High Voltage Photo-Switches – HV-PhotoSw

A process flow, that is to say the succession of technological steps for the fabrication of these components, required the development and optimization of some technological steps. This was notably carried out within the framework of the postdoctoral researcher recruited into the project.
To optimize the design of the targeted components, numerical simulations, i.e. Technology Computer-Aided Design or TCAD type was carried out. The doctoral student recruited for the project contributed significantly to these numerical simulations for the optimization of the elementary cell which sets the current gain and the peripheral protection, which in turn determines the voltage withstand.
A set of masks was created thanks to scripts (Klayout/Ruby) allowing great flexibility. Numerous electrical and optical test fields have been carried out and their exploitation will allow partners to publish original results with a very fine level of analysis.
Two manufacturing lots were produced using the clean rooms of ESIEE (ESYCOM UMR CNRS 9007 laboratory) and NanoMat/UTT (L2n EMR CNRS 7004 laboratory).

Electrical characterizations on wafers were carried out at ISL in a high-voltage vacuum chamber for reverse characterizations and at Ampère for on-state characterizations. Electrical and optical characterizations on packaged components were carried out.
After analysis, an adjustment of the design of the components was made between the 2 batches (Klayout/Ruby).
The last manufacturing batch was finalized at the end of 2023 and will be subject to measurements outside the project.

The main objectives of the HV-PhotoSw project were achieved:
• design and manufacturing of a 10 kV BJT. We obtained 11.2 kV as the best voltage withstand and the on-state characterizations showed the good operation of the BJT with a current gain of 15 at Ic=15 A.
• optical control of the high-voltage BJT. We have successfully optically controlled optical BJTs with a pulsed laser and with a UV LED.
We thus hold a world record for a BJT component with high-voltage optical control, namely 10 kV.

Our paper received the “Best Student Paper” award at the IEEE CAS 2022 International Semiconductor Conference.

We plan to present our results, a priori in 2024, to institutions and industrials who have shown interest in our work.
The HV-PhotoSw project is now complete and we have consumed almost all of the allocated resources. The second batch will be completed in December 2023.
This leaves us with many interesting measurements to make and publish
• Characterization in switching of electrical BJTs, on a 1000 V bench and on a 10 kV bench,
• Characterization in switching and optical control of a 10 kV BJT,
• Characterization by µ-obic bench in order to judge the effectiveness of peripheral protections on specific test patterns,
• Testing a BJT in a DC-breaker function.

AMMAR, Ali, PHUNG, Luong Viêt, PLANSON, Dominique, et al. Design and Methodology of Silicon Carbide High Voltage Termination Extension for Small Area BJTs. In : Materials Science Forum. Trans Tech Publications Ltd, 2022. p. 613-618. hal 03703994

AMMAR, Ali, LAZAR, Mihai, VERGNE, Bertrand, et al. Design, Fabrication and Characterization of 10 kV 4H-SiC BJT for the Phototransistor Target. SCIENCE AND TECHNOLOGY, 2023, vol. 26, no 2, p. 193-204. hal-04276812

AMMAR, Ali, LAZAR, Mihai, VERGNE, Bertr, et al. Optimized Junction Termination Extension and Ring System for 11 kV 4H-SiC BJT. In : 2022 International Semiconductor Conference (CAS). IEEE, 2022. p. 191-194. hal-03856578

BROSSELARD, Pierre, PLANSON, Dominique, TOURNIER, D., et al. Design and Characterization of an Optical 4H-SiC Bipolar Junction Transistor. In : ICSCRM 2023-International Conference on Silicon Carbide and Related Materials. 2023. hal-04222211

AMMAR, Ali. Design, conception, and fabrication of high-voltage bipolar devices based on 4H-SiC. 2023. Thèse de doctorat. Lyon, INSA.

The deployment of renewable energy involves a mutation of the electric grids to account these new sources. Particularly, HVDC and MVDC grids are new technologies that demand high voltage power electronic converters. Silicon carbide (SiC) is a good material for high-voltage power devices while Gallium Nitride (GaN) is limited in voltage range and diamond remains a long-term target. 10 kV and above SiC-devices are targeted in research projects while 3.3 kV devices are under industrialization. However, to achieve high voltage power converters, it is needed to associate in series power modules or power components. So, one main issue is the design of highly insulated drivers (up to several 100 kV). There exists some solutions as the self-supplying but these solutions are complex and may decrease the global reliability of the system.
On the other hand, the photo-transistor is a classic device, and it is very common for the silicon ones. If some optical controlled SiC thyristors have been demonstrated, to our knowledge an optical controlled SiC Bipolar Junction Transistors (BJT) has not yet been demonstrated for high voltage applications. The advantage of the photo-transistor is to avoid the use of a circuit driver because of the electrical insulation may be easily achieved by an optical fiber.
So, the project proposes the design and fabrication of a high-voltage phototransistor. The interest of such a power device is very high because of the simplification of the drive insulation. Moreover, for a given die area, the BJT is the best component with the lower on-loss among all other SiC power semiconductor devices.
The consortium consists of 2 academic partners (Ampere Lab. and ISL) and one SME (NovaSiC).
NovasiC will optimize a fast chemical vapor deposition (CVD) for the low-doping thick epitaxial layer (n-type) needed to reach 10 kV devices. The target is to reduce the cost of this process which is a significant part of the global process costs for such high-voltage devices. The obtained results will be compared to commercial epitaxial-layers.
The project proposes the fabrication and the test of high voltage photo-transistors, BJTs, on two batches of fabrications. To reduce the risks in the project, some transistors on the wafers will have a base metalization to enable electrical tests. This type of cells may be used for the 4 targeted devices of the project. All of them are high voltage vertical devices with the target of 10 kV. This blocked voltage target has the advantage to be interesting for numerous applications (e-transformer, modules for HVDC or MVDC applications, DC breakers …) and not to be so difficult to be fabricated in an academic clean-room. However for a efficient BJT, the requirement for accurate process is high in terms of photolithography, dry etching, passivation and packaging. So technological facilities of the ESIEE are targeted which is recognized for its professionalism and support of research projects developed on its 600 m2 clean-room, offering a full 100 to 150 mm technology which is ideal for the actual SiC wafer sizes (4 to 6 inches).
To avoid strong constraints on the device fabrication, a solution is to develop all the high voltage BJTs on the same wafer (2 lots).
Depending on the project results, some targeted products are expected
• Large 10 kV photo-transistors for efficient and isolated converters (HVDC, MVDC), DC Breaker (solid state).
• Large 10 kV Photo-Darlington for all applications where the efficiency is not critical : single pulse applications, DC breakers (hybrid), active snubber ...
• 10 kV optical switch 10 kV, i.e. a large BJT with a low voltage silicon phototransisor and a self-supply : a bypass of the first target to compare phototransisor to more classical solution.
• Large 10 kV (even 15 kV) or 3.3 kV BJTs, as an alternative of SiC-MOSFETs without the gate oxide reliability issues for applications where a high efficiency is required: HVDC, MVDC, electric traction…