A route to form initial hydrocarbon pool species in methanol conversion to olefins over zeolites

Article abstract of DOI:10.1016/j.jcat.2014.05.015

The formation mechanism of the original C-C bond in methanol conversion to hydrocarbons over zeolite catalysts remains a grand challenge, although many researchers have done a lot of work and made significant progress. Here, a convincing route for formation of initial hydrocarbon pool (HCP) species involving original C-C bonds from dimethyl ether (DME) and/or methanol is illustrated by combining coincident experimental and theoretically calculated results. Elaborate experimental results gave strong evidence for predominant direct mechanism in the initial methanol-to-olefins process catalyzed by SAPO-34. A critical intermediate of the methoxymethyl cation was detected and theoretically verified through the reaction of the methoxy group and DME. This intermediate species subsequently reacted with DME or methanol to produce C-C bond-containing compounds 1,2-dimethoxyethane or 2-methoxyethanol. Further formation of oxonium cations led to generation of ethers or alcohols, and further to propene as the primary alkene product that induced the occurrence of the HCP mechanism.2014 Elsevier Inc. All rights reserved.

Full text of DOI:10.1016/j.jcat.2014.05.015

Journal of Catalysis 317 (2014) 277–283
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
A route to form initial hydrocarbon pool species in methanol conversion
to olefins over zeolites
Junfen Li a, Zhihong Wei a, Yanyan Chen a, Buqin Jing a, Yue He a, Mei Dong a, Haijun Jiao a, Xuekuan Li b,
Zhangfeng Qin a, Jianguo Wang a, Weibin Fan a,
⇑
a State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, 27 South Taoyuan Road, Taiyuan 030001, China
b Applied Catalysis and Green Chemical Industry Laboratory, Institute of Coal Chemistry, Chinese Academy of Sciences, 27 South Taoyuan Road, Taiyuan 030001, China
a r t i c l e i n f o
a b s t r a c t
Article history:
The formation mechanism of the original C–C bond in methanol conversion to hydrocarbons over zeolite
catalysts remains a grand challenge, although many researchers have done a lot of work and made sig-
nificant progress. Here, a convincing route for formation of initial hydrocarbon pool (HCP) species involv-
ing original C–C bonds from dimethyl ether (DME) and/or methanol is illustrated by combining
coincident experimental and theoretically calculated results. Elaborate experimental results gave strong
evidence for predominant direct mechanism in the initial methanol-to-olefins process catalyzed by
SAPO-34. A critical intermediate of the methoxymethyl cation was detected and theoretically verified
through the reaction of the methoxy group and DME. This intermediate species subsequently reacted
with DME or methanol to produce C–C bond-containing compounds 1,2-dimethoxyethane or 2-methoxy-
ethanol. Further formation of oxonium cations led to generation of ethers or alcohols, and further to pro-
pene as the primary alkene product that induced the occurrence of the HCP mechanism.
Ó 2014 Elsevier Inc. All rights reserved.
Received 23 February 2014
Revised 19 May 2014
Accepted 20 May 2014
Keywords:
Methanol conversion to olefins
SAPO-34
Original C–C bond
Direct reaction mechanism
Methoxymethyl cation
1. Introduction
roles, and significant progress has been made, proposing olefin-
based and aromatic-based cycles for formation of olefins [17–24].
Methanol is the most important platform compound and energy
carrier for conversion of carbon resources such as coal, natural gas,
and biomass to fuels and commodity chemicals [1,2]. The principal
issue in methanol conversion to hydrocarbons, such as olefins/pro-
pene (MTO/MTP), aromatics (MTA), and gasoline (MTG), on solid
acid zeolites [1,2] is control of product selectivity. This needs a
clear and deep understanding of the catalytic mechanism, particu-
larly of the transformation pathway of C–O bonds to C–C bonds.
Although more than 20 direct mechanisms, including oxonium
ylide, carbene, carbocation, and methane–formaldehyde mecha-
nisms, have been proposed [2–5], the computed energy barriers
are unrealistically high and the proposed intermediates are
remarkably unstable [6,7]. Therefore, the hydrocarbon pool (HCP)
mechanism has been considered to govern the methanol-to-
hydrocarbon conversion process because of its reasonable inter-
pretation of the induction period at the early stage [8–19]. Thus,
most experimental and theoretical researchers focus on the
identity of hydrocarbon pool species and the illustration of their
However, the HCP mechanism did not account for the origin of
initial HCP species involving the formation of the first C–C bond.
Thus, the organic residual in the calcined zeolite catalyst was
assumed to be the initial HCP species [25–28]. Regardless of this,
the 13C MAS NMR and IR spectroscopy results for the conversion
of methoxy species over acidic zeolites support the existence of a
direct mechanism, although no direct evidence was provided
[29–31]. In addition, it was found that propene could be formed
from methoxy groups and dimethyl ether (DME), although no evi-
dence was obtained in this case, either [32], and carbene species
existed in the methylation of ethene over HZSM-5 [33]. This incon-
sistency shows that the most active and controversial debate in the
last forty years on the formation of the first C–C bond in methanol
conversion is still going on. The origin of the HCP species remains a
grand challenge due to its extreme complexity and the up-to-date
limited characterization techniques. Thus, attempts are made here
to show a convincing route for forming the original hydrocarbon
pool species at the initial MTO reaction catalyzed by SAPO-34.
The direct mechanism predominates the initial MTO process via
the formation of CH3OCH+2 intermediate species and propene is
the first alkene product that induces the HCP mechanism.
⇑
Corresponding author. Fax: +86 351 4041153.
E-mail address: fanwb@sxicc.ac.cn (W. Fan).
http://dx.doi.org/10.1016/j.jcat.2014.05.015
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

278
J. Li et al. / Journal of Catalysis 317 (2014) 277–283
The unit cell of HSAPO-34 (a = b = c = 9.421 Å,
a = b = c = 94.2°)
2. Methods
2.1. Experimental
was derived from the Silicalite-CHA structure (all Si atoms are
symmetrically equivalent), in which all Si atoms are alternatively
replaced by P and Al atoms, and one P atom is replaced by one Si
atom to generate one Brønsted acid site per cage [42]. This corre-
sponds to a Si/Al ratio of 0.17 in HSAPO-34. In the DFT calculations,
the p(1 ꢀ 1 ꢀ 1) cell was used for C1 reactions such as methoxy
and DME formation, while the p(2 ꢀ 1 ꢀ 1) cell was used for C2
reactions, e.g., C1–C1 coupling, in order to avoid the interaction
between molecules, as shown in the Supplementary Material
(Fig. S1 in the Supplementary Material). The proton is located at
the site of O(884), which refers to the part of 8-, 8-, and 4-mem-
bered rings (MR) in the framework. The other sites of O(864) and
O(844) are also involved in the reaction of C1–C1 coupling. The cal-
culations show that the relative energy differences of protons
bonded to the four nonequivalent O sites are less than 4.0 kJ/mol.
This indicates that the proton can shift in these four sites. Never-
theless, the acidic strength based on the adsorption energies of
NH3 increases in the order H(864) (ꢁ123.5 kJ/mol) > H(884)
(ꢁ117.7 kJ/mol) > H(844) (ꢁ100.3 kJ/mol). In the simulation of all
the reactions, all atoms in the cell are allowed to relax with the lat-
tice constants fixed.
SAPO-34 with a Si/Al ratio of 0.15 was synthesized with trieth-
ylamine as a template. The as-synthesized sample was calcined in
air for 5 and 10 h to obtain non-fully-calcined SAPO-34 (NFC-
SAPO-34) and fully calcined SAPO-34 (FC-SAPO-34; acid amount:
0.7 mmol/g).
The MTO catalytic properties of these two samples in the induc-
tion period were tested in a fixed-bed pulse reactor. Typically,
100 mg of catalyst was loaded and pretreated at 550 °C for 3 h
before evaluation. The products were analyzed by a Shimadzu
gas chromatograph (GC-2014C) or a Shimadzu gas chromato-
graph–mass spectrometer (GC–MS QP 2010), both of which were
equipped with a HP-PLOT/Q column (30 m ꢀ 0.32 mm ꢀ 20
lm).
The IR spectra of the samples were measured on a Bruker TENSOR
27 FT-IR spectrometer equipped with a MCT detector. Before the
spectra were recorded, the self-supported SAPO-34 wafer (30 mg)
was treated at 500 °C and 0.1 Pa for 2 h.
The procedures for preparing methoxy groups are as follows:
the pretreated SAPO-34 first adsorbed methanol at 30 or 50 °C
until it was saturated. Then the physically adsorbed methanol
was flushed with Ar or pumped out. Finally, the temperature was
ramped to 300 °C at a rate of 5 °C/min under flushing (in pulse
experiments) or pumping (for IR spectroscopy) conditions.
3. Results and discussion
3.1. Evidence for the existence of a direct reaction mechanism
2.2. Density functional theory calculation methods
In the methanol conversion process, DME is readily produced
over the acidic zeolite catalysts [2,43,44]. This is also confirmed
by the computational result that the energy barrier for DME forma-
tion from two methanol molecules through the interaction with
acid sites is not high (95.5 kJ/mol), and the rate constant is
1.3 ꢀ 104 sꢁ1 at 400 °C (R6, Fig. S2 in the Supplementary Material).
Therefore, we initially compared the reaction behavior of methanol
and DME on a fully calcined SAPO-34 (FC-SAPO-34) catalyst
(Fig. S3 in the Supplementary Material) using a pulse reactor. The
conversion of methanol was only 0.7% in the first injection, while
it quickly increased to 44.9% at the fifth pulse (Fig. 1a). This reac-
tion pattern is intimately associated with the accumulation of
HCP species, mainly methyl-substituted benzene and naphthalene
(Figs. S4 and S5 in the Supplementary Material) in the catalyst. In
contrast, when DME was injected, the conversion readily reached
6.8% at the first pulse, but it increased only moderately to 14.5%
at the sixth pulse. One might think that this is due to the ready for-
mation of the HCP species from DME at the initial time, but a slow
increase in the amount during the reaction process. However, even
when the HCP species was first generated in the catalyst by intro-
ducing methanol, a much lower conversion was still obtained for
DME (Fig. 1a). After five successive injections of methanol, its con-
version reached only 23.0%. This suggests that (1) a direct mecha-
nism probably predominates in the conversion of DME in the initial
process irrespective of the existence of the HCP mechanism, and/or
that (2) DME is converted via another type of HCP mechanism that
is largely different from that occurring in the conversion of meth-
anol—namely, the transformation of methanol and DME needs dif-
ferent types of HCP species.
