Simulation of Nanostructures on Parallel Architectures

在计算机模拟的帮助下，纳米结构的设计和小型化可以取得重大进展，特别是考虑到并行算法和计算机体系结构的最新发展。最广泛使用的纳米结构是外延生长的半导体器件，然而嵌入在多孔材料中的微团簇也显示出相当大的技术潜力。通过计算机模拟，研究纳米级器件和嵌入的微团簇在沸石中的行为。嵌入材料将包括IV类和二进制I-VII、II-VI和III-V材料。分子动力学、动态模拟退火、量子分子动力学和变分蒙特卡罗方法将用于研究纳米结构的性质和过程。此外，新的模拟技术，如模拟进化，遗传算法和变分分子动力学将被开发来研究这些量子受限系统。提出的仿真工作将通过在SIMD（单指令多数据）或MIMD（多指令多数据）机器上执行它们来利用固有的并行性。

Significant advances in the design and miniaturization of nanostructutes can be made with the aid of computer simulation, especially in light of recent developments in parallel algorithms and computer architectures. The most widely used nanostructures are epitaxially grown semiconducting devices, however embedded microclusters in porous materials also show considerable technological potential. By means of computer simulation, the behavior of nanoscale devices and embedded microclusters in zeolites will be studied. The embedding materials will include group IV and binary I-VII, II-VI and III-V materials. Molecular dynamics, dynamical simulated annealing, quantum molecular dynamics, and variational Monte Carlo methods will be used to investigate the properties and processes in nanostructures. In addition, new simulation techniques such as simulated evolution, genetic algorithms, and variational molecular dynamics will be developed to study these quantum confined systems. The proposed simulation work will exploit the inherent parallelism by performing them on SIMD (single instruction multiple data) or MIMD (multiple instruction multiple data) machines.