Journal of Hazardous Materials Advances 10 (2023) 100260

Contents lists available at ScienceDirect

Journal of Hazardous Materials Advances

journal homepage: www.elsevier.com/locate/hazadv

Identification of biodegradable plastics using differential scanning

calorimetry and carbon composition with chemometrics
Menglin Zhang a,c,1, Yongzhi Zhang b,1, Chunlin Li b, Nie Jing b, Shengzhi Shao b, Fang Wang b,
Hanyi Mei b, Karyne M. Rogers b,d, Xiangdong Kong a,c,∗, Yuwei Yuan b,∗
a
Institute of Smart Biomaterials, School of Materials Science and Engineering, Zhejiang Sci-Tech University, Hangzhou 310018, Zhejiang, China
b
Institute of Agro-Products Safety and Nutrition, Zhejiang Academy of Agricultural Sciences; Key Laboratory of Information Traceability for Agricultural Products,
Ministry of Agriculture and Rural Affairs of China, Hangzhou 310021, China
c
Zhejiang-Mauritius Joint Research Center for Biomaterials and Tissue Engineering, Zhejiang Sci-Tech University, Hangzhou 310018, Zhejiang, China
d
National Isotope Centre, GNS Science, Lower Hutt 5040, New Zealand

a r t i c l e i n f o a b s t r a c t

Keywords: Biodegradable plastics have become an important commodity to replace non-degradable plastics and alleviate
Biodegradable plastic environment pollution. It is important to identify instances of counterfeit and mislabeled biodegradable plastics
DSC in the marketplace to avoid sustainability fraud and improve plastics waste recycling. In this study, differential
EA-IRMS
scanning calorimetry (DSC) and elemental analyzer-isotope ratio mass spectrometry (EA-IRMS) were used to
Chemometrics
analyze biodegradable plastics (polylactic acid (PLA), poly butylene adipate-co-terephthalate (PBAT), PBAT/PLA
Verification
Carbon isotopes blend (PBAT+PLA)) and non-degradable plastics (polyethylene terephthalate (PET), high density polyethylene
(HDPE), low density polyethylene (LDPE), polypropylene (PP)). A verification model was built using orthogonal
partial least squares-discriminant analysis (OPLS-DA). Results showed that the melting point of biodegradable
and non-degradable plastics ranged from 120 °C to 176 °C and 105 °C to 253 °C, respectively. The melting points
of biodegradable PLA and non-degradable PP overlapped. The 𝛿 13 C values of biodegradable and non-degradable
plastics ranged from -25.8 ‰ to -11.6 ‰ and -32.3 ‰ to -25.4 ‰, respectively. The % C of biodegradable and
non-degradable plastics ranged from 36.0% to 61.0% and 56.4% to 85.3%, respectively. Three OPLS-DA models
were established by combining DSC variables, 𝛿 13 C values and % C. The PLA discriminant accuracy of these
models was 100%, suggesting this approach holds promise for the regulation and identification of biodegradable
plastics in the marketplace.

1. Introduction label claims and substitution of a large proportion of non-degradable

plastics as biodegradable products in the marketplace. These fraudu-
Plastics are widely used in agriculture, packaging, construction, lent activities hinders the healthy development of the plastics industry
medical care and other industries (Andrady and Neal, 2009), provid- and impacts waste disposal options (Rillig et al., 2019). Therefore, it
ing an unparalleled convenience for everyday life. In 2021, the an- is important to find a convenient and accurate verification method for
nual output of plastics reached 390.7 million tons. The plastic prod- biodegradable plastics to enhance environmental security and green in-
ucts industry is one of the fastest growing industries in the world dustry development.
(Plastics Europe, 2022), yet the non-degradable nature of traditional Biodegradable plastic refers to a kind of environmentally friendly
plastics has led to serious pollution problems (Lehner et al., 2019). plastic that can be degraded by natural microorganisms under natu-
Biodegradable plastics have received extensive attention in many coun- ral soil, sandy soil conditions or specific conditions such as compost-
tries and regions due to their ability to breakdown over time (Flury and ing, anaerobic digestion or water-based culture solution, and eventually
Narayan, 2021). Biodegradable plastics research and production began completely degraded into CO2 or CH4 , H2 O, inorganic salts and new
later in China than in western countries. The lack of unified national biomass (Qi et al., 2017). The biodegradability of plastics is determined
standards and testing methods in China has led to the appearance of by whether the carbon (C) elements can be completely converted into
non-degradable plastics masquerading as biodegradable plastics, false CO2 through the process of microbial action (ISO 17556, 2019), but the

∗
Corresponding authors at: 198 Shiqiao Road, Shangcheng District, Hangzhou 310021, Zhejiang Province, China
E-mail addresses: kongxd@zstu.edu.cn (X. Kong), ywytea@163.com (Y. Yuan).
1
Joint first authors

https://doi.org/10.1016/j.hazadv.2023.100260
Received 4 January 2023; Received in revised form 4 February 2023; Accepted 14 February 2023
2772-4166/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

istics and melting peak area using DSC (Shabaka et al., 2019). PE, PP
and PET microplastic fragments extracted from three types of indus-
trial wastewater were quantitatively and identified using DSC. DSC was
found to be an effective method to improve plastics detection accuracy
by determining preheating and cooling steps (Bitter and Lackner, 2021).
In addition, DCS has been used in combination with thermogravimetric
(TG) or optical microscope images to analyze and identify plastic frag-
ments in water samples (Schindler et al., 2017; Chialanza et al., 2018).
EA-IRMS is an effective method for tracing the geographical origin
of agricultural products in the environment (Yuan et al., 2018). Stud-
ies show that stable C isotopes (𝛿 13 C) values of beet and rice ranges
from −28 ‰ to −23 ‰, as they absorb CO2 by the Calvin cycle (C3
cycle), and corn and sugarcane use CO2 via the Hatch-Slack cycle (C4
cycle) with 𝛿 13 C values ranging from −14 ‰ to −10 ‰ (Zhang et al.,
2016). The 𝛿 13 C values of petroleum range from −32.5 ‰ to −23.3 ‰
Fig. 1. Plastics identification quadrant classifying plastic type, raw material
(Yeh and Epstein, 1981). Previous studies have shown that the 𝛿 13 C val-
source and biodegradability (Rogers et al., 2021). ues of plastic was consistent with that of its raw precursor material, and
this technique held the promise of distinguishing between bio-based and
petroleum-based plastics (Suzuki et al., 2010). On this basis, the appli-
process is frequently complicated and the cycle is long. Biodegradable cability of C isotopes in the identification of bio-based and petroleum-
plastics can be made from both fossil-fuel or bio-based raw materials based plastics was utilized, and results showed that the 𝛿 13 C values of
(Nazareth et al., 2019, Fig. 1). Polylactic acid (PLA) and poly butylene HDPE, LDPE and PLA plastics were significantly different, and not af-
adipate-co-terephthalate (PBAT) are the most widely used biodegrad- fected by the color of plastics (Berto et al., 2017). In order to improve
able plastics in the market (Haider et al., 2019). PLA is an aliphatic on the single isotope identification method, dual C and hydrogen (H)
polyester; its synthetic raw material, lactic acid, can be obtained from isotope characteristics combined with partial least squares discriminant
fermented corn, straw and other renewable plants, in contrast to tra- analysis (PLS-DA) were used to distinguish bio-based, petroleum-based
ditional petroleum raw materials (Fukushima et al., 2009). PBAT is and mixed bio-petroleum plastics (Rogers et al., 2021).
an aliphatic-aromatic-co-polyester with high flexibility and excellent EA-IRMS provides a highly sensitive and automated analysis for plas-
impact strength (Kargarzadeh et al., 2020), yet it is derived mainly tics authentication with no sample color analytical biases conversely to
from petroleum resources (Iwata, 2015). PBAT/PLA (bio-petroleum colorimetric methods. It can be used as an investigative tool for plastics
blend) plastics eliminates the disadvantages of PLA, such as poor tough- identification and environmental monitoring (Birch et al., 2021).
ness, providing improved physical performance and durability, and Although previous plastic authentication studies have focused more
has been widely used in shopping bags (Ding et al., 2018). Common on the use of 𝛿 13 C values to distinguish between bio-based and
non-degradable plastics such as polyethylene terephthalate (PET), high petroleum-based plastics (Suzuki et al., 2010; Rogers et al., 2021), this
density polyethylene (HDPE), low density polyethylene (LDPE), and study expands on these studies to investigate for the first time, the appli-
polypropylene (PP) are mainly synthesized from petroleum (Geyer et al., cability of DSC and EA-IRMS analysis for the identification of biodegrad-
2017), while PE and PET can also be bio-based (Iwata, 2015; Fig. 1). Due able and non-degradable plastics.
to the different thermal properties and sources of plastics (Shim et al., Different plastic thermal properties and raw material ingredients
2017), differential scanning calorimetry (DSC) and elemental analyzer- lead to different DSC results and C isotope compositions. However, plas-
isotope ratio mass spectrometry (EA-IRMS) have attracted wide atten- tics biodegradability is a specific attribute that is still under-developed
tion to achieve the identification of plastics. for plastics identification. In this study, a combination of stable iso-
DSC is a method to analyze the physical and chemical changes (such tope and DSC analytical methods were investigated as a tool to identify
as melting, crystallization and glass transition) of materials due to phase biodegradable plastics and improve methods to assuring their authen-
transformation and exothermic or endothermic phenomena (Vyazovkin, ticity. This research evaluated the thermal properties and C isotopes
2002). DSC has the advantage of easy calibration, is a highly sensitive, of analytical grade reference polymers sourced from different chemi-
technique when used to analyze basic thermodynamic properties of cal suppliers (including biodegradable plastics (PLA, PBAT, PBAT+PLA)
materials, and can be used for plastics identification (Majewsky et al., and non-degradable plastics (LDPE, HDPE, PP, PET)) using DSC and EA-
2016). DSC only requires a small amount of sample for analysis (about IRMS. We use our reference dataset to propose an OPLS-DA (Orthogonal
1 to 20 mg) and does not require complex data processing. Partial Least Squares Discriminant Analysis) model using a combination
DSC is commonly used to identify plastic waste in aquatic environ- of DCS and C isotopes (𝛿 13 C values and C content (% C)) to identify in-
ments but is mainly used to identify non-degradable plastics such as market biodegradable and non-degradable plastics. This work provides
HDPE, LDPE and PP (Shabaka et al., 2019), rather than this study’s fo- an analytical pathway for the identification and purity quantification of
cus on biodegradable plastics. biodegradable plastics to improve regulatory investigations and reduce
Previous studies on plastic identification also combined a variety of plastics fraud.
analysis techniques, such as the combination of spectral analysis and
microscopic images, although spectral analysis is an indirect analytical 2. Materials and methods
method, and can be complicated to interpret the spectral data outputs
(Mecozzi et al., 2016), and microscopic images are mainly based on sub- 2.1. Sample collection and processing
jective judgment (Frére et al., 2016). Compared to other plastic identi-
fication techniques such as Fourier transform infrared spectrum (FTIR) A number of different plastic samples were purchased directly
and Raman spectroscopy (Pakhomova et al., 2020; Schymanski et al., from plastics manufacturers, including 23 biodegradable plastic sam-
2018), DSC analysis does not need complicated spectral preprocessing, ples (PLA, PBAT, PBAT+PLA) and 47 non-degradable plastic samples
so the data output is faster and less complex. (LDPE, HDPE, PP, PET). Samples were rinsed with deionized water to
Previously, eleven microplastics from the Mediterranean coast of remove contaminants, oven dried at 35 °C and cut up into small pieces
Egypt, including LDPE, HDPE, PP, PET, etc., were identified for their (about 3 × 3 cm) for later use. Details of plastic samples are shown in
polymer composition and proportion based on their melting character- Table 1 and Fig. S1.

2
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Table 1
Plastics manufacturing information.

