation tests. Examination of
several components of the MSRE after operation of

. the test reactor was terminated revealed that all

surfaces in contact with the fuel salt were embrittled
to depths of 100 to 250 um. Investigations of samples
from the MSRE and bench tests gave strong support
to the hypothesis that the fission product tellurium
caused the embrittlement by diffusion preferentially

.along the grain boundaries where it caused grain

boundary embrittlement.”” However, it was not

possible at that time to provide conclusive direct
evidence that tellurium caused the observed embrit-
tlement.

A piece of Hastelloy N foil about 100 um thick
which had been attached to one of the surveillance
assemblies was severely embrittled by exposure in the
MSRE; this occurrence provided an opportunity to

. use recently developed high-resolution Auger elec-

tron spectroscopic techniques to determine the
composition of the grain boundaries in some detail,
as described below.

“Monitor‘ing was done with the close cooperation of R.L.
Manning, Analytical Chemistry Division.

88

6.10.1 Sample History

A piece of thin Hastelloy N (heat 5075) sheet was
attached to a strap on the fourth group of surveil-
lance specimens which was exposed to the MSRE
core 7203 hr at temperatures above 500°C. The
material was rolled to a thickness of 100 um and
annealed 0.5 hr at 1180°C prior to insertion in the
MSRE. The chemical composition of the material is
given in Table 6.16. The total thermal neutron
fluence is estimated to have been 5.1 X 10%
neutrons/cm’, and a detailed account of the thermal
and radiation history of this group of surveillance
specimens can be found in ref. 22. About 6000 hr of
the exposure was at 650°C, and 3000 hr was in
contact with fuel containing fresh fission products:

The foil was wrapped around a thicker Hastelloy N
strap to hold it in place; thus the outside of the foil
was exposed to flowing salt and the underside to salt
that flowed less rapidly. The foil was severely
embrittled -and broke while being removed from the
strap. Micrographs of the foil are included in ref. 22.
In some bent areas, cracks formed which often
extended completely through the foil. The sample
used in this study was from a straight unstressed
region which presumably had few, if any, open
cracks. Detailed examination confirmed that the area
examined was indeed a fresh fracture not previously
open to air.”’ The size of the sample wasabout 0.3 cm
wide X 1 cm long X 100 um- thick. The sample
radioactivity was less than 1 R/hr at contact.

6.10.2 Sampie Fracture and Analysis Techniques

The sample was extremely brittle and was
therefore easily fractured by applying a bending force
in the protective ultrahigh vacuum environment of
the Auger analysis system. The residual gas pressure
in the system was below 1 X 107'° torr at the time of
fracture, and the release of helium as the sample
fractured was monitored to provide an estimate of

Table 6.16. Average composition
of MSRE foil sample

Basis
wt % at. %

Element

Mo 16.4 10.5
Ni Balance 75.3
Cr 6.6 7.8
C 0.07 0.25
Fe 4.0 4.4
Mn 0.46 0.51
Si 0.58 1.26

the concentration of helium in the grain boundaries.
The system used for these studies has been described
previously.”*

The Auger analysis of the sample began within a
few minutes after the fracture to avoid errors due to
absorption of gases on the fracture surface. An
electron beam about 5 um in diameter with a current
of ~0.5 uA at 4 to 10 keV was used for these
measurements.

6.10.3 Observations

Figure 6.37 shows the helium partial pressure in
the system during the time of the fracture. Under the
conditions of this measurement the initial increase in
helium pressure times the volume of the system gives
the amount of helium released. The exponential
pressure decay after the fracture is simply the
characteristic system pump down for helium after
admission of the gas pulse and is in agreement with
the previously measured rates. The amount of gas
released was enough to fill 3.4% of the grain
boundary sites exposed if we assume a surface
roughness factor of two (i.e., the actual area of the
fracture is assumed to be twice as large as the cross
section of the sample since the fracture surface is not
smooth and planar).

Figures 6.38 and 6.39 show scanning electron
micrographs of the fracture surfaces and identify the
areas analyzed. The fracture was completely inter-
granular and along grain boundaries filled with

FRACTURE

4);‘!0-H

He PARTIAL PRESSURE (Torr)

89

carbide precipitate particles. There is no evidence of
deformation in these micrographs. The fracture is
apparently along the interface between the matrix
and the carbide precipitates rather than through the
carbides. This was confirmed by the change in the
composition as a function of depth below the
exposed fracture surface.

Several types of Auger analyses were made
including (1) two-dimensional elemental maps,
which can be directly compared with scanning
electron micrographs; (2) line scans, which are
graphic outputs of the Auger signal strength as a
function of position along a line across the sample;
and (3) complete Auger spectral analysis in a selected
area of the sample surface. These measurements
become progressively more quantitative in the order
listed. Reference 23 has a complete set of the maps,
line scans, and point analyses made on this sample.
Figure 6.40 is a composite of several maps showing
the distribution of tellurium, nickel, and molybde-
num on the as-fractured surface. These images are of
the area shown in Fig. 6.38a. Topographical features
may be identified in the absorbed sample current
image and easily related to the SEM images. These
SEM images provide information on the microstruc-
tural features of the fracture surface. The topographi-
cal information in the elemental maps helps to orient
the map relative to the absorbed sample current and
SEM images but makes the map less quantitative.
Nevertheless the strong, uniformly intense tellurium

ORNL-DWG 76-7597

o

|-—20 sec——|

TIME

Fig. 6.37. Helium release during fracture of the MSRE foil sample. The amount of helium released is estimated to be equal to
about 3.4% of the grain boundary sites exposed during the fracture.

90

Fig. 6.38. Scanning electron micrographs of part of the fracture surface of the MSRE sample. (2) The 200X magnification shows
a particularly interesting area where one grain goes entirely through the thickness of the foil, and the exposed surface is nearly flat
and perpendicular to the original foil surface. The numbers on the figure indicate areas investigated in detail and shown at higher
magnification — at 1000X in parts (b), (c), and (d) and at 5000X in parts (¢), (/) and (g). Note that these images reveal that the grain
boundary is lined with small particles (likely carbides) and that they appear to almost completely fill the grain boundary area. The
fracture appears to be simply a separation of the particles from the matrix with little or no ductile behavior.

signal from the entire fracture surface of this sample
is evident. The molybdenum and tellurium maps
show similar distributions, while the nickel map
shows a distribution somewhat complementary to
that of molybdenum and tellurium. This is best seen
in the central region slightly to the right and left of the
area labeled 1. The similarity of the tellurium and
molybdenum distributions and the complementary
nickel distribution is easily seen in the line scans
shown in Fig. 6.41. These scans are along the path
indicated by the vertical white trace in Fig. 6.40.
Similar relationships have been noted among other
samples.

Complete Auger analyses of several selected spots
were made immediately after the sample was frac-
tured. Figure 6.42a shows the Auger spectrum from
the center of the sample in area 1 shown in Fig. 6.38,
and this was also typical of spectra taken in areas
2, 3, and 4. The strong tellurium Auger signals are
evident. The compositions (in atomic percent) in
areas | through 4 are listed in Table 6.17. Figure 6.42
compares the Auger spectra taken from area 1
immediately after fracturing the sample and the
composition at the same area taken after sputter
etching was used to determine the distribution of
tellurium as a function of depth below the original

91

| 75 150 MCRONS, 450 575
- o
5008 | ' INCHES | D.015

MICRONS A < © 5000x 2

L—1000X L
0.001 INCHES 16 MICRONS

Fig. 6.39. Scanning electron micrographs of a part of the fracture surface of the MSRE sample immediately to the left of the
area shown in the previous figure. (a) The 160X magnification shows a more typical cross section through the foil with two grain
diameters across the foil thickness. Higher magnification images of area 4 are shown in parts (b) and (c), at 1000 and at 5000X
respectively. They show a somewhat rougher fracture surface than that shown in Fig. 6.38. In this case, some of the particles appear
to remain on the observed surface while others are partially pulled out and still others appear to have adhered to the other half of the
fracture interface. This would be expected when the adhesion of the particles is more or less equal between the grain on either side of
the boundary. This adhesion is quite likely a function of crystallographic morphology, depending perhaps on the details of tellurium
embrittlement as well as the mechanical properties of the embrittled interface. None of the particles appear to be fractured; instead,
they have separated from the matrix. The particles in this region appear to be more or less equiaxed, while those shown in Fig. 6.38
are more platelike. Note especially the comparison between Fig. 6.39¢ and Fig. 6.38¢ In Fig. 6.38g, some of the flat particles have
been pulled out, leaving shallow flat-bottomed depressions

92

¥-134216 ¥-134217

Te (483 eV)

v-134218 ¥-134219

Fig. 6.40. This composite shows (a) an absorbed sample current image of the region shown in Fig. 6.38 and elemental maps of

this region obtained using Auger electron images for (b) tellurium (483 €V), (c) nickel (848 V), and (d) molybdenum (221 eV). All

\ images are at about 250X. Note the rather uniform tellurium composition over the entire surface, the superposition of topographical
features on the image, and the somewhat complementary nature of the tellurium and nickel maps in the regions to the right and left

of point 1

‘ Ni (848 eV) Mo (221 eV)

ORNL-DWG 76-7599

SPECIMEN EDGES ——|

INTENSITY —

DISTANCE —=

Fig. 6.41. This composite shows Auger signal intensity along the line shown in Fig. 6.40a. The scan proceeds from top to bottom
of the image and from left to right on the figure. The intensity of the signals is modulated by both topography and composition so
that both features appear in these plots; however, it is quite apparent that the tellurium and molybdenum signals vary in the same
general way, while the nickel signal contains the same topographical features but is somewhate complementary to the tellurium and

molybdenum scans.

ORNL-DWG 76-7598

(@)

(o)

ON(EY/dE (arbitrary units)

/MH
\

o

Mo

Mo

SPOT {-FRESH FRACTURE

Te

SPOT{-SPUTTERED 5 min

Ni

ENERGY ———=

Fig. 6.42. Auger spectra taken from the center of the area of the sample shown in Fig. 6.37 near the spot marked 1. The upper
spectrum labeled (¢) was obtained immediately after the sample was fractured, while (b) was taken after sputtering about five atom
layers from the exposed fracture surface. The tellurium peaks are quite prominent in spectrum (2) and represent about 14 at. %
tellurium on the surface, but tellurium is almost undetectable in spectrum (b), indicating that the tellurium was nearly completely
removed by the ion etching of only five atomic layers from the surface. The other elements remain high. The molybdenum and
carbon actually increase, as they should if bulk carbides were uncovered by removing the tellurium.

fracture surface. Table 6.18 is a tabulation of data
taken from several spots durihg_ sputter etching, and
Fig. 6.43 isa graphical representation of the changes
in composition near spot 4.

6.10.4 Discussion of Observations

The release of helium from this sample when 1t
fractured was estimated to be sufficient to fill about
3.4% of the grain boundary sites exposed by the
fracture. The most likely sources of this helium is the
"B(n,a)’Li reaction. Boron is present in the grain
boundaries of the Hastelloy N as the result of the
melting practice and heat treatment of the alloy prior
to irradiation and will also be concentrated in the
carbides formed during irradiation. This amount of

helium should not be enough to cause the observed
intergranular embrittlement at room temperature. If
it were, severe embrittlement would have been found
throughout the entire thickness of surveillance
samples and other Hastelloy N parts from the
MSRE. This was not the case, and it can be
concluded that this level of helium in the grain
boundaries is not severely detrimental to the
mechanical properties at room temperature.

The fracture surface was completely intergranular.
There was no reduction of area at the fracture site,
and no significant plastic deformation was apparent
in the scanning electron micrographs. The appear-
ance of the fracture surface suggests that the grain
boundaries were filled with precipitates and that the
Table 6.17. Auger electron spectrographic (AES)
analysis of MSRE foil sample?

Spot Flement (at. %)¢
number? e Mo Ni Cr C
1 126 325 298 44 20.7
2 9.0 175 46.8 5.9 20.7
3 10.6 182 43.8 5.1 22.4
4d 12.3 6.8 546 58 20.5

9Grain boundary regions exposed by fracture in
AES system.

bsee Fig. 6.38.

CValues are derived using technique and sensitivities
from Handbook of Auger Electron Spectroscopy,
Physical Electronics Industries, Inc., 1974, and nor-
malized so that the reported values total 100% during
analysis.

dCarbon adjusted to compensate for contamination
during analysis.

- Table 6.18. Auger electron spectrographic
analysis of sputtered MSRE foil sample

_Sputter time Element (at. %)

Area \
(min) Te Mo Ni Cr C
1 126 325 298 44 207
1 5 0.6 31.0 21.0 1.3 462
1 20 37.0 209 1.6 40.6
1 60 26.7  20.8 48.1
1 150 269 191 21 51.3
2 9.0 175 469 59 207
2 20 222 356 3.3 389
2 60 19.3 27.8 26 49.1
2 150 16.1 34.1 3.4 446
4 12.3 6.8 546 5.8 20.5
4 5 0.5 166 49.1 4.4 293
4 20 219 456 3.2 29.3
4 60 131 40.7 3.1 433
4 150 159 406 2.6 404

fracture occured between' the precipitates and the
matrix. Figures 6.38¢ and g show areas where most of
the particles were retained in the surface examined.
These areas are rich in molybdenum and carbon
(Table 6.17). In some areas, some particles have
apparently been pulled from the surface examined
(Figs. 6.38f and 6.39¢). Areas such as these tend to
have somewhat less molybdenum and carbon and to
be richer in nickel; hence the complementary nature
of the elemental images and line scans reveals the
presence of carbides and shows the surfaces to which
they adhere. In no case have we seen fractured

94

ORNL-DWG 763.;1231

MOLYBDENUM )
CAREON 40 at%
5 16 at %
[
EI
= CHROMIUM
= N
Q S
21 at%
ST~
TELLUKIUM
k | | | ] | |
0 o5 50 75 10 125 150
SPUTTERING TIME (minutes)

(APPROXIMATELY 1 ATOM LAYER REMOVED /MINUTE)

Fig. 6.43. Auger signal intensity as a function of sputtering
time. These curves show the change in-concentration of several
elements as a function of depth below the original fracture
surface. The numbers at the 150-min point give the estimated
atomic percent at a depth estimated to be 150 atomic layers
below the original fracture surface. These data are from the area
near point 4 on Fig. 6.39,

precipitate particles. There is no doubt that these
precipitates are carbides since the fracture surfaces
are rich in molybdenum and carbon and the
concentrations of these elements remain high after
sputter etching (Table 6.18 and Fig. 6.43) which
indicates an average bulk composition in the vicinity

- of the fracture surface consistent with the presence of

large amounts of carbide precipitates. The presence
of the carbides in the boundaries reduces the ductility
and tends to promote intergranular fracture, but not
to the extent observed here. Many samples of
standard Hastelloy N, including both irradiated and
unirradiated control samples, have grain boundaries
containing carbides and still retain elongations
greater than 309% and reductions in area greater than
20% (ref. 22). Many of the samples exposed in the
case of the MSRE show extensive intergranular
surface cracking while maintaining good overall
ductility, a fact which further supports the argument
that the carbide precipitation in the grain boundaries
1s not the cause of the observed embrittlement. This
last observation also argues that a combination of
carbide precipitation and helium generation does not
cause the observed embrittlement. :

The Auger electron spectrum from the fracture
surface, Fig. 6.42a, shows the presence of large
amounts of tellurium. The concentrations reported in

‘Table 6.17 are based on the assumption of uniform
composition in the substrate. In the case of a
monolayer or less of contaminant on an interface that
is torn apart during the sample fracture, the
concentrations obtained would be too small by more
than a factor of 2. In this case the real concentration
at the interface could be substantially greater than 30
at. %. The composition-depth profile in Fig. 6.43
shows that the very large tellurium concentration on
the as-fractured surface does not extend into the
bulk; thus, tellurium is present in large amounts on
the fracture surface but 1s concentrated very strongly
at the grain boundary interfaces and may indeed be
present as a fractional monolayer at the grain
boundary. It is known from previous experiments”*
that grain boundaries in Hastelloy N can be
embrittled by tellurium at levels below those
observed here. No other fission products or other
embrittling agents were detected on the fracture
surface, and one must conclude that tellurium does
cause the observed intergranular embrittlement of
Hastelloy N in the MSRE.

6.11 HIGH-RESOLUTION FRACTOGRAPHY
OF HASTELLOY N ALLOYS
EXPOSED TO TELLURIUM

D. N. Braski

This investigation is aimed at providing additional
information concerning the mechanism of tellurium
embrittlement of Hastelloy N alloys. Of particular
interest were the fractographic details of samples
exposed in several different ways to tellurium. These
details were revealed by use of a modern transmission
electron microscope with a high-resolution scanning
attachment. The results for two standard Hastelloy N
(heat 405065) samples are presented first, followed by
the results for two modified alloys. Although
different delivery systems were used, the same
tellurium source, CriTes at 700°C, was used in all
tests. After exposure to tellurium the samples were
tensile tested to failure at room temperature. During
the test a number of cracks opened up along the gage
lengths of the samples. The number and depth of
cracks were measured optically on metallographi-
cally prepared sections. Table 6.19 lists the alloy, heat
treatment, and tellurium exposure for each sample.
Details of the exposure tests may be found in Sects.
6.9 and 6.12.

The first sample, 7943, was actually a smaller
Auger spectroscopy specimen that was exposed for
190 hr to molten salt saturated with Cr;Tes. A high-
temperature solution anneal was previously done on

95

‘mentioned previously, Rellicketa

this sample to grow large grains for the Auger studies.
After testing, the fracture mode was observed at
relatively low magnification to be intergranular (Fig.
6.44a). Tellurium was detected in area X by both
Auger and energy-dispersive x-ray spectroscopy
(EDS). Higher-magnification micrographs of area X
are shown in Figs. 6.44b and ¢. The nodular features
present on the surface were similar to those observed
in other materials that failed from stress corrosion or
hydrogen embrittlement.”® These nodular features
were probably corrosion products that contained
tellurium. Several faceted areas were also observed at
points A and B in Fig. 6.44¢. Faceting is characteris-
tic of brittle failure and has previously been observed
in iron-tellurium alloys.”’” Figure 6.44d is a higher-
magnification fractograph taken of area Y, in which
no tellurium was detected. This area displayed some
ductility during fracture, as evidenced by the dimpled
surface caused by microvoid coalescence. The
microvoids were initiated by the particles located at
the bases of the dimples. These particles were
probably the MsC-type carbides that are usually
found in the grain boundaries of this alloy.

Standard Hastelloy N was also exposed to the
vapor above Cr;Tes at 700°C in a closed system (see
SN 13515 in Table 6.19 for conditions). The fracture
mode for this sample was complicated and difficult to
analyze. The central areas were dimpled (Fig. 6.45q).
Areas near the edge of the sample showed mixed
modes of fracture (Figs. 6.45b, ¢, and d). Near the
specimen edge in Fig. 6.45b, the first area resemblesa
cleavage fracture with small steps. For the area
further in, the mode shows characteristics of ductile
fracture as evidenced by the dimples. At the base of
these dimples is unusual microfaceting (Fig. 6.45¢, at
X and what appears to be a second-phase or cor-
rosion product (Fig. 6.45¢ and d, at Y. This latter
area also shows an unusual geometric faceting.
Other, more conventional looking particles were
found in the same area (Fig 6.45b at Z); these may be
MC-type carbides or possibly corrosion product.
Information on the composition and: structure of
these phases has not yet been obtained. Initial EDS
analyses did not detect any tellurium in any of these
areas, but that does not preclude the possibility of
small concentrations being present.

Faceting has now been observed in both standard
Hastelloy N samples using two different tellurium
delivery systems and therefore may be a characteris-
tic of tellurium embrittlement in this material. As
1.’ found extensive
faceting in an iron-tellurium alloy which was cooled
slowly from an austenitizing temperature of 925°C.
Table 6.19. Tellurium exposure conditions and tellurium susceptibility

Exposure Exposure Average
Specimen Heat Solution . P posu Cracks crack
Tellurium exposure time temperature
number number anneal {number/cm) depth
(hr) cO
()
7943 405065 2hrat 1260°C Molten salt 190 700 Auger sample — not analyzed
(LiF + BeF;3 + ThF4) + Cr3Tey for number and depth of cracks
SN 13515 405065 1 hrat1177°C Cr3Teq vapor 1000 700 94 44.5
SN 15728 413 1hrat 1177°C Molten salt + CryTeq 243 700 102 374
(1% Nb + 1% Ti)
SN15725 411(1% Nb) 1hrat1177°C Molten salt + Cr3Teq 243 700 17 30.5

96
97

¥-137724

20 um

Y-137628

Fig. 6.44. Fractograph of standard Hastelloy N (heat 405065) after 190 hr in molten salt + Cr3Te and tensile testing at room
temperature. Parts, b, and ¢ were taken near the fracture edge, and part d is a middle region

98

Y-137726

Y-137620

Fig. 6.45. Fractographs of standard Hastelloy N (heat 405065) after 1000 hr in Cr3Te, vapor at 700°C and tensile testing at
100m temperature. Part  is a middle region, and parts b, ¢, and d are views at the fracture edge

However, Henry et al.”* observed faceting in nickel

fractures after thermal treatments between 650 and
800°C. There was no tellurium present in the system
of these workers; they attributed the faceting to the
absorption of oxygen in the grain boundaries. Since
decarburized layers have been observed in some of
our Hastelloy N samples, it is possible that oxygen as
well as tellurium influenced the fracture mode near
the surface. Other evidence in support of this idea is
presented in Fig. 6.46. This micrograph was taken
from a transverse section of pilgered (tube-reduced)
29 titanium-Hastelloy N which had been heated at
538°C for 6 hr and air cooled. This particular
treatment caused fire cracking of the material. Note
that the mode of cracking through the decarburized
layer was mixed transgranular and intergranular,
while it was entirely intergranular beyond the layer.
The difference in fracture in this case was probably
caused as much by the absence of carbides in the
grain boundary as by the presence of absorbed
oxygen. Thus, decarburization can be a factor in the
fracture mode of thermally treated Hastelloy N and
should be considered when interpreting the present
investigation.

The fractures of two modified Hastelloy N alloy
were compared. The first, alloy 411, contained | wt %
niobium and was considerably more resistant to
tellurium attack than the second, alloy 413, which
contained | wt % niobium plus 1 wt % titanium

(Table 6.19). Both alloys demonstrated intergranular
fracture modes in the central portions of the samples
(Figs. 6.47a and b). One normally associates low
ductility with intergranular fracture, but alloys 411
and 413 had total elongations of 54.8% and 45.8%
respectively. The fracture surfaces illustrate this
ductile behavior by the dimpling on surfaces normal
to the tensile load and striations on the surfaces
parallel to the tensile load. The striations are
characteristic of an extended slip process or glide.
The areas more affected by the tellurium were at the
outside edges. Figure 6.48a shows fractographs of an
area near the edge of the 411 sample. The fracture
mode was also intergranular, but the surface features
reflect a more brittle behavior. Higher magnification
of this area showed the absence of dimples and the
presence of corrosion products (Fig. 6.48h). The tiny
holes in this surface (Fig. 6.48¢) were created by the
pullout of carbide particles or possibly a corrosion
product. Tellurium was not detected on the surfaces
in Fig. 6.48 using EDS. Examination of alloy 413
near the outside fracture edge revealed a tendency for
this material to fail along the precipitate-matrix
interface of the grain boundary (Fig. 6.49). This series
shows an array of dendritic precipitates lying on the
grain boundary which were aligned preferentially in
crystallographic directions. The angles between
dendrites were not measured accurately, but they
would appear to correspond to the angles between

v-135214

| Decarburized
Zone

Fig. 6.46. Transverse section of 2% titanium-modified Hastelloy N annealed at 538°C for 6 hr and air cooled. Firecracking is due

to thermomechanical treatments.
Y-137718

100

Y-137717

Fig. 6.47. Fractographs from central region of alloys 411 (a) and 413 (b) fracture surfaces after 243 hr in molten salt + Cr3Teq

and tensile testing at room temperature.

intersecting (111) planes. These are the habit planes
for the MC-type carbides which are found in the
alloy, where (111) MC|| (111) matrix. Evidence
suggests an epitaxial relationship between the
dendrite particles and the matrix (Fig. 6.49a). The
dendrite particles were found to end abruptly at the
twin boundary. Close inspection of the twin band
showed that similar particles were originally present
but were retained in the mating fracture surface. The
orientation of the peaks of the dendrite pyramids in
the twin band was opposite to those in the adjacent
area (Fig. 6.49b). Since no tellurium was detected by
EDS in any of the areas in Fig. 6.49, one might
presume that only extremely small amounts of
tellurium are needed to weaken the grain boundaries.
This conclusion agrees with Auger spectroscopy
studies®” that find tellurium in small quanties (<10 at.
%) and monolayer thicknesses on tellurium-exposed
Hastelloy N fractures.

In summary, high-resolution scanning electron
microscopy has characterized the fracture surfaces of
several Hastelloy N samples which had been
embrittled by exposure to tellurium at 700°C. The
central regions, which were apparently not affected
by the tellurium, generally demonstrated a highly
ductile intergranular failure. Regions near the outer
edge (the specimen surface) were affected by the

tellurium exposure and failed in a brittle intergranu-
lar mode. Minute features of the fracture surface in
this latter region varied among alloys; these features
included microfaceting, cleavage, and a number of
unusual second-phase particles and corrosion prod-
ucts. The observation of faceting in Hastelloy N is
particularly interesting because it has been seen by
other investigators in iron-tellurium alloys. The
effect of oxygen, either absorbed in the grain
boundary or involved in decarburization, may also
have influenced the fracture mode at the outer edge.
Tellurium was detected by EDS on the fracture
surfaces of one 405065 sample exposed to molten salt
plus Cr;Tes, but was not found in the same alloy
exposed to CriTes vapor or the tetanium-modified or
titanium plus niobium-modified alloys.

6.12 METALLOGRAPHIC EXAMINATION OF
SAMPLES EXPOSED TO TELLURIUM-
CONTAINING ENVIRONMENTS

H. E. McCoy B. McNabb J. C. Feltner

Samples of modified Hastelloy N were exposed to
tellurium-containing environments. These samples
were deformed to failure at 25°C, a procedure which
forms surface cracks if the grain boundaries are

101

Y-137719

Fig. 6.48. Fractographs of alloy 411 (Hastelloy N modified with 1% titanium) after 243 hr in molten salt + Cr3Teg followed by
tensile testing at room temperature.

102

Fig. 6.49. Fractographs of alloy 413 (Hastelloy N modified with 1% niobium plus 1% titanium) after 243 hr in molten salt +
Cr3Te, followed by tensile testing at room temperature.

brittle; a metallographic section of each sample was
prepared to permit determination of the extent of
cracking. These tests have two objectives. The first is
to develop a method for exposing samples to
tellurium to produce a reaction rate comparable to
that anticipated for an MSBR. This rate is thought to
be a flux of tellurium of about 10" atoms cm ™ sec .
The second objective is to compare the cracking
tendencies of various alloys of modified Hastelloy N.

