Polymer Degradation and Stability 98 (2013) 972e979

Contents lists available at SciVerse ScienceDirect

Polymer Degradation and Stability

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

Analysis of recycled PET bottles products by pyrolysis-gas chromatography

Nino Dimitrov a, Ljerka Kratofil Krehula b, *, Anita Pti
cek Siro
ci
c b, Zlata Hrnjak-Murgi
cb
a
Croatian National Institute of Public Health, Object of Common Use and Medical Device Department, Rockefellerova 7, HR-10000 Zagreb, Croatia
b
Faculty of Chemical Engineering and Technology, University of Zagreb, Marulicev trg 19, P.O. Box 177, HR-10000 Zagreb, Croatia

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

Article history: The impact of the contaminants in post used PET bottles on the decomposition of polymer itself after
Received 27 November 2012 recycling process was analyzed. For that reason sorption experiments of harmful substances to PET
Accepted 17 February 2013 bottles were performed. In order to monitor the presence of contaminants in recycled PET and evolution
Available online 4 March 2013
of thermal decomposition products of samples: virgin PET, contaminated PET flakes (PET K) and recycled
PET (PET RM) were studied. The thermal degradation products of PET were achieved by pyrolysis using
Keywords:
Pyrolysis-Gas Chromatography/Mass Spectrometry (Py-GC/MS), temperature-programmed evolved gas
Pyrolysis-gas chromatography
analysis-mass spectrometry (EGA-MS) and thermogravimetric analysis (TGA). Reactive pyrolysis with
Evolved gas analysis
Reactive pyrolysis
tetramethyl ammonium hydroxide (TMAH) was also performed to gather additional information about
Recycling of PET bottles degradation products in studied samples. The pyrolysis products analysis of the studied PET samples
Poly(ethylene-terephthalate) indicate some changes in decomposition mechanism of PET samples that pass through contamination
Tetramethyl ammonium hydroxide process due to some diversity in their chemical and structural composition.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction detection of penetrated harmful substances in recycled PET. On the

Poly(ethylene-terephthalate) (PET) is a polymer in widespread
use in the production of bottles and containers of water and
carbonated drinks [1]. In order to minimize polymer waste and
ensure environmental protection the methods of polymer recycling
are continuously developing. The quantity of collected PET waste is
enormously growing but in 2007 just about 40% of all PET bottles in
the market were collected for recycling [2]. Unfortunately, the
properties of the reclaimed materials of waste polymers, subjected
to recycling, are often partly deteriorated in comparison with those
of the virgin materials. It is very important to overcome these de-
ficiencies during recycling process and obtain satisfied mechanical
properties, durability, aesthetics and especially the health suit-
ability of the recycled material [3,4]. The main concern in recycling
of post consumer PET for food contact is secondary use of plastic
packaging on the consumer side. It is assumed that in usual
distribution and use of PET bottles one of five bottles may be a non-
food bottle [5], which is explained by its secondary use for storage
of various types of chemicals (cosmetics, solvents, detergents,
motor oils etc.). For that reason many published studies [6,7] were
focused on PET bottles cleaning efficiency before recycling and to
* Corresponding author. Tel.: þ385 14597121; fax: þ385 4597142.
E-mail addresses: nino.dimitrov@hzjz.hr (N. Dimitrov), krehula@fkit.hr
(L. Kratofil Krehula), apticek@fkit.hr (A. Pti
cek Siro
ci
c), zhrnjak@fkit.hr (Z. Hrnjak-
Murgi c).
other side, it should be considered the question: does the con-
taminants in recycling process have the impact on material final
properties and its structural stability? For this purpose, to obtain
more realistic estimation of the concentration of contaminants in
polymer, the simulation of the contamination by secondary use is
performed in this work by test mixture of hazardous chemicals.
Test mixture contains contaminants [8], which simulate objective
conditions that occur during the use of PET bottles. After the
contamination procedure and recycling process PET bottles are
analyzed but some traces of chemicals are sometimes undetectable
due to their very low concentration. Even using advanced spec-
troscopic methods, such as high-field solution NMR, the charac-
terization of the contaminants traces that can penetrate in recycled
polymer has not been an easy task.
That gives us the reason to study contamination of PET bottles to
gather more information about PET polymer structure stability by
analyzing pyrolysis products that are result of chemical structure of
the sample. In this work, recycled PET was studied by Py-GC/MS in
terms of establishing the potential remaining contamination and
degradation products after material recycling process. Moreover,
reactive Py-GC in the presence of organic alkali (thermally assisted
hydrolysis and methylation (THM) such as tetramethyl ammonium
hydroxide (TMAH) often provides useful information for the
condensation type of polymer materials containing ester and/or
carbonate linkages [9,10]. Thus, the possible very small amount of
contaminants, which penetrated into the structure of the polymer,

0141-3910/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.polymdegradstab.2013.02.013
 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979 973

can redirect the pyrolysis decomposition, which can be indentified Table 1

by following the pyrolysis products. Chemical properties of six contaminants in the test mixture.

Model CAS Molecular Properties Conc.

