A low-cost, low-fiber-breakage, injection molding process for long sisal fiber reinforced polypropylene.

INTRODUCTION

Cellulosic natural fibers are being incorporated into polymeric matrices to reduce cost and to improve mechanical properties, this trend being driven mainly by government regulations that promote the use of renewable and recyclable materials. At first it was expected that natural fibers might replace higher-priced synthetic fibers and reduce the environmental impact of much agricultural fiber waste. Nevertheless, published work has demonstrated that natural sisal, jute, flax, hemp, aspen, coconut or banana fibers cannot reach the mechanical properties of synthetic glass, carbon or DuPont Kevlar in applications of high performance. However, for applications where the production volume is large, if the mechanical properties do not need to be outstanding but have to be better than the properties of unfilled plastics, some composites reinforced with natural fibers may be a cost-efficient option. The development of these markets has advanced slower and faced more obstacles than expected, and the most relevant customer nowadays is the automotive industry.

Sisal fiber (SF) is extracted from the plant Agave Sisalana and is available in South Africa, Central America, and Asia. It is renewable, nonabrasive, and biodegradable and does not cause respiratory irritation. SF shows moderately high specific strength and stiffness, and has a very low price and enhanced energy recovery. SF has already been used as an effective reinforcing material in polymeric resin matrices to make useful structural composite materials (1, 2).

The mechanical and physical properties of sisal fiber reinforced composites are very sensitive to processing methods, because these determine important variables like fiber impregnation and mixing, fiber mechanical interaction with the polymeric matrix, final length and diameter, fiber orientation, and the maximum volume fraction of fibers that can be incorporated to the composites. Mechanical interaction between fibers and the polymeric matrix depends on fiber impregnation and on the polymer adhesion to the fiber surface. Fiber impregnation depends mainly on processing variables, and the polymer adhesion to the fibers depends on chemical identities (3).

Polypropylene (PP) is an attractive matrix for natural fiber reinforced polymeric composites. It has a low price, low density, high softening temperatures, good surface hardness, resistance to abrasion, low electric conductivity, low processing temperatures that avoid fiber thermal degradation, and good thermal stability, and it can be recycled by the same process (4, 5). Most of the problems faced by the processors when attempting to use SF/PP composites result from the poor mechanical properties often obtained (tensile, impact, and elastic modulus). Poor mechanical properties measured for some SF/PP composites are assumed to be due to the PP high melt viscosity and to the weak interfacial bonding between fibers surface polar groups and the hydrophobic matrix. The PP high viscosity is assumed to be responsible for poor penetration of the liquid PP into the fiber bundles, thus diminishing the SF/PP contact area. For better interaction at the SF/PP interface, the presence of a compatibilizer or coupling agent is required. An improvement is observed for the case of addition of maleic anhydride-grafted polypropylene (MA-g-PP) to the composite matrix (6). This coupling agent supplies functional groups that increase the fiber-matrix adhesion by an esterification reaction between cellulosic fiber hydroxyl groups and the anhydride functionality of MA-g-PP (6-8).

Most of the SF/PP composites used in industry are made out of short fibers (< 2 mm), dispersed in the matrix made up of PP and MA-g-PP using a twin-screw extruder. This blending process breaks the fibers in length and in diameter, and the aspect ratio changes, and long fiber composites cannot be obtained by this method. Consequently, the improvement in mechanical properties for these composites is only moderate. On the other hand, considerable quantities of very expensive MA-g-PP incorporated by this technique are homogeneously dispersed in the matrix because of its high solubility in PP. Therefore, only a small fraction of the MA-g-PP anhydride groups will reach the surface of the sisal fiber. The rest of the very expensive MA-g-PP is not used for its specific purpose.

