Difference between the mechanisms of propylene production from methanol and ethanol over ZSM-5 catalysts

Article abstract of DOI:10.1016/j.apcata.2013.07.064

We investigated the difference between the mechanisms of methanol conversion and ethanol conversion to propylene over ZSM-5 catalysts. The mechanism of methanol conversion depended on the Si/Al₂ ratio, whereas the mechanism of ethanol conversion did not. We believe that at a high Si/Al₂ ratio, propylene was produced from methanol via dimethyl ether by means of the hydrocarbon pool mechanism. In contrast, at a low Si/Al ₂ ratio, dimethyl ether was dehydrated to ethylene, which was then directly converted to propylene. In the case of ethanol conversion, propylene was produced from ethylene, which was formed by intramolecular dehydration. We propose that ethylene was transformed to a carbene species, which was a transient intermediate for direct production of propylene.

Full text of DOI:10.1016/j.apcata.2013.07.064

Applied Catalysis A: General 467 (2013) 380–385
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Difference between the mechanisms of propylene production from
methanol and ethanol over ZSM-5 catalysts
Atsushi Takahashi, Wei Xia, Qiang Wu, Teruyuki Furukawa, Isao Nakamura,
∗
Hiromichi Shimada, Tadahiro Fujitani
Research Institute for Innovation in Sustainable Chemistry, National Institute of Advanced Industrial Science and Technology (AIST), 16-1 Onogawa,
Tsukuba, Ibaraki 305-8569, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
We investigated the difference between the mechanisms of methanol conversion and ethanol conversion
to propylene over ZSM-5 catalysts. The mechanism of methanol conversion depended on the Si/Al2 ratio,
whereas the mechanism of ethanol conversion did not. We believe that at a high Si/Al2 ratio, propyl-
ene was produced from methanol via dimethyl ether by means of the hydrocarbon pool mechanism. In
contrast, at a low Si/Al2 ratio, dimethyl ether was dehydrated to ethylene, which was then directly con-
verted to propylene. In the case of ethanol conversion, propylene was produced from ethylene, which
was formed by intramolecular dehydration. We propose that ethylene was transformed to a carbene
species, which was a transient intermediate for direct production of propylene.
Received 18 February 2013
Received in revised form 17 July 2013
Accepted 31 July 2013
Available online xxx
Keywords:
Propylene
Ethanol
Methanol
©
2013 Elsevier B.V. All rights reserved.
ZSM-5
Mechanistic difference
1
. Introduction
However, they are rapidly deactivated by coke deposition. Makarfi
et al. found that the activity and product selectivity of H-ZSM-
5 zeolites for conversion of bioethanol [6] depends on the Si/Al
ratio of the zeolites. Aguayo et al. studied the effect of temperature
and space time on the product distribution and catalyst deactiva-
tion over ZSM-5 zeolites [7]. Xia et al. studied the active sites on
ZSM-5 zeolites for transformation of ethanol into propylene [8] and
found that the nature and reactivity of the acidic active sites did not
Because of the impending shortage of petroleum and the need
to minimize CO2 emissions, the development of alternatives to
petroleum has become a high priority worldwide. One of the best
sources of alternative fuels is biomass because it is an abundant
and carbon-neutral renewable resource. Bioethanol is the fuel pro-
duced in the largest quantities from biomass feedstock. Although
bioethanol is currently used mainly as a fuel or fuel additive in
motor vehicles, it is also potentially useful as a feedstock in the
chemical industry [1]. For example, in 2007, Braskem started up a
pilot plant in Brazil to produce ethylene by means of dehydration
of bioethanol produced from sugarcane [2].
Propylene is one of the most important chemical feedstocks,
and the demand for propylene is growing much faster than that
for ethylene because of the need for propylene derivatives such
as polypropylene and propylene oxide [3]. Therefore, the devel-
opment of methods for producing propylene from bioethanol has
attracted considerable attention. Catalytic conversion of ethanol
to propylene on zeolites has been reported [4–12]. Zeolites with
an eight-membered-ring pore system, such as SAPO-34 [4] and
LEV [5], show high selectivity for ethanol-to-propylene conversion.
depend on the Si/Al ratio and that the catalytic activity was propor-
2
tional to the number of acidic sites. To improve the selectivity for
propylene from ethanol, ZSM-5 zeolites have been modified with
metals such as iron [9] or tungsten and lanthanum [10]. Recently,
we reported that modification of ZSM-5 zeolites with phospho-
rus effectively stabilizes the catalyst and enhances the propylene
selectivity of the ethanol conversion reaction [11,12].
