Synthesis of highly active sulfated zirconia by sulfation with SO 3

Article abstract of DOI:10.1016/j.jcat.2005.11.039

Highly active sulfated zirconia (SZ) materials were prepared by sulfation of crystalline zirconia with gaseous SO₃. This direct method circumvents the use of liquid sulfation agents, as well as a final calcination step. The amount of sulfur species retained by SO₃-SZ samples was directly proportional to the fraction of monoclinic phase and the concentration of Bronsted acid sites. However, the catalytic activity in n-butane isomerization at 373 K was proportional to the concentration of labile chemisorbed SO₃, most likely in the form of pyrosulfate species showing a characteristic IR band at 1404 cm^-1. Tetragonal zirconia favors the formation of these labile sulfur species, playing a key role in light alkane activation.

Full text of DOI:10.1016/j.jcat.2005.11.039

Journal of Catalysis 238 (2006) 39–45
www.elsevier.com/locate/jcat
Synthesis of highly active sulfated zirconia by sulfation with SO3
1
2
∗
Xuebing Li , Katsutoshi Nagaoka , Roberta Olindo, Johannes A. Lercher
Technische Universität München, Lehrstuhl für Technische Chemie II Lichtenbergstraße 4, 85747 Garching, Germany
Received 26 October 2005; revised 25 November 2005; accepted 28 November 2005
Available online 22 December 2005
Abstract
Highly active sulfated zirconia (SZ) materials were prepared by sulfation of crystalline zirconia with gaseous SO . This direct method cir-
3
cumvents the use of liquid sulfation agents, as well as a final calcination step. The amount of sulfur species retained by SO -SZ samples was
3
directly proportional to the fraction of monoclinic phase and the concentration of Brønsted acid sites. However, the catalytic activity in n-butane
isomerization at 373 K was proportional to the concentration of labile chemisorbed SO , most likely in the form of pyrosulfate species showing
3
−
1
a characteristic IR band at 1404 cm . Tetragonal zirconia favors the formation of these labile sulfur species, playing a key role in light alkane
activation.
©
2005 Elsevier Inc. All rights reserved.
Keywords: Sulfated zirconia; Butane isomerization; SO sulfation; Activity; Labile sulfate
3
1
. Introduction
tive [6,7]; and (iii) a final calcination at high temperatures is a
crucial step in the formation of active sites [8–10].
In contrast, a recent study found aqueous sulfation of crys-
talline zirconia stabilized in the tetragonal or cubic forms by
formation of a solid solution with Y2O3 to be an effective
method for preparation of catalytically active SZ [11]. In agree-
ment with the literature, sulfated monoclinic zirconia was re-
ported to be hardly active for butane isomerization.
One-step synthesis of moderately active monoclinic SZ cat-
alyst was reported by Stichert et al. [12]. However, the corre-
sponding sample exhibited a lower catalytic activity than tetrag-
onal materials, and the authors concluded that even if the crystal
phase plays a role with respect to the activity of SZ, it is not as
crucial as generally thought. Synthesis of active monoclinic SZ
catalysts has also been reported by others [13,14]; however, the
catalysts had nonnegligible amounts of tetragonal zirconia and
in every case the catalytic activity of the monoclinic samples
was much lower than that of tetragonal SZ.
Sulfated zirconia (SZ) and other sulfated metal oxides have
been studied for more than two decades because of their high
catalytic activity for activation of short alkanes at low temper-
atures. The catalytic properties of sulfated metal oxides de-
pend strongly on subtle variations in the material. The influ-
ence of preparation parameters on catalytic activity has been
thoroughly discussed in the literature, particularly the zirconia
precursor [1], the sulfation procedure [2], and the calcination
treatment [3,4]. Three conditions are generally recognized as
indispensable for the preparation of an active catalyst: (i) the
treatment with liquid sulfating agents must be performed on
amorphous zirconium hydroxide, the sulfation of crystalline
zirconia being ineffective [5]; (ii) only the tetragonal or cubic
ZrO phases are active, whereas the monoclinic is nearly inac-
2
Whatever the ZrO phase used, a final calcination step has
2
*
Corresponding author. Fax: +49 0 89 289 13544.
E-mail addresses: roberta.olindo@ch.tum.de (R. Olindo),
been recognized as being critical for the formation of active
sites. Some claim that high-temperature treatment forms active
sites by effectively binding sulfate groups to the zirconia sur-
face [8,9]. Others proposed that another important function of
the calcination step is the partial removal of sulfate species from
highly uncoordinated sites on ZrO2, creating strong Lewis acid
johannes.lercher@ch.tum.de (J.A. Lercher).