Recycle no. Polymer Molecular formula Number of samples Use of plastics Note

Polyethylene Terephthalate (PET) (C10 H8 O4 )n 11 beverage bottles, spray bottles, Non-degradable

fruit boxes, take-out boxes Petroleum

High Density Polyethylene (HDPE) H(C2 H2 )n H 17 shopping bags, garbage bags, Non-degradable
household containers Petroleum

Low Density Polyethylene (LDPE) H (C2 H2 )n H 8 packaging film, shock-proof Non-degradable

bubble bag, food packaging Petroleum
bags
Polypropylene (PP) (C2 H2 CH(CH3 ))n 11 disposable tableware, Non-degradable
disposable cups Petroleum

Poly (lactic acid) (PLA) (C4 H6 O3 )n 10 Straw (white, black, red, Biodegradable
green), disposable knife, fork, Bio-based
cup lid, tableware bag

Poly (butylene adipate-co-terephthalate) (C10 H10 O4 .C6 H10 O4 .C4 H10 O2 )x 4 shopping bags, film Biodegradable
(PBAT) Petroleum

PBAT+PLA / 9 shopping bags, garbage bags, Biodegradable

agricultural mulch Mixed bio-based and
petroleum

Analytical grade polymers were used as reference samples to verify Where, X is the 𝛿 13 C value of the sample, and Rsample and Rstandard are the
the accuracy of the in-market plastic samples. PLA was purchased from heavy-to-light ratio of element in the sample and reference materials,
Shanghai Aladdin Biochemical Technology Co. LTD.; and PBAT, PET, respectively, i. e. 13 C/12 C. This isotopic delta value is presented as per
HDPE, LDPE, and PP were purchased from Shanghai Maclin Biochemical mil (‰).
Technology Co., LTD. Reference materials were purchased from IAEA for multi-point cal-
ibration of C isotopes: B2155 (protein, 𝛿 13 C=−26.98 ‰), IAEA-CH-6
2.2. DSC analysis (sucrose, 𝛿 13 C=−10.45 ‰) and USGS64 (glycine, 𝛿 13 C =−40.81 ‰).
The 𝛿 13 C value is based on the V-PDB from Vienna.
Thermal analysis of plastics was performed in N2 atmosphere using a
Differential Scanning Calorimetry (DSC) machine (Model 200 F3, NET- 2.4. Data analysis and OPLS-DA model
ZSCH GMBH, Germany). Around 5.0 mg of plastic was weighed into an
aluminum crucible and heated from room temperature to 180 °C (HDPE, Proteus Analysis and Origin 2021 software were used to analyze the
LDPE, PBAT), 220 °C (PP, PLA, PBAT+PLA) or 300 °C (PET). Samples DSC thermograms. SPSS 18.0 software (IBM, New York, USA) was used
were cooled to 30 °C and then heated again to 180 °C, 220 °C or 300 °C to conduct one-way analysis of variance (ANOVA). Orthogonal partial
accordingly (heating/cooling rate was 10 °C /min). The first tempera- least squares discriminant analysis (OPLS-DA) was analyzed by SIMCA
ture rise is used to eliminate the historical heat of the sample, and the 14.1 software (Umetrics, Gottingen, Germany).
second temperature rise is used to analyze the plastic. The crucible mass Three OPLS-DA models were constructed with 𝛿 13 C values, % C
before and after the heating test was compared. If there was no change, and DSC signals at different temperatures as input variables, including
then it was assumed that there was no chemical change such as decom- biodegradable and non-degradable plastics model 1 (45 training sets and
position in the process of thermal analysis. The glass transition tem- 25 test sets), biodegradable plastics model 2 (15 training sets and 8 test
perature (Tg ), melting point (Tm ), and cold crystallization temperature sets) and non-degradable plastics model 3 (30 training sets and 17 test
(Tc ) was obtained from the DSC thermograms (Beltrán-Sanahuja et al., sets).
2019).
3. Results and discussion
2.3. Carbon isotope analysis
3.1. DSC results of biodegradable and non-degradable plastics
The simultaneous determination of C content and isotope compo-
The DSC results of different plastic types changed according to each
sition (% C and 𝛿 13 C value) was carried out on a Vario Isotope Cube
polymer’s molecular structure (Fig. 2). Table 2 summarizes the corre-
elemental analyzer (Elementar, Germany) linked to a BioVision isotope
sponding data. Tm of biodegradable and non-degradable plastics ranged
ratio mass spectrometer (Elementar, Germany). Around 2.0 mg of plas-
from 120 °C to 176 °C, and 105 °C to 253 °C, respectively. The DSC ther-
tic was cut up finely, weighed and folded into a 6 × 4 mm tin capsule
mograms of non-degradable plastics had better repeatability than that
in duplicate and combusted at 1150 °C in a WO3 reactor. Gasses were
of biodegradable plastics.
reduced as N2 and CO2 in a reduction reactor at 850 °C. The gasses
Fig. 2a-d shows the DCS thermograms of PLA, where Tm of PLA
were purified, separated on a GC column, passed through a diluter and
ranged from 146 °C to 176 °C. Different PLA starting materials and pro-
finally measured by the mass spectrometer. The stable isotopic compo-
duction technologies lead to three different types of DSC thermograms
sition was expressed as delta notation (𝛿) and the calculation is showed
for the PLA plastic samples (Fig. 2b-d). The two melting peaks in Fig. 2c
as the following Eq. (1):
were related to the imperfect 𝛼’ type crystal and perfect 𝛼’ type crystal
Xsample (‰) = (𝑅𝑠𝑎𝑚𝑝𝑙𝑒 ∕𝑅𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 ) − 1 (1) structures of PLA, respectively (Hu et al., 2019; Arruda et al., 2015).

3
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Fig. 2. DSC thermograms of biodegradable plastics. (a) PLA reference sample; (b) to (d) PLA plastics (n=10); (e) PBAT reference sample; (f) PBAT plastics (n = 4);
(g) to (o) PBAT+PLA plastic samples (n = 9). Tg is the glass transition temperature, Tm is the melting temperature, Tc is the cold crystallization temperature, EXO
means exothermic and PLA 1 to 10 represents 10 PLA plastic samples.

4
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Table 2
Melting point data of biodegradable and non-degradable plastics from DSC.

