Thank you for your reply ! I tried this strategy fingers crossed, but it didn’t work, as could be expected. On a single unit cell short of 2 electrons, they mostly left from the 3 I atoms (+0.5 each compared to the “normal” case), and just +0.25 for the Sn atom. Furthermore, as I want to model it as a defect, I will eventually want to have a supercell (maybe 3x3x3), with only one of the Sn being in the unusual oxidation state. So I guess that would make that approach even less realistic, as there is no reason that the code should decide to arbitrarily deplete one of the atoms rather than the others.
I will try to have a closer look to the discussion link you sent in your other message and see if I can apply the idea to my case. Julien From: Marzari Nicola [mailto:nicola.marz...@epfl.ch] Sent: mercredi 24 avril 2019 12:38 To: Quantum Espresso users Forum; julien_barb...@sjtu.edu.cn Subject: Re: [QE-users] forcing Sn(iv) or Sn(0) charges in a MASnI3 crystal Sorry, different issue - you can change the number of electrons in the unit cell, and hope they go or leave from the right place. Sent from a tiny keyboard... Contact info: http://theossrv1.epfl.ch/Main/Contact On 24 Apr 2019, at 00:32, Marzari Nicola <nicola.marz...@epfl.ch <mailto:nicola.marz...@epfl.ch> > wrote: Here is a discussion http://theossrv1.epfl.ch/Main/OxidationStates Sent from a tiny keyboard... Contact info: http://theossrv1.epfl.ch/Main/Contact On 24 Apr 2019, at 00:28, JULIEN, CLAUDE, PIERRE BARBAUD <julien_barb...@sjtu.edu.cn <mailto:julien_barb...@sjtu.edu.cn> > wrote: Dear users, I am working on a MASnI3 crystal. In this structure, the Sn can usually be considered as a Sn2+ cation. I ran some calculations on the system, and performed a Bader charge analysis on an all-electron paw charge density. It seems to confirm that the tin is in Sn(2) configuration with bader charge 48.3 instead of 50 (this is, by the way, inconsistent with the results of Lowdin analysis as implemented in projwfc.x, which gives pretty much the full 4d10 5s2 5p2 orbitals of Sn(0) ). So everything is as expected so far (from the Bader point of view at least). However, I would like to model a MASnI3 with a “defect” consisting of an “oxidized or reduced” tin atom given by Sn(iv) or Sn(0) in this material. Indeed, it was reported in J. Mater. Chem. A, 2018,6, 21389-21395 that some degradation mechanisms can lead to the presence of such states, and I want to explore their consequence on the material properties. However, I am not sure how to tackle this. My first idea was that I probably needed to create a pseudopotential with 2 missing or additional valence electrons. But on second thought, this method might be valid if we have missing core electrons, but not for valence. I highly doubt that it would give me the expected result once I place it in the crystal lattice, given that there is really no reason for those additional electrons to gently “stay on their starting atom”, so to speak. So is there a reliable method to study such a system by “forcing” an oxidization state for an atom in a crystal ? This task seems to be made difficult by the very subjective definition of an “atomic charge” in the framework of quantum mechanics and DFT… Thanks in advance Julien _______________________________________________ Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso <http://www.max-centre.eu/quantum-espresso> ) users mailing list users@lists.quantum-espresso.org <mailto:users@lists.quantum-espresso.org> https://lists.quantum-espresso.org/mailman/listinfo/users _______________________________________________ Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso <http://www.max-centre.eu/quantum-espresso> ) users mailing list users@lists.quantum-espresso.org <mailto:users@lists.quantum-espresso.org> https://lists.quantum-espresso.org/mailman/listinfo/users
_______________________________________________ Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso) users mailing list users@lists.quantum-espresso.org https://lists.quantum-espresso.org/mailman/listinfo/users
