Reactions of n-Butenes on Palladium Films Evidence for n-Allylic Species BY MARC J. LEDOUX* AND FRANCOIS G. GAULT Laboratoire de Catalyse, UniversitC Louis Pasteur, Strasbourg, France AND ALAIN BOUCHY AND GEORGES ROUSSY Laboratoire de Chimie Thiorique, Universite Nancy I, Villers-les-Nancy, France Received 27th February, 1978 The contact reactions of all three n-butenes have been investigated on palladium films in the presence of perdeuteropropene or deuterium, using jointly mass spectrometry and microwave spectroscopy. Conclusive. evidence has been found for an allylic mechanism in exchange and double-bond migration. Part of the exchanged molecules rcsult from dissociativc vinylic-type adsorption and isomerization also takes place according to an associative Horiuti-Polanyi mechanism or by direct processes without deuterium incorporation.A proposal was made which correlates the various reaction mechanisms with various types of surface' metal atoms, each having a different coordination number. Since the introduction of n-allylic species into heterogeneous catalysis 9 there has been controversy about their existence and significance for hydrocarbon reactions on metal surfaces. The exchange distribution patterns of a number of cyclic hydro- carbons can be accounted for by this intermediate, but only by assuming that n-allylic species can convert into n-olefinic species in two different ways, reacting either with adsorbed hydrogen atoms or with molecular hydrogen coming from the gas p h a ~ e . l ' ~ While it was soon recognized that n-allylic complexes are very coinmon in organometallic cheniistry and convert easily into n-olefinic complexes by insertion into a metal-hydrogen bond, the reaction of molecular hydrogen with n-allylic species to form n-bonded olefins has never been observed in homogeneous catalysis.On the other hand, an alternative mechanism, the roll-over mechanism, based on adsorbed species doubly attached to the metal by three-centre bonds, has been this mechanism could account for all the observed exchange distribution patterns, as well as the n-allylic mechanism. Later on, a key compound that allowed one to distinghish between the roll-over mechanism and a n-allylic mechanism with transoid stddition of molecular hydrogen was investigated for exchange. The results unambiguously favoured the former mechanism? While the question of the involvement of n-allylic species in the exchange of cycloalkaiies remained unsettled, some illdirect proofs were provided that z-allylic species might phy an important role in exchange and isoinerization of o i e i h 011 pa1ladiuix8* in the presence of 2652 Moreover it was shown that on iron,loiM .J . LEDOUX, F . G . GAULT, A . BOUCHY AND G . ROUSSY 2653 perdeuteropropene, most of the reactions of butenes could not be accounted for by the classical Horiuti-Polanyi mechanism ; no fewer than five reaction mechanisms could be identified, among which are a vinylic-type dissociative mechanism for exchange (the Farkas mechanism), direct cis-trans isomerization, and double-bond migration without deuterium incorporation. Because olefin reactions on metals are so complex, it is illusory to characterize the various mechanisms in detail by simply looking at the deuterium distributions as obtained by mass spectrometer measurements.Location of the deuterium in the reaction products is required and that is better achieved by microwave spectroscopy. This technique was first introduced into catalytic study by Hirota, studying the exchange of propene on various metal catalysts.14* l 5 However, interpretation of the results might be difficult in that case, since no distinction can be made between the exchanged and the isomerized molecules. For this reason, we decided to reinvestigate the reaction mechanisms of olefins on metals, by combining mass spectrometry and microwave spectroscopy and choosing butenes as model The present paper is devoted to reactions on palladium to decide whether or not n-allylic species may be formed on this metal.On the other hand, an attempt has recently been made to correlate the various reactions of olefins with various types of sites on a metal surface.l* To each reaction mechanism was associated a superficial metal atom with specific configuration and coordination number. If such a theory were correct, one should observe, when modifying the roughness of the surface, an effect on the relative contribution of the various reaction paths. A second aim of this work, then, was to determine the effect on the reaction mechanisms of variation in the conditions of film preparation. EXPERIMENTAL CATALYSTS Palladium films of different weights were prepared by evaporating, under a vacuum of 10-6-10-7 Torr, onto a glass vessel maintained at 0°C (films of type I) or 470°C (films of type 11), a specpure filament of palladium (diam.0.2 m) tightly wound on a 15 cm long wire of tungsten (diam 0.3 mm). Before film preparation, the reaction vessel was outgassed at 500°C for at least three hours and the filament heated for 30 min at the limit of evaporation, to eliminate dissolved and adsorbed gases. MATERIALS The three n-butenes, Fluka puriss, were used without further purification. We checked their purity (>99.995 %) by gas chromatography before each experiment. Deuterium, from Air Liquide Co., was purified by diffusion through a palladium thimble. Its isotopic purity was 99.4 %.Perdeuteropropene, from Merck, Sharpe and Dohme, was purified by semipreparative gas chromatography. Mass spectrometer measurements revealed an isotopic purity of 99.1 %. APPARATUS AND PROCEDURE We used a conventional static Pyrex apparatus like that previously described byKembal1,l but all the valves, taps and joints were of Teflon or Viton without any grease and the dead volumes were as small as possible to avoid loss of products. The mixture of but-Zene and perdeuteropropene (or deuterium), prepared in a separate part of the apparatus, was introduced on a fresh film at 0°C in the case of C3De and at -84°C in the case of deuterium. In the presence of C3D6, the total conversion was kept as low as possible in order to avoid consecutive reactions.2654 REACTIONS OF n-BUTENES ON PALLADIUM FILMS ANALYSIS CHROMATOGRAPHY The products were separated and purified on a semi-preparative gas-chromatograph operating under a pressure of 170 Torr and equipped with a 5-m, $in.column of 30% dimethylsulpholane deposited on fire-brick. The working temperature was - 30°C and a catharometer was used as the detector. Sicce the deuteromolecules are slightly separated by gas-liquid chromatography according to their deuterium content, several careful con- secutive chromatographies were made to avoid any loss of product; each moiecule was obtained with a purity >99%. MASS SPECTROMETRY The various deuterated but-2-enes were analysed with a Varian CH7 mass spectrometer. 70V electrons were used to ionize the molecules. The usual corrections were made for naturally occurring isotopes, and for the first four C-H or C-D fragmentations.For these latter corrections, we used the method developed by Gault and Kemball,20 introducing the correct values for C-D fragmentations, as obtained from the mass spectrum of C4Ds molecules. MICROWAVE SPECTROSCOPY The microwave analysis of [2Hl]b~t-l -enes and ci~-[~H~]but-2-enes has been described previou~1y.l~ The position of the last non-exchanged hydrogen atom in [2H,]but-l-enes was also determined. To do this, we had to assign some transitions for [2H7]but-l-ene isomers, whose rotational spectra had not previously been studied. The assignment of the transitions of [1,1 ,3,3,4,4,4-2H7]but-1-ene, which is reported in table 1, was obtained TABLE 1 .-TRANSITIONS OF [I , I ,3,3,4,4,4-'H,]BUT-l-ENE.transition 414-+515 5 1 4 ' 6 6 1 5 5 1 5 3 6 1 6 calculated observed frequency/MHz 32 785.3 39 902.0 39 341.5 frequencylMHz 32 781.6 39 896.2 39 335.9 after we had predicted the spectrum using the method of Nosberger 21 and available data from the following molecules : [2Ho], cis-['Hl], tran~-[~H~], [2-2M1], [3-*Hl], [l ,3-2H2], [2,3-2H2], [4-'H 1], [5-'H l]but- 1 -enes, [2HO] but- 1 -ene 1 3C and [ 'Hs] but-1 -em. The 1as t two were prepared specifically for this purpose. The [2Hs]bUt-l-ene was prepared using the method of Larson et aZ.,22 and the but-l-ene-l-13C obtained from 13CH31 and propion- aldehyde by the Grignard reaction. The microwave study will be fully described in a separate publication. RESULTS PALLADIUM FILMS CONDENSED AT 0°C All three butenes were reacted at 0°C in the presence of perdeuteropropene, each time on a fresh palladium film that had been condensed at the reaction temperature (films of type I).The deuterium distributions of the various reaction products and the location of the label in some specific deuterated molecules (hyperfine distributions) are given in table 2. All these distributions may be considered as initial or quasi- initial, except the one obtained for the exchanged but-1-ene. Since readsorption and isotopic dilution, in this case could have scrambled the labelling of the [2H,] species, an experiment was made at smaller conversion. The corresponding fine and hyperfine distributions, reported in brackets, show that such a scrambling does not occur.M .J . LEDOUX, F. G . GAULT, A . BOUCHY AND G . ROUSSY 2655 REACTIONS OF BUT- 1 -ENE The exchange of but-1-ene, faster than isomerization, yielded mainly the mono- deuterated molecule [2H,], but also more extensively deuterated molecules in decreasing amounts up to [2H7]. The microwave analysis showed that the deuterium in the [2Hl] molecule was located mainly on carbon atom 3 (21 60 %) and 1 ( ~ 4 0 %), with position 1 -trans largely predominant over 1 -cis. The deuterium distributions in both isomers, trans-but-2-ene and cis-but-2-ene were very similar and included all the deuteromolecules from [2H0] to [2H,], in decreasing amounts. [2Ho] and [2H,] molecules were the major products (60 %), and the microwave analysis of the [2H,]-cis-but-2-ene showed that all the deuterium was located on the same carbon atom 1.TABLE 2.