Negative Scattering Rates

[ankit.gujarat]

Hi,
I am currently calculating the shortrange scattering rate for alpha-Ga2O3, but scattering rates at some energy values are coming out to be negative. I donot understand what could be the possible reason for this result.
Here is the input:
--
&inputepw
prefix = 'aGa2O3'
amass(1) = 69.723
amass(2) = 15.999
outdir = './scratch/'

iverbosity = 1

elph = .true.
epbwrite = .true.
epbread = .false.

etf_mem = 0

epwwrite = .true.
epwread = .false.

nbndsub = 1
nbndskip = 44

wannierize = .true.
num_iter = 500
iprint = 4
proj(1) = 'f=0.6435000,0.6435000,0.6435000:s'
dis_froz_min= 12.5287
dis_win_min = 12.5287
wdata(1) = 'num_bands = 45'

efermi_read = .true.
fermi_energy= 12.1866

asr_typ = 'crystal'

lpolar = .true.
shortrange = .true.

elecselfen = .true.
nest_fn = .false.
phonselfen = .false.
a2f = .false.

scattering = .true.
scattering_serta = .true.

fsthick = 10 ! eV
eptemp = 300 ! K (same as PRB 76, 165108)
degaussw = 0.01 ! eV

dvscf_dir = './save/'

filkf = './kpoints.dat'

nqf1 = 20
nqf2 = 20
nqf3 = 20

nk1 = 6
nk2 = 6
nk3 = 6

nq1 = 6
nq2 = 6
nq3 = 6
/
32 cartesian
0.000000000000000E+00 0.000000000000000E+00 0.000000000000000E+00
-0.177208793294881E+00 -0.102311544511585E+00 0.661588117531448E-01
-0.354417586589762E+00 -0.204623089023169E+00 0.132317623506290E+00
0.531626379884643E+00 0.306934633534754E+00 -0.198476435259434E+00
-0.177208793294881E+00 0.102311544511585E+00 0.132317623506298E+00
-0.354417586589762E+00 0.000000000000000E+00 0.198476435259443E+00
0.531626379884643E+00 0.511557722557924E+00 -0.132317623506281E+00
0.354417586589762E+00 0.409246178046339E+00 -0.661588117531359E-01
0.177208793294881E+00 0.306934633534754E+00 0.888178419700125E-14
-0.354417586589762E+00 0.204623089023169E+00 0.264635247012597E+00
0.531626379884643E+00 0.716180811581093E+00 -0.661588117531270E-01
0.354417586589762E+00 0.613869267069508E+00 0.177080572427712E-13
0.531626379884643E+00 -0.306934633534754E+00 -0.396952870518895E+00
0.000000000000000E+00 0.000000000000000E+00 0.198476435259443E+00
-0.177208793294881E+00 -0.102311544511585E+00 0.264635247012588E+00
0.708835173179524E+00 0.409246178046339E+00 -0.661588117531359E-01
0.531626379884643E+00 0.306934633534754E+00 0.885402862138562E-14
0.354417586589762E+00 0.204623089023169E+00 0.661588117531536E-01
-0.177208793294881E+00 0.102311544511585E+00 0.330794058765742E+00
0.708835173179524E+00 0.613869267069508E+00 0.177358128183869E-13
0.531626379884643E+00 0.511557722557924E+00 0.661588117531625E-01
0.354417586589762E+00 0.409246178046339E+00 0.132317623506307E+00
0.708835173179524E+00 -0.409246178046339E+00 -0.330794058765750E+00
0.531626379884643E+00 -0.511557722557924E+00 -0.264635247012606E+00
0.354417586589762E+00 -0.613869267069508E+00 -0.198476435259461E+00
0.531626379884643E+00 -0.306934633534754E+00 -0.198476435259452E+00
0.000000000000000E+00 0.000000000000000E+00 0.396952870518886E+00
0.886043966474405E+00 0.511557722557924E+00 0.661588117531625E-01
0.708835173179524E+00 0.409246178046339E+00 0.132317623506307E+00
0.886043966474405E+00 -0.511557722557924E+00 -0.264635247012606E+00
0.708835173179524E+00 -0.613869267069508E+00 -0.198476435259461E+00
0.000000000000000E+00 0.000000000000000E+00 -0.595429305778329E+00

The total number of valence bands are 44 and I am just interested in the 1st conduction band for the scattering rate calculation hence nbndskip equals 44.
Any help in this regard will be appreciated.

Thank You,
Ankit Sharma
University at Buffalo

[ankit.gujarat]

The fermi energy is calculated separately using the DOS calculation from Quantum espresso and for a doping density of 1e17 cm-3(assuming complete ionization). The top of the valence band in around 10 eV and the bottom of the conduction band is around 12.5 eV obtained under the LDA.
For any more information, please let me know.

Thank You,
Ankit Sharma

[ankit.gujarat]

Hi,
Update:
In the transport.f90 module, the matrix element are read as only the real part of the square of the matrix element in case of shortrange and not Gamma point. I am not able to figure out why this is the case, because in my case some of the shortrange matrix elements are pure imaginary and hence the square is a negative number. I think this might be the reason for the negative shortrange scattering rates.
Any clarification in this regard will be greatly appreciated. Below is the code snippet present in the transport.f90 and selfen_elec.f90 files.

IF ( shortrange .AND. ( abs(xqf (1, iq))> eps8 .OR. abs(xqf (2, iq))> eps8 &
.OR. abs(xqf (3, iq))> eps8 )) THEN
! SP: The abs has to be removed. Indeed the epf17 can be a pure imaginary
! number, in which case its square will be a negative number.
g2 = REAL( (epf17 (jbnd, ibnd, imode, ik)**two)*inv_wq*g2_tmp )
ELSE
g2 = (ABS(epf17 (jbnd, ibnd, imode, ik))**two)*inv_wq*g2_tmp
ENDIF

Thank You,
Ankit

[wangshaofei18]

hello.
I have some questions to consult you. How to read the scattering_rate file?
the third list is E (eV). does it is the electron energy with reference to the Fermi energy E=E(electron)-E(Fermi)? or it just the absolute electron energy.
I got a very strange E (eV).
I did not set the fermi energy in the epw.in file.
hoping to receive your letters.
Thanks a lot!