Recycling of high impact polystyrene in coextruded sheet: influence of the number of processing cycles on the microstructure and macroscopic properties.
High impact polystyrene (HIPS) is one of the most widely used thermoplastics, especially in those products where brittle polystyrene (PS) can not be used. HIPS can be structurally defined as a multiphase system that exhibits a continuous rigid PS phase together with disperse rubber particles, which vary from 0.5 to 10 [micro]m in diameter. The rubbers commonly used in the HIPS synthesis are polybutadiene or styrene/butadiene copolymers. The synthesis conditions generate rubber particles that are mainly crosslinked and contain grafted and free PS occluded within the particles. Thus, the synthesis process determines the particle's inner structure, which can vary considerably from one set of synthesis conditions to another .
There are numerous studies in the literature related to the HIPS mechanical properties and all of them suggest that the dispersed rubber phase greatly contributes to the high performance of HIPS. The influence of different parameters like the rubber content (RC), the volume fraction of the disperse rubber phase ([PHI]), the crosslinking density of the rubber phase, and the molecular weight distribution of the PS matrix have been evaluated on the impact strength in HIPS by Wagner and Robeson  who established that both the elongation and the impact strength increased with [PHI]. The Young's modulus increased with an increase in the crosslinking density of the rubber phase, whereas the elongation and impact strength decreased with this last parameter. Moreover, the swelling index (SI) of the rubber phase (associated with the rubber phase crosslinking density) had an optimum value in the interval of 10-14, in order to prevent the coalescence of PS occluded in the rubber particles during melt processing . The importance of the crosslinking density of the rubber phase was demonstrated by Bucknall et al. , who reported that the tensile strength increased and the impact strength decreased drastically with the crosslinking density.
Taking into account that the MFI is a very important parameter that describes the flow characteristics of thermoplastics, in particular HIPS, Nikitin et al.  reported the influence of the molecular characteristics such as [bar.M.sub.w] of the PS matrix and parameters of the rubber phase like RC, GD, GC, the rubber particle diameter (d) and distribution ([chi]), on the MFI. These authors indicated that the MFI increased with a decrease in [bar.M.sub.w] and depended on [chi] and GD in a complex way.
All the studies mentioned above involved virgin HIPS. Only few papers, like the one reported by Kalfoglou and Chaffey , paid attention to the effects of repetitive extrusion processing and extrusion at high temperature on HIPS characteristics and properties. In this sense, Kalfoglou reported that the rheological and tensile properties changed significantly while the elongation and the impact resistance changed only at high processing temperature (290[degrees]C), decreasing to 57 and 29%, respectively. These changes were attributed to the breakage of the interactions between the matrix and the rubber particles as well as to the rubber degradation, which is a consequence of the number of extrusion cycles and/or to the high temperature employed during processing.
On this respect, Szabados and Pukanszky  reported the variation of the solution viscosity of virgin and recycled HIPS samples versus time of processing--of these two samples--in a two roll mill. The solution viscosity of the virgin HIPS increased markedly with the time of milling, reaching a maximum after ca 25 min (due to the high crosslinking level in the rubber phase, which in turn is due to the strong thermo-mechanical treatment) and then, as the milling continued, the solution viscosity decreased down to approximately the same initial value after ca 60 min of milling, remaining approximately constant up to 150 min of processing. On the other hand, the solution viscosity of the recycled HIPS remained practically constant throughout the 150 min of milling.
The present study focuses on the degradative effect of coextrusion processing on the physical and chemical properties of commercial HIPS (Basf Polystyrol 2710, specially used in the manufacture of refrigerator parts via thermoforming of coextruded sheet).
During the thermoforming process, it is common that 25-30% of the original laminate goes to scrap, which then becomes material for recycling. This recycled HIPS is used in a composition of 30 wt% in the formulation of coextruded sheet along with 70 wt% of virgin HIPS. Due to the importance of the elastomeric phase on the HIPS properties, GC, GD, SI, morphology of the rubber phase and [bar.M.sub.w] of the PS matrix were characterized after each processing cycle. The rheological and performance properties of the produced HIPS were evaluated and correlated with the physical-chemical properties.
Two HIPS, BASF 2710, and Avantra 585K, both from BASF, were used in this study, and their characteristics are summarized in Table 1.
Coextrusion of HIPS sheet was performed in an industrial plant using an AMO equipment having three extruders (Fig. 1) under the processing conditions shown in Table 2. The extrusion conditions were the same for all processing cycles with an increasing temperature profile along the screw (175-220[degrees]C) and a constant die temperature of 210[degrees]C.
