Redox evolution and mass transfer during serpentinization: An experimental and theoretical study at 200 °C, 500 bar with implications for ultramafic-hosted hydrothermal systems at Mid-Ocean Ridges

Abstract

Highly reducing and high-pH vent fluids characterize moderately low temperature ultramafic-hosted hydrothermal systems, such as the recently discovered Lost City hydrothermal field at 30°N Mid-Atlantic Ridge Ridge (MAR). To better understand the role of mineral reaction rates on changes in fluid chemistry and mineralization processes in these and similar systems, we conducted an experimental study involving seawater and peridotite at 200 °C, 500 bar. Time series changes in fluid chemistry were monitored and compared with analogous data predicted using experimental and theoretical data for mineral dissolution rates. Although there was qualitative agreement between predicted and measured changes in the chemical evolution of the fluid for some species, the rate and magnitude of increase in pH, dissolved chloride and H 2 did not agree well with predictions based on theoretical modeling results. Experimental data indicate that dissolved H 2 abruptly and intermittently increased, reaching a value only approximately 20% of that predicted assuming magnetite as the primary Fe-bearing alteration phase. The distribution and valence of Fe in primary and secondary minerals reveal that the most abundant secondary mineral, serpentine, contained significant amounts of both ferric and ferrous Fe, with the less abundant brucite, also being Fe-rich ( X Fe = 0.3). Surprisingly, magnetite was present in only trace amounts, indicating that H 2 generation was largely accommodated by the formation of Fe–chrysotile. Accordingly, the diversity of Fe-bearing secondary minerals together with rates of serpentinization less than theoretically predicted, account best for the relatively low dissolved H 2 concentrations produced. Thus, the experimental data can be used to obtain provisional estimates of thermodynamic data for Fe-bearing minerals, enhancing the application of reaction path models depicting mass transfer processes during serpentinization at mid-ocean ridges. Similarly, the observed differences between theoretically predicted and experimentally measured pH values result from constraints imposed by complex patterns of mass transfer inherent to the experimental system. In particular, the experimental observation of a late stage increase in Na/Cl ratio likely results from the dissolution of a Na 2O component of clinopyroxene, which causes pH to increase sufficiently to induce precipitation of a Ca-bearing phase, perhaps portlandite. As with the redox variability observed during the experiment, this event could not be predicted, underscoring the need to use caution when modeling alteration processes in the chemically complex ultramafic-hosted hydrothermal systems at elevated temperatures and pressures.