The mechanism of aromatic dealkylation in methanol-to-hydrocarbons conversion on H-ZSM-5: What are the aromatic precursors to light olefins?

Article abstract of DOI:10.1016/j.jcat.2013.11.003

Co-reactions of 7.5-9.3 kPa of DME with 4 kPa of toluene, p-xylene, and 4-ethyltoluene on H-ZSM-5 at 523-723 K at low conversions (<10 C%) with varying isotopic feed compositions of ¹³C/¹²C show that carbons originating from the aromatic ring are incorporated into ethene and propene. A comparison of the predicted ¹³C-contents of ethene and propene postulated on the basis of the paring, side-chain, and ring-expansion aromatic dealkylation mechanisms based on the experimentally observed isotopologue distribution of 1,2,4-trimethylbenzene, 1,2,4,5-tetramethylbenzene, and 4-ethyltoluene reveal that the predicted ¹³C-content of ethene and propene from 1,2,4,5-tetramethylbenzene via the paring mechanism most closely match the experimentally observed ¹³C-contents of ethene and propene (<10% mean relative error), compared to the other mechanisms and aromatic precursors examined. This work quantitatively shows that aromatic dealkylation to form ethene and propene occurs through the paring mechanism and that 1,2,4,5-tetraMB is the predominant aromatic precursor for light olefin formation for MTO conversion on H-ZSM-5 for a 200 K range in temperature.

Full text of DOI:10.1016/j.jcat.2013.11.003

Journal of Catalysis 311 (2014) 6–16
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
The mechanism of aromatic dealkylation in methanol-to-hydrocarbons
conversion on H-ZSM-5: What are the aromatic precursors to light
olefins?
⇑
Samia Ilias, Aditya Bhan
Department of Chemical Engineering and Materials Science, University of Minnesota, Twin Cities, 421 Washington Avenue SE, Minneapolis, MN 55455, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Co-reactions of 7.5–9.3 kPa of DME with 4 kPa of toluene, p-xylene, and 4-ethyltoluene on H-ZSM-5 at
523–723 K at low conversions (<10 C%) with varying isotopic feed compositions of 13C/12C show that car-
bons originating from the aromatic ring are incorporated into ethene and propene. A comparison of the
predicted 13C-contents of ethene and propene postulated on the basis of the paring, side-chain, and
ring-expansion aromatic dealkylation mechanisms based on the experimentally observed isotopologue
distribution of 1,2,4-trimethylbenzene, 1,2,4,5-tetramethylbenzene, and 4-ethyltoluene reveal that the
predicted 13C-content of ethene and propene from 1,2,4,5-tetramethylbenzene via the paring mechanism
most closely match the experimentally observed 13C-contents of ethene and propene (<10% mean relative
error), compared to the other mechanisms and aromatic precursors examined. This work quantitatively
shows that aromatic dealkylation to form ethene and propene occurs through the paring mechanism and
that 1,2,4,5-tetraMB is the predominant aromatic precursor for light olefin formation for MTO conversion
on H-ZSM-5 for a 200 K range in temperature.
Received 22 August 2013
Revised 23 October 2013
Accepted 6 November 2013
Keywords:
Methanol-to-olefins
Zeolite
Aromatic dealkylation
Paring mechanism
Side-chain mechanism
Ethylene
Propylene
Isotopic experiments
Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction
[11–15]. Isotopic experiments on H-ZSM-5, H-BEA, and H-MOR in
which methanol/dimethyl ether (DME) is co-reacted with aromat-
The conversion of methanol over acid zeolite catalysts to a wide
variety of hydrocarbons, including gasoline-range hydrocarbons
(methanol-to-gasoline, MTG) and light olefins (methanol-
to-olefins, MTO) has been extensively studied because methanol
can be produced from any carbon-based feedstock via a syngas
(CO + H2) intermediate [1–3]. Methanol-to-hydrocarbons (MTH)
proceeds through an indirect ‘‘hydrocarbon pool’’ mechanism in
which hydrocarbon formation occurs through repeated methyla-
tion and cracking of olefins and aromatics entrained within the
zeolite pores [4–8]. Isotopic switching experiments in which a
12C-methanol feed is switched with a 13C-methanol feed during
steady-state reaction on H-ZSM-5 at 623 K have shown that after
the switch the 13C-incorporation of ethene closely matched that
of polymethylbenzenes (polyMBs), while the 13C-incorporation of
C3 and higher olefins matched each other [9,10]. This result
showed that two different catalytic cycles are at work on H-ZSM-
5, one involving ethene and polyMBs and the other involving C3
and higher olefins. Similar work has also been done on other
zeolite and zeotype frameworks, either showing the prevalence
of both cycles or showing the dominance of one cycle over another
ics have also shown that ethene and propene contain carbon atoms
originating from the aromatic co-feed [16–18]. While these results
show that both ethene and propene are mechanistically linked to
polyMBs, the immediate precursor to ethene and propene, and
the mechanism of aromatic dealkylation to form light olefins is still
debated.
Experimental and theoretical work has postulated that light
olefin formation from polyMBs occurs through a paring, side-chain,
or ring-expansion mechanism (Schemes 1 and 2). The paring
mechanism, first proposed by Sullivan et al. [19] for the hydro-
cracking of hexaMB on a nickel sulfide catalyst, is initiated by
gem-methylation of a MB which is followed by ring contraction
to a 5-membered ring. As a result, an isopropyl or isobutyl group
is formed on the 5-membered ring, which can subsequently crack
to form propene or isobutene, containing one aromatic ring carbon.
Arstad et al. [20] proposed a ring-expansion mechanism for ethene
and propene formation from heptamethylbenzenium. Similar to
the paring mechanism, the ring-expansion mechanism is also
initiated by gem-methylation of a polyMB and the eliminated
olefin contains one aromatic ring carbon. Experimental work form
Bjørgen et al. [21] in which 13C-methanol is co-reacted with
12C-benzene on H-BEA reported that at 543 K at least 50% of pro-
pene and isobutane (coming from isobutene) molecules contained
⇑
Corresponding author.
E-mail addresses: ilias003@umn.edu (S. Ilias), abhan@umn.edu (A. Bhan).