-REACTIONS OF BUTENES IN THE PRESENCE OF C3D6 ON PALLADIUM FILMS CONDENSED AT 0°C (TYPE-I FILM) reacting hydrocarbon but-1-ene temp. time O'C 5 min CSD 6/C& ratio 15.4 weight of film 70.3 mg exch. /isom. 2.6 products B1 trans B2 trans B2* cis B2 (calculated) conversion 83.3 % 5.6 % 11.1 2 isom. ratio by m.s. t cis B2/trans B2 = 2 [ZH 01 57.2 (79.2) 29.1 0.3 37.3 12H 11 26.4 (16.3) 24.3 2.1 27.4 t2&1 8.9 (3.2) 13.4 6.3 13.5 kZHd 3.6 (1.2) 11.9 10.7 9.1 IZH41 2.1 10.7 10.7 6.3 12Hsl 1.2 7.1 7.1 4.0 [2H61 0.5 2.9 2.75 1.9 PH71 0.1 0.6 0.6 0.5 t2Hsl - 0.05 0.05 0.05 ~#1/100 0.73 1.87 1.42 microwave B 1 -[zHl] ci~-B2-[2H 11 analysis trans-but-Zene 16.25 81.9 mg 0.8 0.5"C 10 min B1 trans B2 cis B2 2.5 % 95.9 % 1.4 % 2 96.6 9.2 2.5 1.0 24.6 4 0.3 4.5 5 0.2 2.2 4 0.2 3.2 5 0.4 5.9 16.5 1.1 15.4 39 0.2 22.3 22.5 0.05 12.6 6.18 0.13 4.32 B1-[2H7] cis-B2-[zH1] cis-but-2-ene 0°C 5 min 12.4 26.8 m g 2 B1 trans B2 cis B2 4.6 % 2.0 % 93.4 % Bl/trans B2 = 2.3 1.7 24.6 86.7 8.1 31.0 7.0 12.8 5.3 2.0 14.0 4.1 1.3 13.6 5.2 1 .o 15.9 8.2 0.8 16.6 11.1 0.8 13.3 8.3 0.3 3.4 2.1 0.1 4.29 2.57 0.31 cis-B2-[zH 11 --- I;;\ /--- 71 * Contribution of the rc-allylic mechanism. t Mass spectrometry.R E A C TI 0 N S 0 F franS-B U T - 2 - EN E A N D CiS-B U T - 2 - EN E Both exchange and isomerization were slower for trans-but-2-ene than for but-1 -ene and there was less exchange than isomerization. Two pronounced maxima at C2H1] and [2H6] appeared in the deuterium distribu- tion of the exchanged trans-but-2-ene, while the deuterium distributions of the isomers, but- 1 -ene and cis-but-2-ene, were very different.The main deuteromolecules obtained by double-bond migration were the [2H6], [2H7] and ['HJ but-I-enes ; ['H7] was predominant and the last non-exchanged hydrogen atom in this molecule was located on carbon atom 2. The deuterium distributions of the cis-but-2-ene extended up to [2H8] with two maxima at [2H,] and ['H7] ; the deuterium atom in the mono- deuterated cis-but-2-ene was located exclusively on carbon atom 2. The exchange and isomerization rates of cis-but-2-ene lay between those of2656 REACTIONS OF n-BUTENES ON PALLADIUM FILMS but- 1 -ene and trans-but-2-eneY with an exchange to isomerization ratio very similar to that obtained for but-1-ene.The monodeuterated species [2Hl] was the major exchange product, but the deuterium distribution extended up to [2H,] with a sharp break after [2H6]. In the [2Hl] molecule, about 70 % of the label was located on carbon 2 (or 3) and 30 % on carbon 1 (or 4). The deuterium distributions of the two isomers, but-1 -ene and trans-but-2-ene, were very similar to the distributions of the but-1-ene and cis-but-2-ene obtained by isomerization of trans-but-2-ene, except for the second maximum which lay at [2H6] instead of [2H7]. The shift towards the less deuterated species in the deuterium TABLE 3.-REACTION OF BUTENES IN THE PRESENCE OF C3D6 ON A PALLADIUM FILM CONDENSED AT 470°C (TYPE-IT FILM) reacting hydrocarbon temperature and time C,D6/C4& weight of film exch./isom. product conversion isom. ratio m.s.* [2H,] 12H11 L2H2I [2H31 E”H41 f2H51 12H71 [2H81 [ 2 ~ 6 1 4/1m deuterium content microwave analysis cis-but-2-ene 0°C 10 min 16.4 43.5 mg 0.2 B1 trans B2 cis B2 1.0 % 1.1 % 97.9 % Bl/trans B2 = 0.9 11.0 34.3 16.3 8.0 8.7 9.2 7.4 4.2 0.9 27.8 99.5 56.6 0.5 3.2 2.5 - 2.6 3.3 - 2.4 - 1.2 - 0.4 I - 2.53 1.23 0.005 but- 1-ene 0°C 15.7 65.6 mg 1 trans B2 B1 94.9 % 2.8 % cisltrans = 0.8 94.9 44.9 3.8 33.4 0.7 9.1 0.3 6.5 0.2 3.8 0.1 1.7 - 0.5 - 0.1 - 0.1 0.07 0.99 B 1 -[2H 11 7 10 min cis B2 2.3 % 48.9 34.4 7.8 4.6 2.6 1.2 0.4 0.1 - 0.88 * Mass spectrometry distributions, as well as the decrease in the deuterium content of each isomer, might be due to higher isotopic dilution when passing from the trans-but-2-ene to the cis-but-2-ene experiment.Indeed, the number of hydrogen atoms released on the surface was 0.54 per molecule in the reaction of cis-but-2-ene instead of 0.34 in the reaction of trans-but-2-ene. PALLADIUM FILMS CONDENSED AT 470°C Some reactions were effected on palladium films deposited by evaporation on a glass substrate maintained at 470°C (type I1 films). The deuterium distributions of the reaction products obtained from