Concentrations of dissolved O2 are indicated with the blue line (right y-axis), Fe(II) with the green line (right y-axis) and S(-II) with the orange line (left y-axis) in borehole water, starting 50 days before the first H2 injection that occurs at day 0. Initially, H2 injection was done through a small reservoir at the surface via a gas permeable membrane when recirculating porewater. After 50 days, H2 input was suboptimal because of membrane clogging (Supplementary Fig. 2). It is the reason why, starting day 134, H2 was regularly injected directly in the borehole, thus creating a gas phase. Shortly thereafter, it was also decided to stop porewater circulation because of H2 loss at the peristaltic pump. The experiment procedure is indicated at the top of the plot: the light blue line shows when water was recirculated, the dark red line shows when H2 was optimally delivered from the surface, the light red line shows when H2 was sub-optimally delivered from surface, the dark crosses show when the H2-permeable membrane was changed, and red diamonds show when H2 was injected directly into the borehole. Samples are also indicated at the top of the plot: green squares show samples for 16S rRNA gene sequencing, red squares show samples for metagenomic sequencing, and blue square show samples for metaproteomic analysis.