CAREER: Efficient DFT-based computational approach for correlated systems

点击翻译按钮获取中文摘要

TECHNICAL SUMMARYThis CAREER award supports theoretical and computational research, and education to develop, validate, and disseminate a new first-principles density-functional-theory-based computational approach that will be able to describe correlated and weakly correlated materials at a higher level of accuracy than existing density-functional-theory-based calculations. This objective will be pursued through a correction to the approximate total energy that selectively acts on 'correlated' electronic states. The new correction functional will be designed to capture specific many-body properties. This functional will be shaped on second-quantization model Hamiltonians that have been formulated and broadly used to study the behavior of correlated electrons. The correction will be incorporated into the density-functional-theory energy functional and will include effective interactions computed from first principles. The extension of computational methods already in use with standard density-functional-theory methods will further expand the descriptive power of the new computational tool. This project is envisioned to proceed along the following steps:1) Formulation of a generalized functional that acts differently on occupied and semi-occupied states. This generalization will be used to study optimal refinements to improve the flexibility and the general applicability of the functional.2) Definition of an efficient and accurate computational method for the run-time calculation of effective electronic couplings from the interaction kernel.3) Development of a general corrective Hamiltonian able to reintroduce, within controlled approximations, selected many-body terms of the electronic energy into approximate exchange-correlation functionals. The new corrective functional is aimed to greatly improve the accuracy and the numerical efficiency of electronic structure calculations enabling the screening and optimization of materials for technological applications. Of particular interest is predicting novel materials for solar cells, catalytic conversion of alkanes, functional metallic alloys and oxides, parent compounds of high temperature superconductors, oxides for advanced electronics, and photo-catalysis.This award also supports educational activities with the aim of making first-principles electronic-structure calculations accessible to students and senior scientists who are not familiar with computational techniques. Specific objectives include: developing a course on how to perform reliable density functional theory calculations accessible to graduate students who work in experimental groups, and developing hands-on computational sessions and integrating them into other courses. The PI will continue to participate in summer schools and tutorials organized by the developers of the first-principles electronic-structure code Quantum-ESPRESSO. NON TECHNICAL SUMMARYThis CAREER award supports theoretical and computational research, and education to develop computational techniques for efficient and accurate modeling of materials that contain electrons that interact with each other particularly strongly. Notable examples of these correlated materials are transition-metal and rare-earth compounds that usually contain electrons that are very localized around a transition metal or rare earth atom. These electrons play an important role in determining essential properties of strongly correlated materials. Correlated materials may appear in a broad spectrum of technological applications that includes: high temperature superconductors, solar cells, catalysts, energy conversion and storage systems, and emerging electronic device technologies that, in addition to charge, exploit magnetic properties of the electron for their operation. Accurate computational modeling of correlated materials is a fundamental challenge; the capability would greatly facilitate the design and optimization of materials with desired properties through both the precise characterization of the microscopic factors controlling their behavior, and the efficient use of computers to screen many candidate materials to predict materials which will be optimally suited for a particular application. Most available predictive computational approaches to perform materials specific calculations are either very computationally expensive or not accurate enough to capture the physical properties of materials with strongly interacting electrons. This award supports developing and implementing new techniques to predict accurately and efficiently the properties of correlated materials. This award also supports educational activities to teach students and senior scientists who are not familiar with computational techniques how to use effectively advanced computational materials modeling tools. Specific objectives include: developing a course on how to perform reliable predictive materials modeling calculations, focusing particularly on graduate students who work in experimental groups, and developing 'hands-on' computational sessions integrating them into other courses. The PI will continue to participate in summer schools and tutorials organized by the developers of the first principles electronic structure code Quantum-ESPRESSO.