warning: The distance between some ions is very small

[maria_dresser]

Hello,

I took a well converged structure from materials project and directly removed the Li atoms and replaced them with H atoms. However, when trying to do a standard relaxation calculation, I keep getting the warning: The distance between some ions is very small please check the nearest neigbor list in the OUTCAR file. Visualizing the POSCAR in vesta, the structure looks just fine to me. Even with this warning I still get a vasp.out file which says the structure is converged at the end, but the energy of the structure is much larger than I think it should be. Thank you for any help you can provide for how I can get rid of this warning and get a more reasonable energy. I have attached a zip of the relevant files.

Best,
Maria

[martin.schlipf]

I think you duplicated the Fe atoms. There are 32 Fe atoms and the second 16 have the same coordinates as the first 16. When you looked at the structure, a bunch of O atoms should be missing?

One other comment, take care whether the calculations are really converged. In the output you provided the calculation took 50 electronic steps, but you also limited the calculation to 50 steps, so the energy was not really converged.

[maria_dresser]

I see, thank you! This has fixed the issue.

[martin.schlipf]

For future reference, VASP reports the nearest neighbors of all atoms. Specifically in your case

Code: Select all

 ion  position               nearest neighbor table
[...]
   9  0.121  0.129  0.625-  25 0.00  42 1.91  49 1.91  45 1.94  41 1.94  46 1.97  18 2.91  19 2.91
                            10 2.91  12 2.91  21 2.95  22 2.95

The 9th atom and the 25th atom have a distance of 0.