Emanation thermal analysis study of synthetic gibbsite: Evaluation of experimental ETA results by mathematical modelling

Article abstract of DOI:10.1023/A:1023313907122

Emanation thermal analysis (ETA) was used for thermal characterization of microstructure changes taking place during heating of synthetic gibbsite sample in argon in the range of 25-1200°C. Microstructure development and the increase of the surface area u

Full text of DOI:10.1023/A:1023313907122

Journal of Thermal Analysis and Calorimetry, Vol. 71 (2003) 773–782
EMANATION THERMAL ANALYSIS STUDY OF
SYNTHETIC GIBBSITE
Evaluation of experimental ETA results by mathematical
modelling
V. Balek1*, J. Šubrt 1, J. Rouquerol 2, P. Llewellyn2, V. Zeleòák3,
I. M. Bountsewa4, I. N. Beckman4 and K. Györyová3
1Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, 250 68 Øe ,
Czech Republic
2MADIREL, UMR CNRS - Provence University, Centre de St Jérôme, 13397 Marseille,
Cedex 20, France
3Institute of Chemical Sciences, Department of Inorganic Chemistry, Faculty of Sciences,
P. J. Šafárik University, 041 54 Košice, Slovak Republic
4Chemical Faculty, Moscow State University, 119899 Moscow, GSP-3, Russia
Abstract
Emanation thermal analysis (ETA) was used for thermal characterization of microstructure changes
taking place during heating of synthetic gibbsite sample in argon in the range of 25–1200°C.
Microstructure development and the increase of the surface area under in-situ conditions of the sam-
ple heating were characterized. The increase of the radon release rate from 130–330°C monitored
the increase of the surface area due to the dehydration of Al(OH)3. During heating of the sample in
the range 450–1080°C the ETA results characterized the annealing of surface and near surface struc-
ture irregularities of intermediate products of gibbsite heat treatment. The mathematical model for
the evaluation of the ETA experimental results was proposed. From the comparison of the experi-
mental ETA results with the model curves it followed that the model is suitable for the quantitative
characterization of microstructure changes taking place on heating of gibbsite sample.
Keywords: ETA, gibbsite, mathematical modelling, microstructure changes
Introduction
Thermal behaviour of aluminium hydroxide, called gibbsite, was investigated both as
natural and synthetic mineral by a number of authors [1–13]. Thermogravimetry
(TG), differential thermal analysis (DTA), XRD, IR-spectroscopy and other tech-
niques were used to elucidate the mechanism of the thermal behaviour of gibbsite un-
der different conditions including the effect of mechanochemical activation [14–18].
It was shown by Rouquerol et al. [10, 19], by using a form of controlled rate thermal
analysis (CRTA) which we now call controlled rate evolved gas detection (CR-EGD),
*
Author for correspondence: E-mail: bal@ujv.cz
1388–6150/2003/ $ 20.00
Akadémiai Kiadó, Budapest
© 2003 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

774
BALEK et al.: SYNTHETIC GIBBSITE
that microporous alumina with tailored porosity can be prepared by thermal decomposi-
tion of gibbsite under controlled low pressure of the residual water vapour.
The aim of this study was to characterize by means of ETA [20, 21] the
microstructure changes taking place during heating of the synthetic gibbsite in argon
and to evaluate the experimental ETA results by using a mathematical model. The
ETA results were compared with the results of TG and DTA.
Experimental
Material studied
A synthetic gibbsite Al(OH)3 obtained by the Bayer process and supplied by
Pechiney (France) was used in the experiment. Its grain size is between 50 and
80 µm, its krypton BET surface area is 0.14 m2 g–1 and it contains 0.21% sodium,
0.08% iron and 0.21% CO2.
Methods used
ETA [20, 21] involves the measurements of radon release rate from samples previ-
ously labelled. The samples for ETA were labelled by acetone solution containing
traces of 228Th and 224Ra nitrates. The specific activity of a sample after labelling was
Fig. 1 TG/DTG, DTA and ETA results of gibbsite measured during heating in argon at
the rate 6 K min–1
J. Therm. Anal. Cal., 71, 2003