Thermoplastic pultrusion is a process recently developed to obtain thermoplastic composites with high fiber concentrations (higher than 50% v/v). This process also allows preservation of the whole original length of the fibers. For the case of sisal fibers, the average length of the composite fibers that can be obtained by this process is higher than the so-called critical fiber length of the composite. The pultrusion process can be used to produce long rods, which can be cut to obtain pellets. These processes have been used to produce glass fiber-thermoplastic pellets, which have been used for injection molding, obtaining composites with better mechanical properties (9, 10).

Some authors have suggested the incorporation of a small quantity of MA-g-PP directly onto the sisal fibers, followed by pelletization, for later use of the pellets to make SF/PP composites. This would allow diminishing the quantity of necessary bonding agent to obtain much cheaper composites with better mechanical properties (11, 12).

For this work, a single-screw extruder and a specially designed die were used to produce SF/MA-g-PP continuous rod by thermoplastic pultrusion. The expectation is to improve fiber bonding, reduce the fraction of MA-g-PP in the final composite, and reduce fiber breakage by not using a twin-screw extruder blending process. Pellets obtained by cutting the pultruded rods were dry-blended at the injection machine hopper, and SF/MA-g-PP/PP composites were directly obtained by injection molding. The composite fiber content was determined, and the fibers were extracted from the composites to study the length distribution. Mechanical and impact properties of the composites were determined.

EXPERIMENTAL

Materials Used

Sisal fiber (Agave Sisalana) yarn with a fiber diameter of about 100-200 [micro]m was obtained from Brascorda Co. (Brazil). The fibers were used for this work as received, without any surface modification or chemical treatment. Mechanical properties are reported in Table 1.

Maleic anhydride-grafted PP (1.0 wt% MA) (Polybond 3200, produced by Crompton Europe) with a melt flow index of 110 g/10 min and melting point of 157[degrees]C was used as a coupling agent. Typical properties are reported in Table 2.

Commercial polypropylene (Stamylan CX02) with a melt flow index of 30 g/10 min was supplied by DSM. Mechanical properties are reported in Table 3.

Sample Preparation

Continuous dry sisal yarns were pulled through a melt impregnation die designed for this work. The extrusion die consisted of a single rotating disk, powered by an external electric motor, rotating into a polymer melt pool. The action of the single powered disk impregnated the sisal yarn with molten MA-g-PP supplied by a single-screw extruder (21 mm diameter, L/D = 24) at low pressure (about 0.2-0.4 MPa). The low melt pressure operation significantly reduces fiber breakage, and allows a fast yarn impregnation speed (between 10 and 20 m/min). A simple schematic of the process is shown in Fig. 1.

The thermoplastic pultrusion line was operated with an extruder mass flow of 700 g/h. The line speed of the sisal yarn was between 10 and 20 m/min, and the rotating speed of the disk was 250 rpm.

As a result of the process, a continuous SF/MA-g-PP rod of about 1.5 mm diameter is obtained. This continuous rod is then chopped into uniform pellets, typically 13 mm long. The fiber content of the pellets was measured as described below, and averaged 70% (w/w).

The pellets with high fiber concentration were dried in air in a controlled oven, at 85[degrees]C for 24 h, to reduce bubbles in the injected parts. The dried pellets were then dry-blended at the injection machine hopper with the required amounts of unfilled PP pellets. These blends were injection molded to make tensile specimens as specified by UNE-EN-ISO 527/1-2. A Margarit JM 110 injection molding machine was used to obtain composite tensile specimens. The injection machine used has a barrel diameter of 45 mm. A regular injection nozzle was used, with a diameter of 3 mm. Tensile specimens were molded with specified SF, MA-g-PP and PP concentrations. The SF concentrations ranged from 0% to 24.5% (w/w). The injection process was carried out at 180[degrees]C, with a mass flow of 45 g/s, reaching a maximum filling pressure of 5.5 MPa.

A blend of 24.5% (w/w) SF and 75.5% (w/w) PP was prepared in a Brabender mixer. The fibers were added to a polypropylene melt at 180[degrees]C, and the rotor speed was 60 rpm, blending for 10 min. The blend was taken from the mixer while hot, and then compression-molded. Tensile specimens were machined from the SF/PP composite sheet.

