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

Understanding the nanoscale structure of polymeric micelles is challenging: their relatively small size tests the limits of most experimental techniques, while the great conformational flexibility of the individual polymer chains makes deriving insight from computer simulations difficult. Pluronics and Tetronics are amphiphilic block copolymers based on poly(ethylene oxide) and poly(propylene oxide) blocks that self-assemble into micelles, which have been widely studied experimentally given their extensive use as excipients in drug formulations and as biomaterials. In contrast to these wide-ranging applications, the characterization of their nanoscale structure and dynamics is still incomplete. In particular, how the architecture of the blocks in linear Pluronics and four-arm Tetronics influences the arrangement of the chains within a core-shell morphology is not well understood. We apply unsupervised machine learning techniques to provide an unprecedented level of detail regarding the distribution of polymer conformations within the micelles and identify the underlying structure in the seemingly disordered micellar corona. The methodology applied in this work improves our understanding of the structure of these industrially relevant nanoparticles and establishes a general methodology for investigating the conformational distribution of polymers in self-assembled structures.

Automated summary

Understanding the nanoscale structure of polymeric micelles is challenging due to their small size and high conformational flexibility, but applying unsupervised machine learning techniques can provide unprecedented detail and improve our understanding of these nanoparticles.