ORNL-3594.txt

 

 

£S5

I

 

 

 

 

LOCKHEED MARTIN ENERGY RESEA
l |

[T

 

 

 

 

 

 

 

 

4 445k OHaA?L?L b

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 
ORNL-35%94

Contract No. W~7405-eng-26

Reactor Chemistry Division

MOLTEN~SALT SOLVENTS FOR FLUORIDE VOLATILITY PROCESSING
OF ALUMINUM~-MATRIX NUCLEAR FUEL ELEMENTS

R. E. Thoma
B. J. Sturm
E. H. Guinn

AUGUST 1964

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPCRATION
for the
U.S5. ATOMIC ENERGY COMMISSION
1ii

CONTENTS

-A?D st mc‘t L . & " e ” » - . . - - . - ®

Introduction .« « « o ¢ o ¢ o o o o o o

Choice of Constituent Fluorides as AlFi3 Solvents

Survey of Potential Solvent Systems
Procedures . « & & » ¢ o o o ¢ s 5 o
Materials e s & 6 s e s s e o o s e o
Results and Discussion . ¢« « « o « o &

AlF3 Melting Point . . « ¢« « .« + .«

Systems Based on LiF-KF . . . . . .

The System LiF-NaF~AlF3 « ¢« o o o

The System NaF-KF-AlF3 + o o « « &

The System KF-ZxrFs~=AlF3 . « o « o &
Conclusions e s s s e s e e & s e v .

References « o o o o o s o o o o o o &

Page

oW M

18]

O 0 =3

10

11
12
34
MOLTEN-SALT SOLVENTS FOR FLUORIDE VOLATILITY PROCESSING
OF ALUMINUM~-MATRIYX NUCLEAR FUEL ELEMENTS

R. E. Thoma, B. J. Sturm, E. H. Guinn

ABSTRACT

The results of a search for molten-salt solvents for use
in fluoride wvolatility processing of aluminum-matrixz fuel ele-
ments are presented. The solubility of aluminum fluoride in
various mixtures of fluorides was determined in order to esti-
mate the feasibility and cost of processing methods. OSufficient
data were accumulated to construct equilibrium phase diagrams
of the solution systems, IiF-NaF-AlFj3, LiF-KF-AlF3, LiF~K3AlFg-
MFo {(where MF, is CaFp, SrFp, or Zans, and XKF-ZrFs~Al1F3. New
and revised phase diagrams were determined for the limiting
binary systems of the alkali fluorides with AlFa by use of a
new visual method for determining the occurrence of liquidus
transitions. This method provided several advantages not
available in classical methods of obtaining liquidus data.

For example, it was observed for the first time that immisci-
ble ligquids are formed at high temperatures in Al¥s-based
systems. The temperatures at which such liguids form are,
however, higher than is feasible for adoption in most current
chemical technologies. Of the various materials evaluated as
solvents for the volatility process, the greatest potential
for application was displayed by the KF-ZrF;-AlF3 system.
High solubility and good dissolution rates are afforded by
the inexpensive solvent salt KzZrFs. At operating tempera-
tures, approximately 600°C, the AlF3 capacity of the solvent
is in excess of 25 mole %.

 

INTRODUCTION

Development of the fluoride volatility process is sought as a useful
and effective alternative to conventional aqueous processing for recovery
of uranium from spent nuclear fuels. The process depends on dissolving
fuel elements in a suitable molten fluoride solvent by passage of HF gas
followed by fluorination of the resulting melt to volatilize uranium as
UFa.l The UFs is purified by selective sorption on solid NaF or other
fluorides. Molten~salt Tluoride wolatility processing of nuclear rfuels
offers advantages not available with other chewmical reprocessing methods:
(1) the process is sjmyle,z requiring only a hydrofluorination step and
a fluorination step instead of the many steps--dejacketing, acid dissolu~
tion, precipitation, filtrafiion,3m5 gte.~~characteristic of the usual
aqueous procedures; (2) the uranium is recovered as UPFg, the form reguired
for isolope separaticn,e and (3) disposal is made of essentially all fis-
sion products as water-insoluble solids. It is also one of the Tew methods
that can be used Tor processing certain ceramic fuels.{

The process has been previously applied successfully to zirconium-

1,8,9

matrix fuels. Currently, it is considersd for alwminum-matrix fuels

10,11 and because of the need

because of theilr use in numercus reactors
anticipated for processing large quantities of these fuels.

One of the most important aspectis of volatility process development
in adapting the process to a particular kind of fu=l element is the selec-
tion of a suitable molten~salt solvent into which to pass the HF gas during
dissolution. A preliminary swrvey of prospective AlF3 solvents reported
by Boles and Thoma,l2 showed that a Bel;~-LiF solvent provides moderately
good solubility for AlFs;, bul,; because of expense and toxicity, an slter-
native solvent system is vpreferred. They considered KF-Zr¥, solvents to
be of potential use. The preliminary data obtained at that time were too
limited for a critical selection of optimal solvents for use in volatility
processing. Accordingly, a more intensive search for sclvents was dniti-
ated using newer methods which permit rapid accuwmlation of large numbers
of liguid~solid transition data.

Desirabple characteristics of the solvent for the process include the
following:

1. TIow cost.

2. Tigquidus temperatures below 600°C.

3. Substantial solubility of Al¥F3 at 500 to 600°C.

4. Low vapor pressure.

5. Low viscosity.

6

. Noncorrosive with respect to the INOR-8 container.
In reprocessing aluminum-matrix fuel elements, the temperature must
be kept below the melting point of alumimm (600°C)13 to avoid forming
liguid metal, which is corrosive to the container alloy, INOR~-8. The sus-
ceptibility of INOR-8 to corrosion by liguid aluminum is offset by its
excellent structural properties at high temperatures and its resistance
to corrosion by fluorine and HF. It is desirable to prevent the maximum
operating temperature from exceeding a limit of at least 50°C lower than
the melting point of aluminum; consequently, 600°C has been tentatively
chosen as the maximum temperature for the process. The economic feasibil-
ity of the process is highly dependent upon the cost of the solvent.
Solvents in which the saturating concentrations for AlFi3 are as low as
8 mole % may be useful if the solvent costs less than 50 cents a pound.
I, however, higher-priced solvent components are required in order to
obtain the desirable solvent characteristics, the economic feasibility

of the process may then require higher saturating concentrations of AlFj3.

