Difference between Van der Waals (DFT-D and DFT-D3) corrections in ab-initio calculations

Question

In Kohn-Sham DFT calculations, Van der Waals corrections are often implemented in the structure optimization calculations because the typical functionals such as LDA and GGA are found to not treat long-range dispersion forces faithfully. But even in the context of Van der Waals corrections, there are some options such as the DFT-D and DFT-D3 schemes (refer quantum espresso's documentation - under 'vdw-corr' tag).

What is the difference between these two corrections? Also, is it possible to conclude which one is more suited for a certain application?

Answer 1

As you mention, there are many empirical dispersion corrections for density functional theory.

Generally, the term "DFT-D" refers to a generic dispersion-corrected density functional calculation, regardless of the specific method used for the dispersion correction used.

The D3 dispersion model is a specific dispersion correction method and is now something of a family, started by Grimme's 2010 article: "A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu" J. Chem. Phys. 132, 154104 (2010)

In most cases, it's recommended to use Becke-Johnson damping - D3BJ: J. Comput. Chem., 32: 1456-1465 (2011)

There have been a few efforts to improve these:

• D3M - J. Phys. Chem. Lett. 2016, 7, 12, 2197–2203
• D3(op) - J. Chem. Theory Comput. 2017, 13, 5, 2043–2052
• D4 - J. Chem. Phys. 150, 154122 (2019)

There are a wide variety of other methods, particularly those working to add many-body dispersion. Grimme wrote a review, although obviously it's been an active field:

"Dispersion-Corrected Mean-Field Electronic Structure Methods" Chem. Rev. 2016, 116, 9, 5105–5154