Samples PLA PBAT PBAT+PLA (n = 9) PET HDPE LDPE PP

(n = 70) (n = 10) (n = 4) (n = 11) (n = 17) (n = 12) (n = 11)
PBAT PLA

Min 146 120 121 147 241 130 105 154

Max 176 121 129 173 253 136 112 166
Mean – 120 124 – 248 132 110 163
SD – 0.5 2.6 – 3.9 1.6 2.6 3.8

Note: PLA, PBAT and PBAT+PLA are biodegradable plastics; PET, HDPE, LDPE and PP are non-degradable plastics.

The mechanism of melting, recrystallization and remelting proved the values and % C of plastics were analyzed using EA-IRMS. The mean and
rational for a bimodal phenomenon in polyester (Ma et al., 2008). Due standard deviations (SD) of 𝛿 13 C and % C for different plastic types are
to the high Tg of PLA, at around 60 °C, and the low crystallization rate, it listed in Table 3 and their distributions are plotted in Fig. S2.
cannot fully crystallize during the molding process (Fig. 2d) (Wen et al., The 𝛿 13 C values of biodegradable and non-degradable plastics
2019). The DSC thermograms of PBAT had good repeatability (Fig. 2e, ranged widely from −25.8 ‰ to −11.6 ‰ and −32.3 ‰ to −25.4
f), but samples all showed a wide melting heat absorption temperature ‰, respectively. The% C of biodegradable and non-degradable plas-
(Signori et al., 2009). The Tm of PBAT was 120 °C±0.5 °C. Fig. 2g-o tics ranged from 36.0% to 61.0% and 56.4% to 85.3%, respectively
shows the DSC thermograms of PBAT+PLA blends. The Tm of PBAT and (Table 3). Biodegradable plastics can be derived from both biological
PLA was 125 °C±2.7 °C and 146 °C to 173 °C, respectively. Due to im- and/or petroleum sources (Rogers et al., 2021; Fig. 1), so the range of
perfect crystallization of PLA, a bimodal phenomenon also appeared. 𝛿 13 C values is wide. At present, non-degradable in-market plastics are
Compared to single source PLA plastics, the Tm of PLA after PBAT ad- mainly derived from petroleum raw materials, so their 𝛿 13 C values are
dition was lower as the regularity of PLA molecular chain was further more negative, consistent with a petroleum source.
reduced (Yu and Li, 2014). The general mixing ratio of PBAT and PLA Notably, PLA plastics were significantly different from other plastics
of the raw plastic products varied during the trials, resulting in differ- (p < 0.05) with more positive 𝛿 13 C values and lower % C. The mean 𝛿 13 C
ent signal sizes generated by PBAT and PLA during thermal analysis value of PLA plastics was −13.9 ‰±1.6 ‰, which was higher than PBAT
(Çoban et al., 2018). (−25.1 ‰±0.7 ‰) and PBAT+PLA (−23.8 ‰±0.6 ‰) plastics (Table 3).
PET has the highest Tm at 248 °C±3.9 °C (Fig. 3a, b). Only some The mean 𝛿 13 C value of PET plastics was −27.6 ‰±1.4 ‰, HDPE was
PET plastics showed obvious Tg (around 80 °C) and Tc (around 150 °C) −29.2 ‰±1.2 ‰, LDPE was −29.5 ‰±1.2 ‰ and PP was −28.0 ‰±1.7
(D’Amico et al., 2008). High crystallinity of PET and low amorphous ‰ (Table 3). There was no significant difference in mean 𝛿 13 C values
content leads to insignificant glass transition and cold crystallization among the four non-degradable plastics (p > 0.05). The mean 𝛿 13 C value
peaks (Ruvolo-Filho and Curti, 2008). The DSC thermograms of HDPE of PLA was the highest of all plastics in this study, because its raw in-
had good repeatability, and Tm was 132 °C±3.5 °C (Fig. 3c, d). The DSC gredients (lactic acid) come mainly from corn starch (a C4 plant), while
thermograms of the plastic samples labeled as LDPE appeared as two PBAT comes from petroleum. These isotope results are slightly higher
melting peaks in the thermograms (Fig. 3f). The former was from LDPE than other previous studies (Rogers et al., 2021), but may be related to
and the Tm was 110 °C±2.1 °C, which was consistent with the LDPE the different international ingredients from plant and petroleum sources
reference sample result (Fig. 3e). The latter was from linear low-density (Birch et al., 2021).
polyethylene (LLDPE) and the Tm was 123 °C±1.0 °C, which was similar PLA had a mean % C of 44.4%±4.9%, the lowest among all plas-
to PBAT. This was consistent with the results of a previous LDPE/LLDPE tics. The mean % C of PBAT plastics was slightly higher (57.2%±4.9%)
blending study (Shieh and Chuang, 2001). LDPE, LLDPE and HDPE have than PBAT+PLA plastics (51.2%±4.6%). The mean % C of PET was
the same molecular formula, but the degree of branching is different 60.4%±1.4%, HDPE was 80.5%±3.4%, LDPE 83.8%±1.0% and PP was
(LLDPE>LDPE>HDPE). The higher the degree of branching, the higher 84.1%± 0.7% (Table 3). The mean % C of PET, PBAT and PBAT+PLA
the melting point (Salakhov et al., 2021). Fig. 3h, g showed the DSC was significantly different from other non-degradable plastics (p <
thermogram of PP, and the Tm was 163 °C±3.8 °C, which was consistent 0.05). Due to the existence of ester bonds in the molecular structure of
with previous research (Larsen et al., 2021). PLA (Auras et al., 2004), the mean % C was lowest. PBAT and PET have
According to DSC melting peak results, PLA was significantly differ- ester bonds and a benzene ring structure (Su, 2021; Nisticò, 2020), so
ent from non-degradable plastics except for PP, and PBAT was different their mean % C is slightly higher. LDPE, HDPE and PP are alkane poly-
from non-degradable plastics except for LLDPE. However, the Tm of the mers with only C and H elements in the molecular chain (Millican and
same type of plastic was shown to fluctuate. For example, in this study, Agarwal, 2021), so their mean % C is highest.
the difference between the maximum and minimum melting point of Results show that the 𝛿 13 C values of biodegradable and non-
PLA was found to be up to 30 °C. This was because the characteris- degradable plastics are different, especially for PLA (Table 3). How-
tic phase transition temperature is affected not only by the degree of ever, there is an overlap for the 𝛿 13 C values of biodegradable and non-
polymer branching, but also by impurities or additives, and particle size degradable plastics at around −25 ‰ (Fig. S2), mainly because PBAT is
(Menczel and Prime, 2009; Boey et al., 2021) which can cause confusion derived from petroleum. The % C of biodegradable and non-degradable
for data reading and plastic analysis. The authenticity of in-market PLA, plastics overlaps between 55% and 65% (Fig. S2), which is mainly re-
PBAT and PBAT+PLA plastics cannot be accurately assessed using only lated to the benzene ring and ester bonds in both PET and PBAT molec-
melting point analysis. A single melting point for a plastic can be subject ular chains. Therefore, the classification of plastics by C isotopes alone
to multiple interpretations. Therefore, the classification effect of plastic is also limited.
by DSC alone should be verified by another method (Baidurah, 2022).

