Furadav Discuss. Chem. Soc., 1990, 89, 31-40 In Situ Structural Studies of the Passive Film on Iron and Iron/ Chromium Alloys using X-Ray Absorption Spectroscopy Moussa Kerkar and James Robinson* Department of' Physics, Universit-v of Warwick, Coventry, CV4 7AL A. John Forty Universit-v of Stirling, Stirling, Scotland The structure of the passive film that protects iron and iron/chromium alloys from corrosion has been investigated in siru in an aqueous environment using X-ray absorption spectroscopy, with fluorescence yield detection. From an analysis of the EXAFS and XANES it has been shown that the film formed on iron consists of FeO, octahedra, linked together by sharing edges to form what are probably sheets, or chains. When the iron is alloyed with chromium, or when it is exposed to a solution containing chromate ions, the passive film becomes more disordered and the distance between neighbouring iron atoms increases slightly.More significantly, from the passivation point of view, it has been shown that the chromium is incorpor- ated into the passive layer as a phase essentially identical to Cr(OH),, i.e. CrO, octahedra linked by hydrogen bonds into an amorphous three- dimensional phase. I t is this phase that appears to give rise to the enhanced corrosion resistance that chromium imparts to iron alloys. The corrosion of iron and mild steels (rusting) is of course very familiar to us all and it is a major economic, safety and environmental problem. It is well known that by alloying the iron with various metals, in particular chromium, to form steels this corrosion can be inhibited in all but the most aggressive media.Similar inhibition can also be achieved by immersion in certain solutions containing, for example, chromate ions. This inhibition of corrosion, or passivation, is brought about by the formation of a thin 'oxide'-containing passive layer on the metal surface which prevents further chemical attack on the metal. Surprisingly, despite its obvious importance, the chemical nature and structure of this film are still not well understood and are the subject of considerable debate. Clearly an understanding of the structure of the passive film and particularly of the role played by alloying elements in modifying this structure would be very beneficial. The establishment of these properties of the passive film has been hindered by the extreme thinness of the film (typically 2 nm), and also because it is probably significantly disordered.Until very recently such studies as have been made have largely been conducted ex situ, using electron diffraction and X-ray photoelectron spectroscopy (XPS). The results of these studies have been far from conclusive. Most electron diffraction data' ' suggest that the film is mainly y-Fe,O, with possibly an inner layer of Fe,O, (duplex model). The XPS data,".' on the other hand, suggest that the film is either y-Fe,O, or y-FeOOH, the latter being observed when there were iron(ii) ions present in solution, a conclusion also reached by some of the electron diffraction studies.' A major problem with these electron diffraction and XPS studies, however, is that they require that the sample is transferred from the corrosive environment to ultrahigh vacuum ( U H V ) for investigation. It is very likely that this will lead to dehydration and recrystalli- sation of the film, particularly in the presence of an electron beam which can cause local heating, therefore there must be considerable doubt that the phases investigated 3132 Passive Films ex situ are the same as those present in situ.Clearly there is a need for in situ structural investigations if this problem is to be resolved. Probably the first technique capable of providing in situ structural information to be applied to this problem was Mossbauer spectroscopy. Recent in situ studies of the passive film formed on Fe at controlled electrode potentials,' and ex situ studies of dried films,'*' using this technique, have demonstrated that there do indeed appear to be structural changes on drying, indicating the importance of in situ investigation.The in situ Mossbauer spectroscopy study concluded that the passive film had a disordered iron oxyhydroxide-like structure, possibly consisting of chains or sheets of FeO, octahedra linked by their edges. Mossbauer spectroscopy does not, of course, provide direct structural information, and therefore the above conclusions were largely drawn from comparisons with data from known structures. Direct structural information can come only from diffraction or extended X-ray absorption fine structure (EXAFS) studies. Whilst in situ X-ray diffraction studies of thin surface films at the electrode/solution interface are possible,"." the disordered nature of the passive layer on Fe means that this approach is probably not well suited to the study of these films.EXAFS studies are, however, feasible. A particular advantage of the EXAFS technique is