I came across a paper which talks about $Cs_2SnI_6$ the material is well researched and there are dozens of experimental papers which say that the material exists in Fm3m space group(FCC), the particular paper I am talking about talks about the phase transition from cubic to monocilinic as the tempreature decreases although this theoretical calculation does not really go with experimental evidence. My point is that the they show the phonon band structure as enter image description here

where the red lines show unstable modes. So is this calculation wrong(as I have heard this would be a dynamically unstable mode, but the material does exist) or can materials with negative phonon modes also exist?

P.S there are studies which talk about the anharmonic lattice dynamics of Cs2SnI6 saying that the material shows quite a lot of anharmonicity.

  • 3
    $\begingroup$ Does this answer help? In particular, the part that says that the structure with the soft modes is likely a saddle point in the potential energy surface but might be a minima in the free energy surface above a certain critical temperature. $\endgroup$
    – CW Tan
    Nov 21, 2022 at 4:24
  • $\begingroup$ Yes thanks that actually did answer my question. $\endgroup$
    – Chan
    Nov 21, 2022 at 15:11
  • $\begingroup$ @CWTan can you write it up briefly as an answer? it is actually an interesting question, and others may bump into something similar. $\endgroup$
    – Greg
    Nov 22, 2022 at 7:09

1 Answer 1


My answer will be a more general one about soft modes and not too specifically-related to the system OP cited as these ideas are probably generally applicable to other structures.

Firstly, I would say that phonon calculations can be qualitatively "wrong" if the soft modes arise from convergence problems rather than anything physical. This could be a result of

  • Poor electronic convergence. The forces might not be well-converged.
  • Poor geometry relaxation convergence. The forces might not small enough, which would be problematic for say computing the force constants matrix with finite differences.
  • Poor supercell convergence. The phonon dispersion values at the $\mathbf{q}$-points commensurate with the supercell used may not exhibit soft modes, but Fourier interpolating those points for values in between those commensurate $\mathbf{q}$-points may lead to artificial soft modes that might disappear when larger supercells are used. (see this Q&A)

If the soft modes are still present, they can be physically interpreted as indicative of dynamical instability as OP pointed out. This Q&A provides a comprehensive discussion relevant to OP's question. In short, we can rationalize the experimental observation of a structure with a soft mode by thinking of the structure as a saddle point in the potential energy landscape that could become a minima in the free energy landscape above some critical temperature.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .