MRI: Acquisition of a High Performance Computing Cluster for Undergraduate Chemistry Research and Teaching by the Midwest Undergraduate Computational Chemistry Consortium (MU3C)

Hope College, Holland MI

Investigators

Abstract

The acquisition of a high-performance computing system by the Midwest Undergraduate Computational Chemistry Consortium (MU3C) will enable undergraduate students and their faculty mentors at a diverse set of 17 colleges and universities to use computational chemistry to advance our understanding of multiple basic and applied phenomena in terms of the properties of individual atoms and molecules. Fundamental research topics will advance prediction of environmental chemistry, explanation of reaction mechanisms, design of new materials and new catalysts, understanding of spectroscopy and excited state dynamics, and the development of new molecular dynamics methods. During the three years of the grant, MU3C faculty will train over 140 undergraduates to conduct computational chemistry research and to communicate their results at professional scientific conferences. Over 5000 undergraduates will use the computing system as part of their study of general, organic, inorganic, physical, and biological chemistry. MU3C includes institutions serving large numbers of first-generation college students, students eligible for Pell grants, and underrepresented minority students. The project will provide 40 new CPU-based compute nodes and four new GPU-based compute nodes (32 GPUs) combined with 40 existing CPU-based compute nodes to yield a computer cluster well-suited to the broad range of research undertaken by MU3C's collaborative teams of faculty and undergraduate students. The new CPU-based compute nodes will have 256 GB of RAM, compared to 24 GB in the existing CPU-based nodes, enabling a variety of demanding quantum chemical (QC) calculations that cannot be executed on the current MU3C cluster. The new GPU-based compute nodes will replace existing GPU-based nodes that, because of their age, will soon be unable to run current versions of computational software. The existing CPU-based nodes are still useful for a variety of modestly-demanding QC calculations. The combined power of all three types of compute nodes will provide enough capacity to support the growing MU3C consortium, which has doubled in the number of research groups since its last MRI award. The research projects undertaken by these groups will contribute to a number of fields as described above. For example, calculations using DLPNO-CCSD(T) and CASSCF/NEVPT theories will aid the design of new catalysts in which expensive 2nd and 3rd row transition metals are replaced by inexpensive 1st row transition metals; calculations using EOM-CCSD(dT) theories will model complex reaction mechanisms to aid prediction of the evolution of atmospheric peroxides; calculations using time-dependent density functional theory will help explain the novel enhanced luminescence of ?magic-sized? quantum dots; and calculations using Monte Carlo methods will aid the design of new zeolite materials for the selective capture and sequestration of atmospheric CO2. Over two dozen different projects will advance our understanding of chemistry while preparing a broad set of undergraduate students for careers in STEM fields. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.