Water and Cellulose

Adsorption and diffusion of water in the pores of amorphous cellulose.

Amorphization of a cellulose chain.

Cellulose is an amazing biopolymer — present in plants' cell wall, it is the most abundant polymer on Earth. Mostly in crystalline state, cellulose has the longitudinal Young's modulus of approximately 160 GPa, which is not very much less than that of steel! Cellulose crystals are mechanically anisotropic with hexagonally-faced microfibrils that are amphiphilic. In cell walls cellulose forms, together with other non-crystalline polymers, a true bio-composite that reacts to its environment's humidity.

Molecular Dynamics structure of amorphous cellulose: chain and pores of bulk state. [Macromolecules 2015]
Sorption site in glucose ring.

Amorphous cellulose, in contrast, has much smaller modulus (of ~10 GPa) and is characterized by a nano-porous structure with exposed hydroxyl sites which makes it strongly water-absorbing. The adsorbed water is known to rearrange the structure of amorphous cellulose and to drive changes in its mechanical properties and geometry. Because moisture sorption occurs mostly at the atomic scale, the Molecular Dynamics (MD) simulations are probably the best tool to investigate these phenomena. Despite their limited spacial and temporal capabilities, MD simulations are fully capable to reproduce the hygroscopic swelling and moisture softening effects in natural polymeric systems.

Clusters of water adsorbing in the pores of amorphous cellulose. [Polymer science... 2016]

The MD model of cellulose is simulated with GROMACS package using Gromos 53a6 force field. The atoms interact with each other through Coulomb and Van der Waals forces, and additionally through covalent bonds within chains. The studied structures are initially constructed in the dry state and then subsequently hydrated to different water concentrations, up to the respective saturation value. The models are in full periodic boundary conditions in order to minimize the finite-size effects. The simulations are carried out at 300 K regulated by Nosé-Hoover thermostat and at stress-free conditions by the applied Parinello-Rahman barostat. The amount of adsorbed water molecules (or concentration) depends on the external chemical potential μ that is linked to the relative humidity RH through Kelvin equation,

RH = exp(-(μ-μsat)/kBT) .

Water density profiles in amorphous cellulose depending on moisture content.

The higher the chemical potential, the higher the RH and thus the higher the water concentration. The dependence between chemical potential (or RH) and concentration is the adsorption isotherm that is a characteristic of a material. Their typical exponential-like shape is a result of the interaction with other water molecules. At the initial stage of adsorption, when the material is nearly dry, the water molecules are dispersed more or less uniformly within the material’s pores such that the water molecules do not form clusters or feel the presence of each other. This phenomenon and the availability of the sorption sites, whose binding energy is strongly negative, makes the chemical potential at the initial stage of adsorption differ by 20 kJ/mol or more from the saturation state chemical potential. Arrival of next water molecules into the pores increases the chemical potential as the molecules start forming clusters and the ratio of sorption sites to water molecules can easily exceed unity. This means that there are two types of adsorbed water molecules: bound at sorption sites and those that are not. The not-bound molecules are in a state close to bulk liquid and their relative amount increases with concentration by which the average chemical potential approaches that of saturated vapor or, equivalently, liquid water.

Swelling of amorphous cellulose upon increasing the number of adsorbed water molecules.

The adsorption of water in hydrophilic polymers such as cellulose is a complex process that is mainly driven by the formation and breaking of the hydrogen bonds. Due to porous character and flexibility of biopolymers, water molecules can be easily adsorbed attracted by strong energy of hydrogen bonds. It is worth noting that breaking of hydrogen bonds is a reversible process enabling the polymers to regain their original shape and stiffness upon desorption and is one of the main mechanisms in moisture-induced shape memory effects of wood tissue.

Dynamics of hydrogen bonds near cellulose surface. [Langmuir 2017]

The diffusion coefficient of water adsorbed in hydrophilic porous materials, such as noncrystalline cellulose, depends on water activity. Faster diffusion at higher water concentrations is observed in experimental and modeling studies. An increasing water concentration leads to significant changes in the free energy landscape due to the perturbation of local electrostatic potential. Smoothening of strong energy minima, corresponding to sorption sites, and formation of layered structure facilitates water transport in the vicinity of cellulose. The transition probabilities and hydrogen bond stability reflect the changes in the energy landscape. As a result of a concentration increase, the emerging basins of attraction and spreading out of those existing in the diluted state lead to an increase in water entropy. Thermal fluctuations of cellulose rearrange the landscape in the diluted or low moisture content state, increase adsorbed water entropy, and decrease the water-cellulose hydrogen bond lifetime.

