Development And Application Of Multi-Scale Computational Methods For Complex Systems

With the development of computer technology and the enhancement of computing power,multiscale simulations are being applied in the field of computational chemistry increasingly.Since chemical reactions and the properties of biomolecules often involve physical and chemical behaviors at different time and spatial scales,multiscale simulation method can combine information from these scales using multi-level computational methods from quantum mechanics to classical mechanics,thus more accurately predicting and explaining the properties and behaviors of the system.Compared with traditional computational methods,multiscale simulations can more accurately simulate the physical and chemical behaviors of complex systems such as macromolecules and materials,and therefore have been widely used in computational chemistry research.Especially in the design of new catalysts,drug molecules,and other fields,the accuracy and efficiency advantages of multiscale simulations are even more pronounced,providing powerful support and guidance for experimental research.This paper focuses on the multi-scale computational methods and applications for studying complex systems.In the first chapter,the basic principles of quantum chemistry and density functional theory,as well as the basic theory of molecular dynamics simulation,are introduced.In the second chapter,a new density functional,CF22D（Chemistry Functional 2022 with damped Dispersion）,developed for more widespread applications in theoretical research in various fields of chemistry,is presented.Compared with existing functionals,CF22D functional uses physical descriptors,extensive databases,and a supervised learning strategy.Firstly,by combining the global hybrid non-separated meta general gradient approximation and the damped dispersion,which depend on the electronic density and occupied orbitals,with a flexible functional form,the search for new density functional forms is expanded to a more broadly defined new energy functional search.By simultaneously optimizing the form of the density functional and the parameters of the molecular mechanics damping dispersion term,CF22D provides a universally reliable prediction of non-covalent interactions.Secondly,a performance-triggered iterative supervised training strategy is used to optimize the energy functional.CF22D functional can be widely applied to chemical research areas such as bonding energies,chemical properties,reaction barriers,and non-covalent interactions in main group element and transition metals.Studying chemical reaction mechanisms is beneficial for researchers to understand the bonding between atoms or groups.However,the structures of reaction transition state exist for an extremely short time,making it difficult to capture experimentally.Therefore,computational chemistry methods have become an effective means of studying reaction transition states.Density functional theory（DFT）is one of the commonly used methods for studying reaction mechanisms.In the third chapter of this paper,the high-precision DFT calculations were investigated to study the chemical reaction mechanism of the formation of benzene ring from the T₁/T₂states of acetylene and ground-state acetylene through photochemical treatment on the Pluto’s surface.The structures of the reactants,transition states,intermediates,and products were optimized at B3LYP/6-311G**level.More precise results were obtained for the T₁/T₂states structures by performing calculations at CCSD（T）/CBS and CASPT2/aug-cc-p VTZ level,respectively.The study revealed that the reaction only requires a small energy barrier to occur,with the energy barrier exhibiting a decline in acetylene ice environment.The investigation holds significant implications regarding the abundance of benzene and polycyclic aromatic hydrocarbons on Pluto’s surface,given that they may form before haze particles with acetylene attachment land on the surface of Pluto.Therefore,this study provides theoretical support for a deeper comprehension of the chemical reaction mechanism of benzene and polycyclic aromatic hydrocarbons on Pluto’s surface.Research on chemical reactions in systems with a small number of atoms can be studied using high-precision DFT calculations to investigate their reaction mechanisms.With the increase of the time scale and space scale of the research system,the study of chemical reactions in bio-molecules is typically carried out using molecular mechanics methods,and the biological behavior of biological macro-molecules can be observed in a long time scale.In the fourth chapter of this paper,molecular dynamics simulations of biomolecular systems were conducted using the constant pH molecular dynamics（Cp HMD）method based on molecular mechanics.By titrating the internal ionizable residue GLU8,the conformational stability ofβ-endorphin amyloid-like fibrils under different pH conditions（from acidic to neutral）was studied,and the pK_avalues of GLU8 on different peptide chains were systematically investigated.This study reveals that the mechanism of gradual release ofβ-endorphin monomers is related to the degree of pK_ashift of GLU8 at different positions.It provides a clear idea for understanding the structural transformation of amyloid proteins in hormone secretion processes,and provides a theoretical basis for designing intelligent bio-tools for efficient storage and rapid release of analgesic hormones.