Combination of Vertical and Horizontal GaN Transistors for Efficient Power Conversion Electronics – C-PI-GaN

The main objective of this project is the development of a new PECs technology based on GaN vertical high-power switch monolithically integrated with HEMT GaN-device. To achieve the goal of this project we divided it to two complementary objectives: First, fabrication of vertical transistors using freestanding GaN substrate. The state-of-the-art low defect density of this last allowed us to push the breakdown voltage (VBR) of the transistor over the highest performance achieved so far. Scalability of the VBR and RON is a side knowledge to the project, but mandatory to address specific market applications. Second, integrating monolithically a horizontal (standard) GaN-HEMT along with the Vertical GaN based transistor. The horizontal device is the first essential elementary device for the design and fabrication of the monolithically integrated driver to be co-integrated with the vertical high-power switch on a same die. This advanced approach is novel in the GaN field and is one of the strong research part of this project.
In these two main objectives, an extensive research work has been performed on the epi-layer growth conditions, TCAD simulation, device fabrication process and material and device characterization.

In this project most of our objectives were achieved and deliverables honored. It's worthy to notice that in the framework of this project, the first GaN vertical device is achieved in France. Indeed, the manufacturing of this innovative device is 100% fabricated in national laboratories starting from the substrate manufacturing to the characterization of the final device. As continuity of this encouraging results, CPiGaN partners started several PhD students projects dealing with different aspects of the vertical GaN device, starting from the epi-growth to the characterization of the final device. On the other hand, two ANR projects also dealing with GaN vertical device were granted in 2022 (ANR-ELEGANT ant ANR-ASGEIR).

The work started in C-Pi-GaN project is pursued in the projects below :

Ø ANR-ASGIER-PRC-2023 project where we will develop compliant substrate to accommodate the lattice mismatch between Silicon and GaN epi-layers. The consortium of this project is : CRHEA (PI), LN2, GREMAN, GREYC

Ø ANR-ELEGANT-PRC-2023 project where we will develop vertical GaN device on silicon substrate. The consortium of this project is : LETI (PI), AMPERE, CRHEA, LN2

Ø The consortium of CpiGaN, started Several PhD students projects dealing with different aspects of the vertical GaN device starting from the epi-growth to the characterization of the final device.

The quantity and quality of scientific production of researchers involved in CPiGaN project is the most important impact of this collaborative work. Even with all complications and delays induced by different COVIDs episodes that we faced during this project, the research work carried on the development of GaN vertical transistor technology was very fruitful. Indeed, thanks to the involvement of all partners, CpiGaN project produced five papers published in international journals and more than four conferences where the project results were presented.

In a world where electrical energy plays a more and more important role in the energy mix in reason of the worldwide increasing use of renewable energies, the market of Power Electronic Conversion Systems (PECS) is constantly growing in size and in complexity to address the need and challenges of electric power conversion at all levels. Today’s PECS market is largely dominated by silicon-based technology, as it is low-cost and mature. However, this technology limits the trade-off between size, energetic efficiency and maximum operating voltage due to the intrinsic limitation of the Si semiconductor. While emerging technologies such as electric vehicles and further applications keep demanding higher performance for operating voltage up to the kV range, it appears crucial to overcome these limitations to follow these major technological changes.
One promising solution to this problem is the use of wide-bandgap semiconductor technologies Gallium Nitride (GaN) which is far superior in terms of electron mobility and breakdown voltage. GaN technology is presently being used in applications up to 650V by means of horizontally-configured high electron mobility transistors (HEMTs) on foreign substrates (Si or SiC). Although GaN HEMT technology is exhibiting impressive performance, we are still not using all the capabilities of this material owing to several limitations caused by the growing of GaN epi-layer in on foreign substrate.
On the other hand, recent progress in material science have enabled the fabrication of GaN substrates at reasonable costs. Therefore, in this project, we target to overcome these limitations by using a GaN substrate and a homo-epitaxial growth approach to fabricate Vertical GaN transistors (MOSVFETs). Indeed, previous works in literature have demonstrated the tremendous gain of power density with the vertical approach on GaN bulk substrate.
The cutting-edge idea driving this proposal is the monolithic combination of horizontal HEMT and vertical MOSVFET transistors within the same chip. What makes this proposal stand out from previous works is the fact that, for the first time, we will demonstrate the monolithic integration of a high-speed HEMT-based driver with a high-voltage MOSVFET-based power switch.
This approach enables to benefit from the superiority of normally-off GaN MOSVFETs in terms of specific ON-resistance while enabling the integration of a driver circuit featuring GaN HEMTs for very high speed switching. The result will be a state-of-the-art PECS capable of handling voltage up to 1200V with exceptional characteristics in terms of robustness, reactivity, power density and efficiency, all thanks to the combination of both configuration: high-voltage handling and current density of vertical devices and high cut-off frequency of horizontal devices. This technological challenge represents thus a significant breakthrough while being achievable by current expertise and technological capabilities of the partners of this project.