Shanghai Jiao Tong University, 1954 Huashan Road, Shanghai 200030, People's ... and High Performance Materials, The University of Southern Mississippi,.
In Situ Fourier Transform Infrared Spectroscopic Study of the Thermal Degradation of Isotactic Poly(propylene) PENG HE, YAN XIAO, PUMING ZHANG, NA ZHU, XINYUAN ZHU,* and DEYUE YAN Institute of Polymer Materials, College of Chemistry and Chemical Technology, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, People’s Republic of China (P.H., Y.X., N.Z., X.Z., D.Y.); Department of Biomedical Engineering, Shanghai Jiao Tong University, 1954 Huashan Road, Shanghai 200030, People’s Republic of China (P.Z.); and Shelby F. Thames Polymer Science Research Center, School of Polymers and High Performance Materials, The University of Southern Mississippi, Hattiesburg, Mississippi 39406 (P.H.)
The conformational change of isotactic poly(propylene) (iPP) during the thermal degradation process has been carefully studied by in situ Fourier transform infrared (FT-IR) spectroscopy. This new method of studying thermal degradation of iPP not only shows the conventional kinetic parameter information of thermal degradation such as the degradation activation energy DE and the degradation factor n, which are in accord with the results of traditional thermogravimetry experiments, but also indicates that many significant molecular structure changes occur during the thermal degradation process that come from some characteristic absorption band changes of in situ FT-IR. A multivariate approach, principal components analysis (PCA), is applied to the analysis of infrared (IR) data, and the results further confirm the multi-step processes of the thermal degradation of iPP. Above all, this is a new application to polymer thermal degradation by in situ FT-IR that connects the intermediate conformational change with final results during thermal degradation. Index Headings: Fourier transform infrared spectroscopy; In situ; FT-IR spectroscopy; Thermal degradation; Isotactic poly(propylene); Multivariate analysis.
INTRODUCTION Resource reuse of waste plastic has attracted more and more attention in recent years due to the requirements of environment protection. Poly(propylene) (PP) is one of the major components of waste plastic, and its thermal degradation has been extensively investigated for some decades.1–7 Most of these works used the traditional thermogravimetry (TG) technique to characterize the thermal decomposition behaviors because of the simple experimental principle and process, accurate experimental data, and straightforward graph analysis. Generally, it is found that poly(propylene) degrades via random chain scissions, and the first-order kinetic model can be used to describe the degradation behavior. The apparent degradation activation energy is estimated as ;260 KJ/mol for isotactic poly(propylene) (iPP) and 230 KJ/mol for atactic poly(propylene) (aPP). Combining the TG technique with the interfaced pyrolysis gas chromatography (PyGC), Kiang and his co-workers analyzed the degradation products of PP and identified that the main products in decreasing yields were 2,4-dimethyl-1-heptene, 2-pentene, propylene, 2-methyl-1-pentene, and 2,4,6-trimethyl1-nonene. 2 This result is similar to that reported by BockReceived 22 January 2004; accepted 18 August 2004. * Author to whom correspondence should be sent. E-mail: xyzhu@ sjtu.edu.cn.
Volume 59, Number 1, 2005
horn et al., who found that the main degradation products were C12 compounds, like 2,4,6-trimethyl-1-nonene, and C14 compounds, like 2,4,6,8-tetramethyl-1-decene.5 Unfortunately, in spite of the popular application of TG in thermal degradation, a major drawback of this analysis method is that it cannot directly provide molecular structure change during the degradation process. There are several experimental methods to characterize the molecular structure of polymers, and infrared (IR) spectroscopy is one of the most commonly used techniques. Some absorption bands in the IR spectra are very sensitive to the molecular structure and physical state of polymers. If the IR spectra can be recorded continually during the heating process, the conformation and molecular structure change of polymer chains can be inferred by the change of various absorption bands. In this work, both the traditional TG method and a new time-resolved Fourier transform infrared (FT-IR) spectroscopic technique are used to study the thermal degradation of isotactic poly(propylene) (iPP). It will be found that compared with the TG analysis, the FT-IR technique can give much important information regarding molecular structure change in the thermal degradation process. EXPERIMENTAL The iPP sample used in this study was kindly supplied by Shanghai Jinshan Petrochemical Corp., Shanghai, P.R. China. The isotacticity of the sample is approximately 94.5%, and the melt flow index is 3.0 g/10 min. The IR spectra were collected at a resolution of 4.0 cm21 using a Bruker Equinox-55 FTIR spectrometer equipped with a variable temperature cell under a flowing nitrogen atmosphere. The film sample, about 20 mm thick, was heated from room temperature (20 8C) to 450 8C at a heating rate of 10 8C/min. At the same time, the FT-IR spectra were recorded at the rate of 1 spectrum per 4 s. Thus, the relationship of IR intensity with temperature can be obtained. Since each band has its own intensity coefficient, the IR intensities of observed bands have been normalized to the maximum values. The temperature had been carefully calibrated and the scanned wavenumber range was 4000 to 400 cm21. The iPP sample began to degrade at about 240 8C and the degradation ended at about 400 8C. The IR data were also treated by using principal components analysis (PCA), one of the most commonly used
0003-7028 / 05 / 5901-0033$2.00 / 0 q 2005 Society for Applied Spectroscopy
APPLIED SPECTROSCOPY
33
FIG. 1. FT-IR spectra for the thermal degradation of iPP at a heating rate of 10 8C/min.
FIG. 2. Freeman-Carroll plots of d[Ln(da/dt)]/d[Ln(1 2 a)] vs. d(1/ T)/d[Ln(1 2 a)] for the thermal degradation of iPP at various absorption bands.
multivariate analysis techniques. The PCA model was built by using the Matlab 6.5 software. Thermogravimetric analysis was performed with a TA Instruments Model 2050 thermogravimetric analyzer in nitrogen at a gas flow rate of 40 mL/min with a sample weight of 0.6690 mg. The heating rate was 10 8C/min, and the iPP sample was degraded from 240 8C to 400 8C. Differential scanning calorimetry (DSC) traces were recorded with a Perkin-Elmer Pyris-1 Series differential scanning calorimeter under a flowing nitrogen atmosphere. The DSC was carefully calibrated with In and Zn standards before measurement. The heating rate was 10 8C/min and the weight of the sample used in DSC was approximately 2.0 mg. The obtained result shows that the melting temperature of iPP sample used in IR measurements is 161 8C.
tion bands in the FT-IR experiment, a is also proportional to the intensity change of various absorption bands during the degradation process at time t. The Freeman-Carroll, Friedman, and Chang techniques are three single heatingrate treatment methods for TG and FT-IR curves to obtain the kinetic parameters of thermal degradation.8–10 The respective equations are:
RESULTS AND DISCUSSION
Chang:
The IR spectra of iPP were measured at different temperatures at a heating rate of 10 8C/min. With increasing temperature, some absorption bands become stronger and stronger, while the intensities of other bands are considerably weakened. Based on the relationship of band intensity with temperature, the kinetic parameter of various bands in the degradation process can be obtained. Figure 1 gives FT-IR spectra for the thermal degradation of iPP at the heating rate of 10 8C/min. Here, we use three different data treatment methods (Freeman-Carroll,8 Friedman,9 and Chang10) to calculate the degradation activation energy DE and the degradation factor n. The thermal degradation kinetics related to TG weight loss data came from the kinetic equation: da/dt 5 Z 3 (1 2 a) ne2D E/(RT)
Volume 59, Number 1, 2005
1 2 1 2@DLn(1 2 a)
(2)
1 dt 2 2 n 3 Ln(1 2 a) 1 (R 3 T)
(3)
DLn(da/dt) DE 1 5n2 3D DLn(1 2 a) R T Friedman: Ln(Z) 5 Ln
Ln
[
DE
da
]
(da/dt) DE 5 Ln(Z) 2 n (1 2 a) (R 3 T)
(4)
By plotting DLn(da/dt)/DLn(1 2 a) against D(1/T)/ DLn(1 2 a), Ln(da/dt), or Ln(1 2 a) against 1/T, and
(1)
where a is weight loss of the polymer undergoing degradation at time t, da/dt means the degradation rate or weight loss rate, Z is the frequency factor, n stands for the degradation factor, DE represents the degradation activation energy, R is the gas constant (8.3136 J mol 21 K21), and T denotes the absolute temperature (K).8–11 Because the weight loss of the polymer in the TG experiment can be correlated to the intensity change of absorp34
Freeman-Carroll:
FIG. 3. Friedman plots of Ln(da/dt) vs. 1/T for the thermal degradation of iPP at various absorption bands.
