J. CHEM. SOC. FARADAY TRANS., 1994, 90(4), 659-665 Catalytic Reactions of o-Xylene and m-Xylene with Deuterium on Metal Films Robert J. Harpert and Charles Kernball* Department of Chemistry, University of Edinburgh, West Mains Road, Edinburgh, UK EH9 3JJ The exchange and deuteriation of o-xylene and rn-xylene have been followed by mass spectrometric analysis using evaporated metal films of iron, palladium, platinum or tungsten as catalysts, usually at temperatures in the range 273-350 K. Side-group exchange was rapid in all cases but the rates of exchange of ring atoms depended on the metal and were generally lower for atoms ortho to a methyl group. The formation of dimethylcyclohexanes was accompanied by further exchange, the nature of which varied with the metal.Significant amounts of trans-1,2-dimethylcyclohexanes were produced from o-xylene over palladium and iron. The mechanisms of the reactions are discussed in terms of dissociative adsorption for all types of exchange and a contribution of a ' roll-over' reaction of adsorbed cyclohexenes in the formation of dimethylcyclohexanes, particularly with palladium and iron. The experimental work described in this paper was carried out at the Queen's University of Belfast between 1962 and 1965.' Reactions of alkylbenzenes with deuterium over metal catalysts can provide interesting information about the rela- tive rates of exchange of different types of hydrogen atoms in the molecules and also about the nature of the addition process which can yield substituted cyclohexanes with a wide range of isotopic content.In the 1960s, results with both unsintered and sintered nickel films2 showed that the relative rates of exchange of the aromatic hydrogen atoms depend on steric rather than electronic factors and that hydrogen atoms ortho to alkyl side groups are less readily replaced. Studies with p-xylene on evaporated films of a number of metals3 showed considerable variation with the nature of the metal but at that stage there were doubts about whether the exchange of the aromatic hydrogen atoms occurred by a dis- sociative or an associative mechanism. For the xylenes, it is helpful to define symbols to describe the different kinds of hydrogen atoms in the molecules. We use S for the atoms in the methyl side groups and R,, R, and R, for the aromatic hydrogen atoms where the subscript indi- cates the numbers of methyl groups ortho to the hydrogen atom.p-Xylene3 proved to be a reactant which was relatively easy to study because it contains only two kinds of hydrogen atoms, i.e. S and R,, but for the same reason the amount of information that it can provide about the activity of the dif- ferent metals for the exchange of aromatic hydrogen atoms is limited. The case for looking at the behaviour of o-xylene and m-xylene as well is that both compounds contain more than one type of aromatic hydrogen atom in addition to methyl hydrogen atoms, in fact, rn-xylene contains all three types, R, to R,. Much of the work published in the last 30 years on the reactions of aromatic hydrocarbons with deuterium over metal catalysts has been limited to benzene.The reactions of benzene and deuterium have been examined on a range of metal catalysts such as evaporated Pd-Au alloy films,4 single crystal^,'.^ supported ni~kel,~ platinum/alumina8~9 and nickel/ZnO.'O Extensive studies on the labelling of aromatic compounds with deuterium or tritium have been carried out by Garnett and co-workers, often using heavy water as the source of the label and including homogeneous as well as heterogeneous catalysis."