CORDIERITE-GARNET-SILLIMANITE-QUARTZ EQUILIBRIUM .1. NEW EXPERIMENTAL CALIBRATION IN THE SYSTEM FEO-AL2O3-SIO2-H2O AND CERTAIN P-T-XH20 RELATIONS

The equilibrium in which hydrous Fe-cordierite breaks down to almandine, sillimanite, quartz, and water was previously experimentally determined by Richardson (1968) and Holdaway and Lee (1977) using QMF buffer and by Weisbrod (1973) using QIF buffer. All these studies yielded similar results - a negative dP/dT slope for the equilibrium curve. However, based on theoretical arguments, Martignole and Sisi (1981), and based on Fe-Mg partitioning experiments on coexisting cordierite and garnet in equilibrium with sillimanite and quartz, Aranovich and Podlesskii (1983) suggested that this equilibrium curve has a positive dP/dT slope and its position depends on the water content of the equilibrium cordierite. We have redetermined this equilibrium using a much improved technique of detecting reaction direction, and cordierite starting material that contained virtually no hercynite. Hercynite was present as a contaminant in the cordierites of previous experimental studies and possibly reacted with quartz during the experimental runs to expand the apparent stability field of Fe-cordierite. We synthesized Fe-cordierite from reagent grade oxides at 710-degrees-C and 2 kbar (using QMF buffer) with two intermediate stages of grinding and mixing. The cordierite has a unit cell volume of 1574.60 angstrom3 (molar volume = 23.706 J/bar) and no Fe3+ as indicated by X-ray diffraction and room temperature Mossbauer studies respectively. Reaction direction was concluded by noting greater-than-or-equal-to 20% change of the ratios of intensities of two key X-ray diffraction peaks of cordierite and almandine. Our results show that the four-phase equilibrium curve passes through the points 2.1 kbar, 650-degrees-C and 2.5 kbar, 750-degrees-C. This disagrees with all previous experimental studies. H2O in the Fe-cordierite, equilibrated at 2.2 kbar and 700-degrees-C and determined by H-extraction line in the stable isotope laboratory, is 1.13 wt% (n = 0.41 moles). H2O content of pure Mg-cordierite equilibrated under identical conditions and determined by thermogravimetric analysis is 1.22 wt% (n = 0.40). Similar determinations on Fe-cordierite and Mg-cordierite equilibrated at 2.0 kbar and 650-degrees-C show 1.27 wt% (n = 0.46) and 1.47 wt% (n = 0.48) of H2O respectively. Thus, H2O content appears to be independent of Fe/Mg ratio in cordierite, a conclusion which supports previous experimental determinations. The experimentally determined equilibrium curve represents conditions of P(H2O) = P(total). From this we calculated the anhydrous curve representing equilibrium under conditions of X(H2O)V = 0.0. A family of calculated equilibrium curves of constant n(H2O)Cord cut the experimentally determined curve at a very small angle indicating a slight variation in n(H2O)Cord in cordierite in equilibrium with almandine, sillimanite, and quartz under the conditions of constant X(H2O)V. Another set of calculated equilibrium curves, each representing constant a(H2O)V demonstrate that the slopes of the curves vary with X(H2O)V, and are all positive in the full range of 0.0 less-than-or-equal-to X(H2O)V less-than-or-equal-to 1.0.