Thermal behavior of aluminum fluoride trihydrate

Article abstract of DOI:10.1016/S0040-6031(99)00436-0

The thermal decomposition of aluminum fluoride trihydrate was investigated by TG-DTG. The process was recognized to involve three stages. It was found that the apparent activation energy, E, and the preexponential factor, A, for stage I reaction were 41.9 kJ mol^-1 and 10^2.49 S^-1, respectively. For stage II reaction, the value of E was 47.5 kJ mol^-1 and the most probable differential mechanism function was in the form of (1-α), and magnitude of 80.4 kJ mol^-1 of E and a function in the form of 3/2(1-α)[-1n(1-α)(1/3)] were derived for stage III reaction. From the measurements conducted under increasing temperature conditions, contents of fluorine, C(F)(%), on moisture-free basis and water, C_W(%), on current mass basis contained in the AlF₃·3H₂O sample were determined and correlation between the temperature range dictated by the quality specification of AlF₃, C(F)≥61.0%, C_W<0.5%, and the heating rate were presented. The content of fluorine and water were also determined under isothermal conditions and correlation between the time duration of the treating temperature necessary for the quality specification to be met, and the magnitude of temperature were presented. (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0040-6031(99)00436-0

Thermochimica Acta 352±353 (2000) 47±52
Thermal behavior of aluminum ¯uoride trihydrate
a,*
Xu Delong , Luo Yongqin , Ji Ying , Zhou Longbao , Gai Wenkui
a,1
a
b
b
a
Institute of Powder Engineering, Xi'an University of Architecture & Technology, Xi'an 710055, Shaanxi, PR China
b
Bai Yin Fu Hua Yan Trade Co. LTD, Bai Yin 730901, Gansu, PR China
Received 2 June 1999; received in revised form 20 August 1999; accepted 16 September 1999
Abstract
The thermal decomposition of aluminum ¯uoride trihydrate was investigated by TG-DTG. The process was recognized to
involve three stages:
ꢀ
1
08� 277 C
I
AlF3 Á 3H2O
!
AlF3 Á 0:5H2O ‡ 2:5H2O
ꢀ
2
77� 550 C
II AlF3 Á 0:5H2O
!
AlF3 ‡ 0:5H2O
ꢀ
>380 C
III AlF3 Á 3H2O ! Al2O3 ‡ 6HF
It was found that the apparent activation energy, E, and the preexponential factor, A, for stage I reaction were 41.9 kJ mol
�
1
2
.49 � 1 � 1
and 10 s , respectively. For stage II reaction, the value of E was 47.5 kJ mol and the most probable differential
mechanism function was in the form of (1� a), and magnitude of 80.4 kJ mol of E and a function in the form of 3/
�
1
1
/3
2
conditions, contents of ¯uorine, C (%), on moisture-free basis and water, C (%), on current mass basis contained in the
(1� a)[� ln(1� a) ] were derived for stage III reaction. From the measurements conducted under increasing temperature
F
W
AlF
AlF
3
Á3H
, C
2
O sample were determined and correlation between the temperature range dictated by the quality speci®cation of
<0.5%, and the heating rate were presented. The content of ¯uorine and water were also determined
3
F
ꢁ61.0%, C
W
under isothermal conditions and correlation between the time duration of the treating temperature necessary for the quality
speci®cation to be met, and the magnitude of temperature were presented. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: AlF
3
2
Á3H O; Isothermal; Kinetic parameters; Non-isothermal; TG-DTG; Thermal behavior
1
. Introduction
Anhydrous aluminum ¯uoride is used in the elec-
papers [1±4]. Some of the works treated the thermal
behavior of AlF Á3H O in aqueous solution [1]. Isto-
3
2
min et al. [2,3] reported qualitatively the dehydration
process of AlF Á3H O and relevant inferences, and
trolytic reduction for production of metal aluminum.
