| Abstract Scope |
Extended space travel in low Earth orbit (LEO) demands materials capable of withstanding extreme environments with a main source of material degradation being from atomic oxygen (AO) erosion. This work presents a machine learning (ML) framework trained on experimental data from NASA’s MISSE (Materials International Space Station Experiment) missions to predict AO erosion of polymers. Key chemical and processing features driving erosion resistance are identified using interpretable ML models. These predictive tools are integrated with generative algorithms, including a genetic algorithm and virtual forward synthesis, to design novel, synthetically accessible polymers optimized for LEO durability. This approach enables efficient screening and discovery of AO-resistant polymers critical for more robust space station components. By combining predictive modeling, materials design, and LEO-specific constraints, this research supports the goals of extended space travel desired for both commercial and government applications as well as contributes to the broader commercial space manufacturing ecosystem. |