0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.11.003

S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
7
propene. Work by Sassi et al. [24] at 723 K on H-BEA shows that
pulse reactions of 13C-methanol with various 12C-MBs result in
ethene and propene having similar 13C-contents as the overall
13C-content of the methanol and aromatic methyls in the feed,
suggesting that olefin formation at these conditions occurs via a
side-chain mechanism. Reactions of 13C-methanol with 12C- ethyl,
propyl, and butylbenzenes on H-BEA at 623 K produced high
fractions (>50%) of completely 12C-labeled ethene, propene, and
butenes, respectively [24,25]. While the high fraction of completely
12C-labeled olefins shows that the side-chain mechanism does
occur under conditions where aromatics with C2+ alkyl groups
are co-fed with methanol, it is unclear to what extent the side-
chain mechanism contributes to olefin formation under typical
MTH conditions where the majority of aromatics are methylbenz-
enes [13]. Work by Sassi et al. at temperatures above 623 K on
H-BEA suggests that the side-chain mechanism is occurring while
the results reported by Bjorgen et al. at temperatures below
573 K suggest that either the paring or ring-expansion mechanism
predominantly contributes to olefin formation. These results are
reconcilable if a change in the dealkylation mechanism occurs with
temperature.
Theoretical studies by McCann et al. [26] using ONIOM methods
on 46T clusters of H-ZSM-5 show that the highest barriers for iso-
butene formation from toluene are aromatic methylation steps
(Ea = 150–162 kJ molꢀ1). Lesthaeghe et al. [27] investigated the
side-chain mechanism for ethene formation from o-xylene on
46T clusters of H-ZSM-5 using ONIOM methods and found
that the ethene elimination step had the highest barrier
(Ea ꢁ 200 kJ molꢀ1). The high barrier to ethene elimination led
Lesthaeghe et al. [27] to hypothesize that side-chains may grow
longer to eliminate propene or may eventually cause deactivation
of the catalyst. Additionally, Seiler et al. [28] used in situ 13C
NMR for the continuous flow reaction of 13C-methanol followed
by a switch to 12C-methanol on H-ZSM-5 at 548–573 K and ob-
served that the peak at 10–30 ppm decreased by 40% after the
switch. This range in chemical shift was assigned to alkyl groups
on aromatics, and therefore, this 40% reduction in peak area was
attributed to olefin formation via the side-chain mechanism. Alter-
natively, the 10–30 ppm region can also be assigned to methyl
groups on aromatics [29], implying that the paring, side-chain,
and ring-expansion mechanism are all consistent with the peak
reduction following a switch to 12C-methanol. While theoretical
work suggests that barriers to aromatic dealkylation are lower
for the paring mechanism on H-ZSM-5, experimental evidence
has not been able to distinguish between the different aromatic
dealkylation mechanisms on this zeolite.
Understanding the mechanism of aromatic dealkylation to form
light olefins in MTH is key to understanding how ethene and
propene selectivity can be controlled. In this work, we use the
effluent isotopologue distributions of 1,2,4-triMB, 1,2,4,5-tetraMB,
and 4-ethyltoluene for the co-reactions of DME with toluene,
p-xylene, and 4-ethyltoluene, where either DME or the aromatic
co-feed is labeled with 13C-atoms, on H-ZSM-5 at 523–723 K to
predict the 13C-contents of ethene and propene based on the par-
ing, side-chain, and ring-expansion mechanisms. 1,2,4-TriMB,
1,2,4,5-tetraMB are examined as precursors to ethene and propene
formation because isotopic switching experiments on H-ZSM-5
have shown that the 13C-content of ethene matches that of 1,2,4-
triMB and 1,2,4,5-tetraMB, but not the 13C-content of penta- or
hexaMB [9,10]. Additional isotopic experiments involving metha-
nol/DME with aromatic co-feeds on H-ZSM-5 where either the
aromatic or the methylating agent is 13C-labeled show that a sig-
nificant fraction of these aromatics (<30%) are not direct products
of methylation of the aromatic co-feed and contain both 12C and
13C-atoms, leading to the conclusion that 1,2,4-triMB and 1,2,4,5-
tetraMB are rebuilt from repeated methylations and dealkylations
Scheme 1. Several mechanisms have been proposed for ethene and propene
formation from MBs. The paring mechanism for ethene (1e) and propene (1p)
formation from 1,2,4,5-tetraMB is shown in path 1. The side-chain mechanism for
ethene (2e) and propene (2p) formation from 1,2,4,5-tetraMB is shown in path 2.
Path 3e shows an alternative paring-type mechanism for ethene formation from
1,2,4,5-tetraMB that only requires protonation, not methylation; (–) aromatic ring
carbon, (s) aromatic methyl carbon, and ( ) methyl carbon from DME/methanol.
Scheme 2. Ring-expansion mechanism for ethene (4e) and propene (4p) formation;
(–) aromatic ring carbon, (s) aromatic methyl carbon, and ( ) methyl carbon from
DME/methanol.
one 12C-atom, showing that an aromatic ring carbon is incorpo-
rated into light olefins. Similar experiments by Westgård Erichsen
et al. [22] at 523 K on H-SAPO-5 showed that at least 60% of ethene
and propene molecules also contained one 13C-atom. Although the
incorporation of an aromatic ring carbon into light olefins shows
that on these large-pore catalysts at low temperatures, aromatic
dealkylation occurs via either
a paring or ring-expansion
mechanism, the experimental results reported by Bjørgen et al.
and Erichsen et al. did not distinguish between the two mecha-
nisms and the mechanism of aromatic dealkylation was not dis-
cernible for similar isotopic experiments on H-BEA at 603 K
[22,23].
The side-chain mechanism is also initiated by gem-methylation
of a MB, however, in this mechanism gem-methylation is followed
by deprotonation and results in the formation of an exo-cyclic dou-
ble bond. The exo-cyclic double bond can then undergo side-chain
methylation to form a side-chain that can crack to form ethene or

8
S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
on H-ZSM-5 [17,18]. 4-Ethyltoluene is examined as an aromatic
precursor because protonation of 4-ethyltoluene results in the
formation of an intermediate to aromatic dealkylation for the
side-chain mechanism. By comparing the predicted 13C-contents
of ethene and propene from the effluent isotopologue distributions
of three aromatic precursors for the different aromatic dealkylation
mechanisms that have previously been postulated to the experi-
mentally observed 13C-content of ethene and propene, both the
dominant aromatic precursor and dominant aromatic dealkylation
mechanism for ethene and propene formation was determined.
The predicted 13C-content of ethene and propene from 1,2,4,5-tet-
raMB via the paring mechanism mostly closely matches the exper-
imentally observed 13C-content of ethene and propene, showing
for the first time that aromatic dealkylation in MTH conversion
on H-ZSM-5 occurs through the paring mechanism and that
1,2,4,5-tetraMB is the dominant precursor to ethene and propene
formation at 523–723 K.
