Sisal fibre and its composites: a review of recent developments

Composites Science and Technology 60 (2000) 2037±2055

www.elsevier.com/locate/compscitech

Sisal ®bre and its composites: a review of recent developments Yan Li, Yiu-Wing Mai *, Lin Ye Centre for Advanced Materials Technology (CAMT), Department of Mechanical & Mechatronic Engineering J07, The University of Sydney, Sydney, NSW 2006, Australia Received 4 December 1998; received in revised form 4 April 2000; accepted 19 April 2000

Abstract Sisal ®bre is a promising reinforcement for use in composites on account of its low cost, low density, high speci®c strength and modulus, no health risk, easy availability in some countries and renewability. In recent years, there has been an increasing interest in ®nding new applications for sisal-®bre-reinforced composites that are traditionally used for making ropes, mats, carpets, fancy articles and others. This review presents a summary of recent developments of sisal ®bre and its composites. The properties of sisal ®bre itself, interface between sisal ®bre and matrix, properties of sisal-®bre-reinforced composites and their hybrid composites have been reviewed. Suggestions for future work are also given. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: B. Electrical property; B. Interface; B. Mechanical property; B. Surface treatment; Ageing; Degradation; Natural±®bre composite; Sisal ®bre

1. Introduction Sisal ®bre is one of the most widely used natural ®bres and is very easily cultivated. It has short renewal times and grows wild in the hedges of ®elds and railway tracks [1]. Nearly 4.5 million tons of sisal ®bre are produced every year throughout the world. Tanzania and Brazil are the two main producing countries [2]. Sisal ®bre is a hard ®bre extracted from the leaves of the sisal plant (Agave sisalana). Though native to tropical and sub-tropical North and South America, sisal plant is now widely grown in tropical countries of Africa, the West Indies and the Far East [3]. A sketch of a sisal plant is shown in Fig. 1 and sisal ®bres are extracted from the leaves. A sisal plant produces about 200±250 leaves and each leaf contains 1000±1200 ®bre bundles which is composed of 4% ®bre, 0.75% cuticle, 8% dry matter and 87.25% water [1]. So normally a leaf weighing about 600 g will yield about 3% by weight of ®bre with each leaf containing about 1000 ®bres. The sisal leaf contains three types of ®bres [3]: mechanical, ribbon and xylem. The mechanical ®bres are * Corresponding author. Tel.: +61-2-9351-2290; fax: +61-2-93513760. E-mail address: [email protected] (Yiu-Wing Mai).

mostly extracted from the periphery of the leaf. They have a roughly thickened-horseshoe shape and seldom divide during the extraction processes. They are the most commercially useful of the sisal ®bre. Ribbon ®bres occur in association with the conducting tissues in the median line of the leaf. Fig. 1 shows a cross-section of a sisal leaf and indicates where mechanical and ribbon ®bres are obtained [3]. The related conducting tissue structure of the ribbon ®bre gives them considerable mechanical strength. They are the longest ®bres and compared with mechanical ®bres they can be easily split longitudinally during processing. Xylem ®bres have an irregular shape and occur opposite the ribbon ®bres through the connection of vascular bundles as shown in Fig. 2. They are composed of thin-walled cells and are therefore easily broken up and lost during the extraction process. The processing methods for extracting sisal ®bres have been described by Chand et al. [2] and Mukherjee and Stayanarayana [1]. The methods include (1) retting followed by scraping and (2) mechanical means using decorticators. It is shown that the mechanical process yields about 2±4% ®bre (15 kg per 8 h) with good quality having a lustrous colour while the retting process yields a large quantity of poor quality ®bres. After extraction, the ®bres are washed thoroughly in plenty of clean water to remove the surplus wastes such as chlorophyll, leaf juices and adhesive solids.

0266-3538/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0266-3538(00)00101-9

2038

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

Fig. 2. Cross-section of a ribbon-®bre bundle.

Fig. 1. (a) A sketch of sisal plant and the cross-section of a sisal leaf [3]; (b) photograph of a sisal plant.

The chemical compositions of sisal ®bres have been reported by several groups of researchers [4±7]. For example, Wilson [4] indicated that sisal ®bre contains 78% cellulose, 8% lignin, 10% hemi-celluloses, 2% waxes and about 1% ash by weight; but Rowell [5] found that sisal contains 43±56% cellulose, 7±9% lignin, 21±24% pentosan and 0.6±1.1% ash. More recently, Joseph et al. [6] reported that sisal contains 85±88% cellulose. These large variations in chemical compositions of sisal ®bre are a result of its di€erent source, age, measurement methods, etc. Indeed, Chand and Hashmi [7] showed that the cellulose and lignin contents of sisal vary from 49.62±60.95 and 3.75±4.40%, respectively, depending on the age of the plant. The length of sisal ®bre is between 1.0 and 1.5 m and the diameter is about 100±300 mm [8]. The ®bre is actually a bundle of hollow sub-®bres. Their cell walls are

reinforced with spirally oriented cellulose in a hemi-cellulose and lignin matrix. So, the cell wall is a composite structure of lignocellulosic material reinforced by helical micro®brillar bands of cellulose. The composition of the external surface of the cell wall is a layer of lignaceous material and waxy substances which bond the cell to its adjacent neighbours. Hence, this surface will not form a strong bond with a polymer matrix. Also, cellulose is a hydrophilic glucan polymer consisting of a linear chain of 1, 4-b-bonded anhydroglucose units [9] and this large amount of hydroxyl groups will give sisal ®bre hydrophilic properties. This will lead to a very poor interface between sisal ®bre and the hydrophobic matrix and very poor moisture absorption resistance. Though sisal ®bre is one of the most widely used natural ®bres, a large quantity of this economic and renewable resource is still under-utilised. At present, sisal ®bre is mainly used as ropes for the marine and agriculture industry [1]. Other applications of sisal ®bres include twines, cords, upholstery, padding and mat making, ®shing nets, fancy articles such as purses, wall hangings, table mats, etc. [10]. A new potential application is for manufacture of corrugated roo®ng panels that are strong and cheap with good ®re resistance [11]. During the past decade (1987±1998), the identi®cation of new application areas for this economical material has become an urgent task. The use of sisal ®bre as a reinforcement in composites has raised great interest and expectations amongst materials scientists and engineers. The major studies on sisal ®bres carried out during this 10-year period can be broadly divided into the following topics: . Properties of sisal ®bres: Mechanical, thermal and dielectric properties of sisal ®bre have been studied in detail. X-ray di€raction, IR, TG, SEM, DSC, DMA, etc., have been used to determine the characteristics of sisal ®bre and provide theoretical

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

support for processing and application of this ®bre. . Interface properties between sisal ®bre and matrix: The main purpose here is to modify the ®bresurface structure by using chemical and thermal treatment methods in order to enhance the bond strength between ®bre and matrix and reduce water absorption of sisal ®bre. . Properties of sisal-®bre-reinforced composites: The matrix used in sisal-®bre-reinforced composites include thermoplastics (polyethylene, polypropylene, polystyrene, PVC, etc.), thermosets (epoxy, polyester, etc.), rubber (natural rubber, styrene±butadiene rubber, etc.), gypsum and cement. The e€ects of processing methods, ®bre length, ®bre orientation, ®bre-volume fraction and ®bre-surface treatment on the mechanical and physical properties of sisal-®bre-reinforced composites have been studied. Also, several theoretical models are given to predict the properties of the composites. . Sisal/glass-®bre-reinforced hybrid composites: To take advantage of both sisal and glass ®bres, they have been added conjointly to the matrix so that an optimal, superior but economical composite can be obtained. The hybrid e€ect of sisal/glass ®bres on the mechanical properties have been studied and explained. Papers published between 1987 and 1998 related to sisal ®bres are listed in Table 1. It can be seen that research interest has changed from the ®bre itself to sisal-®bre-reinforced composites and hybrid composites. The study of interface between sisal ®bre and matrix, however, remains an important topic. 2. Properties of sisal ®bre 2.1. Price Compared to synthetic ®bres, the price of sisal ®bre (0.36 US$/kg) is very low [3]. It is about one-ninth of that of glass ®bre (3.25 US$/kg) and one-®ve hundredth of carbon ®bre (500 US$/kg). For speci®c price (mod-

2039

ulus per unit cost), it (41.67 GPakg/$) is almost the best next to jute (43.33 GPakg/$) amongst all the synthetic and cellulosic ®bres. 2.2. Properties Generally, the strength and sti€ness of plant ®bres depend on the cellulose content and the spiral angle which the bands of micro®brils in the inner secondary cell wall make with the ®bre axis. That is, the structure and properties of natural ®bres depend on their source, age, etc. [12]. The tensile properties of sisal ®bre are not uniform along its length [3]. The root or lower part has low tensile strength and modulus but high fracture strain. The ®bre becomes stronger and sti€er at midspan and the tip has moderate properties. Table 2 shows the properties of sisal ®bres as reported by di€erent researchers. Note that except for the structure and properties of the natural ®bre itself, experimental conditions such as ®bre length, test speed, etc., all have some e€ects on the properties of natural ®bres [13,14]. Mukherjee and Satyanarayana [1] studied the e€ects of ®bre diameter, test length and test speed on the tensile strength, initial modulus and percent elongation at the break of sisal ®bres. They concluded that no signi®cant variation of mechanical properties with change in ®bre diameter was observed. However, the tensile strength and percent elongation at the break decrease while Young's modulus increases with ®bre length. With increasing speed of testing, Young's modulus and tensile strength both increase but elongation does not show any signi®cant variation. However, at a test speed of 500 mm/min, the tensile strength decreases sharply. These results have been explained in terms of the internal structure of the ®bre, such as cell structure, micro®brillar angle (20±25 ), defects, etc. In rapid mechanical testing, the ®bre behaves like an elastic body, i.e. the crystalline region shares the major applied load resulting in high values of both modulus and tensile strength. When the testing speed decreases, the applied load will be borne increasingly by the amorphous region. However, at very slow test speeds, the ®bre behaves like a viscous liquid. The amorphous regions take up a major

Table 1 Number of papers published in the period (1987±1998) related to sisal ®bresa Number of papers Classi®cation Properties of sisal ®bres Interface between sisal ®bre and matrix Properties of sisal-ibre composites Sisal/glass hybrid composites a

1987±1990

1991±1994

1994±1998

Total

4 [1,2,10,23] 1 [9] 3 [29,33,63] ±

6 [15±17,20±22] 1[8] 13 [3,11,34±37,54, 57±59,61,62,64] ±

± 2 [6,24] 11 [26,32,33,42±44, 51±53,55,56] 3 [67±69]

10 4 27 3

Data taken from Compendex (Computerised Engineering Index).

