Carbonylation of benzene in a zeolite catalyst [22]

Full text of DOI:10.1021/ja970325o

J. Am. Chem. Soc. 1997, 119, 5077-5078
5077
The data shown in Figure 1 were obtained following
adsorption of benzene and CO on pure PQ(7). The spectrum
shown in Figure 1a is the result of exposure of the zeolite to
1000 µmol/g of pure, unenriched benzene. A single, isolated
feature is observed at 129.4 ppm from TMS, similar to pure
liquid benzene at 128.5 ppm. The 0.9-ppm shift in the resonance
frequency is due to introduction of the benzene molecule into
the environment of the zeolite cavity.11 Figure 1b shows the
results obtained after exposure of the zeolite sample to 1000
µmol/g of benzene, followed by exposure to 200 Torr of carbon
monoxide. The sample was allowed to stand at room temper-
ature for 24 h prior to transfer to the NMR rotor. The only
signal observed after this treatment corresponds to unreacted
benzene, 129.5 ppm.
The system shows new features at 178.2 and 186.0 ppm on
heating to 110 °C, Figure 1c, or on heating to 150 °C, Figure
1d. These features are most likely due to the formation of
benzoic acid, as in ref 2. Even though the H-Y samples were
thoroughly outgassed under high vacuum at high temperature,
we cannot rule out the presence of trace amounts of water which
would lead to the formation of benzoic acid and account for
the observed features. One cannot interpret these features as
being due to benzene12 since we do not observe signal at these
frequencies for the case of pure benzene, as shown Figure 1a.
Carbon monoxide, either chemisorbed or freely translating in
the zeolite pores, is a another possibility.13 However, we do
not observe signal on exposure of the zeolite to pure CO, Figure
1e, probably due to desorption during sample transfer.
The results of the coadsorption experiment were different
when the PQ(7) sample was doped with AlCl3. Catalyst samples
were prepared by using a variation of published procedures,14
by transferring a small amount of AlCl3 into a Teflon autoclave
liner under an inert atmosphere. A small amount of degassed
zeolite was placed into a 5-mL Pyrex beaker, which was then
placed into the autoclave liner. The autoclave was then sealed
and heated in a convection oven at 150 °C for 90 min. After
being cooled to room temperature, a portion of the zeolite was
reintroduced to the vacuum manifold, and then evacuated to
the background pressure at room temperature. This sample was
then exposed to benzene and CO, as discussed above.
Figure 2a shows the spectrum obtained after exposure of the
AlCl3-promoted PQ(7) to 1000 µmol/g of benzene, followed
by exposure to 400 Torr of carbon monoxide. The spectrum
shows a feature at 129.3 ppm, due to pure unreacted benzene,
as well as a feature at 205.4 ppm, due to the formation of
benzaldehyde. The spectrum also shows spinning sidebands
at 265.3, 145.3, and 85.3 ppm, consistent with the magic angle
spinning frequency of 3015 Hz. Finally, Figure 2b is the result
obtained after exposure of the PQ(7) sample to pure benzalde-
hyde. The feature at 205.5 ppm is due to the carbonyl carbon
in benzaldehyde, and spinning sidebands at 265.5 and 146.7
ppm are observed as distinct features.
Carbonylation of Benzene in a Zeolite Catalyst
T. H. Clingenpeel and A. I. Biaglow*
Department of Chemistry, United States Military Academy
West Point, New York 10996
ReceiVed January 31, 1997
ReVised Manuscript ReceiVed March 25, 1997
The application of zeolites as acid catalysts in the production
of specialty chemicals has significant commercial importance.
The broad range of electrophilic agents that react with aromatic
compounds in zeolites provide many potentially useful prod-
ucts.1 It has been shown recently that CO and H2O can react
with small alkenes in the zeolite H-ZSM-5 to form carboxylic
acids.2 Under the same conditions but in the absence of water,
it was also shown that ethene, isobutene, and 1-octene react to
form ketones instead of carboxylic acids.3 Under more strongly
acidic conditions such as in the presence of HCl/AlCl3, CO
reacts directly with aromatics such as benzene.4,5 In this paper,
we demonstrate the formation of benzoic acid in the zeolite H-Y
and benzaldehyde in AlCl3-doped H-Y from the direct reaction
of benzene and CO.6
Carbonylation of benzene with carbon monoxide is known
to require the presence of Lewis and Bro¨nsted acids.4,5 Steamed
zeolites are widely believed to have both Bro¨nsted and Lewis
acid sites. Furthermore, cooperative effects between Bro¨nsted
and Lewis acid sites in zeolite catalysts have been discussed
for some time.7-9 As an example, activity in alkane cracking
reactions can be enhanced by a factor of 5 or more in H-Y that
has been dealuminated with steam versus ammonium hexafluo-
rosilicate. The H-Y sample used in this study, denoted PQ(7),
has been characterized previously9 and was shown to exhibit
this type of activity enhancement, suggesting that this sample
might be active as a carbonylation catalyst.