Basins of attraction for water near cellulose surface. [Langmuir 2017]

At low moisture content, a state difficult to obtain in experiments without damaging the material, the behavior of a porous material is different than at higher moisture content. Near the percolation threshold, many properties of the hydrated system undergo a qualitative change. The moisture content, at which water network starts to percolate, separates two regimes. The low moisture content regime, up to the inflection point of the sorption curve, is characterized by low diffusion, little swelling, and no change in elastic modulus. The water clusters comprise up to 10 water molecules, and the number of clusters increases with moisture content. The water molecules are mainly located in the existing voids, strongly bonded to the cellulose sorption sites. This causes the molecular transport to be very slow, resulting in a small diffusion coefficient. The water present in amorphous cellulose does not weaken the van der Waals and Coulomb interactions and causes little swelling — so it doesn't impact the bulk modulus.

Percolation threshold visible in the qualitative transition in modulus (left) and diffusion coefficient (right). [ACS Macro Letters 2014]

Above certain moisture content water network percolates and the system enters the second regime. The clusters start to merge, decreasing in number and increasing in size. A linear swelling curve is measured as the adsorbed water molecules push away the cellulose chains, making space for additional molecules, and every inserted molecule increases the volume equally. Swelling occurs because the solvation pressure of the fluid is significantly larger than the bulk pressure and results from an interplay between the adsorption energy, steric repulsion, and cohesive forces (van der Waals or H-bonds).

Reaching percolation threshold by increasing moisture content.

In contrast to swelling, a decrease in stiffness is a more complex process. It involves breaking of hydrogen bonds and weakening van der Waals and Coulomb interactions, by increasing both the chain-to-chain distance and the surface area of the voids occupied by water. Apart from water trapped at the sorption sites, there are molecules that are not so strongly bonded to cellulose. This, in turn, leads to greater mobility and consequently an increase of the diffusion coefficient. In the case of pure amorphous cellulose, contrary to systems containing crystalline cellulose, swelling is not restrained because the amorphous matrix is soft and can easily yield to increase the volume of the pores. (For wood in general, the diffusion coefficient increases with relative humidity throughout the whole range.)

We find it remarkable that the percolation co-occurs with the observed changes, although there is no direct physical causal effect between the percolation and the change of mechanical properties that are more related to the breaking of hydrogen bonds and other structural changes.

The presented research partially fills up the missing knowledge in sorption-induced phenomena occurring in natural systems containing hydrophilic polymers (e.g. wood). The applied methodology can be particularly useful in understanding the process of functionalization of natural polymers (hydrophobization of sorption sites) and facilitates upscaling to continuum methods.

If you would like to know more details about this topic, feel free to contact us or check the provided links and references.

author: Dr. Karol Kulasinski

References & External Links


Kulasinski, K., (2016) Effects of water adsorption in hydrophilic polymers, in Polymer science: research advances, practical applications and educational aspects. Eds. Méndez-Vilas, A. and Solano-Martín, A., pp. 217–223.

Kulasinski, K., (2017) Free energy landscape of hydrophilic polymers as a driving factor in water diffusion. Langmuir 33: 5362-5370.

Kulasinski, K., Derome, D. and Carmeliet, J. (2017) Modeling of hydration effects on wood cell wall secondary layer using molecular dynamics simulations. Journal of the Mechanics and Physics of Solids 103: 221-235.

Kulasinski, K. and Guyer, R. (2016). Quantification of nanopore networks: application to amorphous polymers. Journal of Physical Chemistry C 120: 28144-28151.

Kulasinski, K., Salmén, L., Derome, D. and Carmeliet, J. (2016). Moisture adsorption of glucomannan and xylan hemicelluloses. Cellulose 23: 1629–1637.

Kulasinski, K., Guyer, R., Derome, D. and Carmeliet, J. (2015). Water diffusion in hydrophilic systems: a stop and go process. Langmuir 31: 10843–10849.

Kulasinski, K., Guyer, R., Derome, D. and Carmeliet, J. (2015). Water adsorption in wood microfibril: role of the crystalline-amorphous cellulose. Biomacromolecules 16: 2972–2978.

Kulasinski, K., Guyer, R., Derome, D. and Carmeliet, J. (2015). Poroelastic model for adsorption-induced deformation of biopolymers obtained from molecular simulations. Physical Review E 92: 022605.

Kulasinski, K., Guyer, R., Keten, S., Derome, D. and Carmeliet, J. (2015). Impact of Moisture Adsorption on Structure and Physical Properties of Amorphous Biopolymers. Macromolecules 48: 2793–2800.

Kulasinski, K., Keten, S., Churakov, S., Guyer, R., Derome, D. and Carmeliet, J. (2014). Molecular Mechanism of Moisture-Induced Transition in Amorphous Cellulose. ACS Macro Letters 3: 1037–1040.

Kulasinski, K., Keten, S., Churakov, S. V., Derome, D. and Carmeliet, J. (2014). A comparative molecular dynamics study of crystalline, paracrystalline, and amorphous states of cellulose. Cellulose 21: 1103–1116.