Energy landscapes for electronic structure
Orbital-optimized multiple self-consistent-field (SCF) solutions are increasingly being interpreted as mean-field approximations of diabatic or excited electronic states. However, surprisingly little is known about the topology of the electronic energy landscape from which these multiple solutions e...
| Main Authors: | , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
American Chemical Society
2020
|
| _version_ | 1797086905232261120 |
|---|---|
| author | Burton, HGA Wales, DJ |
| author_facet | Burton, HGA Wales, DJ |
| author_sort | Burton, HGA |
| collection | OXFORD |
| description | Orbital-optimized multiple self-consistent-field (SCF) solutions are increasingly being interpreted as mean-field approximations of diabatic or excited electronic states. However, surprisingly little is known about the topology of the electronic energy landscape from which these multiple solutions emerge. In this contribution, we extend energy landscape methods, developed for investigating molecular potential energy surfaces, to investigate and understand the structure of the electronic SCF energy surface. Using analytic gradients and Hessians, we systematically identify every real SCF minimum for the prototypical H4 molecule with the 3-21G basis set, and the index-1 saddles that connect these minima. The resulting SCF energy landscape has a double-funnel structure, with no high-energy local minima. The effect of molecular symmetry on the pathways is analyzed, and we demonstrate how the SCF energy landscape changes with the basis set, SCF potential, molecular structure, and spin state. These results provide guiding principles for the future development of algorithms to systematically identify multiple SCF solutions from an orbital optimization perspective. |
| first_indexed | 2024-03-07T02:28:35Z |
| format | Journal article |
| id | oxford-uuid:a673875a-ef88-49ac-bb85-362d85bacfd3 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T02:28:35Z |
| publishDate | 2020 |
| publisher | American Chemical Society |
| record_format | dspace |
| spelling | oxford-uuid:a673875a-ef88-49ac-bb85-362d85bacfd32022-03-27T02:47:31ZEnergy landscapes for electronic structureJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:a673875a-ef88-49ac-bb85-362d85bacfd3EnglishSymplectic ElementsAmerican Chemical Society2020Burton, HGAWales, DJOrbital-optimized multiple self-consistent-field (SCF) solutions are increasingly being interpreted as mean-field approximations of diabatic or excited electronic states. However, surprisingly little is known about the topology of the electronic energy landscape from which these multiple solutions emerge. In this contribution, we extend energy landscape methods, developed for investigating molecular potential energy surfaces, to investigate and understand the structure of the electronic SCF energy surface. Using analytic gradients and Hessians, we systematically identify every real SCF minimum for the prototypical H4 molecule with the 3-21G basis set, and the index-1 saddles that connect these minima. The resulting SCF energy landscape has a double-funnel structure, with no high-energy local minima. The effect of molecular symmetry on the pathways is analyzed, and we demonstrate how the SCF energy landscape changes with the basis set, SCF potential, molecular structure, and spin state. These results provide guiding principles for the future development of algorithms to systematically identify multiple SCF solutions from an orbital optimization perspective. |
| spellingShingle | Burton, HGA Wales, DJ Energy landscapes for electronic structure |
| title | Energy landscapes for electronic structure |
| title_full | Energy landscapes for electronic structure |
| title_fullStr | Energy landscapes for electronic structure |
| title_full_unstemmed | Energy landscapes for electronic structure |
| title_short | Energy landscapes for electronic structure |
| title_sort | energy landscapes for electronic structure |
| work_keys_str_mv | AT burtonhga energylandscapesforelectronicstructure AT walesdj energylandscapesforelectronicstructure |