BALEK et al.: SYNTHETIC GIBBSITE
775
105 Bq g–1. Atoms of 220Rn were formed by the spontaneous α-decay of 228Th and
224Ra. The 224Ra and 220Rn atoms were incorporated into the sample due to the recoil
energy (85 keV/atom), which the atoms gain by the spontaneous α-decay. The maxi-
mum depth of 220Rn penetration was calculated by TRIM code [22]. It was determined
that the surface and near surface layers up to the maximal depth of 80 nm were la-
belled. The values of the radon release rate E (in relative units) were calculated from
the measured radioactivity values Aα of the Rn atoms released from the sample, di-
vided by the total radioactivity of the parent nuclides Aγ used for the sample labelling
i.e. E=Aα/Aγ as described in [20]. Semiconductor detector was used for the measure-
ment of the total γ-radioactivity. The schematic drawing of the ETA part of the appa-
ratus is given as Fig. 1 in [20].
The ETA measurements were performed during heating in argon at the rate of
6 K min–1, using a modified Netzsch ETA-DTA instrument, model 404. During the ETA
measurements the labelled sample of 0.1 g was situated in a corundum crucible and over-
flowed with the constant flow of argon (flow rate 50 cm3 min–1), which carried the radon
released from the sample into the measuring chamber of radon radioactivity. Details of
the ETA measurements and data treatment has been described elsewhere [20].
TG-DTA measurements were carried out using NETZSCH STA 409 apparatus
in the flow of argon (flow rate 50 cm3 min–1). The sample mass was 0.1 g, the sample
was placed into corundum crucible and heated at the rate of 6 K min–1.
Results and discussion
TG/DTG, DTA and ETA results
Figure 1 presents the TG/DTG, DTA and ETA results of synthetic gibbsite obtained
during heating in argon. The total mass loss observed on TG curve in the temperature
range 220–820°C was 34.2%. This value is in agreement with the theoretical value
34.6% for the reaction (1). The dehydration of the sample was characterized by three
DTA endothermic effects at 222, 303 and 498°C.
Al(OH)3→1/2Al2O3+3/2H2O
(1)
These DTA peaks were interpreted by several authors [1–4] who agree on the gen-
eral interpretation although not on the more detailed mechanism. For instance, the first
peak, observed here at 222°C, is usually considered to correspond to a partial transforma-
tion of gibbsite into boehmite, according to the Scheme (2) given by Mackenzie:
Al(OH)3→χ-Al2O3+γ-AlO(OH)
(2)
Now, it looked logical to most authors that the transformation starts on the sur-
face of the gibbsite crystal. Nevertheless, Rouquerol et al. [9] showed that this step
did not correspond to any increase in the krypton or nitrogen BET surface area (hence
no transition alumina is formed in the outer parts of the crystal) whereas droplets of
free water were shown (by NMR) to develop in the core of the crystal: this leads to
the idea that boehmite is formed in the center of the crystal, where hydrothermal con-
J. Therm. Anal. Cal., 71, 2003

776
BALEK et al.: SYNTHETIC GIBBSITE
ditions favourable to this transformation exist, and that this boehmite (which is 23%
denser than the starting gibbsite) leaves room for small cavities (located in the core of
the boehmitic phase) able to accomodate, in standard liquid form, part of the water
produced by reaction (2).
This first endothermic DTA peak corresponds to a mass loss of ~7.4% (TG
curve in Fig. 1). The second, more intense endothermic peak at 303°C, was accompa-
nied by a mass loss of ~22.1%.
A further mass loss of ~5.1% was completed at 820°C and corresponds to the total
dehydration of boehmite and formation of a second type of anhydrous aluminium oxide.
The latter dehydration step can be formally expressed by the chemical reaction (3).
AlO(OH)→1/2Al2O3+1/2H2O
(3)
This transformation is reflected by the endothermic effect on DTA curve at
498°C. The temperature of this effect is in accordance with the temperature of de-
hydroxylation of boehmite, stated by Mackenzie [8].
The ETA characterized microstructure development of the sample and brought
about additional information about microstructure development under in-situ condi-
tions of heating. The ETA results in Fig. 1 enabled us to determine the temperature
intervals at which the microstructure changes, due to respective dehydration steps
took place in comparison with the results of TG and DTA.
The microstructure changes which accompanied the first step of the dehydration
(reflected by the DTA effect at 222°C) were reflected by ETA as a small effect only, in
the same temperature range. The observed small intensity of the effect was in agreement
with the mechanism of the dehydration of gibbsite proposed by Rouquerol [10, 19]. He
suggested that the first step of the gibbsite dehydration takes place in the core of the
grains under hydrothermal conditions. Taking into account that the gibbsite crystals were
labelled by 220Rn and its 224Ra parent nuclide to the maximum depths of the 80 nm, the in-
formation obtained by ETA must be considered as the description of behaviour of surface
layers up to this depth. Consequently, from the observed small ETA effect it was as-
sumed no drastic microstructure changes took place in this near surface layers of the
gibbsite sample in the mentioned temperature range. This confirmed the hypothesis pub-
lished in [10, 19] that the water formed during the dehydration of gibbsite escaped from
the crystal through very narrow ‘structural channels’ formed in the gibbsite structure.
These channels are parallel to each other and perpendicular to the reaction interface [19].
We supposed that the surface area acting at this dehydration was much smaller in com-
parison to the surface area of the whole crystal. Consequently, only small increase of the
radon release rate could be observed on ETA in the temperature range 200–260°C.
The steep increase of the radon release rate observed on sample heating from 230 to
333°C indicated drastic microstructure changes which took place in this temperature
range as the result of rupture of the core on the particle surface and that water was re-
leased through overall surface of the grains. This gave rise to the large surface area of the
intermediate product and formation of new paths for radon release from the sample.
From the break observed at 330°C on the ETA curve and the subsequent de-
crease of the radon release rate, E, it followed that the surface area of the intermediate
J. Therm. Anal. Cal., 71, 2003