Choice of Constituent Fluorides as AlF3 Solvents

 

The cholce of possible solvent constituents can be rapidly narroved
to one group: cheap fluorides which are stable in the presence of gaseous
HE and fluorine. Cationic constituents should either have only one va-
lence state as the flucride, or both the lower fluoride, existing during
dissolution, and the higher fluoride, formed during fluorination, should
be noncorrosive, possess sufficiently low vapor pressure, and have suit-
able melting characteristics for use in the solvent.

Previous to the ORNL work on the volatility process there was little
published information concerning the attack of aluminum metal by molten
fluorides, and none regarding the relationship of HF or other dissolved
oxidant gases to this attack. It is known that aluminum reacts with molten
alkali metal fluorides and with molten cryolite to form free alkali metal

14 The metal also reacts with molten

and the appropriate fluorocaluminate.
KaThFg, forming z Th-Al alloy and a potassium fluorcaluminate. Aluminum
metal reacts with solid CeFs at 1000°C to form cerium metal and AlF3z.
With KaTiFe¢, aluminum reacts to form KpAlFg, free titanium metal, and an
opaque phase believed to be a complex of TiFg.ls Published free-energy

valuest® favor the formation of TiFs3 by the reaction:
Al + Zw, = 3TiFs3 + AlF3 .

The rree~energy values suggest that the alkalil metal {luorides ThFgz, Cels,
AlFa, and TiFz may serve as potential sclvent constituents. Mclien Snka
and NizHF2 both attack aluminum.m&%&il7m19 but have disadvantages which
preclude thelir use as poteptial volatility solvents. The former is very
corrosive to structural metals;zo the latter presents a vapor prcblem

and decomposes during fluorination. Aluminum reacts vigorously when heated

14 Adumima reacts with molten alkali

with fluorides of Ni, Cao, Fe, or Os.
metal fluorcoborates and fluorosilicates to form, respectively, aluminum
boride and silicon or silicides. The use of fluoroborates and fluorosili-
cates in the volatility process presumably should be avoided because
vorides and silicides are s0 inert that they are likely to remain in the
processing solvent in the form of an annoying sludge.

Relatively few Tluorides possess the properties reguired for their
use as major constituents of a solvent. Fluorides of nonmetals, semimetals,
inert gases, and the platinum metals either have too high a vapor pressure
or are too corrosive to be considered. (See Tsble 1 for boiling points.)
These objections apply as well to fluorides of Cu; Mo, Ag, W, Au, Hg, Nb,
Ta, V, Cr, Ma, Co, T1l, Pb, and Sn. Scarcliiy would eliminate consideration
of most rare earths, all transuranium elements, also Sc, Y, Re, Hf, Te,
Fr, Ra, Ac, and Pa. Fluorides of Zn, Ga, In, and Cd do not qualify because
their reduetion by aluminum would form corrosive liguid metal. Uranium
fluoride of natural isotopic camposition is objectionable because its use
would alter the isotoplic composition of the fuel being processed. Thus
the list of possible solvent constituents is therefore narrowed to the

following fluorides:

Li¥ BeFa Al¥3 TiFs FeFo
NaF MgFos LaFs ZxFe NiF'as
KF Cal» Cel's Thi'y

RbF Sr¥2

CsF BaFp

Most of the fluorides in this group are suitable only as minor constitu-~
ents of the solvent for the following reasons:

1. EKbF, CsF, LaF3, and ThF4 are moderately expensive.
n

2. TFeFz and NiFz in high concentrations may present corrosion
problems during fluorination.

3. MgFp, CaFa, SrFa, BaFz, and CeF3 have very high melting
points (see Table 1).

4, TiFs in high concentrations may exert somewhat excessive
vapor pressure.
These compounds were, therefore, considered not as major solvent constit-
uents but merely as possible additives for possibly depressing the liquidus
of a promising solvent. The remaining six compounds--LiF, NaF, KF, BeFp,
AlF3, and ZrFi--were given the principal consideration as solvent compo-

nents.

survey of Potential Solvent Systems

 

Except for BeFp, which is too viscous, none of the promising fluoride
constituents individually has the necessary low melting point (below 600°C)
to be used directly as the solvent (see Table 1). A mumber of binary mix-
tures of these fluorides do, however, form adeguately low melting eutectics
(see Table 2). The only Al¥s binary system which provides sufficiently
low melting mixtures for possible use in the process is the KF-AlF3 sys-
tem; its capacity for additional AlFj3 at the process temperature is,
however, limited to about 5 mole %, too low for process use.

In order to find low-melting solvent systems suitable for the proc-
ess, phase relationships were studied in the ternary systems formed by
dissolving Al¥3 in a molten binary solvent. Little concern was given to
more complex solvent systems because the study of polycomponent systems
delineating the phase reactions occurring as AlF3 dissolves in such sol-
vents iz too inveolved to permit adequate characterization in a reasonable
length of time. In addition, the probability of diagnosing the cause of
off-performance difficulties in engineering tests by identification of
crystallized solids is remote for multicomponent salt systems unless de-
tailed investigation of the phase behavior has been made. Accordingly,
when a fourth component was considered, it was usually only as @& minor
addition to a promising ternary system, e.g., A~B~AlF3, included in oxrder
to lower the liguidus temperatures enough to meet process requirements.

Although the fluoride velatility process is intended for reprocessing
o

fuels conteaining both uranium and alumimen, consideration was given only
to the solubility of the resulting Al¥F3z in evaluating 2 solvent. The
uranium content of these fuels, generally less than 1 at. %, yields too
low a UFg concenvration in the solvent to affect the ligquidus temperature

significsently.

PROCEDURES

2,21

. . - - l ~a
The initial phase studies of the solvent systems were performed

,._J

primarily by classical procedures wiich proved to be generally inadeguate
for systems containing Al¥i., Because AlFz and alkali metal fluoroaluwni-
nates are frequently not microscopically distinguishable from each other,
these systems were not amenable to studies employing the guenching tech-
niqus. Alsc; the thermal changs at the liouidus temperature was often
too small to be jesdily deteected by thermal analysis. Even when a thermal
change was detected,; its interpretation, without visual observaltion or
accompanying quench data, was equivoeal..

