The influence of catalyst acid strength on the methanol to hydrocarbons (MTH) reaction

Article abstract of DOI:10.1016/j.cattod.2013.03.017

The methanol to hydrocarbons (MTH) reaction was studied over two isostructural zeotype catalysts of different acid strength, H-SAPO-5 and H-SSZ-24. Conversion of methanol alone was performed at 350-450 C and WHSV = 0.31-2.48 h^-1. The product selectivities of the two catalysts were compared at similar conversion. The strongly acidic H-SSZ-24 was found to be more selective towards aromatic products and C₂-C₃ hydrocarbons as compared to the moderately acidic H-SAPO-5, which produced more non-aromatic C₄₊ hydrocarbons. Co-reactions of ¹³CH ₃OH and benzene at 250-300 C with low conversion of both reactants revealed that both catalysts produced ethene and propene from polymethylbenzenes via a paring mechanism. However, this reaction proceeded more readily in H-SSZ-24 than in H-SAPO-5. Furthermore, isobutene formation was found to be mainly associated with aromatic intermediates in H-SSZ-24, whereas isobutene produced over H-SAPO-5 was mainly formed via alkene intermediates. Overall, the results obtained in this study suggest that a lower acid strength promotes an alkene-mediated MTH reaction mechanism.

Full text of DOI:10.1016/j.cattod.2013.03.017

Catalysis Today 215 (2013) 216–223
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
The influence of catalyst acid strength on the methanol to
hydrocarbons (MTH) reaction
Marius Westgård Erichsen, Stian Svelle, Unni Olsbye∗
inGAP Centre of Research Based Innovation, Department of Chemistry, University of Oslo, N-0315 Oslo, Norway
a r t i c l e i n f o
a b s t r a c t
Article history:
The methanol to hydrocarbons (MTH) reaction was studied over two isostructural zeotype catalysts
of different acid strength, H-SAPO-5 and H-SSZ-24. Conversion of methanol alone was performed at
350–450 ◦C and WHSV = 0.31–2.48 h−1. The product selectivities of the two catalysts were compared
at similar conversion. The strongly acidic H-SSZ-24 was found to be more selective towards aromatic
products and C2–C3 hydrocarbons as compared to the moderately acidic H-SAPO-5, which produced more
non-aromatic C4+ hydrocarbons. Co-reactions of 13CH3OH and benzene at 250–300 ◦C with low conversion
of both reactants revealed that both catalysts produced ethene and propene from polymethylbenzenes
via a paring mechanism. However, this reaction proceeded more readily in H-SSZ-24 than in H-SAPO-
5. Furthermore, isobutene formation was found to be mainly associated with aromatic intermediates
in H-SSZ-24, whereas isobutene produced over H-SAPO-5 was mainly formed via alkene intermediates.
Overall, the results obtained in this study suggest that a lower acid strength promotes an alkene-mediated
MTH reaction mechanism.
Received 16 November 2012
Received in revised form 5 March 2013
Accepted 14 March 2013
Available online 16 April 2013
Keywords:
Methanol conversion
MTH
H-SAPO-5
H-SSZ-24
Acid strength
Paring reaction
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
hydrocarbon reactions, where several classes of reactions compete,
could be sensitive to a change in acid strength.
Acid catalysis is important in many chemical reactions, not
only in the petrochemical industry where acidic zeolite catalysts
are used in major processes, such as in catalytic cracking [1]. Thus,
the influence of acid strength on reaction mechanisms and rates
is potentially an important topic. However, the term acidity in
solids is somewhat ambiguous as both its definition and how to
measure it has been a subject of debate [2]. In zeolites and related
materials (zeotypes), the most important acid sites for catalysis
are Brønsted type (proton donating). A calculated deprotonation
energy for an isolated site is one measure of acid strength which
has the advantage of being rigorously defined. Unfortunately,
reliable deprotonation energies are not available for many zeotype
catalysts. Another measure of acid strength is the interaction of
the acidic OH group with basic probe molecules as monitored by
e.g. vibrational spectroscopy [3] or NMR [2]. A recent study by
Macht et al. [4] revealed that the activation barriers of both the
dehydration of butanol and the isomerisation of hexane depended
linearly on acid strength (quantified by deprotonation energy).
However, the sensitivity to changing acid strength was different for
the two reactions. This means that the selectivity of acid-catalysed
The conversion of methanol to hydrocarbons (MTH) over acidic
zeotype catalysts is a flexible reaction route to produce light alkenes
or gasoline from alternative hydrocarbon feedstocks such as natu-
ral gas, biomass or coal [5]. Several processes based on this reaction
are being commercialised, such as the UOP/Hydro methanol to
olefins (MTO) process, Lurgi’s methanol to propene (MTP), Mobil’s
methanol to gasoline (MTG) process and the Topsøe TIGAS process
[5]. Depending on catalyst used and conditions employed three
main product groups, in addition to the simple dehydration of
methanol to form dimethyl ether, can be obtained in these pro-
cesses. These are: light alkenes such as ethene and propene; higher
alkenes (C4–C8); and mixtures of aromatic and alkane products.
Direct reactions between methanol molecules are insignificant
under steady-state MTH conditions [6,7]. Instead, product for-
mation proceeds via a hydrocarbon pool mechanism [8–10]. The
identity of the hydrocarbon pool has been studied intensively, and
it has been found that both aromatics and alkenes are important
hydrocarbon pool compounds [11–18]. The current mechanistic
understanding of the MTH reaction is schematically illustrated
in Scheme 1: Alkenes and aromatics are both methylated with
methanol (or dimethyl ether) to form larger hydrocarbons and sub-
sequently crack or dealkylate to form light alkenes and regenerate
the starting compounds. This division between two classes of inter-
mediates is often referred to as the dual-cycle concept. The two
cycles are interrelated: Higher alkenes may undergo cyclisation
∗
Corresponding author. Tel.: +47 22855456.
