Carbonate mineralization of volcanic province basalts

Flood basalts are receiving increasing attention as possible host formations for geologic sequestration of anthropogenic CO2, with studies underway in the United States, India, Iceland, and Canada. Basalts from the United States, India, and South Africa were reacted with aqueous dissolved CO2 and aqueous dissolved CO2–H2S mixtures under supercritical CO2 (scCO2) conditions to study the geochemical reactions resulting from injection of CO2 in such formations. Despite the basalt samples having similar bulk chemical composition, mineralogy and dissolution kinetics, long-term static experiments show significant differences in rates of mineralization as well as compositions and morphologies of precipitates that form when the basalts are reacted with CO2 and CO2–H2S mixtures in water. For example, basalt from the Newark Basin in the United States was by far the most reactive of any basalt tested to date. Reacted grains from the Newark Basin basalt appeared severely weathered and contained extensive carbonate precipitates with significant Fe content. In comparison, the post-reacted samples associated with the Columbia River basalts from the United States contained calcite grains with classic “dogtooth spar” morphology and trace cation substitution (Mg and Mn). Carbonation of the other basalts produced precipitates with compositions that varied chemically throughout the entire testing period. The Karoo basalt from South Africa appeared the least reactive, with very limited mineralization occurring during the testing with CO2-saturated water. Compositional differences in the precipitates suggest changes in fluid chemistry unique to the dissolution behavior of each basalt sample reacted with CO2-saturated water. No convincing correlations were identified between basalt reactivity and differences in bulk composition, mineralogy, glassy mesostasis quantity or composition. Moreover, the relative reactivity of different basalt samples was unexpectedly different in the experiments conducted with aqueous dissolved CO2–H2S mixtures versus those with CO2 only. For example, the Karoo basalt was highly reactive in the presence of aqueous dissolved CO2–H2S, as evident by nodules of carbonate coating the basalt grains after 181 days of testing. However, the most reactive basalt in CO2–H2O, Newark Basin, formed only iron sulfide coatings in tests with a CO2–H2S mixture, which inhibited carbonate mineralization.