DIY adapting SEM for low-voltage TEM imaging

Abstract

Electron microscopy is essential for examining materials and biological samples at microscopic levels, providing detailed insights. Achieving high-quality imaging is often challenged by the potential damage high-energy beams can cause to sensitive samples. This study compares scanning electron microscopy (SEM) and transmission electron microscopy (TEM) to evaluate image quality, noise levels, and the ability to preserve delicate specimens. We used a modified SEM system with a transmitted electrons conversion accessory, allowing it to operate like a TEM but at lower voltages, thereby reducing sample damage. Our analysis included quantitative assessments of noise levels and texture characteristics such as entropy, contrast, dissimilarity, homogeneity, energy, and correlation. This comprehensive evaluation directly compared traditional TEM and the adapted SEM system across various images. The results showed that TEM provided images with higher clarity and significantly lower noise levels, reinforcing its status as the preferred method for detailed studies. However, the modified SEM system also produced high-quality images at very low acceleration voltages, which is crucial for imaging samples sensitive to high-energy exposure. The texture metrics analysis highlighted the strengths and limitations of each method, with TEM images exhibiting lower entropy and higher homogeneity, indicating smoother and more uniform textures. This study emphasizes the importance of selecting the appropriate electron microscopy method based on research needs, such as sample sensitivity and required detail level. With its conversion accessory, the modified SEM system is a versatile and valuable tool, offering a practical alternative to TEM for various applications. This research enhances our understanding of the capabilities and limitations of SEM and TEM. It paves the way for further innovations in electron microscopy techniques, improving their applicability for studying sensitive materials. RESEARCH HIGHLIGHTS: Our study introduces a modified SEM adapter enabling TEM-like imaging at reduced voltages, effectively minimizing sample damage without compromising image resolution. Through comparative analysis, we found that images from the modified SEM closely match the quality of traditional TEM, showcasing significantly lower noise levels. This advancement underscores the SEM's enhanced capability for detailed structural analysis of sensitive materials, broadening its utility across materials science and biology.

Keywords: SEM vs. TEM; electron microscopy; image quality; low‐voltage imaging; noise analysis; sample sensitivity.

FIGURE 1

Adapter changes made. Image adapted from Eisaku et al. The dashed lines indicate the modifications we implemented to reduce the height of the adapter, ensuring it does not interfere with the backscattered electron sensor.

FIGURE 2

The 3D sketch illustrates the modifications made to the adapter to reduce its height from 40 to 10 mm, ensuring compatibility with the scanning electron microscopy model (FEI XL‐30) without interfering with the backscattered electron sensor.

FIGURE 3

(a) The green line represents the electron beam passing through the sample on the TEM grid. After the electron beam traverses the sample, it hits the gold surface, emitting secondary electrons. These secondary electrons are then captured by the Everhart–Thornley detector. (b) Actual implementation of the adapter. In the top image, a rendering; in the bottom image, a real photo.

FIGURE 4

In the image's upper right and lower left corners, muscle fibers can be identified, within whose sarcoplasm the contractile elements align to form myofibrils composed of myofilaments arranged in the register. In the center of the image, the pericellular space is visible, where loose connective tissue (endomysium) with sparse collagen fibers organizes. A capillary can be seen within this connective tissue, distinguished by its endothelial lining. Adjacent to the capillary, a cellular structure is present, clearly showing the elongated nucleus of a fibroblast.

FIGURE 5

The pairs of images acquired, the left transmission electron microscopy (TEM) and the right scanning electron microscopy (SEM). Efforts were made to capture the same images at the same magnification to facilitate easy comparison. The left column displays images obtained using TEM, while the right column shows corresponding images taken with the modified SEM.

FIGURE 6

Image processing process in Fiji: Importing, cropping, registering, and recrop. (a) The original image. (b) The image was cropped to remove any elements interfering with registration, creating a virtual stack. (c) The images are rigidly registered to ensure correct orientation and positioning. Finally, the individual cropped frames are exported, resulting in corresponding scanning electron microscopy and transmission electron microscopy images for comparison. (d) SEM images registered and cropped to remove any black spaces. (e) TEM images are similarly registered and cropped.

FIGURE 7

Example of a noise image in white areas. White areas where only the resin is present were sampled to evaluate sensor noise. Each sampled area had a minimum size of 50 × 50 pixels and was taken from the same position in both scanning electron microscopy and transmission electron microscopy images to ensure consistency in the noise level assessment.