Rheological and calorimetric properties of recycled bisphenol A poly(carbonate)

Polymer Degradation and Stability 82 (2003) 99–104 www.elsevier.com/locate/polydegstab Rheological and calorimetric properties of recycled bisphenol A poly(carbonate) J.F. Feller*, A. Bourmaud Polymers and Processes Laboratory, South Brittany-University, 56 321 Lorient, France Received 12 February 2003; received in revised form 4 April 2003; accepted 23 April 2003 This paper is dedicated to A. Feller for his action in environment preservation. Abstract Mechanical recycling of engineering polycarbonate wastes provides an interesting way to decrease the impact on environment, if the degradation due to successive processing is properly characterized. We analyzed polycarbonate (PC) degradation generated by eight successive recycling, using rheological, calorimetric and colorimetric tools at each step. From the first to the fifth recycling, an increase of storage modulus E 0 (at T=30  C and f=1 Hz) in both bending and tension mode is observed, together with a decrease : of the Newtonian limit viscosity Z0 (at T=240  C and
=0.1 s1). Moreover, a decrease of Tg of about 3.5  C and of
Cp at Tg is noticed from the first to the eighth recycling. Accelerated aging under UV irradiation shows a more rapid coloration of recycled PC. All these results evidence a decrease of molar mass resulting from thermomecanical stress encountered by the PC during recycling. This modification of polycarbonate properties, which remaining good, must be taken into account to adjust the processing conditions and determine the possible applications. # 2003 Elsevier Ltd. All rights reserved. Keywords: Polycarbonate; Recycling; Rheology; Injection molding 1. Introduction In the past decades, a great deal of attention has been focused on plastics recycling. Due to an increase of consideration of the impact of human activity on the environment, it is meaningful to introduce recycling in the product design. Several ways to use alternative feedstock for production of plastics are known resulting from, thermal, mechanical [1–3], chemical and enzymatic recycling [4]. Thermal degradation of polymeric materials into low molecular weigh compounds generally requires high temperatures and thus does not lead to interesting energy balance. Chemical treatment can be rather expansive and do not apply to all plastics. Enzymatic depolymerization of esters and carbonates seem to be one promising answer for green polymer production although extensive study is still needed. Mechanical recycling of post consumer commodity products and engineering plastics wastes can also provide a interesting feedstock provided that few * Corresponding author. E-mail address: jean-francois.feller@univ-ubs.fr (J.F. Feller). 0141-3910/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0141-3910(03)00169-1 contaminants are introduced and that degradation due to successive processing is properly characterized. In the present study, we investigated the changes in rheological and calorimetric properties of polycarbonate (PC) as a function of the number of recycling steps. 2. Experimental 2.1. Materials Production wastes of Bayloy1 (Axxis) Polycarbonate (PC) sheets were obtained from Self-Signal company. This formulation is based on Makrolon1 3103 (Bayer). The main characteristics of the virgin PC are recalled in Table 1 and its chemical structure in Scheme 1. The processing cycle we have used consisted in a first grinding followed by a subsequent injection. After each cycle, one part of the samples was ground for injection and the other was taken for analysis. Samples of cycles nos. 1, 2, 3 and 4, were tested by calorimetry, dynamic mechanical analysis, and cone/plate rheometry whereas 100 J.F. Feller, A. Bourmaud / Polymer Degradation and Stability 82 (2003) 99–104 Table 1 Polymer characteristics from producer Poly(carbonate) Tg ( C) glass temperature transition E (GPa) Young modulus "r (%) Strain at break  r (MPa) Stress at break l (W m1, K1) Thermal conductivity d Density at 23  C 148 2.3 100 65–70 0.21 1.2 (from feeding zone to die): 270/290/290/290  C corresponding to processing temperature of 290  C. All samples were cut from extruded tapes. Photodegradation was studied in a device allowing ultra violet irradiation 24 h/24 h at a controlled temperature of 40  C, to simulate aging due to sun exposure but in accelerated conditions. In fact, 40 days exposure in artificial conditions are expected to correspond to about 200–300 days exposure in real conditions. Samples used for accelerated aging studies were either ground and extruded for color analysis or ground and injected for dynamic mechanical analysis before they were introduced in the irradiation device. The surface color changes were analyzed with a Minolta CR 500 colorimeter using L a b coordinates. Scheme 1. Bisphenol A polycarbonate chemical formula. 3. Results and discussion samples of cycles no. 4, 5, 6 and 7 were only analyzed by the first two techniques. 