Activation of sulfated zirconia catalysts Effect of water content on their activity in n-butane isomerization

Article abstract of DOI:10.1039/a706892a

The effect of water content in sulfated zirconia catalysts on their activities for n-butane isomerization was investigated using catalytic testing, Fourier transform infrared (FTIR) spectroscopy and thermogravimetry analysis. It was found that minor amounts of water promote the catalytic activity while excess water diminishes it. The FTIR spectrum of pyridine adsorbed on the catalysts showed that a decrease in water content resulted in a decrease in Bronsted acidity with a concurrent increase in Lewis acidity. In the most active catalyst, approximately two-thirds of the acid sites are Bronsted sites, and the rest are Lewis sites; an appropriate amount of water is needed to produce this ratio. The Bronsted acid sites presumably contribute to the catalysis by protonating butene formed at redox sites on the catalyst to create carbocations. The high catalytic capability of the Bronsted acid sites requires the presence of adjacent Lewis acid sites, which withdraw electrons from bisulfate SO - H bonds through an induction effect, giving rise to more acidic protons.

Full text of DOI:10.1039/a706892a

View Article Online / Journal Homepage / Table of Contents for this issue
Activation of sulfated zirconia catalysts
E†ect of water content on their activity in n-butane isomerization
Steven X. Song and Ronald A. Kydd
Department of Chemistry, University of Calgary, Calgary, Alberta, Canada T 2N 1N4
The e†ect of water content in sulfated zirconia catalysts on their activities for n-butane isomerization was investigated using
catalytic testing, Fourier transform infrared (FTIR) spectroscopy and thermogravimetry analysis. It was found that minor
amounts of water promote the catalytic activity while excess water diminishes it. The FTIR spectrum of pyridine adsorbed on the
catalysts showed that a decrease in water content resulted in a decrease in Bronsted acidity with a concurrent increase in Lewis
acidity. In the most active catalyst, approximately two-thirds of the acid sites are Bronsted sites, and the rest are Lewis sites; an
appropriate amount of water is needed to produce this ratio. The Bronsted acid sites presumably contribute to the catalysis by
protonating butene formed at redox sites on the catalyst to create carbocations. The high catalytic capability of the Bronsted acid
sites requires the presence of adjacent Lewis acid sites, which withdraw electrons from bisulfate SOwH bonds through an
induction e†ect, giving rise to more acidic protons.
Sulfated zirconia (SZ) has attracted extensive attention
because of its strong acidity and high activity in light alkane
conversions at relatively mild temperatures.1h6 Much work
has been devoted to studies of the preparation, character-
ization and application of this catalytic material. However,
owing to the sensitivity of the material to the preparation and
activation conditions, a number of themes concerning its
physicochemical properties and catalytic reactivity still remain
uncertain.6 One such example is the e†ect of the catalyst
surface hydration on its reactivity. Both beneÐcial1,7 and
detrimental2,8h11 e†ects of water have so far been reported in
the literature. In addition, some authors also suggested that
an intermediate water content is necessary for the catalyst to
reach its maximum activity.12h17
whether some other e†ect of the activation is important.
The objective of this work was to clarify the e†ect of water
on the activity of sulfated zirconia catalysts. We carried out
n-butane isomerization over sulfated zirconia catalysts activat-
ed in Ñowing dry air at various temperatures (150È500 ¡C) and
thus having di†erent water contents, and also characterized
the acidity of these catalyst samples using infrared (IR) spec-
troscopy of adsorbed pyridine. A correlation between the
catalytic results and the acidity measurements was attempted.
We noticed that so far most of the reported IR data for acidity
characterization were obtained over catalysts pretreated under
vacuum at elevated temperatures, while most reaction studies
were carried out over catalysts activated in an inert gas or dry
air stream under atmospheric pressure at elevated tem-
peratures. We believe that a correlation between acidity and
activity data obtained using di†erent activation conditions for
each of these measurements is not appropriate because the
di†erent activations may lead to di†erent dryness of the cata-
lysts. Thus, in this work we used the same activation pro-
cedure for both catalytic experiments and acidity
characterization.
