Photoanodic dissolution

After demonsfrafing fhe sfabilizafion of CdS- and CdSe-based PEC, using sulfide- or polysulfide-confaining elecfrolytes, Ellis el al. [51] proceeded lo show dial fhe photoanodic dissolution of single-crystal n-type CdTe, which was found to be unstable in a polysulfide electrolyte, could be completely quenched by adding Na2Te in the alkaline solution of NaOH. The photoelectrochemistry in their cell was considered to be consisting of the reactions... 

The photoanodic dissolution also occurs in the electrochemistry and photoelectrochemistry of compact electrodes of these materials. In fact, it is the most serious obstacle to the practical use of semiconductors such as CdS in photoelectrochemical cells The product of corrosion in the absence of oxygen is sulfur. In the presence of oxygen, sulfate ions are formed as in the case of the colloidal particles... 

The photoanodic dissolution of sulfides is described for CdS by the following equations ... 

Methyl viologen (l,T-dimethyl-4,4 -bipyridylium dichloride, MV " ) promotes photoanodic dissolution in aerated CdS solution Figure 8 shows how the rate of dissolution depends on the concentration. The colloid has a weak fluorescence at 620 nm which is quenched by. The curves for fluorescence and dissolution in Fig. 8 are symmetric, which indicates that the two processes have a common intermediate that reacts with M. These effects are explained by the following mechanism ... 

Fig. 8. Intensity of fluorescence and rate of photoanodic dissolution of colloidal CdS as function of the concentration of added methyl viologen...

Photoanodic dissolution in the presence of air and its promotion by methyl viologen was also observed for alkaline solutions of colloidal cadmium phosphide, CdjPj and cadmium arsenide, CdjAsj Bismuth sulfide, Sb Sj, photo-dissolves in the presence of air mainly according to... 

Most striking is the increase in the fluorescence intensity of a CdS colloid as it undergoes photoanodic dissolution. As the colloidal particles become smaller they fluoresce with a greater quantum yield Very small CdS particles prepared by the methods described in section 5.1, fluoresce with a quantum yield of 3 %... 

The recombination of trapped electrons and holes produces the fluorescence. Adsorbed oxygen scavenges electrons producing O2" which also is adsorbed. OJ is a much better quencher than Oj. Its accumulation under illumination therefore leads to the decrease in fluorescence intensity. During the dark period disappears. During the illumination in the presence of oxygen, the colloid undergoes photoanodic dissolution (see Sect. 3.2). The ZnS particles become smaller in this way, and this finally leads to an increase in fluorescence yield as already described for CdS. 

Upon irradiation of n-type GaP in alkaline solution, electron flow from GaP (bandgap = 2.24 eV) to the counter electrode leads to photoanodic dissolution by the following reactions ... 

Wilson JR, Park SM (1982) Photoanodic dissolution of n-CdS studied by rotating ring-disk electrodes. J Electrochem Soc 129 149-154... 

Ellis AB, Bolts JM, Wrighton MS (1977) Characterization of n-type semiconducting indium phosphide photoelectrodes stabilization to photoanodic dissolution in aqueous solutions of telluride and ditelluride ions. J Electrochem Soc 124 1603-1607... 

The photoanodic dissolution of n-silicon in acidic fluoride media provides an example of the complexity of multistep photoelectrochemical reactions [33, 34]. The reaction requires the transfer of four electrons, but it is clear that not all of the steps involve photogenerated holes because the photocurrent quantum efficiency is between 2 and 4. The explanation of the high quantum efficiencies is that the initial hole capture step can be followed by a series of steps in which intermediates with low electron affinity inject electrons into the conduction band. These intermediates can be assigned nominal oxidation states as shown in the following scheme. 

Fig. 8.16. Reaction scheme for photoanodic dissolution of silicon in low intensity limit illustrating the competition between hole capture steps (rate constants k to k ) and electron injection steps (rate constants k to k,). The nominal valence states of the silicon intermediates are indicated. The final product Si(IV) is the soluble hexafluorosilicate species.

HF or H2O. A wide range of processes, including pore formation in n- and p-type silicon in HF solutions, pore formation in n-type silicon in HF solutions under illumination, and photoanodic dissolution of n-type silicon in NH4F solutions, can be explained by these models. In addition, they are consistent with the models developed for open-circuit etching of silicon in fluoride solutions, discussed in Sec. 2.2.2. 

Assuming that the single Si-F ligand at the kink site is sufficiently polar to allow successive reaction of the two Si-Si backbonds with HF, then the overall reaction is consistent with pore formation in n- and p-type silicon where two charges are transferred per silicon atom. For the case of photoanodic dissolution of n-type silicon in NH4F [120-124] the characteristic photomultiplication effect at low light intensities is also consistent with the hole capture/electron injection sequence shown in Fig. 29 a. 

Electron injection from reaction intermediates of the oxidation of n-type semiconductors can be observed as quantum efficiency larger than unity in photocurrent-potential measurements. There are two striking examples in the literature the photoanodic dissolution of n-type silicon in HE solution [91, 92] and of n-type InP in HCl solution [54]. In these cases the quantum efficiency at low light intensity is exceptionally high, four for silicon and two for InP. In the case of Si, this means that only one photon (and thus one hole) is required to dissolve each silicon atom three electrons are injected into the conduction band... 

The assumption of pinned band edges is very often not valid. Lincot and Vedel [50] in an early study of the photoanodic dissolution of -type CdTe used EIS to show that the Fermi level becomes pinned in a wide potential range positive with respect to Vfb. This means that in this range the band-bending within the semiconductor remains constant while the potential changes across the Helmholtz layer. In their analysis, Lincot and Vedel consider the rate constants for change transfer to be exponentially dependent on the Helmholtz potential. 

Photoanodic dissolution does not proceed homogeneously over the surface. 

Here, n is the electron stoichiometry (equivalents/mole) of the photoanodic dissolution reaction, W is the width (cm) of each groove, a is the angle of the groove face with respect to the (100) surface, p is the angle of the crystal slice with respect to the (100) surface, and N is the number of grooves/cm. 

---