E-mail address: unni.olsbye@kjemi.uio.no (U. Olsbye).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.03.017

M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
217
Scheme 1. General scheme of the dual-cycle mechanism. Both the relative propagation of each cycle and the exact structure of the intermediates depend on the catalyst
employed and the reaction conditions. This in turn means that all products shown here are not necessarily observed in all systems.
and hydride transfer reactions leading to formation of aromatics
and alkanes. Aromatics, on the other hand, contribute to for-
mation of light alkenes that may subsequently take part in the
alkene cycle. It has been found that the relative activity of the
two classes of intermediates and their exact structure depends
on the catalyst employed as well as the reaction conditions (see
[5] and refs. therein): beta zeolite, with 12-ring zeolite channels,
favour the heptamethylbenzenium cation as the aromatic reac-
tion intermediate, leading preferably to formation of propene and
isobutane [14,15,18–20]. MFI zeolites, with 3D 10-ring channels
favour tri- and tetra-methylbenzene as aromatic intermediates,
leading preferably to ethene and propene formation [16–19]. In
each case, the alkene cycle contributes to product formation beside
the arene cycle. TON, with 1D 10-ring channels of slightly smaller
diameter than MFI, strongly favour the alkene cycle while the arene
cycle is suppressed [21,22]. In addition to influencing which hydro-
carbon pool species can be formed and converted, the available
channel space inside the catalyst influences which products are able
to diffuse out of the structure [23]. As such, the largest aromatic
compound found in the product effluent of BEA is hexamethyl-
benzene, while in MFI it is 1,2,4,5-tetramethylbenzene. Aromatic
compounds formed in TON are retained in the channels, and its
effluent consists mainly of alkenes [5]. Evidence exists that a change
in reaction conditions, or acid site densities, can change the relative
propagation rates of the alkene and arene cycles [24,25]. How-
ever, the influence of acid strength on the MTH performance of
the catalyst is not yet fully understood.
The effects of acid strength in MTH were first investigated
by Yuen et al. [26], who compared aluminosilicate and silicoa-
luminophosphate versions of AFI (H-SAPO-5 versus H-SSZ-24)
and CHA (H-SAPO-34 versus H-SSZ-13) structured catalysts. Their
experiments were performed under similar conditions giving full
conversion. They observed a more rapid breakthrough of methanol
over the strongly acidic H-SSZ-13 than over the moderately acidic
H-SAPO-34. Comparing the performance of the AFI and CHA topolo-
gies, they concluded that the pore-size was a more important factor
than acid strength in this reaction. More recently, our group pub-
lished a detailed comparison of H-SAPO-34 and H-SSZ-13 [27]. This
study included a series of tests at less than full conversion at differ-
ent temperatures, in addition to characterisation of the retained
hydrocarbons after use. While it was found that H-SSZ-13 was
more active and deactivated faster than the less acidic H-SAPO-34,
the methanol conversion capacity (i.e. the cumulative methanol
production before full deactivation) was surprisingly found to be
higher in H-SSZ-13 at temperatures below 375 ◦C. Further mech-
anistic insights proved difficult due to the severe restrictions on
diffusion in the CHA structure leading to a high degree of product
shape selectivity [28].
In contrast to the CHA structure, where narrow pores restrict
diffusion into and out of the large cavities, the AFI structure has
one-dimensional tubular 12-ring channels that, while similar in
diameter to the cavities of the CHA structure, are open all the
way to the exterior. This allows bulky molecules to diffuse in and
out of the structure, and provides an opportunity for monitoring
and manipulating the hydrocarbon-pool by performing co-reaction
experiments. Furthermore, the simple structure of the AFI frame-
work makes it well-suited as a model system for the MTH reaction
in general, as the absence of cavities and intersections facilitates
estimation of the space available for reactions. In a recent study [29]
we performed co-feeding and isotopic labelling experiments over
the moderately acidic H-SAPO-5 catalyst (AFI structure, 12-ring
pores) and found that this catalyst differed from large-pore zeolites
studied previously. It was shown that aromatic polymethylben-
zene (PolyMB) intermediates could produce ethene and propene
under conditions where their concentration was high. However,
we concluded that higher alkene intermediates were more impor-
tant hydrocarbon pool species under normal MTH conditions. This
conclusion contrasted previous findings over H-Beta, which has a
similar pore size but is a stronger acid. In H-Beta, polyMBs were
found to be the dominant intermediates [14,15,18,19] under similar
conditions.
The present study is aimed to further elucidate the influence
of acid strength of zeotype catalysts in the MTH reaction. For
this purpose the zeolite analogue to H-SAPO-5, called H-SSZ-
24, was synthesised and the performance of the two catalysts
in MTH reaction was compared. Both product selectivities dur-
ing methanol conversion and the reaction patterns of polyMBs in
methanol/benzene co-reactions were investigated. Such a direct
comparison between two isostructural catalysts of differing acid
strength provides a unique opportunity to study the effect of acid
strength alone.
2. Materials and methods
2.1. Catalyst synthesis and characterisation
H-SAPO-5 was synthesised hydrothermally from water, tri-
ethylamine (TEA) (Fluka, 99.5%), orthophosphoric acid (Merck,
85%), Cab-O-Sil M5 (Riedel-de Haën) and Catapal B (Vista). A mix-
ture with the composition: 50H2O:1Al2O3:1P2O5: 0.1SiO2:3.9TEA
was prepared and crystallised for 4 h at 200 ◦C in a rotating Teflon
lined steel autoclave. Removal of the structure directing agent (TEA)
was performed by calcination in air at 600 ◦C for 2 h.
The synthesis of H-SSZ-24 was based on the work by Kubota et al.
[30,31]. A sample of zeolite H-Beta was first synthesised accord-
ing to [32] (without adding Ti) using water, tetraethylammonium

218
M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
hydroxide (Aldrich, 40% in water), Cab-O-Sil M5 and aluminium
nitrate (Al(NO3)3·9H2O, Fluka, 98%). This sample was then cal-
cined at 550 ◦C (heating ramp ∼50 ◦C/h) for 36 h and used as
aluminium source in the synthesis of H-SSZ-24. The structure
directing agent for H-SSZ-24 synthesis, N(16)-methylsparteinium
hydroxide (MeSpa+OH−), was synthesised similarly to the proce-
dure reported by Lobo et al. [33] and Kubota et al. [34]. Briefly,
sparteine was extracted from (−)sparteine sulphate pentahydrate
(ABCR) by stirring with 10% NaOH and extracting with CH2Cl2. The
solid (-)sparteine was then dissolved in acetone and methylated
with methyl iodide (Sigma-Aldrich, 99%). The crude product was
recrystallised from boiling 2-propanol and ion-exchanged to OH-
form with Amberlite IRN 78 (Supelco) ion exchange resin.
Fig. 1. Powder XRD patterns of the catalysts employed in this work.
The SSZ-24 sample was prepared by mixing 5.88 g of water,
2.27 g of a 0.66 mmol g−1 solution of MeSpa+OH− and 0.24 g of a
6.3 mmol g−1 aqueous solution of NaOH. After stirring for 10 min,
0.18 g of the BEA precursor and 0.43 g Cab-O-Sil was added. This
mixture was stirred for 4 h before being set to crystallise in a Teflon-
lined steel autoclave at 170 ◦C for 25 h. The structure directing agent
was removed by calcination at 550 ◦C (heating ramp ∼50 ◦C/h) in a
flow of 25% O2 in N2 for 10 h. The Na-SSZ-24 thus formed was ion-
exchanged three times with an excess of 1 M NH4NO3 (30 g solution
per gram catalyst) at 80 ◦C before calcining again at 550 ◦C in 25%
O2 in N2 for 10 h again to make H-SSZ-24.
