J. MATER. CHEM., 1991, 1(2), 191-195 Depth Profiles and Electrochemical Properties of IrO, Electrocatalysts Stabilized with TiO, Achille De Battisti,*" Andrea Barbieri," Anna Giatti," Giancarlo Battaglin,b Sandro Daolio" and Angelo Boscolo Boscolettod aDipartimento di Chimica dell'Universita and Centro CNR per lo Studio della Fotochimica e Reattivita S.E.C.C.,via L. Borsari 46, 1-44 100 Ferrara, Italy bUnita CIFM, Dipartimento di Chimica Fisica, Universita di Venezia, Calle Larga S. Marta 2137, 30 123 Venezia, Italy "lstituto di Polarogra fia ed Elettrochimica Preparativa del CNR, C.so Stati Uniti 4, 35100 Padova, Italy dMONTEDIPE, via della Chimica, 5,Porto Marghera (Ve), Italy lrO,/TiO, mixed-oxide electrocatalysts of different composition have been studied by Rutherford backscattering spectrometry (RBS) and cyclic voltammetry. Obtained results indicate that enrichment of titanium oxide in the outermost part of the coatings occurs.Comparison with data for the RuOJTiO, system suggests that the enrichment is partially controlled by the concentration of the noble-metal ion in the precursor salt mixture and by the nature of the anions present in it, while it is practically independent of the nature of the noble-metal ion itself. Plots of voltammetric charge against iridium concentration in the coatings exhibit maxima for a sample with 30 mol% lr02. An interpretation of the electrochemical behaviour on the basis of the shape of composition depth profiles has been attempted. Keywords: Electrocatalyst; Depth profile; Iridium dioxide; Titanium dioxide Electrocatalysts based on complex mixtures of Group 8 metal oxides (eg.Ru02, IrO,) and metal oxides of Groups 4A and 5A, have found wide application in industrial preparative electrochemistry. 1-5 They are generally obtained by thermal decomposition of suitable precursor salt mixtures supported on corrosion-resistant metals such as Ti and some of its alloys. Group 8 oxides give good electronic conductivity and catalytic properties to the materia1,'*2*4-7 while the other components are added in order to increase its corrosion resistance under industrial cell conditions. Many different compositions of mixed-oxide electrocatalysts have been pro- posed by the patent literature. They are generally quite complex, and any rationalization in this field is a difficult task. In fact, most of the existing works concern pure active components: RuO,, Ir02, and their mixtures.The role of the stabilizing component is examined in a limited number of papers, essentially devoted to the electrochemical behav-1*13714iOur8,10.12 or bulk microstru~ture~~' of Ru02/Ti02 electrodes. In the present work the investigation is extended to Ir0,-based materials, for which only few results are avail- able.15 The surface morphology of the samples and their microstructures have been studied by scanning electron microscopy (SEM), and by wide-angle X-ray scattering (WAXS), respectively. Considering that the electrochemical charging process involves not only their surface but also the outermost part of the bulk ~olume,'~-~~ concentration depth profiling of the metal ions and oxygen in the films has been carried out.For this purpose, Rutherford backscattering spec- trometry (RBS)26,27 has been used. For comparison, data for the analogous Ru02/Ti02 system have also been reported. The results of the ex situ characterization have been correlated with the electrochemical results, obtained by cyclic voltamme- try (CV). Experimental The compositions of the investigated layers, expressed in IrO, mol% were: 20, 30, 50, 70 and 80. Layer thicknesses were in the range 200-300 nm. Each composition consisted of four samples, belonging to the same preparation batch. They were prepared by a thermal decomposition method, outlined else- here.