A novel algorithm is introduced for coding all Slater determinants in the covalent space with conserved SZ, the z component of total spin S for a classical valence bond (VB) model. It effectively minimizes the search time and the storing space in the central memory of the computer. In cooperation with symmetry reductions based on molecular point group and spin inversion, the VB calculations have been extended to benzenoid hydrocarbons of up to 28 π-electrons that have 4×10 7 configurations. The low-lying states of benzenoids with 24, 26, and 28 π-electrons have been obtained for 62 species. To rationalize the aromaticity of benzenoids in a VB scheme, the resonance energy per hexagon (REPH) is defined. A linear correlation between the REPH and the energy gap of the ground (singlet) state and the first excited (triplet) state for 89 benzenoids is established.

We have carried out density matrix renormalization group calculations on the T1 state of linear polyenes applying the Pariser-Parr-Pople (PPP) Model. The geometry optimization for the polyene oligomers C2nH2n12 (n54,5,6,...,15) shows that the S0 to T1 excitation region is composed of a soliton-antisoliton pair located symmetrically away from the center of the chain and leads to singleand double-bond interconversions in between. The distance between the soliton and antisoliton centers in T1 state changes with the length of the chain, contradictory to earlier conclusions obtained with PPP-SDCI or ab initio SCI methods. The inconsistency most possibly comes from the insufficient consideration of the electron correlations in small-scale CI methods.

The chain length dependence of excitation energies of oilgomers of polyene, p-phenylene, pentafulvene, cyclopentadiene, pyrrole, furan, silole, phosphole, and thiophene was studied employing the time-dependent density functional theory with B3LYP functional. Band gaps and effective conjugation lengths of the corresponding polymers were obtained by extrapolating vertical excitation energies of trimers through pentamers to infinite chain length. Polypentafulvene, polycyclopentadiene, polysilole, and polyphosphole were predicted to have smaller band gaps than polythiophene, and polypentafulvene has the smallest band gap. Among the studied cyclopentadiene-based polymers, those with the third row heteroatoms were found to have narrower band gaps than their second row counterparts.

The mechanism of H2O2 dismutation catalyzed by the dibenzotetraaza[14]annulene-Fe(Ⅲ) complex ([Fe(C24H22N4O4)]+) which was recently reported (Paschke, J.; Kirsch, M.; Korth, H. G.; de Groot, H.; Sustmann, R. J. Am. Chem. Soc. 2001, 123, 11099) has been investigated by density functional theory using the B3LYP hybrid functional. The quartet potential energy profile of the catalytic reaction has been explored. In the whole catalytic cycle, the ratedetermining step is found to be the O-O bond homolytic cleavage, without the assistance of solvent molecules in the second coordination shell. The calculated free energy barrier for this step is 10.8kcal/mol, which is in reasonable agreement with the experimental facts. The calculations also show that the hydroxyl and hydroperoxyl radicals may be generated in the reaction processes, but they can be efficiently quenched in strongly exothermic steps. This study provides a satisfactory explanation to the observed efficiency of the H2O2 dismutation catalyzed by this complex.

A linear scaling local correlation approach is proposed for approximately solving the coupled cluster doubles (CCD) equations of large systems in a basis of orthogonal localized molecular orbitals (LMOs). By restricting double excitations from spatially close occupied LMOs into their associated virtual LMOs, the number of significant excitation amplitudes scales only linearly with molecular size in large molecules. Significant amplitudes are obtained to a very good approximation by solving the CCD equations of various subsystems, each of which is made up of a cluster associated with the orbital indices of a subset of significant amplitudes and the local environmental domain of the cluster. The combined effect of these two approximations leads to a linear scaling algorithm for large systems. By using typical thresholds, which are designed to target an energy accuracy, our numerical calculations for a wide range of molecules using the 6-31G or 6-31G* basis set demonstrate that the present local correlation approach recovers more than 98.5% of the conventional CCD correlation energy.