Choosing the right functional

[Mohammed]

Hello,

I am running a DFT simualtion on anthracene molecules with a phosphonate anchor group. In order to get an agreement with experimental results I wanted to use the B3LYP but I understand its not included in the functional library in ATK. I used BLYP instead but I get a much smaller band gap (2.2eV sim compared to 3.5eV exp). I not familiar with all the functionals included but I am trying different sets to see if I can retain the correct band gap.

But I would greatful if you can guide me on how to get the correct value.

Thank You

[Petr Khomyakov]

You may try using Meta-GGA as described in the following tutorial http://docs.quantumwise.com/tutorials/inas_2d_mgga/inas_2d_mgga.html.

I would like to note that, for semiconductor and insulator materials, the Kohn-Sham energy band gap that is actually calculated in density functional theory is not the same as fundamental energy gap, in general.  So, Meta-GGA is one of a possible way to resolve this issue.

Using hybrid functionals would be another option, e.g., the HSE06 functional is available in the FHI-AIMS calculator,  http://docs.quantumwise.com/tutorials/fhiaims_hse/fhiaims_hse.html, but you should keep in mind that hybrid functional calculations are much more expensive computationally, compared to the LDA, PBE and Meta-GGA functionals.

[mlee]

I recommend to use the exchange correlation, Hybrid-B3LYP in the FHI-aims. And you should check your spin state. If you have unpaired spin on your molecules, you should set to the spin polarized, SHybrid-B3LYP. I hope you get a comparable result.

[Mohammed]

Thank you for your replies.

I will try with the Meta-GGA functional and see the output. I understand that the band gap would be underestimated compared to experimental values but a difference of more than 1 eV seems quite large to me.

On a different note I wont be able to use the FHI-Aims calculator as it is not included in our license.