Ilya G. Ryabinkin

When do we need to account for the geometric phase in excited state dynamics

We investigate the role of the geometric phase (GP) in an internal conversion process when the system changes its electronic state by passing through a conical intersection (CI). Local analysis of a two-dimensional linear vibronic coupling (LVC) model Hamiltonian near the CI shows that the role of the GP is twofold. First, it compensates for a repulsion created by the so-called diagonal Born-Oppenheimer correction. Second, the GP enhances the non-adiabatic transition probability for a wave-packet part that experiences a central collision with the CI. To assess the significance of both GP contributions we propose two indicators that can be computed from parameters of electronic surfaces and initial conditions. To generalize our analysis to N-dimensional systems we introduce a reduction of a general N-dimensional LVC model to an effective 2D LVC model using a mode transformation that preserves short-time dynamics of the original N-dimensional model. Using examples of the bis(methylene) adamantyl and butatriene cations, and the pyrazine molecule we have demonstrated that their effective 2D models reproduce the short-time dynamics of the corresponding full dimensional models, and the introduced indicators are very reliable in assessing GP effects.

Accurate and efficient approximation to the optimized effective potential for exchange.

We devise an efficient practical method for computing the Kohn-Sham exchange-correlation potential corresponding to a Hartree-Fock electron density. This potential is almost indistinguishable from the exact-exchange optimized effective potential (OEP) and, when used as an approximation to the OEP, is vastly better than all existing models. Using our method one can obtain unambiguous, nearly exact OEPs for any reasonable finite one-electron basis set at the same low cost as the Krieger-Li-Iafrate and Becke-Johnson potentials. For all practical purposes, this solves the long-standing problem of black-box construction of OEPs in exact-exchange calculations.

Geometric phase effects in dynamics near conical intersections: symmetry breaking and spatial localization.

We show that finite systems with conical intersections can exhibit spontaneous symmetry breaking which manifests itself in spatial localization of eigenstates. This localization has a geometric phase origin and is robust against variation of model parameters. The transition between localized and delocalized eigenstate regimes resembles a continuous phase transition. The localization slows down the low-energy quantum nuclear dynamics at low temperatures.

Geometric phase effects in low-energy dynamics near conical intersections: A study of the multidimensional linear vibronic coupling model

In molecular systems containing conical intersections (CIs), a nontrivial geometric phase (GP) appears in the nuclear and electronic wave functions in the adiabatic representation. We study GP effects in nuclear dynamics of an N-dimensional linear vibronic coupling (LVC) model. The main impact of GP on low-energy nuclear dynamics is reduction of population transfer between the local minima of the LVC lower energy surface. For the LVC model, we proposed an isometric coordinate transformation that confines non-adiabatic effects within a two-dimensional subsystem interacting with an N - 2 dimensional environment. Since environmental modes do not couple electronic states, all GP effects originate from nuclear dynamics within the subsystem. We explored when the GP affects nuclear dynamics of the isolated subsystem, and how the subsystem-environment interaction can interfere with GP effects. Comparing quantum dynamics with and without GP allowed us to devise simple rules to determine significance of the GP for nuclear dynamics in this model.

Reduction of Electronic Wave Functions to Kohn-Sham Effective Potentials.

A method for calculating the Kohn-Sham exchange-correlation potential v(XC)(r) from a given electronic wave function is devised and implemented. It requires on input one- and two-electron reduced density matrices and involves construction of the generalized Fock matrix. The method is free from numerical limitations and basis-set artifacts of conventional schemes for constructing v(XC)(r) in which the potential is recovered from a given electron density, and is simpler than various many-body techniques. The chief significance of this development is that it allows one to directly probe the functional derivative of the true exchange-correlation energy functional and to rigorously test and improve various density-functional approximations.