2040

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

Table 2 Properties of sisal ®bres reported by di€erent researchers Density (kg/m3)

Moisture content (%)

Tensile strength (MPa)

Tensile Modulus (GPa)

Maximum strain (%)

Diameter (mm)

Reference

1450 1450 ± 1030 1410 1400 ± 1450

11 ± ± ± ± ± ± ±

604 530±640 347 500±600 400±700 450±700 530±630 450±700

9.4±15.8 9.4±22 14 16±21 9±20 7±13 17±22 7±13

± 3±7 5 3.6±5.1 5±14 4±9 3.64±5.12 4±9

50±200 50±300 ± ± 100±300 ± 100±300 ±

[14] [2] [65] [28] [44] [43] [1] [53]

portion of the applied load giving a low ®bre modulus and a low tensile strength. But at very high strain rates (500 mm/min), the sudden fall in tensile strength may be a result of the presence of imperfections in the ®bre causing immediate failure. In Ref. [1], the micro®brillar angle and number of strengthening cells in the sisal ®bres did not show any appreciable variation in ®bre diameter. Hence, no appreciable change in values of Young's modulus and tensile strength were observed. As the test length increases, the number of weak links or imperfections also increases, thus resulting in reduction in tensile strength. However, with increasing ®bre length, sisal o€ers a higher resistance to applied stress as a result of the involvement of more oriented cellulosic ®bres. This probably also accounts for the higher modulus of the ®bres at longer test lengths. The reason for such behaviour by the characteristics of natural ®bres such as multi-cellular structures, visco-elastic nature and nonuniform structural inhomogeneity. Chand et al. [12] reported the e€ects of testing speed and gauge length on the mechanical properties of other kinds of natural ®bres (sun-hemp ®bres). Their results support the ®nding of Ref. [1], though the magnitudes are much lower than those of sisal ®bres. (For example, when the gauge length is 50 mm and testing speed is 50 mm/min, the tensile strength of sisal ®bre is 759 MPa. However, for sun-hemp ®bre, the tensile strength is only 283 MPa.) The mechanical properties of sisal ®bres obtained from di€erent age at three di€erent temperature were investigated by Chand and Hashmi [15]. The tensile strength, modulus and toughness (de®ned as energy absorption per unit volume) values of sisal ®bre decrease with increasing temperature. The relative e€ect of plant age on these mechanical properties is less prominent at 100 C than at 30 C. This is attributed to the more intense removal of water and/or other volatiles (at 100 C) originally present in the ®bres, which otherwise act as plasticising agents in the chains of the cellulose macromolecules. It is, however, noted that at 80 C both tensile strength and modulus decrease with age of the plant. This trend is di€erent to testing at 100 C. Table 3

shows the results of sisal ®bres of di€erent age at different temperature. For electrical applications, the dielectric properties of sisal ®bre at di€erent temperature and frequency have also been studied [16]. Increase of frequency decreases the dielectric constant "0 value, while increase of temperature increases "0 at all frequencies. Increasing the plant age shifts the dissipation factor ( tan ) peak to higher temperature. These phenomena were explained on the basis of structural changes. Water absorbed by sisal ®bres has OHÿ anions which act as dipoles. Other than OHÿ anions, there are several impurities and ions on the ®bre. These dipoles and ions contribute to the "0 and tan  behaviours of sisal ®bres. At low frequencies, high "0 and tan  values in sisal ®bre are caused by the dipolar contribution of absorbed water molecules. "0 values at intermediate frequencies are the result of contributions from space charge polarisation. At high frequencies, the contribution of polarisation of absorbed water molecules and space charge decreases and electronic and atomic polarisation becomes operative. Increase in temperature a€ects the mobility of ions and consequently changes the ionic contributions. The temperature dependence of the dielectric behaviour of sisal ®bre is shown in Fig. 3. Yang et al. [17] used IR, X-ray di€raction and TG to study the e€ect of thermal treatment on the chemical structure and crystallinity of sisal ®bres. They concluded that the IR spectrum did not change below 200 C treatment while density and crystallinity increased. This means that the chemical structure of sisal ®bres will not change below 200 C while the degree of crystallinity can be increased and hence the density. There was a slight weight loss (2%) below 200 C probably caused by the evaporation of water absorbed by sisal ®bres, substances of low boiling point and others that can be decomposed below this temperature. However, the large amounts of cellulose, semi-cellulose and glucans were not lost. Also, they found that thermal decomposition of sisal ®bre could be divided into three stages. The thermal behaviours are essentially identical for heat treatment between 150 to 200 C. Hence, thermal treatment of sisal ®bre can be carried out below 200 C.

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

2041

Table 3 Comparisons of mechanical properties of sisal ®bres with di€erent age at di€erent temperature [15] Age of plant

3 5 7 9

Toughness per unit volume (MJ/m3)

Tensile strength (MPa)

30 C

80 C

100 C

30

80 C

100 C

30 C

80 C

100 C

4.8 5.5 6.0 7.4

4.9 7.8 5.2 5.4

4.1 4.3 4.7 5.2

452 508 500 581

350 355 300 316

303 300 280 339

26 29 34 37

29 ± 22 17.5

21 22 17 21

Tensile modulus (GPa)

Table 4 Results of TG of treated and untreated sisal ®bresa,b Sisal ®bre

t1 ( C) w1 (%) t2 ( C) w2 (%) t3 ( C) w3 (%)

Untreated

69

98

150 C/4 h treated

69

98

200 C/0.5 h treated 61

98

302 334 297 333 300 327

80 57 83 51 79 56

478

8.2

478

6.8

478

7.3

a

Three-stage thermal decomposition [17]. ti ith stage decomposition temperature; wi ith stage percentage weight remnant. b

able levels by better wetting and chemical bonding between ®bre and matrix. 3.1. Treatment of sisal ®bre surface

Fig. 3. Variation of (a) dielectric constant "0 and (b) dissipation factor tan  with temperature in 2-year-old plant ®bres [17].

Table 4 shows the TG results of untreated and heattreated sisal ®bres. 3. Interface modi®cations Interfaces play an important role in the physical and mechanical properties of composites [18,19]. The hydroxyl groups which occur throughout the structure of natural ®bres make them hydrophilic, but many polymer matrices are hydrophobic so that sisal-polymer composites have poor interfaces. Also, the hydrophilic sisal ®bres will absorb a large amount of water in the composite leading to failure by delamination. Adequate adhesion across the interface can be achieved at desir-

Most previous studies were focused on ®bre-surface treatment methods and the resultant e€ects on the physical and mechanical properties of di€erent ®bre-matrix composite systems. But what will happen to the ®bres after being treated? Several investigators [20±22] have studied the surface morphology, mechanical and degradation properties of the treated ®bres. Yang et al. [20] studied the relationship of surface modi®cation and tensile properties of sisal ®bres. Their modi®cation methods include: alkali treatment, H2SO4 treatment, conjoint H2SO4 and alkali treatment, benzol/ alcohol dewax treatment, acetylated treatment, thermal treatment, alkali-thermal treatment and thermal-alkali treatment. The results are summarised in Table 5. Thermal treatment (at 150 C for 4 h) seems to be the most desirable method in terms of strength and modulus properties because of the increased crystallinity (from 62.4% for untreated to 66.2% for 150 C/4 h treated) of sisal ®bres. When the temperature reaches 200 C, the tensile properties will drop greatly as a result of the degradation of ®bres. Other treatments increase the ductillity of sisal ®bres substantially but decrease the modulus. To improve the moisture-resistance, Chand et al. acetylated sisal ®bre and studied its tensile strength [23]. It was shown that acetylation could reduce the moisture

2042

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

Table 5 E€ect of treatment methods on tensile properties of sisal ®bres [20]

. Coupling agent 3:

Treatment methods Tensile strength Tensile modulus Elongation (g/tex) (103 g/tex) at break (%) Untreated Benzol/alcohol Acetic acid+alkali Alkali Acetylated Thermal Alkali-thermal Thermal-alkali

