Semi-Automated Creation of Density Functional Tight Binding Models through Leveraging Chebyshev Polynomial-Based...

Area of Science:

• Computational materials science
• Quantum simulations
• Condensed matter physics

Background:

• Density Functional Tight Binding (DFTB) offers computational efficiency for quantum simulations compared to DFT.
• Developing accurate DFTB models, especially for metallic and interfacial systems, is challenging due to complex bonding and electronic states.
• Existing methods require significant effort for system-specific DFTB potential development.

Purpose of the Study:

• To create a rapid-screening approach for systematically improvable DFTB interaction potentials.
• To develop transferable DFTB models applicable across various thermodynamic conditions.
• To streamline the creation of reliable DFTB models for materials simulations.

Main Methods:

• Leveraged a reactive molecular dynamics force field using Chebyshev polynomial representations for many-body interactions.
• Employed a rapid-screening workflow for efficient generation of multi-center representations.
• Utilized a small training set of DFT calculations for model development, focusing on TiH2 as a model system.

Main Results:

• Successfully generated a systematically improvable DFTB model using a small training set.
• The developed DFTB model demonstrated accuracy for both bulk and surface properties of TiH2.
• The approach proved effective for a range of thermodynamic conditions, validating its transferability.

Conclusions:

• The rapid-screening approach enables efficient and reliable development of transferable DFTB models.
• This method reduces the reliance on extensive DFT calculations for model creation.
• The developed DFTB models enhance the simulation of condensed matter systems, particularly those with complex bonding.