Tag: Ensemble Machine Learning Model
The NICKEFFECT project has successfully bridged the gap between laboratory research and industrial application with the publication of CWA 18302:2026, a landmark CEN Workshop Agreement developed in collaboration with CEN-CENELEC. This document establishes the first harmonized European protocol for the electrochemical characterization of non-noble, porous metal-based electrodes for hydrogen generation in acidic media. By standardizing testing cells, activity parameters, and data reporting, this success story—led by CIDETEC, the Spanish Association for Standardization (UNE), and project coordinator Aliona Nicolenco—provides the scientific community with a critical tool to accelerate the development of cost-effective, PGM-free water electrolysis technologies, ensuring that early-stage material innovations can be reliably benchmarked and scaled for a net-zero future.
The NICKEFFECT project leverages cutting-edge computational strategies to revolutionize the design of magnetic Ni-Co alloys for RAM applications. While traditional first-principles methods like Density Functional Theory (DFT) and modern universal machine-learning interatomic potentials have significantly accelerated the characterization of known structures, the search for entirely new materials has historically been limited by the vast dimensionality of chemical space. To overcome this, we are integrating Generative Artificial Intelligence (GenAI), which encodes complex crystal structures into lower-dimension latent spaces to suggest novel compounds. Moving beyond a simple "generate-then-screen" workflow, our latest research—developed in collaboration with Matgenix—explores co-modality alignment. By synchronizing latent spaces for specific physical properties and synthesis parameters, we can directly optimize the generation of compounds that meet precise performance targets and manufacturing requirements, setting a new standard for efficient, objective-driven materials discovery.
In the pursuit of materials innovation, an accurate machine learning model is only the beginning. Within the NICKEFFECT project, we are bridging the gap between isolated experiments and scalable scientific impact by integrating robust MLOps (Machine Learning Operations). By utilizing tools like MLflow to ensure every dataset, parameter, and model version is traceable and reproducible, we’ve transformed complex data into a dependable digital decision-support system. Discover how our structured MLOps workflow—from automated tracking to REST API integration—is setting a new standard for transparency and efficiency in materials science research.
Towards Safe and Sustainable by Design Nanomaterials: Functionality-based Screening at the Nanoscale
The NICKEFFECT project is redefining material innovation by applying the European Commission’s Safe-and-Sustainable-by-Design (SSbD) framework. By replacing costly platinum-group metals with advanced nickel-based coatings, the project demonstrates how early-stage screening tools—including LCA and Risk Assessment—can guide the development of greener, more efficient materials for water electrolysis and fuel cells.
The NICKEFFECT project highlights a transformative leap in fuel cell efficiency through Advent Technologies’ Ion-Pair™ HT-PEM MEAs. Developed in collaboration with Los Alamos National Laboratory, this next-generation technology overcomes the thermal and durability limitations of traditional systems by utilizing a unique bond between charged polymers and phosphoric acid. Capable of operating across a wide temperature range and delivering significantly higher current densities, these membranes enable compact cooling architectures and simplified thermal management. This breakthrough is set to redefine high-power applications in aviation, marine, and heavy-duty transport, providing a robust, lightweight, and durable solution for the global transition to sustainable energy.
The NICKEFFECT project, represented by partner Singulus Technologies AG, successfully exhibited its cutting-edge advancements at SEMICON Europa 2025 in Munich, focusing on green and sustainable data storage. Aligned with the event’s theme of European economic resilience, the exhibition highlighted the latest results on Platinum (Pt)-free, non-volatile MRAM stacks, a key development in reducing reliance on critical raw materials. Product Manager Matthias Landmann shared these innovative process developments at the Singulus stand, underscoring the project's commitment to enabling a technologically advanced and environmentally sustainable future for the microelectronics industry.