Scanning Electron Microscope (SEM) Testing Tutorial Series——Sample Requirements and Preparation——Conventional Sample Preparation

Sample preparation is also crucial for scanning electron microscopy observation. Poorly prepared samples can significantly affect observation results.

Ideally, samples should have the best possible conductivity to avoid charging phenomena that can prevent normal SEM observation. Additionally, samples need good thermal conductivity; otherwise, the temperature at the impact point may rise, causing low-melting-point components in the sample to volatilize, resulting in irradiation damage and affecting accurate morphology observation. For quantitative EDS/WDS/EPMA analysis, the sample surface should also be as smooth as possible.

1. Sampling

For SEM experiments, samples should be as small as possible while still being representative. Especially for non-conductive samples, small samples can improve both conductivity and thermal conductivity. Large samples placed in the sample chamber release more gas, especially in porous materials, affecting vacuum levels and significantly increasing the time required to achieve a vacuum, potentially introducing substantial contamination. Therefore, for porous materials, pre-treatment such as using a hairdryer, infrared lamp baking, or placing in an oven at low temperatures to release gases can reduce the vacuuming time when placed in the SEM.

For film cross-sections, it is advisable to cut, embed, and polish. During embedding, the sample can be split into two halves, with the observed film face inward, then glued and embedded before grinding. This avoids film detachment or the formation of cracks and gaps during polishing due to mechanical property differences between the film and embedding material, as shown in Figure 1. By gluing the film face-to-face, the mechanical properties on both sides are equal, improving this issue.

Figure 1: Damage Caused by Asymmetric Mechanical Properties of Single Film Faces

For softer samples, avoid direct cutting with scissors, which can cause deformation and poor-quality cross-sections. Use sharp blades like surgical blades to make a clean cut, or freeze the sample in liquid nitrogen before breaking it. A small incision can be made before freezing to ensure it breaks along the cut with minimal force. When conditions allow, use cross-section ion beam or FIB polishing.

For powder samples, a small amount should be used to avoid affecting conductivity and stability due to powder overlap. If powder samples are severely agglomerated, they can be mixed with a volatile solvent (e.g., pure water, ethanol, hexane, cyclohexane) to form a suspension. After ultrasonic dispersion, place a small drop on a sample holder or a silicon chip, copper (aluminum) conductive tape. Avoid using carbon conductive tape, as it is not dense and can embed the sample in gaps, affecting observation. After the solvent evaporates, the powder adheres to the substrate by surface adsorption, as shown in Figure 2.

Figure 2: Powder Sample Preparation via Ultrasonic Dispersion

Note the choice of solvent, as it should not affect the sample to be observed, otherwise it may alter the sample's initial morphology and cause image distortion. As shown in Figure 3, polymer sphere samples remain spherical when dispersed in water, but change shape when dispersed in anhydrous ethanol.

Figure 3: Effects of Water (Left) and Ethanol (Right) Dilution on Morphology

2. Cleaning

Ensure samples are fresh and free from contamination, particularly oil. Avoid direct hand contact to prevent grease contamination. Cleanliness applies not only to the sample but also to the sample stage, which should be frequently cleaned with anhydrous ethanol.

3. Sample Mounting

ASamples should be mounted smoothly and securely to minimize contact resistance, improving conductivity and thermal conductivity. For samples with uneven bottom surfaces, silver paste should be used to fill gaps and ensure good contact. For EBSD testing, silver paste is recommended to provide stable electrical connections under tilting and gravitational stress.

For extreme resolution observations, silver paste can further improve conductivity beyond carbon conductive tape.

Mounting should ensure a conductive path between the sample surface and the sample stage. Even highly conductive surfaces will experience charging if not properly connected to the sample stage. This is particularly important for irregular samples, as shown in Figure 4. The left and middle images show poor mounting without conductive paths, while the right image shows a proper conductive path.

Figure 4: Proper (Right) vs. Improper (Left, Middle) Mounting

Even for regular samples like blocks or thin films, improper mounting can occur. Some may think a conductive sample can be directly mounted on conductive tape, as shown in Figure 5 left. However, there may still be no path between the sample surface and the stage, causing charging or image drift under high beam currents. In Figure 5 right, extra conductive tape is folded back onto the sample surface, ensuring a conductive path regardless of the sample's intrinsic conductivity, reducing contact resistance and enhancing conductivity.

For powder samples, use small amounts to avoid affecting conductivity and thermal conductivity. Powder can be lightly sprinkled on double-sided carbon conductive tape on the sample holder, pressed with a flat object like a glass slide or the backing of conductive tape, then blown off with an air blower to remove loose particles, as shown in Figure 5 left. For very small powder samples, carbon conductive tape can be directly used to pick up the powder, as shown in Figure 5 right.

Figure 5: Methods for Mounting Powder Samples

4. Coating

For samples with poor conductivity, coating is usually used. Gold (Au) coating is common, but for higher resolution, platinum (Pt), chromium (Cr), or iridium (Ir) coatings can be used. For strict quantitative EDS analysis, metal coatings are avoided due to their strong X-ray absorption, and carbon coatings are used instead.

Modern coating equipment can accurately control coating thickness, usually applying 5nm coatings to improve conductivity. For special structured samples like sponges or foams, even thick conductive layers may not form continuous paths. Thus, coating thickness is typically limited to 10nm; if conductivity doesn't improve with 10nm coatings, further thickening is ineffective. Gold particles are visible at about 100,000x magnification, platinum at about 200,000x, and chromium or iridium at about 300,000x. After coating, the metal layer replaces the sample for secondary electron emission, enhancing signal strength and contrast for better image quality. As long as the coating doesn't obscure actual sample details, full coating treatment can be used, unless specific uncoated observations are required. For high magnification or precise measurements, coating should be carefully considered, as it can obscure details or cause measurement deviations. As shown in Figure 6, left shows SEM measurements of PS spheres after gold coating, right shows direct TEM measurements. SEM shows about 195nm, while TEM shows about 185nm, indicating gold coating increased the diameter by about 10nm.

Figure 6: Comparison of PS Spheres Observed in SEM with Coating (Left) and Directly in TEM (Right)