The issue of the water e†ect is closely related to another
important topic, which has been studied extensively but is still
under debate, i.e., the nature of the catalytically active acid
sites. Most of the recent reports indicated the simultaneous
presence of both Lewis and Bronsted acid sites and the
reversible transformation between them upon hydration/
dehydration, in spite of the earlier studies which showed the
existence of only Lewis acidity.18h21 Several authors came to
contradictory conclusions about which type of acid site is
responsible for the activity of sulfated zirconia. Some corre-
lated the catalytic activity to Lewis acidity,8h11,22,23 whereas
others emphasized the role of Bronsted acid sites.1,7,24h26 In
addition, there is another proposal that attributes the activity
to a synergistic e†ect between both types of acid sites.12,27h31
Research into the water e†ect on the reactivity of sulfated
zirconia catalysts is of signiÐcance. This is not only because
many reactions (e.g., FischerÈTropsch synthesis32 and those
involving alcohols14,33,34) over this type of catalyst may
involve the formation of water as a by-product; such research
can also help us understand the activation process of the cata-
lysts. These types of catalyst are usually prepared by impreg-
nating zirconium hydroxide powders with an aqueous
solution of H SO or (NH ) SO followed by drying and cal-
Experimental
Sulfated zirconia catalysts used in this study were prepared
using the conventional method.
A
0.45
M
solution of
ZrOCl É 8H O (600 ml) was hydrolyzed by dropwise addition
2
2
of 75 ml aqueous ammonia (28È30 wt.% NH ) with vigorous
3
mechanical stirring to produce Zr(OH) precipitate. After the
4
hydrolysis, the pH of the mixture was about 9, and the
mixture was stirred for another 30 min. The Zr(OH) precipi-
4
tate was then separated from the liquid by Ðltration and
washed thoroughly with distilled water until no chloride ions
were detected. The Zr(OH) was then dried overnight and
4
powdered to below 100 mesh. Sulfation was carried out by
adding the Zr(OH) powder to a 0.5 M solution of H SO (1 g
4
2
4
solid per 10 ml liquid) followed by stirring the mixture for 1 h.
The solid was separated from the liquid by Ðltration without
washing and then dried overnight.
2
4
4 2
4
cination in air. The calcined sample is usually cooled and
stored in the ambient atmosphere before use, and may adsorb
moisture from the atmosphere during its cooling and storage.
An activation or pretreatment process is usually necessary
prior to reactions to make the material catalytically active.
Since the dryness of the catalyst strongly depends on the acti-
vation temperature, the question arises as to whether this
dryness is the factor that determines catalytic activity, or
The dried sulfated Zr(OH) was calcined either ex situ or in
4
situ. In the ex situ calcination, sulfated Zr(OH) was calcined
4
in a furnace at a given temperature (450È750 ¡C) for 3 h in
static air and then cooled and stored under ambient condi-
tions before being placed in the reactor. In the in situ calcina-
tion, sulfated Zr(OH) was calcined in the reactor, in which
4
J. Chem. Soc., Faraday T rans., 1998, 94(9), 1333È1338
1333

View Article Online
the catalytic testing was to be done, at a given temperature
(500È700 ¡C) for 3 h under Ñowing dry air (30 ml min~1). The
catalyst ex situ calcined at 620 ¡C for 3 h has a surface area of
77 m2 g~1.
n-Butane isomerization was carried out in a Ðxed bed
downward-Ñow quartz reactor (9 mm id) at 90 ¡C under atmo-
spheric pressure. The reaction feedstock was a mixture of n-
butane (Instrument Grade, 99.5%) with helium (High Purity).
Prior to the reaction, ex situ calcined catalyst samples were
activated in the reactor at a given temperature (150È500 ¡C)
for 2 h and cooled to reaction temperature in Ñowing dry air
(30 ml min~1). In situ calcined catalysts were cooled from the
calcination temperature directly to the reaction temperature
under the dry air stream; in this case, care was taken to
prevent contact of the catalyst with moisture before activity
testing. All the catalyst samples were purged with Ñowing
helium (30 ml min~1) at the reaction temperature for 10 min
before being exposed to the reaction feedstock. The reaction
products containing propane, isobutane, pentanes and uncon-
verted n-butane were analyzed on-line using an HP 5890 gas
chromatograph equipped with a stainless-steel column (3 mm
od, 3 m long) packed with Porapak Q (80/100 mesh) and a
Ñame ionization detector. In all experiments, the selectivity to
isobutane was above 94%.
