NCSI: MD Simulations of Defect Dynamics & DFT Calculations for Photovoltaics: Visualization & Analysis

Status	Completed
Mentor Name	Stephen Harnish
Mentor's XSEDE Affiliation	Research and Education Allocations: Have participated in NCSI workshops since 2003 and served as EMPOWER mentor 7 times since 2017
Mentor Has Been in XSEDE Community	4-5 years
Project Title	MD Simulations of Defect Dynamics & DFT Calculations for Photovoltaics: Visualization & Analysis
Summary	Students assist mathematics, computer science and physics faculty in visualizations and analyses of molecular dynamics simulations and density functional theory calculations. The two nano-scale applications are tracking acoustically-controlled defect transitions and simulating thin-film materials to aid photovoltaic technologies. Students adapt code developed by previous student and faculty researchers, and test them on a local Linux cluster before analysis of large-scale production runs on OSC's Owens cluster and DFT calculations using VASP on the University of Toledo's Antec3.
Job Description	The first ongoing research project uses cluster computing for multiple MD simulations over large parameter spaces of varied temperatures, pressures, interatomic potentials, and acoustic standing waves within highly symmetric crystalline lattices. One on-going research line tests theoretical and kinetic Monte Carlo predictions that vacancies and other point defects tend to locate near acoustic wave anti-nodes. This prediction offers an analog of Born's principle. Such classical analogs of this fundamental principle of QM are of interest to researchers of analog models of quantum gravity. The second sets of MD and DFT runs extend previous simulations of thin-film materials of interest in developing new photovoltaic technologies. These are based on work of colleagues at regional universities, especially The University of Toledo. As part of this spring 2022 XSEDE EMPOWER research, the student(s) will adapt previous LAMMPS, C and Python code for FCC lattices with LJ, SW and broader classes of radially-dependent interatomic potentials. The final phases of the work expose the student researchers to visualization tools such as Paraview and Ovito. While scientific motivations are partly theoretical, through this exposure to visualization tools, parallel algorithms and core concepts of solid-state physics and photovoltaic technologies, the student(s) will also gain a foundation for research and development in materials science & engineering.
Computational Resources	Besides resources of faculty and those gleaned from SC'11 and SC'12 education programs, as well as BWSIP training materials, the student researchers will utilize resources available at the HPC University. Besides VASP runs on the University of Toledo's Antec3, an education allocation from the OSC will be used for most production runs.
Contribution to Community	By providing early training experiences for young computational scientists, this project prepares them for more advanced work and independent research, and hence contributes back to the XSEDE and other computational science communities.
Position Type	Learner
Training Plan	Because of the strengths of the proposed students, I would primarily start by exposing them to the excellent tutorial resources of the OSC. I would also direct them to the BW Petascale Institute and HPC University resources while they adapt past code and contribute new programming tools for the larger project.
Student Prerequisites/Conditions/Qualifications	The student researchers must have strong backgrounds or demonstrated abilities in mathematics and/or computer programming, core knowledge in physics, and basic knowledge of C, FORTRAN or Python. Candidates should have the ability or motivation for self-directed code adaptation, and maturity conducive to self-directed as well as collaborative research. The proposed students would be ideal assistants for these research projects.
Duration	Semester
Start Date	01/14/2022
End Date	04/29/2022