Comparison of one‐parameter and linearly scaled one‐parameter double‐hybrid density functionals for noncovalent interactions

Abstract

We have compared the performances of the one‐parameter and linearly scaled one‐parameter double‐hybrid density functionals (1DH‐DFs and LS1DH‐DFs) for noncovalent interactions. The only one parameter related to the Hartree–Fock (HF) exchange for each of the tested 1DH‐DFs and LS1DH‐DFs has been fitted with the well‐designed S66 database. The obtained DHDFs are dubbed as 1DH‐PBE‐NC, LS1DH‐PBE‐NC, 1DH‐TPSS‐NC, LS1DH‐TPSS‐NC, 1DH‐PWB95‐NC, and LS1DH‐PWB95‐NC, where “NC” denotes noncovalent interactions. With a specific combination of exchange and correlation functionals, the dependent parameters related to the nonlocal second‐order perturbative energies are nearly identical for the 1DH and LS1DH models. According to our benchmark computations against the S66, S22B, NCCE31, ADIM6, and L7 databases, we suggest that the 1DH‐PWB95‐NC and LS1DH‐PWB95‐NC functionals are much more suitable for evaluating noncovalent interaction energies. Unlike the versatile DHDFs with dispersion corrections for general purpose, our optimized 1DH‐DFs and LS1DH‐DFs only aim at noncovalent interactions. © 2016 Wiley Periodicals, Inc.

Small Basis Set Allowing the Recovery of Dispersion Interactions with Double-Hybrid Functionals

“…Owing to these differences in reference data we have opted to discuss the results with respect to two different reference sets: the original one (QCISD­(T)/CBS) and that obtained with the DLPNO-CCSD­(T)/CBS approach using a “Normal” threshold, which has been already used for comparison purposes. ,, …”

“…Owing to these differences in reference data we have opted to discuss the results with respect to two different reference sets: the original one (QCISD(T)/CBS) and that obtained with the DLPNO-CCSD(T)/CBS approach using a "Normal" threshold, which has been already used for comparison purposes. 34,36,82 The results obtained with the PBE-QIDH functional are reported in Table 4 together with the two references data sets and literature values obtained at the B2PLYP-D3(BJ) level. The MAE for the PBE-QIDH/DH-SVPD approach is 1.90 kcal/mol with respect to the DLPNO-CCSD(T)/CBS data set and 1.25 kcal/mol if the QCISD(T)/CBS reference is considered.…”

DSD-PBEP86-NL and DOD-PBEP86-NL functionals for noncovalent interactions: Basis set effects and tentative applications to large noncovalent systems

“…According to previous works, the replacement of NL correlation for D3BJ correction is also reasonable for general applications other than noncovalent interactions. On this point, the DSD‐PBEP86‐NL and DOD‐PBEP86‐NL functionals are much better than the specifically parameterized DHDFs only for noncovalent interactions …”

“…On this point, the DSD-PBEP86-NL and DOD-PBEP86-NL functionals are much better than the specifically parameterized DHDFs only for noncovalent interactions. [72,73]…”

Crystal Structure and Solid-State Packing of 4-Chloro-5H-1,2,3-dithiazol-5-one and 4-Chloro-5H-1,2,3-dithiazole-5-thione

“…The methods used in this work are proven to represent similar heterocyclic compounds. B3LYP is a standard of DFT for gas phase molecules and is a well proven compromise between computational cost and accuracy [ 68 , 69 ]. BLYP and PBE are also common functionals applied to gas phase molecules and yield comparable results where the difference is that BLYP and PBE are gradient corrected functionals and B3LYP is a hybrid function.…”