Rtx®-DHA-150 column (150 m, 0.25 mm i.d., stationary phase
thickness 1 ␮m). Hydrogen (purity 6.0) was used as carrier gas. For
isotope analysis of effluent in the benzene/methanol co-feed exper-
iments, the same GC and conditions were used, but the effluent was
routed to an Agilent 5975C MS detector.
3. Results and discussion
3.1. Catalyst characterisation
Both zeotype samples were characterised by powder XRD,
SEM, N2 adsorption, and FT-IR. For powder XRD measurements,
a Bruker D8 Discover diffractometer with Bragg–Brentano geom-
Catalyst characterisation data are compiled in Table 1. X-ray
diffraction patterns of H-SAPO-5 and H-SSZ-24 are shown in Fig. 1.
Both diffractograms correspond to pure and highly crystalline AFI
structures, only differing slightly in relative intensities and peak
positions as expected due to the difference in composition. No
crystalline impurities were detected. SEM analysis of the samples
revealed the H-SAPO-5 samples to consist of hexagonal crystals
roughly ∼1 ␮m in diameter and 1–2 ␮m in length, while the H-
SSZ-24 sample consisted of more irregular hexagonal rod-shaped
crystals less than 1 ␮m long. The BET surface areas of the samples
were determined to be 341 m2/g and 284 m2/g for H-SAPO-5 and
H-SSZ-24 respectively.
EDS analysis of the two samples showed (Al+P)/Si = 80 for H-
SAPO-5 and Si/Al = 35 in H-SSZ-24 (Table 1). These data suggest a
higher density of acid sites in the zeolite sample. The consequences
of this finding for the present study are discussed in Section 3.2
below.
FT-IR spectra of the H-SAPO-5 sample subjected to increasing
amounts of CO at −196 ◦C are presented in Fig. 2. The activated sam-
ple (black line) shows three absorption bands in the O-H stretching
region: at 3678, 3630 and at 3530 cm−1. Similar bands in H-SAPO-
5 have been described in numerous previous reports [35–38]. The
band at 3678 cm−1 is ascribed to P-OH groups at the external sur-
face (or defect sites), while the two bands at 3630 and 3530 cm−1
˚
etry using Cu K␣ radiation (ꢀ = 1.5406 A) was employed. SEM was
performed with a FEI Quanta 200 FEG-ESEM instrument. N2 adsorp-
tion isotherms were measured at -196 ◦C, using a BELSORP-mini II
instrument. Samples were outgassed in vacuum for 1 h at 80 ◦C and
3 h at 300 ◦C. Specific surface area was calculated using the BET
equation based on p/p0 data in the range 0.01–0.15. FT-IR mea-
surements were performed in transmission mode on a Vertex 80
instrument with MCT detector. Before measurement the samples
were pressed into a self-supporting wafer, and were pre-treated
under vacuum by heating to 120 ◦C for 1 h, 300 ◦C for 1 h and 450 ◦C
for 1 h.
2.2. Catalytic tests
All catalytic tests were performed in a fixed-bed glass reactor of
8 mm inner diameter. The catalyst powder was pressed and sieved
to obtain particles between 0.25 and 0.42 mm. Prior to the intro-
duction of reactants, for all experiments, the pressed and sieved
catalyst was calcined in situ at 550 ◦C under a flow of oxygen for
1 h before cooling to reaction temperature. Reactants used were 12
C
methanol (VWR, 99.8%), 13C methanol (Cambridge Isotope Labora-
tories, 99%) and benzene (Riedel-de Haën, 99.5%). These reactants
were fed over the catalyst by passing a stream of helium through a
saturator at constant temperature.
Both catalysts were tested for the conversion of methanol at
350 ◦C and 450 ◦C. 100 mg of H-SAPO-5 and 25 mg of H-SSZ-24
was used. The methanol partial pressure in these experiments
was 40 mbar (saturator temperature 0 ◦C) and the flow of helium
was adjusted to obtain WHSV between 0.31 and 2.48 h−1. Product
selectivities in the effluent were compared for the two catalysts at
similar conversion and temperature.
Co-reaction experiments of 12C benzene and 13C methanol were
performed at feed partial pressures of 76 mbar and 25 mbar, respec-
tively (saturator temperatures of 20 ◦C and 10 ◦C). 100 mg were
used of both catalysts and WHSV was 1.0 and 0.8 h−1 respectively
over H-SAPO-5 and 8.0 and 6.4 h−1 over H-SSZ-24. The effluents
were analysed after 2 min on stream.
Fig. 2. FT-IR spectra of increasing dosages of CO on H-SAPO-5 at −196 ◦C. In (a)
the OH stretching region is shown, while (b) shows the CO stretching region (back-
ground subtracted). The black bold curves correspond to the activated sample, while
the grey bold curves correspond to the spectrum of highest CO loading.
The reactor effluent was analysed quantitatively by online GC
analysis (Agilent 7890 with flame ionisation detector) on a Restek

M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
219
Table 1
Summary of characterisation data for the two catalysts employed in this work. Acid site densities are the ratios of (Al+P)/Si (in H-SAPO-5) or Si/Al (in H-SSZ-24) as determined
by EDS.
Sample
Structure
Crystal size (␮m)
BET surface
Acid site density
ꢁ(OH) shift
ꢁ(CO) shift
H-SAPO-5
H-SSZ-24
AFI
AFI
1 × 2 ␮m
<1 ␮m
340 m2/g
284 m2/g
80
35
−265 cm−1
−317 cm−1
+34 cm−1
+38 cm−1
are ascribed to bridging Si–OH–Al groups with protons located in
the 12- and 6-rings of the AFI structure, respectively [36].
Again, the corresponding ꢁCO region of the sample (Fig. 3b)
mirrors the observations from the ꢁOH region. The frequency of
CO adsorbed on the acidic Si–OH–Al groups are 2177 cm−1, giving
ꢂꢁCO = +38 cm−1. The interaction of the Si–OH groups in H-SSZ-24
with CO is also visible in the CO region, giving rise to a band at high
The grey lines in Fig. 2 are spectra of the catalyst samples
exposed to CO. The O–H stretch region is shown in Fig. 2a. From
Fig. 2a it can be seen that as CO is adsorbed on the sample, the
Si–OH–Al band at 3630 cm−1 is eroded while two new bands grow
at 3464 cm−1 and 3365 cm−1. With a further increase in CO pres-
sure, the P–OH band at 3678 cm−1 is also eroded. This change is
accompanied by a decrease in intensity around 3500 cm−1. It is
uncertain whether these two events are related or not since sev-
eral bands shift slightly as the CO loading increases. The appearance
of two new bands when the Si–OH–Al band at 3630 cm−1 erodes
is surprising since H-SAPO-5 is expected to contain only one type
of strong acid site. This could mean either that two distinctly dif-
ferent acid sites are present in H-SAPO-5, or it could be due to a
fermi-resonance effect similar to that proposed by Chakarova et al.
