Thermochimica Acta 679 (2019) 178340

Contents lists available at ScienceDirect

Thermochimica Acta
journal homepage: www.elsevier.com/locate/tca

Thermal degradation study of waste polyethylene terephthalate (PET) under T

inert and oxidative environments
Pallab Das , Pankaj Tiwari
⁎

Department of Chemical Engineering, Indian Institute of Technology Guwahati, Guwahati, Assam 781039, India

ARTICLE INFO ABSTRACT

Keywords: Thermogravimetric analysis (TGA) of polyethylene terephthalate obtained from waste soft drink bottles (PET-
Polyethylene terephthalate SDB) was carried out under inert (nitrogen) as well as reactive (synthetic air) environments to evaluate the
Waste soft drink bottles degradation behaviour and degradation kinetics. Non-isothermal TGA experiments at heating rates of 5, 10, 20,
Isoconversional method 40 and 50 °C/min were performed and the kinetic parameters such as activation energy (Eα) and pre-exponential
Non-isothermal TGA
factor (Aα) were determined at various extent of conversions using advance isoconversional (AIC) method.
Isothermal TGA
Criados’ master plot technique established F1 (first order) as possible reaction model ( f ( ) ) for both inert and
oxygenated non-isothermal TGA data. Under inert (N2) environment a residue accumulation of around 20% was
observed whereas complete conversion of PET-SDB occurred in oxygenated environment. The activation energy
values were found in the range of 203–355 kJ/mol and 145–218 kJ/mol for inert and oxygenated environment
respectively. Isothermal TGA data at various temperatures were analysed for the estimation of kinetic para-
meters. In the considered range of conversions, the range of activation energy obtained for multiple isothermal
TGA data was found in the range of 196–217 kJ/mol.

1. Introduction commercial PET applied in textile industry and analysed the volatile
products with the help of gas chromatography. At 700 °C, pyrolysis of
Polyethylene terephthalate (PET) is one of the most common ther- PET produced CO, CO2, ethylene-acetylene, acetaldehyde, benzene,
moplastic polymer resin of polyester family with multidimensional benzoic acid, vinyl benzoate, divinyl terephthalate, acetophenone and
applications as one-dimensional fiber, two-dimensional films to three- p-acetyl vinyl benzoate. Martı́n-Gullón et al. [9] carried out the pyr-
dimensional bottles or containers including various composites with olysis of PET waste at the temperature of 200, 300, 400, 500, 600 and
other polymers, glass and ceramics etc. Two third of the PET produced 700 °C for 20 min duration in a ceramic boat (crucible) inside a tubular
is utilized for fiber and staple filaments [1–4]. Films and bottles are furnace under both nitrogen and air flow (0.025 m3/h) conditions.
mainly used as packaging materials in the food and shipping industry. Their study revealed that PET degrades within the range of 400–700 °C.
Most products from PET are mechanically recyclable and can be re- Most of the earlier work on thermal degradation of PET covers the
cycled back to its pure form by cleaning the contamination and foreign thermal stability and effect of flame retardant chemicals in PET com-
substances [5,6]. However, the PET recyclate cannot be used directly in posites [10,11].
food packaging, because of hygiene concerns. Other recycling processes To utilize the thermal processes in the disposal of PET waste a de-
like solvolysis, glycolysis, methanolysis, bio-degradation [7] etc. can tailed kinetic study is necessary to understand the reaction mechanism
produce terepathalic acid (TPA), dimethylterephthalate (DMT), ethy- under different temperature program. Buxbaum [12], established that
lene glycol (EG) bis-2-hydroxyethylene terephthalate (BHET), mono- PET follows a random scission of the ester links under thermal treat-
mers and oligomers etc. [2–4]. Plastic waste including post-consumer ment which involves a six centred cyclic transition state to produce
PET has become a worldwide concern due to the huge generation and vinyl ester and carboxylic acid. These primary products of pyrolysis
the complexity involved in the recycle and disposal processes. The PET then undergo secondary reaction to give various products like CO, CO2,
waste can also be treated in high temperature processes like combustion acetaldehyde, aromatic acid and their vinyl esters. Saha et al. [13]
and pyrolysis. These processes are conducted at high temperature carried out TGA based kinetic study with multiple heating mass loss
(300 °C – 1000 °C) and at low pressure (1 atm). data of waste PET obtained from different soft drink bottles. Both nth
Bednas et al. [8] carried out extensive study on the pyrolysis of order kinetics and ASTM E698 models were used and the activation

