Forces on certain atoms are zero in Materials Project

[Emily_Amonde]

When trying to extract forces from certain Materials Project tasks, I am getting atoms where forces are exactly zero, which seems strange. Are the VASP forces really zero in these cases or due to some other reason?

For example:

doc = mpr.tasks.search(task_ids = ['mp-1098692'], fields=["task_id", 'calcs_reversed'])

n=1
for i in range(0, len(doc[0].calcs_reversed[n].output.ionic_steps)):
    print(doc[0].calcs_reversed[n].output.ionic_steps[i].e_fr_energy, np.mean(np.absolute(doc[0].calcs_reversed[n].output.ionic_steps[i].forces)), np.min(np.absolute(doc[0].calcs_reversed[n].output.ionic_steps[i].forces)))

Yields this output:
-376.74321268 0.17382322041666667 0.0
-378.05058349 0.08167673489583334 0.0
-378.04921049 0.08319385666666666 0.0
-378.32745488 0.055105734999999996 0.0
-378.3424853 0.05601396833333334 0.0
-378.39719075 0.0175318371875 0.0
-378.39737717 0.020061451041666668 0.0
-378.40761167 0.01817606625 0.0
-378.40771002 0.0169736115625 0.0

And printing out the complete forces shows atoms where forces are (0,0,0).

[Joseph_Montoya]

Someone who actually knows something about VASP and DFT should answer this question - but my understanding is that primitive cell calculations with e.g. one atom or those with certain cubic symmetries will have zero forces by virtue of the symmetry alone, e.g. the imposed symmetry ensures that there cannot be a force on the constituent atoms, only a stress on the unit cell. See this VASP thread.

That said - your example is tetragonal but it looks like there’s some broken symmetry in at least one of the bonds so I’m surprised the forces are zero - it doesn’t look it would have that property unless VASP is reducing the symmetry internally (which it often does, but I would be surprised if it was in this case).