SCF not convergent in GEO_OPT

[Jonathan Campeggio]

Hi, 

 I am doing a geometry optimization of a Pd slab with a molecule of ethanol. I have controlled both the coordinates and the periodicity of the problem and they seem ok. ALso, a single SCF run converges. The problem is in the geometry optimization. Where I do errors? Thank you

[Nicholas Winner]

Please attach your output as so we can see the progression of the GEO_OPT.

[Jonathan Campeggio]

First, thank you for your rapid response and kindness. In attachment there is the ongoing output.

Best regards,

Jonathan

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/4b743979-4a1e-4a5c-9754-d438036f903bn%40googlegroups.com.

[Lucas Lodeiro]

There is something I do not understand. You said that you can converge a scf calculation... but the geo_opt is not converging. As I can see, in your geo_opt calculation you cannot converge not even the first scf calculation. Are you using different scf setup for the last calculation?  

If you can converge a scf calculation, I would use the same input for the geo_opt calculation, and read the wavefunction as the guess.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CADLcJcbna9KS8OUncNEK_po-67tU9x_6h_2Q7MeKngEYdjOXVw%40mail.gmail.com.

[Jonathan Campeggio]

To be more complete I will post in attachment the convergent SCF input. I have commented the MIXING section because I have wrongly thought that it would become faster. I am trying with your suggestions, thanks