Kinetics Study of Polypropylene Pyrolysis by Non-Isothermal Thermogravimetric Analysis

Polypropylene (PP) is considered as one of six polymers representative of plastic wastes. This paper attempts to obtain information on PP polymer pyrolysis kinetics with the help of thermogravimetric analysis (TGA). TGA is used to measure the weight of the sample with temperature increases at different heating rates—5, 10, 20, 30, and 40 K min−1—in inert nitrogen. The pyrolytic kinetics have been analyzed by four model-free methods—Friedman (FR), Flynn–Wall–Qzawa (FWO), Kissinger–Akahira–Sunose (KAS) and Starnik (STK)—and by two model-fitting methods—Coats–Redfern (CR) and Criado methods. The values of activation energies of PP polymer pyrolysis at different conversions are in good agreement with the average of (141, 112, 106, 108 kJ mol−1) for FR, FWO, KAS and STK, respectively. Criado methods have been implemented with the CR method to obtain the reaction mechanism model. As per Criado’s method, the most controlling reaction mechanism has been identified as the geometrical contraction models—cylinder model.


Introduction
Many researchers are focused on finding suitable ways to deal with huge plastic wastes in order to recycle such vast wastes. In our previous publication (Dubdub and Al-Yaari (2020, 2021) [1,2]), the problem of growing plastic wastes throughout the world has been outlined, as well as how it can be sorted by a "primary", "secondary", or "tertiary" recycling method with their limitations. The last option, "tertiary", usually needs some advanced investigations, and one of the available methods is the "pyrolysis" method. This method has some advantages compared with the rest of the methods such as the combustion by the reduction in the volume of gaseous products (Kaminsky et al. (1996) [3]).
In this research direction, there is a limited amount of reported literature on PP as one of the major plastic waste constituents and their pyrolysis kinetics. Wu et al. (1993) [4] applied the pyrolysis for six polymers mixtures-high-density polyethylene (HDPE), low-density polyethylene (LDPE), PP, polystyrene (PS), Polyvinyl chloride (PVC), and acrylonitrile-butadiene-styrene (ABS) of municipal solid waste (MSW)-by TGA at three sets of heating rates (1, 2, and 5.5 K min −1 ). They found that the interaction between these polymers of MSW in the pyrolysis was insignificant. Diaz Silvarrey and Phan (2016) [5] studied the kinetic cracking of different polymers-PP, PS, HDPE, LDPE, and Polyethylene terephthalate (PET)-for TGA. They used Malek, KAS, and linear model fitting methods to obtain the pyrolysis mechanism. They ran the TGA with four heating rates (5, 10, 20, and 40 K min −1 ) and a temperature range of 30-700 • C. Their results showed that all the polymers followed the same one-step decomposition mechanism with the increase in temperature, and their order of decomposition was: PS < PET < PP < LDPE < HDPE. They also reported that the values of the activation energy and pre-exponential factor by the KAS method for PP are 261.22 KJ mol −1 and 3.03 × 10 21 S −1 . Chowlu et al. (2009) [6] studied the pyrolysis behavior for the mixture of two polymers, PP and LDPE, with five different mixture compositions and heating rates. They used the Vyazovkin (VYK) method (modelfree technique) to evaluate the change in activation