Papers by Author: Christiane Dubois

Abstract: Implantation of 18O into highly B-doped and undoped silicon provides the possibility to investigate the effect of B-doping and to distinguish the processes of in-diffusion and out-diffusion of oxygen by profiling of 16O and 18O, respectively. The simultaneous in- and outdiffusion of oxygen was observed at 1000°C under oxidizing conditions. For silicon, heavily Bdoped to concentrations of 􀀀 1019 B cm-3, oxygen tends to diffuse out toward the surface. Moreover, a fraction of the oxygen from both sources, implanted 18O and in-diffused 16O, also migrates deep into the substrate and is trapped far beyond the mean ion range RP in the depth of x
3RP at the so-called trans-RP gettering peak. In undoped silicon oxygen accumulation only takes place at vacancy-type defects introduced by ion implantation at a position shallower than RP. The mobility of oxygen implanted into B-doped Si is higher than for implantation into undoped Si. Highly mobile defects are suggested to be formed in B-doped silicon beside the common mobile interstitial oxygen, Oi, and the immobile SiOX precipitates. These I OXBY defects may involve selfinterstitials, I, and O and B atoms. The trans-RP peak appears due to the decay of these defects and the segregation of their constituents.

Abstract: 6H and 4H–SiC epilayers were Al-implanted at room temperature with multiple energies (ranging from 25 to 300 keV) in order to form p-type layers with an Al plateau concentration of 4.5×1019 cm-3 and 9×1019 cm-3. Post-implantation annealing were performed at 1700 or 1800 °C up to 30 min in Ar ambient. During this process, some samples were encapsulated with a graphite (C) cap obtained by thermal conversion of a spin-coated AZ5214E photoresist. From Atomic Force Microscope measurements, the roughness is found to increase drastically with annealing temperature for unprotected samples while the C capped samples show a preservation of their surface states even for the highest annealing temperature. After 1800°C/30 min annealing, the RMS roughness is 0.46 nm for the lower fluence implanted samples, slightly higher than for unimplanted samples (0.31 nm). Secondary Ion Mass Spectroscopy measurements confirm that the C cap was totally removed from the SiC surface. The total Al-implanted fluence is preserved during postimplantation annealing. A redistribution of the Al dopants is observed at the surface which might be attributed to Si vacancy-enhanced diffusion. An accumulation peak is also observed after annealing at 0.29 9m, depth corresponding to the amorphous/crystalline interface that was determined on the as-implanted samples by Rutherford Backscattering Spectroscopy in channeling mode. The redistribution of the dopants has an impact on their electrical activation. A lower sheet resistance (Rsh= 8 k
) is obtained for samples annealed without capping than for samples annealed with C capping (Rsh= 15 k
).

Abstract: Al-Si patterns were formed on n-type 4H-SiC substrate by a photolithographic process including wet Al etching and Si/SiC reactive ion etching (RIE) process. RF 1000°C annealing under C3H8 flow was performed to obtain p+ SiC layers by a Vapour-Liquid-Solid (VLS) process. This method enables to grow layers with different width (up to 800 µm) and various shapes. Nevertheless the remaining Al-based droplets on the largest patterns are indicators of crack defects, going through the p+ layer down to the substrate. SIMS analyses have shown an Al profile with high doping concentration near the surface, high N compensation and Si/C stoechiometry variation between the substrate and the VLS layer. The hydrogen profile follows the Al profile in the VLS layer with an overshoot at the VLS/substrate interface. I-V measurements performed directly on the semiconductor layers have confirmed the formed p-n junction and allowed to measure a sheet resistance of 5.5 kW/ı