Trace element partitioning and Lu–Hf isotope systematics in spinel peridotites from the Rio Grande Rift and Colorado Plateau: Towards improved age assessment of clinopyroxene Lu/Hf–176Hf/177Hf in SCLM peridotite

Abstract

Our study of Colorado Plateau and Rio Grande rift spinel peridotite xenoliths determined bulk Lu–Hf budgets and cpx, opx, and whole rock 176Hf/177Hf to evaluate the potential age significance of Lu/Hf–Hf isotope correlations in sub-continental lithospheric mantle (SCLM) derived xenoliths. The samples have fertilities (spinel Cr# range from 0.1 to 0.5), and equilibration temperatures (950–1050°C) that overlap with those of spinel peridotites commonly used in Lu–Hf dating studies. The Lu/Hf of clinopyroxene (cpx) and its associated whole rock (WR) are similar in fertile samples, but cpx has a lower Lu/Hf than that of the WR in refractory samples. Orthopyroxene (opx) has systematically higher Lu/Hf than cpx, but in many samples, the opx and cpx have identical 176Hf/177Hf, suggesting that the Hf isotope composition of these minerals is supported by the Lu/Hf of the WR rather than that of the individual minerals. Many opx fractions and whole rocks have less radiogenic 176Hf/177Hf than their corresponding cpx (but similar to that of the host magma), reflecting contamination. Because the 176Hf/177Hf of cpx is supported by the WR Lu/Hf, it is necessary to use the latter rather than the cpx Lu/Hf for isochron dating. However, because the actual WR Lu/Hf values are susceptible to secondary overprinting, we recommend using model WR Lu/Hf values (determined from mineral Lu and Hf concentrations and modal abundances) to construct external cpx isochrons. Model WR and measured cpx 176Hf/177Hf are correlated (r2=0.98) in a suite of eastern Colorado Plateau xenoliths and yield an apparent age of 1.2Ga (versus 1.5Ga when – incorrectly – using cpx Lu/Hf). Despite the good correlation between Lu/Hf and 176Hf/177Hf, we do not find correlations between indicators of melt depletion and Lu/Hf or 176Hf/177Hf. This suggests that Lu/Hf–176Hf/177Hf correlations are not the result of melt depletion. Instead, we propose that Lu/Hf–176Hf/177Hf correlations are the result of mixing between depleted and enriched components (e.g. metasomatism of depleted lithospheric mantle). We emphasize that additional data should be utilized (e.g. whole rock Lu–Hf budget, degree of metasomatism, opx 176Hf/177Hf) when interpreting apparent Lu–Hf isochrons for age significance.