Hexane isomerization and cracking activity and intrinsic acidity of H-zeolites and sulfated zirconia-titania

Article abstract of DOI:10.1021/jp055587o

Adsorption of N₂ was studied on zeolite H-Y, ultrastabilized H-Y (H-USY), H-mordenite, H-ZSM-5, H-β, and on sulfated zirconia-titania (SZT) mixed oxide by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) at 298 K and at N₂ pressures up to 9 bar. The adsorption-induced Δv_OH red-shift of the v_OH bands was used as a measure of the intrinsic acid strength of the Brnonsted acid sites. The intrinsic acid strength of the solids follows the order of H-ZSM-5 ≈ H-mordenite ≈ H-β > H-USY > SZT ≈ H-Y. The solids were characterized by their hexane conversion activities at 553 K and 6.1 kPa hexane partial pressure. The reaction was shown to proceed predominantly by a bimolecular mechanism, while the reaction was first order in hexane and zero order in alkenes. The site-specific apparent rate constant of the bimolecular hexane conversion was shown to parallel the intrinsic acid strength of the samples, suggesting that the ratio of the apparent and the intrinsic activity, that is, the K_A′ equilibrium constant of alkane adsorption on the hydrocarbon-covered sorption sites, is hardly dependent on the catalyst structure.

Full text of DOI:10.1021/jp055587o

J. Phys. Chem. B 2006, 110, 1711-1721
1711
Hexane Isomerization and Cracking Activity and Intrinsic Acidity of H-Zeolites and
Sulfated Zirconia-Titania
Ferenc Lo´nyi, Anita Kova´cs, and Jo´zsef Valyon*
Institute of Surface Chemistry and Catalysis, Chemical Research Center, Hungarian Academy of Sciences,
Pusztaszeri u. 59-67, Budapest, Hungary 1025
ReceiVed: September 30, 2005; In Final Form: NoVember 10, 2005
Adsorption of N2 was studied on zeolite H-Y, ultrastabilized H-Y (H-USY), H-mordenite, H-ZSM-5,
H-â, and on sulfated zirconia-titania (SZT) mixed oxide by diffuse reflectance infrared Fourier transform
spectroscopy (DRIFTS) at 298 K and at N2 pressures up to 9 bar. The adsorption-induced ∆νOH red-shift of
the νOH bands was used as a measure of the intrinsic acid strength of the Brnønsted acid sites. The intrinsic
acid strength of the solids follows the order of H-ZSM-5 ≈ H-mordenite ≈ H-â > H-USY > SZT ≈
H-Y. The solids were characterized by their hexane conversion activities at 553 K and 6.1 kPa hexane
partial pressure. The reaction was shown to proceed predominantly by a bimolecular mechanism, while the
reaction was first order in hexane and zero order in alkenes. The site-specific apparent rate constant of the
bimolecular hexane conversion was shown to parallel the intrinsic acid strength of the samples, suggesting
that the ratio of the apparent and the intrinsic activity, that is, the KA′ equilibrium constant of alkane adsorption
on the hydrocarbon-covered sorption sites, is hardly dependent on the catalyst structure.
1. Introduction
Recently, significant progress was achieved in understanding
the relationship between the alkane conversion activity and the
Provided that the energy of all interactions between a weak
base adsorbate molecule and its adsorption environment is
negligible relative to the energy of its H-bonding to the
hydroxyls of a solid acid, any parameter that characterizes the
interaction parallels the OH deprotonation energy (DE) and,
thereby, the intrinsic acid strength of the sites.1-17 For instance,
the adsorption-induced 1H NMR chemical shift of the protons,
δH, or the ∆νOH shift of the νOH infrared (IR) band is related to
the strength of H-bonding. Studies using adsorptives, such as
CO, benzene, or CD3CN, for probing the acid strength suggested
that H-zeolites, including H-ZSM-5, H-mordenite, H-â, and
H-Y, were much stronger Brønsted acids than sulfated zirconia
(SZ).13-19 As far as zeolites and SZ are concerned, this result
contradicted the found activities of the solids in the conversion
of alkanes.16-19 Moreover, the adsorption heat of strong bases
acidity of H-zeolites.23-27 It was shown that adsorption of
reactant alkane contributes to the apparent kinetics of alkane
isomerization and cracking. Thus, the apparent rate constant,
related to a single active site, cannot be expected to correlate
with the intrinsic acid strength of the Brønsted acid sites.
Attempts to determine intrinsic catalytic activities and to relate
them to the intrinsic acid strengths led to conflicting results.23-26
23,24
Lunsford et al.
found that intrinsic activation energies of
unimolecular hexane cracking were different for H-zeolites,
having different structure and composition. However, this
variation was not fully consistent with the variation of the
intrinsic acidities of the catalysts, determined by the mentioned
IR spectroscopic method. In contrast, Williams et al.25,27 claimed
that the intrinsic acid strength of the active Brønsted sites were
similar in different zeolites, and the different activities in the
unimolecular alkane cracking could be accounted primarily for
the differences in the alkane adsorption properties of the
catalysts.
such as ammonia showed better correlation with the catalytic
19-21
activity than δH or ∆νOH
.
These results inferred that
catalytic activity was not determined solely by the acidity of
the catalysts. Another plausible explanation is that the measured
parameters were inappropriate to characterize the acidity,
effective under the reaction conditions.
In most previous IR works, the spectra of the surface species,
obtained from adsorption of said diatomic molecular probes,
were recorded at low-temperature (<100 K) in order to get
detectable coverage of the sorption sites. This temperature was
far lower than that of the catalytic alkane conversions. At lower
temperature, the mobility of the protons is lower and their acid
strength is higher than at higher temperature.28 In the present
study, high-pressure DRIFT spectroscopy was used to study N2
adsorption over acid catalysts at 298 K. The obtained ∆νOH
values were taken as a measure of the intrinsic acid strength of
the Brønsted sites. Virtually identical ∆νOH values were obtained
for high-silica zeolites, such as H-ZSM-5, H-mordenite, and
H-â, and significantly lower for zeolite H-Y, SZ, and SZT.
Unexpectedly, the apparent site-specific rate constant of bimo-
lecular hexane cracking was found to show fair correlation with
the intrinsic acidity of the catalysts.
One important aspect of probing intrinsic Brønsted acidity
is that the adsorptive base should have insignificant interaction
with the solid except for the H-bonding to the OH groups.6,7
Spatial hindrance of adsorption must be also avoided.22 For the
examination of microporous materials, the application of small
diatomic molecules, such as CO, N2, O2, or H2, are preferred
because these adsorptives comply best with above require-
ments.2-12 The nitrogen has the advantage that it does not go
through chemical changes on most catalysts. It is also preferable
for IR spectroscopic examination because the N2 gas is IR
inactive, and therefore, it does not contribute to the spectrum.3-7,12
* Corresponding author. E-mail: valyon@chemres.hu. Fax: +36 1 325
7750.
10.1021/jp055587o CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/05/2006

1712 J. Phys. Chem. B, Vol. 110, No. 4, 2006
Lo´nyi et al.
TABLE 1: Characterization of the Zeolites
type
composition
note
NH4-ZSM-5
(NH4)5.75Na0.05(AlO2)5.8(SiO2)90.2
prepared from Na, H-ZSM-5
(T-3ZIPC, Chem. Kombinat, VEB,
Bitterfeld-Wolfen, former GDR) by ion
exchange
prepared from Na-Y (Linde Co.,
LZ-Y62) by ion exchange
LZ-Y82, Linde Co.;
NH4-Y
(NH4)>54.6Na<0.2(AlO2)54.8(SiO2)137.2
NH4-USY
(NH4)<29.3Na<1.7(EFAl)24.5(AlO2)31(SiO2)161
EFAl ) extraframework Al; the amount
of EFAl was taken from ref 60
LZ-M6, UOP
H-mordenite
H-â
H
H
5.97 Na0.23(AlO2)6.2 (SiO2)41.8
4.68 Na0.02(AlO2)4.7(SiO2)59.3
Valfor CP, PQ Corp.
2. Experimental Section
were obtained at 298 K by averaging 512 scans at a nominal
resolution of 2 cm-1. Because the level of the DRIFTS signal
depends on the scattering coefficient, which is function of the
particle size and the packing density of the sample, only
qualitative comparisons of the band intensities were made for
the different samples. However, the quantitative evaluation of
the signals is justified for a single sample at different gas
loading.
Materials. The NH4-zeolites, used in the present study, were
either commercial products or were obtained from Na-form
zeolites by conventional ion exchange. The studied H-zeolites
were either commercial products or were prepared from the
NH4-zeolites by in situ thermal deammoniation (Table 1).