1
Present address: Department of Chemical Engineering, University of Cali-
fornia at Berkeley, Berkeley, CA 94720, USA.
2
Present address: Department of Applied Chemistry, Faculty of Engineering,
Oita University Dannoharu 700, Oita 870-1192, Japan.
0021-9517/$ – see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.11.039

40
X. Li et al. / Journal of Catalysis 238 (2006) 39–45
sites. The strength of these (potentially catalytically active) acid
sites is said to be enhanced by the electron-withdrawing effect
of the neighboring sulfate groups [10,11].
Preparation of SZ by sulfation of crystalline metal oxides
with gaseous SO3 has been reported. Yamaguchi et al. [15]
demonstrated that strong acid sites are generated on Fe2O3 by
sulfation with SO3 at 573 K, with the final catalyst being active
for cyclopropane isomerization. Recently, Haw et al. [16,17]
noted the formation of Brønsted acid sites on zirconia by SO3
chemisorption using NMR spectroscopy and theoretical stud-
ies.
The present study explores new materials based on sul-
fated crystalline zirconia with gaseous SO3. The procedure that
we have developed is the first to allow us to independently
vary solid-state properties, such as phase composition and sul-
fate loading, which can provide insight into more conventional
sulfate-based catalysts. In addition, materials prepared by in
situ sulfation with SO3 were compared with materials prepared
by ex situ sulfation, to investigate the influence of very labile
sulfate groups on the catalytic activity for n-butane isomeriza-
tion.
Fig. 1. Scheme of sulfation ex situ with gaseous SO on crystalline zirconia
3
materials.
Sulfation of SZ with SO3 was also performed in the reactor
(
“in situ sulfation”), to investigate the influence of rehydration
and activation between sulfation and catalytic tests. The in situ
sulfated SZ sample is denoted as SZ-SO3 in situ. First, 0.2 g of
−1
SZ was heated for 2 h at 673 K (at a rate of 10 K min ) in He
(
−1
10 ml min ) and then cooled to 373 K. Subsequently, 50 ml
of He saturated with SO3 at room temperature were flushed
2
. Experimental
−1
over the catalyst at 373 K. After purging with He (10 ml min
)
for 0.5 h, n-butane isomerization was studied at this tempera-
ture.
2
.1. Catalyst preparation
Sulfate-doped zirconium hydroxide was obtained from Mag-
2
.2. Catalyst characterization
nesium Electron (XZO 1077/01). The received material was
−
1
calcined up to 873 K at a rate of 10 K min in static air and
kept at this temperature for 3 h. The final sample is denoted as
SZ henceforth.
The sulfate content of the catalysts was determined by ion
chromatography (IC) as described previously [18]. Typically,
0.02 g of SZ was suspended in 20 ml of 0.01 M NaOH for
20 min. The solution was filtered through a 0.45-µm filter, and
the sulfur content in the liquid was determined by IC using a
Metrohm 690 ion chromatograph. The BET surface area and the
pore volume of the samples were determined by physisorption
of N2 at 77 K using a PMI automated BET-sorptometer.
X-Ray diffraction (XRD) patterns were collected with a
Philips X’Pert-1 XRD powder diffractometer using Cu-Kα ra-
diation. The fraction of tetragonal zirconia (Xt) was calculated
using the following equation:
To wash this calcined material, 20 g of SZ was suspended in
00 ml of bi-distilled water for 20 min and then filtered. This
washing procedure was repeated 3 times. Finally, the filter cake
was dried at room temperature. The resulting powder is denoted
as SZ-WW.
The sulfate species were extracted from 20 g of SZ using
00 ml of NaOH 0.05 M. After filtration, the cake was washed
carefully for several times with bi-distilled water to remove the
residual NaOH. The resulting material is pure zirconia, denoted
as SZ-BW.
4
4
Plain zirconia (ZrO2) used in this study was prepared by
calcination of commercial zirconium hydroxide (Zr(OH)4,
Aldrich) in flowing air at 873 K for 3 h at an increment of
¯
Xt = It(111)/(Im(111) + It(111) + Im(111)).