Fourier transform infrared (FTIR) spectra were recorded on
a Nicolet 8000 FTIR spectrometer at 2 cm~1 resolution using
a liquid nitrogen cooled mercuryÈcadmiumÈtelluride (MCT)
detector. SZ samples calcined at 620 ¡C were powdered below
200 mesh and pressed (3 ] 108 Pa, 2 min) into self-supporting
wafers (11 mg cm~2, diameter \ 13 mm). The activation of the
samples was conducted in a quartz IR cell in the same way as
for the catalyst activation in the catalytic testing. In brief, a
wafer was heat-treated at a given temperature (150È500 ¡C) for
2 h and then cooled to room temperature, both steps being
carried out in Ñowing dry air (130 ml min~1). After 30 min
purge of the IR cell with Ñowing helium (130 ml min~1) at
room temperature, an FTIR spectrum was recorded. The
wafer in the cell was then heated to 100 ¡C in Ñowing helium
and pyridine adsorption was performed by admitting a helium
stream (80 ml min~1) saturated with pyridine at room tem-
perature into the cell for 30 min. The physisorbed pyridine
was removed by purging the wafer at 100 ¡C with Ñowing
helium (130 ml min~1) for 2 h, and then another IR spectrum
was recorded.
Fig. 1 n-Butane conversion vs. time on stream over SZ catalysts acti-
vated at various temperatures (indicated on the right-hand side of
each curve). Before activation, the catalysts were ex situ calcined at
620 ¡C for 3 h. Reaction conditions: 90 ¡C, n-butane \ 1 ml min~1,
helium \ 22 ml min~1, catalyst \ 0.25 g.
Thermogravimetry analysis (TG) was carried out using a
Perkin-Elmer TGS-1 thermobalance. About 5 mg of SZ cal-
cined at 620 ¡C was heated in a platinum pan under Ñowing
dry air from room temperature to 800 ¡C at a rate of 20 ¡C
min~1, and the sample weight vs. temperature was recorded.
In a separate experiment, in order to simulate the catalyst
activation conditions used for the catalytic testing, an SZ
sample was heat treated in the thermobalance under a dry air
stream at incrementally increasing temperatures (150È550 ¡C)
for 2 h, and the sample weight after the treatment at each
temperature was monitored.
using higher reaction temperatures (e.g., above 150 ¡C)
because it was too short to be observed at these reaction tem-
peratures. It is seen from Fig. 1 that the induction period
decreased with the increasing catalyst activation temperature
from 150 to 250 ¡C and then stayed essentially unchanged at
higher catalyst activation temperatures. It is also seen from
Fig. 1 that the maximum n-butane conversion increased with
the catalyst activation temperature from 150 to 300 ¡C and
then decreased at higher catalyst activation temperatures. This
is more clearly shown in Fig. 2(a). Fig. 2(b) shows the n-butane
conversion data at a reaction temperature of 45 ¡C. Fig. 2
clearly indicates that there is an optimum temperature for
catalyst activation, which in our study lies in the range 250È
300 ¡C.
It is known that, among many factors, the water content of
SZ catalysts is important in inÑuencing their activity. The
catalysts used in Fig. 1 and 2 were ex situ calcined at 620 ¡C
for 3 h in static air in a furnace. The calcined catalysts were
exposed to the ambient atmosphere during cooling and
storage without any special protection. In the temperature
range for the catalyst activation (150È500 ¡C), the sulfate ions
will not be removed from the catalyst, but the water content
in the catalyst (i.e., catalyst dryness) will be a†ected by the
temperature of activation; the higher the activation tem-
Results
Fig. 1 shows n-butane conversion vs. time on stream over SZ
catalysts, which were Ðrst ex situ calcined at 620 ¡C and then
in situ activated in Ñowing dry air at temperatures from 150È
500 ¡C. As the reaction was carried out at a low temperature,
90 ¡C, an induction period in which n-butane conversion
increased with time on stream was observed in all the experi-
ments. After the conversion reached a maximum, it started to
decrease with the time on stream. Throughout this paper, we
take the maximum n-butane conversion as a measure of the
catalytic activity, rather than simply taking the conversion at
some Ðxed time. The induction period in n-butane isomer-
ization over SZ-based catalysts was usually ignored in studies
1334
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
catalysts were found to be similar to those over the ex situ
calcined catalysts shown in Fig. 1. The maximum n-butane
conversion vs. the activation temperature is depicted in Fig. 4,
which closely resembles Fig. 2(a) with a minor di†erence
believed to arise from the minor temperature di†erence
between in situ and ex situ calcinations. The induction periods
corresponding to the catalytic testing data in Fig. 4 varied
from 38 to 48 min, depending on the catalyst activation tem-
perature.
IR spectroscopy was performed to obtain an indication of
the change of the catalyst acidity with the activation tem-
perature. Consistent with the literature data, a band at 1300È
1400 cm~1 was observed, which generally has been assigned
to the asymmetric stretching of SxO double bonds. As shown
in Table 1, the wavenumber at which the band appears
strongly depends on the activation temperature: the higher
the activation temperature, the higher the wavenumber. After
pyridine adsorption, the band shifted to a lower wavenumber,
with higher activation temperatures producing larger shifts.