BALEK et al.: SYNTHETIC GIBBSITE
777
products slightly decreased on heating in spite of the fact that the dehydration of the
AlO(OH) continued in this temperature range and formation of Al2O3 was indicated
by DTA endothermic effect at 498°C: this is well in agreement with the earlier find-
ing that the transition alumina produced has relatively low thermal stability and starts
to loose its nitrogen BET surface area at 300°C under 5 mbar [19].
The decrease of E in the temperature range 615–1100°C reflected the decrease
of surface area and annealing of near surface structure irregularities in anhydrous
Al2O3. The enhanced increase of radon release rate E above 1100°C was ascribed to
the diffusion of the radon atoms from bulk of the sample.
Mathematical modelling and evaluation of ETA results
It was supposed in the modelling that 228 Th and 224Ra do not migrate in the solid at the
temperatures used for the ETA measurements. The value of the total emanating rate
ETOTAL of a solid is composed of two terms, namely the emanating rate, ER, due to re-
coil, and the emanating rate, ED, due to diffusion. The emanating rate ER depends on
the shape and size of the sample, the recoil range of 224Ra and 220Rn in the sample and
on the initial distribution of 228Th.
According to the theory proposed recently by Beckman and Balek [23], the total
emanating rate, E, during solid-state transitions of solids can be schematically written as:
E(T)=ER+ED(T)Ψ(T)
(4)
i.e. the second term is a product of two functions; ED characterizing the radon perme-
ability of diffusion channels and the Ψ(T) function characterizing the microstructure
changes in the solid.
In Fig. 2 the experimental ETA results (presented as dots) are compared with the
temperature dependences of radon release rate obtained by modelling (curves 2–4). It
was supposed, that following chemical processes took place in course of heating of
Fig. 2 Comparison of experimental ETA results (dots) with the results of the mathe-
matical modelling (curves 2–4) of the radon release rate from the gibbsite sam-
ple in segments II–IV. Segments I–IV were determined considering, that in each
segment only one type of solid-state change to be described by mathematical
modelling takes place
J. Therm. Anal. Cal., 71, 2003

778
BALEK et al.: SYNTHETIC GIBBSITE
gibbsite: the dehydration gibbsite to boehmite AlO(OH) and the dehydration of
boehmite AlO(OH) to aluminium oxide Al2O3, both accompanied by release of re-
maining water from the sample. Therefore the model of parallel diffusion was used
for the description of radon mobility in the synthetic gibbsite and the products of its
thermal decomposition. Four temperature ranges were distinguished during the mod-
elling of the ETA curve (Fig. 2 and Table 1).
In the ranges I and II, i.e. from 130 to 270°C and from 270 to 395°C, respec-
tively Eqs (5)–(7) were used in the modelling:
E1(T)=E1D(T)*Ψ1(T)
(5)
where
p3
p1exp –
p3
3
T
E (T ) = p1p2exp –
coth
–
(6)
ID
T
p3
3
p1exp –
T
and
p4
T–p5
2p6
ψ 1(T )=1–
1+erf
(7)
2
In the ranges III and IV i.e. from 400 to 615°C and from 615 to 1080°C, respec-
tively the Eqs (8) and (9) were used for the modelling:
E2(T)=Ψ2(T)
(8)
where
p4
T – p5
2p6
ψ 2 (T )= p7 1–
1+erf
(9)
2
and p1–p7 are parameters. In the calculation of these parameters we assumed that the
process of dehydration and structure defects annealing can be formally described as
first order kinetics.
From the parameters obtained during the modelling, the activation energy
QD [J mol–1] of radon diffusion as well as enthalpy ∆H [J mol–1] of the solid-state pro-
cesses were calculated:
QD=2Rp3 [J mol–1]
(10)
where R molar gas constant
Tmax
∆H(T )=Tmax 69.5+4.6log10
[J mol–1]
(11)
β
J. Therm. Anal. Cal., 71, 2003

