Machine-learning-guided exploration of domain walls in the hybrid improper ferroelectric Ca3Ti2O7
Abstract
Ruddlesden-Popper phases are highly tunable and naturally layered structures, in which polarization can arise via a hybrid improper ferroelectric mechanism. This enables a complex domain wall (DW) structure where multiple order parameters, like octahedral rotations, polar distortions and strain, interact. In this work, we explore the rich set of DW structures in prototypical Ca3Ti2O7, mapping out the DWs in the {100}, {110} and {001} pseudo-tetragonal planes using group theory and machine-learned interatomic potentials (MLIPs). The trained potential reproduces the density functional theory (DFT) order parameter and polarization profiles for all wall types and orientations considered. A charge-aware training framework combined with reference Born effective charges reduces the prediction errors for the DW formation energies by 70%, revealing the importance of including long-range electrostatics to model symmetry-broken interfaces. Finally, the MLIP is used to identify minimum energy pathways at the atomic scale with nearly the precision of DFT calculations, revealing a low-energy antipolar configuration for polarization switching.
Related
- Machine-Learning Exploration of Defect Topologies and Thermodynamic Stability in Graphene with Atomic Vacancies
- Towards Extended Active Learning for Modelling Ferroelectric Switching: the Need for 'Gold Standards'
- Revealing tilt-driven structural heterogeneity in hybrid improper ferroelectrics by spatially-resolved crystallography
- Quantum spectral thermodynamics and active learning enable million-scale exploration of high-entropy ceramics
- Reinforcement Learning in High-Energy Physics: A Survey from Accelerator Control to Theory Exploration
BibTeX
@misc{skogvoll2026machinelearningguided,
title = {{Machine-learning-guided exploration of domain walls in the hybrid improper ferroelectric Ca3Ti2O7}},
author = {Ida C. Skogvoll and Erik Fransson and Leo Ö. Westin and Benjamin A. D. Williamson and Nicholas C. Bristowe and Sverre M. Selbach and Paul Erhart},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.35346},
eprint = {2609.35346},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.35346},
}
RIS
TY - UNPB
TI - Machine-learning-guided exploration of domain walls in the hybrid improper ferroelectric Ca3Ti2O7
AU - Ida C. Skogvoll
AU - Erik Fransson
AU - Leo Ö. Westin
AU - Benjamin A. D. Williamson
AU - Nicholas C. Bristowe
AU - Sverre M. Selbach
AU - Paul Erhart
PY - 2026
DA - 2026/09/28
PB - arXiv
DO - 10.48550/arxiv.2609.35346
UR - https://arxiv.org/abs/2609.35346
AB - Ruddlesden-Popper phases are highly tunable and naturally layered structures, in which polarization can arise via a hybrid improper ferroelectric mechanism. This enables a complex domain wall (DW) structure where multiple order parameters, like octahedral rotations, polar distortions and strain, interact. In this work, we explore the rich set of DW structures in prototypical Ca3Ti2O7, mapping out the DWs in the {100}, {110} and {001} pseudo-tetragonal planes using group theory and machine-learned interatomic potentials (MLIPs). The trained potential reproduces the density functional theory (DFT) order parameter and polarization profiles for all wall types and orientations considered. A charge-aware training framework combined with reference Born effective charges reduces the prediction errors for the DW formation energies by 70%, revealing the importance of including long-range electrostatics to model symmetry-broken interfaces. Finally, the MLIP is used to identify minimum energy pathways at the atomic scale with nearly the precision of DFT calculations, revealing a low-energy antipolar configuration for polarization switching.
ER -