A comparison of the behavior of functional/basis set combinations for hydrogen-bonding in the water dimer with emphasis on basis set superposition error

We evaluate the performance of ten functionals (B3LYP, M05, M05-2X, M06, M06-2X, B2PLYP, B2PLYPD, X3LYP, B97D, and MPWB1K) in combination with 16 basis sets ranging in complexity from 6-31G(d) to aug-cc-pV5Z for the calculation of the H-bonded water dimer with the goal of defining which combinations of functionals and basis sets provide a combination of economy and accuracy for H-bonded systems. We have compared the results to the best non-density functional theory (non-DFT) molecular orbital (MO) calculations and to experimental results. Several of the smaller basis sets lead to qualitatively incorrect geometries when optimized on a normal potential energy surface (PES). This problem disappears when the optimization is performed on a counterpoise (CP) corrected PES. The calculated interaction energies (ΔEs) with the largest basis sets vary from -4.42 (B97D) to -5.19 (B2PLYPD) kcal/mol for the different functionals. Small basis sets generally predict stronger interactions than the large ones. We found that, because of error compensation, the smaller basis sets gave the best results (in comparison to experimental and high-level non-DFT MO calculations) when combined with a functional that predicts a weak interaction with the largest basis set. As many applications are complex systems and require economical calculations, we suggest the following functional/basis set combinations in order of increasing complexity and cost: (1) D95(d,p) with B3LYP, B97D, M06, or MPWB1k; (2) 6-311G(d,p) with B3LYP; (3) D95++(d,p) with B3LYP, B97D, or MPWB1K; (4) 6-311++G(d,p) with B3LYP or B97D; and (5) aug-cc-pVDZ with M05-2X, M06-2X, or X3LYP.     

Assessment of density functionals with long‐range and/or empirical dispersion corrections for conformational energy calculations of peptides

Density functionals with long‐range and/or empirical dispersion corrections, including LC‐ωPBE, B97‐D, ωB97X‐D, M06‐2X, B2PLYP‐D, and mPW2PLYP‐D functionals, are assessed for their ability to describe the conformational preferences of Ac‐Ala‐NHMe (the alanine dipeptide) and Ac‐Pro‐NHMe (the proline dipeptide) in the gas phase and in water, which have been used as prototypes for amino acid residues of peptides. For both dipeptides, the mean absolute deviation (MAD) is estimated to be 0.22–0.40 kcal/mol in conformational energy and 2.0–3.2° in torsion angles ? and ψ using these functionals with the 6‐311++G(d,p) basis set against the reference values calculated at the MP2/aug‐cc‐pVTZ//MP2/aug‐cc‐pVDZ level of theory in the gas phase. The overall performance is obtained in the order B2PLYP‐D ≈ mPW2PLYP‐D > ωB97X‐D ≈ M06‐2X > MP2 > LC‐ωPBE > B3LYP with the 6–311++G(d,p) basis set. The SMD model at the M06‐2X/6‐31+G(d) level of theory well reproduced experimental hydration free energies of the model compounds for backbone and side chains of peptides with MADs of 0.47 and 4.3 kcal/mol for 20 neutral and 5 charged molecules, respectively. The B2PLYP‐D/6‐311++G(d,p)//SMD M06‐2X/6‐31+G(d) level of theory provides the populations of backbone and/or prolyl peptide bond for the alanine and proline dipeptides in water that are consistent with the observed values. © 2010 Wiley Periodicals, Inc. J Comput Chem, 2010     

Accurate prediction of vertical electronic transitions of Ni(II) coordination compounds via time dependent density functional theory

Time dependent density functional theory calculations are completed for five Ni(II) complexes formed by polydentate peptides to predict the electronic absorption spectrum. The ligands examined were glycyl‐glycyl‐glycine (GGG), glycyl‐glycyl‐glycyl‐glycine (GGGG), glycyl‐glycyl‐histidine (GGH), glycyl‐glycyl‐cysteine (GGC), and triethylenetetramine (trien). Fifteen functionals and two basis sets were tested. On the basis of the mean absolute percent deviation (MAPD), the ranking among the functionals is: HSE06 ∼ MPW1PW91 ∼ PBE0 > ω‐B97x‐D ∼ B3P86 ∼ B3LYP ∼ CAM‐B3LYP > PBE ∼ BLYP ∼ BP86 > TPSS > TPSSh > BHandHLYP > M06 ≫ M06‐2X. Concerning the basis sets, the triple‐ζ def2‐TZVP performs better than the double‐ζ LANL2DZ. With the functional HSE06 and basis set def2‐TZVP the MAPD with respect to the experimental λ_max is 1.65% with a standard deviation of 1.26%. The absorption electronic spectra were interpreted in terms of vertical excitations between occupied and virtual MOs based on Ni‐d atomic orbitals. The electronic structure of the Ni(II) species is also discussed.     

