Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing

arXiv:2601.09627 · cond-mat.mtrl-sci, cond-mat.mes-hall · Submitted 2026-01-14 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: Today's paper: "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing".

Mira: This scientific paper presents a novel Electron Backscatter Diffraction (EBSD) reindexing technique designed to accurately map local ferroelectric polarization directions in polycrystalline materials, overcoming significant challenges posed by crystallographic pseudosymmetries.

Kai: First, who's behind it and why it matters.

Title and authors: Kai: So we're diving into this paper today, "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing." It sounds like they're tackling a really tough problem in materials science where you need to map out local polarization directions in these switchable materials.

Mira: It does sound complex. The title hints that the core challenge they are addressing is how to get accurate maps of polarization when the underlying crystal structure has these certain symmetries, which I think is where the real theoretical meat of the issue lies.

Lev: From my side, I'm curious about what kind of data processing we're talking about. If we can actually get stable local domain information from EBSD in these materials, that opens up a whole new avenue for testing error-correction codes on fabricated ferroelectric devices.

Kai: Exactly. And the summary of the paper points out that they are focusing on barium titanate and lead zirconium titanate polycrystals because those materials show these issues clearly. They’re aiming to overcome what they call crystallographic pseudosymmetries that mess up standard EBSD indexing methods, which is a big hurdle for getting this data in the first place.

Mira: That's what I'm focused on—the physics of why those patterns are so similar; it’s because the underlying symmetry allows for very close polarization variants to look almost identical under diffraction conditions, which makes distinguishing them extremely hard.

Lev: If the signal is this weak due to dynamic diffraction rather than true Bragg diffraction, then any error correction scheme we build on top of that map will be incredibly sensitive to noise in the input data.

Kai: And the paper addresses that head-on by proposing several new ways to process and index the data. The key takeaway from their summary is that they’re developing a novel technique specifically designed to handle these tricky pseudosymmetric materials where conventional methods just fall apart.

Mira: I'm interested in how they are handling the noise reduction aspect, because when you deal with near-identical patterns, you need something much more sophisticated than simple pattern matching to avoid introducing artifacts.

Lev: Robustness is crucial; if the indexing method introduces a systematic bias based on which variant it picks, that translates directly into correlated errors in any subsequent simulation we run on real hardware.

Kai: They introduce a few specific methodological improvements. For instance, they use Bayesian optimization with Gaussian processes to automatically find the best pattern processing parameters by maximizing a normalized cross-correlation score between simulated and experimental patterns.

Mira: That sounds like an intelligent way to handle the parameter tuning; using that kind of optimization routine should help them systematically select the right filters without relying on manual guesswork.

Lev: I wonder how much computational overhead that adds, though if it saves us from having to discard half the data due to poor processing, it might actually be worth it for high-fidelity simulations.

Title and authors: Kai: They also have this "Pseudo-Symmetry-Sensitive Neighbor Pattern Averaging," or PSS-NPA scheme, which uses a selection criterion based on a detectable jump in NCC score to decide which neighboring patterns contribute to the average, instead of just using the correlation as a simple weight.

Mira: That sounds like they are trying to isolate the signal from true domain boundaries while intelligently ignoring those similar patterns that cause noise around interfaces.

Lev: If you can define a cutoff based on a statistical jump in the NCC score, it suggests they’re building robustness right into the averaging process, which is exactly what we need for reliable results.

Kai: Then there's this new confidence index called the Weighted Correlation Metric or WCC, which they say significantly reduces values compared to just using the standard NCC metric when trying to select a domain variant out of six possibilities.

Mira: I've seen how correlation metrics can be overly optimistic in these situations, so if the WCC truly reduces those values and makes it easier to distinguish between polar domains, that’s a big step forward for classification accuracy.

Lev: A metric that extends the range of distinction is very valuable because it means the uncertainty around our model parameters shrinks significantly when we try to identify a specific domain orientation.

Kai: Building on that, they then use the Pseudo-Symmetry Confidence Index or CIPS, which leverages the WCC and computes a penalty term based on the mean absolute difference between experimental and simulated curves for all six possible PS variants.

Mira: So, if CIPS is used to pick the orientation with the highest score, it means they are explicitly quantifying how well each variant matches what they expect to see, which is much more rigorous than just a raw correlation number.

