Melt rheological behavior of starch-based matrix composites reinforced with short sisal fibers.
In recent years, natural fiber-reinforced thermoplastic composites have found wide use in structural and nonstructural applications exploiting several advantages over more traditional glass fibers (1-7): biodegradability, availability of renewable resources and subproducts of other industrial sectors, low density, medium to low costs, low abrasion, etc. (it must be conceded, however, that they offer also some severe disadvantages, lower durability and processing temperatures being the main drawbacks). The use of biodegradable matrices allows the development of a new class of fully biodegradable composites. Biodegradable polymers constitute a new family of polymers designed to be degraded by living organisms. They offer a possible alternative to traditional nondegradable polymers when recycling is impractical or not economical, and they can be composted together with food and yard waste (8, 9). The design of the most suitable processing conditions is guided mainly by the rheological behavior of the composites (10). A number of investigations on the rheological behavior of short natural fiber-reinforced thermoplastic and elastomers have been reported (11-13): incorporation of fillers in thermoplastics and elastomers will increase the melt viscosity, which may result in unusual rheological effects.
Suspensions with short fibers in a polymer matrix may show a complex behavior characterized by one or more of the following phenomena (14):
* Viscosity increase as fiber concentration and/or aspect ratio increase (15, 16).
* Shear rate at which the shear thinning behavior of the suspending fluid starts can be reduced by the presence of fibers (15-17).
* Stress overshoot phenomena observed in transient flows of suspensions of fiber increases as fiber concentration or its aspect ratio increase.
* Rheological properties of concentrate suspensions are highly shear dependent (15, 16).
* Bulk rheological behavior of fiber suspensions is sensitive to fiber length distribution and fiber orientation (i.e. at higher fiber length, they are difficult to orient in the direction of flow, and the viscosity increases) (18).
* Fiber breakage in natural fiber occurs in length and diameter, and results in a redistribution of fiber aspect ratio. Fiber breakage can lead to unexpected rheological properties of fiber suspensions (19).
Whereas processability is a main advantage of short-fiber composites, the mechanical properties of the final product are strongly related to the fiber distribution and orientation, which are determined during processing. Moreover, these materials are generally processed with high-speed injection techniques where the viscoelasticity of the compound is strongly dominated by its transient behavior. The extent to which these composites can be used is currently limited to a large degree by our understanding of their rheological behavior and by our ability to exploit their rheological properties in the design and implementation of processing operations. The possibility to predict and control the mold filling process and the distribution of fiber orientations in the final product depends on the rheological description of these suspensions and on the ability to simulate numerically their flows (20).
Crowson et al. (21) reported in detail the flow behavior of short glass fibers reinforced thermoplastic during injection molding. Composite viscosity is considerably influenced by fiber content and fiber length, and this effect is more evident at low shear rates than at high shear rates.
Barbosa et al. (22) have reported a comprehensive statistical study of the complex injection molding behavior of short glass fiber composites in which the effects of the rheological behavior of the compounds on the final structure and properties of the composites were highlighted.
The rheological behavior of short jute fiber composites was reported by Murthy et al. (23). Melt flow behaviors of short sisal fiber in different polymeric systems have been studied by Joseph et al. (24, 25). They studied viscoelastic properties of low-density polyethylene reinforced with short pineapple fibers as a function of shear rate at different fiber orientations. It was found that longitudinally oriented fiber composites have the maximum value of storage modulus. Prasantha Kumar et al. (26) determined the morphology and melt rheological behavior of short-sisal reinforced SBR composites. It was found that the composites behave as pseudo-plastic materials. Different treatments of fibers were done and the treated fibers have strong interfacial adhesion between the fiber and rubber matrix. As a consequence of that, the viscosity of the treated fiber suspension increased.
The viscous behavior of molten starch is known to depend on temperature, moisture content, and thermomechanical treatment (27-30). Viscosity data and rheological models that take into account these variables have been reported (31-35). They all report a thermoplastic behavior of low hydrated starch within a two orders of magnitude shear-rate range, with an Arrhenius dependence on temperature and similar for moisture content. They showed that transformation-like gelatinization and retrogradation affect the viscoelastic behavior of molten starch.