Spin-polarized DFT calculations for periodic HSAPO-34 catalysts
were carried out with the Vienna ab initio simulation package
(VASP) [34,35] using the projector-augmented wave (PAW)
method [36,37] and the generalized gradient approximation with
the Perdew–Wang exchange-correlation function (GGA-PBE) [38].
Frequency calculations were carried out to verify that the obtained
stationary points are minimum structures with real frequencies
alone or transition states with only one imaginary frequency along
the reaction coordinates. The vibrational frequencies and normal
modes were calculated by diagonalization of the mass-weighted
force constant matrix, which was obtained using the method of
finite differences of force, as implemented in VASP. The ions are
displaced in the directions of each Cartesian coordinate by 0.02 Å.
The zero-point-energy (ZPE) corrections were calculated using
statistical mechanics based on the Boltzmann distribution. The
enthalpy, entropy, and Gibbs free energy were derived from the
partition functions. Activation energy is attained with the ZPE cor-
rection. The partition functions were calculated in the temperature
range of 250–400 °C, which was selected on the basis of the exper-
imental conditions. The rate constant k obtained using transition-
state theory (TST) is defined as follows [39,40]:
D
S–0 =R
ꢁ
G–0 =RT
kBT
h
kBT
h
–
0
DH
=RT
D
eꢁ
e
k ¼
¼
e
;
where kB is Boltzmann’s constant, h is Planck’s constant, and
H–0 , and S–0 are the changes of standard molar Gibbs free energy,
D
G–0 ,
D
D
enthalpy, and entropy between the transition state (TS) and the ini-
tiation state (reactant, IS), respectively.
To clarify this point, DME and methanol were pulsed to non-
fully-calcined SAPO-34 (NFC-SAPO-34), in which a certain amount
of template residue was present (Fig. S3 in the Supplementary
Material). It was found that the conversion of methanol reached
4.9% in the first injection, as high as seven times that obtained
on the FC-SAPO-34. This indicates that the template residue indeed
can act as HCP species. However, a different result was obtained for
DME; its conversion in the first pulse was about 7.5%, which is very
close to that (6.8%) attained on the FC-SAPO-34. This shows that
All the reaction energy barriers over the SAPO-34 were calcu-
lated by the nudged elastic band (NEB) method [41] with eight
equally spaced images along the reaction pathway. The adsorption
energy were calculated with the equation Eads = E(molecule@HSAPO-34)
ꢁ [E(molecule) + E(HSAPO-34)], where E(molecule@HSAPO-34), E(molecule), and
E(HSAPO-34) are the total energies of the HSAPO-34 unit cell with
adsorbate (methanol) in the pores, free adsorbate (methanol)
molecule, and HSAPO-34 unit cell, respectively.

J. Li et al. / Journal of Catalysis 317 (2014) 277–283
279
the template residue has only a marginal effect on the conversion
of DME, suggesting that the conversion of DME follows a direct
mechanism, instead of the HCP mechanism, in the initial process.
To further confirm this inference, highly purified 13C methanol
was pulsed to the FC-SAPO-34 and NFC-SAPO-34, and the pulse
dosages were adjusted to ensure that the amounts of produced ole-
fins were comparable. For the NFC-SAPO-34, a methanol conver-
sion of 47.3% was attained with the formation of 12C and 13C
olefins (Fig. 1b and Fig. S6 in the Supplementary Material). This
indicates that the carbon (C) atoms were transferred from hydro-
carbon pool species (template residue) to the product. Over the
FC-SAPO-34, however, only 13C olefins were produced regardless
of the relatively low methanol conversion (1.6%) (Fig. 1c). In order
to exclude the possible influence of impurities in reagents such as
13C ethanol, 13C methanol was prepared in situ by reacting a 13C
methoxy group with H2O according to the reported protocol
[16,43,44]. As a consequence, a similar product distribution (13
C
olefins only) was observed (Fig. 1d), further verifying the direct
mechanism.
