Aerogel and xerogel catalysts based on θ-alumina doped with silicon for high temperature reactions

Article abstract of DOI:10.1002/ejic.200400657

Numerous materials (supports and catalysts) based on alumina have been prepared using the sol-gel process and carbon dioxide supercritical drying. In this work two types of solids, i.e. xerogels and aerogels, were systematically compared and a way of introducing platinum metal with a content of 5% percent by weight was examined. The structural data, the surface area, Pt dispersion and catalytic activity for the decomposition of the propellant were measured for the various samples. The (Al₂O₃)_0.88(SiO ₂)_0.12 samples prepared show very interesting porosity values, especially for the aerogel. For this reason, they were chosen as supports for the synthesis of 5 wt% platinum on alumina catalysts. The results presented in this work allowed us to obtain an overall view of the influence of the preparation mode on the properties of platinum on alumina supported catalysts. The dispersion of the metal phase is directly dependent on the specific surface of the support. A significant value for the surface area implies a large amount of centers for interaction with the metal precursor and, consequently, the appearance of more centres of simultaneous germination. Although aerogels obtained by carbon dioxide supercritical drying always show superior properties compared with xerogels, for catalytic decompositions the xerogels still remain superior. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2005.

Full text of DOI:10.1002/ejic.200400657

FULL PAPER
Aerogel and Xerogel Catalysts Based on θ-Alumina Doped with Silicon for
High Temperature Reactions
[a]
[b]
[b]
[b]
Aurelien Florin Popa, Laurence Courth e´ oux, Eric Gautron, Sylvie Rossignol,* and
Charles Kappenstein[
b]
Keywords: Aerogel / Xerogel / Catalyst / Platinum / Alumina
Numerous materials (supports and catalysts) based on alu-
of the influence of the preparation mode on the properties of
mina have been prepared using the sol-gel process and car- platinum on alumina supported catalysts. The dispersion of
bon dioxide supercritical drying. In this work two types of
solids, i.e. xerogels and aerogels, were systematically com-
pared and a way of introducing platinum metal with a con-
tent of 5% percent by weight was examined. The structural
the metal phase is directly dependent on the specific surface
of the support. A significant value for the surface area implies
a large amount of centers for interaction with the metal pre-
cursor and, consequently, the appearance of more centres of
data, the surface area, Pt dispersion and catalytic activity for simultaneous germination. Although aerogels obtained by
the decomposition of the propellant were measured for the
various samples. The (Al 0.12 samples prepared
show very interesting porosity values, especially for the aero-
gel. For this reason, they were chosen as supports for the
synthesis of 5 wt% platinum on alumina catalysts. The results
presented in this work allowed us to obtain an overall view
carbon dioxide supercritical drying always show superior
properties compared with xerogels, for catalytic decomposi-
tions the xerogels still remain superior.
2
3 2
O )0.88(SiO )
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
Introduction
barium or silicon) can be introduced during the sol-gel
[
6Ϫ10]
process.
Moreover, for catalytic applications, it is gen-
The sol-gel method represents a way of synthesising ma-
terials which display excellent properties suitable for many
applications such as catalysis, ceramics and glasses, and
nano-materials.[ Therefore, numerous catalyst supports or
catalysts have been prepared by this procedure which leads
to high purity, homogeneous materials with large surface
areas. These properties are key factors for the application
erally necessary to add an active phase (a metal phase such
as platinum, palladium etc.) to improve the properties.
Numerous authors have used alumina-supported plati-
num catalysts, prepared generally by the impregnation of
commercially available supports. The literature concerning
the sol-gel method was less important until the 1990Јs dur-
ing which time more papers appeared. Samples have been
generally prepared by conventional drying processes (293 to
1]
[
2,3]
of these materials.
Their properties are comparable and,
in some cases, superior to some catalysts prepared using
conventional procedures such as co-precipitation or con-
ventional ceramic methods. In particular, we have focussed
our interest in recent years on the preparation and evalu-
ation of alumina-based catalysts for applications concern-
ing the decomposition of monopropellants.[ For this
specific purpose, it is important to obtain catalytic materials
or supports displaying high thermal stability (i.e. at 1200
3
93 K) leading to xerogels. The use of supercritical drying,
generally in carbon dioxide, leads to aerogels but the corre-
sponding reports are much less numerous and focus mainly
on pure alumina supports. The first paper concerning a Pt
catalyst supported on alumina (aerogel and xerogel) and
prepared by the sol-gel method was published in 1993 by
Balakhrishnan et al.
aluminium tri-sec-butoxide Al(O-sec-C H ) (ATB) follow-
ing the work of Yoldas and the platinum was introduced
during the sol-gel transformation using H PtCl as a pre-
cursor. The supercritical conditions for obtaining the aero-
gels were T ϭ 295 °C and P ϭ 85 bar with methanol. The
4]
[11]
The alumina precursor used was
4
9 3
°
C or even higher). The selected supports for achieving this
[12]
[5]
purpose are aerogel aluminas or doped aluminas. In the
latter case, the doping element (lanthanum, praseodymium,
2
6
[
a]
Alexandru Ioan Cuza University, Faculty of Chemistry, samples were then treated at 773 K under helium. The main
Department of Inorganic Chemistry,
conclusions from this work were: (i) during the treatment,
6
600 Iasi, Romania
[
b]
University of Poitiers, LACCO UMR 6503, Laboratoire de a reduction of the surface area can be observed and (ii)
Catalyse par les M e´ taux,
aerogels display a lower Pt metal dispersion in comparison
4
0 Avenue du Recteur Pineau, 86022 Poitiers Cedex, France
with xerogels of similar loading. In 1998, Cho et al.[ pre-
13]
Fax: (internat.) ϩ33-5-49454157
E-mail: sylvie.rossignol@univ-poitiers.fr
pared the same materials using aluminium tripropoxide Al-
Eur. J. Inorg. Chem. 2005, 543Ϫ554
DOI: 10.1002/ejic.200400657
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
543

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
(
OC H ) (ATP) and H PtCl , with a final treatment at enable the determination of evolved water and thus the for-
3 7 3 2 6
773 K. They demonstrated that the sol-gel (xerogel) catalyst mula of the initially dried solids. At the beginning of the
displays a higher thermal stability of the metal particles measurements, the dried samples always contain an excess
than the conventionally impregnated catalyst. For the sol- of water and are known as hydrated boehmites. For both
gel sample, the exposed Pt metal particles are anchored to samples presented in Figure 2, the complete formulas are
the alumina surface by metalϪoxygen bonding. Other xero- AlOOH·0.94H O and (AlOOH)0.94-(SiO ) ·1.0H O.
