Comparing ab initio density-functional and wave function theories: the impact of correlation on the electronic density and the role of the correlation potential

The framework of ab initio density-functional theory (DFT) has been introduced as a way to provide a seamless connection between the Kohn-Sham (KS) formulation of DFT and wave-function based ab initio approaches [R. J. Bartlett, I. Grabowski, S. Hirata, and S. Ivanov, J. Chem. Phys. 122, 034104 (2005)]. Recently, an analysis of the impact of dynamical correlation effects on the density of the neon atom was presented [K. Jankowski, K. Nowakowski, I. Grabowski, and J. Wasilewski, J. Chem. Phys. 130, 164102 (2009)], contrasting the behaviour for a variety of standard density functionals with that of ab initio approaches based on second-order M?ller-Plesset (MP2) and coupled cluster theories at the singles-doubles (CCSD) and singles-doubles perturbative triples [CCSD(T)] levels. In the present work, we consider ab initio density functionals based on second-order many-body perturbation theory and coupled cluster perturbation theory in a similar manner, for a range of small atomic and molecular systems. For comparison, we also consider results obtained from MP2, CCSD, and CCSD(T) calculations. In addition to this density based analysis, we determine the KS correlation potentials corresponding to these densities and compare them with those obtained for a range of ab initio density functionals via the optimized effective potential method. The correlation energies, densities, and potentials calculated using ab initio DFT display a similar systematic behaviour to those derived from electronic densities calculated using ab initio wave function theories. In contrast, typical explicit density functionals for the correlation energy, such as VWN5 and LYP, do not show behaviour consistent with this picture of dynamical correlation, although they may provide some degree of correction for already erroneous explicitly density-dependent exchange-only functionals. The results presented here using orbital dependent ab initio density functionals show that they provide a treatment of exchange and correlation contributions within the KS framework that is more consistent with traditional ab initio wave function based methods.     

Ab initio and DFT benchmark study for nucleophilic substitution at carbon (SN2@C) and silicon (SN2@Si)

To obtain a set of consistent benchmark potential energy surfaces (PES) for the two archetypal nucleophilic substitution reactions of the chloride anion at carbon in chloromethane (S(N)2@C) and at silicon in chlorosilane (S(N)2@Si), we have explored these PESes using a hierarchical series of ab initio methods [HF, MP2, MP4SDQ, CCSD, CCSD(T)] in combination with a hierarchical series of six Gaussian-type basis sets, up to g polarization. Relative energies of stationary points are converged to within 0.01 to 0.56 kcal/mol as a function of the basis-set size. Our best estimate, at CCSD(T)/aug-cc-pVQZ, for the relative energies of the [Cl(-), CH(3)Cl] reactant complex, the [Cl-CH(3)-Cl](-) transition state and the stable [Cl-SiH(3)-Cl](-) transition complex is -10.42, +2.52, and -27.10 kcal/mol, respectively. Furthermore, we have investigated the performance for these reactions of four popular density functionals, namely, BP86, BLYP, B3LYP, and OLYP, in combination with a large doubly polarized Slater-type basis set of triple-zeta quality (TZ2P). Best overall agreement with our CCSD(T)/aug-cc-pVQZ benchmark is obtained with OLYP and B3LYP. However, OLYP performs better for the S(N)2@C overall and central barriers, which it underestimates by 2.65 and 4.05 kcal/mol, respectively. The other DFT approaches underestimate these barriers by some 4.8 (B3LYP) to 9.0 kcal/mol (BLYP).     

Oxidative addition of the ethane C-C bond to Pd. An ab initio benchmark and DFT validation study

