Creation of Bronsted acidity by grafting aluminum isopropoxide on silica under controlled conditions: Determination of the number of Bronsted Sites and their turnover frequency for m-xylene isomerization

Article abstract of DOI:10.1002/cctc.201300824

There is not a unique Bronsted acid site for aluminosilicates (ASAs). IR spectroscopy following CO adsorption proves the creation of Bronsted acid sites on Al/SiO₂ ASAs, which are synthesized by the deposition of aluminum species on hydroxylated silica. These sites are active for ethanol dehydration and m-xylene isomerization. Controlled deposition under anhydrous conditions optimizes the number of sites, whereas the presence of water leads to alumina agglomerates with no Bronsted acidity. The turnover frequency for m-xylene isomerization (4.3×10^-4s ^-1 site^-1 at 350°C, atmospheric pressure, and 0.6cm³ h^-1 of m-xylene) is approximately 3times lower than that of the Bronsted acid sites of Si/Al₂O₃ and 75times lower than that of an ultrastable Y-type zeolite without extra-framework aluminum. Ethanol as the acid test: The thermodesorption of ethanol demonstrates the formation of Bronsted acid sites on aluminum-grafted silica. Knowing the number of sites, the turnover frequency (TOF) for m-xylene isomerization is calculated. Bronsted sites of Al/SiO₂ are weaker than those of Si/Al₂O₃ and ultrastable Y-type (USY) zeolites, which is reflected in the ethanol dehydration temperature.

Full text of DOI:10.1002/cctc.201300824

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300824
Creation of Brønsted Acidity by Grafting Aluminum
Isopropoxide on Silica under Controlled Conditions:
Determination of the Number of Brønsted Sites and their
Turnover Frequency for m-Xylene Isomerization
[
a, b]
[b]
[b]
Maxime Caillot,
Alexandra Chaumonnot, Mathieu Digne, and
[a, c]
Jeroen A. Van Bokhoven*
There is not a unique Brønsted acid site for aluminosilicates
ASAs). IR spectroscopy following CO adsorption proves the
presence of water leads to alumina agglomerates with no
(
Brønsted acidity. The turnover frequency for m-xylene isomeri-
ꢁ
4
ꢁ1
ꢁ1
creation of Brønsted acid sites on Al/SiO ASAs, which are syn-
zation (4.3ꢀ10
s site at 3508C, atmospheric pressure, and
2
3
ꢁ1
thesized by the deposition of aluminum species on hydroxylat-
ed silica. These sites are active for ethanol dehydration and
m-xylene isomerization. Controlled deposition under anhy-
drous conditions optimizes the number of sites, whereas the
0.6 cm h of m-xylene) is approximately 3 times lower than
that of the Brønsted acid sites of Si/Al O and 75 times lower
2
3
than that of an ultrastable Y-type zeolite without extra-frame-
work aluminum.
Introduction
[
11,12,14]
Owing to the increasing need for middle distillates, the devel-
opment and optimization of catalysts to convert the heavy
fraction of crude oil are of fundamental importance. Thanks to
their Brønsted acidity, amorphous aluminosilicates (ASAs) are
materials of interest. However, their diversity in the distribution
of silicon and aluminum species throughout the material and
their amorphous character render a straightforward correlation
between surface structure and acidity difficult. Therefore, the
nature and number of Brønsted sites on ASAs are regular
genation.
Hensen et al. found that the Brønsted acidity
of Al/SiO , prepared by homogeneous deposition–precipitation
2
of aluminum nitrate on silica, results from the diffusion of a lim-
ited amount of tetrahedral aluminum atoms through the silica
lattice, which leads to negative charge compensation by labile
acidic protons, hence Brønsted acidity, similar to what occurs
[14]
in zeolites. The choice of the preparation technique for ob-
taining ASAs—co-gelation, grafting, and so on—strongly influ-
ences the spread of silicon and aluminum atoms through the
[
1–4]
[15]
topics of discussion.
particles and can lead to dissimilar structures.
[5]
We have described the synthesis of ASAs by grafting. Al/
The nature of the OH groups upon the grafting of aluminum
can be monitored by observing the OH stretching region
SiO was obtained by depositing Al(OiPr) on an amorphous
2
3
ꢁ
1
silica gel under controlled conditions. Such Al/SiO materials
(ꢀ3000 to 4000 cm ) by IR spectroscopy. The adsorption of
2
ꢁ1
display Brønsted acidic character, which is not the case of the
CO leads to changes in the 2100 to 2250 cm range (CO
stretching); each band thus obtained is ascribable to a specific
type of acidic site.
[
6–12]
parent support.
cracking of cumene.
tional catalysts, which are active in hydrogenation/dehydro-
For example, Al/SiO materials catalyze the
2
[
9,10,13]
The deposition of Pd leads to bifunc-
We already demonstrated the relevance of ethanol adsorp-
tion followed by thermogravimetry to evaluate the develop-
[16]
ment of Brønsted acid sites (BAS) on Si/Al O ASAs. The de-
2
3
[a] Dr. M. Caillot, Prof. Dr. J. A. Van Bokhoven
hydration of ethanol to ethylene is catalyzed by the Lewis acid
Chemistry and Bioengineering
ETH Zurich
sites (LAS) of g-alumina and the Brønsted acid sites (BAS) of
[17–22]
Wolfgang-Pauli-Strasse 10, 8093 Zꢀrich (Switzerland)
Fax: (+41)44-632-1162
E-mail: jeroen.vanbokhoven@chem.ethz.ch
zeolites and ASAs.
Knowing the number of BAS enables
the calculation of the turnover frequency (TOF) in Brønsted
acid demanding reactions, such as m-xylene isomerization. On
[
b] Dr. M. Caillot, Dr. A. Chaumonnot, Dr. M. Digne
ꢁ
3
ꢁ1
ꢁ1
Si/Al O , the TOF is 1.4ꢀ10
s
site , and on an ultrastable
Direction Catalyse et Sꢁparation
2
3
IFP Energies nouvelles
Rond-point de l’ꢁchangeur de Solaize, BP 3, 69360 Solaize (France)
Y-type (USY) zeolite without extra-framework aluminum the
ꢁ
2
ꢁ1
ꢁ1 [16]
TOF is 3.1ꢀ10
s site .
