Some Remarks on the Interpretation of the REE-in-two-Mineral Thermobarometers

Abstract:

Distributions of REE between mantle minerals depend on temperature, pressure and mineral compositions, and can be used as thermobarometers. We have developed a REE-in-two-pyroxene thermometer, a REE-in-garnet-clinopyroxene thermobarometer, and a REE-in-plagioclase-augite thermometer for ultramafic and mafic rocks. These trace element based two-mineral thermobarometers are built on parameterized lattice strain models for mineral-melt REE partitioning that we independently calibrated using published data from mineral-melt partitioning experiments. The physical meanings of calculated temperatures and pressures can be understood in terms of the average closure temperatures and closure pressures of REE in cooling bi-mineralic systems. Because trivalent REE diffuse slower than divalent Ca-Fe-Mg in minerals, the REE-based thermobarometers may record early thermal event(s), whereas the major element based thermometers may record more recent thermal events in closed systems that experienced cooling. A number of geological and geodynamic processes can result in decreases in temperature and possibly pressure over time: mantle upwelling beneath mid-ocean ridge spreading centers or in response to lower crust delamination, exhumation, heat loss to country rock from crustal magma chambers. For shallow level intrusive or extrusive mafic rocks, temperatures derived from the REE-in-plagioclase-augite thermometer may correspond to plagioclase saturation temperature. Hence application of REE- and Ca-Fe-Mg based thermobarometers to the same samples can shed new insight into the thermal history of mafic and ultramafic rocks.

Initial applications of the REE-in-two-mineral thermobarometers to Earth and planetary samples have lead to a number of interesting observations which we will summarize in this presentation. However, care and caution must be exercised when processing and interpreting data from REE-in-two-mineral thermobarometers. Common issues include inaccuracies in REE data, deviation of LREE data from a linear regression line, rotation in the inversion diagram, and abnormally high temperature and/or low pressure. We will address these common problems through examples from actual samples and numerical simulations.