Abstract
Today's electronic structure theory depends equally upon density functional theory (DFT) and wavefunction theory (WFT). The interconnections between the two has long been of interest, but has not been developed explicitly enough to enable the best of both worlds to be used to obtain improved results. ab initio dft has been pursued at QTP over a few years as a partial solution. ab initio dft combines elements of correlated WFT, by taking orbital dependent functionals from coupled-cluster and many-body perturbation theory with optimised potential ideas to offer a seamless connection. The emphasis has been on the correlation potentials used in Kohn–Sham theory rather than the functionals themselves, as the potential is more sensitive to various approximations. This contribution summarises that work and suggests some future developments.
Acknowledgements
As the citations indicate, our efforts in ab initio dft have covered a few years and depended upon the exceptional talents of several students and postdocs at QTP. These include Nevin Oliphant, who first added the numerical integration step to ACES II for DFT and showed that a post-HF correlation correction ala Becke, based on BLYP provided solutions competitive with CCSD(T) Citation57. Stan Ivanov joined our efforts from the DFT world to help get us started Citation58. He was followed by So Hirata, Irek Grabowski and Victor Lotrich whose roles were pivotal (see their contribution to this special issue). We also benefitted from several visits by Jim Talman. My students who were most important in enabling us to understand the theory and do rather involved OEP calculations were Igor Schweigert and Denis Bokhan, with some early contributions from Tom Henderson. Very recent work is being done by Prakash Verma. As in all of our work, Ajith Perera had an important role behind the scenes. All our work in this area has been supported by the US Air Force Office of Scientific Research under Dr. Michael Berman. I particularly thank Irek Grabowski and Andrew Teal for permission to show their recent results for correlated potentials () before publication.