Catalysis by Group lk Metals Part 1 ,-Reaction of Buta-1,3-diene with Hydrogen and with Deuterium catalysed by Alumina-supported Gold BY DOUGLAS A. BUCHANAN AND GEOFFREY WEBB* Chemistry Department, The University, Glasgow G12 SQQ, Scotland Received 23rd May, 1974 The reaction of buta-l,3-diene with hydrogen and deuterium has been studied using gold supported on y-alumina and a mixed boehmite+y-alumina in the temperature range 443-533 K. The reaction is completely selective for butene formation and, although both catalysts are active for the hydro- isomerisation of but-l-ene at 473 K, they do not catalyse the hydrogenation of n-butenes to n-butane. The kinetics and activation energy have been determined together with the variation in product distributions with hydrogen uptake, initial reactant pressures and temperature.The distribution of deuterium in each of the products has been determined and in the reaction with deuterium the prod- ducts contain an excess of hydrogen over that expected from hydrogen-deuterium mass balance. Buta-1,3-diene is adsorbed on the surface of the gold and subsequently reacts with hydrogen which migrates from the support to the metal. Type A hydrogen equilibrates with gas phase deuterium, whereas type B hydrogen does not react with gas phase deuterium but interacts with adsorbed hydrocarbon. Values for the sizes of the type A and B hydrogen pools are quoted. Two types of hydrogen exist on the alumina support. The low activity of gold for many catalysed reactions has been attributed to its lack of a partially filled d-band at the usual temperatures, and its consequent inability to chemisorb simple molecules.Gold will not adsorb or activate molecular hydrogen at temperatures below 473 K,I although ethylene can be chemisorbed in a weakly held form on gold surfaces.2 This chemisorption of olefins is consistent with the observa- tion that gold, when dispersed on alumina or magnesium oxide, possesses apprec- iable activity for the equilibration of benzene + cyclohexane mixtures. Gold powder will catalyse the skeletal isomerisation of ne~pentane,~ although, compared with Group VIII metal catalysts, the temperature required for this reaction is rather high being around 800 K, where thermal promotion of electrons from the 5d to the 6s level may occur.'* 4-6 Gold, when electroplated on to the surface of a palladium-silver alloy, will catalyse the hydrogenation of but-1-ene and cy~lohexene,~* but only when hydrogen atoms are provided to the gold surface by diffusion through the palladium-silver alloy.This pre-requisite, for hydrogenation activity, of the provision of atomic hydrogen at the gold surface has also been demonstrated in a study of the hydrogenation of ethylene on evaporated gold films, using the gold-catalysed dehydrogenation of formic acid as the source of hydrogen atom^.^ that in the reactions of unsaturated hydrocarbons using supported Group VIII metal catalysts, the support plays a significant role in deter- mining the catalytic behaviour. In particular the migration of adsorbed species between metal and support, and the role of hydrogen associated with the support have been discussed.In view of the apparent inability of gold to activate molecular hydrogen, the present studies were undertaken in an attempt to separate the catalytic effects of the metal and the support in hydrogenation reactions, particularly with reference to the availability of hydrogen and the migration of adsorbed species be- tween metal and support. It has been shown 134D. A. BUCHANAN AND G. WEBB 135 EXPERIMENTAL CATALYSTS Two catalyst supports were used; y-alumina (Degussa Ltd.), and y-alumina, from the same source, which had been pretreated by refluxing with a 1 mol dm-3 aqueous sodium acetate solution for 24 h, followed by washing with distilled water, until all traces of acetate ion had been removed, and finally drying in an air oven at 393-423 K.This latter support is designated alumina-A. X-ray analysis of alumina-A showed that it consisted of a mixture of y-alumina and the alumina hydrate, boehmite, although the relative proportions of each component could not be satisfactorily determined. The catalysts, containing 1 % w/w gold, were prepared by adding an aqueous solution of chloroauric acid (HAuCI,), containing the required weight of gold, to an aqueous suspension of the support. The excess water was removed by evaporation and the catalyst fmally dried in an air oven at a temperature between 393 and 423 K. During the preparation the supported salt was observed to undergo thermal decomposition resulting in the formation of a mauve coloured product containing metallic gold. In addition to the above catalysts, 1 % wlw gold supported on a-alumina (I.C.I. Ltd.) and Aerosil silica (Degussa Ltd.) were prepared.However, under the reaction conditions used in this study, neither of these catalysts was observed to possess any catalytic activity. MATERIALS Buta-l,3-diene (Matheson Co.) contained no impurities detectable by gas chromato- graphy and was merely degassed before use. Cylinder hydrogen was purifi5d by diffusion through a palladium-silver alloy thimble. Deuterium (Norsk Hydro) was of purity 99.8 atom % D and contained less than 10 p.p.m. oxygen impurity. It was used as supplied. APPARATUS A N D PROCEDURE Reactions were carried out with the catalyst, usually 0.5 g, resting on the bottom of a cylindrical Pyrex reaction vessel (capacity ca. 100 cm3).The reaction vessel was connected to a conventional high vacuum system maintained at Nm-2 or better by means of an oil diffusion pump backed by an oil rotary-pump. The catalyst was activated by treatment with three successive volumes of hydrogen, at a pressure of 2.66 x lo4 N m-2, for a total of 12 h at 523 K. The catalyst was then evacuated for 30 min and cooled to the reaction temperature. The reaction mixture was admitted to the reaction vessel and the reaction followed by the pressure fall observed on a mercury manometer. At the required pressure fall the reaction products were extracted for analysis. In those reactions where deuterium wasused, following the activation in hydrogen, the catalyst was treated with three successive volumes of deuterium (pressure 3.33 x lo4 N m-2) at 673 K, each sample of deuterium being left in contact with the catalyst for 12 h.Mass spectro- metric analysis of the deuterium after the pretreatment showed that this treatment was effect- ive in exchanging all of the exchangeabk hydrogen in the catalyst. ANALYSIS OF REACTION PRODUCTS The analysis of reaction products was effected by gas-liquid chromatography using an 8 m column packed with a 40 % w/w dispersion of hexa-2,5-dione supported on 30-60 mesh Silocel firebrick. In the experiments using deuterium each reaction product was, on elution from the gas chromatograph, condensed out of the helium carrier gas stream in a trap cooled in liquid nitrogen. The products were then analysed for deuterium content using an A.E.I.MS20 mass spectrometer. For the hydrocarbon analyses an ionizing beam voltage of 15 V was used, while for the analysis of residual deuterium a beam voltage of 70 V was used.136 B u TAD I E N E H Y D R o GEN A T I ON OVE R Au/AI,O, RESULTS THE REACTION OF BUTA-I,3-DIENE WITH HYDROGEN The variation in the distribution of reaction products with pressure fall was studied using both Au/y-alumina and Au/alumina-A at 473 K and a pressure of 6.7 x lo3 N rn-, of buta-1,3-diene and 1.33 x lo4 N rn-, of hydrogen. Under these conditions the reaction was completely selective for the formation of n-butene; no butane was ever observed as a reaction product and the reaction ceased at a pressure fall corresponding to the uptake of 1 mole of hydrogen per mole of buta-1,3-diene.The variation in the butene distribution with hydrogen uptake per mole of buta- 1,3- diene (AH) is shown in fig. l(a) and (b) for Au/y-alumina and Aulalumina-A res- pectively. X I I I I 1 I I 0 0.2 0.4 0.6 0.8 1.0 1 I I I I 0 0.2 0.4 0.6 0.8 1.0 hydrogen consumed (AH)/mol (mol diene)-l FIG. 1.