My research interests are mainly focused on the theoretical description of
various physical and chemical events using the tools of Quantum Mechanics
and Quantum Chemistry. At very beginning in China, I did some researches
on empirical electronic theory of solids and molecules with my honors
supervisor, Prof. Ruihuang Yu, member of the Chinese Academy of Sciences.
With this empirical theory, I developed the theoretical method for predicting
the enthalpy of formation of binary liquid and solid phases in the variants
of Miedema's semi-empirical model. And also, by using this theoretical method,
I calculated valence electron structure of superconductor YBa2Cu3O7
and gave the correlation of structure transformation, doping and oxygen content.
After I graduated from Jilin University with my Ph.D, I worked in Institute of
Material Science, Jilin University, and Institute of Solid State Physics, Siping Normal College of China,
having several projects, for example, the magnetic properties of nanometer material F2O3
and F3O4, and seed polymer coatings. In 1996, I began to code developments of theoretical
chemistry in the institute of theoretical chemistry, Erlangen-Nürnberg University, Germany, working with
Prof. Ladik and Prof. Otto. In there, I developed the package, which is used to
calculate band structures of one-dimensional systems in the basis of Dirac-Hartree-Fock equations
with Gaussian basis functions. In 1998, I joined in the ParaGauss team of the institute of
theoretical chemistry, Technological University of Munich, Germany. Working with Prof. Rösch,
I developed an algorithm for analytically evaluating integral elements of Gaussian pseudopotentials
with solid harmonic Gaussian basis functions. This algorithm is also adaptive to analytically evaluate
their derivatives because Hobson's theorem of solid harmonics leads to recursion relation to calculate
derivatives of those integral matrix elements. The algorithm has been coded in ParaGauss 2.1.
At beginning of May, 1999, I started to work with Prof. York, in Department of Chemistry, University of
Minnesota, as a postdoctor research fellow and a research fellow in Minnesota Supercomputing
Institute. Working together with Prof. York, I developed a quantum chemistry program ( in the moment called
F95-Parallel DFT Package ),
which is written by using F95 with the object-oriented style and parallelized
by generic programming interface for MPI and PVM implementing its own automatic administration of send and receive buffers.
In F95-Parallel DFT Package, both Gaussian and numerical basis functions can be used. While with numerical
basis sets, it is quite same as Dmol package in many ways since it is based
on Delley's idea, for example, density partitioning into atomic spheres with
rapidly multipolar expansion providing reliable and accurate techniques
to treat Coulomb integrals. With Gaussian basis functions, density fitting technique is used in order
to avoid calculating Coulomb integrals. Now, the parts of this code with numerical basis sets are modularized.
It means that it consists of some modules, for example, normal scf,
direct scf to save memory, coupled Kohn-Sham response, magnetic shielding of
continuous transformation of gauge origin, polarization density,
magnetization density, and conductivity density tensors, and graphic tools
to show those density tensors, and so on. As following, it shows components of dynamic polarization density,
static polarization density, magnetization density, and conductivity density tensors of some molecules and
fregments, calculated by using F95-Parallel DFT Package recently.