IR spectroscopy of adsorbed pyridine provides information
about Bronsted and Lewis acidities on the SZ catalysts. Both
Bronsted and Lewis acid sites were observed in the range of
activation temperature used in this study. An estimation of the
acidity was performed based on the integrated molar absorp-
tion coefficients (IMACs) determined by Emeis,35 which are
considered to be the most reliable reported so far.36 The
IMACs of bands characteristic of pyridine adsorbed on
Bronsted sites (1540 cm~1 band) and Lewis sites (1445 cm~1
band) are 1.67 cm lmol~1 and 2.22 cm lmol~1, respectively.
Fig. 5 shows the dependence of the acidity on the activation
temperature. As expected, an increase in activation tem-
perature resulted in a decrease in Bronsted acidity with a con-
current increase in Lewis acidity. The total acidity also
decreased with the increasing activation temperature (Fig. 5A).
Fig. 2 Dependence of the catalyst activity on its activation tem-
perature. The catalysts were the same as in Fig. 1. (a) Reaction condi-
tions as in Fig. 1. (b) Reaction conditions: 45 ¡C, n-butane \ 1 ml
min~1, helium \ 4 ml min~1, catalyst \ 0.3 g.
perature, the drier the catalyst. Therefore it is suggested that
the change of n-butane conversion vs. the catalyst activation
temperature is due to the change in the catalyst dryness.
To investigate the role of water, in a series of experiments
we loaded sulfated Zr(OH) samples into the reactor, calcined
4
them in situ at various temperatures (500È700 ¡C), and then
cooled them in a dry air stream to the reaction temperature
(90 ¡C). Using this approach, the samples maintained the high
degree of dryness reached during the high-temperature calcin-
ation because they did not have chance to adsorb moisture
from the atmosphere. The reaction results are shown in Fig.
3(a). The n-butane conversion over the dry catalyst samples
was very low. In contrast, the ex situ calcined SZ catalysts,
which were wetted by atmospheric moisture during cooling
and storage and then activated at 250 ¡C in dry air, exhibited
much higher activity [Fig. 3(b)].
To clarify further the role of water in SZ catalysts, we
carried out the following experiment. After sulfated Zr(OH)
samples (0.25 g) were in situ calcined at 620 ¡C in a dry air
4
stream for 3 h and then cooled to 100 ¡C, a helium stream (22
ml min~1) passing through a bubbler containing water at
11 ¡C was admitted into the reactor for 10 min. The wetted
catalyst was then activated in a dry air stream at a certain
temperature (150È500 ¡C) for 2 h and cooled to the reaction
temperature (90 ¡C). The catalyst testing results over these
Fig. 4 Dependence of the catalyst activity on its activation tem-
perature. The catalysts were in situ calcined at 620 ¡C for 3 h,
hydrated at 100 ¡C and activated in dry air at various temperatures.
Reaction conditions as in Fig. 1.
Table 1 E†ect of activation temperature on the IR band of the
asymmetric stretching of SxO double bond
SxO stretching frequency/cm~1
activation
temperature/¡C
before pyridine adsorption
after pyridine adsorption
150
250
300
400
500
1301 (not well developed)
1300 (not well developed)
1317
1334
1365
1387
1310
1313
1317
1320
Fig. 3 Dependence of the catalyst activity on its calcination tem-
perature. (a) In situ calcination. (b) Ex situ calcination followed by
activation at 250 ¡C in dry air. Reaction conditions as in Fig. 1.
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1335

View Article Online
300 ¡C) contains about 200 lmol g~1 of water, and 67% of
Bronsted sites out of the total acid sites.
Discussion
In the above, we have shown that the activity of SZ catalysts
for n-butane isomerization strongly depends on their water
content under their working conditions. In the present study,
the catalyst water content was controlled by varying the acti-
vation temperature of the catalyst samples, which were cal-
cined (ex situ or in situ) and then hydrated (by atmospheric
moisture or by water from a bubbler). The activation tem-
peratures were chosen so that the di†erent catalyst drynesses
could be obtained, but the sulfur species in the catalyst
samples would stay intact. It was found in this study that
minor amounts of water promote the catalyst activity while
excess water diminishes it. This result is di†erent from that
reported by some other researchers, who found a poisoning
e†ect of water on SZ catalysts,8h11 but agrees with the data
recently reported by Dumesic and co-workers.15h17 It is
worth noting that the water content in working SZ catalysts is
a function of a number of factors including their preparation
method, calcination/activation conditions, and even the cata-
lyst testing method. In addition, the promoting e†ect of water
was only observed at a very low level of water content, which
make its observation difficult. In the studies of water e†ect
reported so far, di†erent conditions for catalyst preparation,
calcination/activation and catalytic testing were used. These
di†erences may have caused the di†erent observation in water
e†ects previously reported.
