ORNL-3913.txt

LOCKHEED MARTIN ENEI AESEARCH LIBHARIES

il

345 05159761

 
 
ORNL-3913
UC.4 — Chemistry
TID-4500 (47th ed.)

Contract No, W-7405-eng-26

REACTOR CHEMISTRY DIVISION ANNUAL PROGRESS REPORT
For Period Ending December 31, 1965

Director

W. R. Grimes

Associate Directors

E. G. Bohlmann
H. F. McDuffie
G. M. Watson

Senior Scientific Advisors

F. F. Blankenship
C. H. Secoy

MARCH 1966

OAK RIDGE NATIONAL LABORATORY
Dak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the
1. 5. ATOMIC ENERGY COMMISSION

LOGCKHEED MARTIN ENERGY RESEARCH LIBRAR

UAADIRALT

 

 

1l

4 445k 0515976 1

 

 

 

 

 

 

 

 

 

 

 

 

 

i
i

[
i
Contents

PART |, MOLTEN-SALT REACTORS

1. Phase Equilibrium ond Crystallographic Studies

LIQUID-LIQUID IMMISCIBILITY IN THE SYSTEM L:’F—Ber-ZrF4
H. A. Friedman and R. F. TROMG ..t e ee et bbb ey s b s aae e e e et nns 3
Composition~temperature limits of liquid-liquid immiscibility in the system Lil“~'BeF?—ZrF4 have been partially

established by application of high-temperature centrifugation.

S0DIUM FLUORIDE-SCANDIUM FLUORIDE PHASE EQUILIBRIA
R. H. Karraker and R. E. Thoma...........c.oceiccevinann e e ettt ee e ah et eees e aeeeeenaeeians ceiei i ianeeraaei e nean s e e 4
Results of a completed investigation of the high-temperature reactions of Na¥ and S(:F3 show that the system
is characterized by two eutectic and two peritectic reactions which occur in association with the intermediate
phases 31‘~Ta3['?~-Sr:F3 and NaF-SCFs.-

COMPOSITIONAL VARIABILITY IN SODIUM FLUCRIDE -LANTHANIDE TRIFLUORIDE COMPLEX
COMPOUNDS
R. E. Thoma, H. Insley, and G. M. Hebert i e e e er e e e e e e et e s e e et enemceen 6
Monotonic expansion of the compesition limits of the complex NaF»l’.,-nF3 crystalline phases has been ascribed
te a reduction in the polarizability of the lanthanide ions with increasing atomic number accompanied by a cor-

responding increase in free space within the crysial lattices,

PHASE EQUILIBRIUM STUDIES IN THE U02~Zr02 SYSTEM
K. A. Romberger, H. H. Stone, and C. F. Baes, Jr. i e 8
The U02-2r02 phase diagram has been revised in accord with the measured low solubilities of Z’rO2 in cubic
UO2 and of UO,2 in tetragonal ZrO:! and monoclinic ZrOZ. These solubilities are, respectively, 0.4, 1, and 0,15

mole % at the eutectic temperature of 1116°C.

THE CRYSTAL STRUCTURE OF LIUF5
G D BrifI O s et sarae s ar s e e b a e etaeeeMeeeseitieasereeiseesieaereteaeeseseemeeeaetieaeeaaneneiiaas 10
An x-ray structure analysis has ascertained the formula of this compound, previcusly designated 7LiF‘—6UF4.

In it are found UE‘8 polyhedra which are quite similar to those in crystalline UF4.

THE CRYSTAL STRUCTURE OF Li3A!F6
J- H. Burns and A. C. TenniSSBIl .o e e 12

- +
This compound crystallizes with A!.F‘S3 octahedra joined together by 1L.i ions.

REFINEMENT OF THE CRYSTAL STRUCTURE OF (NH jj MnF

The previously determined structure of (NH4}2MnF5 was refined by least squares.

HIGH-TEMPERATURE X-RAY STUDIES
G. D. Brunton, D. R. Sears, and J. Hl. BUINS i e 16
The furnace attachment for the x-ray diffractometer has heen used to study phase transformations in rare-earth
trifluorides, the U(')Q-'-Zr(i);Z system, and LisAIFG.

i1l
iv

CRYSTALLOGRAPHIC DATA ON NEW COMPOUNDS
J. H. Burns, D. R. Sears, and G. I BrURLON oot e et ettt ettt et et e e aae e 17
Unit-cell dimensions and space groups have been determined for Na3SCF6, BI-KLaF4, Rb}”aFé, Li4UF

LiU4F17, and NaBiF4.

THE CRYSTAL STRUCTURE OF I\e'chZréF_.’,r
J. H. Burns, R. D. Ellison, and H. AL L evy o e e e e e e e e it e 17
Six formula weights of NaF and six of Z’rF4 comprise a structure which has fluorine-bridged zirconium octahedra

containing a seventh fluorine atom within; a seventh sodium atom is also present in the structure and has 12

fluorine neighbors.

2. Chemical Studies of Molten Salts

OXIDE CHEMISTRY OF LfF«BeFZ-ZrF4 MELTS
C. F. Baes, Jr., and B. F. HitCh o e e e e e e e 20
Measurements of the solubility of ZrO2 in simulated MSRE fuel salt and flush salt mixtures are being continued

by means of an improved technique of oxide analysis.

THERMODYNAMICS OF MOLTEN L.iF-BeF2 SOLUTIONS
O R C T Vet £ PRSP 21
Enthalpies and free energies of formations, electrode potentials, and activity coefficients are estimated for

various solutes in 2Lil~'-13el-""_j3 from existing chemical and thermochemical data.

VAPOR PRESSURES OF FLUORIDE MELTS
S. Cantor, D. 8. Hsu, and W. T Ward e et et e e 24
Rodebush-Dixon and boiling-point techniques were used to measure vapor pressures of the LiF-BeF2 system,

the MSRE fuel solvent, and Ber. A transpiration apparatus was constructed and tested with pure LiF.

VISCOSITIES OF MOLTEN FLUORIDES
B Cantor and W. L W ard oo e e e e s e ey e et e e et i 27
Studies on the LiF-BeF2 system were extended. The compound NaBF4 was found to be fluid.

ESTIMATING DENSITIES OF MOLTEN FLUORIDE MIXTURES
Bt I O L st it et e e e e eee et et e eeets e ereeetaee et eeea e e ee et ee e ettt eee bt e en et e et ae e e e enianeras 27
Examination of newer data of molar volumes resulted in revised values for use in estimating densities of

molten fluorides.,

ESTIMATING SPECIFIC HEATS AND THERMAL CONDUCTIVITIES OF FUSED FLUORIDES
B A O i e it it ittt e e e e et e e e e e e e e et e e e et ee s e oo e e —eeeeee e e eea e eeeneeeneeetnnn et et et et a et a e e 29
By modifying the Dulong-Petit rule to equal 8 cal (OK)'"1 (gTam-atom)_l, recasonable agreement with experi-
mental data was obtained. A semitheoretical method used to estimate thermal conductivities agreed with experi-
mental results; the method depends on estimating sonic velocity, which itself can be calculated from thermo-

dynamic parameters.

SOLUBILITY OF DF AND HF IN LiFvBeF2 (66-34 MOLE %)
P. E. Field and J. H. SRaffer e e e e e 33
Solubilities of DF and HF were determined for the temperature interval 500—700°C to provide a basis for

estimating the solubility of tritium fluoride in the blanket of a thermonuclear breeder reactor.

3. Chemical Separatien and Irradiatien Behavior

IN-PILE MOLTEN-SALT IRRADIATION ASSEMBLY
H. C. Savage, M. J. Kelly, E. L.. Compere, J. M. Baker, and E. G. Bohlmann ..., 34

An in-pile molten-salt irradiation experiment is being constructed for operation in bearm hole HN-1 of the ORR.
EVAPORATIVE-DISTILLATION STUDIES ON MOLTEN-SALT FUEL COMPONENTS
M o BBLLY oottt et e e et e e e 35

Mass-~rate variations with temperature for distillation of "LiF and MSRE fuel solvent and the volatility of NdF |

relative to that of MSRE fuel solvent have been determined experimentally.

EFFECTIVE ACTIVITY COEFFICIENTS BY EVAPORATIVE DISTILLATION OF MOLTEN SALTS
M. J. Kelly
Activity coefficients of BeF2 and ZrF4 have been obtained from data on vacuum distillation of these com-

pounds from molten LAF solvent.

REMOVAL OF IODIDE FROM LiF-BeFZ MELTS
B. F. Freasier, C. F. Baes, Jr., and H. H. Btone .. e e e e, e 38
Tedide was readily removed from molten LiF-Bel", mixtures by HF-H sparging; this process should prove a

valuable means of removing the precursor of 135ye (1351) from molfen—fluondc reactor fuels,

REMOVAL OF RARE EARTHS FROM MOLTEN FLUORIDES BY EXTRACTION INTO MOLTEN METALS
J. H. Shaffer, F. A. Doss, W. K. R. Finnell, W. P. Teichert, and W. R. Grimes. . ..., 40
Rare-carth fluorides have been removed from solution in a fluoride mixture which simulates the barren fuel

solvent for a molten-salt breeder reactor by reduction and extraction into molten bismuth.

REMOVAL OF PROTACTINIUM FROM MOLTEN FLUORIDES BY OXIDE PRECIPITATION
J. H. Shaffer, F. A. Dossg, W. K. R. Finnell, W. P, Teichert, and W. R. Grimes ... ........cooiuiiiimmimiiineeaan 41
Previous studies were extended to demonstrate that protactinium could be precipitated as its oxide by the

addition of ZrO? to a fluoride solvent known to have Zr()2 as a stable oxide phase.

REMOVAL OF PROTACTINIUM FROM MOLTEN FLUORIDES BY REDUCTION PROCESSES
J. H. Shaffer, F. A. Doss, W. K. R. Finnell, W. P. Teichert, and W. R. Grimes
In the reference-design MSBR, protactinium can be removed from solution in the proposed fluoride blanket

mixture by reduction with thorium or with alloys of thorium in lead or bismuth.

SOLUBILITY OF THORIUM IN MOLTEN LLEAD
J- H. Shaffer, F. A. Doss, W. K. R. Finnell, and W. P, T eiChert et e, 43
The solubility of thorium in lead over the temperature interval 400-600°C was determined in order to support

studies of the extraction of protactinium from molten fluorides.

PROTACTINIUM STUDIES IN THE HiGH-ALPHA MOLTEN-SALT LABORATORY
e e BB O it iitie it et L L b s L bttt et e eot £k e e e f e e £at te ettt et e eun e esfn e et tn ettt et tneettn e stntntarnsrtnarrrn e bares 44
Seven glove boxes have been interconnected and equipped for studies of protactinium recovery from molten-

fluoride breeder-blanket mixtures.

SEGREGATION ON FREEZING LiCI-KC] EUTECTIC MELTS CONTAINING SOLUBLE SOLUTES
H. A. Friedman and F. F. Blankenship e e e et te et s e et et e s ettt e e e n e e e e e eanennns 45
The feasibility of purifying salt melts by freezing with rapid stirring to facilitate diffusion of rejected solute

from the freezing front was explored.

4. Direct Support for MSRE

PREPARATION AND LOADING OF MSRE FLUORIDES
J. H. Shaffer, W. K. R, Finnell, F. A. Doss, and W. P. Teichert ..o e 47
The production of all the component fluoride mixtures for the MSRE fuel was completed; these mixtures were

added to the MSRE as required during the precritical and critical testing phases of MSRE operation.
CHEMICAL BEHAVIOR OF FLUORIDES DURING MSRE OPERATION
Results of chemical tests with MSRE fuel and coolant salts, obtained during the precritical test, the zero-

power test, and the initial period of the full-power test, indicated that the reactor salts are of excellent purity

and that no appreciable corrosion of the interior of the reactor has occurred.
vi

MEASUREMENT OF DENSITIES OF MOLTEN SALTS
B. J. Sturm and R. E, TROMA i i e e e e e e 50
Density of the MSRE fuel salt was determined with molten mixtures using an electrical probe to measure volume
of the salt. The fuel-salt density (in g/cm”) is described by the expression d = 2.848 — 7.693 x 10~ " ¢ (°C).

PART 1l. AQUEOUS REACTORS

5. Corrasion and Chemica! Behavior in Reactor Environments

MECHANISM OF ANODIC FILM GROWTH ON ZIRCONIUM AT ELEVATED TEMPERATURES
Al L. Bacarella and Al L SUEEON et e st e ekt e a e e e et e r b 55
Considerations of experimental data for zirconium oxidation and of theory provided evidence that the pre-
exponential term in the anodic-film-growth equation can be interpreted in terms of paraineters of fundamental sig-

nificance.

MECHANISM OF RADIATION CORROSION OF ZIRCONIUM AND ZIRCALOY-2
R. J. Davis and G. H. Jems o e ettt e e et e et 57
The previously reported radiation effect on the postirradiation corrosion was confirmed, and exploratory work

toward new methods of evaluating film protective properties is in progress.

EFFECTS OF REACTOR OPERATION ON HFIR COOLANT

.................................................................................................................................................................. 58
Final evaluations were completed of the probable concentrations of excess oxidant needed to stabilize HNO3

in the HFIR coolant-moderator and of the expected steady-state concentration of the decomposition products of

water.

NASA TUNGSTEN REACTOR RADIATION CHEMISTRY STUDIES
G. H. Jenks, H. C. Savage, and E, G. BohImanm ... .o e e et et e e 58
Designs, equipment, and procedures are being developed for experiments to test the effects of irradiation on

loss of cadmium from CdSO4 solution under conditions of interest in the NASA Tungsten Water-Moderated Reactor.

CORROSION SUPPORT FOR VARIOUS PROJECTS
J. C. Griess, J. L. English, L. 1.. Fairchild, and P. D. Neumanm .. .................ccccooiiiiiriiioiiieie i 60
Corrosion studies were conducted on materials to be used in the High Flux Isotope, Advanced Test, and
Argonne Advanced Research reactors; some studies of Hastelloy N and nickel! for use in gas-phase fluidized+bed

fuel-element process equipment were also carried out.

6. Chemistry of High-Temperature Aqueous Solutions

ELECTRICAL CONDUCTANCE MEASUREMENTS OF AQUEQOUS SODIUM CHLORIDE SOLUTIONS
TO 800°C AND 4000 BARS
A. S. Quist, W. Jennings, Jr., and W. L. Marshall e et 63
The electrical conductances of aqueous sodium chloride sclutions from 0.001 tc 0.1 m were measured at tem-
peratures from 100 to 800°C; sodium chloride solutions still behaved as moderately strong electrolytes at high

temperatures and densities,

AQUEOUS SOLUBILITY OF MAGNETITE AT ELEVATED TEMPERATURES
F. H. Sweeton, R. W. Ray, and C. F. Baes, Jr. i e ittt ter e et tv et a e tee e et 64
The solubility of Feao4 is being studied as a function of pH at temperatures between 150 and 260°C.
yii

SOLUBILITIES OF CALCIUM HYDROXIDE AND SATURATION BEHAVIOR OF CALCIUM
HYDROXIDE-CALCIUM CARBONATE MIXTURES IN AQUEOUS SODIUM NITRATE
SOLUTIONS FROM 0.5 TO 350°C
L. B. Yeatts, Jr., and W. L. Marshall ..., et et e e a e nn et aaen e e et n e e e e e 65
The solubilities of calcium hydroxide and of calcium hydroxide—calcium carbonate mixtures in water and
aqueous sodium nitrate solutions were determined and used hoth to test further the applicability of an extended

Debye-Hickel equatlion and to obtain thermodynamic functions.

7. interaction of Water with Particulate Solids

SURFACE CHEMISTRY OF THORIA
C. H. Secoy
Heats of Immersion and Adsorption
H. F. Holmes, E. L. Fuller, Jr., and J. E. SHUCKEY .ottt e e e 68
Net differential heats of adsorption for several high- and low=-surface-area thoria samples have been calculated
from calorimetrically determined hezats of immersion.
Water Vapor Adsorption and Desorption
E. L. Fuller, Jr., and H. Fu HoOe S, i et ir e e e et r e et ie et e e e e rr e e eart b eaeeeereeetanrterareneees 69
Careful determination of adsorption and desorption isotherms for water vapor on thoria has revealed the complex
details of both.the chemisorption and physical adsorption processes,
Infrared Spectra of Adscrbed Species on Thoria
Co 80 BShoup, Jre oo e e ettt teeanwaseee e et eeiiEedieseeemeeeeeraseaeaseeeetsaeaeeeattensean e aaenetn e enain e 70
The infrared spectra of a thin thorium oxide disk as functions of pretreatment conditions substantiate the com-~
plex nature of the chemical water-adsorption process,
Elvctrokinetic Phenomena at the Thorium Oxide~Aguecus Solution Interface
C. 8. Shoup, JIr., and H. F. HOIMES ..o oo e e 71
An electrical study of the interface between thoria and agueous solutions of several electrolvtes has disclosed

surface conductances two or three orders of magnitude greater than those predicted by classical theory.

GAS EVOLUTION FROM SOL-GEL URANIUM-THORIUM OXIDE FUELS
D, N. Hess and B. A, Soldano........ccoiiiiiiii ettt A e eeet e eeen e eLhee e e eeentaeeeeee e e e e reinrae s aaeeneat s 72
The gas-adsorption capacity of thorium sol-gel is shown to be reduced by exposure to heated mixtures of COZ
and HZO’

PART . GAS-COOLED REACTORS

8. Diffusion Processes

TRANSPORT PROPERTIES OF GASES
Thermal Transpiretion. Rotational Relaxation Numbers for Nitrogen and Carbon Dioxide
A I, M A A UE KA it ettt et ettt e e e eeee ettt e e et eeeeeeeeen te ettt eeeeaseeeet A teeaeteniaaea e ea it tae i eataen e teearen e raraa e ehe s 77
The analysis of thermal transpiration data in accordance with the dusty-gas model appears teo provide a simple
method for the investigation of inelastic collisions.
Goseous Diffusion in Noble Gas Systems
A . M AU S A S ittt i et ttieetie ertit et e et et eee et e e eeet ettt ea et teaeentt ettt tetneer At e e nrar e te e eeentnrn e ee e enaanne s - 77
Preliminary analysis of diffusion data of the systems He-Kr, Ar-Kr, and Xe«Kr reaffirms the feasibility of em-

vloying viscosity measurements to derive diffusion coefficients at high temperatures.

Gaseous Diffusion in Porous Media
A. P. Malinauskas, F. A. Mason, and R. B, Evans HI. ... et e et e et h e ettt anee e iieee e eataeeaaaneareas 78
The additivity of diffugsive and viscous fluxes has been found to be theoretically justified; as a result, the

previous treatment of gas transport in the presence of a pressure gradient has been improved.
viii

SOLID-STATE TRANSPORT PROCESSES IN GRAPHITIC SYSTEMS
Recoil Phenomena
R. B. Evans 1III, J. L. Rutherford, and R. B. Perez. ... e 79
The recoil range for light fission fragments in General Electric pyrolytic carbon is greater than the range for
the heavy fragments (13 (land11 (), and the straggling factor to range value, /R, is 0.126.
Actinide Diffusion
R. B. Evans III, J. L. Rutherford, and F. L. Carlsen, Jr. ... e 80
Actinide diffusion at low concentrations in General Electric pyrocarbon was found not to be concentration-
dependent as in experiments with High Temperature Materials pyrocarbon, but preliminary constant-potential data
indicate high coefficients and low activation energies.
Self-Diffusion
R S T - R o £ 1) L U ORI 81
Equipment and technigues are being developed to study the diffusion of 14¢ in pyvrolytic carbens and various

graphites.

9. Reactions of Reactor Components with Oxidizing Gases

I.. G. Overholser

REACTIVITY OF ATJ GRAPHITE WITH LOW CONCENTRATIONS OF OXIDIZING AND REDUCING GASES
P S =1 =033 O OO 83
Experimental studies were made of the reactivity of AT]J graphite with low concentrations of water vapor,

carbon dioxide, methane, and carbon monoxide.

COMPATIBILITY OF PYROLYTIC-CARBON-COATED FUEL PARTICLES WITH WATER VAPOR
e M. Blood o ettt e ettt en et aa e e 85
Reaction-rate, coating-failure, and surface-area data were obtained for seven batches of pyrolytic-carbon-
coated fuel particles during and after exposure at 1000°C to partial pressures of water vapor of 4.5, 45, and 570

torrs.

COMPATIBILITY OF METALS WITH LOW CONCENTRATIONS OF CARBON MONOXIDE
Jo B BaK T e e e U 86
The effect of carbon monoxide at 250 to 300 ppm in He on molybdenum, gold-plated stainless steel, and mild
steel at 450 to 850°C has been evaluated.

10. lrradiation Behavior of High-Temperature Fuel Material

Oscar Sisman and J. G. Morgan

FISSION-GAS RELEASE FROM PYROLYTIC-CARBON-COATED FUEL PARTICLES
P. E. Reagan, J. W. Gooch, J. G. Morgan, T. W. Fulton, and C. D. Baumann...................co, 88
Fission-gas release rates from pyrolytic-carbon-coated UO2 fuel particles were low even at elevated tem-

peratures after high burnups.

POSTIRRADIATION EXAMINATION OF COATED FUEL PARTICLES
P. E. Reagan and E. L. L ong, Tl i e e e e e e e e 90
Several pyrolytic-carbon-coated UO2 particles have shown very little radiation damage at temperatures up t{o

1600°C and burnups to 25 at, % of the heavy metal.

POSTIRRADIATION TESTING OF COATED FUEL PARTICLES
M. T. Morgan, R. L.. Towns, and C. D. BaUmMaNTI ..o e e et e a e 94
Metallic fission product release studies of pyrolytic-carbon-coated fuel particles show possible effects of

burnup, coating density, and fuel composition on the fission product release rates.
ix

POSTIRRADIATION EXAMINATION OF FUEL.ED GRAPHITE SPHERES
D. R. Cuneo, J. G. Morgan, H., E. Robertson, C. D. Baumann, and E. L. Long, Jr. .....nnni. 95
Postirradiation examination of AVR-type spheres fueled with coated particles led to the conclusion that, after

10% burnup, the spheres had retained acceptable structural integrity.

POSTIRRADIATION EXAMINATIOM OF EGCR FUEL ELLEMENT PROTOTYPE CAPSULES
M. F. Osborne, E. L. Long, Jr., H. E. Robertson, and J. G. MOFZAam ... ..ottt et a e eeeneas a8
An EGCR prototype capsule that contained productien~-run UO2 pellets was examined after irradiation fo a

purnup of 10,000 Mwd/metric ton UOQ, and we found no major detrimental changes.

11. Fission-Gas Release During Fissioning of UO,

R. M. Carroll, Oscar Sisman, T. W. Fulton, R. B. Perez, and G. M. Watson

EXPERIMENTAL .. e e e e eeer e aas 99
The medified in-pile assembly has been used te study fission-gas release rates while the fission rate (at
constant temperature) or the temperature {at constant fission rate) was varied sinusoidally in a carefully contrelled

wWay.

A T HE M AT AL MO D E L e it ittt tttaaaae s ee i aaee e trtrs neeaaaaae eesastatsterarnesea e s e eenaaraanessnsrebes sanaeaaaanes 100
The mathematical model of a defect-trapping theory of fission-gas behavior has been refined and tested in a

preliminary way with our experimental data.

12. Miscellaneous Studies for Solid-Fueled Reactors

EQUILIBRIUM STUDIES IN THE SYSTEM ThOZ-UOZ-Uflg
Lo O. Gilpatrick and Cu He SOy i i e e b 103
Althcugh the general features of the phase diagram for this system in equilibrium with air from 700 to 1600°C
tad been reported previously, several details have begen either confirmed or slightly revised by sdditional experi-

ments with improved techniques.

BEHAVIOR OF REFRACTORY-METAL CARBIDES UNDER IRRADIATION
. W. Keilholtz, R. . Moore, and M. F. Osborne o i s e s 104
The effects of fast neutrons on monocarbides of Ti, Zr, Nb, Ta, and W are being investigated at temperatures
from 100 to 1400°C in order to evaluate their potential use in nuclear systems requiring extremely high power

densities,

EFFECTS OF FAST-MEUTRON IRRADIATION ON OXIDES

G. W. Keitholtz, R, E. Moore, and M. F. Os8bBOME e T 105
Trradiation effects on sintered MgQ, AIEO%, and BeQ were investigated over the temperature interval 100~
1100°C and the fast-neutron dose range of 0.2 to 5.1 X 104! peutrons/cm’ in srder te determine realistic conditions

for uze of these oxides in nuclear systewms and to establish mechavisms of neulron damage.

ANNEALIMG OF IRRADIATION-INDUCED THERMAL CONDUCTIVITY CHANGES OF CERAMICS

C. D. Boppoooooeeie e e e e 109
The annealing of the neutron-induced thermal conductivity change was measured in nine ceramic materials

after an irradiation dose uwp to 2 ¥ 1017 fast neutrons/cmz.
PART IV. OTHER ORNL PROGRAMS

13. Chemical Support for Saline Water Program

THERMODYNAMICS OF GYPSUM IN AQUEOQUS SODIUM CHLORIDE SOLUTIONS
W. L. Marshall and Ruth SIuS her e et et e ettt e e e s 113
From an extensive study of the phase behavior of gypsum in aqueous sodium chloride and of those solutions
cosaturated with sodium chloride or a double salt, NaZSO4-5CaSO4-3H20, thermodynamic functions were calculated

both for a standard state and at high ionic strengths.

THE OSMOTIC BEHAVIOR OF SIMULLATED SEA-SALT SOLUTIONS AT 123°C
P. B. Bien and B. A. Soldano .. ... e e e 115
Varying the nature of the multivalent ionic components of seawater gave rise to significant changes in the

osmotic behavior of saline solutions at elevated temperature.

ALUMINUM- AND TITANIUM-ALLQY CORROSION IN SALINE WATERS AT ELEVATED TEMPERATURES
E. G. Bohlmann, J. C. Griess, F. A. Posey, and J. F. Winesette e 117
Continuing studies of the corresion of aluminum and titanium alloys in saline water have demonstrated the

superiority of 5454 aluminum and substantial inverse temperature dependence of the titanium pitting potential.

CHEMISTRY OF SCALE CONTROL
E. L. Compere and J. E. Savolaiflen . 119
Considerations of chemical factors related to the economic control of scale deposition in sSeawater distillation
equipment have included the possibility of carbon dioxide addition to prevent alkaline scale formation, the tend-
encies of certain elements to form ion pairs in seawater, and the rate factors in therinal-precipitation processes for

calcium sulfate.

14. Effects of Radiation on Grganic Materials

W. W. Parkinson and QOscar Sisman

EFFECT OF RADIATION ON POLYMERS
W. W. Parkinson, W. K. Kirkland, and R. M. K eV S O .o e 121
The tensile properties of polytetrafluoroethylene are retained to doses in excess of 2 x 107 rads in a vacuum;

elongation at break shows a maximum at dosages near 3 X 10* rads whether irradiated in air or vacuumnl,

RADIATION-INDUCED REACTIONS OF HYDROCARBONS
R. M. Keyser and W. W, Parkini s O o e e 121
Gamma radiation of saturated solutions of ammonia in n-hexane and in hexene=1 produced less than the de-

tectable limit (about 0.2 molecule per 100 ev absorbed) of amines.

ADDITION REACTIONS OF FURAN DERIVATIVES
C. D. Bopp, W. D, Burch, and W. W, Parkins Om o e e e e 123
Scolutions of dihydrofuran and cyclohexene in saturated furan derivatives were irradiated in a survey of reactions
utilizing isotopic-decay radiation, and it was found that yields of adducts or dimers and trimers approached 10

molecules per 100 ev.

15. Fivoride Studies for Other ORNL Programs

THE CHEMISTRY OF CHROMIUM IN THE FLLUORIDE VOLATIILITY PROCESS
B. J. Sturm and R. E. T aomia . o i e e e e e e et e e ans 125
Fluorides and oxyfluorides of chromium in the oxidation states Cr(III) to Cr(VI) were synthesized and partially

characterized; free energy of formation for solid Cr02F2 was found to be 189 kcal/mole at 298°K.
xi

PREPARATION OF LiF SINGLE CRYSTALS BY THE MODIFIED STCCKRBARGER METHOD
12, (5. IR085 AN R B, oM. it et et ettt e e et ra e e et b ann e as b reras s n et e a e e e e n e e na et aren 126
A large (270-g) single crystal of Li¥, grown in a platinum-lined capsule, was found to contain a lower con-

centration of heavy-metal contaminants than is currently detectable by activation analysis, that is, <1.86 pph.

16. Chemical Support for the Controiled Thermonuclear Research Frogram
R. A, Strehlow

VACUUM ANALYSIS IN AN EXPERIMENTAL PLASMA DEVICE
B A B0 Lottt irii et oo oeed e et en ettt nee e n e et et teek e et e e e e A eaeha e A eRs e nra bt e e e ea s 127
A relatively simple mass analyzer installed on DCX-2 and its counterpart on a much simpler similar vacuum

system are being used to define the gaseous environmenti in this experimental plasma device,

INTERACTION OF TRITIUM WITH THERMONUCLEAR-REACTOR MATERIALS
Bt B Al L S ert i ivren ettt et ee e e et e e et e n b e et oo aa e e e eeata i aeaaeaan ety e e 128
Literature information and 2 simple diffusion model were used to estimate the steady-«state heldup of tritium in

the atom-bombarded wall of a proposed thermonuclear reactor.

HYDROGEN SURCHARGING OF MOLYBDENUM IN A GLOW DISCHARGE
D M RICRAP S OG0l s o iee et v e te s ettt e et et e e ettt et ettt eeeee e et nen e eea et ee e n e e et et an aean e 129
Molvbdenum at elevated temperatures picks up relatively little hydrogen when pombarded with protons in a glow
discharge; bombardment al lower temperatures of molybdenum whose surface is contaminated leads to veclusion of

large guantities of the gas.

MEASUREMENT OF GAS LOAD FROM SOURCE OF ELECTROMAGNETIC SEPARATOR
B2 Al LB LMW oot ettt et oo e e4eeis et eeeeeeee e eeeeeeeestaeeeneeed ettt e e tneanan et en e e e aeetn s 129
The flow rate of chlerinating agent (0014) (rom the ion source of a calutron has been measured during operation,
and the data have been applied in design of a differential pumping system for the scurce of a new electromagnetic

separator.

APPEARANCE-POTENTIAL MEASUREMENTS FROM TIME-OF-FLIGHT MASS SPECTROMETRY
Jo D REUMATL. o o e et e e e v e e hus e e r e e e s e a e 130
The time-of-flight mass spectrometer with a retarding-potential circuit has been calibrated with several per-
manent gases and shown to be capable of precize evaluation of appearance potentials with source pressures ag

low as 5 » 1077 torr.

PART V. NUCLEAR SAFETY

17. Nuclear Safety Tests in Major Facilities

FISSION PRODUCTS FROM FUELS UNDER REACTOR-TRANSIENT CONDITIONS
G. W. Parker, R, A. Lorenz and J. G. Wilhelm o e e 135
Fission product release from UO2 in stainless steel or Zircaloy cladding has been successfully determined

during exposure o transients in TREAT of specimens in water or high-pressure steam.

FISSION PRODUCTS FROM SIMULATED LOSS-OF.-COCLANT ACCIDENTS IN ORR
W. E. Browning, Jr., C. E. Miller, Jr., W. H. Montgomery, B. F. Roberts, R. P. Shields,
0. W. Thomas, A. F. Roemer, and J. G. WIlhelm .. 137
Effects of atmosphere, cladding material, fuel burnup, maxzimum fuel temperature, and seresol aging on be-

havior of fission products are being investigated in in-pile experiments.
X11

FI5510N PRODUCTS FROM HIGH-BURNUPF U02
G. W. Parker, W. M. Martin, G, E. Creek, R. A, Lorenz, and C. J. Barton................ooiii i 138
Behavior of fission products from UO2 previously irradiated to 1000 Mwd/ton and melted in the Containment

Mockup Facility was similar to that observed in similar experiments with simulated high-burnup fuels.

THE CONTAINMENT RESEARCH INSTALL ATION
G, W, Parker and W. Jo Martinl o e e et e e e e e et e e e e e 139
The Containment Research Installation, an enlarged and improved version of the Containment Mockup Facility,

is nearly complete, and component testing is expected to begin early in 1966,

ANALYSIS OF PLANS FOR SCALE-UP IN NUCIL.EAR SAFETY PRUGRAM
C. E. Miller, Jr., and W E. BrowWmiiig, J o i ettt e ettt e e e ettt e r e e s 141
Qur analysis of experiments planned in the U.S. to study the behavior of fission products following a reactor

accident suggests fhat additional intermediate experiments, 1% and 10% of the size of the LOFT, are necessary.

18. Laboratory-Scale Supporting Studies

RETENTION OF RADIOIODINE AND METHYL IODIDE BY ACTIVATED CARBONS
W. E. Browning, Jr., R. D. Ackley, J. E. Attrill, G. W. Parker, G. E. Creek, F. V. Hensley, R. E. Adams,

J. D. Dake, D. C. Evans, and A. F ermelh oo e e e e e 142
It has been found that activated carbon impregnated with inactive 12712 or K127I removes iodine activity as
CHSISII from gas streams under a variety of conditions.

IDENTITY OF VAPOR FORMS OF RADIOIODINE
W. E. Browning, Jr., R. E. Adams, R. D. Ackley, J. E. Attrill, J. D. Dake, and D. 1. Ford ..... ...l 144
Gas chromatography with simultaneous electron-capture and radiation detection has been utilized in determining
the identity of various iodine vapor forms encountered and in ascertaining the purity of elemental iodine and methyl

iodide socurces employed in laboratory iodine-behavior studies.

DEVELOPMENT OF METHODS FOR DISTINGUISHING AND MEASURING GAS-BORNE FORMS OF
FISSION PRODUCTS
W. E. Browaing, Jr., R. E. Adams, R. D. Ackley, R. L. Bennett, M. D. Silverman, J. Truitt, W. H. Hinds,
A. F. Roemer, Jr., B. A. Cameron, J. D. Dake, and D. T, Ford. ... e, 144
Characterization devices such as composite diffusion tubes, fibrous-filter analyzers, gas chromatographs, and
May packs are being examined and modified for application in high-humidity experiments designed to study the be-

havior of the various forms of gas-borne fission products under reactor accident conditions.

REMOVAL OF PARTICULATE MATERIALS FROM GASES UNDER ACCIDENT CONDITIONS
W. E. Browning, Jr., R. E, Adams, J. S. Gill, and G. L. KoChanny ... e e 145
l.aboratory investigations are being continued of the filtration efficiency of high-efficiency filter media for

particulate aerosols which simulate those expected to be generated during reactor accidents.

IGNITION OF CHARCOAL ADSORBERS BY FISSION PRODUCT DECAY HEAT
W. E. Browning, Jr., C. E. Miller, Jr., B. F. Roberts, and R. P, Shields ... 146
Preliminary results of a study to determine the effects of iodine-decay heat on the ignition temperature of
charcoal adsorbers show that the ignition temperature is not greatly affected by a decay-heat load greater than

those expected at various power reactors.

P B L LA T LN S e ettt ettt bttt ettt 147
PAPERS PRESENTED AT SCIENTIFIC AND TECHNICAL MEETINGS ... 153
Part |

Molten-Salt Reactors

 
1. Phase Equilibrium and Crystallographic Studies

LIQUID-LIQUID IMMISCIBILITY
IN THE SYSTEM LiF-BeF -ZiF

H. A. Friedman R. E. Thoma

Beryllium fluoride systems have been of interest
because they provide “‘weakened models™! of Si0,
The weakened model contains cations
and anions of the same radius as those in the
silicate system, but with half the ionic charge. In
MgO-S10,, a liquid immiscibility region is found;
in the weakened model LiF-BeF, this has not been
found, although metastable glasses, mutually in-
soluble, are encountered at 60 to 90 mole % BeF, ,
where the melts are highly viscous. Our recent
work on the Lil-BeF -ZiF system” reveals that
stable liquid-liguid regions do exist in this ternary,
although they do not extend stably to the LiF-Bel
binary.

systems.

Silicate and BeF glagses are structurally anal-
ogous and are composed of networks of bridging
ions, oxide and fluoride ions respectively. Other
cations, particularly the low velent ones, if added
to these glasses, break the bridges and are referred
to as modifiers; their effect on viscosity is quite
pronounced. The possible role of ZrF, as a net-
work modifier or network former has not previously
been examined.

Our liquid-liquid immiscibility studies were made
with a new high-temperature centrifuge that devel-
oped a centrifugal force of 530 x ¢ and gave pood
layer separation in sealed metal capsules. Results

 

1V. M. Goldschmidt, *'Geochemische Verteilungs-

gesetze der Flemente: VIII, Untersuchungen uber Bau
und Eigenschaften von Krystallen®® (Geochemical Dis-
tribution L.aws of the Elements: VIII, Researches on
Structure and Properties of Crystals), Skrifter Norske
Videnskaps-Akad, Oslo, It Mat.-Naturv. KI. 1926, No. 8,
pp. 7—156 {1927); Ceram. Abstr. 6(7), 308 (1927).

2Reactor Chem. Div. Ann. Progr. Rept. Jan. 31,
1965, ORNL.~3789, p. 3.

were based on chemical analyses of portions of
quenched samples obtained after 1-hr periods at
elevated temperature in the centrifuge. In pre-
liminary trials with PbO-B O, glasses, good agree-
ment was obtained with the work of Geller and
Bunting.” lsothermal tie lines at 550, 650, and
700°C for the immiscibility region in LiF-Bel, -
ZrF , are shown in Fig. i.1.

The possibility of stable liquid-liquid immis-
cibility in the LiF-BeF, system was given special
attention. This was in part prompted by results on
the chemical activity of BeF, as a function of
composition in the LiF-BeF, binary, as obtained
from HF-H, 0 equilibria,” where there was an
indication that immiscibility might occur for 80
mole % BeF , at temperatures above 700°C.

Ten samples representing six binary composi-
tions in the range from 80 to 34 mcle % Bel, were
centrifuged and examined; no phase separation was
found. Further, visual observations of an 80 mole
% BeF | melt were made under helium in a glove
box and furmace assembly by Cantor and Ward.’
Again the liquid was single phase, although on the
initial heating to 700°C a lower viscous layer and
a thinner, easily stirred, upper layer persisted;
only after heating to 850°C did the melt become
homogeneous. The melt remained homogeneous
through subsequent temperature cycles ifrom room
temperature to 850°C. The conclusion was reached
that there is no stable immiscibility above the
liguidus in the LiF-BeF , hinary.

 

3R. F. Geller and E. N. Buonting, J. Res. Nail, Bur.
Std. 18, 585 (1937).

*A. L. Mathews and C. F. Baes, Oxide Chemistry and
Thermodynamics of Molten Lithium Fluoride—Beryilium
Fluoride by Egquilibration with Gaseous Water—Hydrogen
Fluoride Mixtures, ORNL-TM-1129, p. 104 (May 7,
1965).

5, Cantor and W. T. Ward, MSRP Semiann. Progrc
Rept. Aug. 31, 1965, ORNL-3872 (in press).
Zrfa

 

3LIF-4ZrF, J N

 

Z 2 LIQUIDS ~_
2LIF-ZrF, /0 N

 

\/

LiF

Fig. 1.1. Liquid-Ligquid Immiscibility in the System LiF-BeF,-ZrF,.

(c) 700°C isotherm.

SODIUM FLUORIDE-SCANDIUM
FLUORIDE PHASE EQUILIBRIA

R. H. Karraker® R. E. Thoma

One of the most important factors which influence
the number and types of intermediate compounds

 

SORINS research participant, Memphis State University,
Memphis, Tenn.

ORNL-DWS 66 - 9614

    
    

 

 

BeF,

(a) 550°C isotherm; (b) 650°C isotherm;

formed between pairs of highly ionic compounds
such as the alkali fluorides and group III fluorides
is the relative size of the cations.” Size effects
are of diminishing significance as bonding becomes
less ionic or when unusual ligand effects, such as
those induced by the lanthanide and actinide
contractions, become of importance. In the course

 

’R. E. Thoma, Inorg. Chem. 1, 220 (1962).
  
  
 

 

| |
1300 - —1 Nt/ ludt =146

|

{°C}

Hoo

-

900 N

 

MPERATUR

 

e
1
=

5 Naf + 9 Lui

700 b N e e d e

 

 

 

500

 

 

 

 

 

300 L——

’ (C‘)J { (
NaY/Sc o et 2EE e b ‘Naf/m“: 248 — T

ORNL—-DWG 66 ~962

e e

     

 

 

 

 

 

W’ JF}}

= E] . |
—dhal

= “} |
o JI’_J By AR |

Fig. 1.2. Comparison of the Systems {a) NaF-LuF3, (b) NaF-ScFj3, and {c} NaF-AlF4.

of the recently completed investigation of the
sodium fluoride—lanthanide trifluoride systems (see
*#Compositional Variability in Sodium Fluoride—
Lanthanide Fluoride Complex Compounds,” this
report), it became evident that within the series
cation charge density was beginning to obscure the
effect of size. It was therefore of great interest to
obtain ioformation concerning the comparative
phase behavior among the following group of sys-
tems, which are ranked below in order of increasing
difference of the tripositive cation to alkali cation
radius ratio:

T\IaJF«LaF3 0.923
NaF-BiF3 1.054
NaF-Lqu 1.156
]\I::flf"-InI*“3 1.210
NaF-8clk 3 1.256
K.F-LaFS 1.254
NaF-AIFS 2.178

Investigation of the system NalF-ScF,, reported
previously in preliminary form,® has now been
completed. The equilibrium phase diagram of the
system is compared with those of NaF-LuF, and
Nab-AlF, in Fig. 1.2. Equilibrium reactions of
NaF and SLFJ bear closer resemblance to those
of NaF and AlF, than to NaF and LuF
the

5o despite
fact that the scandium ion radius (D.78) is

 

8Rrea(:tor Chemn, Div.
1965, ORNL-378%9.

Ann. Progr. Rept. Jan. 31,

nearer to Lu®? (0.848) than to A1** (0.45). Similar
to the Nal-AlF , system, molten NaF-ScF  mix-
tures form a cryolite-like phase, 3NaF.Sck
which, like cryolite, is dimorphic. Crystals of each
form are isomorphous with their cryolite analogs.
The phase 3NaF -ScF, undergoes a strongly
exothermal solid-state inversion when cooled below
680°C and inverts to poorly crystallized and highly
twinned material. Single crystals of the low-
temperature form of 3Nal - ScF , were obtained by
growth from NaF-ScF, mixtures of 70-30 mole %
composition. X-ray diffraction analysis of these
single crystals indicated that they are of mono-
clinic symmetry (see “*Crystallographic Data on
New Compounds,*’ this report) and are isomorphous
with cryolite. A high-temperature foem of cryolite
is reported” to occur as a face-centered cubic
phase. Preliminary results of high-temperature
x-ray diffractometric experiments with 3NalF . ScF
indicate that the high-temperature forms of 3Nal -
5c¢F, and 3NaF - AlF | are isomorphous.

Whereas the intermediate phase in the 3Naf -
AlF -AlF , subsystem is of 5:3 composition, only
a 1:1 phase is formed in the 3NaF -ScF -5cF,
subsystem. The compound NaF - 5ck | melts incon-
gruently to ScF, and liquid at 660°C. X-ray dif-
fraction data obtained from single crystals of
NaF - 5¢F, revealed it to be of hexagonal ' sym-
metry (see ““Crystallographic Data on New Com-
pounds,”’ this report).

 

9E, G. Steward and H. P. Rooksby, Acta Cryst. A,
48 (1953),
Table 1.1. Inveriont Equilibria in the System NaF-ScF

 

Compositicn Invariant

(mole % SCFS) Temperature (°C)

L refers to liquid

Type of Equilibrium

3

Phase Reaction at Specified Temperature

 

17 800 Eutectic L. == NaF 4 @ -3NaF - S(:F3
25 885 Congruent melting L &= a-3NaF -Sch
point
35 680 Inversion of 3NaF -ScF 4-3NaF -ScF ; === f3-3NaF -ScF
peritectic
38 650 Eutectic L= B~3NaF . S(:F3 + NaF . ScF3
42 660 Peritectic

 

We believe that the greater resemblance of the
system NaF-SCF3 to NaF-AlF, rather than to
NaF-LuF3 arises as a result of the lanthanide
contraction, producing the Lu®" ion of extra-
ordinarily high charge density. Quantitative rela-
tionships of the combined effects of ion size and
charge density have been developed by Dietzel.1?
They require accurate information of interionic
distances, which for these complex fluorides must
await determination of the crystal structures.

The composition and temperatures of invariant
equilibrium points in the system Na\F-ScF3 are
listed in Table 1.1,

COMPOSITIONAL VYARIABILITY IN SODIUM
FLUORIDE-LANTHANIDE TRIFLUORIDE
COMPLEX COMPOUNDS

R. E. Thoma H. Insley
(. M. Hebert

Investigation of the sodium fluoride—~lanthanide
trifluoride systems'! has been completed. A
salient feature of the NaF-LnF , binary systems
is the unusual behavioral sequence produced by
the compositional variability of the fluorite-like
phases. This behavior shows the remarkable effect
of cation size and polarizability in lanthanide
systems. As shown in Table 1.2, the cubic phase
is not formed at all in the systems NaF-LaF and

IOA, Dietzel, Z, Elektrochem, 48, 9 (1942),

11I'\’e-ac:tor Chem. Div. Ann. Frogr. Rept. Jan. 31,
1965, ORNL-3789, p. 13.

 

L+ SCF3 == I. + NalF -SCF3

NaF~CeF3, is formed by Nal' and PrFS, and is
extended to an increasingly broad composition
range throughout the rest of the system sequence.
A trend also develops toward broadening the com-
position limits and extension of the temperature
range through which the cubic and orthorhombic
phases are stable. The trends toward increasing
number, compositional variability, and stability of
the NaF-LnF, crystal phases appear to be dis-
tinct, not only with respect to the crystal chem-
istry of the complex compounds, but also in their
modes of crystallization. No satisfactory theory
has as yet been developed to explain the quanti-
tative aspects of compositional variability of
NaF-Lnk , phases. Qualitatively, the trends are
believed to indicate an increase in the lattice
energies of the crystalline phases resulting from
the lanthanide contraction and a corresponding
decrease in polarizability. The decrease in polar-
izability is probably the most influential factor in
enabling compositional variability of the phases to
be extended with increasing atomic number of the
lanthanide. As represented by the Lorentz-Lorenz
equation, molar refraction approximates the volume
occupied by the constituent ions in crystalline
phases. A measure of the crystal free space can
therefore be estimated from the difference between
molar volume, as computed from unit-cell dimen-
sions, and the molar refraction. As the atomic
number increases, small increases in the free
space fraction are noted for the lanthanide tri-
fluorides and for the hexagonal NaF - LnF , phases.
A more pronounced increase in free space fraction
is found for the fluorite-like NaF-LnF3 phases,
Table 1.2. Optical and X-Ray Diffraction Data for NuF-LnF3 Crystalline Phoses

 

 

A. Hexagonal NoF * Lnf-’3
Ln ag (A) T(ay) cq (A) T(cy) N N

 

 

 

 

 

 

 

 

 

 

¢ o
La 6.157 C.006 3.822 0.008 1.486 1.500
Ce 6.131 0.006 3.776 0.004 1.493 1.514
Pr 6.123 0.604 3.743 0.002 1.494 1.516
Nd 6.100 0.002 3.711 0.002 1.493 1.515
Pm (6.056) (3.670) (1.492) (1.515)
Sm 6.051 0.61¢ 3.640 G.007 1.492 1.516
Eu 6.044 0.603 3.613 0.003 1.492 1.516
Gd 6.020 0.003 3.001 0.008 1.483 1.507
Th 6.003 0.004 3.580 0.002 1.486 1.506
Dy 5.085 0.004 3.554 0.003 1.486 1.510
Ho 5.851 0.601 3.528 0.002 1.486 1.510
Er 5.959 0.002 3.514 0.002 1.482 1.504
Tm 5.953 0.6032 3.494 0.002 1.476 1.496
Yb 5.929 0.002 3.471 0.002 1.482 1.504
Lau 5.912 0.003 3.458 0.003 1.484 1.506
(Y5 5.967 0.002 3.523 0.002 1.464 1.486

8. Cubic NqF-LnF3 Phases ot Composition Limits
NaF-Rich Composition Limit SNal - QLnFS Composition Limit
b Compositioa X-Ray Density ~ X-Ray Density
(mole % LnFS) R. L 2o (4) (g/ce) R.L g (a) (g/cc)
Pr 55.5 1.512 5.710 1.526 5,720 5.768
Nd 55.0 1.506 5.670 4.431 1.524 5.685 £.962
Pm (54.5) (1.5C03 (5.630) (4.620) (1.522) (5.5655) (6.131D
Sm 53.5 1.485 5,605 4,702 1.520 5.627 5.315
Eu 52.5 1.486 5.575 4.809 av 1.519 5.616 6.392
Gd 51.5 1.470 5.550 4,978 1.502 5.594 6.620
Th 0.0 1.472 5.535 5.051 1.504 5.565 6.772
Dy 48.5 1.462 5.505 5,205 1.504 5.547 £6.939
Ho 47.0 1.458 5.490 5.286 1.504 5.5125 7.093
r 45.0 1.440 5.475 5.387 1.493 5.514 7.263
Tr 43.5 1.427 5.460 5.4&6 1.494 5,493 7.336
i 41.5 1.415 5.444G 5.611 1.488 5.480 7.510
La 29,0 1.4G2 5.425 5.6938 1.482 5. 463 7.580
. Orthorhomhic SNoF - ‘?Lan
Kefractive Index Lattice Constants (A)
Lo . SO . )
N, N, N a, b, cq
Dy 1.514 5.547 27.74 4.804
Ho 1,510 5.525 27.63 4.784
Fr 1.504 1.506 1.506 5.514 27.57 4.775
I'm 1.501 5.493 27.47 4.757
Yh 1.482 1.494 1.495 5.480 27.40 4.746
L 1.480 1.496 1.487 5.463 27.32 4,731

 
Table 1.2 {(continued)

 

Refractive Index

 

Lattice Constants (A)

 

 

 

Ln Symmetry Density Ref.
NgorNa NworN,y a, .!b0 o (g/ec)
La Hexagonal 1.597 1.603 7.186 7.352 5.936 a
Ce Hexagonal 1.607 1.613 7.112 7.279 6.157 b
Pr Hexagonal 1.614 1.618 7.075 7.238 6.14 C
Nd Hexagonal 1.621 1.628 7.030 7.200 6.500 a
Sm Orthorhombic 1.577 1.608 6.669 7.059 4,405 6.643 a
Eu Orthorhombic 1.572 1.600 6.622 7.019 4.396 6.793 d
Gd Orthorhombic 1.570 1.600 6.571 6.985 4.393 7.056 e
Tb Orthorhombic 1.570 1.600 6.513 6.949 4.384 7.236 e
Dy Orthorhombic 1.570 1.600 6.460 6,906 4.376 7.465 e
Ho Orthorhombic 1.566 1.598 6.404 6.875 4.379 7.644 e
Er Orthorhombic 1.566 1.598 6.354 6.848 4.380 7.814 e
Tm Orthorhombic 1.564 1.598 6.283 6.811 4,408 7.971 e
¥b Orthorhombic 1.558 1.568 6.216 6.786 4,434 8.168
Lu Orthorhombic 1.554 1.558 6.181 6.731 4.446 8.44

“E. Staritzky and L. B. Asprey,
PASTM X-Ray Diffraction Card No. 8-45.
“ASTM X-Ray Diffraction Card No. 6-0325.

 

 

Anal. Chem. 29, 857 (1957).

9a, Zalkin and D. H. Templeton, J. Am. Chem. Soc. 75, 2453 (1953).

®ASTM X-Ray Diffraction Card No. 12-788.

with an even greater increase in the free space
fraction for the equimolar cubic phases than for the
cubic SNaF :9LnF , phases as Z increases. Sub-
stitutional solid solution of Ln®" into the fluorite
unit cell gives rise to cation vacancies but is
partly compensated by filling of interstitial posi-
tions with fluoride ions, as described by Roy and
Roy.!?  The fact that the free space fraction
occupied by the ions in the cubic structure at the
S5NaF - 9LnF3 saturation limit is even greater than
composition lends additional
evidence to the validity of the previous interstitial
fluorine model for solution mechanism.

We conclude that the combined effect of the
reduction in polarizability of the lanthanide ions
and the increase in free space within the crystal
lattice is to reduce the specificity of Na® and
Ln®"' ions with regard to the cation sites they
occupy in fluorite-like, orthorhombic, and hexagonal

at the equimolar

 

'2b. M. Roy and R. Roy, J. Electrochem. Soc.
421 (1964).

111,

crystals. The consequence of this effect is the
increase observed in compositional variability of
the crystal phases in the NaF-LnF , systems with
increasing atomic number of the lanthanide.

PHASE EQUILIBRIUM STUDIES
IN THE U0,-ZrO, SYSTEM

K. A. Romberger H. H. Stone
C. F. Baes, ]Jr.

In a previous report, a new phase diagram was
suggested for the UO,-ZrO, system in which
essentially no solid solution formation was indi-
cated below about 1100°C.'°
which is in variance with previously published

This conclusion,

 

13 A. Romberger et al., Reactor Chem. Div. Ann.
Progr. Rept. Jan. 31, 1965, ORN1.-3789, p. 243.
 

 

    
   

  
     

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL.-DWG 5&-~963
MOLE PERCENT U0,
N0 99.5 990  98.5 100 90 80 70 €0 50 40 30 20 10 O 05 1.0 0.5 0
I T T 7T T Ty T [T
1go - CUBI —gsoc L1QuID ZBO()/ ~ TETRAGONAL s.s. 1180
o \
i
140~ © CUBICs.s. — 2400 SLIQUID - 1040
o + \ +
o TETRAG. 5.5, \ CURIC s.5.
\ ‘.
HOO |- 2000 - - \ 100
4 CLBIC s. 8.
—_— 4 -
o 4 TETRAG.| |/ CUBIC s.s. 2 o
— ' S5.5. -
o DE0 — 1800 |- ~11600 |- ~ 1080 1
: CUBIC s.5. \ | 7T s s I CUBIC < 5. o iz
1 -+ . + et
2 MONCCLINIC .. " li VETRAGGNAL 5.5 il | MONOCLINIC s.5. =
414020 - 1200 ) Y200 w1020 ¢
= o * $ % x4 O =
= o J | I A &
Tl l C s s
230 | - 800 | CUBIL 5.5, l »‘;L 500 |- | 980
i ! \
/ | MONDCLINIC s.¢. | /, \
/ | | /, \ MONGCLINIC 5.8,
240 it — 400+ | |/ ] 1400 - - 340
. // I i! ’ \
L e e L \
500 [ | ol v e ey N A Jsoo
0 0.5 1.0 £5 0 10 20 30D 40 S0 60 TO BC 90 100 385 920 995 100

MCLE PERCENT Zr0,

LEGEND :

LIQUID - LAMBERTSOMN AND MUELLER (23)
COHEN
{161}

MURMBPTON AND ROY (22)

 

(ABCVE 1600°C]
——— (BELOW $600°C)

-~

)

O "WET" CHEMICAL ANALYSIS
X ACTIVATION ANALYSIS FOR U
AND SCHANER

(BELOW 1800°C ) ORNL PROPOSED

Fig. 1.3. U0,-Zr0, Phase Diagram.

phase diagrams for this s;ystem,””w was based

primarily on x-ray and petrographic examination
ot UO,-Zr0, solids equilibrated with fluoride
melts al temperatures from 500 to 700°C'? and
frem 900 to 1020°C."?

