ArticlePDF Available

Miniaturized Near-Infrared (MicroNIR) Spectrometer in Plastic Waste Sorting

MDPI
Materials
Authors:

Abstract and Figures

Valorisation of the urban plastic waste in high-quality recyclates is an imperative challenge in the new paradigm of the circular economy. In this scenario, a key role in the improvement of the recycling process is exerted by the optimization of waste sorting. In spite of the enormous developments achieved in the field of automated sorting systems, the quest for the reduction of cross-contamination of incompatible polymers as well as a rapid and punctual sorting of the unmatched polymers has not been sufficiently developed. In this paper, we demonstrate that a miniaturized handheld near-infrared (NIR) spectrometer can be used to successfully fingerprint and classify different plastic polymers. The investigated urban plastic waste comprised polyethylene (PE), polypropylene (PP), poly(vinyl chloride) (PVC), poly(ethylene terephthalate) (PET), and poly(styrene) (PS), collected directly in a recycling plastic waste plant, without any kind of sample washing or treatment. The application of unsupervised and supervised chemometric tools such as principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) on the NIR dataset resulted in a complete classification of the polymer classes. In addition, several kinds of PET (clear, blue, coloured, opaque, and boxes) were correctly classified as PET class, and PE samples with different branching degrees were properly separated.
This content is subject to copyright.
materials
Article
Miniaturized Near-Infrared (MicroNIR) Spectrometer
in Plastic Waste Sorting
Monika Rani 1, 2, , Claudio Marchesi 1, 2, , Stefania Federici 1, 2, * , Gianluca Rovelli 3,
Ivano Alessandri 2,4,5 , Irene Vassalini 2,4,5, Serena Ducoli 2,4, Laura Borgese 1,2,
Annalisa Zacco 1,2, Fabjola Bilo 1,2, Elza Bontempi 1,2 and Laura E. Depero 1,2
1Department of Mechanical and Industrial Engineering, University of Brescia, via Branze,
38-25123 Brescia, Italy
2Consorzio Interuniversitario Nazionale per la Scienza e Tecnologia dei Materiali (INSTM),
50121 Firenze, Italy
3ABCS Srl-Scientific Instruments & Materials Science, viale Regina Giovanna, 37-20129 Milan, Italy
4Department of Information Engineering, University of Brescia, via Branze, 38-25123 Brescia, Italy
5CNR-INO, Unit of Brescia, 25123 Brescia, Italy
*Correspondence: stefania.federici@unibs.it
These authors are equally contributed.
Received: 24 July 2019; Accepted: 23 August 2019; Published: 27 August 2019


Abstract:
Valorisation of the urban plastic waste in high-quality recyclates is an imperative challenge
in the new paradigm of the circular economy. In this scenario, a key role in the improvement of
the recycling process is exerted by the optimization of waste sorting. In spite of the enormous
developments achieved in the field of automated sorting systems, the quest for the reduction
of cross-contamination of incompatible polymers as well as a rapid and punctual sorting of the
unmatched polymers has not been suciently developed. In this paper, we demonstrate that a
miniaturized handheld near-infrared (NIR) spectrometer can be used to successfully fingerprint and
classify dierent plastic polymers. The investigated urban plastic waste comprised polyethylene (PE),
polypropylene (PP), poly(vinyl chloride) (PVC), poly(ethylene terephthalate) (PET), and poly(styrene)
(PS), collected directly in a recycling plastic waste plant, without any kind of sample washing or
treatment. The application of unsupervised and supervised chemometric tools such as principal
component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) on the NIR
dataset resulted in a complete classification of the polymer classes. In addition, several kinds of PET
(clear, blue, coloured, opaque, and boxes) were correctly classified as PET class, and PE samples with
dierent branching degrees were properly separated.
Keywords:
urban plastic waste; near-infrared (NIR) spectroscopy; chemometrics; principal
component analysis (PCA); partial least squares-discriminant analysis (PLS-DA); circular economy
1. Introduction
The huge amount of urban plastic waste and the continuous growth in human plastic consumption
require a high valorisation of the collected waste in the direction of a whole-system economic
sustainability. By the numbers, European plastic production reached almost 60 million tonnes in 2016,
and 335 million tonnes globally (Source: PlasticsEurope Market Research Group (PEMRG)/Conversio
Market & Strategy GmbH). In this framework, over 8.4 million tonnes of plastic waste were collected
in order to be recycled inside and outside the EU [
1
]. The recycling of polymer waste has valuable
environmental benefits due to the substitution of primary production, and a key role in the improvement
of the recycling process is exerted by the optimization of waste sorting [2,3].
Materials 2019,12, 2740; doi:10.3390/ma12172740 www.mdpi.com/journal/materials
Materials 2019,12, 2740 2 of 13
Generally, automated sorting processes are widespread in plastic recycling facilities [
4
].
These systems are usually based on vibrational spectroscopy techniques for polymer identification [
5
9
]
and camera systems for optical recognition of clear and coloured products [
10
,
11
]. Other sorting
technologies include UV–Vis spectroscopy [
12
] or mass spectroscopy (Py-GC/MS) for identification [
13
],
hyper-spectral imaging methods [
14
], X-ray detection for the separation of specific containers [
15
],
fluorescence spectroscopy for identifying halogens and heavy metals [
16
], and laser-induced plasma
spectroscopy to detect additives [
17
]. This approach has increased the purity of the output plastic
over the years, reaching a high percentage of recyclates in the production of secondary materials.
Nevertheless, there are severe limitations of these techniques in handling mixed plastic, which requires
additional sorting elsewhere and can aect the quality of the recyclate, if not properly assigned.
The positive cost/benefit balance can take place only if the separated fractions of polymers match a
high purity grade, fulfilling the market requirement of high recyclate quality. Therefore, a key step in
post-consumer recycling is reducing the cross-contamination of incompatible polymers [10].
Besides the automated sorting of the plastic stream, manual sorting is still necessary for the full
separation of the collected waste. However, manual sorting based only on plastic code recognition is
inecient and time consuming. In some cases, plastic goods can be badly crushed or have missing or
unreadable codes. Manufacturers can also use dierent polymers for the same product, depending on
marketing choices. In addition, the high turnover rate can be aected by an improperly trained or
unskilled workforce.
In this scenario, the possibility of exploiting a well-established technique for polymer identification
combined with a miniaturized, portable, low-cost, and real-time spectrometer for local and punctual
semi-automated sorting is highly desirable. Urban plastic waste normally contains plastics such as
polyethylene (PE), polypropylene (PP), poly(vinyl chloride) (PVC), poly(ethylene terephthalate) (PET),
and poly(styrene) (PS). These polymers have characteristic near-infrared (NIR) spectral fingerprints that
can be used for accurate sorting and separation [5]. The progress in miniaturization technologies has
generated specific instrumental features that have been reported in polymer identification applications,
but results have been reported only for a selection of polymers from commercial libraries, and are not
fully representative of real plastic waste [18].
In the present work, we challenged a compacted and miniaturized NIR spectrometer for a rapid
identification of urban plastic waste collected directly from a selection and recycling plant, without
any sample treatment. We collected a large amount of plastic samples directly from the plant and we
created a robust database of NIR spectra. Dierent chemometric tools were applied for data evaluation
of all the recorded samples. Principal component analysis (PCA) for preliminary data exploration and
partial least squares-discriminant analysis (PLS-DA) as supervised pattern recognition were applied to
fingerprint and classify the dierent plastic polymers.
2. Materials and Methods
2.1. Polymer Samples Collection
Plastic samples were collected at the selection division of the recovery and recycling plant Montello
SpA (Bergamo, Italy), which receives post-consumer plastic in the form of urban waste for recycling.
