How should one add diffuse functions to the basis set? Diffuse functions are known to be critical in describing the electron distribution of anions (as discussed in my book), but they are also quite important in describing weak interactions, like hydrogen bonds, and can be critical in evaluating activation barriers and other properties.

The Truhlar group has been active in benchmarking the need of basis functions and their recent review¹ summarizes their work. In particular, they recommend that for DFT computations a minimally augmented basis set is appropriate for examining barrier heights and weakly bound systems. A minimally augmented basis set would have s and p diffuse functions on heavy atoms for the Pople split-valence basis sets and the Dunning cc-pVxZ basis sets.

For wavefunction based computations, they recommend the use of the “jun-“ basis sets. The “jun” basis set is one of the so-called calendar basis set derived from the aug-cc-pVxZ, which includes diffuse functions of all types. So, for C in the aug-cc-pVTZ basis set, there are diffuse s, p, d, and f functions. The “jun-“ basis set omits the diffuse f functions along with all diffuse functions on H.

The great advantage of these trimmed basis sets is that they are smaller than the fully augmented sets, leading to faster computations. And since trimming off some diffuse functions leads to little loss in accuracy, one should seriously consider using these types of basis sets. As Truhlar notes, these trimmed basis sets might allow one to use a partially augmented but larger zeta basis set at the same cost of the smaller zeta basis that is fully augmented.

References

(1) Papajak, E.; Zheng, J.; Xu, X.; Leverentz, H. R.; Truhlar, D. G., "Perspectives on Basis Sets Beautiful: Seasonal Plantings of Diffuse Basis Functions," J. Chem. Theory Comput., 2011,
7, 3027-3034, DOI: 10.1021/ct200106a

Bally and Rablen have followed up their important study of the appropriate basis sets and density functional needed to compute NMR chemical shifts¹ (see this post) with this great examination of procedures for computing proton-proton coupling constants.²

They performed a comparison of 165 experimental coupling constants from 66 small, rigid molecules with computed proton-proton coupling constants. They use a variety of basis sets and functionals. They also test whether all four components that lead to nuclear-nuclear spin coupling constants are need, or if just the Fermi contact term would suffice.

The computationally most efficient procedure, one that still provides excellent agreement with the experimental coupling constants is the following:

1. optimize the geometry at B3LYP/6-31G(d)
2. Calculate only the proton-proton Fermi contact terms at B3LYP/6-31G(d,p)u+1s[H]. The basis set used for computing the Fermi contact terms is unusual. The basis set for hydrogen (denoted as “u+1s[H]”) uncontracts the core functions and adds one more very compact 1s function.
3. Scale the Fermi contact terms by 0.9155 to obtain the proton-proton coupling constants.

This methodology provides coupling constants with a mean error of 0.51 Hz, and when applied to a probe set of 61 coupling constants in 37 different molecules (including a few that require a number of conformers and thus a Boltzmann-weighted averaging of the coupling constants) the mean error is only 0.56 Hz.

Bally and Rablen supply a set of scripts to automate the computation of the coupling constants according to this prescription; these scripts are available in the supporting materials and also on the Cheshire web site. It should also be noted that the procedure described above can be performed with Gaussian-09; no other software is needed. Thus, these computations are amenable to synthetic chemists with a basic understanding of quantum chemistry.

References

(1) Jain, R.; Bally, T.; Rablen, P. R., "Calculating Accurate Proton Chemical Shifts of Organic Molecules with Density Functional Methods and Modest Basis Sets," J. Org. Chem., 2009, 74, 4017-4023, DOI: 10.1021/jo900482q.

(2) Bally, T.; Rablen, P. R., "Quantum-Chemical Simulation of ¹H NMR Spectra. 2. Comparison of DFT-Based Procedures for Computing Proton-Proton Coupling Constants in Organic Molecules," J. Org. Chem., 2011, 76, 4818-4830, DOI: 10.1021/jo200513q

For those of you interested in learning about dispersion corrections for density functional theory, I recommend Grimme’s latest review article.¹ He discusses four different approaches to dealing with dispersion: (a) vdW-DF methods whereby a non-local dispersion term is included explicitly in the functional, (b) parameterized functional which account for some dispersion (like the M06-2x functional), (c) semiclassical corrections, labeled typically as DFT-D, which add an atom-pair term that typically has an r^-6 form, and (d) one-electron corrections.

The heart of the review is the comparison of the effect of including dispersion on thermochemistry. Grimme nicely points out that reaction energies and activation barriers typically are predicted with errors of 6-8 kcal mol^-1 with conventional DFT, and these errors are reduced by up to 1.5 kcal mol^-1 with the inclusion fo the “-D3” correction. Even double hybrid methods, whose mean errors are much smaller (about 3 kcal mol^-1), can be improved by over 0.5 kcal mol^-1 with the inclusion of the “-D3” correction. The same is also true for conformational energies.

Since the added expense of including the “-D3” correction is small, there is really no good reason for not including it routinely in all types of computations.

(As an aside, the article cited here is available for free through the end of this year. This new journal WIREs Computational Molecular Science has many review articles that will be of interest to readers of this blog.)

References

(1) Grimme, S., "Density functional theory with London dispersion corrections," WIREs Comput. Mol. Sci., 2011, 1, 211-228, DOI: 10.1002/wcms.30