H-SAPO-5 as methanol-to-olefins (MTO) model catalyst: Towards elucidating the effects of acid strength

Article abstract of DOI:10.1016/j.jcat.2012.11.004

The methanol-to-hydrocarbons (MTH) reaction was studied over a moderately acidic zeotype material, H-SAPO-5, at 350-450°C and with WHSV = 0.3-5 h ^-1. C₃-C₅ alkenes were the main products of reaction, followed by C₆₊ aliphatics. Conversion-selectivity plots from experiments conducted at various contact times revealed that coking did not influence product selectivity significantly. Steady-state isotope transient experiments (¹²CH₃OH//¹³CH₃OH) were performed at 450°C. ¹³C incorporation was more rapid in the alkene products than in the polymethylated benzene molecules that were retained inside the catalyst after testing, suggesting that polymethylbenzenes contribute only to a minor extent to alkene formation in H-SAPO-5. Co-feed studies of methanol and benzene at 350°C revealed that benzene shifts the product selectivity towards ethene and aromatic products. Co-feeding ¹³CH₃OH and benzene at 250°C, giving <2% conversion of both reactants, indicated that polymethylbenzenes, when present in excessive amounts, may contribute to ethene and propene, but not to C₄₊ alkene, formation. Furthermore, the isotopic labelling pattern of ethene provided the first direct experimental evidence for ethene formation by a paring-type reaction from polymethylated benzene intermediates. Overall, the results obtained in this study suggest that a lower acid strength promotes an alkene-mediated MTH reaction mechanism, and that acid strength is therefore an important design parameter for selectivity optimisation in zeotype catalysis.

Full text of DOI:10.1016/j.jcat.2012.11.004

Journal of Catalysis 298 (2013) 94–101
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
H-SAPO-5 as methanol-to-olefins (MTO) model catalyst: Towards elucidating
the effects of acid strength
⇑
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:
Received 6 July 2012
Revised 21 September 2012
Accepted 3 November 2012
Available online 27 December 2012
The methanol-to-hydrocarbons (MTH) reaction was studied over a moderately acidic zeotype material,
H-SAPO-5, at 350–450 °C and with WHSV = 0.3–5 hÀ1. C3–C5 alkenes were the main products of reaction,
followed by C6+ aliphatics. Conversion-selectivity plots from experiments conducted at various contact
times revealed that coking did not influence product selectivity significantly. Steady-state isotope tran-
sient experiments (12CH3OH//13CH3OH) were performed at 450 °C. 13C incorporation was more rapid in
the alkene products than in the polymethylated benzene molecules that were retained inside the catalyst
after testing, suggesting that polymethylbenzenes contribute only to a minor extent to alkene formation
in H-SAPO-5. Co-feed studies of methanol and benzene at 350 °C revealed that benzene shifts the product
selectivity towards ethene and aromatic products. Co-feeding 13CH3OH and benzene at 250 °C, giving <2%
conversion of both reactants, indicated that polymethylbenzenes, when present in excessive amounts,
may contribute to ethene and propene, but not to C4+ alkene, formation. Furthermore, the isotopic label-
ling pattern of ethene provided the first direct experimental evidence for ethene formation by a paring-
type reaction from polymethylated benzene intermediates. Overall, the results obtained in this study sug-
gest that a lower acid strength promotes an alkene-mediated MTH reaction mechanism, and that acid
strength is therefore an important design parameter for selectivity optimisation in zeotype catalysis.
Ó 2012 Elsevier Inc. All rights reserved.
Keywords:
Methanol conversion
MTO, H-SAPO-5
Isotopic labelling
Acid strength
1. Introduction
Generally speaking, apart from the simple dehydration to form
dimethyl ether, three main product groups can be obtained. These
As world oil reserves are becoming increasingly difficult to
extract, the interest in utilisation of alternative feedstocks, such
as natural gas and biomass, for production of petrochemicals, is
increasing. The conversion of methanol-to-hydrocarbons (MTH)
represents a family of flexible processes for production of either
light alkenes or gasoline, depending on catalyst used and
conditions employed. As methanol production from a variety of
sources is well established, commercial interest in methanol con-
version processes is on the rise [1]. Several technologies are being
commercialised, such as the UOP/Hydro methanol-to-olefins
(MTO) process, Lurgi’s methanol-to-propene (MTP) process,
Mobil’s methanol-to-gasoline (MTG) process and the Topsøe TIGAS
process [1].
are light alkenes such as ethene and propene, higher alkenes
(C4–C8) and mixtures of aromatic and alkane products. Which
product groups are favoured depend strongly on the catalyst used,
as changes in the available internal channel space influence both
which intermediates can be formed, and which products are able
to diffuse out of the structure [2]. Tuning the reaction so that only
a select group of products is produced is highly desirable, but re-
quires a high degree of insight into both the fundamental reaction
mechanisms and their interplay with the sterical limitations of the
employed catalysts.
The MTH reaction has been known for more than 30 years. It
was first published by Chang and Silvestri [3], and mechanistic
studies have been a core research topic for many years. As an
example, more than 20 mechanisms have been proposed for the di-
rect formation of carbon–carbon bonds from methanol [4], even
though it now appears as if such direct coupling of C–C bonds is
not an important part of the reaction under steady-state conditions
[5,6]. A major breakthrough in mechanistic understanding was the
proposal by Dahl and Kolboe [7–9] that alkene formation proceeds
by sequential methylation and elimination from an initially unde-
fined hydrocarbon pool. Later studies revealed that multiply meth-
ylated aromatics play an important role as hydrocarbon pool
species in the catalysts H-SAPO-34 [10–12] and H-Beta [13,14].
The MTH reaction is carried out over zeolite or zeotype cata-
lysts, which are Brønsted-acidic, crystalline materials with internal
pores and cavities of molecular dimensions. A key issue in MTH is
product selectivity. A complex network of reactions is in operation
during the conversion of methanol into the observed products.
⇑
Corresponding author. Fax: +47 22855441.
E-mail addresses: m.w.erichsen@kjemi.uio.no (M. Westgård Erichsen), stian.
svelle@kjemi.uio.no (S. Svelle), unni.olsbye@kjemi.uio.no (U. Olsbye).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.11.004

M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
95
Hexamethylbenzene and its further methylated cation, the hep-
tamethylbenzenium cation (heptaMB+), are believed to be particu-
larly important intermediates for alkene formation in these
catalysts.
tored and can be manipulated through co-reaction experiments.
