Faraday Discuss. Chem. Soc., 1985, 80, 7-15 Twenty-second Spiers Memorial Lecture The Gas-Solid Interface BY JAMES A. MORRISON McMaster University, Hamilton, Ontario, Canada Received 20th September, 1985 It is a sobering and humbling experience to be asked to give this 22nd Spiers Lecture in memory of the first secretary of the Faraday Society. The high standard which has been set by preceding Lecturers is nothing short of intimidating. Thus, Mr President, I should like at the outset to express my deep gratitude for the confidence which has been expressed in me by the Council of the Faraday Division. My involvement with the former Faraday Society began at an early stage in my career when the Society held Discussion No. 14 in Canada in September, 1952. Several of us were ‘requisitioned’ by the late President of the National Research Council of Canada, Dr E.W. R. Steacie, to look after the forty-five members of the Society who came from Europe to attend the Discussion, and who visited a number of Canadian universities and industries. I believe that this was the first occasion when the Society held one of its meetings outside of the United Kingdom. The expedition was supported by a grant of S250 from the Royal Society, which was a generous offering for the time! The main reason for the Discussion being held in Canada will not appear in any of the records. It was, in brief, that Dr Steacie wanted Canadian societies to aspire to higher standards for their scientific meetings, and he invited the Faraday Society to set an example. There is much evidence that the strategy worked and, moreover, a tradition was established. We are now beginning the fifth Faraday Discussion to be staged in Canada. I hope you will not find it too peculiar if I approach the subject of the gas-solid interface from a rather distant and superficially unrelated point.One of my col- leagues in the Department of Music, William Wallace, who is a composer, used to have lunch regularly with a group from the Department of Physics. I was not privy to the conversations which were held over lunch, but I should like to describe one of their consequences for you briefly. At some point, Wallace was introduced to the idea of random number selection and was challenged to use it to compose a piece of music. He accepted the challenge and produced a composition with the title Ancaster Serenade, named after the village where I live on the western edge of Hamilton.In constructing his composition, he assigned notes to particular digits, e.g. C to 1, D to 2, E to 3 etc., and then drew numbers at random from the Ancaster section of the local telephone directory. The digit 9 was taken to be wild and could be used to satisfy any needs of harmony. All of the numbers in the Ancaster section have seven digits and. are of the form: 648-abcd. Thus, a theme or motif was automatically established by the recurring sequence 648. When the music received its first public performance, residents of Ancaster who were in the audience were advised to listen carefully because their numbers might come up. 78 SPIERS MEMORIAL LECTURE Table 1.Preceding related Discussions no. date topic 40 1965 Intermolecular Forces 52 1971 Surface Chemistry of Oxides 55 1973 Molecular Beam Scattering 58 1974 Photoeffects in Adsorbed Species 59 1975 Physical Adsorption in Condensed Phases 66 1978 Structure and Motion in Molecular Liquids What has this odd tale got to do with a discussion of physical interactions and energy exchange at the gas-solid interface? We shall soon see that there is a repetitive theme: the accurate prediction of adsorptive behaviour depends critically on a detailed knowledge of adsorbate-substrate and adsorbate-adsorbate interaction potentials. Professor Steele may be excused if he develops a feeling of d&h uu. His fine and useful book' of ten years ago is constructed around the same theme.BACKGROUND The Faraday Society, and now the Faraday Division of the Royal Society of Chemistry, has held a series of Discussion on surfaces and intermolecular forces over the past twenty years. Their titles and dates are summarized in table 1. Several of the contributors to the present Discussion participated in one or more of the earlier ones. In his Spiers Memorial Lecture at Discussion No. 40, Professor H. C. Longuet-Higgins2 presented a thoughtful account of the fundamental origin of intermolecular interactions in terms of basic forces. Several of the accompanying papers were concerned with the main themes of the present Discussion: physical interactions and energy exchange at the gas-solid interface. For instance, Professor Douglas Everett3 dealt with