Abstract Iron–minerals–water interactions are of primary importance in the contexts of underground structure engineering (e.g. reactive barriers or deep geological storage) and for the understanding of secondary alteration processes in primitive meteorites. To improve our understanding of these systems, we determine the mineralogical transformations induced by the association of iron and silicates during a cooling through an experimental simulation of iron–clay interactions with a step-by-step procedure in the range of 90°C to 40°C. The run products and solutions are well characterised, by means of different techniques (X-ray diffraction, scanning and transmission electron microscopy, manocalcimetry, inductively coupled plasma optical emission spectrometry and ion chromatography), and the thermodynamic data concerning Fe-bearing phyllosilicates are well-tested comparing the modelling and experimental results. Therefore, the main mineralogical modifications observed include the remarkable formation of cronstedtite and greenalite, as well as the formation of magnetite at all temperatures, along with a significant dissolution of quartz, mixed-layer illite–smectite clays, illite (affecting more than 70% of each mineralogical phase) and a partial alteration of chlorite, kaolinite and dolomite. The experimental results confirm the reaction path predicted by thermodynamic modelling, i.e. the formation of iron-rich T–O phyllosilicates (cronstedtite and greenalite) and magnetite at the expense of metal iron and silicates. Both the experimental and thermodynamic results presented in this study provide important constraints to well predict the impact of nuclear waste canister corrosion in a claystone media and to better understand secondary alteration processes, which could also affect the mineralogical and chemical composition of primitive meteorites.