Ideal Semiconductor Electrodes

It should be emphasized that the metal doi-coated semiconductor electrodes can meet all the above-mentioned requirements simultaneously and have the properties of the ideal semiconductor electrode. The key point is that, for metal dot-coated electrodes, the reaction-proceeding part is limited to the narrow regions of metal dots and the remaining major semiconductor surface is kept free from surface states. On the contrary, for normal semiconductor electrodes with homogeneous surfaces, interfacial reactions occur over the entire surface, producing reaction intermediates (surface recombination centers) all over the surface. 

Typical values of transfer coefficients a and ji thus obtained are listed in Table 4 for single crystal and polycrystalline thin-film electrodes [69] and for a HTHP diamond single crystal [77], We see for Ce3+/ 41 system (as well as for Fe(CN)63 /4 and quinone/hydroquinone systems [104]), that, on the whole, the transfer coefficients are small and their sum is less than 1. We recall that an ideal semiconductor electrode must demonstrate a rectification effect in particular, a reaction proceeding via the valence band has transfer coefficients a = 0, / =l a + / = 1 [6], Actually, the ideal behavior is rarely the case even with single crystal semiconductor materials fabricated by advanced technologies. Departure from the ideal semiconductor behavior is likely because the interfacial potential drop is located in part in the Helmholtz layer (due e.g. to a high density of surface states), or because the surface states participate in the reaction. As a result, the transfer coefficients a and ji take values intermediate between those characteristic of a semiconductor (0 or 1) and a metal ( 0.5). 

Figure 18.2.6 Schematic representation of the variation of electron-transfer rate, and transfer coefficient, a, with electrode potential for an ideal semiconductor electrode. The current is is equivalent to that defined in (18.2.9) or (18.2.10). At sufficiently extreme potentials (not shown) mass transfer would lead to a limiting current on the right side of the diagram. [Reprinted with permission from B. R. Horrocks, M. V. Mirkin, and A. J. Bard, 7. Phys. Chem., 98, 9106 (1994). Copyright 1994, American Chemical Society.]...

Herein, criteria are developed for ideal polarizable semiconductor electrode-solution interfaces. A variety of experimental studies involving metal dichalcogenide-solution interfaces are discussed within the context of these criteria. These interfaces approach ideality in most respects and are well suited for fundamental studies involving electron transfer to solution species or adsorbed dyes. 

Based on the discussion above, it seems evident that a detailed understanding of kinetic processes occurring at semiconductor electrodes requires the determination of the interfacial energetics. Electrostatic models are available that allow calculation of the spatial distributions of potential and charged species from interfacial capacitance vs. applied potential data (23.24). Like metal electrodes, these models can only be applied at ideal polarizable semiconductor-solution interfaces (25)- In accordance with the behavior of the mercury-solution interface, a set of criteria for ideal interfaces is f. The electrode surface is clean or can be readily renewed within the timescale of... 

Let us consider in more detail, using the above concepts, how a photocorrosion process occurs under the illumination of a semiconductor. Suppose that electron transitions at the interface between the semiconductor and solution do not take place in darkness in a certain potential range (the semiconductor behaves like an ideally polarizable electrode). This range is confined to the potentials of decomposition of the semiconductor and/or solution. The steady state potential of a semiconductor is usually determined in this case by chemisorption processes (e.g., of oxygen) or, which is the same in the language of the physics of semiconductor surface, by charging of slow surface states. It is these processes that determine the steady state band bending. 

Despite extensive studies, the photovoltage or the solar-to-chemical energy conversion efficiency still remains relatively low. The main reason is that it is very difficult to meet all requirements for high efficiency. For example, high catalytic activity and sufficient passivation at the electrode surface are incompatible. It was found, however, that a semiconductor electrode modified with small metal particles can meet all the requirements and thus becomes an ideal type semiconductor electrode. Cu, Ag, and Au were chosen because they were reported to work as efficient electrocatalysts for the C02 reduction. p-Si electrodes modified with these metals in C02-staurated aqueous electrolyte under illumination produce mainly methane and ethylene.178 This is similar to the metal electrodes but the metal-particle-coated electrodes work at approximately 0.5 V more positive potentials, contrary to continuous metal-coated p-Si electrodes. 

Although several single-crystal, wide-band gap semiconductors provide electrochemical and optical responses close to those expected from the ideal semiconductor-electrolyte model, most semiconducting electrodes do not behave in this manner. The principal and by far overriding deviation from the behavior described in the previous section is photodecomposition of the electrode. This occurs when the semiconductor thermodynamics are such that thermal or photogenerated valence band holes are sufficiently oxidizing to oxidize the semiconductor lattice [8,9]. In this case, kinetics routinely favor semiconductor oxidation over the oxidation of dissolved redox species. For example, irradiation of n-CdX (X = S, Se, or Te) in an aqueous electrolyte gives rise exclusively to semiconductor decomposition products as indicated by... 