-Variation of butene distribution with hydrogen uptake over Au/y-alumina (a), and Au/ alumina-A (6) at 473 K. (&4&)0 = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.33 x 104N m-2. (0, but-1-ene ; 0, cis-but-Zene ; 0, trans-but-Zene). (a) (6) time/h FIG. 2.-Variation in butene distribution with time of contact of but-1-ene+ hydrogen mixture with Au/alumina-A at 473 K. ( P ~ 4 ~ s ) o = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.33 x 104N m-2. Broken lines indicate the thermodynamic equilibrium values.(0, but-1-ene ; 0, cis-but-Zene ; (>, tvans-but-2- ene).D . A. BUCHANAN AND G . WEBB 137 These results show that the butene distribution was independent of hydrogen uptake up to AH = 0.6 for Au/y-alumina or AH = 0.90 with Au/alumina-A. The initial but- 1 -ene/but-2-ene ratio was greater with Au/y-alumina. In the latter stages of the reaction the butenes effectively competed with the buta-1,3-diene for the catalyst surface and underwent isomerisation, although no n-butane was detected. This effect was more marked with Au/y-alumina than with Aulalumina-A. If the reaction products were left in contact with the catalyst for a further 18 h, the butenes underwent extensive isomerisation, eventually attaining their thermodynamic equi- librium proportions, as indicated by the points shown at AH = 1.0 in fig.l(a) and (6) although again no n-butane was detectable. Under comparable conditions, Au/ alumina-A will effect the hydroisomerisation of but-1-ene (see fig. 2), although but-l- ene hydrogenation does not occur. It was noticeable throughout this work that the catalytic activity progressively decreased from reaction to reaction until, after about twenty reactions, a low limiting value was attained. By adopting a standard run technique, in which alternate reac- tions were carried out using a standard butene/hydrogen mixture, it was possible to correct for changes in catalytic activity and hence, using catalysts which had already been subjected to a number (usually 8-10) of reactions, to determine initial rate orders with respect to both hydrogen and buta-1,3-diene over the pressure range 6.7 x lo3 to 3.33 x 104N m-2 using each catalyst at 473 K.The results for the two catalysts were similar, the reaction following the rate expression ; This rate expression is also consistent with the observed pressure fall against time curves, which were first order with respect to the total pressure. The product distributions were independent of initial reactant pressures ; again the reaction was completely selective, no n-butane being observed under the condition used. The effect of temperature in the range 473 to 533 K (Au/y-alumina) and 443 to 533 K (Au/alumina-A) was studied using 6.7 x 103N m-2 buta-1,3-diene and 1.33 x 104Nm-2 hydrogen. Using a standard run technique to correct for variations in catalyst activity, an apparent activation energy of 36.5 4.0 kJ mol-1 was obtained for each catalyst.The reaction products were extracted at a pressure fall of (2 & 0.1) x 103N m-2 and analysed. With both catalysts the but-1 -ene/but-2-ene ratio appeared to be independent of temperature, although the trans/& ratio in the but-2-ene yield increased with increasing temperature (see table 1) THE REACTION BETWEEN BUTA-I ,3-DIENE AND DEUTERIUM The variation in the distribution of deuterium in the reaction products with pres- sure fall was determined using both Au/y-alumina and Au/alumina-A. The varia- tions of deuterated product distributions with pressure fall are shown in table 2, in which D.N. represents the deuterium number, defined as the average number of deuterium atoms per hydrocarbon molecule. With both catalysts the three n-butenes exhibit very similar deuterium distributions, although the extent of deuteration is surprisingly low.The extent of deuteration of both the butenes and the buta-l,3-diene is substantially less using Au/alumina-A than using Au/y-alumina, although the amounts of hydrogen exchange are comparable for the two catalysts. At first sight it would appear that, as expected in view of the increasing amount of hydro- gen exchange, the degree of deuteration of the butenes decreased with increasing conversion of buta-I ,3-diene. However, consideration of the sequence in which the Several interesting features emerge from these results.138 BUTADIENE HYDROGENATION OVER Au/A1,0, reactions were performed shows that any effects of pressure fall are completely masked by the variation of the extent of deuteration with reaction number.The variation is such that the deuterium number of each butene increased from one reaction to the next. These observations are further substantiated by studying the variation of the extent of deuteration with reaction number in a series of reactions, carried out under TABLE 1 .-VARIATION OF THE DISTRIBUTION OF BUTENES WITH TEMPERATURE butene distribution (%) temperature/K 1 -B t-2-B C-2-B (t-2-B/c-2-B) Au /alumina- A 473 50.5 17.6 31.9 0.55 493 51.3 16.8 31.9 0.53 518 50.9 19.7 29.4 0.66 533 51.9 20.6 27.5 0.75 Au/y -alumina 443 58.4 12.8 27.8 0.46 473 58.1 14.3 27.6 0.52 488 58.7 14.5 26.8 0.54 503 59.2 14.7 26.1 0.56 533 59.7 15.4 24.9 0.62 (Pc.,H& = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.3 x 1 0 4 ~ m-2.Pressure fall at analysis = (25 0.1) x 103N m-2. TABLE 2.-vARIATION OF DEUTERATED PRODUCT DISTRIBUTIONS WITH PRESSURE FALL catalyst Au/y-alumina reaction no. pressure fall/ lO2N m-2 product 1 -B DO 25.6 D1 44.2 D2 25.7 D3 3.9 D4 0.6 D5 0.1 D.N. 1.10 x in HxD2-x catalyst reaction no. pressure fall/ 102N m-2 product DO D1 D2 D3 D4 D5 D.N. x in HxD2-x A3 16.0 t-2-B C-2-B 25.2 24.8 46.6 44.3 26.4 24.6 4.2 4.7 1.2 1.3 0.4 0.4 1.15 1.14 0.23 1.3-B 1 -B 82.9 42.0 14.7 39.6 1.8 15.6 0.6 2.4 0.0 0.4 0.0 0.0 0.20 0.79 Aulalumina-A A1 41.6 t-2-B C-2-B 39.6 42.9 38.5 38.7 18.5 15.1 3.0 2.8 0.5 0.5 0.0 0.0 0.86 0.79 0.27 AS 20.8 - - 1,3-B 1 -B 77.1 19.5 17.4 42.0 3.9 31.7 1 .l 5.9 0.5 0.9 0.0 0.0 0.30 1.27 0.22 1-B t-2-B 48.3 45.9 40.4 39.5 10.4 12.8 0.7 1.2 0.2 0.6 0.0 0.0 0.64 0.71 C-2-B 45.8 42.2 11.3 0.5 0.2 0.0 0.67 0.24 F6 18.0 1,3-B 1-B 97.3 52.6 2.1 37.7 0.6 9.1 0.0 0.5 0.0 0.0 0.0 0.0 0.03 0.65 F3 42.0 t-2-B C-2-B 51.0 53.7 36.5 37.6 11.2 8.2 1.2 0.6 0.0 0.0 0.0 0.0 0.69 0.63 0.30 1,3-B 95.1 4.0 0.7 0.2 0.0 0.0 0.06 Initial buta-1,3-diene pressure = 6.7 x 103N m-2 ; initial deuterium pressure = 1.33 x 104N m-2 ; temperature = 473 K.No species above Ds were observed.D. A . BUCHANAN AND G. WEBB 139 identical conditions to those used above, and analysed at a constant pressure fall of (2 +_ 0.1) x 103N m-2. The results show that, for each catalyst, although the butene distribution remained constant, the deuterium number of each n-butene increased with increasing reaction number as shown in fig.3. It is also of interest to compare the extent of deuteration, as indicated by the deuterium number, with the initial rate of deuteration. The results for but-1-ene are shown, using each catalyst, in fig. 4. Similar trends were observed for cis- and trans-but-2-ene. I L2- 1.0- 5 8 0.8- .