Faraday Discuss., 1993,95, 55-64 X-Ray Surface Diffraction Studies of the Restructuring and Electrodeposition of Pb Monolayers on Au( 100) Single Crystals K. M. Robinson and W. E. O'Grady Code 6170 Chemistry Division, Naval Research Laboratory, Washington, DC 20375, USA High-intensity synchrotron X-ray radiation has increased the feasibility of studying the electrochemical solid/liquid interface in situ. Recent work on the interfacial structure of Au single-crystal electrodes has shown the need to merge appropriate electrochemical cell designs to work on standard four-cycle diffractometers. In this report, an electrochemical-based methodology for use of in situ X-ray diffraction from the electrode/electrolyte interface is applied to electrodeposition of Pb on Au(l00).The cell in this experiment, unlike traditional reflection cells, provides the necessary electrochemical control of the electrode surface, as determined by the cyclic voltammagram, in conjunction with the X-ray studies. Growth of the Pb monolayer begins with diffusion of Pb+' ions to the surface at potentials below the Au(00)-(5 x 20) P.Z.C.Electrodeposition causes a lifting of the (5 x 20) reconstruction. Pb atoms are deposited on the (1 x 1) surface in a c(2 x 2) structure. Domain sizes are small due to constriction by Au islands formed when the (5 x 20) is lifted. Irreversible surface defects are observed at -0.32 V vs. SCE. Potentials at which Pb on Pb deposition occurs result in a roughened surface and alloy formation.The interfacial structure of Au single-crystal electrodes in contact with various electrolytes is fundamentally important to the understanding of electrochemical systems. It has been shown by many different techniques that the primary faces of Au reconstruct under potential In the presence of Pb2+ ions, there is also electrodeposition at potentials anodic (more positive) than bulk Pb dep~sition.*~'~~~ The cyclic voltammagrams, CVs, of the Pb electrodeposition are very face dependent and reflect a large amount of restructuring of the interface. The Pb deposition on Au(lO0) has been studied by LEED,*,18J9STM14-16and by electrochemical methods.Io-l2J7 The results of these studies show that Pb on Au(lO0) involves a complex system of restructuring and formation of AuPb alloys.UHV studies of Pb on Au(100) show a considerable number of structures upon desorption and lateral and normal growth of alloys at submonolayer coverages. l9 The CVs of the Pb deposition on Au(lO0) contain several transitions with narrow potential widths as well as a broad transition which is assumed to be caused by Pb blocking the adsorption sites.12 Because of these transitions, any study of this system must be capable of demonstrating appropriate potential control. In situ X-ray diffraction has the capability of a surface-sensitive probe and has good potential control. Synchrotron radiation has become a powerful tool in electrochemical surface studies. Various techniques, such as surface diffraction, EXAFS, X-ray standing waves and fluorescence, have been used to characterize the electrode/electrolyte interface.20 The focus of this work has been on in situ X-ray diffraction of single-crystal electrodes.Surface diffraction has frequently been used in UHV environments.21,22 The high intensity produced by the synchrotron source has made in situ surface X-ray diffraction possible, thus allowing the potential-dependent surface structure to be probed. 55 Pb Monolayers on Au (100) Single Crystals There have been several systems studied with in situ X-ray scattering techniques which concern fundamental electrochemical systems such as Au, Pt and Cu ~xidation~~-~~ and Au( 100) surface reconstruction,' as well as electrochemical deposition on Au and Ag( 1 1 1) fa~es.'~,~~,~* all the in situWith the exception of X-ray reflectivity studies at Cu ~xidation,~~ studies used a reflectance-geometry electrochemical cell.There are two possible diffraction geometries for studying in situ electrodes which have been labelled reflectance and transmission. The X-ray scattering advantages of the transmission geometry over the reflectance geometry have been discussed previ~wsly.