Study Of The Growth Of SiC Crystals With Different Morphologies In Matal-silicide Fluxes

As a third generation wide bandgap semiconductor, silicon carbide (SiC) has been developed accompanying the progress of elemental semiconductors, e.g. Si, of the first generation and compound semiconductors, e.g. GaAs, GaP and InP, of the second generation. Due to the excellent physical and chemical properties, SiC bulk crystals and SiC nanomaterials have tremendous potential applications in the fabrication of optic and electronic devices, especially those suitable for operation under harsh environment such as high temperature and high irradiation. Focusing on the syntheses of SiC crystals with different morphologies using metal silicide melts as fluxes, this dissertation is composed of two parts: the study on the nucleation and growth of SiC single crystal in metal-silicide melts in the first part, and the synthesis, microstructure and property of SiC nanometerials by solution methods in the second part.In the first part two experiment routes were performed to study the nucleation and growth of SiC crystal from fluxes: spontaneous melt infiltration (MI) of the metal-silicide melts into porous SiC powder preforms and solution method using metal-silicide alloy as the fluxes and graphite crucible providing carbon source. The results showed that Fe_xSi_y (FeSi, Fe₃Si, Fe₅Si₃) melts could infiltrate the SiC powder preforms and produced densified SiC/Fe_xSi_y composites with mechanical properties superior to monolithic silicides. More importantly, with the decrease in SiC particle dimension (<0.5μm), the dissolution of SiC in some of the melts resulted in anomalous grain ripening and single SiC crystal growth, strongly indicating that Fe_xSi_y can be used as an appropriate flux to grow SiC single crystal. Because graphite precipitation, an undesirable phenomenon for SiC crystal growth, was found in the Fe₃Si melt, the 1600℃ isothermal section of the Fe-Si-C phase diagram is constructed to investigate how the melt composition influences the SiC precipitation. Over the isothermal section, X_Si=27mol% in the Fe_xSi_y system is determined as the critical value, over which SiC crystal growth can occur, otherwise only leading to free carbon precipitation. Among the Fe_xSi_y (FeSi, Fe₃Si, Fe₅Si₃) system used in the melt infiltration study, Fe₅Si₃ alloy is first regarded as a proper melt for SiC crystal growth.Based on the above analysis, solution growth of SiC crystals is practiced by heating Fe₅Si₃, FeSi, and FeSi₂ in graphite crucibles. The experimental results revealed that carbon from the crucible can be dissolved and SiC crystals can grow out of the melts. All the facts confirm that the Fe_xSi_y (X_Si > 27 at.%) melt have a proper carbon solubility and are suitable for the nucleation and growth of SiC. Similar experiments were performed for the Co_xSi_y (CoSi, Co₂Si, CoSi₂) fluxes. Only CoSi melt was able to spontaneously infiltrate into SiC powder preforms where SiC crystals with dimension of 0.5 mm were formed, and the maximal growth velocity of SiC crystal was 120μm/h. For Co₂Si and CoSi₂ melts, they could not spontaneously infiltrate into SiC powder preforms, and no SiC precipitation was found. Adding chromium (Cr) in CoSi melt can increase the carbon solubility in the melt and obtained SiC crystals with larger dimension, which proved that adding the metal whose electron number of the 4f layer is small to the Si melt can increase the carbon solubility. Ti-Si melt was also tried for SiC crystal growth, in which Ti-Si with 77atom% Si content is suitable for SiC crystal growth.All the XRD patterns of the SiC crystals growth in metal-silicide melts demonstrated that the crystals mostly belong to zinc blende structure, i.e. 3C-SiC (p-SiC), with a few of 6H-SiC which are usually ascribed to the stacking defaults or other faults in 3C-SiC. The Raman spectra affirm again that the SiC crystals with some stacking faults are basically 3C-SiC. The LO phonon mode's shifting in the Raman spectrum also shows that the SiC crystals have a large charge carrier concentration, which we think due to high level metal doping in the SiC crystals. According to the growth SiC crystal microstructure and crystal growth theory, the multi-nuclei growth mode of SiC crystal growth in the melts is proposed.In the study of solution growth SiC crystal, SiC nanowires were found on the surface of the melts when the relatively high oxygen impurity was present in the furnace. Inspired by the important phenomenon, the syntheses of SiC nanowires and SiC/SiO₂ nanocables were investigated by the simple vapor-reaction approach in the second part of this dissertation. Metal silicides were employed as the starting materials, and graphite plate acted as the carbon resource. SiC nanowires can bealways found at the surface of Fe-Si and Ni-Si solidified layers when the heat-treatment temperature is higher than their melting points. With the enhancing of heat-treatment temperature and prolonging of dwelling time, the diameter of SiC nanowires has the trend of increasing. Due to the poor wetting between Ni-Si alloy and graphite, the Ni-Si melt did not spread but forming liquid balls on the graphite plate at the temperature 200℃ above its melting point. SiC nanowires grew at the surface of liquid ball and formed villiform nano wire-balls. During the cooling stage in some cases, the SiO reacted with the CO in the furnace to form SiO₂ which then deposited on the surface of SiC nanowires, forming the SiC/SiO₂ nanocable with SiC as the core and SiO₂ as the wrapping layer.The SiC nanowire synthesized on the surface of silicide liquid films were characterized by XRD, SEM, TEM, Raman and FTIR. The results show that the prototype of SiC nanowires is also 3C-SiC (β-SiC), with round and hexagonal cross sections. Solidified liquid droplets were often found attaching to the tips of these SiC nanowires, a well recognized evidence of vapor-liquid-solid (VLS) reaction responsible for the nanowire growth. Base upon these, a growth mechanism combining solid-liquid-solid (SLS) mode and VLS mode is proposed. The carbon in the graphite dissolved in the melt layer, and the supersaturated carbon then reacted with the Si to form SiC embryos by the SLS reaction. Because of its low density compared to the melt, the small SiC embryos just nucleated floated over the surface of the melt, and pushing up the melt to form small alloy droplets on the top. Then the vapor phases (CO and SiO) were dissolved in the alloy droplets and reacted under the catalytic action of Fe or Ni to make SiC embryos grow along a preferential crystallographic direction. The metal elements (Fe and Ni) played two roles: increasing the carbon solubility and catalyzing the reaction.Silicon carbide (3C-SiC) nanorods were also synthesized in a special vapor-solid (VS) process. In this process, Si powder and carbon black were separated, but still reacted via VS reaction to form SiC at 1470℃ for 1-9 hours, and the morphologies of the products change from particles, non-regular short bar to regular and even hexagon prism shaped nanorods. XRD, IR, SEM and TEM were employed to characterize theSiC nanorods. Based on the characterizations, a vapor-solid formation mechanism of SiC nanorods is proposed, i.e. the gaseous silicon reacted with carbon black to form the SiC nanorods in a proper heat treatment conditions.