From Molecules to Plants: AI and Multiscale Simulation for Specialty Materials Discovery

Speaker:

Jean Yves Delannoy
Chief Digital Officer
Arkema

Abstract:

Specialty-materials R&D sits on decades of data, yet new products are still largely found by trial and error. This talk traces a path from molecular modeling to industrial AI. I will first revisit multiscale simulation of aqueous guar solutions, where automated coarse-graining combined with machine learning helps predict polymer rheology across chemical space. I will then show how data science is reshaping R&D at Arkema: fingerprint-based molecular recommendations, models flagging biodegradable candidates with over 80% accuracy, and generative AI for literature mining and formulation design. Finally, I will introduce NEOMAT, a European program coupling a materials knowledge graph, generative molecular design and physics-informed multiscale simulation with active-learning loops driving automated labs. Throughout, I will reflect on what it takes to move these tools from the laptop to the plant.
 

Speaker Biography:

Jean-Yves Delannoy is Scientific Director of Digital R&D and Chief Digital Officer at Arkema, where he leads the use of AI, data science and simulation to accelerate materials innovation and speed up the business transformation. A physicist by training, he earned his PhD from École normale supérieure de Lyon and the University of Waterloo. He began his industrial career at Michelin, applying molecular modeling to tire performance. At Solvay, he led polymer modeling and simulation, managed joint research laboratories with CNRS, including the COMPASS lab with the University of Pennsylvania, and later drove R&I digital transformation. He then served as Head of Design and Representation Systems at Siemens before joining Arkema in 2023. His work spans polymer physics, multiscale simulation, machine learning and generative AI for chemistry. He also completed executive training at INSEAD.