TY - JOUR
T1 - Nonadiabatic dynamics study of methaniminium with ORMAS
T2 - Challenges of incomplete active spaces in dynamics simulations
AU - West, Aaron C.
AU - Barbatti, Mario
AU - Lischka, Hans
AU - Windus, Theresa L.
N1 - Funding Information:
ACW and TLW are indebted to Michael W. Schmidt and Mark S. Gordon for help in using the capabilities of GAMESS and MCSCF. This material is based upon work (ACW) supported by the National Science Foundation under Grant No. OISE-0730114 for the Partnerships in International Research and Education (PIRE). This research (TLW) is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and Biosciences through the Ames Laboratory. The Ames Laboratory is operated for the U.S. Department of Energy by Iowa State University under Contract No. DE-AC02-07CH11358. Support was also provided to HL by the Robert A. Welch Foundation under Grant No. D-0005 . TLW acknowledges computing resources purchased through funds provided by Ames Laboratory and Iowa State University.
PY - 2014/7/15
Y1 - 2014/7/15
N2 - The issues associated with the use of multiconfigurational wave function methods, especially those that use active spaces with less than a full-valence orbital space or incomplete excitations between multiple active spaces or both, in nonadiabatic dynamics is carefully examined. Toward this end, the dynamics package NEWTON-X and the electronic structure suite GAMESS are interfaced for nonadiabatic and adiabatic "on-the-fly" dynamics simulations. In particular, this interface allows for the first study of nonadiabatic dynamics with the occupation restricted multiple active space (ORMAS) approximation, which is unique to GAMESS. Several dynamics simulations using methaniminium as an example were performed with various computationally feasible active space choices (or schemes) in order to test the qualitative accuracy and relative expense of different active space choices. Overall, for ORMAS orbital subspace divisions, schemes with no excitations between orbital subspaces give qualitatively incorrect state populations while schemes with single excitations between orbital subspaces recover the qualitatively correct state populations relative to the CASSCF level of theory at a lower computational expense. However, larger number of excitations between ORMAS subspaces cause more issues with orbital integrity within the active spaces, especially at the initiation of the trajectories. In addition, all active spaces show a large number of trajectories with an orbital integrity issue that is not caught by the energy conservation checks. So, trajectories must be monitored carefully and more overall trajectories are likely to be needed to obtain quantitative statistical information.
AB - The issues associated with the use of multiconfigurational wave function methods, especially those that use active spaces with less than a full-valence orbital space or incomplete excitations between multiple active spaces or both, in nonadiabatic dynamics is carefully examined. Toward this end, the dynamics package NEWTON-X and the electronic structure suite GAMESS are interfaced for nonadiabatic and adiabatic "on-the-fly" dynamics simulations. In particular, this interface allows for the first study of nonadiabatic dynamics with the occupation restricted multiple active space (ORMAS) approximation, which is unique to GAMESS. Several dynamics simulations using methaniminium as an example were performed with various computationally feasible active space choices (or schemes) in order to test the qualitative accuracy and relative expense of different active space choices. Overall, for ORMAS orbital subspace divisions, schemes with no excitations between orbital subspaces give qualitatively incorrect state populations while schemes with single excitations between orbital subspaces recover the qualitatively correct state populations relative to the CASSCF level of theory at a lower computational expense. However, larger number of excitations between ORMAS subspaces cause more issues with orbital integrity within the active spaces, especially at the initiation of the trajectories. In addition, all active spaces show a large number of trajectories with an orbital integrity issue that is not caught by the energy conservation checks. So, trajectories must be monitored carefully and more overall trajectories are likely to be needed to obtain quantitative statistical information.
KW - GAMESS
KW - MCSCF
KW - NEWTON-X
KW - Nonadiabatic
KW - ORMAS
KW - Surface hopping
UR - http://www.scopus.com/inward/record.url?scp=84903879699&partnerID=8YFLogxK
U2 - 10.1016/j.comptc.2014.03.015
DO - 10.1016/j.comptc.2014.03.015
M3 - Article
AN - SCOPUS:84903879699
SN - 2210-271X
VL - 1040-1041
SP - 158
EP - 166
JO - Computational and Theoretical Chemistry
JF - Computational and Theoretical Chemistry
ER -