Predictive Reliability Modelling and Characterization of Silicon Carbide Power MOS-Transistors in Grid-Connected Voltage Source Converters

Modern society's reliance on electrical energy is almost as critical as its reliance on food and water. In the UK, majority of the electrical energy is generated by electrical machines powered by fossil fuels. The principles of sustainability require that the energy consumption pattern changes since fuel reserves are finite. Furthermore, shifting away from fossil fuels is integral to the de-carbonisation of the economy which is critical for tackling global warming. To this end, substantial progress has been made on harnessing wind, solar and other renewable energy sources. However, change is also required in the manner in which electricity is transmitted and distributed through the grid. Renewable energy is usually intermittent and unpredictable, characteristics which make it unsuitable for direct connection with the electric grid. Renewable sources like wind and solar energy can only interface with the electrical grid through power electronics. Power electronics is required for the processing and conditioning of electrical energy so as to make it complaint with the grid. At the heart of power electronics, we have power semiconductor devices which have traditionally been fabricated out of silicon bipolar technology. However, silicon is reaching its fundamental limits in terms of energy density, hence, moving to advanced power materials like silicon carbide can give added impetus to the field of power electronics. Silicon carbide is a wide bandgap semiconductor with a higher critical electric field and higher thermal conductivity. In this project, the reliability of power converters implemented in Silicon-Carbide MOS-Transistor technology is investigated. These power converters will typically be used in off-shore wind-farms for power conversion in high voltage DC transmission (HVDC) systems. The converters can also be used in flexible AC transmission systems like STATCOMS (Static Compensators). The overall objective is to characterize the reliability of power converters implemented in silicon-carbide MOS transistors.

Society will benefit from this project because it feeds into the energy efficiency and economic de-carbonisation theme currently embarked on by both government and industry. Energy security not only relies on availability of energy sources but also the physical reliability of the power network infrastructure. The increasing use of power electronics in the national grid requires more effort in predictive reliability modelling of the semiconductors in the power converters.

Running power semiconductor devices hotter and faster will inevitably reduce the lifetime of the devices. In applications like electric vehicles, air-crafts and ships as well as grid-connected voltage sourced converters, reliability is critical because the consequences of premature failure are very serious indeed. The primary industrial impact of this project will be to give confidence to the power electronics community in the application of silicon carbide devices. Silicon carbide devices are very expensive and their credibility is staked on the ability to withstand higher temperatures while increasing power density. However, before industry can deploy SiC devices, there must sufficient understanding of the performance limits.

Working with ALSTOM and Dynex, the project will have an immediate impact in the power conversion industry. In applications like off-shore wind converters, where access for maintenance is expensive, accurate reliability prediction techniques will be a very useful design tool. The increasing use of power electronics in critical systems requires an increased focus on the principles of "design for durability". The results of this project will not be specific to silicon carbide but will be applicable to silicon devices as well.

Another sector that will be positively impacted by the project is the automotive sector which is currently embarking on increased vehicle powertrain electrification. Replacing the current silicon IGBTs and PiN diodes in electric vehicle 3-phase inverters with SiC MOSFETs and Schottky diodes will rapidly increase the power density and improve the energy conversion efficiency of the vehicles. Before embarking on such a significant change, the automotive sector would like to gain some confidence in SiC devices. The high cost of the devices can be offset by system level benefits of faster switching which will include shrinkage of the passive components. However, these system level benefits require faster switching which impacts reliability negatively. So the question for the automotive sector is how to strike the right balance between energy density/efficiency and reliability. This problem also extend to the electric train, ship and air-craft industries where power electronic devices play an increasingly important role. The results of this project will help in answering that question.

The project will also have an impact in the packaging community where modules are designed for various applications. The mechanical design of wire-bonds and die attach directly impact the electro-thermal and thermo-mechanical reliability. Deploying fast switching silicon-carbide devices in power electronic systems will place more focus on parasitic package inductances which coupled with the device capacitances will cause resonance and electromagnetic interference (EMI). At the moment, fast switching devices can be limited by parasitic inductances, hence, significantly more effort is required in the development of methods of suppressing EMI. Furthermore, packaging is a bottleneck in high temperature power electronics and this project will impact positively on the research efforts in high temperature packaging. High temperature gate drive circuitry and increased integration with the power module is another research area that will be impacted positively by this project.