Subduction Cycling of C-O-H Volatiles Constrained by Near-Solidus Phase Relations of Water-undersaturated, Carbonated Pelite at 3 GPa

While the petrology of water release from subducting slab to magmatic arcs is reasonably well-understood, similar cycle of carbon is a poorly understood process. Subducting sediments globally contribute to the geochemical characteristics of arc magmas1, but sedimentary carbonates are predicted to survive shallow dehydration2. Previous experiments on Fe-calcareous clay of Lesser Antilles3 and dry, carbonated pelite composition relevant for Central American trench4 suggested that melting also may not be efficient in releasing CO2 to mantle wedge. To add to the discussion, here we present results from new high pressure experiments on water-undersaturated, CO2-saturated pelitic compositions. Piston cylinder experiments using Au capsules were carried out at 3.0 GPa and 800-1150 °C, using two starting compositions, both containing 5 wt.% CO2 and 1 wt.% H2O, The SiO2 contents of the two bulk compositions were varied (HPLC2: ~60 wt.% and HPLC3: ~51 wt.%) to capture variable loss of silica during shallow dehydration. The subsolidus phase assemblages for both the compositions are identical and at 800 °C comprise cpx + garnet + coesite + Fe-Mg-calcite + phengite + rutile. The solidi are located between 800 and 850 °C and phengite disappears and K-spar appears at 900 °C. Calcitic carbonate disappears at 950-1000 °C for both the compositions. Melts in all the experiments at ≥900 °C are fluid-saturated and are rhyolitic in composition. Silicate melt compositions show a decrease in SiO2 and an increase of TiO2, FeO, and CaO with increasing temperature. Al2O3, MgO, Na2O, and K2O in the melt varied little with temperature. The temperatures of solidus, phengite-out, and carbonate breakdown are respectively ~50 °C, ~50 °C, and ~75 °C lower compared to those determined for a calcareous clay3 at similar pressures. Our study suggests that calcitic carbonate and phengite are likely cycled into the deep Earth via relatively cool subduction (<800 °C slab-surface temperature at 3 GPa), whereas dehydration melting of phengite might be encountered for warm subduction zones (800-950 °C at 3 GPa5). Melting of downgoing sediments in warm subduction zones, although is efficient in releasing water to mantle wedge, unlikely strips-off all the carbon as calcitic carbonate remains stable in the residue 100-200 °C above the solidus. Main pulse of slab-CO2 at sub-arc depths may come from sediment diapirs that, owing to lower densities with respect to peridotite, upwell and undergo super-adiabatic melting in the mantle wedge. 1Plank, T. and Langmuir, C. 1993, Nature 362, 739-743. 2Kerrick, D. M. and Connolly, J. A. D. 2001, Nature 411, 293-296. 3Thomsen, T.B. and Schmidt, M.W., EPSL 267, 17-31. 4Tsuno, K. and Dasgupta, D. 2010, Contrib Mineral Petrol doi: 10.1007/s00410-010-0560-9 5Syracuse, E. M., van Keken, P. E. and Abers, G. A. 2010, PEPI doi:10.1016/j.pepi.2010.02.004