arXiv Science⌕ Search

arXiv subjects

Fintan G. Hardy

Publications and source records attributed to Fintan G. Hardy.

2 recordsLinked to original sources

Quantitatively Probing the Tilt Angle Failures in 4D-STEM Polarisation Analysis

Four-dimensional scanning transmission electron microscopy (4D-STEM) is increasingly used to map polarisation in ferroic textures such as domain walls, vortices, merons and skyrmions. These textures are accompanied by local strain, lattice rotation and misorientation, which also modulate the diffraction features used to infer polar order. Even small deviations from the zone axis can therefore produce contrast that is indistinguishable from a polarisation signal. Here we quantify the tilt sensitivity of two polarisation-sensitive signals, diffraction disc intensity and Kikuchi band intensity, in three ferroelectrics with distinct polar distortions: PbTiO3, BaTiO3 and Cu3B7O13Cl boracite. We find that tilts of only 0.01-0.05 degrees reverse most Friedel-pair disc asymmetries, and even the most robust disc pairs reverse by 0.04 degrees in BaTiO3, 0.14 degrees in PbTiO3 and 0.18 degrees in boracite, producing contrast equal and opposite to that of the polarisation. Kikuchi band contrast is the only signal that reports absolute polarity without knowledge of the specimen thickness, but it does so only up to 0.09 degrees in PbTiO3, 0.015 degrees in BaTiO3 and 0.02 degrees in boracite. From these results we define tilt tolerances for each signal and material, and propose practical criteria for distinguishing polarisation contrast from misorientation. The analysis also applies to scanning electron microscopy-based diffraction, including electron backscatter diffraction.

cond-mat.mtrl-sci↗

Polarization Domain Mapping From 4D-STEM Using Deep Learning

Polarization in ferroelectric domains arises from atomic-scale structural variations that govern macroscopic functionalities. The interfaces between these domains known as domain walls host distinct physical responses, making their identification and control critical. Four dimensional scanning transmission electron microscopy (4DSTEM) enables simultaneous acquisition of real and reciprocal-space information at the atomic scale, offering a powerful platform for domain mapping. However, conventional analyses rely on computationally intensive processing and manual interpretation, which are time consuming and prone to misalignment and diffraction artefacts. Here, we present a convolutional neural network that, with minimal training, classifies polarization directions from diffraction data and segments domains in real space. We further introduce an adaptive sampling strategy that prioritizes images from domain wall regions, reducing the number of training images required while improving accuracy and interpretability. We demonstrate this approach for domain mapping in ferroelectric boracite, Cu3B7O13Cl.

cond-mat.mtrl-sci↗