⁎
Corresponding author.
E-mail address: [email protected] (P. Das).

https://doi.org/10.1016/j.tca.2019.178340
Received 24 April 2019; Received in revised form 16 July 2019; Accepted 17 July 2019
Available online 20 July 2019
0040-6031/ © 2019 Elsevier B.V. All rights reserved.
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

energy in the range of 320–338.98 kJ/mol was reported. The same re- Table 1
search group [14] also carried out TGA based kinetic analysis utilizing Proximate analysis of PET-SDB sample (%w/w).
model free approach to find out the variable activation energy of PET Moisture content Volatile matter Ash content Fixed carbon
degradation reaction and suggested that the variable kinetic parameters (MC) (VM) (AC) (FC)
are more certain to predict the degradation mechanism of plastics like
0 97.31 0.209 2.47
PET under simulated heating condition. Martin-Gullon [9] predicted
the degradation reaction model based on two separate portion of the
mass loss curve; first one where original PET sample thermally de-
possible heating rate (100 °C/min) was chosen to reach the desired
graded (around 80%) quickly into char and volatiles at around
temperature with less induction period. The gas flow rate for all TGA
377–527 °C. In the second part, further increase in temperature slowly
experiments was maintained at 60 ml/min (40 ml/min as purge and
decomposed (6–7%) the char. The char (residue) may be combusted to
20 ml/min as protective gas) and the mass vs time/temperature data
reduce the residual amount.
were recorded.
The kinetic analysis of PET (waste) has been reported in the lit-
erature, however very few studies estimated all three kinetic para-
meters, viz. activation energy (Eα), pre-exponential factor (Aα) and 3. Theories on kinetic analysis
reaction model (f( )) and the probable variations in the parameters
with the progress of the reaction. Most of the previous work is limited to 3.1. Advanced isoconversional method (for non-isothermal TGA)
the prediction of single value kinetic parameters like activation energy
for the degradation reaction of PET/Post consumer PET. However, a The kinetic analysis was carried out based on Arrhenius law of re-
single activation energy unable to capture the complexity involve in the action kinetic, which was manipulated mathematically to fit the TGA
solid state reactions like PET degradation [15,16]. In the condensed data. The Arrhenius rate equation for non-isothermal process is given as
phase reactions like plastic degradation, the overall rate is a function of
several simultaneously occurring single step reactions with their own d E
= A exp f( )
energy barriers. As a result, the effective activation energy value varies dT RT (1)
with the progress of the reaction or the temperature. In solid-state de-
Where α is the conversion value, = (W0 W)/(W0 W ) . Where, W0 ,
gradation reaction like thermal degradation of PET, the prediction of
W , W are the initial weight, instantaneous weight and residual weight
kinetics is difficult, as the reaction dynamic is dominated by product
of the sample respectively and β is the heating rate.
diffusion rate through the solid matrix. One approach to solve this
Advance isoconversional method (AIC) [7] utilizes the numerical
problem involves the use of isoconversional kinetic analysis methods
integration of experimental data denoted by I (E , T ) = p (x ) , where
like advance isoconversional method (AIC) proposed by Vyazovkin and
p (x ) = 0 e x x 2dx andx = E RT .
his team [7]. Isoconversional methods recommended by the Interna-
The detailed aspects of AIC method has been explained in various
tional Confederation for Thermal Analysis and Calorimetry (ICTAC) are
studies by Vyazovkin et al. [20–24]. The values of activation energy
able to capture the complexity involved in the thermal degradation of
(E ) at various conversion can be determined by minimizing the func-
solid [17–20]. AIC method utilizes advance numerical integration of the
tion (E ),
TGA data and successfully predict the activation energy at various level
of the conversion. In the present study, AIC method was applied to n n
I [E , Ti (t )]
predict the variable activation energy in the degradation reaction of (E ) = = min
I [E , Tj (t )] (2)
polyethylene terephthalate from waste soft drink bottles (PET-SDB) i=1 j i