 

141
(1963).

154
160

Cohen and B. E. Schaner, J. Nucl., Mater. 9, 18

M. Wolten, J. Am. Chem. Scc. 80, 4772 {1958).
E. Evans, f. Am. Ceram. Soc. 43, 443 (1960).

R. Wright, D. E. Kizer, and D. L. Keller, Studies
U()Z-Z,r()2 System, BMI-1689 (Aug. 27, 1964).

17T.
in the
185, M. Voronov, E. A. Voitekhova, and A. S. Damlin,
Proc, U.N. Intem. Conf. Peaceful Uses At. Energy,
2nd, Geneva, 1958 6, 221 (1958).

19¢. F. Baes, Jr., J. H. Shaffer, and H. F. McDuffie,
Trans. Am. Nucl. Soc. 6, 393 (1963).

Equilibrations of the oxides with fluoride melts
have since been completed for temperatures up to
1150°C. The oxide solids from these latter equil-
ibrations were not only examined by x-ray and
petrographic techniques as before, but also these
and all of the previously equilibrated solids have
been chemically analyzed io determine the U-Zr
composition of the individual phases. (The separa-
tion of the phases was performed using hot nitric
acid, a liquid in which UO, is readily soluble and
Zr0O, is virtually insoluble.)

Incorporation of this new informalion gave the
more detailed diagram for the U0 ,-ZtO, system
shown in Fig. 1.3, The monoclinic-tetragonal
transformation of the Zr0 -rich solid solutions
occurred at 1110 1 5°C. At this temperature
Z:0, is soluble in cubic UO, to about 0.40 mole
%, while UO, is soluble in monoclinic ZiO, to
about 0.15 mole %, and in tetragonal ZrQ, to about
1.0 mole %.
temperature for pure ZrO, has been given by
Mumpton and Roy?® as 1170°C. No attempt was
made to determine this inversion temperature in
the present work. However, microscopic investiga-
tion of a sample of pure ZrO, equilibrated at
1190°C has shown that it definitely had been
tetragonal. Hence, the Mumpton and Roy value of
1170°C falls within our limits of 1110 to 1190°C.

Below the transition temperature of 1110°C, the
saturating solids are the cubic UQ,-rich and the
monoclinic ZrO _-rich solid solutions. However,
even at 1110°C, these solid solutions are very
dilute.
ities will decrease exponentially as the tempera-

The monoclinic-tetragonal transition

Moreover, it is expected that the solubil-

ture is reduced. Hence, for all practical purposes,
the equilibrium solids at lower temperatures are
the pure phase materials, that is, pure U0, and
pure Zr()z.

The correlation between the data given in the
present work and results presented previously by
other authors'*~ 182921 {5 al50 shown in Fig. 1.3.
For temperatures below 1700°C, solid lines have
been used to represent the solubility limits as
deterinined from log XMO2 vs 1/7T plots of our
data and all other pertinent data presently avail-
able in the literature.!*~ 1829 The dashed lines
those from the recently published diagram
of Cohen and Schaner.'* These lines begin to
deviate from each other at about 1600°C. As the
temperature is further reduced, the deviations

are

increase rapidly.

We believe the differences in the results reported
here and those reported previously by others
reflect the inability of the other investigators to
obtain equilibrinm products at temperatures below
1600°C solely via solid-state reactions. These
temperatures are more than a thousand degrees
below the melting points of the pure oxides. At
such temperatures, solid-state
reactions of the oxides should be very slow. In
the present investigation the use of a fused fluo-
ride phase provided a path by which the slowness
of the solid-state reaction could be overcome and
equilibrium could be achieved.

relatively low

 

20p. A Mumpton and R. Roy, J. Am. Ceram. Soc.
43, 234 (1960).

21W. A. Lambertson and M. H. Mueller, J. Am. Ceram.
Soc. 36, 365 (1953).

10

THE CRYSTAL STRUCTURE
OF LiUF,

G. ID. Brunton

The compound LiUF_ was originally described

as Li,{,U6F31 in the molten-salt

system Lik-
UF4.22"25 The crystal structure was determined
in order to resolve the stoichiometry and because
LiUF _ is the uranium salt which would be deposited
if fue! should freeze in the Molten-Salt Reactor
Experiment.

Tetragonal LilJE crystallizes in space group
I41/a with a = 14.884 + 0.002 A and ¢ = 6.5467 +
0.0003 A. The calculated density is 6.23 g/cc
with 16 formula weights per unit cell,
parameters are listed in Table 1.3.

Atomic

Figure 1.4 is an illustration of two asymmetric
units of LiUF | related by a center of symmetiy —
approximately one-eighth of a uunit cell.
1.5 is a stereoscopic pair showing the contents
of one unit cell.

Figure

Each uranium is surrounded by
nine ftluorine ions which form the corners of a
14-faced polyhedron having the form of a prism
with triangular bases and with a four-faced pyramid
on each of the three prism faces. Each lithium
ion is surrounded by six fluorine ions at the corners
of a distorted octahedron.

The structure of LiUF . consists of 16 uranium-
containing polyhedra and 16 lithium-containing
octahedra linked together so that every uranium
polyhedron shares an end with its centrosym-
metrical neighbor, shares corners with five other
uranium polyhedra, and shares a face, an edge,
and two corners, respectively, with four lithium
octahedra. Each lithium octahedron shares edges
with three other lithium octahedra, and shares
an edge, a face, and two corners, respectively,
with four uranium polyhedra. This linkage gives
spiraling cross-connected chains parallel to the
C axis.

 

221.. A. Harris, The Crystal Structures of 7:6 Type
Compounds of Alkali ¥Fluorides with Uranium Tetra-
fluoride, CF-58-3-15 (March 6, 1958).

23¢, J. Barton et al., J. Am. Ceram. Soc. 41, 63
{1958),

241, A. Harris, G. D. White,
Phys. Chem. 63, 1974 (1959).

25C. F. Weaver ef al., J. Am. Ceram. Soc. 43, 213
(1960).

and R. E. Thoma, J.
11

F5 ORNL - DWG 65 ~{2543

 

 

N

o
-

o

Fig. 1.4. Two Centrosymmetrically Related Asymmetric Units of LiUdFg.
Table 1.3, Atomic Parameters far LiUFS

 

 

Atom  x + % 10°

 

U 0.06176(0.4) 0.05640(0.4) 0.24623(1.7)  0.00026(0.3)
F,  0.0361(9)  0.0523(9)  0.8734(25)  0.0017(2)
F, 0.2008(9)  0.1754(9)  0.7602(28)  0.0018(2)
F,  0.10859)  0.1831(9)  0.7504(22)  0.0013(2)
F, 0.2048(10) 0.0918(10)  0.3576(26)  0.0020(3)
F, 0.0484(%)  0.1677(10)  0.4787(26)  0.0018(2)

Li  0.068(30)  0.163(31)  0.773(84) 0.0015(%)

*Isotropic temperature factors:

822 'Bll '
c*2

B .~ R .
337 L7

12 =77 13 23

ORMNé_—DVWG 65 12542

 

Fig. 1.5. One Unit Cell of LiUF; (Stereographic Pair).

THE CRYSTAL STRUCTURE OF Li AlF,

26

J. H. Burns A. C. Tennissen

A recent phase diagram study on the ternary
systems involving AlF ; and LiF, NaF, and KF 27

 

26Research Participant from Lamar State College of
Technology, Beaumont, Tex.

2’R., E. Thoma, B. ]J. Sturm, and E. H, Guinn, Molfen-
Salt Solvents for Fluoride Volatility Processing of
Aluminum-Maérix Nuclear Fuel Elements, ORNIL.-3594
(August 1964).

indicated that LisAlFfi, NaSAlFé, and K3A1F‘6
are each components of these systems; and while
the latter two were known to have the cryolite
structure, LiSAIF() had not been studied. The
existence of sixfold and twelvefold coordination
for the alkali-metal ions in cryolite was not ex-
pected to hold for Li*, and indeed a new structure
type was found.

Crystals of L13A1F6 were prepared by melting
together under vacuum a 3:1 mixture of LiF and
AlF .. Some of the compound distilled to a cold
finger in the vessel, but the larger crystals were
obtained from the residue in the bottom. Powder
x-ray diffraction examination showed both to be
virtually the same phases, with a little excess
AlF , on the cold finger,

X-ray precession photographs were used to
establish that the crystals were otthorhombic with
unit-cell dimensions: a = 9.54, b = 8.23, ¢ = 4.88
(£0.02 A). Systematically absent reflections: hOf,
L = 2n, and 0k, k + | = 2n, indicated the probable
space groups Pama and Pna? .» which differ by the
presence of a center of symmetry. The density,
calculated with four formula weights in the unit
cell, is 2.80 g/cm?®. The density was measured
to be between 2.6 and 3.0 by flotation in heavy
liquids. |

Intensity data for determination of atomic posi-
tions was obtained by photographing Akl zones,

13

with I =0, 1, 2, 3, and 4, employing the multiple-
{ilm Weissenberg technique. The intensities of
the reflections were estimated by comparison with
calibrated film strips. Some 238 independent
reflections were measured and their intensities
reduced to structure factors, F .

The Patterson vector method was applied to
locate the Al and F atoms, and calculation of a
difference Fourier map then was successful in
determining the Li positions. All atoms of the
structure occupy fourfold general sites of space
group PnaZl. Refinement of atomic positions and
individual isotropic temperature factors was car-
ried out by least squares.
the discrepancy factor,

The present value of

ZlF, |-

 

 

 

£ VE R,

z

is 0.13,

ORNL—DW3 66 — 259

 

Fig. 1.6. Crystal Structure of LizAlF,.
It is expected that somewhat better agreement
will be achieved with further refinement, but the
atomic sites given in Table 1.4 are not expected
to change appreciably.

A drawing of the Li AII'_ structure is shown in
Fig. 1.6. The presence of A1F63“‘ ions is apparent;
these are joined together in such a manner as to
provide octahedral coordination for each of the Lit
ions. Bond distances are within the normal range,
but further refinement of the structure will be
awaited before reporting them.

Table 1.4. Atomic Porameters for Li3A1F6

 

Atom X y z

Li{1) 0.357 0.372 0.525
Li(2) 0.131 0.407 0.499
Li{3) 0.349 0.527 0.017
Al 0.128 0.246 0°

F(1) D.222 0.086 0.145
F({2) 0.026 0,249 0.299
F(3) 0.238 0.212 0.694
F{4) 0.022 0.390 0.856
F(5) 0.244 0.391 0.185
F(6) Q0.020 0.091 0.837

 

Chosen arbitrarily to fix the origin on 2 .

REFINEMENT OF THE CRYSTAL
STRUCTURE OF (NH ) ,MnF,

D. R. Sears

The manganic ion is unstable with respect to
disproportionation, and as a consequence, simple
salts of trivalent manganese are rare. Neverthe-
less, it is desirable to collect definitive informa-
tion on the structures of manganic salts when
possible, for complex anions containing trivalent
manganese can be expected to exhibit significant
crystal field effects. In particular, octahedrally

14

coordinated Mn>" should be tetragonally distorted
as a result of the Jahn-Teller effect.?8

The compound (NH ) MnF _ crystallizes in the
centrosymmetric orthorhombic space group D;:-
Pnma, with a = 6.20 = 0.03, b = 7.94 + (.01, and
c = 10.72 £ 0.01 A.
spectrometric zonal =x-ray intensity data were
collected in 1957 from specimens of (NH,),MnF _
prepared by T. S. Piper. The data had resulted in
a structure determination,?® but it had not been
possible to perform a satisfactory refinement with
the computing facilities available then. The
opportunity arose to perform a modern full matrix
least-squares with the

existing AOI and Ok! intensity data, using a modifi-
cation of the Busing-Martin-Levy program ORFLS.??
Anomalous dispersion effects were included in the
calculations.

Final least-squares-adjusted atomic positions
and thermal parameters resulted in an agreement
factor,

Excellent photographic and

anisotropic refinement

between observed and calculated structure factors,
lFO | and ’FC , as low as 0.068 for all observable
data. The contents of one unit cell and adjacent
atoms are shown in Fig. 1.7. Each octahedron has

 

the composition MnF , with manganese occupying
the center and fluorines the vertices. Ammonium
groups appear as stippled spheres and form infinite
hexagonal nets in mirror planes at y - 1/;, ?‘,;_ In-
finite kinked chains of MnF  octahedra, sharing
opposite vertices, pass through these ammonium
nets. The manganese atoms lie on symmetry
centers, and consequently there are but three
crystallographically distinct fluorines.

Bond distances and angles are presented in
Table 1.5. In this table, £ is the fluorine atom
shared by adjacent octahedra. Nonbonding contacts
between adjacent fluorines in the octahedra were
in the range 2.59 to 2.84 A. Although nitrogen-
fluorine distances as low as 2.81 were discovered,

25"See*_, for example, R. Dingle, Inorg. Chem. 4, 1287
(1965).

29D. R. Sears, Ph.D, thesis, Cornell University,
1G658.

30, R. Busing and K. O. Martin, ORFLS, a Fortran
Crystallographic Least-Squares Program, ORNL~TM-305
(1962).
15

ORNL-DWG 65-5847

 

 

 

Fig. 1.7. Schematic View of Unit Cell of (NH,);MnF s and Selected Atoms of Adjocent Ceils. Manguanese occurs

at the center of sach octabedron, ond fluorine at earch vertex. Nitrogen atoms are stippled spheres.

Table 1.5. Bond Distances™ and Angles

Distances

 

Bond Type Length {A) Standard Error {(A)

 

(2) Mn-F 2.091 (2.101) 0,005 (0.006)

(&) ¥n-F | 1838 (1.853) 0,009 (0.009)

(2) MooF 1.842 (1.853) 0.004 (0.004)

 

 

 

 

Angles
Bond Angle Angle Standard Error
FoMn-F 89.4° 90.5° 0.5°
F Mo-F o 37.7° 92.3° G.3°
F MaeF 89.7° 90.3% 0,3°
143.4° 0.8~

I’vIn-FI-Mn

 

Mistances in parentheses are corrected for themmal
metion, with fluorine assuamed to **ride’” op the manganese.

a high coordination number of nitrogen (7 to 9) and
the infrared spectrum®® of {NK~§4)2MnF5 suggest
strongly that N-H-I' hydrogen boonding is abszent
ot very weak.

The bond distasces and angles involving man-
ganese reveal that there iz indeed a pronounced
tetragonal disiortion of the octahedral manganese
coordination shell. Doubtless, sharing of the F
atoms between adiacent octahedra contributes to
the substaniiel elongation of the Mn-F_bonds, but
a Jahn-Teller digtortion must contribute signifi-
cantly. However, quontitative separation of
effects is not practicable.

the

Dingle®® has discussed both the crystal spectrum
of (NH 4),}M11F . and the solution spectrum of trivalent
in concentrated HF with excess K7

manganesea ,
presumed to contain MnF63’".

The nearly perfect
correspondence  of the spectra in the 10,000 to
28,000 cm™! region suggests that tetragonally
distorted MnFé3m octahedra appear in solution;
such a disftortion in aqueous MnF63” solutions
surely cannot result from chain formation. We
infer that the distortion arises largely from ligand
field effects in solution and, by extension, does
so also in the crystal.

HIGH-TEMPERATURE X-RAY STUDIES

G. D. Brunton
J- H

D. R. Sears
. Burns

A Materials Research Corporation high-tempera-~
ture x-ray diffractometer furnace has been employed
in several studies of phase change at elevated
temperatures, with the following resulis.

The orthorhombic lanthanide trifluorides SmF
through LuF , and YF  have been shown to undergo
a phase transition below their melting points.® Of
these compounds, SmF . through HoF , invert to
the hexagonal LaF , structure while YF , and ErF,
through Lul , invert to a crystalline form which
is probably hexagonal but whose unit-cell dimen-
sions suggest that it is different from the LaF
structure. The unit-cell dimensions of the hexagonal
phases and the approximate inversion temperatures
are summarized in Table 1.6.

Table 1.6, UWnit-Cell Parameters and Approximdte

Inversion Temperatures

 

 

Transition
“0 (4 “o (A) Temperature (OC)
SmkF 3 7.07 7.24 355
EuF‘3 7.04 7.26 700
GdF 7.06 7.20 900
TbF 7.03 7.10 950
DyF3 7.01 7.05 1030
HoF 7.01 7.08 1070
ErF3 6.97 8.27 1075
TmF3 7.03 8.35 1030
YbF 6.99 8.32 985
LuF3 6.96 8.30 945
YF 7.13 8.45 1052

 

16

Phase equilibrium studies in the UO -ZiO,
system have been continued, the principal objects
being: (1) to investigate the composition-tempera-
ture region in which Thoma®! has postulated the
existence of a compound of approximate composi-
tion UO 3Zr02, (2) to furnish x-ray diffraction
data for the continuing fluoride-flux c‘qulhbratmn
experiments of Romberger, Stone, and Baes;*?
and (3) to develop
constant—composition relationships for the sys-
tem UO, (cubic)-ZrO (cubic).

Tentative results to date suggest that the sys-
tem cubic UOzwcubic Zi0, at room temperature
displays a linear relationship between a  and
mole fraction Z10,, at least to 50% Z10,. The
data can be extrapolated to a, = 5.08 A at 100%
ZtO,, in reasonable agreement with the results
obtained by extrapolation of the CeO,-Zr0O, data
(5.11 A) due to Duwez and Odell.??

Attempts to establish the existence of a compound
of composition near UO,k -3Z:0, above 1600°C
have not been successful. We have not obtained
pure tetragonal starting material near this mole
ratio. X-ray diffraction experiments in the temper-
ature range 1400 to 1650°C at a ZrO,:UO, ratio
of 1.36, using a pure tetragonal solid-solution
starting material, have resulted in exsolution, to
form nearly pure cubic UO, and tetragonal ZrO -
UQ, solid solution richer in Z10 .

Our high-temperature diffractonieter attachment
is being modified extensively to permit operation
at higher temperatures in better vacuum and to
diminish problems in window opacity caused by
filament and sample volatilization.

room-temperature lattice-

Some high-temperature x-ray diffraction meas-
urements were made in order to relate our single-
crystal study of Li3A1F6 to the reported* poly-
morphism:

Qo sy ——> O i
225° 475° 575° 705°

€ .

The pattern for the Li,Al¥  phase whose single
crystals were analyzed (see Burns and Tennissen,

BIR. E. Thowa, Reactor Chem. Div. Ann. Progr. Repi.
Jan. 31, 1965, ORNL.-3789, p. 245,

32K, A. Romberger et al., Reactor Chem. Div. Ann.
Praogr. Rept. Jan. 31, 1965, ORNL.-3739, p. 244,

33p, Duwez and F. Odell, J. Am. Ceram. Soc.
247 (1950).

34P. Gross, Fulmer Research Institute, Stoke Poges,
FEngland, unpublished results.

33,
17

Table 1.7. Crystallographic Data

 

Compound .'P~TaSc1:*‘4 NagScF 6 BI-KLaF4 I_,i4UF§3 RbPaFG
System Hexagonal Monoclinic Hexagonal Orthorhombic Orthorhombic
Unit-cell a=12.97 0,03 a=560 £0.02 2=6.53010.002 a-=9.960 £0.001 a-8.06 £0.02

dimensions, A

c=9.27 T 0.02 bh=581 12002 c=23.8001+0.002 5:=0.882 L0.001 b=12.00 =0.03

c=8.12 *0.02 c=5.986 1 0.002 c:= 585 3 0.02

[3=90%5" 57

P3 12) P3 21\
Space group P3212) or P3221/ PZl/n
Formula weights 2
par unit cell
Calculated 2.87
density, g/cm3
Notes a a b

P53 or P63/m Prnma or PnaZl Cnuna or Abm?2

3/2 4 4
4,51 4.71 5.05
c d

3ee “Sodium Flusride—Scandinm Fluoride Phase Equilibria,’’ this report.

bIsostmctural with cryolite, NasAIFS,

“Previously studied by W. 1. Zachariasen, Acfa Cryst. 1, 265 (1948).

dPrepared by O, L. Keller, ORNI, Chemistry Division.

this report) was calculated. With it and the knowl-
edge of the sluggishness of the a —~ § transition,
it was concluded that the preparation of the com-
pound resulted in a mixture of f§ and y. This
mixture was heated to 720°C in the diffractometer
futnace, where both of the original components
disappeared; by slow cooling the component cor-
responding to the single crystals reappeared at
about 500°C. From these data we tentatively
concluded that the single-crystal study is of
v-Li AIF . Further analysis of this compound will
be carried cut when the diffractometer is modified,

CRYSTALLOGRAPHIC DATA
ON NEW COMPOUNDS

J. H. Burns D. R. Sears
G. D. Brunton

Unit-cell dimensions and symmetry information
for five compounds have been obtained by single-
crystal x-ray diffraction and are summarized in
Table 1.7. These data are necessary in order to

determine whether each compound has a known
structural type; or, if it does not, they are a
prerequisite  for proceeding with the structure
determination. Of these compounds, Na ScF
was found to be of the cryolite type (N33A1F6),
while the others represent new structure types.
Three of these, 8 -KLaF , Li, UF _, and RbPaFfi,
are being subjected to complete structure analyses.

THE CRYSTAL STRUCTURE
OF Na,]ZiréF:”

J. H. Burns R. D. Ellison®?
H. A. Levy?®®

This compound is representative of a structural
type which has occurred in numerous moften-salt
systems. Among those which have been observed?®

 

PORNL Chemistry Division.

36p . K. Thoma, ed., Phase Diagrams of Nuclear
Reactor Matsrials, ORNL-~-2548 (Nov. 2, 1959),
18

 

 

 

Table 1.8. Struciural Parameters for No7Zr6F3]a

Atom X y z f))“ [‘322 633 BIQ _['313 623
Na(l) 0.0793 0.3038 0.4827 0.0033 0.0041 0.0050 0.0020 --0.0006 —0.0001
Na(2) 0 0 1/2 0.0046 0.0046 0.0054 0.0023 0 0

Zr 0.189¢6 0.0515 0.1791 0.0017 0.0013 0.0020 0.0007 0.G002 --(,0001
F 0.3556 0.1114 0.0016 0.0020 0.0022 0.0033 0.0010 0.0005 —0.0006
F{2) 0.1837 0.0554 0.3944 0.0032 0.0023 0.0026 0.0012 ---0.0000 —0.0002
¥ (3 0.2734 0.3706 0.4243 0.0028 0.0017 0.0036 0.0009 —0.0008 --3.0005
F (4) 0.2087 C.1586 0.0020 0.0038 0.0035 0.0028 0.0023 --0.0002 0.0006
F(5) 0.2433 0.5417 0.4413 0.0030 (.0045 (.0072 0.0028 0.0016 —-0.0020
F(6) 0 0 0 0.0176 0.01706 0.0947 0.00D88 0 0

A exagonal coordinates for all atoms in general positions, 18(f), of space group R3, except F(6) and Na(2),
which are in 3{a) and 3(b) respectively.

and for which the x-ray powder diagrams have been ically by a PDP-5 computer.*’ The structure was

reported®’ are the 7:6 compounds of NaF, KF, and solved by an analysis of the Patterson function
RbE with UK, and ThF , as well as 7NH I - 6UF‘4.38 computed with these data and was refined by
The existence of this curious stoichiometry has least squares. The positional and anisotropic
been suggested as being due to the possibility of  thermal parameters are given in Table 1.8.
an NaZrF _ structure with the presence of sites A portion of the structure is shown in Fig. 1.8.
where one extra Nal per six NaZrF_ units could The atoms are represented by ellipsoids showing
be accommodated.®?® In order to check this hypoth- their thermal motion. The very large ellipsoid
esis and/or ascertain the structural causes of corresponds to the ‘““extra’’ fluorine, which is
this formula, a complete crystal-structure deter- indeed contained in a polyhedron of six Zr atoms
mination was cairied out. bridged by F atoms. The large motion of this
About 900 x-ray intensity data from a small F atom means that either it is rattling about quite
spherical single crystal were collected with a freely or else there is a statistical disorder in
Picker full-circle goniometer controlled automat- ~ which this atom is bonded randomly to one of the

six Zr atoms at a time. The “‘extra’’ Na atom is
Na(2); it is closely coordinated by six F neighbors
and by six additional F neighbors at a little greater
distance. Each Zr atom has eight F atoms about
it and each Na(l) has seven F atoms.

37G. D. Brunton et al., Crystallographic Data for
Some Metal Fluorides, Chlorides, and Oxides, ORNL-
3761 (February 1965). o

38
R. A. Penneman et al., Inorg. Chem. 3, 309 (1964). Wy R Busing, R. D. Ellison, and H. A. Levy,

%0, A. Agron and R. D. Ellison, J. Phys. Chem. 63, Chemistry Div., Ann. Progr. Rept. May 1965, ORNL-~
2076 (1959). 3832, p. 128.

 
19

ORNL — DWG 66 ~ 260

 

 

Fig. 1.8. A Portion of the NazZr,Fy Structure Showing Probability Ellipsoids of Thermal Motion of the Atoms.
2. Chemical Studies of Molten Salts

OXIDE CHEMISTRY OF LiF-BeF -ZrF , MELTS

C. F. Baes, ]r. B. F. Hitch

In previously reported studies of the oxide

chemistry of LiF-BeF 2=~ZrF4 melts, equilibrium
quotients were determined by the transpiration
method?! for the following reactions:

H,0(8) + 2F7(d) == 07 (d) + 2HF (@) ,
00 = P/ P07 5

HIO(g) + F7(d) == OH™(d) + HF(g) ,
Qp= (PHF/PI-I O)[0H~] (D)
2

2 2
H, C(g) + — MF _(d)== - MO_, (s) + 2HF (&),
z z

2
- PHF

(M?+ = Be?t, Zt**, U

Oy /Py o M7127%)
- 2

(3)

Reactions (1) and (2) were of interest because of
their role in the contamination of molten fluorides
with oxide, their role in the removal of oxides by
HF during salt purification, and, more recently,
because of their use in oxide analysis.® Reaction
(3) has been of interest because it relates the
thermodynamic properties of the dissolved {luoride
MF _ to the available thermodynamic data for
MO , (), H,0(¢), and HF(g). Estimates of the
soluéility product of the oxide MOZ/2 have been
obtained from the equilibrium quotients QO and

 

1A, L. Mathews, C. F. Baes, ]Jr., and B. F. Hitch,
Rcactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNI1-3789, pp. 56—65; alsc ref, 7, this section.

MsR Program Semiann. Progr. Repé. Sept. 30, 1965,
ORNI.-3897 (in press).

20

O

uo_,, - (047077 = IM?T[03]2/2 | (4)

However, these estimates have been of limited
accuracy because of the difficulties in determin-
ing QO by the transpiration method,

During the past year, a more direct method for
determining oxide solubilities has been adopted.
A measured volume of salt, presaturated by
equilibration with excess solid oxide, is filtered
through a sintered nickel filter into a heated re-
ceiving vessel, It is then sparged with an H -HF
mixture and the total amount of water evolved
determined by Karl-Fischer titration, Thus far,
essentially complete removal of oxide from the
filtered salt samples has been accomplished in
3 hr or less by use of an influent HF pressure of
0.05 atm. The blank has been equivalent to
~2 x 1072 mole/kg of oxide (32 ppm), the sensi-
tivity has been ~.5 x 107 * mole/kg (8 ppm), while
the experimental values ranged from 1.5 x 1072 to
1.5 x 107 * mole/kg.

This method presently is being used to redeter=-
miie the solubilities of ZrO_ in (2LiF-BeF2) +
ZrE  mixtures, which simulate MSRE fuel salt and
flush salt mixtures. The results obtained thus far
are compared in Fig. 2.1 with previously reported’
estimates based on measured values of the ratio
(_)O/Qzr in such salt mixtures. When completed,
these measurements should permit the establishe
ment of more reliable tolerance limits for oxide

levels in the MSRE and in future molten-salt
reactors. In addition, the more accurate solubility
values - when combined with er values [reaction

(3)] — should yield cortespondingly more reliable
values of ¢ . These will be useful in optimizing
oxide removal procedures during salt purification
and salt analysis,
21

0.05
=

L

>

2

3 002
£

zZ

]

5 om
o

'.‘

=z

LiJ

Q

& 0.005
(W]

wl

o3

=

<o

0.002 2

0.0
0.0

~
T~
<y -b/ofl
2 8g ¥
4/*0

 

o.coz G.005 Q.4 002

ORNL-DWG $6-964

~ Oe
%ay,
/O/V

0.05 04 02 05 i

ZrF, CONCENTRATION (moles/kg)

Fig. 2.1. Variation of Oxide Concentration with ZrFy Concentration in (2LiF-BeFy) + ZrFy Melts Saturated

with BeO or Zr05. Dushed curves represent previous estimates based on transpiration measurements of Qp: Oger

and Q7

THERMODYNAMICS OF MOLTEN LiF.BeF,
SOLUTIONS?

C. F. Baes, Jr.

In connection with the past development of the
MSRE, a considerable number of heterogeneous
eguilibria have been studied which involved a
molten fluoride solvent of the approximate composi-
tion 2LiF-BeF2. These equilibria have included:
(1) reduction by hydrogen of dissolved NiF , Fek,
oLl BeFZ,S and ¥JF4;6 (2) metathetic reactions
of gaseous HF with dissolved oxide,” sulfide,®
and iodide;? (3) metathetic reactions between the
solid oxides and dissolved fluorides of Be(Il),
U(IV), Z(IV),'? and Th(IV);'! and (4) solubility

 

3Based upon a paper presented at the IAEA Symposium
on Thermodynamics with Emphasis on Nuclear Materials
and Atomic Transport in Solids, Vienna, Austria,
July 22--27, 1965,

4(",. M. Blood, Solubility and Stability of Structural
Metal Difluorides in Molten Fluoride Mixtures, ORNL-
CFe61-5-4 (Sept, 21, 1961).

G, Dirian and K. A. Romberger, Reactor Chem. Div.
Ann, Progr. Rept. Jan. 31, 1965, ORNL~3789, pp.
7679,

7,10,12

of sparingly soluble oxides and flue-

rides,*-13-14

 

°a. Long, Reactor Chem. Div. Ann. Progr. Rept.
Jan. 31, 1965, ORNL=3789, pp. 65-72.

7A. L. Mathews and C. F, Baes, Jr., Oxide Chemistry
and Thermodynamics of Molten Lithium Fluoride—
Berylliam Fluoride by Equilibration with Gaseous Water—
Hydragen Fluoride Mixtures, ORNL-TM-1129 (May 7,
1965); A, L. Mathews, Ph.D, thesis, The Chemistry and
Thermodynamics of Malten Lithium Fluoride—-Beryllium
Fluoride by Equilibration with Gaseous Water--Hydrogen
Filuoride Mixtures, University of Mississippi, Oxford,
Miss. (June 1965),

8H, 11, Stone and C. F. Baes, Jr., Reactor Chem. Div.
Ann. Progr., Rept. Jan. 31, 1985, ORNL=3789, pp.
7276,

B, F. Freasier, C. ¥, Baes, Jr., and H. H, Stone, in
this report, p. 38,

mj. E. Eorgan et al.,, Reactor Chem. Div. Ann,
Progr. Rept. Jan. 31, 1964, ORNL-3591, pp. 45--46,

“J’. H. Shaffer and G. M. Watsen, Reactor Chem. Div,
Ann. Progr. Rept., jan, 31, 1960, ORNL.-2931, p. 90,

]2C. F, Baes, Jr., and B, ¥, Hitch, Reactor Chem.

Div, Ann. Progr. Rept. fan. 31, 1965, ORNL-3789,
pp. 61-85; *QOxide Chemistry of LiF-BeFQv»ZrF4

Melts,?! this report.

3w, 1. Ward et al., Solubility Relations Among Rare-
Earth Fluorides in Selecfed Molten Fluoride Solvents,
ORNL-2749 (Qet. 13, 1959); see also W. R. Grimes
et al., Chem. Eng. Progr. 55(27), 6570 (1859).

14(:, J. Barton, J. Phys. Chem. 64, 3069 (1960),
 

 

Table 2.1. Formation Heats ond Free Energies in 2LiF-BeF, (773 to 1000°K)
a —Afif _AG! Estimated T Error
Solute (kcal) (kcal) (kcal)

1 Li +F 142,70 124.79 0.8
2 La3t43F 405, 5 351.8 7
3 cedtuzE” 400,5 347.0 7
4 sm3ta3p” 389.5 337.5 7
5 Be?t 4or” 242.75 211.80 1
6 Th*' 1aF 424.0 5
7 7zt 4 ar 451.85 386.21 1.7
8 UtTaar” 444,61 386,48 1.8
9 Ut 13r” 336.73 296.19 1.8
10 ce2tvor 171.82 150,41 0.9
11 Fe? 4 2F 154.69 132.92 0.8
12 Ni2T 4 2F” 146.87 110.61 0.8
13 Be?t 1 02 131.91 105,10 0.6
14 el i 20u” 170.54 159.35 0.7
15 Be?t 421”7 96.46 63.56 1
16 Belt +s2” <79 (873°K)

 

“The standard state of the ions is the hypotheticalmole fraction solution in 2LiF-BeF2, with the

. Lt
exception of Li , Be?

During the past year, this information has been
reviewed and summarized by thermodynamic methods
This
seems fitting and proper at this stage in the

as a means of extending its usefulness.

development of the molten-salt reactor concept.

Future molten-salt reactors may well employ salt
mixtures similar to those of the MSRE considered
here; hence, a knowledge of the thermodynamics
of these solutions could prove generally useful.

A consistent set of formation heats and free
energies has been calculated for the solvent
components and for various solutes at low cone

centration in 2LiF-BeF  (Table 2.1). This was

+
, and F , for which the standard s*ate is 2LiF—BeF2.

done by combining the observed equilibrium con-
stants (a measure of the free energy difference

between the reactants and the products of a given
reaction) with existing thermochemical data'®~17

for the solid and gaseous reactants and products.

15]ANAF (Joint Army-Navy-Air Force) Interitn Ther-
mochemical Tables, Revised Mar. 31, 1965, Thermal
Research lL.ab., Dow Chemical Co., Midland, Mich.

161, Brewer, pp. 76--192 in The Chemistry and
Metallurgy of Miscellaneous Materials; Thermodynamics,
ed, by L. L. Quill, McGraw-Hill, New York, 1950,

M. H. Rand and O. Kubaschewski, The Thermal

Properties of Uranium Compounds, p. 71, Interscience,
New York, 1963,
23

ORNL-DWS 65~-8880

 

 

 

 

ACTIVITY COEFFICIENT

- REFERENCE
COMPOSITION

 

 

 

 

 

 

0.4

3 R - zfr= ?7/

    

 

0.3 033 035 037 0.39

XEeFZ

   

1C0

 

 

 

60 -

 

 

 

40

 

 

 

 

 

 

 

 

 

 

 

 

 

002 004 085 008

Fig. 2.2. Variation of Activity Coefficients in LiF-BeF, at 606°C. bLeft, as a function of XBer at low solute

concentration; right, VZrF4 as a function of XZrF4 and }’UF4 as a function of XUF4 added to 2LiF-BeF2. In this

figure z/r is the ratic of cation charge to radius.

The partial molal free energy of formation, AG, of
a solute, MFZ, at a given mole fraction, XMF;

may be calculated f{rom the AGT value listed in
Table 2.1 by use of the expression

AG = AG! 4 RT 1n (X (%)

ME TME )
Z z

wherein

)‘MFZ = “MP{,/(“MF? TLE T ger ?) - 6

g i - e . - -y .
Ihe activity coetficient, yMFZ’ is defined as
unity if X

x

is 2LiFwBeF2. Variation of

is low and the solvent composition
d

YME with XE@FO an

with solute concentration is shown for a few sol-
utes in Fig., 2.2. These curves were estimated
from observed dependence of heterogensous equi-
libria on melt composition. From these it appears
that sctivity coefficient variations may be corres

lated with the cation charge () to cation radius
(r) ratio, but more such data are needed to estab-
lish the generalily of the correlation.

Table 2.2 lists standzid electrede polentials for
various half-cell reactions, arbitrarily referred fo
the following half-cell reaction:

HF(g) + o w2 F () + V2 (@, E°=0. (@)

The manner in which these potentials are calcu-
lated is meost easily seen by peinting out that any
combination of half-cell reactions which gives a
complete reaction of the form

will vield the corresponding AGE value in Table
2.1 (n is the number of equivalents of charge and
F is the faradayv).
24

Table 2.2. Calculated Electrode Potentials in
2LiF.BeF23 (773 to 1000°K)

 

 

 

 

 

 

Temperature
Half-Cell Reactions'b E° (IOOOOK) (v) Coefficient

(mv/°C)
Li'+ e = Li(s) ~2.541 0.821
La®® 1 3¢ == Lags) ~2.21 0.821
Ce i 3e = Ce(s) ~2.14 0.817
Sm >t 1 3e = Sm(s) 2,01 0,795
Bel® 4 2e = Be(s) ~1.721 0,715
Th*t 1 4e = Th(s) ~1.73
Zr*t i de = Zr(s) ~1.316 0.755
Uttt de = U(s) 1,319 0.674
U3ty 3e += U(s) 1,410 0.630
vitte = Ut —1.044 0.807
Cr2t s 2e = Cr(s) ~0.390 0.505
Felt= 2e &= Fe(s) —0.011 0.516
NiZt+ 2¢ == Nils) +0,473 0.830
HF(g)+ e = F~ + ,H (&) 0 0
BL@+e =1 —0.343 0.002
1/ng(g) + 2 = 8§77 <0,10
70, @ + 20 = 0% 0.558 0.124
1/Qc,z(,g) + 1/2112(@ + e v OH 1.734 0.472
1/2F2(g) +e = F 2,871 0.044

 

 

€Calculated from Aéf values in Table 2.1.

bStandard states for ions are defined in footnote a of Table 2.1.

VAPOR PRESSURES OF FLUORIDE MELTS species, and to obtain data of importance to the
Molten-Salt Reactor FProgram.

S. Cantor D.S. Heu'® Rodebush-Dixon'? and boiling point methods were

W. T. Ward used to measure total pressure over 16 melts

covering the entire composition range. Each melt
was measured over at least a 185° temperature
Toto!l Pressure Measurements range; the pressure range usually covered 1 to
100 mm Hg. For all cases except one, the linear
The vapor pressures of the system LiF.BeF  expression
are being investigated to derive thermodynamic
activities, to determine the significant vapor log p(mm) = A — B/T(°K) (9

18 gummer employee, 1963, from the University of 9. 1. Rodebush and A. L. Dixon, Phys. Rev.
California, Berkeley, 26, 851 (1925).
Table 2.3. Vapor Pressure in the LiF‘-Ber System

 

Melt Composition

Equation: log p(mm) = A — B/T("K)

 

 

{(mole %) Temperature Range Measured

ST o e z 5
190 7791147 10.491 10,953
84.99 15.01 826-1116.5 19.648 11,230
75.01 24.99 890--1112.5 10.255 10,756
70.00 30,00 8431150 10,296 10,879
65.00 35.00 857--1112 9.787 10,266
57.54 42.46 866---1121 8.817 10,449
530,00 50,00 886-1071 S.134 9,738
45.04 54.96 9321155 9.096 9,044
39,95 60.06 930--1224 8.993 16,058
36.00 64,00 9501214 9.279 10,688
30.00 70.00 9661281 B.660 10,138
25.00 75.00 10201272 8.836 10,526
11.00 89,00 891.-1239 6.711 8,175
7.00 93.00 0781234 8.702 11,042
3.00 97.00 1020~1270 10.062 13,178

100 10261272 See below”

 

?The equation for pure LiF is log p = 3,619 ~ 15,4530/T - 6.039 log T.

provided adequate fit to the data (4 and B are
constants). The data are summarized in Table
2.3, The constants for each equation were obe
tained by the method of least squares, Isotherms
at 1000 and 1100°C are shown in Fig, 2.3.

To obtain some unotion of the vapor species,
vapor was collected in the tubes of the vessel at
the conclusion of vapor pressure measurements
for several compositions. Chemical analysis
showed only traces of lithium ion in condensates
where the composition of the melt was 75 or
greater mole % Ber. For melts with 70 or less
mole % BeF , considerable quantities of lithium
were found in the condensates. All that may be
concluded from these analyses is (1) at greatet
than 75 mole % BeF,, the vapor is virtually pure
BEFZ, and (2) at less than 70 mole % Bel’*‘z, the
vapor becomes more complicated; the vapor prob-

ably contains the compounds LiBeF, and

LiQBeF,‘,zo The only activity coefficients that
may be calculated from these data alone are those
of Be¥ at melt concentrations of 75 mole % or
greater in Ber. For these concentrations the
activity coefficients of BeF  at temperatures of
1000 and 1100°C were found to be greater than
unity {varying between 1.02 and 1.10), These
values of the activity coefficient are in reasone
able accord with results of the H, 0-HF equilibrae
tion studies?? of the LlF—BeFZ system,

Manometric vapor pressures were also measured
for the MSRE fuel solvent (composition: 64.7~
30.1-5.2 mole % LiFmBeFZ-Zerr) in the tempera-
ture range 960 to 1167°C. The data fit the equa-
tion log p(mm) = 8,803 - 9936/T(°K).

20 A, Bichler and J. 1. Stauffer, TAEA Symposium on
Thermodynamics, July 1855, paper SM-66/26.

1A, L. Mathews and €. ¥. Baes, Jr., ORNL-TM-
1129 (May 7, 1665),

Linear
ORNL-DWG 65-10170

 

 

 

 

 

 

 

 

 

 

 

400 o :
| 5
'E“‘_*Efi-hau:_‘::;_ ,,]7)_ — ii
bt |
200 - —f— et — — —
N ‘
| | \ |
10 - ‘\ —IT T T
o— N T T
r °
ol TN N ]
- o K**———-"-IX\HOO"C - ]
- L #Ai_.... e R
. 20 F - : — \. -
£ | *
: \
. .
10 | T 10000(;\3; i ———

 

 

 

]
!
t‘[
qs
-
JL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.5 |-
0.2 L ‘
o 20 40 c0 80 100
BeFy mole % LLiF

Fig. 2.3. Vapor Pressures in the LiF-BeF, System.

26

extrapolation of the data to 663°C, the highest
temperature of MSRE fuel salt at normal power
operation, yields a vapor pressure of 0,015 mm,

Transpiration Studies

Apparatus constructed
pressures by the carrier-gas method in order to
determine (1) vapor composition in the LiF-Bel
system (these measurements will complement the
manometric pressure data already obtained for
this system), and (2) rare-earth vapor conceatra-
tions in equilibrium with liquid mixtures of im~-
portance to the molten-salt reactor distillation
process (from these concentrations, more accurate
decontamination factors for the rare-earth fission
products will be obtained).

The apparatus, shown schematically in Fig. 2.4,
closely resembles that of Sense et al.?? To test
the reliability of the apparatus at clevated tempera~
tures and the procedures for removing condensed

was to obtain wvapor

vapors, several runs were carried out with LiF,
Satisfactory agreement with the reliable transpita=
tion data obtained by Sense?? attained if

argon flow rates were kept below 50 c¢m’/min and

was

if the condensers were washed for about 12 hr with
the solvent (a dilute solution of disodium Ver-
senate) at about 70°C.,

K A, Sense, M. J. Snyder, and J. W. Clegg, J.
Phys. Chem. 58, 22324 (1954,
23 A. Sense and R, W, Stone, J. Phys, Chem, 62,
1411 (1958).

 

ORNL—-DWG 65—13115

— MOLECULAR SIEVE DRYING COLUMN

— HEATED COPPER TURNINGS

\ >}
qj VENT
FLOW
METER /rPQESSURE GAGE (O-—
ARGON
e

 

@
[——
NICKEL J
{5-in. LONG

 

THERMOCOUPLE NO1
{LIQUID
SALT TEMPERATURE }—

HIGH TEMPERATURE ALUMINA TUBE

{36-in LONG)

101In Hzo)

  

 

 

 

THERMOCOUPLE NO. 3-
STAINLESS STEEL SLEEVE )
(5Vzin LONi; THERMOCOUPLE /
ND. 2.
\ y
M‘__f ] A : |
- 4 | MOLTEN
/ SALT —
S INSULATOR ™ /
/
f/
MARSHAL FURNACE (16-in LONG)-: (

   
 

ARGON TO DRYING

/ 2 COLUMN; THEN
LAVITE ./ S ON TO WET
TEST METER

. CONDENSER TUBE
(CONDENSER HAS 0.021-in. OPENING IN
END ABCVE CENTER OF SAMPLE BOAT)

Fig. 2.4. Transpiration Apparatus.
YISCOSITIES OF MOLTEN FLUORIDES

3. Cantor W, T. Ward

Li!':--BeF’2

Viscosity measurements of the LiF-BeF sys-
tem were extended to include three more mixtures
(36, 40, and 45 mole % BeF ). The viscosity of
pure BeF  was measured again, this time over a
410° temperature range. As previously reported,??
all measurements were carried out with the Brook-
field LLVT viscometer,

For the three mixtures, viscosity decreased with
decreasing Ber concentration. The degree of
decrease with concentration, however, was not
very large; below 50 mole % BeF  concentration,
melts are more typically ionic with relatively low
vizcosilies; the clusters of beryllium-fluorine
linkages that accounted for the very high viscos-
ities at higher BeF, concentrations are greatly
diminished when the concentration of BeF  is less
than 50 mole %.

The viscosity of pure BeF, was remeasured
to determine if pronounced deviation from Arrhenijus
behavior occurs. The plot of log n vs 1/T showed
only slight curvature; hence, BeF, is Arrhenius.
Usually the physical behavior of BeF  is analogous
to that of Si0_. Macedo and Litovitz?® postulated
that the Arrhenius behavior of 5i0, is due to the
constancy of density with temperature which, in
turn, means that the free volume of 510 , does not
change with temperature. To establish that den-
sity constancy with temperature is also why Bel
is Arrthenius will be difficult, because density
measurements of Bel”_  are a difficult experimental
task; nevertheless, we are currently considering
methods for measuring the density of pure BeF .

The data on the LiF-BeFZ system are summarized
in Table 2.4, The constants for the viscosity-
temperature equations were obtained by the method
of least squares.

NGBFA

To provide some information for heat transfer
calculations on the MSBR secondary coolant, pre-

 

243, Cantor and W. T. Ward, Reactor Chem. Div.
Ann. Progr. Rept. Jan., 31, 1965, ORNL.-3739, p. 81,

5p, B, Macedo and T. A, Litovitz, J. Chem. Phys.
42, 245 (1965),

27

liminary measurements were begun to determine
the viscosities of fluoroborates. The first measure=
ments were made on NaBF ; the results were:
7 + 2 centipoises at 466°C and 14 1 3 centipoises
at 436°C. The precision is poor because the
Brookfield viscometer is primarily designed to
measure higher viscosities,

No mixtures of alkali fluorides with NaBF | were
measured because such melts would most likely
have been even less viscous than pure NaBF ;
the poor precision in this lower viscosity range
discouraged us from attempting further measure-
ments with the instrumentation at hand.

ESTIMATING DENSITIES OF MOLTEN
FLUORIDE MIXTURES

S, Cantor

Several vears ago,2® the author, after examining
the published data on density of molten fluorides,
proposed that the simple rule of additivity of
molar volumes was very useful for estimating
densities of fluoride meits., To facilitate calcula-
tions, a table of empirical molar volumes was
derived from the published data. Since the earlier
report, there have been several experimental
studies of densities of fluoride melts. In this
repott, we reexamine the previous method of esti-
mation taking into account the newer measurements.

First, it appears that the rule of additivity of mo-
lar volumes held for all binary fluoride systems
except one; the one exception was the NaF-UF
system,?’ where positive deviations from addi-
tivity were as great as 6%; the rule held in two
studies of the L.iI*"»’I‘hF4 system?®:*? and in the
LiF-UF,,”® NaF-ThF,,*® and LiF-KF’? systems.

It was also found that the measured densities?®
of the eutectic compositions for the KF-NaF,

 

265. Cantor, Reactor Chem. Div. Aan. Progr. Rept,
Jan. 31, 1962, ORNL-3262, pp. 38—41.

27E, A, Brown and B. Porter, U/.S5. Bur. Mines Rept.
Invest, 6500 (1964).

28D. G. Hill and S, Cantor, Reactor Chem. Div. Ann.
Progr. Rept, Jan. 31, 1963, ORNI.-3417, pp. 4748,

e W, Mellors and S. Senderoff, **The Density and
Surface Tension of Molten Fluorides,* p. 578 in Pro-
ceedings of the First Australian Conference on Elace
trochemistry, Pergamcn, New York, 1964; also in
Union Carbide Research Summary 1JRS«70, Union Care
bide Corp., Parms, Ohio,
28

Table 2.4, Summary of Data and Constants for Viscosity-Temperature Equation

log 717 (centipoises) = A/TCK) - B for the System LiF-Ber

 

Composition Temperature Range

Viscosity at

 

 

(mole % BeF ) Measured (°C) 4 o 600°C (centipoises)
36,00 462-600 2,000 1.226 11.6
40.00 441637 2,203 1.367 14.3
45.00 419-638 2,589 1.685 19.1
50.00 376577 3,066 2.073 27%
55,01 389.-584 3,378 2,203 16%
60.00 437584 3,785 2.376 91°
65.00 451724 4,198 2.507 200
70.00 480704 4,695 2.695 480
75.00 490~705 5,362 3.036 1,275
79.99 558745 5,983 3.223 4,260
85.00 539.-747 6,551 3.340 14,600
90.02 564882 7,528 3.766 71,800
91.02 545.-832 7,640 3.702 112,000
93.01 572842 8,198 3,977 258,000
94,91 557837 8,604 4,099 569,000
96.01 601—844 8,928 4,218 1,016,000
97.00 601897 9,587 4.636 2,208,000°
98.01 632--917 10,359 5.179 4,840,000%
99.01 692—-967 11,358 5.816 12,560,000

100 7021112 See below” 63,800,000?
aExtrapoIated.

bThe equation for pure BeF2 is log # (centipoises) = 14,148/T — 18.345 — 3.382 log T.

NaF-LiF, NaF-LiF-CaF , and KF-NaF-LiF sys-
tems were all within 1% of densities estimated
by the additivity rule (using the molar volumes
given in Table 2.5). Two ternary systems,
NaF-LiF-ZrF43° and  KF-LiF-ZrtF ,?°  both
studied by the same investigators and by the
same method, were each consistently additive in
molar volumes. However, the molar volume of
ZrF4 that fits one system did not fit the other
system (e.g., at 800°C, the molar volume of ZrF‘4
in LiF-KF eutectic 3. LiF-NaF

was 55 cm in

 

3053, W. Mellors and S. Senderoff,
Soc. 111, 1355 (1964).

J. Electrochem,

cutectic, the molar volume was 45 cm’). For
density measurements carried out by Sturm®! on
four fluoride mixtures, the additivity estimate
differed from the experimental results by 5 to 6%
for three mixtures; the discrepancy between ex-
perimental and estimated density of fluoride mix-
tures seldom exceeds 3%. Although errors in the
molar volumes of 'E-L"rl:‘4 and KF may be responsible,
the reasons for these larger discrepancies are as
yet unresolved.

318. J. Sturm and R. E. Thoma, *f*Measurement of
Densities of Molten Salts,*®’ this report,
Thus, it appears that, although there may be
exceptions, the rule of additive molar volumes
describes the experimental data on molten fluorides
quite well and remains the simplest, most accurate
method for estimating densities of fluoride melts.

Table 2.5 gives a revised and enlarged set of
empirical molar volumes. The molar volumes of
AlF, were derived from the studies of Edwards
et al.®? on cryolite. Molar volumes of the alkaline-
earth fluorides (other than Ber), YFS, and the
rare-earth fluorides are based on measurements. of
the pure components carried out by Kirshenbaum
and co-workers.>%+3* The values for CsF were ob-
tained from Yaffe’s measurements.®® The molar
volumes of UF in Table 2.5 are based on measure-
ment of the pure liquid;®® these volumes, when
used additively with those listed for LiF and NaF,
provided good agreement between calculated and
measured?’ densities in the LiF-UF, and NaF-
UF, systems. The previous dual values for molar
volumes of UF, (see ref. 26), based on measure-
ments of mixtures,?”**® were probably in error
because much of the UF, was removed from solu-
tion by precipitation of UQ,.

{To estimate a density expression of the form

d=a-—bt, (10)

first solve for two densities by using the equation

Y )
i=1

d,
LWy ol

=1

(11)

where Ni and M, are the mole fraction and molecu-
lar weight of component v, and Vl,(t) is the molar
volume of component i at temperature t. Substitute
molar volumes from Table 2.5 at the two different

 

32]. D. Edwards et al., J. Electrochem. Soc. 100,
508 {1953).

33A. D. Kirshenbaum, J. A. Cahill, and C. S. Stokes,
J. Inorg. Nucl, Chem. 15, 297 (1960).

A, D. Kirshenbaum and J. A. Cahill, J. Chem. Eng.
Data 7, 98 (1962).

351. S. Yaffe, Chem. Div. Semiann.,
June 20, 1956, ORNL-2159, p. 79.

3% A, D. Kirshenbaum and J. AL Cahnill, J. Inorg. Nucl.
Chem. 19, 65 (1961).

37TR, C. Blanke ef al., MLM-1076 (April 1956).
3813 . Blanke of al., MLM-1086 (December 1956).

Progr. Rept.

Toble 2.5, Empirical Molar Volumes of Fluorides

 

Molar Volume (cm3/mole)

 

 

At 600°C At 800°C

LiF 13.46 14.19
NaF 19.08 20,20
KF 28.1 30.0
RbF 33.9 36.1
CsF 40,2 43.1
BeF 23.6 24,4
Mgk 22.4 23.3
CaF 27.5 28.3
StF 30.4 31.6
BaF 35.8 37.3
ALF 26.9 30.7
YF, 34.6 35.5
Lak 37.7 38.7
CeF 36.3 37.6
PrF 36.6 37.6
SmF 39.0 39.8
ZrF, 47 50
Th¥, 46.6 47.7
UF, 45.5 46,7

 

temperatures to obtain the two values of d,; since
density is linear with temperature, substitution of
the two values of dt in Eq. (10) provides the solu-
tion for the constants a and b.}

ESTIMATING SPECIFIC HEATS AND THERMAL
CONDUCTIVITIES OF FUSED FLUORIDES

8. Cantor

Specific Heats

The simplest rule for estimating high-tempera-
ture heat capacities in the condensed states is
that of Dulong and Petit, in which the heat
capacity per grameatom is approximately equal to
30

6 cal/°K. Accordingly, the experimental heat
capacities of molten fluorides were examined in
order to modify the Dulonge«Petit value for molten
fluorides. The data for pure compounds (see Table
2.6) indicate that the heat capacity per gram-atom
is approximately 8 cal/°K. The data for fluoride
mixtures would seem to indicate the average value
to be somewhat higher than 8, However, the heat
capacities of the mixtures are probably uncertain
by 10 to 20%, whereas the experimental uncertainty
for most of the pure compounds is less than 5%.
Hence, 8 cal (°K)™! (gram-atom)™ ! is the value
chosen for estimating specific heats of fluoride
melts,

The temperature variation of C_ in liquids, when
determined accurately, is very small; whether this
variation is negative or positive has not been
adequately established. It is, therefore, prudent
to assume that C is temperature independent.

[The specific heat is estimated from the follow~
ing expression:

8Y (Wp)
cm T (12)
2: (NiMi)

where ¢ is the specific heat, p. is the number of
atoms in a molecule of component 7, and Ni and
Mi are the mole fraction and the molecular weight,
respectively, of component 1,

Sample calculation: For MSRE coolant, 66-34
mole % LiF-Ber,

T p,) = 0.66(2) + 0.34(3) = 2.34
(VM )~ 0.66(26) + 0.34(47) = 33.1,

8(2.34)
- —— = 0.57 cal (°K)" ! g7 1.]
© 7 33 CK)” e

Thermol Conductivities
At present, accurate measurement of thermal
conductivities of fused fluorides is very difficult;
hence, reliable methods for estimating thermal
conductivities are extremely useful, especially if

one wishes to predict heat transfer rates in

circulating systems. Gambill?® has published an
empirical method for estimating thermal conductivi-
ties which agrees with the available data. This
report will describe a new semitheorefical method
based on Bridgman’s theoty of energy transport
in liquids.*®

Bridgman proposed that energy is transferred by
collision from one molecular layer to the next at
a rate equal to the local sonic velocity and that
the distance traveled between collisions is equal
to some characteristic molecular distance, DBridg-
man obtained the equation for A, the thermal con-
ductivity,

A, (13)

where k is Boltzmann’s constant, g is the velocity
of sound, and A is characteristic molecular dis
tance between collisions, To evaluate A, Bridg-
man substituted the average distance between
molecular centers, (V/N)!/3 (V is the molar
volume and ¥ is Avogadro’s number), thus deriving

the expression
A= 3k N) v
= V "J_ .

Equation (14) is in good agreement with the ex=
perimental data on covalent compounds.

(14)

However,
this equation predicts low thermal conductivities
for molten fluorides because the characteristic
collision distance is too large. The collision
distance for molten salts should be less because
the ions (the ‘“‘molecular’’ entities in the liquid)
are much less compressible than covalent mole~
cules, Rather than attempt to estimate collision
distances in fused salts a priori,*! Eq. (14) was
empirically adjusted to fit the available data, the
equation obtained being:

A 4431<<N)2/3
= (4.4) v o

(15)

39W. R. Gambill, Chem. Eng. 66(14), 129 (1959).

0p w, Bridgman, Proc. Am. Acad. Arts Sci. 39,
162 (1923).

41Perhaps the collision distance may be derived from
either the molecular free velume or the compressibility.
31

Table 2.6. Heat Copacities of Molten Fluorides

 

 

 

 

 

 

A. Pure Compounds
Compound Cp per Mole Cp per Atom Reference
ik 15,50 7.75
NaF 16.40 8,20 a
KF 16.00 8.00
13eF2
825" K 18.58 6.19 b
1200°K 22,70 7.57 b
MgF2 22.60 7.53 a
Can 23.90 7.97 a
UF4 40 8 c
Na3A1F6 93.4 9,34 a
B, Mixtures
Composition (mole %)
C per Atom Reference
NaF LiF KF BeF ZF,  ThF,  UF, ?
65 35 7.82 d
61 39 7.79 d
57 43 7.91 d
53 47 7.50 d
50 50 8.22 d
50 46 8.15
50 25 25 9.15
65 15 20 7.52 &
53 43 4 8.33 2
56 39 B 8.12 e
53.5 40 6.5 7.60 e
76 12 12 9.86 e
25 60 15 8.86 e
53 46 1 7.95 8
11.5 46,5 42 9.35 e
50 30 89.24 h
46.5 26 27.5 9.85 &
10,9 44.5 43.5 1.1 9,64 e
11.2 45.3 41 2.5 9.26 i
38.4 57.6 4 10.93 e
48 48 4 9,54 h
20 55 21 4 8.36 h
4,8 50.1 41.3 3.8 9.26 e
53 46 1 7.75 é
62 37 1 7.99 J
70 10 20 7.74 k
62 36,5 1 0.5 7.69 k
67 18.5 14 0.5 8.89 1
71 16 13 7.80 k
67 i8 15 8.77 m
32

Table 2.6 (continued)

Composition (mole %)

 

 

 

C per Atom Reference
NaF LiF KF BeF ZrF ThF UF P
2 4 4 4
70 23 5 1 1 8.06 n
68 32 8.32 o
K. K. Kelley, ‘*High~Temperature Heat-Content, Heat-Capacity, and Entropy Data for the Elements and In=-

organic Compounds,’* U.S. Bur.
R. Taylor and T. E. Gardner, U.S. Bur. Mines Rept. Invest. 6664 (1965).

G. King and A, U, Christensen, U.S. Bur. Mines Rept. Invest. 5709 (1961).
D. Powers, ANP Quart, Progr. Rept. Mar. 10, 1956, ORNIL.~2061, pp. 176-78.

b,
°E.
dy,

Mines Bull. 584 (1960).

*W. D, Powers and G. C. Blalock, Enthalpies and Heat Capacities of Solid and Molten Fluoride Mixtures, ORNL-

1856 (Feb, 1, 1956).

'W. D. Powers, ANP Quart. Progr. Rept. Mar, 31, 1957, ORNI.~2277, p. 87.
€w, D, Powers, MSR Program Quart. Progr. Rept. Oct. 31, 1958, ORN1.=-2626, pp. 4445,
hg, 1 Cohen, W. D, Powers, and N, D. Greene, A4 Physical Property Summary for ANP Fluoride Mixtures, ORNL~

2150 (Aug. 23, 1956).

'w. D, Powers, ANP Quart. Frogr. Rept. June 30, 1957, ORNL-2340, p. 103.
jW. D. Powers, MSR Program Quart. Progr. Rept. June 30, 1958, ORNL-2551, p. 31.

kusr Program Quart. Progr
IMSR Program Quart. Progr
“IMSR Program Quart, Progr
"MSR Program Quart. Progr

. Rept. Apr. 30, 1959, ORNT1.-2723, p. 38.
. Rept. Apr. 30, 1960, ORNL-2973, p. 23

. Rept, Oct. 31, 1959, ORNL-2890, p. 21,
. Rept. Feb, 21, 1961, ORNL.-3122, p. 139.

°MSR Program Semiann. Progr. Rept. Aug. 31, 1961, ORNL-3215, p. 132.

The agreement between Eq. (15) and the experi-
mental data is shown in Table 2.7.

The velocity of sound was itself estimated from
the equation:

2

pe o= [_(f\p/cv) -1 Cp

[ (16
a*TH )

where C_and C_ are the molar heat capacities at
constant pressure and volume, respectively, « is
the expansivity, T is the absolute temperature, and
M is the molecular weight. The ratio C /CV
varies between 1.2 and 1.3 for most fused salts

(1.2 was chosen for making the estimation), Cp,

as indicated above, is very close to 8 cal (“°K)™!
(gram-atom)” ', The expansivity is the negative of
the temperature coefficient of density divided by
the density, that is, a = ~(1/p) (<3p,/c9T)p. By using
the additivity rule of molar volume (see ‘‘Estima~
ting Densities of Molten Fluoride Mixtures,’’ this
report), o is easily estimated.

Table 2.7. Therma!l Conductivities in Fluoride Melts

 

Al Btu fe 7 et CF)™ 1]
Melt (mole %) -

 

 

Experimental® Calculated

RbF-ZrF4-UF4 1.0 0.94
(48-48~1)

LiF-RbF 1.2 1.39
(43-57)

Na¥-KF-[iF 2.6 2.10
(11.5~42-46,5)

NaF»KF-LiF-UF4 2.3 2,01
(10.9-43,5-44.5-1,1)

NaF-ZrF4~UF4 1.3 1.24
(50-46-1)

NaF-ZrF4 -U¥F, 1.2 1,31
(53.5<40-6.5)

NaF-BeF2 2.4 3.20
(57-43)

 

%Originally regarded as good to 125%.
33

Finally, combining Egs. (15) and (16) and evaluat-
ing constants, where possible, one ends up with
a simplified equation for themal conductivity:

3.78 « 10° -

)x[ergs sec” 1 cm 1 (OK)“ 1] = --7{}“2-73:;— ':‘;: » (17)

where ¢ is the specific heat [cal CK)" ! g7, T
is tempetature (VK), ¥V is molar volume (cm3), and
2 is expansivity (“K) L,

SOLUBILITY OF DF AND HF IN LiF-BeF2
(66-34 MOL.E %)
P. £, Field*? J. H. Shaffer

The solubilities of HF and DF in the molten
mixture LiF»BeF2 (66-34 mole %) were determined
the temperature range 500 to 700°C and
at gas saturation pressures between 1 and 2 atm.
An extrapolafion of the solubility values provides
an approximation of the solubility of tritium
fluoride in the molten fluoride mixture., The be-
havior of tritium fluoride, formed by neutron irradize-

over

tion of lithium, in the fluoride mixture would be of
interest in the molten-salt reactor concept as well
as in the proposed uge of a molten fluoride breeder
blanket for a thermonuclear reactor.*?

The experimental procedure has been previously
described and employed for systematic studies of
gas solubilities in molten fluoride miztures.*?
Anhydrous hydrogen fluoride was obtained from a
commercial source and was used without further
purification.  Anhydrous deunterium fluoride was
prepared by the Technical Division, ORGDP, by
reaction of elemental deuterium and fluorine.*®

 

2 summer ORINS Researcn Participant with Reactor

Chemistry Division, 1965 Assistant Professor of
Chemisiry, Virginia Polytechnic Institute.

430, 3. Rose and M. Clark, Jr., Plasmas and Con-
trolled Fusion, p. 296, M.L.°T, Press, Cambridge, Mass.
1661, :

44j. H. Shaffer, W. R. Grimes, and G, M. Watson,
J. Phys., Chem. 63, 1999 (1959).

458. T, Benton, R. L. Farrar, Jr., and K. M. McGili,
Preparation of Anhydrous Deuterium Fluoride by
Direct Combination of the Elements, K«1585 (Jan. 29,
1964).

ORNL—-OWG 65--2420R

 

 

 

(8]

 

NOTE:
(KH} TAKEN FROM LEAST

YALUES OF SLOPES

GAS SOLUBILITY (moles solute/mole soivent ¥ 10%)

 

 

 

 

/e i T SQUARE FIT OF EXPERI - -
MENTAL DATA OF 1 A
vs Yy
O T et i mm m mn wa
0 1 2 3

SATURATION PRESSURE (atm)

Fig. 2.5. Henry's Law Plot of HF and DF Solubilities
in LiF-BeF, (66-34 mole %}

As shown by Fig. 2.5, the solubilities of the two
gases obeyed Hemry's law, within the experi-
mental precision, over the ranges of temperature
and pressure that were studied. Values obtained
for Henry’s law constants expressed as moles of
HF pec mole of melt per atmosphere at 500, 600,
and 700°C were 3.37 x 107%, 2.16 x 107*, and
1.51 x 10™* respectively. Corresponding values
for the solubility of DF were 2.97 x 1077%, 1,83 «
107%, and 1.25 x 107% at 500, 600, and 700°C
respectively,

Differences in the solubilities of DF and HF are
outside the 95% confidence level attributed fo the
experimental data. Evaluation of the temperature
dependence of the Henry’s law constants by a
least-squares {it of the data indicated that, within
experimental precigsion, the enthalpies of solution
of the two gases are equal to about —6.0 keal/mole
for HF and -6.4 kcal/mole for DF.
3. Chemical Separation

IN-PILE MOLTENSALY IRRADIATION
ASSEMBLY

H. C. Savage E. L. Comperse
M. J. Kelly J. M. Baker
. G. Bohlmann

Development of an in-pile molten-salt irradiation
assembly! to provide supporting information for
an understanding of both short-term and long-term
effects of irradiation and fissioning on fuels and
materials for molten-salt reactors has continued
during the past year. The irradiation experi-
ments are to be conducted in beam hole HN-1
of the ORR, and it is anticipated that in-pile
operation of the first experiment will begin in
the middle of calendar year 1966.

An autoclave-type (capsule) experiment with
thermally induced salt flow has been designed,
and some 4500 hr of operation have been accu-
mulated in mockup tests with two prototype as-
semblies. Test operation has been at a nominal
salt temperature of 1200°F with a salt mixture
whose composition is LiF-BeF -ZtF -UF, (65-
29-5.1 mole %, liquidus temperature ™~ 840°F).
Salt circulation rates of ~10 c¢cm®/min are achieved
by maintaining a median temperature gradient
of 100 to 180°F between the autoclave and the
cold leg. The flow rate is monitored by heat
balance measurements around the cold leg.

Results of these mockup tests indicate that
the present design of the autoclave is suitable
for use in a molten-salt irradiation program with
the irradiation objectives of (1) 200 w/cm? fuel
fission power and (2) up to 50% 23°U burnup and
long-term in-pile operation.

 

lReactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNIL-3789, pp. 45—48.

and lrradiation Behavior

34

Earlier models of the autoclave had a horizontal
return line connecting the cold leg with the main
autoclave.? Test operation with helium cover gas
revealed that gas buildup occurred in this return
tine and caused a loss of salt flow after about
16 hr. Flow recovery after evacuation and re-
pressurization but
since continuous salt flow in the experiment is
gsirable, the autoclave was redesigned to elim-
inate the horizontal return line (Fig. 3.1). The
prototype model was modified by removing all
but about 1 in. of the 5-in.-long horizontal line.
This modification increased the operating time

was invariably successful,

without loss of flow to several hundred hours.
Complete elimination the horizontal return
line in the autoclave assembly to be operated
in-pile is expected to correct the flow loss from
gas accumulation. However, test operation of
the first in-pile experiment will be carried out
in the mockup facility to demonstrate satisfactory

of

performance prior to in-pile operation.

Experiment facilities associated with beam hole
HN-1 of the ORR are being modified for the
molten-salt experiment. Instrument and control
previously used to operate in-pile cor-
rosion test loops,® are being revised for the
higher operating temperatures and lower pressure
requirements of the molten-salt experiment.

Necessary auxiliary equipment now being de-

panels,

(1) a new alu-
minum liner for beam hole HN-1, (2) a radiation-
shield plug which is part of the experiment pack-
age, and (3) revisions to an existing equipment
chamber, located at the face of the reactor shield-
ing, which will contain tanks and valves necessary

signed and constructed includes:

 

2Reactor Chem. Div. Ann. Progr. Rept., Jan. 31, 1965,
ORNL-3789, Fig. 2.3, p. 46.

3H. C. Savage, G. H. Jenks, and E. G. Bohlmann,
In~Pite Corrosion Test Loops fot Aqueous Homogeneous
Reactor Solutions, ORNL-2977 (Nov, 10, 1960).
1

SALT RESBERVOI

 
 

OGRNL-DWG 86--965

 

 
   
  
   
  

GAS SAMPLE LINE—"

THERMOCOUPLE //
WELL —f—-

~PRESSURE MONITOR LINME
— BALT FLOW PASSAGE

/ e THERMOCOUPLE WELL

 

EENSSN —GRAPHITE CORE

 

 

 

 

 
 
 
 

o
HEATER €

HEATER---

COOLING COIlL-—

AUTOCLAVE “‘\1;

 

 
 
   

il ‘-"-u % ,f -5

%;_'\‘ ’\1
HEATER -~ I \f-"'.’c;;

0

AN @

 

 

 

  
   
 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 3.1. In-Pile Molten-Salt Convection Loop.

to remove salt samples and add makeup salt to
the autoclave during in-pile operation,

After in-pile operation, the experiment package,
consisting of the molten-salt autoclave assembly
in its container, the shield plug, and the inter-
connecting lines, will be removed from the beam
hole into a shielded carrier and transported to
hot-cell facilities to be cut up and examined.

EVAPORATIVE-DISTILLATION STUDIES
ON MOLTEN-S3ALT FUEL COMPONENTS

M. J. Kelly

Vacuum distillation separation of molten-salt
fuel or fuel components from the rare-earth fission

products is an attractive method of decreasing
neutron losses by capture.  To design process
equipment for this task, both the mass rate of
distillation and the relative volatility of the rare
earths must be known for the particular salt
system used. Completed experiments concermn
the planned process demonstration for MSRE
fuel but are also directly applicable to any pro-
posed thermal MSBR fuel.

For MSRE fuel, removal of the uranium by fluo-
rination is proposed. The fuel solvent and re-
maining fission products would then be fed at
the distillation rate to a wvacuum still charped
with LiF-BeF -ZiFF, at that composition which
will yield the fuel solvent as distilled product;
the rare-earth fission products would concentrate
in the still. The residue would be discarded or
processed when necessitated by heat from the
fission products or by carry over of rare earths.
A 10-ml graphite cylinder, containing ~17 g
of salt with a free surface (when molten) of 1
cm?, was used for the still pot. This cylinder
fitted in a “[1’’ shaped INOR-8 tube heated elec-
trically to a given temperature as measured by
four thermocouples in the graphite cylinder. When
the desired temperature was reached, distillation
wasg initiated by evacuating the assembly. Vapor-
ized salt then passed up and over the top of
the *“N1’? with a small portion (attributed to thermal
reflux) collecting at the base of the graphite
cylinder. The product salt condensed in a cooler
(~450°C) collecting cup in the opposite leg.
Distillation was stopped by helium addition after
preselected time periods.

Mass-rate data for ‘Lil® were determined first,
and then MSRE solvent was added to the 7LiF
and distilled in aliquot portions from it. Each
repetition brought the BeF,  and ZrF, conceu-
trations in the pot closer to those which would
yield fuel solvent (LiF-BelF -Zr¥ ; 65-30-5 mole
%) as product. After the equilibrium concentra-
tion was approached, 2200 ppm of neodymium
(as NdF ) was added to the still bottom and
several distillate Then
the neodymium concentration was raised to 22,000
ppm, and the sequence was repeated. For both

samples were taken.

(g/cmz- sec)

=
e

 

DISTILLATION RA”

 

Fig. 3.2.

Ohbserved Distillation Rate vs Temperature.

 

36

"LiF and the
the effect of temperature on mass rate was deter-
mined.

The data of Fig. 3.2 show the effect of tempera-
ture and mass rate; single experiments are con-
siderably scattered, so bands are
The slope of the bands is consistent
with the heat of vaporization of the components
and strongly suggests that ebullition does not
occur. Solvent surface heat flux is <1/100 of
the available heat to the graphite cylinder; it
is doubtful that the distillation is heat limited.
The fact that the observed rate for ‘LiF is only

nominal equilibiium composition,

inclusive
shown,

~10% of theoretical is unexplained, but it has
been observed in many distillations using several
experimental configurations.

The solvent for any proposed MSBR should
distill at rates above that shown for LiF with
the possible exception of ThF  -bearing systems.
The effect of ThY", is being studied.

Effectiveness of separation from rare earths
was determined by activation analysis for nco-
dymium in the product, the still bottom, and the
refluxed salt deposited around the base of the
graphite. The data are shown in Table 3.1.

The equilibtium composition in the still pot
undoubtedly changes with temperature due to the
change in activity coefficients of the components.
The composition found
after several solvent additions of 5 to 7% by
weight and after distillations at 1030°C was
LiF-Bel" -ZrF, (85.4-10.7-3.9) in the presence
of 22,000 ppm of neodymium as fluoride.

at nominal equilibrium

 

Table 3.1. Concentration of Needymium in Fractions
from Vacuum Distillations
Nd Concentration®
Fraction (ppm)

 

2200 ppm Added 22,000 ppm Added

 

Still bottom 2570 22,000
Product 21 600"
Reflux 133 34

 

“Mean values from several determinations; data show
little scatter except for reflux specimens.

bThese product samples also showed cerium and
lanthanum, which undoubtedly represent contamination
during grinding and handling of specimens; neodymium
analysis, therefore, may well be too high.
EFFECTIVE ACTIVITY COEFFICIENTS

BY EVAPORATIVE DISTILLATION
OF MOLTEN SALTS

M. J. Kelly

Evaporation into a vacuum from a quiescent
molten-salt mixture of low vapor pressure should
not permit wvapor-liquid equilibrium at the sur-
face of the melt; transport from the surface should
be controlled by the evaporation rates of the
individual components.  The amount vaporized
is a function of the equilibrium vapor pressure,
molecular mass, temperature, and surface area;
according to Langmuir:*

grams of Z
L KPZWZ/T , (1)

z cm”™ sSecC

where P, is the vapor pressure of component Z,
M is molecular weight of component Z, T is
absolute temperature, and K is a constant which
is dependent on units.

In a multicomponent system at equilibrium,

P_=y,N, P, (2)
and

grams of Z .....

W, = — 5 =Ky, N, PSVM,/T. (3

If we use a {ixed mechanical configuration and,
for simplicity, define the activity coefficient of
the major component (in this case LiF) as unity,
the activity coefficient for any other constituent
may be determined using the ratio

 

_W/F . KyZNZP%\/M /T @
WL-i,F K”)N[ i L1FV W n«‘/T
altered to
v, = NLiF . Wz . PEiF ] MLiF (5)
z = : ] 5
z Nz WLW P; v M,

Proceeding with this technigque, the data shown
in Fig. 3.3 have been taken from LiF-BelF  and

 

1. Langmuir, Phys. Rev. 2, (Ser. 2), 329 (1913} (and
subsequent papers).

37

Table 3.2, Vapor Pressures for Pure Fluorides

 

Vapor Pressure

 

 

Compound (mm He)
1000°C 900°C
LiF 0.47 0.072
HeF , 65 12.0
ZrF4 2700 780

 

LiF-BeF -ZrF | melts. For comparison purposes,
selected data from other sources are included,®~7

Although internal consistency exists, the valnes
are dependent upon vapor-pressure data for the
pure components. Calculations were made using
the values (P°) for the pure components shown
in Table 3.2.8—1°

The least-precise measurements are from MSRE-
solvent distillations; these are included to sub-
stantiate the surmise that, when 2‘,'rF4 is included
in the melt, it effectively removes LiF from the
solvent, The 1000°C LiF-BeF, line has been
extended to 100 mole % LiF since the initial
LiF-BeF -ZrF, experiment contained <0.35 mole
% ZrF,, a quantity so small that the system
can he assumed to be LiF-BeF . On the other
hand, as ZrF, builds up in the mixed system
(to approximately 3 mole %), enhancement of
the Bel activity is noted; this is consistent
with results from MSRE solvent (LiF-BeF ,~LrF
65-30-5 mole % initially). The temperature% re-
ported are those of the bulk melt, and it is rec-
ognized that the surface temperature may be
significantly lower, causing an indeterminate (for
the present) error. Both surface-temperature
effects and the pure LiF-ZrF system are being
studied., It is interesting to note that the as-
sumption of unit activity coeffictent for LiF doeg
not lead to incompatibility with the data of others.

 

*Reactor Chem. Div. Ann. Progr, Rept., Jan., 31, 1965,
ORNIL.-3789, pp. 5952,

6. A. Sense and R. W. Stone, J. Phys. Chem. 62,
96 (1958).

“A. Buchler and J. L. Stauffer, Symposium on Thermo-
dynamics with Emphasis on Nuclear Materials and
Atomic Transport in Solids, Vieana, July 2227, 19653,
paper SM-6b-20, p. 15.

SHandbook of Chemistry and Physics, 44th ed., p.
2438, Chemical Rubber Publishing Company.

9B. Porter and E. A. Brown, J. Am. Ceram. Scc. 45,
49 (1662).

10 . .
8. Canter, personal communication,
38

ORNL-DOWG 66-967

 

   

 

 

 

 

 

 

 

 

ICIENT

-

 

 

 

 

 

 

 

 

ACTIViTY COEF:

 

 

 

 

 

o b
1073 e _J _‘I,
[ —-—-LiF - Bef,, BAES, 700°C ..
- LiF-ZrF,, SENSE efal |~
g| ——2 LiF-BeF, + ZrF,, BAES
 ———DERIVED FROM MSRE SOLVENT
DISTILL ATIONS 300-050°C
- © LiF-Bef,, BUCHLER, 600°C
®  LiF-BeF,, 1000°C |
27 o LiF-BeF,, 300°C T
p,8 Lif+Bef,-7rF,, 1000°C
10_4 \__----.,...._....J.._....._..._...._..L_,h,_____....“,*_
65 70 75 80

 

 

 

e

 

 

 

 

85 20 35 100

LiF IN MELT (imole %)

Fig. 3.3.

Compositions.

REMOVAL OF {ODIDE FROM LiF-BeF, MELTS

B. F. Freasier'? C. F. Baes, ]Jr.

H. H. Stone
The removal of the 6.7-hr fission product 13°I
from a molten-salt fuel would reduce the amount

11ORINS Summer Participant, 1965, from Louisiana
Polytechnic Institute, Ruston.

Effective Activity Coefficients Calculated from Evaporative-Distillation Data for M5BR-Solvent

of daughter '®°Xe, a major fission product poison,
formed in the fuel. In measurements not yet
completed,'? it has been found that iodide (the
chemical form of iodine expected to be present)
can be readily removed from LiF-BeF, melts

 

12MSR Program Semiann. Progr. Rept. Aug. 31, 1965,
ORNL-3872, p. 127, MSR Program Semniann. Progr, Rept.
Feb, 28, 1965, ORNL-3812, p. 137,
by sparging with mixtures of HF and H
by the reaction

,» evidently

HF(g) + I7(d) == F~(d) + HI{g) ,

Q= PHI/PHF[I“] kg/mole . (6)

The experiments were performed in all-nickel
vessels which did not react with HI in the pres-
ence of H, at the temperatures studied (474 to
635°C). Such reaction was avoided at the lower
temperatures of the gas exit system by use of
gold-lined nickel and Teflon tubing. The iodine
was added initially as Nal; tests were made of
the effluent gas, and HI, but no elemental io-
dine, was found. The HI was trapped in an
Nz«OH scrubbing solution and determined by stand-
ard iodometric methods. Only about 80% of the
initially added iodide appeared in the NaOH scrub-
ber; it is believed that the remaining iodide
escaped as a result of adsorption of HI on par-
ticulate matter which was not effectively trapped
by the scrubber. When radicactive iodine was

used in tracer experiments and a filter was
placed in the effluent gas stream just down-
stream of the reaction vessel, all the iodine

passing the filter was caught in the NaOH scrub-
ber soclution and appeared as iodide ion. The
discrepancy in material balance will be iovesti-
gated further, but it does not seem likely that
the present estimates of @ for the reaction will
be substantially altered.

Results were consistent with the following in-
tegrated rate equation which may be derived from
reaction (b):

In ((I-1/01719 = ~Qu /W) . (7)
The terms [I171° and [I7] are, respectively, the
concentrations of iodide present initially and
of iodide present after n_ . moles of HF have
been passed through W kg of melt. This rela-
tion was used to determine the equilibrium quo-
tient, @, for reaction (6). The value of ¢ so
obtained was found not to be significantly de-
pendent on the HF flow rate (0.35 to 1.57 milli-
moles min™?! kg™!'), on the partial pressure of
HF (0.02 to 0.1 atm), ot on the initial iodide
concentration (0.004 to 0.04 mole/kg).  This
indicated that reaction (0) was indeed the only
one of significance, that eguilibrium sparging
condifions were obtained, and that the activity

39

coefficient of the iodide ion did not wvary ap-
preciably over the concentration range of iodide
employed.
For 2LiF-BeF, melts (Fig. 3.4) the presently
incomplete results give
log Q = —1.094 + 2.079(10°%/T) (8)

with error limits of perhaps £15% in Q. Thus,

"Q increases ~ the ease of jodide removal by HF

increases - with decreasing temperature. With
increasing BeF, content of the solvent, Q ap-
pears to go through a maximum.

Mole % BeF2
O 482°C)

33.3
51

41.5
71

49,7
32

The number of moles of HF necessary to remove
half the iodide present in 1 kg of melt by equi-
librium sparging is, by Eq. (7), simply 0.693/(.
For a reactor system in which a side stream of
the fuel is continuailly treated with HF, the mini-
mum amount of HF passed through the fuel per

ORNL~DWG 65-9871
TEMPERATURE (°C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

550 £00 550
8O pop T [
R Py
Yo R O e
a0 e
°
- 3¢ - ", T
T f, ‘
—(—"; ® |
g b AN D5 kip] ]
{ |
= yd , |
= . | \
o 1 !
& 20 . .._.._........; ............. b
i
[ R R D]
\ ?
\ f
| |
| |
| |
P e ]
1.07 1.44 145 1.19 1.23 1.27 1,34 1.35
1000,
7P
Fig. 3.4, VYariation with Temperature of the Equilib-

rium Quotient for lodide Removal $rom QLEF-Ber,
40

hour to give a specified overall removal half-
time (tl /2) is

moles of HF /hour > (0.693/1‘1/2)WT/Q )
where WT is the total weight of the fuel. Since
the half-time for decay of !3° to !*°Xe is 6.7

hr, iodide-removal half-times of the order of an
hour might be desired. With Q = 40 kg/mole
(2LiF-BeF, at 500°C), half the iodide present
in a reactor fuel could be removed in 1 hr by
the passage of a minimum of 0.0173 mole of HF
(388 std cm?) per hour per kilogram of fuel.

REMOVAL OF RARE EARTHS FROM MOLTEN
FLUQGRIDES BY EXTRACTION
INTD MOLTEN METALS

J. H. Shaffer W. K. R. Finnell
F. A. Doss W. P. Teichert
W. R. Grimes

In a two-region molten-salt breeder reactor, the
fuel mixture will require routine reprocessing to
reduce concentrations of those fission products
which have high neutron-capture cross sections.
Of the various fission products which form sta-
ble chemical compounds in the fluoride fuel
rare earths will comprise the major
poison fraction, The extraction of selected
rare earths from solution in a molten fluoride
mixture into immiscible molten metals is being
studied as a possible chemical reprocessing
Experiments conducted thus far have
cerium,

mixture,

me thod.
examined the extraction of lanthanum,
neodymium, and europium from the fluoride sol-
LiF-BeF, (66-34 mole %), into bismuth
metal. This fluoride mixture simulates the fuel

solvent proposed for a molten-salt breeder reac-

vent,

tor; UK, is here presumed to have been removed
from the fuel by fluorination.

Fluoride mixtures having rare-earth concentra-
tions of about 10™% m.f. together with their ap-
propriate radioisotopes (for analytical purposes)
were prepared in nickel vessels. The mixtures
were further treated at 600°C with anhydrous
HF and H,, according to established fluoride
purification techniques, until the gamma activity
in two or more consecutive filtrate samples of

Table 3.3. Distribution of Rare Earths
Between LiF-BeF, (66-34 Mcle %)
and Molten Bismuth when Reduced

with Beryllium Metal at 600°C

 

 

Rare Earth Rare Earth
Remaining Dissolved
Rare Earth in Salt Phase in Metal Phase
(%) (%)
Lanthanum 0 41
Cerium 0.1 90
Neodymium 1 49
Europium 3 90

 

the salt mixture became constant and approached
anticipated values. The liquid-metal extractants
were prepared in stainless steel extraction ves-
sels with low-carbon steel liners; pretreatment
with hydrogen at 600°C reduced oxide impurities
in the liquid metals. An extraction experiment
was started by transferring a portion of the pre-
into the Each
extraction experiment involved about 2.35 kg
of molten bismuth and 1 to 2 kg of salt. The
two phases were agitated by sparging helium
through a tube that extended into the metal phase.

Two types of extraction experiments have been
conducted. In the initial experiments, beryllium
metal was added as machined turnings to the
extraction vessel during preparation of the molten
metal. Following introduction of the salt mix-
ture, filtered samples of each phase were taken

pared salt extraction vessel.

at periodic intervals and analyzed radiochemically
for their respective rare-earth content. The dis-
tributions of rare earths at the conclusion of
each summarized in Table 3.3,
These results demonstrate the effective removal
of rare earths from the salt phase.

experiment are

The incom-
plete dissolution of rare earths in the metal
phase may indicate that a third, seolid, phase
was formed; it is not unlikely that this insoluble
phase rare-earth beryllide such as have
been observed by others at this Laboratory,?!3

Spectrographic analyses of samples taken from
the metal phase showed the presence of dissolved

is a

 

13 -
M. E. Whatley, private communication.
lithium in the molten bismuth, suggesting that
the beryllium metal had caused reduction of part
of the lithium. In two additional experiments,
extractions of cerium and neodymium were ac-
complished by adding lithium metal directly to
the molten bismuth. Samples of the salt and
metal phases were withdrawn under assumed equi-
librium conditions after each addition of lithium
metal. The distribution of cerium and neodymium
between the two phases at the conclusion of
the experiments essentially duplicated that found
in the beryllium-reduction experiments.
centration of lithium in the metal phase of each
experiment increased linearly with the quantity
of lithium added; however, only 1/4 to 1/3 of
the added lithium appeared in the metal phase,
and it may be that a reduction of beryllium ion
to its elemental form accounted for the missing
lithium.  Furthermore, as shown in Fig. 3.5,
the concentration of rare earths in the metal
phase in each experiment became independent of
the lithium concentration found in the metal phase.

Subsequent extraction experiments will be ex-
tended to include other rare earths and to c;tudy
these inferred reaction equilibria.

The coa-

ORNL-DWGE 66-988

 

 

 

 

 

 

 

 

 

 

 

 

 

° } | t T
‘OO% (,e H&MD\:’AL FR’OM SAL
——— Pt
w ;
4 ol
?EI 4 feeeeeeeeees R T —
o
=2 o CERIUM 1
oo
w
F = . e b
zs ° . |
Qb -
Z oo 100% Nd REMOVAL FRUM S[\LI
g5 ey — e e D T e .
L. g 2o s — j e e
r £ |
?_._‘ —
i
<
1l o
L) ] OETIRNTTIQ] O e
c AT
/ |
. ; NECGCTYMIUM
VoM
0 | | _
G 0.2 1.0 1.5 2.0 2.5
LITHIUM FOURND IN METAL PHASE
{ mole fraction x 107)
Fig. 3.5. Extraction of Cerium and Neodymium from

LiF'Ber {66-34 Mole %) into Molten Bismuth by Addi-
tion of Lithium Metal at 8009 C.

41

REMOVAL OF PROTACTINIUM FROM MOLTEN
FLUORIDES BY OXIDE PRECIPITATION

J. H. Shaffer
. A. Doss

W. K. E. Finnell
W. P, Teichert
W. R. Grimes

Since a single-region molten-salt breeder reactor
would incorporate the fertile material in the reac-
tor fuel mixture, chemical reprocessing schemes
for recovering 23U could be made more effective
if “%Pa, its precursor, could be removed with-
alteration of the relatively large uranium
concentration in the fuel mixiure. The precipi-
tation of an oxide of protactinium by the delib-
erate addition of oxide ion may provide the basis
for such a reprocessing method if the simultane-
ous precipitation of UO, can be avoided. Previ-
ous studies have demonbtratfld the chemical
feasibility of oxide precipitation for removing
protactinium from a flucride mixture, LiF-BeF -
ThF4 (67-18-15 mole %), proposed as the blanket
of a two-region molten-salt breeder reactor.!?

In the fluoride fue! mixture of the MSRE, sui-
ficient ZrF, has been added to accommodate
gross oxide contamination without loss of uranium
from solution as UO,_. Earlier studies had dem-
onstrated that UC, would not precipitate at 700°C
from the solvent, LlF Bel', (66-34 mole %) with
added UF, and ZcF , unhl the concentration
ratio of ZtF, to UF, dropped below about 1.5.'°
Therefore, a preliminary study was made ot the
precipitation of PaO, from a fluoride mixture
known to have Zr0, as the stable oxide phase.!®

The f{luoride mixture consisted of LiF-BeF
(66-34 mole %) with added Z:F, equivalent to
0.5 mele of zirconium per kilogram of salt mix-
ture. About 1 mc of **°Pa was included in the
salt prepatation as irradiated ThO_. The mix-
ture was pretreated with anhydrous HF and H,
to convert oxides to fluorides. )

The deliberate introduction of sclid-phase oxide
to the melt was made by adding Zr0O, in small
increments. Filtered samples of the salt mix-
ture were taken at assumed equilibrium condi-
tions after each oxide addition and were analyzed

out

 

14+ H. Shaffer ef al., Nucl. Sci. Eng. 18, 177 (1964).

Ppescior Chem. Div. Ann. Progr. Rept. Jan. 31, 1961,
ORNL-3127, p. 8.

Yo Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNL-3789, p. 56.
42

The results
showed that approximately

for 23%Pa by gamma spectrometry.
of these analyses
80% of the 2°°Pa activity was removed after
the addition of about 67.5 g of Zr0, (equivalent
to 0.125 mole per kilogram of salt) to the mixture.
labile solid
solution with ZrO, or was removed from solu-
tion in the salt mixture by surface adsorption
on ZrQO,, then its distribution coefficient should
have remained constant. The fraction of Pa
remaining in the liquid phase could then be ex-

If protactinium either formed a

pressed as a linear function of added ZtO, by
the equation

- LW
y ZYOZ ?
FPa W’salt

(10)

 

where D = [Pa]oxide/[Pa]salt’ Fpa = fraction of
Pa in salt, and W is the weight of the designated
interpretation of the experimental
function,

phase. An

data according to this linear shown

ORNL-DWG €66-969

W
- _ 1 B ZFOQ
Fpg Lo ( )

a WsaLT

 

 

WHERE £, = FRACTION OF #33pa IN LiQUID

 

 

 

 

a0
2
W“*-‘-‘WEIGHT RATIO OF PHASES
SALT
& | _ CONC. OF Pa IN SOLID PHASE |
! CONC. OF Pa IN LIQUID PHASE P
b
. *
3 : - s
| ~
i /
4 +— ‘1 o

 

 

HREp

Waarr = 3.5 kg
./ ) 0= 237

 

 

 

 

 

RECIPROCAL FRACTION OF Pe REMAINING IN LIQUID PHASE

 

 

 

 

 

 

O 10 20 30 40 50 60 70
WEIGHT OF Zr0Q, ADDED AS SOLID PHASE (g)

Fig. 3.6. Removal of 233Pa from Solution in LiF-BeF,
(66-34 Mole %) with Added ZrF, (0.5 Mole per Kilo-
gram of Salt) by Addition of Z:0, ot 600°C.

in Fig. 3.6, illustrates the constancy of the
distribution coefficient at a calculated value of
about 237. The data further suggest that ahout
7 g of the initial ZrO, addition partially dis-
solved in the salt phase or was otherwise lost
from the reaction mixture. Further experiments
will include studies of the effect of ZiO, sur-
face arca on protactinium removal from salt mix-
tures proposed for a single-region molten-salt
breeder reactor.

HEMOVAL OF PROTACTINIUM FROM MOLTEN
FLUORIDES BY REDUCTION PROCESSES

W. K. R. Finnell
W. P. Teichert
W. K. Grimes

J. H. Shaffer
F. A. Doss

The effective recovery of ?23Pa from a molten-
salt breeder reactor will provide more economic
production of fissionable #33U by substantially
reducing blanket inventory and equipment costs
and by improving neutron utilization. Accord-
ingly, chemical development efforts supporting
the reference-design MSBR are concerned with
the removal of protactinium from the blanket mix-
ture, LiF—BeFZ-ThF’4 (73-2-25 mole %), by methods
which can be feasibly adapted as chemical proc-
esses. An experimental program has been ini-
tiated to study the reduction of protactinium
fluorides from this salt mixture by molten lead
or bismuth saturated with thortum metal at about
400°C. It was hoped that protactinium, as Pak
in the salt phase, would be reduced to its metal-
lic state by thorium metal and could he recovered
in the molten lead or bismuth.

The primary objective of initial experiments
with this program has been the study of protactin-
ium removal from the salt phase of the extraction
For these experiments sufficient 233Pa
was obtained for radiochemical analysis by neu-
tron irradiation of a small quantity of ThO,.
The simulated blanket mixture was prepared from

together with the irradiated
ThO,, in nickel equipment. This mixture was
treated at 600°C with an HF-H, mixture (1:10
volume ratio) to remove oxide ion and at 700°C

system.

its components,

with H, alone to reduce structural-metal impuri-
ties. The metal-phase extractant, lead or bis-
muth with added thorium metal, was prepared in
the extraction vessel {304l stainless steel
with a low-carbon steel liner) by treatment with
H, at 600°C. 'The extraction experiments were
started by transferring a known gquantity of the
prepared salt mixture into the extraction vessgel
containing the metal-phase extractant. Filtered
samples of the salt phase were taken periodically
for radiochemical analysis of dissolved protac-
tinium. In each experiment, ?3°Pa was rapidly
removed from the salt phase and remained absent
from the solution during the approximately 100
ht at 600°C while tests were made. Subsequent
hydrofluorination of the extraction system with
an HF-H, mixture (1:20 volume ratic) showed
that “%3Pa could be rapidly and almost quanti-
tatively returned to solution in the salt phase.

Typical results of these experiments are shown
in Fig. 3.7. In this experiment thorium metal
was added after the salt mixture was introduced
into the extraction vessel in order to demonstrate
the necessity of the reduction teaction.

The objective of experiments now in progress
iz to examine methods for recovering 2®°Pa from
the extraction system, The proposed use .of
macro quantities of 2?!Pa may be required to

ORNL-UWG 66-970

 

 

 

 

 

 

 

 

 

 

100 w—-T——-—-—-—-—-—---,‘.______q
: !

80 }—~ ------------------- — -
W WEIGHT OF SALT: 6004
w WEIGHT OF LEAD: 900 ¢
2 | | L
T b‘o e b e ) .._.........._.,A:z e e = 2 emeenm < < o o ]
& \ : |
5 |
< !
o
< a0 |- _ [ e
¢ |
oy WITHOUT ADDED WITH Gdwt %o
N THORIUM & Th IN Pb ADDED

20l ol ol ]

s a :
o
1=
J -
0 Ai,__ VR”;\‘A. . o)
0 20 40 O 20 40

EXTRACTION TIME (hr)

Fig. 3.7. Effect of Thorium Metal on the Extraction
of 233Pa from LiF-BeF,-ThF, (73-2-25 Mole %) in Salt-
Lead System at 603°C.

circumvent the anticipated adsorption of the micro
guantities of ?°3Pa, currently used, on the walls
of the container, or on other inscluble species
in the system. Additional studies of the delib-
erate precipitation and adsorption of *33Pa on
solid, stationary beds such as steel wool will
be made for comparative evaluation.

SOLUBILITY OF THORIUM IN MOLTEN LE.AB

J. H. Shaffer
F. A. Doss

W. K. K. Finnell
W. P. Teichert

Current studies of the removal of protactinium
from a molten fluoride mixture, which simulates
the blanket of the reference design MSBR, have
been directed toward the development of a liquid-
liquid extraction process. The method proposes
that protactinium, as PaF, in a salt phase, can
be reduced to its metallic state and extracted
into a molten-metal phase. The possible use
of a molten mixture of thorium in lead would
provide a convenient method for combining: the
reducing agent with the metal-phase extractant
and for replenishing thorium to the fluoride blanket
mixture. The objective of this study has been
to establish the solubility of thorium in lead
over the temperature range of interest tc this
program and to provide a lead-thorium solution
of known composition for subsequent protactin-
ium-exfraction experiments. :

The experimental mixture, contained in low-
carbon steel, consisted of approximately 3 kg
of lead and 100 g of thorium-metal chips. Values
for the solubility of thorinm were obtained by
analyses of filtered samples withdrawn from the
melt at selected temperatures over the interval
400 to 600°C under assumed equilibrium condi-
tions. Samples wete withdrawn during two heat-
ing and coeling cycles and submitted for acti-
vation and spectrographic analyses. These re-
sults, plotted as the logarithm of the solubility
vs the reciprocal of the absolute temperature
in Fig. 3.8, indicate that the heat of solution
of thorium in lead is approximately 19 kcal/mole
and that its solubility at 606°C is about 1.85 x
104 m.f.
CRNL— OWG 66 - 971
TEMPERATURE (°C}

  

 

 

600 550 500 480 400
30 — - ——
T—Ffi* T
& | |
o . B | |
% 2.0 T o BY ACTIVATION ANALYSIS
g 0 8Y SPECTROGRAFPHIC ANALYSIS
o l
QO
£ |
= 10— N — e —
S e o
_J_j. . <1___ - =
; T
- | ‘ |
F - e _ _..._‘r—
1. ; (
O ] o _ i -
= \
9 |
— . _ -
<I 1 |
r ®
=
L o B
Q)
O |
o
(W] ! i }
0.2 L. _L_qL LJ_
1 i2 13 14 5 16
«o,ooo/nom

PROTACTINIUM STUDIES IN THE HIGH-ALPHA
MOLTEN-SALT LABORATORY

C. J. Barton

The Reactor Chemistiy Division has, for the
past few lacked facilities for research
with significant quantities of alpha-active mate-
rials 23%py and ?®*'Pa. Interest in
using the latter isotope as a stand-in for the
more radicactive 23%Pa led to construction of
the High-Alpha Molten-Salt Laboratory which is
shown in Fig. 3.9. Seven interconnected glove
boxes are presently installed in the laboratory,
and one end of the train is connected to a glove-
equipped hood. The glove boxes are maintained
at a negative pressure with respect to the labora-
tory by an and the
room is at a negative pressure with respect to
the surrounding area. The glove-box exit air
is doubly filtered, and all air exhausted from
the room discharges through the plant central
off-gas system.

The glove boxes are equipped for, or are adapt-
able to, a variety of operations, but the study

years,

such as

automatic control system,

14

Fig- 3-8-
of Therium in Melten Lead.

Temperature Dependence of the Solubility

of the removal of protactinium from molten-fluoride
breeder-blanket and supporting basic
research are the principal uses planned for this
laboratory. A Zeiss polarizing microscope, a
Mettler balance, and a quenching-furnace arrange-
ment are mounted in separate glove boxes. The
stainiess steel box at the right side of the angle
of the train in Fig. 3.9 is fitted with a stain-
less steel heating well that is welded to the
bottom of the box and is surrounded by a 5-in,
tube furnace. Most of the high-temperature studies
other than quenching will be performed in this
box which is also egquipped with a manifold to
control application of vacuum or admission of
helium, hydrogen, or HF to flanged pots in the
heating well.

An experiment was performed in this box to

mixtures

confirm the oxide precipitation of protactinium
from molten salts reported earlier.'* Protactin-
ium at tracer concentration (<1 ppb ?°’Pa) was
completely precipitated by addition of thorium
oxide to molten LiK¥-BeF -ThF, (73-2-25 mole %),
and treatment the oxide-contaminated melt
with a dry mixture of HF and H, redissolved the
protactinium, in agreement with the results of the

of

previous investigators.

An ion exchange method reported by Chetham-
Strode and Keller!?” was used to purify about
0.1 g of 23113’.21205. The oxide was dissolved
in 2.5 M HF and loaded on an anion exchange
bed. Elution with 17 M HF gave a purified
fraction essentially free of alpha-emitting daugh-
ters and, according to the originator of the proc-
free of niobium which was present as an
impurity in the oxide received from England.
The purified 23!'Pa fraction will be used for
further studies of protactinium
breeder-blanket mixtures.

ess,

recovery from

174, Chetham-Strode, Jr., and O. L. Keller, Jr., *“Ion
Exchange of Protactinium{V) in HF Solutions,’”’ paper
presented at the International Conference on Protactin-
ium Chemistry, Orsay, France, July 2-8, 1965.

 
 

 

 

45

 

PHOTO 80771

Fig. 3.9. View of Glove Boxes in High-Alpha Moiten-3alt Laboratory.

SEGREGATION ON FREEZING LiCl-KC!
EUTECTIC MELTS CONTAINING
SOLUBLE SOLUTES

H. A. Friedman . F. Blankenship

Purification by freezing is possible under some
circumstances.  In the absence of solid-solution
formation and with freezing rates sufficiently
slow that rejected solute can diffuse away from
the advancing freezing front, hipgh degrees of
putification can be achieved; in principle, this
process  might be applicable to reprocessing
fuel melts from a molten-salt reactor. Slowness
is one of the obvious disadvantages of this method.
to teduce the thickness of the fixed
film through which diffusion must occur, strongly
influences the rate at which freezing can be
A few experiments were carried out
to determine whether the stirring available “with

Stirring,

efiective.

ordinaty laboratory equipment would be adequate
to give good separation of salts at freezing rates
sufficiently high to be of poteatial practical
interest.

For convenience and visibility, the LiCl-KC
cutectic, contained in an open 400-m! Pyrex

beaker, was used as the melt to which impurities
were added. Stirring was accomplished with a
Pyrex propeller stirring rod revolving at 700
tpm. The beaker was lowered from a hot zone
to a cold zone at a rate of about 2 cm/hr; this
was near the lower limit of rates that were deemed
of practical interest and was not changed. After
some preliminary trials, the segregations shown
in Table 3.4 were obtained while using a coiled

Calrod heating element. Beneath the heating
element there was a ring with holes through
which cooling air was blown; this helped fix

the location of the freezing front.

Part of each ingot containing PbCl, was ex-
posed to moist H S to convert the lead to the
sulfide, thereby ‘fleveloping the concentration
gradient for visual observation. The upper ingot
in Fig. 3.10 is the control which f{roze witheut
To the left and righi are the second
and fourth ingots listed in Table 2.4. The sharp
segregation of the impurity into the last liquid
to freeze is clearly evident.

stirring.

The experiments were regarded as a demonstra-
tion of the successful purification of a salt mix-
ture by freezing with rapid stirring.
46

Table 3.4, Chemical Analyses of Frozen Ingots

 

 

 

Approximate Chemical Analysis
Concentration (ppm)
Solute of Solute Added —_ Remarks
(ppm) Bottom Center Top
PbClza 900 760 850 1000 Heated with gas-ring burner; no sfirring
PbCI2 1100 <30 140 6700 Started using method described as the final
apparatus; used double-propeller stirrer
PbCI2 1100 <60 2650 Same method
PbClz 960 <60 <250 4940 Same method except used single-propeller
stisrrer
Ninb 790 455 391 Same method as directly above
SInF3 1000 500 1900 Melt remained cloudy; results may have been
confused by insoluble oxide
UF4 1160 35 15 3130 Slight clouding of melt

 

 

a ; . . .
Conirol experiment to show effect of no stirring.

bRegarded as failure due to solid-sclution formation.

PHOTO 65545

 

 

Fig. 3.10. Segregation of PbCl, in Frozen Ingots of LiCl-KCl Eutectic. Ingot fragments were treated with
moist H,5 to develop the concentration profile. Upper sample: contral, without stirring. l.eft and right: typical

segregation with stirring.
4. Direct Support for MSRE

PREPARATION AND LOADING
OF MSRE FLUORIDES

J. H. Shatfer
W. K. R. Finneil

F. A, Doss
W. P. Teichest

The preparation of all MSRE {luoride mixtures
and the loading of these mixtures intc the reacior
drain tanks have been completed by Reactor Chem-
istry  Division personnel. Preparation of the
secondary-coolant and the flush sait (totaling
15,000 b of 7LiF~BeF2 mizxfure containing 66
mole % "LiF) was completed and the mixture was
foaded into the MSRE during 1964.':? MSRE
fuel was prepared as the following salt composi-
tions: fuel solvent ("LiF-BeF -ZiF , 64.7-30.1-5.2
mole %), depleted fuel concentrate (LiF-*370F |
73-27 mole %3, and enriched fuel conceatrate
("LiF-?%5UF , 73-27 mole %).

The enriched fuel concentrate (six batches, sach
containing about 33 1b of 275U} and most of the
11.000 1bh of fuel solvent were prepared during
1964.7°7  The few remaining batches (eachk of
about 275 1b) of this material and the two batches
{of about 600 ib) of depleted fuel concentrate were
prepared during early 1965 by the technique de-
scribed previcusly.?

Loading of the fuel mixfures isto the MSRE to
achieve final fuel composition for criticality and
for eventual power operation was completed during
this reporting period. Initial loading consisted in
the transfer of 10,050 1b of fuel solvent to the fuel
drein tank and the addition of 320 (b of depleted
fuel concentrate.  This mixture,
the ultimate fuel composition but essentially with-

quite near to

EMSR Program Semiann. Progr. Rept. July 31, 1964,
CRML-370G8, p. 288.

*Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1963,
ORNL-3789, p. 99.

47

oul enriched uranium, was then transfetred to the
reactor circuit, was circulated there for some 250
hr in a precritical test (PC-2), and was returned to
the MSRE drain tank. There, enriched fuel con-
centrate was added. Four additions of the con-
centrate (with criticality tests of the fuel in the
reactor core interspersed) brought the **5U inven-
tory of the drain-tank contents to 68.76 kg. These
operations  had been accomplished in a routine
manner by previously described techniques ' ? by
late May 1965.

Additions of enriched fuel concentrate to the
MSRE tank had, as anticipated, brought the 23°U
conceniration to more than 95% of that required for
criticality. The remaining ?°°U was added to the
fuel circulating in MSRE as solid pellets of enriched
fuel concentrate in small nickel capsules.® These
capsules, each containing about 85 g of 27°U
in about 148 g of the "LiF-UF, sutectic, were
filled from a =single lurge batch of enriched fuel
concentrate, Capsules were construcied from 6-in.
lengths of nickel tubing with 3/&-=ifi. outer diameters,
0.038-in. walls, and hemisgherical bottoms., The
top plug was penetrated by two Yin.-CD x 0,025
in.-wall nickel fill tubes. Seven capsules were ¢on-
pected in series by their %-in. fill tubes and
clugtered within @ 4-in-diem chamber which was
externally heated. The inlet and cutlet {ill tubes
to the cluster were connected by tube fittings to
the salt transfer line and te an overflow reservoir,
The assembly was heated to 600°C; helium pres-
was applied to the salt storage container to
flow the melten fluoride mixture into the clustered
capsules. Displaced gases were vented through
the top of the overflow reservoir. Liquid levels
in the capsules were visually observed by radiog-
raphy with a portable x-ray unit and a TVX camera.

sure

 

e, WL Haubenreich, private communication (Aug. 16,
1963).
 

Fig. 4.1,

The capsule cluster and salt transfer line were
cooled to near room temperature while back flowing
helium through the system. The filled capsule
cluster was disconnected from the assembly and
its exposed tubes were capped. The net weight of
salt mixture in the cluster was determined, and
the cluster was sealed in a water-tight can. A
photograph of the equipment during a typical filling
operation is shown in Fig. 4.1, The 161 capsules
were filled, at a rate of about five clusters per
day, during a five-day period, wotking two shifts
per day.

Although the accountability of 235U was main-
tained for each capsule cluster, each capsule was
photographed by x rays before and after filling.
Thus, minor variations in weight from capsule to
capsule could be calculated from measurements on

the contact prints. This examination revealed no

 

Filling of Fuel-Enriching Capsules for MSRE.

defects in the capsule clusters nor any variation
in the frozen salt mixture.

