Wood's piezoelectric potential
Wood is not known for its conducting ability, yet researchers at the University of Texas are using loading analysis and finite element modelling techniques to unlock its properties. Dr Dan Wheat and Dr Loukas Kallivokas apply the current thinking.
Imagine the novelty of taking an electrocardiogram of a wood structure. Electrodes on the surface would tell of surface voltages or charges while the structure is under stress, and the pattern and magnitudes of these voltages and charges would then be calibrated to relay information about what is beneath in terms of growth characteristics that can affect strength and the grade of the wood. Even selected information about stresses could be ascertained. While such simplistic reckoning is far from reality at the moment, much of the technical infrastructure exists, but needs laboratory development and refinement, plus a healthy dose of attendant mathematics.
Wood is endowed with a property that makes it a self-sensing material – it is a piezoelectric polymer. Within its cell wall structure are geometrically near-regularised regions of microfibrils – groups of polymer chains with sufficient crystalline-like behaviour that electrical polarisation occurs when stress is applied. The result is a potential difference on various surfaces. The electric polarisation of wood under stress and its converse effect – deformation when exposed to an electric field – qualify it as a true piezoelectric material.
Under a general state of three-dimensional stress, polarisation occurs in the longitudinal (L), radial (R), and tangential (T) directions. These are the orthotropic directions normally assumed at any point in wood. At a cut cross-section of a log and at a specific point, the L-direction refers to the direction of trunk growth, the R-direction is radial to a growth ring, and the T-direction is tangent to a growth ring. Conventional wisdom assigns a preference for selected shear stresses as the dominant contributors to the polarisation. This electrical endowment, once captured and tamed, literally may signal the presence of strength-affecting growth characteristics, threatening levels of in situ stress, or a myriad of other possibilities.
The first reports of the piezoelectric effect in wood were made in 1941 in the former Soviet Union. Since that time, there have been many studies of the effect by wood chemists, wood physicists, and engineers, with the largest number of researchers in Japan. Research in the piezoelectric properties of wood continues, from the atomic level up to small-sized clear wood specimens, with all manner of research themes. In the 1980s, research was conducted on larger-sized dimension lumber members, and there have been measurements – although few – of electric charge and voltage potentials near knots of stressed statically-loaded and vibrated specimens, an especially promising development. Potential applications in grading of wood were cited.
Despite the continued research in piezoelectricity in wood, the applications to the larger structural members inexplicably nearly ceased. It may be posited that the reasons may have been related to the difficulty in measuring charges or voltages, the perceived relevance of the topic in general, or the real or perceived hardware and software needs to bring the area to full fruition. Whatever the cause, research on dimension lumber members faded away.
We want to see its return on a larger level, and in a humble effort to revive interest in applications of piezoelectricity to wood, with structural health monitoring and grading in mind especially, attention turned to a set of small wood specimens that had been tested in Japan. The team at the University of Texas attempted a multiphysics finite element modeling process in order to predict the piezoelectric responses, namely surface charges, seen in some of the Japanese tests. Such modelling required a combination of specific piezoelectric moduli measured by the Japanese researchers as well as some conjectured values for dielectric properties in each of the L, R, and T directions along with some of the orthotropic elastic properties. Anyone familiar with three-dimensional modelling of wood knows that some educated conjecturing is essential, since for any given piece of wood, knowledge of all 12 elastic constants as well as the full menu of electrical properties, is unfeasible in practical time constraints.
The small specimens tested by the Japanese were dry 8 x 15 x 15mm blocks with the load directed as shown. There were two sets of specimens and in both the electric polarisation occurred in the 8mm direction, which was the T direction for one set of four specimens and the R direction for the other set of four. Moreover, in order to maximise the shear stress and presumably also the piezoelectric response, as was noted above, the L direction in each is at 45° with respect to the applied load.
The modeling efforts were promising, since predicted values of electric charge were nearly identical for four of the eight specimens. The differences in the remaining four ranged from 23–28%. Computed values were lower than the measured values for seven of the eight specimens.
Knots and volts
The extant literature in the field included several studies in the US, Japan and South Africa on the piezoelectric effect in larger wood members that possessed growth defects, such as knots. A small number of reports of voltage and charge distributions near knots were among these studies, so attention was devoted to continued finite element modeling to a small board containing a knot. The objective was to perform numerical experiments similar to the test specimens in order to detect any correspondence in predictions and measurements. The board, lying in the L-T plane, had a cross section of 2.5x10cm and a length of 25cm. A tensile stress of 6.9MPa was applied in the L direction. The electrical and mechanical properties of the board were chosen to be representative of the realistic values of those used in modelling the blocks tested in Japan and described previously, while the knot itself was assumed to be electrically neutral. A number of parameters were included in the parametric studies: slope of grain in the board, knot size, shape, location, and angle of penetration through the board. In all cases of knot penetration, the knot was assumed to be parallel to the 2.5cm side of the board. Selected results of the case of a board with a single knot and zero slope of grain are shown.
Finite element modelling of predicted voltages for two boards - Top: Elliptical knot piercing board perpendicular to the broad side. Bottom: Knot piercing board at 60 degree angle
The images show a wide face of the board and predicted voltages for two cases: an elliptical knot at the centre of the board piercing it perpendicular to its broad side, a similar knot, but piercing the board at an angle of 60 degrees from the perpendicular. In the first case, symmetry of predicted voltage patterns is evident, while in the second the higher voltages are redistributed toward the left and the acute angle of incidence of the knot. Again, results are encouraging and are consistent with voltage and charge patterns near knots that have been reported. Other computational experiments were also done.
This provides a brief glance at an area that is felt to have great potential by recognising the not-so-well-known electrical properties of wood. For many years, electrical properties such as dielectric constants and capacitance of wood have been used to determine moisture contents. More recently, German researchers have attempted to use piezoelectric responses caused by drying deformations to discern moisture content in wood. However, it is recognised that there are many years of laboratory and computational research required to bring about even greater engineering utility of the piezoelectric endowments of wood. Piezoelectric properties of wood are influenced greatly by moisture content, temperature, and even static vs cyclic and impulsive loads, to mention only a few critical parameters.
Much work awaits researchers to perform the required laboratory and computational studies of mechanical and electrical properties of wood. These include work on various scales, from atomic level investigations of cellulose to stress analysis of larger structural members having growth defects. Further clarification of the tensor relationships of polarisation and stress is essential. In any event, the tools exist for further study that could make an electrocardiogram of wood a reality.
Dr Dan Wheat and Dr Loukas Kallivokas, Civil, Architectural and Environmental Engineering Department, the University of Texas at Austin, USA.