Polymer Degradation and Stability 92 (2007) 189e196 www.elsevier.com/locate/polydegstab Thermogravimetric analysis of PVC/ELNR blends M.N. Radhakrishnan Nair a, George V. Thomas b, M.R. Gopinathan Nair b,* b a Department of Chemistry, D.B. College, Thalayolaparambu P.O., Kottayam, Kerala 686 605, India School of Chemical Sciences, Mahatma Gandhi University, Priyadarshini Hills P.O., Kottayam, Kerala 686 560, India Received 21 September 2006; received in revised form 24 November 2006; accepted 27 November 2006 Available online 17 January 2007 Abstract The thermal behaviour of a series of solution-cast blends of polyvinyl chloride/epoxidised liquid natural rubber (ELNR) of different mole percentage of epoxidation has been examined using thermogravimetric analysis. Thermal degradation is found to occur through a two-step route in which the first step corresponds to the dehydrochlorination of PVC to form a polyene and the second step is attributed to the decomposition of the ELNR and the polyene. Introduction of 20 and 50 mol% of epoxy group into the liquid NR is found to enhance the thermal stability of PVC. Probable mechanisms of degradation have been suggested on the basis of the kinetic analysis of the degradation studies. It is found that the mechanism is influenced by the epoxy content of the blend system. Activation energy for the degradation and the entropy change have also been reported. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Polymer blends; Thermogravimetry; Degradation kinetics; Activation energy; PVC; ELNR 1. Introduction It is well known that PVC being a proton-donating polymer interacts with oxirane rings of epoxidised oils and hydrocarbons resulting in the plasticization of the resin. Thus epoxidised oils have been used as partial replacement of plasticizers, which also act as thermal stabilizers for PVC [1]. Polymeric additives with epoxy groups are found to be better choices compared to the low molecular weight epoxy compounds for improving thermal characteristics of chlorinated polymers in general [2e5]. The oxirane rings lead to exothermic interaction with the chlorinated structural units in the polymer. Thus, epoxidised natural rubber has been used as a compatibiliser in blends consisting of chlorinated polyethylene and polyvinyl chloride [6,7]. This was found to be highly effective in getting a wide spectrum of mechanical properties for the blend system. The compatibility behaviour of epoxidised natural rubber with hydroxyl containing polymeric resins * Corresponding author. Tel.: þ91 481 2731036; fax: þ91 481 2731009. E-mail address: mrg.nair@rediffmail.com (M.R. Gopinathan Nair). 0141-3910/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2006.11.014 such as novolac, resole and epoxy resins was also studied and found to be much effective in diversifying their properties [8]. Epoxidised polybutadiene with a higher mole percent of epoxidation was found to be miscible with polyvinyl chloride. This was established by applying a scheme based on copolymere copolymer miscibility theory [9]. In the present work it is proposed to study the thermal behaviour of a few series of blends consisting of polyvinyl chloride and epoxidised liquid natural rubber (ELNR). ELNR is expected to act as a polymeric additive to PVC. Its low molecular weight will enable greater diffusion into the PVC matrix leading to molecular level interaction between the blend components. 2. Materials PVC with K value 65, M W ¼ 1:4  105 , and density 1.37 (M/ s Chemicals and Plastics, Chennai, India) was used for blending. Natural crumb rubber (ISNR-5) with M W ¼ 8:2  105 , and with intrinsic viscosity in benzene at 30  C ¼ 4.45 dL/g, was received from Rubber Research Institute of India (RRII), 190 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 Kottayam, India. Liquid natural rubber with M V ¼ 4900 (LNR) is obtained by depolymerisation of NR [10]. Epoxidised LNR (ELNR) of 20 mol% epoxidation (ELNR-20) and 50 mol% epoxidation (ELNR-50) were prepared by the epoxidation of LNR in our laboratory [11]. 