Methylbenzene hydrocarbon pool in methanol-to-olefins conversion over zeolite H-ZSM-5

Article abstract of DOI:10.1016/j.jcat.2015.10.001

The formation and reactivity of a methylbenzenes (MBs) hydrocarbon pool in the induction period of the methanol-to-olefins (MTO) reaction over zeolite H-ZSM-5 was investigated and the mechanistic link of MBs to ethene and propene was revealed. Time evolution analysis of the formed MBs and ¹²C/¹³C methanol-switching experiments indicate that in the induction period bulkier compounds such as tetraMB and pentaMB have higher reactivity than their lighter counterparts such as p/m-diMB and triMB. By correlating the distribution of MBs trapped on H-ZSM-5 with ethene and propene, we found that tetraMB and pentaMB favor the formation of propene, while p/m-diMB and triMB mainly contribute to the formation of ethene. On the basis of this relationship, the olefin (ethene and propene) selectivity can be controlled by regulating the distribution of trapped MBs by varying the silicon-to-aluminum ratio of ZSM-5, reaction temperature, and space velocity. The reactivity of MBs and the correlation of MBs with olefins were also verified under steady-state conditions. By observation of key cyclopentenyl and pentamethylbenzenium cation intermediates using in situ solid-state NMR spectroscopy, a paring mechanism was proposed to link MBs with ethene and propene. P/M-diMB and triMB produce ethylcyclopentenyl cations followed by splitting off of ethene, while tetraMB and pentaMB generate propyl-attached intermediates, which eventually produce propene. This work provides new insight into the MBs hydrocarbon pool in MTO chemistry.

Full text of DOI:10.1016/j.jcat.2015.10.001

Journal of Catalysis 332 (2015) 127–137
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Methylbenzene hydrocarbon pool in methanol-to-olefins conversion
over zeolite H-ZSM-5
a
b
a
a
a
Chao Wang a, Jun Xu a, , Guodong Qi , Yanjun Gong , Weiyu Wang , Pan Gao , Qiang Wang ,
⇑
Ningdong Feng a, Xiaolong Liu a, Feng Deng a,
⇑
a State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, National Center for Magnetic Resonance in Wuhan, Wuhan Institute of Physics and
Mathematics, Chinese Academy of Sciences, Wuhan 430071, People’s Republic of China
b State Key Laboratory of Heavy Oil Processing, CNPC Key Laboratory of Catalysis, China University of Petroleum – Beijing, Beijing 102249, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
The formation and reactivity of a methylbenzenes (MBs) hydrocarbon pool in the induction period of the
methanol-to-olefins (MTO) reaction over zeolite H-ZSM-5 was investigated and the mechanistic link of
Received 17 July 2015
Revised 30 September 2015
Accepted 2 October 2015
MBs to ethene and propene was revealed. Time evolution analysis of the formed MBs and 12C/13
C
methanol-switching experiments indicate that in the induction period bulkier compounds such as
tetraMB and pentaMB have higher reactivity than their lighter counterparts such as p/m-diMB and
triMB. By correlating the distribution of MBs trapped on H-ZSM-5 with ethene and propene, we found
that tetraMB and pentaMB favor the formation of propene, while p/m-diMB and triMB mainly contribute
to the formation of ethene. On the basis of this relationship, the olefin (ethene and propene) selectivity
can be controlled by regulating the distribution of trapped MBs by varying the silicon-to-aluminum ratio
of ZSM-5, reaction temperature, and space velocity. The reactivity of MBs and the correlation of MBs with
olefins were also verified under steady-state conditions. By observation of key cyclopentenyl and
pentamethylbenzenium cation intermediates using in situ solid-state NMR spectroscopy, a paring mech-
anism was proposed to link MBs with ethene and propene. P/M-diMB and triMB produce ethylcyclopen-
tenyl cations followed by splitting off of ethene, while tetraMB and pentaMB generate propyl-attached
intermediates, which eventually produce propene. This work provides new insight into the MBs hydro-
carbon pool in MTO chemistry.
Keywords:
Olefins
Zeolites
Mechanism
Carbocations
NMR spectroscopy
Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction
reaction barrier is involved from the theoretical point of view
[10–12]. An indirect mechanism, known as the hydrocarbon pool
Olefins such as ethylene and propene that are mainly produced
from cracking of crude oil are building blocks for producing
polyethylene and polypropylene. The methanol-to-olefins (MTO)
process based on acidic zeolite catalysts is becoming a promising
alternative to the oil route. The prominent advantage of MTO is
that the feed methanol can be obtained from a wide range of
sources, such as coal, natural gas, and biomass [1–4]. Mechanistic
understanding of the MTO reaction is essential in achieving selec-
tivity control of specific olefins, which, however, is a big challenge.
Although intensive experimental and theoretical effort has been
dedicated to this issue, the exact mechanism underlying methanol
conversion has remained poorly understood over the past decades
[4–9].
(HP), has been widely accepted for olefin formation [3,4,8–10].
The HP mechanism that was introduced by Dahl and Kolboe
describes a catalytic reaction center (HP species) formed in zeolite,
on which repeated C–C bond forming and breaking occurs to pro-
duce olefin products [13–15]. Polymethylbenzens (MBs) and cyclic
carbenium ions were identified as active HP species on zeolites
(ZSM-5, MOR, and Beta) and zeotype materials (SAPO-34 and
SAPO-5) [2,3,16–23]. On the basis of HP species, two different
routes have been proposed to explain the operation of the HP
mechanism, namely, the side-chain methylation route [16,17,24–26]
and the paring route [27,28]. A dual-cycle mechanism was recently
claimed on H-ZSM-5 with ethene being related to light MBs and
C+3 olefins to propene [29], and thus both MBs and olefins could
serve as HP species.
Direct formation of C–C bonds by C1 entities was initially con-
sidered, but is still debated, because a prohibitively high-energy
The co-reaction of methanol and/or dimethyl ether (DME) with
aromatics on H-ZSM-5, H-MOR, and H-Beta has shown that ethene
and propene can be related to MBs [23,30,31]. This is evidenced by
isotopic scrambling experiments, which reveal that some carbons
⇑
Corresponding authors.
E-mail addresses: xujun@wipm.ac.cn (J. Xu), dengf@wipm.ac.cn (F. Deng).
http://dx.doi.org/10.1016/j.jcat.2015.10.001
0021-9517/Ó 2015 Elsevier Inc. All rights reserved.

128
C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
of ethene and propene originate from co-fed aromatics. Ilias and
Bhan further showed that MBs such as tetraMB could be precursors
to ethene and propene under steady-state conditions [31]. Another
important catalyst, SAPO-34, with small pore size and large cage
structure, does not permit the entrance of aromatic reactants,
which can be only formed in place in the MTO reaction. On the
basis of 13C isotopic analysis, Arstad and Kolboe demonstrated that
pentaMB and hexaMB in SAPO-34 were active species responsible
for the formation of ethene and propene [32].