30.7 38.8 9.3 31.7 33.2 42.0 27.6 25.7

1.18 0.99 0.39 0.53 0.35 1.22 0.70 0.71

2.5 3.7 2.6 7.5 8.3 3.5 4.7 4.4

content from 11 to 5.45%. However, the tensile strength of acetylated sisal ®bre was reduced from 445 to 320 MPa caused by the loss of the hemi-cellulose in the ®bre during acetylation. Surface modi®cation of sisal ®bre using coupling agents was also studied by Singh et al. [24]. Sisal ®bres were cut into short lengths (6 cm) and washed with distilled water to remove water solubles from the ®bre surface completely. After being air-dried they were heated to 80 C for 8 h to remove excess surface moisture. Then these ®bres were treated by dipping into the coupling-agent solution and stirring slowly for 0.5 h. After that, the treated ®bres were washed with solvent to remove compounds not covalently bonded to the ®bres and then kept at 80 C in an oven for 4 h to constant weight. The four coupling agents used were: N-substituted methacrylamide (coupling agent 1); gamma-methacryloxypropyl trimethoxy silane (coupling agent 2); neopentyl(dially)oxy, tri(dioctyl)pyro-phosphato titanate (coupling agent 3); and neopentyl(diallyl)oxy, triacryl zironate (coupling agent 4). Their chemical structures and possible interactions with sisal are as follows: . Coupling agent 1:

. Coupling agent 2:

. Coupling agent 4:

The hydroxyl groups attached to the glucose units of the cellulose will react with these coupling agents in the presence of moisture by substituting the right part of the dashed lines shown on the chemical structures of the coupling agents. The e€ects of these coupling agents on the moisture content in sisal ®bre are obtained and discussed. It is clear that moisture absorption of surface-treated ®bres has been reduced signi®cantly by providing hydrophobicity to the surface via long-chain hydrocarbon attachment. In addition, these coupling agents penetrate the cell wall through surface pores and deposit in the inter®brillar regions and on the surface, restricting further ingress of moisture. 3.2. Treatment of ®bre/matrix interfaces To make good use of sisal-®bre reinforcement in composites, ®bre-surface treatment must be carried out to obtain an enhanced interface between the hydrophilic sisal ®bre and the hydrophobic polymer matrices. Modi®cations of interfaces between sisal ®bre and polyester, epoxy, polypropylene, etc., have been studied. Both mechanical and moisture absorption resistance properties can be improved. These results are described in the following sub-sections. 3.2.1. Sisal/polyester composites The properties of sisal-®bre-reinforced polyester composites can be improved when sisal ®bres were suitably modi®ed with surface treatment [24]. The modi®cation methods have been discussed in Section 3.1. In the work by Singh et al. [24], it was explained that the modi®ed `interphase' is much less sti€ than the resin

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

matrix and provides a deformation mechanism to reduce interfacial stress concentration [25]. Further, it may also prevent ®bre/®bre contacts, hence removing the sources of high stress concentrations in the ®nal composites. The e€ect of ®bre-surface treatment on the mechanical properties under wet environment has also been studied in [24]. The results in Fig. 4 clearly show that by improving interfacial adhesion the moisture-induced degradation of composites can be reduced. Treated ®bre composites absorb moisture at a slower rate than the untreated counterparts, probably because of the formation of a relatively more hydrophobic matrix interface region by co-reacting organo-functionality of the coupling agents with the resin matrix. Though signi®cant reductions in tensile strength (30±44%) and ¯exural strength (50±70%) were observed for both untreated and surface-treated sisal composites (Table 6 and Fig. 4), the strength retention of surface treated composites is higher than that of composites containing untreated sisal ®bres. Fig. 4 also shows that N-substituted methacrylamide treated sisal-®bre-reinforced polyester composites generally exhibit better mechanical properties under dry and wet conditions. It should, however, be noted in Table 6 that with sisal ®bres treated by N-substituted methacrylamide and silane have relatively lower void contents compared to other ®bre-treated composites.

2043

3.2.2. Sisal/epoxy composites For the interface between sisal ®bres and epoxy, Bisanda and Ansell [8] adopted silane treatment methods to improve the adhesion and moisture resistance properties. The ®bres were ®rst dewaxed in 2000 ml solution of benzene and alcohol (methylated spirit), ratio 1:1, by soaking batches of about 200 g of sisal ®bres, 300 mm long, in a sealed glass vessel. The ®bres were soaked for 24 h, rinsed in alcohol and distilled water. These ®bres were then mercerized by soaking in a 0.5 N solution of sodium hydroxide, for about 72 h, rinsed in distilled water and dried. To study the moisture resistance of composites, the ®bres were treated using silane that had been diluted to 5% in methylated spirits. A 0.1 M solution of ceric ammonium nitrate (CAN) was used as catalyst. The ®bres were treated in batches by soaking about 150 g of the mercerized ®bres in 1500 ml of the silane/CAN solution (2:1), for about 24 h at room temperature. The ®bres were then rinsed in distilled water and dried. The reaction mechanisms are as follows. First, silane reacts with water to form a silanol and an alcohol: NH2 …CH2 †3 Si…OC2 H5 †3 ‡ 3H2 O ˆ …HO†3 Si…CH2 †3 NH2 ‡ 3…C2 H5 OH† Then, in the presence of moisture, the silanol reacts with the hydroxyl groups attached to the glucose units (G) of the cellulose molecules in the cell wall, thereby bonding itself to the cell wall with further rejection of water: NH2 …CH2 †3 Si…OH†3 ‡ H2 O ‡ GOH ˆ NH2 …CH2 †3 Si…OH†2 OG ‡ 2H2 O

Fig. 4. E€ect of wet environment on tensile strength of sisal/polyester composites (®bre content 50 vol.%) for 35 days: 1, untreated; 2, Nsubstituted methyacrylamide treated; 3, silane treated; 4, titanate treated; 5, zirconate treated. RH, relative humidity [24].

It is found that the treatment of sisal ®bres in silane, preceded by mercerization, provides improved wettabillity, mechanical properties and water resistance of sisalepoxy composites. The results are shown in Table 7.

Table 6 E€ect of surface treatments of sisal ®bres on the properties of sisal/polyester compositesa [24] Property

Untreated

N-substituted methacrylamide treated

Silane treated

Titanate treated

Zirconate treated

Density (g/cm3) Void content (%) Tensile strength (MPa) Elongation (%) Tensile modulus (GPa) Energy to break (MJ/m2) 105 Flexural strength (MPa) Flexural modulus (GPa)

0.99 16.11 29.66 9.52 1.15 7.96 59.57 11.94

1.05 5.88 39.48 9.75 2.06 11.06 76.75 15.35

1.12 3.01 34.14 5.71 1.75 4.78 96.88 19.42

1.02 12.55 36.26 8.00 1.67 8.60 75.59 15.13

1.00 13.30 34.69 9.51 1.39 10.13 72.15 14.46

a

Fibre content  50 vol%.

2044

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

Table 7 Mechanical and physical properties of sisal/epoxy compositesa [8] Material/Treatment

Compressive strength (MPa)

Flexural strength (MPa)

Flexural modulus (GPa)

Relative density

Water absorbed (%) after 72 h

Epoxy resin Untreated/dry Mercerised/dry Silane/dry Untreated/wet Mercerised/wet Silane/wet

120.0 148.0 183.1 184.8 62.53 75.83 98.53

95.0 266.5 262.1 244.5 221.7 200.7 237.2

3.1 15.93 17.63 17.36 9.15 10.12 12.33

1.17 1.14 1.24 1.25 ± ± ±

± ± ± ± 15.6 9.0 5.0

a

Fibre volume fraction: 0.4.

Mercerization has greatly improved resin pick-up, or wettability, of the ®bres because of the increase in density of the composite. It is believed that, by providing additional sites of mechanical interlocking, this treatment leads to improvement of interfacial bonding, hence promoting resin/®bre interpenetration at the interface. The high hydrophobic resin pick-up could also account for the reduction in water absorption and hence improved mechanical properties under wet conditions. Though treatment of sisal ®bres in silane preceded by mercerization produces very little change in the mechanical properties of dry composites, mechanical performance of wet composites, and hence water resistance, can be improved. The treatment in 100% silane produces ®bres that are almost hydrophobic. This may be a result of improved interfacial bonding arising from the use of the silane. Water molecules at the interface tend to replace the resin-®bre covalent bond by weaker hydrogen bonds, hence silane plays an important role in reducing water absorption in cellulosic-®bre-reinforced composites. Sisal ®bres have a central hollow region, the lumen, which gives access to water penetration by capillarity, especially when composites have high ®bre content. So, although silane treatment can create a hydrophobic ®bre surface, it is not possible to prevent water from entering the composite by capillary action, as long as the ®bre ends are exposed. It is recommended that, for practical purposes, it may be necessary to seal o€ the external surfaces by water-repellents so as to keep water uptake in the composite to a minimum. 3.2.3. Sisal/phenol formaldehyde composites Yang et al. [26] using NaOH, silane (3-aminopropyltriethoxy silane), chemical grafting and thermal treatment methods studied the e€ects on sisal/phenol formaldehyde composites. Alkali and thermal treatment methods have already been discussed in Section 3.1 [20]. For silane treatment, the ®bre was ®rstly treated by alkali and then immersed in the silane alcohol solution. After that, it was air-dried and followed by heating at 100 C. While for chemical grafting the sisal ®bre is

initially alkali treated and then grafted with vinyl acid in an acid medium under N2 atmosphere with a grafting rate of 5.7%. The mechanical behaviour, water-resistance and ¯exible fracture morphology of short sisal/ phenol formaldehyde composites have all been measured. The results show that cohesiveness between rigid sisal ®bre and brittle phenol formaldehyde can be improved when sisal ®bre is treated with silane so that the mechanical properties of the composites can be increased. Water-resistance of the composites can be improved after being treated by grafting and silane treatment. 3.2.4. Sisal/polyethylene composites Owing to the increasing use of thermoplastics, sisal ®bre reinforced-thermoplastics have become increasingly important. Joseph et al. [6] reported the e€ect of chemical treatment on the tensile properties of shortsisal-®bre reinforced-polyethylene (PE) composites (both randomly and unidirectionally oriented) and analysed the mechanisms of di€erent treatment methods. The treatment methods and their mechanisms are: a. Alkali treatment: Alkali treatment can remove natural and arti®cial impurities and produce a rough surface topography. In addition, alkali treatment leads to ®bre ®brillation, i.e. breaking down the ®bre bundle into smaller ®bres. This increases the e€ective surface area available for wetting by the matrix resin. Hence, increasing the ®bre aspect ratio caused by reduced ®bre diameter and producing a rough surface topography o€er better ®bre/matrix interface adhesion and increase in mechanical properties. b. Isocyanate treatment: The hydrophilic nature of sisal ®bres can be reduced by treating the ®bres surface with urethane derivative of cardanol (CTDIC) because of the linkage of the long chain structure of CTDIC to the cellulosic ®bres. This makes sisal ®bres compatible with the PE matrix, thus resulting in a strong interfacial bond between these two constituents with improved mechanical properties.

