Platinum Metals Rev., 1983, 27, (1), 16
Metal-Support and Metal-Additive Effects in Catalysis
A Report on an International Symposium
The advantages of using supported metal catalysts in preference to unsupported metals such as colloidal platinum or platinum black have long been appreciated. The support facilitates the formation of extremely small metal particles having a high proportion of their atoms at the surface; the particles have remarkable thermal stability and the presence of the support allows the incorporation of beneficial additives, usually known as promoters. For many years supports were thought to be catalytically inert, except in the special case of bifunctional platinum-alumina catalysts for petroleum reforming. Recently there has arisen evidence that in some systems at least the support can exert a marked influence over the properties of metal particles residing on it. When the platinum metals are supported on titania and reduced in hydrogen at 500°C they lose most of their ability to chemisorb hydrogen and carbon monoxide, and for certain reactions they also lose much of their catalytic activity. This has been attributed to a strong metal-support interaction (SMSI), and its discovery has awakened interest in the whole question of the nature of the interaction of metal particles with the support and with any promoters which may be present.
The first International Symposium devoted specifically to this aspect of catalysis was held at Ecully, near Lyon, from the 14th to the 16th September 1982, and was organised by the Institut dc Recherches sur la Catalyse with financial support from the c.n.r.s. It was attended by some 180 persons from a number of countries including the People’s Republic of China; six plenary lectures and thirty-five papers were presented during the three-day meeting. Two of the plenary lectures and at least twelve of the papers bore directly on the SMSI question. Rather than attempt to summarise each paper, it is perhaps better to identify their principal themes and to indicate the current state of understanding of the problem by the use of selective examples.
Strong Metal-Support Interactions
As mentioned above, the prime indicator of a SMSI is the loss of the metal’s ability to chemisorb molecules which normally react without difficulty at metal surfaces: hydrogen and carbon monoxide are the usual shibboleths. The effect of reduction in hydrogen at about 500°C is not to cause particle growth by sintering, so the loss of chemisorption capacity cannot be attributed to this. However transmission electron microscopy has revealed suggestions of alteration in the morphology of metal particles, as exemplified by the work of R. T. K. Baker at Exxon Corporate Research Science Laboratories with platinum on titania. However the appearance of ‘pill-boxes’ of metal is not restricted to systems showing SMSI and there is dispute among the experts as to the proper interpretation of what is seen.
Another notable feature of SMSI is that it is reversed by oxidising conditions. Oxygen chemisorption is not suppressed, and any explanation of the effect must take account of this.
It has often been suggested that the origin of the SMSI effect, at least as it applies to titania, lies in a modification of the metal’s electronic character. There is evidence that spillover of hydrogen atoms from the metal to the titania can occur at quite low temperatures, generating first Ti3+ and OH ions, the latter subsequently forming anion vacancies and water, according to the equations:
where the symbol indicates a surface anion vacancy. It is then postulated that electron transfer takes place in the sense
As with all explanations of SMSI so far proposed, this idea is not without its critics since the first electron to move to the metal particle will tend to repel others, and so the transfer is thought to be implausible on the grounds of simple electrostatics. Nevertheless small changes in the binding energies of the 4d5/2 and 4f7/2 electrons of platinum in the expected sense have been detected by T. Huizinga and R. Prins of the Eindhoven University of Technology using X-ray photoelectron spectroscopy, while the work of J.-M. Herrmann, J. Disdier and P. Pichat at the Institut de Recherches sur la Catalyse, Villeurbanne, on the electrical and photoconductivity of platinum on titania provides supporting evidence.
The response of the specific rates of catalysed reactions to variation in metal particle size, and hence it is thought to variation in surface topography, can be placed in one of two categories. If there is no response, the reaction is said to be structure-insensitive: many simple hydrogenations, for example, of alkenes and of benzene, appear to fall into this class. If there is a response, either positive or negative, the reaction is called structure-sensitive: hydrogenolysis of hydrocarbons, and carbon monoxide methanation and the Fischer-Tropsch reaction, behave in this way, showing higher and lower specific activities, respectively, with increase in the metal dispersion. Now structure-insensitive reactions show relatively small activity changes when the metal moves from the normal to the SMSI state, while hydrogenolysis reactions generally show very large activity decreases. The work of D. E. Resasco and G. L. Haller, of Yale University, demonstrates that this effect is particularly associated with the smaller metal particle size fraction. The reactions of carbon monoxide proceed if anything more rapidly on metals in the SMSI state, a fact attributed by P. B. Wells, R. B. Moves, S. R. Morris and R. Whyman at Hull University to the role of the product water in neutralising the SMSI.
Studies on Insulator Supports
The phenomena of SMSI are not confined to titania although most of the work so far carried out has used this oxide as support. According to S. E. Wanke, J. Adamiec, B. Tesche and U. Klengler at the University of Alberta, platinum on magnesia shows similar effects, not yet so deeply studied; ceria behaves analogously, but in these systems the loss of reactivity is more reasonably attributed to the formation of inter-metallic compounds such as Pt5Ce.
There is still much to learn about the interaction of the platinum metals with insulator oxide supports such as gamma-alumina. E. Lesage, H. Dexpert, E. Freund and J. P. Lynch at the Institut Français du Pétrole presented the results of an elegant study of palladium on this support, showing the occurrence of epitactic growth of the metal, not previously suspected. Transition metal carbides and polyamides as supports for noble metals featured in the plenary lecture by S. J. Teichner, C. Hoang Van and M. Astier, Université Claude Bernard (Lyon I), Villeurbanne.
The second part of the Symposium was concerned with promoters and poisons. Their effects are not easily distinguished, as certain promoters which are used to improve selectivity actually diminish the rate, and are sometimes referred to as selective poisons. This is not a well-charted area, but several of the papers presented at this Symposium threw considerable light on the subject. The synthesis of methanol from hydrogen and carbon monoxide catalysed by supported palladium catalysts under mild conditions is notoriously sensitive to the acidity/basicity balance of the support. Despite the helpful paper by J. W. Geus, V. Poneč, E. K. Poels and R. Koolstra from the State University Leiden and the State-University Utrecht, the role of the basic element is still far from clear.
There is still considerable interest in improving the selectivity of the partial reduction of alkynes and alkadienes catalysed by palladium. J. P. Boitiaux, J. Cosyns and G. Martino at the Institut Français du Pétrole showed that selective poisons such as carbon monoxide and pyridine are effective in improving the alkene yield from reactants such as acetylene or isoprene. The selective dechlorination of polychloroanilines to 3,5-dichloroaniline can be controlled by the addition of hydrochloric acid.
The texts of the plenary lectures and papers are now published by Elsevier as Volume 11 of their ‘Studies in Surface Science and Catalysis’ series. Unfortunately the discussion of the papers could not be incorporated in it, the editors preferring to publish the texts at the earliest opportunity. The volume will however be read with interest by all concerned with the structure and behaviour of supported metal catalysts.