[
c] Prof. Dr. J. A. Van Bokhoven
Laboratory for Catalysis and Sustainable Chemistry (LSK)
Paul Scherrer Institut (PSI)
The aim of this study was to determine the acidity of Al/SiO2
by grafting aluminum isopropoxide on silica. The catalytic ac-
tivity to m-xylene isomerization was measured and indicated
the emergence of Brønsted acidity upon grafting of silica with
aluminum species. This was confirmed by recording the IR
5
232 Villigen, (Switzerland)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300824.
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 832

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
spectra of the CO stretching region after CO adsorption. We
performed the adsorption of ethanol followed by thermog-
ravimetry. This enabled the determination of the surface cover-
Table 2. Surface density of ethanol retained after purge following pre-
treatment at 5008C.
ꢁ
2
[
16]
Entry
Sample
Ethanol retained after purge [nm ]
age of aluminum species, similar to Si/Al O .
The TOF of
2
3
Al/SiO for m-xylene isomerization was calculated.
1
2
3
4
5
6
7
8
9
silica
1.3
1.3
1.5
1.6
2.0
2.2
2.6
1.6
1.4
2.8
2
AS(3/anh)
AS(6/anh)
AS(10/anh)
AS(15/anh)
AS(25/anh)
AS(35/anh)
AS(16/1eqW)
AS(17/3eqW)
g-alumina
Results
Adsorption of ethanol followed by thermogravimetry
Surface coverage of aluminum species on Al/SiO2
1
0
The main characteristics of Al/SiO and its composition were
2
tuned by adapting the amount of precursor in the reaction
mixture (Table 1). Anhydrous conditions led to a uniform depo-
sition of aluminum species (entries 2–7), whereas the presence
The theoretical coverage of Al/SiO by aluminum species
2
[16]
was estimated as follows [Eq. (2)]:
[5]
of water favored a heterogeneous deposit (entries 8 and 9).
In the rest of the document, blue labels correspond to alumi-
na, silica, and materials prepared under anhydrous conditions
Theoretical density of surface Al atoms ½%ꢂ
density of grafted Al
ð2Þ
¼
ꢃ 100
(entries 1 to 7 and 10) and purple/pink labels correspond to
density of Al atoms on alumina
materials prepared under aqueous conditions (entries 8 and 9).
Following pretreatment at 5008C, ethanol was adsorbed at
room temperature, and weakly bonded molecules were desor-
bed by purging. With increasing alumina content (Table 2, en-
tries 1–7), a higher amount of ethanol was retained after the
In the plot of the experimental density of aluminum atoms
(
y axis) versus the theoretical plot (x axis, Figure 1), the dashed
line corresponds to the experimental density, which is the
same as the theoretical density, as found upon depositing the
aluminum atoms uniformly on the surface. Thus, samples that
were prepared under anhydrous conditions (entries 3–7) dis-
play a homogeneous deposition of species. The higher the sur-
face density of aluminum atoms, the higher the amount of
strong adsorption sites for ethanol. Conversely, entries 8 and 9
show inhomogeneous deposition, which confirms previous ob-
[
16]
purge. Similar to our previous observations, this indicated
a greater number of strong ethanol adsorption sites on the
alumina surfaces. Despite a similar aluminum loading, entries 8
and 9, synthesized in the presence of water, showed a smaller
amount of retained ethanol than entry 5, which was prepared
under anhydrous conditions.
[
16]
[5]
As for the Si/Al O materials, the surface coverage by alu-
2
3
servations. The higher the water content during synthesis,
the lower the density of the aluminum atoms accessible to
ethanol, which is indicative of the presence of clusters of alu-
mina species.
minum species on the Al/SiO2 materials was calculated by
comparing the amount of ethanol retained after the purge to
the minimum and maximum amounts on pure silica
ꢁ
2
ꢁ2
(
1.3 EtOH nm ) and pure alumina (2.8 EtOH nm ), respectively
Eq. (1)]:
The difference in ethanol retained after purge between silica
and AS(3/anh) is not accurate enough to calculate the surface
ratio of surface aluminum; thus, it is not reported in Figure 1,
and the surface ratio of aluminum species of AS(3/anh) used
further in the document corresponds to the theoretical one.
[
Experimental density of surface Al atoms ½%ꢂ
ꢁ
2
ð1Þ
amount of adsorbed ethanol ½nm ꢂꢁ1:3
¼
ꢃ 100
2
:8ꢁ1:3
Dehydration of ethanol
[
5]
2
Table 1. Synthesis conditions and main characteristics of Al/SiO .
The derivative of the weight loss
of silica, g-alumina, and various
Entry Sample Sample
Al
2
O
3
S
BET
2
Synthesis
technique
T [8C],
[
a]
ꢁ1
nature
name
[wt%] [m g
]
Pretreatment Synthesis
Al/SiO2 (prepared under anhy-
drous conditions) during the
thermoprogrammed desorption
(TPD) of adsorbed ethanol from
room temperature to 4008C
shows a desorption feature at
approximately 808C, associated
with ethanol desorption from
medium/strong sites, visible for
all materials (Figure 2a). Above
1
2
3
4
5
6
7
8
9
1
silica
silica
AS(3/anh)
AS(6/anh)
AS(10/anh)
AS(15/anh)
AS(25/anh)
AS(35/anh)
AS(16/1eqW)
AS(17/3eqW)
0.0 550
2.8 505
5.7 460
–
–
–
Al/SiO
Al/SiO
Al/SiO
Al/SiO
Al/SiO
Al/SiO
Al/SiO
Al/SiO
2
CLD anhydrous
CLD anhydrous
CLD anhydrous
CLD anhydrous
CLD anhydrous (ꢀ2)
CLD anhydrous (ꢀ3)
CLD anhydrous (1 equiv. H
CLD anhydrous (3 equiv. H
–
30, vacuum
30, vacuum
30, vacuum
30, vacuum
30, vacuum
30, vacuum
O) 30, vacuum
O) 30, vacuum
–
110, toluene
110, toluene
110, toluene
110, toluene
110, toluene
110, toluene
110, toluene
110, toluene
–
2
2
2
2
2
2
2
10.3 430
15.0 400
24.8 330
34.9 185
16.1 420
16.7 420
100.0 235
2
2
0
alumina g-alumina
[
1
a] AS(x/y): Al/SiO
or 3 equiv. water per precursor molecule: 1eqW and 3eqW).