FIG. 4. Friedman plots of Ln(1 2 a) vs. 1/T for the thermal degradation of iPP at various absorption bands.
Ln[(da/dt)/(1 2 a) n] against 1/T, respectively, the values of the kinetic parameters such as the degradation activation energy DE and the degradation factor n can be determined from the slope and the intercept, respectively. Figures 2 through 5 show the three different techniques to determine the kinetic parameters, and the results are summarized in Table I. For comparison, the kinetic parameters obtained by TG are also listed. It can be found that both IR and TG analysis give identical kinetic parameters: the degradation activation energy DE is approximately 260 KJ/mol and the degradation factor n is 1.1. These results confirm that the apparent reaction order of thermal degradation is 1.1 and that iPP degrades via
FIG. 5. Chang plots of Ln[(da/dt)/(1 2 a) n] vs. 1/T for the thermal degradation of iPP at various absorption bands.
random chain scissions, which is in agreement with the reports in the literature.1–6 The agreement of the timeresolved FT-IR technique with TGA results supports the conclusion that FT-IR can be used for kinetic studies on the thermal degradation of polymers. On the other hand, the difference of activation energy DE for various absorption bands provides important information about the molecular structure change in the degradation process. The intensity of the 1289 cm21 band increases greatly after iPP crystals are melted, and increases with further increasing temperature. Thus, it can be inferred that the 1289 cm21 band is related to the amorphous state, representing the disordering of polymer
TABLE I. Kinetic parameters of thermal degradation of isotactic polypropylene calculated by three different techniques.a FT-IR wavenumber (cm21) Parameter
1720
1460
1380
1289
1153
1100
973
TGA
Friedman DE n Ln(Z) Standard error
253 1.1 51.4 0.0100
253 1.0 53.2 0.0060
279 1.0 61.2 0.0070
208 1.0 38.2 0.0080
297 1.2 64.9 0.0030
235 1.0 48.9 0.0060
236 1.0 49.4 0.0070
264 1.0 51.2 0.0060
Freeman-Carroll DE n Ln(Z) Standard error
257 1.2 52.3 0.0152
252 1.1 53.0 0.0230
270 1.1 59.2 0.0141
231 1.1 46.8 0.0221
286 0.9 62.2 0.0032
221 1.1 45.4 0.0076
236 1.1 49.5 0.0157
250 1.1 48.4 0.0150
Chang DE n Ln(Z) Standard error
264 1.2 53.9 0.0275
258 1.2 54.4 0.0338
285 1.2 62.7 0.0465
233 1.2 47.2 0.0039
299 1.2 65.3 0.0822
238 1.2 49.2 0.0939
245 1.2 51.6 0.1282
281 1.2 55.0 0.0600
Average DE n Ln(Z) Standard error
258 1.2 52.5 0.0176
254 1.1 53.5 0.0209
278 1.1 61.0 0.0225
224 1.1 44.1 0.0113
294 1.1 64.1 0.0295
231 1.1 47.8 0.0358
239 1.1 50.2 0.0503
265 1.1 51.5 0.0270
Overall Average DE n Ln(Z) Standard error a
IR 254 1.1 53.3 0.0268
TG 265 1.1 51.5 0.0270
The units of DE and Ln(Z) are KJ/mol and min21, respectively, and the heating rate was 10 8C/min.
APPLIED SPECTROSCOPY
35
TABLE II. Spectral assignments of isotactic polypropylene during the thermal degradation process. FT-IR wavenumber (cm21) 1720 1460 1380 1289 1153 1100 973
Spectral assignments carbonyl group scission of the main chain itself methyl scission from main chain amorphous state regular head-to-tail sequences with a certain length regularity band of short isotactic helices regularity band of short isotactic helices
Band shape changes intensity intensity intensity intensity intensity
increases decreases decreases increases decreases
intensity decreases intensity decreases
conformation. The disordering of iPP chains occurs before the degradation, so the activation energy DE1289 of the 1289 cm21 band (224 KJ/mol) is the lowest. Both the 973 cm21 and 1100 cm21 bands belong to the regularity
FIG. 6.