2. Experimental compound number mass, mg/l
g mol1

2.1. Materials
108-88-3 92.1 - aromatic hydrocarbon 1000
The post consumed PET bottles were milled into PET flaks - volatile, non-polar
(intrinsic viscosity 0.7495 dl/g) then were contaminated and - used in adhesives and
further were used in recycling process to obtain recycled PET dyes; constituent of
samples. Irganox 1076 (CIBA, phenolic primary antioxidant for motor oils and
automotive
processing) and Zn-stearate (Chemson Polymer Additive AG,
compounds
Austria, lubricant for extrusion) were used during recycling. For the
contamination of PET flakes the test solvent mixture [11] was used,
containing: toluene (99.5%, Carlo Erba), D-limonene (99.9%,
Fluka) 1,1,1-trichloroethane (97%, Sigma Aldrich), phenyl- 138-86-3 136.2 - hydrocarbon 1000
- volatile, non-polar
cyclohexane (97%, Sigma Aldrich), benzophenone (99.5%,
- often found in recycled
Dr.Ehrensorfer), and chlorobenzene, (99.7%, Sigma Aldrich). PET (main Compound
Dichloromethane (DCM) (p.a., Kemika, Zagreb) was used as a sol- of citrus oil based
vent for the test mixture and reagent of tetramethyl ammonium flavors in fruit drinks)
hydroxide (TMAH) (electronic grade 99.9999%, Alfa AesarÒ) in
methanol solution, 25 wt% was used to induce reactive pyrolysis. 25323-89-1 133.4 - halogenated aliphatic 1000
hydrocarbon substitute
2.2. Contamination procedure - very volatile, polar

PET flakes (200 g) were contaminated with the test solvent 827-52-1 160.3 - aromatic hydrocarbon 1000
mixture to simulate the worst possible scenario of contamination of - non-volatile, non-polar
bottles before recycling into new PET bottles for food application. - substance proposed
by ILSI-Europe for
The test solvent mixture was prepared by mixing the chemicals in
application as a model
DCM (152.5 ml) and the concentration of each contaminant in the compound in
solution was 1000 ppm, Table 1. The contamination process was Challenge tests
carried out in 1 L reactor with reflexive heat sink to prevent mass
loss of contaminants because the suspension was heated up to
119-61-9 182.2 - aromatic ketone 1000
40  C for 24 h with continuous mixing. After then suspension was
- non-volatile, non-polar
left in the reactor for another 10 days and mixed each day for 2 h. - substance proposed
Thus, contaminated PET flakes were filtered from suspension, by ILSI-Europe for
washed with DCM and dried at air atmosphere for 3 days and for application as a model
another 6 h under controlled temperature (180  C) before recycling compound in Challenge
tests
by melt extrusion. Obtained contaminated PET flakes were denoted
as PET K samples and further used in analysis.
108-90-7 112.6 - halogenated aromatic 1000
2.3. Recycling process hydrocarbon
- volatile, non-polar
- swelling agent
Contaminated PET flakes were recycled by melt extrusion that
was carried out on the Rondol Bench Top 21 twin-screw extruder to
obtain recycled PET samples (PET RM). PET flakes were homoge-
nized with thermal stabilizer Irganox 1076 (0.5 mass %) and the Lab, Ultra ALLOYþ-5; 30 m long  0.25 mm i.d, 0.25 mm film thick-
lubricant Zn-stearate (1.0 mass %) by dry mixing before extrusion. ness) under a programming temperature condition; GC oven was
Temperature profile in the extruder was set to 220/230/230/240/ programmed at 40  C with the heating rate of 20  C/min up to
240/245  C (from hopper to die) and screws rotation speed was set 300  C. The pyrolysis interface and GC injector were kept at 320  C.
to 100 rpm. GC Injection port was in constant linear velocity mode 48.8 cm/s
with split ratio 1/60. Mass spectrometer was operated in the EI mode
2.4. Py-GC/MS analysis with mass range of 29e600 amu, scanned at 1250 amu/sec.
The reactive pyrolysis with tetramethyl ammonium hydroxide
The pyrolysis of virgin PET, PET K (contaminated flakes) and PET (TMAH) was carried out by the same Py-GC/MS system under the
RM (recycled PET) samples were carried out in a multifunctional same conditions. About 2 ml of TMAH was added to powdered PET
double-shot pyrolyzer (Frontier Lab., PY-2020iD). The vertical micro- sample (0, 1 mg) into stainless steel sample cup first set at the
furnace pyrolyzer mounted on a Shimadzu GCeMS-QP2010 Plus was upper position of the pyrolyzer at around room temperature, and
used under a flow of He carrier gas, and data search libraries of then introduced into the centre of the pyrolyzer heated at 300  C
Frontier Lab., F-search (MS 08) and NIST/EPA/NIH (NIST 05) were and for the second measurement at 400  C.
used for the peak identification. The flash pyrolysis temperature was
fixed at 600  C and the resulting pyrolyzates of the PET granule 2.5. EGA-MS analysis
samples of 0.2 mg were placed into the deactivated stainless steel
sample cup separated by a metal capillary separation column coated Evolved gas analysis-Mass spectrometry (EGA-MS) measure-
with immobilized 5% diphenyl-95% dimethylpolysiloxane (Frontier ments have been carried out in a multifunctional double-shot
974 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979