Fig. 5 Dependence of the catalyst acidity on catalyst activation tem-
perature. Before activation, the catalysts were treated as described in
the caption to Fig. 1.
Thermogravimetry data of SZ catalysts from our study and
from the literature37 indicate that during the thermal treat-
ment below 500 ¡C, the weight loss of the samples is essen-
tially due to its dehydration. The removal of sulfur species
only occurs at temperatures higher than 550 ¡C. Therefore, as
considered by Kobe et al.,16 we assume the catalyst activated
at 500 ¡C contains no water, and the water contents of the
catalysts activated at the lower temperature were calculated
based on the weight di†erence between these samples and the
one activated at 500 ¡C. The results are shown in Fig. 6. The
catalyst dehydration occurred in three stages. In the Ðrst stage
(below 220 ¡C), there was a fast dehydration due to the
removal of physically adsorbed water. A comparison of Fig. 6
with Fig. 5 indicates that the removal of this portion of water
does not strongly a†ect the Bronsted sites. During the second
stage (220È400 ¡C), dehydration was slow, leading to a
gradual decrease in Bronsted sites. In the third stage (400È
Fig. 2 indicates that the tendency for the reactivity di†er-
ence induced by the catalyst thermal pretreatment observed at
two reaction temperatures (90 and 45 ¡C) is the same. In addi-
tion, Kobe et al.16 carried out reactions at a higher tem-
perature (150 ¡C) and observed the same tendency as we have.
Therefore, we believe that our conclusion about the water
e†ect is appropriate, at least at reaction temperatures up to
150 ¡C.
Pretreatment (i.e., calcination and activation) of SZ is nor-
mally necessary to make the material catalytically active. For
ex situ calcined samples, the purpose of the activation step is
to achieve a preliminary surface dehydration. Then, can one
consider that, if the surface dehydration achieved in a calcina-
tion step is maintained (e.g., for in situ calcination where the
calcined catalyst is not exposed to the atmosphere), the cata-
lyst will necessarily be e†ective?38 The answer based on our
data is no. Fig. 3(a) indicates that in situ calcined SZ catalysts
without any hydration are inactive. In other words, an active
SZ catalyst needs a certain degree of hydration. The cooling
and storage of ex situ calcined samples provide a chance for
the samples to adsorb moisture from the atmosphere. Once an
appropriate water content in the catalyst is created by
removal of excess water in activation, the catalyst becomes
active [Fig. 3(b)]. In contrast, for in situ calcined catalyst, as
the high degree of surface dehydration achieved in its calcina-
tion is maintained, the catalyst is inactive. Therefore, it is con-
cluded that an activation step is needed for both ex situ and in
situ calcined SZ catalysts to become active. For ex situ cal-
cined samples, the activation is to remove excess water to
reach an appropriate water content, whereas for in situ cal-
cined samples, the activation is to add an appropriate amount
of water to the samples to make them active.
500 ¡C), dehydration became faster, leading to
a sharp
decrease in Bronsted sites. A comparison of Fig. 2 with Fig. 5
and 6 indicates that the most active catalyst (calcined at 250È
The form of water in a SZ catalyst depends on the degree of
its dehydration. According to Bolis et al.,31 upon contact with
a highly dehydrated SZ where coordinatively unsaturated
(cus) Zr4` cations are present, H O may dissociate on very
2
strongly acidic cus Zr4` cations, giving rise to a surface OH
group. Excessively adsorbed water may exist in the form of
undissociated water molecules. We believe that heating a fully
hydrated SZ sample involves Ðrst a fast removal of undis-
Fig. 6 Water content of SZ catalysts activated at various tem-
peratures. Before activation, the catalysts were treated as described in
the caption to Fig. 1.
1336
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
sociated water molecules at lower temperatures and then a
slow elimination of surface OH groups at higher temperatures,
as shown in Fig. 6. An active SZ catalyst should contain no
undissociated water but possess strongly acidic OH groups.
During the preparation of this manuscript, the work by
Dumesic and co-workers15h17 was published, which showed a
similar e†ect of water content in SZ to that described above.
