Enhanced para-xylene selectivity in the toluene alkylation reaction at ultralow contact time

Article abstract of DOI:10.1021/ja046421o

Dramatic improvements in the para-xylene selectivity of the toluene alkylation reaction can be effected by operating the catalytic reaction at ultralow contact time. Unexpectedly, the rate of alkylation is sustained, while unwanted side reactions are suppressed. By demonstrating that contact time directly influences the fate of para-xylene, which is known to form and diffuse preferentially within the zeolite catalyst, we conclude that external mass transfer is a key parameter in controlling selectivity. Even non-optimized catalysts can be made to achieve near-perfect selectivity, without sacrificing conversion. Copyright

Full text of DOI:10.1021/ja046421o

Published on Web 03/17/2005
Enhanced para-Xylene Selectivity in the Toluene Alkylation Reaction at
Ultralow Contact Time
John Breen,† Robbie Burch,† Manisha Kulkarni,† Paul Collier,*,‡ and Stan Golunski‡
CenTAcat, School of Chemistry, Queen’s UniVersity Belfast, DaVid Keir Building, Belfast BT9 5AG, U.K, and
Johnson Matthey Technology Centre, Blount’s Court, Sonning Common, Reading RG4 9NH, U.K.
Received June 17, 2004; E-mail: collipj@matthey.com
We are able to achieve close to 100% para-xylene selectivity
by operating the catalyzed gas-phase methylation of toluene at high
space velocity. The near-perfect selectivity is the direct result of
low contact time between the gas stream and the catalyst, which
avoids re-isomerization of para-xylene once it has been formed
within the internal structure of the zeolite catalyst.
para-Xylene is a key intermediate in the synthesis of terephthalic
acid and terephthalates, which are used to make polyesters. Of the
competing routes to xylene, toluene methylation could ultimately
prevail, if the para selectivity of the process can be pushed well
above 90% while the price of methanol declines as expected.
Despite several phases of catalyst improvement over the last 30
years, however, the highest selectivities reported generally lie
between 80 and 90%.1 When methanol and toluene are passed
through a heated bed of ZSM-5 microporous aluminosilicate zeolite,
an equilibrium mixture of xylene isomers is formed. At temperatures
between 350 and 650 °C, the mixture is approximately 23% para-
xylene, 51% meta-xylene, and 26% ortho-xylene.2 Non-equilibrium
operation, for example by zeolite modification, typically leads to
Figure 1. Ternary xylene isomer plot illustrating the beneficial effect on
increased para-xylene yields. Techniques used routinely include
para-xylene selectivity by operating at ultralow contact time. MgO-ZSM-5
catalyst at 440 °C and molar feed ) 44.4% C6H5CH3, 5.6% CH3OH, 50%
(i) addition of an oxide, such as MgO, B2O3, or P2O5,3 (ii) pre-
coking by high-temperature anaerobic treatment with a carbon-
aceous material,2 (iii) silanization,4 and (iv) steaming.5 Each
treatment is thought to reduce the Brønsted acidity of the zeolite,
so that isomerization is suppressed, causing an inversion in the
isomerization/alkylation ratio (which is 10/1 in unmodified ZSM-
56). By constraining the channels and intersections inside the zeolite,
the modifiers also suppress the formation of the ortho and meta
isomers and enhance the relative rate of diffusion of para-xylene,
though usually at the expense of toluene conversion.2,3,7 Further
modification of the zeolite can be achieved online, by operating
the methylation reaction at relatively high temperature (>500 °C)
and at low contact time.8 These conditions promote coking,9,10 which
can result in additional improvement in para selectivity, but a
consequence of this is loss of toluene conversion and poor methanol
utilization in the formation of xylene.2,10
We have studied the alkylation of toluene using ZSM-5 (SiO2/
Al2O3 ) 80/1), modified by the addition of 10% (by weight) of
either Mg or B, which is present as the oxide in the final catalyst.