The method used for crack counting and measur-
ing was described previously, and data were
presented for .several alloys.”® The experimental
conditions associated with the experiments to be

103

discussed in this report are summarized in Table 6.20.
The chemical compositions of the alloys studied are
givenin Table 6.21. In all cases the sample wasa small
tensile specimen /s in. in diameter by 1.78 in. long
having a reduced section s in. in diameter by 1 in.
long. Unless specified otherwise in Table 6.21, the
specimens were annealed 1 hr at 1177°C in argon
prior to exposure to tellurium. The results of crack
measurements and data resulting from tensile testsat
25°C which were used to open the embrittled grain
boundaries are shown in Table 6.22.
Photomicrographs of the samples from experi-
ment 75-11 show that the extent of reaction for these

Table 6.20. General description of tellurium-Hastelloy N exposures

Exl:-»erim-ent Experimenters Exp9§ure . Alloys General .
designation conditions included comments
75-11 McNabb 100 hr at 700°C in vapor 405065,471-114,474-534, 474-535,
McCoy above Te at 300°C 600600
75-12 McNabb 250 hr at 700°C in vapor 405065 with different grain sizes Specimens discolored
McCoy above Te at 300°C
75-13 McNabb 1000 hr at 700°C in vapor 405065,421543, 295, 298, 345, 348,
McCoy above Te at 300°C 411,413,421,424,425(2)
75-14 Keiser 500 hr at 700°C in salt + 405064,470-835 Thin reaction layers
Cr,Te,
75-15 Keiser 243 hr at 700°C in salt + 411,413 Thin reaction layers
Cr,Te,
75-16 McNabb 1000 hr at 700°C in vapor 405065, 600600, 471-114,474-533, Reaction layers, some spalling
McCoy above Cr, Te, at 700°C 474-534,474-535, 237, 295, 298,
' 303, 345, 348, 417,421,422,
424
75-17 Keiser 649 hr at 650°C in salt + 405065
' LiTe,
75-18 McNabb 1000 hr at 700°C in vapor Ni, 30488, Incoloy 601, NX 6372G, Little evidence of reaction
McCoy above Te at 300°C NX 7187G, NX 7353, 405065,
421543,472-186,474-533,
474-534,474-535,474-539,
474-557,474-558,474-901,
475-421, 62, 63, 234, 295, 296,
297, 303, 306, 345, 348,411,
413,417,422, 456,460, 462, 503,
504, 505, 506
75-19 Keiser 500 hr at 700°C in salt + 503, 504, 505, 506
Cr, Te, '
75-20 Keiser 1000 hr at 700°C in salt + 405065
Ni, Te,
75-21 Keiser 500 hr at 700°C in salt + 405065
Cr,Te, (half in salt, half
in vapor)
75-22 Brynestad 2500 hr at 700°C in vapor 405065, 295,424
above Ni + Ni, Te,
75-23 Brynestad 2500 hr at 750°C in vapor 405065, 295,424
above Ni+ Ni, Te,
75-24 McNabb 287 hr at 700°C in vapor 450,513,574, 515,516,517,
McCoy above Te at 300°C 518,523,524

%Where no comments are given, there were no visible reaction layers.
104

Table 6.21. Chemical analyses of alloys used in tellurium cracking studies (wt %)

Heat number Mo Cr Fe Mn C Si Ti Nb Al QOther
62 11.34 7.52 a 020 0.042 0.01 a 1.9 a
63 11.45 7.33 a 020  0.135 0.01 a 2.5 a
180 11.2 7.0 0.040 022 0.046 001 <002 184 &
181 11.5 6.84 0.054 0.23  0.045 0.01 0.50 185 a 0.03W
234 16.0 1.2 40 <002 0.05° 0.13 <002 <005 <005 0.75Ce
237 12.0 6.7 43 049  0.032 a 0.04 1.03 <0.05
295 11.4 8.06 4.02 028 0057 <0.02 <0.02 085 a 0.05 W
296 11.5 8.09 3.96 028 0059 <002 <0.02 1.2 a 02 W
297 1202 70° 4.0° 02° 0.6 0.02 0.24 057 4
298 12.0° 7.0° 4.0° 02° 006 002 <001 2.0 a
303 12.0° 7.0° 4.0% 02%  0.06° 0.02 0.49 084 a
304 12.0 6.9 5.0 0.2 0.06° 0.05 <001 <005 <001 O08W
305 11.2 8.25 4,16 022 0.072 0.09 0.88 1.3 a
306 10.6 8.04 3.11 0.18  0.065 0.27 0.01 055 a
345 11.0 7.1 3.8 026 005 0.22 0.02 045 a
346 11.0 6.7 3.7 0.18  0.05° 0.48 0.02 049 4
147 12.0 7.6 4.3 025 0.0s° 047 <0.02 088 a
348 12.0 7.2 0.07 0.19 0.0s° 047 <0.02 062 a4
411 11.71 6.78 a 0.1 0.043 p a 1.15  a
413 11.82 6.75 a 0.1 0.045 a 0.9 113 a4
417 11.48 6.82 0.1 0.1 0.056 a 0.01 229 a4
421 11.43 6.75 a 0.2 0.048 a 1.9 1.04 007
422 11.27 6.78 0.09 0.15  0.059 a 1.0 229 007
424 11.32 6.76 a 0.15  0.063 a 1.8 1.3¢ 010
425 11.42 6.72 a 0.15  0.037 a 19 048 008
450 13.6 8.1 0.3 <0.1 0.051 a 2.1 <0.1 a
456 120 7.0° a 0.2  0060° o 2.1 1.03 018
460 120 100° a 02°  0.060° o 2.36 121 017
462 12.0° 150 a 02> 0060 4 2.21 1.0 0.16
503 11.92 7.40 a 0.2 0.029 a <0.01 <002 ~0.
504 11.84  10.2 a 0.17  0.029 a <0.01 <0.02 ~0.1
505 12.08 11.9 a 022 0.034 a <0.01 <002 ~0.1
506 1197 145 a 022 0.037 a <00l <002 ~0.1
513 1140 7.2 <0.05 0.20 0040 <002  <0.02 045  0.05
514 11.50 7.05  <0.05 020 0035 <002 <0.02 069  0.05
515 11.53 7.20 <0.05 020 0038 <002 <0.02 020 005
516 10.58 730  <0.05 0.20 0.049 <002 <002 <010 0.05
517 11.45 7.10  <0.05 <001 0045 <002  <0.02 110 003
518 11.55 7.18  <0.05 020 0.040 0.05 0.95 096  0.10
523 11.40 1040  <0.05 023 0037 <002  <0.02 047  0.10
524 1165 1520 <0.05 0.17 0040 <002 <0.02 069  0.10
Ni-280¢ 70 2 10 5 15 <1 5 <03 100 Ta
30455-139969 18.50 Balance  1.51  0.027 0.45 9.3 Ni
Inconel 601 21.91°  15.01° 0.15¢  0.03¢ 0.26° 0.31°¢ 1.32° 0.03Cu°
Inconel 600 15.5% 8.0” 0.5  o0.08? 0.25°
600600

NX 6372G 1513 9.17° 0.27°  0.10° 0.09¢ 0.31 Cu®

NX 7187G 15.82° 1.57°¢ 0.11° 0.07¢ 0.17¢ 0.28 Cu®

NX 7353 15.04°  9.44° 0.38° 0.07° 0.14¢ 0.31 Cu®
421543 12.4 7.31 0.038 0.08 0.05 0014 0003 0.70  0.02
405065 16.0 7.1 4.0 0.55 0.06 057 <0.01 a <0.03
469-344 13.0 74 4.0 0.56 0.11 a 0.77 1.7 - a 0.019 Zr
469-648 12.8 6.9 0.30 0.34  0.043 a 0.92 195 &
469-714 13.0 8.5 0.10 0.35  0.013 a 0.80 160 2
470-786 12.2 7.6 041 043  0.044 p 0.82 062 a 0.024 Zr
470-835 12.5 7.9 0.68 0.60  0.052 a 0.71 260 a4 0.031 Hf
471-114 12.5 74 0.062 0.02 0.058 0026 1.5 a 0.07
472-166 14.65° 15.49°  0.56° 0.51° 0.01° 0.27° 0.14° 0.01La‘0.01Bf

0.06 W.£ 0.05 Co°
472-503 12.9 6.79 0.089 <001  0.066 0.089  2.16 0.05  0.09
474-533 11.37 1.3 0.03¢ 0.04 0.091 0.15 1.75 a 0.54  0.14 WS 0.03 Co¢
474534 11.7 7.1 0.06 <001 008 . 003 2.09 a 0.53 0.14W,0.013 La
474535 11.8 7.3 0.05 <001 0.08 0.03 2.13 a 0.55  0.10 W,0.010 La,
0.03 Ce

474-539 11.38 74 0.07° 0.03  0.041 0.10 2.55 a 0.62 0.11 W,0.06 Co,0.02 W
474-557 11.91 7.23 0.08 <001 0.04 0.03 2.14 a 0.02
474-558 12.00 7.20 0.07 <0.01 0.07 0.03 2.05 P 0.02 002W,0.021La
474901 12.5 6.96 0.05¢ 0.02° 0.034 0.10 2.00 a a .
475421 11.93 7.10 0.06 0.12 007 0.04 1.90 a 0.12 0.02Co

2Not analyzed, but no intentional addition made of this element.
Not analyzed, but nominal concentration indicated.

“Vendor’s analysis.
Concentrations in ppm.
Table 6.22. Intergranular crack formation and tensile properties? of samples exposed to

tellurium and strained to faiture at 25°C

Cracks per Depth (1) Standard 95% confidence Weight?  Yield Ultlm;il te Fracture Uniform  Fracture Redl’wtlon
Exp. No. Heat No. unit length ———— deviation interval change stress t::;: stress  elongation  strain a;:a
Cracks/in. Cracksiom ' M @) @) (mg) (107 psi) (¥ o (107 psD (%) %) s
75-11 405065 105 41 203 33.0 6.6 29 0.50 54.2 132.7 126.5 37.3 42.1 40.3
471-114 17 7 13.0 21.6 5.5 5.0 0.29 51.5 118.9 1134 54.1 56.6 47.6
474-534 363 143 225 412 6.5 1.2 -0.51 58.1 129.9 117.5 41.7 44.6 48.8
474-535 0 0 ~1.59 52.8 1249 1194 474 49.4 47.0
600600 67 26 170 27.5 5.3 24 +0.62 39.2 102.9 76.2 29.6 348 61.2
75-12 405065 .
as swaged 370 146 17.8  36.8 5.6 1.9 +0.97 1029 159.8 149.7 26.3 29.0 34.6
405065
1 hr at 982°C 350 138 164 25.8 49 1.6 +0.85 79.9 150.5 132.1 314 34.3 44.1
405065 '
1 hrat 1038°C 330 130 132 271 5.2 1.8 +0.43 62.8 141.3 125.7 348 37.2 44.8
- 405065
1 hrat1093°C 320 126 189 363 5.8 1.7 +0.26 56.9 137.6 125.3 37.0 40.3 43.1
405065 :
1 hrat1150°C 330 130 31.8 59.1 12.7 4.4 +0.34 53.4 1304 122.6 39.5 41.8 48.1
405065 . .
1hrat1177°C 280 110 48.0 92.0 18.0 6.8 +0.29 52.3 1260 1154 40.8 433 40.3
405065
1 hr at 1204°C 200 79 69.6 1074 18.3 8.2 +0.60 50.8 121.6 119.7 374 384 31.6
405065 '
1 hrat 1232°C - 220 87 73.9 130.0 26.5 11.3 +2.06 504 111.3 109.6 333 33.6 26.6
405065 :
1hrat1260°C ‘110 43 87.3 1794 34.4 14.7 +1.96 49.3 104.2 100.9 31.2 315 27.6
75-13 405065 215 85 229 541 9.9 3.0 -0.07 53.6 1324 120.3 39.5 41.7 31.7
295 12 5 10.0 23.7 7.9 6.5 +0.08 534 121.6 112.6 436 46.3 45.3
298 70 28 le.4 383 8.1 2.7 +0.03 54.3 124.6 108.3 45.3 49.2 51.1
345 16 6 9.9 16.2 3.5 2.5 +0.04 59.6 121.8 111.1 39.6 42.7 47.1
348 74 29 85 1758 13.0 4.3 +0.11 52.7 121.9 105.3 43.5 46.8 51.3
411 44 17 124 355 8.5 3.6 +0.30 45.1 114.9 103.5 47.7 49.9 46 .4
413 280 110 13.5 41.8 9.3 2.0 —0.06 53.7 125.3 117.0 45.7 48.0 422
421 380 150 35.1 89.0 20.2 6.6 0.00 53.8 121.3 118.9 44.7 46.7 45.7
424 350 138 369 77.1 15.3 5.2 0.00 55.3 1324 125.0 42.1 44 .4 497
425 130 51 38.2 740 18.5 10.3 +0.15 52.2 124.2 1184 43.2 446 40.8
425 280 110 30.2 65.9 15.2 5.7 0.00 52.6 124.7 118.7 43.1 45.0 41.1
421543 22 9 179 41.8 11.9 7.2 +0.14 519 117.0 106.7 45.0 417.7 50.5
75-14 405065 350 138 504 904 18.9 6.4 +0.44 52.5 124.5 116.2 38.0 40.3 394
' 470-835 247 97 242 94.7 20.3 4.7 -0:90 56.8 136.9 129.8 434 449 39.3
75-15 411 43 17 30.5 66.6 17.5 8.5 —4.13 41.5 112.7 97.9 51.7 54.8 534
413 260 102 374 936 23.6 6.5 —4.18 49.5 1204 113 43.5 45.8 42.7
195 717 29.1 629 13.9 4.5

SOT
Table 6.22 (continued)

Cracks per Depth (u) Standard 95% confidence Weight?®  Yield Ultlm.ate Fracture Uniform Fracture Redl‘lctlon
Exp. No. Heat No. unit length —— deviation interval change stress Ze:::;l: stress  elongation  strain area
Cracksfin. Cracks/em ¥ M2 () () (mg) (10> ps) o3’ o (10°psi) (%) (%) p
75-16 405065 240 95 445 89.9 16.2 4.7 +3.89 525 125.7 119.9 38.5 40.6
600600 115 45 358 589 11.2 4.7 +18.25 34,5 96.2 78.2 31.5 35.3
471-114 225 89 56.8 120.0 25.1 7.5 +57.33 48.6 111.2 102.7 499 526
474-533 305 120 726 1425 376 9.6 +32.90 504 114.8 107.2 48.0 50.7
474-534 380 150 40.5 71.17 18.4 4.2 +17.70 564 124.9 115.0 40.8 43.2
474-535 310 122 574 103.8 24 4 6.2 +35.14 51.8 117.0 106.3 47.1 49.9
237 145 57 57.2 1273 26.0 9.6 +1.28 55.1 118.4 106.0 434 46.0
295 70 28 229 50.8 12.4 54 +8.44 54.2 124.6 113.5 39.9 42.5
298 310 122 35.0 135.0 19.1 4.8 +11.46 53.5 122.3 108.4 46.1 49.0
303 295 116 349 835 14.8 3.8 +11.46 51.8 117.6 105.8 46.5 49.6
345 83 33 60.8 105.2 23.5 8.2 +7.63 50.2 113.5 104.5 436 46.5
348 145 57 68.7 120.5 24.4 9.1 +41.14 47.2 112.3 99.9 45.0 - 48.4
417 155 61 37.6 B5.5 16.9 6.1 +36.91 518 122.4 113.1 426 448
421 205 81 534 898 22.5 7.0 +4.79 52.0 124.3 116.2 45.2 47.1
422 250 98 59.1 122.0 23.6 6.7 +21.72 53.2 125.2 114.8 434 46.1
424 265 104 62.9 113.2 27.0 7.4 +44.12 544 128.8 120.6 45.5 476
75-17 405065 74 29 15.2 28.2 7.0 2.3 51.8 131.9 122.2 41.5 43.7 428
75-18  Ni 280 90 35 10.9 32.0 5.8 1.7 +7.41 13.2 67.0 47.6 35.9 41.2 428 .
304L SS 0 0 +0.48 27.1 86.6 64.9 69.4 74.9 . 73.8
INC 601 10 4 1.0 8.5 1.8 1.6 +0.16 449 105.8 96.9 33.7 353 42.1
NX 6372G 0 0 -0.15 47.7 1174 71.3 234 29.2 68.3
NX 7187G 0 0 —0.02 48.0 118.7 814 26.2 31.1 67.4
NX 7353 0 0 -0.06 46.1 115.2 76.8 25.3 31.2 66.3
405065 24 9 8.0 15.1 3.1 1.8 -0.03 534 133.9 124 4 425 45.0
421543 45 18 9.6 20.4 34 1.6 +1.39 574 133.8 123.9 38.6 40.5 441
472166 10 4 1.0 254 9.8 9.8 -0.41 45.6 99.2 98.4 38.8 38.8 278
474-533 165 65 42.8 758 16.3 5.7 +0.10 53.0 1229 112.3 49.0 52.8 47.4
474-534 270 106 24.3  45.2 9.8 2.7 +0.08 57.6 129.5 116.3 44.6 47.4 49.5
474-535 195 77 31.5 55.5 12.6 4.0 +0.06 53.6 123.2 115.5 46.6 48.9 46.1
474-539 285 112 204 37.7 8.0 2.1 +0.11 56.4 131.1 118.1 439 46.7 479
474-551 260 102 40.8 58.0 10.8 4.2 0.00 58.5 130.9 119.1 43.3 46.1 46.7
474-558 0 0 +0.04 58.1 125.5 115.6 41.6 45.0 47.0
474-901 2 1 5.8 5.8 +0.05 56.5 128.3 118.5 47.6 509 48.4
475421 127 50 139 31.3 7.3 2.0 0.00 51.0 119.0 110.8 50.8 539 47.5
62 22 9 88 17.6 33 20 +0.18 52.8 122.2 114.2 43.0 436 39.3
63 345 136 39.0 100.1 19.8 48 +0.17 54.5 140.0 119.2 319 349 48.3
234 170 67 30.8 74.3 18.0 8.7 +0.30° 57.5 124.9 122.2 34.5 36.6 358
295 130 51 209 161.5 29.3 11.5 +0.03 54.6 127.0 110.7 40.0 436 433
295¢ 13 5 9.0 2145 8.7 7.8 +0.21 55.5 1294 115.0 384 414 478
2959 6 2 10.6 20.8 7.0 6.3 +0.01 544 125.3 110.1 40.6 44.2 484
296 16 6 5.5 8.2 1.9 1.3 +0.15 55.1 124.9 110.7 46.0 48.8 51.3
297 125 49 23.8 1149 23.5 9.4 +0.24 58.5 129.3 111.4 47.1 50.8 52,1
303 37 15 8.3 135 2.8 1.5 +0.26 528 123.2 110.6 458 48.7 51.7

901
Table 6.22 (continued)

Cracks per Depth (u) Standard 95% confidence Weight®  Yield Umm:; te Fracture Uniform Fracture Redl.lct;on
Exp. No. Heat No. unit length ————— deviation interval change stress t;n: - stress  elongation  strain .
Cracks/in._Cracksiom A Max () () (mg) (1% psi) 5 :fsi) (10° psi) (%) (%) if,/,)a
306 143 56 13.1  26.3 58 1.8 +0.14 59.0 131.4 1189 36.0 38.3 45.1
345 6 2 4.8 5.3 0.4 0.5 +0.23 59.2 123.1 109.2 39.9 43.0 48.7
348 6 2 5.6 6.5 0.7 0.8 +0.19 50.1 122.5 114.4 442 45.6 45.7
411 18 7 5.9 9.1 2.5 1.7 +0.24 48.4 119.5 109.1 45.5 47.7 42.1
413 20 8 5.6 8.6 2.0 1.4 +0.12 533 128.2 116.5 46.2 48.7 453
417 12 5 8.8 158 4.7 38 +0.27 55.7 131.5 120.1 41.8 44.1 46.9
422 12 5 19.0 371 12.1 9.9 +0.07 56.4 1359 124.9 44.7 47.5 48.7
456 30 12 10.3  20.1 48 2.8 +0.06 56.7 135.7 1234 46.9 49.8 49.2
460 60 24 6.7 155 2.5 1.0 -0.10 69.3 156.3 150.7 34.5 36.3 40.3
462 74 29 10.3 326 5.8 1.9 —-1.05 111.4 201.2 201.2 24.7 24.7 248
503 6 2 85 116 2.7 31 +0.30 39.4 108.6 95.3 513 54.4 538
504 2 1 9.1 9.1 +0.27 45.9 113.6 102.0 48.2 51.0 53.0
505 2 1 10.8 10.8 +0.20 49.0 116.8 104.6 48.5 52.1 51.0
506 12 5 28.1 88.7 31.5 25.7 +0.24 47.6 114.0 106.3 45.0 47.8 47.7
75-19 503 120 47 356 815 18.6 7.6 37.6 103.5 89.4 51.8 55.2 54.1
504 70 28 4]1.3 144.8 29.8 13.0 -43.5 108.7 55.8 47.7 50.6 521
505 180 71 43.7 919 214 7.1 45.9 110.5 99.9 472 50.1 52.2
506 145 57 489 93.2 22.0 8.2 47.0 110.0 98.3 45.1 47.8 47.5
7520 405065 220 87 19.7 692 14.1 4.3 539 132.1 126.3 39.2 40.8 35.6
75-21 405065 310€ 122 74.8 160.5 35.6 12.8 51.9 123.4 115.7 38.2 40.3 35.6 -
3107 122 624 1223  30.3 10.9 ' _
75-22 405065 290 114 17.6 55.3 11.1 4.1 -0.07 53.7 130.8 122.6 38.4 40.3 40.3
295 135 53 13.2 239 4.9 1.9 -0.08 538 119.6 -~ 109.5 43.2 45.5 426
424 530 209 25.7  73.0 11.7 3.2 -0.09 55.8 133.4 124.2 44 .4 46.8 45.0
75-23 405065 200 79 21.4 145.4 24.7 7.8 —0.13 52.7 130.5 127.9 34.0 35.3 30.4
295 66 26 14.7  39.1 9.2 3.2 -0.20 50.8 121.2 110.6 46.3 49.2 49.2
424 430 169 16.4 404 10.0 3.1 -0.22 579 139.7 129.6 43.6 46.2 46.7
75-24 450 127 50 21.9 584 11.9 3.9. +0.40 50.1 112.1 106.9 55.7 58.5 48,5
513 20 8 108 184 4.2 2.6 +0.47 39.5 105.6 96.4 55.6. 61.6 51.9
514 14 6 134 188 5.2 4.0 +0.38 39.3 105.9 96.2 56.8 59.8 55.1
515 10 4 147 179 3.5 3.1 +0.24 36.5 102.4 91.5 58.1 60.6 55.3
516 34 13 149 440 10.1 4.9 +0.15 40.5 107.4 93.9 51.7 54.0 522
517 6 2 12.7 256 11.2 12.9 +0.26 38.9 103.7 93.0 56.4 58.9 526
518 167 66 9.6 21.0 3.9 1.1 +0.43 50.7 120.0 108.5 515 54.2 51.1
523 40 16 94 21.6 4.5 2.6 +0.33 47.5 112.2 102.7 45.5 47.3 48.4
524 74 29 12.1  20.8 4.7 1.6 +0.48 50.4 116.2 108.6 49.7 52.0 44.5

9Tensile tests run at 25°C at a strain rate of 0.044 min

bTotal weight of specimens, 6.1 t0 6.2 g.

€Polished with Al;O3 powder prior to exposure.

doxidized 10 min at 650°C in air prior to exposure
€Half exposed to liquid salt containing Te.

THalf exposed to vapor above Te-containing salt,

-1

L0T
samples was within expected limits with no visible
reaction products on any specimens (Fig. 6.50). The
crack severity varied among the five specimens, with
the standard alloy ‘(heat 405065) and one
titanium—lanthanum-modified alloy (heat 474-534)
showing extensive cracking and the other three alloys
showing much less cracking. Inconel 600 was
observed previously to be more resistant to intergran-
ular cracking than standard Hastelloy N. The
variations among the three heats of modified
Hastelloy N are probably insignificant.

Experiment 75-12 was an evaluation of the
influence of grain size on the depth and frequency of
intergranular cracking. Samples of standard Hastel-
loy N (heat 405065) were made from Y4~in.-diam
swaged rod and were given various annealing
treatments prior to exposure for 250 hr at 700°C to
the vapor above tellurium at 300°C. The specimens
were then strained to failure at 25°C and sectioned
metallographically to determine the extent of
cracking (Table 6.23). Photomicrographs of the
strained portions of the specimens (Fig. 6.51) can be
compared with etched views of unstrained portions
of the same specimens (Fig. 6.52). Such comparisons
show qualitatively that the frequency of cracking
decreases and the depth of cracking increases with
increasing annealing temperature (or increasing
grain sizes).

These trends may be defined quantitatively (Fig.
6.53). With small grain sizes the number of cracks is
proportional to L™, and at larger grain sizes the
number of cracks is proportional to L™, (L is the

108

mean distance between grain boundaries.) The depth
of cracking goes through a slight decrease with the
first few low-temperature anneals but then increases
with increasing grain size. In the intermediate range
of grain sizes the crack depth (average and maxi-
mum) is proportional to L"*.

The specimens in experiment 75-13 were exposed
for 1000 hr to the vapor above tellurium metal at
300° C. This experiment is compared with a compan-
ion experiment (75-10) which operated for only 250
hr’® (Table 6.24). The comparison is made on the
basis of the product of the crack number and average
depth. The ranking of the alloys on the basis of this
parameter is similar for the two experiments. Alloy
413 looks considerably worse in the longer experi-
ment. There is a very distinct break between the
alloys modified with 2% or less niobium and those
containing various titanium and niobium combina-
tions. The alloys containing only niobium look far
superior to those containing niobium and titanium.
Photomicrographs of these specimens clearly reflect
the beneficial effects of niobium (Figs. 6.54 and 6.55).

Experiments 75-14 and 75-15 exposed specimens
to fluoride salt containing Cr;Tes. Heat 405065, of
standard Hastelloy N, cracked more severely in
experiment.-75-14 than did alloy 470-835, of modified
Hastelloy N (Fig. 6.56 and Table 6.22). Both
specimens showed some evidence of very shallow
surface reaction layers. In experiment 75-15 two
specimens were exposed, one containing nominally
1% niobium (alloy 411) and the other containing 1%
niobium and 1% titanium (alloy 413). The alloy

Table 6.23. Effect of grain size on the depth and frequency of
tellurium-induced intergranular cracking in Hastelloy N
(heat 405065) experiment 75-12

Crack depth (u)
Grain size® Crack 95%
Anneal — frequency . Standard
G 7 (number/cm) Average Maxlmum deviation cqnfidence
interval

As swaged 9.9 10.5 146 17.8 36.8 5.6 1.9
1 hr at 982°C 8.9 12.8 . 138 164 25.8 49 1.6
1 hr at 1038°C 7.9 20.0 130 13.2 271 52 1.8
1 hr at 1093°C 7.6 22.5 126 18.9 363 58 1.7
1 hr at 1149°C 6.1 37.4 130 31.8 59.1 12.7 4.4
1 hrat 1177°C 6.1 38.0 110 48.0 92.0 180 . 6.8
1 hr at 1204°C 51 53.0 - 78 69.6 107 18.3 8.2
1hrat1232°C 4.0 69.5 87 73.9 130 26.5 11.3
1 hr at 1260°C 1.9 160 43 87.3 179 344 147

9Grain size estimated by the intercept method (J. E. Hilliard, “Estimating Grain Size by the Intercept Method,” Metal
Progr. 85(5),99—102 (May 1964). G = ASTM grain size number: L = average linear intercept (u).
109

m
i
7~

7 p ¢ P : g ’ %

o % £ % s .
v-134541

100 200 MICRONS 600 700
100X:

0.005 0.010 INCHES 0020 0.025

Fig. 6.50. Several materials exposed to tellurium (experiment 75-11). (a) Standard Hastelloy N heat 405065; (b) 2% titanium-modified
Hastelloy N, heat 471-114; (c) titanium-lanthanum-modified Hastelloy N, heat 474-534: (d). titanium~cerium-lanthanum-modified
Hastelloy N, heat 474-535; and (e) Inconel 600, heat 600600

Y-134550 Y-134560

Y-134551
®) . i

Y-134563

()

¥-134553

Y-134564

Y-134555

) : T AR . v-134568
¥-134557

Y= = = »

100 200 MICRONS 600 700
L L 100X L TR —
0.005 0.010 INCHES 0.020 0.025

Fig. 6.51. Standard Hastelloy N (heat405065) given the indicated anneal and exposed for 250 hrat 700° C to the vapor above tellurium at
300° (experiment 75-12). Strained to failure at 25° C and sectioned for metallographic examination. (a) As swaged, (b)l hrat982°C,(c) I hr
at 1038°C, (d) | hrat 1093°C, (¢) | hrat 1149°C, (/) 1 hr at 1177°C, (g) | hr at 1204°C, (k) | hr at 1232°C, and (i) 1 hr at 1260°C.

01T
111

100 200 MICRONS 600 700
Q0290 |, M, 4 |
0005 0010 INCHES 0020 0.025

Fig. 6.52. Standard Hastelloy N (heat 405065) annealed at the indicated temperature. Etched with glyceria regia. (a) As swaged, (b) | hr
at982°C, (c) | hrat 1038°C, (d) | hrat 1093°C, (¢) 1 hrat 1149°C, (/) | hrat 1177°C, (g) | hrat 1204°C, (h) 1 hrat 1232°C, and (i) | hrat
1260°C.

112

ORNL-DWG 76-B273R

200
MAX. DEPTH (microns) __-®
e
‘\\ ¢
100 p S
fl AR el ©
fi AN AVG. DEPTH (microns)
5 X TTY
= o /A \\
wl
2 50 / st SN
E pa¥
\ NO. OF CRACKS (cracks/cm)
\, o)
S "4
.~_."? /
20
o //f’
Q
10
10 20 ~ 50 100 200 500
L {(microns)

Fig. 6.53. Variation of crack parameters with the average linear intercept (L) for standard Hastelloy N exposed to tellurium
(experiment 75-12).

Table 6.24. Relative ranking of specimens from experiments
75-10% (250 hr) and 75-13 (1000 hr)

Experiment 75-10 Experiment 75-13
Heat Conqentra.tionb (%) Product of number of cracks _ Product of number of cracks _
number - and average depth Relative and average depth Relative
Ti Nb Other (cracks ) severity® (cracks severity®©
X u X u)
cm cm
424 1.8 1.34 4512 1 5085 2
421 219 1.04 3245 2 5251 1
425 198 048 3198 3 3322 3
405065 2328 4 1938 5
425 1.98 0.48 2157 5 1955 4
298 2.0 713 6 452 7
413 1.0 1.13 336 7 1488 6
348 062 04756Si 209 8 248 8
411 1.15 133 9 215 9
295 : 0.85 106 10 47 12
421543 0.7 : 76 11 155 10
345 045 0.228i 45 12 62 11

9See pp. 108—122, ORNL-5078.
bGee Table 6.21 for detailed chemical analyses.
Severity of cracking ranges from a maximum of 1 to a minimum of 12.

113

(@)

¥-134379

100 200 MICRONS 600 700
1 it L 00X —L 1 L
T T T T
0.005 0.010 INCHES 0020 0.025

Fig. 6.54. Strained specimens from experiment 75-13. Specimens were exposed for 1000 hrat 700° C to the vapor above tellurium metal
at 300°C. (a) Alloy 300, (b) alloy 299, (c) alloy 301, () alloy 292, (e) alloy 302, and (/) alloy 294. As polished

114

m

(a)

“

Y-134385

© i
(@)

(e) Fa

%)

100 200 MICRONS 600 700
il 1 j00X—L i
0.005 0.010 INCHES 0.020 0.025

Fig. 6.55. Strained specimens from experiment 75-13 exposed for 1000 hr at 700°C to the vapor above tellurium metal at 300°C (a)
Alloy 298, (b) alloy 296, (c) alloy 297, (d) alloy 293, (e) alloy 303, and (/) alloy 295. As polished.

Y-134346

(b)

Y-134347

100 200 MICRONS 600 700 20 40 60 MICRONS 12|O 1‘]10
L L 1 400X—1 TR 1 } 1 L1 500Xx— T
0.005 0.010 INCHES 0.020 0.025 0.001 INCHES 0.005

Fig. 6.56. Specimens exposed to salt plus CrsTes for 500 hr at 700° C (experiment 75-14). Hastelloy N (heat 405065): (a) Gage section
and (b) shoulder. Heat 470-835: (c) gage section and (d) shoulder. As polished.