36
band and are associated with the presence of short isotactic helices.12–19 The critical length of helical isotactic sequences of the 973 cm21 band is shorter than that of the latter.15–17 With increasing temperature, the intensities of the 973 and 1100 cm21 bands decrease, indicating the destruction of the short 31 helix structure. The activation energy DE1100 of the 1100 cm21 band (231 KJ/mol) is slightly lower than that of the 973 cm21 band (DE973 5 239 KJ/mol) because the long 31 helix structure is more easily destroyed. The activation energy DE1460 of the 1460 cm21 band (254 KJ/mol) is almost similar to that of the 1720 cm21 band (DE1720 5 258 KJ/mol), implying the occurrence of thermal degradation of iPP main chains. Since methyl scission from the main chain is more difficult than scission of the main chain itself, the activation energy DE1380 of the 1380 cm21 band (278 KJ/mol) is a little higher than that of the 1460 cm21 band. Chain decomposition of the iPP main chain might be induced by the scission of carbon–carbon bonds as in the following reactions:
Principal components analysis applied to the IR data (top) and plots of PC1 (lower left) and PC2 (lower right).
Volume 59, Number 1, 2005
–[–CH2–CHMe–CH2–CHMe–]– → –[–CH2–C·H–CH2–CHMe–]– 1 ·Me
(5)
–[–CH2–CHMe–CH2–CHMe–]– → –[–CH2–C·HMe] 1 [·CH2–CHMe–]–
(6)
It has been reported 2,5 that the first reaction (Eq. 5) needs higher energy than the second one (Eq. 6), which means that methyl scission from the main chain is more difficult than scission of the main chain. This conclusion is further confirmed by our work. Finally, the disappearance of the 1153 cm21 band signals the end of iPP thermal degradation. The absorption at 1153 cm21 comes from the regular head-to-tail sequences with a certain length. At high temperature, iPP main chains are broken and the long regular sequences are destroyed, which leads to the complete decomposition. As a result, the activation energy of the 1153 cm21 band (294 KJ/mol) is the highest in all of the bands that we have studied. The spectral assignments along with information on data treatment of isotactic polypropylene during thermal degradation are summarized in Table II. Consequently, the molecular structure change of iPP chains in thermal degradation can be described as follows: As the temperature increases, the disorder of molecular chains increases. The short helix conformation existing in the melt becomes shorter and shorter, and is gradually destroyed due to the thermal disturbance. Further increasing the temperature leads to the degradation of iPP main chains via random chain scission. Finally, the iPP chains are completely degraded into small molecules. A multivariate analysis technique, principal components analysis (PCA), was also used to analyze the IR data. This approach enables the analysis of the variance in data to be simplified by extracting the principal components. Figure 6 shows the scores plot of PC1, which accounts for 64.31% of the variance in the spectral data, against PC2, which accounts for 29.60%. It can be found that an obvious transition occurs in the vicinity of 160 8C. DSC measurement has proven that the melting point of this iPP sample is 161 8C, so the transition at 160 8C in the scores plot corresponds to the melting of iPP crystals. From the melting point to 280 8C, the variation of the scores plot in Fig. 6 is not linear, indicating the multiple transition of macromolecular structure in this temperature region. Above 290 8C, the random chain scissions predominate, and therefore the scores plot changes linearly. The multivariate analysis of IR data confirms the degradation mechanism mentioned above. Furthermore, the IR spectra of degradation products also give some important information. For example, despite its low intensity, the 973 cm21 band still exists in degradation products. This result demonstrates that some degradation products of iPP have short 31 helix conformation. This conclusion is also in agreement with the literature, in which it had been found that the major degradation products of iPP include 2,4,6-trimethyl-1-heptene, 2,4,6-trimethyl-1-nonene, and 2,4,6,8-tetramethyl-1decene, etc. 2,5 The absorption band at 1720 cm21 indicates that although the degradation carries through in nitrogen