Lev: Quantifying that penalty term gives us a measurable quantity we can use to assess the fidelity of our input data before we even start running complex error correction algorithms.

Kai: Finally, they tackle the geometry problem with this "DIC-based global geometry refinement" algorithm, which computes a displacement field from averaging vectors across patterns to simultaneously solve for all six sample-detector geometry parameters.

Mira: That addresses a major systematic error source; if the orientation identification is shifted because of a slight misalignment in the detector, then we have to fix that globally, not just locally on each pattern.

Lev: Solving for linear superposition of parameter changes based on displacement sensitivity sounds like a very powerful way to decouple orientation dependence from global calibration issues.

Kai: So, to wrap up the paper's summary and improvements: they’ve moved toward using Bayesian optimization for processing, PSS-NPA for noise reduction, WCC and CIPS for variant selection, and DIC refinement for geometry. It sounds like a comprehensive pipeline designed specifically to overcome the limitations of conventional EBSD indexing on ferroelectrics.

Title and authors: Mira: I think what they've done is provide a much more rigorous statistical framework than just relying on standard pattern matching when dealing with materials like PZT where the patterns are so close together.

Lev: And for us, it means we have a clearer path to acquiring reliable microstructure maps, which is the necessary first step before we can even think about running complex error correction schemes on real hardware.

Kai: This work by Griesbach et al., "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing," provides a new toolset for characterizing the local domain microstructure in ferroelectrics, which is vital for things like high-density memory storage and low-power transistors.

Mira: I think the implication here is that we can finally move past the inability to map polarization direction in three dimensions using conventional experimental approaches when dealing with these materials.

Lev: And if we can get that three dee mapping reliably, it fundamentally changes how we approach designing quantum error correction codes for these devices because we'll have a much better understanding of the physical switching dynamics.

Kai: So, as they conclude this paper, they show superior distinguishability in their validation studies when comparing their CIPS values to the best correlation values obtained using just NCC.

Mira: That comparison shows that CIPS is significantly better than the NCC metric, which really backs up their claim about its superior ability to differentiate between close polar domain variants.

Lev: If they can achieve one to two orders of magnitude higher confidence scores compared to the standard metric, that gives us a quantifiable measure of how much noise we can expect in our final microstructural data.

Kai: This paper by Griesbach et al., "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing," successfully addresses the challenge of mapping polarization in these materials, and it lays out a very detailed methodology for how to achieve high accuracy.

Mira: The overall impact seems to be providing a practical, data-driven solution for a problem that has been intractable with previous methods when dealing with polycrystalline ferroelectric samples.

Lev: For the future work, I expect they’ll move toward applying this pipeline in more complex, real-world samples and testing its ability to handle even more extreme distributions of grain orientations.

Kai: That seems like the natural next step; moving from single crystals to highly disordered polycrystals is where these kinds of reindexing techniques really prove their worth in practical applications.

Mira: And I'm looking forward to seeing how they integrate this into larger automated characterization workflows, which will be crucial for high-throughput analysis.

Lev: If this technique proves robust across different ferroelectric families, it could become a standard preprocessing step for any material science study involving domain mapping in these systems.

Kai: We’ve covered the core of "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing," and it really highlights how targeted methodology can unlock information currently hidden by crystallographic complexities.

The paper's summary: Kai: So, to recap, these authors have developed a new pipeline for EBSD reindexing that specifically targets those tough materials like BTO and PZT where the crystal symmetry makes it nearly impossible to tell different polarization directions apart using standard methods.

Mira: Exactly, and what's really important is how they tackle that fundamental issue with pseudosymmetry; they’re not just tweaking existing software, but building a whole new statistical framework around pattern matching.

Lev: From my perspective as an error-correction researcher, if the input data can be processed with this level of confidence, it means our subsequent simulations aren't starting from fundamentally biased maps caused by poor initial data quality.

Kai: It’s about moving from guesswork to a method that quantifies uncertainty using metrics like the CIPS index, which gives a measurable score for how confident the AI is in its choice of domain variant.

Mira: That quantification is what really sets this paper apart; they show that their new WCC and CIPS metrics are substantially better at separating those closely related polarization patterns than just using the standard NCC metric.

Lev: If we can get that kind of quantifiable certainty, it gives us a much better baseline for designing quantum error-correction codes because we know the physical state we're simulating is derived from high-fidelity measurements.