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

c. Peroxide treatment: The function of peroxide treatment is that it can graft PE on to the cellulose surface. The grafting reaction is a result of the peroxide-initiated free radical reaction between the PE matrix and cellulose ®bres as shown below: RO ÿ OR ! 2RO: RO: ‡ PE-H ! ROH ‡ PE: RO: ‡ Cellulose-H ! ROH ‡ Cellulose: PE: ‡ Cellulose: ! PE-Cellulose The tensile strength of the composites increases with peroxide concentration up to a critical level depending on the ®bre content and then remains constant. The existence of a critical concentration of peroxide suggests that the peroxide-initiated grafting reactions terminate when the ®bres are covered with grafted PE and excess peroxide causes some crosslinking of the PE molecules themselves. d. Permanganate treatment: Permanganate is thought to induce grafting reactions between sisal ®bres and PE matrix [6]. Joseph et al. have attempted to explain the initiating graft copolymerisation but the conditions for this to occur are quite critical and cannot be ful®lled easily. (It is also dicult to justify the di-valency of H in reactions given in [6].) Instead, oxidisation between permanganate and sisal ®bre will take place. So, the mechanism for improvement of the interfacial properties by permanganate treatment is that the permanganate roughens the ®bre surface and produces mechanical interlocks with the matrix. Hence, the interfacial bonding between permanganate treated sisal ®bre and matrix is improved. The permanganate concentration used is a critical factor for the mechanical properties of the composite. It is observed that the tensile strength reaches a maximum at a permanganate concentration of 0.055% and then decreases sharply with further increase in permanganate concentration. This is caused by the degradation of cellulosic ®bres at high permanganate concentration. Tensile properties of these sisal/PE composites with di€erent treatment methods were compared. It appears that the increase in properties as a result of these treatments are in the order: DCP (dicumyl peroxide) >CTDIC>BP(benzoyl peroxide)>KMnO4>alkali.

. . . .

2045

Sisal-®bre-reinforced Sisal-®bre-reinforced Sisal-®bre-reinforced Sisal-®bre-reinforced

thermosets thermoplastics rubbers cement and gypsum

The number of papers published since 1987 are listed in Table 8. It is shown that the most widely used matrix for sisal-®bre-reinforced composites has changed from gypsum and cement to rubber and polyethylene. Thermosetting matrices such as epoxy and polyester have also been used for sisal-®bre-reinforced composites. 4.1. Sisal-®bre-reinforced thermoset matrices The most widely used thermosetting matrix reinforced by natural ®bres is polyester [27±30]. Compression moulding is the most widely used and convenient method to make these composites, whether the ®bre is long or short. The tensile and impact properties of this kind of composites have been obtained by Sanadi et al. [27]. It is shown that the tensile strength and elastic modulus of the composites containing up to 40% ®brevolume fraction (Vf ) increase linearly with Vf in good agreement with the rule of mixtures. The work of fracture, as determined by Izod impact test, also increases linearly with Vf . Analysis of the energyabsorption mechanisms during impact fracture shows that ®bre pull-out and interface fracture are the major contributors to the high toughness of these composites. This result indicates that sisal ®bres have potential to produce inexpensive materials with high toughness. Comparison of the impact properties of di€erent natural ®bre-reinforced composites, including sisal, pineapple, banana and coir shows that sisal-®bre composites possess the highest impact toughness owing to the optimal micro-®brillar angle of the ®bre (21 for sisal, 12 Table 8 Number of papers related to sisal ®bres published in the period (1987± 1998)a Matrix

1987±1990

Polyethylene

0

Polystyrene

0

Rubber

0

Epoxy

0

Polyester

4. Properties of sisal-®bre-reinforced composites

Gypsum

According to the types of matrices used in sisal-®brereinforced composites, they can be divided into the following categories:

Cement a

3 [30,32,33] 0 3 [63,65,66]

1991±1994 4 [34±37] 0 4 [54,57±59] 1 [8] 0 3 [60,61,64] 1 [62]

1995±1998

Total

6 [6, 44,51±52,67±68] 1 [43] 3 [53,55,56] 0

10

1 [24] 0

4 3

0

4

1 7 1

Data taken from Compendex (Computerised Engineering Index).

2046

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

for banana, 14 for pineapple and 45 for coir). It has been proven by Gordon and Jeronimidis [31] that the toughness of composites increases with the micro®brillar angle of the ®bres and reaches a maximum at 15±20 . It will then decrease with increasing angle. The optimal micro-®brillar angle of sisal ®bre (21 ) leads to better impact resistance with a work of fracture of 98.7 KJ/m2 when the ®bre-volume fraction is 50%. For the same volume fraction of pineapple ®bres, this is 79.5 KJ/m2; for banana and coir ®bres, they are 51.6 and 43.5 KJ/m2, respectively. When compared to synthetic ®bre composites, the speci®c impact work of fracture for the natural ®bre composites is not much worse. The speci®c work of fracture (ie toughness per unit density) of 60% volume fraction sisal ®bre/polyester composites is 115 KJ mÿ2/g, while for ultra-high-modulus polyethylene (UHMPE) and E-glass ®bres, these values are 125 and 165 KJ mÿ2/g, respectively. Rong et al. studied the e€ect of ®bre pre-treatment and water absorption on the impact properties of sisal ®bre reinforced polyester and epoxy matrices [32,33]. Three ®bre-surface treatment methods including alkali treatment, coupling agent treatment and heat treatment were used. They indicated that ®bre-surface treatment has a strong e€ect on the impact behaviour of the composites and the e€ects are di€erent for di€erent matrices. It is observed that ®bre pull-out is the major contributor to the energy absorption. Increased ®bretensile strength promoted by thermal treatment can increase the impact performance of the composites. Water absorption in sisal ®bre composites is mainly caused by the ®bres and leads to a very poor interface between the sisal ®bre and the matrix. Di€erent matrix systems have di€erent interface characteristics. Generally, water absorption in sisal/polyester composite is two to three times that in sisal/epoxy composite and this leads to their di€erent impact properties. For sisal/epoxy composite, the impact strength improves with water absorption as a result of an acceptable level of interface debonding, but for sisal/polyester composites, the impact strength decreases through the complete destruction of the interface. From the viewpoint of interface enhancement, Bisanda and Ansell [8] studied the mechanical and physical properties of sisal/epoxy composites and Yang et al. [26] investigated sisal/phenol formaldehyde composites. The properties of these two composite systems have already been discussed in Sections 3.2.2 and 3.2.3. 4.2. Sisal-®bre-reinforced thermoplastics matrices In recent years, sisal-®bre-reinforced thermoplastics composites have gained much more interest among materials scientists and engineers than thermosets because of their low cost and recyclable properties. Many papers have now been published to study the properties

of these composites rather than sisal-®bre-reinforced thermosets. Such properties include mechanical, environmental, electrical and dynamic. 4.2.1. Processing methods The mixing methods so far used are: melt mixing and solution mixing. In melt mixing, the ®bre is added to a melt of thermoplastics and mixing is performed using a mixer at a speci®ed temperature and speed for a speci®ed time. Then the mix is taken out from the mixer while hot and is extruded using an injection moulding machine as long and thick rods. In the solution mixing method, the ®bres are added to a viscous solution of thermoplastics in a solvent in a stainless-steel beaker with a stainless-steel stirrer. The temperature is maintained for some time and the mix transferred to a ¯at tray and kept in a vacuum oven to remove the solvent. The solution-mixing procedure avoids ®bre damage that normally occurs during blending of ®bre and thermoplastics by melt-mixing [34]. Generally, randomly oriented sisal-®bre-reinforced composites are prepared by standard injection moulding of the blends. Oriented sisal composites are processed by aligning the long extruded rods with compression molding. Polyethylene is the most widely used thermoplastics matrix [34±37]. For other natural ®bre composites, polypropylene is also a major matrix material [38±41]. Recently, sisal ®bre-reinforced polystyrene and PVC have also been studied [42,43]. 4.2.2. Properties of sisal ®bre reinforced polyethylene 4.2.2.1. Mechanical properties. Joseph et al. [35] studied the tensile properties of short sisal ®bre/polyethylene composites in relation to processing methods and the e€ects of ®bre content, length and orientation (see Tables 9 and 10). As expected, the tensile properties show a gradual increase with ®bre length reaching a maximum at about 6 mm (12.5 MPa) and then decrease (e.g. 10.24 MPa at 10 mm). Unidirectional short ®bres achieved by extrusion enhance the tensile strength and elastic modulus of the composites along the axis of ®bre alignment by more than two-fold compared to randomly oriented ®bre composites. Di€erent processing methods lead to di€erent extent of ®bre damage, di€erent ®bre-length distribution and hence di€erent mechanical properties. The e€ect of ®bre length on the mechanical properties can be explained by the fact that long ®bres tend to bend or curl during moulding. This causes a reduction in the e€ective length of the ®bre below the critical ®bre length in a particular direction and hence a decrease of mechanical properties. The experimental tensile properties of short sisal®bre-reinforced LDPE with di€erent ®bre-volume fractions have been compared with existing theories of reinforcement [44]. The models selected were series and

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

2047

Table 9 Variation of tensile properties of melt mixing composite (MMC) and solution mixing composite (SMC) with ®bre content [35]a MMC

SMC

Fibre weight (%)

Tensile strength (MPa)

Modulus (MPa)

Elongation at break (%)

Fibre weight (%)

Tensile strength (MPa)

Modulus (MPa)

Elongation at break (%)

0 10 12 20

9.2 9.65 ± 11.25

140 276 ± 408

200 42 ± 22

0 10 12 20

30

10.2

346

8

30

9.2 10.8 (15.61) 12.5 (21.66) 14.7 (31.12)

140 324 (1429) 453 (2088) 781 (3086)

200.0 27.0 (4) 10.2 (3) 7.1 (1.8)

a

Average ®bre length: 5.8 mm; values given in parentheses are the properties of unidirectionally aligned ®bre composites.