−1
IR spectra of the samples were collected using a Bruker IFS
8 (or, alternatively, a Perkin–Elmer 2000) spectrometer at
cm resolution. Self-supporting wafers with a density of 5–
0 mg cm were prepared. The wafers were placed into in a
1
0 K min
Sulfation by gaseous SO3 was performed in a quartz tube
“ex situ sulfation” with respect to the catalytic tests). Be-
fore gaseous SO3 was admitted, SZ or zirconia (pellets, 355–
10 µm, 0.5–2 g) were preheated in the quartz tube under N2
.
8
4
1
−
1
(
−
2
7
stainless steel cell with CaF2 windows, gradually heated at a
−1
−1
−1
flow (50 ml min ) at 673 K for 2 h, then cooled to 473 K.
The SO3 cylinder was then filled with 50 ml of N2 via valve 1
rate of 10 K min to 673 K in He flow (10 ml min ), and
held at that temperature for 2 h. A spectrum was recorded after
the temperature was stabilized at 373 K. For adsorption of pyri-
dine, the experiments were conducted in an evacuable cell. The
(
Fig. 1) and subsequently connected to the flow system. SO3-
saturated nitrogen (50 ml) was flown through the quartz tube
−
1
−1
sample wafer was heated at a rate of 10 K min to 673 K and
together with the carrier gas (N2, 50 ml min ) for 1 h. Samples
of all materials were sulfated following the above-described
procedure and finally stored in a desiccator. We call these sam-
ples SZ-SO3, SZ-WW-SO3, SZ-BW-SO3, and ZrO2–SO3.
−
6
held at that temperature (base pressure of 10 mbar) for 2 h.
The sample was exposed to 0.1 mbar of pyridine at 373 K and
outgassed at the same temperature.

X. Li et al. / Journal of Catalysis 238 (2006) 39–45
41
2
.3. Butane isomerization
n-Butane isomerization was carried out in a quartz micro-
tube reactor (8 mm i.d.) under atmospheric pressure. 0.2 g of SZ
pellet (355–710 µm) was loaded into the reactor and activated
−
1
in situ at 673 K for 2 h in He flow (10 ml min ). The catalyst
was cooled to 373 K, and the reactant mixture (5% n-butane in
−
1
He; total flow, 20 ml min ) was fed through the catalyst bed.
Butane (99.5%, Messer) was passed through an olefin trap con-
taining activated H-Y zeolite (20 g) before being mixed with
He. After purification, even traces of butenes were not detected
in the reactant mixture (detection limit, 1 ppm). The reaction
products were analyzed using an on-line HP 5890 gas chro-
matograph equipped with a capillary column (plot Al2O3, 50 m
Fig. 3. n-Butane reaction rate versus time on stream on sulfated zirconia at
373 K (Q) SZ-BW-SO and (") ZrO –SO .
3
2
3
×
0.32 mm × 0.52 µm) connected to a flame ionization detec-
tor.
3
. Results
3
.1. Catalytic activity for n-butane skeletal isomerization
Fig. 2 shows the catalytic activity versus time on stream for
n-butane skeletal isomerization at 373 K on the parent sample
SZ), the water-washed sample (SZ-WW), and the ex situ SO3
(
sulfated samples (SZ-SO3 and SZ-WW-SO3). SZ showed mod-
erate catalytic activity with a selectivity of 96% to isobutane.
Washing SZ with water (SZ-WW) removed the labile sulfate
fraction and resulted in a completely inactive sample [19]. Sul-
fation of SZ with gaseous SO3 (SZ-SO3) drastically increased
the catalytic activity, with the maximum activity of SZ-SO3 be-
ing around 10 times that of SZ. After sulfation of the inactive
SZ-WW sample with gaseous SO3, the catalytic activity was
lower, but close to that of SZ-SO3.
The ex situ SO3 sulfation technique was also applied to
the two zirconia samples containing high concentrations of the
monoclinic phase, ZrO2 (70% monoclinic) and SZ-BW (94%
monoclinic). The catalytic activity for n-butane isomerization
of the two corresponding samples (ZrO2–SO3 and SZ-BW-
SO3) at 373 K is shown in Fig. 3. Contrary to the previous
conclusion that the sulfation of crystalline monoclinic zirconia
Fig. 4. n-Butane reaction rate versus time on stream at 373 K on in situ SO
sulfated SZ.
3
is not capable of generating active sites for n-butane isomeriza-
tion [5], ZrO2–SO3 and SZ-BW-SO3 exhibited slightly higher
catalytic activity than that of SZ obtained by calcination of an
amorphous sulfated zirconium hydroxide.
