Imaginary rotational and translational modes

Dear ASE developers,

I’m performing vibrational analysis using the Vibrational class and noticed that some modes corresponding to the rotation and translation of my molecules are imaginary, and some of them with unexpectedly high frequencies like 20i or 50i. I tried making the geometry optimization and single-point calculation criteria stricter, but the issue persists. Do you have any idea what might be causing this and how I could resolve it? The molecules I’m working with are relatively large and floppy, containing around 20 to 40 atoms.

Best regards,

Since the energy is constant with respect to displacements and rotations, you should generally expect 6 eigenvalues that do not correspond to proper modes.

See also the tutorial Vibrational modes — ASE documentation

In general it is wise to check small well-known systems like H2O before doing big and floppy production calculations.