~~~~ In brief, the transmission geometry, in which a drop of solution is maintained on the surface by capillary action, allows for precise potential control of the sample at the expense of a slightly longer absorption length and an increased noise in the scattering, both of which can be reduced by operating at higher X-ray energy, 9.5-1 1 keV.In addition, the drop assures a constant absorption length for both large-angle and glancing-incidence angle scattering. This geometry permits all of the electrochemical experiments (CVs and a.c. impedance) to be carried out under diffraction conditions in a manner that allows comparison with what already exists in the literature and at the same time assures that a high-purity experiment has been conducted. The reflectance geometry traps a thin layer, <20 pm, of solution between the surface and a thin polypropylene or Mylar window. The assumption in this design is that the electrolyte will remain uniform in thickness, hence it will not flow when the cell is rotated on the diffractometer. The reflection geometry enhances the scattering, by allowing a larger sample and less absorption by the solution at high incidence angles.However, at angles approaching glancing incidence, this cell suffers several major disadvantages: (1) the actual glancing experiment cannot be achieved because the X-rays are reflected from the window; (2) as the glancing angle is approached, the absorption length in the solution, which is proportional to sin-' a, becomes extremely long, leading to excessive signal loss. The inability of these reflectance cells to provide well defined potential control at the surface leaves in doubt the potential at which the reported structural transitions occur. Furthermore, it is not possible to obtain a high-quality cyclic voltammagram under conditions where diffraction is being carried out and hence there are always questions about the state of the surface and the purity of the experiment.In particular, electrodeposited systems, such as Pb on Au, the quantity of material deposited cannot be precisely determined and coverages based on ex situ CVs, should not be assumed. Surface X-Ray Diffraction Surface X-ray diffraction is a well developed technique.21.22 Details concerning the reconstruction, roughness and faceting of the surface can be determined by measuring the intensity variations along the crystal truncation rods, CTRs21%22,29,30and the surface or non- integer scattering rods. The scattering vector is defined as Q =q, +q,,where q, =c*l, and q =a*h +b*k. The vectors a*, b* and c* are the reciprocal lattice vectors for the bulk crystal and h, k, I are the Miller indices.For CTRs, h and k have integer values while the vertical index, I, is not constrained to integer values due to the termination of the bulk crystal. For an abruptly terminated, or flat surface the intensity of the rods drops as 1 /Aq:, where Aq, is the distance from the bulk Bragg reflection. If the surface termination deviates from a flat surface to a rough surface the intensity of the rod drops more sharply. Changes in the surface charge or atomic density and spacing, caused by surface reconstructions, also cause changes in the intensity of the CTRS.~,~~.~~ By modelling the sum of the interferences from successive layers of the crystalline surface, a reflectivity profile based on single-layer deviations from an ideal termination can be constructed: K.M. Robinson and W. E. O'Grady The first part of eqn. (1) contains the Fresnel transmission coefficient, T(q,),for a low angle of incidence,21 atomic form factor F(q,)and a bulk Debye-Waller factor W(=+qf(u2)).The data reported within are Lorentz, area and polarization corrected. The second part of eqn. (1) contains the crystalline layer interference terms. The layer density, pm, is relative to the (loo)-( 1 x 1) layer density. The layer fluctuation parameter, o,, acts as a layer-dependent Debye-Waller factor. The layer displacement parameter, E,, allows the individual layers to deviate from the average multilayer spacing, c. A more direct approach to the reconstructed structure can be made by measuring the non-integer diffraction rods.These non-integer rods are related solely to the surface structure and not the bulk. Variations in the intensity of the non-integer rods describe how the reconstruction is arranged with respect to the bulk structure as well as determining atomic positions within the surface lattice. This latter measurement is far more difficult to perform, the intensity of the non-integer rods is small and they frequently are lost in the noise from the thermal diffuse scattering from the bulk crystal and the scattering from the solution. Transmission Electrochemical Cell A detailed description of the transmission geometry cell is discussed el~ewhere.