In the case of the sheet coextruded after six cycles, the sheet composition was 70% [R.sub.1], 21% [R.sub.2], 6.3% [R.sub.3], 1.89% [R.sub.4], 0.56% [R.sub.5] and 0.25% [R.sub.6], where [R.sub.i] indicates the extrusion cycle. The coextruded sheet was formed by a layer (60 wt%) which consists of a blend of 50/50 of virgin and recycled HIPS, sandwiched between two layers (19 wt% each) of virgin HIPS, plus a thin layer of Avantra 585K HIPS (2 wt%) on one side, for the sole purpose of obtaining a glossy finish, as is shown in Fig. 1.
Characterization of the Coextruded HIPS
The products of each cycle were recycled and processed by injection molding at constant conditions (Table 3) to obtain standard specimens to be evaluated with respect to their physical-chemical, rheological, and mechanical properties. All the measurements were repeated three times in the case of the physico-chemical properties and seven times in the case of the mechanical and rheological properties, and the standard deviation (SD) was calculated in all cases.
[FIGURE 1 OMITTED]
Determination of Physical-Chemical Properties
Gel Content. The gel (insoluble) fraction was isolated from the soluble fraction (free PS) dissolving 1 g of HIPS sample in 25 ml of toluene using the centrifugation technique for 45 min at 20,000 rpm at -20[degrees]C. The soluble fraction was precipitated with methanol and both fractions were dried under vacuum at 50[degrees]C to constant weight and gravimetrically calculated as mentioned in Ref. .
Grafting Degree. Grafting degree (GD) was determined according to Eq. 1 reported by Gasperowicz and Laskawski .
GD = ([P[S.sub.graft]]/[[P.sub.calculated]PB]) X 100 (1)
where P[S.sub.graft] = amount of PS chemically and physically bonded to the PB, and [P.sub.calculated] PB = percent of polibutadiene (PB) determined by infrared spectroscopy = 8.3 wt%
Swell Index. The swell index (SI) was determined placing the dry gel, previously determined as described for GC determination, into 50 ml of toluene during 24 h until saturation. The swollen gel was separated from the supernatant and weighed. The SI was then calculated as the ratio of the wet (swollen) gel to the dry gel.
Morphology. Internal morphology and particle size were analyzed from injection molded specimens by TEM from ultramicrotome-sectioned samples, previously stained with Os[O.sub.4].
Molecular Weight of PS Matrix. The molecular weight of the PS matrix after each processing cycle was determined by size exclusion chromatography (SEC), using a Hewlett Packard chromatograph fit with a set of ultrastyrogel column (of nominal pore [10.sup.5], [10.sup.4], and [10.sup.3] [Angstrom]) using THF as a solvent and at room temperature.
Determination of Rheological Properties
Melt Flow Index. The melt flow index (MFI) determinations were carried out in a Tinius Olsen plastometer according to standard ASTM 435, at 200[degrees]C and at a constant weight of 5 kg.
Capillar Rheometry. An Instron 2700 capillary rheometer fitted with a die with an L/D ratio of 30 and a radius of 0.5 mm was used to determine the melt viscosity under the following test conditions: temperature = 210[degrees]C and a shear rate interval from 6 to 5000 [s.sup.-1].
Dynamic Mechanical Analysis. Dynamic mechanical analysis (DMA) was performed in flexure mode, using a TA Instruments DMA 983 equipment, at a temperature interval from -140 to 110[degrees]C, using a frequency of 0.1 Hz, an amplitude of 0.5 mm, and a heating rate of 5[degrees]C/min.
Determination of Performance Properties
Tensile Stress at Break and Izod Impact Strength. The tensile stress at break tests were carried out according to ASTM D-638 on a Universal Instron machine at a constant cross-head speed of 5 mm/min, at 25[degrees]C, and 50% relative humidity (RH) where the specimens were first conditioned at 25[degrees]C for 24 h.
The Izod impact strength was determined according to ASTM D 256 using a CSI 137 equipment.
RESULTS AND DISCUSSIONS
Figure 2a and 2b show respectively the GC and GD variation as a function of the number of processing cycles.
The gel content (GC) represents the insoluble fraction of the disperse rubber phase and takes into account the initial RC plus the grafted and occluded PS. It can be observed in Fig. 2a that it slightly decreases or remains constant with increasing number of processing cycles, except for the case where the HIPS has been subjected to only one processing cycle, in which case, it reaches a GC value of 26%. The slight increase in GC after the sample has been subjected to the first processing cycle can be attributed, as mentioned by Stzabados and Pukanszky , to crosslinking and/or grafting reactions in the elastomeric phase, precisely during processing. Grafting reactions cause an increase in the total mass of the rubber phase (GC).