^^,,^ Both precursor salts were dissolved in isopropyl alcohol (Fluka purissimum).The concentration of IrC13 3Hz0 was 1% wjv. The solution of the titanium salt was prepared at the same molar concentration. Solutions were mixed in convenient ratios, then 'painted' on the titanium foil (1 cm x 1 cm plates). After solvent evaporation and short age- ing, the precursor salt deposits were thermolysed at a tempera- ture of 400 "C under oxygen. Supports were polished mechanically with emery paper, then with diamond pastes of different coarseness (7-1 pm) and, finally, chemically polished. Ru02/Ti02 films (70 mol% of RuO,) were prepared, using the couples Ru(NO)(N03),/TiC14 and RuCl, 3H,0/TiC14 as precursor salts.The thermolysis temperature was the same as for the preparation of Ir0,-based materials. Nominal IrO, concentrations in the oxide coatings were evaluated on the basis of data relative to the original chemicals used in the preparation, and may consequently be affected by a certain degree of uncertainty, in the range of a few mol%. Ir, Ti and 0 concentration-depth profiles were determined by RBS using a 4He' beam with energies of 1.0 or 2.2 MeV.? Scattered particles were detected at 160" by a surface-barrier silicon detector. Depth resolution in the very-near-surface region was 15-20 nm. Depth profiles were obtained by a fitting procedure of the experimental spectra. This procedure makes use of a computer code developed in the University of Padova by A. Carnera,30 based on the principles of backscattering analy- sis,26*27which synthesizes spectra from given concentration profiles.Profiles were represented as segmented lines, since the depth resolution of RBS broadens edges, hindering a detailed study near the points of sharp change in slope. The assumed concentration profiles were varied until a satisfactory fit of the synthesized spectrum to the experimental one was obtained. The estimated uncertainties in Ir, Ru and Ti relative t 1 eVz1.602 x 1O-l9J. concentrations were ca. 1-2% in the range of maximum concentration values. The uncertainty relative to oxygen was larger, owing to the low elastic scattering cross-section, and to the overlap of the oxygen signal with that of the titanium substrate.The depth scale is expressed in atomscm-,, the natural units of RBS analysis. Conversion of this scale to that in usual length units was carried out by dividing by the molecular density of the oxide film. This was assumed to be the weighted average of the densities of the noble-metal oxide and TiO, (rutile). This procedure may introduce systematic errors in the reported depth values, which we estimate to be not larger than 10%. Much larger uncertainties affect the compositions at the interface region between oxide layer and support, owing to the impossibility of evaluating separately the effect of interdiffusion of species, variations in thickness of the layer and roughness of surfaces on which they are formed.At this stage of the research, however, detailed features of the oxide layer/metal support interface are not strictly needed, surface properties of relatively thick coating being the main target of the present work. Electrochemical experiments were performed with a Solartron 1286 electrochemical interface. Automatic data acquisition and elaboration were performed with in-house ~oftware.~' Cyclic voltammetry measurements were carried out in 1 mol dmP3 perchloric acid solutions, following widely adopted procedures. The potential range explored was 0.00-1.20 V (us. SCE) and potential sweep rates were chosen between 0.005 and 0.1OOV s-'. The anodic charges were obtained by integration of the anodic part of the voltammo- grams performed at a potential sweep rate of 0.050 V s-'.The estimated uncertainty for voltammetric charges was ca. 1Yo. Results A qualitative indication of the surface morphology of the supported layers has been obtained by SEM. Results for samples with IrO, concentrations of 30 and 70 mol% are given in Fig. 1. In both cases larger-sized particles can be seen within a structure which essentially consists of very small grains. An examination of SEM images of all the investigated samples shows the number of larger particles decreasing with IrO, content. X-Ray diffraction data indicate that only one phase, an Ir0,/Ti02 solid solution with rutile structure, exists in the investigated composition range. Therefore, the forma- tion of the phase can be reasonably assumed to take place according to the mechanism proposed for pure TI-O,.