Abstract

A difficulty in performing high-temperature (>900 °C) experiments on near-liquidus hydrous mafic melts in gas-medium cold-seal pressure vessels (CSPV) is the tendency for H2O in the fluid phase to dissociate and H2 to diffuse through capsule material, leading to progressive oxidation of sample material. Negative consequences include premature stabilization of Fe-Ti oxide phases and commensurate deviation of the liquid line of descent toward silica enrichment. Moreover, time-variance of an intensive variable equal in importance to temperature or total pressure is an unwanted feature of any experimental study. Methodologies commonly employed to mitigate the oxidation problem, not without their own drawbacks, include incorporating CH4 into the pressurizing gas, limiting run duration to 24 h, enclosing samples in Au-alloy capsules, and incorporating solid buffering assemblages to serve as indicators of fO2 excursion. Using the Co-Pd-O system as a fO2 sensor, we investigated progressive oxidation of basaltic andesite at 1010 °C and PH2O = 150 MPa. Our time-series of 12, 24, 36, 48, and 60 h run durations reveals that oxidation occurs at a very high rate (~3–4 log unit change in fO2 in 48 h). Both the variability of fO2 and magnitude of dehydration-oxidation are considered unacceptable for phase equilibria work. Incorporation of additional CH4 serves only to offset the progressive oxidation trend toward a lower absolute range in fO2. Ultimately, rapid oxidation in CSPV hinders the chemical equilibration of experimental charges. To mitigate the issue, we propose the following solution: Incorporation of a substantial mass of Ni metal powder as an O2 getter to the outer capsule successfully: (1) slows down oxidation; (2) stabilizes fO2 at the nickel-nickel oxide (NNO) buffer after ~20 h; and (3) allows compositions to approach equilibrium. Runs much longer than 48 h may require one or more steps involving quenching and re-filling the pressure system with CH4.

You do not currently have access to this article.