cis-but-2-ene and but-1-ene at 0°C in the presence of C3D6 are reported in table 3. In the case of but-1-ene, both isomerization and exchange reactions were strongly affected by the thermal treatment, the latter muchM. J . LEDOUX, F . G . GAULT, A . BOUCHY AND G . ROUSSY 2657 more than the former. A value of 1 was obtained for the exchange to isomerization ratio, instead of 2.6 on films of type I.Similarly, the deuterium content of either isomer was much lower than on films of type I : 0.99 and 0.83 for trans-but-2-eiie and cis-but-2-ene instead of 1.87 and 1.42, respectively. The hyperfine distribution of the [2H1]-but-l-ene was also very different from the one obtained on type-I films. The label was located not only on carbon 1 (27 %) and 3 (54 %), but also on carbon 2 (20 3/). TABLE %-REACTIONS OF fl-BUTENES IN THE PRESENCE OF D2 ON PALLADIUM FILMS CONDENSED AT 470°C (TYPE-11 FILM) reacting product trans-but-2-ene cis- but-2-ene temperature and time - 84°C 2 min 30 s - 84OC 2 min 30 s DzIC~HS ratio weight of film product conversion ms.* [2HO] m 1 1 I”H21 m 3 1 E”H41 I”H51 12H61 12H71 I ” b 1 E2H91 [“HI01 4/100 deuterium content microwave analysis B1 1.7 % 7.5 15.8 20.3 19.1 15.2 10.8 6.5 3.4 1.4 10.9 34.5 mg trans B2 cis B2 68.9 % 18.9 % 44.4 11.7 11.8 30.4 12.0 15.5 9.5 12,.8 7.7 10.4 5.9 7.8 4.3 5.8 2.9 3.9 1.3 1.7 3.33 1.81 2.56 ci~-B24’Ht 11 butane 16.5 % 1.1 6.9 15.0 15.9 15.2 13.3 10.7 8.1 6.1 4.8 2.8 4.53 B1 5.5 % 5.4 11.5 15.7 16.0 14.9 13.3 10.9 8.3 4.0 - - 9.9 9.8 mg trans B2 cis B2 butane 50.4 % 33.0 % 11.1 % 6.3 40.0 0.0 19.1 10.2 3.1 12.6 9.8 8.7 11.7 8.4 10.2 11.1 7.6 11.0 11.2 7.3 11.6 11.7 7.1 12.1 10.6 6.1 13.0 5.7 3.2 13.8 - - 11.0 - I 5.4 3.72 3.70 2.33 5.80 B1 -[”Hi] ci~B2-I ’Hi] ?q / 34 * Mass spectrometry, In the case of cis-but-2-eneY exchange and double-bond migration were much more decreased than cis-trans isomerization, with an exchange to isomerization ratio of 0.2, instead of 2 on type-I films, and a but-l-ene/tram-but-2-ene ratio of 0.9 instead of 2.3.The deuterium content of the but-1-ene isomer was decreased (2.5 instead of 4.3), and the monodeuterated species became predominant in the deuterium distribution, although a secondary maximum at I2H5] was still present. The deuterium pattern of the tramisomer was very similar to the one obtained on type-I films, with [2Ho] and [2H,] predominant, but the perdeuterated molecules [2M,]-[2H,] were in much smaller amounts (15.6 as coinpared with 44.3). Some experiments were done at -84°C on type-I1 films in the presence of deuterium. The reactions were very fast and deuteration to butane took place to some extent (11-17 %).Although the readsorption processes could no longer be neglected, the deuterium distributions of the various products, reported in table 4, are reminiscent of the distributions obtained in the presence of perdeuteropropene.2658 REACTIONS OF n-BUTENES ON PALLADIUM FILMS On account of the larger amounts of material available, the microwave analysis of some monodeuterated species were made, which could not be done in the previous experiments. ~is-[2-~H,]but-2-ene is the only monodeuterated species obtained from trans-but-2-ene. In the [2Hl]but-l-ene obtained from trans-but-2-ene, all the deuterium was located on carbon 3. The specificity of the labelling is all the more significant as the hyperfine distribution of the exchanged close to equilibrium.DISCUSSION EXCHANGE (TYPE-I FILMS) EXCHANGE OF BUT- 1 -ENE cis-[*Hl]but-2-ene is very The hyperfine distribution of the [2H 1] species allows the Horiuti-Polanyi mechanism for exchange to be ruled out. (1) If the alkyl-alkene reversal were operative, the [3-2Hl]but-l-ene could only be explained by assuming two consecutive steps and very high isotopic dilution, [2-2Hl]but-2-enes should then be observed simultaneously, but they are not : SCHEME 1 (2) Since the classical Horiuti-Polanyi mechanism involves as intermediate a half- hydrogenated state, with both hydrogen atoms of the CHzD group equivalent, equal amounts of cis-[ 1-2Hl] and trans-[ 1-2Hl] species should be observed, whereas the labelling of carbon 1 in but-1-ene is actually highly dissymmetrical. H SCHEME 2 (3) According to the Horiuti-Polanyi mechanism, the observed very poor exchange- ability of the hydrogen atom located on carbon 2 would imply that adsorbed sec-butyl species are much more reactive than n-butyl species.Although differences have been observed between the reactivities of these two 24 (most often the n-butyl form is more reactive), the differences are never pronounced enough to explain the very high observed [1-2H,]/[2-2H,] ratio. The present results thus force us to conclude that the associative Horiuti-Polanyi mechanism does not operate for but- 1 -ene exchange and that dissociative