In both cases, an aqueous precursor (magnesium nitrate; boric acid)
was used to impregnate the zeolite, which was then dried (120 °C,
2 h) and calcined (500 °C, 2 h) to form the oxide on the external
and internal surfaces.
H2O, and variable H2 to alter contact time. t1 ) 2.54 s, t2 ) 1.1 s, t3 ) 0.86
s, t4 ) 0.63 s, t5 ) 0.17 s. Toluene conversion was typically between 10
and 12%.
did not greatly exceed the equilibrium level. However, we were
able to increase the para selectivity, for a fixed toluene/methanol/
steam ratio, by (a) proportionally increasing the feed rate of each
component, (b) adding a gas-phase diluent such as hydrogen or
nitrogen to the feed (see Figure 1), (c) reducing the catalyst charge
while keeping the bed size constant (by adding cordierite as an
inert solid diluent), or (d) coating a smaller charge of catalyst on
a ceramic foam substrate. The common effect in each instance was
to lower the contact time between the catalyst charge and the
reactants, as calculated by dividing the active catalyst volume by
the total feed rate (see example in Supporting Information). We
were not able to increase the para selectivity of unmodified HZSM-5
much above the equilibrium level in the same way.
Although the long-term durability of B2O3-ZSM-5 was limited
by a very slow loss of boron, it was possible to make direct
comparisons between its performance and that of MgO-ZSM-5
even after one week of testing. For a toluene/methanol mole ratio
of 8/1, the maximum possible toluene conversion by methylation
alone is 12.5%, which was achieved by B2O3-ZSM-5 with 99.9%
para selectivity at a contact time of 0.4 s. The contact time had to
be decreased by half again before MgO-ZSM-5 matched this
performance (Figure 1).
The modified zeolites were tested under conditions designed to
simulate the first stage of a known manufacturing route,11 in which
a large excess of toluene is used to limit methanol condensation to
dimethyl ether, and steam is added to inhibit methanol dehydration
to ethylene. Under these initial conditions, the yield of para-xylene
Our observation that para selectivity increases in a linear manner
with decreasing contact time contradicts previously accepted opinion
in this area, namely that the selectivity of ZSM-5 based catalysts
is largely insensitive to contact time (a view reinforced by a neural
† Queen’s University Belfast.
‡ Johnson Matthey Technology Centre.
9
5020
J. AM. CHEM. SOC. 2005, 127, 5020-5021
10.1021/ja046421o CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
velocity through the bed is raised, and the contact time lowered,
the incidence of readsorption and hence isomerization of para-
xylene on the external sites is reduced; that is, the Thiele modulus
is increased. To test this model, we examined the effect of contact
time on the isomerization of para-xylene by our catalysts in the
presence of methanol, steam, and hydrogen at our normal operating
temperature. As Figure 2 confirms, the probability of isomerization
decreases as contact time is lowered, and this trend becomes even
more pronounced at very low contact times (below about 0.3 s in
the case of MgO-ZSM-5).
Up until now, academic and patent literature in the field of
toluene methylation has indicated a tradeoff between selectivity
and activity. This has led to an emphasis on catalyst design and
modification as potential routes to achieving complete conversion
with 100% para selectivity. For example, reaction models predict
that the ideal zeolite catalyst should have an internal/external site
ratio of 10/1 or more.13 Our results show that, through a combination
of noncoking conditions and ultralow contact time, even non-
optimized catalysts can be made to perform exceptionally.
Figure 2. para-Xylene isomerization using a MgO-ZSM-5 (80) catalyst
(b para-xylene and 2 ortho- and meta-xylene). 440 °C and molar feed )
44.4% C6H4(CH3)2, 5.6% CH3OH, 50% H2O, and variable H2 to alter contact
time.