3.3. Establishment and verification of an OPLS-DA model and regulatory

3.2. Carbon content and isotope composition of biodegradable and protocols
non-degradable plastics
As previously seen, single variable DSC or EA-IRMS analysis methods
According to different raw starting materials, the C content and iso- have certain limitations. DCS analysis is unable to distinguish plastics
tope composition of plastics have certain notable differences. The 𝛿 13 C with similar Tm ’s (e.g., the Tm of PLA and PP are both around 160 °C),

5
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Fig. 3. DSC thermograms of non-degradable

plastic. (a) PET reference sample; (b) PET plas-
tics (n = 11); (c) HDPE reference sample; (d)
HDPE plastics (n = 17); (e) LDPE reference sam-
ple; (f) LDPE plastics (n = 8); (g) PP reference
sample; (h) PP plastics (n = 11).

Table 3
Carbon isotope and elemental composition of biodegradable and non-degradable plastics.

Samples PLA PBAT PBAT+PLA PET HDPE LDPE PP

(n = 70) (n = 10) (n = 4) (n = 9) (n = 11) (n = 17) (n = 8) (n = 11)

𝛿 13 C (‰) −13.9 ± 1.6 −25.1 ± 0.7 −23.8 ± 0.6 −27.5 ± 0.4 −29.2 ± 1.2 −29.5 ± 1.2 −27.8 ± 1.7
Mean ± SD
C (%) 44.4 ± 4.9 57.2 ± 4.9 51.2 ± 4.6 60.4 ± 1.4 80.5 ± 3.4 83.8 ± 1.0 84.1 ± 0.7
Mean ± SD
Note: PLA, PBAT and PBAT+PLA are biodegradable plastics; PET, HDPE, LDPE and PP are non-degradable plastics.

6
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Fig. 4. Score-loading plot of different OPLS-DA models; (a) biodegradable and non-degradable plastics Model 1; (b) biodegradable plastics Model 2; (c) non-
degradable plastics Model 3.

providing an ambiguous result. However, when combined with the 𝛿 13 C The OPLS-DA model 2 for biodegradable plastics is shown in Fig. 4b
values or % C, it may be possible to differentiate plastics with similar and Table 4. The scores of the training set and test sets for this OPLS-DA
Tm ’s (e.g., the 𝛿 13 C values of PLA and PP is −13.9 ‰ and −28 ‰, re- model are shown in Fig. 4b, and the fitting parameters R2 X(cum) and
spectively). R2 Y(cum) were 0.919 and 0.703, respectively. PLA plastics were con-
In this study, three OPLS-DA models were constructed to explore centrated and distributed in the first and fourth quadrants. PBAT plas-
the classification and identification of different biodegradable and non- tics and PBAT+PLA plastics were distributed in the second and fourth
degradable plastics; 1) a biodegradable and non-degradable plastic quadrants with partial crossover. The overall accuracies of the training
model, 2) a biodegradable plastic model and 3) a non-degradable plas- and test sets of plastics were 100% and 87.5%, respectively (Table 4).
tic model. OPLS-DA is a widely used supervised pattern recognition In this test set, one PBAT plastic was misclassified as PBAT+PLA. The
method, often a powerful alternative to unsupervised methods such as classification ability of the training set was better than that of the test
principal component analysis (PCA) (Chung et al., 2019). set. Model 2 was useful to distinguish PBAT and PBAT+PLA plastics, but
The OPLS-DA model 1 for biodegradable and non-degradable plas- especially PLA plastics.
tics is shown in Fig. 4a and Table 4. The scores of the training and test The OPLS-DA model 3 for non-degradable plastics is shown in Fig. 4c
sets for this OPLS-DA model are shown in Fig. 4a, and the fitting pa- and Table 4. The scores of the training and test sets for this OPLS-DA
rameters R2 X(cum) and R2 Y(cum) were 0.865 and 0.946, respectively. model are shown in Fig. 4c, and the fitting parameters R2 X(cum) and
Biodegradable plastics were distributed in the first and fourth quadrants, R2 Y(cum) were 0.917 and 0.922, respectively. LDPE plastics and PP
while non-degradable plastics were distributed in the second and third plastics were distributed in the first quadrant. PET plastics were dis-
quadrants. The overall accuracies of the training and test sets of plastics tributed in the third quadrant and HDPE plastics were distributed in
were 100% (Table 4). the fourth quadrant. The overall accuracies of the training and test sets

Table 4
Discriminant accuracy of OPLS-DA; Model 1 for biodegradable and non-degradable plastics, Model 2 for biodegradable plastics, Model 3 for non-degradable plastics.

Model Data Set Type Number of samples Test result Discriminant Overall
accuracy/% accuracy/%
Biodegradable Non-degradable
Plastic Plastic

Model 1 Training Set Biodegradable Plastic 15 15 0 100% 100%

Non-degradable Plastic 30 0 30 100%

Test Set Biodegradable Plastic 8 8 0 100% 100%

Non-degradable Plastic 17 0 17 100%

PLA PBAT PBAT+PLA

Model 2 Training Set PLA 7 7 0 0 100% 100%

PBAT 2 0 2 0 100%
PBAT+PLA 6 0 0 6 100%

Test Set PLA 3 3 0 0 100% 87.5%

PBAT 2 0 1 1 50%
PBAT+PLA 3 0 0 3 100%

PET HDPE LDPE PP

Model 3 Training Set PET 7 7 0 0 0 100% 100%

HDPE 11 0 11 0 0 100%
LDPE 5 0 0 5 0 100%
PP 7 0 0 0 7 100%

Test Set PET 4 4 0 0 0 100% 94.1%

HDPE 6 0 5 1 0 83.3%
LDPE 3 0 0 30 0 100%
PP 4 0 0 0 4 100%

7
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

The present study only covers a narrow range of biodegradable plas-

tics. As more types of biodegradable plastics become available with the
advent of new raw material sources, further studies will be required to
confirm the ongoing effectiveness of DSC and EA-IRMS for the identifi-
cation of biodegradable plastics.

