We are planning on comparing the results of Quantum dots and Bulk materials for our systems so we are planning to use Gaussian for QD and VASP for Bulk. Considering that we will be comparing their results we should perform the same calculations(same corrections, same functionals and all) so should we hard code the constants used in Dispersion correction(DFT-D2)? So should we use the same constants such as $C_6$, vdW_radius, scaling factors, etc ??

If setting them the same is perhaps the correct path. Then we will be using the ones given by Gaussian as we can directly print them out using IOp(3/33=3).

Thanks and any help is greatly appreciated !!

  • $\begingroup$ I can't actually see a question here. I am voting to close on that basis, however, if you can edit the question to have a well defined question then I will retract it. I should say that you should have some caution in comparing VASP and Gaussian though since one is a planewave basis set and the other is an orbital basis set. $\endgroup$ Commented Jun 17, 2022 at 5:54
  • $\begingroup$ Hi @TristanMaxson I have tried to make my question a bit more clear. Basically I am trying to ask whether the coefficients used for bulk and QD calculations in Dispersion correction are same or not?? I do realize that the basis set vary a lot but we are trying to have as much similarity in calculations as much as possible. $\endgroup$ Commented Jun 17, 2022 at 6:25
  • 1
    $\begingroup$ If you want to compare, you need to run the simulations using the same code. In VASP (or any other similar periodic package) you can define a sufficiently big cell with you system at the center to be considered like an isolated molecule. $\endgroup$
    – Camps
    Commented Jun 17, 2022 at 18:35

1 Answer 1


The error due to the basis set will likely be larger than any difference in the DFT-D2 treatment. This is because you cannot use the same basis set in a (non-periodic) Gaussian calculation and a VASP calculation, as the former uses Gaussian functions while the latter uses plane waves. Thus, for maximum comparability of the results, you should use VASP also for the QD calculation, by specifying a periodic box that is much larger than the size of the QD so as to minimize the spurious interactions of the QD with its periodic images.

On the other hand, if comparability is not a concern, then using Gaussian to calculate the QD is a good idea, since less computational resources are wasted in describing the empty space surrounding the QD. In that case, however, you may use DFT-D3 instead of DFT-D2, since the former is also supported by Gaussian and offers higher accuracy. If you use ORCA, you can even go to DFT-D4.

  • $\begingroup$ I was actually leaning towards using VASP for both the calculations but my advisor has told me to use Gaussian. And as for using DFT-D3, that is not possible because the version of Gaussian that I have access to is old and only allows DFT-D2. Furthermore VASP does not have $C_6$ parameter for elements about Xe due to which I have to calculate/obtain the parameters so that is why I am asking whether I should use the ones given by gaussian or not ?? $\endgroup$ Commented Jun 17, 2022 at 11:49
  • 1
    $\begingroup$ Further, the Grimme group often provides stand-alone code/binaries, possibly also for periodic systems. $\endgroup$
    – TAR86
    Commented Jun 17, 2022 at 15:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .