J. Chem. Sac., Faraday Trans. I, 1984, 80, 1689-1704 Sulphur-35 Radiotracer Studies of the Effect of Hydrogen Sulphide on a Molybdenum Disulphide Catalyst in the Hydrogenation of Buta- I ,3-diene BY KENNETH C. CAMPBELL,* MOHAMMAD L. MIRZA, SAMUEL J. THOMSON AND GEOFFREY WEBB Department of Chemistry, University of Glasgow, Glasgow G12 SQQ, Scotland Received 20th June, 1983 Hydrogenation of buta-l,3-diene on a molybdenum disulphide catalyst has been studied in order to determine the change in the pattern of butene product distribution with catalyst treatment (with hydrogen, a buta-l,3-diene + hydrogen mixture, air, hydrogen sulphide or a thiophene + hydrogen mixture). Catalyst treatment with sulphur compounds (H,S or thiophene + H,) resulted in a butene product distribution closer to thermodynamic equilibrium proportions than for a catalyst treated with hydrogen or air.35S has been used as a tracer in H,S to measure the quantity of H,S necessary to change the catalytic function on fresh and hydrogen-treated catalysts and to observe the extent of sulphur isotopic exchange of gas-phase H,S with MoS,-lattice sulphur and with adsorbed H,S. Evidence is presented for the participation in the hydrogenation reaction of hydrogen present as adsorbed HS groups on the MoS, surface, and the change of catalytic function with H,S treatment has been interpreted in terms of the change in distribution of these HS groups. Extensive use has been made of thiophene in catalytic hydrodesulphurisation studies as a simple but typical heterocyclic sulphur compound. Much information has been obtained from cobalt-molybdenum sulphide systems,l> and results with other mixed sulphide systems show that it is possible to generalise to a large extent from results obtained on cobalt-molybdenum cataly~ts.~ An interesting feature to emerge from these studies comes from a consideration of the generally accepted mechanism for thiophene hydrodesulphurisation put forward by Kolboe and Amberg.* They suggested that the first step in the primary reaction pathway is C-S bond cleavage to form b~ta-1,3-diene.~ This idea opens a vast field of interest in using buta-l,3-diene as a probe, by studying its catalytic hydrogenation on an unsupported molybdenum disulphide catalyst, with particular attention being paid to the isomeric butene products. Investigation of the hydrogenation of butadiene (which term will be used throughout this paper to refer exclusively to buta-l,3-diene) over a molybdenum disulphide catalyst has shown that two distinct types of catalyst surface can be characterised in terms of the pattern of distribution of the various n-butenes in the initial products, typically as shown in table 1, where we have adopted the nomenclature of Wells and coworkers.6.The type-A surface is characteristic of a freshly prepared catalyst which has been subjected to several butadiene hydrogenation reactions. Type-B behaviour resulted when a type-A catalyst was sulphided, either by being treated with hydrogen sulphide or by being allowed to serve as a catalyst for the hydrodesulphurisation of thiophene.The rate of butadiene hydrogenation was very similar on both types of surface. 16891690 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE Table 1. Butene distributions type A type B but- 1 -ene 58% 30% trans-but-2-ene 28% 50% cis-but-2-ene 14% 20% This paper is concerned with a quantitative study in which radioactive sulphur has been used as a tracer in hydrogen sulphide to observe its uptake by the catalyst, its exchange with sulphur in the molybdenum sulphur lattice and exchange between adsorbed hydrogen sulphide and inactive gas-phase hydrogen sulphide. EXPERIMENTAL CATALYST PREPARATION Ammonium thiomolybdate was prepared by passing hydrogen sulphide into a solution of 110 g AnalaR ammonium molybdate in 400 cm3 of 36 wt% ammonia solution at 65 "C.The ammonium thiomolybdate was filtered off, dried and decomposed by being heated to 1000 "C in flowing oxygen-free nitrogen for 48 h. A yellow sublimate, presumed to be sulphur, was formed in the cooler regions of the tubing outside the furnace. The nitrogen flow was maintained while the product cooled to room temperature. Electron diffraction confirmed the product to have the molybdenite (MoS,) structure. The surface area (B.E.T.) was 4.3 m2 g-l. MATERIALS Hydrogen (British Oxygen Co.) was purified by passing it over reduced 5% Pd/W03 catalyst at ambient temperature to remove oxygen, then through anhydrous magnesium perchlorate to remove water. Hydrogen sulphide (B.D.H. Laboratory Chemicals Ltd), 99.7% pure, was degassed and purified by trap to trap distillation.Radioactive [35S]hydrogen sulphide was obtained in 5 mCi samples from Radiochemical Centre, Amersham, with a specific activity of 10 mCi mmol-'(0.37 GBq mmol-l) and was suitably diluted before use with inactive hydrogen sulphide. Buta-1,3-diene (B.D.H. Laboratory Chemicals Ltd), 99% pure, was condensed at 77 K and degassed before use. No impurities were then detectable by gas chromatography. APPARATUS The apparatus was a conventional vacuum system employing a mercury diffusion pump and consisting of a ca. 100 