Ms Dipanwita Dutta
Paul Scherrer Institute


Atomic-Scale Analysis of the SiC/Oxide Interface to Improve High-Power MOSFET (Metal-Oxide-Semiconductor-Field-Effect-Transistor) Devices


Silicon carbide (SiC) is currently the most promising material to be employed in power MOSFET as it exhibits superior performance and lower power dissipation over silicon. Its wide band gap, thermal conductivity, high critical field, and the ability to grow native oxide would make it the perfect MOSFET material for high power devices. There is a growing market for the more energy efficient SiC power devices benefitting from SiC MOSFET technology. Specifically this includes converters for the more decentralized grids connecting renewable energy generators to the electricity grid, traction power-modules to control trains, etc. A persistent obstacle towards the wider use of silicon carbide (SiC) technology, in particular for advanced high power (>~10 kV) and high temperature applications is the deteriorated MOS-channel mobility after thermal growth of the gate oxide insulator. Thus, the more energy efficient SiC power MOSFET is currently unable to replace the Schottky diode based power devices.

We report the first unambiguous evidence that it is the removal of carbon (from silicon carbide) during the oxidation process which leaves behind characteristic structural defects, carbon clusters or condensates at the interface, impairing the channel mobility of the semiconductor. Also we report a unique combined theory and experiment approach to analyze the near-interface defects in the form of impurities, crystal-defects and carbon clusters, their formation during oxidation and their detrimental influence on the electronic performance of the MOS-channel. On the basis of our mechanistic analysis we propose and discuss different measures to be taken to minimize or avoid the carbon accumulation, in order to obtain higher carrier-mobilities in the inversion channel below the SiC/SiO2 interface. This includes processing parameters as well as post processing passivation steps.


Dipanwita Dutta, during her Master thesis, worked on graphene and carbon nanomaterials where she developed a flexible transistor from turbostratic graphite obtained from pencil-drawing-on-paper. This raised her curiosity about the microscopic origins of semiconducting properties and diode and transistor-like physics. After completing her MSc degree in Materials Science and working as an R&D on innovative transparent heaters and solar cells from JNCASR, Bangalore she joined PSI for a PhD project sponsored by the Swiss Nanoscience Institute. She investigated the increased complexity of the carbon removal and oxidation process of thermal oxides on SiC vs. Si wafers towards improving SiC power MOSFETs.

Ms. Dutta, towards her PhD performed experiments mainly with force microscopy, atom probe and extensive Raman spectroscopy, electron microscopy to identify different defects and modifications of C accumulates present at SiO2/SiC interfaces. In an experiment and theory approach with partners from the University of Basel she unambiguously identified Carbon leftovers (clusters) in their chemical composition, their bonding and their impact on the near interface semiconducting behavior (e.g. mobility) in SiC. The work was presented in last year’s International Conference on Silicon Carbide and Related Materials (ICSCRM) and attained a wide resonance in the research community.