- 8 - El ?-I Q) c) 3 0.6- 0.4 - - I I I t i 1 2 3 4 5 reaction number FIG. 3.-Variation in deuterium number of but-l-ene (filled symbols), trans-but-Zene (open symbols) and cis-but-2-ene (half-filled symbols) with reaction number over Au/y-alumina (circles) and Au/ alumina-A (squares). [(PC.,HJ~ = 6.7 x 103N m-2 ( P D ~ ) ~ = 1.33 x 104N m-2 ; temperature = 473 KI] 1 I I 1 I I 0 5.0 10.0 initial rate/N m-2 min-l FIG. 4.-Variation of deuterium number of but-1-ene with initial rate of deuteration over AuJp alumina (0) and Aulalumina-A (a) (conditions as in fig.3).140 BUTDAIENE HYDROGENATION OVER Au/A1,03 Although the “ reaction number ” effect tended to mask the effect of temperature upon the deuterobutene distribution and the extent of buta-l,3-diene exchange, it was possible to detect a temperature effect with both catalysts. From the results, shown in table 3, it is possible to see that, with both Au/y-alumina and Au/alumina-A, increasing temperature caused a decrease in the deuterium content of each butene, whereas the extent of hydrogen exchange tended to increase with increasing temper- ature. TABLE 3 .-THE VARIATION OF THE DEUTERATED PRODUCT DISTRIBUTIONS WITH TEMPERATURE catalyst Au /alumina-A temperature/K reaction no.product D.N. x in H,D2-x catalyst temperature/K reaction no. product 443 W 3 1-B t-2-B C-2-B 48.0 43.4 49.8 39.3 37.4 39.5 11.4 15.0 9.5 0.6 2.0 0.6 0.0 1.2 0.4 0.0 0.0 0.2 533 H/5 1.3-B 1-B t-2-B ~-24B 1,3-B 98.0 64.3 63.0 63.4 97.8 1.5 31.0 31.2 30.9 1.6 0.4 4.4 5.2 4.9 0.6 0.0 0.3 0.3 0.5 0.0 0.0 0.0 0.3 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.67 0.83 0.63 0.03 0.41 . 0.33 0.43 0.03 0.22 0.54 Au /y-alumina 473 D /2 533 D/3 1-B t-2-B C-2-B 1,3-B 28.7 27.7 28.4 80.4 45.1 44.1 45.7 17.1 22.5 23.8 22.2 2.2 3.2 4.0 3.5 0.3 0.6 0.5 0.2 0.0 0.0 0.0 0.0 0.0 1-B t-2-B C-2-B 33.8 33.1 31.3 43.3 43.2 44.2 19.1 19.7 19.6 3.2 3.3 3.8 0.5 0.7 0.8 0.1 0.0 0.3 1,3-B 79.7 17.6 2.5 0.1 0.0 0.0 D.N.1.02 1.05 1.02 0.22 0.94 0.95 0.99 0.23 0.18 0.28 Initial buta-1,3-diene pressure = 6.7 x 103N m-2 ; initial deuterium pressure = 1.3 x 104N m-’ ; pressure fall at analysis = 2.0 x 103N m-2. It was found that in all the buta-l,3-diene + deuterium reactions it was not possible to obtain a satisfactory mass balance for hydrogen and deuterium in the reactants and products. In all cases the products were considerably hydrogen rich. It was also noticeable that with Au/y-alumina the effect was less than with Au/alumina-A. Furthermore, with both catalysts, the mass imbalance became less from reaction to reaction, as the catalyst activity decreased. THE CATALYST-DEUTERIUM EXCHANGE REACTION The apparent lack of a good hydrogen-deuterium mass balance, together with the low deuterium content of the n-butenes suggest that the catalyst contains a source of hydrogen atoms capable of taking part in the ‘‘ hydrogenation ” reaction.In order to investigate this possibility, the interaction of deuterium with the catalyst was examined to determine the extent of hydrogen-deuterium exchange at 473 K following catalyst activation at 523 or 673 K. The catalyst (0.75 g) was first activated in hydrogenD. A. BUCHANAN AND G. WEBB 141 (2.66 x 104N m-2) for 12 h. The vessel containing the catalyst was then evacuated for 30 min at the activation temperature and the sample allowed to cool in vacuo to 473 K. The catalyst was then allowed to equilibrate with 2.66 x 104N m-2 of deuterium for varying times. Following each equilibration the deuterium was analysed for hydrogen content, the catalyst evacuated for 30 min and a further 2.66 x 104N m-2 of deuterium allowed to equilibrate with the catalyst for the same time as was used in the first reaction.For each catalyst this procedure was repeated twice and the results are shown in table 4. TABLE 4.-cATALY ST-DEUTERIUM EXCHANGE REACTION hydrogen number reduction temp./ exchange time/ K m n reaction 1 reaction 2 reaction 3 Au/ y-alumina 523 720 1.30 0.98 0.68 523 230 0.65 0.47 0.35 673 720 0.35 0.29 0.22 Au/alumina-A 523 720 0.82 0.65 0.60 523 30 0.24 0 . 1 7 0.10 673 720 0 . 1 4 0.10 0.08 Wt. of catalyst = 0.75 g; temperature of exchange = 473 K ; initial deuterium pressure = 1.33 x 104N m-2. CATALYST STRUCTURE Using an ultramicrotoming technique in which catalyst samples were embedded in Araldite, six to eight thin sections of each catalyst were examined by electron micro- scopy to determine the size distribution of the gold particles.Approximately 200 particles were counted from each section; the results for both Au/y-alumina and Au/alumina-A are shown 3 3 2 .- 4 4 a cd - c, c, x FIG. 5.-Distribution of metal particle size/A lines). particle sizes in Au/y-alumina (broken line) and Au/aulmina-A (full Some catalyst samples were also examined after use and it was found that, although these used catalysts had lost most of their activity, there was no detectable change in the particle size distributions relative to the unused catalysts.142 BUTADIENE HYDROGENATION OVER Au/A1203 DISCUSSION The results presented above show that gold, when supported in a finely divided state on alumina will effectively catalyse the reaction between buta-l,3-diene and hydrogen without, as reported previously,'* * the necessity of continuously providing hydrogen in atomic form to the gold surface.The activity of these gold catalysts is, not surprisingly, somewhat lower than that of the more conventional supported Group VIII metals such as palladium or platinum. Furthermore, the gold catalysts tend to progressively lose activity in use, the activity after around twenty reactions being approximately two orders of magnitude below the initial activity. The gold-catalysed reactions show some similarities to those observed with supported Group VIII metals and copper.13 Thus all three n-butenes are formed as initial products and the butene distribution and trans/cis ratio in but-2-ene are similar to those observed with ~1atinum.l~ However, unlike copper or the Group VIII metals, alumina-supported gold will not, under the conditions used in this study, catalyse the hydrogenation of n-butenes to butane, although it will catalyse the slow isomerisation of but-1-ene to cis- and trans-but-2-ene as shown in fig.2. This apparent lack of hydrogenation activity may be kinetic in its origin since studies of the hydrogenation of pent- 1 -ene over these catalysts, using hydrogen/hydrocarbon ratios of ca. 700, show that at 373 K both pentene hydrogenation and isomerisation occur. One of the most significant features to arise from the studies of the interaction of buta-l,3-diene with deuterium is the close relationship between the extent of deuter- ation of the products and the catalyst activity as