cm3 flat-bottomed reaction vessel in an electrically heated furnace, a mercury manometer, storage vessels for reactants, a spiral trap which could be cooled in liquid nitrogen to separate hydrocarbon products from unreacted hydrogen and sample vessels for transferring these hydrocarbons to a separate gas-chromatography apparatus.Two chromato- graphy columns in series were used, a 5 m column packed with 33% dimethylsulpholane on 30-60 mesh firebrick (to separate n-butane, but-1-ene, trans-but-2-ene, cis-but-2-ene and buta-1,3-diene) and a 9 m column containing 5% S.E. 30 silicone oil on 30-60 mesh firebrick (to resolve trans-but-2-ene and hydrogen sulphide). Helium (35 cm3 min-l) was used as carrier in conjunction with a katharometer detector, and the columns operated at 20 "C. Calibrations were made with standard samples. The effluent from the detector was passed through a flow cell to measure [35S]H,S activity by means of a Mullard MX 168/01 thin end-window Geiger-Muller counter coupled to an I.D.L. Ltd autoscaler-timer frequency ratemeter and potentiometric chart recorder.Counter reproducibility was verified and decay corrections made (35S, ti = 87 d) by frequent calibration with standard [35S]H,S samples.K. C. CAMPBELL, M. L. MIRZA, S. J. THOMSON AND G. WEBB 1691 60 h 50 20 0 10 20 30 LO 50 60 -AP Fig. 1. Variation of butene distribution with extent of reaction on a fresh molybdenum disulphide catalyst at 350 "C: 0, but-1-ene; a, trans-but-2-ene; 0, cis-but-2-ene. PROCEDURE A weighed sample of molybdenum disulphide catalyst (ca. 500 mg) was distributed over the flat bottom of the reaction vessel and evacuated for 30 min during which the temperature was raised to 350_+2 "C. A mixture of butadiene (50+ 1 Torr*) and hydrogen (l50+2 Torr) was admitted (either immediately or after catalyst treatment with H,S etc.) and the progress of hydrogenation observed with the manometer.Products were separated from unreacted hydrogen when the pressure had fallen 10 Torr and examined by gas chromatography. This standardised procedure was adopted in all experiments except where otherwise indicated in the results section. RESULTS HYDROGENATION OF BUTADIENE ON A FRESH CATALYST Five successive butadiene hydrogenation reactions were performed on a fresh sample (0.576 g) of molybdenum disulphide. The product distribution for the first reaction was but-1-ene, 30%; trans-but-2-ene, 44%; and cis-but-2-ene, 26%. During the course of the first three reactions there was then a transition to conventional type-A behaviour (but-1-ene, 58%; trans-but-2-ene, 24%; cis-but-2-ene, 18 %).After the catalyst had achieved this type-A state, further butadiene hydrogenation reactions were performed in which the products were extracted at varying pressure falls. The reactions were carried out in a random order. Within the range of pressure fall up to 30 Torr, the product distribution was found to be constant, independent of the extent of reaction (fig. 1). CATALYST PRETREATMENT WITH HYDROGEN A fresh catalyst (0.576 g) was heated to 350 "C under vacuum and was then treated with two changes of hydrogen (300 Torr) at 350 "C for a total of 4 h. A series of four standard butadiene hydrogenation reactions was carried out. The catalyst showed * 1 Torr = 101 325/760 Pa.1692 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE Table 2.Change of butadiene hydrogenation product distribution with catalyst pretreatment no. of but-1-ene but-2-ene but-2-ene type catalyst treatment results (%) (%> translcis assignment H2 H2S H,, then air H,S, then H, H,S, H,, air H,S, H,, air, H, H,S, H,, air, H,, H,S thiophene + H, thiophene + H2, then N, at 750 "C H,, air, H,S 4 5 8 7 5 4 4 10 3 4 62.6 54.6 27.9 28.4 63.2 61.9 32.9 30.0 63.4 32.7 37.4 45.4 72.1 71.6 36.8 38.1 67.1 70.0 36.6 67.3 1.54 1.18 2.42 2.44 2.34 2.75 2.15 2.47 2.00 1.62 A B reproducible type-A behaviour throughout the series, with the average distribution as shown in table 2, row (a). The rapid transition of the fresh catalyst to type-A behaviour can evidently be brought about either by hydrogen or a hydrogen+butadiene mixture, which suggested that the initial different behaviour might have been due to surface oxygen acquired by exposure to the atmosphere.To investigate this, another sample of catalyst (0.570 g) was treated with hydrogen (400 Torr, 6 h, 350 "C), evacuated (30 min, 350 "C), then exposed to air (45 Torr, 2 h, 350 "C). Five standard butadiene hydrogenation reactions then gave the result shown in table 2, row (b). The but- 1 -ene and trans-but-2-ene showed a typical type-A distribution, except that the trans-but- 2-enelcis-but-2-ene ratio was considerably lower than normal. However, the results give no basis for the belief that surface oxygen caused different selectivity. CATALYST PRETREATMENT WITH HYDROGEN SULPHIDE A fresh sample of catalyst (1 .OO g) was heated under vacuum at 350 "C for 30 min.A large excess of hydrogen sulphide (20 Torr) was admitted, allowed to remain for 1 h, then pumped away and the catalyst evacuated for 30 min. Eight standard butadiene hydrogenation reactions, with analysis of products after 20 Torr pressure fall, showed