The use of methanol rather than ethanol for the production of
hydrocarbons on acidic zeolite catalysts has also attracted consid-
erable attention [13–17]. ZSM-5 catalysts are effective for selective
propylene production from methanol [13,14], and more than 20
mechanisms (e.g., the oxonium-ylide mechanism and the ethyl-
ene methylation mechanism) have been proposed for the reaction
[14]. During the past decade, the hydrocarbon pool mechanism
has gained increasing acceptance [15–17]. However, the reaction
mechanism and the active sites for ethanol conversion over ZSM-5
are not well understood. Whether the mechanism and the active
site for ethanol conversion are the same as those for methanol
conversion is an interesting question.
∗ Corresponding author at: 16-1 Onogawa, Tsukuba, Ibaraki 305-8569, Japan.
Tel.: +81 29 861 8377; fax: +81 29 861 8374.
E-mail address: t-fujitani@aist.go.jp (T. Fujitani).
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.07.064

A. Takahashi et al. / Applied Catalysis A: General 467 (2013) 380–385
381
In the present work, we investigated the mechanistic differences
between methanol conversion and ethanol conversion over ZSM-5
catalyst by evaluating the relationship between product distribu-
tion and contact time.
4
3
2
0
0
0
80
60
40
20
0
C H
3
6
2
. Experimental
2.1. Catalyst preparation
DME
C
4
H
8
>
C hydrocarbons
5
NH -ZSM-5 zeolites with Si/Al ratios of 80 (CBV 8014, spe-
4
2
2
−1
cific surface area = 425 m g ) and 280 (CBV 28014, specific surface
area = 400 m g ) obtained from Zeolyst International were used
2
−1
10
0
C H
2 4
C H
4
10
as catalyst precursors. Prior to the conversion reactions, the NH -
4
◦
ZSM-5 samples were calcined in air at 600 C for 4 h to obtain
H-ZSM-5 (ZSM-5) samples.
C H
3
8
0
2
4
6
8
10
2.2. Catalyst testing
-3
-1
W/F / 10 g•cc •min
Conversion reactions of methanol, ethanol, dimethyl ether
DME), and ethylene were carried out in a continuous-flow fixed-
bed tubular quartz reactor at atmospheric pressure, and the
reaction temperature was monitored with a thermocouple in
the catalyst bed. ZSM-5 samples used to catalyze the reactions
were pressed, crushed, and sieved with a 14–22 mesh sieve. Prior
Fig. 1. Effect of contact time on product distribution for methanol conversion over
ZSM-5 at Si/Al2 = 280 (reaction temperature 500 C, methanol concentration 67%):
◦
(
ꢀ DME, ꢁ ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocar-
bons.
3. Results and discussion
to each reaction, the catalyst was pretreated in a flow of N
100 cm min ) at 600 C for 1 h, and then the reactor was cooled
2
3
−1
◦
(
3.1. Conversion of methanol and ethanol over ZSM-5 at
to the reaction temperature. Methanol (Wako Pure Chemical Indus-
tries Ltd., purity >99.5%) or ethanol (Wako Pure Chemical Industries
Ltd., purity >99.5%) was supplied by a plunger pump and vapor-
ized with a heater. DME or ethylene was supplied from a cylinder,
diluted with N2 to maintain a given concentration, and then
introduced to the reactor. The reaction products were analyzed by
means of an online gas chromatograph equipped with a hydrogen-
flame ionization detector and an Rt-Alumina PLOT column (Restek
Si/Al = 280
2
3.1.1. Methanol conversion
We evaluated the effect of contact time on the product
distribution for methanol conversion at 500 C over ZSM-5 at
◦
Si/Al = 280 and a feed gas methanol concentration of 67% (Fig. 1).
