296 P. J. Hillson and E. K. Rideal If the velocity of what is believed to be the rate-controlling process could be altered by some independent method an unambiguous decision as to which is in fact the rate-controlling process could be made. The photovoltaic action of light incident on polarized electrodes has been examined with this object in view, and as a result it has been possible to make such a decision and show that, with certain modifications, the catalytic and electrochemical mechanisms are those which are actually operative at polarized electrodes and to define under what conditions the former or latter process controls the reaction.
On overpotential 297 The electrode, which had been previously cleaned in 40 % nitric acid and washed with distilled water, was prepolarized for 2 or 3 hr. to reduce any surface oxide-layer and to bring it to a reproducible condition. Then the electrolyte was renewed.
298 P. J. Hillson and E. K. Rideal incident light were determined separately. The experiments were repeated on the mercury cathode using a 1000 W B.T.H. high-pressure mercury lamp. The light intensities obtained with this lamp were greater by a factor of from 5 to 40, depending on the wave-length, than those obtained previously. The photocurrent, at constant wave-length and polarizing current, were proportional to the light intensity, within the experimental error, showing that the term quantum efficiency, or the number of extra ions discharged per quantum of incident light, is of physical significance when applied to this reaction.
On overpotential 299 Owing to the rather large currents that were employed for the high current pulse (4 mA) there was a noticeable potential drop due to the resistance of the solution. This was apparent as a short vertical line in the potential-time pattern. The capacity of the double layer, and hence the area of the electrode, was inversely proportional to the slope of the potential-time pattern during the charging of the double layer. Since the cell was not adapted to taking a mercury electrode, the capacities were compared with that of a heavily amalgamated copper electrode whose area was assumed to be equal to its apparent area. This assumption was justified by the determination of the ratio real area: apparent area for an amalgamated silver electrode. The value obtained for the ratio was 0*95. The area of the mercury elec trode was assumed to be equal to its apparent area.
300 P. J. Hillson and E. K. Rideal and copper electrodes, while the mercury electrode gives a curved line of approxi mately the same slope. The results may be expressed by the equation logio y = const. + log10 i + fihv/2'<i0kT, where 1/30.
302 P. J. Hillson and E. K. Rideal the reflectivity of the metal, and hence the amount of heat absorbed by the electrode, changes rapidly with the wave-length, e.g. the reflectivity of silver falls from 70 to 10 % between 3500 and 3000 A.
304 P. J. Hillson and E. K. Rideal Mercury, and possibly other high overpotential metals, may behave differently in that the heats of adsorption of oxygen and hydrogen are so low that water is not dissociated but adsorbed as molecules. Kemball (1946) found that water vapour was adsorbed on to mercury as molecules which were probably associated on the surface. In such a case, the potential at which sensible evolution of hydrogen begins depends primarily on the displacement of water, whose chemical potential does not greatly vary with increasing polarization, by hydrogen: Hg—HaO = Hg—H + £02 + £H20 - 34 kcal. - (AtfHs0 - AtfH), where A//Hi0 and A#H are the heats of adsorption of water and hydrogen. If these heats are small and of the same order, the required potential is about —1*4 V. Mercury, then, should exhibit a high overpotential for the deposition of hydrogen.