The preparation method and the properties of the sulfated
zirconia-titania (SZT) sample was described in ref 17. An
equimolar mixture of Zr-n-propoxide/Ti-i-propoxide was
hydrolyzed by NH3 solution. The obtained precipitate was dried
at 383 K. The dry gel was contacted with 0.25 M H2SO4
solution, dried again, and calcined in air at 823 K. The sulfate
content and the specific surface area of the sample were 10.2
wt % and 258 m2/g, respectively. For comparison, zirconia (Z)
and sulfated zirconia (SZ) samples were prepared from Zr-n-
propoxide by following a similar preparation procedure. The
SZ sample was shown to have much lower surface area than
the SZT sample, but the surface-specific properties of the
samples were found to be virtually the same.17
Catalytic Hexane Conversion. The catalyst powder was
compressed to pellets, and the pellets were crushed and sieved.
The 0.25-0.50 mm sieve fraction was used as a catalyst.
Depending on the activity, 10-500 mg of catalyst particles were
filled into a quartz tube microreactor (i.d. 4 mm). The catalyst
was activated in situ in the reactor in an O2 flow of 20 cm3/min
at 773 K overnight. Then the reactor was flushed with a 20
cm3/min He flow and cooled to the 553 K temperature of the
reaction. The reaction was started by switching the He flow to
a flow of a hexane/He mixture. The mixture was prepared by
saturating a He flow with hexane vapor at 273 K. Thus, the
hexane partial pressure was 6.1 kPa. The hexane feed, fed on a
unit mass of the catalyst (F/W, µmol g-1 s-1), was varied to
get conversions below about 20%. The reactor effluent was
analyzed by gas chromatography (GC) first after 5 min time
on stream (TOS) and then in about 15-min intervals. The total
hexane conversion, X, was calculated from the composition of
the reactor effluent with exclusion of the n-hexane
In the DRIFTS experiments, UHP grade (>99.995%) He, N2,
and O2 were used after further purification. The helium flow
was passed through a 77 K column that was filled with activated
carbon. The N2 and O2 flows were contacted with dehydrated
molecular sieve 5A adsorbent. In the catalytic experiments,
analytical grade hexane (Aldrich, 99+ %) was used without
purification.
6
as X ) (1/6) ∑1 jXi, where Xi and j are the mole fraction and
DRIFT Spectroscopy. The DRIFTS measurements were
carried out by using a Nicolet 5PC spectrometer equipped with
a COLLECTOR II diffuse reflectance mirror system and a high-
temperature/high-pressure DRIFTS cell (Spectra-Tech, Inc.). The
sample cup of the cell was filled with finely powdered catalyst.
The catalyst was heated to 773 K at a rate of 10 K/min in a 50
cm3/min flow of 50% O2/N2 at atmospheric pressure. At 773
K, the sample was contacted with the gas flow for 1 h, then the
cell was flushed with a He flow for 10 min and cooled in the
He flow to 293 K. In this treatment, the catalysts were
dehydrated, the NH4-form zeolites were also deammoniated and,
thereby, converted into H-form.
A single-beam spectrum was determined while flushing the
catalyst sample in the DRIFTS cell with pure helium. Then the
He flow was switched to a flow of 50 cm3/min N2 flow. After
flushing the cell for 5 min, a second single-beam spectrum was
taken up. Under identical conditions, a single-beam spectrum
was recorded also with an alignment mirror in the sample cup.
The spectrum of the catalyst was obtained as the ratio of the
single-beam spectra recorded for the catalyst and the alignment
mirror. The spectrum of the adsorbed N2 and the adsorption-
induced spectral changes were obtained by generating the ratio
of the single-beam spectra of the catalyst, in contact with N2
and He. The gas pressure was controlled between 1 and 9 bar
by using a pressure regulator, downstream of the cell. All spectra
the carbon number of the ith component, respectively. The
activity of the catalyst on stream decreased in time. Initial
conversions (X0), were determined by extrapolating the conver-
sion results to zero TOS, using the empirical X vs TOS relation,
applied also by Williams et al.26 Assuming that the reaction is
first order in the reactant hexane (vide infra) the apparent rate
constant of the reaction, kB,app, was obtained as slope of -ln(1
- X0) vs W/F plots (Figure 1)
3. Results
The N2 adsorption coverage of the sorption sites was increased
by increasing the nitrogen pressure above the catalyst. The
increasing coverage was paralleled by the erosion of the original
νOH bands and by the concomitant rise of new νOH bands, shifted
to lower frequencies (Figures 2-7). The appearance of the
shifted bands indicated that the H-bonding of N2 weakened the
O-H bonds of some hydroxyl groups. As a rule, the OH groups
in a H-bond interaction have a higher absorption coefficient
and give a broader IR band than the unperturbed hydroxyls.
From the spectra, recorded at different N2 pressures, it is
sometimes difficult to resolve all the bands, which are diminish-
ing and emerging due to N2 adsorption (Figures 2A to 7A).
However, if only the adsorption-induced spectral changes are
presented, these bands can be clearly distinguished (Figures 2B
to 7B).

H-Zeolites and Sulfated Zirconia-Titania
J. Phys. Chem. B, Vol. 110, No. 4, 2006 1713
Figure 3. DRIFT spectra of zeolite H-mordenite (A) under N2 at
1-8 bar pressures and under He at 1 bar, and (B) the difference of
these spectra. Spectra correspond to those in Figure 2.
Figure 1. First-order plots for n-hexane isomerization and cracking
at 6.1 kPa n-hexane partial pressure and 553 K over H-ZSM-5, H-â,
H-mordenite (H-MOR), H-USY, H-Y, and sulfated zirconia-titania
(SZT). The insert shows an enlarged plot for the high-silica zeolites.
Figure 4. DRIFT spectra of zeolite H-â (A) under N2 at 1-8 bar
pressures and under He at 1 bar, and (B) the difference of these spectra.
Spectra correspond to those in Figure 2.
Figure 2. DRIFT spectra of zeolite H-ZSM-5 in the νOH frequency
region. Spectra (A) were recorded at 298 K, while the sample was in
equilibrium with nitrogen at 1, 2, 4, 6, and 8 bar or under He at 1 bar.
Spectra (B) were obtained by subtracting the spectrum recorded in He
from each spectrum recorded in N2. An arrow points to the band of
the bridging OH groups, eroding due to N2 adsorption, and to the
concomitantly developing, red-shifted νOH band.
was 2:1. There is no doubt that hydroxyl groups in the main
channels can H-bond nitrogen, but the participation of the 8-MR
hydroxyls in the N2 adsorption is a matter of discussion.3a,30 In
the present study, neither the unperturbed nor the shifted
stretching band of the bridging OH groups of H-mordenite,
shown in Figure 3, was resolved to components.
Zeolite H-Y and H-USY. Besides the band of the silanol
groups at 3743 cm-1, the bridging OH groups in the R and â
cages of zeolite H-Y give characteristic bands at 3640 and 3545
cm-1, respectively (Figure 5A).32 The OH band structure is more
complex for the ultrastabilized H-Y, referred to as H-USY
because, in this zeolite, a fraction of the bridging OH groups is
under the influence of cationic extraframework aluminum
(EFAl) species.33,34 The bands at 3625 and 3604 cm-1 were
unequivocally assigned to OH groups in the large supercages
(R cages), while those at 3545 and 3525 cm-1 were assigned
to hydroxyls in the small sodalite cages (â cages). The bands
at 3604 and 3545 cm-1 were attributed to hydroxyl groups,
affected by the presence of the Lewis acid EFAl species (Figure
6).35
Zeolite H-ZSM-5 and H-mordenite. Two characteristic
νOH bands were observed. The band near to 3745 cm-1 was
assigned to the νOH mode of terminal silanol groups. Terminal
silanol groups can be present either on the external surface of
the zeolite crystallites or at framework defects. At about 3610
cm-1, the stretching band of those OH groups appeared that
are bridging between framework Al and Si atoms (Figures 2A
and 3A).3,4,22,29 The N2 adsorption resulted in a 19-20 cm-1
red-shift of the silanol band. For silica fume (Aerosil), the
corresponding shift was found to be only 11 cm-1 (not shown).
The adsorption-induced drop of the stretching vibration fre-
quency was significantly higher for the bridging hydroxyls. A
new band appeared in the spectrum of both zeolites at 3529
cm-1, corresponding to a red-shift of 82-83 cm-1 (Figures 2B
and 3B).
In previous studies, the slightly asymmetrical νOH band of
H-mordenite was resolved to component bands at 3616 cm-1
and at about 3590 cm-1, using a curve-fitting computer
program.3a,30 The bands were assigned to the hydroxyl groups
present in larger main channels, having 12-membered rings (12-
MR), and in narrow 8-MR pores or at the 8-MR access to the
side pockets of the main channel.3a,31 The integrated absorbance
ratio of the high-frequency and low-frequency component bands
On H-bonding N2 the silanol stretching frequencies suffered
a 15-18 cm-1 red-shift (Figures 5 and 6). Generally larger red-
shifts were detected for the bridging OH groups, present in the
R cages, than for those in the â cages. The shifts were 45 and
25 cm-1, respectively, for those hydroxyls, which were not
perturbed by the electronic effect of the EFAl species. Those
hydroxyls, which were under the influence of cationic EFAl

1714 J. Phys. Chem. B, Vol. 110, No. 4, 2006
Lo´nyi et al.
