First-principles method for carrier transport in polycrystalline materials

Carrier transport in polycrystalline materials is often studied using the Mayadas-Shatzkes (MS) model, originally developed for 3D metals with certain simplifications. However, the MS model relies on phenomenological parameters that are unknown without fitting to experimental data or making assumptions, which limits its accuracy and applicability. Here, we develop a fully ab-initio method that improves accuracy and broadens applicability across various materials. Using this method, we investigate carrier transport in polycrystalline 2D MoS₂. Taking mirror twin boundaries (MTBs) as a case study, we find that accurate modeling of grain boundary (GB) scattering requires careful consideration of state-dependent and non-uniform scattering effects. We also observe that certain MTBs exhibit stronger band bending, which leads to reduced transmission near the band valley minimum and, consequently, lower mobility. Our results demonstrate that both the MS model and Matthiessen’s rule fail to accurately predict mobility due to their inherent assumptions, highlighting the need for models that capture the complex, state-dependent scattering behavior in polycrystalline materials.