[ERROR]Fractional e/h occupations found. BS coupling forced.

[c3031929]

Hey all,

I've been trying to do some calculations of my own on metals in Yambo and I've come up with the error:

[ERROR]Fractional e/h occupations found. BS coupling forced.

All I've done is follow the procedure in the Yambo paper, the exact same procedure used to produce the Bulk Si results. The only difference between my procedure and the paper is that I simply accept all of Yambo's recommendations. I'm not exactly sure what has gone wrong, but I have included both the yambo.in and the log file from one of the runs (all the runs produce these files identically and have identical results). Any ideas?

Any help will be most appreciated.

Cheers,

Danno

[andrea marini]

As you can see from the l_ file Yambo has requested a finite temperature to converge the Fermi level. This means that your "silicon" is very wierd as no temperature should be required for a semiconductor.

What is the gap of your system ? Attach also the report file.

Moreover most of the input variables are not reasonable, number of BS bands, number of bands in the response and so on. Please run the tutorials first and/or use the exact input file we provided along with the CPC paper.

[c3031929]

My system is not the Si system from the paper, I've just been using the same general procedure to do BSE calculations on my systems. The system to which the attached files belong is Au.

I have now attached the requested report file. As for the bandgap of the system, I'm not quite sure at present how to determine that from abinit, but I will try and find that out soon.

[andrea marini]

My system is not the Si system from the paper, I've just been using the same general procedure to do BSE calculations on my systems. The system to which the attached files belong is Au.
(....)
I'm not quite sure at present how to determine that from abinit, but I will try and find that out soon.

Dear Danno,

why do you want to calculate excitonic properties on Gold ? Do you know the Gold, as a metal, has zero gap ?

Let me ask you: Do you expect a metal to exhibit bound and/or continous excitons ?

As a hint I point you to a paper a wrote several years ago.

[c3031929]

The thing I'm really trying to calculate is simply epsilon. The whole aim of my honours project is to calculate that for various metals using the different approaches Yambo provides. For the minute, I'm just trying to get the basics of how to produce the results I want for the materials I want. Quite honestly I am finding it all a little bit overwhelming. I have managed to reproduce the results from the Yambo paper for Bulk Si, and I am now just trying to obtain results for my own systems being Au, Ag and Al, and if I have time a few more.

Now that you mention the point about Gold having zero energy gap I should have realised that, but I didn't. Again, feeling a little bit overwhelmed.

As far as it goes for Au exhibiting bound and/or continuous excitons, I honestly haven't a clue. I shall have to look that stuff up.

Thanks for pointing me to your paper, I shall have a read of it and do my best to understand it if I can. I appreciate all the help I can get!

Cheers,

Danno

[andrea marini]

The thing I'm really trying to calculate is simply epsilon. The whole aim of my honours project is to calculate that for various metals using the different approaches Yambo provides. For the minute, I'm just trying to get the basics of how to produce the results I want for the materials I want. Quite honestly I am finding it all a little bit overwhelming. I have managed to reproduce the results from the Yambo paper for Bulk Si, and I am now just trying to obtain results for my own systems being Au, Ag and Al, and if I have time a few more.

Now that you mention the point about Gold having zero energy gap I should have realised that, but I didn't. Again, feeling a little bit overwhelmed.

As far as it goes for Au exhibiting bound and/or continuous excitons, I honestly haven't a clue. I shall have to look that stuff up.

Dear Danno, please take your time to understand the physics BEFORE running any code (Yambo, pw or whatever). Codes are like pencils, the you can use to write (calculate) something. But if you have no idea what theory is behind is like trying to write down a complicate theory before studying the basics. In this case the pencil won't' help you at all!

Plase take your time to go trough the lectures notes we provide. Ask your collegues for some basic solid state physics book.

I understand that it could be overwhelming... but it is a complete nonsense to run Yambo on a system you do not even know if it is a metal or a semiconductor (there are NO excitons in alluminum, the BSE there is useless!). And it is even worst to run the Bethe-Salpeter equation without even knowing if your system has or not excitons.