Theoretical calculations and in situ NMR experiments provide
more insight into the exact route to light olefins from MBs. The
DFT calculations of McCann et al. showed that an aromatics-
based paring cycle operates on H-ZSM-5, where methylation of
aromatics produces C5 and C6 cyclic carbocation intermediates
and isobutene is generated by the ring contraction of the C6 carbo-
cation [33]. Using a combined in situ solid-state NMR, GC–MS, and
DFT calculation approaches, we recently correlated olefins with the
HP species on ZSM-5 [34] and revealed an aromatic-base
paring route for the formation of propene, in which MBs are
mechanistically linked to propene via pentamethylbenzenium
and 1,3,4-trimethylcyclopentenyl cations. The formation of ethene
on H-ZSM-5 from aromatics was theoretically studied by
Lesthaeghe et al. [35]. A side-chain route was proposed, but the
higher barrier of 200 kJ/mol for the ethene elimination step makes
this route unfavorable. Most recently, we identified ethylcyclopen-
tenyl carbocations as intermediates in the aromatics-based cycle
on ZSM-5 and demonstrated that the elimination of ethyl groups
from the ethylcyclopentenyl cations provides a viable route for
ethene formation [36].
Among the produced hydrocarbons, a wide range of aromatics
from benzene to hexamethylbenzenes are readily formed on
H-ZSM-5 in the typical MTO reaction. Although previous results
demonstrated the correlation between light olefins and MBs [16,
17,29–31,37–39], the mechanistic link of specific MBs to product
selectivity is not established. It is also noted that the MTO reaction
is well studied under steady-state conditions when olefins are
steadily produced and complex secondary reactions consisting of
methylation, alkylation, oligomerization, cracking, etc. are prevail-
ing [5–7,40,41]. The secondary reactions bring about considerable
difficulty in distinguishing the exact role of MBs from other com-
pounds in the formation of light olefins. We previously found the
formation of MBs and cyclic carbocations on H-ZSM-5 [34]. Impor-
tantly, at lower temperatures, these HP species were observed
prior to the formation of olefins and the boom of secondary reac-
tions. Since the induction period is of vital importance in MTO
reactions that are associated with the formation of the first C–C
bonds in olefin products, investigation of MBs in the induction per-
iod in addition to the steady-state period would allow us to gain
more insight into the MTO chemistry.
heating treatment in all samples (Fig. S1 in the Supplementary
Material). The Brønsted acid sites concentrations are 860, 130,
and 65 lmol/g on H-ZSM-5 (15), H-ZSM-5 (100), and H-ZSM-5
(200), respectively, as determined by 1H solid state NMR. The
H-ZSM-5 zeolites were pressed into pellets between 60 and 80
mesh and the pellets (0.2 g) were activated at 400 °C in flowing
helium for 1 h prior to the reactions. The pulse-quench reactions
were carried out in a pulse-quench reactor [42] for a preset period
using a carrier gas (He) flow of 500 mL/min; then they were
thermally quenched by pulsing liquid nitrogen onto the catalyst
bed, using high-speed valves controlled by a GC computer (<1 s).
Typically, 490 lmol (20 lL) methanol was pulsed into the reactor
for each pulse reaction. For the continuous-flow reaction, methanol
with a weight hourly space velocity (WHSV) of 2 hꢀ1 was reacted
over the H-ZSM-5 (0.2 g) pellets in a fixed-bed reactor.
In 12C/13C methanol isotope transient experiments, methanol in
13C natural abundance was reacted on H-ZSM-5 for a predeter-
mined period, followed by switching to 13C-methanol (99% 13C),
and was allowed to react for a certain time. The evolution of the
13C components in the products was determined by GC–MS
analysis.
2.2. Analysis of trapped species
The catalyst with trapped products was dissolved in 20 wt.% HF
solution and then extracted with CH2Cl2. The bottom layer contain-
ing the organic phase of the extracted solution was separated and
analyzed by gas chromatography (GC). In addition, the catalyst
with trapped products was also directly analyzed by solid-state
NMR spectroscopy (see the following).
2.3. Gas chromatography
The effluent was analyzed quantitatively by online GC
(Shimadzu GC-2010 plus) equipped with a flame ionization detec-
tor and a Petrocol DH 100 fused silica capillary column (100 m,
0.25 mm i.d., 0.5 lm film thickness). The temperature program-
ming started at 50 °C (maintained for 15 min), followed by a rate
of 15 °C minꢀ1 to a final temperature of 200 °C. The isotopic
compositions of the trapped species were analyzed by GC–MS
(Shimadzu GCMS-QP2010) with the same capillary column as used
for GC analysis. The following temperature programming was
applied: maintained at an initial temperature of 50 °C for 1 min,
followed by a rate of 10 °C minꢀ1 to a final temperature of
250 °C, maintained for 10 min.
In this work, the reactivity of MBs produced in the induction
period of MTO reactions over H-ZSM-5 was studied in detail. We
found that the initially formed MBs were intrinsically correlated
with light olefins. The specific link of lighter MBs (p/m-diMB and
triMB) with ethene and bulkier MBs (tetraMB and pentaMB) with
propene were identified. Moreover, the mechanism underlying
the relation between MBs and ethene and propene was discussed.
2.4. NMR experiments
After the reaction was quenched, the pulse-quench reactor con-
taining the catalyst was sealed. The sealed reactor was then trans-
ferred to a glove box filled with pure N2 and the catalyst was
packed into an NMR rotor for NMR measurements.
All solid-state NMR spectroscopy experiments were carried out
at 11.7 T on a Bruker-Avance III-500 spectrometer, equipped with a
4 mm probe, with resonance frequencies of 500.57 and
125.87 MHz for 1H and 13C, respectively. Single-pulse 13C MAS
2. Experimental
experiments with 1H decoupling were performed using a
p
/2 13C
2.1. Materials and catalysis experiments
pulse width of 3.9 ls and a repetition time of 5 s. The magic angle
spinning rate was set to 10.6–12.5 kHz. For the 1H ? 13C CP/MAS
NMR experiments, the Hartmann–Hahn condition was achieved
using hexamethylbenzene (HMB), with a contact time of 4 ms
and a repetition time of 2 s. The 13C chemical shifts were refer-
enced to HMB (a second reference to TMS).
Before the MTO reactions, the H-ZSM-5 (Si/Al = 15, 100, and
200) zeolites were calcined at 550 °C under flow air for 5 h to
remove organic impurities that would affect the MTO reaction.
GC analysis shows that no organic species are present after the

C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
129
Table 1
3. Results and discussion
Product selectivity (C%) and methanol conversion (%) obtained from one pulse
reaction of methanol on H-ZSM-5 at 350 °C. The selectivity of MBs is indicated in
parentheses.
3.1. Formation of methylbenzenes on H-ZSM-5
RT (s)a
Con.
0
P/Eb
C2
C3
–
C4
C5
C6+
To monitor the formation of MBs and primary olefins in the ini-
tial stage, the MTO reactions were carried out on H-ZSM-5 (15) at a
relatively lower temperature (350 °C) and quenched by liquid
nitrogen within a short period (<1 s). Moreover, a high feed rate
(500 mL/min) of methanol was utilized to suppress the secondary
reactions. Fig. 1 shows the GC chromatogram of the effluent prod-
ucts obtained from methanol reacting for 2–64 s at 350 °C. When
the reaction time is very short (<4 s), the conversion is only ca.
