Transferable density functional tight binding for carbon, hydrogen, nitrogen, and oxygen: Application to shock...

Area of Science:

• Computational Chemistry
• Materials Science
• Theoretical Physics

Background:

• Density Functional Tight Binding (DFTB) is a computationally efficient quantum mechanical method.
• Accurate parameterization is crucial for DFTB's predictive power in molecular simulations.
• Existing methods may have limitations in accuracy for certain molecular systems and properties.

Purpose of the Study:

• To develop and validate a new DFTB parameterization, lanl31, for molecules containing carbon, hydrogen, nitrogen, and oxygen.
• To assess the transferability and accuracy of the lanl31 parameterization against high-level ab initio calculations and experimental data.
• To apply the lanl31 parameterization to simulate the principal Hugoniot of various energetic materials.

Main Methods:

• Developed the lanl31 parameterization by optimizing Hubbard U values, on-site energies, and interatomic potentials using simulated annealing.
• Validated the parameterization using the CHNO subset of the QM-9 database for atomization energies and interatomic forces.
• Performed molecular dynamics simulations of principal Hugoniots for several liquid organic materials.

Main Results:

• The lanl31 parameterization demonstrates small errors in atomization energies and interatomic forces, comparable to DFT methods.
• It shows good correlations for molecular dipole moments and HOMO-LUMO gaps.
• Simulations of principal Hugoniots for various materials show excellent agreement with experimental data, outperforming ReaxFF-lg.

Conclusions:

• The lanl31 DFTB parameterization offers a significant improvement in accuracy for C, H, N, and O containing molecules.
• It provides a reliable and efficient tool for molecular dynamics simulations, particularly for energetic materials.
• The study highlights the potential of DFTB for accurately predicting shock wave phenomena.