Reflectance Studies of Adsorption on a Platinum Electrode BY M. A. BARRETT AND ROGER PARSONS Dept. of Physical Chemistry The University Bristol BS8 1TS Received 1 st September 1970 A simple technique for measuring the intensity of polarized light reflected from an electrode is described. Results for the adsorption on platinum of oxygen hydrogen halides methanol formal- dehyde and formic acid are discussed. The method is particularly sensitive for oxygen and halide5 and the optical properties of the former can be determined within limits. Tronstad demonstrated that adsorbed films of fatty acids on mercury could be detected by the change in the ellipticity of light reflected from the surface. Recently ellipsometry has been used to study adsorbed layers of oxygen 2-5 on platinum elec- trodes as well as the specific adsorption of anions ; and there has been developed a method of studying adsorption on electrodes using the measurement of the intensity of reflected light.7 Although this method necessarily provides less information in a given measurement it has advantages in that the sensitivity can be greatly increased by modulation techniques,'.it is easily adapted for measurements at varying wave- lengths of light and kinetic measurements can be made without the great elaboration required for automatic ellipsometry. The object of the present work is to study the intermediates adsorbed on platinum electrodes during the oxidation of organic molecules. The technique chosen com- bines features from both of those described above. The intensity of the two components of the reflected light is measured.It is also possible to measure the phase difference between the components although this was not usually done. EXPERIMENTAL The scheme of the optical measurements is shown in fig. 1. The light source was tungsten (Osram 48 W) or a deuterium lamp with stabilized power supplies. The light was passed through an Optica plane grating monochromator and then collimated by a quartz lens. In the full arrangement a polarizer could be placed before the cell and a compensator and analyzer after it. The polarizer and analyzer were Glan-Taylor prisms (Lambrecht Crystal Optics) mounted in Bellingham and Stanley circles calibrated to 0.01'. The compensator was a Babinet-Soleil (Gaertner L.135 W). Intensities were measured with a photomultiplier (EMI type 6256B) with a stabilized power supply (Keithley type 245) and recorded on a pen- recorder (Servoscribe RE 511).The cell was a rectangular Suprasil 4.5 x 2.5 x 6 cm or occasionally a glass cell of similar dimensions. The platinum electrode was formed from a bright rolled sheet bent to a U-section to form two parallel sheets about 2 mm apart. The two sheets were slightly offset to allow entry and exit of the light beam. An even number of reflections was used so that the reflected light beam emerged parallel to the incident beam. The number of reflections was dependent on the distance between the sheets and the angle they made with the beam. At first as many as forty reflections were used but it appeared that better results were obtained with the greater intensity remaining after a much smaller number and most of the work described here was done with two reflections.The potential of the electrode was controlled by a potentiostat (Chemical Electronics type TR40/3A) using a 72 M. A . BARRETT AND R. PARSONS 73 platinum auxiliary electrode and a saturated calomel reference electrode. A manual switching unit with potentiometric circuits allowed the electrode potential to be switched from one value to another for example in the cleaning cycle and the kinetic runs. Potentials are quoted with respect to the standard hydrogen electrode. Working Collimating Electrode Calomel Soleil sator \ Photo- / - Silica Cell Electrode multiplier FIG. 1 .-Schematic diagram of the apparatus. METHOD OF OPERATION The apparatus could be used as an ellipsometer. If the analyzer was kept with its axis at 45" to the plane of incidence tan i,b can be determined from the azimuth of the polarizer and A from the retardation of the Babinet-Soleil compensator with its axis permanentIy set at vertical and horizontal positions.In this arrangement the polarizer setting is indepen- dent of the compensator setting which is a decided advantage for rapid measurements. With the components described the apparatus can be used in the u.-v. down to 200nm but the sensitivity is not particularly good in comparison with other ellipsometers. Greater sensitivity of changes in A and continuous recording can be achieved by putting one com- ponent e.g. the compensator out of balance. The intensity then changes with any change in A and the sign of this depends on the direction of the deviation from balance.Any simultaneous differences in absolute reflection intensities will be superimposed on this. 4 FIG. 2.