Real Space Method for First Principles Calculation of Electronic, Magnetic, and Spectroscopic Properties of Halogen-Bridged Transition Metal Compounds

9520319威尔金斯这个奖项是通过1995财政年度计算方法真实的材料竞争。 资金通过材料研究、数学科学和高级科学计算司以及数学和物理科学局多学科活动办公室提供。 该研究将算法和代码开发与材料理论相结合，以解决计算突破可以产生巨大回报的重要材料问题。 因此，该研究涵盖了高性能计算和通信以及先进材料和加工的战略举措。 卤素桥连的过渡金属直链化合物在它们的基态和能隙方面具有宽的化学可调性。 动量空间计算表明，这些材料可能具有技术重要性。 真实的空间计算将允许研究更大的系统，这是由于从按所考虑的粒子数的立方缩放的计算过渡到按粒子数线性缩放的计算。 更大的系统将能够计算更精确的光学特性，并能够精确处理缺陷，包括极化子和孤子，这将允许光学和磁性的实际计算。 计算方面的工作包括专门的并行稀疏矩阵算法的开发;使用和开发负载平衡算法;快速傅立叶变换和其他矩阵运算的并行磁盘算法的开发;以及使用高性能Fortran提供的共享数据空间模型。 代码将在各种平台上运行。 %此奖项是通过FY 95计算方法真实的材料竞赛获得的。 资金通过材料研究，数学科学和高级科学计算部门以及数学和物理科学理事会多学科活动办公室提供。 该研究将算法和代码开发与材料理论相结合，以解决计算突破可以产生巨大回报的重要材料问题。 因此，该研究涵盖了高性能计算和通信以及先进材料和加工的战略举措。 这项研究将集中在一类被称为卤素桥接化合物的材料上。 这些化合物可以通过改变它们的结构和组分容易地改性，以实现所需的电子、光学和磁性。 因此，它们是设计材料。目前的计算技术只暗示了这些可能性。 目前的研究将发展计算技术，以处理更现实的系统，这些化合物。 因此，这项研究的成功完成将推进最先进计算机上的计算，将有助于对这些化合物性质的基本理解，并将使一类材料的准确设计能够用于各种应用。 ***

9520319 Wilkins This award is one made through the FY95 Computational Approaches to Real Materials competition. Funding is provided through the Divisions of Materials Research, Mathematical Sciences and Advanced Scientific Computing, as well as by the Office of Multidisciplinary Activities in the Mathematical and Physical Sciences Directorate. The research combines algorithm and code development with materials theory to address important materials problems for which computational breakthroughs can produce large payoffs. As such, the research spans the strategic initiatives in High Performance Computing and Communications and Advanced Materials and Processing. The halogen-bridged transition-metal linear-chain compounds are unique in their wide chemical tunability with respect to their ground state and their energy gap. Momentum-space computation has indicated that these materials may have technological importance. Real space computation will allow much larger systems to be studied due to transitioning from calculations scaling as the cube of the number of particles being considered to those scaling linearly with the number of particles. The larger systems will enable more accurate optical properties to be calculated and enable accurate treatment of defects, including polarons and solitons, which will allow realistic calculations of optical and magnetic properties. Computational aspects of the work include development of specialized parallel sparse matrix algorithms; use and development of load-balancing heuristics; development of parallel disk-based algorithms for fast Fourier transforms and other matrix operations; and, use of shared data-space model such as provided by High Performance Fortran. Codes will be run on a variety of platforms. %%% This award is one made through the FY95 Computational Approaches to Real Materials competition. Funding is provided through the Divisions of Materials Research, Mathematical Sciences and Advanced Scientific C omputing, as well as by the Office of Multidisciplinary Activities in the Mathematical and Physical Sciences Directorate. The research combines algorithm and code development with materials theory to address important materials problems for which computational breakthroughs can produce large payoffs. As such, the research spans the strategic initiatives in High Performance Computing and Communications and Advanced Materials and Processing. The research will focus on a class of materials called halogen-bridged compounds. These compounds can be readily modified by varying their structure and components in order to achieve desired electronic, optical and magnetic properties. Thus, they are Odesigner materials.O Present computational techniques only hint at these possibilities. The present research will develop computational techniques in order to deal with much more realistic systems of these compounds. Consequently, successful completion of this research will advance computation on state-of-the-art computers, will assist in fundamental understanding of the properties of these compounds and will enable the accurate design of a class of materials for a variety of applications. ***