5% and a considerable fraction of MBs (ca. 96%), mainly toluene
and p/m-diMB, is produced, whereas the light olefins such as
ethene and propene are negligible (Table 1). After 8 s, the forma-
tion of olefins including C+3 olefins is prevailing while the content
of MBs is decreasing, indicating that the consumption of MBs gen-
erates olefins. This provides evidence that the MBs serve as precur-
sors to the light olefins [31,34]. The secondary reaction is
prevailing at longer reaction times, reflected by the increase in
MBs after 16 s. This is due to the preferential transformation of
the formed C+3 olefins and C5–C10 hydrocarbons on the acid sites
of H-ZSM-5. The lower temperature and short reaction time allows
the observation of MBs evolution in relation to olefins. In compar-
ison, the reaction at a higher temperature of 400 °C produces a con-
siderable amount of light olefins at a conversion of 74% even
within 2 s of reaction time (Fig. S2 and Table S1). The sum of
ethene and propene is over 50%, but the MBs are very low (ca.
1%). At a similar conversion (ca. 76%), the selectivity to MBs
(2.23%) at 350 °C is almost twice that (1.26%) at 400 °C. This is
likely due to the enhanced reactivity of MBs at the higher temper-
ature. In addition, the enhanced formation of light olefins could
also be associated with the secondary reactions at higher temper-
ature. The cracking of the formed longer alkenes such as C+3 com-
pounds (Tables 1 and S1) into ethene and propene with high
activation energy could be prevalent at higher temperatures. It
should be noted that the remarkably reduced induction period at
2
4
8
16
32
64
–
–
–
–
–
–
– (–)
4.72
0.72
2.44
2.80
3.49
5.42
5.55
1.77
95.79 (95.79)
43.57 (39.54)
18.97 (2.23)
32.15 (11.56)
57.29 (48.56)
37.55
76.32
89.54
97.31
11.01
12.95
6.52
30.84
45.18
35.33
30.15
8.04
18.93
15.30
4.40
14.74
13.43
11.81
2.61
5.43
a
Reaction time after one methanol pulse.
Propene-to-ethene ratio.
b
higher temperature leads to a rapid conversion of formed olefins
and hydrocarbons back into MBs along with the reaction time
being increased (MBs amount to ca. 67% at 64 s).
The reactive MBs in the very early period would have a great
impact on olefin formation. As is well known in various zeolite
and zeotype catalysts, the interaction of MBs with the zeolite
framework would generate a catalytic center of HP in the MTO
reaction [20,43–46]. Thus, the analysis of organic species trapped
inside H-ZSM-5 will give more insight into the role of MBs in the
formation of olefins.
3.2. Stability and reactivity of trapped methylbenzenes
The trapped organic compounds on H-ZSM-5(15) were obtained
by thermally quenching the methanol pulse reactions rapidly at a
preset reaction time. Since the MBs show higher reactivity at
higher temperature, the analysis was performed on the trapped
species obtained at 400 °C. Fig. 2a shows the GC–MS data of the
extract of trapped species formed on H-ZSM-5(15). MBs ranging
from p/m-diMB to pentaMB are found in the very early stage
(2 s). Differently from what is observed in the effluent, bulkier spe-
cies with more methyl substituents dominate the trapped prod-
ucts, particularly before 16 s. It is noteworthy that hexaMB is not
present in our trapped products, whereas it was abundant under
the steady-state conditions [29]. The previous work of Svelle and
co-workers indicated that hexaMB formed on ZSM-5 was virtually
unreactive in the MTO reaction [29]. Additionally, hydrocarbons
larger than hexaMB, which were considered as coke species, are
not formed on our catalysts. Therefore, our experimental results
indicate that the formation of MBs is dynamically favored in the
induction period and the large and unreactive molecules would
be accumulated over a long reaction time.
The evolution of trapped MBs was further analyzed from 2 to
32 s of reaction time. As shown in Fig. 2b, the normalized pentaMB
reaches a maximum at 4 s, followed by a rapid decrease. In com-
parison, 1,2,3,5-tetraMB exhibits a relatively slow increase and
decline rate. The light MBs, such as p/m-diMB and 1,2,4-triMB,
share a similar trend, increasing and decreasing at a much lower
rate. This allows us to give a stability order for the observed
MBs: pentaMB < 1,2,3,5-tetraMB < 1,2,4-triMB ꢁ p/m-diMB. The
stability of MBs can reflect to some extent the reactivity, which
is related to their formation, transformation, and diffusion in zeo-
lite channels. For a qualitative analysis of the reactivity, the forma-
tion, transformation, and diffusion rates of MB with i methyl
groups are defined by vifor vtira and vdi iff , respectively. Thus the
,
change in concentration of trapped MBs with time can be
described by the formula
dni
dt
¼
vifor
ꢀ
vitra
ꢀ
vidiff
ð1Þ
where ni denotes the amount of MB with i methyl groups. The MBs
having higher reactivity can be more easily transformed into other
Fig. 1. GC analysis of effluent products obtained from methanol pulse reaction over
H-ZSM-5(15) at 350 °C for various times.

130
C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
the sum of transformation and diffusion rates, the rate of change
in MB concentration is zero, indicating that the concentration of
trapped species reaches a maximum. For pentaMB, the diffusion
rate is nearly zero; thus the concentration changing rate can be
dn5
dt
described by
¼
m5for
ꢀ
m5tra. The negative growth of residual pen-
taMBs after 4 s indicates that the transformation rate is higher than
the formation rate (Fig. 2b). This leads to the conclusion that pen-
taMB is an active species and could be transformed into other
hydrocarbons, at least in the initial period. A similar trend is found
for 1,2,3,5-tetraMB, but with relatively lower rates of increase and
decrease. Since the diffusion rate of 1,2,3,5-tetraMB can be consid-
ered to be zero, it can be concluded that 1,2,3,5-tetraMB is an active
species as well, but its reactivity is lower than that of pentaMB.
P/m-diMB and 1,2,4-triMB exhibits almost the same trend.
Although easy diffusion will contribute to the reduction of these
two compounds in the zeolite channels, the decrease could be due
to some extent to the transformation of these two compounds.
Indeed, the reactivity of p/m-diMB and triMB has been noted in
the MTO reaction over H-ZSM-5 and related to the formation of
light olefins such as ethene under steady-state conditions [21,37].
It has been established in the MTO reaction that the reactivity of
organic species trapped in the zeolite voids can be assessed by a
transient 12C/13C methanol-switching experiment [29,47]. To gain
further insight into the reactivity of MBs formed in MTO reaction,
a series of 12C/13C methanol-switching experiments were con-
ducted. Methanol in natural abundance was reacted on H-ZSM-5
(15) for 32 s at 400 °C. After the 12C methanol was completely con-
verted, the reaction was rapidly switched to 13C methanol and
allowed to proceed for another 4 s. Analysis of the incorporation
of 13C atoms into the MBs would reveal the reactivity: more incor-
porated 13C atoms indicate a higher reactivity of MBs and vice
versa. Fig. 3 shows the 13C isotopic distribution of p/m-diMB,
triMB, tetraMB, and pentaMB trapped in the H-ZSM-5(15) chan-
nels. All these 13C-compounds were formed by gradually labeling
the 12C methanol-generated MBs rather than directly from 13C
methanol, because few fully labeled products were observed. Obvi-
ously, all the MBs show reactivity to methanol reactant. A large dif-
ference is demonstrated, however, in both 13C distribution and
total 13C content among the MBs. The incorporation of 13C
increases with the increase in the number of methyl substituents.
