Rare earth element diffusion in apatite

Diffusion of rare earth elements (REEs) in natural and synthetic fluorapatite has been characterized under anhydrous conditions. Three types of experiments were run. In the first set of experiments, Sm was introduced into the apatite by means of ion implantation, with diffusivities extracted through measurement of the "relaxation" of the implanted profile after diffusion anneals. The second group consisted of "in diffusion" experiments, in which apatite was immersed in reservoirs of synthetic REE apatite analogs of various compositions. The final set of experiments was "out-diffusion" experiments run on synthetic Nd-doped apatite immersed in a reservoir of synthetic (undoped) fluorapatite. REE depth profiles in all cases were measured with Rutherford Backscattering Spectrometry. Diffusion rates for the REE vary significantly among these sets of experiments. For the ion-implantation experiments, the following Arrhenius relation was obtained for Sm, over the temperature range 750 degreesC to 1100 degreesC: D-imp = 6.3 x 10(-7) exp(-298 +/- 17 kJ/mol/RT) m(2)/s Diffusion of a series of REE, from light to heavy, was investigated in the "in-diffusion" experiments. Over the temperature range 800 degreesC to 1250 degreesC, the following Arrhenius relations are obtained for La, Nd, Dy, and Yb, for in-diffusion experiments using REE silicate oxyapatite sources: D-La = 2.6 X 10(-7) exp(-324 +/- 9 kJ/mol/RT) m(2)/s D-Nd = 2.4 X 10(-6) exp(-348 +/- 13 kJ/mol/RT) m(2)/s D-Dy = 9.7 X 10(-7) exp(-340 +/- 11 kJ/mol/RT) m(2)/s D-Yb = 1.3 X 10(-8) exp(-292 +/- 23 kJ/mol/RT) m(2)/s Diffusivities of the REE in these "in-diffusion" experiments are all quite similar, suggesting little difference in diffusion rates in apatite with increasing ionic radii of the REEs. The "out-diffusion" experiments on the Nd-doped synthetic apatite, over the temperature range 950 degreesC to 1400 degreesC, yield the Arrhenius law. D-out = 9.3 X 10(-6) exp(-392 +/- 31 kJ/mol/RT) m(2)/s The differences in REE diffusion among these three sets of experiments (i.e., ion implantation, in-diffusion, and out-diffusion) may be attributable to the differences in substitutional processes facilitating REE exchange. The fastest diffusion, found in the ion-implantation experiments, is likely largely governed by simple light REE+3 <-> REE+3 exchange, with no charge compensating species necessary. REE transport in the indiffusion experiments requires movement of an additional charge-compensating species, either through the substitutions REE+3 + Si+4 <-> Ca+2 + P+5 or REE+3 <-> Na+1 <-> 2 Ca+2, and thus proceeds more slowly than simple REE exchange. Slowest of all is Nd out-diffusion from the synthetic Nd-doped apatite, for which neither charge compensating species are present nor REEs available in the surrounding reservoir to facilitate Nd exchange. This observed dependence of REE diffusion rates on the exchange process involved has important geochemical implications. These findings indicate that REE isotope and chemical signatures can become decoupled in apatite, with light REE isotope exchange proceeding much more rapidly than REE chemical diffusion altering total REE concentrations. Under temperatures typical of thermal events, REE zoning (involving differences in REE concentration across zones) of a given dimension might persist over time periods two orders of magnitude greater than those under which zoning of REE isotopes (without significant changes in total REE) on similar scale is preserved. Copyright (C) 2000 Elsevier Science Ltd.