-' However, apart from this work, t Present address : Lorimont Enterprises BV, Hertog Hendriklaan 2,4817 JV Breda, The Netherlands. there is relatively little published information relating specifi- cally to the exchange reactions of alkylbenzenes with deute- rium on metal catalysts.Results for p-xylene have been reported for various supported Pt catalysts,16 and tert-butylbenzene and p-tert-butyltoluene have been examined over a series of evaporated metal films.17 The relative rates of exchange and deuteriation of toluene and deuterium have been determined for a wide range of evaporated metal films.18 In view of the limited amount of information in the literature, there seemed to be a case for re-examining the experimental data on the reactions of o-xylene and m-xylene with deuterium over evaporated metal films' to see whether the results provided a clearer indication of the mechanism of aromatic exchange and of the process of deuteriation.Experimental The apparatus and the technique for evaporating wires to make films have been described previo~sly.~.~ The essential feature of the apparatus was the connection of the reaction vessel (198 cm3) by a capillary leak to a Metropolitan-Vickers MS2 mass spectrometer so that continuous analysis of the reaction mixture could be made. The xylenes, 99.95% purity, were purchased from the National Chemical Laboratory, dried over molecular sieves and distilled in uacuo. The charge of hydrocarbon was 164 Pa which corresponded to 8.6 x lo'* molecules in the reaction vessel with a 22 : 1 ratio of deuterium :hydrocarbon. Gas chromatographic analyses were carried out on samples collected from the reaction vessel at the end of the reactions.The vessel was cooled in liquid nitrogen for 10 min and then the residual mixture of hydrogen and deuterium pumped off. The condensable compounds were then distilled to a removable trap, dissolved in several drops of isopentane and analysed using a 'Pye' argon chromatograph with a 1.22 m column of Celite (100-120 mesh B.S.S.) impregnated with 10 wt.% silicone oil and operated at 323 K with a flow rate of 60 cm3 min-'. Mass Spectrometric Analysis Analyses were made using low accelerating voltage, 17 eV, electrons in order to minimise fragmentation. Parent ions in the range of values of m/z from 106 to 116 were used to deter- mine the composition of the isotopic xylenes after correcting for natural isotopes and for fragmentation.The main frag- ment ions for which correction was made were those formed J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 by loss of one H or D atom from the xylenes; these ions were less than 14% of the parent ions. Corrections were made on the assumption of random loss of H or D atoms. Parent ions in the range rn/z 112-128 were used for the analysis of the dimethylcyclohexanes. There was no difficulty about the overlap of the mass ranges because negligible amounts of the lighter dimethylcyclohexanes were formed. The relative sensitivity of the mass spectrometer for xylene :dimethyl-cyclohexane was 4.95 for o-xylene and 5.50 for rn-xylene, respectively. Although the fragmentation of the dimethyl- cyclohexanes by loss of hydrogen or deuterium was only a few per cent, substantial quantities of fragment ions were formed by loss of CX,, CX, and CX, (X representing H or D).These ions amounted to typically 300, 50 and 1%, respec-tively, of the parent ions and gave peaks in the range rn/z 95-110. The size of these corrections did not cause major problems because in most cases rather small amounts of the heavier dimethylcyclohexanes were formed and it was still possible to obtain values for the lighter xylenes. With Pd, which catalysed exchange much more efficiently than addi- tion, the amounts of all the lighter xylenes had become negli- gible before the dimethylcyclohexanes were formed. Results The usual method of carrying out experiments involved fol- lowing the reactions for approximately 1 h at 273 K (or sometimes at a lower temperature) and then continuing for periods of ca.30 min at one or more higher temperatures. Initial or earliest-measurable distributions of exchanged xylenes obtained mainly at 273 K, are given in Table 1. Graphs of the yields of the various isotopic products with time, such as those shown in Fig. 1-5, were used to divide the hydrogen atoms in the molecules into groups A, B, C etc. according to the ease of exchange. With o-xylene over