Synthesis, characterization and catalytic properties of sol-gel derived mixed oxides

Article abstract of DOI:10.1016/S0022-3697(03)00276-2

Alumina and 1:1 mixed oxides of Al₂O₃-ZrO ₂, Al₂O₃-TiO₂, SiO ₂-TiO₂ and ZrO₂-TiO₂ were synthesized via a sol-gel process by the so-called neutral amine route, followed by calcination at 600 °C. The mixed oxides were characterized before calcination by X-ray diffraction, revealing ordered hexagonal (Al ₂O₃, Al₂O₃-ZrO₂, ZrO ₂-TiO₂) and lamellar (SiO₂-TiO₂, Al₂O₃-TiO₂) structures. After calcination, all samples exhibited amorphous structures. Surface areas and pore characteristics of all materials were determined by nitrogen adsorption isotherms. The calcined oxides are active catalysts in the epoxidation of cyclooctene with tert-butyl hydroperoxide as oxidant. Epoxide yields from 14 to 45% were found for reaction times of 24 h. The titanium oxide containing catalysts are the most active and selective ones. On the other hand, the poisoning of acidic centers yields a decreasing activity while increasing selectivity.

Full text of DOI:10.1016/S0022-3697(03)00276-2

Journal of Physics and Chemistry of Solids 64 (2003) 2385–2389
www.elsevier.com/locate/jpcs
Synthesis, characterization and catalytic properties of sol–gel
derived mixed oxides
*
Robson F. de Farias , Ulrich Arnold, Leandro Martınez, Ulf Schuchardt,
´
Marcelo J.D.M. Jannini, Claudio Airoldi
´
˜
Instituto de Quımica, Universidade Estadual de Campinas, CP 6154, 13083-970 Campinas, Sao Paulo, Brazil
Received 16 January 2003; revised 11 June 2003; accepted 11 June 2003
Abstract
Alumina and 1:1 mixed oxides of Al2O3–ZrO2, Al2O3–TiO2, SiO2–TiO2 and ZrO2–TiO2 were synthesized via a sol–gel
process by the so-called neutral amine route, followed by calcination at 600 8C. The mixed oxides were characterized before
calcination by X-ray diffraction, revealing ordered hexagonal (Al2O3, Al2O3–ZrO2, ZrO2–TiO2) and lamellar (SiO2–TiO2,
Al2O3–TiO2) structures. After calcination, all samples exhibited amorphous structures. Surface areas and pore characteristics
of all materials were determined by nitrogen adsorption isotherms. The calcined oxides are active catalysts in the epoxidation of
cyclooctene with tert-butyl hydroperoxide as oxidant. Epoxide yields from 14 to 45% were found for reaction times of 24 h. The
titanium oxide containing catalysts are the most active and selective ones. On the other hand, the poisoning of acidic centers
yields a decreasing activity while increasing selectivity.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
materials, the so-called redox molecular sieves [7], were
extensively investigated.
Besides a wide variety of crystalline silicalites, zeolites,
aluminophosphates and silicoalumiminophosphates, amor-
phous metallosilicates and mixed oxides, prepared by sol–
gel techniques, were found to be useful heterogeneous
oxidation catalysts [8–12]. The use of TS-1 and related
catalysts is limited to substrates which can enter the well
defined pores of the crystalline materials. Sol–gel tech-
niques, on the other hand, allow a wide structural variation
and adaptation of the catalysts to the requirements of the
catalytic reaction.
The sol–gel process [1] is an useful synthetic approach
for the preparation of amorphous, as well as structurally
ordered materials. Through sol–gel process, it is possible to
control the properties of the synthesized samples, such as
porosity and surface area, to obtain homogeneous matrices.
It has been shown that metal-doped lamellar silica [2–4]
and mixed oxides of ZrO2–TiO2 [5] with hexagonal
structure can be synthesized via the sol–gel method by
the so-called neutral amine route.
The synthesis and industrial application of titanium
silicalite 1 (TS-1) [6] have been focused on the development
of new heterogeneous oxidation catalysts. In this context,
transition metal containing micro- and mesoporous
The present investigation focus on sol–gel synthesis of
alumina and 1:1 mixed oxides of Al2O3–ZrO2, SiO2–TiO2,
ZrO2–TiO2 and Al2O3–TiO2. Due to their species with
¨
acidic sites of different types (mainly Bronsted sites for SiO2
and Lewis sites for the other oxides) and strengths, the
mixed oxides are interesting matrices from a catalytic point
of view. These oxides were synthesized through the neutral
amine route and were characterized by X-ray diffraction,
nitrogen physisorption and thermogravimetry. The catalytic
*
Corresponding author. Address: Instituto de Quimica,
Universidade Federal de Roraima, Av. Venezuela s/n, Boa Vista,
Roraima 69310-270, Brazil. Tel.: þ55-19-3788-3109; fax: þ55-
19-3788-3023.