determined from the initial rates of reaction, together with the considerable hydrogen-deuterium mass imbalance between reactants and products.Clearly, even after the extensive pretreatment in deuterium, the freshly prepared catalysts still contain a source of hydrogen which can partake in the hydrogenation reaction. Examination of the results in table 4 reveals that, with both the Au/y-alumina and the Au/alumina-A, extensive exchange occurs between catalyst hydrogen and gas- phase deuterium. Furthermore, the amounts of catalyst hydrogen are dependent upon the catalyst pretreatment temperature. Thus for Au/y-alumina the total numbers of exchangeable H-atoms per gram of catalyst following pretreatment at 523 and 673 K are 1.93 x 1021 and 5.59 x 1020 respectively.The corresponding values for Au/ alumina-A are 1.45 x 1021 and 2.08 x 1020 atoms per gram of catalyst. This effect of high temperature pretreatment suggests that the surface undergoes an irreversible change with loss of hydrogen, probably due to either dehydration or progressive dehydroxylation of the alumina surface. The results in table 3 show that the extent of deuterium incorporation in the pro- ducts was less with Au/alumina-A than with Au/y-alumina, although the initial rate of " hydrogenation " was higher with the former catalyst. In view of the prolonged catalyst pretreatment in deuterium during activation, and the lower total amount of exchangeable hydrogen on the Au/alumina-A catalyst, the results indicate that, on both catalysts, at least two types of surface hydrogen exist : type A, which undergoes equilibration with deuterium, and type B, which does not undergo ready exchange with deuterium, but which can readily partake in the hydrocarbon hydrogenation reaction.The relative amounts of type A and type B hydrogen must be different on the two catalysts such that the ratio (type A/type B) is greater on Au/y-alumina than on Au/alumina-A. Since with both catalysts the extent of deuterium incorporation in the products increases as the activity decreases, the pool of type B hydrogen must gradually become depleted and not replaced by " hydrogen " from the gas-phase.D.A. BUCHANAN AND G . WEBB 143 The magnitude of the type B pool can be estimated for each catalyst from the amounts of excess hydrogen in the products as determined from the hydrogen-deuterium mass balance between reactants and products. Using this method the respective values for Au/y-alumina and Au/alumina-A are 7.36 x 1020 atoms per gram and 2.62 x 1021 atoms per gram. With regard to the identity of the types A and B hydrogen, comparison of the results presented in table 4 with results obtained for tritium exchange of y-alumina and alumina-A at 523 K l6 suggests that the type A hydrogen may be identified with the surface hydroxyl groups on the alumina. The identity of the type B hydrogen is less clear. The occlusion of hydrogen in Group VIII metal powders,13 platinum black l7 and palladium black l8 and the reaction of occluded hydrogen with olefins is well established.However, comparison of the magnitude of the type B hydrogen pool with the gold concentration [-3 x 1019 atom (g catalyst)-l] rules out any possi- bility of the type B hydrogen being occluded in the metal. Clearly type B hydrogen is associated with the support, although its precise chemical identity is yet to be estab- lished. In view of the complete absence of activity of either the y-alumina or alumina-A for buta-l,3-diene hydrogenation, under the conditions used in the present study, the results presented above are explicable in terms of a mechanism in which buta-1,3- diene is adsorbed at the gold surface, and subsequently undergoes reaction with hydrogen atoms which migrate from the support to the metal. From the particle size distributions of the gold particles, it seems unlikely that the gold itself will contain special sites capable of activating molecular hydrogen, unless the activity resides in extremely small particles not sensed by electron microscopy.Thus the support plays a significant role in that it provides a source of atomic " hydrogen " for reaction. Such a conclusion is in agreement with the suggestions of previous workers regarding the catalytic activity of gold for hydrocarbon hydrogenation.'. The kinetics, high selectivity and the butene isomerisation activity of these catalysts is also consistent with the above conclusion, since it is envisaged that the steady state surface concentra- tion of hydrogen atoms on the metal will be very small under these conditions.The butene distributions show that, over both catalysts, the but- l-ene/but-2-ene ratio is slightly greater than unity and the cisltrans ratio in but-2-ene is around 1.5. Furthermore, the butene distribution is invariant with conversion until most of the buta- 1,3-diene has reacted. These observations together with the close similarities in the deuterium content of all three n-butenes suggest that the butenes are all formed directly from adsorbed buta-l,3-diene by a mechanism similar to that proposed by Phillipson et al. In conclusion, the results presented above show that gold when supported on y-alumina, or a mixed y-alumina-boehmite support, will catalyse the hydrogenation of buta-1,3-diene to butene, although the support plays a significant role by providing a source of atomic hydrogen to the gold surface.Further, at least two types of reactive hydrogen exist on the support, although the precise chemical identity of the various forms of support hydrogen is not established. Further investigations of the hydrogen associated with the support are at present being carried out. for the copper-catalysed reaction. The authors thank the S.R.C. for the award of a maintenance grant to one of us (D. A. B.), and for a grant to purchase the A.E.I. MS20 mass spectrometer. R. J. Mikovsky, M. Boudart and H. S. Taylor, J. Arner. Chem. Soc., 1954, 76, 3814. G. Parravano, J. Catalysis, 1970, 18, 320. * N. W. Cant and W. K. Hall, J. Phys. Chem., 1971,75,2914.144 BUTADIENE HYDROGENATION OVER Au/A1203 M.Boudart and L. D. Ptak, J. Catalysis, 1970, 16,90. D. D. Eley and D. R. Rossington, Chemisorption, ed. W. E. Garner (Butterworth, London, 1957), p. 137. R. P. Chambers and M. Boudart, J. Catalysis, 1966, 5, 517. B. J. Wood and H. Wise, J. Catalysis, 1966, 5, 135. W. M. H. Sachtler and N. H. De Boer, J. Phys. Chem., 1960,64,1579. * R. S. Yolles, B. J. Wood and H. Wise, J. Catalysis, 1971, 21, 66. lo G. Webb and J. I. Macnab, J. Catalysis, 1972, 26,226. l1 G. Webb and J. A. Altham, J. Catalysis, 1970, 18, 133. l2 J. U. Reid, S. J. Thomson and G. Webb, J. Catalysis, 1973, 29, 421 ; 1973, 30, 372. l3 P. B. Wells, Surface and Defect Properties of Solids (The Chemical Society, London, 1971), l4 G. C. Bond, G. Webb, P.B. Wells and J. M. Winterbottom, J. Chem. SOC., 1965, 3218. l5 G. C. Bond, P. A. Sermon, D. A. Buchanan, G. Webb and P. B. Wells, Chem. Comm., 1973 l 6 P. A. Sermon, G. C. Bond and G. Webb, Chem. Comm., 1974,417. vol. 1, p. 236. 