Large-area Mapping Workflows in Scanning Electron Microscopy (SEM)

Large-area imaging in scanning electron microscopy (SEM) is a technique for exploring expansive regions of a sample at high resolution, making it invaluable in fields such as materials science, electronics, and life sciences. However, capturing high-resolution images across large sample areas presents unique challenges, including time constraints and data management. This is where image stitching algorithms come into play, enabling researchers to seamlessly combine multiple high-magnification images into a single, comprehensive view (often referred to as a large-area map). In this blog, we’ll dive into how large-area mapping works, explore various image stitching software and workflows, and discuss how these techniques can improve efficiency and enhance imaging results.

How does large-area imaging work?

Sometimes, it is necessary to capture a high-resolution image of an area that exceeds the maximum field of view provided by the SEM optics. This can be achieved using a technique called image stitching, which involves acquiring multiple smaller images (also referred to as “tiles”) and combining them into a single, high-resolution large-area map (Figure 1).

Image stitching algorithms allow users to systematically collect and merge these images, creating a detailed view of large sample areas ¹. The performance of this process depends on several adjustable parameters within the algorithm, which can affect image resolution, acquisition time, and overall data quality. Understanding these parameters is essential for optimizing the image stitching process to suit your specific application.

Automated Image Mapping with Phenom SEM

Phenom Desktop SEMs come with the capability to capture large-area SEM images with the Automated Image Mapping (AIM) plugin. This feature, accessible within the Phenom User Interface (UI), enables image stitching using data from either the backscattered electron detector (BSD) or the secondary electron detector (SED).

To specify the imaging area, users can draw a shape directly on the low-magnification optical image displayed through the navigation camera view (Figure 2). The AIM settings panel lets users customize basic image stitching parameters, including the tile field of view, the amount of frame averaging (to enhance the signal-to-noise ratio), option to autofocus on every tile, and the choice of detector for image acquisition.

Before initiating the AIM data collection run, users are provided with the estimated duration, the number of tiles, and the pixel size (which corresponds to image resolution). This preview allows for easy optimization of settings to balance acquisition speed and image quality.

Large-area mapping with MAPS 3 for Phenom SEM

Some image stitching parameters cannot be adjusted through the AIM plugin. For a more advanced and customizable workflow, MAPS 3 is the ideal solution for large-area mapping with Phenom Desktop SEMs (Figure 3). MAPS 3 provides enhanced options for image stitching parameters, layer overlays, and multi-modal data visualization, significantly reducing manual effort and enhancing data analysis abilities.² In addition to SEM imaging, MAPS 3 can generate large-area elemental maps using energy-dispersive X-ray spectroscopy (EDS). MAPS 3 automates the entire data collection process and includes a built-in visualization suite, enabling users to overlay and align different datasets for comprehensive, multi-modal analysis.

How is MAPS 3 Different from AIM?

MAPS 3 offers 11 additional user-defined parameters for large-area mapping, providing greater flexibility and control. These include tile resolution, contrast, brightness, tile overlap, beam intensity, vacuum level, and tile acquisition sequence. Unlike AIM, MAPS 3 allows users to customize automated functions like focus, contrast, and brightness, applying them to the first tile, every tile, or a selected set. This flexibility ensures high image quality while minimizing unnecessary acquisition time.

Maintaining focus across large-area SEM maps is a common challenge, and MAPS 3 addresses this with customizable “focus strategies.” Users can fix the working distance, interpolate focal lengths across tiles, or apply autofocus at defined intervals. Fixed focus works best for flat samples, while interpolated focus is ideal for planar topographies. For non-planar samples, periodic autofocus ensures accurate imaging. Additionally, MAPS 3 allows individual tiles to be recollected and updated, eliminating the need to recapture the entire map—a significant improvement over AIM.

Different Types of Imaging Workflows

When conducting large-area imaging in SEM, the choice of workflow plays a critical role in achieving the desired balance between resolution, coverage, and data complexity. This section explores three common workflows for large-area imaging: (1) large-area overviews, ideal for quick assessments of extensive sample regions; (2) multiple tile sets, designed for higher-resolution imaging of specific areas of interest or when analyzing multiple samples; and (3) multi-modal datasets, which integrate complementary imaging and analytical techniques to deliver richer, more comprehensive insights. Each workflow offers unique advantages and customization options, allowing users to tailor their imaging strategy to specific research needs.

Workflow 1: Large-area overviews

Large-area overviews involve collecting a high-resolution SEM image over an expansive sample area, sometimes covering the entire sample (Figure 4). This workflow is ideal for applications where a broad perspective is needed to identify regions of interest or assess overall sample uniformity. By providing detailed context, large-area overviews allow researchers to efficiently pinpoint areas for further high-magnification analysis, saving time and resources in the imaging process.

Why is this workflow important?

• Analyzing unknown samples
• Ease in locating regions of interest for further analysis
• Reduces time and effort required to obtain high resolution images
• Zoom in from the full macroscopic view to see nanoscale details

Workflow 2: Automated image acquisition of multiple samples

Automated image acquisition over multiple samples ensures consistent, high-throughput data collection while minimizing user effort and variability. Large-area imaging complements this by capturing comprehensive views of entire sample regions, enabling efficient analysis and comparison across multiple specimens. Figure 5 shows several samples loaded onto a SEM sample stub that were imaged within minutes using MAPS 3.

Why is this workflow important?

• Automate processes/tasks to acquire data more efficiently
• Ensure reproducibility and reduce user error
• Use case for quality control
• Overnight analysis

Workflow 3: Multi-modal datasets

Multi-modal datasets combine multiple imaging and analytical techniques to provide a more comprehensive understanding of a sample. These types of datasets can include SEM images, showing sample topography and microstructure, as well as EDS maps that show the elemental distribution (Figure 6). MAPS 3 for Phenom Desktop SEMs facilitates automated acquisition of SEM and EDS data over expansive sample areas and provides an environment for visualization with its offline data viewer.

Why is this workflow important?

• Correlate multiple modes of data (e.g., elemental composition with surface morphology)
• Enhanced insights and provide meaningful interpretations
• Reduce time required to align and visualize multiple datasets
• Locate area(s) with elements of interest

Key Takeaways

Large-area mapping algorithms enable the creation of high-resolution SEM images across large sample areas with minimal effort. These composite images are made by stitching together individual high-resolution SEM tiles, allowing users to analyze both macroscopic structures and nanoscale details within the same dataset. For researchers interested in large-area mapping, the AIM plugin for Phenom Desktop SEM offers basic mapping capabilities. For more advanced functionality, MAPS 3 provides greater control over the stitching process, along with the ability to automate multi-sample imaging and acquire multi-modal datasets, making it a powerful tool for more complex workflows. Table 1 summarizes the key differences between the two large-area mapping solutions that were discussed throughout this article.

References

1. D. Burian, “Automated stitching for scanning electron microscopy images of integrated circuits,” Diplom-Engineer, TU Wien, 2022. ↩︎
2. J. O. Buckman, “Use of automated image acquisition and stitching in scanning electron microscopy: Imaging of large scale areas of materials at high resolution,” Microscopy and Analysis, vol. 28, no. 1, pp. 13-15, 2014. ↩︎