COH FLUIDS AT UPPER-MANTLE CONDITIONS: AN EXPERIMENTAL STUDY ON VOLATILE SPECIATION AND MINERAL SOLUBILITY IN THE MS+COH SYSTEM

COH fluids play a fundamental role in many geological processes, controlling the location of melting in subduction zones and promoting mass transfer from the subducting lithosphere to the overlying mantle wedge. The properties of COH fluids are strictly dependent on the composition of the fluid in subduction systems, i.e., the speciation of the volatile components of the fluid itself and the presence of solutes deriving from the dissolution of rock-forming minerals. In the scientific literature, the speciation of COH fluids has been generally determined through thermodynamic calculations using equations of state of simple H2O–non-polar gas systems (e.g., H2O–CO2–CH4), equations that do not consider the complexity related to dissolution processes, which are substantially unexplored in COH fluids and limited so far to aqueous fluids (Newton & Manning, 2002). The aim of this work is to experimentally investigate the speciation of COH volatile components of the fluid and the dissolution of mantle minerals in carbon-saturated COH fluids at buffered fO2 conditions. Our experimental approach relies on two different techniques: i) analysis by means of quadrupole mass spectrometer (QMS) of the COH fluid from pierced run capsules to retrieve speciation of volatile components and ii) analysis of frozen COH fluid with laser-ablation inductively coupled plasma-mass spectrometry (LA-ICP-MS) to measure the amount of solutes. Experiments were conducted at P = 1–3 GPa and T = 700–1200 °C using a rocking piston-cylinder apparatus. Mantle minerals in equilibrium with COH fluids are represented by synthetic forsterite, enstatite and natural magnesite. fO2 conditions were controlled employing the double capsule technique and nickel–nickel oxide (NNO) buffer. We also performed a series of experimental runs in the COH-only system in single or double capsules, varying the packing material that surrounds the capsule and the oxygen buffer, to evaluate the differences in COH volatile speciation determined by the choice of the experimental setup. Quantitative analyses of COH volatile speciation were retrieved by piercing the capsule in a gas-tight vessel at T = 80 °C and convoying evolved gases to a QMS through a heated line to avoid the condensation of water. Our experimental results on COH volatile speciation highlighted the importance of the experimental verification of volatile speciation, which can diverge considerably compared to the thermodynamic model (Perplex; Connolly, 1990) depending on the experimental strategies adopted. In particular, when single capsules are employed, the packing material that surrounds the capsule exerts a major control on the COH volatile speciation. Double capsule experiments provided similar COH volatile speciation compared to thermodynamic modeling for what concerned the COH-only system. However, the addition of mantle minerals in the experimental charge at the same P–T–fH2 conditions determines a shift in the COH fluid composition toward more CO2-rich terms. At P = 1 GPa, data show an increase in CO2 of + 11 mol% at T = 800 °C and of + 26 mol% at T = 900 °C in the COH fluid in equilibrium with forsterite + enstatite compared to a pure COH fluid. To evaluate if this shift could be determined by interactions of the COH fluid with solid phases, we retrieved the solubility of mantle minerals in COH fluids through the cryogenic LA-ICP-MS technique described by Kessel et al. (2004). With this method the COH fluid is trapped into a diamond layer, the aqueous part of the COH fluid is frozen to avoid any precipitation of solutes and is analyzed through LA-ICP-MS. Experimental results on mantle minerals solubility in COH fluids suggest that the amount of dissolved material in COH fluids is similar compared to mantle mineral solubility in H2O-only fluid and ranges from the 2 wt.%, expressed as MgO + SiO2, at P = 1 GPa and T = 800 °C, to 12 wt.% at P = 2 GPa and T = 1100 °C for the forsterite + enstatite assemblage. The formation of dissolved species containing carbon, such as CO32- and Mg(HCO3)+ lead to an increase in the amount of carbon in the fluid, but not in CO2 species. In order to get the increase of CO2 that we observed in experiments analyzed through quadrupole mass spectrometry, we suggest a series of possible dissolution reactions involving Mg-solutes, which could lead to increase the amount of CO2 in the fluid. As a consequence, the quantity of CO2 infiltrating into the mantle-wedge could be remarkably high, compared to the COH fluid composition predicted by thermodynamic modeling.