that it is possible to investigate the local structure about a particular atom type by analysing the fine structure above its X-ray absorption edge. Thus, for example, in a sample containing both Fe and Cr it is possible to investigate the structure about each of these atoms. In addition to the structural information that may be obtained from an analysis of the EXAFS it is often also possible to obtain useful information from the near-edge region (X AN ES).The possibility of using EXAFS to probe the structure of the passive film on iron was first demonstrated in 1982."' This first study was conducted ex situ, but there have subsequently been a few in situ investigation,"'" including two where the sample has formed part of an electrochemical cell. Some of these studies have merely demonstrated that there is a change in the X-ray absorption spectrum on passivation, without any serious attempt at obtaining structural information (probably largely due to the poor quality of the data). Indeed, on occasion this system seems to be being used as a test of experimental, particularly detection, techniques rather than as the subject of a structural study. Such analyses ' O . ' ' as have been performed indicate Fe--0 bond lengths lying between those of y-Fe,O, and y-FeOOH with a suggestion that the presence of Cr increases this bond length. On the basis of these results the authors proposed that the film was not a stoichiometric oxide, or hydroxide, and that water was incorporated into the film. The uncertainties associated with the bond lengths in these studies were, however, very large, in part due to the data quality, and in part to the presence of a contribution to the EXAFS from the unpassivated metal.The recording of the X-ray absorption spectrum of a passive film, on iron or one of its alloys, of a quality adequate for reliable analysis of the EXAFS, presents a number of problems in regard to the design of the experiment. It is clearly not possible to use conventional transmission techniques to record the X-ray absorption spectrum of a passive layer.Ex situ studies of surface films, including those of passive layers on Fe,I5 have largely relied on electron-yield detection to record the spectra, thus taking advantage of the limited sampling depth of this technique (50 nm) to achieve a degree of surface selectivity. For in situ studies this approach cannot easily be followed, and therefore fluorescent photon detection, which is well suited to the study of what are effectively dilute systems, is used. The sampling depth of this technique is, however, very much greater than the thickness of any passive layer and therefore, typically, for a bulk sample with a thin passive layer on the surface, fluorescent photons will arise from both the bulk metal and the passive layer.The absorption spectrum will thus be a mixture of contributions from the passive layer and the metal, and it is then very difficult, if not impossible, to analyse. There are two ways this problem may be overcome.M. Kerkar, J. Robinson and A. J. Forty 33 Firstly, angles of incidence less than the critical angle may be used and in this way the X-ray beam is totally externally reflected after sampling only a thin surface layer. However, even with very flat samples, the sampling depth would still be greater than the thickness of the passive layer. An additional problem for in situ work is that the glancing-incidence configuration leads to long solution pathlengths for the X-ray photons, and at the iron and chromium K-edges, in which we are interested, the X-rays are not very penetrating. The second approach, first adopted by Long et af."' and used in the work to be described here, is to use very thin layers of metal so that when they are passivated essentially all the metal is converted to the passive film, thereby minimising the contribution from the metal substrate.In this paper we describe how X-ray absorption spectroscopy (both EXAFS and XANES), at both the Fe and Cr K-edges, may be used to study, in situ, the structure of the passive film that forms on iron, and some iron/chromium alloys, prepared as thin films. These films were passivated both chemically and electrochemically and X-ray absorption data have been obtained as a function of alloy composition, electrode potential and solution composition, with a view to determining the structure of the film and identifying the role played by Cr in improving the corrosion resistance of steels.In order to study the electrochemically passivated samples a cell design that reconciles the requirements of a free path for the X-ray photons to, and from, the sample, and a good electrochemical configuration is required and this is also discussed. Experimental Thin films of iron and its alloys with chromium can be easily produced by thermal evaporation onto a variety of substrates. Glass microscope slides have been used by others,"'-" but in our experience these contain