Electrical and optical simulation of a Buried Tunnel Junction VCSEL

The last decade, and above all the last year affected by the pandemic, made us acknowledge how much our everyday life is dependent on internet. Nowadays communication but also entertainment, news, healthcare, shopping and so on, rely on web-conveyed data. This exponential increase of information traffic demands more and more computational resources, for processing and transportation of data. In such a cloud-based world, the already consolidated paradigm is that of centralized computation, which involves large-scale data centres where most of the data traffic occurs. It is understood, then, that the role of short range intra-datacenter interconnections is essential: bitrates up to Tbit/s and negligible energy consumption ultimately led to the employment of optical links and in particular Vertical Cavity Surface Emitting Lasers (VCSELs) for data transmission. VCSELs recently overcame the historical Edge-Emitting Lasers (EEL), being potentially faster and more energetically efficient, alongside fabrication advantages. Indeed, they have radically changed the fabrication and testing processes: being vertically-emitting, arrays of devices can be built onto a single wafer slice, enabling mass-production and testing. However, almost all commercialized VCSELs rely on a standard pin diode configuration and an oxide aperture in the proximity of the active region to provide confinement both for current (hence lowered threshold) and light (optical modes engineering). This approach presents two critical points: hole transport and wet oxidation process. For what concerns the first one, the issues are due to the p-doped DBR: the lower mobility of holes with respect to electrons leads to a higher resistivity, and thus higher losses and worse dynamic (tau = RC). Regarding the second, wet oxidation is the most difficult fabrication step: poorly controllable on the radial direction and potentially critical also along the growth one, due to compositional grading of the layers. A breakthrough is implementing a buried tunnel junction (BTJ) in such devices, as hole injector. The BTJ allows to effectively replace the entire p-DBR with a second n-DBR, reducing at once optical losses and resistivity, and providing a brand-new and built-in confinement method, thanks to a selective etch of its surroundings. However convenient the fabrication of VCSELs is with respect to EEL, the prototyping and testing stage of fabrication is still expensive and potentially material-wasting: computer-aided design (CAD) is of fundamental importance to have a first structure design to focus the work on, in the subsequent fabrication phase. The aim of the present Master's Thesis is integrating an already existing software developed at Politecnico di Torino (VENUS, or better, its 1D version D1ANA) based on drift-diffusion, with the possibility of including a model for the already cited BTJ, modeled with NEGF formalism. The connection between the two worlds is accounted by means of a local generation-recombination rate (GR), which gives the contribution of the interband tunnelling process. Thanks to these methods, a whole BTJ-VCSEL device can be simulated, and its results compared to those of a standard oxide-confined VCSEL, ultimately demonstrating the better performances of the BTJ version.