Nuclear waste repositories are being installed in deep excavated rock formations in some places in Europe to isolate and store radioactive waste. In France, the Callovo-Oxfordian formation (COx) is a potential candidate for a nuclear waste repository. After the closure of the underground nuclear waste repository, aqueous corrosion of the steel canister and, to a lesser extent, radiolysis of water would produce significant amounts of H2. This H2 can interact with materials from the repository and with the surrounding clay host formation. The COx formation contains pyrite (FeS2), which has been demonstrated to react with Hydrogen gas (H2) (Truche et al. 2010) at temperature ranging from 90°C to 180°C. This work aims at understanding these interactions within the range of temperature prevailing into COx using electrochemical techniques and the analysis of solutions and interfaces. After pyrite electrodes had been assembled, various electrochemical disturbances were applied to this material (and to platinum for comparison) while it was submerged in a partially reconstituted solution of COx pore water, enclosed in a High Pressure Thermo-Reactor (HPTR) at 90 °C (present work; Betelu et al., 2012), or in a Low Pressure Thermo-Reactor at 25°C (see poster session; Ignatiadis et al., 2012), in the absence and in the presence of pyrite grains (particle size between 40 and 63 µm) and H2 (0 partial pressure or 1 bar). In addition to the electrochemical behaviour of the platinum and the pyrite, the pH, temperature and pressure of the liquid medium were monitored. Analyses of the surface and the inside (cross-section) of pyrite electrodes subjected to the reactivity of H2 under the various conditions described above enable us to compare the alterations undergone by the pyrite. It is clear that the pyrite reacts with H2 and that its alteration is indicated by the presence, at the surface of electrodes, of a black scale whose Fe/S ratio is near 1 (FeS), while the ratio within the pyrite is 0.5 (FeS2). The growth of this iron sulphide scale in the pyrite dissolution pits, the proximity of the source and the iron sulphide suggests that pyrrhotite easily forms a core and increases more rapidly than the pyrite dissolution. The intensity of the pyrite alteration by H2 seems to increase with the imposed cathodic potential on the pyrite electrode. Potentiometric measurements demonstrated that EPt and EPy decrease in the presence of H2 to reach a stable redox potential controlled by the H+/H2 couple. However, EPt and EPy drop lower, but less rapidly, in the presence of pyrite compared to the case without pyrite grains. Pyrite goes through a period during which EPy is greater than EPt. This period corresponds to the (alkaline dissolution, then) reduction of pyrite by H2, but without the pyrite's being entirely covered by pyrrhotite. EPt and EPy are identical at this time. The corrosion currents are strong on pyrite in the presence of H2 and increase with the imposed cathodic potentials. They are much stronger than those measured without pyrite grains. The initial reduction reaction rate is rapid due to the high reactivity of pyrite surface, hence the rapid increase in the sulphide content in the bulk solution, which seems to hinder the progress of the reaction up to a rapidly reached sulphide concentration plateau. Pyrite dissolution and pyrrhotite precipitation are coupled reactions. The fluid composition remains fairly steady when pyrite dissolution balances pyrrhotite precipitation, the iron content remaining very low in the solution. The simulation of the data acquired throughout Electrochemical Impedance Spectrometry enabled to determine an equivalent electrical circuit (EEC). The EEC parameters modelled at different moments and under different conditions made it possible to follow the phenomena occurring on the electrolyte/pyrite interface. In general, Rt is much higher on pyrite in the presence of grains than without grains. The impedance curves measured on pyrite with grains have a different shape than those without grains: appearance of a loop at high frequencies followed by a diffusion segment towards low frequencies. The impedance curves for imposed cathodic potentials (that normally accelerate the reduction processes at the pyrite electrode), compared to those done at the corrosion potential (Ec), have shapes that progressively slant towards the Zreal-axis, with progressive loss of the diffusion segment chronologically as follows: - Rt, high without H2, decreases progressively in its presence (this means that the charge transfer of the reduction reaction progressively impeded by its consequences); - the diffusion processes (G parameters) accelerate from the electrode to the solution (this means that produced S2- progressively diffuses to the bulk solution); - the diffusion processes are hindered (Rc and C parameters) by increasingly thick FeS scale that covers the pyrite surface. Under these conditions, three steps seem to control the kinetics of the pyrite reduction reaction: - The diffusion of H2 and H2S(aq) between the reaction front and the reaction medium through the porous microstructures of the pyrrhotite - The pyrite dissolution reaction at the surface - The pyrrhotite precipitation reaction at the surface The future exploitation of this electrochemical database will make it possible to determine the kinetics of FeS2 reduction by H2 in a HPTR. The prospects for this work are, therefore, the exploitation of these data and their extrapolation to storage conditions. The apparent consequences of FeS2-H2 interaction into an actual repository cannot yet be discussed, particularly the impact of HS-/S2- freed in solution. Further investigations should be realized, notably at lower temperature and in the presence of COx, in order to determine the reactivity and competitivity of other sources of sulphur.