2-Butanone, was supplied by E. Merck (India) Ltd, was dried over anhydrous calcium chloride and distilled before use. Glacial acetic acid, toluene, methanol and hydrogen peroxide (30 w/v) were of reagent grade supplied by BDH, Mumbai. Toluene and methanol were dried and distilled before use. 3. Experimental 3.1. Preparation of liquid natural rubber (LNR) It was prepared by a reported procedure [10] given as follows. Natural crumb rubber was masticated for 8 min at 40  C. About 100 g of it was dissolved in 1 l toluene and the solution was charged into a flat-bottomed borosilicate glass flask of 2 l capacity. The flask was fitted with a water condenser and a mechanical stirrer. Hydrogen peroxide solution (100 ml) was added and thoroughly mixed with the solution. The mixture was then homogenised to a certain extent by adding 150 ml of methanol. The whole assembly was placed in sunlight. After about 50 h of irradiation, the clear supernatant toluene layer was decanted. Toluene was removed by distillation under reduced pressure. The residual viscous liquid rubber containing some toluene was kept tightly corked in the absence of light. 3.2. Preparation of epoxidised liquid natural rubber 3.2.1. ELNR-50 LNR (25 g) dissolved in 166 ml of toluene to obtain a 15% solution was stirred at 50  C for 10 h with 1.1 mol of H2O2 and 0.35 mol of glacial acetic acid. At the end of the reaction, the product was isolated by precipitating from methanol and dried in vacuum. It was then characterized by titrimetric, IR, and NMR techniques. 3.2.2. ELNR-20 LNR (25 g) dissolved in 166 ml of toluene to obtain a 15% solution was stirred at 50  C for 3 h with 0.55 mol of H2O2 and 0.35 mol of glacial acetic acid. The product was isolated as before. 3.3. Preparation of blends A series of blends of PVC and liquid natural rubber/epoxidised liquid natural rubber were prepared with the help of a common solvent, viz., 2-butanone as follows. A 3% w/v solution of PVC in 2-butanone was prepared and added to the rubber solution made in the same solvent (30% w/v) at compositions such as 70/30, 50/50 and 30/70 of PVC and the rubber. It was thoroughly mixed by a magnetic stirrer for 5 h and cast on a glass plate. The samples were then dried in vacuum at 70  C for two days to remove the traces of residual solvent. 3.4. Thermogravimetric analysis The thermogravimetric analysis (TGA) was carried out on a Dupont 2000 TGA in N2 atmosphere at a heating rate of 10  C/min. The samples were scanned from 30  C to 600  C. 3.5. Designation of blends The blends were designated as follows. PVC/LNR (70/30) means a blend of 70 parts PVC and 30 parts LNR; PVC/ ELNR-20 (70/30) means a blend of 70 parts PVC and 30 parts epoxidised liquid natural rubber of 20 mol% epoxidation; PVC/ELNR-50 (70/30) means a blend of 70 parts PVC and 30 parts epoxidised liquid natural rubber of 50 mol% epoxidation, etc. 4. Results and discussion The blends were prepared in three series, viz., PVC/LNR, PVC/ELNR-20 and PVC/ELNR-50. Samples were made in different compositions such as 70/30, 50/50 and 30/70 of PVC and the rubber. All the samples were subjected to thermogravimetric analysis and the data were used to assess the effect of epoxy content of ELNR on the degradation pattern of PVC. For the sake of comparison thermogravimetric traces of PVC, LNR, ELNR-20, and ELNR-50 are given in Fig. 1(a)e(d). The corresponding data are tabulated in Table 1. A two-stage degradation pattern is seen in the case of PVC. The first stage begins at 180  C (T0) and ends at 375  C with a peak temperature at 300  C (T1max). This corresponds to a weight loss of 59.5%, which is attributed to the elimination of HCl molecules leaving behind longer polyene chains. The second stage of degradation begins at 375  C (T20) and ends at 490  C with a peak temperature of 466  C (T2max). Thermal degradation of the polyene sequences occurs during this stage yielding volatile aromatic and aliphatic compounds by the intramolecular cyclisation of the conjugated sequences [12]. The total weight loss at this stage is found to be 90%. Liquid natural rubber