CO2 Solubility in Rhyolitic Melts as a Function of P, T, and fO2 - Implications for Carbon Flux in Subduction Zones

Understanding the balance between subduction inputs vs. arc output of carbon is critical for constraining the global carbon cycle. However, the agent of carbon transfer from slab to sub-arc mantle is not constrained [1]. Partial melt of ocean-floor sediments is thought to be a key agent of mass transfer in subduction zones, accounting for the trace element characteristics of arc magmas [2]. Yet the carbon carrying capacity of rhyolitic partial melts of sediments remains unknown at sub-arc depths. In our previous work [3], we constrained CO₂ solubility of natural rhyolite from 1.5-3.0 GPa, 1300 °C and logfO₂ at FMQ×1.0. However, the effects of T and fO₂ on CO₂ solubility remain unconstrained. In particular, for sediments with organic carbon, graphite stability is expected and the fO₂ of C-dissolution can be lower, which may affect the solubility. Thus it is critical to constrain the CO₂ solubility of sediment partial melts under graphite-saturated conditions. We determined CO₂ solubility of a model rhyolite composition, similar to partial melt composition of natural metapelite [4], at graphite saturation, using Pt/Gr capsules and a piston cylinder device. Experiments were conducted at 1.5-3.0 GPa and 1100-1400 °C. FTIR was employed to measure the concentrations of CO₂ and H₂O in doubly polished experimental glasses. Raman and SIMS were used to determine the presence of reduced carbon species and total carbon, respectively. FTIR spectra reveal that CO₂ is dissolved as both molecular CO₂ (CO₂^mol.) and carbonates (CO₃^2-). For graphite-saturated, hydrous melts with measured H₂O ~2.0 wt.%, CO₂^tot. (CO₂^mol.+CO₃^2-) values increase with increasing P from ~0.6 to 1.2 wt.% from 1.5 to 3.0 GPa at 1300 °C. These values are lower than more oxidized melts with the same water content, which were 0.85 to 1.99 wt.% CO₂ as P increased. At 3 GPa, graphite-saturated experiments from 1100 to 1300 °C yield CO₂^tot. value of 1.18-1.20 wt.%, suggesting minor effect of temperature in bulk CO₂ solubility. To meet the minimum requirement of 3000 ppm CO₂ in primary arc magma [5,6], the required sediment melt contribution is 0.18-0.28 wt.% CO₂, which is distinctly lower than the solubility limit of graphite-saturated melt. However, 1.7 wt.% CO₂ in primary arc basalts [5] exceeds the solubility limit of reduced, hydrous melts, which is in contrast to more oxidized, hydrous melts which can contribute up to 2 wt.% CO₂. We determine that ~1.7-15% of sediment melt would be required to meet 3000 ppm CO₂ in the primary arc basalt depending on the depth of melting (1.5-3.0 GPa) and the degree of mantle wedge melting (15-30%). This contribution is higher than that previously calculated for the more oxidized melts, but still may not be an unreasonable slab flux. [1] Dasgupta (2013) RiMG, 75, 183-229; [2] Plank and Langmuir (1993) Nature, 362, 739-743. [3] Duncan and Dasgupta. (in review) GCA; [4] Tsuno and Dasgupta (2011) CMP, 161, 743-763; [5] Blundy et al. (2010) EPSL, 290, 289-301; [6] Wallace (2005) JVGR, 140, 217-240.