Extrapolation and Scaling of the DFT-SAPT Interaction Energies toward the Basis Set Limit

An efficient method by combining different basis sets and SAPT levels

In symmetry-adapted perturbation theory (SAPT), accurate calculations on non-covalent interaction (NCI) for large complexes with more than 50 atoms are time-consuming using large basis sets. More efficient ones with smaller basis sets usually result in poor prediction in terms of dispersion and overall energies. In this study, we propose two composite methods with baseline calculated at SAPT2/aug-cc-pVDZ and SAPT2/aug-cc-pVTZ with dispersion term corrected at SAPT2+ level using bond functions and smaller basis set with δ MP2 corrections respectively. Benchmark results on representative NCI data sets, such as S22, S66, and so forth, show significant improvements on the accuracy compared to the original SAPT Silver standard and comparable to SAPT Gold standard in some cases with much less computational cost.

Graphene as Nanocarrier for Gold(I)-Monocarbene Complexes: Strength and Nature of Physisorption

Gold(I) metal complexes are finding increasing applications as therapeutic agents against a variety of diseases. As their potential use as effective metallodrugs is continuously confirmed, the issue of their administration, distribution and delivery to desired biological targets emerges. Graphene and its derivatives possess attractive properties in terms of high affinity and low toxicity, suggesting that they can efficaciously be used as drug nanocarriers. In the present study, we computationally address the adsorption of a gold(I) N-heterocyclic monocarbene, namely, IMeAuCl (where IMe = 1,3-dimethylimidazol-2-ylidene), on graphene. The Au(I) N-heterocyclic carbene family has indeed shown promising anticancer activity and the N-heterocyclic ring could easily interact with planar graphene nanostructures. By means of high-level electronic structure approaches, we investigated the strength and nature of the involved interaction using small graphene prototypes, which allow us to benchmark the best-performing DFT functionals as well as assess the role of the different contributions to total interaction energies. Moreover, realistic adsorption enthalpies and free energy values are obtained by exploiting the optimal DFT method to describe the drug adsorption on larger graphene models. Such values (ΔHads = -18.4 kcal/mol and ΔGads= -7.20 kcal/mol for the largest C150H30 model) indicate a very favorable adsorption, mainly arising from the dispersion component of the interaction, with the electrostatic attraction also playing a non-negligible role.

Comprehensive Basis-Set Testing of Extended Symmetry-Adapted Perturbation Theory and Assessment of Mixed-Basis Combinations to Reduce Cost

Hybrid or "extended" symmetry-adapted perturbation theory (XSAPT) replaces traditional SAPT's treatment of dispersion with better performing alternatives while at the same time extending two-body (dimer) SAPT to a many-body treatment of polarization using a self-consistent charge embedding procedure. The present work presents a systematic study of how XSAPT interaction energies and energy components converge with respect to the choice of Gaussian basis set. Errors can be reduced in a systematic way using correlation-consistent basis sets, with aug-cc-pVTZ results converged within <0.1 kcal/mol. Similar (if slightly less systematic) behavior is obtained using Karlsruhe basis sets at much lower cost, and we introduce new versions with limited augmentation that are even more efficient. Pople-style basis sets, which are more efficient still, often afford good results if a large number of polarization functions are included. The dispersion models used in XSAPT afford much faster basis-set convergence as compared to the perturbative description of dispersion in conventional SAPT, meaning that "compromise" basis sets (such as jun-cc-pVDZ) are no longer required and benchmark-quality results can be obtained using triple-ζ basis sets. The use of diffuse functions proves to be essential, especially for the description of hydrogen bonds. The "δ(Hartree-Fock)" correction for high-order induction can be performed in double-ζ basis sets without significant loss of accuracy, leading to a mixed-basis approach that offers 4× speedup over the existing (cubic scaling) XSAPT approach.

