The work presented in this thesis is focused on the photocatalytic water splitting reaction. In particular the overall reaction for production of both hydrogen and oxygen has been investigated using water vapor and several light sources. The enormous amount of energy irradiated by the sun and the consequences of its efficient storage are the main motivations behind this research. Indeed, the achievement of solar-driven water splitting could represent a big step for humanity towards a sustainable society based on renewable energy. The energy problem associated with the increasing world population is discussed in the beginning of this thesis followed by an introduction to the basics of photocatalysis. The experimental setup used in this study and the silicon based μ-reactor technology is described afterwards. Almost the entire work presented in the thesis has been done loading the catalysts in these μ-reactors and analyzing the products of the reaction with a quadrupole mass spectrometer (QMS). Several catalysts have been tested for photocatalytic water splitting: GaN:ZnO, SrTiO3, TiO2, NaTaO3…, most of them in the form of nanopowder and loaded with cocatalyst nanoparticles. In particular, as an introduction to the water splitting experiments, the results obtained with SrTiO2 and TiO2 are presented. These semiconductors are well known examples of materials active under UV illumination. However to achieve high efficiency of solar energy conversion the catalysts needs to be active for longer wavelength. GaN:ZnO is one of the few photocatalysts that is able to achieve overall water splitting with visible light. Therefore the reaction has been studied focusing on this material. GaN:ZnO loaded with Rh2-yCryO3 showed high activity and hydrogen and oxygen could even be detected under illumination with a solar light simulator (A.M. 1.5). The dependence of the activity as a function of light intensity showed a linear behavior for the initial rate and has been studied in comparison with the water consumption rate that can be also detected using our μ-reactor. The effects of the temperature of the reactor and the partial pressure of water are also presented and explained in terms of relative humidity. This has been found to be a key parameter for the gas phase water splitting reaction. The results of this study have been combined in simple expression for the rate of photocatalytic hydrogen production. Hydrogen oxidation experiments have been performed in order to study the water splitting back reaction and explain the high activity of the Rh2-yCryO3/GaN:ZnO. The water formation rate at room temperature was measured with the QMS introducing in the reactor a stoichiometric mixture of hydrogen and oxygen. GaN:ZnO without cocatalyst and loaded with Rh, Pt, Cr2O3/Rh, Cr2O3/Pt, and Rh–Cr mixed oxide was used for this study and the results are compared with their photocatalytic activities. The water splitting back reaction has been tested both in the dark and under illumination and the results clearly show how the water formation is suppressed for the GaN:ZnO loaded with Cr2O3/Rh, Cr2O3/Pt, and Rh2-yCryO3 and at the same time the activity is strongly increased respect to Pt/GaN:ZnO and Rh/GaN:ZnO. The last chapter discusses the efforts that have been done to achieve the goal of obtaining in situ supplementary information to the products detection using μ-reactors. In particular a new kind of μ-reactor that has a Pyrex lid on both sides is presented. With this reactor is possible to measure the absorbance of the materials deposited inside the μ-reactor and to combine optical measurements and spectroscopy with the detection of activity with the QMS. Finally an attempt to combine the μ-reactor with surface science techniques working under UHV is presented. The strategy is based on replacing the standard μ -reactor with an open μ-reactor that can be closed by pressure. The design and realization of the setup and the preliminary problems encountered are described.