1.10 Natural Organic Fibers HANS LILHOLT Risù National Laboratory, Roskilde, Denmark and J. MARK LAWTHER Royal Veterinary and Agricultural University, Copenhagen, Denmark 1.10.1 HISTORIC BACKGROUND

1

1.10.2 INTRODUCTION

4

1.10.2.1 Classification 1.10.2.2 Natural Organic Fibers with High Strength 1.010.2.2.1 Bond energy 1.010.2.2.2 Structure and strength 1.010.2.2.3 Practical strength 1.010.2.2.4 Stiffness 1.10.2.3 Morphological Characteristics Specific to Plant Fibers

4 6 6 7 7 10 11

1.10.3 CHARACTERISTICS OF NATURAL ORGANIC FIBERS 1.10.3.1 Basic Chemical Composition 1.10.3.2 Specific Types of Natural Organic Fibers 1.10.3.3 Natural Organic Fibers and Synthetic Fibers

12 12 14 14

1.10.4 SURFACE CHARACTERIZATION AND MODIFICATION

18

1.10.4.1 Basic Physical±Chemical Concepts 1.10.4.2 Approaches to Surface Modification

18 18

1.10.5 PROCESSING TO COMPOSITES 1.10.5.1 1.10.5.2 1.10.5.3 1.10.5.4

20

Forming Fibers into Semiproducts Processing Limits Matrix Polymers Processing Techniques

20 20 21 22

1.10.6 POTENTIAL APPLICATIONS AND OUTLOOK

22

1.10.7 REFERENCES

23

1.10.1

HISTORIC BACKGROUND

old site in Turkey (Chawla, 1998), the linen being the cellulosic fibers of the flax plant. Natural as well as synthetic fibers have been and are being used both as fibers and fiber assemblies and as reinforcements in composites. The use of natural fiber directly is the oldest application in the form of textiles and related products. Therefore, an extensive

Natural organic fibers have been around for a very long time, from the beginning of life on earth if their mere existence is used as criterion, from about the earliest prehistoric periods if archaeological evidence is used as basis, since fabrics of linen have been found in a 9000-year 1

2

Natural Organic Fibers Table 1 Natural fibers (1930±1960). Fiber

Stiffness E (GPa)

Density r (g cm73)

E/r

103 57 85 51 90 159 70 70

1.50 1.50 1.50 1.50 1.50 2.60 2.54 2.70

69 38 57 34 60 61 28 26

Flax Hemp (dry) Hemp (dried under tension) Ramie (dry) Ramie (dried under tension) Asbestos (chrysotile) Glass (E-type) Aluminum

experience and practical knowledge relates to the textiles and the textile industry. Besides being of vast importance for these areas of application, much of the knowledge could beneficially be introduced in the other and newer area of application, namely the use of fibers as reinforcements in composites. The textile and the reinforcement areas have both developed separately and on different timescales and inspired and supported each other. The textile area has used natural fibers directly, for many different types of products; the main focus is on the appearance, esthetics, and physical performance, e.g., the temperature and moisture control of many types of textile products. The focus has been less on the mechanical aspects such as strength, although the textiles need to resist the load of their use. A related application with more focus on the mechanical performance of natural fibers has been their use in ªstrengtheningº of soft, brittle materials such as clay, which was mixed with straw to reduce the practical crack sensitivity in e.g., walls of mud-built houses. These microstructurally coarse and primitive composite materials were probably the first deliberate use of natural fibers in combination with a matrix. A direct use of the strength of natural fibers is in lines, ropes, and other one-dimensional products; the many uses include early suspension bridges for on-foot passage of rivers and rigging for naval ships in early times and into the nineteenth century. The importance of plant fibers for ropes and rigging in naval ships is illustrated by the rope testing recorded by Samuel Pepys, secretary (clerk of the acts) to the Navy Board in England, in his diary for March 2, 1663 (Pepys, 1967). Here the established Riga hemp is compared to a new sort of Indian grass (sisal?) in a direct tensile test of the ropes. The protection of the fibers against environmental degradation, i.e., rotting, is important and is done by the use of natural tar for impregnation. Pepys reflects on the ease of such impregnation in relation to the various rope qualities. It was thus already at that time clear to the practical

users of natural fibers that strength, environmental stability, and interfacial aspects all were of importance for efficient use. Early attempts to use composites as structural materials for load-bearing applications were made in the late 1930s, probably without recognition of the composite principles and the importance of fibers as the reinforcing part of composites. The wish to minimize structural (load-bearing) weight in aircrafts led to the development and study of materials based on polymers and natural fibers. Such fibers were cellulose, in particular flax, as well as asbestos and glass, and the matrix was a phenolic resin. The practical components and their performance in service led slowly to appreciation of the microstructural aspects, in particular fiber orientation, as parameters of importance for achieving sufficient stiffness and strength. It was demonstrated that a delta wing for an aircraft could perform at least as well as its metallic counterpart. This also indicated clearly that composite materials should have better specific stiffness (stiffness/density) than metal alloys if worthwhile structural weight savings were to be obtained. Such improvement in material properties could be achieved by development of high-stiffness, lightweight fibers and by arranging for a good and controlled fiber orientation so that fibers were preferentially aligned with the (major) load directions. It is noteworthy that the (early) natural fibers have rather high specific stiffnesses, better than that of aluminum, but in the creation of a composite with nonaligned fibers a great deal of this potential is lost. Table 1 shows natural fibers from the period 1930±1960 in comparison with glass fibers and aluminum. The research effort in the 1950s was generally on strong, stiff, and light fibers, and new fibers appeared such as high-strength glass fibers, whiskers based on ceramics and inorganics, carbon/graphite fibers, and later polyaramid fibers. In all these developments a fundamental understanding of the molecular and atomic scale of the material

Historic Background

3

Table 2 Examples of applications of natural fibers. Time period Prehistoric/historic

Present

Future

Material/product Textiles Ropes Primitive composites Textiles Early composites (e.g., inner door in cars) Insulation; tissue Plates, boards, laminates Furniture Advanced composites Improved interface Stronger fibers Load-bearing products More versatile products Mixture of short and long fibers

micro- and nanostructure was of great significance in elucidating the principles of high stiffness and strength and low density. Such research efforts were apparently not directed towards natural fibers; these were assumed to have properties and characteristics given by nature and likely to be variable. With these often rather nonuniform characteristics of natural fibers, due to plant type, growth conditions, weather, and handling after harvesting, it is perhaps not surprising that the research efforts for better reinforcement fibers were not directed towards natural fibers but rather towards the less variable and perhaps structurally simpler inorganic materials mentioned above. The increasing development and use of composites based on polymer matrices and synthetic fibers have dominated the period since the 1960s. Recently, a renewed interest has developed in the use of natural fibers in composites. This relates to (perhaps) the nature of the composite principles, to the potential for improvement of natural fiber properties, and to the increased concern of society towards nature and the environment. A few observations are given to these three aspects. The basic principle of composites is that of a bundle of strong fibers held together by a matrix of low shear resistance, such that stresses of a crack are not focused on the fibers but rather are blunted; this makes the crack less harmful to the survival of the composite. A special and important mechanism is the pullout of fibers from the matrix under energy consumption. Both these concepts give a quasiductile behavior to the materials systems, where in certain cases both the fibers and the matrix could be brittle themselves. An example is the composite

Dimensionality

Matrix

2 1 2

+

2 2 3 2 2±3 1 1±2 1, 2, 3 3

Binder

+ + + + + + +

+

+

system of cellulosic fibers and melamine polymer, where the brittle melamine is made quasiductile as a composite. The potential for improvement of the natural fibers is based on an increasing knowledge of their internal structure, such as cellulose content, crystallinity, and microstructural perfection, all of which could lead to better, stronger natural fibers. The surfaces of natural fibers are better understood, and are being studied with respect to improved surface and interface characteristics in the composites. The potential for improved fiber configuration, in particular orientation and alignment of fibers, is not developed to any great extent. The concern for nature and its status and development has dramatically increased the interest in resources, renewable materials, and biodegradable/recycling aspects of nature, and has thus focused interest on the nonfood applications of natural resources. It may not necessarily be the best way to protect and maintain nature, because this will interact with nature's life by using the resources for ªotherº purposes. Uses of natural fibers and related composites have been plentiful over the long period of prehistory and history, and the materials and/ or products have often been used without recognizing their composite nature, in whatever wide context this may be interpreted. To give an indication of the various types of materials and products and their characteristics, the following list (Table 2) of examples is presented, relating to old, present, and future uses. The dimensionality refers to the arrangement of the natural fibers in the material or product; the matrix is a (often) polymer holding the fibers together with a minimum of porosity, ideally approaching zero, while the binder is a (normally) polymer

4

Natural Organic Fibers

Figure 1

Classification of fibers according to origin. Table 3 Chemical classification.

