An Evaluation On Printed Circuit Boards Separability From Bulk Electronic Waste Using Hyperspectral Imaging

Electronic industry has considerably rose in recent years, promoting the development of new technologies and requiring at the same time a high demand for raw materials. The increasing demand of new technological products dramatically contributed to increase the volume of waste consisting of electrical or electronic equipment (WEEE) to be discarded as defective, unused, or obsolete. However, a significant amount of valuable materials is contained in WEEE, such as metals, precious metals, high-quality plastic and other profitably recoverable materials. In this study, the attention was focused on the development, the set-up and the implementation of non-ferrous metal concentration strategies to apply at the end of an industrial recycling process. To reach this goal hyperspectral imaging (HSI) - based sensing technique can be applied. HSI is one of the main emerging innovative technologies that can be profitably appliable to fulfil an in-depth characterization of WEEE by-products as well as for the implementation of sensor-based sorter [1, 2]. According to different authors, HSI can be very useful for evaluating different kind of plastics or resins for sorting them from mixed WEEE [3, 4, 5]. The main target of this study was to set-up a model, based on a chemometric approach, in order to recognize embedded metals PCBs, with particular reference to the recovery of non-ferrous metals from small and medium appliances WEEE bulk by-product, obtained by an innovative Magnetic Density Separation (MDS) process. The resulting and investigated product belongs to an MDS output density fraction ranging between 1300 and 2200 kg/m3, with - 10 mm size. The product is mainly constituted by printed circuit boards (PCBs), plastics and glass. A small amount of metals still occurs in this output. Even if, at least theoretically, in this density range metals should not occur. Hyperspectral dataset were obtained from different scans carried out by SisuCHEMA XL™ Chemical Imaging Workstation, capable of acquiring in a range of the spectrum ranging from 1000 nm to 2500 nm. The different material classes were firstly evaluated by a visive analysis. Then, a training Dataset was prepared performing an exploratory analysis with PCA, using the PLS_Toolbox 8.1 (Eigenvector Research Inc.), in Matlab™ environment. Collected dataset on training material in SWIR range were first pre-processed with standard normal variate scaling (SNV) by setting an offset value equal to 1 and mean center. The set classes were: black plastic, wires plastic, wood, white plastic, PCBs and glass. A PLS-DA test for classes prediction was carried out on material in petri dish. Two algorithms for preprocessing were used: the Probabilistic Quotient Normalization (PQN) for data normalization and Mean Center. A good identification was obtained for printed circuit boards especially with reference to epoxy resin layer below the copper layer and solder mask on the top (protective layer for the welds). Sporadic misclassifications are present, they are probably due to light scattering, samples heterogeneity and presence of impurities. Infrared spectral range of does not allow the full identification of metallic materials spectra. Indeed, only PCBs are recognized. By applying this model, particular attention must be paid for particle size. E.g., an HSI-based separator or a quality control system may work better with coarse particles: those in which metal is not completely liberated from the embedding material. References: [1] A. Picon et al., Automation of waste recycling using hyperspectral image analysis. Bilbao, IEEE, pp. 1 - 4, 2010. [2] N. Picone & M. Colledani, Characterization of fine metal particles using hyperspectral imaging in automatic WEEE recycling systems. s.l., POLIMI, 2015. [3] R. Palmieri, G. Bonifazi & S. Serranti, Recycling-oriented characterization of plastic frames and printed circuit boards from mobile phones by electronic and chemical imaging. Waste Managment, Volume 34, p. 2120, 2014. [4] H. Masoumi, , S. Safavi & Z. Khani, Identification and Classification of Plastic Resins using Near Infrared Reflectance Spectroscopy. International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, 6 (5), pp. 877, 2012. [5] J. Beigbeder, D. Perrin, J. -F. Mascaro, & M. Lopez-Cuesta, Study of the physico-chemical properties of recycled polymers from waste electrical and electronic equipment (WEEE) sorted by high resolution near infrared devices. Resources, Conservation and Recycling, Volume 78, pp. 105, 2013.