Table 10 Variation of tensile properties of solution mixing composite (SMC) with repeated extrusion of the blends [35]a

Mc ˆ x…Mm Vm ‡ Mf Vf † ‡ …1 ÿ x†

No. of extrusions

Tensile strength (MPa)

Modulus (MPa)

Elongation at break (%)

Tc ˆ x…Tm Vm ‡ Tf Vf † ‡ …1 ÿ x†

1 3 5 6

12.5 11.24 10.8 10.4

453 388 316 285

10.2 13.4 20.2 23.0

a

Average ®bre length: 5.8 mm, ®bre content: 20 wt.%.

parallel [45], Hirsch [46], Halpin-Tsai [47], modi®ed Halpin-Tsai [48], Cox [49] and modi®ed Bowyer and Bader [50] models. These models are given below for easy reference: . Parallel and series model [45]:  Parallel model …1†

Tc ˆ Tf Vf ‡ Tm Vm

…2†

 Series model Mc ˆ

Tc ˆ

Mm Mf Mm Vf ‡ Mf Vm

Tm Tf Tm Vf ‡ Tf Vm

…3†

…4†

where M, T and V are Young's moduli, tensile strength and volume fraction, respectively. The subscripts c, m and f represent composite, matrix and ®bre, respectively. . Hirsch's model [46]

Tf Tm Tm Vf ‡ Tf Vm

…5† …6†

where x is a parameter which determines the stress transfer between ®bre and matrix. It is always assumed that x is determined mainly by ®bre orientation, ®bre length l and stress ampli®cation e€ect at the ®bre ends. This model is in fact a combination of the parallel and series models. . Halpin-Tsai model [47] 1 ‡ AVf † Mc ˆ M m … 1 ÿ Vf

…7†

1 ‡ A Vf † 1 ÿ  Vf

…8†

Tc ˆ T m …

Mc ˆ Mf V f ‡ Mm V m

Mf Mm Mm Vf ‡ Mf Vm

ˆ

Mf =Mm ÿ 1 Mf =Mm ‡ A

…9†

Tf =Tm ÿ 1 Tf =Tm ‡ A

…10†

 ˆ

where A is a factor determined by ®bre geometry, ®bre distribution and ®bre-volume fraction.  and  account for the relative moduli and strength of ®bre and matrix, respectively. . Modi®ed Halpin-Tsai equation [48] 1 ‡ AVf Mc ˆ M m … † 1 ÿ  Vf

…11†

1 ‡ A Vf † 1 ÿ  V f

…12†

Tc ˆ T m …

2048

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

ˆ1‡…

1 ÿ max †Vf 2max

…13† …14†

K ˆ 1 ‡ 2l=d

…15†

The modi®ed Halpin-Tsai equation includes the maximum packing fraction, max , of the reinforcement. max equals 0.785 for square arrangement of ®bres, 0.907 for hexagonal array of ®bres and 0.82 for random packing of ®bres.  and  are given by Eqs. (9) and (10), respectively, and depends on the particle packing fraction. l=d is the ®bre aspect ratio. . Cox model [49]

Tc ˆ Tf Vf …1 ÿ ˆ‰

tanh l=2 † ‡ Mm V m l=2

tanh l=2 † ‡ Tm Vm l=2

2Gm 1 Š2 Mf Af ln…R=r†

…16† …17† …18†

where r is the radius of the ®bre, Gm is the shear modulus of matrix, Af is the area of the ®bre and R is the centre-to-centre distance of the ®bres. For hexagonally packed ®bres 2r2 1 R ˆ …p †2 3V f

…19†

For square packed ®bres  12 † R ˆ r… 4Vf

…20†

. Modi®ed Bowyer and Bader's model [50] Bowyer and Bader's model indicates that the tensile strength of short-®bre-reinforced thermoplastic composites is the sum of contributions from sub-critical and super-critical ®bres and that from the matrix. M c ˆ Mf K 1 K 2 V f ‡ Mm V m

…21†

Tc ˆ Tf K1 K2 Vf ‡ Tm Vm

…22†

where K1 is the ®bre orientation factor and K2 is the ®bre length factor. If ®bre length l > lc , K2 ˆ l ÿ lc =2l If ®bre length l < lc ,

…24†

where lc is the critical ®bre length.

AˆKÿ1

Mc ˆ Mf Vf …1 ÿ

K2 ˆ l=2lc

…23†

Experimental results and theoretical values obtained from the above models were compared. It is observed that all models except the series and parallel models show reasonable agreement with experimental data of longitudinally oriented composites, especially at low ®bre-volume fractions. But, the Hirsch and modi®ed Bowyer and Bader equations show the closest agreement with test results. However, agreement with the experimental data of randomly oriented composites cannot be accurately predicted with these two models as shown in Figs. 5 and 6. All theoretical models indicate that the tensile properties of short-®bre-reinforced composites strongly depend on ®bre length, ®bre volume fraction, ®bre dispersion, ®bre orientation and ®bre/ matrix interfacial strength. 4.2.2.2. Dynamic mechanical properties. The e€ects of ®bre length, orientation, volume fraction and ®bre surface treatment on the dynamic mechanical properties of sisal-®bre-reinforced PE including storage modulus, loss modulus and damping characteristics have also been studied [34]. It is found that addition of 10% of short sisal ®bres into LDPE increases the storage moduli and loss moduli of the composites, levelling o€ at higher volume fraction. This is believed to be caused by the increasing ®bre-to-®bre interaction at high volume fractions resulting in poor dispersion. Both storage and loss moduli decrease with increase of temperature. The damping properties of the composites decrease with addition of ®bres and are strongly in¯uenced by ®bre orientation. The storage and loss moduli of randomly oriented composites were intermediate between those of longitudinally and transversely oriented composites. The in¯uence of ®bre length indicates that a critical length of 6 mm is needed to obtain maximum dynamic moduli. This suggests that a critical length exists for maximum stress transfer between ®bre and matrix. The storage and loss moduli of the isocyanate-treated composites are higher than those of the untreated composites as a result of the improved ®bre/matrix interface adhesion. Some dynamic mechanical results are shown in Figs. 7 and 8. 4.2.2.3. Electrical properties. The electrical properties of sisal-®bre-reinforced LDPE have been studied with respect to the e€ects of frequency, ®bre content and ®bre length on the dielectric constant, volume resistivity and dielectric loss factor [51]. The dielectric constant increases steadily with increasing ®bre content for all frequencies in the range 1 to 107 Hz. It is also noted that dielectric constant decreases with increase of ®bre length and frequency. Maximum dielectric constant values are

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

Fig. 5. Variation of experimental and theoretical tensile properties of randomly oriented composites as a function of ®bre-volume fraction [44] ( Ð ) strength (- - - -) modulus.

Fig. 6. Variation of experimental and theoretical tensile properties of randomly oriented composites as a function of ®bre length [44]. For keys, see Fig. 5.

Fig. 7. Storage moduli and mechanical-loss factor versus temperature for untreated (randomly oriented) sisal/LDPE composites at di€erent ®bre content. Fibre length=6 mm, frequency=10 Hz [34].

obtained at low frequencies. Sisal/LDPE composites of 1 mm ®bre length and 30% ®bre content have the highest values of dielectric constants at all frequencies. The volume resistivity values decrease with increase of frequency and ®bre content, i.e. the electric conductivity of composites are greater than neat LDPE. When compared to glass/LDPE composites, the same trend in electrical properties is observed, but the changes of

2049

Fig. 8. Variation of loss moduli with temperature for untreated (randomly oriented) sisal/LDPE composites at di€erent ®bre content. Fibre length=6 mm, frequency=10 Hz [34].

dielectric constants of the latter composites with frequency and ®bre content are smaller as a result of their lower interfacial polarisation. The dielectric constant, dielectric loss factor and electrical conductivity of 25% sisal/LDPE can be increased considerably by adding 5% carbon black. The electrical conductivity of hydrophobic LDPE can also be improved by mixing it with hydrophilic ligno-cellulosic ®bres. Solution mixing technique has no adverse e€ects. Finally, it is important to mention that a 25% sisal/LDPE composite containing 5% carbon black can be used in anti-static applications to dissipate static charges. 4.2.2.4. Ageing properties. The environmental properties of sisal-®bre composites are very important because, as a natural ®bre, sisal ages and causes degradation. The e€ects of aging on the physical and mechanical properties of sisal-®bre-reinforced polyethylene composites have been studied [52]. The tensile properties and dimensional stability are evaluated under two di€erent ageing conditions: one is by immersing samples in boiling water for 7 h under atmospheric pressure; and the other is by heating the samples at 70 C in an air circulating oven for 7 days. Both cardanol derivative of toluene di-isocyanate (CTDIC) treated and untreated sisal®bre-reinforced composites have been studied. The ageing properties of the sisal composites are also compared to those of glass-®bre composites aged under identical conditions. It is concluded that CTDIC-treated composites showed better mechanical properties and dimensional stability as compared to untreated composites as a result of the existence of an e€ective interfacial bond between ®bre and matrix. Better dimensional stability is o€ered by glass/LDPE composite because of the hydrophobic nature of the glass ®bre. With suitable ®bre-surface treatment, mechanical properties such as strength and elastic modulus of sisal/LDPE composites can be improved to comparable levels as those of glass/ LDPE.

