No abstract

The physical and chemical composition of p-InP surfaces prepared with HC1 and Br2/NH3(aq) etches have been investigated using SEM and XPS, and the interfacial energetics have been studied using capacitance and open-circuit photovoltage measurements. The Br2/NH3(aq) etch produced a microscopically smooth surface with a surface layer containing less than a monolayer of impurities, composed primarily of InPO4 with some In(OH)3 and some adsorbed hydroxyl species. The HC1 etch produced a rough, crystalline surface with a one to two monolayer indium-rich surface layer which was heavily hydrated and contained In20~ or In(OH)3 and adsorbed hydroxyl species. The HC1 etch also resulted in greater surface hydrocarbon contamination than the Br2/NH3(aq) etch. Capacitance measurements of the p-InP/acetonitrile interface reveal that the interfacial energetics of Br2/NH3(aq) etched electrodes are controlled by filling of empty surface states by solution redox couples, while those of the HCl-etched electrodes are controlled by electrochemical growth of a surface film. The BrJNH3(aq) etch yields a p-InP surface which allows efficient electron transfer to the solution resulting in a deep depletion energetic condition under extreme reverse bias conditions, while HCl-etched electrodes enter an inversion region where excess electrons accumulate in the space charge region. Empty surface states in the bandgap of Br2/NH3(aq)etched electrodes were found to be evenly distributed at a density of ca. 1.7 • 1012 cm -2 V -1. The open-circuit photovoltages of Br2/NH3(aq)-etched electrodes approached the bulk recombination limited value of 800 mV for redox couples with E(0') near -1.0V, but for the most negative and the most positive redox couples, deviations in the ~xpected behavior were observed. These are explained by bandedge movement which alters the barrier height. These studies reveal that the behavior of p-InP photoelectrochemical cells is determined largely by the method of surface pretreatment and only in part by bulk properties of the semiconductor material. Even when the surface films are only a few monolayers thick, their effect on the surface energetics are dramatic.

No abstract

One of the most extensively studied small bandgap p-type semiconductors is p-InP. Nevertheless, uncertainty remains about the interfacial energetics of this material under reverse bias conditions. Part of the problem is that several etching procedures have been reported, yet none have been shown to yield ideal behavior.Several explanations for the photoelectrochemical behavior of p-InP have been offered. Constant open circuit photovoltages (Vor observed at p-InP/CH3CN for redox couples with a wide range of formal reduction potentials (E o') has been attributed to Fermi-level pinning. 1 FLP can be the result of surface state charging 2 or of carder inversion, 3 which result in band edge movement (BEM). Studies of InP/metal junctions 4 generally support the view that p-InP undergoes FLP; however, Heller et al. reported that p-InP is not pinned in aqueous solutions 5. Measurement of the space charge capacitance (Csr vs. applied potential (E) offers information about the interfacial energetics for this system. Mott-Schottky (M-S) plots (1/Csr 2 vs. E ) can be used to show that p-InP/CH3CN exhibits either BEM or deep depletion behavior, depending upon the surface preparation used, the light intensity, and the presence of redox couples in solution (vide infra). Herein, we report results for p-InP electrodes etched by two different procedures leading to markedly different capacitance and photoelectrochemical behavior. EXPERIMENTAL Reagents and Materials.-UV-grade acetonitrile (BurdickJackson) was distilled over P205 and over Call 2 under a nitrogen atmosphere.Tetrabutylammonium fluoroborate (Southwestern Analytical Chemicals) which had been doubly recrystallized from ethanol and dried under vacuum for two days was used as the supporting electrolyte (0.1 M). Decamethylferrocene, DFER (Strem) was purified by sublimation. DFER has an E o' of-0.50 V vs. the E o' of ferrocene.Capacitance Measurements.-All experiments were performed in a helium glove box using the three electrode cell, instrumentation, and electrode construction described previously 6 . All scans were done in stirred solutions, and used a 0.15 cm 2 single crystal InP electrode (Crysta-Comm), with a doping density of 2x1017cm -3. A 0.5 mW, 543.5 nm green HeNe laser (Melles-Griot) was used for illuminated experiments.*Electrochemical Society Student Member.Intensities of 0.5 mW/cm 2 were generally used; higher intensities lead to complex Mott-Schottky behavior. All capacitance data has been analyzed assuming a series R-C circuit. Electrode Surface Preparation-Two different treatmentswere used to prepare the p-InP electrode surfaces.Etching three times for one minute in 50% HC1/methanol produced a grey microscopically rough surface. A five step procedure utilizing a chemomechanical etch in 0.05% bromine in methanol followed by 50% aqueous ammonia, which was developed by Aspnes 7 to yield the most abrupt interface as determined by ellipsometry, produced a smooth mirror finish. RESULTS and DISCUSSIONInterpretation of Mott-Schottky Plots: For an interface that disp...