2
2 0.06
2
[
14]
gel Pt/Al O catalysts prepared by Castillo et al.
Manasilp et al.
and Therefore, in the low temperature range (20Ϫ800 °C), the
using the same precursors (ATB-ATP, behaviour of the doped samples is similar to that of pure
H PtCl ) and treated at 823 K and 773 K, respectively, boehmites. On the other hand, the TDA curves reveal an
2
3
[
15]
2
6
demonstrate that the co-gelation of alkoxides and noble me- important difference at high temperature: the exothermic
tals induces the stabilisation of the metallic phase and its peak at about 1200 °C which characterises the transform-
incorporation into the alumina network producing better ation into alpha alumina is clearly visible in the case of the
metal-support interactions. Recently, using ATB with ace- pure boehmites but is shifted to a higher temperature for
[16]
tylacetonate and LiBH as precursors, Sault et al.
dem- the doped samples. In this case, the boehmite is transformed
4
onstrated that after calcination at 873 K, the metal particles into transition alumina but its structure is more stable and
are not encapsulated or occluded into the support. All these leads to alpha alumina at much higher temperatures (Ͼ1350
observations confirm that the metal-support interactions °C). This thermal behaviour is the same for the aerogel and
strongly depend on the precursors used, on the thermal xerogel samples. Consequently, we expected that the doped
treatment and on the method of drying. In order to extend samples would not be transformed into alpha alumina, even
the use of such samples to high temperature applications, after a hard thermal treatment (5 h calcination at 1200 °C).
it seemed important to study the influence of the drying
conditions and to then compare xerogels and aerogels.
In this paper, we present our results on the sol-gel prep-
aration of alumina or alumina doped with silica as a sup-
port. A systematic comparison has been made between sup-
port samples dried under conventional conditions (xerogels)
and under supercritical conditions using CO (aerogels) in-
2
cluding the behaviour after the thermal treatment. The in-
troduction of platinum was carried out using the conven-
tional impregnation of dried or calcined sol-gel supports or
during the sol-gel process. Powder X-ray diffraction (XRD),
transmission electron microscopy (TEM) and thermal
analyses (TDA-TGA) were used to determine the structural
properties. To determine the effects of the synthesis on the
catalytic activity, an evaluation of the properties of the cata-
lyst towards the propellant decomposition has been per-
formed with aqueous solutions of hydroxylammonium ni-
ϩ
Ϫ
trate NH OH NO (HAN). The preliminary results are
3
3
Figure 1. X-ray diffraction patterns of pure and doped boehmites,
xerogels and aerogels (lines: PDF file 21Ϫ1307).
presented here.
Results and Discussion
A. Supports
The diffraction patterns of the four different types of sup-
ports, after drying, are presented in Figure 1. They are all
composed of nanocrystalline hydrated boehmite γ-
[
17]
AlO(OH)·nH O, see PDF file 21Ϫ1307 mentioned in ref.
2
The peak positions are essentially identical for all samples
which also show the same cell parameters. The shift of the
[
18]
first diffraction peak has been discussed previously.
The thermal analyses curves (Figure 2) indicate two
major weight losses accompanied by endothermic effects
which correspond firstly to the dehydration of the hydrated
boehmite and then to the transformation of boehmite into
gamma alumina. Assuming that at the end of the analyses
Figure 2. Thermal analyses of dried samples of AlO(OH) (x) and
the final product is the anhydrous α-alumina, the TG curves AlO(OH)-Si12 (x)
544 © 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjic.org
Eur. J. Inorg. Chem. 2005, 543Ϫ554

Aerogel and Xerogel Catalysts Based on Doped θ-Alumina
FULL PAPER
The diffraction patterns of the calcined supports (Fig- thermore, the X-ray mapping revealed the atomic distri-
ure 3) confirm this fact, revealing that the transformation bution of Al and Si across the samples. For the sake of
into alpha alumina is delayed by doping. After the thermal clarity, we present a previously published diagram (Fig-
[
17]
treatment, the structure of the Al O Si12 xerogel or aerogel ure 4).
For ASi12 (a) (aerogel sample) we can see that
2
3
corresponds to θ alumina (monoclinic, PDF file 23Ϫ1009) the distribution of both aluminium and silicon atoms is
with a well ordered structure. By comparison, the non- homogeneous over the entire zone of the sample which was
doped aerogel (Al O ) is entirely transformed into alpha analysed. Other explored zones show that this homogeneity
2
3
alumina after the same treatment (Figure 7).
is a characteristic of the aerogel samples (see 2 in Figure 4).
Xerogel sample ASi12 (x) contains some zones showing ob-
vious microscale heterogeneity (see 1 Figure 4). This latter
sample contains domains where the silicon concentration is
enhanced. From EDX analyses, the Si and Al atomic com-
positions were quantified. For the aerogel, the atomic Si:Al
ratio of 6:94 is in excellent agreement with the expected
value (6.3:93.7). For the xerogel, this ratio strongly depends
on the analysed zone (200 nm). In Figure 4 (1), the analysed
zone exhibits a very important excess of silicon: Si:Al ϭ
33:67. In fact, the large (200 nm) dark particle clearly vis-
ible on the image of the xerogel (see 2 in Figure 4) contains
only silicon and oxygen atoms which correspond to pure
amorphous silica. The aluminium content is smaller in the
zones where this silicon agglomeration appears. In compen-
2 3
Figure 3. X-ray diffraction patterns of doped samples of Al O -
Si12 (a) and (x) and nondoped Al
2
O
3
(a), calcined at 1200 °C for sation, other zones are reduced in their silicon content lead-
5
h (θ Al
2
O
3
, PDF: 23Ϫ1009; α Al
2
O
3
, PDF: 46Ϫ1212)
ing to a higher content of aluminium in comparison with
the calculated value. In these zones, the aluminium content
can reach values as high as 95.7%.
