Multicomponent Diffusion in Seven- and Eight-Component Silicate Melts

Natural silicate melts (i.e., magmatic liquids) contain 5 to 10 major oxide components. Hence, diffusion in natural melts is always multi-component diffusion with manifestations such as uphill diffusion. However, complexities in rigorous treatment of multicomponent diffusion made geochemists shy away from treating such diffusion in the past. My group has been working on multicomponent diffusion in seven-and eight-component silicate melts for about 10 years. We started with multicomponent diffusion in a seven-component system (SiO 2 -TiO 2 -Al 2 O 3 -MgO-CaO-Na 2 O-K 2 O) and obtained the 66 diffusion matrix [1]. Then we focused on a synthetic mid-ocean ridge basalt with eight components (SiO 2 -TiO 2 -Al 2 O 3 -FeO-MgO-CaO-Na 2 O-K 2 O) [2,3] because basalt is the most abundant volcanic melt in the Earth. From experimental data, we obtained 77 diffusion matrix at three temperatures, and found that the eigenvector matrix is roughly invariant with temperature and each of the seven eigenvalues depends on temperature following the Arrhenius relation. This provides a formulation to calculate the diffusion matrix at any temperature within the experimental temperature range. Furthermore, we hypothesize that the diffusion eigenvectors are independent of melt compositions [4,5]. Therefore, we can examine diffusion in silicate melts in eigen-component space, and each eigenvalue is the diffusion coefficient for its corresponding eigen-component. Our preliminary examination of literature data shows that most data are consistent with the hypothesis. We are beginning to develop an online tool to model multicomponent diffusion in natural silicate melt using the eigen-component approach [5]. Here, I report these developments for the broader diffusion community and present future perspectives.