We have computed a state-of-the-art benchmark potential energy surface (PES) for the archetypal oxidative addition of the ethane C-C bond to the palladium atom and have used this to evaluate the performance of 24 popular density functionals, covering LDA, GGA, meta-GGA, and hybrid density functionals, for describing this reaction. The ab initio benchmark is obtained by exploring the PES using a hierarchical series of ab initio methods [HF, MP2, CCSD, CCSD(T)] in combination with a hierarchical series of five Gaussian-type basis sets, up to g polarization. Relativistic effects are taken into account either through a relativistic effective core potential for palladium or through a full four-component all-electron approach. Our best estimate of kinetic and thermodynamic parameters is -10.8 (-11.3) kcal/mol for the formation of the reactant complex, 19.4 (17.1) kcal/mol for the activation energy relative to the separate reactants, and -4.5 (-6.8) kcal/mol for the reaction energy (zero-point vibrational energy-corrected values in parentheses). Our work highlights the importance of sufficient higher angular momentum polarization functions for correctly describing metal-d-electron correlation. Best overall agreement with our ab initio benchmark is obtained by functionals from all three categories, GGA, meta-GGA, and hybrid DFT, with mean absolute errors of 1.5 to 2.5 kcal/mol and errors in activation energies ranging from -0.2 to -3.2 kcal/mol. Interestingly, the well-known BLYP functional compares very reasonably with a slight underestimation of the overall barrier by -0.9 kcal/mol. For comparison, with B3LYP we arrive at an overestimation of the overall barrier by 5.8 kcal/mol. On the other hand, B3LYP performs excellently for the central barrier (i.e., relative to the reactant complex) which it underestimates by only -0.1 kcal/mol.     

Density functional theory and the correlation consistent basis sets: the tight d effect on HSO and HOS

The HSO and HOS isomers have been revisited using the DFT functionals, B3LYP, B3PW91, and PBE, in combination with tight d-augmented correlation consistent basis sets, cc-pV(x+d)Z and aug-cc-pV(x+d)Z for second-row atoms. Structures, vibrationally averaged structures, relative energies, harmonic and anharmonic frequencies, enthalpies of formation of HSO and HOS, and the barrier for the HSO/HOS isomerization have been determined. These results were compared with results from previous DFT and ab initio studies in which the standard correlation consistent basis sets were used. The relative energies of the two isomers converge more rapidly and smoothly with respect to increasing basis set size for the tight d-augmented sets than for the standard basis sets. Our best calculations, B3PW91/aug-cc-pV(5+d)Z, for the relative energy of the isomers are in excellent agreement with previous CCSD(T) results given by Wilson and Dunning.     

Ab initio calculations on halogen-bonded complexes and comparison with density functional methods

A systematic theoretical investigation on a series of dimeric complexes formed between some halocarbon molecules and electron donors has been carried out by employing both ab initio and density functional methods. Full geometry optimizations are performed at the Moller-Plesset second-order perturbation (MP2) level of theory with the Dunning's correlation-consistent basis set, aug-cc-pVDZ. Binding energies are extrapolated to the complete basis set (CBS) limit by means of two most commonly used extrapolation methods and the aug-cc-pVXZ (X = D, T, Q) basis sets series. The coupled cluster with single, double, and noniterative triple excitations [CCSD(T)] correction term, determined as a difference between CCSD(T) and MP2 binding energies, is estimated with the aug-cc-pVDZ basis set. In general, the inclusion of higher-order electron correlation effects leads to a repulsive correction with respect to those predicted at the MP2 level. The calculations described herein have shown that the CCSD(T) CBS limits yield binding energies with a range of -0.89 to -4.38 kcal/mol for the halogen-bonded complexes under study. The performance of several density functional theory (DFT) methods has been evaluated comparing the results with those obtained from MP2 and CCSD(T). It is shown that PBEKCIS, B97-1, and MPWLYP functionals provide accuracies close to the computationally very expensive ab initio methods.     

Energy landscapes of nucleophilic substitution reactions: a comparison of density functional theory and coupled cluster methods

We have carried out a detailed evaluation of the performance of all classes of density functional theory (DFT) for describing the potential energy surface (PES) of a wide range of nucleophilic substitution (SN2) reactions involving, amongst others, nucleophilic attack at carbon, nitrogen, silicon, and sulfur. In particular, we investigate the ability of the local density approximation (LDA), generalized gradient approximation (GGA), meta-GGA as well as hybrid DFT to reproduce high-level coupled cluster (CCSD(T)) benchmarks that are close to the basis set limit. The most accurate GGA, meta-GGA, and hybrid functionals yield mean absolute deviations of about 2 kcal/mol relative to the coupled cluster data, for reactant complexation, central barriers, overall barriers as well as reaction energies. For the three nonlocal DFT classes, the best functionals are found to be OPBE (GGA), OLAP3 (meta-GGA), and mPBE0KCIS (hybrid DFT). The popular B3LYP functional is not bad but performs significantly worse than the best GGA functionals. Furthermore, we have compared the geometries from several density functionals with the reference CCSD(T) data. The same GGA functionals that perform best for the energies (OPBE, OLYP), also perform best for the geometries with average absolute deviations in bond lengths of 0.06 A and 0.6 degrees, even better than the best meta-GGA and hybrid functionals. In view of the reduced computational effort of GGAs with respect to meta-GGAs and hybrid functionals, let alone coupled cluster, we recommend the use of accurate GGAs such as OPBE or OLYP for the study of SN2 reactions.     

