High-energy photoelectron C 1s and O 1s shake-up spectra of CO

The core electron shake phenomenon

“…To our knowledge, the only applications [57,58] of Green's function methods to core electron shake-up spectra are due to the so-called algebraic diagrammatic construction, A D C (~) and A D C (~) , methods, which are, respectively, complete thirdand fourth-order approximation schemes to the one-particle Green's function [57]. Within this scheme, the ionization energies and spectroscopic factors are obtained as eigenpairs of a Hermitian secular matrix and computed with techniques similar to those used in diagonalizing large CI matrices.…”

“…The construction and solution of the secular matrix is actually simplified in the core electron problem since one is allowed to neglect the core-valence correlation (cf. singleoccupancy restriction, neglect of semi-internal CI, see Section 3.2), which leads to a decoupling of the secular problem [58,87]. The secular matrix is then defined by a configuration space comprising lh, 2h-lp, and 3h-2p states with exactly one core-level vacancy.…”

“…The low-order particle states are treated at higher level ( l h states by the fourth perturbation order) than higher-order particle states (e.g., 3h-2p states by the first perturbation order). Lower shake-up states are thus better described than are higher lying ones [58]. Starting out from a set of ground-state orbitals, there is also the problem of interpretation of what in the excitation pattern should be interpreted as relaxation and what are "true" excitations referring to the final state.…”

“…The calculation of shake-up spectra can be subdivided into two main categories: Green's function methods (here ADC(~) and ~& 4 ) [57,58]) and wavefunction methods (here OSRHF, CI, and MCSCF). The latter can be subdivided into explicit Hamiltonian methods, active space methods, configuration selection methods, and amplitude selection methods.…”

Interaction between theory and experiment in molecular core electron spectroscopies

“…A straight application of RPA on X-ray emission leads to errors larger than 10 eV for first-row molecules. Furthermore, some core electron processes are delicately dependent on the precise nature of the electron correlation and would require higher-order propagator methods if the unrelaxed ground state is used as the reference state [16]. Higherorder expansions and/or level shifting techniques are necessary to obtain solutions with proper core occupations.…”