[39] when they adsorbed CO on H-ZSM-5.
The corresponding spectra of the CO stretching region of the
sample (Fig. 2b) mirror the observations above, as the CO stretch-
ing frequency is blue-shifted relative to free CO molecules when
adsorbed on the acid sites of the materials. The frequency of CO
adsorbed on the acidic Si–OH–Al groups are 2174 cm−1, giving
ꢂꢁCO = +34 cm−1. The P–OH groups are, however, more difficult to
observe in this region. The observation of only one band for CO
adsorption at low coverage suggests that the observation of two
bands in the OH region is spectral rather than representing two
distinct sites.
CO coverage at around 2158 cm−1
.
Overall, the key difference between the H-SAPO-5 and H-SSZ-24
is the significant difference in magnitude of the shift in OH stretch-
ing frequency for the Si–OH–Al groups when they interact with
CO. The largest shift observed in H-SAPO-5 is ꢂꢁOH = −265 cm−1
while the shift in H-SSZ-24 is ꢂꢁOH = −317 cm−1, meaning that the
acid sites of H-SSZ-24 are significantly stronger than those of H-
SAPO-5. Both this conclusion and the values of the shifts correspond
fairly well with the differences found by Bordiga et al. [41] between
the isostructural zeotypes H-SAPO-34 (ꢂꢁOH = −270 cm−1) and H-
SSZ-13 (ꢂꢁOH = −316 cm−1), signifying that an acidity difference on
this scale could be a general trend for high-silica zeolites and SAPO
materials.
It is further interesting to note that in both samples the bridging
Si-OH-Al groups present in the 6-rings of the AFI structure are not
accessible to CO molecules at −196 ◦C, as the bands at 3530 cm−1
in H-SAPO-5 and 3488 cm−1 in H-SSZ-24 do not disappear upon
adsorption of CO. These sites have previously been shown to inter-
act with stronger bases at higher temperatures [36,38], suggesting
mobility of the protons involved. On the other hand, the Si–OH–Al
groups in the 12-rings of the structure interact strongly with CO
even at low coverage. The surface P–OH in H-SAPO-5 and Si–OH in
H-SSZ-24 show a weak interaction with CO at high coverage only.
IR spectra of H-SSZ-24 are presented in Fig. 3. The general
spectroscopic features of H-SSZ-24 are very similar to those of H-
SAPO-5: Three absorption bands are visible in the O-H stretching
region of the activated H-SSZ-24 sample (black line), at 3747, 3612
and 3488 cm−1. These bands are ascribed to Si–OH groups and
to Si–OH–Al groups located in 12- and 6-rings respectively. The
band positions correspond well with those reported by Martinez-
Triguero et al. [40] at room temperature, (in fact reproducing our
measurements at room temperature gave complete overlap with
their data). When the sample was cooled to −196 ◦C, the values
reported here were obtained. As CO is adsorbed (grey lines), the
Si–OH–Al band at 3612 cm−1 is eroded and a new band grows at
3295 cm−1. As more CO is adsorbed, the Si–OH band at 3747 cm−1
3.2. Methanol conversion over H-SSZ-24 and H-SAPO-5
The scope of this contribution was to elucidate the effects
of a change in acid strength on the product selectivity and the
favoured mechanism of the MTH reaction over wide pore zeotypes
with AFI topology. Selectivity comparisons of the two catalysts at
similar conversions (20 and 60%) at 350 ◦C and 450 ◦C are shown
in Figs. 4 and 5, respectively. It should be noted that due to the
activity differences between the two materials, the H-SSZ-24 sam-
ple was tested at four times higher weight hourly space velocity
(WHSV) than the H-SAPO-5 sample in order to start at similar
conversion. The employed WHSV for H-SSZ-24 and H-SAPO-5
was 2.48 h−1 and 0.62 h−1 respectively at 450 ◦C, and 1.24 h−1 and
0.31 h−1 respectively at 350 ◦C. For H-SAPO-5 at 350 ◦C, data at low
conversion was obtained in a separate experiment performed at
is also eroded, and a new band appears at 3650 cm−1
.
WHSV 0.62 h−1 due to a very slow deactivation at WHSV 0.31 h−1
.
Furthermore, the reported data are not initial conversions, but
correspond to partly deactivated catalysts. Previous studies of
H-SAPO-5 showed that deactivation had a negligible influence on
the product selectivity versus methanol conversion correlations
in the conversion ranges reported [29]. A similar conclusion has
previously been reached for the title reaction both over H-ZSM-5
[42] and over several 1-dimensional 10-ring structures [43].
Distinct selectivity differences were observed between the two
materials, exemplified in Figs. 4 and 5 at two conversion levels
and two temperatures; i.e.; 20% and 60% conversion at 350 ◦C and
450 ◦C. A more detailed breakdown of the C4, C5 and aromatics
fractions from Figs. 4 and 5 is included as Supplementary material
(section S1). Over H-SAPO-5, C4 was the major product under all
Fig. 3. FT-IR spectra of increasing dosages of CO on H-SSZ-24 at −196 ◦C. In (a) the OH
stretching region is shown, while (b) shows the CO stretching region (background
subtracted). The black bold curves correspond to the activated sample, while the
grey bold curves correspond to the spectrum of highest CO loading.

220
M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
in the C4 and C5 fractions, respectively, at the highest conversion
(60%) and only around 10% at 20% conversion. At 450 ◦C, the amount
of alkanes was even lower over H-SAPO-5 (<10%), while a signif-
icant alkane fraction was still found in the effluent of H-SSZ-24
(>40%).
Concerning the non-aromatic fraction, the relative percentages
of the C2-C6 hydrocarbons were quite independent of conversion
at each temperature over H-SSZ-24, but varied significantly with
conversion over H-SAPO-5 (Figs. 4 and 5). The most striking exam-
ple is the C4:C3 ratio, which changed from 4.3 at 20% conversion
to 2.0 at 60% conversion over H-SAPO-5 at 350 ◦C, while it changed
only from 2.0 to 1.6 over H-SSZ-24. At 450 ◦C, the corresponding
numbers were: 2.5–1.5 for H-SAPO-5, and 0.9 at both conversions
over H-SSZ-24.
