Stalking the Synthetically-Accessible Polymer Universe with RxnChainer: Digital Synthesis, Retrosynthesis and Beyond

Polymers enable countless modern technologies, yet vast regions of their chemical space remain unexplored. Traditional polymer discovery relies on chemical intuition, ingenuity, and experience (with a healthy dose of serendipity), yet it fails to leverage millions of potentially accessible and synthesizable polymer structures. Here, we present RxnChainer, a digital methodology integrating virtual polymer generation, retrosynthetic analysis, and post-polymerization modification to systematically explore synthetically accessible polymer space. Using commercially available monomers from the Toxic Substances Control Act (TSCA) and ChEMBL databases and RxnChainer, we generated over 289 million hypothetical polymers across 44 polymerization pathways spanning 32 polymer classes, including polyamides, polyimides, polyesters, and polyethers. Comparison with the known (i.e., previously synthesized) spectrum of polymers revealed that a significant portion of these new synthesizable structures are novel, i.e., previously unknown and unexplored. We demonstrate the methodology's versatility through automated retrosynthetic planning for 30,000 polyesters and targeted functionalization via four post-polymerization modification pathways incorporating vinyl and nitrile pendant groups. The resulting datasets enable downstream tasks such as property-driven screening, application-specific design, and training of generative models.