Figure 7. DRIFT spectra of sulfated zirconia-titania (A) under N2 at
1-8 bar pressures and under He at 1 bar, and (B) the difference of
these spectra. Spectra correspond to those in Figure 2.
Figure 5. DRIFT spectra of zeolite H-Y (A) under N2 at 1-8 bar
pressures and under He at 1 bar, and (B) the difference of these spectra.
Spectra correspond to those in Figure 2.
Sulfated Zirconia-Titania (SZT). The major absorption
bands, observed in the OH region, were at 3643 cm-1 and at
3741 cm-1 (Figure 7A).
Zirconia (Z) and sulfated zirconia (SZ) samples were also
investigated for comparison (not shown). The νOH bands of the
Z were at 3772 cm-1 and at 3668 cm-1. Except for the lower
band intensities, the spectrum of the SZ was similar to that of
SZT. The 3643 cm-1 band was assigned to acidic OH groups,
while the 3741 cm-1 band was assigned to nonacidic OH
groups.13,14
N2 adsorption induced a 14-15 cm-1 red-shift in the OH
stretching frequencies of the Z sample. The small shift reflects
the weak acid character of these OH groups. In contrast, the
shifts of the 3643 cm-1 and the 3741 cm-1 bands of the SZT
were 48 and 18 cm-1, respectively (Figure 7). It is also important
to note that the νSdO band of the SZT sample at 1391 cm-1
was hardly effected by the N2 adsorption.
Spectrum of the Adsorbed N2. The vibration transitions of
homonuclear diatomics, such as N2, are IR inactive in the gas
phase. However, adsorption in the zeolite cavities lowers the
symmetry of the molecule and results in the appearance of an
interaction-induced IR spectrum.3-10,12,40
The integrated absorbance of νNN stretching band in the
2300-2350 cm-1 frequency region was found to increase
parallel with the erosion of the hydroxyl bands and the
appearance of the corresponding shifted OH bands, indicating
that the observed changes are related to the interaction between
the OH groups and the N2 molecules. The changes were fully
reversible for all the studied samples. These observations are
in agreement with earlier findings.3,4,7-9
Spectra of nitrogen adsorbed over H-zeolites and SZT are
shown in Figure 8. The 2330 cm-1 Raman frequency of the
gas-phase N2 is given as a reference value.41 The band of N2,
weakly bound to silica (Aerosil), was at 2332 cm-1 (not shown),
i.e., hardly apart from the gas-phase Raman frequency.
Two νNN absorption bands were distinguished (Figure 8), a
weak band in the 2340-2360 cm-1 region and a major band
near to 2330 cm-1. The weak band, which could not be traced
only in the spectrum of zeolite H-MOR and SZT, has been
assigned to N2 adsorbed on Lewis acid sites. The band at about
2330 cm-1 stems from N2, adsorbed on Brønsted acid sites.3-6,8
From the major band, two component bands could be resolved
at frequencies slightly above and below the gas-phase Raman
frequency by using a curve fitting computer program (Figure
8, thin lines).
Figure 6. DRIFT spectra of zeolite H-USY (A) under N2 at 1-8 bar
pressures and under He at 1 bar, and (B) the difference of these spectra.
Spectra correspond to those in Figure 2.
species, suffered a larger frequency shift. The 3604 cm-1 band
shifted by about 67 cm-1 to 3537 cm-1, while the band of the
â-cage hydroxyls at 3545 cm-1 shifted by about 45 cm-1 to
3500 cm-1 (Figure 6). The larger shifts are in line with the
higher acid strength of these hydroxyls.36 The N2 adsorption
and the Brønsted acidity of H-faujasites were discussed in a
recent DRIFT spectroscopic study.35
Zeolite H-â. In accordance with previous reports,37,38 bands
of silanol and bridging hydroxyl groups were observed at 3744
and 3609 cm-1, respectively (Figure 4A). The additional band
at 3663 cm-1 was associated with the OH groups linked to EFAl
species.32,33 A weak OH band was also detected as a shoulder
at 3780 cm-1 (Figure 4A, not labeled). This band was suggested
by Kiricsi et al.37 to stem from the intermediates of the EFAl
formation process.
The adsorption of nitrogen affected all the νOH bands. The
bands of the Si- and EFAl-bound hydroxyls shifted to lower
frequencies by 17-18 cm-1. Two shifted bands were generated
from the bridging OH groups: one at about 3570 cm-1 and
another at 3529 cm-1 (Figure 4B). Thus, the red-shifts, brought
about by N2 adsorption, were 39 cm-1 and 80 cm-1, respec-
tively. The results suggest that bridging hydroxyls of the H-â
zeolite can be characterized by two distinctly different acid
strengths. Because the bands resemble those of the R-cage
hydroxyls of H-USY, it was suggested that the acid strengths
of some of hydroxyls were enhanced by Lewis acid species.39

H-Zeolites and Sulfated Zirconia-Titania
J. Phys. Chem. B, Vol. 110, No. 4, 2006 1715
activities could not be calculated and compared. It was assumed
that the number of active sites is related to the total number of
Brønsted acid sites in the catalyst. The total number of Brønsted
acid sites was taken as equal to the framework Al content of
the zeolites and to the sulfate content of the SZT. Because the
fraction of the sites, which are accessible for the reactant hexane
and also active in the reaction, is not known, apparent site-
specific bimolecular rate constants were determined by relating
the apparent bimolecular rate constant, kB,app, to the aluminum
content of the catalyst sample (Table 3).
4. Discussion
Understanding the relation of the acidity and the catalytic
properties of catalysts helps in learning about the underlying
reasons of the catalytic activity. However, it is difficult to bridge
the gap between the equilibrium thermodynamic concept of
acidity and the composite kinetic concept of catalytic activity.42
The free energy change of the reaction between different solid
acids and a base is often used to locate the acid on an acidity
scale. However, this energy is determined by all the adsorption
interactions. A Lewis base can coordinate to Lewis acid sites,
polarize O-H bonds, or even abstract protons from OH groups.
Van der Waals forces and, in some cases, additional H-bonds,
contribute to the binding of the base to the surface of a solid.
Latter effects, generally referred to as solvating and medium
effects, are responsible for the deviation of the intrinsic and
the measured acidity of the OH groups. The medium effects
depend on the structure and composition of the adsorbent and
the base, moreover, on the measurement conditions, such as
the temperature and the coverage. For example, the effects are
more pronounced in micropores than in larger pores or over
open surfaces. As a result, solids can present varying relative
acidities against different bases.4 This is particularly important
to keep in mind when we compare the acidities of very different
solids, such as zeolites and sulfated zirconia, on the acidity scale
of a given base and want to draw conclusions about the behavior
of the same solids against another base, which can be, for
instance, a reactant in an acid-catalyzed reaction. In relation to
a catalytic reaction, the catalyst acidities are best characterized
by the interaction of the catalysts with the reactant itself.
However, an acidity scale can be established that is independent
of the chemical nature of any base on the bases of the
deprotonation energy (DE) of Brønsted acid sites. The difficulty
of using the DE scale is that the DE of a real catalyst cannot be
obtained by theoretical calculation; moreover, the experimental
characterization of intrinsic catalyst acidity is still a matter of
discussion.
The catalytic performance can be correlated with the acidic
property if the same catalytic mechanism dominates over each
compared catalyst and the correlation properly includes all
adsorption and diffusion effects that contribute to the activity.43
Thus, usually, not the apparent, but the intrinsic parameters of
the reaction, such as the intrinsic rate constant or activation
energy of an acid-catalyzed reaction, are the sufficient measures
of the intrinsic acid strength.
The Intrinsic Acid Strength of the Catalysts. It was
suggested that the ∆νOH shifts of the νOH IR band, induced by
N2 adsorption, was proportional with the unperturbed, intrinsic
Brønsted acid strength of the hydroxyl groups.6,7,12 Thus, the
∆νOH is a parameter that shows correlation with the deproto-
nation energy of the acid.1 This infers that, besides the
H-bonding interaction with the acidic hydroxyl groups, any
adsorption interaction of the N2 is virtually negligible. Usually,
low temperature (<100 K) was used to get a detectable amount
Figure 8. The DRIFT spectra of N2 adsorbed on H-ZSM-5,
H-mordenite, H-â, H-Y, H-USY, and sulfated zirconia-titania at
8 bar pressure in the νNN frequency region. Component bands were
resolved by using a curve fitting computer program.