Y-134372

STT
containing only niobilum was more resistant to
intergranular cracking (Fig. 6.57). However, both
alloys formed surface reaction productsand did form
cracks.

In experiment 75-16, specimens of several alloys
were sealed in a capsule with Cr;Tes at 700°C and
exposed to the vapor above the CriTes for 1000 hr.
The appearance of the specimens after removal from
the capsule (Fig. 6.58) and the high weight gains
(Table 6.22) are indicative of the relatively high
reaction rate. Further investigation revealed that this
lot of Cr;Tes had not been sufficiently homogenized
since sufficient tellurium was present to be detected
by x-ray diffraction. Thus the activity of tellurium in
this capsule would have been quite high, resulting in
the observed high reaction rates. The extent of
cracking was quite severe in this experiment, how-
ever there were detectable differences among the
various alloys. The alloys are ranked on the basis of
the product of the numberand depth of cracks (Table
6.25). The alloys containing additions of (.45 to 2%
niobium showed less extensive cracking, as did
Inconel 600 (alloy 600600). Alloys containing
titanium and various amounts of rare earth elements
showed somewhat more cracking. Photomicro-
graphs of the unstressed portions of the alloys show
reaction layers of varying thicknesses on all speci-
mens (Fig. 6.59). Photomicrographs of the strained

116

sections of several of the specimens are shown in Fig.
6.60. The depth and frequency of cracking vary
markedly in the various alloys. ‘

In experiment 75-17 a standard Hastelloy N
specimen was exposed for 649 hr at 650°C to salt
containing LiTes. Electrochemical probes in this melt
failed to indicate the presence of a soluble tellurium
species, but the Hastelloy N specimen was embrittled
sufficiently to form intergranular cracks (Fig. 6.61).

Experiment 75-18 used tellurium metal as a source
of tellurium, as was done in several other experi-
ments; but the tellurium activity was not reproduci-
ble, probably because the temperature of the
tellurium metal at 300° C could not be controlled. In
experiment 75-18, there was little evidence of
reaction of the tellurium with the specimens. The
specimens after exposure (Fig. 6.62) had the same
colorations present before insertion into the capsule,
and the weight changes were extremely small (Table
6.22). The limited supply of tellurium into the region
of the capsule where the specimens were located
probably accounts for anomalous behavior of this
entire group of specimens.

Experiment 75-19 was run to examine the
influence of chromium content on the extent of
cracking in the basic Hastelloy N alloy. Four alloys
containing from 7.4 to 14.5% chromium were
exposed to salt containing CrsTes for 500 hr at

Table 6.25. Ranking of alloys from experiment
75-16 specimens annealed 1000 hr at 700°C
in the vapor above Cr, Te, at 700°C

Product of number of cracks
and average depth

Alloy Concentration (%)
(cmks X p) number  Ti  Nb Other
cm

8712 474-533 1.75
7003 474-535 2.13 0.01 La, 0.030 Ce
6542 424 1.8 1.34
6075 474-534 2.09 0.013 La
5792 422 1.0 2.29
5055 471-114 1.75
4325 421 1.9 1.04
4270 298 2.0
4228 405065
4048 303 049  0.84
3916 348 0.62
3260 237 1.03
2294 417 2.29
2006 345 0.45
1611 600600 15 Cr

641 295 0.85

117

100 200 MICRONS 600 700
1 v L M O ? ?

0005 0010 INCHES 0020 0.025

20 40 60 MICRONS 120 140
500X
0.001 INCHES 0.005
(5) 3
V125437
100 200 MICRONS 600 700
P Lo TR e e A
0.005 0.010 INCHES 0020 0.025
(¢) ¥
V135041
100 200 MICRONS 600 700
t L—L——— 100X—y ey
0.005 0.010 INCHES 0.020 0.025
(d)
y.135442
20 40 60 MICRONS 120 140
; 1 500X —t——Lr—1
N 0.001 INCHES 0.005
100 200 MICRONS 600 700
X

0005 0.010 INCHES 0020 0025

Fig. 6.57. Modified-Hastelloy N specimens exposed to salt plus Cr;Tes for 243 hr at 700°C (experiment 75-15). (a), (b) Unstressed
shoulder and (c) stressed gage section of alloy 411. (d), (¢) Unstressed shoulder and (/) stressed gage section of alloy 413. As polished.

237

40
4065

295 298

118

Y-134977
426 600  471- 4TA- 4Th- AT4-
600 114 533 534 535

Fig. 6.58. Specimens after exposure for 1000 hr at 700°C to the vapor above CrsTes at 700° (experiment 75-16).

700°C. The crack statistics (Table 6.22) and the
photomicrographs (Fig. 6.63) show that there was no
detectable influence of these variations in chromium
concentration on the severity of cracking.

In experiment 75-20, standard Hastelloy N was
exposed to salt containing NisTe; at 700°C. The
crack statistics (Table 6.22) and the photomicro-
graphs (Fig. 6.64) show that the material was
embrittled by the exposure.

In experiment 75-21, standard Hastelloy N was
exposed to salt containing CriTes for 500 hr at
700° C. The specimen was positioned partly in the salt
and partly in the vapor. The sample cracked
extensively, with the cracking being much greater in
the portion exposed to salt (Table 6.22 and Fig. 6.65).

In experiments 75-22 and 75-23, three alloys were
exposed to the vaporabove Ni+ NisTe: for 2500 hrat
700 and 750°C respectively (Fig. 6.66). The cracking
statistics (Table 6.22) indicate several trends that may
or may not be significant. The frequency of cracking
is slightly less for all three alloys when exposed at
750° C compared with 700° C. The depth of cracking
appears greater at 750° C for two alloys and less for
one alloy. A clear trend at both conditions is that
alloy 295, which contains 0.85% niobium, formed
fewer and shallower cracks than the other two alloys
(Figs. 6.67 and 6.68).

In experiment 75-24, nine alloys were exposed for
287 hrat 700° C to the vapor above tellurium metal at
300°C. These specimens all exhibited small weight
gains and shallow crack formation (Table 6.22). The
alloys are ranked in Table 6.26 on the basis of the
product of crack frequency and depth. Alloys 450

and 518 had the poorest resistance to cracking. All
the alloys with niobium additions but not titanium
had good resistance to cracking. No systematic
variation with niobium content was detected,
probably because of the very limited extent of
cracking in this rather short exposure. Alloys 523 and
524, which contained more than the normal 7%
chromium, did not appear particularly good in this
test. Typical appearances of the specimens from this
experiment are shown in Fig. 6.69.

Although these experiments involved several
techniques and methods for exposing specimens to
tellurium-containing environments, they lead to four
important conclusions.

1. Hastelloy N modified with about 0.5 to 2%
niobium has good resistance to embrittlement by
tellurium.

. Hastelloy N modified with niobium and titanium
has poor resistance to embrittlement, indicating
that titanium counteracts the beneficial effects of
niobium.

3. Although Inconel 600, which contains 16%
chromium, exhibited good resistance to tellurium
embrittlement, Hastelloy N modified to contain
15% chromium was not resistant to embrittlement
by tellurium.

. Although a soluble tellurium species was not
detected by electrochemical techniques in any of
the salt-telluride mixtures, standard Hastelloy N
was embrittled in all such systems.

(c)

(8)

Y-138348

(b)

(@)

Y-138270

Y-138344

(h)
100 200 MICRONS 600 700
| B 1 1 L 1 1 1

T T 100X T T ™
0.005 0.010 |INCHES 0.020 0.025

Fig. 6.59. Several materials exposed for 1000 hr at 700°C to the vapor above Cr;Te, at 700° C (experiment 75-16). Unstrained portions
in as-polished condition. (a) Alloy 474-533, (b) alloy 422, (c) alloy 298, (d) alloy 405065, () alloy 348, (f) alloy 345, (g) alloy 600600, and (h)
alloy 295.

Y-138260

Y-138268

[ v-138262

611
(&)

Y-138349

Y-138345

=
100 200 MICRONS 600 700
100 X—1 1 1

T T
0.005 0.010 INCHES 0.020 0.025

Fig. 6.60. Stressed portions of several materials exposed for 1000 hr at 700° C to the vapor above Cr3Teq at 700° C (experiment 75-16)
and strained to failure at 25°C. (a) Alloy474-533, (b)alloy 422, () alloy 298, (d) alloy 405065, (e) alloy 348, (/) alloy 345, (g) alloy 600600, and
(h) alloy 295. As polished.

Y-138277

Y-138261
(@) - e

Y-138531

Y-138532

100 200
1 1

MICRONS i

BCl)O 7(‘)0

T T
0.005 0.010

10 '
INCHES 0.020 0.025

Y-138533

-
.. o« e .
© - L]
20 40 60 MICRONS 120 140
—1— 500X
0.001 INCHES 0.005

Fig. 6.61. Hastelloy N (heat 405065) exposed to salt containing LiTe, for 649 hrat 650° C (experiment 75-17). (a) Stressed section, 100X;

(b) unstressed section, 100X; (c) unstressed section, 500X

6.13 SALT-TELLURIUM CREEP STUDIES

T. K. Roche B. McNabb J. C. Feltner
The eight creep-rupture machines for testing in
MSBR fluoride salt (LiF-BeF»-ThF.-UFs, 72-16-
11.7-0.3 mole %) are all now in operation. The test
program has been set up to run duplicate tests on
eachalloy, one specimen being tested in clean saltand
the other in tellurium-containing salt to obtain an
indication of susceptibility to embrittlement while

under stress in the presence of tellurium. Tellurium s
supplied to the salt by additions of milligram
quantities of CriTes through valves in the top of each
salt chamber. Provision has been made for salt
sampling during the course of a test by drawing a
small quantity into a copper tube that can be inserted
into the chamber and pushed below the salt level.
Also, these machines have been equipped with
electrochemical probes for making voltammetric
measurements to monitor U*'/U" ratios of the salt
(Sect. 5.1, this report).

62 63 234 295 295 295 269 297 303 306 5

Y-136181

413 417 822 456 460 462 405
065

348

1]

Wb 474~ 476~ 474
5% 535 539 557

421-
543

I

166 533

474
558

474
901

475-
421

304
£

X6
2

NX7
187

NX7 N
353

Fig. 6.62. Specimens exposed for 1000 hr at 700°C to the vapor above tellurium at 300°C (experiment 75-18).

Table 6.26. Ranking of alloys from experiment
75-24 specimens exposed for 287 hr at 700°C
to the vapor above tellurium at 300°C

Product of number of cracks

and average depth Alloy Concentration (%)
number cr Ti Nb

1095 450 8.1 21
634 518 7.18 0.95 0.96
351 524 15.20 <0.02 0.69
194 516 7.30 <0.02 <0.10
150 523 10.40 <0.02 047
86 513 720 <0.02 045
80 514 7.05 <0.02 0.69
59 515 720 <0.02 020
25 517 7.10 <0.02 110

At present, all tests are being carried out at 650°C
and 30.0 X 10° psi. The alloys include both titanium-
and niobium-modified Hastelloy N. Two tests have
been completed on specimens of 2% titanium-
modified Hastelloy N (heat 74-533). A test in clean
salt and its counterpart in tellurium-containing salt
were discontinued after 4175 hr and 2380 hr
respectively. Approximately 150 mg of Cr;Tes was
added to the tellurium-containing salt test in a series
of three additions. About this same quantity hasalso

been added to other tests in progress. The salt from
these tests has been sampled for analysis, but results
are incomplete. Time-dependent strain measure-
ments recorded for tests completed or still in progress
in the salt environment are presented in Table 6.27.
The differences observed forany one alloy either with
or without tellurium being present are not considered
significant. The most significant part of these tests
will be the posttest metallographic examination to
determine the crack severity.

123

Y-138526

(a)
“
(b)

Y-138528

(e)
m
@)

100 200 MICRONS 600 700
11 1 00X—1 Y
0.005 0.010 INCHES 0.020 0.025

Fig. 6.63. Specimens exposed to salt containing CrsTeafor 500 hrat 700° C (experiment 75-19). (a) Alloy 503, (b) alloy 504, (c)alloy 505,
and (d) alloy 506.

124

" 1A .
.
? (Y 3 SUPS v v iy {
4 . T emegiAe
ai
100 200 MICRONS 600 700
i 1 1 L 100X—L TS S

T
0.005 0.0‘40 INCHES 0.020 0.025

Fig. 6.64. Hastelloy N (heat 405065) exposed to salt plus NisTe for 1000 hr at 700°C (experiment 75-20).

Table 6.27. Time-dependent strain measurements on niobium —titanium-
modified Hastelloy N alloys being creep-tested in fluoride
salt at 650°C, 30 X 10° psi

Composition® i 3 g
Heat Wt %) Time-dependent strain after indicated hours (%)

number Nb Ti 100 250 500 1000 1500 2000 2500 3000 3500 4000

474533 217 021 045 135 307 393 493 572 645 7.5 7.79°
474-533° 217 033 155 326 436 543°

421543 0.7 023 027 045 089 146 204 276 349 423
421543° 0.7 007 011 027 081 133 200 273 354 428

411 115 030 038 060 116 176 249 333 417 506

411¢ 115 046 057 066 115 163 222 288 362 440

413 113 09 -009 001 036 146 228 3.14 396 470 559

413¢ 113 09 005 008 057 147 216 278 346 421 496

62 19 026 042 089 187

62¢ 19 0.14 021 050 129

9Base composition of Ni~12% Mo—7% Cr—0.05% C; see Table 6.21 for detailed chemical analysis.
Ppest discontinued.
“Tellurium-containing salt test.

Y-138539
Y-138538

(a) ‘ : HO)

20 40 60 MICRONS 120 140
100 200 MICRONS ' 600 700 P8P M ox i e

1 T
o R Y 0.005
0.005 0.010 INCHES 0020 0.025 0.001 INCHES

| Y-138535
Y-138537

STl

Y-138636 |

(©

()]
{00 200 MICRONS 600 700
1 1 L 100 X—L 1 |y
T T T |
0.005 0.010 INCHES 0.020 0.025

Fig. 6.65. Hastelloy N (heat 405065)exposed to salt containing CrsTe, for 500 hrat 700° C (experiment 75-21). (a), (b) exposed to vapor,
unstressed portion; (¢) exposed to vapor, stressed portion; (@) (e) exposed to salt, unstressed portion; and (f) exposed to salt, stressed portion.
126

Y-136901

-

Aems

295 42_11 405065 295 424 405065
¢ ommem
Fig. 6.66. Specimens exposed to the vapor above NisTe; for 2500 hr. Experiment 75-22 was run at 700° C, and experiment 75-23 was run

at 750°C.

Y-138519

¥-138520

(b)

()

100 200 MICRONS 600 700
1 1 L 400 X—L i i
T T T T T
0.005 0.010 INCHES 0.020 0.025

i

Fig. 6.67. Specimens exposed to the vapor above NisTe at 700°C for 2500 hr (experiment 75-22) and strained to failure. (a) Alloy
405065, (b) alloy 295, and (c) alloy 424. As polished.

127

Y-138522

Y-138523

()

100 200
0.005 0.010

MICRONS
100X:

600 700

INCHES 0.020 0.025

Fig. 6.68. Specimens exposed to the vapor above NisTe: at 750°C for 2500 hr (experiment 75-23) and strained to failure. (a) Alloy

405065, (b) alloy 295, and (c) alloy 424. As polished

6.14 SALT PREPARATION AND FUEL PIN
FILLING FOR MSR PROGRAM CAPSULE
IRRADIATION EXPERIMENT TeGen-2

M. R. Bennett A. D. Kelmers

Fueled irradiation experiments were initiated in
1972 to investigate the behavior of fission products,
particularly tellurium, on structural materials being
considered for the Molten-Salt Breeder Reactor. The
fuel pins were made of 4-in. lengths of '/-in.-diam
tubing of the various alloys under investigation and
were filled with fuel salt containing ***U. Preparation

of these fuel pins required that a suitable salt charge
be prepared and transferred into the fuel pins and
that the fuel pin be separated, sealed, and inspected
for subsequent irradiation. Because of the use of U,
these pins were prepared in a glove box in the
Plutonium Laboratory of the Metals and Ceramics
Division.

The first fuel pins were prepared in 1972, but the
MSR Program was discontinued before the pins were
irradiated. The glove box was retained, but much of
the equipment was disposed of. When the MSR
Program was reinstated in 1974, the facility for

Y-138525 | Y-138183

Y-138189 I

(&)

Y 138186

Y-138168
(e)
“
: . . =~ : : Y-138180
(2) : ‘
— ’

@ (h)
100 200 MICRONS 800 700
| ' : L—100X—y —

T T
0.005 0.010 INCHES 0.020 0025

Fig. 6.69. Specimensexposed to the vaporabove tellurium metalat 300° Cfor 287 hrat 700° C (experiment 75-24) and strained to failure
at 25°C. (a) Alloy 450, (b) alloy 518, (c) alloy 524, (d) alloy 516, (e) alloy 523, (f) alloy 513, (g)alloy 514, (g) alloy 515, and (i) alloy 517. As
polished.

preparing the fuel pins had to be rebuilt. The original
design for the salt preparation vessel was used, but
the design of the salt receiving vessel was modified
considerably.

One batch of salt was prepared for filling several
pins. The individual sets of pins (six each) were
assembled in the salt receiving vessel by C. K.
Thomas and J. W. Woods of the Metals and
Ceramics Division. The operating work associated
with the salt preparation and filling the pins was done
by W. B. Stines and W. H. Miller of the Metals and
Ceramics Division. Various chemical analyses of the
salt were performed by members of the Analytical
Chemistry Division. The filled fuel pins were made
available to C. R. Hyman of the Reactor Division for
assembly into an irradiation capsule.

" The following material will cover the preparation
of the salt and the filling of the fuel pins for TeGen-2.
There are several other experiments in this series,and
some of the details concerning these experiments will
be discussed in Sect. 6.15.

6.14.1 Salt Preparation

The fluoride fuel mixture for the TeGen-2
experiment was prepared in a nickel reaction vessel

129

fabricated from a 10-in. length of 2'-in. IPS,
schedule 40 pipe in accordance with Dwg. M-10613-
RM-004E (Fig. 6.70). Following receipt of the vessel,
it was installed in the glove box furnace well (Fuel
Cycle Alpha Facility, Building 4508). A schematic of
the apparatus is shown in Fig. 6.71,and a photograph
of the gas delivery system is shown in Fig. 6.72.
Preparations were then made for loading the fuel
mixture. The total fuel mixture (714 g) consisted of
600 g carrier salt, "LiF-BeF-ZrF4(65.4-29.6-5.0 mole
). 73.0 g 2*UF,, and 41.1 g *’UQ,. The carrier salt
contained “hafnium-free” (<20 ppm) ZrF. and
isotopieally pure (99.9+%) "Li. The **U0O; contained
less than 7 ppm *U.

The following procedure was used to minimize
contamination of the glove-box and component
equipment duringaddition of the UF., and UQ:. Four
6-in. lengths of '/>-in.-OD metal tubing were inserted
into appropriate lengths of flexible tygon tubing and
taped together securely. These, in turn, were taped
securely to polyethylene bottles, two containing the
carrier salt (half, or 300 g, in each bottle) and the
remaining two bottles containing the UFs and UQz.
By then adding the contents of each bottle through
the large riser of the receiving vessel in the order (1)

(Y

= PARTS 1481 ’“'; 8
AT | D 0. [ RGO | DEXCRPTION STOCK STT_| WATERAL
OB RE TR INT) JORETY -_'.EF_“'
[1E8] [y j j‘ 1 ai vy
w18 ) B I wajw 37 ]
=24 th | sk p 1
xf1-s 1 [RATE- 3V, oA x Ty 1o A
=[2¢ 7 BATT WELD MALL GO, MAMILOM S §00--Cw | 31§ 137
u|er T 1w L_ AL _1
x|18 ) T %oz
afey N TR T ST B

NOYES:
T MATED WELDS SaAU RAVE “Puut REATY RFFORT
AADIGERASTY ~ PRAL PROJECT RMAN.
b sfunsa WA ORCK COMMATED ASLY RY YALUUM
TEATINA - wITH
SECY. 7 OF \WSPUTTION EMGIMEEIINA QUALITY
T ASMRARCE MROCEDURE Ay . LEAR
RATE BmAL wOT BACLED | X 107 31D Viac,
3. AW PATH WAL, B vidBLY EAKL OF POREIGN
MATERVALE § ODRAREALEID WiTH ETAYL ALLOWOY
REFORL  WALLDAWG -
- 4  pEbhG compriTIONS

FREALAL ¥ P
Yunr 100"

b cur )
} ~

e ..‘

L ALL MICREL MATERIAL Sk WAVE Y4 THAW
&.01 RGOH .

[T RTARRTOR o g oo i1

- ot A LAY
o .
Tuaes Gassiat Cazrendriny
T

TaGen -1 FURL PN = 1042

SALT PREPARATION
VESSEL WELDMENT
b RH 2.~ 1
[u[1061% Jam

I
[oc4]e}s]

Wiy

Fig. 6.70. Salt preparation vessel.
ORNL-DWG. 75-12930

FLOW DIAGRAM FOR TeGen-2

Argon —»
Argon —»
HF ~Hy ~Argon —» — 1
” >4
Vent
¢S5 v-q
@V“' &,Av-m N
V-1A AV-2 , Vent Hz
, L V-3 (0.G.)
AV-3V
Electrod | Av-4
ectrodes Qb § Electrodes 1
FM No.2 C3 Conductivity C-4
Probe
-/ -
QPA\H % AV-2 777 X
C-1 = [\
V-2 Salt Transfer AV-3A
1/2" 0D Tube
Riser C-2
(e | ]
- Salt Fill
=Salt § Vessel (SFV)
Preparation +Sa it
Sparge Vessel 5 Overflow
Tube ML (SPV) fig Receiver
=] =E5 (SOR)
Qo
y | il Ry
Rlaniie Nickel I 3% Fuel Pi
— . i ° - Fuel Pins
Wool LS —
Receiving Reservoir

Fig. 6.71. Diagram of glove box, furnaces, vessels, and gas lines.

carrier salt, (2) UO, (3) UF,4, and (4) carrier salt, the
UF,; and UO; were effectively “scrubbed” from the
inside walls of the riser and covered with a layer of
carrier salt. '

After the riser-tube was recapped, the salt mixture
was heated to 600° C (melting point of 450° C) under
~200 sccm argon flow and sparged for 1 hr in
preparation for electrochemical measurements of the
oxide content of the salt by personnel of the
Analytical Chemistry Division. The oxide measure-
ment was done by retracting the sparge tube and
lowering a gold-iridium electrode to a depth of '/, in.
into the molten salt. The volt-ammograms obtained
from these measurements were for calibration of an
experimental oxygen procedure and were not used in
the TeGen-2 preparation.

6.14.2 Hydrofluorination

After the voltammetric measurements were made,
the sparge tube was lowered into the salt mixture, and

a 36-hr hydrofluorination was carried out at 600°C
using a HF-H; (20-80 mole %) mixture at a flow rate
of 500 sccm. The procedure stipulated a hydrofluori-
nation time of 48 hr; however, time scheduling
necessitated reducing this to 36 hr. At the end of this
time, the HF and H, flows were cut off, and the salt
was sparged with argon atabout 500 sccm for 12 hrto
remove dissolved HF. Two samples of the salt melt
(one filtered and one unfiltered) were then taken and
submitted to the Analytical Chemistry Division for
an isotopic uranium assay, measurement of the
U /**U ratio, and determination of the concentra-
tions of total U, Fe, N1, Cr, and oxide.

The analytical results for the uranium inventory
(Tables 6.28a,b,and c)were in excellent agreement
with the theoretical values, showing less than 109
deviation. The discrepancies observed in the iron and
total oxide analyses are difficult to evaluate. It
appears unquestionable, however, that oxidation
occurred at some point during analysis of the No. 1
PHOTO 127276

Fig. 6.72. Gas delivery system.

samples since the No. 2 samples with considerably
lower oxide values were duplicates taken five days
later.

In both cases the oxide content was too high to
continue with the next step in the operational
procedure, and further hydrofluorination was
required. The salt mixture was therefore reheated to
600° C under argon flow and hydrofluorinated with
HF-H; for an additional 24 hr using the conditions
previously employed. Again, samples of the salt melt

were taken before and after the hydrofluorination.
As before, extreme care was taken during and after
the sampling to preclude the possibility of contact
with air; the results are shown in Table 6.29.

A comparison of the analytical data given in Table
6.29 with those obtained earlier (Table 6.28) gives
further support to the speculation that the discrepan-
cies in the data probably resulted from sample
contamination, probably with moisture, during
analytical procedures. The oxide content of both

Table 6.284. Analytical results for uranium
isotopes in the TeGen salt charge” (at. %)

132

Table 6.29. Oxide concentration (ppm) after
second hydrofluorination

Mass Filtered Unfiltered
233 39.82 39.85
234 0.511 0.512
235 0.157 0.156
236 0.007 0.007
238 59.51 59.48

2 After first hydrofluorination.

Table 6.28b. Analytical results for total uranium
in the TeGen salt charge? (at. %)

Salt Uranium total (mg/g) 233238y
sample Theoretical Observed Theoretical Observed
Filtered 127.5 123.4 0.653 0.676
Unfiltered 127.5 118.6 0.653 0.677

2 After first hydrofluorination.

Table 6.28¢. Analytical results for total metallic
impurities in the TeGen salt charge” (at. %)

After
Impurity Prehydrofluorination hydrofluorination
Filtered Unfiltered

Fe 129 255 213
Ni 318 <25 <25
Cr 56 34 33
Oxide 1680 £ 512 2100° 2000
678% 607

4gample No. 1, taken Aug. 13, 1975.
bSampie No. 2, taken Aug. 18, 1975.

samples taken prior to additional hydrofluorination
was lower by a factor of about 2 than that obtained
from identical samples taken previously from the
same salt. Also, the observation that the oxide
content remained relatively constant after an
additional 24 hr of hydrofluorination is difficult to
accept as valid, and it seems highly probable that the
actual oxide content of the final salt mixture was less
than 300 ppm.

6.143 U*/U’ Ratio Adjustment

The U*/ U™ ratio was adjusted by heating the salt
mixture to 700°C under argon flow and then
introducing hydrogen at 500 sccm for 6 hr. Attheend
of this time, the hydrogen flow was cut off, and the
salt was sparged for 1 hr with argon at about 200

Concentration
Salt sample
After 36 hr After 60 hr
Filtered 420 357
Unfittered ’ 332 603

sccm. Preparations were then made for evaluation of
the U*"/ U’ ratio by retracting the sparge tube above
the salt level and lowering the gold-iridium electrodes
into the melt to adepth of '/sin. Analytical Chemistry
Division personnel then made volt-ammogram
recordings. A final U*'/ U™ value of 58 (1.72% of the
total U as U’") was obtained, which was within the
desired range of 1.0 to 1.8% of the total U as ur.
After retraction of the electrodes from the melt, the
salt mixture was cooled to room temperature under
argon and maintained in storage status for subse-
quent transfer to fuel pins.

6.14.4 Preparation of Salt Fill Vessel

A drawing of the assembled salt fill vessel is shown
in Fig. 6.73. The vessel is shown in two different
stages of assembly in Figs. 6.74 and 6.75. The salt 1s
first transferred to the short vessel in the center and
then pushed by gas pressure into the individual fuel
pins, with the overflow going into the long vessel. The
top Y in. of the salt from the fuel pins is then blown
back to the initial vessel where the salt can be
analyzed. The outer flanged vessel remains in place,
but a new inner assembly is made for each filling.

‘In preparation for receiving the salt transfer from
the salt preparation vessel, the salt fill vessel was
pretreated with hydrogen. The salt fill vessel was
heated to a temperature of 665°C under an argon
flow of about 200 sccm. The operational procedure
stipulated a pretreatment temperature of 700°C;
however, because of failure of one of the Calrod
heating units, the maximum temperature obtainable
was 665° C. The argon flow was cut off,and hydrogen
was allowed to flow through the vessel and pins at
about 500 sccm for 4 hr. At this time the hydrogen
was cut off, and the vessel was cooled to room
temperature under argon. All valves and connections
were then closed off.