3.2. Formation of methoxy groups
Upon methanol adsorption on FC-SAPO-34, methoxy groups are
readily formed as the first intermediate species and they are reac-
tive for a number of probe molecules such as water, methanol, and
aniline [2,43,44]. The asymmetric C–H stretching vibration band of
chemisorbed methanol gradually shifted from 2958 to 2977 cmꢁ1
with the increase of temperature from 50 to 300 °C, showing the
formation of methoxy species (Fig. 2a and Fig. S7 in the Supple-
mentary Material) [45,46]. However, the chemisorbed DME cannot
be transformed into the methoxy groups; most of the DME are des-
orbed upon evacuation and increase of temperature due to the
lower adsorption energy of DME (33.8 kJ/mol) than of methanol
Fig. 1. Experimental evidence for the direct mechanism. (a) Methanol and DME
conversions obtained at different pulses in the MTO process catalyzed by FC-SAPO-
34 and DME conversions attained after successively injecting various times of
methanol; e.g., the DME conversions at pulses 3, 4, 5, and 6 refer to the values
obtained by pulsing DME after successive injection of 2, 3, 4, and 5 times of
methanol. 100 mg of FC-SAPO-34 was first pretreated at 550 °C for 2 h in air before
the reaction, which was carried out at 400 °C with Ar as carrier gas, the flow rate of
which was 300 mL/min. Every time 0.07 mmol of methanol or 0.035 mmol of DME
was injected. The methanol and DME conversions were calculated by considering
both of them in the effluents as unreacted substrates. (b) Reaction results of 13C-
methanol over NFC-SAPO-34, which was activated at 550 °C under Ar before
0.005 mmol of 13C-methanol was pulsed. (c) Reaction results of 13C-methanol over
FC-SAPO-34, to which 0.02 mmol of 13C-methanol was pulsed. (d) Reaction results
of 13C methoxy group (13CH3O) and 5
lL of H2O. The green-bar plots and the red-bar
Fig. 2. IR spectra for the formation of methoxy groups from methanol and DME. (a)
200 Pa of methanol or DME was introduced into the IR cell equipped with 30 mg of
FC-SAPO-34 self-supported wafer at 50 °C. The spectra were recorded from 50 to
300 °C with a temperature ramp rate of 5 °C/min. (b) DME reacted at 300 °C for 30 s
at different pressures.
plots in (b–d) are the mass spectra of produced ethene and propene, respectively.
The labeled numbers correspond to the mass/charge ratio (m/z). (For interpretation
of the references to color in this figure legend, the reader is referred to the web
version of this article.)

280
J. Li et al. / Journal of Catalysis 317 (2014) 277–283
(67.5 kJ/mol; see Fig. 2a). The methoxy group could only be
detected when DME was reacted for more than 30 s at 300 °C
and under higher DME pressure (Fig. 2b). Computational results
show that the formation of methoxy groups is realized through
the decomposition of a trimethyl oxonium (TMO) cation formed
from two DME molecules (R1 and R2, Fig. S2 in the Supplementary
Material). The energy barriers and rate constants for the formation
and decomposition of TMO cations are 101.3 kJ/mol and
1.3 ꢀ 104 sꢁ1 (400 °C) and 59.8 kJ/mol and 1.9 ꢀ 1010 sꢁ1 (400 °C),
respectively. In contrast,
a remarkably high energy barrier
(198.8 kJ/mol) and a very low rate constant (8.0 ꢀ 10ꢁ2 sꢁ1) were
obtained for the formation of methoxy groups from a single DME
molecule (R5 in Fig. S2 of the Supplementary Material), indicating
that it is difficult for this to occur. These calculated results are in
good agreement with the dependence of methoxy group formation
on the DME pressure (Fig. 2b). This is also supported by the results
calculated by Blaszkowski and Lesthaeghe with 1T and 5T cluster
models, although much higher energy barriers (141 and 131.3 kJ/
mol) were obtained for generation of TMO species [4,6], which is
the rate-determining step for formation of methoxy groups from
two DME molecules over zeolite acidic sites. A higher energy bar-
rier (59.8 kJ/mol) obtained by the periodic method for the decom-
position of TMO to methoxy groups and DME than those calculated
with 1T (27 kJ/mol) and 5T (47.7 kJ/mol) clusters is due to the sta-
bilization of the TMO species by the HSAPO-34 framework.
3.3. Formation of original C–C bond
The much higher reactivity of DME than of methanol in the ori-
ginal MTO process (Fig. 1a) and the conversion of DME via the
direct mechanism suggest that methoxy groups and DME may
react to form the intermediates for generating hydrocarbons. Com-
putations show that the energy barrier and rate constant of this
reaction are 135.1 kJ/mol and 4.2 ꢀ 104 sꢁ1 (400 °C) with the for-
mation of methoxymethyl cations (CH3OCH+2) stabilized inside
the SAPO-34 framework (Fig. 3a). The stabilization energy of CH3-
OCH+2 species is about 276 kJ/mol (relative to CH3+) [47], being sim-
ilar to that of tertiary carbocations as a result of lone pair-electron
donation from the oxygen to carbon. The distance between H
atoms in CH3OCH+2 and O atoms in the framework is 1.936 Å, indic-
ative of interaction between the zeolite framework and CH3OCH+2.