Samples for NMR10 were prepared by placing 50-100 mg
of the PQ(7) sample into glass sample tubes and evacuating to
a pressure of less than 1.0 × 10-5 Torr at 750 K for 12 h. After
being cooled to liquid nitrogen temperature, samples were dosed
with controlled volumes of benzene and to various pressures
of 13 C-labeled carbon monoxide. Samples were sealed in glass
tubes and allowed to stand at room temperature for varying
lengths of time or were heated prior to being transferred to NMR
rotors under an inert atmosphere. Zeolite samples containing
pure carbonyl-13C-labeled benzaldehyde for comparison were
prepared following a similar procedure.
(1) Venuto, P. B. Microporous Mater. 1994, 2, 297.
(2) Stepanov, A. G.; Luzgin, M. V.; Romannikov, V. N.; Zamaraev, K.
I. J. Am. Chem. Soc. 1995, 117, 3615.
(3) Luzgin, M. V.; Romannikov, V. N.; Stepanov, A. G.; Zamaraev, K.
I. J. Am. Chem. Soc. 1996, 118, 10890.
(4) Gattermann, L.; Koch, J. A. Ber. 1897, 30, 1622. Gattermann, L.
Ann. 1906, 347, 347. Crounse, N. N. Org. React. 1949, 5, 290. Dilke, M.
H.; Eley, D. D. J. Chem. Soc. 1949, 2601, 2613. For a review, see: Olah,
G. A.; Kuhn, S. J. In Friedel-Crafts and Related Reactions; Olah, G. A.,
Ed.; Wiley: New York, 1964; pp 1153-1256.
The assignment of the feature at 205.4 ppm in Figure 2a as
benzaldehyde is supported by results from several previous
investigations in addition to the data shown here. First, Olah
and co-workers observed the chemical shift of benzaldehyde
in magic acid solutions as 205.9 ppm.15 Second, we observed
sites in steamed H-Y which are capable of polarizing the
carbonyl group in acetone to nearly the same extent as magic
(5) Olah, G. A.; Arpad, M. In Hydrocarbon Chemistry; Wiley: New
York, 1995; pp 276-277.
(6) Kenvin, J.; Schiraldi, D. Improved Method for the Production of
Aromatic Aldehydes. U.S. Patent Application, January 15, 1997.
(7) Beyerlein, R. A.; McVicker, G. B.; Yacullo, L. N.; Zemiak, J. J. J.
Phys. Chem. 1988, 92, 1967.
(8) Lunsford, J. H. In Fluid Catalytic Cracking 2; ACS Symp. Ser. 452;
American Chemical Society: Washington, DC, 1991; pp 1-11.
(9) Biaglow, A. I.; Parrillo, D. J.; Kokotailo, G. T.; Gorte, R. J. J. Catal.
1994, 148, 213.
(11) Pfeifer, H.; Meiler, W.; Deininger, D. Annu. Rep. NMR Spectrosc.
1983, 15, 291.
(10) All NMR spectra shown here were acquired on a Bru¨ker MSL 200
spectrometer at room temperature with 1H-13C cross polarization, a proton
90° pulse of 5 µs, a contact time of 3 ms, an acquisition time of 40 ms, and
a repetition time of 3 s. The magnet was shimmed with adamantane until
a line width of less than 2.5 Hz was obtained. Adamantane was used as an
external frequency reference, and showed daily frequency variations of less
than 0.1 Hz. Lorentzian line broadening of 20 Hz was added to each
spectrum prior to the Fourier transform.
(12) The time-averaged chemical shift of the benzenium ion is 242 ppm
from TMS; see: Olah, G. A.; Schlosberg, R. H.; Porter, R. D.; Mo, Y. K.;
Kelly, D. P.; Mateescu, G. D. J. Am. Chem. Soc. 1972, 94, 2034.
(13) Anderson, M. W.; Klinowski, J. J. Am. Chem. Soc. 1990, 112, 10.
Munson, E. J.; Lazo, N. D.; Moellenhoff, M. E.; Haw, J. F. J. Am. Chem.
Soc. 1991, 113, 2783.
(14) Getty, E. G.; Drago, R. S. Inorg. Chem. 1990, 29, 1186 and ref-
erences contained therein, particularly refs 2 and 7-13.
S0002-7863(97)00325-9 CCC: $14.00 © 1997 American Chemical Society

5078 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Communications to the Editor
favorable in the AlCl3-free zeolite, while the presence of AlCl3
promotes the formation of benzaldehyde. Both pathways can
be explained by the formation of an acylium cation via initial
protonation of benzene.3 Another possibility is the formation
of a formyl cation from CO,5 followed by direct electrophilic
attack on benzene by the formyl cation to form a benzenium
ion. The NMR data shown here cannot distinguish between
these two processes.
