4H-SiC epitaxial growth on 2˚ off-axis substrates using trichlorosilane (TCS) is presented. Good surface morphology was obtained for epilayers with C/Si ratios of 0.6 and 0.8 at a growth temperature of 1600°C. The triangle defect density was reduced to a level below 5 cm -2 at 1600°C and below 1 cm -2 at 1625°C for a C/Si ratio of 0.8. Photoluminescence (PL) measurements were carried out with band-pass filters of 420 nm, 460 nm, and 480 nm to detect stacking faults. A stacking fault density of below 5 cm -2 was achieved at 1600°C and 1625°C with a C/Si ratio of 0.8. The optimal conditions for TCS growth were a C/Si ratio of 0.8 and a growth temperature of 1600°C. The evaluation of stacking faults and etch pit density indicated that the use of 2˚ off-axis substrates and TCS is effective for reducing basal plane dislocations. Comparing these results to those using silane (SiH 4 ) with HCl added, it was demonstrated that TCS is much more suitable for obtaining high-quality epilayers on 2º off-axis substrates.
The stacking fault formation during physical vapor transport growth of heavily nitrogen-doped (mid-10 19 cm −3 ) 4H-SiC crystals was investigated. Low-voltage scanning electron microscopy (LVSEM) observations detected the stacking fault formation on the (000-1) facet of heavily nitrogen-doped 4H-SiC crystals. Stacking faults showed characteristic morphologies, and atomic force microscopy (AFM) studies revealed that these morphologies of stacking faults stemmed from the interaction between surface steps and stacking faults. Based on these results, the stacking fault formation mechanism in heavily nitrogen-doped 4H-SiC crystals is discussed.
The strain fields in a 4H-SiC homo-epitaxial layer deposited on a nitrogen-doped 4H-SiC substrate were studied using Raman scattering microscopy. The cross-sectional (1-100) and (11-20) surfaces of the epitaxial substrate were examined through the peak shifts of several Raman-active phonon modes for 4H-SiC, and tensile strain was found along the direction of 4° off the c -axis at the epilayer/substrate interface. The effect of the facet trace in the substrate, which has a higher nitrogen concentration than the other parts of the substrate, was also studied. The tensile strain at the epilayer/substrate interface was found to be hardly enhanced for the epilayer deposited on the facet trace.
The etching mechanism of SiC single crystals by molten KOH has been investigated. The etching process is significantly affected by the etching ambience: the etching rate is greatly reduced by a nitrogen gas purge. This result clearly suggests an essential role of dissolved oxygen in the melt. SiC{0001} surfaces show a large surface polarity dependence, where the etching rate of SiC(0001)C is about four times larger than that of SiC(0001)Si. The etching rate of SiC(0001)C exhibits an Arrhenius type temperature dependence with an activation energy of 15–20 kcal/mol. The obtained activation energy and selectivity between the (0001)C and the (0001)Si surfaces are quite similar to those for thermal oxidation, which implies that the surface oxidation process occurs during molten KOH etching of SiC and is the rate-limiting step for the etching. We have conducted a comparative study of molten KOH etching with thermal oxidation in regard to the crystal orientation, polytype and carrier concentration dependence.
The defect structure at the growth front of 4H-SiC boules grown using the physical vapor transport (PVT) method has been investigated using high resolution x-ray diffraction and x-ray topography. The crystal parameters such as the c -lattice constant exhibited characteristic variations across the growth front, which appeared to be caused by variation in surface morphology of the as-grown surface of the boules rather than the defect structure underneath the surface. X-ray topography also revealed that basal plane dislocations are hardly nucleated at the growth front during PVT growth of 4H-SiC crystals.
'/%3e%3cuse xlink:href='%23a' transform='rotate(60)'/%3e%3ccircle r='17' fill='%23000095'/%3e%3ccircle r='15' fill='%23fff'/%3e%3c/svg%3e)