The source of lithium in connate fluids: Evidence from the geothermal reservoir at Soultz-sous-Forêts, Upper Rhine Graben, France

Geothermal brines in the Upper Rhine Graben have been used as a spa or for salt production since Roman times. Heat and power are generated in geothermal power plants since 2007. Recently, their elevated Li-content has additionally attracted economic interest. This increased interest is in contrast with our understanding of the geological-hydrothermal evolution. We use petrology, major and trace element mineral chemistry and mass balance calculation from drill cores that intersect granitic geothermal reservoir rocks at Soultz-sous-Forêts between Strasbourg and Karlsruhe to shed light on fluid-rock interaction in a reservoir that is actively used for heat and power generation. The alkali feldspar and the two-mica granite in the reservoir have a typical plagioclase, K-feldspar, quartz, biotite and muscovite assemblage with some accessories of titanite, apatite and zircon. Two hydrothermal alteration events are distinguished: (1) albitization of the feldspars; (2) distal replacement of feldspars by sericite and calcite, of biotite by chlorite and titanite; and proximal to hydrothermal veins replacement of the feldspars by sericite and kaolinite. Event 2 feldspar alteration quantitatively releases Pb and Ba to the fluid, whereas Rb, Cs, Sr and Zn show different behaviour depending on whole-rock and mineral composition. Event 2 biotite-chlorite alteration releases Li, Rb, Cs, Sr, Ba, Zn and Pb to the fluid. Mass balance calculation indicates that Si, Fe, Ca, K, Rb, Sr, Zn and Pb contents of the Soultz-sous-Forêts geothermal brine may be explained by fluid-rock interaction in the reservoir. However, the reservoir rock volume that needs to be leached in order to reach recent brine composition varies by several orders of magnitude between the different elements. Many of the elements may be leached during hydrothermal alteration, however in particular Li and Cs require unrealistic fluid-rock ratios of >1/300. These considerations indicate that Na, Ca, Li, Cs and Ba need an additional external source. Based on this, we propose a model where Middle Triassic bittern brines already enriched in Li, Rb, and Cs reacted with the reservoir rocks during hydrothermal event 2 and subsequently mixed with Jurassic-Cretaceous marine water that dissolved evaporites during downward migration. This agrees with Jurassic-Cretaceous illite ages from various sites in the Black Forest and indicates a complex ∼150 m.y. hydrothermal evolution for the brines. There is likely no single source of Li, and it is likely derived from complex fluid-rock interaction with the sedimentary (evaporite) and, less importantly, the crystalline strata of the Upper Rhine Graben. Critical for Li-resource development is the complex hydrothermal history of connate fluids that interacted with sedimentary strata and the preservation in deep-seated reservoirs.