Conversion of methanol into hydrocarbons over zeolite H-ZSM-5: Ethene formation is mechanistically separated from the formation of higher alkenes

Article abstract of DOI:10.1021/ja065810a

The widely debated reaction mechanism for the conversion of methanol to hydrocarbons over acidic zeolite H-ZSM-5 has been investigated using isotopic labeling. The mechanistic findings for H-ZSM-5 are clearly different from those previously described at a detailed level for H-β and H-SAPO-34 catalysts. On the basis of the current set of data, we can state that, for H-ZSM-5, ethene appears to be formed exclusively from the xylenes and trimethylbenzenes. Moreover, propene and higher alkenes are to a significant extent formed from alkene methylations and interconversions. This implies that ethene formation is mechanistically separated from the formation of higher alkenes, an insight of utmost importance for understanding and possibly controlling the ethene/propene selectivity in methanol-to-alkenes catalysis. Copyright

Full text of DOI:10.1021/ja065810a

Published on Web 10/26/2006
Conversion of Methanol into Hydrocarbons over Zeolite H-ZSM-5: Ethene
Formation Is Mechanistically Separated from the Formation of Higher Alkenes
†
‡
‡
†
†
†
Stian Svelle, Finn Joensen, Jesper Nerlov, Unni Olsbye, Karl-Petter Lillerud, Stein Kolboe, and
,
‡
Morten Bjørgen*
Haldor Topsøe A/S, NymølleVej 55, DK-2800 Lyngby, Denmark, and Centre for Materials Science and
Nanotechnology, Department of Chemistry, UniVersity of Oslo, P.O. Box 1033 Blindern, N-0315 Oslo, Norway
Received August 10, 2006; E-mail: mobj@topsoe.dk
Understanding and obtaining product control have always been
a major challenge in the development of heterogeneous catalysts.
On the basis of fundamental mechanistic findings, this work outlines
ways of controlling the product selectivity in the conversion of
methanol to hydrocarbons. Methanol can be produced from natural
gas or any gasifiable carbonaceous material, such as petroleum,
coal or biomass. The conversion of methanol to hydrocarbons is
generally denoted as the MTH reaction, but as process conditions
and catalyst choice affect the selectivities, the abbreviations MTO
Scheme 1
time. The time evolution of the 13C content in the effluent was
determined from GC-MS analyses performed 0.5, 1.0, and 2.0 min
after the 12C/ C methanol switch. Corresponding isotopic data for
the organic material retained in the zeolite pores during reaction
were obtained from three parallel experiments by thermally
(
(
methanol to olefins), MTP (methanol to propene), and MTG
methanol to gasoline) are often used.
13
The MTH reaction mechanism has been debated intensively for
1
decades. It is well documented that methanol molecules cannot
be coupled directly into hydrocarbons at rates relevant to steady-
2-4
13
state conversion, and recent results from the groups of Haw and
Kolboe are clearly in favor of an indirect reaction route.5 An
indirect route, known as “the hydrocarbon pool mechanism”, was
outlined in the early 1990s.7 The hydrocarbon pool was described
as a catalytic scaffold, constituted by large organics adsorbed in
the zeolite, to which methanol is added and olefins are eliminated
in a closed catalytic cycle. Hence, an active MTH catalyst can be
viewed as an organic-inorganic hybrid material constituted by the
quenching the reaction after 0.5, 1.0, and 2.0 min of C methanol
,6
reaction. These three catalyst samples were dissolved in 15% HF,
and the liberated organics were extracted from the aqueous phase
with CH Cl and analyzed with GC-MS.
,8
2
2
The major products in the effluent, as determined by GC-FID at