Four types of active carbocations, viz., 1,3-dimethylcyclopente-
nyl, pentamethylcyclopentenyl, pentamethylbenzenium, and hep-
tamethylbenzenium cations, have been observed in the methanol
conversion process over zeolites by Haw and Liu with 13C MAS
NMR spectroscopy [16,23,24,48,49]. This suggests that it is possi-
ble to detect active CH3OCH+2 species since they can be stabilized
in the CHA cage of SAPO-34. Indeed, the formation of CH3OCH2+
intermediate species was experimentally substantiated by IR spec-
troscopy. When 6.5 Pa of DME was introduced into the IR cell
equipped with CH3O-formed FC-SAPO-34 self-supported wafers
at 30 °C, an intense band at 2964 cmꢁ1, attributed to asymmetric
C–H stretching vibration in the CH3 group of DME [50], was
observed (Fig. 3b). After the temperature was increased to
180 °C, a new band appeared at 2960 cmꢁ1, while the 2964 cmꢁ1
band decreased in intensity. This newly formed 2960 cmꢁ1 band
is characteristic of asymmetric C–H stretching vibration of the
CH2 group in CH3OCH2OZ (Z = zeolite; Fig. S8 in the Supplementary
Material) [50]. The absence of this new band in the IR spectrum of
the fresh FC-SAPO-34 wafer purged with DME under the same con-
ditions confirms that the CH3OCH+2 species indeed comes from the
reaction of methoxy groups and DME.
Fig. 3. (a) Route for the formation of the first C–C bond. (b) IR spectra for the
formation of CH3OCH+2 by the reaction of methoxy groups and DME. The blue
numbers represent the energy barriers (kJ/mol), while the orange and green
numbers in parentheses are the rate constants (sꢁ1) at 250 and 400 °C, respectively.
6.5 Pa of DME was introduced into an IR cell equipped with 30 mg of CH3O-formed
FC-SAPO-34 self-supported wafer at 30 °C. The spectra were recorded from 30 to
180 °C with a temperature ramp rate of 5 °C/min. The spectra recorded at 30 and
180 °C were denoted as CH3O + DME (30 °C) and CH3O + DME (180 °C), respectively.
The spectra of DMM (30 °C) and DMM (180 °C) were obtained by introducing 6.5 Pa
of dimethoxymethane into the IR cell equipped with 30 mg of the pretreated FC-
SAPO-34 self-supported wafer at 30 °C C, and then increasing the temperature to
180 °C. The spectra of DME (30 °C) and DME (180 °C) were obtained according to
the same procedures as those employed for measuring DMM (30 °C) and DMM
(180 °C), except that DME was introduced. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
these two reactions are 94.6 kJ/mol and 2.4 ꢀ 105 sꢁ1 and
102.3 kJ/mol and 3.7 ꢀ 105 sꢁ1 (400 °C), respectively. To our
knowledge, this is the most energetically favorable route reported
for the formation of the first C–C bond.
3.4. Route for formation of initial olefins
1,2-Dimethoxyethane or 2-methoxyethanol was then methyl-
ated by methoxy groups to give 2-methoxyethyldimethyl,
2-hydroxyethyldimethyl, or 2-methoxyethylmethyl oxonium
cations, which further decomposed into 2-methoxyethoxy or 2-
hydroxyethoxy species (Fig. 4 and Fig. S9 in the Supplementary
Material). Because these two species have similar reaction pro-
cesses, the formation of olefin products from 2-methoxyethoxy
species will be, as an example, illustrated in the following. The 2-
methoxyethoxy species was first deprotonated to give methyl vinyl
ether (Fig. 4b), which was methylated to generate methyl propenyl
ether, resulting in growth of the carbon chain from one to two C–C
bonds. This was followed by formation of dimethyl propenyl
oxonium cations and subsequent decomposition into propenoxy
species, which could easily transform into allyloxy species through
spontaneous isomerization. The formation of propenoxy species
The CH3OCH+2 species further couples with another DME or
methanol molecule to give 1,2-dimethoxyethane or 2-methoxy-
ethanol, forming the first C–C bond (Fig. 3a and Fig. S10 in the Sup-
plementary Material). The energy barriers and rate constants of

J. Li et al. / Journal of Catalysis 317 (2014) 277–283
281
Fig. 4. Routes for formation of ethene (a) and propene (b). The blue numbers represent the energy barriers (kJ/mol), while the orange and green numbers in parentheses are
the rate constants (sꢁ1) at 250 and 400 °C, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)
was supported by the finding that allyl alcohol promoted methanol
conversion more strongly than ethanol and 2-propanol as a result
of easier formation of propene (Fig. S10 in the Supplementary
Material). The allyloxy species further reacted with DME to pro-
duce propene and CH3OCH2+. The CH3OCH+2 species then coupled
with DME molecules to maintain the reaction cycle and avoid the
repeated generation of methane, while the propene quickly aroma-
tized into the HCP species, and thus greatly accelerated the meth-
anol conversion process (Fig. 5a and Fig. S11 in the Supplementary
Material), soon causing the HCP mechanism to dominate the meth-
anol conversion process.
formation of methyl ethyl ether from 2-methoxyethoxy groups
and DME molecules reached 163.1 kJ/mol with a rate constant of
only 5.4 ꢀ 101 sꢁ1 at 400 °C. This shows that the formation of eth-
ene is far more difficult than that of propene. Indeed, the first pulse
of methanol to FC-SAPO-34 at 400 °C gave a propene selectivity of
72.6%, 13.2 times the ethene selectivity (Table S1a in the Supple-
mentary Material). This difference was more significant when
DME was injected, as also observed by Yamazaki and co-workers
[32], although the relative propene amount sharply decreased with
increasing conversion for both DME and methanol (Table S1 in the
Supplementary Material).
