different phonon structure between QE and EPW

[sevenel]

Dear all,
I calculate the phonon structure using QE and EPW, and I found that they have a little difference at the Gamma position.
How to deal with it, and if it will have a larger impact when I calculate the mobility?
Another question if I get the negative mobility, is it trustworthy？

[sevenel]

this is my input file

Code: Select all

&inputepw
 prefix='GeSe'
 outdir='./'
 dvscf_dir='../ph/save'
 
 elph=.true.
 epbwrite = .true.
 epbread = .false.
 epwwrite=.true.
 epwread=.false.
 lpolar = .true.
 vme = 'wannier'
 use_ws = .true.
 
 wannierize=.true.
 nbndsub=32
 bands_skipped='exclude_bands = 1-20'
 num_iter = 200
 dis_froz_max = 3.0
 dis_froz_min = -16.0
 proj(1)='Ge:s,p'
 proj(2)='Se:s,p'

 efermi_read = .true.  
 fermi_energy= -1.5392
 fsthick = 100
 nstemp = 1
 temps = 0.075
 degaussw = 0.001
 
 band_plot = .true.
 filkf='KPOINTS'
 filqf='KPOINTS'
 nk1 = 6 
 nk2 = 6
 nk3 = 1
 nq1 = 3
 nq2 = 3
 nq3 = 1
/

[sponce]

Dear sevenel,

The phonon frequency of polar materials is ill defined exactly at q=Gamma.
However it is well defined for q-->0 along a particular direction.

Therefore you can safely remove the q=0 point from your plot and it should look exactly like the one obtained from matdyn.x

You should not get negative mobility (at least the diagonal of the mobility tensor should not be negative).
Off diagonal component (if very small) can be negative.
It also depend if you do drift or Hall mobility. In the case of Hall mobility, it is possible to have negative values depending how you orient your B-field.

Best,
Samuel

[sevenel]

Dear Samuel
Thank you very much for your reply
Best wishes
zhenqing