Papers by Author: Francis Baillet

Abstract: Raman spectroscopy and photoemission microscopy were coupled as two complementary non-destructive optical techniques in order to study biased 4H-SiC pin diodes. These two characterization tools have been largely used for the study of semiconductors but the combination of these two techniques has hardly been reported so far. Some structural defects inducing the same electrical damage could be discriminated and identified. Temperature could be measured in operating devices and the influence of the diode operating mode on the Raman signal could be evidenced.

Abstract: Structural defects in SiC crystals were investigated and 4H-SiC pin devices were characterized by micro-Raman scattering and photoemission. With the experimental set-up presented, defects could be successfully detected in SiC crystals but stacking faults could not be detected with micro-Raman scattering, although they could be detected by photoemission. Residual stress could be evaluated in 4H-SiC devices, as well as the temperature increase associated with the devices powering. A good correlation was found between the characterization techniques used: micro-Raman scattering and photoemission.

Abstract: Modeling and simulation of the SiC growth process is sufficiently mature to be used as a training tool for engineers as well as a decision making tool, e.g. when building new process equipment or up-scaling old ones. It is possible to simulate accurately temperature and deposition distributions, as well as doping. The key of success would be the combined use of simulation, experiments and characterization in a "daily interaction". The main limitation in SiC growth modeling is the accurate knowledge of physical, thermal, radiative, chemical and electrical data for the different components of the reactor. This is the weakest link in developing completely predictive models. In addition, the link between the thermochemical history of the grown material and its structure and defects still needs further development and input of experimental data.