Description of Chiral Complexes within Functional-Group Symmetry-Adapted Perturbation Theory-The Case of (S/R)-Carvone with Derivatives of (-)-Menthol

Symmetry-adapted perturbation theory (SAPT) and functional-group SAPT (F-SAPT) are applied to examine differences in interaction energies of diastereoisomeric complexes of two chiral molecules of natural origin: (S/R)-carvone with (-)-menthol. The study is extended by including derivatives of menthol with its hydroxy group exchanged by another functional group, thus examining the substituent effect of the interaction and the interaction differences between diastereoisomers. The partitioning of the interaction energy into functional-group components allows one to explain this phenomenon by the mutual cancellation of attractive and repulsive interactions between functional groups. In some cases, one can identify dominant chiral interactions between groups of atoms of carvone and menthol derivatives, while in many other instances, no major interaction can be distinguished and the net chiral difference results from subtle near cancellation of several smaller terms. Our results indicate that the F-SAPT method can be faithfully utilized for such analyses.

HFLD: A Nonempirical London Dispersion-Corrected Hartree-Fock Method for the Quantification and Analysis of Noncovalent Interaction Energies of Large Molecular Systems †

A nonempirical quantum mechanical method for the efficient and accurate quantification and analysis of intermolecular interactions is presented and tested on existing benchmark sets. The leading idea here is to focus on the intermolecular part of the correlation energy that contains the all-important London dispersion (LD) interaction. To keep the cost of the method low, essentially at the level of a Hartree-Fock (HF) calculation, the intramolecular part of the correlation energy is neglected. We also neglect the nondispersive parts of the intermolecular correlation energy. This scheme that we denote as Hartree-Fock plus London dispersion (HFLD) can be readily realized on the basis of the recently reported multilevel implementation of the domain-based local pair natural orbital coupled-cluster (DLPNO-CC) theory in conjunction with the well-established local energy decomposition (LED) analysis. The accuracy and efficiency of the HFLD method are evaluated on rare gas dimers, on the S66 and L7 benchmark sets of noncovalent interactions, and on an additional set (LP14) consisting of bulky Lewis pairs held together by intermolecular interactions of various strengths, with interaction energies ranging from -8 to -107 kcal/mol. It is first shown that the LD energy calculated with this approach is essentially identical to that obtained from the full DLPNO-CCSD(T)/LED calculation, with a mean absolute error of 0.2 kcal/mol on the S66 benchmark set. Moreover, in terms of the overall interaction energies, the HFLD method shows an efficiency that is comparable to that of the HF method, while retaining an accuracy between that of the DLPNO-CCSD and DLPNO-CCSD(T) schemes. Since the underlying DLPNO-CCSD method is linear scaling with respect to the system size, the HFLD approach also does not lead to new bottlenecks for large systems. As an illustrative example of its efficiency, the HFLD scheme was applied to the interaction between the substrate and the residues in the active site of the cyclohexanone monooxygenase enzyme. The excellent cost/performance ratio indicates that the HFLD method opens new avenues for the accurate calculation and analysis of noncovalent interaction energies in large molecular systems.

Symmetry-Adapted Perturbation Theory Decomposition of the Reaction Force: Insights into Substituent Effects Involved in Hemiacetal Formation Mechanisms

The decomposition of the reaction force based on symmetry-adapted perturbation theory (SAPT) has been proposed. This approach was used to investigate the substituent effects along the reaction coordinate pathway for the hemiacetal formation mechanism between methanol and substituted aldehydes of the form CX₃CHO (X = H, F, Cl, and Br), providing a quantitative evaluation of the reaction-driving and reaction-retarding force components. Our results highlight the importance of more favorable electrostatic and induction effects in the reactions involving halogenated aldehydes that leads to lower activation energy barriers. These substituent effects are further elucidated by applying the functional-group partition of symmetry-adapted perturbation theory (F-SAPT). The results show that the reaction is largely driven by favorable direct noncovalent interactions between the CX₃ group on the aldehyde and the OH group on methanol.