2050

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

leads to the poor immersion of ®bres in the PVC matrix. Also, both treated and untreated sisal/PVC composites have quite good moisture resistance.

4.2.3. Properties of sisal ®bre-reinforced polystyrene composite For other thermoplastics matrices, Manikandan et al. [43] studied the tensile properties of short sisal-®brereinforced polystyrene composites. Untreated and benzoylated sisal ®bres were used to produce the composites and the in¯uences of ®bre length, ®bre content, ®bre orientation and ®bre benzoylation were investigated. Variation in ®bre length produces no considerable change in the modulus of the composites but gives maximum tensile strength (25 MPa) at a ®bre length of about 10 mm (aspect ratio=82). This critical ®bre length is quite di€erent to the sisal/PE composite which is 6 mm. Table 11 shows the e€ect of ®bre-volume fraction on the mechanical properties. There is an initial reduction in tensile strength at Vf ˆ 10% followed by an increase to Vf ˆ 20% and remains constant at even higher Vf. These results are also di€erent to sisal/PE composites which conform to the rule of mixtures. The orientation e€ects on the mechanical properties of sisal/ PS composites are also given in Table 11.

4.3. Sisal-®bre-reinforced rubber matrix Table 8 shows that rubber is the second most widely used matrix for sisal ®bre composites behind polyethylene [53±59]. Rubber matrices include natural rubber and styrene±butadiene rubber. The main research areas concern the e€ect of ®bre length, orientation, loading, type of bonding agent and ®bre/matrix interaction on the properties of the composites which include mechanical properties, rheological behaviour, thermal ageing, gamma-radiation and ozone resistance. Experimental results show that, for best balance of properties, the ®bre length is about 6 mm. This is the same as the sisal/PE composites. Orientation e€ects are as expected. Addition of short sisal ®bres to rubber o€ers good reinforcement, which can be further strengthened by a suitable coupling agent such as a resorcinol/heca bonding system. A two-stage stress relaxation has been observed in acetylated sisal-®bre-reinforced natural rubber composites. Initial relaxation occurs at short times (200 s), and second-stage relaxation takes much longer to complete. The initial mechanism is a result of the ®bre/rubber attachments and the latter to the physical and chemical relaxation processes of the natural rubber molecules. The relaxation is in¯uenced by the coupling agent indicating that the ®bre/rubber interface is involved. Gum vulcanite shows only one relaxation process, the rate of which is almost independent of the strain level. For the composite without a coupling agent, the rate of relaxation increases with strain level and vice versa. The initial rate of the stress-relaxation process diminishes after ageing (at 70 and 100 C for 4 days). Varghese et al. [57] studied the e€ect of acetylation and bonding agent on the ageing properties of sisal-®brereinforced natural rubber composites which include

4.2.4. Properties of sisal-®bre-reinforced PVC composite Yang et al. [42] studied sisal/PVC composites with respect to the e€ects of ®bre and matrix modi®cation, processing parameters on the mechanical and water resistance properties. Their main objective is to obtain the best processing parameters and interface modi®cation to make novel sisal/PVC composites. To make good use of sisal ®bre and PVC, it is important to improve the interface so that better mechanical properties of the composite can be obtained. But, unfortunately,their results show that thermal treatment, acetylation and coupling agent improve neither the interface nor the mechanical properties. On the contrary, the untreated sisal-®bre-reinforced PVC composite possesses better mechanical properties. These results have been explained by the small ®bre-volume fraction (18.5%) of their composites and the melting processing method that

Table 11 Tensile properties of sisal-polystyrene composites as a function of ®bre content and ®bre orientation [43]a Fibre Content (wt%)

Ultimate tensile strength (MPa)

Young's modulus (MPa)

L

L

T

R

0

34.9

34.9

34.9

10

29.41 (21.30)

24.49 (14.75)

24.72 (18.16)

800 (629)

660 (596.8)

20

45.52 (43.20)

22.96 (12.37)

31.14 (25.98)

1026 (999.9)

30

48.3 (45.06)

22.08 (11.04)

30.09 (20.42)

1125 (998)

a

390

T 390

Elongation at break (%) R 390

L

T

R

9

9

9

457.4 (516.8)

6 (9)

4 (5)

6 (7)

664.7 (487.9)

543.3 (553.7)

7 (8)

5 (3)

6 (6)

701.3 (577.6)

710.7 (624.3)

8 (7)

4 (2)

6 (4)

Fibre length was 6 mm. The values in parentheses give the properties of untreated ®bre composites. L, longitudinal; T, transverse; R, randomly oriented.

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

2051

®re behaviour of sisal short-®bre-reinforced gypsum using specially designed testing equipment. Though gypsum itself has good combustion-resistance, with increasing water reduction as a result of increasing temperature, there is a progressive shrinkage process that promotes surface ®ssuration. Adding sisal ®bres into the gypsum matrix increases the ®re insulating performance and delays the occurrence of surface ®ssuration. Swift [66] studied sisal/cement composites and their potential for rural Africa. He studied the mechanical properties including ¯exural, energy absorption and pointed out that a composite material formed by sisal ®bre and cement is suitable for applications in several structures. For example, cladding walls to produce earthquake-resistant adobe structures for houses, roofing sheets and tiles, grain storage bins and water ducts.

thermal ageing, gamma radiation and ozone resistance. High ®bre-volume fraction shows better resistance to ageing, especially with ®bre-surface treatment. Fibre orientation is also found to reduce the extent of degradation under these ageing conditions. Increasing the dosage of gamma radiation was found to increase the extent of the ageing process. 4.4. Sisal-®bre-reinforced gypsum and cement matrices For applications as building materials, sisal-®brereinforced gypsum and cement composites have long been studied before 1994 [60±65]. The majority of these works are focused on interface, mechanical, ®re and environmental properties and their applications. Bessell and Mutuli [65] studied the interfacial bond strength of sisal/cement composites using a tensile specimen containing a single ®lament in the brittle matrix. The crack spacing in the matrix was measured and used to evaluate the ®bre/matrix bonding strength. It is shown that the interfacial bond of sisal/cement composite is lower than that of other composites because sisal ®bre absorbs moisture from cement thus leading to a very poor interface. The following aspects of sisal-®bre-reinforced cement or gypsum composites have also been studied previously. For example, Hernandez et al. [61] studied the

4.5. Other matrix systems Bisanda and Ansell [3] used cashew nut shell liquid (CNSL) as a matrix to make sisal-®bre-reinforced composites. CNSL is a natural monomer blend that has been condensation polymerized with formaldehyde in the presence of an alkaline catalyst to produce a thermosetting resin. It can be used to bind sisal ®bres to produce a cheap and useful composite. The resin is

Table 12 Variation of mechanical properties with volume fraction of GRP in sisal/glass hybrid composites [68]ab Hybrid designation

Volume fraction of ®bres

Tensile strength (MPa)

Tensile modulus (GPa)

Elongation at break (%)

Tear strength (N/mm)

Hardness (Shore D)

Sisal

Glass

VSRP

VGRP

SRP 1 2 3 4 5 6 7

0.3 0.27 0.24 0.21 0.18 0.15 0.09 0.06

± 0.02 0.04 0.06 0.08 0.10 0.14 0.16

1 0.9 0.8 0.7 0.6 0.5 0.3 0.2

± 0.1 0.2 0.3 0.4 0.5 0.7 0.8

12.37 (7.25) 15.97 (7.72) 16.68 (7.83) 16.92 (7.97) 17.6 (8.11) 17.76 (8.26) 19.98 (8.7) 20.98 (9)

133.3 (130) 160.8 (147.3) 171.4 (152) 187.5 (160) 192.3 (175.23) 196.6 (189.6) 200 (190.3) 210 (194.6)

8.8 (13.55) 7.64 (12.52) 7.56 (11.79) 7.16 (10.89) 6.9 (10.39) 6.6 (6.8) 5.88 (6.0) 5.62 (9.56)

52.57 60.66 70.45 71.32 72.16 73.23 79.78 81.33

43 47 50 52 54 57 60 63

GRP

±

0.2

±

1

22.94 (9.28)

220 (200)

4.81 (9.37)

86.83

64

a b

SRP, sisal-®bre-reinforced plastics; GRP, glass-®bre-reinforced plastics. Values in parentheses are properties of randomly oriented composites.

Table 13 Properties of 50:50 sisal/glass hybrid composites containing alkali-treated and untreated sisal ®bre [68]a

Untreated Treated a

Tensile strength (MPa)

Young's modulus (MPa)

Elongation at break (%)

L

R

L

R

L

R

17.76 19.66

8.26 8.89

196 210

189 200

6.6 5.2

10.12 9.9

L, longitudinal oriented; R, random oriented.

