For centuries, mankind has mostly used fossil fuels, i.e., natural gas, coal, and oil for its energy needs. With the fast rise of the world population and the rising standards of living in the developing countries, the amount of energy the world is going to need in the coming decades will grow enormously. Due to environmental concerns and the need to secure the energy supply, more actions have to be taken in the development and implementation of cleaner technologies, based on solar, wind, geothermal power, and biomass, combined with storage technologies such as batteries and hydrogen. Moreover, to fulfill the future global energy demand, all of the available energy sources will be needed. Despite the fact that solar energy is abundant, clean, and widely available, only a small percentage of this energy is utilized through conversion to electricity by photovoltaic systems. This is partly due to the need for smart technological inventions to make the alternative energy cost-effective and competitive with the conventional energy production. The Netherlands intends to increase the percentage of energy produced by sustainable sources from 4% nowadays to 20% in 2020. When considering sustainable energy production, also energy storage has to be investigated, since especially solar and wind energy are not continuous. In the chain of solar energy production and the need for energy storage, hydrogen is a promising storage candidate due to its high gravimetric energy density and its nonpollutive combustion product, water. Nowadays, hydrogen can be produced by a number of processes, such as electrolysis of water, steam reforming of natural gas, and biomass conversion, which directly or indirectly release large amounts of carbon dioxide into the atmosphere. An elegant alternative will be the use of solar energy for the electrolysis of water in a photoelectrochemical cell. Direct photoelectrolysis has the potential to be economically more attractive than coupled systems of photovoltaic cells and electrolysers. Additional information regarding this subject can be found in Chapter 1, as well as the necessary requirements for the photoelectrodes. In this thesis three different metal oxides, TiO2, InVO4, and Fe2O3 are investigated as photoanodes for water splitting applications. The aim of the research presented in this thesis is two-fold. First, the use of low-cost deposition techniques for the preparation of efficient thin-film photoanodes is explored. Special attention is given to the possibility to introduce dopants in a controlled manner. Second, the influence of the presence of ionic point defects on the photoelectrochemical performance of the materials is investigated. Titanium dioxide has long been considered as one of the most promising semiconductors for photoelectrolysis applications due to its low cost, non-toxicity, and excellent stability against corrosion. However, because of its wide bandgap (3.2 eV for anatase) the utilization of TiO2 typically remains confined to the UV light region, which constitutes only a small fraction (2-3%) of the solar radiation reaching the earth surface. Therefore, shifting its photoresponse into the visible range of the solar spectrum would enhance its potential for chemical solar energy conversion. Attempts to achieve this performance are typically focused on adding dopants. In Chapter 2 Fe- and C-doped TiO2 thin films have been investigated. Pulsed spray pyrolysis is employed to deposit high quality dense films. The prepared films are crystalline with an anatase structure and a post-deposition anneal does not change the morphology. For the Fe-doped TiO2 a small sub-bandgap photoresponse is observed, which is attributed to the presence of additional states located just above the valence band. Little or no visible-light photoresponse is observed for the C-doped anatase TiO2, which is attributed to a (too) low carbon content. However, the photocurrent at h?>Eg is significantly larger for C-doped TiO2 than for undoped TiO2. The strong enhancement of UV photoresponse is most likely caused by a change in the electronic structure of the material due to the presence of carbon and/or related defects. Photoluminescence measurements suggest that the defects present in oxidized carbon-doped anatase resemble those present in undoped, partially reduced TiO2. Although the exact nature of these defects is unknown, impedance measurements reveal a donor density of 1019-1020 cm-3 in C-doped TiO2. The high UV photoresponse is rather surprising for such a high donor density, and illustrates the intriguing, but still poorly understood properties of anion dopants. From the synthesis point of view, the preparation of carbon-doped TiO2 photoanodes was found to be challenging. A more simple approach is to prepare C-doped powders by solid-state reactions at high temperatures. Such a system can serve as a convenient screening method for selecting suitable anion dopants by analyzing the evolved gases, even if the efficiencies are very low. In Chapter 3 the focus is on the addition of carbon as dopant to anatase TiO2 by a post-deposition thermal treatment in a hexane-rich environment. Both thin films prepared by spray pyrolysis and mesoporous TiO2 photoelectrodes prepared by doctor-blading a paste of TiO2 nanoparticles are investigated. It is found that the carbon is mainly located at the surface of the TiO2. While it causes a black coloring of the mesoporous TiO2 film, it does not enhance the photocatalytic activity in the visible part of the spectrum. Only a small amount of carbon (<0.1 at%) diffuses into the bulk of the material where it causes a small shift (0.05-0.1 eV) of the absorption edge towards higher wavelengths. An intriguing observation is that the presence of carbon in anatase TiO2 increases the anatase to-rutile transformation temperature from 600 °C to beyond 800 oC. Based on the temperature at which this transformation occurs, it has been concluded that films sprayed under a CO2 atmosphere contain significantly more carbon compared to the TiO2 samples subjected to the post anneal hexane treatment, although in both cases the amount of carbon incorporated into the TiO2 films is not sufficient to generate a visible light response. To increase the concentration of C in TiO2 the oxidation of TiC films may be a more promising route. Although the amount of C as dopant present in TiO2 is not sufficient to generate a significant visible light response, the rather large increase in the anatase-to-rutile phase transformation temperature enables one to retain the anatase structure during high temperature heat treatments. Such