Band gap engineering via doping: A predictive approach

Abstract

We employ an extension of Harrison's theory at the tight binding level of approximation to develop a predictive approach for band gap engineering involving isovalent doping of wide band gap semiconductors. Our results indicate that reasonably accurate predictions can be achieved at qualitative as well as quantitative levels. The predictive results were checked against ab initio ones obtained at the level of DFT/SGGA + U approximation. The minor disagreements between predicted and ab initio results can be attributed to the electronic processes not incorporated in Harrison's theory. These include processes such as the conduction band anticrossing [Shan et al., Phys. Rev. Lett. 82, 1221 (1999); Walukiewicz et al., Phys. Rev. Lett. 85, 1552 (2000)] and valence band anticrossing [Alberi et al., Phys. Rev. B 77, 073202 (2008); Appl. Phys. Lett. 92, 162105 (2008); Appl. Phys. Lett. 91, 051909 (2007); Phys. Rev. B 75, 045203 (2007)], as well as the multiorbital rehybridization. Another cause of disagreement between the results of our predictive approach and the ab initio ones is shown to be the result of the shift of Fermi energy within the impurity band formed at the edge of the valence band maximum due to rehybridization. The validity of our approach is demonstrated with example applications for the systems GaN{sub 1−x}Sb{sub x}, GaP{sub 1−x}Sb{sub x}, AlSb{sub 1−x}P{sub x}, AlP{sub 1−x}Sb{sub x}, and InP{sub 1−x}Sb{sub x}.