In total, 250 samples belonging to various classes of polymer were used (see Supplementary Materials
Figure S1). The polymers considered were: several kinds of PET (clear, blue, coloured, opaque,
and boxes), PE with dierent branching degrees (low-density polyethylene, LDPE, and high-density
polyethylene, HDPE), PVC, PP, and PS. The collection included bottles, containers, and packets of
dierent shapes, sizes, and colours. The samples were flattened or partially flattened at the sampling
site, and were not washed prior to testing. Black samples were not considered, due to the very low
reflectance in the NIR spectral region and the low signal-to-noise ratio of NIR sensors, which hamper
successful black polymer spectral acquisition [19].
Materials 2019,12, 2740 3 of 13
2.2. NIR Measurements
Spectra acquisition was realized using MicroNIR On-site (Viavi Solutions Inc., CA, United States
in reflectance mode without any sample preparation. The instrument is a miniaturized palm-sized
portable spectrometer of about 250 g in weight, less than 200 mm in length, and 50 mm in diameter.
This handheld spectrometer employs a linear variable filter (LVF) as dispersing element, in contrast
to traditional diraction-based spectrometers (see Figure 1for an instrument operating scheme).
The controlling parameters for spectral data collection were set at 10 ms integration time and 50 scans,
resulting in a short measurement time of 0.25 s. Not less than five replicates in dierent positions were
acquired by a point-and-shoot technique, in order to minimize eects driven by the non-uniformity of
samples. A total of 1303 NIR spectra were collected. A dark and a reference scan were carried out
approximately every 10 min. Data acquisition was realized through MicroNIR
TM
Pro v3.0 software
(Viavi Solutions Inc., CA, United States).
Materials 2019, 12, x FOR PEER REVIEW 3 of 13
Spectra acquisition was realized using MicroNIR On-site (Viavi Solutions Inc., CA, United
States in reflectance mode without any sample preparation. The instrument is a miniaturized
palm-sized portable spectrometer of about 250 g in weight, less than 200 mm in length, and 50 mm in
diameter. This handheld spectrometer employs a linear variable filter (LVF) as dispersing element,
in contrast to traditional diffraction-based spectrometers (see Figure 1 for an instrument operating
scheme). The controlling parameters for spectral data collection were set at 10 ms integration time
and 50 scans, resulting in a short measurement time of 0.25 s. Not less than five replicates in different
positions were acquired by a point-and-shoot technique, in order to minimize effects driven by the
non-uniformity of samples. A total of 1303 NIR spectra were collected. A dark and a reference scan
were carried out approximately every 10 min. Data acquisition was realized through MicroNIRTM
Pro v3.0 software (Viavi Solutions Inc., CA, United States).
Figure 1. (a) Miniaturized near-infrared (microNIR) device and (b) instrument operating scheme.
2.3. Signal Pretreatment
Signal correction methods are mathematical tools employed to remove additive and/or
multiplicative effects on spectroscopic data, which often affect the performance of chemometric
analysis [20]. These phenomena are typical of light scattering effects, which induce a photon loss
(addictive effect) and an increased path length (multiplicative effect), among others. These methods
are “row-wise”, and therefore the correction is made sample-by-sample. This is in contrast to mean
centring and autoscaling, which are “column-wise” treatments. In most cases, the combination of the
two types of method is required in order to make a satisfactory data analysis [21]. In this study,
spectra were collected into a single 1307 × 125 matrix (samples × scans), and several pre-processing
tools were applied to the raw data. The best result was obtained by applying Savitzky-Golay
second-derivative procedure of seven data points and a second-order polynomial followed by
standard normal variate (SNV). Moreover, after the preprocessing, spectra were normalized by
mean centring.
2.4. Data Analysis
Different chemometric tools were applied in order to have a correct data evaluation of all the
analysed samples. The first step consisted of an exploratory analysis by PCA aiming to explore the
data structure. PCA was carried out with a total of 1303 NIR spectra for all five polymer classes.
Subsequently, PLS-DA was used as supervised pattern recognition with the aim of separating the
different commodities, and it was performed on a total of 1287 NIR spectra (16 spectra were
Figure 1. (a) Miniaturized near-infrared (microNIR) device and (b) instrument operating scheme.
2.3. Signal Pretreatment
Signal correction methods are mathematical tools employed to remove additive and/or
multiplicative eects on spectroscopic data, which often aect the performance of chemometric
analysis [
20
]. These phenomena are typical of light scattering eects, which induce a photon loss
(addictive eect) and an increased path length (multiplicative eect), among others. These methods
are “row-wise”, and therefore the correction is made sample-by-sample. This is in contrast to mean
centring and autoscaling, which are “column-wise” treatments. In most cases, the combination of the
two types of method is required in order to make a satisfactory data analysis [
21
]. In this study, spectra
were collected into a single 1307
×
125 matrix (samples
×
scans), and several pre-processing tools were
applied to the raw data. The best result was obtained by applying Savitzky-Golay second-derivative
procedure of seven data points and a second-order polynomial followed by standard normal variate
(SNV). Moreover, after the preprocessing, spectra were normalized by mean centring.
2.4. Data Analysis
Dierent chemometric tools were applied in order to have a correct data evaluation of all the
analysed samples. The first step consisted of an exploratory analysis by PCA aiming to explore the
data structure. PCA was carried out with a total of 1303 NIR spectra for all five polymer classes.
Subsequently, PLS-DA was used as supervised pattern recognition with the aim of separating the
Materials 2019,12, 2740 4 of 13
dierent commodities, and it was performed on a total of 1287 NIR spectra (16 spectra were removed as
outliers). All calculations were performed using MATLAB (R2018a) with chemometric toolboxes [
22
,
23
].
3. Results and Discussion
NIR spectroscopy has the key advantage of being a rapid-response analytical tool, recording
spectra with no prior manipulation and predicting physical and chemical parameters from a single
spectrum [
24
]. In the NIR region, the absorption bands occur due to overtones or combination bands
of mainly carbon–hydrogen vibrations and oxygen-hydrogen vibrations. A correct band assignment is
challenging because it may be due to dierent combinations of fundamental vibrations. Moreover,
overtone vibrations are highly overlapped [25].
The NIR reflectance spectra of the five polymers (i.e., PET, PP, PVC, PE, and PS) in the range
900–1700 nm, obtained by the MicroNIR On-site, are shown in Figure 2.
Materials 2019, 12, x FOR PEER REVIEW 4 of 13
removed as outliers). All calculations were performed using MATLAB (R2018a) with chemometric
toolboxes [22,23].
3. Results and Discussion
NIR spectroscopy has the key advantage of being a rapid-response analytical tool, recording
spectra with no prior manipulation and predicting physical and chemical parameters from a single
spectrum [24]. In the NIR region, the absorption bands occur due to overtones or combination bands
of mainly carbon–hydrogen vibrations and oxygen-hydrogen vibrations. A correct band assignment
is challenging because it may be due to different combinations of fundamental vibrations. Moreover,
overtone vibrations are highly overlapped [25].
The NIR reflectance spectra of the five polymers (i.e., PET, PP, PVC, PE, and PS) in the range
900–1700 nm, obtained by the MicroNIR On-site, are shown in Figure 2.
Figure 2. Near-infrared (NIR) spectra of five classes of plastics: (a) representative raw spectra of the
five classes; (b) corresponding spectra after pre-treatment by second derivative and standard normal
variate (SNV). PE: polyethylene; PET: poly(ethylene terephthalate); PP: polypropylene; PS:
poly(styrene); PVC: poly(vinyl chloride).