6.14.5 Salt Transfer

With both the salt preparation vessel and the salt
fill vessel at room temperature, a preshaped Y4-in.-
OD nickel salt-transfer line was connected between
133

QEML PG - 766659

WELD LTERY:
MUMBERS | wiL DI

WEPECT WOAL _PROCLLDUEES |

WELD/MAF AsdD
TAMY

MRS, DA T wmm
==

wyesos

%
s

IU.a 0 BLf Iy ‘
Ja[#mo lrriaar el ]

the sparge tube exit of the salt preparation vesseland
the salt transfer inlet line to the salt fill vessel (Fig.
6.76). A 750-W Calrod heating unit was fastened
securely to the salt-transfer line, and the entire
assembly was then heavily insulated. With argon
flowing through, both vessels and the salt-transfer
line were heated to 650° C in preparation for the salt
transfer. The objective in this part of the procedure
was to transfer about 85 cm’® (203 g) of the salt
mixture from the salt preparation vessel into the
receiving reservoir of the salt fill vessel. A conductiv-
ity probe was positioned at a predetermined level in
the receiving reservoir to monitor the salt level.
Contact of the rising salt with the probe was indicated

by a red light on the instrument panel. At this point, -

argon was allowed to flow through the bypass line
leading to the salt preparation vessel at a rate of
about 100 sccm (under 5 psig pressure). After the vent
valves of the salt fill vessel were opened, the bypass
valve to the salt preparation vessel was closed. Salt
was then observed to transfer as evidenced by
“smoking” of the insulation. Within about | min, the
red light went on, indicating completion of the
transfer. The argon flow was cut off, and the
appropriate valves were closed. A determination of

. Salt fill vessel.

the salt level in the receiving reservoir was then made
by raising the probe and measuring the point at which
the red light went off. This was shown to be s in.
above the specified level, or the equivalent of a 7.6
cm’ excess of salt. The total transfer was thus about
92.5 cm’, which was less than 10%deviation from the
stipulated transfer volume of 85 cm’.
In preparation for salt transfer from the receiving
_reservoir into the fuel pins, the temperature of the salt
fill vessel was reduced to 600°C, and argon was
flowed at a rate of about 25 sccm through the bypass
line leading to the receiving reservoir. The bypass line
valve was then closed, at which time the pressure
slowly increased to 5 psig. After a few minutes, the
pressure decreased slowly to zero, indicating comple-
tion of the transfer. This was confirmed by lowering
the conductivity probe into the receiving reservoir to
a position about s in. from the bottom of the vessel
and observing that the red light failed to go on. All
valves on the salt fill vessel were then closed, and the
salt was allowed to equilibrate in the fuel pins for 24
hr at a temperature of 600°C.
After 24 hr of equilibration, preparations were
made for transfer of excess salt in the upper regions of
the fuel pins into the electrode cup of the receiving
134

PHOTO 202475

TERS

20 2 )
RN AR ARG,

Fig. 6.74, Salt fil vessel partially assemled. Outside container is on the right, and the partial inner assembly is on the left. The short

vessel initially receives the salt. The longer vessel is the salt overflow receiver, which catches any salt in excess of that required to fill the fuel

pins.

reservoir for final evaluation of the U*'/ U™ ratio. At
a temperature of 600° C,argon was flowed at a rate of
about 25 sccm through the bypass line leading to the
fuel pin overflow reservoir. The bypass line valve was
then closed, pressurizing the overflow reservoir to
about 5 psig. In less than 1 min, the pressure dropped
to about 0.1 psig, indicating that the salt had
transferred and that argon was bubbling up through
the salt in the receiving reservoir. The argon flow was
then cut off, and all valves to the salt fill vessel were
closed. After lowering the iridium electrode ' in. into
the salt in the receiving reservoir cup, final determi-
nation of the U/ U* ratio was made by Analytical

Chemistry Division personnel. Interpretation of the
volt-ammograms obtained gave a U*'/U* ratio of
200 (0.5% of the total Uas U*). It seems possible that
hydrogen pretreatment of the salt fill vessel and fuel
pins at the lower temperature (665° C) than the 700° C
stipulated in the operating procedure may have
contributed somewhat to the reoxidation of some of
the U™

It was decided to make a second measurement of
the U*'/ U ratio since there was some uncertainty as
to the proper functioning of the instrumentation
during the volt-ammogram recordings. This was
done under the same conditions as the previous

135

PHOTO 202575

AN N HANNNN

Fig. 6.75. Salt fill vessel showing outside portion on right and assembled internal assemly on left. All joints are welded in the lines

connecting the small vessels and the fuel pins.

determination. A U*'/ U™ ratio of 205 (0.49% of the
total U as U") was obtained, which is in excellent
agreement with the initial value of 0.50%. Since this
value was considered acceptable, the salt fill vessel
was cooled to room temperature under argon for
subsequent disassembly. All valves and connections
were closed, and the glove box was then decontami-
nated.

6.15 SALT PREPARATION AND
FILLING OF TeGen-3 AND 4

H. E. McCoy W. H. Miller
B. McNabb W. B. Stines

The same salt prepared for the filling of TeGen-2
was used to fill the fuel pins for TeGen-3 and 4. The
same basic procedure was also used for transferring
the salt from the preparation vessel into the fuel pins,

but some changes were made which shortened the
time required.

During the interval between the filling of the pins
for TeGen-2 and for TeGen-3, the salt in the
preparation vessel was under a positive pressure of
argon. Before the pins for TeGen-3 were filled, the
transfer line and the salt fill vessel (including the fuel
pins) were purged with hydrogen for 4 hr at 700°C.
The salt preparation vessel was then heated, and
approximately 100 cm”® of salt was transferred into
the receiving reservoir. At this time the U/ U*" ratio
was determined to be 0.001. The salt was transferred
into the pins and allowed to remain at 600° C for 8 hr.
The sample blown back for analysis did not indicate a
detectable change in the U™/ U* ratio. Although the
salt was more oxidizing thandesired, it was judged to
be suitable for these experiments. The salt fill vessel
was disassembled, the pins separated, closure welds

136

PHOTO 1273-76

Fig. 6.76. Salt-transfer line from salt preparation vessel to salt filling vessel.

made, pins radiographed and leak checked, and the
pins given to the Reactor Division for assembly into
the irradiation capsule.

It was decided that prior to filling the pins for
TeGen4 the salt remaining in the preparation vessel
would be retreated with hydrogen to increase the
U™/ U* ratio. Since the new salt fill vessel and the
transfer line were to be pretreated with hydrogen
also, the entire system was heated and pretreated with
hydrogen for 4 hr at 750°C. The U**/ U** ratio in the
salt preparation vessel could not be determined,
probably because a conductive film had formed on
the surface of the salt. Approximately 100 cm’ of salt
was transferred to the salt fill vessel where the
U*/U" ratio was determined to be 0.010. The salt
was then transferred into the individual pins, held at
600°C for 8 hr, and partially transferred back to the
receiving reservoir. The U''/U*" ratio was deter-
mined to be in the range 0f0.013 to 0.010. The salt fill
vessel was disassembled, the pins separated, closure
welds made, pins radiographed and leak checked,
and pins given to the Reactor Division for assembly
into the irradiation capsule.

6.16 OPERATION OF TeGen-2 AND -3
C. R. Hyman

TeGen-2 and -3 are the second and third ORR
poolside experiments designed to irradiate prospec-
tive MSBR vessel materials. The experiments were
designed to produce a fission product inventory of at
least 5 X 10" tellurium atoms/cm” at the metal-to-
salt interface, together with a representative mix of
other fission products. The test device is an
irradiation capsule containing three 'h-in.-OD X
0.035-in.-wall, 4-in.-long tubular fuel pins, partially
filled with fuel salt. The three fuel pins in TeGen-2
were 2% titanium-modified Hastelloy, Inconel-600,
and titanium-lanthanum-modified Hastelloy for the
top, middle, and bottom fuel pins respectively. In
TeGen-3 the three fuel pins were 2% titanium—'2%
niobium-modifed Hastelloy, 2% titanium-1%
niobium-modified Hastelloy, and 1% titanium-1%
niobium-modified Hastelloy for the top, middle,and
bottom pins respectively. The fuel pins were filled
with approximately 7.14 cm’ of fuel salt, leavinga >~
in. void at the top of the fuel pin, which was backfilled

with helium after salt filling. The fuel salt was a
mixture of LiF-BeF,-ZrF4-" UF-**UF, (63.5-29.0-
5.0-1.0-1.5 mole %).

The fuel pins were arranged vertically inside a
double-walled type 304 stainless steel vessel. The void
between the fuel pins and the inner wall of the vessel
was filled with NaK for improving heat transfer. A
schematic of the experiment is shown in Fig. 6.77.
Small design changes were made in TeGen-2 and -3,
which should be noted. Small hubs were attached to
the end plugs inside each fuel pin, and four foils 0.005
X 0.125 X 0.375 in. were attached to each hub. The
bottom hub was solid and was0.265 in. in diameter X

137

0.125 in. thick. The top hub was 0.267 in. in OD X
0.135 in. in ID X 0.125 in. thick. The foils and hubs -
for each fuel pin were made of the same material as
their respective fuel pin. The foils were included to
provide samples that could possibly have sufficiently
low activity for Auger analysis.

The capsules were instrumented with four
Chromel-Alumel thermocouples per fuel pin. One
electrical resistance heater was wrapped around each
fuel pin, with the heaters positioned to minimize axial
temperature gradients. The heaters also maintained
the fuel pins abouve 150° C during periods when the
reactor was shut down and while the capsule was

ORNL-DWG 75-7864

3 HEATER LEADS

4 GAS LINES
15 THERMOCOUPLES

TN

ARGON
HELIUM

NaK

-~

TE-101

P A A VA A A A VA

TE-102

8ULKHEAD

2z
b A A S A S A S— 4

304 SS PRIMARY CONTAINMENT

le—— 304 SS SECONDARY CONTAINMENT

TeGen-2 — 2% Ti-Modified
Hastelloy N

Z

FUELPIN1 —
MSR FUEL SALT

TE-103

REACTOR

TE-104

NOTE —~TE-101, TE-105, AND
TE-109 FACE SOUTH

L L VL W W . A WA W L W, .

TE-105

TE-106

W0 W W W W, W W L WL WL W ¥

rd
—r———r—T T

//:>HEATERCOWI
TeGen-3 — 2% Ti-0.57% Nb-Modified

Hastelloy N

TeGen-2 — Inconel 600
FUELPIN2 -

v -

AN MSR FUEL SALT

TE-107

TE-108

Tz
W W . ¥ N Y

T

FTrrrzz

HEATER COIL 2

)TeGen-B — 2% Ti-1% Nb-Modified

Hastelloy N

TeGen~2 — Ti-La~Modified
Hastelloy N

fi?::—FUELNN3—
MSR FUEL SALT

WL ", "W L W W W Y N W W W . W W W W

HEATER COIL 3

TeGen-3 — 1% Ti-1% Nb-Modified
Hastelloy N

TE-115

. Drawifig of TeGen-2 and -3 capsules.
removed from the reactor after completion of the
irradiation but prior to hot cell disassembly. Fission
product decay radiation in frozen salt (<150°C)
causes dissociation of the fuel salt components at a
faster rate than recombination at that temperature.
At temperatures higher than this, recombination
exceeds evolution, and no net fluorine is generated.

The design operating temperature for the speci-
mens was 700°C at the salt-to-metal interface, and
this temperature was maintained as uniform as
possible over the length of the capsule throughout the

irradiation. Bulk temperature control was main-
tained by adjusting the experiment position with
respect to the reactor by means of the movable track
on which the experiment was placed in the ORR
poolside facility. This movement changed the
neutron flux at the experiment and thus varied the
fission heat produced. Fine temperature control and
uniform axial temperatures were attained by heater
power adjustment. A more complete description of
the experiment design may be found in an earlier
report.”’ Figures 6.78 and 6.79 are pictures of TeGen-

PHOTO 4735-75

T e——— e e ST =

Fig. 6.78. TeGen-2 subassembly prior to insertion into the primary containment. The bottom of the assembly is on the right, and the fuel

pins from left to right are numbered 1, 2, and 3 respectively.

PHOTO 4047-76

Fig. 6.79. TeGen-3 subassembly prior to insertion into primary containment. The bottom of the assembly s on the right, and the fuel

pins from left to right are numbered 1, 2, and 3 respectively.

2 and -3 subassemblies before insertion into the
primary containment.

6.16.1 Operating History of
TeGen-2 and -3

TeGen-2 was installed into the ORR poolside
irradiation facility position P<4A on November 3,
1975, and full-power irradiation began on November
4, 1975. The experimental assembly was irradiated
for a total of 1492 hr but was retracted away from the
reactor several times during this period. Irradiation
was ended on January 21, 1976, at which time the
experimental assembly was retracted to 16 in. away
from the face of the reactor. Since TeGen-3 was ready
for insertion at the same time TeGen-2 was being
removed, the gas lines of the TeGen-2 capsule were
cut one day before transferring it to the hot cell.
However, during this time the heater and thermocou-
ple instrumentation remained intact, and the heater
power kept the capsule at >150°C. On January 28,
1976, TeGen-2 was transferred to the ORR hot cell
and disassembled. Flux monitors and fuel pins were
retained.

One unexpected problem developed after irradia-
tion of the capsule had been completed. While
removing the capsule and transferring it to the hot
cell, it was necessary to depressurize the gas systems
in the capsule. The secondary (argon) gas system was
depressurized with no problems, but the primary
(helium) system could not be depressurized. The lines
leading to and away from the capsule were depressur-
ized so that if only one plug existed, the capsule
would depressurize. Forward purging and back
purging were attempted, but the plug was not cleared.
Finally the gas lines were severed and welded. The
next day the capsule was transferred to the hot cell,
where dissassembly began. No unusual events or
physical characteristics were observed when the
experimental assembly was dismantled. The pins
appeared to be intact.

The most likely cause of the plug in the primary gas
line was impurities in the NaK system. The carrier
used to transfer the NaK from the source tank to the
capsule filling apparatus was not filled with the
required amount of NaK and thus had to be refilled
two more times. In addition, the procedure for filling
the experimental assembly with-NaK from the carrier
had to be done twice before the NaK fill was
completed. It is possible that during the filling of the
carrier from the source tank, air may have been
introduced into the NaK. This would have formed
metallic oxides, which could have plugged the lines

139

during or after transfer of the NaK to the experimen-
tal assembly. To prevent this problem in the future, a
larger NaK carrier will be used and a better procedure
followed in filling the carrier from the source tank.

TeGen-3 was inserted on January 28, 1976, and

began full-power operation at design temperature on

January 30, 1976. Since that time, TeGen-3 has been

‘operating at design temperature. There have been

periods when the temperatures indicated by the
safety thermocouples Te-113, -114,-115(Fig. 6.77) at
the bottom of the primary containment have
oscillated rather severely. The amplitudes of the
oscillations have been as high as 100°C, which is
much higher than was recorded for the previous two
capsules. Smaller oscillations occurred at the middle
and upper regions of the capsule as well.

At one time there was a 200° C drop in the readings
of the safety thermocouples of TeGen-3. The reactor
and experiment were operating under unusual
conditions when the drop occurred. The previous two
TeGen capsules experienced temperature oscillations
that were probably due to NaK currents inside the
primary containment. These currents were undesir-
able because they cooled the bottom of the capsule by
convection. The bottom fuel pin had an axial
temperature difference of approximately 70°C, and
the safety thermocouples were 50° C cooler than the
lowest thermocouple on the lower fuel pin. To
alleviate this problem, TeGen-3 was constructed with
a baffle positioned on the bottom of the lowest fuel
pin. However, the temperature distribution was
worse, with an axial gradient of 200° C on the bottom
fuel pin and the safety thermocouples about 100°C
cooler. These data, in addition to the oscillating
pattern observed, suggest that NaK currents may still
exist in the capsule. In any event, the temperature
abberations noticed have not posed any threat to the
experiment or to the reactor.

6.16.2 Data Analysis for
TeGen-2 and -3 .

Tables 6.30 and 6.31 give typical operating data for
TeGen-2 and -3. The average temperature of TeGen-
2 over the length of the capsule was 693.7°C, which
corresponds to a fuel-salt to fuel-pin interface
temperature of about 700° C. In TeGen-3 the average
temperature 1s 677.3°C, and the corresponding
interface temperature is approximately 684°C. The
maximum temperature variation occurs in the
bottom fuel pin in each capsule. This variance occurs
because of NaK currents existing in the primary
containment. Even though TeGen-3 had a baffle to
140

Table 6.30. Typical operating conditions for irradiation capsule TeGen-2°

Metal-gas or

Fuel pin 'I'hermocouple Metal'salt i.nterface Heater power
Location Material number temperatures ("C) W
Observed Average
Top 2% Ti-modified 101 717¢ 228.8
Hastelloy N 102 706
103 700 701
104 697
Middle Inconel 600 105 704€ 57.1
106 718
107 710 698
108 667 ).
Bottom  Ti-La-modified 109 722°¢ 183.1
Hastelloy N 110 711 :
111 683 682
112 653

2Data taken Dec. 4, 1975, at 8:07 and 7.45 in. away from the reactor.

bgee Fig. 6.77 for locations.

“Metal-gas interface temperatures. All others are metal-salt interface temperatures.

Table 6.31. Typical operating conditions for irradiation capsule TeGen-3?

Metalgas or
Fuel pin Thermocouple l\t/[etal-saltt mtegag)e Heater power
Location Material number empetatires W)
Observed  Average
Top 2% Ti and 0.5% Nb 101 594€ 230.3
Hastelloy 102 692
103 714 709
104 720
Middle 2% and 1% Nb 105 682° 8.6
Hastelloy 106 726
107 Open 726
108 Open
Bottom 1% Ti and 1% Nb 109 700¢ 173.8
Hastelloy 110 695
111 665 630
112 530
- 9Data taken Mar. 4, 1976, at 12:55 and 5.5 in. away from the reactor face.

bSee Fig. 6.77 for locations.

“Metalgas interface temperatures. All others are metal-salt interface temperatures.

i - obstruct the currents, a favorable geometry still

~ existed for the occurrence of very local violent NaK
currents, which were evidenced by the oscillating
temperatures recorded during operation of the
capsule. Figure 6.80 gives the relative locations of the
thermocouples, both vertically and circumferen-
tially, as well as their orientation relative to the face
of the reactor.

Several sets of temperature data are plotted in
Figs. 6.81 and 6.82 for TeGen-2 and -3 respectively.
The top set of curves respesents the normal operating
conditions with the heater power as noted in Tables
6.30 and 6.31, the middle curves denote the
temperatures at the same capsule position but
without the heater power, and the bottom curves
denote temperatures at the retracted position with
ORIENTATION TE No.

101

102
103

104

REACTOR

TE-103,107, 11

105

106
107

TE—104,
108, 112

TE-101,
105, 109

108
TE-102, 106, 110

109

110
mm

112

141

ORNL-DWG 75-7865A

a

AXIAL LOCATION |

DISTANCE FROM HMP

FUEL PINS
{(in.)

T

Iyl

i

5.00

4.00
3.25

304 SS

1.75

0.50

~0.50
-1.25

- REACTOR HORIZONTAL
MIDPLANE {HMP) '

HASTELLOY-N

—2.75

—4.00

-5.00
-5.75

INCONEL-601

-7.25

Fig. 6.80. Thermocouple location and orientation information. Materials shown on the right were those in TeGen-1. The same physical
layout was used in TeGen-2 and -3 although the materials were different.

the heater power as noted in Tables 6.30 and 6.31.
This last condition represents negligible gamma and
fission heat. '

Figure 6.83 presents the design temperature
distribution in the fuel region of the pins where the
fission heat production is considered the major
source of heat flux. Figure 6.84 presents the
temperature distribution in the heater region of the
pin and assumes no fission heat and an average heat
rate of 26.5 W/g in the region of the heaters,
thermocouples, and NaK. The maximum amount of
power that a heater must produce was estimated from
the 26.5 W/g mentioned above to be approximately
300 W/heater. The actual heater power used during
operation was much less because of the conduction of
fission heat from the ends of the fuel pins.

Figure 6.85 gives an estimate of the relative radial
temperature profile for the top, middle, and bottom
fuel pins respectively. Although exact temperature
distributions are not known, the differences in the
respective fission reaction rates in the fuel pins cause
the centerline temperatures of the middle and bottom

fuel pins to be higher than for the top fuel pin. Asa
result, the temperature gradients in the fuel pins
differ substantially, radially as well as axially, and
therefore may influence the corresponding fission
product deposition rates on the surfaces of the fuel
pins. This occurs in spite of having the same
temperature along the outside of the fuel pins. Note
that the temperatures at the surfaces of the fuel pins
are the respective averages of the lower three
thermocouple readings on each fuel pin in Table 6.30.

These thermocouple readings represent actual fuel
region temperatures and not those of the void region
above the salt. Also note that the heater power
provides heat to the ends of the fuel pins where little
or no fission heat is produced. This has the effect of
leveling out the temperatures over the length of the
experiment.

6.16.3 Preliminary Results
of TeGen-2

Using a crude heat balancing technique, the
steady-state fission heat generation powers were
142

ORNL—DWG 76—668B8E

900 ! 1 | | T T I i
NORMAL CPERATING CONDITIONS:
800 - POSITION =7.45in. FROM REACTOR _
ELECTRICAL HEAT FACE = 469.0 watts
700 — /\ \ )
600 — /\ —
g N oo o
2
W 500 (— FISSION AND GAMMA HEAT ONLY: /\ ]
x POSITION=7.45in. FROM REACTOR FACE
L=
@
W
a
- HEATER POWER ONLY:POSITION = 7.45in.
300 |— FROM REACTOR FACE _ N
ELECTRICAL HEAT = 469.0 watts
THERMOCOUPLE POSITIONS AND NUMBERS
200 — TE TE TE TE TE TE TE TE TE TE TE TE
101 102 103 104 105 106 107 108 109 110 144 12
100 — 2% Ti- MODIFIED INCONEL-600 Ti-La MODIFIED
HASTELLOY HASTELLOY
o 1 | | | | l | 1
16 12 8 4 o -4 -8 -12 -16 -20

AXIAL DISTANCE FROM HORIZONTAL MIDPLANE {cm)

Fig. 6.81. Temperature distribution in TeGen-2.

calculated as 163, 258, and 173 W for the top, middle,
and bottom fuel pins respectively. These were
calculated knowing heater powers, gamma heats,
thermal conductivities, temperatures, and several
other parameters and iterating the GENGTC fission
heat values to calculate the observed temperatures.
Values produced by this method of evaluating fission
rates will be compared with those produced from
neutronics support of TeGen-2.

The tellurium production was calculated from the
above estimates of fission heat rates. Assuming a
yield of 0.04 stable tellurium atoms per fission,
tellurium inventories of 1.21 X 10'¢, 1.92 X 10'®, and
1.29 X 10"* tellurium atoms were produced in the top,
middle, and bottom fuel pins respectively. Assuming
a surface area of 29.7 ¢cm’ in contact with the salt and
an even deposition rate on all surfaces, the telluriuin
deposition should be approximately 4.07 X 10'¢,6.46
X 10", and 4.34 X 10" tellurium atoms/cm” for the
top, middle, and bottom fuel pins respectively. These

values compare favorably with the design level of 5X
10'® atoms/cm’. More-detailed tellurium production
estimates will be made when the flux monitors from
the experiment are analyzed.

6.16.4 Future Irradiations

Future irradiation of capsules will be very similar
to that of capsules run previously in this series.
However, the baffle structure placed in the bottom of
the primary containment will be eleminated because
of the temperature profile and fluctuations seen in
TeGen-3. In future experiments, fine quartz wool will
be placed between the primary and secondary
containment to reduce heat radiation from the
bottom of the capsule. This insulation should reduce
the heat loss and thus raise the temperatures in the
lower end of the capsule. The heater coils will be
arranged as they were in TeGen-2,
143

ORNL-DWG 76-6687

900
R
800 |— NORMAL OPERATING CONDITIONS:
POSITION =5.5 in. FROM REACTOR FACE ]
ELECTRICAL HEAT = 412.7 waits
700 |— ———— ]
600 |— / —
o FISSION AND GAMMA HEAT ONLY
& POSITION =5.5 in. FROM REACTOR FACE
Y 500 |— ' —
2
g HEATER POWER ONLY: ,
i POSITION =48 in. FROM REACTOR FACE
< 400 f— ELECTRICAL HEAT:=412.7 watls —
= k/
300 — —
\‘--”,
THERMOCOUPLE POSITIONS AND NUMBERS
200 |— ]
TE TE TE TE TE TE TE TE TE TE TE TE
101 102 103 104 105 106 107 108 109 110 114 142
00 L | = i i
2% Ti-2% Nb 2% Ti- 1% Nb 1% Ti—- 1% Kb
HASTELLOY HASTELLOY HASTELLOY
A T T T O O O
16 12 8 4 ) -4 -8 -12 -t6 - -20

AXIAL DISTANCE FROM THE HORIZONTAL MIDPLANE (cm)

Fig. 6.82. Temperature distributions in TeGen-3.

6.17 EXAMINATION OF TeGen-2
B. McNabb H. E. McCoy

The TeGen experiments were designed to evaluate
the resistance to cracking of several materials of
interest to MSBR when they are exposed to molten
salt containing fission products. The Y:-in.-OD fuel
pin is made of the material being evaluated. TeGen-1
containment materials consisted of type-304 stainless
steel, standard Hastelloy N, and Inconel 601. As
reported previously,’”” of the three materials tested,
Inconel 601 was the most resistant to cracking.

TeGen-2 is almost identical in design to TeGen-l,
but the fuel pin materials were 2% titanium-modified
Hastelloy N heat 74533 (top fuel pin), Inconel 600
(middle fuel pin), and 2% titanium-modified Hastel-
loy N plus 0.013 lanthanum (bottom fuel pin). The
chemical compositions are given in Table 6.32. The
irradiation of TeGen-2 (approximately 1578 hr at
700°C) was concluded on January 26, 1976, and
disassembly in the ORR hot cells was done on

January 28. The operating conditions of TeGen-2 are
reported in detail in Sect. 6.16.

Visual inspection of the disassembled equipment
showed the parts to be unchanged in their appearance
after the irradiation. The top and bottom fuel pins of
modified Hastelloy N (pin 2, heat 74533, and pin 6,
heat 74534, respectively) had a slight gold tint, and
the middle pin (pin 4, Inconel 600) had a silver tint

_ after annealing but before assembly in the experi-

mental equipment. The top fuel pin had a small
scratch along its length, probably from a sheathed
thermocouple pulled across it during disassembly.
The small scratch did not affect the tensile properties
of rings cut from the pin and tensile tested, and it was
undetected in subsequent metallographic examina-
tion. After disassembly, the fuel pins were heated ina
small furnace 150° C or higher to prevent evolution of
fluorine from the fuel salt. The pins were maintained
at this temperature except during transfer to the High
Radiation Level Examination Laboratory
(HRLEL), during gamma scanning and measuring,
144

ORNL-DWG 76-6689

9
N
5 $
4 A 7
X S& S¢
g % §6? N
A <& &, T 2
| ~ AL o AY ¢y O¢cy
4 NG & > W %
© - DA O3 (2 A
CL ~ & > O RS ' d e
“ & g Ok & AN © &
CAPSULE O g S §~2~§- &4 > £4
! & o SR X S S
' & & YEX &S & SS
& £ 4 IIE o < 5y
974 °C
.
s 743°C
713°CT Y
- i 678°C
708°C _1!—___._—EEG°C
[ ]
(-]
88°C 83°C
I‘-—q__.—.
— W , w +
--0.546—-|
—~—0.547 —=
———0.635
0997
— 1,087
1149
1,266

RADIAL DISTANCE (cm)

Fig. 6.83. Design temperature distribution of fuel pin in salt region of pin.
145

ORNL- DWG 76-6690

88°C

696°C

699°C

703°C

704°C

O.546-|

—— (.547—"

= 3

—(0.635 ——==

0.997

1.087

1.149

-1.266

¢
CAPSULE

RADIAL DISTANCE {cm)
Fig. 6.84., Temperature distribution in heater region of fuel pins.

146

ORNL~DWG T6-6691

A,
o
-
&
é &
V;t ékt?,’ éuf?/
R, SN
& $& N
A & (%] ~ =
< A Yo o P4
l X \ Q"{f}- - Q‘,o Uy
- S o ) A7
Lo & S NS N
capsule ¥ & & Q¢ & R N WY
| & S N NI o 3.
: & & qv .ztf’é—i\' &g &£ &
P T G INEE T ¥ '3
851 °CT \
! 702 | 70
702 @ =@ . 692
FISSION HEAT |
=0.65 kW /ft |
87 82
S —— o
A A A A A A
.r L ¥ ¥ ¥ ¥ ¥ ¥
933 °C 'O
| 702
702 ¢—¢ 698
FISSION HEAT
=1.03 kW/ft
$ 38 83
\ \.
A A A A A A A
¥ ¥ ¥ | J ¥ ¥ ¥
|
843 °C e
168
| coeg~eI 4582 g
0—-—._" 673
FISSION HEAT )
=0.69 kW /ft
A A A A A A A
¥ ¥ ] ¥ ¥ ¥ ¥
-0‘546’-|
- 3,547 —m

- 1.266

RADIAL DISTANCE (cm)

Fig. 6.85. Relative radial temperature profiles.
Table 6.32. Compositions of alloys used in TeGen-2 (in wt %)

Alloy Heat Source Ni | Mo Cr Fe Mn Si C Ti Al W v Co Cu S B N La H 0
number
Modified Hastelloy N 74533 A Balzance 11.67 7.02 0.03 <0.01 0.03 0.05 2.17 048 0.14 . 0.03 <001 <0.002 0.002 0.005
B 11.37 17.30 0.04 0.15 009 1.75 054 0.02 0.1 0.0015 0.0001 0.18
Inconel 600 NX 3752 A 75.58 15.03 8.62 0.27 0.16 0.03 0.28 0.007 -
Modified Hastelloy N 74534 A Balance 11.66 7.12 0.06 <0.01 0.03 008 2.09 053 0.14 003 002 <0.002 <0.002 0.007 0.013
B 11.58 7.67 0.02 0.1 0.05 1.77 054 0.02 0.05 0.0012 0.010 0.0002 0.0020

%A — vendors; B — ORNL.

Lyl
and until their final sectioning for tensile and
analytical specimens.

The three fuel pins were gamma-scanned in the
HRLEL with a sodium-iodide detector using a 17-
in.-thick lead collimator with a 0.010-in. X I-in.-wide
slit (Fig. 6.86). The pins were first scanned using a
single-channel analyzer and the differential from0.55
MeV to 0.75 MeV with 5000 counts/ sec full scaleand
20% standard deviation. An aluminum spacer s in.
long was placed between each pin. Each fuel pin
recording has a spike which indicates sudden
increased activity at the top surface of the fuel salt;

148

the top pin (No. 2, heat 74533) has three spikes for
this region but did not have a spike for the bottom of
the fuel, as did the other two pins. None of the pinsin
TeGen-1 had a spike for the bottom of the fuel. The
only difference in TeGen-1 and TeGen-2 construc-
tion was that TeGen-2 had a hub with four small foils
s in. wide X ¥g in. long X 0.005 in. thick welded to itat
the top (gas plenum) and bottom (immersed in salt).
The top pin (No. 2, heat 74533) had a void at the
bottom of the pin, with no salt around the bottom
hub and foils.