a small, but significant, iron impurity. The samples used in this study were therefore prepared by thermal evaporation, under vacuum, onto Mylar.This plastic contains no iron or chromium impurities, it scatters X-rays only weakly, and metals adhere well to it. The film thickness was continuously monitored during deposition and the typical thickness used was between 1.2 and 2.0 nm. The composition of the films was determined by subsequent EDAX analysis in a transmission electron microscope. For the electrochemical work a thin film of gold was evaporated onto the Mylar prior to depositing the Fe film. This was to ensure that good electrical contact to the film was maintained. The gold does not absorb X-rays sig- nificantly in the energy range of interest. 7 keV photons do not penetrate far in water, and therefore electrochemical cells must be designed with this in mind. We have used two types of electrochemical cell that permit the electrode potential to be maintained whilst recording X-ray spectra, and these designs are shown in fig.1. In the first [fig. l ( a ) ] the electrode (alloy on gold on Mylar) was mounted in such a way that it could be floated on the solution surface whilst it was being passivated. When passivation was complete the electrode was raised, whilst maintaining the electrode potential, against a vertical Mylar window, thus trapping a thin capillary layer between the electrode and the window. This thin solution layer was easily penetrated by the incident and fluorescent X-ray photons and minimised the amount of scattered radiation reaching the fluorescence detector. In the second design [fig. l ( h ) ] the electrode was also the window through which the X-rays entered and left.It had better electrochemical characteristics, did not require the electrode to be moved and it permitted much easier solution deoxygenation. In both cells the secondary electrode was a platinum gauze, whilst the saturated calomel reference electrode (SCE) was mounted in a separate compartment. For the electrochemical work the electrode potential was maintained with a potentiostat (Hi-Tek DT 2101) in conjunction with a function generator (Hi-Tek PPR1) and all potentials are referred to the SCE. All34 Passive Films C 0 .- c W aJ J) \ IM. Kerkar, J. Robinson and A. J. Forty 35 solutions were prepared from AnalaR reagents with distilled, deionised, water and solutions were deoxygenated with argon.In situ X-ray absorption spectra at both the Fe and Cr K-edges were recorded for Fe and Fe/Cr alloys (up to 25% Cr) passivated by immersion in 0.1 mol dm-3 sodium nitrite or 0.005 mol dmP3 potassium chromate, and samples electrochemically passivated in borate buffer (pH 8.4) and 0.1 mol dm-' sodium perchlorate. The metal films to be passivated by immersion were placed in solution immediately on removal from the vacuum evaporation chamber and left for several days prior to their spectra being recorded. Electrochemical samples were mounted in the cell and were cathodised at negative potentials, to remove any air-formed oxide, prior to anodic passivation. When the cell shown in fig. 1 ( b ) was used, then prior to anodisation it was flushed with clean electrolyte, whilst maintaining the electrode potential and purging with argon, in order to eliminate any iron(l1) ions that had been introduced into the solution by the cathodic pretreatment.All the X-ray absorption results described here were obtained on beamline 8.1 at the S.R.S. Daresbury. The monochromator on this station" was a double-crystal instru- ment with a bent first crystal and a post-monochromator Pt coated mirror to provide a high photon flux at the sample. The spectra of model compounds (Fe and Cr oxides and oxyhydroxides) were obtained in the conventional transmission mode whilst fluores- cence detection was used for the passive films. Whilst initial work was conducted with a scintillation detector the majority of the results were, however, obtained with a 13-element solid-state Ge detector (Canberra). Each element of this detector was equipped with its own amplifier and discriminator and, to prevent pile-up effects, the count rate in any single channel was limited to <lo4 counts s-'.The fluorescence yield X-ray absorption spectra of the passive films were obtained by mounting the sample vertically at an angle of ca. 45" to the incident X-ray beam and positioning the detector in the plane of the storage ring making an angle of 45" to the sample. For most of this work the storage ring was operating at 2 GeV with a current of 200 mA and good-quality spectra were obtained by summing ca. 8 scans, each of 45 min. duration. Results and Discussion Whilst evaporated thin films have been used before for EXAFS studies of passive films, there do not appear to have been any comparisons of the behaviour of this type of sample and bulk material. In particular, it is not clear whether it is justifiable to extend conclusions drawn from