DynMat calculation DISPLACECAR

[as2123]

Hello,

I would like to use the DynMat code to calculate harmonic frequencies by finite differences (I cannot use IBRION=5 because the calculation takes longer than the maximum wall time of my jobs).

I read the manual here: https://theory.cm.utexas.edu/vtsttools/dymmat.html
Then I tried a test calculation for a CO2 molecule, so I have a POSCAR file with 3 atoms, I set NSW=10 (3*3 +1) in the INCAR file, and I use a DISPLACECAR file with a 3x3 matrix filled with the number 0.02 (the displacement in Ångström).

However, when I try to run it, it only calculates displacements for the first atom in the POSCAR, so it ends after 4 ionic steps instead of 10 (one to calculate the forces of the initial geometry, and three for displacements of the C atom in each coordinate).

Maybe I am misunderstanding how to write the DISPLACECAR file? Please find attached INCAR, POSCAR, and DISPLACECAR, as well as the XDATCAR so you can see the displacements.

Thank you for your help.

[graeme]

Try changing the line:

EDIFF = -1E-20
to
EDIFFG = -1E-20

[as2123]

Hello Prof. Henkelman,

I had not noticed that typo in the INCAR. That was it, now it works fine. Thank you and sorry for the trouble.

[graeme]

Fantastic, happy to help!