For use in the MSRE, the capsules were detached
from the cluster, the inlet and outlet tubes were
sealed, and the capsules were opened and inserted
through an appropriate mechanism into the circu-
lating fuel in the MSRE pump bowl as needed.?

in uranium density

CHEMICAL BEHAYIOR OF FLUORIDES
DURING MSRE OPERATION

R. E. Thoma

The necessary service operations and the eval-
uations to determine whether the molten fluorides
retain their chemical purity during MSRE operation

 

 
form an important, integral part of the MSRE sup-
port program. In preparation for this effort, inves-
tigations have been made — and in some cases are
continuing ~— in such diverse lines as crystalliza-
tionn behavior of reactor salts,*'® solubility of
possible contaminant oxides as a function of
temperature,® compatibility of fluorides with their
metallic environments,” and the effects of irradia-
tion on molten and frozen salts.® In a closely
cooperative effort, the ORNL Analytical Chemistry
Division has developed methods for routine-basis
analyses for the several species whose concentra-
tions must be known if chemical integrity of the
system is to be assured.

Initial, precritical operation of the MSRE (run
PC-1) was performed with flush salt (7LiF‘-BeF2,
66 mole % 7LiF) in the fuel circuit. This flush
salt, therefore, constituted the frozen-salt seal in
the freeze flanges and the freeze valves of the core
circuit. After circulation for a 1000-hr test (and
for removal of oxide scale, if any, from the reactor
circuit), the flush salt was drained to its own
storage tank., There it was treated with HF and H,
to remove oxide before reuse; a reassuringly small
quantity, corresponding to 115 ppm of 0?7, was
recovered as H O in this step.

The fuel mixture was then constituted — with
an interruption for the PC-2 test — in the fuel
drain tank and the reactor circuit as described
in the section immediately preceding. Since the
reactor citcuitry does not drain perfectly, the not
unexpected result was the dilution of the fuel
mixture by some flush salt left in the reactor from
PC-1. This dilution resulted in a disparity between
nominal and analytical values for uranium con-
centrations in PC-2 and in differences between
nominal and measured 238U enrichment values
during the zero-power experiments. We believe
that approximately 140 lb (less than 1.5%) of the
"LiF-BeF | flush-salt diluent remained in the fuel
circuitry at the end of the PC-1 test and was

 

*H. A. Friedman and R. E. Thoma, Reactor Chem.
Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL.-3789, p. 7.

“MSR Program Semiann. Progr. Rept. Jan. 31, 1964,
ORNL-3626, p. 117.

A. L. Mathews, C. F. Baes, and B. F. Hitch, Reactor
Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-
3789, p. 56.

W. R. Grimes, MSR Program Semiann. Progr. Rept.
July 31, 1964, ORNL-3708, p. 238.

83F. F. Bilankenship, MSR Program Semiann. Progr.
Rept. July 31, 1964, ORNL-3708, p. 252.

49

responsible for the apparent discrepancy. Careful
chemical analysis and mass-spectrometric deter-
mination of the 235U/?38U ratio indicate that
the critical concentration in the MSRE was 4.453%
by weight of U (1.39% %35U). This figure is lower
by 1.26% than the book value of 4.51% U.

Analysis samples for metallic corrosion
products continues to indicate that, as expected,
corrcsion of the INOR-8 by the fluoride melts is
insignificant.  The compounds NiF, and FeF
are unstable toward reduction by Cr in INOR-8;
these fluorides would be corrosion agents, therefore,
if present in the fluids. Neither FeF nor NiF
is stable toward reduction by Hz at elevated
temperatures; the pretreatment of fuel, flush, and
coolant salts with HF-H  and then H = should,
therefore, have removed other fluorides completely.
Chemical analysis of the tluoride mixtures before
introduction into the reactor showed about 5 ppm
of Ni and about 125 ppm of Fe. We believe that
these materials are present largely, if not entirely,
as colloidally suspended metals. On the other
hand, CrF, is not reduced by H,; the analyzed
concentration of Cr (about 30 ppm initially) is
believed to be present as Ccr?”,

During the PC-1 test, with "LiF-BeF , (66 mole
% 7LiF) in the fuel and coolant circuits, chromium
concentrations in the melt increased by about
25 ppm in the fuel circuit and by about 10 ppm
in the coolant circuit; thus, about 100 g of Cr
was removed from the interior walls of the reactor
circuit. Nickel concentrations remained at about
5 ppm during this test, while Fe concentrations
dropped by about 50 ppm. Occurrence of 30 ppm
of Fe?” in the initial melt is {aside from the
material-balance difficulty) scarcely credible. The
decrease in Fe concentration, therefore, was
believed to be due to settling of some 200 g of
Fe powder in the drain tank or in the circuit. The
increase in chromium concertration represents, we
believe, a teal increase in Cr?'; it is probably
due to dissolving of oxide films from the reactor
circuit walls and the subsequent reduction by Cr
of the C1%%, Fe3™, and Ni?" so obtained.

Analyses for Fe, Cr, and Ni in the reactor fuel
of PC-2 and of subsequent critical and zero-power
runs (totaling about 1100 hr) show no appreciable
concentration changes. The failure of Ct2¥ to
increase in concentration is not surprising since
the oxides should have been well removed in PC-1
and since the fuel mixture was thoroughly treated

of
with H2 (to the extent that about 2% of the uranium
was present as U®?) before addition to the re-
actor.

Purity of the fuel during PC-2 and the zero-power
tests was assured by chemical analysis of daily
samples. Values for the concentration levels of
all contaminants except oxide were uniform, credible,
and satisfactorily low. Ozxide values, obtained by
the KBrF4 method,® were sporadic and high, pre-
sumably as a result of water adsorbed in samples
Verification that oxide
concentration did not seriously exceed the solubility

during sample preparation.

limits was made by regular examination of salt
specimens employing petrographic methods, which
under can detect the
presence of a few hundred ppm of crystalline
oxide,

In recent months, an improved method for deter-
mining the oxide concentration of MSRE salts has
developed. !° The the

reaction

favorable circumstances

been method employs

0%~ | 2HF(g) = H,0(8) + 2F ™ .

Oxide concentration is regarded as equivalent to
the quantity of water evolved as a molten-salt
specimen is purged with an Hz~HF gas mixture.
The method has been applied for assay of the
flush-
initial operations of the MSRE full-power tests.
Oxide concentration was found to be approximately
100 ppm in both the flush and fuel salts.'?

From a chemical standpoint, the operation of the

and fuel-salt specimens obtained during

MSRE during the precritical and zero-power experi-
ments was a success. There is every reason to
believe that the salts were maintained in an ex-
cellent state of purity during all transfer, fill,
and circulating operations.

 

°G. Goldberg, A. 5. Mevyer, Jr., and J. C. White, Anal.
Chem. 32, 314 (1960).

10ysr Program Semiann. Progr. Rept, Aug. 31, 1965,
ORNL-3872, p. 140.

g S. Meyer, Jr., unpublished work and personal
communication.

50

MEASUREMENT OF DENSITIES
OF MOLTEN SALTS

B. J. Sturm R. E. Thoma

Relatively few measurements of the densities
of liquid salt mixtures have been made in the
development of ORNL molten-salt technology.
Values have usually been obtained from estimates
such as the relatively imprecise method of mixtures
employed by Cohen and Jones,'? based on room-
temperature densities of the components, or the
more satisfactory method employed by Cantor,'3
which assumes additivity of molar volumes. In
an effort to obtain more accurate experimental
values, we have measured directly the volumes of
molten-salt mixtures at various temperatures. Initial
results of these experiments were reported pre-
viously.'* By adopting several experiinental in-
novations, we have improved the accuracy of
experimental data appreciably. One such measure
is to prevent salt mixtures from freezing until all
volume wmeaswrements have been completed. In
this way, errors in the volume measurement arising
from distortion of the container vessel upon freez-
ing and remelting the salts are avoided. Values
obtained by this experimental procedure are com-
pared with previous values in Table 4.1. The
equation for density, d, as a function of temperature
is

d-a- bt.

The new values are considered to be in agreement
with the recent measurements of the salts stored
in drain tanks at the MSRE site.?

 

128. I. Cohen and T. N. Jones, A Sumunary of Density
Measurements on Molten Fluoride Mixtures and a Cor-
relation for Predicting Densities of Fluoride Mixtures,
ORNL-1702 (July 19, 1954, decl. Nov. 2, 1861).

3p. B. Bien, S. Cantor, and F. F. Blankenship,
Reactar Chem. Div. Ann. Progr. Rept. Jan., 31, 1961,
ORNL-3127, pp. 24--25; S. Cantor, Reactor Chem. Div.
Ann, Progr. Rept. Jan. 31, 1962, ORNL-3262, pp. 38--41,

l4g J. Sturm and R. E. Thoma, Reactor Chem. Div,
Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 83—
84.
51

Table 4.1. Densities of Molten-Salt Mixtures

 

Density Parameters

 

Melt and Composition Method (g/cm 3)
in Mole % a b
MSRE Coolant: Estimate
66 LiF, 34 BeF Method of mixtures” 2.24 0.0006
Sum of molar volumes” 2.152 0.000391
Lindauer® ¢ 2.160 0.00040
Measurement
Mound Laboratory © 2.158 0.00037
Pressure probed’f
This work 2.296 0.000482
MSRE Fuel: Egtimate
65.0 LiF, 20.17 BeF ), Method of mixtures® 2.61 0.0007
5.00 ZrF4, 0.83 UF4 Sum of molar volumes? 2.670 0.000594
Measurement
Haubenreich?
This work 2.848 0.000769
Volatility Solvent It Estimate
62.5 KF, 20.8 ZrF4, Method of mixtures® 2.92 0.0007
16.7 AlF, Sum of molar volumes” 3.298 0.000992
Measurement
This work 3.178 0.00106
Volatility Solvent II: Estimate
57.7 KF, 19.2 ZrF4, Method of mixtures® 2.91 0.0007
23.1 AIF Sum of molar volumes” 3.350 0.001038
Measurement
This work 3.202 0.00105

Density (g/cma)

650°C

1.921

1.954
1.986

1.983

2.15
2.384

2.331¢
2.348

600°C

2.50
2.703

2.54

2.49
2.728

 

2ORNL-1702.
PORNL-3262, pp. 38—41.

“Peter Patriarca, private communication (Juae 1965).

dOriginally reported in Fnglish units but converted to metric for comparison in this report.

"MLM-1086.

fj. R. Lngel, private communication (February 1965),

Ep, N, Haubenreich, private communication (Aug., 14, 1965),
Part 1l

Aqueous Reactors

 
5. Corrosion and Chemical Behavior

in Reactor Environments

MECHANISM OF ANODIC FILM GROWTH ON
ZIRCONIUM AT ELEVATED TEMPERATURES

A. L, Bacarella A. L. Sutton

it was shown previously!-? that the anodic-film-
growth current for zirconium in oxygenated 0.05 m
H S0 at temperatures from 200 to 284°C can be
represented by a hyperbolic sine function of the
field strength,

i-i lexp (BV/X ) —exp (- BV/X )l
= 2i sinh (BV/X ). (1)

The first term in the exponential form of the equa-
tion represents the current flow assisted by the
field, and the second term the cumrent flow against
the field.
thicknesses (high field strengths), the second term
becomes negligible, and the ‘‘high-field’’ approxi-
mation is valid:

At low temperatures and small film

i io exp (BV/X ), (2)
where V/X ~is the average macroscopic tield
across the film thickness XL,' B is the field coef-
ficient, and {_ contains terms which are independ-
ent of the field strength and includes the concen-
tration of the mobile ions (anion vacancies) and

 

1

A, L. Bacarella and A, L. Sution, Reactor Chem.
Div. Ann. Progr. Rept. Jan. 31, 1865, ORNIL-3789,
pp. 135-38.

ZA. L. Bacarella and A, L. Sutton, Electrochemical
Technology, in press.

55

It was further shown
that B could be interpreted in terms of parameters
which have fundamental significance:

the activation-energy terms.

€+ 2)
p e+
kT 3

 

; (3)

where ¢ is the charge on the anicn vacancy, a is
the activation distance, k7 is the DBoltzmann
energy factor, ¢ is the dielectric constant of the
oxide at temperature T, and the term[(¢ + 2)/3] x
(V/X ) is the local (Lorentz) field effective in the
transport of anion vacancies, The temperature
dependence of the dielectric constant € was shown
to give values for (gqa)¥ which were reasonable
and independent of temperature.

During the past year the anodic~film-growth re-
action was studied further, and the measurements
were extended to 174°C. A large number of meas~
urements were performed to obtain a better statisti-
cal estimate of the parameters, particularly those

implicit in 7. On the basis of the model used,

iy =2 x10%aqu C(X) exp (~¢,/RT)

% exp (~¢,/RT), (4

where 10° coulombs per faraday converts transport
of charge to unils of amperes, 2a is the jump
distance, v (10'? to 10'? per second) is the fre-
quency factor for anion vacancies, C is the con-
centration of anion vacancies (a function of posi-
tion in the oxide), and ¢, and ¢, are the activation
energies for the formation and transport of vacan-
cies.
Table 5.1.

i =2 sinh (BV/X); B= -

56

Paragmeters in the Rate Equation for Anodic Fiim

Growth

(qa) (€ + D)

m——n A

kT 3

V=225v;g=1c¢

 

 

 

Temp. 1'0 L] B .« kT i (qa)V {(qa) = a
°K) (amp/cm %) (e /v) (e % cm) = (cAv) (eA = A)

447 2.4 % 107 ° 5.02 % 107° 19.3 % 1078 36 3.44 1.53

474 1.15 x 1078 6.00 % 107° 24.4 % 1078 46 3.44 1.53

505 8.7 % 1078 7.33 % 1078 31.8% 1078 59 3.55 1.58

525 2.6 %107 8.54 » 107 ° 39.1 x 1078 73 3.52 1.56

557 a,7x 1077 10,2 % 1078 48.8 x 1072 (96) (3.59) (1.59)

208 4.75 % 1078 12,2 % 1078 28 2,75 1.22

Earlier papers®** have presented a description through anion-deficient zirconium dioxide tilms

of the electrochemical cell and the experimental
methods. Experimentally, we obtain the anodic-
film-growth current 1 in amperes per square centi-
meter as a function of time at constant potential
and temperature. The correlative film thicknesses
aie obtained by integrating the anodic current with
time and by assuming an initial thickness X0 =

100 A:

X, =2.88 x 10° f[idmlOU.
0

Dielectric constants previously reported! were
obtained from but a few measurements and were
in error by about 20% because of an unsatisfactory
experimental design. Those to be reported here
are based on a larger number of measurements and
are believed to he more reliable estimates.

Table 5.1 contains a more recent compilation of
the parameters in the rate eguation,
is an Arrhenius plot of the preexponential current
1'0. The revised value for the activation energy,
¢ + ¢, obtained from these data is 29.5 + 3
kcal/mole, compared with our previously reported
26 + 2 kcal/mole, Tennyson Smith® has determined
the activation energy for

Figure 5.1

diffusion of oxygen

 

3A. L. Bacarella and A. 1., Sutton, J. Electrochem.
Soc. 112, 546 (1965).

1A, L. Bacarella, Reactor Chem. Div. Ann. Progr.
Rept. Jan. 31, 1962, ORN1.-3262, pp. 8689,

5'".[‘em'lyson Smith, J. Flectrochem. Soc. 112, 560
(1965).

using the “‘interrupted-kinetics’’ technique® and
has reported ¢ = 31.1 kcal/mole, a value in near
agreement with our revised value. This agreement
suggests further that the rate equation used to
fit the data provides a satisfactory basis for an
explanation of the film growth.

Considerations of the concentration temm C(X)

in Eq. (4) were made, in which it was assumed that

in one case and that
C(X) = Co exp (—8X)

in another case, with CO/CL a constant independ-
ent of X . The field was taken as the general
function of X, E(X)., It was found that either dis-
tribution of vacancies led to an equation in which
the vacancy concentration C(X) is CO:

7)
RT

1=2x10%agy CO exp (—

 

(ga) (¢ + 2)
X eXp —— EX) , (5
P27 3 (X)y » (5
®A. Rosenburg, J. Electrochem. Soc. 107, 795

{1960). ,
 

 

 

 

10761

 

 

 

 

 

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

 

f. CURRENT DENSITY {amp/cm?)

 

 

 

      

 

 

 

 

 

 

 

 

 

2 feeeeereeneb e, 25000 23D°C  200°C 1 Nereeeeeepeeeeeeneanes
A T | ]\
1079 ! J -
1.70 1.80 1.80 2.00 210 2.20 2.30 2.40
1000/7_ (°K)

Fig. 5.1. Arrhenius Plot of the Preexponential Cur-
rent iy in the Rate Equation i = iy exp (BV/X).

where E(X)  is the field at X equal to zero, Coms-
parisons between the values of C(X) obtained using
the data in Table 5.1 and those evaluated by other
workers® 7 using different approaches show reason-
able agreement considering the probable uncer-
tainties of the values,

Although the form of E(X)  is not yet known, we
believe that the results of the considerations
indicate that the empirical equation (1) can be
interpreted in terms of parameters of fundamental
significance.

 

R, F. Domagala and D, 5, McPherson, J. Metals 6,
138 (1054).

57

MECHANISM OF RADIATION CORROSION
OF ZIRCONIUM AND ZIRCALOCY.2

R. J. Davis G. H. Jenks
Oxide Growth and Film Capacitance on
Preirradiated Specimens

Previous wortk,® which showed that reactor ir-
radiation of Zircaloy-2 caused accelerated post-
irradiation corrosion, was extended to crystal-bar
zirconium. Specimens were: (1) pickled in HF-
HNO_, (2) vacuum (1 to 3 x 10~° mm Hg) annealed
at 700°C to remove traces of fluoride, and (3)
exposed to 300°C steam-oxygen for 1 hr and heated
in helium at 500°C for 16 hr to provide an oxygen-
rich layer near the surface. The subsequent
treatments were the same as in previous experi-
ments.® Data from two specimens, one irradiated
and one contrel, show that the irradiated specimen
corrodes about twice as fast as the control. [t
indicated that this radiation effect?® is not
related to the additives in Zircaloy-2 nor to fluo-
ride left by pickling. '

is

Search for New Methods to 3tudy
Oxide-Fiim Properties

It can be posiulated that the irradiation damage
results in less protective oxide.® Altemating-
current impedance data were collected?® in the hope
that they would distinguish, on an empirical basis,
between oxides of varying protective quality.
They did not. Some exploratory work has there-
fore been done to find a better method to evaluate
film protective properties. The two prevalent
theories of oxidation of zirconium are: (1) diffu-
sion of anion vacancies through a concentration
gradient and (2) diffusion of an oxygen species
through pores. Methods of measuring vacancies
ot pores were therefore sought.

The Anion-Yacancy Grodient. — Recently an
indirect measure of the anion-vacancy concentra-
tion gradient in tantalum oxide was reported by
others.? Film capacitance vs temperature was
measured; above a critical temperature a rapid
rise in capacitance occurred from the increased

 

°g. J. Davis, Reactor Chem. Div. Ann. Progr. Rept.
Jan. 31, 1965, OCRNL~-3789, pp. 126 ff.

D, M Smyvth et al., J. Electrochem. Soc. 111, 1331
(1964),
electronic conductivity produced by thermal ioniza-
tion of vacancies. The results enabled a calcula-
tion of the vacancy gradient, This technique works
if the electronic conductivity can be made high

Our calculations using reported data®®
11

enough.
and observations the required
conductivity can be realized in anodic zirconia
films after heating in inert environments, Explor-

indicate that

atory experimental work in which specimens were
heated for short times (<1 hr) in helium or oxygen
did not show the effect. Additional work is planned.
Direct-Current Resistance. — The possibility
was explored of determining dc resistance of anodic
films on zirconium using contacts of mercury,
evaporated silver, and gallium. The apparent
resistances with silver, mercury, or unrubbed
gallium drifted to higher values with time., Ac-
cordingly, it was concluded that these contacts
are not suitable, at least without additional work,
Contacts made by rubbing gallium onto the surface
were stable, they obeyed Ohm’s law, they were
the same at either polarity, and they were propor-
tional to filin thickness; but they were more than
four orders of magnitude lower than can be recon-
ciled with ac impedances measured in electrolytes.
The ac impedances with the rubbed gallium contacts
are consistent with the dc resistance and fit a
simple parallel resistance-capacitance equivalent
circuit. (1) the
gallium filling pores (4 x 107'* cm? pore area/
cm?) or (2) the gallium being rubbed through an
outer, to an inner
layer of resistivity about 5 x 107 ohm-cm. Since
the significance of the resistances is uncertain,
work on this will probably not continue,
Permeability, — Zirconia membranes made by
dissolving the metal substrate from 100-v-anodized
films are permeable to potassium nitrate in solu-
tion.

This behavior may result from:
2

high-resistance oxide layer

We have confirmed reported!? findings, but
we consider the interpretations uncertain because
it is possible that pores are formed during removal
of the substrate,

Zirconia films are permeable to hydrogen'?

under some conditions and the permeation can be

 

1D‘Tennyson Smith, J. IZlectrochem. Soc. 111, 1020
(1964);, 111, 1027 (1964) 112, 560 (1963),

Ty, L. Young ef al.; Some Electrical Properties of

Zr02 Film, AERE-R-4957 (June 1965).
12 A, H. Mitchell and R. E, Salomon, J. Electrochemn.
Soc. 112, 361 (1965).

138, A. Gulbransen and K. F. Andrew, J. Electro-
chem. Soc. 101, 348 (1954); 101, 560 (1954),

58

measured with the metal substrate intact because
hydrogen dissolves rapidly!? in the metal. An
appartatus to measure hydrogen uptake rates under
controlled designed and

conditions has been

partially assembled. Specimens will be exposed
to hydrogen at 100°C, and the hydrogen volume will
be periodically measured to about +107° cm?
(STP)/cm?. Correlation of permeability with cor-
rosion rates will indicate whether or not corrosion
is a process of permeation through pores.

EFFECTS OF REACTOR OPERATION
ON HFIR COOLANT

G. H. Jenks

Final evaluations'? (1) the
concentration of excess oxidant required in the
HFIR reasonable
assurance that the concentration of hydrogen ions
near will not differ ap-
preciably from that in the bulk of the solution and
(2) the expected steady-state concentrations of
the decomposition products of water at particular
concentrations of excess oxidant. It was con-
cluded that the excess-oxidant concentration
should be about 1073 M O2 or the equivalent H,0,.
The expected steady-state concentrations of H,

were completed of:
coolant-moderator to provide

fuel-element surfaces

0,, and H,0, with this excess oxidant are near
1072 M.
experimental values in the HFIR awaits power

Comparison between the predicted and

operation of the reactor.

NASA TUNGSTEN REACTOR RADIATION
CHEMISTRY STUDIES

G. H. Jenks H. C. Savage
E. G. Bohlmann

Poison control solutions of CdSO, are being
considered for possible use in the NASA Tungsten
Water-Moderated Reactor,
the effects of irradiation on the stability of these
solutions toward loss of cadmium is needed for a
complete evaluation of this poison control system.

Information regarding

 

Coolant, ORNL-3848 (Cctober 1965).
We have planned and are developing experiments
to test the stability of CdSO, solutions under
electron with intensity and other
conditions such that they either simulate those in
the reactor or provide a severe test of the stability
under irradiation.

Test solutions in contact with Zircaloy-2 at
temperatures in the range 60 to 120°C will be ir-
radiated at power densities up to 150 w/cm?®, The
solution will be static in one type of experiment.
In another type, the solution will be circulated in
otdet to simulate fluid-film conditions in the re~

irradiation,

actor. The temperatures, power densities, solu-
tion compeositions, and contlainer material will be
those of possible interest in the reactor, The use
of electrons rather than reactor radiations is ex-
pected, from theoretical considerations,
crease the chance that an irradiaiion effect will
occuy at a given power density. The two ditferent
types of experiments are believed necessary to
assure that the radiation stability is tested under
conditions at least as severe as those in the re-
actor. Theoretical considerations did not enable
us to predict reliably whether agitation would
have any effect on the solution stability o, in
fact, the direction of an effect if one occurred.

The static iradiation cell is comprised of a
single loop of small-bore Zircaloy-2 tubing sur-
rounded by a coolant jacket, as shown in Fig,
5.2. The Zircaloy-2 tubing contains a fine filter
at one end of the leop. In operation, the test solu-
tion will be exposed to radiation in the Zircaloy-2
- {ube, It will then be forced through the filter,
collected, and analyzed. The temperature will
be controlled by passing controlled-temperature
water through the jacket at a rapid rate. The re-
sults of component testing of the static system
have provided reasonable assurance of design
feasibility and adequacy. The final test system
iz under construction.

Dynamic experiments will be conducted with a
small, high-speed (35,000 rpm), centrifugal pump
with which solution is circulated through a small
tube which fomms a loop in front of the cover plate
of the pump. The diameter of the pump is about
1/2 in., and the total liquid volume is about 1,:} cm®.
All the solution will be irradiated continuously
during an exposure. The tube is included in order
to provide a channel in which the flow is well
defined and in which film coefficients can be cal-
culated and, if necessary, measured. Calculations

{o ine

: PHOTO 80937

   

 

of Static-Ceil

Mockup,
0.403 in.; inner
= 0,423 in.; outer diom-

eter of the irradiation coil = 0.503 in.

Fig. 5.2. X-Ray Photogrophs
Inner diometer of the cooling coil =

diameter of the irradiation coil

show that the required maximum velocity in the
tube depends upon diameter and is <25 fps for a
diameter of 26 mils.

Tests of the performance of a pump design’®
were made using a stainless steel mode!l designed
by L. V. Wilson of the Reactor Division., The
results showed that the head decreased with in-
creasing flow, being, at 33,000 rpm, about 77 ft at
shutoff and 51 ft at 220 cm®/min, The results
of other tests have shown that the latter head-flow
capability is sufficient to produce the required
maximum velocity in a loop of 26-mil fubing. Some

 

15Thompson Ramo Wooldridge, Inc., Jet=Cenirifugal
Mercury Pump Design and Performance Analysis, NAA~
5R-6314, TRW Report No. ER-3420 (February 1964)
{confidential).
problems with the shaft seal and with wear be-
tween the shaft and housing were recognized and
remain to be corrected before a final Zircaloy-2
unit is constructed.

The results of the tests with the pump model
and of other completed tests of the dynamic system
indicate that the design is feasible and adequate,

CORROSION SUPPORT FOR VARIOUS
PROJECTS

I.. L. Fairchild
P. D, Neumann

Griess
. L. English

Beryllium Corrosion

Three high-flux reactors either being constructed
or designed will use metallic beryllium, exposed
directly to the coolant, as moderator. The reactors
are the High Flux Isotope Reactor (HFIR), the
Advanced Test Reactor (ATR), and the Argonne
Reactor (AARR). In the
first two reactors, the coolant will be water ad-
justed to a pl of 5.0 with nitric acid at a maximum
temperature of about 100°C; in the AARR, the
coolant will be deionized water at a slightly lower
temperature.

Test results reported previously'® showed that,
in water adjusted to a pH of 5.0 and tlowing at 12
to 80 fps, beryllium corroded at a constant rate of
1.9 mils/year over periods that lasted as long as
17,600 hr. A few small pits were noted on some
of the specimens. In these tests, the ratio of

Advanced Research

exposed beryllium surface area to volume of water
was 10 cm?/liter and the system contained 22.5
liters of water, which was continuously deionized
by passing a side stream through a cation ex-
changer (hydrogen form) at the rate of 3 liters/hr.

In one of the specimens in the above series of
tests, two cracks originated from stenciled identi-
These
were first detected after 10,440 hr of exposure.
Prior to that time, all specimens had been ex-
amined several times and no cracks had been found.
This particular specimen was removed from test after

fication marks on one edge of a specimen.

11,980 hr and subjected to metallographic exami-

nation. Figure 5.3 shows the larger of the two

 

16]. C. Griess et al., Reactor Chem. Div. Ann. Progr.
Rept. Jan. 31, 1965, ORNL-3789, pp. 122-23.

60

cracks, which penetrated into the metal a distance
of 7 mils. Examination of the polished specimen
under polarized light showed the crack to be trans-
This appears to be the first

incident of stress~corrosion cracking of beryllium

granular in nature,

ever reported,

In recent tests conducted for the AARR in de-
ionized water at 93°C (200°F) at 44 fps, the effect
of the ratic of surface area of exposed beryllium
to volume of water was examined. In tests that
lasted in excess of 1000 hr, it was shown that
when the exposed beryllium area was 0.45 cm’
per liter of water, the beryllium corroded at a
constant rate of 2.7 mils/year. Under identical
conditions with a surface-area-to-volume ratio of
4.5 cm?/liter, the corrosion rate was 1.1 mils/
vear. In both of the above tests, the total volume
of water in the system was 27.6 liters and a side
stream was passed through a mixed-bed deionizer
at a rate of 4.9 liters/hr, This is the same relative
rate of deionization that will be used in the AARR.
The ratio of exposed surface area of beryllium to
volume of water in the AARR is 2.3 cm?/liter,
essentially midway between the cxtremes tested,

Although one isolated case of apparent stress-
corrosion cracking was observed and a few small
isolated pits were found, the test results indicate
that unclad beryllium will have adequate corrosion
resistance in the three new reactors,

Chemical Development for HFIR Operation

G. H. Jenks has concluded from a study of the
expected radiation chemistry of the HFIR coolant
(water adjusted to a pH of 5.0 with nitric acid)
that either oxygen or hydrogen peroxide at some
moderate level must be present in the coolant to
assure the stability of the nitrate ions.!” It is,
therefore, desirable to monitor continuously the
excess oxidizing capacity of the solution, that
is, the concentration of oxygen and hydrogen
peroxide in excess of the equivalent concentration
of hydrogen, During precritical hydraulic testing
of the HFIR system, the usefulness of a con-
based on the reaction
metal was
However, the hydrogen peroxide

tinuous oxygen analyzer'®
of dissolved oxygen with thallium
demonstrated.

 

17G. H. Jenks, Effects of Reactor Operation in HFIR
Coolant, ORNL=3848 {October 1965),
18R, s. Greeley et al., Checkout of Bettis Dissolved

Oxygen Analyzer, ORNI.-CF-60-1-57 (Jan. 14, 1960).
 

61

Y-56635

 

003

05

Q0e

 

 

Fig. 5.3. Crack in QMV Beryllium Specimen After 11,980 hr in pH 5.0 Water at 100°C Temperature and 23 fps

Flow. As polished; magnification 500x.

present in the reactor water reacts with thallium
metal relatively slowly aad irreproducibly, so it
was necessary to include a catalytic chamber cons
taining a large surface area of platinum to de-
compose the peroxide and permit measurement
of total available oxygen. Laboratory studies
showed that the platinum effectively recombines
dissolved hydrogen and oxygen, so that in the
HFIR system only the excess oxidant will be
measured, This measurement will be used to
conirol the addition rate of oxidant (either oxygen
or hydrogen peroxide) to the HFIR primary coolant.

With the above analyzer and confirmatory chemi-
cal analyses, the decomposition rate of hydrogen
peroxide in the primary coolant was determined in
the absence of radiation. The results indicated
that the decomposition is a homogeneous solution
reaction with a half-life of about 6 hr at 105 to
110°F.

Hydriding of Zircaloy-2 and Tantalum
In the Transuranium Processing Facility, the

aluminum=-clad target rods irradiated in the HFIR
will be dissolved in hydrochloric acid, Both

Zircaloy-2 and tantalum have adequate corrosion
resistance to the acid solution for equipment use,
but both can suffer hydrogen embrittlement under
some conditions. Therefore, a series of tests with
both materials was conducted to determine the
amount of hydrogen entering the metal from the
corrosion process and its effect on the ductility
of the material.

Zircaloy~2 strips, 10 mils thick, and tantalum
specimens of similar thickness were exposed to
8 M HCl at several temperatures for periods of
1000, 2000, and 3000 hr, After exposure, each
specimen was given a 180° bend test as a qualita-
tive measure of its ductility. Originally, the
Zircaloy-2 contained 8 ppm hydrogen and the
tantalum 1 ppm.
rosion

The hydrogen content and cor-
rate of the Zircaloy-2 specimens after
various exposure periods and the results of the
bend tests are shown in Table 5.2

At each temperature, the hydrogen content ap-
peared to be approaching a limiting value after
2000 to 3000 hr. Generally, the corrosion rate
and the corresponding hydrogen content increased
with temperature. For the particular geometry
involved, the hydrogen concentration required to
Table 5.2.

62

The Corrosion Rate and Hydrogen Pickup of Zircaloy-2 Exposed to 8 M HC!

 

Temperature

Corrosion Rate (mils/year)

Hydrogen Content (ppom)

 

°C)

 

 

1000 hr 2000 hr 3000 hr 1000 hr 2000 hr 3000 hr
35 0.4 0.8 0.7 18 170 220
45 0.5 1.1 1.0 A2 350 380
55 0.6 1.4 1.1 150 730 700
65 2.2 2.3 1.9 430 680 820
75 3.4 3.2 2.4 760 13007 1600°
85 5.0 3.8 2.9 740 17007 19007
95 4.0 3.5 2.5 660 18007 1500

105 6.0 4,7 3.1 34007 40007 4200°

 

4Brittle fracture of specimen during 180° bend test.

cause embrittlement was between 1000 and 1500
pprm.

The tantalum specimens exposed under the same
the Zircaloy-2
completely resistant to the acid and consequently
did not pick up hydrogen.

The target-rod dissolver in the Transuranium
Processing Facility will be made of Zircaloy-2
and will operate at about 45°C, a temperature
where corrosion and embrittlement should not be
a problem, For evaporators in which acidic solu-
tions will be concentrated at the boiling point,
tantalum liners will be used.

conditions as specimens were

Corrosion Testing in Support of
Power-Reactor Fuel-Element Reprocessing

During the past year, the effort in this program
has been concerned solely with the behavior of
pure nickel and Hastelloy N in various gas-phase
head-end treatments. In the gas-phase processes,
the fuel elements are suspended in a fluidized bed
of aluminum oxide, and, depending on the specific

process, are subjected cyclically to fluorine, 60-
40 mixtures of oxygen and hydrogen fluoride, and
oxygen alone at temperatures as high as 800°C.
In the tests conducted, specimens of both materials
were exposed in the fluidized bed and in the gas
phase above it.

Both Hastelloy N and nickel underwent only
minor corrosion during 31 cycles of exposure to a
60-40 mixture of O -HF at 625°C. Each cycle
consisted of a period of 1 hr for heatup, 6 hr at
temperature, and 1 hr for cooling, Both alloys
were heavily oxidized on exposure to oxygen at
800°C. However, only Hastelloy N was signifi-
cantly corroded during eight cycles of a process
consisting of 6 hr in oxygen at 550°C followed by
6 hr in fluorine at 550°C. Additional testing in a
cycle consisting of 6 hr in a 60-40 mixture of
O,-HF at 625°C followed by 6 hr in fluorine at
550°C showed that the attack on nickel was light
but that Hastelloy N specimens were heavily
attacked on the edges.
an integral part of the process development, and
the results are included in various reports by the
Chemical Technology Division.

This testing program is
6. Chemistry of High-Temperature Aqueous Solutions

ELECTRICAL CONDUCTANCE MEASUREMENTS
OF AQUEQUS SODIUM CHLORIDE SOLUTIONS
TO 800°C AND 4000 BARS |

A. S. Quist W. Jennings, Jr.
W. L. Marshall

The electrical conductances of sodium chloride
solutions, ranging in concentration from 0.001 to
0.1 m, have been measured at temperatures from
100 to 800°C and at pressures from 1 to 4000
Two different conductance cells were used
One of the cells

bars.
for this series of measurements.
has been described previously.'™?  However,
because of experimental difficulties with the
Bridgman type of pressure seal in the high-tem-
perature region of the conductance cell, a new
cell was constructed which eliminated the pressfire
seals in the high-tempetature region. This new
cell was constructed from a cylinder of Udimet
700, 24 in. long and 1 in. in diameter. A hole
0.250 in. in diameter was drilled the length of
the bar, and . platinum-iridium liners were inserted
to prevent corrosion of the Udimet 700 by the
high-temperature aqueous solutions. The other
parts of the conductance cell are of the same
design as described earlier. !> :
Both conductance cells were used for measure-
ments on the most dilute (0.001 to 0.02 m) solu-
tions. ‘The measurements from the two cells were
in agreement to within experimental error (1 to
29). The sesults for 0.01 m Na(Cl solutions are
shown in Fig. 6.1, where specific conductances are

 

1E. U. Franck et al., Reactor Chem. Div. Ann. Progr.
Repi. Jan., 31, 1961, ORNL-3127, pp. 5052, _
A, 8. Quist ef al., Reactor Chem. Div. Ann. Progr.

Jan. 31, 1962, ORNL-3262, pp. 7375,
¥, Franck et al., Rev. Sci. Instr. 33, 115 (1962).

Rupf.

3

63

plotted as a function of temperature at pressures to
4000 bars. The type of behavior shown in Fig. 6.1
is typical for a strong electrolyte, and is similar to
the results obtained previously with potassium sul-
fate.®'S A
300°C, iofiic mobilities increase rapidly due to
the rapid decrease in the viscosity of water.
Therefore the conductance of the sodium chloride
solutions increases also, However, the dielectric
constant of water is also decreasing with increas-
ing temperature. This lowering of dielectric con-
stant allows ion pair formation fo begin to occur.
Apparently, at approximately 300 to 450°C (de-
pending on the pressure) the two effects (increas-

As the temperature increases from 0 to

ing mobility of the ions; increasing association
between the ions) begin to offset one another, and
the conductance begins to diminish with increasing
temperature. It should also be noted that when
the temperature of an aqueous solution is in-
creased, at constant pressure, the density de-
creases so that thete are fewer ions per cubic
centimeter, and consequently a smaller conduct-
ance.

In another study, measurements have been made
on 0.01 demal (¥ 0.01 m) potassium chloride solu-
tions (this is used as a standard solution for
conductance cell constant deferminations at 25°C)
to 800°C and 4000 bars. It is thought that this
would be the logical solution to use for a reference
at elevated temperatures and pressures; thus, as
more researchers make conductance measurements
at high temperatures and pressures, the results
from the different laboratories can be more easily
compared.  Figure 6.2 shows a comparison of
previously ' determined conductances of K,80,,

 

*A. S, Quist et al., Reactor Chem. Div. Ann. Progr.
Rept. Jan., 31, 1963, ORNL-3417, pp. 77~82.

SA. 8. Ouist ef al., J. Phys. Chem. 67, 2453 (1063
ORNL-DWG 66-974

700

 

   

  
  

600 - - 4000 bars

500
400
300

200

 

100 i e e

 

SPECIFIC CONDUCTANCE {ohm™ ! em™1) x 10°

400 600
TEMPERATURE (°C)

Fig. 6.1. The Specific Conductance of 0.01 m NoCli
from 100 to 800°C ot Pressures to 4000 Bars.

Solution

ORNL-DWG 66-973
DENSITY (g/cm?)
Q.30

1.04 Q.76

 

 

0.01 m NoCL . \ ;

400
300
200

100

0.0022 m K550, S ?'”'

 

 

SPECIFIC CONDUCTANCE (chm™*!

 

o i .
200 400 &00
TEMPERATURE

 

(°C)

Fig. 6.2. Comparison of the Specific Conductances
of K,50,, KHSO,, H2504, KCl1, and NaCl Solutions as

Functions of Temperature.

H,SO,, and KHSO, solutions with the new values
for NaCl and KCl as a function of temperature at
a constant pressure of 4000 bars. The different
behavior of the KHSO, and H,SO, solutions as
compared to the other solutions?'*~7 is due to
changes the first and second dissociation
constants of sulfuric acid with temperature and
pressure.® ’ Both dissociation constants decrease

in

 

®A. s. Quist ef al., Reactor Chem. Div. Ann. Progr.
Rept. Jan. 31, 1964, ORNL-3591, pp. 84—-88.

7A. S. Quist et al., J. Phys. Chem. 69, 2726 (1965).

64

with increasing temperature, and increase with
increasing pressure. The difference between the
conductances of NaCl and KCl is due to the
difference in mobility of the potassium and sodium
ions.

AQUEQUS SOLUBILITY OF MAGNETITE
AT ELEVATED TEMPERATURES

F. H. Sweeton R. W. Ray
C. F. Baes, Jr.

The solubility of magnetite (Fe,0,) in agueous
solutions at elevated temperatures is of special
interest in pressurized-water reactor systems, in
which Fe, 0, is a main component of the corrosion
film. It also is of interest to geologists in under-
standing the origin of Fe O, deposits.®

The flowing system for measuring the solubility
that was reported earlier® has been modified in
several ways to improve the precision of the data,
One important change has been the substitution
of 66 g of nonradioactive Fe O, for 2.7 g of
radioactive material, thus making possible a longer
contact time with the solution. This Fe,0, was
prepared by oxidizing a carbony! iron powder with
steam at 400 to 500°C. Its final specific surface
area was 0.12 m?/g, about twice that of the radio-
active material.

Water for the tests was purified and equilibrated
with H, as before. All batches had conductivities
of less than 0.08 micromho/cm at room tempera-
tures, and concentrations of dissolved O, below
10 ppb (and usually less than 5 ppb).

The purified water, sometimes containing added
HCl, was pumped first through a recombiner in
which it made contact with platinoum black at
260°C (to catalyze the reaction of the remaining
O, with H,), and on through the bed of Fe 0, in
a column held at a controlled temperature. The
solution was then cooled and passed through a
Millipore filter with 0.1-u pores. The recombiner,
column, and filter holder were made of gold-plated
stainless steel. Platinum tubing was used to

connect the units. The equilibrated solution was

 

8. T. Holser and C. J. Schneer, Geol. Soc. Am.
Bull. 72, 36986 (1961).

9F. H. Sweeton, C. F. Baes, Jr., and R. W. Ray,
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNI1.-3789, pp. 120-21,
ORMNI. ~DWG 66 --975
20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\ * r
\ »
o \
\\
.
o ° \\
10
0 n THEDRETICAL.\___
SLOPE _
B3 5 (NO H‘1’D'ROLYSIS)\_.__..,..,....__.-__;._
5 ° | \
2
: \
BB [T e e e
\
o \
\
2 o 5o ] \
= 4 | e 150°C \C
5 © 200°C \
g ]
8
&
[
®
2
®
®
%
1 _ ®
5.4 5.6 5.8 6.0 6.2
pH AT TEMPERATURE (CALCULAYED FROM
25°C pH DATA)
Fig. 6.3. The Observed Solubility of Fe30, in

Aqueous Selutions Presaturoted at 25°C with Hy  at

1 atm.

passed through flowing conductivity and pH cells
and then through beds of cation exchanger to
collect the dissolved iron, which was later removed
and assayed spectrophotometrically by the
phenanthroline method.

The results are shown in Fig. 6.3. The pH
rmeasured at 25°C has been converted to pH at the
aquilibration temperature, using dissociation con-
stants of water at 25, 150, and 200°C of 0.01,
2.24, and 5.01-107 1'% (molal units), respectively,
at any tem-

S o-

and assuming no hydrolysis of Fe?®
perature. Sixfold changes in the flow rate produced
no significant change in the composition of the
product solution, indicating that the Fe O, had
reached equilibrium with the flowing solution.

65

We have used the upper two groups of data in
Fig. 6.3 to calculate the solubility product (in
molal and atmospheric units) for the reaction

Y% Fe,0,(s) + 2H (soln) + %11,(g)

= Fe?*(soln) + 4@1‘120(301"1) :

The resulting values of log K at 150 and 200°C
are 7.00 and 6.08 with standard deviations of
0.11 and 0.06 respectively. We think the other
150°C points are low due to a slight leakage of
0, into the pH cell to give the reaction

a2t ]
Fe~" + /402+H20

- I/zFezfis + 2HT,

and we are now modifying the pH cell to prevent
this. The solubility product estimated for 200°C
predicts that the solubility of Fe,0, in pure H,0
should be 1.4 pm; this is in good agreement with
our previcusly reported® figure of 1.5 um.

We will continue measurements to cover the
range of 150 to 260°C with HC! solutions from
0 to 30 pm with the intention of examining the
data to see if the dissolved iron is hydrolyzed.

SOLUBILITIES OF CALCIUM HYDROXIDE AND
SATURATION BEHAVIOR OF CALCIUM
HYDROXIDE~CALCIUM CARBONATE MIXTURES
IN AQUEOUS 50DIUM NITRATE SOLUTIONS
FROM 0.5 TO 350°C

L. B. Yeatts, Jr. W. L. Marshall

The solubilities of calcium hydroxide and the
saturation behavior of calcium hydroxide—calcium
carbonate mixtures were determined at tempera-
tures from 0.5 to 350°C in aqueous solutions of
sodium nitrate from 0 to above 3 m in concen-
tration.

Commercially available reagent grade calcium
hydroxide wand calcium carbonate were digested
in boiling deionized water to remove soluble im-
purities. The hot solutions were filtered and the
recovered product dried at 100 to 110°C. The dry
calcium hydroxide was ignited at 1150°C for about
i6 hr to convert calcium carbonate impurities to
calcium oxide; the carbon dioxide content of the

freshly ignited oxide was about 250 ppm. Excess
solid calcium hydroxide and mixtures of calcium
hydroxide and calcium carbonate were equilibrated
with aqueous sodium nitrate, prepared with de-
ionized water, using rocking equipment. The
containing vessels for the 0.5 and 25°C experi-
ments were Pyrex stoppered bottles; for the 50
to 350°C experiments, they were titanium alloy
bombs. The solutions were filtered during sample
withdrawal at the temperature of the experimeit
in order to remove suspended solids. A portion
of each sample was treated with excess concen-
trated nitric acid, evaporated to dryness at 95 to
100°C, and weighed as calcium and sodium ni-
trates. Another portion of each sample solution
was used for the determination of calcium ion
concentration by a potentiometric titration with
standard EDTA solution.
tration in the solution was determined by difference
from the results of these two analyses.

The dissolution of calcium hydroxide was as-
sumed to reach the following equilibrium;:

The electrolyte concen-

Ca(OH),(s) = Ca?*(ag) + 20H(aq) .

The thermodynamic solubility product expression
for this equilibrium is given by:

3..3
v 748}’i,

Kg - m 2+m2 B
P oH

2
ca? ) OH

where m is molal concentration, y is the activity
coefficient, and s is the molal solubility of calcium
hydroxide. Converting to logarithms, substituting
an extended Debye-Hiickel expression for log y,

and rearranging the equation leads to the form:

Sit7?

log K, - log K2 + 6+ 3BI+ 3CI?,
(1 AI'/?%

where

S = theoretical Debye-Hiickel limiting slope
for a given temperature,

+ 3mCa(OH)2’

0 1; oy _
KSp = the solubility product constant at I = 0,

! = ionic strength = nzNaN03

A, B, C = constants.

Since Ksp - 4s? and Ksop: Hs3, then

511/2
4+ BI+ CI*.

log s = log sOyr2_
(1 + AIY/2)

66

The dependence of the calcium hydroxide solu-
bility upon the ionic strength function, I'/2/(1 +
AI'/?), at various temperatures is shown in Fig.
6.4. These curves represent the best least-squares
fit of the experimental data, using the A values
listed in the legend. The intercepts for the curves
are the log s° The inverse relationship
between solubility and temperature and the direct
relationship between solubility and ionic strength
are clearly seen. The poorest fit of the curves
with the data occurs at the higher temperatures,

values.

where the solubility of calcium hydroxide is suf-
ficiently low to make precise analytical measure-
ments difficult. By using the values of Kgp
determined from the data of Fig. 6.4, assuming
that ACp - D + ET(°K), where D and E are con-
stants, and then by determining the change in
log Kgp as a function of 1/7(°K) with the van’t
Hoff expression, the standard
quantities were calculated and are presented in

Table 6.1.

thermodynamic

CRNL-DWG 6512010

 

 

Ca(OH}, {(molality)

 

 

 

 

 

03 05

72144 1%2)

06

Fig. 6.4. Solubility of Ca(OH), in Aqueous NaNO;
from 0.5 to 350°C.
67

The saturation behavior of calcium hydroxide—  alone over the same range of temperature and
calcium carbonate mixtures was identical, for  ionic strength.
all practical purposes, to that of calcium hydroxide

Table 6.1. Standard Thermodynamic Properties of the Equilibrium: Ca(OH),(s) =2 Cq‘2+(aq) + 20H " (ag)

 

Temperature AF® Af"” As®

 

 

(OC‘.) Kgp (kcal/mole) (kcal /fmole) (cal mole ! deg™ 1)
0 1.32 % 1073 6.10 ~1.26 ~26.9
25 9.61 x 1077 6.84 ~2.97 ~32.9
50 5.81 % 107° 7.74 4,79 ~38.8
100 1.46 x 10~° 9,96 —~8.74 ~50.1
150 2,60 % 1077 12.7 ~13.1 )
200 3.75 x 1078 16.1 ~17.9 ~71.9
250 4,70 % 1072 10.9 —023.2 ~82.4
ano 5.34 x 10710 24.3 ~28.9 -92.8

350 5,69 x 10711 20,2 -35,0 ~103

 
7.

SURFACE CHEMISTRY OF THOR!A

C. H. Secoy

Heats of Ilmmersion and Adsorption

H. F. Holmes E. L. Fuller, Jr.
J. E. Stuckey!

The caleorimetric investigation of the interaction
of water with the surface of ThO, is continuing,
The apparatus and procedure have been de-
scribed.?™ 3

Net differential heats of adsorption (AH_) of
water on ThO, have been derived from calori-
metrically determined heats of immersion (h) of
ThO, samples containing known amounts of pre-
sorbed water (adsorbed after outgassing at 500°C
but prior to the immersion experiments). These
studies have been completed for three samples
of ThO : A (650°C, 14.7 m?/g); B (800°C, 11.5
m?/g); and E (1600°C, 1.24 m?/g). Data in pa-
rentheses refer to the calcining temperature and
specific surface area respectively. Initial values
of AH_ ranged from -20 to —30 kecal/mole.
no case did z.i\Ha decrease to zero with completion
of the f{irst adsorbed monolayer. These large
exothermic values clearly indicate that the water
is in a chemisorbed state, most probably as sut-

In

 

!professor of Chemistry, Hendrix College, Conway,
Ark.

‘c. H. Secoy, H. F. Holmes, and E. I.. Fuller, Jr.,
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNL-3789, pp. 164—69.

3¢. H. Secoy and H. ¥, Holmes, Reactor Chem. Div,
Ann. Progr. Rept. Jan. 31, 1963, ORNL-3417, pp. 12429,

‘H. F. Holines and C. H. Secoy, J. Phys, Chem. 69,
151 (1G665).

SH. F. Holmes, E. IL.. Fuller, Jr., and C. H. Secoy,
to be published in J. Phys. Chem. (February 19686).

Interaction of Water

68

with Particulate Solids

resulting from hydrolysis
At larger coverages addi-

face hydroxyl
of surface oxide ions.
tional contributions to AH_ arise from hydrogen
bonding between the surface hydroxyl groups and
the subsequently adsorbed layers of water. Sam-
ples A and B gave AHa vs coverage curves which
are typical of adsorption on a heterogeneous sur-
face.® The corresponding curve for sample E
exhibhited three distinct regions in which water
is adsorbed with a constant AH_.  The most
plausible explanation for this unusual behavior
is the formation of successive immobile homo-
geneous monolayers. Such an idealized adsorp-
tion is more likely to occur with sample E because
of its relatively large crystallite size (>2500 A)
as compared to samples A and B (200 A},

The dependence of the heat of immersion on
physical properties such as specific surface area,
particle size, and crystallite size is a complex
4.5.7  The heats of
immersion of a series of low-surface-area ThO,
samples are shown in Fig. 7.1. In addition to
sample E, these include samples D (1200°C,
2.20 m?/¢); ] (1200°C, 2.96 m?/g); K (1400°C,
1.55 m?*/g); and 1. (1600°C, 0.95 m?/g). Maxi-
mum deviation of the experimental points from
the curves in Fig. 7.1 is about 2%.
sideration of the combined

groups

and unresolved question.

From a con-
uncertainty in the
surface area and heat measurements, one must
conclude that samples E, J, K, and L have iden-
This is the first known
case of such behavior. It would thus appear that

in order to have reproducible idealized surface

tical heats of immersion.

®a. C. Zettlemoyer and J. J. Chessick, Advan, Chem.
Ser., No. 43, p. 88, American Chemical Society, Wash-
ington, D.C., 1964.

’W. H. Wade and N. Hackerman, Advan: Chem. Ser.,
No. 43, p. 222, American Chemical Society, Washington,
D.C., 1964,

 
ORNL-0OWG G6--9786

 

 

 

 

TOO
600
<2
&
2
- "
1 h
50G
400
OUTGASSIMG TEMPERATURE {°C)
Fig. 7.1. Heats of Immersion of Low-Surface-Area

Thoria Samples.

reactions, one must work with samples having
low specific surfaces and a relatively large crys-
tallite size (the crystalliie size of these samples
ranged from 1500 to >2500 A). This does not,
however, the results obtained with
sample D. The only physical difference in sam-
Samples E, J, K,
and L. have a mean particle diameter of about
1.5 p, while the corresponding value for sample
D is 3.0 y. On this basis it appears that the
number of crystallites per particle is an important

account for

ple D is the particle size.

factor.  Conceivably, this could influence the
number and type of crystal faces exposed for
surface reactions. Further experiments are re-

quired to resolve this important point,

Water Vapor Adsorption and Desorption

€. L. Fuller, Jr. H. F'. Holmes

The complex nature of water adsorption on tho-
rium oxide has been studied by use of a sensitive
microbalance. High-temperature sintering (1200°C)
appears to produce a material which predominantly
presents the 100 cubic face in the surface, upon
which there are three distinct modes of adsorption.
There is an initial rapid chemisorption, formiung
surface hydroxyl groups which are slowly hydrated.
In addition and as a precursor for surface hydra-
tion, physical adsorption occurs,

constructed at 25.00°C reveal the
type II physical adsorption isotherms character-
istic of polar adsorbates. In addition, the slow
hydrating process occurs as a perturbing effect:
each successive isotherm fails to close upon
the preceding one by a decreasing amount until,
after six months of repetitive adsorption and de-
reproducible

Isotherms

sorption, a isotherm is achieved.

The rate of irreversible binding is a complex
function of both the water vapor pressure and the
amount of water previously bound. The rate is
increased markedly with the first increments of
pressure, but higher pressure
less accelerating effects. The rate diminishes
appreciably as more water is bhound under the
physically adsorbed water.  The final vacuum
weight of adsorbed water corresponds exactly to
the aforementioned stoichiometry.

increments have

The 1000°C calcination decreases the number
of pores but does not alter the distribution (there
are siill pores of radii near 10 A); whereas the
small pores anneal out at the 1200°C calcination,
leaving a minimum radius of 30 A.  The fact
that thorium oxide must be heated to 1000°C
in vacuo to remove all water indicates that there
are two different types of sintering involved.

The amount of bound water is stoichiometrically
equivalent to that required to form the surface
analog of a hydrated bulk hydroxide. In addition,
this is the amount of water required to build up
one completed face-centered cubic lattice unit
on the surface above the 100 Th()é plane, with
the hydroxide and water oxygen occupying the
image pesitions of the substrate oxide ions. Thus
this bound water may be similar to the cubic form
of ice, where the O—Q spacing is nearly equal to
that present in bulk Th()z.

The shape of the nitrogen isotherms at --195°C
changes, showing that the partial pressure at
which monolayer (BET) coverage occurs increases
from below 0.05 to 0.13 as more water is: pre-
adsorbed. This is undoubtedly due to the fact
that the water-covered surface is less energetic
and has less affinity for the nitrogen.

