Solubility and Speciation of CO2 in Natural Rhyolitic Melts at 1.5-3.0 GPa - Implications for Carbon Flux in Subduction Zones via Sediment Partial Melts

Sediment partial melt derived from the subducting plate is thought to be a critical agent in transport of trace elements and water to arc basalt source regions. Sediment melts may also act as significant carrier of CO₂ released from subducting carbonates. However, the CO₂ carrying capacity of rhyolitic melts, similar to those derived from partial melting of subducting pelitic sediments, at sub-arc depths remains unconstrained. Here we measured the solubility of CO₂ in a model sediment partial melt; experiments were conducted at 1.5-3.0 GPa and 1300°C with variable water contents. The rhyolitic melt compositions were constructed with reagent grade oxides, carbonates, and hydroxides, with carbonates as the source of excess CO₂ and Al(OH)₃ as the source of variable water content, and were contained in AuPd capsules. All experiments produced glasses with bubbles, the latter being taken as the evidence of equilibrium vapor saturation at experimental P-T conditions. Micro-Fourier Transform Infrared spectroscopy was employed to measure the concentrations of CO₂ and H₂O in doubly polished, bubble-free sections of the glasses. The total CO₂ solubility, CO₂^tot. (= CO₂^mol. + CO₃^2-), of experimental melts increases with pressure and water content. For melts with H₂O of ~0.5 wt.%, CO₂^tot. values increase and CO₂^mol./CO₂^tot. values decrease with increasing pressure from 0.68 to 1.2 wt.% and 0.95 to 0.60 from 1.5 to 3.0 GPa, respectively. In contrast to the water-poor melts, the hydrous melts with ~2 and 3.5 wt.% H₂O showed the opposite CO₂^mol./CO₂^tot.-pressure trend (0.18 to 0.55 for 2 wt.% H₂O and 0.04 to 0.18 for 3.5 wt.% H₂O from 1.5 to 3.0 GPa). Total CO₂ contents of hydrous melts were also higher at a given pressure and ranged from 0.8 to 1.5 wt.% from 1.5 to 3.0 GPa for experiments with 2 wt.% H₂O and 0.9 to >1.6 wt.% from 1.5 to 3.0 GPa for experiments with 3.5 wt.% H₂O. Measurements of melt inclusions and gases determined that primary arc magmas contain ≥0.3 wt.% CO₂ and may be as high as 1.4-1.7 wt.% CO₂ [1, 2]. If 0.3 wt.% CO₂ in primary arc basalts derives from 15-30% melting [3, 4] of the mantle wedge, a mantle concentration of 450-900 ppm is required. Assuming that the pre-subduction modification mantle wedge contains 37-246 ppm CO₂ [5-9], addition of sediment melt amounting ~2-9% of the mantle and containing 1 wt.% CO₂ is sufficient to elevate the mantle budget to the required level. If the sediment partial melts are more hydrous, the flux of the melt to deliver the required CO₂ budget would be even lower. Therefore, our CO₂ solubility data suggest that hydrous sediment melt could be the agent of CO₂ delivery from slab to arc source mantle wedge. [1] Blundy et al. (2010) EPSL, 290, 289-301; [2] Wallace (2005) JVGR, 140, 217-240; [3] Grove et al. (2002) CMP, 142, 375-396; [4] Stolper and Newman (1994) EPSL, 121, 293-325; [5] Cartigny et al. (2008) EPSL, 265, 672-685; [6] Dasgupta and Hirschmann (2010) EPSL, 298, 1-13; [7] Hirschmann and Dasgupta (2009) CG, 262, 4-16; [8] Marty (2012) EPSL, 313, 56-66; [9] Saal et al. (2002) Nature, 419, 451-455.