E-mail address: robsonfarias@aol.com (R.F. de Farias).
0022-3697/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0022-3697(03)00276-2

2386
R.F. de Farias et al. / Journal of Physics and Chemistry of Solids 64 (2003) 2385–2389
activity of these materials towards cyclooctene epoxidation
with tert-butyl hydroperoxide was tested.
2.5 mmol of n-decane (Aldrich, 99%) as internal standard
for gas chromatography (GC) was magnetically stirred at
80 8C for 24 h. Aliquots were taken from the reaction
mixtures after 3 and 24 h. The samples were analyzed, using
an Hewlett Packard HP 5890 Series II gas chromatograph
equipped with an HP Ultra 2 capillary column (50.0 m £
2. Experimental
0.2 mm
£
0.33 mm film thickness; crosslinked by 5%
2.1. Synthesis and characterization of the oxides
phenylmethylsilicone) and a flame ionization detector
(FID). Products were quantified using calibration curves
obtained with standard solutions.
Tetraethoxysilane, titanium(IV) butoxide, zirconiu-
m(IV) butoxide and aluminum sec-butoxide were used as
alkoxide precursors (all of them from Aldrich). The
compounds 1,10-diaminodecane, 1,12-diaminododecane,
and dodecyltrimethylammonium bromide were used as
organic templates, and n-propanol was employed as solvent
for all reactions. All reagents were of analytical grade and
were used without further purification.
In order to increase the selectivity of the catalytic
reaction, 100 mg of Al2O3 and the mixed oxides Al2O3–
ZrO2 and SiO2–TiO2 were treated with 10 cm3 of a
0.5 mol dm23 lithium acetate solution for 4 h under
magnetic stirring. Then the powders were filtered under
vacuum, washed with 40 cm3 of distilled water and then
dried at 120 8C for 12 h. Such an experimental procedure
13.0 mmol of each metal precursors were mixed and
magnetically stirred under a nitrogen atmosphere for 72 h.
The resulting mixture was added to a solution of 10.0 mmol
of one of the organic template in 500 mmol of propanol.
Here 1,12-diaminododecane was used for titanium–zirco-
nium oxides, dodecyltrimethylammonium bromide was
employed for silicon–titanium oxide, and 1,10 diaminode-
cane was used as template for all other oxides. During this
step is critical to prevent water from reaction in order to
control the reaction rate and avoid undesirable hydrolysis
and polycondensation reactions. Different template mol-
ecules were employed to provoke variations in the pores
diameters of the prepared catalysts. Then, in this stage,
1.0 mol of water was added at once to the resultant solution.
The formed powder was aged for 48 h and then was washed
with distilled water and was aged again at room temperature
for 48 h. The materials were dried under vacuum at 80 8C
for 8 h and finally calcined at 600 8C (with a heating rate of
50 8C min21) for 6 h.
¨
was also used to neutralize the Bronsted acidic sites on the
matrices [13].
3. Results and discussion
3.1. Characterization of the mixed oxides
All oxides are fine white powders. Powder X-ray
diffraction patterns of the uncalcined materials
reveal hexagonal structures for Al2O3, Al2O3–ZrO2 and
ZrO2–TiO2 and lamellar structures for Al2O3–TiO2 and
SiO2–TiO2. X-ray data for the uncalcined materials are
summarized in Table 1 and the diffraction pattern of Al2O3–
TiO2, which exhibits a lamellar structure [14], is given as
example in Fig. 1. Upon calcination the template molecules
are removed, as shown by thermogravimetric analysis,
causing the collapse of the hexagonal or lamellar structures,
which were sustained by the presence of the template
molecules, then producing amorphous samples. It is worth
noting that at 600 8C TiO2 sol–gel derived materials will be
crystallized after calcinations. However, as in this case, the
crystallization of a mixed SiO2–TiO2 matrix will occur at
much higher temperatures than the single titanium oxide, as
verified by Kumar and co-workers [15].
