Introducing PAIMP, the polarizable ab initio model potential method for embedded clusters.

Area of Science:

• Computational Materials Science
• Quantum Chemistry
• Solid-State Physics

Background:

• Accurate modeling of point defects in ionic solids is crucial for understanding material properties.
• Existing methods like the ab initio model potential (AIMP) framework have limitations in describing polarization effects.

Purpose of the Study:

• Introduce a novel polarizable ab initio model potential (PAIMP) to enhance calculations of point defects.
• Investigate the impact of polarization on defect properties, particularly for surface and charge-transfer phenomena.

Main Methods:

• Developed PAIMP by augmenting AIMP with a self-consistent induced-dipole operator and site-resolved polarizability tensors.
• Benchmarked PAIMP using Cr-doped α-Al2O3 (bulk and surfaces) and a bulk oxygen vacancy (F-center).
• Employed linear-response complete active space self-consistent field (LR-CASSCF) and multi-state complete active space second-order perturbation theory (MS-CASPT2) calculations.

Main Results:

• PAIMP and AIMP yielded similar results for bulk Cr:Al2O3 d-d transitions due to high symmetry.
• Significant PAIMP effects observed for under-coordinated surface defects, including red-shifted charge-transfer bands with enhanced structure and oscillator strengths.
• LR-CASSCF and MS-CASPT2 with PAIMP qualitatively captured the F-center doublet but remained blue-shifted compared to experimental data.

Conclusions:

• Polarization effects are negligible for bulk d-d transitions but crucial for charge-transfer excitations and low-coordination environments.
• PAIMP provides more structured spectra and higher intensities for these systems.
• PAIMP represents a significant advancement for computationally efficient and robust descriptions of localized defects in ionic materials.