Overstepping the garnet isograd: a comparison of QuiG barometry and thermodynamic modeling

Abstract

The consequences of overstepping the garnet isograd reaction have been investigated by comparing the composition of garnet formed at overstepped P–T conditions (the overstep or “OS” model) with the P–T conditions that would be inferred by assuming garnet nucleated in equilibrium with the matrix assemblage at the isograd (the equilibrium or “EQ” model). The garnet nucleus composition formed at overstepped conditions is calculated as the composition that produces the maximum decrease in Gibbs free energy from the equilibrated, garnet-absent, matrix assemblage for the bulk composition under study. Isopleths were then calculated for this garnet nucleus composition assuming equilibrium with the matrix assemblage (the EQ model). Comparison of the actual P–T conditions of nucleation (the OS model) with those inferred from the EQ model reveals considerable discrepancy between the two. In general, the inferred garnet nucleation P–T conditions (the EQ model) are at a lower temperature and higher or lower pressure (depending on the coexisting calcic phase(s)) than the actual (OS model) nucleation conditions. Moreover, the degree of discrepancy increases with the degree of overstepping. Independent estimates of the pressure of nucleation of garnet were made using the Raman shift of quartz inclusions in garnet (quartz-in-garnet or QuiG barometry). To test the validity of this method, an experimental synthesis of garnet containing quartz inclusions was made at 800 °C, 20 kbar, and the measured Raman shift reproduced the synthesis conditions to within 120 bars. Raman band shifts from three natural samples were then used to calculate an isochore along which garnet was presumed to have nucleated. Model calculations were made at several temperatures along this isochore (the OS model), and these P–T conditions were compared to those computed assuming equilibrium nucleation (the EQ model) to estimate the degree of overstepping displayed by these samples. A sample from the garnet isograd in eastern Vermont is consistent with overstepping of around 10 degrees and 0.6 kbar (affinities of around 2 kJ/mole garnet). A sample from the staurolite–kyanite zone in the same terrane requires overstepping of around 50 °C and 2–5 kbar (affinities of around 10–18 kJ/mole garnet). A similar amount of overstepping was inferred for a blueschist sample from Sifnos, Greece. These results indicate that overstepping of garnet nucleation reactions may be common and pronounced in regionally metamorphosed terranes, and that the P–T conditions and paths inferred from garnet zoning studies may be egregiously in error.