MOLECULAR STRUCTURE AND ULTRASTRUCTURE
Chemical analysis reveals the existence of four constituents and provides data on their relative proportions. This information may be summarised as in Table 1 proportions are for timber in general and slight variations in these can occur between timber of different species, or in different parts of a single tree.
Thus, Bertaud and Holmbom (2004), for example, record that the heartwood of Norway spruce contained significantly more lignin and less cellulose than the sapwood; differences in the amounts of specific hemicelluloses were also found to vary not only between heartwood and sapwood, but also between earlywood and latewood.
Cellulose: Cellulose (C6H10O5)n occurs in the form of long slender filaments or chains, these having been built up within the cell wall from the glucose monomer (C6Hl2O6). While the number of units per cellulose molecule (the degree of polymerisation) can vary considerably even within one cell wall, it is thought that a value of 8000–10 000 is a realistic average for the secondary cell wall, while the primary cell wall has a degree of polymerisation of only 2000–4000 (Simson and Timell, 1978).
The anhydroglucose unit C6Hl0O5, which is not quite flat, is in the form of a six-sided ring consisting of five carbon atoms and one oxygen atom (Fig. 1); the side groups play an important part in intra- and intermolecular bonding as will be noted later. Successive glucose units are covalently linked in the 1,4 positions giving rise to a potentially straight and extended chain; i.e. moving in a clockwise direction around the ring it is the first and fourth carbon atoms after the oxygen atom that combine with adjacent glucose units to form the long-chain molecule.
The anhydroglucose units comprising the molecule are not flat, as noted above; rather they assume a ‘chair’ configuration, with the hydroxyl groups (one primary and two secondary) in the equatorial positions and the hydrogen atoms in the axial positions (Fig. 1).
Glucose, however, can be present in one of two forms depending on the position of the –OH group attached to carbon 1. When this group lies above the ring, i.e. on the same side as that on carbon 4, the unit is called α-glucose and when this combines with an adjacent unit with the removal of H–O–H (known as a condensation reaction) the resulting molecule is called starch, a product which is manufactured in the crown and stored in the parenchyma cells.
When the –OH group lies below the ring, the unit is known as β-glucose, and on combining with adjacent units, again by a condensation reaction, a molecule of cellulose is produced in which alternate anhydroglucose units are rotated through 180°: it is this product that is the principal wall-building constituent of timber. Cellulose chains may crystallise in many ways, but one form, namely cellulose I, is characteristic of natural cellulosic materials.
Over the years there have been various attempts to model the structure of cellulose I. The model that has gained widest acceptance, is that proposed by Gardner and Blackwell (1974). Using X-ray diffraction methods on the cellulose of Valonia, these authors proposed an eight-chain unit cell with all the chains running in the same direction.
Forty-one reflections were observed in their X-ray diffractions and these were indexed using a monoclinic unit cell having dimensions a = 1.634 nm, b = 1.572 nm and c = 1.038 nm (the cell axis) with b = 97°; the unit cell therefore comprises a number of whole chains or parts of chains totalling eight in number.
All but three of the reflections can be indexed by a two-chain unit cell almost identical to the earlier model by Meyer and Misch (1937), though this model had adjacent chains aligned in opposite directions. These three reflections are reported as being very weak, which means that the differences between the four Meyer and Misch unit cells making up the eight-chain cell must be small.
Gardner and Blackwell therefore take a two-chain unit cell (a = 0.817 nm, b = 0.786 nm and c = 1.038 nm) as an adequate approximation to the real structure. Their proposed model for cellulose I is shown in Fig. 2, which shows the chains lying in a parallel configuration, the centre chain staggered by 0.266 × c (= 0.276 nm).
Cellulose that has regenerated from a solution displays a different crystalline structure, and is known as cellulose II; in this case there is complete agreement that the unit cell possesses an antiparallel arrangement of the cellulose molecule. Within the structure of cellulose I, both primary and secondary bonding are represented and many of the technical properties of wood can be related to the variety of bonding present.