under inert (N2) environment and oxygenated (Air: 20% O2, 80% N2)
Where the temperature integral:
environment. To give a holistic approach to the degradation kinetics of
the PET-SDB, the current study also includes the determination of ki- t E
netic parameters for the degradation data obtained for isothermal TGA I [E , Ti (t )] = exp dt
t RTi (ti ) (3)
at multiple degradation temperatures. Utilizing thermal degradation
process as an alternative approach in the disposal of post-consumer PET i and j represent the various temperature program (ranging from 1
demands efficient process design and optimization of reaction condi- to n, n being the total number of temperature program) and i j .
tions. In this context, the current study present an elaborate analysis of Eq. (2) was solved numerically and minimization was repeated for
the kinetic parameters of the thermal degradation post-consumer PET, every conversion to obtain a relative dependency between E and .
which will help in developing efficient process design and lead to The possible reaction model was estimated using Criados’ master-
widespread implementation of the thermal processes. plots technique [25]. The details of the method has been reported in our
previous study [26]. The mathematical form of Criados’ masterplots is
2. Materials and methods
2
z( ) f ( )g ( ) T (d / dt )
= =
Samples were obtained from empty soft drink/water bottles. The z (0.5) f (0.5) g (0.5) T0.5 (d /dt )0.5 (4)
collected bottles, identified with their resin identification code
(PET = 1) were cleaned after removing all the foreign materials (labels, The left hand side of Eq. 4 is the theoretical masterplots which
sticker, caps etc.). Clean and dry samples were chipped into small signifies characteristic of each reaction mechanism mentioned in
pieces of 2–3 mm and were mixed uniformly. The proximate analysis of Table 1S. The right hand side of Eq. (4), which is also known as reduced
the sample was carried out considering the standard ASTM methods rate, was calculated from the experimental results. Comparison of the
(D317, D3173, D3175) and the obtained data are summarized in theoretical masterplots with the experimental reduced rate plot sug-
Table 1. The thermal degradation study of the sample was performed gests the most suitable reaction model f(α). For multi-heating rate data,
using TGA (NETZSCH 209 F1 TG) equipment in both inert (N2 gas) and the master plots were obtained for all the heating rates and the most
reactive (Synthetic air: 20% O2 and 80% N2) environments. Non-iso- suitable model has been selected by comparing the linearity coefficient
thermal experiments were conducted at heating rates of 5, 10, 20, 40 (R2) between theoretical master plots and experimental reduced plots.
and 50 °C/min to scan the samples in the temperature range of 30 °C to Finally, the values of pre-exponential factor (A) can be determined
700 °C. Isothermal analysis was carried out at four temperatures (325, by the use of Constable plot [26–28]. The logarithm form of rate ex-
350, 375 and 400 °C) with holding time of 8 h (480 min). The maximum pression can be derived from Eq. (1):

