Speaker
Description
Understanding the surface behaviour of polymer composites in Low Earth Orbit (LEO) environments is critical for the design of durable space materials. In this study, the evolution of wettability in poly(2,5-benzimidazole) (ABPBI) composites reinforced with carbon nanotubes (CNTs) under UV--ozone exposure was investigated as a proxy for oxidative processes encountered in LEO, where solar ultraviolet radiation dissociates molecular oxygen to generate highly reactive atomic oxygen species responsible for surface oxidation and degradation of polymeric materials \citep{Banks2004AO, deGroh2008AO}. Water contact angle (WCA) and surface free energy (SFE) measurements were obtained for composites with varying CNT loadings (0--3 wt.\%) subjected to controlled UV--ozone treatment times.
The experimental results reveal a distinctly non-linear evolution of wettability, characterised by an initial rapid decrease in WCA followed by fluctuations at longer exposure times \citep{NkosiInPrep2026}. This behaviour is attributed to competing surface mechanisms, including the formation of polar oxygen-containing functional groups through photo-oxidation, and the concurrent etching or removal of oxidised surface layers due to reactive oxygen species and UV-induced degradation processes \cite{Fattahi2020UVOzone, Satoh2025UVOzone, Feldman2002PhotoOxidation}.
To capture this complex behaviour, a machine learning framework was developed to model the relationship between exposure conditions and wettability response. Feature engineering techniques, including non-linear transformations of exposure time and interaction terms with CNT loading, were incorporated to reflect the underlying physical processes. Ensemble-based regression models were trained to predict WCA, demonstrating strong agreement with experimental observations and successfully reproducing the observed non-monotonic trends.
The results highlight the importance of non-linear modelling approaches in understanding surface evolution under oxidative environments. This work demonstrates that integrating experimental data with machine learning provides a powerful pathway for interpreting complex physicochemical behaviour in advanced polymer composites without relying on simplified linear assumptions.
| Apply for student award at which level: | PhD |
|---|---|
| Consent on use of personal information: Abstract Submission | Yes, I ACCEPT |