convergence with respect to "wscut"
Posted: Thu Sep 19, 2019 3:06 am
Dear EPW developer,
When we solve the isotropic Eliashberg equation, we recently found that at a given temperature, the gap value strongly depends on the cutoff of imaginary Matsubara frequency.
For a particular example, at a given temperature (T = 0.5 K), the gap is equal to 0.14 meV when "wscut = 1 eV" (default value). As we increase "wscut", the gap value increases. When we use a ridiculous "wscut = 100 eV", the gap value is 0.45 eV and still is not converged.
Is this physically reasonable? We would think that a "wscut" should be large enough so that the gap value does not depend on "wscut" (i.e. converged).
Your insight is greatly appreciated.
Thank you very much.
Hanghui Chen
Department of Physics
New York University Shanghai and New York University
When we solve the isotropic Eliashberg equation, we recently found that at a given temperature, the gap value strongly depends on the cutoff of imaginary Matsubara frequency.
For a particular example, at a given temperature (T = 0.5 K), the gap is equal to 0.14 meV when "wscut = 1 eV" (default value). As we increase "wscut", the gap value increases. When we use a ridiculous "wscut = 100 eV", the gap value is 0.45 eV and still is not converged.
Is this physically reasonable? We would think that a "wscut" should be large enough so that the gap value does not depend on "wscut" (i.e. converged).
Your insight is greatly appreciated.
Thank you very much.
Hanghui Chen
Department of Physics
New York University Shanghai and New York University