[Pw_forum] Negative frequencies in phonon calculation

[Paolo Giannozzi]

On Thursday 06 October 2005 13:35, Alberto Milani wrote:

> I wonder if the code fails in some other way in stabilizing the
> linear chain (e.g. choice of the functional, pseudopotential, ...).

the choice of the functional in linear chains is for sure delicate.
Once upon a time I did some calculations in conjugate chains,
only to discover that there is no hope to obtain a decent
bond alternation with either LDA or GGA. Also forget about
decent interchain interactions. 

Anyway, this is unlikely to be relevant to your problem.
You may try two things:
- repeat the calculation with norm-conserving pseudopotentials,
just to check if there is something wrong in phonon calculations
with ultrasoft PPs
- make a frozen-phonon calculation for one of the modes that
displays a large negative frequency, if possible (i.e. if you do not
need a too large supercell)

Paolo

-- 
Paolo Giannozzi             e-mail:  giannozz at nest.sns.it
Scuola Normale Superiore    Phone:   +39/050-509876, Fax:-563513 
Piazza dei Cavalieri 7      I-56126 Pisa, Italy