Calorimetric study of thermal decomposition of lithium hexafluorophosphate

Article abstract of DOI:10.1023/A:1025125306291

Enthalpy of formation of lithium hexafluorophosphate was calculated based on the differential scanning calorimetry study of heat capacity and thermal decomposition. It was found that thermal decomposition of LiPF₆ proceeds at normal pressure in

Full text of DOI:10.1023/A:1025125306291

Journal of Thermal Analysis and Calorimetry, Vol. 73 (2003) 71–83
CALORIMETRIC STUDY OF THERMAL
DECOMPOSITION OF LITHIUM
HEXAFLUOROPHOSPHATE
1
K. S. Gavritchev , G. A. Sharpataya , A. A. Smagin , E. N. Malyi
and V. A. Matyukha
1
*
1
2
2
2
Kurnakov Institute of General and Inorganic Chemistry, Leninsky pr.31, Moscow, 119991, Russia
Federal Unitary State Enterprise ‘Siberian Group of Chemical Enterprises’ (SGChE),
Kurchatov str.1, Seversk, Tomsk Region, 636000, Russia
2
(
Received May 14, 2002; in revised form November 20, 2002)
Abstract
Enthalpy of formation of lithium hexafluorophosphate was calculated based on the dif-
ferential scanning calorimetry study of heat capacity and thermal decomposition. It was
found that thermal decomposition of LiPF proceeds at normal pressure in the tempera-
6
ture range 450–550 K. Enthalpy of LiPF decomposition is ∆ H(LiPF , c, 298.15 K)=
6
d
6
–
1
8
4.27±1.34 kJ mole . Enthalpy of formation of lithium hexafluorophosphate from ele-
0
–1
ments in standard state is ∆ H (LiPF , c, 298.15 K) = –2296±3 kJ mol .
f
6
Keywords: DSC, enthalpy of formation, heat capacity, thermal decomposition
Introduction
Study of thermodynamic properties of lithium hexafluorophosphate is governed by
its application in the process of lithium batteries and accumulator production. Avail-
able in literature data on the results of experimental study of formation enthalpy of al-
kali metal hexafluorophosphates contain only ∆ H values for sodium, potassium, ru-
f
bidium and cesium salts [1]. No information about the study of LiPF enthalpy of
formation was found in the literature. Calculation of LiPF formation enthalpy based
6
6
on the known models gives a very rough quantity results.
In this work the enthalpy of LiPF formation was calculated based on experi-
6
mental value of enthalpy of decomposition of LiPF studied by differential scanning
6
calorimetry (DSC) as DSC is the effective tool for the study of decomposition with
the formation of gaseous products [2].
*
Author for correspondence: E-mail: gavrich@igic.ras.ru
1
388–6150/2003/ $ 20.00
Akadémiai Kiadó, Budapest
©
2003 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

7
2
GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
Experimental
Sample
Crystalline lithium hexafluorophosphate was synthesized at the SGChE by the direct
reaction of lithium fluoride with phosphoric pentafluoride in the HF medium. Analy-
sis of prepared sample showed that it contains 99.7–99.8 wt.% LiPF . Content of im-
6
purities is: LiF–(0.1÷0.2 wt.%), HF–(0.02÷0.07 wt.%), H O–(0.001÷0.009 wt.%).
2
Molecular mass of LiPF is equal to 151.905.
6
Method
The thermal decomposition of lithium hexafluorophosphate has been studied using a
SETARAM DSC 111 instrument (temperature reproducibility, ±0.1 K; uncertainty in
the temperature scale, ±0.5 K; DSC measuring sensitivity, 15 µW). At temperatures
below 500 K the error of thermal effects determination is within 1%.
Two different experimental ways were used to study the decomposition of lith-
(c) → LiF(c) + PF (g):
5
ium hexafluorophosphate according to the reaction LiPF
6
a) in hermetically sealed calorimetric crucible
b) in calorimetric crucible with the remove of gaseous products of decomposition.
Because the gaseous phosphoric pentafluoride forms at the decomposition of LiPF
,
6
these two ways differ radically. In the first case the thermal effect is determined at con-
stant volume and variable (increasing) pressure, i.e. measured heat of reaction relates to
the internal energy change ∆ U(T). In the second case pressure is near to constant and
r
equal to standard conditions, so the enthalpy of reaction ∆ H(T) is measured.
r
To carry out the experiments we used stainless steel calorimetric crucibles with the
nickel seal. Schemes of two used types of calorimetric crucibles are given in Fig. 1.
Fig. 1 Schemes of stainless steel calorimetric crucibles used in the study of thermal de-
6
composition of LiPF A – sealed crucible, B – crucible with remove of gaseous
products of decomposition. 1 – nickel seal, 2 – cap, 3 - crucible body, 4 - sam-
ple, 5 – tube for gas remove, 6 – heat buffer, 7 – tube for argon blowing, 8 – cru-
cible holder, 9 – argon input, 10 – gaseous products output
J. Therm. Anal. Cal., 73, 2003

GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
73
Testing experiment in the temperature range 300–600 K showed that no reaction
of lithium hexafluorophosphate and decomposition products with crucible occurs.
Experimental setup based on SETARAM DSC 111 was built up to carry out the study
of LiPF decomposition at normal pressure. To pump out all formed gaseous products
6
the system was evacuated to bring the pressure a little down (200 Pa) the normal
value. Scheme of experimental setup is plotted in Fig. 2.
Fig. 2 Scheme of experimental setup. 1 – argon input, 2 – DSC 111 measuring block,
3
–
– argon output, 4 – gaseous products output, 5 – cartridge with active carbon, 6
liquid nitrogen trap, 7 – liquid nitrogen trap with active carbon, 8 – pump
Since lithium hexafluorophosphate is extremely hygroscopic substance, all manipu-
lations with sample (including filling of calorimetric crucibles, preliminary and interme-
diate operations at crucible weighing before and after LiPF decomposition) were carried
6
out in glove box with dried nitrogen atmosphere. Tightness of sealed calorimetric cruci-
bles was controlled by weighing during the day after sealing a calorimetric crucible.
Values of thermal effects (∆ U or ∆ H) were calculated by graphic integration of
r
r
peaks in the DSC curves using the SETARAM software. The error of integration var-
ied from 0.1 to 3% depending on peak shape and area. Temperatures of thermal ef-
fects were determined as extrapolated peak onset temperature T and as maximum
e
peak temperature Tp.
Results and discussion
Study of LiPF decomposition in sealed crucible
6
Thermal behavior of lithium hexafluorophosphate in sealed stainless steel crucible
3
V=0.15 cm ) was studied in the temperature range 298–513 K. Stainless steel cruci-
(
bles with the nickel seal can stand a pressure up to 35 MPa at 573 K.
DSC curves were recorded following the regime: heating in the range 298–513 K
–
1
–1
rate 2 K min ) – isothermal stage (10 min) - cooling from 513 to 343 K (rate 2 K min ).
(
Results of decomposition investigation of LiPF samples for four different batches
6
are presented in Table 1. The heating-cooling cycle for sample 3 is shown in Fig. 3.
From Table 1 one can see that decomposition of LiPF in sealed crucible pro-
6
ceeds in the temperature range 410(425)–480 K. The endothermic peaks were fixed at
T = 467.2±0.1 K and T = 470.1±0.2 K (from four runs). Heating up to 550 K showed
e
p
that no thermal effects appear above the decomposition event.
Inasmuch evolved gaseous phosphoric pentafluoride formed at LiPF decompo-
6
sition does not remove from the reaction zone, the exothermic effect connected with
the reverse reaction is observed at cooling in the range 468–420 K with T =
e
4
62.4±0.1 K and T = 459.8±0.3 K (from four runs).
p
J. Therm. Anal. Cal., 73, 2003

Table 1 Parameters of thermal effects in DSC curves for lithium hexafluorophosphate (sealed calorimetric crucibles)
Endothermic effect in heating curve Exothermic effect in cooling curve
No., mass
scanning rate
Integrating
range/ K
–1
Integrating
range/ K
–1
T
e
/ K
T
p
/ K
∆
r
U/ J g
e
T / K
T
p
/ K
r
– ∆ U/ J g
No. 1, 0.0722 g
2
–
1
408–483
408–483
423–482
423–482
467.2
466.9
467.5
467.2
469.8
470.0
470.4
470.2
25.92±0.06
20.62±0.07
20.41±0.15
19.67±0.06
468–418
468–418
468–413
468–413
462.5
462.5
462.4
462.3
460.1
460.1
459.5
459.5
17.70±0.31
18.73±0.94
20.06±0.04
19.66±0.01
K min
No. 2, 0.0856 g
2
–
1
K min
No. 3 0.0979 g
2
–
1
K min
No. 4, 0.0948 g
2
–
1
K min

GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
75
6
Fig. 3 DSC curves of heating and cooling for sample 3 LiPF in sealed crucible
Values of thermal effects of forward (decomposition) and reverse (formation)
reactions for LiPF are very near (Table 1). It confirms that decomposition process in
6
the sealed crucible is suppressed by the pressure influence of forming gaseous PF5
and that the reverse reaction of lithium hexafluorophosphate synthesis from lithium
fluoride and phosphoric pentafluoride takes place at cooling of sample in the sealed
crucible (Fig. 3). It is necessary to note that values of endothermic effect presented in
Table 1 can not be attributed to heat of decomposition of the whole of the specimen.
Study of LiPF decomposition in crucible with the remove of gaseous products of
6
decomposition
DSC curves of two samples of LiPF were recorded by discrete «heating-cooling» cy-
6
cles. Measurements for each sample were carried out by series consisting of three cy-
cles. Results of study of LiPF decomposition heat are given in Tables 2, 3 and Figs 4,
6
5
. Values of thermal effects in Tables are presented as well in mJ as in J per gram of
initial sample mass. The last value is more convenient for comparison of the results
obtained in consecutive cycles.
Sample LiPF (1)
6
The initial study of sample 1(cycle 1) was performed following the program used ear-
lier for study of LiPF decomposition in hermetically sealed stainless steel calorimet-
6
ric crucible (mass 0.1602(5) g, series 1).
Cycle 1 (Fig. 4a): heating in the range 297–548 K (rate 2 K min ) – isothermal
–
1
–
1
stage (10 min) - cooling from 548 to 323 K (rate 2 K min ).
Weighing the sample 1 after the cycle 1 showed that it is heavier (0.04741 g)
than theoretical value of LiF mass (0.0274 g) for the full decomposition of LiPF -1. It
6
can be considered as the incomplete decomposition of LiPF sample. From other side,
6
J. Therm. Anal. Cal., 73, 2003

Table 2 Parameters of endothermic effects in the heating curves of lithium hexafluorophosphate (crucible with gas remove)
Peak I
Peak II
/ K
Sample LiPF
Peak III
/ K
-1; mass 0.1602 g (series 1)
Peak IV
r
∆ H
Cycle,
Scanning
rate
Integrating
range/ K
–1
J g
T
e
/ K
T
p
/ K
T
p
T
e
T
p
/ K
T
p
/ K
mJ
6
Cycle 1
2
–
–
–
1
467.3
465.9
467.4
470.5
473.8
472.9
489.3
491.3
486.2
486.5
499.6
503.0
499.3
520.5
457–544
431–543
430–538
73492±97
458.61±0.60
68.84±0.65
K min
Cycle 2
1
1
11032±105
5
K min
Cycle 3
K min
1624±11
10.13±0.07
5
Total effect of three heatings
-2; mass 0.1221 g (series 2)
86148±213
537.58±1.32
Sample LiPF
6
Cycle 1
–
–
–
1
1
1
466.2
472.8
522.0, 532.7
433–576
59331±30
485.92±0.24
43.01±0.86
5
K min
Cycle 2
K min
Cycle 3
K min
477.4
477.1
490.9
491.2
405–532
408–518
5251.2±105.5
5
1587.4±24.7
13.00±0.20
5
Total effect of three heatings
66170±160
541.93±1.30