exhibits only a single stage degradation that occurs between 240 and 450  C. The peak temperature is observed at 384  C and the weight loss at the end of this stage is found to be 98%. Fragmentation of polyisoprene chains must occur during the degradation yielding volatile fragments such as isoprene and dipentene [13,14]. The degradation of ELNR-20 and ELNR-50 is found to follow similar pattern and shows only single stage degradation. The onset temperature rose from 240  C to 275  C and to 295  C, respectively, as the epoxy content is increased from 0 to 20 and to 50 mol%. Similarly the peak temperatures increased from 384  C to 396  C and to 415  C, respectively, with the corresponding increase in epoxy content. Weight loss for ELNR20 and ELNR-50 was 96% and 93%, respectively. The data clearly indicate that the breakdown of the rubber molecules is delayed by the presence of epoxy group, which could be due to the higher extent of interaction among the rubber molecules through the polar epoxy groups. 191 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 Fig. 1. Thermogravimetric curves of (a) PVC, (b) LNR, (c) ELNR-20 and (d) ELNR-50. Thermogravimetric curves of PVC/LNR blends at 70/30, 50/50 and 30/70 compositions are given in Fig. 2(a)e(c). The data obtained are noted in Table 2. All the samples undergo a two-stage degradation pattern. The first stage degradation with a peak temperature around 260  C for the different compositions corresponds to the dehydrochlorination of PVC. Weight loss at this stage appropriated to the PVC content in the blend and was found to be 43.4, 30, and 17.7%, respectively, which is consistent with that of PVC as described above. The table also shows that the peak temperature for the second stage degradation occurs around 450  C which is again the degradation temperature corresponding to that of the NR and the residual polyene formed after the dehydrochlorination of PVC. Over 90% weight loss occurs at this stage indicating the completion of the degradation process. All the above observations correspond to those of the constituent polymers irrespective of the sample composition suggesting that the degradation pattern of one constituent is not Table 1 Thermogravimetric data of PVC, LNR, ELNR-20 and ELNR-50 Sample PVC LNR ELNR-20 ELNR-50 First stage Second stage Onset temperature,  C (T10) Temperature range,  C Peak temperature,  C (T1max) Wt. loss (%) Peak temperature,  C (T2max) Wt. loss (%) 180 240 275 295 180e375 240e450 275e500 295e500 300 384 396 415 59.5 98 96 93 466 e e e 90 e e e 192 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 Fig. 2. Thermograms of (a) PVC/LNR (70/30), (b) PVC/LNR (50/50) and (c) PVC/LNR (30/70). influenced by the presence of the other. This is due to the lack of interaction between them. The TG and DTG curves of the PVC/ELNR-20 blends are given in Fig. 3(a)e(c). Data collected from them are tabulated in Table 2. Compared to PVC/LNR samples described above, peak temperature of the first stage degradation of this series occurs at higher temperatures, which seem to be increasing with the ELNR content of the sample. An increase from Table 2 Thermogravimetric data of PVC/LNR, PVC/ELNR-20 and PVC/ELNR-50 blends Blend PVC/LNR (70/30) PVC/LNR (50/50) PVC/LNR (30/70) PVC/ELNR-20 (70/30) PVC/ELNR-20 (50/50) PVC/ELNR-20 (30/70) PVC/ELNR-50 (70/30) PVC/ELNR-50 (50/50) PVC/ELNR-50 (30/70) First stage T(50)( C) Second stage Onset temperature,  C (T10) Temperature range,  C (T1max) Peak temperature,  C Wt. loss (%) Peak temperature,  C (T2max) Wt. loss (%) 125 140 148 150 185 186 189 216 217 125e380 140e380 148e390 150e350 185e375 186e350 189e385 216e400 217e400 256 261 266 250 281 277 276 283 284 62 60 59 59 55 52 55 49 48 440 452 440 433 474 462 450 480 471 92 93 94 89 85 83 86 82 81 327 313 326 296 341 415 351 365 385 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 193 Fig. 3. Thermograms of (a) PVC/ELNR-20 (70/30), (b) PVC/ELNR-20 (50/50) and (c) PVC/ELNR-20 (30/70).   