Shear viscosity of liquid Sb and Bi from pseudopotential theory: The temperature-insensitive soft-tail component

Abstract

A static, pairwise pipeline, a Bretonnet–Silbert (BS) pseudopotential with Rice–Allnatt transport theory, reproduces the magnitude of liquid-metal shear viscosity, while its temperature gradient marks where the static description reaches its boundary. We compute the shear viscosity η(T) of liquid Sb and Bi from each melting point to at least 500 K above it, using five local-field corrections and the Zahid pseudopotential parameters, and resolve the temperature gradient |dη/dT| into kinetic, hard-core, and soft-tail components. The soft-tail component η_s accounts for 54–59% of the total viscosity but barely changes with temperature, so the calculated gradient is systematically too weak: η is underestimated at the melting point (−16% for Sb, −32% for Bi) and overestimated at high temperature, crossing the reference data near 1025 K (Sb) and 850 K (Bi). Because η_s is an integral over the pair structure g(r), replacing the hard-sphere form g^HS(r) with a modified hypernetted-chain (MHNC) one raises η_s by 7–21% but does not recover the gradient; across the hard-sphere and MHNC closures tested here, improving the static pair structure alone leaves the slope deficit largely unchanged, pointing to the Rice–Allnatt decomposition rather than these g(r) models. The local-field correction changes η by only 3–4%, an order of magnitude less than the spread induced by the non-unique BS parameters; the sparse ab initio molecular-dynamics data do not contradict this weaker-than-experiment gradient. These results delimit the static, pairwise framework for Sb and Bi and provide a component-resolved benchmark for dynamic, many-body, or screening extensions of liquid-metal transport theory.

The authors' abstract, as published at the source. Fluid Phase Equilibria, 2026 · DOI ↗