The values found for the BET surface area of the samples
Table 1) confirm that the doping retards the transform-
ation into alpha alumina. The doped xerogel retains a good
(
2
Ϫ1
specific surface area (65 m ·g ) whereas the aerogel has a
2
Ϫ1
much better value of 120 m ·g
(compared with 1Ϫ3
2
Ϫ1
m ·g for the nondoped alumina). This value is similar to
that published by Mizushima,[8,9] (114 m ·g ) for a 10%
doped sample prepared in an alcoholic medium. The po-
rous volume is slightly larger for the ASi12 (a) sample (0.24
2
Ϫ1
3
Ϫ1
3
Ϫ1
cm ·g ) than for the ASi12 (x) sample (0.16 cm ·g ). The
mean pore sizes of the two doped aluminas are comparable
at 9.5 and 9.7 nm, respectively.
The transmission electron microscopy images enabled us
to analyse the morphology of these calcined supports. Fur- for the: (1) Al O -Si12 (x) and (2) Al O Si12 (a) samples
2 3 2 3
Figure 4. Electron microscopy (X-ray mapping and EDX analyses
Table 1. Samples prepared, mode of platinum introduction, thermal treatment, nomenclature, specific surface area (BET) and experimen-
tal platinum load (expected value 5 wt.-%)
Sample
aerogel (a); xerogel (x)
Pt introduction
impregnation sol-gel
Nomenclature
BET
[m /g]
Pt
[%]
Pt size
[nm]
2
Al
Al
2
2
O
O
3
(a)
(x)
1
1
1
1
A (a)
A (x)
ASi12 (a)
ASi12 (x)
SG-Pt-B-c (a)
SG-Pt-B-c (x)
SG-Pt-BSi12-c (a)
SG-Pt-BSi12-c (x)
Pt-A-r (a)
Pt-A-r (x)
Pt-ASi12-r (a)
Pt-ASi12-r (x)
3
1
3
(
(
Al
Al
2
O
O
3
)
)
0.88(SiO)0.12 (a)
0.88(SiO)0.12 (x)
120
65
41
44
130
41
3
1
120
68
2
3
5
5
5
5
5
5
5
5
% Pt/Al
% Pt/Al
% Pt/(Al
% Pt/(Al
% Pt/Al
% Pt/Al
2
O
O
3
(a)
(x)
SG-1
SG-1
SG-1
SG-1
Imp-2
Imp-2
Imp-2
Imp-2
3.0
3.0[
3.5
2.6
2.2[
2.6
2.4
3.5
7.0
11.0
5.6
4.3
2.6
3.0
a]
a]
2
3
2
O
O
3
)
)
0.88(SiO)0.12 (a)
0.88(SiO)0.12 (x)
2
3
2
O
O
3
(a)
(x)
2
3
% Pt/(Al
% Pt/(Al
2
O
O
3
)
)
0.88(SiO)0.12 (a)
0.88(SiO)0.12 (x)
2
3
3.8
[
a]
Values obtained by XRD quantification. 1. calcination at 1200 ° (5 h); 2. reduction at 400 ° (3 h).
Eur. J. Inorg. Chem. 2005, 543Ϫ554
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
545

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
The silica segregation can explain the lower specific sur-
face area (Table 1) of the xerogel sample 1-Asi12(x) after
the thermal treatment at high temperature. During the
transformation of alumina into the alpha form, small sites
are first formed by a nucleation process. The corundum for-
mation then continues by a growing and sintering process
which requires contact between the former crystallites and
migration of cations or cationic vacancies from the alumina
particles of the previous transition.[25]
The stabilising effect of silicon occurs at the surfaces of
the crystallites. When the silicon distribution is homo-
geneous, the silicon atoms are bonded tetrahedrally through
µ-oxo bridges AlϪOϪSiϪOϪAlϪ and cannot take part at
the sintering process because the α-Al O structure contains
2 3
Figure 5. Enlargement of the diffraction patterns of the Al O Si12
samples X: (a) ϭ aerogel, (x) ϭ xerogel
2
3
the cations in octahedral sites. It is much more difficult for
these sites to be occupied by silicon. The surface silicon
layer behaves as a barrier between the growing areas and
limits the transformation. Another explanation for the de-
This cell parameter difference does not exist for the non-
lay of the transformation into alpha alumina is related to calcined samples displaying the boehmite pattern (Fig-
the crystalline structure of the transition alumina. Beyond ure 1). No displacements can be seen, in agreement with the
600Ϫ700 °C, the quantity of surface OH groups contained boehmite structure which displays only octahedral alu-
in alumina is no longer sufficient to maintain the stability minium sites. These sites are able to host aluminium atoms
of the structure. The removal of these hydroxyl groups leads but not silicon atoms. During the sol synthesis, boehmite
to the formation of tetrahedral vacancies which destabilise nanocrystallites are formed first by hydrolysis of the alu-
the structure. In the presence of well distributed silicon minium precursor. After addition of the silicon precursor,
atoms, the influence of these tetrahedral vacancies is low- hydrolysis occurs later and produces silicon species which
ered and the transformation is delayed.[
19Ϫ22]
may be bonded to the surface aluminium atoms of the
The diffraction peaks of the two samples after the high boehmite particles by a µ -oxo bridge, this being a conse-
2
temperature treatment have been enlarged to focus on the quence of an oxolation reaction:ϪSiϪOH ϩ HOϪAlϪ Ǟ
peak displacement (Figure 5). We observed a systematic ϪSiϪOϪAlϪ ϩ H O
2
and slight but visible displacement of the aerogel peaks to
Thus, before calcination, silicon atoms are located mainly
higher 2θ values. This displacement varies between 0.09 and at the surface of the crystallites and only during the trans-
0
.29° and can be explained by the cell parameters shrinking formation of boehmite into γ alumina can they participate
due to the replacement of aluminium by silicon atoms in in the transition γ-alumina structure by substituting tetra-
the structure. Silicon atoms are slightly smaller than Al hedral aluminium atoms. In Figure 6, which shows a pos-
˚
atoms (bond length SiϪO ϭ 1.91 A compared with sible structural rearrangement during the transformation of
˚
[23]
AlϪO ϭ 2.03 A).