Accurate and Efficient ab Initio Calculations for Supramolecular Complexes: Symmetry-Adapted Perturbation Theory with Many-Body Dispersion

Symmetry-adapted perturbation theory (SAPT) provides a chemically meaningful energy decomposition scheme for nonbonded interactions that is useful for interpretive purposes. Although formally a dimer theory, we have previously introduced an "extended" version (XSAPT) that incorporates many-body polarization via self-consistent charge embedding. Here, we extend the XSAPT methodology to include nonadditive dispersion, using a modified form of the many-body dispersion (MBD) method of Tkatchenko and co-workers. Dispersion interactions beyond the pairwise atom-atom approximation improve total interaction energies even in small systems, and for large π-stacked complexes these corrections can amount to several kilocalories per mole. The XSAPT+MBD method introduced here achieves errors of ≲1 kcal/mol (as compared to high-level ab initio benchmarks) for the L7 data set of large dispersion-bound complexes and ≲4 kcal/mol (as compared to experiment) for the S30L data set of host-guest complexes. This is superior to the best contemporary density functional methods for noncovalent interactions, at comparable or lower cost. XSAPT+MBD represents a promising method for application to supramolecular assemblies, including protein-ligand binding.

Local Energy Decomposition of Open-Shell Molecular Systems in the Domain-Based Local Pair Natural Orbital Coupled Cluster Framework

Local energy decomposition (LED) analysis decomposes the interaction energy between two fragments calculated at the domain-based local pair natural orbital CCSD(T) (DLPNO-CCSD(T)) level of theory into a series of chemically meaningful contributions and has found widespread applications in the study of noncovalent interactions. Herein, an extension of this scheme that allows for the analysis of interaction energies of open-shell molecular systems calculated at the UHF-DLPNO-CCSD(T) level is presented. The new scheme is illustrated through applications to the CH₂···X (X = He, Ne, Ar, Kr, and water) and heme···CO interactions in the low-lying singlet and triplet spin states. The results are used to discuss the mechanism that governs the change in the singlet–triplet energy gap of methylene and heme upon adduct formation.

Understanding non-covalent interactions in larger molecular complexes from first principles

Non-covalent interactions pervade all matter and play a fundamental role in layered materials, biological systems, and large molecular complexes. Despite this, our accumulated understanding of non-covalent interactions to date has been mainly developed in the tens-of-atoms molecular regime. This falls considerably short of the scales at which we would like to understand energy trends, structural properties, and temperature dependencies in materials where non-covalent interactions have an appreciable role. However, as more reference information is obtained beyond moderately sized molecular systems, our understanding is improving and we stand to gain pertinent insights by tackling more complex systems, such as supramolecular complexes, molecular crystals, and other soft materials. In addition, accurate reference information is needed to provide the drive for extending the predictive power of more efficient workhorse methods, such as density functional approximations that also approximate van der Waals dispersion interactions. In this perspective, we discuss the first-principles approaches that have been used to obtain reference interaction energies for beyond modestly sized molecular complexes. The methods include quantum Monte Carlo, symmetry-adapted perturbation theory, non-canonical coupled cluster theory, and approaches based on the random-phase approximation. By considering the approximations that underpin each method, the most accurate theoretical references for supramolecular complexes and molecular crystals to date are ascertained. With these, we also assess a handful of widely used exchange-correlation functionals in density functional theory. The discussion culminates in a framework for putting into perspective the accuracy of high-level wavefunction-based methods and identifying future challenges.

Effect of Electron Correlation on Intermolecular Interactions: A Pair Natural Orbitals Coupled Cluster Based Local Energy Decomposition Study

