Cu(₁₋y)AgyInS₂(₁₋ᵪ)Se₂ᵪ as a Prototype of the Pentenary Chalcopyrite Semiconductor Systems

Abstract

<p>The group Ill-V alloy quarternary semiconductors, such as Ga(₁₋y)InyAs(₁₋ᵪ)Pᵪ have been extensively employed in lattice matching different semiconducting layers (at specific bandgaps) to form heterojunction electro-optical devices. However, these cover only a limited set of direct bandgap/lattice constant combinations. The analogous pentenary alloys, consisting of the ternary chalcopyrite groups l-lll-VI₂ and II-IV-V₂, have the potential of similar applications as they cover a even wider band/lattice range. As, a prototype of such alloys, samples the pentenary Cu(₁₋y)AgyInS₂(₁₋ᵪ)Se₂ᵪ have been synthesize and studied.</p> <p>Samples were prepared by reacting stoichimetric powder mixtures at about 900ºC. X-ray diffractometry tests suggest the compounds maintained complete powder solid solubility throughout the system in the chalcopyrite crystal structure. The intrinsic conductivity type of the alloys appear to follow a trend towards n-type for silver and sulfur rich compounds, while forming p-type for copper and selenium rich materials.</p> <p>The bandgap of these samples were measured using cathodoluminescence techniques, which generally have some ambiguity in their resulting estimates. To generate better values of the band parameters extensive computer modeling for emission spectra from heavily doped direct bandgap materials was done. The effect of band tails and Gaussian impurity states on the luminescence spectra was studied for changes in doping densities, temperature and carrier injection levels. Formulae were derived from these models to obtain better estimates of the bandgap and impurity activation levels. Algorithms were developed to obtain the impurity spreading energy of a tailed or Gaussian band, and the quasi-Fermi energy levels for injected current in a material with a specific band structure.</p> <p>Cathodoluminescence measurements were made at 300 and 77 K on the samples. As predicted by the models, it was found easier to generate good estimates from the 300 K results due to the obscuring effects of the impurity bands at cooler temperatures. Using these band estimates least squares fits were made on the data to generate the topological bandgap/lattice constant versus compositional value maps for the studied alloy.</p>