Fiber Content Determination

If the reinforcing fibers lose weight only at temperatures higher than that of the polymeric matrix (as is the case for glass or carbon fibers), then thermal gravimetric analysis (TGA) becomes an easy way to calculate the composite fiber content. The thermal decompositions of SF and thermoplastic polyolefins occur in the same temperature interval, and for this reason TGA is not a reliable method to determine the natural fiber content in the thermoplastic matrix used here. Figure 2 shows the TGA output for the SF, the MA-g-PP, and for the compounded pellets.

[FIGURE 1 OMITTED]

Looking at Fig. 2 in the 100[degrees]C-150[degrees]C region, we observe that: a) the SF loses weight by water evaporation, b) the MA-g-PP does not lose weight. A quick calculation of fiber content can be done on this basis, and it coincides quite well with another method described below. Events occurring at higher temperatures cannot be safely assigned separately to any of the composite components.

To overcome these difficulties, the composite fiber content and the average fiber length after injection molding were determined by dissolving the polymeric matrix in boiling xylene. The fibers were then recovered by filtration, washed, dried, weighted and measured. The fiber concentration results obtained in this way are reported in Table 3. The fiber length distribution curve shown in Fig. 3 was constructed with the measured length values of 650 individual fibers--taken from a single tensile specimen--after extraction in hot xylene, washing and drying.

[FIGURE 2 OMITTED]

Characterization of the Interface Adhesion

Fourier transform infrared (FT-IR) spectra of SF before and after the two-step extrusion and injection molding process were obtained. The sisal fibers were recovered from injected tensile specimens by dissolving the polymeric matrix in boiling toluene for several hours. The infrared spectra were recorded in the range between 2000 and 1500 [cm.sup.-1] at a resolution of 2 [cm.sup.-1] and 64 scans per each spectrum.

Scanning electron microscopy (SEM) was used to study the fracture surfaces of the composite samples.

Mechanical Properties

Stress-strain curves were calculated from force-displacement graphs measured in an Instron 5500 R 6025 Universal Mechanical Testing Machine. Tests were conducted under UNE-EN-ISO 527/1-2 specifications. Charpy impact tests were performed under UNE-EN-ISO 179 specifications. Heat distortion temperatures were measured under UNE-EN-ISO 75 "AF," with plane 1A type specimens.

RESULTS AND DISCUSSION

Figure 3 shows the number length distribution of the individual SF for injection molded composites. A large fraction of the population (91%) has lengths in the exact range of the length of the SF/MA-g-PP pellets; and the rest (9%) show varied smaller lengths. Here we must point out that--this being a number distribution--the results show that less than 5% of the total number of fibers in the pellets loaded at the injection machine break into pieces with shorter lengths. This is a very important feature of this process, since it suggests that most of the fiber breakage found for regular processes--mixing in a twin-screw extruder and pelletizing, followed by injection molding--occurs inside the twin-screw extruder (13).

[FIGURE 3 OMITTED]

Table 3 shows the effects of sisal fiber content on the tensile properties of virgin and reinforced polypropylene. The incorporation of approximately 25 wt% of sisal fiber (and a corresponding 10 wt% of MA-g-PP) causes the composite tensile modulus to increase by more than two times, and the tensile strength to improve by 45%.

To explain these effects, it is convenient to keep in mind the values for the mechanical properties for the sisal fiber and for the PP. The SF shows a tensile modulus of about 15 GPa, and an elongation at break of about 2%, while the PP shows a tensile modulus of 1.46 GPa, and an elongation at break higher than 100%.

Figure 4 is used to compare the stress-strain curves for all the composites compatibilized with MA-g-PP used for this work. The increase of fiber content in the composites causes three simultaneous effects, which clearly show that the fibers are strongly bonded to the PP matrix: a) the composite tensile modulus increases, b) the tensile strength increases, and c) the elongation at break is reduced.

