Antisites in silicon carbide

Ten years ago, deep-level-transient-spectroscopy (DLTS) signals, assigned to centers labeled as $H1\mathrm{}$, $H2\mathrm{}$, $H3\mathrm{}$, and $E2\mathrm{}$, have been detected in neutron-irradiated 3C SiC. The $H$ centers were believed to be the primary point defects and the $E2\mathrm{}$ center a secondary defect, which forms after the $H$ centers start to migrate. A conclusive identification of these signals has not been presented so far. We present computational evidence that the $H$ centers are due to silicon antisite defects $\left({\mathrm{Si}}_{\mathrm{C}}\right)$. In both cubic (3C) and hexagonal (2H) polytypes, the silicon antisite has several ionization levels in the band gap. The positions of these ionization levels in 3C SiC have been calculated accurately with the plane wave pseudopotential method using a large 128-atom site supercell, and compared with the DLTS spectrum. A very good agreement with experimental data indicates that $H$ centers are due to the formation of ${\mathrm{Si}}_{\mathrm{C}}$ during neutron irradiation. The formation energies and local geometries of the antisite defects in SiC are also reported.