The aim of this work is to study the effect of sisal short-fiber content, processing method, and fiber treatment on the rheological behavior of starch-based matrix composites in dynamic and steady shear conditions.
Equations that predict the shape of flow curves representing the steady viscosity of thermoplastics need at least four parameters. In common use are the Cross model (36), Carreau model (37), and power-law model, which can be obtained by a simple redefinition of parameters:
[eta] = [K.sub.2][dot.[gamma].sup.n-1] (1)
where n is called the power-law index and [K.sub.2] is called flow consistency. This two-parameter model is the most-used rheological equation for polymers, but it must be used carefully, in reduced shear rate intervals, as it can introduce very high inaccuracy into rheological results, out of the measurement range. Moreover, these rheological models represent the viscosity changes due to shear thinning, and do not take into account elastic behavior, thermal and stress degradation, phase segregation or melt rupture.
In order to analyze the relation between the dynamic and the steady shear data, [eta]* versus [omega] and [eta] versus [dot.[gamma]] can be superimposed, and the results can be graphically represented. The exact relationship in the lower limits of frequency and shear rate is:
[eta]'([omega])[.sub.[omega][right arrow]0] = [eta]([dot.[gamma]])[.sub.[gamma][right arrow]0] (2)
The previous equation states that the viscosity measured in oscillatory shear in the zero frequency limits is equal to the low shear viscosity measured in the steady shear in the zero shear rate limits. An empirical relation-ship between [eta] and [eta]' at other than the low limits of shear rates and frequency has been developed by Cox-Merz for polystyrene, which proposes that [eta] should be the same function of [dot.[gamma]] as |[eta]*| is of [omega], where |[eta]*| is the modulus of the complex viscosity:
|[eta]*| = [([eta]')[.sup.2] + (G'/[omega])[.sup.2]][.sup.1/2] (3)
In addition, such a superposition can be justified by Bueche's theory stating that a macromolecule can be approximated as rotating inside an envelope of angular velocity of half the shear rate ([eta] = [eta]* where [dot.[gamma]] = 2.[omega]).
A commercial biodegradable thermoplastic compound named MaterBi-Y, kindly supplied by Novamont (Novara, Italy), was used as matrix in this research. This material is commercialized as a starch-based polymer blend. In previous research (38) we reported a degradation study of MaterBi-Y and its composites and we reported a possible composition compatible with TGA results: 38% starch, 38% cellulose derivatives; 3% water, and 21% additives.
Sisal fibers were obtained from Brascorda (Bayeux-Paraiba, Brazil). Fiber length and diameter were measured with an optical microscope on over 100 fibers. The average diameter was 300 [micro]m [+ or -] 50 and the average length was 7.2 mm [+ or -] 0.6.
Fibers were treated with Na(OH) aqueous solution (5% p/v) for 24h with continuous stirring at 25[degrees]C. After washing and drying, the matrix and fibers were mixed in an intensive mixer for 6 min at 180[degrees]C and then compressed in a hot press at 180[degrees]C and 700 MPa. Compressed plates (3 mm thickness) were annealed at 60[degrees]C in order to relieve thermal stresses produced during cooling.
Injection and compression molding were used to analyze the effects of the processing technique on the properties of the matrix and its sisal fiber composites. For the case of injection molding, matrix and fibers were fed directly into a 60-ton injection machine (Sandretto, Collegno, Italy).
After processing, fibers were extracted from the matrix, and diameter and length were measured by optical microscopy.
Measurements of rheological properties were performed using a Rheometric Scientific ARES N2 with parallel-plate geometry. Tests were carried out in steady rate and dynamic frequency modes at three temperatures: 180[degrees]C, 190[degrees]C and 200[degrees]C. The steady rate tests were realized in a range of rates from 0.1 to 10 [s.sup.-1]. Dynamic shear properties were determined as a function of angular shear rate of 0.1 to 500 rad/sec. For all experiments, the strain amplitude was maintained constant at 0.5%.