[FIGURE 2 OMITTED]
In subsequent processing cycles, GC decreases due to particle fragmentation, caused by the shear stress and the high temperature during processing. This induces the occluded PS to migrate to the continuous PS phase. And then, it can be extracted along with the soluble fraction during centrifugation resulting in the observed GC decrease. These changes can be observed in Fig. 3, where for virgin HIPS (Fig. 3a), the rubber phase particle is of the salami type and has a particle size in the order of 5 [micro]m, with occlusions in the range of 0.5-2 [micro]m in size. However, after processing, (particularly after the second and the following processing cycles), the occluded PS within the rubber particles becomes exposed to the PS matrix because the wall of the rubber particle was broken and the occlusions became smaller (Fig. 3b). This behavior could be corroborated with the results obtained by the SI values, which increase up to 16 (Fig. 4) with the number of processing cycles, suggesting a degradation of the crosslinked three-dimensional network (and then the particle fragmentation ocurred) that is in accordance with Wagner and Robeson's  report. This degradation is caused by the thermo-mechanical treatment during extrusion, as reported by Szabados and Pukanszky  and Yenalyev et al. .
On the other hand, the decrease in SI from R0 to R1 and R2 from a value of 12 to a value of 10 (Fig. 4) suggests an increase in the crosslinking density along the first two processing cycles, which can be caused by crosslinking reactions between the double bonds of the rubber exposed to high temperature during the processing.
Figure 5 shows how the weight average molecular weight of the PS matrix ([bar.M.sub.w]), decreases with the number of processing cycles. This is indicative of the polymer chain scission due to the thermo-mechanical treatment during processing. This behavior and the strong alteration of the rubber phase, observed and described above, are responsible for the changes that occurred in the rheological properties, as it is demonstrated in the following section.
[FIGURE 3 OMITTED]
With respect to the rheological properties, Fig. 6 shows the variation of the MFI as a function of the number of processing cycles, where it is observed that the MFI initially decreases, after the first processing cycle, from 2.88 g/10 min (virgin HIPS) to a value of 1.7 g/10 min, but then, gradually increases after the second processing cycle up to 3.4 [+ or -] 0.2 g/10 min.
[FIGURE 4 OMITTED]
It must be taken into account that the flow properties of heterogeneous systems like HIPS depend on several factors such as [bar.M.sub.w] and polidispersity index (I) of the PS matrix, temperature, as well as volume fraction, and GD of the rubber phase.
In this particular case, the I of the PS matrix remains without change from R0 to R1 (3.3 and 3.4, respectively), and the [bar.M.sub.w] decreases only by 7% so that the variations after the first processing cycle can be attributed mainly to changes in the rubber phase, that is, to the increase in the crosslinking and grafting reactions that occur during processing, which yield a more rigid particle and the increase in the GD that produces an increase in the GC that in turn make the MFI to decrease. This result is in accordance with that reported by Nikitin et al. , where the MFI decreases with increasing ratio between the GC and the RC.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
To explain the flow properties' behavior during the first processing cycle, the behavior of the viscosity at low shear rate can be considered. In this sense Eq. 2 can be used to estimate the HIPS viscosity [[eta].sub.0,HIPS] at low shear rate [10,11]
[[eta].sub.0,HIPS] = [[eta].sub.0,PS] X [2.5[PHI]/[e.sup.[1-[PHI]/0.68]]] (2)
where [[eta].sub.0,PS] is the zero shear viscosity of the PS matrix and [PHI] is the volume fraction of the disperse rubber phase (considering both, the grafted and occluded PS). Equation 2 takes into account two main contributions: one from the continuos PS matrix and another one from the dispersed rubber phase. In this respect, Kruse and Southern  established the following relationship to estimate [[eta].sub.0,PS] (Eq. 3)
ln[[eta].sub.0,PS] = -20.95 + 3.4 ln([M.sub.w]/33,000) + 11,000/T (3)
where [bar.M.sub.wPS] is the weight average molecular weight of the PS matrix (in g/mol) and T is the temperature (in K). In this way, the HIPS viscosity is influenced by two properties: [bar.M.sub.wPS] and [PHI]. Any Increase in [bar.M.sub.wPS] and/or [PHI], will increase the viscosity and as a consequence the MFI will decrease.
