The influence of Fe3+ on garnet–orthopyroxene and garnet–olivine geothermometers

Abstract

Applying Fe2+–Mg exchange geothermometers to natural samples may lead to incorrect temperature estimates if significant Fe3+ is present. In order to quantify this effect, high-pressure experiments were carried out in a belt apparatus in a natural system close to CFMAS at 5 GPa and 1,100–1,400 °C. The oxygen fugacity in the experiments was at or below the Re–ReO2 buffer. This is at significantly more oxidized conditions than in previous experiments, and, as consequence, higher Fe3+/Fe2+ ratios were generated. The Fe3+ content of garnet in the experiments was quantified by electron microprobe using the flank method. Making the usual assumption that Fetotal = Fe2+, the two-pyroxene thermometer of Brey and Kohler (J Pet 31:1353–1378, 1990) reproduced the experimental temperature to ±35 °C and the garnet–clinopyroxene Fe2+–Mg exchange thermometer of Krogh (Contrib Miner Pet 99:44–48, 1988) overestimated the temperatures on average by only 25 °C. On the other hand, application of the garnet–olivine (O’Neill and Wood in Contrib Miner Pet 70:59–70, 1979) and garnet–orthopyroxene (Harley in Contrib Miner Pet 86:359–373, 1984) exchange geothermometers yielded an underestimation in calculated temperatures of >200 °C. However, making explicit accounting for Fe3+ in garnet (i.e. using only measured Fe2+) leads to a vast improvement in the agreement between calculated and experimental temperatures, generally to within ±70 °C for the garnet–orthopyroxene geothermometer as well as noticeable improvement of calculated temperatures for the garnet–olivine geothermometer. Our results demonstrate that the two-pyroxene and garnet–clinopyroxene thermometers are rather insensitive to the presence of Fe3+ whilst direct accounting of Fe3+ in garnet is essential when applying the garnet–olivine and garnet–orthopyroxene thermometers.