An absorption band at 2338 cm-1 with a shoulder at 2328
cm-1 was observed on the spectrum of SZT. Interestingly, weak
νNN bands were obtained at the same frequencies from N2
adsorption on Z (not shown), while the N2-induced ∆νOH was
much smaller for Z than for SZ or SZT.
Catalytic Hexane Conversion. When in contact with the
hexane flow, catalysts became saturated with species formed
from hexane. Unconverted hexane appeared in the product
mixture only after an initial transient period of about 2-3 min
TOS. After that, each catalyst was rapidly losing activity. The
comparison of the catalytic activities involves some uncertainty
due to the activity change and the different time dependence of
this change. An extrapolation procedure was applied to the
results, obtained at different space velocities, to estimate the
conversions at zero time on stream. The conversion and product
distribution data, given in Table 2, were obtained at 5 min TOS.
Alkene product could not be detected. Pentane was formed over
each catalyst, but much less, if any, methane was found
compared to pentane. Propane and butanes were the major
products. Butanes were always in large excess to ethane. The
high selectivity for isobutane formation indicated that isomer-
ization and secondary cracking processes also affected the
product selectivity. The complex product distribution suggests
that, at the applied relatively low reaction temperature and high
hexane pressure (553 K and 6.1 kPa), the bimolecular mecha-
nism prevails.
The high-silica zeolites, namely H-ZSM-5, H-mordenite,
and H-â, were more active than the rest of the studied catalysts.
The hexane conversion over H-USY and SZT was comparable
to that over the most active H-â, but at 10 and 25 times longer
space times, respectively. Under similar reaction conditions, the
activity of H-Y was hardly measurable (Table 2). The first-
order plots of the high-silica zeolites are very close to each other
(Figure 1). Zeolite H-Y showed low activity. The apparent
bimolecular rate constant, kB,app, of zeolite H-USY and sample
SZT was intermediate between those of the high-silica zeolites
and H-Y (Table 3). Because only the 12-MR pores (0.67 nm
× 0.70 nm) of H-mordenite are accessible for alkanes, the pore
system of mordenite is one-dimensional with regard to the
alkane transport. Zeolite H-â has a three-dimensional system
of pores, having 12-MR apertures. The comparable activities
of these catalysts (Figure 1 and Table 3) substantiate that the
diffusion effect cannot reduce significantly the catalytic ef-
ficiency of the H-mordenite. Because the number of the
effective active sites (S′) is not known, exact site-specific

1716 J. Phys. Chem. B, Vol. 110, No. 4, 2006
Lo´nyi et al.
SZT
TABLE 2: Hexane Isomerization and Cracking over H-Zeolites and Sulfated Zirconia-Titaniaa
H-ZSM-5
H-mordenite
H-â
H-Y
H-USY
F/W, µmol‚gcat.-1 s-1
total conversion, mol %
cracking, mol %
20
20
20
0.8
2.0
16.39
6.93
9.46
0.8
5.82
4.99
0.83
8.52
6.03
2.48
12.68
7.35
5.33
0.86
0.23
0.63
15.41
11.13
4.28
isomerization, mol %
Product Distribution, mol %
CH4
C2H6
C3H8
C4H10
i-C4H10
C5H12
i-C5H12
i-C6H14
<0.02
0.9
25.7
25.9
20.6
10.9
7.0
2.4
<0.02
25.9
15.8
18.5
7.2
<0.02
<0.02
<0.02
10.1
0
12.1
6.1
1.1
<0.02
17.1
3.5
19.0
1.8
<0.02
1.0
37.4
3.9
26.3
1.7
10.8
18.9
0.4
14.1
6.2
29.0
5.7
12.0
32.6
9.7
20.5
6.1
65.6
10.7
46.8
9.0
a The space velocity of hexane, F/W, was decreased for the less active catalysts in order to get measurable conversions. The reaction was carried
out at a hexane partial pressure of 6.1 kPa at 553 K. The conversions in cracking and isomerization reactions, as well as the total conversion, were
calculated from the effluent composition, determined at 5 min time on stream as X ) (1/6) ∑16 jXi, where where Xi and j are the mole fraction and
carbon number of the ith component, respectively.
TABLE 3: Activity of Bimolecular Hexane Conversion and
Intrinsic Acidity (∆νOH) of Zeolite and Sulfated
Zirconia-Titania Catalysts
those in ref 7, N2 was found to H-bond also to the very weak
acid silanol groups. However, even at room temperature, the
â-cage protons of zeolite H-Y were of limited accessibility
for N2. It can be understood if we consider that the nitrogen
molecule has a kinetic diameter of 0.36 nm, while the diameter
of six-member ring windows on the â-cage cage is 0.26 nm.
Because the PA of nitrogen is only 493.8 kJ mol-1, much
smaller than the 928.8 kJ mol-1 PA of pyridine, N2 cannot pull
out protons from the â-cages. Surprisingly, a very small fraction
of the â-cage hydroxyls was found to H-bond N2. This was
attributed to the partly defected faujasite structure of the zeolite
Y sample.35 A relatively large number of the â-cage hydroxyl
groups were accessible in the zeolite H-USY, indicating that
ultrastabilization generates a strongly defected faujasite struc-
ture.35,45
Lewis acid sites are easily distinguished from Brønsted acid
sites in zeolite samples by analyzing the νNN absorption bands
in the 2340-2350 cm-1 frequency region3-6,8. The dominant
band around 2330 cm-1 was assigned to Brønsted-bound
nitrogen. The interpretation of the 2330 cm-1 band, containing
more than one component, is not straightforward. On adsorption,
the stretching band of diatomic homonuclear molecules, like
N2, becomes IR active and shifts relative to the Raman band of
the gas molecule. Cohen de Lara et al.46,47 associated the shift
with the orientation of the adsorption-induced dipole moment
with respect to the electric field. Accordingly, the shift is in
the direction of the higher frequencies (blue-shift) if the
molecular axis of the adsorbed molecule is oriented parallel with
the E electric field, and is in the direction of lower frequencies
(red-shift) if the molecule is oriented perpendicular or inclined
to the direction of E.
kB,app
)
kB,app/S′ ×
∆νOH for
acidic OH
kBKA′S′,a
103,b
Si/Al µmol‚gcat.-1 s-1
s-1
groups,c cm-1
H-â
12.6
15.5
7.38
6.11
7.06
0.60
0.025
1.32
6.03
6.06
3.28
39; 80
83
82
H-ZSM-5
H-mordenite 6.7
H-USY
H-Y
SZT
2.9 (5.2d)
0.13 (0.24e) (0.29j) 45;g 67h 25;i 45j
2.5
0.005
1.24
45g
48
n. a.k
a Apparent bimolecular rate constant, calculated as the slope of
-ln(1 - X0) vs W/F plots of Figure 1. b Apparent site-specific
bimolecular rate constant, obtained by relating the kB,app value to the
Al content of zeolites or to the sulfate content of the SZT, expressed
in µmol‚gcat.
.
c Shift of the νOH frequency of the acidic hydroxyl
-1
groups, induced by N2 adsorption at 298 K. d Si to framework-Al ratio.
e The kB,app was related to the framework Al content. f The kB,app was
related to the EFAl content. g ∆νOH of R-cage hydroxyls. h ∆νOH of
i
R-cage hydroxyls under the influence of EFAl species. ∆νOH of â-cage
hydroxyls. j ∆νOH of â-cage hydroxyls under the influence of EFAL
species. k Note: n.a. ) not applicable.
of adsorbed N2 at the applied relatively low pressure (<1 bar).
In harmony with IR and 1H NMR results, theoretical calculations
suggested that the mobility of the bridging protons in H-zeolites,
their H-bonding to neighboring framework oxygen atoms, and
as a consequence, their acid strength, depend on temperature.28
At higher temperatures, H-zeolites are weaker acids than at
lower temperatures, but show more homogeneous acid strength
distribution. At low temperature, the reduced mobility of the
adsorbate and the proton may hinder the adsorption on the less-
accessible sites. For example, it was found that even small
molecules, like N2, O2, or CO, did not interact with OH groups
located in the small sodalite cages of Y-zeolite.5-7,24 In contrast,
these hydroxyls protonated the much larger pyridine molecule
at room temperature.33,36 It was argued that the higher proton
affinity (PA) of the adsorbate and the higher adsorption
temperature made the interaction of pyridine and the â-cage
hydroxyls feasible.24,44
From adsorption at low-temperature on H-mordenite, νNN
component bands were obtained at 2333 and 2330 cm-1, which
were assigned to adsorption on the Brønsted acid sites in the
main mordenite channels and on the less-accessible acid sites
in the side pockets, respectively.30a At higher N2 coverage, a
third component was observed at 2325 cm-1, which was
assigned to N2 bound to silanol OH groups.7,44,30a
The temperature, that must be low to get detectable amount
of adsorbed N2 at low adsorption pressures, can be increased if
the equilibrium pressure is also increased. In the present study,
DRIFT spectra obtained from adsorption of nitrogen was
recorded at room temperature in the 1-9 bar pressure range.
