Characterization, modeling and aging behavior of GaN power transistors

Seeking green cities, reducing the greenhouse effect and finding a substitute to oil uses are all strong reasons that attract research and industry to find other alternatives. One of these solutions is a broader use of electrical energy. Over the years, this energy vector has been widespread by using power electronics in many automotive, aeronautical and military applications. A power electronics system is mainly composed of power components such as diodes and transistors. Depending on the applications, these components may have specific characteristics to withstand severe operating conditions such as high temperature, high power and high-frequency operations. In order to reach these characteristics, the power devices must be made of specific semiconductor materials. Initially, the most common semiconductor material (Silicon) is used, but it shows some performance limitations for power electronics applications. Over the last two decades, with research and development, there has been new materials that have demonstrated better characteristics than Silicon. Silicon Carbide (SiC) and Gallium Nitride (GaN) have shown many advantages for use in power electronics. They represent the materials estimated to be widely used for power electronics shortly; they have attracted significant interest benefiting from their excellent semiconductor properties. Power components based on these materials have recently attracted much attention. They have presented the fashion in which the scientific community has not ceased to work on it.The presence of advantages for a device is not sufficient before its emergence in real applications. The reliability issue is essential and represents an inevitable phase before validating its use. Although GaN components have shown many advantages, users are wondering: “Are we ready to use these components in real applications?” Much current research is underway to answer this question. In this thesis, we will participate in the answer. The main objective of this thesis is to study some elements of the GaN High Electron Mobility Transistors (HEMT) reliability for power applications. The power transistors reliability study is carried out by a lot of theoretical and experimental works. It can be performed at the semiconductor, packaging or system level. In this thesis, we present theoretical and experimental works carried out only at the semiconductor level.GaN transistors suffer from a reversible physical phenomenon called “trapping”, this phenomenon represents a source of instability of the device physical and electrical characteristics. It has presented an essential element for the reliability works of GaN HEMTs. The full understanding of the effect of this phenomenon and its evolution with ageing is not yet fully accomplished. In this thesis, we present theoretical and experimental works to understand better the reliability linked to this phenomenon and the possible interaction with ageing in power cycling conditions. This work was carried out within the framework of the French MEGaN project, founded by the public investment bank (bpi France). These studies were carried out on two different components; the first one is manufactured and supplied by a project partner CEA-LET, the second one is a commercial power device fabricated by GaN-Systems Inc.