In the wet process, the anhydrous ¯uoride is prepared
by heating trihydrated aluminum ¯uoride. Its thermal
decomposition behavior has been reported in previous
3
2
they revealed the rate-limiting stages in the dehydra-
tion process of AlF Á3H O. For preparing AlF from
3
2
3
AlF Á3H O, with speci®ed quality and reduced energy
3
2
consumption, the technical parameters are required,
however it has rarely been reported. This investigation
project intends to reveal the kinetic behaviors of
*
Corresponding author. Fax: ‡86-29-5535724.
1
Fax: ‡86-29-5535724.
0040-6031/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 9 9 ) 0 0 4 3 6 - 0

4
8
X. Delong et al. / Thermochimica Acta 352±353 (2000) 47±52
dehydration of AlF Á3H O and hydrolytic reaction of
3
2
AlF with water and to determine the values of
3
relevant kinetic parameters.
2
. Experimental
2
.1. Materials
AlF Á3H O used in the experiments was for com-
3
2
mercial use, produced by Baiyin Fu Hua Yan Trade
Co. Ltd., Baiyin, Gansu, China. The samples were
dried in an oven at the temperature of 100±1058C for
2 h and sieved through a 120 mesh sieve before test.
MgO used was analytically pure reagent. It was
Fig. 1. TG-DTG curves of AlF
3
Á3H
2
1
O and system AlF
3
2
Á3H O/
�
MgO at a heating rate of 1108C min
.
ignited at 9008C for 2 h before use.
good agreement with the theoretically predicted value
of 6.52%, indicating the anhydrous aluminum ¯uoride
result. As the temperature exceeds 3808C the third-
stage reaction begins. This stage is a hydrolytic reac-
2
.2. Experimental apparatus and conditions
A thermogravimetry analysis system of TGA7 (Per-
kin-Elmer Co., USA) was used to study the thermal
decomposition process. Inert atmosphere was pro-
1
vided by ¯owing N of 60 ml min . Sample mass
tion of AlF with water. On the basis of the experiment
3
carried out and the calculated results, the following
reactions are suggested for de®ning the thermal
decomposition processes of AlF Á3H O.
�
2
of about 6 mg was used in each test. A series of non-
isothermal TG-DTG experiments with heating rates of
5
the program. Another series of isothermal TG-DTG
experiments were carried out at the temperature of
5
3
2
ꢀ
1
08� 277 C
�
1
, 10, 20, 50, 80 and 1308C min were conducted in
AlF3 Á 3H2O
!
AlF3 Á 0:5H2O ‡ 2:5H2O
277� 550ꢀC
AlF Á 0:5H O
!
AlF ‡ 0:5H O
3
2
3
2
ꢀ
>
380 C
00, 550, 600, and 6508C.
2
AlF Á 3H O ! Al O ‡ 6HF "
3
2
2
3
3.2. Kinetic parameters determined from the ®rst-
stage dehydration reaction
3
. Results and discussion
3
.1. Thermal behavior and decomposition process
In order to determine the kinetic parameters that is
the apparent activation energy, E, and preexponential
factor, A, of the ®rst-stage dehydration reaction for the
title compound, a multiple heating method (Kis-
singer's method [4]) was employed. As shown in
�
1
Typical curves, at the heating rate of 108C min
,
for aluminum ¯uoride trihydrate, AlF Á3H O and the
3
2
system AlF Á3H O/MgO are shown in Fig. 1 and the
3
2
data are listed in Table 1. A three-stage mass loss
process is clearly observable from the TG curve for
AlF Á3H O. In the ®rst stage, with the temperature
Table 1
3
2
Mass loss vs. temperature for AlF
3
2
Á3H O
being in the range 108 and 2778C, a mass loss of
3
2.69% is observed. This value is in good agreement
with the calculated one of 32.61% for the loss of 2.5
Stage
Temperature
Mass loss (%)
range (8C)
Exp.
Theory
molecules of water from AlF Á3H O. In the second-
3
2
First-stage
Second-stage
Third-stage
108±277
277±550
>380
32.69
6.96
32.61
6.32
stage, the temperature varied from 277 to 5508C,
meanwhile the sample further decomposed. The value
of mass loss of 6.96% found from this stage is again in

X. Delong et al. / Thermochimica Acta 352±353 (2000) 47±52
49
formulas of g(a), 20 in total, listed in Ref. [6] results
�
1
in the value of E of 47.5 kJ mol and the most prob-
able differential mechanism function, (1� a), by the
method of logical choices.