Funding information

This research did not receive any specific grant from funding agen-
cies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.
Fig. 5. A regulatory protocol to verify biodegradable and non-degradable plas-
tics. Data availability

Data will be made available on request.

of plastics samples were 100% and 94.1%, respectively (Table 4). One
HDPE plastic in the test set was misidentified as LDPE. The model was Acknowledgements
found to effectively distinguish LDPE, HDPE, PP and PET plastics.
Following the outcomes of this research, we propose a verification This work was supported by the Special Fund of Discipline
method to determine and classify biodegradable and non-degradable Construction for Traceability of Agricultural Product (2022-ZAAS),
plastics. A protocol is proposed in Fig. 5 and is described as follows. DSC Zhejiang International Science and Technology Cooperation Project
variables, 𝛿 13 C values and % C of various plastic types are evaluated us- (2019C04020) and International Scientific and Technological Cooper-
ing OPLS-DA, which can effectively distinguish biodegradable plastics ation Base of Intelligent Biomaterials and Functional Fibers, Hangzhou,
from non-degradable plastics. OPLS-DA Model 1 was constructed to dis- 310018, China. We also thank the manufacturing suppliers and research
tinguish between biodegradable and non-degradable plastics. On this institutions that provided plastic samples.
basis, biodegradable plastics could be correctly assigned as PLA, PBAT
and PBAT+PLA by OPLS-DA model 2, and non-degradable plastics could Supplementary materials
be correctly assigned as PET, HDPE, LDPE and PP by OPLA-DA model
3. Supplementary material associated with this article can be found, in
The combination of DSC variables and 𝛿 13 C values provides an im- the online version, at doi:10.1016/j.hazadv.2023.100260.
proved strategy to achieve a more accurate identification of petroleum,
References
bio-based and biodegradable plastics. DSC and EA-IRMS analysis are not
only simple techniques, but the data output is rapid and easy to inter- Andrady, A.L., Neal, M.A., 2009. Applications and societal benefits of plastics. Phil. T rans.
pret. This study is the first to use OPLS-DA chemometrics to combine R. Soc. B 364, 1977–1984. doi:10.1098/rstb.2008.0304, 1526.
DSC results with 𝛿 13 C values and % C to establish a model for distin- Arruda, L.C., Magaton, M., Bretas, R.E.S., Ueki, M.M., 2015. Influence of chain extender
on mechanical, thermal and morphological properties of blown films of PLA/PBAT
guishing biodegradable and non-degradable plastics. This technique can blends. Polym. Test. 43, 27–37. doi:10.1016/j.polymertesting.2015.02.005.
be employed for regulatory agencies and industry to reduce fraud, and Auras, R., Harte, B., Selke, S., 2004. An overview of polylactides as packaging materials.
plastic waste recyclers to improve recycling and make plastic identifi- Macromol. Biosci. 4 (9), 835–864. doi:10.1002/mabi.200400043.
Baidurah, S., 2022. Methods of analyses for biodegradable polymers: a review. Polymers
cation easier. (Basel) 14, 4928. doi:10.3390/polym14224928.
Beltrán-Sanahuja, A., Casado-Coy, N., Simó-Cabrera, L., Sanz-Lázaro, C., 2019. Monitoring
polymer degradation under different conditions in the marine environment. Environ.
4. Conclusion Pollut., 113836 doi:10.1016/j.envpol.2019.113836.
Berto, D., Rampazzo, F., Gion, C., Noventa, S., Ronchi, F., Traldi, U., Giova-
Biodegradable (both petroleum and bio-based) plastics (PLA, PBAT, nardi, O., 2017. Preliminary study to characterize plastic polymers using elemen-
tal analyser/isotope ratio mass spectrometry (EA/IRMS). Chemosphere 176, 47–56.
PBAT+PLA) and non-degradable petroleum plastics (PET, HDPE, LDPE, doi:10.1016/j.chemosphere.2017.02.090.
PP) found in the marketplace were identified using a combination of Birch, Q.T., Potter, P.M., Pinto, P.X., Dionysiou, D.D., Al-Abed, S.R., 2021. Isotope ratio
DSC and EA-IRMS. DSC and EA-IRMS analyses found that different plas- mass spectrometry and spectroscopic techniques for microplastics characterization.
Talanta 224, 121743. doi:10.1016/j.talanta.2020.121743.
tic types have different thermodynamic properties, 𝛿 13 C values and%C Bitter, H., Lackner, S., 2021. Fast and easy quantification of semi-crystalline microplas-
due to their different polymer properties and raw precursor materials. tics in exemplary environmental matrices by differential scanning calorimetry (DSC).
An OPLS-DA discrimination method was used to combine DSC variables, Chem. Eng. J. 423, 129941. doi:10.1016/j.cej.2021.129941.
Boey, J.Y., Mohamad, L., Khok, Y.S., Tay, G.S., Baidurah, S., 2021. A review of the ap-
𝛿 13 C values and % C characteristics. Three effective identification mod- plications and biodegradation of polyhydroxyalkanoates and poly(lactic acid) and its
els were able to contrast and classify biodegradable and non-degradable composites. Polymers (Basel) 13, 1544. doi:10.3390/polym13101544.
plastics (model 1), biodegradable plastics (model 2) and non-degradable Chialanza, M.R., Sierra, I., Pérez Parada, A., Fornaro, L., 2018. Identification and
quantitation of semi-crystalline microplastics using image analysis and differ-
plastics (model 3) and had accuracies of more than 85%. The discrimi-
ential scanning calorimetry. Environ. Sci. Pollut. Res. 25 (17), 16767–16775.
nant accuracy for PLA reached 100%. doi:10.1007/s11356-018-1846-0.
The combination of DSC variables, 𝛿 13 C values and % C is found Chung, I.M., Kim, J.K., Han, J.G., Kong, W.S., Kim, S.Y., Yang, Y.J., Kim, S.H., 2019.
Potential geo-discriminative tools to trace the origins of the dried slices of shiitake
to be a useful approach to identify biodegradable from non-degradable
(Lentinula edodes) using stable isotope ratios and OPLS-DA. Food Chem 295, 505–
plastics and can provide new information and technical support for the 513. doi:10.1007/s12517-016-2621-5.
verification of biodegradable plastics. This method may be useful for Çoban, O., Bora, M.Ö., Kutluk, T., Özkoç, G., 2018. Mechanical and thermal properties
market regulations of biodegradable plastics, and it could be helpful of volcanic particle filled PLA/PBAT composites. Polym. Coposite. 39 (S3), E1500–
E1511. doi:10.1002/pc.24393.
for waste plastics processing to separate inert, non-degradable materials D’Amico, T., Donahue, C.J., Rais, E.A., 2008. Thermal analysis of plastics. J. Chem. Educ.
(fossil plastics) from bioplastics in composts and waste recycling. 85 (3), 404. doi:10.1021/ed085p404.