2
−4
−3
At W/F = 1.0 × 10 g cm min, methanol conversion was 65%, DME
was the main product (yield 40%), and the yields of propylene
and butenes were 8% and 5%, respectively. As the contact time
increased, methanol conversion increased, and the yield of DME
decreased. At W/F = 1.0 × 10 g cm min, methanol conversion
reached 100%, and no DME was observed. In contrast, the yields of
ethylene, propylene, and butenes increased with increasing con-
Co.) for C –C hydrocarbons, and with a thermal conductivity
1
8
detector and a ShinCarbon ST column (Shinwa Chemicals Industries
Ltd.) for N2 and H . The yields of hydrocarbons were calculated on
−3
−3
2
a carbon number basis.
−3
−3
2.3. Catalyst characterization
tact time, reaching their maxima at W/F = 2.5 × 10 g cm min.
The maximum yield of propylene was 37%. The effect of contact
time on the propylene yield was much larger than the effects on
the yields of ethylene and butenes. In contrast to the yields of
ethylene, propylene, and butenes, the yields of >C5 hydrocarbons
and paraffins (propane and butanes) increased continuously with
increasing contact time. Among the >C5 hydrocarbons, aromatics
(benzene, toluene, and xylene) were the main products, constitut-
ing 80% of the >C5 hydrocarbons at all contact times. These results
suggest that methanol first underwent dehydrative condensation
to form DME, which was then converted to ethylene, propylene,
and butenes. Ethylene, propylene and butenes were subsequently
converted to aromatics by cyclodehydrogenation and to paraffins
by hydrogenation.
We attempted to confirm that DME was an intermediate in
methanol conversion by evaluating the effect of contact time on the
product distribution for DME conversion (Fig. 2). The DME concen-
tration in the feed gas was maintained at 33%, so that the amount
of carbon supplied to the reaction was the same as the amount sup-
plied to the methanol reaction. The yields of propylene and butenes
The acidic site properties of the catalysts were investigated by
means of ammonia temperature-programmed desorption with a
BELCAT-32 (BEL JAPAN, Inc.) chemisorption analyzer. Each catalyst
sample (0.1 g) was placed in a small quartz tube and dried under a
3
−1
◦
He flow (99.99%, 30 cm min ) at 500 C for 1 h. The sample was
◦
3
−1
cooled to 100 C, the adsorption of NH /He (30 cm min ) was
3
allowed to proceed for 1 h, and then the catalyst was flushed with
3
−1
He (30 cm min ) at the same temperature for 1 h to remove NH3
that was physically adsorbed on the sample surface. Temperature-
programmed desorption measurements were carried out from 100
◦
◦
−1
to 600 C at a heating rate of 10 C min with He as the carrier
gas. The desorbed NH3 was quantified with a thermal conductivity
detector.
The X-ray diffraction patterns of the catalyst samples were
obtained with an X-ray diffractometer (RINT2000, Rigaku) oper-
ated at 40 kV and 40 mA using Cu K␣ monochromatized radiation
(
2
ꢀ = 0.154178 nm). The diffraction patterns were measured in the
◦
◦
◦
−1
ꢁ range of 20 –60 at a scan speed of 2.0 min and a step width
◦
−4
−3
of 0.02 .
at W/F = 1.0 × 10 g cm min were 14% and 10%, respectively. The
yields of ethylene, propylene, and butenes initially increased with
increasing contact time. The propylene yield reached a maximum
of 36% and then began to decrease with increasing contact time,
along with the yields of ethylene and butene. In contrast, the yields
of >C5 hydrocarbons and paraffins increased continuously with
Surface areas and pore volumes were determined by the
N2 adsorption–desorption method (Brunauer–Emmett–Teller
◦
method) at −196 C by means of a physisorption analyzer (ASAP
2
020, Micromeritics). Prior to the adsorption measurements, each
◦
sample was degassed at 350 C for 10 h under reduced pressure.