0.52
diphenyl, 95% methyl-siloxane capillary column (HP-5, 50.0 m ꢂ
320 m) connected to a mass spectrometer. The total
pressure of the reactor was 130 kPa. Similar reactions were also
performed at 623 K and 723 K, using 0.5 mg and 0.1 mg of catalyst,
respectively, diluted in approximately 150 mg of quartz sand. Iso-
topologue distributions were determined from mass fragmentation
patterns using the method outlined by Price and Iglesia [30].
lm) connected to a flame ionization detector and a 5%
l
m ꢂ 0.52
l
2.3. Predicting 13C-content of ethene and propene from aromatics
In Section 1, three different proposed mechanisms for aromatic
dealkylation in MTH were discussed and these mechanisms are
examined in this work. For both the side-chain mechanism and
the ring-expansion mechanisms, routes for ethene and propene
formation have been previously proposed [16,20,24]. For the par-
ing mechanism, however, only routes for propene and isobutene
formation have been proposed [19,21]. Arstad et al. also performed
calculations involving free carbocations investigating if a paring
mechanism depicted in 3e of Scheme 1 was possible, but deter-
mined that the secondary carbenium ion formed was not stable,
and therefore, this mechanism will not form ethene [31]. Recent
work from Westgård Erichsen et al. [22] on the co-reaction of
13C-methanol with 12C-benzene on H-SAPO-5 shows that 60% of
the ethene produced contains one ring carbon, which is consistent
with the ring-expansion mechanism or a paring-type mechanism.
We propose mechanism 1e (Scheme 1) to account for the possibil-
ity of ethene coming from a paring-type mechanism. In mechanism
1e, a methyl from the isopropyl group on the 5-membered ring
shifts to the ring accompanied by a simultaneous hydrogen shift
from the ring to the isopropyl group. Note that in this mechanism,
ethene should always contain one ring carbon, but the other car-
bon may come from either methyl carbon on the isopropyl group.
In this work, multiple dealkylation mechanisms for ethene and
propene formation are compared to determine which mechanism
best fits the experimentally observed 13C-content of ethene and
propene based on the effluent isotopic composition of 1,2,4-triMB,
1,2,4,5-tetraMB, and 4-ethyltoluene. The 13C-content of aromatic
ring carbons and methyl carbons must be known to correlate the
13C-content of these aromatic precursors to that of ethene and pro-
pene. To determine the 13C-content of the ring and methyl carbons,
we assume that methyl carbons on aromatics are the first to incor-
porate DME carbons. The limitations of this assumption for the iso-
topic results presented in this work are discussed further in
Section 3. For a particular MB with a total of t methyl groups, mi
is the number of 13C methyl carbons, and ri is the number of
13C ring carbons in a particular isotopologue of the MB with i
13C-atoms. For i 6 t, mi = t and ri = 0; for i > t, mi = t and ri = i ꢀ t. If
the overall fraction of the aromatic species with i 13C-atoms is xi,
then the total 13C-content of the methyl groups (13M) in the MB is
2. Materials and methods
2.1. Catalyst preparation
The catalyst, H-ZSM-5 (CBV8014), Si/Al = 42.6, was obtained in
the ammonium form from Zeolyst International. Structural and
chemical characterization of the commercial H-ZSM-5 sample used
in this study is reported in Section S.1 of the Supplemental
information. The silicon-to-aluminum ratio of the commercial
H-ZSM-5 sample was determined by ICP-OES elemental analysis
(performed by Galbraith Laboratories). The ammonium-form zeo-
lite was sieved to obtain aggregate particle sizes between 180
and 425 l
m (40–80 mesh) and treated in 1.67 cm3 sꢀ1 of dry air
(20–21% O2, <10 ppm H2O, Minneapolis Oxygen) at 773 K for 4 h
(heating rate of 0.0167 K sꢀ1) to convert it to the proton-form zeo-
lite. The catalyst was pre-treated in situ in 1.67 cm3 sꢀ1 of helium
flow (99.995% purity, Minneapolis Oxygen) at 773 K overnight
using a heating rate of 0.0167 K sꢀ1 prior to reaction.
2.2. Catalytic reactions of DME with aromatic co-feeds over H-ZSM-5
A stainless steel packed-bed reactor (0.25 in o.d.; 0.215 in i.d.)
equipped with a concentric thermal well (0.0625 in o.d.; 0.0485
in i.d.) aligned along the tube center was used for the conversion
of DME. The catalyst bed was supported between quartz wool
plugs and operated at isothermal conditions using an ARI heating
coil regulated by a Watlow Temperature Controller (96 Series).
Reactions were performed using 10 mg of catalyst mixed with
approximately 150 mg of quartz sand at 523 K with either
12C2- or 13C2-DME at 7.5–9.3 kPa (12C2-DME from Matheson Tri-
Gas, 99.5% purity; 13C2-DME from Aldrich Chemistry, 99% 13C-atom
purity; 0.024–0.030 cm3 sꢀ1) with 4 kPa of an aromatic co-feed
and a balance of helium so that the total feed flow rate was
0.417 cm3 sꢀ1. The aromatic co-feeds used were 12C8-p-xylene
(Sigma Aldrich, >99% purity), 13C2-p-xylene where the 13C-atoms
were on the methyl groups (Isotec, 99% 13C-atom purity), 12C7-tol-
uene (Sigma Aldrich, >99.9% purity), and 12C9-4-ethyltoluene
(Fluka Analytical, >95.0% purity). 13C NMR was performed on the
p-xylene sample from Isotec to confirm that the 13C-atoms were
located only on the methyl carbons (Section S.2 of the Supplemen-
tal information). The aromatic co-feeds were fed as a liquid using a
Cole Parmer EW-74900-00 syringe pump at rates of 0.61, 0.67, and
tþ6
X
13M ¼
mixi;
i¼0
and the total 13C-content of the ring carbons (13R) in the MB is
P
tþ6rixi
6
13R ¼
:
i¼0
This method is also used to calculate the 13C-content of the ring car-
bons and the exocyclic carbons for 4-ethyltoluene. For the specific
case of 4-ethyltoluene, 13M is the total 13C-content of methyl and
ethyl carbons.
0.72 l
L sꢀ1 for toluene, p-xylene, and 4-ethyltoluene, respectively.
Heat traced lines (423 K) were used to transfer the aromatic
co-feed to the reactor and the reactor effluent to a gas chromato-
graph–mass spectrometer (Agilent 7890-5975C) equipped with a
Based on the four mechanisms for ethene formation (1e, 2e, 3e,
and 4e) and the three mechanisms for propene formation (1p, 2p,
4p) shown in Schemes 1 and 2 and the effluent isotopologue
distributions for the three aromatic precursors, the 13C-contents
methyl-siloxane capillary column (HP-1, 50.0 m ꢂ 320 m ꢂ
l

S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
9
for ethene and propene can be predicted. The predicted 13C-con-
tent depends on both the mechanism and the aromatic precursor.