2 2 3
with x wt% of Al O and obtained by technique y (CLD anhydrous: anh; CLD aqueous with
2
008C, ethylene desorbs from
[17,23]
the dehydration sites.
Dehy-
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 833

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 1. Fraction of available alumina surface: theoretical x values (based
on the composition data) versus experimental y values (based on of ethanol
adsorption). Numbers in parentheses correspond to entries in Table 1.
CLD=chemical liquid deposition.
dration occurs only on materials possessing surface aluminum
species, and the amount increases with the aluminum content.
The dehydration peak of Al/SiO is constantly located at ap-
2
proximately 2708C, except for AS(3/anh) (entry 2), the dehydra-
tion temperature of which is slightly higher (2758C). The dehy-
dration temperature is notably higher than that of g-alumina
Figure 2. Derivative of the TPD weight loss of ethanol on Al/SiO
samples after pretreatment at 5008C: a) prepared by anhydrous CLD; b) pre-
). The inten-
sities are normalized by the surface area of each sample. Numbers in paren-
theses correspond to entries in Table 1.
2
-grafted
pared by various methods with similar loading (ꢀ15–17% Al
2 3
O
(entry 10), which is indicative of a different surface.
For Al/SiO prepared in the presence of water (Figure 2b),
2
the higher the water content during synthesis, the higher the
dehydration temperature [2808C for AS(16/3eqW), entry 9] and
the lower the intensity. Compared to that of AS(16/anh)
site (Table 3, column 5). At a higher aluminum loading (en-
tries 3–7), the number of dehydration sites per exposed alumi-
num atoms decreases. Consistent with their small amount of
exposed aluminum, AS(16/1eqW) and AS(17/3eqW) show
a large number of dehydration sites per exposed atom, espe-
cially AS(17/3eqW), for which one third of the exposed alumi-
num atoms are associated to a dehydration site. For entries
2–6, 8 and 9, the more aluminum atoms present on the sur-
face, the greater the percentage of adsorbed ethanol mole-
cules that undergo dehydration (Table 3, column 6). For AS(35/
anh) (entry 7), only 20% of the ethanol is dehydrat-
(entry 5), the dehydration peak of AS(16/3eqW) is particularly
asymmetric.
For samples prepared under anhydrous conditions, high alu-
minum loading is associated to a higher amount of dehydrated
ethanol (Table 3, column 4). Samples in entries 8 and 9, pre-
pared by aqueous grafting, have fewer dehydration sites than
the sample in entry 5. At low aluminum loading [AS(3/anh),
entry 2], one fifth of the exposed atoms form a dehydration
ed; thus, the number of aluminum atoms that form
medium/strong adsorption sites (at which ethanol
2
Table 3. Ethanol dehydration over Al/SiO -grafted samples as a function of the sur-
adsorbs but does not react) increases more than the
number of aluminum atoms that form dehydration
sites.
face coverage of aluminum species.
Entry Sample
Surface coverage
of aluminum
EtOH undergoing dehydration Fraction of
[
b]
Isosurface
Per Al atom
EtOH dehy-
ꢁ
2
[a]
[c]
species [Eq. (1)] [%] [nm
]
drated [%]
1
2
3
4
5
6
7
8
9
1
silica
0
ꢀ5
11
15
46
56
83
0.00
–
0
10
13
20
22
28
20
23
19
30
[
d]
CO adsorption followed by IR spectroscopy
IR spectra of the OH stretching region
AS(3/anh)
AS(6/anh)
AS(10/anh)
AS(15/anh)
AS(25/anh)
AS(35/anh)
AS(16/1eqW) 21
AS(17/3eqW)
g-alumina
0.12
0.20
0.26
0.45
0.61
0.65
0.38
0.27
0.83
ꢀ0.2
0.15
0.14
0.08
0.09
0.05
0.14
0.33
0.07
In Figure 3, pure silica (entry 1) shows a main band at
ꢁ
1
[24]
3
745 cm corresponding to isolated silanols. The
ꢁ
1
low-wavenumber tail between 3400 and 3700 cm
[
24]
7
100
corresponds to weakly interacting silanol groups,
0
which is consistent with the high density of surface
ꢁ2
[a] Based on the total surface area. [b] Based on the surface coverage of aluminum
OH groups (3.3 OHnm ). The first deposited alumi-
num atoms (entry 2) do not change the global shape
of the spectra, except for a small bump at approxi-
species. [c] Relative to total EtOH desorbed during TPD. [d] Based on the theoretical
density of grafted atoms, by considering a uniform deposition.
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 834

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 4. IR spectra of Al/SiO
2
prepared by various grafting techniques:
a) AS(16/1eqW) and b) AS(17/3eqW). Numbers in parentheses correspond to
entries in Table 1. For each spectrum the intensity is adapted so that all
spectra can be compared.
3
eqW) (entry 9) is close to that of entry 2 (with 3% Al O ). Both
2 3
ꢁ
1
samples show a small bump at approximately 3600 cm .
IR spectra of the CO stretching region upon CO adsorption
Upon progressive dosing of CO until saturation on pure silica,
the CO stretching region displays two main bands (Figure 5a).
In the CO stretching region, the higher the CO stretching fre-
quency of the OH groups, which are characterized by bands in
ꢁ
1
[27]
the 2140 to 2180 cm region, the higher the acidity. CO ad-
sorption first causes an increase in the band at approximately
ꢁ
1
2
155 cm . This corresponds to adsorption on weak OH
[3]
groups. Further, CO adsorption yields a shoulder at approxi-
ꢁ
1
mately 2140 cm , which corresponds to physisorption on sila-
[3,28–30]
Figure 3. IR spectra of Al/SiO
2
prepared by anhydrous CLD: a) silica, b) AS(3/
nols.