X-ray diffraction patterns of the materials were recorded
on a Shimadzu XD-3A apparatus using Cu Ka radiation
(35 kV, 25 mA). Specific surface areas of the calcined
samples were determined by the BET method using nitrogen
adsorption isotherms recorded on a Micromeritics Flowsorb
II 2300 equipment. Pore diameters as well as micro and
mesopore volumes were determined from the corresponding
Horvath–Kawazoe pore size distribution curves. Thermo-
gravimetric data were obtained on a Shimadzu TGA-50
instrument under argon atmosphere with a heating rate of
Table 1
X-ray diffraction data for hybrid matrices consisting of aluminum,
silicon, titanium or zirconium oxides
5 8C min21
.
2.2. Catalytic epoxidations
Sample
2u (8)
d (nm)
Diffraction Structure
planes
The reactions were carried out at ambient atmosphere
using a glass tube placed in a temperature equilibrated oil
bath and fit with a reflux condenser.
Al2O3
3.2
4.5
2.80
1.96
2.88
100
100
100
Hexagonal
Hexagonal
Hexagonal
Lamellar
ZrO2–TiO2
Al2O3–ZrO2 3.3
A mixture of 10.0 mmol of cyclooctene (Aldrich, 95%),
15.0 mmol of tert-butyl hydroperoxide (TBHP, 88% in
cyclohexane; Nitrocarbono SA), 50.0 mg of catalyst and
Al2O3–TiO2
SiO2–TiO2
3.8; 7.6 2.32; 1.16 001; 002
4.1; 8.2 2.20; 1.10 001; 002
Lamellar

R.F. de Farias et al. / Journal of Physics and Chemistry of Solids 64 (2003) 2385–2389
2387
Fig. 1. X-ray diffraction pattern of Al2O3–TiO2 oxide.
Surface areas, pore volumes and pore diameters were
determined by nitrogen physisorption and are summarized
in Table 2. Surface areas between 100 and 160 m2 g21 were
determined for the mixed oxides and a remarkably higher
surface area of 343 m2 g21 was found for Al2O3. The mixed
oxides exhibits mesopores with pore diameters between
1.64 and 3.54 nm, whereas a pore diameter of 6.61 nm was
determined for Al2O3.
Two adsorption isotherm types for the calcined materials
were found. The oxides Al2O3, Al2O3–ZrO2, Al2O3–TiO2
and ZrO2–TiO2 presented MCM-41 like isotherms [16]. On
the other hand, the isotherm of SiO2–TiO2, Fig. 2(a), is a
type IV isotherm, which is typical of capillary condensation
taking place in mesoporous. It exhibits a type H4 hysteresis
loop, often associated with narrow slit-like pores [17]. For
SiO2–TiO2 the medium pore diameter is 1.64 nm. The
Horvath–Kawazoe differential pore volume plot, Fig. 2(b),
indicates that the pore volume attributable to micropores is
low, showing that the material is mainly non-porous.
Adsorption on amorphous surface is thermodynamically
more favorable than capillary condensation, thus giving a
type I isotherm. It is worth noting that, as shown in Fig. 2(b),
the SiO2–TiO2 matrix exhibits two kinds of pores: some in
the range 1.0–1.5 nm (with maximum values at ,1.15 nm)
and others with diameters in the range 1.9–3.0 nm (with
Fig. 2. Adsorption isotherm (a) and differential pore volume plot (b)
for SiO2–TiO2.
maximum value at ,2.25 nm). So, the ‘mean’ value shown
in Table 2 has the following physical meaning: if all pores of
the considered sample have the same value, (which is not
true, of course) this value will be around 1.64 nm. The same
argument is valid for the other samples. So, the values
shown in Fig. 2 could be used to compare the different
samples porosity.
The thermogravimetric curve for non-calcined SiO2–
TiO2 is shown in Fig. 3. Two clearly distinguishable weight
losses due to the release of water and the template molecules
Table 2
BET surface area ðSÞ; micro ðVmicroÞ and mesopores ðVmesoÞ; and
mean pore diameter ðdpÞ of sol–gel derived mixed oxides
determined by the corresponding Horvath–Kawazoe pore size
distribution curve
Oxide
SBET
(m2 g21
Vmicro
(cm3 g21
Vmeso
(cm3 g21
dp
)
)
)
(nm)
Al2O3
343
154
107
115
124
0.1359
0.0550
0.0408
0.0417
0.0508
0.4784
0.1062
0.0278
0.0482
0.0427
6.61
3.54
1.74
2.31
1.64
Al2O3–ZrO2
ZrO2–TiO2
Al2O3–TiO2
SiO2–TiO2
Fig. 3. Thermogravimetric and derivative curves for SiO2–TiO2.