Pd at 273 K, Fig. 1, products up to D, were formed readily but further exchange was very slow. However reactions at higher temperatures showed that two further hydrogen atoms were replaced more rapidly than the final two.In contrast over Fe at 273 K, Fig. 2, products up to D, appeared readily with o-xylene. The results in Fig. 3 for W gave some indication that the number in the groups A :B :C were 2 : 6 :2. Fig. 4 and 5 show that, over both Pt and Fe, ready exchange of seven hydrogen atoms of rn-xylene occurred but that further exchange was even slower over Fe than over Pt. The clearest evidence that the 10th hydrogen atom of rn-xylene was not easy to exchange was obtained over Fe; at 325 K when 24% of D, had been formed there was no trace of Dlo. Small amounts of the D,, product were observed over Pd but only 60 h Y 0 10 20 30 40 50 t/min Fig. 1 The exchange of o-xylene on 15.9 mg Pd at 273 K at 373 K. The groupings of hydrogens according to rate of exchange are brought together in Table 2.In order to confirm that the least reactive hydrogen atom in m-xylene was in the position ortho 8o r I 60 h s v v) .-0 40 .-0 P 0 4-.-0, 20 to the two methyl 0 2 4 6 tlmin Fig. 2 The exchange of o-xylene on 12.3 mg Fe at 273 K Table 1 Initial or earliest-measurable distributions of exchanged xylenes at 273 K product (%) mean deuterium D5 D6 D, Dt3 D9 DIO content(MJ catalyst D, D2 D3 D4 o-xylene Pd 69 21 7 2 Fe” 13 11 14 16 Pt 77 11 6 2 W 83 10 3 2 rn-xy lene Pd 40 24 19 9 Feb 2 2 2 3 Pt 74 12 10 2 W‘ 27 21 13 11 (I Distribution after 1 min when Do had decreased to 80%.for reaction at 266 K when Dohad fallen to 88%. 0.5 ----1.41 17 10 0.4 4.01 1 0.4 0.1 --1.45 0.3 0.1 ---1.32 3 ----2.18 31 50 0.1 --6.06 0.2 ----1.47 7 7 4 1 -3.29 Distribution after 2 min when Do had decreased to 27%. ‘Distribution after 1 min J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 60'r \, 501 .-0 P I/ \ \c .-2 20 lo 0 5 10 15 20 tlmin Fig. 3 The exchange of o-xylene on 7.7 mg W at 273 K 0 10 20 30 t/min Fig. 4 The exchange of rn-xylene on 8.9 mg Pt at 273 K 8o r 0 4 8 12 tlrnin Fig. 5 The exchange of rn-xylene on 15.2 mg Fe at 273 K; the D, and D, species have been omitted for the sake of clarity 661 groups, some experiments were carried out with 1,3,5-tri-methylbenzene using an evaportated nickel film.Nickel was chosen for this test because it was known' to catalyse ring exchange of other alkylbenzenes without the complications of deuteriation. As expected a rapid exchange, 14 atoms (100 molecules)-' min-' (10 mg)-', took place at 273 K but was limited to the nine D atoms in the methyl groups. On raising the temperature to 423 K, 71% of the D, compound was formed but no more highly exchanged products were detected; exchange of the remaining three atoms occurred at 473 K. Determination of Rates of Exchange The methods for the determination of the rates of exchange for molecules containing different types of hydrogen atoms have been described previ~usly.~ Under favourable circum- stances it is possible to evaluate each of the terms in the equation, k, = kA + kB + kc (1) where k, is the initial rate of entry of deuterium atoms into 100 molecules of reactant, and k,, k, and k, are the corre- sponding rates for the various types of hydrogen atoms in the molecule.For example, with o-xylene over Pd which gives exchange of the side group atoms more rapidly than the first two ring atoms, the overall rate, k, is determined from the amounts of all isotopic products and the rate for the second group, k,, from the relative amounts of the D6-D8 products. The validity of the method has been tested by studies using computer-generated data'' and shown to be reliable provid- ed that the ratio of the rates of successive groups of atoms exceeds a factor of five, i.e.kdk, > 5 and k,/k, > 5. In the results reported pre~iously,'*~~'~*'~ an attempt was made to estimate the mean number of hydrogen atoms of each kind replaced initially for molecules undergoing routine reactions (i.e. MA,M, and