In the absorbance spectra (Figure 2a), the principal absorbance band for PET was found at 1660
nm due to the first overtone of CH stretching [26], with two other small peaks around 1130 nm and
1415 nm. In the case of PVC, this main peak was shifted and was not recorded in the instrumental
operating window, and only the two small peaks around 1190 nm and 1420 nm were seen [26]. The
same behaviour appeared for PE and PP, assigning two bands around 1210 nm and 1430 nm for PE
and around 1190 nm and 1400 nm for PP. PS exhibited a dominant band around 1675 nm and two
small peaks around 1140 nm and 1205 nm [5]. Figure 2b refers to the corresponding spectra shown in
Figure 2a after pre-treatment by second derivative and SNV. In general, by comparing the spectra of
samples made of one type of polymer, a variation in the absorbance values was noted and this was
ascribed to variations in sample thickness [5]. However, the overall form of the spectra was
preserved, concluding that dust or liquid contaminations randomly characterizing each sample did
not affect the quality of NIR spectra (see Appendix A, Figure A1) [27]. In addition, NIR spectra of
samples of different colours did not display considerable differences (Figure A1), meaning that the
colorants present in the plastic resin matrix did not significantly influence the NIR spectra.
Moreover, no differences in band positions were found when studying the spectra of several kinds
of PET (clear, coloured, opaque, box, and blue) (see Appendix A, Figure A2), implying that the
spectra can be successfully exploited to discriminate PET samples, despite their differences. For this
reason, hereinafter, only one PET class is considered.
3.1. Principal Component Analysis
Principal component analysis is a very powerful chemometric tool for analysing data structure
[28]. The aim of PCA is to extract the information encoded in a certain number of variables into a
Figure 2.
Near-infrared (NIR) spectra of five classes of plastics: (
a
) representative raw spectra of the five
classes; (
b
) corresponding spectra after pre-treatment by second derivative and standard normal variate
(SNV). PE: polyethylene; PET: poly(ethylene terephthalate); PP: polypropylene; PS: poly(styrene); PVC:
poly(vinyl chloride).
In the absorbance spectra (Figure 2a), the principal absorbance band for PET was found at 1660 nm
due to the first overtone of C–H stretching [
26
], with two other small peaks around 1130 nm and
1415 nm. In the case of PVC, this main peak was shifted and was not recorded in the instrumental
operating window, and only the two small peaks around 1190 nm and 1420 nm were seen [
26
].
The same behaviour appeared for PE and PP, assigning two bands around 1210 nm and 1430 nm for PE
and around 1190 nm and 1400 nm for PP. PS exhibited a dominant band around 1675 nm and two
small peaks around 1140 nm and 1205 nm [
5
]. Figure 2b refers to the corresponding spectra shown in
Figure 2a after pre-treatment by second derivative and SNV. In general, by comparing the spectra of
samples made of one type of polymer, a variation in the absorbance values was noted and this was
ascribed to variations in sample thickness [
5
]. However, the overall form of the spectra was preserved,
concluding that dust or liquid contaminations randomly characterizing each sample did not aect the
quality of NIR spectra (see Appendix A, Figure A1) [
27
]. In addition, NIR spectra of samples of dierent
colours did not display considerable dierences (Figure A1), meaning that the colorants present in
the plastic resin matrix did not significantly influence the NIR spectra. Moreover, no dierences in
band positions were found when studying the spectra of several kinds of PET (clear, coloured, opaque,
box, and blue) (see Appendix A, Figure A2), implying that the spectra can be successfully exploited to
discriminate PET samples, despite their dierences. For this reason, hereinafter, only one PET class
is considered.
Materials 2019,12, 2740 5 of 13
3.1. Principal Component Analysis
Principal component analysis is a very powerful chemometric tool for analysing data structure [
28
].
The aim of PCA is to extract the information encoded in a certain number of variables into a smaller
set of new orthogonal variables called principal components. PCA calculation was performed over
the entire spectral region and using all collected spectra. The first three components accounted for
87.11% of the total variability (PC1 56.75%, PC2 15.63%, and PC3 14.73%). Figure 3displays the score
plots of PC1 vs. PC2 (Figure 3a) and PC1 vs. PC3 (Figure 3b). In the score plot displayed in Figure 3a,
a clear separation between the five commodities can be seen. Indeed, in the new space determined
by the first two components, polymers formed very tight and homogeneous clusters. Along the first
component, a clear dierentiation between PET, PVC, PP, and PE was highlighted, while the second
component allowed the dierentiation of PS from the other polymers, and PE from PP. On the other
hand, Figure 3b shows the score plot of PC1 vs. PC3. In this case PET was dierentiated from the other
polymers along PC1, while PC3 distinguished between the remaining sets of polymers.
Materials 2019, 12, x FOR PEER REVIEW 5 of 13
smaller set of new orthogonal variables called principal components. PCA calculation was
performed over the entire spectral region and using all collected spectra. The first three components
accounted for 87.11% of the total variability (PC1 56.75%, PC2 15.63%, and PC3 14.73%). Figure 3
displays the score plots of PC1 vs. PC2 (Figure 3a) and PC1 vs. PC3 (Figure 3b). In the score plot
displayed in Figure 3a, a clear separation between the five commodities can be seen. Indeed, in the
new space determined by the first two components, polymers formed very tight and homogeneous
clusters. Along the first component, a clear differentiation between PET, PVC, PP, and PE was
highlighted, while the second component allowed the differentiation of PS from the other polymers,
and PE from PP. On the other hand, Figure 3b shows the score plot of PC1 vs. PC3. In this case PET
was differentiated from the other polymers along PC1, while PC3 distinguished between the
remaining sets of polymers.
Figure 3. Results of principal components analysis carried out with spectral data of the different
commodities. (a) The score plot of the first two components is shown, as well as (b) the score plot of
PC1 vs. PC3 EV: explained variance.
Figure 3.
Results of principal components analysis carried out with spectral data of the dierent
commodities. (
a
) The score plot of the first two components is shown, as well as (
b
) the score plot of
PC1 vs. PC3 EV: explained variance.
Materials 2019,12, 2740 6 of 13
Furthermore, the results of PCA applied to the restricted dataset of PET samples showed no
separation among the five kinds of PET samples (Figure 4), confirming the interpretation of the band
assignment of NIR spectra.
Materials 2019, 12, x FOR PEER REVIEW 6 of 13
Furthermore, the results of PCA applied to the restricted dataset of PET samples showed no
separation among the five kinds of PET samples (Figure 4), confirming the interpretation of the band
assignment of NIR spectra.
Figure 4. Principal component analysis (PCA) analysis of several kinds of poly(ethylene
terephthalate) (PET) samples for a total of 659 spectra considered. EV: explained variance.
3.2. HDPE versus LDPE
The discrimination between HDPE and LDPE is an important issue in terms of plastic recycling
[3]. PE is largely used in the food packaging sector, thanks to its remarkable properties in terms of
mechanical and optical performances, water-vapor-resistance, and heat-sealing attitude. The
amount of side branches that are attached to the main PE polymer chain determines the
morphological properties, which in turn affects optical, physical, and thermal properties [29,30]. NIR
spectroscopy successfully discriminates between HDPE and LDPE [31,32], even if severe limitations
can occur depending on the characteristics of the measuring instruments [27]. In the sampling
campaign, we collected 25 samples of HDPE and 25 samples of LDPE. Raw NIR spectra did not
show any significant feature to discriminate the two plastics, nor did a clear separation in the score
plots resulting from the previous PCA analysis (Figure 3). However, by limiting the PCA analysis to
the PE samples, a separation between HDPE and LDPE was obtained; the first two components
explicated 61.22% of the total variability (PC1 43.22%, PC2 18.00%), confirming the successful
application of NIR spectroscopy in discriminating these two polymers. Figure 5 displays the score
plots of PC1 vs. PC2.
Figure 4.
Principal component analysis (PCA) analysis of several kinds of poly(ethylene terephthalate)
(PET) samples for a total of 659 spectra considered. EV: explained variance.
3.2. HDPE versus LDPE
The discrimination between HDPE and LDPE is an important issue in terms of plastic recycling [
3
].
