Reactions of Halobenzenes with Methanol on the Microporous Solid

5946

Langmuir 2004, 20, 5946-5951

Reactions of Halobenzenes with Methanol on the Microporous Solid Acids HBeta, HZSM-5, and HSAPO-5: Halogenation Does Not Improve the Hydrocarbon Pool David M. Marcus, Weiguo Song, Saifudin M. Abubakar, Emma Jani, Alain Sassi, and James F. Haw* Contribution from the Loker Hydrocarbon Research Institute and Department of Chemistry, University of Southern California, University Park, Los Angeles, California 90089-1661 Received October 19, 2003. In Final Form: February 25, 2004 The reactions of fluorobenzene, 3-fluorotoluene, and three isomers of difluorotoluene, chlorobenzene, and bromobenzene with excesses of methanol were investigated on the large-pore catalysts HBeta (*BEA) and HSAPO-5 (AFI), and on the medium-pore HZSM-5 (MFI). Flow reactor studies in pulse mode with GC-MS detection revealed that the fluorobenzene derivatives were readily methylated at, for example, 375 °C, but not even pentamethylfluorobenzene was obviously active as a reaction center for methanolto-olefin (MTO) catalysis. Carbon-labeling studies revealed that small amounts of methylbenzenes were formed by defluorination, and these aromatic hydrocarbons seemed to account for the small yields of olefins (and their secondary reaction products) observed. Loss of one fluorine was also evident in the products for one of the difluorotoluene isomers. On HSAPO-5 the activity order for ring-methylation of halobenzenes was F > Cl . Br. On HZSM-5, chlorobenzene and especially bromobenzene lost halogen through a route forming halomethane. These largely negative results will nevertheless be useful in testing theoretical models of the detailed reaction steps in the hydrocarbon pool mechanism for MTO catalysis.

Introduction The groups of Kolboe1-5 and Haw6-10 have independently shown by a variety of methods that the mechanism of methanol-to-olefin (MTO) catalysis11,12 on zeolites and related catalysts involves a hydrocarbon pool of intermediates. The essential feature of the hydrocarbon pool mechanism is that larger organic species in the catalyst act as scaffolds onto which C-C bonds can be formed and broken without recourse to smaller, higher energy intermediates and transition states. Scheme 1 shows in a general way how this might happen for the specific example of methylbenzenes as the active components of the hydrocarbon pool. Formation of an alkyl chain (isopropyl in the example shown) is assumed to occur by either side-chain alkylation13,14 or a process of ring contraction-expansion termed the paring mechanism.15 On a working catalyst, methylbenzenes and other pool species self-assemble by secondary reactions of the primary * Corresponding author. E-mail: [email protected]. (1) Dahl, I. M.; Kolboe, S. J. Catal. 1996, 161, 304-309. (2) Dahl, I. M.; Kolboe, S. J. Catal. 1994, 149, 458-464. (3) Mikkelsen, O.; Ronning, P. O.; Kolboe, S. Microporous Mesoporous Mater. 2000, 40, 95-113. (4) Arstad, B.; Kolboe, S. Catal. Lett. 2001, 71, 209-212. (5) Arstad, B.; Kolboe, S. J. Am. Chem. Soc. 2001, 123, 8137-8138. (6) 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, 2651-2652. (7) Haw, J. F.; Nicholas, J. B.; Song, W.; Deng, F.; Wang, Z.; Heneghan, C. S. J. Am. Chem. Soc. 2000, 122, 4763-4775. (8) Song, W.; Haw, J. F.; Nicholas, J. B.; Heneghan, K. J. Am. Chem. Soc. 2000, 122, 10726-10727. (9) Song, W.; Fu, H.; Haw, J. F. J. Am. Chem. Soc. 2001, 123, 47494754. (10) Sassi, A.; Wildman, M.; Ahn, H. J.; Prasad, P.; Nicholas, J. B.; Haw, J. F. J. Phys. Chem. B 2002, 106, 2294-2303. (11) Chang, C. D. Catal. Rev. 1983, 25, 1-118. (12) Sto¨cker, M. Microporous Mesoporous Mater. 1999, 29, 3-48. (13) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal. 1983, 82, 261266. (14) Mole, T.; Bett, G.; Seddon, D. J. Catal. 1983, 84, 435-445. (15) Sullivan, R. F.; Egan, C. J.; Langlois, G. E.; Sieg, R. P. J. Am. Chem. Soc. 1961, 83, 1156-1160.