~ Briefly, the geometry of the cell uses the traditional thin-layer cell concept, with the counter electrode directly opposite the working electrode, with the reference electrode tip situated in between.However, instead of a thin layer of solution squeezed between the electrodes, a larger drop of solution is suspended from a funnel-shaped reservoir between the two electrodes. The larger drop provides clearance for the incident and exit X-ray beams, Fig. 1. Multiple electrolyte solutions can be introduced to the crystal surface in the cell to allow tests of varying concentrations. The cylindrical shape of the cell provides 360" rotational symmetry and a 0-90" 26 range for the specular reflection measurements. The crystals used in this cell are 2.5 mm in diameter by 5 mm long and are held in a collette which seals about the base of the crystal. The collette can be mounted directly on the four-circle diffractometer.A groove in the lower half of the cell is filled with water-saturated filter paper to provide a constant 100% humidity inside the cell. This prevents the electrolyte drop from evaporating during the long exposure time needed for diffraction. In tests this mechanism kept a 2.5 mm diameter cylindrical drop stable for a period of two days even while undergoing the multiple rotations of the four-circle diffractometer. Experimental Prior studies of the electrodeposition of Pb on Au( 100) showed little more than a surface disorder in the [0, 0, 13 direction.31LEED results predict a c(2 x 2) structure for a coverage of 0.5 monolayer.8,12J8 Fig.l(b) shows the expected diffraction pattern for the c(2 x 2) pattern using the f.c.c. coordinate system and the positions at which data were collected. Because of the similarities of the atomic form factor of Au and Pb, for the c(2 x 2) structure it is expected that it is very difficult to distinguish between the surface (1 ,1) position and the (1, 1, I) CTR.32 The intensity relationship: 1 1O),, = [1 -FPb/FAu(I( 1 1 O)o,opb/I( Y~coverage)]~ developed for the c(2 x 2) at the (1, 1,0) position,32 predicts a minimal intensity variation for increasing coverage; however, comparisons of the (0,0, I), (1,1,I) and (2,0, I) CTRs and the glancing-incidence measurements of the (1, 0), (0, 1) positions of the c(2 x 2) structure allowed a determination of the Pb structure.The crystal was removed from the cell, repolished and the procedures outlined above were repeated to ensure reproducibility. The measurements were made on beamline X-23B at the NSLS at an energy of 9.5 keV, which represents a compromise between maximum beamline intensity and X-ray absorption by the electrolyte. Pb Monolayers on Au (100) Single Crystals The Au(100) crystal was cut from an Au boule, bulk mosaic <0.05" (1, 1, l), and polished down to 1 pm alumina powder. The crystals were further electropolished in a 1:1 :2 mixture of glycerol-ethanol-HCl(conc.). The crystals were then flame-annealed according to standard procedures30 and flushed with distilled water. This procedure has been shown to produce atomically smooth surfaces with a (1 x 1) surface stru~ture.~.~.~,~~ Only upon emersion at 0.4 V vs.SCE was the (5 x 20) reconstruction obser~ed.~.