[QE-users] Slab calculation doesn't converge when Environ is added to the job

[Reinaldo Pis Diez]

Dear folks

I'm trying to simulate a large Iron oxide nanoparticle as a slab to 
calculate the energy associated to some atom replacements. 
Experimental data is obtained in aqueous solution, so, it seems 
reasonable to include solvent effects into my calculations.

I've started optimizing the bulk structure of the Iron oxide using 
DoJo PPs both for Fe and O. Then, I added some vacuum above and 
below, and some OH^{-} groups above because those structures are 
easily hydrated.

Slab jobs (relax only) converge well without solvent, both neutral 
(unreal) and with 6 negative charges. But when I add solvent effects 
through Environ (scf only), important oscillations happen for total 
energy, for electrostatic embedding and for the one-el term when 
environ contributions are added to the local potential and no SCF is 
achieved. I've tried both for the negatively charged system and for 
the neutral (unreal) one.

I've tested almost all the possible combinations in the environ.in 
file (attached is the last try) without success. I'm using QE 6.5 
and Environ 1.1, and as I said above, DoJo PPs from 
http://www.pseudo-dojo.org/. Attached are the PWScf and Environ 
input file and the (interrupted) output. I'd appreciate any hints to 
go forward with these calculations.

Regards,

Reinaldo Pis Diez
Center of Inorganic Chemistry
Department of Chemistry
Natl Univ of La Plata
Argentina
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ironOxide.tgz
Type: application/x-compressed-tar
Size: 9044 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20200729/8bae647b/attachment.bin>