The extremely low vapor pressure of the initially
adsorbed waler and the complex kinetics of hydra-
tion must be considered before evaluating thermo-
dynamic properties or specific surface areas from
water isotherms. One must be equally cauntious
when attempting to predict the amount of adsorp-
tion from these isotherms.
TRANSMISSION —a=

AMBENT

 

 

 

S b
2800 200C 1800

—

—_l

4000 3600 3200

70

ORNL-DWG G5 -10810

         

7

1800

FREQUENCY (o )

Fig. 7.2. Infrared Spectra of ThO, (Sample A) in Vacuo After 24 hr Outgassing at Indicated Temperatures.

Ambient spectrum obtained in laboratory envirenment at a relative humidity of 30%.

Infrared Spectra of Adsorbed Species on Thoria

C. S. Shoup, ]Jr.

Exploratory infrared spectra of a thin, self-sup-
porting pressed disk® of thorium oxide (sample
A)? were recorded in vacuo (107 torr) as a func-
tion of pretreatment conditions. The disk was
held in a nickel sample holder within an infrared
cell and, by means of an external magnet, could
be raised to a section of the cell away from the
silver chloride windows and heated as high as
500°C. After pretreatment at a known temper-
ature for at least 24 hr, the cell was sealed off
under a vacuum, the sample cooled and lowered
to the optical section, and the cell placed in the
spectrophotometer.®  All spectra were obtained
at about 35°C, and the sample was not exposed
to the atmosphere between measurements.

Several features of interest are shown in the
infrared spectra of Fig. 7.2, The most obvious

 

8¢c. 4. Secoy and C. 8. Shoup, Jr., Reactor Chem.
Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp.
172-73.

IPerkin-Elmer model 521 infrared spectrophotometer
provided through the courtesy of the Chemistry Division
of ORNL.

feature is the general decrease in intensity of
the various absorption bands with increasing out-
gassing temperature, particularly between 200
and 500°C. In addition, several of the absorption
bands are displaced in frequency, and in some
new bands appear with increasing out-
gassing temperature. This indicative of the
formation of structurally different adsorbed species
as some of the previously adsorbed species are
desorbed.

The broad bands around 3500 cw~! and less,
due to perturbed O—H stretching modes of vibra-
tion, show that a considerable amount of hydrogen-
bonded adsorbed water remains on the surface
after evacuating at 200°C. Even after outgassing
at 500°C, the thoria is not free of surface O—H
groups, as shown by the rather sharp bands above
3600 cm™!'. Spectra obtained with an improved
signal-to-noise ratio display three sharp bands
at 3715, 3650, and 3520 cm ' when the quantity
of adsorbed water corresponds to approximately
The relative intensities

cases
13

one monolayer or less.
of these bands and their changes under a variety
of conditions are such as to indicate the presence
of three different types of surface O—H groups
at low coverage.
Some contamination of the surface is evident
from the bands due to C—H stretching vibrations
in the 2850 to 2950 cm™! region. This hydro-
carbon-like material is easily removed by suffi-
cient heat or by treating briefly with oxygen at
high temperatures and has only minor effects on
the rest of the spectrum,

A prominent absorption band at 1630 cm™
in the spectrum of the thoria—water
vapor interface in the presence of waler vapor
at a partial pressure of about 14 torr. Pumping
at toom temperature was sufficient to remove this
band, however, confirming it to be due to the
H—O—H bending motion of physically adsorbed
hydrogen-bonded water molecules.

A series of in vacuo infrared spectra were oh-
tained after the 500°C-out-gassed thoria had been
exposed to water vapor for periods of 24 hr to
as long as 11 days each before subsequent evacu-
ation at room temperature. Each successive
spectrum revealed the presence of a greater quan-
tily of adsorbed water than the previous spectrum,
“irreversible’ adsorp-
tion of water that had been originally detected

b was

revealed

thus confirming the slow

gravimetrically. 10

Electrokinetic Phenomena at the Thorium

Oxide—Aquecus Solution Interface!!

. S. Shoup, Jr. H. F. Holmes
Electrokinetic effects of agqueous solutions in
porous plugs of thorium oxide have been investi-
pated from the standpoint of irreversible thermo-
dynamics.'® A kinetic electroosmotic technique
was used to determine the electrokinetic poten-
tial. ' This method gave results in agreement
with streaming potential data except in the pres-
ence of acidic solutions, In general, the electro-
kinetic zeta potential was observed to decrease
with incteasing pH, varying from about +25 mv
to —55 mv with an isoelectric point near pH = 9.4.

 

10-. H. Secoy, E. L. Fulier,
Reactor Chem. Div. Ann.
ORNL-3789, pp. 16972,

Y4, F. Holmes, C. S. Shoup,
J. Phys. Chem. 69, 3148 (1965).

125 R, deGroot, Themodynamics of [rreversible Prr)(—
esses, Inferscience, New York, 1951.

3¢, H. Secoy and H. F. Holmes, Chem. Tech. Div.
Ann. Progr. Rept. Aug. 31, 1959, ORNL-Z?’SB, pPp. 85—86.

Jr., and H.
Progr., Rept. Jan.

F. Holmes,
31, 1965,

er., and C. H. Secoy,

71

ORM DWG p5-2979R

  
    

HaO AND
D NH@GH IN POROUS £1LUGS

- v M0 AND KCI IN GAPILLARIES
No.?ro3 IN POROUS PLUGS

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 7.3. Specific Surface Conductivity as a Function
of Electrolyte Specific Conductivity.

An investigation of ThO, plugs from a variety
however, indicates that the electro-
influenced by the

of sources,
kinetic potential is strongly
method of preparation of the thoria.
one batch of thoria calcined at 1600°C gave no
indication of a positive zeta potential over the
pH range of 5.9 to 12. The complex behavior of
the electrical double layer has been observed in
other oxide~aqueous sclution systems!*=1% and
appears to be highly dependent on the chemical
and thermal history of the sample. Until this
dependence can be put on a concrete basis, zeta-

For example,

potential mesasurements for such systems must
be assumed to be valid only for the specific sam-
ples investigated. |

Measutements of the electrical conductance of
aqueous solutions flowing through porous plugs
of ThO, were correlated with measurements using
ceramic caplllarles* with known geometries.'” This
made it possible to measure the cell constant
associated with the surface conductance®! and
thus to determine the specific surface conduc-
tivity for several electrolytes. The variation of
the specific surface conductivity, A, with the
specific conductivity of the bulk solution, A,,
is illustrated in Fig. 7.3 for a variety of porous
plugs and capillaries of thorium oxide.

 

Hp,
Chem.

1 SD.
Trans.

J. O’Connor and A. S. Buchanan, Australian J.

6, 278 (1953).

J. O°Connor, P. G. Jobansen, and A.S. Buchanan,
Faraday Soc, 52, 229 (1956).

100, Street, Australian J. Chem. 17, 8§28 (1964).

17¢. H. Secoy and C. 8. Bhoup, Jr., Reacfor Chem,
Div., Ann. Progr. Rept. Jan. 31, 1964, ORNL-3591,

“pp. 1089,
An important feature of the results shown in
Fig. 7.3 is the fact that they are some two to
three orders of magnitude larger than would be
predicted from the classical theories of surface
conductance.!® In addition, there seems to be
no obvious relationship between the surface con-
ductivity and the electrokinetic potential as pre-
dicted by these theories. Classically, the expo-
nential dependence of surface conductivity on
the electrokinetic potential is based on the pres-
ence of excess ions in the double layer due to
the potential difference. In the present system,
however, it appears that the contribution of the
excess ions in the double layer is masked by a
second mechanism which is much larger in magni-
tude.

It is believed that this second mechanism is the
ionization and subsequent conductance of surface
hydroxyl groups. This mechanism is further indi-
cated by the relatively slight dependence of the
surface conductivity on the bulk ionic concentra-
tion, implying that the concentration of the sur-
face conducting species is also only slightly
dependent on the bulk ionic concentration. If
the major mechanism for surface conductance is
ionization of surface hydroxyl groups, it should be
dependent on pH, as shown by the consistently
high values obtained with NH OH. In the case of
basic solutions of Na,CO,, however, the slight but
consistent reduction in surface conductivity may
be due to strong specific adsorption of carbonate

ions.

GAS EVOLUTION FROM SOL-GEL
URANIUM-THORIUM OXIDE FUELS

D. N. Hess B. A. Soldano
Thoria-3% UO, sol-gel material prepared for
use as a reactor fuel has been found to evolve
gases when heated or when in reactor service.
Such gases could develop excessive pressure
within fuel elements or could react with the fuel
element cladding to weaken it or to affect its
Previous study of gas
of this

heat transfer capability.

evolution from representative samples

 

181 Th. G. Overbeek in Colloid Science (ed. by H. R.
Kruyt), wvol. 1, chap. V, Elsevier, New York, 1952,

material !® had shown that particle size, or sur-
face area, was an important parameter, with respect
to both the amount and composition of the gas
evolved, and that CO, was reversibly absorbed
at temperatures below 500°C and easily desorbed
at higher temperatures.

Studies during the past year have been directed
toward the development of treatments which would
reduce the surface area of the prepared sol-ge!l
and thus reduce the amount of gas which it could
release in service. Batch exposures at temper-
atures of 1000°C to CO, or H,0 alone did not
give any substantial change in the surface area
of the sol-gel or in its capacity for reversible
adsorption of COQ. On the other hand, mixtures
of the two gases, CO, and H,O, at 1000°C were
found to give approximately 50% decreases in
the surface area and absorption capacity of the
sol-ge!l for each treatment. The duration of the
treatment did not appear to be important. Thus,
two treatments would reduce the surface area to
25% of the original — three to 12.5% of the orig-
inal. Tests in which the sol-gel pellets were
ground to successively smaller particle sizes
between successive batch treatments with mix-
tures of H,O and CO, at 1000°C suggested that
the effect of the treatment permeated the entire
material and was not localized at the superficial
surface exposed by grinding. Table 7.1 illus-
trates these results; a threefold reduction
BET surface area was accompanied by a fourfold
reduction
the particle

in
in CO, adsorption capacity, although
reduction would have, in an
untreated sol-gel, caused an eightlold increase

size

in both surface area and gas evolution with an
accompanying increase in CO2 adsorption capacity.

Since batch treatments are not convenient for
larger-scale concepts and since single batch
treatments, however extended did not
give more than 50% reduction in area, the effect
of treatment in a flowing stream of H,0 + CO,
at 1000°C was studied. As shown in Fig. 7.4, the
flowing-stream technique was much more effective;
the capacity of the sol-gel for gas adsorption
appeared to be a predictable function of treatment
time, regardless of whether the treatments were
continuous or interrupted. A 16-hr treatment ap-
peared appropriate for a 90%

in time,

reduction in the

 

9H. N. Hess, W. T. Rainey, and B. A. Soldano,
Reactor Chem. Div., Ann. Progr. Rept. Jan. 31, 1965,
ORNI.-3789, p. 177.
original gas-adsorption capacity of the material.
Corresponding reductions in the BET surface
area of the material were obtained. The linearity
of the semilogarithmic plot in Fig. 7.4 sugpgests
that there may be a simple first-order process
that is responsible for the surface area reduction,
but the detailed mechanisms of the chemical reac-
tions involved have not been elucidated.

A new series of laboratory samples of sol-gel
material, prepared with various furnace atmos-
phetes during heating, calcination, and cooling
and designated the PL series, was received and

Table 7.1. Effect of Treatment with HzO«CO;Z

on Sol-Gel Properties

 

 

BET Reversible CD2
Mesh Size Sequence of Surface Absorptive
of Tests and Avea Capacity
Particles Treatment (mz/'g) (std co/e)
40=8G Original material: 2.57 0.20
depassed at
1000°C
4080 After first surface 0.10
area reduction
140 Ground and sized 1.58 0.10
to 80--140 mesh
(otherwise untreated)
80-~-140 After second sur- D.08
face area reduction
270 round and sized 1.67 Q.06
to 200-270 mesh
(otherwise untreated)
200270 After third surface 0.78 0.05

area reduction

 

73

evaluated with respect to gas evolution and revers-
ible gas adsorption. All PL samples gave much
less gas evolution upon being ground io the as-
received condition and heated to 1000°C in vacuo;
about 0.30 std cc/g of gas was evolved, compared
with 1.30 for the H-II samples previously studied.
An even more striking reduction was noted in the
capacity of the FL series for reversible CO,
adsorption; values of 0.02 std cc/g were obtained,
compared with 0.27 for the H-II material. Thus
the new material, if its preparation proves to be
reproducible, should be much superior from the
point of view of gas evolution and might not re-
quire any treatment fo reduce its surface area.

ORNL-DWG B6-977

 

 

 

 

 

 

 

 

 

 

0.5 e ; e g
} o e MULTIPLE, CUMULATIVE EXPOSURES
= r | & SINGLE EXPOSURES -
~
£ ]
>
t:
[
Q b e
% M NG e
A N Q\\ “ —_ _
- L SN e -
g 0.05 L _i\ N\ oeim e o
SO
S ——— \\. SN ol
| '
E{-_’ 0.02 —t e -
Lé \\
u
T 00 L- | -4 — A&
RN S L T s
i R J -
0.005 ——-—Jr— e
0 4 8 12 16 20 24 28

TREATMENT TiME (hr)

Fig. 7.4. Effect of Flowing COy and H, 0 Mixture at
1000°C on the Reversible CO, Capacity of ThO;-3%
U0, Sol-Gel.
Part il

Gas-Cooled Reactors

 
8. Diffusion Processes

TRANSPORT PROPERTIES OF GASES

Rotational Relaxation
Numbers for Nitrogen and Catbon Dioxide

Thermal Transpiration.

A. P. Malinauskas

It is well known that serious errors in pressure
measurement can result if the manometer is
maintained at a temperature other than that of
the region whose pressure is of interest. The
phenomenon responsible for the discrepancy is
known as thermal transpiration (or the “thermo-
molecular pressure effect®’) and is defined as the
transport of mass as the result of a temperature
gradient. Unlike thermal diffusion, however, which
is similarly defined, the thermal transpiration phe-
nomenon is not restricted to distinguishable par-
ticles. In fact, ever since its discovery in 1879
by Reynolds,! investigations of the effect have
been conducted exclusively with pure gases.

Unfortunately, the utility of the phenomenon was
thought to be solely that of a pressure correction;
as a result, the scientific community appears to
have been content with the empirical and semi-
empirical formulations which were proposed and
shown experimentally to be qualitatively correct
at  best. Recently, however, Mason and co-
workers,?*? in an attempt to describe gas trans-
port in the region where free-molecule and hy-

3

 

1o, Reynolds, Phil. Trans. Roy. Soc. London, Ser, B
170, 727 (1880); paper No. 33, Scientific Papers, pp-
257.~390, The University Press, Cambridge.

R, B. Evans III, G. M. Watson, and E. A. Mason,
J. Chem. Phys. 35, 2076 (1961) and 36, 1894 (1962);
E. A. Mason and A. P. Malinauskas, f. Chem. Phys. 39,
5272 (1963).

g, A. Mason, G. M. Watson, and R. B. Evans III,
J. Chem. Phys. 38, 1808 (1963); E. A. Mason, J.
Chem. Phys. 39, 522 (1963).

77

drodynamic mechanisms compete, were led to
reinvestigate the thermal transpiration effect and
uncovered a relation by which it appears possible
to obtain information regarding inelastic mo-
lecular collisions from thermal transpiration data.?

The first real test of the theories involved have
been completed in this laboratory during the past
year.” Thermal transpiration studies were con-
ducted with the gases Ar, Xe, N, and C'Oz‘
The phenomenon was generated by maintaining
temperatures of about 290°K and 545%K on the
opposite ends of fourteen O.l-mm-ID Pyrex glass
capillaries arranged in parailel.

Although several apparently minor discrepancies
between theory and experiment were noted, the
utility of the method for studies of inelastic col-
lisions has been demonstrated to be extremely
promiging. For example, the rotational collision
numbers for N, and CO, which were obtained from
the experimental data were found to be 4.4 and
2.4, respectively, and are in excellent agreement
with reported values which were determined by
the more conventional, elaborate,
methods. *

and more

Gaseous Diffusion in Noble Gas Systems

A. P. Malinauskas

Unlike related transport properties,
diffusion, the transport of mass or identity as the
result of a concentration gradient, is
totally govemed by unlike-molecule interactions,
that is, like-molecnle collisions affect the mecha~
nism only as small perturbations. However, the
relative insensitivity of diffusion to the choice

gaseous

almost

 

‘A, P Malinauskas, to be published in The [fournal
of Chemical Physicsa.
of the intermolecular potential characteristic of
the interactions necessitates measurements either
over a wide range of temperature or of an accuracy
far better than that currently possible. Of the
two alternatives, the former appears to be the
more feasible. Indeed, some diffusion measure-
ments have been reported at temperatures as high
as 1100°K without a corresponding loss in
accuracy,5 but the experimental requirements are
so stringent and the procedure ig so elaborate
that a more simple approach is desirable. Such
an approach appears to have been provided by
theory, in that it is possible to relate the com-
position dependence of viscosity to the diffusion
coefficient characteristic of a given gas pair.®
Thus, because viscosity coefficients at high
(or low) temperatures can be readily determined
experimentally, the evaluation of diffusion data
via the theoretical relationship is worthy of in-
vestigation. Although the objectives of the dif=-
fusion program at the Laboratory are primarily
oriented toward an elucidation of intermolecular
interactions, the initial studies have been con-
with an affirmation of
diffusion interrelationship.

cerned the viscosity-
The first series of
investigations in this regard were made with the
systems He=Ar, He-Xe, and Ar-Xe and have
resolved a previous discrepancy between theory
The second phase of the
program involves the systems He-Kr, Ar-Kr, and
Xe-Kr. Analysis of the latter data has not yet
been completed; however, the preliminary results
tend to further validate the theoretical relation-
ship between the viscosity and diffusion phe-
nomena.

and experiment.’

Gaseous Diffusion in Porous Media

A. P. Malinauskas E. A. Mason®
R. B. Evans III

One of the major developments regarding the
migration of gases within a porous medium has
been the formulation of the ‘‘dusty-gas’’ model.

 

SR. E. Walker and A. A. Westenberg, J. Chem. Phys.
31, 519 (1959),

3. Weissman and E. A. Mason, J. Chem. Phys. 37,
1289 (1962).

7A. P, Malinauskas, J. Chem. Phys. 42, 156 (1965).

8Consultant, University of Maryland, Institute for
Molecular Physics.

78

This model, which is particularly useful for a
description of gas transport in the theoretically
difficult
and hydrodynamic mechanisms are of equal im-
portance, utilizes the somewhat unorthodox view
that the solid surfaces of a porous medium or of
the walls of capillaries may be mathematically re-
garded as agglomerates of giant gas molecules
(dust). As a result, a gas-surface interaction is
treated as a special type of molecular encounter
as would occur between two gas molecules when
one of the interacting molecules is immobile and
preponderantly larger and heavier than the other.

Use of the mode!l has yielded the first con-
sistent

transition region where f{ree-molecule

treatments of gas transport in porous
septa for the following cases:?:? (1) diffusion
due to a composition gradient; (2) gas flow as
the result of gradients in composition and pres-
sure; (3) thermal transpiration, in which pressure
and temperature gradients are involved; and (4)
migration (and separation) of gases under the
combined influence of composition and tempera-
ture gradients. However, our inability, at the
time, to cope with the combination of diffusive
and viscous modes of transport in a manner which
was satisfactory from a theoretical viewpoint
necessitated the introduction to a certain degree
of empirical methods in those cases in which
pressure gradients were at least partly respon-
sible for the particular transport phenomenon.

Somewhat erroneously, perhaps, we sought a
clarification of this so-called ‘‘forced-flow”’
problem in higher kinetic theory approximations.
The approach was fruitful in an indirect manner
in that it led to the discovery of the proper
procedure of combining diffusive and viscous
flows, and prebably to a complete solution of the
gas transpoit problem.

Stated briefly, one can separate (mentally)
motion in a gas mixture into a diffusive flux and
a viscous flux. What we have learned, however,
is that the total flux is simply the zum of these
two fluxes; that is, there is no direct interaction
between viscous and diffusive flows; the only
““cross term’’ of consequence can be absorbed
into a change in the pressure-diffusion coeffi-
cient.? Although this simple additivity relation
had been used previously by others, its theoretical

 

°S. Chapman and T. G. Cowling, Proc. Roy. Soc.
London, 8er. A 179, 159 (1941); V. Zhdanov, Yu.
Kagan, and A. Sazykin, Soviet Phys. JETP (English
Transl,) 15, 596 (1962).
® appears to have been unnoticed.

justification
To date, the theoretical reinvestigation of dif-
fusion in the presence of a pressure gradient has
been the most fruitful aspect of this work; a
semiempirical equation which had been derived
earlier has now been developed completely from
theory, with an explict expression for what had
been an empirical (and adjustable) coefficient.
Moreover, the new results lead directly to the
prediction of the Kramers-Kistemaker'? or Kirken~
dall'! effect, in which a pressure drop is gen-
erated when the net flux in a diffusing mixture
is zero. '

SOLID-STATE TRANSPORT PROCESSES IN
GRAPHITIC SYSTEMS

Recoil Phenomena

K. B. Evans IlI  J. L. Rutherford
R. B. Perez?

Utilizatien of oxide fuel kernels and multilayer
. pyrocarbon coatings enhances the efficiency of
coated fuel particles. Tendencies for fuel mi-~
gration, which are greatest during coaling op-
erations, are reduced by the use of oxide!®
fuels. Spearhead propagation is minimized'® by
applying a porous inner coating followed by an
outer coating with good retention properties. It
is not unlikely that the success of the multi-
layer be partially attributed to
sacrificial absorption by the inner coating of
recoiled fission fragments and knocked-on fuel

concept can

atoms.
For maximum absorption, ratios of inner-coating
thickness to recoil range should be greater than

unify. Our objective is the specification of this

 

1043, A. Kramers and J. Kistemaker, Physica 10,

£99 (1943).

By, p. McCarty and E. A, Mason, Phys. Fluids
3, 908 (1960).

IzConsultant, University of Florida.

R B. HEvans I, J. O. Stiegler, and J. Truitt,
Actinide Diffusion in  Pyrocarbons and Graphite,
QOENL-3711 (December 1964).

4R L. Hamner, GCR Program Semiann. Progr. Rept.
Sept, 30, 1965 (in press).

130, Sisman et al., GCR Program Semiann. Progr.
Rept. Sept. 30, 1965 (in press).

79

ratio for various carbonaceous materials. Thus
we are engaged in the determination of the dis-
tribution of fission fragments that have recoiled
to average positions within pyrocarbon specimens.
From the curves related to the distribution we can
extract values of the range R and a distribution
parameter associated with the recoil range,
called the straggling factor a. Although special
low-density pyrocarbon specimens have been
prepared for experiments to be conducted in the
near future, present experiments have been re-
stricted to General Electric pyrocarbon speci-
mens.

To initiate a series of experiments, pyrocarbon
specimens are bombarded with 2?3U% at 40 kev
to form fission-fragment sources.
specimens, accompanied by an uncontaminated
pyrocarbon target specimen,
thermal-neutron fluxes.
incrementally

The source

are exposed to

All specimens are then
and assayed by gamma
counting so that integral curves related to the
distribution of fragments can be con-
structed.  Applicable transport equations predict
that plots of F.R., the total activity remaining
in a specimen after grinding to a penetration z,
vs z can be approximated by a straight line over
most of the interval between zero and E. The
only contribufion of « ig reflected in the small
tail beyond /R = 0.95. Extrapolations to the z
value corresponding to F.R. = 0 will indicate the
true range value. Some resulis appear in Fig.
8.1.

Several blank experiments were performed using
2327, to establish the as-deposited condition of
the actinide on source specimens and to provide
a z = 0 correction for the range dala. DBased on
the tlank data, we conclude that the source layer

is

ground

various

very thin (an important requirement for data
correlation) and that the average layer position
is 0.8 u beneath the source-specimen surface.

As demonstrated by the recoil data in Fig. 8.1,
the range value for the light fragments (13.0 u)
is greater than the range value for the heavy
fragments (11.0 p); the ratio of the range values
for the two groups is 1.18.
conducted 1o
order to

Similar experiments
in air’'” reveal a smatic of 1.32. 1In
egtimate a value of the straggling-
factor/range-value ratio, results for three of our
best range experiments have been replotted in

 

16, D. Evans, The Afomic Nucleus, p. 668, McGraw-
Hill, New York, 1955,
CRNL-DWG 65-12434

 

   
 

 

 

 

 

1O e
RANGE VALUES (u)
EXP A-1 EXP. A-2
09 N (CLOSED PTS.) (CPEN PTS.)
' “O. 104 1.5 }11 o
v s %oe 2 10.7 H
% v 27 12.6 13.5 }
0813 ' ’ 13.
X oy 125 3.8 3.0p
= |
— x i
g 07|
o .
g .
g 06 |y - Ny ! : :
v \
&
> X ™ RANGE DATA
- x 4
o N '3 : ALL CURVES
< j . 1 REFERRED TO
g 04— NN (meos
G % ‘ 1 \‘ i
— X . WO
L &
© o3 7 BN
“ X A
Q
£ \
§ X *
£ 0%y SLANK DATA
//
%
OfF- %
X ;J( A.\ D DV
% \sl o
A, x “m a v O
0 ’f.:‘_’f_fl_‘_y.’});&":effixm!xwx_..__ xm........f,\l 73:&\1};?’?
C 0.2 04 06 08 10 12 t.4
Ir
Fig. 8.1. Comparison of Range Measurements for
Average Light and Heavy Fission Fragments from 235U

in General Electric Pyrographite. Data for five blank

experiments are also shown.

terms of concentration; that is, (A activity/Az)
Correlation of these plots with applicable
equations suggests an /R value of 0.126. We
note that the « and R values cited are automati-
cally averaged with

vs Z.

respect to the <a> and
<c¢> directions by the conditions of our experi-
ments, in which fragments enter the target at all
solid angles from 0 to 27 steradians. This point
has been verified experimentally using <a> and
< ¢ > direction specimens.

Actinide Diffusion

R. B. Evans III J. L. Rutherford
F. L. Carlsen, Jr.'7

Investigations of uranium and thorium diffusion
in graphite matrices have been carried out on

 

17F0rmerly with ORNL Metals and Ceramics Di-
vision; present address Stellite Division, Union Car-
bide Corporation, Kokomao, Ind.

80

High Temperature Materials pyrocarbon (IITM-
PyC) and General Electric pyrocarbon (GE-
PyC). Although superficial examinations suggest
similar structures and properties for these ma-
terials,!® we find a distinguishing feature of the
GE-PyC to be continuity of structure across
basal planes. The integrity of this structure is
retained at all temperatures up to 2400°C without
delamination from crystalline rearrangement and
growth.!? Under reasonable conditions of tem-
perature and low actinide concentrations, ac-
quisition of reliable <a> direction results {(never
obtained with HTM-PyC) is assured. This and
the fact that GE-PyC has been studied intensely

elsewhere 20

constitute prime reasons for our in-
terest in this material.

Present results evolved from two distinct ex-
periments in which actinide invaded pyrocarbon
from either a thin layer at low concentrations or
a constant-potential source at the maximum con-
centration C .- the apparent solubility. In each
case, determinations of D, and Di (coefficients
for diffusion in the <a> and <c> directions)
were attempted. Estimates of C_ ~ could be ex-
tracted only from constant-potential results since
surface concentrations in thin-layer experiments
diminish with time. Some average results obh-
tained at comparable diffusion times are pre-

sented in Table 8.1. Clearly, uranium diffuses

faster than thorium — particularly in the <a>
direction.
Additional thorium data reveal a DL<+C>/

D, <--c> ratio of 1.5 at 2065°C. This may be a
valuable clue for an interpretation of the well-
20,21 <c> direction diffusion
From both thoriuim and uranium diffusion
data (not shown), we found the coefficients to be
invariant with time over the

x/2+/Dt ranges investigated.

known nonuniformn

patterns,

temperature and
In other words,

18B5th G.E.

stratic-columnar

and HTM pyrocarbons possess turbo-
{sometimes called granular) strucs

tures which lead to high matrix densities (™2.2 g/cn12)
and abnormally high anisotropic ratios for many proper-
ties.,

Vg, 1. Carlsen, Jr., ‘*Effects of Pyrolytic-Graphite
Structure on Diffusion of Thorium,"” thesis submitted at
the University of Tennessee, issued as ORNL-TM-
1080 (June 1965).

20_]. R. Wolfe, D. R. McKenzie, and R. J. Borg,
The Diffusion of Non<Volatile Metallic Elements in
Graphite, UCRL~7324 (April 1964).

2R, B. Evans Il et al., Reactor Chem. Div. Ann,
Progr. Rept. Jan. 31, 1965, ORNI.-3789, pp. 193--07,
81

Table 8.1. Diffusion Coefficients and Related Results for Actinides at Low Concentrations

in General Eleciric Pyrocarbons (Columnar)

 

 

D Coefficients (cm'z/sec)

Uranium

 

 

 

Thorium
D-Lgx'Jf‘ C> DII D_L D”

Temperature, e

2065 2.1 % 10770 1.5 x 1078 3.0x 10710 2.2 x 1077

1865 2.0 x 10711 1.2x1077? 4.3 1071 1.4%x 1078

1697 1.0x 10712 6.8 x 10741 4.6 x 10772 1.6 x 107°

Related Results

[cl?, g/cm? 3.1 < 1074 2.8 x 10”4 2.1x 107% 5.0 x 10™7
AE®, cal/mole 1.2 = 10° 1.6 % 10° 1.0 x 10° 1.2 % 10°
AES, cal/mole 1.1 % 10° 1.4 % 10° 1.3 % 10° 1.2 x 10°

 

A verage actinide concentration, Q, (rsz‘)"l/z

bActivation energy for present results.

exXp '-—(n’)"l; Gy = 2 pglem”.

3

“Activation energy reported by J. R. Wolfe eof al.,, UCRL~7324 (1964).

ORNL-DWG 635-12433

TEMPERATURE (°C)

 

 

 

 

 

2400 2000 8GO GO0 1400 1300
T """“‘T‘""""""“”‘[]““"‘“"]’"'T""—"‘“‘
s ORNL
om0 [JORI,
. N |
3 e e N
2 oy e
~ N
o X
£ \
o
o
O
o
2

 

 

 

 

 

 

 

 

5.0

10,000/, o)

Fig. 8.2. Comparison of Uranium Diffusion Coeffi-

cients Reported for General Electric Pyrogrophite

(Columnar).

the coefficients are true constants, and structural
changes induced by high-temperature exposure
(sometimes catalyzed by high actinide concentra-
tions) do not occur at the conditions of these ex-
periments. Dramatic verification (see Fig. 8.2)
of this is evidenced by the fact that our uranium

 

coefficients, obtained at concentrations danger-
ously near the anticipated C value, show ex-
cellent agreement?? with uranium coefficients *°
obtained at practically zero actinide concentra-
tions (carrier-free 2°2U).

In opposition to thin-layer results, preliminary
constant-potential data suggest comparatively high

coefficients and low activation energies. How-
ever, these results may agree with the high-
concentration results?? for HTM-PyC. Con-
siderations of combined structural and

concentration effects might enable reconciliation
of the anomalous diffusion behavior as observed
at high and low concentration levels. It is clear,
however, that thin-layer data are not indicative
of fuel migration in carbide fuel-particle coatings
since the actinide concentrations
carbide interfaces are quite high.

at coating~

Self-Ditfusion

K. B. Evans III

Several dramatic changes occur in pyrolytic
carbons when subjected to temperatures above

 

22 . . ..

Similar comparisons indicate a less-than=excellent
agreement of thorium coefficients. ‘The magnitudes of
ceefficients reported by Wolle: ef al., are greater than
ours by a factor of 5; however, .D”/D ratios and AE
vajues are comparable, L
those at which they are deposited. At high tem-
peratures the structures ‘‘improve’ in that they
tend to assume a graphitic structure; values of
the thermal conductivity and impurity-atom dif-
fusivity decrease. The tate and mode of this an-
nealing process are diffusion controlled and are
best interpreted through self-diffusion measure-
ments.

In systems other than graphite, self-diffusion
measurements may be circumvented, since the
self-diffusion coefficient has been correlated with
other properties more amenable to experimenta-
tion.  Analogous correlations for graphite are
undetermined; the necessity of conducting the
formidable self-diffusion experiments still re-
mains, if for no other reason than to obtain cor-
relations of the type mentioned.

We have initiated a systematic study of the
role of self-diffusion in the annealing phenomena
associated with graphites and carbons. From
this information we hope to estimate diffusion

82

parameters for perfect graphite crystals via
extrapolation of data for improved structures. A
series of special thorium-pyrocarbon diffusion
experiments has been completed pursuant to the
selection of materials and annealing conditions
for use in the present investigation. Observa-
tions of actinide diffusion patterns are quite
valuable in surveys of this sort, since actinides
migrate preferentially along defective paths in
carbons. As a result of these experiments, sev-
eral possible carbon candidates were chosen for
our initial self-diffusion studies.

Equipment and techniques used to study dif-
fusion of actinides and fission product species in

graphites and carbons are being modified as re-

quired in order to pemmit study of '*C diffusion.
The development of methods for accurate place-
ment of tracer on surfaces and the demonstration
of certain means of assay for the diffused isotope
pose the most immediate problems.
9. Reactions of Reactor Components

with Oxidizing Gases

I.. G. Overholser

REACTIVITY OF ATJ GRAPHITE WITH LOW
CONCENTRATIONS OF OXIDIZING AND
REDUCING GASES

J. P. Blakely

Rates of reaction of a 1-in.-diam sphere of AT]
graphite with low concentrations of water vapor
and carbon dioxide in flowing helium (1 atm), ob-
tained from continuously recorded weight changes
and compositions of the effluent gases as estab-
lished by a sensitive gas chromatograph, have
been reported.’’? These studies have been ex-
tended {o gas mixtures containing both water vapor
Addition of catbon
dioxide to the water vapor increased the reaction
rates above those observed for water vapor alone,
but the rates found for the mixed oxidants were
less than the sum of the rates for the two oxidants.
Addition of hydrogen inhibited the reactions of
both the oxidants with graphite.

In view of the retarding effect of hydrogen on the
water vapor—graphite reaction noted earlier,’'?
similar studies were performed? to establish what

. . . - 3
and carbon dioxide in helium,

 

'L, G. Overholser and J. P. Blakely, GCR Program
Semiann. Progr, Rept. Sept. 30, 1964, ORNI.-3731, pp.
161-67.

2J. P. Blakely and L. G. Overholser, “Oxidation of
AT] Graphite by Low Concentrations of Water Vapor
and Carbon Diexide in Helium,’' Carbon (in press).

*L.. G. Overholser and J. P. Blakely, GCR Program
Semiann. Progr, Rept, Mar, 31, 1963, ORNL.-3807, pp.
150-35, .

.. G. Overholser and J. P. Blakely, GCR Program
Semiann., Progr, Rept, Sept, 30, 1965, ORNL-3885 (in
press ).

83

effect, if any, the addition of methane might have
on this reaction. Data obtained from these studies
are given in Table 9.1. Carbon removal calcu-
lated from CO, + CO found in the effluent gases
shows that the water vapor—graphite reaction was
retarded by the addition of methane, although to a
lesser extent than by an equal concentration of
hydrogen.? If only weight changes are detennined,
one might conclude that the oxidation of graphite
by water vapor was completely suppressed. Data
obtained from the effluent gases, however, show
that both oxidation and deposition occurred. The
calculated and observed weight changes agreed
satisfactorly in those cases where weight losses
or small weight gains were found. The calculated
weight changes were significantly larger than the
measured weight changes in those instances where
large weight gains prevailed. 'This behavior could
be due to spalling of carbon from the graphite sur-
face. Both the oxidation and deposition rates de-
creased with decreasing temperature and were not
measurable at temperatures much below 700°C.
Carbon deposition rates from methane-helium mix-
tures were determined in the absence of water
vapor to establish what effect water vapor might
have had on the cracking of methane. A compari-
son of the calculated deposition rates listed in
Table 9.1 with those given in Fig. 9.1 shows that
water vapor has very little, if any, effect on the
cracking of methane. The overall rates are dif-
ferent, however, in the two cases due te loss of
carbon by oxidation in the presence of water vapor.
The slopes of the plots given in Fig. 9.1 cor-
respond to activation energies of ™~ 50 kcal/mole
for all concentrations of methane examined. The
Table 9.1. Carbon Deposition from Methane —Water Vapor—He!lium Mixtures

HZO Concentration of 110 ppm

 

 

 

 

CH4 Concentration Loss and Gain of Carbon {mg em” % hr™h Observed
Temperature Flow Influent Effluent Effluent Concentration (ppm) Loss Gain Net Weight
o lem? (STP)/min] Gases Gases CO, CO TotalH, Corrected H, Calculated from Calculated from Gain or Change
{ppm) {ppm) C02 + CO Correcied H2 Loss (mg cm 2 hrkl)
% 1073 % 107 x 1073 x 1077
800 280 18 2 37 8.1 8.1 --8.0
BOO 275 142 132 15 <1 46 16 6.0 3.2 2.8 -3.9
800 300 270 235 i4 <1 46 18 6.1 3.0 -2.2 -1.8
800 275 335 320 i4 1 63 34 6.9 6.7 +3.7 +0.8
800 300 560 520 11 <1 62 490 4.8 8.6 +3.8 +2.0
830 300 635 630 9 <1 63 45 4.0 9.7 +5.7 +2.8
800 275 680 650 10 <1 78 58 4.0 11.1 +7.1 +4.7
750 285 290 280 © <1 20 S 2.5 1.7 —0.8 0.6
750 285 560 550 6 <1 29 17 2.5 3.5 +1.0 1.2
700 360 260 260 2 <1 13 6 0.9 1.3 +0.4 2.7

 

rg
ORNL-OWG 65-11488
TEMPERATURE (°C)

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_1850 800 750 700
| - .
O T T T T
}i%‘ """""""""""""""""""" I ]
o | T T T
T ............. b Lo .
f O 560 ppm ZH
) .\ ,,,,,,,,,,,,,, e _ ppmCHy L
I \\ [ ® 300 ppm CHy
T i A 120 ppm CHy,
\‘.f\ 1O~2 Ad.‘:.’xi, ”””””””””” ki T
. :
o™ 5
E
tud
*..
=
. 2
5
i
W
g 103
L
&
5
l
ol S R R
} |
. |
89 9.4 Q3 95 9.7 9.9 104 t0.3
10,000/, e
Fig. 9.1. Effect of Temperature on Carbon Deposition

Rates at Yarious Methane Concentrations.

apparent order of the deposition reaction with
respect to methane falls between 0.5 and 0.8 in
the temperature range of 750 to 850°C,

Studies® also were made of the disproportiona-
tion of carbon monoxide on AT] graphite, using
1501500 ppm of carbon monoxide in helium at tem-
peratures of 550 to 850°C. The complete lack of
agreement between deposition rates calculated
from the carbon dioxide found in'the effluent gases
and those obtained from measured weight changes
is not understood. Using values for the deposition
rates calculated from the carbon dioxide formed,
it was found that deposition rates increased with
increasing carbon monoxide concentrations. The
deposition rates at constant carbon monoxide con-
centration increased upon raising the temperature
from 550 to 750°C and then remained constant or
decreased slightly when the temperature was
raised to 850°C. The low rates found suggest that

85

AT] graphite is not an efficient medium for pro-
moting the disproportionation of carbon monoxide
under the experimental conditions examined.

COMPATIBILITY OF PYROLYTIC-CARBON.
COATED FUEL PARTICLES WITH
WATER YAPOR

C. M. Blood

Current developments indicate that fuel particles
coated with pyrolytic carbon will be present in all-
ceramic cores of high-temperature gas-cooled re-
actors. Inleakage of water vapor during operation
of such a reactor poses hazards because of the
high temperatures of the core components, Oxida-
tion of the fuel particle coatings could cause coat-
ing failures and a resultant release of volatile fis-
sion products into the helium coolant.

The compatibility of wvarious batches of fuel
particles with water vapor has been examined® in
an attempt to evaluate the hazards. In these
studies, fuel particles were exposed at 1000°C to
flowing helium--water vapor mixtures having partial
pressures of water vapor of 4.5, 45, and 570 torrs.
The fuel particles had various types of pyrolytic
carbon coatings and cores of UC, or (U, TH)C,.
The reactivity of the fuel coatings was determined
from weight changes and effluent gas composi-
tfions, The incidence of failure of the coatings
was obtained from the quantities of wranium and
thorium removed by acid leach following exposure
fo water vapor. Microscopic and metallographic
examinations were made following exposure to
Surface
area measurements were made on both oxidized

water vapor and also after acid leach.

and unoxidized fuel particles.

Reaction rates for different batches of fuel par-
ticles obtained at 1000°C using various partial
pressures of water vapor are given in Fig. 9.2,
These data show that the apparent order of the
water vapor—coating reaction increased with in-
creasing partial pressure of water vapor in virtu-
ally all cases. This could be due, in part, to dif-
ferences in burnoff at the various partial pressures.

 

C. M. Blood and L. G. Overholser, GCK Program
Semiann., Frogr. Rept, Mar. 31, 1965, ORNL.3807, pp.
140—42; GCR Program Semiann, Progr., Rept, Sepi. 30,
1965, ORNL-3885 (in press).
QRNL-DWG 55-11362

     

300 e ——— s o
200 - ' =
160 -
50
_ 20
=
> 10
o
E
1 5
z -
G
=z
o 2 - :
E LAMINAR 1
1 o LAMINAR I
e 1 m CRANULAR IV -~ 00
. 7 ISOTROPIC V
as - e ISOTROPIC VI T
4 [SOTROPIC VII .
: o NCC-217 :
01 ! [ . cea . oo - - . - -
1 2 5 10 20 50 100 200 500 1000
H,0 PARTIAL PREZSSURE (torr)
Fig. 9.2. Effect of Partial Pressure of Water Vapor

on Reactivity of Fuel Particles at 1000°C.

The incidence of failure of the coatings also may
be responsible for this behavior.
a few particles failed at a partial pressure of 4.5
torrs, whereas at 570 torrs appreciahle fractions
failed in all cases. If the core residues catalyzed
the reaction, this could account for the change in
apparent order observed. This behavior makes an
extrapolation to other partial pressures hazardous.

Surface area data obtained for oxidized particles
from the various bhatches of fuel particles were

In general, only

extremely variable. In general, no correlation of
Also,

no consistent relationship of incidence of failure

reactivity with surface area was possible.

of coatings to reaction rate, surface area, or burn-
off Photomicrographs of oxidized
particles from various batches showed that the
mode of attack varied from batch to batch; this
might be anticipated from the variable reaction
rate and surface area data.

Studies of the effect of support media for the fuel
particles on the reaction rate showed that rates
obtained with alumina or platinum were essentially
the same. Use of a graphite sleeve to support the
fuel particles reduced the attack of water vapor
on the coatings; the degree of protection increased
with increasing sleeve length, This suggests that a

was evident.

86

graphite body can afford protection for the particle
coatings but that the configuration of the fuel as-
sembly will determine, to a large degree, the extent
of protection obtained.

The studies are continuing, with emphasis being
placed on the effect of higher temperature (1200
to 1400°C) and of prior irradiation on the reactivity
with water vapor.

COMPATIBILITY OF METALS WITH LOW
CONCENTRATIONS OF CARBON MONOXIDE

J. E. Baker

In the BeO-graphite compatibility tests,® metals
are used as heat shields in an inert-gas atmosphere
containing ~ 200 ppm of carbon monoxide at tem-
peratures ranging from 400 to 800°C. Possible
deposition of carbon on the metal surfaces and the
resulting decrease in reflectance prompted a lab-
oratory study of the metals in an atmosphere ap-
proximating that prevailing in the engineering tests.

Specimens of molybdenuni, gold-plated stainless
steel, and mild steel were exposed to flowing
helium (1 atm) containing 250 to 300 ppm of carbon
monoxide at temperature intervals of ~100°C in
the temperature range of ~450 to 850°C. Expo-
sure times averaged ™30 hr at each temperature.
Weight changes were continuously recorded by a
sensitive analytical balance, and the influent and
effluent gases were analyzed by a sensitive gas
chromatograph. X-ray studies were made of the
exposed specimens in an attempt to identify any
deposits formed.

The molybdenum specimen gained ~10 pg/cm?
during an exposure time of ~150 hr. The weight
changes were so small that the effect of tempera-
ture could not be measured. The exposed speci-
men had a very thin dark-gray film and a few spots
suggesting localized attack. The film is probably
an oxide, which could have been produced by very
low concentrations of oxygen and/or water vapor
present as contaminants in the helium stream.
There was no evidence of carbon deposition.

The gold-plated stainless steel specimen showed
an overall weight gain of ~ 100 ug/cm? for an ex-
posure time of ~150 hr. The rate of weight gain

®C. A. Brandon and J. A. Conlin, GCR Program
Semiann. Progr. Rept, Sept. 30, 1964, ORNI.-3731, pp.
202 --6.
varied by about a factor of 2 over the temperature
range examined, with no consistent effect of tem-
perature evident. The exposed specimen had a
dull green color, suggesting failure of the gold
plate.  X-ray analysis showed the film to be
Cr,0, with no carbon present.

A bright specimen of mild steel lost ~ 170 pug/cm?
during ~ 150 hr. The rate of weight loss increased
with temperature. The surface was bright at the
end of the test and showed no evidence of carbon
deposition. A rusty specimen of mild steel lost
~ 1500 pg/cm? during ~150 hr, with no consistent
effect of temperature on the rate of weight loss

87

evident. The surface brightened during the ex-
posure, and x-ray analysis found only FeO present
in the surface film. Since the original specimen
had Fe O, and Fe O -H,O present, carbon monox-
ide reduction of the oxides seems responsible for
the large weight losses.

The results indicate that carbon is not deposited
on these metal surfaces in this temperature range
from low concentrations of carbon monoxide in
helium containing essentially no hydrogen or water
vapor. The situation could be different in the
presence of the latter.
10.

Irradiation Behavior

of High-Temperature Fuel Material

QOscar Sisman

FISSION.GAS RELEASE FROM PYROLYTIC-
CARBON-COATED FUEL PARTICLES

P. E. Reagan J. G. Morgan
J. W. Gooch T. W. Fulton
C. D. Baumann

2 we showed that ra-

In previous experiments'
diation damage to the pyrolytic carbon coating
was due primarily to fission fragments originating
at the core-coating interface. This damage can
be alleviated by providing a gap at the core-~
coating interface, by a porous carbon layer, or by
a sacrificial pyrolytic carbon layer. Particles
of dense UO2 have been coated (in the ORNL
Metals Division) by techniques
which provide one of these protective measures.
We have measured fission-gas release rates from
these three kinds of coated particles during
irradiations in the A9, B9, and Cl facilities in
the ORR.? The irradiation temperature, burnup,
and fission-gas release data are given in Table
10.1.

Particles in batch OR-298 had heen heat treated
to introduce a gap between the uranium oxide core
and the inner layer of pyrolytic carbon. The coat-
ing on these particles consisted of an anisotropic

and Ceramics

 

Iy, E. Reagan, F., L. Carlsen, and R. M. Carroll,
“Fission~-Gas Release from Pyrolytic Carbon Coated
Fuel Particles During Irradiation,’”’ Nucl. Sci. Eng.
18(3), 30118 (1964).

’p. E. Reagan, J. G. Morgan, and O. Sisman, “Fis-
sion-Gas Release from Pyrolytic Carbon Cgated Fuel
Particles During Irradiation at 2000 to 2500°F,* Nucl.
Sci. Eng. 23(2), 210—23 (1965).

3p. E. Reagan ef al., **Fission Gas Release from
Coated Particles,’”® GCR Program Semiann. Progr.
Rept. Sept. 30, 1965 (in press).

88

J. G. Morgan

inner layer and a granular outer layer. After a
little more than 26% bumup at 1400°C, a few
bursts of fission gas and the
fractional release (see Table 10.1) increased by
a factor of about 7.

Three experiments were conducted using par-
ticles with a porous carbon buffer layer between
the uranium oxide core and the pyrolytic carbon
coating. Particles from batch OR-354 had a
relatively thick (82 u) isotropic coating over the
The fractional fission-gas release,
as shown in Table 10.1, remained very nearly con-
stant throughout the test. Particles from batch
OR-348 were similar to those of OR-354 (porous,
isotropic on an oxide core), but the isotropic
coating was more dense. Initial values for frac-~
release from the OR-348 particles were
near 10~ %; the values increased to near 107 by
the end of the test. No bursts of fission gas were
observed. The third batch of particles (batch
OR-HB23) with a porous inner layer were made
with sol-gel uo, cores;* these were coated with
an isotropic layer next to the porous casbon
buffer layer and a granular outer layer. These
particles operated for 25 at. % heavy metal burnup
at 1600°C. The fission-gas telease rates were
quite low (see Table 10.1); they increased by
less than a factor of 2 during the entire test.

Particles from batch OR-343 were made with an
inner coating which would shrink away from the
uranium oxide core. These particles had a low-
density (~1.5) pyrolytic carbon coating applied
directly to the core, followed by a higher-density

were released,

porous buffer,

tional

4J. P. McBride, *‘Preparation of UO  Microspheres by
a Sol-Gel Technique,’’ ORNL~3874 (in"press).
Table 10.1. Fission-Gas Release from Pyrolytic-CarbonsCoated Uronium Oxide Fuel Particles During Irradiation

 

Average Fractional Fission+Gas Release, R/B®

 

 

 

Capsale Batch Coating Burnup Temperature 5 ; _ )}
(ate % U) {°C) MK B8 87y 13ixe Piixe
AY-2 OR-298 Gap + 27.9 1400 9.3% 107" 7ox1077"  57x10777 11,7 x1077F 6sx 10778
anisotropic + 8.5%X107°%° 49x107%  3.4x107% d o
granular
BY-26 OR~354 Porous carbon + 12.1 1350 4.7 % 1079 4,2 % 1078 3.9 x107° 2.7x107% 17 x 10”8
isctropic
‘BO=27 OR=348 Porous. cathon + 5.4 1580 16X 1670 1.3x 107" 1.2 x 1078 1.0 x 10~° 0.6 x 10~ F
dense isotropic
Clel6  OR-343 Isotropic I + 11.9 1400 t1x1077  3.8x1077  06x 1077 0.9x1077 0.4 x 1077
isotropic It 2.8 1500 1.9 %1077 1.4 x 1077 1.2% 1077 1,7 x 1077 g7 x1077
13,7 1400° 1.1 x 1977 0.8 % 1077 9.7 x 1077 d d
C1-17  OR-HB23  Porous carbon + 25.2 1600 6.2x107%  a3x107%  3.0x1078 7.6 x107%  s.4x1078
isotropic +
granular
fRelease rate /birth rate; average for last week of test unless otherwise noted.

PBefore bursts of fission gas,

CAfter bursts of fission gas.

dData not available.

€After 2.8% buenup at 1500°C.

6%
    
 
 

CRNL-DWG 65-118435
ST OIS GA-3G of ¢
T Qo

>

T aggee -

    

 

CUTTTTT B9-26 OR-354 g 11
LT 43500¢ ""1“‘7%“ P
L . \/« ,,%,77‘,7?,,‘ .

  

 

 

 

L B00°C e
2 L . e e
1078 U e -
1o’ 2 5 1978 2 5 107>
CORTAMINATION, BASZC OM ALPAA COUNT (grams URANIJM]
Fig. 10.1, Relationship Between 98Kr Relcase and

Cooting Contamination.

(~1.8) pyrolytic carbon outer layer. As shown in
Table 10.1, the fractional fission-gas release
nearly doubled when the temperature was increased
from 1400 to 1500°C, but it returned to about the
original value when the temperature wasg again
decreased to 1400°C.

We observed that a correlation could be made
between the degree of contamination in the coat-
ing and the fission-gas release from particles
with unbroken coatings (see Fig. 10.1), This
leads to the conclusion that (below 1400°C) the
fission-gas release from unbroken coatings is
predominantly from contamination in the coating
rather than from the fuel core.? We have found
that coated oxide particles have less contami-
nation in the coating than coated carbide par-
ticles.

POSTIRRADIATION EXAMINATION OF COATED
FUEL PARTICLES

P. E. Reagan E. L. Long, Jr.5

The irradiated uranium oxide particles described
in the previous section were examined with a

 

5l‘.h:,-tal's and Ceramics Division.

90

microscope (at 30x), and about 100 particles from
each experiment were selected for metallographic
examination.? The particles made with an in-
tentional gap at the core-coating interface (batch
OR-298) revealed no failed coatings, although
their behavior during irradiation suggested that
a few particles did fail. Metallographic exami-
nation (see Fig. 10.2) showed that the inner
pyrolytic carbon coating had undergone severe
delamination but that the outer coating had re-
mained intact. No reaction product was obsetrved
at the core-coating interface.
cores showed small metallic inclusions in the
grain boundaries and a collection of fission-gas
bubbles,

Postimadiation examination of the uranium oxide
patticles from batch OR-354 revealed that only
minor microstructural changes had occuired as a
result of the irradiation. No failures or evidence
of potential failures was noted. The only changes
apparent densification of the
porous carbon inner coating and the presence of
fission-gas bubbles and small metallic inclusions
No
coating failures and only minor microstructural
changes were noted when irradiated particles from
batch OR-348 were examined. The only change in
the coating was a continuous gap that formed at
the core-coating interface. The third batch of
particles with a porous inner layer (OR-HB23)
had been irradiated to 25 at. % uranium burnup at
1600°C. An apparent densification of the porous
catbon layer, as shown in Fig. 10.3, was the only

The urasium oxide

observed were

in the grain boundaries of the uranium oxide.

microstructural change noted in these coatings
after irradiation.

The particles of batch OR-343 were coated in

~an experiment to determine whether a low-density

coating applied directly to the UO, core would
shrink during high-temperature irradiation and
provide the gap necessary to relieve the stress
on the coating. This did indeed occur, as is
shown in Fig. 10.4. Although about 20% of the
coatings showed short fractures spiraling into the
inner coating, no failed coatings were observed.

The advantages of the UO2 cores over UC
cores became more obvious when it was observed
that the U0, wonld not convert to UC, as long as
the coating was intact. From our experience with
a variety of coated UO, particles, we conclude
that they are superior to the UC, particles for
the following reasons: (1) the oxide does not
flow at high burnup and expand into voids or
91

PHGTO 282346

E

Y.58967

L

ot a et
e
X

e
S

i ""&és’w:' T
"%?';'-&, "'5.9 o

 

 

 

 

 

1
-

T IS
T
i

L

!

 

 

 

Fig. 10.2. Pyrolytic-Carbon-Coated Uranium Oxide Particles from Batch OR-298. Magnification 200x: (a) unir-
cadiated; (b} irradiated to 28% burnup at 1400°C in capsule A9-2.
92

PHOTO 82347

Y-62437

 

 

 

 

 

15

T

 

 

 

Fig. 10.3. Pyrolytic-Carbon-Coated Uranium Oxide Particles from Batch OR-HMB23. Magnification 200x:
(a) unirradiated; (b)irradiated to 25% burnup at 1600°C in capsule C1-17.
 

PHOTO 82348

93

e

2

...fiu’tufl”uu

 

 

 

 

f”%r

!