a consistent type-B behaviour with product distributions as in table 2 (c). The initial rate, however, diminished during the series from 18.2 Torr min-l in the first reaction to 9.0 Torr min-l in reactions 7 and 8. The catalyst was treated with two changes of hydrogen (ca. 200 Torr) at 420 "C for a total period of 10 h, in an attempt to remove the sulphur which had been introduced as hydrogen sulphide. The catalyst was evacuated for 15 min at 420 "C and while the temperature was lowered to 350 "C.A further series of standard reactions revealed [table 2(d)] no change in the type-B behaviour. A marked change was observed in the catalytic function, however, when the catalyst was exposed to air (100 Torr) at 350 "C for 1 h followed by evacuation at 350 "C for 30 min. The catalyst had reverted to type A, and further treatment with hydrogen (200 Torr, 350 "C, 4 h) did not alter this [table 2(e) and cf)]. Re-exposure to hydrogen sulphide (20 Torr, 350 "C, 30 min) followed by evacuation (15 min) restored the type-B behaviour [table 2 ( g ) ] .K. C. CAMPBELL, M. L. MIRZA, S. J. THOMSON AND G. WEBB 1693 0 10 20 30 LO H2S uptake/pmol g-I Fig. 2. Butene product distribution as a function of hydrogen sulphide u>-ake: 0, but-1-ene; a, trans-but-2-ene; 0, cis-but-2-ene. INTRODUCTION OF SULPHUR via THIOPHENE HYDRODESULPHURISATION A fresh sample of catalyst (1.002 g) was treated with two changes of a 1 : 4 mixture of thiophene vapour and hydrogen (250 Torr, 350 O C , 1.5 h).The reaction products were not analysed at this stage. The catalyst was evacuated at 350 "C for 30 min and a series of five standard butadiene reactions performed, with analysis of products after 20 Torr pressure fall. The results, table 2(h), show type-B behaviour the same as if the catalyst had been treated with hydrogen sulphide. An attempt was then made to remove the sulphur (introduced from the thiophene) under more drastic conditions. The catalyst was placed in a vitreous silica tube in a furnace and oxygen-free nitrogen passed over it (1 h at ambient temperature, then 5 h at 750 "C).A pale yellow sublimate, presumed to be sulphur, was observed on the cooler parts of the tube. The nitrogen flow was maintained while the catalyst was cooled completely to ambient temperature; the catalyst was then transferred back to the reaction vessel. A further series of butadiene hydrogenation reactions [table 2 ( j ) ] showed that the catalyst had reverted to type-A behaviour. EFFECT OF HYDROGEN PRESSURE ON PRODUCT DISTRIBUTION Using a hydrogen-treated type-A catalyst a series of butadiene hydrogenation reactions (butadiene pressure 50 Torr) with initial hydrogen pressures varying in the range 50-280 Torr showed no variation in product distribution. Similarly, with a hydrogen sulphide-treated type-B catalyst the product distribution was independent of initial hydrogen pressure over the range 100-250 Torr.TRACER EXPERIMENTS WITH [35S]H2S QUANTITY OF HYDROGEN SULPHIDE TO CHANGE CATALYTIC FUNCTION (i) Fresh molybdenum disulphide catalyst. In these experiments a fresh sample of molybdenum disulphide (0.5704 g) was treated with successive small doses of [35S]H2S and several standard butadiene hydrogenation reactions performed after each dose. It was thus possible to determine the amount of hydrogen sulphide necessary to change1694 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE 0 5 10 15 20 H2 S u&-ike/pmol gzl Fig. 3. Butene product distribution as a function of hydrogen sulphide uptake on catalysts which had been used for butadiene hydrogenation (round symbols) or had been treated with hydrogen at 350 "C (square symbols).(Key as fig. 2.) the catalytic behaviour completely from type A to type B. The results are presented in fig. 2 in terms of the butene distribution relative to the hydrogen sulphide uptake in pmol per g of catalyst. Although 27 pmol g-l was sufficient to change the catalytic behaviour completely from type A to type B, hydrogen sulphide uptake continued beyond this amount. However, doses of hydrogen sulphide up to 27 pmol g-l were completely taken up, but beyond this limit each sample was only partially adsorbed and adsorption may have been pressure dependent. The catalyst was treated with hydrogen (300 Torr, 500 "C, 5 h), evacuated for 15 min, then the temperature was lowered to 350 "C and three successive standard butadiene hydrogenations performed. This was followed by exposure to air (25 Torr, 350 "C, 1 h), 30 min evacuation, then two further standard butadiene hydrogenation reactions which revealed a return to type-A behaviour.No 35S activity was released either in the products of the butadiene hydrogenation reactions or as a result of the treatment of the catalyst with hydrogen or air. It was also observed in the hydrogen sulphide adsorption that, after the adsorption limit had been reached, the specific activity of the hydrogen sulphide which was not taken up by the catalyst remained unchanged. This showed that there was no exchange of sulphur between the catalyst and the gas-phase hydrogen sulphide; i.e. the sulphur in the hydrogen sulphide had not been isotopically diluted by inactive sulphur from the molybdenum disulphide lattice.