Theoretical, Experimental, and Computational Study of Carriers Localization in Nitride Devices – TECCLON

Modern (opto)electronic devices rely on band engineering capabilities, made possible by the use of semiconductor heterostructures in which band gaps, offsets and strains in the active layers are optimized by associating pure compounds and alloys. High electron mobility transistors create two-dimensional electron gases at the interface between two semiconductor layers which possess a very high mobility, unsurpassed by any single material (homo)structure. Similarly, double heterostructures, through electron and optical confinement effects, give rise to efficient light emitting devices. For those alloy materials, however, intrinsic compositional disorder resulting from the random distribution of atoms is unavoidable. Although semiconductor alloys are ubiquitous in many modern semiconductor devices, the effects of intrinsic alloy disorder on device performances are still poorly understood, because of the huge computational power required for direct, “brute force” computations and of the complexity of the various resulting effects which must be considered for the description of disorder-induced phenomena. For most semiconductors, alloy disorder effects can be treated perturbatively and modify smoothly materials parameters because of the small amplitude of the potential fluctuations compared to the thermal energy at room temperature. For instance, electrical properties of AlInGaAs can still be described by drift-diffusion equations with a well-defined mobility. In nitrides alloy semiconductors this is no longer valid. The energy gap difference between pure materials is large enough to induce large potential fluctuations in alloys, strongly affecting the operation of nitrides based devices even at room temperature. For instance, strong localization effects are expected having a major impact on carrier transport. An added complication came from the fact that nitride semiconductors have large spontaneous and piezoelectric electric fields. Moreover, the numerical modeling of devices operations must be done in a self-consistent scheme. Using a “standard” approach resolving the Schrödinger equation for the quantum states and quantum statistical methods for carrier phenomena would require extremely long computational times, which has made so far such an approach unattainable for cases of practical interest. In our previous project CRIPRONI we proposed a new, effective approach for simulating real-world devices including the effects of disorder, based on a theoretical model developed by one of the members of the CRIPRONI consortium: the localization landscape theory. We demonstrated during that project the potentiality of our approach, obtaining for the first time, as an example, a realistic estimation of the turn on voltage of nitride-based blue LEDs. The computation time in our approach is typically 1,000 times shorter than a Schrödinger-based model, finally making the self-consistent simulation of nitride-based devices possible. In parallel, we developed a scanning tunneling luminescence (STL) experiment which made possible to observe the predictions of the modeling tool, namely single localized states in InGaN quantum wells at a few nanometer scale. The method allows to measure the local, intrinsic disorder separately from the long-range growth-induced compositional fluctuations. The objective of TECCLON is to further develop these modeling and experimental tools, in order to provide a better understanding of the limiting factors to the performances of nitride devices, and to get access to the parameters of the local electronic structure and to the carrier transport in alloys. This further improved modeling and the better knowledge of materials properties will help in solving still open questions that must be addressed in view of quasi-doubling the wallplug efficiency of visible green, yellow and amber LEDs for lighting, and of improving significantly the performance of UV LEDs for disinfection and water purification systems.