Chemical structure Main group

Subgroup

Examples

Cellulose

Polypeptide

Fibroin (simple) Fibrous proteins (complex)

giving some connection between fibers without constituting a fully dense phase filling the space between the fibers.

1.10.2

INTRODUCTION

The natural organic fibers are basically characterized by the same parameters and properties as all other fibers, and fundamentally they are expected to offer the same reinforcing effects in a (polymer) matrix as their synthetic counterparts, although the efficiency and level of reinforcement may be different. While the traditional, synthetic fibers such as glass fibers have very well defined geometry, morphology, surface characteristics, mechanical, physical, and chemical properties as well as relative easy composite fabrication aspects, the natural organic fibers have the same principal characteristics but with an often large degree of nonuniformity of the various parameters. This situation leads to a limited characterization profile of most natural fibers today, giving a rather large variability in fiber characteristics and behavior, and thus much larger difficulties in fabricating composite materials of an acceptable reproducibility. There is thus a need for improved characterization, sorting of fibers in quality groups, and efficient fabrication routes related to fiber quality. The natural fibers show nonuniformity in

bast fibers leaf fibers stem fibers seed fibers wood fibers fruit fibers silk wool hair

most characteristics: chemical composition, crystallinity, surface properties, diameter, cross-sectional shape, length, strength, and stiffness. This poses both problems of quality grouping and difficulties in using traditional composite theory, which is in nearly all cases based on (nearly) uniform (single-valued) parameters of the above mentioned types.

1.10.2.1

Classification

Natural organic fibers are derived from either plant or animal sources. The majority of useful natural textile fibers are plant derived with the notable exceptions of wool and, to a lesser extent, silk. The first degree of classification is therefore facile (see Figure 1). The natural organic fibers may be classified according to several principles and treated from various points of view. The basic principles are chemical and physical structure, natural origin and plant classification, or fiber length. The practical points could be availability, processing aspects, need for natural composites, or actual applications. The classification according to chemical nature is relatively simple, but contains many specific types in each group. The two main chemical structures are celluloses and proteins. Table 3 shows this classification, partly based on Evans (1948).

Introduction

5

Fibres Synthetic

Natural Mineral

Animal

Wool

Silk

Natural polymer

Metal

Ceramic

(Other)

Asbestos

Hair

Synthetic polymer

Vegetable

Bast

Seed

Leaf

Grass Stem

Wood

Flax

Sisal

Cotton

Soft wood

Reed canary grass

Hemp

Banana

Coir

Hard wood

Cereal straw (wheat straw) (rye straw)

Jute

Manila hemp (Abaca)

Oil palm

Ramie

Kenaf

Figure 2 Fibers according to origin.

Fibres Continuous fibres

Short fibres

Yarn Woven

Knitted

Mats Braided

(filament winding)

Felt

Non-woven

Figure 3 Fibers according to length.

The cellulose chain is composed of glucose units joined by oxygen bridges. The polypeptides are composed of a basic chain of amide (-COÐNH-) units with ÐCH3 side radicals in the complex subgroup. The classification according to natural and plant origin is shown in Figure 2 which is based on various sources of classification, often made for practical reasons of a selected group of fibers. This classification takes the two main groups of natural fibers and synthetic or man-made fibers as the basis. There are ªoverlapsº in e.g., ceramic fibers which may be of natural origin, e.g., asbestos, or of man-made origin, e.g., oxide fibers; organic fibers are both derived from their natural origin, e.g., bast fibers, or are of man-made origin, e.g., regenerated cellulose (rayon). The man-made synthetic polymer fibers in the overview represent the enormous field of (synthetic) textile fibers. The fibers, natural or synthetic, are in this context treated with respect to their potential as reinforcements in composites. This does

not exclude their other properties and their use in other contexts, processes, materials, or products. The classification according to length is of some importance in relation both to composites (their properties and performance) and to handling of the natural fibers from plant to material. The traditional group with respect to length is mainly based on the use of (natural) fibers in the textile industry, but is also useful for handling and performance of natural fibers in composites. Figure 3 is based on the simplified classification used by Chawla (1998). This overview also indicates (in the lower lines) the various forms of semiproducts used in composite manufacturing, where filament winding is included because this is a direct way of handling yarns (fiber bundles) for composite fabrication. The classification shown in Figure 3 only distinguishes between continuous and noncontinuous (ªshortº) fibers. The actual length of natural fibers is much more variable, depending

6

Natural Organic Fibers Table 4 Dimensions of natural fibers. Vegetable fibers

Average length (mm)

Average diameter (mm)

Aspect ratio

33.0 25.0 3.0

19 25 20

1700 1000 150

1.4

15

90

18.0

20

900

2.7

14

200

3.3 1.0

33 20

100 50

Bast fibers flax (single) hemp (single) jute (single) Stem fibers straw Seed fibers cotton (single) Grass fibers bamboo Wood fibers soft hard Source: Olesen and Plackett, 1999.

on fiber type and fiber handling. The actual dimensions are listed in Table 4 as length, diameter, and aspect ratio = length/diameter. The values refer to the natural fibers as they are made available more or less directly from their natural origin; they have not been through a composite processing route. The wide range seen for natural fibers gives potentially many ways of using the fibers in composites, ranging from aligned fiber, high-strength composites to random fiber, moderate-strength composites. The possibility of shortening the long fibers for specific processing requirements is also an option. The potentially very high aspect ratios for some of the natural fibers make them suitable for continuous fiber handling, such as filament winding or pultrusion. The natural fibers are being, and will probably increasingly be, used for many different purposes and reasons. The various routes to the exploitation of natural fibers will be briefly discussed. Their natural and plentiful availability is in itself a strong drive to use them, as history shows, and increasingly sophisticated and advanced applications will be found within the field of composites and elsewhere. Their basically good properties, dictated by their chemical and physical structural characteristics, present them as promising reinforcements in composites. The practical application of natural fibers for load-bearing, structural composites relies strongly on efficient processing routes from plant to fiber to semiproduct and to material and product. Efficient, in terms of materials and energy, manufacturing technology is in great need for a wider use of natural fiber composites, both in their own right and as an

expansion of the spectrum of materials available for structural applications and products. The need for natural composites, partly to replace existing materials, but probably more to supplement them, is also governed by a political and society driven interest in nature's state of health at present and in the future.

1.10.2.2

Natural Organic Fibers with High Strength

Natural fibers have been used in composites and composite-like materials because experience has shown that such fibers were or could be strong. To achieve strong composites it is important to start from strong fibers; these are not themselves sufficient to achieve strong composites, because good bonding to the matrix and good orientation are also needed. Nevertheless, strong fibers are a prerequisite to strong composites, and it is therefore of considerable interest to elucidate the potentials and limitations of strong natural organic fibers. This approach is similar to the concept taken by Kelly and Macmillan (1986) in establishing the limits for strong solids.

1.10.2.2.1

Bond energy

The chemical structure of natural organic fibers is based on the cellulose molecule and the protein molecule. Since these chemical compounds contain covalently bonded atoms of the types C, N, O, and H, the bond energies of relevant combinations of these atoms form the basis for the (potential) high strength of

Introduction Table 5 Bond energies for atom±atom combinations. Energy (kJ mol71)

Bond C C C C C C

(aliph.)ÐC (aliph.) (aliph.)ÐO (aliph.)ÐN (arom.)ÐC (arom.) (arom.)ÐO (arom.)ÐN

347 389 343 410 448 460

Source: Kelly and Macmillan, 1986.

materials/fibers composed of these atoms and molecules. The bond energies for both aliphatic and aromatic atom combinations are given in Table 5, which is based on the data in Kelly and Macmillan (1986). The bond energies form one basic part in estimations of maximum/theoretical strength and stiffness. Also important are the deformation mechanisms of the assembly of bonds in a given molecule and the density of bonds in the material. The deformations refer to stretching of the atom-to-atom bonds and to changing of the bond angle, and the density refers to the number of molecular chains per cross-sectional area transverse to the loading direction. An estimate of the force to stretch and break a carbon±carbon bond has been given by Kelly and Macmillan (1986). The aliphatic CÐC bond has an energy of 347 kJ mol71, or 5.8610719 J atom71 (Table 5), and from this the force to break the CÐC bond is calculated to be F = 6.1610719 N. This value F, in combination with bond density B, allows a simple estimate of strength when this is based on bond breaking: sth = F6B.