A theoretical analysis of substituted aromatic compounds

The substituents ? CH₃, ? F, ? NO₂, ? OCH₃, and ? CH₂?CH₂ were placed at the ortho, meta, and para positions on the aromatic molecules aniline, benzaldehdye, nitrobenzene, and phenol. MMFF94, AM1, B3LYP, M06, M06‐2X, ωB97X, ωB97X‐d, and RI‐MP2 using cc‐pVDZ and cc‐pVTZ and CCSD(T) with cc‐pVDZ basis sets were used to calculate the geometries and energies of all regiomers of the molecules. Relative energies of the ortho and meta regiomers relative to the para regiomers were calculated and compared to the CCSD(T) values. A good basis set correlation between cc‐pVDZ and cc‐pVTZ was observed in RI‐MP2. Overall, RI‐MP2 gave the best correlation with the CCSD(T) results. All of the hybrid functionals showed similar accuracy and could effectively describe the intramolecular hydrogen‐bonding interactions of these compounds. The methoxy group at the para position in methoxyaniline, methoxyphenol, methoxynitrobenzene, and methoxybenzaldehyde was rotated around the phenyl‐O bond. HF, along with the cc‐pVDZ basis with the other methods, generated inaccurate energy profiles for p‐methoxyphenol. For the density functional theory methods, it was necessary to use improved grids to get smooth curves. © 2013 Wiley Periodicals, Inc.     

分子筛限域孔道中吡啶的吸附结构和能量

以吸附于ZSM-5孔道中的吡啶分子为例,利用量子化学理论方法考察了计算模型和密度泛函方法的选择对吡啶吸附结构和吸附能的影响,从而为准确计算分子筛限域孔道中客体分子吸附态结构和能量参数提供了依据.计算结果表明,吡啶吸附能随着所选用的分子筛的计算模型(从8T到128T)增大而增大,当选用的孔道结构能够将整个分子筛的孔道结构完全包括进来的时候(72T)达到收敛.与常规的密度泛函方法(B3LYP和M06-2X)相比较,考虑到色散作用校正的B97D泛函方法能够很好地处理分子筛体系中主客体间的长程相互作用和弱相互作用,计算得到的能量数据与实验结果符合得很好.     

Computational studies on ethylene addition to nickel bis(dithiolene)

The density functionals B3LYP, B3PW91, BMK, HSE06, LC-ωPBE, M05, M06, O3LYP, TPSS, ω-B97X, and ω-B97XD are used to optimize key transition states and intermediates for ethylene addition to Ni(edt)(2) (edt = S(2)C(2)H(2)). The efficacy of the basis sets 6-31G**, 6-31++G**, cc-pVDZ, aug-cc-pVDZ, cc-pVTZ, and aug-cc-pVTZ is also examined. The geometric parameters optimized with different basis sets and density functionals are similar and agree well with experimental values. The ω-B97XD functional gives relative energies closest to those from CCSD, while M06 and HSE06 yield results close to those from CCSD(T). CASSCF and CASSCF-PT2 calculation results are also given. Variation of the relative energies from different density functionals appears to arise, in part, from the multireference character of this system, as confirmed by the T1 diagnostic and CASSCF calculations.     

