in polar materials the long range nature of Coulomb interaction makes
the dynamical matrix non analytical in the q->0 limit. This originates
the so called LO-TO splitting.
In cubic semiconductors this splitting is such that the LO frequency is
direction independent but in systems of low enough symmetry the
dielectric response and the effective charges tensors are not simply
multiples of the identity. Hence the LO (and TO) frequencies may depend
on the
direction of approach to the gamma point.
see for instance the phonon dispersion in wurtzite semiconductors
Bungaro and de Gironcoli, "Ab initio study of phonons in wurtzite
AlxGa1-xN alloys", Applied Physics Letters 76, 2101-2103 (2000),DOI:
10.1063/1.126268
stefano
On 27/04/2016 14:29, Atanu Paul wrote:
Hello all,
I am new to use PWscf code. I have been trying to calculate the phonon
dispersion of a 2d structure using Pwscf code. Dynamical matrices have
been calculated with 4x4x1 mesh. Then I have used 'q2r.x' and
'matdyn.x' to get the phonon dispersion curves. The structure of the
unit cell is orthorhombic and the high symmectric kpoint path is taken
as G-Y-S-X-G. Now If I look at the dispersion curve, the frequencies
at Gamma(G) point ( approaching from G-Y, i.e along y-axis in
reciprocal space) is different from the frequencies at Gamma(G) point
(approaching from X-G, i.e along x-axis in reciprocal space). How can
frequency be different at Gamma point? Though I have got positive
frequency at any k-points after execution of 'matdyn.x'. Am I doing
anything wrong?
Thanking you in advance.
atanu
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum
