.
.
.
.
.
.
.
1
.
i
1 OF !
ORNLP
2202

.
in
.
.
.
..
.
.
.
.
30
budet
.
.
.
.
:
11
.
TEETESE
.
.
.
.
1
.
..
=
i
11:25 || 1.4 İLE
.
in
MICROCOPY RESOLUTION TEST CHART
NATIONAL BURF, AU OF STANDAROS - 1963
. .
Conf- 660907-1
"
MASTER 2

CFST: !
ZU
H.C. $
1.
mm
THE SOLUBILITIES OF HF AND DF IN MOLTEN FLUORIDES
.
JUN 27 1966
Paul E. Fiela b
V Department of Chemistry
Virginia Polytechnic Institute
Blacksburg, Virginia 24061
and
James H. Shaffer
Oak Ridge National Laboratory+
Oak Ridge, Tennessee 37830
=
-
2
-
.
-
ET
S
ABSTRACT. The solubilities of HF and DF in molten
LiF-BeF2 · (66-34 mole %) were determined over the temperature
range 500-7000C and at solute gas pressures between 1 and 2
atm. Using previously established experimental methods, the
solubilities of both gases were found to obey Henry's Law.
The Henry's Law constants, Kg (10-4 moles HF/mole melt-atm)
for HF and DF respectively at 500, 600 and 700°C were:
3.37+0.13, 2.96+0.07; 2.16+0.05; 1.83+0.03; and 1.51+0.06,
1.25+0.03; where the values given were obtained by a linear
least squares fit of the experimental data as in X vs. 1/T
and the uncertainties are at the 95% confidence interval.
The heats of solution, AHS, obtained from the least squares
evaluation were -5.98+0.19 and -6.43+0.15 kcal/mole for HF
and DF rest actively. The comparison of AHS for HF in this
melt composition with those obtained previously in melts
ranging from 54 to 89 moi.e % Lir in BeF, reveals linear
dependence of AHS on the mole fraction of Lif above and below
67% with a maximum at 67% LiF. Interpretacion of the iso-
tope effect is made by comparison of the difference in the
entropies of solution between DF and HR with the difference
in the calculated values of the entropies of the two gases at
6000C.
rs
.
1
.
7

LEGAL NOTICE
The report was prepared as an account of Government sponsored work. Neither the United
States, por un Commiuston, por LAT person ucting on behalf of the Commission:
A. Wakes may warranty or representation, expressed or implied, with respect to the accu-
rusy, completeness, or wellness of the information contained la to report, or that the use
of any lalormation, apparatus, method, or proceso disclosed la wis roport may not infringe
pointly owned right; er
B. Assumus uny Habilues with respect to the use of, or for damages resulting from the
.. un of may taformation, appuntes, moltsd, or proctus disclosed in IMS report.
Au urad in the above, "pernco actions on behalf of the Commirdion" includes any em•
.. plogue or contractor of the Commission, or employee of such contractor, to the extent that
much emplogue ur coatractor of the Commission, or employee of such contractor prepares,
... disseminates, or provide soca to, any information portant to his employmeat or coatract
with the Commission, or he employmeal with much contractor.
RELEASED FOR ANNOUNCEMENT
A MIRAR SCIENCE ARSIRACTS
.
.
.
.
-
.
::.
,
;
.
..
2
-
-
.
COM
HN
W
e
in a relat
www
oner til at
eine er los
ON
4
WW14
0
.
.
THE SOLUBILITIES OF HF AND DE IN MOLTEN FLUORIDES
Z
INTRODUCTION
Analysis of the isotope effect on the sol'ibility of HF in
molten fluoride solvents is of particular interest in order to'
estimate the solubility of tritium fluoride in a molten salt
containing Lif. The behavior of tritium fluoride, formed by
neutron lrradiation of lithium in the fluoride mixture, would be
of interest ir: the molten salt reactor concept as well as in the
proposed use of a molten Fluoride breeder blanket for a thermo-
nuclear reactor.
Previous studies of the solubility of HF in molten
fluoride mixtures have shown interesting solvent effects on the ::
solubility of the gas. Shaffer* has shown an empirical corre-
lation of the solubility of HF in melts of NaF-BeF, anů NaF-Zra
to the "free" fluoride ion concentration. The "free" fluoride
concentration was calculated as the difference between the equi-
valent concentration of the complex formed, i.e. Bef, or art,
and the concentration of NaF. A third solvent studied by these
workers was the LiF-BeF, system. Solubility determinations were
made over a temperature range of a few hundred degrees around 700°.
and between 1 and 2 atm. pressure of the solute gas phase in com-
position of 54, 5s, 69, 80 and 89 mole % Lif in BeFg. Whereas
the solubility of HF in the Naf systems showed a strong dependence
: on the concentration of Naf in the melt, the HF solubility in the
LiF system did not. No attempt was made to treat the HF-Lir-Bef,
system by the "free" fluoride correlation.
In the present study an attempt is made to elucidate 'che
solvent characteristics which effect the solubility of HF in molten
fluoride melts containing Lif by investigating the isotope effect
of the solute and dependence of the HF solubility on the Lif
concentration. The solubilities of HF and DF in molten Lif-BeF,
(66-34 mole %) were determined at the temperatures 500, 600, and
700°c and at solute gas saturation pressures of approximately
1.3, 1.6 and 2.0 atm.
wewniartseitserii