PentaMB displays much broader 13C atoms distribution and con-
tains about 42% 13C atoms, tetraMB having slightly fewer 13C atoms
(36%), while much fewer 13C atoms are incorporated into p/m-
diMB and triMBs (ca. 20% 13C atoms). This allows us to conclude
that the reactivity of MBs follows the order pentaMB > 1,2,3,5-
tetraMB, >p/m-diMB ꢁ 1,2,4-triMB. Thus, the rapid decrease of
pentaMB and tetraMB observed in Fig. 2b can be explained by their
high reactivity, which results in quick transformation. It is worth
noting that light MBs could be formed as products in this process.
In the proposed paring mechanism in the MTO reaction over
H-ZSM-5, light MBs are produced along with splitting off olefins
from bulkier MBs [6,33]. Additionally, the formed olefins could
regenerate light MBs via polymerization and cyclization. Both con-
tribute to the accumulation of p/m-diMB and triMBs. Fig. 2b shows
that the formation rate of the light MBs seems to fit well with the
transformation rate of the bulky MBs. The lower reactivity of the
light MBs is seen in the lower rate of decrease of these compounds,
as compared with the bulky MBs.
Fig. 2. GC–MS analysis of trapped products obtained from one pulse reaction of
methanol over H-ZSM-5(15) at 400 °C (a) and the evolution of the normalized
trapped MBs (b).
compounds, giving a higher vitra. In a certain zeolite with a well-
defined channel system and pore size, the diffusion rate is largely
affected by the molecular size of the respective MB. The MB with
a larger molecular size would have a lower diffusion rate due to
the confinement effect imposed by the zeolite framework. PentaMB,
1,2,3,5-tetraMB, and 1,3,5-triMB were not found in the effluent
products (Fig. S2), indicating that these MBs are too big to diffuse
out of the zeolite channels. Therefore, their diffusion rates are
defined to be zero. As for the lighter MBs such as diMBs, 1,2,4-
triMBs, and 1,2,4,5-tetraMBs detected in the effluent, they should
have higher diffusion rates than the above-mentioned bulkier
MBs. By comparing the molecular sizes, we deduce that their diffu-
sion rate follows the order diMBs > 1,2,4-triMB > 1,2,4,5-tetraMB.
Taken together, the diffusion rates of MBs in the zeolite channels
would follow the order p/m-diMB > 1,2,4-triMB > 1,2,4,5-
tetraMB > 1,3,5-triMB = 1,2,3,5-tetraMB = pentaMB = hexaMB.
The rate of change in certain MB concentration decreases when the
3.3. The relation of methylbenzenes to propene and ethene
sum of transformation and diffusion rates ðmtira
þ
vidiff Þ is greater
The hydrocarbon pool of MBs has been previously identified on
H-ZSM-5 under steady-state conditions [16,31,37,48]. The above
observation of MBs prior to light olefins and their reactivity
implies a distinctive role of MBs in the induction period. Valuable
than the formation rate ðviforÞ, and vice versa. In the reaction, the
methanol reactant on H-ZSM-5 is decreasing after the pulse. At
the moment when the formation rate of a certain MB is equal to

C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
131
Fig. 3. Isotopic analysis of trapped MBs in H-ZSM-5(15) channels. The catalyst was first exposed to 12C methanol for 32 s, and thereafter to 13C-methanol for 4 s.
information about MBs in the formation of light olefins can be
obtained by linking MBs to propene and ethene in the induction
period. We used the average number of methyl groups per
benzene ring (Meave) to describe the distribution of MBs trapped
in H-ZSM-5 [32]:
5
X
Meave
¼
nfn
ð2Þ
n¼2
fn denotes the fraction of MB with n methyl groups. Accordingly, a
higher concentration of bulkier MBs gives a larger Meave and vice
versa. The selectivity to ethene and propene is assessed by the
propene-to-ethene ratio (P/E). Fig. 4a shows the calculated P/E
and Meave of trapped MBs obtained from the reactions in Fig. 2, as
a function of reaction time. It is interesting to note that the P/E
shows a trend similar to that of Meave, gradually decreasing before
16 s and leveling off after 16 s. This indicates that the formation of
propene and ethene in the initial period is linked to the distribution
of MBs. At a very short time (2 s), the Meave is 4.3, implying that
bulkier MBs like pentaMB and tetraMB dominates the trapped
MBs (also see Fig. S3). At the same time, a high selectivity to
propene is obtained with P/E = 6.0. In comparison, an increase in
selectivity to ethene occurs with P/E dropping to 4.4 when the
Meave decreases to 3.1 after 16 s. The lower Meave indicates that
p/m-diMB and triMB are becoming dominant and responsible for
ethene formation. A similar correlation between Meave and P/E is
confirmed at a lower temperature of 350 °C (Fig. S4). Therefore,
our experimental observations clearly demonstrate that pentaMB
and tetraMB favor propene formation, while p/m-diMB and triMB
favor ethene formation in the induction period.
The previous work of Haw and co-workers indicated that
trapped MBs in SAPO-34 catalyst were essential for it to be active
in MTO reaction [49]. The HP center of MBs formed after the first
pulse of the methanol reaction would have a profound effect on
the subsequent reaction. Here, the H-ZSM-5(15) was subsequently
treated with an identical second pulse of methanol after 32 s of the
first pulse (Fig. S5 and Table S2). The conversion of methanol
Fig. 4. Evolution of Meave and P/E in a single-methanol-pulse experiment (a) and a
double-pulse experiment (b) at 400 °C. For a double-pulse experiment, the H-ZSM-5
(15) catalyst was first exposed to 20
another 20 L of methanol for 4 s. Meave refers to the average number of methyl
groups per benzene ring. P/E refers to the propene-to-ethene ratio.
lL of methanol for 32 s, and thereafter to
l

132
C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
quickly reaches 8% after 2 s of reaction. Consistent with the work of
Haw et al. [45], the pre-established HP shortens the induction time
of the following MTO reaction. The relationship of the MB distribu-
tion to ethene and propene is further analyzed. As shown in Fig. 4b,
the concentration of MBs with more methyl groups reflected by
higher Meave is increasing and reaches a maximum of 4.1 at 4 s, fol-
lowed by a steady decrease. The increase of Meave is associated
with formation of bulkier MBs due to the methylation of light
MBs by methanol from the second pulse. However, the high reac-
tivity of bulkier MBs leads to their rapid transformation to light
MBs and thus a decrease in Meave again at longer reaction time.
Analysis of ethene and propene shows that the evolution of calcu-
lated P/E has a trend similar to that of Meave. This agrees well with
the observation in Fig. 4a and provides direct evidence that the
MBs in the range from p/m-diMB to pentaMB are formed as reac-
tive catalytic centers in the initial reaction period. Importantly,
the selectivity of light olefins such as ethene and propene can be
correlated with the distribution of MBs.
tivity to propylene in the MTO reaction, based on which the com-
mercial MTP (methanol-to-propene) process has been developed
on high-silica ZSM-5 [51]. A series of MTO reactions by pulsing
methanol were conducted over H-ZSM-5 with two different Si/Al
ratios (15 and 100) at 400 °C (Fig. S7). The distribution of MBs
and P/E was analyzed and the data were listed in Table 2. For
ZSM-5 with a high Si/Al ratio, bulkier MBs are preferentially
formed compared with light MBs, as reflected by the higher Meave
.