energy with the conversion. They found Materials 2023, 16, 584 2 of 12 that this change in the reaction occurs in three different zones: slow at low conversion, slightly high at the middle conversion, and strongly high until the end of the decomposition. They concluded that the best mixture of the PP/LDPE weight ratio is 65%/35% since it has a low activation energy. Aboulkas et al. (2010) [7] calculated the activation energy and the reaction model of the pyrolytic reaction of HDPE, LDPE, and PP from non-isothermal TGA data. They obtained the values of activation energy by FR, KAS, and FWO between 179 and 188 kJ mol −1 for PP. They found the "geometrical contraction models-cylinder" to be the conversion model using the CR and Criado methods. Yu et al. (2016) [8] reviewed and compiled the various reported results on the pyrolysis of PVC mixed with PP/PE/PS. It has been reported that the interaction between the polymers depends mainly on the nature of the polymer itself. They also studied the effect of this different mixture of PVC and one of these polymers (PP/PE/PS) with the respect to: the onset temperature, maximum peak decomposition, residue weight, and quantity. Anene et al. (2018) [9] researched the cracking of LDPE/PP mixtures with different compositions in order to investigate the effect of the addition of PP to LDPE. They found that the cracking starts with a lower temperature for LDPE/PP compared to LDPE alone, informing the interaction between the two polymers. Finally, Mumbach et al. (2019) [10] performed the decomposition of plastic solid waste (PSW) by TG under inert conditions from 25 to 1000 • C with four heating rates (5, 10, 20, and 30 K min −1 ). They estimated that the feedstock of PSW will include: 17.28% LDPE, 7.41% HDPE, 51.85% PP, 17.28% plastics for PET, PS, and PVC, and 6.18% lignocellulosic materials. They evaluated the activation energy using KAS, STK, FWO, VYA, and STK (isoconversional methods) and the frequency factor by the compensation effect factor, while they evaluated the reaction model by the master plot. They identified that three main stages of reaction occur: the first decomposition occurs by the main decomposition of holocellulose and the minor decomposition of the first degradation stage of PVC (dichlorination); the second decomposition occurs mainly by a decomposition of a mixture of polymers, such as PS and some adhesive acrylic-based resins and PVC (dichlorination); and the third decomposition occurs by PP, HDPE, and LDPE as the major fraction and the second PVC thermal decomposition as the minor fraction. Galiwango and Gabbar (2022) [11] studied the co-pyrolysis of the PP polymer and paper wastes. They found that the activation energy values for the PP-paper mixture are much lower than those of pure PP, and this is because of the synergistic interactions between PP and paper wastes.
The objective of this study is to obtain the kinetic parameters of PP polymer pyrolysis using TGA data. Since the kinetic parameters are required for any further study, such as setting any type of reactor, the activation energy has been calculated using six methods (four model-free methods and two model-fitting methods). Dubdub and Al-Yaari (2020, 2021) [1,2] started a comprehensive study of the calculation of kinetics parameters for the pyrolysis of different polymers using TGA.

