Page 1 of 1

phonon spectrum for rhombohedral crystal

Posted: Tue Jan 17, 2012 7:53 am
by dps
Dear all,
I am trying to simulate Sb2Te3 phonon spectrum as described in tutorial rf2. As the calculations are quite time consuming for converged ecut, I used rhombohedral primary cell:

Code: Select all

#Definition of the unit cell
#from previous cell optimization
  acell    1.6656138762E+01  1.6656138762E+01  1.6656138762E+01
  rprim   2.9828161671E-01  0.0000000000E+00  9.5447790814E-01
            -1.4914080836E-01  2.5831945755E-01  9.5447790814E-01
            -1.4914080836E-01 -2.5831945755E-01  9.5447790814E-01

#Definition of the atom types
  ntypat 2          # There are two types of atoms
  znucl 51 52     # Sb, Te

#Definition of the atoms
  natom 5           # There are 5 atoms
  typat 1 2 2 2 1  # 1 - Sb, 2 - Te             
  # from previous cell optimization
  xred
                       3.9999873689E-01  3.9999873689E-01  3.9999873689E-01
                       2.0966603731E-01  2.0966603731E-01  2.0966603731E-01
                       0.0000000000E+00  0.0000000000E+00  0.0000000000E+00
                      -2.0966603731E-01 -2.0966603731E-01 -2.0966603731E-01
                      -3.9999873689E-01 -3.9999873689E-01 -3.9999873689E-01


Then, how should I plot phonon spectrum in anaddb? In particular, anaddb help file said that brav=1 is for all lattices? Then, can I use it with my results for rhombohedral one? And what is the difference between brav=1 and brav=4, e.g. for hexagonal lattice, or brav=1 and brav=2 for FCC?
I am asking because I have plotted phonon spectrum for Si as an exercise. At first I calculate ddb using FCC cell and then plot phonon spectrum using brav=2 and brav=1 - they look similar but for brav=1 - curves look less smooth. But in general: is it correct at all?

Thank you,
dps

Re: phonon spectrum for rhombohedral crystal

Posted: Thu May 24, 2012 2:06 pm
by mverstra
brav > 1 is more accurate if you have the appropriate symmetry (fcc bcc or hcp) _AND_ if you have the correct rprim and alignment of the lattice. This is described in the anaddb help page for the brav input variable. For your Rhombohedral cell there is no equivalent brav coded in anaddb, so brav 1 is your best bet.

Matthieu