ABINIT Discussion Forums

Dear colleagus,

I would like to find out the eigendisplacements in the cubic ZrO2, for example at some point between Gamma and X. Motivation is the instability at X point. Since the phonon eigenvalues turn out wrong for me (2 TA modes are unstable at X instead of TO (but I don't know if it's degenerate)), I would like to know are these phonon modes I get really acoustic or am I just reading it wrong.

But when I get the results from anaddb, I get for example these at Gamma:

Phonon frequencies in cm-1 :
- 0.000000E+00 0.000000E+00 0.000000E+00 2.767236E+02 2.767236E+02
- 2.767236E+02 5.956055E+02 5.956056E+02 5.956056E+02

Mode number 1 Energy 0.000000E+00
Attention : low frequency mode.
(Could be unstable or acoustic mode)
; 1 1.17976621E-05 -2.08707954E-05 2.10982444E-03
; 0.00000000E+00 0.00000000E+00 0.00000000E+00
; 2 1.17978818E-05 -2.08712849E-05 2.10982562E-03
; 0.00000000E+00 0.00000000E+00 0.00000000E+00
; 3 1.17977383E-05 -2.08712329E-05 2.10982538E-03
; 0.00000000E+00 0.00000000E+00 0.00000000E+00

I'm worried about the displacements in x and y directions. Shouldn't they be 0 exactly? Is it because of incompletely converged variables like ecut and ngkpt (I used ecut 50 Ha and ngkpt 3*12)?

Similarly, for q-point 0.01875 0.01875 0.03750 close to Gamma in the Gamma-X direction, I get:

Phonon frequencies in cm-1 :
- -9.842997E+01 -8.312513E+01 7.530105E+01 2.643965E+02 2.870527E+02
- 5.938648E+02 5.953870E+02 5.989505E+02 6.829342E+02

Mode number 1 Energy -4.484800E-04
Attention : low frequency mode.
(Could be unstable or acoustic mode)
; 1 5.65833328E-04 1.75619733E-03 7.70260089E-04
; 7.74029616E-05 -3.30273232E-04 4.42049210E-04
; 2 8.40020422E-04 1.51462318E-03 6.99001493E-04
; 5.36760655E-04 -8.07336369E-04 8.80414463E-04
; 3 7.34390945E-04 1.42349495E-03 4.87154580E-04
; 9.50544684E-04 9.66364543E-05 8.96842780E-04


Could anyone help me and explain why this happens? How can I identify phonon modes with these results? Should I ignore eigenvectors with lengths lower the 10^(-3) and consider them as 0?
I'm attaching my rf and anaddb input file.

Thank you all in advance!

Yours,
Igor Lukacevic

Dear Igor,

As discussed at the CECAM tutorial in Lausanne with you, a few weeks ago, this was indeed a bug...
ABINIT versions before 2006 worked just fine for ZrO2 , but a bug had been introduced then.
This bug has been fixed in ABINITv6.0.4, available today. I hope you will be able to correctly
reproduce the phonon band structure of ZrO2 that was published a long time ago.
Thanks for your careful testing, and reporting.

Best regards,
Xavier

Dear Xavier,

Don't mention it. Besides, it's a kind of obligation of all of us to contribute, this way or another. It was my pleasure to work with You and the others in Lausanne. It was a great benefit to me.
I'm sure it won't be a problem anymore, since the bug is fixed. My attention now is slowly shifted toward other phases and I'll be on it as soon as I get back to Osijek.
Thank You again for Your effort.

Cheers!

Igor