CO adsorption on the sample with the lowest
anh), c) AS(6/anh), d) AS(10/anh), e) AS(15/anh), f) AS(25/anh), and g) AS(35/
anh). Numbers in parentheses correspond to entries in Table 1. For each
spectrum the intensity is adapted so that all spectra can be compared.
number of grafted aluminum [AS(3/anh), entry 2] causes the
ꢁ1
appearance of a band at 2179 cm , which corresponds to CO
[3]
adsorption on strong BAS. Such high wavenumbers are usu-
[31]
ally obtained for the strong BAS of zeolites. Thus, for this
sample, the BAS are the only sites that differ from the sites of
pure silica. A higher aluminum loading (entries 3–7) induces
ꢁ
1
mately 3600 cm , which is also detected at higher loading
(
entry 5). A broad band of low intensity at approximately
ꢁ
1
ꢁ1
3
850 cm , also visible for entry 4, corresponds to a torsional
the appearance of strong (2230 cm ) then weak/medium
[
24]
ꢁ1
[3,28]
mode resonance of silanols. Increasing the aluminum load-
ing (entries 3–7) causes a decrease in the tail from 3400 to
(ꢀ2200 cm ) LAS.
Their number increases with the alumi-
num content. At the same time, CO adsorption on the BAS in-
creases, and CO physisorption on silanols decreases.
ꢁ
1
3
700 cm and, hence, less bonding between silanol groups.
At the same time, a broad band centered at approximately
AS(16/1eqW) and AS(17/3eqW), prepared in the presence of
water, are characterized by the presence of strong and weak/
medium LAS and BAS, even at the first doses of CO (Figure 6).
The number of strong LAS on AS(16/1eqW) is particularly high.
The peak corresponding to the BAS is similar in both samples.
The peak corresponding to CO physisorption on silanols is
better defined on AS(17/3eqW) than on AS(16/1eqW).
ꢁ
1
3
650 cm increases. This was observed for our Si/Al O materi-
2 3
[
16]
als
and corresponds to interacting AlꢁOH and SiꢁOH
groups. The acidic hydroxy groups of ASAs could be located in
[
25,26]
this area.
Even at high loadings, the band corresponding
to isolated silanols is still visible.
Figure 4 gives the spectra for AS(16/1eqW) and AS(17/
3
eqW), which were prepared in the presence of water. The
spectrum of AS(16/1eqW) (entry 8) is similar to that of entry 4
with 10% Al O ) in Figure 3, which has a similar coverage of
The integration of the peaks for CO adsorption on the
ꢁ
1
strong LAS (2230 cm ) and the weak/medium LAS
ꢁ
1
(
(2200 cm ), normalized by the weight of the pellet and the
surface area of the sample, yields a comparative estimate of
2
3
aluminum species (Figure 1). Similarly, the spectrum of AS(17/
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 835

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
2
Figure 6. IR spectra of CO adsorbed on Al/SiO prepared in the presence of
water: a) AS(16/1eqW) and b) AS(17/3eqW). Numbers in parentheses corre-
spond to entries in Table 1. Full spectra are shown in Figure S1.
Figure 7. Evolution of the fraction of strong Lewis acid sites (LAS) of the
total number of LAS. Numbers in parentheses correspond to entries in
Table 1.
Materials stemming from anhydrous synthesis show a de-
crease in the fraction of strong LAS as the coverage by alumi-
num species increases (Figure 7). Under such preparation con-
ditions, the first LAS are strong (entries 2 and 3). At a high alu-
minum loading, the percentage of strong LAS remains con-
stant and low. The percentage of strong LAS of AS(16/1eqW) is
consistent with its coverage of aluminum species. Given its
even lower coverage, the structure of the alumina deposit of
AS(17/3eqW) differs, with a low percentage of strong LAS. Pure
g-Al O (entry 10) shows no strong LAS.
2
Figure 5. IR spectra of CO adsorbed on Al/SiO prepared by anhydrous CLD:
a) silica, b) AS(3/anh), c) AS(6/anh), d) AS(10/anh), e) AS(15/anh), f) AS(25/
anh), and g) AS(35/anh). Numbers in parentheses correspond to entries in
Table 1. Full spectra are shown in Figure S1.
the number of sites. The fraction of strong LAS of the total LAS
[
16]
is estimated as follows [Eq. (3)]:
2
3
Fraction of strong LAS in total LAS
Aband strong LAS
ð3Þ
m-Xylene isomerization
¼ Aband strong LAS
þ Aweak=medium LAS
The grafting of aluminum species under anhydrous conditions
first led to an increase (entries 2–6) and then to a decrease
(entry 7) in the isosurface rate of m-xylene conversion (Table 4,
column 3). AS(16/1eqW) and AS(17/3eqW) showed values con-
sistent with their surface ratio of exposed aluminum. The rates
per exposed aluminum atom (Table 4, column 4) decreased
in which A is the area of the CO peak of species i, normalized
i
by the weight of the wafer and the surface area of the material
ꢁ
1
(
i=strong LAS: peak at 2230 cm ; weak/medium LAS: peak at
ꢁ
1
2
200 cm ).
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 836

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
ꢁ
1
bumps at approximately 3600 to 3650 cm that emerge from
a broader band of OH groups. These may result from the inter-
action of deposited aluminum with the silanols of the silica
2
Table 4. Conversion of m-xylene over Al/SiO -grafted samples.
Entry Sample
Rate of m-xylene converted at 10 min
Isosurface
Per Al atom
[ꢀ10 moleculeh
[3,32]
support.
ꢁ
2
ꢁ1
ꢁ2 [a]
ꢁ2
ꢁ1 [b]
[
ꢀ10 moleculeh nm
]
]
1
2
3
4
5
6
7
8
9
1
silica
2.9
–
AS(3/anh)
AS(6/anh)
AS(10/anh)
AS(15/anh)
AS(25/anh)
AS(35/anh)
AS(16/1eqW) 43.7
AS(17/3eqW) 30.8
g-alumina
20.7
31.2
47.1
59.1
78.9
42.2
34.5
23.6
25.0
10.3
11.2
4.1
16.3
37.4
0.6
Brønsted acid sites as active sites for ethanol dehydration
and m-xylene isomerization
Active sites for ethanol dehydration and m-xylene
isomerization
As probed by CO, grafting of the first aluminum species
(entry 2) causes a steep increase in the number of BAS (Fig-
ure 8a). With increasing aluminum loading, the number of BAS
increased more slowly.
0
7.9
[a] Based on the total surface area. [b] Based on the surface coverage of
aluminum species.
continuously with increasing surface coverage of aluminum
species. The highest rate was obtained on AS(17/3eqW).