444. Z. Paal and S. J. Thomson, J. CataZysis, 1973, 30,96. L. V. Babenkova, N. M. Popova, D. V. Sokol’skii and V. K. Solynserkova, Doklady Akad. Nauk S.S.S.R., 1973, 210, 888. l9 J. J. Phillipson, P. B. Wells and G. R. Wilson, J. Chem. SOC. A, 1969, 1351. Catalysis by Group lk Metals Part 1 ,-Reaction of Buta-1,3-diene with Hydrogen and with Deuterium catalysed by Alumina-supported Gold BY DOUGLAS A. BUCHANAN AND GEOFFREY WEBB* Chemistry Department, The University, Glasgow G12 SQQ, Scotland Received 23rd May, 1974 The reaction of buta-l,3-diene with hydrogen and deuterium has been studied using gold supported on y-alumina and a mixed boehmite+y-alumina in the temperature range 443-533 K.The reaction is completely selective for butene formation and, although both catalysts are active for the hydro- isomerisation of but-l-ene at 473 K, they do not catalyse the hydrogenation of n-butenes to n-butane. The kinetics and activation energy have been determined together with the variation in product distributions with hydrogen uptake, initial reactant pressures and temperature. The distribution of deuterium in each of the products has been determined and in the reaction with deuterium the prod- ducts contain an excess of hydrogen over that expected from hydrogen-deuterium mass balance. Buta-1,3-diene is adsorbed on the surface of the gold and subsequently reacts with hydrogen which migrates from the support to the metal.Type A hydrogen equilibrates with gas phase deuterium, whereas type B hydrogen does not react with gas phase deuterium but interacts with adsorbed hydrocarbon. Values for the sizes of the type A and B hydrogen pools are quoted. Two types of hydrogen exist on the alumina support. The low activity of gold for many catalysed reactions has been attributed to its lack of a partially filled d-band at the usual temperatures, and its consequent inability to chemisorb simple molecules. Gold will not adsorb or activate molecular hydrogen at temperatures below 473 K,I although ethylene can be chemisorbed in a weakly held form on gold surfaces.2 This chemisorption of olefins is consistent with the observa- tion that gold, when dispersed on alumina or magnesium oxide, possesses apprec- iable activity for the equilibration of benzene + cyclohexane mixtures.Gold powder will catalyse the skeletal isomerisation of ne~pentane,~ although, compared with Group VIII metal catalysts, the temperature required for this reaction is rather high being around 800 K, where thermal promotion of electrons from the 5d to the 6s level may occur.'* 4-6 Gold, when electroplated on to the surface of a palladium-silver alloy, will catalyse the hydrogenation of but-1-ene and cy~lohexene,~* but only when hydrogen atoms are provided to the gold surface by diffusion through the palladium-silver alloy. This pre-requisite, for hydrogenation activity, of the provision of atomic hydrogen at the gold surface has also been demonstrated in a study of the hydrogenation of ethylene on evaporated gold films, using the gold-catalysed dehydrogenation of formic acid as the source of hydrogen atom^.^ that in the reactions of unsaturated hydrocarbons using supported Group VIII metal catalysts, the support plays a significant role in deter- mining the catalytic behaviour. In particular the migration of adsorbed species between metal and support, and the role of hydrogen associated with the support have been discussed.In view of the apparent inability of gold to activate molecular hydrogen, the present studies were undertaken in an attempt to separate the catalytic effects of the metal and the support in hydrogenation reactions, particularly with reference to the availability of hydrogen and the migration of adsorbed species be- tween metal and support.It has been shown 134D. A. BUCHANAN AND G. WEBB 135 EXPERIMENTAL CATALYSTS Two catalyst supports were used; y-alumina (Degussa Ltd.), and y-alumina, from the same source, which had been pretreated by refluxing with a 1 mol dm-3 aqueous sodium acetate solution for 24 h, followed by washing with distilled water, until all traces of acetate ion had been removed, and finally drying in an air oven at 393-423 K. This latter support is designated alumina-A. X-ray analysis of alumina-A showed that it consisted of a mixture of y-alumina and the alumina hydrate, boehmite, although the relative proportions of each component could not be satisfactorily determined.The catalysts, containing 1 % w/w gold, were prepared by adding an aqueous solution of chloroauric acid (HAuCI,), containing the required weight of gold, to an aqueous suspension of the support. The excess water was removed by evaporation and the catalyst fmally dried in an air oven at a temperature between 393 and 423 K. During the preparation the supported salt was observed to undergo thermal decomposition resulting in the formation of a mauve coloured product containing metallic gold. In addition to the above catalysts, 1 % wlw gold supported on a-alumina (I.C.I. Ltd.) and Aerosil silica (Degussa Ltd.) were prepared. However, under the reaction conditions used in this study, neither of these catalysts was observed to possess any catalytic activity.MATERIALS Buta-l,3-diene (Matheson Co.) contained no impurities detectable by gas chromato- graphy and was merely degassed before use. Cylinder hydrogen was purifi5d by diffusion through a palladium-silver alloy thimble. Deuterium (Norsk Hydro) was of purity 99.8 atom % D and contained less than 10 p.p.m. oxygen impurity. It was used as supplied. APPARATUS A N D PROCEDURE Reactions were carried out with the catalyst, usually 0.5 g, resting on the bottom of a cylindrical Pyrex reaction vessel (capacity ca. 100 cm3). The reaction vessel was connected to a conventional high vacuum system maintained at Nm-2 or better by means of an oil diffusion pump backed by an oil rotary-pump. The catalyst was activated by treatment with three successive volumes of hydrogen, at a pressure of 2.66 x lo4 N m-2, for a total of 12 h at 523 K.The catalyst was then evacuated for 30 min and cooled to the reaction temperature. The reaction mixture was admitted to the reaction vessel and the reaction followed by the pressure fall observed on a mercury manometer. At the required pressure fall the reaction products were extracted for analysis. In those reactions where deuterium wasused, following the activation in hydrogen, the catalyst was treated with three successive volumes of deuterium (pressure 3.33 x lo4 N m-2) at 673 K, each sample of deuterium being left in contact with the catalyst for 12 h. Mass spectro- metric analysis of the deuterium after the pretreatment showed that this treatment was effect- ive in exchanging all of the exchangeabk hydrogen in the catalyst.ANALYSIS OF REACTION PRODUCTS The analysis of reaction products was effected by gas-liquid chromatography using an 8 m column packed with a 40 % w/w dispersion of hexa-2,5-dione supported on 30-60 mesh Silocel firebrick. In the experiments using deuterium each reaction product was, on elution from the gas chromatograph, condensed out of the helium carrier gas stream in a trap cooled in liquid nitrogen. The products were then analysed for deuterium content using an A.E.I. MS20 mass spectrometer. For the hydrocarbon analyses an ionizing beam voltage of 15 V was used, while for the analysis of residual deuterium a beam voltage of 70 V was used.136 B u TAD I E N E H Y D R o GEN A T I ON OVE R Au/AI,O, RESULTS THE REACTION OF BUTA-I,3-DIENE WITH HYDROGEN The variation in the distribution of reaction products with pressure fall was studied using both Au/y-alumina and Au/alumina-A at 473 K and a pressure of 6.7 x lo3 N rn-, of buta-1,3-diene and 1.33 x lo4 N rn-, of hydrogen.Under these conditions the reaction was completely selective for the formation of n-butene; no butane was ever observed as a reaction product and the reaction ceased at a pressure fall corresponding to the uptake of 1 mole of hydrogen per mole of buta-1,3-diene. The variation in the butene distribution with hydrogen uptake per mole of buta- 1,3- diene (AH) is shown in fig. l(a) and (b) for Au/y-alumina and Aulalumina-A res- pectively. X I I I I 1 I I 0 0.2 0.4 0.6 0.8 1.0 1 I I I I 0 0.2 0.4 0.6 0.8 1.0 hydrogen consumed (AH)/mol (mol diene)-l FIG.1.