olefinic products, but they can also be formed from suitable precursors in the feed.16 In addition to methylbenzenes, other active hydrocarbon pool species include methylcyclopentenyl carbenium ions (on some catalysts) and methylnaphthalenes.17 Fundamental research objectives in MTO catalysis include a more complete understanding of the hydrocarbon pool mechanism and, in particular, factors influencing olefin product selectivity. One possible strategy for modification of the hydrocarbon pool is heteroatom functionalization. Recently we studied methylphenols as possible functionalized centers for the conversion of methanol to olefins on the silico-aluminophosphate catalyst HSAPO-34. The CHA topology of that catalyst precludes the entrance or exit of any aromatic compound, even benzene. Therefore, we devised a ship-in-a-bottle route for the synthesis of methylphenols in HSAPO-3418 based on selective oxidation of methylbenzenes (that were self-assembled in the usual way). This synthesis is described in Scheme 2. Unfortunately, even though we were easily able to form pentamethylphenol in the cages of HSAPO-34, this species would not further react to form olefins under MTO conditions. Hydroxyl is an activating group for electrophilic alkylation of benzene rings, but the detailed mechanism operating in Scheme 1 and its corresponding potential energy surface is not sufficiently well-known to predict a priori the effect of any given substituent on the rate of olefin synthesis. Furthermore, methylphenols are strong bases in the gas phase. It may be that the strategy in Scheme 2 failed because the catalyst protonated the hydroxyl group instead of catalyzing a process analogous to Scheme 1.18,19 (16) Song, W.; Marcus, D. M.; Fu, H.; Ehresmann, J. O.; Haw, J. F. J. Am. Chem. Soc. 2002, 124, 3844-3845. (17) Song, W.; Fu, H.; Haw, J. F. J. Phys. Chem. B 2001, 105, 1283912843. (18) Fu, H.; Song, W.; Marcus, D. M.; Haw, J. F. J. Phys. Chem. B. 2002, 106, 5648-5652. (19) Venuto, P. B.; Wu, E. L. J. Catal. 1969, 15, 205-208.

10.1021/la035944r CCC: $27.50 © 2004 American Chemical Society Published on Web 06/08/2004