The development of post-Hartree-Fock (post-HF) energy decomposition schemes that are able to decompose the HF and correlation components of the interaction energy into chemically meaningful contributions is a very active field of research. One of the challenges is to provide a clear-cut quantification to the elusive London dispersion component of the intermolecular interaction. London dispersion is well-known to be a pure correlation effect, and as such it is not properly described by mean field theories. In this context, we have recently developed the local energy decomposition (LED) analysis, which provides a chemically meaningful decomposition of the interaction energy between two or more fragments computed at the domain-based local pair natural orbitals coupled cluster (DLPNO-CCSD(T)) level of theory. In this work, this scheme is used in conjunction with other interpretation tools to study a series of molecular adducts held together by intermolecular interactions of different natures. The HF and correlation components of the interaction energy are thus decomposed into a series of chemically meaningful contributions. Emphasis is placed on discussing the physical effects associated with the inclusion of electron correlation. It is found that four distinct physical effects can contribute to the magnitude of the correlation part of intermolecular binding energies (Δ E_int^C): (i) London dispersion, (ii) the correlation correction to the reference induction energy, (iii) the correlation correction to the electron sharing process, and (iv) the correlation correction to the permanent electrostatics. As expected, the largest contribution to the correlation binding energy of neutral, apolar molecules is London dispersion, as in the argon dimer case. In contrast, the correction for the HF induction energy dominates Δ E_int^C in systems in which an apolar molecule interacts with charged or strongly polar species, as in Ar-Li⁺. This effect has its origin in the systematic underestimation of polarizabilities at the HF level of theory. For similar reasons, electron sharing largely contributes to the correlation binding energy of covalently bound molecules, as in the beryllium dimer case. Finally, the correction for HF permanent electrostatics significantly contributes to Δ E_int^C in molecules with strong dipoles, such as water and hydrogen fluoride dimers. This effect originates from the characteristic overestimation of dipole moments at the HF level of theory, leading in some cases to positive Δ E_int^C values. Our results are apparently in contrast to the widely accepted view that Δ E_int^C is typically dominated by London dispersion, at least, in the strongly interacting region. Clearly, post-HF energy decomposition schemes are very powerful tools to analyze, categorize, and understand the various contributions to the intermolecular interaction energy. Hopefully, this will eventually lead to insights that are helpful in designing systems with tailored properties. All analysis tools presented in this work will be available free of charge in the next release of the ORCA program package.

On the Nature of σ-σ, σ-π, and π-π Stacking in Extended Systems

Stacking interactions have been evaluated, employing computational methods, in dimers formed by analogous aliphatic and aromatic species of increasing size. Changes in stability as the systems become larger are mostly controlled by the balance of increasing repulsion and dispersion contributions, while electrostatics plays a secondary but relevant role. The interaction energy increases as the size of the system grows, but it does much faster in π-π dimers than in σ-π complexes and more remarkably than in σ-σ dimers. The main factor behind the larger stability of aromatic dimers compared to complexes containing aliphatic molecules is related to changes in the properties of the aromatic systems due to electron delocalization leading to larger dispersion contributions. Besides, an extra stabilization in π-π complexes is due to the softening of the repulsive wall in aromatic species that allows the molecules to come closer.

Comparison of the DFT-SAPT and Canonical EDA Schemes for the Energy Decomposition of Various Types of Noncovalent Interactions

Interaction energies computed with density functional theory can be divided into physically meaningful components by symmetry-adapted perturbation theory (DFT-SAPT) or the canonical energy decomposition analysis (EDA). In this work, the decomposition results obtained by these schemes were compared for more than 200 hydrogen-, halogen-, and pnicogen-bonded, dispersion-bound, and mixed complexes to investigate their similarity in the evaluation of the nature of noncovalent interactions. BLYP functional with D3(BJ) correction was used for the EDA scheme, whereas asymptotically corrected PBE0 functional for DFT-SAPT provided some of the best combinations for description of noncovalent interactions. Both schemes provide similar results concerning total interaction energies and insight into the individual energy components. For most complexes, the dominant energetic term was identified equally by both decomposition schemes. Because the canonical EDA is computationally less demanding than the DFT-SAPT, the former can be especially used in cases where the systems investigated are very large.

Local energy decomposition analysis of hydrogen-bonded dimers within a domain-based pair natural orbital coupled cluster study

The local energy decomposition (LED) analysis allows for a decomposition of the accurate domain-based local pair natural orbital CCSD(T) [DLPNO-CCSD(T)] energy into physically meaningful contributions including geometric and electronic preparation, electrostatic interaction, interfragment exchange, dynamic charge polarization, and London dispersion terms. Herein, this technique is employed in the study of hydrogen-bonding interactions in a series of conformers of water and hydrogen fluoride dimers. Initially, DLPNO-CCSD(T) dissociation energies for the most stable conformers are computed and compared with available experimental data. Afterwards, the decay of the LED terms with the intermolecular distance (r) is discussed and results are compared with the ones obtained from the popular symmetry adapted perturbation theory (SAPT). It is found that, as expected, electrostatic contributions slowly decay for increasing r and dominate the interaction energies in the long range. London dispersion contributions decay as expected, as r-6. They significantly affect the depths of the potential wells. The interfragment exchange provides a further stabilizing contribution that decays exponentially with the intermolecular distance. This information is used to rationalize the trend of stability of various conformers of the water and hydrogen fluoride dimers.

GW100: Comparison of Methods and Accuracy of Results Obtained with the WEST Code

The reproducibility of calculations carried out within many-body perturbation theory at the G₀ W₀ level is assessed for 100 closed shell molecules and compared to that of density functional theory. We consider vertical ionization potentials (VIP) and electron affinities (VEA) obtained with five different codes: BerkeleyGW, FHI-aims, TURBOMOLE, VASP, and WEST. We review the approximations and parameters that control the accuracy of G₀ W₀ results in each code, and we discuss in detail the effect of extrapolation techniques for the parameters entering the WEST code. Differences between the VIP and VEA computed with the various codes are within ∼60 and ∼120 meV, respectively, which is up to four times larger than in the case of the best results obtained with DFT codes. Vertical ionization potentials are validated against experiment and CCSD(T) quantum chemistry results showing a mean absolute relative error of ∼4% for data obtained with WEST. Our analysis of the differences between localized orbitals and plane-wave implementations points out molecules containing Cu, I, Ga, and Xe as major sources of discrepancies, which call for a re-evaluation of the pseudopotentials used for these systems in G₀ W₀ calculations.

An optimized charge penetration model for use with the AMOEBA force field

The principal challenge of using classical physics to model biomolecular interactions is capturing the nature of short-range interactions that drive biological processes from nucleic acid base stacking to protein-ligand binding. In particular most classical force fields suffer from an error in their electrostatic models that arises from an ability to account for the overlap between charge distributions occurring when molecules get close to each other, known as charge penetration. In this work we present a simple, physically motivated model for including charge penetration in the AMOEBA (Atomic Multipole Optimized Energetics for Biomolecular Applications) force field. With a function derived from the charge distribution of a hydrogen-like atom and a limited number of parameters, our charge penetration model dramatically improves the description of electrostatics at short range. On a database of 101 biomolecular dimers, the charge penetration model brings the error in the electrostatic interaction energy relative to the ab initio SAPT electrostatic interaction energy from 13.4 kcal mol^-1 to 1.3 kcal mol^-1. The model is shown not only to be robust and transferable for the AMOEBA model, but also physically meaningful as it universally improves the description of the electrostatic potential around a given molecule.

Dissecting the concave-convex π-π interaction in corannulene and sumanene dimers: SAPT(DFT) analysis and performance of DFT dispersion-corrected methods

The characteristics of the concave-convex π-π interactions are evaluated in 32 buckybowl dimers formed by corannulene, sumanene, and two substituted sumanenes (with S and CO groups), using symmetry-adapted perturbation theory [SAPT(DFT)] and density functional theory (DFT). According to our results, the main stabilizing contribution is dispersion, followed by electrostatics. Regarding the ability of DFT methods to reproduce the results obtained with the most expensive and rigorous methods, TPSS-D seems to be the best option overall, although its results slightly tend to underestimate the interaction energies and to overestimate the equilibrium distances. The other two tested DFT-D methods, B97-D2 and B3LYP-D, supply rather reasonable results as well. M06-2X, although it is a good option from a geometrical point of view, leads to too weak interactions, with differences with respect to the reference values amounting to about 4 kcal/mol (25% of the total interaction energy). © 2017 Wiley Periodicals, Inc.