3
82
A. Takahashi et al. / Applied Catalysis A: General 467 (2013) 380–385
40
30
20
10
0
100
C H4
2
C H
3
6
8
6
4
0
0
0
C H
>C hydrocarbons
5
4
8
C H4
2
C H
20
0
C H
3 6
4
10
C H8
4
C H
3
8
0
2
4
6
8
10
0
2
4
6
8
10
-3
-1
W/F / 10 g•cc •min
-3
-1
W/F / 10 g•cc •min
Fig. 2. Effect of contact time on product distribution for DME conversion over ZSM-
Fig. 3. Effect of contact time on product distribution for ethanol conversion over
ZSM-5 at Si/Al2 = 280 (reaction temperature 500 C, ethanol concentration 50%): ꢁ
ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocarbons.
◦
5
at Si/Al2 = 280 (reaction temperature 500 C, DME concentration 33%): ꢀ DME, ꢁ
◦
ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocarbons.
increasing contact time. The similarity between the contact time
dependence of the product distribution for DME conversion and
that for methanol conversion (Fig. 1) confirmed that methanol con-
First, we evaluated the effect of contact time on the product dis-
tribution at 500 C and a feed gas ethanol concentration of 50%, so
◦
that the amount of carbon supplied to the reaction was the same
as the amount supplied to the methanol reaction (Fig. 3). Ethanol
was completely consumed by the reaction. At a short contact time
version over ZSM-5 at Si/Al = 280 proceeded via DME.
2
It has been suggested that propylene is produced from DME by
means of an oxonium-ylide mechanism [18,19]. This mechanism
−
3
−3
(W/F = 1.0 × 10 g cm min), no ethers (e.g., diethyl ether) were
observed, and ethylene was the only product, indicating that the
initial reaction was not a dehydrative condensation but rather
an intramolecular dehydration. The yield of ethylene decreased
continuously with increasing contact time, whereas the yields of
propylene and butenes increased with increasing contact time.
+
involves the methylation of DME by CH3 to form the trimethy-
loxonium cation, which is transformed to the dimethyloxonium
methylide by proton abstraction. Propylene is then produced by
methylation of the dimethyloxonium methylide and subsequent
proton abstraction. However, Lesthaeghe et al. demonstrated that
the dimethyloxonium methylide is a highly energetic species
and is not formed in zeolites, and these investigators thus sug-
gested that oxonium ylides are unlikely to play an important role
in methanol-to-olefin conversion [20]. Furthermore, hydrogen or
paraffins should be produced if propylene is produced by the
oxonium-ylide mechanism, but in the current study, hydrogen and
paraffins were not produced in the range of contact times at which
the propylene yield increased. Therefore, we concur that propylene
production by means of the oxonium-ylide mechanism is unlikely
to occur.
−
3
−3
Even at a long contact time (W/F = 10 × 10 g cm min), the pro-
pylene yield was only 14%, which is much lower than the maximum
yield from methanol (37%). No aromatics or paraffins were pro-
duced at any contact time. These results imply that in the reaction
over ZSM-5, ethanol was converted to propylene and butenes via
ethylene and that the mechanisms of propylene production from
methanol and ethanol were different.
The yields of propylene from methanol and ethanol differed sub-
stantially, suggesting that the number of acidic sites on ZSM-5 at
Si/Al = 280 was sufficient to activate methanol but insufficient to
2
Dessau et al. have suggested that propylene is produced by
methylation of ethylene with CH3+ [21,22]. However, Kolbe et al.
reported that ethylene methylation is only a minor pathway in
the methanol-to-hydrocarbon reaction [23], and they suggested a
hydrocarbon pool mechanism for methanol conversion over zeo-
lites [24]. Recently, this mechanism has gained experimental and
theoretical support [15–17,23–25]. Haw et al. used 13C magic-angle
spinning NMR spectroscopy to study the reaction of DME over H-
ZSM-5 zeolites and found that a 1,3-dimethylcyclopentenyl cation
was produced first and that the product distribution showed a pre-
ponderance of propylene relative to other olefins [25]. These results
provide solid proof that propylene can be formed from DME by the
hydrocarbon pool mechanism over ZSM-5. Therefore, we believe
that in our system, propylene was produced by means of a similar
activate ethanol. We previously reported that the catalytic activity
of ZSM-5 for ethanol conversion to propylene strongly depends on
the number of acidic sites [8]. The details of the mechanism of the
ethanol conversion reaction are described in Section 3.2.2.