2.2. Techniques Rheological properties of polymers were determined with a ThermoHaake RheoStress 1 rheometer with cone/plate geometry in both steady and dynamic mode. Cone and plate of 20 mm diameters were separated by a 60 mm gap. All experiments in dynamic mode were proceeded at 1 Hz. The amplitude of oscillations was fixed at 2% after determination of the linear viscoelasticity range. Dynamic mechanical analysis was performed on a T.A. Instruments 2980 DMA with the three-points bending accessory, at a frequency of 1 Hz, with an amplitude of 50 mm. A static force of 0.01 N and an autotension of 120% were applied to 21040 mm3 samples. The distance between supports was 20 mm. Temperature was scanned from 30 to 170  C with a 3  C min1 heating rate. For tension mode, samples of 21010 mm3 were used. Calorimetric measurements were made on a PerkinElmer Pyris 1 differential scanning calorimeter (DSC) with Pyris V3.0 software for data collection and treatment. The calibration was done with indium and zinc. The base line was checked every day. Aluminum pans with holes were used and the samples mass was approximately 10 mg. All the temperatures measured from a peak maximum (Tc, Tm) are determined to better than  0.5  C and from a sigmoid (Tg) at less than  1  C. Sample processing. After grinding and 24 h drying at 120  C, polycarbonate wastes pellets were introduced in a Boy injection press (screw diameter: =24 mm, maximal locking force: 220 kN) with the temperature profile: 280/285/290 (from feeding zone). Extrusion was performed using a Fairex single screw extruder (L=600 mm, =30 mm) with the following temperature profiles As mechanical recycling is expected to degrade macromolecular structure, we have followed the evolution of both moduli and viscosity as a function of the number of processing cycles. 3.1. Dynamic mechanical analysis (DMA) Dynamic mechanical analysis carried out at 1 Hz on injected samples of polycarbonate from 30 to 170  C show interesting features. In Fig. 1, the storage modulus (E0 ) curves, obtained in three-point bending mode, have been overlaid to emphasize the influence of recycling. Three zones can be identified: zone 1 corresponds to the solid state behavior, zone 2 to the end of rubber-like Fig. 1. Storage modulus (E 0 ) in bending mode versus temperature as a function of number of recycling (NR) in DMA. J.F. Feller, A. Bourmaud / Polymer Degradation and Stability 82 (2003) 99–104 plateau and zone 3 to the polymer flow. Zone 1 analysis provides the evolution of E0 with the number of recycling (NR) at 30  C, results are presented in Fig. 2. It can be seen that E 0 (30  C) increases from the first cycle (PC1) to the fifth cycle (PC5) and then remains constant up to eighth cycle (PC8). This phenomenon could be attributed to a decrease of molecular weight for two reasons: the first one is that Kavano et al. [5] observed the same phenomenon with three different polycarbonates. The second reason is that as it will be seen in the following, viscosity decreases with the number of recycling steps and it is well known that viscosity is related to molecular weight [6]. Moreover, these authors noticed that the decrease of the molecular weight was followed by an increase of the yield strength and a decease of the impact strength. Curves of E 0 versus NR obtained by additional experiments, in tension mode 101 with measurements in both transversal and longitudinal direction, have been overlaid to the previous one in bending mode (cf. Fig. 2). The same stiffening is observed in both directions and over the fourth recycling, E 0 in the longitudinal direction becomes much more important than in the transverse one. This suggests that macromolecular chains shortened by high shearing during injection can more easily align themselves during the processing, leading to an increase of anisotropy in the sample. Zone 2 of Fig. 1 features a phenomenon that could have something common with stress relaxation but which needs further investigation to be fully understood. Magnification of zone 3 analyzed in Fig. 3, shows a slight decrease of glass transition temperature (Tg) with NR, from 156.5 to 153  C. 3.2. Differential scanning calorimetry (DSC) Fig. 2. Evolution of E0 as a function of number of recycling (NR), for three points bending, longitudinal and transversal tension, at T=30  C and f=1 Hz. In Fig. 4, it can be seen that DSC experiments confirm DMA results: the glass transition temperature (Tg) of recycled PC decreases with the number of recycling (NR). Both technique show that Tg decreases of about 3.5  C from the first to the eighth cycle. Despite the fact that the two techniques show the same change of Tg versus NR, in Fig. 3, a more precise analysis shows that differences can be found. First, the curve obtained by DSC is shifted to lower temperatures, of about 7  C. This difference in Tg determination is not surprising since it is well known that in DMA relaxation temperatures depend on frequency (an agreement with DSC values would have been found for a lower frequency). Second, the curves do not have exactly the same shape, but considering the precision on Tg determination from an inflection point ( 1  C), this difference may not be significant. In DSC experiments, Tg decreases smoothly from the first to the fifth cycle and then decreases sharply over the sixth cycle. In DMA experiments, Tg Fig. 3. Tg and