Kai: And they don't stop there; they’ve included a global geometry refinement step using displacement fields to fix sample-detector issues, which is huge because those systematic errors often get baked into orientation maps.

Mira: That decoupling of orientation from global geometry is critical because it means the local polarization map isn't just a reflection of the crystal structure, but also accurately reflects the actual spatial arrangement on the chip or in the bulk material.

Lev: If we can reliably map those domains in three dimensions with this level of precision, it really changes how we model switching dynamics under external fields, which is something we need for robust quantum devices.

Kai: So what they’ve built here is a comprehensive toolset that handles processing optimization, noise reduction through PSS-NPA, and geometry calibration all in one pipeline tailored for ferroelectrics.

Mira: It’s a practical solution to a problem that has been intractable because the underlying physics—the pseudosymmetry—was being ignored by previous indexing algorithms.

Lev: This approach suggests that even when dealing with materials where the patterns are almost identical, we can still extract meaningful microstructural information necessary for our error-correction models.

Kai: And this work sets a high bar for how we should be processing complex experimental data in condensed matter physics involving these kinds of materials.

The paper's improvements: Kai: So, we've seen how they’ve created a new indexing pipeline for EBSD data that handles those tricky pseudosymmetric materials, and now we’re looking at what they suggest for improving that process.

Mira: They propose several methodological tweaks to make their original approach even more robust, particularly focusing on automated parameter selection and refined noise reduction techniques.

Lev: If the authors are suggesting ways to automate the pattern processing using Bayesian optimization, that could mean we can build an AI system that doesn't rely on manual tuning anymore; that’s a big step toward scalable characterization.

Kai: Right, so they are suggesting this AI system can automatically find the best filters—like dynamic background subtraction or FFT settings—to maximize the signal contrast before even looking at the patterns.

Mira: They also introduce this Pseudo-Symmetry-Sensitive Neighbor Pattern Averaging, or PSS-NPA, which is a smarter way to reduce noise by only averaging neighbors that show a statistically significant jump in their correlation scores.

Lev: That sounds like they're building an adaptive filter into the noise reduction step itself, which would make the resulting microstructural map much cleaner for our error-correction simulations.

Kai: And on top of that, they are proposing this Weighted Correlation Metric and the Pseudo-Symmetry Confidence Index to give us a much more reliable way to pick the correct domain variant than just using raw correlation numbers.

Mira: The CIPS metric is particularly interesting because it calculates a penalty based on the mean absolute difference between experimental and simulated patterns for all six variants, giving us a direct measure of how well each variant actually matches reality.

Lev: That provides a much stronger statistical foundation for our input data; if we can quantify the mismatch explicitly, we have more control over the uncertainty in our simulations.

Kai: Finally, they suggest using the DIC-based global geometry refinement to simultaneously solve for all those six sample-detector parameters, which fixes systematic shifts across the entire map at once.

Mira: That addresses a major source of error—the geometry calibration—by using displacement fields to find a consistent shift across patterns rather than treating each pattern in isolation.

Lev: If we can ensure the geometric calibration is this robust, it makes running large-scale simulations on real hardware much more predictable and reduces the need for extensive pre-calibration steps.

Kai: The overall implication here is that they’re moving toward an automated, self-optimizing pipeline where the AI handles the heavy lifting of tuning parameters and selecting variants with high statistical confidence.

Mira: This suggests that in materials science, we can expect to see more sophisticated machine learning methods integrated directly into data processing workflows to handle complex symmetries.

Lev: It means we can expect higher fidelity inputs for quantum simulations, which is precisely what’s needed when trying to understand the physical mechanisms driving decoherence in real devices.

Kai: This work by Griesbach et al., "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing," shows that targeted methodological improvements can make high-quality microstructural data accessible even in materials where traditional indexing fails.

Conclusion: Kai: To wrap up, we’ve covered how Griesbach et al.'s paper on "Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing" presents a robust new pipeline for extracting local domain information from complex ferroelectrics like BTO and PZT.

Mira: It really highlights how crucial it is to develop specialized tools when standard techniques fail due to crystallographic complexities, showing that targeted statistical improvements can bridge that gap between raw data and physical reality.

Lev: From my side, it confirms that getting high-fidelity input data for quantum simulations of ferroelectric switching dynamics is achievable if we use these sophisticated processing methods.