20 min on stream were C (10 C %, mainly ethene), C (30 C %,
2
3
mainly propene), butenes (14 C %), butanes (5 C %), C5+ alkenes
and aromatics (40 C %), and traces of methane. The conversion of
methanol/dimethyl ether was 73% (see Supporting Information,
Table S1 for further details). These results are typical for low
aluminum H-ZSM-5 under the conditions employed.
9
inorganic zeolite and the organic hydrocarbon pool. For H-â and
H-SAPO-34,10 the main catalytic engines of the hydrocarbon pool
are the higher methylated benzenes or their cationic derivatives.
For H-â, the heptamethylbenzenium cation displays the greatest
reactivity, whereas hexamethylbenzene behaves similarly in H-
SAPO-34. A simplified representation of this cycle, exemplified
by propene formation from heptamethylbenzenium over H-â, is
outlined in Scheme 1.
The material retained within the catalyst was also analyzed after
20 min on stream by the HF dissolution procedure, and the GC-
MS chromatogram of the CH Cl extract is shown in Figure 1. The
2
2
chromatogram is dominated by a few compounds, all of them are
methylbenzenes (MBs), ranging from minor amounts of xylenes
via triMBs to larger amounts of tetra-, penta-, and hexaMB.
TetraMB (durene) is the largest hydrocarbon able to diffuse out of
the zeolite pores (detectable in the effluent), implying that penta-
and hexaMB are truly trapped within the catalyst, probably residing
in the channel intersections. The chromatogram in Figure 1 is
representative for a wider range of reaction conditions and H-ZSM-5
catalyst samples. Finally, for this sample, the compositions of the
gas phase and the retained material remain stable for several days
under these conditions.
Clearly, the detailed reaction mechanism varies with, for
example, pore architecture and possibly also catalyst acid strength.
H-ZSM-5 is the archetype MTG catalyst and also the catalyst in
Lurgi’s MTP process, and in this report we show that for H-ZSM-5
(
1) the higher methylbenzenes are present in the zeolite pores, but
they are virtually unreactive; (2) ethene appears to be formed
exclusively from the lower methylbenzenes; (3) propene and higher
alkenes are to a considerable extent formed from alkene methyla-
tions and interconversions (e.g., cracking reactions). The two latter
statements, implying that the ethene formation is mechanistically
separated from the formation of higher alkenes, are of utmost
importance for understanding and possibly controlling the ethene/
propene selectivity in MTO/MTP catalysis.
The pathway of methanol incorporation into both gas-phase
13
products and retained compounds was investigated using
C
13
labeling as described above. The time evolutions of total C content
for the gas phase and retained hydrocarbons are shown in Figure
-
1
-1) was reacted from gas-phase
Methanol (WHSV ) 7.0 gg
h
12 13
2
a and 2b, respectively. Strikingly, 0.5 min after C/ C methanol
(130 hPa) over 60 mg H-ZSM-5 (Si/Al ) 140) in a fixed bed reactor
switch, the gas-phase products are divided into two groups (Figure
12
at 350 °C. C methanol was fed for 18 min before switching to
13
2
a). The C contents in the C
3
-C
6
alkenes are very similar (87-
13
C methanol (99% purity) and reacting further for a predetermined
9
1%), whereas the second distinct group, that is, ethene and the
13
aromatics, contain considerably less C (70-75%). The same
†
Center for Materials Science and Nanotechnology.
Haldor Topsøe A/S.
‡
grouping is discernible also after 1 min, but the difference has
14770
9
J. AM. CHEM. SOC. 2006, 128, 14770-14771
10.1021/ja065810a CCC: $33.50 © 2006 American Chemical Society

