一些密度泛函方法预测电子亲和势

选取了杂化泛函B3LYP, B3PW91, O3LYP, PBE0, 以及与之相对应的GGA泛函BLYP, BPW91, OLYP和PBE, 还选取了能更好地兼顾强相互作用和弱相互作用的X3LYP泛函和在预测NMR的化学位移有较好表现的OPBE泛函, 以及两种meta-GGA泛函VSXC和TPSS, 共12种泛函, 详细地考察了这些泛函在预测EA方面的准确性.

Performance of Some Density Functional Methods in the Prediction of the Adiabatic Electron Affinities

In this paper, we examined the performance of 12 density functionals(B3LYP, X3LYP, O3LYP, PBE0, B3PW91, BLYP, OLYP, OPBE, PBE, BPW91,VSXC and TPSS), combined with two basis setsDZP++ and 6-311+G(3df,2p)] in the prediction of the adiabatic electron affinities against a set of 91 well-established experimental values. We have reached a conclusion that all these modern functionals are ge-nerally good, giving a mean absolute deviation(MAD) less than 0.24 eV. B3LYP and X3LYP are among the most reliable methods for this test set. When the DZP++ basis set is adopted, X3LYP is the best with a MAD of 0.14 eV, while B3LYP leads to a MAD of 0.16 eV. When the basis set is extended to 6-311+G(3df,2p), B3LYP improves its MAD to 0.13 eV, while the MAD of X3LYP remains to be 0.14 eV. Most hybrid functionals outperform the corresponding GGA functinals, with the exception of PBE0. The meta-GGA functionals(VSXC and TPSS) are not necessarily better than the GGA functionals. Our calculations also reveal that there are flaws associated with the OPTX exchange functional in the calculations of EA.