3.4. Kinetic parameters of the hydrolytic reaction of
AlF with water
3
For de®ning the de¯uorine process in the hydrolytic
reaction of AlF with water in the temperature range
3
3
content, C , on moisture-free basis preserved at the
80ꢂ6158C, Eq. (1) is used to calculate the ¯uorine
F
end of each temperature interval from 3808C to Ti
MAlF Á3H O
MAlF3
Fig. 2. Kissinger's plot of the ®rst-stage dehydration reaction peak
temperature deduced from DTG curves.
A:M:
A
i
3
2
C ˆ 67:88 � ꢃmi
� m †
(1)
Fi
A.M
Where mi
A
and mi are the mass fraction, %, of
2
p
� 1
Fig. 2, the relationship of ln(b/T ) versus T presents
system AlF Á3H O/MgO and AlF Á3H O preserved at
p
3
2
3
2
a straight line, where b is the heating rate, T is the
temperature T , read from Fig. 1, respectively;
i
p
maximum peak temperature applied during the reac-
tion mentioned above. From the experimental data in
Fig. 1, the calculated value of apparent activation
M
and AlF respectively; and the value of 67.88 is the
and MAlF3 are molcule mass of AlF Á3H O
AlF3Á3H2O
3
2
3
¯uorine content in AlF . Data of the de¯uorine in
3
�
1
energy is determined to be 41.9 kJ mol and of the
s . The correlation
dehydrolytic process of AlF with water are shown in
3
2.49 � 1
preexponential factor to be 10
coef®cient is 0.998.
Table 3. The data are determined in the temperature
range 380±6158C, with the de¯uorine fraction being
unity for the whole temperature range.
3
.3. Apparent activation energy and mechanism
By substituting the calculated data in Table 3 into
the formulas of g(a) mentioned above, the value of E
of 80.4 kJ mol and the most probable differential
function of the second-stage reaction for the title
compound
�
1
1/3
mechanism function, 3/2(1� a)[� ln(1� a)
] are
Based on the total mass loss in the temperature
range, 277±5508C, of the second-stage dehydration
reaction of AlF Á0.5H O, the mass loss fraction, a, in
derived by the method of logical choice.
3.5. Relationships between the content of ¯uorine
and water and temperature
3
2
various temperature intervals are calculated by refer-
ring to the TG curve shown in Fig. 1. The calculated
data are shown in Table 2. In accordance with Phadnis'
method [5], substitution of the data of a into the
Contents of water, C , on current mass basis and
W
¯uorine, C , are expressed as relationships versus
F
Table 3
Table 2
Data of the de¯uorine reaction in the hydrolytic process of AlF
with water
3
Mass loss fraction of the second-stage dehydration of AlF
determined from TG curve
3
2
Á0.5H O
Data point
T
i
(K)
a
i
Data point
T
i
(K)
a
i
Data point
T
i
(K)
a
i
Data point
T
i
(K)
a
i
1
2
3
4
5
6
673
707
733
767
793
811
0.01
0.05
0.10
0.20
0.30
0.40
7
8
826
836
851
863
876
0.50
0.60
0.70
0.80
0.90
1
2
3
4
5
562.4
572.5
588.3
604.8
625.1
0.1185
0.1990
0.3016
0.4005
0.5007
6
7
8
9
648.8
680.7
716.4
762.2
0.6000
0.7000
0.8000
0.9000
9
10
11

5
0
X. Delong et al. / Thermochimica Acta 352±353 (2000) 47±52
Fig. 3. (a) C
W
and (b) C
F
vs. T relationships under different heating rate for AlF
3
2
Á3H O.
temperature, T, with heating rate, b, as parameters, see
Fig. 3 and Table 4. Shown in Fig. 3, the values of CF
T ꢄ 308:60 ‡ 134:41 ln b
for C ꢁ 61:0% jrj ˆ 0:9794
F
(3b)
and CW are calculated by using Eqs. (1) and (2)
m Ai :M: � m
B
CW ˆ
 100%
(2)
3.6. Thermal behavior of AlF Á3H O under
A:M:
3
2
mi
isothermal conditions
where m is the ®nal mass fraction of the system AlF
B
3
3
H O/MgO, read at the end of the heating process.
