Application of near-infrared spectroscopy for the fast detection and sorting of wood–plastic composites and waste wood treated with wood preservatives

2016 
The aim of this work was to increase the amount of recycled waste wood in the wood-processing industry. Contaminations such as plastics and wood preservatives therefore have to be identified and removed. A promising technique to detect this foreign matter is near-infrared (NIR) spectroscopy. For several organic and inorganic active compounds for wood preservation, the NIR reflectance spectra acquired with sensors with the capability of on-line measurements were characterized. Some of them were significantly different from those of previously published spectra. Moreover, untreated wood was distinguishable from wood treated with preservatives. Confounding factors, such as the water content, wood species and solvent of the preservative, were examined which had so far massively interfered the NIR detection and classification. Chemometric-based solutions are presented in this work to cope with analytical challenges arisen from the complexity of waste wood samples. This formed the basis for a sorting trial in which treated wood chips were detected in real time with an NIR spectrometer and automatically separated by pneumatic nozzles, showing that the laboratory experiments can be transferred into small industrial scale. Additionally, the experiments showed that wood–plastic composites (WPC) of different plastic types were distinguishable with NIR spectroscopy and can be sorted by ejection based on this information. Unmixed material is a major prerequisite for a high-quality recycling of WPC and has never been demonstrated on a technical scale. Furthermore, three different NIR devices (hyperspectral imaging camera, miniaturized spectrometer and sorting plant) were evaluated regarding their applicability in the waste wood recycling process.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    17
    References
    21
    Citations
    NaN
    KQI
    []