This award supports computational and theoretical research and education on the development and applications of a new quantum Monte Carlo method to answer materials-specific questions, focusing on solids ranging from strongly correlated transition metal compounds to moderately correlated prototypical semiconductors to fundamental models such as the electron gas and multi-band Hubbard-like models. The phaseless auxiliary-field quantum Monte Carlo method was invented by the PI and collaborators. Recent efforts have developed it into one of the most accurate ground-state many-body methods available for electronic structure calculations. The PI will apply this method to solids, and continue its development. Questions that will be addressed include electronic structure and excited states calculations in solids, computations of electronic density and ionic forces, the use of pairing wave functions and calculations of magnetic order and unconventional phases.
The phaseless auxiliary-field quantum Monte Carlo method offers a new, general framework for many-body calculations in lattice models of correlated systems as well as in real materials. Implemented algorithms on parallel computers have excellent scalability. The research aims to take advantage of the new sustained petascale computing facilities to carry out breakthrough calculations.
The PI will continue efforts to integrate research with education and outreach activities, foster interdisciplinary collaboration in computational science education and research on campus, and play an active role in training students, post-docs, and senior researchers through summer and winter schools, and workshops.
NON-TECHNICAL SUMMARY
This award supports computational and theoretical research and education to advance our understanding of materials properties, and improve our ability to do predictive calculations on computers. Interesting effects and new phenomena in materials can originate from the complicated correlations in the motion of the electrons that result from the interactions among them. A quantum mechanical description of the electrons in a material is required to capture the intricate ballet of the electrons and its consequences for the properties of the material.
This project supports research to develop a promising computer simulation technique pioneered by the PI, one that can describe materials down to the constituent electrons and capture their correlated motion. The PI's method offers new opportunities for more accurate and predictive computer simulations of materials and new states of matter. The PI aims to advance the study of materials with strongly interacting electrons, establish accurate benchmark calculations, and enhance the capabilities of materials simulation based on fundamental quantum mechanics. The theoretical framework and techniques developed may also have impact on chemistry and other branches of physics.
The PI will continue to integrate research with education and outreach activities by mentoring both undergraduate and graduate students in research, incorporating materials from this project into new courses, reaching out to minority students, and fostering interdisciplinary collaboration in computational science education and research on campus. In the larger scientific community, the PI will continue to play an active role in training students, post-docs, and senior researchers through schools and workshops, and developing software and tutorials for hands-on learning of new theoretical and computational approaches.
Understanding and predicting materials properties require robust and reliable calculations at the most fundamental level. Often the desired effects originate from electron interaction, and small errors in its treatment can result in crucial and qualitative differences in the calculated properties. While enormous progress has been achieved in recent years, spurred in part by rapid advances in computing power, there are many systems of strong fundamental and technological interests for which we do not have reliable computational approaches. The focus of this project is to develop computational capabilities for accurate and reliable calculations in quantum matter, and to apply them to important problems in condensed matter physics. Major accomplishments in this funding cycle include: new method for many-body excited-state and qusi-particle band-structure calculations in solids; the development of a post-processing finite-size correction scheme for many-body calculations in solids that have magnetic order; advances in incorporating Bardeen-Cooper-Schrieffer (BCS) trial wave functions in the the auxiliary-field quantum Monte Carlo (AFQMC) framework; with the BCS AFQMC development, exact calculations to determine the so-called Bertch parameter in the unitary Fermi gas, which was in excellent agreement with a simultaneous experimental measurement in ultracold atomic gases; the study of magnetic orders in quantum system including itinerant ferromagnetism and antiferromagnetic spin-density waves; elucidation of the energetics and magnetic states of cobalt adsorption on graphene; development of a finite-T method for bose-fermi mixtures; formulations of symmetry preservation and constraint release in AFQMC; study of exotic pairing states in spin-imbalanced fermion systems and their connection with magnetic order; and pseudopotential-free calculations in solids with a frozen-orbital formalism. Methods that we have developed have been adopted in other disciplines such as nuclear physics. In education, our goal was to integrate the research projects above with curriculum development and mentoring of both undergraduate and graduate students, and postdoctoral researchers. Activities include outreach to local schools, training of young researchers in the scientific community by organizing and teaching at international summer/winter schools, development of open source hands-on learning softwares and pedagogical materials for the research projects, conference organization and national and international collaborative research team-building. Many of the research projects involved large-scale computing at national supercomputing facilities, providing an invaluable opportunity to train the next generation of researchers with expertise in high-performance computing.