(ii) Used or hydrogen-treated molybdenum sulphide catalysts. Fig. 3 shows the results of other series of reactions with samples of catalysts which had been used for butadiene hydrogenation (round symbols, 0.577 g catalyst) or had been treated with hydrogen at 350 "C (square symbols, 0.575 g catalyst) before being exposed to hydrogen sulphide. The results are comparable and reveal that the transition to type-B behaviour required less hydrogen sulphide than for the fresh catalyst. Finally, a sample of fresh catalyst (0.578 g) was treated with hydrogen (400 Torr, 350 "C, 6 h), evacuated, then exposed to air (20 Torr, 350 "C, 2 h). Five standard butadiene hydrogenation reactions were performed.9.74 pmol of hydrogen sulphideK. C. CAMPBELL, M. L. MIRZA, S. J. THOMSON AND G . WEBB 1695 Table 3. Interaction of hydrogen sulphide with molybdenum disulphide catalyst expt. 1 expt. 2 expt. 3 weight of MoS,/g [35S]H,S admitted/pmol [35S]H,S taken up/pmol H, returned to gas phase/pmol H, formed: H,S taken up H,S admitted for exchange/pmol H,S remaining in gas phase/pmol relative specific activity exchangeable 35S on catalyst/pmol g-' total H,S uptake at saturation/pmol g-l 35S uptake as percentage of total H,S to saturate /pmol g-l 0.308 47.7 44.5 144.2 11.5 13.3 13.3 30.1 144.2 100 0.26 0.41 0.598 0.471 39.7 65.8 39.7 65.8 66.3 139.7 14.1 18.7 45.3 11.5 12.1 122.8 54.0 0.36 0.28 0.27 were then admitted, of which 8.85 pmol were taken up (15.3 pmol g-l) and a further four standard butadiene hydrogenation reactions carried out ; the results are shown in table 2(k).The feature of only partial uptake of hydrogen sulphide, after the transition to type-B behaviour was complete, was observed with these catalysts also. HYDROGEN SULPHIDE ADSORPTION AND ISOTOPIC EXCHANGE [35S]hydrogen sulphide was used as a tracer to study (i) the limit to the extent of hydrogen sulphide uptake by the catalyst, (ii) the amount of hydrogen produced as a result of this hydrogen sulphide uptake and (iii) the extent to which sulphur introduced to the surface by the adsorption of hydrogen sulphide could be exchanged with gas-phase hydrogen sulphide. In experiment 1 a fresh sample of catalyst was heated to 350 "C and kept at this temperature under vacuum for 1 h.It was exposed to an excess of [35S]hydrogen sulphide for 2 h and the gas-phase material was transferred quantitatively to the gas chromatograph for analysis. Non-radioactive hydrogen sulphide was then introduced into the catalyst vessel to investigate whether exchange would occur with the 35S now present on the surface from the adsorbed [35S]H,S. The hydrogen sulphide was allowed to remain in the vessel for 30 min before it was removed for analysis, and its specific activity was then determined relative to that of the original [35S]hydrogen sulphide. The results are summarised in table 3. Experiment 2 was conducted under similar conditions to experiment 1 except that the amount of [35S]hydrogen sulphide was deliberately chosen to be only about half that necessary to saturate the surface.After allowing interaction to take place for 2 h, non-radioactive hydrogen sulphide was admitted in sufficient quantity to saturate the surface and leave a small excess in the gas phase which could be examined for exchanged radioactive sulphur. These results also are summarised in table 3. Table 3 also shows the results of experiment 3 which was conducted to verify the amount of hydrogen produced by the interaction of hydrogen sulphide with the catalyst under conditions of near-saturation with hydrogen sulphide. The investigation as in experiment 1 was then conducted on a sample of molybdenum disulphide which had been pretreated with hydrogen. A fresh sample of catalyst was kept at 350 "C under vacuum for 1 h, then a measured amount of hydrogen was1696 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE Table 4.Hydrogen sulphide uptake on molybdenum disulphide pretreated with hydrogen weight of catalyst/g 0.758 1 hydrogen uptake/pmol 40.70 /pmol g-l 53.70 hydrogen sulphide admitted/pmol 99.29 hydrogen sulphide uptake pmol 24.96 /pmol g-' 32.9 admitted and allowed to remain in contact with the catalyst at 350 "C for 1.5 h. After the hydrogen uptake had been measured, the catalyst vessel was evacuated for 30 min and a measured amount of hydrogen sulphide was introduced and left for 2 h at 350 "C. Table 4 shows the results which were obtained. BUTADIENE HYDROGENATION REACTION RATES Butadiene hydrogenation on a fresh catalyst proceeded typically with an initial rate in the range 3-10 Torr min-l per g of catalyst.Treatments with hydrogen, air or hydrogen sulphide all increased this rate (but only by a factor of 2 or less). In the experiments in which the catalyst was progressively treated with several small doses of hydrogen sulphide, the gradual change in the butene distribution was accompanied by a gradual change in the hydrogenation activity. When at a certain stage of sulphur uptake no further modification in the butene distribution was observed, then at the same stage the hydrogenation activity also ceased to increase. Thereafter, on