Normal coordinate analysis (NCA) is a useful tool for studying the dynamical and equilibrium properties of polymer, biological, and other systems. NCA on systems with hundreds of thousands of atoms has two major challenges. First, the extreme sensitivity of the Hessian matrix to even small changes in structure near mechanical Equilibrum creates in the noumeral results the presence of imaginery vibrational frequencies and the mixin of modes. By computing a Hessian matrix averaged over a molecular dynamics trajectory rather than at a single configuration, it is possible to avoid these unphysical results. This method has been successfully used for systems as large as 36,000 atoms. Second, NCA requires the diagonalization of very large sparse matrices, limiting the size of system that can be treated due to the computational effort. By using eigenvalue binning techniques, it is possible to reduce the computational effort by two or three orders of magnitude while retaining almost the same accuracy for the computation of thermodynamic properties as full diagonalization.
|Presenter:||Robert Tuzun (Faculty)|
|Time:||3:45 pm (Session V)|