Although the ex situ SO sulfated samples showed signif-
3
icantly enhanced catalytic activity compared with the parent
material, we speculated that some of the very labile sulfate
species could be lost during the activation treatment before test-
ing the catalytic activity. To investigate this aspect, samples of
the parent SZ were modified ex situ and in situ with SO , and
3
the catalytic activities for n-butane isomerization were com-
pared. As shown in Fig. 4, the SZ-SO in situ sample showed
3
very high initial catalytic activity, approximately two orders of
magnitude higher than the maximum activity of SZ. However,
all catalysts exposed in situ to SO deactivated rapidly, related
3
to oligomerization of olefins formed at higher concentration in
the more active samples.
3
.2. Physical properties
The BET area and sulfate content of the different zirconia
samples are reported in Table 1. Both calcined commercial SZ
2
−1
and ZrO had specific surface areas >100 m g . Washing
2
SZ with water removed approximately 40% of the initial sulfur
species [19], whereas washing with dilute NaOH completely re-
moved the sulfur species. As expected, sulfation with gaseous
Fig. 2. n-Butane reaction rate versus time on stream on sulfated zirconia sam-
ples at 373 K (1) SZ-WW, (E) SZ, (2) SZ-WW-SO and (F) SZ-SO . SO
gaseous sulfation was performed ex situ.
3
3
3

42
X. Li et al. / Journal of Catalysis 238 (2006) 39–45
Table 1
Physical properties of zirconia and sulfated zirconia samples
Sample
BET
area
Sulfate
Tetragonal Acid sites
−
1
2− phase
(mmol g
)
content SO
mmolg
4
)
2
−1
(m g ) (
−1
(%)
Brønsted
Lewis
SZ
SZ-SO
SZ-WW
SZ-WW-SO n.d.
SZ-BW-SO n.d.
ZrO
ZrO –SO
109
n.d.
n.d.
0.44
0.64
0.25
0.91
1.53
n.d.
100
100
58
58
6
0.050
0.071
0
0.113
0.135
0
0.106
0.073
0.163
0.054
0.037
0.206
0.089
3
3
3
123
n.d.
30
30
2
1.15
0.173
2
3
Fig. 6. Correlation among fraction of tetragonal phase T/(T + M), (") sulfur
content, (2) Brønsted acid site concentration BAS and (Q) maximum catalytic
activity for the SO sulfated samples: SZ-SO , SZ-WW-SO , ZrO –SO and
3
3
3
2
3
SZ-BW-SO .
3
of sulfate groups on almost pure monoclinic zirconia (SZ-BW-
SO3) was nearly twice that of tetragonal zirconia (SZ-SO3).
This indicates that the monoclinic phase is able to retain more
SO3 than the tetragonal phase, suggesting that the monoclinic
form may be more basic than the tetragonal.
3
.3. Characterization of acid base properties by in situ IR
spectroscopy
3
.3.1. IR spectra of surface hydroxyl groups
Fig. 7A compiles the IR spectra, normalized by the wafer
thickness, in the region of the surface OH stretching bands. Be-
fore the spectra were recorded, the samples were activated in
He at 673 K for 2 h. The IR spectrum of SZ showed a mod-
−
1
erately strong, asymmetric band at 3634 cm with a shoulder
−
1
−1
at 3660 cm and very weak bands at 3740 and 3710 cm . In
addition, a weak band located at 3578 cm was also observed.
Both samples that did not contain labile sulfate surface groups
−
1
(
SZ-WW and ZrO2) showed a very strong IR band at 3600–
Fig. 5. XRD profiles of sulfated zirconias and zirconias. T: tetragonal, M: mon-
oclinic.
−
1
3700 cm , assigned to bridging OH groups [21,22], as well
as OH bands between 3700 and 3800 cm assigned to termi-
−
1
SO3 increased the sulfate concentration on the parent material
from 0.44 to 0.64 mmol g (SZ-SO3). According to the surface
nal OH groups [21,22]. SO sulfation reduced the intensity of
the OH bands above 3600 cm , as shown in the spectra of SZ-
3
−
1
−1
area that a sulfate group occupies based on its kinetic diameter
3 3 3 2 3
SO , SZ-WW-SO , SZ-BW-SO , and ZrO –SO . The decrease
2
(
0.31 nm ) [20], the sulfate concentrations of all SO3 sulfated
in OH group concentration after SO sulfation is tentatively at-
3
−
1
samples exceed the monolayer coverage (0.58 mmol g for
tributed to the substitution of the surface OH groups of zirconia
by SO . Thus SZ-BW-SO , which has the highest sulfate con-
3 3
−
1
SZ-SO3 and 0.66 mmol g for ZrO2–SO3).