Reactions of Halobenzenes with Methanol

Langmuir, Vol. 20, No. 14, 2004 5947 Scheme 1

Scheme 2

For the present investigation, we evaluated the possibility that halogen substituents might modify the hydrocarbon pool in some useful way. We first tried an in situ chlorination analogous to Scheme 2 to form chlorobenzenes in the cages of HSAPO-34 without success. An easier way to test mechanistic ideas about the hydrocarbon pool is to pulse gaseous mixtures of methanol and possible pool species into large pore zeolites such as the 12-ring catalyst HBeta (*BEA topology).10,20 We synthesized the 12-ring silico-aluminophosphate HSAPO-5 (AFI) so that we could also screen halogenated compounds on a catalyst with a framework elemental composition similar to the more useful catalyst HSAPO-34. Although hexamethylbenzene cannot enter the medium-pore (10-ring) zeolite HZSM-5 (MFI topology), 1,2,4,5-tetramethylbenzene can. As HZSM-5 is an important methanol to hydrocarbon catalyst, we considered this as a third material. These materials are depicted in Figure 1. We were again disappointed in our search for active, heteroatom-functionalized reaction centers. Fluoro and chloro substitutents were reasonably stable under MTO conditions, even at 400 °C, and we were able to form up to pentamethylfluorobenzene and dimethyl- (or higher) difluorobenzene on some of the catalysts. However, we also observed small amounts of methylbenzenes, and 13C labeling studies showed that their rings originated with the halobenzene; i.e., dehydrohalogenation occurred to a small extent in the catalyst. Olefins were synthesized in low yields in most experiments, but this activity was accounted for by the methylbenzenes. Thus, we conclude that halobenzenes have negligible activity as MTO reaction centers. This result suggests a critical test for future studies of the MTO reaction mechanism using theoretical modeling. Also some of the detailed observations here, including the observation of a reaction channel to methyl halides, appear to relate, albeit imperfectly, to gas-phase studies of halobenzenes promoted by gaseous Brønsted acids.21 Experimental Section Materials and Reagents. Methanol-13C (99%) was obtained from Cambridge Isotope Laboratories. Reagent grade methanol was from Mallinckrodt. Fluorobenzene (99%), chlorobenzene (99.9%), bromobenzene (99%), toluene (99.8%), 2,5-difluorotoluene (98%), 2,6-difluorotoluene (99%), and 3,4-difluorotoluene (99%), were obtained from Aldrich. Zeolyst International supplied the samples of zeolite HBeta (*BEA) and zeolite HZSM-5 (MFI). These were CP811E-75 (SiO2/Al2O3 ) 75) and CBV 3024E (SiO2/ Al2O3 ) 30), respectively. Samples of silicoaluminophospate catalyst HSAPO-5 (AFI) were synthesized using a published procedure.22 Several compositions were employed in this investigation, but the material used most frequently had an acid site (20) Sassi, A.; Wildman, M. A.; Haw, J. F. J. Phys. Chem. B. 2002, 106, 8768-8773. (21) Speranza, M.; Cacace, F. J. Am. Chem. Soc. 1977, 99, 30513055.

Figure 1. Illustration comparing the structures of the three catalysts used: (a) HBeta (*BEA) has two distinct sets of intersecting channels defined by 12-member rings (6.6 × 6.7 Å and 5.6 × 5.6 Å). (b) HSAPO-5 (AFI) has one-dimensional channels formed by 12-member rings (7.3 × 7.3 Å). (c) HZSM-5 (MFI) has two perpendicular sets of intersecting channels defined by 10-member rings (5.1 × 5.5 Å and 5.3 × 5.6 Å). In each case the view is down a channel (three are shown to suggest other details of the structure). density of ca. 0.4 mmol g-1, and hence, it was most closely related to the HBeta sample with a SiO2/Al2O3 ratio of 75. Catalysis. In common with many of our other recent studies, we used a benchtop microreactor system for all experiments (22) Lok, B. M.; Messina, C. A.; Patton, R. L.; Gajek, R. T.; Cannan, T. R.; Flanigen, E. M. U.S. Patent 4,440,871, 1984.