ORNL-DWG 7610075

MODIF|
HEAT
PIN N

ED HASTELLOY N
74533
0. 2-TOP PIN

\

INCONEL 600

PIN ROC.4 —MIDDLE P

N

)

i

VERTICAL SCAN

!

MODIFIED HASTELLOY N
HEAT 74534
PIN NO. 6 -BOTTOM PIN

FULL SCALE : 300,000 counts /min
STANDARD DEVIATION: 2.0 %
ENERGY RANGE COUNTED: 0.55 TO O.75 MeV

r"

Qo 10 20 30 40

%

50
SCALE

60 70 80 90 100

Fig. 6.86. Gamma scans of fuel pins from the TeGen-2 experiment. Scans made vertically from bottom to top of each pin. Numbers

denote positions for activity vs energy scans.
149

Each of the fuel pins was scanned for elements highest for the middle pin, as expected, slightly lower
present at the bottom, middle, and top of the saltand for the bottom pin, and considerably lower for the
in the gas plenum, using the 512 channel analyzer top pin. There were differences in the intensities in the
with a sodium-diode detectorin HRLEL. A counting 0.6- to 1.0-MeV range where the peaks are contribut-
time of 15 sec was used for all scans except for those ed to by *Zr, ®Nb, **Co, and '*I. These differences
of the gas plenum and the background scans, which are not presently understood but may be related to
lasted 10 min because of the reduced activity. Figure the void space at the bottom of pin 2. These scans
6.87 is a plot of energy or activity counts per minute were made four days after removal from the reactor
per channel for each of the three fuel pins at the flux, and some of the short half-life fission products
bottom of the fuel. The scans were on the spike of are in evidence, whereas TeGen-1 had a longer (about
increased activity on the bottom of pins 6 (bottom) two months) cooling period and no evidence of short
and 4 (middle) and at an equivalent position at the half-life fission products.

bottom of the fuel in pin 2 (top). The activity was

ORNL-DWG 76-7666

" — T T T T |
[~ —taag, 99, 99 — - =10 ,140g, 103y ' 2
— [ e '_MQCe,ggMo. 991, -— —MOLU.MOBU.‘OBRU,u}[ . — 3 1
1 190 o 140g, 103p, 580, — — 59,
S _: }MOBO e
— | | =---- _}0321 —
| —_—me. || 7T —_——60r, |
_—_:—} Te — . — 957, 95y, 132 — — 132, 60¢,
S YL - T 132; 60,
21— | emmue= 952.,. ub —
140L°
—_3% —_——
58
...... Co —_— — | 140
go‘ — n ____}151] — c— ————— } L°~
K A — e } 132 —]
I I R A A fou -~
"
_ "
: Ia!
- { :' :\ ’\ I — 132 ]
SN b A ) St —
— iy 7 \ I\l i 1{/ —
~id: Y [PV
! 0 ;n :‘. ':'l' ’In“ +— i ——""Co e
! gt j S { 1 — —13,8%,
34 1 " : \
> !f’: Y / l'.. K \t gt L ‘ .....
. l"‘ U\./ ‘.'.,.-, I , \‘
€ T ) H
g ’ ; qj.' F
Kl . ) ;!
: |05 'l ‘/ h ! / \\ N
5 N - 5 W
2 4 | W ;
o y . v e
N ] 7 R L4
. L d Wil ‘
‘ L)
l .\ 1'\1* N\ i
1]
v
| VAN \
: \ " :
I ‘. e “. ! et '
2 } / "‘\ IA-‘V‘.VA *
| \ :
I HACKGROUND (CURVE A) ]
- == HASTELLOY N HEAT 74533 (CURVE B}
102 i e — INCONEL 600 {CURVE C)
} ----- HASTELLOY N HEAT 74534 (CURVE D) \
\—i
|
| o b
5
|| -
| "
1=
SCAN IN GAS SPACE \ %
: i
10}
o] 0.2 0.4 0.6 o8 10 1.2 1.4 1.6 1.8 2.0

ENERGY {Mev)

Fig. 6.87. Elemental scans of the TeGen-2 capsules with the detector located near the bottom of the capsule. Fifteen-second scan made
through a 0.010 X 1 X 17-in.deep slit.
Each of the fuel pins had a spike of activity at the
top of the fuel salt; Fig. 6.88 shows a plot of activity vs
energy at the peak of the spike. Some of the fuel salt is
present in each of the scans at the top of the salt, as
shown by the high '“’Lapeak at 1.596 MeV in the first
three scans. These scans are similar to those at the
middle of the salt, except that the activity is slightly
higher at the peak. The fourth scan in Fig. 6.88 was
made on the third spike above the salt in the gas
plenum on pin No. 2, shown as point 11 in Fig. 6.86.

150

Note the lower peak for “’La at 1.596 MeV,
indicating little, if any, fuel salt, and the higher peaks
for iodine and tellurium, which should migrate to the
gas space. This curve is similar to the ones at the
bottom of the salt of the middle and bottom pins,
except for the lathanum peak, shown in Fig. 6.87.
The differences in these scans and the reason for
deposition in layers, leading to spikes of activity in
the gas space in the top pin, are not understood at this
time. Analytical Chemistry samples were taken from

QORNL-DWG 76- 7667

5 MOLa'lO}Ru.MOBu'SBCO"BSI
07— . am 140; (103 1405 133 —
I —__: —_—— MOLO,‘OSHU,MOBO,SBCo,usl —
— T T 184ce P, 99, mm oo 190, o 103, 140g 5B 133 S
5 |—— 132 —
::gao ‘:ICe e so(l:o
———— Bo,'%ce 132, 59
...... 140, _ 14 s 432 — 7], Y Fe
Bo," Ce — } v - 12, ]
952r 95Nb 1321
» — 1327, M. 957, 95y 1321 58¢, —
...... — 957, 95y 132] — 50,
_ St 95,5, 99y, 132 —_— 132
| == e -l
______ 60
w0 — Co
5 — :—}N‘OLO —]
H
o 2 'v‘
.
=
2
5
4 10° |
| §
| ]
t
? ]
| \\
102 \
1
vl \
\ |
i Wi \
HASTELLOY N HEAT 74533, 1st SPIKE (CURVE A} ‘L! . \
b == .e= HASTELLOY N HEAT 74533, 3rd SPIKE (CURVE B) ¥ 1
— e JNCONEL 600 (CURVE C) h \
T e HASTELLOY N HEAT 74534 (CURVE DI "\k
V&
2
10'
o} 0.2 0.6 o8 1.0 1.2 1.4 1.6 2.0

ENERGY (MeV)

Fig. 6.88. Elemental scans of the TeGen-2 capsules with the detector located at the spike at the salt-gas interface. Fifteen-second scan
made through a 0.010 X 1 X {7-in.-deep slit.
these various areas, and the results may help to
explain the counting data.

Scans were taken in the gas space of each of the fuel
pins, but the peaks are similar to the one shown in
Fig. 6.88, scan 11, except that the activity was much
lower in areas not on activity spikes. Scans were also
made at the middle of the fuel of each pin, but they
were very similar to scan 8 at the bottom of the No. 2
fuel pin shown in Fig. 6.87. Analytical Chemistry has
a multichannel analyzer with a germanium detector
that can better separate the peaks of the different
fission products, and these analyses will better define
which elements are present.

The fuel pins were measured in HRLEL for
diametral change by using a V-block and dial gage
arrangement. Table 6.33 shows the changes in
dimensions to be very small, with a maximum of
0.3% 1increase in the diameter in the gas space of pin2
at the top. Prior to measuring, the pins had been
maintained at 150°C in a furnace, removed, and then
allowed to cool. It 1s possible that they were still
slightly warm when measured. An unirradiated
control pin measured in the same way showed a slight
negative change, -0.10%, and it had not been heated
previously. These changes are very small and are
probably not significant. The pins were sectioned in
the same manner as TeGen-1 by gripping the pin in a

TYPE DESCRIPTION AND USE

A Y46 in. RING FOR MECHANICAL

PROPERTIES
1/ain. FOR LEACH (2 STEP)
' SECTION TO BE RETAINED

151

rotating chuck and using an abrasive cut-off saw. The
sections cut and their respective depositions are
shown in Fig. 6.89. The salt surfaces, or menisci, were
photographed during sectioning (Fig. 6.90). The salt
level varied in pin 2. Since the salt level was higher in
pin 2 due to the void at the bottom, pin 2 was
photographed after cut A-2, and pins 4 and 6 were
photographed after cut A<4. The surfaces look
different for each of the pins, as did TeGen-1, but the
reason for these differences is not known. It could
possibly be related to slightly different cooling rates
for each fuel pin during solidification as they were
removed from the irradiation flux. Chemistry
samples of the salt were taken from the surface of
each of the fuel pins, and the results of these analyses
may help explain these differences.

The pins were sectioned through cut A-7 from the
top of the pin, then reversed, and section C-1 placed
in the rotating chuck and the bottom end cap cut off
with ring A-16 still attached so that the foils could be
removed without damage; then cuts A-16 through A-
8 were made. The top pin (No. 2, heat 74533) had a
void at the bottom of the pin, and no salt wasaround
the foils welded to the hub. Figure 6.91a shows a
photograph of the bottom hub and foils surrounded
by the tube (segment A-16). There is no salt around
the foils or in the bottom of the pin on top of the end

ORNL-DWG 75-9062R

DN

_______

DN

Fig. 6.89. Schematic diagram of individual fuel pin showing the locations of test specimens.
cap. Figure 6.916 is a photograph of the bottom of
the salt that rested on top of the hub in Fig. 6.91a.
Note the particles lying on the hub (Fig. 6.91a) and
the matching impressions in the salt(Fig. 6.915). The
salt receded from the bottom area in a meniscus, and
the saw accidentally cut into it part of the way
around. Itis notunderstood how this void could have

152

occurred, removing the salt so cleanly from the area
of the foils and raising the entire volume of salt
greater than s in. higher in the tube. X-ray
photographs were taken of the fuel pin before it was
installed in the experiment, and there was not a void
in the bottom of the pin at that time. The fuel pins
were maintained at 700°C approximately 1600 hr,

Table 6.33. Diameter measurements of fuel pins from TeGen-2

Pin Postirradiation Average Preirradiation Diameter
number Position diameter diameter average diameter change
(in.) (in.) (in.) (%)
2 Top 0.5008 0.5006 0.5001 0.10
Top 0.5004
Middle 0.5006 0.5007 0.5004 0.06
Middle 0.5007
Bottom 0.5003 0.5000 0.4991 0.18
Bottom 0.4997
Gas 0.5017 0.5010 0.4995 0.30
Gas 0.5003
Average 0.5006 04998 0.16
4 Top 0.5003 0.5005 0.5003 0.04
Top 0.5007
Middle 0.5012 0.5012 0.5003 0.18
Middle 0.5012
Bottom 0.5015 0.5013 0.5002 0.22
Bottom 0.5011
Gas 0.4999 0.4999 0.5002 -0.06
Gas 0.4999
Average 0.5007 0.5003 0.08
6 Top 0.5007 0.5010 0.5005 0.10
Top 0.5012
Middle 0.5008 0.5008 0.5009 -0.02
Middle 0.5007
Bottom 0.5002 0.5004 0.5009 -0.10
Bottom 0.5005
Gas 0.5006 0.5007 0.5004 0.06
Gas 0.5008
Average 0.5007 0.5007 0.00
39 Top 0.4994 0.4995 0.5001 -0.12
Top 0.4996
Middle 0.4995 0.4995 0.5002 —-0.14
Middle 0.4994
Bottom 0.4994 0.4995 0.5001 -0.12
Bottom 0.499%6
Gas 0.4994 0.4998 0.5001 -0.06
Gas 0.5001
Average 0.4996 0.5001 -0.10

%Unirradiated control.
R-70701

Fig. 6.90. The meniscus of the fuel salt with the top of the pin ‘
removed. 6X. (@) Modified Hastelloy N, heat 74533, top pin 2; (b)
Inconel 600, heat NX 3752, middle pin 4; (¢) modified Hastelloy N,
heat 74534, bottom pin 6.
PST

| 0.17in |
I 4.2mm |

Fig. 6.91. TeGen-2 fuel pin 2 (top pin). 6X. (a) Bottom of pin
showing hub and foils free of salt, (b) bottom of salt that rested on
hub showing void space around edge of tube, and (c) ring A-14 cut
from pin and showing bubbles in the salt (light green).

and the salt would most certainly have been molten
during that time. Figure 6.91¢ shows the next ring
above the bottom of the salt shown in 6.91b after
it was cut off with the abrasive cut-off saw. (The salt
with saw markings in it and bubbles around it is still
in the ring.) The bubbles were very light green,
contrasted with the dark green of the rest of the fuel
salt. Evidently, some of the salt was melted by the
heat generated by the uncooled cut-off wheel ( ;s in.
away) while cutting the back side of the ring. Some
bubbles were noted in the cutting of each ring near
the bottom, but not as many formed near the top of
pin 2, and very few were noted in cutting the other
two fuel pins. Fuel pin 2 (top) and fuel pin 6 (bottom)
are almost identical in composition, and the same
batch of salt was used to fill all the fuel pins; so this
variable behavior is not understood. Pin 6 (heat
74534) contained 0.01% lanthanum, but this should
not change the thermal conductivity appreciably, and
the salt was well mixed by sparging with hydrogen
and argon before the fuel pins were filled.

155

Some of the A rings (Fig. 6.89) were tensile tested
on an Instron tensile machine in the Building 3026D
hot cells. The ring fixture was modified so that the
ring could be held in place. Figure 692 is a
photograph of the fixture with a 0.5-in.-diam X 0.35-
in.-wall X /js-in.-long ring in place and the top clamp
partially rotated to hold the ring to prevent its sliding
out. During testing, as the two halves of the fixture
were pulled apart, both clamps were rotated to the
closed position. This modification worked well and
alleviated the problems encountered in testing of
TeGen-1 rings. Tensile data for the top, middle, and
bottom irradiated fuel pins are given in Tables 6.34,
6.35, and 6.36 respectively. An arbitrarily assumed
gage length of 0.5 in. used in the calculation of
the yield strength and the uniform and total
elongations. The values given in the tables are not
absolute values but are useful for comparisons within
the series. The yield strengths appear to be slightly
increased and the elongations and reductions in area
slightly decreased by the irradiation and exposure to

¥-136070 |

Fig. 6.92. The fixture for testing rings in tension, showing ring in position for testing.

156

Table 6.34. Tensile properties of modified Hastelloy N, heat 74533,
rings sawed from pin 2 and tested at 25°C

Cross 3 .
Specimen sectional” Stress (107 ps) Elongation” (%) Reduction
number area Proportional Yield® Ultimate ¢\ ture  Uniform Totat in area (%)
1072 in.% limit tensile
2A-1 4.27 65.6 71.0 1246 121.3 29.7 28.5 279
2A-2 5.80 55.2 61.6 108.4 104 4 19.5 204 19.7
2A-3 7.07 47.4 533 110.6 104.7 19.9 214 27.2
2A4 5.33 619 67.5 1234 108.7 24.3 26.3 12.2
2A-5 5.03 60.7 66.7 117.8 110.0 23.8 249 16.8
2A6 6.50 57.7 66.2 117.0 111.2 19.8 214 27.0
2A-7 5.16 66.9 70.8 122.9 118.2 19.3 204 18.8
2A-8° 5.19 53.0 574 83.6 684 7.0 10.7 17.7
2A-11 5.07 62.2 67.1 120.2 1154 25.8 27.2 26.7
2A-13 3.50 62.3 68.0 114.9 1094 21.8 22.7 31.1
2A-16b 341 85.1 93.3 126.7 121.7 7.9 8.8 26.3
Control 579 118.3 116.9 49.8 51.3 36.7
Control 57.1 1154 1133 48.6 492 40.5
9Based on gage length of 0.5 in.
bRing notched during sawing.
Table 6.35. Tensile properties of Inconel 600, heat NX 3752,
rings sawed from pin 4 and tested at 25°C
Cross 3
Specimen sectional 5 Stress (10 pSl). Elongation® (%) Reduction
number area Proportional  y qa  Ulimate o o e Uniform  Totar  inarea (%)
‘ (1 0—3 in.2) limit tensile '
4A-1 5.11 45.0 509 93.0 85.5 22.0 23.7 28.6
4A-2 6.13 37.5 42.1 84.5 73.4 249 27.9 27.1
4A-3 3.52 34.1 38.9 84.1 57.7 21.0 26.6 429
4A4 3.96 36.9 389 76.5 64.4 19.0 22.7 28.1
4A-5 3.97 40.3 443 85.4 69.8 204 249 14.6
4A-6 5.72 34.1 393 86.4 59.5 204 254 443
4A-7 4.76 36.4 40.5 86.6 79.8 21.0 234 334
4A-8 4.06 40.6 44 .8 840 77.6 18.9 218 327
4A-11 5.88 434 46.8 86.6 77.2 22.6 26.4 36.4
4A-13 5.34 35.6 39.7 86.7 76.2 255 29.6 20.6
4A-16 4.13 45.3 48.9 94.2 843 18.9 22.6 22.9
3A-1° 4.86 35.4 38.4 83.0 77.1 30.6 33.0 31.6
3a4-2° 4.65 36.6 39.8 83.7 78.5 22.6 25.0 22.9
3A-37 494 39.5 47.5 89.0 779 24 3 27.6 16.0
3A-4° 5.11 45.0 479 87.7 79.3 202 234 18.8
3A-5b 4.33 379 42.7 88.2 79.6 245 27.3 21.1
Control 35.8 82.5 69.3 40.6 445 44 .1
Control 36.2 82.3 68.4 39.7 437 441

?Based on gage length of 0.5 in.
Control pin filled with salt at same time but not irradiated or tested in hot cell.

157

Table 6.36. Tensile properties of modified Hastelloy N, heat 74534,
rings sawed from pin 6 and tested at 25°C

Cross 3.
Specimen sectional , Shrem (0 pSl). Elongation” (%) Reduction
number area Prop 0 r?lonal Yield” Ultm!ate Fracture Uniform Total in area (%)
a 03 in.z) limit tensile -
6A-1 4.60 65.2 71.1 118.7 © 1179 17.9 18.2 32.9
6A-2 5.24 64.9 70.3 118.8 114.6 226 231 22.1
6A-3 5.09 60.3 66.8 1199 114.0 21.2 21.9 18.3
6A4 5.13 63.3 71.7 126.5 124.7 23.0 233 19.4
6A-5 5.29 59.0 66.8 120.7 117.3 21.7 226 27.5
6A-6 4.68 66.2 72.2 1224 117.5 17.6 18.2 14.1
6A-7 - 4.28 67.0 72.1 125.8 123.7 254 25.9 23.0
6A-8 6.11 55.2 63.0 119.5 116.2 21.1 21.7 13.5
6A-11 4.91 54.0 59.1 117.8 114.7 219 224 28.0
6A-13 3.68 66.7 72.7 127.1 1216 = 214 22.6 18.8
6A-16 4.50 60.7 65.6 106.2 103.3 16.7 17.1 25.0
Control 67.9 121.4 118.2 33.6 35.1 37.3
Control 67.7 122.7 121.9 355 36.2

40.0

9Based on gage length of 0.5 in.

fission products. The elongations are still approxi-
mately 20% (except for some rings that had small
notches accidentally sawed in them), which is
acceptable. These changes are due to the combined
effects of being at 700° C for approximately 1600 hr
of exposure to the thermal neutrons and of the
shallow cracking caused by the fission products.
The rings were numbered consecutively from the
top of each fuel pin (Fig. 6.89). The fuel salt level was
higher in the top pin (approximately at ring A3) than
the middle and bottom pins (approximately at ring
A6) due to the void at the bottom of the top pin
elevating the salt. As discussed previously, ***’ the
most severe cracking in the MSRE occurred near the
salt-to-gas interface in the pump bowl. In the
standard Hastelloy N pin in TeGen-1, the rings at the
top of the gas plenum (A-1 and A-2) cracked almost
as severely as at any other location (approximately
2.0 mil deep). Figure 6.93 shows photomicrographs
of the top pin (modified Hastelloy N, heat 74533)
rings after tensile testing at 25° C. Rings A-1 (top) and
A-16 (bottom) did not develop cracks. Both rings
were shielded to some extent by the foils welded to
the hub in the center (Fig. 6.91a). (Ring A-16 was
later cut from the area surrounding the foils.) Ring A-
1 was surrounded by similar foils at the top or gas
space of the same fuel pin. As previously noted, the
void at the bottom of the pin would indicate that both
rings (A-1 and A-16) were in a gas space not exposed
to the fuel salt. Ring A-2 (Fig. 6.916) was just above
the salt level, ring A-3 (Fig. 6.91¢) was at the surface

of the salt, ring A-8 (Fig. 6.91d) was at the middle of
the salt. Ring A-2, justabove the salt surface, appears
to have deeper cracks than the other rings.

Rings of the middle pin (Inconel 600, heat NX
3752), after tensile testing at 25° C, are shown in Fig.
6.94. Ring A-1 was at the top of the gas plenum, ring
A-5 was just above the salt, ring A-6 was at the
surface of the salt, ring A-8 was at the middle of the
salt, and ring A-16 was at the bottom of the salt. Ring
A-1 (Fig. 6.944) at the top of the gas plenum did not
crack on the inside (exposed to gaseous fission
products) or on the outside exposed to NaK. Rings
A-15 (Fig. 6.94b), A-6 (Fig. 6.94c), and A-8 (Fig.
6.94d) cracked on both the inside, exposed to saltand
fission products, and on the outside, which was
exposed to NaK. Ring A-16 (Fig. 6.94¢) did not crack
as severely on the inside as the rings near the top and
middle of the salt and did not crack on the outside,
exposed to NaK. Unless there was a severe tempera-
ture gradient along the length of the fuel pin, no
explanation for this varied behavior can be given at
this time.

Rings of the bottom pin (modified Hastelloy N,
heat 74534) tested at 25°C are shown in Fig. 6.95.
Ring A-1 (Fig. 6.95a) was located at the top of the gas
plenum, A-5 (Fig. 6.95b) was just above the salt
surface, A-6 (Fig. 6.95¢) was just below the salt
surface, A-8 (Fig. 6.954) was at the middle of the salt,
and A-16 (Fig. 6.95¢) was at the bottom of the salt.
Again, note that ring A-1 at the top of the gas plenum
did not crack, as it was probably masked to some
R-71020

R-71026

R-71075

8ST

MICRONS
00X—Y .

) 690 7?0
T T
0.005 0.010 INCHES 0.020 0.025

100 200
1 1

Fig. 6.93. Fractures of rings tensile tested at 25° C from the top
fuel pin in TeGen-2, modified Hastelloy N, heat 74533. As
polished. 100X. (@) Location A-1 in the gas space, (b) location A-2
just above salt, (¢) location A-3 just below surface of salt, (d)
location A-8 at middle of salt, and (e) location A-16 at bottom of
pin but free of salt. In each case the lower surface was exposed to
the inside of the capsule, and the upper surface was exposed to the
NaK.

R-71042

R-71037

R-71087

6ST

MICRONS
L {00X—1y

R-71096 100
1

T

0.(;05 0.0|10 INCHES 0.020 0.025

!
—

200 600 700
1 3 1

Fig. 6.94. Fractures of rings tensile tested at 25°C from the
middle fuel pin of TeGen-2, Inconel 600, heat NX 3752.
(a)Location A-1 in the gas space, (b) location A-5 just above the
salt, (¢) location A-6 just below the surface of the salt, (d) location
A-8 at the middle of the salt, and (e) location A-16 at bottom of
salt. As polished. 100X. In each case the lower surface was exposed
to the inside of the capsule, and the upper surface was exposed to
the NaK.

R-71060

R-71101

R71111 |

100 200 MICRONS 600 700
i TR 100X . TR
0.005 0.010 INCHES 0.020 0.025

Fig. 6.95. The fracture of rings tensile tested at 25°C from the
bottom fuel pin in TeGen-2, modified Hastelloy N, heat 74534. (a)
Location A-1 in the gas space, (b) location A-5 justabove salt level,

salt, and (e) location A-16 at bottom of salt. As polished. 100X. In
each case the lower surface was exposed to the inside of the capsule,
and the upper surface was exposed to the NaK.

(¢) location A-6 just below salt level, (d) location A-8 at middle of

091

extent by the foils. Ring A-16 at the bottom of the salt
appears to have the most severe cracking of any of the
rings, which is opposite to the trends noted in the
middle and top pins.

The cracking in the Inconel 600 pin has slightly
different characteristics than the cracking in either of
the modified Hastelloy N pins. In our experiments
with Inconel 600 exposed to tellurium-containing
environments, some heats of Inconel 600 were
observed to crack, and other heats in the same
capsule did not. All the heats had almost identical
chemical compositions, and it appears that 15%
chromium is borderline for crack resistance in this
alloy. Inconel 601 with 23% chromium resisted
cracking much better in TeGen-1 than Inconel 600
did in TeGen-2.

Analytical chemistry samples were taken from
each fuel pin in the B sections (Fig. 6.89; designations
are from top to bottom). Specimen B-1 was from the
gas plenum, B-2 was from near the salt surface and
B-3 was from near the bottom of the salt. Fuel salt
from the A rings containing the surface of the salt in
each fuel pin was mechanically pressed from the rings
and sent for analysis. This surface salt probably
contained the elements causing the spike of activity in
this region. These chemistry samples should help in
understanding the distribution of the fission products
along the length of the fuel pins and the reason for the
different cracking behavior in different areas of the
fuel pins.

Also being investigated in the TeGen series of
experiments are other alloys which are more resistant
_to tellurium attack in our experiments and which
should be more resistant to fission products. Some of
these alloys were included in TeGen-3, scheduled for
completion of the irradiation period in mid-March,
and in TeGen-4, scheduled for insertion in the ORR
in mid-April.

REFERENCES

1. T. K. Roche et al., MSR Program Semiannu. Progr. Rep.
Aug. 31, 1975, ORNL-5078, pp. 65-69.

2. T. K. Roche, H. E.McCoy,and J. C. Feltner, MS R Program
Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 74-78.

3. T. K. Roche, B. McNabb, and J. C. Feltner, MSR Program
Semiannu. Progr. Rep. Feb. 28, 1975, ORNL-5047, pp. 63—66.

4, H. E. McCoy and J. R. Weir, Jr., “Development of a
Titanium-Modified Hastelloy N with Improved Resistance to
Radiation Damage,”
Structural Alloys for Thermal and Fast Reactors, Spec. Tech.
Publ. 457, American Society for Testing and Matenials, Philadel-
phia, 1969.

161

pp. 290-311 in [rradiation Effects in .

5. D. N. Braski, J. M. Leitnaker, and G. A. Potter, MSR
Program Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, pp.
84-91.

6. J. M. Silcock and W. J. Tunstall, “Partial Dislocations
Associated with NbC Precipitation in Austenitic Stainless Steels,
Phil. Mag. 10, 360-89 (1964).

7. D. N. Braski, J. M. Leitnaker, and G. A. Potter, MSR
Program Semiannu. Progr. Rep. Aug. 31, 1974, ORNL-5011, p.
74.

8. E. R. Stover and J. Wulff, Trans. Met. Soc. AIME, 215, 127
(1959).

9, R. E. Gehlbach and S. W. Cook, MSR Program Semiannu.
Progr. Rep. Feb. 28, 1970, ORNL-4548, p. 231.

10. J. R. Keiser, J. R. DiStefano, and E. J. Lawrence, MSR
Program Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, p.
94,

1. L. M. Toth and L. O. Gilpatrick, The Equilibrium of Dilute
UF; Solution Contained in Graphic, ORNL/TM-4056 (December
1972).

12. B. McNabb and H. E. McCoy, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 97-100.

13. B. McNabb and H. E. McCoy, MSR Program Semiannu.
Progr. Rep. Feb. 28, 1975, ORNL-5047, pp. 94-101.

14. C. F. Weaverand J. D. Redman, MSR Program Semiannu.
Progr. Rep. March 31, 1973, ORNL-4832, p. 48.

i5. D. R. Stull and G. C. Sinke, “Tellurium,” p. 200 in
Thermodynamic Properties of the Elements, Amer. Chem. Soc.,
Washington, D.C., 1956.

16. P. Budininkas, R. K. Edwards, and P. G. Wahlbeck,
“Dissociation Energies of Group Vla Gaseous Homonuclear
Diatomic Molecules 111. Tellurium,” J. Chem. Phys. 48, 2870
(1968).

17. S. Dushman, “Flow of Gases Through Tubes and Orifices,”
p. 94 in Scientific Foundations of Vacuum Technique, (Chapter 2,
2nd ed) ed. by J. M. Lalferty, Wiley, New York, 1962.

18. R. Hultgren, et al., Selected Values the Thermodynamic
Properties of the Elements, p. 19, Amer. Soc. Metals, Metals Park, .
Ohio, 1973.

19. The lithium tellundes used in these experiments were
prepared by A.D. Kelmer and D.Y. Valentine of the Chemistry
Division. ‘

20. J. R. Keiser, et al., MSR Program Semiannu. Progr. Rep.
Aug. 31, 1975, ORNL-5078, p. 102

21. J. Brynestad, MSR Program Semiannu. Progr. Rep. Aug.
31, 1975, ORNL-5078, p. 100.

22. H. E. McCoy and B. McNabb, /ntergranular Cracking of
INOR-8 in the MSRE, ORNL-4829 (November 1972).

23. R. E. Clausing and L. Heatherly, An Auger Electron
Spectroscopic Examination of Grain Boundary Fracture Surfaces
of Hastelloy N Irradiated in the MSRE, ORNL-TM (in prepara-
tion).