As shown in Fig. 5, the zirconia in the SZ sample was
mainly tetragonal. Water washing reduced the fraction of the
tetragonal phase and induced transformation to the mono-
clinic phase (42%) [19]. Moreover, after removal of all sulfate
groups by washing with NaOH, the resulting material, SZ-
BW, was nearly exclusively monoclinic (94%). Sulfation with
SO3 did not influence the XRD pattern of the parent sample
SZ-BW.
Fig. 6 shows that the sulfate content of samples after SO3
sulfation was directly proportional to the fraction of monoclinic
phase present. The higher the fraction of monoclinic phase, the
higher the concentration of sulfate groups. The concentration
3
tent among the SO -sulfated samples, did not exhibit OH bands
−
1
above 3600 cm . The samples containing labile sulfate species
−
1
showed a band at 3570 cm
.
3.3.2. IR spectra of surface sulfate groups
The IR spectra of sulfate groups of the samples activated
at 473 K are compiled in Fig. 7B. All of the samples showed
stretching vibrations of the S=O group between 1300 and
−
1
1450 cm . In accordance with previous assignations, the
bands are attributed to highly covalent sulfate groups [23,
24]. The IR spectrum of SZ had the maximum of this band
−
1
at 1404 cm . After sulfation with SO3 (trace SZ-SO3), the

X. Li et al. / Journal of Catalysis 238 (2006) 39–45
43
Fig. 7. IR spectra in the regions of surface OH group (A) and surface sulfate group (B) of sulfated zirconias and zirconia.
corresponding band was broader, but otherwise unchanged.
A broad band appeared simultaneously in the 1300–1200 cm
region.
concentrations of Brønsted and Lewis acid sites derived in this
way are compiled in Table 1. For quantification, the molar ab-
sorption coefficients of the bands of adsorbed pyridine were set
equal to those determined for zeolites [25].
−
1
After washing with water (SZ-WW), the part of this com-
−
1
plex band at higher wavenumbers (i.e., that at 1404 cm
)
The results compiled in Table 1 show that the water wash-
ing treatment removed all Brønsted acid sites together with
water-soluble sulfate groups from SZ. Simultaneously, the con-
centration of Lewis acid sites increased [19]. This suggests that
the labile sulfate groups are responsible for generating Brønsted
acid sites and that they are at least partially bound to Lewis acid
sites. In contrast, sulfation with SO3 led to the restitution of the
Brønsted acid sites (compare SZ-WW-SO3 and SZ-SO3 to SZ-
WW and SZ) and the decrease in the concentration of Lewis
acid sites. Thus, at least some of the chemisorbed SO3 mole-
cules lead to the labile sulfate groups coordinatively bound to
Lewis acid sites. Note that SO3 sulfation also induced Brønsted
acidity on pure ZrO2, which otherwise have only Lewis acid
sites. With the exception of SZ-WW (having no Brønsted acid
sites), the concentration of Brønsted acid sites of SZ is directly
related to the sulfate concentration (see Fig. 6). We speculate,
therefore, that the sulfate groups remaining after the washing
procedure on SZ-WW either are not located at the outer surface
or are bound in a from not permitting hydroxylation.
disappeared; as a result, the maximum of the band was shifted
−
1
downward to 1391 cm [19]. However, sulfation of the washed
sample (SZ-WW-SO3) restored the labile sulfate groups, as
−
1
indicated by the reappearance of the peak at 1404 cm . In
the region of S–O vibration, two bands located at 1011 and
−
1
1
043 cm were observed in the spectra of SZ, SZ-SO3, and
−
1
SZ-WW-SO3. The strong bands in the 1100–1300 cm range
in the IR spectra of SO3-sulfated samples were ascribed to the
surface polysulfate groups, which are related to the high sulfate
content.
3
.3.3. Determination of acid sites by IR spectroscopy of
adsorbed pyridine
Pyridine is widely used as qualitative and quantitative probe
to identify the concentration and nature of acid sites using the
−
1
characteristic bands of pyridinium ions at 1540 cm , as well
−
1
as the band between 1440 and 1455 cm characteristic for
pyridine bound coordinatively to Lewis acid sites [25]. The

44
X. Li et al. / Journal of Catalysis 238 (2006) 39–45
4
. Discussion
Calcined SZ was active for n-butane skeletal isomerization
at 373 K. Removal of the water-soluble sulfate by washing with
water rendered the catalyst inactive even though approximately
6
0% of the sulfate species were retained [19]. In contrast, sul-
fation of the parent SZ with SO3 induced very high catalytic
activity, indicating that SO3 deposition on the properly acti-
vated SZ generates more active sites. Sulfation with SO3 on the
inactive water-washed sample (SZ-WW) induced high activity.
