From Manual Control to Automated Insights: Versatile FIB-SEM Workflows for Semiconductor Research

As semiconductor devices become increasingly complex and three-dimensional, researchers need more than one way to access and understand the structures hidden beneath the surface.

Some investigations begin with hands-on FIB-SEM operation, where researchers need the freedom to explore an unfamiliar device, locate a region of interest, and optimize the preparation strategy. Once that workflow is established, however, repeatedly performing the same operations manually may not be the best use of expert time.

The Thermo Scientific Scios 3 FIB-SEM bridges these two needs, providing the flexibility of manual operation together with automated workflows that can improve efficiency, repeatability, and accuracy.

The result? Less time spent on tedious operations and more time focused on semiconductor insights.

From manual exploration to automated FIB-SEM cross-sections

Cross-sectioning is fundamental to semiconductor research, providing access to buried device structures, interfaces, and defects that cannot be fully understood from surface imaging alone.

Manual operation gives researchers precise control when developing a new workflow. But once the preparation method is established, the Scios 3 FIB-SEM can automate cross-sectioning across multiple sites, reducing repetitive operator interaction while applying the workflow consistently.

This creates a natural progression:

Explore manually → Establish workflow → Automate repetitive preparation → Analyze results

Automation does not replace researcher expertise. It allows that expertise to be applied where it creates greater value.

SEM images of automated cross-sections on a MOSFET device using the Thermo Scientific Scios 3 FIB-SEM and Thermo Scientific Auto Cross Section (AXS) Software.
Figure 1. Automated multi-site cross-section preparation using Thermo Scientific Auto Cross Section (AXS) Software. (a) SEM image of a 5 × 5 array of cross-sectioned sites prepared on a MOSFET device using a Scios 3 FIB-SEM operated with AXS Software.(b) Enlarged view of a representative site from the array, showing the prepared cross-section and exposed device region. (c) Higher-magnification SEM image of the cross-section, revealing detailed device structures and interconnect features for subsequent characterization.

Three TEM lamella preparation approaches, one FIB-SEM platform

Advanced semiconductor characterization often requires site-specific TEM samples, but one preparation geometry does not fit every analytical question.

The Scios 3 FIB-SEM supports automated top-down, inverted, and planar TEM lamella preparation, giving researchers the flexibility to select the approach according to the device structure and information required:

  • Top-down preparation provides a conventional route to site-specific cross-sectional TEM specimens.
  • Inverted preparation offers an alternative geometry for challenging structures and can help reduce curtaining and seek for thinner lamella during preparation.
  • Planar preparation provides a complementary lateral perspective when the information of interest cannot be fully captured through a conventional cross-section.

The key is versatility: The analytical question determines the preparation methodology, not the limitations of a single workflow.

Once the appropriate methodology has been established, automation can reduce repetitive preparation steps while supporting more consistent execution from sample to sample.

Planar, top-down, and inverted TEM lamella preparation methods for semiconductor characterization.
Figure 2. Three TEM lamella preparation methodologies for semiconductor characterization: planar (left), top-down (right), and inverted (bottom). Each approach provides a different sample orientation and thinning direction, enabling researchers to select the preparation geometry best suited to the region of interest and analytical objective.

STEM imaging: See more before moving to TEM

Preparing a TEM lamella does not necessarily mark the end of the FIB-SEM workflow. With 30 kV STEM imaging, the Scios 3 FIB-SEM enables researchers to acquire transmission images from electron-transparent semiconductor specimens directly within the system. This provides an opportunity to inspect the prepared lamella, verify the region of interest, and evaluate key device structures before proceeding to further analysis.

Complementary STEM imaging modes provide different contrast mechanisms for visualizing device structures and material features, giving researchers additional information to assess sample quality and determine the most appropriate next step. Rather than simply preparing and transferring a lamella, researchers can make a more informed decision based on what they observe directly within the FIB-SEM.

The workflow therefore evolves from prepare and transfer to a more insight-driven process:

Prepare → Inspect → Decide → Analyze

30 kV STEM images of a semiconductor device with the Scios 3 FIB-SEM using bright-field, dark-field, and HAADF imaging modes.
Figure 3. 30 kV STEM imaging of a semiconductor device using the Scios 3 FIB-SEM. Representative STEM images acquired from an electron-transparent semiconductor specimen using (a) bright-field (BF), (b) dark-field (DF-S), and (c) high-angle annular dark-field (HAADF) imaging modes.

From 2D slices to 3D understanding

Modern semiconductor devices are inherently three-dimensional. A single cross-section may therefore reveal only part of the story.

With serial FIB-SEM sectioning, sequential images can be collected through a volume of interest. These individual slices can then be assembled and visualized using Thermo Scientific Avizo Software to reveal the structure in three dimensions.

Instead of asking only, “What does the device look like here?” researchers can investigate, “How does this structure evolve through the volume?”

That is the transition from image acquisition to deeper structural insight.

Spend time on insights, not repetition

The versatility of the Scios 3 FIB-SEM is not about performing every task automatically.

It is about giving semiconductor researchers the choice to use manual control when exploration and judgment are required and automation when the workflow becomes established and repeatable.

Across automated cross-sectioning, top-down, inverted, and planar TEM lamella preparation, 30 kV STEM imaging, and serial sectioning for 3D visualization, researchers can select and combine workflows according to the characterization challenge.

The progression is simple:

Explore → Automate → Prepare → Image → Reconstruct → Understand

By moving repetitive operations toward automation, the Scios 3 FIB-SEM can help improve efficiency, repeatability, and accuracy—while allowing researchers to devote more of their expertise to what matters most: Understanding the device, interpreting the data, and finding the next insight.

FIB-SEM serial imaging and 3D reconstruction of semiconductor structures using Avizo Software.
Figure 4. FIB-SEM image acquisition and 3D reconstruction workflow. Serial cross-sectional images are acquired by sequential FIB milling and SEM imaging. The resulting image stack is imported into Avizo Software for image processing, alignment, and segmentation of the features of interest. The segmented slices are subsequently reconstructed into a three-dimensional (3D) volume, enabling visualization of the internal microstructure and quantitative 3D metrology measurements.

The Scios 3 FIB-SEM provides semiconductor researchers with a versatile platform that combines the flexibility of manual operation with automation to support a wide range of characterization challenges. Researchers can explore and optimize new workflows manually, then automate repetitive processes to improve efficiency, repeatability, and accuracy.

From automated, multi-site cross-sectioning and multiple TEM lamella preparation approaches to in-system 30 kV STEM imaging and serial sectioning for 3D reconstruction with Avizo Software, the Scios 3 FIB-SEM connects sample preparation, imaging, and analysis within a flexible workflow.

This progression of Explore → Automate → Prepare → Image → Reconstruct → Understand helps researchers spend less time on repetitive operations and more time interpreting data, understanding complex semiconductor devices, and uncovering the next insight.

Melissa Sundar

Leave a Reply

Your email address will not be published. Required fields are marked *