Large Scale Parallel Software for Realistic Models of Conductive and Nonlinear Optical Polymeric Materials

Theoretical chemistry techniques used to provide fundamental understanding of the molecular requirements for achieving conductive and nonlinear optical properties in polymeric materials are implemented in large scale parallel software enabling investigations on realistic models. The algorithms are based on techniques for simultaneous optimization of the electronic and nuclear coordinates. Linear response of a molecular system is used to obtain a direct expression for the conductivity tensor. Nonlinear response theory (quadratic, cubic, ...) allows for the calculation of hyperpolarizabilities. We will exploit joining different dynamics based methods and design and plan development of a Quantum Theory Integral Package (QTIP) that would make programming integral evaluations comparable to programming linear algebra operations using LAPACK. Previously validated method and software, ENDyne a dynamics approach based on coherent states and production software by parallelization. Phase I will focus on parallelization and efficient use of memory for molecular integral evaluation. The resulting software will consist of compatible integrated modules applicable to a wide class of problems and be well structured to link to the QTIP library and lend itself to generalizations during later phases of this work. OPTION: None submitted.