2, The Ohio State University, Columbus, Ohio, United States
The family of heterovalent ternary nitride semiconductors offers an intriguing opportunity to expand the range of materials properties and potential for innovative device designs beyond those available solely with the binary nitrides and their homovalent alloys. The heterostructures formed by III-nitride and II-IV-nitride can potentially address key issues that are currently faced by the III-nitride community. One example is the close values predicted for the spontaneous polarization coefficients for the Zn-IV-nitrides (IV=Si,Ge,Sn), which leads to much smaller built-in polarizations for heterostructures grown along the c-axis. Other possibilities arise with the large band offsets predicted for GaN-ZnGeN2 heterointerfaces, which allow for the design of novel LED structures that are calculated to improve electron-hole wavefunction overlap by more than a factor of 2 compared to conventional InGaN-GaN device structures. The large band offsets, coupled with the close lattice match and similar optimal growth temperatures for GaN and ZnGeN2, are also advantageous for the design of quantum cascade laser structures for emission in the near-infrared intersubband transition wavelength range. Furthermore, several of the heterovalent ternary nitrides are composed entirely of abundant, nontoxic, inexpensive elements. One of these, ZnSnN2, has garnered considerable interest as a potential solar photovoltaic material. The increased complexity of the heterovalent ternary nitride lattice, compared to the binary nitride lattice, allows for octet-rule-preserving polytypes and random stacking of two orthorhombic phases, as well as for cation exchange defects. This situation is especially pertinent to ZnSnN2, as the two orthorhombic phases are predicted to have very close energies of formation and band gaps. Octet-rule-preserving polytypes are also possible for mixtures of the binary and heterovalent ternary nitrides. While the proposed LED and quantum cascade structures noted above do not require p-doping of the ZnGeN2 layer, p-doping will be necessary for may applications. Ab initio calculations of defect energies and concentrations indicate that p-doping by substitutional elements or by adjustments in cation stoichiometry may be problematic for the heterovalent ternary nitrides, and successful doping strategies to achieve high p-doping may have to rely on the formation of defect complexes.
This work was supported by the NSF DMREF SusChEM award 1533957 and by the NSF SusChEM award 1409346.