2
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

E d values of degradation temperatures (To and Tm) and residual weight left
ln (Af ( )) = + ln
RT dt (5) after each TGA experiment are given in Table 2. The onset temperature
(To) and peak degradation temperature (Tm) increased with the in-
In a more generalised approach Eq. (4) can be written as crease of the heating rates. The degradation temperatures (To and Tm)
b = ln (Af ( )) aE , where, a = 1/RTα and b = ln((dα/dt)α). The have lower values for oxygenated pyrolysis then in comparison to the
activation energy (E ) and (Af ( )) at various stages of conversion fol- pyrolytic degradation, which implies that the thermal degradation of
lows a logarithmic linear relationship. The values of ln (A f ( )) can be PET-SDB under oxygenated environment occurs at a faster rate than
determined from the intercept of a linear plot of b and aE using iso- that of the degradation under pyrolytic inert condition.
conversion principle. Substituting the known reaction model f ( ) Fig. 2 represents the isothermal degradation (under N2) curves of
predicted by Criados’ masterplots, the values of A (pre-exponential PET-SDB at four temperatures 325, 350, 375 and 400 °C. The weight
factor) were predicted. loss during non-isothermal induction period (100 °C·min−1) before the
temperature reached the final temperature, can be neglected as the
3.2. McCallum method (for isothermal TGA) heating time is very less (high heating rate) compared to the isothermal
holding time. The residual weight left after holding the isothermal
The values of activation energy obtained from non-isothermal TGA temperature for 480 min varied with the increase of isothermal tem-
can be compared with the values of activation energy obtained from perature (Table 2). At low temperature (325 °C), incomplete conversion
isothermal TGA. The calculation method was proposed by McCallum led to the accumulation of almost 63% residual weight (8.65 mg),
[29]. Hill et al. [30] carried out the thermal degradation kinetics of whereas at high isothermal temperature of 400 °C the residual weight
polymethacrylonitrile under several isothermal temperature by uti- reached to 10% of the feed (1.43 mg). Although, the conversion process
lizing the method recommended by McCallum [29]. The Arrhenius rate in the isothermal run at 375 and 400 °C was completed in the later
equation (Eq. 1) can also be expressed as, stages but the residual accumulation were 18% and 10% respectively.
d E More residue yield at lower temperature was probably due to the for-
= A exp f( ) mation of stable ring structure with less weight loss [30].
dt RT (6)
Further integrating Eq. 6 results 4.3. Kinetic analysis
d E E
= A exp dt ( ) = A exp t The kinetic triplets (Eα, Aα, and f(a)) for non-isothermal degradation
f( ) RT RT (7)
were calculated with the help of isoconversional principle. The variable
Which can be written as, activation energy (E ) values were calculated using advance iso-
conversional method (AIC). The variation of activation energy (E ) with
E
ln (t ) = ln [ ( )] ln (A) + respect to conversion are shown in Fig. 3 for the degradation reaction of
RT (8)
PET-SDB under inert (N2) and oxygenated (air) environment. The var-
Where, ( ) represents the integral expression. For multiple isothermal iation of activation energy was shown over the conversion range of 0.05
process the logarithmic time ( ln (t )) is the time taken to reach a par- – 0.8 and 0.05 – 0.95 for N2 and air environments respectively. The
ticular conversion α, was plotted against 1/T for each temperature. The estimated values of activation energies were found higher in case of
slopes of the plots ln (t ) vs. 1 T provided the activation energy at dif- pyrolytic degradation than the oxidative degradation. The calculated
ferent extent of conversion. The McCallum method was used based on values of activation energies were in the range of 203–355 kJ/mol and
the assumption that over a range of isothermal temperatures ( ) has 145–218 kJ/mol under inert and oxygenated environment respectively.
the same value for a given conversion. In both the TGA environment (N2 and Air) the trend of activation en-
ergy started at lower values and increased over the period of de-
4. Results and discussion gradation while maintaining nearly constant profile in the intermediate
stages. However, a rapid growth of activation energy was observed
4.1. Proximate analysis of PET-SDB between the conversion values of 0.7 – 0.8 as the material left in the
crucible converted into stable char residue. A sudden fall of activation
The results of the proximate analysis of PET-SDB samples are shown energy from 218 kJ/mol at = 0.8 to 145 kJ/mol at = 0.91 was ob-
in Table 1. Dry samples have no moisture and negligible ash content. served in case of oxygenated degradation (Air), which implies the rapid
The proximate analysis revealed significant amount of volatile matter combustion of char (residue) at higher temperature.
(97.31%) present in the sample. Under thermal degradation, higher The values of pre-exponential factor or frequency factor always
yield of volatile matter can be expected due to high organic content. compensate the values of activation energy in the balance kinetic
Presence of plasticizer, colour, additives and other inorganic substances equation (Eq. 5) [7]. The logarithmic values of pre-exponential factor
may produce the trace amount of ash content. (A ) should be linear with the activation energy across the entire
conversion range [27]. By replacing the values of (d dt ) from the
4.2. TGA and DTG curves experimental data and activation energy (Eα) obtained by AIC method
in Eq. (5), the values of Af ( ) was determined over the range of con-
TGA weight loss data for PET-SDB samples obtained under inert version considered. The Criados’ masterplot technique was utilized to
(N2) and oxygenated (Air) atmosphere are shown in Fig. 1-I (a) and -I calculate the reaction model f ( ) . The Criados’ masterplots for the
(b) respectively. Complete conversion of the sample was observed degradation data of PET-SDB obtained under the inert atmosphere are
under oxygenated TGA because of the combustion of char at high shown in Fig. 4(a). The reaction models were determined based on the
temperatures, whereas degradation reaction ceased at a residue accu- best linear fit between experimental and theoretical masterplots. The
mulation of 20% when thermal degradation takes place under inert linearity coefficients were estimated for each heating rate and the lin-
environment (Fig. 1-I(a)). Further degradation seems continued with earity coefficient values were plotted (Fig. 4(b)) with respect to each
negligible rate. Shape of the TGA profiles were found consistent with heating rate. Close examination of Fig. 4(b) revealed that the PET-SDB
the increase of the heating rate. The DTG plots are shown in Fig. 1-I(b) degradation under inert condition follows more than one reaction
and -II(b) for both inert and oxygenated atmosphere respectively. The models (f ( )) those are listed in Table 1S. The models corresponded to
method of calculation of To and Tm, was discussed in our previous study equal linearity coefficient (R2) for A2, A3, A4 and F1 for the data ob-
[26]. The adopted temperature programs of the TGA, along with the tained for all the heating rates considered. To identify the most suited