During this reaction, m-xylene was either isomerized (I) or dis-
proportionated to toluene and trimethylbenzenes (D). The
latter can further form isomers of xylene. The ratio isomeriza-
tion/disproportionation increased with the isosurface catalytic
activity; AS(3/anh) displayed the lowest value (I/D=0.49) and
AS(15/anh) displayed the highest value (I/D=0.85).
Discussion
Surface occupancy and distribution of aluminum species on
silica
Anhydrous conditions during synthesis yield regular and uni-
form deposition of aluminum species on silica: the coverage is
proportional to the surface density of the grafted species. The
IR spectra of the CO stretching region also show the disap-
pearance of silanols physisorbing CO, which is directly related
to the coverage of aluminum species. Water during synthesis,
even at low concentration (1 equiv. per precursor molecule),
favors the polymerization of the precursor to aggregates,
which causes a lower coverage of the silica surface: the higher
the water concentration, the higher the degree of aggregation.
These findings agree with our previous findings based on
time-of-flight secondary ion mass spectrometry, NMR spectros-
Figure 8. Evolution of the number of a) Brønsted acid sites (BAS) and
b) strong Lewis acid sites (LAS) (c) and weak or medium LAS (a) of Al/
SiO , probed by CO adsorption followed by IR spectroscopy. Numbers in pa-
2
[
5]
copy, and TEM data.
The IR spectra of the OH stretching region (Figure 3) show
that grafting of the silica surface caused a disruption of
H-bonding among the silanol groups. Considering a “chain”
rentheses correspond to entries in Table 1.
[
3]
model of silanols that interact with each other, the grafting
of aluminum species not only consumes silanols by condensa-
tion with the OH groups, but it also disturbs the long-distance
interaction of these groups. When this occurs, the remaining
silanols are, for the most part, isolated; they remain visible
even at high aluminum loading, and this confirms that the
coverage onto silica is not complete. New OH groups, typically
found on ASAs, become visible in the OH region upon grafting
of enough aluminum. Interestingly, three of the samples (en-
tries 2, 8, and 9), which have low alumina coverage, show
For samples prepared under anhydrous conditions, the
number of strong LAS first increased [particularly pronounced
between entry 2 (3% Al O ) and entry 3 (6% Al O )] and then
2
3
2
3
decreased above 10% Al O (Figure 8b). Up to 10% Al O , no
2
3
2
3
weak/medium LAS were visible. Although the number of
strong LAS decreased, weak and medium LAS appeared and
their number increased. Thus, increasing the aluminum loading
caused a weakening of the LAS. Contrary to the BAS, the evo-
lution of the number of LAS was strongly related to the struc-
ture of the deposit. Samples prepared in the presence of water
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 837

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
on sites other than weak LAS.
The evolution of the number of
strong LAS, which was highest
at approximately 10% Al O , did
2
3
not parallel the amount of dehy-
drated ethanol, which was high-
est at approximately 25% Al O .
2
3
Thus, strong LAS cannot be the
exclusive sites for ethanol dehy-
dration.
It can thus be postulated that
the ethanol dehydration sites are
a combination of BAS and LAS.
Figure 9 gives a good correlation
between the number of BAS and
the amount of ethanol dehydra-
tion sites for entries 1–6, 8,
and 9.
Figure 9. Evolution of the m-xylene isomerization conversion rate (c), the amount of dehydrated ethanol
(
a), and the number of Brønsted acid sites (BAS), as probed by CO (g). Numbers in parentheses correspond
to entries in Table 1.
showed a high number of LAS, which were strong LAS, even at
In another study, we showed that on Si/Al O the tempera-
2
3
low coverage. g-Al O3 (entry 10) showed neither BAS nor
ture of the ethanol dehydration peak is associated with the
2
[16]
strong LAS but a large number of weak LAS.
nature of the active site. In the present study, all the samples
prepared under anhydrous conditions showed similar ethanol
dehydration temperatures (ꢀ2708C). Thus, the ethanol dehy-
dration sites of entry 2, with BAS but almost no LAS, are the
same as those of entries 3 to 6. Furthermore, the dehydration
temperature was higher than that of g-Al O (2528C), for which
[
33–37]
The isomerization of m-xylene requires BAS.
This was
confirmed by the good correlation for entries 1–6, 8, and 9 be-
tween the rate of m-xylene conversion and the number of BAS
as a function of the coverage of aluminum species (Figure 9):
the BAS of Al/SiO are responsible for m-xylene isomerization.
2
2
3
[16]
g-Al O , which has no BAS, did not catalyze m-xylene isomeri-
ethanol dehydration occurs on LAS. Therefore, we assume
2
3
zation. At high aluminum loading, AS(35/anh) (entry 7) showed
discrepant results: Whereas the number of BAS probed by CO
still increased, the rate of conversion of m-xylene decreased.
The reason for this is not immediately clear.
that the BAS of Al/SiO , detected by CO adsorption and
2
probed by ethanol dehydration, are the active sites for
m-xylene isomerization and that for samples 2–6 their number
is probed by the amount of dehydrated ethanol.
The dehydration of ethanol can take place on both Lewis
[16]
acid sites of alumina and Brønsted acid sites of Si/Al O .
2
3
Turnover frequency for m-xylene isomerization
AS(3/anh) (entry 2) was active in both ethanol dehydration and
m-xylene isomerization (Figure 9), despite the absence of weak
LAS and a low number of strong LAS (Figure 8b). Conversely,
the number of BAS in this sample is high, despite its low alu-
minum loading (Figure 8a). Although entries 2 and 3 do not
have weak LAS, these materials still dehydrated ethanol
We already showed that the BAS of ASAs dehydrate ethanol,
which enabled us to estimate their number and to calculate
[16]
their TOF for m-xylene isomerization. On the basis of the
number of ethanol dehydration sites (Table 5, column 3) and
by assuming that they are exclusively BAS, the TOF of BAS in
ꢁ
1
ꢁ1
(Figure 9). This showed that ethanol dehydration can be done
s site was calculated as follows [Eq. (4)]:
Table 5. Density of m-xylene conversion sites and ethanol conversion sites over Al/SiO
2
-grafted samples.