The failure to detect 1,2-dimethoxyethane with GC–MS by
adjusting reaction temperature, contact time, and DME amount
seems difficult to understand. In fact, the small pore size
(3.8 ꢀ 3.8 Å) of HSAPO-34 would impose a steric effect on the dif-
fusion of 1,2-dimethoxyethane. This makes its diffusion rate far
lower than its rate of transformation to (2-methoxy) ethyl
dimethyl oxonium cations (rate constant 3.6 ꢀ 105 sꢁ1 at 400 °C),
which converts into (2-methoxy)ethoxyl groups and DME at an
extremely high rate (rate constant 2.7 ꢀ 108 sꢁ1 at 400 °C). In addi-
tion, the rate of transformation of 1,2-dimethoxyethane is higher
than its formation rate. Therefore, 1,2-dimethoxyethane is impos-
sible to detect by GC–MS.
Fig. 4 and Fig. S9 (R3) in the Supplementary Material show that
the energy barriers and the rate constants at 400 °C of all the ele-
mentary steps for the production of propene are lower than
135.1 kJ/mol and larger than 4.2 ꢀ 104 sꢁ1, respectively. In con-
trast, in the pathway to ethene (Fig. 4a), the energy barrier to the
However, the selectivity of ethene was unexpectedly higher
than that of propene when 1,2-dimethoxyethane reacted with
methoxy groups formed in the FC-SAPO-34 (Fig. S12 in the Supple-
mentary Material). This seems to contradict the computational
results and the generation of more propene in the initial period.
Actually, it does not. A very short contact time of 1,2-dimethoxy-
ethane with the FC-SAPO-34 does not allow it to diffuse into the
CHA cage, and consequently, its reaction mainly occurred on the
external surface, giving ethene as the major olefin product.
Fig. 5b shows that the propene/ethene ratio in the product expo-
nentially increased from 0.05 to 0.22 with decreasing carrier gas
flow rate from 300 to 7.1 mL/min (0.03 mmol of 1.2-dimethoxy-
ethane was pulsed), indicating that only 1,2-dimethoxyethane
molecules inside the CHA cages may preferentially transform into
propene. This is confirmed by the result that a further increase in
the propene amount (propene/ethene = 0.31) was achieved by
pulsing 0.03 mmol of 1.2-dimethoxyethane and allowing it to

282
J. Li et al. / Journal of Catalysis 317 (2014) 277–283
Table 1
Computational results for methane–formaldehyde and methoxymethyl cation-based
mechanisms.
Model
1st step
2nd step
b
b
E1 (kJ/mol)a
k1 (sꢁ1
)
E2 (kJ/mol)a
k2 (sꢁ1
)
Methane–formaldehyde mechanism
1T [4]
171
3T [3]
148.2
147.1
149.6
185.0
183.1
124.5
5T [6]c
Periodicd
4.0 ꢀ 100
2.1 ꢀ 103
1.7 ꢀ 10ꢁ6
7.4 ꢀ 10ꢁ1
Methoxymethyl cation mechanism
Periodicd
135.1
1.1 ꢀ 104
94.6
6.8 ꢀ 104
a
b
c
Energy barriers.
Rate constants.
Rate constants were calculated at 447 °C.
Rate constants were calculated at 350 °C.
d
two steps involved in the methane–formaldehyde mechanism
were also calculated by the same periodic method, and the results
are listed in Table 1. The calculated energy barrier (149.6 kJ/mol) of
the first step is similar to that obtained with the cluster model, but
that of the second step significantly decreases to 124.5 kJ/mol due
to the stabilization of the transition state by the SAPO-34 frame-
work. Nevertheless, the two steps involved in the methoxymethyl
cation-based mechanism not only are energetically more favorable,
but also proceed much faster. The formation of methoxymethyl
cations from methoxy groups and DME (first step) has an energy
barrier lower than that of the formation of methane and formalde-
hyde by about 14.5 kJ/mol, and the rate constant also increases
about four times. This difference is more significant for the second
step. The energy barrier for generation of 1,2-dimethoxyethane
from methoxymethyl cations and DME (second step) is only
94.6 kJ/mol, lower than that for formation of ethanol from CH4
and HCHO by about 30 kJ/mol. In particular, the rate constant
increases about 105 times. This shows that it is much more possible
for the methoxymethyl cation-based mechanism to occur than the
methane–formaldehyde mechanism.