These authors16 suggested that the changes in catalytic activ-
ity with drying temperature are not directly related to a trans-
formation of Bronsted to Lewis acid sites, rather, the surface
sulfate species may participate in butane isomerization via an
oxidationÈreduction cycle, and water may promote the cata-
lytic activity by altering the oxidationÈreduction properties of
the surface species. In the present work, the FTIR spectrum of
pyridine adsorbed on the catalysts shows that the decrease in
water content resulted in a decrease in Bronsted sites with a
concurrent increase in Lewis sites. The most active catalyst
possesses an intermediate amount of Bronsted acidity. There-
fore, we believe that the contribution of the catalyst acidity to
the catalytic activity should not be ignored.
The acidity of SZ catalysts has been extensively studied.6 It
has been observed that both Bronsted and Lewis acid sites
can be present and their densities depend on, among many
factors, the degree of dryness of the catalysts. The two types of
acid sites are reversibly transformed upon hydration/
dehydration. Our observation (Fig. 5) is in agreement with the
literature data.
SigniÐcant change in the IR spectra in the range of SxO
stretching frequency of sulfate ion was caused by activating
SZ at di†erent temperatures (Table 1). The band at 1300È1400
cm~1 is attributed to the asymmetric vibration of SxO bonds
of the bidentate sulfate ion coordinated to the metal cation.18
The band position relates to the strength of the Lewis acid
sites of the metal cations in sulfated metal oxides. The higher
the frequency of the asymmetric vibration of the SxO band,
the stronger the Lewis acid sites.26 It is seen from Table 1
that, with an increase in the activation temperature, this band
shifted to higher frequencies, indicating that the Lewis acid
sites were strengthened. When the catalysts were exposed to
pyridine, the SxO bands shifted to lower frequencies; the
stronger the Lewis acid sites, the larger the shift.18
Lewis acid site adjacent to an SwOwH group. These
bisulfate groups act as highly acidic Bronsted sites because the
neighboring Lewis acid sites tend to withdraw electrons from
the bisulfate group, thus weakening the SOwH bond. It is
inferred that the combination of the bisulfate protons with
adjacent Lewis acid sites is responsible for the strong acidity.
It is believed that the bisulfate protons act as active Bronsted
acid sites. In catalyst samples with excess water, although the
Bronsted acid sites are abundant, their strength is low because
the adjacent Lewis acid sites are scarce. With the increasing
activation temperature, more adsorbed water is removed and
the strength of the Lewis acid sites is enhanced, which resulted
in an increased electron-withdrawing ability of the Lewis acid
sites. This, in turn, increases the strength of the Bronsted acid
sites and thus the catalytic activity. When the catalyst is acti-
vated at progressively higher temperatures, the loss of the
Bronsted acid sites becomes prodominent, and the drastic
reduction of the amount of the Bronsted acid sites leads to a
decrease of the catalytic activity.
In the above, we have discussed the importance of the cata-
lyst acidity on the catalytic activity in n-butane isomerization.
As to how the acidity contributes to the activity, there are
di†erent considerations. Some researchers believed that SZ-
based catalysts possess superacidity that can easily generate
carbocations from butane molecules to initiate the reaction.
However, this consideration has been challenged by several
recent studies using various techniques for acidity character-
ization,13,39h43 which showed that the acid strength of SZ is
only about equivalent to that of such zeolites as HY, ZSM-5,
etc., which are not superacids. The exceptionally high activity
of SZ catalysts stems from their ability to facilitate the forma-
tion of butene molecules.44,45 It was proposed that the cata-
lyst acidity contributes to n-butane conversion by protonating
the butene molecules formed to generate carbenium ions,
which then either go through further transformations46 or
participate in a bimolecular reaction process in which the
carbenium ions Ðrst react with butene molecules to form C
entities, which then go through skeletal rearrangement and b-
8
scission to form isobutane.44,47 The formation of butene,
according to Ghenciu and Farcasiu,46 may take place through
a redox process involving a one-electron oxidation followed
by trapping of the cation radicals on the surface, which leads
to sulÐte esters. The esters can either ionize to generate carbo-
cations or eliminate to form butene. Therefore, the catalyst
acidity contributes to the conversion of n-butane, not through
direct activation of butane to create carbocations, but by
protonating butene, which is formed from butane at a di†er-
ent type of sites (redox). Therefore, a synergism between the
redox capability and the acidity of SZ catalysts can be visual-
ized, and the Bronsted acid sites are essential for the proto-
nation.
