<!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html>   <blockquote> <blockquote> <blockquote> <blockquote> <center>Some Aspects of Wood Structure and Function by Ephraim Segerman (for publication in the Journal of the Catgut Acoustical Society, 2001)</center> </blockquote> </blockquote> </blockquote> </blockquote> Introduction A scientific theory is an assumed picture of what is happening that can reasonably explain all of the relevant evidence. That evidence can be just simple observations of phenomena. Careful measurements under controlled conditions are good to make. They create more evidence that challenges a theory to try to explain, and if it fails, the theory is disgarded. Only a failure to adequately explain evidence can invalidate a theory, not (as some assume) a lack of supporting careful measurements. A series of theories are presented here that offer explanations of observed phenomena in terms of the wood structure, either physical or chemical (including adsorbed water). They include explanations of permanent and recoverable inelastic bending, sound absorption during playing-in and reduced sound absorption in stewed and aged wood. Basic wood chemistry Wood is a remarkably simple material, both chemically and physically. Chemically, if we ignore the adsorbed water, 99% of wood is comprised of three types of chemicals: about half is cellulose, and about a quarter each is of hemicellulose and lignin. Cellulose molecules are long linear (unbranched) polysaccharide polymers, hemicellulose molecules are shorter branched mixed polysaccharide and polyuronide polymers, and lignin molecules are phenolic polymers. Cellulose is largely crystalline, organised into microfibrils, and is very stable in normal environments. Hemicellulose and lignin are not crystalline, with hemicellulose being rather unstable and lignin very stable. Hemicellulose is the only component that absorbs water to any extent. All of the changes in the dimensions of wood with changing weather are due to how much the hemicellulose is swollen with adsorbed water. Small ions like lithium and sodium can join with and stabilise more adsorbed water molecules (sodium ions can raise the equilibrium moisture content by up to 2% - significant at low moisture contents). When dry, hemicellulose breaks down into carbon dioxide and water. At 20 degrees C, this would reduce wood weight by 1% per century. When there is water present, acid breaks down the hemicellulose somewhat faster by ‘hydrolysis’, mostly into sugar molecules. Added acid speeds this up, but this happens normally because of the natural acidity of wood. These processes of hemicellulose breakdown are called ‘degradation’, and they get very much faster with higher temperatures1. The basic physical structure of wood Physically, wood is a collection of long thin pointed cells made up of cell walls on the outside and air in the inside. Each cell wall has four layers, with the ‘primary’ layer on the outside and three ‘secondary’ layers inside. The amount of hemicellulose is about the same in each layer because the layers need to swell and contract together without stress between them when moisture content changes. The cellulose content of each layer increases steadily from the outer ‘primary’ layer to the innermost ‘secondary’ layer. The lignin content consequently decreases in that sequence. The cellulose microfibrils lie parallel to each other within each layer, and spiral around the cell’s long direction. Different layers have different angles of spiralling. The hemicellulose combined with the lignin acts as a glue that holds together the layers and the cellulose microfibrils within each layer. Most cells have their long directions parallel in the direction of tree growth (some bundles of cells lie perpendicular to this majority, forming the ‘rays’ seen in a radial section). The wall of each cell has a cross-sectional shape that is rectangular with slightly rounded corners. Adjacent cells have their walls glued to one-another by a mixture of about three-quarters lignin and a quarter hemicellulose. The glue layer is about as thick as an individual cell-wall layer except at the rounded corners, where it fills the space. Pairs of glued-together adjacent cell walls act as structural units. There are evenly-spaced holes in the cell walls called ‘pits’ which are usually lined up with similar holes in adjacent cell walls. These holes allow the passage of water or air between cells, and ultimately between the inside and outside of a piece of wood. Bending and taking a set When wood is bent at normal temperatures, the cells on the convex side of the bend tend to be stretched, and those cells on the concave side tend to be compressed. The cell walls can hardly be stretched. They lie along the grain (long direction), the