For example, 4-ethyltoluene if protonated has the same structure
as an intermediate of the side-chain mechanism and the ring-
expansion mechanism. Therefore, in both these mechanisms, the
predicted 13C-content of ethene from 4-ethyltoluene would simply
be 13M, the overall 13C-content of the exocyclic carbons in 4-ethyl-
toluene. In the ring-expansion mechanism, the 13C-content of eth-
ene will depend on which methyl in the gem-methyl group is
inserted into the 7-membered ring as well as if there is a methyl
group ortho to the gem-methyl group. The equations used for the
predicted 13C-contents of ethene and propene for each aromatic
precursor and mechanism examined in this work are summarized
in Table 1, where 13D is the fraction of 13C-atoms in the methylat-
ing agent (DME or methanol). The absolute error for the predicted
13C-content of the olefin formed (13Cpred) compared to the experi-
mentally observed 13C-content of the olefin formed (13Cobs) is
distributions are shown in Section S.4 of the Supplemental infor-
mation. The total carbon converted in these experiments was less
than 10 C% for all reactions (Table 2). For all the reactions per-
formed, at 1 min time-on stream, ethene, propene, and C8+ hydro-
carbons accounted for 90–96% of the products formed, indicating
that olefin methylation and cracking were suppressed (Table 2).
Unless otherwise specified, all results shown in this work were
collected at 1 min time-on-stream. While results at 1 min
time-on-stream may still be in the induction period for MTH, the
isotopologue distributions for the aromatic precursors examined
at this time-on-stream contain fewer DME carbons incorporated
into the aromatic ring compared to later times-on-stream. It is
important to minimize the incorporation of DME carbons into
the aromatic rings because it is unlikely that the incorporation of
DME and aromatic methyl carbons into the ring is random. The
models used for predictions of the 13C-contents of ethene and pro-
pene, however, assume that 13C-atoms in the aromatic ring follow
a statistical distribution, and therefore, unless otherwise specified,
only isotopic results at 1 min time-on stream are used.
ꢀ
ꢀ
13
13
ꢀ
ꢀ
C
ꢀ
C
;
obs
pred
and the relative error is
3.1. Co-reaction of 13C2-p-xylene and 13C-DME on H-ZSM-5
ꢀ
ꢀ
13
13
ꢀ
ꢀ
C
ꢀ
C
pred
obs
13
:
A major observable difference distinguishing the side-chain
mechanism from either the paring or ring-expansion mechanisms
for aromatic dealkylation reactions is that the paring and ring-
expansion mechanisms result in the incorporation of an aromatic
ring carbon into the olefin formed, whereas the side-chain mecha-
nism only incorporates aromatic methyl carbons and carbons from
the methylating agent (DME or methanol) into the olefin formed.
In the co-reaction of 13C-DME with 13C2-p-xylene, all methyl car-
bons in the feed are 13C-labeled and aromatic ring carbons in the
feed are 12C-atoms.
C
obs
3. Results and discussion
Isotopic experiments with feed compositions comprising of
4 kPa of toluene, p-xylene, or 4-ethyltoluene with 7.5–9.3 kPa of
13C-DME in which the number of 12C-methyls on the aromatic feed
was varied were conducted and are enumerated below:
(1) 13C2-p-xylene: 0 12C-aromatic methyls.
(2) 12C7-toluene: 1 12C-aromatic methyl.
(3) 12C8-p-xylene: 2 12C-aromatic methyls.
(4) 12C9-4-ethyltoluene: 1 12C-aromatic methyl and 1 12C-aro-
matic ethyl.
3.1.1. Isotopic composition of aromatics
Fig. 1c–f shows the isotopologue distribution of p-xylene, 1,2,4-
triMB, 4-ethyltoluene, and 1,2,4,5-tetraMB for the co-reaction of
13C-DME with 13C2-p-xylene at 523 K on H-ZSM-5. Nearly all of
the p-xylene in the effluent contains only two 13C-atoms, showing
that the p-xylene in the effluent is almost entirely unreacted feed
(Fig. 1c). Both 1,2,4-triMB and 1,2,4,5-tetraMB are predominantly
methylation products of 13C2-p-xylene, with 85.6% of 1,2,4-triMB
and 76.8% of 1,2,4,5-tetraMB containing only three and four
13C-atoms, respectively (Fig. 1d and f), suggesting that for the
majority of these polyMBs, the benzene ring is comprised of all
12C-atoms. The most abundant isotopologue of 4-ethyltoluene also
contains three 13C-atoms, showing that 4-ethyltoluene may form
Reactions 3 and 4 were also performed at higher temperatures
(623–723 K) to determine if the mechanism of aromatic dealkyla-
tion changes with temperature. Additionally, 13C2-p-xylene was
co-reacted with 12C-DME at 523 K so that the only 13C-methyls
were the two methyls on p-xylene. The isotopic compositions of
olefin and aromatic products of reactions (1) and (4) are discussed
in depth in Sections 3.1 and 3.2 and the time-on-stream depen-
dence for reaction (1) is shown in Section S.3 of the Supplemental
information. For the other reactions, the product isotopologue
from methylation of 13C2-p-xylene.
A significant fraction of
Table 1
Models used to predict the 13C-content of olefins formed from 1,2,4-triMB, 1,2,4,5-tetraMB, and 4-ethyltoluene via different aromatic dealkylation mechanisms.
Mechanism
Aromatic precursor
Olefin formed
Ethene (e)
Propene (p)
13M þ 13D þ 213
R
13M þ 13D þ 13R
Paring 1
All
4
3
13M þ 13D
13M þ 213
D
Side-chain 2
1,2,4-triMB and 1,2,4,5-tetraMB
2
3
13M
213M þ 13D
4-Ethyltoluene
All
3
–
13M þ 13R
Paring 3
2
13M þ 513D þ 1213
R
Ring expansion 4
1,2,4-triMB
3
13M þ 13D þ 213R
7
24
6
513M þ 313D þ 813
R
313M þ 13D þ 213
R
1,2,4,5-tetraMB
4-Ethyltoluene
16
6
213M þ 13D
3
13M

10
S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
Table 2
Conversion and product selectivity for co-reactions of 7.5–9.3 kPa of DME with 4 kPa of aromatic at 523–723 K.
