The M2DESCO has being running along the three expected technical paths, as responding to the overall Project workflow; namely i) the Modelling approach including the development of individual materials and process models, their enhancement (faster, better performance) through AI/ML tools, and their interconnection (multiscale), ii) the Safe and Sustainable by Design in steps 1-5, encompassing safety/risk, environmental, and social-economical assessments and iii) the Experimental approach, stressing over the need to enhance the performance of specimen/test result production (high throughput testing).
The modelling activities have progressed toward the complex challenge of developing reliable physical methods to compute and predict the intricate structures of high entropy multielemental materials (HEAs), including those synthesized as thin films. The project has implemented models for HEA properties and phase stability using three main approaches: thermodynamic phase equilibrium calculations (CALPHAD), ab-initio ground-state stability via density functional theory (DFT), and molecular dynamics (MD) simulations for large atomic systems. During this initial phase, M2DESCO focused on validating these individual models and establishing an effective multiscale modelling architecture. DFT serves as the primary benchmark for energetic and mechanical properties, enabling the refinement of interatomic potentials used in CALPHAD and MD to extend predictions across longer time and length scales.
Traditional trial-and-error methods are impractical for modelling complex HEAs with more than five elements. To overcome this, the project is developing AI and machine learning tools to identify and select interatomic potentials (MACE). Using cluster expansions, quasi-harmonic calculations, and CPA/KKR methods supported by MACE, the project has begun generating predictive results for HEA properties such as phase stability, elastic modulus, friction, and wetting.
Another key modelling direction involves HEAs produced under non-equilibrium conditions, such as coatings formed via physical vapor deposition (PVD). M2DESCO is exploring advanced techniques like Arc-evaporation and High Power Impulse Magnetron Sputtering (HIPIMS), which are highly relevant to the coating industry. The strategy combines macroscopic computational fluid dynamics with atomistic non-equilibrium Monte Carlo models, previously developed for conventional sputtering. These models are enriched by input from industrial partners and the materials modelling team, with parameters like atomic sticking coefficients helping to bridge processing and material behavior.
On the experimental side, two main approaches are underway. First, test HEA materials and coatings are being produced and validated to generate training datasets for the modelling teams. Two coating families have been selected: hard phases (AlTiCrMoW+N) and corrosion-resistant compositions (CoNiFeMoCr), developed both as bulk targets and coatings. Second, high-throughput testing procedures are being designed and implemented to accelerate development timelines by a factor of five. Two experimental furnaces and two industrial reactors have produced over 100 test specimens, which are now undergoing compositional, mechanical, and microstructural evaluation, with results documented in D3.1. The project is moving toward more complex HEA coating structures and has initiated coating tests on industrial components, with three test runs planned using optimized materials.
Finally, the Safe and Sustainable by Design (SSbD) activities have been successfully launched and aligned with both project and industrial objectives. The full five-step SSbD framework has been designed, covering value chain mapping, benchmark identification, and socio-economic impact analysis. Industrial partners have provided essential data for technical assessments, enabling the construction of safety, environmental, and socio-economic evaluation criteria. AI and ML tools are being applied to enhance safety and risk assessments through SAR models for multicomponent systems, with promising preliminary results. Life cycle assessments have also progressed, focusing on methodology development, data quality evaluation, and gap identification as outlined for this first reporting period.