Covalent bonding both within the glucose rings and linking together the rings to form the molecular chain contributes to the high axial tensile strength of timber. There is no evidence of primary bonding laterally between the chains; rather this seems to be a complex mixture of (fairly strong) hydrogen bonds and (weak) van der Waals forces. The same –OH groups that take part in this hydrogen bonding are highly attractive to water molecules, which explain the affinity of cellulose for water.
Whereas some earlier workers, though placing the intermolecular hydrogen bonds in the ac plane, recorded that the intramolecular hydrogen bonds were on a diagonal plane, thereby linking different layers, Gardner and Blackwell (1974) identified the existence of both intermolecular and intramolecular hydrogen bonds, all of which, however, are interpreted as lying only on the ac plane (Fig. 2); they consider the structure of cellulose as an array of hydrogen-bonded sheets held together by van der Waals forces across the cb plane.
The degree of crystallinity of cellulose is usually assessed by X-ray and electron diffraction techniques, though other methods have been employed. Generally, a value of about 60% is obtained, though values as high as 90% are recorded in the literature. This wide range in values is due in part to the different techniques employed in the determination of crystallinity and in part to the fact that wood is comprised not just of crystalline and noncrystalline constituents, but rather of a series of substances of varying crystallinity.
Regions of complete crystallinity and regions with a total absence of crystalline structure (amorphous zones) can be recognised, but the transition from one state to the other is gradual. The length of the cellulose molecule is about 5000 nm (0.005 mm) whereas the average size of each crystalline region determined by X-ray analysis is only 60 nm in length, 5 nm in width and 3 nm in thickness. This means that any cellulose molecule will pass through several regions of high crystallinity – known as crystallites or micelles – with intermediate non-crystalline or low-crystalline zones in which the cellulose chains are in only loose association with each other (Fig. 2c).
Thus, the majority of chains emerging from one crystallite will pass to the next, creating a high degree of longitudinal coordination; this collective unit is termed a microfibril and has ‘infinite’ length. It is clothed with chains of cellulose mixed with chains of sugar units other than glucose (see below), which lie parallel but are not regularly spaced.
This brings the microfibril in timber to about 10 nm in breadth, and in some algae, such as Valonia, to 30 nm. The degree of crystallinity will therefore vary along its length and it has been proposed that this could be periodic. A more comprehensive account of the structure of cellulose can be fond in Chapter 1 of Dinwoodie (2000).
Hemicelluloses and lignin: In Table 1 reference is made to the other constituents of wood besides cellulose. Two of these, the hemicelluloses and lignin, are regarded as cementing materials contributing to the structural integrity of wood and also to its high stiffness. The hemicelluloses, like cellulose itself, are carbohydrates built up of sugar units, but unlike cellulose in the type of units they comprise; these units differ between softwoods and hardwoods and generally, the total percentage of the hemicelluloses present in timber is greater in hardwoods than in softwoods (Table 1).
Both the degree of crystallisation and the degree of polymerisation of the hemicelluloses are generally low, the molecule containing fewer than 200 units; in these respects, and also in their lack of resistance to alkali solutions, the hemicelluloses are quite different from true cellulose (Siau, 1984).
Lignin, present in about equal proportions to the hemicelluloses, is chemically dissimilar to these and to cellulose. Lignin is a complex, three-dimensional, polymeric, aromatic molecule composed of phenyl groups with a molecular weight of about 11 000. It is non-crystalline and the structure varies between wood from a conifer and from a broadleaved tree.
About 25% of the total lignin in timber is to be found in the middle lamella, an intercellular layer composed of lignin and pectin together with the primary cell wall. Since this compound middle lamella is very thin, the concentration of lignin is correspondingly high (about 70%). Deposition of the lignin in this layer is rapid.
The bulk of the lignin (about 75%) is present within the secondary cell wall, having been deposited following completion of the cellulosic framework. Initiation of lignification of the secondary wall commences when the compound middle lamella is about half completed and extends gradually inwards across the secondary wall (Saka and Thomas, 1982). Termination of the lignification process towards the end of the period of differentiation coincides with the death of the cell.
Most cellulosic plants do not contain lignin and it is the inclusion of this substance within the framework of timber that is largely responsible for the stiffness of timber, especially in the dried condition. A recent and comprehensive account of the structure and influence of the hemicelluloses and lignin in determining certain aspects of wood quality is given by Pereira et al. (2003).
