ABINIT Discussion Forums

Hi, can't quite find that anywhere, but is there a Grimme-like dispersion correction available with Abinit?

thanks

Jonas

Hi Jonas,
I totally understand where you're coming from, I've asked on these forums a couple times about things that were in places I wasn't expecting. The Grimme DFT-D2 pair wise dispersion correction is available. It's listed under the the features page (http://www.abinit.org/documentation/hel ... tures.html), and there is a test for it (under ~abinit/tests/vdwxc/t10.in). The two key parameters are vdw_xc =5 which activates DFT-D2 and vdw_tol, the cut off for when you stop counting the vdw interaction energies. This works with both PAW and NC psuedopotentials. There is also the method of Silvestrelli et al. using maximally localized wave functions (vdw-WF1 & 2) activated by vdw_xc =10 & 11, but I don't know anything about the Sivestrelli method beyond that.
Cheers,
James

much appreciated

Jonas