5
$\begingroup$

Recently I am trying to calculate the binding energy between CO2/ C2H2 and MOF (containing Ni, C, H, O, N). I want to know that do I have to use DFT+U for Ni element in this binding energy calculation? Since I read some articles that they only use DFT+U in electronic structure research.
Here is my MOF POSCAR:

C1.625 H0.875 N0.25 Ni0.125 O0.75
1.0
       16.6688003540         0.0000000000         0.0000000000
       -8.3344001770        14.4356045572         0.0000000000
        0.0000000000         0.0000000000         7.0012998581
   Ni    O    N    C    H
    3   18    6   39   21
Direct
     0.000000000         0.320639998         0.500000000
     0.679360032         0.679360032         0.500000000
     0.320639998         0.000000000         0.500000000
     0.116499998         0.447899997         0.500000000
     0.552100003         0.668600023         0.500000000
     0.331400007         0.883499980         0.500000000
     0.447899997         0.116499998         0.500000000
     0.668600023         0.552100003         0.500000000
     0.883499980         0.331400007         0.500000000
     0.924300015         0.174300000         0.500000000
     0.825699985         0.750000000         0.500000000
     0.250000000         0.075700000         0.500000000
     0.174300000         0.924300015         0.500000000
     0.750000000         0.825699985         0.500000000
     0.075700000         0.250000000         0.500000000
     0.083800003         0.560100019         0.500000000
     0.439899981         0.523699999         0.500000000
     0.476300001         0.916199982         0.500000000
     0.560100019         0.083800003         0.500000000
     0.523699999         0.439899981         0.500000000
     0.916199982         0.476300001         0.500000000
     0.000000000         0.320499986         0.801199973
     0.679499984         0.679499984         0.801199973
     0.320499986         0.000000000         0.801199973
     0.000000000         0.320499986         0.198800027
     0.679499984         0.679499984         0.198800027
     0.320499986         0.000000000         0.198800027
     0.000000000         0.084299996         0.500000000
     0.915700018         0.915700018         0.500000000
     0.084299996         0.000000000         0.500000000
     0.917299986         0.000000000         0.500000000
     0.000000000         0.917299986         0.500000000
     0.082699999         0.082699999         0.500000000
     0.000000000         0.173600003         0.500000000
     0.826399982         0.826399982         0.500000000
     0.173600003         0.000000000         0.500000000
     0.306499988         0.572799981         0.500000000
     0.427200019         0.733700037         0.500000000
     0.266299993         0.693500042         0.500000000
     0.572799981         0.306499988         0.500000000
     0.733700037         0.427200019         0.500000000
     0.693500042         0.266299993         0.500000000
     0.239700004         0.601499975         0.500000000
     0.398500025         0.638200045         0.500000000
     0.361799955         0.760299981         0.500000000
     0.601499975         0.239700004         0.500000000
     0.638200045         0.398500025         0.500000000
     0.760299981         0.361799955         0.500000000
     0.078699999         0.358399987         0.901499987
     0.641600013         0.720300019         0.901499987
     0.279699981         0.921299994         0.901499987
     0.078699999         0.358399987         0.098500013
     0.641600013         0.720300019         0.098500013
     0.279699981         0.921299994         0.098500013
     0.358399987         0.078699999         0.098500013
     0.720300019         0.641600013         0.098500013
     0.921299994         0.279699981         0.098500013
     0.358399987         0.078699999         0.901499987
     0.720300019         0.641600013         0.901499987
     0.921299994         0.279699981         0.901499987
     0.140300006         0.529699981         0.500000000
     0.470300019         0.610599995         0.500000000
     0.389399976         0.859699965         0.500000000
     0.529699981         0.140300006         0.500000000
     0.610599995         0.470300019         0.500000000
     0.859699965         0.389399976         0.500000000
     0.861523032         0.999998987         0.500000000
     0.000001000         0.861523986         0.500000000
     0.138475999         0.138476998         0.500000000
     0.287968993         0.510095000         0.500000000
     0.489905000         0.777873993         0.500000000
     0.222126007         0.712031007         0.500000000
     0.510095000         0.287968993         0.500000000
     0.777873993         0.489905000         0.500000000
     0.712031007         0.222126007         0.500000000
     0.134962007         0.385520995         0.836829007
     0.614479005         0.749441028         0.836829007
     0.250558972         0.865037978         0.836829007
     0.134962007         0.385520995         0.163170993
     0.614479005         0.749441028         0.163170993
     0.250558972         0.865037978         0.163170993
     0.385520995         0.134962007         0.163170993
     0.749441028         0.614479005         0.163170993
     0.865037978         0.250558972         0.163170993
     0.385520995         0.134962007         0.836829007
     0.749441028         0.614479005         0.836829007
     0.865037978         0.250558972         0.836829007

Hope someone could help me. Thanks in advance!

$\endgroup$
2
  • 2
    $\begingroup$ No, it is not necessary. It depends on what physics of the material you want to get correct, and if desired outcome isn't achieved, you can try whether DFT+U can solve the problem. I have seen people doing it to open a band gap/improve band gap/to account for the d- or f-orbital localization. Keep in mind that calculating U parameter is 20-30 times more expensive too (talking from Quantum ESPRESSO experience) and it also needs to be converged with respect to lattice parameters, energy cutoff, k-grid etc. So, it might be a good idea to check if you get your desired outcome with regular DFT $\endgroup$ Oct 9, 2023 at 8:10
  • $\begingroup$ @AbdulMuhaymin that looks like the start of a good answer! $\endgroup$ Oct 9, 2023 at 11:35

1 Answer 1

2
$\begingroup$

I think the number of $\ce{Ni}$ atoms is too few compared to the organic species, I am afraid the use of an inconsistent $U$ makes the problem more complicated besides the computational cost mentioned in the comments.

I would have recommended to use a LCAO basis-set, a small basis set on $\ce{Ni}$ will be sufficient to ensure a localization and a hybrid function will reduce the self-interaction error and give a more accurate energy gap. Since Vasp is used for this calculation, instead of using $U$, I will suggest a sufficiently large cut-off first and then add this correction if the result is really inconsistent, but $U$ can be avoided in this case.

$\endgroup$

You must log in to answer this question.

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