Most of these difficultlies were overcome Dy melting the mixtures in
an inert-atmosphere glove box (Figs. 1 and 2) and observing phase changss
through a window. GLigquidus temperaturcs were determined accurately (nsu-
ally * 2°C) by noting the temperature at which the first crystals were
obgerved in a cooling melt. The melts were stivred manually to prevent
supercooling and to ensure uniTormity of composition and temperature.
Usually aboub & mole of sall contalned in a hydrogen~fired; polished nickel
crucible was used for the study. Intense illumination was provided by a
Zirconarce photomicrographic lamp (Fish-Schurman Corporavion, New York).
The lignt Trom this instrument overrides the near infrared hackground
radiation from the melt to temperatures of aboul 1200°9C and thus facili-
tates determination of liguidus under conditions where other methods would
be difficult or impossible. Atmosplicre control was obtalined by evacuating

the glove box to 30 u and refilling with helium purified by passage through

9

activated charcozal cooled with liquid nitrogen. 'The procsdure is ra

As many as sixteen composibions in & given system can be studied in an

nt

8=nhr period by sequential additions of prewsighed specimens. Since the
melt may be observed as phase changes occur, the apparatus 1s also useful
for obtaining interpretable thermal analysis data. To enswre accuracy,
the temperature recorder was periodically standardized against LiF melt-
ing at 848 * 100.22"24

The precision of temperature measurements obtainable in routine use
of Chromel-Alumel thermocouples is generally believed to be *59C. Occa~
sional calibrations with pure salts have shown that the accuracy of thermal
transition temperatures reported here is within the #*5°C precision limits.
Correspondingly, the accuracy of visual transition data reported here is
within *3°C.

A similar visual procedure called "visual polythermal method"” has

been used by Russian investigators,25

but thelr procedure seems inferior
to that used here in that it apparently does not permit (1) agitation of
the observed melt, (2) addition of salt during the run to alter the com-
position, or (3) use of vacuum to control the atmosphere.

A sharp increase in viscosity 1s often displayed by molbten salts as
they cool to temperatures approaching the liguidus. This effect was noted
in most of the systems discussed in this report and was useful in signal-
ing the onset of crystallization. Such a sharp change in viscosity has
been observed also by Velyukov and Sipriya26 for NaszAlFs and L13A1Fg and
by 8111527 for “nCla, A sharp change in the electrical conductance at
phase~-trnasition temperatures was also noted. Preliminary measurement of
electrical conductance using an ohlmmeter (Model 630, Triplett Electrical
Instrument Company, Bluffton, Ohio) indicated that it may also provide a

procedure useful for studying AlF3 systems.

MATERIALS

Purity of the fluorides used in the phase studies was very important.
The molten~-salt systems were studied primerily by a visual procedure which,
in determination of a liquidus temperature, depended on the appearance of
precipitate, oftten in such a form that 1t clouded the melt. Accordingly,
any impurity which clouded the melt interfered with the study. Because
they reacted to form very sparingly soluble phases, hydroxides and water
vapor proved Lo e especizlly objectionable and in some cases initially
veyy difficult to remove or avold in the preparation of thess [luorides.
Hydroxides and moisture were additionally objectionable because they
attacked the metal comtalner, contaminating the melt with highly colored
nickel ion. Often such melts gave rise to lrreproducible Liquidus Tenper -
atures due, presumably. Lo a progressive increase in oxide concenbration
as the hydroxide reacted.

ree methods were found to e useful for preparing flusrides of low
oxygern content:

1. Vacuum sublimaiion: Applicable to purdifying cafimer*ial
AlF3 and ZrFs, as the corrasponding oxides have exbremely

low volatility. ZrFs was obtained with as 1little as
250 ppm oxygen using the apcaratus shown in Flg. 3.

4

 

2. Vacuunm distilletion after vrecipitation of oxide:
Applicable to KF. The vapor preasure of KOM =t 850 to
1000°C¢ is high enough to preclude reduction of oxygen
impurity to less than 1200 ta 1500 v Dy distillatifin.
However, in moliten potassium fluoride, KOH reacts with
various m;tal fluorides to prc:10¢tafe metal oxiuea as
Tollows:

 

+OXHET XKF

ZMF  + xKOH — 2ZMO
X x/2]
and purification of KF is possible by vacuum distile
lation from the remaining wmolten mixture The use of

Fel: and reFa to Pfflfilp¢uatu oxlde pirod iiced\crystals

of KF which combained 900 ppm oxygen: the use af
2.3 mole % UF, gave a product containing only 300 PP

3. Ammonium bifluoride Tusion: Fasion of alkali metal
fluorides with hydrated AlF; in the presence of molten
N W2 proved to bhe useful for preparing ILi:A1Fg,
NezhlFg, KallFsg. KA1F,, and CszA1Fg. Slowv cooling of
the melts to promote crystal growith and selection of
the better-crystellized portion served to provide
additional puritication. Fusion of NHyHF,; with hy-
drated AlFa formed (NH4)3§ ..... 1Fg. 1us thermal decampo-
sition at 600°C in a helium stream yielded anhydrous
AlFa which was comparable in purcity to the sublime=d
produect. Alkalins earth fluorides were also purified
by NHzIF; treatment.

 
RESULTS AND DISCUGSION

In the search for high~capacity solvent systems for use in the fluo-
ride volatility process, equilibrium solubility data for aluminum fluoride
in seven fluoride systems were obtained. The systems examined included
LiF-KF-AlF3, K3AlFe-1iF-CaF,, K3Al1Fg-LiF-SrFy, K3AlFg~LiF-ZnF,, LiF-NaF-
AlF3, NaF-KF-AlFj3, and KF-ZrF4-~AlFs. The phase diagrams constructed fram
the data obtained in these examinations show that only one of these systems,
KF-Zr¥Fas-AlFi;, can be expected to have practical application as a solvent.
New data were obtained for the limiting binary systems LiF-AlF3, NaF-Al¥Fs,
KF-AlF3, and KF-ZrF¥4. New phase diagrams of each of these systems are
shown in Figs. 8 to 11. Experimental data for the systems reported here
were collected simultaneously for several systems, thus making it possible
to curtail the efforts on any one system as the development in another

system showed promise.

AlF3 Melting Point

Previously reported experimental values for the melting point of
Al¥3 range from 98612 to 104000.28 All of our visual observation data
indicate that these values are low and that the melting point is higher
than the reported sublimation temperature, 127000,29 though not as high
(19200C) as was regarded by Steunenberg and ngel.BO Because high liquidus
temperatures and vapdr pressures prevented visual study of mixtures con~
taining over 56 mole % AlFj3, too little data were obtained to permit a
good extrapolation of its melting point. Since AlF3 is of similar struc-
ture to CngBl and of comparable size relationship,32 we surmise that its

equilibrium melting point at 1 atm is close to that for CrFs, 1404°C.