~, Accordingly, the formation of larger particles can be due to the existence of specific sites at which nucleation and further growth occur under more favourable conditions.Considering that such sites have to be based on Ir species (Ir polynuclear complexes for instance3,), it seems reasonable that the number of particles in question decreases with decreasing Ir02 content in the layers. Most of the reacting mass, on the other hand, will undergo the transformation to oxide solid solution in a, more or less, aspecific way, creating the morphologically more homogeneous part of the layer. Fig. 2 shows, as an example, the RBS spectrum of the sample containing 30 mol% of IrO, and 70 mol% of Ti02.For the sake of comparison, the spectrum synthesized from the concentration profiles shown in Fig. 3 is also shown in Fig. 2. Profiles of the film containing 70 mol% of IrO,, exemplifying the situation for greater noble-metal content, are presented in Fig. 4.In both Fig. 3 and Fig. 4 the non-uniform- ity of the distribution of metal components across the oxide layers is evident. A quite similar shape of composition depth profiles has been found also for the films with an 11-0, nominal content of 20, 50 and 80 mol%. In particular, enrichment with titanium dioxide is observed in the near-surface region. J. MATER. CHEM., 1991, VOL. 1 Fig. 1 SEM image of the surface of IrO,/TiO, mixed-oxide films.(a) 30 mol% IrO,; (b) 70 mol% IrO,. Magnification of the outer part of the photographs is four times less than in the central part n Ir I I 0.2 0.4 0-6 0.8 energy/MeV Fig. 2 Experimental RBS spectrum (histogram) (4He+, 1.0 MeV, 8= 160"),with superimposed simulation (continuous line) of a IrO,/TiO, coating (30 mol% of IrO,) prepared at 400 "C on a mirror-finished Ti plate Data obtained by secondary-ion mass spectrometry (SIMS) for a film containing 80 mol% of IrO," are in agreement with these results. This feature, found also in the case of Ru0,-based system^,^^-^^ is not restricted to the very-near- surface region of the films, but extends to the order of 10 nm below the surface itself. According to the results in Fig.3 and 4 and to those for the other film compositions studied, the coating thickness across which the compositional change is J. MATER. CHEM., 1991, VOL. 1 0.4 -Ir -/0.2 0.0 I I 0 200 400 600 th ickness/n m Fig. 3 Depth profiles of Ir, Ti and 0, for a Ti-supported IrO,/TiO, with iridium content corresponding to 30 mol% of 11-0, (from data in Fig. 2) 2" 1.4-1 \ 1600 th icknesdn rn Fig. 4 Depth profiles of Ir, Ti and 0,for a Ti-supported IrO,/TiO, coating (70 molo/o of 11-0,) observed depends on the nominal composition, being larger for the sample containing 30 mol% of Ir02. In none of the samples does oxygen stoichiometry appear to show anomalies, the expected value of 2 being found in any case within the experimental uncertainties, independent of the nominal noble- metal content of the supported layers or their depth.At this point it may be useful to discuss the possible effect of the surface morphology of the films on RBS results. The general aspects of this problem have been discussed in the literature, in the case of both supported films36 38 and bulk ~ samples.39 Dramatic effects have been observed, as expected, when the film is discontinuous, broken off into small 'islands', and part of the substrate remains partially uncovered. For bulk rough samples, on the other hand, almost negligible effects are observed when incident particles strike the sample surface in the normal direction, and the backscattered particles are detected at angles typical of RBS analysis (larger than 150').Since the oxide films studied in the present work completely cover the substrate, and only one phase is present, their spectra should correspond to a situation closer to that for bulk rough samples. Variations in thickness should influ- ence the tailing edges of the signals, thus hindering the reliable analysis of the interface between substrate and coating. In order to justify this assumption, we made many attempts to reproduce the experimental spectra by supposing the sample to consist of 'islands' of different thicknesses, covering different percentages of the area of the film. These attempts were successful only if the profiles of all the islands were very similar to each other, and very close to those reported in Fig.3-5. We can conclude that, if we restrict our analysis to the outermost part of the coatings, depth profiles are practi- 0 nominal ___, 100 300 500 thicknesshrn Fig. 5 Depth profiles of Ru, Ti and 0 for Ti-supported RuO,/TiO, coatings. Composition 70 mol% RuO,. Solid lines refer to a film prepared by thermal decomposition of a RuC1, -3H20/TiC1, salt mixture, while dotted lines refer to a film obtained from Ru(NO)(NO,), and TiCl,. In both cases the pyrolysis of salt mixtures was carried out at 400"C in an oxygen atmosphere, as for the case of IrO,/TiO, coatings cally unaffected