Thus, the pronounced promoting effect of sulfation with SO3
on the catalytic activity of SZ gives strong evidence that the la-
bile sulfate species, generated by the adsorption of SO3, are
essential for the formation of active sites. This indicates that
SO3 must be adsorbed on or close to an existing, albeit inactive,
sulfate group. Theoretical calculations suggest that the active
sulfate species formed is a pyrosulfate group. These functional
groups have been shown to initiate the catalytic cycle by provid-
ing olefins via oxidative dehydrogenation of n-butane [26,27].
A kinetic analysis suggests that the initial rate of isomerization
is directly proportional to the rate at which these oxidizing sul-
fate sites generate butene [26].
The paramount role of the labile active sulfate species is also
demonstrated by the difference between the catalysts sulfated in
situ and those sulfated ex situ. The extremely higher initial ac-
tivity of SZ after in situ sulfation compared with the catalytic
material sulfated ex situ indicates that the concentration of the
active sulfate species induced by chemisorbed SO3 is dramati-
cally increased. We attribute the difference to the significantly
higher concentration of the oxidizing pyrosulfate groups in the
case of the in situ sulfation. It should be noted that these labile
sulfates are very sensitive toward thermal treatment, especially
after rehydration under ambient conditions.
Fig. 8. IR spectra in the region of S=O stretching vibration of sulfated zirco-
nias: SZ-SO , SZ-WW-SO , ZrO –SO and SZ-BW-SO .
3
3
2
3
3
the parent SZ. Nevertheless, it should be kept in mind that this
activity is an order of magnitude lower that of SZ-SO3, which
was obtained by sulfation of the tetragonal sample SZ. Thus,
it appears that the catalytic activity is primarily a function of
the concentration of the labile sulfate groups, which in turn in-
creases with increasing concentration of tetragonal zirconia.
This supposition is supported by the data in Fig. 8 indicating
that the intensity of the band characteristic of the labile sul-
−
1
fate/pyrosulfate group (1404 cm ) varies in parallel with the
catalytic activity. The sample with the most intense band was
the most active; that is, the activity for n-butane isomerization at
−
1
−1
s
3
73 K was 0.175 µmol g
on SZ-SO3. SZ-BW-SO3, hav-
−
1
ing the lowest 1404 cm band, was the least active sample
−
1
−1
).
(
0.028 µmol g
s
The finding that catalytic activity depends strongly on the
concentration of labile sulfate suggests that sulfate groups on
monoclinic SZ should be more stable than sulfate groups on
tetragonal SZ. This trend is also seen in the uptake of SO3 on the
various samples. Having exposed all samples to the exact same
number of SO3 molecules at the same temperature and SO3
partial pressure, the uptake of SO3 increased with increasing
concentration of the monoclinic phase (Fig. 6). This suggests
that stronger basic sites allow binding SO3 more effectively on
monoclinic zirconia than on tetragonal zirconia, but these more
basic sites in monoclinic zirconia also reduce the formation of
labile groups. In line with this reasoning, the catalytic activity
and the fraction of tetragonal zirconia of the SO3 sulfated sam-
ples are directly proportional.
The role of coexisting Brønsted and Lewis acid sites has
been discussed extensively in the SZ literature [29–31]. The
present results clearly show that the role of Lewis acid sites
may be only to influence sulfate group formation, and it might
not be actively involved in the catalytic process. In fact, pure
zirconia (ZrO2) and water-washed SZ (SZ-WW), which have
only Lewis acid sites, are completely inactive. The adsorption
of SO3 leads to the decrease in the concentration of Lewis acid
sites and the increase/formation of Brønsted acid sites. This
suggests that a fraction of labile sulfate species is coordinately
bound to Lewis acid sites. For a given group of supports, the
concentration of Brønsted acid sites increases in parallel to the
The importance of adsorbed SO3 becomes clearer when
comparing the activities of SZ calcined in a shallow bed with
that calcined in a deep bed [28]. These calcination procedures
were conducted under exactly identical conditions, with heating
−
1
up to 873 K at a rate of 10 K min in static air and maintain-
ing this temperature for 3 h. The only difference was the depth
of the catalyst bed; in the shallow-bed configuration, the cata-
lyst was spread out evenly on a ceramic plate, whereas in the
deep-bed mode the catalyst had a bed depth of approximately
1
cm. SO3 evolves during calcination; it largely escapes during
shallow-bed calcination, but is at least partially captured by the
upper layers in the bed. Note that, interestingly, the latter cat-
alyst was an order of magnitude more active than the former.
