🔬 materials science

Orientation refinement for nanobeam precession electron diffraction patterns using spot intensity profiles

This paper presents a novel method for refining crystallographic orientations in nanobeam precession electron diffraction by analyzing internal spot intensity profiles, achieving significantly higher angular precision (0.044°–0.066°) and correcting systematic errors compared to standard automated crystal orientation mapping.

Original authors: Arthur Després, Muriel Vérion, Edgar F. Rauch

Published 2026-10-05

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Arthur Després, Muriel Vérion, Edgar F. Rauch

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

In the world of materials science, understanding the tiny architecture of matter is essential for everything from building faster microchips to creating stronger alloys. To see this hidden structure, scientists use a powerful tool called a transmission electron microscope, which fires a beam of electrons through a sample so thin that the particles pass right through it. As the electrons exit the other side, they scatter and form a pattern of bright spots on a detector, much like light passing through a sieve. These spots, known as diffraction patterns, act as a unique fingerprint for the crystal arrangement inside the material. By analyzing the position and brightness of these spots, researchers can determine the exact orientation of the crystal grains, a process vital for mapping how materials behave under stress or heat. However, when scientists try to measure these orientations with extreme precision, especially in very small areas, the standard methods often hit a wall of inaccuracy, leaving them unable to see the subtle twists and turns that occur at the atomic level.

A team of researchers at the University of Grenoble Alpes has developed a new way to sharpen this view, pushing the limits of how precisely they can measure the direction of crystals. They focused on a technique called nanobeam precession electron diffraction, where a tiny, focused beam of electrons is tilted in a circle as it scans the sample. This tilting, or precession, helps create clearer patterns, but the standard way of reading them still misses fine details. The team realized that the bright spots in these patterns are not just uniform blobs of light; they often contain subtle, internal variations in brightness, with some parts of a spot glowing more intensely than others. These variations are not random noise but are actually caused by the specific way the electrons interact with the crystal as the beam tilts. By building a computer model that simulates these internal brightness patterns and matching them to the real images, the researchers created a method to refine the orientation measurements far beyond what was previously possible.

To test their idea, the scientists took a single crystal of silicon, a material commonly used in electronics, and tilted it step-by-step around a specific axis. They captured hundreds of diffraction patterns, some with long exposure times to ensure every detail was clear, and others with very short exposure times to mimic the faster speeds used in modern scanning. When they applied their new refinement method to these images, the results were striking. The standard method, which is widely used in the field, could only determine the tilt angles with a precision ranging from 0.25 to 0.33 degrees. In contrast, the new method, which accounts for the internal intensity profiles of the spots, narrowed that uncertainty down to a range between 0.044 and 0.054 degrees. This improvement is significant because it allows scientists to detect much smaller misalignments in the crystal structure, such as those found near defects or boundaries between different grains. Even when the researchers used shorter exposure times, which introduce a natural graininess to the data, the new method remained robust, maintaining a precision between 0.054 and 0.066 degrees, still far superior to the standard approach.

One of the most important discoveries was that the new method fixes a specific flaw in the standard technique that occurs when the crystal is aligned with a major axis, known as a zone axis. In these specific alignments, the standard method often fails to find the correct orientation, creating gaps in the data where no measurement is possible. The new refinement method successfully bridges these gaps, providing a continuous and accurate map of the crystal's orientation. The researchers also found that their approach works well even when the data is noisy, suggesting it can be applied to rapid scanning techniques used to map large areas of material in a short time. This capability opens the door to more detailed studies of how materials deform and change, potentially leading to better designs for everything from microelectronics to structural metals. By turning the subtle, internal variations of light in a diffraction spot into a precise measuring tool, the team has provided a clearer window into the atomic world, proving that even the smallest details in a pattern can hold the key to the most accurate measurements.