by their surface topography. This justifies the conclusions reported above. Complementary to earlier preliminary data,33*3" RBS depth profiles relative to a Ru02/Ti02 coating are shown in Fig.5. Ruthenium concentration is 70 mol%. Solid lines indicate the dependence of Ru, Ti and 0 stoichiometry on depth for a coating obtained by thermolysis of hydrated ruthenium tri- chloride/titanium tetrachloride salt deposit. Dashed lines correspond to the mixture obtained by thermolysis of a Ru(NO)(NO~)~/T~C~~salt deposit. The profiles in Fig. 4 and 5 exhibit similar features. An enrichment with Ti species in the outermost part of the layers is observed in each case and the thickness across which the phenomenon occurs is quite similar in the two cases. The concentration of noble metal at the surface is slightly larger for the case of the layer obtained by thermal decomposition of Ru(NO)(NO~)~.The shapes of the profiles of oxygen stoichiometry in Fig. 4 and 5 differ considerably, oxygen content being larger in the case of Ru0,- based films. By means of nuclear reaction analysis, based on the'H(15N,sry)'2C reaction, maxima in hydrogen concentration us. depth profiles have been dete~ted.~' Furthermore, by thermoanalytical methods it has been shown that the elimin- ation of a chemically bound water from Group 8 metal-oxide films is slow and inc~mplete.~"~~ In agreement with these experimental results, the maxima in the oxygen concentration profiles in Fig. 5, can be tentatively attributed to slow water migration from the underlying part of the coatings during the thermal treatments, needed to bring the overall thickness to the required value. However, further experimental evidence is required to explain both the influence of the nature of the ruthenium salt on the segregation of Ti species and the high level of average oxygen stoichiometries, such as that shown in Fig.5. As mentioned previously, the electrodes characterized by different ex situ techniques were also studied in situ by CV. As described in the Experimental, the CV experiments were performed in 1 mol dmP3 HC104. The voltammograms show a couple of moderately pronounced peaks (anodic, cathodic). Comparison of their potentials with the literature data2 indicates that the peaks are due to the solid-state redox process: IrO,+H++e-+IrOOH By integration of the anodic part of the voltammograms the anodic charge, q* can be obtained, which reflects the micro- structural texture of oxide electrodes,’ and is considered a measure of their catalytic activity.’3l2 In Fig. 6, the dependence of q* on the film composition is shown.A maximum of anodic charge is observed around the IrOz bulk concentration of 30 mol%. More generally, larger charge values are associated with intermediate/low noble-metal contents. Electrodes con- taining more Ir species, exhibit lower charge-storage capacity. Similar results have been obtained for electrode materials J. MATER. CHEM., 1991, VOL. 1 formation of solid solutions. On the other hand, enrichment with one component in the outermost part of several different systems seems to be caused by the different reactivity of precursors. The different stability of the oxides formed in the bulk of the phases and at their surface is not considered to be an important factor in these cases. The thermal oxidation of some binary alloys such as Sn-Pb46 and Cu-Ni4’ results in a surface oxide layer the composition of which depends mainly on the rate of oxidation of the metal components of the alloy.The same has been observed for stainless as well as for anodic oxidation of metal ~ystems.~~.~~ As far as supported coatings are concerned, White and co-workers dernon~trated~l.~~how migration of Ti species takes place in Rh and Pt thin films deposited on flat TiO, surfaces. This causes what the authors define as ‘encapsulation’ of the noble- metal film by titanium oxyspecies.In our case, precursor salts of the elements of Groups 4 and 5 certainly exhibit a larger reactivity towards oxygen, compared with the precursor salts of Ir or Ru and this could explain, according to the above considerations, the observed based on and R~0~/Zr0~.