Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
Super junctions (SJs) have enabled unprecedented performance in Silicon power devices, which could be further improved by applying this concept to wide bandgap semiconductors like gallium nitride (GaN). Currently, polarization super junctions (PSJs) are the most promising candidates for GaN SJs. Yet, until now, p-type doping of the GaN cap layer was required to ensure the presence of a 2-D hole gas (2DHG), and the proper charge matching between the 2DHG and 2-D electron gas (2DEG), which is fundamental to the operation of SJs. This approach, however, requires precise control of the p-GaN doping level, which is very challenging and has hindered the demonstration of high-performance PSJs. Besides, while PSJs are particularly promising for multichannel structures aimed at reducing the sheet resistance, achieving proper charge matching combined with large electron density for all the buried channels is challenging. Here, we propose a simple and robust platform for intrinsic PSJs (i-PSJs) that enables excellent charge matching for a wide range of structures, without relying on doping. We show that surface donor states are the origin of charge mismatch and provide a strategy to minimize their impact. Simulated devices based on this structure show optimal carrier depletion with a flat electric field profile in the whole drift region. Finally, we extend this concept to multichannel i-PSJ structures. We demonstrate a much-reduced sheet resistance down to 58 Ω /sq and present a robust strategy to achieve charge balance, which enables reducing the on-resistance without degrading the off-state performance, thus greatly improving the device figure-of-merit.
Edoardo Charbon, Claudio Bruschini, Ekin Kizilkan, Utku Karaca, Myung Jae Lee
Giovanni De Cesare, Sylvain Dunand, Jonathan Emanuel Thomet, Mikaël Martino, Luca Massimiliano Antognini, Matthew James Large
Mohammad Samizadeh Nikooytabalvandani