M,) as well as the rates. Subsequent workZo showed that for fast reactions, i.e. with rates of exchange exceeding 2 or 3% min-', the methods of evaluating M were unreliable and tended to exaggerate the values. Some rates of reaction are given in Table 2. Arrhenius parameters are reported in Table 3 for the exchange of the various groupings of hydrogen atoms and they provide a convenient method of summarising the data obtained on rates of reaction; the errors were +2 kJ mol-' for E and f0.8for log A.It was not possible to measure with accuracy rates of exchange of the fastest group of hydrogen atoms at temperatures above 273 K and so on Arrhenius parameters were obtained for group A atoms except for Pt. Likewise, because of the rapid deuteriation over W no Arrhenius parameters could be derived for exchange. Deuteriation In all cases, deuteriation gave rise to a range of isotopic dimethylcyclohexanes containing a high average number of deuterium atoms; typical initial distributions are shown in Table 4 and some results in Table 2. A feature which was common to all the deuteriation results was that the dimethyl- cyclohexanes contained more deuterium than expected for a simple addition process, i.e. some further exchange took place during the formation of the saturated products.The relevant data to demonstrate this point are given in Table 5 which includes the average deuterium content of the xylenes from which the dimethylcyclohexanes were formed and the expected number of deuterium atoms which would have been gained for a simple addition process. Arrhenius parameters for deuteriation are given in Table 6. Measurement of the J. CHEM. SOC. FARADAY TRANS., 1994,VOL. 90 Table 2 Groupings of hydrogen atoms by rate of exchange and summary of rates for exchange and deuteriation at 273K initial rates no. of hydrogen /atoms (100molecules)-' atoms in groups min-' (10 mg)-' rate of deuteriation catalyst A B C D kA kB k/% min-' (10mg)-' 0-x ylene Pd 6 2 2" -23 0.01 (4x 10-3)* --55Fe 8 2 0.06 0.06 ---6 -Pt 10 0.11 W 2 6 2 -76' 27' 1.4 m-xy lene Pd 1 2d 1 85 0.01 (0.06)b Fe 2 1 -300 0.15 (0.01)b Pt 2 1 -18 0.8 2.6 W 1 --30' -18' " The value of kdk, was 10at 348K.Rates estimated by extrapolation from higher temperatures. Approximate rates. The value of kdk, was 5 at 323K. 'Rates at 266K. cis :trans ratios of the dimethylcyclohexanes were made after The basic mechanism for the exchange of aliphatic C-H the reactions had been followed at the upper temperatures, bonds in saturated molecules involves reversible dissociative see Table 6. The percentages of trans-1,2-dimethyl-adsorption, initiated by the formation of adsorbed alkyl rad- cyclohexane formed were 50% for Pd, 69% for Fe and 11% icals.Such a mechanism will also hold for side-group for the two other metals; the percentages of the trans-1,3-exchange with alkylbenzenes. However, for the exchange of dimethylcyclohexanes ranged from 8% for Fe to 15% for Pd. aryl C-H bonds, both dissociative and associative mecha- nisms have been proposed and the early literature has been reviewed.21 It has been arguedI8 that a dissociative mecha- Discussion nism might be slow because of the high bond-dissociation General Concepts energy of phenyl-H compared with benzyl-H for which values of 464 and 368 kJ mol-', respectively, have been given.22 But In order to discuss the present results adequately, it is neces- two other factors will influence the relative rates of disso- sary to review briefly a number of concepts that have been ciative processes.The first will be the strength of the metal- established for the catalytic behaviour of aromatic molecules carbon bond of the adsorbed dissociated species which is over metal catalysts. likely to be greater for metal-aryl than for metal-benzyl Table 3 Arrhenius parameters for exchange group of temperature E log[A/molecules temperature/K where catalyst of atoms /K /kJ mol-' s-l (10mg)-'] k = 1% min-' (10mg)-' o-x ylene Pd B 273-348 63 25.2 330 Fe B 273-323 51 23.6 319 Pt all