Using Valence Bond Theory to Understand Electronic Excited States:  Application to the Hidden Excited State (2Ag) of C2nH2n+2(n= 2−14) Polyenes

461

619

We present a detailed investigation of the acene series using high-level wave function theory. Our ab initio density matrix renormalization group algorithm has enabled us to carry out complete active space calculations on the acenes from napthalene to dodecacene correlating the full -valence space. While we find that the ground state is a singlet for all chain lengths, examination of several measures of radical character, including the natural orbitals, effective number of unpaired electrons, and various correlation functions, suggests that the longer acene ground states are polyradical in nature.

Section: The Nature Of Bonding In the Acene Polyradical Statementioning

confidence: 99%

The radical character of the acenes: A density matrix renormalization group study

Hachmann

Dorando

Aviles

et al. 2007

461

619

“…7,13 Here, the name "covalent" derives from the dominant covalent Rumer diagrams appearing in valence bond descriptions (see, e.g., Refs. [16][17][18]. The covalent excitations are the 2A − g , 1B − u , and 3A − g states as classified by the irreducible representations of the C 2h point groups of the all-trans polyenes and by the "Pariser alternancy symmetry" (PAS), which provides the additional labels "+" and "−."…”

Section: A Theory Of π-Electron Excitations In Long Polyenesmentioning

confidence: 99%

“…Here, the question arises, which level of theory one should choose for the computation of ground state geometries. This question is important, because calculated vertical excitation energies will differentially and sensitively depend on the bond length alternation in the chosen ground state geometry 7,18,45 (cf. also Sec.…”

Section: Om2 Optimized Geometriesmentioning

confidence: 99%

“…For comparison with further results on excitation energies in long polyenes, which, however, only cover either the covalent 18,26 or the ionic 56 excitations and were obtained by DMRG/CASSCF, 26 valence bond DFT, 18 PPP/DMRG, 56 and PPP/CCSD, 56 respectively, Sec. IV B of the SM provides graphical representations of corresponding data in Figs.…”

Section: B Causes and Effects Of Lacking Size Consistencymentioning

confidence: 99%

See 1 more Smart Citation

Electronic excitations in long polyenes revisited

Schmidt

Tavan

2012

106

We apply the valence shell model OM2 [W. Weber and W. Thiel, Theor. Chem. Acc. 103, 495, (2000)] combined with multireference configuration interaction (MRCI) to compute the vertical excitation energies and transition dipole moments of the low-energy singlet excitations in the polyenes with 4 ≤ N ≤ 22π -electrons. We find that the OM2/MRCI descriptions closely resemble those of Pariser-Parr-Pople (PPP) π -electron models [P. Tavan and K. Schulten, Phys. Rev. B 36, 4337, (1987)], if equivalent MRCI procedures and regularly alternating model geometries are used. OM2/MRCI optimized geometries are shown to entail improved descriptions particularly for smaller polyenes (N ≤ 12), for which sizeable deviations from the regular model geometries are found. With configuration interaction active spaces covering also the σ -in addition to the π -electrons, OM2/MRCI excitation energies turn out to become smaller by at most 0.35 eV for the ionic and 0.15 eV for the covalent excitations. The particle-hole (ph) symmetry, which in Pariser-Parr-Pople models arises from the zero-differential overlap approximation, is demonstrated to be only weakly broken in OM2 such that the oscillator strengths of the covalent 1B − u states, which artificially vanish in ph-symmetric models, are predicted to be very small. According to OM2/MRCI and experimental data the 1B − u state is the third excited singlet state for N < 12 and becomes the second for N ≥ 14. By comparisons with results of other theoretical approaches and experimental evidence we argue that deficiencies of the particular MRCI method employed by us, which show up in a poor size consistency of the covalent excitations for N > 12, are caused by its restriction to at most doubly excited references.

“…They are not expected to be well described by any nearest-neighbor model, which neglects long range electronelectron repulsion, including the Hubbard model. The Hubbard model does agree with CASPT2 and most 58,70 (but by no means all 71,72 ) ab initio calculations in placing the 1 B u charge transfer states below the 1 A g homopolar states in these molecules. The homopolar state is spectroscopically dark, but there is no doubt that it becomes the lowest-lying singlet excited state in longer polyenes, 73 with the crossover probably coming at octatetraene in which the two states are nearly degenerate.…”

Section: B Excitation Energiesmentioning

confidence: 72%

A distance-dependent parameterization of the extended Hubbard model for conjugated and aromatic hydrocarbons derived from stretched ethene

Schmalz

Serrano‐Andrés

Sauri

et al. 2011

The Hubbard model, which is widely used in physics but is mostly unfamiliar to chemists, provides an attractive yet simple model for chemistry beyond the self consistent field molecular orbital approximation. The Hubbard model adds an effective electron-electron repulsion when two electrons occupy the same atomic orbital to the familiar Hückel Hamiltonian. Thus it breaks the degeneracy between excited singlet and triplet states and allows an explicit treatment of electron correlation. We show how to evaluate the parameters of the model from high-level ab initio calculations on twoatom fragments and then to transfer the parameters to large molecules and polymers where accurate ab initio calculations are difficult or impossible. The recently developed MS-RASPT2 method is used to generate accurate potential energy curves for ethene as a function of carbon-carbon bond length, which are used to parameterize the model for conjugated hydrocarbons. Test applications to several conjugated/aromatic molecules show that even though the model is very simple, it is capable of reasonably accurate predictions for bond lengths, and predicts molecular excitation energies in reasonable agreement with those from the MS-RASPT2 method.