all catalysts the rate diminished gradually with the number of butadiene hydrogenation reactions performed, and this may possibly be attributable to the accumulation of carbonaceous material on the surface.Some reactivation could be brought about by treatment with hydrogen at 500 O C , but in general the activity could not be fully restored to its original value. To obtain some insight into the reliability of the initial rate measurements, a series of butadiene hydrogenation reactions was performed at various temperatures in the range 3 14-424 "C on a run-in sample of molybdenum disulphide which showed type-A behaviour. An Arrhenius plot, using the initial rates measured from pressure against time plots, gave 48.5 3.0 kJ mol-l for the energy of activation. Using this value it can readily be shown that between the limits of the temperature control in standard butadiene hydrogenation reactions (350 f 2.0 "C) a 6% variation in reaction rate would be expected.In view of the uncertainties imposed by temperature control, surface deposits and the difficulty of measuring initial rates with good accuracy from pressure against time plots, we do not attempt to attach any significance to the rather small changes in rate produced by the various catalyst treatments. However, there is certainly no marked change in hydrogenation activity accompanying the change from type-A to type-B product distributions. DISCUSSION PRODUCT DISTRIBUTIONS IN RELATION TO CATALYST TREATMENT Apart from the anomalous behaviour during the first three hydrogenation reactions performed on a fresh molybdenum disulphide catalyst, the product distributions conformed to a clear pattern. Catalysts which had been treated with sulphur showed a type-B product distribution, while those which had not gave a type-A distribution.K.C. CAMPBELL, M. L. MIRZA, S. J . THOMSON AND G. WEBB 1697 Table 5. Distribution of butene isomers but-1 -ene trans-but-2-ene cis-but-2-ene (%> (%> (%I observed (fig. 1) 21 47 calculateda 16 52 32 32 a Data from ref. (9). It was immaterial whether the sulphur was introduced by the adsorption of hydrogen sulphide, or by allowing the hydrodesulphurisation of thiophene to take place on the catalyst. Once type-B behaviour had been established by such sulphur treatment, the catalyst could be made to revert to type A by treatment with air at 350 "C or by heating it to 750 "C in a stream of oxygen-free nitrogen. Removal of elemental sulphur was evident at 750 "C. Air at 350 "C possibly removed sulphur as sulphur dioxide, although none was detected in the gas phase in the 35S-tracer experiments.A similar effect of air, but in the opposite sense (favouring but-2-ene formation), has been observed in the hydrogenation of butadiene on a nickel phosphide catalyst.* The results shown in fig. 1 reveal that the hydrogenation of butadiene on type-A molybdenum disulphide proceeds, during the course of a single reaction, with marked preference for but-1-ene formation in the early stages, but that the distribution later changes abruptly to acomposition close to the thermodynamic equilibrium proportions calculated for 350 0C,9 as shown in table 5. For the standard hydrogenation mixture used (50 Torr butadiene + 150 Torr hydrogen), a pressure fall of 50 Torr represents complete conversion of butadiene to butenes, so that the completion of the change in the product distribution corresponds to the stage in the reaction when nearly all of the butadiene has been used up.Similarly, significant amounts of n-butane were formed only when the total pressure had fallen by 45 Torr or more. Evidently the sites of activity for butadiene hydrogenation are also active for isomerisation and hydrogenation of butenes, but when both are present the butadiene competes favourably for the available sites and effectively prevents re-equilibration or further hydrogenation of the butene isomers. However, but-2-enes do occur in the initial products and the but- l-ene/but-Zene ratio is independent of the extent of reaction in the pressure-falLrange 0-25 Torr.The but-2-enes are therefore not formed by the sequence butadiene -+ but-1-ene --+ but- 2-enes, but are formed simultaneously with but-I-ene. This view is supported by the results of Wells and co-workers,1° who found identical deuterium distributions in the three butene products when butadiene reacted with deuterium over MoS, at 573-623 K. It is also consistent with our finding that the product distributions are independent of initial hydrogen pressure, which would not be the case if butadiene hydrogenation to butenes and butene isomerisation occurred simultaneously. We conclude that discrimination is in favour of but- 1-ene, and that thermodynamic equilibrium of the butenes can only occur when the amount of butadiene remaining has decreased sufficiently for significant butene readsorption to become possible. Hydrogen treatment was without effect in changing the function of either the type-A or type-B catalyst: in particular it did not interact with a type-B catalyst to form hydrogen sulphide and cause it to revert to type-A behaviour. The fresh molybdenum sulphide catalyst therefore