Calculating and Characterizing the Charge Distributions in Solids

Area of Science:

• Computational materials science
• Solid-state chemistry
• Quantum chemistry

Background:

• Accurate partial atomic charges on metal centers are vital for understanding material electronic and catalytic properties.
• Discrepancies in charge calculation methods are well-documented for molecules but less so for solid materials.

Purpose of the Study:

• To systematically investigate the impact of various density functionals and charge models on partial atomic charges of transition metal centers in solids.
• To compare the performance of different computational approaches for calculating metal center charges in diverse solid-state systems.

Main Methods:

• Density functional theory (DFT) calculations were performed using multiple functionals (PBE, PBE+U, TPSS, revTPSS, HLE17, revM06-L, B3LYP, B3LYP*).
• Four charge models were employed: Bader, Hirshfeld, CM5, and DDEC6.
• A test set of 18 solids was used, including 12 nonmagnetic metal oxides, sulfides, and selenides (MX2), and 6 ferromagnetic transition metal complexes.

Main Results:

• Bader charges were consistently the highest, while Hirshfeld charges were the lowest across all tested systems and functionals.
• Charge model selection exhibited a greater variation in results than the choice of density functional.
• Metal center charges were sensitive to Hubbard U parameters in PBE+U and Hartree-Fock exchange percentage in modified B3LYP functionals, generally increasing with higher values.

Conclusions:

• The choice of charge model is a dominant factor influencing partial atomic charge calculations for transition metal centers in solids.
• PBE+U with appropriate U parameters and certain meta-GGA functionals like HLE17 show promise for accurate charge estimations, comparable to hybrid functionals like B3LYP.
• Further studies are needed to establish standardized best practices for charge calculations in solid-state materials.