In order to compare mechanical properties for SF/PP composites without MA-g-PP coupling agent, a composite containing 24.5% (w/w) of SF was prepared, and tensile properties were also included in Table 3. The two-step process could not be used to prepare SF/PP. Without adding MA-g-PP, the composites containing 24.5% (w/w) of long SF could not be injection molded below 230[degrees]C. A higher barrel temperature was necessary to reduce the very high shear viscosity of melt PP with the long SF dispersed. Tensile specimens injected at high barrel temperature show very serious SF thermal degradation and odor emission. For this reason, a compression molding technique was used to obtain SF/PP composites to compare the tensile mechanical properties of sisal composites with and without the incorporation of MA-g-PP. A method to apply the two-step process to obtain long SF/PP composites without adding MA-g-PP is being developed, and will be described in a future publication.

[FIGURE 4 OMITTED]

The measured long SF/PP composites tensile modulus is only 1.5 times larger than for unfilled PP, while the tensile strength decreases by 39%.

If MA-g-PP is used as a coupling agent, the tensile modulus and tensile strength increase because the real local deformation of the PP close to the fibers is severely restricted by the much higher modulus fibers that are strongly bonded to the PP matrix. The improvement of tensile properties by addition of MA-g-PP was reported by many authors (1, 2, 5-8, 11, 12, 14). These authors also suggest that the better bonding between the sisal fiber surface and PP is caused by the esterification of anhydride groups of MA-g-PP with the hydroxyl groups of SF. A strong adhesion between PP and SF allows a better stress transfer and distribution inside the composite, from the matrix to the fibers acting as reinforcement, and vice versa. Sisal fiber shows mechanical properties different from those of polypropylene, and the resulting composites display intermediate properties.

Figure 5 shows the FT-IR spectra--in the 2000-1500 [cm.sup.-1] region--for the SF as received (before the extrusion and injection process) and for SF extracted from SF/MA-g-PP/P tensile specimens. The characteristic peak at 1742 [cm.sup.-1] has been assigned to an ester bond (11, 15). The esterification of the maleic acid anhydride groups in MA-g-PP and the cellulosic hydroxyl groups in SF takes place.

[FIGURE 5 OMITTED]

Figures 6A and 6B show SEM micrographs of the fracture surfaces of the composite tensile specimens (24.5% SF). For the sample shown in Fig. 6A, MA-g-PP was not used for the composite preparation. The SF surface shows no bonded PP, and it looks like the fibers have been pulled out from the PP matrix, consuming very little energy in the process. Figure 6B corresponds to composites containing MA-g-PP. We can observe that the pulled SF shows a bonded PP matrix. This image corresponds to a strong interfacial adhesion, and explains the improvement measured for the tensile properties of SF/MA-g-PP/PP composites when compared with SF/PP composites. A large amount of energy is used to pull the SF out, and in some cases the SF breaks before being completely released from the PP matrix.

The heat distortion temperature results for composites containing MA-g-PP are also shown in Table 3. HDT for the composites increases by up to 35 K. This is again a consequence of the good adhesion between the sisal fibers and the PP matrix, because of the way HDT is determined. HDT is not actually a material property, but it is a measure of a beam bending speed, when a specified bending stress is applied and the beam temperature is being increased at a specified rate. The temperature at which the specified bending is reached is reported as the HDT. The matrix-bonded, high-modulus SF slows down the beam bending, thus resulting in a higher HDT.

