Abstract
Owing to the high oil price and environmental issues, the use of renewable energy have attracted much attention; particularly, using photoelectrochemical (PEC) cell for water splitting. PEC cell is a device that can convert the intermittent solar energy directly into that storable and clean chemical energy (H2) by means of dissociating water. PEC water splitting uses the semiconductor as photoelectrode, which are mainly transition metal oxides including TiO2, WO3, and Fe2O3. Upon the light incidence, minority carriers (holes in photoanode) may be generated and react with OH- of alkaline aqueous solutions to release oxygen molecules at anode and hydrogen molecules simultaneously at cathode. To maximize the cell performance, photoelectrode must have the appropriate band gap energy to absorb as many as possible of photons from solar irradiation, and generate electron-hole pairs that bear sufficient energy to split the water molecule. Among the potential materials, hematite (α-Fe2O3) has been recognized to have the most appropriate band gap energy (2.1 eV). In addition, it also appears to have high photoelectrochemical stability in electrolyte. However, the performance of hematite has not been satisfactory, and shows large performance discrepancy among hematite prepared with different methods. In this study, we prepared two hematite films using respectively the anodic electrodeposition and Fe oxidation method. The differences of their characteristics in water splitting were then analyzed with XRD, SEM, UV-vis absorption spectra, photocurrent, Mott-Schottky and electrochemical impedance spectroscopy (EIS) measurements. From XRD analysis, both samples appear with hematite crystal structure, and their morphology appears to be the aggregate of nanoparticles with the size of about 30 nm. In UV-vis absorption check, the sample with anodic electrodeposition III always shows better absorption than that with Fe oxidation. Nevertheless, the performance of solar water splitting can only occur in cell with photoanode prepared with Fe oxidation method, the photocurrent can reached 0.126 mA/cm2 (at 1.23 VRHE, AM1.5G, 92 mW/cm2). In contrast, no water splitting was observed from the cell using anodic electrodeposition films, no matter how high intensity of incident light. From Mott-Schottky plots, high level concentration of electrons (~1×1020 cm-3) was derived from anodic electrodeposition sample, and no capacitance change can be detected between dark and illumination condition. On the other hand, Fe oxidation method shows lower concentrations which is about 6×1019 cm-3, and capacitance tends to vary with light incidence. Besides, the Fermi level pinning was observed when the sample is subjected to a positive bias. From the aforementioned results, it is evident that the concentration of bulk defects may be the cause for no water splitting observed in the sample prepared with anodic electrodeposition, and has nothing to do with morphology and crystal structure of iron oxides. From the results of EIS analysis, it is clear that Fermi level pinning is directly caused by the high surface state density. Despite that, there seems no explicit correlation between the magnitude of photocurrent, the rate of water splitting, and surface defect state level and its density. From the change of band diagram with bias, it is found that the excess minority carrier concentration concerns more about the onset voltage of water splitting.