Density functional and chemical model study of the competition between methyl and hydrogen scission of propane and ��-scission of the propyl radical

In this work, we study the competence between the reactions of hydrogen and methyl scission during thermal cracking and combustion of propane, the emergence of the two isomers of the propyl radical, n-propyl and i-propyl, and their subsequent ??-scission reaction to ethene and methyl radical. The purpose of the study was to analyze the accuracy of density functional (DFT) methods as applied on this relatively well-known subset of the reactions implied in the production of propylene oxide from propane and propene. Conventional (B3LYP, B3PW91) and state-of-the-art (PBE0, M06, BMK) DFT methods were employed, and their accuracy checked against experimental data and calculations performed using model chemistries (complete basis set CBS-4M, QB3, and APNO, and G4 methods) and ab initio methods (MP2, CCSD(T) with a large 6-311 ++G(3df,2pd) basis set). The results obtained at the BMK level for the thermodynamics of the reactions are closer to experimental data than those afforded by any other DFT method and very similar actually to CBS or CCSD(T) results, even if a medium size basis set is used. Activation energies determined using two- and three-parameter Arrhenius equations are also very good, but the preexponential factors are incorrect. Tunneling and internal rotation corrections must be applied to obtain semiquantitative results.     

Oxidative addition of the fluoromethane C-F bond to Pd. An ab initio benchmark and DFT validation study

We have computed a state-of-the-art benchmark potential energy surface (PES) for two reaction pathways (oxidative insertion, OxIn, and S(N)2) for oxidative addition of the fluoromethane C-F bond to the palladium atom and have used this to evaluate the performance of 26 popular density functionals, covering LDA, GGA, meta-GGA, and hybrid density functionals, for describing these reactions. The ab initio benchmark is obtained by exploring the PES using a hierarchical series of ab initio methods (HF, MP2, CCSD, CCSD(T)) in combination with a hierarchical series of seven Gaussian-type basis sets, up to g polarization. Relativistic effects are taken into account through a full four-component all-electron approach. Our best estimate of kinetic and thermodynamic parameters is -5.3 (-6.1) kcal/mol for the formation of the reactant complex, 27.8 (25.4) kcal/mol for the activation energy for oxidative insertion (OxIn) relative to the separate reactants, 37.5 (31.8) kcal/mol for the activation energy for the alternative S(N)2 pathway, and -6.4 (-7.8) kcal/mol for the reaction energy (zero-point vibrational energy-corrected values in parentheses). Our work highlights the importance of sufficient higher angular momentum polarization functions for correctly describing metal-d-electron correlation. Best overall agreement with our ab initio benchmark is obtained by functionals from all three categories, GGA, meta-GGA, and hybrid DFT, with mean absolute errors of 1.4-2.7 kcal/mol and errors in activation energies ranging from 0.3 to 2.8 kcal/mol. The B3LYP functional compares very well with a slight underestimation of the overall barrier for OxIn by -0.9 kcal/mol. For comparison, the well-known BLYP functional underestimates the overall barrier by -10.1 kcal/mol. The relative performance of these two functionals is inverted with respect to previous findings for the insertion of Pd into the C-H and C-C bonds. However, all major functionals yield correct trends and qualitative features of the PES, in particular, a clear preference for the OxIn over the alternative S(N)2 pathway.     

