Modeling

Efficient searches through chemical space with alchemy

with Anatole von Lilienfeld, Uni-Basel 
and John R. Kitchin, Carnegie-Mellon University

Current Students: Karthikeyan Saravanan

Ongoing work: We are evaluating the performance of chemical alchemy in accurately predicting descriptors for catalysis. A manuscript on this topic is forthcoming in early 2017.


Improved modeling of chemical interactions in solvent phase

jp-2014-07872d_0010

Current Students: Mitch Groenenboom, Yasemin Basdogan, Nguyen Vo, Ethan Henderson

Overview: Our work in this field aims to unravel complicated reaction mechanisms where processes are not only determined by chemical transformations of intermediates in solution, but also interactions between the intermediates, the solvent, and other species in the solvent (e.g. counter ions from the electrolyte).  

Atomistic potential development for nanoscale simulations

with John Kitchin (CMU)

Current Students: Mitch Groenenboom


Overview: Our work in this field aims to develop computational methods that are accurate at predictive for nanoscale materials modeling and chemical reactions on realistic nanostructured systems containing 10,000s of atoms.  


Published work: In collaboration with John Kitchin's (Carnegie Mellon University-ChemE) group, we have studied how ReaxFF and Behler–Parrinello neural network (BPNN) atomistic potentials should be trained to be accurate and tractable across multiple structural regimes of gold as a representative example of a single-component material. …

Assessments of classical forcefields for conformational searches

with Geoffrey Hutchison (Pitt)


Current Students: Karthikeyan Saravanan

Overview: Classical forcefields are widely used to drive molecular dynamics simulations and for rapidly identifying low-energy confirmations of different molecules.  While such forcefields are generally considered sufficiently accurate for qualitative comparisons, extensive and systematic assessments of their accuracy have not yet been carried out.