3
3
.2. Conversion of methanol and ethanol over ZSM-5 at Si/Al = 80
2
.2.1. Methanol conversion
We also evaluated the effect of the Si/Al2 ratio on the
catalytic activity of ZSM-5 for methanol conversion by inves-
tigating the product distribution over ZSM-5 at Si/Al = 80
2
as a function of contact time (Fig. 4). At short contact times
−4
−3
(
W/F = 1.0 × 10 g cm min), the main product was DME (70%),
and ethylene was also produced. The DME yield decreased
continuously with increasing contact time, whereas the
mechanism on ZSM-5 at Si/Al = 280.
2
ethylene yield initially increased, reached
a maximum at
−
4
−3
3.1.2. Ethanol conversion
W/F = 5.0 × 10 g cm min, and then decreased. The yields of
We also investigated the mechanism of ethanol conversion to
determine whether it differed from that of methanol conversion.
propylene and butenes increased with increasing contact time,
−
3
−3
reached their maxima at W/F = 2.5 × 10 g cm min, and then

A. Takahashi et al. / Applied Catalysis A: General 467 (2013) 380–385
383
4
3
2
1
0
0
0
0
0
80
60
40
20
0
40
30
20
10
0
80
60
DME
C H4
2
3
C H6
C H6
3
>
C hydrocarbons
C H
5
2
4
>C hydrocarbons 40
5
C H8
4
C H
4
10
C H8
4
C
H
20
0
4
10
C H
8
C H8
3
3
0
2
4
6
8
10
0
2
4
6
8
10
-3
-1
-3
-1
W/F / 10 g•cc •min
W/F / 10 g•cc •min
Fig. 4. Effect of contact time on product distribution for methanol conversion over
ZSM-5 catalysts at Si/Al2 = 80 (reaction temperature 500 C, methanol concentration
Fig. 6. Effect of contact time on product distribution for ethanol conversion over
ZSM-5 at Si/Al2 = 80 (reaction temperature 500 C, ethanol concentration 50%): ꢁ
ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocarbons.
◦
◦
6
7%): ꢀ DME, ꢁ ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5
hydrocarbons.
−
3
−3
W/F = 2.5 × 10 g cm min, the yields of propylene and butenes
reached their maxima; the maximum yield of propylene was 27%,
which was similar to that for methanol conversion (Fig. 4). Sub-
sequently, the yields of propylene and butenes decreased with
increasing contact time, whereas the yields of >C5 hydrocarbons
and paraffins increased continuously with increasing contact time.
These results confirm that production of propylene from methanol
decreased. The maximum yield of propylene was 27%. The
yields of >C5 hydrocarbons and paraffins increased continuously
with increasing contact time, and aromatics were the main >C5
hydrocarbon products. These results suggest that over ZSM-5 at
Si/Al = 80, methanol underwent dehydrative condensation to form
2
DME, which was then converted to ethylene by intramolecular
dehydration. Subsequently, ethylene was converted to propylene
and butenes, which were converted to aromatics by cyclodehydro-
genation and to paraffins by hydrogenation. On the basis of these
proceeded via ethylene at Si/Al = 80.
2
3
.2.2. Ethanol conversion
To clarify the effect of the Si/Al2 ratio on ethanol conver-
sion, we examined the product distribution of the reaction over
ZSM-5 at Si/Al = 80 as a function of contact time (Fig. 6). At
results, we believe that at Si/Al = 280, propylene production from
2
DME occurred by means of the hydrocarbon pool mechanism,
whereas at Si/Al = 80, ethylene was directly transformed to
2
2
propylene.
−4
−3
short contact times (W/F = 5.0 × 10 g cm min), the main prod-
uct was ethylene, and the ethylene yield decreased with increasing
contact time. In contrast, the yields of propylene and butenes
increased with increasing contact time, reaching their maxima at
Next, we examined the effect of contact time on the reac-
tion of DME to confirm that propylene production from methanol
occurred via ethylene over ZSM-5 at Si/Al = 80 (Fig. 5). At short
2
contact times, ethylene was the only product. As the contact time
increased, the ethylene yield decreased, propylene and butenes
formed, and their yields increased with increasing contact time. At
−3
−3
W/F = 2.5 × 10 g cm min. The maximum propylene yield was
2
7%, which was similar to that for methanol conversion (Fig. 4).