The P/E increases from 4.92 (Si/Al = 15) to 9.25 (Si/Al = 100), which
can again be correlated with the distribution of MBs. Indeed, the
distribution of MBs is associated with the Brønsted acid concentra-
tions on different samples. On a low-silica H-ZSM-5, there are more
Brønsted acid sites on which MBs might be tightly adsorbed, while
on a high-silica sample, the lower concentration of Brønsted acid
sites would result in weaker adsorption of MBs. Taking into
account the larger confinement effect imposed by the channel of
H-ZSM-5 on the bulkier MBs, on the high-silica H-ZSM-5, the
formed light MBs could easily transport and diffuse out of zeolite
channels, leading to a higher concentration of bulkier MBs (larger
Meave). Thus, the predominant bulkier MBs formed on the high-
silica H-ZSM-5 contribute to the higher selectivity to propene.
3.4. The factors affecting methylbenzene distribution and olefin
selectivity
3.4.3. Reaction temperature
3.4.1. Space velocity
Reaction temperature also influences the distribution of MBs
and the selectivity of olefins. At 350 °C, Meave is 4.3 and P/E is 11
at 8 s of reaction time. In comparison, Meave and P/E drop to 3.5
and 4.9, respectively, at 400 °C (Table 2), which is consistent with
the correlation between MB distribution and selectivity to ethene
and propene. The effect of the temperature on the reaction was fur-
ther analyzed by comparing P/E at different reaction times when
Meave is similar. It is noted that at 400 °C and 2 s of reaction time,
the Meave increases to 4.31, similarly to that at 350 °C and 8 s of
reaction time. The corresponding P/E increases to 6.0, but still is
much lower than that (P/E = 11) at 350 °C. Thus, it can be con-
cluded that higher temperature favors the formation of ethene
on H-ZSM-5. Similarly, Svell et al. demonstrated that P/E was sen-
sitive to temperature and a higher temperature contributed to
ethene on both H-SAPO-34 and H-SSZ-13 catalysts [52]. It is clear
that the distribution of MBs can be related to ethene and propene
at different temperatures. However, the higher selectivity to
ethene observed at 400 °C and 2 s of reaction time suggests that
additional factors might affect olefin formation. This may be
understood by considering the thermodynamics of the formation
of ethene and propene at different temperature [20]. We assume
that ethene is produced through light MBs with a higher energy
barrier than that for propene from bulkier MBs. The formation
rates of both ethene and propene will increase with reaction tem-
perature. However, ethene is more favored than propene at higher
temperatures due to the higher energy barrier, which leads to a
decreased P/E.
The space velocity is a key factor in the production of olefins in
MTO reactions [20,50]. Here, the pulse reaction was carried out
with two space velocities by varying carrier gas (He) flow rate from
200 to 500 sccm (keeping the methanol content constant). The P/E
was calculated and the distribution of MBs was analyzed by Meave
from the trapped MBs (Fig. S6 and Table 2). The experimental data
indicate that a higher He flow rate of 500 sccm gives a slightly
higher Meave of 3.51 as compared with 3.19 at a lower He flow rate
of 200 sccm. Preferred formation of propene is evident at higher He
flow rates, with P/E being 4.92 at 500 sccm and 2.91 at 200 sccm.
This is consistent with the above observation that the distribution
of MBs affects the selectivity of propene and ethene. The different
diffusion properties of MBs in zeolite channels at different space
velocities should result in variation of MB distribution. The MBs
with more methyl groups, such as tetraMB and pentaMBs, would
experience much more diffusion hindrance in the H-ZSM-5 chan-
nels than the light MBs. A high space velocity promotes the diffu-
sion of light MBs out of zeolite channels but has little effect on
bulkier MBs. Accordingly, a relatively higher concentration of bulk-
ier MBs would be trapped on H-ZSM-5 and higher Meave can be
expected. Thus the selectivity of ethene and propene could be con-
trolled by the He flow rate, which influences the diffusion of
formed MBs.
3.4.2. Si/Al ratio
The Brønsted acidity associated with the Si/Al ratio of zeolite is
essential to the MTO reaction. Chang et al. demonstrated that
reducing the Brønsted acidity of H-ZSM-5 could improve the selec-
3.5. The methanol-to-olefins reaction under steady-state conditions
After analyzing the reactivity of MBs in the induction period, we
extended our study to the reaction under steady-state conditions.
The reactions were preformed over H-ZSM-5 with Si/Al ratios of
15, 100, and 200. Under 100% conversion at 400 °C, the turnover
frequencies (TOF) were measured as 0.018, 0.12, and 0.27 sꢀ1 for
H-ZSM-5(15), H-ZSM-5(100), and H-ZSM-5(200), respectively.
The TOF value for our low-silica H-ZSM-5(15) is comparable to that
(0.016 sꢀ1) determined for H-ZSM-5 with a Si/Al ratio of 25 in a
previous report [53]. Fig. 5 shows the GC–MS chromatogram of
trapped MBs on H-ZSM-5 after 15 min of reaction at 400 °C. MBs
up to hexaMBs are observable on the three samples. The
p/m-diMB and triMB decrease as the Si/Al ratio increases, while
tetraMBs and pentaMBs exhibit the reverse trend. This can be
Table 2
Meave and P/E ratio obtained from methanol pulse reactions over H-ZSM-5 at two
flow rates and temperatures. The Si/Al ratio of zeolite is given in parentheses.
P/Ec
b
Sample
Flow rate
(sccm)
Temperature
(°C)
Time
(s)a
Meave
ZSM-5(15)
ZSM-5(15)
ZSM-5(15)
ZSM-5(15)
ZSM-5(100)
200
500
500
500
500
400
400
350
400
400
8
8
8
2
8
3.19
3.51
4.30
4.31
4.32
2.91
4.92
11.01
6.0
9.25
a
b
c
Reaction time after one methanol pulse.
Average number of methyl groups per benzene ring.
Propene-to-ethene ratio.

C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
133
Fig. 5. GC–MS analysis of trapped products obtained from methanol reaction at
400 °C for 15 min over H-ZSM-5 with Si/Al = 15 (a), 100 (b), and 200 (c).
Table 3
Meave and P/E ratio obtained from methanol reaction over H-ZSM-5 under steady-
state conditions at 400 °C for 15 min.
a
Zeolite
Meave
P/Eb
2.68
5.95
11.32
ZSM-5(15)
ZSM-5(100)
ZSM-5(200)
3.49
3.64
3.84
Note: The Si/Al ratio of zeolite is given in parentheses.
a
Average number of methyl groups per benzene ring.
Propene-to-ethene ratio.
b
explained by considering both Brønsted acid site and zeolite
framework confinement effects on the transport and diffusion
behavior of MBs with different molecular sizes. The analysis of
Meave and P/E over H-ZSM-5 catalysts with different Si/Al ratios
shows a correlation of the MB distribution with the selectivity to
ethene and propene (Table 3), which demonstrates that propene
is favored when bulkier MBs are predominant while light MBs
mainly contributes to ethene under steady-state conditions, con-
sistent with the results derived from the induction period. It
should be noted that the competing interconversions are proceed-
ing under steady-state conditions, in which the methylation/crack-
ing of C+3 olefins might also contribute to propene formation
[21,37]. The dependence of olefins on the MBs trapped on zeolite
catalysts in both the induction period and the steady-state period
definitely implies the importance of MBs in controlling the forma-
tion of both ethene and propene.