Aromatic feed
DME feed label
13C2-p-xylene
13C
12C7-toluene
13C
12C8-p-xylene
13C
12C9-4-ethyltoluene
13C
13C
13C
12C
13C
Temperature (K)
Conv. (C%)a
523
2.7
623
2.1
723
4.4
523
2.9
523
10.0
523
3.6
523
4.1
723
8.4
Selectivity (C%)
Ethene
Propene
11.9
14.8
4.8
25.2
24.7
6.4
29.7
24.5
7.6
10.1
11.8
3.4
2.1
2.3
2.3
93.3
8.2
9.0
4.4
78.4
45.7
9.7
3.4
44.9
15.9
9.7
C4–C7 aliphatics
C7+ aromaticsb
68.6
43.7
38.2
74.4
41.2
29.4
a
Based on overall conversion of DME and the aromatic co-feed.
The aromatic co-feed present in the reaction effluent was not included in the assessment of product selectivity.
b
Fig. 1. Isotopologue distributions for the co-reaction of 13C-DME with of 13C2-p-xylene at 523 K for (a) ethene, (b) propene, (c) p-xylene, (d) 1,2,4-triMB, (e) 4-ethyltoluene,
and (f) 1,2,4,5-tetraMB.
4-ethyltoluene (65.8%), however, does not contain exactly three
13C-atoms, which shows that there are multiple routes to 4-ethyl-
toluene formation.
and 65.9% of propene contains only two 13C-atoms (Fig. 1b),
therefore, nearly 2/3 of both ethene and propene contain only
one aromatic ring carbon (12C), which is consistent with paring
mechanisms 1e, 3e, and with the ring-expansion mechanism 4e
in Schemes 1 and 2. While the incorporation of ring carbons into
ethene and propene shows that the side-chain mechanism is not
the dominant aromatic dealkylation mechanism under these
3.1.2. Isotopic composition of ethene and propene
For the co-reaction of 13C-DME with 13C2-p-xylene at 523 K on
H-ZSM-5, 65.3% of ethene contains only one 13C-atom (Fig. 1a)

S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
11
conditions, the paring and ring-expansion mechanisms cannot be
differentiated and the predominant aromatic precursor to ethene
and propene cannot be determined based solely on these results.
A quantitative assessment of both the dominant aromatic dealky-
lation mechanism and the predominant precursor to ethene and
propene formation is discussed in Section 3.3.
1,2,4,5-tetraMB. Additionally, 92.2% of 4-ethyltoluene contained
at least four 13C-atoms (Fig. 2e), showing that at higher tempera-
tures (ꢁ723 K), p-xylene methylation is not a major route for
4-ethyltoluene formation.
The majority of ethene and propene produced by the co-reac-
tion of 13C-DME with 13C2-p-xylene on H-ZSM-5 at 723 K contains
only one 12C-atom, originating from the aromatic ring of the
13C2-p-xylene feed (Fig. 2a and b). Compared to the isotopologue
distributions of ethene and propene and 523 K, there is a higher
fraction of all 13C ethene and propene at 723 K, most likely as a re-
sult of higher 13C content of the aromatics at 723 K compared to
523 K. The high fraction of ethene and propene containing only
one 12C-atom at both 523 K and 723 K shows that aromatic dealky-
lation on H-ZSM-5 when a methylbenzene is co-reacted with DME
results in the incorporation of an aromatic ring carbon into the
olefin formed.
3.1.3. Effect of temperature
The co-reaction of 13C-DME with 13C2-p-xylene on H-ZSM-5
was also performed at 723 K, and the isotopologue distributions
for light olefins and aromatics in the effluent of this reaction are
shown in Fig. 2. At these conditions, 89.7% of p-xylene in the
effluent contained two 13C-atoms, showing that the majority of
p-xylene in the effluent was unreacted feed (Fig. 2c). Similar to
the results at 523 K, a significant fraction of 1,2,4-triMB (55.7%)
and 1,2,4,5-tetraMB (43.5%) produced at these reaction conditions
are methylation products of the p-xylene co-feed, containing three
and four 13C-atoms, respectively (Fig. 2d and f). At 723 K, however,
there is a higher fraction of isotopologues of these polyMBs
containing more 13C-atoms than the number of methyl groups,
showing that at these reaction conditions, more 13C-atoms are
being incorporated into the aromatic ring of 1,2,4-triMB and
3.2. Co-reaction of 12C-4-ethyltoluene and 13C-DME on H-ZSM-5
4-Ethyltoluene was used as a co-feed to understand if and to
what extent the side-chain mechanism contributes to light olefin
production in MTH. The isotopologue distributions for light olefins
Fig. 2. Isotopologue distributions for the co-reaction of 13C-DME with of 13C2-p-xylene at 723 K for (a) ethene, (b) propene, (c) p-xylene, (d) 1,2,4-triMB, (e) 4-ethyltoluene,
and (f) 1,2,4,5-tetraMB.

12
S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
Fig. 3. Isotopologue distributions for the co-reaction of 13C-DME with of 12C-4-ethyltoluene at 523 K for (a) ethene, (b) propene, (c) p-xylene, (d) 1,2,4-triMB, (e) 4-
ethyltoluene, and (f) 1,2,4,5-tetraMB at (j) 1, ( ) 5, and (h) 10 min time-on-stream.
and aromatics for the reaction of 13C-DME and 12C-4-ethyltoluene
at 523 K are shown in Fig. 3 at 1, 5, and 10 min time-on-stream.
do not follow this same pattern and have a significant fraction of
isotopologues with 1–4 and 2–5 13C-atoms, respectively (Fig. 3c
and d). The isotopologue distribution of toluene is not shown in
Fig. 3 because the concentration of toluene in the reaction effluent
was too low to accurately quantify the isotopologue distribution.
Based on the results shown in Fig. 3c and d, it is unclear what
mechanistic pathways form p-xylene and 1,2,4-triMB here.
3.2.1. Isotopic composition at 1 min time-on-stream
Based on the side-chain mechanism (2e in Scheme 1), 12C-4-
ethyltoluene should dealkylate and form all 12C-ethene or the
side-chain may undergo methylation by 13C-DME and subse-
quently eliminate propene with only one 13C-atom. At 1 min
time-on-stream, 81.8% of ethene does have only 12C-atoms and
38.2% of propene contains only one 13C-atom (Fig. 3), which is con-
sistent with the side-chain mechanism. For propene, however, the
most abundant isotopologue contains two 13C-atoms, which would
be consistent with the paring mechanism assuming that incorpora-
tion of the 12C-atom in propene results from activation of the aro-
matic ring.
