Multi-scale QM/QM/MM molecular dynamics through quantum embedding.

Abstract

Physical and chemical processes in bio-molecular systems typically span a wide range of time and space scales. These coupled processes at various scales must be taken into account concurrently in simulations depending on the problems of interest, yet they usually do not need to be depicted with the same accuracy. Large-scale protein conformational changes, for example, can be represented by carefully calibrated force fields, whereas chemical reactions at an enzyme's catalytic center necessitate a quantum mechanical approach. However, the number of atoms involved in the reaction center may still be quite large, limiting the applicability of some high-accuracy quantum chemistry methods, but the challenging part of the electronic structure may be even more local. A transition metal in enzymes, for example, may require some high-level multi-reference approaches, but a low-level mean-field description, such as the density functional theory, may be sufficient for metal ligands, serving as its environments. In this study, we used quantum embedding theory to rigorously partition a quantum system into multiple fragments, each of which could be solved using high-level quantum chemistry methods, allowing us to develop a multi-scale QM(high-level)/QM(low-level)/MM approach to achieve varying levels of accuracy in different scaled phenomenon.