If the structure of θϪΑl Ο can be boehmite, the surface Si atom is linked to a surface HO
2 3
viewed as a distorted spinel structure,[ we can approxi- group of boehmite and to an external hydroxyl group. The
mate the peak displacement on the basis of a cubic struc- transformation of the hydroxyl group and a slight silicon
ture. Assuming that all the silicon atoms are located in the displacement leads to the position of silica in a tetrahedral
tetrahedral sites of the θϪΑl Ο structure, we can estimate site of the transition alumina. The presence of these silicon
24]
2
3
the 2θ displacement from the bond length difference to be atoms stabilises the alumina defect structure and can ex-
of the order of 0.2 to 0.5°. The experimental shifts between plain the delay in the transformation into corundum.
the aerogel and the xerogel peak positions range from 0.1
The results presented in this section show that the drying
to 0.3°, indicating that the silicon content incorporated into technique has an important influence on the features of
the structure is far superior in the case of the aerogel (of doped alumina. The morphology, thermal stability, homo-
the order of 50%).
geneity, fine structure and the textural properties are depen-
Figure 6. Model of silicon insertion into the gamma alumina structure during the boehmite transformation
46 www.eurjic.org Eur. J. Inorg. Chem. 2005, 543Ϫ554
5
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Aerogel and Xerogel Catalysts Based on Doped θ-Alumina
FULL PAPER
dent on this. The ASi12 samples prepared show very inter- 2. Physico-Chemical Characterisation
esting porosity values, especially for the aerogel. For this
reason, they were chosen as supports for the synthesis of 5
wt.-% platinum on alumina catalysts.
a) Catalysts on Nondoped Supports
The diffraction patterns of both aerogel catalysts (sol-gel
and impregnated) are presented in Figure 7. As expected,
the structure of the support is α-Al O and the metal phase
2
3
appears as large sized crystallites. The average size values
calculated from peak breadths are around 70 nm for the
sol-gel catalyst and 26 nm for the impregnated one. For the
xerogel sample, the respective values are of the same order
B. Catalysts
1. Effective Catalyst Metal Load
(
Table 1). The sintering is slightly more pronounced for the
For a selection of catalysts (6 samples), the platinum con-
sol-gel samples compared with the samples prepared by im-
pregnation.
tent was analysed in order to establish precisely the effective
metal load after the thermal treatment. For two other
samples, SG-Pt-B-c (x) and Pt-A-r (a), the metal load was
estimated from a quantitative analysis of the X-ray diffrac-
For the Pt-A-r (a) sample, transmission electron mi-
croscopy agrees well with this average size. The image re-
veals a rather large distribution of the crystallite sizes, rang-
ing from a few tenths to several hundred angstroms. We
present two different characteristic zones of the sample in
Figure 8.
tion patterns with respect to their homologues SG-Pt-B-c
25]
(
a) and Pt-A-r (x).[ The results are presented in Table 1.
The metal content is between 47 and 71% of the calculated
value, so that a quantity of platinum is lost during the prep-
aration of the catalysts. This loss probably takes place dur-
ing the thermal treatment processes (calcination and re-
[
26]
duction).
During the introduction of the platinum pre-
cursor, the anions present in solution ([PtCl6]2
Ϫ
,
Ϫ
[
Pt(H O)Cl ] or others) are adsorbed into the surfaces of
2 5
the crystallites and this fixation is mainly due to electro-
static interactions between negative complexes and surface
ϩ
OH2 groups of the support present in acidic solutions or
Ϫ
to ligand substitution around the platinum atoms (Cl by
surface OH groups). The fraction of platinum atoms which
is not firmly anchored on the surface can become detached
Figure 8. Two transmission electron microscopy images of the Pt-
and leave during the thermal treatments, probably in the A-r (a) sample
form of more volatile oxychloride species. Nevertheless,
hexachloroplatinic acid remains the best precursor because
it allows the preparation of catalysts with high metal disper-
b) Catalysts on Doped Supports
[
11,27]
sions.
We expect also a lower metal percentage for the
Catalysts Prepared by Sol-Gel Synthesis: The diffraction
patterns indicate the presence of θ alumina for the support
and the same large metal crystallites (Figure 9) of average
size 43 nm for the xerogel and 56 nm for the aerogel. This
is the only case when the aerogel crystallites are larger than
the xerogel ones. This fact is probably related to the particu-
lar texture of the sample of SG PtϪBSi12-c (a). In fact, the
precursor anions can be adsorbed onto the boehmite sites
concurrently with the silicon species. As a consequent, the
materials displaying the α-Al O structure due to the strong
decrease of the BET surface area (Table 1).
For the samples supported on doped alumina, this plati-
num content decrease could be related to the surface
2
3
SiϪOH groups which can prevent the metal from bonding
firmly to the surface.[28]
Figure 7. Diffraction patterns of catalysts on nondoped supports:
(
1) SG Pt-B-c (a) sample; (2) Pt-A-r (a) sample. (PDF: Pt: Figure 9. Diffraction patterns of (1) SG Pt-BSi12-c (a) and (2) SG
0
4Ϫ0802; αAl
2
O
3
: 46Ϫ1212)
2 3
Pt-BSi12-c (x) samples (θ Al O , PDF: 23Ϫ1009)
Eur. J. Inorg. Chem. 2005, 543Ϫ554
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
547

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
surface nature and the interactions between crystallites may ticle, revealed a better silicon repartition compared with the
be different from the case of a pure boehmite sol. xerogel. Instead, the signal due to the platinum atoms is
After the thermal treatment at 1200 °C, the sample SG practically invisible over the zone showing that the metal is
Pt-BSi12-c (a) exhibits the largest specific surface area (130 present only as isolated crystallites.
2
Ϫ1
2
Ϫ1
m ·g ), compared with only 41 m ·g for its xerogel hom-
Catalysts Prepared by Impregnation on Doped Alumina:
ologue SG Pt-BSi12-c (x). The porous volume is 0.45 The diffraction patterns of the catalysts impregnated on
3
Ϫ1
cm ·g , five times larger than for the xerogel (0.09 doped alumina are presented in Figure 12. In the case of the
3
Ϫ1
cm ·g ) and the mean pore sizes are 14 and 7 nm, respec- aerogel, the peaks corresponding to platinum are practically
tively. Consequently, the platinum crystallites located inside indiscernible from the background indicating a very small
the aerogel pores can develop and reach larger sizes.
crystallite size. For the xerogel, the platinum peaks are weak
The TEM pictures (Figure 10) show, for the xerogel but visible. By comparing the intensity of the isolated (311)
sample, some zones with a normal aspect of the porous sup- peak (36 counts degree) with the SG-Pt-BSi12 c (x) sample
port and metal crystallites with sizes ranging from 7 to (which has practically the same metal load, 304 counts de-
100 nm and some other singular zones particles with differ- gree) we can estimate that only about 10 wt% of the plati-
ent textures can be observed. In order to identify the nature num contributes to the diffraction peaks in the Pt-ASi12-r
of the particles visible on the B image, we recorded three (x) sample. By using the width of the (311) peak, which
EDX spectra, shown in Figure 11. The first (a) corresponds appears well isolated at 2θ ϭ 81.3°, we can estimate a size
to a homogeneous zone. We could establish the presence of of 40 nm. In fact, for a large size distribution, the mean
aluminium and silicon but not platinum. The second reveals value calculated on the basis of peak’s width is largely over-
platinum crystallites and, finally, the third (c) indicates a estimated since the small crystallites make very little contri-
large particle containing only aluminium atoms. The ap- bution and the peak profile is mainly influenced by the
[
27,29]
pearance of such particles can be explained by the presence large crystallites.
