Title:	Computational design of novel functional materials for sustainable technologies
DNr:	NAISS 2025/3-31
Project Type:	NAISS Large
Principal Investigator:	Johanna Rosen <johanna.rosen@liu.se>
Affiliation:	Linköpings universitet
Duration:	2026-01-01 – 2027-01-01
Classification:	10304 20501 20506
Homepage:	https://liu.se/en/research/materials-design
Keywords:

Abstract

We apply for a continued 1-year NAISS Large Compute and Storage project to advance theoretical materials research in the Materials Design division at Linköping University. Our team has a proven track record of high-impact publications (Science, Nature Synthesis, Materials Today, Nature Communications, Angewandte Chemie, Matter) with steadily increasing output from 2019-2025, demonstrating both productivity and scientific merit from using NAISS resources. The division combines state-of-the-art computational theory with experimental synthesis to discover novel materials and enable emerging applications in energy storage, catalysis, hard coatings, water remediation, structural materials, sensors, and carbon capture. Building on our current NAISS allocations (2024/3-15 Compute, 2024/6-326 Storage), where we utilized >4 million core-hours/month on Tetralith and Dardel and managed 20 TB (NSC) and 10 TB (PDC), we will expand projects employing ab-initio calculations, molecular dynamics, and machine learning to explore 3D to 2D material conversions, MXenes, surface reactions, on-surface synthesis, and identification of synthesizable materials with targeted properties. Professor Johanna Rosén serves as principal investigator, supported by Assoc. Profs. Jonas Björk and Martin Dahlqvist, with a team of ~30 active theoretical and experimental researchers. The group’s computational work has produced 42 peer-reviewed publications, 2 PhD theses, and 1 Licentiate thesis between October 2023–October 2025. Our main tools include VASP, USPEX for crystal structure prediction, LAMMPS for molecular dynamics, LOBSTER for bonding analysis, and ASE/Pymatgen-based high-throughput workflows. For large disordered alloys and supercells (200–500 atoms), pretrained machine-learned interatomic potentials (e.g., MACE) reduce computational load, and Fat nodes are essential for memory-intensive tasks. We request continued allocations at current levels: 2 000 000 core-hours/month on Tetralith and 2 000 000 core-hours/month on Dardel. Active storage needs are 25 TB and 7 500 000 files on NSC, 15 TB and 5 000 000 files on PDC, reflecting high-throughput workflows for multiple concurrent researchers. All data are managed with user- or project-specific directories, ASE databases, and automated scripts. Non-active data will be continuously removed, and key datasets archived locally and shared publicly according to FAIR principles. Scientific focus: WP1: Discovery of novel laminated materials (MAX, MAB, s-MAX, and beyond) with tailored mechanical, electronic, and chemical properties using DFT, cluster expansion, and evolutionary algorithms, including A-site substitutions for novel 2D precursors. WP2: Identification and functionalization of 2D materials via chemical exfoliation from layered 3D precursors, combining high-throughput simulations and machine learning to predict selective etching, surface terminations, and exfoliation energetics. WP3: Surface reactions and catalysis studies on 2D carbides and single-atom catalysts, including tandem CO₂ reduction/olefin formation, alcohol dehydration, and on-surface synthesis of organic nanostructures such as graphene nanoribbons, integrating DFT, NEB/Dimer methods, and ML-driven potential energy landscape exploration. This allocation is essential for sustaining ongoing projects, supporting new research directions, and ensuring efficient execution of large-scale, computationally demanding studies to advance functional materials for sustainable technologies.