Trouble or unbeliveable results obtained while calculating the gap function

[since2012]

Dear all,

When I calculating the gap function and the superconducting transition temperature, the main results as following:

Electron-phonon coupling strength = 2.6437681

Estimated Allen-Dynes Tc = 67.665860 K for muc = 0.10000

Estimated w_log in Allen-Dynes Tc = 34.392584 meV

Estimated BCS superconducting gap = 10.262547 meV

Estimated Tc from machine learning model = 99.710918 K

The results of lambda and the AD-Tc is very close to the one calculated by QE directly (lambda~2.7, Tc~70.9 K without AD revision). And the gap seems acceptable. But when i further solve anisotropic Eliashberg equations on imaginary-axis, the results is too bad. Taking the results of 20 K as example, the plotted gap distribution as following:


The convergence seems to be badly poor! What are the reasons for this ? Can someone give me some advices ? BTW, I have tested the degaussw, fshik, conv_thr_iaxis, what's else?

Any advices and comments are welcomed.

Thanks and Best Regards!

Jianguo Si
-------------------------------------------------
Songshan Lake Materials Laboratory
Building A1, University Innovation Park, Songshan Lake, Dongguan, Guangdong, CHINA

[hmori]

Hi Jianguo,

Unfortunately, I am unable to view the attached image, so I can only provide general comments.

To achieve converged results, it is necessary to test wscut and the sizes of the k and q grids (nkf1, nkf2, nkf3, nqf1, nqf2, and nqf3).
It is advisable to set wscut to at least 5 times the maximum phonon frequency. By default, wscut is configured to be 10 times the maximum phonon frequency.

Best regards,
Hitoshi

[since2012]

Hi Jianguo,

Unfortunately, I am unable to view the attached image, so I can only provide general comments.

To achieve converged results, it is necessary to test wscut and the sizes of the k and q grids (nkf1, nkf2, nkf3, nqf1, nqf2, and nqf3).
It is advisable to set wscut to at least 5 times the maximum phonon frequency. By default, wscut is configured to be 10 times the maximum phonon frequency.

Best regards,
Hitoshi

Dear Hitoshi,

Thanks for your kind reply.

According to some Chinese user's advice, I increased the degaussw and degaussq, and descrease the fsthik, the gap looks more convergence and acceptable.

Best and Regards!
Jianguo