Technical Summary: Orientation Refinement for Nanobeam Precession Electron Diffraction Patterns Using Spot Intensity Profiles

Problem Statement
Automated Crystallographic Orientation Mapping (ACOM) is the standard method for orientation mapping in transmission electron microscopy (TEM), particularly when using nanobeam precession electron diffraction (4D-SPED) to achieve high spatial resolution. However, standard ACOM processing, which correlates experimental patterns with kinematical templates, typically achieves an angular precision of 0.5° (or 0.12° with advanced algorithms). This level of precision is insufficient for characterizing small misorientations at defects such as dislocations, subgrain boundaries, and inclusions. Furthermore, existing methods that utilize more physically accurate intensity descriptions (e.g., Palatinus et al.) are optimized for parallel beam conditions and treat diffraction spots as punctual or Gaussian, failing to account for the finite size and internal intensity variations inherent to nanobeam precession patterns. Current strain mapping techniques often attempt to cancel these internal intensity variations rather than exploiting them, missing an opportunity to improve measurement precision.

Methodology
The authors propose a refinement method that explicitly models and utilizes the internal variations of intensity within diffraction spots to refine crystallographic orientations. The methodology proceeds as follows:

Intensity Modeling: The method distinguishes between the homogeneous intensity distribution (related to the convergence angle) and the heterogeneous distribution caused by variations in excitation error due to local convergence angles and precession.
- The homogeneous distribution is modeled using a generalized Lorentz function to simulate the spot radius and decay.
- The heterogeneous distribution is calculated by integrating the intensity over the precession phase, assuming two-beam conditions. This results in a "camel plot" intensity profile where intensity peaks at specific excitation errors ( $\pm g\beta$ ).
- The final simulated spot intensity ( $I_{het}$ ) is the element-wise product of the precession intensity and the homogeneous intensity profile.
Refinement Procedure:
- An initial orientation is obtained via standard ACOM.
- Out-of-plane refinement: The orientation is rotated out-of-plane to minimize the mean error between experimental and theoretical spot intensities.
- In-plane refinement: The in-plane rotation is refined by calculating the deformation gradient tensor ( $F$ ) that transforms theoretical spot positions to experimental positions. The in-plane rotation is derived from the antisymmetric part of this tensor.
- The agreement between theoretical and experimental spots is quantified using phase cross-correlation in Fourier space, which provides position shifts and a normalized root mean square error invariant to intensity scaling.

Key Results
The method was validated using a tilt series of a silicon single crystal, where the sample was tilted in 1° steps around the goniometer axis.

Angular Precision: With long exposure times (10s effective, minimizing Poisson noise), the refinement method achieved a precision of 0.044° to 0.054°. This represents a significant improvement over the standard ACOM method, which yielded a precision of 0.25° to 0.33° for the same dataset.
Noise Robustness: When simulating shorter exposure times (10ms) to introduce Poisson noise, the precision degraded only marginally to a range of 0.054° to 0.066°, demonstrating the robustness of the phase cross-correlation approach.
Systematic Error Correction: The standard ACOM method exhibited a systematic error near zone axes, resulting in gaps in indexed orientations and oscillations in misorientation angles. The proposed refinement method successfully corrected this, providing continuous and accurate orientation data even near zone axes.
Misorientation Axis Accuracy: The point-to-point misorientation axis calculated using the refined orientations showed a root mean square deviation of only 1.7° from the goniometer tilt axis, compared to 12.1° for standard ACOM.

Significance and Claims
The paper claims that internal intensity variations in nanobeam precession diffraction spots, previously considered noise or ignored, can be accurately modeled and utilized to significantly enhance angular resolution. The proposed method:

Achieves angular precision (0.044°–0.054°) that surpasses current state-of-the-art ACOM processing and approaches the precision of parallel beam methods, but is specifically adapted for nanobeam conditions.
Corrects systematic indexing errors near zone axes that plague standard ACOM.
Is robust to Poisson noise, suggesting it is applicable to high-speed 4D-SPED orientation mapping (acquisition rates ~10ms) without significant loss of precision.
Offers potential for improving strain mapping by utilizing internal intensity variations to locate diffraction spots with higher accuracy, paralleling strategies used in non-precession datasets but leveraging the natural diffraction conditions of precession.

The authors conclude that this refinement technique is a viable path toward high-resolution orientation and strain mapping using large 4D-SPED datasets.

More like this