1. Introduction
Wood, a porous natural polymer composite material, typically exhibits ample water sorption sites, specifically hydroxyl groups, when the saturated vapor partial pressure on the surface of the water contained in the pore is less than the environmental pressure [
1]. In response to fluctuations in the external humidity, wood adjusts its moisture content (MC) through absorption or desorption to attain equilibrium, and this affects the quality of wood products. Thus, the viscoelastic properties of unmodified and modified wood samples have been assessed through dynamic vapor sorption (DVS).
The moisture absorption characteristics of wood are contingent on the temperature and relative humidity of the external environment [
2], and moisture expedites surface photodegradation during outdoor exposure [
3]. Adsorption isotherms have been used to evaluate these characteristics, illustrating the correlation between the relative humidity (RH) and equilibrium moisture content (EMC) of the wood environment under a steady temperature and pressure. This correlation is pivotal in modifying the service lifespan of outdoor exposed wood.
Phenol formaldehyde (PF) resin, a water-soluble, low-molecular-weight polymer, bolsters the dimensional stability of wood materials by infiltrating and inflating their cell wall [
4]. Upon modification, a nonbonding network is established between wood and resin, reducing the hygroscopicity of wood. The hydrophilic hydroxyl groups in cell walls are partially obstructed, and the cell wall nanopores (nanocapillaries) become filled with PF resin. This process reduces the number of sorption sites for water and minimizes the space available for external moisture. The hygroscopic structure of wood is, thus, impeded by PF resin, yielding drier cell walls and an improved dimensional stability [
5].
Beech wood is classified as a softwood [
6]. In response to environmental conditions such as wind, softwoods exhibit adaptive growth, bending, and tilting of the trunk or branches under pressure, a phenomenon known as pressure wood; the non-stressed part is referred to as the corresponding wood [
7]. Compared with the corresponding or normal wood, pressure wood demonstrates a thicker tracheid wall, reduced cellulose content, and increased lignin content. However, the propensity of beech toward cracking and warping in practical processing, which are deemed as defects, hinders the efficient use of plantation wood [
8]. The water adsorption characteristics of wood correlate with the content of chemical components and the pore size of the cell wall. Consequently, understanding the water adsorption characteristics of pressure wood has both theoretical and practical implications for promoting the development and utilization of plantation wood.
Various theoretical models have been proposed to interpret the complex water adsorption process, including the Hailwood–Horrobin model, the Brunauer–Emmett–Teller (BET) model, the Guggenheim–Anderson–de Boer (GAB) model, and the Halsey model [
9,
10,
11]. Hills et al. [
12] used the parallel exponential kinetics model to analyze the water vapor sorption behavior of flax fibers, while Krabbenhoft et al. [
13] proposed a model for non-Fickian moisture transfer in wood. Moreover, Xie et al. [
14] investigated the water vapor sorption behavior of three celluloses using the Kelvin–Voigt viscoelastic model. Each model, however, exhibits deviations between the actual and measured values under varying RH conditions.
This study aimed to analyze the sorption behavior of untreated and treated beech samples, using PF resin of differing molecular weights at various concentrations. A DVS apparatus was used to record the sorption dynamics, sorption rate, and hysteresis in both unmodified and modified beech specimens.
3. Results
3.1. Dynamic Vapor Sorption (DVS) Behavior of Beech Wood
The DVS characteristics of both the unmodified and modified beech samples were investigated to determine the impact of PF resin modification. The response of a wood sample to a change in the set RH produces an asymptotic curve that approaches the EMC after an infinite time of exposure at a given RH. As depicted in
Figure 1, the EMC of all the wood samples increased with the RH in the adsorption stage, with the EMC first increasing rapidly before increasing slowly to equilibrium at each humidity level. In the desorption stage, the EMC gradually decreased with diminishing humidity for all samples, with the EMC decreasing rapidly at first and then slowly dropping until reaching equilibrium at each RH level. At the 95% stage, the EMC of the sample increases rapidly and cannot be completely balanced. As evident from
Figure 1, the PF resin modifications induced a decline in the measured MC during the sorption process, with the MC decreasing more noticeably with the increasing molecular weight of the PF resin. Compared with the EMC (23.7% at RH 95%) of the unmodified sample, that of the modified beech wood at M
w = 286 g/mol and M
w = 410 g/mol demonstrated a reduced MC (MCR) of 11.73 and 14.26%, respectively. The results suggest that PF resin modification considerably reduces moisture adsorption [
15]. Reportedly, low-molecular-weight resins penetrate the wood cell wall more easily, and, ultimately, bulk the wood material to a greater extent, thereby offering a superior dimensional stability; however, high-molecular-weight resins fill only the cell lumina [
16,
17], which diminishes the adsorption of hydrate water in low RH ranges. Moreover, PF resin can crosslink microfibrils, thereby constraining swelling when modified samples are exposed to humidity. In moist environments, water diffusion into the crosslinked cell wall creates a compressive stress because it cannot swell, contrary to its unmodified counterpart. Furthermore, the incorporation of PF resin into the nanopores of the cell wall pre-swells the cell wall, consequently decreasing the size (free space) of the nanopores as well as the cell wall swelling caused by water adsorption. Eventually, the accommodation for water (primarily dissolved and condensed water) diminishes in the high RH range [
18,
19].