Table 3 Parameters of exothermal effects in cooling curves of lithium hexafluorophosphate (crucibles with gas remove)
–
r
∆ H
Cycle,
scanning rate
e
T / K
T
p
/ K
LiPF
–1
Integrating range/ K
-1; mass 0.1602 g (series 1)
mJ
J g
6
–
–
–
1
1
1
Cycle 1, 2 K min
Cycle 2, 5 K min
Cycle 3, 5 K min
425.4
425.9
417.3
420.5
413.0
406.8
442–393
454–386
421–397
3657±17
22.82±0.11
5.509±0.107
0.300±0.002
28.63±0.22
882.9±17.1
48.04±0.34
4587.9±34.4
Total effect of three coolings
LiPF -2; mass 0.1221 g (series 2)
6
–
–
–
1
1
1
Cycle 1, 5 K min
Cycle 2, 5 K min
Cycle 3, 5 K min
411.5
413.1
424.0
403.9
399.1
397.8
435–367
437–360
432–367
3014.9±8.5
1189.3±28.0
369.5±13.0
4573.7±49.5
24.69±0.07
9.74±0.23
3.03±0.11
37.46±0.41
Total effect of three coolings

7
8
GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
6
Fig. 4 DSC curves of heating and cooling of sample LiPF -1: a – cycle 1, b – cycle 2, c – cycle 3
at sample cooling the exothermic effect in DSC curve was observed (Fig. 4a), which
can be assumed as the reverse reaction of formed LiF with gaseous PF remained in
5
crucible volume. Lithium fluoride has no phase transition in this temperature region.
In four days the cycle 2 for sample 1 (Fig. 4b) was conducted (sample mass is
–
.04741 g): heating from 293 to 573 K (rate 5 K min ) – isothermal stage (5 min) –
1
0
cooling from 573 to 323 K (rate 5 K min ).
–
1
Just after this study the cycle 3 (Fig. 4c) was carried out following the program
described for the cycle 2. Mass of sample after decomposition in the cycles 2 and 3
(
0.0285 g) was 4% more than theoretical value of LiF mass (0.0274 g).
As one can see from Table 2 and Fig. 4a, at the first heating (cycle 1) with rate 2
–
1
K min the decomposition in the temperature range 460–542 K takes place. The total
endothermic effect consists of four non–resolved peaks I–IV. Peaks I–III relate to the
ascending peak slope. Peak I temperature T = 467.3 K can be considered as the de-
e
J. Therm. Anal. Cal., 73, 2003

GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
79
6
Fig. 5 DSC curves of heating and cooling of sample LiPF -2: a – cycle 1, b – cycle 2; c – cycle 3
composition onset. The highest peak III was observed at T = 499.6 K, and the main
p
peak IV at T = 520.5 K.
p
In the cycles 2 and 3 peak III is the main (Table 2 and Figs 4b, 4c) and peak I
corresponds to the onset of substance decomposition.
The heat of endothermic effect of cycle 1 is 458.6 J g , while the total heat of
–
1
–
1
three cycles is 537.6 J g (Table 2). In the first cycle 85% of total decomposition heat
was released. Calculation of mass balance of LiPF decomposition by equation (1–x)
6
LiPF (c) → x LiF (c) + PF (g) gives the sample mass 0.04729 g after the decomposi-
6
5
tion of 85% of sample mass which is near to experimental value 0.04741 g.
Sample LiPF (2)
6
Study of the second sample (mass 0.1221 g, series 2) was carried out in three identical
cycles one after another without the intermediate sample weighing: heating from 293
J. Therm. Anal. Cal., 73, 2003

