LASER DIODES AND THEIR RELIABILITY. PHYSICAL MODELS AND EXPERIMENTAL VALIDATION.

This PhD thesis focuses on laser diodes, and reports on three years spent within the Optoelectronic Reliability team of prof. Massimo Vanzi and prof. Giovanna Mura, at the Department of Electric and Electronic Engineering (DIEE) of the University of Cagliari. The relevance of laser diodes in several technological fields increases progressively with the evolution of their technology, and then of their performances. A summary of such history will be given in detail at the beginning of Ch.1. Telecommunications in particular have been both the most intense prompt and the hardest challenge to solid state semiconductor light emitters. Faster and faster, and more and more powerful devices have been progressively developed, with the conflicting constraints of single-mode operation combined with tenability, and the continuous request for integrated optical functions. The team at DIEE is historically involved, since decades, in studying the survival of laser diodes, that is, their Reliability. Reliability is in itself a manifold activity, that involves statistics, failure physics, modeling, microscopic analysis, strategy. This work witnesses three years of intense evolution of a new physical model for describing laser diodes, that was initially developed for giving account of the measured degradation of electric and optic parameters of failed devices. The model became progressively more and more inclusive, and also caused several deep revisions of the widespread approaches to laser diode physics and reliability. At the beginning of this period, the model was existing, and had been yet applied to several studies of specific failure cases. Anyway, it left some open questions, that limited its application to the totality of real cases. Along the three years of this PhD course, I had the chance to cooperate in first refining the fundamental model, and then in solving one by one all the unresolved issues, and in signing as an Author all the relevant papers published by the DIEE team in this field. The Thesis is accordingly organized in three parts and some appendices. The first resumes the general model, including a quite new formulation of optical gain that has been only partially published, at the date of the presentation of this work. Also the most relevant non-idealities are here discussed and solved. The second part collects the several experimental studies that have been carried out and published on real devices, looking for physical validation of the theoretical predictions, or, conversely, measuring the magnitude of some relevant quantities requested for calibrating the model. The third part faces the issue of failures. Experimental cases, with the interpretation of the most puzzling ones, are complemented with a sort of dictionary of failure modes, that is the systematic prediction of the measurable electric and optic degradations, to be compared with the real world for sorting and classifying effects and then addressing the physical analysis to discover their root causes. The final appendices resume some more theoretical details, or report on still incomplete studies that require further refinement and calibration.