1.10.2.2.2

Structure and strength

The chemical structure governs the bond types and the bond density in the different compounds/materials, and thus the values of theoretical strength, which can be estimated according to the models used over the years by various authors. For the natural organic fibers cellulose is the building block for most agrofibers, and it is also the most extensively studied molecule in relation to strength and stiffness. A comparison will be made to the basic CÐC chain, which is included in a complex way in the cellulose chain, and which is the backbone of diamond, graphite, and polyethylene. Polyethylene is based on the simple CÐC chain, and the PE fibers contain these chains in alignment; the crystallinity is established by the

7

parallel chains, either separate chains or backfolded chains; PE fibers are described in Chapter 1.09, this volume. Graphite is based on hexagonally, planarly arranged C atoms; this hexagonal pattern can be viewed as made up of CÐC chains, which are linked from chain to chain. The diamond is based on a three-dimensional lattice of C atoms, this crystallographic structure can be viewed as composed of CÐC chains with linking between chains in a threedimensional pattern. Cellulose is based on the glucose molecule, and the basic unit is made up of two glucose molecules linked by an oxygen atom. The unit cell is composed of alternate glucose rings, related to each other by a 180 8 rotation, and has a repeat distance of 1.03 nm (Evans, 1948). The cellulose appears in two crystallographic forms, cellulose I with parallel chains, and cellulose II with antiparallel chains. Calculation of the theoretical strength and stiffness of materials is based on the ideal (perfect) structure, which in most cases is crystalline. This allows a number of structural parameters and atom±atom interactions to be accounted for. Several such calculations have been made since the 1930s; and the basic ideas and methods are collected and described critically by Kelly and Macmillan (1986). The structural parameters are the atom-toatom force and the density of bonds; these refer to number of bonds, i.e., chains, per cross-sectional area taken perpendicular to the fiber axis, which in most cases will be the loading direction of the individual fibers in composites. Relevant values for the force and the chain density are collected in Table 6, and are taken from various sources; the values for F shown in parentheses are implied from other data because they were not given in the reference. It is seen that the atom-to-atom force is comparable for the four materials, while the chain density varies more, and this variation shows up in the theoretical strength values. It is clear that the theoretical strength for these (and other) materials are extremely high compared to any experimental values. This discrepancy is in general caused by imperfections of the (ideal) structures, but they also indicate the potentially very large scope for improvement of the fibers and their structure in order to achieve stronger materials.

1.10.2.2.3

Practical strength

The strength of real fibers, natural and polymeric, is normally quite low compared to any theoretical estimate; the strength is also

8

Natural Organic Fibers Table 6 Bond density and theoretical strength. Area/chain Distance between Chains/area Force chains (nm) (1018 m72) (1079N) (nm)2

Material (loading direction)

sth GPa

Reference

Diamond (5111>)

0.055

0.23

18.2

6.1

111

Kelly and Macmillan (1986) p. 8

Graphite (CÐC in basal plane)

0.071

0.27

14.0

8.3

116

Perepelkin (1972)

Polyethylene

0.182

0.43

5.5

6.1

34

(CÐC chain)

0.182

0.43

5.5

(3.4±3.6) 19±20 Perepelkin (1966)

Cellulose (along chain)

0.323

0.57

3.1

(3.9±6.1) 12±19 Perepelkin (1966)

Cellulose triacetate (along chain)

0.641

0.80

1.56

(3.8±6.1) 6±9.5 Perepelkin (1966)

Kelly and Macmillan (1986) p. 22

Table 7 Parameters and strength for polyethylene and celluloses. Parameter to,s U0, kJ mol±1 (mechanical fracture) g, kJ/mol/MPa sth, GPa smax.att, GPa sexp., GPa spract., GPa a

Polyethylene ± 250±293 0.013±0.015 19±20 13±14 0.8±0.9 ±

Cellulose

Cellulose triacetate

10713 167 0.009±0.014 12±19 6±10 5.2a 1.3±1.4 0.3±0.5

± (167) 0.018±0.028 6±9.5 3±5 0.5±0.6 ±

100% crystallinity, 100% orientation, degree of polymerization 400.

dependent on temperature and rate of loading (strain rate). Typical strength values at temperatures below room temperature show, by extrapolation, strength values at 0 K which are about two times higher than the room temperature values (Perepelkin, 1966). The reference level for strength is the theoretical value, which refers to the ideal structure of the material, natural fiber, polymer, or another substance. Several methods have been used to calculate the theoretical strength, one of which is mentioned above. Perepelkin (1966) based a calculation on the expression for the long-term strength, developed by Zhurkov (1965): s = U0/g 7 (RT/g) ln t/t0

(1)

where U0 is the bond dissociation energy, g is a structural parameter of the polymer7molecular chains, t is the time to rupture, and t0 is a constant equal to the period of thermal vibrations; R is the gas constant and T is the absolute temperature. This equation defines

the theoretical strength for an ideal structure at T = 0 8K and deformed to a rupture time of t = t0 : sth = U0/g

(2)

A high bond energy U0 and an ideal structure, i.e., low g, lead to high theoretical strength. Details of the calculation and the numerical data involved are given by Perepelkin (1966, 1970), from where data for polyethylene and cellulose are collected and presented in Table 7. The strength equation (1) can be used to evaluate the maximum attainable strength, which, somewhat arbitrarily, is taken at room temperature and moderate deformation rate; the values T = 293 K and t = 10 s are used to obtain smax.att. = sth 7 80/g

which with Equation (2) gives smax.att. = sth (1 7 80/Uo)

(3)

Introduction These strengths will be the maximum attainable values for an ideal structure for the polymer chain, i.e., fully crystalline, fully orientated along the fiber axis, and with a high degree of polymerization. As an example, quoted by Perepelkin (1966), the strength of hydrated cellulose fibers of 100% crystallinity, 100% orientation, and a degree of polymerization of 380± 400, is about 5200 MPa, which is obtained by extrapolation of experimental data. The maximum attainable strengths calculated from Equation (3) are listed in Table 7; these are, as expected, close to the values listed by Perepelkin (1970), while they are somewhat lower than the values listed by Perepelkin (1966). In general, the maximum attainable strengths are reasonably consistent, and they are between 50 and 70% of the theoretical strengths. The maximum recorded strength for many polymer fibers under ªnormalº test conditions are listed by Perepelkin (1966) and selected values are given in Table 7 for polyethylene and celluloses. The high value of 5.2 GPa for cellulose refers to a near-ideal structure and is close to the estimated values for maximum attainable strength, thus agreeing with this estimate. The other experimental values are significantly lower than the estimates, and are of the order of 15% of the maximum attainable strength. These experimental values refer to fibers probably tested directly as fibers. In composites the fibers act as an assembly, and the practical strength values for cellulose fibers in the form of flax and jute have been estimated (Toftegaard and Lilholt, 1999) from composite experimental strength values and composite theory by back-calculation to estimate the effective/practical fiber strengths for in situ fibers. Such values are, at present, even lower than the free fiber strengths. From the data in Table 7 it is clear that the structure and its imperfection for real fibers, both natural and polymeric, are of great significance for the strength which can be attained, and that there is great scope for improvements of the fiber structures, whether the ªgoalº is the maximum attainable strength or even the theoretical strength. Some of the imperfections which are responsible for the low strength values are briefly discussed by Perepelkin (1966) and shall be mentioned here to indicate ways of improving the structure of natural organic fibers. Within the structure of natural (and other) fibers, regions with noncrystalline structure, i.e. amorphous regions, are potential ªweak points,º mainly because the number of chains per cross-sectional area is lower, resulting in low local strength.

9

The strength of real fibers is strongly dependent on the orientation of their structural, crystalline elements; high strengths are normally achieved with alignments within about +5 8 of perfect orientation. Strength estimates for fully oriented and fully random (isotropic) structures, respectively, indicate that the strength of an isotropic structure is 0.17 of the strength of an ideally oriented structure, while in practice the strength of real structures is 0.30±0.50 of the ideal strength. The fiber structure may, and most likely will, contain various imperfections and defects; cracks will formally cause local stress concentrations which will reduce the strength. The severeness of cracks depends on the anisotropy of the fiber structure, and for high anisotropy the stress concentration of transverse cracks is low and thus not a decisive factor for fiber strength. This is supported by experiments on, e.g., wood cellulose, and is well known for fiber reinforced composites. When anisotropy decreases, transverse cracks become more serious, i.e., for not well oriented fibers. The degree of polymerization will affect the strength of highly oriented fibers; in general the strength is inversely proportional to the degree of polymerization (Perepelkin, 1966) for fibers with ideal structure. The effect of, e.g., amorphous regions will increase the dependence and reduce strength further, although for a high degree of polymerization the correction is about 10%. It is thus clear that several structural aspects can and will affect the strength of natural and other polymer fibers: (i) crystallinity/amorphousness; (ii) orientation of molecular chains; (iii) imperfections, defects, cracks; (iv) degree of polymerization. Some of these interact or cooperate, e.g., high orientation leads to anisotropy, which reduces the severeness of cracks; e.g., the effect of degree of polymerization is affected by amorphous regions. The general presentation of strengths of materials of high molecular weight indicate that the strength values, theoretical, maximum attainable, and attained, are only weakly dependent on the molecular interaction, chain conformation, and flexibility and are directly related to the number of chains per cross-sectional area. This discussion shows the structural aspects which affect and reduce the strength of natural fibers, but it also gives methods for improving strength by perfecting the structure of the fibers; whether this is possible, in particular for natural organic fibers, is not clear at present; there is both a materials science line and an economic route to consider.