[FIGURE 6 OMITTED]

Table 3 also shows results for the Charpy impact tests for composites containing MA-g-PP. The impact resistance of the composites falls by up to 30%. The impact resistance is diminished by the incorporation of the fragile MA-g-PP. Thermoplastic mechanical properties depend on having a high molecular weight. During the PP functionalization to prepare MA-g-PP, the molecular weight of PP falls dramatically. MA-g-PP shows a brittle impact behavior, and therefore its inclusion reduces PP impact resistance. On the other hand, the good SF/PP adhesion turns out to be inconvenient. If the fibers are not strongly bonded the matrix, on impact many of them are pulled out the matrix because the fiber resistance is larger than the interfacial force. This fiber pullout process consumes part of the energy received during the impact, and contributes to an increase in the impact resistance of the composite. If the fiber/matrix adhesion is too good, the fiber pullout process does not take place, and cracks grow, breaking the fibers. For this case of strong fiber-matrix bonding there is not an additional energy-consumption mechanism. The composite becomes brittle, the absorbed impact energy diminishes, and the Charpy impact resistance diminishes too. The addition of fibers also increase the probabilities of fiber agglomeration and stress concentration. Since the fiber ends can act as crack initiators, under these conditions a high concentration of fiber ends could cause crack initiation and potential composite failure.

CONCLUSION

In this work, a non-conventional system for producing long fiber SF/PP composites by injection molding has been tested. An adequate extrusion die was designed to perform a thermoplastic pultrusion process without any noticeable fiber breakage. The pultrusion line consists of a single-screw extruder, the impregnation die, a puller and a pelletizer system. By this simple process, SF/MA-g-PP pellets with very high fiber concentration could be obtained. These pellets--containing 70% w/w of long SF and 30% w/w of MA-g-PP coupling agent--could be dry-blended with regular PP pellets at the injection machine hopper, and SF/MA-g-PP/PP composites were directly obtained by injection molding. The injection molding process was carried out using a standard machine operating under normal conditions of barrel temperature and mass flow, without any specially designed injection nozzle. No thermal degradation was observed that was due to the injection molding process.

The use of this non-conventional two-step process allows the processing of SF/PP with control of the final mechanical properties and costs. The mechanical properties can be controlled by fiber content, coupling agent content, and fiber length. The raw materials' cost fraction of the composites can be considerably reduced by decreasing the amount of the very expensive MA-g-PP needed in the SF/MA-g-PP pellets to obtain the desired fiber bonding and composite properties.

The tensile mechanical properties obtained by testing the injection molded composites show a significant improvement when compared with unreinforced PP. The improvement is taken as an indication of two important facts: 1) the MA-g-PP acts as an efficient coupling agent; 2) long SF act as a good reinforcement for the PP matrix.

Table 1. Properties of Sisal Fibers.

             Tensile    Elongation  Tensile   Fiber       Density
             Strength*  at Break*   Modulus*  Diameter
             [MPa]      [%]         [GPa]     [[micro]m]  [g/[cm.sup.3]]

Sisal Fiber  640        1.5         15        100-200     1.45

* ASTM D 3379-75.

Table 2. Typical Properties of MA-g-PP.

Properties                 Typical Values  Method

Physical form              Pellets
Melt flow rate (190/2.16)  110 g/10 min    ASTM D1238
Density at 23[degrees]C    0.91 g/cc       ASTM D792
Maleic anhydride level     1.0 weight%
Melting point              157[degrees]C   (DSC)

Table 3. Measured Properties for All Composites.

Blends of PP + SF/MA-g-PP Pellets

Polypropylene         wt%           100      75.5     95      85
SF/MA-g-PP content    wt%             0       0        5      15
Sisal fiber content   wt%             0      24.5      3.5    10.5
MA-g-PP content       wt%             0       0        1.5     4.5
Sp. gravity of blend  g/[cm.sup.3]    0.916   1.024    0.930   0.962
Raw materials price   US$/Kg          1.08    0.9036   1.17    1.36
Tensile strength
  UNE-EN-ISO 527/1-2  MPa            24.29   14.82    24.87   26.70
Tensile modulus
  UNE-EN-ISO 527/1-2  GPa             1.460   2.196    1.608   1.905
Tensile elongation
  UNE-EN-ISO 527/1-2  %               5.05    3.21     4.41    3.43
Notched Izod impact
  UNE-EN ISO 179      KJ/[m.sup.2]    7.00   --        6.47    5.04
HDT
  UNE-EN ISO 75 "AF"
  plane specimen 1A   [degrees]C     55.5    --       57.2    62.9

