Phonon Calc at Gamma

Message

Tiger-paw · #1 Post by **Tiger-paw** » Sun Dec 04, 2005 9:27 pm

1.
To calc Phonon Freq in a crystal supercell at Gamma, I use IBRION=50, NFREE = 2, but how to choose a reasonable POTIM? Will the calculated phonon freq be sensitive to POTIM?

2.
In my calc, all freq I obtained for Gamma are nonzero. Am I supposed to see 3 zero freq and 3N-3 nonzero normal mode freq? What can I do for this?

Many thanks for your help,

tiger-paw

#2 Post by **admin** » Mon Dec 05, 2005 9:59 am

1) The Hessian Matrix is constructed via displacements from the equilibrium positons as set by POTIM, therefore the valuehas to be small enough to ensure that you are in the harmonic regime (a few hundereths of Ã…).
btw: IBRION =5 (not 50)
2) please choose a reasonable input in POSCAR (selective dynamics, T/F): if all positions are free to be displaced, the number of degrees of freedom is 3N, including the translation of the whole species as well.

dfcox · #3 Post by **dfcox** » Sat Feb 11, 2006 7:40 pm

Admin-

Could you be a little more specific about point 1) when IBRION=5? What value of POTIM is appropriate to give a displacement of a few hundreths of an Angstom? Does POTIM represent the magnitude of the displacement in Angstroms?

Thanks

#4 Post by **admin** » Mon Feb 13, 2006 12:46 pm

1) please read the manual (http://cms.mpi.univie.ac.at/vasp/vasp/node101.html) concerning the
valid settings of IBRION.
2) POTIM is the displacement in Ã… you are right

Tiger-paw · #5 Post by **Tiger-paw** » Fri Sep 15, 2006 4:10 pm

Admin-

Could the direct/supercell phonon code be used to obtain all normal mode freqs for a surface slab, with vacuum and even adsorbent on both sides, and hence total free energy? (assuming all atoms are free to be displaced)

Is Gamma-point-only enough (or exact) for this C_1 sysmetry surface slab?

Many thanks!

Tiger-paw · #6 Post by **Tiger-paw** » Fri Sep 15, 2006 4:15 pm

Sorry, "adsorbent" should be "adsorbates", i.e. adsorbed atoms or moleculers.

bandy · #7 Post by **bandy** » Sat Sep 16, 2006 6:51 am

why you want to study adsorbates on both sides(surfaces) of slab? Is there any specific reason? because the calculation will be very expensive?

Tiger-paw · #8 Post by **Tiger-paw** » Thu Sep 21, 2006 3:48 pm

surface dipoles on two sides can be cancelled.

#9 Post by **admin** » Tue Sep 26, 2006 11:35 am

to correct for the dipol moments, you can also set IDIPOL=3 if it helps to spare computing time
in principle, you can leave all atoms free to relax and calculate all vibration modes, however, this will be very time-consuming