BALEK et al.: SYNTHETIC GIBBSITE
779
The Tmax=p5 corresponds to the temperature of the maximal rate of the
solid-state process. The calculated values of the activation energy QD of radon diffu-
sion as well as enthalpy ∆H the solid-state process in the respective temperature inter-
vals are given in Table 1.
Table 1 The values of the activation energy QD/J mol–1 of radon diffusion and enthalpy
∆H/J mol–1 of the structural process in the respective temperature intervals
I
II
III
395–615
IV
615–1080
Temp. interval of the ETA curve/°C
130–270
270–395
Onset and final temperature of
microstructure changes determined from
Ψi(T) function/°C
130
340
310
370
430
615
615
1100
QD/J mol–1
∆H/J mol–1
Tmax/°C
1.541 105 2.396 105
4.126 104 4.504 104 6.358 104
–
–
9.98 104
253
338
529
972
ψ i(T )
max /%
1
3.5
3
10.2
ψ (T )
The comparison of the experimental ETA results with the results of the mathe-
matical modelling for the dehydration of gibbsite in the range up to 1100°C is pre-
sented in Fig. 2.
In order to quantitatively describe microstructure changes taking place during
heating of gibbsite in the investigated temperature ranges, we selected four segments
(Fig. 2) supposing for each segment the existence of one type of solid-state change. In
the segment I (from 130 to 270°C) no remarkable effect corresponding to microstrure
changes in the near surface layers of gibbsite during dehydroxylation was observed,
in the segment II (from 270 to 395°C) the surface area increased, characterizing the
drastic microstructure changes as the result of rupture of the core on the surface of the
decomposing gibbsite particles, in the segment III (from 400 to 615°C) and IV (from
615 to 1080°C) the annealing of surface and near surface layers of the intermediate
product of thermal decomposition were supposed.
The model curve 2 in Fig. 2 characterized the rupture of the core on the surface of
the particles and the water release through all the surface of the grains. The activation en-
ergy QD of radon diffusion as well as enthalpy ∆H during this solid-state process, which
were calculated using the proposed model are 239.6 and 45.04 kJ mol–1, respectively. The
model curve 3 in Fig. 2 characterized the final step of the sample dehydration. The de-
crease of the radon release rate reflected the fact that the surface area of boehmite did not
increase during heating in spite of the fact that the dehydration of the γ-AlO(OH) contin-
ued. The calculated enthalpy ∆H of this process is 63.58 kJ mol–1. Model curve 4 in Fig. 2
characterized the decrease of surface area and annealing of near surface structure irregu-
larities in anhydrous Al2O3. The calculated value of enthalpy characterizing this process
is 99.8 kJ mol–1.
J. Therm. Anal. Cal., 71, 2003

780
BALEK et al.: SYNTHETIC GIBBSITE
In addition, in order to better characterize the process of formation of boehmite
under hydrothermal conditions, which was reflected as a small effect on experimental
ETA curve, the modelling of the experimental ETA results in segment I was carried
out separately.
The model curves 2–4 were subtracted from the experimental ETA curve and
curve presented in Fig. 3 as dots resulted. The result of the mathematical treatment of
these data is presented in Fig. 3 as curve 1. The following values of activation energy
of radon diffusion QD and enthalpy ∆H of this solid-state process were determined
from the modelling: QD=154.1 and ∆H = 41.26 kJ mol–1, respectively.
The Ψi(T) functions used in the mathematical modelling for the characterization
of the microstructure changes and/or the annealing of radon diffusion paths in the
gibbsite sample in the above determined four segments are presented in Fig. 4. The
functions Ψi(T) (curves 1–4) as well as their first derivatives dΨi/dT (curves 1’–4’)
Fig. 3 Results of mathematical modelling (curve 1) compared with the experimental
ETA data (dots) characterizing microstructure changes in segment I
Fig. 4 Temperature dependences of functions Ψi(T) (curves 1–4) and their derivatives
dΨi/dT (curves 1’–4’) in different segments of ETA curves as determined in Fig. 2
J. Therm. Anal. Cal., 71, 2003