Vibrational frequency scale factors for density functional theory and the polarization consistent basis sets

Calculated harmonic vibrational frequencies systematically deviate from experimental vibrational frequencies. The observed deviation can be corrected by applying a scale factor. Scale factors for: (i) harmonic vibrational frequencies [categorized into low (<1000 cm^?1) and high (>1000 cm^?1)], (ii) vibrational contributions to enthalpy and entropy, and (iii) zero‐point vibrational energies (ZPVEs) have been determined for widely used density functionals in combination with polarization consistent basis sets (pc‐n, n = 0,1,2,3,4). The density functionals include pure functionals (BP86, BPW91, BLYP, HCTH93, PBEPBE), hybrid functionals with Hartree‐Fock exchange (B3LYP, B3P86, B3PW91, PBE1PBE, mPW1K, BH&HLYP), hybrid meta functionals with the kinetic energy density gradient (M05, M06, M05‐2X, M06‐2X), a double hybrid functional with Møller‐Plesset correlation (B2GP‐PLYP), and a dispersion corrected functional (B97‐D). The experimental frequencies for calibration were from 41 organic molecules and the ZPVEs for comparison were from 24 small molecules (diatomics, triatomics). For this family of basis sets, the scale factors for each property are more dependent on the functional selection than on basis set level, and thus allow for a suggested scale factor for each density functional when employing polarization consistent basis sets (pc‐n, n = 1,2,3,4). A separate scale factor is recommended when the un‐polarized basis set, pc‐0, is used in combination with the density functionals. © 2012 Wiley Periodicals, Inc.     

用GGA密度泛函及其长程、色散校正方法计算各类氢键的结合能

应用广义梯度近似(GGA) (PW91和PBE)、含动能密度的广义梯度近似(meta-GGA) (M06-L)、杂化泛函(hyper-GGA)(M06-2X、X3LYP和B3LYP)及其长程校正泛函LC-DFT(CAM-B3LYP、LC-ωPBE和ωB97X)和色散校正密度泛函(DFT-D)(ωB97X-D和B97-D),用多种基函数对15种不同强度的传统氢键和非传统氢键体系的结合能进行了系统的计算与分析.并与高精度的CCSD(T)/aug-cc-pVQZ结果比较发现:在上述各类泛函中,对于氢键结合能的计算M06-2X和ωB97X-D泛函较为精确与可靠,且没有必要使用过大的基函数,6-311++G(2d,2p)或aug-cc-pVDZ水平的基组就已足够,各类泛函所计算结合能的基组重叠误差(BSSE)均较小,除ωB97X和ωB97X-D外,其它9种泛函不经BSSE校正也能得到同样甚至更准确的结果.     

Comparison of some dispersion-corrected and traditional functionals as applied to peptides and conformations of cyclohexane derivatives

We compare the energetic and structural properties of fully optimized α-helical and antiparallel β-sheet polyalanines and the energetic differences between axial and equatorial conformations of three cyclohexane derivatives (methyl, fluoro, and chloro) as calculated using several functionals designed to treat dispersion (B97-D, ωB97x-D, M06, M06L, and M06-2X) with other traditional functionals not specifically parametrized to treat dispersion (B3LYP, X3LYP, and PBE1PBE) and with experimental results. Those functionals developed to treat dispersion significantly overestimate interaction enthalpies of folding for the α-helix and predict unreasonable structures that contain Ramachandran φ and ψ and C = O[ellipsis (horizontal)]N H-bonding angles that are out of the bounds of databases compiled the β-sheets. These structures are consistent with overestimation of the interaction energies. For the cyclohexanes, these functionals overestimate the stabilities of the axial conformation, especially when used with smaller basis sets. Their performance improves when the basis set is improved from D95?? to aug-cc-pVTZ (which would not be possible with systems as large as the peptides).     

New theoretical insight on the acid sites distribution,their local structures and acid strength of the SAPO‐11 molecular sieve

ONIOM(DFT:PM3) calculations were carried out to investigate and characterize possible acid sites of SAPO‐11 molecular sieve. Two functionals were employed: B3LYP and ωB97X‐D. This last functional includes dispersion effects that are absent in the former. Benzene, pyridine, and ammonia interaction energies as well as the OH stretching frequencies of the POH, SiOH, and bridged Si(OH)Al groups were used to characterize the acid sites. This work shows that the adsorption of benzene on the surface is as strong as the adsorptions inside main channel of SAPO‐11. Pyridine adsorption on the surface is weaker than the one corresponding to the main channel. NH₃ molecule interacts strongly with all OH groups or acid sites present in SAPO‐11. Moreover, the results reveal that it is possible to adsorb two NH₃ molecules at only one Brønsted site. The adsorption of the second NH₃ molecule is energetically favorable mainly due to the hydrogen bond formation between the NH₃ molecules. In general the interaction energy increases with the type of functional used, according to the trend ωB97X‐D > B3LYP. The results show that ONIOM methodology seems to be suited to investigate the acid sites in SAPO‐11.     