Fig. 6. Isotopic distribution and total 13C content (%) in trapped organic species
obtained from transient 12C/13C isotopic switch experiments over H-ZSM-5 with
Si/Al = 15 and 100. 12C methanol was first reacted on a freshly activated catalyst
(0.2 g) at 400 °C for 15 min at a WHSV of 2 hꢀ1, and then the reaction was switched
to 13C methanol and allowed to proceed for 0.5 min.
13C-labeled diMB and 1,3,5-triMBs are also broadly distributed, but
the incorporation of 13C atoms is less. In addition, hexaMB is dom-
inated by isotopomers containing fewer 13C atoms and has the
lowest 13C content. Analysis of the 13C content among the MBs sug-
gests the difference in their reactivity toward the addition of
methanol reactant. Thus, the reactivity of the various MBs has
the order pentaMB > 1,2,3,5-tetraMB > p/m-diMB > 1,3,5-triMB >
1,2,4-triMB > hexaMB. This is consistent with that obtained from
the induction period. The low reactivity of hexaMB is also in agree-
ment with the previous report [29], confirming the smaller impor-
tance of such species in the MTO reaction on H-ZSM-5. It is noted
Transient 12C/13C isotopic switch experiments were conducted
to analyze the reactivity of trapped MBs under steady-state condi-
tions. Fig. 6 shows the isotopic data on MBs trapped on ZSM-5 with
Si/Al = 15 and 100. Different 13C distributions and total 13C content
are observable for the various MB compounds. On H-ZSM-5 with
Si/Al = 15, 1,2,3,5-tetraMBs and pentaMBs display a broad distribu-
tion centered on the isotopomers highly enriched in 13C atoms. The

134
C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
Fig. 8. Gas chromatography analysis of trapped products obtained from methanol
pulse reaction over H-ZSM-5(15) at 350 (a) and 400 °C. Each sample was prepared
by injecting 20 l
L of 13C methanol onto a freshly activated catalyst bed (0.2 g) while
He was flowed at 500 ml minꢀ1, and the reaction was allowed to occur for 8 s
followed by a rapid thermal quench.
moment, it is still unclear why the higher 13C incorporation rate
of the light MBs is observed.
3.6. Mechanistic understanding of the relationship between
methylbenzenes and olefins
In the co-reaction of methanol/DME with benzene, toluene, or
p/m-diMB, the transfer of carbon atoms from the MBs’ phenyl ring
to ethene and propene suggests a paring mechanism responsible
for the formation of light olefins [23,31,39]. In our previous work,
by the observation and identification of the key active intermedi-
ates such as cyclic carbenium ions with in situ solid-state NMR, a
detailed aromatic-based paring route was depicted for the forma-
tion of propene and ethene [34,36]. Here, we are seeking to gain
insight into the mechanistic link between MBs and ethene and pro-
pene in the initial stage of the MTO reaction. Fig. 7 shows the 13C
CP MAS NMR spectra of trapped products obtained from pulse-
quench experiments on H-ZSM-5(15) in which 13C-methanol was
injected at 350 and 400 °C and reacted for between 8 and 32 s
before thermal quench. After 8 s of reaction at 350 °C (Fig. 7a),
the conversion of methanol (50 ppm) and dimethyl ether
(60 ppm) is evidently accompanied by the formation of unsatu-
rated compounds such as MBs at 130–140 ppm and cyclic carboca-
tions at 240–260 ppm [34,45,54]. According to our previous work
[34,36], the characteristic signals at 244, 247, 250, 252, 255, and
257 ppm evidence the formation of 1,3-dimethylcyclopentenyl,
1,2,3-trimethylcyclopentenyl, 1,3,4-trimethylcyclopentenyl, 1-
methyl-3-ethylcyclopentenyl, 1,5-dimethyl-3-ethylcyclopentenyl,
and 1,4-dimethyl-3-ethylcyclopentenyl carbocations. It is noted
that the signal intensities of these species are much lower than
Fig. 7. 13C CP/MAS NMR spectra of trapped products obtained from methanol
pulse–quench reaction over H-ZSM-5(15) at 350 (a–c) and 400 °C (d–f). Each
sample was prepared by injecting 20 l
L of 13C methanol onto a freshly activated
catalyst bed (0.2 g) while He was flowed at 500 ml minꢀ1, and the reaction was
allowed to occur for different time followed by a rapid thermal quench.
that there exists a difference in the reactivity of 1,3,5-triMBs and
1,2,4-triMBs, with the former being more reactive than the latter.
This indicates that the reactivity of MBs is sensitive to the geome-
try of the trapped molecules, though both triMBs have the same
number of methyl substitutions. The confinement effect imposed
by the zeolite channels is assumed to result in such a difference
for the two triMBs.
The 13C incorporation was also analyzed on the high-silica H-
ZSM-5(100). Interestingly, the content of 13C atoms in light MBs
such as diMB and triMBs is notably increased and higher than that
in tetraMB and pentaMB. The enhanced selectivity to propene on
the high-silica sample compared with the low-silica sample (see
Table 3) suggests that the higher 13C incorporation rate of light
MBs does not necessarily reflect the increase of reactivity. Other-
wise, the selectivity to ethene would increase according to the
above-established relationship between MBs and olefins. At the

C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
135
Scheme 1. Proposed paring routes for the formation of ethene and propene.
those under steady-state condition [34,36], indicating the low con-
centration of these carbocations formed in the induction period.
Besides these C5-cyclic cations, the formation of C6-cyclic cations
such as pentamethylbenzenium ions is unambiguously demon-
strated by the signals at 205 and 189 ppm [55] (Fig. 7a and b). To
further support the assignment of these species, the trapped com-
pounds on the catalysts were isolated by dissolving the catalyst
with HF and extracting with CH2Cl2 and then were analyzed by
GC–MS. The GC spectra of trapped products show the presence of
methyl- and ethylcyclopentadienes, such as 1,3-dimethylcyclo-
pentadiene (1,3-DiMCP), 1,2,3-trimethylcyclopentadiene (1,2,
3-TriMCP), 1,2,4-trimethylcyclopentadiene (1,2,4-TriMCP), 1-
methyl-3-ethylcyclopentadiene (1-M-3-ECP), 1,4-dimethyl-2-ethy
lcyclopentadiene (1,4-DiM-2-ECP), and 1,2-dimethyl-4-ethylcyclo
pentadiene (1,2-DiM-ECP) (Fig. 8). Indeed, we have previously
identified all these species obtained under steady-state conditions
on H-ZSM-5 as well [34]. Since the cyclopentadienes can be consid-
ered as the conjugated bases of the corresponding cyclopentenyl
cations [34,43], the methyl- or ethylcyclopentenyl cations (1,3-
increase of signals at 130–140 ppm and at 240–260 ppm. At the
same time, the formation of light olefins speeds up as observed
in Fig. 1, suggesting the intermediate role of the cations in the for-
mation of olefins. All these cyclopentenyl cations can be observed
at higher reaction temperature (400 °C) but at lower intensity
(Fig. 7d–f). Comparison of reactions at different temperatures
(325–400 °C) shows that the lower temperature definitely favors
the stabilization and observation of the active carbocations espe-
cially pentamethylbenzenium cation (Fig. S8). Under steady-state
conditions, similar C5- and C6-cyclic cations have been identified
as key intermediates involved in the aromatics-base cycle for the
formation of light olefins via a paring mechanism that connects
MBs with isobutene [33], propene [34,56], and ethene [36].
