Identifying native point defect configurations inα-alumina.

Area of Science:

• Materials Science
• Solid-State Physics
• Computational Chemistry

Background:

• Point defects and their atomic/electronic structures are crucial for material properties like corrosion resistance and optoelectronic device performance.
• Complex energy landscapes with metastable defect configurations pose challenges for accurate first-principles modeling.

Purpose of the Study:

• To reevaluate native point defect geometries in α-Al2O3 using various computational sampling methods.
• To identify novel defect configurations and assess the effectiveness of different geometry sampling techniques.

Main Methods:

• Density Functional Theory (DFT) calculations were employed to study point defects in α-Al2O3.
• Three distinct methods for sampling candidate geometries were compared: atomic displacement, Voronoi decomposition initialization, and Bayesian optimization.

Main Results:

• Symmetry-breaking distortions were observed for oxygen vacancies in certain charge states.
• Several distinct oxygen split-interstitial geometries were identified, potentially resolving literature discrepancies.
• A previously unreported trigonal geometry for aluminum interstitials was discovered.

Conclusions:

• The identified novel defect configurations can significantly impact understanding of defect migration in protective aluminum-oxide scales.
• The Voronoi scheme demonstrated superior effectiveness in sampling interstitial sites, consistently finding the lowest-energy geometries.
• Defect geometry strongly influences defect level positions within the band gap, emphasizing the need for thorough ground-state geometry searches.