2
Samples of AlF Á3H O were also tested under iso-
3
2
For the aluminum ¯uoride quality speci®cations,
thermal conditions with the temperature regulated to
500, 550, 600 and 6508C respectively. The results are
shown in Fig. 4.
i.e. C <0.5%, and C ꢁ61.0%, to be met by a certain
W
F
heating rate, allowable temperature range is solved
from the equations in Table 4. Values solved are
tabulated in Table 5 and expressed with empirical
equations, as Eqs. (3a) and (3b). From the point of
view of controlling the heating process conditions for
preparing the quali®ed AlF from AlF Á3H O, these
Fig. 5 depicts the contents of ¯uorine and water
varying with time under constant temperature condi-
tion. Table 6 summarizes the relationships of linear
regression between the contents of ¯uorine and water
andthetimeelapsedunderspeci®ctemperature. Table7
presents the time duration required to have the
quality speci®cation to be met. Eqs. (4a) and (4b) give
the regression formulas to predict the required time
range.
3
3
2
data and equations should be of practical importance.
T > 441:89 ‡ 31:12 ln b
for CW < 0:5% jrj ˆ 0:9914
(3a)
Table 4
Relationships of T vs.C
W
and C
F 3 2
under different heating rates for AlF 3H O
�
1
2
2
R
b (8C min
)
C
W
(%)±T (8C)
R
C
F
(%)±T (8C)
5
2
� 2
� 2
� 2
� 2
� 2
� 2
� 4
2
� 2
5
C
W
C
W
C
W
C
W
C
W
C
W
ˆ7.45Â10� T � 8.53Â10 T‡24.47
0.9997
0.9998
0.9944
0.9970
0.9857
0.9960
C
F
C
F
C
F
C
F
C
F
C
F
ˆ� 2.25Â10 T ‡17.22Â10 T‡32.41
0.9869
0.9982
0.9984
0.9992
0.9973
0.9986
�
5
2
� 4 2 � 2
10
20
50
80
ˆ7.65Â10 T � 9.55Â10 T‡29.34
ˆ� 2.40Â10 T ‡19.82Â10 T‡26.50
� 5 2 � 2
�
�
�
5
2
ˆ7.38Â10 T � 9.10Â10 T‡27.99
ˆ� 5.06Â10 T ‡4.04Â10 T‡59.86
� 5 2 � 2
5
2
ˆ3.94Â10 T � 5.36Â10 T‡18.16
ˆ� 3.25Â10 T ‡2.42Â10 T‡63.47
� 5 2 � 2
5
2
ˆ4.60Â10 T � 6.32Â10 T‡21.70
ˆ� 3.34Â10 T ‡3.09Â10 T‡60.55
� 5 2 � 2
5
2
1
30
ˆ3.48Â10� T � 4.90Â10 T‡17.31
ˆ� 2.48Â10 T ‡2.01Â10 T‡63.86

X. Delong et al. / Thermochimica Acta 352±353 (2000) 47±52
51
Table 5
Allowable temperature ranges for the quality speci®cation to be
met
�
1
b (8C min
)
T (8C) for C
W
<0.5%
T (8C) for C
F
ꢁ61.0%
5
>495.3
>510.5
>529.1
>570.5
>581.5
>588.9
ꢄ521.8
ꢄ575.9
ꢄ769.1
ꢄ838.5
ꢄ910.3
ꢄ933.5
10
20
50
80
1
30
Fig. 4. Typical TG curve of AlF
3
2
Á3H O at 5008C.
Fig. 5. `C
F
±t' and `C
W
±t' relationships of AlF
3
2
Á3H O at (a) 5008C; (b) 5508C; (c) 6008C; (d) 6508C.
