Machine-Learned Force Fields for Lattice Dynamics at Coupled-Cluster Level Accuracy
This work addresses the problem of high-accuracy force field development for materials science, offering incremental improvements in predicting solid-state vibrational properties.
The researchers tackled the challenge of accurately modeling lattice dynamics in solids by developing machine-learned force fields trained on coupled-cluster theory, which improved vibrational frequency predictions for optical modes compared to density functional theory, with results better aligning with experimental data.
We investigate Machine-Learned Force Fields (MLFFs) trained on approximate Density Functional Theory (DFT) and Coupled Cluster (CC) level potential energy surfaces for the carbon diamond and lithium hydride solids. We assess the accuracy and precision of the MLFFs by calculating phonon dispersions and vibrational densities of states (VDOS) that are compared to experiment and reference ab initio results. To overcome limitations from long-range effects and the lack of atomic forces in the CC training data, a delta-learning approach based on the difference between CC and DFT results is explored. Compared to DFT, MLFFs trained on CC theory yield higher vibrational frequencies for optical modes, agreeing better with experiment. Furthermore, the MLFFs are used to estimate anharmonic effects on the VDOS of lithium hydride at the level of CC theory.