Résumé
Wood behaviour is characterized by a high sensibility to humidity, swelling when humidity increases and shrinkage during drying. From green condition to ovendry condition, shrinkage ranges from 0.05 % to 0.3 % along the grain direction, however, two cases exist where longitudinal shrinkage starts to be more important: in reaction wood (tension wood of hardwoods and compression wood of conifers) and juvenile wood. In these two types of wood, axial shrinkage can reach 1 % or more. These changes in longitudinal shrinkage can be explained by the wood fibre structure. The knowledge of the wood cell structure, as a multi-layer fibre composite, allows the modelling of the longitudinal shrinkage, considering that the cell wall is reduced to S2 layer. These models give a good understanding of macroscopic axial shrinkage for different values of microfibril angle, for normal, compression and juvenile wood. However, they cannot explain the behaviour of tension wood. Tension wood cell wall is characterised by a very low microfibrils angle. In some species, important changes occur and part of the secondary wall is replaced by a distinct layer. This layer is described as purely or highly cellulosic, highly crystalline and with microfibril angle very low or nil. However, although a very low microfibril angle in G-layer, macroscopic longitudinal shrinkage is high in tension wood. Some years ago, we shown the experimental evidence of the G-layer shrinkage, thinks to the new observation techniques such as Atomic Force Microscopy but we were enable to explain it. Recent observations of some cutting arfefact in G-layer evidenced the weakly bonding of microfibrils in G-layer transversally to the fibre axis. Latter X-ray diffraction experiments proved that cellulose microfibrils d-spacing is very lightly shortened during axial shrinkage compare to macroscopic one. These observations conclude that cellulose was subject to buckling during G-layer shrinkage and cannot be the driving force of G-layer longitudinal shrinkage. The origin of G-layer shrinkage is then still in question. In this study, we concentrate on the characterisation of the non-cellulosique component of the G-layer, its organisation and its behaviour during drying.