Skip to Main Content

Title ImagePublic Abstract

 
Collapse

DE-SC0021058: Feasibility of a Novel Density Functional Method Outside the Kohn-Sham Framework for Modeling Global Potential Energy Surfaces of Molecular Chemical Reactions

Award Status: Inactive
  • Institution: The Pennsylvania State University, University Park, PA
  • UEI: NPM2J7MSCF61
  • DUNS: 003403953
  • Most Recent Award Date: 08/25/2021
  • Number of Support Periods: 2
  • PM: Holder, Aaron
  • Current Budget Period: 09/01/2021 - 08/31/2022
  • Current Project Period: 09/01/2020 - 08/31/2022
  • PI: Knizia, Gerald
  • Supplement Budget Period: N/A
 

Public Abstract

Feasibility of a novel density functional method outside the Kohn-Sham framework for
modeling global potential energy surfaces of molecular chemical reactions


Gerald Knizia, The Pennsylvania State University (Principal Investigator)

We aim towards the construction of computational methods capable of modeling the global potential energy surface (PES) of small molecules, molecular ions, and radicals — including the parts of the PES which correspond to chemical reactions, and the reaction paths connecting reactants, intermediates, and reaction products. This goal may seem humble at first glance, but for chemical systems with more than about six atoms in total, at present there are no established theoretical methods capable of simulating such systems reliably, not even for small molecules in the gas phase which are electronically benign. This restriction severely hampers our ability to model and control chemical processes under harsh conditions.

The central issue is that such systems require an accurate description of all feasible transformations of their electron systems, including of processes (e.g., homolytic bond breaking) which established electronic structure methods cannot handle adequately. In particular, in practice these are inaccessible to Kohn-Sham Density-Functional Theory (DFT), which earned the 1998 Nobel Price in chemistry, and which is nowadays ubiquitous in theoretical and computational chemistry, even despite this limitation.

With the goal of constructing a method capable of modeling such chemical systems, we here pursue a novel approach towards a Multi-Configuration (MC) DFT outside the traditional frameworks of Kohn-Sham theory and other methods of coupling wave function theory with DFT. Rather than being a complete active space (CAS) method, the proposed DOCI-DFT employs a special restricted form of the active space wave function, called Doubly-Occupied CI (DOCI). This wave function form is sufficient to describe not only heterolytic, but also homolytic bond dissociation processes at the zeroth order (i.e., as active space wave function). There are no other standard mean field methods which can do so. Additionally, these DOCI wave functions are capable of describing exact ground states of certain classes of interacting Hamiltonians. We will show that the latter property allows for a consistent formal theory in full analogy to Kohn-Sham DFT, which provides a clean variational principle for a DOCI-based self-consistent MC-DFT. Additionally, we will argue that the approach affords an exact determination of its exchange correlation functional in the local density approximation (LDA). Combined with highly efficient coupled cluster-type (pCCD/pCCDQ) approximations to the DOCI active space wave function, this DOCI-DFT should provide an efficient and accurate description of chemical processes including homolytic and dissociative reaction channels, and should be particularly suitable for computing geometries and reaction paths.

If successful, the pursued method will significantly advance theoretical methodology for computationally modeling chemical processes in nature, engineering, and industry, particularly under harsh conditions characterized by high temperature, high pressure, strong electric fields, or the presence of ionizing radiation. Examples include combustion processes (e.g., of fossil fuels in engines and turbines) and processes surrounding electric arc discharges (e.g., in power electronics), but also processes in industrial chemical reactors or nuclear reactors.



Scroll to top