Computer simulation has been firmly established as a "third way" - complementing experiment and theory - in the study of molecular systems.  Using state-of-the-art high-performance computers and theoretical models, it is now possible to create highly accurate and detailed atomic-level "movies" of the structure and dynamics of complex systems involving millions of atoms.   The simulated molecular dynamics can be used to address problems of great importance to chemistry, physics, and biology, including the exploration of properties of novel materials, the function of complex molecular motors, and the aggregation of proteins underlying disease.

A fundamental property of materials and biological systems is the process of "phase change," in which a system maintains the same molecular composition but displays dramatic change in its structure and physical properties. Examples include the freezing of liquid water to solid ice, the aggregation of soluble proteins to form assembled fibrils associated with disease, and the organization of disordered fats into large-scale, ordered structures such as cell membranes.

At this time, the computer simulation of systems exhibiting "phase change," involving large-scale molecular structures (such as membranes and protein assemblies) and dramatic structural reorganization (such as water freezing), presents outstanding challenges to existing simulation approaches.  The research supported by this award was designed to address these outstanding challenges and provide new computational models and methods for exploring diverse complex molecular systems.

Over the three years of this award, our research developed the generalized Replica Exchange Method (gREM), as an effective approach for the computer simulation of strong phase change in complex molecular systems. The central idea behind gREM is to introduce a "bias" in the sampling of possible structures of the system in a way that allows one to explore not only stable states (such as a liquid or solid) but also "metastable" and unstable states (such as water crystals nucleating growth on a pathway to ice formation, bubbles forming in liquids on the way to a boiling transition to vapor, and small, short-lived "oligomeric" aggregates of protein that seed the growth of large protein fibril assemblies).  As such, gREM allows for the simulation of stable phases as well as the transitions between them.  The key ingredients of gREM are parameterized effective sampling weights that bias the simulation and smoothly join stable phases. The research led to a straightforward method for the application of this inventive computational approach, providing a sure way to achieve efficient computer simulation of diverse molecular systems.

Various applications have been explored, including studies of the liquid-to-liquid phase change involved in metallic glass formation, the mechanism of vapor-liquid transition in liquids including water, the surprising behavior of water freezing when confined to a "nanoslit," and the organizational transitions of peptides in membranes and micelles.   In addition, a highly efficient method for statistical analysis of the "big data" generated by these large-scale simulations was developed, and has already found broad application in the field allowing for the comprehensive characterization of thermodynamic and structural properties of materials and biomolecules.

This award also supported an ambitious outreach program carried out in collaboration with the Pinhead Institute, a non-profit organization devoted to K-12 science education and outreach in Southwestern Colorado.  The outreach involved the promotion of "molecular ...

Please report errors in award information by writing to: awardsearch@nsf.gov.