Additional insight into the first step in the reaction mechanism
can be obtained from an examination of the relative gas-phase
proton affinities of benzene and CO. In zeolites, this approach
has proved to be useful for scaling the relative energies of
reaction intermediates and for predicting the favorable reaction
pathway when multiple pathways are possible.18 The gas-phase
proton affinity of benzene is somewhat higher than that of CO,
185.8 vs 139.0 kcal/mol.19 This suggests that the first step in
the reaction should probably be viewed as the formation of a
benzenium ion, although this interpretation clearly depends on
the nature of the active site.
We are currently examining the active sites in the Lewis acid-
promoted zeolites using a variety of techniques based on
incorporation of carefully controlled amounts of metal halides
into the zeolite. Using thermogravimetric analysis to measure
the weight uptake, we have currently been able to form
stoichiometric, 1:1 adsorption complexes of BCl3, AlCl3, and
SbF5 at the Bro¨nsted sites in the zeolite H-ZSM-5. These data,
along with preliminary computations and NMR data, suggest
that we are observing a tetrahedral complex at the Bro¨nsted
site, 1,20 similar to that reported by Markova and coworkers.21
Figure 1. Undoped H-Y with the following: (a) 1000 µmol/g of
benzene with the sample held at room temperature for 19 days (32768
scans; MAS frequency 3017 ( 10 Hz), (b) 1000 µmol/g of benzene
and 200 Torr of CO with the sample held at room temperature for 1
day (16384 scans, MAS frequency 4049 ( 10 Hz), (c) 1000 µmol/g of
benzene and 200 Torr of CO with the sample held at 110 °C for 4 h
and at room temperature for 5 days (34934 scans, MAS frequency 3004
( 10 Hz), (d) 1000 µmol/g of benzene and 457 Torr of CO with the
sample held at 150 °C for four 7-h periods over 8 days (32768 scans,
MAS frequency 3017 ( 3 Hz), and (e) 200 Torr of CO with the sample
held at room temperature for 11 days (32768 scans, MAS frequency
3017 ( 10 Hz).
It is possible that this complex is the active site in the
carbonylation reaction. We also note that the 1:1 complexes
are stable above 700 K, depending on the specific zeolite and
Lewis acid combination and pretreatment conditions. This
suggests that variations of these materials might be useful as
heterogeneous catalysts.
Finally, we recognize that trace amounts of water can react
with AlCl3 to produce HCl gas, which can then react with an
additional AlCl3 to form HAlCl4.4 This would still allow the
reaction to proceed, but would not involve participation of the
zeolite Bro¨nsted sites. Under these conditions, the system may
be thought of as a high-surface area form of AlCl3 with the
zeolite as a support. However, comparison of the signal-to-
noise ratios in Figure 2, parts a and b, suggests that 200 µmol/g
of benzaldehyde is forming, and this cannot all be accounted
for by the presence of trace amounts of water. As discussed
above, further experiments are in progress to resolve this issue.
Figure 2. (a) AlCl3-promoted H-Y with 1000 µmol/g of benzene and
404 Torr of CO. The sample was held at room temperature for 9 days
(16384 scans, MAS frequency 3014 ( 10 Hz). (b) Undoped H-Y with
210 µmol/g of pure 13 C-carbonyl-labeled benzaldehyde (16384 scans,
MAS frequency 3020 ( 10 Hz).
Scheme 1
Acknowledgment. The authors would like to acknowledge many
helpful discussions with Jeff Kenvin, David Schiraldi, Mark White,
and Tania Wessel. We would like to thank the Hoechst Celanese
Corporation and the National Science Foundation (grant CTS-95-20920)
for financial support.
acids.16 Also, by examining pure AlCl3, Nicholas and co-
workers have characterized several carbocation structures and
found the chemical shifts to be nearly identical with those seen
in magic acid solution.17
Scheme 1 illustrates the different pathways that are available
for the reaction of benzene with carbon monoxide and how the
reaction pathway depends upon the type of acid sites present
in the zeolite. The formation of benzoic acid seems to be
JA970325O
(18) Biaglow, A. I.; Sepa, J.; White, D.; Gorte, R. J. J. Catal. 1995,
154, 208.
(19) Aue, D. H.; Bowers, M. T. Gas Phase Ion Chemistry; Academic
Press: New York, 1979; Vol. 2, p 1.
(15) Olah, G. A.; Rasul, G.; York, C.; Surya Prakash, G. K. J. Am. Chem.
Soc. 1995, 117, 11211. See structure 2.
(16) Biaglow, A. I.; Gorte, R. J.; White, D. J. Catal. 1994, 150, 221.
(17) Nicholas, J. B.; Xu, T.; Barich, D. W.; Torres, P. D.; Haw, J. F. J.
Am. Chem. Soc. 1996, 118, 4202-4203.
(20) Full structural optimization at the Hartree-Fock level with the
6-31g* basis set (Gaussian 94, Version D).
(21) Markova, M. A.; Bates, S. P.; Dwyer, J. J. Am. Chem. Soc. 1995,
117, 11309.