A low-cost LIBS detection system combined with chemometrics for rapid identification of plastic waste

Highlights

• Application of LIBS for the rapid identification of plastic waste.

• Potential of low cost compact LIBS detection system is demonstrated.

• Data acquisition with a single laser pulse demonstrated.

• Identification rates ~97% achieved with the application of chemometric.

• Random forest based feature selection reduced the analysis time by more than 10 times.

Abstract

We present, rapid and efficient identification of ten different types of post-consumer plastics obtained from a local recycling unit by deploying a low cost, compact CCD spectrometer in laser-induced breakdown spectroscopy (LIBS) technique. For this investigation, spectral emissions were collected by an Echelle spectrograph equipped with an intensified charge-coupled device (ES-ICCD) as well as a non-gated Czerny Turner CCD spectrometer (NCT-CCD). The performance is evaluated by interrogating the samples in a single-shot as well as accumulation mode (ten consecutive laser shots). The results from principal component analysis (PCA) have shown excellent discrimination. Further, the artificial neural network (ANN) analysis has demonstrated that individual identification accuracies/rates up to ~99 % can be achieved. The data acquired with ES-ICCD in the accumulation of ten shots have shown average identification accuracies ~97 %. Nevertheless, similar performance is achieved with the NCT-CCD spectrometer even in a single shot acquisition which reduces the overall analysis time by a factor of ~15 times compared to the ES-ICCD. Furthermore, the detector/collection system size, weight, and cost also can be reduced by ~10 times by employing a NCT-CCD spectrometer. The results have the potential in realizing a compact and low-cost LIBS system for the rapid identification of plastics with higher accuracies for the real-time application.

Introduction

Laser-Induced Breakdown Spectroscopy (LIBS) is an atomic emission spectroscopic technique which can provide qualitative and quantitative information of the elemental composition (Buckley et al., 2000, Costa et al., 2017b, Yao et al., 2020). Detection of multiple elements in the single-shot acquisition and minimal or no sample preparation are the key features of the LIBS technique. Rapid analysis and nearly non-destructive nature with standoff capability have enabled it as a promising application in various fields like explosive detection, space exploration, and archaeological investigations, etc (Gundawar et al., 2017, Pořízka et al., 2018). It has been widely used in combination with various multivariate methods for the identification of materials in different applications (Gundawar et al., 2017, Myakalwar et al., 2015, Pořízka et al., 2018, Sreedhar et al., 2014, Vrábel et al., 2020).

Every year, several million tons of plastic waste is generated worldwide, leading to a landfill problem (Cao et al., 2016, Chen et al., 2014, Geyer et al., 2017, Rochman et al., 2013). Their higher durability adds more to the disposal problem and the traditional disposal methods such as burying and incineration are leading to the wastage of natural resources as well (Chen et al., 2014, Yang et al., 2012). The ever-increasing domestic and industrial plastic waste raises more concerns over their recycling which offers an efficient solution to minimize waste output and reduces the usage of natural resources. Initially, the plastics waste is cleaned in water streams to remove the debris and hazardous contaminants which is otherwise harmful to the operators/workers working in the recycling plant (Ragaert et al., 2017, Singh et al., 2017). Sorting is a crucial step in the recycling process as it affects the quality and cost of the final products (Hopewell et al., 2009, Milios et al., 2018). Currently, manual sorting is the most used technique albeit, it is time-consuming and error-prone (Liu et al., 2018, Ruj et al., 2015). Further, the employees performing the task of manual sorting need to be properly protected as they may come in contact with the hazardous contaminants of the waste materials. Sink floating measurement is another conventional technique used for separation based on their density (Carvalho et al., 2009). However, it requires different reagents for efficient classification. Different spectroscopic methods are reported in the literature for the identification of the plastics. They include Raman spectroscopy (Shameem et al., 2017), and near-infrared spectroscopy (NIR) (Huth-Fehre et al., 1995), etc (Kassouf et al., 2014, Sharma et al., 2019b, Vrancken et al., 2017). NIR and Raman measurements have limitations with the black and dark-coloured samples (Huth-Fehre et al., 1995, Shameem et al., 2017).

LIBS can overcome many of the limitations of conventional and other spectroscopic methods (Gundupalli et al., 2017). However, it has the issue of matrix effect which can mainly reduce the accuracy of quantitative measurements. The works reported for the identification of plastics using the LIBS technique are given in Table 1.

Various intensity ratios involving the atomic and molecular emissions were evaluated for sorting the different types of plastics in which H/C and C₂/C are found to be the prominent (Gondal and Siddiqui, 2007, Junjuri et al., 2019). Further, the improvement in accuracy with the aid of the ratiometric approach was accomplished by utilizing the buffer gas (Barbier et al., 2013). Various statistical and multivariate methods utilized for the identification of the plastics are in given Table 1 which have shown promising results. Nevertheless, most of these studies described in the literature were mainly reported on the standard plastics purchased from the companies/industries. Only a very few studies were performed on used/post-consumer plastics as listed in Table1. Since there are chances of contamination of the samples from the time of disposal by a consumer to the waste collection center, it increases the difficulty of identification. To address these issues, it is necessary to build a data model using the real-time-post-consumer plastics obtained from the recycling center. Moreover, 20 works out of 22 in Table 1 have acquired spectral data by accumulating the 10–50 consecutive laser pulses depending on the experiments. Though the accumulation of the spectra increases the signal strength to noise ratio at the same time, it increases the analysis time. Furthermore, 60 % of the total works reported in the literature (details are given in the last column of Table 1) have employed the Echelle based ICCD spectrometers and remaining studies have used the multichannel (2–6) gated CCD spectrometers.

The main motive of our work is to minimize the cost, size, weight, and the analysis time of the LIBS detection system. To investigate that, the performance of a single channel non-gated low-cost CCD spectrometer (NCT-CCD) based system operated in single-shot mode is compared with those obtained from an Echelle spectrograph coupled with an ICCD (ES-ICCD). The experiment was performed on ten different post-consumer plastics obtained from the local recycling unit as their identification would be more relevant to the actual scenario. The spectra were acquired with the single shot and ten shot accumulation modes to evaluate the performance of multivariate techniques in both the conditions. Further, multivariate methods like PCA, random forest (RF), and ANN were deployed for treating the data and to validate their application for a bigger set of ten different types of post-consumer plastics. Finally, the RF algorithm is used to find the important features of the LIBS spectra.