J. Woods Halley, Fellow

Numerical Studies of Fluids and Disordered Solids

  This group studied a variety of systems including the electrode-electrolyte interface, amorphous polymers, superfluid helium, and Bose-Einstein condensates using molecular dynamics, and electronic structure methods-both separately and together in direct dynamics methods for the former two, and quantum Monte Carlo and field theoretic models to study the latter. These studies have technical applications to corrosion, solar cell development, batteries, fuel cells, and nuclear waste disposal.

  The simulation of electrochemical interfaces is one of a larger class of materials-related simulation challenges in which one needs to couple calculations spanning about 10 orders of magnitude in length and time scale in order to produce macroscopic predictions. Generally, methods within each scale are available, but robust and reliable methods for coupling one scale to another are not. In the group's work, methods for coupling the electronic structure scale to the atomic scale and the atomic scale to higher length scales were developed. At the electronic scale, both quantum chemical, Hartree-Fock-based methods, and solid-state, density functional methods-while working from the same principles-are used. Results from these electronic scale calculations were then used to parametrize models at the atomic scale, either as force fields for classical molecular dynamics calculations or as self consistent tight binding models. Using these methods, the group studied such features of electrochemical interfaces as electron transfer rates, capacitance, and oxide structure and conductivity. Intermediate scale tight binding methods also made possible new kinds of study of magnetic solids and highly disordered solids, occuring in electrodes but also elsewhere. In the electrochemical studies, it is often also necessary to connect the atomic scale calculations to continuum models in order to make predictions of the experimental quantities of interest.

  The study of polymer electrolytes is motivated by a world-wide search for better ion conducting polymers for use in advanced batteries. The Halley group elucidated the mechanism of ion transport in one of the prototype electrolytes. This led to new directions in the search for better ion conducting polymers at Lawrence Berkeley Laboratory, which recently produced very promising new polymer materials for this application. The polymer simulations, which use classical molecular dynamics at the atomic scale, are connected to longer time scales than are accessible by the traditional methods by a variety of techniques, some of them developed in this group, including a form of renormalized molecular dynamics and an innovative application of the parallel replica method.

  The work on Bose-Einstein condensates was motivated by a search for better experimenal probes of the condensate, particularly in liquid helium four. To evaluate a proposal by this group for a new kind of helium atom scattering experiment to study the condensate, they carrried out both strong coupling variational calculations and weak coupling calculations at the Bogoliubov level. Helium is a strongly coupled fluid, but the weak coupling calculations permit evaluation of the validity of the strong coupling results, because they involve approximations that are better controlled. Also, weak coupling calculations apply directly to a newly discovered set Bose-condensed alkali gasses. The results of recent weak coupling calculations showed that there are transparency effects reminiscent of those conjectured to occur for strongly coupled systems.

  In another category of calculations associated with the condensate studies, members of the group simulated the behavior of pulsed atomic helium beams, which are used in the experiment to study the condensate in liquid helium four. Using a gas dynamics code to model these pulsed low energy helium beams, the workers found a good fit with the experimental results at low source power. The fit implied that significant collisional effects occur in these beams. The code could also be used to simulate signals to be expected in the experiment (not yet carried out) in which the beams will be used to study the condensate in liquid helium four. Very recently, these gas dynamics codes were used in a computationally innovative way to predict the behavior of atomic helium pulses of much higher density, which behave anomalously in the experiments. To simulate the millions of atoms involved, the group (in particular, Aaron Wynveen) wrote a hybrid code in which the molecular dynamics portion was used to determine a boundary condition for the continuum hydrodynamic description of the gas used in most of the volume of the simulated space. This simulation recently produced results quite similar to the anomalous experimental results and suggests an explanation in terms of shock waves in the pulse.