The Effect of Composition on the Volatile Solubility of Mixed H2O-CO2 Fluids in Arc Basalts

The volatiles H2O and CO2 are abundant in melt inclusions from arc lavas and play an important role in magmatic processes associated with subduction zones. While there are a significant number of studies into the solubility behavior of these volatiles, they do not always cover the necessary range of geologic conditions needed to accurately interpret pre-eruptive conditions recorded by melt inclusions. Most studies have considered only a pure, single component fluid, and are at either very high (greater than 1.0 GPa) or low (less than 200 MPa) pressure. The few studies that have investigated mixed volatiles (H2O and CO2) in melts have only rarely measured the fluid composition in equilibrium with the melt. In addition, the overall effect of bulk melt composition is not well constrained. Therefore, because melt inclusions found in a single lava commonly have a wide bulk compositional variation, the intensive parameters (P-T) cannot be accurately modeled for the solubility surface using the bulk composition of the host lava. We present here the initial results of an experimental study aimed at defining the mixed volatile solubility surface in P-T-Xi space, for two associated melt compositions from Central America; one representing a "typical" lava composition and the other a Ca-rich melt observed in inclusions found in the same lava type. Superliquidus experiments (0.4-0.7 GPa; 1200° C) were conducted in a non end-loaded piston cylinder using a double capsule method. An outer Pt capsule containing an undersaturated hydrous rhyolite was used to prevent the brittle failure of an inner Au-Pd capsule containing basalt and added volatiles (variable H2O:CO2 ratios; XH2O(fluid) ~0.3-0.70). Oxygen fugacity was monitored in the inner capsule using a thin Pt strip and measuring the resulting Pt-Fe alloy. After equilibration and quenching, fluid compositions were measured by vacuum manometry, and the volatile content of the glasses was measured by high T vacuum manometry, secondary ion mass spectrometry, and infra-red spectrometry. Results at 0.4 and 0.6 GPa and XH2O fluid of ~0.6 show the dramatic influence of calcium on CO2 solubility. We observe an ~2-fold increase in CO2 content when the basalt CaO content is increased by ~20 percent (~6000 versus 2,500 ppm at 0.6 GPa), whereas H2O solubility is unchanged or even diminished. Our results indicate that current solubility models for do not accurately predict H2O and CO2 abundance in high Ca basalt at these pressures, and by extension, the intensive parameters predicted by these models for measured volatile abundances in high-Ca melt inclusions are inaccurate. The high solubility of CO2 in calcic basalts also has significant influence on arc geochemistry and volcanic behavior. For example, high-Ca basalts could act as carriers of volatiles and additional alkaline earths. Through underplating and mixing, high-calcic basalts could transport CO2 into magma chambers, thereby contributing to fluid saturation and other physical conditions leading to eruption.