Crystal surface integrity and diffusion measurements on Earth and planetary materials

Abstract

Characterization of diffusion behavior in minerals is key to providing quantitative constraints on the ages and thermal histories of Earth and planetary materials. Laboratory experiments are a vital source of the needed diffusion measurements, but these can pose challenges because the length scales of diffusion achievable in a laboratory time are commonly less than 1 μm. An effective strategy for dealing with this challenge is to conduct experiments involving inward diffusion of the element of interest from a surface source, followed by quantification of the resulting diffusive-uptake profile using a high-resolution depth-profiling technique such as Rutherford backscattering spectroscopy (RBS), nuclear reaction analysis (NRA), or ion microprobe (SIMS). The value of data from such experiments is crucially dependent on the assumption that diffusion in the near-surface of the sample is representative of diffusion in the bulk material. Historical arguments suggest that the very process of preparing a polished surface for diffusion studies introduces defects—in the form of dislocations and cracks—in the outermost micrometer of the sample that make this region fundamentally different from the bulk crystal in terms of its diffusion properties. Extensive indirect evidence suggests that, in fact, the near-surface region of carefully prepared samples is no different from the bulk crystal in terms of its diffusion properties. A direct confirmation of this conclusion is nevertheless clearly important. Here we use transmission electron microscopy to confirm that the near-surface regions of olivine, quartz and feldspar crystals prepared using careful polishing protocols contain no features that could plausibly affect diffusion. This finding does not preclude damage to the mineral structure from other techniques used in diffusion studies (e.g., ion implantation), but even in this case the role of possible structural damage can be objectively assessed and controlled. While all evidence points to the reliability of diffusivities obtained from in-diffusion experiments, we do not recommend experiments of this type using a powder source as a means of obtaining diffusant solubility or partitioning information for the mineral of interest.