Scheme 1. Side Reactions of Methanol
Acknowledgment. This work forms part of Controlling the
Access of Reactant Molecules to Active Centres (CARMAC), which
is a program supported by the U.K.’s Engineering and Physical
Sciences Research Council. P.C. and S.G. thank Mike Capaldi,
Marianne Field, and Andrew Nunn (of Johnson Matthey Group
Patent Department) for interesting and stimulating discussions
during the writing of this communication.
network analysis of a large number of publications1). However,
interpretation of the results of previous studies is often complicated
by the fact that toluene methylation can be accompanied by side
reactions, such as the dehydration of methanol to form ethylene
(Scheme 1), which in turn can lead to coke formation on the internal
and external catalyst sites. The conditions we used (440 °C; high
steam content in feed) lay outside those normally associated with
either dehydration or coke formation. Significantly, in our experi-
ments, there was no induction phase at each new contact time and
no hysteresis between selectivity trends measured during different
cycles. Because the changes in para selectivity cannot be attributed
to online modification of our catalysts by coke or its precursors,
the underlying effect must be associated with the contact time
between the catalyst and the gas phase.
Wei12 has proposed a mathematical model for toluene alkylation,
which correlates para selectivity with the Thiele modulus (es-
sentially the ratio of the alkylation rate to the isomerization rate).
High values of the Thiele modulus can result from strong diffusional
or kinetic control of the isomerization reaction and lead to high
para-xylene selectivity. Traditionally, this has been achieved by
chemical modification of the zeolite (by Mg, P3, or coke2), which
causes site blocking and narrowing of the channels.
Supporting Information Available: Example calculation of contact
time in the toluene alkylation reaction and experimental procedure. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Petit, J.; Zupan, J.; Leherte, L.; Vercauteren, D. P. Comput. Chem. 2002,
26, 557.
(2) Kaeding, W. W.; Chu, C.; Young, L. B.; Weinstein, B.; Butter, S. A. J.
Catal. 1981, 67, 159.
(3) Minachev, K. M.; Isakov, Y. I.; Shpiro, E. S.; Isakova, T. A.; Tkachenko,
O. P.; Zelinsky, N. D. Proc. Int. Congr. Catal. 9th, 1988, 461.
(4) Halgeri, A. B.; Das, J. Catal. Today 2002, 73, 65.
(5) Ashton, A. G.; Batmanian, S.; Dwyer, J.; Elliott, I. S.; Fitch, F. R. J.
Mol. Catal. 1986, 34, 73.
(6) Young, L. B.; Butter, S. A.; Kaeding, W. W. J. Catal. 1982, 76, 418.
(7) Xie, Y.; Zhao, B.; Long, X.; Tang, Y. In Shape SelectiVe Catalysis; Song,
C., Garce´s, J. M., Sugi, Y., Eds.; ACS Symposium Series 738; American
Chemical Society: Washington, DC, 2000; p 188.
(8) Weinstein, B. U.S. Patent 3,965,207, 1976. Assigned to Mobil Oil
Corporation.
(9) Bibby, D. M.; Howe, R. F.; McLellan, G. D. Appl. Catal., A 1992, 93, 1.
(10) Sotelo, J. L.; Uguina, M. A.; Valverde, J. L.; Serrano, D. P. App. Catal.,
A 1994, 114, 273.
(11) Gentry, J. Toluene to ParaxylenesSelective Conversion Processes Lead
the Way to Quick, Low Cost PX Production; CMAI World Petrochemical
Conference (www.gtchouston.com/articles/CMAI%202000%20Joe.pdf),
March 2000.
Our proposed model for toluene methylation is consistent with
Wei’s by assuming the preferential diffusion of para-xylene through
the channel structure of modified zeolites. However, as the product
emerges it can come into contact with unconstrained (and in some
cases unmodified) sites, where isomerization can occur. As the gas
(12) Wei, J. J. Catal. 1982, 76, 433.
(13) Vayssilov, G.; Yankov, M.; Hamid, A. Appl. Catal., A 1993, 94, 117.
JA046421O
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5021