CCSD(T) anharmonic force fields may be the 'gold standard' (if 5 cm-1 accuracy is acceptable), but their computational cost is prohibitive for all but very small molecules. Therefore, the combination of CCSD(T) harmonic frequencies with DFT anharmonic force fields (obtained from numerical differentiation of analytical second derivatives) has gained some currency. For a fairly large benchmark set, we assess a very large number of basis set and exchange-correlation functional combinations. We show that double-hybrid functionals such as Grimme's B2PLYP [1] and our own DSD-PBEP86 [2] can yield frequencies between CCSD and CCSD(T) quality without any further correction.For CCSD(T) harmonic frequencies, cc-pVQZ or larger basis sets are usually required for converged results, which may present an obstacle for some systems. We explore several of the factors that influence the quality of explicitly correlated alternatives: the approximation to the full CCSD(T)-F12 equations, the basis set, and the scaling of (T). We show that CCSD(T*)(F12*)/cc-pVDZ-F12 is already acceptable for most applications, and CCSD(T*)/cc-pVTZ-F12 typically comes within 1 cm-1 of the basis set limit. Benzene, with its notorious intramolecular BSSE problem for the bending frequencies, is discussed as a cautionary tale.The lecture touches on several research interests of the late lamented Prof. Handy: anharmonic force fields (including of C6H6), density functional theory, and explicitly correlated methods.[1] Grimme S., J. Chem. Phys. 2006 , 124, 034108 [2] Kozuch S., Martin J.M.L., Phys. Chem. Chem. Phys. 2011 , 13, 20104-20107; J. Comput Chem. 2013 , 34, 2327

Explicitly correlated coupled cluster theory and double hybrid density functionals: An excellent team for anharmonic force fields of small molecules

248th National Meeting (2014)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.