Noncovalent interactions between cisplatin and graphene prototypes

Cisplatin (CP) has been widely used as an anticancer drug for more than 30 years despite severe side effects due to its low bioavailability and poor specificity. For this reason, it is paramount to study and design novel nanomaterials to be used as vectors capable to effectively deliver the drug to the biological target. The CP square-planar geometry, together with its low water solubility, suggests that it could be possibly easily adsorbed on 2D graphene nanostructures through the interaction with the related highly conjugated π-electron system. In this work, pyrene has been first selected as the minimum approximation to the graphene plane, which allows to properly study the noncovalent interactions determining the CP adsorption. In particular, electronic structure calculations at the MP2C and DFT-SAPT levels of theory have allowed to obtain benchmark interaction energies for some limiting configurations of the CP-pyrene complex, as well as to assess the role of the different contributions to the total interaction: it has been found that the parallel configurations of the aggregate are mainly stabilized around the minimum region by dispersion, in a similar way as for complexes bonded through π-π interactions. Then, the benchmark interaction energies have been used to test corresponding estimations obtained within the less expensive DFT to validate an optimal exchange-correlation functional which includes corrections to take properly into account for the dispersion contribution. Reliable DFT interaction energies have been therefore obtained for CP adsorbed on graphene prototypes of increasing size, ranging from coronene, ovalene, and up to C₁₅₀ H₃₀ . Finally, DFT geometry optimizations and frequency calculations have also allowed a reliable estimation of the adsorption enthalpy of CP on graphene, which is found particularly favorable (about -20 kcal/mol at 298 K and 1 bar) being twice that estimated for the corresponding benzene adsorption. © 2017 Wiley Periodicals, Inc.

The role of the σ-holes in stability of non-bonded chalcogenide⋯benzene interactions: the ground and excited states

Different characters of molecular electrostatic potential (MEP) in the ground and excited states of chalcogenides are responsible for changes in conformer stability of T-shape and stacked non-bonded chalcogenide–benzene complexes upon electronic excitation.

Dispersion Interactions in QM/EFP

The dispersion energy term between quantum-mechanical (QM) and classical (represented by effective fragment potentials, EFP) subsystems is developed and implemented. A new formulation is based on long-range perturbation theory and uses dynamic polarizability tensors of the effective fragments and electric field integrals and orbital energies of the quantum-mechanical subsystem. No parametrization is involved. The accuracy of the QM-EFP dispersion energy is tested on a number of model systems; the average mean unsigned error is 0.8 kcal/mol or 13% with respect to the symmetry adapted perturbation theory on the S22 data set of noncovalent interactions. The computational cost of the dispersion energy computation is low compared to the self-consistent field calculation of the QM subsystem. The dispersion energy is sensitive to the level of theory employed for the QM part and to the electrostatic interactions in the system. The latter means that the dispersion interactions in the QM/EFP method are not purely two-body but have more complex many-body behavior.

Estimating and modeling charge transfer from the SAPT induction energy

Recent studies using quantum mechanics energy decomposition methods, for example, SAPT and ALMO, have revealed that the charge transfer energy may play an important role in short ranged inter-molecular interactions, and have a different distance dependence comparing with the polarization energy. However, the charge transfer energy component has been ignored in most current polarizable or non-polarizable force fields. In this work, first, we proposed an empirical decomposition of SAPT induction energy into charge transfer and polarization energy that mimics the regularized SAPT method (ED-SAPT). This empirical decomposition is free of the divergence issue, hence providing a good reference for force field development. Then, we further extended this concept in the context of AMOEBA polarizable force field, proposed a consistent approach to treat the charge transfer phenomenon. Current results show a promising application of this charge transfer model in future force field development. © 2017 Wiley Periodicals, Inc.

Noncovalent Interactions in Specific Recognition Motifs of Protein-DNA Complexes

In view of the importance of protein-DNA interactions in biological processes, we extracted from the Protein Data Bank several one-to-one complexes of amino acids with nucleotides that matched certain geometric and energetic specificity criteria and investigated them using quantum chemistry methods. The CCSD(T)/CBS interaction energies were used as a benchmark to compare the performance of the MP2.5, MP2-F12, DFT-D3, and PM6-D3H4 methods. All methods yielded good agreement with the reference values, with declining accuracy from MP2.5 to PM6-D3H4. Regardless of the site of interaction, the minima found after full optimization in implicit solvent with high dielectric constant were close to the structures experimentally detected in protein-DNA complexes. According to DFT-SAPT analysis, the nature of noncovalent interactions strongly depends on the type of amino acid. The negatively charged sugar-phosphate backbone of DNA heavily influences the strength of interactions and must be included in the computational model, especially in the case of interactions with charged amino acids.

