Carrier injection modelling in buried tunnel junction VCSELs

After conquering the data-communication and sensing markets, vertical-cavity surface-emitting lasers (VCSELs) are now playing a key role in imaging applications in smartphones and automotive contexts. In fact, among their "competitors", these lasers exhibit unique features in terms of power consumption, reliability, testability, packaging costs and array-oriented manufacturability, making them the ideal light source for portable applications. In order to export these advantages to the widest range of upcoming applications, the VCSEL community is exploring novel concepts and materials, aiming to move the emission from the consolidated 850nm wavelength towards mid-infrared and/or blue ranges. A criticality of VCSEL design is the injection of holes in the active region, which is controlled, in classical AlGaAs VCSELs, by an aperture obtained through wet oxidation processes. A pitfall of the to-be-explored material systems is the technological unavailability of such processes, combined with poor hole transport processes, which demand for alternative strategies. A very interesting approach concerns the employment of a buried tunnel junctions (BTJ). Tunnel junctions consist of heavily doped pn junctions where, under reverse bias operation, the conduction and valence bands are overlooking, and carrier transport can take place through band-to-band tunneling (BTBT). In addition to replacing the oxide aperture, BTJs allow to eliminate p-doped contacts, with huge benefits to the VCSEL electrical conductivity and giving room to better thermal control and enhanced modulation speed overcoming the state of the art. Despite the great interest in this subject, the frantic times-to-market discourage investments such a revolution of the VCSEL architecture, which would require extensive and expensive prototyping campaigns. In this view, computer-aided design (CAD) tools play a starring role as alternative prototyping frameworks for research and development departments. The purpose of this Master's thesis is to include the BTBT quantum corrections within the drift-diffusion carrier transport picture implemented in the multiphysics Vcsel Electro-opto-thermal NUmerical Simulator VENUS, developed by CNR-IEIIT and Politecnico di Torino. Two BTJ-VCSELs are investigated as case studies: an InGaAsP long-wavelength device, and an AlGaAs short-wavelength device manufactured and characterized at Chalmers University of Technology. In the first part of the thesis Hurkx's work has been reviewed and implemented, which is based on local generation-recombination (GR) rates accounting for the junction electric field. Even though this model is quite consolidated for silicon electronics and implemented in commercial simulators such as Sentaurus Device by Synopsys, it doesn't appear to be suitable to these material systems. Therefore, the second part of the thesis explores a non-local GR model derived from a 4-band non-equilibrium Green's functions (NEGF) formalism. It is demonstrated that the ballistic nature of BTBT allows to apply the NEGF solver in the coherent limit, lowering its staggering computational cost. This observation is at the basis of the self-consistent NEGF-VENUS simulation framework, enabling to couple a semiclassical picture of the bulky sections of the device with a genuine quantum description of BTBT.