By employing a combination of co-feeding and isotopic labelling
experiments, this work provides detailed mechanistic insight for
the MTH reaction over H-SAPO-5.
While the studies that led to the acceptance of the hydrocarbon
pool mechanism and the focus on aromatic intermediates were
performed using catalysts with large pores or internal voids, some
later studies have focused on zeolites where the internal volumes
are too small to allow dealkylation reactions of the largest
polymethylbenzenes (polyMBs). In H-ZSM-5, which consists of
3D 10-ring pores, isotope transient experiments showed that light-
er polyMBs (Xylene, TriMB and TetraMB in particular) were the
main intermediates in the polyMB-based, or arene, cycle [15,16],
leading mainly to ethene formation. In addition, those experiments
revealed that C3+ alkenes were the dominating reaction intermedi-
ates for C3+ formation, by a methylation – cracking cycle [15,16].
Overall, the observed reaction mechanism, where alkenes and are-
nes function side by side as reaction intermediates, and are con-
nected through hydride transfer reactions leading to
transformation of alkenes into aromatics and alkanes, was called
the dual-cycle mechanism [15–17].
Recently, it has been demonstrated that the arene cycle can be
further suppressed by using narrow pored 1D 10-ring catalysts
such as H-ZSM-22 to produce product streams rich in higher al-
kenes and poor in aromatics, alkanes and ethene [18,19]. On the
other extreme, a thorough investigation into three large-pore zeo-
lites revealed similar behaviour and reactivity of the polyMBs as in
H-Beta [20]. Yet other studies showed that it is possible to alter the
mechanistic pathway in large-pore zeolites towards the alkene cy-
cle when operating at very low temperature and high pressure
[21,22].
2. Experimental
2.1. Catalyst synthesis and characterisation
H-SAPO-5 was synthesised hydrothermally from water, trieth-
ylamine (TEA) (Fluka, 99.5%), orthophosphoric acid (Merck, 85%),
Cab-O-Sil M5 (Riedel-de Haën) and Catapal B (Vista). A mixture
with the composition 50 H2O:1 Al2O3:1 P2O5:0.1 SiO2:3.9 TEA
was prepared and crystallised for 4 h at 200 °C in a rotating Tef-
lon-lined steel autoclave. Removal of the structure directing agent
(TEA) was performed by calcination in air at 600 °C for 2 h. Cal-
cined samples were characterised by powder XRD, SEM, N2 adsorp-
tion, NH3 TPA and FT-IR.
For powder XRD measurements, a Siemens D-500 with primary
Ge monochromator and Bragg–Brentano geometry using Cu K
a
radiation (k = 1.5406 Å) was employed. SEM was performed with
a FEI Quanta 200 FEG-ESEM. N2 adsorption isotherms were mea-
sured at À196 °C, using a BELSORP-mini II instrument. The sample
was 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 measurements were performed
in transmission mode on a Vertex 80 instrument with MCT detec-
tor. Before measurement, the sample was deposited on a silicon
wafer and was pretreated under vacuum by heating to 120 °C for
1 h, 300 °C for 1 h and 450 °C for 1 h.
Temperature programmed adsorption of NH3 was performed in
a TGA setup, using a Mettler Toledo TGA/SDTA851e. Samples were
heated to 500 °C under a flow of N2 at a rate of 10 °C minÀ1 and
held there for 15 min before switching to a flow of 2% NH3 in N2.
After a further 40 min, the temperature was decreased to 400 °C
(10 °C minÀ1). This temperature was again held for 40 min before
decreasing to 300 °C, which was held for another 40 min before
decreasing to 200 °C. The weight gain after each temperature de-
crease was logged and used for calculating the amount of adsorbed
NH3 per gram catalyst.
The zeotype catalyst used in commercial MTO plants [1],
H-SAPO-34 (CHA topology), has a significantly lower acid strength
than conventional zeolites such as H-ZSM-5 (MFI topology). The
difference in acid strength has been demonstrated by FTIR
experiments using CO as probe molecule at À196 °C, showing a
weaker interaction in the case of H-SAPO-34 than of H-ZSM-5
(D
m
OH = À270 cmÀ1 [23] and > À300 cmÀ1 [24], respectively). A
long-standing, but not yet fully resolved issue, is the influence of
acid strength on the MTH performance of the catalyst. This issue
was first addressed by Yuen et al. [25], who compared aluminosil-
icate and silicoaluminophosphate versions of both AFI (H-SAPO-5
versus H-SSZ-24) and CHA (H-SAPO-34 versus H-SSZ-13) struc-
tured catalysts. They compared the materials under conditions giv-
ing full initial methanol conversion and observed a more rapid
breakthrough of methanol over the zeolite catalysts compared to
their silicoaluminophosphate analogues. More recently, a detailed
comparison between H-SAPO-34 and H-SSZ-13 was published
[26]. The study included a series of tests performed at less than full
conversion at different temperatures, as well as characterisation of
retained hydrocarbons after testing. As expected, the more acidic
H-SSZ-13 had a higher activity, and deactivated faster, than the
less acidic H-SAPO-34. However, it was surprisingly found that
at one test temperature, the methanol conversion capacity (i.e.
the cumulative methanol production before full deactivation)
was higher over H-SSZ-13 than over H-SAPO-34. Further
mechanistic insight proved difficult due to the severe diffusional
constraints in the CHA structure leading to a high degree of
product shape selectivity [27].
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
12C methanol (VWR, 99.8%), 13C methanol (Cambridge Isotope Lab-
oratories, 99%) and benzene (Riedel-de Haën, 99.5%).
Three types of experiments were performed as part of this
study:
1. Ordinary methanol reactions, where 12C methanol was reacted
over H-SAPO-5 at 350 or 450 °C. 100 mg of catalyst was used,
and methanol was fed by passing a stream of helium through
a methanol saturator at 0 °C (methanol partial pressure of
40 mbar). Gas flows were varied in order to obtain WHSV
between 0.31 and 5.0 hÀ1, and the effluent was analysed by
on-line GC during the reaction.
The present work employs a zeotype catalyst of similar acid
strength as H-SAPO-34, namely H-SAPO-5 (AFI structure).