10

Natural Organic Fibers Table 8 Theoretical and experimental stiffness. Eth Chains/area (1018 m72) (GPa)

Material Diamond

18.2

900 1210

Graphite Graphite (thread-like) Polyethylene

14.0 5.5

Cellulose I

3.1

Cellulose II

(3.1)

182 340 123 56 56

Reference

Eexp (GPa)

Reference

Kelly and Macmillan, 1986 (p. 22) Kelly and Macmillan, 1986 (p. 6)

1100±1200 Kelly and Macmillan, 1986 (p. 22) 1100±1200 Kelly and Macmillan, 1986 (p. 6) 170±700 Perepelkin, 1972 570±1000 Perepelkin, 1972 Treloar, 1960 240±360 Holliday, 1975 Shimanouchi et al., 1962 250 Sakurada et al., 1966 Meyer and Lotmar, 1936 130 Sakurada et al., 1966 Treloar, 1960 Jaswon et al., 1968 90 Sakurada et al., 1966

Table 9 Experimental stiffness values for practical natural fibers. Fiber

Conditions

Eexp (GPa)

Flax

dry dried under tension 65% relative humidity, unstressed dry dried under tension 65% relative humidity, unstressed 65% relative humidity, stressed dry dry dried under tension

80 110 74 70 87 83 93±103 66 50±70 92

Hemp

Jute Ramie

1.10.2.2.4

Stiffness

The stiffness of polymers in general and natural organic fibers in particular is of importance for several reasons (Holliday, 1975). It cannot be expected that high, and in the limit theoretical, strengths can be reached before the high or theoretical stiffness (modulus) has been achieved. Second, it is normally much easier to relate stiffness to molecular structure than it is to relate strength and other (complex) mechanical properties to structure. Third, although strength is important in engineering design, stiffness is often as important because many engineering structures are limited by the allowable deflection. This is further of importance for natural fibers, because the attainable stiffness values are moderate compared to other materials such as synthetic fibers like graphite and aromatic polymers. The concept for calculating the theoretical stiffness of materials based on linear polymer chains loaded along the chain direction is considering the stretching and bending of the relevant atom±atom bonds. The presentation by Kelly and Macmillan (1986) collects several

Reference Meyer and Lotmar, Meyer and Lotmar, DeVries, 1953 Meyer and Lotmar, Meyer and Lotmar, DeVries, 1953 DeVries, 1953 Meyer and Lotmar, Meyer and Lotmar, Meyer and Lotmar,

1936 1936 1936 1936 1936 1936 1936

such estimates. Meyer and Lotmar (1936) pioneered this type of calculation and estimated the stiffness of cellulose. Other calculations based on the same ideas have been made by several authors, and an overview is given by Holliday (1975). Experimental values for the basic structural elements have been obtained by testing carefully prepared and characterized fibers, using test methods such as X-ray and Raman techniques (Sakurada et al., 1966). Relevant data for theoretical and related experimental values of stiffness for diamond, graphite, polyethylene, and celluloses are presented in Table 8. It is clear that the agreement between theory and experiments is much closer for stiffness than it is for strength. This further indicates that experimental measurement of stiffness can be a strong indicator of the molecular structure, and in particular its degree of perfection or lack of defects. It may therefore be one way to improve and document the strength of natural organic fibers. The stiffness of practical cellulosic fibers is (also) of relevance and some data are collected in Table 9. All values are in the range

Introduction

11

cellulose I. This increase may indicate the potential straightening and/or aligning of the molecular chains under load. The initial slopes (moduli) are 250 GPa for polyethylene and 125 GPa for cellulose I, and are thus close to the values given in Table 8.

1.10.2.3

Figure 4 Stress±strain curves for polyethylene and cellulose I (ramie); strains refer to crystalline regions; the curves are redrawn from data of Sakurada et al. (1996).

70±100 GPa, and the individual fibers show the effect of dry/humid conditions and tension. The increased stiffness for dry fibers is related to the chemistry of the cellulose, and the increased stiffness under tension may indicate the ease of orienting and aligning the molecular chains in the fibers. It seems that under dry conditions and under tension, stiffness values of more than 100 Gpa should be attainable. This is an encouraging aspect of natural organic fibers when they are to be used as reinforcements in (polymeric) composites. The effect of moisture on the stiffness of cellulose has been studied by Sakurada et al. (1966), who recorded the stiffness of selected natural fibers (ramie, cellulose I; rayon, cellulose II) both for the crystalline regions (lattice modulus measured by X-ray technique) and for the specimen. The stiffness of the crystalline regions was not affected by an increase in moisture content, while the specimen stiffness was reduced by a factor of about two for an increase of relative humidity from approximately 10% to approximately 40±50%. The much lower and changeable modulus of the specimen may indicate the influence of the amorphous regions in the fiber structure, and that the modulus of these regions are (much) lower than the modulus of the crystalline regions. The effect of tension or straining on the modulus is illustrated in Figure 4 for polyethylene and ramie (cellulose I), respectively, redrawn as conventional stress±strain curves from the experimental data and curves of Sakurada et al. (1966). The curves refer to the crystalline regions (strain measured by X-ray technique), and for both fibers the slope, i.e. the modulus, increases beyond a strain of 0.06% for polyethylene and of about 0.25% for

Morphological Characteristics Specific to Plant Fibers

Definition of the term ªfiberº when applied to plant derived materials needs to be clarified. Strictly, plant fibers are single cells. For example, when softwoods such as pine or spruce are pulped or defibrated, separation into individual fiber cells occurs (as in paper making). These fibers are generally 1±3 mm in length and constitute single cells (Ilvessalo-Pfaffli, 1995). However, in the case of the longer bast ªfibersº obtained from flax, hemp, and jute, or leaf ªfibersº obtained from sisal, the long fiber is actually a ªfiber bundleº or a bundle of individual fiber cells. In flax, e.g., the individual fiber cell is typically 25 mm long, whereas the fiber bundle can be 750 mm long. In jute, fiber bundles of around 3 m in length are readily obtained. Such fiber bundles are typically less than 0.5 mm in diameter, producing aspect ratios for flax and jute fibers ranging from 1500 to 3000. In contrast, the aspect ratios of single fibers in softwood pulp are around 100 (Bolton, 1992). Plant fibers supplied to and utilized by the composites sector are therefore either single fibers (wood) or fiber bundles (flax, hemp, jute, sisal, coir). The stronger fibers with greatest potential for reinforcing engineering plastics are really fiber bundles of length ranging from 500 to 3000 mm and diameter typically less than 0.5 mm. This is illustrated schematically in Figure 5. Commercially important plant fibers are often derived from two cell types: schlerenchyma cells, which are thick-walled and lignified, and tracheid cells, most commonly found in woods. Within schlerenchyma cells, a thick secondary wall is laid down over the primary wall. The fiber bundles obtained from flax, hemp, jute, and other bast fiber bearing plants are strands of overlapping, schlerenchyma cells. Tracheids in wood are the dead remnants of the conducting cells of vascular plants. The mature tree contains mostly dead tracheids with very highly lignified and thick secondary walls. The nature of lignin is discussed in Section 1.10.3.1. Within individual plant fiber cell walls there is a further level of complexity: the individual

12

Natural Organic Fibers

Figure 5 The structure and morphology of bast fibers and fiber bundles.

cells are themselves composites comprised of cellulose microfibrils embedded in a matrix of other less crystalline polymers. This point is further elaborated in Section 1.10.3. Animal fibers such as wool and silk are protein based and as such are single fibers rather than fiber bundles. Proteins are natural heteropolyamides with complex primary, secondary, and tertiary chemical structures.

