Self consistent field convergence not reaching

[Seba Darshan]

Dear Abinit users

I am very sad to say that for my last few posts in the forum, I didn't get any reply.
I am posting my problem now with a hope that someone will help me with some suggestions.

I am trying to optimize water molecule on aluminium doped graphene. Even after 182 iteration steps, the self consistent field is not converging.
I have used toldff 5.0d-5, nstep as 200.

Could someone tell me what could be the problem that leads to this non-convergence.

Regards
Seba

[jzwanzig]

Well, you probably have a difficult system. Did you do lots of control experiments? Did you make sure that your model of Al-doped graphene itself gives predictions that agree with expt or intuition? Did you make sure you can accurately model an H2O molecule on a simpler surface, say pure Al or pure graphene?

[Seba Darshan]

Well, you probably have a difficult system. Did you do lots of control experiments? Did you make sure that your model of Al-doped graphene itself gives predictions that agree with expt or intuition? Did you make sure you can accurately model an H2O molecule on a simpler surface, say pure Al or pure graphene?

Dear Sir

Thank you for the reply.
In the reported research articles, GGA of PBE approximation was used for modelling H2O molecule on Al-doped or pure graphene. That's why I tried using the same and with GGA, I am getting the warning message that "nstep =200 was not enough SCF cycles to converge".

I was able to model Aluminium doped graphene alone using GGA of PBE, the results are close to previously reported theoretical results. But I am having convergence issues on optimizing the geometry of H2O on Al-doped graphene. Can you please suggest what could be possible reasons for this SCF convergence issue.

Thanks in Advance
Seba