Enhancing the accuracy of density functional tight binding models through ChIMES many-body interaction potentials

Area of Science:

• Computational Chemistry
• Materials Science
• Quantum Mechanics

Background:

• Semi-empirical quantum models like Density Functional Tight Binding (DFTB) offer scalability for simulations but lack systematic parameterization methods.
• Developing accurate DFTB models often requires significant manual effort and expertise.

Purpose of the Study:

• To present a novel approach using the Chebyshev Interaction Model for Efficient Simulation (ChIMES) for rapid DFTB model parameterization.
• To demonstrate the transferability and accuracy of ChIMES-parameterized DFTB models across different material systems.

Main Methods:

• Utilized the ChIMES framework to develop parameterized DFTB models, incorporating many-body interactions.
• Applied the developed models to simulate silicon polymorphs and reviewed existing work on titanium hydride.
• Created a general-purpose DFTB/ChIMES model for organic molecules and compounds.

Main Results:

• ChIMES-parameterized DFTB models achieved accuracy comparable to underlying quantum methods but with significantly reduced computational cost.
• The developed models demonstrated strong transferability across diverse chemical systems.
• Achieved accuracy approaching hybrid functional and coupled cluster methods with orders of magnitude fewer parameters than neural network approaches.

Conclusions:

• The DFTB/ChIMES approach provides a computationally efficient and accurate method for quantum simulations.
• This methodology facilitates the study of materials under extreme thermodynamic conditions, aiding in the interpretation of experimental results.