Low-temperature halo-carbon homoepitaxial growth of 4H-silicon carbide

New halo-carbon precursor, CH3Cl, is used in this work to replace the traditional C3H8 gas as a carbon precursor for the homoepitaxial growth of 4H-SiC. The traditional SiH4-C 3H8-H2 systems require high growth temperatures to enable the desirable step-flow growth for high-quality epilayers.;A well known problem of the regular-temperature growth is the homogeneous gas-phase nucleation caused by SiH4 decomposition. However, the degree of Si cluster formation in the gas phase and its influence on our low-temperature epitaxial growth was unknown prior to this work.;Growth at temperatures below 1400°C was demonstrated previously only for a limited range of substrate surface orientations and with poor quality. Mirror-like epilayer surface without foreign polytype inclusions and with rare surface defects was demonstrated at temperatures down to 1280-1300°C for our halo-carbon growth. Quantitatively different growth-rate dependences on the carbon-precursor flow rate suggested different precursor decomposition kinetics and different surface reactions in CH3Cl and C3H 8 systems.;Photoluminescence measurement indicated the high quality of the epilayers grown at 1300°C. A mirror-like surface morphology with rare surface defects was demonstrated for the growth on low off-axis substrates at 1380°C.;The most critical growth-rate limiting mechanism during the low-temperature epitaxial growth is the formation of Si clusters, which depleted the Si supply to the growth surface, in the gas phase. Presence of chlorine in the CH 3Cl precursor significantly reduces but does not completely eliminate this problem.;The addition of HCl during growths improved the growth rate and surface morphology drastically but also brought up some complex results, suggesting more complex mechanisms of HCl interaction with the gas-phase clusters.;These complicated results were explained partly by an additional mechanism of precursor depletion enhanced in presence of HCl.;Complex changes in the effective silicon to carbon ratio in the growth zone indicated that the supply of carbon species may also be enhanced at least at low HCl flow rates. This fact allowed us to suggest that the gas-phase clusters may contain a significant amount of carbon. The new model assuming coexistence of the silicon and carbon in the gas-phase clusters enabled the explanation of the complex experimental trends reported in this work.