Conclusions
By controlling the formation and the thickness of PSII layer immobilized onto TiO2/ITO electrodes and by studying the behavior of the photocurrent in the presence and absence of external mediators and an QB-site inhibitor, we have shown that electrons are transferred to the TiO2 directly from QA −•. This is not unexpected since it is a relatively low potential electron carrier that is close to an exposed surface of the protein and electron transfer by this route has been suggested, though not demonstrated, earlier [30]. Unexpectedly, the rate-limiting step for photocurrent formation is electron transfer through the TiO2. Mobile electron carriers (DCBQ, PpBQ and quercetin) are able to take electrons from the TiO2 to the ITO thereby enhancing the photocurrent. The slow rate of electron transfer through the nanostructured TiO2 is due to its conduction band (Ec) being far above both the reduction potential of QA in PSII (Fig. 7) and also Fermi level (Ef) which is imposed by the applied bias, resulting in very low electron mobility in the nanoporous TiO2. In these circumstance electrons arrive at the semiconductor (in the Fermi level) at an energy level well below the conduction band edge they are thus slow to enter the conduction band if at all. Instead they may remain close to the surface of the material in lower energy states, available for interactions with mediators and slow to migrate to the conducting electrode. It has been suggested that a driving force of at least ΔGinj = −0.2 eV is required in order to obtain efficient electron injection from an excited dye-molecule into the conduction band of a semiconductor [65]. It seems likely that a similar requirement will apply to electrons injected from biological systems. It can be seen that even the short-lived Pheophytin anion radical (Phe/Phe-• Em ~ −500 mV) would be a poor electron donor to TiO2.
