PUTTING THE BRAKES ON GEOSPEEDOMETRY – WHY SIMPLE DIFFUSION PROFILES YIELD MAXIMUM COOLING RATES

Commonly, geospeedometry seeks to model diffusion profiles on crystal edges in terms of simple chemical or isotopic exchange with the matrix. This approach assumes that matrix kinetics, such as grain boundary transport, do not restrict loss or gain of material from the crystal rim. However, trace element depth profiles collected for Zr, Hf, Ta, Nb, U and Sn in amphibolite-facies rutile grains of the Catalina Schist, show significant variability within a single rock. For example, Zr profiles among different grains have significantly different slopes, and grains with similar Zr profiles can have vastly different Nb profiles. These differences imply that textural and kinetic idiosyncrasies of the matrix affect the ability of specific crystals to accept or release trace elements. These idiosyncrasies should result in flatter profiles than the true cooling history would predict. For example, if grain boundary Zr transport or zircon formation limits the loss of Zr from a rutile crystal, Zr may “back-up” at the crystal face, leading to a flatter diffusion profile. Inversion of a flatter profile using standard models will yield a spuriously fast cooling rate, reflecting less loss of Zr than would occur if the matrix could accept any amount of outward diffusing Zr. In the extreme case, an inclusion isolated in a diffusionally inert host would exhibit no diffusion profile, which would theoretically imply quenching at an infinite cooling rate. We recognize diffusive limitations for inclusions and may be able to improve models for systems with known matrix sources and sinks (e.g., garnet-biotite Fe-Mg, rutile-zircon-quartz Zr, etc.). Other systems, however, are poorly constrained – where exactly do Pb, Ta, Nb, W, Sn, etc. reside in the matrix? Can matrix minerals and grain boundary diffusion keep pace with the crystal’s attempt to reequilibrate? Overall, crystals exhibiting the greatest diffusive loss or gain of an element will conform most closely with the assumptions of theoretical models, and provide the best (but still maximum) estimate of cooling rate. Averaging profiles from multiple grains will increase calculated cooling rates. Because different elements have different mobilities and reactivities in the matrix, each has the potential to yield a different estimate of cooling history.