RESULTS AND DISCUSSION
Viscosity of Matrix and Composites
Before starting dynamic frequency sweep tests, the linear viscoelastic range was determined through a strain sweep test. G' (storage modulus) and G" (loss modulus) were registered as functions of deformation ([gamma]). Figure 1 shows this curve for the injected starch based matrix and for its composite with 15 wt% of sisal fibers. It is clearly observed that G' and G" remain constant until 1% of deformation, showing the linear viscoelastic range. It appears that this range is mainly controlled by the flow behavior of the matrix and it is not strongly affected by the fibers. Following these results, we selected a deformation of 0.5% for all tests. G' and G" are higher for the composites than the matrix, showing, as expected, that the stiffness increases.
[FIGURE 1 OMITTED]
The variations of complex viscosity ([eta]*) and storage modulus (G') with angular frequency for the injected starch-based matrix and its composites with sisal short fibers are shown in Fig. 2 (for all the experimental temperatures) and Fig. 3 respectively. The temperature used for Fig. 3 was 180[degrees]C, but a similar behavior was observed at all temperatures.
Figure 2 shows that the viscosity of the material decreases with frequency and increases with fiber content. The increase of viscosity and storage modulus with fiber content is not linear, showing a saturation effect at higher fiber concentrations.
Effects of the Processing Technique
Figure 4 shows the variation of complex viscosity as a function of frequency and temperature for matrix and composites processed by injection molding and by compression molding. From these figures it is possible to observe that the processing technique has almost no influence on the neat matrix properties, while, for composites, compression-molded samples show higher viscosity than injected ones. This result suggests that it is more difficult to process compression-molded material than injected material. The difference in viscosity is more evident at low temperature, and may be due to the high viscosity of the matrix, which retains the different orientations of the fibers. Another effect that should be taken into account is that the specimen testing process in parallel plate will introduce an inherent randomness on a macroscopic level.
The distributions of fiber length and diameter are shown in Figs. 5 and 6 respectively. It can be deduced that the main difference in dimensional parameters is related to the final fiber diameter. It appears that the intensive mixing used before compression molding is more efficient in separating the bundles of fibers, leading to a lower average diameter and higher aspect ratio. The final aspect ratios obtained for compression-molded and injection-molded specimens are in agreement with the viscosity differences obtained: 17.9 [+ or -] 3.0 for compression-molded samples and 16.2 [+ or -] 2.2 for injected samples. The results suggest that the aspect ratio effect is more important in the rheological characterization than fiber orientation.
Effects of the Fiber Treatment
Alkaline treatment is one of the most commonly used chemical treatments on natural fibers. Joseph et al. (39) reported the effect of chemical treatment on the tensile properties of short-sisal-fiber reinforced-polyethylene (PE) composites (both randomly and unidirectionally oriented) and analyzed the mechanisms of different treatment methods. This treatment was also used by our group previously for sisal fibers (40).
[FIGURE 2 OMITTED]
Figure 7 shows the variation of complex viscosity as a function of frequency for composites with 15 wt% sisal fibers with and without NaOH treatment, both processed by compression molding. Although a similar behavior can be observed, treated fiber composites show higher viscosity values for all studied temperatures. The aspect ratio (length/diameter) was 17.9 [+ or -] 3.0 for untreated fibers, as reported before, and 19.5 [+ or -] 2.2 for treated fibers (Fig. 8), in agreement with the viscosity differences observed. The same tendency was exhibited for all fiber contents. The observed effects of the fiber treatment can be attributed to the fiber fragmentation (40), because the modulus of the composites was not improved (41). Again, the viscosities of the composites with treated fibers were found to be higher than corresponding values of untreated samples at all shear rates studied, which suggests the need for more rigorous processing conditions.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
The superposition of dynamic and steady rheological tests results for the starch-based matrix and its composites is reported in Fig. 9 showing the applicability of Bueche's theory ([eta] = [eta]* where [dot.[gamma]] = 2.[omega]). These results allow the possibility to obtain models and parameters from dynamic curves instead of steady curves. If high fiber-fiber interaction is present in the composites, the Cox-Merz rule fails (42). As consequence, it is possible to say that the fiber concentration is low and no fiber-fiber interaction or weak interaction was observed.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
A power-law equation can be used for only the linear part of the viscosity curve. From this model we can obtained [K.sub.2] and n. The correlation between experimental data and the power-law model is also shown in Fig. 9. A similar general behavior was obtained for all the materials studied; Table 1 shows the model parameters for the three materials analyzed: injection-molded, compression-molded, compression-molded/treated fibers. The results reflect the differences already reported and attributed to the different aspect ratio of the fibers in the composites. In general, the values of n increase with temperature but decrease with fiber content, indicating that composites have a more pseudoplastic character than the pure matrix. On the other hand, [K.sub.2] decreases with temperature and increases with the incorporation of fibers.