250 C to 281 C is observed as the ELNR-20 content is increased from 30 to 50 parts and remains around this level even at 70 parts. The weight loss after the first stage degradation was found to be 59, 55, and 52%, respectively, for the samples with 30, 50 and 70 parts of the rubber and the values appropriated to the PVC content are 41.3, 27.5, and 15.6% for each sample. It is to be noted that the mass loss undergoes a decrease as the ELNR-20 content increases in the blend, which could be an indication of its stabilising effect on PVC. However, the second stage degradation occurs as before without any marked variation in the temperature range. Similar results are obtained in the case of the other series of samples, viz., PVC/ELNR-50. Thermogravimetric curves are given in Fig. 4(a)e(c) and the data are tabulated in Table 2. As shown in the table, peak temperature of the first stage degradation is higher at around 280  C and the weight loss appropriated to PVC content is decreasing from 38.5 to 14.4% as the ELNR-50 content is increased from 30 to 70%. Here again the second stage degradation proceeds as usual. It was well established that epoxy group absorbs hydrogen chloride (Scheme 1), there by stalling its catalytic activity for further dehydrochlorination of the PVC [1]. The temperature at which 50% weight loss happens is tabulated in Table 2. It is noted that the values are highest for the ELNR-50 blend reflecting maximum thermal stability. Percentage mass loss at three different temperatures selected at random, viz., 250, 350 and 450  C (Dm1, Dm2, Dm3) for all the three series of blends are shown in Table 3.These data are intended to follow the stability of the blends in terms of the mass loss. Values are lowest for the different compositions involving ELNR-50 indicating its role in the thermal stabilisation of PVC. A comparison of (T10) and (T1max) values for the blends in each series shows that 70/30 blends possess the lowest value, and 50/50 blends possess the highest value which tends to stabilise with the 30/70 composition. Hence it is found that 50 parts of ELNR with 50 mol% of epoxidation is a suitable composition which brings about maximum stabilisation to the PVC. 4.1. Degradation kinetics Solid-state thermal decomposition reactions were generally homogenous in nature. The most prominent rate controlling 194 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 Fig. 4. Thermograms of (a) PVC/ELNR-50 (70/30), (b) PVC/ELNR-50 (50/50) and (c) PVC/ELNR-50 (30/70). process operative in a particular case is chosen and used for deriving the rate equation. The decomposition kinetics of the blends was derived from the TG curves by applying an analytical method proposed by Coats and Redfern [15]. The integral equation used has the form      gðaÞ AR 2RT E ln ¼ ln 1  T2 FE E RT where g(a) is the kinetic model function, a is the fraction decomposed at any temperature, and F is the heating rate. A is the pre-exponential factor (Arrhenius parameter) which is calculated from the intercept by the relation intercept ¼ ln ðAR=FEÞ The slope of the plot of left hand side against 1/T is a straight line, from which the energy of activation, E, is calculated. Different authors, based on the integral method have derived several mechanistic equations to determine the decomposition kinetics. Satava [16] has chosen the nine equations Scheme 1. based on nine possible mechanisms [i.e., nine different forms of g(a)]. The form of g(a) that best represents the experimental data gives the proper mechanism. In the present study all TG data were analysed using the nine mechanistic equations proposed by Satava. Results of the analysis were used to derive the best fit curves with the highest correlation coefficients. The data are summarised in Table 4. The observed mechanisms of decomposition of the homopolymers are as follows. First stage decomposition of PVC involves the mechanism, viz., one-dimensional diffusion according to the mechanistic equation obtained in the form (a)2 ¼ kt. This step is due to the dehydrochlorination of PVC which is an autocatalytic process. Second stage involved the scission of the polyene chains and the