Systems Based on LiF-KI¥

 

Mixtures of the lightest alksli fluorides, ILiF, NeF, and KF, are not
so low melting as those obtainable with EbF or CsF; the cost of these latter
two materials, however, precludes their economic use in process development.
The binary mixture of the cheaper fluorides which affords the lowest~-melting
solvent is the cguimolar LAF-KF eutectic wixvure which melts at 5007°C (sec
Fig. 4). ‘The phase diasranm of the system LAF-XF-AlFa, consbructed on the
basis of the datva shown 1n Table 3, is given in Fig. 5. Ilovariant equilib-
ia are listed in Tsble 4. The diagvan shows clearly that LiF-KF mixtures
cannot provide useful solvents because off the extent to which the primars;
phase fields of the high-melting caonpounds K3Allg and KzLiAl¥e approach the
limiting binary system L[iF-Al¥Fa3. The iternary system 1s canprised of the
subsystems KE-LiF-Kz81Fs, KaAlFe~-11i2AlFg~-1aF, and KaAlF,-TiA1F5.  Both of
the composition seclions KallFg-LialAlFs and K3Alls-TaiT are apvarently quasi-
binary. The minimum ligquidus temperature zlong the composition section
L

K:fi..’fi.l.st“Lj..P; is I?QOOC; \.«l-a— Jrl‘i t}A\. Seb L. L\.J.LL I

/\
< §>
- ,.J

b
= N

&

high liguidus proiiie for the system 1i
fram possiblis use as a sclvent. 'The possibility that the liguidus Tor some

3 o Y S £ —— - L e o S e -
inexpenglive Tour~-compoient cambinations mi

for LiF-KF-A1F3 gsve impebtus to an investigatlon of the effect of the addi-
tives Ja¥y, Srify, and “nlb's. Accordingly, an investigation wag made of thc
extent to which scie of the Group 1l fluorides depressed the ternaxy lia-
uidus. Resulbs of these experiments are given in Taebles 5 to 7. Ths minor
benefits of adding thess components to the leitnsry mixbtures were insufflfi-
cient to sugzes! thet sxiensive igvestigation of the mullicomponent systems
was practical.
At AlFa conceutrations sbove 50 mole % in the LiF-KF~-AlFa syatem we

cbsarved imniscible liquids. Thelr phase relationships are not yet ade-

quately explzined; cither twoe tiue (i.e,, igsotronic, liauids or one true
» s 33-34 -
liguid and one liguid rstalline phase (resoPhase ’77) ecould be present.

The System LiF-Nap-AlFa

 

An examinabtion was made of The liguldus surface of the LiF-Nab-AlFa

t Al¥3 concentrations between 0 and 35 mole %. As in the TiF-KF-
1¥3 system, the composition area at wnich liguidus temperstures arc bhelow

500°C is much too small for the systen to be of practical value in the

volatility process. The vhase disgram, shown in Fig. 6, is dominated by
the cryolite phase NazdlFg, whieh erystallizes {ram LiF-Nap-AlFi3 ligul

as o high~-melting phase Tor much of the lower AlFj3 part of the syslewm.

Data obtained Por the system are given in Table &.
The System NaF-KF-A1F4

 

Preliminary investigation of the system NaF-KF-AlFj3 made by Barton
et gi.35 indicated that aluminum fluoride solubility was negligible in
NaF-KF mixtures except at temperatures above the NaF-KF eutectic. This
inference was corroborated by additional experiments conducted as part

of the present investigation.

The System KF-ZrFgz-Al1F4

 

Binary systems of the alkali fluorides with Zx¥s; afford low-melting
solvent mixtures for the heavy-metal fluorides UFy; and ThFg and can be
expected to provide useful solvents for AlF3 as well. Othef solvents
would be preferred because of the high cost of ZrF, and because of the
volatility of Zr¥Fgi-rich ligquids at high temperatures. Nevertheless, the
liguidus temperatures at concentrations of 35 to 45 mole % ZrFs; in the
KF-Z2rFs system and the availability of KpZrFe¢ as an inexpensive reagent
suggested the use of the reagent in the preliminary evaluation of the
aluminum solvent systecms.l2 The experimental data obtained in this inves-
tigation are shown in Table 9, The phase diagram of the system is shown
in Fig. 7. Crystallization reactions within the system KF-ZrF4-AlF3 have
been characterized in detail except for those involving AlFj3 and ZrF,.
Both of these coamponents are high melting and volatile. Their phase reac-
tions are extremely difficult to exsmine at high temperatures because of
this volatility end their low heat of fusion, which preclude most dynamic
methods for obtaining phase data. Their crystalliization reactions in the
ternary mixtures suggest that the only interaction cccurring between them
at high temperatures is the formation of a eutectiec. For these reasons,
we have cmitted investigation of the limiting binary system AlF3-ZrF,;. At
600°C, the maximm acceptable temperature for the process, a solvent, KF-
zrFs (63-37 mole %), was found to have 15 mole % AlF3 capacity. Idguidus
temperatures In the binary system KF-ZrFs exclude the use of solvents
richer in X¥F. It can be seen from the phase diagram of the system KF-ZxrFs-
AlF3 (Fig. 7), constructed in this study, that by a single addition of KF
after partial dissolution of the fuel element the solubility of AlF3 is
increased from 15 to 26 mole %.
172
Two dnmisaible ligulds or & liguid and a liquid ecrystalline phase
o — o o g e B —— T i - QDA
were found in the KF-Al¥a binary systam above 53 mole % Al¥s at 980°C.
The two-ligquid region apparently cxtends iaho the KE-AlF3-7Z1xFs teprnasy
systerm hut not Lo compositions curreally of joterest as volatllity sol-
VauLe.
CONCTUSTONS
The resulis of the investigstions reported here togelther with the

L - -

results of dissolulion rate and corrosion rate tests made by Chemdcal

£

Technology Trivision persomnel indicais conclusively that the syshom KE-
1

fuels. They alss shov that thic essential ceriteria necessarily imposed in

sciecting a solvent systan, il.e.; maximm eguilibrium solubility, mexionug
rates of dissolution, minimal rsies of container vessel corrosion, and
minimus solvent costs, are not met (competitively) by sny of the other
systems considersd in preliminary or cwrrent studies. Accordingly, more
caniplete dava have been obtained Tor the system KP-Zrke-Al¥s thae Tor ang

of the other systems reported here. It wss observed for the first tims
that dvmiscible liquids are formed at high temperatures in AlFs3-based
systems. The Lemperatures at which such liquids form ave; however, higher

A

in most current chemical technclogies.