adsorption takes place. In a previous paper,18 it was suggested that metal atoms of high coordination number (isolated adatoms or corner atoms) could promote dissociative adsorption.In these sites, C, the presence, at least, of three free valencies allows exchange according to a vinylic or an allylic mechanism. We believe that both exchange mechanism operate concurrently on palladium films. The allylic-type mechanism,M. J . LEDOUX, F . G . GAULT, A . BOUCHY AND G . ROUSSY 2659 first proposed by molecules (scheme could account for 1 Rooney and Webb,25 would explain the formation of [3-2Hl] 3). The vinylic-type mechanism, first introduced by Farkas et d.,' :he dissymmetrical labelling of positions cis-1 and trans-1 (scheme 4). SCHEME 3 H-M-0 - 1 ..c & ~is-[l-~Hi]but-l-ene H-M-D SCHEME 4 According to the Farkas mechanism, there is no reason for having equal amounts of tran~-[l-~H~] and ci~-[l-~H,] molecules.On the contrary, a difference in reactivity is expected between the two hydrogen atoms of carbon 1, cis-1 and trans-1, on account of the steric interaction between the substituent and the carbon-metal bond in the resulting a-vinylic adsorbed species. I \- + -Fee -c + I ,Fe, .I. + i -- -Pd- L/ .. . . . -cc + I -Pd- FIG. 1 .---Free energy diagrams showing differences in dissociative adsorption of but-1-ene between iron and palladium. The very small amount of [2-2Hl]but-l-ene could also be due to a vinylic-type mechanism of exchange. [2-2Hl]but-l-ene was the major product in the exchange of but-1-ene on iron l6 and this metal is very suitable for vinylic dissociative adsorption,2660 REACTIOKS OF n-RUTENES ON PALLADIUM FILMS as shown by the very fast and exclusive exchange of the three vinylic hydrogeil atoms of but-l-ene.'* The difference in behaviour between iron and palladiuix, where dissociative adsorption of but-l-ene involves carbon atoms 2 and 1, respectively, could be explained by the free energy diagrams represented in fig.1. In these diagrams, we assume that the free energy (F.E.) levels of the activaied co:-i.iyicxes corresponding to the three adsorbed species, although arranged in the same order on both metals, are considerably raised in the case of iron relative to the F.E. levels of allylic mechanism associative mechanism FIG. 2.-Mechanisms for formation of ~is-[l-~H~]but-2-ene and [l-2Hl]but-l-ene, tram-[ 2-2Ii.11 but-2-ene associative mechanism r._-_-_.__---_____----~------------- FIG. 3.-Mechanisiii of formation of ~is-[2-~H~]but-2-ene. the reactants.That would explain why on iron the exchange at carbon 1 could not be detected in the [21-Il]b~it-l-er,e iiiolecufes. In the case of paIladium, the very poor exchange at carbon 2 could be exp!ained if desorption is thc ratedetermining step and involves too high a free energy o f activation. EXCHANGE OF CiS-BUT-2-ENE The results obtained in the case of cis-but-2-ene confirin well the existence of two dissociative mechanisms for exchange, allylic and vinylic, responsible for thc two reaction products, [ 1 -2H and [2-2H,]cis-buf-2-ene, respectively. We believe that ei~-[l-~H~]but-2-ene is obtained by an allylic mechanism. It is difficult indeed to accaunt for this species by the classical Horiuti-Polanyi mechanism, except by assuming isotopic dilution and a number of consecutive steps which should also provide [1-2Hl]but-l-ene : this species is not observed in the reaction products (fig.2). Nor can the major exchange product, .ci~-[2-~H~]but-2-ene be explained byM. J . LEDOUX, F . G . GAULT, A . BOUCHY A N D G. ROUSSY 266 1 a simple Horiuti-Polanyi mechanism, unless one assumes at least two consecutive alkyl-alkene reversals (fig. 3). Since formation of tran~-[~H~]but-2-ene involves only one of these steps, one should observe more trans-[2Hl]isomer than exchanged ~ i s - [ ~ H ~ ] . The reverse was true : the amount of ~is-[~H,]but-2-ene obtained by exchange (4.5 %) was much larger than the amount of tran~-[~HJbut-2-ene obtained by cis-trans isomerization (0.6 %).We believe, therefore, that ~is-[2-~H,]but-2-ene is formed by a vinylic type dissociative mechanism. If the vinylic but-2-enyl adsorbed species is placed in the free energy diagrams at a level very close to the internal but-1-enyl species and if the F.E. level of the reacting cis-but-2-ene is much below the level of but-1-ene, one would expect that, on iron, cis-but-2-ene exchange would be much slower than but-1-ene exchange. Indeed, the exchange of cis-but-2-ene was almost non-existent on iron, and that is a good confir- mation of the above proposals. 0 2 + + A2 FIG. 4.