CCL: H-bond and dispersion energy

From: "Stefan Grimme" <grimme,thch.uni-bonn.de>
Subject: CCL: H-bond and dispersion energy
Date: Mon, 22 Jun 2015 01:59:48 -0400

 Sent to CCL by: "Stefan  Grimme" [grimme:+:thch.uni-bonn.de]
 >Stefan,
 >
 >What are your thoughts on D3 vs D3BJ?
 >Sam Abrash
 Dear Sam,
 D3BJ(BJ=Becke-Johnson damping scheme) is our default since 4 years. It
 has the physically correct short-range behavior (i.e., the dispersion energy
 goes to a constant for small inter-atomic distances) and yields better
 structures than D3(0). The interaction energies with both versions are
 very similar. Note that the (non-empirical) dispersion coefficients
 are the same in both versions. D3(0) should only be used with
 M06... type functionals which already include the medium range part
 and for which double-counting effects with B3(BJ) would occur.
 Best wishes
 Stefan