Thermal degradation kinetics investigation on Nano-ZnO/IFR synergetic flame retarded polypropylene/ethylene-propylene-diene monomer composites processed via different fields

2.1 Materials and devices

The polypropylene (PP-T30S) was obtained from Sinopec Maoming. Co., Ltd. China. The ethylene-propylene-diene monomer (EPDM-2504N) was obtained from Exxon Mobil Co., Ltd. USA. Both the poly ammonium phosphate (APP-101) and the dipentaerythritol (DPER-HY) were obtained from Yiju Chemical Co., Ltd. Guangzhou, China. The melamine (MEL-AR) was supplied by Kemiou Chemical Reagent Co., Ltd. Tianjin, China. Furthermore, The ZnO (ZnO-DZZn-01) with the particle size of 20 nm was supplied by Dezhao Chemical Co., Ltd. Guangzhou, China. The silane coupling agent (SCA-A172) was supplied by Shouzheng Chemical Technology Co., Ltd. Guangzhou, China. The polyethylene wax (H110) was supplied by S.Q.I Co., Ltd. Thailand.

2.2 Preparation of flame retardant PP nanocomposites

Firstly, the SCA was weighed and dissolved in an appropriate amount of methanol. The solution of SCA was sprayed onto the surface of IFR and Nano-ZnO, and the mixture was stirred in a high-speed mixer for 30 min. Then the mixture was placed in a drying oven for 2 h at 100°C. Afterwards, PP, EPDM, IFR, Nano-ZnO, and PE wax in different ratios were extruded using the vane extruder (VE) and three-screw extruder (TSE) separately. To ensure the credibility of the contrast, the rotational speeds of VE and TSE were respectively 7 r/ min and 25.5 r/ min, so that the yields were the same. The processing temperatures from feeding section to the extrusion section were shown as follows. VE: 160°C, 175°C, 185°C, 185°C; TSE: 160°C, 175°C, 185°C, 185°C, 185°C, 185°C, 185°C, 185°C, 185°C, 185°C. Formulations of the specimens are presented in Table 1. The IFR contained APP, DPER, and MEL was prepared by us. After extruding, the polymer composite samples were pressed under 15 MPa for 20 min at 185°C

into laminates with suitable thickness and size, and they were kept under the condition of 50% relative humidity and 25°C for at least 48 h before testing.

2.3 Thermogravimetric test

All the samples were examined under nitrogen flow in a thermogravimetric analysis apparatus (model 209, Netzsch, Germany), which was used with crucible sample holders, at the heating rate of 5°C/min, 10°C/min, 15°C/min, 20°C/min, separately. Whereas the final temperature was set to be 600°C.

All thermal degradation kinetics methods were based on non-isothermal kinetic theory combined with Arrhenius equation (16), the thermal degradation kinetics was expressed as:

Based on Eq. 1 and the thermogravimetric curves, the activation energy E, pre-exponential factor A of the thermal decomposition kinetics were calculated by Kissinger method and Flynn-Wall-Ozawa method (17).

3.1 Thermogravimetic analysis

Figure 1 and Figure 2 present the TG curves of the flame retardant PP nanocomposites processed via VE and TSE at different heating rates. Combining with data in Table 2, it was found that with the increasing content of Nano-ZnO, the decomposition temperature at 0.05% weightlessness first rose then fell at a heating rate of 5°C/min; The decomposition temperature at 0.05% weightlessness kept falling at the heating rates of 10°C/min and 15°C/min; While the decomposition temperature at 0.05% weightlessness exhibited a disorder state at a heating rate of 20°C/min. Among them, the decomposition temperature at 0.05% weightlessness was basically in inversely proportional to the heating rate. Took data from samples that processed by VE for instance, the maximum and minimum temperatures were separately 248.5°C and 155.7°C at the heating rate of 5°C/min, and the maximum and minimum temperatures were separately 140.3°C and 32.3°C at the heating rate of 20°C/min. However, this principle was not applicable for samples which processed by TSE. All the decomposition temperatures at 0.05% weightlessness appeared at a rather low temperature with an average number of 75.9°C except for the five temperatures with a heating rate of 5°C/min. All of them maintained at rather high values which around 240°C, and it seemed that they were not even influenced by the content variation of IFR and Nano-ZnO.