5948

Langmuir, Vol. 20, No. 14, 2004

Figure 2. Comparison of the reactivities of either toluene or one of the monohalobenzenes with excess methanol on HSAPO-5 at 375 °C. These GC-MS total ion chromatograms identify the volatile products exiting the catalyst beds following pulsed introduction of 10:1 mol ratios of methanol to benzene derivative. (a) Toluene and methanol yielded xylenes, pentamethylbenzene, and hexamethylbenzene. (b) Fluorobenzene was methylated up to pentamethylfluorobenzene, but methylbenzenes also formed. (c) Chlorobenzene and methanol were less reactive, forming only chlorotoluenes and chloroxylenes. (d) Bromobenzene was almost completely unreactive with methanol on this catalyst. reported here. All metal components contacting the catalyst, reactants, or products were stainless steel. He (600 sccm) was used as the carrier gas in all experiments. The reactor (5 cm long, 0.65 cm internal diameter) was loaded with 300 mg of fresh catalyst for each experiment, and the catalyst was rigorously calcined and activated in place immediately prior to use by heating at 600 °C for 30 min in flowing air (500 sccm), followed by 15 min at the reaction temperature (usually 375 °C) under helium. We have sometimes observed small amounts of oxidation products following the first reagent pulse onto rigorously calcined catalysts, and we have found that a water pulse after calcination eliminates oxidation. A 10.2 µL pulse of water was delivered onto the catalyst after each rigorous calcination. All subsequent reagent injections were 10.2 µL, and these were typically 1 mol halobenzene:10 mol methanol or methanol-13C. The tubing downstream of the reactor was heated to prevent condensation. In every case, volatile products were sampled 3.6 s after reagent induction using a Valco valve. This time was selected to maximize the signals obtained with the mass spectrometric detector; no significant changes in product selectivity were observed with small variations in sampling time. Gas chromatography (Agilent 6890 Series GC system) with mass spectrometric detection (Agilent 5973) was used to analyze all reaction products. The ionization voltage was 70.1 eV and the source temperature was 230 °C. The products were separated on a 50 m Supelco DH fused silica capillary column (0.25 mm diameter, 1.0 µm film thickness). A temperature program maintained the oven temperature at 35 °C for an initial 3 min followed by a ramp of 20 °C/min to a final temperature of 290 °C.

Marcus et al.

Figure 3. GC-MS total ion chromatograms comparing the reactivities of methanol and fluorobenzene (10:1 mol ratio) on the three catalysts at 375 °C: (a) On HBeta fluorotoluenes and small amounts of fluoroxylenes and pentamethylfluorobenzene were formed. Hexamethylbenzene was also observed. (b) On HSAPO-5, there was a slightly higher yield of olefins. (c) HZSM-5 gave the highest yields of methylfluorobenzenes, and products as large as trimethylfluorobenzene exited the catalyst. Small amounts of methylbenzenes were also produced, and olefins and alkanes (the latter being secondary reaction products) are clearly evident.

Results Halobenzenes on HSAPO-5. The GC-MS total ion chromatograms in Figure 2 compare the reactions of excess methanol with toluene or halobenzenes on HSAPO-5. Solutions of methanol and various substituted benzenes (10:1 mol ratio) were injected onto freshly activated 300 mg beds of HSAPO-5 at 375 °C. A solution of toluene and methanol (Figure 2a) served as a baseline for MTO chemistry on HSAPO-5 under these conditions. Conversion of methanol and dimethyl ether was low for all experiments on this catalyst at 375 °C, but higher temperatures promoted dehalogenation (vide infra). Toluene was methylated to xylenes and tri-, tetra-, penta-, and hexamethylbenzenes. Pentamethylbenzene was the major product, while the levels of tri- and tetramethylbenzene were very low. Fluorobenzene (Figure 2b) was comparable to toluene in its reactivity with methanol, forming even pentamethylfluorobenzene, but small amounts of methylbenzenes, including penta- and hexamethylbenzene, were also present. Chlorobenzene (Figure 2c) was less reactive than fluorobenzene, but did form chlorotoluenes and chloroxylenes. Bromobenzene and methanol (Figure 2d) showed almost no reactivity on HSAPO-5. Fluorobenzene on Various Catalysts. The GC-MS total ion chromatograms in Figure 3 compare various

