This work addresses a novel technique for selecting the best process parameters for the 4H–SiC epitaxial layer in a horizontal hot-wall chemical vapor reactor using a transient multi-physical (thermal-fluid-chemical) simulation model and combined with a machine-learning model. An experiment was performed to validate the feasibility of the numerical model. Secondly, a single-factor analysis was conducted to investigate the effects of process parameters, including the deposition temperature, inlet-flow volume, rotational speed of the susceptor, and cavity pressure, on the quality of the 4H–SiC epitaxial layer. Finally, a machine learning algorithm, the ant colony optimization-back propagation neural network (ACO–BPNN), was employed to develop the input/output model and optimize process parameters for obtaining a high-quality epitaxial layer and reducing the optimization cycle and costs. Notably, the optimized process was validated by real experiments, where the error between calculation and experiment is 4.03 % for deposition rate and 0.49 % for coefficient of variation, respectively. The results highlight the model as reliable and lay the foundation for the CVD growth of the 4H–SiC epitaxial layer.

@en

In this work, 4H-SiC homoepitaxial layers were grown on 4°off-axis substrates at different susceptor rotation speeds by using a hot-wall horizontal CVD reactor. The effect of different susceptor rotation speed on the quality of 4H-SiC epitaxial layers in terms of thickness, thickness uniformity, crystallinity, surface morphology and morphological defects was investigated via Fourier transform infrared spectroscopy (FTIR), high-resolution X-ray diffraction (XRD), atomic force microscopy (AFM), confocal differential interference contrast microscopy (CDIC), ultra-violet photo-luminescence spectroscopy (UV-PL), scanning electron microscopy (SEM), and micro-Raman spectroscopy, respectively. A flow field simulation was performed to explain the impact of susceptor rotation speed on the film deposition. The FTIR results suggested that the susceptor rotation speed could be an important factor to adjust thickness uniformity and deposition rate. The XRD patterns showed that crystallinity was independent of the susceptor rotation speed. The surface morphology can be improved by changing the susceptor rotation speed. According to CDIC scans, the down-fall related defects were reduced through the increase in the susceptor rotation speed. The origin of down-fall related defects was interpreted by Raman spectroscopy and speculative models. To sum up, the susceptor rotation speed is a crucial factor in increasing growth rate and improving uniformity. Also, the faster susceptor rotation speed helps reduce the number of down-fall related defects in the hot-wall CVD reactor.@en

Silicon carbide (SiC) epitaxial process is a key step in the fabrication of power devices, and the temperature field inside the reactor chamber plays an essential role in this process. In this paper, the temperature field in the horizontal chemical-vapor-deposition reactor chamber used for growing homo-epitaxial 4H-SiC material is studied using the finite-element method. A three-dimensional time-dependency model is built for the accuracy of simulation, and the effects of 11 relative coil locations (−50, −28, −18, −10, −4,0,4,10,18,28, and 50 mm) on heating efficiency and temperature uniformity of the substrate are analyzed. Results indicate that the suitable relative location between the center of coil and that of the substrate to achieve optimum temperature uniformity is −4 mm, and 18 mm to obtain the highest heating efficiency. To increase the heating efficiency and temperature uniformity of the substrate, the structure of the reactor was analyzed and optimized. It is observed that both heating efficiency and temperature uniformity can be effectively improved by adding a graphite pillar inside the down susceptor.@en

A fan-out panel-level packaging (FOPLP) with an embedded redistribution layer (RDL) via interconnection reduces the size, thermal resistance, and parasitic inductance of power module packaging. In this study, the effect of the RDL via size on the reliability of a FOPLP SiC MOSFET power module was investigated. To improve the thermal management and thermal cycling reliability of the designed SiC module, genetic algorithm (GA)–assisted optimization methods were proposed to optimize the RDL via size. First, the heat dissipation and the plastic work density of the SiC MOSFET module with various via diameters and depths were simulated using finite element simulations. Next, both the ant colony optimization-backpropagation neural network (ACOBPNN) with finite element simulation and the nondominated sorting genetic algorithm (NSGA-II) with theoretical model were developed to optimize the RDL via size. The results revealed that: (1) smaller via depth and size reduce the heat dissipation and thermal cycling reliability of the RDL via; (2) through both the ACO-BPNN and NSGA-II, the same optimal heat dissipation and plastic work density can be achieved in the designed module. (3) ACO-BPNN with assist of finite element simulation can provide a more effective optimization in complex packaging structure.@en