1.10.3 1.10.3.1

CHARACTERISTICS OF NATURAL ORGANIC FIBERS Basic Chemical Composition

To understand the properties of plant fibers it is necessary to understand the fine structure (ultrastructure) and chemical composition of the cell wall. It should be appreciated that most of the scientific work available in the literature is based on studies of either wood fibers or cotton because of the commercial significance of these materials. However, a number of recent studies of flax are reaching the publication stage and further information is becoming available about fiber structure and cell wall chemistry (Sharma and Van Sumere, 1992; Akin et al., 1996). The following constitutes a brief review of the plant fiber cell wall. The basic information is derived from studies of wood fibers. As much as possible, any contrasting properties of bast and leaf fibers will be highlighted. Fully developed plant fiber cell walls contain four main types of polymers in varying amounts according to species, variety, and other aspects. These are cellulose, hemicelluloses, pectic materials, and lignin. The pectins occur in most mature plant cell walls with the exception of woody tissues, wherein extensive secondary wall thickening replaces almost all of

the pectin with lignin. The other components are universal to plant fibers. Cellulose is the reinforcing material within the cell wall. Plant cell walls can be envisaged as ªspirally wound compositesº with microfibrils of highly crystalline cellulose as the reinforcer within the matrix of hemicelluloses and lignin. Cellulose is defined chemically as a linear, crystalline polymer composed of (1±4) linked b-Dglucopyranose residues (see Figure 6). It is a high molecular weight homopolymer of the pyranose ring form of the common sugar, glucose. The biosynthesis of cellulose is complex and has only recently become understood. For more fundamental information, details can be found in the text of Delmer and Stone (1988). However, it is well established that cellulose molecules are laid down in microfibrils in which there is extensive hydrogen bonding between cellulose chains producing a strong crystalline structure. In the wood cell wall, the microfibril is reported to be ovoid or almost square in cross-section and 3±4 nm in diameter (Fujita and Harada, 1991). It is difficult to precisely estimate microfibril length, but values of several micrometers occur. The aspect ratio of the individual microfibril is therefore high. The microfibrils in bast fibers from flax, jute, and hemp are larger than those commonly found in wood. There is currently much debate as to the precise nature of microfibrils, however, it is clear that they are inherently very strong, with tensile strength superior to steel. The fact that plant fibers in general are weaker reflects weaknesses in other domains of the cell wall. The isolation and utilization of microfibrils themselves in composites is an attractive thought that has received some recent attention (Dufresne et al., 1997). The secondary wall found in wood cells is composed of two or three layers, known as S1,

Characteristics of Natural Organic Fibers

13

Figure 6 The molecular structure and arrangement of cellulose.

S2, and S3, respectively. In each of these layers, the cellulose microfibrils are ªspirally-woundº at a different angle to the major axis of the tracheid. This variation in microfibril angle imparts strength to the fiber structure in a variety of directions. Within the bast or schlerenchyma cells found in flax, hemp, jute, and kenaf, the secondary wall is less thick than that of wood, but contains layers of similarly spirally-wound microfibrils embedded in a hemicellulose and pectin-rich matrix. This ªcomposite structureº imparts potentially high strength to regions of the cell wall. Hemicelluloses cover the surface of the microfibrils, hydrogen bonding to outer cellulose chains. They are a group of heteropolysaccharides occurring within plant cells. They are ªcopolymersº of a number of sugars, most commonly glucose, mannose, xylose, galactose, and arabinose. Hemicelluloses are of lower molecular weight than cellulose, are occasionally branched, and are invariably much less ordered than cellulose as a consequence of their more heterogeneous structures. Hemicelluloses are very hydrophilic polymers and are largely responsible for the water sorption behavior exhibited by plant fibers (along with pectins when present). In many plant and wood species, the hemicelluloses tend to interface between the cellulose and the lignin (see below). Further discussion on the

nature and role of hemicelluloses is given by Shimizu (1991). The third component of wood and plant cell walls is lignin. Lignin is generally regarded as the adhesive within the cell wall and is the final polymer to be laid down during cell development. Lignin is essentially a disordered, polyaromatic, and cross-linked polymer arising from the free radical polymerization of two or three monomers structurally related to phenylpropane. The common monomers are transconiferyl alcohol and trans-synapyl alcohol. Many nonwood lignins also contain units resulting from the copolymerization of a third monomer trans-p-coumaryl alcohol (Sakakibara, 1991). Free radical coupling of the lignin monomers gives rise to a very condensed, reticulated, and cross-linked structure. The lignin matrix is therefore analogous to a thermoset polymer in conventional composite terminology. The lignin polymer is laid down between the hemicellulose zones surrounding microfibrils, conferring rigidity, a degree of hydrophobicity, and decay resistance to the cell wall. In most nonwood fiber cells, pectin is a major matrix component within the cell wall. Pectins are polysaccharides and can have complex structures; they can be branched. The main chain is a polymer of (1±4)-a-D-glucuronic acid, in which the acid groups are partially esterified with methanol. At frequent intervals,

14

Natural Organic Fibers Table 10

Fiber/fiber bundle

Comparative chemical and mechanical properties of available plant fibers.

Type

Cellulose content (%)

Lignin content (%)

Pectin content (%)

Tensile strength (MPa)

Stiffness (GPa)

Other comments Fiber bundle Fiber bundle Fiber bundle Fiber bundle Fiber bundle Fiber bundle Fiber bundle High cellulose Low cost Low cost Available fibrous waste material

Flax

Bast

65±85

1±4

5±12

500±900

50±70

Hemp

Bast

60±77

3±10

5±14

350±800

30±60

Jute

Bast

45±63

12±25

4±10

200±450

20±55

Kenaf

Bast

45±57

8±13

3±5

±

±

Sisal

Leaf

50±64

±

±

100±850

9±22

Abaca

Leaf

60

12±13

1

±

±

Coir

Seed

%30

40±45

±

±

±

Cotton

Seed

85±90

±

0±1

±

±

Softwood Hardwood Wheat straw

Wood Wood Cereal stem (grass)

40±45 40±50 38±41

26±34 20±30 12±16

0±1 0±1 0±1

98±170 ± low

10±50 ± low

Sources: Fengel and Wegener, 1984a; Robson et al., 1993; Lawther et al., 1995; Rowell et al., 1997; Sun et al., 1998.

residues of the sugar, rhamnose, are included in the main chain and side chains rich in arabinose and galactose sugars are attached to the rhamnose residues. In addition, acid groups between adjacent pectin chains are often cross-linked by calcium ions. This confers a degree of structural integrity and rigidity to regions of the cell wall rich in pectin. Pectin is an important component of nonwood fibers, particularly the important bast fibers from flax, hemp, and jute (see Figures 7±10).

1.10.3.2

Specific Types of Natural Organic Fibers

The natural organic fibers of most interest with respect to composite materials are cellulose-based fibers from plants. The most important of these are the bast (schlerenchyma type) fiber bundles obtained from plants such as flax, hemp, jute, and kenaf. Leaf fibers such as sisal, pineapple, and abaca (manila hemp) are also of interest. The two most prominent of the seed fibers are cotton and coconut (ªcoirº). Some comparative properties of these fibers are shown in Table 10. Of the values shown in Table 10, many are approximate and can vary according to different varieties of a species, growing conditions,

and a number of other factors including methods of fiber preparation from the plant raw materials. Such variability is common among plant derived fibers, and property values quoted should be treated with caution. This is generally in contrast to synthetic fibers, which tend to be produced to tight specifications.

1.10.3.3

Natural Organic Fibers and Synthetic Fibers

Natural fibers can be used as reinforcements in composites, both in their own right and in competition with synthetic fibers. The fundamental characteristics which qualify fibers as reinforcements are diameter, density, stiffness, and strength. These properties, in turn, depend on the structure and technical treatments, and the natural fibers are particularly variable and not yet controllable in terms of such characteristics. Typical values for both synthetic and natural fibers are presented in Table 11. The often rather large range for stiffness and strength values of natural fibers reflect both the natural variation of plant type, growth conditions, and harvest procedures, and the relatively early stage at present for the industrial use of natural fibers as reinforcements in composites. Specific

Characteristics of Natural Organic Fibers

Figure 7 Flax fibers.

Figure 8 Hemp fibers.

15

16

Natural Organic Fibers

Figure 9 Jute fibers.

Figure 10

Birch fibers.

Characteristics of Natural Organic Fibers

17

Table 11 Practical properties of fibers for composites: synthetic inorganic fibers, synthetic organic fibers, and natural organic fibers.