atoms 273-298 40 23.5 251 m-x ylene Pd B 273-373 40 21.3 337 Pd C 323-373 28 18.6 426 Fe B 273-325 56 25.1 296 Pt B 273-298 44 23.5 275 Table 4 Isotopic composition of the dimethylcyclohexanes produced initially product (%) temperature catalyst /K D5 D6 D7 D8 D9 DIO Dll D12 D13 D14 D15 D16 1,2-dirnethylcyclohexanesfrom o-xylene Pd 348 -----7 23 26 21 11 8 4 --16Fe" 273 6 10 8 21 9 19 4 4 3 Pt 273 5 25 25 8 11 7 8 6 5 ---W 273 8 13 12 14 15 8 12 4 7 3 2 2 1,3-dimethylcyclohexanesfrom rn-xylene Pd 323 -----4 17 24 23 19 11 2 Fe 325 -------0.2 17 50 28 5 -Pt 273 -20 19 13 15 13 12 3 5 Wb 266 -11 23 14 13 12 10 7 7 3 a Distribution after 3min.Distribution after 2min. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 5 Evidence for exchange during the formation of the dimethylcyclohexanes average D contents temperature of xylenes for sample exchange of dimethylcyclohexanes catalyst /K converted addition with addition formed o-x ylene Pd 348 5.9" 5.1 1.5 12.5 Fe 273 3.3b 5.8 1.6 10.7 Pt 273 0 6.0 2.1 8.1 W 273 0 6.0 3.0 9.0 m-xy lene Pd 323 59' 5.1 1.8 12.8 Fe 325 7.e 5.0 2.2 14.2 Pt 27 3 0 6.0 2.5 8.5 W 266 0.4 6.0 2.6 9.0 Main xylenes were: a D, 13%, D, 82%, D, 5%. D, 54%, D, 13%, D, 25%.D, 15%, D, 79%, D, 6%. D, 19%, D, 62%, D, 18%. bonds and which will also depend on the nature of the metal. Anderson23 demonstrated that there was a correlation between the rate of exchange of ethane over various metal films and the metal-metal bond strength as measured by the latent heat of vaporisation of the metal. Tungsten with a high metal-metal bond strength was one of the most active cata- lysts and palladium one of the least active.The second factor is the activation energy involved in the dissociative adsorp- tion step involving the breaking of the C-H.bond. This will probably be less for aryl C-H than for alkyl C-H because of the close approach of the metal to the aromatic hydrocar- bon through coordination of the ring parallel to the surface. An associative mechanism for the exchange of benzene will involve the reversible formation of adsorbed cyclohexadienyl radicals but there is a complicating factor because of the likely orientation of the adsorbed species with the ring paral- lel to the surface. Two reaction paths will be needed to effect e~change.~.~'If the D atom is added to the lower side of the ring then the removal of H must take place from the upper side of the ring (or vice uersa, upper addition of D and lower removal of H).There is general agreement24-27 that the rate-determining step in the hydrogenation (or deuteriation) of benzene is the formation of adsorbed cyclohexadiene because subsequent steps will be rapid. However, stereochemical considerations are also important3p2' because the rings of all the interme- diates from adsorbed benzene to adsorbed cyclohexyl species are likely to be oriented more or less parallel to the surface and this leads to the concept of 'simple deuteriation'.2* This refers to the addition of deuterium to the aromatic com- pound without any accompanying exchange of the hydrogen atoms in the benzene and the product is mainly D6-cyclohexane.No exchange or redistribution of the 6 H atoms in the reactant occurs because addition of the D atoms is assumed to take place on the lower side of the ring of the various intermediates. Interconversion between adsorbed Species such as C6H6D, and C6H6D,+' may Well OCCUr but because of the orientation of the species no replacement of the H atoms takes place. Results for the C6H6-D2 reaction at low temperatures over a number of metal films conformed closely to this behaviour.28 For xylenes, the concept of 'simple deuteriation' implies the absence of any exchange of the four ring atoms and the formation of only cis-dimeth ylcyclohexanes. There are many example^,^',^^-^^ h owever, which show that other processes play a part in the conversion of aro-matics to the corresponding cyclohexanes as well as 'simple addition'.There is evidence for the formation of cyclohexenes as