3.2. DVS Isotherm Hysteresis Plot
Hygroscopic hysteresis, a common phenomenon observed in the water adsorption process of wood materials, is driven by a reduction in the number of accessible hydroxyl groups within the wood material. The number of available hydroxyl groups in wood that can bind to water molecules decreases with increasing water content, thereby diminishing the EMC value attained through moisture absorption. The modification of wood with the same exposure history with four PF resins of different molecular weights reduced the sorption hysteresis compared to that of unmodified samples (
Figure 2). Modified beech samples with a higher molecular weight (M
w = 410 g/mol) displayed less hysteresis at RH levels of 80% and 90% compared with unmodified beech samples. At a lower molecular weight (M
w = 286 g/mol), the sorption hysteresis of the modified beech was reduced across the full hygroscopic range, particularly in the high RH range, and the maximum hysteresis point shifted from 80% RH in the unmodified wood to 70% RH in the modified wood.
The reduction in hysteresis in beech woods can be attributed to the changes in the cell wall structure caused by PF impregnation. The hysteresis effect has been previously interpreted based on sorption phenomena on a glassy solid below the glass transition temperature [
20]. In the adsorption stage, the adsorption of the water molecules onto the sorption sites of cell walls is an exothermic process. The thermal motion of incoming water molecules causes the expansion of cell wall nanopores, and thereby creates a new internal surface. During desorption, the relaxation of the cell wall matrices to their state during adsorption is kinetically hindered [
21].
Modification with PF resin leads to crosslinking and bulking in the cell walls due to chemical incorporation [
22,
23]. Crosslinking reduces the degree of freedom of cell wall polymers, and bulking pre-swells the cell wall. Both effects contribute to an increase in wood stiffness after modification, leading to less deformation during the sorption process.
3.3. DVS Isotherm Adsorption and Desorption Plot
During the adsorption process, water molecules sever the hydrogen bonds between cellulose macromolecular chains, producing new free hydroxyl groups, which continue to interact with water molecules to form new hydrogen bonds. Moreover, the distance between molecular chains increases during the adsorption process. In the desorption process, hydrogen bonds between adjacent molecular chains break, and the distance between molecular chains decreases due to the elimination of water molecules. These phenomena result in changes to the physico-mechanical properties of beech samples, ranging from the cellular to the macroscopic level.
Figure 3 shows the adsorption and desorption plots of untreated and treated beech samples. When the RH exceeds 60%–70%, the water adsorption isotherm exhibits an upward-curving trend, displaying an S-shaped curve, indicative of a class II isotherm with multimolecular layer suction characteristics [
23]. The EMC of all wood samples increased with the increasing RH. Moreover, the adsorption curve of both the unmodified and modified wood samples steepened with the increasing RH. This phenomenon can be attributed to the adsorption of water vapor molecules onto the adsorption sites on the sample surface, causing the EMC value to increase during the early stage of adsorption. As adsorption persists, a monolayer forms, and a multimolecular layer gradually materializes, causing a swift increase in the equilibrium water content. In the later stage of adsorption, the adsorption gradually saturates, capillary condensation occurs, and the water content increases further. Compared with the modified wood, the unmodified wood samples exhibited a higher EMC upon adsorption and desorption.