[FIGURE 9 OMITTED]
As the power-law model is not able to predict the entire rheological curve, the applicability of the Cross model was tested:
[eta] = [[eta].sub.0]/[1 + ([tau].[omega])[.sup.b]] (4)
where [[eta].sub.0] is the asymptotic value of viscosity at very low shear rates, [tau] is the relaxation time, and b is a dimensionless constant. Normally, the zero-shear viscosity is related to the molecular structure of the polymer in terms of the possibilities of intermolecular interactions (43). For composite systems the changes in the zero shear rate viscosity are also affected by the interaction between matrix and fibers (44).
The correlation between the Cross model and experimental data are shown in Fig. 10. From this figure, it is possible to see that model prediction and experimental data are in good agreement. Obtained parameters are shown in Table 2.
The temperature dependence of the viscosity can be described in the range of studied temperatures by an exponential-type equation:
[eta] = A. exp ([E.sub.a]/R.T) (5)
[FIGURE 10 OMITTED]
where [E.sub.a] is an activation energy, R is the universal gas constant, T is the test temperature and A is the pre-exponential constant. Although the physical meaning of this equation is very relative for the composite systems studied, the value of the activation energy, obtained from the linear representation of In [eta] vs. 1/T, is a useful parameter to demonstrate the effects of temperature on the rheological properties of the composites studied.
Table 3 shows the activation energies for the injected starch matrix and its composites with different sisal short fiber concentrations at different frequencies. In general, it can be observed that the activation energy of the materials decreases as the angular frequency increases. It is possible to see that activation energy decreases with fiber content except for composites with 15 wt% short-sisal fibers, which could be due to weak fiber-fiber interaction.
Inspection of the results in Table 3 can be useful for determining the relative influence of material composition and processing conditions on the rheological behavior of these natural fiber composites. In the range of fiber contents studied, the activation energy values presented higher differences in function of the frequency than fiber content.
A systematic analysis of the melt rheological behavior of a commercial starch-based (MaterBi[R]) matrix composite reinforced with short sisal fibers was performed. The effects of shear rate, temperature, fiber content, and treatment were analyzed and classical non-Newtonian models were applied. The model was able to predict experimental data, and, as expected, the pseudoplastic behavior increased with the fiber content.
The systematic study demonstrated that shear rate was the most influential processing condition in the range of fiber content studied, while from the point of view of the material structure, the intercalation effectiveness of the polymer matrix in the fibers was directly linked to the rheological behavior. In fact, processing techniques with high stresses and more efficient mechanical mixing promote the opening of fiber bundles, increasing the aspect ratio of the fibers and the average viscosity of the molten composite. A similar effect on the increase of the aspect ratio was observed when treated fibers were used.