Fig. 5. (a) Methanol conversion obtained over FC-SAPO-34 at different pulses
before and after injection of propene. At 250 °C, 0.07 mmol of methanol or propene
was pulsed into a reactor filled with 100 mg FC-SAPO-34 at a carrier gas (Ar) flow
rate of 300 mL minꢁ1. (b) Effect of carrier gas flow rate on the propene/ethene ratio
in the product obtained in the reaction of 1,2-dimethoxyethane and methoxy
groups formed over FC-SAPO-34. 0.03 mmol of 1,2-dimethoxyethane was pulsed
into a reactor filled with 100 mg of CH3O-formed FC-SAPO-34 at different carrier
gas (Ar) flow rates. The effluent gases were analyzed by GC.
adsorb on the CH3O-formed catalyst for 20 min at 50 °C before the
reaction.
4. Conclusions
3.5. Comparison with methane–formaldehyde mechanism
The reaction results of methanol and dimethyl ether (DME) over
HSAPO-34 and the isotope labeling experiment on the formation of
olefins from methanol evidenced the existence of a direct mecha-
nism in the process of conversion of methanol to olefins. It was
identified that the reaction of methoxy groups on SAPO-34 and
DME gave methoxymethyl cations (CH3OCH+2) that subsequently
coupled with DME or methanol to produce C–C bond-containing
compounds of 1,2-dimethoxyethane or 2-methoxyethanol. This is
followed by the formation of oxonium cations and ethers, which
led to generation of propene as the primary product that induced
the occurrence of the HCP mechanism. This finding provides new
fundamental insights into the mechanism in the formation of ini-
tial C–C bonds and original hydrocarbon pool species, and would
significantly promote study of the mechanism of methanol to
hydrocarbons that facilitates the design of highly selective metha-
nol conversion catalysts.
A similar reaction route proposed for the formation of the first
C–C bond is the methane–formaldehyde mechanism [3], which is
supported by the results of Kubelková and co-workers that DME,
methane, and formaldehyde were first formed before generation
of aromatics from 2 Pa of methanol on HZSM-5 at 400 °C [5].
According to this mechanism, the CH3 group of a methanol mole-
cule approaches a surface methoxy group and generates a CH4
molecule through the hydride transfer reaction, while the H of
the OH group in methanol simultaneously transfers to a zeolite
surface and gives a formaldehyde molecule. This is followed by
the formation of ethanol through the decomposition of methane
into H+ and CHꢁ3 , which then transfer to the basic oxygen atom of
ZOꢁ and the carbon atom of formaldehyde, respectively. Blaszkow-
ski and co-workers reported that formation of CH4 from methanol
and surface methoxy groups (first step) has an energy barrier of
171 kJ/mol (Table 1) [4]. An increase in the cluster model from
1T to 3T decreased the energy barrier to 148.2 kJ/mol. However,
the 3T cluster calculation shows that the energy barrier for forma-
tion of ethanol (second step) is as high as 185.0 kJ/mol [3]. Similar
results were obtained by Lesthaeghe and coworkers with a 5T clus-
ter model [6], showing that it is very difficult for the second step to
occur.
Acknowledgments
This research is supported by the National Basic Research
Program (2011CB201403 and 2011CB201406), the National Natu-
ral Science Foundation of China (21003148, 21103216, 21273263
and 21273264), the Natural Science Foundation of Shanxi Province
of China (2013021007-3), and an Autonomous Research Project of
SKLCC (2013BWZ002). The computational results are obtained on
In order to reasonably compare our proposed route (designated
as a methoxymethyl cation-based mechanism) with the methane–
formaldehyde mechanism, the energy barriers and reaction rates of

J. Li et al. / Journal of Catalysis 317 (2014) 277–283
283
[22] K. De Wispelaere, K. Hemelsoet, M. Waroquier, V. Van Speybroeck, J. Catal. 305
(2013) 76–80.
[23] J. Li, Y. Wei, J. Chen, P. Tian, X. Su, S. Xu, Y. Qi, Q. Wang, Y. Zhou, Y. He, Z. Liu, J.
Am. Chem. Soc. 134 (2012) 836–839.
the ScGrid of the Supercomputing Center, Computer Network
Information Center of Chinese Academy of Sciences.
[24] S. Xu, A. Zheng, Y. Wei, J. Chen, J. Li, Y. Chu, M. Zhang, Q. Wang, Y. Zhou, J.
Wang, F. Deng, Z. Liu, Angew. Chem. Int. Ed. 52 (2013) 11564–11568.
[25] W. Song, D.M. Marcus, H. Fu, J.O. Ehresmann, J.F. Haw, J. Am. Chem. Soc. 124
(2002) 3844–3845.
[26] Z. Cui, Q. Liu, S. Bain, Z. Ma, W. Song, J. Phys. Chem. C 112 (2008) 2685–2688.
[27] D.M. Marcus, K.A. McLachlan, M.A. Wildman, J.O. Ehresmann, P.W. Kletnieks,
J.F. Haw, Angew. Chem. Int. Ed. 45 (2006) 3133–3136.
[28] J.F. Haw, D.M. Marcus, P.W. Kletnieks, J. Catal. 244 (2006) 130–133.
[29] W. Wang, A. Buchholz, M. Seiler, M. Hunger, J. Am. Chem. Soc. 125 (2003)
15260–15267.