the determination of the second virial coefficient for the two-dimensional adsorbed gas and with the restriction it placed on the form of the interaction potential between two adsorbed molecules.The data which were analysed (for noble gases and CH, on graphitized carbon black) bear a haunting relation to our consideration of the structure of overlayers of simple molecules, their mobility and the scattering of atomic and molecular beams from them as they are held in the vicinity of the surface of graphite. Does this mean that our understand- ing of these simple adsorbing systems has advanced rather little over a period of twenty years? The answer to that question is unequivocally no, as the papers to follow will clearly demonstrate. In the first place, powerful new experimental techniques now give us exquisitely detailed information about the structure of overlayers and the dynamics of adsorption at solid surfaces.Quite generally, experiments with beams and jets are having a profound effect on many areas of physical chemistry/chemical physics. Surface studies are among those being greatly affected. In the second place, we can deal successfully theoretically with interactions between spherical atoms and we know precisely the form of long-range interactions for more complex particles. However, as will come out in the course of the Discussion, it is still troublesome for us to deal adequately with short-range interac- tions between molecules and a surface. To complete these general introductory remarks, I should like to draw your attention to the founding of the Faraday Society in 1903 'to promote the study of electrochemistry, electrometallurgy, chemical physics, metallography and kindredJ. A.MORRISON 9 Table 2. Approximate energies (in kJ mol-’) for the krypton-graphite system energy Kr-graphite 11.5 Kr-Kr 1.7 graphite corrugation 0.3 subjects’. That, incidentally, could easily be a description of much of modern materials research. The tradition it established and which carries on is to encourage multidisciplinary studies. Thus, as a group, we are physicists and chemists, theorists and experimentalists who share a deep interest in the gas-solid interface. To affirm the merit of this kind of multidisciplinary interaction, we need only think of how rapidly the innovative theory of Kosterlitz and Thouless4 on two-dimensional melting (the purest of pure physics) has led us to accept that phase diagrams for physisorbed layers can be extraordinarily complex. While the subject of phase transitions in adsorbed layers is the theme of the next Faraday Symposium to be held in December 1985, we shall be discussing here the equilibrium structure of certain overlayers and the consequences of that on the scattering of atomic beams.Throughout, we shall be conscious of the pivotal role now being played by molecular simulations in mediating the results of theory and experiment. INTERACTIONS At the beginning of a closer look at the substance of this Discussion, we should note the relative magnitudes of the interaction energies for our two-dimensional systems.To take the particular example of the krypton-graphite system, approximate values are listed in table 2. Here, the value given for the Kr-graphite interaction corresponds to a surface coverage of about a monolayer. It would be larger for a single atom on a bare surface. The Kr-Kr interaction is estimated as the heat of fusion. In the adsorbing system, a balance of interactions is struck which, as we now know, is often delicate and leads to subtle phase changes, partial wetting and other fascinating phenomena. The variation of the interaction energy with the amount adsorbed can be consider- able even in simple cases such as in the one illustrated in fig. 1. Here, the isosteric heat of adsorption of xenon on graphite, as determined ~alorimetrically,~ is plotted against the amount adsorbed, n,. These results were used to refine the well depth for a xenon-carbon interaction potential.6 The graphite surface is obviously not perfectly homogeneous (regularly periodic) as the higher heats of adsorption at the lowest surface coverages indicate.Nevertheless, an accurate value of the interaction energy of a xenon atom with a carbon surface can be obtained by extrapolation to n,=0, as is indicated by the dashed line in fig. 1. The contribution of the Xe-Xe interactions is given by the excess of qst over that for n , = 0 . The monolayer is marked by the abrupt fall in qst at higher surface coverages. Had the experiments been carried far enough, qst would have fallen to approximately the heat of vaporiz- ation of liquid xenon: 12.5 kJ mol-’.This example is an uncomplicated one from the points of view of both theory and experiment. An encouraging contribution to this Discussion takes us a step further. Talbot et u1.’ have investigated N2 absorbed on graphite by means of10 SPIERS MEMORIAL LECTURE 24 22 - I d : 20 \ c": 18 16 0 0.5 1.0 1.5 2.0 2.5 n,/ lo-) moI Fig. 1. Isosteric heat of adsorption of xenon on Grafoil at T = 195.5 K5 (with permission of Chemical Physics Letters). molecular dynamics simulation with Lennard-Jones-type potentials for both molecule-surface and molecule-molecule interactions. A gratifying degree of agree- ment is achieved with the experimental values of the heat of adsorption* over a range of surface coverages less than a monolayer. We shall see soon that measure- ments of heats of adsorption (and of adsorption isotherms) can possibly give us more than just numbers for theory to shoot at.In particular, effects observed in the course of the measurements can alert us to subtleties of the balance of interactions. The first three contributions to the programme of the Discussion are concerned with light-atom scattering from surfaces. Danielson et 0 1 . ~ have investigated a prototypic system with a minimum number of electrons from which the scattering of the He atoms takes place. By using very low-energy beams, they are able to make a close comparison with a semiempirical pairwise potential which accounts for nearly all of the observations. By contrast, the attempt by Jonsson and Weare" to refine elastic scattering from physisorbed overlayers ends on an uncertain note. The logical introduction of a long-range triple-dipole correction for non-additive effects seems to worsen the correlation between theory and experiment.We are thus given both good news and bad news about our progress in evolving satisfactory potentials. It will be surprising indeed if the role of many-body interactions does not become a strong element in the scientific debate to follow. MOBILITY OF ADSORBED LAYERS AND WETTING OF SOLID SURFACES Let us now shift our attention to other aspects of physisorption which give us bits of information to be used for refining interaction potentials. The pioneering studies of Thorny, Duval and their collaborators" delineated phase diagrams for sub-monolayer physisorbed films on graphite and stimulated theoretical efforts to characterise the two-dimensional liquid-vapour critical point.One such calculation is described in a contribution by Klein and Cole to this Discussion,'* in which potentials of the Lennard-Jones form are used with modifications such as theJ. A. MORRISON 11 introduction of a triple-dipole three-body contribution. The essential result is that the reduced liquid-vapour critical temperature for two dimensions should be close to 0.50 for some rare gases and light methane. The agreement for the rare gases is satisfactory, but TZ for adsorbed CH,, as estimated from heat-capacity experi- m e n t ~ , ~ ~ is appreciably smaller than calculated. Some recent measurements of heats of adsorption of CH, on graphite indicate', either that unexpected clustering of adsorbed molecules is taking place at low surface coverages or that the two-phase critical temperature is higher than has been estimated.Diffusion of atoms or molecules over a solid surface must obviously depend upon the corrugation of the surface, and this is either the main or the subsidiary concern of several contributions to the Discussion. The experimental approaches, e.g. quasielastic neutron scattering or diffraction of atomic or molecular beams, yield well defined results, but their analysis to provide quantitative descriptions of the interaction potentials is proving to be difficult to accomplish. Here, the repulsive part of the potential is the crucial element, and we have no general theory on which to build its description, as we have for long-range attractive interaction.This is perhaps an appropriate place to digress briefly to consider physisorbed multilayers and the attempts which are being made to estimate their stability in relation to that of the bulk solid adsorbate. The point at issue is whether adsorption on well defined surfaces of graphite or simple metals proceeds relentlessly with increasing pressure unless it is limited by geometric factors such as a finite pore volume of a substrate. In other words, are the surfaces always wetted completely by the adsorbed layers? It is now known that several systems display partial or incomplete wetting below a characteristic wetting temperature, and one of them, ethylene-graphite, has been investigated exhaustively by X-ray and neutron diff rac- tion.15 The adsorption