intermediate products and for xylenes, the production of some dimethylcyclohexenes can provide a route to the forma- tion of trans-dimethylcyclohexanes. The extent of interme- diate cyclohexene formation is dependent on the reactant, the temperature, the nature and the dispersion of the metal, and also on the nature of the support. However, it is important to emphasise that the actual desorption of a cyclohexene inter- mediate product is not essential to bring about ring exchange alongside deuteriation or the formation of trans-products. The concept of a 'roll-over' process which involves the 'turnover' of an adsorbed alkene on the surface and was defined in relation to the exchange of cycloalkanes with de~terium~~,~'provides an adequate mechanism for both further ring-atom exchange and the formation of trans-dimet hylcyclohexanes.Exchange The results in Table 1 and 2 show the characteristics of each metal for the exchange of the two xylenes but it is also useful Table 6 Arrhenius parameters for deuteriation catalyst temperature /K temperature/K where k = 1% min-' (10 mg)-' o-x ylene Pd 298-348 36 19.7 423 Fe 273-373 40 21.5 331 Pt 298-348 32 20.3 325 W 273-298 24 19.9 264 rn-xylene Pd 323-373 36 20.8 338 Fe Pt 325-348 273-298 (46)"33 22.0(22)" (3 34)" 256 a Approximate values. to look at the relative rates for the S, R,, R, and R, hydro- gen atoms, derived from those in Table 2 and given in Table 7.Palladium shows an exceptionally high value for the exchange ratio S/R,, i.e. good activity for side-group exchange and relatively poor catalysis even for unhindered ring positions. The other three metals give similar rates for S and R, hydrogen atoms, except that the R, atoms in o-xylene exchange more rapidly than the side-group atoms over tung- sten. Thus, the general pattern for ring exchange relative to side-group exchange is W > Pt = Fe % Pd. This sequence is consistent with a dissociative mechanism for aryl exchange assuming that the rate of reaction will be faster for metals having greater metal-carbon bond strengths which in turn will parallel metal-metal bond strengths.A recent review3* has drawn attention to the trend in surface reactivity of benzene on moving across the transition series-the sequence running from non-dissociative adsorption, through partial dissociation, to complete dissociation on W(100) at low coverage.39 Cinneide and Clarke4 claimed further evidence to support a dissociative mechanism for benzene exchange over palladium-gold alloy films. They argued that exchange and addition had to involve different mechanisms because deu- teriation fell to zero at median alloy compositions whereas exchange persisted to Pd-lean alloys. Iron is the only metal to give a high value for the exchange ratio R,/R, and the other three metals show ratios between 1 and 20.All metals, as would be expected for a dissociative mechanism, show very low activity for the exchange of the R, hydrogen atom located between the methyl groups of rn-xylene. The earlier work3 with p-xylene showed that with tungsten and platinum there was a tendency to complete the exchange of the first methyl group to react because similar amounts of D, and D3 compounds were found in the initial distributions of products. Table 1 shows that a substantial degree of multi- ple exchange occurs, particularly with iron and for rn-xylene over tungsten although these reactions are too fast to permit any deductions to be made about the mechanisms. However, the similar amounts of D, and D3 products over platinum with m-xylene, and to a lesser extent with o-xylene, provide further evidence for a tendency to complete the exchange of the first group to react.These facts suggest a possible role for an a,a-diadsorbed species, =CHC,H4CH3, in the overall mechanism of the exchange in addition to adsorbed benzyl radicals. Deuteriation Tables 4 and 5 provide evidence of further exchange during deuteriation with both reactants over all four metals but the results can be divided into two groups. With palladium and iron, side-group exchange was substantially complete before Table 7 Relative rates of exchange for different types of hydrogen atoms at 273 K catalyst react ant S/R 0 Pd o-xylene 2300 10" Pd Fe m-xylene 0-xylene 8500 ca. 1 5b lo00 Fe m-xylene ca.1 2000 Pt o-xylene ca. 1 ca. 1 Pt m-x ylene ca. 1 1:a. 20 W W o-xylene m-xvlene' 0.3 ca. 1 >3 ca. 1 a At 348 K. At 323 K. 