“…We note that, for a given concentration of carbon dopants (number of carbon dopants per supercell), there are several ways to divide the carbon at nitrogen and boron sites. Since the experimental samples are n-doped, we only consider configurations with scriptN normalB ≥ scriptN normalN namely, C B (1.5% doping) (and C N for reference), 2 C B (3%), C B C N (3%), 2 C B 1 C N (4.7%), 2 C B 2 C N (6.25%), 3 C B 1 C N (6.25%), 3 C B 2 C N (7.8%), and 4 C B 1 C N (7.8%). Also, under N-rich conditions, the formation energies of C B and C N substituents were shown to be comparable, with a slight preference for the formation of the former defect than for the latter one .…”