A theoretical study of complexes formed between cations and curved aromatic systems: electrostatics does not always control cation–π interaction

Cation–π interactions in curved aromatic systems are not controlled by electrostatics; induction and dispersion dominate in most cases studied.

Non-covalent interactions in anisole–(CO2)n (n = 1, 2) complexes

Non-covalent interactions are a ubiquitous binding motif and a challenge for theory and experiments.

Prediction of cyclohexane-water distribution coefficient for SAMPL5 drug-like compounds with the QMPFF3 and ARROW polarizable force fields

We present the performance of blind predictions of water-cyclohexane distribution coefficients for 53 drug-like compounds in the SAMPL5 challenge by three methods currently in use within our group. Two of them utilize QMPFF3 and ARROW, polarizable force-fields of varying complexity, and the third uses the General Amber Force-Field (GAFF). The polarizable FF's are implemented in an in-house MD package, Arbalest. We find that when we had time to parametrize the functional groups with care (batch 0), the polarizable force-fields outperformed the non-polarizable one. Conversely, on the full set of 53 compounds, GAFF performed better than both QMPFF3 and ARROW. We also describe the torsion-restrain method we used to improve sampling of molecular conformational space and thus the overall accuracy of prediction. The SAMPL5 challenge highlighted several drawbacks of our force-fields, such as our significant systematic over-estimation of hydrophobic interactions, specifically for alkanes and aromatic rings.

Intermolecular symmetry-adapted perturbation theory study of large organic complexes

Binding energies for the complexes of the S12L database by Grimme [Chem. Eur. J. 18, 9955 (2012)] were calculated using intermolecular symmetry-adapted perturbation theory combined with a density-functional theory description of the interacting molecules. The individual interaction energy decompositions revealed no particular change in the stabilisation pattern as compared to smaller dimer systems at equilibrium structures. This demonstrates that, to some extent, the qualitative description of the interaction of small dimer systems may be extrapolated to larger systems, a method that is widely used in force-fields in which the total interaction energy is decomposed into atom-atom contributions. A comparison of the binding energies with accurate experimental reference values from Grimme, the latter including thermodynamic corrections from semiempirical calculations, has shown a fairly good agreement to within the error range of the reference binding energies.

General Model for Treating Short-Range Electrostatic Penetration in a Molecular Mechanics Force Field

Classical molecular mechanics force fields typically model interatomic electrostatic interactions with point charges or multipole expansions, which can fail for atoms in close contact due to the lack of a description of penetration effects between their electron clouds. These short-range penetration effects can be significant and are essential for accurate modeling of intermolecular interactions. In this work we report parametrization of an empirical charge-charge function previously reported (Piquemal J.-P.; J. Phys. Chem. A2003, 107, 10353) to correct for the missing penetration term in standard molecular mechanics force fields. For this purpose, we have developed a database (S101×7) of 101 unique molecular dimers, each at 7 different intermolecular distances. Electrostatic, induction/polarization, repulsion, and dispersion energies, as well as the total interaction energy for each complex in the database are calculated using the SAPT2+ method (Parker T. M.; J. Chem. Phys.2014, 140, 094106). This empirical penetration model significantly improves agreement between point multipole and quantum mechanical electrostatic energies across the set of dimers and distances, while using only a limited set of parameters for each chemical element. Given the simplicity and effectiveness of the model, we expect the electrostatic penetration correction will become a standard component of future molecular mechanics force fields.

On the interaction between the imidazolium cation and aromatic amino acids. A computational study

Phe, Tyr and Trp form parallel complexes with cation⋯π interactions. His complexes are the strongest, but without making contact with the aromatic cloud.

