DFT band gap underestimation

[sadegh]

Hello everyone,

The well-known band gap underestimation in LDA and GGA-PBE calculations is confirmed. I am trying to get accurate band gap for ASiNRs and AGNRs. As you may know, using GW or HSE03 can give a derivation of 20-30% compared to LDA and GGA. But there is not this opportunity to calculate band gap using these methods by ATK. I tried MGGA and got a more accurate band gap for the structures compared to LDA and GGA ( no significant changes). Is there a way to get an accurate band gap (in good agreement with experiments) using ATK?

Thanks in Advance 

[Jess Wellendorff]

Unfortunately ATK does not support exact exchange (therefore no HSE) or GW. However, MGGA calculations should give you significantly better band gaps than with LDA or GGA. The MGGA calculations need a c-parameter, which can be calculated self-consistently, or can be set such as to reproduce experimental band gaps for some well known case.

[sadegh]

Thanks Sir

 I employed MGGA for the calculation. I determined c parameter ( 0.917)  by calculating the band gap of bulk silicon ( about 1.13 eV). [ I have no idea why this value is different from tutorial (( http://www.quantumwise.com/documents/tutorials/latest/NanowireDevice/index.html/chap.nanowirebulk.html#sect2.nanowirebulk.job.cvalue )) where c is 1.0364!! and gives a band gap of 1.58 eV for Si bulk based on my calculations]

Then, I used both parameters to calculate ASiNR band gap with large vacuum regions. But, the results show that the bang gap does not change significantly:

--- ( Eg for 4ASiNR using MGGA = 1.004 eV) when c=0.917
--- ( Eg for 4ASiNR using MGGA = 1.02 eV) when c=1.0364

while ( Eg for 4ASiNR using GGA = 0.957 eV)

Enclosed to this message you can find the python files. Due to the file uploading restrictions, only python files are attached. Can you please help me with that?

[Jess Wellendorff]

Your scripts look ok, so it would appear that you are working with a structure for which the TB09 MGGA is not able to give significant corrections to the GGA band gap.