Discrimination of man-made vitreous fibres and asbestos fibres using hyperspectral imaging

Man-made vitreous fibres (MMVFs) are synthetic materials made from processed rock, glass, or minerals, characterized by non-crystalline inorganic silicates with a length-to-width ratio of at least 3:1. Depending on their chemical and geometrical properties, particularly their respirable characteristics critical for lung penetration, MMVFs can be classified as either hazardous or non-hazardous [1]. These fibres have widespread use in thermal and acoustic insulation as well as fire protection. However, distinguishing MMVFs from hazardous asbestos fibres remains a significant challenge, especially during building demolition. This study aims to evaluate hyperspectral imaging (HSI) as a non-destructive and non-invasive method to classify and predict the presence of asbestos within MMVF samples, providing a rapid, efficient alternative to traditional methods. The study employed hyperspectral imaging (HSI) to analyse mixed samples containing known percentages of asbestos and MMVF. Hyperspectral data were processed using chemometric techniques, specifically principal component analysis (PCA), to develop a predictive model for asbestos and MMVF’s identification [2]. This job will show the preliminary analytical results obtained using HSI integrated by PCA and traditional SEM images coupled with EDS on the same samples for distinguishing both asbestos and MMVF’s fibre and to compare their data. Unlike traditional scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (EDS), which requires meticulous sample preparation and skilled personnel [3], HSI integrates imaging and spectroscopy to facilitate the rapid identification and spatial mapping of MMVF and asbestos components without manipulating the sample. The application of hyperspectral imaging demonstrated its capability to accurately and efficiently classify MMVF and asbestos components. By using PCA in the chemometric analysis, the method showed promise in predicting asbestos presence within mixed samples, effectively distinguishing hazardous elements. Hyperspectral imaging presents a reliable, non-invasive, and preparation-free alternative to traditional SEM-EDS techniques for MMVF analysis. This approach not only reduces analysis time and exposure risks but also enhances the rapid and accurate characterization of fibrous materials. The findings support the potential of HSI to improve waste management practices and address the environmental and human health risks associated with hazardous materials in demolition and construction waste.