Reactions of Halobenzenes with Methanol

Langmuir, Vol. 20, No. 14, 2004 5949

Figure 5. Bar graphs reporting ion-mass distributions from the vicinities of the molecular ions for selected products from the reactions of fluorobenzene derivatives with methanol-13C at 375 °C. (a) Pentamethylbenzene formed from fluorobenzene on HSAPO-5 very clearly had both 12C and 13C in the ring. (b) In the same experiment as discussed previously, pentamethylfluorobenzene had exactly five 13C atoms, showing that no incorporation into the ring occurred. (c) Trimethylbenzene formed from 3-fluorotoluene on HZSM-5 very clearly had both 12 C and 13C in the ring. (d) In the same experiment as discussed previously, trimethylfluorobenzene had no 13C incorporated into the ring. Figure 4. Experiments similar to the previous ones except probing the reactions of 3-fluorotoluene with methanol on HZSM-5 at various temperatures. (a) At 350 °C, fluoroxylenes and trimethylfluorobenzene were formed. (b) At 375 °C, substantially more fluoroxylenes and trimethylfluorotoluene were produced, but small amounts of methylbenzenes were also formed. (c) At 400 °C, methylbenzenes were formed at the expense of fluoroxylenes and trimethylfluorotoluene.

catalysts for the reaction of methanol and fluorobenzene (10:1 mol ratio) at 375 °C. HBeta (Figure 3a) and HSAPO-5 (Figure 3b) produced similar ring-methylation products, with HSAPO-5 slightly being more active. HSAPO-5 produced more olefins, but also substantially more methylbenzenes, most notably pentamethylbenzene. HZSM-5 (Figure 3c) showed the highest conversion of methanol and dimethyl ether to hydrocarbon products, including secondary reaction products such as isobutane and isopentane. The yields of fluorotoluene, fluoroxylenes, and trimethylfluorobenzene were all much higher on HZSM5. However, as expected, tetra- and pentamethylfluorobenzene did not exit this medium pore zeolite (cf., Figure 1). Figure 4 reports a study of the effect of reaction temperature using the more active HZSM-5 catalyst. 3-Fluorotoluene was used as the aromatic species, and again there were 10 equivalents of methanol in each case. At 350 °C (Figure 4a) there was almost no production of olefins, and a low yield of ring-methylation products. There were no obvious levels of methylbenzenes. Increasing the temperature to 375 °C (Figure 4b) resulted in large increases in the production of olefins, alkanes, and ringmethylation products. Small amounts of methylbenzenes were observed. Repeating this experiment at 400 °C (Figure 4c) resulted in nearly complete conversion of methanol and dimethyl ether, but the yields of methylbenzenes were elevated at the expense of fluoroxylenes and trimethylfluorobenzene.

Difluorobenzenes on Various Catalysts. Similar experiments were used to study the reactions of methanol with 2,5-difluorotoluene (10:1 mol ratio) at 375 °C on the three catalysts. As with a single fluorine, HBeta was the least active catalyst, and HSAPO-5 was of intermediate activity, giving higher levels of difluoroxylenes and triand tetramethyldifluorobenzenes. In both cases, the two difluoroxylene isomers were formed in similar yields. HZSM-5 was again the most active catalyst, and in this case one of the difluoroxylene isomers was highly favored over the other. Olefin yields were negligible on all catalysts. These experiments can be found in Figure S1 of the Supporting Information. The reactions of three different isomers of difluorotoluene with methanol on HZSM-5 were studied, again at 375 °C. The three isomers did give slightly different product distributions, suggesting that equilibration did not occur on the catalyst, but the only notable difference was a somewhat greater degree of fluorine loss for 3,4difluorotoluene. Here, loss of a single fluorine was evidenced by the formation of di- and trimethylfluorobenzenes. Figure S2 of the Supporting Information details these results. Carbon Isotope Tracer Experiments. The experiments in Figure 3b (fluorobenzene on HSAPO-5) and Figure 4b (3-fluorotoluene on HZSM-5) were actually carried out using methanol-13C as the only methanol source. The bar graphs in Figure 5 report the carbon isotope distributions (near the molecular ion) for two products from each of those experiments. The pentamethylbenzene from the reaction of fluorobenzene on HSAPO-5 (Figure 5a) clearly has significant amounts of both 12C and 13C incorporated in its rings; no other interpretation can match the observed mass distribution. In contrast, the pentamethylfluorobenzene from the same experiment (Figure 5b) most simply has five (and only

