In our laboratory, we are developing novel ways of carrying out quantum chemistry calculations. Some of the projects that we are engaged in this area are summarized below:

General-Purpose Graphics Processing units for Quantum Chemistry

We are exploring the use of video graphics cards (general-purpose graphics processing units or GPGPUs) to carry out correlated electronic structure calculations (Moller-Plesset Perturbation theory) with speedups of up to 20X. We are closely collaborating with Q-Chem, Inc.

Distributed computing and quantum chemistry.

In the context of The Clean Energy project and in collaboration with IBM's World Community Grid, we are working on carrying out tens of thousands of simultaneous quantum chemistry calculations using the BOINC infrastructure.

Quantum mechanics in complex environments.

In the context of our DARPA SERS fundamentals interaction with the groups of Ken Crozier and Eric Mazur, we are working on the development of novel electronic structure approaches that take nano-scale environments into account.This is a collaboration with Dr. Michael Stopa, from Harvard's Center for Nanoscale Sciences.

Novel electronic structure approaches

Electronic structure theory allows us to understand and predict the behavior of chemical processes. Nevertheless, practical methods are approximate and fail in certain cases. We are working on improving and developing methods for the determination of the electronic structure of atoms and molecules. In particular, we are working on new methods for density functional theory (DFT) and on improvements to the quantum Monte Carlo method. These techniques are complementary: DFT is fast and reasonably accurate, and quantum Monte Carlo can be used in benchmark studies, where a definitive answer is required.

Development of semi-classical dynamics methods.

In collaboration with Dr. Michele Ceotto, from the University of Milan, our group is developing practical first-principles semi-classical methods for nuclear propagation. We are working in integrating semi-classical propagation into mainstream quantum chemistry packages, and in particular finding ways to make it as efficient and convenient as regular classical first-principles dynamics propagation.