3.2.2. Isotopic compositions at 5 and 10 min time-on-stream
With increasing time-on-stream, the 13C-content of ethene,
propene, p-xylene, 1,2,4-triMB, and 1,2,4,5-tetraMB increases
(Fig. 3). The 13C-content of 4-ethyltoluene, however, remains
unchanged with increasing time-on-stream. The changing isotopo-
logue distributions with time-on-stream show that the methyl-
benzenes in the aromatic hydrocarbon pool as well as the light
olefins evolve with time-on-stream and become more 13C-rich. In
contrast, 90% of isotopologues of a methylation product of 4-ethyl-
toluene contained only one 13C-atom at all times-on-stream. This
product was determined to be a methylation product of 4-ethyltol-
uene based on its isotopologue distribution and because its
concentration increased significantly when 4-ethyltoluene was
co-fed with DME, compared to the co-reaction of DME with other
At 1 min time-on-stream, 4-ethyltoluene contains a natural
abundance of 1.1% 13C, showing that the 4-ethyltoluene in the
effluent is primarily unreacted feed (Fig. 3e). In contrast, 75.1% of
1,2,4,5-tetraMB at 1 min time-on-stream has three 13C-atoms, sug-
gesting that 4-ethyltoluene eliminates ethene to form toluene,
which subsequently is methylated three times by 13C-DME to form
13C3-1,2,4,5-tetraMB (Fig. 3f). However, p-xylene and 1,2,4-triMB

S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
13
methylbenzenes. The absence of 13C-atoms in 4-ethyltoluene and
its methylation product show that these aromatics are not re-
formed at a rate as fast as 1,2,4-triMB or 1,2,4,5-tetraMB, suggest-
ing that these ethylated aromatics are not co-catalytic. While the
side-chain mechanism can occur in the presence of a 4-ethyltolu-
ene, ethylated aromatics do not participate as co-catalysts in the
MTH hydrocarbon pool mechanism on H-ZSM-5.
13C-DME and 12C-4-ethyltoluene at both 523 K and 723 K suggest
that while the side-chain mechanism does occur in the presence
of ethylated aromatics, evidenced by the fraction of all 12C-labeled
ethene, a paring mechanism is also occurring in parallel, supported
by the high fraction of ethene and propene containing both 13C and
12C-atoms.
In the co-reaction of 4-ethyltoluene with DME at 523 K, 4-eth-
yltoluene initiates the aromatic hydrocarbon pool and dealkylation
of 4-ethyltoluene to form toluene, which is subsequently methyl-
ated, results in the formation of MBs. These MBs can continue to
react, undergoing repeated methylations and dealkylations and
thus increasing the 13C-content with increasing time-on-stream.
As a result, both ethene and propene, which are formed primarily
from MBs and not 4-ethyltoluene at later times-on-stream (5 and
10 min) also incorporate more 13C-atoms. The evolution of the
hydrocarbon pool for the co-reaction of 12C-4-ethyltoluene and
13C-DME at 523 K is depicted in Scheme 3.
3.3. Comparison of aromatic dealkylation mechanisms
Co-reactions of 13C-DME were also performed with 12C-toluene
at 523 K, 12C-p-xylene at 523 and 623 K, and 12C-DME with
13C2-p-xylene at 523 K (isotopologue distributions are shown in
Figs. S.5 and S.6 of the Supplemental information) in addition to
the co-reaction of 13C-DME with 13C2-p-xylene and 12C-4-ethyltol-
uene at 523 and 723 K on H-ZSM-5 discussed above. In this set of
reactions, the 13C/12C composition of methyls (coming from both
DME and the aromatic) in the feed is varied to distinguish between
the paring and ring-expansion mechanism.
Parity plots of the different dealkylation mechanisms compar-
ing the experimentally observed 13C-content of ethene to the pre-
dicted 13C-content of ethene from 1,2,4-triMB, 1,2,4,5-tetraMB, and
4-ethyltoluene from the different aromatic dealkylation mecha-
nisms are shown in Figs. 5 and S.12 of the Supplemental informa-
tion. Due to the change in dealkylation mechanism for ethene
formation when 4-ethyltoluene is used as a co-feed compared to
p-xylene as a co-feed (Section 3.2), the parity plots for ethene for-
mation only show reactions in which in toluene or p-xylene were
used co-feeds at temperatures ranging from 523 K to 723 K. While
the isotopic results for ethene presented in Section 3.1 showed that
nearly 2/3 of the ethene produced from the co-reaction of 13C-DME
with 13C2-p-xylene contained only one ring carbon (Fig. 1e), the
paring and ring-expansion mechanisms could not be distinguished
3.2.3. Effect of temperature
The co-reaction of 13C-DME and 12C-4-ethyltoluene on H-ZSM-5
was also performed at 723 K. Due to rapid deactivation of the
catalyst at these conditions (within 3 min time-on-stream), the
product isotopologue distributions are only shown at 1 min time-
on-stream (Fig. 4). For ethene, the isotopologues with zero and
one 13C-atoms have a similar fraction (41.4% and 35.1%, respec-
tively, Fig. 4a), which is consistent with both the side-chain and
paring mechanism. For propene, the most abundant isotopologue
has only one 12C-atom, consistent with the paring or ring-expan-
sion mechanism (Fig. 4b), showing that while the side-chain mech-
anism does occur when 4-ethyltoluene is co-reacted with DME to
form ethene, the paring or ring-expansion mechanism is also
occurring in parallel to form propene. The most abundant isotopo-
logue of p-xylene, 1,2,4-triMB, and 1,2,4,5-tetraMB contains 1, 2,
and 3 13C-atoms, respectively, indicating that these methylbenz-
enes may form from ethene elimination from 4-ethyltoluene to
form toluene, followed by sequential methylations of toluene by
13C-DME (Fig. 4c, d, and f). Due to the observed fast deactivation
of the catalyst at 723 K when 4-ethyltoluene is co-reacted with
DME, it is likely that 4-ethyltoluene is a precursor to coke on
H-ZSM-5, which is in agreement with results from Schulz [32],
showing that ethylated and propylated benzenes were the pre-
dominant aromatics retained inside the deactivated H-ZSM-5 cat-
alyst after the reaction of methanol at 543 K based on GC–MS
analysis of organics extracted from the zeolite by dissolution in
HF after reaction. Results in our work for the co-reaction of
from each other based on those results. By varying the 13C/12
C
compositions of methyls (coming from both DME and the aro-
matic) and predicting the 13C-content of ethene based on the four
dealkylation mechanisms shown in Schemes 1 and 2, the parity
plot in Fig. 5 and the mean absolute and relative errors shown in
Table 3 show that the predicted 13C-content of ethene coming from
1,2,4,5-tetraMB via the paring 1e mechanism is consistent with the
observed 13C-content of ethene and results in the lowest mean
absolute and relative errors (Table 3) compared to the other aro-
matic dealkylation mechanisms and other aromatic precursors
examined in this work. Both the mean absolute and relative error
for the predicted 13C-content of ethene coming from 4-ethyltolu-
ene via the paring 1e mechanism are lower compared to the pre-
dicted 13C-content of ethene from 4-ethyltoluene via other
aromatic dealkylation mechanisms, showing that at these reaction
conditions, 4-ethyltoluene most likely dealkylates via the paring 1e
mechanism (Scheme 1), and not the side-chain 2e mechanism
(Scheme 2). Comparing the predicted 13C-content of ethene from
the paring 1e mechanism, the mean absolute and relative error
for 1,2,4,5-tetraMB is lower than that predicted when considering
1,2,4-triMB and 4-ethyltoluene as the predominant precursors to
ethene, showing that while other aromatics may contribute to eth-
ene formation, 1,2,4,5-tetraMB is the predominant precursor to
ethene in MTH conversion on H-ZSM-5.