studies of thin films to bulk alloys.Fig. 2 shows polarisation curves obtained in borate buffer for bulk iron, an evaporated gold film and a 2 nm film of iron evaporated onto gold. I t can be seen straight away that the curve for the Fe film on Au is essentially a combination of the curves for Au and Fe. In other words, whilst some of the gold is clearly exposed to the solution, the behaviour of the iron is essentially the same as that of bulk iron. The polarisation curves for alloy films have also been compared with those of bulk material and again the behaviour is essentially identical.In addition to recording polarisation curves these film electrodes have also been studied using photocurrent spectroscopy," and again the behaviour of the films i s almost identical to that of electrodes made from bulk alloys. On the basis of these results it can therefore be concluded that the thin films appear to behave in the same way as bulk alloys, and that structural information obtained by studying these films should be equally applicable to bulk alloys. The solutions used for immersion passivation, potassium chromate and sodium nitrite, were chosen as it has been shown that films formed in these solutions appear to have structure, composition and formation kinetics similar to those formed by anodic oxidation in solutions free from iron( 1 1 ) ions.' These films were therefore studied prior to undertaking the rather more difficult electrochemical experiments. In general terms36 Passive Films /' 1., 1 1 I I I -0.4 0.0 0.4 0.8 E / V us. SCE Fig. 2. Polarisation curves in borate buffer solution (pH 8.4) for, ( a ) iron, ( b ) gold and ( c ) a 0.20 nm film of iron on gold. I I 7100 7120 7140 7160 7180 energy / e V Fig. 3. XANES spectra at the Fe K-edge for ( a ) a 0.2 nm film of Fe at 0.8 V in 0.1 mol drn.-, sodium perchlorate solution, ( h ) y-FeOOH, ( c ) a-FeOOH, ( d ) y-Fe203 and ( e ) a-Fe,O,. these immersion passivated films were indeed found to have a structure similar to that of the electrochemically formed ones, and it is therefore convenient to discuss both types of sample together.Fig. 3 shows the near-edge (XANES) spectrum, at the Fe K-edge, of an electrochemi- cally passivated Fe film, and for comparison the spectra for the model compounds. It can be seen that the spectrum for the passive film is more like that for the oxyhydroxides than the oxides (in fact detailed examination shows it to be most like that for the y-oxyhydroxide). These results therefore lend support to the idea that the passive film has a structure like that of an oxyhydroxide. More detailed structural information, however, can be obtained from an analysis of the EXAFS.M. Kerkar, J. Robinson and A. J. Forty 37 -6 t 1 ' 1 I I 1 1 0.4 0.6 0.8 wavevector ( k )/nm ~ ' 0 0.2 0.4 0.6 0.8 1.0 r/nm Fig. 4. ( a ) k3-weighted EXAFS spectrum for an Fe film at 0.8 V in 0.1 mol dm-3 sodium perchlorate and the best-fit theoretical curve.( b ) The corresponding Fourier transforms. The theoretical curves are shown dashed. X-Ray absorption spectra at the Fe K-edge were recorded for a wide range of samples, and the EXAFS oscillations were extracted in the usual way. In order to permit the analysis of the EXAFS it was first necessary to obtain accurate phase-shift data. Using the program E X C U R V ~ ~ , which implements full curved wave theory,18 the EXAFS shown by a-Fe203 was modelled using known crystallographic parameters, and ab initio calculated phase shifts were refined to optimise the fit to the experimental data. These refined phase shifts were then checked by analysis of the EXAFS for the other model compounds, whose structures were also well known, before being used to deter- mine the structures of the passive films.Fig. 4 shows, for example, the best fit between experimental and calculated data for an Fe film passivated at 0.8 V in sodium perchlorate solution. To fit the EXAFS two shells were required, six oxygen atoms at 0.201 nm and six iron atoms at 0.302 nm. These shell radii do not of themselves permit easy iden- tification of the structure as in most iron oxides and oxyhydroxides the basic building unit is an octahedron of oxygen atoms surrounding the iron centre, and therefore the radii of the first shells are very similar. Whilst the radii of the second shells range from 0.295 nm for cr-Fe20, to 0.308 nm for y-FeOOH the greatest variation occurs in the third shell. This arises from the different ways the FeO, octahedra are joined to form the extended three-dimensional structure.For the a-oxyhydroxide, and both forms of the oxide, there is a shell containing iron atoms lying between 0.33 and 0.35 nm and its presence is clearly observed in the EXAFS. For the y-oxyhydroxide no such shell exists, nor was any evidence for such a shell obtained for the passive films. This therefore lends further support to the idea that the structure of the passive film on iron is similar to that of y-FeOOH, i.e. FeO, octahedra