|
|
|
|
Research Interests
1. Dynamics of Very Large Molecular Systems We are interested in simulating the dynamic processes of large complex systems, such as polymer aggregates or films, biological molecules and nano-structures. To simulate the electronic dynamics of these systems, we have developed a linear-scaling method for electronic ground and excited states, the localized-density-matrix (LDM) method. It has been employed to determine the absorption spectra of large polyacetylene oligomers, PPV aggregates, carbon nanotubes and light-harvesting systems. The largest system that has been calculated is a polyacetylene oligomer that contains 30,000 carbon atoms, and the PPP Hamiltonian was employed in the calculation. The LDM method has been so far implemented at semiempirical CNDO/S, INDO/S, PM3 and AM1, and the time-dependent density functional theory (TDDFT). Efforts are under the way to calculate other dynamic properties, for example, emission spectrum, circular dichroism spectrum, conductivity and NMR signals.
Quantum dissipation is a subject of wide spread interests in many fields of physics, chemistry and materials science. Various quantum dissipation theories (QDTs) have been developed to investigate the dynamic properties of open systems, for instance, the Bloch-Redfield theory, Fokker-Planck equation and Lindblad semi-group formalism. The key physical quantity in all these theories is the reduced system density matrix. The computational costs of these theories are very expensive, and the calculations are thus limited to the small model systems. We propose a new formalism to simulate the electronic dynamics of open systems with the ultimate objective to investigate the complex open systems from the first-principles. Unlike the existing QDTs where the reduced density matrices of systems are followed, we develop the equation of motion for the reduced single-electron density matrix while considering quantum dissipation explicitly. This enables us to study much larger systems than before. Combined with the LDM method, the new formalism is expected to be employed to simulate the electronic dynamics of very large open systems. The thermal baths can be the nuclei of system and/or the environment. The traditional phenomenological dissipation coefficient g in the equation of motion for the reduced single-electron density matrix is replaced by explicit quantum dissipation terms. These terms include the electron-nuclei couplings and energy/material exchanges with the environment. The semi-empirical Hamiltonians like CNDO/S, INDO/S and PM3 are to be employed to describe the dynamics of combined system (the system plus environment). We will also seek to implement the new formalism within the framework of the TDDFT.
We have developed a neural network based software to construct the quantitative structure-activity relationship (QSAR) using the existing experimental data. It has been used to search for drug candidates of aldose reductase inhibitors (ARIs). Combined with Cerius2, Autodock and QM/MM program, it provides an efficient and powerful tool to design drug candidates on computer.
|
|
|