Parity plots for the three different mechanisms for propene
formation are shown in Figs. 5 and S.13 of the Supplemental
information. These parity plots include experimentally observed
13C-contents for propene for the co-reaction of DME with toluene,
p-xylene, and 4-ethyltoluene between 523 and 723 K. For propene
formation from 1,2,4,5-tetraMB, the model predictions for the par-
ing 1p mechanism fit the best with the observed 13C-content of
propene (Fig. 5). Both the mean absolute and relative errors for
propene formation from the paring 1p mechanism are lower
compared to the side-chain 2p and ring-expansion 4p mechanisms
Scheme 3. Depiction of the evolution of the hydrocarbon pool with increasing
time-on-stream for the co-reaction of 12C-4-ethyltoluene with 13C-DME at 523 K on
H-ZSM-5; (
)
13C-atoms.

14
S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
Fig. 4. Isotopologue distributions for the co-reaction of 13C-DME with of 12C-4-ethyltoluene at 723 K for (a) ethene, (b) propene, (c) p-xylene, (d) 1,2,4-triMB, (e) 4-
ethyltoluene, and (f) 1,2,4,5-tetraMB.
the paring 1p mechanism, the mean absolute and relative error
for 1,2,4,5-tetraMB is lower than that of 1,2,4-triMB and 4-ethyltol-
uene (Table 4), leading us to postulate that 1,2,4,5-tetraMB is the
predominant aromatic precursor to propene formation.
The parity plots in Figs. 5 and S.12 and S.13 of the Supplemental
information quantitatively show that aromatic dealkylation in
MTH conversion on H-ZSM-5 occurs primarily through the paring
mechanism (1e and 1p in Scheme 1) and the paring mechanism
is the dominant mechanism for light olefin formation through aro-
matic dealkylation on H-ZSM-5 for a 200 K range in temperature.
Although the isotopic results in this work are only presented for
data collected at 1 min time-on-stream, we expect that there
Fig. 5. Parity plots for the predicted vs. experimentally observed 13C-content for
should not be any significant change in the mechanism of aromatic
dealkylation with increasing time-on-stream because tetraMBs are
ethene (squares) and propene (circles) from 1,2,4,5-tetraMB via the paring 1e and
1p mechanisms, respectively, for the co-reactions of (d) 12C-DME with
still retained within the zeolite pores even after 14 days time-on-
stream for the reaction of methanol on H-ZSM-5 at 643 K [9].
The results in this work show that ethene is formed from one aro-
matic ring carbon and the other carbon has a 50% chance of coming
from the methylating agent and a 50% chance of coming from an
aromatic methyl carbon. Propene is formed from one aromatic ring
carbon, one aromatic methyl carbon, and one carbon coming from
the methylating agent. Aromatic dealkylation studies for MTO
13C2-p-xylene, (
) )
13C-DME with 13C2-p-xylene, ( 13C-DME with 12C-toluene,
(
(
)
)
13C-DME with 12C-p-xylene at 523 K, (
13C-DME with 13C2-p-xylene at 723 K, (
)
13C-DME with 12C-p-xylene at 623 K,
13C-DME with 12C-4-ethyltoluene at
)
523 K, and (
)
13C-DME with 12C-4-ethyltoluene at 723 K.
(Table 4). This result is in agreement with ethene formation occur-
ring through paring 1e mechanism, as both the paring 1e and 1p
mechanisms share common intermediates. Similar to ethene, for

S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
15
Table 3
Absolute and relative error for the model predictions of the 13C-content of ethene formed from various aromatic precursors via the different aromatic dealkylation mechanisms
compared to the experimentally observed 13C-content of ethene when varying isotopic compositions of DME are co-reacted with p-xylene or toluene at 523–723 K. Units of
absolute and relative error are percentage points of 13C and%, respectively.
Mean absolute error
124triMB
Mean relative error
124triMB
1245tetraMB
4ET
1245tetraMB
4ET
15.6
110.9
45.0
Paring 1e
Side-chain 2e
Paring 3e
8.0
30.1
16.8
9.5
4.0
33.3
9.1
5.8
27.8
11.2
27.8
21.9
87.6
61.2
28.5
8.9
82.1
34.3
14.0
Ring exp. 4e
4.8
110.9
Table 4
Absolute and relative error for the model predictions of the 13C-content of propene formed from various aromatic precursors via the different aromatic dealkylation mechanisms
compared to the experimentally observed 13C-content of propene when varying isotopic compositions of DME are co-reacted with p-xylene, toluene, or 4-ethyltoluene at 523–
723 K. Units of absolute and relative error are percentage points of 13C and%, respectively.
Mean absolute error
124triMB
Mean relative error
124triMB
1245tetraMB
4ET
1245tetraMB
4ET
Paring 1e
Side-chain 2e
Ring exp. 4e
6.9
25.5
13.4
3.8
26.9
6.9
10.3
23.1
23.1
15.5
47.2
44.3
6.6
45.2
27.6
19.2
52.1
52.1
conversion on large-pore materials H-BEA and H-SAPO-5 have also
shown that aromatic ring carbons are incorporated into light ole-
fins. The methods used in this study can also be applied to these
catalysts to examine if the aromatic dealkylation mechanism
changes with zeolite structure – for example, does a larger catalyst
pore-size allow for the space-demanding ring-expansion mecha-
nism to operate?