Nickel-catalyzed alkyl coupling reactions: evaluation of computational methods

The B3LYP, M06, M06L, M062X, MPW1K, and PBE1PBE DFT methods were evaluated for modeling nickel-catalyzed coupling reactions. The reaction consists of a nucleophilic attack by a carbanion equivalent on the nickel complex, S(N)2 attack by the anionic nickel complex on an alkyl halide, and reductive elimination of the coupled alkane product, regenerating the nickel catalyst. On the basis of CCSD(T)//DFT single-point energies, the B3LYP, M06, and PBE1PBE functionals were judged to generate the best ground state geometries. M06 energies are generally comparable or superior to B3LYP and PBE1PBE energies for transition state calculations. The MP2 and CCSD methods were also evaluated for single-point energies at the M06 geometries. The rate-determining step of this reaction was found to be nucleophilic attack of a L(2)NiR anion on the alkyl halide.     

Density functional theory and post-Hartree-Fock studies on molecular structure and harmonic vibrational spectrum of formaldehyde

Density functional theory (DFT) with the Becke's three-parameter exchange correlation functional and the functional of Lee, Yang and Parr, gradient-corrected functionals of Perdew, and Perdew and Wang [the DFT(B3LYP), DFT(B3P86) and DFT(B3PW91) methods, respectively], and several levels of conventional ab initio post-Hartree-Fock theory (second- and fourth-order perturbation theory M?ller-Plesset MP2 and MP4(SDTQ), coupled cluster with the single and double excitations (CCSD), and CCSD with perturbative triple excitation [CCSD(T)], configuration interaction with the single and double excitations [CISD], and quadratic configuration interaction method [QCISD(T)], using several basis sets [ranging from a simple 6-31G(d,p) basis set to a 6-311+ +G(3df, 2pd) one], were applied to study of the molecular structure (geometrical parameters, rotational constants, dipole moment) and harmonized infrared (IR) spectrum of formaldehyde (CH₂O). High-level ab initio methods CCSD(T) and QCISD(T) with the 6-311+ +G(3df, 2pd) predict correctly molecular parameters, vibrational harmonic wavenumbers and the shifts of the harmonic IR spectrum of ¹²CH₂ ¹⁶O upon isotopic substitution. Received: 30 January 1997 / Accepted: 7 May 1997     

Ab initio, density functional theory, and continuum solvation model prediction of the product ratio in the S(N)2 reaction of NO2(-) with CH3CH2Cl and CH3CH2Br in DMSO solution

The reaction pathways for the interaction of the nitrite ion with ethyl chloride and ethyl bromide in DMSO solution were investigated at the ab initio level of theory, and the solvent effect was included through the polarizable continuum model. The performance of BLYP, GLYP, XLYP, OLYP, PBE0, B3PW91, B3LYP, and X3LYP density functionals has been tested. For the ethyl bromide case, our best ab initio calculations at the CCSD(T)/aug-cc-pVTZ level predicts product ratio of 73% and 27% for nitroethane and ethyl nitrite, respectively, which can be compared with the experimental values of 67% and 33%. This translates to an error in the relative DeltaG* of only 0.17 kcal mol(-1). No functional is accurate (deviation <0.5 kcal mol(-1)) for predicting relative DeltaG*. The hybrid X3LYP functional presents the best performance with deviation 0.82 kcal mol(-1). The present problem should be included in the test set used for the evaluation of new functionals.     

Density functional calculations of difluorodiazete structures with Gaussian-orbital-type approach

The comparison of local nonlocal and hybrid DFT methods with RHF , MP 2, CCSD , and CPF ab initio methods in generating geometries and relative energies of cis- and trans-difluorodiazete, SVWN , BLYP , BP 86, BECKE 3LYP , and BECKE 3P 86 DFT methods with 6-311 + g(2d) and 6-311 + + g(3df) basis sets. The geometries generated with RHF ab initio models are quite different from experimental values and energy evaluation prefers the wrong isomer. The hybrid methods give superior geometries while energies evaluated with nonlocal DFT methods are better than the one obtained with MP 2 or CCDS ab initio methods. The results suggest DFT as the method of choice of studying similar systems. © 1996 John Wiley & Sons, Inc.     