Blends of PP + SF/MA-g-PP Pellets

Polypropylene         75      65
SF/MA-g-PP content    25      35
Sisal fiber content   17.5    24.5
MA-g-PP content        7.5    10.5
Sp. gravity of blend   1.0     1.027
Raw materials price    1.55    1.73
Tensile strength
  UNE-EN-ISO 527/1-2  30.06   34.98
Tensile modulus
  UNE-EN-ISO 527/1-2   2.512   3.306
Tensile elongation
  UNE-EN-ISO 527/1-2   2.37    1.71
Notched Izod impact
  UNE-EN ISO 179       4.79    4.83
HDT
  UNE-EN ISO 75 "AF"
  plane specimen 1A   73.7    89.6

ACKNOWLEDGMENTS

The authors would like to thank Juan Asarou for the extrusion die careful machining, and Laura Izaguirre for kindly measuring mechanical and impact properties. Financial support from ANPCYT (PICT 99-14-07247) and CYTED (Proyecto VII-11) is acknowledged with gratitude.

REFERENCES

1. Y. Li, Y. Wing Mai, and L. Ye, Com. Sci. Tech., 60, 2037 (2000).

2. P. V. Joseph, K. Joseph, and S. Thomas, Com. Sci. Tech., 59, 1625 (1999).

3. V. P. Cyras, S. Iannace, J. M. Kenny, and A. Vazquez, Polym. Compos., 22, 104 (2001).

4. C. Edwards, Reinforced Plastics, 34 (April 2001).

5. K. Van De Velde and P. Kiekens, Comp. Struct, 54, 355 (2001).

6. R. Gauthier, C. Joly, A. C. Coupas, H. Gauthier, and M. Escoubes, Polym. Compos., 19, 287 (1998).

7. P. V. Joseph, M. S. Rabello, L. H. C. Mattoso, K. Joseph, and S. Thomas, Com. Sci. Tech., 62, 1357 (2002).

8. R., Karnani, M. Krishnan, and R. Narayan, Polym. Eng. Sci., 37, 476 (1997).

9. R. Marissen, L. Th. van der Drift, and J. Sterk, Com. Sci. Tech., 60, 2029 (2000).

10. T. Hartness, G. Husman, J. Koening, and J. Dyksterhouse, Comp. Part A, 32, 1155 (2001).

11. M. Kazayawoko, J. J. Balatinecz, and L. M. Matuana, J. Mat. Sci., 34, 6189 (1999).

12. K. L. Fung, X. S. Xing, R. K. Y. Li, S. C. Tjong, and Y. W. Mai, Comp. Sci. Tech., 63, 1255 (2003).

13. K. Jayaraman, Comp. Sci. Tech., 63, 367 (2003).

14. A. K. Rana, A. Mandal, B. C. Mitra, R. Jacobson, R. Rowell, and A. N. Banerjee, J. Appl. Polym. Sci., 69, 329 (1998).

15. J. M. Felix and P. Gatenholm, J. Appl. Polym. Sci., 42, 609 (1991).

L. M. ARZONDO (1*), A. VAZQUEZ (1,2), J. M. CARELLA (1,2), and J. M. PASTOR (3)

(1) Departamento de Ingenieria en Materiales

(2) Instituto de Investigaciones en Ciencia y Tecnologia de Materiales

(INTEMA) (UNMdP-CONICET)

Facultad de Ingenieria

Universidad Nacional de Mar del Plata 

Juan B. Justo 4302, (7600) Mar del Plata, Republica Argentina

(3) Departamento de Fisica de la Materia Condensada

Escuela Tecnica Superior de Ingenieros Industriales

Universidad de Valladolid

Paseo del Cauce s/n (47011), Valladolid, Espana

*To whom correspondence should be addressed. E-mail: [e-mail]