Accelerated screening of diamines for polyamide synthesis

Biobased polyamides offer an attractive route to more sustainable plastics, but identifying monomer combinations that yield targeted polymer properties remains challenging. Here, we screened diamines to identify monomers that enhance the thermal stability of polyamides synthesized from dimethyl furan-2,5-dicarboxylate. We described 38 commercially available diamines using molecular descriptors and transformed them into a low-dimensional latent representation that captured dominant diversity in diamine structure, size, flexibility, hydrophobicity and cyclicity. We systematically selected a representative subset of nine diamines for polyamide synthesis and regressed measured polyamide properties against the latent embeddings. The resulting regression models predicted thermal degradation onset and 5% mass-loss temperatures with mean validation errors of 14-25 °C for independent polyamides. Model interpretation indicated that higher degradation onset temperatures were associated with larger, more flexible and more hydrocarbon-rich diamines with reduced cyclicity and hydrogen-bonding capacity. Guided by these predictions we prepared polyamide films and found that a dodecane-1,12-diamine-based polyamide with higher thermal stability formed a self-standing transparent film with low water wettability, whereas a 1,2-cyclohexane-based polyamide with one of the lowest predicted stabilities formed a yellowish, fragile film with near-complete wetting. These results demonstrate how latent molecular representations can guide diamine selection and the design of biobased polyamides with targeted thermal and film-forming properties. Our approach can be used in other polymer synthesis problems where molecular descriptors can be linked to target properties.