We can therefore assume that a large
of the metal precursor in the boehmite sol in the earlier part of the crystallites are small in size and only a small
stages of the synthesis. The competition of adsorption on other part of the better developed crystallites will be re-
boehmite crystallites between the platinum and silicon spec- sponsible for the diffraction signal.
ies leads to crystallites covered by more platinum and less
silicon. The X-ray analysis reveals that a small fraction of
platinum atoms is rather well dispersed over the sample but
the other fraction exists as isolated crystallites. In the same
way as observed for the supports, some zones with a silicon
agglomeration can be identified.
Figure 12. Diffraction patterns of the (1) Pt-ASi12-r (a) and (2) Pt-
2 3
ASi12-r (x) samples (the circles correspond to θ Al O , PDF:
23Ϫ1009)
2
Ϫ1
The specific surface areas of these catalysts are 68 m ·g
2
Ϫ1
for the xerogel and 120 m ·g for the aerogel. This is prac-
Figure 10. Two transmission electron microscopy images for the tically the same as in the supports and indicates no pore
sample SG Pt-BSi12-c (x)
blocking which is in agreement with the small sizes of most
of the crystallites.
For the aerogel, we observed similar behaviour with plati-
Transmission electron microscopy confirmed these re-
num isolated crystallites and α-Al O particles. The X-ray sults. For the aerogel, almost all the platinum particles are
2
3
mapping, performed on a zone around the α alumina par- smaller than 4 nm (Figure 13, left). The xerogel exhibits a
Figure 11. EDX spectra corresponding to different zones of the sample SG Pt-BSi12-c (x)
548
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 543Ϫ554

Aerogel and Xerogel Catalysts Based on Doped θ-Alumina
FULL PAPER
larger size distribution with crystallites smaller than 4 nm Table 2. Comparison between average crystallite sizes determined
but with larger particles (Figure 13, right). Both histograms by transmission electron microscopy or by metal dispersion meas-
urements
are shown in Figure 14 as well as the average crystallite
sizes. From the metal dispersion measurements, an average
surface crystallite size can be determined (Table 2). Two
models may be used used, a cubic model with one face in
contact with the support and a hemispherical model. The
average sizes refer to the cube parameter or the diameter.
The results for the aerogel are in good agreement with the
average sizes determined from the electron microscopy his-
togram. For the xerogel, this agreement is poorer because
of the larger size distribution revealed by electron mi-
croscopy and X-ray diffraction (90% of the platinum ap-
pears as crystallites with average sizes of about 4 nm, i.e.
10% larger and visible by XRD). It is also possible that a
part of the metal crystallites, located in closed pores, is not
accessible for the hydrogen adsorption.[
30,16]
C. Discussion
. Influence of the Preparation Mode by Impregnation
1
versus the One-Step Method (Sol-Gel)
For the samples prepared by the sol-gel method, the cal-
cination at 1200 °C induces an important sintering of the
metal phase, whatever the nature of the support (xerogel or
aerogel). The SG Pt-BSi12-c (a) sample is of particular
note, displaying a very large specific surface area of 130
m ·g . This important difference is due to the specific syn-
thetic conditions (simultaneous presence of silicon and
platinum species in the sol) as well as to the supercritical
drying which allows the retention of the gel morphology.
The metal dispersion is not related to the porosity of the
support because the mean size of the platinum crystallites
for the SG Pt-BSi12-c (a) sample is still greater than 50 nm.
Therefore, the metal dispersion is mainly determined by the
growth and sintering of the particles.
Figure 13. Transmission electron microscopy photos of the (left)
Pt-ASi12-r (a) and (right) Pt-ASi12-r (x) samples
2
Ϫ1
In contrast, the catalysts prepared by impregnation on
calcined alumina contain smaller metal particles because
sintering cannot occur during the reduction process. In this
case, the dispersion of the metal phase is directly dependent
on the specific surface of the support. A large value of the
surface area implies a large amount of adsorption centers
within the metal precursor (i.e. H PtCl ) and, consequently,
2
6
the formation of numerous centers for germination of plati-
num particles during the reduction. The catalysts supported
Figure 14. Histogram of the size repartition of the platinum crystal- on doped alumina also exhibit special characteristics, es-
lites for the (1) Pt-ASi12-r (a) and (2) Pt-ASi12-r (x) samples (Tv,
pecially for the aerogel sample Pt-ASi12-r (a), namely a
Ts: see Table 2)
2
Ϫ1
specific surface area higher than 120 m ·g after thermal
treatment, good metal dispersion (Ͼ 50%) and a homo-
geneous distribution of silicon and platinum atoms over the
sample. As a consequence, this sample could be tested as
an interesting catalytic material for applications requiring
1) high temperatures.
(
Nevertheless, it must be stated that the great difference
in dispersion between sol-gel and impregnated catalysts is
related to the temperature difference applied to the plati-
Eur. J. Inorg. Chem. 2005, 543Ϫ554
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
549

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
num precursor. For the sol-gel samples, the platinum pre- platinum content is also reported. The decomposition of
cursor is decomposed to platinum crystallites during the the HAN79%/water mixture alone appears at a temperature
high temperature treatment (1200 °C, 5 h) in an oxidising of 165 °C.
atmosphere leading to a strong sintering effect. In the other
All the supports (without the active phase) display a cata-
case, the impregnated sample is reduced at low tempera- lytic effect. This effect is more pronounced in the case of
tures (see Exp. Sect.) and the maximum temperature experi- the xerogel samples (106 °C for A (x) and 96 °C for ASi12
enced by the platinum particles is only 400 °C. To confirm (x)). For the aerogels, the decomposition temperatures are
the critical role of the temperature, the xerogel sample (Pt- higher and similar for each sample (140 °C). The very
ASi12-r (x)) was treated with the same high temperature strong difference in the surface area between nondoped and
conditions after reduction at 400 °C. A small decrease in Si-doped alumina seems to have a limited influence on this
2
Ϫ1
the surface area could be observed from 68 m ·g to 43 reaction. The introduction of platinum leads to a further
2
Ϫ1
m ·g , accompanied by a strong sintering of the platinum catalytic effect in the case of the impregnated catalysts. A
particles. The average size was between 4 nm and 30 nm further decrease in the decomposition temperature, in com-
before treatment and 40 nm after treatment. It must be said parison with the support, is of the order of 25 °C except
that the real catalytic conditions used will not be so drastic, for the Si-doped aerogel which display a decrease of 60 °C.
the high temperatures being reached over very short time For the sol-gel catalysts, the effect of platinum is much
ranges.
more limited.