Table 1. Parameters [K.sub.2] (Consistency) and n (Power-Law Index) Obtained for the Power-Law Model. Fiber Content Temperature Injected (wt%) ([degrees]C) [K.sub.2] (kPa.s) n 0 180 125.5 0.330 190 34.1 0.441 200 13.9 0.500 5 180 250.1 0.283 190 67.3 0.391 200 35.2 0.417 10 180 371.2 0.252 190 141.8 0.347 200 85.8 0.362 15 180 440.2 0.255 190 195.4 0.316 200 96.9 0.381 Fiber Content Temperature Compressed Compressed-Fiber Treated (wt%) ([degrees]C) [K.sup.2] n [K.sub.2] n (kPa.s) (kPa.s) 0 180 137.1 0.318 137.1 0.318 190 37.6 0.417 37.6 0.417 200 14.3 0.489 14.3 0.489 5 180 304.2 0.272 402.8 0.213 190 80.4 0.366 132.4 0.313 200 36.8 0.404 40.4 0.402 10 180 436.0 0.248 500.7 0.191 190 168.7 0.324 212.5 0.374 200 90.11 0.361 102.3 0.358 15 180 537.7 0.246 686.2 0.187 190 228.8 0.303 303.7 0.256 200 102.7 0.380 114.7 0.384 Table 2. Parameters Obtained for the Cross Model. Fiber Content Temperature ([degrees]C) [tau] (s) [[eta].sub.0] b (wt%) (kPa.s) 0 180 4.00 399.3 0.713 190 31.26 249.5 0.573 200 36.96 88.7 0.505 5 180 8.21 1213.7 0.729 190 32.91 391.5 0.570 200 36.82 250.5 0.527 10 180 15.94 3094.4 0.757 190 34.09 1395.0 0.648 200 33.42 911.3 0.637 15 180 16.31 3892.1 0.758 190 36.35 1666.0 0.668 200 32.25 961.2 0.643 Table 3. Activation Energies for Injected Starch-Based Matrix and Its Composites With Sisal Short Fibers at Different Frequencies. [E.sub.a] (kJ/mol) Wt% Fibers [omega] = 0.1 rad/s [omega] = 1 rad/s [omega] = 10 rad/s 0 253.6 191.4 149.9 5 251.7 158.8 138.8 10 229.4 118.5 100.5 15 234.9 124.9 101.1
Financial support from Fundacion Antorchas, the National Research Council of Argentina (CONICET), Secretaria de Ciencia, Tecnologia e Inovacion Productiva. PICT12-08011 and the National Research Council of Italy (CNR) is gratefully acknowledged.
1. A. K. Bledzki and J. Gassan, J, Prog, Polym, Sci, 24, 221 (1999).
2. S. Iannace, L. Nocilla, and L. Nicolais, J. Appl. Polym. Sci., 73, 585 (1999).
3. H. Hanselka and A. S. Herrman, 7th International Tech. Textile Symposium 1994 (June 1995).
4. V. P. Cyras, S. Iannace, J. M. Kenny, and A. Vazquez, Polym. Compos., 22, 1 (2001).
5. A. Dufresne and M. R. Vignon. Macromolecules: 31, 2693 (1998).
6. J. Kuruvilla and L. H. Mattoso, Natural Polymers and Composites, 333 (1998).
7. C. Bastioli, "Starch-polymer composites," in Degradable Polymers, Principles and Applications, pp. 112-137. G. Scott and D. Gilead, eds., Chapman & Hall, London (1995).
8. C. Bastioli, Biodegradable Materials. The Wiley Encyclopedia of Packaging Technology, Second Edition, pp. 77-83, A. L. Brody and K. S. Marsh, eds., John Wiley & Sons, Inc. (1987).
9. C. Bastioli, "Biodegradable materials state of art and future perspectives," in Frontiers in the Science and Technology of the Polymer Recycling. Antalya, Turkey (June 1997).
