Benchmark Many-Body GW and Bethe–Salpeter Calculations for Small Transition Metal Molecules

Abstract

We study the electronic and optical properties of 39 small molecules containing transition metal atoms and 7 others related to quantum-dots for photovoltaics. We explore in particular the merits of the many-body GW formalism, as compared to the ΔSCF approach within density functional theory, in the description of the ionization energy and electronic affinity. Mean average errors of 0.2-0.3 eV with respect to experiment are found when using the PBE0 functional for ΔSCF and as a starting point for GW. The effect of partial self-consistency at the GW level is explored. Further, for optical excitations, the Bethe-Salpeter formalism is found to offer similar accuracy as time-dependent DFT-based methods with the hybrid PBE0 functional, with mean average discrepancies of about 0.3 and 0.2 eV, respectively, as compared to available experimental data. Our calculations validate the accuracy of the parameter-free GW and Bethe-Salpeter formalisms for this class of systems, opening the way to the study of large clusters containing transition metal atoms of interest for photovoltaic applications.

Excitation spectra of aromatic molecules within a real-spaceGW-BSE formalism: Role of self-consistency and vertex corrections

“…Our results concerning self consistency stand in contrast to past calculations using atomcentered basis sets, which suggest that eigenvalue, QS, and fully self-consistent GW can improve spectral properties for molecules. [2][3][4][5]10,13,15,17,22,24,52,55 At the same time, the systematic deterioration in accuracy in our QSGW also differs from a plane-wave implementation of self-consistent GW , which does not show a clear trend for increasing or decreasing accuracy with the same DFT-LDA starting point. 19 We observe that for self-consistent GW , numerical considerations such as the choice of a quasiparticle basis for self-consistency, as well as the basis set chosen to represent wave functions, must be better understood before a consensus can be reached on the theoretical accuracy of self-consistent GW for molecules.…”

“…In several examples with atom-centered basis sets, the accuracy of selfconsistent GW is competitive with G 0 W 0 predictions using Hartree-Fock and hybrid functional mean-field starting points. [2][3][4][5]10,13,15,17,22,24,52,55 However, recent results for fully self-consistent GW on a plane-wave basis set report slightly larger errors. 19 Overall, self-consistent GW pushes the IPs upward compared to G 0 W 0 with a DFT starting point, and it shifts IPs slightly downward when using a Hartree-Fock starting point.…”

Structural and excited-state properties of oligoacene crystals from first principles

“…This method is often very successful, but nevertheless it is somewhat dependent on the DFT starting point. GW can be evaluated, in principle, selfconsistently by different approaches [68,[68][69][70][71][72][73][74], mitigating the starting-point dependence by iterating over eigenvalues and wavefunctions. Given the computational demands associated with acene crystals, in the following we limit our study to the diagonal part of and, if going beyond G 0 W 0 , we only update the eigenvalues in G and W , retaining the original DFT wave functions under the assumption that they are close to the true QP wave functions [66,[75][76][77].…”

Evaluating theGWApproximation with CCSD(T) for Charged Excitations Across the Oligoacenes

“…hybrid functionals. [18][19][20][54][55][56][57] In this work, we will consider some promising hybrid functionals that have been identified in previous studies, 54,58 including PBE0 59 and BHLYP, 47 with 25% and 50% exact exchange, respectively. We will also consider the OTRSH functional 33 described above.…”