I am new to using the SQS method. I want to generate a structure for a doped perovksite with formula ABO3. I want to dope the B site with some transition metal and study the workfunction of the surface.
My question is: will it be possible to use as an input the structure of the slab I am interested in and let corrdump and mcsqs populate the sites based on my defined concentrations?
The workfunction will be highly dependent on the distribution of the dopants on the surface? How can I make sure the choice of SQS does not bias the results?
Yes, you can do this. The -2d option of mcsqs will generate supercells only along the surface (assuming the surface normal in along "c" or "z"). You can also force a given cell to be used, by providing your own sqscell.out file and specifying the -rc option on the command line. See mcsqs -h for more help on file formats.
You can study convergence by using SQS of increasing size (and better match of the random state correlations).
The trickiest part of this is if there is surface segregation of the dopants, the SQS formalism won’t account for that.