James B. Anderson

Quantum Chemistry/Molecular Dynamics/Reaction Kinetics

The theoretical chemist is accustomed to judging the success of a theoretical prediction according to how well it agrees with an experimental measurement. Since the object of theory is the prediction of the results of experiment, that would appear to be an entirely satisfactory state of affairs. However, if it is true that "the underlying physical laws ... for the whole of chemistry are ... completely known" (Dirac, 1929), then it should be possible to predict the results of experiment more accurately than they can be measured. If the theoretical chemist could obtain exact solutions of the Schrödinger equation for many-body systems, the experimental chemist would soon become accustomed to judging the success fo an experimental measurement by how well it agrees with a theoretical prediction.

In fact, it is now possible to obtain exact solutions of the Schrödinger equation for systems of a few electrons. These systems include the molecular ion H3+, the molecule H2, the reaction intermediate H-H-H, the unstable pair H-He, the stable dimer He2, and the trimer He3. The Quantum Monte Carlo method used in solving the time-independent Schrödinger equation for these systems is exact in that it requires no physical or mathematical assumptions beyond those of the Schrödinger equation. As in most Monte Carlo methods, there is a statistical or sampling error which is readily estimated.

For larger systems, neither Quantum Monte Carlo methods nor any other methods at present provide such exact results. However, for many of these larger systems, Quantum Monte Carlo calculations provide the lowest-energy, most accurate results available. The systems include atoms such as Fe; molecules such as H2O, CH4, and HF; larger molecules such as C20; and condensed materials such as diamond and solid N2.

Professor Anderson and his co-workers are investigating these and other methods for improved predictions of quantum chemistry. Their current emphasis is in the area of high-performance computing in materials physics and chemistry, with the aim of more accurate predictions for larger organic systems and diamond-like materials.

The group is also active in the areas of reaction kinetics, chemical dynamics, and molecular dynamics. Projects in these areas include studies of Monte Carlo methods for the direct simulation of reaction systems with nonthermal distributions, with coupled gas-dynamic and reaction effects, and with many other effects difficult to treat in any other way. Also included is research in the combination of transition-state theory and molecular dynamics known as rare-event theory and used for the simulation of rare events such as simple reactions in the gas phase, exchange reactions in solution, enzyme-catalyzed reactions, and protein rearrangements.