Fig. 4. Product selectivities for H-SAPO-5 (dark grey) and H-SSZ-24 (light grey) at
350 ◦C at 20% (a) and 60% (b) conversion. Data were obtained during deactivation of
H-SSZ-24 at WHSV = 1.24 h−1 and of H-SAPO-5 at WHSV = 0.62 h−1 (a) and 0.31 (b).
Corresponding numbers for the C3: C2 ratio, was ∼10 and 8–15
over H-SAPO-5 and 350 ◦C and 450 ◦C, respectively, and ∼5 at 350 ◦C
and 5–6 at 450 ◦C, over H-SSZ-24. The C2 and C3 fractions were
under all conditions dominated by ethene and propene respec-
tively. The C4: C2 ratio was significantly higher for H-SAPO-5 than
for H-SSZ-24, being ≥20 over H-SAPO-5 and ≤10 over H-SSZ-24
under all reported conditions.
Overall, the data in Figs. 4 and 5 show that the selectivity to aro-
matic products is higher over H-SSZ-24 than over H-SAPO-5. The
observed difference could be ascribed to the different acid strength
of the two materials, but it could also be related to the higher den-
sity of acid sites in H-SSZ-24 compared to H-SAPO-5 (Table 1): a
higher density of acid sites has previously been shown to favour
sequential reactions when performing the MTH reaction over H-
ZSM-5 for Si/Al ratios lower than 50 [42]. A similar influence of
acid site density in propene oligomerisation over H-ZSM-5 has been
reported for low Si/Al ratios, but Si/Al ratios of 40 and 140 gave very
similar selectivity at a given conversion [44]. In order to elucidate
the influence of acid site density on the selectivity pattern, a num-
ber of H-SAPO-5 batches with (Al+P)/Si ratios in the range 20–80
were tested and compared to H-SSZ-24. The test data are shown as
Supplementary material, Figs. S1 and S2. Although individual selec-
tivity differences were observed among the H-SAPO-5 batches,
the selectivity towards aromatic products was distinctively higher
over the H-SSZ-24 batch than over each of the H-SAPO-5 batches.
The observed difference in selectivity towards aromatic products
between H-SSZ-24 and H-SAPO-5 is therefore ascribed to the dif-
ference in acid strength between the two materials.
Fig. 5. Product selectivities for H-SAPO-5 (dark grey) and H-SSZ-24 (light grey) at
450 ◦C at 20% (a) and 60% (b) conversion. Data were obtained during deactivation of
H-SSZ-24 at WHSV = 2.48 h−1 and of H-SAPO-5 at WHSV = 0.62 h−1
.
conditions (∼38–52% selectivity), with a volcano-like product dis-
tribution in the C2–C6+ non-aromatic region. Isobutene represented
>50% of the C4 fraction under all reported conditions. Further-
more, H-SAPO-5 gave low selectivity to aromatic products; < 14%
at 350 ◦C and < 5% at 450 ◦C, respectively, in the conversion range
covered by the two figures. Turning to H-SSZ-24, its most dis-
tinct feature was a much higher selectivity to aromatic products
than observed over H-SAPO-5. H-SSZ-24 gave > 26% and > 19% aro-
matics selectivity at 350 ◦C and 450 ◦C, respectively. Among the
non-aromatic products, C4 was the main fraction also over H-SSZ-
24 at 350 ◦C, while C3 was the main fraction at 450 ◦C. Isobutyl
species dominated the C4 fraction also over H-SSZ-24. In line with
the higher selectivity towards aromatic products, a higher percent-
age of the non-aromatic products was paraffinic rather than olefinic
over H-SSZ-24, compared to H-SAPO-5: At 350 ◦C, the C4, C5 and
C6+ fractions of H-SSZ-24 were all composed of > 60% isoalkanes,
while the percentage of alkanes over H-SAPO-5 was 22% and 34%
Yuen et al. have previously reported that H-SSZ-24 and H-SAPO-
5 gave similar product selectivities in the MTH reaction, in apparent
contradiction to our results. A possible reason for the difference
is that their tests were performed at full conversion for all sam-
ples, allowing the products to undergo sequential reactions in the
direction of thermodynamic equilibrium.
Table 2
Experimental details, conversion and product selectivities from co-reactions of 12C benzene and 13C methanol.
Temperature
Catalyst
250 ◦
C
300 ◦
C
H-SSZ-24
H-SAPO-5
H-SSZ-24
H-SAPO-5
WHSV benzene
WHSV methanol
Conversion (%)
6.4 h−1
8.0 h−1
7.5
0.8 h−1
1.0 h−1
1.8
6.4 h−1
8.0 h−1
21.0
0.8 h−1
1.0 h−1
16.1
Selectivity (%)
Methane
C2
C3
C4
0.4
0.9
1.9
1.1
5.0
37.6
36.6
16.5
0
0.2
1.9
4.2
3.8
3.6
34.6
45.8
5.9
0.1
2.5
6.7
21.3
17.7
15.4
24.4
12.1
0.4
1.2
3.5
1.7
74.5
11.2
7.5
C5+
Toluene
Other MBs
HexaMB

M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
221
Fig. 6. Distribution of 13C atoms in ethene (left), propene (middle), isobutene (right)
after 2 min of co-reacting methanol with benzene (molar ratio 3:1) at 250 ◦C (top)
and 300 ◦C (bottom) over H-SAPO-5. At 250 ◦C, the total 13C content in ethene was
53% and 67% in propene, while it was 88% in i-butene. At 300 ◦C, the total 13C content
was ∼80% in ethene and propene and 91% in i-butene. The molecules above the
graphs are drawn with stars signifying labelled carbon atoms representative of the
most abundant isotopologues found at 250 ◦C.
Fig. 7. Distribution of 13C atoms in ethene (left), propene (middle), isobutene (right)
after 2 min of co-reacting methanol with benzene (molar ratio 3:1) at 250 ◦C (top)
and 300 ◦C (bottom) over H-SSZ-24. At 250 ◦C, the total 13C content in ethene and
propene was ∼60%, while it was 75% in i-butene. At 300 ◦C total 13C contents was
∼70% in ethene and propene and 78% in i-butene. The molecules above the graphs
are drawn with stars signifying labelled carbon atoms representative of the most
abundant isotopologues found at 250 ◦C.
(originating mainly from methanol) is competitive with benzene
methylation under these conditions.
