No abstract

IV and V on convergence of the iteration procedure and uniqueness of the solution for S'{k)/S 2 (k). 26 The condition (B4) is also satisfied by the exact g(r)

The efficiency of anodic oxide formation, Cox, on nGaAs in an electrolyte formulated by Hasegawa and Hartnagel (1) depends on the applied current density (2). In contrast, there is no evidence of parasitic reactions associated with the semiconductor-oxide interfacial charge transfer (2, 3), hence the faradaic efficiency, Cf = 1. The less than 100% efficiency of the growth process may be attributed to: (i) transport of reaction product(s) to bulk electrolyte prior to completion of surface coverage (1, 2), (ii) dissolution of oxide layer on contact with an electrolyte (1, 2) and (iii) transport of nonparticipating species across the growing oxide layer into the bulk electrolyte (2).This preliminary note is concerned with the effect of growth rate on the accumulation of diffusing substances in the omde layer and on the oxide structure. ExperimentalAnodic oxides were grown galvanostatically, employing standard procedures and apparatus. Potential decays on open circuit, following current interruption, were c~ispiayed on a tfewlett-Packard, Model 7001 AM X-Y recorder. Oxide layers were examined with a I-litachi HU-125 electron microscope operated at 125 kV. Specimens for the microscopic examination were prepared by ion milling from the GaAs substrate side until a small hole was observed. Areas around the hole were sutliciently thin to yield good images. Results and DiscussionE~ectrode behavior on approach to Vres~.--An approach to the electrode rest potential upon the termination of oxide layer growth to a predetermined thickness, as a function of growth rate, is illustrated in Fig. 1. Qualitatively, the difference in the electrode transient behavior between the fast (solid lines) and the slow (dashed lines) growth rates increases with an increase in the oxide layer thickness. The slow approach to Vrest and its dependence on the oxide layer thickness rules out the participation of potential affecting processes within both the electric double layer and the depletion and/or space charge layer. Consequently, the processes responsible for the observed behavior ought to be confined to the oxide layer.It was shown (2) that anodic oxidation of n-GaAs obeys a relationship j = A exp(~E/RT) [1] The empirical constant A is related to molecular diffu-* Electrochemical Society Active Member.