Self-supervised machine learning framework for high-throughput electron microscopy

Poseidon AX Publication Alert in Science Advances on Nanomaterial Synthesis!

Tuesday publication post! 📖 In this latest work, published in #ScienceAdvances, in situ liquid-phase TEM was used to show a new self-supervised denoising framework, SHINE, to enhance the technique!

Using the #PoseidonAX holder, the hydrogen evolution reaction (HER) on monolayer MoS₂ catalysts was followed. SHINE was used to enabling high-contrast, high-throughput denoising using only the raw noisy data; no ground-truth images needed. In these low-dose conditions, this approach:

👀 Reveals nanoscale bubble formation and morphological changes without frame averaging
💧 Distinguishes MoS₂ layers, silicon nitride, and evolving H₂ bubbles with unprecedented clarity
🔬 Maintains sample integrity—critical for studying synthesis pathways and nucleation events in liquid

Beyond HER, this method opens the door to more accurate simulations and analyses of #nanoparticle formation during liquid-phase EM experiments.

Want to read the entire work? Find it here!
https://www.science.org/doi/10.1126/sciadv.ads5552

Related Posts

Poseidon AX publication alert in Small for Batteries!
Fusion AX Publication Alert in Small on Nanomaterials Synthesis