PE is largely used in the food packaging sector, thanks to its remarkable properties in terms of
mechanical and optical performances, water-vapor-resistance, and heat-sealing attitude. The amount of
side branches that are attached to the main PE polymer chain determines the morphological properties,
which in turn aects optical, physical, and thermal properties [
29
,
30
]. NIR spectroscopy successfully
discriminates between HDPE and LDPE [
31
,
32
], even if severe limitations can occur depending on the
characteristics of the measuring instruments [
27
]. In the sampling campaign, we collected 25 samples of
HDPE and 25 samples of LDPE. Raw NIR spectra did not show any significant feature to discriminate
the two plastics, nor did a clear separation in the score plots resulting from the previous PCA analysis
(Figure 3). However, by limiting the PCA analysis to the PE samples, a separation between HDPE and
LDPE was obtained; the first two components explicated 61.22% of the total variability (PC1 43.22%,
PC2 18.00%), confirming the successful application of NIR spectroscopy in discriminating these two
polymers. Figure 5displays the score plots of PC1 vs. PC2.
Materials 2019,12, 2740 7 of 13
Materials 2019, 12, x FOR PEER REVIEW 7 of 13
Figure 5. Results of principal components analysis carried out with spectral data of polyethylene (PE)
class polymers. Score plot of the first two components is shown.
3.3. Partial Least Squares-Discriminant Analysis
After the exploratory PCA analysis, a supervised classification tool was applied in order to
distinguish the different plastic classes. In PLS-DA, the well-known PLS regression algorithm is
modified with a classification goal. The response variable is categorical, reflecting the belonging
class of the statistical units. PLS-DA returns the prediction in a vector of size equal to the number of
classes in the predictor variables, with values ranging from 0 to 1 [33].
Validation tools are very useful methods in chemometrics, and they are used to verify the
capability of the model prediction. Before the model calibration, data were split into training and test
sets containing 901 and 386 samples, respectively. The training set was used to calibrate the model
both in fitting and in cross-validation, while the test set was employed only at the end of the
procedure in order to evaluate the true predictive capability of the calibrated model [22].
During the calibration, the selection of the optimal number of latent variables (LVs) is a crucial
point, which is performed based on the cross-validation procedure [22]. This should allow the
optimization of the complexity of the multivariate model according to the predictive capacity of the
model itself. Cross-validation is generally carried out by dividing the training set into different
cross-validation groups, and during each round one group is removed from the training set. Over
the rounds, the model is calibrated on the remaining training samples and then used to predict
samples of the cross-validation group. The cross-validation procedure was based on a Venetian
blinds approach with 10 groups. In the Venetian blinds procedure, the calibration group and the
cross-validation group are selected by choosing every n-th sample from the dataset starting from the
first one [22]. The selection of the number of cross-validation groups was set to 10, as a reasonable
number in order to avoid the so-called overfitting” (i.e., the overestimation of the predictive
capability of the model). Figure 6 shows the error rate (Figure 6a) and the degree of not assigned
samples as a function of the number of latent variables (Figure 6b). The optimal number of LVs was
set to 6 (explaining 94% of the total variability and 0.0099% of not assigned samples), which was
associated with the minimum error rate and, simultaneously, with the minimum percentage of not
assigned samples. Four LVs reached the same error rate, but 0.019% of not assigned samples.
Figure 5.
Results of principal components analysis carried out with spectral data of polyethylene (PE)
class polymers. Score plot of the first two components is shown.
3.3. Partial Least Squares-Discriminant Analysis
After the exploratory PCA analysis, a supervised classification tool was applied in order to
distinguish the dierent plastic classes. In PLS-DA, the well-known PLS regression algorithm is
modified with a classification goal. The response variable is categorical, reflecting the belonging class
of the statistical units. PLS-DA returns the prediction in a vector of size equal to the number of classes
in the predictor variables, with values ranging from 0 to 1 [33].
Validation tools are very useful methods in chemometrics, and they are used to verify the capability
of the model prediction. Before the model calibration, data were split into training and test sets
containing 901 and 386 samples, respectively. The training set was used to calibrate the model both in
fitting and in cross-validation, while the test set was employed only at the end of the procedure in
order to evaluate the true predictive capability of the calibrated model [22].
During the calibration, the selection of the optimal number of latent variables (LVs) is a crucial
point, which is performed based on the cross-validation procedure [
22
]. This should allow the
optimization of the complexity of the multivariate model according to the predictive capacity of
the model itself. Cross-validation is generally carried out by dividing the training set into dierent
cross-validation groups, and during each round one group is removed from the training set. Over the
rounds, the model is calibrated on the remaining training samples and then used to predict samples of
the cross-validation group. The cross-validation procedure was based on a Venetian blinds approach
with 10 groups. In the Venetian blinds procedure, the calibration group and the cross-validation
group are selected by choosing every n-th sample from the dataset starting from the first one [
22
].
The selection of the number of cross-validation groups was set to 10, as a reasonable number in order
to avoid the so-called “overfitting” (i.e., the overestimation of the predictive capability of the model).
Figure 6shows the error rate (Figure 6a) and the degree of not assigned samples as a function of the
number of latent variables (Figure 6b). The optimal number of LVs was set to 6 (explaining 94% of the
total variability and 0.0099% of not assigned samples), which was associated with the minimum error
rate and, simultaneously, with the minimum percentage of not assigned samples. Four LVs reached
the same error rate, but 0.019% of not assigned samples.
Materials 2019,12, 2740 8 of 13
Materials 2019, 12, x FOR PEER REVIEW 8 of 13
Figure 6. Error rate (a) and not assigned samples (b) as a function of latent variables calculated in the
partial least squares-discriminant analysis (PLS-DA) model. Six latent variables (LVs) was the
optimal number, marked in red, with 0.0099% of not assigned samples.
After the determination of the optimal number of LVs, the calculation of the PLS-DA model was
performed selecting six latent variables and 10 cross-validation groups for internal validation. In
Table 1, the two confusion matrices achieved in fitting and in cross-validation were presented. Both
in fitting and in cross-validation, the performance of the model was very high, proved by the
correlation between observed and predicted classes. All the investigated samples were correctly
designated, and only four samples in fitting and five in cross-validation were not assigned.
Table 1. Confusion matrices obtained from the PLS-DA model, both in fitting and in cross-validation
(based on Venetian blinds with 10 groups). The “Not Assignedcolumn contains samples which
were not assigned to any of the considered classes.
Experimental
Class
Calculated Class
PET PE PP PVC PS Not
Assigned
Fitting - - - - - -
PET 459 0 0 0 0 2
PE 0 181 0 0 0 1
PP 0 0 82 0 0 0
PVC 0 0 0 84 0 1
PS 0 0 0 0 91 0
Cross-Validation - - - - - -
PET 460 0 0 0 0 1
PE 0 181 0 0 0 1
PP 0 0 82 0 0 0
PVC 0 0 0 82 0 3
PS 0 0 0 0 91 0
The performance of the classification model could be also evaluated by considering the
classification parameters derived from the confusion matrix: sensitivity (Sn), specificity (Sp), and
Figure 6.
Error rate (
a
) and not assigned samples (
b
) as a function of latent variables calculated in the
partial least squares-discriminant analysis (PLS-DA) model. Six latent variables (LVs) was the optimal
number, marked in red, with 0.0099% of not assigned samples.
After the determination of the optimal number of LVs, the calculation of the PLS-DA model
was performed selecting six latent variables and 10 cross-validation groups for internal validation.
In Table 1, the two confusion matrices achieved in fitting and in cross-validation were presented.
Both in fitting and in cross-validation, the performance of the model was very high, proved by the
correlation between observed and predicted classes. All the investigated samples were correctly
designated, and only four samples in fitting and five in cross-validation were not assigned.
Table 1.
Confusion matrices obtained from the PLS-DA model, both in fitting and in cross-validation
(based on Venetian blinds with 10 groups). The “Not Assigned” column contains samples which were
not assigned to any of the considered classes.
