Granite-water interactions in a flow-through experimental system with applications to the Hot Dry Rock geothermal system at Rosemanowes, Cornwall, U.K.

Abstract

A suite of laboratory experiments reacting granite with fluids typical of those circulated in a Hot Dry Rock (HDR) geothermal system at Rosemanowes, Cornwall has been conducted in the temperature range 60–100°C under flow-through conditions. Experiments of three types were performed: variable flow rate, using a streamwater-type fluid; variable temperature, using pH-buffer solutions; and variable fluid composition, using synthetic analogues of circulation fluids from the HDR site at Rosemanowes. The results of the variable flow rate experiments revealed gains in Si, Na, Ca, Li, B, F and Ba in the output fluids, the magnitudes of which generally decreased with increasing flow rate, whereas Mg, Sr, Fe and Mn were either lost from fluids or showed gains that were invariant with changes in flow rate. The principal reactant minerals were calcite, plagioclase, biotite and quartz, and a smectite clay was the “sink” for chemical components lost from the fluids. Output fluids from the variable temperature, pH-buffered experiments had chemical compositions dominated by dissolution, with few effects attributable to precipitation processes. Although increased temperature increased the net gains of most chemical components in output fluids, this was particularly so for Si, suggesting that the rate of quartz dissolution was increased relative to that of other minerals. Variations in the chemical composition of the fluids from the variable fluid composition experiments showed that only Mg was removed from typical Rosemanowes “injection” fluid, whereas Mg, Ca, Sr, Fe, Mn and, in some instances, F were removed from “production” fluids. Potassium was removed from production fluids with elevated initial concentrations of this component. Smectitic clays, zeolites, calcite and fluorite are inferred products of these reactions. A comparison of processes revealed by the laboratory data with the chemical exchanges observed during HDR circulation tests indicated that the latter involved sorption/precipitation processes to a greater extent, particularly for components such as Si and Li. The determination of the bulk rate of dissolution of the granite by chemical analysis of the fluids from the laboratory experiments varied between 10 −12 and 10 −10 kg/m 2/s at any temperature. These measurements emphasise that rock dissolution is a heterogeneous process, involving the selective removal of the most reactive mineral phases. Rate constants of plagioclase dissolution estimated from the release of Na from the granite (∼10 −11 mol/m 2/s at 80°C) are consistent with previous determinations on isolated pure minerals available in the literature. The rate of tourmaline dissolution has been estimated in a similar manner, using data for B, as ∼10 −11 mol/m 2/s at 80°C. The use of a bulk granite dissolution rate derived from the laboratory experiments to predict the output of Na and Si from the reservoir at Rosemanowes would have under-estimated Na and over-estimated Si as a result of the absence or small contribution of processes such as fluid mixing and the precipitation of secondary solids in the laboratory experiments.