2388
R.F. de Farias et al. / Journal of Physics and Chemistry of Solids 64 (2003) 2385–2389
are observed. These mass losses are evidenced in the
derivative forms of the curve, with the presence of two
distinct peaks, respectively. The curves for other materials
show a similar profile. Since the template molecules are
completely removed only at temperatures around 600 8C,
this temperature was chosen for calcination, to obtain
template-free oxide samples. On the other hand, this
temperature is already critical for the mixed oxides, since
pure ZrO2, TiO2 and Al2O3 are known to loose surface area
and pore volume at temperatures above 350 8C [18].
Table 4
Epoxidation of cyclooctene with TBHP catalyzed by mixed oxides
by using the reaction conditions: 10.0 mmol cyclooctene,
15.0 mmol TBHP (88% in cyclohexane), 2.5 mmol n-decane,
50.0 mg of catalyst at 80 8C
Catalyst
Cyclooctene
conversion
(%)
Selectivity
(%)
Epoxide
yield
(%)
3 h
3
24 h
36
3 h
24 h
3 h
3
24 h
14
Without
catalyst
100
39
3.2. Catalytic epoxidation tests
Al2O3
5
10
22
35
30
48
100
100
100
65
47
70
5
10
25
18
20
45
The results of catalytic cyclooctene epoxidation with
tert-butyl hydroperoxide are given in Table 3. Epoxide
selectivities of 100% were observed after 3 h, except for the
reaction catalyzed by SiO2–TiO2, for which a selectivity of
93% was observed (ter-butylic alcohol is the byproduct).
However, selectivities decreased significantly with pro-
longed reaction times. On the other hand, epoxide yields are
remarkably higher after reaction times of 24 h compared to
only 3 h reactions. Epoxide yields around 22% were
obtained after 24 h, using Al2O3 and Al2O3–ZrO2 as
catalysts. The observed catalytic activity of Al2O3 is in
agreement with previously reported results [19]. Epoxide
yields of up to 45% after 24 h were obtained using titanium
containing mixed oxides. Cyclooctene conversions and
Al2O3–ZrO2
SiO2–TiO2
All oxide samples were previously treated with a lithium acetate
solution.
and 4 data, it can be observed that, as a general behavior, the
treatment with lithium acetate (neutralization of the
¨
Bronsted acidic sites) leads to an increase in the selectivity
of the oxide matrices. On the other hand, the cyclooctene
conversion and the epoxide yield are reduced, which is an
¨
evidence of the prominent role of the Bronsted acidic sites
on the considered epoxidation reaction.
The better results obtained for titanium containing
catalysts in the present work, that is, higher cyclooctene
conversion (consumption of the main reagent) and epoxide
yeld (formation of the main product), could be explained
taking into account the well known catalytic activity of TiO2
due to its strong oxidant power, as observed also for SrTiO3
samples [20]. Furthermore, for WO3/TiO2 systems [21] it
was observed a photocatalytic activity 2.8–3 times higher
than those observed for pure TiO2 (gas phase decomposition
of 2-propanol). The higher acidity of the double oxide
system, in comparison with pure TiO2 could be used to
explain such facts [21].
epoxide selectivities increase in the order ZrO2–TiO2
,
Al2O3–TiO2 , SiO2–TiO2. The relatively low catalytic
activity of ZrO2–TiO2 could be interpreted by the lower
mesopore volume of the oxide (Table 2), and these results
suggest that at least part of the cyclooctene molecules react
inside the mesopores. However, some differences in surface
activity could not be disregarded.
The catalytic results for the samples treated with lithium
acetate solution are shown in Table 4. Comparing Tables 3
Table 3
Epoxidation of cyclooctene with TBHP catalyzed by mixed oxides
by using the reaction conditions: 10.0 mmol cyclooctene,
15.0 mmol TBHP (88% in cyclohexane), 2.5 mmol n-decane,
50.0 mg of catalyst at 80 8C
The low catalytic activity of Al2O3 despite its higher
mesoporousity, could be explained taking into account that
alumina exhibits almost only Lewis acidic sites, and, as
verified by comparison of Tables 3 and 4 data, the presence
¨
of Bronsted acidic sites is of prominent importance to
Catalyst
Cyclooctene
conversion
(%)
Selectivity
(%)
Epoxide
catalyze the considered reaction. This fact is also confirmed
by the higher activity of the silica–titania catalyst, since
silica have a large number of OH groups (an average value
of 5.0 OH groups per nm22) [1] whereas alumina and
zirconia are, in essence, Lewis acidic oxides, exhibiting a
minor amount of surface OH groups.