Figure 1

TG curves of flame retardant PP nanocomposites processed via VE at different heating rates.

Figure 2

TG curves of flame retardant PP nanocomposites processed via TSE at different heating rates.

Besides, the decomposition temperature at 0.1% weightlessness for VE presented the same principle, which first rose then fell at the heating rates of 5°C/min and 10°C/min. Afterwards, it went into a mess with heating rates of 15°C/min and 20°C/min. For TSE, the same situation occurred as well as the decomposition temperature variation at 0.05% weightlessness. The weight loss between 99.9% and 99.95% was so tiny that the tendency wouldn’t change.

In addition, referring to VE, the decomposition temperature at 0.5%, 0.7%, and 5% weightlessness with a heating rate of 5°C/min showed the same tendency, though the temperature gap became smaller and smaller with the increasing content of Nano-ZnO. Then it obeyed a principle of temperature falling continuously with the heating rates of 10°C/min and 15°C/min. The temperatures exhibited irregular mess with the heating rate of 20°C/min. On the other hand, referring to TSE, the decomposition temperature at 0.5%, 0.7%, and 5% weightlessness showed a weak tendency of first rising then falling with a heating rate of 5°C/min. The former two values were very close which owing to the small decomposition amplitude. The tendency altered to first falling then rising with the heating rates of 10°C/min and 15°C/min. Again, the

temperatures exhibited irregular mess with the heating rate of 20°C/min, which was the same phenomenon as the data of VE.

Finally, regardless of the heating rate or content of flame retardant, the decomposition temperatures at 50% weightlessness for VE were nearly the same and showed no obvious rule as well as that for TSE. Such phenomenon reflected that all the decomposition reactions were in an intense stage which mainly dominated by the esterification of IFR (18).

According to the results above, the decomposition procedure presented certain rules in VE at 5°C/min, 10°C min, and 15°C/min, meanwhile it presented certain rule in TSE only at 5°C/min. Considering the same formula of flame retardant processed respectively by VE and TSE, the only reason can be confirmed that the different dispersion characteristics which endowed by different devices. In other words, the better dispersion and distribution of Nano-ZnO/IFR particles in PP/EPDM blend by VE (19), the lower decomposition rate and a longer time to reach the decomposition peak.

Figure 3 and Figure 4 present the DTG curves of flame retardant PP nanocomposites processed via VE and TSE at different heating rates. Combing with Table 2, there were not so much difference with the values of θ and T_max between VE and TSE. Meanwhile, the value of θ declined with the rising of the heating rate, the value of T_max was just the opposite. dα/dt increased with the rising of the heating rate, namely more heat energy released in unit time, which produced more char layer at the beginning of the esterification, the char layer separated oxygen and heat from the composites and atmosphere to postpone the process of complete combustion (20). The longer of the ignition time, the higher temperature of the composites at the most intense degradation reaction, and the more exhaustive ignition.

Figure 3

DTG curves of flame retardant PP nanocomposites processed via VE at different heating rates.

Figure 4

DTG curves of flame retardant PP nanocomposites processed via TSE at different heating rates.

Moreover, the average values of V_max for VE were a little lower than that for TSE, which meant that the esterification reaction peaks for VE were milder than that for TSE. It can be inferred that better dispersion and distribution of synergistic flame retardants make a relatively moderate reaction rate.

3.2 Thermogravimetic kinetics

3.2.1 Kissinger method

Kissinger method is a method of differentiation. Activation energies E were calculated according to different

temperatures which corresponding to peaks of DTG curves at four heating rates (21). Function equation is expressed as follow:

(2)lnβ/Tp2=ln(AR/E)−E/RTp

According to the relationship between T_p and β, curve fitting is undertaken. We plotted straight lines based on ln β / T²_pand 1/T_p with gradients of −E/R to calculate the activation energy E, and to obtain the pre-exponential factor A in line with intercept ln(AR/E). All the curves and data are shown in Figure 5 and Table 3.