Mampel equation (ln(1  a) ¼ kt) best represents the experimental data which give the mechanism, viz., random nucleation with one nucleus on each particle. LNR decomposition also followed the onedimensional diffusion mechanism. ELNR-20 and ELNR-50 decomposed by the three-dimensional diffusion pathways, which may be caused by the introduction of the oxirane rings into the chain. PVC/LNR blends showed the same decomposition mechanism as that of PVC, viz., one-dimensional diffusion for the first stage and random nucleation for the second stage indicating little interaction between the components. PVC/ELNR-20 (70/30) blend showed two mechanisms, viz., random nucleation for the first stage and three-dimensional diffusion for 195 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 Table 3 Mass loss due to the blends at different temperatures Blend Mass loss Dm1(%) (250  C) Dm2(%) (350  C) Dm3(%) (450  C) PVC/LNR (70/30) PVC/LNR (50/50) PVC/LNR (30/70) PVC/ELNR-20 (70/30) PVC/ELNR-20 (50/50) PVC/ELNR-20 (30/70) PVC/ELNR-50 (70/30) PVC/ELNR-50 (50/50) PVC/ELNR-50 (30/70) 32.7 15.6 17.5 12.9 11.0 9.8 11.9 9.7 9.3 69.7 55.6 52.5 67.4 54.8 51.2 48.8 47.6 46.3 77.8 76.3 67.5 74.2 67.4 58.1 66.7 65.1 56.1 the second stage which are due to PVC and ELNR-20, respectively. Existence of two different mechanisms indicates insufficient interaction between the blend components. For all other blends, viz., PVC/ELNR-20 (50/50), PVC/ELNR-20 (30/70), PVC/ELNR-50 (70/30), PVC/ELNR-50 (50/50) and PVC/ ELNR-50 (30/70), the Mampel equation best represents the experimental data. This corresponds to the random nucleation mechanism. It is noteworthy that the above data show a change in mechanism with an increase in the mol% of epoxidation and also with an increase in the quantity of epoxidised rubber present in the blends. From this it is obvious that the blend characteristics undergo a change with respect to the epoxy content which could be due to the greater interaction between the blend components enabled by the higher epoxy content. Activation energy for the first stage degradation of PVC, LNR, ELNR-20 and ELNR-50 was 72.94, 42.7, 78.56, and 83.94 kJ/mol, respectively. It could be noted that the activation energy of liquid rubber sharply increases from 42.7 to 78.56, and to 83.94 kJ/mol with an increase in mol% of epoxidation from 0 to 20% and then to 50%. This is a case of the influence of epoxy group on the degradation of the rubber molecules. Regarding the blends, the activation energy for the first stage degradation of PVC/LNR is in the order 46.1, 42.8, 35.18 kJ/mol for the 70/30, 50/50 and 30/70 blends, respectively. For PVC/ELNR-20 blends the values are 40.9, 72.84, 77.25 kJ/mol for the three compositions. Higher values of activation energy at higher ELNR content indicate greater degree of interaction between the blend components which slows down the dehydrochlorination process. PVC/ELNR-50 blends showed still higher values, viz., 64, 123 and 109 kJ/ mol, respectively. The marked increase in activation energy values in this case is also in agreement with the observation that the liquid rubber with 50 mol% of epoxidation leads to greater stabilisation of PVC against the dehydrochlorination. Activation energy for the second stage degradation of PVC is 102.9 kJ/mol. The values remain high for all the blend systems (Table 4). Higher activation energy is construed by the fact that this stage involves chain breaking. The entropy of activation (DS ) was also calculated for each stage of thermal decomposition for the various blends using the relationship,