treatments are sometimes employed to improve the contact between small particles, and this can now be achieved while avoiding the transformation of anatase to the less photo-active rutile phase. In Chapter 4 the ternary oxide InVO4 is investigated as photoelectrode, since InVO4 is one of the few ternary oxides that shows hydrogen evolution from pure water under visible light illumination. Spray pyrolysis is explored as a novel method to produce thin films of nearly phase-pure orthorhombic InVO4 at low temperatures. In contrast, most of the complex oxide photocatalyst powders described in the literature are prepared using high-temperature solid state reactions. Optical absorption analysis shows that InVO4 has an indirect bandgap of ~3.2 eV with a sub-bandgap absorption at 2.5 eV, which is attributed to the presence of deep donors in the space charge region of the material. The InVO4 photoresponse is mainly in the UV region of the solar spectrum and the photon-to-current conversion efficiency (IPCE) is less than 1%. The amount of visible light absorption scales with the total surface area, which explains why InVO4 powders absorb much more light than compact films. Detailed electrochemical impedance analysis of InVO4 thin-film photoelectrodes shows that the high concentration of deep donors causes a very narrow space charge region (few nm) in which the photo-generated electron and holes are separated. Ni and Cu dopants added to the InVO4 films do not improve the response and seem rather to act as recombination centers than as acceptor-type dopants. A dielectric constant of 50 was calculated for the first time and a flatband potential of 0.04 eV vs. RHE was determined, which confirms that the InVO4 is able to evolve H2. The main bottleneck of the thin film InVO4 compared to powders appears to be the high concentration of (deep) donors and a concomitant small depletion layer width. Moreover, accurate control over the stoichiometry of the metal ions for InVO4 remains challenging compared to simple binary oxides. This is an important and general aspect that needs to be considered when designing photoelectrodes based on ternary and more complex metal oxides. Iron oxide (Fe2O3) electrodes are the subject of Chapter 5. Hematite Fe2O3 has potential as a photoanode for water photoelectrolysis. It is stable in aqueous solutions with pH >3 and it has a near-optimal bandgap of 2.1 eV that enables up to 32% sunlight absorption. It is also an abundant and inexpensive semiconductor. However, certain intrinsic properties of Fe2O3 limit its performance as a photoelectrode. It shows a poor conductivity, which often leads to a high recombination rate, the energy level of the conduction band is too low for hydrogen evolution, it has slow kinetics for water oxidation, and a modest optical absorption coefficient. The main goal of the research in this chapter is on the preparation of Fe2O3 nanorods perpendicular to the substrate. Such a morphology could improve the efficiency of Fe2O3 photoanodes by decreasing the diffusion path length of the photo-generated holes. To achieve this, electrodeposition of Fe followed by thermal oxidation (EDOX) is explored as a new method to obtain these photoelectrodes efficiently in a short time and with the possibility to control the morphology. It is observed that after reaching a certain thickness or at a sufficiently high deposition current (2 mA/cm2) nano-sized needle- like features start to form. A further increase of the deposition current to 3 mA/cm2 results in a columnar growth consisting of small irregular crystallites stacked on top of each other. In contrast, it has been observed that when a Si precursor is present in the electrolyte solution during the electrodeposition, compact dense Fe2O3 films consisting of spherical particles are formed after the post-deposition oxidation. XRD analysis and Raman spectroscopy show that the Fe2O3 films mainly consist of the ?- Fe2O3 hematite phase. An important advantage of the EDOX method is that low-cost equipment can be used. Moreover, the direct thermal oxidation of electrodeposited Fe avoids the formation of undesired intermediate phases, such as FeOOH. Finally, the electrodeposition process takes place in a non-aqueous solvent, which allows the use of certain dopant precursors (e.g. TiCl4) that cannot be used in aqueous systems. This is an important advantage of this approach compared to other efforts in the literature. Although the EDOX method has been successfully demonstrated, an important remaining challenge is to improve control over the morphology and the doping density. Further improvements of the morphology may be possible by adding additives to the electrolyte solution, or by controlling the process parameters during oxidation of the electrodeposited iron. In conclusion, it has been demonstrated in this thesis that high-quality metal oxides can be prepared by low-cost techniques. Concomitantly, the impact of the ionic point defects on the performance of those photoanode materials is addressed and important material properties are determined. The attempts to enhance the visible-light response does not always result in higher efficiencies, due to low dopant concentrations in TiO2 films, poor electronic properties of InVO4 thin films, and morphology challenges for Fe2O3 to overcome its limitation as photoanode material. Nevertheless, important properties of the thin film photoanodes have been investigated and determined. No binary oxide solely can split water with a reasonable efficiency unless a tandem cell approach with suitable oxides is used. In principle, a lot can be gained by directing future research efforts towards ternary and even more complex metal oxides. In this case, however, the control over the stoichiometry of the metal ions remains challenging compared to simple binary oxides as has been demonstrated by the results on InVO4. While for some of the alternative energy sources technological breakthroughs have been achieved and new technologies are developed, for the hydrogen economy and in particular for the photoelectrochemical water splitting with metal oxides, more research is needed in developing low-cost techniques to make highly structured electrodes with large aspect ratios, to develop synthesis and/or doping strategies to suppress undesired defects present in the ternary oxides with non-stoichiometric metal:metal ratios, and not the least, in finding new and improving existing materials. When these challenges are met, photoelectrochemical water splitting may emerge as an economically viable and truly renewable pathway towards clean hydrogen.