With regard to the supports, the effect of the surface area
remains very limited for the catalysts except for the impreg-
nated aerogel samples Pt-ASi12-r (x) which display an im-
portant decomposition temperature decrease (from 125 °C
2
. Influence of the Drying Mode: Xerogel versus Aerogel
The main advantage of the supercritical drying is the pos-
sibility of maintaining the microporosity of the wet gel and
obtaining materials with high specific surfaces and porosit-
ies. Indeed, the specific surface area of the aerogel supports
and catalysts is about 120 m ·g even after a thermal treat-
ment at 1200 °C (5 h). By comparison, the xerogel catalysts
reveal a specific surface area respectively for the sol-gel, 41
2
Ϫ1
to 85 °C) related to the surface area increase (1 m ·g to
2
Ϫ1
6
8 m ·g ). For both samples the platinum is of the same
2
Ϫ1
order. The higher decomposition temperatures of the sol-
gel catalysts compared with the impregnated catalysts are
mainly due to the platinum particles sintering during the
calcinations leading to a very low platinum accessibility.
This effect is clearly observed for the samples Pt-ASi12-r
(x) and SG-Pt-BSi12-c (x) which have the same effective
metal content (3.5%) but exhibit a decomposition tempera-
ture difference of 20 °C. This indicates the importance of
the preparation mode.
2
Ϫ1
2
Ϫ1
m ·g and for the impregnated, 68 m ·g . Those values
are related to the high porosity of the aerogel materials
3
Ϫ1
3
Ϫ1
(0.24 cm ·g for the doped alumina and only 0.16 cm ·g
for the corresponding xerogel). An other important aspect
is the platinum size distribution which is shorter for the
catalyst prepared by impregnation on doped aerogel than
for the xerogel catalyst. Furthermore, the platinum particles
are more homogeneously distributed on the aerogel sup-
port. All these remarks display the importance of the drying
mode for the final properties of the materials.
The comparison between aerogel and xerogel supports
and catalysts shows that the xerogel samples are always
more efficient (Figure 15). A possible explanation for this
difference can be related to the higher insulation properties
[
31]
of the aerogels.
Due to its lower thermal conductivity
and higher porosity, the temperature within the aerogel is
lower than the measured crucible temperature, leading to a
temperature difference higher than for xerogels and thus to
a higher measured decomposition temperature. Moreover,
when the exothermic decomposition starts on the external
surface of the catalyst powder, the release energy propagates
more slowly to the centre of the aerogel powder resulting in
a delay in the decomposition reaction of the monopropell-
ant.
D. Catalytic Tests
The decomposition temperatures of the HAN/water mix-
tures, with and without the presence of supports or catalysts
are reported in Figure 15. For the catalysts, the effective
Catalysts supported on silicon-doped alumina are pro-
mising for high temperature applications. Moreover, xerogel
catalysts have been revealed to be better than aerogel cata-
lysts despite the heterogeneous dispersion of silicon. To
confirm this hypothesis, studies must be carried out on the
decomposition characteristics, i.e. the temperature increase,
pressure increase and ignition delay.
Conclusion
Figure 15. Decomposition temperatures of HAN/water mixtures
with and without supports or catalysts and metal content for aero-
gels and xerogels materials
We have synthesised sol-gel materials based on pure or
doped alumina using classical and supercritical drying con-
550
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 543Ϫ554

Aerogel and Xerogel Catalysts Based on Doped θ-Alumina
FULL PAPER
ditions. For the supports, the morphology, thermal stability, stirred at 60 °C for 1 h. Concentrated hydrochloric acid (0.28 mL,
Ϫ3
Ϫ1
3
(
.3 10
mol, 12 mol·L ) and, if necessary, tetraethoxysilane
homogeneity and the structure depend strongly on the dry-
ing technique used. The ASi12 sample displays a high
parosity and it was selected as support of catalyst.
A general characteristic of all catalysts is that a fraction
of the platinum is lost during the preparation of the cata-
lysts and the real metal content is between 47 and 71% of
the calculated value (5 wt.-%).
Ϫ3
0.51 mL, 2.24 10 mol, 98%, Aldrich) as a doping precursor were
then added and the temperature was raised and maintained at 80
C for 2 h.
The xerogel samples were obtained after drying overnight at 120
C in a ventilated oven. For the aerogel samples, about 75% of the
water content evaporated at 80 °C over 3 h affording a viscous sol
which was poured in the drying mould (stainless steel, d ϭ 3.6 cm,
°
°
The comparison between the aerogel and xerogel samples h ϭ 3.5 cm). A 5-mm hole in the bottom of the mould enables the
displays, respectively, a fine distribution on the surface and solvent exchange. After 15 h, the solvent exchange started: water
a heterogeneous distribution of the metallic phase. This fea- was replaced first by acetone (complete miscibility) and then ace-
2
tone was replaced by CO just before the supercritical drying. The
ture can be linked to the larger specific surface area of the
aerogels.
The introduction mode of platinum by impregnation
leads to smaller particles which are well dispersed. On the
other hand, the sol-gel method results in the formation of
large particles.
solvent exchange procedure was as follows: the wet gel was covered
with acetone (about 25 mL) and the solvent was evacuated through
the bottom hole before the addition of acetone (25 mL, three times
a day over 5 days). The mould was then introduced into the auto-
clave (300 mL, Parr Instruments) and acetone (100 mL) was added.
