J. Chem. SOC., Faraday Trans. I, 1984,80, 1 I 1-1 17 Electron Mean Free Paths from Quantitative X-Ray Photoelectron Spectroscopic Studies of a Modified Thionine Electrode BY W. JOHN ALBERY* AND A. ROBERT HILLMAN Department of Chemistry, Imperial College of Science and Technology, South Kensington, London SW7 2AY AND RUSSELL G. EGDELL AND HOWARD NUTTON Inorganic Chemistry Laboratory, South Parks Road, Oxford OX1 30R Received 22nd April, 1983 Thionine multilayer films deposited on Pt and Sn-doped In,O, electrodes at oxidizing potentials display electrochemical activity similar to that of free thionine; the film thickness may be measured from cyclic voltammetry of the surface-bound redox centres. Attenuation by the thionine films of substrate core photoemission in X-ray photoelectron spectroscopy is shown to conform to a Beer-Lambert law.Electron mean free paths in the films for different electron kinetic energies have been measured and found to be in reasonable agreement with theoretical predictions. This agreement may be contrasted with pathlengthsmeasured in Langmuir-Blodgett films, which are considerably larger than those predicted by theory. The surface sensitivity of X-ray photoelectron spectroscopy (X.P.S.) derives from the short pathlength (typically < 100 A) of low-energy electrons in solids; development of the technique as a method for quantitative analysis demands a knowledge of the attenuation behaviour of electrons in solids. With a continuum model for the solid, the flux of electrons with energy E, I(E), emerging from a slab of material of thickness 1 may be related to the incident flux Io(E) through a Beer-Lambert relationship of the type where A(E) is the mean free path of electrons with energy E.Several compilations of mean free pathlengths may be found in the literature,lP3 along with critical discussion of their method of determination. Particularly attractive is the overlayer technique, wherein one measures the attenuation of a substrate core photoelectron peak as a function of the thickness of a well characterized overlayer. Recent interest in this method has centred on multilayer films of medium- or long-chain carboxylic acids and their salts deposited on substrates by Langmuir-Blodgett techniques.l-' These films can be characterized by capacitance measurements4 or by ellipsometry .5 Pathlengths of ca.50 A at an electron energy of ca. 1000 eV appear typical for these films, although a pathlength closer to 100 A has been reported for 1 keV electrons in a polydiacetylene carboxylic acid film.6 These values are all substantially larger than the mean free path of ca. 10 A estimated from the empirical 'universal curve' of Seah and D e n ~ h ; ~ they are also larger than measured pathlengths in graphite.8* In the present work we use the electrochemical oxidative polymerisation of thionine1*l3 to prepare films for attenuation studies of known thickness. Thionine = exp [ - W91 (1) 111112 MODIFIED THIONINE ELECTRODES (Th) and leucothionine (L) form a two-electron redox couple as shown in reaction (1) Th L At potentials > 1.0 V [relative to the saturated calomel electrode (SCE)] thionine may be irreversibly oxidized, leading to the build up of a surface-bound polymer film on the electrode.As described below the total number of surface-bound redox centres may be found by integrating the charge from a cyclic voltammogram obtained at a scan rate of < 100 mV s-l. In the preparation of the electrodes it is found that the longer the coating time the greater is the number of redox centres attached to the electrode. Thus this system provides a method for preparing reproducible films of different and known thicknesses. We report here attenuation studies for thionine overlayers on platinum and tin-doped In,O, electrodes. These two substrates provide reasonably strong core signals in X.P.S.over a wide range of electron-kinetic energies, enabling us to explore the energy dependence of the electron mean free paths. EXPERIMENTAL All electrochemical experiments were carried out at 25 OC and all potentials are quoted relative to the SCE. Polymer films were deposited on either platinum or tin-doped In,O, by holding the electrodes at a potential of 1.1 V in 0.05 mol dm-, H,S04 containing ca. 30 pmol dm-, thionine for varying times. The thickness of the film was estimated by cyclic voltammetry of the washed coated electrode in 0.05 mol dm-3 H,SO, electrolyte. At slow scan rates the peak height varied linearly with sweep rate, showing that the whole of the layer was reduced or oxidised.ll Following correction for double-layer charging, the number of surface-bound redox centres could be obtained from the area of the cyclic voltammograms.We have previously shown12 that there is good agreement between the number of redox centres and a measurement from the ring current of a ring-disc electrode of the amount of thionine deposited on the disc electrode. To convert the number of redox centres (as measured from the charge) into an overlayer thickness we estimate from the molecular geometry that each redox centre in the film occupies a volume of 200 A. We assume that there are sulphate counterions in the 1a~er.l~ Preliminary X-ray photoelectron spectra were measured in an AEI ES200B spectrometer with a Mg anode operating at 228 W and a base pressure of 2 x lo-@ Torr.* It was established in the preliminary experiments that desorption and decomposition of the films was of minor importance: cyclic voltammograms taken after X.P.S.experiments indicated that exposure to u.h.v. for ca. 12 h and X-irradiation for up to ca. 8 h only reduces the coverage of surface redox centres by ca. 10%. Quantitative attenuation studies were conducted in an ESCALAB 5 spectrometer (V. G. Scientific) again using Mg Ka radiation at an anode power of 240 W. The total time for accumulation of spectra was typically < 4 h per sample. The spectrometer base pressure was 5 x lo-" Torr, although owing to rapid sampling throughout typical operating pressures were closer to 5 x Torr. Photoelectrons emerging within a cone 5' from the surface normal werecollected by the analyser input lens; we assume strictly normal photoemission in the derivation of pathlengths. Following stripping of satellite structure and the background of inelastically scattered electron^,'^ areas were estimated by numerical integration of narrow (25 eV) scans over the peaks of interest.* 1 Torr = 101 325/760 Pa.W. J. ALBERY, A. R. HILLMAN, R. G. EGDELL AND H. NUTTON 113 RESULTS AND DISCUSSION Typical wide-scan photoelectron spectra for thionine films of various thicknesses on a platinum electrode are shown in fig. 1 along with the cyclic voltammograms obtained in 0.05 mol dm-3 H,SO, background electrolyte. Increasing film thickness is monitored by the increasing area of the cyclic voltammograms. Note the progressive attenuation of Pt 4fand Pt 4d signals in X.P.S., together with the growth of the S 2p I I 1 100 PA I H 100mV 0 250 500 750 binding energy/eV Fig.1. Wide-scan photoelectron spectra of Pt coated with thionine films of varying thickness: (a) 4; (b) 19 and (c) 49 A. Cyclic voltammograms of the films are shown adjacent to the spectra: film thickness is proportional to the area of the voltammogram. and N 1s signals characteristic of the thionine overlayer. Fig. 2 shows a Beer-Lambert plot for the attenuation of the Pt 4fsigna1, demonstrating that eqn (1) is obeyed over almost two orders of magnitude of variation in the electron flux I(E). These data yield L(E) = 16 A at E = 1180 eV. We can also use the N 1s and S 2p peaks, which are caused by atoms in the overlayer itself. In fig. 1 these peaks increase in intensity as the coat becomes thicker.However,114 3 - MODIFIED THIONINE ELECTRODES (0) - 1 - -2 - n s c 3 - 3 - -4 - - 5 - 1 1 I 1 0 20 LO 60 Fig. 3. Data for the intensities of S 2p and N 1s peaks plotted according to eqn (2). the strength of these signals is not linear in film thickness as self-attenuation of the photoelectron flux from atoms in the film leads to a limiting intensity Im(E). For a film thickness I the photoelectron intensity I(E) is given byW. J. ALBERY, A. R. HILLMAN, R. G . EGDELL AND H. NUTTON 115 Hence a plot of log [Ioo/(loc - I ) ] against I should be linear with slope l / A . Fig. 3 confirms that the N 1s and S 2p peaks conform to this behaviour, enabling us to derive values for the electron pathlengths at the kinetic energies characteristic of atoms in an overlayer itself.The range of electron kinetic energies is further extended by the study of thionine on Sn-doped In,O,; similar results were obtained to those above. Here the metal 4d, 3d and 3p photoelectron and MNN Auger signals span a range of kinetic energies from 1230 to 400 eV. The results are collected in table 1. In fig. 4 we plot the variation with Table 1. Kinetic energies and pathlengths 2 . 