A Theoretical Study On The Molecular-Photonics Of Complex Systems

The investigation of structures and properties of complex systems is an important science research direction.It is a hot topic to employ theoretical simulations to study molecular photonics of complex systems,so as to acquire important information of structural evolution,photo-electron interactions,and other physical and chemical properties.With the rapid development of high-performance computers and quantum mechanical calculation methods,density functional theory(DFT)based on first principles has become an indispensable method to study complex systems.Simultaneously,the development of molecular dynamics theory(MD)also provides effective ways to capture the dynamic structural evolutions of large-scale systems.The boom of artificial intelligence such as machine learning(ML)inspires new ideas for analysis and processing of big data of complex systems.Photo-absorption and emission spectra play important roles in molecular photonics.In principle,the microscopic nature is photo-electron interaction,which is the key bridge connecting their micro-structures and optical properties.However,it also brings out the bottleneck problem for theoretical simulations and interpretations.Especially,for some complex molecular systems,the excited states induced by photo-electron interactions can hardly be simulated accurately.On the other hand,there's an alternative way to employ the divide-conquer strategy which break the complex system into fragments,and then calculate each fragments and further construct the exciton Hamiltonian matrix by considering interactions between excitons in every fragment.Furthermore,for the study of photoemission by complex systems,we have proposed a novel strategy of aggregate-enhanced molecular photo-emission,with which we can tune photo-excitations and develop functional materials with enhanced performance.In this dissertation,we have employed DFT,MD and ML methods to study molecular photonics of complex systems,covering the simulation of photo-absorption spectra and molecular luminescence.To be specific,one topic is about the development and application of a computation protocol to predict the ultraviolet(UV)photo-absorption spectroscopy of proteins;another topic is to investigate the influence of molecular aggregations on the photo-emission properties.This thesis is mainly divided into the following five chapters:Chapter 1 contains two sections:the first one mainly introduces the recent progress of various types of protein photo-spectroscopic characterizations.Proteins are the building blocks of life and their structural characterization is helpful for understanding the functions and underlying mechanisms.One can determine protein structures and reveal atomic and electronic kinetics by analyze experimentally measured photo-spectroscopic signals with the help of theoretical interpretations and signal assignments.However,the theoretical simulations of proteins spectra are often limited by the huge computational costs to consider those complex and flexible structures,real-time dynamic evolution and environmental fluctuations.The idea of divide-conquer allows us to calculate the individual fragments of the whole system,and later combine the electrostatic interactions between fragments based on the Frenkel model to build the exciton Hamiltonian matrix.Inspiringly,the recent quick advancement of artificial intelligence technologies such as machine learning(ML)and deep learning(DL)method have exhibited great advantages in dealing with nonlinear problems with simple rules and high complexities,enabling the prediction of properties based on the careful training of big data obtained from high-throughput quantum chemistry calculations.This offers possibilities of efficiently simulating protein photo-responses and spectroscopic signals at the quantum chemistry level by using the prediction power of ML-based methods.The second part of Chapter 1 introduces the recent advances of molecular luminescence of aggregated molecular systems.Compared to the traditional way of adjusting molecular functional groups for optimized photo-emission properties,the tuning of molecular aggregation level holds advantages of design flexibility and broad application potential.The luminescence mechanism in aggregated systems paves a promising way for the optimization of molecular luminescence properties.In Chapter 2,we introduce the first-principles method of density functional theory(DFT).Based on the Born-Oppenheimer approximation,DFT is suitable for the calculation of electronic structures and properties of systems with dozens or hundreds of atoms.Electron density is the key parameter of DFT calculations,for which the complex interacting many-electron system is simplified into the movement of non-interacting single-electron in a mean field,based on which the Kohn-Sham equation is established and to be solved.Finally,the ground state electron density and total energy of the system are attained by considering the exchange-correlation functional approximation and electron self-consistent iteration.Time-dependent density functional theory(TDDFT)can be carried out by introducing the concept of instantaneousness into DFT,which allows a relatively accurate simulation for excited states.In Chapter 3,we introduce the basic idea of molecular dynamics theory(MD).MD is one of the tools to simulate the time-dependent behavior of a molecular system.A MD simulation explores the macroscopic properties of a relatively complex system through the calculations of atomic motions at microscopic scale,based on the Newtonian equation of particle motion.We can acquire a trajectory in a MD simulation,attain positions and velocities as time goes,and further determine evolution information at any time.Usually it requires one to calculate the forces or interactions of atoms in the system.By carrying out an equilibrium simulation for a long enough time,we can obtain the trajectory of the system and further analyze the interested properties.In Chapter 4,we introduce the investigation of photo-electron interactions of complex systems,focusing on the simulation of protein far-ultraviolet(FUV)photo-absorption spectroscopy.A protein is split into a series of fragments;afterwards,a ML procedure is employed to construct the structure-property relationship for all fragments,to predict excitation energies and transition dipole moments of peptide bonds and ground state dipole moments of amino acid residues based on descriptors of geometric information;finally the properties of the whole protein are computed by solving the Frenkel exciton Hamiltonian combining excitons of all fragments and their interactions.FUV spectroscopy is acquired by diagonalization of Hamiltonian.The predicted FUV spectra of proteins can then be used to characterize their structures,which is very important for understanding their functions.Importantly,the ML protocol based on QM and MD simulation shows a favorable transferability,which has been successfully applied to the study of protein folding and mutation.In Chapter 5,we introduce the research work on a novel mechanism of aggregation enhanced intersystem transition(ISC)and its effect in promoting molecular phosphorescence.ISC plays a crucial role in molecular luminescence.Conventionally,one can tune ISC ability by incorporating heavy-atoms,which is suffered to disadvantages of high price,dark-toxicity,and low flexibility.Here we propose the design of aggregated pure organic molecules,in which the intermolecular interactions affect the ISC processes.Our DFT simulation results found that the intermolecular interactions induces a strong energy splitting effect on the molecular orbitals,decreasing the energy gap between triplet states and singlet excitation states and resulting in the emergence of more ISC channels.It enhances the overall ISC rate and therefore promotes phosphorescence performance.Such a novel aggregation-induced intersystem crossing(AI-ISC)mechanism is of great potential in improving molecular phosphorescence intensity,lifetime and the tuning of emission wavelength.