cannot be a true sulphur-rich type-B catalyst because the change to type-A behaviour could be brought about at 350 "C by treatment with 56 FAR 11698 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE Fig.4. Disposition of sulphur atoms about a molybdenum atom. 5-S --- -- I 1 s (a) (b) Fig, 5. Molybdenum disulphide adsorption sites: (a) B or 2M site; (b) C or 3M site. hydrogen alone. Although on the fresh catalyst the but-1-ene: but-2-ene ratio was similar to type-B behaviour, the trans:cis ratio for the but-2-enes (1.69) was much nearer to the thermodynamic equilibrium value (1.63) than for typical type-B behaviour.Thus it seems likely that the fresh catalyst is really showing type-A behaviour but with subsequent isomerisation of the but-1-ene, whilst on the run-in catalyst this isomerisation has become so inhibited as to be negligible. CATALYST STRUCTURE The hydrogenation of olefins on molybdenum disulphide has been extensively studied by Tanaka and Okuharall and has recently been reviewed.12 To explain the different function of molybdenum disulphide catalysts after treatment with hydrogen sulphide we begin with their model, which was developed from the concept of coordinative unsaturation introduced by Siegel.13 We introduce the refinement of making a distinction between sulphur atoms coordinated to three molybdenum atoms (as in the bulk of the material) and those coordinated to only two or one (as on exposed edges in the layer structure).Using the symbol X to indicate a fully coordinated sulphur atom and S to indicate an unsaturated sulphur atom, the disposition of sulphur atoms about any molybdenum atom in the bulk can be represented as in fig. 4(a), but a six-coordinate molybdenum on an edge, with exposed face represented by the shaded area, will be as in fig. 4(b). Some of these S-sulphur positions are likely to be unoccupied. Two vacancies will be common and will constitute a B or 2M site, while three vacancies, as at a corner, will result in a C or 3M site (fig.5). Sulphur atoms in X situations need not be considered so it is convenient to indicate the reactive sites as a plane view of the exposed face, as in the lower part of fig. 5.K. C. CAMPBELL, M. L. MIRZA, S. J. THOMSON AND G . WEBB 1699 J--i H S Scheme 1. INTERACTIONS OF H, AND H,S WITH THE CATALYST We now make the postulate that the S-sulphur atoms are capable of bonding to hydrogen atoms and that hydrogen in such SH groups is reactive, as we have shown in an earlier p~b1ication.l~ It may be that the circumstances under which this can occur demand the presence of vacant hydrogen adsorption sites at molybdenum atoms, e.g. S S HS S and this is implicit in the interpretation of earlier work where the presence of ,MH or 3MH sites is invoked13 if these sites are formed from dissociated H, molecules.Treatment of the catalyst with hydrogen sulphide provides an alternative means of producing HS groups on the surface, and the interaction of H, with ,M and 3M sites on H,S-treated and untreated catalysts is compared in scheme 1 (the sulphur atoms originating from MoS, are ringed for clarity, but are chemically identical with those introduced from hydrogen sulphide). Points of contrast between H, and H,S adsorption (in equivalent amounts) are that H,S increases the population of HS groups on the surface more so than H,, and that H,S reduces the coordinative unsaturation with respect to sulphur of the molybdenum by introducing another sulphur atom. Since olefin hydrogenation has been inferred to occur on 3M sites,15-18 hydrogen sulphide would be expected to act as a catalyst poison by conversion of these sites to ,M or 'M, and this has indeed been found16 to be the case for ethylene hydrogenation at ambient temperatures : ethylene hydrogenation and ethylene-D, exchange were negligible after the catalyst had been exposed to hydrogen sulphide.In contrast, for the hydrogenation of butadiene at the higher temperature used in our 56-21700 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE P S ' butadiene , + but -1 -ene HS H Scheme 2. work, we found no poisoning (rather a slight increase in rate) when the catalyst was exposed to hydrogen sulphide. Two other points of contrast were apparent between our high-temperature work and the published ambient-temperature results for butadiene hydrogenation.First, at room temperature the product was found to be but-1-ene with nearly 100% selectivity until all the butadiene had reacted: only then did isomerisation to but-2-enes and hydrogenation to n-butane commence.18 However, at 350°C a yield of 40-50% but-2-ene (fig. l), the typical type-A behaviour, was observed from the beginning of the reaction. Secondly, at ambient temperature, maintenance of the molecular identity of the hydrogen was observed, i.e. [,H,]but-l-ene from D, and [,Hl]but-l-ene from HD as the sole products.ls9 l9 Deuterium, introduced to the catalyst as D,S, became more widely distributed than this when a butadiene + hydrogen mixture was allowed to interact on the catalyst at 350 OC.14 TYPE-A AND TYPE-B BEHAVIOUR : THE REACTION MECHANISM All of the foregoing observations are evidence for the greater