Nanoscale Mechanism of Moisture-Induced Swelling in Wood Microfibril Bundles

Abstract

Understanding nanoscale moisture interactions is fundamental to most applications of wood, including cellulosic nanomaterials with tailored properties. By combining X-ray scattering experiments with molecular simulations and taking advantage of computed scattering, we studied the moisture-induced changes in cellulose microfibril bundles of softwood secondary cell walls. Our models reproduced the most important experimentally observed changes in diffraction peak locations and widths and gave new insights into their interpretation. We found that changes in the packing of microfibrils dominate at moisture contents above 10-15%, whereas deformations in cellulose crystallites take place closer to the dry state. Fibrillar aggregation is a significant source of drying-related changes in the interior of the microfibrils. Our results corroborate the fundamental role of nanoscale phenomena in the swelling behavior and properties of wood-based materials and promote their utilization in nanomaterials development. Simulation-assisted scattering analysis proved an efficient tool for advancing the nanoscale characterization of cellulosic materials.

Keywords: Cell wall nanostructure; Cellulose crystallinity; Molecular dynamics; Wood-water interactions; X-ray scattering.

Figure 1

Cartoon representation of the hierarchical structure of spruce wood. The wood tissue (bottom right) consists of elongated tracheid cells, whose secondary cell walls bear the load of the tree. Cellulose microfibril bundles (top right) that consist of microfibrils with a diameter of 2–3 nm are a central component of the secondary cell wall structure. Lignin domains between microfibril bundles are not depicted.

Figure 2

X-ray scattering at different moisture conditions, as observed experimentally (a–c) and computed (e,f) from molecular models such as in (d). (a) Equatorial anisotropic SAXS and WAXS intensities on double-logarithmic axes with the isotropic scattering contribution shown with light-colored lines. A schematic representation of the experiment is shown above the image. The main reflections from cellulose I_β and the equatorial and meridional integration sectors are indicated in the two-dimensional scattering patterns shown as insets. (b) Equatorial and (c) meridional anisotropic WAXS intensities, labeled with the corresponding moisture content (MC). The peak marked by a dashed line in (a) and (b) is due to unsuccessful background subtraction (Kapton windows), see SI for details. (d) Molecular model of a microfibril bundle of four microfibrils in a periodic simulation domain (domain outlined in black). MC_c refers to the moisture content relative to the carbohydrates. (e) Equatorial and (f) meridional scattering intensities computed fibril-by-fibril from the model shown in (d), excluding water, labeled with the MC_c (see Figure S7 for scattering intensities computed from all models).

Figure 3

Moisture-dependence of the crystalline parameters based on WAXS experiments and molecular models. (a,b) Lattice spacings d_hkl and (c,d) crystal size L_hkl corresponding to different directions hkl in the crystal, as determined experimentally and from the periodic bundle model either directly (see Figure S4) or from computed scattering intensities. The MC for the models (MC_c) has been multiplied by 0.7 (approximate mass ratio of carbohydrates and total dry mass) to make it comparable with results from complete cell walls. The experimental values shown for the 200 peak correspond to a weighted average of a bimodal size distribution of crystallites (see SI for details). Note the separate vertical axes for experimental (right) and model-based (left) results for hkl = 004 in (a,c). See Figures S9, S10, and S11 for more results.

Figure 4

Moisture-related swelling of microfibril bundles based on scattering analysis and modeling. (a) Results of SAXS fits using the WoodSAS model (SI eq S3) (see Figure S13 for all results). (b) Microfibril packing distance as a function of MC (0.7 × MC_c) as determined from the periodic bundle model (see Figure S15 for non-periodic model) and a comparison to experimental SAXS and SANS results^,,,,, (MC of 41% used for Norway spruce in water-saturated state; error bars correspond to standard deviation between samples). Snapshots of the periodic model with density of water in blue are shown on the right. (c) Number of hydrogen bonds as a function of MC_c (see Figure S16 for all results). (d) Fractions (f in %) of primary alcohol group conformers (tg, trans–gauche; gt, gauchetrans) as a function of MC_c, shown separately for all cellulose chains and chains at microfibril core and surface (see Figure S17 for all results). In panels c and d, the lines correspond to the mean of all four fibrils and the filled areas to ±1 standard deviation of the means of each fibril. (e) Simulation snapshots of two microfibrils of the periodic model (cellulose in green and orange, hemicelluloses in gray), which illustrates the deformation of cellulose crystals in the dry state. This partially explains the changes in lattice spacings and experimentally determined crystal size (Figure 3).

Figure 5

Summary of structural changes taking place in microfibril bundles in spruce wood cell walls at different MC ranges, illustrated in connection with a sorption isotherm (circles indicating the data points). In the dry state (MC close to 0%), the cellulose microfibrils are tightly packed and locked into a position where the cellulose crystallites are deformed due to interaction with the matrix polymers and neighboring fibril surfaces. Once some water becomes available (MC a few per cent), it penetrates the matrix and forms small clusters. Further increase of the MC allows the interfibrillar matrix to reorganize and to accommodate more water, which releases the deforming stresses on the microfibrils and allows them to adopt a higher level of crystalline order. At around MC 10–15%, the individual water clusters have grown and merged enough to occupy the spaces between the microfibrils, and the matrix softens in a process reminiscent of a glass transition. Above this point, the matrix can deform more easily, and it expands to accommodate more water in continuous channels, and the fibril bundles swell causing an increase in the microfibril packing distance and allowing faster diffusion of water. Simulation snapshots with MC referring to 0.7 × MC_c are shown on the right (see Figure S19 for all).