BAS
Entry
Sample
EtOH dehyd. sites
TOF
Fraction of BAS
in EtOH dehyd.
sites [%]
Surface coverage
of Al species
[%]
Isosurface
Per exposed Al
Per exposed Si
ꢁ
2
ꢁ4 ꢁ1
ꢁ1
ꢁ2 [a]
ꢁ2 [b]
ꢁ2 [b]
[c]
[nm
]
[ꢀ10
s
site
]
[nm
]
atom [ꢀ10
]
atom [ꢀ10
]
2
3
4
5
6
7
8
9
AS(3/anh)
AS(6/anh)
0.12
0.20
0.26
0.45
0.61
0.50
0.38
0.27
4.7
4.4
5.0
3.6
3.7
–
0.14
0.20
0.31
0.39
0.52
0.28
0.29
0.20
22.5
15.4
16.3
6.7
7.3
2.7
1.1
1.8
2.9
5.7
9.4
12.8
2.9
1.7
ꢀ100
ꢀ100
ꢀ100
ꢀ100
ꢀ100
ꢀ5
11
15
46
56
83
21
7
AS(10/anh)
AS(15/anh)
AS(25/anh)
AS(35/anh)
AS(16/1eqW)
AS(17/3eqW)
[
[
[
d]
e]
e]
55
76
73
–
–
10.7
24.5
ꢁ
4
ꢁ1
ꢁ1
[a] Calculated on the basis of an average TOF value of 4.25ꢀ10
s
site obtained on samples 2–6 for m-xylene conversion. Values were obtained by as-
suming that all the EtOH dehydration sites of these materials are BAS [Eq. (4)]. [b] Calculated on the basis of the surface coverage of aluminum species.
c] Calculated by dividing the number of BAS by the total number of dehydration sites. [d]<100 as a result of accessibility: a higher number of sites is ac-
cessible to EtOH than to m-xylene. [e]<100 as a result of the composite nature of the EtOH dehydration sites.
[
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 838

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
TOF ½sꢁ site ꢂ
1
ꢁ1
ure 8b). Nonetheless, BAS are major active sites and represent
between 73 and 76% of the dehydrated ethanol.
ð4Þ
rate of m-xylene converted at 10 min
1
3600
¼
ꢃ
Table 5, columns 6 and 7 give the density of BAS per surface
aluminum and silicon atoms. If the surface density of alumi-
num species increases, the density of silicon atoms decreases;
thus, the number of BAS per surface silicon increases. Given
that the number of surface aluminum species increases faster
than the number of BAS, the number of BAS per surface alumi-
num decreases. At low aluminum loading [AS(3/anh)], one out
of five aluminum atoms is related to a BAS.
EtOH dehydration sites
ꢁ1
ꢁ2
in which the rate is in [moleculeh nm ].
Entries 2 to 6, that is, materials showing consistent results
between the number of BAS probed by CO, the amount of de-
hydrated ethanol, and the rate of conversion of m-xylene
(
Figure 9), and having similar ethanol dehydration tempera-
There is a strong correlation, even for samples with dissimi-
lar surface structures (entries 1–6, 8, and 9), between the
number of BAS, probed by different means, and the coverage
by aluminum species (Figure 9 and Table 5). Conversely, de-
spite a similar coverage, the number of LAS is higher on mate-
rials with alumina clusters (entries 8 and 9) than on materials
with well-dispersed aluminum species (entries 3 and 4) (Fig-
ure 8b). Thus, BAS are associated with well-dispersed alumi-
num species, whereas clusters favor the presence of LAS.
tures (Figure 2a), yielded comparable TOFs (Table 5, column 4),
ꢁ4
ꢁ1
the average of which (4.3ꢀ10
5 times lower than that obtained for a USY zeolite without
extra-framework aluminum (CBV720 from Zeolyst): 3.1ꢀ
s
per BAS) is approximately
7
ꢁ
2
ꢁ1
[16]
10
s
per BAS.
Assuming that all the Al/SiO2 materials
have the same TOF, this average value enabled the calculation
of the number of BAS (Table 5, column 5). The number of BAS
in entries 2 to 6 is the same as the number of ethanol dehydra-
tion sites. For entries 7 to 9, the number of dehydration sites is
higher than the number of BAS. A non-negligible percentage
(
24 to 45%; Table 5, column 8) of the ethanol dehydration sites
Comparison with Si/Al O grafting and zeolite
2
3
of these materials is not active in m-xylene isomerization (see
below).
Table 6 gives a comparison between the main features of Al/
On AS(35/anh) (entry 7), this difference might be due to the
accessibility of the reactant molecules to the active sites:
Figure 9 shows that although CO probes a higher number of
BAS if the alumina loading was increased from 25 (entry 6) to
SiO2 and Si/Al O3 from our previous study and those of a
2
H-USY zeolite that is free of extra-framework aluminum (EFAL)
[16]
species. In both Si/Al O and Al/SiO , the number of BAS de-
2
3
2
creases at high loading of grafted species, but for different rea-
sons: On Si/Al O , new species are grafted on top of the BAS
3
5% Al O (entry 7), ethanol dehydration and m-xylene isomer-
2 3
2
3
ization indicate an opposite trend. The adsorption of 2,6-luti-
dine, with a similar size to that of m-xylene, followed by IR
spectroscopy, also showed a decrease in the number of BAS if
the alumina loading was increased from entry 6 to entry 7
at the highest loadings, whereas on Al/SiO , the accessibility of
2
large molecules, such as m-xylene, to the BAS might be diffi-
cult if the alumina deposit is too thick. The number of BAS per
2
nm is generally higher on Al/SiO than on Si/Al O , but these
2
2
3
(Supporting Information, Figures S2 and S3).
sites are less active: the TOF for m-xylene isomerization is
AS(16/1eqW) (entry 8) and AS(17/3eqW) (entry 9) both
3.3 times lower than that on Si/Al O and 75 times lower than
2
3
showed a higher ethanol dehydration temperature than the
materials prepared under anhydrous conditions as well as
asymmetric ethanol dehydration peaks. This indicates that the
active sites for ethanol dehydration are different to those of
entries 2 to 7. We assume that these samples have additional
sites, other than BAS, that dehydrate ethanol and contribute to
a composite ethanol peak resulting from their combined ef-
fects. These sites might be LAS, which are more numerous on
water-prepared samples than on the other samples (Fig-
that on a USY zeolite without EFAL. This is also reflected by
the average ethanol dehydration temperature (2698C), which
is higher than that of the BAS of Si/Al O (2488C) and much
2
3
higher than that of the zeolite (2118C). The dehydration peak
is also broader on Al/SiO , which suggests a larger spread in
2
local structure. On g-alumina, the LAS on the (100) surface are
responsible for the large number of ethanol dehydration sites.