C-H···π interactions and the nature of the donor carbon atom

The influence of multiple substituents (F, CH3, NO2, CN, Cl, OH and NH2) on the C-H···π interaction in benzene-ethylene complex was investigated using the estimated CCSD(T) method and complete basis set limit. The results were compared with our earlier reported complexes of benzene-acetylene and benzene-methane, thus completing the sp, sp(2) and sp(3) series of C-H donors. The stabilization energy values for multiple fluoro-substituted benzene-ethylene complexes are found to be very close to those of the multiple fluoro-substituted benzene-methane complexes. Expectedly, the stabilization energies for the multiple methyl-substituted benzene-ethylene complexes lie between those of the multiple methyl-substituted benzene-methane and benzene-acetylene complexes. Energy decomposition analysis using the DFT-SAPT method predicts the dispersion energy to be dominant, similar to the benzene-methane complexes. For the symmetrically disubstituted complexes (-OH, -Cl, -NH2, -CN and -NO2), additional C-H···X interaction was observed, possibly due to the angular orientation of the ethylene molecule. Multidimensional correlation analysis between the electrostatic, dispersion and exchange-repulsion with the C-H···π interaction distance (r), Hammett constant (σ) and the molar refractivity (MR) revealed strong correlation between dispersion energy and the C-H···π interaction distance (r) as well as molar refractivity (MR).

On the origin of the substantial stabilisation of the electron-donor 1,3-dithiole-2-thione-4-carboxyclic acid···I2 and DABCO···I2 complexes

The stabilisation energies of the crystal structures of 1,3-dithiole-2-thione-4-carboxyclic acid···I2 and DABCO···I2 complexes determined by the CCSD(T)/CBS method are very large and exceed 8 and 15 kcal mol(-1), respectively. The DFT-D method (B97-D3/def2-QZVP) strongly overestimates these stabilisation energies, which support the well-known fact that the DFT-D method is not very applicable to the study of charge-transfer complexes. On the other hand, the M06-2X/def2-QZVP method provides surprisingly reliable energies. A DFT-SAPT analysis has shown that a substantial stabilisation of these complexes arises from the charge-transfer energy included in the induction energy and that the respective induction energy is much larger than that of other non-covalently bound complexes. The total stabilisation energies of the complexes mentioned as well as of those where iodine has been replaced by lighter halogens (Br2 and Cl2) or by hetero systems (IF, ICH3, N2) correlate well with the magnitude of the σ-hole (Vs,max value) as well as with the LUMO energy. The nature of the stabilisation of all complexes between both electron donors and X2 (X = I, Br, Cl, N) systems is explained by the magnitude of the σ-hole but surprisingly also by the values of the electric quadrupole moment of these systems. Evidently, the nature of the stabilisation of halogen-bonded complexes between electron donors and systems where the first non-zero electric multipole moment is the quadrupole moment can be explained not only by the recently introduced concept of the σ-hole but also by the classical concept of electric quadrupole moments.

Quantum Mechanical Calculation of Noncovalent Interactions: A Large-Scale Evaluation of PMx, DFT, and SAPT Approaches

Quantum mechanical (QM) calculations of noncovalent interactions are uniquely useful as tools to test and improve molecular mechanics force fields and to model the forces involved in biomolecular binding and folding. Because the more computationally tractable QM methods necessarily include approximations, which risk degrading accuracy, it is essential to evaluate such methods by comparison with high-level reference calculations. Here, we use the extensive Benchmark Energy and Geometry Database (BEGDB) of CCSD(T)/CBS reference results to evaluate the accuracy and speed of widely used QM methods for over 1200 chemically varied gas-phase dimers. In particular, we study the semiempirical PM6 and PM7 methods; density functional theory (DFT) approaches B3LYP, B97-D, M062X, and ωB97X-D; and symmetry-adapted perturbation theory (SAPT) approach. For the PM6 and DFT methods, we also examine the effects of post hoc corrections for hydrogen bonding (PM6-DH+, PM6-DH2), halogen atoms (PM6-DH2X), and dispersion (DFT-D3 with zero and Becke–Johnson damping). Several orders of the SAPT expansion are also compared, ranging from SAPT0 up to SAPT2+3, where computationally feasible. We find that all DFT methods with dispersion corrections, as well as SAPT at orders above SAPT2, consistently provide dimer interaction energies within 1.0 kcal/mol RMSE across all systems. We also show that a linear scaling of the perturbative energy terms provided by the fast SAPT0 method yields similar high accuracy, at particularly low computational cost. The energies of all the dimer systems from the various QM approaches are included in the Supporting Information, as are the full SAPT2+(3) energy decomposition for a subset of over 1000 systems. The latter can be used to guide the parametrization of molecular mechanics force fields on a term-by-term basis.