8
M. Zhang, Y. Zhang, C. Li et al. Journal of Hazardous Materials Advances 10 (2023) 100260

Ding, Y., Lu, B., Wang, P., Wang, G., Ji, J., 2018. PLA-PBAT-PLA tri-block Pakhomova, S., Zhdanov, I., van Bavel, B., 2020. Polymer type identification of marine
copolymers: effective compatibilizers for promotion of the mechanical and rhe- plastic litter using a miniature near-infrared spectrometer (MicroNIR). Appl. Sci. 10
ological properties of PLA/PBAT blends. Polym. Degrad. Stab. 147, 41–48. (23), 8707. doi:10.3390/app10238707.
doi:10.1016/j.polymdegradstab.2017.11.012. Qi, X., Ren, Y., Wang, X., 2017. New advances in the biodegradation of poly (lactic) acid.
Flury, M., Narayan, R., 2021. Biodegradable plastic as an integral part of the solution to Int. Biodeterior. Biodegrad. 117, 215–223. doi:10.1016/j.ibiod.2017.01.010.
plastic waste pollution of the environment. Curr. Opin. Green Sustainable Chem. 30, Rillig, M.C., Lehmann, A., de Souza Machado, A.A., Yang, G., 2019. Microplastic effects
100490. doi:10.1016/j.cogsc.2021.100490. on plants. New Phytologist 223 (3), 1066–1070. doi:10.1111/nph.15794.
Frére, L., Paul-Pont, I., Moreau, J., 2016. A semi-automated raman micro-spectroscopy Rogers, K.M., Turnbull, J.C., Dahl, J., Phillips, A., Bridson, J.H., Raymond, L.G., Liu, Z.,
method for morphological and chemical characterizations of microplastic litter. Mar. Yuan, Y., Hill, S.J., 2021. Authenticating bioplastics using carbon and hydrogen stable
Pollut. Bll. 113 (1–2), 461–468. doi:10.1016/j.marpolbul.2016.10.051. isotopes–an alternative analytical approach. Rapid Commun. Mass Spectrom. 35 (9),
Fukushima, K., Tabuani, D., Camino, G., 2009. Nanocomposites of PLA and PCL e9051. doi:10.1002/rcm.9051.
based on montmorillonite and sepiolite. Mater. Sci. Eng. C 29 (4), 1433–1441. Ruvolo-Filho, A., Curti, P.S., 2008. PET Recycled and processed from flakes with different
doi:10.1016/j.msec.2008.11.005. amount of water uptake: characterization by DSC, TG, and FTIR-ATR. J. Mater. Sci.
Geyer, R., Jambeck, J.R., Law, K.L., 2017. Production, use, and fate of all plastics ever 43 (4), 1406–1420. doi:10.1007/s10853-007-2282-6.
made. Sci. Adv. 3 (7), e1700782. doi:10.1126/sciadv.1700782. Salakhov, I.I., Shaidullin, N.M., Chalykh, A.E., Matsko, M.A., Shapagin, A.V., Batyr-
Haider, T.P., Völker, C., Kramm, J., Landfester, K., Wurm, F.R., 2019. Plastics of the shin, A.Z., Shandryuk, G.A., Nifant’ev, I.E., 2021. Low-temperature mechanical prop-
future? The impact of biodegradable polymers on the environment and on society. erties of high-density and low-density polyethylene and their blends. Polymers (Basel)
Angew. Chem. Int. Ed. 58 (1), 50–62. doi:10.1002/anie.201805766. 13 (11), 1821. doi:10.3390/polym13111821.
Hu, C., Lv, T., Li, J., Huang, S., Li, H., Chen, J., Yu, D., de Claville Christianse, J., Schindler, A., Doedt, M., Gezgin, Ş., Menzel, J., Schmölzer, S., 2017. Identification of
Jiang, S., An, L., 2019. Conformational energy settled crystallization behaviors of polymers by means of DSC, TG, STA and computer-assisted database search. J. Therm.
poly (L-lactic acid). ACS Appl. Polym. Mater. 1 (9), 2552–2560. doi:10.1021/acsapm. Anal. Calorim. 129 (2), 833–842. doi:10.1007/s10973-017-6208-5.
9b00722. Schymanski, D., Goldbeck, C., Humpf, H.U., Fürst, P., 2018. Analysis of microplastics in
ISO 17556, 2019. Plastics—determination of the ultimate aerobic biodegradability of plas- water by micro-raman spectroscopy: release of plastic particles from different packag-
tic materials in soil by measuring the oxygen demand in a respirometer or the amount ing into mineral water. Water Res 129, 154–162. doi:10.1016/j.watres.2017.11.011.
of carbon dioxide evolved. ISO https://www.iso.org/cms/render/live/en/sites/ Shabaka, S.H., Ghobashy, M., Marey, R.S., 2019. Identification of marine microplastics in
isoorg/contents/data/standard/07/49/74993.html. eastern harbor, mediterranean coast of Egypt, using differential scanning calorimetry.
Iwata, T., 2015. Biodegradable and bio-based polymers: future prospects of eco-friendly Mar. Pollut. Bull. 142, 494–503. doi:10.1016/j.marpolbul.2019.03.062.
plastics. Angew. Chem., Int. Ed 54 (11), 3210–3215. doi:10.1002/anie. Shieh, Y.T., Chuang, H.C., 2001. DSC and DMA studies on silane-grafted and
Kargarzadeh, H., Galeski, A., Pawlak, A., 2020. PBAT green composites: effects of kraft water-crosslinked LDPE/LLDPE blends. J. Appl. Polym. Sci. 81 (7), 1808–1816.