-Recorder trace during cleaning cycle and experiment on a Pt electrode in 0.5 M H2S04 containing 1 M formic acid. The lower part of the figure gives the potential with respect to the normal hydrogen electrode while the upper part shows the relative intensity of the p component of light of wavelength 350 nm at an angle of incidence of 60". (a) 0.5 M HzSOj ; (b) the same with 1 M formic acid. 1s indicates a duration of 60 s. 74 ADSORPTION ON Pt BY REFLECTANCE Thus to interpret the changes it is necessary to separate the two effects. Two methods are available (a) to make equivalent traces with the direction of deviation reversed; (b) the compensator and analyzer are removed and the total intensity recorded which can be sub- tracted from the off-balance curve.Here the polarizer must be left in position of balance in order to give the correct weight to the two polarized components. In either case the sequence of events at the electrode surface must be duplicated accurately. In most of the work however simply the intensity of one or both of the components of the reflected light was recorded. For this the polarizer and the compensator were removed and the analyzer kept with its axis either parallel or perpendicular to the plane of incidence. Reproducibility of the electrode surface was achieved by pre-electrolysis and by carrying out a cleaning cycle of the type recommended by Gilman lo before each measurement. This is shown in fig. 2 together with the corresponding intensity Ip of the parallel component.For the first ten cycles there was an increase in the amplitude of the changes but thereafter they no longer varied with the number of cycles. The oxidizing potential was kept down to 1.5 V to avoid any possibility of oxygen evolution which would greatly affect the intensity of reflected light as the bubbles formed tend to stick to the surface. Besides preparing the electrode surface in a reproducible way this cleaning cycle provides several reference points for the measurement of intensity changes as well as a continual check on the state of the electrode. Intensity changes of 0.01 % were readily detected and the overall reproducibility was about 0.2 %. RESULTS THE REFRACTIVE INDEX OF PLATINUM Published values of the complex refractive index of platinum 4* 6* l-l vary over a wide range as shown in fig.3a and 3b. It therefore seemed worthwhile to determine this quantity which was required for further calculations in this work. This was done using the apparatus as an ellipsometer to measure tan t,9 and A. Measurements were done with the platinum electrode immersed in 0.5 M H2S04 and held at +0.3 V with respect to the hydrogen electrode. Angles of incidence from 60 to 68" were used. The scatter in the results is still disappointingly large and an analysis of the errors involved leads to the conclusion that the main source of error is the non- uniformity of the Babinet-Soleil compensator. Better results could be obtained with a mica retardation plate but this does not transmit below 300nm. Some results were also obtained for the same electrode in air; these are also shown in fig.3a and 3b. However it seems likely that these data are affected by contamination. Since adsorption on platinum at 0.3 V in sulphuric acid is negligible data obtained under these conditions are more likely to be reliable. ADSORPTION OF OXYGEN The variation of intensity of the parallel component (Ip) normal component (In) and of delta after two reflections as a function of potential is shown in fig. 4 for a platinum electrode in 0.5 M H,S04. In agreement with previous work there is a marked lowering of both intensity and A at potentials greater than 0.75-0.8OV corresponding to the adsorption of oxygen. Above 0.80V the intensity drops linearly with potential up to about 1.5 V. These changes d(IJ d(In) and d(A) consequent on the adsorption of the oxygen film were used to narrow down the possible values of the refractive index of the film.The relationships between these changes varies in a complicated way with the refrac- tive index of the adsorbed film (nfilm) so initially computations were made giving a survey of the optical changes to be expected over the entire range of likely values of nfilm for the particular angle of incidence and wavelength employed. These were n I*! I M. A . BARRETT A N D R . PARSONS I I I I X A X o Q 0 OO 0 0 x A +* . X 0 A + + + + 0 + ++ i- 0 0 + 0 a + 0 I I 0 I I I 1 300 4 0 0 500 6 0 0 0 0 0 0 A + + + + O + . + + % + A e x % + I ' 11 I T + I+ 75 3 0 0 4 0 0 5 0 0 60( A/nm (4 FIG. 3.-The complex refractive index of Pt fi = n-ik. (a) n ; (b) k as a function of wavelength 0 in 0.5 M H2S04 at 0.3 V ; 0 in air; 0 Visscher (ref.