S
e

e

S
wm unw

I

 

ey

/.Nmm.hv
%

ication 200x: {a) unir-

nif

Mag

icles from Batch OR-343.

ium Oxide Part

Coated Uran
burnup at 1400

Carbon-

Pyrolytic-

4.
)

. 10

Fig

le C1-16.

in capsu

C

C and 3% at 15007

iated to 12%

adiated; {&) irrad

-
1
cracks as the carbide does, (2) uranium from the
oxide will not diffuse into the pyrolytic carbon
coatings even at high temperatures, (3) uranium
contamination in the coating can be kept to a
much lower level during fabrication with oxide
cores, and (4) the oxide is not reactive and is
much easier to handle during fabrication of the
coated particle.

POSTIRRADIATION TESTING OF COATED
FUEL PARTICLES

M. T. Morgan R. L. Towns
C. D. Baumann

Annealing studies of irradiated coated fuel par-
ticles were continued to determine the stability
of the coatings at temperatures higher than that

94

of the irradiation and to measure their ability to
contain fission products. The particles tested
had fuel cores of UC,, (Th,U)C , or UO, with
duplex or triplex pyrolytic carbon coatings. They
were heated to temperatures up to 2000°C in a
flowing helium atmosphere; the fuinace com-
ponents were periodically removed for analysis to
measure fission product release as a function of
time and temperature. ®

The release rates obtained in five annealing
experiments at 1370, 1700, and 2000°C are com-
pated in Table 10.2. The values shown were
averaged over 6 hr of anneal following an initial
1-hr heating period at each temperature. A graph
of the cumulative releases vs time for barium and
strontium from three batches of particles is shown
in Fig. 10.5.

 

5M. T. Morgan, Reactor Chem. Div. Ann. Progr. Rept.
Jan. 31, 1965, ORNL-3789, pp. 212~13.

Toble 10.2, Fissjon Product Release Rates fream Pyrolytic-Carbon-Coated Fusl

Particles During Postirradiation Annealing

 

 

Annealing

T

 

 

Sample® - Fission Product Release Rates (%/hr)
emperature
/ No. (OC) 140Ba 898r 144Ce 13'.7CS 91Y
______ ; < < -
B9-20 1370 0.1 0.3 =0,003 =0.03
B9-21 0.1 0.2 So0.01 0.01
C1-15 0.1 0.04 0.04 0.02
< <
B9-20 1700 0.8 0.6 =0,002 =01
Bo-21 0.6 2. 0.04 0.02
Cl-15 0.5 0.4 0.3 0.08 0.3
Cl-l6 6. 2. 0.0006 0.3 0.02
BO-20 2000 2. 2. 0.03 §0.2
BOo-21 6. 7. 5 0.3
Cl-15 3. 5 3. 1 3
C1l-16 10. 7 0.005 7 0.01

®B9-20 and C1-16 samples are duplex-coated UO2 with uranium burnups of 4.6 and 14.7% respectively. Cl-16 is

the sample with the loweredensity outer coating. B9-21 and Cl-15 are from the same batch of triplexscoated (Th,U)C2

particles, but with burnups of 0.29 and 8.9% respectively.
ORNL—DWG 65 ~123486
CAPSULE B92-2G, BATCH OR - 201 U0, CORE ~Ba
Sr

CAPSULE B9-21, BATCH GA-314,(Th,U) G, CORE - rE;!cj
’ 3r

CAPSULE C{~15, BATCH GA-314,(Th U] C, CORE--Ba
ANNEALING TIME (nr)

P OHE O8

Sr

FISSION - PROCUCT RELEASE ()

 

 

 

 

 

 

 

0 S0 20 30 :-10 50
SQUARE ROOT OF ANMEALING TIME (min2) '

Fig. 10.5.

During Postirradiation Anneal of Coated Fuel Particles

at 2000°C,

Accumulated Fission Product Release

The experiments so far have been performed
with fuel materials of current interest. Though
no systematic and basic study in fission product
release has been made, some trends are evident.
The behavier of triplex pyrolytic-carbon-coated
(Th,U)Cz particles shows some change with
burnup. Particles irradiated to 8.9%
burnup suffered 2% coating failures during the
initial hour at 1700°C or at 2000°C, while par~
ticles irradiated to 0.29% uranium burnup had no
failures during 19 hr at 2000°C. The lower re-
lease of barium and strontium at the higher burnup
remains without explanation.

Particles of UQ_ with an outer coating density
of 1.8 g/cm? had higher release rates for barium,
strontium, and cesium, but lower release rates
for cerium and yttrium than did {Th,U)C2 particles
with outer coating densities of about 2 g/cm?.
The higher release rates of barium, strontium, and
cerium may be due to the lower-density coating,
while the lower release rates of cerium and
ytirium in the UQO_ particles may indicate better
retention by the fuel core itself.

uranium

POSTIRRADIATION EXAMINATION OF
FUELED GRAPHITE SPHERES

D. R. Cuneo H. E. Robertson
J. G. Morgan C. D. Baumann
' E. L. Long, Jr.

We have completed the evaluation of five irra~
diation experiments that contained fueled graphite
elements in the form of spheres. These elements
were fueled with pyrolytic-carbon-coated particles
400 p in diameter dispersed in a graphite matrix.
A summary of the irradiations is shown in Table
10.3.  All of the spheres reported in the table
were 6 cm in diameter except those in experiment
8B-5, which were 1% in. in diameter. The spheres
of interest to the German AVR Pebble Bed Re-
actor Program were fabricated with unfueled outer
shells of machined graphite or shells molded
around the core matrix. Two of the spheres were
completely fueled (no unfueled shell).

In addition, we examined two spheres of interest
to the TARGET program. These were spheres of
graphite that contained several drilled holes into
which were poured loose coated particles. Ex-
periment 8B-5 contained eight 17 -in.-diam spheres
which underwent the highest burnups of any
spheres tested (™25 at. % of the heavy metal).
Metallographic examinations were carried out for
four spheres in this experiment; the remaining
four spheres were duplications of those examined
and were used for compression testing only. We
observed a high percentage of failure of the
laminar-coated particles in two spheres of this
experiment. Laminar-coated particles fabricated
by the same manufacturer but irradiated to about
1,45 the burnup were undamaged. Laminar-coated
(U,T‘ry)if,2 particles made with normal uranium
were found te have ceating fractures at the fuel-
coating interface and evidence of loss of crystal-
line detail in the fuel core. This is the only
case of damage we have observed in particles
containing normal uranium. Figure 10.6 is a
photomicrograph of a mnormal and an enriched
(U,Th)Cz particle from this experiment. We
found no broken coatings in the spheres that con-
tained either duplex- or triplex-coated fuel par-
ticles. Production-run Carbon Products Division
spheres, experiment O1A-8, operated successiully.
No damaged particles were found, and the non-
volatile fission products found in the graphite
 

Fig. 10.6.

96

 

CHES

g N

A

~

 

 

Pyrolytic-Carbon-Coated (Th,U)C2 Particles in Sphere Mo. 1 from Experiment 8B-5. The enriched

particles were from batch 3M-120; the normal particles were from batch 3M-119, The enriched particle (right) reached

a burnup of 25.6% heavy metal, but the normal particle (left) reached a burnup of approximately 0.2% heavy metal.

powder packed around the spheres showed a
relatively small amount of migration through the
unfueled shells of the spheres,

Relationships compression strength and
shrinkage as related to types of sphere manu-
facture are given in Table 10.4 for spheres itra-
diated to high burnups in experiment 8B-5. All
spheres had molded fuel cores. The two
spheres that were completely fueled (no unfueled
shell) underwent large shrinkages in diameter
and an apparent gain in compression strength.
The two spheres with molded unfueled shells
were found to exhibit the same degree of shrink-
age and, for one sphere tested, a decrease in

for

the

The machined unfueled
had the least outside diameter
However, from the large losses in
compression strength (irradiated vs unirradiated
spheres), we conclude that the molded cores
underwent considerable shrinkage, so that the
compression tests reflect only the strength of
the unfueled shell. Such loss of contact between
the core and shell adversely affects the heat
transfer from the core as well as the mechanical
strength of the sphere. The same trends can be
shown for 6-cm-diam spheres, although not to this
extent because of their low burnup.

compression strength.
shell spheres
shrinkage.
97

Table 10.3, Fueled Graphite Sphere ltradiations

 

 

ORR Fuel Particle Type of Sphere Estimated a Burnup Percent
Fxperiment Composition Coating Shell Temperature (% heavy 'metal) Failed b
No. (°C) Particles
AVR-type
spheres
8B-5 (U,Th)Cz, Laminar 1 molded, 700-~1000 25.6-27 67100
2 spheres M) 1 machined
8B-5 (U,Th}Cz, Triplex Machined 800 26.6 D
1 sphere (GA)
8B-5 (U, Th)C , Duplex No shell 800 27. 0
1 sphere (CPD)
O1A-8° uc,, Duplex Machined 7501200 9.1-10.8 0
3 spheres (CPD)
O1-8 ucC 2 Duplex Machined 700 2.1 Q
1 sphere (ORNL)
Q8-8 ucC 2 Puplex Machined 750--1200 3.5-4 0
2 spheres {ORNL)
08-8 uc o Duplex Molded 900 4.2 0
1 sphere (3M)
05-8 (U,Th)CZ, " Laminar Molded 900--1200 7.3-8.6 0
2 spheres (3M) '
TARGE T-type
spheres
O1-8 (U,Th}CZ in Duplex Solid sphere 7751200 2.1 0
18 holes, with 19 holes
UO2 + ThO2
in one hole;
1 sphere
o1-8 uc 2 Duplex Solid sphere 850 2.2 0
1 sphere (CPD) with 18 holes

 

AWhere two temperatures are given, the second is central temperature; only one

with central thermocouple.

bBy metallographic examination.

CAVR productionerun spheres.

sphere per experiment is equipped
98

Toble 10.4. Shrinkages and Compression Strengths of Graphite Spheres lrradiated
to Burnups of 25 to 27 at. % in Experiment 8B-5

 

 

Unfueled Shells of Spheres Average
Diameter
Thickness Shrinkagea
Type (in.)
' (%)
Machined 0.2 0.85
Machined 0.2 0.77
Machined Q.2 0.61
Machined 0.2 0.44
No shell 0 1.49
No shell 0 2.32
Molded 0.2 1.54
Molded 0.2 1.38

 

 

Unirradiated
Compression Loading Equivalent Spheres,
to Failure (1b) Compression loading

toe Failure (1h)

 

1725 (2)°
900 1500 (2)
225 1725 (2)
275 1500 (2)
2650 1740 (4)
3050 1740 (4)
1250 2200 (3)

2200 (3)

 

FAverage of readings taken at pole, equator, and temperate regions of spheres,

Number in parentheses indicates number of unirradiated spheres tested. Value given for lecading to failure is an

average.

POSTIRRADIATION EXAMINATION OF EGCR
FUEL ELEMENT PROTOTYPE CAPSULES

M. F. Osborne
E. L. Long, ]r.5

H. E. Robertson
J. G. Morgan

Except for one EGCR prototype capsule which
is still being irradiated, all the elements in this
series have been examined. Of those elements
irradiated in the ETR,’ six were examined vis-
ually and three were evaluated in detail, The
capsules coantained UQ, pellets fabricated at
ORNL; these pellets were either solid, hollow,
or hollow with a BeQ bushing. The design power
rating during irradiation was 35,000 Btu hr™!
ft—!, and the stainless steel cladding tempera-
ture varied from 1250 to 1550°F. The fuel re-
ceived burnups of from 4500 to 14,300 Mwd per
metric ton of uo,. Only one element experienced
cladding failure. The other elements had only
minor cladding deformation such as circumferential
ridges at pellet inteifaces and a collapse of the
cladding against the fuel. These effects were
observed in previous tests.?®

The element with cladding failure contained
solid UO, pellets. A longitudinal tear in the

cladding was apparently caused by overpowering
of the element early in the irradiation. Grain
growth in the cladding at the site of failure indi-
cated the presence of a hot spot. Unlike earlier
instanceg of cladding failure in this series, there
was no evidence of nitride formation in the clad-
ding or columnar grain growth in the fuel.

We have completed the evaluation of an EGCR
piototype capsule which had been irradiated in
the ORR. This element was of special interest
as it contained actual EGCR production-run UO,
pellets and 304H stainless steel cladding. The
irradiation was carried out at a heat rating of
31,000 Btu hr™! ft~! and a cladding temperature
of 1300°C. The dished-end hollow fuel pellets
achieved a burnup of 10,000 Mwd per metric ton
of UOz. Under these conditions, the cladding
had collapsed around the fuel but no circumferen-
tial ridges were formed. Bowing of the element
was nominal, and there was no shifting of the
fuel. The element had not failed, and its per-
formance was satisfactory under EGCR design
conditions.

 

M. F. Osborne et al., Reactor Chem. Div, Ann.
Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 218--19,

8c. D. Baumann, Irradiation Effects in the EGCR
Fuel, ORNL~3504 (June 1965).
1. Fission-Gas Release

R. M. Carroll
Oscar Sisman

G. M.

EXPERIMENTAL

From the resuits of in-pile experiments on single-
crystal UO, specimens, we have concluded that
fission gas is not released by classical diffusion
We conclude that the release is con-
2,3 These observa-

PIOCESSEes.
trolled by a trapping process,
tions are consistent with those by other experi-
menters?'® and have led us to a defect-trap theory
of fission-gas release.® This theory postulates
that a defect in the UO, crystal structure will trap
migrating fission gas; the rate of fission-gas es-
cape, accordingly, is controlled by the number of
traps present. Gas-trapping defects consist of in-
herent flaws, such as grain boundaries and internal
pores, in the UO, as well as point defects and
clusters of point defects created as a consequence
of the fission process. Steady-state experiments
with single crystals’ and with polycrystalline
UO, (ref. 8) have supported this defect-trap theory.

 

l¢onsultant from the University of Florida.

’R. M. Carroll and Oscar Sisman, Nucl, Sci. Eng. 21,
147--58 (1965).

5R. M. Carroll and Oscar Sisman, Fuels and Materials
Development Program Quarterly Progress Report, Septem-
ber 1965 (in press),

TR, M. Carroll, ‘““The Behavior of Fission-Gas. in
Fuels,”” AIME: Conference on Radiation Effects, Septem-
ber 1965 (1o be published by AIME). '

SR. M. Carroll, Nucl. Safety 7(1) (in press).

‘R, M. Carroll, R. B. Perez, and Oscar Sismun, J. Am.
Ceram. Soc. 48(2), 5559 (1965).

"R, M. Carroll and P. E. Reagan, Nucl, Sci. Eng, 21,
14146 (1965).

3R. M. Carroll and Oscar Sisman, “‘In-Pile Fission-
Gas Release from Fine-Grain UO,,"" J. Nucl. Mater. (in
press). -

During Fissioning of UO,

T. W. Fulton
R. B. Perez!

Watson

99

We have modified the experimental assembly to
permit a slow, controlled oscillation of the speci-
men through the flux of neutrons.”? '? These modi-
fications permit a controlled sinusoidal variation
of neutron flux (fission rate) or specimen tempera-
ture. The fission gas released during the oscilla-
tions is monitored continuously by a gamma-ray
spectrometer. Time dependence of the temperature
(or the fission rate) and of the fission-gas release
rate is measured and recorded by punch-tape read-
out.'1-12 A computer program has been developed
to analyze the fission-gas release waves in terms
of their Fourier components; this analysis yields
amplitude and phase-shift information as a function
of frequency of the oscillation.

The effect of temperature oscillations was ob-
tained by comparing the steady-state release (at
zero frequency) with that at different frequencies.!?
The gas release was found to increase as the fre-
quency of oscillations increased; at very slow
oscillations the release rate approached steady-
state levels. This result is predicted by the defect-
trap model; '? that is, as the oscillation frequency

 

IR, M. Carroll and Oscar Sisman, Trans. 4Am. Nucl.
Soc. 8(1), 22 (June 1965), accepted for publication in
Nuclear Applications.

]'OR. M. Carroll and Oscar Sisman, Fuels and Materials
Development Program Quarterly Progress Report, Septem-
ber 1965 (in press).

"R, M. Carroll et al., GCR Program Semiann. Progr.
Rept, Mar, 31, 1965, pp. 8087, ORNI.-3805.

IR, M. Carroll et al., GCR Program Semiann. Progr.
Rept, September 1965 (in press).

3R, B. Perez, Trans. Am. Nucl. Soc. 8(1), 22—23
(1965); to be published in Nuclear Applications.
EMISSION {atoms fsec)

-
o

88y,

 

 

ORNL-DWG 85- 4305

 

 

3 4 (x10'3)

NEUTRON FLUX (neutrons fem? - sec)

Fig. 11.1. Knockout Release During Fission Rate Oscillations.

increases, the frequency-dependent factors over-
come the trapping effect; the release rate in-
creases and approaches that predicted by models
hased upon classical diffusion theory.

Fission-gas release at temperatures below 600°C
occurs primarily by a recoil process wherein fis-
sion fragments passing through the surface of the
UO, specimen eject or ‘‘knock out’” UO, molecules
along with any fission-gas atoms in the knockout
zone. Knockout release during fission-rate oscil-
lations at a constant temperature of 575°C is
shown in Fig. 11.1. The data points of Fig. 11.1
form a sort of hysteresis curve; arrows by the
data points indicate whether the fission rate was
increasing or decreasing when the data were ob-
tained, The hysteresis curve shows the gas re-
lease rate to be higher during fission-rate oscilla-
tions than when the fission rate was constant.

We suggest that the data of Fig. 11.1 reflect the
rhythmically changing concentration of fission-gas
traps created by the fission process.
destruction and creation of traps allow the fission
gas to reach the specimen surface in surges. These
surges account for both the hysteresis and the in-

The cyclic

creased amount of gas at the specimen surface
available for knockout.

MATHEMATICAL MODEL

Material balance equations for fission-gas re-
lease {parent and daughter nuclides) from a thin
slab of fissionable material are based on produc-
tion, loss, and diffusion-leakage terms. In general:

Rate of change of concentration
— diffusion-leakage contribution -- loss terms

+ production terms .

For the diffusion-trapping model the loss terms in-
volve radioactive decay and trapping by intrinsic
flaws and point defects. The production terms in-
volve the fission rate, fission yield, and, for the
daughter, radioactive decay of the parent. The
concentration of point defects is a balance be-
tween their rate of formation as a consequence of
fission and their destruction by annealing at high
temperature.
101

Three material balance equations (for which the
symbols are defined in Table 11.1) result; they
are coupled and nonlinear.

oM 3?2

Parent: e D, 57T (A + {20)

— BN, (B | M(z,0) + B, F(D) .

oH a2
Daughter: =5 D, Pl Ay + 8,)

— hiV () | H{2,1) + B, F(O + A, M(z 1) .
o,
Point defects: m?zf—i =aF() —v.N, .
; ,

At steady state (OM/dt = OH/dt = ON /dt = 0)
the equations become linear and can be solved by
conventional methods. The steady-state solutions
for the release rates yield expressions which are
similar in appearance to the expressions obtained
from the old diffusion model,’? except that the
decay constant, A, is replaced by an effective
decay constant, A’, which includes the effect of
trapping. An increase in temperature will increase
the rate of trap annealing and will decrease the
effective decay constant, A",
tron flux will increase the value of the A”, will tend
to neutralize the increase in production, and will
make the release rate fairly insensitive to flux

An increase in neu-

Table 11.1. Definition of Symbols

 

 

Symbol Definition

M, H Atomic concentration of parent and daughter
nuclides, respectively, atoms,/cm3

DW’ 1 Diffusion coefficients, cmz/sec, = DG[EXP

(~AE/RT))
Fission vields, fission fragments /fission

24 Rate constant for trap annealing = L/O[exp
(-“AE“/R’I')]

a Traps formed per fission

gfi Trapping probability, due to permanent de-
fects, s.ec'"1

er Concentration of points defects, defects/cm3

R Second-order rate conslant, (cm3)/(dei‘ect)
(sec)

 

changes. This behavior agrees qualitatively, at
least, with the observed phenomena at steady state.

An approximate solution has been obtained for
the case where the flux is varied sinusoidally
while the temperature remains constant. To com-
pate the diffusion-trapping model with the simple
diffusion model, the analytical expressions for the
total release rate transfer functions for each model
were coded for numerical computations. The ampli-
tude of the release rate transfer function vs fre-
quency is shown in Fig. 11.2a. When the diffusion
model is used this amplitude decreases with the
square root of the frequency; when the diffusion-
trapping model is used this amplitude exhibits an
extended plateau in the low frequency range. As
the frequency increases, the frequency-dependent
factors overcome the trapping effects and both

ORNL-DWSG 85~1165

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

-3 b /T,-"“‘a/ e L
- 1 7
o / i P
o } ! A7
= 30 L b . Al
[ | * o
-,
L e DIFFUSION-TRAP THEDRY ~
& e DIFFUSION THEORY /{/’i’
E)J ~2Q b b R /./ . -
< 2
& &
; A |
O =00 S
| | | |2 T=100 °C ‘
{ ! ; j
b=
w2 5 1073 2 5 107E

o [radians x sec™)

Fig. 11.2. Release Rate Transfer Function vs Fre-

quency.
models tend to coincide. For lower frequencies
the trapping effect predominates; hence the fre-
quency-dependent terms do not significantly affect
the magnitude of the transfer function.

Equivalent comparisons were made with the phase
shift (Fig. 11.2b) of the transfer function. In the
diffusion model, the phase shift quickly reaches
an asymptotic value of —45° and is independent of

102

temperature. The phase shift for the diffusion-
trapping model tends
slowly toward the --45° asymptote, and is a func-
tion of temperature.

Data have been obtained by oscillating the tem-

is always smaller, more

perature at constant flux; for this case, the equa-
tions do not linearize readily and the solution has
not been completed.
12.

EQUILIBRIUM STUDIES IN THE SYSTEM
ThO,-U0,-UO,

L. O. Gilpatrick C. H. Secoy

Equilibrium studies of the system ThO,-U0,-UO,
were continued at partial pressures of O, equal
to that found in the atmosphere and at temperatures
ranging from 1200 to 1550°C. Previous work has
established the general features of the phase dia-
gram for the system. '—*

A study was made to determine if the scatter in
analytical compositions could be due to experi-
mental techniques. A temperature of 1550°C and
a composition of 90 mole % urania and 10 mole
% thoria were chosen for this work.
in grinding, storage conditions, and lapsed time
had little or no effect on the observed equilibrium
composition. Initial oxidation state and degree
of initial solid solution formation had a very small
effect on the equilibrium composition (less than
2.5%). When gquench conditions were varied by
substituting liquid N, submersion in place of
rapid air cooling, a reduction in UQO, content
from 29 to about 26 mole % was obtained. Com-
positions the unit cell size*

Variations

calculated from
derived from x-ray powder patterns also confirmed
that the fcc phase had a UO, composition near
25 mole %. These findings would indicate some
surface oxidation during the quench period by the
older technique,

 

13, A. Friedman and R. E. Thoma, Reactor Chem.
Div. Ann. Progr. Rept. Jan. 31, 1963, ORNL-3417,
pp. 13034,

21.. O. Gilpatrick, H. H. Stone, and C. H. Secoy, Reac-
tor Chem. Diyv. Ann. Progr. Rept. Jan. 31, 1963, ORNL-
3417, pp. 13430,

3. o. Gilpatrick, H. H. Stone, and C, H. Secoy, Reac-
tor Chem. Div. Ann. Progr., Rept. jan. 31, 1944, ORNL~
3581, pp. 16064

4. o Gilpatrick and C. H. Secoy, Reactor Chem.
Div. Ann. Progr. Rept. Jan. 31, 1965, ORNIL.-3789,
pp. 239-43,

103

Miscellaneous Studies for Solid-Fueled Reactors

Unit cell dimensions were computed from x-ray
powder patterns using a least-squares fitting pro-
gram written by Williams® in place of the more
tedious graphical procedure. Test cases showed
no bias due to this change. Comparisons were
also made between the Debye-Scherrer camera
and the bench diffractometer methods of measur-
ing lattice parameters by the powder technique.
Good agreement was found in this case also.

An effort was made to better define the transition
temperature at which orthorhombic U30 4 converts
to the face-centered cubic UO,, phase in air.
The transition is sluggish, but the first sign of
the fcc phase appeared at 1528°C and progressed
fairly rapidly (50% in 2 hr) at 1540°C. No fcc
phase was cbserved after 2 hr at 1514°C. Tem-
perature calibration was felt to be reliable to
+5°C as measwred by two independent methods.
the year, Cohen and Berman of the
Westinghouse Bettis Laboratory reported on this
system® and found a lower urania composition
(50 mole %) for the two-phase boundary at 1200°C
than had been found in this study (~63 mole %).?
In view of this difference, a remeasurement at
1200°C was made using more recent technigues.
No effect could be ascribed to the use of pellets
at 60 mole % urania. This composition showed
some orthorhombic phase (V2 to 5%). Composi-
tions of 56, 53, and 45% urania displayed no
detectable orthothombic phase.  This indicates
that the phase boundary is between 56 and 60
mole % urania. Composition of 90% urania showed
little change from the older work.

Apparatus and equipment have been designed and
built to extend this study to lower partial pres-
sures of oxygen in the range of 107% atm, where

During

 

D, R. Williams, LRC~2: A Forlran Tattfice Constant
Refinement Program, 1S-1052 (November 1964).,

%1, Cohen and R. M. Berman, Am. Ceram. Soc. Bull,
44(4), 391 (1965).
shifts in the oxygen-metal ratio should be large
enough to be easily measured.
starting compositions have also been prepared
by more extreme procedures
ous material in the solid
oxides have been reduced in H, at temperatures
in excess of 1650°C, and reduced mixtures have
been fused in helium at temperatures above 2200°C.

A new series of

to assure a homogene-

golution form. Mixed

BEHAVIOR OF REFRACTORY-METAL
CARBIDES UNDER IRRADIATION

G. W, Keilholtz R. E. Moore
M. F. Osborne

Refractory-metal carbides of groups IV to VI
have potential applications in high-performance
nuclear power plants and in reactors for special
applications requiring extremely high power den-
sities. A series of experiments in progress is
aimed at determining the changes in physical
of monocarbides of

and mechanical properties

Ii, Zr, Nb, Ta, and W during fast-neutron irradi-

104

the form of 1/2-in. X l,é-in. cylinders over the

temperature interval 100 to 1400°C, the neutron
flux range 0.5 to 3.0 x 10!* neutrons cm™? sec™!
(>1 Mey), and the neutron dose range 0.5 to 5 x
102! peutrons/cm? (>1 Mev). Included in the
experimental series are specimens of each of
the monocarbides made by three different methods:
(1) hot pressing, (2) slip casting and sintering,
and (3) explosion pressing. Three low-temperature
(100 to 400°C) uninstrumented assemblies and
one instrumented high-temperature (1100°C) assem-
bly containing these carbides are undergoing iria-
diation in the ETR. Other experiments are planned
to achieve temperatures up to 1400°C.

Examination is complete for an assembly of one
sample of each of the five hot-pressed monocar-
bides irradiated (see Table 12.1) in the ORR at
about 100°C.7 Most of the specimens sustained
very little gross damage. Metallographic examina-
tions revealed no evidence of grain-boundary sepa-
ration in any of the samples. The volume increase
of the carbide specimens calculated from dimen-
sional measurements and the volume increase calcu-
lated from the lattice parameter expansions are

 

7G. W. Keilholtz, R. E Moore, and M. F. Osborne,

 

 

ation. The specimens are being irradiated in ORNL-TM-1350 (in preparation) (classified).
Table 12.1. Gross Volume Expansion and Lattice Parameter Expansion
of Refractory-Meta! Carbides Irradiated at 100°C
Volume Increase Volume Increase
Crystal Fast-Neutron Thermal-Neutron Aa/aoa /\c/cea from Lattice from Dimensional
Material Structure Dose, nvt ]?ose, nvt {7) (%) Parameters®? Measurements
(>1 Mev) (>1 Mev) (%) )
x 1021 x 10°1
WwWC Hexagonal 1.7 3.4 0.17 0.35 0.7 0.3
TiC Cubic 1.7 3.4 0.58 1.7 2.5
NbC Cubic 1.7 3.4 0.16 0.5 (3.6)°
TaC Cubic 1.4 2.7 2.7
ZrC Cubic 1.4 2.7 0.33 1.6 2.6

 

4, L. Yakel, ORNL Metals and Ceramics Division, personal communications.
obtained from x-ray diffraction patterns of unirradiated specimens

and ¢

in each case were based on values for a o

0
from the same batches.

The values for Aa/ao and AC/CO

"The volume increases were calculated from the equation AV/V0 = 2Aa/ao + Ac/c0 for hexagonal crystals and

from the equation f.\V,/VO = SAa/aO for cubic crystals.

©This value was extrapolated from dimensional measurements of the diameter of the niobium carbide specimen;
the length of the specimen could not be measured because one end was found to be broken off after irradiation.
given in Table 12.1. The lattice parameters were
measured from x-ray patterns obtained from reflec-
tions from polished surfaces, The gross volume ex-
pansion of the tungsten carbide specimen, the
only one of the five carbides which does not have
an isotropic crystal structure, was only about
0.3%. The gross volume expansions of the cubic
carbides of titanium, tantalum, and zirconium
were much greater (™2.6%); the lattice expansion
accounts for about 60 to 75% of the gross ex-
pansion of the titanium and zirconium carbides.
An x-ray pattern could not be obtained on tan-
talum carbide. The niobium carbide sample ap-
peared o expand more than the others, but its
lattice parameter expansion was very small,

These preliminary results indicate that refrac-
tory-metal carbides are =ufficiently resistant to
fast neutrons to merit consideration for nuclear
applications involving high neutron doses. Tung-
sten carbide, in particular, is attractive because
of its small lattice parameter expansion and
small gross volume expansion. The magnitudes
of these expansions are generally reliable indi-
cations of the degree of neutron damage.

EFFECTS OF FAST-NEUTRON IRRADIATION
ON OXIDES

G. W, Keilholtz R. E. Moore
M. ¥, Osbome

A systematic investigation of the effect of fast-
neutron irradiation on sintered compacts of MgO
and Al O, has been completed, and the results
have been compared with those previously ob-
tained® 16 for BeO. Several hundred cylindrical
specimens of each oxide with various grain sizes
and densities were irradiated in the Engineering
Test Reactor over the temperature range 100 to
1100°C and the fast-neutron flux range 0.5 fo

 

812, P. Shields, J. E. Lee, Jr., and W. E. Browning, Jr.
Effects of Fast Neutron Irradiation and High Tempers-
ture on Beryllium Oxide, ORNL-3164 (March 16, 1962).

R, P. Shields, J. E. Lee, Jr., and W. E. Browning, Jr.,
Trans. Am. Nucl. Soc. 4(2), 338 (1961),

1064 Ww. Keilholtz et al., Radiation Damage Solids,
Proc. Symp., Venice, 1962, vol. 11 (Vienna, IAEA, 1962).

e, w. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
The Effect of Fast-Neutron Irradiation on Beryllium
Oxide Compacts at High Temperatures, ORNL-TM-741
(Dec. 11, 1963).

3.0 « 10%* neutrens cm™? sec™! (>1 Mev). The
results are reported elsewhere, 717720

Among the three oxides, the greatest difference
in behavior was between MgQ, which has a cubic
structure, and BeQO, with a hexagonal crystal
structure. Previous resulis have shown that
the primary mode of damage to BeO is grain-
boundary separation caused by anisctropic crystal
The degree of damage, that is, frac-
turing or powdering, iucreases with increasing
dose and decreases with increasing temperature.
The cubic structure of MgQO precludes anisctropic
expansion, and the lattice parameter expansion
is much smaller than that of BeO. Accordingly,
irradiated MgO exhibits virtually no grain-bound-
ary separation, no powdering is observed, and
the pross volume expansion is small relative to
that of Bel,

Transgranular fracture is severe in MzO. For
example, about 40% of specimens irradiated at
150°C over the dose range 0.2 to 1.4 x 107!
neutrons/cm? were fractured. Unlike BeO, dam-
age to MgQ specimens was not related to the
neutron dose. Approximately 40% of the specimens
irradiated at 800°C and about 0% of the speci-
mens irradiated at 1100°C over the dose range 0.5
to 5.1 = 104! neutrons/cm?® (>1 Mev) were frac-
The randomness of fracture can

expansion.

tured randomly.

 

12(}. W. Keilholte ef al., Behavior of Be( Under Neu-
tron Irradiation, ORNL-TM-742 (Dec. 11, 1963).

13G. W. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
J- Nucl. Mater. 11(3), 25364 (1964).

14, W, Keilholtz ot al., J. Nucl, Mater. 14, 87—95
(1964).

15(}. W. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
Ylrradiation Damage to Sintered Beryllium Oxide as a
Function of the Fast-Neutron Dose and Flux at 110, 650,
and 11007¢C,*" submitted to Nuclear Science and Fngi-
neering for publication.

16()1. W. Keilholtz, J. E. Lee, Jr., and R. E. Mocore,
Reactor Chem. Div, Ann. Progr. Rept. Jan. 31, 1965,
ORNL-378Y9, pp. 23338,

17G. W. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
ORNL~TM-1050 (Mar. 26, 1965) (classified).

13G. W. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
ORNL-TM-1140 (July 9, 1965) («olassified).

196, W. Keilholtz, J. E. Les, Jr., and R. E. Moore,
ORNL-~TM-1300 {in press) (classified),

204, W. Keilholtz, J. E. Lee, Jr., and R. E. Moore,
“Properties of Mapnesium, Aluminum, and Beryllium
Oxide Compacts, . Irradiated to Fast-Neutron Doses
Greater than 107 Neutrons cm  ° at 150, 800, and
11007 (2,** accepted for pablication in Proceedings of
Joint Division Meeting of the Materials Science and
Technology Division of the American Nuclear Society
and the Refractories Division of the American Ceramic
Society, May 8—11, 1966, Washington, D.C.
106

Table 12.2. Lottice Parameter Expansion of MgQ, CL-A1203, and B0 lrradicted

at Low Temperatures to Comparable Fasi-Meutron Doses

 

 

 

 

Crystal Irradiation Fast-Neutron
Material . . , , ‘1, @
Structure Temperature Dose, > 1 Mev /,,\a/ao Ac = AvV/v
) (neutrons/cmz)
% 1021
Mg Cubic 150 1.1 0.0005 0.0015
CL-A1203 Hexagonal 150 1.0 0.0023 0.0024 0.0070
BeD Hexagonal 110 1.0 0.0013 0.0312 0.0338

9The fractional volume incrcase, AV/VO, for BeO and Al,O0; was calculated from the equation .«f\V/VO =

Q(Aa/ao) + (,t'\c/co). The equation AV/VO o S(g:\a/ao} was used for the case of cubic MgQ.

be explained by postulating that a minimal neu-
tron dose can weaken the crystals encugh to per-
mit the propagation of cracks from randomly dis-
tributed sites within MgQO compacts. The minimal
dose is probablv near the lower limit of our ex-
periments (~2 x 102° neutrons/cm?, >1 Mev),

since increases in strength have been reported for

21,22

lower doses by other experimenters. Direct

evidence is lacking that higher doses produce
weakening of MgO crystals, but electron micro-
graphs of MgQO irradiated to doses greater than
102! npeutrons/cm? (Figs. 12.1 and 12.2) show
a severe deterioration which, it seems likely,
would be accompanied by a loss in strength.

In-pile ancealing at high temperatures reduces
the volume expansion and lattice parameter ex-
pansion of MgO. Thermal stresses within neutron-
damaged compacts probably account for the mark-
edly greater gross damage observed in irradiations
at 1100°C.

The crystal stiucture of Al O, is anisotropic,
but the behavior of A1203 on exposure to fast
neutrons resembles that of MgO rather than BeO.

2lp A, J]. Sambell and R. Bradley, Phil. Mag. 9,
161 (1964).

22_]. Elston, ‘*Behavior of Neutron-Irradiated Beryllium
Oxide,’ Saclay Center of Nuclear Studies Report
DM-1182 {(1662).

 

A comparison of the lattice parameter expansions
of the three oxides irradiated at low temperatures
to a dose of ~102! neutrons/cm? is shown in
Table 12.2. The a and ¢ parameters of Al,O,
expanded by about the same amount under these
conditions; this is in sharp contrast to the behav-
ior of Be(), in which nearly all the expansion is
in the ¢ parameter. Recent irradiations of trans-
lucent Al O, specimens (lucalox) to somewhat
higher doses (~1.4 x 102! neutrons/cm?) pro-
duced a greater expansion of ¢ parameter, how-
ever, and resulted in an anisotropic expansion
ratio (Ac/co)/(Aa/ao) of about 3.8. 'This ratio,
however, is small compared with that for BeO.
Therefore, the mechanisms through which Al O,
is fractured during fast-neutron irradiation appear
to be the same as in MgO.

Neutron damage to all three oxides, as judged
from lattice parameter expansion and gross vol-
ume expansion, decreases with increasing irradi-

Because of in-pile thermal
gross fracturing of Mg0O and

ation temperature.
stress,
Al,0, increased markedly at high temperatures
(1100°C). Therefore, to minimize damage to these
oxides in nuclear reactor application, the temper-
ature should be maintained as high as practicable,
and the system should be designed so as to mini-
mize thermal stress. In the case of BeO, thermal
cycling, which tends to promote grain-boundary
separation, should also be avoided.

however,
107

F PHOTO 81629

S

 

Fig. 12.1. Electron Micrograph of Unirradiated MgQ of Density 3.4 g/cm3 ond Grain Size 10 ¢t at 32,000x Magni-
fication. Reduced 11%.
108

*:
-

A,

 

Fig. 12.2. Elzctron Micrograph of Mg0 of Density 3.4 g/cm3 and Grain Size 10 ;¢ hrradioted to a Fast-Neutron

Dose of 1.2 x 102! Neutrons/cm? (>1 Mev) at 150°C at 32,000x Magnification Showing Genera! Deterioration and

Small Cracks. Reduced 12.5%.
ANNEALING OF IRRADIATION-INDUCED
THERMAL CONDUCTIVITY CHANGES
OF CERAMICS

C. D. Bopp

The annealing of the neutron-induced thermal
conductivity change has been measured in several
ceramic materials. The annealing temperatures are
listed in Table 12.3. A fluxmeter apparatus?3
was used with the materials of high conductivity,
and a mercury-contact apparatus?? was used with
the materials of lower conductivity. Previously,?®
the mercury-contact apparatus was used for pre-
and postirradiation measurements of these same
materials, but annealing was not studied. Com-
parison with the present results shows good
agreement except in the instance of thin speci-
mens of beryllia, for which it appears that the
mercury-contact apparatus is unsuitable because
of the extremely high conductivity of this mate-
rial. The annealing temperatures for beryllia
shown in Table 12.3 are in good agreement with
those found by other workers?® for material given
nearly the same irradiation dosage.

 

23¢. D. Bopp and O. Sisman, Reactor Chem. Div. Ann.
Progr. Rept, Jan. 31, 1965, ORNL-3789, p. 232.

24 b, Bopp, J. Am. Ceram. Soc. 47, 154 (1964).

250. Sisman, C. D. Bopp, and R. L. Towns, Solid
State Div. Ann. Progr. Rept. Aug. 31, 1957, ORNL-2413,
p. 80,

26y Elston and C. Labbe, J. Nucl. Mater. 4, 143
(1961).

109

 

Taoble 12.3. Annealing of Irradiation-lnduced
Therma! Resistance
Temperatureb at Which
the Irradiation-Induced
Irradiation Thermal Resistance
Material Dosage” Is Decreased by

the Indicated Percentage

 

 

20% 40% 60%  80%
Sintered beryllia  2x 10'% 600 650 750 900
Sintered alumina 2% 101° 600 1050 c
Sapphire 4-8x 107 600 850 1000 1250
Spinel 2x16Y 556 900
Porcelain 2x10'® 5350 950 ¢
Forsterite 2 x 101° 750 1250 c
Zircon 2x10'% 1000 1050 1100 1250
Cordierite 2x10'% 755 1000 1100 1250
Steatite 2x10*% 550 800 ¢

 

“The dosage unit is fast neutrons /em? (>1 Mev) ex-
cept in the_instance of =zircon, for which the unit is
fissions/cm” (since the presence of a trace of uranium
impurity dominated the radiation effect in zircon; see
C. D. Bopp et al., Reactor Chem. Div. Aan. Progr. Rept.
Jan. 31, 1965, ORNL-3789, p. 231).

bAnnealing was conducted isochronally; the sample
was heated for periods of 1 hr at successively higher
temperatures using 50°C increments, except that the
pericd of heating at 1250°C, the highest temperature
used, was 4 hr.

®This amount of annealing was not attained after the
1250°C heating.
Part IV
Other ORNL Programs

 
13. Chemical Support for Saline Water Program

THERMODYNAMICS OF GYPSUM IN AQUEOUS
SODIUM CHLORIDE SCLUTIONS

W. L. Marshall Ruth Slusher

In an extensive investigation of the equilibrium
phase relationships of gypsum (CaSO4-2[-120) in
NaCl-H_ O solutions continuing from the previous
work,! solubilities were determined at 12 dif-
ferent temperatures from 0 to 110°C and in solu-
tions of sodium chloride from very dilute to those
saturated with sodium chloride or a new solid
phase. All results and available literature data
were evaluated by an equation,

log K, = log K2 +8S\I/(1 + AVD)

sp

—~ Bl - CI* -2 log a, , (1)

where 85 is the Debye-Hiickel limiting slope for a
2-2 electrolyte corrected for molal units, a_ is the
activity of water, I is the ionic strpngth, A, B,
and C are adjustable parameters, K is the sol-
ubility product constant (at I = 0), and K __ is the
o product [(Ca?™) (80,77 DetermmedpvaTues
of K”p are shown in Fi 1g 13.1, and the parameters
A, B, and C are included in Table 13.1. With the
obtained parameters, the differences between most
calculated and experimental values of K op Were
between 0.5 and 3% over the entire ranges of con-
ceatration and temperature,

Standard Thermodynamic Yalues for Gypsum

Assumptions were made that the standard heat
of solution, AHS, and the standard change in heat

 

lw, L. Masrshall, Ruth Slusher,
J. Chem. Eng. Data 9, 187 {1964).

and E, V. Jones,

113

ORML--DWG A5-- 11415

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7{*C)
Lo &0 80 40 20 0
BT T { ‘ """" T
,,,,, . ] I Lo
i . L L_v
/..—P‘ i
a4 R ot . '-\.\\i' Jl
A ™
oo N
// : -
° % 4'5 e —— L - b _— —_
X /
g /s
1 /P’ T Tty T T T T
4-6 S / ........
/ X
*/“ - 1% OF ABSOLUTE VALLE o | ——
a7 lf —— —
26 28 o) 32 24 15
Y X107
Fig. 13.1. Logarithm of Kfp vs 1/T(°K) for Ca$0,s

2H,0 (Gypsum), 0~110°C.

capacity at constant pressute, ACZJ could be ex-

pressed by

AH® = E + {ACS dT )

and
AC; =F + GT, 3)
where &, F, and G are parameters, These ex-

pressions were substituted into the van’t Hoff
equation,

d1n Kgp/d(l/Tj =—-AHYT , {4
which was then integrated over all values of Kgp
and T(°K) to obtain a four-parameter equation.
With the sepatately determined values of K:
(unsmoothed but obtained using smoothed values
114

Table 13.1. Standard Thermodynamic Quantities and Parameteis 4, B, and C Used in Eq. (1)
for the Equilibrium CaSO ;2H,0(s) e=2Ca?(aq) + $0,2(aq) + 2H,0

 

 

T (°C) AF® Ar® As® A B c
(kcal/mole) (kcal/mole) (cal mole ! °C~ 1)

0 5.58 +2.50 -11.3 1.450 —0.0680 0.0264
10 5.71 +1.55 —14.7 1.468 —-0.0274 0.0212
20 5.87 +0.68 -17.7 1.490 —~0.0072 0.0178
25 5.96 +0.27 —-190,1 1.500 1+ 0.0006 0.0164
30 6.06 —0.12 —20.4 1.510 +0.0076 0.0160
40 6.28 -0.85 ~22.8 1.530 +0.0178 0.0150
50 6.52 —1.50 —24.8 1.544 +-0.0200 0.0138
60 6.77 —2.07 ~-26.5 1.558 +0.0200 0.01256
70 7.05 -2.57 —-28.0 1.570 +0.0200 0.0112
80 7.33 -3.00 ~20.2 1.580 +0.0200 0.0094
90 7.63 —-3.35 —30.2 1.588 + 0.0200 0.0072

100 7.94 —~3.62 --31.0 1.594 +0.0200 0.0050
110 8,25 —3.82 —31.5 1.595 +0.0200 0.0038

 

of Asp) from Fig. 13.1, the four parameters were
evaluated by the method of least squares to
obtain the equation

log Kgp = 390.9619 — 152.6246 log T
— 12545.62/7 + 0.081849287T . (5)

The average deviation from this equation of the
experimentally determined values of Kg shown in
Fig. 13.1 was 10.6%. Values of AHI® at each
temperature were obtained by differentiating Eq.
(5) with respect to 1/T(°K) and substituting the
the result into the van’t Hoff equation, (4), while
those values of AC® were obtained by differen-
tiating with respect to T the resulting expression
for AH®  While separate values of AC® might
be expected to be somewhat inaccurate, the
average value of AC® from 0 to 110°C of —57 cal
mole™! °C~! is believed to be significant.
Values of AS® and AF® were obtained from the
standard thermodynamic equation,

AF®= AH® — T AS°. (6)

Representative calculated thermodynamic values
obtained by these procedures are
Table 13.1.

included in

The Additional Sslid Phase

At concentrations of NaCl above 3 to 5 m and
at temperatures from 70 to 95°C, a saturating
solid other than CaSO4-2H20 or NaCl was found.
In special experiments at 70°C, this second
saturating solid phase was identified by petro~
graphic examination? to be N32804-5CaSO4-3H20,
found previously in the system CaClz-NazSO4-
H2O.3 By the formation of N82$O4-5Ca804-31{20
from solutions initially only of NaCl and CaSO,,
the system becomes a four-component system,
tather than three, and must be defined by the
components CaSO4, NaCl, CaCIz, and H20.

2Thanks are due G. D. Brunton, Reactor Chemistry
Division, for these examinations.

3A. E. Hill and J. H. Wills, J. Am. Chem. Soc. 60,
1647 (1938).
THE OSMOTIC BEHAVIOR OF SIMULATED
SEA-SALT SOLUTIONS AT 123°C

P. B. Bien B. A, Soldano

The isopiestic technique previously developed
at ORNL* has been used to test the behavior of
thtee simulated sea-salt solutions at 123°C.
The tests were made to assess the applicability
of calculation methods proposed for determining
thermodynamic properties, including the wvapor
pressures, of solutions of interest to the vacuum-
distillation process for
water.”

desalination of sea-

The three sea-salt solutions were prepared to
simulate closely the ‘‘standard seawater’” defined
by K. S. Spiegler.® In order to avoid the evolution
of gases and possible corrosion in the vapor
chamber, the troublesome univalent anions HCOs‘
and Br™ were replaced by equivalent quantities of
chloride ion. Solution A was further modified by
replacing the calcium with magnesium, thereby
increasing the magnesium concentration to
0.06451 m; the sulfate concentration in solution A
was kept at 0.02856 m. Solution B was modified
further by completely removing the calcium to-
gether with an equivalent amount of sulfate; the
sulfate concentration in solution B was thus re-
duced to a concentration of 0.01822 m. Saline C
was not modified except by the replacement of
chloride for the bicarbonate and bromide. Thus,
the compositions of the three solutions were:

A (m) B (m) C {m)
cat? 0,01084
Mgt 0.06451 0.05417 0.05417
gt 0.01007 0.01607 0.01007
Na 0.47564 0.47564 0.47564
50,7 0.02856 0.01822 0.02856
1™ 0.55761 7.55761 0.55761
I 0.7078 0.6664 0.7078

Four dishes of each solution, four dishes of NaCl
standard solution, and four dishes containing
standard weights for internal calibration of the
balance were loaded into the chamber together
with a larger dish containing NaCl solution which

 

8. A. Soldano and G. S. Patterson, J. Chemn. Soc.,
1962, 937,

“R. W. Stoughton and M. H. Lietzke,
Data 10, 254 (1965).

6K. S. Spiegler, Sea Water Purification, Wiley, New
York, 1962,

J. Chem. Eng.

115

gerved as a buffer reservoir, After the air was
evacuated from the sealed chamber, the apparatus
was brought to and held at 123°C. Each day, the
weights of every dish were recorded as readings
on an electrically operated magnetic balance,
after which a small amount of steam was vented
from the apparatus, thus increasing the concen-
trations of all the test, standard, and buffer solu-
tions. The changes in the concentrations of the
test solutions were inferred from the changes in
the weights of the dishes. The jonic strengths of
the solutions were then calculated from the formula

1 2
I :;21111.2‘. ,

with the concentrations m, deduced from the
changes in the solution weights. Table 13.2
presents average values of I for the NaCl standard

Toble 13.2. Experimenta! Yalues for lonic Strength,

 

 

}:mj.ziz/?.

Date, 1965 [NaCJ. IA IB IC

March 9 0.6030 0.8550 $.7965 0.9130
10 0.6683 0.8228 0.7884 0.8714
11 0.6734 0.8318 0.8019% 0.8580
12 0.6987 0.8662 0.8087 0.5134
15 0.6693 0.8758 0.8358 0.%911
17 0.7744 0.9611 0.9278 1.0011
12 0.8658 1.0229 1.0165 1.1158
23 1.4601 1.6585 1.5788 1.6892
24 1.5750 1.7812 1.7510 1.8034
25 1.6342 1.8484 1.7464 2.0445
26 1.8466 2.1308 2,0421 2.0568
26 1.9674 2.2016 2.1102 2.2416
23 2.0728 2.2008 2.1406 2.3070
31 2.6250 2.6484 2.5905 2.9369

April 1 2.6496 2.7254 2.6530 2.9761
2 2.7894 2.8929 2.8513 2.9808
5 3.1081 3.0117 2.9846 3.3108
5 3.3583 3.3064 3.2420 3.5966
6 32.6166 3.5664 3.4352 3.7656
7 3.6801 3.3466 3.4851 3.8864
5 4.3018 4.0551 4.0087 4.43950
2 4.6016 4,1999 4.0G093 4.4954
9 6.1244 5.5620 5.2566 6.2986
12 5.1064 4.5226 4.4029 3.0060
13 7.9924 6.3590 6.5914 7.6788
14 7.9248 6.5999 6.2487 6.9634

 
, ISOP{ESTIC RATIOS

 

Yygct
Igact

A=
o
w0

0.8

116

 

  

ORNL—DWG €6—278

  

..—-“""—'—-
| /,_,-«-F

SALINE A~

   
  

SALINE ¢

e

e

 

 

e S
1 2 3 4 5 6 7 8
IONIC STRENGTH QF NaCl
Fig. 13.2. lsapiestic Ratios of Salines A, B, and € Compared.

solution and the three test solutions at 26 dif=-
ferent levels of water activity., Values for the
isopiestic ratios R for each of the test solutions
were obtained at each point from the ratio INaCl/
Itest soln * The equations fitted to R as a func-

ion iven as foll :
tion of [Nac1 are giv s follows

R, =0.7701 + 0.08797 - 0.00471%; o, =0.0292,

A

RB = 0.8082 + 0.0889! — 0.0052[%; Op = 0.0270 ,
B
and

R_ = 0.7427 + 0.0784/ — 0.00551% o, = 0.0344
C

Plots of these equations, as shown in Fig. 13.2,
suggested that solution C was clearly different
in behavior from solutions A and B. The con-~
siderable overlap between solutions A and B
suggested that a more sensitive test could be
made of the possible significance of any dif-
ference between them. To this end, the hypothesis
was made that the solutions did not differ, and
a t test was applied to the 26 values of I, — IB
obtained from the data in Table 13.2. An average
value of the difference found was 0.077 unit of
ionic strength (not units of R); this difference

was found to be well above the 0.001 level of
significance.

Marshall al.” have recently shown that
calcium sulfate should begin to precipitate from
seawater at about 117°C. Therefore, it is likely
that all the tests with solution C were performed
in the presence of solid CaSO, (anhydrite) and
that, consequently, the data for solution C should
not be suitable for analysis until and unless
accurate values for the solubility of the com-
ponents make it possible to infer the true solu-
tion composition. It may be noted, however, that
the values of R calculated from the true solution
composition should be higher than those plotted
in Fig. 13.2, probably bringing the data more
closely in line with those of solutions A and B.
The difference between A and B was observed
to be quite consistent over the whole concentra-
tion range, and the average value of this dif-
terence, 0,077 ionic strength unit, is about twice
the value calculated for the solutions as made up
at room temperature, 0.0414. The solubilities of

et

7W. L. Marshall, Reacfor Chem. Div. Ann. Progr.
Rept. Jan. 31, 1965, ORNL-3789, p. 294.
the components of solutions A and B have not
yet been determined as a function of concentration,
and it is still assumed that no precipitation took
place in the dishes containing these solutions.

With respect to the original objective, that of
testing the proposed method of calculating vapor
pressures of saline waters, solutions A and B
may not be sufficiently different to permit good
evaluation of such a method. Further tests should
be made with solutions known to be phase stable
at elevated temperatures and probably should
begin with simpler systems such as those con-
taining mixtures of NaCl and Na2SO4 or mixtures
of NaCl and MgCl1,.

ALUMINUM- AND TITANIUM-ALLOY CORROSION
IN SALINE WATERS AT ELEVATED
TEMPERATURES

E. G. Bohlmann
J. C. Griess

F. A. Posey®
J. F. Winesette

The studies of the corrosion of aluminum and
titanium alloys in sodium chloride solutions have
continued. They have included investipations in
the 100-gpm titanium loops,® the small titanium
efectrochemical loop,'? and conventional labora~
tory equipment. It was found necessary to modify
the electrode assembly in the small titanium loop
to reduce troublesome IR drops; this entailed
ptoviding coaxial polarizing electrodes and dual
reference electrodes as shown in Fig. 13.3.

Galvanostatic polarization studies of the cor-
rosion of the 5454 (27% Mg, 0.8% Mn, 0.1% Cr,
<0.01% Cu) and 6061 (1.0% Mg, 0.6% Si, 0.25% Cu,
0.25% Cr) aluminum alloys in 1 M NaCl at 150°C
have been catried out in this equipment.!! The
results obtained were generally similar to those
obtained by other workers at lower temperatures.
Thus, a pronounced minimum exists in the cor-
rosion rate of aluminum and its alloys in chloride
solutions in the vicinity of neutrality. Changes

 

BORNL Chemistry Division.

YSaline Water Conversion Report for 1962, U.S, Dept.
Intericr, pp. 12, 16, 10962,

g, a. Bohlimann, F. A. Posey, and J. F. Winesette,
Reactor Chem. Div. Ann. Progr. Rept. fan. 31, 19635,
ORNIL-3789, p. 296.

Ny G. Bohlmann and F. A. Posey, *“‘Aluminum and
Titanium Cortosion in Saline Waters at Elevated Tem-
peratures,’’ Proc. First Intemational Symposium  on
Water Desalination (in press).