protection, oxidative degradation cannot be totally avoided due to the existence of small amounts of oxygen. With the appearance of the 1720 cm21 band, the thermal degradation of iPP takes a turning point in which carbon– carbon bonds and carbon–hydrogen bonds change into carbon–oxygen double bonds and the degradation is transferred to a new covalent bond coming into being from old Van der Waals interactions. Compared with the TG experimental results, the FT-IR experimental data indicate that the scission of main chain carbon–carbon bonds represents the start of thermal degradation of iPP chains. CONCLUSION In situ Fourier transform infrared spectroscopy offers a new and comprehensive way to investigate the thermal degradation of polymers. FT-IR provides information on the thermal degradation process, not only including the important kinetic parameters such as the degradation activation energy DE and the degradation factor n, but also involving information on molecular structure changes during the thermal degradation process. In our study, various absorption band changes show that the disordering of iPP chains occurs (the increase of the 1289 cm21 band), followed by the destruction of the short 31 helix structure (the decrease of the 1100 and 973 cm21 bands), followed by the appearance of oxidative degradation of iPP main chains (the increase of the 1720 cm21 band) and chain decomposition (the decrease of the 1460 and 1380 cm21 bands), followed by complete decomposition (the disappearance of the 1153 cm21 band). This degradation mechanism is further supported by the principal components analysis of IR data. The advantage of applying the time-resolved FT-IR technique to thermal degradation is that both macroscopic physical behaviors and microcosmic molecular movements can be measured at the same time. ACKNOWLEDGMENTS This work is supported by the National Natural Science Foundation of China (20104004). The authors also gratefully acknowledge support from the State Key Laboratory for Modification of Chemical Fiber and Materials.
1. L. A. Wall and S. Straus, J. Polym. Sci. 44, 313 (1960). 2. J. K. Y. Kiang, P. C. Uden, and J. C. W. Chien, Polym. Degrad. Stab. 2, 113 (1980). 3. B. Dickens, J. Polym. Sci., Polym. Chem. Ed. 20, 1169 (1982). 4. C. Vasile, L. Odochian, and I. Agherghinei, J. Polym. Sci., Polym. Chem. Ed. 26, 1639 (1988). 5. H. Bockhorn, A. Hornung, U. Hornung, and D. Schawaller, J. Anal. Appl. Pyrolysis 48, 93 (1999). 6. T. Ishikawa, T. Ohkawa, M. Suzuki, T. Tsuchiya, and K. Takeda, J. Appl. Polym. Sci. 88, 1465 (2003). 7. Z. Gao, T. Kaneko, I. Amasaki, and M. Nakada, Polym. Degrad. Stab. 80, 269 (2003). 8. E. S. Freeman and B. Carroll, J. Phys. Chem. 62, 394 (1958). 9. H. L. Friedman, J. Polym. Sci., C 6, 183 (1964). 10. W. L. Chang, J. Appl. Polym. Sci. 53, 1759 (1994). 11. X. S. Wang, X. G. Li, and D. Y. Yan, Polym. Degrad. Stab. 69, 361 (2000). 12. G. Zerbi, F. Ciampelli, and V. Zamboni, J. Polym. Sci., C 7, 141 (1963). 13. T. Miyamoto and H. Inagaki, J. Polym. Sci. A-2 7, 963 (1969).
APPLIED SPECTROSCOPY
37
14. Y. V. Kissin, Adv. Polym. Sci. 15, 92 (1975). 15. X. Y. Zhu, D. Y. Yan, H. X. Yao, and P. F. Zhu, Macromol. Rapid Commun. 21, 354 (2000). 16. X. Y. Zhu, D. Y. Yan, and Y. P. Fang, J. Phys. Chem. B 105, 12461 (2001).
38
Volume 59, Number 1, 2005
17. X. Y. Zhu, Y. P. Fang, and D. Y. Yan, Polymer 42, 8595 (2001). 18. X. Y. Zhu, Y. J. Li, D. Y. Yan, Q. H. Lu, and P. F. Zhu, Colloid Polym. Sci. 279, 292 (2001). 19. X. Y. Zhu, D. Y. Yan, Y. P. Fang, and L. Chen, Appl. Spectrosc. 57, 104 (2003).