Computational studies of new Dzyaloshinsky-Moriya interactions

Projektinformationen

New-DMI

ID Finanzhilfevereinbarung: 101152613

DOI

10.3030/101152613

Projekt abgeschlossen

EK-Unterschriftsdatum 7 März 2024

Startdatum 1 April 2024

Enddatum 31 März 2026

Finanziert unter

• Marie Skłodowska-Curie Actions (MSCA)

Gesamtkosten

Keine Daten

EU-Beitrag

€ 173 847,36

Koordiniert durch

MAX-PLANCK-GESELLSCHAFT ZUR FORDERUNG DER WISSENSCHAFTEN EV
Germany

Periodic Reporting for period 1 - New-DMI (Computational studies of new Dzyaloshinsky-Moriya interactions)

Berichtszeitraum: 2024-04-01 bis 2026-03-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Chiral magnets host strikingly ordered, swirling magnetic textures - such as skyrmions - that are of intense scientific interest and hold promise for future low-energy data storage and spintronic devices. These textures arise from a delicate competition between magnetic energy terms, the most important being the Dzyaloshinskii-Moriya interaction (DMI), an antisymmetric exchange interaction that emerges in materials lacking inversion symmetry. The form the DMI takes depends on the crystal symmetry of the material, and each symmetry class can stabilise different, and sometimes unexpected, magnetic configurations.

Although the DMI is central to this field, widely used open-source micromagnetic simulation tools have historically implemented only a small number of DMI forms, covering a limited set of crystal symmetries and so restricting the range of materials and phenomena that researchers can model. The New-DMI project addresses this gap by deriving and implementing the DMI terms for the remaining non-centrosymmetric crystallographic point groups, integrating them into established open-source micromagnetic calculators, and combining these capabilities with machine-learning methods to automatically explore and classify the resulting magnetic phases. The overall objective is to broaden the materials space that can be studied computationally, accelerate the discovery of novel chiral magnetic textures, and deliver these capabilities as openly available, reusable software - while training the fellow as an independent, open-science-minded researcher within the European Research Area.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

The project extended established open-source micromagnetic software with new energy terms describing the Dzyaloshinskii–Moriya interaction (DMI). Working within OOMMF, accessed through the Python-based Ubermag framework, the implementation now covers all crystallographic point groups — including symmetry classes not previously available in mainstream open-source micromagnetic tools — substantially broadening the range of chiral magnetic materials that can be simulated. These terms were released as open-source software (BSD-3-Clause), with source code on GitHub and installation via pip and conda, so the wider research community can reuse them directly.

To make sense of the large and complex space of magnetic configurations these interactions produce, machine-learning workflows were developed that combine an auto-encoder neural network with unsupervised clustering to identify and classify distinct magnetic phases automatically. Using these tools a wide range of magnetic structures was explored and several novel textures identified, including a detailed study of conical-helix textures in thin films hosting different kinds of DMI.

The fellowship also extended the mag2exp package for simulating "virtual experiments", adding techniques such as ferromagnetic resonance, so that simulated states can be compared directly with experimental measurements and used to predict the experimental signatures of novel structures.

Beyond the core software, the fellow contributed to the development of NeuralMag, an open-source nodal finite-difference code for inverse micromagnetics, and carried out methodological work characterising discretization anisotropy — numerical artefacts arising from the finite-difference grid — which improves the accuracy and reliability of micromagnetic simulations. Ongoing work applies machine learning to predict the extrinsic magnetic properties of isolated cubic grains.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

For the first time in mainstream open-source micromagnetics, the project makes the full set of crystallographic point-group Dzyaloshinskii–Moriya interaction (DMI) terms available in a single, freely usable framework (OOMMF via Ubermag). Previously only a handful of symmetry classes could be modelled; researchers worldwide can now simulate chiral magnetic materials across symmetry classes that were effectively inaccessible without bespoke code, lowering the barrier to studying novel chiral and topological textures.
The project also advances the methodology of the field. Machine-learning workflows replace the manual, expert-driven mapping of magnetic phase diagrams with automated, reproducible discovery and classification of phases, making it feasible to screen large simulation datasets and accelerating materials discovery. The extended mag2exp "virtual experiment" capabilities go beyond existing tools by predicting the experimental signatures, tightening the loop between simulation and experiment and reducing experimental cost. Complementary methodological work quantifying discretization anisotropy improves the accuracy and reliability of finite-difference simulations, and a contribution to NeuralMag adds open inverse-micromagnetics capability to the community toolset.
The potential impact lies in enabling the discovery of new magnetic textures relevant to spintronics and energy-efficient data storage, supporting the EU's digital and green transitions. All results are released as open-source software (BSD-3-Clause, via GitHub, pip and conda) at TRL 3–4, maximising uptake under open-science and FAIR principles.
Key needs to ensure further uptake and success include: further research to exploit the newly accessible DMI material classes; experimental demonstration and validation of the predicted textures; continued integration and maintenance within widely used community packages, together with standardisation of the methods; and sustained international, cross-sector collaboration within the European Research Area to move these fundamental capabilities toward longer-term spintronic applications.