-Energy diagram for double bond migration in but-1-ene. DOUBLE-BOND MIGRATION (TYPE-I FILMS) BUT-2-ENE -+ BUT- 1 -ENE ISOMERIZATION The characteristic deuterium distribution of the but- 1-enes obtained from truns- but-2-ene included three highly deuterated species [2H6], and [2H8].This distribution should be related to that of the [2H6]- [2H,]cis-but-2-enes and to that of multiply exchanged trans-but-2-enes with a pronounced maximum at C2H6]. We believe that all these highly deuterated molecules are obtained by an allylic mechanism involving several interconversions between n-olefinic and n-allylic species. Such a mechanism was proposed previously to explain the complete exchange of cycloalkanes on palladium 1* and was rejected on the basis of a single e~periment.~ As will be shown, the labelling of the C2H7]but-l-ene and the [’H6] maximum in the exchange pattern of trans-but-2-ene are definite proofs of the existence of such a mechanism in the case of olejn reactions on palladium.The break after [2H6] in the distribution of truns-but-2-enes, very similar to the break after [2H,] in the exchange pattern of 1,2 dimethylcyclopentene, implies a very fast exchange of all allylic hydrogens, best explained by very fast interconversions2462 REACTIONS OF n-BUTENES ON PALLADIUM FILMS between allylic species and n-olefinic species having the same but-2-ene structure ; in this process the conversion into olefinic adsorbed species with but-1-ene structure is a difficult step. In order to explain double-bond migration and the deuterium pattern of but-1-ene, then, we have to admit that two different asymmetrical n-allylic species exist on the surface, interconverting through a transition state which could be a symmetrical n-allylic species.In fig. 4, species A, represents an asymmetrical allylic species where thep, orbitals of carbons 2 and 3 interact more strongly with the CT metal orbitals than does the p , orbital of carbon 1. Such a species would then part 1 y retain the cis- but -2-ene configuration. In species A,, conversely, the C, and C2 pz orbitals interact more strongly with the metal than does the C3 pz orbital. The high energy barrier between species A, and A2, which corresponds to the symmetrical n-allylic species, prevents fast inter- conversions between the two n-olefinic species 0, and 02. But-2-ene -+ but-1-ene isomerization may thus be represented by the following succession of steps (fig. 5). 4 D3 M s k w [ Hs]-cis-but-2- ene 4 HD D3 4 ki=-!-b-&+D~~ D U ?JJ [1,1 .3,3,4,4,4-2H7]but-1-ene * -.-.4 - *2 H H H D H D H D 2 H 02 Al 01 *I 01 A! A2 O2 A2 A i Oi A i (1) Fast interconversions between O2 and A, exchange the six hydrogen atoms of (2) Species A2, quintuply deuterated, converts into species Al. (3) Interconversions between A, and O1 exchange a sixth and a seventh hydrogen atom on carbon 3. Desorption at this stage yields C2H7]but-l-ene with the last hydrogen remaining on carbon 2. (4) In order to form [2H8]bUt-l-ene, two more interconversions between A,-type and A2-type allylic species are required. Interconversions between .n-olefinic and n-allylic species, 0 1 and Ai, should be faster than olefin desorption, since the amounts of [1,1,2,3,4,4,4-2H7]but-l-ene are negligible. Consequently, [2H5] and [2H6]bUt-l-eneS derive from adsorbed trans- but-2-enes that have exchanged only part of their methyl hydrogens.(5) Free rotation around the C2-C3 bond in O1 and reversal to adsorbed cis- but-2-ene could account for the ci~-[~H,]- [2H8]but-2-enes. The above mechanism thus explains the distribution of deuterium in but-I-ene, including the location of the deuterium atoms on the [,H,] species, the multiple exchange of trans-but-2-ene, and the highly deuterated cis-but-2-enes. FIG. 5.-Steps in but-2-ene-tbut-1-ene isomerization. the methyl groups. Desorption at this stage yields tran~-[~H,]but-2-ene. BUT- 1 -ENE 4 BUT-2-ENE ISOMERIZATION Since double-bond migration is accompanied by extensive exchange of but- 1-ene, the but-2-ene deuterium distributions are strongly altered by isotopic dilution.To take this effect into account, the deuterium patterns of the more highly deuteratedM. J . LEDOUX, F . G . GAULT, A . BOUCHY AND G . ROUSSY 2663 but-2-ene isomers have been recalculated, using as true initial distribution that of the but-1-ene obtained from trans-but-2-ene and using several values of the deuterium mole fraction x . The best fit is obtained with x = 0.51, and the corresponding distributions, calculated by adjusting [2H4] and [2H5], are reported in the fourth column of table 2. For the [2H,]-[2H,] species, excellent agreement with the observed values was obtained. One may consider, then, that these molecules, and also most of [2H3] and half of [2H2], are obtained by the same allylic mechanism Al as proposed above for the but-2-ene + but-1-ene isomerization.However, part of the trans- but-2-ene distribution, including E2H0] = 28.8, ['H1] = 22.2, [2H,] = 7.2 and [2H3] = 1.2, was obtained by some other process. ,y/-- ;?I-.