The N2 was found to coordinate to the bridging OH groups of
H-zeolites and to the OH groups of zirconia. In agreement with
results of low-temperature IR measurements, for instance with
At variance with above results, we found component bands
at 2330 and 2325 cm-1, having an integrated absorbance ratio
of about 2:1 (Figure 8). One of the νNN component bands of
N2, bound to the Brønsted acid sites of H-mordenite, is near
to the gas-phase Raman frequency, while the other is at a lower
frequency (Figure 8). In accordance with the band assignment
of Cohen de Lara et al.,46,47 the higher-frequency band could
reflect parallel orientation of the induced N2 dipole to the electric

H-Zeolites and Sulfated Zirconia-Titania
J. Phys. Chem. B, Vol. 110, No. 4, 2006 1717
TABLE 4: Adsorption-Induced Shift of the OH bands
(∆νOH, cm-1) Assigned to Acidic OH Groups in Different
Zeolites and Sulfated Zirconia
consistent picture emerges for the zeolites. The H-ZSM-5,
H-mordenite, and H-â contain Brønsted acid sites, having
nearly identical acid strength. These zeolites are stronger acids
than zeolite H-Y. Some hydroxyls of the H-USY have
significantly stronger Brønsted acidity than the bridging hy-
droxyls of zeolite H-Y. Studies testing the acidity of sulfated
zirconia and sulfated zirconia-titania by using benzene and CO
as a basic adsorptive suggest that these materials are weaker
Brønsted acids than any of the mentioned zeolites. In contrast,
zirconia samples showed similar acid strength against nitrogen
than zeolite H-Y. The acidity of the samples against the
weakest base nitrogen was compared with the activity of the
catalysts in the isomerization and cracking of hexane. Higher
activity was obtained for the samples, characterized by larger
∆νOH (Table 3). However, the apparent rate constant of the
bimolecular hexane conversion showed variance for the most
active high-silica H-zeolites, characterized by similar intrinsic
acid strength. These differences may come from real differences
in the intrinsic activities of the active sites in the different
zeolites or, more probably, from the fact that the number of
active sites is neither equal nor proportional with the total
number of the Brønsted acid sites.
C6H6
(298 K)
CO
(77 K)
N2
(77 K)
N2
(298 K)m
H-ZSM-5
H-mordenite
H-â
350a,b
308-340d,e,f 121-122e,l 83
290-389f,g 107 l
82
352-365c 307-319e,h 119-126e,l 39; 80
H-Y
280-320a,b 273-296d,e,i 93-110e,l 45
H-USY
327-382d,e 139e
45; 67 (R-cage)
25; 45 (â-cage)
SZ or SZT
185-200a,b 140-220 j,k
-
48
a Reference 15. b Reference 17. c Reference 37. d Reference 50.
e Reference 24. f Reference 61. g Reference 11. h Reference 8. i Refer-
ence 2. j Reference 14. k Reference 13. l Reference 4. m This work.
field in the main channels, while the lower-frequency band may
correspond to adsorption in inclined orientation. The inclined
orientation could appear because of precessional motion of the
adsorbate molecule around the anchoring sorption center. This
motion may be more intense if the interaction is weaker and
concerns relatively more molecules at higher temperatures.47
The abundance of N2, bound inclined, can be increased also by
spatial constraints around the sorption sites, for instance around
the Brønsted acid sites in the side pockets of H-mordenite.
The band at 2325 cm-1, observed by Zecchina et al.7,30a and
Corma et al.,19 was well below the 2330 cm-1 gas-phase Raman
frequency. Instead of considering the possibility of red-shift,
the frequency of 2324 cm-1 (ref 7) or 2321 cm-1 (ref 30a) was
selected as a reference value. The 2324 cm-1 was obtained as
the average of the values, calculated for the condensed state,
and measured in rare gas matrixes.7 The 2321 cm-1 is the
measured Raman frequency of N2 adsorbed on silicalite, that
is, a purely siliceous ZSM-5.30a All the νNN component bands
were then considered as blue-shifted. We found that the νNN
band of N2, adsorbed on SiO2 (Aerosil) at room temperature,
appeared blue-shifted by 2 cm-1 relative to the 2330 cm-1
reference frequency. Result show that, in the absence of spatial
constraints, N2 is bound with a molecular axis parallel to the
electric field on the surface of SiO2 and a positive shift can be
observed. In our opinion, it is more rational to use the gas-
phase Raman frequency as a reference value and attribute any
red-shift to inclined orientation of the molecule with respect to
the electric field of the adsorbing surface, as suggested by Cohen
de Lara.47
On adsorption, a Lewis base can coordinate to Lewis acid
sites, polarize the O-H bonds, or even abstract protons from
the OH groups. Van der Waals forces and, in some cases,
additional H-bonds, contribute to the stabilization of the base
on the surface. Latter effects, generally referred to as solvating
and medium effects, are responsible for the deviation of the
intrinsic and the measured acidities. The medium effects are
more pronounced in micropores than in larger pores or on open
sufaces. The adsorption of a base on a solid acid must increase
the electron density on the conjugated base of the acid and,
thereby, affect both the basicity of the adsorbate and the acid
strength of the Brønsted acid sites. The νSdO band of the sulfate
species in the sulfated zirconia samples was found to indicate
these changes.13,17 If the electron density in the SdO bond
increases, the bond loses its covalent character and becomes
more ionic. As a result, the νSdO band shifts to lower
wavenumbers. It has been suggested that this shift takes place
on the inductive effect of the adsorbate, bound to coordinatively
unsaturated Zr4+ surface sites.13,48 The extent of the shift
depends on the basicity of the adsorptive. The adsorption of
benzene induced a 15 cm-1 ∆νSdO red-shift.17 Upon CO
adsorption, the shift was 14 cm-1 13 In contrast, adsorption of
.
Downward- and upward-shifted νNN component bands, similar
to those found for H-mordenite, were observed on all the
studied samples (Figure 8). The assignment of the component
bands to different Brønsted acid sites seems to be ambiguous
at present. Unexpectedly, two νNN component bands were
obtained from adsorption on the mesoporous sulfated zirconia-
titania sample (Figure 8). Adsorption of N2 on the nonsulfated
zirconia, not containing strongly acidic OH groups, gave two
similar bands. The bands below and above the Raman frequency
indicate that the N2 is adsorbed with the molecular axis having
both parallel and inclined orientation relative to the electric field.
For the studied solid acids having different structures, no
correlation was found between the shifts of the corresponding
νNN and νOH bands.
In Table 4, ∆νOH data are given for different acid-base
interactions. Our results are shown together with related data,
collected from the literature, for adsorptives such as benzene
(PA ) 750.4 kJ mol-1), CO (PA ) 594.0 kJ mol-1), and N2
(PA ) 493.8 kJ mol-1). The small differences in the corre-
sponding ∆νOH shifts can be attributed to differences of the
samples, studied in different laboratories. From the data, a
N2 on sulfated zirconia-titania resulted in about a 1 cm-1 red-
shift of the νSdO frequency. These results unambiguously show
that the basicity of the conjugate base increases during the
interaction with benzene and CO, while N2 virtually does not
affect the conjugate base of the Brønsted acid. From this aspect,
N2 can be considered as a good basic probe to characterize the
intrinsic acid strength of the Brønsted sites.
According to the extent of the ∆νOH shift, measured by room-
temperature adsorption of N2 at elevated pressures (Table 4,
last column), the following order of intrinsic Brønsted acidity
was obtained: H-ZSM-5 ≈ H-mordenite ≈ H-â > H-USY
> SZT ≈ H-Y. For similar samples, using molecular probes
of different basicity, Paze et al.8 obtained linear correlation
between the ∆νOH of the stretching band of the strong Brønsted
acid groups and that of the silanol groups. These plots,
corresponding to BHW plots,49 suggested the same acidity order.
It was also concluded that the Brønsted sites of H-ZSM-5,
H-mordenite, and H-â are energetically uniform.4,24 However,
the acid strength measurements of the present work do not
support the uniformity of the H-â hydroxyls.

1718 J. Phys. Chem. B, Vol. 110, No. 4, 2006
Lo´nyi et al.
By using the IR method and CO or N2 as an adsorptive probe,
Kotrel et al.24 concluded that H-USY was a stronger Brønsted
acid than any of the mentioned zeolites. However, the activity
of the H-USY in the hexane isomerization and cracking was
in conflict with this finding.24,50 The present study seems to
resolve this contradiction. The ∆νOH shift, induced by N2
adsorption, was smaller for H-USY than for H-â, which is in
agreement with the hexane conversion activities (Table 3).
of Aa governs the rate of carbenium ion and product formation.
kU
kU
9
H2
Aa + H+ 98 [A+]
8 E+ +
(3)
( )
a
Symbol A+ represent the transition state of the reaction that, as
a comfortable convention, can be visualized as protonated alkane
CR2H3+. E+ stands for CR2H+ and CRH2+ carbenium ions, and
a for product alkanes.
