Practical Electron Microscopy and Database

When conducting electron microscopy analysis of SiC materials, several specific considerations are essential due to the material's unique properties. Sample preparation is particularly challenging, as SiC is a very hard and brittle material. Techniques like focused ion beam (FIB) milling are often necessary to produce thin samples for transmission electron microscopy (TEM) while minimizing damage. For SEM analysis, commercially available SiC samples, such as 4H-SiC MOSFETs, are typically bonded to aluminum specimen holders, approximately 12.5 mm in diameter, and coated with a thin conductive layer—typically 20 nm of gold [1] —to enhance image quality and prevent charging under the electron beam. Final thinning with low-energy ion polishing is recommended to avoid mechanical damage that standard polishing may induce, particularly for TEM sample preparation where thickness should ideally be below 100 nm to ensure sufficient electron transmission.

Imaging conditions for SiC require careful optimization, as this material exhibits some resistance to electron beam damage, though prolonged exposure to high-energy beams can lead to localized heating and potential carbon contamination. Determining the optimal beam voltage is critical for achieving high-quality SEM images. For instance, cross-sectional samples can be cut at 90° via FIB-SEM at a high voltage (e.g., 30 kV), and then imaged at an angle, such as 54° to the beam, using an in-column secondary electron (SE) detector. By varying the electron beam voltage in small increments (500 eV steps between 1.0 and 5.0 kV) with a moderate beam current (e.g., 2 nA) and a dwell time per pixel of around 12.8 µs, the most suitable imaging conditions can be determined as shown in Figure 657. Through visual examination and analysis of the grey value profiles across critical transitions, such as the N+ emitter to P- body, the optimal imaging voltage can be identified. For example, while the sharpest transition was observed between 2.5 and 3.0 kV, a voltage such as 2.0 kV offers the best contrast for specific features, such as implant patterns, making it the preferred choice for tomography runs.

SiC also presents unique challenges in crystallographic analysis due to its various polytypes, such as 4H-SiC and 6H-SiC, each with distinct stacking sequences. Proper interpretation of electron diffraction patterns is essential for accurate polytype identification. High-resolution TEM (HRTEM) is particularly useful for identifying stacking faults and dislocations, which are common in SiC devices exposed to high stresses. Orienting the crystal correctly relative to the electron beam is crucial for capturing accurate images of these structural features.

For elemental and chemical analysis, EDS and EELS are invaluable techniques. While EDS can be used to map silicon and carbon distributions, its sensitivity to carbon is limited due to the element’s low atomic number. EELS complements EDS by providing detailed information on carbon bonding states, enabling differentiation between SiC and contaminants like graphite. EELS is also beneficial for in-situ analyses, especially when paired with a heating stage to assess SiC’s thermal stability. With its high melting point, SiC remains stable at elevated temperatures, but phase transformations can occur, warranting careful observation in thermal studies.

Surface and interface analyses of SiC also benefit from techniques like atomic force microscopy (AFM), which complements electron microscopy by examining properties such as roughness, elasticity, and electrical characteristics. Due to its hardness, SiC is suitable for AFM, providing additional surface information that electron microscopy may not capture. However, surface oxidation, particularly during beam exposure, can affect results, suggesting a need for controlled environments or protective coatings during analysis.

[1] Heiko Stegmann, Greg Johnson, David Taraci, Andreas Rummel, 3D Visualization and Characterization of SiC MOSFET Junctions Using EBIC and FIB-SEM Tomography, access in 2023.