*
I.LT
• Vt
.....
.
4
.
ULI
- 3 -
EXPERIMENTAL
Materials.--Anhydrous HF was obtained from cylinders supplied
by Farshaw Chemical Company. The liquid HF had a specified mini-
mum purity of 99.9%. Anhydrous DF was prepared by the Technical
Division, Oak Ridge Gaseous Diffusion Plant a:id has been described
elsewhere.' Both gases were used directly from their cylinders
without further purification.
The fluoride mixture of 66 mole % 'Lif and Bef, was a sample
of the batch production facilities of the Oak Ridge National
Laboratory, The 'Lif used was at least 99.99% pure "Li. The
removal of oxides, sulfides, and structural metal impurities was
accomplished by treating the fluoride melts with a gak mixture of
10 mole % HF in H,. Average chemical analysis of the salt batches
had the following impurities (ppn) : Cr(16), Ni(39), Fe (123),
s(less than 5), oxide removed as water (1650).
Apparatus and Procedure.--A schematic diagram of the apparatus
is shown in Figure 1. Ali materials in contact with the molten
salt or HF were Grade A nickel. Approximately 3.5 kg of molten
salt was used. The general procedure consisted of saturating the
salt in the larger vessel with Hr at constant pressure for 5-6
hours, transferring approx.imately 1 liter under cover of HF into
the smaller vessel and then stripping out the dissolved HF with
He at a flow rate of 8 1/hr. Overnight. The He-HF mixture was
buboled through a 2.00 liter volume of standard KOH having a
bubble path of approximately 80 cm. The number of milliequivalents
of HF per volume of salt transferred was then determined by
titration of aliquots of the KOH solution with standard HCl.
Details of the apparatus and procedure are given in reference 2.
RESU.TS
. The raw solubility data obtained by titration of the standard
KOH solution had the units of moles HF/cc melt. Using the data
of Cantor for the partial molal volumes of LiF and BeF, the solu-
..bilities, Ci were expressed in moles HF, mole melt. The Henry's
en stop
p en met primaid
.
les he
r.com
corte
.
**
*
o elementos para mais il
permet de beste van die ma
y
m
i
..
.
!
.
-
- 4 -
Law constants, Kyse were calculated as C/? where P was the satura-
tion pressure of the HF. Figure 2 shows the Henry's Law plot of
the data taken at 500, 600 and 700°c. Additional data for the HF
solubility were obtained at 524 and 595°c. The slopes of the curves
shown in Figure 2 were calculated by assuming a temperature inde-
pendent heat of solution for both solutes and making linear least
squares calculations of In Ky versus 1/T for all solubility deter-
minations of each of the yases.
The uncertainties in the calculated values of Ky were evaluated
at the 95% confidence level and are listed in Table 1 along with
the respective Henry's Law constants and the total number of solu-
bility determinations, n. It can be seen that the maximum per cent
uncertainty is less than 4% in all cases for K, and well within the
estimated 5% experimental accuracy of previous HF solubility studies.'
t
.7
5
.
.
--
-
-
S
0
DISCUSSION
In order to compare the present results with those obtained
previously for HF in Lif-Bef, mixtures, the Henry's Law values of
Shaffer and Watson were recalculated by the method úf least ·
squares. The five melt compositions studied by them were: 54,
59, 69, 80 and 89 mole % Lif in Berg. Because the variation of
the HF solubility isotherms with Lif concentration were not
simple functions (see Figure 3), a comparison of the heats of
solution vvs made.
The heat of solution, AH®, was calculated from the slope of
the least squares equation according to the van't Hoff relationship..
The calculated slopes are shown along with the experimental data
for both solutes in Figure 4.. 'The values of ApS and Ass and their
uncertainties evaluated as one standard deviation are listed in .
Table 2. The calculation of Ass is discussed below. The plot of
AHS (HF) vs. mole % Lit is shown in Figure 5. The results show an
extremely good linear fit over the two regions of composition
defined by the pure components and the Li,Bef, composition (66,7%).
Intercepts of the curves permit an evaluation of the AFP (HF). in the
pure components. These are -2.35 and +0.68 kcal/mole for Lif and

