Deposition of silicon carbide coatings on alumina particles in a microwave plasma-assisted spouted bed reactor

Abstract

Alumina particles were successfully coated with silicon carbide (SiC) layers in a microwave plasma-assisted spouted bed reactor. Methyltrichlorosilane (MTS) was used as precursor for the SiC deposition reaction, and argon served as both carrier and plasma gas. The microwaves were guided from a generator along a rectangular waveguide. A quartz tube was mounted between two support flanges and positioned perpendicular to the metallic waveguide. A graphite nozzle was inserted into the bottom of the quartz tube to bring about the spouting action of the bed. A metallic grid was installed at the top of the tube to prevent particles exiting the reaction zone. Process parameters under investigation were enthalpy, pressure and the hydrogen-to-MTS ratio. The design of experiments (DoE) followed that of a Box-Wilson 3-dimensional central composite design (CCD), covering a wide range of experimental parameters, within the capability of the system. The measured growth rates ranged from 50 μm/h to 140 μm/h, with mass deposition rates from 19.1 g/(h·m2) to 331 g/(h·m2). Response surface methodology (RSM) and analysis of variance (ANOVA) yielded 3D surface contour plots for navigating the design space, and the models indicated the most significant term to be pressure. Characterisation techniques for the SiC layers included scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD) spectroscopy, X-ray photoelectron spectroscopy (XPS), energy-dispersive X-ray (EDX) spectroscopy and Fourier-transform infrared (FTIR) spectroscopy. These techniques assisted in developing a colour chart as well as a similar morphological chart used to indicate the change in morphology of the layers throughout a 2D design space. High enthalpy and pressure values tended to produce dark-coloured layers, often accompanied by carbon-rich deposits. XPS characterisation also indicated the presence of organosilicons, likely the remnants of unreacted or partially-reacted MTS compounds embedded in the layers. It is evident that within the design space, the optimal region for SiC deposition requires high enthalpy (~ 5 MJ/kg) and pressure (> -60 kPag), with reasonable hydrogen-to-MTS ratios (~ 5:1). The quality (i.e. crystallinity, particle size, Si:C ratios) of the layers appears to improve at these conditions, at the expense of decreased deposition rates. Brunauer–Emmett–Teller (BET) analysis indicated a pore size distribution showing the majority of pore sizes to be below 20 nm, classified as mesoporous. Heat treatment of the particles (in air at 1 000 °C) yielded an increased oxygen content within the layers, likely a result of oxidation of the Si and SiC. By contrast, the chlorine content decreased, indicative of the organosilicon reactions forming volatile hydrogen chloride (HCl). SEM images demonstrated that the layer boundaries became sharper and more defined after heat treatment. A 2D finite-element model was developed to assist in investigating the effect of the plasma parameters on the SiC deposition mechanism. Limited access to computing power required careful consideration and simplification of the model, while still successfully coupling relevant physical processes, such as laminar flow, heat transfer, plasma chemistry and chemical vapour deposition (CVD) processes. Some experimental observations were successfully predicted by the model, such as the total volume of the quartz tube occupied by the plasma zone, gas temperatures during operation (~ 1 000 °C), gas velocities (~ 2 m/s to 12 m/s), and the effect of a SiC layer growing on the reactor walls. Other results comparable with literature values include the electron densities (~ 5.93 × 1019 m-3) and energies (~ 1.22 eV). The deposition rates on the particles were not in agreement with experimental results, likely due to a simplified MTS decomposition mechanism, as well as the use of a temperature dependent sticking coefficient.