EAGER: Rung-Reduced Density Functionals for Cost-Capped Ab Initio Molecular Dynamics

EAGER：用于成本上限从头算分子动力学的梯级约简密度泛函

• 批准号: 1515307
• 负责人: Samuel Trickey
• 金额: $ 16.08万
• 依托单位: University of Florida
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-09-01 至 2018-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1515307&HistoricalAwards=false
• 关键词: EAGER; Rung; Reduced; Density; Functionals

NONTECHNICAL SUMMARYThe project involves the development of new exchange correlation potentials that are computationally cheap and would enable ab initio molecular dynamics on larger systems. This unconventional, promising, but risky approach to developing these correlation potentials qualifies the project for EAGER. Context is the growing importance of ab initio molecular dynamics simulations for predictive simulation of novel material properties. This gives detailed insight into atomic-scale processes that affect bulk properties, even for experimentally inaccessible conditions. The PI's constraint-based (not fitted to data bases) functionals reduce molecular binding energy errors by a factor of two compared to the most popular methods with vibration frequencies, ionization potentials, etc. An enticing and highly unusual route to be explored in this proposal is provided by a recent major advance based on the first fully non-empirical orbital-free kinetic energy functional at the GGA rung. There are four areas of broader impact. Success will transform ab initio molecular dynamics simulations of complicated, challenging materials, particularly on department and group-level machines by keeping standard Kohn-Sham cost-scaling as low as possible. Certain fundamental aspects of pure Density Functional Theory itself (beyond simulations) will be substantially clarified, including spacing of the Jacobs' ladder rungs. International science will be enhanced by involvement with collaborators and their student in Mexico. The postdoctoral scholar will benefit from a high-level academic apprenticeship.TECHNICAL SUMMARYThe project involves the development of new exchange correlation potentials that are computationally cheap and would enable ab initio MD on larger systems. The proposed solution is to approximate exchange-correlation density functionals dependent only on the electron number density and its spatial derivatives, and not explicitly on the Kohn-Sham orbitals. This unconventional, promising, but risky approach qualifies the project for EAGER. Context is the growing importance of ab initio molecular dynamics simulations for predictive simulation of novel material properties. This gives detailed insight into atomic-scale processes that affect bulk properties, even for experimentally inaccessible conditions. The PI's constraint-based (not fitted to data bases) functionals reduce molecular binding energy errors by a factor of two compared to the most popular methods with vibration frequencies, ionization potentials, etc. This "rung reduction" strategy will remove the orbitals from meta-generalized gradient approximation (meta-GGA) functionals to get highly refined GGA functionals. An enticing and highly unusual route to be explored in this proposal is provided by a recent major advance based on the first fully non-empirical orbital-free kinetic energy functional at the GGA rung. There are four areas of broader impact. Success will transform ab initio molecular dynamics simulations of complicated, challenging materials, particularly on department and group-level machines by keeping standard Kohn-Sham cost-scaling as low as possible. Certain fundamental aspects of pure DFT itself (beyond simulations) will be substantially clarified, including spacing of the Jacobs' ladder rungs. International science will be enhanced by involvement with collaborators and their student in Mexico. The postdoctoral scholar will benefit from a high-level academic apprenticeship.

非技术总结该项目涉及开发新的交换关联势，这种势在计算上很便宜，并将使更大系统上的从头计算分子动力学成为可能。这种非常规的、有前途的、但有风险的方法来开发这些关联势，使该项目有资格成为EIGER。背景是从头算分子动力学模拟对于新材料性质的预测模拟的重要性日益增长。这提供了对影响整体性质的原子尺度过程的详细洞察，即使在实验上无法达到的条件下也是如此。与最流行的振动频率、电离势等方法相比，PI的基于约束的(不适合于数据库)泛函将分子结合能误差减少了两倍。这一建议中探索的一条诱人和非常不寻常的路线是由最近的一项重大进展提供的，该进展基于GGA梯级的第一个完全非经验的无轨道动能泛函。有四个领域会产生更广泛的影响。Success将改变复杂、具有挑战性的材料的从头分子动力学模拟，特别是在部门和小组级别的机器上，通过将标准的Kohn-Sham成本比例保持在尽可能低的水平。纯密度泛函理论本身的某些基本方面(超越模拟)将被实质地阐明，包括雅各布斯梯级的间距。国际科学将通过与合作者及其在墨西哥的学生的参与而得到加强。博士后学者将从高水平的学术见习中受益。技术总结该项目涉及开发新的交换关联势，这种势在计算上很便宜，并将使从头算MD能够在更大的系统上实现。所提出的解是近似交换相关密度泛函，只依赖于电子数密度及其空间导数，而不是显式地依赖于Kohn-Sham轨道。这种非常规的、有前途的、但有风险的方法使该项目有资格成为渴望的项目。背景是从头算分子动力学模拟对于新材料性质的预测模拟的重要性日益增长。这提供了对影响整体性质的原子尺度过程的详细洞察，即使在实验上无法达到的条件下也是如此。与最流行的振动频率、电离势等方法相比，PI的基于约束(不适用于数据库)的泛函将分子结合能误差减少了两倍。这种“阶跃减少”策略将从超广义梯度近似(META-GGA)泛函中去除轨道，得到高度精炼的GGA泛函。这一提议中将要探索的一条诱人和极不寻常的路线是由最近的一项重大进展提供的，该进展基于GGA梯级的第一个完全非经验的无轨道动能泛函。有四个领域会产生更广泛的影响。Success将改变复杂、具有挑战性的材料的从头分子动力学模拟，特别是在部门和小组级别的机器上，通过将标准的Kohn-Sham成本比例保持在尽可能低的水平。纯DFT本身的某些基本方面(不限于模拟)将被充分阐明，包括雅各布斯梯级的间距。国际科学将通过与合作者及其在墨西哥的学生的参与而得到加强。博士后学者将受益于高水平的学术学徒制。