117

with pH in the polarization curves of anodic and
cathodic processes occurting at the aluminum-
electrolyte interface provide a kinetic basis for
understanding this and other aspects of the cor-
rosion behavior. At low potentials, the rate of
the anodic or corrosion reaction is independent of
the electrode potential, but it increases with
increasing pH. The rate of the anodic process is
controlled by the rate of mass transport of hy-
droxide ions to the oxide-solution interface. At
higher potentials, in the presence of chloride
ions, the anodic-polarization curve exhibits a
pitting potential which independent of the
anodic current density. The pitting potential
does not vary with pH, but decreases with in-
creasing chloride concentration. The cathodic
reaction in alkaline solution consists of the re-
duction of water molecules to {form molecular
hydrogen; this process is pH independent. With
increasing acidity, reduction of hydrogen ions be-

18

comes increasingly important. The minimum cor-~
rosion rate represents a compromise between the
decrease in the rate of the transport-controlled
anodic reaction with increasing acidity and the
increase in the rate of the cathodic hydrogen-
evolution reaction. Oxygen in solution may also
increase the corrosion rate by providing an addi-
tional cathodic process.

Comparison of polarization curves of the 5454
and 6061 alloys shows that the rate of the cathodic
hydrogen-evolution reaction of the 6061 alloy is
considerably greater than that of the 5454 alloy.
The enhanced rate of the cathodic process on the
6061 alloy accounts for its greater corrosion rate
at any pH and for its susceptibility to pitting
attack. Catalysis of the cathodic process on the
6061 alloy may be attributable to its copper
content. It is hypothesized that accumulation of
copper at the metal-oxide interface as corrosion
progresses results in pit formation.

These studies stemmed from results obtained
in the 100-gpm titanium loops, which showed
gross pitting of 6061 alloy specimens after 248-hr
exposure to pH 6.0, 1 ¥ NaCl at 150°C. Under
comparable conditions, the 5454 alloy showed
uniform attack after 1620-hr exposure; continuing
corrosion rates were "~ 6 mils/year at 7 to 25 fps.
Similar results are being cbtained at 100°C; 6061
specimens show pitting attack after 500 to 1500
hr, whereas 5454 specimens show uniform attack
with continuing corrosion rates of ~1 mil/year
atter ~2800-hr exposure,
ASBESTOS WICK
FROM CALOMEL
ELECTRODE

   
    
 
   
   

GASKET
INSUL ATOR

ORNL-DWG 65-42264

POLARIZING
ELECTRODE

CALOMEL
REFERENCE
ELECTRODE

   

SPECIMEN
NQ. 3

 

 

 

 

 

—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Investigations of the severe corrosion of titanium
in saline waters, reported last year, have also
Three racks of specimens are being
exposed in the water box of the No. 1 effect at
the Freeport Desalination Plant. In this location,

continued.

the specimens are exposed to near-normal con-
centration, pH ~7.5 brine (sulfuric acid treated
and neutralized for scale prevention) at 129 to
138°C. Racks have been examined at 12-, 32a,
and 82-day intervals and have all shown negligible
to severe attack in areas of contact with Teflon;
no attack was observed in metal-to-metal contact
areas., This was consistent with the previously
noted promotion of attack by contact with Teflon.
It appears likely that this results from the release
of small amounts of fluoride ion!? by the Teflon.
Anodic polarization studies have shown that as
little as 12 ppm fluoride substantially increases
the active corrosion rate and the rate of activation
of titanium.

It is also likely that the absence of more general
attack on the rack specimens stemmed from the
somewhat low temperature of exposure — the bulk
of the exposure was at temperatures less than
138°C. The previously noted dominant importance

 

12k, G. Bohlmann and J. C. Griess, Reactor Chemn.
Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3780,
p. 297.

 

 

 

 

 

 

 

of temperature was reemphasized by recent ex-
perience with 100-gpm titanium-loop piping. For-
mation of several large pits was observed during
operation at temperature 2150°C. Subsequently,
the loop has been operated at 100°C for ~2800 hr
with no apparent further attack. A similar pit
initiated in another part of the loop penetrated the
0.2-in. pipe wall after approximately 5400 hr at
150°C and after 300 hr at 200°C, 1!

A particularly significant point conceming the
very sharp temperature dependence has just been
discovered. This is that the breakdown or pitting
potential for titanium, reported as ~10 v in room-
temperature chloride solutions,!® has a very sharp
temperature dependence.  Studies in the small
titanium electrochemical loop showed pitting po-
tentials at plus 500 mv (vs S.C.E.) or less for
commercially pure titanium at temperatures of
150°C and above. Laboratory and loop electro-
chemical studies have shown this sharp tempera-
ture dependence over the range 25 to 190°C. The
actual potentials measured and temperature de-
pendence, however, are greatly influenced by
alloy composition and surface condition, so no
quantitative data will be presented at this time.

 

13N, Hackerman and C. D. Hall, Je., J. Electrochem.
Soc. 101, 321 (1954).
CHEMISTRY OF SCALE CONTROL

E. L. Compete  J. E. Savolainen

The tendency of seawater to deposit scale in
distillation limits
range and brine~concentration factor of the process
and affects the efficiency. The problem
economic: scale control costs probably should
not exceed a few cents per 1000 gal. This elimi-
pnates many technologically feasible processes,
unless salable by-products are produced.

Raw-materials costs alone appear to be too high
for most processes except those using compounds
of carbon or sulfur. Sulfuric acid addition is
conventionally used to prevent the formation of
calcium carbonate or magnesium hydroxide scale
and permits operation up to temperatures near
250°F. Calcium sulfate may precipitate above
this temperature.

The prevention of alkaline scale in seawater
heaters by means of carbon dioxide additions
has been supgested. Ellis'* observed the effect
of carbon dioxide pressure on the solubility of
calcite in various sodium chloride solutions at
elevated temperatures. Thermodynamic equilibrium
constants and mean activity coefficients of cal-
cium and bicarbonate ions were reported. We
adapted these findings to the composition and
alkalinity of standard seawater, assuming that
the solutions and activity coefficients are similar
at equal ionic strengths. It was thus estimated
that the addition of 0.9 1b of CO, per 1000 gal
of seawater feed, corresponding to a CO, partial
pressure of .09 atm at 25°C, might prevent
calciie precipitation up to 125%C (257°F). In a
once-through flash-evaporator system, alkaline
scale would precipitate in the brine bulk as
catbon dioxide flashed in the first stage. Heating
surfaces would not be fouled, and C02 might be
recycled without compressors.

Data were not available permitting consideration
of magnesium hydroxide precipitation.

The assumption that seawater resembles sodium
chloride solutions of equal ifonic strength neg-
lected consideration of ion-pair formation. The
chemical model of seawater at 25°C of Garrels!®
and Thompson showed a significant degree of

equipment the temperature

is

 

144, J. Eliis, Am. J. Sci. 261, 239~67 (1963).

15R. M. Garrels and Mary E. Thompson, Am. J. Sci.
260, 57—66 (1962).

119

association of sulfate, bicarbonate, and carbonate
jong with magnesium and sodium and also with
ions. Our preliminary estimates for
100°C have indicated that magnesium sulfate pair
formation will be greater at the increased tempera-
ture. Calcium sulfate is similatly affected to a
lesser extent. These phenomena may also affect
the application to seawater problems of the studies
of Marshall, Slusher, and Jones!® on the solu-
bility of calcium sulfate in sodium chloride solu-

calcium

tions.

The economic removal of calcium sulfate
scaling tendencies may be possible by thermal
precipitation techniques in which seawater is
heated to a temperature of calcium sulfate super-
saturation; precipitation may be facilitated by
contacting with calcium sulfate particles. An
understanding of ionic and solubility equilibria
and precipitation kinetics are needed to provide
a rational basis for an economic process using
this technique.

The rate of precipitation from a supersaturated
solution is affected by mass transfer to the crystal
surface, reaction with surface sites, and nu-
cleation of new particles. Without nucleation, the
rate cannot exceed that of mass transfer. Co-
efficients for mass transfer to particles were
related by Harriott!? to that for a single freely
falling sphere in the given liquid. The coefficient
decreases as particle size increases up to about
100 ;2 and changes little thereafter. For anhydrite
crystallization from standard seawater with about
10°F superheating, we estimated that the mass-
transfer limitation would permit growth rates up
to several centimeters per hour. Such rates would
permit an effective unit for removing calcium sul-
fate from seawater to be designed.

Neilsen!? has indicated that, as supersaturation
is decreased and the surface reaction becomes
limiting, crystallization rates decrease more than
the concentration by a substantial proportion.
On the other hand, if the concentration is in=-
creased and high levels of supersaturation are
reached, undesirably rapid nucleation is to be
anticipated. Thus, the supersaturation must be

 

16w, L. Marshall, Ruth Slusher, and E. V. Jones, [.
Chem. Eng. Data 9, 18791 (1964).

17peter Harriott, A.L.Ch.E. (Am. Inst. Chem. Engrs.)
7. 8(1), 93-102 (1962).

185, E. Neilsen, Kinetics of Precipitation, Mac~
millan, New York, 1264,
maintained in the range controlled by mass trans-
fer for the most effective operation of “‘contact-
stabilization’ equipment.

In addition to capital and heat costs of the
contact-stabilization process, pumping-energy costs
may be appreciable. At higher temperatures, a
higher head is required in order to pressurize the

120

feed to prevent boiling. For example, at 330°F
the vapor pressure of water is 103 psi, and about
1 kwhr is required to pump 1000 gal of feed
against this head. The maximum economic tem-
perature is thus limited by the exponential rise
in vapor pressure.
14. Effects of Radiation on Organic Materials

W. W. Parkinson

EFFECT OF RADIATION ON POLYMERS

W. W. Parkinson W. K. Kirkland
R. M. Keyser

Polytetrafluorcethylene (Tetlon) degrades at very
moderate radiation doses in air, but it has been
found to show much less sensitivity when irradi-
ated in an inert atmosphere ot vacuum. 1.2 Because
of the usefulness of Teflon in deleterious environ-
ments and because the films used in previous
work ? would be extremely sensitive to atmosphere,
sheet stock of 1[3,) and 1/6 in. thickness has been
investigated. Tensile specimens have been ir-
radiated at about 25°C in air and vacuum over
a range of doses up to 7 x 107 rads. Tensile
properties were measured in air at room tempera-
ture. The specimens in vacuum were outgassed
at 5 > 107° torr and 140°C and sealed in glass
capsules. The dose rate was 1.2 x 10° rads /hr.

The tfensile properties of the 1/1 ;-in. sheet ir-
radiated in air and in vacuum are plotted vs dose
in Fig. 14.1. It is significant that in vacuum the
tensile strength decreases to 40-45% of its orig-
inal value at 2 x 10° tads but remains almost
constant from this dose to perhaps 10° rads or
more, It is seen that useful mechanical properties
are refained in an inert atmosphere to doses in
excess of 3 » 107 rads. In air both the "/16- and

1/$,)=-ixl.-t}1i(:1': specimens retain 33% of their original
tensile strength at 10° rads, in contrast to the

films of the earlier work,? where the tensile

 

el b Bopp and O. Sisman, Nucleonics 13(7), 28
{1955).

20, A. Wall and R. E. Florin, J. Appl. Polymer Sci. 2,
251 (1959).

121

Oscar Sisman

strength decreased to very low values in this
dose range.

The maximum in the elongation at break, ob-
served in the 1/32--1'.11. as well as the 1/1 gmin. spec-
imens, is interesting in the light of “zero strength
time’’ measurements, demonstrating that scission
predominates at doses above 10% rads.® Probably,
radiation-induced defects in the crystal structure,
indicated by a minimum in the density-dose curve
at 10? rads,® increase the ductility and account
for the maximum in the elongation.

RADIATION-INDUCED REACTTIONS
OF HYDROCARBONS

R. M. Keyser W, W. Parkinson

Two processes are under study as possibilities
for utilizing the fission energy which appears as
kinetic energy of the fission fragments. A *°Co
assembly ig used currently as a more convenient
source of ionizing radiation than a nuclear reactor.
One of the processes is the hydrogenation and
alkylation of coal in mixtures with alkanes. The
other process is the synthesis of amines in mix-
tures of ammonia with alkanes and alkenes.

The radiation-induced reactions of naphthalene
in hexane are being investigated initially as a
model system for the coal process.® A tempera-
ture-programmed gas chromatograph is being recon-
ditioned, and high-temperature columns suitable

 

*A. Nishioka et al., J. Appl. Polymer Sc¢i. 2, 114
(1959).

4W. W. Parkinson et al.,, Keactor Chem. Div. Ann.
Progr, Rept. Jan, 31, 1965, OENL-3789, p. 320.
122

ORNL.-DWG 66-979

 

 

 

 

 

 

 

 

 

 

 

7000 -\ | ‘ e
. ; | :
J | A TENSILE STRENGTH, IRRADIATED IN AIR
6000 - Lo el o
| A ELONGATION, IRRADIATED IN AIR
O TENSLE STRENGTH,IRRADIATED IN VACUUM
® ELONGATION, IRRADIATED IN VACUUM
(Y4—in. SHEET)
5000 | J ~~~~~~ ©
A | /"—m"h':’\ | |
4000

 

 

 

 

 

 

 

 

 

3000

 

 

 

ENSILE STRENGTH ( psi

2000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, i L |
—NL ;

10°

 

 

LY

RADIATION DCSE ({rads)

Fig. 14.1.

for analysis of the expected products are being
procured,

The effect of strong bases on reactions involving
ionic intermediates in irradiated hydrocarbons has
been demonstrated recently.® These results have
led us to search for amines in the radiolysis
products from solutions of ammonia in n-hexane
and in hexene-1. Solutions for irradiation were
prepared by condensing known amounts of an-
hydrous ammonia into deaerated samples of n-
hexane and hexene-l.
most solutions were of the order of 2 to 3 mole %,
which is about the limit of ammonia solubility
in these hydrocarbons at room temperature and

Ammonia concentrations in

atmospheric pressure. Irradiations were carried

out in a 20,000-curie ®°%°Co source at a dose rate

 

SW. R. Busler, D. H. Martin, and T. F. Williams, Dis-
cussions Faraday Soc. 36, 102 (1963).

Tensile Properties of liradiated Teflon.

of 8.5 x 10'% ev g7 ! min~?!. Total doses were in
the range 5.0 x 102% to 8.6 x 102! ev/g.

After irradiation the samples were opened and
subjected to gas chromatographic analysis using
a column with a liquid phase consisting of tetra-
hydroxyethylenediamine with
amine added as a tail reducer.

tetraethylenepent-
This column gives
good separation with only slight peak tailing of
calibration mixtures containing expected products
such as the 1-, 2-, and 3-aminchexanes and lower-
molecular-weight amines.

The results obtained so far have not been prom-
ising. No amines have been detected in the ir-
radiated solutions, the limit of detection corre-
sponding to a G(amine) of 0.2 molecule per 100 ev,
Mechanistic considerations indicate that a possible
reaction path leading to amine formation in ir-
radiated ammonia-hexane solutions may be formu-
lated as follows:
123

+ —_
CH,,—>CH, S +e, (1
4 + +
CH,~-—>CH " +H, 2)
+ +
CH  +NH, —>I[CH NH]I, (3)

. 34 _
[Cfil:{laNHSj +e” —> CH NH +H-, (4)

C H

o
ellgg * 07 > CGH, g ()

In view of the telatively low concentrations of
ammonia possible at atmospheric pressure, reac-
tion (5) may occur before the C6H1_3+ carbonium
ion encounters an ammonia molecule as in (3).
Experiments with ammonia at pressures above
atmospheric are currently in progress in an effort
to increase the ammonia concentration in the
system.

ADDITION REACTIONS OF FURAN
DERIVATIVES

C. D. Bopp W. D. Burch®
W. W. Parkinson

Polar organic compounds have been observed to
add to olefinic groups in chain reactions initiated
by chemically produced free radicals or ionic
intermediates.  Such chain reactions offer the
possibility of the high vyields required for com-
utilization of the radiation from radio-
isotopes. Materials which may be candidates for
upgrading through such reactions are the furan
derivatives. The preducts, hicyelic and tricyclic
ethers, would be useful for solvents and for further
processing into chelating agents, surface-active

mercial

agents, ete.

A survey is being performed of the radiation-
induced addition of saturated furan and similar
ethers to unsaturated furan denvatives and al-
kenes. Compounds which have been investigated
are listed in Table 14.1 and grouped as saturated
compounds (telogens) and unsaturated compounds.
Solutions containing a telogen and an unsaturated
compound in concentrations of 10 to 1 by volume
irradiated and partially analyzed by gas
In many cases, infrared spectra

were
chromatography.

 

5Temporary employee from Kansas Stafte University,
Manhattan.

were recorded of the starting materials and the
product mixture.

The telogen was present in greater concentration
in the mixtures since it was desired to promote
the formation of 1:1 telomers in the reactions
(1)—(3) below and to minimize the addition of
second and third molecules of the unsaturated

compound.
Hp e, He [P SO
L/H ~AM - He
Hal X Hz N
0 TCHy o~ “CH
(1)
HE H2 H ™~ oI
L e T
2 ~. 2
0 TcH o E
-0
A Lo
H Tj» H
Ha Ha
~o7 (2)
H') HE
_ =, —— sl
¢ H:T“ T Hyp ~ L?H
3
Hp ~y Hyp
M Mo
. H. . H
. |__\ j\/\uu e .
o T Ha Ha %
, 07 CHa
H?. ~ F{Z *
0 {3)

Solutions containing 2-methyltetrahydrofuran were
studied over a range of concentrations and doses
since the hydrogen abstraction, step (3), proceeds
with greater ease in the case of a tertiary hy-
drogen.

The chromatographic analyses show that con-
siderable guantities of products are formed in
the molecular-weight range of dimers and simple
telomers. The results for solutions of dihydro-
furan in Z-methvltetrahydrofuran indicate that the
major product changes trom a dimer or 1:1 adduct
to a trimer or 1:2 adduct when the dose exceeds
about 8 x 107 rads. Increasing the concentration
of dihydrofuran in the solution favors the lower-
molecular-weight product at the higher doses.
Comparison of solutions of tetrahydrofuran and
124

Table 14.1. Compounds lrradiated in a Survey of Addition Reactions

 

Saturated Compounds (Telogens)

Unsaturated Compounds

 

 

 

Name Structure
HE o H2
Tetrahydrofuran Hy L 1 H2
0
1‘12 s H2
2-Methyltetrahydrofuran L H
Hp L
0~ CHj
Hp Ha
2,5-Dimethyltetrahydrofuran H H
CH3 Q CH3
CHgz CHz
Diisopropy! cther H({—‘O"CIZH
[ [
CH3 CH3
g
Isopropyl alcohol HC —OH
I
Cl"f3

 

 

Structure

 

Name

Cyclohexene

H H
Dihydrofuran \[—-.—.,j/

Ho H,
\O

H H

2,5-Dimethylfuran
CHgz 0 CHz
H H

Furan

 

2-methyltetrahydrofuran in both dihydrofuran and
cyclohexene suggests that tetrahydrofuran gives
higher yields than the 2-methyl compound.

Evaporation of the irradiated mixtures indicated
that over 20% of the original materials had been
converted to nonvolatile residue at doses of about
4 % 10® rads. Initial yields in terms of consump~
tion of dihydrofuran in mixtures with methyltetra-
hydrofuran were estimated at G values (molecules
per 100 ev) of about 10.

Infrared spectra indicated the formation of small
quantities of hydroxy and carbonyl compounds.
Mixtures of telogens with furan and 2 5-dimethyl-
furan gave very little 1:1 or higher telomer,
probably because of resonance stabilization of the
furan ring.

Further work will involve identifying the products
indicated by gas chromatography and measurement
of yields at elevated temperatures.
15. Fluoride Studies for Other ORNL Programs

THE CHEMISTRY OF CHROMIUM IN THE
FLUORIDE VOLATILITY PROCESS

B. J. Sturm R. E. Thoma

In earlier versions of the Fluoride Volatility
Process, volatile UF _ was obtained by fluorina-
tion of fluoride melts in which solid fue!l elements
had been dissolved by treatment with HF. For
zirconium-’ aluminum-based® fuel elements
the method worked well because the available low-
melting salt mixtures permitted the fluorination
at temperatures of 600°C or lower, where corrosion
was not severe. Fuel elements with matrices and
claddings of stainless steel proved less tractable;
a process in which UF _ is recovered by volatiliza-
tion from a fluidized bed instead of a melt® looks
promising for these fuels. Uranium hexafluoride
from both processes is purified by adsorption-
desorption cycles on beds of NaF,

In the fluidized-bed procedure the fuel elements
are supported in a fluidized alumina powder, which
acts as a heat transfer medium, and are decom-
posed by treatment with HF-0, mixtures. Uranium
is then recovered as UF 5 by fluorination. When
chromium, an important constituent of stainless
steel, reacts with the HF-0O, mixture, it forms
nonvolatile oxides and fluorides. On fluorination,
chromium is oxidized to higher valence, and large
quantities of its volatile fluorides and oxyfluorides
form. These compounds might volatilize with
and contaminate the recovered UF , Little is
known of these volatile compounds of chromium,
Reactor Chemistry Division researches performed

or

-

(38

 

lohem. Tech. Div. Ann. Progr. Rept. May 31, 1963,
ORNL-3452, p. 26,

2Chem. Tech. Div. Ann. Progr. Rept. May 31, 1564,
ORNL-3627, p. 40.

SChem. Tech. Div. Ann. Information Meeting Program,
Nov. 10--12, 1965, pp. 5~7.

125

in support of the ORNL Fluoride Volatility Process
program have, accordingly, been devoted to the
chromium compounds which may affect the new
method.

Synthesis of CrF4

Freesenergy estimates® indicated that the re-
action

AgF 4+ CiF, 2 CF , + AgF , AF = =29 keal ,

might be used for synthesizing CeF,. In laboratory
tests, we found that dense blue-black CiF, vapors
were evolved above 400°C, the boiling point of the
compound, when equimolar amounts of the re-

actants were present or when there was an excess
of AgF .

Reaction of Chromium Fluerides with NaF

Since UF_ is purified by absorption on NaF
beds, the behavior of chromium fluorides with NaF
was studied, Two compounds, 3NaF - CrF, and
SNaF - 3CiF |, formed during the crystallization
of molten NaF-CtF, mixtures containing less
than 50 mole % CrF,. The former compound (ap-
parently a structural analog of cryolite) exists in
two crystal modifications. The high-temperature
form melted reproducibly at 1090°C and exhibited
a range of solid-solution stoichiometry, The low-
temperature form is biaxial, with an average re-
fractive index of 1.411. X-ray diffraction pattems
from 5NaF~3CrF3 powder suggest that the coms
pound, melting incongruently at 800°C, is iso-
morphous with chiolite (5NaF=3A1F3). In the

 

4A‘ Glassner, The Thermochemical Prog’:)erties of the
Oxides, Fluorides, and Chlovides at 25007 K, ANL-5750
(1957).
composition region 0 to 50 mole % CrF,, the sys-
tem NakF-CrF | has two invariant points: a eutectic
at 14 mole % CrF, and 873°C and a peritectic near
42 mole % CrF , and 800°C,

An NaF-CiF phase was synthesized by expos-
ing granulated NaF to CrF, vapor at 400 to 500°C
in a controlled-atmosphere glove box. This product
had a face-centered cubic structure with a lattice
constant of 7.90 A, Since it is isostructural with
the high-temperature form of KZCIFG,S
pound is believed to be Na,CrF .

the coms-

Behavior of Cr02F2

Chromium oxyfluoride has been prepared® by
reaction of CoF, and CrO, and has been exposed
in the vapor state for 0.5 hr at 80°C to several of
the materials expected to contact the Fluoride
Volatility Process gas streams. Sodium fluoride
reacted with the CrO I, to produce an (as yet
unidentified) yellow-orange compound. Aluminum
fluoride, UF,, and UO, showed no evidence of
reaction.

Equilibrium data’~? for the reaction

2HF () + Cr0,(s) & CrO,F (&) + H,0(8)

suggest that AG for solid CrO,F, at 298°K should
be near — 189 kcal/mole. From this value one
predicts that CrOF should oxidize CeF, to CrF,
uo, to UO,, and (by 2 kcal per gram atom of F)
UF, to UF,. No tests of the first two reactions
have been made; UF4 was not oxidized to UF5
during the relatively mild exposures above.

PREPARATION OF LiF SINGLE CRYSTALS
BY THE MODIFIED STOCKBARGER
METHOD

R. G. Ross R. E. Thoma

As part of the AEC Pure Materials Program,
sustained efforts have been made within the Re

 

SH. C. Clark and Y. N. Sodana, Can. J. Chem. 42, 50
(1964).

6G. D. Flesch and H. J. Svec, J. Am. Chem. Soc. 80,
3189 (1958).

A, Engeibrecht and A, V. Grosse, J. Am. Chem. Sac.
74, 5262 (1952).

89. A, Munter, O. T, Sepli, and R. A. Kossatz, Ind.
Eng. Chem. 39, 427 (1947).

9R. A, Oriani and C. P. Smyth, J. Am. Chem. Soc. 70,
125 (1948).

126

actor Chemistry Division to develop techniques
and apparatus required for the production of very
pure, large (350 g) single crystals of LilF with
selected isotopic ratios. In a previous report10 we
indicated that by refinement of standardized
techniques’! for preparing lithium fluoride single
crystals, we could routinely produce such crystals
in which metallic impurity concentration does not
exceed 30 ppb. Crystals were grown from molten
LiF in capsules of grade ‘A’ nickel, The single
contaminant detected in the LiF product by the
best analytical methods available is manganese,
its source is the nickel capsule wall, from which
Mn® and/or Mn2* diffuse into the melt.

The principal application of these LiF crystals
has been; in testing theoretical models telating
the variation of the thermal conductivity of the
crystals with their ®Li-71.i ratio. For this purpose,
workers at the Cornell Materials Science Center!?
have found that crystals must be sufficiently pure
that the impurity does not mask the isotopic effect.
While adequate for most studies which require
very pure crystals of LiF, material which contains
as much as 1 ppm Mn?" is insufficiently pure for
testing theoretical models under consideration.
In attempts to improve the purity of LiF crystals
further, we have modified the Stockbarger ap-
paratus to grow LiF crystals in a capsule lined
with laboratory-grade platinum,
has proved to be very effective in improving the
chemical purity of crystalline LiF. A 270-g crys-
tal of LiF, designated ORNI.-11, and containing
99,99 at. % ‘Li, was grown in the platinum-lined
capsule. As in previous ORNIL preparations, the
crystal was vacuum annealed and handled sub-
sequently only in vacuum or inert atmosphere,

This innovation

The crystal was found to contain a lower concen-
tration of heavy-metal contaminants than is cure
rently detectable by activation analysis. 'That is
the
most likely contaminant, is <1,86 parts per billion,
and the crystal is purer than any of our previous
efforts.

to say that the concentration of manganese,

 

10¢, F. Weaver et al., The Production of LiF Single
Crystals with Selected Isotopic Ratios of Lithium,
ORNL-3341 (March 1964),

llp B, Thoma et al., Reactor Chem. Div. Ann. Progr.
Rept, Jan. 31, 1965, ORNL-3789, p, 323.

12p, D, Thacher, Thermal Conductivity Studies of
Phonon Scattering by Boundaries and Isotopes in
Lithium Fluoride Crystals (thesis). Report 369, Ma-
terials Science Center, Cornell Univ,, Ithaca, N.Y.
(June 1965),
16. Chemical Support for the

Controlled Thermonuclear Research Program

R. A. Strehlow

Whether controlled thermonuclear devices with
useful power densities are feasible is a question
which can be conveniently divided into two prin-
cipal parts. The first part asks whether a plasma
fuel element can be constructed; that is, whether

the physics of plasma confinement can be solved’

in a practical way. The second part, which is more
obscure, asks whether there is any fundamental
bar {o construction of a thermonuclear reactor if
the confinement problem is solvable. We have
continued to study chemical aspects of both parts
of this question.

Our studies in support of the confinement prob-
lems are directed toward development of tech-
niques to diagnose the quality of the plasma
environment in experimental devices and to assist
in attempts to improve this quality. Such pursuits
have dominated our chemical studies.!'™® In addi-
tion, an extensive literature survey and an experi-
mental study have been devoted to possible prob-
lems of fritium inventory and the fuel cycle of a
thermonuclear reactor. Some of our special equip-
ment was also used to assist in the design (by
ORNL Isotopes Division personnel) of an isotope
separator of a new and improved type. Each of
these portions of our effort is described briefly
in the following.

 

lReactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNI~3789.

2rhermonuclear Div. Semiann. Progr. Rept. Apr. 30,
1964, ORNL-3652, pp. 13845,

3Thermonuclear Div. Semiana. Progr. Rept. Oct. 31,
1964, ORNL=3760, p. 95,

127

YACUUM ANALYSIS IN AN EXPERIMENTAL
PLASMA DEVICE

R. A, Strehlow

A simple mass-spectrometric assembly has been
installed and operated routinely to yield a continu-
ing record of vacuum conditions during operation
of the plasma device DCX-2. About 2500 mass
spectta have been obtained in this study since
installation of this residual-gas analyzer in Feb-
ruary 1965. Such data have proved helpful in
assaying various pumping and operating parameters
of DCX-2; such species as acetylene, trichloro-
ethylene, methane, ethane, alcohols, and vapors
of metals (such as Zn) can be detected unambig-
uously. The mass spectra alone are insufficient
to yield a real diaguosis of the DCX-2 vacuum
environments since, in many cases, the origin of
the observed species is more important than the
fact of its presence. Accordingly, a study is under
way of the effect of processes within DCX-2 upon
the mass spectrum obtained from pertinent materials.

For this study a second mass-analyzer assembly,
similar to that on DCX-2, has been installed on a
simple, large vacuum system whose construction
materials, assembly methods, pumps, pump .oils,
etc., are similar to those of DCX-2. This assembly
pemmits study of the effects of various analyzer
instrument parameters (electron-~accelerating volt-
age and ion-source assembly procedure, for example)
on the spectrum observed. More important, how-
ever, it permits study of the effect of processes
such as electron and ion bombardment, temperature
fluctuations, titanium getter evaporation, and
others used in DCX-2 on the mass spectrum ob-
served.

With this assembly, mass-spectrometric studies
have been made of the effect of warming liquid-
nitrogen-cooled surfaces,'*? of hot-filament reac-
tions and electron bombardment of contaminated
surfaces,? and of reactions of titanium in vacuum
systems,®> Recent studies have dealt with the
formation of products from bombardment with hydro-
gen ions during glow discharges within the as-
sembly. These studies suggest that acetylene is
observed, in addition to methane, only when a
trisiloxane oil is present in the bombarded materials.
Acetylene, previously observed in DCX-2 during
ion injection of I12+ into the assembly, is now be-
lieved to be evidence of contamination by this
pump-oil component,

INTERACTION OF TRITIUM WITH
THERMONUCLEAR-REACTOR MATERIALS

S. S. Kirslis

For a thermonuclear reactor to be feastble, the
holdup of tritium in the metal of the containment
vessel (probably molybdenum or tungsten) and in
that of the ion sources (alloys of nickel or iron)
must not result in excessive tritium inventory and
decay losses. The equilibrium solubility of hydro-
gen in these metals at low pressures is low. How-
ever, the metals can be cathodically ‘‘surcharged”’
with hydrogen to amounts of the order of 100 c¢m?
(NTP)/g; this is sufficient to cause, in some
cases, blistering and cracking of the metal. In
order to judge more clearly whether the metal in
a reactor under bombardment by energetic tritium
atoms might be surcharged with tritium, a literature
study of the various possible surface and interior
diffusion processes has been made; this survey
will be published separately.

The 50- to 100-kev
pected to penetrate the metal to their range depth
of 1 to 2 x 107* cm. They presumably diffuse
inward (and also back to the entrance surface)
according to normal diffusion laws. If there is
no appreciable barrier to their egress at the sur-

tritium atoms may be ex-

face, the concentration of tritium atoms just in-
side the metal will be low or zero. If all this is
true, the steady-state concentration of tritium in

 

D, M. Richardson and R. A. Strehlow, Trans. Natl.
Vacuum Symp., 10th, 1963, p. 97.

128

the metal will have a peak at the range depth and
will fall off to low values at the bombarded sur-
face and at the far surface of the metal. The
average concentration in the metal would be half
the peak concentration. If a barrier to egress of
tritium from the bombarded surface exists, the
whole steady-state concentration curve would cor-
respondingly rise to higher tritium concentrations,
The factors determining the peak concentration of
tritium are the flux of atoms to the metal, the dif-
fusion constant of tritium in the metal, and the
boundary condition for the degassing of the bom-
barded metal surface.

Information on hydrogen ditffusion in the metals
of interest was derived mainly f{rom published
permeability and solubility studies. For informa-
tion on the degassing boundary condition, the
literature was searched in the fields of chemisorp-
tion and desorption, catalysis, hydrogen-electrode
behavior, and hydrogen-atom recombination, Other
peitinent information was obtained from published
studies of ion bombardment, spuftering of metal
surfaces, and saturation of metals by proton or
deuteron bombardment.

The conclusions of the literature study are as
follows:

1. Hydiogen is easily degassed at low pressures
and moderate temperatures, even from metals
(such as tungsten) for which the chemisorption
heats are high.

Cathodic surcharging depends on the poisoning
of the metal surface for hydrogen-atom recombi-
nation. In a gaseous environment on clean
metals, hydrogen-atom recombination is rapid.
Diffusion rates for hydrogen in molybdenum at
600 to 800°C are sufficiently rapid to deliver
injected atoms rapidly back to the bombarded
surface, thus maintaining a very low (and from
a practical viewpoint, negligible) concentration
of hydrogen in the metal, The same is expected
to be true for tritium. If the hydrogen concen-
tration will indeed be as low as estimated from
the simple diffusion model, it should cause no
hydrogen embrittlement problems. Since the
diffusion constants for nickel and iron are much
higher than those of molybdenum, high internal
hydrogen concentrations are not expected even
at much lower temperatures.

Some recent work on deuteron bombardment of
metals indicates higher of
deuterium just inside the metal surface than

concentrations
129

predicted by the simple model. Some evidence
indicates this may be the effect of surface
contamination on the metal,

HYDROGEN SURCHARGING OF MOLYBDENUM
IN A GLOW DISCHARGE

D. M. Richardson

A study of hydrogen occlusion by a molybdenum
cathode in a glow discharge has been conducted.
The principal objective of this work was to assess
the significance of hydrogen surcharging to design
considerations for themmonuclear reactors. Initial
studies had indicated a possible hydrogen content
of one or more atmospheric cubic centimeters per
cubic centimeter of metal after bombardment in a
discharge. Though this concentration of hydrogen
is less than 0.1 at. %, even this low value could,
if it were typical of the concentration in a large
fraction of the reactor, present serious problems
to the reactor designer. The work described here
and the literature study summarized above, how-
ever, have led to the conclusion that, at elevated
temperatures, occlusion of hydrogen is not large.
At lower temperatures it appears that surface
contamination can matkedly impede the recombina~
tion of hydrogen during even low-enetgy bombard-
ment; very high hydrogen concentrations in the
metal are able, therefore, to be achieved.

For these studies, a 45-cm length of molybdenum
wire 1 mm in diameler was used as a cathode in a
cylindrical tank with a volume of 200 liters. The
ends of the tank were electrically shielded with
sheet Teflon, The center of the tank was cone
nected to a vacuum system through a 6-in. isolation
gate valve and a gate valve with a small orifice
which had a speed of 2 x 107* the speed of the
vacunm-system manifold., This allowed a steady
introduction of hydrogen gas through a palladium
leak to the glow discharge chamber at a pressure
of 1072 to 1 torr with continuous mass analysis
of the effluent gas from this chamber. The gas
during the usual discharge contained not more than
1 part per thousand of nonhydrogen impurity.

The conditions for the usual low-temperature
discharge are summarized in Table 16.1. The total
bombardment for the conditions shown was 12
amp-sec or about 1 to 3 atm-cm® of hydrogen gas.
After opening the valve and pumping down to a

Table 16.1. Conditions Employed in Experiments
with Glow Discharges

 

Wire volume, cm® 0.35
Wire area (apparent), em? 14
Pressure, torrs 0.25
Discharge current, ma 10
Discharge time, min 20
Applied voltage to cathode, v 500

 

pressure of about 1 x 1077 torr, the wire was
heated to redness. Mass-spectrometric and ione
gage observations were made of the hydrogen
evolved by the depassing process. Pressute rises
of as much as 6 x 10™* torr were observed during
the 4-sec degassing procedure, Since the system
pumping speed for hydrogen was about 1000 liters/
sec, this pressure rise corresponded to about 2
atm-cm® of hydrogen,

Stringent cleaning of the wire and ion bombard-
ment of the chamber walls for several hours, fol-
lowed by repetition of the discharge, led to a value
of only 0.017 atm-cm®. Continued repetitions of
the discharge-degas cycle led to increasing amounts
of hydrogen being occluded. Since the wire was
not heated past 1000°C, a gradual increase in
surface contamination could have been responsible
for the very high values.

Several attempts were made at higher current
densities (and consequent high temperatures) to
determine the amount of hydrogen occluded during
a usual discharge. None of these attempts led to
an evolution of hydrogen in detectable amounts,

Molybdenum surfaces which are clean or which
are heated do not seem to occlude large quantities
of hydrogen during low~intensity bombardment
with protons,

MEASUREMENT OF GAS LOAD FROM SOURCE
OF ELECTROMAGNETIC SEPARATOR

R. A. Stehlow

Scientists of the ORNL Isotopes Division are
designing an improved electromagnetic isotope
separator. This separator, a considerably modified
calutron, is to have separate differential pumping
systems for the source, collector, and main-vacuums-
tank regions. For design of these pumping systems,
130

it was necessary to know the gas load upon each;
that from the source was judged to require experi-
mental measurement.

ORNL electromagnetic separators use, where
possible, vapor of the chloride of the element
whose isotopes are to be separated., This chloride
is generated within the source assembly by chlori-
nation with CCl, of the appropriate metal or oxide.
The gas load from the source assembly, therefore,
consists almost entirely of unreacted CCl1 .

A flow-rate meter with a constant (regulated)
pressure and variable volume, capable of operation
with condensable pases and at low pressures, was
adapted for this study., This device, designed to
measure flow rates as low as 10™° torr-liter/min
for the thermonuclear support propram, was installed
near the source region of a calutron. The CCI4
flow rate from this unit was monitored during a run
in which the isotopes of cerium were being sep-
arated. The measured flow rate, higher by nearly
an order of magnitude than that previously esti-
mated, has been used in design of the differential
pumping system for the source of the new unit,

ELECTRON ENERCY
5.4 15.3

14.7 14.9

ION INTENSITY

 

2.7
ELECTRON ENERGY FOR ETHAME FRAGMENTS (ev)

12.5 2.2

 

APPEARANCE-POTENTIAL MEASUREMENTS
FROM TIME-OF-FLIGHT MASS SPECTROMETRY

J. D. Redman

Polymeric species are commonly observed in the
vapor of simple and of mixed metal halides, We
hope to assist in interpreting mass spectra from
such halide vapors by determining appearance
potentials of the various ions formed by electron
impact with the wvapor species evolving from an
effusion cell. A preliminary stage of this work
has been completed with the construction of a
retarding-potential circuit and its application to
the study, with a time-of-flight mass spectrometer,
This
assessment was considered necessary, even though
application of retarding-potential-difference methods

of various gases to assess its reliability.

to this type of spectrometer has been demonstrated,
to determine whether gases introduced at the low
pressures corresponding to the expected flux of
salt-vapor species would yield appearance poten-~
tials with satisfactory precision.

ORNL- DWG £6-980

FOR KRYPTCN (ev)
15.5 15.7 15.2

6.7

 

o
\goa
—— ) \\

 

 

 

A e
ors S|
s
0.44 ){/ ! :
! L
7 |
|
—— 1
134 133 135 3.7 139 44 143 145

Fig. 16.1. Normalized lonization-Efficiency Curves for Krypton and for C2H4+, CoHg +, and C?_Hé'+ from Ethane.
The assembly was standardized by a series of
careful appearance-potential (AP) measurements
using krypton. We obtained the value 13.51 % 0,04
ev from these measurements; this is less by 0.5
ev than the generally accepted value obtained
by other investigators.>”'! We have ascribed
this discrepancy to (reproducible) contact potentials
and, perhaps, other systematic errors in vur as-

 

C. E, Melton and W, H, Hamill, J. Chem. Phys. (to
be published).

®R. E. Fox et 2l., Phys. Rev. 89, 555 (1L953),

D, C, Frost and C. A, MeDowell, Proc, Roy. Soc,
(London), Ser. A 232, 227 (1955).

8]. F., Burns, Nature 192, 651 (1961).
Y. Kaneko, J, Phys. Soc. Japan 16, 1587 (1961).

195, N, Foner and B. H. Hall, Phys. Rev. 122, 512
(1961).

g, 1, Field and J. L. Franklin, Electron Impact
Phenomena, Academic, New York, 1957,

131

sembly. Appearance potentials for other gases
were obtained from ionization-efficiency curves,
normalized for this systematic deviation at onset,
obtained in the standardization with krypton.
Typical normalized ionization-efficiency plots are
shown in Fig. 16.1 for krypten and ethane frag-
ments. Our values for other appearance potentials
are compared in Table 16.2 with those obtained
by others. We fail to check the published values
for the C2H4+ and C21-15+ fragments from butane,
but agree quite satisfactorily with published values
for all others studied.

Since the operating source pressure was main-
tained at 5 x 1077 to 1 x 107 % torr in these studies,
we believe that vapor fluxes from effusion cells
containing metal halides will prove adequate for
the planned investigation. It is likely that temper-
ature control of the effusion-cell system will be
of prime importance,
132

Table 16.2. lonization Potentials for Fragment Peaks of Argon, Nitrogen, Ethane, and Butane

Mormalized for Yariation of Krypton AP from Literature Value

 

Number of

 

Fragment Reactant D . . Ionization Potentials (ev) Reference
eterminations
Art Ar 6 15.65 + 0.15 16.00 + 0.10 16.32 + 0.14 This work
Art Ar 15.74 a
ar’t Ar 15.77 b
, * , 6 15.65 = 0.18 16.85 + 0.19 This work
+
, ) 15.60 b
C,H, * C,H, 5 12.28 + 0.08 12.76 + 0,20 13.42 + 0.10 This work
+ .
C,H, C,H, 12,10 12.70 13.20 c
+
C,H, C,H, 12.10 5
JHo * C,H, 3 12.80 + 0.2 13.39 + 0,14 13.78 + 0.23 This work
+
JH, C,H, 12.10 12.55 13.40 c
+
C,H, C,H, 12.80 b
C,H, * C,H, 4 11.66 * 0.17 12.00 + 0,10 12.84 + 0.15 This work
+
C,H, C,H_ 11.60 12.10 12.65 c
+
C 1, C,H, 11,60 b
, 4* C,H,, 2 12.26 * 0.24 12.86 + 0,1 13,52 + 0.3 This work
+
H, CH 11.40 B
C2H5+ C,H, 3 12.75 +0.25 13.27 + 0,02 13.77 +0.09 This work
+
C,H, C,H, 12.10 b
+ e
C,H, C,H 1 11.40 11.60 12.10 This work
+
C,H, C,H 11.50 11.80 12.20 c
+
C,H, C,H 11.70 B
+
\ ,{, _J‘._ -
C,H,, C,H 4 10.68 + 0.22 11.11 + 0.20 This work
+
C,H, C,H,, 10.60 11.00 c
+
C,H,, C,H,, 10.80 b

 

2C. E. Mclton and W, H. Hamill, J. Chem. Phys. (to be published).
bE. H. Field and J. L. Franklin, Electron Impact Phenomena, Academic, New York, 1957,
°C, E. Melton and W, H. Hamill, J. Chem. Phys. 41(7), 546 (1964),
Part V
Nuclear Safety

 
17. Nuclear Safety Tests in Major Facilities

FiSSION PRODUCTS FROM FUELS
UNDER REACTOR-TRANSIENT CONDITIONS

G. W. Parker R. A. Lorenz
J. G. Wilhelm*

Miniature fuel elements are melted in a special
assembly in the TREAT to study the release of
radioactive material when the fuel cladding and
the fuel are melted or vaporized rapidly, as in a
nuclear accident resulting from a reactor tran-
The program attempts to measure and inter-
pret the effect upon fission product release. of
fuel type, cladding, atmosphere during the tran-
sient, inventory of fission products, and charac-
teristics of the transient. The extent of reaction
of the cladding and of the UO _ fuel with the cover-
ing atmosphere is also determined.

sient.

Behavior on Melting in Steam

Analysis of two experiments which used atmos-
pheres of 1000 psia steam (285°C) has been re-
ported elsewhere.? One experiment used a spec-
imen of UQ, with stainless steel cladding; during
exposure in TREAT the cladding melted, oxidized
extensively, slumped in a spongy mass, and ad-
hered to the unmelted UO, pellets. Fission product
release from fuel and cladding was much less
than that in an earlier experiment, which used
previously unirradiated stainless-steel-clad UO_ in
45-psi argon atmosphere, where the cladding flowed
completely off of the unmelted U0, pellets. Appar-
ently, the layer of oxidized cladding formed by the

 

10n assignment from Karlsruhe Center for Nuclear
Research and Development, Karlsruhe, West Germany.

2. WL Parker, R. A. Lorenz, and J. G. Wilhelm, Nucl.
Safety Program Semiann. Progr. . Rept. June 30, 19635,
ORNL-3843, pp. 39-67.

135

high-pressure steam provided an effective barrier
to fission product release.

The second experiment nsed a specimen with
Zircaloy-2 cladding on the UO, fuel; after the

exposure, fuel and cladding were found to be
fragmented and dispersed within the alumina
crucible. Direct comparison of fission product

release from the stainless-steel- and the Zircaloy-
clad specimens cannot be made since plugged
tubing prevented release of steam from the auto-
clave after the transient with the Zircaloy-clad
specimen. However, the distribution of fission
ptoducts within the fuel autoclave showed that
the release was greater for the Zircaloy-clad
specimen; this must be considered at present to

be a result of the fragmentation.

Behovior on Melting Under Water

The first two of a series of experiments have
been performed with the TREAT reactor in order
to study the release of fission products from fuel
melted underwater. The experiments used stainless-
steel- and Zircaloy-2-clad UQ, fuel specimens
irradiated to 18 Mwd/metric ton. Aninitial specimen
temperature of 70°C was used, and operating and
reactor-transient conditions for the two experiments
were identical. Heat input to the fuel during the
transients. was approximately 500 cal per gram of
UO2 (50‘?0: greater than any previously used); the
U0, was heated to well above its melting point.

The experiments were designed to simulate tran-
sient accidents with water-cooled reactors in which
steam and water are expelled from the core vessel.
Valves were opened immediately after the transient,
and the transient-generated steam and a purge of
argon gas were permitted to flow through a con-
denser, water-collection traps, filters, and a
charcoal bed into a gas-collection tank. The fuel-
containing autoclave was then electrically heated
136

to 350°C maximum for 1 hr to slowly boil excess . 2 . o vesers
water out of the fuel autoclave. '
The release and distribution of fission products
were essentially the same in the two experiments.
The fuel specimens melted completely (Figs. 17.1
and 17.2). Fission product release to the fuel
autoclave was high, but transport of fission products
out of the autoclave by steam release and argon
purge was relatively small. The '?°Te, '*7Cs,
and 13! carried out of the fuel autoclave ranged
from 2 to 7%. The transported release of non-
volatile materials (°°Zr, '*'Ce, and UO ) was
only 0.1%. Approximately 0.005% of the 311
reached the filter papers and charcoal bed.

Fig. 17.1. Puffy-Appearing Cake of Melted Fuel and
Cladding from Stainless-Steel-Clad UO?_ Melted Under
Water by Transient Heat Input of 504 cal per Gram of
UO2 in TREAT Experiment 7 (Front Malf of Crucible

Remowved). ' .

 

PHOTO B1912

 

Fig. 17.2. End Caps ond Melted Fuel and Cladding in Sample Holder from TREAT Experiment 8 in Which
Zircaloy-2-Clad UC)2 Was Melted Under Water by Transient Heat Input of 511 cal per Gram of U02 (Flux Monitor

Capsule in Place on Back Half of Crucible). An unmelted fuel specimen is shown dlongside for comparison.
137

Most of the transient-generated steam condensed
in the fuel autoclave, and only a small amount
was immediately discharged through the condenser.
The water inside the fuel autoclave was an excellent
trap for the fission products since slow evapora-
tion of the remaining water (simulating accident
after-heat) did not transport large amounts of
fission products. Approximately 1 liter of hydrogen
was produced in each experiment by reaction
between c¢ladding and water; this resulted in a
chemically reducing atmosphere for most fission
products.

Metallographic examination of portions of the
melted fuel samples showed the formation of a
eutectic phase in the specimen with stainless
steel cladding. The major phase in the sample
from the Zircaloy-2-clad specimen was a solid
solution of approximately 75% UO, and 25% Zr:()z.

Two additional experiments were recently per-
formed with fuel specimens and reactor transients
similar to those described above. The initial
temperature was increased and the argon purge was
eliminated in order to have quicker and more com-
plete release of steam from the fuel autoclave in
the first minute following the transients. Examina-
tion of these experiments is in progress.

FISSION PRODUCTS FROM SIMULATED
LOSS-OF-COOLANT ACCIDENTS IN ORR

W. E. Browning, Jr.
C. E. Miller, Jr.

W. H. Montgemery
B. F. Roberts

R. P, Shields

0. W. Thomas?
A. F. Roemer?
J. G. Wilhelm®

Release of fission products and their subsequent
behavior under a variety of conditions which might
occur in loss-of-coolant accidents to nuclear re-
actors is the subject of a continuing study. The
loss-of-coolant accidents are simulated in a versa-
tile experimental assembly in the Oak Ridge
Research Reactor.®*®  This assembly permits

 

3General Engineering and Construction Division.
4Analytical Chemistry Division.

SW. E. Browning, Jr., ef al., Reactor Chem. Div. Ann.
Progr. Rept. Jan. 31, 1961, ORNL-3127, pp. 149~52;
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962,
ORNI.-3262, pp. 172-76.

‘w. E. Browning, Jr., et al., Reactor Chem. Div. Ann.
Progr. Rept. Jan. 31, 1965, ORNL-3789, p. 263.

attainment of controlled temperatures (through
nuclear heating of the small fuel specimen) to and
above the melting point of UO, with any of sev-
eral atmospheres. The release and the subsequent
deposition of eight fission products (I, Te, Cs, Ru,
Sr, Ba, Zr, and Ce) are determined in each of the
experiments. Results of these investigations are
published in detail elsewhere. ”

The atmospheres that have been investigated
include helium, moist helium, steam-helium-hydro-
gen mixtures, dry air, moist air, and steam-air
mixtures, Of the eight elements measured, only
iodine and ruthenium showed significant variations
in behavior with changes in atmosphere. lodine
and ruthenium were both more volatile in a moist
air than in any other atmosphere studied and were
transported farther under these conditions. lodine
passed through absolute filters but was retained
on charcoal; ruthenium was more easily filtered
in the volatile form (from moist air) than in the
less volatile form. Ruthenium was not volatile
in an atmosphere containing appreciable guantities
of steam. The release of cesium from the 1000°C
zone was lowest in experiments in which steam
was a component of the atmosphere.

Cladding material near the heated fuel is an
abundant, strongly reducing reagent and can be
expected to affect fission product behavior. Stain-
less steel appears to retain ruthenium and, under
oxidizing conditions, to lower the melting point
of UO,. Experiments with Zircaloy cladding are
in progress. Out-of-pile experiments* on the
effects of stainless steel and Zircaloy cladding
have shown that the efficiency of trapping iodine
by steam condensation is increased by the presence
of these materials.

Three experiments have been performed to study
the effect of high burnup of fuel (>>20,000 Mwd/ton)
on fission product behavior. Suchfuel was examined
metallographically and was found to be typical
of high-burnup fuel used in power reactors of
advanced design. Cesium and ruthenium were the
only elements affected by burnup; the fractional
release of ruthenium decreased with increasing
burnup, while that of cesium increased.

Effects of maximum temperature of the fuel are

also being investigated. In these experiments,

 

7W. E. Browning, Jr., et al., Nucl. Safety Program
Semiann. Progr. Reptf. June 30, 1965, ORNL-3843,
pp. 3—39; and Nucl. Safety Program Semiann. Progr.
Rept. Dec. 31, 1965, in preparation.
performed under oxidizing conditions in which the
maximum fuel temperature was maintained at ap-
proximately 2000°C, the UO  appeared to have
melted due to the formation of a [ow-melting eutectic
with stainless steel oxide. The releases of the
eight elements in such intermediate-temperature
experiments were similar to releases in experiments
in which complete melting occurred at approximately
2900°C. Future experiments are planned at less
than 2000°C and under reducing conditions.

The primary purpose of these in-pile experiments
is, of course, to permit prediction of behavior of
the fission products in possible reactor accidents
with reasonably assumed conditions. Accordingly,
assessment of the form and the physical and
chemical characteristics of the released materials
is an essential part of the program.
data with a fibrous filter suggest that particles of
two size ranges are important in transport of fission
product activity. Analysis by composite diffusion
tubes shows three nonelemental iodine species in
the gas. Studies with composite diffusion tubes
have also shown that desorption of iodine from
the apparatus after a simulated accident is small;
only about 0.1% of the original inventory was
desorbed,
form.

Preliminary

and that primarily in a nonelemental

The aging of fission product aerosols following
release from the fuel may substantially alter the
form and therefore the behavior of the fission
products as a function of time. In order to investi-
gate this effect, a simulated reactor containment

vessel is being built into the in-pile facility, Future
experiments with this extended facility will treat
aerosol aging.

FISSION PRODUCTS FROM HIGH-BURNUP UO,

G. W, Parker R. A. Lorenz
W. M. Martin C. J. Barton
G. E. Creek

In the first experiment of a series designed
piimarily to test the effect of burnup on release
and behavior of fission products, a re-irradiated
specimen of stainless-steel-clad UO, previously
irradiated to 1000 Mwd/ton was melted in the
Containment Mockup Facility (CMF).® The specimen
was melted by induction heating in a steam-air
atmosphere; the released fission products were
aged in a stainless steel tank filled with pres-
surized steam-air mixture.

Distributions of released ‘‘real’’ fission products
are compared in Table 17.1 with those for simulated
fission products from previous experiments in the
CMF. The comparison tends to validate the use
of simulated high-burnup UO  fuel. Deposition
of fission products before they reached the stainless
steel aging tank was probably affected by the
different furnace-tube geometry of this experiment

 

8G. W. Parker et al., Reactor Chem. Div, Ann. Progr.
Rept. Jan. 31, 1965, ORNL-3789, p. 251.

 

 

 

 

 

Table 17.1. Distribution of Fission Products Released from Simulated Fuel (8)
and from High-Burnup U02 (HB)
Fission Products Found (% of total inventory)
L . Cesium Tellurium Ruthenium Strontium Iodine
ocation o N e _
HB S HB S HB S 1B S HB S
Fumace tube and duct 25.0 3.0 18.1 0.3 Q.26 0.07 0.047 0.01 15.0 10.1
Tank walls and
deposition samples 8.5 15.1 12.4 6.9 0.054 0.35 0.015 0.04 27.0 33.8
Condensate 12.6  43.6 0.7 0.8 0.001 0.12 0.005 0.0003 57.6 56.0
Filters 2.1 0.8 0.5 0.45 0.002 0.0006 0.0001 0.0004 0.35 0.05
Total release from fuel 48.2 62.5 31.7 8.4 0.32 0.54 0.07 0.05 ~100 ~100
from that of earlier experiments. However, few
differences in the distribution of fission products
reaching the tank were noted. The volatile fission
products, lodine, cesium, and tellurium, were the
only activities other than rare pases reaching the
tank in significant amounts; a large fraction of
these activities remained on the tank wall or was
collected in the condensate after 3 hr aging in the
tank with gradually decreasing pressure and
temperature.