This demonstrates not only how important SO3 is in forming
an active catalytic material, but also how subtle variations of
the catalyst preparation conditions may dramatically influence
the activity, and hence the stability, of SZ.
Having demonstrated that the generation of the labile oxi-
dizing sulfate groups by chemisorbing SO3 is critical for gener-
ating an active catalyst, we now turn our attention to the role of
the underlying substrate (i.e., the crystalline form of zirconia)
in the formation of active pyrosulfate species. Sulfation with
SO3 of an almost monoclinic zirconia sample (SZ-BW) led to
a sample (SZ-BW-SO3) with catalytic activity close to that of

X. Li et al. / Journal of Catalysis 238 (2006) 39–45
45
concentration of sulfate species (see Fig. 6). At the same time,
Acknowledgments
the intensity of the bands corresponding to the terminal and
bridging hydroxyl groups of ZrO2 decrease dramatically af-
ter SO3 sulfation. Although we do not wish to speculate on
the nature of the newly formed OH groups, we would like to
point to quantum chemical calculations [32] and suggest that
the hydroxy groups in these models are related to hydroxylated
pyrosulfate or bisulfate species [33,34].
In our study, sulfation with gaseous SO3 has been success-
fully applied to already crystallized samples despite the pres-
ence/absence of sulfate groups before calcination and the crys-
tallographic form of zirconia. SO3 sulfation has been purposely
applied to already crystallized samples to gain insight into the
nature of the real sulfating agent (SO3) and the function of the
calcination procedure. It has been claimed that calcination of
SZ precursors at high temperature by aqueous sulfation of an
amorphous zirconium hydroxide or crystalline zirconia is a nec-
essary step in generating active sites. This calcination is claimed
to convert the amorphous zirconium hydroxide to tetragonal
zirconia, to bind the sulfate groups with the zirconia surface
Financial support from the Deutsche Forschungsgemein-
schaft (DFG) in the framework of DFG priority program 1091,
“Bridging the gap in Heterogeneous Catalysis,” is gratefully
acknowledged. The authors are indebted to Dr. C. Breitkopf,
Dr. S. Wrabetz, M. Standke, Dr. K. Meinel, and Dr. A. Hof-
mann, and thank Professor J. Sauer, Professor H. Papp, and
Dr. F. Jentoft for valuable discussions.
References
[
[
[
[
1] T. Yamaguchi, K. Tanabe, Y.C. Kung, Mater. Chem. Phys. 16 (1986) 67.
2] J.R. Sohn, H.W. Kim, J. Mol. Catal. 52 (1989) 361.
3] D. F a˘ rca s¸ iu, J.Q. Li, Appl. Catal. A 175 (1998) 1.
4] M.T. Tran, N.S. Gnep, G. Szabo, M. Guisnet, Appl. Catal. A: Gen. 171
(1998) 207.
[5] M. Hino, S. Kobayashi, K. Arata, J. Am. Chem. Soc. 101 (1979) 6439.
[
[
[
[
6] R.A. Comelli, C.R. Vera, J.M. Parera, J. Catal. 151 (1995) 96.
7] C. Morterra, G. Cerrato, F. Pinna, M. Signoretto, J. Catal. 157 (1995) 109.
8] D. F a˘ rca s¸ iu, J.Q. Li, Appl. Catal. A: Gen. 128 (1995) 97.
9] F. Lonyi, J. Valyon, J. Engelhardt, F. Mizukami, J. Catal. 160 (1996) 279.
[
8,9,35], and in part to eliminate the excess sulfate groups from
[10] C. Morterra, G. Cerrato, M. Signoretto, Catal. Lett. 41 (1996) 101.
4+
Lewis acid sites, that is, coordinately unsaturated Zr [10,11].
The present study shows that sulfation with SO3 at 473 K can
generate active sites without calcination and that Lewis acid
sites are primarily needed for binding SO3. Thus, in the con-
ventional preparation method using a liquid sulfating agent, the
calcination of the sulfated precursor at high temperatures serves
only to generate the SO3 in situ so that it can be captured, form-
ing pyrosulfate groups.
