eptemp and smearing parameters in EPW

[andreyl]

Dear EPW community,

Smearing technique is used multiply times in EPW, so there
are number of parameters controlling it.
I understand it this way (please correct me if I am wrong):

(Let's assume we don't use delta_approx approximation)

eptemp controls fermi and bose occupation smearing and enters:
Phonon-self energy (f(e_nk)-f(e_mk+q))
Electron self-energy (n(w_qnu)+f(e_ml+q), n(w_qnu)+1-f(e_mk+q))

degaussw controls electronic states delta in lambda calculation
and also used as a smearing parameter in calculation of DOS at E_fermi.

degaussq controls phonon states delta in lambda calculation.

What I am confused about are the following parameters: nqsmear, nsmear, delta_qsmear, delta_smear.
I understand this parameters allow to vary the number and step How are they
connected with degaussw and degaussq and are they connected to them at all?

Next, which of this parameters enter a2F calculation.

Also, if I calculate the Tc, say using Allen-Dynes-McMillan equation for simplicity,
what is the appropriate way to treat eptemp? Should I set it to (expected)
critical temperature value (say 10 K) or treat it like a purely computational
parameter and converge the calculation with changing eptemp?

Best wishes,
Andrey

[liu xinbao]

I have the same confusion.
I found the test method in another post viewtopic.php?f=3&t=26
Should I test all the parameters such as nqsmear, nsmear, delta_qsmear, delta_smear in that way?
Really looking forward to any reply.

[roxana]

Hi,

degaussw = smearing value of electronic states
delta_smear = change in the smearing energy (e.g., step change in degaussw)
nsmear = number of different smearing values for degaussw

For instance if

degaussw = 0.025 eV
delta_smear = 0.05 eV
nsmear = 3

you perform a calculation for 3 smearing values (nsmear = 3). First with a smearing value of 0.025 (degaussw = 0.025), second with a smearing value of 0.075 (degaussw + delta_smear = 0.075), and third with a smearing value of 0.125 (degaussw+2*delta_smear = 0.125)

degaussq = smearing value of phononic states
delta_qsmear = change in the smearing energy (e.g., step change in degaussq)
nqsmear = number of different smearing values for degaussq

eptemp = temperature for the electronic Fermi occupations in the e-p calculation (units of Kelvin). This eptemp has nothing to do with the Tc in Allen-Dynes or McMillan equations.

Best,
Roxana

[liu xinbao]

Thank you very much.
Really help a lot !