Papers by Author: Jian Wei Wan

Abstract: Improvements in the quality and consistency of 4H-SiC epitaxy wafers are now starting to enable growth of commercial SiC power device applications in areas such as inverters for photo-voltaic systems and power supplies. Recent work has achieved very low epitaxy surface roughness and very low BPD (Basal plane dislocation) in the on 4 degree off-axis substrates. In this paper, we report characterization of the very low BPD epitaxy wafers and a newly observed triangular defect.

Abstract: Effective recombination lifetimes of 4H-SiC PiN epitaxy wafers are measured by -PCD (microwave photoconductive decay) system at wafer level. Lifetimes measured in presence and absence of the p+ layer show lower lifetime values with p+ layer present. This is attributed to high recombination rate at p+/n- interface. Lifetimes at various buffer thicknesses show lower values at the buffer layer of about 50 m due to high interface recombination rate resulting from rougher surface of the buffer layer. Lifetimes of PiN wafers from interrupted and continuous p+/n- growth are very comparable.

Abstract: Excess carrier lifetimes in 4H SiC epitaxial wafers were characterized by microwave photoconductive decay (o/PCD). The measured decay compromised of surface and bulk recombination curves have fast and slow components. Measured lifetimes are not changed with various surface passivation techniques. High resolution lifetime maps show good correlation with stress birefringence images and lower lifetime around extended material defects like grainboundaries, defect clusters, edge defects and polytype switching bands. Chlorosilane based CVD epiwafers show higher bulk lifetime values than standard silane based CVD materials due to less bulk lifetime defect density.

Abstract: A SiC epitaxy process based on chlorosilane/propane chemistry has been successfully transferred from a single-wafer R&D system to a multi-wafer CVD reactor. The optimized process results in very smooth epi surface (RMS~0.24nm) and minimum surface pits (less than 0.5/cm2). Both n-type and p-type doping in a wide range are demonstrated using nitrogen and aluminum, respectively. The high performance benchmarks for thickness uniformity (intra-wafer variation <1% and inter-wafer variation <1%) and doping uniformity (intra-wafer variation <6% and inter-wafer variation <3%) are achieved on 5 x 3-inch wafers. The carrier lifetime in these epilayers measured by μ-PCD is over 5 μs, the longest value reported so far for SiC epitaxial wafers.

Abstract: Gas phase etching of 4H SiC n+ substrates was performed utilizing chlorine containing etch chemistries in a hot wall CVD system. Carbon and silicon vapor were added to explore selective etching reactions on the wafer surface. The impact of the etch on the bare wafer surface as a function of temperature and etch chemistry is investigated. Selection of the etch chemistry and temperature are critical to ensure a smooth etched surface on which to begin epitaxial deposition. Etching also influences defect propagation from the substrate into the epitaxial layer. The results show etch chemistry reactions will influence the conversion of micropipes in the epi buffer layer.

Abstract: Micropipes are considered to be a major device killer in SiC wafers. Developing a method to count and map micropipes efficiently and accurately has been a challenging task to date. In this work, a new method based on KOH etching and full wafer, high resolution digital imaging is developed to map and count micropipes in both conductive and semi-insulating SiC wafers. This method is also compared with a non-destructive method based on laser light scattering and a good agreement between the two methods is demonstrated.

Abstract: Micropipe density (MPD) is a crucial parameter for silicon carbide (SiC) substrates that determines the quality, stability and yield of the semiconductor devices built on these substrates. The importance of MPD is underscored by the fact that all existing specifications for 6H- and 4H-SiC substrates set upper limits for it. Several methods for measuring the MPD are known, however, their reliability and applicability to various types of substrates (e.g. semiinsulating, conducting, etc.) has not been systematically studied. The subject of this paper is a comparative study of various techniques used for the MPD measurement accompanied by statistical analysis of the results. The study was initiated by several organizations working in the immediate field of silicon carbide or in closely related fields and included SiC substrate manufacturers, substrate consumers, equipment manufacturers and universities. The study represented a round robin experiment in which MPD was measured on thirty SiC wafers of various pedigrees. The values of MPD have been determined using both destructive and non-destructive techniques. The repeatability of each technique is analyzed and compared with that of other techniques.

Abstract: Epitaxial growth of SiC films was performed on 4H SiC n+ substrates utilizing a chlorosilane/propane chemistry in both single wafer and batch CVD systems. Variations of the chlorosilane flow under fixed conditions of gas composition, temperature and pressure resulted in growth rates between 4 to 20 μm/hr. Fixing the chlorosilane flow rate to achieve a growth rate of approximately 4 μm/hr, the effects of temperature, pressure and gas composition on background dopant incorporation, epitaxial layer uniformity and epitaxial defect generation were investigated. Intentional n and p-type doping has been demonstrated over the carrier range 1×1018-1×1020/cm3. This paper presents the first reported of use of chlorosilane precursors to grow high quality undoped, n and p doped SiC epilayers.