Subsequently, the yields of propylene and butenes decreased
with increasing contact time, whereas the yields of >C5 hydro-
carbons and paraffins increased continuously with increasing
contact time; aromatics were the main >C5 hydrocarbons at
4
3
2
1
0
0
0
0
0
80
60
40
20
0
−
3
−3
W/F > 5.0 × 10 g cm min. These results indicate that ethanol
was first dehydrated to ethylene, which was then converted to
propylene or butenes. Subsequently, propylene and butenes were
C H4
2
converted to aromatics and paraffins on ZSM-5 at Si/Al = 80. Thus,
we found that both methanol and ethanol conversion over ZSM-5
at Si/Al = 80 proceeded via ethylene.
To clarify the mechanism of ethanol conversion, we examined
the elementary step of propylene production from ethylene. First,
we determined the product distribution for ethylene conversion as
2
C H
3
6
2
>
C hydrocarbons
5
C H8
4
a function of contact time on ZSM-5 at Si/Al = 80 and a feed gas eth-
2
ylene concentration of 33%, so that the amount of carbon was the
same as the amount supplied to the ethanol reaction (Fig. 7). The
yields of propylene and butenes increased with increasing contact
time, reaching their maxima at W/F = 2.5 × 10 g cm min. In con-
trast, the yields of >C5 hydrocarbons and paraffins increased with
increasing contact time. The dependence of the product distribu-
tion on contact time for ethylene conversion was in good agreement
with that for ethanol conversion (Fig. 6), which strongly suggests
that propylene was directly produced via ethylene during ethanol
conversion.
C H
4
10
C H
3
8
−
3
−3
0
2
4
6
8
10
-3
-1
W/F / 10 g•cc •min
Fig. 5. Effect of contact time on product distribution for DME conversion over ZSM-
◦
5
at Si/Al2 = 80 (reaction temperature 500 C, DME concentration 33%): ꢀ DME, ꢁ
ethylene, ᭹ propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocarbons.

3
84
A. Takahashi et al. / Applied Catalysis A: General 467 (2013) 380–385
ethylene only. Therefore, we considered that propylene was not
40
30
20
10
0
produced by the olefin metathesis on ZSM-5. Baba et al. reported
that on SAPO-34, trimerization of ethylene produces hexene, which
then undergoes ␤-fission to produce propylene [4]; and Murata
et al. reported that propylene is produced by means of a sim-
ilar reaction mechanism on ZSM-5 [10]. Recently, Wang et al.
reported the formation of polyenic species, (CH CH )n, by means
of oligomerization of ethylene on ZSM-5, as indicated by IR spec-
troscopy [3]. These investigators speculated that propylene is
produced by the cracking of the polyenic species. We consider
this mechanism to be similar to the ethylene trimerization and
C H6
3
>
C hydrocarbons
5
C H8
4
␤-fission mechanism described above. In contrast, Corma et al. sug-
gested that propylene is produced from ethylene and ethane via a
C-butonium ion on zeolites, as indicated by quantum chemistry
calculations [27]. In all these mechanisms, propylene is indirectly
produced via some intermediate. Therefore, none of these mech-
anisms can explain our results for direct production of propylene
from ethylene.
C H
4
10
C H8
3
Recently, Basset et al. reported that ethylene is transformed
directly into propylene on W(H) /Al O [28]. They proposed that
0
2
4
6
8
10
3
2
3
-3
-1
W/F / 10 g•cc •min
a carbene (CH ) complex formed on tungsten reacts with ethyl-
2
ene to form propylene by cross-metathesis. In contrast, Beran et al.
reported that interaction of ethylene with the hydroxyl groups of
HY zeolites, which leads to the formation of a stable ␲-complex,
weakens the C C bond [29,30]. Because the acid strength of the OH
groups of H-ZSM-5 zeolites is higher than that of faujasites, the for-
mation of a ␲-complex on H-ZSM-5 zeolites is energetically more
favorable than the formation of such a complex on faujasites. Thus,
we believe that a carbene species was formed as a result of weak-
ening of the C C bond on ZSM-5 and that the carbene species acted
as a transient intermediate in the direct production of propylene.