~~’~~ segregation phenomena. As Fig. 5 shows, change in the Ru Discussion and Conclusions The results presented in this work provide details of some general physicochemical features of Ti-supported mixed-oxide coatings based on Ir02. As far as depth profiling is concerned, RBS results confirm that following the preparation method described in the Experimental, surface enrichment with Ti species occurs. Comparison of Fig. 4and 5 also indicates that, for approximately the same bulk noble-metal concentration the extent of Ti surface enrichment is comparable.Reasons for this could be tentatively sought in the surface activity of one of the two oxide components; this type of segregation has been met in oxide systems of a different nature.4s In our case, however, the situation seems to be more complex. In fact, the observed enrichment with titanium dioxide species takes place within a finite thickness, and not only at the surface, in two or three molecular layers (multilayer forma- tion). As previously observed, this enrichment occurs in the case of one-phase system^^^.^' and also for the Ru02/Ta205 system,33 for which microstructural analysis excludes the 0 0 0 0 0 1 I I 1 I I 20406090 mot% IrO, Fig.6 Dependence of the anodic charge density, obtained by inte-gration of the anodic part of cyclic voltammograms (potentialsweep rate: 50 mV s-l), on electrode composition salt from chloride to nitrosyl nitrate is sufficient to cause a less pronounced compositional anisotropy, although the final product is the same oxide mixture in both cases. As far as the electrochemical charge-storage capacity is concerned, we have to consider that the charging process, whatever the direction of the potential sweep, can be con- sidered as due to two separate contributions: (i) the double- layer charging, bound to the capacity of the electrode/solution interface, and (ii) the oxidation-state changes involving noble- metal ions.The first of these should be essentially related to the real electrode surface area, the second to the concentration of sites capable of undergoing the redox process. The former cannot be correlated easily to any simple compositional variable. Even assuming that the maxima of real surface area can be generated by segregation of new phases, which is bound, in turn, to changes in bulk composition,12 we have to bear in mind that, of the four mentioned binary mixtures, it is only in our case and in that of Ru02/Ti02 layers,’ that solid solutions are formed. The number of sites capable of undergoing redox processes, on the other hand, cannot be measured simply by the noble- metal concentration in the electrode surface region.In fact, as previously noted, electron and ion mobility are involved in the complex charging mechanism of this kind of oxide electrode. In the range of iridium concentrations explored in the present work electron conductivity is rather high and, in agreement with the conclusions drawn by other author^;^' the control of the process is due rather to the proton mobility. This, in turn, will be larger with increasing number of defects in the outermost part of the films. The observed changes of the Ir/Ti atom ratio with depth can themselves be considered among the physical causes of defects in local microstructure. Deeper compositionally perturbed regions should therefore imply larger charge-storage capacities. Along these lines the larger charge-storage capacity characteristic of low/intermedi- ate iridium contents can be tentatively allowed for, on the basis of the shape of concentration depth profiles obtained by RBS.The results presented here certainly witness to the com- plexity of a meaningful, exhaustive characterization of mixed- oxide electrocatalysts. Non-electrochemical techniques, however, although raising new questions, supply important information complementary to that available from merely electrochemical approaches. References 1 S. Trasatti and G.Lodi, in Electrodes of Conductive MetaI Oxides, Part A, ed. S. Trasatti, Elsevier, Amsterdam, 1980, p. 301. J. MATER. CHEM., 1991, VOL. 1 195 2 S. Trasatti and G. Lodi, in Electrodes of Conductive Metal Oxides, 28 R.Amadelli, Thesis, Ferrara, 1975. 3 Part B, ed. S. Trasatti, Elsevier, Amsterdam, 1981, p. 521. A. Nidola, in Electrodes of Conductive Metal Oxides, Part B, ed. 29 G. Lodi, C. Bighi and C. De Asmundis, Mater. Chem., 1976, 1, 177. 