C O M M U N I C A T I O N S
incorporation by another mechanism, and the only satisfactory
explanation is alkene methylations followed by cracking of the
1
2
larger alkenes thus formed. For example, immediately after C/
1
3
12
C methanol switch, a
C
3
propene may be methylated repeatedly
four times, yielding a 2C
1
13
C
heptene, which may in turn be
3
4
cracked into 13C enriched propene and butene. This cycle of C3+
alkenes methylation/cracking reactions is related to the scheme
originally proposed by Dessau.16 However, according to that scheme
ethene is reequilibrated with the higher alkenes, and cracking is
the only route for ethene formation. This is not in compliance with
Figure 1. Chromatogram (GC-MS) of retained compounds. Prior to HF
dissolution, the catalyst was exposed to methanol for 20 min at 350 °C.
13
our data as the discrepancy between the C content in ethene and
3 6
the C -C alkenes excludes any measurable ethene formation via
this alkene based cycle, that is, ethene is not an alkene cracking
product.
The above discussion leads to an interesting question regarding
mechanistic understanding and selectivity control: Do the C3+
alkene cycle and the aromatics/ethene cycle run independently?
This is probably not the case for H-ZSM-5, as the constant
production of aromatics during methanol conversion means that
some of the C3+ alkenes continuously form new aromatics. This
means that for H-ZSM-5 the aromatics/ethene cycle cannot run
without the C3+ alkene cycle. Inversely, on the basis of the low
reactivity of ethene toward methanol relative to that of propene
and butenes,17 the contribution to the C3+ alkenes involving ethene
is very small and might not be required for the C3+ alkene cycle to
occur. Finally, the significance of propene formation from the
methylbenzenes cannot be evaluated at this point.
Figure 2. Time evolution of 13C content in effluent (a) and retained material
1
2
13
(b) after C/ C methanol feed switch.
In conclusion, if it were possible to separate these two cycles,
by sterically supressing the formation of the larger aromatics (e.g.,
by altering the topology) and allowing product formation to occur
only via the C3+ alkenes, formation of ethene might be avoided in
an MTP application.
become smaller and disappears after 2 min of 13C methanol reaction,
as the system approaches complete isotope exchange.
Xylenes and triMB are present both in the effluent and among
the retained material (Figure 2b). The isotopic composition of the
13
xylenes is virtually identical in both cases, whereas the C content
in triMB is slightly lower for the retained molecules. Further, Figure
Supporting Information Available: Product details. This material
is available free of charge via the Internet at http://pubs.acs.org.
1
3
2
b shows an evident trend in the C content in the higher
1
3
methylbenzenes: The incorporation of C is slower as the number
of methyl groups increases. The low rate of 13C incorporation of
hexaMB in particular, is in stark contrast to data obtained in a
References
(
1) St o¨ cker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
similar manner for H-â and H-SAPO-34 where hexaMB unques-
(2) Song, W.; Marcus, D. M.; Fu, H.; Ehresmann, J. O.; Haw, J. F. J. Am.
Chem. Soc. 2002, 124, 3844-3845.
tionably has the highest reactivity.9
-11
As a direct implication of
(
3) Marcus, D. M.; McLachlan, K. A.; Wildman, M. A.; Ehresmann, J. O.;
this, it may be stated that a hydrocarbon pool mechanism involving
the highest methylbenzenes, proven to be all dominating for H-â
and H-SAPO-34, cannot be applicable to the H-ZSM-5 catalyst.
Rather, on the basis of the practically identical 13C contents and
Kletnieks, P. W.; Haw, J. F. Angew. Chem., Int. Ed. 2006, 45, 3133-
3136.
(
4) Lesthaeghe, D.; Van Speybroeck, V.; Marin, G. B.; Waroquier, M. Angew.
Chem., Int. Ed. 2006, 45, 1-10.
(5) Haw, J. F.; Song, W.; Marcus, D. M.; Nicholas, J. B. Acc. Chem. Res.
2003, 36, 317-326.
time evolutions, ethene formation proceeds via the xylenes and/or
the triMBs. Ethene formation from xylenes and triMBs is in accord
with results from Haw and co-workers,1 showing that the lower
methylbenzenes yield predominantly ethene, whereas the higher
analogoues favor propene. The exact mechanism for ethene
elimination cannot be assessed based on the current data, but Haw
and co-workers1 have also shown that methylated cyclopentenyl
cations are integral to alkene formation on H-ZSM-5 and suggested
that these compounds are equilibrated with other, active carbon
species. No five rings are found among the retained compounds in
our experiments, even so it seems likely that both five- and six-
(6) Olsbye, U.; Bjørgen, M.; Svelle, S.; Lillerud, K.-P.; Kolboe, S. Catal.
Today 2005, 106, 108-111.
(
(
(
7) Dahl, I. M.; Kolboe, S. Catal. Lett. 1993, 20, 329-336.
2,13
8) Dahl, I. M.; Kolboe, S. Stud. Surf. Sci. Catal. 1995, 98, 176-177.
9) Bjørgen, M.; Olsbye, U.; Petersen, D.; Kolboe, S. J. Catal. 2004, 221,
1-10.
(
10) Arstad, B.; Kolboe, S. J. Am. Chem. Soc. 2001, 123, 8137-8138.
(11) Bjørgen, M.; Olsbye, U.; Kolboe, S. J. Catal. 2003, 215, 30-44.
(
(
12) Song, W.; Fu, H.; Haw, J. F. J. Am. Chem. Soc. 2001, 123, 4749-4754.
13) Arstad, B.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 2004, 123, 2991-
3001.
4,15
(
14) Haw, J. F.; Nicholas, J. B.; Song, W.; Deng, F.; Wang, Z.; Xu, T.;
Heneghan, C. S. J. Am. Chem. Soc. 2000, 122, 4763-4775.
(15) Goguen, P. W.; Xu, T.; Barich, D. H.; Skloss, T. W.; Song, W.; Wang,
Z.; Nicholas, J. B.; Haw, J .F. J. Am. Chem. Soc. 1998, 120, 2650-2651.
(
16) Dessau, R. M. J. Catal. 1986, 99, 111-116.
(17) Svelle, S.; Rønning, P. O.; Olsbye, U.; Kolboe, S. J. Catal. 2005, 234,
385-400.
9
membered rings are involved in alkene formation.
The higher 1 C content in the C
3
3
6
-C alkenes (Figure 2a)
compared to ethene (and the aromatics) must be attributed to
13
C
JA065810A
J. AM. CHEM. SOC.
9
VOL. 128, NO. 46, 2006 14771