Table 3

Activation energies and pre-exponential factors of all components.

Sample code	Activation energy (E) kJ/mol	Pre-exponential factor (lgA)	Correlation coefficient (r)
Lγ5/10/15/20	257.56	17.83	0.95219
Lδ5/10/15/20	211.21	14.49	0.91662
Lε5/10/15/20	193.14	13.08	0.97818
Lζ5/10/15/20	188.40	12.75	0.95510
Lη5/10/15/20	240.73	16.50	0.92772
Sγ5/10/15/20	158.42	10.69	0.79893
Sδ5/10/15/20	217.35	14.93	0.98227
Sε5/10/15/20	201.23	13.74	0.81770
Sζ5/10/15/20	167.99	11.22	0.54710
Sη5/10/15/20	220.32	12.08	0.92189

Figure 5

Plots of lnβ/Tp2    and1/Tpof flame retardant PP nanocomposites processed via VE and TSE.

In Figure 5, each straight line is fitted by four data points, the gradient of −E/R and intercept ln(AR/E) are shown in it with exact values. Moreover, in Table 3, all the correlation coefficients r are very close to 1, especially for VE, which represented high reliabilities of E and A (3).

The activation energy of Lη5/10/15/20 is higher than that of Sη5/10/15/20 by 20.41 kJ/mol, the percentage difference reaches 9.26%. Besides, the activation energy of Lζ5/10/15/20 is higher than that of Sζ5/10/15/20 also by 20.41 kJ/mol, the percentage difference reaches 12.15%. Furthermore, the activation energy of Lγ5/10/15/20 is higher than that of Sγ5/10/15/20 by 99.14 kJ/mol, the percentage difference reaches 62.58%. While the activation energy of Lε5/10/15/20 is less than that of Sε5/10/15/20 by 8.09 kJ/mol, the activation energy of Lδ5/10/15/20 is less than that of Sδ5/10/15/20 only by 6.14 kJ/mol. For most synergistic flame retardants, the best flame retardancy performance appeared when the content of synergist was 1-3 wt% (4,6,22). It can be inferred that more homogeneous samples owned higher thermal stability, namely VE endowed samples with 2, 3, 5 wt% Nano-ZnO better dispersion and distribution effects in processing.

The pre-exponential factor A, which represented the effective collision times among activated molecules, has a positive correlation exponential function relationship with the activation energy E (23). It can also be proved in Table 3 that the tendency of A was the same as E.

3.2.2 Flynn-Wall-Ozawa method

Flynn-Wall-Ozawa method is a method of integration. Activation energies E_o were calculated according to different temperatures which corresponding to the same degree of reactivity α in curves with different heating rates. We used integral transformation and Doyle approximation to change Eq. 1 into Eq. 3, in which, T was separately T₁, T₂, T₃, and T₄ that corresponding to the heating rates of 5°C/min, 10°C/min, 15°C/min, and 20°C/min. The activation energy E_o was confirmed by the linear relation between 1gβ and 1/T, there was no need to consider the difference of reaction mechanism (24).