yield (%)
3 h
3
24 h
36
3 h
24 h
3 h
3
24 h
Without
catalyst
100
39
14
Al2O3
7
8
43
37
41
54
62
100
100
100
100
93
53
56
63
69
73
7
8
23
21
26
37
45
Al2O3–ZrO2
ZrO2–TiO2
Al2O3–TiO2
SiO2–TiO2
4. Conclusions
11
25
28
11
25
26
A wide variety of mixed oxides can be prepared via
the sol–gel method by the neutral amine route.

R.F. de Farias et al. / Journal of Physics and Chemistry of Solids 64 (2003) 2385–2389
2389
[2] R.F. de Farias, C. Airoldi, Colloids Surf. A 172 (2000) 145.
[3] R.F. de Farias, C. Airoldi, J. Non-Cryst. Solids 261 (2000) 181.
[4] R.F. de Farias, C. Airoldi, J. Solid State Chem. 149 (2000) 113.
[5] R.F. de Farias, C. Airoldi, J. Colloid Interface Sci. 220
(1999) 255.
The characteristics of the materials can be tuned
by the choice of the monomeric precursors as well as
the preparation parameters. It is possible to vary the
composition of the materials within a wide range, provided
that appropriate organometallic precursors are available.
Without optimizing the preparation parameters it was
possible to synthesize epoxidation active catalysts via a
simple and fast preparation route. Modifying the materials,
it should be possible to obtain useful heterogeneous
oxidation catalysts even for industrial purposes. It is
worth noting that the titanium containing catalysts are the
most active and selective ones. On the other hand, the
poisoning of acidic centers yields a decreasing activity
while increasing selectivity.
[6] M. Taramasso, G. Perego, B. Notari, US Patent, 4 410 501
1983.
[7] I.W.C.E. Arends, R.A. Sheldon, M. Wallau, U. Schuchardt,
Angew. Chem. Int. Ed. Engl. 36 (1997) 1143.
[8] R. Hutter, T. Mallat, A. Baiker, J. Catal. 153 (1995) 177.
[9] W.F. Maier, J.A. Martens, S. Klein, J. Heilmann, R. Parton, K.
Vercruysse, P.A. Jacobs, Angew. Chem. Int. Ed. Engl. 35
(1996) 180.
[10] R. Neumann, M. Levin-Elad, Appl. Catal. A 122 (1995) 85.
[11] M. Rogovin, R. Neumann, J. Mol. Cat. A 138 (1999) 315.
[12] R. Neumann, M. Levin-Elad, J. Catal. 166 (1997) 206.
[13] I.W.C.E. Arends, R.A. Sheldon, M. Wallau, U. Schuchardt,
Angew. Chem. Ed. Engl. 36 (1997) 1144.
Acknowledgements
[14] R.F. de Farias, C. Airoldi, J. Solid State Chem. 158 (2001) 134.
[15] S. Kumar, et al., Mater. Lett. 38 (1999) 161.
[16] P.T. Tanev, T.J. Pinnavaia, Science 271 (1996) 1267.
[17] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A.
The authors thank the Conselho Nacional de Desenvol-
´ ´
vimento Cientıfico e Tecnologico (CNPq) for fellowships
`
and the Fundac¸a˜o de Amparo a Pesquisa do Estado de Sa˜o
´
Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 67
Paulo (FAPESP) for financial support. UA thanks FAPESP
and the Deutscher Akademischer Austauschdienst (DAAD)
for a scientific exchange grant within a FAPESP-DAAD
cooperation.
(1985) 603.
[18] W.F. Maier, I. Tilgner, M. Wiedorn, H. Ko, Adv. Mater. 5
(1993) 726.
[19] D. Mandelli, M.C.A. van Vliet, R.A. Sheldon, U. Schuchardt,
Proceedings of the XVIIth Iberoamerican Symposium on
Catalysis, Porto, Portugal, C39, 2000.
References
[20] M. Miyauchi, A. Nakajima, A. Fujishima, K. Hashimoto, T.
Watanabe, Chem. Mater 12 (2000) 3.
[21] K.Y. Song, M.K. Park, Y.T. Kwon, H.W. Lee, W.J. Chung,
W.I. Lee, Chem. Mater. 13 (2001) 2349.
[1] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic
Press, New York, 1990.