Along with preserving crystallinity and minimizing FIB-induced damage, controlling lamella thickness is essential for reliable TEM analysis. For most 100 to 200 keV TEM work in materials science, a lamella should be below 150 nm; above this, it is often too thick for routine imaging. For HR TEM/STEM, the target is even tighter, with an ideal thickness of approximately 30 nm to support high-resolution, high-contrast results.
1. In the FIB-SEM, you can estimate lamella thickness by monitoring its transparency using the SE detector at a certain accelerating voltage. When the primary-beam interaction volume becomes larger than the lamella thickness, the lamella begins to appear transparent in the SE image. This threshold can be supported with Monte Carlo simulations (for example, in CASINO); for tungsten at 10 keV, the first signs of transparency (blue lines show scattering of secondary electrons in bulk) typically correspond to ~130 nm thickness. Keep in mind this approach has a tolerance and can be less reliable in multicrystalline materials, where grain-to-grain channeling affects contrast.
2. In a TEM equipped with EELS, lamella thickness can be estimated using the electron mean free path (MFP) of a known reference material within the sample. A common approach is to use a silicon substrate beneath the region of interest, since silicon has a well-characterized MFP. By measuring the relative thickness (t/λ) from the EELS signal and applying the reference MFP, you can convert this directly to nanometers.
For example, the images below show that the EELS spectrum acquired from the silicon substrate region (red) provides a practical thickness reference for the lamella. Using a known silicon mean free path of 150 nm and the measured average t/λ = 0.11, the lamella thickness is estimated at 16.5 nm.
3. In the FIB-SEM, you can estimate lamella thickness by adding X-shaped reference marks on the lamella surface, for example, using SEM Pt deposition, and tracking the spacing as you polish. This method depends only on the size of the marks (which determines your measurement precision) and the skills of the FIB operator. The image below shows dots of SEM Pt deposition arranged in an X shape with a 60° angle, forming two opposing equilateral triangles. With this geometry, the distance between the outer dots directly relates to the distance to the triangle apex (the central point), making thickness progression easy to gauge during polishing. To improve visibility and endpoint control, add additional dots on each side before the central point.
Along the marked line (green), a cross-section should reveal two groups of three peaks, each peak corresponding to the cross-section of a Pt dot. In this example, once three dots are reached, the remaining distance to the central dot is 50 nm. Polishing the lamella to the same three-dot endpoint from both sides therefore yields a final thickness of 100 nm.
Using two X-marks also helps you keep the lamella aligned during polishing, which is especially useful when you need to endpoint a specific feature within the section. As you polish, compare the distances between the dot groups on both sides: they should remain equal. If one side shows a smaller distance, it indicates the polishing line has rotated toward that side, and the milling angle or position should be corrected to restore symmetry.
4. Finally, there is another, destructive way to estimate the lamella thickness. you can prepare a cross-lamella from the original section once the required analysis is complete and the lamella can be sacrificed. This approach enables a direct thickness measurement and is a practical way to confirm estimates by any of three previous methods. In the example shown, the thickness measured from the cross-lamella in TEM correlates well with the value previously derived from the EELS-based estimate(2).
We hope these practical tips help you achieve more consistent lamella thickness and more reliable TEM results. If you would like support optimizing your workflow or discussing a specific sample challenge, please get in touch with the Tescan team. We are ready to help you move from preparation to high-quality data with confidence.
Written by Maksym Klymov
Head of Applications Department, Tescan