What's the point of using extremely denser k-grid for extracting curvature effective masses? I saw in some publications that they use extremely dense k-grid for band structure calculations for reliable effective masses. However, in some publications they just extract effective masses from a band structure that determined with relatively coarse k-grid. I could not find so many information about effective mass convergence. Is it something really reliable? Thanks for your comments
The effective mass is related to the second derivative of the E-k curve as
So an accurate value of effective mass requires an accurate dispersion curve at all k points. If enough k values are not taken to plot the curve, plotting tools will interpolate the values to have a smooth curve, and the values may vary significantly from the actual values. This may lead to an inaccurate value of effective mass in those regions of brillouin zone.