The reduction in the EMC can be attributed to the increased quantity of PF resin deposited in beech samples, confirming that the EMC is considerably affected by the presence of PF resin in wood [
24,
25,
26,
27]. Typically, PF resin does not interact with wood components and deposits as a coating on the internal cell wall structure, as visualized via scanning electron microscopy. The impregnated PF resin can form irreversible hydrogen bonds within hemicellulose and amorphous cellulose, which are dispersed in cell wall voids. The decrease in the EMC of modified beech wood can be primarily attributed to the reduction in sorption sites. In conclusion, the modified beech wood samples exhibited the lowest hygroscopicity.
3.4. Mechanism of Adsorption and Desorption of Beech Wood Modified with PF
The process of wood production is extensive and subject to fluctuating ambient temperature and humidity, resulting in wood products undergoing varying moisture absorption or desorption processes and dynamic water-content changes.
Figure 4 depicts the mechanism of adsorption and desorption in beech wood modified using PF. Beech is a crucial broadleaf wood type. In broadleaf wood, a higher density corresponds to thicker wood fiber cell walls, smaller cell cavities, and a lower total porosity, resulting in slower water movement. The moisture absorption rate serves as a vital parameter to characterize the moisture absorption capacity of the wood. The measurement and investigation of this rate is critical to understanding the wood adsorption mechanism.
The moisture expansion rate of wood increases with the increasing RH of its environment. This can be attributed to the minimal variations in MC during the hygroscopic stage at a low RH, resulting in a thin hygroscopic water layer between the wood cell wall microfibers and minor alterations in the distance between microfibers. However, when the RH exceeds 95%, a substantial quantity of water infiltrates the space between the wood microfibrils, increasing the thickness of the water layer and resulting in a sharp increase in the distance between microfibrils. Therefore, the wet expansion rate at a high RH is substantially higher than that at a low RH, and large variations in water content remarkably increase the wet expansion rate and reduce the dimensional stability. At each stage of moisture absorption, the partial pressure of the water vapor inside the container is higher than that on the wood surface, inducing rapid water vapor absorption by the wood and a swift increase in the MC. During the hygroscopic process, the partial pressure difference of water vapor decreases, and the hygroscopic rate of wood gradually decreases until dynamic equilibrium is achieved. In all cases, the maximum adsorption rate occurs in the initial stage.
In the desorption process, the partial pressure of water vapor on the wood surface exceeds that inside the container, enabling wood to rapidly expel water, which results in an increased desorption rate. As the desorption process continues, the partial pressure difference in the water vapor decreases, gradually reducing the desorption rate of the wood.
Initially, water molecules form a single layer of molecular water adsorption by interacting with polar groups at the adsorption sites. However, the number of adsorption sites on cellulose and hemicelluloses inside wood is limited, with most binding sites occupied at a low RH. When the RH increases, the single-layer molecular adsorption curve remains relatively gentle owing to the limited adsorption capacity. The single-layer molecular adsorbed water then interacts with water molecules again to form multilayer molecular adsorbed water, which increases with the increasing RH until it reaches equilibrium with the single-layer molecular adsorbed water. As the RH increases, the adsorption capacity of multilayer molecules increases rapidly, causing a swift ascent in the multilayer adsorption curve. Therefore, studying the dynamic change in the MC of beech wood under different environmental temperatures, humidities, and moisture absorption processes is vital to guide the use, storage, and drying of wood products.
4. Conclusions
The modification of beech wood using PF resins of various molecular weights led to a decrease in the EMC. The results demonstrated that, compared with unmodified beech samples, all PF resins of different molecular weights exhibited a reduced sorption hysteresis. Furthermore, all sorption isotherms for unmodified wood and modified beech wood with PF resins of different molecular weights displayed characteristic sigmoidal shapes. The decrease in the EMC value of the modified beech samples can be attributed to the deposition of PF resin in the nanopores of the beech wood cell walls, which reduces sorption sites.
Furthermore, the results revealed that the water content absorbed by the sample increases in stages with increasing humidity. In each RH stage, the water content of the sample initially increases rapidly, then decreases gradually until equilibrium is attained. In the desorption stage, the MC of the sample decreases in stages with decreasing humidity. In each RH stage, the MC of the sample rapidly declines, then gradually decreases until equilibrium, in contrast to the adsorption stage. Impregnating beech samples with PF resin improves their hygroscopicity. DVS can provide valuable insights into the structure and cell wall composition of beech wood fibers during their development.