3
$\begingroup$

How to make Deuterium POTCAR file? i have Hydrogen POTCAR file.

$\endgroup$
7
  • 1
    $\begingroup$ For DFT studies, the mass doesn't matter. For hydrogen and deuterium, the electronic properties should be the same. If you are doing molecular dynamics, then you need to change the mass accordingly. But for geometry optimization, or anything else for that matter, using DFT, you can even set the mass to 0 since it has no effect. See Phil Hasnip's answer to this question. VASP is also a plane wave pseudopotential based code. So everything he said there applies to this question as well. $\endgroup$ Feb 5 at 9:13
  • $\begingroup$ i don't completely understood that answer but i understood that the POMASS tag doesn't affect energy calculations. To calculate D2 molecule energy, optimize its structure like H2. Differentiate POTCAR or POSCAR files for Deuterium (D) and Hydrogen (H) by specifying the tags corresponding in the POTCAR or using distinct POSCAR files for each molecule. I have observed that the dissociation energies of D2 and H2 are identical, contrary to information suggesting differences(almost 2 KJ/mol). here i am not doing MD calculation here. $\endgroup$ Feb 5 at 10:40
  • 2
    $\begingroup$ From which source are you referring? As explained in the linked answer, the nuclear mass doesn't enter any of the equations that these DFT codes solve. For H2 and D2, the DFT results should be identical as you have observed. Could it be the case that you are taking the dissociation energy values from other type of calculation/experimental results instead of a DFT calculation? In any case, please cite the paper from where you got the 2 KJ/mol value so that the methodology used there can be checked. $\endgroup$ Feb 5 at 11:13
  • $\begingroup$ pubs.acs.org/doi/10.1021/acs.jpcc.2c04567 here they calculate adsorption energy on Pd(111) surface in Table 1 which is different. If i allow your logic in this, then how we calculate this adsorption energy if we use same poscar and potcar file (not depend on mass) $\endgroup$ Feb 5 at 14:08
  • 1
    $\begingroup$ In the Table 1 of the paper you referred, they did not calculate those values using DFT. By running simple SCF or geometry optimization calculation, it is not possible to get isotope dependent result. I think you need some kind of phonon calculations to get the ZPE. $\endgroup$ Feb 5 at 15:34

1 Answer 1

4
$\begingroup$

Hydrogen and deuterium are basically identical electronically (*) and you should use the same pseudopotential (POTCAR) for each. You should not see any difference between hydrogen and deuterium (or tritium) in energies, forces, stresses, force-constants (Hessian) etc. nor, consequently, optimised Born-Oppenheimer geometries at 0 K.

There are some differences which affect certain types of calculation, principally:

  • Mass
    Deuterium has twice the mass of hydrogen, and this is important for its dynamics and quantum effects. The mass affects MD and phonon calculations, as well as zero-point motion and energies. Note: if you wish to obtain quantum-corrected geometries you will need to either use some form of path-integral MD, or phonon calculations beyond the usual harmonic approximation (ie you need at least 3rd order force constants).

  • Spin
    The nucleus can have intrinsic spin (and quadrupoles, octopoles...) which is important for some properties - most obviously NMR chemical shifts.

In general, different isotopes have different spin-orbit coupling as well, but this is negligible for such light nuclei as these.

(*) There is actually a very, very small difference due to the change in reduced mass, but we almost always use the Born-Oppenheimer approximation anyway which neglects this.

$\endgroup$

You must log in to answer this question.

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