PTST: A polar topological structure toolkit and database
Authors:
Guanshihan Du,
Yuanyuan Yao,
Linming Zhou,
Yuhui Huang,
Mohit Tanwani,
He Tian,
Yu Chen,
Kaishi Song,
Juan Li,
Yunjun Gao,
Sujit Das,
Yongjun Wu,
Lu Chen,
Zijian Hong
Abstract:
Ferroelectric oxide superlattices with complex topological structures such as vortices, skyrmions, and flux closure domains have garnered significant attention due to their fascinating properties and potential applications. However, progress in this field is often impeded by challenges such as limited data-sharing mechanisms, redundant data generation efforts, high barriers between simulations and…
▽ More
Ferroelectric oxide superlattices with complex topological structures such as vortices, skyrmions, and flux closure domains have garnered significant attention due to their fascinating properties and potential applications. However, progress in this field is often impeded by challenges such as limited data-sharing mechanisms, redundant data generation efforts, high barriers between simulations and experiments, and the underutilization of existing datasets. To address these challenges, we have created the Polar Topological Structure Toolbox and Database(PTST). This community driven repository compiles both standard datasets from high throughput phase field simulations and user submitted nonstandard datasets. The PTST utilizes a Global Local Transformer (GL Transformer) to classify polarization states by dividing each sample into spatial sub blocks and extracting hierarchical features, resulting in ten distinct topological categories. Through the PTST web interface, users can easily retrieve polarization data based on specific parameters or by matching experimental images. Additionally, a Binary Phase Diagram Generator allows users to create strain and electric field phase diagrams within seconds. By providing ready-to-use configurations and integrated machine-learning workflows, PTST significantly reduces computational load, streamlines reproducible research, and promotes deeper insights into ferroelectric topological transitions.
△ Less
Submitted 6 April, 2025;
originally announced April 2025.