Experimental Class Calculated Class
PET PE PP PVC PS Not Assigned
Fitting - - - - - -
PET 459 0 0 0 0 2
PE 0 181 0 0 0 1
PP 0 0 82 0 0 0
PVC 0 0 0 84 0 1
PS 0 0 0 0
91
0
Cross-Validation - - - - - -
PET 460 0 0 0 0 1
PE 0 181 0 0 0 1
PP 0 0 82 0 0 0
PVC 0 0 0 82 0 3
PS 0 0 0 0
91
0
The performance of the classification model could be also evaluated by considering the classification
parameters derived from the confusion matrix: sensitivity (Sn), specificity (Sp), and non-error rate
(NER). These parameters represent the ability of the model to correctly identify the samples of the
Materials 2019,12, 2740 9 of 13
given class, the ability of a classifier to reject the samples of other classes, and the average of the class
sensitivities, respectively [
34
]. Table 2collects the parameters of the PLS-DA model. In both fitting
and cross-validation, NER, Sn, and Sp were equal to 1, meaning that 100% of the polymers were
correctly classified.
Table 2.
Classification parameters (non-error rate (NER), class sensitivity (Sn), and specificity (Sp))
calculated in fitting and in cross-validation.
-- PET PE PP PVC PS
NER Sn Sp Sn Sp Sn Sp Sn Sp Sn Sp
Fitting 1 1 1 1 1 1 1 1 1 1 1
Cross-validation 1 1 1 1 1 1 1 1 1 1 1
Furthermore, we analysed the confusion matrix and the classification parameters for the test set,
used as an additional trial to validate the model. Table 3shows the confusion matrix derived from
the test set; again, 100% of the samples were classified, and only four samples were not assigned.
Table 4summarizes the classification parameters obtained from the confusion matrix, with the NER,
Sn, and Sp of each class equal to 1.
Table 3.
Confusion matrix obtained from the PLS-DA model fitted on the validation set. The “Not
Assigned” column contains samples which were not assigned to any of the considered classes.
Experimental Class Calculated Class
PET PE PP PVC PS Not Assigned
PET 197 0 0 0 0 1
PE 0 78 0 0 0 1
PP 0 0 33 0 0 1
PVC 0 0 0 35 0 1
PS 0 0 0 0
39
0
Table 4.
Classification parameters (non-error rate (NER), class sensitivity (Sn), and specificity (Sp))
calculated on the validation set.
-- PET PE PP PVC PS
NER Sn Sp Sn Sp Sn Sp Sn Sp Sn Sp
Test 1 1 1 1 1 1 1 1 1 1 1
In the case of the restricted data set composed of HDPE and LDPE spectra, the optimal number
of LVs was set to six, assigning 90% of total variability (see Supplementary Materials, Figure S2).
With respect to the analysis of the polymer classes, the degree of not assigned samples as a function
of the number of latent variables was equal to zero, meaning that all the samples were assigned to a
class. After the determination of the optimal number of LVs, the calculation of the PLS-DA model
was performed, selecting six latent variables and 10 cross-validation groups for internal validation.
Both in fitting and in cross-validation, all the investigated samples were correctly designated (see
Table S1 in the Supplementary Materials for the two confusion matrices achieved in fitting and in
cross-validation).
4. Conclusions
In the present study, the use of a handheld NIR spectrometer combined with robust chemometrics
tools to fingerprint dierent urban plastic waste collected directly from a recycling plant was presented.
Plastic samples included clear, blue, coloured, opaque, and boxes of PET, PE with dierent branching
Materials 2019,12, 2740 10 of 13
degrees (HDPE and LDPE), PP, PVC, and PS. In a first step, PCA was applied to NIR spectra of all
polymer classes, and polymers were correctly separated in clusters. Moreover, by limiting the analysis
on PE samples, HDPE and LDPE were correctly discriminated.
In addition, by comparing samples of dierent colours, no dierences in band positions were
found, suggesting that colorants present in the plastic resin matrix do not significantly influence NIR
spectra. Moreover, even if samples were not washed or treated prior to the analysis, NIR spectra were
not aected by dust contamination or liquid residuals, meaning a successful application directly to the
urban plastic waste.
After the exploratory PCA analysis, PLS-DA was applied. Before the model calibration, data were
split into a training and test set, using the training set to calibrate the model. In both fitting and
cross-validation, the performance of the model was very high, proved by the correlation between
observed and predicted classes. All the analysed samples were also correctly designated in the case of
the restricted data set composed of HDPE and LDPE spectra.
The results show promising outcomes in order to improve the reliability and eciency of the
manual sorting of plastic waste, increasing the volume of recycled plastic and the management of
multi-component polymer types.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1944/12/17/2740/s1.
Figure S1: Systematic organization of the collected urban plastic waste for laboratory measurements, Figure S2:
Error rate as a function of latent variables calculated in the PLS-DA model for HDPE and LDPE classification. Six
LVs was the optimal number, Table S1: Confusion matrices obtained from PLS-DA model, both in fitting and in
cross-validation (based on Venetian blinds with 10 groups) restricted to the PE class.
Author Contributions:
Conceptualization, S.F. and L.E.D.; Funding acquisition, G.R., E.B., and L.E.D.;
Investigation, M.R., C.M., I.V., S.D., A.Z., and F.B.; Resources, E.B. and L.E.D.; Supervision, S.F. and L.E.D.;
Writing—original draft, M.R., C.M., and S.F.; Writing—review and editing, S.F., I.A., L.B., and L.E.D.
Funding:
This research was funded by PON “R&I” 2014-2020: SIRIMAP—SIstemi di Rilevamento
dell’Inquinamento MArino da Plastiche e successivo recupero-riciclo (No. ARS01_01183) CUP D86C18000520008.
Conflicts of Interest: The authors declare no conflicts of interest.
Appendix A
Appendix A.1 Evaluation of Sample Contaminations and Colour Influence
Figure A1 shows a representative collection of NIR spectra of coloured PET samples. The coloured
PET batch was chosen as a demonstrator set to compare samples with random contaminations and
to evaluate the presence of colorants. A total of 136 spectra were collected (25 samples with at least
5 replicates in dierent positions for each sample). The samples obtained from the recycling plant
were flattened or partially flattened directly on site and were not washed prior the testing. Liquid
residuals, pieces of labels, and environmental dust were largely present on the surface and inside the
samples. Nevertheless, NIR spectra showed only a variation in the total absorbance values, attributable
to variations in sample thickness, according to Lambert–Beer law [
35
]. Moreover, no dierences in
band positions were found despite the dierent colours of the commodities.
Materials 2019,12, 2740 11 of 13
Figure A1. Representative NIR spectra of coloured PET samples.
Appendix A.2 PET Samples
Several kinds of PET were considered. In particular, clear, coloured, opaque, blue, and box samples
of PET were collected at the recycling plant (25 samples for each kind). In order to investigate the NIR
contribution of the dierent kinds, we acquired at least five replicates for each sample (for a total of 659
spectra) to make spectra comparison and multivariate analysis. Figure A2 shows representative NIR
spectra of a restriction of two random samples for each kind, in order to better visualize the spectra in
the plot. The overall form of the spectra was preserved and no significant dierences are noted.
Materials 2019, 12, x FOR PEER REVIEW 11 of 13
Figure A1. Representative NIR spectra of coloured PET samples.
Appendix A.2. PET Samples
Several kinds of PET were considered. In particular, clear, coloured, opaque, blue, and box
samples of PET were collected at the recycling plant (25 samples for each kind). In order to
investigate the NIR contribution of the different kinds, we acquired at least five replicates for each
sample (for a total of 659 spectra) to make spectra comparison and multivariate analysis. Figure A2
shows representative NIR spectra of a restriction of two random samples for each kind, in order to
better visualize the spectra in the plot. The overall form of the spectra was preserved and no
significant differences are noted.