Kai: We’ve talked about the new metrics like CIPS and how they improve upon the traditional NCC metric for variant selection, which really shows us a more rigorous way to quantify our certainty in the results.

Mira: And that quantification is vital because it means we can better assess where our theoretical assumptions are most likely to break down based on experimental noise levels.

Lev: If we can reliably map those domains, it gives us a clearer picture of the local polarization landscape, which is exactly what we need to build more accurate models for error-correction codes in these systems.

Kai: So, the implication is that this technique could become a standard preprocessing step for any advanced AI system trying to analyze microstructure from EBSD data in these challenging materials.

Mira: That’s right, and it pushes us toward a future where AI tools are not just interpreting raw data but actively optimizing the physical measurement process itself.

Lev: I think this focus on robust indexing is going to be a major step for our field because it solves a fundamental data acquisition problem that plagues many of our experimental setups.

Kai: Indeed, and we’re really excited about what this means for building more reliable quantum hardware prototypes based on these materials.

Mira: It’s encouraging to see such a detailed statistical framework applied to something as fundamentally material-dependent as ferroelectric domain structure.

Lev: Next time, we want to look at how this improved mapping data translates directly into designing specific error-correction protocols for memory devices.

Kai: We are looking forward to that discussion and seeing how these new maps translate into actual device performance metrics.

Mechanics & Materials Laboratory, Department of Mechanical and Process Engineering, ETH Zürich · NEAT Lab, Department of Materials, ETH Zürich

cond-mat.mtrl-sci, cond-mat.mes-hall

Submitted: 2026-01-14

Updated: 2026-01-14

DOI: 10.1016/j.actamat.2026.122386

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 83/100

The gist: This scientific paper presents a novel Electron Backscatter Diffraction (EBSD) reindexing technique designed to accurately map local ferroelectric polarization directions in polycrystalline

Key concepts

Pseudosymmetry (PS): This occurs when different crystal orientations or domain variants have Kikuchi diffraction patterns that are very similar, making it hard for standard EBSD pattern matching to distinguish between them. The paper shows how this similarity can be extremely high in materials like BTO and PZT.
Normalized Cross-Correlation (NCC): NCC is a metric used in EBSD pattern matching to compare an experimental diffraction pattern against simulated patterns of different domain variants. While useful, it often fails when patterns are very similar, leading to poor discrimination between close PS variants.
Pseudosymmetry Confidence Index (CIPS): A new index that leverages the Weighted Cross-Correlation (WCC) metric to assess how well an experimental pattern matches all possible domain variants. It is superior to NCC because it provides wider margins, making it much better at reliably selecting the correct polarization direction.
DIC-based Global Geometry Refinement: A method that uses a displacement field calculated across many patterns to find global mismatches in sample-detector geometry. This decouples orientation effects from geometry errors, ensuring accurate pattern matching even when dealing with complex PS variations.

Terminology

Calibration — 26× in this paper · explained in 7 episodes
Bayesian Optimization — 11× in this paper · explained in 3 episodes
Domain Walls — 7× in this paper · explained in 2 episodes
Piezoresponse Force Microscopy — 6× in this paper
Accuracy — 5× in this paper · explained in 1 episode
Fidelity — 5× in this paper · explained in 2 episodes
Aperture — 2× in this paper · explained in 1 episode
Core — 2× in this paper · explained in 2 episodes
Robustness — 2× in this paper · explained in 5 episodes
Adam — 1× in this paper · explained in 2 episodes
Computational overhead — 1× in this paper · explained in 2 episodes
Decoupling — 1× in this paper · explained in 2 episodes
Fast-Fourier Transform — 1× in this paper
Initial Data — 1× in this paper · explained in 2 episodes
Neighbor Pattern Averaging and Re-Indexing — 1× in this paper
Non-Local Pattern Averaging Reindexing — 1× in this paper
Systematic Bias — 1× in this paper · explained in 3 episodes

Summary

This scientific paper presents a novel Electron Backscatter Diffraction (EBSD) reindexing technique designed to accurately map local ferroelectric polarization directions in polycrystalline materials, overcoming significant challenges posed by crystallographic pseudosymmetries. This method is critical because understanding the local domain microstructure—the nucleation and evolution of domains—is essential for advancing key technologies such as high-density memory storage, low-power transistors, and high-speed fiber optic communication. By successfully distinguishing between close polarization variants in materials like barium titanate (BTO) and lead zirconium titanate (PZT), the research provides a tool to obtain spatially resolved microstructural information currently unattainable with conventional methods.