Materials and Thermogravimetry of PP
PP supplied from Ipoh SY Recycle Plastic/Malaysia has been used. PP pellets were ground into powder before putting them in the thermogravimetric analyzer. The pyrolysis of PP was performed by a 1020 series thermogravimetric TGA-7, manufactured by PerkinElmer Co., Waltham, MA, USA. A total of 10 mg of PP powder samples was used throughout all the TGA experiments in an inert atmosphere of pure N 2 (99.999%) gas flowing at 100 cm 3 /min at five different heating rates (5,10,20,30, and 40 K min −1 ). Each experiment was run more than once to ensure the reproducibility of the collected data. The international confederation for thermal analysis and calorimetry (ICTAC) recommends a ratio between the lowest and highest non-isothermal heating rates within 10-15 (Osman et al. (2022) [12]). In this work, a ratio of 8 (lowest = 5, highest = 40) is close to the minimum limit of 10. Both of the analysis results (ultimate and proximate) are presented in Table 1. The 1020 series Thermogravimetric Analysis TGA7 has been employed to achieve the objective of this work. It has the following components: the controller with thermal analysis software and the thermogravimetric analyzer.

Kinetic Theory
There are two ways of running a TGA instrument: isothermal and non-isothermal. In this work, non-isothermal TGA is used over isothermal TGA because the non-isothermal TGA needs less time, and the isothermal TGA is not convenient at higher temperatures for large non-isothermal heat-up and cool-down times (Nawaz and Kumar (2022) [13]). During the pyrolysis, polymer chains break, producing volatile products and releasing them from the original polymer, which causes a loss of mass. Therefore, TGA is suitable for obtaining the kinetic parameters of the polymer (Hayoune et al. (2022) [14]). The estimation of kinetic variable with the data collected from the TGA can be performed by different methods. Those methods include one single TGA data (called the "model-fitting" method) or multiple TGA data at different heating rates (called the "model free" or "isoconversional" methods) and calculating the kinetic equation with differential methods or integrals.
The reaction kinetics of the PP pyrolysis can be started with the following equation: where α, t, E a , R, A 0 , and T stand for the reaction conversion, the time, the activation energy, the universal gas constant, the frequency factor, and the absolute temperature, respectively. For non-isothermal pyrolysis, β (heating rate) can be inserted into Equation (1): The derivation of the model-free methods (FR, FWO, KAS, and STK) and model-fitting methods, starting from Equation (2) with all the assumptions until the final equation, can be found in Aboulkas et al. (2010) [7]. Tables 2 and 3 present the kinetic equations of the four commonly used model-free methods and the two model-fitting methods.
The Criado equation connects the reduced theoretical curve representing the characteristic of each reaction mechanism (left side) and the experimental data (right side). Therefore, a comparison between these two sides will inform which exact kinetic model will describe the experimental reaction. Table 4 shows the common solid-state thermal reaction mechanisms-f (α) and g(α)-used in the CR method and the Criado method. Table 2. Equations for four model-free methods.

Thermogravimetry of PP
The TG and the DTG with the set of heating rates of PP polymer pyrolysis are shown in Figures 1 and 2, respectively. The thermograms were identical, and the thermal decomposition characteristics (onset, peak, and final temperatures) shown in Table 5 were shifted to a higher temperature with a higher heating rate; this shifting was also reported by different researchers (Wu et al. (1993)

Thermogravimetry of PP
The TG and the DTG with the set of heating rates of PP polymer pyrolysis are shown in Figures 1 and 2, respectively. The thermograms were identical, and the thermal decomposition characteristics (onset, peak, and final temperatures) shown in Table 5 were shifted to a higher temperature with a higher heating rate; this shifting was also reported by different researchers (Wu et al. (1993)      There are two ways to calculate the kinetics parameters from TGA data: either a single thermogram or multiple thermograms at a set of heating rates, differentially or integrally (Chan and Balke (1997) [17]. In this paper, six methods have been implemented and compared between them. One set includes four model-free methods (FR, FWO, KAS, and STK), while the second set, model-fitting, implies one single thermogram (CR and Criado equation). The Criado equation method is considered the "masterplot" since it determines the kinetic model of the process.

Model-Free Kinetics Calculation
Model-free methods are preferable to model-fitting methods based on precision, proficiency, and reliability. Model-free methods need at least three heating rate runs in order to reckon the kinetic parameters without any assumptions for the reaction mechanism (Nawaz and Kumar (2022) [11]). In this section, four types of isoconversional methods have been performed to calculate the activation energy Ea in kJ mol −1 . The fitted linear equations of FR, FWO, KAS, and STK at conversions ranging from 0.1 to 0.9 for the five sets are shown in Figure 3, and the summaries of the calculated values of the activation energies are presented in Figure 4 and Table 6, respectively. Generally, model-free methods are more confident than model-fitting methods because their activation energy values are consistent with the non-isothermal data. This also helps to highlight the complexity of multi-reaction since the activation energy calculation relates to the conversion (Vyazovkin and Wight (1999) [18]). Table 6 [20] supported this finding by obtaining the activation energy of an individual corn cob and PE pyrolysis by FR, which was higher than that of KAS, FWO, and STK. In addition, Figure 4 shows that all the values of the activation energies calculated by the FR method (differential method) are slightly higher than those of the rest of the methods (FWO, KAS, and STK) (integral method), and this finding is in agreement with Aboulkas's finding.   There are two ways to calculate the kinetics parameters from TGA data: either a single thermogram or multiple thermograms at a set of heating rates, differentially or integrally (Chan and Balke (1997) [17]. In this paper, six methods have been implemented and compared between them. One set includes four model-free methods (FR, FWO, KAS, and STK), while the second set, model-fitting, implies one single thermogram (CR and Criado equation). The Criado equation method is considered the "masterplot" since it determines the kinetic model of the process.