We propose the following mechanism for the ethylene conver-
sion reaction:
Fig. 7. Effect of contact time on product distribution for ethylene conversion over
ZSM-5 at Si/Al2 = 80 (reaction temperature 500 C, ethylene concentration 33%): ᭹
propylene, ꢀ propane, ꢂ butenes, ꢃ butanes, ꢄ >C5 hydrocarbons.
◦
Next, we examined the product distribution at the initial stage
of the reaction, that is, at very short contact times (Fig. 8). Pro-
pylene and butenes were the only products observed at this stage.
The yields of both products were linearly related to contact time,
and the straight lines passed through the origin. In contrast, we
observed an induction period for the production of C5 and C6
olefins, and no aromatics or paraffins were produced during the
initial stage of the reaction. These results strongly suggest that eth-
ylene was directly converted to propylene and butenes over ZSM-5.
Several researchers have studied the mechanism of propylene
production from ethylene. For example, Iwamoto et al. reported
that on Ni-MCM-41, the metathesis of ethylene with butene pro-
ceeds to form propylene [26]. We examined the reaction between
C2H4 → (CH2)2
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
CH + C H → C H
2
2
4
3
6
2
C H → C H
2
4
4
8
ethylene and butene over ZSM-5 at Si/Al = 80 to confirm whether
2
the olefin metathesis occurred on ZSM-5. However, propylene
production was not promoted compared with the reaction of
C H + C H → C5H
3
6
2
4
10
C H + C H → C H
4
8
2
4
6
12
0
0
0
0
0
.5
.4
.3
.2
.1
0
2
C H → C H + 3H
3
6
6
6
2
C H + C H → C7H + 3H
3
6
4
8
8
2
2
C H → C H + 3H2
4
8
8
10
C H6
3
CnH2n + H → CnH
2
2(n+1)
C H
5
10
Ethylene was transformed to a carbene species via a ␲-complex
on ZSM-5 (Eq. (1)), and the carbene reacted with ethylene to form
propylene (Eq. (2)). Ethylene was also converted to butene by
dimerization (Eq. (3)). Propylene or butene subsequently reacted
with ethylene to form C5 or C6 olefins, respectively (Eqs. (4) and
C H
4
8
C H
6
12
(5)). In addition, cyclodehydrogenation reactions between two
molecules of propylene, two molecules of butene, or between
propylene and butene yielded C6–8 aromatics and hydrogen (Eqs.
0
0.02
0.04
0.06
0.08
0.1
(
(
6)–(8)), and the hydrogen reacted with olefins to form paraffins
Eq. (9)). We believe that the carbene species (Eq. (1)) formed more
-
3
-1
W/F/ 10 g•cc •min
efficiently over ZSM-5 at Si/Al = 80 than at Si/Al = 280. However,
further investigations are required to clarify the mechanism for
propylene formation from ethylene.
2
2
Fig. 8. Effect on contact time on product distribution at the initial stage of eth-
ylene conversion over ZSM-5 at Si/Al2 = 80 (reaction temperature 500 C, ethylene
concentration 33%): ᭹ propylene, ꢂ butenes, ꢁ pentenes, ꢀ hexenes.
◦

A. Takahashi et al. / Applied Catalysis A: General 467 (2013) 380–385
385
4
. Conclusions
[6] Y.I. Makarfi, M.S. Yakimova, A.S. Lermontov, V.I. Erofeev, L.M. Koval, V.F.
Tretiyakov, Chem. Eng. J. 154 (2009) 396–400.
[
7] A.T. Aguayo, A.G. Gayubo, A.M. Tarrío, J. Bilbao, J. Chem. Technol. Biotechnol.
7 (2002) 211–216.
[8] W. Xia, A. Takahashi, I. Nakamura, H. Shimada, T. Fujitani, J. Mol. Catal. A 328
2010) 114–118.
9] M. Inaba, K. Murata, I. Takahara, React. Kinet. Catal. Lett. 97 (2009) 19–26.
10] K. Murata, M. Inaba, I. Takahara, J. Jpn. Petrol. Inst. 51 (2008) 234–239.
We elucidated the difference between the mechanisms of
7
methanol conversion and ethanol conversion over ZSM-5 catalysts.
