Introduction

Wood, as a lignocellulose material, has served human in field of building and construction for many years. Recent increasing concerns on the environment and sustainability have urged people to develop renewable materials as substitute for the nonrenewable (Li et al. 2022). Compared to steel and cement, wood is more renewable, sustainable and safer in some cases (Ramage et al. 2017; Guo et al. 2022). As natural forests decrease, fast-growing wood attracts increasing attention for its short growth cycle and relatively abundant resources. However, the physical–mechanical performances of fast-growing wood are not satisfied for direct application in buildings, constructions or others. Developing new technologies to improve the performance of fast-growing wood is significantly important for sustainable material science. Preparing fast-growing wood into wood-polymer composites was an efficient method to upgrade wood. Some monomers or prepolymers can be introduced into the hierarchical structures of wood, such as low molecular weight formaldehyde-based resin (FBR), 1,3-dimethylol-4,5-dihydroxy-ethyleneurea (DMDHEU) and furfuryl alcohol (FA) (Xie et al. 2013).

Compared with FBR and DMDHEU, furfuryl alcohol (FA) is derived from abundant biomass residuals including rice hulls, bagasse, and corncobs, etc. (Gong et al. 2017; Yang et al. 2019). Under the action of heat or catalyst, FA can in-situ polymerize into hydrophobic furfural resin in wood cell walls steadily. The wood composite with in-situ polymerized furfural resin (WFC) usually presents better hydrophobicity, dimensional stability, compressive strength and bio-resistance, etc. (Lande et al. 2008; Yang et al. 2022). Importantly, WFC shows low toxicity and releases less volatile organic compounds or polycyclic aromatic hydrocarbons after combustion (Lande et al. 2004; Pilgård et al. 2010). In other words, WFC has slight impacts on environment comparatively. However, the toughness decrease of WFC is the major issue to be addressed. Similar to other resins, like phenolic resin and urea resin, the high crosslinking density and steric hindrance of molecules provide the resin with inherent brittleness (Zheng et al. 2015; Lu et al. 2022). When the brittle furfural resin polymerizes in wood cells and pores, the stress concentration would easily occur and accelerate the crack propagation. Many researches have been conducted to alleviate the toughness reduction caused by furfurylation. The introduction of long-chain polymer molecule is an efficient method, such as epoxidized soybean oil, polyethylene glycol and polyvinyl alcohol (PVA) (Liu et al. 2021; Wang et al. 2022; Yang et al. 2022). PVA is a water soluble, non-toxic and odorless polymer with flexible long molecular chains. PVA possesses excellent flexibility and toughness (Wu et al. 2012; Rao et al. 2019). Besides, PVA has better biodegradability and can be degraded in natural environment without aggravating white pollution when discarded (Stevens 2002; Wu et al. 2012). As a polyol, PVA can participate in chemical reactions, such as esterification, etherification and acetalization. Lin et al. (2008) introduced the PVA into urea–formaldehyde furan resin and found that PVA interrupted the cross-linking of resin and toughened the resin. Wang et al. (2022) reported that PVA could also interrupt the self-crosslinking of furfural resin and alleviate the reduction of impact toughness. However, PVA has a high molecular weight, which negatively influences the penetration of FA into cell walls of wood and further affects the properties of materials.

Recently, many studies have used delignification in preparation of advanced functional materials or composites (Chen et al. 2020), such as transparent wood and magnetic wood. Delignification is a process to remove the lignin from lignocellulose materials by chemical, physical or biological treatments (Zhao et al. 2006). Lignin is one of the main components of wood and makes up about 20–30% of the dry weight of wood (Côté 1968; Sjöström 1981). After a certain delignification, a large number of new mesopores appear in wood cell walls (Liang et al. 2020). Due to the porosity enrichment of wood after delignification, it is much easier to impregnate modifiers or functional agents into wood. In addition, lignin exhibits relatively stiffness in wood matrix (Kumar et al. 2021), and a decrease in lignin content can soften or weaken the rigidity of matrix components and increase the fracture toughness (Zhang et al. 2006; Zhang et al. 2011). Besides, previous researches showed that the removal of lignin from the cell wall increased proximities (or aggregation) of cellulose fibrils and thus helped the increase of mechanical performances of wood (Salmen and Burgert 2009; Kumar et al. 2021).

Based on above, we proposed a new technology to fabricate wood-polymer composites by combining the synergistic advantages of introduction of FA and PVA and delignification. Hydrophobic FA resin and PVA were hypothesized to bulk wood cell walls and hinder the wood-water interactions. PVA with flexible long chain was also expected to interact with FA and improve the toughness of composites. Delignification was used to facilitate the penetration of FA and PVA into cell walls and reduce the rigidity of materials. The objective of this study was to process low quality fast-growing wood into wood-polymer composites with excellent hydrophobicity, dimensional stability and mechanical strength and reveal the modification mechanism by analyses of Scanning electron microscope (SEM), Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD). The fabrication and mechanism analysis of wood-polymer composites were helpful of developing new technologies of upgrading low-quality wood products.

Materials and methods

Materials

Planted poplar (Populus cathayana Rehd.) was collected from Hebei province, China. Sapwood was selected and the average air-dried density was about 380 kg·m−3. Wood blocks without any defects were processed into size of 4 mm (longitudinal) (L) × 20 mm (tangential) (T) × 20 mm (radial) (R), 120 mm (L) × 10 mm (T) × 4 mm (R) and 80 mm (L) × 10 mm (T) × 4 mm (R), respectively.