Hierarchy of model Kohn–Sham potentials for orbital-dependent functionals: A practical alternative to the optimized effective potential method

We describe a method for constructing a hierarchy of model potentials approximating the functional derivative of a given orbital-dependent exchange-correlation functional with respect to electron density. Each model is derived by assuming a particular relationship between the self-consistent solutions of Kohn-Sham (KS) and generalized Kohn-Sham (GKS) equations for the same functional. In the KS scheme, the functional is differentiated with respect to density, in the GKS scheme--with respect to orbitals. The lowest-level approximation is the orbital-averaged effective potential (OAEP) built with the GKS orbitals. The second-level approximation, termed the orbital-consistent effective potential (OCEP), is based on the assumption that the KS and GKS orbitals are the same. It has the form of the OAEP plus a correction term. The highest-level approximation is the density-consistent effective potential (DCEP), derived under the assumption that the KS and GKS electron densities are equal. The analytic expression for a DCEP is the OCEP formula augmented with kinetic-energy-density-dependent terms. In the case of exact-exchange functional, the OAEP is the Slater potential, the OCEP is roughly equivalent to the localized Hartree-Fock approximation and related models, and the DCEP is practically indistinguishable from the true optimized effective potential for exact exchange. All three levels of the proposed hierarchy require solutions of the GKS equations as input and have the same affordable computational cost.

Removal of Basis-Set Artifacts in Kohn–Sham Potentials Recovered from Electron Densities

Kohn-Sham effective potentials recovered from Gaussian-basis-set electron densities exhibit large oscillations and asymptotic divergences not found in exact potentials and in functional derivatives of approximate density functionals. We show that the detailed structure of these oscillations and divergences is almost exclusively determined by the basis set in terms of which the reference density is expressed, and is almost independent of the density-functional or wave function method used for computing the density. Based on this observation, we propose a smoothening scheme in which most basis-set artifacts in a Kohn-Sham potential recovered from a given density are removed by subtracting the oscillation profile of the exchange-only local-density approximation potential computed in the same basis set as the reference density. The correction allows one to obtain smooth Kohn-Sham potentials from electron densities even for small Gaussian basis sets and greatly reduces discrepancies between the original (input) density and the density obtained from the reconstructed potential.

Why Do Mixed Quantum-Classical Methods Describe Short-Time Dynamics through Conical Intersections So Well? Analysis of Geometric Phase Effects

Adequate simulation of nonadiabatic dynamics through conical intersection requires accounting for a nontrivial geometric phase (GP) emerging in electronic and nuclear wave functions in the adiabatic representation. Popular mixed quantum-classical (MQC) methods, surface hopping and Ehrenfest, do not carry a nuclear wave function to be able to incorporate the GP into nuclear dynamics. Surprisingly, the MQC methods reproduce ultrafast interstate crossing dynamics generated with the exact quantum propagation so well as if they contained information about the GP. Using two-dimensional linear vibronic coupling models we unravel how the MQC methods can effectively mimic the most significant dynamical GP effects: (1) compensation for repulsive diagonal second-order nonadiabatic couplings and (2) transfer enhancement for a fully cylindrically symmetric component of a nuclear distribution.

Average local ionization energy generalized to correlated wavefunctions

The average local ionization energy function introduced by Politzer and co-workers [Can. J. Chem. 68, 1440 (1990)] as a descriptor of chemical reactivity has a limited utility because it is defined only for one-determinantal self-consistent-field methods such as the Hartree-Fock theory and the Kohn-Sham density-functional scheme. We reinterpret the negative of the average local ionization energy as the average total energy of an electron at a given point and, by rewriting this quantity in terms of reduced density matrices, arrive at its natural generalization to correlated wavefunctions. The generalized average local electron energy turns out to be the diagonal part of the coordinate representation of the generalized Fock operator divided by the electron density; it reduces to the original definition in terms of canonical orbitals and their eigenvalues for one-determinantal wavefunctions. The discussion is illustrated with calculations on selected atoms and molecules at various levels of theory.

Efficient construction of exchange and correlation potentials by inverting the Kohn–Sham equations

Given a set of canonical Kohn-Sham orbitals, orbital energies, and an external potential for a many-electron system, one can invert the Kohn-Sham equations in a single step to obtain the corresponding exchange-correlation potential, vXC(r). For orbitals and orbital energies that are solutions of the Kohn-Sham equations with a multiplicative vXC(r) this procedure recovers vXC(r) (in the basis set limit), but for eigenfunctions of a non-multiplicative one-electron operator it produces an orbital-averaged potential. In particular, substitution of Hartree-Fock orbitals and eigenvalues into the Kohn-Sham inversion formula is a fast way to compute the Slater potential. In the same way, we efficiently construct orbital-averaged exchange and correlation potentials for hybrid and kinetic-energy-density-dependent functionals. We also show how the Kohn-Sham inversion approach can be used to compute functional derivatives of explicit density functionals and to approximate functional derivatives of orbital-dependent functionals.