Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-ready design of realistic 2D materials and interfaces with Mat3ra-2D

Vsevolod Biryukov, Kamal Choudhary, Timur Bazhirov

Abstract

Artificial intelligence (AI) and machine learning (ML) models in materials science are predominantly trained on ideal bulk crystals, limiting their transferability to real-world applications where surfaces, interfaces, and defects dominate. We present Mat3ra-2D, an open-source framework for the rapid design of realistic two-dimensional materials and related structures, including slabs and heterogeneous interfaces, with support for disorder and defect-driven complexity. The approach combines: (1) well-defined standards for storing and exchanging materials data with a modular implementation of core concepts and (2) transformation workflows expressed as configuration-builder pipelines that preserve provenance and metadata. We implement typical structure generation tasks, such as constructing orientation-specific slabs or strain-matching interfaces, in reusable Jupyter notebooks that serve as both interactive documentation and templates for reproducible runs. To lower the barrier to adoption, we design the examples to run in any web browser and demonstrate how to incorporate these developments into a web application. Mat3ra-2D enables systematic creation and organization of realistic 2D- and interface-aware datasets for AI/ML-ready applications.

AI-ready design of realistic 2D materials and interfaces with Mat3ra-2D

Abstract

Artificial intelligence (AI) and machine learning (ML) models in materials science are predominantly trained on ideal bulk crystals, limiting their transferability to real-world applications where surfaces, interfaces, and defects dominate. We present Mat3ra-2D, an open-source framework for the rapid design of realistic two-dimensional materials and related structures, including slabs and heterogeneous interfaces, with support for disorder and defect-driven complexity. The approach combines: (1) well-defined standards for storing and exchanging materials data with a modular implementation of core concepts and (2) transformation workflows expressed as configuration-builder pipelines that preserve provenance and metadata. We implement typical structure generation tasks, such as constructing orientation-specific slabs or strain-matching interfaces, in reusable Jupyter notebooks that serve as both interactive documentation and templates for reproducible runs. To lower the barrier to adoption, we design the examples to run in any web browser and demonstrate how to incorporate these developments into a web application. Mat3ra-2D enables systematic creation and organization of realistic 2D- and interface-aware datasets for AI/ML-ready applications.

Paper Structure

This paper contains 24 sections, 7 figures.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Components
  5. Standards
  6. Commonly-used data
  7. OOD abstraction layer
  8. Materials design implementation
  9. Notebook examples and browser access
  10. Open-source ecosystem approach
  11. Results
  12. Build procedures
  13. Example 1: constructing slabs with specific terminations
  14. Example 2: constructing interfaces without strain matching
  15. Example 3: strain-matched interfaces
  16. ...and 9 more sections

Figures (7)

  • Figure 1: Interface transformation stages: define, refine, and build for selecting and constructing a strain-matched interface with recorded metadata.
  • Figure 2: SrTiO$_3$(110) slab construction with termination control: (a) bulk crystal from the reference data repository, (b) and (c) alternative surface terminations obtained with the same Miller indices, layer count, and vacuum by changing the termination formula.
  • Figure 3: Ge(001)/Si(001) interface without strain matching: (a) and (b) film and substrate slabs defined independently with Miller indices (001), layer count, and vacuum; (c) combined interface after applying a target interfacial gap and optional lateral shift.
  • Figure 4: Graphene/Ni(001) interface construction workflow: (a) Starting graphene monolayer film, (b) Ni(001) slab substrate, (c) ZSL analyzer output showing strain--size trade-off for candidate configurations (each point represents a commensurate match ranked by strain percentage and interface area), (d) Selected interface structure with commensurate matching applied. The workflow follows define--refine--build: define film and substrate slabs, refine by enumerating and ranking matches, then build the selected configuration with recorded metadata.
  • Figure 5: The JupyterLite environment hosting the transformation notebooks. The left panel shows the file system browser with the notebooks listed, the right panel shows the editor with the Introduction notebook open. The Introduction notebook contains the index and allows to navigate to the other ones. The notebooks are organized by the M-CODEbiryukov2026mcode tag for quick navigation.
  • ...and 2 more figures