linked together by sharing edges and not faces or corners. Fe/Cr alloy samples were studied in a similar way to the pure iron films and the results of the analysis of the Fe K-edge EXAFS for immersion passivated samples are given in table 1, as are those for 7-FeOOH.The data presented are the shell radii and the value of A, which is defined as twice the Debye-Waller factor, in all cases there were six atoms of oxygen in the first shell and six of iron in the second. It can be seen38 Passive Films Table 1. Results of the analysis of the EXAFS of immersion passivated sample sample Fe-O/nm A,/10-4 nm' Fe-Fe/nm A,/10-4 nm' y- FeOO H 0.204 0.8 0.309 1.4 Fe Fe 5%Cr Fe 10% C r Fe 15% C r Fe 25% Cr samples passivated in 0.1 mol dm- NaNO, 0.199 1.9 0.303 5.7 0.200 2.2 0.302 3.9 0.200 1.9 0.305 3.9 0.201 2.6 0.305 5.0 0.202 2.8 0.307 5.8 samples passivated in 0.005 mol dm-j K'CrO, Fe 0.200 2.2 0.302 4.8 Fe 5%Cr 0.199 1.8 0.302 4.6 Fe 10% Cr 0.199 2.1 0.302 5.1 Fe 15% Cr 0.200 2.7 0.304 4.5 Fe 25% C r 0.200 2.5 0.302 5.8 Table 2.Results of the analysis of the EXAFS for samples passivated electrochemically sample Fe-O/nm A,/ 10- nm' Fe-Fe/nm A,/ lo-" nm' Fe Fe 25% Cr Fe Fe 25% C r samples passivated in borate solution 0.202 2.0 0.305 3 .O 0.202 2.0 0.305 4.4 0.201 2.4 0.302 3.6 0.202 2.1 0.305 4.6 samples passivated in perchlorate solution that the only significant changes with composition are for samples passivated in sodium nitrite solution where the second shell, initially much shorter than that in 7-FeOOH, increases in radius with increasing chromium content, whilst the first shell radius also increases slightly. No similar trend is seen for passivation in potassium chromate (note that whilst absolute radii are subject to an error of ca.0.002 nm any trends are significant at least at the 0.001 nm level). Note also that the disorder, particularly of the second shells, as indicated by the Debye-Waller factors, is significantly greater in the passive films than in the crystalline model compound. The results of analysis of EXAFS data for Fe and alloys containing 25% Cr electro- chemically passivated in borate buffer and sodium perchlorate solutions are shown in table 2. As was pointed out earlier it has been suggested',' that the passive film formed in iron( 1 1 ) ion-free solutions may be different from that formed in the presence of these ions. The results obtained with both types of cell described earlier were identical, even with electrolyte flushing, and therefore no evidence was found to support this hypothesis.From table 2 it can be seen that the samples passivated in perchlorate solution behave in a similar way to those passivated in sodium nitrite, i.e. the radius of the second shell increases with increasing Cr content. For samples passivated in borate buffer; however, the film structure seems to be independent of the Cr content, the second shell radius being long even for the pure Fe film. This lack of sensitivity to the Cr content is probably due to the incorporation of B into the film (such incorporation is well established). Thus the effect of the introduction of either Cr- or B-containing species into the passive film is t o slightly increase the Fe-Fe distance in the second shell with an accompanying increase in the Debye-Waller factors.The films on pure Fe have also been studied asM. Kerkar, J. Robinson and A. J. Forty 39 10 5 - 4 G O 4 - 5 -1 0 1 I 1 I I I I 0.4 0.6 0.8 1 .o wavevector ( k ) / n m - ' 0.4 0.6 0.8 1 . o wavevector ( k )/nm ' Fig. 5. ( a ) The k3-weighted EXAFS spectrum at the Cr K-edge of an Fe foil passivated in 0.005 mol dmP3 potassium chromate and the best-fit theoretical curve. ( b ) The corresponding curves for Cr(OH),. The theoretical best fit curves are shown dashed. a function of the electrode potential and there appears to be no variation in structure with potential, though of course there is an increase in the film thickness with increasing passivation voltage. The overall conclusion from these results is that for the passive film on Fe the basic structure is similar to that of y-FeOOH except that the second shell radius is shorter and the overall structure is rather disordered. The contraction of this second shell radius may be due to the lack of long-range, three-dimensional order in the passive film, i.e.the octahedra are joined into chains, sheets or clusters but not into an extended three-dimensional structure. The incorporation of other species into the film e.g. Cr or B, leads to a lengthening of the second shell radius but no other detectable changes. However, it is unlikely that this dilation of the second shell alone explains the improved corrosion resistance imparted by alloying with Cr or treating with chromate-containing solution, and therefore to investigate this further the EXAFS at the Cr K-edge spectra were recorded.For all samples passivated in potassium chromate solution it was necessary to wash and dry