Calculation of the rates of m-xylene conversion per unit of sur-
face stresses the impact of the high surface area of zeolites:
2 2 3
Table 6. Main catalytic properties of Al/SiO and Si/Al O materials prepared under anhydrous conditions and H-USY without (w/o) extra-framework
aluminum (EFAL).
[
3
16]
[16]
Al/SiO
2
Si/Al
2
O
H-USY w/o EFAL (CBV720)
Evolution of the number of BAS upon grafting guest species under mild conditions
increases
0.52
AS(25/anh)
0.61
goes through a max.
0.18
SA(17/4eqW)
0.83
–
0.18
ꢁ2
Max. number of BAS [nm
Sample name
Max. number of EtOH dehyd. sites [nm
Sample name
]
ꢁ
2
]
0.18
AS(25/anh)
pure g-alumina
ꢁ
1
ꢁ1
ꢁ4
ꢁ3
ꢁ2
TOF for m-xylene isomerization [s active site
]
4.3ꢀ10
1.4ꢀ10
3.1ꢀ10
2640
ꢁ
1
catꢁ1
Max. rate of m-xylene converted at 10 min [mmolh
g
]
43
29
Sample name
AS(25/anh)
269
SA(17/4eqW)
248
Average EtOH dehyd. temperature (if exclusively catalyzed by BAS) [8C]
211
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 839

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
The isoweight rate of the USY zeolite is 61 and 91 times higher
than that on Al/SiO and Si/Al O , respectively.
with a flame ionization detector (FID) and an FFAP column. The cal-
culation of the rate of conversion was based on the conversion of
m-xylene into its products (o- and p-xylene, toluene, and trimethyl-
benzenes) after 10 min on-stream [Eq. (6)]:
2
2
3
Conclusions
Rate ½molecule hꢁ nm
1
ꢁ2
"
!#
Controlled deposition of aluminum species on silica occurred
under anhydrous conditions and created Brønsted acid sites,
the number of which increased with the aluminum loading.
Ethanol dehydration followed by thermogravimetry, combined
with IR spectroscopy after CO adsorption, yielded the number
of Brønsted acid sites and the TOF of these sites for m-xylene
species
species
X
X
Ai
NC atoms i
Ai
A
mꢁxylene
8
¼
=
þ
NC atoms i
ð6Þ
i
i
flowrate ꢃ density
NA
ꢃ S
mꢁxylene
ꢃ
M
ꢃ m
mꢁxylene
catalyst
BET
ꢁ4
ꢁ1
ꢁ1
isomerization: 4.3ꢀ10
s
site . This value is 3.3 times lower
in which A
is the area of the GC peak of species i (i=toluene,
i
p-xylene, o-xylene, trimethylbenzenes), flow rate is in [cm h ], N
is Avogadro’s number (6.02ꢀ10 mol ), and S [nm g ] is the
surface area of the catalyst.
3
ꢁ1
than that of Si/Al O and 75 times lower than that of a USY
A
2
3
2
3
ꢁ1
ꢁ2 ꢁ1
BET
zeolite without extra-framework aluminum. This confirms that
the nature of the Brønsted acid sites strongly depends on the
method of ASA synthesis. Strong Lewis acid sites also ap-
peared during grafting; they became weaker as the concentra- Acknowledgements
tion of the guest species was increased. Under aqueous condi-
tions, grafting was less controlled and alumina agglomerates
possessing weak and medium Lewis acid sites but no Brønsted
acidity were formed.
The authors thank Anne-Agathe Quoineaud, Emmanuel Soyer,
and Laurent Lemaꢂtre from IFP Energies nouvelles for CO IR spec-
troscopy measurements. We acknowledge financial support from
IFP Energies nouvelles.
Experimental Section
Keywords: acidity · coverage degree · ethanol · silicates ·
thermogravimetry
For IR spectroscopy studies, samples were pressed into pellets and
ꢁ5
pretreated in situ under vacuum (10 mbar) for 10 h at 4508C
ꢁ
1
(
heating rate of 58Cmin ) including a plateau at 1508C for 1 h.
[
1] a) D. G. Poduval, J. A. R. van Veer, M. S. Rigutto, E. J. M. Hensen, Chem.
Commun. 2010, 46, 3466–3468; b) E. J. M. Hensen, D. G. Poduval, V. De-
girmenci, D. A. J. M. Ligthart, W. Chen, F. Maugꢂ, M. S. Rigutto, J. A. R.
van Veen, J. Phys. Chem. C 2012, 116, 21416—21429.
Samples were then cooled down at liquid-nitrogen temperature
and increasing amounts of CO, up to a total amount of 600 mmol,
were admitted into the chamber. Spectra were recorded after each
dose of CO, by a Nexus Fourier transform apparatus (resolution
[2] C. Chizallet, P. Raybaud, Angew. Chem. 2009, 121, 2935–2937; Angew.
Chem. Int. Ed. 2009, 48, 2891–2893.
ꢁ
1
4
cm ).
[
3] G. Crꢂpeau, V. Montouillout, A. Vimont, L. Mariey, T. Cseri, F. Maugꢂ, J.
Phys. Chem. B 2006, 110, 15172–15185.
Adsorption and desorption of ethanol, followed by thermogravim-
etry, were performed in a similar manner as the method described
[17]
[16]
[4] M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A. Guercio, R.
by Kwak et al. and previously used for our Si/Al O materials.
2
3
Catani, R. J. Willey, J. Catal. 1998, 179, 581–596.
Samples were first activated under a flow of helium at 5008C for
h. After cooling to room temperature, helium was passed
[
5] M. Caillot, A. Chaumonnot, M. Digne, C. Poleunis, D. P. Debecker, J. A.
van Bokhoven, Microporous Mesoporous Mater. 2014, 185, 179–189.
6] R. Mokaya, W. Jones, Z. H. Luan, M. D. Alba, J. Klinowski, Catal. Lett.
1996, 37, 113–120.