'At 266 K. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 deuteriation began and the additional exchange which accompanied deuteriation involved replacement of ring hydrogen atoms: both of these metals gave significant amount of products in the D13 to D,, range. On the other hand, with platinum and tungsten an appreciable fraction of the exchange which accompanied deuteriation must have been associated with side-group exchange and only relatively small amounts of the high exchanged D13 to D,, products were formed. The cis :trans ratios of 1,2-dimethylcyclohexanesprovide further evidence of a difference of behaviour between palla- dium and iron, and the other two metals.High percentages of trans- 1,2-dimethylcyclohexane were observed over palladium and iron but only low percentages were formed over plati- num and tungsten. These results suggest that roll-over of the various substituted cyclohexenes, leading to both ring exchange and the formation of trans- products, occurs more readily with palladium and iron, than with platinum and tungsten. However, temperature may also be a factor since most of the results in Table 4 for palladium and iron were taken at higher temperatures than with platinum and tung- sten. Recent results4' on the exchange of methylcyclopentane with deuterium have confirmed that increase of temperature favours roll-over and that the chance of this reaction depends on the metal with the sequence being Pd > Pt > Rh.The cis : trans ratios for the 1,3-dimethylcyclohexanesare less useful because the cis-product is the more stable form and so there is less tendency for trans-formation to occur. Other comments can be made about some of the features of the distributions in Table 4. The maxima at D,, and D13 for the 1,2-dirnethylcyclohexanesformed over iron probably arise from complete exchange of one methyl group and either two or four ring atoms as well as the six D added. A similar interpretation would apply to the results for tungsten where there was a partial cut-off in the distribution after the D,, product. In most cases, the rates of deuteriation were a factor of 10 faster for rn-xylene than for o-xylene although the results for iron were different.The pattern of activity of the metals for deuteriation was W > Pt > Fe GZ Pd, which is similar to results for benzene,' and toluene.18 For most systems, the energies for deuteriation were in the range 30-40 kJ mol-'. These values are somewhat smaller than those reported for various supported palladium catalysts27 for which the activa- tion energies were in the range 46-62 kJ mol-'. The turnover frequencies for reactions over palladium evaluated at 413 K using data on the area of the films41 are 1.4 x s-l for o-xylene and 17 x s-' for rn-xylene. These are close to the values reported by Rahaman and Vannice for palladium powder of 4.9 x s-' and 27 x s-' for the two xylenes measured using pressures CQ.10-fold greater than those in the present work. Conclusions These investigations establish the relative activities of the four metals (Fe, Pd, Pt and W) for both exchange and deu- teriation of the xylenes without complications associated with the supports of supported-metal catalysts, The results for the three xylenes, taken together, provide comparisons between the rates of exchange of side-group hydrogen atoms and the rates for the various types of ring atoms, classified by the number of neighbouring methyl groups. The variations exhibited by the metals, for the rates of exchange of the different hydrogen atoms are consistent with a dissociative mechanism for aryl-H exchange.The conversion