A number of models have been proposed to describe the
sulfate species on the catalyst surface. Of these models, the one
proposed by ClearÐeld et al.27 seems to correlate best with
our experimental data. In this model (Scheme 1), the
uncalcined catalyst (species I) contains protons as bisulfate
and as hydroxy groups bridging two Zr ions. During calcina-
tion, water is lost to form species II and III. In both struc-
tures, Lewis acid sites are formed as indicated by asterisks, but
in species II the bisulfate remains intact, which results in a
According to the synergistic reaction process, an induction
period in which n-butane conversion increases with time on
stream can be predicted in the early stages of the reaction.46,48
Indeed, an induction period was commonly observed in most
of the studies of n-butane isomerization over Fe and Mn pro-
moted SZ catalysts at low temperatures (usually below
100 ¡C). At relatively high reaction temperatures, it becomes
too short to be observed. The induction period over
unpromoted SZ catalysts was rarely reported in the literature
because the reaction temperature used there was usually
higher than 150 ¡C. In this work, we used a reaction tem-
perature of O90 ¡C. Clear induction periods were observed
(Fig. 1), which are consistent with the proposal that butene is
involved in the reaction.
Conclusions
It was shown in this work that small amounts of water
Scheme 1
increase the activity of SZ catalysts in n-butane isomerization,
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1337

View Article Online
18 T. Jin, T. Yamaguchi and K. Tanabe, J. Phys. Chem., 1986, 90,
4794.
while excess water diminishes it. Both Bronsted and Lewis
acid sites were present on the active catalysts, and a decrease
in water content resulted in a decrease in Bronsted acidity
with a concurrent increase in Lewis acidity. In the most active
catalyst, approximately two-thirds of the acid sites are Bron-
sted sites, and the rest are Lewis sites; an appropriate amount
of water is needed to produce this ratio. We believe that the
Bronsted acid sites (i.e., bisulfate protons) are primarily
responsible for the formation of carbenium ions by proto-
nating butene formed at redox sites on the catalyst. The
19 T. Yamaguchi, K. Tanabe and Y. C. Kung, Mater. Chem. Phys.,
1986, 16, 67.
20 M. Bensitel, O. Saur, J. C. Lavally and G. Mabilon, Mater. Chem.
Phys., 1987, 17, 249.
21 A. Corma, V. Fornes, M. I. Juan-Rajadell and J. M. Lopez Nieto,
Appl. Catal. A: Gen., 1994, 116, 151.
22 F. Pinna, M. Signoretto, G. Strukul, G. Cerrato and C. Morterra,
Catal. L ett., 1994, 26, 339.
23 R. Srinivasan, R. A. Keogh and B. H. Davis, Catal. L ett., 1996,
36, 51.
butene and the carbenium ions react to form C entities which
24 D. A. Ward and E. I. Ko., J. Catal., 1994, 150, 18.
25 K. B. Fogash, G. Yaluris, M. R. Gonzalez, P. Ouraipryvan, D. A.
Ward, E. I. Ko and J. A. Dumesic, Catal. L ett., 1995, 32, 241.
26 K. Ebitani, J. Konishi and H. Hattori, J. Catal., 1991, 130, 257.
27 A. ClearÐeld, G. P. D. Serrette and A. H. Khazi-Syed, Catal.
T oday, 1994, 20, 295.
8
then go through skeletal rearrangement and b-scission to
produce isobutane. According to the model of surface sulfate
species proposed by ClearÐeld et al.,27 the acid strength of the
bisulfate protons depends on the number of adjacent co-
ordinatively unsaturated acid sites (i.e., Lewis sites), which
weaken the bisulfate SOwH bonds by withdrawing electrons
from the bonds through an inductive e†ect.
28 P. Nascimento, C. Akratopoulou, M. Oszagyan, G. Coudurier, C.
Travers, J. F. Joly and J. C. Vedrine, Stud. Surf. Sci. Catal., 1993,
75, 1185.
29 F. R. Chen, G. Coudurier, J-F. Joly and J. C. Vedrine, J. Catal.,
1993, 143, 616.
30 C. Morterra, G. Cerrato, V. Bolis, S. Di Ciero and M. Signoretto,
J. Chem. Soc., Faraday T rans., 1997, 93, 1179.
31 V. Bolis, G. Magnacca, G. Cerrato and C. Morterra, L angmuir,
1997, 13, 888.
We gratefully acknowledge Ðnancial support from the Natural
Sciences and Engineering Research Council of Canada
(NSERC). S.X.S. is thankful to NSERC for awarding a Post-
doctoral Fellowship.