Overcoming Challenges of EBSD on Ferroelectrics

The paper first establishes the limitations of state-of-the-art EBSD data processing when applied to ferroelectrics, identifying several key hurdles. These challenges include:

Inadvertent domain evolution caused by the electron beam due to charge build-up in insulating ferroelectrics.
The difficulty in identifying polar domain orientations using traditional Hough-based indexing because minute differences between Kikuchi patterns of polar domains (rotated by 180°) are often from dynamic diffraction rather than geometrically defined Bragg diffraction.
The failure of pattern matching techniques to distinguish between close pseudosymmetry (PS) variants, as materials like PZT and BTO exhibit greater similarities between their Kikuchi patterns compared to single crystals like lithium niobate.

Improved Pattern Pre-processing and Indexing

To address these limitations, the authors developed several improved pre-processing and indexing methodologies tailored for pseudosymmetric (PS) materials:

(1) Automatic Pattern Processing Optimization:

The authors introduced a method using Bayesian optimization with Gaussian processes to automatically determine the optimal pattern processing parameters. This process aims to maximize the normalized cross correlation (NCC) between a matched simulated and experimental pattern, selecting from eight individual parameters (e.g., dynamic background subtraction, adaptive histogram equalization, FFT filters).

(2) PSS-NPA Scheme:

The Pseudo-Symmetry-Sensitive Neighbor Pattern Averaging (PSS-NPA) scheme was developed to reduce noise while preserving domain division near interfaces of abrupt crystallographic change. Unlike simpler methods that use the normalized cross-correlation as a weight, PSS-NPA uses a selection criterion based on a detectable jump in NCC score to define a cutoff, ensuring that only patterns similar to the central pattern contribute to the averaging.

Novel Confidence Index for Variant Selection

To select the correct domain variant out of six possibilities after orientation refinement, the authors introduced a new metric:

(1) Weighted Correlation Metric (WCC):

The WCC metric is defined as:

(2)

W A B =     − −  - – - –╝

The WCC is superior to the NCC metric because it significantly reduces the values compared to NCC, extending the range and making it easier to distinguish between polar domains.

(2) Pseudo-Symmetry Confidence Index (CIPS):

The CIPS metric leverages the WCC metric. It assesses the match between an experimental pattern and simulated patterns of all PS variants by computing a penalty term, denoted as:

(7)

CI c PS = −, where  is the mean absolute difference between the experimental and simulated curves. The optimized variant orientation with the highest CIPS is taken as the correct one.

DIC-Based Global Geometry Refinement

A final crucial step involves accurately determining the sample-detector geometry, which is essential for pattern matching. The authors developed a DIC-based global geometry refinement algorithm to decouple orientation dependence from global sample-detector geometry calibration:

A displacement field is computed by averaging displacement vectors across all patterns selected from a subset of the map.
This field represents the consistent shifts between experimental and simulated patterns corresponding to a global mismatch in geometry parameters.
The algorithm then uses displacement sensitivity to changes in each geometry parameter (e.g., pcx, pcy, pcz) to solve for the linear superposition of parameter changes that maximizes reduction in this average displacement field, iteratively refining the geometry until convergence is achieved.

Case Studies and Validation

The developed method was validated on two challenging systems:

For a single-crystal BTO sample, six orientation refinement passes were completed using Kikuchipy's function, followed by the application of the CIPS metric to select the correct variant for each pixel. The results showed that the CIPS values are one to two orders of magnitude higher than the best CVM values when using the NCC metric, demonstrating superior distinguishability.

Improvements for AI systems

As a fastidious researcher, I have analyzed this paper, Ferroelectric polarization mapping through pseudosymmetry-sensitive EBSD reindexing, and identified several high-impact areas where its novel methodology—specifically the development of a new pattern-matching technique and confidence index—can be directly integrated to enhance AI systems.

Here are the specific improvements for AI systems:

)1. Improvement in Materials Science/Data Interpretation (AI System Goal: Automated Microstructure Analysis)

The core improvement is the creation of a highly robust, automated pipeline for extracting local ferroelectric domain orientation and polarization direction from Electron Backscatter Diffraction (EBSD) data, especially in materials exhibiting pseudosymmetry (PS).