~~ Fresh 0.1 mol dm-3 HCIO, and 1.O mmol dm-3 Pb(CIO,), electrolyte, made from Ultrex grade HCIO,, Aldrich grade Pb(CI04),.3H,0 and 18 MSZ cm distilled water, was deoxygenated with 99.9995% pure N, and pumped into the reservoir.A small drop, 2.5 mm diameter, was brought into contact with the electrode surface with the potential held at 0.4 V vs. SCE. An Au/AuO, reference electrode was used in the measurements; however, data are reported vs. SCE. The CV from the outlined preparation technique is shown in Fig. 2. The letters indicate potential regions at which the X-ray diffraction data were measured. The result of holding the potential in a particular region will be discussed further.The CV is identical to those reported in the literaturesJ2 for Pb on Au( 100). The CV was recorded on the first cycle and is not a steady-state CV. Multiple cycles were avoided as expected alloy formation had been observed by STM.14*15 The CV clearly supports the high purity achieved in the preparation procedure and that the cell provides excellent potential control with no IR drop. In situ Diffraction Results and Discussion The integrated intensity for the (0, 0, I), (1, 1, I) and (2,0, I) CTRs for the three potential regions of Fig. 2 are shown in Fig. 3-5. There are considerable differences with the fits (a) ~ N2in,et ,,-reference electrode I-counter electrode electrolyte inlet reservoir water filter paper working electrode I X (1,2)s (220) Fig.1 (a) Transmission geometry cell for in situ X-ray diffraction. The entire cell is further encompassed in an He-filled bag. (b)The c(2 x 2) diffraction pattern. Arrows indicate positions examined by X-ray diffraction. K. M. Robinson and W. E. O’Grady 2 1 $0 -1 -2 2 1 3r=-0 -1 -2 -0.4 -0.2 0.0 0.2 0.4 EIV us. SCE Fig. 2 Au(100) cyclic voltammagram recorded in the transmission cell, (a) first CV 0.1 mol dm-3 HCI04-l mmol dm-3 Pb(CIO,), 25 mV s-I, (b)same CV except after extended period of time at potentials negative of the irreversible peak at -0.32 V vs.SCE 1 3.0 i,I3.5 2.5 -4.A I t 0.0 0.5 1.0 1.5 2.0 2.5 3.0 (090, I) Fig. 3 (0, 0, I) CTR and fits of eqn. (1) for the regions A( 0,-), B( A,-* -) and C( m, - - -) of Fig. 2 Pb Monolayers on Au (100) Single Crystals oooooo I 100 000 1000 10 000 100 10 0.0 0.2 0.4 0.6 0.8 1.0 (17 17 1) Fig. 4 (1, 1, I) CTR for the regions B (A)and C (M) of Fig. 2 10 000 1000 AU._v) gU ._ 100 inI-0.0 0.1 0.2 0.3 0.4 0.5 0.6 (2, 0, I) Fig. 5 (2, 0, l) CTR for the regions B (A)and C (U)of Fig. 2 and a calculation (-) for the ideally terminated surface in region A associated with each of the CTRs. Table 1 contains the results from a fit of eqn. (1) to the (0, 0, I) data. The (I, 1, I) and (2,0, I) CTRs are discussed qualitatively.Each potential region will be discussed separately. + 0.440 V vs. SCE The (0, 0, 1) profile was best fit with a model which depicts the (5 x 20) reconstruction and adsorption of ca. 60% (1 x 1) coverage of the Pb2+ ions. There is no significant current measured in the CV to suggest that the Pb has undergone electrodeposition. In addition, the Pb layer has a large fluctuation term which suggests that the Pb is not strongly adsorbed, which would not be possible if the Pb had been electrodeposited on the Au surface. This supports a diffusion of the Pb2+ ions to the surface as the potential is stepped below the K. M. Robinson and W.E. O'Grady Table 1 Fit parameters of eqn. (1) to the (0, 0, l) data in Fig.3 €2 (uIi2)2EIV vs. SCE 'Jl PI €1 'J? Pr + 0.40 to 0.0 0.2 0.60 -0.03 0.1 1.20 0.22 0.08 0.0 to -0.32 0.1 0.50 -0.10 0.1 1.0 0.05 0.08 -0.32 to -0.40 0.2 1.4 0.15 0.1 1.1 -0.60 0.08 P.Z.C.for Au(100)-(5 x 20).3.4 The (1, 1, I) and (2, 0, I) CTRs depict a smooth surface termination. 0.0 to -0.32 V VS. SCE. The (0, 0, I) profile is best fit with a (1 x 1) surface and 50% surface converge on Au and electrodeposited Pb layer. The (2, 0, I) profile also shows an increase in the surface roughness; however, the (1, 1,I) profile remains ideally smooth. As Pb is electro-deposited, the (5 x 20) reconstruction is lifted and forms Au-Pb surface mosaics on top of the Au (1 x 1) structure. This potential region is consistent with only 25% (1 x 1) total Pb electrodeposited.When the 20% additional Au atoms from the reconstruction are included, this accounts, within error, for the surface fit density of 50%. The (2,0, I) shows an increase in the surface disorder; however, at higher values of I, the CTR becomes nearly ideal in profile, consistent with the (1, 1, I). Rocking curves taken above and below the critical angle, Fig. 6, provide some insight