I'm a little confused how to make a VASP input file to calculate the energy for the isolated atom:

  1. Do I need to change the lattice parameter of my unit cell or can keep the value for the whole system?

  2. If I have 4 atoms of one type in my unit cell, can I consider the position of any of them or must it be in the middle of the unit cell?

  3. I observed that some people use K point set $1\times1\times1$ for calculating a single atom. Is it necessary to use this set or can I use the normal K points set that is used at the time of convergence?

To boil it down to one question, what do I need to change about typical VASP input file to calculate the energy of a single atom?


1 Answer 1


Isolated atom itself mean that there is no interaction from other nuclei , Hence box needs to be large enough so that atom doesn't interact with periodic image. In general 10-15 Angstrom size of box is enough. Once box is defined large (cubic) enough, placing atom anywhere inside box will not result any interaction from periodic image. If real space is large enough, reciprocal image will shrink to a point and number of kpoints will boil down to single point, Hence 1x1x1 is used for energy calculations. Better to look at OUTCAR to check reciprocal space lattice parameter and volume

  • $\begingroup$ Just curious: suppose I'm calculating the adsorption energy of an atom on a surface, where the surface is modelled by a periodic slab. Is it advantageous to calculate the energy of the atom using the same cell as the periodic slab calculation? This way the interaction of the atom with its periodic images is non-zero, but one may expect that it leads to better error cancellation, since the adsorbed atom also interacts with its periodic images, and computing the isolated atom in the same periodic cell tends to cancel this interaction as we subtract the energies. $\endgroup$
    – wzkchem5
    Commented Nov 28, 2021 at 15:45
  • $\begingroup$ This is the way I learned it, but from time to time I also encounter the practice that you are talking about, i.e. choose the cell such that it is large enough, but it does not need to be exactly as large as the cell used in periodic calculations. I'm curious whether the method that I learned is wrong, or is it also acceptable? If it is acceptable, is it inferior to the method that you describe here? $\endgroup$
    – wzkchem5
    Commented Nov 28, 2021 at 15:50
  • $\begingroup$ @wzkchem5 this may warrant its own question. It will make it easier to preserve the question/answer and it will be easier for Pranav to add an answer without the constraints of a comment. $\endgroup$
    – Tyberius
    Commented Nov 28, 2021 at 18:14
  • $\begingroup$ @wzkchem5 we can chat further in vasp chat room chat.stackexchange.com/rooms/109983/vasp $\endgroup$ Commented Nov 29, 2021 at 3:23
  • $\begingroup$ @pranavkumar Thanks. I've re-posted my question in the chat room $\endgroup$
    – wzkchem5
    Commented Nov 29, 2021 at 12:24

You must log in to answer this question.

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