0 ’ core peak kinetic energy/eV pathlength/A 1237 1227 1180 1089 8 54 809 80 1 767 758 587 550 400 25.5 29.0 15.6 15.8 12.6 17.8 18.4 13.9 13.7 14.6 8.5 8.3 V V 2 . L 2.6 2.8 3 .O 3.2 log (EIeV) Fig. 4. Electron mean free paths in thionine films on Pt (0) and In203 (0) electrodes. Experimental data for graphite (A)8, and Langmuir-Blodgett multilayers (V) are also The empirical ‘universal curve’ of Seah and Dench is shown by the broken line assuming that rn = 3 %i in eqn (3).The full line is calculated for the model of Leckey and ~ o w o r k e r s ~ ~ * ~ ~ from eqn (4). The following parameter values are used: p = 2.70 g A = 325, Z = 112 and Eg = 2.1 eV. 5 FAR 1116 MODIFIED THIONINE ELECTRODES kinetic energy of the electron mean free path ( A ) in thionine films on both Pt and In,O, substrates together with additional experimental data points for Langmuir-Blodgett m~ltilayers~-~ and for graphite.s. We have also plotted the empirical ‘universal curve’ of Seah and D e n ~ h , ~ which is given by ( 3 ) where A(E) is in A, E in eV and m is the thickness of one monolayer. This empirical curve is constructed from the scattered data for some 300 different compounds.Our data lie within the scatter of the points used to construct the curve. A more fundamental equation based on electron scattering by bulk jellium which relates the electron pathlength to the kinetic energy has been derived by Leckey and coworkers:15y l6 In this equation all energies are in electronvolts and Ep is the plasma energy, given A(E) = 538 mE++0.13 (m3 E): A(E) = 1.8 E E f / E t . (4) Ep = 28.8 ( p Z / A ) : ( 5 ) by where p is the density, A is the molecular weight and 2 is the number of valence electrons per molecule. In eqn (4) Eis the centroid of the optical loss function, which for the present purpose can be approximated byl5t l6 E = Ep+Eg (6) where the bandgap Eg is taken as the energy of the first U.V.absorption band. A plot of eqn (4) is shown in fig. 4, and it is gratifying that our data for films of the same material, but interrogated with electrons of different energy, lie close to the theoretical line, thereby confirming the theoretical model. The similarity of the pathlengths in our films to those in graphites* is striking, and provides support for a model of the polymer layer in which the aromatic units are oriented with the plane of their rings parallel to the surface. It is also apparent from fig. (4) that the pathlengths in thionine films are very much shorter than those in Langmuir-Blodgett multilayers. This is in agreement with the hypothesis that the long pathlengths in Langmuir-Blodgett films are caused by channelling effects peculiar to systems where long-chain molecules are oriented normal to a surface and are not a general feature of organic films.In conclusion, this work has shown first that electron scattering in compact organic films obeys a simple Beer-Lambert model and secondly that electron pathlengths for thionine films are in reasonable agreement with the model of Leckey and coworkers.15~ l6 The equipment was funded by the S.E.R.C. We thank Shell Research Ltd and I.C.I. for the provision of two Research Fellowships for R. G. E. and A. R. H., respectively. This is a contribution from the Oxford Imperial Energy Group and the Wolfson Unit for Modified Electrodes. C. J. Powell, SUI$ Sci., 1974, 44, 29. C. R. Brundle, J . Vac. Sci. Technol., 1974, 11, 212. M. P. Seah and W. A. Dench, Surf: Interface Anal., 1979, 1, 2. D. T. Clark, Y. C. T. Fok and G . G. Roberts, J. Electron Spectrosc., 1981, 22, 173. S. M. Hall, J. D. Andrade, S. M. Ma and R. N. King, J . Electron Spectrosc., 1979, 17, 181. B. Hupfer, H. Schupp, H. Ringsdorf and J. D. Andrade, J . Electron Spectrosc., 1981, 23, 103. C. R. Brundle, H. Hopster and J. D. Swalen, J. Chem. Phys., 1979,70, 5190. R. G. Steinhardt, J. Hudis and M. L. Perlman, Phys. Rev. B, 1972, 5, 1016. 13 K. Jacobi and J. Holzl, Surf: Sci., 1971, 26, 54.W. J. ALBERY, A. R. HILLMAN, R. G. EGDELL AND H. NUTTON 117 lo W. D. Albery, A. W. Foulds, K. J. Hall, A. R. Hillman, R. G. Egdell and A. F. Orchard, Nature l 1 W. J. Albery, W. R. Bowen, F. S. Fisher, A. W. Foulds, K. J. Hall, A. R. Hillman, R. G. Egdell and l2 W. J. Albery, A. W. Foulds, K. J. Hall and A. R. Hillman, J. Electrochem. SOC., 1980, 127, 654. l 3 W. J. Albery, M. G. Boutelle, P. J. Colby and A. R. Hillman, J. Electroanal. Chem., 1982, 133, 135. l4 N. Beatham and A. F. Orchard, J. Electron Spectrosc., 1976, 9, 129. l5 J. Szajman and R. C. G. Leckey, J. Electron Spectrosc., 1981, 23, 83. l6 J. Szajman, J. Liesegang, J. G. Jenkin and R. C. G. Leckey, J. Electron Spectrosc., 1981, 23, 97. (London), 1979, 282, 793. A. F. Orchard, J. Electroanal. Chem., 1980, 107, 37. (PAPER 3/642) 5-2