Effects of U-Th-rich grain boundary phases on apatite helium ages

Single-grain apatite (U-Th-Sm)/He ages (He ages) from non-detrital samples sometimes show larger dispersions than can be explained by known sources of age dispersion, such as grain size, radiation damage, parent zonation, fracturing, or intracrystalline inclusions. We present observations and model age bias effects of an additional source of apatite He age dispersion observed in some cases: U-Th-rich grain boundary phases (GBPs) precipitated on apatite crystal surfaces. Apatite grains from several samples with dispersed apatite He ages are coated or partially coated by reddish-orange GBPs rich in Fe, U, and Th. These GBPs are heterogeneous in thickness, grain coverage, and composition, and have effective U concentrations ([eU]) up to similar to 1000 ppm. These phases may have large effects on the bulk [eU] and He-4 compositions of apatite grains and can produce significant He age biases. The direction and magnitude of this age bias is primarily a function of four factors: 1) the host grain's size, 2) the GBPs' thickness and [eU] relative to the host grain, 3) the timing of GBP formation relative to the cooling age of the host grain, and 4) whether the GBPs are preserved and analyzed with the grain during He dating. Some of the most severe effects occur when GBPs formed before or near the time of the apatite cooling age but were lost just prior to analysis, for example during mineral separation. In this case, GBPs of commonly observed thicknesses (1-10 mu m) with [eU] 2-10 times that of the host apatite grain implant enough He-4 into typical-sized apatite grains to produce hundreds of percent positive age biases. In contrast, when U-Th-rich GBPs are preserved and analyzed with apatite grains, the resulting He ages can be negatively biased. Though the heterogeneity, variable preservation, and ambiguous formation ages of GBPs preclude a general, quantitative, and practical solution to this problem, our model demonstrates that observed age dispersions in some samples are consistent with the effects of U-Th-rich grain boundary phases and illustrates the conditions under which, and by how much, they will bias He ages. (C) 2014 Elsevier B.V. All rights reserved.