Page 20 - Read Online

Page 14 of 17 Schertzer et al. J. Mater. Inf. 2025, 5, 5 https://dx.doi.org/10.20517/jmi.2024.69

5 138 ± 34.7 34.6 ± 44.7 22.6 ± 18.8

WU: Water uptake; SR: swelling ratio.

offers an immediate pathway to identifying stable, high-conductivity polymers without relying on fluorinated monomers.

To summarize the findings and contributions presented in this work:
• Although the body of AEM data remains relatively small, it provides untapped potential for exploring the vast copolymer space, offering practical solutions
for material challenges.
• Using theoretical IEC as a key descriptor in ML models, we bypass the need for synthesized samples, accelerating the screening process.
• Incorporating several properties in a MT framework empowers us to pinpoint materials that portray high hydroxide conductivity, low WU, and low SR. In
the future, mechanical properties and performance under alkaline conditions should be considered to identify polymers with long-term stability for practical
application.
• Despite the dominance of fluorinated monomers in top-performing polymers, we identified hundreds of novel, fluorine-free copolymers with strong
predictive confidence, marking a significant step toward more sustainable materials.
• Future efforts will focus on expanding our dataset using natural language processing and advanced molecular modeling, paving the way for even more robust
and generalizable models.

DECLARATIONS
Authors’ contributions
Conceived and guided the work: Ramprasad, R.
Curated the training dataset from the literature: Schertzer, W., Shukla, S., Rafiq, R.
Designed, trained, and evaluated the machine-learning models: Schertzer, W., Shukla, S.
Generated the candidates, predicted their properties, and ranked them: Schertzer, W.