🔬 materials science

The role of absorption in three-dimensional electron diffraction dynamical structure refinement

This paper demonstrates that while absorption can be analytically modeled and significantly improves the refinement accuracy of high- $Z$ materials like $\text{CsPbBr}_3$ , it remains negligible for the routine dynamical structure refinement of low- $Z$ materials.

Original authors: Benjamin Colmey, Tiarnan A. S. Doherty, Shreshth A. Malik, Paul A. Midgley

Published 2026-02-10

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Benjamin Colmey, Tiarnan A. S. Doherty, Shreshth A. Malik, Paul A. Midgley

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

Imagine you are trying to figure out the exact shape of a complex, microscopic LEGO castle by looking at the shadows it casts on a wall. In the world of science, researchers use a technique called 3D Electron Diffraction to do exactly this—they shoot electrons at tiny crystals and study how they bounce off to map out where every atom is located.

However, this paper identifies a "glitch" in the way scientists have been doing these calculations: they’ve been ignoring "absorption."

Here is a breakdown of the paper using everyday analogies.

1. The Problem: The "Vanishing" Light

In a perfect world, when you shine a light on an object, the light bounces off predictably. But in the real world, some of that light gets "soaked up" by the object itself. It doesn't just bounce; it gets swallowed.

In electron diffraction, as electrons travel through a crystal, some of them get "lost" because they bump into vibrating atoms or lose energy. Scientists have traditionally been using math that assumes every electron that goes in comes back out (the "Elastic Model"). This is like trying to calculate the shape of a castle by looking at shadows, while assuming the castle is made of perfectly reflective mirrors, even though it’s actually made of slightly fuzzy, light-absorbing wood.

2. The "Zone Axis" Chaos: The Spotlight Effect

The paper points out a specific problem: when the crystal is tilted at certain angles (called Zone Axes), the math goes haywire.

The Analogy: Imagine you are walking through a forest with a flashlight. Most of the time, the light hits the trees and bounces back to you clearly. But if you walk into a very dense, dark thicket where the trees are packed tightly together, your flashlight beam gets swallowed up by the shadows almost immediately.

In the past, when scientists saw these "dark thickets" (high-error data points near zone axes), they would simply throw that data away, thinking their equipment was broken or the crystal was bad. This paper proves that the data isn't "bad"—the math was just missing the "absorption" part of the equation.

3. The Experiment: Heavy vs. Light Materials

To see if this "absorption" actually matters, the researchers tested three different "castles":

CsPbBr3 (The Heavyweight): A material made of heavy atoms (like Lead). This is like a castle made of thick, dark velvet. It absorbs a lot of light.
Quartz (The Middleweight): A moderate material. Like a castle made of stained glass.
Borane (The Lightweight): A material made of very light atoms. Like a castle made of thin lace.

The Result: For the "Lace" (Borane) and the "Glass" (Quartz), ignoring absorption didn't really change much. The math was "close enough." But for the "Velvet" (the heavy Lead material), ignoring absorption caused noticeable errors. By adding the absorption math, they were able to make their structural models much more accurate.

4. The Conclusion: When should you worry?

The researchers aren't saying scientists need to change everything they do. Instead, they are providing a "Rule of Thumb."

The Metaphor: If you are photographing a piece of white paper, you don't need to worry about how much light the paper absorbs; it's negligible. But if you are photographing a black velvet cloak in a dim room, you must account for how much light is being swallowed, or your photo will be wrong.

The Takeaway:

If you are studying light, simple materials, you can keep using the old, easy math.
If you are studying heavy, complex materials (like many modern high-tech semiconductors), you need to include "absorption" in your math, or your "map" of the atoms will be slightly blurry and incorrect.

Technical Summary: The Role of Absorption in 3D Electron Diffraction Dynamical Structure Refinement

Problem Statement

In three-dimensional electron diffraction (3D ED), structural models are refined by iteratively matching simulated diffraction intensities to experimental data. While current state-of-the-art methods utilize dynamical refinement (accounting for multiple scattering via Bloch-wave simulations), they almost exclusively rely on elastic scattering models.

Inelastic scattering processes—specifically thermal-diffuse scattering (TDS), plasmon excitations, and core-loss ionization—are known to attenuate coherent beams and redistribute intensity. Historically, these effects (collectively termed "absorption") have been modeled in Convergent Beam Electron Diffraction (CBED) but have been largely neglected in 3D ED. This omission potentially introduces systematic errors (residuals) into structural refinements, particularly in high-atomic-number ( $Z$ ) materials or near high-symmetry zone axes.

Methodology

The authors employ a multi-faceted approach to quantify the impact of absorption:

Analytical Theory: They derive a two-beam expression for absorbed integrated intensity. This shows that for thicknesses ( $t$ ) much smaller than the extinction distance ( $\xi_g$ ), reflections follow a uniform exponential decay governed by the mean absorptive potential $U'_0$ , supplemented by a reflection-specific "anomalous" term.
Bloch-Wave Simulations: Using a differentiable Bloch-wave framework, the authors performed many-beam simulations with and without absorption. They compared these against two-beam models to observe how absorption affects the redistribution of intensity.
Residual Analysis: They developed a mathematical model to predict how the refinement residual ( $R_1$ ) scales with thickness, proposing that $R_1$ should increase linearly with $t$ due to the mismatch between elastic models and the actual absorptive decay.
Experimental Refinement: To validate the theory, the authors performed full dynamical refinements on three distinct materials: $\text{CsPbBr}_3$ (high- $Z$ ), $\alpha$ -quartz (medium- $Z$ ), and borane (low- $Z$ ).

Key Contributions

Formalism for 3D ED: Established a theoretical link between the absorptive potential and the integrated rocking-curve intensities used in 3D ED.
Identification of the "Zone Axis Problem": Demonstrated that high residuals near zone axes, often dismissed as crystal imperfections or "bad data," are actually a predictable consequence of unmodeled absorption and many-beam coupling.
Quantification of Error: Provided a closed-form estimation of how neglecting absorption causes the refinement residual to grow linearly with specimen thickness.

Results

Material Dependency:
- $\text{CsPbBr}_3$ : Absorption is significant. Including it in the refinement improved the residual ( $R_{obs}$ ) from 6.4% to 5.3%.
- $\alpha$ -quartz and Borane: Absorption effects were negligible, with almost no change in refinement residuals, confirming that the impact scales with atomic number and thickness.
Thickness and Orientation: The residual $R_1$ increases linearly with thickness. The effect is most pronounced near zone axes, where many-beam interactions are strongest and the anomalous absorption term diverges.
Structural Stability: Despite the improvement in fit, the inclusion of absorption did not significantly alter the recovered atomic coordinates or displacement parameters, suggesting that while absorption affects the precision (residuals), it does not inherently bias the structural solution for most routine materials.

Significance

This paper provides a definitive guide for crystallographers regarding the necessity of absorption corrections in 3D ED. The primary conclusion is that absorption can be safely neglected for routine refinement of most inorganic and molecular crystals, provided they are not high- $Z$ materials and the thickness does not approach the extinction distance ( $\xi_g$ ).

However, for high-precision work on heavy-atom materials or when analyzing data near zone axes, the inclusion of an absorptive potential is critical. This work also justifies the retention of "near-zone-axis" data frames in datasets, which were previously discarded due to high residuals, thereby increasing the statistical power of 3D ED refinements.