Q-Chem contains an ever-growing suite of many-body methods for describing open-shell molecules and excited states.172 The EOM-CC224–226 (link) and ADC227,241 (link) formalisms are two powerful approaches for describing multiconfigurational wave functions within a black-box single-reference formalism. Target states |Ψex⟩ are described as excitations from a reference state |Ψ0⟩, |Ψex=R^|Ψ0, where R^ is an excitation operator parameterized via amplitudes that are determined by solving an eigenvalue problem. In EOM-CC, these amplitudes are eigenvectors of the effective Hamiltonian H¯=eT^H^eT^, in which T^ is either the CC or the MP2 operator for the reference state. Currently, EOM-CCSD and EOM-MP2 models are available. In ADC, an effective shifted Hamiltonian is constructed using perturbation theory and the intermediate state representation (ISR) formalism,227,241 (link) similar to Eq. (10), to afford M=Ψex|H^E0|Ψex, where E0 is the energy of the MPn reference state. Diagonalization of the Hermitian matrix M yields excitation energies, and the ADC eigenvectors give access to the excited-state wave function. Second-order standard ADC(2), extended ADC(2)-x, and ADC(3) are available.241 For the second-order ADC schemes, spin-opposite-scaled (SOS) variants are also implemented.242 (link)
Various EOM-CC and ADC models are defined by the choice of reference state |Ψ0⟩ and excitation operator R^ , as illustrated in Fig. 10. The following models are available:224,227,241 (link) EE (excitation energies), IP (ionization potentials), EA (electron affinities), SF (spin–flip, for triplet and quartet references), 2SF (double SF, for quintet references); DIP (double IP), and DEA (double EA). At present, the 2SF, DIP, and DEA variants are only available in combination with an EOM treatment.243 (link)
Analytic gradients244,245 (link) and properties246–248 (link) are available for most of these models, including transition properties between different target states (e.g., transition dipoles, angular momentum, and electronic circular dichroism rotatory strengths),249 (link) nonadiabatic couplings,250 (link) spin–orbit couplings,220,251,252 (link) and nonlinear optical properties, including two-photon transition moments and (hyper)polarizabilities for both ground and excited states.253–256 (link) Extensions of these theories to metastable states257 (link) (resonances) and to core-level excitations258–260 (link) are also available and are highlighted in Sec. V.
The IP and EA variants of these models afford spin-pure descriptions of ground and excited doublet states and are useful for modeling charge-transfer processes. EOM-SF and SF-ADC methods are suitable for treating diradicals, triradicals, and conical intersections. The DEA and DIP ansätze further expand the scope of applicability.243 (link) Spin–flip methods can be used to treat strongly correlated systems within an effective Hamiltonian formalism,221,261,262 (link) with applications to single-molecule magnets and even infinite spin chains.222 (link)
For visualization purposes, both Dyson orbitals264 (link) and natural transition orbitals265 (link) (NTOs) are available,15,88,220,266–269 (link) including NTOs of the response density matrices for analyzing two-photon absorption270 (link) and resonant inelastic x-ray scattering.271 (link)
Figure 11 highlights the application of these tools to model magnetic properties and spin-forbidden chemistry. Exciton analyses,267,268,272–274 (link) bridging the gap between the quasiparticle and MO pictures of excited states, enable the calculation and visualization of electron–hole correlation.89,267,268,272,273 (link)
Free full text: Click here