6
$\begingroup$

So, I'm looking for a benchmark of DFT functionals not including dispersion corrections. Something similar to Grimme's and coworkers GMTKN30 database, but for benchmark calculations done without dispersion corrections.

As in the case of GMTKN30 database, I'm interested in the same properties: general main group thermochemistry & kinetics. But

  • I already have calculations done many years ago without dispersion corrections since this feature simply was not implemented in most of the packages that time.
  • My workhorse is G3/G4 method which relies on geometry obtained without dispersion corrections anyway.
  • I quickly run few calculations including dispersion corrections and found no noticeable difference. Don't have time to redo all the work anyway.

I basically need to justify the choice of just a few DFT functional for my own benchmarks based on thorough literature reviews.

$\endgroup$
  • $\begingroup$ Maybe you could ask Prof. Grimme if he knows sth like that. $\endgroup$ – pH13 - Yet another Philipp Aug 14 '15 at 12:09
  • 2
    $\begingroup$ I would guess that Truhlar has some papers like this on DFT thermochemistry. Maybe in the M06 era? $\endgroup$ – Geoff Hutchison Aug 14 '15 at 12:21
  • $\begingroup$ @GeoffHutchison, good hint. The M06 paper + references there in seems to be the thing I'm looking for. $\endgroup$ – Wildcat Aug 14 '15 at 13:24
5
$\begingroup$

There is a benchmark of the GMTKN30 database with and without the D3 dispersion correction. See Figure 4 here: "Effect of the damping function in dispersion corrected density functional theory"

However, your statement

I quickly run few calculations including dispersion corrections and found no noticeable difference.

is not supported by this.

Another example is this paper: "Improving intermolecular interactions in DFTB3 using extended polarization from chemical-potential equalization" where we compare PBE to PBE-D3 for 11 different datasets (see the SI). For many dispersion-dominated complexes, DFT (no D3) is qualitatively wrong, and in many other cases, the error is comparable to the mean interaction energy.

In general, I would strongly recommend against DFT without an empirical dispersion correction unless you are only focusing on electronic properties, but then you wouldn't be very interested in GMTKN30 anyway.

Since you say you are using G3/G4 you might also be interested in these recent benchmark papers:

$\endgroup$
  • $\begingroup$ Many thanks for the references! I know that today it it generally not recommended to do DFT without a dispersion correction, but as I said I have a huge pile of calcs done few years ago at G4 level. :( And what I mean by saying that there is no noticeable differences is that overall G4 numbers agrees with, say, M062x-D3. So, in other words I do not compare DFT against DTF-D3, rather G4 against DTF-D3. But I have to investigate few non-covalently bonded complexes more closely... $\endgroup$ – Wildcat Aug 15 '15 at 8:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.