Element partitioning and stabilization for impurities removal between liquid silicon and silicate melts: Ab initio insights into electronic structure

Abstract

The partitioning and stabilization of common impurity elements challenge the impurity removal process during slag refining for developing high-purity metals. The role of electronic properties remains indistinct in the stability of non-metallic, alkaline earth metal, and transition metal impurity clusters in silicon and silicate melts, impeding the advancement of separation medium and purification process. Herein, local structures of impurity atoms are explored through ab initio molecular dynamics, and notable configurations are [PSi4], [TiSi10], [MgSi12] in silicon melts and [PO4], [TiO5], [MgO5] in silicate melts. A novel partition model based on impurity coordination predicted the impurity atoms’ distribution behavior consistent with industrial experiments. Furthermore, this model proposes a critical parameter k that evaluates impurity oxidation levels in coexisting systems and establishes a quantitative link between melt structures and distribution ratio, enabling precise predictions from two-phase ab initio calculations. Innovatively, the electronic structure analysis assessed element partitioning through bonding characteristics and interactions within impurity clusters via parameters: electron density distribution, partial density of states, electron localization function, and Mayer bond order. This yields a distinctive mapping relationship between impurity distribution ratio and stability, evidenced by quantitative associations with 12 key ab initio-derived descriptors through Pearson correlation analysis. This work unveils the electronic properties governing partitioning and endows modeling approach for distribution, driven by cluster structure analysis and ab initio calculations.