Predictive Modeling and Machine Learning for Functional Nanoporous Materials Consortium/End-Station

PI J. Ilja Siepmann
Functional Nanoporous Materials

Functional Nanoporous Materials. Image: J. Ilja Siepmann

An interdisciplinary, collaborative team will use predictive hierarchical modeling and machine learning to accelerate the discovery and design of materials for a variety of energy-related applications and to advance scientific and technological capabilities with innovative discoveries. The research objectives of this ALCC proposal are aligned with the goals of the DOE-funded Nanoporous Materials Genome Center (NMGC). Research focus is on four topics: 

  • Hierarchical screening and machine learning for adsorption in nanoporous materials aimed at the discovery of materials with superior performance, including the development of transferable force fields for aluminosilicates and aluminophosphate zeotypes enabling future hierarchical screening studies, and the development of machine-learned models for the prediction of mixture isotherms from single-component isotherms. 
  • First principles Monte Carlo calculations for adsorption systems aimed at improving characterization of catalytic acid sites through modeling of reactive adsorption of propene and amines in acidic zeolites. 
  • First principles simulations for nanoparticle assembly aimed at understanding hydration forces and providing a molecular-scale view of the double layer that govern nanoparticle assembly and synthesis. 
  • Development of periodic wave function in density functional theory embedding methods aimed at improving accuracy and efficiency of predictive modeling for the selective oxidation of alkanes to alcohols in metal-organic frameworks. 

Work enabled by prior ALCC allocations has led to the award of one US patent for the ethanol purification.