BALEK et al.: SYNTHETIC GIBBSITE
781
were used for the evaluation of the microstructure changes in the individual segments
determined in Fig. 2. The temperatures of Tmax, which indicate the maximal rate of the
annealing of radon diffusion paths, were determined from the curves 1’–4’ in Fig. 4
and are summarized in Table 1 along with the values of the parameter
Ψi (T )max/Ψ(T ) characterizing the relative intensity of the annealing processes in
the respective segments. From Fig. 4 it follows that the lowest intensity of the
microstructure changes was observed in the temperature range from 130 to 270°C,
whereas the relative highest intensity (more than 10%) corresponds to the change
above 615°C characterizing the annealing of alumina particles.
Conclusions
ETA results brought about additional information about development of the micro-
structure during thermal decomposition of gibbsite. The mathematical model used for
the evaluation of ETA experimental results was found suitable and can be recom-
mended for further application.
* * *
The results were supported by the Ministry of Education of the Czech Republic in the frame of the
projects No LN00A028 and ME 344 (Program Kontakt). The authors thank for the financial support
of the bilateral cooperation between Czech Republic and France (Barrande project, No 2000-014).
The authors (I.M.B and I.N.B) express their thanks to NATO science fellowship programme for the
financial support of the research stay in the Czech Republic.
References
1 R. C. Mackenzie, in ‘The Differential Thermal Investigation of Clays’ (R. C. Mackenzie, Ed.)
Mineralogical Society, London 1957, p. 299.
2 W. F. Bradley and R. E. Grim, Am. Miner., 36 (1951) 182.
3 J. F. Brown, D. Clark and W. W. Elliott, J. Chem. Soc. London, 13 (1953) 84.
4 J. H. de Boer, J. M. H. Fortuin and J. J. Steggerda, Proc. Kon. Ned. Akad. Wetensch.,
Ser. B 57 (1954) 170.
5 V. P. Ivanova, Zap. vses. Miner. Obshch., 90 (1961) 50.
6 B. A. Scott and W. H. Horsman, Trans. Brit. Ceram. Soc., 69 (1970) 131.
7 W. Lodding, in ‘Thermal Analysis’, R. F. Schwenker Jr. and P. D. Garn Eds, Academic Press,
New York 1969, p. 1249.
8 R. C. Mackenzie, in ‘Differential Thermal Analysis’ Vol. I., (R. C. Mackenzie, Ed.), Academic
Press, London 1970, p. 279.
9 J. Rouquerol, F. Rouquerol and M. Ganteaume, J. Catal., 36 (1975) 99.
10 J. Rouquerol, F. Rouquerol and M. Ganteaume, J. Catal., 57 (1979) 222.
11 F. Paulik, J. Paulik and R. Nauman, Thermochim. Acta, 64 (1983) 1.
12 K. J. D. MacKenzie, J. Thermal Anal., 5 (1973) 19.
13 R. C. T. Slade, J. C. Southern and I. M. Thompson, J. Mater. Chem., 1 (1991) 563.
14 J. T. Kloprogge, H. D. Ruan and R. L. Frost, J. Mater. Sci., 37 (2002) 1121.
J. Therm. Anal. Cal., 71, 2003

782
BALEK et al.: SYNTHETIC GIBBSITE
15 M. N. Danchevskaya, O. G. Ovchinnikova, Y. D. Ivakin and G. P. Muraveva, Russ. J. Phys.
Chem., 74 (2000) 1250.
16 A. P. Ilin, V. Y. Prokofev, T. V. Sazanova and S. P. Kochetkov, Russ. J. Appl. Chem.,
70 (1997) 93.
17 U. Steinike, H. Geissler, H. P. Hennig, K. Jancke, J. Jedamzik, U. Kretzschmar and
U. Bollmann, Z. Anorg. Allg. Chem., 590 (1990) 213.
18 K. J. D. MacKenzie, J. Temuujin and K. Okada, Thermochim. Acta, 327 (1999) 103.
19 J. Rouquerol and M. Ganteaume, J. Thermal Anal., 11 (1977) 201.
20 V. Balek, Thermochim. Acta, 192 (1991) 1.
21 V. Balek, J. Šubrt, T. Mitsuhashi, I. N. Beckman and K. Györyová, J. Therm. Anal. Cal.,
67 (2002) 15.
22 J. F. Ziegler and J. P. Biersack, The stopping and range of ions in solids, Pergamon Press,
New York 1985.
23 I. N. Beckman and V. Balek, J. Therm. Anal. Cal., 67 (2002) 49.
J. Therm. Anal. Cal., 71, 2003