Hyperspectral imaging and chemometrics to investigate the chemical wood modification

Abstract

Wood possesses an anisotropic hierarchical structure that causes a heterogeneous distribution of chemical reagents in modified wood at different spatial scales. Due to the heterogeneity in chemical distribution, localised regions of wood may remain susceptible to moisture uptake, dimensional instability, and fungal decay. The identification of regions with insufficient chemical uptake is necessary to develop efficient treatment processes, but standard gravimetric methods are insensitive to the location of chemical reagents within the wood. The primary objective of this thesis was to analyse the suitability of spectroscopic-based imaging methods to reveal the distribution of chemical reagents in modified wood at different length scales. The studies focused on the chemical modification of wood with acetic anhydride, paraformaldehyde, and thermosetting resins. The added chemical reagents are known to either react with cell wall polymers to create covalent bonds or to polymerize macromolecules within the cell wall space.

To analyze the chemical changes caused by the modification agents on different spatial scales, the studies combined two chemical imaging techniques that differ in their lateral resolution to identify the process-dependent heterogeneity in modified wood. Near-infrared (NIR) hyperspectral imaging identified and quantified the distribution of chemical reagents and the corresponding moisture content at a macroscopic scale of a few millimeters. Chemometric analysis not only revealed the sample-to-sample variations in chemical uptake and the associated moisture content but also highlighted the localised variations, most notably earlywood and latewood differences. Confocal Raman imaging validated the differences between earlywood and latewood on the cellular level and visualised chemical differences between cell wall regions. Following this, the moisture uptake and the consequent swelling of the modified samples were determined by the dynamic measurements of mass and dimensions within the hygroscopic range. The results indicated the effectiveness of chemical modifications in reducing the moisture content of untreated wood.

Overall, the results in this thesis demonstrated the ability of chemical imaging techniques to localise chemical reagents in small woodblocks and larger board sections. The findings provide a step forward in understanding the chemical changes caused by wood modification in different hierarchical structures in wood on different length scales. In the future, the methods may be used to characterise other treatments and processes that affect the wood composition.