H-SAPO-5 features a one-dimensional pore system of tubular chan-
nels that, while similar in diameter to the H-SAPO-34 cavities, are
open to the exterior and allows bulky compounds to diffuse in and
out of the structure. Importantly, H-SAPO-5 can thus be employed
as a model system in which the hydrocarbon pool is easily moni-
2. Steady-state isotope transient experiments, where 12C metha-
nol was reacted over 50 mg of catalyst at 450 °C for 18 min
before switching to a feed of 13C labelled methanol. Two identi-
cal lines with methanol partial pressures of 40 mbar and WHSV

96
M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
0.93 hÀ1 were used for this purpose. After the feed to the reactor
was rapidly changed, the isotopic compositions of both effluent
and compounds retained inside the catalyst were monitored.
3. Co-reactions of methanol and benzene over 100 mg of catalyst
were performed with feed partial pressures of 40 mbar and
13 mbar, respectively (saturator temperatures of 4 °C and
10 °C for the two lines), for experiments with unlabelled reac-
tants (WHSV 1.5 and 1.2 hÀ1). The experiment was carried out
at 350 °C. When 13C methanol and 12C benzene were used, par-
tial pressures were 76 mbar and 25 mbar, respectively (satura-
tor temperatures of 20 °C and 10 °C, WHSV 1.0 and 0.8 hÀ1).
These experiments were carried out at 250 and 275 °C.
Fig. 2. FT-IR spectra of increasing dosages of CO on H-SAPO-5 at À196 °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 clean 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
RtxÒ-DHA-150 column (150 m, 0.25 mm i.d., stationary phase
correspond well with observations from previous studies of
H-SAPO-5 and have been assigned to Si–OH–Al hydroxyls located in
the 12- and 6-rings of the structure, respectively [28]. The hydrox-
yls in the 6-rings are reported to be inaccessible to weakly basic
probe molecules at low temperatures, but will react with stronger
bases and/or at elevated temperatures [28,29]. Thus, they are not
expected to interact with CO at À196 °C, and, indeed, no interac-
tion is observed here. However, the broad absorption for hydroxyls
in the 12-rings (centred around 3626 cmÀ1) erodes upon CO
adsorption, mirroring the growth of two bands at lower wavenum-
bers (3460 cmÀ1 and 3358 cmÀ1). These two bands may represent
different types of acid sites, as previously suggested [30]. Alterna-
tively, the two bands may be caused by a similar fermi resonance
effect as that proposed by Chakarova and Hadjiivanov [24] when
adsorbing CO on H-ZSM-5. Regardless of the nature of this
thickness 1 lm). Hydrogen (purity 6.0) was used as carrier gas.
For isotope analysis of effluent in the benzene/methanol co-feed
experiments, the same GC and conditions were used, but the
effluent was routed to an Agilent 5975C MS detector. The isotopic
composition of the gas-phase effluent from transient experiments
was analysed by injection into a HP6890 GC with a GS GasPro
column (30 m, 0.32 mm i.d.) and HP 5973 MS detector running
on helium (purity 5.0) carrier gas.
In order to analyse aromatic compounds trapped inside the
zeolite voids, catalyst samples quenched at predetermined times
were dissolved in 15% hydrofluoric acid. The aromatics were then
extracted using CH2Cl2 (Merck, 99.9%) and subsequently analysed
by an Agilent 6890 GC equipped with an Agilent 5973 MS detector.
The column used was a HP-5ms (60 m, 0.25 mm i.d., 0.25
stationary phase), and carrier gas was helium (purity 5.0).
lm
other absorption band, the largest observed red-shift (DmOH =
À268 cmÀ1) is comparable to what has been found in similar
OH = À270 cmÀ1 [23]). Thus, a sim-
3. Results and discussion
experiments for H-SAPO-34 (
ilar acid strength is also inferred. In addition, this value is signifi-
cantly lower than the shifts
measured for high-silica zeolites [24]. The other
D
m
OH > À300 cmÀ1 commonly
3.1. Physical characteristics of the catalyst
D
m
DmOH value
(À196 cmÀ1) is much lower, suggesting a weakly acidic site. The
CO stretching region of the spectra (Fig. 2b) suggests similar con-
clusions, as the observed absorptions at 2138 cmÀ1 (free, liquid-
like CO molecules) and 2173 cmÀ1 (CO interacting with Brønsted
sites) correspond well with spectra of H-SAPO-34 [23]. In addition,
an absorption is observed between these two possibly indicative of
CO interacting with a weaker acid site. None of the absorptions
normally assigned to Si–OH, Al–OH or P–OH hydroxyls were ob-
served for this sample, implying that it contains few defects.
Acid site density of the material was determined by NH3 TPA.
Uptake of ammonia per gram catalyst corresponded to an
(Al + P)/Si ratio of 43. The BET surface area of the sample was found
to be 341 m2 gÀ1, which is in the upper range of values reported in
literature [31–35].
Fig. 1 displays the powder XRD pattern and a representative
SEM micrograph of the sample. The diffractogram corresponds to
that of a pure highly crystalline AFI structure. SEM revealed hexag-
onal crystals, mostly around 1
lm in diameter, with lengths of the
crystals varying between 1 and 2
lm.
FT-IR spectra of the activated sample exposed to increasing
dosages of CO at À196 °C are shown in Fig. 2. The clean sample
displays two absorptions in the O–H stretching region (Fig. 2a):
one at 3626 cmÀ1 and another at 3530 cmÀ1. These absorptions
3.2. Methanol conversion over H-SAPO-5
Methanol was reacted over the H-SAPO-5 sample at tempera-
tures of 350 °C and 450 °C, within a range of WHSVs between 0.3
and 4.9 hÀ1. Table 1 lists representative product selectivities at
65% and 20% methanol conversion for the two temperatures stud-
ied. HTI values listed are hydrogen transfer indexes: the ratios be-
tween the yield of alkanes and the yield of alkenes and alkanes
combined. Iso values represent the fraction of branched versus
linear hydrocarbons (alkanes and alkenes). Details on methanol
conversion, deactivation and product yields can be found in the
Supplementary information (Figs. S1 and S2).
