期刊
JOURNAL OF COMPUTATIONAL CHEMISTRY
卷 38, 期 19, 页码 1693-1703出版社
WILEY
DOI: 10.1002/jcc.24813
关键词
Frozen-density embedding (FDE); continuum solvation model; Lagrangian methods; analytical nuclear gradients; excited states
资金
- Deutsche Forschungsgemeinschaft DFG [HO-4605/2-1]
- Research Council of Norway through a Centre of Excellence [179568/V30]
- Tromso Forskningsstiftelse (SurfInt grant)
We present the explicit derivation of an approach to the multiscale description of molecules in complex environments that combines frozen-density embedding (FDE) with continuum solvation models, in particular the conductor-like screening model (COSMO). FDE provides an explicit atomistic description of molecule-environment interactions at reduced computational cost, while the outer continuum layer accounts for the effect of long-range isotropic electrostatic interactions. Our treatment is based on a variational Lagrangian framework, enabling rigorous derivations of ground- and excited-state response properties. As an example of the flexibility of the theoretical framework, we derive and discuss FDE+COSMO analytical molecular gradients for excited states within the Tamm-Dancoff approximation (TDA) and for ground states within second-order MOller-Plesset perturbation theory (MP2) and a second-order approximate coupled cluster with singles and doubles (CC2). It is shown how this method can be used to describe vertical electronic excitation (VEE) energies and Stokes shifts for uracil in water and carbostyril in dimethyl sulfoxide (DMSO), respectively. In addition, VEEs for some simplified protein models are computed, illustrating the performance of this method when applied to larger systems. The interaction terms between the FDE subsystem densities and the continuum can influence excitation energies up to 0.3 eV and, thus, cannot be neglected for general applications. We find that the net influence of the continuum in presence of the first FDE shell on the excitation energy amounts to about 0.05 eV for the cases investigated. The present work is an important step toward rigorously derived ab initio multilayer and multiscale modeling approaches. (c) 2017 Wiley Periodicals, Inc.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据