-. \f/ -. [2Ho] M-rr M-H M 4 FIG. 6.-Mechanism accounting for [2Ho]-[zH3] distribution. A combination of an intramolecular hydrogen shift and the Horiuti-Polanyi mechanism could account for the above [2Ho]-[2H,] distribution. 1-2 double-bond migration takes place almost exclusively without deuterium incorporation on iron films and was interpreted by a sigmatropic mechanism.' ' A non-repetitive Horiuti- Polanyi mechanism accounts for the [ 2H 1] species obtained by cis-trans isomerization on most metals.ll However, an alternative mechanism could be proposed, accounting for the whole [2Ho]-[2H3] distribution.Such a mechanism, A2, basically of the same nature as the allylic mechanism Al, would involve, instead of C sites, B sites with only two free valencies available. A decreasing [2Ho]-[2H3] deuterium distribution is to be expected according to fig. 6 , provided that replacement of adsorbed hydrogen by adsorbed deuterium is as fast as or slower than olefin desorption. CiS-tranS I S 0 M ERI Z A TI ON The distribution of deuterium in the cis- or trans-isomer obtained in this reaction includes, besides the highly deuterated species discussed previously, [2H,] and [2Hl] molecules. Direct cis-trans isomerization, without deuterium incorporation, is commonly observed on transition metals.l1. 26 The mechanism for this reaction involves a weakening of the C2-C3 bond and free rotation, either by some electron transfer to the metal [scheme 5(b)] or by partial hydrogen displacement [scheme 5(a)].The latter mechanism is suggested by the result of nb initio calculations 29 showing that, during the course of a sigmatropic migration, the migrating hydrogen interacts with carbon 2 before any bonding with carbon 3 becomes significant. It was proposed that highly coordinated surface atoms, the ones of low Miller index faces, with one single free valency available, were responsible for these reactions (scheme 5a). Isomeric2664 REACTIONS OF n-BUTENES ON PALLADIUM FILMS monodeutero-cis-but-2-ene, as shown by microwave analysis, retains all the label on carbon 2 or 3, which is consistent with a simple Horiuti-Polanyi mechanism. The break after [,H1] in the deuterium distribution suggests that, in contrast to what is conventionally believed, very few alkyl-alkene reversals take place before desorption.That three consecutive steps are required to form traizs-[*H,], instead of one step to form [2H,], makes the break more pronounced than it would be in the case of the but-2-enes obtained from but-1 -ene by double-bond migration. In the latter case, 9 M free rotation L SCHEME S(u).-Signiatropic mechanism. SCHEME S(b).-Charge transfer to the metal, I Dv ci~-[~Hi]but-2-ene M-H double bond migration SCHEME 6 -___________-_______---.------------- two and one steps are required to form the di- and mono-deuterobut-2-enes respec- tively (scheme 6). Thus formation of [,H,]- [2H,]but-2-enes from but-1-ene may be simply explained by the classical Horiuti-Polanyi mechanism and does not require the introduction of an additional allylic mechanism A,.As already pointed out,18 it is believed that the associative mechanism requires sites with two free valencies available (sites B) such as the sites arising at the intersection of two-secant low-index faces (edge atoms) (scheme 7). SCHEME 7M. J. LEDOUX, F. G. GAULT, A. BOUCHY AND G. ROUSSY 2665 GENERAL DISCUSSION REACTIONS ON FILMS CONDENSED AT 470°C A combination of mass spectrometric and microwave analyses allows one to characterize a number of reactions, which may be classified as follows : (A) direct cis-trans isomerization and sigmatropic double-bond migration ; (B) Horiuti- Polanyi mechanism for cis-truns isomerization and double-bond migration.This associative mechanism yields mainly singly deuterated molecules ; (C) dissociative mechanisms including a vinylic adsorption operative for exchange, and an allylic adsorption involved in exchange and in double-bond migration with extensive deuterium incorporation. FIG. 7. According to the number of ligands required in the various reaction mechanisms, these three classes of reactions were related to three types of sites A, B and C , differently coordinated and having respectively one, two, or three free valencies available. Some of these sites may accommodate two different species. Allylic and vinylic adsorptions for instance, compete on C sites. If one assumes that the free energy level of the activated complexes corresponding to the allylic species (dotted line in fig.1) are placed above the ones corresponding to but-2-enyl species and below the ones corresponding to but-1-enyl species, the inversion of the ratio between allylic- type and vinylic-type exchange when passing from but-1-ene to but-2-ene