ITR: Science and Software for Predictive Simulation of Chemo-Mechanical Phenomena in Real Materials

This award is in response to a medium proposal submitted to the FY03 Information Technology Research (ITR) Solicitation. It is jointly supported by the Division of Materials Research and the Chemistry Division. In addition to primary research at Florida, research will also be done at MIT and Arizona.

Themes that distinguish this research on molecular simulations are that it is quantum mechanically based; predictive; chemically accurate; and, electronic state specific. The initial focus will be on the effect of water on the properties of a silica nanorod and on electron transport in nanostructures. Computer modeling of materials can be no better than the forces used to describe the interaction among the atoms involved. These are usually based on classical physics, which permits the rapid generation of forces required for large-scale simulations. But because bond breaking and formation, optical properties, and chemical reactions are poorly described classically, reliable modeling must ultimately be based on quantum mechanics (QM). But quantum calculations are computationally intensive, leading to a multi-scale approach in which quantum mechanics is used in critical regions, which are then embedded in a classical simulation. The inclusion of quantum mechanics is necessary to make materials modeling predictive enough to guide experiment.

Simulations must be chemically accurate, a necessary feature if modeling is to succeed on long unsolved problems such as the reason why silica, when wet, is weaker by several orders of magnitude than when dry, while addition of ammonia to silica shows no such effect. A proper, quantum mechanically based simulation should reflect these differences, qualitatively and quantitatively. Another theme is that, by using QM in the simulations, electron state specificity is achieved. Classical models do not distinguish between ground and excited electronic states.

A transfer Hamiltonian (TH), developed in work preparatory to this research, has a functional form that permits simplified QM to retain predictive quality at increased computation speed, and is a generalization of the frequently used tight-binding (TB) approximation that cures TB's inability to describe bond breaking. The TH differs from the popular density functional theory methods in that it fits to the true QM Hamiltonian rather than to densities or energies.

The interface between quantum regions and their classical embedding will be quantified using a density-operator Liouville-von Neumann dynamics that offers a framework for separating a fully QM system into two parts, modeling one in a classical or dielectric manner.

The key results of the project will be new theoretical methods for chemically accurate and realistic materials modeling, and their software implementation, applied to challenging problems. A large number of graduate students and postdoctoral associates will be supported, as well as undergraduate students with an emphasis on underrepresented groups. Sessions on materials simulation results will be integrated into the annual Sanibel meetings. Software produced will be made available to the wider community through the Materials Computation Center at the University of Illinois.

This award is in response to a medium proposal submitted to the FY03 Information Technology Research (ITR) Solicitation. It is jointly supported by the Division of Materials Research and the Chemistry Division. In addition to primary research at Florida, research will also be done at MIT and Arizona.

Themes that distinguish this research on molecular simulations are that it is quantum mechanically based; predictive; chemically accurate; and, electronic state specific. The initial focus will be on the effect of water on the properties of a silica nanorod and on electron transport in nanostructures.

The key results of the project will be new theoretical methods for chemically accurate and realistic materials modeling, and their software implementation, applied to challenging problems. A large number of graduate students and postdoctoral associates will be supported, as well as undergraduate students with an emphasis on underrepresented groups. Sessions on materials simulation results will be integrated into the annual Sanibel meetings. Software produced will be made available to the wider community through the Materials Computation Center at the University of Illinois.