Harnessing Large-Scale Quantum Calculations for Predicting Material and Chemical Properties

Abstract

Density functional theory (DFT) is a powerful method for probing chemical and material properties and guiding the design of novel materials from first principles. However, the computational demands of large-scale and long-time DFT calculations can be challenging. To address this limitation, this presentation discusses the approximate DFT method, density functional tight-binding (DFTB), as an alternative approach. The first part of the talk demonstrates the role of DFT in describing the electronic structure and transport properties of doped carbon nanotubes. In the second part of the talk, I will discuss the accuracy and efficiency of DFTB in performing large-scale electronic structure calculations. To achieve this goal, we have interfaced DFTB with the cluster approach to statistical mechanics (CASM) program, which allows for the efficient evaluation of formation energies and convex hull. Furthermore, we have extended the DFTB approach to perform long-timescale (~10 ns) metadynamics calculation on biochemical systems with the help of GPUs. GPU-DFTB allows for an efficient and accurate description of the free energy surfaces and provides valuable insight into the transition pathways. In summary, we aim to accelerate quantum-based computations by enabling accurate and efficient prediction of material properties with DFTB.