3.3. Light alkene formation from aromatic intermediates
The isotopic distribution in the light alkenes over H-SAPO-5 is
shown in Fig. 6. It reveals that one 12C atom from benzene was
incorporated into the ethene and propene produced over H-SAPO-5
at 250 ◦C. On the other hand the isobutene produced over this cat-
alyst contained a significantly larger fraction of 13C atoms than the
lighter compounds, and the all 13C isotopomer was dominant. At
300 ◦C, all alkenes contained more 13C and the all-13C isotopomer
dominated in all three compounds. The systematic incorporation
of 12C atoms originating from benzene in ethene and propene at
250 ◦C points to polyMBs as intermediates for the production of
these alkenes. On the other hand, the lack of such a systematic pat-
tern and the low incorporation of 12C in isobutene indicate that this
alkene is formed by a mechanism different from the lighter alkenes
in H-SAPO-5.
As presented in the Introduction, aromatic species represent
both end products and hydrocarbon pool species in the MTH reac-
tion. In order to investigate the reactivity of aromatic molecules
in the two catalysts, co-feed experiments of 12C-benzene and
13C labelled methanol were performed at 250–300 ◦C. During
these experiments the concentration of polymethylated benzene
molecules (polyMBs) was greatly enhanced inside the catalysts
compared to ordinary MTH experiments, thereby enabling an
investigation of their reactivity. Again, a much higher WHSV of
methanol and benzene combined was used over H-SSZ-24 than
over H-SAPO-5. However, even with a factor 8 in contact time
difference, significantly higher conversion was obtained over H-
SSZ-24 (7.5%) compared to H-SAPO-5 (1.8%) after 2 min on stream
at 250 ◦C (Table 2). At 300 ◦C the conversion after 2 min on stream
was more similar over the two catalysts (21.0 vs. 16.1%), but rela-
tive polyMB yields differed significantly (Table 2). A short time on
stream was necessary to obtain distinct isotopic labelling patterns
in the products (Figs. 6 and 7). For this reason data collected after
2 min on stream are reported in spite of the conversion differences.
It is apparent from Table 2 that, in line with the observations from
conversion of methanol alone, H-SAPO-5 is much more selective
towards isobutene/isobutane than H-SSZ-24. At 250 ◦C the ratios
found between the yield of C2, C3 and C4 hydrocarbons were 1:3:9
in H-SAPO-5, while they were 1:2:1 in H-SSZ-24. Upon increasing
the temperature to 300 ◦C, these ratios did not change much, except
for an increase in C4 selectivity for H-SSZ-24, bringing the ratios to
1:2:2.
A plot of isotopic distribution in products formed over H-SSZ-
24 is shown in Fig. 7. Similarly to what was found for H-SAPO-5
at 250 ◦C, the major isotopomers of ethene and propene contained
one 12C atom. However, in contrast to H-SAPO-5, the dominating
isobutene isotopomer also contained one 12C atom at this tem-
perature. At 300 ◦C the alkenes contained a larger fraction of 13C
atoms, and the isotopic distribution observed at 250 ◦C was less
clear. However, a significant amount of alkenes containing one 12
C
atom was still observed. Overall, significantly more 12C was present
in all three alkenes than what was found over H-SAPO-5 at both
temperatures. These results suggest that C2-C4 alkene formation at
250 ◦C over H-SSZ-24 mainly proceeds through polyMB intermedi-
ates, and that this same mechanism contributes even at 300 ◦C.
The isotopic distribution of the polyMBs produced by the
co-reaction of 13C-methanol and 12C-benzene are shown as Sup-
plementary material (Figs. S3–S6). The dominating isotopomer of
all polyMBs contained the same number of 13C atoms as methyl
groups. This is consistent with sequential methylations of benzene
being the dominant pathway for formation of polyMBs under these
conditions. However, it is worth noting that at 300 ◦C (Fig. S4), the
13C distribution in the polyMBs produced by H-SAPO-5 became
bimodal. The surprisingly high abundance of highly 13C substituted
isotopomers may suggest that formation of polyMBs from alkenes
3.4. On the mechanism of arene de-alkylation
The data reported in Figs. 6 and 7 further shed light on another
mechanistic aspect of the MTH reaction: PolyMB de-alkylation.
This topic has been debated for many years, and two main mech-
anistic proposals have been put forth: The paring reaction and
the exocyclic methylation mechanism. A simplified scheme of
these two reaction pathways is presented in Scheme 2. The latter
mechanism was first proposed by Mole et al. [45,46] and later
refined by Haw et al. [23] and involves deprotonation of HeptaMB+

222
M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
ring -contraction or -expansion step must be involved in the growth
and elimination of alkyl side-chains.
3.5. Summary and mechanistic interpretations
The main findings of this study may be summarised as follows:
In Section 3.2, it was found that the strongly acidic H-SSZ-24 is
more selective towards ethene and aromatics than the moderately
acidic H-SAPO-5 during methanol conversion. Furthermore, it
was found that the relative distribution of C2-C4 alkenes was
more sensitive to the conversion of methanol in H-SAPO-5 than
in H-SSZ-24. This observation may suggest that C2-C4 alkenes
are mainly formed from a common precursor in H-SSZ-24, but
probably not in H-SAPO-5.
The co-feed experiments in Section 3.3 showed that H-SAPO-5
was significantly more selective towards isobutene than H-SSZ-24.
Furthermore, as pointed out in Section 3.4, isobutene showed an
isotopic labelling pattern consistent with an alkene cycle over H-
SAPO-5, but with an arene cycle over H-SSZ-24 at 250 ◦C. C2 and
C3 showed an isotopic labelling pattern consistent with formation
from an arene cycle over both catalysts at 250 ◦C. These data match
the selectivity patterns reported in Section 3.2 well, and point to
the alkene cycle being a dominating source of product formation in
H-SAPO-5, while the arene cycle dominates in H-SSZ-24.
Scheme 2. Expected labelling patterns when propene is split off from a selectively
labelled heptaMB+ via paring or exocyclic mechanisms, respectively. Stars are used
in order to keep track of the carbons originally present as methyl groups. The paring
pathway involves systematic scrambling of these into the benzene ring, while the
exocyclic methylation does not. Scheme from [29].
If one further considers the data obtained at 300 ◦C, showing the
transition from a clear labelling pattern at 250 ◦C to nearly complete
scrambling for all alkenes in H-SAPO-5 (Fig. 6) at 300 ◦C, and com-
pares it to the situation in H-SSZ-24, where the original pattern
was still visible (Fig. 7) at the higher temperature, it strengthens
the argument that polyMBs are less important hydrocarbon pool
species in H-SAPO-5 than in H-SSZ-24. Finally, it should be noted
that while these isotopic distributions were easily reproducible
over a wide conversion range over H-SSZ-24 at 250 ◦C, obtaining
paring-type labelling patterns in H-SAPO-5 required a very low
conversion.