Typically, MBs initiates the paring cycle, in which the key step is
the ring contraction of C6 to C5 species and splitting off of an alkyl
group from the C5 compound to form olefin. Thus, it can be
expected that a similar mechanism proceeds in the induction
period. We have established above that lighter MBs with two or
three methyl groups are related to ethene, while bulkier MBs favor
propene. Therefore, ethene and propene could be formed via a
dimethylcyclopentenyl,
1,2,3-trimethylcyclopentenyl,
1,3,4-
trimethylcyclopentenyl, 1-methyl-3-ethylcyclopentenyl, 1,5-dime
thyl-3-ethylcyclopentenyl, and 1,4-dimethyl-3-ethylcyclopentenyl
cations) should be formed and presented on H-ZSM-5 in the initial
stage of MTO reaction, giving the low-field NMR signals (240–
260 ppm). The MBs and methyl- or ethyl-substituted cyclopen-
tenyl cations grow up after 8 s of reaction time, reflected by the
different aromatic-based paring route. Scheme 1 shows the
proposed routes from MBs to olefins via the observed carbocations.
The methylation of light MBs such as toluene would result in
benzenium cations [57], and its contraction gives rise to
ethylcyclopentenyl cations, which can split off ethene products
(Scheme 1a). The observation of different ethylcyclopentenyl

136
C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
cations (1-methyl-3-ethylcyclopentenyl, 1,5-dimethyl-3-ethylcy
clopentenyl, and 1,4-dimethyl-3-ethylcyclopentenyl cations)
suggests that the methylation of MBs could be different
(Scheme 1a–d), but the following ring contraction produces ethene
in a similar way. The methylated bulkier benzenium ions that gen-
erate propene could be generated by gem-methylation of triMB
and tetraMB [33]. The ring contraction of tetramethylbenzenium
and pentamethylbenzenium ions generates propyl-attached 1,3,5,
6,6-pentamethylbicyclo[3.1.0] and 1,3,6,6-tetramethylbicyclo[3.1.
0] hexenyl cation transient state species, which subsequently split
off propene with simultaneous formation of methyl cyclopentenyl
cations (Scheme 1e and f). Based on energy calculations, this kind
of reaction route has been rationalized in the formation of propene
under steady-state conditions [34].
C6-cyclic carbocations were identified, suggesting the presence of
a paring mechanism in the induction period. The intimate correla-
tion between MBs and ethene and propene can be rationalized by
the proposed paring routes. The ethyl and propyl group-attached
cation intermediates are generated by ring contraction of lighter
and bulkier MBs, respectively, from which the elimination of alkyl
groups accounts for the formation of ethene and propene. These
results confirm the essential role of MBs in the MTO reaction and
shed new light on the MTO chemistry, particularly in the induction
period. The established mechanistic correlation between MBs and
ethene and propene is helpful for the design of zeolite catalysts
for achieving selective production of light olefins.
Acknowledgment
All these paring routes for the formation of ethene and propene
are working in the full catalytic cycles and are interconnected.
Along with the formation of ethene and propene, the active
methylcyclopentenyl cations are readily rearranged back into aro-
matics that will be involved in a new catalytic cycle [33]. For exam-
ple, the 1,3,4-trimethylcyclopentenyl cation produced by route (f)
could be expanded to p/m-diMB, from which trimethylbenzenium
ions are formed by methylation, subsequently leading to ethene
via route (b) in Scheme 1.
This work was supported by the National Natural Science Foun-
dation of China (Grants 21210005, 21173254, 21473245, and
21473246).
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.2015.10.001.
The bulkier benzenium cations can be formed by a sequence of
methylation of light MBs from methanol [58]. Waroquier et al. del-
icately demonstrated the striking effluence of the zeolite frame-
work in the methylation reaction in the confined space in terms
of reaction barriers and reaction energy [59]. The calculated results
showed that the increase of methyl groups leads to more reactive
benzenium cations. Pentamethylbenzenium ion was predicted to
be the most stabilized cation that could be formed by methylation
of tetraMB or protonation of rearranged pentaMBs [60]. This is con-
sistent with our observation that bulkier MBs such as tetraMB and
pentaMB have higher reactivity than their light counterparts
because of the facile formation of stable benzenium, such as pen-
tamethylbenzenium ions. Although the further methylation of
pentaMB may also occur to form hexamethylbenzenium ions, this
transformation would be unfavorable due to the higher reaction
energy of hexamethylbenzenium ions in the confined space of
the zeolite channel [59]. Additionally, the much larger hep-
tamethylbenzenium ion that was observed in zeolites such as Beta
[18] and SAPO-34 [43] with large pores or cavities is not likely to
be formed over ZSM-5 by methylation of hexaMB because of the
strong geometric constraint of the zeolite framework on the bulky
molecule. This also explains the low reactivity observed for hex-
aMB in the MTO reaction.
References
[1] C.D. Chang, Catal. Rev. 25 (1983) 1.
[2] R.M. Dessau, R.B. LaPierre, J. Catal. 78 (1982) 136.
[3] R.M. Dessau, J. Catal. 99 (1986) 111.
[4] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3.
[5] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003) 317.
[6] 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.
[7] K. Hemelsoet, J. Van der Mynsbrugge, K. De Wispelaere, M. Waroquier, V. Van
Speybroeck, ChemPhysChem 14 (2013) 1526.
[8] S. Ilias, A. Bhan, ACS Catal. 3 (2012) 18.
[9] J.F. Li, Z.H. Wei, Y.Y. Chen, B.Q. Jing, Y. He, M. Dong, H.J. Jiao, X.K. Li, Z.F. Qin, J.G.
Wang, W.B. Fan, J. Catal. 317 (2014) 277.
[10] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Angew. Chem.
Int. Ed. 45 (2006) 1714.
[11] W.G. Song, D.M. Marcus, H. Fu, J.O. Ehresmann, J.F. Haw, J. Am. Chem. Soc. 124
(2002) 3844.
[12] D.M. Marcus, K.A. McLachlan, M.A. Wildman, J.O. Ehresmann, P.W. Kletnieks, J.
F. Haw, Angew. Chem. Int. Ed. 45 (2006) 3133.
[13] I. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329.
[14] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304.
[15] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458.
[16] T. Mole, G. Bett, D. Seddon, J. Catal. 84 (1983) 435.
[17] T. Mole, J.A. Whiteside, D. Seddon, J. Catal. 82 (1983) 261.
[18] W.G. Song, J. Nicholas, A. Sassi, J.F. Haw, Catal. Lett. 81 (2002) 49.
[19] B.E. Langner, Appl. Catal. 2 (1982) 289.
[20] W.G. Song, H. Fu, J.F. Haw, J. Am. Chem. Soc. 123 (2001) 4749.