Overall, the dominant products of methanol conversion over the
H-SAPO-5 catalyst are C3–C5 hydrocarbons (Table 1), mainly in the
Fig. 1. Powder XRD diffractogram (top) and SEM micrograph (bottom) of the
employed H-SAPO-5 sample.

M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
97
Table 1
H-SAPO-5 sample is displayed in Fig. 3. One minute after the
switch, the 13C content of the alkene fraction was 70–85%, while
it was only 4–30% in the retained aromatics fraction. With increas-
ing time after the switch (2 min), the 13C content of both fractions
increased, to 85–95% in the alkene fraction and 10–50% in the re-
tained aromatics fraction. The hydrocarbon pool mechanism is
far too complex to allow for quantitative determination of individ-
ual reaction rates from the transient experiment. However, the
much slower incorporation of 13C in the aromatic compounds
compared to the alkenes, as seen in Fig. 3, clearly suggests that
polyMBs are not major reaction intermediates for alkene formation
in the H-SAPO-5 catalyst under the given conditions.
Selectivity of H-SAPO-5 samples at 20% and 65% methanol conversion for temper-
atures 350 °C and 450 °C. HTI refers to hydrogen transfer indexes: yield of alkanes/
(yield of alkenes + alkanes). Iso refers to the fraction of branched hydrocarbons,
covering both alkanes and alkenes.
Compounds
350 °C
20%
450 °C
20%
350 °C
65%
450 °C
65%
Methane
C2
C3
C4
0.2
1.3
3.2
1.5
0.2
2.7
2.6
1.7
12.0
51.5
15.7
14.0
5.3
19.1
43.1
17.0
13.6
2.6
21.2
38.1
11.9
10.6
15.4
28.4
37.4
13.9
11.1
4.9
C5
C6+ aliphatics
Aromatics
A comparison of the fraction of 13C in the retained polyMBs and
light alkene effluent after 18 min of 12C methanol feed and 2 min of
13C methanol over H-SAPO-5 and previously reported similar
experiments over the catalysts H-Beta, H-ZSM-5 and H-ZSM-22 is
included in the Supplementary information (Fig. S3). Starting from
the sample with the narrowest pore size, H-ZSM-22 (4.6 Â 5.7 Å,
no intersections), via the sample with intermediate pore size,
H-ZSM-5 (5.3 Â 5.6 Å for the largest pores, with 3D intersections)
to the sample with largest pore size, H-Beta (7.7 Â 6.6 Å for the
largest pores, with 3D intersections), a clear trend is observed: In
H-ZSM-22, 13C incorporation is rapid in the light alkene fraction,
while it is very slow in the retained polyMB fraction, thus suggest-
ing that the polyMB fraction does not contribute to alkene forma-
tion in this structure [18]. In H-ZSM-5, 13C incorporation is
significantly more rapid in the C3+ alkenes compared to the re-
tained polyMBs, but the polyMBs are significantly more active than
in H-ZSM-22. Furthermore, the incorporation of 13C in ethene is
very similar to that of the polyMBs, which led to the conclusion
that ethene is formed from polyMB intermediates, while C3+ al-
kenes are formed mainly from alkene intermediates in H-ZSM-5
[15,16]. Finally, in H-Beta zeolite, the 13C incorporation with time
after switch is very similar in the alkene fraction and the aromatics
fraction, suggesting that polyMBs are even more important
reaction intermediates in this structure [17,37]. In conclusion,
the comparison shows a clear trend with respect to an increasing
importance of PolyMB intermediates with an increasing pore/
cavity size.
C4 HTI
C5 HTI
0.07
0.14
0.03
0.06
0.27
0.40
0.03
0.07
C4 iso
C5 iso
0.77
0.80
0.67
0.71
0.73
0.84
0.59
0.69
form of alkenes. However, the aromatics fraction becomes signifi-
cant at high conversions (65% versus 20% in Table 1) and low reac-
tion temperature and is accompanied by an increasing alkane
fraction; shown in Table 1 as an increasing HTI. Branched hydro-
carbons are found in higher abundance than linear ones across
all product fractions, especially at low conversion levels (Table 1).
Among the main products, the ratio between C4 and C3 hydrocar-
bons decreases sharply as methanol conversion increases. A higher
selectivity to C3 than C4 products is obtained at higher conversions
than those listed here (at 450 °C), while at lower conversion levels,
isobutene becomes increasingly dominant. Conversely, the ratio
between C4 and C5 hydrocarbons appears largely unaffected by
changes in conversion level, remaining nearly unchanged over
the entire conversion range studied. C4 hydrocarbons are favoured
over both C3 and C5 hydrocarbons when the temperature is de-
creased from 450 °C to 350 °C.
The selectivity versus conversion data shown in Supplementary
information (Figs. S1 and S2) indicate that product selectivities de-
pend mainly on reaction temperature and the degree of methanol
conversion, while being mostly independent of both the feed
rate and the degree of deactivation. This observation corresponds
well with the proposal by Janssens [36] that deactivation is non-
selective and can be viewed simply as the loss of active sites,
without any influence on the product selectivity. An exception to
this general trend is observed for severely deactivated samples,
where the selectivity to methane becomes higher than if the same
conversion levels were obtained over a fresh catalyst.
Another prominent feature when comparing H-Beta zeolite
with H-ZSM-5 (Fig. S3) is that 13C is more rapidly incorporated in
hexamethylbenzene among the polyMBs in H-Beta zeolite (in
agreement with a higher alkene formation rate with increasing
3.3. Steady-state methanol isotope transient experiments
Steady-state isotope transient experiments are commonly used
in catalysis research to distinguish between reaction intermediates
and spectator molecules. In general, an active intermediate will
incorporate the labelled component more rapidly than, or equally
fast as, the product molecules. On the other hand, a spectator mol-
ecule will incorporate the labelled component more slowly than
the products. In the MTH process, two main classes of possible
intermediates are commonly considered, that is, alkenes and are-
nes, respectively (see Section 1).
In this study,
a steady-state switch from unlabelled to
13C-labelled methanol was performed after 18 min on stream. This
test duration should allow unlabelled hydrocarbon pool species to
be built up before the switch. At the time of the switch, the conver-
sion of methanol was 75%.
Fig. 3. Evolution of 13C in effluent and retained aromatics after switching from 12C
to 13C methanol. The experiment was performed at 450 °C and WHSV 0.93 hÀ1, and
the isotopic switch executed after 18 min on stream.
