Ferromagnetic insulating behaviour has recently been demonstrated in cobalt-substituted SrTiO 3 at room temperature [1]. Experimentally, it was found that a well-defined hysteresis loop only occurs at high Co concentrations of 30-40%. X-ray photoelectron spectroscopy also indicated that Co substitutes for Ti with Co being in high-spin +2 oxidation state. In this work, we employ density-functional theory to explain the experimentally observed properties of cobalt-substituted SrTiO 3 . We examine in detail the role of oxygen vacancy (OV) defects and their formation of defect complexes with the Co ions as the origin of the ferromagnetic insulating behaviour. Our first-principles thermodynamic calculations indicate that OV defects are much more likely to occur next to Co atoms where their formation energies could be reduced by as much as 1.28 eV compared to that in bulk SrTiO 3 . We also find that Co in these Co-OV complexes occurs in the high-spin state in agreement with core level spectroscopy, and identify a linear arrangement of the Co-OV defect complexes to be the most energetically favourable structure. These defect complexes are also shown to interact ferromagnetically and that their magnetic interaction is found to be short-ranged, consistent with the relatively high Co concentrations needed experimentally for ferromagnetism to be observed in cobalt-substituted SrTiO 3 .
I.Introduction The discovery of room-temperature ferromagnetism in doped transition metal oxides by Matsumoto et al. [2] has initiated many theoretical and experimental studies, owing to potential applications of these materials in spintronics. For example, , SnO 2 [5-6], In 2 O 3 [7][8] and TiO 2 [9][10], have all been reported to show room-temperature ferromagnetism when doped with magnetic ions. Increasing the Curie temperature and being able to control the spin degrees of freedom remain major challenges in this field. In this respect, defects in these materials, both intrinsic and extrinsic, play an important role. Defects provide additional carriers to the system (electrons or holes) which could significantly alter magnetic interactions and the Curie temperature in these materials.