10. G. Jayamol, R. Janardhan. J. Anand, S. Bhagawan, and S. Thomas, Polymer, 37, 24, 5421 (1996).
11. M. Fujiyama and J. Kawasaki, J. Appl. Polym, Sci., 42, 481 (1991).
12. K. Wang and J. Lee, J. Appl. Polym. Sci., 33, 431 (1987).
13. J. Crownson, M. Folks, and P. Bright, Polym, Eng, Sci., 20, 925 (1980).
14. A. Ramazani, A. Ait-Kaidi, and M. Grmela, J. Non-Newtonian Fluid Mech., 73, 241 (1997).
15. M. Kamal and A. Mutel, Polym. Eng., 5, 4, 293 (1985).
16. M. Kamal and A. Mutel, Polym, Compos., 10, 337 (1989).
17. L. Czarnecki and J. White, J. Appl. Polym, Sci., 25, 1217 (1980).
18. J. George, R. Janardhan, J. S. Anand, S. S. Bhagawan, and S. Thomas, Polymer, 37, 24, 5421 (1996).
19. B. Franzen, C. Klaso, J. Kubat, and T. Kitano, Composites, 20, 65 (1989).
20. S. Barbosa, D. Ercoli, M. Bibbo, and J. M. Kenny, Composites Structures, 26, 259 (1993).
21. R. Crowson and M. Folkes, Polym, Eng. Sci., 20, 63 (1980).
22. S. Barbosa and J. M. Kenny, Polym, Eng. Sci., 39, 1880 (1999).
23. V. Murthy, B. Gupta, and S. De, Plast, Rubber Process, Appl., 5, 307 (1985).
24. K. Joseph, B. Kuriakose, S. Thomas, and C. Premalatha, Plast, Rubber Process, Appl., 21, 238 (1994).
25. J. George, K. Joseph, S. Bhagawan, and S. Thomas, Mater, Lett., 18, 161 (1993).
26. R. Prasantha Kumar, K. C. Manikandan Nair, S. Thomas, S. C. Schit, and K. Ramamurthy, Composites Science and Technology, 60, 1737 (2000).
27. T. J. McMaster, A. Senouci, and A. C. Smith, Rheologica Acta, 18, 145 (1987).
28. M. Padmanabhan and M. Bhattacharya, J. Rheology, 35, 315 (1991).
29. R. Parker, A. L. Ollet, R. Lai-Fook, and A. Smith, Rheology of Food, p. 53, R. E. Carter, London (1990).
30. R. Parker, A. L. Ollet, R. Lai-Fook, and A. Smith, Rheology of Food, p. 1290, P. Zeuthen, London (1990).
31. J. Harper, T. Rhodes, and L. Wanninger, AIChE Symposium, 67, 40 (1971).
32. N. Cervone and J. Harper, J. Food Processing Engineering, 2, 83 (1978).
33. B. Vegnes, G. Della Valle, and J. Tayeb, Rheologica Acta, 32, 465 (1993).
34. J. Willet, B. Jasberg, and C. Swason, Polym. Eng. Sci., 35, 202 (1995).
35. G. Della Valle, P. Colonna, A. Patria, and B. Vergnes, J. Rheology, 40, 347 (1996).
36. M. Cross, J. Colloid, 20, 417 (1958).
37. P. Carreau, Trans, Soc, Rheol., 16, 99 (1972).
38. V. Alvarez and A. Vazquez, Polymer Degradation and Stability, accepted (September 2003).
39. K. Joseph, S. Thomas, and C. Pavithran, Polymer, 37, 5139 (1996).
40. V. P. Cyras, C. Vallo, J. M. Kenny, and A. Vazquez, "Effect of Chemical Treatment on the Properties of Short Sisal Fiber-Reinforced PCL/Starch Blend," J. Composite Materials, in press.
41. V. A. Alvarez, R. A. Ruseckaite, and A. Vazquez, J. Composites Materials, 37, 1575 (2003).
42. I. A. Kinloch, S. A. Roberts, and A. H. Windle, Polymer, 43, 7481 (2002).
43. C. Clansen and W. Kulicke, Prog, Polym, Sci., 26, 1839 (2001).
44. S. Barbosa, D. Ercoli, M. Bibbo, and J. M. Kenny, Composites Structures, 27, 83 (1994).
V. A. ALVAREZ (1), A. TERENZI (2), J. M. KENNY (2), and A. VAZQUEZ(*1)
(1) Research Institute of Material Science and Technology (INTEMA) Mar del Plata University Juan B. Justo 4302 (7600) Mar del Plata, Argentina
(2) Materials Engineering Centre University of Perugia Loc. Pentima Bassa 21, Terni, Italy
*To whom correspondence should be addressed.
|Printer friendly Cite/link Email Feedback|
|Author:||Alvarez, V.A.; Terenzi, A.; Kenny, J.M.; Vazquez, A.|
|Publication:||Polymer Engineering and Science|
|Date:||Oct 1, 2004|
|Previous Article:||Dynamics of foam growth: bubble growth in a limited amount of liquid.|
|Next Article:||Measurement and prediction of diffusion coefficients of supercritical C[O.sub.2] in molten polymers.|