If a decrease in [bar.M.sub.wPS] is accompanied by an increase in [PHI] or vice versa, the variation in MFI will depend on the relative contribution of each parameter. For example, if the once processed virgin and processed HIPS are compared (A and B, respectively); and if it happens that [bar.M.sub.wPS(A)] = 211,000 g/mol, [bar.M.sub.wPS(B)] = 200,000 g/mol, and [[PHI].sub.(A)] = 0.22, and [[PHI].sub.(B)] = 0.26, it can be observed that, by applying Eqs. 2 and 3, [[eta].sub.0,HIPS(A)] = 1.24 X [10.sup.3] Pa s and [[eta].sub.0,HIPS(B)] = 1.31 X [10.sup.3] Pa s. In this case, therefore, an increase in [PHI] determines an increase in viscosity and a decrease in MFI.
[FIGURE 7 OMITTED]
On the other hand, after the second and up to the sixth processing cycle, the MFI increases due to a decrease in GD of the rubber phase as well as to a decrease in the molecular weight of the PS matrix.
Figure 7 shows that the viscosity of the material subjected to one processing cycle is higher than that of the virgin material, especially at low shear rates, but as shear rate increases, all curves tend to concur to a similar value and this difference in viscosities becomes very small.
[FIGURE 8 OMITTED]
This tendency suggests once again a change in the molecular structure during the first processing cycle followed by bond scission during the following cycles. Since there was not significant changes in the viscosity values evaluated in the molten state, the power law index--a more sensitive response and calculated through a linear regression analysis from the flow curves--was plotted as a function of the number of processing cycles (Fig. 8). It must be mentioned that the regression analysis of each flow curve has a correlation factor [R.sup.2] = 0.99.
The nature of the molten plastic can be inferred from the n values. All n values resulted in values less than 1, characteristic of pseudoplastic materials. For virgin HIPS n is equal to 0.311, then it decreases after the first processing cycle to 0.29, indicating an increase in the pseudoplastic behavior due to an increase in the GD--and then it increases gradually up to a value of 0.35, after the sixth processing cycle--evidencing a lesser pseudoplastic character with respect to the virgin and the once processed HIPS as a consequence of the decrease in GD values.
The behaviors described above can be also corroborated with DMA (Fig. 9) where the storage modulus (E') of the virgin HIPS is 17,500 GPa, whereas for R1 it increases up to 20,000 GPa. This increase after the first processing cycle is again an evidence of crosslinking reactions that occur due to the thermo-mechanical treatment that increases the rigidity of the rubber phase and in consequence, the rigidity of the composite material. On the other hand, the tan [delta] peak that corresponds to the Tg of the rubber phase (Fig. 10) increases from -96 to -77[degrees]C. This displacement in tan [delta] is associated with an increase in the GD and crosslinking density of the elastomeric phase, which diminishes the flexibility of the rubber phase [2, 12].
Figure 11 shows that the izod impact strength decreases slightly after the first processing cycle (from 80 to 69 J/m) due to a decrease in the flexibility of the particle caused by an increase in the crosslinking density of the rubber phase. On the other hand, the storage modulus (Fig. 9) and the tensile stress at break (Fig. 12) both increase and the impact strength decreases as the GD increases.
[FIGURE 9 OMITTED]
However, from the second to the last processing cycle, the impact strength increases continuously up to 96.19 J/m due to an increase in the SI that indicates a decrease in the crosslinking density in the rubber phase.
After the last processing cycle, the results can be interpreted as the concurrency of two main factors: the degradation of the rubber phase structure and the decrease of the [bar.M.sub.w] of the PS matrix caused by the shear stress and high temperature used in the processing. It must be observed that the decrease in the [bar.M.sub.w] has no effect on the tensile stress at break.
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
The variations in the physical and chemical properties like the GC, the GD, the SI, and the [bar.M.sub.w] of the PS matrix confirm a change in the microstructure of the HIPS caused by the repetitive coextrusion processing, especially after the first processing cycle. This change is principally attributed to a degradation mechanism that takes place in the rubber phase as well as in the PS matrix, influencing the mechanical and rheological properties of HIPS.
[FIGURE 12 OMITTED]
The authors thank MABE Company, from Queretaro, Mexico for their support and for allowing the use of their equipment and to Pablo Acuna for his technical support.