2
through a saturator of ethanol for approximately 30 min. Ethanol
molecules on weak adsorption sites were then evacuated by purg-
ing with helium at room temperature for 1 h. Strongly adsorbed
molecules were desorbed during the thermoprogrammed desorp-
tion of ethanol (TPD), from room temperature to 4008C, at a heat-
[
[7] R. Ryoo, S. Jun, J. M. Kim, M. J. Kim, Chem. Commun. 1997, 2225–2226.
[
8] A. Auroux, R. Monaci, E. Rombi, V. Solinas, A. Sorrentino, E. Santacesaria,
Thermochim. Acta 2001, 379, 227–231.
[9] M. V. Landau, E. Dafa, M. L. Kaliya, T. Sen, M. Herskowitz, Microporous
Mesoporous Mater. 2001, 49, 65–81.
ꢁ
1
ing rate of 108Cmin . The intensity of the two main desorption
features (ethanol and ethylene) was measured by calculating the
first derivative of the weight loss during TPD with respect to tem-
perature. The amount of dehydrated ethanol was calculated as fol-
lows [Eq. (5)]:
[
[
10] M. Rozwadowski, M. Lezanska, J. Wloch, K. Erdmann, J. Kornatowski,
Preparation and catalytic characterization. of Al-grafted MCM-48 materi-
als, 2002, pp. 755–762.
11] A. Martins, J. M. Silva, J. Blanchard, P. Massiani, M. Breysse, F. R. Ribeiro,
M. F. Ribeiro, React. Kinet. Catal. Lett. 2004, 82, 139–147.
Ethanol dehydrated ½nmꢁ
2
ꢂ
[12] X. Y. Chen, M. Jia, G. Z. Liu, X. W. Zhang, L. Wang, Z. T. Mi, Appl. Surf. Sci.
010, 256, 5856–5861.
[13] R. Mokaya, W. Jones, Chem. Commun. 1997, 2185–2186.
2
Apeakethylene
ð5Þ
¼
ethanol retained after purge ꢃ A
peakethylene
þ Apeakethanol
[
14] E. J. M. Hensen, D. G. Poduval, P. C. C. M. Magusin, A. E. Coumans, J. A. R.
van Veen, J. Catal. 2010, 269, 201–218.
[
15] A. Chaumonnot in Catalysis by Transition Metal Sulfides: From Molecular
Theory to Industrial Application (Eds.: H. Toulhoat, P. Raybaud), Editions
Technip, Paris, 2013, pp. 225–247.
in which Apeak i is the area of the peak corresponding to the de-
sorption of species i upon calculating the derivative of the weight
loss of ethanol during TPD.
[
16] M. Caillot, A. Chaumonnot, M. Digne, J. A. van Bokhoven, ChemCatChem
The m-xylene isomerization test was performed by passing gas-
2
013, 5, 3644–3656.
17] J. H. Kwak, D. H. Mei, C. H. F. Peden, R. Rousseau, J. Szanyi, Catal. Lett.
011, 141, 649–655.
3
ꢁ1
eous m-xylene (0.6 cm h ) through a bed of catalyst (0.5 g) at
508C. The catalyst had been preheated in an air flow at 3508C.
[
3
2
Analysis of the products was performed by online gas chromatog-
raphy by means of a Hewlett Packard 6850 apparatus, equipped
[18] S. Roy, G. Mpourmpakis, D. Y. Hong, D. G. Vlachos, A. Bhan, R. J. Gorte,
ACS Catal. 2012, 2, 1846–1853.
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 840

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[
[
[
19] W. E. Farneth, R. J. Gorte, Chem. Rev. 1995, 95, 615–635.
20] C. B. Phillips, R. Datta, Ind. Eng. Chem. Res. 1997, 36, 4466–4475.
21] J. N. Kondo, K. Ito, E. Yoda, F. Wakabayashi, K. Domen, J. Phys. Chem. B
[30] O. Cairon, T. Chevreau, J. C. Lavalley, J. Chem. Soc. Faraday Trans. 1998,
94, 3039–3047.
[31] O. Cairon, Phys. Chem. Chem. Phys. 2010, 12, 6333–6336.
[32] B. Xu, C. Sievers, J. A. Lercher, J. A. R. van Veen, P. Giltay, R. Prins, J. A.
van Bokhoven, J. Phys. Chem. C 2007, 111, 12075–12079.
[33] V. S. Nayak, V. R. Choudhary, Appl. Catal. 1982, 4, 333–352.
[34] J. Pꢂrez-Pariente, E. Sastre, V. Fornes, J. A. Martens, P. A. Jacobs, A.
Corma, Appl. Catal. 1991, 69, 125–137.
2
005, 109, 10969–10972.
22] R. Barthos, A. Szechenyi, F. Solymosi, J. Phys. Chem. B 2006, 110, 21816–
1825.
[
2
[23] S. Golay, R. Doepper, A. Renken, Appl. Catal. A 1998, 172, 97–106.
[24] B. A. Morrow, A. J. McFarlan, J. Phys. Chem. 1992, 96, 1395–1400.
[25] M. I. Zaki, H. Knøzinger, Mater. Chem. Phys. 1987, 17, 201–215.
[26] F. Leydier, C. Chizallet, A. Chaumonnot, M. Digne, E. Soyer, A. A. Quoi-
neaud, D. Costa, P. Raybaud, J. Catal. 2011, 284, 215–229.
[
[
[
35] G. Mirth, J. Cejka, J. A. Lercher, J. Catal. 1993, 139, 24–33.
36] C. W. Jones, S. I. Zones, M. E. Davis, Appl. Catal. A 1999, 181, 289–303.
37] S. Al-Khattaf, M. N. Akhtar, T. Odedairo, A. Aitani, N. M. Tukur, M. Kubu,
Z. Musilova-Pavlackova, J. Cejka, Appl. Catal. A 2011, 394, 176–190.
[27] W. Daniell, U. Schubert, R. Glockler, A. Meyer, K. Noweck, H. Knozinger,
Appl. Catal. A 2000, 196, 247–260.
[
28] A. Zecchina, E. E. Platero, C. O. Arean, J. Catal. 1987, 107, 244–247.
29] T. P. Beebe, P. Gelin, J. T. Yates, Surf. Sci. 1984, 148, 526–550.
Received: September 27, 2013
Published online on January 21, 2014
[
ꢁ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 832 – 841 841