Moderately doped diamond demonstrates almost ideal semiconductor behavior in inert background electrolytes (linear Mott -Schottky plots, photoelectrochemical properties (see below), etc.), which provides evidence for band edge pinning at the semiconductor surface. By comparison in redox electrolytes, a metal-like behavior is observed with the band edges unpinned at the surface. This phenomenon, although not yet fully understood, has been observed with numerous semiconductor electrodes (e.g. silicon, gallium arsenide, and others) [113], It must be associated with chemical interaction between semiconductor material and redox system, which results in a large and variable Helmholtz potential drop. 

Electron transfer rate constants of outer sphere redox reactions can be measured relatively easily at n-type semiconductor electrodes. This is because electrons are withdrawn from the surface under depletion conditions, so that their concentration is lower than in the bulk. Under ideal... 

The attainment of equilibrium between the semiconductor and solution, however, is somewhat hindered. Many semiconductor electrodes behave over a broad range of potentials as ideally polarizable electrodes. 

In addition to the rather indirect a.c. techniques discussed above, a more direct technique for the electrical characterisation of the depletion region is available even though this is, primarily a d.c. technique, it is included here since the information is frequently highly complementary. In this technique, ohmic contacts are made to the working surface of the semiconductor electrode and then shielded from the electrolyte as shown in Fig. 36 they are then used as probes to measure the surface conductivity. Ideally,... 

Figure 29. Calculated current-potential characteristics for direct (dashed lines, 0/cm ) and surface state mediated electron transfer between an -type semiconductor electrode and a simple redox system. The plots show the transition from ideal diode behavior to metallic behavior with increasing density of surface states at around the Fermi-level of the solid (indicated in the figures). This is also clear from the plots below, which show the change of the interfacial potential drop over the Helmholtz-layer (here denoted as A(Pfj) with respect tot the total change of the interfacial potential drop (here denoted as A(p). Results from D. Vanmaekelbergh, Electrochim. Acta 42, 1121 (1997).

The foregoing discussion strictly refers to semiconductors in single-crystal form. Amorphous and polycrystalline counterparts present other complications caused by the presence of defects, trap states, grain boundaries and the like. For this reason, we orient the subsequent discussion mainly toward single crystals, although comparisons with the less ideal cases are made where appropriate. The distinction between metal and semiconductor electrodes also becomes important when we consider the electrostatics across the corresponding solid-liquid interfaces this is done next. 

The semiconductor electrode must be ideally polarizable over the potential range of interest. This means that there is no leakage current or Faradaic reaction to allow charge transfer across the semiconductor-electrolyte interface. This restriction is not too important if measurements are taken at sufficiently high frequency that the effects of Faradaic reactions are suppressed. 

The system comprising the resistor Re and capacitor C in series provides an example of a class of systems for which, at the zero-frequency or dc limit, current cannot pass. Such systems are considered to have a blocking or ideally polarizable electrode. Depending on the specific conditions, batteries, liquid mercury electrodes, semiconductor devices, passive electrodes, and electroactive polymers provide examples of systems that exhibit such blocking behavior. 

Photoelectrochemical (PEC) reduction of CO2 with a p-type semiconductor electrode can be regarded as one of the solar energy conversion technologies and is important from a view-point of the global environmental problems. The reaction proceeds by essentially the same mechanism as photosynthesis and is of much interest as an artificial model for it. A number of studies have been made [1], but the photovoltage or the solar-to-chemical energy conversion efficiency still remains relatively low. We reported [2-4] that a p-Si electrode modified with small metal (Cu, Au and Ag) particles worked as an ideal-type electrode for the PEC reduction of CO2 in aqueous solutions. In the present paper we will report that the electrode of this type is also effective for the PEC reduction of CO2 in non-aqueous solutions which have high CO2 solubility. 

Although the very short lifetime and photodecomposition of the bis-pyridyl polymer render it far from an ideal case, we have used it to obtain sensitized photocurrents at a semiconductor electrode (37). Thick films (T 10 ° moles/cm2) of the polymer on n-T102 show an obvious emission when Irradiated with visible light. In acetonitrile solution with hydroquinone as supersensi-tizer, stable photocurrents (0.2-0.3 pA/cm2) are observed and the... 

Although this description is frequently given of the semiconductor-solution junction, in fact, such reversible behavior of a semiconductor electrode is rarely found, especially for aqueous solutions. This lack of equilibration can be ascribed to corrosion of the semiconductor, to surface film (e.g., oxide) formation, or to inherently slow electron transfer across the interface. Under such conditions, the behavior of the semiconductor electrode approaches ideal polarizability (see Section 1.2). 

A schematic representation of the ideal electron-transfer rate and transfer coefficient as functions of potential for a semiconductor electrode is shown in Figure 18.2.6. Although there have been numerous studies with semiconductor electrodes, such ideal behavior is rarely seen (45, 47, 49, 57-59). Difficulties in such measurements include the presence of processes in parallel with the electron-transfer reaction involving dissolved reactant at the semiconductor surface, such as corrosion of the semiconductor material, effects of the resistance of the electrode material, and charge-transfer reactions that occur via surface states. 

E.C. Dutoit, RLv Melrhaeghe, F. Cardon, W.P. Gomes, Investigation on the frequency-dependence of the impedance of the nearly ideally polarizable semiconductor electrodes CdSe, CdS and Ti02. Berichte der Bunsengellschaft fur physikalische Chemie 79, 1206-1213... 

---