8
0
GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
–
to 573 K (rate 5 K min ) – isothermal stage (5 min) – cooling from 573 to 323 K (rate
1
–
1
5
K min ).
At the first heating with scanning rate 5 K min (Table 2 and Fig. 5a) decompo-
sition proceeds in the range 440–570 K. Two peaks were observed in the DSC curve
–
1
of first heating (cycle 1: peak I (T = 466.2 K) and non-resolved peak IV (T = 522.0
e
p
and 532.7 K). Peaks II (Table 2 and Figs 5b, 5c) which relate to peak II of endother-
mic effect in heating curve of sample LiPF -1 (cycle 1) were observed at the second
6
(
cycle 2) and third (cycle 3) heatings. Mass of sample 2 after cycles 1–3 was 0.0240
g. This value is 15% more than theoretical one (0.02085 g) for LiF.
Values of endo- and exothermic effects in DSC curves at consecutive heat-
ing/cooling cycles decrease sequentially (Tables 2, 3).
Fig. 6 Comparison of endothermic effects in DSC curves at heating with different
scanning rates: 2 (1) and 5 (2) K min
–1
DSC curves for initial heating of samples 1 and 2 at different rates (2 and 5 K
–
1
min ) are presented in Fig. 6. One can see that position of peaks I (which relate to de-
composition onset) is unchanged.
It is interesting to compare DSC curves of LiPF decomposition in hermetically
6
sealed crucible and crucible with removing of gaseous products of reaction. Both
–
curves were recorded at the same scanning rate – 2 K min . Heating–cooling curves
1
of LiPF decomposition in different crucibles are presented in Fig. 7. One can see that
6
the sole peak in DSC heating curve for sample studied in sealed crucible (LiPF sam-
6
ple 3 from Table 1) coincides with peak I of sample LiPF -1 as for temperature as for
6
heat effect taking into account the sample masses.
So, based on the results of cooling in cycle 1 of LiPF -1, the endothermic peak in
6
the heating curve of lithium hexafluorophosphate studied in sealed crucible (Fig. 3)
may be identified only with the decomposition of LiPF and not with the structural
phase transition in this substance as no thermal effect at temperature 459.5 K (tem-
6
J. Therm. Anal. Cal., 73, 2003

GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
81
Fig. 7 Comparison of DSC curves of heating and cooling of LiPF
sealed crucible (1) and crucible with gas remove (2) for the same heating rate
6
decomposition in
–
2 K min )
1
(
perature of reverse peak for LiPF sample 3 from Table 1) was found although few
6
percents of LiPF mass was not decomposed in this run (sample LiPF -1; series 1).
6
6
Heat capacity of LiPF in the temperature range 298–470 K
6
To derive the reaction enthalpy to the standard temperature 298.15 K it is necessary
to know the heat capacity of initial substances and final products of reaction.
Heat capacity measurements of LiPF were carried out at continuous heating in
6
–
1
the temperature range 298–450 K with the rate 2 K min . The sample mass was
Table 4 Smoothed heat capacity and enthalpy change of LiPF
6
0
0
0
H (T) – H (298.15 K)/
0
0
H (T) – H (298.15 K)/
0
C (T)/
–
C (T)/
–1
J K mol
p
p
T/ K
T/ K
1
J K mol
–1
–1
–1
–1
J mol
J mol
2
3
3
3
3
3
3
3
3
3
98.15
151.1
151.9
155.8
159.6
163.2
166.7
170.1
173.3
176.4
179.4
0
390
400
410
420
430
440
450
460
470
182.2
184.8
187.3
189.6
191.7
193.6
195.2
196.6
197.7
15420
17260
19120
21000
22910
24830
26780
28740
30710
00
10
20
30
40
50
60
70
80
303.0
1841
3418
5033
6682
8367
10080
11830
13610
J. Therm. Anal. Cal., 73, 2003

8
2
GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
0.2204 g (almost full fill of crucible by powder sample). It was found that heat capac-
ity anomaly of lithium hexafluorophosphate decomposition appears at 410 K. Thus,
the regular heat capacity at decomposition temperature (T = 467 K) can be evaluated
e
by extrapolation of regular heat capacity determined for the region 298–410 K.
Heat capacity data in the temperature region 307–410 K were approximated by
–
2
2
3
4
the Meier-Kelly type equation C
(T) = a+bT – cT +dT +eT +fT , where the equa-
p
–
4
tion coefficients are a = 113.9758; b = 0.2213438; c = 2122892; d = –5.352194⋅10 ;
–
6
–9
e = 2.600709⋅10 ; f = –3.328746⋅10 .
Scatter of experimental data was less than 0.5%. Smoothed heat capacity values
0
C (T) and the enthalpy change H (T) – H (298.15 K) are given in Table 4.
0
0
p
One must be constantly bear in mind that heat capacity data studied by DSC may
differ from studied in equilibrium state. But in this study we are interested not in real
heat capacity values but in the values of enthalpy increment from reaction tempera-
ture to standard one.
Calculation of enthalpy of decomposition and enthalpy of
formation
Decomposition enthalpy of lithium hexafluorophosphate was estimated based on the ex-
perimental DSC data (Tables 2, 3). In calculating scheme the enthalpy of decomposition
will be equal to the sum of endothermic effects in all three cycles minus the sum of all
exothermic effects. As it was written above, the thermal events in the cooling curves were
attributed to the reverse reaction of LiPF formation from LiF and PF .
6 5
The preliminary calculation of decomposition enthalpy using the initial sample
masses gives the following results (results of intermediate calculations and errors are
given with the excess of significant figures):
–
1
Sample 1: ∆ H = 77.301±0.234 kJ mole
d
–
1
Sample 2: ∆ H = 76.632±0.260 kJ mole .
d
As well as the sample mass after decomposition differs from theoretical LiF
mass value, the correction of decomposition enthalpy was made:
–
1
Sample 1: ∆ H = 77.855±0.234 kJ mole
d
–
1
Sample 2: ∆ H = 78.661±0.260 kJ mole .
d
The average value of decomposition enthalpy at temperature 467 K (taken as the
decomposition temperature) is
–
1
∆ H (LiPF , c, 467 K) = 78.258±0.494 kJ mole .
d 6
–
1
Table 5 Enthalpy change of reaction products and initial substance/ kJ mol
∆
H (LiF, c) [2]
.723
∆H (PF
5
, g) [3]
∆H (LiPF
6
, c)
∆H
7
16.386
30.117
- 6.008
J. Therm. Anal. Cal., 73, 2003

