Design of new high voltage SuperJunction power switches – SUPER SWITCH

Several technological steps have to be studied, improved and compared : deep etching (LAAS and GREMAN), doping (LAAS, GREMAN and IBS) and dielectric deposition into deep trenches (LAAS).

Deep vertical etchings were obtained by ICP-DRIE at LAAS and by electrochemistry at GREMAN. PIII implantations were performed on top wafer at IBS: it remains to translate them into deep trenches at LAAS and GREMAN.

The following is the P-doping by PIII tests into the trenches; these tests have to be validated by the design of Superjunction diodes in a first time.

No scientific production to date.

The research and development in the fields of green renewable energies are one of the main targets for the future of humanity. The yield improvement of the energy chain absolutely needs a yield improvement of each part of it. The transport sector, for example, consumes 25% of world energy and uses more than half of the oil produced in the world. Consequently, the research orientations related to the use of alternative energies for road transport (hybrid and electric vehicles for instance) are of the most importance, but it will also be vital in the future to modify the ways of life associated to the use of private vehicles and the energy consumption.
Research activities in power electronics fully address these future changes. There are numerous power electronics applications (electrical traction, industrial drives, distribution network management, electrical household appliances, transport and portable units,…) using a large variety of power devices. The drastic performance improvement of these power devices, in terms of energy saving, cost, size, weight and reliability, is a key factor for the safeguard of the energy.
This is the context of the « SUPER SWITCH » research project here proposed .
The main objective of this project is to propose alternative solutions to the IGBT in power converters in the 600-1200 V breakdown voltage range. Today, up to 600 V, the main IGBT competitor is the MOSFET. Indeed, the MOSFET exhibits many interesting properties for power applications (high switching speed, high input impedance, thermal stability, internal free-wheeling diode) but, in high voltage range (600-1200 V), it is limited by its high specific on-resistance and, therefore, its important on-state voltage drop, which induces more on-losses, compared to bipolar devices: to find the best trade-off between these two parameters (specific on-resistance / breakdown voltage) is one of the major challenges for unipolar switches in this voltage range.
For this reason, several innovative unipolar structures have been recently proposed to overcome the theoretical limit of the “specific on-resistance / breakdown voltage” trade-off of conventional unipolar power devices. The best structure is the well-known SuperJunction MOSFET firstly manufactured by Infineon (COOLMOS™) and STMicroelectronics (MDMESH™). These devices are fabricated by multi-epitaxy, which is an expensive process.
In this project, we propose new technological ways for the realization of SuperJunction devices (diodes and MOSFETs) in the 600-1200 V voltage range, using a single epitaxy proccess: the realization of very performant power switches, at a lower cost compared to conventional processes of Infineon and STMicroelectronics, will be possible if the project succeeds.
Actually, based on complementary competences of LAAS-CNRS, LMP, IMS and IBS, the « SUPER SWITCH » project proposes a new technological process, using only one epitaxy for the realization of Deep Trench SuperJunction Diodes (DT-SJDiodes) and MOSFETs (DT-SJMOSFETs) including an original termination, called Deep Trench Termination (DT2). Since SuperJunction device performances are very sensitive to charge balance, several critical process steps have to be studied and developed: deep etching with a perfect verticality of the trench sidewalls, doping technique along trench sidewalls and trench filling with a suitable dielectric able to fill deep and wide as well as shallow trenches. The influence of these steps on electrical and thermal performances will be also investigated by means of simulation tools. Process simulations will allow to define the optimal fabrication process of Deep Trench SuperJunction devices. Finally, DT-SJDiodes, DT-SJMOSFETs and DT2 will be fabricated and characterized in terms of electrical performance, thermal performance and robustness.