into the surface disorder. Near the surface, the (200) planes are displaced f0.45" from the bulk (200) planes. There is no bulk roughening as observed by the (1 ,1,I) CTR. An exact Au-Pb exchange would require an Au: Pb ratio in a single layer of 25 : 1 to account for a 0.45" distortion, calculated for rpb/rAu= 1.2. The amount of Pb electrodeposited gives a surface ratio of Au: Pb = 4.There are two possible positions for the excess of Pb: (1) AuPb alloys on the surface or (2) a structure that slightly distorts the (1 x 1) surface. In a c(2 x 2) configuration, a Pb atom would have to be pressed only 0.02 A, based on hard-sphere calculations, into the (1 x 1) Au structure to cause the 0.45" distortion. There is no evidence, via alloy diffraction peaks, that there is any alloy formation at this potential. Secondly, there is a slight increase in the (1, 1, 0) surface diffraction, Fig. 6, consistent with the intensity predicted by eqn. (2). This suggests that there is a c(2 x 2) structure. The lack of any (1 ,0) or (0, 1) scattering, associated with the c(2 x 2), can be accounted for by the fact that there are Au islands, from the lifted (5 x 20), occupying 20% of the available surface.It has been shown that these Au atoms form small islands randomly situated on the (1 x 1) surface.33 These islands would prevent growth of large domains of c(2 x 2) and therefore weak in-plane diffraction. -0.32 to -0.40 V VS. SCE The structures reported for potentials above -0.32 V vs. SCE appear reversible in both the X-ray data and the CV. The surface distortion peaks, Fig. 6, are lost if the potential is swept positive before the CV peak at -0.32 V vs.SCE. If the potential is stepped below -0.32 V vs. SCE, the distortions in the (2, 0, I) profile are permanent, Fig. 6, and the CV is also changed, Fig. 2. The (0, 0, I) profile is best fit with a AuPb layer, possibly hexagonal, due to the high density from the fit, on top of a (1 x 1) surface, also with a slightly higher density, which could possibly be due to Au-Pb interdiffusion.A dense hexagonal overlayer has been observed in LEED for coverages >60°h.18,34There is no in-plane diffraction observed to allow a determination of the surface structure, A visual improvement to the fit would allow slight distortions to the third layer; however, with the present error associated with the intensity integration, there is no statistical improvement. Interestingly, the (2, 0, I) CTR Pb Monolayers on Au (100) Single Crystals 500 to) oo ++ 00 300 Q 100-._ 1 & v OO*v) d 00 070 00 0 0 00-00 0 OO Q? Ooe:0 Oe,0 0 oooo 0 O 0 0 t oooo + + 30 -2 -1 0 1 2 4 Fig.6 Rocking curves for the (a)(2,0,0.l), (0),(2,0,0.15),( +) positions at -0.3 V vs.SCE and (b) the (1, 1, 0.1) position at -0.3 V vs. SCE, (0)and at 0.4 V us. SCE, (+) above the critical angle and the entire (1, 1, I) CTR remain unchanged, which means the permanent distortions involve only the first two to three layers. Below the critical angle, the (2, 0, I) contains many structural features. More information is necessary to fit these complex fluctuations. -0.4 to -0.5 V VS. SCE At the potential at which Pb will deposit on Pb, the surface structure changes rapidly. Fig. 7 shows the (0, 0, I) profile in this region. The only fit to the data is a highly disordered surface.The (1, 1,l)and (2,0,1)profiles are so disordered that the rods are not distinguishable from the background. Extra peaks have appeared in the data which are powder rings of the AuPb2 alloy. An extended period, 15-30 min, at this potential increased the size of the alloy peaks, which remained after cycling the potential back to + 0.4V vs.SCE. At this point the crystal had to be removed from the cell, etched in an HNO,-HCl acid mixture, repolished with alumina powder and annealed to regain a smooth surface. Conclusions Three distinct stages of Pb electrodeposition on Au( 100) have been studied by in situ X-ray diffraction. Initially, Pb2+ ions diffuse to the surface as the potential is stepped below the K. M. Robinson and W.E. O’Grady I I I 1.0 1.5 2.0 2.5 3.0 (0,090 Fig. 7 (0, 0, I) CTR for region D, -0.4 to -0.5 V vs. SCE. The new peaks are powder rings for the AuPb, alloy. P.Z.C. for the