(4)) ; X Landolt-Bornstein (ref. (1 1)) for solid Pt ; + Landolt-Bornstein for deposited Pt ; A Greef (ref. (1 3)) ; I Rideout and Wemple (ref. (12)). 76 ADSORPTION ON Pt BY REFLECTANCE based on the usual equations for reflection coefficients R of filmed surfaces as given e.g. by Winterbottom l5 (for two reflections I R l4 = I ) . dR, dR and dA were computed using the range of film thickness 0-5A since variation from linearity is negligible in this small thickness range. Many addtional computations were necessary to narrow down the possible values of refractive index. For convenience the complex film index is represented in fig. 5 as a plot in the complex plane with possible values marked out as areas.I*O+*++$ h+ o + *99 I O 0 O 0 o + FIG. EIV (NHT 4.-Intensity of p (0) and n (+) components and A (A) for light reflected twice from electrode in 0.5 M H2S04. Wavelength 350 nm ; angle of incidence 69". a Pt The experimental values of d(I,) d(1,) and d(A) were based on the differences in the values of each quantity between 1.0 and 1.5 V. This eliminates the necessity of an assumption as to the film free value. The choice of 1.5 V could lead to a small error because there is evidence of an inflection in the (intensity potential) curve in this region similar to that found by Biegler and Woods l6 in the coulometric curve. The areas in the complex refractive index plane were narrowed down in the following steps (i) using the experimental values of d(IJ/d(I') together with the fact that both d(1,) and d(1,) are negative; (ii) comparison of dA with d(IJ; this was the most helpful in spite of the poor accuracy of dA ; (iii) comparison of dR at two angles of incidence (the range used was 45-60' with some data up to 68").Fig. 5 shows the resulting areas defined by this procedure at four wavelengths. These determinations are on optical grounds alone and from one end of the segment to the other results in thickness determination ranging from 2 A/V to infinity as the segment approaches the point describing the refractive index of the platinum M . A . BARRETT AND R . PARSONS 77 substrate. Towards this point the various changes become vanishingly small and the computed data are not written out to enough sigruficant places to offer any reliable guide. This is equivalent to extremely large thicknesses in any case so that segments drawn cut off at the point representing 20a/V in fig.5. The experience of the analysis shows that although the accuracy of the experimental data leaves much to be desired improvement of the accuracy would tend to narrow down the segment I 2 3 4 I 2 3 4 k FIG. 5.-Complex refractive index (4 = n-ik) for oxide layer on platinum electrode in 0.5 M H2S04. The shaded area shows values which are compatible with the available optical data. The assumed refractive index of Pt is marked with X . (a) A = 250 nm; (b) A = 290 nm ; (c) A = 350 nm ; (4 h = 400nm. rather than to indicate which portion along its length is the correct one. Resort has to be made to bulk properties to achieve this and following the same assumptions as Visscher about the density and coulometrically determined quantity of PtO or PtOz leading to 6.6A/V and 5.2&V respectively the following results for the refractive index are obtained at 250 nm 2.4- 2i ; for 290 nm 2.7 - 1.6i ; for 350 nm 3.1 - 1.4i (all numbers 1 0.3) and at 400 nm 3.0 - 1.7i (& 0.4) with the latitude in the results covering both types of oxide.These results seem consistent with the value of 3.4- 1.3ikO.5 which is the average value obtained by Visscher at 546 nm. ADSORPTION OF HYDROGEN At potentials more negative than 0.2 V there is again a decrease in intensity but by a much smaller amount than found in the oxygen region. We attribute this to the formation of adsorbed hydrogen which occurs in this region.le At lower angles of incidence there is also evidence of a small maximum in I at about 0.15 V.This maximum occurs also in the normal component I which is greater than its value on a bare surface in the whole of the hydrogen region (fig. 4). In the lower potential region where the changes involve weakly adsorbed hydrogen both d(l,) d(1,) and d(A) are negative for all wavelengths studied in the u.