The autoclave was equipped with two entries (liquid CO
2
and ar-
Finely, from our evaluation experiments on HAN solu-
gon) and two outflows (gas and liquid) and the temperature and
tions, the best catalytic activity is displayed by platinum the pressure can be controlled as shown in Figure 17.
catalysts prepared by impregnation on Si-doped xerogel al- The reactor was filled with liquid carbon dioxide using several fill-
umina.
ing and purging steps in order to quantitatively evacuate the ace-
tone and the air contained in the autoclave. Upon completely filling
the autoclave with liquid CO
tained between the bottle (room temperature) and the autoclave
water cooling system). The temperature difference (about 4Ϫ5 °C)
was sufficient to push liquid CO into the reactor because the press-
2
, a temperature difference was ob-
Experimental Section
(
2
A. Preparations
ure in the bottle (55 bar at 20 °C) exceeded the equilibrium pressure
in the reactor (49 bar at 14 °C). After complete filling, the pressure
was about 55 bar (Figure 18, point 1). The reactor was isolated and
the temperature increased from 15 °C to 40 °C which led to an
important pressure increase (step 1Ϫ2) up to 140Ϫ150 bar (inside
the supercritical domain). After 1 h, the supercritical fluid was
evacuated over 1 h at the same temperature and the pressure was
decreased to 1 bar (step 2Ϫ3). The heating was then turned off and
argon was flushed through the reactor (step 3Ϫ4). The aerogel was
then taken out of the reactor.
a) Support: Pure Alumina and Doped Alumina
The preparation procedure for pure alumina and Si-doped alumina
2 3 2
[(Al O )0.88(SiO )0.12] by the sol gel method has been described in
[6]
previous work and is shown in Figure 16. The ASi12 label for
this material (Table 1) indicates the final molar composition (after
drying and firing) as SiO /(SiO ϩAl O ) ϭ 0.12. The silica weight
2 2 2 3
percent in the dried sample is 7.4%, corresponding to an atomic
ratio Si:Al ϭ 6.3:93.7.
Pure boehmites and Si-doped boehmites (aerogels and xerogels)
Ϫ1
were calcined at 1200 °C under air for 5 h (ramp of 5 °C·min ) to
obtain pure or doped aluminas.
(
2)
b) Supported Platinum Catalysts (5 wt.-%)
The catalysts were prepared by two procedures: (i) a one step pro-
cedure using the sol-gel method or (ii) an impregnation procedure
on a calcined support (pure or doped alumina).
2 6
Sol-Gel Procedure: The platinum precursor solution (H PtCl ) was
introduced during the preparation of the support in the sol with
the hydrochloric acid (see Figure 16). The preparation procedure
follows the same steps used for the support (drying, calcination at
1
200 °C). The high temperature calcination induces the formation
[1]
of metal by a redox reaction:
(
3)
Figure 16. Scheme of the sol-gel preparation procedure
To prepare 2 g of support, aluminium tri-sec-butoxide (8.48 g,
Impregnation Procedure: A volume of the metallic solution
PtCl ) corresponding to the porous volume of the support
0.0334 mol, 97%, Aldrich) and pure water (60.1 g, 3.34 mol) were (H
2
6
Eur. J. Inorg. Chem. 2005, 543Ϫ554
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
551

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
Figure 17. Diagram of the supercritical drying apparatus
Figure 18. Diagram of the pressure and temperature for supercritical drying
(
1.5 mL·gϪ1) was added to the alumina (wetness impregnation).
B. Characterisations
The concentration of the solution was then calculated giving a me-
tal content of 5 wt.-%. The catalysts were dried overnight at 120
The effective metal load was determined from the CNRS analyses.
The specific surface areas of the samples were determined by the
BET method from the nitrogen adsorption isotherms at Ϫ196 °C
in an automated Micromeritics ASAP 2000 apparatus after evacu-
ation for 1 h 30 min at 350 °C. The method used is a simple BET
Ϫ1
°
C, calcined at 400 °C for 3 h under air (20 mL·min ) and reduced
Ϫ1
under H
2
(20 mL·min ) for 1 h at 200 °C and 2 h at 400 °C to
obtain platinum supported alumina.
7
0
points procedure with P/P values between 0.05 and 0.25. The
(
4)
complete adsorption and desorption isotherm was obtained for a
few samples leading to the porous volume values.
Whatever the procedure, the introduction of the metallic precursor
did not affect the acidity (pH near 1).
Differential thermal analyses (DTA) and thermogravimetric analy-
ses (TGA) were recorded in order to characterise the supports after
552
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 543Ϫ554

Aerogel and Xerogel Catalysts Based on Doped θ-Alumina
FULL PAPER
drying. TDA-TGA experiments were carried out between 25 and
due to the high exothermicity of the decomposition which destroys
a small amount of catalyst.
1
350 °C using a thermal analyst 2100 TA instrument with platinum
crucibles. The samples were heated in nitrogen flow
100 mL·minϪ1) with a ramp of 10 °C·minϪ1
a
(
.
Diffraction patterns and particles sizes were determined by X-ray
diffraction (XRD) experiments performed on a Siemens D 5005
powder θ-θ diffractometer using the Cu-Kα radiation (λKα
ϭ
0.15186 nm) and a graphite monochromator. The diffraction pat-
terns were obtained under the following conditions: dwell time: 2
s, step: 0.04°, constant divergence slit: 1°.
Crystalline phases were identified by comparison with PDF stand-
ards (Powder Diffraction Files) from ICDD. (AlOOH: 21Ϫ1307,
2 3 2 3
α-Al O : 46-1212, θ-Al O : 23Ϫ1009, Pt: 04Ϫ0802)
The average crystallite sizes (D) were determined from the Scherrer
equation: D ϭ (0.9·λ)/(B·cos θ), λ ϭ 0.15186 nm, B is the width at
half maximum and θ the Bragg angle for the peak in question (rad).
For boehmite, the peak (020) was used without any correction due
to the very small size of the particles. For platinum, the peaks (311)
and (222) were used and a width correction was introduced, i.e.
2
2 1/2
B
cor ϭ (Bexp Ϫ Binst
)
. To determine Binst, the analysis of stand-
ard LaB
6
powder was performed under the same conditions. The
Figure 19. Example of the catalytic decomposition temperature of
a HAN/water mixture on Pt-ASi-r (x)
widths of the peaks were obtained by profile analyses using a
pseudovoigt function from the diffract AT software (socabim,
France).
The structural pictures were obtained using the software CaRine
[
32]
Crystallography.
The platinum dispersion was obtained by hydrogen chemisorption
at room temperature with a lab-build apparatus using pressure
measurements. The samples (300 to 500 mg) were reduced at 400
Acknowledgments
The authors thank all the people working in the ‘‘service de Mesure
Physique’’ of the LACCO.
