Paper Number
CS1                         My Program 

Session
Colloidal Suspensions and Granular Materials

Title
AI-powered Rheo-SAXS-XPCS analysis of cooperative rearrangement in dense colloidal suspensions

Presentation Date and Time
October 20, 2025 (Monday) 9:50

Track / Room
Track 1 / Sweeney Ballroom A

Authors (Click on name to view author profile)

  1. He, Hongrui (Argonne National Laboratory, Materials Science Division, Center for Molecular Engineering)
  2. Tian, Yuan (New York University, Tandon School of Engineering)
  3. Liang, Heyi (New York University, Tandon School of Engineering)
  4. Chu, Miaoqi (Argonne National Laboratory, Advanced Photon Source)
  5. Jiang, Zhang (Argonne National Laboratory, Advanced Photon Source)
  6. de Pablo, Juan (New York University, Tandon School of Engineering)
  7. Tirrell, Matthew (University of Chicago, Pritzker School of Molecular Engineering)
  8. Narayanan, Suresh (Argonne National Laboratory, Advanced Photon Source)
  9. Chen, Wei (Argonne National Laboratory, Materials Science Division)

Author and Affiliation Lines (in printed abstract book)
Hongrui He1, Yuan Tian2, Heyi Liang2, Miaoqi Chu3, Zhang Jiang3, Juan de Pablo2, Matthew Tirrell4, Suresh Narayanan3 and Wei Chen1
1Materials Science Division, Argonne National Laboratory, Lement, IL 60439; 2Tandon School of Engineering, New York University, New York, NY 11201; 3Advanced Photon Source, Argonne National Laboratory, Lemont, IL 60565; 4Pritzker School of Molecular Engineering, University of Chicago, Chicago, IL 60637

Speaker / Presenter 
He, Hongrui

Keywords
artificial intelligence; colloids; methods; machine learning

Text of Abstract
Cooperative dynamics in materials like colloidal suspensions, gels, and polymer arise from complex surface interactions or structural variations. These collective motions drive critical behaviors such as yielding, failure, and avalanches, which are key to material performance. Traditional models, based on statistical mechanics and assuming uniform random particle motion, often miss the complexity of these systems. As a result, cooperative rearrangement regions (CRRs) in X-ray Photon Correlation Spectroscopy (XPCS) are frequently mistaken for noise due to limited coherence and inadequate theoretical models. To tackle this, we developed an AI-powered framework that treats dynamic bursts in two-time correlation functions as spatiotemporal “objects.” Using deep learning, our approach detects and tracks these events, identifying the occurrence, duration, and intensity of CRRs. A lightweight deep-learning detector ensures efficient analysis. Validated through theory, simulations, and experiments, this method accurately captures CRRs and cooperative dynamics in soft matter. Our framework provides clear insights into these complex behaviors and serves as a robust tool for studying soft matter systems.