Cellulose, the world's most abundant natural, renewable, biodegradable polymer, is a major component of plants, wood, algae, bacteria and is widely used in food, pharmaceutical, paper, textile production, or in wastewater treatment applications. A remarkable property of cellulose-based materials is that they are hygroscopic: they absorb huge amounts (typically 25% of the sample dry mass) of water from ambient vapor, inside their solid structure, thanks to the high affinity of water molecules to the polarized OH groups of cellulose molecules. Among others: the sorption of bound water in the structure is at the origin of the swelling or shrinkage of wood; the extraction of the residual bound water fraction is the major energy consumer in the paper mill; bound water transfers controls the moisture buffering characteristic of bio-based construction materials; and it plays a significant role in the moisture storage and transport, and heat loss due to sweating in textiles.

Despite its ubiquitous importance, the transport properties of bound water in cellulosic structures, which govern the absorption or desorption dynamics, are poorly known. Here, we show how straightforward measurements of bound water diffusion in various materials may be carried out with the help of specific desorption set-ups with controlled boundary conditions and original NMR or MRI techniques. Focusing on water transport through a cellulose fiber piling, we demonstrate that the bound water is able to move all along the network made of fibers in contact. The transport dynamics can then be quantified precisely through drying experiments (inducing concentration gradients) under different conditions. Surprisingly, although the bound water molecules are confined in nanometric pores between cellulose microfibrils, the diffusion coefficient of bound water along a fiber axis appears to be very close to the self-diffusion coefficient of water in liquid bulk. Similar results are obtained with wood samples. Finally, we discuss how these results, completed by further tests for determining the vapor diffusion coefficient through such porous networks, can be used to predict water transfers through a model textile or a bio-based insulating material.