pyrolyzer micro-furnace was directly coupled with an ion source of

MS via a deactivated and uncoated stainless steel transfer tube
(UADTM-2.5N, 0.15 mm i.d.  2.5 m length, Frontier Lab). The
column oven temperature was maintained isothermal at 300  C
and the temperature of the micro-furnace pyrolyzer was pro-
grammed from 100 to 700  C at a heating rate of 20  C/min under a
helium flow. PET granules of about 0.2 mg were placed into steel
sample cup. Injection port was in constant linear velocity mode at
50 kPa with split ratio 1/50. Micro-furnace interface temperature
was kept at 320  C in auto mode and interface temperature of MS
kept at 280  C.
By EGA-MS method the reactive pyrolysis was performed in the
presence of TMAH under the same operating conditions for the all
studied PET samples. About 2 ml of TMAH was added to powdered
PET sample (0, 1 mg) in the sample cup, and then introduced into
the centre of the micro-furnace pyrolyzer.

2.6. TG analysis

The pyrolysis and reactive pyrolysis of the studied PET K and PET
RM samples were monitored by thermogravimetric analysis (TGA)
using TA Instrument Q500. Several granules of samples weighted at
w10 mg were heated from room temperature to 600  C at heating
rates of 5  C/min in a nitrogen gas flow of 100 ml/min. About 10 mg
of the studied PET samples were also chemically derivatized by
addition of 35 mL of TMAH reagent in the sample cup.

3. Results and discussion

3.1. Py-GC/MS analysis

The pyrograms and pyrolysis products obtained at 600  C for

virgin PET, PET K (contaminated flakes) and PET RM (recycled PET)
samples are shown in Fig. 1 and Table 2. The pyrograms describe the
peaks seen in the total ion chromatogram (TIC), which represents
the number of compounds detected by GCeMS prior to the pyrolysis
step. For calculation of area percentage of each compound, area
normalization calibration with amount of 0.2 mg PET sample was
performed. The main pyrolysis products for studied PET samples are:
carbon dioxide/acetaldehyde (1), benzene (8), vinyl benzoate (29),
benzoic acid (32), divinyl terephthalate (52), 4-(vinyloxycarbonyl)
benzoic acid (55), ethan-1,2-diyldibenzoate (80). It has been well
accepted that the first step in the degradation of PET polymer during
pyrolysis is the classical ester scission reaction producing carboxylic
acid and olefin end groups [12]. This scission may then yield many
different gaseous substances (primary products), which further
react react among themselves or with oxygen, if the decomposition
has taken place in air atmosphere (secondary products) [13].
The obtained products showed some similarity in pyrograms
(Fig. 1aec) although observed mechanism of the thermal degrada-
tion differs for various PET samples, which can be concluded from
different concentration of degradation products, Table 2. The highest
carbon dioxide/acetaldehyde concentration shows sample PET K
(43.93%) then follows PET RM (33, 61%) and virgin PET 20.73%. The
opposite trend is observed for the second main pyrolysis product 4-
(vinyloxycarbonyl) benzoic acid: PET K 3.66%, PET RM 14.11% and
virgin PET 27.08%. Obtained results show that the degradation of PET
K sample mostly results with the formation of low molecular prod-
ucts (CO2, acetaldehyde) caused by accelerated pyrolytic degradation
due to catalytic effect of remained contaminants (like D-limonene,
chlorobenezene, benzophenone), Scheme 1. On the other hand, vir-
gin PET decomposes mostly on the products with higher molecular
mass (4-(vinyloxycarbonyl) benzoic acid, divinyl terephthalate and
phthalate fragment) due to absence of such contaminants. The
recycled sample PET RM shows product concentrations in between
Fig. 1. Pyrograms of the studied samples: a) Virgin PET, b) PET K, c) PET RM.
those of PET K and virgin PET samples. During the recycling process
the chemicals from the test solvent mixture almost completely
evaporate [14] but their remained very low concentration affects the
mechanism of pyrolytic degradation. In the recycled sample some
traces of contaminants from the solvent test mixture can be observed
as phenylcyclohexane and D-Limonene. It is assumed that phenyl-
cyclohexane may be detected as dehydrogenation to biphenyl (43)
that is formed due to pyrolysis temperature but has not been
detected with PY-GC/MS. D-Limonene was also detected in PET
K ¼ 0.22% and in very small amount in PET RM ¼ 0.09%. The fact that
limonene and phenylcyclohexane are detectable in the recycled
sample indicates penetration of the contaminants deeply in the PET
polymer structure that is not possible to remove efficiently by
cleaning and recycling process. On the other hand, in sample PET K
some obtained pyrolysis products are toxic and hazardous: benzene,
toluene, styrene, phenol, benzofuran, biphenyl and others (Table 2)
due to higher concentration of contaminants. Further differences
between the studied samples can be seen through the formation of
cyclic compounds (cyclopropane (m/z 68), cyclobutanone (m/z 42)
for virgin and PET K samples while in PET RM linear hydrocarbons
(pentadiene m/z 67, 1-hexene m/z 41, 1-heptene m/z 41, and 3-
eicosene m/z 83) are formed. Also, it is assumed that the obtained
product 3-phenyl-2H-chromene (m/z 207) is the consequence of the
crosslinking reaction of PET chains because its presence in increased
concentration is results of the recycling.
3.2. Py-GC/MS analysis in the presence of TMAH
The reactive pyrolysis is a well-established method for the
characterization of complex and intractable samples [15] where the
 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979 975

Table 2
Pyrolysis products of studied PET samples and their concentration.