°
C over 1 h under a static pressure of hydrogen, degassed at this
temperature over 1 h and cooled to ambient temperature. The ad-
sorption isotherms enabled the determination of reversibly and ir-
reversibly adsorbed hydrogen. The dispersion was calculated using
the stoichiometric ratio H/Pt ϭ 1.[
[1]
C. Wang, C. Yeh, Appl. Catal. 2001, 209, 1Ϫ9.
E. Seker, N. Yasyerli, E. Gulari, C. Lambert, R. H. Hammerle,
4]
[2]
Platinum particle sizes and size distributions were obtained from
transmission electron microscopy using a Philips CM 120 micro-
scope with a linear resolution of 3.5 A. Analyses by EDX enabled
us to determine the distribution and quantification of atomic spec-
ies (Al, Si and Pt). The lines used were Lα at 9.441 keV (Pt), Kα
at 1.486 keV (Al) and Kα at 1.739 keV (Si), respectively.
Appl. Catal. B 2002, 37, 27Ϫ35.
[
[
[
3]
4]
5]
R. D. Gonzalez, T. Lopez, R. Gomez, Catalyis Today 1997,
˚
35, 293Ϫ317.
R Eloidi, S. Rossignol, C. Kappenstein, D. Duprez, N. Pillet,
J. Prop. Power 2003, 19, 213Ϫ219.
G. M. Pajonj, Catal. Today 1999, 52, 3Ϫ13.
S. Rossignol, C. Kappenstein, Intern. J. Inorg. Mat. 2001, 3,
[6]
5
1Ϫ58.
J. S. Church, N. W. Cant, D. L. Trimm, Appl. Catal. A. 1993,
01, 105Ϫ116.
[
7]
1
C. Catalytic Tests
[
[
8]
9]
Y. Mizushima, M. Hori, J. Mater. Res. 1993, 8, 2993Ϫ3014.
Y. Mizushima, M. Hori, J. Non-Cryst. Solids 1994, 167, 1Ϫ8.
C. K. Long, J. W. Richardson, M. Ozawa, J. Alloys and Com-
pounds 1997, 250, 356Ϫ359.
[10]
All samples were tested by the decomposition of an aqueous solu-
tion of HAN (hydroxylammonium nitrate, 79 wt.-%) prepared by
water evaporation of a 20 wt.-% HAN solution [from SME (SME:
French society of explosive materials)] in a rotating evaporator ap-
paratus at 50 °C under vacuum conditions (water pump), the con-
centration was controlled by density measurements. The experi-
ments were carried out in a TDA-TGA apparatus which gives a
good estimation of the decomposition temperature.[ The pro-
cedure was as follows. After placing the sample (15 mg) in an alu-
minium pan with a cover, the propellant was added with a micro-
[11]
K. Balakrishnan, R. D. Gonzales, J. Catal. 1993, 144,
395Ϫ413.
[
[
12]
13]
B. E. Yoldas, J. Am. Ceram. Soc. 1982, 65, 387Ϫ393.
I. H. Cho, S. B. Park, S. J. Cho, R. Ryoo, J. Catal. 1998, 173,
295Ϫ303.
[14]
S. Casillo, M. Moran-Pineda, V. Molina, R. Gomez, T. Lopez,
Appl. Catal. B 1998, 15, 203Ϫ209.
A. Manasilp, E. Gulari, Appl. Catal. 2002, 37, 17Ϫ25.
A. G. Sault, A. Martino, J. Catal. 2000, 191, 474Ϫ479.
A. F. Popa, S. Rossignol, C. Kappenstein, J. Mater. Chem.
2002, 12, 2866Ϫ2868.
M Nguefack, A. F. Popa, S. Rossignol, C. Kappenstein, Phys.
Chem. Chem. Phys. 2003, 5, 4279Ϫ4289.
M. Trombetta, G. Busca, R. J. Wiley, J. Colloid & Interf. Sci.
4]
[
[
15]
16]
Ϫ1
[17]
pipette (10 µL). The system was heated (10 °C·min ) to 250 °C
3
Ϫ1
under an argon flow (100 cm ·min ). An example of the catalytic
decomposition is given in Figure 19. The inflexion point of the tem-
perature curve indicates the decomposition temperature. The analy-
ses by TDA-TGA allowed us to obtain the following information:
[
[
18]
19]
1
997, 190, 416Ϫ426.
(i) the onset temperature of the decomposition, (ii) the concen-
[20]
[21]
J. B. Miller, E. I. Ko, Catal. Today 1998, 43, 51Ϫ67.
T. Horiuchi, T. Osaki, T. Sugiyama, J. Non-Cryst. Solis 2001,
291, 187Ϫ198.
tration of the HAN solution at decomposition and (iii) the ef-
ficiency of the catalyst, i.e. the exothermic peak. The weight loss
after a completed reaction is often different from the expected value
[22]
T. Horiuchi, L. Chen, T. Osaki, Catal. Letters 1999, 58, 89Ϫ92.
Eur. J. Inorg. Chem. 2005, 543Ϫ554
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
553

A. F. Popa, L. Courth e´ oux, E. Gautron, S. Rossignol, C. Kappenstein
FULL PAPER
[
23]
[28]
L. E. Sutton, Tables of Interatomic Distances and Configuration
in Molecules and Ions, The Chemical Society, London, 1965,
A. Douidah, P. Mar e´ cot, S. Szabo, J. Barbier, Appl. Catal.
2002, 225, 21Ϫ31.
W. L. Smith, J. Appl. Cryst. 1972, 5, 127Ϫ130.
A. Martino, A. G. Sault, J. S. Kawola, J. Catal. 1999, 187,
30Ϫ38.
[29]
1
1 and 18.
[
[
[
[
24]
25]
26]
27]
[30]
R. S. Zhou, R. L. Snyder, Acta Crystallogr., Sect. B 1991, 47,
17Ϫ630.
A. F. Popa, PhD Dissertation, University of Poitiers, France,
002.
C. K. Lambert, R. D. Gonzales, Appl. Catal. 1998, 172,
33Ϫ239.
L. Khelifi, A. Ghorbel, E. Garbowski, J. Chim. Phys. 1997,
4, 2016Ϫ2026.
6
[31]
L. W. Hrubesh, J. Non-Cryst. Solids 1998, 2225, 335Ϫ342.
CaRine Crystallography 3.1 software, Boudias and Monceau,
France.
[32]
2
2
Received July 28, 2004
Early View Article
Published Online December 6, 2004
9
554
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 543Ϫ554