Diffusion of helium in olivine at 1 atm and 2.7 GPa

Diffusion of helium has been characterized in natural Fe-bearing olivine (Fo90) and pure synthetic Mg2SiO4. Polished slabs of olivine, oriented normal to a, b or c-axes, were implanted with 3He, either at 100 keV at a dose of 5 × 1015/cm2 or at 3.0 MeV at a dose of 1 × 1016/cm2. A set of diffusion experiments on 3He implanted olivine were run in Pt capsules in 1-atm furnaces. In addition to the 1-atm experiments, experiments on implanted samples were also run at higher pressures (2.6–2.7 GPa) to assess the potential effects of pressure on He diffusion and the applicability of the diffusion data in describing He transport in the mantle. The high-pressure experiments were conducted in a piston-cylinder apparatus using a special “soft” pressure cell, with the implanted forsterite samples directly surrounded by AgCl. 3He distributions following all experiments were measured with nuclear reaction analysis using the reaction 3He(d,p)4He. For diffusion in forsterite we obtain the following Arrhenius relations: D‖a=3.73×10-8exp(-135±5kJmol-1/RT)m2s-1,D‖b=1.33×10-8exp(-137±13kJmol-1/RT)m2s-1,D‖c=3.61×10-6exp(-159±4kJmol-1/RT)m2s-1.The data for diffusion parallel to the c-axis define a single Arrhenius line spanning more than 7 orders of magnitude in values of diffusion coefficients, and a 700 °C range in temperature (250–950 °C). Diffusion parallel to the a-axis appears slightly slower, and diffusion parallel to the b-axis is slower than diffusion parallel to the a-axis by about two-thirds of a log unit. This anisotropy is broadly consistent with observations for diffusion of Ni and Fe–Mg in olivine. Diffusion in Fe-bearing natural olivine (transport parallel to b) agrees within uncertainty with findings for He diffusion in the pure Mg end-member olivine. The higher-pressure experiments yield diffusivities in agreement with those from the 1-atm experiments, indicating that the results reported here can be reasonably applied to modeling He transport in the upper mantle. The insensitivity of He diffusion to pressure over the investigated range of conditions suggests that compression of the mineral lattice is not sufficient to significantly influence migration of the relatively small helium atoms, which likely diffuse via crystal interstices.The He diffusivities measured in this work are generally consistent with results from the study of Futagami et al. (1993), who measured He diffusion in natural olivine by outgassing 4He implanted samples, and with the diffusivities measured by bulk-release of He by Shuster et al. (2004), but are about 2 orders of magnitude slower than the recent findings of Tolstikhin et al., 2010 and Blard et al., 2008. An up-temperature extrapolation of our data also show reasonable agreement with the higher-temperature measurements of Hart (1984); the coincident diffusivities from these studies under the high-temperature conditions used by Hart et al. (2008) in their He diffusion/production modeling supports the applicability of their modeling results.