QP gap of Si

[burkzdemir]

Dear Yambo developers,

I'm trying to calculate dielectric function of Si with GW+BSE, however I find that dielectric function is redshifted by about 0.4 eV with respect to experimental result. In the literature GW+BSE dielectric function is almost exactly on top of experimental result. I performed convergence tests with respest to number of k-ponts, total number of bands, response block size, band indices. Also, I find that LDA band gap is 0.41 eV and PBE gap
0.61 eV and with GW the gap is 0.76 eV, however in the literature the GW gap is slightly higher than the experimental value of 1.12 eV. In summary, there is underestimation of GW gap and GW+BSE dielectric function is redshifted. I cannot find the problem. I am sharing my yambo_GW.in and yambo_BSE.in.

Regards,

[Daniele Varsano]

Dear Burak,
please check your lattice parameter as the gap is very sensible to it.
Next, from your input I can see you are considering very few bands in the calculation of the polarisation and Gw summation, but most importantly very few G vectors in the block of the dielectric matrix.

You can have a look to this paper as reference, where both LDA and GW values are reported and also some convergence parameters:
Brad D Malone and Marvin L Cohen 2013 J. Phys.: Condens. Matter 25 105503
Table3 and Fig.3

D. Wing et al. PHYSICAL REVIEW MATERIALS 3, 064603 (2019)
here you can find again GW calculations and BSE.

Best,
Daniele

[burkzdemir]

Dear Daniele,

Thank you for your answer! I increased the number of bands and the problem is solved.

Regards,
Burak