Extractives: Before leaving the chemical composition of wood, mention must be made of the presence of extractives (Table 1). This is a collective name for a series of highly complex organic compounds that are present in certain timbers in relatively small amounts. Some, like waxes, fats and sugars, have little economic significance, but others, for example rubber and resin (from which turpentine is distilled), are of considerable economic importance.
The heartwood of timber, as described previously, generally contains extractives which, in addition to imparting coloration to the wood, bestow on it its natural durability, since most of these compounds are toxic to both fungi and insects. Readers desirous of more information on extractives are referred to the comprehensive text by Hillis (1987).
Minerals: Elements such as calcium, sodium, potassium, phosphorus and magnesium are all components of new growth tissue, but the actual mass of these inorganic materials is small and constitutes on the basis of the oven-dry mass of the timber less than 1% for temperate woods and less than 5% for tropical timbers.
Certain timbers show a propensity to conduct suspensions of minerals that are subsequently deposited within the timber. The presence of silica in the rays of certain tropical timbers, and calcium carbonate in the cell cavities of iroko, are two examples where large concentrations of minerals cause severe problems in log conversion and subsequent machining.
Acidity: Wood is generally acidic in nature, the level of acidity being considerably higher in the heartwood than in the sapwood of the same tree. The pH of the heartwood varies in different species of timber, but is generally about 4.5 to 5.5; however, in some timbers such as eucalypt, oak, and western red cedar, the pH of the heartwood can be as low as 3.0.
Sapwood generally has a pH at least 1.0 units higher than the corresponding heartwood, i.e. the acidity is at least ten times lower than that of the corresponding heartwood. Acidity in wood is due primarily to the generation of acetic acid by hydrolosis of the acetyl groups of the hemicelloses in the presence of moisture; this acidity in wood can cause severe corrosion of certain metals and care has to be exercised in the selection of metallic fixings, especially at higher relative humidities.
The cell wall as a fibre composite
In the introductory remarks wood was defined as a natural composite, and the most successful model used to interpret the ultrastructure of wood from the various chemical and X-ray analyses ascribes the role of ‘fibre’ to the cellulosic microfibrils while the lignin and hemicelluloses are considered as separate components of the ‘matrix’. The cellulosic microfibril, therefore, is interpreted as conferring high tensile strength to the composite owing to the presence of covalent bonding both within and between the anhydroglucose units.
Experimentally it has been shown that reduction in chain length following gamma irradiation markedly reduces the tensile strength of timber; the significance of chain length in determining strength has been confirmed in studies of wood with inherently low degrees of polymerisation. While slippage between the cellulose chains was previously considered to be an important contributor to the development of ultimate tensile strength, this is now thought to be unlikely owing to the forces involved in fracturing large numbers of hydrogen bonds.
Preston (1964) has shown that the hemicelluloses are usually intimately associated with the cellulose, effectively binding the microfibrils together. Bundles of cellulose chains are therefore seen as having a polycrystalline sheath of hemicellulose material, consequently the resulting high degree of hydrogen bonding would make chain slippage unlikely: rather it would appear that stressing results in fracture of the C—O—C linkage. The deposition of lignin is variable in different parts of the cell wall, but it is obvious that its prime function is to protect the hydrophilic (water-seeking) non-crystalline cellulose and the hemicelluloses, which are mechanically weak when wet.
Experimentally, it has been demonstrated that removal of lignin markedly reduces the strength of wood in the wet state, though its removal results in an increase in strength in the dry state, calculated on a net cell wall area basis. Consequently, the lignin is regarded as lying to the outside of the microfibril, forming a protective sheath. Since the lignin is located only on the exterior it must be responsible for cementing together the fibrils and in imparting shear resistance in the transference of stress throughout the composite. The role of lignin in contributing towards the stiffness of timber has already been mentioned.
There has been great debate over the years as to the juxtaposition of the cellulose, hemicellulose and lignin in the composition of a microfibril, and to the size of the basic unit. One of the many models proposed is illustrated in Fig. 3.