The evolution of 13C content with time after isotopic switch for
light alkenes in the effluent as well as retained polyMBs in the

98
M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
number of methyl substituents on the benzene ring, as reported by
Sassi et al. [13]), while it is more rapidly incorporated in xylene and
trimethylbenzene in H-ZSM-5. This observation suggests that the
pores and intersections of H-ZSM-5 are too narrow to promote
dealkylation of the heavier polyMBs [16]. Now, when turning to
H-SAPO-5, which has pore dimensions similar to H-Beta
(7.3 Â 7.3 Å) but with no intersections, the 13C incorporation pat-
tern among the polyMBs follows the expected reactivity trend,
where 13C incorporation is fastest in the heaviest polyMBs. This
observation demonstrates that the available space in the pores of
H-SAPO-5 is larger than in H-ZSM-5. In spite of this, the 13C incor-
poration in the polyMBs is much slower in H-SAPO-5 than in both
H-Beta and H-ZSM-5, while the 13C content observed for the alkene
fraction is similar in all three materials.
Fig. 4. Comparison of the distribution within the aliphatic products range during
co-reaction of benzene and methanol (WHSV 1.2 hÀ1 and 1.5 hÀ1, respectively) and
for pure methanol feed (WHSV 0.6 hÀ1) at 350 °C. The comparison has been made at
approximately 18.5% conversion of either methanol alone, or methanol and
benzene combined. The total selectivity to aliphatics was 40% and 95%, respectively,
for the two experiments.
The large difference in 13C incorporation between polyMBs and
alkenes in H-SAPO-5 suggests that the reaction mechanism is
shifted in the direction of an alkene-mediated mechanism in
H-SAPO-5, compared to a higher importance of the arene-mediated
mechanism in H-Beta and H-ZSM-5. Taking into account that the
pore size of H-SAPO-5 is larger than of H-ZSM-5 and similar to
H-Beta, the comparison strongly suggest that the shift is due to
the lower acid strength of H-SAPO-5 compared to H-ZSM-5 and
H-Beta.
While these experiments strongly suggest that methylbenzenes
do not contribute significantly to alkene formation in H-SAPO-5
under the conditions tested, some indications that methylbenzenes
could still be active hydrocarbon pool species are found. In partic-
ular, the 13C incorporation is not as low as in narrow-pore zeolites
such as H-ZSM-22 ([18] and comparison in Supplementary infor-
mation Fig. S3). Furthermore, in the effluent, ethene displays a low-
er incorporation of 13C than propene and isobutene. This
observation may indicate a mechanistically separate pathway to
ethene formation, possibly more closely linked with polyMB inter-
mediates than formation of the heavier alkenes, as previously sug-
gested for H-ZSM-5 [15,16].
3.5. Selective labelling of aromatic intermediates
In order to reveal mechanistic information on the polyMB-med-
iated reaction cycle, 13CH3OH was co-fed with 12C-benzene at
250 °C and WHSV 1.0 hÀ1 and 0.8 hÀ1 for methanol and benzene,
respectively (molar ratio 3:1). Under these conditions, the conver-
sion was low (1.8% of methanol and benzene combined), and reac-
tions other than (methyl)benzene methylation and subsequent
dealkylation were suppressed. This was verified both by the low
selectivity to aliphatic products (8%) and by the observation that
the polyMBs contained predominantly the same number of 13C
atoms as methyl groups, with only a small degree of scrambling
observed in the most highly methylated compounds (see Supple-
mentary information, Fig. S4). While these conditions are some-
what removed from conventional methanol conversion
conditions, they were necessary in order to avoid formation of
methylbenzenes from methanol alone, which would render the al-
kene labelling patterns unintelligible. The effluent was analysed
after 2 min on stream and labelling patterns of the light alkenes
are shown in Fig. 5.
From this figure, some interesting differences between the
products are immediately apparent. While the majority of the eth-
ene and propene molecules contain one 12C atom, the isobutene
and isopentene (2-methyl-2-butene) molecules display an excess
of all 13C molecules. The most natural assumption from these re-
sults is that the two groups of alkenes are formed via different
reaction pathways. To elaborate further on this difference, it is
necessary to recapitulate the current understanding of polyMB
dealkylation mechanisms.
3.4. Effects of benzene as co-reactant
The low 13C incorporation rate observed for retained polyMBs
compared to alkenes in H-SAPO-5 (Section 3.3) could either mean
that H-SAPO-5 is not able to catalyse the dealkylation of polyMBs,
or that the rate of polyMB dealkylation is slow compared to alkene
methylation and cracking. A low rate of polyMB dealkylation could
either be due to a low abundance of these products, or to a low
reaction rate, compared to what is observed in the more acidic
zeolites. To elucidate whether polyMB dealkylation is feasible in
H-SAPO-5, several methanol and benzene co-feed experiments
were carried out.
Product selectivities obtained for methanol feed alone and for
methanol – benzene co-feed (molar ratio 3:1, partial pressures
40 and 13 mbar, respectively) at 350 °C are shown in Fig. 4. The
conversion was approximately 18.5% in both experiments. It is
apparent from Fig. 4 that the addition of benzene to the feed in-
creases the relative selectivity towards C2 and C3 hydrocarbons at
the expense of C4 selectivity. This is especially apparent for the eth-
ene selectivity, which is very low (1.3%) when methanol is reacted
alone but increases almost fivefold (to 5.7%) when benzene is
added to the feed. Previous mechanistic studies of the MTH reac-
tion over H-ZSM-5 indicated that ethene is predominantly formed
from dealkylation of polyMBs (Section 3.3), while C3+ alkenes are
more easily formed via an alkene methylation – cracking mecha-
nism [16–18]. The observed change in product distribution upon
benzene co-feed over H-SAPO-5, and in particular the increase in
ethene selectivity, suggests that H-SAPO-5 is indeed able to cata-
lyse polyMB dealkylation.
