Departmental Seminar

Microplastic pollution presents significant challenges for environmental monitoring, especially due to contamination by water and organic matter which complicates in situ spectroscopic detection methods. Currently, monitoring microplastics involves sample collection, cleaning, and classification. Here, we introduce a robust near-infrared (NIR) spectroscopy system optimized for rapid, on-site detection of microplastics in environments contaminated with water and plant material. Utilizing data augmentation techniques and machine-learning classification algorithms, the developed model accurately identifies the polymer compositions of plastic particles as small as 500 μm. The approach uses augmented spectra to incorporate contamination into classifier training, significantly improving model robustness and transferability. Key findings highlight that high model accuracy persists even at challenging signal-to-noise ratios, demonstrating practical applicability for field deployment.

In the second part of this talk, I will describe our ongoing efforts along a few different projects in collaboration with Prof Zhanfei Liu at UTMSI, particularly with respect to understanding the degradation mechanism of plastics in the environment. I will introduce IR spectroscopy as a useful tool to characterize heterogeneous matter including plastics as well as biomolecules