The whole-block method and water diffusion in olivine

Details

Meeting2013 Fall Meeting
SectionMineral and Rock Physics
SessionPlastic and Transport Properties of Deep Earth Materials III Posters
IdentifierMR41A-2352
Authors Ferriss, E*, LDEO, Columbia University, Palisades, NY, USA
Plank, T A, LDEO, Columbia University, Palisades, NY, USA
Walker, D, LDEO, Columbia University, Palisades, NY, USA
Index Terms Mineral and crystal chemistry [1042]
Instruments and techniques [1094]
Experimental mineralogy and petrology [3630]
Optical, infrared, and Raman spectroscopy [3934]

Abstract

Accurate knowledge of the diffusion rates and mechanisms of water (hydrogen) in geologic materials is critical for geologic applications such as understanding the deep earth water cycle and determining ascent rates of pre-eruptive magmas. However, diffusion rates and mechanisms are often poorly known. Obtaining robust diffusion profile data for water with transmission FTIR traditionally requires cutting the sample after an experiment to isolate the central slice. Here we develop a method for interpreting diffusion profile data in three dimensions without cutting the sample. This “whole-block” method is nondestructive, which simplifies the analytical procedure and allows multiple experiments on the same sample (e.g., a time series or reversal). Whole-block data represent concentration values that are integrated through the entire sample in the direction parallel to the infrared beam ray path during the measurement, and they are determined by taking the average value of a non-path-integrated 3 dimensional model (e.g., based on error functions or infinite sums in a rectangular parallelepiped) down the ray path for a given position. The whole-block method was tested by comparing whole-block profiles with profiles cut from the center of an oriented diopside sample after a dehydration experiment [1]. Water profiles were measured in the cut slice by both SIMS and FTIR. The results of the two methods are in good agreement both with each other and with diffusion profiles calculated based on the results of the whole-block method. Interpreting whole block measurements without taking into account the integration effects through the crystal can lead to errors in calculated diffusivities and inferred mechanisms. We have used numerical simulations to demonstrate as much as a half an order of magnitude error (typically indicating diffusivities that are too fast) if whole-block data are interpreted using non-path integrated diffusion models. The largest errors are in short and/or fast directions, in which diffusion profiles are well developed. Whole block data also inevitably involve central values that are contaminated by edge concentrations integrated in the signal. This integration effect results in a plateau in the whole-block data that may complicate the interpretation of the whole-block data. For example, previous experimental work on water diffusion in olivine has identified a central plateau using whole block measurements, and this plateau is interpreted to result from a transitional state between two diffusion mechanisms [2]. However, a whole-block model also produces a reasonable fit to this data using the observed initial concentration of zero and a single-step diffusion mechanism. Thus, whole block effects need to be taken into account for the accurate determination of diffusivities and mechanisms. [1] Ferriss et al 2012 AGU; [2] Demouchy&Mackwell 2006 Phys Chem Mineral 33(5).

Cite as: Author(s) (2013), Title, Abstract MR41A-2352 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.