5950

Langmuir, Vol. 20, No. 14, 2004

Figure 6. GC-MS total ion chromatograms comparing the reactivities of methanol and one of the monohalobenzenes with excess methanol on HZSM-5 at 375 °C. (a) Fluorobenzene, shown again for comparison, formed no fluoromethane. (b) Chlorobenzene and methanol formed small amounts of chlorobenzene. (c) Bromobenzene and methanol formed a substantial amount of bromomethane.

five) 13C atoms, as it would have if the five methyl groups came from methanol-13C, but no ring incorporation occurred. Similarly, in the reaction of 3-fluorotoluene on HZSM-5, the predominant isomer of trimethylbenzene formed clearly contained both 13C and 12C in the ring (Figure 5c) but trimethylfluorobenzene was labeled only in methyl groups (Figure 5d). Halomethane Formation. Figure 6 compares fluorobenzene, chlorobenzene, and bromobenzene on HZSM-5 at 375 °C, as usual with 10 equiv of methanol. The fluorobenzene result (Figure 6a) is familiar and is repeated only for comparison. HZSM-5 was more active for ring methylation of chlorobenzene (Figure 6b) than HSAPO-5 (Figure 2b), but the light products included chloromethane. Bromobenzene and methanol (Figure 6c) gave a substantial yield of bromomethane on HZSM-5 at 375 °C. Note that there was no aromatic product such as benzene or phenol in the volatile products to account for the stoichiometry of the reaction yielding methyl bromide from bromobenzene and methanol. Discussion The Source of the Methylbenzenes and Hence the Olefins. The methylbenzenes observed in this study could in principle have arisen from the secondary reactions of olefins formed on halogenated reaction centers, which would be a successful analogue of the unsuccessful reaction in Scheme 2. Unfortunately, the label tracer experiments show that the methylbenzenes form from the halobenzenes (or methylated halobenzenes), as these are the only abundant sources of 12C in some of the experiments. After

Marcus et al.

forming, these methylbenzenes underwent, to some degree, the ring label scrambling that is diagnostic of participation in the MTO reaction.5,10 The mechanistic significance of ring carbon scrambling is not certain; it could be interpreted to support the so-called paring mechanism (ring contraction-expansion),15 but it can be equally rationalized as the consequence of a benzyltropylium or analogous rearrangement.10 Whatever the exact interpretation, there is plenty of evidence that ring scrambling correlates with MTO activity; it is facile for hexamethylbenzene, but it happens not at all for toluene on common MTO catalysts. The observations in Figure 5 that the methylbenzenes have undergone ring scrambling suggests that it would be unproductive to interpret the low olefin yields in most of the results here in the context of primary olefin synthesis on halobenzene reaction centers. The methylbenzenes alone could besand probably aresresponsible for the olefin yields here. We can be less certain about the negative observation of carbon label scrambling into the rings of fluorobenzenes because again the mechanism of scrambling is not firmly established. It is, however, suggestive that it is not observed, just as it was not observed for methylphenols in HSAPO-34. Activity for Ring Methylation. The activity order for ring methylation F > Cl > Br does coincide with the traditional order in physical-organic chemistry, and the difluorobenzenes were less active than fluorobenzene, again as expected. We actually investigated p-difluorobenzene and methanol in our first study of the hydrocarbon pool mechanism in zeolite HZSM-5. In that case we were observing light hydrocarbon products and not the ring-methylated products. Our intent in that study was simply to use several (presumably) inactive species (heptane was another) in control experiments to contrast the addition of more active species such as cyclic olefins. Here we were not in every case able to assign the specific isomers of various products, and we would in any event expect that o-, p- vs m-directing effects would be masked by equilibration and diffusion. Catalyst Activity Order. We were surprised that the silico-aluminophosphate HSAPO-5 was generally more active than the aluminosilicate zeolite HBeta. This is the reverse of the accepted order of acid strength, and more comparisons between the activities of these two large pore catalysts are in order. HZSM-5 was the most active catalyst for methylation of fluorobenzene rings. The smaller size of HZSM-5 clearly influenced the product selectivity for many of the reactions. There are apparently not many previous studies of the reactions of halobenzenes with methanol on zeolite-type acid catalysts. One report claimed that chlorobenzene reacted with methanol to form anisole,23 but we did not see this. Other work has focused on very specific pretreatment procedures for reactions on mordenite24,25 and no useful comparison can be made with the present work. Relationship of the Present Study to Aromatic Substitution and Dehalogenation in the Gas Phase? Several experimental21,26,27 or theoretical investigations28 have been concerned with the gas-phase reactions of halobenzenes or dihalobenzenes with Brønsted acids such as, for example, CH5+, C2H5+, and H3O+. One study in (23) Palekar, M. G. J. Catal. 1992, 134, 373-377. (24) Kodama, H.; Okazaki, S. J. Catal. 1991, 132, 512-523. (25) Horie, O.; Okazaki, S. Chem. Lett. 1986, 7, 1089-1090. (26) Mason, R. S.; Parry, A. J.; Milton, M. P. J. Chem. Soc., Faraday Trans. 1994, 90, 1373-1380. (27) Tkaczyk, M.; Harrison, A. G. Int. J. Mass Spectrom. Ion Processes 1994, 132, 73-82. (28) Hrusak, J.; Schro¨der, D.; Weiske, T.; Schwarz, H. J. Am. Chem. Soc. 1993, 115, 2015-2020.