Fiber Carbon Glass Aramid Polyethylene Flax Hemp Jute Sisal Banana Pineapple Cotton Softwood Hardwood

A B C A B A B

Diameter (mm)

Density (g cm73)

Stiffness (GPa)

Strength (MPa)

Strain (%)

7 6 10 10±20 12 12 38 27 19 25 20

1.75 1.77 2.18 2.54 1.44 1.45 0.97 0.97 1.4±1.5 1.48 1.3±1.5 1.45 1.4 1.44 1.50 1.4 1.4

235 377 827 72 58 121 117 172 50±70 30±60 20±55 9±22 7±20 35±80 6±10 10±50 10±70

3530 4410 2200 3530 3600 3150 2650 3090 500±900 300±800 200±500 100±800 500±700 400±1600 300±600 100±170 90±180

1.5 1.2 0.27 4.8 3.7 2.0 3.5 2.7 1.5±4.0 2±4 2±3 3±14 1±4 0.8±1.6 6±8

20 33 20

sets of properties have not yet been developed for efficient use of natural fibers for industrial products. From Table 11 it is observed that diameters are comparable for all fibers, synthetic and natural, and may thus be in the right range for composites and efficient processing. The densities are also comparable, with glass and carbon on the heavy side and polyethylene on the light side; although the variation range is not great it may still be of importance in flexural beam constructions where density enters an efficiency parameter as density to the third power. The stiffness and strength values are (still) relatively low and variable, and although the potential for improvement seems large, as described in Section 1.10.2.2, it may be difficult to attain. The stiffness and strength of fibers are the basis for the reinforcement, but also the interfacial strength (bonding, adhesion) is important for efficient reinforcement and for aspects related to structural parameters, strength properties, and processing conditions. The interface strength ti controls the critical fiber length and is therefore of significance for short natural fibers such as wood fibers which may not give efficient reinforcement if ti is too low. At present there seems to be no clear measurements of interfacial strength values related to natural fibers. For strength properties where the interface is loaded directly in shear and/or tension, the interfacial strength enters directly as a controlling parameter; this relates to off-axis loading,

Thermal expansion coefficient (1076 K71) 70.4 0.0 71.45 5.0 70.35

shear loading, transverse tensile loading, and axial compressive loading. These types of mechanical loadings are well defined for unidirectional composites, while they may enter as a mixture for composites with randomly oriented fibers. Such composites with natural fibers are often made from mats of fibers, and are at present the most common types of natural fiber composites; therefore the interfacial strength becomes relatively more important for natural fibers. For natural fiber composites, as well as for all other composites, weak interfaces act as internal defects or cracks and are thus general detrimental and lead to early crack development and growth. It may nevertheless not always be an advantage for a composite to have a strong interface, since this could render the composite as a whole brittle because the cracks will not be deviated from their main path by branching and/or interfacial delamination. It is therefore not only of interest to develop strong interfaces but rather to develop interfaces of controlled strength. The surface characteristics of natural fibers and their possible modifications are described in Section 1.10.4. The compatibility between natural fibers and polymer matrices is described by the wettability or surface tension of the polymer on the fiber. These characteristics will govern the impregnation of a fiber assembly by the (normally) liquid polymer during processing to composites. There seems at present to be no well-defined measurements of relevant wettability parameters for natural fibers and polymer matrices.

18 1.10.4 1.10.4.1

Natural Organic Fibers SURFACE CHARACTERIZATION AND MODIFICATION Basic Physical±Chemical Concepts

Almost all natural organic fibers and particularly plant-derived fibers are hydrophilic in nature, mainly as a consequence of their chemical structures. Plant fibers contain polysaccharides, of which two types are very hydrophilic: the hemicelluloses and the pectins. Hydroxyl and carboxylic acid groups located on these branched heteropolysaccharides are active sites for the sorption of water (Avramidis, 1997). The cellulose component is also hydroxyl group rich, but as a consequence of its linearity and high crystallinity, little water can be accommodated within the microfibrils, so the polymer is essentially less hydrophilic than may be expected. However, free, nonH-bonding OH groups on the microfibril surfaces are available for sorption. It should also be remembered that plant fibers are produced within dynamic, living, aqueous environments and, until the cell wall lignification stage of growth, are designed for optimal performance in the living plant in water-rich environments. The consequence of this hydrophilic nature of plant fibers for matrix polymer reinforcement can be profound. Many of the common matrix polymers in composites ranging from polyolefins to polyesters and epoxies are largely hydrophobic in nature. Thermosets such as phenolformaldehyde and related polymers are less hydrophobic in nature and are therefore less problematic. In general, there is a poor surface wetting of plant fibers by most thermoplastic polymers commonly used in composites. This can lead to the formation of ineffective interfaces between the fiber and matrix phases, with consequent problems such as poor stress transfer, small void spaces, and debonding in the resulting composite materials. Indeed, initial problems encountered in ªwetting outº plant fiber mats and rovings with polyolefins have led to an aversion to these materials within parts of the industry. An additional important factor connected with the use of plant fibers is their inherent dimensional instability in conditions of varying humidity. This is an obvious further consequence of the hydrophilic nature of cell wall polysaccharides and the tendency of plant fibers to swell and shrink as they gain or lose water can be problematic, especially in composites where the relative proportion of fibers is high. In such cases, the swelling of fibers in conditions of high humidity or direct exposure to liquid water can stress the surrounding

matrix and lead to composite damage and eventual failure. Composites containing plant fibers have been produced for many years; however, because of such problems the approach has often been ºlow-techº in nature or confined to a number of specialty products (Ivens et al., 1997); the issue of effectively reinforcing thermoplastics needs to be addressed.

1.10.4.2

Approaches to Surface Modification

The problems described in Section 1.10.4.1 can be largely averted by fiber surface sizing/ coating or chemical modification. Surface coating is a simple but expensive solution. A more satisfactory approach is the transformation of surface hydroxyl groups by chemical modification. The polysaccharide and lignin components of plant fiber cell walls and surfaces are chemically reactive; indeed this is reflected in the water sorption behavior of the materials. This inherent reactivity can be harnessed, and hydrophobic monomers, oligomers, or even polymer chains can be covalently attached to the fiber surface. Such reactive modifiers can be loosely described as compatibilizers. Simple chemical modification of wood and plant fibers has received much attention over the past 20 years or so (Andersson and Tillman, 1989; Rowell, 1992; Banks and Lawther, 1994). Techniques such as acetylation, wherein surface hydroxyl groups are reacted with acetic anhydride to produce more hydrophobic acetyl ester groups, have proved moderately successful in improving the nonspecific compatibility of plant fibers with matrix polymers and the dimensional stability of the fibers themselves (Rowell and Rowell, 1989). However, this has broadend out over recent years and a number of potentially effective modification systems and compatibilizers have been examined and developed. Chemical modification is generally taken to mean a process involving the creation of a chemical (usually covalent) bond between a surface cell wall polymer and an introduced reagent to form a novel adduct. Two broad types of modification can be identified: (i) Fiber surface modification that improves the efficiency of wetting by the matrix polymer. This approach is generally termed compatibilization, and is most suitable for fiber reinforced thermoplastic composites. (ii) Modification enabling chemical reaction to occur between the fiber surface and matrix polymer. This is termed coupling and usually utilizes difunctional chemical reagents; a

Surface Characterization and Modification

Figure 11

19

Flax fiber/polypropylene composite; failure zone after mechanical loading gives information on the interfacial adhesion between plant fibers and polymer matrix.

requirement is that a reactive (generally thermosetting) matrix material is utilized in the composite. The first type of modification has received extensive research attention. Useful compatibilizers are developed and used in systems, wherein plant fibers reinforce thermoplastics such as polyethylene (PE), polypropylene (PP), and polystyrene (PS), including maleic anhydride grafted PP (ªMAPPº), monofuntional isocyanates, and m-phenylene bismaleimide modified fibers (Jacobsen et al., 1995; Snijder et al., 1998) (see Figure 11). Some of the tested compatibilizers have proved effective in improving the flexural and tensile properties of composites. However, the effort has tended to be strongest in extrusion compounding of fibers and plastics ready for injection molding, where the MAPP compatibilizer has been successfully utilized. The shear forces and mixing conditions created within an extruder are ideal for the dynamic application of a compatibilizer ªon-lineº during the production process. A number of research groups have reported on the improved properties of injection molded composites based on PP with

wood, flax, and kenaf fibers. This approach, however, utilizes shortened fibers/fiber bundles (5 4 mm) and 50±100% improvements in the tensile strength of test specimens have been reported (see Table 12). With such short fibers, there is still some debate as to whether the fiber serves mainly as a filler, with a minimal reinforcement role. Little work is reported in the literature about the use of compatibilizers with longer plant fiber bundles in mats prior to composite formation. This is perhaps because of the lack of an obvious way to apply the compatibilizer evenly to the fibers and to ensure reaction. Considerable research effort is needed in this area to improve the prospects for the enhanced utilization of plant derived fibers in higher performance composite products. An important step will be the development of effective methods to attach the modifiers to the fibers during mat forming or other processing stages. The second type, namely the use of reactive coupling agents, has received some attention where more reactive, thermoset resins are used. Examples of coupling agents examined

20

Natural Organic Fibers Table 12 Effect of simple compatibilizers on tensile strength of extruded plant fiber composites.

Composite fiber + matrix Flax + PP Jute + PP Sisal + LDPE

Compatibilizer type

Tensile strength increase over pure matrix

Reference

MAPP MAPP Isocyanate + peroxide

50% 100% 25%

Pott et al., 1997 Karmaker and Youngquist, 1996 Joseph et al., 1996

include reactive diisocyanates, silanes, and other difunctional compounds. The use of isocyanates is interesting as these species are very reactive towards fiber hydroxyl groups. However, most thermosetting systems are relatively reactive themselves to hydroxyl groups and wet the fibers reasonably well. Hence the extra modification step is not cost effective in that performance improvements are not spectacular; use in low matrix composites may, however, be justified.