ID# Name m/z Mwt%

Virgin PET PET K PET RM

1 Carbondioxide/Acetaldehyde 44 20.73 43.93 33.61
2 Cyclopropane, 1,10 -methylenebis- 68 0.28 1.83 e
3 1,3-Pentadiene 67 e e 0.47
8 Benzene 78 2.75 5.80 2.76
11 Toluene 91 0.36 2.05 0.38
16 Phenylethyne 102 4.24 e e
17 Styrene 104 0.29 0.41 0.28
19 Phenol 94 0.10 0.71 0.07
21 1-Hexene, 3,3,5-trimethyl 69 e 0.11 0.15
22 D-Limonene 71 e 0.22 0.09
23 Benzeneacetaldehyde 91 0.22 e 0.12
29 Vinyl benzoate 105 4.82 1.41 3.69
32 Benzoic acid 105 10.10 11.82 17.45
33 Benzofuran, 2,3-dihydro- 120 e 0.15 e
43 Biphenyl 154 1.25 0.89 1.14
52 Divinyl terephthalate 175 4.60 0.69 2.51
55 4-(vinyloxycarbonyl) benzoic acid 149 27.08 3.66 14.11
56 Terephthalic acid 149 2.88 1.58 2.36
60 3-Eicosene, (E)- 83 e e 0.14
62 Cyclopentadecanone, 2-hydroxy- 41 e 0.24 e
66 Benzophenone-2,40 -dicarboxylic acid 149 e 0.26 e
68 (1,10 -Biphenyl)-3-carboxylic acid 181 1.70 e 1.41
72 Dodecanoic acid 73 e 2.03 1.95
79 Octadecanoic acid, 2-(2- 73 0.11 e 1.10
hydroxyethoxy)ethyl ester
80 1,2-Ethanediol, dibenzoate 105 4.93 1.60 4.92
82 3-Phenyl-2H-chromene 207 0.04 e 1.83
83 Phthalic acid, methyl neopentyl ester 163 0.28 1.84 0.48
Volatile products 13.24 17.03 8.98

presence of a alkali reagent like TMAH significantly lowers pyrol-
ysis temperature (400  C). It is a very effective method in analysis of
condensation polymers such as PET, especially in the analysis of
minor and/or micro structures. For the studied PET samples the
results are presented in Fig. 2 and Table 3 and imply that obtained
alkali assisted pyrolysis products are completely different by
chemical structure than those obtained without TMAH. Literature
[16] has demonstrated that the role of TMAH is not only the
methylation of the pyrolysis products of the polymers, but also
assisting in bond cleavage and in transesterification reactions.
Under alkaline conditions, the hydroxide anion attacks the carboxyl
oxygen atom to produce equivalent amounts of hydroxyl and
carboxyl end groups [17]. Thus, the main products obtained by this
organic alkali assisted pyrolysis are: trimethylamine (m/z 58),
dimethyl terephthalate (m/z 163) and methyl (m/z 31) and ethyl
(m/z 45) alcohol, Scheme 1. However, due to differences in the
nature of the samples, the concentrations of the major pyrolysis
pyrolysis
CO2 / CH3CHO + COOH
400 °C
CO2 / acetaldehyde benzoic acid
+ CH2CHOOC COOH
vinyloxycarbonyl benzoic acid
-[OOC
[
COOCH2CH2 -n
PET
N(CH3)3 + H3COOC
+ TMAH
Trimethylamine
Alkaline
medium + CH3OH / CH3CH2OH
methyl / ethyl alcohol
Scheme 1. Products of PET pyrolysis.
Fig. 2. Pyrograms of the studied samples in the presence of TMAH: a) Virgin PET, b)
PET K, c) PET RM.
products differ. Thus the concentration of trimethylamine in virgin
PET, PET RM and PET K is 39.29%, 30.25%, 26.26%, and dimethyl
terephthalate concentration is 14.92, 27.02 and 25.61, respectively.
It can be concluded that the dominant reaction in virgin PET sample
is methylation while the reactions of methylation and trans-
esterification are present in approximately equal proportions in PET
K and PET RM. The heat during alkali analysis induces reaction
between sample acidic functional groups [15] and formation of a
free carboxylic acid and ethenyl ester-group-terminated macro-
molecular fragment leads finally to formation of dimethyl tere-
phthalate. Furthermore, it is obtained that most of chemical
compounds formed by alkali pyrolysis are methyl based, while
noncatalytic pyrolysis results in vinyl based compounds and the
compounds of higher molecular masses. The important degrada-
tion product for the studied PET samples is also methylalcohol,
which concentration is in the range from 4 to 9% as well as the
fragments originating from diethylene glycol (DEG): 2-metoxy
ethanol, 1,2-dimethoxy ethane, 1,2-Ethanediol, Diethylene glycol
Acid
monoethyl ether. Thus, the alkali induced pyrolysis of PET samples
medium
showed the formation of alcohols and corresponding methyl ethers
formed from alcohols and phenols (Table 3). Results showed that
more fragments originated from diethylene glycol in PET RM
sample (m/z 45, 2-metoxy ethanol e 10.94%) is obtained. As
explained above, these residual fragments, especially from dimer-
COOCH3
ized glycols, had a low-scale contribution, but their formation can
dimethyl terephthalate be very useful in explaining the recycling and contamination pro-
cesses on PET sample. Pyrogram of PET RM sample (Fig. 2) also
showed a number of assigned small peaks proposed to originate
from DEG, which is a depolymerization product of dehydrogenation
reaction of ethylene glycol (EG). Possible reason for higher forma-
tion of DEG units is mechanical recycling. Dehydration of ethylene
976 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979