The conflicting results seem to rise from the difficulty of
assigning the corresponding original and shifted νOH IR bands
if the OH envelope is as complex as that of the H-USY (Figure
6). In this relation, it is important to notice that H-USY is a
defective structure wherein the â-cage hydroxyls are accessible
both for small adsorbate probe molecules and reactants.45,51,52
To get the right band assignment, our understanding of the νOH
band structure of zeolite H-Y was useful.35 In the H-USY,
some of the OH groups are inductively affected by cationic EFAl
species. Therefore, the spectrum shows four IR bands. Instead
of the four additional bands, expected to be obtained upon N2
adsorption, only three new bands could be resolved (Figure 5B).
The band at about 3500 cm-1 was attributed to the overlapping
shifted bands of the â-cage hydroxyls. The band shift of â-cage
hydroxyls is generally smaller than that of R-cage hydroxyls.
This finding is in accordance with the conclusion of Biaglow
It is assumed here that each carbenium ion, formed in the
reaction of eq 3, establishes equilibrium with the corresponding
alkene (eq 4)
-1
E+ [KE ] E + H+
(4)
where E is alkene and KE is the equilibrium constant of alkene
adsorption. The active site concentration S was assumed to
remain constant during the reaction and was expressed as the
sum of the carbocation and active proton concentration (eq 5).
S ) [A+] + [E+] + [H+]
(5)
Quantum chemical calculations showed that solid acids
stabilize short alkyl chains as a covalently bound alkoxide.55,56
The alkoxide groups were suggested to be in equilibrium with
the corresponding surface-bound carbenium ion and the H-
bonded alkene. Nevertheless, the rate expressions are indepen-
dent of the identity of the reactive intermediates or transition
states. If, relative to the rate of the processes of eq 3, the alkane
and alkene adsorption-desorption (eqs 2 and 4) are rapid
processes, maintaining the equilibrium concentration of the
concerned species, the rate of the unimolecular reaction, rU, can
be expressed as
53
et al. that the hydroxyls in the smaller â-cage are weaker
Brønsted acids than those in the larger R cage. The acidity of
hydroxyls can be affected by H-bonding interactions with the
framework oxygen atoms. This interaction is obviously more
extensive in the â than in the R cage of zeolite H-Y.10,54
Activity-Acidity Relations. Recent studies of hexane con-
version over H-zeolites demonstrated that products could be
formed in both unimolecular and bimolecular processes. The
kinetics of the reaction was described by the Langmuir rate
law.26 The reaction rate (r) was given as
kUKASpA
rU )
(6)
1 + KE pE + KA pA
r ) kUΘA + kBΘE pA
(1)
At low coverage, the denominator approaches to one, and the
rate is
where kU and kB are the intrinsic unimolecular and bimolecular
rate constants, ΘA and ΘE represents the alkane and alkene
coverage of the active sorption sites, respectively, while pA is
the hexane partial pressure. It was shown that, at low conversion
and alkane partial pressure, the alkene coverage is low and the
unimolecular mechanism dominates.23-27 At low coverage, the
Henry rule applies and ΘA) KA pA, where KA is the equilibrium
constant of the hexane adsorption. Thus, in accordance with
the experimental findings, the reaction is first order in the
concentration of the reactant, and the reaction rate is proportional
with an apparent rate constant kU,app ) kU KA. The appearance
of methane and ethane among the alkane products and formation
of some hydrogen is characteristic of the unimolecular cracking
process. Also, hydrogen and alkanes are formed in an amount
close to equimolar to the alkenes. Eventual alkene excess
indicates significant â scission of the carbenium ion intermediate
of the reaction.
rU ) kUKASpA
(7)
Equations 6 and 7 are identical with those derived by Williams
et al.26,27
According to the measurements of Babitz et al.,25 the site-
specific apparent rate constants and the corresponding activation
energies, obtained for the hexane cracking over different high-
silica H-zeolites, such as H-ZSM-5, H-mordenite, H-â, and
H-USY, could be largely attributed to differences in the heats
of hexane adsorption, while the intrinsic activation energies were
found to be nearly the same. Data suggested that either the
intrinsic activation energy of the reaction is insensitive to
differences in acid strength or that there is no difference in the
Brønsted acid strength of the different H-zeolites. In contrast,
Kotrel et al.23,24 reported distinctly different intrinsic activation
energies for the same reaction over H-zeolite catalysts, having
different structures and compositions, and except for H-USY,
showed correlation between the intrinsic activity and the intrinsic
acid strength of the catalysts, which was characterized by the
red-shift of the νOH IR bands, induced by the H-bonding of weak
base adsorbates.
The product distribution of the unimolecular cracking process
can be interpreted by the protolysis of C-C and C-H bonds.
The process steps are alkane adsorption (eq 2), protonation and
dehydrogenation/cracking (eq 3), and alkene desorption (eq 4)
A [K
\
A] Aa
(2)
Under conditions of bimolecular alkane conversion, the active
catalyst surface must be extensively covered by a carbenium
ion/alkoxy intermediate. These surface species can be considered
as sites forming activated intermediates with the reactant hexane
(eq 8). The reaction is generally referred to as bimolecular,
where A stands for the reactant alkane, CR2H2. The reactant,
adsorbed on the catalyst, is given as Aa. The protonation rate

H-Zeolites and Sulfated Zirconia-Titania
J. Phys. Chem. B, Vol. 110, No. 4, 2006 1719
because it proceeds with the participation of two molecular
species. The product distributions obtained at the relatively high
hexane pressure and conversion and the low reaction temper-
ature, applied in this study, suggested that, over the examined
solid acid catalysts, bimolecular isomerization/cracking mech-
anism prevailed.23,26 The bimolecular rate constant was found
to be about 10 times higher than the unimolecular rate constant,
while both processes were found to be first order in the
hydrocarbon concentration.23
The reaction is zero order in alkene. Notice that formally, the
same rate expression was obtained for the bimolecular cracking
at high alkene coverage as for the unimolecular cracking at low
coverage.
Equation 13 provides an explanation for the finding that the
apparent rate constant kB,app ) kBKA′ changes parallel with the
intrinsic acidity of the catalysts (Table 3) because the surface
concentration of the transition state EA+ carbonium ion is low,
i.e., KA′ must be small relative to kB. Moreover, KA′, character-
izing the alkane/E+-zeolite interaction have, presumably, much
weaker dependence on the zeolite composition and structure
than the alkane/H+-zeolite interaction, characterized by KA.
Consequently, while the intrinsic rate constant of the unimo-
lecular cracking, kU, must show better correlation with intrinsic
acidity than the apparent rate constant, kU,app, both the corre-
A significant finding of the present work was that intrinsic
acidity of the Brønsted acid sites, characterized by ∆νOH
,
induced by N2 adsorption, changed parallel with the apparent
site-specific rate constant of the bimolecular hexane cracking
(Figure 1 and Table 3). The SZT catalyst, showing much higher
activity than that corresponding to its acid strength, is an
exception. It seems quite probable that the criterion of an
identical reaction mechanism does not apply for the compared
SZT and zeolite catalysts.57,58 The correlation found for the
zeolites infers that the apparent rate constant hardly or similarly
deviates from the intrinsic rate constant. The kinetic analysis,
given below, helps to understand the above finding and points
out specific features of the bimolecular cracking process. With
the same symbols, used above, the elementary reactions of the
catalytic cycle can be written as
sponding rate constants of the bimolecular reaction, kB and kB,app
,
are expected to change parallel with the intrinsic acidity,
provided that the alkene coverage is high.
In the simplified kinetic treatment, given above, a represent
different alkane products. Therefore, E+ must represent a
number of different active site carbenium ions (eq 3). In the
bimolecular reaction, the interaction of the E+ active sites with
reactant and product alkanes gives a large variety of EA+ and
Ea+ transition states that are responsible for the found product
distribution (Table 2).
Under conditions of unimolecular transformations, the catalyst
was found to retain its activity for hours, allowing the deter-
mination of Arrhenius plots and, thereby, the apparent preex-
ponential constant (Aapp) and activation energy (Eapp) of the
reaction. If kU,app ) kUKAS, it can be easily shown that
E+ + A [KA′] (EA)+
(8)
and
kB
(EA)+ 98 E+ + a
(9)
where (EA)+ represents the carbonium ions, formed from any
alkane and any kind of E+ surface carbenium ion, and kB is the
intrinsic rate constant of bimolecular cracking. The equilibrium
of the alkenes and the carbenium ions, E and E+, respectively,
is given again by eq 4. The condition of constant active site
concentration is given as
ln AU,app ) ln AU,int + (∆SA/R)
EU,app ) EU,int + ∆HA
(14)
(15)
where Au,int and EU,int are the intrinsic preexponential constant
and activation energy, and ∆HA and ∆SA are the adsorption
enthalpy and entropy, respectively.