24. R.E.Clausingand L. Heatherly, MS R Program Semiannu.
Progr. Rep. Feb. 28, 1975, ORNL-5047, p. 104.

25. R. E. Clausing and E. E. Bloom, “Auger Electron
Spectroscopy of Fracture Surfaces in Irradiated Type 304
Stainless Steel,” pp. 491-505 in Proceedings, 4th Bolion Ldg.
Conf., June 1974, Clartor’s Publishing Division, Baton Rouge,
La, 1974.

26. “Fractography and Atlas of Fractographs, pp. 71,72, ASM
Metals Handbook, vol. 9, ASM Metals Park, Ohio, 1974.
27. J. R. Rellick, C. J. McMahon, Jr., H. L. Marcus,and P, W,
Palmberg, “The Effect of Tellurium on Intergranular Cohesion in
Iron,” Met. Trans. 2, 1492-94 (May 1971).

28. G. Henry, J. Plateau, X. Wache, M. Gerber, |. Behar,and C.
Crussard, “Mise en Evidence de Phenomenes d’Adsorption
Intergranulaire et Etude de Leurs Relations Avec La Fragilite
Intergranulaire,” Mem. Sci. Rev. Met. No. 4, 417-26 (1959).

29. R. E. Clausing and L. Heatherly, Molten-Salt Reactor
Program Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, p.
106.

162

30. H. E. McCoy, B. McNabb, and J. C. Feltner, MSR
Program Semiannu. Progr. Rep. Aug. 31, 1975, ORNL-5078, pp.
108—-122.

31. C. R. Hyman, MSR Program Semiannu. Progr. Rep. Aug.
31, 1974, ORNL 5011, pp. 81-85.

32. B.J. McNabband H. E. McCoy, MSR Program Semiannu.
Progr. Rep. Feb. 28, 1975, ORNL-5047, pp. 123-136.

33. B. J. McNabb and H. E. McCoy, MSR Program Semiannu.
Progr. Rep. Aug. 31, 1975, ORNL-5078, pp. 119-131._
163

7. Fuel Processing Materials Development

J. R. DiStefano

The - processes that are being developed for
isolation of protactinium and removal of fission
products from molten-salt breeder reactors require
materials that are corrosion resistant to bismuth-
lithium and molten fluoride solutions. Past experi-
ence has indicated that although the solubilities of
iron-base alloys in bismuth are low, they mass
transfer rapidly in bismuth at 500—700° C. The most
promising materials for salt reprocessing are mo-
lybdenum, Ta-10% W, and graphite.

Molybdenum has been tested‘in a wide range of
bismuth-lithium solutions for up to 10,000 hr and has
shown excellent compatibility. Thermodynamic data
and literature reports indicate that molybdenum will
also be compatible with molten fluoride mixtures.

Tantalum—109% tungsten also has excellent com-
patibility with bismuth-lithium solutions, but tests
are required to demonstrate its compatibility with
molten fluoride salts. A thermal convection loop was
constructed of Ta—109% W, and a 3000-hr test at
600~700° C with LiF-BeF,-ThF-UF,(72-16-11.7-0.3
mole %) was initiated. |

Graphite has shown excellent compatibility with
both bismuth-lithium solutionsland molten salts.
Although no chemical interaction between bismuth-
lithium solutions and graphite has been found, the
liquid-metal solution tends to penetrate the open
porosity of graphite. Graphite exposed to Bi—2.4 wt
% (42 at. %) Li in a molybdenum thermal convection

loop for 3000 hr at 600—700°C picked up large

quantities of bismuth and srr:1a11 quantities of
molybdenum. Recent tests have focused on evaluat-
ing the extent of dissimilar material interaction in
Mo-(Bi-Li)-C and (Ta—10% W)~(Bi-Li)-C systems.

7.1 SUMMARY OF
COMPATIBILITY STUDIES

J. R. DiStefano :
7.1.1 Ta~10% W Thermal Convection Loop Test

Tantalum-10% tungsten tubing of 0.875-in. OD X
O.QSO-in. wall was fabricated from 3.5-in.-diam rod
using extrusion and tube drawing techniques. This
material was then used to fabricate a thermal
. convection loop (Fig. 7.1). The loop contains
Ta-10% W tubular and tensile specimens in the
vertical hotand cold leg sections. The loop surge tank
located above the hot leg is also constructed of

H. E. McCoy

Ta—10% W, and the fluoride salt fill tank connected
to it is made of nickel. Tantalum resistance heaters
are used to heat the vertical hot leg, portions of the
two horizontal crossover lines, and the surge tank.
Calrod resistance heaters were used in other areas.
The entire loop system was heated to 700°C before
salt was transferred into the loop. After flow was
established, the Calrod heaters were turned off;
however, they will subsequently be used when the
loop is drained. The loop is circulating the salt
mixture LiF-BeF,-ThF,-UF, (72-16-11.740.3 mole
%) and has the temperature distributions shown in
Table 7.1. The test is being conducted in a vacuum

Table 7.1. Temperature distribution in
Ta—10% W thermal convection loop

e Temperature

Location °0)
Bottom of hot leg 610
Top of hot leg 690
Top of cold leg 680
Middle of cold leg 640
Bottom of cold leg 585
Surge tank 690

bell jar at a pressure of 1 X 107" torr, and has operated
for 400 of a scheduled 3000 hr.

7.1.2 Dissimilar Material Tests

It was previously found that mass transfer of
molybdenum to graphite occurred in a thermal
convection loop test' after 3000 hr at 600-700°C.
Since various components of a chemical processing
system might well be constructed of graphite and a
refractory metal, tests were begun to evaluate the
extent of interaction of such dissimilar materials
when they are both in contact with the same solution.
Four test variables were investigated (Table 7.2).
Molybdenum or Ta—109; W specimens were heated
for 1000 hr in a graphite crucible containing bismuth-
lithium at 600 or 700°C. A summary of the results
obtained are shown in Tables 7.3 and 7.4, and typical
metallographic results are shown in Fig. 7.2.

The conclusions that can be drawn from these
results are that (1) the extent of interaction was slight
164

Fig. 7.1. Ta-10% W thermal convection loop.

165

Table 7.2. Test variables in refractory metal-graphite-(Bi-Li) tests

Variable Description
Refractory metal Mo, Ta—-10% W
Graphite AT]J, pyrolytically coated graphite
Temperature 600, 700°C |

Lithium concentration of solution

Bi, Bi—0.01% Li, Bi—2.5% Li

Table 7.3. Results from Mo-~(Bi-Li)-graphite
system after 1000 hr at 600 or 700°C

Mo wt Carbon concentration Concentration
Solution change (ppm) in graphite (ppm)
(mg) Solution Mo sample Mo Bi Li
600°C
Bi 0 34 55 20 <20 <2
-0.3 28 47 <10 <20  <°
Bi—0.01% Li -2.3 51 58 30 30 50
Bi—2.5% Li -0.8 157 85 <10 100 250
~0.6 66 71 <10 30° 2507
700°C
Bi -0.8 42 50 <10 90 <2
-0.2 21 43 <10 <20* <2°
Bi—0.01% Li -22 87 55 50 20 60
Bi—2.5% Li 0 49 260 <10 100 200
0 46 240 <10 209 2007

4Pyrocarbon-coated graphite sample.

Table 7.4. Weight and chemical changes in (Ta—10% W)-graphite(Bi-Li) solution after 1000 hr

Ta-10% W Carbon concentration Concentration in graphite
Solution weight change - (ppm)
(mg) Solution Ta-10% W Ta W Bi L
600°C
Bi —0.8 80 50 <500 <500 150 2
135 67 <500 <500 1707 24
Bi—0.01% Li -0.3 52 62 <500 <500 200 100
Bi—2.5% Li 100 40 <500 <500 500 500
90 45 <500 <500 20% 600
700°C
Bi -0.2 148 40 <500 <500 <20 <2
45 45 <500 <500 504 <24
Bi-0.01% Li +3.1 50 45 <500 <500 120 70
| 120°
Bi—2.5% Li +0.2 90 75 <500 <500 7000 600
105 70 <500 <500 1804 700

9pyrocarbon-coated graphite sample.

Sample not acid cleaned; all other results obtained on Ta—10% W samples after acid cleaning.
166

Molybdenum e iisag

20 40 60 MICRONS |%0 140
k L1 — 500X — ==L i
0.001 INCHES 0.005
Ta—10% W

Fig. 7.2. Metallographic appearance of molybdenum and Ta-10% W after exposure to Bi-2.5% Li in a graphite capsule for 1000 hr at
700°C.

to negligible in all cases, (2) carbon mass transfer
increased with temperature, (3) carbon mass transfer
was greatest in Bi—0.019% Li when tested with
molybdenum, and (4) penetration into graphite was
greatest by Bi—2.5% Li.

167

REFERENCE

I, J.R.DiStefano, MSR Program Semiannu. Progr. Rep. Aug.
31, 1975, ORNL-5078, pp. 133-39.
Part 4. Fuel Processing for Molten-Salt Reactors

J. R. Hightower, Jr.

Part 4 deals with the development of processes for
the isolation of protactinium and for the removal of
fission products from molten-salt breeder reactors.
Continuous removal of these materials 1s necessary
for molten-salt reactors to operate as high-
performance breeders. During this report period,
work on the chemistry of fuel reconstitution was
resumed, and engineering development progressed
on the metal transfer process for rare-earth element
removal, on continuous fluorinators for uranium
removal, on fuel reconstitution, and on salt-bismuth
contactors to be used in reductive extraction.

Studies of the chemistry of fuel reconstitution were
resumed. A test of the effectiveness of smooth
platinum for catalyzing the hydrogen reduction of
U’* to U* in small gold equipment has shown thata
smooth platinum sheet of limited surface area would
provide appreciable catalytic activity in the hydrogen
reduction column of the Fuel Reconstitution
Engineering Experiment. Niobium is an important
fission product with volatile fluorides and would be
carried from the fluorinator to the fuel reconstitution
step. Studies of the hydrogen reduction of NbF,
showed that in the absence of granular platinum the
NbF. was reduced slowly to Nb’. In the presence of
granular platinum the rate of NbF4 was rapid for the
first 2 hr but decreased to a value similar to that
experienced in the uncatalyzed reaction thereafter.
The reason for this behavior is being sought, since, if
it is due to poisoning of the platinum catalyst, it has
significant implications for the use of platinum
catalysts in a reactor-fuel processing plant.

Studies in Metal Transfer Experiment MTE-3B
were continued. In this experiment, all parts of the
metal transfer process for rare-earth element removal
are demonstrated using stirred, nondispersing
contactors and salt flow rates that are about 1% of
those required to process the fuel salt in a 1000-
MW(e) MSBR. During this report period, it was

168

determined that the previously observed entrainment
of the fluoride salt into the LiCl resulted from

_operation of the agitators in the contactor vessel at 5

rps. The entrainment was unexpected since none was
seen in experiment MTE-3 under similar conditions.
Tests showed that no entrainment occurred at
agitator speeds up to 4.58 rps. Also during this
period, two additional runs were made at agitator
speeds of 4.17 and 1.67 rps to determine the effect of
agitation on the transfer rate of neodymium from the
fluoride fuel salt to the Bi-Listripper solution. Before
these runs were made, the LiCl and Bi-Li stripper
solution, contaminated with fluoride salt, were
removed from the process vessels and were replaced
with fresh LiCl and Bi-Li. Results of the two runs
show that the rate of transfer of neodymium
increased 300 to 400% when the agitator speed was
increased from 1.67 to 4.17 rps. However, overall
mass-transfer coefficients for neodymium were lower
than predicted by literature correlations, particularly
at the two LiCl-bismuth interfaces. An increase in
mass-transfer rates at these two interfaces would be
required for reasonably sized process equipment to
remove rare earth elements from a 1000-MW(e)
MSBR. Obtaining the desired increases at these
interfaces could be achieved by dispersing the LiCl
and bismuth since entrainment of bismuth into LiCl
would not be as serious as entraining bismuth into the
fluoride salt that returns to the reactor.
Mechanically agitated nondispersing salt-metal
contactors of the type used in experiment MTE-3B
are of interest because entrainment of bismuth into
the fuel salt returning to the reactor can be minimized
since very high/ratios of bismuth flow rate to salt flow
rate can be more easily handled than in column-type
contactors and since these contactors seem more
easily fabricated from molybdenum and graphite
than column-type contactors. Mass-transfer studies
using water and mercury to simulate molten salt and
bismuth are being done to determine how to
extrapolate results of mass-transfer measurements in
a small salt-bismuth contactor to the large sizes that
would be used in a fuel processing plant. During this
report period, mass-transfer-coefficient measure-
ments were completed for three stirred, nondispers-
ing water-mercury contactors, covering a wide range
of agitator diameters and speeds.

A nonradioactive demonstration of frozen-salt
corrosion protection in a continuous fluorinator
requires an internal heat source not subject to attack
by fluorine. To provide such a heat source for future
fluorinator experiments, studies have continued of
autoresistance heating of molten salt in a fluorinator
mock-up. These tests have shown that a major
problem with the present design is plugging that
occurs in the unheated end of the salt inlet tube,

169

which also serves as an electrode for autoresistance
heating. Preliminary testing of the effectiveness of
frozen-salt films for protecting against fluorine
corrosion has continued.

The uranium removed from the fuel salt by
fluorination would be returned to the processed salt
in the fuel reconstitution step. The equipment is being
installed for an engineering experiment to demon-
strate the fuel reconstitution step. During this report
period the salt metering system was tested by flowing
salt at the design flow rate [100 cm’/ min (1.67 X 107
m’/s)] through the process vessels. Also, the UFs
supply system was tested and calibrated, and the two
gas density cells for off-gas analysis were calibrated.
The equipment now works sufficiently well that gold-
lined equipment can be installed and the fuel
reconstitution step demonstrated.
170

8. Chemistry of Fluorination and Fuel Reconstitution

M. R. Bennett

An initial investigation' of the reaction

2UFs(d) + Ha(g) — 2UF4(d) + 2HF(g) (1)

indicated that the reaction rate was low. Subsequent-
ly, it was shown™’ that the reaction follows zero-
order kinetics with a rate constant of 1.35 milli-
moles/hr and that the reaction can be catalyzed by
platinum. A series of experiments at 550° C (823 K)
has been begun to test the catalytic activity of
platinum in forms that could be used in the fuel
Reconstitution Engineering Experiment (FREE)
reduction column and to investigate the order of
reaction of the catalyzed reaction. In the previous
tests the catalyzed reaction was so rapid that only one
sample could be taken before completion of the
reaction.

During this report period the experimental facility
in Building 4501 was reactivated. After the first tests,
problems were encountered with plugged lines in the
UFs generator system. The plugged lines were
replaced, and additional heating tapes and insulation
were used to prevent the occurrence of cold spots in
the generator system.

To evaluate the catalytic activity of smooth
platinum, a 1.25-in.(31.8-mm)-diam disk of e~
in.(1.6-mm)-thick platinum sheet was pressed into a
1.75-in.(44.5-mm)-diam by ’/3-in.(24-mm)-thick
gold disk so that only the surface of the platinum
(area, 8.6 cm”) would be exposed to the molten fuel

salt. This disk was placed in the bottom of the gold -

reaction vessel; 0.200 kg of fuel salt, LiF-BeF,-ThF,
(72-16-12 mole %), containing about | wt % uranium
as UF, was added; and the system was hydrofluori-
nated for 24 hrat 600° C (873 K) to remove any oxides
present. Gaseous UFs was then introduced to yield a
final UFs concentration of about 2 wt %. The
temperature was reduced to 550°C (823 K), and
hydrogen was bubbled into the salt at 10 cm®/ min
- (1.67 X 107" m*/s) through a "e-in.(64-mm)-diam dip
tube. The melt was periodically sampled. At the end
of 2 hr, all the uranium was reduced to UF,.
Apparently even smooth platinum of limited surface
area affords appreciable catalytic activity; thus
platinum sheet that could readily be incorporated in
the FREE experiment should provide adequate
catalytic activity in an MSBR processing system..
To determine the reaction order of the catalyzed
reaction, two tests were carried out (test 16 UR with
2.04 wt % U and test 17 UR with 3.54 wt 9% U™).

A. D. Kelmers

Hydrogen was introduced at 20 cm’/min (3.3 X 107
m’/s), and samples were taken at 10- or 15-min
intervals. The resulting data (Fig. 8.1), when
normalized and plotted as In (U,’")/(U,”") vs time,
gave a straight line, suggesting first-order kinetics for
the catalyzed reaction with a half-life of 65 min.
Additional tests are planned to evaluate the effect of
hydrogen partial pressure and UFs concentration on
the reaction kinetics to try to determine if the reaction
exhibits true first-order kinetics or if it is pseudo first-
order because of an excess of one of the reactants in
these first experiments. First-order kinetics might be
expected of a reaction that is diffusion limited. This
could be diffusion of a reactant to the catalytic site or
diffusion of an active species or product from the
catalytic site.

In addition to uranium, fission products that
remain in the MSBR fuel salt and have fluorides that
are volatile at 500-600°C (773-873 K) will be
removed from the fuel salt by fluorination and will
appear in the fuel reconstitution step. Therefore, it is
of interest to establish the chemistry of typical fission
product fluorides; in particular, it would be desirable
to know if the zero-order kinetics and concomitant
low hydrogen utilization observed for UFs is unique

ORNL-DWG. 76-3812

half life

1 | i | | i
60 80 100 120
TIME {min)

Fig. 8.1. Catalytic reduction of dissolved UFs at 550°C. Test
16 UR, o;test 17 UR, o.

140
- 171

or represents a general case. Niobium, one of the
major fission products, should be volatilized as NbFs
from the fluorination step and would then react with
the UF, in the fuel reconstitution step for form-
soluble NbF.. Previous work® has shown that the
reaction

NbF.(d) + 2H, = 4HF + Nb° (2)
proceeded very slowly at 700°C (973 K) in Li,BeF..
According to Baes,” under these conditions, Nb*(d)
should be in equilibrium with Nb% that is, no
intermediate oxidation states should exist.

A series of tests was carried out using the gold
reactor and apparatus previously used for the
uncatalyzed UFs reduction experiments. In these
tests, 200 g of fuel saltand about 4 g of niobium metal
chips were placed in the reactor and hydrofluorinated
at 600° C (873 K) with anhydrous HF both to remove
oxides from the salt and to form dissolved NbF,. It
had been previously shown® that reaction of HF and
Nb® proceeds slowly. In four experiments, melts
containing 0.33 to 1.1 wt % NbF, were prepared. The
temperature was adjusted to 550°C (823 K), and
hydrogen was bubbled into the salt at 40 cm®/ min
(6.6 X 107" m’/s) through a '/s-in.(6.4-mm)-diam dip
tube. The melts were periodically sampled, and the
off-gas was scrubbed to trap HF, which was
subsequently determined as fluoride ion by a specific-
ion electrode. The reduction reaction proceeded very
- slowly. Scatter in the molten-salt solution analyses
prevented use of those data; however, the measure of
the HF evolution was useful. After normalization to
correct for an initial surge of HF generated when
hydrogen was initially admitted to the reactor, the
data from all four tests gave a single line of slope of
0.208 millimole niobium reduced per hour when
plotted in the form for zero-order kinetics (Fig. 8.2).
Thus it appears that the reduction of NbF, is similar
to that of UFs:.

A fifth test was then carried out with granular
platinum. The generation rate of HF was increased
for the first 2 hr and then fell to a value similar to that
recorded for the uncatalyzed condition. The quantity
of HF evolved during the first 2 hr was equivalent to
that for the reaction

ORNL-DWG, 76-3813

& | T
5 -
M ©
S g o
£
£ £
3 e
b o
= 3
%4 =)
w2l —os50 ¥
w . o
T rate constant = 0.208 mmole Nb/hr 4
- —o2s
3
| 1 | | L
f 2 3 4 S5 ]
TIME (hr)

Fig. 8.2. Reduction of dissolved NbF,; at 550°C. Test 1: o,
1.1 wt % niobium, pure hydrogen. Test 2: o, 0.92 wt %
niobium, Hj-Ar (30-70 mole %). Test 3: ©, 0.33 wt % niobium,
pure hydrogen. Test 4, », 0.72 wt % niobium, pure hydrogen.

NbFi(d) + '/, Hi(g) = NbFa(d) + HF . (3_5

Possibly an intermediate oxidation state, such as
NbF;, may be metastable and may offer a kinetic
barrier to the complete reduction to Nb®. Conversely,
the platinum catalyst may be poisoned by the
reaction products; so the apparent generation of HF
equivalent to the reduction to NbF; may be
coincidental. If the platinum catalyst is poisoned by
fission products, this could have significant implica-
tions for its practical use..

REFERENCES

. M. R. Bennett and L. M. Ferris, J. Inorg. Nucl. Chem. 36,
1285 (1974).

2. M. R.Bennettand A. D. Kelmers, MS R Program Semiannu.
Progr. Rep. Feb. 28, 1975, ORNL-5047, pp. 150-151.

3. A.D. Kelmersand M. R. Bennett, fnorg. Nucl. Chem. Lets.,
in press.

4. C.F. Weaverand J. S. Gill, MSR Program Semiannu. Prog.
Rep. Feb. 28, 1971, ORNL-4676, p. 85.

5.-C. F. Baes, Jr., J. Nucl. Mater. 51, 149 (1974).

6. C. F. Weaver and J. D. Redman, MSR Program Semiannu.
Prog. Rep. Aug. 31, 1970, ORNL4622, p. 73.
172

9. Engineering Development of Processing Operations
J. R. Hightower, Jr.

9.1 METAL TRANSFER
PROCESS DEVELOPMENT

H. C. Savage

The metal transfer process for the removal of rare-
earth-element fission products from Molten-Salt
Breeder Reactor (MSBR) fuel salt is being studied in
engineering scale experimental equipment designat-
ed MTE-3B.' The experiments carried out in MTE-
3B are to determine the rate of removal of
representative rare-earth elements from molten-salt
breeder reactor fuel salt (LiF-BeF.-ThFs, 72-16-12
mole %) and to measure the overall mass-transfer
coefficients between the salt and bismuth phases in
the mechanically agitated process vessels.

Results obtained from the first two experiments
using the rare-earth element neodymium (Nd-1 and
Nd-2) have been reported.’ During these experi-
ments, fluoride fuel salt was unexpectedly entrained
into the lithium chioride in the mechanically agitated
contactor. Agitator speeds of 5 rps were maintained
throughout both runs Nd-1 and Nd-2, duplicating
conditions that were previously used successfully in
similar equipment.”

During this report period, it was determined that
entrainment occurred when the agitators were
operated at a speed of 5 rps but would not occur at
lower speeds (up to 4.6 rps). The lithium chloride,
contaminated with fluoride salt, was removed from
the process vessels along with the lithium-bismuth in
the stripper vessel. Purified lithium chloride and
lithium-bismuth were added to the system, and two
additional experiments were completed. These
experiments were to determine the effect of agitator
speed on the removal rate and overall mass-transfer

coefficients across.the three salt-bismuth interfaces
for the rare-arth element neodymium. In one run
(Nd-3) the agitator speed was maintained at 4.17 rps.
In the second run (Nd—4) the agitator was maintained
at 1.67 rps.

Results obtained in these two experiments (Nd-3
and 4) are given in the following sections.

9.1.1 Entrainment Studies in
Experiment MTE-3B

Based on previous studies in a water-mercury
syste:m,3 it was concluded that entrainment of
fluoride salt into the bismuth and LiCl phases in the
mechanically agitated contactor would occur if the
agitators were operated at speeds of 5.0 rps or higher.
However, in the first metal transfer experiment,
MTE-3, entrainment was not observed at 5.0 rps but
was seen at 6.7 rps.’ '

Since fluoride salt entrainment occurred at an
agitator speed of 5 rps in experiment MTE-3B, tests
were made to determine the maximum allowable
agitator speed that could be used in experiment
MTE-3B without entrainment. These tests were
conducted by operating the agitators in the contrac-
tor at several different speeds (3.3, 4.2, and 5,0 rps)
for times periods ranging from ~50 to ~140 hr.
During each test at constant agitator speed, samples
of the LiCl were removed from the contactor and
analyzed for fluoride content -since an increase in
fluoride ion concentration would indicate entrain-
ment of fluoride salt into the LiCl. Results are shown
in Fig. 9.1, in which the fluoride ion concentrationas
a function of time is plotted for each agitator speed.
The initial concentration of fluoride ion of ~4 wt %

ORNL DWG 76-349

FLUORIDE ION
CONCENTRATION, wt %

p— 5.0rps-’+—'—-4.2 rps

4.6rps

I | L

X | 1 |

0] 40 80 120

160 200 240 280

TIME, hr

Fig. 9.1. Results of tests to determine the rate of entrainment of fluoride salt into LiCl as a function of agitator speed, MTE-3B.
represents the amount of entrainment that occurred
over a period of ~250 hr during runs Nd-1 and Nd-2.
The sequence of agitator speeds shown in Fig. 9.1
represents the order in which the tests were run.
Entrainment is clearly indicated in the ~50-hr test at
5.0 rps. At agitator speeds of 3.3 and 4.2 rps, no
entrainment (within experimental limits) appears to
have occurred over the ~200-hr combined test
periods.

It was concluded that future experiments could be
carried out at agitator speeds up to about 4.6 rps
without entrainment. It was also concluded that
periodic determination of fluoride ion concentration
in the LiCl to verify that no entrainment was
occurring could be done rapidly and would be
desirable in future experiments. For this purpose an
Orion Model 801A pH/mV meter* equipped with
specific-ion (fluoride) electrodes was obtained for
these experiments. This very accurate meter can also
be used to continuously measure and record the EMF
(~250 mV) between the two bismuth phases in the
contactor and stripper vessels containing different
concentrations of lithium reductant (~0.0015 and
~0.050 atom fraction lithium). Changes in EMF
would indicate a change in the relative lithium
concentrations in these phases, with a resultant
change in the equilibrium-distribution coefficient for
neodymium between the salt and bismuth phases.
Determination of the overall mass-transfer coeffi-
cients is dependent on the equilibrium distribution
coefficients.

9.1.2 Removal of LiCl and Bi-Li
Phases and Addition of
Purified Solutions

The lithium chloride (contaminated with fluoride
salt) in the contactor and stripper vessels and the
bismuth (depleted in lithium reductant) in the
stripper were removed without difficulty. The lithium
chloride and bismuth were removed by transfer into
an external receiver tank through a heated transfer
line and dip tube extending into the process vessels. A
new charging vessel was fabricated for subsequent
use in purification and addition of the lithium
chloride and bismuth—5 at. % lithium solution to the
process vessels.

Chemical analyses of the purified lithium chloride
in the contactor and stripper indicated that the

*QOrion Research Inc., Cambridge, Mass.

173

fluoride ion content had been reduced from ~5 wt %
to ~0.15 wt %. The 0.15 wt % fluoride content of the
purified lithium chloride indicated that a heel of
~180 cm’ (out of an original inventory of ~6100 cm")
of the contaminated lithium chloride was not
removed from the process vessels. This amount of
fluoride (0.15 wt %) would not be expected to affect
the equilibrium-distribution coefficients for neodym-
ium or thorium at the LiCl/ bismuth interface in the
contactor.” Analysis of the purified bismuth-lithium
in the stripper indicated a lithium reductant content
of 0.048 atom fraction Li—near the desired 0.050
atom fraction. However, the bismuth-lithium solu-
tion contained ~3000 ppm (wt) thorium—presuma-
bly from a precipitated thorium compound (ThBi,)
remaining in the stripper after removal of the
bismuth, which was saturated with thorium as a
result of the fluoride salt (LiF-BeF,-ThF,, 72-16-12

mole %) entrainment during runs Nd-1 and Nd-2. No

adverse effect on the transfer of neodymium would be
expected by the presence of thorium in the bismuth in
the stripper.

Following the addition of purified LiCl and Bi—5
at. % Li to the process vessels of Metal Transfer
Experiment MTE-3B, two experiments were con-
ducted as described below. The quantities of salt and
bismuth solutions in MTE-3B are shown in Table9.1.

9.1.3 Experiments Nd-3 and Nd-4

Two experiments, Nd-3 and Nd-4, were completed
in Metal Transfer Experiment MTE-3B. The experi-
ments were to measure the overall mass-transfer
coefficient for neodymium across the three salt-
bismuth interfaces in the metal transfer process and
to determine the effect of agitator speed on these
coefficients. For experiment Nd-3, an agitater speed
of 4.17 rps was maintained in the contactor and
stripper vessels, and for experiment Nd-4, an agitator
speed of 1.67 rps was maintained. The experiments
were made over a period of two weeks in January
1976, during which the process equipment was
operated continuously.

Experiment Nd-3. Prior to the start of experiment
Nd-3, 2030 mg of NdF; (1450 mg Nd) containing 149
mCi of '“'Nd tracer (1 1-day half-life) were added to
fuel salt in the reservoir tank. This addition increased
the neodymium concentration in the fuel salt from 21
ppm (wt), remaining after run Nd-2, to 34 ppm (wt).

Run Nd-3 was started at the following conditions:
agitator speeds (contactor and stripper) = 4.17 rps,
fluoride salt circulation rate = ~35 ¢cm*/ min (5.8 X
174

Table 9.1. Inventory of salt and bismuth phases in Metal
Transfer Experiment MTE-3B for runs Nd-3 and Nd4

Volume? Salt or
Material Vessel at 923 K Weight blsml.lth
( 3) (kg) solution
m (kg-moles)
Fluoride fuel salt? Reservoir 0.0294  97.0 1.535
(72-16-12 mole % LiF-BeF,-ThF4) .
Fluoride fuel salt Contactor 0.0031 10.2 1.61
Bismuth-thorium Fluoride salt side of contactor 0.0028 27.0 1.29
Bismuth-thorium LiCl side of contactor 0.0034 32.8 1.56
LiCl Contactor 0.0029 4.3 1.01
LiCl _ Stripper 0.0032 4.7 1.10
Bi—-5 at. % Li Stripper 0.0043 41.8 2.00

2Densities at 923 K: fluoride fuel salt = 3300 kg/m>, LiCl = 1480 kg/m>, and Bi = 9660 kg/m">.

bMole weight = 0.0632 kg.

