Page 1 of 1

Best strategy for charged defect?

Posted: Tue Dec 18, 2018 3:10 pm
by jbarrat
Dear all,

My work involves the implantation of helium ions in oxides, such as Ga2O3 in the present case. I would like to study a supercell (40 atoms), in which the central Ga atom is replaced by a helium ion. Actually, ideally it should be a helium ion (charge +1e) instead of a gallium ion (charge +3e). This means that I expect the supercell to have a net -2 charge, which should be localized about the helium ion.

I have already converged the supercell without the implanted ion, but I am unsure of what would be a suitable strategy for this little project. My first question would be: can I use the gga pseudopotentials for He and Ga atoms? If I remove a Ga atom entirely and add a He atom instead, the system remains at first uncharged. Would it make sense to then set charge -2, and hope that the charge gets located at the He atom? In this post about charged vacancies (, Boris mentions that the charge keyword can be used in order to charge the supercell, and that usually this charge ends up being located near or on the defect, which would be perfect. Is that true and, if yes, is there a preferred method to follow in order to achieve this?

Or are those ideas not going at all in the right direction, and instead a pseudopotential has to be generated for Ga ions and He ions somehow? (I have no experience with creating pseudopotentials sadly)

Thank you very much in advance for your suggestions. :)