Rapid Identification of Automotive Plastics in Dismantling Operations: Evaluation of Specular-Reflectance Infrared Spectroscopy Systems

Specular-reflectance infrared spectroscopy systems can identify the polymer material of an automotive part in about 5 seconds and are currently commercially available. Issues related to the rapid identification of plastics were recently examined at the Vehicle Recycling Development Center, which is operated by the United States Council for Automotive Research (USCAR). The accuracy of identification is a crucial concern in order to minimize co-mingling or contamination of sorted plastics in dismantling-sorting-recycling operations. Accuracy reports in the literature have ranged from 70% to >99%. Our investigation of the signal-to-noise levels of spectrometers, identification algorithms, and spectral reference libraries indicated that the quality-and-completeness of the reference library is the strongest determinant of accuracy when evaluating current commercial systems. With adequate spectral libraries, identification accuracy of 99% can be achieved. Further, most identification errors tend to be “near misses”, with polymers of similar compositon chosen instead of the correct polymer. Foams and elastomers are problem specimens for specular-reflectance, because they give poor spectral reflectance and hence inadequate signals. Other concerns with specular reflectance systems are rough surfaces versus smooth surfaces and the selection of coating-removal tools.