Cp at Tg evolution of PC with NR determined by DSC and DMA. Fig. 4. DSC thermograms of PC showing Tg evolution with NR (PC2–PC8). 102 J.F. Feller, A. Bourmaud / Polymer Degradation and Stability 82 (2003) 99–104 decreases from the first to the fifth cycle and then stabilizes; this tendency recalls that of the storage modulus E 0 (30  C) already noticed in the previous paragraph or the decrease of
Cp amplitude at Tg with NR (Figs. 3 and 4). Both E0 increase and
Cp amplitude at Tg decrease [7,8] are consistent with a stiffening and probably a more brittle behavior [5] of the recycled PC. Such stiffening could come from a better organization of the amorphous phase resulting either from a higher mobility of shorter chains or the memory of orientation due to the process. The Tg decrease is generally interpreted as a decrease of molar mass increasing the number of chain ends and thus increasing the free volume. 3.3. Cone/plate rheometry Cone/plate rheometry has been chosen for it know sensitivity to molar mass variations which are suspected to result from the recycling process. Viscosity evolution with shear rate at 240  C, has been determined in steady state as a function of NR in Fig. 5, and the values obtained are in good agreement with the data of the literature [9,10]. It appears clearly that Z0, the Newtonian limit viscosity, decreases regularly (in logarithmic scale) : with NR. This change has been recalled for
¼ 0:1 s1 in Fig. 6, together with those of other experiments obtained for 260 and 280  C. This figure shows that the viscosity decrease due to recycling, is much more important at 240  C than at 280  C. In Fig. 7, G0 and G00 , respectively the storage and loss shear moduli have been plotted versus frequency from 1 to 1000 rad s1. These experiments in dynamic mode evidence a decrease of both moduli with NR, in the flow zone corresponding to low frequencies. Moreover, in Fig. 8, the G0 /G00 cross over frequency increases sharply with NR, which could Fig. 6. Viscosity (Z) decrease as a function of number of recycling : (NR) at
¼ 0:1s1 for 240, 260 and 280  C. Fig. 7. Evolution of Log G0 , Log G0 with Log ! as a function of NR at 240  C. : Fig. 5. Evolution of viscosity (Z) with shear rate (
) as a function of  number of recycling (NR) at 240 C. Fig. 8. Evolution of G0 , G0 cross over frequency as a function of NR at 240  C. J.F. Feller, A. Bourmaud / Polymer Degradation and Stability 82 (2003) 99–104 103 be related to an increase of chains length distribution [11], although Bafna [12] have recently underlined this is not always the case. Thus, rheometry in the liquid state provides valuable information on the different consequences of recycling on PC chain structure, i.e., evidence of a molar mass decrease and probably a polydispersity increase which confirm other results obtained with solid state characterization techniques. 3.4. Photodegradation The mechanisms involved into thermal and photodegradation have already been well studied [13–15] and show that PC is rather sensitive to thermal and photooxidization. Consequently, it is not surprising to notice chain scission phenomena, due to high shearing in the machines at high temperature under oxidant atmosphere (air). For that reason, many applications require the introduction of stabilizers into the PC formulation to prevent yellowing and embrittlement. It is interesting to determine not only the chain deterioration level but also the degradation of additives used for the polymer protection. Fig. 9 shows the evolution of surface degradation due to UV irradiation, followed by colorimetry. The L a b coordinates obtained for virgin PC change quickly from the beginning of the test to the fifteenth day and then stabilize. The expression of color difference for both Virgin and recycled PC (ground one time and extruded) are presented in Fig. 10. The increase of color difference is more important and the plateau reached earlier for the recycled PC suggesting that an important part of the stabilizers has been eliminated during the first recycling. This tendency will certainly be accentuated for the following cycles. From a mechanical point of view, Fig. 11 shows that the increase of the storage modulus E0 in bending mode (at 30  C and 1 Hz) follows the color difference increase presented in Fig. 10. Comparison of color difference for virgin and recycled PC as a function of UV irradiation time. Fig. 11. Evolution of bending modulus for PC injected and subsequently extruded one time as a function of UV irradiation time. Fig. 10. Nevertheless, it seems that in a third step, above 30 days of irradiation, the modulus decreases smoothly. Comparing the evolution of modulus with UV irradiation and the evolution of modulus with thermomechanical stress suggests that, in our experimental conditions, UV are more destructive than the recycling process. These results underline the necessity to stabilize the PC after each recycling before use. 4. Conclusion Fig. 9. Evolution of L, a, b, coordinates of a virgin PC Bayloy sheet as a function of UV irradiation time. The new approach brought by Eco-design tends to favor the use of engineering plastic wastes