The Geochemistry of Silica in Icelandic Geothermal Systems

Abstract

The geochemistry of silica in Icelandic geothermal fluids was studied. The data considered span over 60 years of a sample collection of both sub-boiling liquid water and two-phase well discharges, in total ~1650 samples from >30 geothermal areas. The measured concentration of silica in the samples ranged from >10 to ~1000 ppm. It was observed to increase with increasing temperature, typically <200 ppm at <75 °C whereas for boiling and close to boiling water (90-100 °C) concentrations up to 700 ppm were observed. Silica concentrations in two-phase well discharges of high-temperature areas were much greater than for the low-temperature areas, typically ~400-1000 ppm with highest values observed at Krafla, Hellisheiði and Nesjavellir. Aqueous speciation calculations revealed that the dominant Si species in all cases were H4SiO4(aq) and HSiO3- at acid to neutral and alkaline pH values, respectively, with NaHSiO3(aq) also being important at elevated Na concentration. pH was observed to be the major factor controlling aqueous silica species distribution with temperature and salinity being less important. Geothermal fluids with neutral to alkaline pH values and temperatures >20 °C were generally observed to be close to saturation with respect to common silica containing geothermal minerals.

In contrast, acid fluids were observed to be undersaturated. At low temperatures, these are typically clays and zeolites whereas, at temperatures >230 °C feldspars, prehnite and epidote together with quartz are often most important. Geothermal geothermometry relies on the assumption of such fluid-mineral equilibria, that are in turn primarily dependent on the temperature at geothermal conditions. It follows, that equilibrium of the fluids with many silica containing geothermal minerals may be used to predict reservoir temperatures. The simplest is only involving silica like quartz and chalcedony, but more complex reactions involving two or more minerals and many components may also be applied. It is further demonstrated that the silica concentrations in geothermal fluids may be used to predict indirectly spatial reservoir temperature variations, here demonstrated for the low-temperature systems at the southern lowlands and at the high-temperature system at Krafla.