lability of reactive hydrogen at 350 "C, and in particular we believe14 that hydrogen in HS groups takes part in the reaction.As a consequence, on our evacuated MoS, catalyst, as well as 1, 2 addition of hydrogen to give but-1-ene, some 1,4 addition to form but-2-enes can occur by the interaction of an adsorbed isobutenyl intermediate on a 3M site with hydrogen on the more distant HS groups of adjacent ,M sites, as in schemes 2 and 3. Formation of both cis and trans isomers of but-2-ene is permitted by rotation about the C2 to C3 bond in the isobutenyl adsorbed intermediate. Under the standard conditions used in our work (10-20% conversion), once the butene products have been desorbed, they cannot be readsorbed owing to competition from further butadiene adsorption ; hence a steady product distribution independent of conversion over the limits observed is possible even though this is markedly different from the thermo- dynamic equilibrium ratios for butenes.On the hydrogen sulphide-treated catalyst the possibility of 1,4 addition is considerably greater owing to the denser population of HS groups on the surface, hence the typical type-B behaviour showing amounts of but-2-enes nearer to thermodynamic equilibrium. This reactivity of hydrogen in a different adsorption situation is analogous to the participation2O9 21 of support hydrogen in hydrogenation reactions on alumina- or silica-supported metal catalysts. In the present case the two types of site are much more intimately mixed, which should facilitate the transfer processes.If under workingK. C. CAMPBELL, M. L. MIRZA, S. J . 'I'HOMSON AND G. WEBB 1701 S m H, H I fi ,SH s s S S H H S *M %l > + but -2-ene Scheme 3. conditions the S-H bond is weaker than the Mo-H adsorption bond, this raises the interesting possibility that reactions which involve transfer of hydrogen atoms will show a preference for one of these adsorption modes over the other. For example, deuterium exchange or isomerisation of but- 1 -ene proceeding via an isobutenyl intermediate might take up a hydrogen atom from an SH group and subsequently lose it, or another hydrogen atom, to form a Mo-H bond. The continuing activity of the catalyst would then be sustained by transfer of hydrogen atoms from Mo--H sites to S-H groups, and the activation energy necessary might be such that this would be the rate-determining step of the overall process.This is a possible explanation of why we observe large changes in product distribution without any significant change in reaction rate. To explain the lack of a poisoning effect by hydrogen sulphide in spite of the reduction in coordinative unsaturation of the molybdenum, we think it likely that sulphur which is not coordinated to three molybdenum atoms is sufficiently loosely held at 350 "C for the butadiene to compete adequately for adsorption sites; i.e. butadiene is able to create 3M sites by promoting the diffusion of sulphur. We envisage this diffusion as taking place between neighbouring sites rather than as an extensive, long-range mobility, because this is ruled out by the exchange results to be discussed later.There is considerable evidence in support of this. (1) On the basis of experimental observations a similar process has been 23 to explain the hydrogenation of acetylene on a sulphided nickel catalyst at 119 "C. (2) At 500-800 "C molybdenum disulphide is an effective catalyst for the decomposition of hydrogen s ~ l p h i d e ~ ~ to hydrogen and elemental sulphur. (3) Even at 350 "C we have observed hydrogen to be produced slowly as a consequence of hydrogen sulphide adsorption on molybdenum disulphide (table 3). (4) At 350 "C gas-phase hydrogen sulphide will exchange with radioactive sulphur previously taken up on the surface as H,S to the extent of ca.30 pmol g-l (table 3), and this amount is in excess of the amount required to change the catalyst function from type A to type B. This is discussed in the next section in relation to the hydrogen sulphide adsorption and exchange results. HYDROGEN SULPHIDE ADSORPTION AND EXCHANGE A sample of molybdenum disulphide which had attained the steady type-A state (either by being used as a butadiene hydrogenation catalyst or by treatment with hydrogen) showed a limit to its hydrogen sulphide uptake of 32.9 pmol g-l (table 4). In contrast, much more hydrogen sulphide was taken up by a fresh catalyst1702 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE (> 120 pmol g-l, table 3). Hydrogen sulphide uptake is thus markedly reduced by the preadsorption of hydrogen, which lends support to the idea of heterolytic dissociative adsorption H2S+H + SH for which there is ample other e ~ i d e n c e .