Development of an integrated optical E-Probe for GaN power transistor reliability analysis

AlGaN/GaN high electron mobility transistors (HEMT) are a key technology currently envisioned for future radar, satellite and communication applications, ranging from civilian to military use. Although the performance of AlGaN/GaN HEMTs presently reaches power levels up to 40W/mm at frequencies as high as 2-10 GHz, i.e., a spectacular performance enabling disruptive changes for many system applications, long-term reliability of AlGaN/GaN HEMTs is still a serious issue, not only in the UK and Europe, but also in the USA and Japan. There are several key factors affecting AlGaN/GaN HEMT reliability resulting in a variety of different failure mechanisms, including trap generation, metal migration and others. These are accelerated by: (i) device temperature, (ii) local stresses / strains (converse piezo-electric and thermal), (iii) high electric fields. Knowledge of these parameters is essential for reliability testing, in particular, for accelerated lifetime testing to predict mean time to failure (MTTF). The CDTR in Bristol developed and pioneered Raman thermography, to probe temperature and stress/strain with sub-micron spatial and nano-second time resolution in the only a few micron size active device area of AlGaN/GaN HEMTs, but there is presently no non-invasive probe available for experimentally quantifying electric field strength and its lateral distribution in particular when operating devices at high voltages. Therefore presently only simulation can be used to estimate electric field strength. The key aim of this research project is to develop, test and employ a non-invasive novel optical probe (E-probe) to quantify electric field strength and its lateral distribution in the device channel of AlGaN/GaN HEMTs, and to integrate it into Raman thermography, to enable simultaneous electric field, temperature and stress analysis of AlGaN/GaN HEMTs, to develop a unique and highly beneficial analysis technique for AlGaN/GaN HEMT reliability research. Experiments on degrading / stressing of devices to probe the resulting changes in the electric field strength and its distribution will be performed for state-of-the-art reliability research. Charge carrier traps generated during stressing change electric field strength which we expect to be able to probe directly here for the first time. Carrier trapping times range from milliseconds to seconds. We aim, a higher-risk component of this project, to also developing the ability to probing time-dependent changes in the electric field with carrier trapping / detrapping in the devices.

This proposal is focused on the development and application of a new technique, E-Probe, to gain novel insight into the reliability of GaN power electronic devices and their failure mechanisms, in particular of AlGaN/GaN high electron mobility transistors (HEMTs). The coming years will see the progressive adoption by the microwave industry of the GaN HEMT technology for both microwave power transistors and microwave monolithic integrated circuits (MMICs). At present, this technology has already demonstrated impressive microwave performances for high power amplifiers (HPAs) in terms of power level, power density, power-added efficiency (PAE), bandwidth and robustness, clearly exceeding those permitted by the existing Si LDMOS and GaAs PHEMT or HBT technologies, however, device reliability is still a major issue of concern. GaN HEMT reliability therefore needs to be addressed to achieve the implementation of this new technology, and economic impact for the UK. While the US and Japan had build up an impressive lead in GaN HEMT performance (also they still do have considerable reliability challenges), to guarantee an independent supply chain for GaN HEMT technology for the UK and Europe, several large scale European programme were initiated (e.g. KORRIGAN, GREAT2, MANGA), involving several key UK industries in particular QinetiQ Ltd and Selex UK (also partners in this EPSRC application), and also including UK universities such as the CDTR in Bristol. These activities are funded by the European Defense Agency (EDA), EC-FP7, and European Space Agency (ESA), and have moved Europe very much closer to the US and Japan in the status of the GaN technology. Overcoming current device reliability hurdles would therefore result in billions of economic benefit for the UK from GaN HEMT technology from UK partners involved in those programmes (QinetiQ Ltd, Selex UK), with further benefits for other UK aeronautics and defense companies (Airbus, RFMD UK, BAE Systems etc). While achieving reliable GaN HEMT technology is naturally an industry focused activities, there exist huge opportunities for UK academia to contribute to obtaining a reliable devices, in terms of understanding why devices fail, i.e., what are device failure mechanisms and what is the underlying physics, and to develop new analysis techniques and reliability testing concepts for GaN HEMT reliability assessment, all of those areas are patentable, and of commercial benefit. This is the target economic impact area of this proposal. Economic impact will also obviously arise from future GaN reliability improvements benefiting from the analysis technique developments in this proposal. This will benefit UK industry involved in this proposal, QinetiQ Ltd and Selex UK, and via the involved international partners, i.e., there is direct economic impact for the UK. Regular meetings with key industries (QinetiQ, Selex, TriQuint / see support letters) will ensure that research outcome of this project will be transferred to industry. Furthermore, Bristol and UK industries will benefit as new analysis technique developments (and corresponding device reliability improvements achieved with the aid of new techniques) have always great potential to provide the basis for the UK to be partners in future European but also US programmes. A previous successful example of economic impact of Bristol research in the GaN HEMT field include the CDTR / Bristol developed and pioneered measurement technique for the quantification of temperature and strain in devices, Raman thermography, developed within a previous EPSRC research project (GR/S76182/01), which is now considered as most acceptable way to quantify channel temperature in GaN HEMTs, not only in European GaN HEMT reliability programmes but also US and Japan programmes based on the work in Bristol. This is the reason why Bristol is involved e.g. in GREAT2 funded by ESA and DRIFT funded by ONR. Similar impact is envisioned for the proposed technique.