TY - JOUR
T1 - An efficient approximate algorithm for nonadiabatic molecular dynamics
AU - Hanasaki, Kota
AU - Kanno, Manabu
AU - Niehaus, Thomas A.
AU - Kono, Hirohiko
N1 - Funding Information:
K.H. is grateful to Professor Chaoyuan Zhu and Dr. Ling Yue for their hospitality during his visit to their laboratory and for stimulating discussions there. This work is supported in part by JSPS KAKENHI (Grant Nos. JP16H04091, JP18K05022, and JP26810002). Part of the calculations was carried out by using computing resources at the Research Center for Computational Science, Okazaki, Japan.
Publisher Copyright:
© 2018 Author(s).
PY - 2018/12/28
Y1 - 2018/12/28
N2 - We propose a modification to the nonadiabatic surface hopping calculation method formulated in a paper by Yu et al. [Phys. Chem. Chem. Phys. 16, 25883 (2014)], which is a multidimensional extension of the Zhu-Nakamura theory with a practical diabatic gradient estimation algorithm. In our modification, their diabatic gradient estimation algorithm, which is based on a simple interpolation of the adiabatic potential energy surfaces, is replaced by an algorithm using the numerical derivatives of the adiabatic gradients. We then apply the algorithm to several models of nonadiabatic dynamics, both analytic and ab initio models, to numerically demonstrate that our method indeed widens the applicability and robustness of their method. We also discuss the validity and limitations of our new nonadiabatic surface hopping method while considering in mind potential applications to excited-state dynamics of biomolecules or unconventional nonadiabatic dynamics such as radiation decay processes in ultraintense X-ray fields.
AB - We propose a modification to the nonadiabatic surface hopping calculation method formulated in a paper by Yu et al. [Phys. Chem. Chem. Phys. 16, 25883 (2014)], which is a multidimensional extension of the Zhu-Nakamura theory with a practical diabatic gradient estimation algorithm. In our modification, their diabatic gradient estimation algorithm, which is based on a simple interpolation of the adiabatic potential energy surfaces, is replaced by an algorithm using the numerical derivatives of the adiabatic gradients. We then apply the algorithm to several models of nonadiabatic dynamics, both analytic and ab initio models, to numerically demonstrate that our method indeed widens the applicability and robustness of their method. We also discuss the validity and limitations of our new nonadiabatic surface hopping method while considering in mind potential applications to excited-state dynamics of biomolecules or unconventional nonadiabatic dynamics such as radiation decay processes in ultraintense X-ray fields.
UR - http://www.scopus.com/inward/record.url?scp=85059390389&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85059390389&partnerID=8YFLogxK
U2 - 10.1063/1.5046757
DO - 10.1063/1.5046757
M3 - Article
C2 - 30599729
AN - SCOPUS:85059390389
SN - 0021-9606
VL - 149
JO - Journal of Chemical Physics
JF - Journal of Chemical Physics
IS - 24
M1 - 244117
ER -