Model-Free Kinetics Calculation
Model-free methods are preferable to model-fitting methods based on precision, proficiency, and reliability. Model-free methods need at least three heating rate runs in order to reckon the kinetic parameters without any assumptions for the reaction mechanism (Nawaz and Kumar (2022) [11]). In this section, four types of isoconversional methods have been performed to calculate the activation energy E a in kJ mol −1 . The fitted linear equations of FR, FWO, KAS, and STK at conversions ranging from 0.1 to 0.9 for the five sets are shown in Figure 3, and the summaries of the calculated values of the activation energies are presented in Figure 4 and Table 6, respectively. Generally, model-free methods are more confident than model-fitting methods because their activation energy values are consistent with the non-isothermal data. This also helps to highlight the complexity of multi-reaction since the activation energy calculation relates to the conversion (Vyazovkin and Wight (1999) [18]). Table 6 shows the E a values for the four methods (FR, FWO, KAS, and STK) for the (0.1-0.9) conversion ranges. Due to some approximations and mathematical simplification, the results slightly differ from each other with this range of conversions. The average  [20] supported this finding by obtaining the activation energy of an individual corn cob and PE pyrolysis by FR, which was higher than that of KAS, FWO, and STK. In addition, Figure 4 shows that all the values of the activation energies calculated by the FR method (differential method) are slightly higher than those of the rest of the methods (FWO, KAS, and STK) (integral method), and this finding is in agreement with Aboulkas's finding. Aboulkas et al. (2010) [7] reported that the FR method is more sensitive than the rest at low and high conversions. As expected, FR, which is the only differential method using the point value of the overall reaction rate, will produce an activation energy value that is different from the integral method using the history  [22] attributed this difference to the variation in the non-linearity. Janković and Manić (2021) [23] attributed the difference between them to the consequence in the calculation formula, the equations origin, and the final results. Table 7 shows the activation energy values for PP from different published papers. It can be concluded that the activation energy values of this paper are between the highest value (225 kJ mol −1 ) by FWO and the lowest value (67 kJ mol − ) by FR (Paik and Kar (2008) [24]).  [7] reported that the FR method is more sensitive than the rest at low and high conversions. As expected, FR, which is the only differential method using the point value of the overall reaction rate, will produce an activation energy value that is different from the integral method using the history  [22] attributed this difference to the variation in the nonlinearity. Janković and Manić (2021) [23] attributed the difference between them to the consequence in the calculation formula, the equations origin, and the final results.

Model-Fitting Kinetics Parameters Calculation
The first CR has been used by plotting ln [  Table 8. A big difference can be seen in the

Model-Fitting Kinetics Parameters Calculation
The first CR has been used by plotting ln g(α) T 2 versus 1/T with a set of formulas of functions g(α) to represent the solid-state reaction. This plot produces only one reaction straight line, and the values of the activation energy obtained by the CR method with different functions g(α) (F1-P4) are shown in Table 8. A big difference can be seen in the value of E a depending on the model reaction mechanism (F1-P4). This abnormal value (between the lowest 14 kJ mol −1 and highest 380 kJ min −1 ) can be attributed to the fact that the mechanisms of these values do not meet the pyrolysis of the PP polymer. The corrected value for this activation will be decided according to the comparison between the right side and the left side of the Criado equation with the proper model mechanism reaction, as shown in Figure 5 and Table 9. The values of E a for the data A2, A3, A4, P2, P3, and P4 are not taken into consideration because their values are far from the published E a values. The graphs of D1, A1, R1, P2, P3, and P4 are deleted from Figure 5 because their curves are not within the experimental curve range. Table 9 lists the values of E a , ln (A 0 ), and R 2 for each test (five tests). It shows the most controlling model reaction mechanism: the geometrical contraction models-contracting cylinder (R3). It has been noticed that some researchers use either CR only by assuming some mechanism model reaction, such as the Reaction Order Models-First-(F1), Second-(F2), and Third-order (F3) as a first trend or such as the current research, where it will be used with the Criado method with the free-model methods to find the appropriate mechanism model as a second trend.