Table 3 intensities at 443  C (Table 4) for virgin PET, while PET RM and PET
Pyrolysis products obtained with TMAH of studied PET samples and their concen- K samples demonstrate slight shift of the maximum peak to higher
tration at 400  C.
temperatures. The observed temperatures reveal differences in
ID# Name m/z Mwt% polymer structure of the studied samples as well as presence of the
Virgin PET PET K PET RM contaminants, which initialized and changed mechanism of poly-
1 Carbon dioxide 44 8.10 e e mer thermal degradation. This statement is supported by the re-
2 Methyl Alcohol 31 8.95 4.98 8.54 sults of the initial degradation temperatures where peaks at 182  C
3 Trimethylamine 58 39.29 26.26 30.25 for PET K and at 241  C for PET RM can be observed. According to
4 2-Propanone, 1-hydroxy- 43 2.24 e e
corresponded mass spectra of evolved gases (Fig. 4) all samples
5 Methylene Chloride 43 e 1.29 e
6 Diethylene glycol monoethyl ether 45 e e 0.36
showed mass fragments characteristic for PET polymer: the peaks
7 Ethanol, 2-methoxy- 45 0.59 7.66 10.94 at m/z 51, 77, 105 and 122 characteristic for benzoic acid and vinyl
8 Ethane, 1,2-dimethoxy 45 0.07 4.42 e benzoate that origin form PET polymer. However, those m/z values
10 1,2-Ethanediol 31 e 1.06 1.18 can also partially originate from contaminants of chlorobenzene
20 Ethanol, 2-(2-methoxyethoxy)- 45 e 0.50 0.37
and benzophenone. The difference in their intensities shows
21 Ethane, 1,10 -oxybis[2-methoxy- 59 e 0.28 0.48
24 D-Limonene 68 e 0.01 e impact of contamination and recycling process on PET. It is
29 Benzoic acid 105 1.20 1.49 0.04 confirmed in literature [20e22] that benzoic acid and vinyl ben-
37 1,2-Ethanediol, monobenzoate 105 1.24 0.22 e zoate were identified as one of the main gas products of the PET
38 Phenol, 4,6-di(1,1-dimethylethyl 205 e e 2.29 pyrolysis formed from phenyl end group of PET oligomers.
)-2-methyl-
40 dimethyl terephthalate 163 14.92 25.61 27.02
Furthermore, at the temperatures above 300  C, PET pyrolysis
41 1,3-Benzenedicarboxylic acid, 163 1.26 0.23 0.10 proceeds via random-chain scission of the ester links [23,24] to
dimethyl ester yield a mixture of terephthalic acid, vinyloxycarbonyl benzoic acid
47 4-Acetylbenzoic acid 149 1.45 5.65 0.92 and divinyl terephtalate that corresponds to m/z 65, 121, 149, 175
50 1,3-Benzenedicarboxylic acid 149 0.80 2.67 0.23
that are characteristic of a phthalate fragment. Higher intensities of
66 1,3-Benzenedicarboxylic acid, dimethyl 163 e 1.91 e
ester m/z 29 and 44 characteristic of condensable gases such as acetal-
82 2-Ethoxycarbonyl-3-methyl-4- 430 e e 0.03 dehyde and CO2 in PET K sample can be also seen. Acetaldehyde
azafluorenone, 2-fluorenylimime formation from PET must be taken into consideration especially if
83 (Dimethylamino)furfurylidene]-2,2- 163 e 0.11 e the material is in contact with food because in foodstuffs and
dimethyl-1,3-dioxane-4,6-dione
91 2-Sulfamyl-4-nitro-40 - 367 0.05 3.56 0.69
beverages migration of acetaldehyde can change organoleptic
acetamidodiphenyl sulfide properties [25].
93 2,6,10,14,18,22-Tetracosahexaene, 69 5.76 3.96 2.31
2,6,10,15,19,23-hexamethyl-, (all-E)-
Volatile products 14.08 8.13 14.27
3.4. EGA/MS analysis in the presence of TMAH