-v. Further- more d(Ip) decreases markedly with decrease in angle of incidence by a factor of 2 in the range 64-45" at 350 nm. 78 ADSORPTION O N Pt BY REFLECTANCE In the region of more positive potentials corresponding to the more strongly bound hydrogen it is not so easy to draw definite conclusions. The optical effects are smaller so that their magnitude and sign are much more dependent on the assump- tions made about the effect of the charge on the metal and ionic adsorption.The least objectionable procedure seems to be to extrapolate the (I,E) plot from the double-layer region. The signs for d(Z,) and d(IJ then depend on wavelength and angle of incidence. Thus the effect of strongly bound hydrogen appears to be different from that of the weakly bound hydrogen. It should be possible in a manner similar to that used for oxygen to find areas corresponding to possible values of nfilm at least for the weakly bound hydrogen. Unfortunately no area has been found that is qualitatively consistent with all the experimental data. It may be that the system cannot be represented as a single film. HALIDES The adsorption of anions on platinum was stuhed in some detail using the ellipsometric method by Ying-Chech Chiu and Genshaw.6 The results obtained for halides differ markedly from those found coulometrically * or usingradiotracers.l8 Since the adsorption of halides provides a straightforward example of optical studies it was decided to check these results by the present method. The results are shown in fig. 6. For the chloride adsorption is detectable in the whole of the double-layer region at M KCl in 0.5 M H2S04. This appears to be more consistent with the coulometric results than with the ellipsometric data. Similarly the independence of the Br- adsorption in the concentration range to molI.-l agrees with E/V (NHE) FIG. 6.-(0) Relative Ip for Pt in 0.5 M H2S04(line) points are for lo-' (61 lowering of Zp due to Br- adsorption deduced from (a) ; (c) the same as (a) but 2.5 x and M KBr added ; M KCl added ; (d) lowering of Zp due to C1- adsorption deduced from (c).Angle of incidence 60". the coulometric results although the latter indicate a constant maximum coverage from 0.2 V to 1.0 V. The present results as far as the potential dependence is concerned agree with those of Genshaw in indicating an increase in coverage with Br- as the potential is increased. However the interpretation that this necessarily indicates an increase in coverage involves the assumption that a given coverage will result in the same optical effect at different potentials without any complicating factor M. A . BARRETT AND R . PARSONS 79 such as a change in the distance of closest approach of the ions or polarization or on the state of the surface of the platinum. An attempt was made to distinguish between these two alternatives by choosing a concentration low enough M) that adjustments to surface coverage would be slow while adjustments of any of the above-mentioned effects would be extremely rapid under any conditions.Potential steps from 0.4 to 0.6 V and from 0.4 to 0.7 V produced adjustments in the reflectivity somewhat faster than those at the same coverage during adsorption from a clean surface but definitely slower than the recorder would be able to indicate. This is taken to indicate that the change is in fact a variation in coverage and the enhanced rate is most likely due to the bulk concentration extending to the surface where there would be a concentration gradient at this coverage when adsorption starts from a clean surface. For adsorption on a clean surface the results plotted as intensity against (time)% give a reasonably straight line evidence that adsorption is diffusion controlled.0.5 I.0 EIV (NHE) FIG. 7. FIG. 8. FIG. 7.-Relative 1' for Pt in 0.5 M H2S04(line) and with 1 M methanol added (X ). Two reflections angle of incidence 68" ; wavelength 400 nm. FIG. 8.-Relative 1 for potential decreasing stepwise 30 s at each potential in 0.5 M H2S04 (0) and in 0.5 M H2S04+1 M CH30H (X). Conditions the same as those for fig. 7. The nature of the reflectivity changes on adsorption of both Cl- and Br- indicate that the simple optical model employed by Genshaw is not strictly correct in assuming no optical absorption in the film. If this were the case Ip would increase with adsorption at all but small angles of incidence and this was definitely not the case either at 550 nm the wavelength used in the ellipsometric studies or at 350 nm used mainly in the present studies.A values are not affected by small amounts of absorp- tion but d(lp) is sensitive to even small values of the absorption coefficient and reverses sign at k N 0.6 under the conditions used. ORGANIC COMPOUNDS Coulometric measurements of the adsorption of the related compounds ; methanol formaldehyde and formic acid are not in good agreement. Thus e.g. Gilman and Breiter l9 reported that the coverage of methanol in a 1 M solution is constant from 0.1 to 0.65 V from anodic sweeps. In contrast Bagotskii and Vasiliev 2o find a marked dependence on potential between 0.05 to 0.4 V. For this reason and in the hope of elucidating the nature of the adsorbed species these compounds have been studied optically.80 ADSORPTION O N Pt BY REFLECTANCE METHANOL.