GAVRITCHEV et al.: DECOMPOSITION OF LITHIUM HEXAFLUOROPHOSPHATE
83
Kirchhoff equation was used to derive the enthalpy value to standard conditions
∆dH(298.15 K) = ∆ H(467 K) – ∆H, where ∆H is the sum of enthalpy change of reac-
d
tion products minus enthalpy change of initial substance in the temperature range
2
98.15–467 K (Table 5).
Decomposition enthalpy at 298.15 K is ∆ H(LiPF , c, 298.15 K) = 84.266 kJ
d
6
–
1
fluor
mole . And, so, enthalpy of LiPF formation from fluorides is: ∆ H
98.15 K) = –84.27±1.34 kJ mole (the error of this value includes the errors of inte-
(LiPF , c,
6
6
f
–
1
2
grating and determination of thermal effects).
Now we are able to calculate the enthalpy of lithium hexafluorophosphate for-
mation from elements in standard state using the known values of formation enthalpy
for substances in the reaction: LiF (c) + PF (g) = LiPF (c):
5 6
fluor
∆fH(LiPF , c, 298.15) = ∆ H(LiF, c, 298.15 K) + ∆ H(PF , g, 298.15 K) + ∆ H
6
f
f
5
f
(
LiPF , c, 298.15 K).
Enthalpy values for PF and LiF were quoted from [3, 4]: ∆ H(LiF, c, 298.15 K)
6
5
f
–
1
–1
=
nally, the standard enthalpy of LiPF formation from elements in standard state is
∆fH (LiPF , c, 298.15 K) = –2296±3 kJ mol . This value adequately correlates with
–618.3±0.7 kJ mole , ∆ H(PF , g, 298.15 K) = –1593.3±1.3 kJ mole . And, fi-
f
5
6
0
–1
6
the results reported in [1] for other alkali metal hexafluorophosphates.
Conclusions
DSC study of LiPF decomposition showed that it proceeds in the temperature range
6
4
50–550 K at normal pressure. It was shown that decomposition of lithium
hexafluorophosphate was sufficiently suppressed by pressure increase of gaseous
PF . Enthalpy of formation of lithium hexafluorophosphate was calculated based on
5
the DSC study of heat capacity and enthalpy of thermal decomposition.
References
1
2
3
T. C. Ehlert and M.-M. Hsia, J. Chem. Eng. Data, 17 (1972) 18.
B. Ptaszynski, E. Skiba and J. Krystek, J. Therm. Anal. Cal., 65 (2001) 231.
Thermodynamic properties of individual substances. A Handbook. Eds. V. P. Glushko,
L. V. Gurvich et al. Nauka, Moscow, 1982, v.4.
4
Thermodynamic properties of individual substances. A Handbook. Eds. V. P. Glushko,
L. V. Gurvich et al. Nauka, Moscow, 1978, v.1.
J. Therm. Anal. Cal., 73, 2003