The proton transfer to a base is a concerted process wherein
bonds are broken and new bonds are formed. The net energy
of the process is favorably calculated on the route wherein (i)
the Brønsted acid site is deprotonated, (ii) the proton is added
to the alkane, and (iii) the protonated alkane is adsorbed on the
conjugated base of the acid. Processes (i) and (ii) involve the
deprotonation energy (DE) of the acid and the proton affinity
(PA) of the base, respectively. The sum of DE and PA is referred
to as the proton-transfer energy, ∆Hpt. The DE of zeolites was
S′ ) [EA+] + [E+] + [H+]
(10)
From eqs 4 and 9-11, the rate of bimolecular cracking, rB
) kB [(EA)+], is obtained in the form of
kBKEKA′S′pE pA
rB )
(11)
1 + KE pE + KA′KE pE pA
This equation seems to be at variance with the rate equation
given by Williams et al.26 The concentration of the transition
state carbonium ions is necessarily low, i.e., KA′pA , 1. It is
worth examining the extreme cases of low and high alkene
coverage. For low alkene coverage, the denominator approaches
one, and the rate, in accordance with Williams et al.,26 can be
given as
determined to be about 1200 kJ mol-1 59
, and the PA of hexane
is 665 kJ mol-1. The adsorption environment provides energy
to the Brønsted acid site/alkane ensemble to facilitate proton
transfer. The energies associated with process (iii) are generally
referred to as stabilization energy, ∆Hst. The stabilization energy
includes all the energies of adsorbate-adsorbate and adsorbate-
adsorbent interactions. The proton transfer from H-zeolites to
strong bases, such as ammonia, is an exothermic process. For
weak base alkanes, the energy balance gives the Ept activation
energy of the process.
rB ) kBKEKA′S′pEpA
(12)
This equation was shown to describe the dependence of the
reaction rate on the alkene concentration over H-USY catalyst
in the range of 0.1-6.2 kPa hexane pressure at 673 K.26 It seems
rational to think that, under the conditions applied in this study,
i.e., at 6.1 kPa hexane pressure and 553 K, the steady-state
alkene coverage of the active Brønsted acid sites is high, i.e.,
KE pE . 1, and thus, eq 11 can be reduced to get
Ept ) ∆Hpt + ∆Hst
(16)
∆Hst is the energy of interaction between the gas-phase
carbonium ion and the oxide anion of the solid acid. For a given
base, ∆Hpt depends only on the intrinsic acidity of the solid
acid, while ∆Hst includes all the structure-dependent energy
contributions.
rB ) kBKA′S′pA
(13)

1720 J. Phys. Chem. B, Vol. 110, No. 4, 2006
Lo´nyi et al.
Provided that the hexane activation occurs through full proton
transfer, EU,app must be equal to Ept. The measured activation
energies of the unimolecular hexane conversion over different
The intrinsic site-specific rate constant of hexane conversion
is attained if the apparent rate constant is corrected by the
equilibrium constant of hexane adsorption and related to the
number of active sites. In the present study, intrinsic activities
of hexane conversion were not obtained. First-order bimolecular
apparent rate constants were determined and related to the total
number of the acid sites. No correction was made for the
adsorption enrichment of hexane on the catalyst surface. After
all, both the apparent and the site-specific apparent rate constants
showed fair correlation with the intrinsic acidity of the catalysts.
The results suggest that the adsorption effect on the rate of
hexane conversion cannot be strongly dependent on the catalyst
composition and structure if the bimolecular transformation is
the dominating reaction mechanism.
H-zeolites ranged from 149 to 186 kJ mol-1 25
. Thus, the energy,
∆Hst, needed to facilitate full proton transfer to hexane, is
estimated to 349-386 kJ mol-1. ∆Hst is obviously much higher
than the approximately 70-90 kJ mol-1 value of ∆HA.23,25
Comparison of eqs 15 and 16 suggests that, in case of full proton
transfer, EU,int must include a considerable fraction of the
structure-sensitive stabilization energy, and it cannot be expected
to be identical for zeolites, having different structure and
composition. Nevertheless, EU,int was found to be independent
from the structure and composition of high-silica zeolites.25 It
seems quite probable, therefore, that alkane isomerization and
cracking proceeds through a reaction route having an intrinsic
activation energy much lower than ∆Hpt. Data substantiate that
alkane activation for unimolecular cracking and dehydrogenation
(eq 3) does not require full proton transfer, but alkane H-bonding
to acidic hydroxyl groups. The similarity of H-bonding interac-
tion of the weak base reactant alkane and the even weaker base
adsorptive N2 explains the good correlation between EU,int and
the adsorption-induced ∆νOH IR shifts.
Acknowledgment. Thanks are due for the financial support
of the Hungarian Research Fund (OTKA nos. M 36939 and T
043552). One of the authors (F.L.) expresses his appreciation
for the financial support of the Ja´nos Bolyai Foundation,
Hungary.
References and Notes
Because of the rapid deactivation of the catalysts, it was not
attempted to determine the Arrhenius parameters of the bimo-
lecular hexane conversion. It would have been also dubious to
assign the apparent activation energy to any specific step of
the complex bimolecular catalytic chain process. It was shown
by Lunsford et al.23 that the intrinsic rate constant of the
bimolecular hexane conversion was about 10 times larger than
that of the unimolecular process on the same catalyst. In the
present study, zeolites H-ZSM-5, H-mordenite, and H-â,
having stronger intrinsic Brønsted acidity, were found to be more
active than the H-Y, H-USY, and SZT catalysts having weaker
Brønsted acid sites (Figure 1 and Table 3). The found correlation
of kB,app and the intrinsic acidity may indicate that the deviation
of kB,app and kU,app cannot be related exclusively to the deviation
of the preexponential factors only, but also EB,app must be lower
than EU,app. It is to be noted that associated molecules have
higher proton affinity than single molecular species. Therefore,
EA+ transition state of the bimolecular process must have a
higher stability than the A+ transition state of the unimolecular
process (eqs 3 and 8). A decrease of the activation energy as
low as about 10 kJ mol-1 can already account for the found
rate difference of the uni- and bimolecular processes. Theoretical
calculations can support or negate this notion.
(1) Paukshtis, E. A.; Yurchenko, E. N. React. Kinet. Catal. Lett. 1981,
16, 131.
(2) Kubelkova´, L.; Beran, S.; Lercher, J. A. Zeolites 1989, 9, 539.
(3) (a) Wakabayashi, F.; Kondo, J. N.; Wada, A.; Domen, K.; Hirose,
C. J. Phys. Chem. 1993, 97, 10761. (b) Wakabayashi, F.; Kondo, J. N.;
Domen, K.; Hirose, C. Catal. Lett. 1993, 21, 257. (c) Wakabayashi, F.;
Kondo, J. N.; Domen, K.; Hirose, C. Stud. Surf. Sci. Catal. 1994, 90, 157.
(d) Wakabayashi, F.; Kondo, J. N.; Domen, K.; Hirose, C. J. Phys. Chem.
1995, 99, 10573.
(4) Wakabayashi, F.; Kondo, J. N.; Domen, K.; Hirose, C. J. Phys.
Chem. 1996, 100, 4154.
(5) Wakabayashi, F.; Kondo, J. N.; Domen, K.; Hirose, C. Microporous
Mater. 1997, 8, 29.
(6) Kno¨zinger, H.; Huber, S. J. Chem. Soc., Faraday Trans. 1998, 4,
2047.
(7) Zecchina, A.; Area´n, C. O. Chem. Soc. ReV. 1996, 25, 187.
(8) Paze´, C.; Bordiga, S.; Lamberti, C.; Salvalaggio, M.; Zecchina, A.;
Belussi, G. J. Phys. Chem. B 1997, 101, 4740.
(9) Bordiga, S.; Palomino, G. T.; Paze´, C.; Zecchina, A. Microporous
Mesoporous Mater. 2000, 34, 67.
(10) Makarova, M. A.; Ojo, A. F.; Kharim, K.; Hunger, M.; Dwyer, J.
J. Phys. Chem. 1994, 98, 3619.
(11) Jiang, M.; Karge, H. G. J. Chem. Soc., Faraday Trans. 1996, 92,
2641.
(12) Coluccia, S.; Marchese, L.; Martra, G. Microporous Mesoporous
Mater. 1999, 30, 43.
(13) Spielbauer, D.; Mekhemer, G. A. H.; Zaki, M. I.; Kno¨zinger, H.
Catal. Lett. 1996, 40, 71.
(14) Platero, E. E.; Mentruit, M. P.; Area´n, C. O.; Zecchina, A. J. Catal.
1996, 162, 268.