In this widely accepted model, the crystalline core is considered to be about 5 × 3 nm containing about 48 chains in either 4- or 8-chain unit cells. Passing outwards from the core of the microfibril, the highly crystalline cellulose gives way first to the partly crystalline layer containing mainly hemicellulose and non-crystalline cellulose chains, and then to the amorphous lignin, this gradual transition of crystallinity from fibre to matrix results in high inter-laminar shear strength, which contributes considerably to the high tensile strength and toughness of wood.
Cell wall layers
When a cambial cell divides to form two daughter cells a new wall is formed comprising the middle lamella and two primary cell walls, one to each daughter cell. These new cells undergo changes within about three days of their formation and one of these developments will be the formation of a secondary wall. The thickness of this wall will depend on the function that the cell will perform, as described earlier, but its basic construction will be similar in all cells.
Early studies on the anatomy of the cell wall used polarisation microscopy, which revealed the direction of orientation of the crystalline regions. These studies indicated that the secondary wall could be subdivided into three layers, and measurement of the extinction position was indicative of the angle at which the microfibrils were orientated.
Subsequent studies with transmission electron microscopy confirmed these findings and provided some additional information, with particular reference to wall texture and variability of angle. However, much of our knowledge on microfibrillar orientation has been derived using X-ray diffraction analysis. Most of these techniques yield only mean angles for any one layer of the cell wall, but recent analysis has indicated that it may be possible to determine the complete microfibrillar angle distribution of the cell wall (Cave, 1997). The relative thickness and mean microfibrillar angle of the layers in a sample of spruce timber are illustrated in Table 2.
The middle lamella, a lignin–pectin complex, is devoid of cellulosic microfibrils, while in the primary wall (P) the microfibrils are loosely packed and interweave at random (Fig. 4); no lamellation is present. In the secondary wall layers the microfibrils are closely packed and parallel to each other.
The outer layer of the secondary wall, the S1, is again thin and is characterised by having from four to six concentric lamellae, the microfibrils of each alternating between a left- and right-hand spiral (S and Z helix), both with a pitch to the longitudinal axis of from 50° to 70° depending on the species of timber.
The middle layer of the secondary wall (S2) is thick and is composed of 30–150 lamellae, the closely packed microfibrils of which all exhibit a similar orientation in a right-hand spiral (Z helix) with a pitch of 10–30° to the longitudinal axis, as illustrated in Figs 4 and 5.
Since more than three-quarters of the cell wall is composed of the S2 layer, it follows that the ultrastructure of this layer will have a very marked influence on the behaviour of the timber. In later sections, anisotropic behaviour, shrinkage, tensile strength and failure morphology will all be related to the microfibrillar angle in the S2 layer.
Kerr and Goring (1975) were among the first workers to question the extent of these concentric lamellae in the S2 layer; these workers found that though there was a preferred orientation of lignin and carbohydrates in the S2 layer, the lamellae were certainly not continuous. Thus, the interrupted lamellae model proposed by them embraced lignin and carbohydrate entities that were bigger in the tangential than in the radial direction.
Cellulose microfibrils were envisaged as being embedded in a matrix of hemicelluloses and lignin. The existence within the S2 layer of concentric lamellae has been questioned again later in the century. Evidence has been presented (Sell and Zimmermann, 1993) from both electron and light microscopy that indicates radial, or near radial orientations of the transverse structure of the S2 layer. The transverse thickness of these agglomerations of microfibrils is 0.1–1.0 nm and they frequently extend the entire width of the S2 layer. A modified model of the cell wall of softwoods has been proposed (Sell & Zimmerman, 1993).
The S3 layer, which may be absent in certain timbers, is very thin, with only a few concentric lamellae; it is characterised, as is the S1 layer, by alternate lamellae possessing microfibrils orientated in opposite spirals with a pitch of 60–90°, though the presence of the right-handed spiral (Z helix) is disputed by some workers.
Generally, the S3 has a looser texture than the S1 and S2 layers and is frequently encrusted with extraneous material. The S3, like the S1, has a higher concentration of lignin than does the S2 (Saka & Thomas, 1982). Electron microscopy has also revealed the presence of a thin warty layer overlaying the S3 layer in certain timbers. Investigations have indicated that the values of microfibrillar angle quoted in Table 2 are only average for the layers and that systematic variation in angle occurs within each layer.