3
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

Fig. 1. (a) TGA and (b) DTG profiles of PET-SDB in presence I. N2 and II. Synthetic air.

Table 2
Conditions of TGA and values of onset (To), peak degradation temperature (To) and residual weight of all the experiments.
Heating condition β °C·min−1 Ffinal temperature Isothermal holding time Initial weight Onset temperature Peak degradation Residual weight
(T∞)°C (th) min (Wo) mg (To) °C temperature (Tm) °C %

Non-isothermal (N2) 5 700 0 8.94 385 427 11.52

10 6.56 398 438 11.89
20 8.68 408 448 14.75
40 7.84 417 465 14.03
50 6.69 427 470 14.2
Non-isothermal (Air) 5 8.98 362 406 1.066
10 7.77 374 423 2.7
20 7.35 389 437 1.11
40 8.82 398 455 0.47
50 7.54 405 458 0.53
Isothermal (N2) 100 325 480 13.88 – – 62.32
350 11.42 – – 30.21
375 11.52 – – 17.62
400 12.65 – – 11.305

Fig. 2. Isothermal weight loss of PET-SDB at various temperatures (under N2 Fig. 3. Distributed activation energy (Eα) obtained by AIC method for PET-SDB
environment). under inert (N2) and oxygenated atmosphere (Air).

reaction model among the four, the values of the pre-exponential fac-
tors were calculated by substituting the E and f ( ) values in Eq. (6) for models. However, oxygenated non-isothermal TGA profiles fallen in
each value of . The variation of pre-exponential factor (A ) values in two-step degradation, which was clearly observed in the DTG plots
the conversion range considered are shown in Fig. 5, corresponding to (Fig. 2(b)). The division of the stages in the oxygenated DTG plot is
all the four models. The pre-exponential values obtained were found shown in Fig. 6 for the heating rate 10 °C/min. The Criados’ masterplots
marginally different for different reaction models. To estimate the were evaluated for subsequent stages over a normalized conversion
correct set of the kinetic triplets, the f ( ) function was confirmed by values for all the heating rates. In the TGA profiles under N2 atmo-
constable plots of each set of values of A in its logarithmic form and sphere, masterplots (Fig. 7) confirmed four possible reaction models
calculated values ofE .The linearity of the constable plots were esti- (A2, A3, A4 and F1) (Fig. 8). All four models produced four different
mated by measuring the root mean square error (RMSE) (Table 3). The sets of pre-exponential factors as shown in Fig. 9. The RMSE values
model F1 (first order given the lowest RMSE values among the four (Table 3) confirmed the lowest RMSE value for F1 model for PET-SDB

4
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

Fig. 7. Criados’ masterplots obtained from degradation data (in air atmosphere)
Fig. 4. (a) Criados’ masterplots for non-isothermal (N2) TGA of PET-SDB (b) of PET-SDB at (a) stage 1 and (b) stage 2.
Linearity coefficient (R2) between theoretical and experimental masterplots at
different heating rates.

Fig. 5. Variation of pre-exponential factor (A) with conversion considering four

reaction models for waste PET-SDB degradation under N2 atmosphere.

Table 3
Root mean square error (RMSE) values between linear constable plot and cal-
culated Eα vs. ln(Aα) obtained for the four reaction models.
Reaction model RMSE

TGA (inert) TGA (synthetic air) Fig. 8. Comparison of theoretical masterplots with experimental masterplots by
linearity coefficient comparison at different heating rate for (a) stage 1 and (b)
A2 0.458 1.18
stage 2.
A3 0.5 1.221
A4 0.523 1.243
F1 0.385 1.103

Fig. 9. Variation of pre-exponential factor (Aα) with conversion considering

four reaction models for non-isothermal degradation of PET-SDB under air at-
Fig. 6. DTG (from TGA) profile of PET-SDB under air atmosphere at 10 °C/min. mosphere.

