Solar photoelectrolysis of water with translucent nano-structured hematite photoanodes

In the context of the diminishing fossil energy resources and global warming, much research is focused on sustainable energy sources. The research that is presented in this thesis falls in this category as it contributes to the development of a device that stores solar energy into a chemical fuel. This device is based on the photoelectrolysis of water in a tandem-cell that produces hydrogen and oxygen under illumination of sunlight. This can be achieved by connecting a larger band gap semiconductor photoanode for oxygen evolution (such as WO3 with Eg = 2.6 eV or Fe2O3 with Eg = 2.0 eV) in series with two dye sensitized solar cells and a hydrogen evolving cathode. Red and green light transmitted by the photoanode is absorbed by the dye, so that a large part of the solar spectrum can be utilized. Fe2O3 absorbs a larger part of the visible solar spectrum as compared to WO3, which results in a 3 times higher theoretical conversion efficiency for the tandem-cell. However, the small hole diffusion length in Fe2O3 of 2 to 20nm is seen as an important reason for the lower practical conversion efficiencies reported in literature. The aim of this thesis project is to improve the photooxidation activity of iron oxide photoanodes in order to improve the practical conversion efficiency of the tandem-cell. For this purpose we have prepared thin films of nano-structured iron oxide with various deposition methods and characterized their performance on the photooxidation of water. The photoanodes are measured in aqueous solution of 1 M NaOH (pH=13.6) under illumination of simulated sunlight (AM 1.5 global of 1000 W/m2). We report the photocurrent at the reversible water oxidation potential of 1.23 V vs. RHE. The first chapter briefly discusses the potential of hydrogen as an energy carrier in a global hydrogen economy and places the tandem-cell in this context. In chapter two the current status of photoelectrolysis of water with iron oxide is briefly reviewed. The experimental set-up for the measurement of the photocurrent under simulated sun light and for the acquisition of photocurrent action spectra are described. This chapter also deals with the spectral mismatch error that is associated with the solar simulator calibration procedure. Chapter three introduces a new solution strategy for efficient photoanodes based on a nanocomposite structure with a hematite film thickness that is commensurate to the hole diffusion length alleviating the problem of poor charge transport. For this purpose nano-porous films of doped and undoped tinoxide and arrays of perpendicularly oriented nanorods of the same material are conformally coated with a thin film of hematite (2-20nm). These electrodes are characterized by SEM, TEM, XEDS elemental analysis, Raman spectroscopy in addition to the photoelectrochemical response. Although, evidence is presented for the successful preparation of the nano-composite structures with a sufficient optical density, the photocurrents obtained from these electrodes are very low. In chapter four we report on thin silicon-doped nanocrystalline α-Fe2O3 films (thickness of 250-500nm) that have been deposited on F-doped SnO2 substrates by ultrasonic spray pyrolysis (USP) and chemical vapor deposition at atmospheric pressure (APCVD). The photoanodes prepared by USP and APCVD gave a photocurrent of 1.17 and 1.45 mA/cm2 respectively. The morphology of the α-Fe2O3 was strongly influenced by the silicon doping, decreasing the feature size of the mesoscopic film. The silicon-doped α-Fe2O3 nano-leaflets show a preferred orientation with the (001) basal plane normal to the substrate. Chapter five treats the preparation and characterization of hematite photoanodes prepared by an improved APCVD setup. Under illumination water is oxidized at the Fe2O3 electrode with higher efficiency (IPCE = 42% at 370 nm and 2.2 mA/cm2) than at the best reported single crystalline Fe2O3 electrodes. This unprecedented efficiency is in part attributed to the dendritic nanostructure which minimizes the distance photo generated holes have to diffuse to reach the Fe2O3/electrolyte interface while still allowing efficient light absorption. Part of the gain in efficiency is obtained by depositing a thin insulating SiO2 interfacial layer between the SnO2 substrate and the Fe2O3 film, and a catalytic cobalt monolayer on the Fe2O3 surface. A mechanistic model for water photooxidation is presented, involving stepwise accumulation of four holes by two vicinal iron or cobalt surface sites. In chapter six we present a detailed investigation of the dependency of the photocurrent response on film thickness and feature size of doped and undoped hematite photoanodes prepared by the APCVD setup discussed in chapter five, in absence of a redox catalyst. Evidence is presented of an unusual high donor density of 1020 cm-3 in the silicon doped photoanodes which would allow a space charge layer to be formed inside the nano-sized features of the polycrystalline film. This suggests that electric field induced charge separation could be an important reason for the photocurrent improvement reported in chapter five. In addition, a strong correlation between feature size and the intensity of a Raman band at 660cm-1 is presented and discussed in the context of disorder in the crystal structure. Chapter seven presents a study of the influence that the deposition parameters have on the photoresponse, morphology, and crystal structure of films prepared by ultrasonic spray pyrolysis. Films are characterized by SEM, XEDS elemental analysis, TEM, XRD, Mössbauer and Raman spectroscopy. A strong correlation is found between the formation of β-Fe2O3 (Bixbyite) and low photocurrents. Silicon doped films enable the electron transport across a film thickness of 1300nm. The absorbed-photon-to-current-conversion-efficiency (APCE) increases with film thickness below 150nm, which indicates an increased recombination near the substrate interface. This could be an important reason for the low photoresponse of nanocomposite electrodes studied in chapter three. In chapter eight we draw some general conclusions with regards to the future development of these photoanodes.