g:
o
=
M
0
Hy
¢
w
L
v
[e)
D
h
2
=3
o
5
<
o
}- .
O
£

On evaluating the merits of possible AlFa solvent silxtures with
respect to phase; corrosion, snd cost data,; the binary mixture KF-Zx¥Fg

by \ - -
(63wd? mole %) was found to satisfy best the composit

9

eriteria. OCn the
basis of this evaluation, this mixture is vecommendsed by Reactor Chamistiy
and. Unemical Technology Division personnel as the most satisfactory sol-
vent Tor dissolution of alumimun-based materials the fluorids volatiiity

process,
13

Table 1. TFluoride Transition Temperature and Free Fnergy
of Formation

 

 

Compound  Melting Point®  Boiling Point™®  "Fr, 2089k
(oc) (o¢) (keal/mole )¢

HE ~83.36 19,46 64,7
LiF 848,02 1681.0 138.8
BeFg 545,02 115%9.0 207.5
BF'3 -128,72 -99,9 269.5
CFy 151.9
NaF 996, 02 1704.0 129.0
MgFa 1263.0% 2260.0 250.8
AlF3 1273.04 306.0
SiFy -90.3% ~95.5 360, 0
PFs -93.8 w84, 6

KF 857.0% 1502.0 127.4
CaFa 1418, 0% 500.0

SeFa 350.0
TiF, 283, 350.0
VF 3 1406.0% ~1400.0 254,0
Vs 19.5 4.3

CrFa 172.0
CrF; 1404, 0F 250, 0
Cr¥'s ~150.0 ~150.0 327.0
MnF'2 930.0 180.0
MnF 2 223.0
FeFs 950,08 1800.0 158.0
FeFa 1300.08 ~1300.08 219.0
Co¥'z ~1200.0 ~1725,0 147.0
CoF s 174.0
NiFa 1677.0 147.0
CuF'p 118.0
ZnFp 872.0 1500.0 164.0
GalFa ~ 239.0

GeFa 271.0
14

Table 1. (continued)

 

-AF

 

comoound  Melting Point™  Boiling Point’’° £, 298%K

= (oc) (oc) (keal/mole)®
AsFa -5,0 58 189.0
Selg -34 6 45,9 221.8
ROF 798.0°% 1408.0 125.7
ST¥o 1400. 0 2410.0 276.7
YF3 380.0
I 910.0° ~900. 0% 424, 0
NbF 5 78.9 233.3 320.0
MoFs 17.5 35.0 383.0
RnF's 279.0
PAF 3 105.0
Agy 435,0 44,3
CdF, 1110.0 1748.0 153.3
In¥F3 234.0
SnFa 215,05 850. 05 147.0
Sni¥y, 237.0
SbF' 290.0 376.0 200. 0
SoF s 8.3 142.7 286.5
TeFg -37.8 38,9 292,0
CSF 682.0% 1251.0 124.5
BaF, 1290.0 2260.0 274.0
LaFs 403,0
CeFs 1460.0° 398.0
HEF, 413.0
TaFs 95.1 229.2 339.0
WFe g.2 17.0

ReFg 18.8 47,6 258.0
Oskg 3,4 47.3 199.0
PLF2 72.0
AuFs3 84,0
HeF s 645.0 647.,0 83.0
15

Table 1. (continued)

 

b,c L 5F

 

Compound Melti?gcfioint& Boili?gcfioint (kcii/i2§:§c
T1F 327.0 655.0 60.0
TiF 159.0
POF, 824.0° 146.6
BiF3 850. 0 200. 0
BiFs 151.4 230.0
ThFy 1100.0 1680.0 454.0
UF4 1036.0 1417.0 421.5

 

“Landolt-Bdrnstein, Zahlenverte und Funktionen, Vol. 2,
Eigenschgften der Materie in Ihren Aggregatzustidnden, Paxrt 4,
"Kalorische Zuslandsgrdssen, opringer, Berlin, bth ed., 1961,

P, Brewer, 'The Fusion and Vaporization Deta of the
Holides," Paper 7, p 193-275 in The Chemistry and Metallurgy
of Miscellaneous Materials: Thermodynamics, ed. by L. L. Ouill,
MeGraw-Hill, New York, 1950; Metallurgical Laboratory Report
CC-3455 (1946).

 

 

 

A, Glassner, The Thermochemical Properties cf the Oxides,
Fluorides, and Chlorides to 25000K, ANL-5750 (1957).

dSublimaticn point.

 

 

 

®B. J. Sturm and C. W. Sheridan, Inorg. Syntheses, 7,
g7 (1963). =

s, g Sturm, Inorg. Chem. 1, 665 (1962).

Elnreported melting points based on work at ORNL, sometimes
only an approximate value based on preliminary experiments.
16

 

 

 

 

 

Table 2. Binary Fiuoride Systems of Potential Use as FProcess Solvents
Eub»?tlb(gi?Pm%"tGJQ Components Congggfigifiégm{ag §e%bld‘Reference

706 IiF-AlF3 1405 a,b
689 TiF-Al¥ 3 32.5 2,0
685 NaF-AlF 3 45 a,b

570 KF=-AlF 3 40 a,b
370 BelF,~AlF 5 22 c

652 LiF-Na¥ 40 2,4
492 LiF-K¥ 50 a,d

71.0 Nak=-KF 60 a,d
355 LiF-Bela 52 &a,d,e
365 NaF-Bel's 55 a,d
340 NaeF-BeFa 43 a,d
323 KF-Bels 7.5 a,e

330 KF-HeFs, 59 a,e
507 LiF-Zrity 49 d,e

500 Nef-7xr¥Fs 40.5 d,e
430" KE -7k, 42t d,e
720 Na2AlFg~11 2A1Fe 60 a.

936 Na 3A1Fg~KA1F, 39 O

%E. M. levin et al., Phase Diagrams for Ceramists, Am. Ceram. Soc.
1956,
b

e Tlfifififman5=

5 co~-Chenmical Constants of Binary bystems in
Concentrated S0lutioiid

1.
Interscience, New York, 196C.

 

 

°R. L. Poles and K. E. Thomz, Volatility Process Phase Studies - A

Survey of Molten Fluoride Solvent Mixtures Suitable for Dissolution of
AlF3, ORNL TM-400 {oct. 22, 1962); {the binary eutectic composition and
t ratur* were not actually determined butl were cotimated from prelim-
Fo~-AlF 4 ternary sysoes

 

 

 

 

R. E. Thoma, cd., Thase Liagrams of Nuclear Reactor Materials,
ORNL-254& (Nov., 6, 1952).

Levin, op. cit., Part 11, 193%.

on current work.