Surface Charge‐Transfer Doping of Graphene Nanoflakes Containing Double‐Vacancy (5‐8‐5) and Stone–Wales (55‐77) Defects through Molecular Adsorption

The adsorption of six electron donor–acceptor (D/A) organic molecules on various sizes of graphene nanoflakes (GNFs) containing two common defects, double‐vacancy (5‐8‐5) and Stone–Wales (55‐77), are investigated by means of ab initio DFT [M06‐2X(‐D3)/cc‐pVDZ]. Different D/A molecules adsorb on a defect graphene (DG) surface with binding energies (ΔE_b) of about ?12 to ?28 kcal mol^?1. The ΔE_b values for adsorption of molecules on the Stone–Wales GNF surface are higher than those on the double vacancy GNF surface. Moreover, binding energies increase by about 10 % with an increase in surface size. The nature of cooperative weak interactions is analyzed based on quantum theory of atoms in molecules, noncovalent interactions plot, and natural bond order analyses, and the dominant interaction is compared for different molecules. Electron density population analysis is used to explain the n‐ and p‐type character of defect graphene nanoflakes (DGNFs) and also the change in electronic properties and reactivity parameters of DGNFs upon adsorption of different molecules and with increasing DGNF size. Results indicate that the HOMO–LUMO energy gap (E_g) of DGNFs decreases upon adsorption of molecules. However, by increasing the size of DGNFs, the E_g and chemical hardness of all complexes decrease and the electrophilicity index increases. Furthermore, the values of the chemical potential of acceptor–DGNF complexes decrease with increasing size, whereas those of donor–DGNF complexes increase.     

Quantum chemistry benchmarking of binding and selectivity for lanthanide extractants

Computer‐aided screening methods facilitate the discovery of new extractants for heavy and rare‐earth metal separations. In this work, we have benchmarked the accuracy of different quantum chemistry methods for calculating extractant binding energies and selectivities. Specifically, we compare calculated data from different exchange correlation functionals (B3LYP‐D3, ωB97X‐D3, and M06‐L) and different basis sets (including large‐core effective core potentials and all‐electron basis sets). We report aqueous‐phase binding energy and selectivity trends for 1:1 and 3:1 extractant/lanthanide models for the complexes. We find that binding selectivities are not particularly sensitive to model chemistry, but binding energies are sensitive. Furthermore, calculated trends in selectivity using 3:1 extractant/lanthanide models are in better agreement with available experimental trends than trends using 1:1 extractant/lanthanide models. Lastly, we find that the B3LYP‐D3/6‐31 + G* model chemistry with the Stuttgart large‐core relativistic effective core potentials on the lanthanide sufficiently reproduces results from larger basis set calculations and is confirmed as suitable for relatively fast and efficient screening of lanthanide binding energies and selectivities.     

Calculations of ionization energies and electron affinities for atoms and molecules: A comparative study with different methods

In the present work, we examined the performance of 36 density functionals, including the newly developed doubly hybrid density functional XYG3 (Y. Zhang, X. Xu, and W. A. Goddard III, Proc. Natl. Acad. Sci, USA, 2009, 106, 4963), to calculate ionization energies (IEs) and electron affinities (EAs). We used the well-established G2-1 set as reference, which contains 14 atoms and 24 molecules for IE, along with 7 atoms and 18 molecules for EA. XYG3 leads to mean absolute deviations (MADs) of 0.057 and 0.080 eV for IEs and EAs, respectively, using the basis set of 6-311 + G (3df,2p). In comparison with some other functionals, MADs for IEs are 0.109 (B2PLYP), 0.119 (M06-2X), 0.159 (X3LYP), 0.161 (PBE), 0.162 (B3LYP), 0.165 (PBE0), 0.173 (TPSS), 0.200 (BLYP), and 0.215 eV (LC-BLYP). MADs for EAs are 0.090 (X3LYP), 0.090 (B2PLYP), 0.102 (PBE), 0.103 (M06-2X), 0.104 (TPSS), 0.105 (BLYP), 0.106 (B3LYP), 0.126 (LC-BLYP), and 0.128 eV (PBE0).     

