EXPERIMENTAL DETERMINATION OF ONE-ATMOSPHERE PHASE RELATIONS FOR DACITIC TO RHYOLITIC MELTS
All experiments were run at one-atmosphere in Deltec furnaces. Oxygen fugacity was held between the NNO+1 and NNO+2 buffers using mixtures of CO2 and H2 gas. Experiments were conducted over a temperature range of 1000 to 1250°C using three initial melt compositions with silica contents ranging from 67.2 to 78.1 wt%. We believe that these runs achieved equilibrium as neither the phase compositions nor the total sample crystallinity changed appreciably in runs longer than 8 days, despite apparent textural disequilibrium. Additionally, melting experiments duplicated the results of crystallization experiments (i.e. the experiments were reversible).
All samples had liquidus temperatures between 1250 and 1200°C, with plagioclase the liquidus phase. The anorthite content of the plagioclase increased with increasing temperature, and at any given temperature, the composition of the plagioclase was independent of bulk composition. When associated glass compositions are projected into the Quartz-Albite-Orthoclase system they define a new one-atmosphere quartz-feldspar cotectic 5-10 less quartz normative than previous estimates (Schairer, 1950; Tuttle and Bowen, 1958). Glass compositions from each sample plot along this cotectic between 1100 and 1000°C, consistent with the plagioclase-quartz co-crystallization textures found in runs at these temperatures. This cotectic constrains glass compositions to a maximum silica content of 76±1 wt% SiO2. Reported glass compositions in excess of 77 wt% SiO2 in volcanic samples suggest non-equilibrium crystallization, perhaps a consequence of large melt undercoolings.