£y

I .
Values are base
Table 3. LiF-KPF-A1F3 Liguid-Solid Trensition Data

 

Second

Liguidus Temperature (°C) Solidus Temperature (°C)

 

Crystallization
Temperature (°C)
(Thermal Analysis)

Composition (mole %)

1 ; +vg o]
TiF %5 AiT Thermal Blectriecal

Analysis Conductivity

Visual Thermal Electrical
Cbservation Analysis Conductivity

 

 

0.0 848 848 8§48
8.5 11.5 779
86.0 14.0 725 735 711
80.0 20.0 765, 5 766
75.2 24 8 784, 5 785
75.0 25.0 772 771 787 771
71.6 28.4 775 775 699
66.7 33.3 735 734 738 708 711
63.2 36.8 745 772 706
62.5 37.5 728
60,9 39.1 730.5
60.0 40.0 770 770 775
58.8 41,2 V47 747 710
57.2 42.8 g12 812 g24
57.1 42.9 786
55,6 by, L 802 802 709
52,7 47.2 860
50.0 50,0 1035

75.0  25.0 996 995 | 995

LT
Table 3.

(continued)

 

 

 

 

 

conosition (aote 1) —prigude Tomerature Co)  orysiallinanson  Sgpidus Tememate (o
LiF X5 AlF3 goervation Analysis Conductivity (Efmperature (o) Analysis Conductivity
Thermel fnalysis)
57.2  42.8 570 559
54,6 45.4 569 565
50.0  50.0 642 575
42,8 57.2 710
5.9 70.6 23.5 971 714

1.1 66.7  22.2 951

20.0 60.0  20.0 917

27.3 54,5 18.2 889

33.3  50.0 16.7 §65

41,7 43,7 14,6 839.5

45.5 40,9 13.5 821

50.0 37.5 12.5 800

55,6  33.3 11.1 769 718

62.5 28.1 9.4 727

64.1 26.9 9.0 722.5 722 722

714 21.5 7.1 749.5 718

83.4 12.5 4.1 791

67.5 7.5  25.0 731.5 730 647

61.4 13.6  25.C £90

56.25 18.75 25.0 649.5 649

8T
Table 3. (continued)

 

Liguidus Temperature (°C) Second

 

 

Crystallization  Soiidus Temperature (OC)
Temperature (°0) Thermal Electrical
(Thérma} Anslysis) Analysis Conductivity

Composition (mole %)

TiT =5 AiTs Visual Thermal Electrical

Ubservation Analysis Conductivity

 

 

50.0 25.0 25.0 696 645
45.0  30.0 25.0 727
40.9 34,1 25.0 749 749 645
37.5 37.5 25,0 767 762 645
30.0 45,0 25.0 802 778
25.0  50.0 25.0 843 779
33,3 16.7  50.0 976
25.0 25.0 50.0 893
20.0 30.0 50.0 858
16.7  33.3  50.0 824

100.0 854 852 g52
15.0  40.0  45.0 593 589 567
14.3  38.0 47.7 623 622 587 565
13.0  34.8 52,2 970 593 564
12,0 32.0 56.0 1098 592 563
15.4 30.8 53.8 1043 597 561
21.4 28.6 50.0 813 805 605
26.7 26,7  46.6 690 685

25,0 25.0 50.0 858 608 560

61

 
20

Table 4. Invariant Equilibria in the System LiF-KF-AlFj

 

 

 

Comosition (nole §)  Temperalure  py. of pui1ibriu
50.0  50.0 492 Butectic

93.0 7.0 850 Eutectie

56.0 44,0 560 Butectic

50.4 49,6 575 Peritectic
85.5 14.5 711 Eutectic
4.0 36,0 710 Futectic
53,0 47.0 890 Peritectic (2)
56,0 19.0 25.0 648 Butectic®
33.0 42.0 25.0 778.5 Peritectic™
28.1  62.5 9.4 722.5 Butectic’

6.0 48.0 46.0 200 Futectic

45,5 53.0 1.5 490 Futectic

 

%In subsystem KablFg-Td 3AlFg.

bIn subsystem K3AlFg-IiF.
21

 

 

 

 

Table 5, K3AlFg-LiF-CaFs Liquid-Solid Transition Deta
Comoattion (one §) Iy Thesal tesiyts Do
6 at2 Tigquidus Liquidus lization Temp. Solidus
100,0 993.5 992 992
80.0 20.0 971.5 %9
66.7 33.3 955 950 937
57.2 42.8 973 Oy
50,0 50.0 1011 95
40,0 20.0 40D.0 959. 5 905
33.3 33,3 33.3 908 875 682
28.6 42.8 28,6 869,5 858 682
25.0 50,0 25.0 839.5
20.0 40.0 40.0 961
18.2  36.4 45,4 996. 5
13.3 53,3 33.3 912
11.7 58,9  29.4 876 850
8.7 69.6 21.7 781 697 688
8.0 64.0 28.0 848 695 688
7.6  61.6 30.8 870 695
18.9 54,1 27.0 827.5 715 680
16.3  60.5 23.2 799.5
14.3  65.3  20.4 779 759 675
12,7  69.1 18.2 761
10.8 73.8 15.4 730.5 714 680
9.3 77.4 13,3 719 710 680
24,1 69.0 6,9 831 720
22.6  64.5 12,9 830.5 714 680
21.2 60.6 18.2 826
20.0 57.2 22.8 825 712 675

 
22

 

 

 

 

Table 6. K3AlFg~Sr¥y Idquid-Solid Transition Data
- Ligquidus Liguidus lization Temp. Solidus

20,0  50.0 1048 757

60.0  40.0 270 967

66,7  33.3 913

75.0 25.0 832

80.0 20.0 74

83.3 16,7 780 778 767
7.7  76.9  15.4 139 ‘703 692
14.3 71l.4 14,3 776 705 628
20.0 66,7 13.3 829
25.0  62.5 12.5 853 848 703 695
66,7 16,7 16,7 9635.5 964
20,0 25.0 25.0 939 230 686
40.0  30.0 30.0 930 918 690
33,3 33.3 33.3 935 903 690

 
23

 

 

 

 

Table 7. K3AlFg-liF-/nFs Liquid-Solid Transition Data
Campostton (mole B peciime T Bseond Ol
Tiquidus Ligquidus lization Tenp. Solidus
1.2 87.8 722.5 722 722
11.8  84.0 4.2 720 720 655
11.3 80.6 8.1 713 712 663
0.9 77.5  11.6 705 705 668
10.4  74.6  14.9 701 701 670
9.7 69.4 20.8 690 689 670
9.0 65.0 26.0 675 675 668 563
8.5 61.0 30.5 668 668 562
8.0 57.5 34.5 G'73 562
7.5 54,4 38.1 666
7.0 50.0 43.0 667