(3)lgβ=lg(AE0/Rg(α))−2.135−0.4567E0/RT

All the values of E_o were calculated according to the gradients from Figure 6. It can be seen from Figure 6 that the linear fitting degree was high, which implied the method was feasible and the results were reliable. Referring to VE, As for 2, 3, 4, 5, 6 wt% Nano-ZnO/IFR, when 0.2 ≤ α ≤ 0.4, the average values of E_o were respectively144.58 kJ/mol, 179.80 kJ/mol, 195.46 kJ/mol, 183.69 kJ/mol, 168.04 kJ/mol; While 0.5 ≤ α ≤ 0.8, the average values of E_o were respectively 188.13 kJ/mol, 186.66 kJ/mol, 177.01 kJ/mol, 180.02 kJ/mol, 179.35 kJ/mol. Referring to TSE, As for 2, 3, 4, 5, 6 wt% Nano-ZnO/IFR, when 0.2 ≤ α ≤ 0.4, the average values of E_o were respectively 195.77 kJ/mol, 203.18 kJ/mol, 151.74 kJ/mol, 175.51 kJ/mol, 210.63 kJ/mol; While 0.5 ≤ α ≤ 0.8, the average values of E_o were respectively 219.51 kJ/mol, 204.74 kJ/mol, 149.23 kJ/mol, 184.90 kJ/mol, 243.06 kJ/mol.

Figure 6

Plots of 1gβ and 1/T of flame retardant PP nanocomposites processed via VE.

Table 4

Activation energies of all components.

Range of α	Activation Energy (kJ/mol)
	η	ζ	ε	δ	γ
0.2 ≤ α ≤ 0.4(VE)	144.58	179.80	195.46	183.69	168.04
0.2 ≤ α ≤ 0.4(TSE)	188.13	186.66	177.01	180.02	179.35
0.5 ≤ α ≤ 0.8(VE)	195.77	203.18	151.74	175.51	210.63
0.5 ≤ α ≤ 0.8(TSE)	219.51	204.74	149.23	184.90	243.06

According to data above, the activation energy of VE first rose then fell with the increasing content of Nano-ZnO when 0.2 ≤ α ≤ 0.4, and it showed a peak with 4 wt% addition of Nano-ZnO. While the activation energy of VE was not only much higher than that above, but also expressed just the opposite when 0.5 ≤ α ≤ 0.8. This was due to the Nano-ZnO might break the C=O bond of DPER and P-N bond between APP and PP at early stage (25), which destroyed the stability of initial esterification reaction of IFR. However, the increasing concentration of Nano-ZnO started to catalyze the esterification until the addition reached its critical value 4 wt%. Afterwards, the excessive and overweight Nano-ZnO broke the char layer that reduced the reaction rate. The activation energy maintained at a high level when 0.5 ≤ α ≤ 0.8 compared with 0.2 ≤ α ≤ 0.4, and the peak appeared when the addition of Nano-ZnO was 2 wt%. The high level of activation energy was not only ascribed to the polyphosphoric acid, volatile phosphorus oxides, and network polymers with P-N bond, all of which were derived from high temperature decomposed IFR (4,9), but also ascribed to the dehydration, cross-linking, catalytic carbonization by Nano-ZnO, heat, oxygen, and molten drop isolation of char layer (3,4,9,25). These made polymer molecular chain need more energy during thermal degradation. The totally different high activation energy ranges from Figure 6 proved the analysis above. In addition, the concentration of IFR and Nano-ZnO increased so gradually that the activation energies maintained around 200 kJ/mol except for ones with 4, 5 wt% Nano-ZnO in them. It can be inferred that overweight Nano-ZnO migration to the bottom of char layer because of gravity, which destroyed the integrity and compactness of char layer during its formation.

On the other hand, the activation energy of TSE (Figure 7) declined gradually but maintained at a high level with the increasing content of Nano-ZnO when 0.2 ≤ α ≤ 0.4, it showed a peak with 2 wt% addition of Nano-ZnO. While the activation energy of TSE was much higher than that above, and exhibited the same rule when 0.5 ≤ α ≤ 0.8. This is due to the reason that the dispersion and distribution of Nano-ZnO is not so well as in VE, the esterification of IFR at early stage is not disturbed so much as samples processed via VE. Furthermore, with a good foundation that formed by the synergistic flame retardant, the flame retardant effects that described above reach the most intense stage, which make the polymer molecular chains need more activation energy to break. The activation energies of samples that processed via TSE fluctuated much more than samples processed via VE. The reason may due to the heterogeneous of Nano-ZnO/IFR particles in the matrix.

Figure 7

Plots of 1gβ and 1/T of flame retardant PP nanocomposites processed via TSE.