Biotite dehydration, partial melting, and fluid composition; experiments in the system KAlO 2 -FeO-MgO-SiO 2 -H 2 O-CO 2

Biotite solid-solutions are a significant H ₂ O reservoir in the lithosphere, and the assemblage Bt+Opx+Kfs+Qtz is commonly used to estimate a _H2O in high-grade metamorphic and magmatic rocks. Here we report experimental constraints on subsolidus mineral equilibria involving biotite and orthopyroxene in the system KAlO ₂ -MgO-FeO-SiO ₂ -H ₂ O-CO ₂ . Our experiments address the question of stability of biotite of a given X _Fe in the assemblage Bt+Qtz+ or -Sa, or the stability of the assemblage Opx+Sa+ or -Qtz. Clemens (1993) and Clemens et al. (1997) concluded that CO ₂ has no effect other than to lower a _H2O and thereby raise the solidus T. Our data at X ^Fl_H2O < extend these conclusions to encompass Fe-bearing systems more similar to natural rocks. From a comparison of experimental data and calculated isopleths of biotite composition in the divariant assemblage Bt+Opx+Kfs-Qtz+fluid, it appears that phlogopite-annite solid-solutions must be significantly non-ideal (at T<800 degrees C) or that enstatite-ferrosilite solid-solutions must have negative values for their Margules-type parameters. Ignoring these factors would result in any calculated a _H2O values being too low. Although various models allow us to estimate X _H2O in H ₂ O-CO ₂ fluids, we are still unable to use biotite equilibria to estimate a _H2O accurately during high-grade metamorphism and magma crystallization. We also consider qualitatively the effects Fe-Mg biotite solid-solution on partial melting equilibria in fluid-poor (rock-dominated) systems in which hydration-dehydration reactions control the fluid composition.