Vent fluid chemistry of the Rainbow hydrothermal system (36°N, MAR): Phase equilibria and in situ pH controls on subseafloor alteration processes

摘要

The Rainbow hydrothermal field is located at 36°13.8'N-33°54.15'W at 2300m depth on the western flank of a non-volcanic ridge between the South AMAR and AMAR segments of the Mid-Atlantic Ridge. The hydrothermal field consists of 10-15 active chimneys that emit high-temperature (~365°C) fluid. In July 2008, vent fluids were sampled during cruise KNOX18RR, providing a rich dataset that extends in time information on subseafloor chemical and physical processes controlling vent fluid chemistry at Rainbow. Data suggest that the Mg concentration of the hydrothermal end-member is not zero, but rather 1.5-2mmol/kg. This surprising result may be caused by a combination of factors including moderately low dissolved silica, low pH, and elevated chloride of the hydrothermal fluid. Combining end-member Mg data with analogous data for dissolved Fe, Si, Al, Ca, and H_2, permits calculation of mineral saturation states for minerals thought appropriate for ultramafic-hosted hydrothermal systems at temperatures and pressures in keeping with constraints imposed by field observations. These data indicate that chlorite solid solution, talc, and magnetite achieve saturation in Rainbow vent fluid at a similar pH_((T,P)) (400°C, 500bar) of approximately 4.95, while higher pH values are indicated for serpentine, suggesting that serpentine may not coexist with the former assemblage at depth at Rainbow. The high Fe/Mg ratio of the Rainbow vent fluid notwithstanding, the mole fraction of clinochlore and chamosite components of chlorite solid solution at depth are predicted to be 0.78 and 0.22, respectively. In situ pH measurements made at Rainbow vents are in good agreement with pH_((T,P)) values estimated from mineral solubility calculations, when the in situ pH data are adjusted for temperature and pressure. Calculations further indicate that pH_((T,P)) and dissolved H_2 are extremely sensitive to changes in dissolved silica owing to constraints imposed by chlorite solid solution-fluid equilibria. Indeed, the predicted correlation between dissolved silica and H_2 defines a trend that is in good agreement with vent fluid data from Rainbow and other high-temperature ultramafic-hosted hydrothermal systems. We speculate that the moderate concentrations of dissolved silica in vent fluids from these systems result from hydrothermal alteration of plagioclase and olivine in the form of subsurface gabbroic intrusions, which, in turn are variably replaced by chlorite+magnetite+talc±tremolite, with important implications for pH lowering, dissolved sulfide concentrations, and metal mobility.