The mechanisms of alkene elimination from polyMBs in the
MTH reaction have been a subject of debate for many years. The
two main mechanistic proposals are the paring reaction and the
exocyclic methylation mechanism. A simplified scheme of these
two reaction pathways is presented in Fig. 6. The latter mechanism
was first proposed by Mole et al. [38,39] and later refined by Haw
et al. [2] and involves deprotonation of HeptaMB+ to form a neutral
compound with an exocyclic double bond. This double bond can
then 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. [40], 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 depro-
tonation and expansion back to a six-ring. In Fig. 6, stars signify
13C-labelled carbon atoms, that is, the carbons originally present
as methyl groups. Since only the paring reaction involves

M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
99
Fig. 5. Distribution of 13C atoms in ethene (left), propene (middle left), isobutene (middle right) and isopentene (right) after 2 min of co-reacting methanol with benzene
(molar ratio 3:1) at 250 °C. Total 13C content was 53% in ethene and 67% in propene, while it was nearly 90% in both isobutene and isopentene. The molecules above the
graphs are drawn with stars signifying labelled carbon atoms representative of the most abundant isotopologues.
isopentene) from this cycle would necessitate, first, deprotonation
of the ethyl side-chain, followed by another methylation to form an
isopropyl group, which would again need to be deprotonated be-
fore finally being methylated to an isobutyl group (or twice to an
isopentyl group). It is not likely that such a sequence of steps
would lead selectively to isobutene and isopentene formation,
while giving insignificant amounts of ethene and propene. It is
therefore concluded that the dominating all-13C isobutene label-
ling pattern shown in Fig. 5 signifies that isobutene and isopentene
are mainly formed via an ‘all-methanol’ cycle in the co-feed exper-
iments, that is, by the alkene methylation – cracking cycle, where
the abundance of 13C (from methylation of C3+ by methanol) in-
creases with each cycle when originally starting with 12C13C2-
propene.
The results of this experiment reveal again some interesting dif-
ferences to the results previously obtained by Bjørgen et al. [14]
over H-Beta zeolite. A comparison of the results obtained by co-
feeding 13C-methanol and 12C-benzene over H-Beta zeolite in
[14] with those obtained for H-SAPO-5 in this study is shown in
the Supplementary information, Fig. S5. The most striking differ-
ence is observed for isobutene/isobutane, where H-SAPO-5 gives
predominantly all-13C-isobutene at both temperatures, while
H-Beta zeolite gives predominantly isobutane with one 12C. The
more acidic H-Beta zeolite thereby seems to favour the arene-
based mechanism, where iso-C4 is formed by a paring-type
reaction, and where the hydrogen transfer index (HTI) is high
due to significant formation of aromatic products (and coke). The
less acidic H-SAPO-5, on the other hand, favours the alkene-based
mechanism, in which isobutene is formed by cracking of higher
alkenes, and the HTI is low, in agreement with a lower selectivity
to aromatic products over this catalyst. It should be noted that
the labelling pattern of isobutane in H-SAPO-5 was similar to that
of isobutene, but could not be accurately determined due to the
almost negligible selectivity to isobutane.
Turning to the propene labelling pattern, which was only re-
ported at 275 °C for H-Beta zeolite, propene with one 12C domi-
nates for H-SAPO-5 at 250 °C, strongly indicating formation by a
paring-type dealkylation of polyMBs. At 275 °C, however, where
the total methanol and benzene conversion was 6.9% (compared
to 1.8% at 250 °C), the all-13C propene compound dominated,
suggesting a stronger contribution from the alkene methylation-
cracking cycle in propene formation at this temperature. In H-beta
zeolite, on the other hand, propene with one 12C dominates even at
270 °C, at 6.9% conversion, suggesting that the arene-based cycle
still dominates propene formation at this temperature. Finally, in
the ethene fraction, which was not reported for H-Beta zeolite,
the monolabelled compound dominated in H-SAPO-5 at both tem-
peratures, suggesting that ethene is formed predominantly via a
paring-type reaction of polyMB intermediates.
Fig. 6. Expected labelling patterns when propene is split off from a selectively
labelled heptaMB+ via paring or exocyclic mechanisms, respectively. Stars signify
labelled carbon atoms.
scrambling of the benzene ring carbons into the alkene products
(and methyl group carbons into the aromatic ring), the labelling
patterns of the alkene products may be used to distinguish be-
tween the two mechanisms. While the scheme in Fig. 6 depicts
HeptaMB+ as the main intermediate, similar mechanisms are as-
sumed 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).
Now, when returning to the data in Fig. 5, we observe that for
ethene and propene, the labelling patterns are consistent with a
paring mechanism. While a similar labelling pattern has previously
been observed for propene [14], this is, to the best of our knowl-
edge, the first reported evidence of ethene formation via the paring
reaction in any acid catalyst. However, ethene loss from poly MBs
via paring-type mechanisms has been studied by mass spectromet-
ric methods by Uggerud et al. [41,42].
It is interesting to note that the C4+ alkenes show a different
labelling pattern and are dominated by all-13C compounds. Such
a labelling pattern could either be an indication of a shift in the
polyMB dealkylation mechanism towards exocyclic methylation
for higher alkenes, or it could indicate that the C4 alkenes are not
formed via polyMB intermediates, but rather via an alkene methyl-
ation-cracking cycle. Looking closely at the exocyclic methylation
cycle in Fig. 6, it is observed that elimination of isobutene (or

100
M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
3.6. Mechanistic implications
Furthermore, C8 alkenes are the largest alkenes observed in the
effluent of H-SAPO-5 in this study. Finally, the observation that
the ratio between C4 and C5 hydrocarbons formed does not vary
significantly when conditions are changed (apart from tempera-
ture) corresponds well with their formation primarily from a com-
mon intermediate. If this common intermediate is an alkene, it
would be expected to consist of at least eight carbons, since smal-
ler species would be unable to crack into a C5 compound without
the involvement of primary carbocations or elimination of ethene
(which is produced in very small amounts over H-SAPO-5).
In the arene cycle of Fig. 7, ethene and propene are formed from
polyMBs. In line with the observed and previously reported [13],
correlation between reactivity and increasing number of methyl
groups, the largest observed polyMB, hexamethylbenzene, is
drawn as the key intermediate in the figure. The light alkenes are
assumed to be formed in a paring-type mechanism since the strong
preference for incorporation of one aromatic-ring carbon into eth-
ene and propene cannot be explained by an exocyclic methylation
pathway. More recent theoretical work also favours the paring
reaction over the exocyclic methylation reaction mechanism in
zeolite H-ZSM-5 [44].
