Résumé
Natural hydrogen (H2) generation has long been believed to only occur during serpentinization of ultramafic to mafic rocks at mid-oceanic ridges. Nevertheless, several studies have shown that H2 could also be emitted in the middle of continents, within ancient cratonic basements. Here we report the study of drill cores from the H2-emitting DR1-A well (Kansas, USA). In this area, the Precambrian granitoid basement rich in ferromagnesian minerals produces high levels of H2. Petrographic observations of samples from a depth of 452 m within the basement show fractured olivine distributed in a matrix of amphibole, pyroxene, feldspar, quartz, and oxides. As revealed by SEM, these Fe-rich olivines (fayalite, Fe2SiO4) are crosscut by veins filled with two types of phyllosilicates, associated with iron oxides. XANES spectroscopy at the Fe Ledge showed that these phases contain both ferrous and ferric iron: the external part of the veins exhibits almost 30% of ferric iron versus only 20% for the center part. Transmission electron microscopy investigations conducted on ultrathin FIB sections extracted across these veins show Fe-phyllosilicates structurally related to chlorites. Although rare in such geological context, the presence of fayalite and chlorite with a heterogeneous but significant ferric iron content strongly suggests water-rock interactions leading to the local generation of H2. However, an alternative explanation would be the trapping of H2 by the Fe-rich phyllosilicates composing the external veins and having crystallized at high temperature. In fact, as shown by XCT, SEM-EDS, µRaman and NanoSIMS, the external veins are restricted to fayalites crystals while the central veins propagate throughout the entire matrix, strongly suggesting a two-stages history. The further interactions of the external phyllosilicates with subsequent infiltrating fluids could have been responsible for the release of molecular hydrogen. Post-processing of XCT data combined with XANES data allow to estimate the volume fractions of those Fe(III)-rich phases, and thus to quantify the potential H2 production associated to those phyllosilicates. The present findings have important implications for exploring similar Precambrian granitoids present in every continent that might correspond to potential natural H2 production sites.