In Search of an Optimum Reference

With support from the Chemical Theory, Models and Computational Methods Program in the Division of Chemistry, Filipp Furche of the University of California, Irvine (UCI) aims to develop, implement, evaluate, and apply a new family of electronic structure methods of computational chemistry to address critical limitations of current theories. These limitations hamper our ability to predict and simulate accurately energies and properties in areas such as catalysis, drug design, atmospheric chemistry, and quantum materials.

The ‘optimum reference’ (OR) methods proposed by Furche promise to minimize by design the ‘reference bias’ of current methods and are expected to extend the scope of electronic structure calculations to hitherto elusive systems and properties. The methods developed in this project will be made available to the public through the Turbomole quantum chemistry software.

The program will support education and workforce readiness through an interdisciplinary program in quantum science at UCI and a high school outreach program aimed at students who otherwise would have limited opportunities to participate in research.

This project builds on and aims to generalize previous developments such as orbital optimization and Brueckner methods, as well as self-consistent random phase approximation methods. While the development of ever more sophisticated correlation treatments in quantum chemistry has produced exceedingly accurate methods for "well behaved" systems, the Achilles' heel of virtually all electron correlation methods in use today is their unphysical dependence on a reference state, i.e., a qualitatively accurate, and readily computable approximation to the physical quantum state of interest.

Furche's work aims to develop stable OR methods guided by four propositions: (i) The OR should minimize the distance to the physical state of interest in a well-defined metric, (ii) have a well-defined physical meaning and give rise to a model chemistry, (iii) the total energy should be stationary at the OR, (iv) the correlation energy as a functional of the reference must be amenable to explicit, controlled approximations. The resulting methods will be assessed for their ability to capture magnetic and spin properties of open-shell d- and f-element compounds systems that are particularly challenging for available approaches.

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.