~ ~ - ~ ~ On the used catalyst ca. 15 pmol g-l was sufficient hydrogen sulphide to bring about the change of function from type A to type B (fig. 3), whereas ca. 27 pmol g-l was required for a fresh catalyst: the first 8 pmol g-l of this was taken up by the catalyst before it had been exposed to the reaction mixture containing hydrogen. We think it likely that the extra hydrogen sulphide necessary to bring about the change in function on the fresh catalyst is dispersed over sites which are not available on the run-in catalyst. The fresh and the used catalysts have the common feature that further hydrogen sulphide samples were not taken up completely once adsorption had proceeded beyond the limit necessary to produce a complete change in the catalytic function.This, together with the fact that < 25% of the sulphur introduced to the catalyst as [35S]H2S is exchangeable with gas-phase hydrogen sulphide, suggests that there is more than one type of hydrogen sulphide adsorption. Comparison of the results of isotopic tracer experiments 1 and 2 gives further insight into this. If the non-exchangeable sites were occupied preferentially as the available surface was progressively filled with H2S molecules, no exchange would have been expected in experiment 2, because only 54% of the adsorptive capacity had been satisfied when the last of the [35S]H2S was taken up. In fact 3.10 pmol of [35S]H2S appeared in the gas phase in the exchange, so a strict sequential occupation of sites in order of adsorption strength does not occur.It is therefore interesting to compare the behaviour of the radioactive and non-radioactive portions of the hydrogen sulphide adsorption in experiment 2 as in the following calculations. Experiment 1 .-Surface saturated with [35,S'lH2S. Let x pmol g-l be exchangeable surface sulphur : * * H2S admitted for exchange = 13.32 pmol = 43.25 pmol g-l (catalyst). Observed relative specific activity = 0.41 x = 0.41 4 3 . 2 5 - k ~ whence inactive H2S and an excess of H2S added for exchange. x = 30.05 pmol g-l. Experiment 2.-Surface partly covered with [35sJH2S, coverage compIeted with Excess of H2S for exchange = 11.5 pmol = 19.23 pmol g-l catalyst.Let y pmol g-l of H2S introduced as [35S]H2S, and z pmol g-l of that introduced as non-radioactive H2S be exchangeable. Observed relative specific activity = 0.27 = 0.27. Y 19.23 + y + zK. C. CAMPBELL, M. L. MIRZA, S. J. THOMSON AND G. WEBB 1703 By comparison with experiment 1, assuming the exchangeable portion is proportional to the total uptake of H2S, y+z x 122.8 144.2 -=- whence whence and y + z = 25.59 = 0.27 Y 19.23+25.59 y = 12.10 pmol g-l z = 13.49 pmol g-l = 0.54 39.7 [35S]H2S adsorbed - total H,S adsorbed - 39.7+45.3 - 11.5 -- [ 35 S]H 2S exchanged total H2S exchanged - y + z = 0.47. These ratios are so similar that there is no clear evidence that there was any significant difference in the ease of exchange of hydrogen sulphide adsorbed at different coverages.It is inferred that hydrogen sulphide must become adsorbed at the first vacant site which the molecule encounters and is not extensively mobile on the surface. Accordingly, we conclude that in the surface SH groups only the hydrogen is capable of migration on the surface. We are grateful to the Ministry of Education of Pakistan for a Central Overseas Training Scheme Award (to M. L. M.). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 F. E. Massoth, Adv. Catal., 1978, 27, 265. C. H. Amberg, in Proc. Climax Third Znt. Con$ on the Chemistry and Uses of Molybdenum (Climax Molybdenum Co., Ann Arbour, 1980), p. 180. P. Grange, Catal. Rev., 1980, 21, 135. S. Kolboe and C. H. Amberg, Can. J . Chem., 1966,44, 2623. P. J. Owens and C. H. Amberg, Ado. Chem. Ser., 1961, 33, 182. M. George, R. B. Moyes, D. Ramanarao and P. B. Wells, J. Catal., 1978, 52, 486. J. J. Phillipson, P. B. Wells and G. R. Wilson, J. Chem. SOC. A, 1969, 1351. F. Nozaki and R. Adachi, J. Catal., 1975, 40, 166. F. D. Rossini et al., Selected Values of Physical and Thermodynamic Properties of Hydrocarbons and Related Compounds (American Petroleum Inst. Research Project 44, Carnegie Press, Pittsburgh, Pa, 1953), p. 737. M. R. Blake, M. Eyre, R. B. Moyes and P. B. Wells, Proc. 7th Znt. Congr. Catal., Tokyo, 1980 (Elsevier, Amsterdam, 1981), p. 591. K. Tanaka and T. Okuhara, Catal. Rev. Sci. Eng., 1977, 15, 249. P. C. H. Mitchell, in Catalysis (Specialist Periodical Report, The Royal Society of Chemistry, London, 1981), vol. 4, p. 175. S. Siegel, J. Catal., 1973, 30, 139. J. Barbour and K. C. Campbell, J. Chem. SOC., Chem. Commun., 1982, 1371. K. Tanaka, T. Okuhara, S. Sat0 and K. Miyahara, J. Catal., 1976, 43, 360. T. Okuhara, K. Tanaka and K. Miyahara, J. Catal., 1977, 48, 229. T. Okuhara, T. Kondo, K. Tanaka and K. Miyahara, J. Phys. Chem., 1977,81,90. T. Okuhara and K. Tanaka, J. Chem. SOC., Faraday Trans. I , 1979, 75, 1403. T. Okuhara, T. Kondo and K. Tanaka, Chem. Lett., 1976, 717. J. A. Altham and G. Webb, J . Catal., 1970, 18, 133. K. C. Campbell and J. Mooney, J . Chem. Soc., Faraday Trans. 1 , 1980, 76, 2332. A. Takeuchi, K. Tanaka and K. Miyahara, Chem. Lett., 1974, 171.1704 RADIOTRACER STUDIES ON MOLYBDENUM DISULPHIDE 23 A. Takeuchi, K. Tanaka, 1. Toyoshima and K. Miyahara, J . Catal., 1975, 40, 94. 24 K. Fukuda, M. Dokiya, T. Kameyama and Y. Kotera, Ind. Eng. Chem. Fundam., 1978, 17, 243. 25 F. E. Massoth, J. Catul., 1975, 36, 164. 26 C. J. Wright, S. Sampson, D. Fraser, R. B. Moyes, P. B. Wells and C. hekel, J. Chem. SOC., Faraday 27 P. Ratnasamy and J. J. Fripiat, Trans. Faraduy SOC., 1970, 66, 2897. Trans. 1, 1980, 76, 1585. (PAPER 3/ 1042)