Moisture transport processes in Scots pine - Anomalous capillary suction. Nonisothermal diffusion.

Doktorsavhandling, 2007

Moisture transport processes in wood have become increasingly important with the growing use of wood as construction material. This thesis deals with two key processes: capillary suction and nonisothermal diffusion. In the study on capillary suction, several methods were used. In the main study, sticks from pine sapwood and heartwood of different lengths (100-400 mm) were placed with the end grain in water. After 1000 hours and 1900 hours of exposure, the moisture content (MC) distribution in each stick was determined. The results from the main study, regarding the sapwood sticks, were peculiar. For the short sticks (100 mm), the MC distributions started out high in the bottom parts and decreased as expected toward the middle. But from the middle to the top, the MC/ increased/. The longer sticks (200 and 300 mm) showed a similar strange behaviour. The penomenon was seen for both exposure times. In the longest sticks (400 mm), this behaviour was not seen. These results show a very non-Fickian nature. In the transient flow, which gradually approached steady-state, MC gradients developed with the flow going/ from a// low MC to a higher MC/ in the upper parts of the samples. It is from this clear that when dealing with capillary suction, a Fickian approach on the macro scale may be inadequate. An explanation on the fiber level is proposed, where the anomalous behaviour is explained by water penetrating some flow channels, being spread laterally on the top surface and then sucked back into unfilled fibers. In the nonisothermal studies, a new method was developed using samples pre-conditioned to high or low MC to avoid hysteresis. Six series of diffusion cups were mounted in the lid of a cooling box that was placed in a climate room, establishing temperature and moisture gradients over the samples. Simultaneously, other cups with matched samples were placed in the same room, only without the temperature gradient. At steady-state conditions, the moisture fluxes and the MC distributions of the samples ere determined. The temperature of the top and bottom surfaces of the samples was monitored. In the study, the applied boundary conditions resulted in pairs of isothermal and nonisothermal MC distributions and moisture fluxes. The results from the experiments were used in a two-term flux model to evaluate the moisture transport coefficients/// D_u / and/// D_Tu /. From experiments and mathematical derivations, it was shown that the moisture state variable can be chosen at will, none distinctively better than any other. Further, it was shown that two sets of measurements, isothermal and nonisothermal, are needed for an accurate evaluation.