It has previously been suggested that the formation of the light-
est alkenes, and especially ethene, is mainly associated with the
arene cycle and polyMB intermediates [5]. In line with this sugges-
tion, the observation of a higher C3+ selectivity over the moderately
acidic H-SAPO-5 than over the more strongly acidic H-SSZ-24 ties
in nicely with the observed lower importance of polyMB interme-
diates during the co-feed experiments. As the two catalysts are very
similar apart from their acid strength, it appears reasonable to con-
clude that a change in acid strength in turn leads to a shift in the
relative abundance of the alkene or arene cycle, with a high acid
strength favouring the latter cycle. It should be noted, however,
that the co-feed experiments were conducted at a lower temper-
ature than the methanol feed experiments, and that the relative
abundance of the alkene versus the arene cycle in each material
might be different at 350–450 ◦C than at 250–300 ◦C.
There are at least two possible explanations to the observed
effect of acid strength in this work; i.e. the enhanced significance of
the arene versus the alkene cycle to light alkene formation in H-SSZ-
24 than in H-SAPO-5. Either, it could be an effect of the increased
aromatics concentration, due to the higher selectivity towards aro-
matic products, in H-SSZ-24 (see Figs. 4 and 5). Computational
studies suggest that cyclisation reactions are facile [48], meaning
that hydride transfers are likely the rate-determining steps in the
formation of aromatics. A link between acid strength and hydride
transfers has been proposed previously [26,49], and could be at
the core of the observed promotion of the arene cycle in H-SSZ-24.
Another possibility is that the key reaction steps of the alkene and
arene cycles display a different sensitivity to acid strength. Macht
et al. [4] proposed that the main reason why n-hexane isomerisa-
tion is more sensitive to a change in acid strength than butanol
dehydration is the more localised charge in the transition state
to form a neutral compound with an exocyclic double bond. This
double bond can be methylated once or twice to form an ethyl
or isopropyl side-chain, which is subsequently eliminated. On the
other hand, the paring reaction, first proposed by Sullivan et al. [47]
to account for the high yield of isobutene during hydrocracking
of hexamethylbenzene, involves the rearrangement of HeptaMB+
to a five-membered ring with an alkyl substituent. This smaller
ring can then either split off propene directly or reorganise further
to eliminate isobutene before deprotonation and expansion back
to a six-ring. Contrary to the exocyclic methylation mechanism,
the paring reaction induces incorporation of ring carbons into the
alkene product. In Scheme 2, the stars signify 13C-labelled carbon
atoms, i.e. the carbons originally present as methyl groups. While
Scheme 2 depicts HeptaMB+ as the main intermediate, similar
mechanisms are assumed to exist for polymethylbenzenium
cations with fewer methyl groups, and the labelling patterns are
also expected to be similar if the elimination products are ethene
or isobutene instead of propene (i.e. one unlabelled carbon should
be observed in both alkenes if the paring mechanism dominates).
Looking back at Fig. 6, we observe that the labelling patterns of
ethene and propene at 250 ◦C over H-SAPO-5 are consistent with
a paring mechanism. In our recent study of this catalyst [29], we
concluded from these data that these two alkenes could be formed
by a paring-type mechanism, whereas isobutene was formed via
an alkene methylation/cracking cycle even under co-feed condi-
tions which strongly favour the propagation of the arene cycle.
A cycle of successive alkene methylation and cracking reactions
nicely explains the high abundance of 13C in isobutene, as the
abundance of 13C would increase with each cycle if the starting
molecule was e.g. 12C13C2-propene. From Fig. 7, the labelling pat-
terns of all three alkenes at 250 ◦C over H-SSZ-24 are seen to fit with
a paring-type mechanism. It is noteworthy that this result is in full
agreement with the propene and isobutane labelling patterns in
similar experiments previously performed over H-Beta [15]. Since
the labelling pattern of isobutene in H-SAPO-5 can be explained
by the alkene cycle while all other alkenes display labelling pat-
terns in accordance with the paring mechanism, a paring-type
reaction involving systematic scrambling of ring-carbons appears
responsible for all product formation in the arene cycle over these
catalysts. While the de-alkylation mechanism does not necessarily
proceed exactly as shown by the paring mechanism in Scheme 2, a

M. Westgård Erichsen et al. / Catalysis Today 215 (2013) 216–223
223
of the latter reaction compared to the former. It is plausible that
the positive charges of transition states involving arenes are more
diffuse than those involving alkenes, which could mean that the
reactions of the arene cycle are more sensitive to acid strength than
those of the alkene cycle.
[7] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Angewandte
Chemie International Edition 45 (2006) 1714–1719.
[8] I.M. Dahl, S. Kolboe, Catalysis Letters 20 (1993) 329–336.
[9] I.M. Dahl, S. Kolboe, Journal of Catalysis 149 (1994) 458–464.
[10] I.M. Dahl, S. Kolboe, Journal of Catalysis 161 (1996) 304–309.
[11] B. Arstad, S. Kolboe, Journal of the American Chemical Society 123 (2001)
8137–8138.
From the data presented here it is not possible to conclude
which, if either, of these two factors is most important, but the
observation that it was difficult to propagate the arene cycle
exclusively in H-SAPO-5 even during co-reactions of benzene and
methanol (where the concentration of polyMBs is high) may sug-
gest that the difference in the relative activity of the two classes
of intermediates with acid strength are more important than the
concentration of polyMBs.
[12] B. Arstad, S. Kolboe, Catalysis Letters 71 (2001) 209–212.
[13] W.G. Song, J.F. Haw, J.B. Nicholas, C.S. Heneghan, Journal of the American Chem-
ical Society 122 (2000) 10726–10727.
[14] A. Sassi, M.A. Wildman, H.J. Ahn, P. Prasad, J.B. Nicholas, J.F. Haw, Journal of
Physical Chemistry B 106 (2002) 2294–2303.
[15] M. Bjørgen, U. Olsbye, D. Petersen, S. Kolboe, Journal of Catalysis 221 (2004)
1–10.
[16] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.-P. Lillerud, S. Kolboe,
M. Bjørgen, Journal of the American Chemical Society 128 (2006)
14770–14771.
[17] M. Bjørgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, Journal of Catalysis 249 (2007) 195–207.
[18] S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, Journal of Physical Chemistry C 111
(2007) 17981–17984.
4. Conclusions
[19] M. Bjørgen, F. Joensen, K.-P. Lillerud, U. Olsbye, S. Svelle, Catalysis Today 142
(2009) 90–97.
[20] M. Bjørgen, S. Akyalcin, U. Olsbye, S. Benard, S. Kolboe, S. Svelle, Journal of
Catalysis 275 (2010) 170–180.
[21] S. Teketel, U. Olsbye, K.P. Lillerud, P. Beato, S. Svelle, Microporous and Meso-
porous Materials 136 (2010) 33–41.
