[Pw_forum] J.Breeze on:Third order anharmonic dynamical matrix

[Jonathan Breeze]

Dear Michele,

Thank you for your reply.  It was very useful.  I had used the
super-cell approach for the frozen phonon method for comparison with the 
D3 calculation.  So, hopefully it should be relatively easy to
apply your suggestion of zone folding.

Out of interest, I recently read a paper by G. Deinzer, G. Birner and D.
Strauch (Phys. Rev. B 67, 144304, 2003) which reports on calculations of
D^3(-q1-q2,q1,q2).  The form of the calculation was similar to that
reported by Debarnardi and Baroni. (Solid State Comm. 91, 813, 1994),
using the 2n+1 theorem.  The main differences are having to calculate
the electronic term for pairs of q values, a more involved Fourier
transformation of the anharmonic coefficients and a slightly more
complex Ewald transformation for the ionic part.

Anyway, thanks again for your advice.

Jonathan Breeze

Michele Lazzeri wrote:

 >>Using the PWSCF D3 code, should it be possible to calculate the
 >>coefficients for phonons at arbitraty wave-vector?
 >>That is, to calculate D^3(-q1-q2,q1,q2).
 >
 >>Jonathan Breeze
 >
 > Dear Jonathan,
 > As you know the code calculates D^3(0,-q, q).
 > More in general, to compute D^3(-q1-q2,q1,q2) is possible, but the
 > modification to be done to the code could be tedious.
 >
 > In alternative, you could use a super-cell approach:
 > If you want to compute D^3(Q,Q-q, Q+q), you can compute
 > D^3(0,-q,q) in a super-cell such that the gamma-point of the
 > super-cell is equivalent to the Q point of the unit-cell.
 > An example: if you want to compute the broadening of a phonon
 > at the border zone, you can use a crystal obtained doubling the unit-cell
 > in the appropriate direction.
 > The phonon you are interested in is a phonon at Gamma for this
 > double-cell.
 > So, if you compute the D3 matrix at gamma, for this double-cell, you
 > obtain
 > all the information that you need.
 >
 > For the moment, this is the most practical thing you can do if you are
 > interested in high-symmetry points.
 > This procedure cannot be applied to any point Q of the Brillouin Zone,
 > and, depending on the point, can be computationally expensive.
 >
 > Michele Lazzeri
 >
 >
 >
 >


-- 
Jonathan Breeze
Research Fellow
Centre for Physical Electronics and Materials
London South Bank University
103 Borough Road
London
SE1 0AA

Tel: +44(0)20 7815 7582
Fax: +44(0)20 7815 7599