Unsupervised Learning Unravels the Structure of Four-Arm and Linear Block Copolymer Micelles

Ziolek, RM (Ziolek, Robert M.)[ 1 ] ; Smith, P (Smith, Paul)[ 1 ] ; Pink, DL (Pink, Demi L.)[ 1 ] ; Dreiss, CA (Dreiss, Cecile A.)[ 2 ] ; Lorenz, CD (Lorenz, Christian D.)[ 1 ]

The Pluronics and Tetronics are amphiphilic block copolymers that typically self-assemble into micelles in aqueous solution, which have societally important applications in drug delivery and biomaterials.  Studying these micelles is challenging experimentally given their relatively small size, while obtaining detailed insight from computer simulations is difficult since the polymer chains are large and conformationally flexible.

In this work, Ziolek and co-workers employed unsupervised machine learning techniques to provide new fundamental insights into the structure of these micelles. They showed that different polymer conformations exist preferentially in different regions of the micelle core. Additionally, hidden Markov models were used to examine the different states that the polymer arms adopt in the micelle shell, uncovering previously unobserved structure that underlies the seemingly disordered shell region. These results provide a detailed picture of these industrially relevant nanoparticles and supports their rational deployment in various application areas.

DOI: 10.1021/acs.macromol.0c02523


Follow @tyc_london for updates from the Thomas Young Centre.