of the xylenes to dimethylcyclohexanes always involves more than a 'simple deuteriation'. Some J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 665 additional exchange of ring atoms accompanied by the for- mation of trans-1,2-dimethylcyclohexanesfrom o-xylene, par- ticularly over palladium and iron, is also found. The results are interpreted in terms of a roll-over mechanism involving the turning over of the various dimethylcyclohexenes on the catalyst surface; in some cases these may appear as interme- diate products but this is not essential to account for the nature of the dimethylcyclohexanes formed. 15 16 17 18 19 20 M. A. Long, J. L. Garnett and P. G. Williams, J. Chem. SOC., Perkin Trans.2, 1984, 2105. J. W. Hightower and C. Kemball, J. Catal., 1965,4, 363. R. J. Harper, S. Siegel and C. Kemball, J. Catal., 1966, 6, 72. C. Horrex, R. B. Moyes and R. C. Squire, in Proc. 4th Int. Congr. Catal., Akademiai Kiado, Budapest, 1971, vol. 1, pp. R. S. Dowie, P. J. Robertson and C. Kemball, J. Catal., 1974,35, 189. R. A. Rajadhyaksha, unpublished results. 332-340. We thank Dr. R. Brown and Dr. G. S. McDougall for helpful discussion. R.J.H. held a research studentship awarded by the 21 22 S. Siegel, Adv. Catal., 1966, 16, 123. D. F. McMillen and D. M. Golden, Annu. Rev. Phys. Chem., 1982, 33,493. Ministry of Education, N. Ireland and the mass spectrometer 23 J. R. Anderson, Rev. Pure Appl. Chem., 1957,7, 165. was purchased by a grant from DSIR, London.24 25 26 J. Horiuti and M. Polanyi, Trans. Faraday SOC.,1934,30, 1164. A. Farkas and L. Farkas, Trans. Faraday SOC., 1937,33,837. W. F. Madden and C. Kemball, J. Chem. SOC., 1961,302. References 27 M. V. Rahaman and M. A. Vannice, J. Catal., 1991, 127, 251, 1 2 3 4 5 6 7 8 9 10 11 12 13 R. J. Harper, Ph.D. Thesis, The Queen’s University of Belfast, 1965. E. Crawford and C. Kemball, Trans. Faraday SOC., 1962, 58, 2452. R. J. Harper and C. Kemball, in Proc. 3rd Int. Congr. Catal., ed. W. M. H. Sachtler, G. C. A. Schuit and P. Zwietering, North- Holland, Amsterdam, 1965, vol. 2, pp. 1145-1 156. A. 0.Cinneide and J. K. A. Clarke, J. Catal., 1972, 26,233. J. Massardier, G. Dalmai-Imelik and J. Barbier, C. R. Hebd. Seances Acad. Sci.Ser. C, 1976,283,257. M. Surman, S. R. Bare, P. Hofmann and D. A. King, Surf. Sci., 1983,126,349. R. van Hardeveld and F. Hartog, in Proc. 4th Int. Congr. Catal., Akademiai Kiado, Budapest, 1971, vol. 2, pp. 295-308. R. Maurel and J. Barbier, J. Chem. Phys. Phys.-Chim. Biol.,1976, 73, 995. A. Morales, J. Barbier and R. Maurel, Acta Cient. Venez., 1979, 30,377. B. S. Gudkov, Kinet. Katal., 1979, 20, 668. J. L. Garnett, Catal. Rev., 1971, 5, 229. J. L. Garnett and R. J. Hodges, J. Am. Chem. SOC., 1967, 89, 4546. J. L. Garnett and R. S. Kenyon, Aust. J. Chem., 1974, 27, 1023, 28 29 30 31 32 33 34 35 36 37 38 39 40 41 267. J. R. Anderson and C. Kemball, Adu. Catal., 1957,9, 51. F. Hartog and P. Zwietering, J. Catal., 1963,2,79. S. Siegel, V. Ku and W. Halpern, J. Catal., 1963,2, 348. S. Siegel, J. Outlaw Jr. and N. Garti, J. Catal., 1979,58, 370. M. Viniegra, G. Cordoba and R. Gomez, J. Mol. Catal., 1990, 58, 107. N. Martin, G. Cordoba, A. Lopez-Gaona and M. Viniegra, React. Kinet. Catal. Lett., 1991,44, 381. R. Gomez, G. Del Angel and V. Bertin, React. Kinet. Catal. Lett., 1991, 44, 517. G. Del Angel, V. Bertin, P. Bosch, R. Gomez and R. D. Gonza-lez, New. J. Chem., 1991, 15,643. R. L. Burwell Jr., Acc. Chem. Res., 1969,2, 289. H. A. Quinn, J. H. Graham, M. A. McKervey and J. J. Rooney, J. Catal., 1971, 23, 35. B. F. G. Johnson, J. Lewis, M. Gallup and M. Martinelli, Faraday Discuss, 1991,92, 241. A. K. Bhattacharya, J. Chem. SOC., Faraday Trans. I, 1980, 76, 126. R. Brown, A. S. Dolan, C. Kemball and G. S. McDougall, J. Chem. SOC.,Faraday Trans., 1992,88,2405. C. Kemball, Proc. R. SOC.London, Ser. A, 1952,214,413. 1033. 14 J. L. Garnett, M. A. Long, A. B. McLaren and K. P. Peterson, J. Chem. Soc., Chem. Commun., 1973,749. Paper 3/05611B ;Receiued 17th September, 1993