Defect state passivation at III-V oxide interfaces for complementary metal–oxide–semiconductor devices

The paper describes the reasons for the greater difficulty in the passivation of interface defects of III–V semiconductors like GaAs. These include the more complex reconstructions of the starting surface which already possess defect configurations, the possibility of injecting As antisites into the substrate which give rise to gap states, and the need to avoid As-As bonds and As dangling bonds which give rise to gap states. The nature of likely defect configurations in terms of their electronic structure is described. The benefits of diffusion barriers and surface nitridation are discussed.