Thus, Abe et al. (1991) have shown that in Abies sachalinensis the microfibrillar angle of the secondary wall, as seen from the lumen, changed in a LH direction from the outermost S1 to the middle of the S2 and then in a RH direction to the innermost S3. This resulted in the boundaries between the three principal layers being very indistinct, confirming reports by previous workers on other species and suggesting that the wall structure can be viewed as a systematically varying continuum.
Microfibrillar angle appears to vary systematically along the length of the cell as well as across the wall thickness. Thus, the angle of the S2 layer has been shown to decrease towards the ends of the cells, while the average S2 angle appears to be related to the length of the cell, itself a function of the rate of growth of the tree.
Systematical differences in microfibrillar angle have been found between radial and tangential walls and this has been related to differences in degree of lignification between these walls. Openings occur in the walls of cells and many of these pit openings are characterised by localised deformation of the microfibrillar structure. Further information on the variability of microfibrillar angle and its importance in determining wood quality is to be found in Dinwoodie (2000) and Butterfield (2003).
VARIABILITY IN STRUCTURE
Variability in performance of wood is one of its inherent deficiencies as a material. It will be discussed later how differences in mechanical properties occur between timbers of different species and how these are manifestations of differences in wall thickness and distribution of cell types. However, superimposed on this genetic source of variation are both a systematic and an environmental one. There are distinct patterns of variation in many features within a single tree.
Length of the cells, thickness of the cell wall and hence density, angle at which the cells are lying with respect to the vertical axis (spiral grain) and angle at which the microfibrils of the S2 layer of the cell wall are located with respect to the vertical axis, all show systematic trends outwards from the centre of the tree to the bark and upwards from the base to the top of the tree. This pattern results in the formation of a core of wood in the tree with many undesirable properties, including low strength and high shrinkage.
This zone, usually regarded as some ten to twenty growth rings in width, is known as core wood or juvenile wood as opposed to the mature wood occurring outside this area. The boundary between juvenile and mature wood is usually defined in terms of the change in slope of the variation in magnitude of one anatomical feature (e.g. cell length, density) when plotted against ring number from the pith.
Environmental factors have considerable influence on the structure of wood and any environmental influence, including forest management, which changes the rate of growth of the tree will affect the technical properties of the wood. However, the relationship is a complex one; in softwoods, increasing growth rate generally results in an increase in the width of earlywood with a resulting decrease in density and mechanical properties.
In diffuse-porous hardwoods increasing growth rate, provided it is not excessive, has little effect on density, while in ring-porous hardwoods, increasing rate of growth, again provided it is not excessive, results in an increase in the width of latewood and consequently in density and strength. There is a whole series of factors that may cause defects in the structure of wood and consequent lowering of its strength.
Perhaps the most important defect with regard to its utilisation is the formation of reaction wood. When trees are inclined to the vertical axis, usually as a result of wind action or of growing on sloping ground, the distribution of growth-promoting hormones is disturbed, resulting in the formation of an abnormal type of tissue. In softwoods, this reaction tissue grows on the compression side of the trunk and is characterised by having a higher than normal lignin content, a higher microfibrillar angle in the S2 layer resulting in increased longitudinal shrinkage, and a generally darker appearance (Fig. 6); this abnormal timber, known as compression wood, is also considerably more brittle than normal wood.
In hardwoods, reaction wood forms on the tension side of trunks and large branches and is therefore called tension wood. It is characterised by the presence of a gelatinous celulosic layer (the G layer) to the inside of the cell wall; this results in a higher than normal cellulose content to the total cell wall, which imparts a rubbery characteristic to the fibres, resulting in difficulties in sawing and machining.
A more comprehensive description of reaction wood is given by Barnett and Jeronimidis (2003). One other defect of considerable technical significance is brittleheart, which is found in many low-density tropical hardwoods and is one manifestation of the presence of longitudinal growth stresses in large diameter trees. Yield of the cell wall occurs under longitudinal compression with the formation of shear lines through the cell wall and throughout the core wood. More information on the variability in structure and its influence on the technical performance of timber can be found in Chapters 5 and 12 of Desch and Dinwoodie (1996).