 
24

Teble 8. LiF-NaF-AlF3 Liquid-Solid Transition Data

 

 

 

 

Composition (mole #)  pZindy T ford Gt
- Tdquidus liguidus lization Tewp. Solidus
75.0 25,0 1007 1005
55.6 444 720 622
50.0  50.0 853
47.7 52.3 1030
45.4 54,6 1083
g.3 41.7  50.0 858 o174
15.4 38.4 46,2 729 660 599
21,4 35.8 42.8 623 645 600
26.7 33.3 40.0 662 638 601
35.3 22.4 35.3 636 627 603
11.1 66,7 22,2 254 785
20.0 60.0 20.0 919 Thts 092
27.3 54.5 18.2 882 627 694

33.3 50.0 16.7 849 722 694

 
25

Teble 9. The System KF-AlF3~ZrFs

 

 

 

 

cogortsion (mje ) VI7EY Mermal alyere pred
Llquidus Liguidus lization Temp. Solidus

75.0 25.0 920 * 4 920 920

75.0 25.0 935.5 %+ 4 932 932

50.0 50.0 604 t 4 600 448P, 480

45,4 54.6 580 * 6 580 440

40.0 60.0 573 + 4

33.3 66,7 650 * 4

25.0 75.0 760 * 4

50.0 50.0 648 + 3

bl bh4 11.1 975 t 3 540

40,0  40.0 20.0 1030 % 4 520 408

33.3 33.3 33.3 >1030 £ 4 585,442 410

66.7 22.2 11.1 890 * 3 499 480

70.0 10.0 20.0 920 + 3 915

72.7 9.1 18.2 938 * 3 934

72.7 9.1 18.2 942 % 3 G34

76,9 7.7 15.4 932 + 3

71.4 14,3 14.3 934 t 3

66.7 13.3 20.0 847 + 3 482

62.5 12.5 25.0 570 £ 4

62.5 20.8 16.7 649 * 4 505 485

53.6 17.9 28.6 721 * 4

46,9 15,6 37.5 743 4 743

41,7 13,9 bty b 821 t 4 818 582 438

37.5  12.5 50.0 869 £ 5 858 575 428

34.1 11.4 54.5 >1000 £ 5

 

precision limits on thermal transition data are approximately * 5°

bMetastable transition.
26

 

 

 

 

 

 

i - . T . a c - el

Table 1.0. Invariant BEguilibriz in the DYSiEfi;KFmAlfiSMZTFQr
e S e [ fommi | m T

Composition \naie ) Temparatuie - s
o e e 7 vpe of mguilibhrium
KR A1F; Jr¥s (°c) P =

86.0 14.0 765 Futectic

2.0 36.0 590 Peritectic

6C.0 40.0 445 Peritectic

58.0 42.0 430 Eutectic

45,0 55.0 440 Futeectic

63.0  15.0 22.0 490 Butectice

~55.0 ~5.0 ~40.0 400 rutectic
TKE-A1F2 invariant eguilibria are given in Table 4.
27

UNCLASSIFIED
ORNL-LR-DWG 20293

VIEWING PORT

      
    
  

GLOVE PORT WITH
RUBBER GLOVE

—UTILITY
CONDUIT

.

HEATED
COMPARTMENT

 
28

UNCLASSIFIED
PHOTO 63518

 

Fig. 2. Visual Study Apparatus with Accessories.
Filgs 3.

Apparatus for Vacuum Sublimation and

Distillatlion.

 

 

UNCL ASSIFIED
PHOTO 64598

6C
TEMPERATURE (°C)

UNCLASSIFIED

ORNL~LR—DWG 35483

 

t000

800

 

800

1

 

700

 

 

600

500

 

 

400

 

 

 

 

 

 

 

 

 

 

 

LiF

10

20

30

40

Fig. 4.

50 60
KF {(mole %)

The System LiF-KF.

70

80

90

KF

og¢
31

UNCLASSIFIED
ORNL-DWGE £3--16684

AlF3 ~1400

 

  
   

  
  
 

\ \rF 645 \

L!3A|F€
r 64 \
\?\X ~540 ?"')O
T e 62720

¢ 50

 

 

 

T o°
/o
KF ——— o S 7 B s e LiF
856 350 800 7507 700 OOO / \ [“492 600 700 848

£-490 SEE INSET

Fig. 5. The Bystem KF-LiF-AlFj.
32

UNCI.ASSIFIED
ORNL-DWG 64—-4537

Alfy

TEMPERATURE IN °C
COMPOSITION IN mole Y

 

 

 

 

Fig. 6. The System LiF-NaF-AlFj.
33

UNCLASSIFIED
DRNML-OWG 63-41685

AIF3 1400

 

rky
910

 

Fig. 7. The System KF-ZrF,-AlF,.
10.

1.

12.

13.

14.

REFERENCES

Chem. Tech. Div. Ann. Progr. Rept. May 31, 1963, ORNL-3452, p. Z20.

R. P. Milford, Process Development in the ORNL Fluoride Volatilily
Program October 1067 to Septenber 1963, ORML 1TM-717 (Oct. 25, 1963).

 

 

R. E. Blanco and €. D. Watson, "Heat-¥nd Processes for Solid Fuels,”
Chap. 3, p. 23-106 in Reactor Handbook, ed. by S. M. Stoller and
R. B. Richards, Vol. II, Interscience, New York, 1961.

 

C. M. Slansky, "Preparation of Fuels for Processing,” Chap. 3, p. 75~
124 in Chemical Processing of Reactor Fuels, ed. by J. F. Flagg,
Academic Press, New York, 1961.

 

F. S. Mertin and G. L. Miles, Chemical Processing of Nuclear Fuels,
Academic Press, New York, 1958, p. 99.

 

Fundamental Nuclear Energy Research 1962 - A Special Report, USAEC
(Dec. 1962}, p. 305.

F. L. Culler, personal communication.

Chem. Tech. Div. Ann. Progr. Rept. June 30, 1962, CRNL-3314, p. 39.

 

"Research and Development on Nonagueous Process - Volatility Proc-
esses,” Reactor Fuel Processing, 6, 19 (1963).

 

H. F. Sawyer and P. lowenstein, "Fuel-Element Fabrication Facilities,"”
Appendix B, p. 673-700 in Nuclear Reactor Fuel Elements - Metallurgy
and Fabrication, ed. by A. R. Kaufmann, Interscience, New York, 1962
{See especially Table B-2, p. 690-5).