To determine the mechanism of aromatic dealkylation, five
12C/13C feed compositions were used so that incorporation of
DME methyls, aromatic methyls, and aromatic ring carbons into
ethene and propene could be differentiated. Mechanism discrimi-
nation through the methods reported in work is applicable to
any reaction system in which a single product species maybe be
formed through multiple mechanisms or pathways as long the iso-
topic composition of the reactants can be varied in such way to al-
low for mechanism differentiation. Using these methods, we
identify the mechanism and the predominant precursors to light
olefin synthesis in methanol-to-hydrocarbons conversion and pos-
tulate that experimental conditions or co-feeds that enhance the
resulting concentration of tri- and tetra-methylbenzenes in
H-ZSM-5 will result in a concomitant increase in the selectivity
to ethylene and propylene.
Effluent isotopologue distributions of 1,2,4-triMB, 1,2,4,5-tet-
raMB, and 4-ethyltoluene were used to predict the 13C-content of
ethene and propene for co-reactions of DME with toluene,
p-xylene, and 4-ethyltoluene where either DME or the aromatic
was 13C-labeled. Through the use of quantitative models presented
in Table 1 to predict the 13C-content of ethene and propene from
various aromatic precursors, the mechanism of aromatic dealkyla-
tion in MTO conversion on H-ZSM-5 was determined. The results
shown in Tables 3 and 4 quantitatively show that the paring
mechanism for light olefin formation is occurring and that
1,2,4,5-tetraMB is the predominant precursor to ethene and
propene formation for MTO conversion on H-ZSM-5 over a 200 K
temperature range. The paring 1e and 1p mechanisms (Scheme 1)
for ethene and propene formation can predict the 13C-content of
ethene and propene from the effluent isotopic composition
of 1,2,4,5-tetraMB with greater accuracy than either the side-chain
or ring-expansion mechanisms for the reaction of DME with three
different aromatic co-feeds, five different 13C/12C feed composi-
tions and a 200 K range in temperature, showing that the paring
mechanism is the dominant mechanism of aromatic dealkylation
for MTO conversion on H-ZSM-5.
Acknowledgments
4. Conclusions
The authors acknowledge financial support from The Dow
Chemical Company and the National Science Foundation (CBET
1055846). The authors also acknowledge Mr. Rachit Khare for help-
ful technical discussions.
Co-reactions DME with toluene, p-xylene, and 4-ethyltoluene
with varying 13C/12C feed compositions were performed to deter-
mine the mechanism of aromatic dealkylation in MTH conversion
on H-ZSM-5. The co-reaction of 13C-DME and 13C2-p-xylene at
523 K and 723 K resulted in a high fraction of ethene and propene
containing only one 12C-atom, originating from the aromatic ring
carbon of the 13C2-p-xylene feed, showing that the mechanism of
aromatic dealkylation on H-ZSM-5 results in the incorporation of
an aromatic ring carbon into light olefins, which is consistent with
both the paring and ring-expansion mechanisms. For the co-reac-
tion of 13C-DME and 13C-4-ethyltoluene at 523 K and 723 K, the
high fraction of ethene with only 12C-atoms suggests that ethene
can be formed via the side-chain mechanism, however, both 4-eth-
yltoluene and its methylation product incorporated negligible
amounts of 13C-atoms, indicating that while these species can un-
dergo side-chain dealkylation, they are not co-catalytic in that they
are not reformed fast enough to participate in the aromatic hydro-
carbon pool.
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.2013.11.003.
References
[1] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3–48.
[2] 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.
[3] S. Ilias, A. Bhan, ACS Catal. 3 (2013) 18–31.
[4] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003) 317–
326.
[5] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304–309.
[6] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458–464.
[7] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329–336.

16
S. Ilias, A. Bhan / Journal of Catalysis 311 (2014) 6–16
[8] W. Song, J.F. Haw, J.B. Nicholas, C.S. Heneghan, J. Am. Chem. Soc. 122 (2000)
10726–10727.
[19] R.F. Sullivan, C.J. Egan, G.E. Langlois, R.P. Sieg, J. Am. Chem. Soc. 83 (1961)
1156–1160.
[9] M. Bjørgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 249 (2007) 195–207.
[10] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.-P. Lillerud, S. Kolboe, M. Bjørgen, J.
Am. Chem. Soc. 128 (2006) 14770–14771.
[11] S. Teketel, U. Olsbye, K.-P. Lillerud, P. Beato, S. Svelle, Microporous Mesoporous
Mater. 136 (2010) 33–41.
[20] B. Arstad, S. Kolboe, O. Swang, J. Phys. Chem. A 109 (2005) 8914–8922.
[21] M. Bjørgen, U. Olsbye, D. Petersen, S. Kolboe, J. Catal. 221 (2004) 1–10.
[22] M. Westgård Erichsen, S. Svelle, U. Olsbye, J. Catal. 298 (2013) 94–101.
[23] M. Bjørgen, U. Olsbye, S. Svelle, S. Kolboe, Catal. Lett. 93 (2004) 37–40.
[24] A. Sassi, M.A. Wildman, H.J. Ahn, P. Prasad, J.B. Nicholas, J.F. Haw, J. Phys. Chem.
B 106 (2002) 2294–2303.
[12] 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.
[13] M. Bjørgen, F. Joensen, K.-P. Lillerud, U. Olsbye, S. Svelle, Catal. Today 142
(2009) 90–97.
[25] A. Sassi, M.A. Wildman, J.F. Haw, J. Phys. Chem. B 106 (2002) 8768–8773.
[26] 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.
[14] S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, J. Phys. Chem. C 111 (2007) 17981–
17984.
[27] D. Lesthaeghe, A. Horré, M. Waroquier, G.B. Marin, V. Van Speybroeck, Chem.
Eur. J. 15 (2009) 10803–10808.
[15] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137–8138.
[16] T. Mole, G. Bett, D. Seddon, J. Catal. 84 (1983) 435–445.
[17] O. Mikkelsen, P.O. Ronning, S. Kolboe, Microporous Mesoporous Mater. 40
(2000) 95–113.
[28] M. Seiler, W. Wang, A. Buchholz, M. Hunger, Catal. Lett. 88 (2003) 187–191.
[29] W. Song, H. Fu, J.F. Haw, J. Am. Chem. Soc. 123 (2001) 4749–4754.
[30] G.L. Price, E. Iglesia, Ind. Eng. Chem. Res. 28 (1989) 839–844.
[31] B. Arstad, S. Kolboe, O. Swang, J. Phys. Org. Chem. 17 (2004) 1023–1032.
[32] H. Schulz, Catal. Today 154 (2010) 183–194.
[18] S. Ilias, A. Bhan, J. Catal. 290 (2012) 186–192.