1.10.5

PROCESSING TO COMPOSITES

Natural fibers find several applications in composites and composite-like contexts. Use of the fibers in bulk form with random orientation and no binder (matrix) can have applications as insulation materials and filters. Use of the fibers with random orientation and some binder have applications as fiberboards and MDF (medium-density fiber) panels. Natural fibers with full matrix content lead to composites for potential load-bearing applications. The present ways to fabricate composites with natural fibers are generally based on the processing technologies used for conventional synthetic fiber composites. In most cases the fibers need to be made into some form of semiproduct which allows (easy) handling of the fiber assembly for the (final) processing to composite materials or products.

1.10.5.1

Forming Fibers into Semiproducts

The most common and nearly only route to composites at present is the use of some form of fiber mat. The fibers should normally be short, and long fibers, such as flax and hemp, will need to be shortened before being made into mats. Various suspensions in air of the fibers are established before the fibers are allowed to settle on a conveyor belt in order to

form a mat. Mats of pure fibers are possible but difficult to handle, and normally some sort of fixation of the fiber assembly is used. With a set of needles moving through the mats, needle punching, some fibers are pulled through the mat and thus used to stabilize it. With a normally small amount of polymer fibers, typically polyethylene/polypropylene, the mats can be stabilized by thermal fixation whereby the polymer fibers melt and anchor the natural fibers. With a small amount of long fibers, the mats can be stabilized for easy handling. Carding techniques are also used to produce fiber mats. The mats need to allow a liquid polymer to impregnate it, and permeability and wetting are governed by the fiber dimensions and mat density as well as by the surface characteristics of the fibers. The general properties and morphology of the fibers are of importance for the method and degree of compaction attainable for the mat during processing; this in turn governs the (maximum) volume fraction of fibers in the resulting composite (Andersen and Lilholt, 1999). These mats are all essentially matrix free, and the matrix is added in some form during processing into composites. An alternative method is to mix short fibers of the polymer matrix into the mat during mat forming such that a hybrid mat is produced. This allows direct processing of hybrid mats into composites, without further addition of polymer matrix. For conventional composites with synthetic fibers, woven mats are used extensively for bior multiaxially fiber oriented composites, and fiber bundles (rovings, yarns) are used widely for unidirectionally fiber oriented composites, including composites fabricated from prepegs. None of these techniques seem to have been tried or used for natural fibers.

1.10.5.2

Processing Limits

Plant and animal derived fibers are by nature organic materials and are therefore flammable,

Processing to Composites subject to thermal degradation, and in many cases are biodegradable. In terms of composite fabrication and processing, the relatively low thermal stability of plant fibers presents problems. It is generally accepted that prolonged exposure to temperatures greater than 110 8C causes degradative changes to polysaccharides within the cell wall, especially hemicelluloses. At 110 8C, the rate of degradation is low; however, as the temperature increases, this rate increases. Exposure of plant fibers at temperatures in excess of 200 8C for even a few minutes can lead to severe degradation and loss of structural integrity of the fibers (Fengel and Wegener, 1984b). Likewise, plant fibers are inherently unstable at high and low pH. The polysaccharide components are particularly susceptible to hydrolytic changes. The b(1±4) and other glycosidic links common in cellulose and hemicellulose polymers are the vulnerable points in the cell wall. This therefore precludes the use of plant fibers in, e.g., thermoset systems which involve extreme acid or alkaline cures. Although not strictly a ºprocessing problem,º plant fibers are also variable in quality, in that year-to-year variations in, e.g., flax fiber supply can occur. This is partially due to climatic (year-to-year) factors, but is also a consequence of the tendency of industrial users of flax to take the ªwasteº materials or ªtowº fibers from textile production. The issue of the biodegradability of plant fibers should be examined. The degradability of these materials is often put forward as a positive advantage justifying the use of these materials. However, many plant fibers can rapidly degrade in warm, humid conditions, particularly if they are in contact with soil or other decaying biomass. For many applications, outdoor performance for a number of years is required for composite components. For such applications, measures must be taken to decrease or remove the tendency of fibers to biodegrade before the end of a reasonable service period. This can be done using the methods of chemical modification or surface coating already discussed, as the first stage in the degradation process involves wetting of the fibers by water, allowing colonization by microorganisms. If the fibers are modified such that the moisture content is unlikely to rise above 10%, then biodegradation will not occur. In summary, provided the plant fiber can be formed into a mat, roving, or other such intermediate product, then a composite can be produced using standard techniques, as long as the abovementioned limits are respected: prolonged exposure to temperatures in excess of around 160 8C and extremes of pH should be

21

avoided. If biodegradation is to be avoided, then some fiber modification is necessary. The limited temperature stability, with a maximum allowable temperature of about 200 8C, sets a sometimes serious limit to the processing conditions and the choice of polymer matrix. Normally low melting thermoplastics are needed, and polypropylene with a melting temperature of about 160 8C has been used extensively. Thermosets can be used if curing can be performed at moderate temperatures, and several polyesters and epoxies will cure below 100 8C and may thus limit the thermal treatment experienced by the natural fibers during processing. The natural fibers are expected to show reduced stiffness and strength at increasing temperature, as discussed in Section 1.10.2.2; this may offer advantages during processing at higher temperatures. At present no data seem to be available to investigate these aspects.

1.10.5.3

Matrix Polymers

The polymers to be used as matrices for composites must be chosen to fulfil several requirements. They contribute to the composite properties, e.g., strength and fracture toughness; they must form a normally good bonding to the fibers to establish a proper composite; and they must allow processing at moderate temperatures, below about 200 8C. The thermoset polymers can normally be selected to have acceptable curing temperatures. The natural fibers are normally hydrophilic due to the hydroxyl groups on cellulose molecules, and therefore a good bonding is expected if the hydroxyl group can react with the thermoset to form covalent bonds. Work on jute (Roe and Ansell, 1985) and sisal fibers (Joseph et al., 1996a, 1996b) seems to demonstrate such effects. The thermosets have potential disadvantages in that processing time is generally long because of the time-consuming curing, and the polymer cannot (directly) be recycled. The thermoplastic polymers normally have a high processing (melting) temperature. The hydrophobic nature of the polymers is not compatible with the hydrophilic nature of the fibers; it is therefore in most cases necessary to modify the fiber surface characteristics, as discussed in Section 1.10.4. The compatibilizing treatment can also be performed by modifying the matrix; a widely used procedure is to involve maleic anhydride modified polypropylene (MA-PP). The thermoplastic polymers generally allow fast processing (melting, solidification), and seem to offer a

22

Natural Organic Fibers

potential for recycling/reuse of the polymer, although this has not been investigated in detail. The polymer matrices can also be classified according to their origin, in the sense that synthetic polymers are derived from crude oil, and this includes most polymers, while natural polymers can be produced from plant material, e.g., polylactate is such a natural polymer.

1.10.5.4

Processing Techniques

The processing methods used to fabricate natural fiber composites are generally the same as those used for synthetic fiber composites. The most widely investigated processes are the hot press consolidation of either natural fiber mats combined with foils of polymer, e.g., film-stacking technique, or consolidation of hybrid mats. The use of ªdryº mats in resin transfer molding (RTM) techniques is widely used, in various modifications. Injection molding, where fibers and matrix are premixed, allows fast production of simple parts. In all cases used the composites consists of randomly oriented fibers. No serious attempts to develop aligned fiber orientations seem to have been made.

1.10.6

POTENTIAL APPLICATIONS AND OUTLOOK

The use of natural fiber composites is of some volume at present, in particular in the transport (car) industry. It is believed that the argument for applying natural fibers is that their composites have acceptable, although not particularly good, properties for the tasks required, such as interior linings and other facilities in cars. In addition the natural fiber composites are generally of lower density than conventional composites (e.g., glass/polymer) and thus offer a, although small, weight reduction; this is important in modern transport where environmental concerns require reduced fuel consumption and exhaust gases. A third aspect is the general interest of society in the use of renewable resources. In this context natural fibers can be recycled through nature's own growth and degradation processes, and this further implies CO2 neutrality. Natural fibers are biodegradable, and composites may be so to some extent; this is an advantage from the point of removing worn-out products, but may be a disadvantage during the life of a product.