T01
.
.
22
in
.
SI
'
***
. -
. . .
.
.
-
-
5
-
Bef, respectively. Some reliability can be placed in the extra-
polations of the solubility curves through the 0-50% and 90-100%
regions for two reasons. First, the isotherms must intersect at
the Lif concentration having a zero heat of solution, i.e. 7%
Lir. Secondly, the solubilities in the pure components must have
a temperature dependence consistent with the heats of solutions of
HF in the pure components. It should he noted that the broken
curve segments of the solubility isotherms in Figure 3 are extra-
polations below the liquidus temperature and correspond to the
hypothetical solubility of HF in the supercooled melt. Their
usefulness lies in providing a means to estimate the HF solubility
at higher temperatures.
In the evaluation of the entropy of solution, As®, it is con-
venient to separate the solution process into two steps: (I) the
expansion of the solute gas from the saturation pressure, P, to a
pressure pi corresponding to the concentration, c, of the solute
in solution; and (II) the dissolution of the gas into the solvent :
· at constant concentration:
F9 (2) I HF9(P') -_II_> soln (c)?
If the gas is assumed to be ideal, the separate terms for the
first step are readily evaluated:
AH, = 0,
"AS, - - R In (P'/P) = - R In (CRT/VP) = - R In KRTin
At equilibriun., AGE + AGII = 0
therefore AH-1 = AHS = T(AS, + ASIL),
ASTI = ASS = (ARXT) + R In KyRTm.
The difference in the entropy of solution of DF and HF in 66
mole % Lif in BeF, over the temperature range 500-700°c is 0.86+0.37
cal degnole where the uncertainty is taken as one standard
deviation.
In an attempt to account for the differences in solubility
of DF and HF as an isotope effect, we have calculated the entropy
contributions to the gases at 600°C for comparison with the
and
.
IN
.
.
'N
...
...
.
.
.
.....
.
.....
........
..4
19.
-ies
-
-
1
-
-
.
...
.
***** ..
......
-
Av
.
.
.-
-
entropies cé solution at this temperature. The calculations made
for each solute gas included the translational contribution based
on the Sackur-Tetrode equation, the rigid rotational contribution
and the harmonic vibrational contribution. Ali calculations were
made in the usual manner. Values of the spectroscopic constants
for HF and DF were obtained from the compilation by Kelley and
King. The calculated values are summarized in Table 3.
Although no physical significance can be attached to the solute
state having zero degrees of freedom, it serves as a useful hypo-
thetical state. The comparison of the entropy of solution with
the entropy contributions to the gas is the major point to be con-
sidered in treating the isotope effect. The vibrational contri-
butions and the difference between those of HF and DF are negligible
in magnitude compared to the uncertainty in the determined entropy
of solution and will not be considered. The translational contri-
butions to the gases are the major factors in both solutes but
again the DF-HF difference is almost an order of magnitude smaller
than that for the rotational case. Whereas one would expect a
large decrease in the translational entropy on solution of the
gaseous species which cor.tributes significantly to the magnitude
of the entropy cf solution, the difference between the isotopic
species DF and HF would be small.
The remaining contribution to the entropy of the gas is
rotation. It can be seen that the difference in the rotational
