Are Dispersion Corrections Accurate Outside Equilibrium? A Case Study on Benzene

Abstract

Modern approaches to modelling dispersion forces are becoming increasingly accurate, and can predict accurate binding distances and energies. However, it is possible that these successes reflect a fortuitous cancellation of errors at equilibrium. Thus, in this work we investigate whether a selection of modern dispersion methods agree with benchmark calculations across several potential-energy curves of the benzene dimer to determine if they are capable of describing forces and energies outside equilibrium. We find the exchange-hold dipole moment (XDM) model describes most cases with the highest overall agreement with reference data for energies and forces, with many-body dispersion (MBD) and its fractionally ionic (FI) variant performing essentially as well. Popular approaches, such as Grimme-D and van der Waals density functional approximations (vdW-DFAs) underperform on our tests. The meta-GGA M06-L is surprisingly good for a method without explicit dispersion corrections. Some problems with SCAN+rVV10 are uncovered and briefly discussed.<br>

Faraday-Cage Screening Reveals Intrinsic Aspects of the van der Waals Attraction

“…By contrast, in interactions between benzene dimers, MBD significantly outperforms D3 at near-contact distances. 38 These two cases, and the screening reported here, illustrate the competitive (and hard-todefine) role played by Dobson-A and -B terms at atomic scales, e.g., the effect of the horizontal field lines in the right diagram of Fig. 2 can by mimicked by local changes to the response function.…”

“…Of general interest also is how different methods treat the van der Waals force during small excursions from equilibrium. 38 More significantly, the observation that the van der Waals force can be switched off severely challenges the traditional pairwise-additive point of view and hence provides an external reference frame for this discussion. Here, combining the results from the different computational methods reveal that the switching off of the van der Waals interaction is a Faraday cage effect.…”

Double-Hybrid Functionals and Tailored Basis Set: Fullerene (C60) Dimer and Isomers as Test Cases

“…Consequently, single point PBE-QIDH calculations with the latter basis set are roughly 4 times faster than those performed with the larger basis (for the C 60 dimer at the equilibrium distance) if all other technical factors remain the same.Beside the interaction energy at a fixed geometry, it is also interesting to analyze the full potential energy profile along the distance between the centers of mass of the two moieties constituting the C 60 dimer. This test is even more stringent since the dispersion corrections could fail to reproduce the energy profile far from energy minima41 , sometime giving spurious discontinuities42 . In our case, however, the validity of this test should be modulated, since the available reference values are calculated at the CEPA-1/cc-pVDZ level, while larger basis set are normally used to reach energies at convergence with this level of theory (triple-or quadruple-, see for instance reference 43 and 44).…”

NENCI-2021 Part I: A Large Benchmark Database of Non-Equilibrium Non-Covalent Interactions Emphasizing Close Intermolecular Contacts

“…However, we would argue that the next generation of theoretical approaches for describing NCIs would tremendously benefit by addressing the following challenges in an accurate, reliable, and computationally efficient manner: (i ) the need to describe NCIs in large molecular and condensed-phase systems (i.e., collective many-body effects, solvation/solvent effects, simultaneous treatment of short-, intermediate-, and long-range NCIs on the same footing); [39][40][41][42][43][44] (ii ) the need to describe the diverse types of NCIs on the same footing (i.e., similar performance for hydrogen bonding, π-π stacking, dispersion, ion-π interactions, etc); [45][46][47][48] and (iii ) the need to describe NCIs in equilibrium and non-equilibrium systems on the same footing (i.e., similar performance across entire potential energy surfaces (PES)). 49,50 An essential part of developing next-generation theoretical methods for describing NCIs involves testing and/or training new approximations against highly accurate benchmark data. For the increasingly popular suite of ML-based models-which require large amounts of high-quality data to learn the quantum mechanics underlying NCIs-such reference data is of critical importance.…”