EGA/MS analysis was also carried out in the presence of alkali

reagent TMAH and obtained EGA thermograms are presented in
glycol during recycling and polymerization processes results in Fig. 5. In the thermograms various shape of EGA curves particularly
residual amounts of dimerized diethylene glycol which determine for the contaminated and recycled samples are observed. The
the quality of recycling process [18]. maximum of curves occurs in the temperature region from 160  C
Although the reactive alkali pyrolysis results in formation of one to 340  C showing highly increased rate of PET hydrolysis for the
long chain (C30H50) product (tetracosahexaene, ID 93, m/z 69), studied PET samples. EGA thermograms of virgin PET, PET K and
more various types of long chain products are formed by non- PET RM show approximately the same number of maximums
catalytic pyrolysis (3-eicosene, 1-hexene, heptene). The various (Table 4) that slightly differ in temperature values. The major part
concentration and higher number of catalytic pyrolysis products of thermal decomposition of virgin PET occurs from 230  C to
indicate various degradation velocity and mechanism. So, the virgin
PET has the lowest number of pyrolysis products while the PET K
sample, which is the most contaminated, produced higher number
of different compounds. It can be also seen that products of virgin
PET are mostly lower molecular masses compounds (CO2, trime-
thylamine, methylalcohol) and the higher masses compounds like
dimethyl terephtalate are present in significantly lower fraction.
Opposite behavior is observed for PET K and PET RM samples that
were contaminated and recycled. Dimethyl terephtalate is present
in very high ratio while other low molecular mass compounds are
present in lower proportion. Furthermore, from the results it can be
observed that in this catalytic pyrolysis significantly less toxic
compounds are obtained (like methylene chloride) due to lower
temperatures of reactive alkali pyrolysis.

3.3. EGA/MS analysis

EGA/MS method was performed using the programmed heating

of the sample during which gases are evolved and directly trans-
ferred into MS to achieve on-line monitoring of the components
[19]. EGA thermograms (Fig. 3) are recorded in the temperature
interval from 160 to 700  C and are obtained from a position of total
ion current (TIC). EGA thermograms show maximum peak Fig. 3. EGA thermograms for the studied samples: a) Virgin PET, b) PET K, c) PET RM.
 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979 977

Table 4
The temperatures at maximum derivatization rate (Tmax) of EGA data for the studied
PET samples.

Sample Temperature/ C

Tmax1 Tmax2 Tmax3 Tmax4 Tmax5 Tmax6 Tmax7

Virgin PET 443 585
PET K 182 452
PET RM 241 446 598
Virgin PET þ TMAH 170 226 292 302 456
PET K þ TMAH 174 222 278 302 454 516
PET RM þ TMAH 172 222 252 280 302 455

310  C, for PET K sample the most intensive decomposition is from

170  C to 310  C while for PET RM decomposition takes place also
from 170  C to 310  C. As can be seen from the temperature in-
tervals of decomposition, it is evident that the virgin PET degrades
at higher temperature in one step, while the other two samples,
which were contaminated and recycled, begin to decompose at
significantly lower temperatures and in multiple steps. The PET K
sample decomposes in two basic steps, while PET RM sample de-
composes in four steps that are clearly seen from the temperature
maximums in Table 4. The existence of different polymer chemical
Fig. 5. EGA thermograms for the studied samples with the presence of TMAH. a) Virgin
PET, b) PET K, c) PET RM.
structure results in a lower thermal stability and decomposition of
such structures in separate steps. It is obvious from the results of
reactive alkali pyrolysis that remained contaminants and recycling
process strongly affected the thermal degradation of PET polymer.
On the mass spectra of EGA thermograms (Fig. 6) for all studied
samples are detected the main fractions of m/z ratios 30, 42, 58, 59,
76, 103, 135, 149, 163, 179 and 194 suggesting the formation of
trimethylamine and dimethyl terephthalate as well as methyl and

Fig. 4. Average mass spectra of thermograms the studied samples: a) Virgin PET, b) Fig. 6. Average mass spectra of thermograms for the studied samples with the pres-
PET K, c) PET RM. ence of TMAH: a) Virgin PET, b) PET K, c) PET RM.
978 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979

Table 5
The temperatures at maximum degradation rate (Tmax) and char residue of TGA data
for the studied PET samples.

Samples Temperature/ C Char/%

Tmax1 Tmax2 Tmax3 Tmax4 Tmax5

Virgin PET e e e 435 12.0
PET K e 202 234 433 8.3
PET RM e e e 249 434 8.3
Virgin PET þ TMAH e 125 161 e 403 13.0
PET K þ TMAH 46 125 167 267 405 7.3
PET RM þ TMAH 55 125 173 274 406 3.0

At these temperatures about 2% of the sample is decomposed. This

could be explained by the fact that the residual impurities i.e. the
solvent test mixture starts to evaporate at such low temperature,
which corresponds to the boiling points of some contaminants
(toluene 110  C, limonene 176  C and chlorobenzene 134  C). It is
assumed that Tmax at 202  C corresponds to decomposition of
contaminants, which then initiate earlier decomposition of PET K
sample at 234  C. On the other side, it can be seen that present
unstable structures initiate PET RM decomposition at 250  C. The
presence of the unstable structures in PET polymer in recycled
sample is confirmed by Py-GC/MA analysis where higher concen-
tration of 2-metoxy ethanol is obtained that originated from
diethylene glycol (Table 3). It is assumed that small concentration of
captured contaminants in the bulk of PET RM sample induced its
earlier decomposition comparing to virgin PET. It is known from the
literature [28,29] that evolution of volatile products can initiate
degradation and strongly affect the decomposition mechanism.

