Electronic Structure of Condensed Matter on the CM-5

凝聚态的电子结构-拟议的研究旨在发展研究凝聚态的计算方法，并使用这些方法来研究固体和液体的电子性质。主要使用的技术是量子蒙特卡罗和密度泛函理论。QMC的两个主要改进是：找到多电子波函数的准确和有效的表示，以及适合于消除核心电子影响的赝势的发展。这种改进的QMC技术可以用来为DFT提供更好的交换相关能量，这将包括结构不均匀的影响。然后，这些改进的DFT方法可以直接与QMC结果进行比较，以进行精度测试。当对新的DFT方法进行测试时，可以使用它们来生成试函数，从而提高QMC分析的效率。不同计算方法之间的这种耦合被视为拟议研究的核心。使用大规模并行计算机解决这类问题的潜在好处也将被调查。随着这些技术的发展，这些技术将被应用于广泛的凝聚系统的研究，从凝胶到高度相关的材料%拟议的研究是理解和预测真实世界材料性质的长期努力的一部分。他们建议开发互补的计算方法来确定各种固体和液体的电子结构(从而确定其物理性质)。重点是改进和简化现有的技术，以便产生的方法适用于复杂材料。这项研究将包括大规模并行计算机算法的开发和评估。随着改进的计算技术的发展，它们将被应用于与实验物理和材料科学有直接联系的问题-既可以测试方法，也可以理解和预测与实验相关的性质。将研究硅、氢金属表面的凝聚态、液体和非晶态系统等材料。

Electronic Structure of Condensed Matter - The proposed research is aimed at the development of computational methods for studying electronic states in condensed matter and the use of those methods to study the electronic properties of solids and liquids. The primary techniques to be used are quantum Monte Carlo and density- functional theory. Two primary improvements for QMC are contemplated: finding accurate and efficient representations for the many-electron wave function and development of pseudopotentials suitable for elimination of the effects of core electrons. Such improved QMC techniques can be used to provide an improved exchange-correlation energy for DFT which would include the effects of structural inhomogeneity. These improved DFT methods can then be directly compared to the QMC results for tests of accuracy. When the new DFT methods are tested, they can be used to generate trial functions to make QMC analysis more efficient. This coupling between different computational approaches is seen as central to the proposed research. The potential benefit of using massively- parallel computers on this class of problems will also be investigated. Concurrent with their development, these techniques will be applied to the study of a wide range of condensed systems, ranging from jellium to highly correlated materials %%% The proposed research is part of a long-term effort to understand and to predict the properties of real-world materials. They propose to develop complementary computational methods for determining the electronic structure (and hence the physical properties) of a wide range of solids and liquids. The emphasis is to refine and simplify existing techniques, so that the resulting methods are practical for use on complex materials. This research will include development and evaluation of algorithms for massively parallel computers. As improved computational techniques are developed, they will be applied to problems with direct connections to experimental physics and materials science - both to test the methods and to understand and predict experimentally relevant properties. Such materials as silicon, condensed states of hydrogen metal surfaces, and liquid and amorphous systems will be studied.