Cloud-based Low-Scaling Quantum Chemistry Simulations for Materials

Predictive simulations are important in the data-driven approaches for materials development (i.e., materials informatics) because it allows for augmenting experimental data, which are usually small and biased. Solid materials, such as heterogeneous catalysts and electrodes, are modeled by quantum mechanical simulations, but the existing commercial offering has limitations in terms of accuracy. To address this issue, this proposal seeks to develop and commercialize a breakthrough software implementation of the so-called hybrid density functional theory and electron correlation models. The key advancements in this work are two-fold: first, it is based on a novel algorithm that uses mixed Gaussian and plane-wave basis functions as an orbital basis, while Sinc functions are used as auxiliary functions to accelerate the computation; second, our software will be integrated with the software libraries that have been developed by the researchers supported by the Department of Energy Advanced Scientific Computing Research program (ASCR) in order to drastically speed up a core step in the algorithm. In Phase I of the proposed work, the novel working equations for hybrid density functional theory will be translated into efficient, parallel algorithms, which will then be implemented into tightly optimized code. The interface between our software stack and the ASCR-funded libraries will be developed. The program will be benchmarked and tested in the heterogeneous cloud environment. The goal in Phase I is to demonstrate efficient hybrid density functional theory calculations at a cost comparable to the conventional pure density functional theory computations. The resulting software will significantly improve the accuracy and throughput of quantum mechanical simulations in data-driven materials development, which would lead to reduced research and development costs for novel materials. This would in turn lead to cheaper consumer prices, benefiting the public and federal government.