This study provides, as far as we know, the first experimental
evidence reported for ethene formed by a paring-type reaction
from polyMBs. It is worth mentioning that, according to theoretical
work by Arstad et al. [45,46], formation of ethyl chains cannot pro-
ceed via a ring contraction. On the other hand, the same studies
also concluded that it may proceed with the involvement of a
ring-expansion step. The expansion to tropylium-type ions before
alkene elimination has been studied by mass spectrometry
[41,42]. Either way, similar incorporation of ring carbons into the
eliminated alkenes may occur in mechanisms, involving ring
expansion and ring contraction. For this reason, it appears prudent
to refer to the mechanism involved simply as ‘paring-type’ or car-
bocation-based, without stating whether contraction or expansion
is involved.
Together, the data reported in Sections 3.2–3.5 give the follow-
ing key information about the methanol-to-hydrocarbons reaction
over H-SAPO-5:
1. H-SAPO-5 gives a high selectivity to non-aromatic products,
predominantly C3-C5 alkenes, at 350–450 °C.
2. Steady-state isotope transient experiment with 12CH3OH//
13CH3OH feed at 350 °C reveals a more rapid incorporation of
13C in alkene products than in retained polyMBs, indicating that
alkene formation proceeds mainly via an alkene methylation-
cracking cycle in this catalyst.
3. Co-feed studies of methanol and benzene at 350 °C give a shift
in product distribution towards C2 and C3 alkenes compared to
methanol only feed, suggesting that arenes (when present in
sufficient amounts) may act as intermediates for light alkene
formation in H-SAPO-5.
4. Co-feed studies of 13C-methanol and 12C-benzene at conditions
giving low conversions, and high concentrations of polyMBs in
the catalyst, at 250–275 °C give isotope labelling patterns in
the C2–C5 products which suggest that ethene and propene
could be formed by an arene cycle, via a paring-type dealkyla-
tion of polyMBs when PolyMBs are present in sufficient
amounts, while isobutene and isopentene (and partly propene)
are formed mainly via an alkene cycle, via methylation and
cracking of alkenes.
Together, these observations lead to the suggestion of a reaction
scheme, which is presented in Fig. 7. Here, an arene cycle is as-
sumed to account for production of the lightest alkenes (ethene
and, partly, propene), while an alkene cycle produces the heavier
alkenes. Such a ‘dual-cycle mechanism’ is similar to the proposed
mechanism for H-ZSM-5 [15,16]. However, due to the increased
available space inside the channels of H-SAPO-5 relative to
H-ZSM-5, both the proposed alkene and arene cycles consist of
larger compounds in H-SAPO-5 than in H-ZSM-5. Note also that
interconversion between the two cycles would account for many
of the observed minor products, such as cyclic compounds, polyun-
saturated hydrocarbons and alkanes.
Perhaps the most interesting observation of this mechanistic
investigation is the uncovered mechanistic differences from previ-
ously studied medium- and large-pore zeolites; that is, the rela-
tively higher 13C incorporation in alkene products than in
retained PolyMBs in H-SAPO-5, compared to previously studied
zeolites, after a methanol isotope transient (Fig. S3), as well as
the clear difference in labelling pattern of the isobutene/isobutane
fraction in 13C-methanol–12C-benzene co-feed experiments per-
formed over H-SAPO-5 in this study, and over H-Beta zeolite in a
In the alkene cycle of Fig. 7, 2,3,3-trimethyl-2-butene has been
drawn as the end product. While other homologues and isomers
certainly exist, this proposal is in accordance with recent reports
by Bercaw et al. [43] and Iglesia et al. [21,22] for similar systems.
Fig. 7. Scheme of the proposed dual-cycle mechanism in operation during methanol-to-hydrocarbons catalysis over H-SAPO-5.

M. Westgård Erichsen et al. / Journal of Catalysis 298 (2013) 94–101
101
[2] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003) 317–
326.
[3] C.D. Chang, A.J. Silvestri, J. Catal. 47 (1977) 249–259.
[4] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3–48.
[5] W.G. Song, D.M. Marcus, H. Fu, J.O. Ehresmann, J.F. Haw, J. Am. Chem. Soc. 124
(2002) 3844–3845.
previous study (Fig. S5). Together, these observations suggest that
the lower acid strength of the H-SAPO-5 catalyst compared to the
previously studied zeolites induces a shift in reaction mechanism
from predominantly arene-based in wide pore zeolites, to mainly
alkene-based in wide pore H-SAPO-5. However, an influence of
the differences in framework structure among the samples cannot
be ruled out without direct comparisons of isostructural catalysts
with different acid strength. Indeed, as a closing comment on
topology influence, it should be noted that the conditions em-
ployed in the co-feed experiments in this study probably lead to
a hydrocarbon pool composition more similar to that obtained in
H-SAPO-34, because its cavity-window structure leads to trapping
and accumulation of aromatic compounds, in favour of an arene-
based mechanism for light alkene formation [10,11,27,47]. This
observation further underlines that MTH catalyst optimisation is
indeed a multiparameter task.
[6] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Angew. Chem. Int.
Ed. 45 (2006) 1714–1719.
[7] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329–336.
[8] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458–464.
[9] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304–309.
[10] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137–8138.
[11] B. Arstad, S. Kolboe, Catal. Lett. 71 (2001) 209–212.
[12] W. Song, J.F. Haw, J.B. Nicholas, C.S. Heneghan, J. Am. Chem. Soc. 122 (2000)
10726–10727.
[13] A. Sassi, M.A. Wildman, H.J. Ahn, P. Prasad, J.B. Nicholas, J.F. Haw, J. Phys. Chem.
B 106 (2002) 2294–2303.
[14] M. Bjørgen, U. Olsbye, D. Petersen, S. Kolboe, J. Catal. 221 (2004) 1–10.
[15] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.P. Lillerud, S. Kolboe, M. Bjørgen, J.
Am. Chem. Soc. 128 (2006) 14770–14771.
[16] M. Bjørgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 249 (2007) 195–207.
[17] S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, J. Phys. Chem. C 111 (2007) 17981–
17984.
4. Conclusions
[18] S. Teketel, U. Olsbye, K.P. Lillerud, P. Beato, S. Svelle, Micropor. Mesopor. Mater.
136 (2010) 33–41.