[22] S. Teketel, S. Svelle, K.P. Lillerud, U. Olsbye, ChemCatChem 1 (2009) 78–81.
[23] J.F. Haw, W.G. Song, D.M. Marcus, J.B. Nicholas, Accounts of Chemical Research
36 (2003) 317–326.
[24] J.H. Ahn, B. Temel, E. Iglesia, Angewandte Chemie International Edition 48
(2009) 3814–3816.
[25] D.A. Simonetti, J.H. Ahn, E. Iglesia, Journal of Catalysis 277 (2011) 173–195.
[26] L.-T. Yuen, S.I. Zones, T.V. Harris, E.J. Gallegos, A. Auroux, Microporous Materials
2 (1994) 105–117.
[27] F. Bleken, M. Bjorgen, L. Palumbo, S. Bordiga, S. Svelle, K.P. Lillerud, U. Olsbye,
Topics in Catalysis 52 (2009) 218–228.
[28] B.P.C. Hereijgers, F. Bleken, M.H. Nilsen, S. Svelle, K.P. Lillerud, M. Bjorgen, B.M.
Weckhuysen, U. Olsbye, Journal of Catalysis 264 (2009) 77–87.
[29] M. Westgård Erichsen, S. Svelle, U. Olsbye, Journal of Catalysis 298 (2013)
94–101.
Two isostructural zeolite catalysts with differing acid strength
were tested in the MTH reaction, and it was found that the product
mixtures formed were different in the two catalysts. In particular,
the more strongly acidic H-SSZ-24 was more selective towards aro-
matics, ethene and propene than the moderately acidic H-SAPO-5.
Co-feed experiments utilising 12C benzene and 13C methanol indi-
cated differences in the preferred mechanism of alkene formation
between the two catalysts. Together, these results led us to propose
that the arene cycle is more important for product formation in the
strongly acidic H-SSZ-24 than in the moderately acidic H-SAPO-5.
Furthermore, the data imply that the acid strength may be used
to tune the selectivity of zeotype catalysts towards aromatic and
alkane versus alkene products.
Finally, the co-feed studies reported here strongly indicates that
the main mechanistic pathway for alkene formation from polyMBs
involves systematic scrambling of ring- and methyl-carbons. The
best known mechanism of this type is the paring mechanism, but
other ring -expansion or -contraction mechanisms are also feasible.
[30] Y. Kubota, H. Maekawa, S. Miyata, T. Tatsumi, Y. Sugi, Microporous and Meso-
porous Materials 101 (2007) 115–126.
[31] H. Maekawa, Y. Kubota, Y. Sugi, Chemistry Letters 33 (2004) 1126–1127.
[32] D. Cardoso, in: H. Robson (Ed.), Verified Syntheses of Zeolitic Materials, Elsevier,
2001, pp. 118–119.
[33] R.F. Lobo, M.E. Davis, Microporous Materials 3 (1994) 61–69.
[34] Y. Kubota, S. Tawada, K. Nakagawa, C. Naitoh, N. Sugimoto, Y. Fukushima, T.
Hanaoka, Y. Imada, Y. Sugi, Microporous and Mesoporous Materials 37 (2000)
291–301.
[35] J. Chen, P.A. Wright, S. Natarajan, J.M. Thomas, Studies in Surface Science and
Catalysis 84 (1994) 1731–1738.
Acknowledgements
This publication is part of the inGAP Centre of Research-based
Innovation, which receives financial support from the Norwegian
Research Council under Contract No. 174893.
[36] B. Zibrowius, E. Loffler, M. Hunger, Zeolites 12 (1992) 167–174.
[37] S.G. Hedge, P. Ratnasamy, L.M. Kustov, V.B. Kazansky, Zeolites
137–141.
[38] F. Schuth, D. Demuth, B. Zibrowius, J. Kornatowski, G. Finger, Journal of the
American Chemical Society 116 (1994) 1090–1095.
8 (1988)
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
cattod.2013.03.017.
[39] K. Chakarova, K. Hadjiivanov, Journal of Physical Chemistry C 115 (2011)
4806–4817.
[40] J. Martinez-Triguero, M.J. Diaz-Caban˜as, M.A. Camblor, V. Fornés, T.L.M. Maesen,
A. Corma, Journal of Catalysis 182 (1999) 463–469.
[41] S. Bordiga, L. Regli, D. Cocina, C. Lamberti, M. Bjorgen, K.P. Lillerud, Journal of
Physical Chemistry B 109 (2005) 2779–2784.
[42] F.L. Bleken, T.V.W. Janssens, S. Svelle, U. Olsbye, Microporous and Mesoporous
Materials 164 (2012) 190–198.
[43] S. Teketel, W. Skistad, S. Benard, U. Olsbye, K.P. Lillerud, P. Beato, S. Svelle, ACS
Catalysis 2 (2012) 26–37.
[44] A.N. Mlinar, P.M. Zimmerman, F.E. Celik, M. Head-Gordon, A.T. Bell, Journal of
Catalysis 288 (2012) 65–73.
[45] T. Mole, G. Bett, D. Seddon, Journal of Catalysis 84 (1983) 435–445.
[46] T. Mole, J.A. Whiteside, D. Seddon, Journal of Catalysis 82 (1983) 261–266.
[47] R.F. Sullivan, C.J. Egan, G.E. Langlois, R.P. Sieg, Journal of the American Chemical
Society 83 (1961) 1156–1160.
[48] M. Vandichel, D. Lesthaeghe, J. Van der Mynsbrugge, M. Waroquier, V. Van
Speybroeck, Journal of Catalysis 271 (2010) 67–78.
References
ˇ
[1] T. Maesen, The Zeolite Scene, in: J. Cejka, H. van Bekkum, A. Corma, F. Schüth
(Eds.), Introduction to Zeolite Science and Practice, Elsevier B.V., Amsterdam,
2007.
[2] J.F. Haw, Physical Chemistry Chemical Physics 4 (2002) 5431–5441.
[3] A. Zecchina, G. Spoto, S. Bordiga, Physical Chemistry Chemical Physics 7 (2005)
1627–1642.
[4] J. Macht, R.T. Carr, E. Iglesia, Journal of the American Chemical Society 131
(2009) 6554–6565.
[5] U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.V.W. Janssens, F. Joensen, S. Bordiga,
K.P. Lillerud, Angewandte Chemie International Edition 51 (2012) 5810–5831.
[6] W.G. Song, D.M. Marcus, H. Fu, J.O. Ehresmann, J.F. Haw, Journal of the American
Chemical Society 124 (2002) 3844–3845.
[49] N. Hazari, J.A. Labinger, V.J. Scott, Journal of Catalysis 263 (2009) 266–276.