Performance of Density Functional Theory for Transition Metal Oxygen Bonds

The accuracy of density functional theory (DFT) limits predictions in theoretical catalysis, and strong chemical bonds between transition metals and oxygen pose a particular challenge. We benchmarked 30 diverse density functionals against the bond dissociation enthalpies (BDE) of the 30 MO and 30 MO⁺ diatomic systems of all the 3d, 4d, and 5d metals, to test universality across the d-block as required in comparative studies. Seven functionals, B98, B97-1, B3P86, B2PLYP, TPSSh, B3LYP, and B97-2, display mean absolute errors (MAE) <30 kJ/mol. In contrast, many commonly used functionals such as PBE and RPBE overestimate M−O bonding by +30 kJ/mol and display MAEs from 48–76 kJ/mol. RPBE and OPBE reduce the over-binding of PBE but remain very inaccurate. We identify a linear relationship (p-value 7.6 ⋅ 10⁻⁵) between the precision and accuracy of DFT, i. e. inaccurate functionals tend to produce larger, unpredictable random errors. Some functionals commonly deviate from this relationship: Thus, M06-2X is very precise but not very accurate, whereas B3LYP* and MN15-L are more accurate but less precise than M06-2X. The best-performing hybrids have 10–30 % HF exchange, but this can be relieved by double hybrids (B2PLYP). Most functionals describe trends well, but errors comparing 5d to 4d/3d are ∼10 kJ/mol larger than group-wise errors, due to uncertainties in the spin-orbit coupling corrections for effective core potentials, affecting e. g. Pt/Pd or Au/Ag comparisons.     

Benchmark results for empirical post-GGA functionals: difficult exchange problems and independent tests

Many of the most promising new density functionals have improved the treatment of non-local exchange effects with the help of semi-empirical information and more sophisticated recipes for combining Hartree-Fock and local exchange approximations. In order to quantify recent advancements and identify directions for improvement, we have examined a broad spectrum of test problems. We evaluate the performance of several new hybrid density functionals (ωB97, ωB97X, ωB97X-D, LRC-ωPBEh, M06, M06-2X, and M06-HF) on a variety of chemical problems, some sensitive to the treatment of exact exchange (which we have hoped to systematically improve) and some which require a balanced treatment of correlation. Since all of the functionals under consideration are parameterized with ground-state thermochemical data, the benchmark aims to determine the applicability of the new density functionals to cases that have not been considered in the optimization of the semi-empirical parameters. The first class of benchmarks includes the excitation energies of 21 molecules (83 states) primarily from a recent benchmark conducted by Tozer and co-workers, with some additional references from data made available from the groups of Thiel and Truhlar. We briefly examine the conformational preferences of a small peptide and complete our study with two recently published sets of data that have shown large, systematic errors in simple alkane thermochemistry. While our results indicate that the more general hybrids currently under development perform well for problems outside of their parameterization and improve over the standard hybrid density functionals in an essentially systematic way, there is still a significant self-interaction error in the more difficult cases. Functionals based on a range-separation of exchange and functionals depending on the kinetic-energy density both perform comparably, and there is evidence for complementary strengths.     

Assessment of density functionals and force field methods on anion–π interaction in heterocyclic calix complexes

The performance of ten density functionals and four force field methods in describing non-covalent interactions have been assessed by studying the interaction energies and structures of the four anion–π complexes involving tetraoxacalix[2]arene[2]triazine and various anions. Their structures are optimized at MP2/6-311++G(d,p) level, and interaction energies are obtained at DF-MP2-F12/aug-cc-pVDZ level. The result shows that the functional M06-2X predicts the most reliable interaction energy, followed by wB97XD and BHandH. B97D slightly overestimates the interaction energy. Other functionals and force field methods seriously overestimate the interaction energy. For the structures, three functionals M06-2X, wB97XD and BH and H predict the most reliable results, followed by B97D. The force field methods predict the largest deviations. The present work suggests that the functional M06-2X is a reliable method to describe energies and structural properties of the large molecules involving the anion–π interactions.