[19] S. Teketel, S. Svelle, K.P. Lillerud, U. Olsbye, ChemCatChem 1 (2009) 78–81.
[20] M. Bjørgen, S. Akyalin, U. Olsbye, S. Benard, S. Kolboe, S. Svelle, J. Catal. 275
(2010) 170–180.
[21] J.H. Ahn, B. Temel, E. Iglesia, Angew. Chem. Int. Ed. 48 (2009) 3814–3816.
[22] D.A. Simonetti, J.H. Ahn, E. Iglesia, J. Catal. 277 (2011) 173–195.
[23] S. Bordiga, L. Regli, D. Cocina, C. Lamberti, M. Bjørgen, K.P. Lillerud, J. Phys.
Chem. B 109 (2005) 2779–2784.
[24] K. Chakarova, K. Hadjiivanov, J. Phys. Chem. C 115 (2011) 4806–4817.
[25] L.-T. Yuen, S.I. Zones, T.V. Harris, E.J. Gallegos, A. Auroux, Micropor. Mater. 2
(1994) 105–117.
[26] F. Bleken, M. Bjørgen, L. Palumbo, S. Bordiga, S. Svelle, K.P. Lillerud, U. Olsbye,
Top. Catal. 52 (2009) 218–228.
[27] B.P.C. Hereijgers, F. Bleken, M.H. Nilsen, S. Svelle, K.P. Lillerud, M. Bjørgen, B.M.
Weckhuysen, U. Olsbye, J. Catal. 264 (2009) 77–87.
[28] B. Zibrowius, E. Löffler, M. Hunger, Zeolites 12 (1992) 167–174.
[29] F. Schüth, D. Demuth, B. Zibrowius, J. Kornatowski, G. Finger, J. Am. Chem. Soc.
116 (1994) 1090–1095.
[30] C. Halik, J.A. Lercher, H. Mayer, J. Chem. Soc., Faraday Trans. 1 (84) (1988)
4457–4469.
[31] J.M. Campelo, F. Lafont, J.M. Marinas, M. Ojeda, Appl. Catal. A: General 192
(2000) 85–96.
[32] L. Wang, C. Guo, S. Yan, X. Huang, Q. Li, Micropor. Mesopor. Mater. 64 (2003)
63–68.
[33] S. Seelan, A.K. Sinha, J. Mol. Catal. A – Chem. 215 (2004) 149–152.
[34] S.H. Jhung, J.-S. Chang, D.S. Kim, S.-E. Park, Micropor. Mesopor. Mater. 71
(2004) 135–142.
[35] A.K. Sinha, S. Sainkar, S. Sivasanker, Micropor. Mesopor. Mater. 31 (1999) 321–
331.
[36] T.V.W. Janssens, J. Catal. 264 (2009) 130–137.
[37] M. Bjørgen, F. Joensen, K.P. Lillerud, U. Olsbye, S. Svelle, Catal. Today 142
(2009) 90–97.
[38] T. Mole, G. Bett, D. Seddon, J. Catal. 84 (1983) 435–445.
[39] T. Mole, J.A. Whiteside, D. Seddon, J. Catal. 82 (1983) 261–266.
[40] R.F. Sullivan, C.J. Egan, G.E. Langlois, R.P. Sieg, J. Am. Chem. Soc. 83 (1961)
1156–1160.
[41] S. Svelle, M. Bjorgen, S. Kolboe, D. Kuck, M. Letzel, U. Olsbye, O. Sekiguchi, E.
Uggerud, Catal. Lett. 109 (2006) 25–35.
A combination of conventional, co-feed and labelling experi-
ments has been performed for methanol conversion over the
weakly acidic catalyst H-SAPO-5. Together, the experiments have
led to the proposal of a mechanistic scheme where both polymeth-
ylbenzenes (polyMBs) and higher alkenes operate as hydrocarbon
pool species. While the existence of the dual-cycle mechanism is
known in other catalyst systems as well, it is here revealed that
the alkene cycle is significantly more important for product forma-
tion in H-SAPO-5 than in previously studied large-pore catalysts of
higher acid strength. This is mainly based on two observed differ-
ences from previous experiments on H-Beta: The slower incorpora-
tion of 13C into the retained aromatics in H-SAPO-5 during
transient experiments, and the negligible abundance of isobutene
formation from polyMB intermediates in H-SAPO-5. Although not
confirmed to be the sole cause, we propose that the observed dif-
ferences are related to the low acid strength of H-SAPO-5 as com-
pared to H-Beta.
Furthermore, the co-reactions of benzene with 13C methanol
performed in this work provide us with mechanistic details for
the dealkylation of polyMB intermediates. The observed labelling
patterns show that ethene and propene can be formed by a mech-
anism involving ring contraction or expansion of the intermedi-
ates, the so-called paring mechanism.
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.
Mahsa Zokaie and INEOS ChlorVinyls are acknowledged for per-
forming TPA measurements on the H–SAPO–5 sample.
[42] O. Sekiguchi, V. Mayer, M.C. Letzel, D. Kuck, E. Uggerud, Eur. J. Mass Spectrom.
15 (2009) 167–181.
[43] J.E. Bercaw, P.L. Diaconescu, R.H. Grubbs, R.D. Kay, S. Kitching, J.A. Labinger, X.
Li, P. Mehrkhodavandi, G.E. Morris, G.J. Sunley, P. Vagner, J. Org. Chem. 71
(2006) 8907–8917.
[44] D.M. McCann, D. Lesthaeghe, P.W. Kletnieks, D.R. Guenther, M.J. Hayman, V.
Van Speybroeck, M. Waroquier, J.F. Haw, Angew. Chem. Int. Ed. 47 (2008)
5179–5182.
[45] B. Arstad, S. Kolboe, O. Swang, J. Phys. Chem. A 109 (2005) 8914–8922.
[46] B. Arstad, S. Kolboe, O. Swang, J. Phys. Org. Chem. 17 (2004) 1023–1032.
[47] D.S. Wragg, M.G. O’Brien, F.L. Bleken, M. Di Michiel, U. Olsbye, H. Fjellvåg,
Angew. Chem. Int. Ed. 51 (2012) 7956–7959.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.11.004.
References
[1] U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.V.W. Janssens, F. Joensen, S. Bordiga,
K.P. Lillerud, Angew. Chem. Int. Ed. 51 (2012) 5810–5831.