-Variation of butene distribution with hydrogen uptake over Au/y-alumina (a), and Au/ alumina-A (6) at 473 K. (&4&)0 = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.33 x 104N m-2. (0, but-1-ene ; 0, cis-but-Zene ; 0, trans-but-Zene). (a) (6) time/h FIG. 2.-Variation in butene distribution with time of contact of but-1-ene+ hydrogen mixture with Au/alumina-A at 473 K. ( P ~ 4 ~ s ) o = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.33 x 104N m-2. Broken lines indicate the thermodynamic equilibrium values. (0, but-1-ene ; 0, cis-but-Zene ; (>, tvans-but-2- ene).D . A. BUCHANAN AND G . WEBB 137 These results show that the butene distribution was independent of hydrogen uptake up to AH = 0.6 for Au/y-alumina or AH = 0.90 with Au/alumina-A. The initial but- 1 -ene/but-2-ene ratio was greater with Au/y-alumina. In the latter stages of the reaction the butenes effectively competed with the buta-1,3-diene for the catalyst surface and underwent isomerisation, although no n-butane was detected.This effect was more marked with Au/y-alumina than with Aulalumina-A. If the reaction products were left in contact with the catalyst for a further 18 h, the butenes underwent extensive isomerisation, eventually attaining their thermodynamic equi- librium proportions, as indicated by the points shown at AH = 1.0 in fig. l(a) and (6) although again no n-butane was detectable. Under comparable conditions, Au/ alumina-A will effect the hydroisomerisation of but-1-ene (see fig. 2), although but-l- ene hydrogenation does not occur.It was noticeable throughout this work that the catalytic activity progressively decreased from reaction to reaction until, after about twenty reactions, a low limiting value was attained. By adopting a standard run technique, in which alternate reac- tions were carried out using a standard butene/hydrogen mixture, it was possible to correct for changes in catalytic activity and hence, using catalysts which had already been subjected to a number (usually 8-10) of reactions, to determine initial rate orders with respect to both hydrogen and buta-1,3-diene over the pressure range 6.7 x lo3 to 3.33 x 104N m-2 using each catalyst at 473 K. The results for the two catalysts were similar, the reaction following the rate expression ; This rate expression is also consistent with the observed pressure fall against time curves, which were first order with respect to the total pressure.The product distributions were independent of initial reactant pressures ; again the reaction was completely selective, no n-butane being observed under the condition used. The effect of temperature in the range 473 to 533 K (Au/y-alumina) and 443 to 533 K (Au/alumina-A) was studied using 6.7 x 103N m-2 buta-1,3-diene and 1.33 x 104Nm-2 hydrogen. Using a standard run technique to correct for variations in catalyst activity, an apparent activation energy of 36.5 4.0 kJ mol-1 was obtained for each catalyst. The reaction products were extracted at a pressure fall of (2 & 0.1) x 103N m-2 and analysed. With both catalysts the but-1 -ene/but-2-ene ratio appeared to be independent of temperature, although the trans/& ratio in the but-2-ene yield increased with increasing temperature (see table 1) THE REACTION BETWEEN BUTA-I ,3-DIENE AND DEUTERIUM The variation in the distribution of deuterium in the reaction products with pres- sure fall was determined using both Au/y-alumina and Au/alumina-A.The varia- tions of deuterated product distributions with pressure fall are shown in table 2, in which D.N. represents the deuterium number, defined as the average number of deuterium atoms per hydrocarbon molecule. With both catalysts the three n-butenes exhibit very similar deuterium distributions, although the extent of deuteration is surprisingly low. The extent of deuteration of both the butenes and the buta-l,3-diene is substantially less using Au/alumina-A than using Au/y-alumina, although the amounts of hydrogen exchange are comparable for the two catalysts.At first sight it would appear that, as expected in view of the increasing amount of hydro- gen exchange, the degree of deuteration of the butenes decreased with increasing conversion of buta-I ,3-diene. However, consideration of the sequence in which the Several interesting features emerge from these results.138 BUTADIENE HYDROGENATION OVER Au/A1,0, reactions were performed shows that any effects of pressure fall are completely masked by the variation of the extent of deuteration with reaction number. The variation is such that the deuterium number of each butene increased from one reaction to the next.These observations are further substantiated by studying the variation of the extent of deuteration with reaction number in a series of reactions, carried out under TABLE 1 .-VARIATION OF THE DISTRIBUTION OF BUTENES WITH TEMPERATURE butene distribution (%) temperature/K 1 -B t-2-B C-2-B (t-2-B/c-2-B) Au /alumina- A 473 50.5 17.6 31.9 0.55 493 51.3 16.8 31.9 0.53 518 50.9 19.7 29.4 0.66 533 51.9 20.6 27.5 0.75 Au/y -alumina 443 58.4 12.8 27.8 0.46 473 58.1 14.3 27.6 0.52 488 58.7 14.5 26.8 0.54 503 59.2 14.7 26.1 0.56 533 59.7 15.4 24.9 0.62 (Pc.,H& = 6.7 x 103N m-2 ; ( P H ~ ) ~ = 1.3 x 1 0 4 ~ m-2. Pressure fall at analysis = (25 0.1) x 103N m-2. TABLE 2.-vARIATION OF DEUTERATED PRODUCT DISTRIBUTIONS WITH PRESSURE FALL catalyst Au/y-alumina reaction no.pressure fall/ lO2N m-2 product 1 -B DO 25.6 D1 44.2 D2 25.7 D3 3.9 D4 0.6 D5 0.1 D.N. 1.10 x in HxD2-x catalyst reaction no. pressure fall/ 102N m-2 product DO D1 D2 D3 D4 D5 D.N. x in HxD2-x A3 16.0 t-2-B C-2-B 25.2 24.8 46.6 44.3 26.4 24.6 4.2 4.7 1.2 1.3 0.4 0.4 1.15 1.14 0.23 1.3-B 1 -B 82.9 42.0 14.7 39.6 1.8 15.6 0.6 2.4 0.0 0.4 0.0 0.0 0.20 0.79 Aulalumina-A A1 41.6 t-2-B C-2-B 39.6 42.9 38.5 38.7 18.5 15.1 3.0 2.8 0.5 0.5 0.0 0.0 0.86 0.79 0.27 AS 20.8 - - 1,3-B 1 -B 77.1 19.5 17.4 42.0 3.9 31.7 1 .l 5.9 0.5 0.9 0.0 0.0 0.30 1.27 0.22 1-B t-2-B 48.3 45.9 40.4 39.5 10.4 12.8 0.7 1.2 0.2 0.6 0.0 0.0 0.64 0.71 C-2-B 45.8 42.2 11.3 0.5 0.2 0.0 0.67 0.24 F6 18.0 1,3-B 1-B 97.3 52.6 2.1 37.7 0.6 9.1 0.0 0.5 0.0 0.0 0.0 0.0 0.03 0.65 F3 42.0 t-2-B C-2-B 51.0 53.7 36.5 37.6 11.2 8.2 1.2 0.6 0.0 0.0 0.0 0.0 0.69 0.63 0.30 1,3-B 95.1 4.0 0.7 0.2 0.0 0.0 0.06 Initial buta-1,3-diene pressure = 6.7 x 103N m-2 ; initial deuterium pressure = 1.33 x 104N m-2 ; temperature = 473 K.No species above Ds were observed.D. A . BUCHANAN AND G. WEBB 139 identical conditions to those used above, and analysed at a constant pressure fall of (2 +_ 0.1) x 103N m-2. The results show that, for each catalyst, although the butene distribution remained constant, the deuterium number of each n-butene increased with increasing reaction number as shown in fig. 3. It is also of interest to compare the extent of deuteration, as indicated by the deuterium number, with the initial rate of deuteration. The results for but-1-ene are shown, using each catalyst, in fig.4. Similar trends were observed for cis- and trans-but-2-ene. I L2- 1.0- 5 8 0.8- .