isotherms for this system show16 layer-by-layer adsorption up to the equivalent of about three layers. Beyond that, the vapour pressure becomes that of bulk solid ethylene.Recently, similar types of adsorption isotherms have been observed17 for the CH,-graphite system. The excess surface density of the adsorbate seems to have a temperature dependence of the form ( Tw- T)-' with /3 = and the wetting tem- perature Tw = 75.5 K, which is much below the melting temperature of bulk CH4 (90.7 K). Bruch and Nil8 report calculations of the stability of trilayers of rare gases on graphite and the ( 1 11) face of Ag, and thereby illustrate the delicacy of the balance of the different interaction energies involved.Several general treatments of the phenomenon have been a t t e m ~ t e d , ' ~ - ~ ~ but the latest is again that interaction potentials need further refinement before reliable predictions of wetting behaviour can be made. STRUCTURES OF OVERLAYERS Elements of structure determination are contained in several of the contributions which deal with the sensitive and elegant experimental methods that are now available to us: atom and molecule scattering and diffraction, and neutron scattering. To them, we can add LEED and X-ray scattering. Particularly interesting experiments and calculations have been performed on overlayers of the simple diatomic molecules N2 and CO adsorbed on graphite. Hoydoo You and Fain2, have uncovered evidence for the existence of a triangular incommensurate structure for N2 under compression.Existing models for the system25 predict a pinwheel structure, and so the new results provide a severe test of the assumed N2-N2 and N2-C interaction potentials.12 SPIERS MEMORIAL LECTURE 0 0.1 0.2 0.3 0.4 0.5 P I Po Fig. 2. Adsorption/desorption of CH4 on graphite at T = 84.5 K: (1) adsorption, (2) desorp- tion from 4.8 x mol, (4) desorption from 17.4 x mol, (3) desorption from 7.4 x mol and (5) desorption from 17.8 x mol. Refined optical spectra of .heteronuclear diatomics adsorbed on oxides and halides obtained by Plater0 et u Z . ~ ~ also yield information about the structure of adlayers and of lateral interactions of the dipole-dipole type. However, for one system at least (CO on NiO), the interaction with the solid surface is rather large and may lie beyond the range commonly associated with physisorption. Both structure and dynamics enter in the studies by Gibson and Sibener27 of the scattering of helium from overlayers of various thicknesses for rare gases adsorbed on Ag( 11 1).It is interesting and satisfying that the experimental dispersion curves can be accounted for satisfactorily, as Cardini et ~ 1 . ~ ~ have shown, through lattice dynamics and computer-simulation calculations using reasonable Morse and Lennard-Jones potentials. Despite this success, it is perhaps not inappropriate to raise a question about the reversibility of multilayer adsorption on these well defined surfaces which we are discussing. We were much surprised recently to observe an odd type of hysteresis in the adsorption of CH4 on gra?hite,’4”7 and some results are illustrated in fig.2. The monolayer capacity for the particular system is ca. 2.5 x lop3 mol, and so the segment illustrated corresponds to surface coverages between 0.8 and 2.2 layers. The remarkable features are (i) that the magnitude of the hysteresis depends uponJ. A. MORRISON 13 18 16 e I d z 3 14 -2 c;: 12 IC Fig. 3. Isosteric heat of adsorption of CH4 on Grafoil MAT at T = 84.5 K for low surface coverages. the amount adsorbed before desorption, (ii) that there are no boundary curves such as are found for adsorption/desorption in porous and (iii) that the hysteresis disappears abruptly at around P / Po = 0.2. It is very unlikely that the hysteresis is caused by experimental pathology because similar effects have been observed at other temperatures and for another system (krypton-graphite). Moreover, hysteresis is detected in the dependence of qst upon the amount adsorbed for a surface coverage greater than the equivalent of a monolayer. Fig.3 shows qst for CH, on a particular type of exfoliated graphite at T = 84.5 K and, except for the 'spike', the form is similar to that displayed for Xe-graphite in fig. 1. The 'spike' marks the region where fluid and incommensurate solid surface phases coexist. The behaviour of qst at higher surface coverages is illustrated in fig. 4. We see that, after desorption from a high surface coverage, the variation in qst is the same, but shifted to higher coverages. Some additional details can be found elsewhere.It is hard to reach a conclusion from these experiments other than that