-&. 7 shows the effect of adding methanol to 0.5 M H2S0 to make a 1 M solution. In contrast to the effect of oxygen halides and hydrogen the adsorption of methanol results in an increase of Ip at angle of incidence of 60" and above and wavelengths above 350 nm ; in the range 0.4-0.6 V this amounts to 0.3-0.4 %. At potentials at or below 0.05 V the sensitivity more than doubles because methanol displaces hydrogen. The effect of a monolayer of hydrogen is estimated to be a reduction of Zp by about 0.5 %. However according to Gilman l9 and Bagotskii,20 only about 75 % of the hydrogen is replaceable by methanol. Thus after methanol has been allowed to adsorb at say 0.4 V then on switching to 0.05 V there should be a decrease in intensity of 0.125 %.A decrease of at least this amount is usually observed so that the results are consistent with an approximately constant amount of adsorbed species over the range 0.05-0.6V and tend to support the conclusions of Breiter and Gilman. The sequence on stepwise decreasing the potential starting at 0.8 V is shown in fig. 8 together with the equivalent observations without methanol. The oxide does not appear to be entirely reduced until 0.4 V curve B and the maximum coverage does not appear to have been reached until 0.3 V. This is similar to the results of Gilman and Breiter l 9 on cathodic sweeps except that they find the coverage still increasing at 0.2V. The discrepancy may be due to the rapidity of their sweep which although slow is about 3 times the present rate.There may not have been time for equilibrium to be established. The determination of coverage below 0.6 V by Bagotskii and Vasiliev 2o is analogous to the cathodic sweep of Gilman and Breiter.19 It is claimed that below 0.6V the oxide is completely reduced very rapidly but this leaves the difference between the anodic sweep and cathodic sweep hard to explain. According to the optical results the adsorbed oxygen is not completely gone until 0.4 V in a cathodic stepped sequence and this would suggest that a reduced number of sites are available for methanol adsorption above 0.4 V. Biegler and Koch 21 found that the rapidity with which oxide is reduced is dependent on the length of time the oxidation potential has been held and adopted 15 ms oxidation time in order to obtain a film sufficiently reduced for their technique.From the slope of the intensity recording the rate of methanol adsorption can be determined if we assume the intensity change to be proportional to the amount adsorbed. Every pre-treatment cycle provides a value for the initial rate at 0.5 V. On the basis of the equilibrium coverage being one monolayer the maximum initial rates were found to lie between 0.03 and 0.1 monolayer/s for 1 M methanol and between 0.1 and 0.2 for 2 M methanol. At 0.025 V and 1 M solution the one determination gave 0.03 monolayer/s. No methanol adsorption occurs for the first few seconds at 0.05 V. Data published by Bagotskii 2o indicate a similar delay at 0.2 V the lowest potential for which data is given. The delay is much shorter at higher potentials.Still using the same anodic cleaning sequence but switching to different low potentials several adsorption rates were determined for 0.5 M methanol in 1 M HC104. The maximum rate occurring just after the initial pause but still at virtually zero coverage was considered the relevant quantity to compare with the initial rates of Biegler and Koch.21 The results are collected in a Tafel plot fig. 9 although the present results do not fall on a straight line. The initial rates reported by Biegler and Koch 21 are for 0.2 M methanol but a line corrected to 0.5 M methanol is included in the figure assuming the rate increases in proportion to the concentration. A limit was put on the useful range of potentials treated in this way by the fact that oxide reduction becomes much slower at high potentials.Biegler and Koch21 M. A . BARRETT AND R. PARSONS 81 altered their anodic cleaning procedure to avoid this problem by oxidizing for only 15 ms but such fast switching mechanism was not possible in this work. Although the result at the lowest potential seems consistent with the extrapolation from Biegler's work the slope is different leading to much higher adsorption rates around 0.1 V. His measurements were taken in 1 M H2S04 as opposed to 1 M HC104 of the present work. One other difference between the two methods is that in his work the measure of the rate is really the difference of the anodic current associated with adsorption of methanol and the cathodic current from reducing the oxide if this is not completed by that time.In contrast the change in intensity is in the same direction for both. Thus any error arising from the tail-end of oxide reduction would be in the opposite direction in the two cases. E/V (NHE) FIG. 9.