sides of the cross-sectional rectangles line up along the radial direction, and lie along but are staggered in the tangential direction. There is thus little scope for deforming cell shapes to respond to a stretching force along the grain and radial directions, a bit more scope along the tangential direction, and a lot more scope in other directions in-between these. There are no directional constraints in changing the shapes of the cell walls in response to a compressing force. If the bending force is released, the cell walls spring back to their previous shapes, but if the force is applied for a period of time, the wood takes a 'set', and only some of the bend would spring back if the force is released. That time can be shortened by higher temperature and moisture content and by internal mechanical stresses such as vibration and moisture gradients due to changes in humidity. It is very short if the temperature is over 90 degrees C, with enough moisture to avoid drying, when the glue between cell walls becomes plastic and flows readily. In taking a set, the individual cell walls in the pairs slip past one another in the direction that tends to relieve the stress caused by the bending force. The glue between cell walls 'gives', allowing sliding. This is called 'creep'. In the grain direction, the long thin pointed cells can slide along their long directions relative to one-another. In directions perpendicular to the grain direction, the slips between adjacent cell walls are towards making new rectangular cell shapes that are shorter in the direction of compression and longer in the direction of stretching. After creep, if the bending force is released at normal temperature, the cell-wall glue holds. When the creep sliding is along the grain direction, there are no residual internal stresses in the wood structure caused by the movement, and so the bend is permanent, with no ‘memory’ of the original relationships between cells. This happens when bows are bent to shape. If the creep sliding changed the shapes of the cells, there are internal stresses that were not there before, as a 'memory' of the original shapes, and under the right conditions, the original shapes can be recovered. These internal stresses could well be in the originally-grown corners of the cell walls having to open out in straight regions, and the regions that were originally straight having to bend at corners. After the original bending force is removed, with time (which can be shortened dramatically by high temperature and moisture content), the original shapes will largely be restored. Thus a 'warped' bridge (which has creep due to stresses only in the radial direction of the tree) will spontaneously straighten with heat and moisture. More destructive aspects of wood bending If the cell walls are distorted, as occurs when wood is dented, it can often be swelled back by moisture and heat, but only if the dent is fresh. The denting compresses by bending some doubled cell walls. A dog’s-leg double-bend zig-zag shape should be common. At each bend, there is more cell wall on the convex side than on the concave side. In the dog’s-leg, each cell wall is on the concave side at one bend and on the convex side at the other. In the denting process, that shape is taken by shearing the glue between the two cell walls in the region between the two bends. The internal stresses are in the bends in cell walls that originally grew straight. These residual stresses can drive a swelling-out of the dent if the surfaces between cell walls is made mobile by heat and moisture. If the dent is not swelled out quickly, the bends can migrate, either to each other or to the cell corners. In either case, the walls straighten out, but at the expense of the cell corners no more being at right angles. Possibly because of cleavage of the primary layer of cell walls at the acute-angle bends, the original cell shapes cannot be restored by heat and moisture. Curvature of wood perpendicular to the grain direction can be made permanent by high heat. The side on which it is applied becomes concave. The scorching or near-scorching heat breaks down much of the hemicellulose in the cell walls, thus contracting the wood on that side. This makes the heated wood lose most of its capacity to absorb water, so the contraction is permanent. Staves of bent-stave English 17th century viol bellies were bent this way with scorching irons, with the bends forming parts of the arching curve. Sound absorption by creep When one first tightens up the strings of a new bowed instrument, there are new bending forces on the structure. The string tensions tend to compress the length of the top plate, increasing the longitudinal arching curvature and raising the arching height. This is complicated by the downward pressure