physi- sorbed multilayers get irreversibly compressed during the adsorption process. Neutron diffraction studies of several years ago for the CD,-graphite system3' showed that the lattice parameter for the adsorbate decreased in the region immedi- ately above monolayer coverage. It would be valuable at this stage if such studies could be extended to much thicker overlayers. Finally, extraordinary sensitivity to structure is demonstrated in the experiments of Poelsema and Comsa3' which capitalise on the large cross-section of isolated atoms and surface imperfections for diffuse scattering of low-energy helium beams. It is especially satisfying to see that the basis of the phenomenon is being investigated theoretically, as described in the contribution by L ~ u .~ ~14 SPIERS MEMORIAL LECTURE Fig. 4. Isosteric heat of adsorption of CH4 on Grafoil MAT at T = 84.5 K for surface coverages greater than the equivalent of a monolayer: 0, initial series; 0, after desorption from n, = 15 x mol. ENERGY EXCHANGE I have left to last any mention of one of the main themes of the Discussion, viz. energy exchange at the gas-solid interface. In part, this is because anything that I might try to contribute to the subject would certainly be overwhelmed quickly by the elegant contributions of Kreuzer, Barker, Auerbach and Tully . Here, especially in the work of Barker and A ~ e r b a c h , ~ ~ we begin to see some of the microscopic detail of the dynamics of collisions between gas atoms or molecules and solid surfaces.It is again very satisfying that these theoretical developments are proceed- ing in close connection with refined experiments. Those who would like to assimilate background in the topic could hardly do better than to consult a recently published review.34 I should like to express appreciation to Dr A. Inaba and Dr M. L. Klein for sharing many happy hours in discussing gas-solid interactions. W. A. Steele, The Interaction of Gases with Solid Surfaces (Pergamon Press, Oxford, 1974). H. C. Longuet-Higgins, Discuss. Furuduy Soc., 1965, 40, 7. D. H. Everett, Discuss. Furuduy Soc., 1965, 40, 177. J. M. Kosterlitz and D. J. Thouless, J. Phys. C , 1973, 6, 1181. J.Piper and J. A. Morrison, Chem. Phys. Lett., 1984, 103, 323. S. F. O’Shea, Y . Ozaki and M. L. Klein, Chem. Phys. Lett., 1983, 94, 355. J. Talbot, D. J. Tildesley and W. A. Steele, Furuduy Discuss. Chem. Soc., 1985, 80,91. * J. Piper, J. A. Morrison, C. Peters and Y. Ozaki, J. Chem. SOC., Furuduy Trans. I , 1983,79, 2863. L. Danielson, J-C. Ruiz, C. Schwartz, G. Scoles and J. M. Hutson, Furuduy Discuss. Chem. Soc., 1985, 80, 47. For example, see A. Thorny, X. Duval and J. Reginer, Surf: Sci. Rep., 1981, 1, 1. A. D. Migone, Z . R. Li and M. H. W. Chan, Phys. Rev. Lett., 1984, 53, 810. lo H. Jonsson and J. H. Weare, Furuduy Discuss. Chem. Soc., 1985, 80, 29. *’ J. R. Klein and M. W. Cole, Furuduy Discuss. Chem. Soc., 1985, 80,71. 11 13J. A. MORRISON 15 A. Inaba, Y.Koga and J. A. Morrison, Faraday Symp. Chem. SOC., 1985, 20, in press. See, for example, M. Sutton, S. G. J. Mochrie and R. J. Birgeneau, Phys. Rev. Lett., 1983, 51, 407; S. G. J. Mochrie, M. Sutton, R. J. Birgeneau, D. E. Moncton and P. M. Horn, Phys. Rev. B, 1984, 30, 263. A. Inaba and J. A. Morrison, Chem. Phys. Lett., to be published. L. W. Bruch and X-Z. Ni, Faraday Discuss. Chem. SOC., 1985, 80, 217. 14 15 l6 J. Menaucourt, A. Thomy and X. Duval, J. Phys. (Paris), 1977, 38, C4-195. 17 18 l 9 D. E. Sullivan, Phys. Rev. B, 1979, 20, 3991. 2o R. Pandit, M. Schick and M. Wortis, Phys. Rev. B, 1982, 26, 5112. 21 R. Pandit and M. E. Fisher, Phys. Rev. Lett., 1983, 51, 1772. 22 M. P. Nightingale, W. F. Saam and M. Schick, Phys. Rev. Lett., 1983, 51, 1275; Phys. Rev. B, 23 S. Dietrich and M. Schick, Phys. Rev. B, 1985, 31, 4718. 25 C. Peters and M. L. Klein, Mol. Phys., 1985, 54, 895. 26 E. E. Platero, D. Scarano, G. Spoto and A. Zecchina, Faraday Discuss. Chem. SOC., 1985,80,183. 27 K. D. Gibson and S. J. Sibener, Faraday Discuss. Chem. Soc., 1985, 80, 203. 28 G. Cardini, S. F. O’Shea and M. L. Klein, Faraday Discuss. Chem. SOC., 1985, SQ, 227. 29 D. H. Everett, in The Gas-Solid Interface, ed. E. A. Flood (Marcel Dekker, New York, 1967), 30 P. Vora, S. K. Sinha and R. K. Crawford, Phys. Rev. Lett., 1979, 43, 704. 31 B. Poelsema and G. Comsa, Faraday Discuss. Chem. SOC., 1985,80, 247. 32 W-K. Liu, Faraday Discuss. Chem. SOC., 1985, 80, 257. 33 J. A. Barker and D. Auerbach, Faraday Discuss, Chem. SOC., 1985, 80, 277. 1984, 30, 3830. H. You and S. C. Fain, Jr., Faraday Discuss. Chem. SOC., 1985, 80, 159. 24 p. 1055. J. A. Barker and D. Auerbach, Surf: Sci. Rep., 1984, 4, 1. 3 4