Optimizing ToF-SIMS data analysis with survey-then-target workflow

Time-of-flight secondary ion mass spectrometry (ToF-SIMS) is a valuable tool for analyzing surface composition, providing material spectra, hyperspectral maps, and depth profiles. However, the sheer volume of data can make analysis difficult.

Oslinker et al. created a practical guide for optimizing machine learning algorithms specifically to handle ToF-SIMS data.

“ToF-SIMS data is information-rich but highly complex, and it can contain millions of pixels acquired from a single scan. Each pixel is represented by a full mass spectrum, which can contain thousands of peaks,” said author Brian Oslinker. “Datasets are often too large and complex to effectively analyze without target analytical tools.”

The authors focused on the self-organizing map with relational perspective mapping toolbox, a suite of analysis tools developed specifically for ToF-SIMS data. They identified and removed several computational bottlenecks to optimize the toolbox, allowing the software to support a survey-then-target workflow.

With this workflow, the software analyzes a low-resolution scan of the sample that researchers can use to identify zones of interest. These zones can then be rescanned at higher resolution. Crucially, the optimization allows this workflow to be performed while the sample is being analyzed, speeding up the process significantly. This improvement will especially benefit researchers working with large or complex samples.

With further optimization, the team hopes this toolbox could provide real-time analysis of ToF-SIMS data.

“We hope that our work provides a guide for researchers attempting to implement real-time machine learning pipelines,” said Oslinker. “This is critical, as there is a significant shift towards data-driven decision-making and machine learning for instrument control.”

Source: “Considerations for implementing real-time machine learning tools to evaluate ToF-SIMS data,” by Brian Oslinker, Wil Gardner, Sarah Elizabeth Bamford, See Yoong Wong, John Allan Webb, Davide Ballabio, Thomas M. Kohl, and Paul J. Pigram, Journal of Vacuum Science and Technology: A (2026). The article can be accessed at https://doi.org/10.1116/6.0005335 .