5
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

Table 4
Kinetic data comparison with the between the earlier studies and the present work.
Reference TGA Condition Method Eα (kJ/mol) Log (Aα) (s−1) f(α)

Cooney, Day and Wiles β = 0.1 – 100 °C/min; Air Kissinger’s Method Stage 1 = 162.9 ± 72.4 – F1 (first order)
[31] Stage 2 = 202.1 ± 6.7
Stage 3 = 147.3 ± 3.5
Freeman-Carroll Method 160.2 – 411.5 (stage 2 only) –
Chatterjee-Conrad Method Stage 1 = 53.5 – 169.3 –
Stage 2 = 147.0 – 301.4
Stage 3 = 200.6 – 292.2
Friedman Method 69.6 – 210.8
Ozawa Method 88.3 – 187.7 8.1 – 16.2
Jenekhe, Lin and Sun [33] β = 0.31 – 10 °C/min; N2 Arrhenius plot 198.61 – 232.46 12.49 – 14.89 (1 - α), first order
Isoconversional method 173.63 – 205.85 – kinetics
Martı́n-Gullón, Esperanza T = 27 – 800 °C, β = 5, 10, Simplex method of Nedler Part 1 (80% weight Part 1 (80% weight For nth order, [f(α) = (1
and Font [9] 20 °C /min, under He and O2 and Mead loss) = 256.4 (100% He) loss) = 16.8 (100% He) – α)n]
(1:0, 3:1, 4:1 and 9:1) 186.0 (33% O2 and 67% He) 11.6 (33% O2 and 67% n = 1.05 for part 1 (80%
He) weight loss and 100%
He)
194.0 (20% O2 and 80% He) 12.2 (20% O2 and 80% n = 1.02 (33% O2 and
He) 67% He)
203.5 (10% O2 and 90% He) 12.9 (10% O2 and 90% n = 1.00 (20% O2 and
He) 80% He)
n = 1.00 (10% O2 and
90% He)
Saha, Maiti and Ghoshal T = 30 – 600 °C; β = 10, 15 and ASTM E698 162.15 – 210.64 26.37 – 34.81 n = 1;
[14] 25 °C /min nth Order model 322.3 – 338.98 54.76 – 57.73 n = 1.72 – 1.82;
Current study T = 30 – 600 °C; β = 5, 10, 20, Advance isoconversional 203 – 355 (N2) 33.3 – 59.3 (N2) F1 (N2)
40 and 50 °C /min under N2 and (AIC) 145 – 218 (Air) 28.5 – 51.8 (Air) F1 (Air)
Air

degradation under air. Therefore, it was observed that the thermal

degradation of PET-SDB under N2 and Air followed the same F1 (first
order) reaction model. A comparison of kinetic data of the present work
and past studies dealing with the degradation kinetics of PET at pure
form [31,32] or from the waste PET bottles [13] is summarized in
Table 4. It was observed that the choice of kinetic model can affect the
obtained kinetic results. Some of the values of the kinetic parameters of
PET degradation reaction obtained in the past were close to the values
obtained by AIC method in the current study.
A variation of activation energy was also determined from the iso-
thermal degradation data of PET-SDB in TGA under N2 atmosphere. The
weight loss curves with respect to time are shown in Fig. 2. Applying Fig. 11. Distribution of activation energy with the degree of conversion ob-
the degradation data to Eq. (8), the ln(t) vs 1/T was plotted and shown tained from isothermal degradation profiles of PET-SDB (under N2 atmosphere).
in Fig. 10. The slope of the regression provided the global activation
energy at each degree of conversion. The variable activation energy
5. Conclusion
with conversion was shown in Fig. 11. The activation energy was found
in the range of 98–183 kJ/mol between conversions of 0.05 to 0.6. The
Polyethylene terephthalate can be found abundantly in any waste
range of activation energy for the degradation of PET-SDB under non-
stream as empty soft drink bottles, containers, waste films and fibers
isothermal TGA in N2 atmosphere was found in the range of
etc. The degradation processes of these type of plastics by thermal
196–217 kJ/mol between the conversion range of 0.05 – 0.6.
means are complex and get affected by the presence of oxygen in the
environment. Kinetic parameters obtained in the inert and oxygenated
environment have shown different values. The activation energy values
obtained by advance isoconversional methods (AIC) were in the range
of 203–355 kJ/mol and 145–218 kJ/mol for N2 and air environment
respectively. The thermal degradation reaction of PET-SDB followed F1
(first order) model for both inert and oxygenated environment.
Activation energy values calculated using multiple isothermal TGA
were estimated in the range of 196–217 kJ/mol in the conversion range
of 0.05 – 0.6. The establishment of kinetic triplets of PET-SDB de-
gradation considering AIC method is a significant for the implementa-
tion of thermal processes like pyrolysis and combustion.