[11] C. Morterra, G. Cerrato, G. Meligrana, M. Signoretto, F. Pinna, G. Strukul,
Catal. Lett. 73 (2001) 113.
[12] W. Stichert, F. Schüth, J. Catal. 174 (1998) 242.
[13] C.R. Vera, J.M. Parera, J. Catal. 165 (1997) 254.
[14] A.F. Bedilo, A.S. Ivanova, N.A. Pakhomov, A.M. Volodin, J. Mol. Catal.
A: Chem. 158 (2000) 409.
[15] T. Yamaguchi, T. Jin, K. Tanabe, J. Phys. Chem. 90 (1986) 3148.
[16] J.F. Haw, J. Zhang, K. Shimizu, T.N. Venkatraman, D.P. Luigi, W. Song,
D.H. Barich, J.B. Nicholas, J. Am. Chem. Soc. 122 (2000) 12561.
[
[
17] J. Zhang, J.B. Nicholas, J.F. Haw, Angew. Chem. Int. Ed. 39 (2000) 3302.
18] P. Canton, R. Olindo, F. Pinna, G. Strukul, P. Riello, M. Meneghetti,
G. Cerrato, C. Morterra, A. Benedetti, Chem. Mater. 13 (2001) 1634.
19] X. Li, K. Nagaoka, J.A. Lercher, J. Catal. 227 (2004) 130.
20] N. Katada, J. Endo, K. Notsu, N. Yasunobu, N. Naito, M. Niwa, J. Phys.
Chem. B 104 (2000) 10321.
5
. Conclusion
[
[
Sulfation with gaseous SO3 has been shown to be an effec-
tive new approach for preparing highly active SZ. This novel
method avoids complex preparation procedures and allows us
to tailor the activity of the catalyst. Labile sulfate species are es-
sential for the formation of active sites, which can be removed
by water-washing treatment and recovered by SO3 sulfation of
crystalline zirconia materials. The decrease in Lewis acidity
after SO3 sulfation indicates that at least some of the labile sul-
fates are coordinately bound to the Lewis acid sites. The labile
sulfate groups induce Brønsted acidity on the zirconia surface.
The increase in Brønsted acid sites and the significant decrease
in zirconia surface OH groups by SO3 sulfation indicate that
Brønsted acid sites are affiliated with pyrosulfate or bisulfate
species. In conventionally prepared materials, the activity of zir-
conia depends critically on the ability to partly decompose and
chemisorb SO3 during calcination and thereby form active py-
rosulfate groups. The direct chemisorption of SO3 circumvents
this step and opens so a wide area of surface modifications for
oxides that has not been available to now.
[
21] W. Hertl, Langmuir 5 (1989) 96.
[22] A.A. Tsyganenko, V.N. Filimonov, J. Mol. Struct. 19 (1973) 579.
[23] T. Yamaguchi, T. Jin, K. Tanabe, J. Phys. Chem. 90 (1986) 3148.
[24] M. Waqif, J. Bachelier, O. Saur, J.C. Lavalley, J. Mol. Catal. 72 (1992)
127.
[25] C.A. Emeis, J. Catal. 141 (1993) 347.
[26] X. Li, K. Nagaoka, L.J. Simon, R. Olindo, J.A. Lercher, A. Hofmann,
J. Sauer, J. Am. Chem. Soc., in press.
[27] X. Li, K. Nagaoka, L.J. Simon, R. Olindo, J.A. Lercher, J. Catal. 232
2005) 456.
(
[
[
[
28] X. Li, K. Nagaoka, L.J. Simon, R. Olindo, J.A. Lercher, in preparation.
29] K. Arata, Adv. Catal. 37 (1990) 165.
30] A. Clearfield, G.P.D. Serrette, A.H. Khazi-Syed, Catal. Today 20 (1994)
2
95.
[31] C. Morterra, G. Cerrato, C. Emanuel, C. Bolis, J. Catal. 142 (1993) 349.
[32] A. Hofmann, J. Sauer, J. Phys. Chem. B 108 (2004) 14652.
[33] L.M. Kustov, V.B. Kazansky, F. Figueras, D. Tichit, J. Catal. 150 (1994)
143.
[
34] V. Adeeva, J.W. de Haan, J. Janchen, G.D. Lei, V. Schunemann, L.J.M.
van de Ven, W.M.H. Sachtler, R.A. van Santen, J. Catal. 151 (1995) 364.
[35] T. Yamaguchi, K. Tanabe, Mater. Chem. Phys. 16 (1986) 67.