Benchmarking Density Functional Theory for Accurate Calculation of Nitride Band Gaps

Area of Science:

• Materials Science
• Computational Chemistry
• Solid-State Physics

Background:

• Accurate calculation of fundamental band gaps is crucial for predicting semiconductor properties.
• Inorganic nitrides are under-represented in existing band gap benchmark studies.
• A comprehensive dataset of nitride band gaps is needed for theoretical benchmarking.

Purpose of the Study:

• To benchmark various exchange-correlation functionals and quasiparticle GW theory for calculating inorganic nitride band gaps.
• To expand the database of inorganic nitrides for band gap benchmarking.
• To identify accurate computational methods for high-throughput screening of nitride materials.

Main Methods:

• Literature survey to collect band gap data for 25 binary and 11 ternary inorganic nitrides.
• Benchmarking of conventional functionals (LDA, PBE), specialized functionals (SLOC, HLE16, AK13), meta-GGAs (TASK), hybrid functionals (mBJ, HSE06), and GW theory.
• Analysis of mean absolute errors (MAE) and mean absolute percentage errors (MAPE) for each method.

Main Results:

• Conventional LDA/PBE functionals significantly underestimate band gaps (MAE > 1.0 eV, MAPE ~ 50%).
• SLOC, AK13LDA, and HLE16 functionals show improved accuracy (MAE 0.5-0.6 eV, MAPE ~ 20-25%).
• mBJ and HSE06 functionals provide the most accurate results (MAE ~ 0.3 eV, MAPE ~ 11-12%).

Conclusions:

• The mBJ and HSE06 functionals are recommended for accurate band gap calculations of inorganic nitrides.
• Accurate functional choice is critical for machine learning models and high-throughput materials screening.
• This work contributes to a more balanced materials database for theoretical benchmarking.