Three principal questions are addressed in this work: (1) What are the energetics of a given system, (2) what are the rates of transitions, and (3) what is the atomic structure at a given temperature. Methods for carrying out computations relevant to these questions are being developed as well as new techniques for analysis and interpretation of the numerical output. The emphasis is on techniques applicable to large, realistic systems.
Studying the energetics of an atomic system generally requires calculating the electron wavefunction. By using density functional theory to describe the electronic interactions, metal and semiconductor systems including up to several hundred atoms are being studied by computations on computer clusters. Examples of current projects include hydrogen atom binding and diffusion in crystals and at crystal surfaces. In some cases the energetics are simple enough that a potential energy function can be developed to describe the energetics. This allows simulations of larger systems and the study of dynamics on a longer timescale. The development of such potential functions, for example for water, is an ongoing effort.
Studies of transitions involve identification of likely paths the system can take from a given initial state to a final state. Evaluation of the free energy barrier to the transition then yields an estimate of the rate constant for the transition within transition state theory. Techniques for determining minimum energy paths and free energy barriers to transitions in classical and quantum systems are being developed and applied studies of chemical reactions and diffusion of atoms on surfaces of solids. A large simulation has been carried out using these new methods in a study of the dissociative adsorption of hydrogen molecules on a metal surface where the effect of quantum mechanical tunneling and zero point energy on the sticking probability was assessed.
Systems chosen for simulation studies in the group tend to fall into one of two categories: Either extensive, high quality experimental measurements have been taken by other workers leading to interesting questions about the microscopic interpretation of the results, or current experimental techniques are unable to provide microscopic information at all because of the complexity of the system. An example of the first is metal crystal growth from vapor, where observations of surprisingly smooth and regular crystal growth at low temperature have been made. The growth of Pt and Cu from vapor has even been observed to improve as the temperature of the crystal is lowered. This is counterintuitive since thermally activated diffusion is required for atoms to find their optimal sites in the crystal lattice. Computer simulation studies have identified several important atomic scale processes in metal crystal growth which help explain these observations. A major goal is to understand how growth conditions can be chosen to optimize the quality of crystals and thin films used in various technological applications.
Amorphous materials and liquid-solid interfaces are an example of the second category. Computer simulations provide insight experimental studies currently cannot give into the microscopic structure and dynamics and thereby contribute to an understanding of the observed macroscopic properties of these materials. An ongoing study involves simulations of water at metal surfaces and growth of ice overlayers, as well as the growth of metal overlayers on oxides.