Dear all,
I have a brief question regarding how to manually change the C6 parameter for DFT-D2 calculations in VASP. I have been conducting some tests and I would like to "switch off" Van der Waals interactions for a particular species, to see how important said species' contribution is to the overall VdW interaction. I had understood that I could do this by using the flag VDW_C6 = [real array] in the INCAR file. I presumed that this means that if I have, say two species, A and B, I can enter VDW_C6 = m n , where m and n are values for C6 for A and B respectively, or alternatively VDW_C6 = p*m q*n, where p is the number of A atoms and q is the number of B atoms in the cell (like what one might do when using the flag MAGMOM). However, I tried both of these approaches, and in each case the calculation ran as normal, using the default C6 values and essentially ignoring this instruction. Does anybody know the correct way to do this? Thanks in advance.
First of all: it should be one (real) value *per atomic species*. And I'm wondering why it does not work because it should definitely overwrite the defaults according to what I've seen in the source code (vdwforcefield.F) ...
Dear Jürgen,
thanks for the response, and for clarifying the input format - some other colleagues suggested the same, so I've been doing some experimentation with the various parameters. I did manage to perform some calculations excluding VdW for a specified species, by including not only the desired C6 values, but also the radii using VDW_R0 (I just set these values to the default values used for a "full" DFT-D2 calculation), although it seems strange that I must specify values that I have no intention of changing. But if as you say the source code permits it, I agree it is a little strange. For these calculations I have been using VASP 5.3.3, if this is relevant.
Thanks again for your advice, it is much appreciated. Any further insights would be most welcome.
Best regards, Michael.
How do you tell VASP to include van der Waals corrections? is it with IVDW=1 (or LVDW=.TRUE.)? I would say that then you have to write the values of the C6 and R0 parameters for each species defined in the POSCAR file, that is, a pair of values {C6,R0} for all the atoms with the same label in the POSCAR. If you want to switch off vdW interactions for a specific atom and there are more of that kind in your system, then you will have to separate them also in the POSCAR and POTCAR files so that they have different labels.
Dear Hugo,
It depends on the VASP version you are using. From what I understand, you can use LVDW=.TRUE. in v. 5.2.11 and later; this automatically applies a DFT-D2 dispersion correction. From version 5.3.4, however, this tag is obsolete, being replaced by the IVDW tag you mention. Setting IVDW=1 gives you D2, =11 gives you D3, =12 gives you D3 with Becke-Jonson damping, and =20 gives you Tkatchenko-Scheffler dispersions. As implemented in VASP, there are default values for C6 (and C8 for D3 corrections) and R0, so VASP uses these if you don't specify values for each species in the POSCAR file as you suggest. I suspect you are correct to suggest that if one desired to change the parameters for different atoms of the same species, then it would be necessary to indicate the relevant species in the POSCAR and POTCAR. I haven't done any tests on this yet, but it would be interesting to try.
Thanks for your response, feel free to add any further comments.
Best regards, Michael.
Hi Michael,
I am facing the same problem that you have crossed so far..
Wondering to hear your experience in solving this problem !!
- Badges
- Science topic
- Similar topics
- Filing