Acknowledgement

Authors would like to acknowledge the financial support (Grant#

Fig. 10. Plot of In (t) versus 1/T for isothermal TGA of PET-SDB. YSS/2014/000837) provided by Department of Science and Technology

6
P. Das and P. Tiwari Thermochimica Acta 679 (2019) 178340

(DST), Government of India. [15] S. Vyazovkin, Reply to “what is meant by the term ‘variable activation energy when
applied in the kinetics analyses of solid state decompositions (crystolysis reac-
tions)?”, Thermochim. Acta 397 (2003) 269–274.
Appendix A. Supplementary data [16] A.R. McLauchlin, O.R. Ghita, Studies on the thermal and mechanical behavior of
PLA-PET blends, J. Appl. Polym. Sci. 133 (2016).
Supplementary material related to this article can be found, in the [17] M.E. Brown, M. Maciejewski, S. Vyazovkin, R. Nomen, J. Sempere, A.K. Burnham,
J. Opfermann, R. Strey, H.L. Anderson, A. Kemmler, R. Keuleers, J. Janssens,
online version, at doi:https://doi.org/10.1016/j.tca.2019.178340. H.O. Desseyn, C. Li, T.B. Tang, B. Roduit, J. Malek, T. Mitsuhashi, Computational
aspects of kinetic analysis Part A: the ICTAC kinetics project-data, methods and
References results, Thermochim. Acta 355 (2000) 125–143.
[18] A.K. Burnham, Computational aspects of kinetic analysis. Part D: the ICTAC kinetics
project-multi-thermal-history model-fitting methods and their relation to iso-
[1] J. Scheirs, W. Kaminsky, Feedstock Recycling and Pyrolysis of Waste Plastics, John conversional methods, Thermochim. Acta 355 (2000) 165–170.
Wiley & Sons Chichester, UK, 2006, pp. 641–660. [19] M. Maciejewski, Computational aspects of kinetic analysis. Part B: the ICTAC
[2] K.R. Arturi, H.U. Sokoli, E.G. Søgaard, F. Vogel, S. Bjelić, Recovery of value-added Kinetics Project – the decomposition kinetics of calcium carbonate revisited, or
chemicals by solvolysis of unsaturated polyester resin, J. Clean. Prod. 170 (2018) some tips on survival in the kinetic minefield, Thermochim. Acta 355 (2000)
131–136. 145–154.
[3] F. Awaja, D. Pavel, Recycling of PET, Eur. Polym. J. 41 (2005) 1453–1477. [20] S. Vyazovkin, Computational aspects of kinetic analysis. Part C. The ICTAC Kinetics
[4] S.S. Idris, N.A. Rahman, K. Ismail, Combustion characteristics of Malaysian oil palm Project – the light at the end of the tunnel? Thermochim. Acta 355 (2000) 155–163.
biomass, sub-bituminous coal and their respective blends via thermogravimetric [21] S. Vyazovkin, A.K. Burnham, J.M. Criado, L.A. Pérez-Maqueda, C. Popescu,
analysis (TGA), Bioresour. Technol. 123 (2012) 581–591. N. Sbirrazzuoli, ICTAC Kinetics Committee recommendations for performing kinetic
[5] F. Welle, Twenty years of PET bottle to bottle recycling—an overview, Resour. computations on thermal analysis data, Thermochim. Acta 520 (2011) 1–19.
Conserv. Recycl. 55 (2011) 865–875. [22] S. Vyazovkin, N. Sbirrazzuoli, Isoconversional kinetic analysis of thermally stimu-
[6] M.N. Siddiqui, H.H. Redhwi, D.S. Achilias, Recycling of poly(ethylene ter- lated processes in polymers, Macromol. Rapid Commun. 27 (2006) 1515–1532.
