Pergamon Geochimica et Cosmochimica Acta, Vol. 61, No. 20, pp. 4281-4295, 1997 Copyright © 1997ElsevierScienceLtd Printed in the USA.All rights reserved 0016-7037/97 $17.00 + .00 eII S0016-7037(97) 00243-3 Water-rock interaction processes in the Triassic sandstone and the granitic basement of the Rhine Graben: Geochemical investigation of a geothermal reservoir L. AQUILINA, * H. PAUWELS,A. GENTER, and C. FOUILLAC BRGM, 1039 rue de Pinville, Direction de la Recherche, 34000 Montpellier, France (Received October 16, 1996;accepted in revised form June 16, 1997) Abstract--Saline fluids have been collected in the Rhine Graben over the last two decades, both from the Triassic sandstone aquifer and the granitic basement down to a depth of 3500m. Their salinities and location are compared in order to distinguish the respective influences of temperature and host-rock mineralogy in the water-rock interaction processes. The comparison shows that sulphates in the sedimentary formations were dissolved by the fluids, which also led to Br enrichment. Mica dissolution has strongly increased the Rb and Cs contents, which then provide an indication of the degree of water-rock interaction. The Sr isotopic ratios are used to compare the fluids with the granite minerals. Two relationships are revealed for the fluids in the sandstone and the granite, one related to widespread mica dissolution, which could have affected both the Bunt- sandstein and the granite, and the other to subsequent plagioclase dissolution, which is observed only in the granite. Computations showed that 12.5g of mica and 1.65g of plagioclase per liter of fluid have been dissolved. The nature of these two relationships suggests two different evolutions for the fluids and the individualization of the two reservoirs during the graben's history. The cation concentrations are mainly controlled by temperature, and are independent of the type of host rock. Equilibrium with the rock mainly caused Ca and K concentration variations, which has induced clear Ca-K and Ca-5 ]80, K-6 '80 correlations. Geothermometric computations indicate that with increasing depth, the cations, the silica and the 6~80 (SO4) geothermometers evolve towards a value close to 230°C. This demonstrates the existence of a hot reservoir in the granite of the graben, at a depth estimated at 4.5-5 km. Copyright © 1997 Elsevier Science Ltd 1. INTRODUCTION The Hot Dry Rock (HDR) geothermal concept has been tested mainly in the U.S. (Fenton Hill), the U.K. (Cammen- ellis ), France (Soultz-sous-For&s), Germany (Bad Urach), and more recently, Japan (Hijiori). Drilling at these sites has made it possible to investigate deep granitic rocks. Natu- ral fluids have been encountered in these environments with salinities of several tens of g/L (Goff et al., 1981, 1985; Edmunds et al., 1985; Matsunaga and Yamaguchi, 1992). In the framework of HDR investigations, several laboratory and natural studies of water-rock interaction processes have been carried out in order to predict fluid circulation patterns in the reservoir, and to understand the origin of these natural saline fluids (Grigsby et al., 1983, 1989; Moore et al., 1983; Savage, 1986; Savage et al., 1992; Richards et al., 1992; Pauwels et al., 1992). These studies have provided a back- ground in natural granite-water interactions. Circulation of saline fluids is well known in the Rhine Graben and saline springs and spas are numerous, e.g., Soultz-sous-For&s, Soultzmat, Niederbronn, Morsbronn, Baden-Baden, etc. Saline fluids were also collected at a depth of about 2 km in a geothermal well at Strasbourg (Fig. 1 ) and in geothermal wells from the eastern part of the graben (Fritz, 1980; Richard et al., 1993; Pauwels et al., 1993). These fluids, which originate from the Buntsandstein aquifer directly overlying the granite, have a salinity which ranges from 100 to 200g/L. Since 1987, Soultz-sous-For&s has *Author to whom correspondence should be addressed (1.aquilina@brgm.fr). 4281 been the French site for the European Hot Dry Rock project. The aim of this project is to create a heat exchanger at a depth of 4-5 km in the granitic basement beneath the sedimentary formations of the Rhine Graben (Kappelmeyer et al., 1991). Two wells (GPK-I and EPS-1) were drilled in 1987 and 1990 to a depth of 2 km in order to investigate the granitic basement. Thick fracture zones allowing the mi- gration of saline fluids in the granite were discovered in the boreholes (Vuataz et al., 1990). All the deep fluids collected in the graben both in the Buntsandstein and in the granite at Soultz are typical sedimentary brines, similar to those found within the same geological formations in the Paris Basin (Fontes and Matray, 1993). Although the Fenton Hill and Carnmenellis fluids are much less saline than those from the Rhine Graben, their study at similar temperatures can reveal some clues as to the water-rock processes at Soultz. During the first phase of the Soultz project, geochemical techniques were used to characterize the chemistry of the fluids sampled in the granite, especially in GPK-1 borehole at 1815m depth (Criaud in Kappelmeyer and Grrard, 1989; Vuataz et al., 1990). Geothermometry results (Pauwels et al., 1993) indicate that the fluids originate from a reservoir at a temperature higher than 200°C, but that the fluid partially re-equilibrates after circulation to an in situ temperature of about 140°C. Geochemists proposed that there was major migration of the fluids within the Buntsandstein aquifer from the eastern deeper part of the graben to the Soultz area (Genter et al., 1989), where the fluids were then injected into the granite along the fault zones (see Fig. 11 ). The deepest eastern part of the graben contains a reservoir at a depth of 5-6 km with a temperature greater than 200°C. As