Multicomponent diffusion in silicate melts: SiO2–TiO2–Al2O3–MgO–CaO–Na2O–K2O System

Abstract

Nine successful diffusion couple experiments were carried out in a 7-component haplobasaltic silicate melt SiO 2 -TiO 2 -Al 2 O 3 -MgO-CaO-Na 2 O-K 2 O system to study multicomponent diffusion at ~1500°C and 1 GPa, typically with compositional gradients in only two components in each experiment. At least two concentration traverses were measured for each experiment. Effective binary diffusion coefficients (EBDC) for monotonic profiles were obtained by an error function fit, and the EBDC of a given component is dependent on its counter diffusing component, especially for SiO 2 . The EBDC's of SiO 2 vary from 15.7 µm 2 /s when diffusing against Al 2 O 3 , to 102.9 µm 2 /s when diffusing against K 2 O. Furthermore, the multicomponent diffusion matrix was obtained by simultaneously fitting profiles of all oxides in all experiments. Most features in the diffusion profiles, for example uphill diffusion, are captured well by this 6×6 diffusion matrix. The slowest diffusing eigenvector is largely due to the exchange between Si and Al, and the fastest diffusing eigenvector is the exchange of Na with all other components. An anorthite dissolution experiment was also conducted to test * Corresponding author. Email address: chguo@umich.edu whether the diffusion matrix can be applied to mineral dissolution experiments. The calculated diffusion profiles in the melt during anorthite dissolution roughly match the measured profiles, demonstrating the validity and utility of the diffusion matrix in this FeO-free aluminosilicate melt system.

Controlled Cooling‐Rate Experiments on Olivine‐Hosted Melt Inclusions: Chemical Diffusion and Quantification of Eruptive Cooling Rates on Hawaii and Mars

“…In the liquid at 1,225 °C, the time scale for homogenization of the slowest diffusing species (e.g., Al 2 O 3 ; Zhang, ; Guo & Zhang, ) is on order 100 min for the largest (~125‐μm radius) inclusion we have studied, and the 24‐hr isothermal heating was expected to be sufficiently long to relax any preexisting compositional gradients in the inclusions. This is corroborated by the representative microprobe profiles of homogenized inclusions shown in Figure , which demonstrate that after the 24‐hr heating step, the glass compositions are homogeneous within the analytical uncertainty of the electron microprobe measurements.…”

“…Chen and Zhang () parameterized D MgO as an effective binary diffusion coefficient (EBDC), which approximates diffusion in a multicomponent system as pseudobinary exchange between the component of interest and a second fictive component that is a combination of all of the other components in the solution (e.g., Chakraborty, ; Cooper, ; Liang, ; Watson & Baker, ). While the effective binary approach is useful in that it foregoes the need for a full diffusion matrix (e.g., Guo & Zhang, , ), the EBDC is in general a function of composition (Cooper, ; Zhang, ) and can vary based on the geometry of the diffusion couple (Liang, ). Accordingly, the EBDC generally will vary across a diffusion profile, will be time‐dependent unless the system has achieved steady state, and will in steady state depend on the geometry of the system (Cooper, ; Liang, ).…”

Aluminum‐enhanced alkali diffusion from float glass to PVD‐sputtered silica thin films

“…For bulk aluminosilicate glasses containing Na 2 O and/or K 2 O, interdiffusion between melts of differing alkali concentrations is mostly governed by the exchange of alkali ions with SiO 2 . Such exchanges, if present, would result in a variation in the Si/Al ratio before and after sodium migration.…”

Seconds after impact: Insights into the thermal history of impact ejecta from diffusion between lechatelierite and host glass in tektites and experiments

“…Other monotonic concentration profiles can be treated similarly. Note, however, that uphill diffusion profiles cannot be reproduced in this way and must either be treated by empirical models (e.g., Zhang, 1993), multicomponent and composition-dependent diffusion matrices (e.g., Trial and Spera, 1994;Guo and Zhang, 2016;Guo and Zhang, 2018), or models that couple the compositional dependence of the chemical potentials of components in silicate melts with diffusion (Richter, 1993). Development of such models is beyond the scope of this work, so we do not attempt quantification of the uphill diffusion observed for K 2 O.…”