32 X. Song and A. Sayari, Energy Fuels, 1996, 10, 561.
33 M. Waqif, J. Bachelier, O. Saur and J-C. Lavalley, J. Mol. Catal.,
1992, 72, 127.
References
1
2
K. Arata, Adv. Catal., 1990, 37, 165.
34 M. Nita, H. Sakoh and K. Aomura, Appl. Catal., 1984, 10, 215.
35 C. A. Emeis, J. Catal., 1993, 141, 347.
K. Tanabe, H. Hattori and T. Yamaguchi, Crit. Rev. Surf. Chem.,
1990, 1, 1.
36 T. Barzetti, E. Selli, D. Moscotti and L. Forni, J. Chem. Soc.,
Faraday T rans., 1996, 92, 1401.
3
4
T. Yamaguchi, Appl. Catal., 1990, 61, 1.
B. H. Davis, R. A. Keogh and R. Srinivasan, Catal. T oday, 1994,
20, 219.
37 R. Srinivasan, R. A. Keogh, D. R. Milburn and B. H. Davis, J.
Catal., 1995, 153, 123.
5
6
7
A. Corma, Chem. Rev., 1995, 95, 559.
38 C. Morterra, G. Cerrato and M. Signoretto, Catal. L ett., 1996, 41,
X. Song and A. Sayari, Catal. Rev.-Sci. Eng., 1996, 38, 329.
M. Y. Wen, I. Wender and J. W. Tierney, Energy Fuels, 1990, 4,
372.
101.
39 B. Umansky, J. Engelhardt and W. K. Hall, J. Catal., 1991, 127,
128.
8
9
C. Morterra, G. Cerrato, F. Pinna, M. Signoretto and G. Strukul,
J. Catal., 1994, 149, 181.
V. M. Mastikhin, A. V. Nosov, S. V. Filimonova, V. V. Terskikh,
N. S. Kotsarenko, V. P. Shmachkova and V. I. Kim, J. Mol.
Catal. A: Chem., 1995, 101, 81.
40 L. M. Kustov, V. B. Kazansky, F. Figueras and D. Ticht, J.
Catal., 1994, 150, 143.
41 V. Adeeva, J. W. de Haan, J. Janchen, G. D. Lei, G. Schunemann,
L. J. M. van de Ven, W. M. H. Sachtler and R. A. van Santen, J.
Catal., 1995, 151, 364.
10 R. A. Keogh, R. Srinivasan and B. H. Davis, J. Catal., 1995, 151,
42 V. Semmer, P. Batamack, C. Doremieux-Morin, R. Vincent and
J. Fraissard, J. Catal., 1996, 161, 186.
43 R. S. Drago and N. Kob, J. Phys. Chem. B, 1997, 101, 3360.
44 V. Adeeva, G. D. Lei and W. M. H. Sachtler, Catal. L ett., 1995,
33, 135.
45 J. E. Tabora and R. J. Davis, J. Am. Chem. Soc., 1996, 118, 12240.
46 A. Ghenciu and D. Farcasiu, J. Mol. Catal. A: Chem., 1996, 109,
273.
47 J. E. Tabora and R. J. Davis, J. Catal., 1996, 162, 125.
48 M. A. Coelho, D. E. Resasco, E. C. Sikabwe and R. L. White,
Catal. L ett., 1995, 32, 253.
292.
11 R. A. Comelli, C. R. Vera and J. M. Parera, J. Catal., 1995, 151,
96.
12 J. H. Lunsford, H. Sang, S. M. Campbell, C-H. Liang and R. G.
Anthony, Catal. L ett., 1994, 27, 305.
13 F. Babou, B. Bigot, G. Coudurier, P. Sautet and J. C. Vedrine,
Stud. Surf. Sci. Catal., 1994, 90, 519.
14 D. Fraenkel, Abstr. Book, 14th North Am. Meeting Catal. Soc.,
Snowbird, UT, June 1995, p. T-48.
15 M. R. Gonzalez, J. M. Kobe, K. B. Fogash and J. A. Dumesic, J.
Catal., 1996, 160, 290.
16 J. M. Kobe, M. R. Gonzalez, K. B. Fogash and J. A. Dumesic, J.
Catal., 1996, 164, 459.
17 M. R. Gonzalez, K. B. Fogash, J. M. Kobe and J. A. Dumesic,
Catal. T oday, 1997, 33, 303.
Paper 7/06892A; Received 23rd September, 1997
1338
J. Chem. Soc., Faraday T rans., 1998, V ol. 94