Fig. 7. TGA thermograms of the studied samples.

ethyl alcohol that are fragments originated from diethylene glycol

due to alkali assisted reaction. Strong intensity of m/z 30, 42, 58
corresponds to the main product trimethylamine formed by TMAH
methylation of PET polymer while m/z 31 and 76 originate from
methyl alcohol. Other detected m/z ratios 103, 135, 149, 163, 194
represent dimethyl terephthalate and benzoic acid.

3.5. TG analysis

Thermogravimetric analysis is one of the intensively used

methods for elucidation of polymer materials thermal decompo-
sition. Fig. 7 shows mass loss (TG) and derivative mass loss (dTG)
curves with the temperature of the studied PET samples. The
thermograms of studied PET samples showed that the overall
degradation undergoes one step of mass loss between 350  C and
500  C. The temperatures of maximum decomposition rate (Tmax)
obtained from the dTG curves at Fig. 7 are presented in Table 5. The
maximum degradation rate for all samples at approximately 430  C
is observed. This mass loss is attributed to PET polymer degradation
process i.e. to a random scission of ester links in the main chain
resulting with formation of different oligomers [26e30]. From the
thermograms and values in Table 5 it can be seen that the small
fraction of polymer in PET K and PET RM samples starts to
decompose at much lower temperature than virgin PET, small
maximums are observed at 202, 234  C and at 249  C, respectively. Fig. 8. TGA thermograms of the studied samples with the presence of TMAH.
 N. Dimitrov et al. / Polymer Degradation and Stability 98 (2013) 972e979
Demertzis et al. [6] showed that smaller, generally volatile chem-
icals more easily penetrate in PET and are found in larger concen-
trations than non-volatile chemicals. Furthermore, the results of
char residue presented in Table 5, are the parameters that also
evaluate thermal stability of polymer structure. In other words,
higher char residue values indicate more thermally stabile chemical
structure of PET polymer.
Alkali assisted pyrolysis of the studied PET samples was also
performed by TG analysis and thermograms are presented on Fig. 8.
From the TG curves it can be seen that TMAH shifts the start of
degradation of all samples to significantly lower temperature. From
the results of Tmax (Table 5) it can be observed that the studied
samples decompose in several steps. It is evident that the virgin PET
sample decomposes in three and the other two samples in five
steps. Such behavior indicates different mechanism of thermal
degradation due to various chemical composition of sample. Thus,
it is seen that virgin PET shows 8% of mass loss at 125  C while
samples PET K and PET RM at the same temperature already have
lost 53 and 76 mass %. The height of maximums presents rate of
mass loss and area under the curves indicates sample mass loss
quantity of the of each degradation step. Results finally show that
more sensitive and deteriorated structure of PET polymer has
significantly low thermal stability under the strong alkaline con-
ditions. This can be seen by comparing virgin PET with other two
samples (PET K and PET RM) that undergone the process of
contamination and recycling.
4. Conclusion
From the pyrolysis products analysis of the studied PET samples
can be seen that the main products of decomposition are: carbon
dioxide, acetaldehyde, benzoic acid and 4-(vinyloxycarbonyl) ben-
zoic acid. From the results it is seen that their finally obtained con-
centration depends on the type of the studied PET samples (virgin
PET, PET K, PET RM). The main products of virgin PET are: 4-(vinyl-
oxycarbonyl) benzoic acid (27%) and low molecular substances CO2/
acetaldehyde (21%) and benzoic acid (10%). PET K and PET RM sam-
ples that pass through contamination procedure have as the main
pyrolysis products low molecular compounds such as CO2/acetal-
dehyde (44 and 34%) and the benzoic acid (17%) while only 4% of 4-
(vinyloxycarbonyl) benzoic acid has been detected. It can be
concluded that the decomposition of PET polymer in PET K and PET
RM was additionally catalyzed by hydrochloric acid formed during
thermal decomposition due to chlorine from the contaminants. Al-
kali assisted pyrolysis performed with TMAH give us insight of
thermal decomposition at lower temperatures. From the pyrolysis
products changed mechanism of decomposition can be seen. In vir-
gin PET methylation reaction predominantes in regard to methyl-
ation and transesterification reaction that are equally represented in
PET K and PET RM samples due to diversity in their chemical and
structural composition that is also supported by TGA results.
Acknowledgments