Figure A2. Representative NIR spectra of two random samples for each kind of PET commodity.
References
1. Plastics Europe. Available online: http://www.plasticseurope.org (accessed on 24 August 2019).
2. Al-Salem, S.M.; Lettieri, P.; Baeyens, J. Recycling and recovery routes of plastic solid waste (PSW): A
review. Waste Manag. 2009, 29, 2625–2643.
Figure A2. Representative NIR spectra of two random samples for each kind of PET commodity.
Materials 2019,12, 2740 12 of 13
References
1. Plastics Europe. Available online: http://www.plasticseurope.org (accessed on 19 June 2019).
2.
Al-Salem, S.M.; Lettieri, P.; Baeyens, J. Recycling and recovery routes of plastic solid waste (PSW): A review.
Waste Manag. 2009,29, 2625–2643. [CrossRef] [PubMed]
3.
Siddique, R.; Khatib, J.; Kaur, I. Use of recycled plastic in concrete: A review. Waste Manag.
2008
,28,
1835–1852. [CrossRef] [PubMed]
4.
Gundupalli, S.P.; Hait, S.; Thakur, A. A review on automated sorting of source-separated municipal solid
waste for recycling. Waste Manag. 2017,60, 56–74. [CrossRef] [PubMed]
5.
Masoumi, H.; Safavi, S.M.; Khani, Z. Identification and classification of plastic resins using near infrared
reflectance spectroscopy. Int. J. Mech. Mechatron. Eng. 2012,6, 877–884.
6.
Masson, J.; Pelletier, L.; Collins, P. Rapid FTIR method for quantification of styrene-butadiene type copolymers
in bitumen. J. Appl. Polym. Sci. 2001,79, 1034–1041. [CrossRef]
7.
Allen, V.; Kalivas, J.H.; Rodriguez, R.G. Post-consumer plastic identification using raman spectroscopy. Appl.
Spectrosc. 1999,53, 672–681. [CrossRef]
8.
Chalmers, J.M.; Everall, N.J.; Measurement, I.C.I. Handbook of Vibrational Spectroscopy: Qualitative and
Quantitative Analysis of Plastics, Polymers and Rubbers by Vibrational Spectroscopy; John Wiley & Sons, Ltd.:
Hoboken, NJ, USA, 2007; ISBN 9780470027325.
9.
Florestan, J.; Lachambre, A.; Mermilliod, N.; Boulou, J.C.; Marfisi, C. Recycling of plastics: Automatic
identification of polymers by spectroscopic methods. Resour. Conserv. Recycl. 1994,10, 67–74. [CrossRef]
10.
Hopewell, J.; Dvorak, R.; Kosior, E. Plastics recycling: Challenges and opportunities. Philos. Trans. R. Soc. B
Biol. Sci. 2009,364, 2115–2126. [CrossRef]
11.
Tachwali, Y.; Al-Assaf, Y.; Al-Ali, A.R. Automatic multistage classification system for plastic bottles recycling.
Resour. Conserv. Recycl. 2007,52, 266–285. [CrossRef]
12.
Safavi, S.M.; Masoumi, H.; Mirian, S.S.; Tabrizchi, M. Sorting of polypropylene resins by color in MSW using
visible reflectance spectroscopy. Waste Manag. 2010,30, 2216–2222. [CrossRef]
13.
Herrera, M.; Matuschek, G.; Kettrup, A. Fast identification of polymer additives by pyrolysis-gas
chromatography/mass spectrometry. J. Anal. Appl. Pyrol. 2003,70, 35–42. [CrossRef]
14.
Ulrici, A.; Serranti, S.; Ferrari, C.; Cesare, D.; Foca, G.; Bonifazi, G. Ecient chemometric strategies for
PET-PLA discrimination in recycling plants using hyperspectral imaging. Chemometr. Intell. Lab.
2013
,122,
31–39. [CrossRef]
15.
Brunner, S.; Fomin, P.; Kargel, C. Automated sorting of polymer flakes: Fluorescence labeling and development
of a measurement system prototype. Waste Manag. 2015,38, 49–60. [CrossRef]
16.
Ernst, T.; Popp, R.; Van Eldik, R. Quantification of heavy metals for the recycling of waste plastics from
electrotechnical applications. Talanta 2000,53, 347–357. [CrossRef]
17.
Anzano, J.M.; Gornushkin, I.B.; Smith, B.W.; Winefofdner, J.D. Laser-Induced Plasma Spectroscopy for Plastic
Identification. Polym. Eng. Sci. 2000,40, 2423–2429. [CrossRef]
18.
Yan, H.; Siesler, H.W. Identification performance of dierent types of handheld near-infrared (NIR)
spectrometers for the recycling of polymer commodities. Appl. Spectrosc. 2018,72, 1362–1370. [CrossRef]
19.
Rozenstein, O.; Puckrin, E.; Adamowski, J. Development of a new approach based on midwave infrared
spectroscopy for post-consumer black plastic waste sorting in the recycling industry. Waste Manag.
2017
,68,
38–44. [CrossRef]
20.
Jansen, J.J.; Downey,G.; Blanchet, L.; Engel, J.; Gerretzen, J.; Szyman, E. Breaking with trends in pre-processing?
Trend. Anal. Chem. 2013,50, 96–106.
21.
Rinnan, Å.; van den Berg, F.; Engelsen, S.B. Review of the most common pre-processing techniques for
near-infrared spectra. TrAC-Trend. Anal. Chem. 2009,28, 1201–1222. [CrossRef]
22.
Ballabio, D.; Consonni, V. Classification tools in chemistry. Part 1: Linear models. PLS-DA. Anal. Methods
2013,5, 3790. [CrossRef]
23.
Ballabio, D. A Matlab toolbox for principal component analysis and unsupervised exploration of data
structure. Chemometr. Intell. Lab. 2015,149, 1–9. [CrossRef]
24.
Blanco, M.; Villarroya, I. NIR spectroscopy: A rapid-response analytical tool. TrAC-Trend. Anal. Chem.
2002
,
21, 240–250. [CrossRef]
Materials 2019,12, 2740 13 of 13
25.
Weyer, L.; Workman, J., Jr. Practical Guide and Spectral Atlas for Interpretive Near-Infrared Spectroscopy;
Workman, J., Jr., Weyer, L., Eds.; CRC Press: Boca Raton, FL, USA, 2012.
26.
Crandall, E.W.; Jagtap, A.N. The near-infrared spectra of polymers. J. Appl. Polym. Sci.
1977
,21, 449–454.
[CrossRef]
27.
Feldho, R.; Wienke, D.; Cammann, K.; Fuchs, H. On-line post consumer package identification by NIR
spectroscopy combined with a FuzzyARTMAP classifier in an industrial environment. Appl. Spectrosc.
1997
,
51, 362–368. [CrossRef]
28. Bro, R.; Smilde, A.K. Principal component analysis. Anal. Methods 2014,6, 2812–2831. [CrossRef]
29.
Martinez-Salazar, J.; S
á
nchez Cuesta, M.; Balt
á
Calleja, F.J. Equilibrium and non-equilibrium melting of
branched polyethylene relating to defect incorporation within crystals. Colloid Polym. Sci.
1987
,265, 239–245.
[CrossRef]
30.
Mart
í
nez-Salazar, J.; Cuesta, M.S.; Plans, J. On phase separation in high- and low-density polyethylene
blends: 1. Melting-point depression analysis. Polymer (Guildf) 1991,32, 2984–2988. [CrossRef]
31.
Miller, C.E. Use of near-infrared spectroscopy to determine the composition of high-density/low-density
polyethylene blend films. Appl. Spectrosc. 1993,47, 222–228. [CrossRef]
32.
