From phase analysis to phase tracking: a robust approach to atomic-scale strain mapping
This paper introduces Geometric Phase Tracking Analysis (GPTA), a robust method that dynamically tracks Bragg reflections in reciprocal space to overcome the limitations of standard Geometric Phase Analysis (GPA), thereby significantly reducing failure rates and improving accuracy in atomic-scale strain mapping across diverse imaging conditions and materials.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Inside the tiniest structures of matter, where atoms arrange themselves into perfect grids, the most important property is often how much those grids are stretched or squeezed. This stretching, known as strain, acts like a hidden dial that controls how a material conducts electricity, how it responds to magnets, and how strong it is. To see this stretching, scientists use powerful microscopes that can image individual atoms. However, measuring the exact amount of stretch is difficult because the atoms are so small and the images are often noisy. For more than twenty years, the standard tool for this job has been a method called geometric phase analysis. It works by looking at the patterns of light and dark in a microscope image to calculate how much the atomic grid has shifted from its perfect shape. While this method has been incredibly useful, it has a critical weakness: it relies on a fixed set of rules that assume the atomic grid is mostly uniform. When the grid is twisted, bent, or when the image is taken with very little light to avoid damaging the sample, the standard tool often fails, producing maps that are full of errors or missing data entirely.
A team of researchers has now introduced a new approach called geometric phase tracking analysis that solves these problems by changing how the measurement is taken. Instead of using a single, unchanging reference point for the entire image, this new method constantly updates its reference as it moves across the picture. Imagine trying to measure the slope of a winding mountain road using a level that only works if the road is straight; the old method is like that level, struggling whenever the road curves. The new method is like a surveyor who carries the level with them, adjusting it at every step to match the local curve of the road. By tracking the position of the atomic patterns in real time, the new technique can follow the grid even when it is rotating, when different parts of the sample are oriented differently, or when the image is too noisy for the old method to work.
The researchers tested this new approach on more than 13,000 simulated images and real experimental photos of platinum crystals. In these tests, the old method failed to produce a correct map in 27% of the cases, particularly when the material was under high strain or when the image was taken with a low dose of electrons to protect the sample. The new tracking method reduced this failure rate to just 11.5%. In situations where the old method broke down completely, the new approach was able to measure the strain with up to 335 times better accuracy. It successfully mapped the strain across boundaries where two different crystals met, identified the exact location of defects in the atomic structure, and worked clearly even when the image was extremely noisy. Crucially, it achieved this without needing to take more pictures or use more powerful equipment, simply by processing the existing data in a smarter way.
The core of this improvement lies in how the method handles the information contained in the image. The old technique looked at the entire image at once, using one average pattern as a guide. If the actual pattern in a specific part of the image moved away from that average, the measurement became unreliable. The new method, however, starts with a guess and then refines it. It looks at the image, calculates where the atoms are actually located, and then uses that new information to update its guide for the next step. This process repeats a few times, allowing the method to lock onto the local structure of the material at every single point. Because it knows exactly where the pattern should be, it can use a much tighter filter to block out noise, something the old method could not do without losing the signal. This ability to adapt to the specimen means that scientists can now study delicate materials that would have been destroyed by the high doses of electrons required for older techniques, opening the door to understanding strain in beam-sensitive biological and chemical samples.
The researchers also found that this new approach is better at handling complex defects, such as the boundaries between different crystal grains or the cores of dislocations where the atomic order is broken. In the old method, if one part of the image was distorted or if a specific pattern disappeared due to the thickness of the sample, the entire measurement could become corrupted. The new method treats each part of the image independently and then combines the results, weighing them by how confident it is in the data. If one part of the image is unclear, the method relies more heavily on the clearer parts to fill in the gaps. This makes the final map robust and reliable, even when the sample is imperfect or the imaging conditions are challenging. The study confirms that by moving from a static analysis to a dynamic tracking process, scientists can now see the atomic-scale strain in materials with a clarity and reliability that was previously impossible, providing a clearer view of the forces that shape the properties of the materials around us.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.