Table 6
±t' and `C
`
C
F
W
±t' relationships of AlF
3
2
Á3H O under isothermal condition
T (8C)
C
F
(%)±t (min)
|r|
C
W
(%)±t (min)
|r |
�
� 2
2
5
00
50
C
C
F
ˆ66.97� 9.73t (C
F
: 60.0±46.5)
0.9909
0.9924
C
C
W
ˆ1.53� 40.96Â10 lnt (C
W
: 0.10±1.20)
: 0.10±0.70)
0.9979
0.9947
0.9999
0.9974
0.9999
0.9849
0.9998
ˆ67.66� 17.64Â10� t (C
ˆ68.86� 25.12Â10� t (C
ˆ68.69� 91.78Â10� t (C
2
ˆ1.11� 29.73Â10 lnt (C
5
F
F
: 60.0±67.0)
F
: 60.0±67.0)
F
: 50.0±66.0)
W
W
3
.76� 0.82t (C
W
: 0.70±1.50)
� 2
2
6
6
00
50
C
F
F
0.9992
0.9959
C
W
ˆ0.65� 2.48Â10 t (C
W
: 0.05±0.55)
2
.11� 0.32t (C
W
: 0.55±1.20)
� 2
2
C
C
W
ˆ0.58� 1.83Â10 t (C
W
: 0.20±0.55)
2.61� 0.55t (C
W
: 0.55±0.80)

5
2
X. Delong et al. / Thermochimica Acta 352±353 (2000) 47±52
1
/3
Table 7
in the form of 3/2(1� a)[� ln(1� a) ] and the
The isothermal time necessary for the quality speci®cations to be
met by constant temperature treatment
� 1
value of E is 80.4 kJ mol
.
For the reason that the activation energy for the
second-stage is significantly less than that for the
third-stage it can be seen that the dehydration rate
should be greater than the hydrolytic reaction rate.
However, as the temperature further increases, the
latter will increase more rapidly than the former.
Based on the TG-DTG measurements the technical
conditions for heating AlF Á3H O to the qualified
T (8C)
t (min)
for C
W
<0.5%
for C
F
ꢁ61.0%
500
550
600
650
>12.36
>7.78
>6.05
>4.37
ꢄ61.36
ꢄ37.75
ꢄ31.29
ꢄ8.38
3
2
AlF have been recommended.
3
t > expꢃ5:84 � 6:74  10� 3T† for
CW < 0:5% jrj ˆ 0:9923
t ꢄ 224:91 � 0:33T for
CF ꢁ 61:0% jrj ˆ 0:9799
Measurements indicated that greater heating rate
(4a)
(4b)
was beneficial to dehydration of AlF Á3H O and
3
2
manifested a relative suppression of hydrolytic
defluorination. Higher degree of constant tempera-
ture resulted in more efficient dehydration of the
AlF Á3H O and more detrimental to preservation
3
2
of fluorine.
4
. Conclusions
Three stages have been recognized in the thermal
decomposition of AlF Á3H O.
References
3
2
For the first-stage dehydration, the values of E and
A have been determined by experiments to be
[
[
1] M. Grobelny, J. Fluorine Chem. 9 (1977) 441±448.
2] S.P. Istomin, A.S. Galkov, L.A. Ivantsov, L.S. Babkina, Tsvet.
Met. 12 (1978) 39±41 (in Russian).
�
1
1.9 kJ mol and 10 s respectively.
2.49 � 1
4
The value of E and the most probable differential
[
3] S.P. Istomin, A.P. Istomin, A.V. Kyun, Zh. Prikl, Khim
(
Leiningrad) 54(3) (1981) 517±521 (in Russian).
mechanism function for the second-stage dehydra-
�
[
[
4] Lu Zhenghua, Qing Jinshu 11 (1983) 22±25 (in Chinese).
5] H.E. Kissinger, Anal. Chem. 29 (1957) 1072.
1
tion are 47.5 kJ mol and (1� a) respectively.
For the hydrolytic reaction of AlF , with water, the
3
[6] A.B. Phadnis, V.V. Deshpande, Thermochim. Acta 62 (1983)
361±367.
most probable differential mechanism function is