Au( 100)-(5 x 20) surface. The reconstruction lifts as electrodeposition occurs and the Pb and Au atoms form a c(2 x 2) surface on an Au( loo)-( 1 x 1) surface. The Pb distorts the (1 x 1) surface. An irreversible deposition peak, at -0.32 V vs. SCE, causes a permanent distortion, and probable Au-Pb atom exchange, after which the AuPb forms a high-density surface, composed of two to three atomic layers, on a (1 x 1) surface. AuPb, alloy forms, probably from the Au-Pb exchange sites, at potentials favourable for Pb on Pb deposition which results in an extremely disordered crystal.These conclusions are in agreement with ex situ LEED studies on Au( 100)and STM observations of AuPb alloys on Au( 11 1). Further study on the effect of stripping the electrodeposited Pb, and of the structures observed by LEED,*-’*are underway. The authors acknowledge the Research Associate Program of the National Research Council. Additional support was provided by the Office of Naval Research. Beamtime was provided on the NRL X-23B beamline, designed and built by Dr. W. T. Elam and J. Kirkland. NSLS is supported by the Department of Energy, Division of Materials Sciences and Division of Chemical Sciences under contract no. DE-AC02-76H000 16. References 1 B. M. Ocko, J. Wang, A. Davenport and H. Issacs, Phys. Rev. Lett., 1990, 65, 1466.2 D. M. Kolb and J. Schneider, Electrochim. Acta, 1986,31, 929. 3 M. S. Zei, G. Lehmpfuhl and D. M. Kolb, Surf. Sci., 1989, 221, 23. 4 A. Hamelin, M. J. Scottomayor, F. Silva, S. Chang and M. J. Weaver, J. Electround. Chem.. 1990, 295, 291. 5 S. Strabc, R. R. Adz2 and A. Hamelin, J. Eleciroanal. Chem., 1988, 249, 291. 6 X. Gao, A. Hamelin and M. J. Weaver, Phvs. Rev. Lett., 1991, 61, 618. 7 X. Gao, A. Hamelin and M. J. Weaver, Phys. Rev. B, 1991,44, 10983. 8 P. Hagens, Ph.D. Dissertation, Case Western Reserve University, 1980. 9 K. M. Robinson and W. E. O’Grady, Rev. Sci. Instrum., in the press. 10 R. R. Adzic, E. Yeager and B. D. Cahan, J. Electrochem. SOC.,1974, 121, 474. 11 J. W. Schultze and D. Diclertmann, Electrochim. Acta, 1976, 22, 489.64 K. M. Robinson and W. E. O’Grady 12 K. Engelsmann, W. J. Lorenz and E. Schmidt, J. Electroanal. Chem.. 1980, 114, 1. 13 M. G. Samant, M. F. Toney, G. L. Borges, L. Blum and 0.R. Melroy, J. Phys. Chem., 1988,92,220. 14 M. P. Green, K. J. Hanson. R. Carr and I. Lindau, J. Electrochem. Soc., 1990, 137, 3493. 15 M. P. Green and K. J. Hanson, Surf Sci. Lett., 1991, 259, L743. 16 N. J. Tao, J. Pan, Y. Li, P. 1. Oden, J. A. Derose and S. M. Lindsay, Surf Sci. Left., 1992, 271, L338. 17 D. A. Koos and G. L. Richmond, J. Phys. Chem., 1992,96, 3770. 18 J. P. Biberian and G. E. Rhead, J.Phys. F., 1973. 3, 675. 19 E. Bauer, Appl. Surf. Sci., 1982, 11/12,479. 20 For an overview of in situ techniques see: Electrochemical Interfaces, ed.H. D. Abruna, VCH, Weinheim, 1991. 21 R. Feidenhans’l, Surf. Sci. Rep., 1989, 10, 105. 22 I. K. Robinson, in Handbook on Synchrotron Radiation, ed. G. S. Brown and D. E. Moncton, Elsevier, Amsterdam, 1991, p. 221. 23 K. M. Robinson and W. E. O’Grady, J. Am. Vac. SOC.,in the press. 24 M. Fleischman and B. W. Mao, J. Electroanal. Chem.. 1987, 229, 125. 25 C. A. Melendres, H. You, V. A, Moroni, Z. Nagy and W. Yun, J. Electroanal. Chem., 1991, 297, 549. 26 G. M. Bommarito, D. Acevedo and H. D. Abruna, J. Phvs. Chem., 1992,96, 3416. 27 M. F. Toney, J. G. Gordon, M. G. Samant, G. L. Borges, 0.R. Melroy, L. Kau and D. G. Wiesler, Phys. Rev. B, 1990,42, 5594. 28 M. F. Toney, J. G. Gordon, M. G. Samant, G. L. Borges, D. G. Wiesler, D. Yee and L. B. Sorensen, Langmuir, 199 1, 7, 796. 29 I. K. Robinson, Phys. Rev. B, 1986, 33, 3830. 30 K. M. Robinson, I. K. Robinson and W. E. O’Grady, Surb Sci., 1992,262, 387. 31 B. Ocko, J. Wang, A. Davenport and H. Isaacs, Extended Abstracts: The Electrochemical Society, Spring 1991 Symposium. 32 P. W. Stephens, P. J. Eng and T. Tse, Surfuce X-Ruy und Neutron Scattering, ed. H. Zabel and I. K. Robinson, Springer-Verlag, Berlin, 1992, pp. 79-82. 33 N. Batina, D. M. Kolb and R. J. Nichols, Langmuir, 1992, 8, 2572. 34 J. Perdereau, J. P. Biberian and G. E. Rhead, J. Phys. F, 1974.4, 798. Paper 3/00266G; Received 14th Januarj,, 1993