lignin particles on the morphological, thermal, crystalline, macro and microme- doi:10.1002/app.1614.
chanical properties. Polymer (Guildf) 203, 122748. doi:10.1016/j.polymer.2020. Shim, W.J., Hong, S.H., Eo, S.E., 2017. Identification methods in microplastic analysis: a
122748. review. Anal. Methods 9 (9), 1384–1391. doi:10.1039/C6AY02558G.
Larsen, Å.G., Olafsen, K., Alcock, B., 2021. Determining the PE fraction in recycled PP. Signori, F., Coltelli, M.B., Bronco, S., 2009. Thermal degradation of poly (lac-
Polym. Test. 96, 107058. doi:10.1016/j.polymertesting.2021.107058. tic acid) (PLA) and Poly (butylene adipate-co-terephthalate) (PBAT) and
Lehner, R., Weder, C., Petri-Fink, A., Rothen-Rutishauser, B., 2019. Emergence of their blends upon melt processing. Polym. Degrad. Stab. 94 (1), 74–82.
Nanoplastic in the Environment and Possible Impact on Human Health. Environ. Sci. doi:10.1016/j.polymdegradstab.2008.10.004.
Technol. 53 (4), 1748–1765. doi:10.1021/acs.est.8b05512. Suzuki, Y., Akamatsu, F., Nakashita, R., Korenaga, T., 2010. A novel method to discrimi-
Ma, P.M., Wang, R.Y., Wang, S.F., Zhang, Y., Zhang, Y.X., Hristova, D., 2008. Ef- nate between plant-and petroleum-derived plastics by stable carbon isotope analysis.
fects of fumed silica on the crystallization behavior and thermal properties of poly Chem. Lett. 39 (9), 998–999. doi:10.1246/cl.2010.998.
(hydroxybutyrate-co-hydroxyvalerate). J. Appl. Polym. Sci. 108 (3), 1770–1777. Su, S., 2021. Prediction of the miscibility of PBAT/PLA Blends. Polymers (Basel) 13 (14),
doi:10.1002/app.27577. 2339. doi:10.3390/polym13142339.
Majewsky, M., Bitter, H., Eiche, E., Horn, H., 2016. Determination of mi- Vyazovkin, S., 2002. Thermal analysis. Anal. Chem. 74, 2749–2762.
croplastic polyethylene (pe) and polypropylene (pp) in environmental sam- doi:10.1021/ac020219r.
ples using thermal Analysis (TGA-DSC). Sci. Total Environ. 568, 507–511. Wen, H., Wang, Y., Wang, D., de Claville Christiansen, J., Yu, D., Jiang, S., Chen, C., 2019.
doi:10.1016/j.scitotenv.2016.06.017. Evaluation of relationship between crystallization structure and thermal-mechanical
Mecozzi, M., Pietroletti, M., Monakhova, Y.B., 2016. FTIR spectroscopy supported by sta- performance of PLA with MCC addition. ChemistrySelect, 4 (34), 10174–10180.
tistical techniques for the structural characterization of plastic debris in the marine doi:10.1002/slct.201902015.
environment: application to monitoring studies. Mar. Pollut. Bull. 106 (1–2), 155– Yeh, H.W., Epstein, S., 1981. Hydrogen and carbon isotopes of petroleum
161. doi:10.1016/j.marpolbul.2016.03.012. and related organic matter. Geochim. Cosmochim. Acta. 45 (5), 753–762.
Menczel, J.D., Prime, R.B., 2009. Thermal Analysis of Polymers: Fundamentals and Ap- doi:10.1016/0016-7037(81)90046-6.
plications. John Wiley & Sons, New Jersey doi:10.1002/macp.200900423. Yu, T., Li, Y., 2014. Influence of Poly (butylenes adipate-co-terephthalate) on the proper-
Millican, J.M., Agarwal, S., 2021. Plastic pollution: a material problem? Macromolecules ties of the biodegradable composites based on ramie/Poly (lactic acid). Composites,
54 (10), 4455–4469. doi:10.1021/acs.macromol.0c02814. Part A, 58, 24–29. doi:10.1016/j.compositesa.2013.11.013.
Nazareth, M., Marques, M.R.C., Leite, M.C.A., Castro, Í.B., 2019. Commercial plastics Yuan, Y., Zhang, W., Zhang, Y., Liu, Z., Shao, S., Zhou, L., Rogers, K.M., 2018. Differ-
claiming biodegradable status: is this also accurate for marine environments? J. Haz- entiating organically farmed rice from conventional and green rice harvested from
ard. Mater. 366, 714–722. doi:10.1016/j.jhazmat.2018.12.052. an experimental field trial using stable isotopes and multi-element chemometrics. J.
Nisticò, R., 2020. Polyethylene terephthalate (pet) in the packaging industry. Polym. Test. Agric. Food Chem. 66 (11), 2607–2615. doi:10.1021/acs.jafc.7b05422.
90, 106707. doi:10.1016/j.polymertesting.2020.106707. Zhang, P., Liu, W., Ma, J., Wu, J., Lu, X., 2016. Carbon dioxide concentration can limit the
Plastics Europe, 2022. Plastics-the facts 2022. https://plasticseurope.org/knowledge-hub/ identification of C4 plants by stable isotope composition. Arabian J. Geosci. 9 (12),
plastics-the-facts-2022/ (accessed 20 November 2022). 1–12. doi:10.1007/s12517-016-2621-5.