Asbestos containing materials detection and classification by the use of hyperspectral imaging.

In this work, hyperspectral imaging in the short wave infrared range (SWIR: 1000-2500nm) coupled with chemometric techniques was evaluated as an analytical tool to detect and classify different asbestos minerals, such as amosite ((Fe2+)2(Fe2+,Mg)5Si8O22(OH)2)), crocidolite (Na2(Mg,Fe)6Si8O22(OH)2) and chrysotile (Mg3(Si2O5)(OH)4), contained in cement matrices. Principal Component Analysis (PCA) was used for data exploration and Soft Independent Modeling of Class Analogies (SIMCA) for sample classification. The classification model was built using spectral characteristics of reference asbestos samples and then applied to the asbestos containing materials. Results showed that identification and classification of amosite, crocidolite and chrysotile was obtained based on their different spectral signatures, mainly related to absorptions detected in the hydroxyl combination bands, such as Mg-OH (2300nm) and Fe-OH (from 2280 to 2343nm). The developed SIMCA model showed very good specificity and sensitivity values (from 0.89 to 1.00). The correctness of classification results was confirmed by stereomicroscopic investigations, based on different color, morphological and morphometrical characteristics of asbestos minerals, and by micro X-ray fluorescence maps, through iron (Fe) and magnesium (Mg) distribution assessment on asbestos fibers. The developed innovative approach could represent an important step forward to detect asbestos in building materials and demolition waste.