Reactions of Halobenzenes with Methanol

particular21 elucidated two reaction channels. In channel A attack of the ring by both CH5+and C2H5+ resulted in protodehalogenation (i.e., formation of HX) in the rate order F , Cl = Br. Reaction channel B involves direct attack of the acid on the halogen, and the relative rate order is F . Cl = Br. The relative extents of dehalogenation in Figure 2 fit channel B. It is not clear how we would use these ideas to explain the formation of chloromethane and bromomethane on HZSM-5 (Figure 6). If we imagine an activated methanol species undergoing electrophilic attack on the halogen (as in channel B), then decomposition of the resulting “halonium cation” would rationalize methyl halide formation, and require a (formal) phenylium (C6H5+) cation. The latter might polymerize in the zeolite. While this explanation is attractive it does not explain the absence of methylfluoride compared with the high yield of methyl bromide. Furthermore, subsequent reactions in the gas phase lead not to methylhalides but other products.20 Nevertheless, the comparison here between gas phase and zeolite chemistry is instructive, both as to the power and limitations of the analogy. Conclusions Halogen substitutents lower rather than raise the activity of benzene rings for MTO reactions. On the positive

Langmuir, Vol. 20, No. 14, 2004 5951

side, this result could be an important test for theoretical modeling, as a correct detailed mechanism (modeled satisfactorily) must reproduce the experimental observation that one fluorine on the ring substantially reduces MTO activity, even with five methyl groups also present. Given that we earlier showed that -OH groups also deactivate a ring for MTO,18 it seems increasingly unlikely that a useful substitutent will be found (other than methyl). Considering the usual substitutents for electrophilic aromatic substitution reactions and factoring in some instinct about solid acid catalysis (e.g., -NH2 may be protonated, aldehydes will react, etc.) the best opportunity might be to test -C6H5 as a substitutent, for example, using methylbiphenyls as reaction centers. Acknowledgment. This work was supported by the National Science Foundation (CHE-0205939). Supporting Information Available: Two figures, Figures S1 and S2, of GC-MS total ion chromatograms detailing the reactions of difluorobenzenes on various catalysts. This material is available free of charge via the Internet at http://pubs.acs.org. LA035944R