to decrease the impact on environment. We have investigated the degradation of poly(carbonate) generated by eight successive recycling using rheological, calorimetric and colorimetric tools at each step. Dynamic mechanical analysis shows that in three point bending mode E 0 (30  C) increases from the first 104 J.F. Feller, A. Bourmaud / Polymer Degradation and Stability 82 (2003) 99–104 cycle (PC1) to the fifth cycle (PC5) and then remains constant up to eighth cycle (PC8). Such phenomenon can be associated to molar mass decrease, certainly leading to an increase of yield strength and a decease of the impact strength). The same stiffening is observed in tension mode in both directions and over the fourth recycling, E 0 (30  C) in the longitudinal direction becomes much more important than in the transversal one. This suggests that macromolecular chains shortened by high shearing during injection can more easily align themselves during the process, leading to an increase of anisotropy in the sample. Differential calorimetric analysis shows that the glass transition temperature (Tg) of recycled PC decreases of about 3.5  C after eight recyclings and that
Cp amplitude at Tg also decreases with the number of recycling. These results confirm a stiffening of the material and probably a more brittle behavior. Cone/plate rheometry shows that Z0, the Newtonian limit viscosity, decreases regularly (in logarithmic scale) with NR and that this phenomenon increases with decreasing temperature (from 280 to 240  C) suggesting a molar mass decrease resulting from chain scission. Moreover, the G0 /G00 cross over frequency increases sharply with NR could signify an increase of chains length distribution. Colorimetric measurements after UV irradiation, realized to simulate aging due to sun exposure, show an increase of color difference more important and the plateau reached earlier for the recycled PC. This suggests that an important part of the stabilizers has been eliminated during the first recycling. All these results evidence a decrease of molar mass resulting from thermomecanical stress encountered by the PC during recycling. These modifications of polycarbonate properties, remaining good, must be taken into account to adjust the processing conditions and determine the possible applications. Acknowledgements The authors would like to thank A. Charpentier, E. Roussière, H. Béllégou, F. Peresse & J. C. Jégo for their contribution to this work. This project was supported by the French Ministry of National Education, Research & Technology (M.E.N.R.T.), Self Signal Society and National Agency for Research Valorization (A.N.VA.R.). References [1] Liu X, Bertilsson H. Recycling of ABS and ABS/PC blends. Journal of Applied Polymer Science 1999;74(3):510–5. [2] Silva Spinacé MA, De Paoli MA. Characterization of poly (ethylene terephtalate) after multiple processing cycles. Journal of Applied Polymer Science 2001;80(1):20–5. [3] Jacob C, Bhattacharya AK, Bhowmick AK, De PP, De SK. Recycling of ethylene propylene diene monomer (EPDM) wastes. III. Processability of EPDM rubber compound containing ground EPDM vulcanisates. Journal of Applied Polymer Science 2003;87:2204–15. [4] Matsumura S. Enzyme-catalyzed synthesis and chemical recycling of polyesters. Macromolecular Bioscience 2002;2(3):105–26. [5] Kavano Y, Keskkula H, Paul DR. Effect of polycarbonate molecular weight and processing conditions on mechanical behaviour of blends with a core-shell impact modifier. Polymer 1996; 37(20):4505–18. [6] Sperling LH. Introduction to physical polymer science. UK: Wiley & Sons; 1986. [7] Wunderlich B. Thermal characterization of polymeric materials: The basis of thermal analysis, 2nd ed. In: Turi EA. New-York: Academic Press; 1997. p. 385. [8] Feller JF, Linossier I, Pimbert S, Levesque G. Carbon black filled poly(ethylene-co-acrylates) composites: calorimetric studies. Journal of Applied Polymer Science 2001;79(5):779–93. [9] Carneiro OS, Covas JA, Bernardo CA, Caldeira G, Van Hattum FWJ, Ting JM, Alig RL, Lake ML. Production and assessment of polycarbonate composites reinforced with vapour-grown carbon fibres. Composites Science & Technology 1998;58(3-4):401–7. [10] Marin N, Favis BD. Co-continuous morphology development in partially miscible PMMA/PC blends. Polymer 2002;43(17):4723– 31. [11] Cruz SA, Zanin M. Evaluation and identification of degradative processes in post-consumer recycled high-density polyethylene. Polymer Degradation and Stability 2003;80(1):31–7. [12] Bafna SS. Is the cross-over modulus a reliable measure of polymeric polydispersity? Journal of Applied Polymer Science 1997; 63(1):111–3. [13] Montaudo G, Carroccio S, Puglisi C. Thermal oxidation of poly(bisphenol A carbonate) investigated by SEC/MALDI. Polymer Degradation and Stability 2002;77(1):137–46. [14] Rivaton A. Recent advances in bisphenol-A polycarbonate photodegradation. Polymer Degradation & Stability 1995;49(1):163–79. [15] Rivaton A, Mailhot B, Soulestin J, Varghese H, Gardette JL. Comparison of the photochemical and thermal degradation of bisphenol-A polycarbonate and trimethylcyclohexane-polycarbonate. Polymer Degradation and Stability 2002;75(1):17–33.