ephthalate) waste through methanolic pyrolysis in a microwave reactor, J. Anal. [23] S. Vyazovkin, C.A. Wight, Estimating realistic confidence intervals for the activation
Appl. Pyrolysis 98 (2012) 214–220. energy determined from thermoanalytical measurements, Anal. Chem. 72 (2000)
[7] M. Barth, A. Honak, T. Oeser, R. Wei, M.R. Belisário-Ferrari, J. Then, J. Schmidt, 3171–3175.
W. Zimmermann, A dual enzyme system composed of a polyester hydrolase and a [24] S. Vyazovkin, Isoconversional kinetics of polymers: the decade past, Macromol.
carboxylesterase enhances the biocatalytic degradation of polyethylene ter- Rapid Commun. 38 (2017) 1600615.
ephthalate films, Biotechnol. J. 11 (2016) 1082–1087. [25] J.M. Criado, Kinetic analysis of DTG data from master curve, Thermochim. Acta 24
[8] M.E. Bednas, M. Day, K. Ho, R. Sander, D.M. Wiles, Combustion and pyrolysis of (1978) 186–189.
poly(ethylene terephthalate). I. The role of flame retardants on products of pyr- [26] P. Das, P. Tiwari, Thermal degradation kinetics of plastics and model selection,
olysis, J. Appl. Polym. Sci. 26 (1981) 277–289. Thermochim. Acta 654 (2017) 191–202.
[9] I. Martı́n-Gullón, M. Esperanza, R. Font, Kinetic model for the pyrolysis and com- [27] F.H. Constable, The mechanism of catalytic decomposition, Proc. R. Soc. Lond. 108
bustion of poly-(ethylene terephthalate) (PET), J. Anal. Appl. Pyrolysis 58–59 (1923) 355–378.
(2001) 635–650. [28] P. Tiwari, M. Deo, Detailed kinetic analysis of oil shale pyrolysis TGA data, AIChE J.
[10] N. Didane, S. Giraud, E. Devaux, G. Lemort, G. Capon, Thermal and fire resistance of 58 (2012) 505–515.
fibrous materials made by PET containing flame retardant agents, Polym. Degrad. [29] J.R. MacCallum, Comprehensive Polymer Science, Pergamon, Oxford, 1989.
Stab. 97 (2012) 2545–2551. [30] D.J.T. Hill, L. Dong, J.H. O’Donnell, G. George, P. Pomery, Thermal degradation of
[11] X.-H. Du, C.-S. Zhao, Y.-Z. Wang, Q. Zhou, Y. Deng, M.-H. Qu, B. Yang, Thermal polymethacrylonitrile, Polym. Degrad. Stab. 40 (1993) 143–150.
oxidative degradation behaviours of flame-retardant thermotropic liquid crystal [31] J.D. Cooney, M. Day, D.M. Wiles, Thermal degradation of poly(ethylene ter-
copolyester/PET blends, Mater. Chem. Phys. 98 (2006) 172–177. ephthalate): a kinetic analysis of thermogravimetric data, J. Appl. Polym. Sci. 28
[12] L. Buxbaum, The degradation of poly (ethylene terephthalate), Angew. Chem. Int. (1983) 2887–2902.
Ed. 7 (1968) 182–190. [32] B.J. Holland, J.N. Hay, The thermal degradation of PET and analogous polyesters
[13] B. Saha, A. Ghoshal, Thermal degradation kinetics of poly (ethylene terephthalate) measured by thermal analysis–Fourier transform infrared spectroscopy, Polymer 43
from waste soft drinks bottles, Chem. Eng. J. 111 (2005) 39–43. (2002) 1835–1847.
[14] B. Saha, A.K. Maiti, A.K. Ghoshal, Model-free method for isothermal and non-iso- [33] S.A. Jenekhe, J.W. Lin, B. Sun, Kinetics of the thermal degradation of polyethylene
thermal decomposition kinetics analysis of PET sample, Thermochim. Acta 444 terephthalate, Thermochim. Acta 61 (1983) 287–299.
(2006) 46–52.