Conical intersections (CIs) of ethylene have been successfully determined using spin-flip density functional theory (SFDFT) combined with a penalty-constrained optimization method. We present in detail three structures, twisted-pyramidalized, hydrogen-migrated, and ethylidene CIs. In contrast to the linear response time-dependent density functional theory, which predicts a purely twisted geometry without pyramidalization as the S(1) global minimum, SFDFT gives a pyramidalized structure. Therefore, this is the first correct optimization of CI points of twisted ethylene by the DFT method. The calculated energies and geometries are in good agreement with those obtained by the multireference configuration interaction (MR-CI) method and the multistate formulation of second-order multireference perturbation theory (MS-CASPT2).

The photoisomerization process of 1,2-diphenylethylene (stilbene) is investigated using the spin-flip density functional theory (SFDFT), which has recently been shown to be a promising approach for locating conical intersection (CI) points (Minezawa, N.; Gordon, M. S. J. Phys. Chem. A2009, 113, 12749). The SFDFT method gives valuable insight into twisted stilbene to which the linear response time-dependent DFT approach cannot be applied. In contrast to the previous SFDFT study of ethylene, a distinct twisted minimum is found for stilbene. The optimized structure has a sizable pyramidalization angle and strong ionic character, indicating that a purely twisted geometry is not a true minimum. In addition, the SFDFT approach can successfully locate two CI points: the twisted-pyramidalized CI that is similar to the ethylene counterpart and another CI that possibly lies on the cyclization pathway of cis-stilbene. The mechanisms of the cis-trans isomerization reaction are discussed on the basis of the two-dimensional potential energy surface along the twisting and pyramidalization angles. Disciplines Chemistry CommentsReprinted (adapted) with permission from

The effects of solvents on electronic spectra can be treated efficiently by combining an accurate quantum mechanical (QM) method for the solute with an efficient and accurate method for the solvent molecules. One of the most sophisticated approaches for treating solvent effects is the effective fragment potential (EFP) method. The EFP method has been interfaced with several QM methods, including configuration interaction, time-dependent density functional theory, multiconfigurational methods, and equations-of-motion coupled cluster methods. These combined QM-EFP methods provide a range of efficient and accurate methods for studying the impact of solvents on electronic excited states. An energy decomposition analysis in terms of physically meaningful components is presented in order to analyze these solvent effects. Several factors that must be considered when one investigates solvent effects on electronic spectra are discussed, and several examples are presented. Disciplines Chemistry CommentsReprinted (adapted) hile multiple absorption and emission spectroscopic experimental studies provide valuable information on the magnitude and dynamics of soluteÀsolvent coupling, calculations on electronic excited states in the condensed phase remain a major challenge to the theoretical chemistry community.1 The increased number of nuclear and electronic degrees of freedom relative to the gas phase makes accurate fully ab initio calculations on a condensed-phase system unfeasible long before the system can approach the bulk. One general approach to this type of problem is to separate a system into two parts, such that one (active, usually solute) part is treated by quantum mechanical (QM) techniques and the other (usually larger, solvent) part is calculated using classical (molecular) mechanics (MM).2 The Hamiltonian of the system then consists of three termŝIn eq 1, H QM/MM is a coupling term between the two levels of theory. Separation of the QM and MM subsystems, in principle, allows one to use any level of theory in both the QM and MM parts.There have been an increasing number of studies devoted to a description of electronic spectroscopy in the condensed phase.3À14 An alternative to the QM/MM approach is to study the electronic excited states of solutes with dielectric continuum methods. 3À5,9,11,12,15 While continuum models are computationally inexpensive, they cannot describe explicit solventÀsolute interactions such as hydrogen bonding. Another promising approach for studying condensed-phase electronic spectroscopy in large molecular systems is to use a fragment-based technique, such as the fragment molecular orbital method (FMO). 16À20The FMO and related methods have the advantage of being close to fully "ab initio", but these methods are still sufficiently computationally demanding that (for example) performing molecular dynamics simulations on excited states in solution is still not feasible.If one is to perform QM/MM calculations to accurately capture solvent effects on electronic excited states, it i...

Fragment molecular orbital molecular dynamics (FMO-MD) with periodic boundary conditions is performed on liquid water using the analytic energy gradient, the electrostatic potential point charge approximation, and the electrostatic dimer approximation. Compared to previous FMO-MD simulations of water that used an approximate energy gradient, inclusion of the response terms to provide a fully analytic energy gradient results in better energy conservation in the NVE ensemble for liquid water. An FMO-MD simulation that includes the fully analytic energy gradient and two body corrections (FMO2) gives improved energy conservation compared with a previously calculated FMO-MD simulation with an approximate energy gradient and including up to three body corrections (FMO3). Disciplines Chemistry CommentsReprinted (adapted) ABSTRACT: Fragment molecular orbital molecular dynamics (FMO-MD) with periodic boundary conditions is performed on liquid water using the analytic energy gradient, the electrostatic potential point charge approximation, and the electrostatic dimer approximation. Compared to previous FMO-MD simulations of water that used an approximate energy gradient, inclusion of the response terms to provide a fully analytic energy gradient results in better energy conservation in the NVE ensemble for liquid water. An FMO-MD simulation that includes the fully analytic energy gradient and two body corrections (FMO2) gives improved energy conservation compared with a previously calculated FMO-MD simulation with an approximate energy gradient and including up to three body corrections (FMO3). INTRODUCTION1.1. Fragment Molecular Orbital Method. Much progress has been made recently in improving algorithms for ab initio calculations on large systems. 1,2 One strategy is to use a fragmentation approach, in which a large system of interest is divided into smaller subsystems, and ab initio calculations are performed on these smaller subsystems. 2 One of the most successful and extensively developed fragmentation methods is the fragment molecular orbital (FMO) method 3 proposed by Kitaura et al. in 1999. The FMO method has been implemented for most ab initio and density functional theory (DFT) methods. 4−8 Numerous approximations to the original FMO method have been implemented to improve the efficiency of the calculation. The two most important of these approximations are the electrostatic point charge (ESP-PC) approximation and the electrostatic dimer (ES-DIM) approximation. 9 The ESP-PC approximation calculates the FMO embedded electrostatic potential using point charges rather than two electron integrals, while the ES-DIM approximation calculates the dimer energy of two fragments using point charges rather than using ab initio methods.FMO gradients were developed soon after the introduction of the FMO method. 10 Improvements to the gradient followed that allowed the use of gradients with the ESP-PC approximation 11 and the ES-DIM approximation. 12 Because the dimer (or trimer) density is not iterated to self-consi...