Posted on the 18th March 2020
- PhD Researcher position
PhD proposal M/F
Exploring self-healing solutions for Aluminum source layers in Power Devices
Industrial Partner: MITSUBISHI ELECTRIC R&D CENTRE EUROPE, 1 allée de Beaulieu, CS 10806, 35708 Rennes Cedex 7, France
Academic Partner: CEMES -Centre d’élaboration de matériaux et d’études structurales
– CNRS 29 Rue Jeanne Marvig, 31055 TOULOUSE Cedex 04
Type of contrat: 3 years contract, as soon as possible
Research topic: Reliability of power devices
Subject proposed by :
Julio BRANDELERO, Mitsubishi Electric R&D Centre Europe, Rennes
Jeffrey EWANCHUK, Mitsubishi Electric R&D Centre Europe, Rennes
And Marc LEGROS, CEMES-CNRS, Toulouse
Power electronics applications have expanded in many fields such as domotics or energy conversion, but their reliability is mostly critical in the transportation field and off-shore applications. Anticipating or even preventing their failure is a key technical issue.
In recent years, several weak spots have been identified in the structure of modern Silicon-based power MOSFETs and IGBTs, and some solutions have been found to increase their resistance to disruption (solder, packaging…). However, the aging of the top metal source, mainly made of Aluminum or Aluminum alloy has persisted as an intrinsic phenomenon, which degrades the electrical performance of the device over time. The deterioration of the Al layer occurs through mechanisms that involve grain boundary diffusion, crack formation and surface oxidation, also driven by stresses arising from thermal mismatch between the metal and the silicon, as described in Fig. 1. Similar oxidizing cracks develop between the metallization and the wire bondings, leading to local failure of the device .
The study will focus on the aluminum layer of the power device that interfaces the silicon die and the electrical circuit.
The objective of this thesis is to find technical ways to slow down the mechanisms that cause the deterioration of the Aluminum layer, either by increasing the metallic contact between the wire bondings and the metal in the initial state and/or to restore this electrical contact during operation.
Thesis conditions and development
Thesis starts: As soon as possible
Presence: Mainly at the CEMES in Toulouse (85% time). Regular visits are planned to Mitsubishi Electric R&D Centre Europe, Rennes.
Planning & reporting: The work program is divided as follows:
- T1: Increase the metallic contact
- T2: Restore the electrical connection
- R: Thesis writing
Monthly reports (1 page) and quarterly meetings (alternately in Toulouse and Rennes)
Education and experience required
- Engineer’s degree or Master’s degree with a focus in Materials Science
- Simulation and programming softwares (COMSOL Multiphysics, Python)
- Competence in materials characterization tools (TEM, FIB)
- Strong general scientific knowledge and multidisciplinary opening (materials science, electrical, mechanical, thermal, chemical, mathematical)
- Power electronics skills would be a plus
- Communication and writing skills in English
- Motivation and dynamism to work in a research environment
- Ability to work in a multicultural and international environment
Send your CV and motivation letter in a pdf format by mail (by specifying in object: your name and the reference JCB_PhD_2020_CEMES): firstname.lastname@example.org
-  D. Martineau, C. Levade, M. Legros, P. Dupuy, and T. Mazeaud. “Universal mechanisms of Al metallization ageing in power MOSFET devices.” In: Microelectronics Reliability 54.11 (2014), pp. 2432–2439.
-  R. Ruffilli et al., Mechanisms of power module source metal degradation during electro-thermal aging. Microelectronics Reliability (2017).