In general, natural fibers have some potential for use in composites in the future both for technical and environmental reasons. At present, the natural fibers are ªbehindº synthetic fibers in terms of technological development for industrial use. Natural fibers need to be developed to produce well-characterized fibers of uniform properties and secured delivery before they can compete with synthetic fibers in a fair way. This leads to a number of steps in the development of natural fibers for composite use. (i) For plants, special cultivation may be needed to produce well-defined raw material. (ii) For fibers, defibration and preparation methods are needed to ensure a stable delivery of well-characterized fibers to the composites industry. (iii) The fibers themselves need improvement in terms of reduced content of defects and thus increased strength and stiffness; fibers also need to be delivered as long fibers, approaching the concept of continuous fibers used for synthetic fibers. (iv) The fiber surfaces and the resulting interfaces in the composites are of decisive importance for good performance of composites; this emphasizes the need for efficient compatibilization procedures. (v) The processing of natural fibers into composites depends heavily on efficient semiproducts, typically fiber mats. In the general development of processing techniques, the direct use of natural fibers should be investigated, such as handling of continuous fibers. (vi) The composite properties must be understood in terms of fiber properties and this link needs to be developed and established so that a rational use of natural fibers in composites can be promoted. (vii) The composite properties need to be improved leading to high performance composites; this requires development of natural fibers with better strength, greater length, and composites with well-oriented fibers. (viii) Composite properties of particular concern are long-term durability under moisture conditions, impact strength, and general fracture toughness, all of which require improved insight into the natural fiber characteristics. (ix) The use of natural fiber composites for products requires design allowables and design procedures, in particular for load-bearing structures. At present secondary structures (panels, boards, linings) are produced with reasonable progress; the design and use of primary structures (beams, plates) is a potential next step which requires extensive knowledge and practical experience.

References 1.10.7

REFERENCES

D. E. Akin, G. R. Gamble, W. Morrison and L. L. Rigsby, J. Sci. Food Agric., 1996, 72, 155±165. T. L. Andersen and H. Lilholt, in `Proceedings of the 7th Euro±Japanese Symposium', Ecole des Mines de Paris, 1999, pp. 1±12. M. Andersson and A. M. Tillman, J. Appl. Polym. Sci., 1989, 37, 3437±3447. S. Avramidis, in `Proceedings of International Conference on Wood±Water Relations', ed. P. Hoffmeyer, Technical University of Denmark, Copenhagen, 1997. W. B. Banks and J. M. Lawther, in `Cellulosic Polymers, Blends and Composites', ed. R. D. Gilbert, Hanser Publishers, Munich, 1994, chap. 7. A. J. Bolton, `The Untapped Potential of Plant Fibers', University of Wales Review, Science and Technology 6, 1992. K. K. Chawla, `Fibrous Materials', Cambridge University Press, Cambridge, UK, 1998. D. Delmer and B. Stone, `The Biochemistry of Plants', Academic Press, New York, 1988, vol. 14, chap. 9. H. DeVries, Ph.D. Thesis, Delft University, 1953. A. Dufresne, J. Y. Cavaille and W. Helbert, in `Polymeric CompositesÐExpanding the Limits', eds. S. I. Andersen, P. Brùndsted, H. Lilholt, Aa. Lystrup, J. T. RheinlaÈnder, B. F. Sùrensen and H. Toftegaard, Risù National Laboratory, Roskilde, Denmark, 1997, pp. 295±300. R. C. Evans, in `An Introduction to Crystal Chemistry', Cambridge University Press, Cambridge, UK, 1948, pp. 353±355. D. Fengel and G. Wegener, `Wood Chemistry, Ultrastructure, Reaction', Walter de Gruyter, Berlin, 1984a, chap. 3. D. Fengel and G. Wegener, `Wood Chemistry, Ultrastructure, Reaction', Walter de Gruyter, Berlin, 1984b, chap. 12. M. Fujita and H. Harada, in `Wood and Cellulosic Chemistry', eds. D. N. S. Hon and N. Shiraishi, Marcel Dekker, New York, 1991, chap. 1. L. Holliday, in `Structures and Properties of Oriented Polymers', ed. I. M. Ward, Applied Science, UK, 1975, pp. 242±263. M.-S. Ilvessalo-Pfaffli, `Fiber AtlasÐIdenfitication of Papermaking Fibers, Part1', Springer Series in Wood Science, Springer, Berlin, 1995. J. Ivens, H. Bos and I. Verpoest, `Renewable BioproductsÐIndustrial Outlets and Research for the 21st Century', ATO-DLO, Wageningen, The Netherlands, 1997. R. E. Jacobsen, D. F. Caulfield, R. M. Rowell and A. R. Sanadi, in `Proceedings of the 2nd Biomass Conference ofthe Americas: Energy, Environment, Agriculture and Industry', National Renewable Energy Laboratory, Portland, OR, 1995. M. A. Jaswon, P. P. Gillis and R. E. Mark, Proc. Roy. Soc., 1968, A306, 389±412. K. Joseph, S. Varghese, G. Kalaprasad, S. Thomas, L. Prasannakumari, P. Koshi and C. Pavithran, Eur. Polym. J., 1996a, 32, 1243±1250. K. S. Joseph, S. Thomas and C. Pavithran, Polymer,

Copyright # 2000 Elsevier Science Ltd. All rights reserved. No part of this publication may be reproduced, stored in any retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without permission in writing from the publishers.

23

1996b, 37, 5139±5149. A. C. Karmaker and J. A. Youngquist, J. Appl. Polym. Sci., 1996, 62, 1147±1151. A. Kelly and N. H. Macmillan, `Strong Solids', Clarendon Press, Oxford, 1986, 3rd edn.. J. M. Lawther, R. Sun and W. B. Banks, J. Agric. Food Chem., 1995, 43(3), 667±675. K. H. Meyer and W. Lotmar, Helv. Chim. Acta, 1936, 19, 68±86. P. O. Olesen and D. Plackett, in `Proceedings of Natural Fibers Performance Forum', ed. U. Blicher-Mathiesen, Ministry of Food, Agriculture and Fisheries, Centre for Plant Fiber Technology, IENICA, Copenhagen, 1999. S. Pepys, in `Diary 1660±1669', ed. O. F. Morshead, G. Bell and Sons, London, 1967, p. 151. K. E. Perepelkin, Mekhanika Polimerov, 1966, 2, 845±856. K. E. Perepelkin, Fiziko-Khimicheskaya Mekhanika Materialov, 1970, 6, 78±80. K. E. Perepelkin, Fiziko-Khimicheskaya Mekhanika Materialov, 1972, 8, 74±78. G. T. Pott, D. J. Heuting, W. A. H. Maassen and J. H. Van Deursen, Materialen, 1997, (May/June), 11±14. D. Robson, J. Hague, G. Newman, G. Jeronimidis and M.P. Ansell, `Survey of Natural Materials for Use in Structural Composites as Reinforcements and Matrices',Commissioned Report, DTI, UK (Crown Copyright), 1993. P. J. Roe and M. P. Ansell, J. Mat. Sci., 1985, 20, 4015± 4020. R. M. Rowell, in `Composites Applications: the Role of Matrix, Fiber and Interface', eds. T. L. Vigo and B. J. Kinzig, VCH Publishers, New York, 1992. R. M. Rowell and J. S. Rowell, in `Cellulose and Wood', ed. C. Shuerch, Wiley, New York, 1989. R. M. Rowell, R. A. Young and J. K. Rowell (eds.), in `Paper and Composites from Agro-Based Sources', Lewis Publishers, New York, 1997, chap. 5. A. Sakakibara, in `Wood and Cellulosic Chemistry', eds. D. N. S. Hon and N. Shiraishi, Marcel Dekker, New York, 1991, chap. 4. H. Sakurada, T. Ito and K. Nakamae, J. Polym. Sci., 1966, C15, 75±91. H. S. S. Sharma and C. F. Van Sumere, `The Biology and Processing of Flax', M Publications, Belfast, 1992. T. Shimanouchi, M. Asahina and S. Enomoto, J. Polym. Sci., 1962, 59, 93. K. Shimizu, in `Wood and Cellulosic Chemistry', eds. D. N. S. Hon and N. Shiraishi, Marcel Dekker, New York, 1991, chap. 5. M. H. B. Snijder, E. J. M. Reinerink and H. L. Bos, in `Proceedings of the Second International Symposium on Natural Polymers and Composites', eds. L. H. C. Mattoso, E. Frollini and A LeaÂo, Embrapa Agricultural Instrumentation, Sao Paulo, Brazil, 1998. R. Sun, J. M. Fang, A. Goodwin, J. M. Lawther and A. J. Bolton, Carbohydrate Polymers, 1998, 37, 351±359. H. Toftegaard and H. Lilholt, in `Proceedings of ICCM12', Paris, eds. T Massard et al., Woodhead Publishing, Cambridge, UK, 1999, vol. 3, paper no. 1115. L. R. G. Treolar, Polymer, 1960, 1, 290±303. S. N. Zhurkov, Int. J. Frac..Mech., 1965, 1, 311±323.

Comprehensive Composite Materials ISBN (set): 0-08 0429939 Volume 1; (ISBN: 0-080437192); pp. 303±325