10”7 m?/s), LiCl circulation rate = ~1200 cm’/ min
(2.0 X 10~ m’/s), temperature = 923 K. After 19 hr of
operation the fluoride salt circulation was stopped,
and the run was continued for a total of 108 hr, at
which time the LiCl salt circulation was stopped
while agitation at 4.17 rps was continued for an
additional 57 hr. The salt and bismuth phases were
sampled throughout the run at 4-to 8-hr intervals for
“'Nd counting and total neodymium analyses to
determine the rate of transfer of neodymium across
the three salt-bismuth interfaces.

The purpose of the initial period of fluoride salt
circulation (19 -hr) was to transfer neodymium
containing '*'Nd tracer into the fluoride salt in the
contactor. When this was accomplished, the fluoride
salt circulation was stopped. This procedure resulted
in an increased rate of change of Nd content in the
relatively small volume of 3.1 X 10”° m’ vs 3.5 X 107
m’ in the contactor and fuel salt reservoir and
improved the measurements of overall mass-transfer
coefficients. When the LiCl circulation was stopped
(after 108 hr), neodymium transfer into the stripper
was stopped, and the equilibrium distribution of
neodymium between the salt and bismuth phases was
established. This procedure was required to deter-
mine the equilibrium-distribution coefficients for
neodymium for calculation of the overall mass-
transfer coefficients.

Experiment Nd-4. At the start of run Nd-4, the
agitator speeds were reduced from 4.17 rps to 1.67
rps, fluoride saltand LiClsalt circulation were at ~35
cm’/min (5.8 X 107 m®/s) and ~ 1200 ¢cm’/ min (2.0 X
10”° m’/s) respectively, and the temperature of all

phases was ~650°C (923 K). The fluoride salt
circulation was continued for 21 hr to replace the
neodymium and '*’Nd tracer in the fluoride salt in the
contactor which was extracted during run Nd-3. The
concentration of neodymium in the fluoride salt in
the contactor had been reduced by about 509 during
run Nd-3. The experiment was continued with LiCl
circulation for 110 hr. Agitation of all phases at 1.67
rps (without salt circulation) was continued for an
additional 55 hr, again to determine the equilibrium-
distribution coefficients for neodymium between the
salt and bismuth phases. Samples of all phases were
taken at about 4- to 8-hr intervals during the
experiment for '“’Nd counting and total neodymium
analyses.

9.1.4 Results

The neodymium concentrations in the fluoride salt
in the contactor and the bismuth-5 at. % lithium in
the stripper during runs Nd-3 and Nd-4, based on
counting of the '“'Nd tracer, are shown in Figs. 9.2 to
9.5. Neodymium concentrations in the bismuth-
thorium and LiCl phases in the contactor and
stripper are not included since the '“’Nd counting
data do not accurately reflect the very low neodymi-
um concentrations (<! ppm) in these phases.
Concentrations in all phases, based on total neodym-
ium analyses using an isotopic dilution mass
spectrometry technique, will be used for the final
determination of overall mass-transfer and
equilibrium-distribution coefficients and will be
included in a final summary report.
175

~

ORNL DWG 76-350 R1

5.0
¢ : L T T 7T T T 17T T 17T T T T T 1
9 — —
< 1 STOPPED FLUORIDE i
@ 4.0 = SALT CIRCULATION S{:%:PEBAH&IJ —
£ | l l _
[=9
B N
- . 0, —
z 3.0 —|,9 .
— ,l_ —
g I
E 20l
z TR N
] - —
= ]
o I ) —)
S 4o —
] i
z J
% B ]
S N S NS SN T T N N N N A A T A
0 20 a0 60 80 100 120 140 160

TIME, bhr
Fig. 9.2. Neodymium concentration in the fluoride salt in the contactor during run Nd-3, MTE-3B.

ORNL DWG 76-35I R1

wn

' 40

o T T T T T T T T T T T T T T
x | ! S
o

< STOPPED FLUORIDE © © ©
E 3.0[~ SALT CIRCULATION ]
o

z- 1

© o0 ' n
E STOPPED LiCl

< CIRCULATION N
'—

=

wl

o ——t
P

Q

5 _
2 I B S e T
5 80 100 120 140 160

TIME, bhr
Fig. 9.3. Neodymium concentration in the bismuth—5 at. % lithium in the stripper during run Nd-3, MTE-3B.

ORNL DWG 76-352

SO T 71T T T T T T T T 1T T T T T 1
© | . ' —
= . STOPPED FLUORIDE : STOPPED LiCI

» 4 0} SALT CIRCULATION CIRCULATION 1
: l l
S — ]
[N
by =]

_ 3.0 , ]
5 | °° o o oy O ©
® 20 o © 7
< '
2k -
S 1ok _|
o
= _— _—
S
N L A S N (S TR T I T

0 20 40 60 80 100 120 140 160

TIME, hr
Fig. 9.4. Neodymium concentration in the fluoride salt in the contactor during run Nd4, MTE-3B.
176

ORNL DWG 76-353

5.0
o R A R U U R N A N AN R N N B
o) - —
<)

3 © O "]
on
.
E
E ]
b=
> . |
1=
'— —
=4
o
= 7
z STOPPED FLUORIDE STOPPED LiCl
S |__SALT CIRCULATION CIRCULATION _
o
o .
a9
=
o | |
¥ .

N NS S VUM W N NS SN SN SO AN S T B S

0 20 40 60 80 100 120 140 160

TIME, hr

Fig. 9.5. Neodymium concentration in the bismuth—5 at.

The lines drawn through the data points in Figs.
9.2t0 9.5 represent the “best-fit” values for the overall
mass-transfer coefficients at the three salt-bismuth
interfaces in the experiment.

Based on analyses for total neodymium,about 140
mg of neodymium was extracted from the fluoride
salt during run Nd-3, and about 40 mg of neodymium
was extracted from the fluoride saltduring run Nd-4.
The operating time for both runs was essentially the
same (110 hr). Thus the rate of removal of
neodymium was reduced by about 300 to 400% when
the agitator speed was reduced from 4.17 to 1.67 rps.
~ Results of five metal transfer process experiments
in which neodymium was used as the representative
rare-earth element fission product are shown in
Tables 9.2 and 9.3. One run, Nd-2, is not shown since
complete consumption of lithium from the stripper
alloy caused by entrainment of fluoride salt into the
lithium chloride precluded measurement of the
overall mass-transfer coefficients. Two of the runs
(EU-6 and EU-7) were conducted in experiment
MTE-3 during 1972, and three runs (Nd-1, Nd-3,and
Nd-4) were conducted in experiment MTE-3B during
1975-1976. .

As seen in Table 9.2, the overall mass-transfer
coefficients increase with increasing agitator speed,
as predicted, with one exception— K3 did not increase
in run EU-7 when the agitator speed was increased
from 3.33 to 5.0 rps. There is also good agreement
between runs EU-7 (MTE-3) and Nd-1 (MTE-3B) at
agitator speeds of 5.0 rps. In runs Nd-3 and Nd-4

% lithium in the stripper during run Nd-4, MTE-3B.

(MTE-3B) the overall mass-transfer coefficients were
decreased by about 300 to 4009% when the agitator
speeds in the contactor and stripper were reduced
from 4.17 to 1.67 rps. Also, the overall mass-transfer
coefficient values at the fluoride salt/bismuth-
thorium interface (K,) for all experiments in MTE-
3B are in the range of 50 to 100% of the values
predicted by the Lewis correlation, compared to the
159% of predicted value observed in the two
experiments in MTE-3. The overall mass-transfer
coefficients at the LiCl/bismuth-thorium and LiCl/

bismuth-5 at. 9% lithium interfaces (K, K3),
“however, are significantly lower than in run Nd-3 at

4.17 rps than would be expected from the values at
5.0 rps seen in runs EU-6, EU-7, and Nd-1.
Because of the limited number of experiments and
the rather large variation in the experimentally
determined overall mass-transfer coefficients, a
meaningful correlation of the data for use in design of
a metal transfer process system foran MSBR has not
been possible. An increase in the mass-transfer rates
would be needed for reasonably sized process
equipment to remove the rare-earth-element fission
products from a 1000 MW (e¢) MSBR, in the case of
neodymium at a rate of about 1.2 g-moles/day. For
run Nd-3 in MTE-3B, an increase of about 500% in
the removal rate of neodymium could be achieved if
the mass-transfer coefficients at the two LiCl/bis-
muth interfaces, K> and Kj;, could be increased
(without increasing the transfer coefficient at the
fluoride salt/bismuth interface) to about 20% and
177

Table 9.2. Overall mass-transfer coefficients for neodymium in the Metal Transfer Experiments MTE-3 and MTE-3B? (mm/sec)

. Amount
Asgll)::t:r Run fi:::; K1 K2 . Ks of Nd
No. Percent of Lewis Percent of Lewis Percent of Lewis  transferred

(rpm) (hr)  Measured correlation Measured correlation Measured correlation (mg)

200 EU-6 45  0.002 15 0.065 10 0.2

300 EU-70 55 0.0032 15 0.11 10 0.2

300 Nd-1¢ 96 0.004- 50-100 0.11 10 0.13 2.5 210

0.010 0.030 0.040
250 Nd-3¢ 110 0.016 {46) (0.014) 4) _ (0.0055) 1 140
100 Nd-4¢ 110 (0.0040) 69 8) (1.2) 41

@Mass-transfer area at each interface: A;, A; = 245 cm?; Aa =186 cm?, The terms K 1» K2, and K 3 are overall mass-transfer coefficients at
interface between phases (1) fluoride salt/bismuth-thorium, (2) LiCl/bismuth-thorium, and (3) LiCl/bismuth—S at. % lithium.

bExperiment MTE-3.
CExperiment MTE-3B.

Table 9.3. Equilibrium-distribution coefficients for neodymium

in the Metal Transfer Experiments MTE-3 and MTE-3B¢
Run No.
Calculated  Experimental  Calculated  Experimental  Calculated?  Experimental
Nd-1 0.0¥  0.03 1.67¢ 1.07 3.5 x 10 >1x 103
Nd-3 0.0179 0'.022 0.949 098 3.5 x 10* >1x 103
Nd-4 0.017¢ (0.022) 0.94¢ 3.5 x 10*
EU-6,7 0.012 0.30 3.5 x 10* 2x 10°

®The terms D 4+ P, Do are equilibrium-distribution coefficients between phases (A) bismuth-thorium/fluo-
ride salt, (B) bismuth-thorium/LiCl, and (C) bismuth—5 at. % lithium/LiCL

bBased on 5 at. % lithium in bismuth.
¢Based on 60 ppm lithium in bismuth.
dBased on 50 ppm lithium in bismuth.
€Experiment MTE-3B.

5%, respectively, of the value predicted by the Lewis
correlation, which is near the values observed in runs
EU-6 and EU-7. Large increases in these two
coefficients might be accomplished by increased
agitation, even to the point of dispersing the LiCl into
the bismuth phase.

9.2 MASS-TRANSFER STUDIES USING
WATER-MERCURY CONTACTORS

C. H. Brown, Jr. J. R. Hightower, Jr.

Mechanically agitated nondispersing salt-bismuth
contactors are being considered for the protactinium
removal step and the rare-earth-element removal step
in the reference MSBR processing plant flowsheet.
These contactors have several advantages over
packed-column salt-bismuth contactors:

1. They can minimize entrainment of bismuth into
the fuel salt returning to the reactor,

2. They can be fabricated more economically from
graphite and molybdenum, and

3. They can be operated more easily with large flow
rate ratios of molten salt.and bismuth.

Experimental development of these stirred contac-
tors has been carried out in two different systems, a
facility in which molten fluoride salt is contacted with
bismuth containing a dissolved reductant and a
system in which mercury and an aqueous electrolyte
phase are used to simulate bismuth and molten salt.
In the first facility, mass transfer of uranium and
zirconium was studied in a single stirred contactor
using actual process fluids.”” The studies with water
and mercury will allow many physical parameters
(such as size and configuration of the contactor
vessels and agitators) to be examined at relatively
small expense over a wide range; then results
measured in the salt-bismuth facility can be extrapo-
lated to the larger contactor sizes that will be present
in the processing plant.
An electrochemical technique has been developed
to measure water-side mass-transfer coefficients in
water-mercury contactors. During this report period
an extensive set of experiments using this technique
has been done by a group* from the MIT School of
Chemical Engineering Practice to determine the
effect of agitator diameter and speed and of cell size
on the water-side mass-transfer coefficient in square
water-mercury contactors.

9.2.1 Experimental Equipment and Procedure

The electrochemical technique for measuring
water-side mass-transfer coefficients is based upon
the diffusion-limited reduction of quinone to
hydroquinone at the mercury surface in the water-
mercury contactor:

@ =0+ 2H  + 26 = HO—@—OH.

(D

_—

O

As has been discussed previously,' reaction (1) can
be driven electrochemically in the forward direction
at the mercury surface (which acts as a cathode in an
electrochemical cell) and in the reverse direction at
another electrode (the anode of an electrochemical
cell) in contact with the aqueous phase of the water-
mercury contactor, The concentration of hydroqui-
none can be made much higher than the concentra-
tion of quinone, and the area of the anode can be
made much greater than the area of the mercury
surface. When this is done and an electric current is
passed between the two electrodes, the mercury
surface is polarized. The magnitude of the electric
current 1s then limited by the rate at which quinone is
transferred to the mercury surface, and the water-side
mass-transfer coefficient is related to the cell current,
the bulk concentration of quinone in the aqueous
phase, and the area of the mercury surface by the
relation

I

k= Arcy

(2

where
k = mass transfer coefficient, mm/sec;
J = diffusion-limited cell current, A,

2
A = area of mercury surface, mm~,.

*]. Hérranz, S. R. Bloxom, J. B. Keeler, and S. R. Roth,

178

Cs = concentration of quinone in the aqueous
phase, kg-mole/mm’,

F = Faraday’s constant, coulombs/kg-equiv,

z number of electrons

equiv/ kg-mole.

discharged, kg-

The defining equation for the mass transfer coeffi-
cient is

J=kCs— C), (3)
where
J = mass flux, kg-mole'mm,
C: = concentration of quinone at the mercury

surface, kg-mole/ mm’.

When the cell is operating under conditions such that

“the mercury surface is polarized,

C=0. (4)

A diagram of the experimental apparatusis shown
in Fig. 9.6. The contactor vessels were Plexiglas
boxes of square cross-section twice as deep as they
were wide; the agitators were flat-bladed turbines
having four blades, with one turbine centered in each
phase. The anode for each cell was made from 1.6-
mm-thick brass sheet, which was formed to fit the
inner perimeter of the cell and was suspended in the
aqueous phase. The anodes were plated with gold or
silver to resist chemical attack by the aqueous
solution.

The potential of the mercury surface was con-
trolled versus a saturated calomel electrode (SCE),
which was suspended in the aqueous phase while
current was passed between the anode and the
cathode. A potentiostat was used which was capable
of automatically varying the impressed voltage
between limits of +2 V(SCE)and —2 V (SCE) at rates
up to I V/min. The current through the cell was
plotted vs the mercury surface potential on a
Hewlett-Packard X-Y plotter.

The composition of the aqueous phase for all runs
was 001-0.05 M hydroquinone, 0.0002-0.001 M
quinone, in a 0.2 M phosphate buffer solution having
a pH of 7.0.

The cell was filled with an appropriate volume of
each phase, and a turbine was positioned at the
midpoint of each phase. The agitator drive was then
started and adjusted to the desired speed. The
variable voltage was.increased from 0 to +1.0 V vs
SCE at a rate of 0.5 V/min, and the cell current and
179

ORNL DOwWG 76-357

<
AGITATOR
CONTACTOR SHAFT
VESSEL \
. —
b, SATURATED
> L CALOMEL
] - REFERENCE
. ELECTRODE
ANODE Ry
— i
] / )
> \ |,
AQUEQUS A
SOLUTION ——{«/|
5 AMMETER \DVOLTMETER
Al
7
- VARIABLE
1 VOLTAGE -
MERCURY , SUPPLY
CATHODE — [ I
Ps
‘/4 -
i s

\— FLAT-BLADED TURBINES
{4 BLADES)

Fig. 9.6. Diagram of equipment used for measuring mass-transfer coefficients in the stirred water-mercury contactor.

voltage were recorded automatically on the X-Y
recorder. The agitator speed was then adjusted to
another value, and the procedure was repeated untila
satisfactory range of agitator speeds had been
investigated.

A typical current-voltage recording is shown in
Fig. 9.7. In this figure, the cell current is recorded
over the range of 0—1 V vs SCE at different agitator
speeds. For each agitator speed, the cell current rises
at low voltages with increasing voltage and reaches a
constant value at higher voltages. This constant
current is the desired diffusion-limited current. The
diffusion current actually oscillates about anaverage
value because of turbulent fluctuations at the
interface. An average value of the diffusion current is
determined from the recordings, and this average
value is recorded and used for the determination of
the mass-transfer coefficient using Eq. (2).

Results of measurements in three contactor vessels

are presented in the following section.

9.2.2 Experimental Results

Water-side mass-transfer coefficients were
measured in three different Plexiglas cells, 102 cm X
102 mm, 203 X 203 mm, and 305 X 305 mm. In each
cell, two or three phase volumes were examined, the
volume of each phase being equal. For each phase
volume, several agitators with different diameters

were tested, one agitator being centered in each
phase. An finally, for each combination of cell size,
phase volume, and agitator diameter, several
determinations of limiting diffusion current were
made over a range of agitator speeds. Agitator speeds
were kept below that speed at which dispersal of the
aqueous phase into the mercury phase began.
Dispersal was detected by small droplets of water
leaving the mercury phase near the edge of the cell. A
summary of the experimental parameters is given in
Table 9.4.

All the mass-transfer data determined in these
experiments are reported elsewhere.'' Reported here
are representative results which indicate the effect of
the experimental parameters on the water-side mass-
transfer coefficient.

Figures 9.8 to 9.10 show the mass-transfer
coefficients as a function of agitator speed for one
value of phase volume in each of the three cells. The
plot for each agitator diameter appears to be
comprised of two regions. At low agitator speeds the
mass-transfer coefficient is proportional to the
agitator speed to the 0.7 power, and at high agitator
speeds the mass-transfer coefficient is proportional
to agitator speed to a power near 2. The change
between the two regions is fairly abrupt for small
agitator diameters, but as agitator diameter in-
creases, the change becomes more gradual. Data
from the water-mercury contactorare compared with
mA

CELL CURRENT,

180

ORNL DWG 76-358

40.
30— /0.8 rps W
o/ %
TR WU W ¥
20+ LN AW LYA L |=
U o J
10— —
/f
o | | | | 1 |
0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -l.4

MERCURY SURFACE POTENTIAL, V vs SCE

Fig. 9.7. Typical current-voltage recording taken to measure diffusion current in stirred water-mercury contactors.

Table 9.4. Summary of experimental parameters for measurement

of aqueous phase mass-transfer coefficients

. Agitator size Range of
(mcrj:luxstznem) (diameter X height) Phas(ec I\;]oal;xmes agitator speed
{mm X mm) (xps)
102x 102 64 X 19 700, 900 0.25-3.0
89 x 19 700, 900 0.25-3.0
203 x 203 89 x 19 3,000, 5,000, 7,000 0.25-2.33
140 x 19 3,000, 5,000, 7,000 0.25-2.33
190 X 19 3,000, 5,000, 7,000 0.25-2.33
305 X 305 89 x 19 9,000, 18,000 0.25-1.08
140 X 19 9,000, 18,000 0.25-1.08
190 x 19 9,000, 18,000 0.25-1.08
240 X 19 9,000, 18,000 0.25-1.08
280 x 19 9,000, 18,000 0.25--1.08

ORNL DWG 76-359

107!
| 0 T TT -~ 1T [T TTL]
) _
Q
[75]
~ — —_
£
£
P.. — —
2
e
O
L
" 2
e -2 L.
W oo ]
Q — |
)y - _
x _
L -
L _
Z |
< {02 x 102mm CELL, 700 c¢cm PHASE VOLUME ]
|.— I po—
. A 89mm DIAM x I9mm WIDE AGITATOR
@ B © 64mm DIAM x 19mm WIDE AGITATOR
g X 77mm DIAM AGITATOR (OLANDER & BENEDICT, 1962) |
ISOBUTANOL TRANSFERRING THROUGH H»0
| @ 77mm DIAM AGITATOR (OLANDER & BENEDICT, 1962) .
HpO TRANSFERRING THROUGH 30% TBP-HEXANE
B 77mm DIAM AGITATOR (OLANDER & BENEDICT, 1972)
H,O TRANSFERRING THROUGH ISOBUTANOL
0-3 RN I B N R
0.1 1.0 10.0

AGITATOR SPEED, rps

Fig. 9.8. Comparison of water-mercury data with aqueous-organic data.

data measured by Olander and Benedict'? using
aqueous-organic systems (Fig. 9.8). The water-
mercury data are bracketed by the aqueous-organic
data (Fig. 9.8). Olander and Benedict also report the
change in the slope of log(k) vs the logarithm of the
agitator speed.

The effect of agitator diameter on mass-transfer
coefficient at constant agitator speed is significant,
but the variation depends on the agitator speed (Figs.
9.8-9.10). The effect that is of most interest for design

" and evaluation of stirred nondispersing contactors is

the dependence of the maximum mass-transfer
coefficient achievable with a given agitator before
dispersion occurs. A relation has been shown
previously’ between the agitator diameter and the
agitator speed at which dispersion (of the water into
mercury) is first noticed. This relationship was also
shown to be essentially independent of the two fluids
being contacted ' since it held for water and mercury,
water and carbon tetrachloride, and water and
dibromoethane. This relation shows that the maxi-
mum agitator speed decreases with increasing
agitator diameter. Therefore, it is necessary to
determine the net effect of lower allowable agitator
speeds and increasing mass-transfer coefficients as
agitator diameter increases.

Although in all runs in the water-mercury
contactor the agitator speed was not increased to the
point where dispersion was seen, the maximum mass-
transfer coefficient measured for each agitator
diameter was plotted vs the agitator diameter on log-
log coordinates (Fig. 9.11). From this information, it
is seen that the maximum mass-transfer coefficient

182

generally increases with increasing agitator diameter.
If it is assumed that this maximum value is
proportional to the agitator diameter raised to some
power, the exponent in that relation probably falls in
the range 0.8-1.6 within the range of agitator
diameters investigated (64-270 mm).

Figure 9.12 shows all the data for the 89-mm-diam
agitator on a single plot. This figure shows that the
variables other than agitator speed and agitator
diameter effect the value of the mass-transfer
coefficient fairly significantly. Careful examination
of Fig. 9.12 shows that for a given cell size the mass-
transfer coefficient varies inversely with the phase

ORNL DWG 76-360

-1
0 T T T 1]
— —
) [ ]
@
(/2]
~ - —
E
E
o B a
2
\u
=
T8
w
o -2 —
&) 10 | —
o [ —
(17} - —
(V'
w | . —
2
< —- —
[
[ | —
v
w)
q | _
=
203 x 203 mm CELL, 5000 cm> PHASE VOLUME _
— © 89mm DIAM x I19mm WIDE AGITATOR —
A 140mm DIAM x I19mm WIDE AGITATOR
@ 190mm DIAM x 19mm WIDE AGITATOR
103 IR N N 0 1 IS N S B A I
. Oll l.o IOIO

AGITATOR SPEED, rps
_Fig. 9.9. Effect of agitator speed and diameter on mass-transfer coefficient in the 203-mm cell with 5000-cm’ phase volumes.
183

ORNL DWG 76 - 36l

o T T T T T T T T T TR

| 305 x 305mm CELL, 9000 c¢cm® PHASE VOLUME ]
— ® 89mm DIAM x |9mm WIDE AGITATOR .

o | 4 140mm DIAM x 1I9mm WIDE AGITATOR ]

® # 190mm DIAM x 19mm WIDE AGITATOR

= | ¢ 240mm DIAM x I9mm WIDE AGITATOR |

€ B 280mm DIAM x 19mm WIDE AGITATOR

'_— O —

<

=

O

L

L

S o

144 — —_

u'l —

W ]

n - ]

Z

< .

g —

-

o — —

@ |

<

p- ~— —

io—a I I | | | I I [
0.1 1.0 _ 10.0

AGITATOR SPEED, rps
Fig. 9.10. Effect of agitator speed and diameter on mass-transfer coefficient in the 305-mm cell with 9000-cm® phase volume.

volume used. This may reflect the effect of distance
between the agitator and the interface since in all
cases the agitators were positioned at the midpoint of
each phase. Systematic variations have not yet been
extracted from the data, but these data serve to show
that details in the cell design can affect the value of
the mass-transfer coefficient by 200% in the case of
these experiments.

The relative effect on the water-side mass-transfer
coefficient of stirring in each phase is shown in Fig.
9.13. One test was performed in the 305-mm cell, with

phase volumes of 18,000 cm’ and one agitator -

diameter of 190 mm. First the mass-transfer
coefficient was measured as a function of agitator
speed, with an agitator in each phase. The test was
then repeated twice, first with an agitator in the
mercury phase only and then with the agitator in the
water phase only. The results indicate that at low
agitator speeds, agitation in the water affects the
water-side mass-transfer predominantly; whereas, as
agitator speed increases, agitation in the mercury
affects the water-side mass-transfer coefficient more
strongly than agitation in the water phase. Moreover,
the mass-transfer coefficient resulting from agitation
ORNL DWG 76-363

T T TTT] T T ]

SLOPE = .6

mm/sec

SLOPE= 0.8

3
TTTTTT]

MAXIMUM MASS TRANSFER COEFFICIENT,
000

[ 1]

1000

o L L1111 | 1]
30 50 100 500
AGITATOR DIAMETER, mm

Fig. 9.11. Effect of agitator diameter on maximum mass-
transfer coefficient achievable in the water-mercury contactors.

184

in both phases is not a simple additive function of the
mass-transfer coefficients resulting from agitation in
one phase only.

These measurements have provided a great deal of
data covering a wide range of physical parameters
which affect mass transfer in stirred cells. Using the
electrochemical technique, the data have been
measured with somewhat more precision than data
based upon transient experiments which have been
used exclusively in the literature. Because of this,
these data will be useful in developing correlations
that can be used to estimate mass-transfer rates in
large-scale nondispersing stirred contactors required
in the MSBR reductive extraction processes.

9.3 CONTINUOUS FLUORINATOR
DEVELOPMENT

R. B. Lindauer

Autoresistance heating test AHT-4 involves the
circulation of molten salt through a simulated
fluorinator (test vessel). A frozen-salt film is formed
by cooling the test vessel wall,-and the salt is kept

ORNL DWG 76-362

10”!
— R R EEREE
124
@ b —
L
E
£
P. [~ —_—
<
w
o
i
w
8 1072 — -
(&) F— —
< L ]
w — —
w
(2] — —
s . e
F_‘ 89mm DIAM x 19mm WIDE AGITATOR
o B CELL PHASE s |
by SIZE, mm VOLUME, cm
g — o 102 700 ]
© 102 900
N A 203 3000 ]
a 203 5000
v 203 7000
® 305 9000
|O-3 | I l I l | ‘Bll 365 1 L IIBOLOOI ‘ l
0.1 1.0 10

AGITATOR SPEED,

rps

Fig. 9.12. Effect of cell size, phase volume, and agitator speed on mass-transfer coefficient produced by the 89-mm-diam

agitator.
185

ORNL DWG T6-159 RI

0.3
Q (oR | — Oll
3 L A ,
E - BOTH PHASES
E | AGITATED o MERCURY ONLY
- AGITATED
= |
=
w
5 =
.y
w — " D’
o
WATER "ONLY
@ - i AGITATED
w
[V
w .
o
[ 4
'—
0.01—
v — A/
o |
< L
E —
) Lol L1 Lo
0.003 =55 0.5 1.0 " 2.0 5.0

AGITATOR SPEED, rps

Fig. 9.13. Relative effect on water-side mass-transfer coeffi-
cient of stirring in each phase in the 305-mm cell using
190-mm-diam agitators and 18,000-cm® phase volumes.

molten by autoresistance heating. The design and
construction have been described pre_:viously.”’16
Five test runs have been made, but a redesigned
electrode seems to be necessary before successful
operation can be achieved. _ '
Equipment for demonstratlng that a frozen-salt
film affords protection against fluorine corrosion has
been previously described.'™® Eight cooling tests
without fluorine have been made in this equipment.

9:3.1 Autoresistance Heating Test AHT-4

Operating experience from the first four runs of
autoresistance heating test AHT-4 indicated two
‘major problems in the forming and maintaining of a
satisfactory frozen-salt film in AHT-4 (Fig. 9.14).
The first problem isthe tendency for the vertical test
section to cool faster and eventually restrict the salt
flow before the electrode side arm has an insulating
salt film, which will allow autoresistance heating to
be started. The second problem is in maintaining a
molten-salt flow through the unheated tip of the
electrode. Before modifying the electrode, a run
(AHT 4-5) was made in which cooling of the vertical

test section was controlled to keep the wall tempera-
ture above that of the electrode side arm. After 8% hr
of cooling, the test vessel liquid level instrument

ORNL DWG 75-498%

GAS—-LIQUID T

SEPARATOR 2

AA

— OFF-GAS@—T‘

FREEZE
VALVE

HEAT
FLOWMETER

ARGON

AUTORESISTANCE
"HEATING POWER
SUPPLY

TEST
VESSEL

Fig. 9.14. Flowsheet for autoresistance heating test AHT4.
began to fluctuate, indicating incipient plugging. At
this time, most of the cooled section of the test vessel
was near a uniform temperature (648 K), and
autoresistance heating was started in an attempt to
maintain a molten core. This attempt was not
successful, and electrode plugging was indicated by a
sharp rise in the gas-liquid separator liquid levels.
The gas lift was stopped when generating 490 W of
autoresistance heating did not clear the plug but did
result in a decrease in resistance, indicating electrical
shorting near the electrode. The final resistance was
0.10 2 and the average cooled wall temperature was
632 K.