Tear strength (N/mm)

Hardness (Shore D)

73.23 78.47

57 62

2052

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

thermally stable up to 230 C and is further cross-linked when exposed to simulated sunlight. The plain woven mats of mercerized sisal ®bre when impregnated with CNSL/fomaldehyde resin produced plain and corrugated laminated composites showing mean tensile strength of 94.5 MPa and Young's modulus of 8.8 GPa [3]. It is recommended that these low-cost corrugated panels can be used for roo®ng applications as a result of their adequate crush bending strength (13.9 MPa). 5. Sisal and synthetic hybrid-®bre composites Reinforcement by two or more ®bres in a single matrix leads to hybrid composites with a great diversity of material properties. It appears that the behaviour of hybrid composites is simply a weighted sum of the individual components so that there is a more favourable balance of properties in the resultant composite material. Sisal and glass ®bres are one good example of hybrid composites [67±69] possessing very good combined properties. For sisal/glass-®bre-reinforced LDPE hybrid composites, the e€ects of ®bre orientation, composition and ®bre-surface treatment on the mechanical properties have been studied and the results are shown in Table 12. Owing to the superior properties of glass ®bres the mechanical properties of the hybrid composites increase with increasing volume fraction of glass ®bres. `Positive' or `negative' hybrid e€ect is de®ned as `larger' or `smaller' than the properties calculated from the rule of mixtures of the two constituent ®bre-reinforced composites. A positive hybrid e€ect has been observed for all mechanical properties except elongation at break. This e€ect is a consequence of increased ®bre dispersion and orientation with increasing volume fraction of glass ®bres. The e€ect of chemical modi®cation (alkali treatment) of sisal ®bres on the mechanical properties of a 50:50 sisal/glass hybrid composite is shown in Table 13 and the improvement is generally less than 10%. Water absorption of the composite is reduced from 11.6 to 3.1% compared to the non-hybridized sisal-®bre composite. Yang et al. [69] studied the mechanical and interface properties of sisal/glass-®bre-reinforced PVC hybrid composites before and after immersion in water. It is found that there exists a `positive' hybrid e€ect for the ¯exural modulus and unnotched impact strength but a `negative' hybrid e€ect for the ¯exural strength. It is suggested that the `negative' hybrid e€ect is caused by the poor interface between sisal, glass ®bres and PVC matrix. They also suggested that water will have a detrimental e€ect on the ®bre/matrix interface leading to reduced properties.

6. Conclusions Sisal ®bre is an e€ective reinforcement of polymer, rubber, gypsum and cement matrices. This has created a range of technological applications beyond its traditional usage as ropes, carpets, mats, etc. The mechanical and physical properties of sisal ®bre not only depend on its source, position and age which will a€ect the structure and properties, but also depend on the experimental conditions, such as ®bre diameter, gauge length, strain rate and test temperature. Fibre-surface treatment can improve the adhesion properties between sisal ®bre and matrix and simultaneously reduce water absorption. Such methods include: (a) silane and other coupling agents to introduce long chain structures onto the sisal ®bre to change its hydrophilic characteristics; (b) peroxide to promote grafting reactions; (c) permanganate and alkali to increase the roughness of ®bre surface hence increasing the surface area available for contact with the matrix; and (d) thermal treatment. Di€erent matrix systems have di€erent properties. The mechanical and physical properties of sisal-®brereinforced composites are very sensitive to processing methods, ®bre length, ®bre orientation and ®bre-volume fraction. Sisal and glass ®bres can be combined to produce hybrid composites which take full advantage of the best properties of the constituents. Almost all the mechanical properties show `positive' hybrid e€ects. 7. Future work From this review, it is clear that chemically treated or modi®ed sisal-®bre-reinforced composites are potential structural materials as a result of their good mechanical, environmental and economic properties. The following areas of research are, however, needed to realise wider applications of sisal ®bres in engineering: . Sisal-textile-reinforced composite is an important area in which little work has been done [3]. Sisal ®bres can be woven into textile preforms and impregnated with resins by resin transfer moulding (RTM) or resin ®lm infusion (RFI) to make superior but more economical composites. . Microstructure of the interface between sisal ®bre and matrix still needs to be investigated and the interfacial properties should be studied with more rigorous single ®bre pullout and fragmentation tests [70]. The relationship between interface and bulk composite properties should be established. . Fracture toughness and fracture mechanisms of sisal-®bre composites do not seem to have been

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055

.

.

.

.

studied in any depth in previous published works. This is important if new improved materials are to be developed for safe usage against crack growth. Mechanical properties of sisal-®bre composites measured from tests quite often disagree with the rule of mixtures. A full explanation can only be obtained if the interface strength and the failure mechanisms are known. Further work is needed particularly to explain `hybrid' e€ects in sisal/glass composites. Economical processing methods must be developed for the composites because of the very low price of sisal ®bres. The relationship between mechanical properties and processing methods should be established. New applications should be found for sisal-®brebased composites. Hybrid ®bre composites with sisal and other ®bres rather than glass may open up new applications. For example, from the economics point of view, sisal ®bres may be hybridised with carbon or aramid ®bres to reduce the costs of these expensive ®bres reinforced composites whilst maintaining their good mechanical performance. Recycling (including burning) characteristics and methods of sisal-®bre-reinforced composites are important aspects of this new material but there are very few published data to date. Recycling is an attractive future research direction that will provide socio-economic bene®ts.

[7] [8] [9] [10]

[11] [12] [13]

[14] [15] [16] [17] [18]

Acknowledgements Y. Li would like to thank the University of Sydney for an Overseas Postgraduate Research Scholarship (OPRS) and an International Postgraduate Award (IPA). Y.-W. Mai also thanks the Australian Research Council for the continuing support on the ®bre/matrix interface research project.

[19] [20] [21] [22]

References [1] Murherjee PS, Satyanarayana KG. Structure and properties of some vegetable ®bres, part 1. Sisal ®bre. Journal of Materials Science 1984;19:3925±34. [2] Chand N, Tiwary RK, Rohatgi PK. Bibliography resource structure properties of natural cellulosic ®bres Ð an annotated bibliography. Journal of Materials Science 1988;23:381±7. [3] Bisanda ETN, Ansell MP. Properties of sisal±CNSL composites. Journal of Materials Science 1992;27:1690±700. [4] Wilson PI. Sisal, vol. II. In Hard ®bres research series, no. 8, Rome: FAO, 1971. [5] Rowell RM. In: Rowell RM, Schultz TP, Narayan R, editors. Emerging technologies for materials & chemicals for biomass, ACS Symposium Ser, 476, 1992. p. 12. [6] Joseph K, Thomas S, Pavithran C. E€ect of chemical treatment

[23] [24] [25] [26] [27] [28]

2053

on the tensile properties of short sisal ®bre-reinforced polyethylene composites. Polymer 1996;37:5139±49. Chand N, Hashmi SAR. Metals Materials and Processes 1993;5:51. Bisanda ETN, Ansell MP. The e€ect of silane treatment on the mechanical and physical properties of sisal±epoxy composites. Composites Science and Technology 1991;41:165±78. Li H, Zadorecki P, Flodin P. Cellulose ®bre±polyester composites with reduced water sensitivity. (1) Chemical treatment and mechanical properties. Polymer Composites 1987;8:199±207. Chand N, Sood S, Rohatgi PK, Sayanarayana KG. Resources, structure, properties and uses of natural ®bres of madhya pradesh. Journal of Scienti®c and Industrial Research 1984;43:489± 99. Bisanda ETN. The manufacture of roo®ng panels from sisal ®bre reinforced composites. Journal of Materials Processing Technology 1993;38:369±80. Chand N, Satyanarayana KG, Rohatgi PK. Mechanical characteristics of sunhemp ®bres. Indian Journal of Textile Research 1986;11:86±9. Hornsby PR, Hinrichsen E, Tarverdi K. Preparation and properties of polypropylene composites reinforced with wheat and ¯ax straw ®bres. Part 1, ®bre characterisation. Journal of Materials Science 1997;32:443±9. Satyanarayana KG, Sukumaran K, Mukherjee PS, Pavithran C, Pillai SG. Natural ®bre-polymer composites. Cement & Concrete Composites 1990;12:117±36. Chand N, Hashmi SAR. Mechanical properties of sisal ®bre at elevated temperatures. Journal of Materials Science 1993;28:6724±8. Chand N, Joshi SK. Temperature dependence of dielectric behaviour of sisal ®bre. Journal of Materials Science Letters 1994;13:156±8. Yang GC, Zeng HM, Zhang WB. Thermal treatment and thermal behaviour of sisal ®bre. Cellulose Science and Technology 1995;3:15±19. Zhou LM, Mai Y-W, Ye L, Kim JK. Techniques for evaluating interfacial properties of ®bre-matrix composites. Key Engineering Materials 1995;104±107:549±600. Kim JK, Zhou LM, Mai Y-W. Interfacial debonding and ®bre pull-out stresses. Part III. Interfacial properties of cement matrix composites. Journal of Materials Science 1993;28:3923±30. Yang GC, Zeng HM, Li JJ, Jian NB, Zhang WB. Relation of modi®cation and tensile properties of sisal ®bre. Acta Scientiarum Naturalium Universitatis Sunyatseni 1996;35:53±7. El-Naggar AM, El-Hosamy MB, Zahran AH, Zondy MH. Surface morphology/mechanical/dyeability properties of radiationgrafted sisal ®bres. American Dyestu€ Reporter 1992;81:40±4. Sabaa MW. Thermal degradation behaviour of sisal ®bres grafted with various vinyl monomers. Polymer Degradation & Stability 1991;32:209±17. Chand N, Verma S, Khazanchi AC. SEM and strength characteristics of acetylated sisal ®bre. Journal of Materials Science Letters 1989;8:1307±9. Singh B, Gupta M, Verma A. In¯uence of ®bre surface treatment on the properties of sisal±polyester composites. Polymer Composites 1996;17:910±8. Kim JK, Lu S, Mai Y-W. Interfacial debonding and ®bre pullout stresses Part IV: In¯uence of interface layer on the stress transfer. Journal of Materials Science 1994;29:554±61. Yang GC, Zeng HM, Li JJ. Study of sisal ®bre/phenol formaldehyde resin composites. Fibre Reinforced Plastics/Composites 1997;3:12±14. Sanadi AR, Prasad SV, Rohatgi PK. Sunhemp ®bre-reinforced polyester Part 1 Analysis of tensile and impact properties. Journal of Materials Science 1986;21:4299±304. Pavithran C, Mukherjee PS, Brahmakumar M, Damodaran AD.