 

 

¥F. 8. Martin and G. L. Miles, Chemical Processing of Nuclear Iuels,
p. 22=3, Academic Press, New York, 1958,

 

R. L. Boles and R. E. Thoma, Volatility Process Phase Studies -~ A
survey of Molten Fluoride Solvent Mixtures Suitable for Dissolution
of AlF3, ORNL TM-400 (Oct. 22, 1962).

 

 

M. Hamsen, Constitution of Binary Alloys, McGraw-Hill, New York, 1958.

 

Deutschen Chemischem Gesellschaft, "Gmelins Handbuch der anorganlchen
Chemie," system No. 35, Aluminum, Part A, Issue 1, 8th ed., Verlag
Chemie, Berlin, 1934, p. 346-8.

N. M. Volkova and G. V. Gaidukov, Izvest. Sibirsk. Ctdel. Akad. Nauk.
SSSR, 1959, 43; Chem. Abstr. 53, 21333 (1959).

 

A. Glassner, The Thermochemical Properties of the Oxides, Fluorides,
and Chlorides to 25000K, ANL-5750 (1957).

 

 
17,
18.

19.

25.

24,

29.

20.

31.

32‘

[
[8 1§

MBRP Quar. Progr. Rept. Jan. 31, 1959, ORNL~2684, p. 114,

 

B. J. Sturm, "Preparation of Inorgenic Fluorides,” p. 186-7 in Reactor
Chem. Div. Ann. Progr. Rept. Jan. 31, 1960, ORNL-2931.

 

B. J. Sturm, Stannous Fluoride as a Component of Molten-Salt Reactor
Fuels (unpublished work).

 

MSRP Progr. Rept. March 1 to Aug. 31, 1961, ORNL-3215, p. 122.

 

R. E. Thoma et al., "Molten Fluoride Mixtures as Possible Fuel Reproc-
essing Solvents,” p. 257-9 in Reactor Chem. Div. Ann. Progr. Rept.
Jan. 31, 1963, ORNL-3417.

 

T. B. Douglas and J. L. Dever, J. Am. Chem, Soc. 76, 4826-9 (1954).

 

landolt-Bornstein, "Elgenschaften der Materie in IThren Aggregatzastan-
den," Part 4 in Kalorische Auslandsgrossen, Springer, Berlin, 1961,
p. 199,

 

A, B. Trenwith, "Lithium Fluoride,” p. 174-7 in Mellor's Comprehensive
Treatise on Inorganlc and Theoretical Chemistry, Vol. II, Suppl. II,
Longmans, Green, London, 1961.

 

 

A. G. Bergman, A. K. Nesterova, and N. A. Bychkova, "A Visual~Polythermal
Method for the Investigation of Silicate Systems,"” Doklady Akad. Nauk.
SSSR, Khimiya, 101, 483-6 (1955); Chem. Abstr. 49, 15418z (1955).

M. M. Vetyukov and G. 1. Sipriya, "Viscosity of Melts of the Systems
LiF-A1F3 and Na3AlFe-1isAlFg," Zhur, Prikl, Khim. 36, 1905-9 (1963).

 

R. B. Ellis, "Surface Tension of Fused Zinc Chloride," Southeast Re-
glonal Am. Chem. Socilety Meeting, Nov. 14-16, 1963, Program and
Abstracts, p. 87.

H. J. Emeleus, "Nonvolatility Inorganic Fluorides,” pp. 10 and 40 in
Fluorine Chemistry, Vol. I, ed. by J. H. Simons, Academic Press, New
York, 1950Q.

 

L. Brever, "The Fusion and Vaporization Data of the Halides," Paper 7,
p. 193-275 in The Chemistry and Metallurgy of Miscellaneous Materials -~
Thermodynamics, ed. by L. L. Quill, McGraw-Hill, New York, 1950;
Metallurgical Lab. Report CC-3455 (1946).

 

R. K. Steunenberg and R. C. Vogel, "Fluoride and Other Halide Vcla-
tility Processes,” Chap. 6, p. 250-312 in Reactor Handbook, Vol. II,
Fuel Reprocessing, ed. by S. M. Stoller and R. B. Richards, Interscience,
New York, 1961.

 

 

A. F. Wells, Structural Inorganic Chemistry, 3rd ed., Oxford, England,

 

L. H. Ahrens, Geochim Cosmochim Acta, 2, 155-169 (1952).

 
33.

32.

4

i'%‘j

G. W. Gray, Molecular Structure and Properties of Liquid Crystals,
Academle Press; New York, 1962.

G. H. Brown and W. G. Shaw, "The Mesomorphic State,” Chem. Reviews,
57, 1049-1157 (1957). T

. J. Barton et al., "The System NaF-KF-AlF3,"” p. 32 in Phase Diagrams
of Nuclear Reactor Materials, ed. by R. E. Thoma, ORNL-2548 (Nov. 2,
1959,
3\

INTERNAL DISTRIBUTION

ORNL~3594
UC~4 — Chemistry
TID~4500 (31st ed.)

1. Biology Library 78-87. E. H. Guinn
2-%4. Central Research Library 82. C. B. Guthrie
5. Reactor Division Library 89. C. A. Hortoun
6~7. ORNL — Y-12 Technical Library 90. R. W. Horton
Document Reference Section 21. C. E. Larson
8-62. Laboratory Records Department 92. H. F. McDuffie
63. Laboratory Records, ORNI R.C. 93. H. G. MacPherson
64. C. J. Barton 2% . Gleb Mamantov
65. J. BE. Bigelow 95. R. P. Milford
66. C. M. Blood 96. M. J. Skinner
67. G. B. Boyd 97. G. P. Smith
68. M. A. Bredig 98-107. B. J. Sturm
69. J. C. Bresee 108. J. A. Swartout
70. W. H. Carr 109-118. R. E. Thoma
7L. W. L. Carter 119. A. M. Weinberg
72. G. I. Cathers 120. J. P. Young
73. 3. W. Clark 121. L. Brewer (consultant)
74. W. BE. Clark 122. F. Daniels (consultant)
75. F. L. Culler 123. R. W. Dayton (consultant)
76. H. A. Friedman 124, E. A. Mason {(consultant)
7. W. R. Grimes
EXTERNAL DISTRIBUTION
125, Research and Development Division, AEC, ORO
126-717. Given distribution as shown in TID-4500 (31lst ed.) under

Chemistry category (75 copies — OTS)