could be explained. One advantage of the above definition of the active sites A, 33 and C is that they may be identified with the normal low-index surface atoms (A sites), or with some of the defects arising in a rough surface : edge atoms (B sites) corner or adatoms (C sites) (fig. 7). If such an identification is correct, any modification of the metal surface should modify the relative amounts of sites A, B arid C, and consequently the relative contributions of the various mechanisms.In this paper, we give an example of such a surface modification when palladium films are deposited on a substrate maintained at high temperature. One might expect that, under such conditions of film preparation, C sites and type C reactions will be drastically suppressed. Comparison between the deuterium distributions obtained on films condensed at 0 and 470°C (tables 2 and 3) shows that this was indeed the case. Ali the highly deuterateci molecules obtained by allylic-type isomerization were2666 REACTIONS OF n-3UTENES ON PALLADIUM FILMS drastically reduced : ['H3]- ['HJbut-1 -enes and trans-but-2-ene from cis-but-2-ene and [2H3]- [2H,]but-2-enes from but-1-ene. Similarly, both allylic- and vinylic- type exchange of but-1-ene and but-2-ene were strongly decreased.Lastly, the very high cisltrans ratio in but-1-ene isomerization (2) and the but-1-eneltrans ratio in cis-but-2-ene isomerization (2.3) on normal films, which could have been explained by the preferential syn-configuration of the .n-allylic species,27* dropped drastically to 0.8 and 0.9, respectively, on treated films; this finding, too is consistent with the disappearance of C sites and type-C reactions. On films condensed at 470°C, the major remaining reactions, then, are those occurring on A and B sites : direct cis-trans isomerization and isomerization according to a sigmatropic and Horiuti-Polanyi mechanism. The latter mechanism even became significant for exchange, accounting for the [2-2H,] species in the hyperfine distribution of but-1-enes, whereas it was negligible on normal rough palladium films.Another possible explanation of the results obtained on films type 11, suggested by one of the referees, is a selective contamination of the surface by carbon deposit during film preparation. However, it was checked by using Leed and Auger that a clean palladium surface [either oriented (1 11) or polycrystalline] does not retain a significant amount of carbon when exposed to hydrocarbon or CO contaminants at temperature lower than 600°C.30 J. J. Rooney, F. G. Gault and C. Kernball, Proc. Chem. SOC., 1960,407. F. G. Gault, J. J. Rooney and C. Kemball, J. Catalysis, 1962, 1, 255. J. J. Rooney, J. Catalysis, 1963, 2, 53. L. Hilaire, G. Maire and F. G. Gault, Bull. SOC. chim. France, 1967, 886. R. L. Burwell and K. Schrage, J. Amer. Chem. SOC., 1966,88,4549. J. A. Roth, B. Geller and R. L. Burwell Jr., J. Res. Inst. Catalysis Hokkaido Unio., 1968, 16, 221. H. A. Quinn, J. H. Graham, M. A. McKervey and J. J. Rooney, J. Catalysis, 1971, 22, 35. L. Hilaire and F. G. Gault, J. Catalysis, 1971, 20, 267. R. Touroude, L. Hilaire and F. G. Gault, J. Catalysis, 1974, 32, 279. lo R. Touroude and F. G. Gault, J. Catalysis, 1974, 32,288. l 1 R. Touroude and F. G. Gault, J. Catalysis, 1974, 32,294. l2 J. Horiuti and M. Polanyi, Nature, 1933, 132, 819 and Trans. Faraday Soc., 1934, 30, 663. l 3 A. Farkas, L. Farkas and E. K. Rideal, Proc. Roy. Soc. A , 1934, 146, 630. l4 K. Hirota and Y. Hironaka, Bull. Chem. SOC. Japan, 1966, 39, 2638 and J. Catalysis, 1965, l5 T. Ueda, J. Hara, K. Hirota, S . Teratani and N. Yoshida, 2. phys. Chem. (Frankfurt), 1969, l 6 M. Ledoux, F. G. Gault, J. J. Masini and G. Roussy, J.C.S. Chem. Comm., 1975, 1034. l7 F. G. Gault, M. Ledoux, J. J. Masini and G. Roussy, Proc. VIth Int. Congr. on Catalysis (The Chemical SOC., London), p. 469, vol. 1. l 8 M. J. Ledoux, Nouv. J. Chim., 1978, 2, 9. l 9 C. Kemball, Proc. Roy. SOC. A, 1951,207,539. 2o F. G. Gault and C. Kemball, Trans. Faraday SOC., 1961, 57, 1781. 21 P. Nosserger, A. Bauder and Hs. H. Gunthard, Chem. Phys., 1973, 1,418. 2 2 J. G. Larson, J. W. Hightower and W. Keith Hall, J. Org. Chem., 1966, 31, 1225. 23 T. Ueda, Proc. Vth Int. Congr. on Cutalysls, Miami, 1972, 1, 431. 24 E. Hirota, M. Ito and T. Weda, Proc. VIth Int. Congr. on Catalysis (The Chemical SOC., London) 2 5 J. J. Rooney and G. Webb, J. Catalysis, 1964, 3,488. 26 M. J. Ledoux, Thisse Etat (Strasbourg, 1977). 27 G. C. Bond and M. Hellier, J. Catalysis, 1965, 4, 1. 28 R. Guisnet, G. Perot and R. Maurel, J. Chim. phys., 1972, 69,1059. 29 J. P. Grima, F. Choplin and G. Kaufniann, J. Organomet. Chem., 1976, 124, 315. 30 P. Legare, Y. Holl and G. Maire, personal communication. 4, 602. 64,64. p. 518, vol. 1. (PAPER 81360)