De Biasio, M.; Arnold, T.; McGunnigle, G.; Leitner, R.; Balthasar, D.; Rehrmann, V. Detecting and
discriminating PE and PP polymers for plastics recycling using NIR imaging spectroscopy. In Proceedings of
the Thermosense XXXII, Orlando, FL, USA, 3 May 2010; Volume 7661, p. 76610V.
33. Barker, M.; Rayens, W. Partial least squares for discrimination. J. Chemom. 2003,17, 166–173. [CrossRef]
34.
Ballabio, D.; Grisoni, F.; Todeschini, R. Multivariate comparison of classification performance measures.
Chemometr. Intell. Lab. 2018,174, 33–44. [CrossRef]
35.
Willard, H.H.; Merritt, L.L.J.; Dean, J.A.; Settle, F.A.J. Instrumental Methods of Analysis, 7th ed.; Wadsworth
Pub Co.: Belmont, CA, USA, 1988.
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
... The use of a handheld NIR spectrometer for the classification of diverse plastic monolayer materials (simple measurements) has already been researched by various authors like Rani et al. and Yang et al. 40,41 In contrast, the clear identification of multilayer plastic materials is more complex, especially if the sample is only to be measured from one side, which has not yet been researched. Furthermore, the low material thickness of films poses an additional challenge as it leads to lower spectral quality, which is crucial for sufficient classification. ...
... 62 PCA aims to extract the information encoded in a certain number of variables into a smaller set of new orthogonal variables called principal components (PC). 40 The PCs are created in order to preserve as much variability as possible. They are new variables that are linear functions of the variables in the original dataset that successively maximize variance and are not correlated with each other. ...
Article
Full-text available
The low thickness of plastic films poses a challenge when using near-infrared (NIR) spectroscopy as it affects the spectral qual-ity and classification. This research focuses on offering a solution to the challenge of classifying multilayer plastic film materialswith a focus on polyolefin multilayer plastics. It presents the importance of spectral quality on accurate classification. The aimis to demonstrate the suitability of the handheld NIR spectrometer in classifying multilayer polyolefin films and assess theimpact of various measuring backgrounds (white tile, Teflon, aluminum, copper, silver, and gold) on classification accuracyin the wavelength range of 1596–2396 nm. Metallic backgrounds have been found to enhance spectral quality and classificationaccuracy. The classification accuracy was consistently high, ranging from 96.55% to 100%, with aluminum and gold back-grounds yielding the best results in theoretical accuracy. In experimental classification, the accuracy reached 100% whenany metallic backgrounds were used. Conversely, Teflon showed a theoretically high accuracy of 96.21% but only achievedan experimental accuracy of 72.2%. These findings suggest that using metallic backgrounds can improve the spectral qualityand classification of plastics with low thickness (films) and complex material composition (multilayers).
... A portable NIR spectrometer [12] or an ultrasonic sensor can also be used for automatic separation of plastic at the source. Technologies such as tracer-based sorting, sorting using digital watermarks, and others can also be noted [13]. ...
Article
Full-text available
One of the pressing global problems in recent years is environmental pollution and inefficient waste management. In this regard, smart bins with automatic waste separation are increasingly being created and introduced, which facilitate recycling and reduce the burden on the environment. The introduction of such bins and simple devices for recycling plastic in educational institutions will not only ensure cleanliness and reduce the amount of waste but will also contribute to the formation of environmental awareness in the younger generation, nurturing in them a sense of responsibility for the environment. The aim of this research is to develop, within the framework of a student startup, and experimentally test smart waste bins with automatic separation of plastic waste and a device for recycling it for secondary use for educational purposes without transportation and additional costs. Based on modern IoT and AI technologies, a smart waste bin with automatic waste separation and a device for recycling collected plastic and creating threads for a 3D printer were developed, implemented in an educational institution, and experimentally tested. Their influence on the formation of environmental awareness among students was confirmed during a survey conducted among students of Abai Kazakh National Pedagogical University (Kazakhstan).
... Additionally, hand-held NIR devices have the potential to be used in rapid identification and inspection of samples on smaller scales with high precision (Schmidt et al., 2020). Rani et al. (2019) presented an application of a microNIR on urban plastic waste and reported no significant loss of accuracy due to contamination. Due to the inherent nature of PVC and the existing schemes for collecting major PVC waste streams, sorting technologies such as NIR are mostly used for flakes from a plastic mixture, and bulky items are sorted manually. ...
Article
Full-text available
Polyvinyl chloride (PVC) is a polymer comprised of more than 50% chlorine that offers unmatched versatility at low expense. PVC is irreplaceable in several applications, such as construction materials, medical applications, and cables. This versatility and tunable properties come at the cost of complex formulations for the product and challenging end-of-life (EoL) options for PVC waste. Pure collected and sorted PVC is already recycled successfully to some extent, yet, when PVC ends up in a mixed plastic waste stream, it can be detrimental to the recycling process. PVC waste and its effects at various concentrations remain a focal point for both scholars and policy makers. In this review, the narrative begins at the naissance of PVC and continues to investigate the EoL valorization options when the products are inevitably discarded. Strategies for PVC waste recycling and the technical and legal challenges regarding each method are discussed, focusing on the European recycling market. An effective solution to handle EoL PVC requires a combination of policies and schemes for proper collection and sorting of specific waste streams and considering all available technologies to select the right tools. This review can support appropriate policies and the selection of suitable methods of recycling PVC waste.
... Moreover, machine learning algorithms act as intelligent assistants, analyzing large datasets to identify patterns and make automated decisions, enhancing automated sorting processes for the faster, more consistent completion of repetitive tasks [7]. These technologies find applications not only in waste management [8][9][10][11] but also in fields like food, pharmaceuticals, forensics, and surface analysis [12][13][14][15]. ...
Article
Full-text available
Waste material classification is critical for efficient recycling and waste management. This study proposes a novel, low-cost material classification system based on a single, voltage-tunable Ge-on-Si photodetector operating across the visible and short-wave infrared (SWIR) spectral regions. Thanks to its tunability, the sensor is able to extract spectral information, and the system effectively distinguishes between seven different materials, including plastics, aluminum, glass, and paper. The system operates with a broadband illuminator, and material identification is obtained through the processing of the photocurrent signal at different bias voltages with classification algorithms. Here, we demonstrate the basic system functionality and near real-time classification of different waste materials.
Article
We report the design, implementation and test of a Mid-Infrared spectrometer proof-of-concept that utilizes an uncooled micro-bolometer array, sensitive in the 3–14 µm spectral range, integrated in a conventional optical dispersive spectrometry setup. Such a spectrometer enables instantaneous measurements across this broad spectral range, comparable to that of a FTIR but with a more compact design and without moving parts. This makes it ideal for integration into portable, battery-powered devices such as handheld scanners. The Mid-IR range offers significant advantages over NIR-SWIR spectrometers, especially for organic compound analysis. A notable application for this instrument: plastic waste sorting—including black plastics—was tested with significant accuracy and effectiveness of plastic classification (on PP, PET and PE samples) with a very simple machine learning algorithm.
Article
Full-text available
Polyethylene terephthalate (PET) is a flexible material commonly used in beverage bottles, and the acceptability of its products is affected by various physical and chemical qualities. Conventional PET plastic resin classification techniques have proven to be expensive and unproductive. This work suggested a unique method for quickly and non-destructively classifying recycled PET (rPET) plastic bottles using near-infrared spectroscopy (NIRS) and machine learning (ML) specifically Waikato Environment for Knowledge Analysis (WEKA). Using selected NIRS wavelength ranges; absorbance values were generated, analyzed, sorted, and compiled into a dataset. This data sheet was then subjected to 43 classifiers using WEKA. The four-fold cross-validation with the J48 sub-classifier resulted in a 100% accuracy, with 300 correct classified instances, zero (0) incorrectly classified instances, and a classifier operating time of 0.00 seconds. It also produced 0% relative absolute error and 0% root mean squared error. This process could significantly enhance production quality control and material assessment, allowing for more accurate matching of PET plastic resin qualities to particular product applications. This study highlighted the potential of NIRS as an effective, non-destructive method for classifying plastic resin in industrial settings.