[30] Y. Jiang, W. Wang, V.R. Reddy Marthala, J. Huang, B. Sulikowski, M. Hunger, J.
Catal. 238 (2006) 21–27.
[31] Y. Jiang, W. Wang, V.R. Reddy Marthala, J. Huang, B. Sulikowski, M. Hunger, J.
Catal. 244 (2006) 134–136.
[32] H. Yamazaki, H. Shima, H. Imai, T. Yokoi, T. Tatsumi, J.N. Kondo, J. Phys. Chem.
C 116 (2012) 24091–24097.
[33] H. Yamazaki, H. Shima, H. Imai, T. Yokoi, T. Tatsumi, J.N. Kondo, Angew. Chem.
Int. Ed. 50 (2011) 1853–1956.
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.2014.05.015.
References
[1] G.A. Olah, A. Goeppert, G.K.S. Prakash, Beyond Oil and Gas: The Methanol
Economy, Wiley-VCH Verlag GmbH & Co. KGaA, 2009.
[2] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3–48.
[3] N. Tajima, T. Tsuneda, F. Toyama, K. Hirao, J. Am. Chem. Soc. 120 (1998) 8222–
8229.
[4] S.R. Blaszkowski, R.A. Van Santen, J. Am. Chem. Soc. 119 (1997) 5020–5027.
[5] L. Kubelková, J. Nováková, P. Jíru˚ , Stud. Surf. Sci. Catal. 18 (1984) 217–224.
[6] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Angew. Chem. Int.
Ed. 45 (2006) 1714–1719.
[34] G. Kresse, J. Hafner, Phys. Rev. B 47 (1993) 558–561.
[35] G. Kresse, J. Furthmüller, Phys. Rev. B 54 (1996) 11169–11186.
[36] P.E. Blöchl, C.J. Forst, J. Schimpl, Bull. Mater. Sci. 26 (2003) 33–41.
[37] P.E. Blöchl, Phys. Rev. B 50 (1994) 17953–17979.
[38] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865–3868.
[39] H. Eyring, Chem. Rev. 17 (1935) 65–77.
[7] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Chem. Phys. Lett.
417 (2006) 309–315.
[8] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329–336.
[9] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458–464.
[10] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304–309.
[11] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137–8138.
[12] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003) 317–
326.
[40] I. Chorkendorff, J.W. Niemantsverdriet, Concepts of Modern Catalysis and
Kinetics, Wiley-VCH, Weinheim, 2003.
[41] H. Jónsson, G. Mills, K.W. Jacobsen, in: B.J. Berne, G. Ciccotti, D.F. Coker (Eds.),
Classical and Quantum Dynamics in Condensed Phase Simulations, World
Scientific, 1998, pp. 385–404.
[13] J.F. Haw, D.M. Marcus, Top. Catal. 34 (2005) 41–48.
[14] U. Olsbye, M. Bjørgen, S. Svelle, K.P. Lillerud, S. Kolboe, Catal. Today 106 (2005)
108–111.
[42] B.M. Lok, J. Am. Chem. Soc. 106 (1984) 6092–6093.
[43] W. Wang, Y. Jiang, M. Hunger, Catal. Today 113 (2006) 102–114.
[44] W. Wang, M. Hunger, Acc. Chem. Res. 41 (2008) 895–904.
[45] T.R. Forester, R.F. Howe, J. Am. Chem. Soc. 109 (1987) 5076–5082.
[46] S.M. Campbell, X.Z. Jiang, R.F. Howe, Micropor. Mesopor. Mater. 29 (1999) 91–
108.
[47] R.H. Martin, F.W. Lampe, R.W. Taft, J. Am. Chem. Soc. 88 (1966) 1353–1357.
[48] T. Xu, D.H. Barich, P.W. Goguen, W. Song, Z. Wang, J.B. Nicholas, J.F. Haw, J. Am.
Chem. Soc. 120 (1998) 4025–4026.
[15] 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.
[16] P.W. Goguen, T. Xu, D.H. Barich, T.W. Skloss, W. Song, Z. Wang, J.B. Nicholas, J.F.
Haw, J. Am. Chem. Soc. 120 (1998) 2650–2651.
[17] K. Hemelsoet, J. Van der Mynsbrugge, K. De Wispelaere, M. Waroquier, V. Van
Speybroeck, ChemPhysChem 14 (2013) 1526–1545.
[18] U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.V. Janssens, F. Joensen, S. Bordiga,
K.P. Lillerud, Angew. Chem. Int. Ed. 51 (2012) 5810–5831.
[19] S. Ilias, A. Bhan, ACS Catal. 3 (2012) 18–31.
[49] W. Song, J.B. Nicholas, A. Sassi, J.F. Haw, Catal. Lett. 81 (2002) 49–53.
[50] F.E. Celik, T. Kim, A.N. Mlinar, A.T. Bell, J. Catal. 274 (2010) 150–162.
[20] M.W. Erichsen, S. Svelle, U. Olsbye, J. Catal. 298 (2013) 94–101.
[21] S. Ilias, A. Bhan, J. Catal. 311 (2014) 6–16.