The mechanism of methanol conversion depended on the Si/Al2
ratio. Propylene was produced from DME by means of a hydro-
carbon pool mechanism at a high Si/Al2 ratio; whereas at a low
Si/Al2 ratio, DME was dehydrated to ethylene, which was directly
converted to propylene. In contrast, the mechanism of ethanol con-
(
[
[
[11] Z. Song, A. Takahashi, I. Nakamura, T. Fujitani, Appl. Catal. A 384 (2010)
201–205.
[
12] A. Takahashi, W. Xia, I. Nakamura, H. Shimada, T. Fujitani, Appl. Catal. A 423–424
2012) 162–167.
(
version was independent of the Si/Al ratio. Propylene was directly
[13] T. Sano, Y. Kiyozumi, S. Shin, Sekiyu Gakkaishi 35 (1992) 429–440.
[14] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3–48.
2
produced via ethylene, which was formed by an intramolecular
dehydration reaction. On the basis of mechanistic analysis of the
initial stage of the reaction, we propose that ethylene was trans-
formed to a carbene, which was a transient intermediate in the
direct production of propylene.
[
15] W. Song, D.M. Marcus, H. Fu, J.O. Ehresmann, J.F. Haw, J. Am. Chem. Soc. 124
2002) 3844–3845.
16] S. Svelle, B. Arstad, S. Kolboe, O. Swang, J. Phys. Chem. B 107 (2003) 9281–9289.
(
[
[17] W. Wang, Y. Jiang, M. Hunger, Catal. Today 113 (2006) 102–114.
18] T.R. Forester, R.F. Howe, J. Am. Chem. Soc. 109 (1987) 5076–5082.
19] O. Dewaele, V.L. Geers, G.F. Froment, G.B. Marin, Chem. Eng. Sci. 54 (1999)
[
[
4
385–4395.
[20] D. Lesthaeghe, V. Van Speybroeck, G.B. Marin, M. Waroquier, Chem. Phys. Lett.
17 (2006) 309–315.
Acknowledgement
4
[
[
21] R.M. Dessau, R.B. LaPierre, J. Catal. 78 (1982) 136–141.
22] R.M. Dessau, J. Catal. 99 (1986) 111–116.
This work was supported by a grant from the New Energy and
Industrial Technology Development Organization (NEDO).
[23] S. Svellea, P.O. Rønning, S. Kolboe, J. Catal. 224 (2004) 115–123.
[
[
24] Ø. Mikkelsen, P.O. Rønning, S. Kolboe, Microporous Mesoporous Mater. 40
2000) 95–113.
(
References
25] J.F. Haw, J.B. Nicholas, W. Song, F. Deng, Z. Wang, T. Xu, C.S. Heneghan, J. Am.
Chem. Soc. 122 (2000) 4763–4775.
[
[
[
[
1] F. Cherubini, Energy Convers. Manage. 51 (2010) 1412–1421.
2] A. Morschbacker, J. Macromol. Sci. Polym. Rev. 49 (2009) 79–84.
3] B. Lin, Q. Zhang, Y. Wang, Ind. Eng. Chem. Res. 48 (2009) 10788–10795.
4] H. Oikawa, Y. Shibata, K. Inazu, Y. Iwase, K. Murai, S. Hyodo, G. Kobayashi, T.
Baba, Appl. Catal. A 312 (2006) 181–185.
[26] M. Iwamoto, Y. Kosugi, J. Phys. Chem. C 111 (2007) 13–15.
[27] M. Boronat, A. Corma, Appl. Catal. A 336 (2008) 2–10.
[28] M. Taoufik, E. Le Roux, J. Thivolle-Cazat, J.M. Basset, Angew. Chem. Int. Ed. 46
(2007) 7202–7205.
[29] S. Beran, P. Jiru, L. Kubelkova, J. Mol. Catal. 12 (1981) 341–349.
[30] S. Beran, J. Mol. Catal. 30 (1985) 95–99.
[
5] T. Inoue, M. Itakura, H. Jon, Y. Oumi, A. Takahashi, T. Fujitani, T. Sano, Micro-
porous Mesoporous Mater. 122 (2009) 149–154.