Article
Full-text available
Principal component analysis is one of the most important and powerful methods in chemometrics as well as in a wealth of other areas. This paper provides a description of how to understand, use, and interpret principal component analysis. The paper focuses on the use of principal component analysis in typical chemometric areas but the results are generally applicable.
Article
For sustainable utilization of raw materials and environmental protection, the recycling of the most common polymers—polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polyvinyl chloride (PVC), and polystyrene (PS)—is an extremely important issue. In the present communication, the discrimination performance of the above polymer commodities based on their near-infrared (NIR) spectra measured with four real handheld (<200 g) spectrometers based on different monochromator principles were investigated. From a total of 43 polymer samples, the diffuse reflection spectra were measured with the handheld instruments. After the original spectra were pretreated by second derivative and standard normal variate (SNV), principal component analysis (PCA) was applied and unknown samples were tested by soft independent modeling of class analogies (SIMCA). The results show that the five polymer commodities cluster in the score plots of their first three principal components (PCs) and, furthermore, samples in calibration and test sets can be correctly identified by SICMA. Thus, it was concluded that on the basis of the NIR spectra measured with the handheld spectrometers the SIMCA analysis provides a suitable analytical tool for the correct assignment of the type of polymer. Because the mean distance between clusters in the score plot reflects the discrimination capability for each polymer pair the variation of this parameter for the spectra measured with the different handheld spectrometers was used to rank the identification performance of the five polymer commodities.
Article
The assessment of the classification performance can be based on class indices, such as sensitivity, specificity and precision, which describe the classification results achieved on each modelled class. However, in several situations, it is useful to represent the global classification performance with a single number. Therefore, several measures have been introduced in literature to deal with this assessment, accuracy being the most known and used. These metrics have been proposed to generally face binary classification tasks and can behave differently depending on the classification scenario. In this study, different global measures of classification performances are compared by means of results achieved on an extended set of real multivariate datasets. The systematic comparison is carried out through multivariate analysis. Further investigations are then derived on specific indices to understand how the presence of unbalanced classes and the number of modelled classes can influence their behaviour. Finally, this work introduces a set of benchmark values based on different random classification scenarios. These benchmark thresholds can serve as the initial criterion to accept or reject a classification model on the basis of its performance.
Article
Waste sorting is key to the process of waste recycling. Exact identification of plastic resin and wood products using Near Infrared (NIR, 1–1.7 µm) sensing is currently in use. Yet, dark targets characterized by low reflectance, such as black plastics, are hard to identify by this method. Following the recent success of Midwave Infrared (MWIR, 3–12 µm) measurements to identify coloured plastic polymers, the aim of this study was to assess whether this technique is applicable to sorting black plastic polymers and wood products. We performed infrared reflectance contact measurements of 234 plastic samples and 29 samples of wood and paper products. Plastic samples included black, coloured and transparent Polyethylene Terephthalate (PET), Polyethylene (PE), Polyvinyl Chloride (PVC), Polypropylene (PP), Polylactic acid (PLA) and Polystyrene (PS). The spectral signatures of the black and coloured plastic samples were compared with clear plastic samples and signatures documented in the literature to identify the polymer spectral features in the presence of coloured material. This information was used to determine the spectral bands that best suit the sorting of black plastic polymers. The main NIR-MWIR absorption features of wood, cardboard and paper were identified as well according to the spectral measurements. Good agreement was found between our measurements and the absorption features documented in the literature. The new approach using MWIR spectral features appears to be useful for black plastics as it overcomes some of the limitations in the NIR region to identify them. The main limitation of this technique for industrial applications is the trade-off between the signal-to-noise ratio of the sensor operating in standoff mode and the speed at which waste is moved under the sensor. This limitation can be resolved by reducing the system’s spectral resolution to 16 cm⁻¹, which allows for faster spectra acquisition while maintaining a reasonable signal-to-noise ratio.
Article
A crucial prerequisite for recycling forming an integral part of municipal solid waste (MSW) management is sorting of useful materials from source-separated MSW. Researchers have been exploring automated sorting techniques to improve the overall efficiency of recycling process. This paper reviews recent advances in physical processes, sensors, and actuators used as well as control and autonomy related issues in the area of automated sorting and recycling of source-separated MSW. We believe that this paper will provide a comprehensive overview of the state of the art and will help future system designers in the area. In this paper, we also present research challenges in the field of automated waste sorting and recycling.
Article
The extensive demand and use of plastics in modern life is associated with a significant economical impact and a serious ecological footprint. The production of plastics involves a high energy consumption and CO2 emission as well as the large need for (limited) fossil resources. Due to the high durability of plastics, large amounts of plastic garbage is mounting in overflowing landfills (plus 9.6 million tons in Europe in the year 2012) and plastic debris is floating in the world oceans or waste-to-energy combustion releases even more CO2 plus toxic substances (dioxins, heavy metals) to the atmosphere. The recycling of plastic products after their life cycle can obviously contribute a great deal to the reduction of the environmental and economical impacts. In order to produce high-quality recycling products, mono-fractional compositions of waste polymers are required. However, existing measurement technologies such as near infrared spectroscopy show limitations in the sorting of complex mixtures and different grades of polymers, especially when black plastics are involved. More recently invented technologies based on mid-infrared, Raman spectroscopy or laser-aided spectroscopy are still under development and expected to be rather expensive. A promising approach to put high sorting purities into practice is to label plastic resins with unique combinations of fluorescence markers (tracers). These are incorporated into virgin resins during the manufacturing process at the ppm (or sub ppm) concentration level, just large enough that the fluorescence emissions can be detected with sensitive instrumentation but neither affect the visual appearance nor the mechanical properties of the polymers. In this paper we present the prototype of a measurement and classification system that identifies polymer flakes (mill material of a few millimeters size) located on a conveyor belt in real time based on the emitted fluorescence of incorporated markers. Classification performance and throughput were experimentally quantified using 3 different types of polymers (Polyoxymethylene (POM), Polybutylenterephthalat (PBT) and Acrylonitrile Styrene Acrylate (ASA)) in colored and uncolored form. Overall, 12 classes of plastic flakes were investigated in this study, where 11 classes were labeled with unique binary combinations of 4 fluorescence markers and class 12 includes unlabeled plastic flakes of various colors. From approx. 68,000 investigated flakes it was found that the developed measurement prototype system achieves an average sensitivity (true positive rate) of 99.4% and a precision (positive predictive value) of 99.5%, while being able to handle up to approx. 1800 flakes per second.
Article
The effectiveness of Hyperspectral imaging (HSI) in the near infrared (NIR) range (1000–1700 nm) was evaluated to discriminate PET (polyethylene terephthalate) from PLA (poly(lactic acid)), two polymers commonly utilized as packaging for foodstuff, in order to improve their further recycling process. An internal calibration based on five reference materials was initially used to eliminate the variability existing among images, then Partial Least Squares-Discriminant Analysis (PLS-DA) was used to distinguish and classify the three classes, i.e., background, PET and PLA. Considering the high amount of data conveyed by the training image, the PLS-DA models were also calculated using as training set a reduced version of the original matrix, with the twofold aim to reduce the computational time and to deal with an equal number of spectra for each class, independently from the initial selected areas. A variable selection procedure by means of iPLS-DA was also applied on both the whole and the reduced matrix. The results obtained on the reduced matrix using only six variables provided a prediction efficiency higher than 98%. Moreover, the possibility to recognize PET and PLA polymers by HSI in the NIR range was further confirmed by using Multivariate Curve Resolution (MCR) as an alternative approach, which also allowed to evaluate the effect of thickness of the transparent plastic samples.