4 5 6 7 8 9 10 S. Trasatti, Elsevier, Amsterdam, 1981, p. 627. D. M. Novak, B. V. Tilak and B. W. Conway, in Modern Aspects of Electrochemistry, ed. B. E. Conway and J. O’M Bockris, Plenum Press, New York, 1982, vol. 14, p. 195. S. Trasatti and W. E. O’Grady, in Advances in Electrochemistry and Electrochemical Engineering, ed. H. Gerischer and C. W. Tobias, Wiley, New York, 1981, vol. 12, p. 177. J. B. Goodenough, in Progress in Solid State Chemistry, ed. H. Reiss, Pergamon Press, Oxford, 1971, vol.5, p. 363. L. F. Mattheis, Phys. Rev., 1976, 13, 2433. E. K. Spasskaya, Yu. B. Makarychev, A. A. Yakovleva and L. M. Yakimenko, Elektrokhimiya, 1977, 13, 327. Yu. E. Roginskaya, B. Sh. Galyamov, V. M. Lebedev, I. D. Belova and Yu. N. Venevtsev, Zh. Neorg. Khim., 1977, 22, 505. D. V. Kokoulina, T. V. Ivanova, Yu. I. Krasovitskaya, Z. I. Kudryavtseva and L. I. Krishtalik, Elektrokhimiya, 1977, 13, 151 1. 30 31 32 33 34 35 36 37 A. Camera, University of Padova, Italy, unpublished results, 1979. A. Barbieri and R. Cenacchi, University of Ferrara, unpublished results, 1988. I. D. Belova, T. V. Varlamova, B. Sh. Galyamov, Yu. E. Rogin-skaya, R. R. Shifrina, S. G. Prutchenko, G. I. Kaplan and M. A. Sevostyanov, Mater. Chem. Phys., 1988, 20, 39. A. De Battisti, R. Brina, G.Gavelli, A. Benedetti and G. Fagher- azzi, J. Electroanal. Chem., 1986, 200, 93. S. Daolio, G. Facchin, C. Pagura and A. De Battisti, Advances in Muss Spectrometry, Heyden, London, 1988, vol. 11B, p. 1734. A. De Battisti, G. Lodi, M. Cappadonia, G. Battaglin and R. Koetz, J. Electrochem. SOC., 1989, 136, 2596. S. U. Campisano, G. Ciavola, E. Costanzo, G. Foti and E. Rimini, Nucl. Znstrum. Methods, 1978, 149, 229. Y. Fujino, H. Kaneko, Y. Takeda and S. Tamaguchi, Nucl. Znstrum. Methods, 1981, 191, 177. 11 W. A. Gerrard and B. C. H. Steele, J. Appl. Electrochem., 1978, 8, 417. 38 J. A. Cairns, J. E. E. Baglin, G. J. Clark and J. F. Ziegler, J. Catal., 1983, 83, 301. 12 13 L. D. Burke and 0.J. Murphy, J. Electroanal. Chem., 1980, 112, 39. Yu. E. Roginskaya, B.Sh. Galyamov, I. D. Belova, R. R. Shifrina, V. B. Kozhevnikov and V. I. Bystrov, Elektrokhimiya, 1982, 18, 1327. 39 40 41 A. R. Knudson, Nucl. Znstrum. Methods, 1980, 168, 163. G. Della Mea, University of Trento, Italy, unpublished results, 1980. G. Lodi, A. De Battisti, A. Benedetti, G. Fagherazzi and J. Kristof, J. Electroanal. Chem., 1988, 256, 441. 14 15 16 17 18 Yu. E. Roginskaya, I. D. Belova, B. Sh. Galyamov, F. Kh. Chibirova and R. R. Shifrina, Mater. Chem. Phys., 1989, 22, 203. S. Daolio, B. Facchin, C. Pagura, A. De Battisti and G. Battaglin, Surfi Znterfbce Anal., 1990, 16, 457. S. Trasatti and G. Buzzanca, J. Electroanal. Chem., 1971, 29, 1. G. Lodi, E. Sivieri, A. De Battisti, S. Trasatti, J. Appf. Electro- chem., 1978, 8, 135. G.Lodi, G. Zucchini, A. De Battisti, E. Sivieri and S. Trasatti, 42 43 44 45 46 G. Lodi, A. De Battisti, G. Bordin, C. De Asmundis and A. Benedetti, J. Electroanal. Chem., 1990, 277, 139. G. Brunoro, A. De Battisti and F. Pulidori, Extended Abs. of the 34th Z.S.E. Meeting, Erlangen, 1983. L. D. Burke and M. McCarthy, Electrochim. Acta, 1984,29, 211. Yu. Ya. Tomashpol’ski, I. Ya. Kolotyrkin, E. N. Lubinin, J. Microsc. Spectrosc. Electron., 1985, 10, 521. R. P. Frankentahl and D. J. Siconolfi, J. Vacuum Sci. Technol., Mater. Chem., 1978, 3, 179. 1980, 17, 1315. 19 S. Hackwood, L. M. Schiavone, W. C. Sautremont-Smith and G. Beni. J. Electrochem. SOC., 1981, 128, 2569. 47 G. G. Scherer, P. Bruesch, H. Devantay, K. Mueller and S. Stucki, J. Electrochem. SOC., 1984, 131, 1336. 20 J. Rishpon and S. Gottesfeld, J. Electrochem. SOC., 1984, 131, 1960. 48 R. P. Frankenthal and D. L. Malm, J. Electrochem. SOC., 1976, 123, 186. 21 P. G. Pickup and V. I. Birss, J. Electroanal. Chem., 1988, 240, 185. 49 H. H. Strehblow, C. M. Melliar-Smith and W. M. Augustyniak, J. Electrochem. SOC., 1978, 125, 915. 22 23 P. G. Pickup and V. I. Birss, J. Electrochem. SOC., 1988, 135, 41. L. D. Burke and D. P. Whelan, J. Electroanal. Chem., 1984, 111, 121. 50 51 J. Perriere, S. Rigo and J. Siejka, J. Electrochem. Soc., 1978, 125, 1549. D. N. Belton, Y. M. Sun and J. M. White, J. Phys. Chem., 1984, 24 S. Gottesfeld and S. Srinivasan, J. Electrochem. Soc., 1978, 125, 88, 1690. 25 26 27 89. B. E. Conway and J. Mozota, Electrochim. Acta, 1983, 28, 1. Wei-Kan Chu, J. W. Mayer and M. A. Nicolet, Backscattering Spectrometry, Academic Press, Orlando, 1978. L. C. Feldman and G. W. Mayer, Fundamentals of Surface and Thin Film Analysis, North Holland, New York, 1986. 52 53 D. N. Belton, Y. M. Sun and J. M. White, J. Phys. Chem., 1984, 88, 5172. S. Hackwood, A. H. Dayem and G. Beni, Phys. Rev. B, 1982, 26. 471. Paper 0/03155K; Received 13th July, 1990