-(u) Rate of adsorption of methanol (0) and formic acid (0) from 0.5 M solutions in 1 M HC104 on Pt. A data of Biegler and Koch (ref. (21)) for 0.2 M methanol in 1 M H2S04; -- recalculated to 0.5 M methanol. (b) Rate of adsorption of formaldehyde on Pt from M HCHO in 0.5 M H2S04. FORMIC AcID.-The formic acid results were identical with those for methanol to within experimental error. This applies to both the optical effect at full coverage and to adsorption rates. FORMALDEHYDE.-Formaldehyde was prepared by refluxing paraformaldehyde in a 0.5 M H2S04 solution with gentle heat for about 5 h.Results were decidedly different from the other two organics. At the same concentration (0.5 M) the In traces were impossible to interpret. In judging the intensity variation due to adsorbed methanol or formic acid two standard intensities were used where the conditions of the surface could reasonably be assumed the same as in the pure electrolyte. These were the hydrogen-covered surface immediately after the oxide was reduced and before there was time for any adsorption of the organic substance and the other was the oxide-covered surface. The results using either of these agreed fairly well. Formaldehyde at the same concentration (0.5 M) gave intensity traces quite different in character from the other two. Formaldehyde adsorption was so much faster that it was impossible to find a point representing full coverage of hydrogen before formaldehyde' had begun to replace hydrogen.Also the shape of the intensity trace during oxidation was 82 ADSORPTION ON Pt BY REFLECTANCE noticeably altered by the presence of formaldehyde. The heavy current needed to oxidize the formaldehyde meant that the potentiostat was unable to bring the potential of the electrode up to 1.5 V in less than 30 s. When this potential was reached there was then no further change in Ip contrary to the normal behaviour during the cleaning cycle. However at low concentrations (0.005 M and 0.05 M) the traces could be interpreted in the same way as for methanol and formic acid and the two possible reference intensities again agree reasonably. Under conditions where a monolayer is expected from coulometric measurements the magnitude of d(Z,) is greater for formaldehyde than for the other two at 350nm though the difference does not show up at 400nm.Adsorption rates were determined for the 0.005 M solution and are plotted in fig. 9b. Even at this concentration adsorption is decidedly more rapid than for 0.5 M methanol or formic acid. Although a few determinations were also made at 0.05 M formaldehyde already the check between the I levels for oxide and hydrogen- covered states was beginning to fail. Loucka and Weber ’’ have likewise reported much more rapid adsorption rates for formaldehyde as compared with the other two organics and also mention a maximum rate at about 0.2 V. If one accepts the intensity at the oxide-covered surface as a standard at 0.5 M the results show a much greater d(1,) than under any other condition and I continues slowly upwards for at least 7 min still showing no signs of ceasing.It seems reason- able to suppose that polymerization is occurring at the surface. The results described so far have been based on the assumption that the observed change in reflected intensity is simply proportional to the amount of adsorbed species. This assumption is only a first approximation and it is necessary to investigate how reliable it is likely to be. When adsorption occurs accompanied by hydrogen displacement a major cause of the increase in I is due to the removal of hydrogen. In the potential range 0.15-0.25 V this would be mainly strongly bound hydrogen while below this potential range it would involve both types of hydrogen with the weakly bound hydrogen giving a proportionately greater effect per atom the proportionality depending on the wavelength and angle of incidence used.Thus the relative influence on I of adsorption on the two types of sites will change with potential. In the potential region above where any adsorbed hydrogen is stable the compli- cation of the optical effect of hydrogen removal is absent. Even here a linear relationship between reflectivity and coverage is dependent on there being only one species or that those present contribute equally to Zp or that the types occur in a constant proportion over the entire range of coverage invariant with time. A further investigation of these factors was made by employing two approaches to coverage determination. This was done with both methanol and formic acid at a concentration of M to give relatively slow adsorption.Between 0.25 and 0.5 V to judge by the simple (reflectivity time) traces made at 60° there was a delay of several seconds before any change occurred followed by a slow rise in intensity fig. 10 ; curves a and 6 check that cathodic pulses do not influence the rate. By the use of cathodic square wave pulses (applied by manual switching) for 1 s or more the hydrogen laid down could be detected by its effect on the intensity. At least 1 s was necessary for the alteration in intensity to be completed and it is not certain to what extent this is due to the recorder. The drop in intensity is a measure of the hydrogen codeposited subject to the same complications of analysis as mentioned above. However regardless of the proportion of the two types of hydrogen co- deposited the definite conclusion from this result is that part of the surface has become blocked to hydrogen codeposition almost immediately on reaching the M .A . BARRETT AND R . PARSONS 83 appropriate potential and sometimes shows overshoot as Smith et aZ.23 found. Thus the species adsorbing initially must either have no optical effect or consist of several species with cancelling optical effects. Further this situation changes after the first few seconds of adsorption since the trend in Ip is upwards for longer times. This argues for the existence of more than one species. It was reasonable to expect that there would be some way of detecting the initial adsorption directly by a trace of Ip. This was in fact found at 250 nm using 45" angle of incidence.The trend was then initially downwards followed as before by a rise. The further conclusions from these observations is that there is a change with time of the nature or proportion of the species. It is possible that the two species involved are the directly adsorbed molecule and some dehydrogenated species like -COH. Oa5 A ] 1.00 A 0.99 -/QQo--o-o C - to I I I i o-o- 1 Q = ~ o - o \ o - ~ 0 c 0 a b 1-00 0 9 9 5 I.0OOb a b 0.995 Bulk methanol has a refractive index close to water and therefore it was expected that even allowing for the effect of dehydrogenation there would be little observable change in the optical parameters. In addition the intensity trace is less stable in the presence of methanol. Although little difference could be detected at 290nm between a methanol covered surface and the film free surface at 400 nm an increase in Ip was observed.As with a positive change in A this positive d(Ip) limits the possible values for the refractive index to two areas but in this case the area inferring 84 ADSORPTION ON Pt BY REFLECTANCE high optical absorption can be considered unlikely. This leads to the conclusion that the refractive index is real and slightly greater than that of the surrounding liquid. As bulk methanol has a slightly lower refractive index this must be due to dehydrogenation producing a more compact film. We are grateful to the Ministry of Defence (Navy) for generously supporting this work and for permission to publish it. We are also indebted to Bhimasena Rao for his valuable preliminary work without which the work described would not have been possible.' L. Tronstad Trans. Faruday SOC. 1935 31 1151. A. K. N. Reddy and J. O'M. Bockris El/@sometry in the Measurement of Surfaces and Thin Flms (Washington 1963) ed. E. Passaglia R. R. Stromberg and J. Kruger N.B.S. Misc. Publ. 256 1964 p. 229. A. K. N. Reddy M. A. Genshaw and J. O'M Bockris J. Chem. Phys. 1968,48 671. W. Visscher Optik. 1967 26 407. R. Greef J. Chem. Phys. 1969,51,3148. ' Ying-Chech Chiu and M. A. Genshaw J. Phys. Chem. 1969,73 3571. ' D. F. A. Koch and D. E. Scaife J. Electrochem. SOC. 1966 113 302. * J. D. E. McIntyre Electrochem. SOC. Meeting. (New York May 1969) Abstract no. 232 ; J. Electrochem. SOC. 1969,116 14OC. A. Bewick and A. M. Tuxford this Symposium. lo S . Gilman Electrochim. Acta. 1964,9 1025.Landolt-Bornstein Zuhlenwerte und Funktionen Band 11 Tie1 8. l2 V. L. Rideout and S. H. Wemple J. Opt. SOC. Amer. 1966 56,749. l 3 R. Greef private communication. l4 A. N. Frumkin Ado. Electrochem. vol. 3 ed. P. Delahay (Interscience. New York 1963). l 5 A. B. Winterbottom Kgl. Norske Videns Skrift. no l. 1955. l6 T. Biegler and R. Woods J. Electroanalyt. Chem. 1969 20 73. V. S. Bagotskii Yu. B. Vassiliev J. Weber and .1. N. Pirtskhalava J. Electroanalyt. Chem. 1970,27 31. N. A. Balashova and V. E. Kazarinov Electroanalyt. Chem. ed. A. J. Bard (Dekker New York) 1969.3 135. l9 S. Gilman and M. W. Breiter J. Electrochem. SOC. 1962 109 622. *O V. S. Bagotskii and Yu. B. Vasiliev Electrochim. Acta. 1966 11 1439. 21 T. Biegler and D. F. A. Koch J. Electrochem. Soc. 1967,114,904. 22 T. Loucka and J. Weber J. Electroanalyt. Chem. 1969,21 329. 23 R. E. Smith H. B. Urbach J. H. Harrison and N. L. Halfield J. Phys. Chern. 1967 71 1250,