entropy is of the same magnitude as the entropy of solution. The
important consideration here is not that non-rotation of the
solute molecules is unreasonable but whether a rotational entropy
contribution of the dissolved state is due to a torsional oscil-
lation. Considering only the rotational entropies, the entropy
of each solute in the gas phase is given by: .
sy so 'n
S - SO = 10.0,
where sº is the entropy in the hypothetical state and equals zero.
Since the entropy of solution is:
.. . ASS = sm-89 .
.
.
.
.
-
-
. .
.
.
.....
....
1
-
TOT
*
m
-9-
where it is the partial molal entropy of the solute, the difference
in the entropies of solution of DF and HF is;
as op - as = (sme - so - (shee - S.=-0.910.4
= (shop - stes - (sor - 3.d
The second term of the last equation is obtained from above :
g) - 10.0 - 8.7 = 1.39
therefore, ..
. (Sport s = 1.3 - 0.940.4 = 0.
This indicates that within the error of one standard deviation
there is no difference between the rotational contribution of HF
and DF in the melt. Inasmuch as their moments of inertia differ
one can conclude that the solute molecules are neither fredly
rotating nor undergoing torsional oscillations in the dissolved
state. There are two possibilities in which the foregoing con-
clusion is possible. Either the molecule is not rotating or it
is complexed as the bifluoride ion, FHF. In the second case
there would be ro isotope effect associated with the rotational
contribution because of the symmetry of the isotopic bifluoride
ions.
Since the rotational barrier would be of the order of RT, 40
äie: approximately 2 kcal/mole, it might be expected that the
formation of the bifluoride ion is the more reasonable alternative.
Although it is well known that LiHF, decomposes at room tempera-
ture, the corresponding potassium compound has a free energy of
formation in supercooled KF of: -1.8 kcal/mole at 600°c. Further,
support for the stability of FHF is the maximum in the heat of
solution of HF at 0.67 mole fraction Lif. It would be at this
composition of the LiF-BeF, melt that there would be available
maximum concentration of free fluoride ion without excess of
lithium ion, i.e: the Bef, provides a stabilizing effect for the
bifluoride formation.
Further consideration of this proposal and analysis of the
maxima found in the solubility isotherms in terms of the equilibria
-
-
-
-
A
WWW.
11
XN..
IT
:
-
*.
.
- 8 -
of the various components in the solution will be published at a
later date.
ACKNOWLEDGEMENT
The authors would like to thank Prof. D. G. Hill for his
valuable discussions during the course of this work.
L
hoe jim pro
blem
*
programos priemones para matangan terbaru tengan pertama
the
mengome proprio pero por mimpi
iv 1
.!
i
'
Y.
*
9-
REFERENCES
(la) To be presented in part at the 152nd National Meeting of the
American Chemical Society, New York, Sept. 1966; (lb) ORINS
Summer Research Participant, 1965; (1c) Operated for the
United States Atomic Energy Commission by the Union Carbide
Corporation.
(2)
J. H. Shaffer, W. R. Grimes and G. li. Watsoin, Le Phys. Chem.,
63, 1999 (1959).
J. H. Shaefer and G. M. Watson, Reactor Chem. Div, Ann. Prog.
Repty, April 29, 1960, ORNL-2584, p. 31.
J. H. Shaffer, ibid., p. 32.
S. T. Benton, R. L. Farrar, Jr., and R. M. McGill, Prepara-
tiori of Anhydrous Deuterium Fluoride by Direct combination of
the Elements, K-1585, Oak Ridge Gaseous Diffusion Plant,
Jan. 29, 1964.