Two different types of samplers were used to
study concentrations of gas-borne activity in the
pressurized tank during the aging period. The
two samplers gave comparable results, and the
total airborne gamma activity (mainly from iodine,
cesium, and tellurium) indicated by these samplers
agreed well with gross-activity values from a
collimated ion chamber adjacent to the side of
the tank. Figure 17.3 shows the variation with
time in concentration of the various gas-borne
activities.

ORNL-—-WG 66-254

N

 

  

T RE

 

 

PRODUCT INVENTORY IN TANK ATMOSPHI

 

 

 

 

 

 

 

 

 

 

 

=
5 \ |
D000 N e
2 . U
e A
L.
< 5r
i~ |
=z
it
OO0 e e e e
T .
Lt
5 ‘ .
0.000M — 1 ; —
0 20 40 o0 80 100 120 140
TIME AFTER MELTING (min)
Fig. 17.3. Composition of the Containment Mockup

Facility Tank Atmosphere as Determined by Gas Samples.

139

Measurements were made of deposition rates
of seven fission products on four different surfaces
(stainless steel, carbon steel, and
vinyl paint) during the experiment. Jodine showed
a significant variation in rate of deposition on
the different surfaces, but the others, with the
possible exception of a faster rate for molybdenum
on carbon steel and vinyl paint, showed no signif-
icant variations. (

aluminum,

THE CONTAINMENT RESEARCH INSTALLATION

G. W. Parker W. J. Martin

A schematic diagram of the Containment Research
Instaliation (CRI) and a brief description of its
principal features were given in the previous
progress report.” All the major components of
the CRI have been received, and construction
is approaching completion. The primary simulation
vessel, equipped with external resistance heaters
(21-kw total capacity), has been mounted in the
hot cell, and the major piping between the furnace
and the containment vessel has been installed
(see Fig. 17.4).

Installation of the containment vessel has been
delayed by a warpage problem (now solved) that
temporarily prevented insertion of one of the re-
movable liners. The liners will be painted with
either an epoxy coating that is of interest to the
LOFT facility at the NRTS or with a Bakelite
coating of the type being considered for use in the
Containment Systems Ezxperiment at Hanford.

Stainless-steel-clad UO_ fuel specimens have
been successfully melted in the Containment Mockup
Facility using the pressurized induction furnace
arrangement that will be used in the CRI. Modifica-
tions of the furnace fittings to allow remote loading
and unloading of highly irradiated fuel are presently
being worked out. Wiring and instrumentation of
the CRI, involving instrument panels on two floor
levels, are about 75% complete.

Preliminary testing of the CRI is expected to
begin early in 1966. '

 

°G. W. Parker et al., Reactar Chem. Div. Ann. Progr,
Rept. Jan. 31, 1965, ORNL-3789, pp. 259-61.
140

PHOTO 81557

. e

—»‘n""lf?‘f"" R
o
LD

 

Fig. 17.4. Primary Simulation Vessel of Containment Research Installation Mounted in Hot Cell.
ANALYSIS OF PLANS FOR SCALE.UP
IN NUCLEAR SAFETY PROGRAM

C. E. Miller, Jr. W. E. Browning, Jr.
The Nuclear Safety Program of the United States
includes firm plans for large-scale experiments —
of which the LOFT!? is the most ambitious — in
which a reactor core would be destroyed and fis-
sion product release and behavior would be meas-
ured. Such experiments will be exceedingly complex
and costly and will, obviously, be few in number.
It will be essential, therefore, to gain from them
the maximum information possible and to be able —
with the help of well-designed smaller-scale
studies — to apply this information to possible
accidents under conditions other than those tested.
At the request of AEC Washington we have attempted
an evaluation of the research and development
program at ORNL and elsewhere and of the varicus
scaled-up experiments for which planning is firm.

11,12 are

As a result of this evaluation, two reports
in the final stages of preparation.

We conclude that experiments to study fission
product behavior in containment and in gas-cleaning
systems are well scaled over the range between
small laboratory experiments and LOFT, but that
experiments to study the release and the transport
of fission products from fuels are not well scaled,
We suggest that intermediate-scale experiments of
the otrder of 1 and 10% of the size of LOFT are

 

1OT. R. Wilson et al., An Engineering Test Program

to Investigaie a Loss-of-Coolant Accident, ID0O-17049
(October 1964),

e g, Mitler, Jr., and W. E. Browning, Jr., A4n
Analysis of the Adequacy of Planned Experiments fo
Test fthe Effects of Scale-Up in Reactor Accidents,
Part [: Scale-Up in the U.5. Nuclear Safety Program,
ORNL-3501, to be published.

Vi, wm Miller, Jr., and W. E. Browning, Jr., 4n
Analysis of the Adeguacy of Planned Experiments. to
Test the Effects of Scale-Up in Reactor Accidents,
Part I: Recommended Additional Nuclear Safety Scale-
Up Experiments, in preparation.

141

needed. These intermediate experiments should
measure fission product release from the fuel and
study fission product behavior in the primary
system following release.

We would propose for consideration four experi-
ments to bridge the pap between those described
in the previous sections of this chapter and the
LOFT assembly. These are:

A. Tests at 1% of LOFT power level:
1.

Investigation of fission product release
and transport in the Containment Research
Installation'?® primary simulator using a
dummy core with irradiated fuel elements.
These would be melted in place by central-

tungsten-resistor methods.

Investigation of fission product release and
transport in an experimental research re-
actor using a multiple-pin subassembly with
irradiated fuel pins. These would be melted
in place using nuclear heating induced by
the parent reactor.

B. Tests at 10% of LOFT power level:

l. Investigation of fission product release and
transport in the Containment Systems Exper-
iment'* primary simulator using a dummy
fuel core with irradiated fuel elements.
These would be melted in place by cedtral-

tungsten-resistor methods.

 

2. Investigation of fission product release
and transport using a fuel subassembly of
irradiated fuel elements in the LOFT f[acil-
ity. These would be melted by nuclear
self-heating induced by neutrons from the
LOFT parent core.

HG. W. Parker and W. J. Martin, Nucl. Safety Program
Semiann. FProgr. Rept. June 30, 1965, ORNL-3843,
po, G297,

14

G. J. Rogers, Program for Containment Systems Ex-
periment, HW-83607 (September 1964).
18. Laboratory-Scale Supporting Studies

RETENTION OF RADIOIODINE AND METHYL
I0DIDE BY ACTIVATED CARBONS

W. E. Browning, Jr. F. V. Hensle
R. D. Ackley R. E. Adams
J. E. Attrill? J. D. Dake?
G. W. Parker D. C. Evans?
G. E. Creek A. Ferreli®

Its relatively high volatility and its great bio-
logical hazard combine to make radioiodine a most
important fission product in nuclear safety con-
Elemental I, is relatively easy to
trap and hold; the great convenience and safety of
ambient-temperature activated-charcoal beds have
led to their general adoption for removing iodine
from air or gas streams. An increasing awareness,
however, that a significant fraction of the iodine
liberated in a nuclear accident might be present as
organic iodides (notably methyl iodide) which are
poorly retained by common activated charcoals
has caused concern in the field for several years.

Careful experiments have served only to con-
firm that methyl iodide is poorly retained by
standard charcoals under high humidity, Removal
of methyl iodide by Pittsburgh type PCB activated
charcoal, 12 x 30 mesh (unimpregnated for CH,I
reaction), was investigated at 24°C; gas chroma-
tography was used to determine the CHBI that
passed the bed. The inlet methyl iodide concen-
tration was varied (inversely with air velocity)
from 8 to 80 mg of CH,I per cubic meter of air to
provide a constant introduction rate, Time periods
over which useful removal efficiencies were ob-

gsiderations.

 

1Analytical Chemistry Division.
2Co-op student, University of Tennessee,

3Noncitizen guest on assignment from the Italian
National Committee for Nuclear Energy.

142

served varied from less than 0.2 hr for a relative
humidity near 100%, an air velocity of 100 fpm,
and a bed depth of 1 in., to about 100 hr for a
relative humidity of less than 3%, an air velocity of
10 fpm, and a bed depth of 3 in.**> Removal of
methyl iodide and of elemental iodine by a variety
of charcoals was investigated with the airecharcoal
system at around 100°C and with the air stream
about 80% saturated with steam. In these tests
the iodine which passed the bed was determined
by radioactivity measurements. Elemental iodine
was trapped with fairly high efficiency, but methyl
iodide was not. Both radiation detection and gas
chromatography were used to test a method based
on dehumidification for improving the efficiency
of methyl iodide removal from steam-air mixiures.
This method was observed to be successful but
would require a rather complex moisture-removal
sys;t*em.6

The recent obsetvation’ "2 that beds of a com-
mercial charcoal (Mine Safety Appliances Company
No. 85851) which is impregnated with several
percent by weight of iodine will effectively remove
methyl iodide is, therefore,
advance. Subsequent tests have shown that other
charcoals when impregnated with I, or KI are
equally effective, We confirmed the observations

a most important

‘w. E. Browning, Jr., et al., Nucl. Safefy Program
Semiann. Progr. Rept. June 30, 1965, ORNIL-3843, pp.
139—-43,

*R. E. Adams et al., Nucl., Safety Program Semiann.
Progr. Rept, Dec. 31, 1965, to be issued.

6W. E. Browning, Jr., et al., Nucl. Safety Program

Semiann. Progr., Rept., June 30, 1965, ORNIL.-3843, pp.
143—-48,

7G. W. Patker and Alberto Ferreli, Nucl. Safety
Program Semiann. Progr. Repft, June 30, 1965, ORNL-
3843, pp. 194-210,

8G. W. Parker et al., Nucl. Safefy Program Semiann,
Progr. Rept. Dec. 31, 1965, to be issued.

IR, E. Adams et al., Nucl., Safety Program Semianit.
Progr. Rept. Dec. 31, 1965, to be issued.
ORNL~DWG 65 -11435RZ

 

00

10 : —— L

O

 

Yo PENETRATION

0.0

Q.00

 

 

  
 
 
 
 
 
 
 

2E72S UNTREATED

25725 + 1, (10 mg/g)

25725 + £, 0. (50 mg/q)

PCB -+ 1, (70 mg/g)

25725 + TE.D. (850 mg/q}+KI (10 mg/y)
BCB + KI (10 mgsy)

PCB-FKI WO mg/g)t 12(10 g /)

2D = TRIETHYLENEDIAMING

 

 

0,000
0 G 02 Q.3

RESIDENCE TIME (sec)

Fig. 18.1.
at 25°C and 70% Relative Humidity with o 20<hr Loading.

reported by British investigators'® that activated
carbon impregnated with triethylenediamine ' is
also a good trapping material for methyl iodide.
Tests conducted with impregnated charcoal under
two different sets of conditions postulated for the
HFIR and LOFT reactor installations gave equally
good retention values; changing the loading level
(weight of methyl iodide per gram of charcoal) over
a wide range had no significant effect on retention
efficiency. A comparison of the methyl iodide
retention (actually the '3'I that was originally
combined as 'CH3 1317 of untreated and impregnated
activated-carbon materials is shown in Fig. 18.1.

The usefulness of the commercially available
impregnated activated charcoal for radicactive

 

IOR,. D. Collins, letter to the edifor, Nucleonics
23(9), 7 (1965). See also D, A, Collins, R. Taylor,
and L. R. Taylor, CONF-650407, p. 853 (September
1965).

  

Effect of Residence Time on Retention of Methy! lodide by Treated and Untreated Coconut Charcoal

methyl iodide trapping was confirmed at 24°C
and for relative humidities up to 71%, and the
mechanism was shown to be due to the exchange
reaction:

CH, Y114 12 7i(on charcoal) » CH, 1?71
+ *¥11(on charcoal) .

The charcoal, as received, contains a form of
iodine, and nonradioactive CH,I is observed emerg-
ing from the charcoal while it is trapping radio-
active CH,I. This charcoal was also found to
trap elemental iodine satisfactorily under the above
conditions.?  Further tests under more severe
conditions are planned.

The investigation of the possibility of impreg-
nating charcoal with chemicals that will react
with methyl iodide to form products retained within
the charcoal bed was done mostly by gas chroma-
tography. Tests were made at 24 to 25°C with
inorganic impregnants for relative humidities around
70% and with organic impregnants for relative
humidities ranging up to 100%. As previously
observed,!? piperidine and piperazine were found
to enhance considerably the retention of methyl
iodide by charcoal so impregnated. Results ob-
tained with several materials suggested by thermo-
chemical calculations as impregnants, such as
bromine, sodium bromate,
were rather encouraging,.

o

and sodium bromide,
Further work is under
way to evaluate these observations,!?

IDENTITY OF VAPOR FORMS OF RADIOIODINE

J. E. Attrill?
J. D. Dake?
D. 1. Ford'?

W. E. Browning, Jr.
R. E. Adams
R. D. Ackley

Gas chromatography has been applied to and
found to be very useful in the analysis of radio-
iodine-vapor samples. By using electron-capture
detectors, which are sensitive to organic halogen-
containing molecules, methyl iodide concentrations
of the order of 1 ppb can be measured. In the anal-
ysis of vapor from radioactive-iodine sources, a
number of unidentified peaks were often observed.
Since some of these peaks may have been due to
small amounts of extraneous impurities, such as
chloroform and fluorocarbons, and since there is con-
siderable variation in sensitivity between dif-
ferent compounds of iodine, there is need, on
for detection that is more specific,
Consequently, a radiation-detector arrangement was

occasion,

constructed and installed so that it was in series
with the electron-capture detector with respect to
gas flow, In this manner, attention was focused
only on those compounds containing '3*'I, and
any interference by extraneous nonradioactive
organic substances present in the sample was
eliminated.

The gas chromatograph, equipped with dual
detection based on radiation and election capture,
was tested with samples of the vapor over an
iodine source that was prepared by the palladium

 

g, E. Adams et al., Nucl, Safety Program Semiann.
Progr. Rept, Dec., 31, 1965, to be issued.

12']?l.ernpv:)rary summer employee, Rice University.

144

iodide method, The vapor samples showed ten
peaks by electron caplure and eight peaks by
radioactivity., = Three of the latter peaks were
identified as methyl, ethyl, and propyl iodides
with methyl iodide accounting for 62% of the
total radioactivity., As time permits, plans are
to accumulate a compilation of relative retention
times for a wide variety of iodine compounds.
This chromatograph was used to determine the
purity of radioactive methyl iodide sources used
in related studies. These sources, as far as their
radioactivity was concerned,
pure methyl iodide.

were essentially
Samples of iodine-containing
gases from other nuclear safety experiments, such
as the in-pile meltdown experiments and the NSPP,
are to be studied. Additional details of this work
appear elsewhere, 7+

DEVELOPMENT OF METHODS FOR
DISTINGUISHING AND MEASURING
GAS-BORNE FORMS OF FISSION

PRODUCTS
W. E. Browning, Jr. J. Truitt
R. E. Adams W. H. Hinds
R. D. Ackley A. F. Roemer, Jr.!
R. L. Bennett B. A. Cameron®?
M. D. Silverman J. D. Dake?
D. I. Ford'?
Characterization devices such as composite
diffusion tubes, fibrous-filter analyzeis, gas

chromatographs, and May packs are needed to
measure various forms of gas-borne fission products
in order to study and understand their behavior
under reactor accident conditions., High humidity
reduces the effectiveness of activated charcoal
for removing various species of radioiodine and
consequently interferes with the performances of
May packs and composite diffusion tubes which
contain charcoal components. Several techniques
have been investigated in an effort to reduce the
moisture effect in diffusion tubes. Heating the
tube to reduce the relative humidity and using

 

Dy, B, Browning, Jr., R. E. Adams, and J. E. Attrill,
Nucl. Safety Program Semiann. Progr, Rept. June 30,
1965, ORNL-3843, pp., 17173,

190 E. Adams et al., Nucl, Safety Program Semiann.
Progr. Rept. Dec. 31, 1965, to be issued.

15’I‘empr;)rary summer employee, Hanover College,
145

refrigerated condensers to remove moisture have
been only partially successful.’® Of several
desiccants which have been tested for drying the
air, Drierite (anhydrous calcium sulfate) appears
to be promising because of its minimal adsorption
of methyl iodide.!” An improved deposition pat=
tern for methyl iodide also has been obtained
under moist conditions by substitution of an todine-
impregnated the normal coconut-
charcoal lining of the diffusion tube.'”

Investigation of iodine behavior in May packs
under various conditions was continued. A test
of a volatile iodine sample in a May pack of the
type used in the Nuclear Safety Pilot Plant was
made with an ait-steam atmosphere at 100°C, 2 atm
absolute pressure, and 90% relative humidity.!®
In general, the behavior was similar to that of
composite diffusion tubes at the same conditions,
but before more precise comparisons can be made,
further experimentation will be required. A facility
for studying the behavior of normal and modified
May packs in moist atmospheres at elevated tem-
peratures is being constructed,

Gas chromatography with dual detection (electron
capture and radiation) has been successfully ap-
plied in the analysis of various radioactive~iodine
aund methyl iodide sources in the laboratory.'?

A comparison of the deposition of various iodine-
vapor species in a composite diffusion tube of a
“design for use in in-pile experiments -with the
deposition in the more usual type of composite
diffusion tube was made.?® Similarities and dif-
ferences were observed; however, with the aid of
this information a useful interpretation of the in-
pile data can be made. In conjunction with this
study, the reactivity of iodine vapor with Dacron
fiber was observed. Deposition on the fiber, which
is used in a fibrous-~filter characterization device
for aerosols, was greater than 90% for elemental
iodine and negligible for the nonelemental species.

charcoal for

 

15y, &, Browning, Jr., and R. B, Adams, Nucl. Safefy
Program Semiann. Progr. Rept. June 30, 1965, ORNL-
3843, pp., 223-24.

172, E. Adams et al., Nucl. Safety Program Semiann.
Progr. Rept, Dec. 31, 1965, to be issued.

1% D, Ackley et al., Nucl. Safety Program Semiann.
Progr. Rept. Dec. 31, 1965, to be issued.

R, E. Adams et al., Nucl. Safety Program Semiann.
Proge, Rept. Dec. 31, 1965, io be issued,

20w, E. Browning, Jr., R. D. Ackley, and A. F. Roemer,
Jr., Nucl., Safety Program Semiann. Progr. Rept. June
30, 1965, ORNL~3843, pp. 224--34,

Application of the fibrous-filter analyzer, devel-
oped previously for study of filtration processes,”!
as an analytical sampler for particulate material
in moist atmospheres is being evaluated.?? Two
problems are currently under study: (1) the physi-
cal response of the analyzer to an aerosol under
conditions of high humidity and (2) the production
of a radicactive, monodisperse aerosol of known
size with which to calibrate the response of the
analyzer under various combinations of operating
parametets. At low sampling velocities of 3 to
11 fpm, the fibrous-filter analyzer shows promise
as an analytical sampler in a wet atmosphere. The
fiber configuration and filtration processes are
not significantly affected by the moisture., A
satisfactory monodisperse aerosol of known size
has been prepared from a Dow polystyrene latex
solution; however, efforts to tag it radiochemically
have not yet been successful,

REMOVAL OF PARTICULATE MATERIALS
FROM GASES UNDER ACCIDENT CONDITIONS

J. S. Gill
G. L. Kochanny??

W. E. Browning, Jr.
R. E. Adams

Laboratory investigations of the filtration of
particulate aerosols which simulate those expected
to be generated during reactor accidents are con-
tinuing. The technique for the generation and
characterization of the test aerosol has been re-
ported previously, ?*

Results of tests in a dry-air (3% relative
humidity) system indicate that the arc-generated
aerosol consists of {wo particulate size groups,
each of which has different tendencies to pene-
trate a commercial (Flanders 700) high~efficiency
filter medium.”® The less penetrating group is
trapped with an efficiency of >989.99% on the

 

21M. D, Silverman and W. E. Browning, Jr., Science
143, 572 (1964).

220, D. Silverman et al.,, Nwucl. Safety Program
Semiann. Progr. Repf. Dec. 31, 1965, to be issued,

B present address: Dow Chemical Company, Midland,
Mich,

4w, E. Browning, Jr., RE. E. Adams, and G. L.
Kochanny, Jr., Reactor Chem. Div. Ann. Proge. Rept.
Jan. 31, 1965, ORNL~-3789, p. 288,

QSW. E. Browning, Jr., et al., Nucl. Safety Program
Semiann. Progr. Rept. June 30, 1965, ORNL~3843,
pp. 148--56.
Flanders filter, while the more penettating group
is trapped at a slightly lower efficiency of >99,9%.
Current using aerosols generated
from UO, clad with either neutron-irradiated stain-
less steel or Zircaloy have produced results similar
to the earlier studies. Similar values were ob-
tained when air velocity ranged from 3.2 to 13 fpm
and when dry helium or argon was substituted for
the air atmosphere.

However, in a recent series of experiments26
in which the relative humidity of the air sweep
ranged from approximately 46 to 100%, the filter
efficiency was noted to drop to as low as 93%.
Fibrous-filter deposition profiles and photomicro-
graphs of an aerosol produced in a 46% relative
humidity experiment indicate that the filtration

experiments

characteristics of the particles of the aerosol are
changed. Future experiments will attempt to de-
termine the role which water vapor plays in this

observed reduction in filter efficiency,

IGNITION OF CHARCOAL ADSORBERS BY
FISSION PRODUCT DECAY HEAT

B. F. Roberts
R. P. Shields

W. E. Browning, Jr.
C. E. Miller, Jr.

In the event of a nuclear accident in which fise
sion products are released into the containment
system, the charcoal adsorbers used for removal of

 

26R. E, Adams, J. S. Gill, and W. E, Browning, Jr.,
Nucl, Safety Program Semiann. Progr. Rept. Dec. 31,
1965, to be issued,

146

iodine are subject to heating by decay of iodine
adsorbed on the charcoal. The heat generated in
pieces of charcoal containing the greatest amount
of adsorbed iodine may raise their temperature to
the ignition point. In order to simulate such an
event, an in-pile experiment that utilizes the f{uel-
melting facility in the Qak Ridge Research Re-
actor has been designed, In this experimeat the
decay of short-lived iodine adsorbed on the char-
coal generates heat. The objective of this ex-
periment is to measure and compare the ignition
temperature of the charcoal with and without fis-
sion product adsorption. A comparison of the cal-
culated decay-heat load on the front surface of the
charcoal in the experimental adsorber with those
of adsorbers at various reactors indicates that
the ignition effects in the reactor adsorbers will
be smaller than those in the experiment.

After the mechanical design and plans for operat-
ing the experiments were shown to be adequate
by tests in a laboratory mockup,?” an initial series
of in-pile experiments was performed.?® In these
experiments the ignition temperature was studied
as a function of both the air flow and the decay-
heat loading., Preliminary results of these studies
indicate that the charcoal ignition temperature is
not greatly affected by the decay heat of adsorbed
fission products,

 

Tw. E. Browning, Jr., et al., Nucl. Safety Program
Semiann. Progr. Repfi. June 30, 1965, ORNL-3843,
pp. 15667,

28W. E. Browning, Jr., et al., Nucl. Safety Program
Semiann., Progr. Rept. Dec. 31, 1965, to be issued.
AUTHOR(s)

Bacarella, A, L., and
A, L. Bution

Baes, C, F., Jr.

Baes, C, F,, Jr.,
N. J. Meyer, and
C. E. Roberts

Bamberger, C. E. L.,
H. ¥. McDuffie, and
C. ¥. Baes, ]Jr.

Barton, C. J., and
W. B. Cottrell

Browning, W. E., Jr.

Brunton, G, D.

Burns, J. H.

Burns, J. H., and
W. R, Busing

Cantor, 8,

Carroll, R. M.

Carroll, R, M.,

R. B. Perez, and

O. Sisman
Caveol]l, R, M,,

and P. E. Reagan
Carroll, R. M.,

and O. Sisman

Publications

JOURNAL ARTICLES

TITLE

Electrochemical Measurements on Zirconium
and Zircaloy-2 at Elevated Temperatures,
2. 200-300°C

The Reduction of Potentiometric Hydrolysis
Data

The Hydrolysis of Thorium(IV) at 0..95°

Purification of Beryllium by Acetylacetone-
EDTA Solvent Extraction — Procedure and
Chemistry

Fission Product Refease and Transport

Removal of Radiviodine from Gases

Crystal Structure of Zirconium Tetrafluoride

Crystal Structure of Hexagonal Sodium
Neodymium Fluoride and Related Compounds

Crystal Structures of Rubidium Lithium

Fluoride and Cesium Lithium Fluoride

The Vapor Pressures of Beryllium Fluoride
and Nickel Fluoride

Fisgion Product Release from Fuels
Release éf Fission Gas During Fissioning of

UC)2

Techuniques for In-Pile Measurements of

Fission Gas Release

In-Pile Fission-Gas Release from Single
Crystal UO:7

147

PUBLICATION

J. Electrochem. Soc. 112(6), 546
(1965)

Inorg. Chem. 4, 588 (19653)

Inorg. Chem. 4, 518 (1965)

Nucl. Sci. Eng. 22(1), 14 (1965)

Nucl. News 8{7), 25 (1965)

Nucl. Safety 6(3) 272 (1965)

J. Inorg. Nucl. Chem. 27, 1173
(1965)

Inorg. Chem. 4, 881 (1965)

Inorg. Chem. 4, 1510 (1965)

J. Chem. Eng. Data 10{3), 237
(1965)

Nucl. Safety 7(1), 34 (19653)

J. Am. Ceram. Soc. 48(2), 55
(1965)

Nucl. Sci. Eng. 21(2), 141
(1965)

Nucl. Sci. Eng. 21(2), 147
(1963)
AUTHOR(s)

Carroll, R. M.,

and O. Sisman

de Bruin, H. J., and
G. M. Watson

Fuller, E. L.,
H., F. Holmes, and
C. H. Secoy

Hoimes, H, F,,
C. S. Shoup, Jr.,
and C, H. Secoy

Johnson, G. L.,
M. J. Kelly, and
D, R, Cuneo

Keilholtz, G. W.,
J. E. Lee, Jr.,
R. E. Moore, and

R. L.. Hamner

Malinauskas, A. P,,
and F, L. Carlsen, Jr.

Marshall, W, L.,
and Ruth Slusher

Miller, C, E., Jr.,

and W, ¥, Hillsmeier

Osborne, M. F,,
E. L. Long, Jt.,
and J. G. Morgan

Overholser, L.. G.,
and J. P, Blakely

Parkinson, W, W., Jr.,
C. D. Bopp, D. Binder,
and J. E, White

Quist, A, S,, and
W, L., Marshall

Quist, A, S,,
W. L. Marshall,
and . R. Jolley

148

TITLE

In-Pile Fission-Gas Release from Fine
Grain UO2

Self-Diffusion of Beryllium in Unirradiated

Beryllium Ozxide

Gravimetric Adsorption Studies of Thorium

Dioxide Surfaces with a Vacuum Microbalance

FElectrokinetic Phenomena at the Thorium

Oxide---Aqueous Solution Interface

Reactions of Aqueous Thorium Nitrate

Solutions with Hydrogen Peroxide

Behavior of BeO Under Neution Irradiation

Gas Transport Characteristics of Uraniume-
Fueled Graphites

Aqueous Bystems at High Temperature, XV,
Solubility and Hydrolytic Instability of
Magnesium Sulfate in Sulfuric Acid-Water
and Deuterosulfuric Acid—Deuterium Oxide
Solutions, 200° to 350°C

Second AEC Conference on Radigactive Fallout

from Nuclear Weapons Tests

Performance of Prototype EGCR Fuel Under

Extreme Conditions

Oxidation of Graphite by Low Concentrations

of Water Vapor and Carbon Dioxide in Helium

A Comparison of Fast Neutron and Gamma
Irradiation of Polystyrene., Part I. Cross-

L.inking Rates

Assignment of Limiting Equivalent Conduct-

ances for Single Ions to 400°

Estimation of the Dielectric Constant of
Water to 800°

Electrical Conductances of Aqueous Solutions
at High Temperature and Pressure. [I, The
Conductances and lonization Constants of
Sulfuric Acid—Water Solutions from 0—800°C
and at Pressures Up to 4000 Bars

PUBLICATION

J. Nucl. Mater. 17(4), 305
{1965)

J. Nucl. Mater. 14, 239 (1964)

Vacuum Microbalance T'ech. 4, 109
{1965)

J. Phys. Chem. 69, 3148 (1965)

J. Inorg. Nucl. Chem. 27, 1787
(1965)

J. Nucl. Mater. 14, 87 (1964)

Am. Ceram. Soc. Bull. 44, 251
(1965)

J. Chem. Eng. Data 10, 353
(1965)

Nucl. Safety 6(3), 283 (1965)

Nucl. Sci. Eng. 22(4), 420 (1965)

Carbon 2, 385 (1965)

J. Phys. Chem. 69, 828 (19653)

J. Prys. Chem. 69, 2084 (1945)

O

J. Phys. Chem, 69, 3165 (1965)

J- Phys. Chem. 69, 2726 (1965)
AUTHOR(s)

Reagan, P. E.,
J. G. Morgan, and

Q. Sisman

Robbins, G. D.,
R. E. Thoma, and
H. Insley

Singh, A. J.,
R. G. Ross, and
R. E. Thoma

Thema, R. E.,
H. Insley,
H. A, Friedman, and
G. M. Hebert

Adams, R. E., and
W. E, Browning, Jr.

Adams, R. E.,
W, E. Browning, Jr.,
Wm, B, Cottrell, and
. W, Parker
Baumann, C, D,
Browning, W, E., Jr.,
and M. E, Davis
Brunton, G. D.,
H. Insley, T. N. McVay,
and R, E. Thoma
Clark, W. E., L. Rice,
and D. N, Hess

English, J. L., and
J. C. Griess

Haynes, V., O.,
. H, Sweeton, and
D, E, Tidwell

Jenks, G, H.

149

TITLE

Fission-Gas Release from Pyrolytic-Carbons
Coated Fuel Particles During Irradiation at
2000 to 2500°F

Phase Equilibria in the System CstZrF4

Vacuum Disiillation of LiF

The Condensed System LiF-NaF-ZrF4 — Phase
Equilibria and Crystallographic Data

REPORTS ISSUED

Iodine Vapor Adsorption Studies for the NS
“Savannah® Project

The Release and Adsorption of Methy! Iodide in
the HFIR Meaximum Credible Accident

{rradiation Effects in the EGCR Fuel

A Standard Surface for Fission Product

Deposition Experiments

Crystallographic Data for Some Metal

Fluorides, Chlorides, and Oxides

Evaluation of Hastelloy I and Other
Corrosion-Resistant Structural Materials
for a Continuvons Centrifuge in a Multi-

purpose Fuel-Recovery Plant

Laboratory Corrosion Studies for the High

Flux Isclope Reactor

Irradiation Deta for the Armmy PM Fuel
Experiment in the ORR Pressurized-Water
Loop for ORR Cycle 52

Irradiation Data for the Army PM Fuel
Experiment in the ORR Pressurized-Water
Loop for ORR Cycles 53 and 54

Effects of Reactor Operation on HFIR Coolant

PUBLICATION

Nucl. Sci. Eng. 23, 215 {1965)

J. Inorg. Nucl, Chem. 27, 559
(1965)

J. Appl. Phys. 36(4), 1367
(1965)

J. Chem. Eng. Data 10(3), 219
(1965)

ORNI.-3726 (February 1965)

ORNL=TM=1201 (October 19635)

ORNL.-3504 {June 1965)

ORNL-TM-1304 (October 1065)

ORNL~3761 (February 1065)

ORNL=3787 (April 1965)

ORNLTM-102%2 (June 1965)

ORNI~-TM-083 (I"ebruary 1963)

ORNL-TM-1034 (April 1965)

ORNI.-3848 (October 1965)
AUTHOR(s)

Jenks, G. H.,
E. G, Bohlmann,
and J. C. Griess

Keilholtz, G. W.,
and C. J. Barton

Mason, E. A,, and
A, P. Malinauskas

Mathews, A, L.,
and C, I, Baes

Rabin, 8, A.,
J. W, Ullmann,
E, L.. T.ong, Jr.,
M. F. Osborne, and
A, E. Goldman

Redman, J. D,

Reed, S. A., and
J. C. Movers

Savage, H. W,,
E. L. Compere,
W. R. Huntley,
R. E. MacPherson,
and A, Taboada

Singh, A. J., R. G.
Ross, and R, E. Thoma

Thoma, R, E,

Yeatts, L. B,, Jr.,
and W, T. Rainey, ]Jr.

150

TITLE

An Evaluation of the Chemical Problems
Associated with the Aqueous Systems in
the Tungsten Water Moderated Reactor,
Addenda 1 and 2

Behavior of Iodine in Reactor Containment

The Effect of Accommodation on Thermal
Transpiration: Limitations of the **Dusty=
Gas®® Model in the Description of Surface

Scattering

Oxide Chemistry and Thermodynamics of
Molten Lithium Fluoride--BeryIlium Fluoride
by Equilibration with Gaseous Water—

Hydrogen Fluoride Mixtures

Irradiation Behavior of High Burup ThOQ—
4.45% UO2 Fuel Rods

A Literature Review of Mass Spectrometric—

Thermochemical Technique Applicable to the

Analysis of Vapor Species over Solid

Inorganic Materials

Estimation of Annual Operating and

Maintenance Costs of Dual Putpose Nuclear-

Electric Sea Water Conversion Stations

SNAP=8 Corrosion Program Quarterly Progress
Report for Period Ending November 30, 1964

SNAP-8§ Corrosion Program Quarterly Progress
Repott for Period Ending February 28, 1965

SNAP-8 Corrosion Program Quarterly Progress
Report for Period Ending May 31, 1965

Zone Melting of Inorganic Fluorides

Rare-Farth Ralides

Purification of Zirconium Tetraflyoride

PUBLICATION

ORNL-TM-278 (March 1965)

ORNL-NSIC-4 (February 1965)

ORNL-3796 (April 1965)

ORNL-TM-1129 (May 1965)

ORNI.-3837 (October 1965)

ORNL-TM-089 (June 1965)

ORNL-TM-1057 (April 1965)

ORNI.-3784 (March 1965)

ORNI.-3823 (June 1965)

ORNI.-3859 (September 1965)

ORNL-3658 (July 1965)

ORNL-3804 (May 1965)

ORNL-TM=1292 (November 1965)
AUTHOR(s)

Adams, R. E., and
W. E. Browning, Jr.

Browning, W. E., Jr.,
R. E. Adams, K. D.
Ackley, M. E, Davis,
and J. E. Attrill

Carroll, R. M.,

and O. Sisman

Compere, E. L., and

J. E. Savolainen

Davis, M. E., W. E.
Browning, Jr., G. E.
Creek, G. W. Parker,
and L. F. Parsly

Joncich, M. J.,
and . F. Holmes

Kelly, M, J.

Miller, C. E., Jr.,
W. E. Browning, Jr.,
R. P. Shields, and
B. F. Roberts

Parker, G. W.,
R. A, Loerenz,
and J. G. Wilhelm

151

BOOKS AND PROCEEDINGS

TITLE

Removal of Iodine and Velatile Iodine

Compounds from Air Systems by Activated

Chearcoal

Identity, Character, and Chemical Behavior of

Vapor Forms of Radioiodine

Fission-Gas Release During Fissioning in
Uranium(IV) Oxide

The Chemistry of Hydrogen in Liquid Alkali

Metal Mixtures Useful as Nuclear Reactor

Coolants.

The Deposition of Fission-Product Iodine on

I. NaK-78

Structural Surfaces

Heat Effects at Electrodes (work performed at

University of Tennessee)

Removal of Rare Earth Fission Products from

Molten Salt Reactor Fuels by Distillation

In-Pile Fission Product Release Experiments -

Atmosphere Effects

Release of

Fuels During Transient Accidents Simulated

Fission Products from Reactor

in TREAT

PUBLICATION

Proc. Intern. Symp.
Fission Product Release and
Transport Under dccident
Conditions, Qak Ridge, Tenn.,
Apr. 5.-7, 1963, CONF-650407
(1965)

Proc. Intem, Symp. Fission
Product Release and
Transport Under Accident
Conditions, Qak Ridge, Tetn.,
Apr, 57, 1965, CONF-650407
(L1965)

Trans. Am. Nucl. Soc, 8(1), 22
(1965)

Trans. Am. Nucl. Soc. 8(1), 170
(1965)

Proc. Intern. Symp. Fission
Product Release and
Transport Under Accident
Condirions, Qak Ridge, Tenn.,
Apr, 5.-7, 1965, CONF-650407
(1965)

Proc. 1st Australian Conf,
Electrochem., Sydney and
Hobart, Ausiralia, February
1963, Pergamoen, New York,
1964

Trans. Am. Nucl. Soc. 8(1), 170
(1965)

Proc. Intern. Symp. Fission
Product Release and
Transport Under Accident
Conditions, Oak Ridge, Tenn.,
Apr. 5.7, 1965, CONF-650407
(1965)

Proc, Intem. Symp. Fission
Product Release and
Transport Under Accident
Conditions, Qak Ridge, Tenn.,
Apr. 5.7, 1965, CONF-650407
(19653)
AUTHOR(s})

Parker, G. W., W, J.
Martin, G. E. Creek,
and C, J. Barton

Perez, R. B.

Roberts, B, F.,
W. E, Browning, Jr.,
R. P. Shields, and
C. E. Miller, Jr.

Mathews, A. L.

152

TITLE

Behavior of Radioiedine in the Containment

Mockup Facility

A Dynamic Method for In-Pile Fission-Gas
Release Studies

Effects of Atmosphere on Behavior of Fission
Products Released by In-Pile Melting of UO2

THESIS

Oxide Chemisfry and Thermodynamics of
Molten Lithium Fluoride--Beryllium Fluoride
by Eqguilibration with Gaseous Water~Hydrogen

Fluoride Mixtures

PUBLICATION

Proc. Intern. Symp. Fission
Product Release and
Transport Under Accident
Conditions, OQak Ridge, Tenn.,
Apr. 5—-7, 1965, CONF-650407
{1963)

Trans. Am. Nucl. Soc. 8(1), 22
(1965)

Trans. Am. Nucl. Soc. 8(1), 139
(1965)

Thesis submitted in partial ful~
fillment of the requirements for
Ph.D. degree, University of
Mississippi, May 1965
Papers Presented at Scientific and Technical Meetings

AUTHOR(s)

Adams, R. E,, and
W, E, Browning, ]Jr.

Bacarella, A. L.

Baes, C, F., I

Baes, C, F., Jr.,
N. J. Meyer, and
C. E, Roberts
Bennett, R, L.,
H. L. Hewmphill, and
W. T. Rainey, Jr.

Bennett, R, L.,
H. L. Hemphill,

W. T. Rainey, Jr., and

G. W. Keilholtz

Blakely, J. P., and
.. G. Overholser

Blankenship, F. F.,
il. F. McDuffie,
R®. E. Thoma, and
W. R. Grimes

Bohimann, E. G.,
and F. A, Posey

TITLE

Removal of lodine and Volatile lodine
Compounds from Air Systems by Activated
Charcoal

Anodic Film Growth on Zirconium at Tempers
atures from 200--300°C

The Chewmistry and Thermodynamics of Molten
Salt Reactor Fluoride Solutions

Acidity Measurements at Elevated Temper=
atures. 2. Thorium(IV) Hydrolysis, 0—95°C

Thermal EMF Drift of Refractory Metal
Thermocouples in Pure and Slightly

Contaminated Helium Atmospheres

Stdbility of Thermoelectric Materials in a

Helium-~Graphite Environment

Oxidation of AT]J Graphite by Low Concentra-
tions of Water Vapor and Carbon Dioxide in
Helium

Molten Fluorides as Nuclear Reactor Fuels

Aluminum and Titanium Corresion in Saline
Waters at Elevated Temperatures

153

PLACE PRESENTED

International Symposiam on
Fission Product Release ard
Transport Under Accident
Conditions, Dak Ridpe, Tenn.,
Apr, 5~7, 1965

Electrochemical Society Meeting,
Buffalo, N.Y., Oct. 1114, 1965

TAEA Symposium on Thermo=
dynamics with Emphasis on
Nuclear Materials and Atomic
Transport in Solids, Vienna,
Austria, July 2227, 1965

American Chemical Society,
Detroit, Mich., Apr, 49, 1965

AEC High-Temperature Thermome
etry Meeting, Washington, D.C,,
Feb. 24-26, 1965

AEC High-Temperature Thermom-
etry Meeting, Washington, D.C.,
Feb. 2426, 1965

Conference on Carbon, 7th
Biennial, Cleveland, June 21-25,
1965

Electrochemical Society, San
Francisco, Calif,, May 9--13,
1965

1st International Symposium on
Water Desalination, Washington,
D.C., Oct. 3—9, 1965
AUTHOR(s)

Browning, W. E., Jr.,
R. E. Adams, R. D,
Ackley, M. E. Davis,
and J. E. Attrill

Buras, J. H.

Buras, J. H., and
E. K. Gordon

Carroll, R. M.

Carroll, K. M., and
PP. E. Reagan

Carroll, R. M., and

O, Sisman

Compere, E. L., and

J. E. Savolainen

Dabbs, J. W. T.,
F. J. Walter, and
G. W. Parker

Davis, M. E,, W, E,
Browning, Jr., G. E.
Creek, G. W, Parker,
and L. F. Parsly

Davis, R. J.

Grimes, W. R.

154

TITLE

Identity, Character, and Chemical Behavior of

Vapor Forms of Radioiodine

Crystal Structure Analysis Applied to

Inorganic Fluorides

Refinement of the Structure of Lithium

Tetrafluoroberyllate

The Behavior of Fission-Gas in Fuels

In-Pile Performance of High-Temperature

Thermocouples

Fission-Gas Release During Fissioning in
Uranium(IV) Oxide

The Chemistry of Hydrogen in Liquid Alkali
Metal Mixtures Useful as Nuclear Coolants.
1. NaK-78

Saddle Point Rotational States from

Resonance Fission of Oriented Nuclei

The Deposition of Fission-Product Iodine

on Structural Surfaces

Oxide Growth and Capacitance on Preirradiated

Zircaloy-2

Chemistry of Molten Fluoride Reactor

Systems

Molten Fluorides as Reactor Material

The Chemistry of the Molten Salt Reactor

PLACE PRESENTED

International Symposium on
Fission Product Release and
Transport Under Accident
Conditions, Qak Ridge, Tenn.,
Apr. 5-7, 1965

ORINS Traveling Lecture
Program, University of
Arkansas, Fayetteville, Dec.
6, 1965

American Crystallographic
Association, Gatlinburg, Tenn.,
June 27—July 2, 1965

AIME Conference on Radiation
Effects, Ashville, N.C,, Sept.
8-10, 1965

AEC High-Temperature
Thermometry Meeting, Washing-
ton, D.C., Feb. 24--26, 1965

American Nuclear Society,
Gatlinburg, Tenn., June 2124,
1965

American Nuclear Society,
Gatlinburg, Tenn. June 21-24,
1945

IAEA Symposium on the Physics
and Chemistry of Fission,
Salzburg, Austria, March 2226,
1965

International Symposium on
Fission Product Release and
Transport Under Accident
Conditions, Oak Ridge, Tenn.,
Apr. 5-7, 1965

Electrochemical Society, San
Francisco, Calif,, May 6-—13,
1965

Gordon Research Conference on
Molten Salts, Meriden, N.H.,
Aug. 30—Sept. 3, 1965

Chemistry Seminar at Argonne
National Laboratory, Nov, 12,
1965

Nuclear Engineering Colloquium
at Brookhaven National Lab-
oratory, Dec. 2, 1965
AUTHOR(s)

Grimes, W. R,

Karraker, R. H.,
and R. E. Thoma

Kelly, M. J.

Keyser, R. M.

Marshall, W. L.

Marshall, W. L.,, and
E. V. jones

Mathews, A. L.,
and C, F, Baes, Jr.

McDuffie, H, F., and
R. P. Atkinson

Miller, C. E., Jr.,
W. E. Browning, Jr.,
R. P. Shields, and
3. F. Roberts

Parker, G. W,
R. A. Lorenz, and
J. G. Wilhelm

Parker, G. W.,
W. J. Martin,
G. E. Creek, and
C. J. Barton

Parkinson, W. W.

155

TITLE

Molten Fluorides as Fuels, Coolants, and

Blankets for Nuclear Reactors

Sodium Fluoride—Scandium Fluoride Phase
Equilibria

Removal of Rare Earth Fission Products from
Molten Salt Reactor Fuels by Distillation

Experiments in Chemonuclear Synthesis at
ORNI.

Solubility and HKlectrical Conductance Meas-
urements of Some Aqueous Electrolytes of

Interest to Oceanography

Second Dissociation Constant of Sulfuric Acid
from Solubilities of Calcium Sulfate in Sule

furic Acid Solutions

Oxide Chemistry and Thermodynamics of

Molten Lithium Fluoride~Beryilium Fluoride

The Development of an Information Center

for Molten Salt Chemistry

In-Pile Fission Product Release Experiments —

Atmosphere Effecis

Release of Fission Products from Reactor
Fuels During Transient Accidents Simulated
in TREAT

Behavior of Radiociodine in the Containment

Mockup Facility

The Effect of Radiation on Plastics and
Rubber

PLACE PRESENTED

Chemistry Seminar at Purdue
University, Lafayette, Ind.,
Dec. 9, 1965

American Chemical Society, SW-SK
Regional Meeting, Memphis,
Tenn., Dec. 2—4, 1965

American Nuclear Society,
Gatlinburg, Tenn., June 21-24,
1965

Chemenuclear Review Meeting at
Onchicta Conference Center,
Tuxedo, N.Y., Oct, 19--20, 1963

Conference on Kinetics and
Mechanism in Aqueous Inorganic
Systems, Miami, Fla., Nov,
18--23, 1965

American Chemical Society,
Detroit, Mich., Apr. 4--9, 1965

American Chemical Society,
SW-SE Regional Meeting,
Memphis, Tenn., Dec. 2~4, 1965

Tennessce Academy of Science,
Qak Ridge, Tenn., Dec. 10--11,
1965

International Symposium on
Fission Product Release and
Transport Under Accident
Conditions, Qak Ridge, Tenn.,
Apr. 57, 1965

International Symposium on
Fission Product Release and
Transport Under Accident
Conditions, Oak Ridge, Tenn.,
Apr, 5--7, 1965

International Symposinm on
Fission Product Release and
Transport Under Accident
Conditiofis, QOak Ridge, Tenn.,
Apr. 5--7, 1965

U.5. Army Nuclear Science

Seminar, Oak Ridge, Tenn.,
July 27, 1965
AUTHOR(s)

Parkinson, W, W,,
and W. D. Burch

Perez, R. B., and
G. M. Watson

Quist, A. S.

Quist, A, S., and
W. L. Marshall

Reed, S. A.

Roberts, B. F.,
wW. E, Browning, ]Jr.,
R. P. Shields, and
C. E. Miller, Jr.

Rutherford, J. L.,
J. P. Blakely, and
I.. G, Overholser

Secoy, C. H,, H. F.
Holmes, C. S. Shoup,
and E. L., Fuller

Soldano, B, A.

Thoma, R. E,

156

TITLE

Synthesis of Organics Through Isotopic

Decay Radiation

A Dynamic Method for In-Pile Fission=-Gas

Release Studies

Electrical Conductances of Aqueous
Potassium Bisulfate and Sulfuric Acid
Solutions to 800°C and 4000 Bars

Electrical Conductances of Aqueous Potas=-
sium Bisulfate and Sulfuric Acid Solutions
to 800°C and 4000 Bars

Some Chemical and Materials Problems

in Sea Water Conversion Plants

Effects of Atmosphere on Behavior of
Fission Products Released by In-Pile
Melting of UO2

Oxidation of Fueled and Unfueled Graphite
Spheres by Steam

The Interaction of Water with the Surface of

Ceramic Oxides

The Osmotic Behavior of Agueous Solutions

at Elevated Temperatures

Methods of Investigating High Temperature
Phase Equilibria

Rare Earth Trifluoride Systems

PLACE PRESENTED

Annual Contractors Conference,
AEC Division of Isotopes
Development, Brookhaven
Mational Laboratory, Oct. 78,
1965

American Nuclear Society,
Gatlinburg, T'enn., June 21-24,
1965

Karlsruhe Technische
Hochschule, Karlsruhe, Germany,
Sept. 13—15, 1965

Central Electricity Research
l.aboratories, L.eatherhead,
England, Sept. 21, 1965

Risgd Research Establishment,
Risg (Roskilde), Denmark,
Sept. 27, 1965

CITCE Symposium on Electro=
chemistry at High Temperatures,
Budapest, Hungary, Sept. 5~10,
1965

American Chemical Society
Regional Meeting, Dayton,
Chio, Dec. 14, 1965

American Nuclear Society,
Gatlinburg, Tenn., June 21-24,
1965

Conference on Carbon, 7th
Biennial, Cleveland, June
21-25, 1965

International Conference on
Tropical Oceanography, Miami
Beach, Fla., Nov. 17—24, 1965

ORINS Traveling L.ecture
Program, Furman University,
Greenville, S,C,, Nov. 1, 1965

ORINS Traveling Lecture
Program, Furman University,
Greenville, 5.C,; Mar, 1, 1965

Chemistry Seminar at Arpgonne
National Laboratory, May 21,
1965
AUTHOR(s)

Truitt, Jack,

J. ©. Stiegler, and

R. B. Evans IIi
Vanderzee, C. E.,l and

E. L. Fuller, Jr.

Watson, G. M.

Yeatts, L. B., Jr.,
and W, L. Marshall

157

TITLE

Actinide Migration in Pyrocarbons as
Influenced by Actinide and Defect

Concentrations

Thermochemistry and Kinetics of the
Hydrolytic Decomposition of Ammonium

Carbamate and Ammonium Dithiocarbamate

The Economics of Nuclear Power

Chemical Aspects of Nuclear Safety

Physico-Chemical Problems of Nuclear Reactors

Solubilities of Calcium Hydroxide and
Saturation Behavior of Calcium Hydroxide —
Calcium Carbonate Mixtures in Aqueous

Sodium Nitrate Seolutions at High Temperatures

 

IResearch performed at University of Nebraska, Lincoln.

PLACE PRESENTED

American Ceramic Society,
Philadelphia, May 2--5, 1963

First Midwest Regional Meeting,
American Chemical Society,
Kansas City, Mo., Nov. 4-5,
1965

AEC International Exhibit at
San Salvador, El Salvador,
Mar, 1922, 1665

AEC International Exhibit at
San Salvador, El Salvador,

Mar. 19..22, 10965

ORINS Traveling Lecture
Program, Tarleton State
College, Stephenville, Tex,,
Deec. 6, 1965

American Chemical Society,
SW-SE Regional Meeting,
Memphis, Tenn,, Dec. 2—4, 1965
T

9-.58.
59,
60.
61.
62.

63-77.
78.
79.
80.
61,
82,
83.
84.
85.
86.
87.
88.
89,
90.
91.
92.
93.
94.
Q5.
96.
97.
98.
99,

100.
101.
102.
103.
104.
105.
106.
107.
108.
109.

159

INTERNAL DISTRIBUTION

Bioclogy Library _
Central Research Library
Reactor Division Library
Laberatory Shift Supervisor
ORNL Y-12 Technical Library
Document Reference Section
Laboratory Records Department

L.aboratory Records, ORNL R.C.

. E. Larson
M. Weinberg
G. MacPherson
E. Boyd

R. Bruce

L
H

. Culler

. Jordan

. Briggs

. Cottrell

. Fraas

. Krous

. Lane

Milier

. Trauger

. Whitman

. Beall

Billingten

. Ferguson

Frye, Jr.

. Kelley

. Taylor

. Bredig

. Corbin

. Cunningham

. Crawford

. King

. Lyon

. Parker
Skinner

. White

. Williams

PEETAAME-TFEZMEI-C00YVQO0PERPEAX>ZIPOF>ETTTOT>O0O
ITHAPIAIMOMOTS->>0EEC

NnO-wmzZx

110.
111
112,
113.
114.
115.
116.
117.
118.
119.
120,
121.
122.
123,
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143,
144,
145.
146.
147.
148.
149,
150,
151.
152.
153.

X PCUIARDYPEOrEBONEPALEOOICMALIMAIACOEOOPO0MROTME

OUIXA>PIPLPETEQOMMI US>V X

OITOmMmMUOS~rrmmImIT e A

TTMOFBr-r<

ORNL-3913
UC.4 — Chemistry
TID-4500 (47th ed.)

. Grimes

. Bohlmann

. McDuffie

. Watson

. Blankenship

. Secoy

. Baes, Jr.

. Bacarella
Barton

. Bopp

. Browning, Jr.

. Brunton

. Burns

antor

. Carroll

. Compere
Davis

. English

. Evans 1]

. Fuller, Jr,

. Griess

. Holmes

. Jenks

. Keilholtz
Kelly
Kirslis

. L.orenz

. Malinauskas

. Marshall

. Miller

. Moore

. Morgan

. Overholser

. Parker

. Parkinson
Quist

. Reed

. Richardson

. Romberger

. Savage

. Sears

. Shaffer
Shor

., Silvermon
160

154, O, Sisman 177. H. Insley (consultant)
155. B. A. Soldano 178. H. R. Jolley (consultant)
156. R. A. Strehlow 179. E. V. Jones (consultant)
157. F. H. Sweeton 180. T. N. McVay (consultant)
158. R. E. Thoma 181. G. Mamantov (consultant)
159-166. G. C. Warlick 182. J. L. Margrave {consultant)
167. C. F. Weaver 183. E. A. Mason (consultant)
168. L. B. Yeatts 184. R. F. Newton (consultant)
169. L. Brewer (consultant) 185. R. B. Perez (consultant)
170. J. W. Cobble (consultant) 186. J. E. Ricci (consultant)
171. F. Daniels (consultant) 187. Howard Reiss (consultant)
172. R. W. Dayton (consultant) 188. G. Scatchard {consultant)
173, P. H. Emmett (consultant) 189. D. A. Shirley {consultant)
174. H. S. Frank (consultant) 190. H. Steinfink (consultant)
175. N. Hackerman (consultant) 191. R. C. Vogel (consultant)
176. D. G. Hill (consultant) 192. T. F. Young (consultant)
EXTERNAL DISTRIBUTION

193. D. F. Bunch, Health Physics Branch, AEC, Washington

194. Paul E. Field, Dept. of Chemistry, Virginia Polytechnic Institute

195. L. R. Zumwalt, General Atomic Division, General Dynamics Corp., San Diego, Calif.

196. Research and Development Division, AEC, ORO

197. Reactor Division, AEC, ORO

198. Assistant General Manager for Research and Development, AEC, Washington

199. Division of Research, AEC, Washington

200. Division of |sotopes Development, AEC, Washington

201. Assistant General Manager for Reactors, AEC, Washington

202. Division of Reactor Development and Technology, AEC, Washington

203. Space Nuclear Propulsion Office, AEC, Washington

204. J. A. Swartout, 270 Park Ave., New York 17, New York

205-533. Given distribution as shown in TID-4500 (47th Ed.) under Chemistry category

(75 copies -- CFSTI)