2054

[29] [30] [31] [32]

[33]

[34]

[35] [36]

[37] [38] [39]

[40]

[41]

[42] [43] [44] [45] [46] [47]

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055 Impact properties of natural ®bre composites. Journal of Materials Science Letters 1987;6:882±4. White NM, Ansell MP. Straw-reinforced polyester composites. Journal of Materials Science 1983;18:1549±56. Pavithran C, Mukherjee PS, Brahmakumar M, Damodaran AD. Impact performance of sisal±polyester composites. Journal of Materials Science Letters 1988;7:825±6. Gordon JE, Jeronimidis. Philosophical Transactions. Royal Society of London, Series A: Mathematical and Physical Sciences 1980;294:545. Rong MZ, Li RKY, Ng CN, Tjong SC, Mai Y-W, Zeng HM. E€ect of ®bre pretreatment on the impact fracture toughness of sisal ®bre reinforced polymer composites. In Proc. First Asian± Australian Conference on Composite Materials (ACCM- 1), 7±9 October 1998, Oseka, Japan. p 433-1±4. Rong MZ, Li RKY, Ng CN, Tjong SC, Mai Y-W, Zeng HM. E€ect of water absorption on the impact behaviour of short sisal ®bre reinforced polymer composites. In Proc. First Asian±Australian Conference on Composite Materials (ACCM-1), 7±9 October 1998, Oseka, Japan. p 438-1±4. Joseph K, Thomas S, Pavithran C. Dynamic mechanical properties of short sisal ®bre reinforced low density polyethylene composites. Journal of Reinforced Plastics and Composites 1993;12:139±55. Joseph K, Thomas S, Pavithran C, Brahmakumar M. Tensile properties of short sisal ®bre-reinforced polyethylene composites. Journal of Applied Polymer Science 1993;47:1731±9. Joseph K, Kuriakose B, Premalatha CK, Thomas S. Melt rheological behaviour of short sisal ®bre reinforced polyethylene composites. Plastics Rubber & Composites Processing & Applications 1994;21:237±45. Joseph K, Thomas S, Pavithran C. Viscoelastic properties of short-sisal-®bre-®lled low-density polyethylene composites: e€ect of ®bre length and orientation. Materials Letters 1992;15:224±8. Karmaker AC. E€ect of water absorption on dimensional stability and impact energy of jute ®bre reinforced polypropylene. Journal of Materials Science Letter 1997;16:462±4. Garkhail S, Heijenrath R, Oever M, van den, Bos H, Peijs T. Natural-®bre-mat-reinforced thermoplastic composites based on ¯ax ®bres and polypropylene. In Scott M, editor. Proceedings of ICCM-11, vol. II, Fatigue, fracture and ceramic matrix composites, Gold Coast, Australia, 14±18 July 1997. p. 794±803. Park BD, Balatinecz JJ. E€ect of impact modi®cation on the mechanical properties of wood-®bre thermoplastic composites with high impact polypropylene (HIPP). Journal of Thermoplastic Composite Materials 1996;9:342±63. Hornsby PR, Hinrichsen E, Tarverdi K. Preparation and properties of polypropylene composites reinforced with wheat and ¯ax straw ®bres. Part II. Analysis of composite microstructure and mechanical properties. Journal of Materials Science 1997;32: 1009±1015. Yang GC, Zeng HM, Li JJ. The study of technology and properties of composites for sisal ®bre reinforced PVC. Fibre Reinforced Plastics/Composites 1995;6:22±6. Manikandan KC, Nair SMD, Thomas S. Tensile properties of short sisal ®bre reinforced polystyrene composites. Journal of Applied Polymer Science 1996;60:1483±97. Kalaprasad G, Joseph K, Thomas S. Theoretical modelling of tensile properties of short sisal ®bre±reinforced low-density polyethylene composites. Journal of Materials Science 1997;32:4261±7. Broutman LJ, Krock RH. Modern composite materials. Reading (MA): Addison Wesley, 1967. Hirsch TJ. Modulus of elasticity of concrete a€ected by elastic moduli of cement paste matrix and aggregate. In Proceedings of the American Concrete Institute 1962;59(12):427±51. Rosen BW. Fibre composite materials. Metals Park (OH): American Society for Metals, 1965. p. 58

[48] Nielson LE. Mechanical properties of polymers and composites, vol. 2. New York: Marcel Dekker, 1974. [49] Cox HL. The elasticity and strength of paper and other ®brous materials. British Journal of Applied Physics 1952;3:72±9. [50] Bowyer WH, Bader MG. On the re-enforcement of thermoplastics by imperfectly aligned discontinuous ®bres. Journal of Materials Science 1972;7:1315±21. [51] Paul A, Thomas S, Pavithran C. Electrical properties of natural®ber reinforced low density polyethylene composites: a comparison with carbon black and glass-®bre ®lled low density polyethylene composites. Journal of Applied Polymer Science 1997; 63:247±66. [52] Joseph K, Thomas S, Pavithran C. E€ect of ageing on the physical and mechanical properties of sisal-®bre-reinforced polyethylene composites. Composites Science and Technology 1995;53:99±110. [53] Prasantha RP, Kumar ML, Amma G, Thomas S. Short sisal ®bre reinforced styrene±butadiene rubber composites. Journal of Applied Polymer Science 1995;58:597±612. [54] Varghese S, Kuriakose B, Thomas S. Stress relaxation in short sisal-®bre-reinforced natural rubber composites. Journal of Applied Polymer Science 1994;53:1051±60. [55] Kumar RP, Thomas S. Short ®bre elastomer composites: e€ect of ®bre length, orientation, loading and bonding agent. Bulletin of Materials Science 1995;18:1021±9. [56] Kumar PR, Thomas S. Tear and processing behaviour of short sisal ®bre reinforced styrene butadiene rubber composites. Polymer International 1995;38:173±82. [57] Varghese S, Kuriakose B, Thomas S. Short sisal ®bre reinforced natural rubber composites: high-energy radiation, thermal and ozone degradation. Polymer Degradation & Stability 1994;44:55± 61. [58] Varghese S, Kuriakose B, Thomas S, Koshy AT. Mechanical and viscoelastic properties of short ®bre reinforced natural rubber composites: e€ects of interfacial adhesion, ®bre loading, and orientation. Journal of Adhesion Science & Technology 1994; 8:235±48. [59] Varghese S, Kuriakose B, Thomas S, Premalatha CK. Rheological behaviour of short sisal ®bre reinforced natural rubber composites. Plastics Rubber & Composites Processing & Applications 1993;20:93±9. [60] Singh M, Garg M. Gypsum-based ®bre-reinforced composites: an alternative to timber. Construction & Building Materials 1994;8:155±60. [61] Hernandez OF, Oteiza IDe, Villanueva L. Fire behaviour of sisal short ®bres reinforced gypsum. In Miraveter A, editor. Proceedings of the Ninth International Conference on Composite Materials (ICCM/9), vol. 5, composites behaviour, Madrid, 12±16 July 1993, p. 951±7 [62] Coutts RSP, Warden PG. Sisal pulp reinforced cement mortar. Cement & Concrete Composites 1992;14:17±21. [63] Mwafongo FG. Sisal-reinforced cement. Batiment International/ Building Research & Practice 1987;20:241±2. [64] Hernandez OF, Oteiza IDe, Villanueva L. Experimental analysis of toughness and modulus of rupture increase of sisal short ®bre reinforced hemihydrated gypsum. Composite Structures 1992; 22:123±37. [65] Bessell TJ, Mutuli SM. The interfacial bond strength of sisal± cement composites using a tensile test. Journal of Materials Science Letters 1982;1:244±6. [66] Swift DG. Sisal-cement composites and their potential for rural Africa. In: Marshall IH, editor. Composite structures 3, Elsevier Applied Science Publisher. London/New York: p. 774±87. [67] Kalapradad G, Joseph K, Thomas S. In¯uence of short glass ®bre addition on the mechanical properties of sisal reinforced low density polyethylene composites. Journal of Composite Materials 1997;31:509±26.

Y. Li et al. / Composites Science and Technology 60 (2000) 2037±2055 [68] Kalaprasad G, Thomas S, Pavithran C, Neelakantan NR, Balakrishnan S. Hybrid e€ect in the mechanical properties of short sisal/glass hybrid ®bre reinforced low density polyethylene composites. Journal of Reinforced Plastics and Composites 1996;15:48±73.

2055

[69] Yang GC, Zeng HM, Jian NB, Li JJ. Properties of sisal ®bre/ glass ®bre reinforced PVC hybrid composites. Plastics Industry 1996;1:79±81. [70] Kim JK, Mai Y-W. Engineered interfaces in ®bre reinforced composites. Oxford: Elsevier, 1998.