Reactive solubility approach to the dissolution of halogens in basaltic melts

Partitioning of halogens between gas and basaltic melts at 1200-1265°C and 10-2000 bar (Alletti, 2008) has been analyzed in detail by means of the CTSFG model (Moretti and Ottonello, 2005) in order to: 1) evaluate redox conditions consistent with related experimental information such as PH2 and S2−/S6+ ratio, 2) estimate the relative abundance of H2O, CO2, KCl, NaCl and HCl of the gas phase, 3) investigate the solubility mechanisms of chlorine and fluorine in (basaltic) melts. For chlorinebased runs, it was found that for input [Na+K]/Cl atom ratios > 0.45 the proportion of gaseous NaCl and KCl exceeds that of HCl. In terms of solubility mechanisms, halogen chemical reactivity can be assessed in terms of dissolution into the melt as chloride or fluoride ion, i.e, Cl− and F−. A complementary solubility mechanism can be invoked to improve the precision and explain somehow unexpected features, such as i) a P-independent background values of halogen solubility, and, for chlorine only, ii) a XCl,melt-squareddependence shown by Cl dissolution. In this additional mechanism, undissociated alkali-halogenides dissolve in the melt up to a solubility limit which depends on structural conditions, i.e. on melt composition for the same T. Although this could be a fascinating hypothesis, somehow recalling the concentration limits of Henry’s law behavior for trace element dissolution into minerals, it is then questioned whether this may reflect or not mixing of a molten salt-like component with the silicate melt. Partition coefficients derived from equilibrium constants of reactions involving Cl− and F− show well the role of water vapor as stripping agent of chlorine. Therefore, the more the CO2 in the system, the more the amount of chlorine getting dissolved into the melt phase.