from the bridge, which tends to straighten out both the longitudinal and sideways arching curvatures in the bridge region. So the longitudinal arching curvature away from the bridge region tends to become even greater. These changes cause other distortions and changes of curvature over the body. Each change of curvature is subject to creep, and while creep occurs, it most probably absorbs vibrational energy whenever the instrument is vibrated. This is suggested because creep in string stretching absorbs vibrational energy, dulling the sound2 . The material uses that energy to speed up the creep. Creep is greatest at the beginning (most during the first week)3, and it slows continuously, eventually settling down to a negligible rate. Vibrating the instrument (as well as heat, moisture and the stresses of moisture gradients during humidity cycling4) will shorten the time it takes to settle down. The speeding-up of creep by vibration appears to be the mechanism by which playing-in helps5 . If an instrument is left with reduced string tension for some time, the changes of curvature along the grain cannot recover, but those perpendicular to the grain direction can to the extent allowed by the constraints (recovery would be helped particularly by the moisture gradients of humidity cycling). Then, when tuned up again, some creep could occur again, so some playing in may be appropriate. Changes of moisture content in equilibrium with a particular relative humidity We should not confuse the above effect of vibrating the wood during playing-in with any improvement of sound that can result from warming-up playing. There is evidence suggesting that vibrating wood has a lower equilibrium moisture content than wood under the same conditions but not vibrating6. The more moisture there is in the wood, the more absorption of vibrational energy there is. So, according to this theory, during warming up by playing, some of the moisture that was in the wood before is freed, and so less sound is absorbed. It is unlikely that this physical effect is large enough to be noticed. Since vibration affects the equilibrium moisture content of wood at constant temperature and relative humidity, we would expect static stresses to do the same. This seems to apply when two pieces of wood with different grain orientations are glued together at one relative humidity, and then the humidity changes. The glue joint holds, so one piece is under compression and the other under extension in the glued area. The wood region in compression can hold less water than it normally could at the new humidity, and that in extension more. It seems likely that wood under static stresses takes the equilibrium water content appropriate for its constrained dimensions rather than that appropriate for unstressed wood at the ambient relative humidity. Thus the weight of plywood varies much less with changes in relative humidity than normal woods. Another example of this principle is the observation that the cracks on centuries-old lute soundboards tend to terminate at the cross-bars glued underneath. The grain direction is along the long dimension of each cross bar, and its length would have varied very little with variation in humidity. This kept the soundboard wood glued to the cross bars from expanding or contracting, and thus it was kept at a relatively constant moisture content. The soundboard wood that was not next to the bars expanded and contracted with humidity changes, and the stresses resulting from moisture gradients associated with these changes enhanced the normal ageing effects of degrading the hemicellulose that contracted the wood. Thus the soundboard cracks between the bars result from contraction there that did not occur at the bars. Sound absorption by water The adsorbed water is a major contributor to the sound absorbed by the wood of musical instruments. Adsorbed water converts some of the energy of sound vibration into heat energy7 . There is typically a 3.5% decrease in damping coefficient for each 1% decrease in moisture content8. Since hemicellulose is the component of wood that adsorbs water, and its capacity to adsorb water depend on how much hemicellulose remains in the wood, the hemicellulose content is directly related to the amount of sound absorption. Thus instrument response, which depends on the sound vibration that is not absorbed, would improve as the hemicellulose degrades. This is probably the main reason why instruments made of matured wood have more response than those made of freshly dried wood, and why old instruments seem to have more response than newly-made instruments. Since hemicellulose itself most probably absorbs sound energy, its loss increases response more than just that due to the reduced moisture content. The closeness of the label date and