Constraints on the loss of matrix-sited helium during vacuum crushing of mafic phenocrysts

Abstract

Vacuum crushing is an efficient technique to selectively release the mantle-derived helium component trapped within olivine and pyroxene phenocrysts. However, contrary to previous assumptions, recent studies have shown that this method may liberate significant matrix-sited cosmogenic 3He ( 3He c) or radiogenic 4He ( 4He ∗). Because this loss may bias both the determination of magmatic 3He/ 4He ratios and the accuracy of 3He c measurements, it is essential to understand what mechanism is responsible and under what conditions matrix helium loss is manifest. To address this question, olivines and pyroxenes with various amounts of matrix-sited 3He (from 10 7 to 10 11 at. g −1) were crushed in air or in vacuum using several crushing devices. Sample temperature was controlled during each crushing experiment, and ranged from 25 to 325 °C. The resulting powders were then sieved to obtain several homogeneous grain fractions ranging between <10 and >300 μm. The 3He c concentrations measured in each fraction clearly show that significant 3He c loss (>20%) affects only the finest fraction (<10 μm) and, importantly, only under hot conditions (here T ⩾300 °C). Even the smallest fractions (<10 μm) quantitatively retain matrix-sited 3He c when crushed under cold conditions ( T ⩽25 °C), regardless of the duration and energy of crushing. These results invalidate the model previously proposed by (Yokochi R., Marty B., Pik R. and Burnard P. (2005) High 3He/ 4He ratios in peridotite xenoliths from SW Japan revisited: evidence for cosmogenic 3He released by vacuum crushing. Geochem. Geophys. Geosyst. 6, doi:10.1029/2004GC000836) that involved spallation tracks and implied that the magnitude of loss was mainly controlled by the grain size. Moreover, new diffusion experiments were carried out to constrain the diffusivity of matrix-sited helium in crushed olivines. When used to model diffusive 3He c loss as a function of grain size during crushing, these new data predict the observed release fairly well. Therefore, we conclude that temperature-enhanced volume diffusion is one of the main mechanisms controlling the release of 3He c during crushing. For future applications, special attention should thus be paid to control the grain size, the crushing duration, and the temperature of the sample.