Theoretical Aspects of Hydrolysis of Peptide Bonds by Zinc Metalloenzymes

Activation and reaction energies for four model systems capturing the essential physicochemical features of the hydrolysis of the peptide bond have been calculated at various level of theory, including the presumably accurate CCSD(T) calculations. The models studied covered a part of the spectrum encountered in biological systems: the hydrolysis in the absence of metal ions (represented by formamide and Ala–Ala) and the hydrolysis in the presence of one and two zinc(II) ions, mimicking the active sites of mono‐ and dizinc metallopeptidases, respectively (by using thermolysin and glutamate carboxypeptidase II as the model catalytic systems and formamide as the model substrate). The results obtained using CCSD(T)/def2‐TZVP and CCSD(T)/aug‐cc‐pVTZ calculations were used as the benchmark values to which the set of cheaper methods, such as (RI‐)DFT, (RI‐)MP2, and SCS‐MP2, were referenced. It was shown that deviations of 3–5 kcal mol^?1 (translating to 2–3 orders in reaction constants) with respect to the reference CCSD(T) barriers are frequently encountered for many correlated methods and most of studied DFT functionals. It has been concluded that from the set of wave‐function methods, both MP2 and SCS‐MP2 methods can be recommended for smaller models (measured by the mean absolute deviation of the activation barriers over the four systems studied), whereas among the popular DFT functionals, B3LYP and especially M06‐2X are likely to be reasonable choices for calculating the activation barriers of zinc metallopeptidases. Finally, with the model of glutamate carboxypeptidase II, issues related to the convergence of the calculated barriers with the size of the model system used as the representative of the enzyme active site were addressed. The intricacies related to system truncation are demonstrated, and suggest that the correlated wave‐function methods may suffer from problems, such as intramolecular BSSE, which make their usage for the larger system questionable. Altogether, the presented data should contribute to efforts to understand enzymatic catalysis more deeply and to gain control of the accuracy and deficiencies of the available theoretical methods and computational approaches.     

Computing reliable energetics for conjugate addition reactions

The performance of various density functionals along with second-order perturbation treatments has been tested for a set of conjugate addition reactions relevant to stereoselective organocatalysis. It is shown that B3LYP predictions seriously underestimate the reaction energies, whereas two newly designed functionals (M05-2X and M06-2X) and the SCS-MP2 method provide very accurate data. These new methods represent promising alternative approaches in future mechanistic studies.     

Kinetics and Thermodynamics of Reactions Involving Criegee Intermediates: An Assessment of Density Functional Theory and Ab Initio Methods Through Comparison with CCSDT(Q)/CBS Data

Reactions involving Criegee intermediates (CIs, R₁R₂COO) are important in atmospheric ozonolysis models. In recent years, density functional theory (DFT) and CCSD(T)-based ab initio methods are increasingly being used for modeling reaction profiles involving CIs. We obtain highly accurate CCSDT(Q)/CBS reaction energies and barrier heights for ring-closing reactions involving atmospherically important CIs (R₁/R₂ = H, Me, OH, OMe, F, CN, cyclopropene, ethylene, acetaldehyde, and acrolein). We use this benchmark data to evaluate the performance of DFT, double-hybrid DFT (DHDFT), and ab initio methods for the kinetics and thermodynamics of these reactions. We find that reaction energies are more challenging for approximate theoretical procedures than barrier heights. Overall, taking both reaction energies and barrier heights into account, only one of the 58 considered DFT methods (the meta-GGA MN12-L) attains near chemical accuracy, with root-mean-square deviations (RMSDs) of 3.5 (barrier heights) and 4.7 (reaction energies) kJ mol⁻¹. Therefore, MN12-L is recommended for investigations where CCSD(T)-based methods are not computationally feasible. For reaction barrier heights performance does not strictly follow Jacob's Ladder, for example, DHDFT methods do not perform better than conventional DFT methods. Of the ab initio methods, the cost-effective CCSD(T)/CBS(MP2) approach gives the best performance for both reaction energies and barrier heights, with RMSDs of 1.7 and 1.4 kJ mol⁻¹, respectively. All the considered Gaussian-n methods show good performance with RMSDs below the threshold of chemical accuracy for both reaction energies and barrier heights, where G4(MP2) shows the best overall performance with RMSDs of 2.9 and 1.5 kJ mol⁻¹, respectively. © 2019 Wiley Periodicals, Inc.     