(6)
S. Cantor, Reactor Chem. Div. Ann._Prog. Rept., March 1966,
ORNL-3913, p. 27.
Since the solubility is expressed in units of moles solute
per mole melt, the units of the solute concentration in th:
gas phase must be moles solute/molal volune of melt, i.e.
nad V. V. fox thi. melt at 600°c is taken as 16.9 cc/mole.°
see for example, G, N. Lewis and M. Randall, "Thermodynamics,"
2nd ed., revised by K. S. Pitzer wind L. Brewer, McGraw-Hill
Book Company, Inc., New York, 1961, Chapter 27.
(9) .K. K. Kelley and E, G, King, Contributions to the Data on
Theoretical Metallurgy. XIV. Entropies of Inorganic Substances,
U. U. Bur. Mines Bull. 592, 1961.
(10) E. :. Moelwyn-Hughes, "Physical Chemistry," 2nd rev. ed.,
Pergamon Press Inc., New York, 1961, p. 694.

"
..
.
- 10 -
Table 1
Calculated values of K, (10%* moles solute/mole melt-atm)
Solute
HF
DF
....................... -
Temperature
500°c
600°c
700°C
.r
3.37+0.13
2.16+0.05
1.51+0.06
2.96+0.07
1.83+0.03
. 1.25+0.03
.
-.
.
.-.....
.......
...
.
..
.
..
.
...
.
.l'any
.
....
.
.
.
.
.
.
.
172
W
!
- 11 -
Table 2
Summary of enthalpy and entropy of
solution calculations at 600°c
HE
DF
3010
-11.888
3237
-12.313
AHS (kcal/mole)
ASS (cal/deg mole)
-5.98+0.19
-20.76+0.22
-6.4340.11
-21.62+0.15
The constants in the equation in K = (m/T) + b were:
obtained by least squares; the uncertainties listed
are taken as one standard deviation.
u
-
u
ot
- 12 -
Table 3
Calculated entropy contributions :f gases at 600°c
HF
DF
DF = HT
Svib
Srot
.
Strans
0.01
8.68
..40.28
48.97
0.08
9.96
40.42
50.46
0.07
1.28
0.14
1.49
$873
-
3
-
-
- W
-
w
*
correcte
-
ya
-
G
4
=
.
.
.
r
*
*.
W
-
LE
5
--
.
.
"
.
L'
IL
.
.
L
-

>
I
02
.
.
.
.
i
-
,
1
-
AND
parents
.
.
WY
.
44::
1
1
L
.
.
.
il.i.:.:.:.:.:::::::
.
1
.
.
1
.
.
.
.
-
KEY: A-HF cylinder in constant temp. bath
B-Surge tank w/ filter
C-Molten salt saturator
D-KOH scrubber
E-Molten sal.t receiver
F-Salt transfer line w/frozen plug
G-Pressure gauges
H=Std. KOH solution
I-Wet test meter
J-Anhydrone/dry ice traps.
Cross-hatched areas-Furnaces and heaters
Figure 1. FF Solubility Apparatus
..
*
-
-
-
-
3.-
Ti
-
v-

ܐܬܝܐܠܐܬܝ ܐܝܐܺܚܪ
Gas solubility (moles HF/mole melt) x 10*
JUILLA1:21IZAI
ITZIELKIMIZI
IIIZYT
TAITZITULLI
DUIVELYY
Experimental DateE+HF,
ITIUIIIIO.DE
Icurvd LaTTINTI.in
Curves: Calculated Tinom
In-K-vs, 41424-by-2.eesti
sauales
TIC
SUICIUD
OO
3.0
1.0
2.0
Saturation pressure, atm. .
.
.
Figure 2. Solubility of HF and DF in Lif-BeF, 166-34 mole %)
:
.
-
.
'
- -
*
Bli
--
----
-
ITه 00
x 10* (moles HF/mole melt-atm)

||
TIT
اا
K
--و800
و.1
.
0.4
0.6
Mole fraction Lif
Figure 3. Solubility of HF as a function of melt
composition in LiF-BeF,.
وا ر
م ا
م
،

IIIIIIIIII
VIINIT
WIIIIIIIIIHIHI
IIIIIIIIIIIIII
IlIIIIIIIMIT
MINUUT
UNIIHIN
QUUMINIAI
OW
Rx 104 (moles HF/mole nelt-atm)
:
...
-
-.
-
sowob
1.0
1.1
1.2
1.3
.
.
odit. W
1/1 x 103. Ox-l ..
Figure $. Semilogarithmic plot of Henry's Law constants
vs. reciprocal absolute temperature for HF and DF in
LiF-BeF, (66-34 mole %).
-
-
-
-
omiro no. comimos-.'..--.
T.
10
•.•.
.

AN
ASS
...

M
-ANS, kcal/mole
1 PM
2
va
.
2
IT:
-
--
-
-
-
-
-
-
.
-
.T
2
2
0.5
0.9
0.7 0.8
Mole fraction Lif
Figure 7. Depenåence of the heat of solution of HF
on melt composition in LiF-BeF,.
-
ZE-
des
ABOVE
UN
.
A
:
E
.
1
WRC
..
.
114
'2EP
.
27
KU
-
*
.
..-
-
-
-
-
- -
-
-
E.
Y
TUA
-
-
M
Mbuloh
'I
.
.
.
.
,
18/31 / 166
DATE FILMED
END










'
1.
1.
1.'
1