[Pw_forum] orthogonality of phonon eigenvectors?

[Sonu Kumar]

On Fri, Feb 10, 2012 at 6:37 PM, Paolo Giannozzi <giannozz at democritos.it>wrote:

>
> On Feb 10, 2012, at 24:32 , Suza W wrote:
>
> >    For example, after the above algebra, we get, these well-
> > arranged eigen displacements
> > [...]
> > ,whereas the initial eigen-vectors that QE code prints directly
> > (which is indeed messy)
>
> if you have degenerate eigenvalues, any linear combination of
> eigenvectors
> in the degenerate subspace is a solution. What you get from numerical
> diagonalization depends upon the phase of the moon.
>
>
Dear Prof. Paolo Giannozzi and ALL QE users,

If i use "dynmat.x", and use q=0,0,0 point (gamma), i get corrected eigen
values and
eigen vectors. It seems to me that ASR is applied to eigen vectors, because
they are different (more or less) from eigen vectors obtained from
conventional
"phonon.x" output. Is it so ?

However this difference could affect properties depending upon eigen
vectors or eigen
displacements. Am i right ?

One Basic question:
since eigen values and eigen vectors are related to eech other through
eigen values equation, Ax-\{Lamda}Ix=O. So, if x is changing(corrected)
after using
dynmax.x (imposing ASR), then why not eigen values are corrected ?
or is there any  linear combination of eigen vectors is used in the
procedure of ASR
so as to get  the same eigenvalue \Lamda.

with kind regards,
SK

==========================================
Sonu Kumar
Phd Student,Physics Department
Indian Institute of Technology ,Delhi-110016, India
web:-http://www.iitd.ac.in/
==========================================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20120211/2615c88e/attachment.html>