- 8 - El ?-I Q) c) 3 0.6- 0.4 - - I I I t i 1 2 3 4 5 reaction number FIG. 3.-Variation in deuterium number of but-l-ene (filled symbols), trans-but-Zene (open symbols) and cis-but-2-ene (half-filled symbols) with reaction number over Au/y-alumina (circles) and Au/ alumina-A (squares). [(PC.,HJ~ = 6.7 x 103N m-2 ( P D ~ ) ~ = 1.33 x 104N m-2 ; temperature = 473 KI] 1 I I 1 I I 0 5.0 10.0 initial rate/N m-2 min-l FIG. 4.-Variation of deuterium number of but-1-ene with initial rate of deuteration over AuJp alumina (0) and Aulalumina-A (a) (conditions as in fig. 3).140 BUTDAIENE HYDROGENATION OVER Au/A1,03 Although the “ reaction number ” effect tended to mask the effect of temperature upon the deuterobutene distribution and the extent of buta-l,3-diene exchange, it was possible to detect a temperature effect with both catalysts.From the results, shown in table 3, it is possible to see that, with both Au/y-alumina and Au/alumina-A, increasing temperature caused a decrease in the deuterium content of each butene, whereas the extent of hydrogen exchange tended to increase with increasing temper- ature. TABLE 3 .-THE VARIATION OF THE DEUTERATED PRODUCT DISTRIBUTIONS WITH TEMPERATURE catalyst Au /alumina-A temperature/K reaction no. product D.N. x in H,D2-x catalyst temperature/K reaction no. product 443 W 3 1-B t-2-B C-2-B 48.0 43.4 49.8 39.3 37.4 39.5 11.4 15.0 9.5 0.6 2.0 0.6 0.0 1.2 0.4 0.0 0.0 0.2 533 H/5 1.3-B 1-B t-2-B ~-24B 1,3-B 98.0 64.3 63.0 63.4 97.8 1.5 31.0 31.2 30.9 1.6 0.4 4.4 5.2 4.9 0.6 0.0 0.3 0.3 0.5 0.0 0.0 0.0 0.3 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.67 0.83 0.63 0.03 0.41 .0.33 0.43 0.03 0.22 0.54 Au /y-alumina 473 D /2 533 D/3 1-B t-2-B C-2-B 1,3-B 28.7 27.7 28.4 80.4 45.1 44.1 45.7 17.1 22.5 23.8 22.2 2.2 3.2 4.0 3.5 0.3 0.6 0.5 0.2 0.0 0.0 0.0 0.0 0.0 1-B t-2-B C-2-B 33.8 33.1 31.3 43.3 43.2 44.2 19.1 19.7 19.6 3.2 3.3 3.8 0.5 0.7 0.8 0.1 0.0 0.3 1,3-B 79.7 17.6 2.5 0.1 0.0 0.0 D.N. 1.02 1.05 1.02 0.22 0.94 0.95 0.99 0.23 0.18 0.28 Initial buta-1,3-diene pressure = 6.7 x 103N m-2 ; initial deuterium pressure = 1.3 x 104N m-’ ; pressure fall at analysis = 2.0 x 103N m-2. It was found that in all the buta-l,3-diene + deuterium reactions it was not possible to obtain a satisfactory mass balance for hydrogen and deuterium in the reactants and products.In all cases the products were considerably hydrogen rich. It was also noticeable that with Au/y-alumina the effect was less than with Au/alumina-A. Furthermore, with both catalysts, the mass imbalance became less from reaction to reaction, as the catalyst activity decreased. THE CATALYST-DEUTERIUM EXCHANGE REACTION The apparent lack of a good hydrogen-deuterium mass balance, together with the low deuterium content of the n-butenes suggest that the catalyst contains a source of hydrogen atoms capable of taking part in the ‘‘ hydrogenation ” reaction. In order to investigate this possibility, the interaction of deuterium with the catalyst was examined to determine the extent of hydrogen-deuterium exchange at 473 K following catalyst activation at 523 or 673 K.The catalyst (0.75 g) was first activated in hydrogenD. A. BUCHANAN AND G. WEBB 141 (2.66 x 104N m-2) for 12 h. The vessel containing the catalyst was then evacuated for 30 min at the activation temperature and the sample allowed to cool in vacuo to 473 K. The catalyst was then allowed to equilibrate with 2.66 x 104N m-2 of deuterium for varying times. Following each equilibration the deuterium was analysed for hydrogen content, the catalyst evacuated for 30 min and a further 2.66 x 104N m-2 of deuterium allowed to equilibrate with the catalyst for the same time as was used in the first reaction. For each catalyst this procedure was repeated twice and the results are shown in table 4.TABLE 4.-cATALY ST-DEUTERIUM EXCHANGE REACTION hydrogen number reduction temp./ exchange time/ K m n reaction 1 reaction 2 reaction 3 Au/ y-alumina 523 720 1.30 0.98 0.68 523 230 0.65 0.47 0.35 673 720 0.35 0.29 0.22 Au/alumina-A 523 720 0.82 0.65 0.60 523 30 0.24 0 . 1 7 0.10 673 720 0 . 1 4 0.10 0.08 Wt. of catalyst = 0.75 g; temperature of exchange = 473 K ; initial deuterium pressure = 1.33 x 104N m-2. CATALYST STRUCTURE Using an ultramicrotoming technique in which catalyst samples were embedded in Araldite, six to eight thin sections of each catalyst were examined by electron micro- scopy to determine the size distribution of the gold particles. Approximately 200 particles were counted from each section; the results for both Au/y-alumina and Au/alumina-A are shown 3 3 2 .- 4 4 a cd - c, c, x FIG.5.-Distribution of metal particle size/A lines). particle sizes in Au/y-alumina (broken line) and Au/aulmina-A (full Some catalyst samples were also examined after use and it was found that, although these used catalysts had lost most of their activity, there was no detectable change in the particle size distributions relative to the unused catalysts.142 BUTADIENE HYDROGENATION OVER Au/A1203 DISCUSSION The results presented above show that gold, when supported in a finely divided state on alumina will effectively catalyse the reaction between buta-l,3-diene and hydrogen without, as reported previously,'* * the necessity of continuously providing hydrogen in atomic form to the gold surface.The activity of these gold catalysts is, not surprisingly, somewhat lower than that of the more conventional supported Group VIII metals such as palladium or platinum. Furthermore, the gold catalysts tend to progressively lose activity in use, the activity after around twenty reactions being approximately two orders of magnitude below the initial activity. The gold-catalysed reactions show some similarities to those observed with supported Group VIII metals and copper.13 Thus all three n-butenes are formed as initial products and the butene distribution and trans/cis ratio in but-2-ene are similar to those observed with ~1atinum.l~ However, unlike copper or the Group VIII metals, alumina-supported gold will not, under the conditions used in this study, catalyse the hydrogenation of n-butenes to butane, although it will catalyse the slow isomerisation of but-1-ene to cis- and trans-but-2-ene as shown in fig.2. This apparent lack of hydrogenation activity may be kinetic in its origin since studies of the hydrogenation of pent- 1 -ene over these catalysts, using hydrogen/hydrocarbon ratios of ca. 700, show that at 373 K both pentene hydrogenation and isomerisation occur. One of the most significant features to arise from the studies of the interaction of buta-l,3-diene with deuterium is the close relationship between the extent of deuter- ation of the products and the catalyst activity as determined from the initial rates of reaction, together with the considerable hydrogen-deuterium mass imbalance between reactants and products.Clearly, even after the extensive pretreatment in deuterium, the freshly prepared catalysts still contain a source of hydrogen which can partake in the hydrogenation reaction. Examination of the results in table 4 reveals that, with both the Au/y-alumina and the Au/alumina-A, extensive exchange occurs between catalyst hydrogen and gas- phase deuterium. Furthermore, the amounts of catalyst hydrogen are dependent upon the catalyst pretreatment temperature. Thus for Au/y-alumina the total numbers of exchangeable H-atoms per gram of catalyst following pretreatment at 523 and 673 K are 1.93 x 1021 and 5.59 x 1020 respectively. The corresponding values for Au/ alumina-A are 1.45 x 1021 and 2.08 x 1020 atoms per gram of catalyst. This effect of high temperature pretreatment suggests that the surface undergoes an irreversible change with loss of hydrogen, probably due to either dehydration or progressive dehydroxylation of the alumina surface.The results in table 3 show that the extent of deuterium incorporation in the pro- ducts was less with Au/alumina-A than with Au/y-alumina, although the initial rate of " hydrogenation " was higher with the former catalyst. In view of the prolonged catalyst pretreatment in deuterium during activation, and the lower total amount of exchangeable hydrogen on the Au/alumina-A catalyst, the results indicate that, on both catalysts, at least two types of surface hydrogen exist : type A, which undergoes equilibration with deuterium, and type B, which does not undergo ready exchange with deuterium, but which can readily partake in the hydrocarbon hydrogenation reaction.The relative amounts of type A and type B hydrogen must be different on the two catalysts such that the ratio (type A/type B) is greater on Au/y-alumina than on Au/alumina-A. Since with both catalysts the extent of deuterium incorporation in the products increases as the activity decreases, the pool of type B hydrogen must gradually become depleted and not replaced by " hydrogen " from the gas-phase.D. A. BUCHANAN AND G . WEBB 143 The magnitude of the type B pool can be estimated for each catalyst from the amounts of excess hydrogen in the products as determined from the hydrogen-deuterium mass balance between reactants and products. Using this method the respective values for Au/y-alumina and Au/alumina-A are 7.36 x 1020 atoms per gram and 2.62 x 1021 atoms per gram.With regard to the identity of the types A and B hydrogen, comparison of the results presented in table 4 with results obtained for tritium exchange of y-alumina and alumina-A at 523 K l6 suggests that the type A hydrogen may be identified with the surface hydroxyl groups on the alumina. The identity of the type B hydrogen is less clear. The occlusion of hydrogen in Group VIII metal powders,13 platinum black l7 and palladium black l8 and the reaction of occluded hydrogen with olefins is well established. However, comparison of the magnitude of the type B hydrogen pool with the gold concentration [-3 x 1019 atom (g catalyst)-l] rules out any possi- bility of the type B hydrogen being occluded in the metal.Clearly type B hydrogen is associated with the support, although its precise chemical identity is yet to be estab- lished. In view of the complete absence of activity of either the y-alumina or alumina-A for buta-l,3-diene hydrogenation, under the conditions used in the present study, the results presented above are explicable in terms of a mechanism in which buta-1,3- diene is adsorbed at the gold surface, and subsequently undergoes reaction with hydrogen atoms which migrate from the support to the metal. From the particle size distributions of the gold particles, it seems unlikely that the gold itself will contain special sites capable of activating molecular hydrogen, unless the activity resides in extremely small particles not sensed by electron microscopy.Thus the support plays a significant role in that it provides a source of atomic " hydrogen " for reaction. Such a conclusion is in agreement with the suggestions of previous workers regarding the catalytic activity of gold for hydrocarbon hydrogenation.'. The kinetics, high selectivity and the butene isomerisation activity of these catalysts is also consistent with the above conclusion, since it is envisaged that the steady state surface concentra- tion of hydrogen atoms on the metal will be very small under these conditions. The butene distributions show that, over both catalysts, the but- l-ene/but-2-ene ratio is slightly greater than unity and the cisltrans ratio in but-2-ene is around 1.5. Furthermore, the butene distribution is invariant with conversion until most of the buta- 1,3-diene has reacted. These observations together with the close similarities in the deuterium content of all three n-butenes suggest that the butenes are all formed directly from adsorbed buta-l,3-diene by a mechanism similar to that proposed by Phillipson et al. In conclusion, the results presented above show that gold when supported on y-alumina, or a mixed y-alumina-boehmite support, will catalyse the hydrogenation of buta-1,3-diene to butene, although the support plays a significant role by providing a source of atomic hydrogen to the gold surface. Further, at least two types of reactive hydrogen exist on the support, although the precise chemical identity of the various forms of support hydrogen is not established. Further investigations of the hydrogen associated with the support are at present being carried out. for the copper-catalysed reaction. The authors thank the S.R.C. for the award of a maintenance grant to one of us (D. A. B.), and for a grant to purchase the A.E.I. MS20 mass spectrometer. R. J. Mikovsky, M. Boudart and H. S. Taylor, J. Arner. Chem. Soc., 1954, 76, 3814. G. Parravano, J. Catalysis, 1970, 18, 320. * N. W. Cant and W. K. Hall, J. Phys. Chem., 1971,75,2914.144 BUTADIENE HYDROGENATION OVER Au/A1203 M. Boudart and L. D. Ptak, J. Catalysis, 1970, 16,90. D. D. Eley and D. R. Rossington, Chemisorption, ed. W. E. Garner (Butterworth, London, 1957), p. 137. R. P. Chambers and M. Boudart, J. Catalysis, 1966, 5, 517. B. J. Wood and H. Wise, J. Catalysis, 1966, 5, 135. W. M. H. Sachtler and N. H. De Boer, J. Phys. Chem., 1960,64,1579. * R. S. Yolles, B. J. Wood and H. Wise, J. Catalysis, 1971, 21, 66. lo G. Webb and J. I. Macnab, J. Catalysis, 1972, 26,226. l1 G. Webb and J. A. Altham, J. Catalysis, 1970, 18, 133. l2 J. U. Reid, S. J. Thomson and G. Webb, J. Catalysis, 1973, 29, 421 ; 1973, 30, 372. l3 P. B. Wells, Surface and Defect Properties of Solids (The Chemical Society, London, 1971), l4 G. C. Bond, G. Webb, P. B. Wells and J. M. Winterbottom, J. Chem. SOC., 1965, 3218. l5 G. C. Bond, P. A. Sermon, D. A. Buchanan, G. Webb and P. B. Wells, Chem. Comm., 1973 l 6 P. A. Sermon, G. C. Bond and G. Webb, Chem. Comm., 1974,417. vol. 1, p. 236. 444. Z. Paal and S. J. Thomson, J. CataZysis, 1973, 30,96. L. V. Babenkova, N. M. Popova, D. V. Sokol’skii and V. K. Solynserkova, Doklady Akad. Nauk S.S.S.R., 1973, 210, 888. l9 J. J. Phillipson, P. B. Wells and G. R. Wilson, J. Chem. SOC. A, 1969, 1351.