Ab initio rate constants from hyperspherical quantum scattering: application to H + CH4 --> H2 + CH3

A general and practical procedure is described for calculating rate constants for chemical reactions using a minimal number of ab initio calculations and quantum-dynamical computations. The method exploits a smooth interpolating functional developed in the hyperspherical representation. This functional is built from two Morse functions and depends on a relatively small number of parameters with respect to conventional functionals developed to date. Thus only a small number of ab initio points needs to be computed. The method is applied to the H + CH4 --> H2 + CH3 reaction. The quantum scattering calculations are performed treating explicitly the bonds being broken and formed. All the degrees of freedom except the breaking and forming bonds are optimized ab initio and harmonic vibrational frequencies and zero-point energies for them are calculated at the MP2(full) level with a cc-pVTZ basis set. Single point energies are calculated at a higher level of theory with the same basis set, namely CCSD(T, full). We report state-to-state cross sections and thermal rate constants for the title reaction and make comparisons with previous results. The calculated rate constants are in good agreement with experiments.     

Hydrogen bonding in the urea dimers and adenine–thymine DNA base pair: anharmonic effects in the intermolecular H-bond and intramolecular H-stretching vibrations

The equilibrium structures, binding energies, vibrational harmonic frequencies, and the anharmonic corrections for two different (cyclic and asymmetric) urea dimers and for the adenine–thymine DNA base pair system have been studied using the second-order Møller–Plesset perturbation theory (MP2) method and different density functional theory (DFT) exchange–correlation (XC) functionals (BLYP, B3LYP, PBE, HCTH407, KMLYP, and BH and HLYP) with the D95V, D95V**, and D95V++** basis sets. The widely used a posteriori Boys–Bernardi or counterpoise correction scheme for basis set superposition error (BSSE) has been included in the calculations to take into account the BSSE effects during geometry optimization (on structure), on binding energies and on the different levels of approximation used for calculating the vibrational frequencies. The results obtained with the ab initio MP2 method are compared with those calculated with different DFT XC functionals; and finally the suitability of these DFT XC functionals to describe intermolecular hydrogen bonds as well as harmonic frequencies and the anharmonic corrections is assessed and discussed.     

Accurate interaction energies of hydrogen-bonded nucleic acid base pairs

Hydrogen-bonded nucleic acids base pairs substantially contribute to the structure and stability of nucleic acids. The study presents reference ab initio structures and interaction energies of selected base pairs with binding energies ranging from -5 to -47 kcal/mol. The molecular structures are obtained using the RI-MP2 (resolution of identity MP2) method with extended cc-pVTZ basis set of atomic orbitals. The RI-MP2 method provides results essentially identical with the standard MP2 method. The interaction energies are calculated using the Complete Basis Set (CBS) extrapolation at the RI-MP2 level. For some base pairs, Coupled-Cluster corrections with inclusion of noniterative triple contributions (CCSD(T)) are given. The calculations are compared with selected medium quality methods. The PW91 DFT functional with the 6-31G basis set matches well the RI-MP2/CBS absolute interaction energies and reproduces the relative values of base pairing energies with a maximum relative error of 2.6 kcal/mol when applied with Becke3LYP-optimized geometries. The Becke3LYP DFT functional underestimates the interaction energies by few kcal/mol with relative error of 2.2 kcal/mol. Very good performance of nonpolarizable Cornell et al. force field is confirmed and this indirectly supports the view that H-bonded base pairs are primarily stabilized by electrostatic interactions.     

Theoretical investigation of clusters of phosphorus and arsenic: fascination and temptation of high symmetries

We present a theoretical study of the energetic and thermodynamic stability of selected phosphorus and arsenic clusters containing 18 to 168 atoms. For this purpose we employ MP2 as well as DFT functionals BP86 and B3LYP with extended basis sets. All procedures predict the family of one-dimensional polymers X18+12n, each with 2n-1 isomers of virtually identical energy, to be more stable than other structures investigated so far. Furthermore, islands of stability result for ring-shaped clusters X24n with Dnd symmetry for n=4 (only for arsenic), 5, 6, and 7. Phosphorus and arsenic show otherwise a very similar behavior. An investigation of basis set effects shows that a doubly polarized triple zeta valence basis (TZVPP) is both necessary and sufficient. In comparison to the reliable spin component scaled MP2 (SCS-MP2) procedure, DFT methods underestimate and MP2 overestimates the stability of larger clusters; the discrepancy increases with the number of atoms. The addition of a long-range dispersion correction to B3LYP energies does not rectify the shortcomings of DFT in comparison with SCS-MP2.