From the results of run AHT 4-5, plugging of the
unheated end of the electrode appears to be the most
immediate problem. A replacement electrode has
been designed which incorporates heating to thé tip
of the electrode. The electrode diameter has been
decreased by using '/,-in. rather than ¥s-in. pipe, and
the heater jacket diameter has been decreased by
using 1',-in. rather than 2-in. pipe. Using these
smaller diameters will increase the clearance between
the jacket and cooled wall from 1.8 in. (46 mm)to 2.1

186

in. (53 mm) to reduce the heating effect on the frozen-
salt film. The completely jacketed electrode will
permit installation of thermocouples at the tip of the
electrode to measure the temperature of the salt
entering the test vessel.

9.3.2 Frozen-Salt Corrosion Protection
Demonstration (FSCPD)

The equipment flowsheet for this demonstration is
shown in Fig. 9.15, and details of the test vessel with
the first cooled tube are shown in Fig. 9.16. Six
cooling test runs defined the conditions of tempera-
ture and cooling air flow necessary for the formation
of an adequate frozen-salt film (Table 9.5). Durmg
these tests a low flow of argon instead of fluorine was
used through the center of the tube. In the last four
runs, a small argon flow was maintained through the
transfer line from the delivery vessel to ensure an
open line for rapid salt transfer from the test vessel at
the end of the test. Transfer of saltat temperatures as
close to the liquidus temperature (778 K) as is
necessary in these tests is usually difficult. The salt

ORNL DWG 75-8948

FCV-3
F, T 100 cc/MIN HASTINGS MASS FLOWMETER
FI
5 PSIG !
| V-9 N S
|
| VALVE CLOSES ON 1 CELL 38
" HIGH Ar BACK | FIRST FLOOR ] g;g
w! ) _ PRESSURE _ D | ouTene
T 0-100 ’ ! BLDG)
@ IN. WATER l /@
W, @ A | | 1 V-5
o L fi
3 - 0-30 ! 1 -
o<! psic @ , V3 el V-4
z! e ‘
w&i I-CFM i
i3! @ 1 X O
LI, .
, 0-15 | ;
Lo—uyd PSIG : v-3
|
(s) v-6X V-7 /®
ARGON ! $ : vd @
2o rse P g /@
AIR ] ' -
I ' A i
l )
1 J TC

DELIVERY

|
| VESSEL
|
I

TEST

ACTIVATED ALUMINA
VESSEL TRAP

Fig. 9.15. Frozen-salt corrosion prgtection demonstration flowsheet.
187

ORNL DWG. 75-89493 R!

FLUORINE IN 1

— AIR OUT
F: - AIR IN
h .

—— ARGON PURGE

v

[ —& FLUORINE OUT

SALT LEVEL——»

Fig. 9.16. Frozen-salt corrosion protection demonstration test vessel showing tube No. 1.
188

Table 9.5. Data summary for FSCPD cooling tests — tube No. 1

Vessel
Cooling air Argon sparge Temperatures (°C) wall
Test Back lov , (cm”/min) Cooled Test cooling Remarks
No. pIessure F owrat.e Transfer ~ Fluorine tube vessel AT rate
(psig) (iters/min) line inlet minimum wall (°/min)
1 2.5 68 17 432 556 124 0.23 Salt transferred out —
very thin film
2 34 79 a 420 552 132 0.40 Salt not transferred —
_ line plugged
3 >4 >80 160 aQ 410 575 165 1.00 Salt transferred out —
very thin film
4 8 129 80 . 60 . 400 540 140 0.60 Inlet plugged
) 7.6 126 100 84 431 543 122 0.53 Inlet plugged
5.2 101 80 >100 417 528 111 0.27 Inlet plugged
4.4 94 150 90 405 512 107 0.12 Salt transferred out —
unable to remove tube
6 6.8 118 100 30 -420 529 109 0.27 Inlet plugged
4.0 89 100 41 446 522 76 0 Inlet plugged
44 - 9% 100 30 439 523 75 0.13 Inlet plugged
4.0 " 89 120 75° 443, 521 78 0 Inlet plugged

9Fluorine inlet plugged.

was transferred from the test vessel after three of the
runs. In one of these cases, the tube could not be

removed for examination, probably because of an -

unusually large salt ring formed above the salt level
from too high a sparge rate. The annular clearance
around the outer tube is 0.3 in. (76 mm). In the other
two cases the salt film was too thin to measure and
did not cover the metal at several points. The main
problem with' this tube design was plugging of the
“fluorine inlet tube at temperatures as high as 719 K.
Since this temperature is well above the solidus
temperature (623 K) of the salt, formation of a
suitable film was not possible.

Tests using the No. 1 cooled tube were ended when
a leak developed in the air cooling annulus of the
tube. The tube was cut off 5 in. (127 mm) from the
bottom, and this section was examined in the
Metallography Laboratory. Figure 9.17 shows the
area where the failure occurred, the grain boundary
attack, and the nickel oxide that was formed. A
horizontal crack developed in the Y3;-in.(0.8-mm)-
thick nickel inner [Y-in. (9.5 mm) OD] tube. The
crack was about Y4 in. (6 mm) long (~20% of the
circumference). The crack was not near a weld but
occurred in metal that had been machined froma 1-
in. nickel rod. Figure 9.17 also shows that the
thickness of the tube was reduced to less than one-
half of the original thickness.

- The second cooled tube was designed to be

sufficiently small to permit the use of argon instead of

air coolant to avoid nickel oxidation. Also, since
introduction of fluorine through the center of the
cooled tube is not required for this demonstration,

“the test vessel dip tubes were modified as shown in
Fig. 9.18. All three tubes are of }5-in.(9.5-mm)-OD

nickel tubing, with the argon coolant inlet tube of Y16-

- in.(4.8-mm)-OD nickel. This cooled tube has only

13% of the cooled area of the first tube and will
require considerably less coolant flow. Air coolant
flow with the first tube had to be restricted to 90
liters/min (1.5 X 107> m*/s) to maintain the bulk of

‘the molten salt above 793 K using the maximum
" power on-the vessel heaters. With the second tube an

argon flow of 35 liters/min (5.8 X 10 m’/s)
(proportionally three times greater) could be used
while maintaining a steady bulk salt temperature.
Two test runs were made with the second cooled
tube. In the first test the tube was cooled by anargon
flow of 35 liters/ min (5.8 X 10~ m’/s) for 3.8 hr. Salt
was then transferred from the test vessel, with the

~ bulk of the salt at 803 K. Cooling air was maintained

on the tube until the test vessel wall temperature was
reduced to 623 K. In the second test a higher argon
flow, 54 liters/ min (9.0 X 10™ m*/s), was used for 5
hr, and the salt was transferred, with the bulk salt

~temperature at 823 K and falling. In neither case was

100 200
1 i

MICRONS
100X—+

600 700
1 1

T
0005 0.010

INCHES 0.020 0.025

J
™

Fig. 9.17. Air-cooled dip tube (tube No. 1) in frozen-salt corrosion protection demonstration.

a satisfactory salt film formed. In the first test,a s~
in.(0.8-mm)-thick film covered the bottom '/ in. (13
mm) of the tube. In the second test, only the tip of the
tube was covered with a salt film. The bulk salt
temperature was probably too high in the second test.

9.4 FUEL RECONSTITUTION
ENGINEERING DEVELOPMENT

R. M. Counce

The reference flowsheet for processing the fuel salt
from a molten-salt breeder reactor (MSBR) is based
upon removal of uranium by fluorination to UFs as
the first processing step.'” The uranium removed in
this step must subsequently be recombined with the
fuel-carrier salt before its return to the reactor. The
method for recombining the uranium with the fuel-
carrier salt (reconstituting the fuel salt) consists of
absorbing gaseous UF; into a recycled fuel salt
stream containing dissolved UF, according to the
reaction

UFe(g + UFy¢ = 2UFsq . )

The resultant UFs would be reduced to UF; with
hydrogen in a separate vessel according to the
reaction:

UFs + ' Hap = UFyq + HF . (©)

Engineering studies of the fuel reconstitution step
are being started to provide the technology necessary
for the design of larger equipment for recombining
UF; generated in fluorinators in the processing plant,
with the processed fuel-carrier salt returning to the
reactor.

During this report period a preliminary hydrody-
namics test of the experimental equipment, in which
salt flow through the system was maintained under
simulated experimental conditions, was successfully
completed. A calibration of the UF; metering system
was completed; a gas density cell used for measuring
concentrations of UF; in argon was calibrated; and
apparatus for producing known concentrations of
HF in hydrogen was developed and was used to
calibrate the gas density cell used for measuring
concentrations of HF in hydrogen. These operations
are discussed in the remainder of the report.

190
ORNL DWG 76-364
ARGON IN ———»

l:—-b ARGON OUT

T
ARGON PURGE —~————®p— |
0 URGE + FLUORINE OUT

FLUORINE IN
7 | l—’ VENT

|

SALT LEVEL—» -~

N——UNCOOLED TUBE

\— COOLED TUBE

Fig. 9.18. Frozen-salt corrosion protection demonstration test vessel showing tube No. 2.

9.4.1 Hydrodynamic Operation

The hydrodynamic operational procedure for the
experiment has been discussed previously.”® During
this report period, five hydrodynamic experiments
were conducted: HR-1, HR-2, HR-3, HR4,and HR-
5. The fourth and fifth experiments, HR<4 and HR-5,
resulted in salt being metered from the feed tank
through the UFs absorption vessel and hydrogen
reduction column to the receiver (Fig. 9.19) at a
constant flow rate of approximately 100 c¢m’/min
(1.67 X 10°° m?/s).

In HR-1, the UFs absorption vessel and -the
hydrogen reduction column were filled. The siphon
was primed in HR-2, and salt flow was started from
the feed tank. Run HR-2 ended abruptly when
plugging occurred in the effluent salt transfer line

191

from the hydrogen reduction column. A third run,
HR-3, did not get under way because of a plug in the
transfer line from the feed tank to the absorption
vessel.

In run HR+4, the UFs absorption vessel and the
hydrogen reduction column were filled in prepara-
tion for.run HR-5. In run HR-5, a salt flow of 100
cm’/min (1.67 X 107 m’/s) was started and was
maintained for 1 hr. The gas streams of UFs and
hydrogen were simulated by two argon flows. The
two argon flow rates of 180 std cm’/min 3X107°
m’/s) simulate the stoichiometric gas flow rates of
UFs and hydrogen necessary to increase the UF,
concentration in the salt from 0.15 mole % in the feed
stream of the UFs absorption vessel to 0.3 mole % in
the ‘effluent stream from the hydrogen reduction
column.

ORNL DWG 74-11666RY

-
r-—'»——fl E I
EBe=—=
UFg HF UFg
ANALYSIS | {ANALYSIS HF TRAP
TRAP
el ,
l x | g BUILDING
OFF-GAS
UFg =l —®= SYSTEM
bé l EFFLUENT
STREAM
Y E = R SAMPLER
FREEZE
VALVES
Ar \
l 1
! ABSORPTION :
- 'VESSEL !
| 1
| ]
| :
]
! ' |
FEED TANK RECEIVER
Ha
\b,

'

M, REDUCTION
COLUMN

Fig. 9.19. Flow diagram of equipment used in first fuel reconstitution engineering experiment.
'y

9.4.2 Calibration of the UF,
Metering System

The UFs supply system®' (Fig. 9.20) provides UFs
at known flow rates to the UF¢ absorption vessel
during the operation of the experiment and to the
UFs-Ar gas density cell for calibration purposes. The

supply system containsa model ALF-500W Hastings

mass flowmeter which indicates and controls the UFs
flow rate. Because the UFs flow rate is an important
experimental parameter, this flowmeter must be
calibrated before the experiment begins.

The flowmeter was calibrated as follows. A steady
flow of UFs was established to the main off-gas trap.
The UF; stream was then diverted through the
calibration trap, a packed bed of NaF pellets
maintained at 373 K. The flow of UFs through the
calibration trap was maintained long enough to
produce a loading of 0.2 kg UFs/kg NaF. The UFs
stream was then diverted back to the main UFs trap,
and the flow of UF, stopped. The system was then
completely purged with argon. The calibration trap
was then removed and weighed again. The weight of
UFs loaded into the trap was determined by
subtracting the original weight from the final weight.
The flow rate of UF¢ was determined then by dividing
the weight of UF¢ by the length of time UF; flowed
into the calibration trap. This process was repeated
for several flow rates of UFs up to 560 std cm’/ min

ORNL DWG 76-365

FLOW
RECORDER - MASS
"CONTROLLER FLOWMETER
: |
1 |
! I
I/P |
CONVERTER |
|
|
l P FLOW
I SENSING
ELEMENT
—D<
UF
SAMPLE UF
TRAP TRAP
—
SUPPLY ’

Fig. 9.20. Diagram of UF4 supply system.

192

(9.33 X 106 m3/s). The results agreed very closely
with the published calibration curves provided by the
manufacturer. '

9.4.3 Calibration of Ar-UFs
Gas Density Detector

Calibration of the gas density cell that is used in
determining the concentration of UFs in the argon
stream leaving the UF; absorption vessel requires
UFs-Ar mixtures of known concentrations. These
mixtures were provided by mixing separate UFs and
argon streams of known flow rates and feeding the
mixture to the gas density cell used for UFs-Ar
analysis. This gas density cell”” was calibrated for
concentrations of 0—50 mole % UF¢ in argon (Fig.
9.21).

ORNL DWG 76-367 Ri

90
80 —
70 |—
>
E
w 60 |—
w
2
O
[/
w
W 50 —
-4
-t
I ]
O 40—
>
=
w
&
s 30—
[/p]
<{ GOW MAC GAS DENSITY CELL
o MODEL 11-373
20 — —
DETECTOR CURRENT — 130mA
DETECTOR TEMP. —— i00°C
REFERENCE GAS ——— NITROGEN
10 REFERENCE GAS 482 x1076 |
FLOW RATE std m3/sec
SAMPLE GAS 5.83 x 1Q"7
FLOW RATE std - m¥sec
o l I I

0 10 20 30 40 50"
UFg IN ARGON, % |

Fig. 9.21. Calibration curve of Gow-Mac gas density cell in
fuel reconstitution engineering equipment for UF¢-Ar analysis.
9.4.4 Calibration of the HF-H:
Gas Density Cell

Calibration of the gas density cell that is used in
determining HF concentration in hydrogen from the
hydrogen reduction column requires a supply of HF-
H, mixtures of known concentrations. Since HF is
not otherwise required in the experiment, a means is
used for producing HF-H; mixtures which does not
require an HF supply and metering system. Fixed
concentrations of HF in hydrogen can be generated
by passing hydrogen gas over NaF-HF in a manner
such that the HF partial pressure is increased to the
dissociation pressure of the complex by using the
equilibrium

NaF-HF — NaF + HF .

In this system the HF partial pressure is a unique
function of temperature as long as both NaF-HF and
NaF are present. Consequently the concentration of
HF in the gas stream in equilibrium with the
NaF-HF-NaF mixture can be set by controlling the
temperature of the system.

The vessel for producing hydrogen saturated with
HF (Fig. 9.22) is 1'/, in. (3.81 X 107 m) sched 40
Monel pipe 0.97 m long, mounted vertically with
standard 150-Ib Monel flanges at either end. This
vessel was filled with a commercial grade of NaF-HF
in the form of '/s-in. (3 mm) right circular cylindrical
pellets.

Hydrogen was passed over the NaF-HF bed in this
vessel at various temperatures at a total pressure of
1 atm, and the concentration of HF in the resulting
mixtures was compared with values calculated from
published values of the NaF-HF dissociation pres-
sure.

The equilibrium concentrations of gaseous HF in
hydrogen over NaF-HF obtained from these experi-
ments were compared to those of Davis®’ and
Froning® for concentrations in the range of 1-50
mole %. In general, concentrations in this range
occurred at temperatures approximately 50° C lower

(7).

193

ORNL DWG 76-366

e e e e

MA
FLOWMETER

—
| —

1

THERMOCOUPLES

\
\

CAL ROD
HEATERS ———al |-

\
)

NaF -HF -
PELLETS —— 1L

\
\
\

lJ \O X} Xd

.Fig. 9.22. Diagram of apparatus for producing mixtures of
HF in hydrogen.

than temperatures predicted from the work of
Froning®* and approximately 75°C lower than
temperatures predicted from the work of Davis.”’ In
Fig. 9.23, the experimental results are definitely
closer to the literature values of HF over NaF-HF
than HF over other known NaF-HF complexes.”
The results of the HF saturator experiment have
shown reproducibility, although the erroris such that
the flow rates in and out of the saturator are used as
the basis in calculating gas composition leaving the
saturator.

Using the HF saturator to generate HF concentra-
tions in hydrogen, the analytical system for H,-HF
analysis was calibrated for concentrations of 0—-50%
HF in hydrogen. This gas analysis system has been
described previously.”” This calibration curve and
pertinent operating conditions are shown in Fig.
9.24. Also shown for comparison in Fig. 9.24 is a
calibration of the system for 0—50% nitrogen in
hydrogen under the same operating conditions.
194

ORNL DWG 76-368

100
] T T T T T ]
— NaF-HF —
(FRONING, etal)
S . NaF-SHF
@ {DAVIS)
©
E b
o NaF-3 3 HF
x (DAVIS)
- NaF-3HF
w — (DAVIS)
a
a
_J
<
5 NaF-2HF
o (DAVIS)
Z 10—
<{ -
Q —
- -
q —
8
et —
- !
o
z NaF-HF
u -

S (DAVIS)
-2 )
2 |
Q
w
I

L 1 |

1.0

103
T(K)

Fig. 9.23. Comparison of equilibrium results of HF saturator with literature values for equilibrium of HF over NaF-HF and
NaF-2NaF.
195

ORNL DWG 76-369RI

40
: ]
36— —
W
w F— —
=
o
a — ]
(7p]
78}
w L
|
a 32—
lJ
o ——
>_
L |
& .
Z GOW MAC GAS DENSITY CELL
Lé—' —— MODEL 11-373
¢»» 28 |— DETECTOR CURRENT— 60mA
=1 DETECTOR TEMP. —— (00°C
© | REFERENCE GAS —— ARGON |
REFERENCE GAS —— 4.00 x 10°©
FLOW RATE std m3/sec
SAMPLE GAS .60 x 10°6 T
FLOW RATE std m3/sec
24 | | | | | |
100 90 80 70 60 50 40

HYDROGEN CONCENTRATION, %

Fig. 9.24. Calibration curve of Gow-Mac gas density cell in fuel reconstitution engineering equipment for H, -HF analysis and

H2 -N 2 analysis.

REFERENCES

I. Chem. Technol. Div. Annu. Progr. Rep. Aug. 31, 1975,
ORNL-5078, pp. 14247

2. Chem. Technol. Div. Annu. Progr. Rep. Mar, 31, 1973,
ORNL-4883, pp. 23-25.

3. L. E. McNeese, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 10, ORNL/TM-
3352, pp. 53-57 (December 1972).

4. Chem. Technol. Div. Annu. Progr. Rep. Mar. 31, 1973,
ORNL-4883, p. 25.

5. L. M. Ferris et al., *“Distribution of Lanthanide and Actinide
Elements Between Liquid Bismuth and Molten LiCI-LiF and
LiBr-LiF Solutions,” J. Inorg. Nucl. Chem. 34, 313-20 (1972).

6. H. C. Savage in Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 23, ORNL/TM-
5252 (in preparation).

7. J. A. Klein et al., MSR Program Semiannu, Progr. Rep.
Aug. 31, 1974, ORNL-5011, pp. 114-18.

8. C. H. Brown, Jr., MSR Program Semiannu. Progr. Rep.
Feb. 28, 1975, ORNL-5047, pp. 152-57.

9. C. H. Brown, Jr., et al., Measurement of Mass-Transfer
Coefficients in a Mechanically Agitated, Nondispersing Contactor
Operating with a Molten Mixture of LiF-Be F»-THF. and Molten
Bismuth, ORNL-5143 (in preparation).

10. C. H. Brown, Jr., MSR Program Semiannu. Progr. Rep.
Feb. 28, 1975, ORNL-5047, pp. 157-62.

11. C. H. Brown, Jr., Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 24, ORNL/TM-
5339 (in preparation).

12. D. R. Olander and M. Benedict, Nucl. Sci. Engr. 14, pp.
28794 (1962).

13. H. O. Weeren and L. E. McNeese, Engineering Develop-
ment Studies for Molten-Salt Breeder Reactor Fuel Processing
No. 10, ORNL/TM-3352, p. 55 (December 1972).

14. J. A. Klein and C. H. Brown, Jr., unpublished data.

15. R, B. Lindauer, MSR Program Semiann. Progr. Rep. Feb.
28, 1975, ORNL-5047, pp. 163-65.
16. R. B. Lindauer, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 22, ORNL/TM-
4863 (in preparation).

17. R. B. Lindaver, MSR Program Semiann. Progr. Rep. Aug.
31, 1975, ORNL-5078, pp. 152-55.

18. R. B. Lindauer, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 23, ORNL/TM-
5252 (in preparation).

19. Chem. Technol. Div. Annu. Progr. Rep. Mar. 31, 1972,
ORNL4794, p. 1. '

196

20. R. M. Counce, MSRProgram Semiannu. Progr. Rep. Aug.
31, 1974, ORNL-5011, p. 128.

21. R. M. Counce, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 23, ORNL/TM-
5252 (in preparation).

22. R. M. Counce, Engineering Development Studies for
Molten-Salt Breeder Reactor Processing No. 22, ORNL/TM-
4863 (in preparation).

23. Wallace Davis, Jr., KLI-2552, p. 3 (Sept. 21, 1953).

24. J. F.'Froning et al., Indus. Engr. Chem., 39, 275-78 (1947).
Part 5. Salt Production

10. PRODUCTION OF FLUORIDE SALT
MIXTURES FOR RESEARCH
AND DEVELOPMENT

F. L. Daley R. W. Horton*

At the beginning of this report period, the repaired
meltdown vessel was returned to service for produc-
tion of MSBR fuel-carrier salt. A total of 450 kg of
fuel-carrier salt was produced in three 150-kg batches
in the new copper-lined treatment vessel, which also
has a protective copper sheet on the inside of the
vessel head. Completed analyses on salt samples from
the first two production batches (FS-108 and FS-109)
show that impurities were present at levels much
lower than normal. The oxygen content of the salt
was below 100 ppm. These results point out the
strong effect of purification vessel corrosion products
on the quality of salt that is produced.

*Consultant.

197

During this report period, the decision to termi-
nate work on the MSR Program by the end of FY
1976 was reached. The need for salt during the
balance of the fiscal year was determined, and this
need was supplied while preparations to place the salt
production facility in a safe standby condition were
proceeding. At the end of the production period, a
total of 1975 kg of salt (of various compositions) had
been produced since activation of the facility in early
1974. Shipments to the program totaled 865 kg, and
678 kg have been stored for possible use.

All production areas were decommissioned and
cleaned to remove potential health hazards. Unused
materials and equipment were either set aside for
burial or were stored; some major property items
were transferred to the Reactor Division. A report on
the final production and decommissioning activities
provides a reference for stored material and records

final comments on the performance of the salt

purification system.
199

MOLTEN-SALT REACTOR PROGRAM

FEBRUARY 1976

L. E. McNEESE, PROGRAM DIRECTOR

MSBR CHEMISTRY

L. M. FERRIS'
. D. KELMERS
MAYA
F.HITCH'
M. TOTH!
. 0. GILPATRICK!
.A.POSEY!
. R.BRONSTEIN
D. E. HEATHERLY!
S.N. RUSSELL3
D.Y.VALENTINE

Inrrer-»

oocoO0coOo0O000

SALT PRODUCTION

F.L. DALEY
W. J. BRYAN
A. L. JOHNSON

CONSULTANT
R. W. HORTON

cT
cT
cT

CT

MSBR DESIGN AND DEVELOPMENT MATERIALS MSBR PROCESSING DEVELOPMENT
J. R. ENGEL? R ‘H. E. McCOY?2 M&C J. R. HIGHTOWER, JR.2 CT
SYSTEMS AND ANALYSIS HASTELLOY N STUDIES CHEMICAL DEVELOPMENT
J.R. ENGEL? R H. E. McCOY? M&C A.D. KELMERS? C
E.J. ALLEN’ R S.L.BENNETT M. R. BENNETT C
H.T. KERR! R D.J. BRADLEY?® M&C
G.T.MAYS R D.N.BRASKI _ M&C ENGINEERING DEVELOPMENT
D. L. REED R J. BRYNESTAD M&C , 2
G.T. YAHR' R R. E. CLAUSING' M&C J.R. HIGHTOWER, JR. cr
R.S. CROUSE‘I M&C C.H. BROWN, JR, CT
SYSTEMS AND COMPONENTS J. R. DISTEFANO? M&C R. M-.ggg,”c's gT
DEVELOPMENT R. L. HEESTAND' M&C C.W. T
C. R HYMAN R R.B. LINDAUER CT
. GUYMON R A .
R.H.GU H. C. SAVAGE CT
1 J. R. KEISER M&C
W. R. HUNTLEY R , 1. BEAMS cT
3 B. C. LESLIE M&C
M. D. SILVERMAN R R. 0. PAYNE CcT
B. McNABB M&C e
R F. BENSON! R T. K. ROCHE1 M&C - b
E.L.BIDDLE R f‘; :’N"H'T;S1 M&C
D. J. FRAYSIER R -W. W00 M&C
H. E. ROBERTSON R
CHEMICAL PROCESSING MATERIALS
J. R. DiISTEFANO? M&C
TECHNICAL SUPPORT
K. W. BOLING! M&C
C. E. DUNN' M&C
AC  ANALYTICAL CHEMISTRY DIVISION W. H. FARMER' M&C
C  CHEMISTRY DIVISION J.C. FELTNER M&C
CT CHEMICAL TECHNOLOGY DIVISION J.L. GRIFFITH' M&C
M&C METALS AND CERAMICS DIVISION L. HEATHERLY" M&C
R REACTOR D!VISION J. W. HENDRICKS! M&C
UT UNIVERSITY OF TENNESSEE T.J. HENSON' M&C
'PART-TIME ON MSRP J. D. HUDSON' M&C
ZDUAL CAPACITY E. J. LAWRENCE M&C
3SCUU STUDENT PARTICIPANT E.H. LEE M&C
SORAU STUDENT G.A.POTTER M&C
L. G. RARDON! M&C
L. J. SHERSKY' R
W. B. STINES M&C
C. K. THOMAS' M&C
C. A. WALLACE! R

J. J. WOODHOUSE'

ANALYTICAL CHEMISTRY

RESEARCH AND DEVELOPMENT

A.S.MEYER?
R.F. APPLE'
J. M. DALE!
D. L. MANNING

AC
AC
AC

AC

ANALYSES

L.T.CORBIN!
J. H. COOPER"
W. R. LAING'
J. A.CARTER

CONSULTANT
G. MAMANTOV

AC
AC

AC

AC

uT

1-10.
. E. J. Allen

51-57.

61.

MSRP Director’s Office

C Beeson

Braskl

Clark

G

. E. Ferguson
.M. Ferris

. O. Gilpatrick

201

INTERNAL DISTRIBUTION

62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.

77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.

93.
94.

95.
96-102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112..

113.

ORNL-5132
Dist. Category UC-76

O.

A. D

HT
W.R.Laing

J. M. Leitnaker
R.B. Lindauer
M. 1. Lundin
H. G. MacPherson
R. E. MacPherson
G. Mamantov
D.L.Manning
W. R. Martin
C. L. Matthews
L. Maya

G.T. Mays
H.E. McCoy

H. F. McDuffie
C. J. McHargue
H. A. McLain
B. McNabb

R. L. Moore
F. H. Neill

P. Patriarca

T. W. Pickel
C. B. Pollock
F. A. Posey

H. Postma
114-

152.

153.
154-155.
156-260.

115.
116. .
“117.
118.
119.
120.
121.
122,
123.
124.
125.
126.
127.
128.
129.
130.
131.

202

A.
F.
G.
I Splewak

J. O. Stiegler
R.E. Thoma

A.J. Thompson
L. M. Toth

132.
133.

134.
135.
136.

137.

138.

139.

140.

141,

142,

143,

144.
145-146.
147,
148-150.
151.

EXTERNAL DISTRIBUTION

Oak Ridge, TN 37830
Director, Reactor Division, ERDA, Oak Ridge Operations Office, Post Office Box E, Oak Ridge, TN 37830
Director, Division of Nuclear Research and Applications, ERDA, Washington, DC 20545

For distribution as shown in TID-4500 under UC-76, Molten-Salt Reactor Technology category (25
copies—NTIS)

. B. Trauger

.Y. Valentine

.N. Washburn

.M. Weinberg
. Weir

. K. Wilkinson

W. R. Winsbro

J. W. Woods

R.G. Wymer

G.T. Yahr

J.P. Young

E. L. Youngblood

Central Research Library
Document Reference Section
Laboratory Records -
Laboratory Records. ORNL R.C.

Research and Technical Support Division, ERDA, Oak Ridge Operations Office, Post Office Box E,