the films prior to analysis otherwise chromate ions in solution would contribute to the X-ray absorption spectrum, otherwise the spectra were recorded in situ. The results obtained with in situ and ex situ samples were, however, essentially the same whilst the structure of the Cr-containing part of the film was also independent of whether the Cr was incorporated from the solution, or from the alloy. XANES spectra at the Cr K-edge of these passive films show clearly that the bulk of the Cr in the passive layer is present in the 3+ oxidation state, and from the similarity in the spectra in the near-edge region it appears that the passive film has a similar structure to Cr(OH), , rather than Cr20,.To obtain further information on the structure of the Cr(OH),-like phase in the passive film the EXAFS were analysed, using a procedure similar to that already described for the Fe K-edge results, and using Cr,O, to determine the phase shifts. Fig. 5 shows the best fit to the data for the passive film and for commercial Cr(OH),. I t can be seen that the spectra are very similar and that40 Passive Films the parameters used to obtain the best fits were, within experimental error, identical except for some variation in the Debye-Waller factors for the second and third shells, implying a possible variation in the disorder. It can therefore be concluded that the structure of the passive film and Cr(OH), are very similar.The material known as Cr(OH), is in fact an amorphous hydrous oxide, of variable water content, in which CrO, octahedra are linked together by hydrogen bonds. This therefore appears to be the structure of the Cr-containing part of the passive film and it is the formation of this amorphous phase that seems to be responsible for the improved corrosion resistance of Cr-containing Fe alloys. Conclusion These in siru measurements have shown that the structure of the passive film on iron may be regarded as a disordered y-FeOOH, which is consistent with that first suggested by O’Crady,‘ on the basis of Mossbauer results. The results presented here, however, represent the first direct in situ structural confirmation of this suggestion. This observa- tion must be reconciled with the conclusion drawn from ex situ electron diffraction studies that the passive film is y-Fe,O,.y-FeOOH can be transformed into y-Fe,O, but this requires elevated temperatures therefore, whilst it would not be expected to occur on drying, it is quite possible that it will occur when the sample is exposed to an electron beam and local heating occurs. The results presented here anyway show that it no longer is necessary to rely on ex situ investigations to determine the structure of thin surface films. The structure of the chromium containing phase in the passive film on Fe/Cr alloys had not previously been identified and it was generally referred to merely as an oxide, or hydrous oxide, phase. The structure of this phase is now established. We are grateful to the S.E.R.C.for the support of this work and to the director and the staff of the SRS, Daresbury, for the provision of synchrotron radiation. References 1 2 3 M. Nagayama and M . Cohen, J. Electrochem. Soc., 1962, 109, 781. C. L. Foley, J. Kruger a n d C. J. Bechtoldt, J. Electrochem. Soc., 1967, 114, 994. K. Kuroda, B. D. Cahan, Gh. Nazri, E. Yeager a n d T. E. Mitchell, J. Electrochem. Soc., 1982, 129, 2163. 4 H. Konno and M. Nagayama, in The C’orrosion and 0.uidation of Metals, ed. U. R. Evans (Arnold, London, 1968), p. 585. 5 M. E. Brett, K. M . Parkin a n d M. J. Graham, J. Electrochem. Soc., 1986, 133, 2031. 6 W. F. O’Grady, J. Electrochem. Soc., 1980, 127, 555. 7 J . Eldridge, M. E. Kordesch a n d R. W. Hoffman, J. Vac. Sci. Technol., 1982, 20, 934. 8 M. Fleischmann, A. Oliver a n d J. Robinson, Electrochim. Acta., 1986, 31, 899. 9 M. G. Samant, M. F. Toney, G. L. Borges, L. Blum a n d 0. R. Melroy, Sue$ Sci., 1988, 193, I29 10 G. G. Long, J. Kruger, D. R. Black and M. Kuriyama, J. Electrochem. Soc., 1982, 129, 240. 1 1 J. Kruger, G . G. Long, M. Kuriyama, I>. R. Black, E. N. Farabaugh, D. M. Sanders and A. I . Goldman, Proc. Inr. Congr. Met. Corros. NRCC: Ottaua, Ont., 1984, 419. 12 G . G. Long, J. Kruger, D. R. Black and M. Kuriyama, J. Electroanal. Chem., 1983. 150, 603. 13 M. E. Kordesch and R. W. Hoffman, Nucl. Inst. Meth., 1984, 222, 347. 14 J. M . Fine, J. J . Rusek, J. Eldridge, M. E. Kordesch, J. A. Mann, R. W. Hoffman and I>. R. Sandstrom, J. Vac. Sci. Technol., 1983, Al, 1036. 15 G. G. Long, I>. A. Fischer, J. Kruger, D. R. Black, I>. K . Tanaka and G. A. Danko, Phj,.s. ReL. 8, 1989, 39, 1651. 16 M. J . Van Der Hoek, W. Werner, P. Van Zuylen, B. R. Dobson, S. S. Hasnain, J. S. Worgan and G. Luijckx, Nucl. Inst. Method., 1986, 246, 380. 17 M. Kerkar a n d J. Robinson, J. Electroanal. Cheni., to be published. 18 S. J. Gurman, N . Binstead a n d I . Ross, J. Phj.s. <-, 1YX4, 17, 143. Puper 9/05380H; Recvired 14th Decrmher; 1989