Convergence wrt # q points and broadening parameter

[physliebe]

Hello all,

I'd like to ask about the convergence of the scattering rates wrt # q points and broadening parameter (\eta).

From Fig. S2(a) of Ponce et al., Phys. Rev. B 97, 121201(R) (2018), I understand that the scattering rates (where the mobility is plotted) will diverge as the \eta approaches zero (I suppose for each data point in that figure, the # q points is large enough). On contrary, according to Zhou et al., PRB 94, 201201(R) (2016), by decreasing the \eta and increasing the # q points, the scattering rates will eventually become converged, i.e., the \eta->0 limit exists.

I fee like these two arguments are very different. Could you please help me clarify? Thanks.

Hyuan

[sponce]

Dear Hyuan,

Thank you for your question.

The Fig.S2 in my paper is done at fixed fine k and q interpolated grids.
If instead you increase the k/q grids when you decrease the smearing, it should indeed converge.
I think that the \eta->0 limit exists for a grid density -> infinity.

Best wishes,
Samuel

[physliebe]

Dear Samuel,

Thank you for your explaination.

Best regards,
Hyuan