No abstract

Organic matter, showing variable degrees of crystallinity and thus of graphitization, is an important source of carbon in subducted sediments, as demonstrated by the isotopic signatures of deep and ultra-deep diamonds and volcanic emissions in arc settings. In this experimental study, we investigated the dissolution of sp 2 hybridized carbon in aqueous fluids at 1 and 3 GPa, and 800°C, taking as end-members i) crystalline synthetic graphite and ii) X-ray amorphous glass-like carbon. We chose glass-like carbon as an analogue of natural "disordered" graphitic carbon derived from organic matter, because unlike other forms of poorly ordered carbon it does not undergo any structural modification at the investigated experimental conditions, allowing approach to thermodynamic equilibrium. Textural observations, Raman spectroscopy, synchrotron X-ray diffraction and dissolution susceptibility of char produced by thermal decomposition of glucose (representative of non-transformed organic matter) at the same experimental conditions support this assumption. The redox state of the experiments was buffered at ∆FMQ ≈-0.5 using double capsules and either fayalite-magnetite-quartz (FMQ) or nickel-nickel oxide (NNO) buffers. At the investigated P-T-fO 2 conditions, the dominant aqueous dissolution product is carbon dioxide, formed by oxidation of solid carbon. At 1 GPa and 800°C, oxidative dissolution of glass-like carbon produces 16-19 mol% more carbon dioxide than crystalline graphite. In contrast, fluids interacting with glass-like carbon at the higher pressure of 3 GPa show only a limited increase in CO 2 (fH 2 NNO) or even a lower CO 2 content (fH 2 FMQ) with respect to fluids interacting with crystalline graphite. The measured fluid compositions allowed retrieving the difference in Gibbs free energy (ΔG) between glass-like carbon and graphite, which is +1.7(1) kJ/mol at 1 GPa-800°C and +0.51(1) kJ/mol (fH 2 NNO) at 3 GPa-800°C. Thermodynamic modeling suggests that the decline in dissolution susceptibility at high pressure is related to the higher compressibility of glass-like carbon with respect to crystalline graphite, resulting in G-P curves crossing at about 3.4 GPa at 800°C, close to the graphite-diamond transition. The new experimental data suggest that, in the presence of aqueous fluids that flush subducted sediments, the removal of poorly crystalline "disordered" graphitic carbon is more efficient than that of crystalline graphite especially at shallow levels of subduction zones, where the difference in free energy is higher and the availability of poorly organized metastable carbonaceous matter and of aqueous fluids produced by devolatilization of the downgoing slab is maximized. At depths greater than 110 km, the small differences in ΔG imply that there is minimal energetic drive for transforming "disordered" graphitic carbon to ordered graphite; "disordered" graphitic carbon could even be energetically slightly favored in a narrow P interval.

Estimates of dissolved CO2 in subduction-zone fluids are based on thermodynamic models, relying on a very sparse experimental data base. Here, we present experimental data at 1–3 GPa, 800 °C, and ∆FMQ ≈ −0.5 for the volatiles and solute contents of graphite-saturated fluids in the systems COH, SiO2–COH ( + quartz/coesite) and MgO–SiO2–COH ( + forsterite and enstatite). The CO2 content of fluids interacting with silicates exceeds the amounts measured in the pure COH system by up to 30 mol%, as a consequence of a decrease in water activity probably associated with the formation of organic complexes containing Si–O–C and Si–O–Mg bonds. The interaction of deep aqueous fluids with silicates is a novel mechanism for controlling the composition of subduction COH fluids, promoting the deep CO2 transfer from the slab–mantle interface to the overlying mantle wedge, in particular where fluids are stable over melts.

The quantitative assessment of COH fluids is crucial in modeling geological processes. The composition of fluids, and in particular their H2O/CO2 ratio, can influence the melting temperatures, the location of hydration or carbonation reactions, and the solute transport capability in several rock systems. In the scientific literature, COH fluids speciation has been generally assumed on the basis of thermodynamic calculations using equations of state of simple H2O–nonpolar gas systems (e.g., H2O–CO2–CH4). Only few authors dealt with the experimental determination of high‐pressure COH fluid species at different conditions, using diverse experimental and analytical approaches (e.g., piston cylinder + capsule piercing + gas chromatography/mass spectrometry; cold seal + silica glass capsules + Raman). In this contribution, we present a new methodology for the synthesis and the analysis of COH fluids in experimental capsules, which allows the quantitative determination of volatiles in the fluid by means of a capsule‐piercing device connected to a quadrupole mass spectrometer. COH fluids are synthesized starting from oxalic acid dihydrate at P = amb and T = 250°C in single capsules heated in a furnace, and at P = 1 GPa and T = 800°C using a piston‐cylinder apparatus and the double‐capsule technique to control the redox conditions employing the rhenium–rhenium oxide oxygen buffer. A quantitative analysis of H2O, CO2, CH4, CO, H2, O2, and N2 along with associated statistical errors is obtained by linear regression of the m/z data of the sample and of standard gas mixtures of known composition. The estimated uncertainties are typically <1% for H2O and CO2, and <5% for CO. Our results suggest that the COH fluid speciation is preserved during and after quench, as the experimental data closely mimic the thermodynamic model both in terms of bulk composition and fluid speciation.