[QE-users] Charged Convergence Issues

[Nicola Marzari]

Thanks Giuseppe!

Indeed, with Oliviero we studied this class of problems (unbound anions, 
and their absorption on metal slabs). The discussion is here:
https://pubs.acs.org/doi/abs/10.1021/acs.jctc.9b00552 (also on arxiv).

In a nutshell, the "correct" bound state is there, but in vacuum the 
unbound state is lower in energy, so that's where DFT-PBE converges to.

In the presence of continuum solvation, the "correct" bound state 
becomes the lowest in energy, so that's where the calculation converges 
to. Since these energies are linear in the dielectric constant of the 
medium, you can extrapolate to the vacuum case.

I say "correct" because the anion would still have self-interaction 
errors - they seem to be smaller than I would have thought, meaning the 
the damaging effect of self-interaction is stabilizing the wrong state, 
but the correct "excited" state has good energies. Same reason why 
penalty functionals or constrained-DFT work well for charge-transfer 
excitation - they do not fix the errors of approximate DFT, but allow 
you to select your "ground" state that approximate DFT makes an excited 
state. Piece of cake...

				nicola



On 19/05/2020 09:45, Giuseppe Mattioli wrote:
> 
> Dear Robert
> The calculation does not converge due to delocalization error (see, 
> e.g., Cohen et al., Science 2008, 321, 792), which affect local (LDA) 
> and semilocal (GGA) functionals. The excess electron of COO- is unbound 
> in your DFT description of the system and hinders scf convergence. On 
> the other hand, COO- is not stable in gas phase, while it is stable in 
> water solution. There are a few possible things to do:
> 
> 1) use a hybrid GGA+EXX functional that partially corrects the 
> delocalization error. It helps but COO- (gas) may still contain an 
> unbound electron.
> 2) embed your system in an implicit solvent with the QE plugin (see 
> quantumenviron.org for details)
> 3) use 2) together with the explicit inclusion of a first solvation 
> sphere (e.g., a few explicit water molecules) which lower the energy of 
> the unbound electron and stabilize COO-, favoring convergence.
> HTH
> Giuseppe
> 
> Quoting Robert Stanton <stantor at clarkson.edu>:
> 
>> Hello,
>>
>>     I'm having an issue in calculating adsorption energies in a system 
>> with
>> charged molecules. I have relaxed the structure with and without the
>> adsorbent. In the case of structure+adsorbent, I introduce a compensating
>> jellium background charge for the COO- molecule, and this converges fine.
>> The structure itself is neutral and converges fine in a neutral cell.
>> However I need to also optimize the lone COO-, and I am running into the
>> issue that this will not converge in the case of a charged cell, but 
>> only a
>> neutral one. The issue then is that if I use the COO- energy converged in
>> the neutral cell, I get adsorption energies that are not accurate.
>>
>>     I am mainly just wondering how the calculation is converging in the
>> latter two cases above, since from what I've read it seems one of them
>> should have a charged cell causing convergence issues? Could cutoffs
>> potentially be causing an issue here? I have tried adding spin
>> polarization, dropping the mixing beta extremely low, and tried other 
>> small
>> molecules with a formal charge and they all have a similar issue. Just
>> looking for any ideas as to how I could get this charged molecule to
>> converge alone in a charged cell.
>>
>> Thank you in advance,
>> Robert Stanton
>> Clarkson University
> 
> 
> 
> GIUSEPPE MATTIOLI
> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
> Via Salaria Km 29,300 - C.P. 10
> I-00015 - Monterotondo Scalo (RM)
> Mob (*preferred*) +39 373 7305625
> Tel + 39 06 90672342 - Fax +39 06 90672316
> E-mail: <giuseppe.mattioli at ism.cnr.it>
> 
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users


-- 
----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
Director, National Centre for Competence in Research NCCR MARVEL, EPFL
http://theossrv1.epfl.ch/Main/Contact http://nccr-marvel.ch/en/project