[21] M. Bjørgen, U. Olsbye, S. Svelle, S. Kolboe, Catal. Lett. 93 (2004) 37.
[22] M.W. Erichsen, S. Svelle, U. Olsbye, J. Catal. 298 (2013) 94.
[23] Ø. Mikkelsen, P.O. Rønning, S. Kolboe, Micropor. Mesopor. Mater. 40 (2000) 95.
[24] M. Seiler, W. Wang, A. Buchholz, M. Hunger, Catal. Lett. 88 (2003) 187.
[25] A. Sassi, M.A. Wildman, H.J. Ahn, P. Prasad, J.B. Nicholas, J.F. Haw, J. Phys. Chem.
B 106 (2002) 2294.
[26] A. Sassi, M.A. Wildman, J.F. Haw, J. Phys. Chem. B 106 (2002) 8768.
[27] R.F. Sullivan, C.J. Egan, G.E. Langlois, R.P. Sieg, J. Am. Chem. Soc. 83 (1961)
1156.
[28] M. Bjørgen, U. Olsbye, D. Petersen, S. Kolboe, J. Catal. 221 (2004) 1.
[29] M. Bjorgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 249 (2007) 195.
[30] X.Y. Sun, S. Mueller, H. Shi, G.L. Haller, M. Sanchez-Sanchez, A.C. van Veen, J.A.
Lercher, J. Catal. 314 (2014) 21.
4. Conclusions
The hydrocarbon pool of MBs in the MTO reaction was investi-
gated over zeolite H-ZSM-5. The reactivity of MBs in the induction
period was revealed by 12C/13C switching experiments. Bulkier
MBs (tetraMB and pentaMB) exhibit higher reactivity than lighter
counterparts (p/m-diMB and triMB). The formation of ethene and
propene was found to be correlated with the MBs trapped on
H-ZSM-5. Lighter MBs such as p/m-diMB and triMB favor ethene
formation, while bulkier MBs like tetraMB and pentaMB favor pro-
pene formation, particularly at low reaction temperature (<400 °C).
The MBs distribution in the induction period directly influences the
selectivity to ethene and propene. Preferential formation of pro-
pene can be achieved by using high-silica H-ZSM-5, a moderate
reaction temperature, and a high space velocity (carrier gas flow).
The same correlation of MBs with ethene and propene was found
to be present under steady-state conditions as well. Using solid-
state NMR spectroscopy in combination with GC–MS, C5- and
[31] S. Ilias, A. Bhan, J. Catal. 311 (2014) 6.
[32] B. Arstad, S. Kolboe, Catal. Lett. 71 (2001) 209.
[33] 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.
[34] C. Wang, Y.Y. Chu, A.M. Zheng, J. Xu, Q. Wang, P. Gao, G.D. Qi, Y.J. Gong, F. Deng,
Chem. Eur. J. 20 (2014) 12432.
[35] D. Lesthaeghe, A. Horré, M. Waroquier, G.B. Marin, V. Van Speybroeck, Chem.
Eur. J. 15 (2009) 10803.
[36] C. Wang, X.F. Yi, J. Xu, G.D. Qi, P. Gao, W.Y. Wang, Y.Y. Chu, Q. Wang, N.D. Feng,
X.L. Liu, A.M. Zheng, F. Deng, Chem. Eur. J. 21 (2015) 12061.

C. Wang et al. / Journal of Catalysis 332 (2015) 127–137
137
[37] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.P. Lillerud, S. Kolboe, M. Bjørgen, J.
Am. Chem. Soc. 128 (2006) 14770.
[50] C. Mei, P. Wen, Z. Liu, H. Liu, Y. Wang, W. Yang, Z. Xie, W. Hua, Z. Gao, J. Catal.
258 (2008) 243.
[38] R. Khare, D. Millar, A. Bhan, J. Catal. 321 (2015) 23.
[39] R. Khare, A. Bhan, J. Catal. 329 (2015) 218.
[40] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3.
[41] S. Ilias, A. Bhan, ACS Catal. (2012) 18.
[42] P.W. Goguen, T. Xu, D.H. Barich, T.W. Skloss, W.G. Song, Z.K. Wang, J.B.
Nicholas, J.F. Haw, J. Am. Chem. Soc. 120 (1998) 2650.
[43] S.T. Xu, A.M. Zheng, Y.X. Wei, J.R. Chen, J.Z. Li, Y.Y. Chu, M.Z. Zhang, Q.Y. Wang,
Y. Zhou, J.B. Wang, F. Deng, Z.M. Liu, Angew. Chem. Int. Ed. 52 (2013) 11564.
[44] M. Bjørgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 249 (2007) 195.
[51] C.D. Chang, C.T.W. Chu, R.F. Socha, J. Catal. 86 (1984) 289.
[52] F. Bleken, M. Bjørgen, L. Palumbo, S. Bordiga, S. Svelle, K.P. Lillerud, U. Olsbye,
Top. Catal. 52 (2009) 218.
[53] F.L. Bleken, S. Chavan, U. Olsbye, M. Boltz, F. Ocampo, B. Louis, App. Catal. A
Gen. 447–448 (2012) 175.
[54] P.W. Goguen, T. Xu, D.H. Barich, T.W. Skloss, W.G. Song, Z. Wang, J.B. Nicholas,
J.F. Haw, J. Am. Chem. Soc. 120 (1998) 2650.
[55] T. Xu, D.H. Barich, P.W. Goguen, W.G. Song, Z.K. Wang, J.B. Nicholas, J.F. Haw, J.
Am. Chem. Soc. 120 (1998) 4025.
[56] U. Olsbye, M. Bjørgen, S. Svelle, K.P. Lillerud, S. Kolboe, Catal. Today 106 (2005)
108.
[57] J. Van der Mynsbrugge, M. Visur, U. Olsbye, P. Beato, M. Bjørgen, V. Van
Speybroeck, S. Svelle, J. Catal. 292 (2012) 201.
[58] A.M. Vos, X. Rozanska, R.A. Schoonheydt, R.A. van Santen, F. Hutschka, J.
Hafner, J. Am. Chem. Soc. 123 (2001) 2799.
[59] D. Lesthaeghe, B. De Sterck, V. Van Speybroeck, G.B. Marin, M. Waroquier,
Angew. Chem. Int. Ed. 46 (2007) 1311.
[45] J.F. Haw, J.B. Nicholas, W.G. Song, F. Deng, Z. Wang, T. Xu, C.S. Heneghan, J. Am.
Chem. Soc. 122 (2000) 4763.
[46] M.W. Erichsen, K.De. Wispelaere, K. Hemelsoet, S.L.C. Moors, T. Deconinck, M.
Waroquier, S. Svelle, V. Van Speybroeck, U. Olsbye, J. Catal. 328 (2015) 186.
[47] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137.
[48] S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, J. Phys. Chem. C 111 (2007)
17981.
[49] W.G. Song, J.F. Haw, J.B. Nicholas, C.S. Heneghan, J. Am. Chem. Soc. 122 (2000)
10726.
[60] M. Bjørgen, F. Bonino, S. Kolboe, K.P. Lillerud, A. Zecchina, S. Bordiga, J. Am.
Chem. Soc. 125 (2003) 15863.