NUK - logo
E-viri
Celotno besedilo
Recenzirano
  • Comparison of the results o...
    Liu, Yanli; Cerezo, Javier; Lin, Na; Zhao, Xian; Improta, Roberto; Santoro, Fabrizio

    Theoretical chemistry accounts, 03/2018, Letnik: 137, Številka: 3
    Journal Article

    We consider the nonadiabatic dynamics of internal conversions (ICs) in systems rigid enough to allow a description of the coupled potential energy surfaces (PES) within the harmonic approximation. Through a hierarchical representation of the Hamiltonian, we define a set of sequentially coupled effective modes and obtain reduced-dimensionality models by truncating the sequence of modes. We systematically investigate the predictions on the electronic populations of these models and of a recently proposed mean-field mixed quantum-classical (MQC) approach, where the most important effective modes are treated at the quantum level and the motion of the remaining ones is approximated with a swarm of classical trajectories. As a test case, we consider a linear vibronic coupling (LVC) model for the π π ∗ / n π ∗ IC in thymine. LVC PES are computed both in gas phase and in water to explore the different performance of the investigated methods for different relative stabilities of the coupled PES. Reference full quantum dynamical (QD) results are obtained with the MultiLayer Multiconfigurational Time Dependent Hartree method. We show that reduced-dimensionality models work very well in the ultrafast time scale (< 100 fs). At longer times, they tend to predict smaller differences between π π ∗ and n π ∗ populations than those computed with full QD simulations because they cannot fully account for trapping mechanisms which are found to involve most of the molecular modes. The proposed MQC model always improves the agreement with reference full QD simulations, even when only few modes are included in the quantum partition. It correctly reproduces the quenching of oscillations in electronic populations and partially recovers the error of reduced-dimensionality models on the long-time populations.