dendrochronological date of some Guarnieri instruments suggests that wood maturation was sometimes considerably shortened, probably by stewing, which greatly accelerates hemicellulose degradation. It was traditional then to ‘salt’ wood to stabilise and preserve it9, and impregnated salts have been found in Guarnieri wood10. The salt helps dimensional stability by raising moisture content at low humidities, but the main effect on stability and sound is due to the hemicellulose degradation of stewing. Some makers today are stewing the wood used in their instruments to give the effect of aged wood. There is additional sound absorption by moisture gradients in the wood11. It appears that the sound energy absorbed is used to speed up the movement of water from regions of higher to lower moisture content. If there has been any change in the relative humidity around an instrument, this is probably a more important reason for warming up an instrument before performing on it than the small lowering of the equilibrium moisture content. Conclusion These theories explain all of the relevant evidence the author is aware of. Such theories are not for believing in, but should be respected unless and until theories that better explain the evidence emerge. In principle, these are all testable. They could be considered rather speculative since most have not been challenged by careful experiments. If such experiments were easy to perform, they would have been performed long ago, and appropriate theories formulated. Theories can result from or precede experiments. It is hoped that these theories will stimulate appropriate experimentation.   -------------------------- 1 A. J. Stamm, Forest Products Journal, Vol. 6 (5) (1956), p. 210. Cited in A. J. Stamm, Preprints of the Contributions to the 1970 New York Conference on Conservation of Stone and Wooden Objects, Second ed. (1971), Vol. 2, pp. 1-11. Almost of the statements made here about the chemistry of wood are derived from this source. 2 Players can often tell when a gut string is going to break soon by it needing more tuning (some fibres are already broken and the remaining fibres stretch more because their share of the tension is increased) and the sound gets dull (the fibres stretching absorb vibration energy). Also, it is well known in harpsichord circles that newly mounted brass strings don’t sound fully when first mounted, and only sound fully when they have stopped their initial stretching. 3 A. Beavitt, The Strad, (Nov. 1996), pp. 916-20. Beavitt claimed that all of the sound improvement in the life of a violin is associated with creep, which is facilitated by humidity cycling. 4 R. Hearmon and J. Paton, Forest Products Journal, Vol. 14 (8) (1964), pp. 357-9. They showed that the rate of creep in stressed wood is increased by humidity cycling. 5 E. Segerman, FoMRHI Quarterly, No. 84, (July, 1996), Comm. 1471, pp. 53-5. When new instruments are strung up, the wood deforms in response to the forces. The creep in this deformation absorbs sound vibrations, reducing response. In ‘playing in’, sound vibration accelerates the creep, making the instrument settle in faster. 6 D. G. Hunt and E. Balsan, Nature, Vol. 379, (22 Feb,.1996), p. 681. Sound absorption increases considerably in the non-equilibrium situation of rising moisture content. I interpret their experiment (Segerman, 1996) as showing that when wood is vibrated, the moisture content in equilibrium with a given outside relative humidity decreases. The effect is small and has only been observed at very high humidity. 7 D. Noak and H. Becker, Wood Science & Technology, Vol. 2 (1968), pp. 213-30. They showed that the damping of sound is strongly increased by increased moisture content. 8 D. G. Hunt and E. Balsan, op. cit. 9 R. Gug, FoMRHI Quarterly, No. 52, (July, 1988), Comm. 881, pp. 44-55. He reported that in 1580, Palissy wrote ‘Salt improves the voice of all sorts of musical instruments’. The impregnation of wood by salts was common practise. It was done by stewing the wood in the salt solution. The purposes usually stated for salting wood were to avoid rot, to repel woodworm and to stabilise it dimensionally (so it reacts less to weather changes). In my interpretation, it is likely that most of the sound improvement and stabilisation was due to hemicellulose degradation by the stewing. 10 J. Nagyavary, Chemical & Engineering News, (May 23, 1988), pp. 24-31. 11 D. G. Hunt and E. Balsan, op. cit.   -- Richard Brekne RPT, N.P.T.F. UiB, Bergen, Norway <A HREF="mailto:rbrekne@broadpark.no">mailto:rbrekne@broadpark.no</A> <A HREF="http://home.broadpark.no/~rbrekne/ricmain.html">http://home.broadpark.no/~rbrekne/ricmain.html</A> <A HREF="http://www.hf.uib.no/grieg/personer/cv_RB.html">http://www.hf.uib.no/grieg/personer/cv_RB.html</A>  </html>