The AI system can be upgraded to natively support the new reindexing scheme. This moves beyond traditional Hough indexing or standard dictionary/spherical pattern matching, which fail when Kikuchi patterns of different domain variants are nearly identical (a major failure mode for BTO and PZT).
The system will utilize the newly developed PS confidence index (CIPS) as its primary decision-making metric rather than relying solely on Normalized Cross-Correlation (NCC).
This allows the AI to accurately distinguish between closely related domain variants (e.g., 180° vs. 90° domains in BTO), which is currently unattainable with other methods.

)2. Improvement in Image/Pattern Pre-processing (AI System Goal: Enhanced Feature Extraction)

The system will incorporate the automated, data-driven optimization of pattern processing steps.

The AI can be equipped with a Bayesian optimization routine using Gaussian Processes to automatically select the optimal set of filters (DBS, AHE, FFT filters) and their parameters based on maximizing the NCC score between experimental and simulated patterns.
This removes reliance on guess work by human operators when tuning processing steps, ensuring that noise is minimized and diffraction signal contrast is maximized for PS materials.

)3. Improvement in Noise Reduction (AI System Goal: High-Fidelity Data Recovery)

The system will integrate the advanced noise reduction technique specifically tailored for ferroelectrics.

The AI can employ the Pseudo-Symmetry-Sensitive Neighbor Pattern Averaging (PSS-NPA) scheme instead of simpler methods like NLPAR. This allows the system to selectively average neighboring patterns based on a statistical jump in NCC scores, effectively ignoring similar patterns that might otherwise introduce noise or lead to misidentification near domain walls or PS interfaces.

)4. Improvement in Geometric Calibration (AI System Goal: Robust Spatial Mapping)

The system will implement the computationally intensive but highly accurate DIC-based global geometry refinement algorithm for EBSD maps.

Instead of performing independent, pointwise optimization of pattern centers, the AI will use this global method to simultaneously refine all six sample-detector geometry parameters (including tilt and center coordinates) by analyzing consistent displacement fields across multiple patterns.
This decouples orientation changes from global geometry errors, leading to significantly more accurate spatial mapping and correcting systematic shifts in domain orientation identification that plague other methods.

)What the Improved AI System Can Do (Specific Capabilities):

A robot or autonomous inspection system could perform high-throughput, automated characterization of ferroelectric samples (like BTO single crystals or PZT polycrystals).
It can generate a full 3D map of local polarization directions and crystal orientations with high confidence, even in complex polycrystalline samples where domain variants are extremely similar.
The system will provide quantitative metrics (CIPS and CVM) to assess the certainty of its own results, allowing it to flag regions where the identification is ambiguous or highly reliable.
It can serve as a diagnostic tool for material science research by identifying specific microstructural features—such as domain walls, grain boundary continuity, and incompatible domain orientations—that are critical for understanding ferroelectric switching behavior under applied fields.

Abstract

Ferroelectric materials exhibit a switchable, spontaneous polarization at the unit cell level--an attractive property utilized in many emerging technologies including, among others, high-density memory storage, low-power transistors, and high-speed fiber optic communication. Understanding the local polarization switching behavior, through domain nucleation and evolution, is critical to advancing these technologies and requires characterization of the local domain microstructure. However, in application-relevant polycrystalline materials exhibiting a distribution of grain orientations, a direct mapping of the polarization direction in three dimensions has remained inaccessible using conventional experimental approaches. Here, taking barium titanate single crystals and lead zirconium titanate polycrystals as our bulk model systems, we map the local polarization directions using a new electron backscatter diffraction indexing technique based on simulated pattern-matching. Through improved pre-processing techniques (including optimized pattern processing, a new pseudosymmetry-sensitive neighbor pattern averaging method, and DIC-based global sample-detector geometry calibration) and a new pseudosymmetry confidence index (which considers not only pattern similarity but pattern dissimilarity trends with other domain variant patterns), we successfully distinguish between the six polarization directions, despite the challengingly small unit cell aspect ratio of the selected materials. The methods developed in this work are not only applicable to ferroelectrics but any material which exhibits close crystallographic pseudosymmetries--extending the current capabilities of EBSD.

Sources

Resolving pseudosymmetry in tetragonal ZrO2 using EBSD with a modified dictionary indexing approach