[Phonopy-users] Negative phonon frequencies

[Artem K. <ar...@ma...>]

Dear Forum,


I am trying to calculate phonon dispersion for CuPbSbS3 (Pmn21, # 31 space
group). In all my attempts, I always get negative frequencies. Several
things were tried - different k-grids; different ISMEAR and SIGMA
parameters in VASP; different ENCUT; different potentials (semi-core
electrons were considered); inclusion of spin-orbit coupling. I also used
Atom Displacements and DFPT approaches.

Nevertheless, the phonon frequencies have the characteristic behavior of
the attached figure (negative modes behaving strangely in the range of -1
THz). My questions are the following:


-   CuPbSbS3 (Pmn21) has been synthesized by my collaborator, which means
the system is stable. However, the negative frequencies indicate the
material is unstable. Why am I getting negative phonon frequencies for a
material that exists?


-   How can I interpret these results providing the material exists?

-   The oxidation states for the atoms are Cu[+1]Pb[+2]Sb[+3]S[-2], Cu[+1]
and Pb[+2]. The common oxidation state for Cu is +2, while for Pb is +3. Is
it possible that unusual oxidation states are responsible for the negative
frequencies? If yes, how can this be avoided?

-   Can we avoid these negative frequencies by making other improvements?

I am attaching one figure showing the phonon dispersion for this material.
I would be happy to share more details about the input atomic positions and
other parameters if necessary.


Thank you for your help in advance and

regards,

Artem K.

-- 
-----------------------------------------------------------------
Artem R. Khabibullin,
Advanced Material and Device Theory Group,
Department of Physics,
University of South Florida, Tampa, FL 33620.