Materialism: A Materials Science Podcast Episode 98: Accelerating Catalyst Research with Meta
Dec 11, 2024
Larry Zitnick, Research Director at Meta's AI team, and Aaike van Vught from VSParticle dive into the intriguing intersection of machine learning and materials science. They discuss the creation of OCx24, an open catalysis database, revealing the challenges of building it from scratch. The guests also explore how autonomous spark ablation techniques streamline materials synthesis. With a holistic look at reproducibility and data-sharing across the community, they illustrate how tech can drive sustainable innovations in catalyst research.
AI Snips
Chapters
Transcript
Episode notes
Meta's Motivation for Materials Research
- Meta's foray into materials science is driven by sustainability concerns and the need for energy-efficient data centers.
- They also research materials for AR devices, highlighting the field's broader applications beyond sustainability.
VSParticle's Synthesis Success
- Meta initially struggled to find companies willing to synthesize small quantities of diverse materials for their catalysis research.
- VSParticle succeeded where other companies failed, providing the needed experimental materials.
Bridging Computational Models and Experiments
- The OCX24 project aims to bridge the gap between computational catalysis models and experimental results.
- This involved creating a diverse dataset with positive and negative results, tested under consistent and industrially relevant conditions.