Conclusions
The TG and DTG thermograms obtained from the TGA study showed the similar shapes and trends at different polymers' compositions. The TGA data confirmed the existence of only one main reaction region. In this paper, two approaches have been achieved for obtaining the TGA kinetics data. In the first approach, four (FR, FWO, KAS, and STK) model-free methods are achieved, and in the second approach, two model-fitting methods, to converge the TGA data with the straight-line, have been used to calculate the activation energy.
In the first one, the average values of the activation energies of PP polymer pyrolysis at different conversions are in good agreement with the average (141, 112, 106, 108 kJ mol −1 )   In the first trend, Ali et al. (2020) [26] used the CR method with the zero-to four-order reaction order models to determine the proper order of the reaction by selecting the most suitable plot. Baloch et al. (2019) [27] considered different orders of reactions: 0, 0.5, 0.741, 1, 1.5, and 2, and the data fitted much better for 0.741. Balsora et al. (2022) [28] used the analytical method to optimize the best value of the reaction order. Bu et al. (2019) [29] presented the kinetic parameters for orders 1, 2, and 3. Lai et al. (2021) [30] and Li et al. (2022) [31] used only the first order of CR because it is considered as the main mechanism. Patidar et al. (2022) [32] revealed that the diffusion model was best suited to reflect the degradation process using the CR method.
In  [10] chose F1 control for the third or last stage, which is mainly considered as the thermal cracking of LDPE, HDPE, and PP as the major fraction and the second PVC thermal decomposition as the minor fraction. They mentioned that the last-stage reaction includes a major decomposition of PP followed by LDPE and HDPE, with a low fraction of PVC dehychlorination. The E values obtained for this stage by the FWO, KAS, STK, and VYA methods were, respectively, 266.82, 268.39, 269.09, and 268.63 kJ mol −1 . Khodaparasti et al. (2022) [33] selected the F3, F2, and F3 models for microalgae Chlorella vulgaris and municipal sewage sludge and co-pyrolysis, respectively, as the best models for predicting the solid-state mechanism. Singh et al. (2021) [20] found the mechanism reaction of the third-order reaction (F3), diffusion Jander (D3), and Ginstling-Brounshtein (D4) models to be best suited for CC pyrolysis, PE pyrolysis, and co-pyrolysis, respectively. Wan and Huang (2021) [34] used the master plot to obtain the first-order reaction model (D1), which was the most suitable mechanism function for describing the pyrolysis of nylon-6 waste.

Conclusions
The TG and DTG thermograms obtained from the TGA study showed the similar shapes and trends at different polymers' compositions. The TGA data confirmed the existence of only one main reaction region. In this paper, two approaches have been achieved for obtaining the TGA kinetics data. In the first approach, four (FR, FWO, KAS, and STK) model-free methods are achieved, and in the second approach, two model-fitting methods, to converge the TGA data with the straight-line, have been used to calculate the activation energy.
In the first one, the average values of the activation energies of PP polymer pyrolysis at different conversions are in good agreement with the average (141, 112, 106, 108 kJ mol −1 ) for FR, FWO, KAS, and STK, respectively.
Model-fitting (CR and Criado) methods have been applied together to obtain the activation energy and the model reaction mechanism. The values of Ea range from 104 to 289 kJ mol −1 , with the geometrical contraction models-contracting cylinder (R3) as the best model reaction mechanism. However, these values are still close to the published results.