Department of Chemistry, Dalhousie University, 6274 Coburg Road, P.O. Box 15000, Halifax, Nova Scotia B3H 4R2, Canada.
J Chem Phys. 2012 Jul 7;137(1):014104. doi: 10.1063/1.4731342.
The deficiency of conventional density-functional theory (DFT) in properly describing van der Waals (vdW) (especially dispersion-bound) complexes has been extensively addressed in the past decade. There are now several new methods published in the literature that are capable of accurately capturing weak dispersion interactions in complexes at equilibrium geometries. However, the performance of these new methods at non-equilibrium geometries remains to be assessed. We have previously published [F. O. Kannemann and A. D. Becke, J. Chem. Theory Comput. 6, 1081 (2010); A. D. Becke, A. A. Arabi, and F. O. Kannemann, Can. J. Chem. 88, 1057 (2010)] that the functional PW86+PBE+XDM for exchange + correlation + dispersion, respectively, is a highly accurate functional for general thermochemistry and vdW complexes at equilibrium geometries. Here, we show that this nonempirical, except for two parameters in the dispersion damping part, functional also performs well for vdW complexes at compressed and stretched intermonomer separations. The mean absolute relative error (MARE) is 9.4% overall for vdW complexes in the "S22×5" database incorporating compressed and stretched geometries [J. Rezac, K. E. Riley, and P. Hobza, J. Chem. Theory Comput. 7, 2427 (2011)]. Our largest MARE on the S22×5 database is 13.3% on the compressed geometry set.
过去十年中,传统密度泛函理论(DFT)在适当描述范德华(vdW)(特别是色散结合)复合物方面的不足已得到广泛解决。现在,文献中已经发表了几种新方法,这些方法能够在平衡几何结构中准确捕获复合物中的弱色散相互作用。然而,这些新方法在非平衡几何结构中的性能仍有待评估。我们之前已经发表过[F. O. Kannemann 和 A. D. Becke,J. Chem. Theory Comput. 6, 1081 (2010); A. D. Becke, A. A. Arabi 和 F. O. Kannemann,Can. J. Chem. 88, 1057 (2010)],交换+相关+色散的 PW86+PBE+XDM 函数分别是平衡几何结构下一般热化学和 vdW 复合物的高度准确函数。在这里,我们表明,这个非经验的,除了色散阻尼部分的两个参数外,功能也很好地适用于压缩和拉伸单体分离的 vdW 复合物。在包含压缩和拉伸结构的“S22×5”数据库中,对于 vdW 复合物,平均绝对相对误差(MARE)总体为 9.4%[J. Rezac, K. E. Riley 和 P. Hobza,J. Chem. Theory Comput. 7, 2427 (2011)]。我们在“S22×5”数据库中的最大 MARE 是在压缩几何数据集上为 13.3%。