Segmented polymer nanoparticles : Synthesis, AI-assisted morphological analysis and regio-specific functionalization

Abstract

This work presents advances in three key areas of colloid science: the synthesis of multi-lobed particles, their AI-aided morphological analysis, and regio-specific functionalization. Using seeded polymerization, anisotropic and Janus polymer nanoparticles were synthesized, and their formation mechanisms were interpreted through revised applications of Flory-Rehner theory for polymer-monomer mixtures. Computer vision and image analysis methods developed here enabled accurate morphological characterization. Regio-specific functionalization was achieved by selectively modifying one lobe of the particles during synthesis, using UV radical initiators to facilitate photopolymerization. This allowed polyelectrolyte brushes to be grafted onto the particles, enhancing their Janus character. All syntheses were scalable and performed on the gram scale, increasing their practical applicability. Particle morphology significantly impacts behavior in self-assembly, suspension rheology, and sedimentation. Anisotropic particles, particularly colloidal dumbbells with two overlapping lobes, are synthesized via seeded polymerization driven by polymer-monomer phase separation. These particles are relevant for applications in catalysis, drug delivery, membranes, and nanomachines. Phase separation transforms each seed into a nanoreactor, with the seed properties critically determining final morphology. Dumbbells can also serve as seeds for a second polymerization round, creating three-lobed particles. The third lobe forms through additional polymer-monomer phase separation, adding complexity and potentially altering anisotropy. The growth direction of the third lobe varies, sometimes reducing anisotropy depending on differences between the initial lobes. While microparticles conform to the Flory-Rehner theory, nanoparticle behavior requires the Flory-Rehner-Morton theory due to their high surface-area-to-volume ratio. This study extends the understanding of seeded polymerization by synthesizing three-lobed nanoparticles and examining how seed surface properties affect phase separation. By varying dumbbell formulations, the formation mechanism of trimers is elucidated. Optimized protocols yield highly anisotropic, linear trimers on a gram scale, offering pathways for assembling complex superstructures. Characterizing these anisotropic particles poses challenges due to their internal complexity. Morphological characterization must account for each internal subdivision, demanding detailed image analysis of electron micrographs. Traditional image processing methods (e.g., Hough transforms, watershed segmentation) often fail for such complex particles. Therefore, AI and neural network-based segmentation offer more reliable analysis of these morphologies. Despite the need, no prior work focuses on analyzing subdivided particles. To address this gap, we developed a novel AI-powered image analysis method capable of extracting features from particle subdivisions and contextualizing them within each particle. Comparisons with conventional datasets validated the method’s accuracy and minimized systemic errors inherent in older approaches. Beyond morphology, these particles can also exhibit chemical anisotropy. Syntheses enable each subdivision to have distinct chemical compositions, enhancing Janus character. These differences allow for regio-specific immobilization of functional monomers, including photopolymerization initiators, enabling chain-growth polymerizations directly from specific surface regions. Photopolymerization initiators immobilized on the particle surface can form polymer brushes—dense polymer chains tethered by one end. Such brushes have applications in anti-fouling, protein resistance, colloid stabilization, and lubrication. Polyelectrolyte brushes can concentrate counterions, creating unique local environments ideal for use as nanoreactors. These nanoreactors can, for instance, reduce metal ions within the brush, enhancing catalytic performance while stabilizing the metal nanoparticles. We demonstrate regio-specific immobilization of photoinitiators on dumbbell seeds, leading to brush formation with spatial control. These strategies create Janus dumbbells with diverse compositions and tailored polyelectrolyte brushes. The use of brushes as nanoreactors was explored for synthesizing catalytic nanoparticles within defined zones. This work contributes major advancements in applying Flory-Rehner-Morton theory to nanoparticle synthesis, analyzing complex lobed nanoparticles, and enabling regio-specific functionalization. Emphasis was placed on understanding reaction mechanisms and creating novel analytical tools rather than direct application development. Future studies could build on these findings to create colloidal crystals with tunable optical properties, potentially usable in photonic devices or logic gates. Functionalized lobed particles could improve catalytic site separation, targeted drug delivery, membrane selectivity, self-assembly, and performance as Pickering surfactants. The image analysis tools developed also pave the way for broader applications in characterizing even more complex colloidal structures.