A ¼ kTs =h eDS=R where A is the Arrhenius parameter, k is the Boltzmann constant, Ts is the peak temperature, DS is the entropy of activation and R is the gas constant. Table 4 shows that DS has negative values. The negative values suggest that the segment undergoes some chain alignment at elevated temperatures forming the activated complex. Entropy of activation in the first stage ranges from 238 to 272 J/degree/mol. Second stage values are little lower and ranges from 186 to 263 J/degree/mol. Arrhenius parameter, A, exhibits a regular trend. Values for second stage degradations are much higher than the first stage values as shown in Table 4. Higher values for the second stage could be accounted by the higher collision frequencies at the degradation temperatures. 5. Conclusion Thermogravimetric data showed that (T10) and (T1max) values of liquid natural rubber increased with an increase in the epoxy content, indicating that the presence of epoxy group delayed the breakdown of the liquid rubber molecules. For the PVC/ LNR blends, the thermal data correspond to those of the Table 4 Kinetic parameters of the blends and the blend components Sample PVC LNR ELNR-20 ELNR-50 PVC/LNR (70/30) PVC/LNR (50/50) PVC/LNR (30/70) PVC/L-ELNR-20 (70/30) PVC/ELNR-20 (50/50) PVC/ELNR-20 (30/70) PVC/L- ELNR-50 (70/30) PVC/ELNR-50 (50/50) PVC/ELNR-50 (30/70) Activation energy E (kJ/mol) Arrhenius parameter A (s1) Entropy of activation DS (J/degree/mol) Stage 1 Stage 2 Stage 1 Stage 1 Stage 2 73 43 79 84 46 43 35 41 73 77.25 64 122.6 109 103 e e e 76 154 117 125 160 183 204 222 200 2 530 649 943 1 0.2 1 714 832 196 609 326 439 246 243 248 271 265 265 270 271.8 238 241 247 247 218 236 e e e 263 199 228 235 186 201 228 218 183 Stage 2 7 e e e 916 574 451 858 2629 553 2500 46 058 4109 196 M.N. Radhakrishnan Nair et al. / Polymer Degradation and Stability 92 (2007) 189e196 constituent polymers irrespective of the sample composition suggesting that the degradation pattern of one constituent is not influenced by the presence of the other. This could be due to the lack of interaction between the blend components. Compared to PVC/LNR samples, the peak temperatures of the first stage degradation of PVC/ELNR-20 and PVC/ELNR-50 series occur at higher temperatures, which seem to be increasing with the ELNR content of the respective blend. Mass loss undergoes a decrease as the ELNR content increases in the blend. This is maximum for ELNR-50, which could be an indication for its stabilising effect on PVC. The second stage degradation of all samples, however, occurs without any marked variation. Higher activation energy indicated from kinetic analysis also supports this observation. From kinetic studies it is seen that PVC/LNR blends exhibited the same decomposition mechanism as that of PVC, viz., one-dimensional diffusion for the first stage and random nucleation for the second stage indicating little interaction between the components. PVC/ELNR-20 (70/30) blend showed two mechanisms; random nucleation for the first stage and three-dimensional diffusion for the second stage, which are due to PVC and ELNR-20, respectively, indicating insufficient interaction between the blend components. For all other blends, the Mampel equation best represents the experimental data. This corresponds to the random nucleation mechanism. The change in mechanism is obviously caused by an increase in the epoxy content of the blend system which offers greater interaction between the blend components. References [1] Penn WS. PVC technology. 3rd ed. London: Applied Science; 1971. p. 188. [2] Margaritis AG, Kalfoglou NK. Polymer 1987;28:497. [3] Kallistis JK, Kalfoglou NK. Angew Makromol Chem 1987;148:103. [4] Varughese KT, Nando B, Dey SK, Sanyal SK. J Mater Sci 1989;24:3491. [5] Varughese KT. Kautsch Gummi Kunstst 1988;41(11):1114e7. [6] Klaric I, Vrandecic NS, Roje U. J Appl Polym Sci 2000;78(1):166e72. [7] Kolesov SV, Kulish EI, Minsker KS. Vysokomol Soed Seriya A & B 2000;42(2):306e11. [8] Kallistis JK, Kalfoglou NK. J Appl Polym Sci 1989;37:453e65. [9] Margaritis AG, Kalfoglou NK. Eur Polym J 1988;24:1043e7. [10] Ravindran T, Gopinathan Nair MR, Francis DJ. J Appl Polym Sci 1987;35:1227. [11] Thomas George V, Gopinathan Nair MR. Kautsch Gummi Kunstst 1997;50(5):398. [12] Wypych J. Polyvinyl chloride degradation. Amsterdam: Elseveir; 1985. [13] McNeill IE, Ackerman L, Guptha SN. J Polym Sci Polym Chem Edn 1978;16:2169. [14] McNeill IE, Guptha SN. Polym Degrad Stab 1985;2:95. [15] Coats W, Redfern JP. Nature 1964;201:68. [16] Satava V. Thermochim Acta 1971;2:423.