The main goal of computer simulation here is the rapid and accurate description or prediction of catalyst structures and their thermal stabilities, which is sometimes difficult to achieve via simulation methods at a single length or time scale. For example, an Au NP with a diameter of only 5 nm owns around 4,000 atoms, it is enormously too large for first principle-based DFT calculations. Therefore, it is expected to use different simulation method to solve the problem at the corresponding scale, and sometimes the multiscale simulation strategies are necessary which bridge the models and simulation techniques across a broad range of length and time scales. 18,19 Based on the study of supported metal nanocatalysts, we mainly introduce here the catalysis application of three widely used simulation methods: the DFT calculation, MD simulation, and ML strategy.
| DFT-based first principle calculation for heterogeneous catalysisDFT is a computational quantum-mechanical modeling method used in chemistry, physics, and materials science to study the electronic structure (principally the ground state) of atoms, molecules, many-body systems, and the condensed phases. 20 The properties can be determined by using functionals of the electron density. Currently, DFT has become an essential complement in catalysis science. As for the research of supported metal nanocatalysts, it can provide insights into the geometric and FIGURE 1 Interaction energies for the density functional theory (DFT) calculations and the fitted Morse potential curve of the structure of (a) weak metalsupport interaction (MSI) and (b) strong MSI DENG ET AL.