• extended defects;
  • ionic semiconductors;
  • nitrides;
  • optoelectronic devices;
  • potential fluctuations;
  • surface states


In this paper I am addressing the fundamental question as to why the performance of optoelectronic devices based on nitride semiconductors is insensitive to high concentration of extended defects, which is not the case for traditional III–V compounds. There are three important differences between these two families of semiconductors, which contribute to this finding: (a) the chemical bonds in nitrides are strongly ionic while are mostly covalent in III–V compounds. This leads to the bunching of the surface states as well as the states associated with dangling bonds in edge dislocations near the band edges, which prevents them from being non-radiative recombination centers. Furthermore, the surface states have less effect on the surface Fermi level position; (b) the nitrides can exist in the wurtzite structure (equilibrium) and the cubic structure (metastable) and the enthalpy of formation of the two allotropic forms differs only by a few meV. Thus, conversions between the two phases occur easily by creation of stacking faults along the closed packed (0001) and (111) planes. As a result basal plane stacking faults are abundant in both compounds and alloys. This leads to strong band structure potential fluctuations; (c) additional band structure potential fluctuations exist in InGaN and AlGaN alloys, due to phase separation and alloy ordering. Both types of potential fluctuations contribute to exciton localization and thus efficient radiative recombination even at room temperature.