(15) Kustov, L. M.; Kazansky, V. B.; Figueras, F.; Tichit, D. J. Catal.
1994, 150, 143.
(16) Adeeva, V.; de Haan, J. W.; Ja¨nchen, J.; Lei, G. D.; Schu¨nemann,
V.; van de Ven, L. J. M.; Sachtler, W. M. H.; van Santen, R. A. J. Catal.
1995, 151, 364.
(17) Lo´nyi, F.; Valyon, J.; Engelhardt, J.; Mizukami, F. J. Catal. 1996,
160, 279.
(18) Song, X.; Sayari, A. Catal. ReV.sSci. Eng. 1996, 38, 329.
(19) Corma, A.; Forne´s, V.; Juan-Rajadell, M. I.; Lo´pez Nieto, J. M.
Appl. Catal., A 1994, 116, 151.
(20) Yaluris, G.; Larson, R. B.; Kobe, J. M.; Gonza´les, M. R.; Fogash,
K. B.; Dumesic, J. A. J. Catal. 1996, 158, 336.
(21) Barthos, R.; Lo´nyi, F.; Onyestya´k, Gy.; Valyon, J. J. Phys. Chem.
B 2000, 104, 7311.
(22) Ward, J. W. In Zeolite Chemistry and Catalysis; Rabo, J. A., Ed.;
ACS Monograph 171, American Chemical Society, Washington, DC, 1976;
p 118.
(23) Kotrel, S.; Rosynek, M. P.; Lunsford, J. H. J. Phys. Chem. B 1999,
103, 818.
(24) Kotrel, S.; Lunsford, J. H.; Kno¨zinger, H. J. Phys. Chem. B 2001,
105, 3917.
(25) Babitz, S. M.; Williams, B. A.; Miller, J. T.; Snurr, R. Q.; Haag,
W. O.; Kung, H. H. Appl. Catal., A 1999, 179, 71.
Conclusions
The aim of studying the relation between the acidity and the
alkane conversion activity of solids is to gain better understand-
ing of the catalytic mechanism. To establish meaningful
correlation for catalysts having different composition and
structure, intrinsic acidities and activities had to be compared.
The controversial results, reported here and in earlier studies,
come from the difficulty of determining intrinsic properties.
The present study shows that the ∆νOH shift of the νOH band,
induced by adsorption of N2 at room temperature, is a good
measure of the intrinsic acid strength of the Brønsted acid sites
of solid acid catalysts. The intrinsic acidity of the high-silica
H-ZSM-5, and H-mordenite, was found to be stronger than
that of H-USY and sulfated zirconia-titania and virtually
independent of the zeolite composition and structure. Zeolite
H-â has two kinds of Brønsted acid sites, which are distinctly
different by their intrinsic acid strengths. The stronger acid sites
are similar to those in H-ZSM-5 and H-mordenite. Zeolite
H-Y is a relatively weak solid acid.

H-Zeolites and Sulfated Zirconia-Titania
J. Phys. Chem. B, Vol. 110, No. 4, 2006 1721
(26) Williams, B. A.; Ji, W.; Miller, J. T.; Snurr, R. Q.; Kung, H. H.
Appl. Catal., A 2000, 203, 179.
(27) Ramachandran, C. E.; Williams, B. A.; van Bokhoven, J. A.; Milller,
J. T. J. Catal. 2005, 233, 100.
(45) Makarova, M. A.; Garforth, A.; Zholobenko, V. L.; Dwyer, J.; Earl,
G. J.; Rawlence, D. Stud. Surf. Sci. Catal. 1994, 84, 365.
(46) Jousse, F.; Cohen de Lara, E. J. Phys. Chem. 1996, 100, 233.
(47) Cohen de Lara, E. Phys. Chem. Chem. Phys. 1999, 1, 501.
(48) Jin, T.; Yamaguchi, T. Y.; Tanabe, K. J. Phys. Chem. 1986, 90,
4794.
(49) Bellamy, L. J.; Hallam, H. E.; Williams, R. L. Trans. Faraday Soc.
1958, 54, 1120.
(50) Makarova, M. A.; Al-Ghefaili, K. M.; Dwyer, J. J. Chem. Soc.,
Faraday Trans. 1994, 90, 383.
(51) Kung, H. H.; Williams, B. A.; Babitz, S. M.; Miller, J. T.; Haag,
W. O.; Snurr, R. Q. Top. Catal. 2000, 10, 59.
(52) Janssen, A. H.; Koster, A. J.; de Jong, K. P. J. Phys. Chem. B
2002, 106, 11905.
(53) Biaglow, A. I.; Parillo, D. J.; Kokotailo, G. T.; Gorte, R. J. J. Catal.
1994, 148, 213.
(54) Jacobs, P. A.; Mortier, W. J. Zeolites 1982, 2, 226.
(55) Kazansky, V. B.; Senchenya, I. N. J. Catal. 1989, 109, 108.
(56) Kazansky, V. B.; Frasch, M. V.; van Santen, R. A. Stud. Surf. Sci.
Catal. 1997, 105, 2283.
(28) Sillar, K.; Burk, P. J. Phys. Chem. B 2004, 108, 9893.
(29) Jacobs, P. A.; von Ballmoos, R. J. Phys. Chem. 1982, 86, 3050.
(30) (a) Geobaldo, F.; Lamberti, C.; Ricchiardi, G.; Bordiga, S.;
Zecchina, A.; Turnes, G.; Palomino, G. T.; Area´n, C. O. J. Phys. Chem.
1995, 99, 11167. (b) Wakabayashi, F.; Kondo, J. N.; Wada, A.; Domen;
K.; Hirose, C. J. Phys. Chem. 1996, 100, 18882. (c) Zecchina, A.; Geobaldo,
F.; Lamberti, C.; Ricchiardi, G.; Bordiga, S. J. Phys. Chem. 1996, 100,
18883.
(31) (a) Alberti, A. Zeolites 1997, 19, 411. (b) Niwa, M.; Suzuki, K.;
Katada, N.; Kanougi, T.; Atoguchi, T. J. Phys. Chem. B 2005, 109, 18749.
(32) Jacobs, P. A.; Uytterhoeven, J. B. J. Chem. Soc., Faraday Trans.
1 1973, 69, 359.
(33) Fritz, P. O.; Lunsford, J. H. J. Catal. 1989, 118, 85.
(34) Sun, Y.; Chu, P.-J.; Lunsford, J. H. Langmuir 1991, 7, 3027.
(35) Lo´nyi, F.; Valyon, J.; Pa´l-Borbe´ly, G. Microporous Mesoporous
Mater. 2003, 66, 273.
(36) Lo´nyi, F.; Lunsford, J. H. J. Catal. 1992, 136, 566.
(37) Kiricsi, I.; Flego, C.; Pazzuconi, G.; Parker, W. O.; Millini, R.;
Perego, C.; Bellussi, G. J. Phys. Chem. 1994, 98, 4627.
(38) Vimont, A.; Thibault-Starzyk, F.; Lavalley, J. C. J. Phys. Chem. B
2000, 104, 286.
(39) Lo´nyi, F.; Kova´cs, A.; Valyon, J. Stud. Surf. Sci. Catal. 2005, 158,
869.
(40) Fo¨rster, H.; Schuldt, M. J. Chem. Phys. 1977, 66, 5237.
(41) Huber, K. P.; Herzberg, G. Constants of Diatomic Molecules,
Molecular Spectra and Molecular Structure; van Nostrand Reinhold: New
York, 1979; Vol. 4.
(42) Farneth, W. E.; Gorte, R. J. Chem. ReV. 1995, 95, 615.
(43) Van Santen, R. A.; Kramer, G. J. Chem. ReV. 1995, 95, 637.
(44) Daniell, W.; Topsoe, N. Y.; Kno¨zinger, H. Langmuir 2001, 17,
6233.
(57) Kim, S. Y.; Goodwin, J. G., Jr.; Fa˜rcas¸iu, D. Appl. Catal., A 2001,
207, 281.
(58) Hong, Z.; Fogash, K. B.; Dumesic, J. A. Catal. Today 1999, 51,
269.
(59) (a) Kazansky, V. B. Acidity and Basicity of Solids: Theory,
Assessment and Utility; Fraissard, J., Petrakis, L., Eds.; Kluwer Academic
Publishers: London, 1994; p 335. (b) Kazansky, V. B. Acc. Chem. Res.
1991, 24, 379.
(60) Umansky, B.; Engelhardt, J.; Hall, W. K. J Catal. 1991, 127, 128.
(61) Garrone, E.; Chiapetta, R.; Spoto, G.; Ugliengo, P.; Zecchina, A.;
Fajula, F. Proc. 9th Int. Zeolite Conf., Montreal 1992; von Balmoos, R.,
Higgins, J. B., Tracy, M. M. J., Eds.; Butterworth-Heinemann: London,
1992; Vol. 2, p 267.