HIPS high impact polystyrene PS polystyrene PB polibutadiene GC gel content GD grafting degree SI swell index [bar.M.sub.w] average molecular weight of the PS matrix MFI melt flow index [eta] shear viscosity N power law index [T.sub.g] glass transition temperature I polidispersity index RH relative humidity DMA dynamic mechanical analysis SD standard deviation
1. F. Haaf, H. Breuer, A. Echte, B.J. Schmitt, and J. Stabenow, J. Sci. Ind. Res., 40, 659 (1981).
2. E.R. Wagner and L.M. Robeson, Rubber Chem. Technol., 43, 1129 (1970).
3. C.B. Bucknall, H.H. Yang, and X.C. Xiang, "Effects of Rubber Cross-Link Density on Deformation Kinetics in HIPS," in The Third Conference Proceedings on Toughening of Plastics, Polymat'94, London (1994).
4. Y.V. Nikitin, L.M. Aleksandrova, and S.L. Moskovskii, Int. Polym. Sci. Technol., 9(11), 58, (1982).
5. N.K. Kalfoglou and C.E. Chaffey, Polym. Eng. Sci., 19(8), 552 (1979).
6. T. Szabados and B. Pukanszky, Int. Polym. Sci. Technol., 4(11), 75 (1977).
7. A. Gasperowicz and W. Laskawski, J. Polym. Sci. Part A: Polym. Chem., 14, 2875 (1976).
8. V.D. Yenalyev, V.I. Melnichenko, O.P. Boukunenko, A. Shelest, N.M. Tchalaya, Y.I. Yegorova, and N.G. Podosyonova, Polym. Sci. Technol., 20, 19 (1983).
9. T. Brenmer, A. Rudin, and G. Cook, J. Appl. Polym. Sci., 41, 1617 (1990).
10. R. Kruse and J. Southern, J. Rheol., 24, 755 (1980).
11. C. Luciani, D. Estenoz, H. Oliva, and G. Meira, Ind. Eng. Chem. Res., 44, 8354 (2005).
12. J.H. Choi, K.H. Ahn, and S.Y. Kim, Polymer, 41, 5229 (2000).
Florentino Soriano, Graciela Morales, Ramon Diaz de Leon
Centro de Investigacion en Quimica Aplicada (CIQA), Blvd. Enrique Reyna No. 140, A. Postal 379. Saltillo, Coahuila, Mexico
*Presented at the National Congress of the Sociedad Polimerica de Mexico. Chihuahua, Chihuahua, Mexico, November, 2004.
Correspondence to: Florentino Soriano; e-mail: firstname.lastname@example.org
TABLE 1. Characteristics of HIPS used in the coextruded sheets. Property BASF 2710 AVANTRA 585K Rubber (wt%) 8.5 16.0 Gel content (wt%) 22 10 Swell index (SI) 12 20 [bar.M.sub.w] (g/mol) 205,000 179,000 MFI (10 g/min) 2.84 3.51 Izod Impact (J/m) 80 106 Tensile strength (kg/[cm.sup.2]) 208 226 [T.sub.g] PB (DMA) ([degrees]C) -94 -92 E' (DMA) (GPa) 15,729 18,140 [bar.M.sub.w], number-average molecular weight; MFI, melt flow index; [T.sub.g], glass transition temperature of the rubber phase; E', storage modulus. TABLE 2. Temperature profile in the coextrusion system. Extruder Temperature Extruder Temperature Extruder Temperature A ([degrees]C) B ([degrees]C) C ([degrees]C) 1 175 1 187 1 167 2 188 2 179 2 167 3 197 3 195 3 166 4 204 4 190 4 167 5 205 5 195 5 168 6 202 6 197 6 168 7 202 7 205 7 168 8 202 8 OFF 8 168 9 OFF 9 200 - - 10 200 10 202 - - 11 205 11 203 - - 12 207 12 201 - - 13 200 13 204 - - 14 200 14 220 - - 15 223 - - - - TABLE 3. HIPS injection molding conditions. Injection parameter Process condition Injection temperature ([degrees]C) 240 Temperature profile ([degrees]C) 240/220/200 Injection pressure (bar) 130 Maintenance pressure (bar) 130/110/1100 Injection rate (mm/s) 50 Injection time (s) 3 Cooling time (s) 18
|Printer friendly Cite/link Email Feedback|
|Author:||Soriano, Florentino; Morales, Graciela; de Leon, Ramon Diaz|
|Publication:||Polymer Engineering and Science|
|Date:||Dec 1, 2006|
|Previous Article:||Investigation of melt manipulation phenomena during injection molding via in situ birefringence observation.|
|Next Article:||The effect of screw geometry on melt temperature profile in single screw extrusion.|