Development of Density Functional Tight-Binding Parameters Using Relative Energy Fitting and Particle Swarm...

Area of Science:

• Computational Chemistry
• Materials Science

Background:

• Density Functional Tight-Binding (DFTB) is a computationally efficient method for electronic structure calculations.
• Accurate parameterization is crucial for DFTB's reliability in predicting molecular properties.
• Existing DFTB parameterization methods may not fully capture the nuances of organic molecule behavior.

Purpose of the Study:

• To develop and validate an optimized parameterization strategy for DFTB applied to organic molecules (H, C, N, O).
• To enhance the accuracy of DFTB in predicting molecular structures and properties, including binding energies, atomic forces, and relative isomer energies.
• To create a more chemistry-driven DFTB parameterization through advanced objective functions.

Main Methods:

• Utilized Particle Swarm Optimization (PSO) to find optimal DFTB parameters.
• Developed an objective function incorporating binding energies, atomic forces (Ballester similarity index), and relative isomer energies (Levenshtein distance-induced similarity).
• Created comprehensive training and testing datasets covering relevant chemical functional groups.

Main Results:

• The new DFTB parameterization demonstrates improved accuracy compared to previous methods.
• Excellent agreement was observed between DFTB results and high-level Density Functional Theory (DFT) data from QM-9 and ANI-1 datasets.
• The optimized parameters accurately predict molecular structures and properties for organic molecules.

Conclusions:

• The proposed strategy offers a robust and accurate method for DFTB parameterization of organic molecules.
• This enhanced DFTB parameterization significantly improves the prediction of molecular structures and properties.
• The approach is validated against large DFT datasets, showing broad applicability and high fidelity.