This paper clarifies why long-range corrected (LC) density functional theory gives orbital energies quantitatively. First, the highest occupied molecular orbital and the lowest unoccupied molecular orbital energies of typical molecules are compared with the minus vertical ionization potentials (IPs) and electron affinities (EAs), respectively. Consequently, only LC exchange functionals are found to give the orbital energies close to the minus IPs and EAs, while other functionals considerably underestimate them. The reproducibility of orbital energies is hardly affected by the difference in the short-range part of LC functionals. Fractional occupation calculations are then carried out to clarify the reason for the accurate orbital energies of LC functionals. As a result, only LC functionals are found to keep the orbital energies almost constant for fractional occupied orbitals. The direct orbital energy dependence on the fractional occupation is expressed by the exchange self-interaction (SI) energy through the potential derivative of the exchange functional plus the Coulomb SI energy. On the basis of this, the exchange SI energies through the potential derivatives are compared with the minus Coulomb SI energy. Consequently, these are revealed to be cancelled out only by LC functionals except for H, He, and Ne atoms.

Chemical reaction calculations were carried out using the long-range correction (LC) scheme, which improves long-range exchange effects in density functional theory (DFT) [J. Chem. Phys. 115, 3540 (2001); 120, 8425 (2004)]. A new determination of the LC scheme parameter mu was made by a root mean square fit of the percent error in calculated atomization energies. As a result, the parameter mu was optimized as 0.47, which is higher than the previous one (mu=0.33). Using this new parameter mu, LC-DFT was firstly applied to geometry optimizations of the G2 benchmark set molecules. Consequently, this new LC-DFT gave more accurate bond lengths and bond angles than previous LC-DFT and hybrid B3LYP results. Following this result, the authors calculated reaction barrier height energies of benchmark reaction sets, which have been underestimated in conventional DFT calculations. Calculated results showed that LC-DFT provided much more accurate barrier height energies with errors less than half those of previous LC-DFT and B3LYP studies. To test the general validity of the new LC-DFT, the authors finally calculated reaction enthalpies. As a result, they found that the LC scheme using the new mu clearly improved the accuracy of calculated enthalpies. The authors therefore conclude that the insufficient inclusion of long-range exchange effects is responsible for the underestimation of reaction barriers in DFT calculations and that LC-DFT using the new parameter is a powerful tool for theoretically investigating chemical reactions.

We report the calculations of core-excitation energies of first-row atoms using the time-dependent density functional theory (DFT) and the long-range correction (LC) scheme for exchange-correlation functionals, including LC-BOP, Coulomb-attenuated method BLYP, and our recently developed LCgau-BOP method, which includes a flexible portion of short-range Hartree-Fock (HF) exchange through the inclusion of a Gaussian function in the LC scheme. We show that the LC scheme completely fails to improve the poor accuracy of conventional generalized gradient approximation functionals, while the LCgau scheme gives an accuracy which is an order of magnitude better than BLYP and significantly better than B3LYP. A reoptimization of the two parameters controlling the inclusion of short-range HF exchange in the LCgau method enables the errors to be reduced to the order of 0.1 eV which is competitive with the best DFT methods we are aware of. This reparametrization does not affect the LC scheme and therefore maintains the high accuracy of predicted reaction barrier heights. Moreover, while there is some loss in accuracy in thermochemical predictions compared to the previously optimized LCgau-BOP, rms errors in the atomization energies over the G2 test set are found to be comparable to B3LYP. Finally, we attempt to rationalize the success of the LC and LCgau schemes in terms of the well-known self-interaction error (SIE) of conventional functionals. To estimate the role of the SIE, we examine the total energy calculations for systems with a fractional number of electrons, not only in the highest occupied molecular orbital but also in the 1s-characterized core orbital. Our conclusion is that the inclusion of short-range HF exchange in LC-type functionals can significantly alleviate the problems of the SIE in the core region. In particular, we confirm that the absence of the SIE diagnostics in the core orbital energies correlates with the accurate prediction of core-excitation energies using the newly optimized LCgau approach.

A new hybrid exchange-correlation functional is presented based on the long-range correction (LC) scheme [H. Iikura et al., J. Chem. Phys. 115, 3540 (2001); Tawada et al., J. Chem. Phys. 120, 8425 (2004)], named LCgau-BOP. The key feature is the use of a two-parameter Gaussian correction to the Coulomb attenuation, which allows a more flexible description of exact exchange at short-range interelectronic separations. The new partitioning preserves 100% exact exchange in the long range, which is known to be important for the success of the LC scheme, with an asymptotic attenuation described by a standard error function with a parameter of 0.42. The LCgau partitioning was optimized for the reproduction of atomization energies over the G2 set and reaction barrier heights over Database/3, and produced results which are superior to B3LYP, CAM-BLYP, and the best LC functionals we are aware of. The results highlight the importance of including a substantial portion of exact exchange in the short range. Using the same parameters, the new functional was tested for the reproduction of geometries, as well as valence, Rydberg and charge-transfer excitations which are known challenges for conventional density functional theory. Our conclusion is that LCgau-BOP can provide a consistently more accurate description of thermochemistries, chemical reactions, and excitation energies than other existing long-range corrected functionals.