The work has been financially supported by the Ministry of
Science, Education and Sports of the Republic of Croatia.
979
References
[1] Barriocanal C, Díez MA, Alvarez R. PET recycling for the modification of pre-
cursors in carbon materials manufacture. J Anal Appl Pyrol 2005;73(1):45e51.
[2] López-Fonseca R, Duque-Ingunza I, de Rivas B, Arnaiz S, Gutiérrez-Ortiz JI.
Chemical recycling of post-consumer PET wastes by glycolysis in the presence
of metal salts. Polym Degrad Stab 2010;95(6):1022e8.
[3] Vilaplana F, Karlsson S. Quality concepts for the improved use of recycled
polymeric materials: a review. Macromol Mater Eng 2008;293:274e97.
[4] Awaja F, Pavel D. Recycling of PET. Eur Polym J 2005;41:1453e77.
[5] Begley TH, McNeal TP, Biles JE, Paquette KE. Evaluating the potential for
recycling all PET bottles into new food packaging. Food Addit Contam
2002;19:135e43.
[6] Demertzis PG, Johansson F, Lievens C, Franz R. Development of a quick
inertness test procedure for multi-use PET containers-sorption behaviour of
bottle wall strips. Pack Tech Sci 1997;10:45e58.
[7] Nir MM, Ram A, Miltz J. Sorption and migration of organic liquids in poly(-
ethylene terephthalate). Pol Eng Sci 1996;36:862e8.
[8] Food and Drug Administration. Points to consider for the use of recycled
plastics in food packaging: chemistry considerations. Washington: Chemistry
Review Branch; 1992.
[9] Challinor JM. Review: the development and applications of thermally assisted
hydrolysis and methylation reactions. J Anal Appl Pyrol 2001;61:3e34.
[10] Wampler TP, editor. Applied pyrolysis handbook. Boca Raton: CRC Press LLC;
2007. p. 177e85.
[11] Franz R, Welle F. Recycled poly(ethylene terephthalate) for direct food contact
applications: challenge test of an inline recycling process. Food Addit Contam
2002;19:502e11.
[12] Deng Y, Wang YZ, Ban DM, Liu XH, Zhou Q. Burning behavior and pyrolysis
products of flame-retardant PET containing sulfur-containing aryl poly-
phosphonate. J Anal Appl Pyrol 2006;76:198e202.
[13] Martín-Gullón I, Esperanza M, Font R. Kinetic model for the pyrolysis
and combustion of poly-(ethylene terephthalate) (PET). J Anal Appl Pyrol
2001;58-59:635e50.
[14] Pennarun PY, Dole P, Feigenbaum A. Overestimated diffusion coefficients for
the prediction of worst case migration from PET: application to recycled PET
and to functional barriers assessment. Packag Techol Sci 2004;17:307e20.
[15] Shadkami F, Helleur R. Recent applications in analytical thermochemolysis.
J Anal Appl Pyrol 2010;89(1):2e16.
[16] Tanczos I, Schöflinger M, Schmidt H, Balla J. Cannizzaro reaction of aldehydes
in TMAH thermochemolysis. J Anal Appl Pyrol 1997;42:21e31.
[17] Hosseini SS, Taheri S, Zadhoush A, Mehrabani-Zeinabad A. Hydrolytic degra-
dation of poly(ethylene terephthalate). J Appl Polym Sci 2007;103:2304e9.
[18] Yoshioka T, Handa T, Grause G, Inomata H, Mizoguchi T. Gasification of PET for
the production of syngas, at: selected papers presented at the third interna-
tional symposium on feedstock recycling of plastics. Karlsruhe:387.
[19] Shin T, Hajime O, Chuichi W. PyrolysiseGC/MS data book of synthetic poly-
mers: pyrograms, thermograms and MS of pyrolyzates. Amsterdam: Elsevier;
2011.
[20] Montaudo G, Lattimer RP. Mass spectrometry of polymers. Boca Raton: CRC
Press LLC; 2001.
[21] Tadashi A, Masuda Y. The effect of humidity on thermal decomposition of
terephthalate polyester. J Anal Appl Pyrol 2004;71(2):525e36.
_ Sınag
[22] Çit I,  A, Yumak T, Uçar S, Mısırlıoglu Z, Canel M. Comparative pyrolysis of
polyolefins (PP and LDPE) and PET. Polym Bull 2009;64(8):817e34.
[23] Balabanovich AI, Pospiech D, Häußler L, Harnisch C, Döring M. Pyrolysis
behavior of phosphorus polyesters. J Anal Appl Pyrol 2009;86:99e107.
[24] Evans JS, Haines PJ, Skinner GA. Pyrolysisegas-chromatographic study of a
series of polyester thermosets. J Anal Appl Pyrol 2000;55:13e28.
[25] Fairgrieve SF. Degradation and stabilization of aromatic polyesters. Shrop-
shire: Smithers Rapra; 2009.
[26] Lecomte HA, Liggat JJ. Degradation mechanism of diethylene glycol units in a
terephthalate polymer. Polym Degrad Stab 2006;91(4):681e9.
[27] Montaudo G, Puglisi C, Samperi F. Primary thermal degradation mechanisms
of PET and PBT. Polym Degrad Stab 1993;42(1):13e28.
 ski J. Volatile products of poly(ethylene terephthalate)
[28] Dzie˛ cio1 M, Trzeszczyn
thermal degradation in nitrogen atmosphere. J Appl Polym Sci 2000;77(9):
1894e901.
[29] McNeill IC, Bounekhel M. Thermal degradation studies of terepht-
halate polyesters: 1. Poly(alkyleneterephthalates). Polym Degrad Stab
1991;34(1e3):187e204.
[30] Jenekhe SA, Lin JW. Kinetics of the thermal degradation of polyethylene
terephthalate. Thermochim Acta 1983;61(3):287e99.