The chemicals including ethanol, benzene, sodium chlorite, acetic acid, FA, maleic anhydride, sodium borate, sodium bromide and potassium carbonate were purchased from the Shanghai Lingfeng Chemical reagent Co., Ltd. Polyvinyl alcohol (PVA) (degree of hydrolysis = 87–89%) was purchased from Alfa Aesar (China) Chemical Co. Ltd. All above chemicals were of analytical grade.

Methods

Delignification

Samples were immersed into the mixture including ethanol and benzene (1:2 volume ratio) for 48 h. Then samples were transferred to a water bath at 60 °C for 3 h. Subsequently, the treated samples were rinsed with running distilled water for 24 h for draining off the residual chemicals. Then all the extracted samples were air-dried and oven-dried at 80 °C in a vacuum oven. A uniform mixture of 967 ml distilled water, 20 g NaClO2, and 13 ml CH3COOH was prepared. Then samples were immersed in the mixture and vacuum-treated under pressure of − 0.1 MPa for 5 h. Subsequently, the samples were transferred to a water bath at 40 °C for 10 h. After that, all the samples were rinsed with running distilled water until the pH of water after rinsed was about 7. Finally, the samples were air-dried and then vacuum-dried at 80 °C until reaching constant weight (Yang et al. 2021).

The mass change (lignin removal) and volume change (VC) was calculated according to Eq. (1) to (2).

$${\text{Mass change}} (\% ) = \frac{{m_{1} - m_{2} }}{{m_{1} }} \times 100\%$$
(1)
$${\text{VC}}(\% ) = \frac{{{\text{(T}}_{1} \times {\text{R}}_{1} \times {\text{L}}_{1} {\text{) - (T}}_{2} \times {\text{R}}_{2} \times {\text{L}}_{2} )}}{{{\text{T}}_{1} \times {\text{R}}_{1} \times {\text{L}}_{1} }} \times {\text{100}}\%$$
(2)

where m1, T1, R1, L1 and m2, T2, R2, L2 are respectively the mass, tangential, radial, longitudinal dimension of oven-dried wood prior to delignification and after delignification.

Polymer impregnation

Based on the solution composition in Table 1, the uniform solutions were respectively prepared under magnetic stirring action. Samples were respectively immersed into the corresponding solutions and vacuum treated under − 0.1 MPa for 4 h. Subsequently, the excessive chemicals on wood surface were wiped off. Each sample was wrapped with aluminum foil and placed aside for 48 h. Then samples were dried at 105 °C for 3 h. Later the aluminum foil was removed and the temperature was set as 60 °C and 80 °C both for 2 h respectively. Finally the temperature was adjusted to 105 °C to attain the oven-dried wood with constant weight. Untreated nature wood was defined as control (NW).

Table 1 Solution composition for PVA treated wood (WP), furfurylated wood (WF), FA-PVA treated wood (WPF), wood after delignification (WD), PVA treated wood after delignification (WDP), furfurylated wood after delignification (WDF) and PVA-FA treated wood after delignification (WDPF)
Full size table

The weight percent gain (WPG) and volume change (VC) of wood were determined with Eq. (3) to (6).

For NW:

$${\text{WPG}}(\% ) = \frac{{{\text{m}}_{4} - {\text{ m}}_{0} }}{{{\text{m}}_{0} }} \times {\text{100}}\%$$
(3)
$${\text{VC}}(\% ) = \frac{{({\text{T}}_{4} \times {\text{R}}_{4} \times {\text{L}}_{4} ){\text{ }} - {\text{ }}({\text{T}}_{0} \times {\text{R}}_{0} \times {\text{L}}_{0} )}}{{{\text{T}}_{0} \times {\text{R}}_{0} \times {\text{L}}_{0} }} \times 100\%$$
(4)

For WD:

$${\text{WPG}}(\% ) = \frac{{{\text{m}}_{4} - {\text{ m}}_{2} }}{{{\text{m}}_{2} }} \times {\text{100}}\%$$
(5)
$${\text{VC}}(\% ) = \frac{{({\text{T}}_{4} \times {\text{R}}_{4} \times {\text{L}}_{4} ) - ({\text{T}}_{2} \times {\text{R}}_{2} \times {\text{L}}_{2} )}}{{{\text{T}}_{2} \times {\text{R}}_{2} \times {\text{L}}_{2} }} \times 100\%$$
(6)

where m0, T0, R0 and L0 are respectively the mass, tangential, radial and longitudinal dimension of oven-dried NW; m4, T4, R4 and L4 are respectively the mass, tangential, radial and longitudinal dimension of oven-dried wood after treatment.

Scanning electron microscope (SEM) analysis

Wood samples (1 mm (L) × 2 mm (T) × 2 mm (R)) were prepared by using a sliding microtome with disposable blades (REM-710, Yamato Kohki industrial Co., Ltd, Japan). SEM (S-3400, Hitachi, Japan) was used for characterizing the morphology of wood with the acceleration voltage of 3 kV. Three replicates were dried at 80 °C for 2 h and sputtered with Pt–Pd before SEM analysis.

Fourier transform infrared (FTIR) analysis

Three replicate samples (100–200 mesh) were prepared and tested by using a FTIR spectrometer (Bruker vertex 70 V, Germany) based on the KBr pellet method (Sing et al. 1985). The resolution was set as 4 cm−1, and 32 scans were carried out within the spectrum range from 400 to 4000 cm−1.

X-ray diffraction (XRD) analysis

The crystalline structure of three replicate samples (100–200 mesh) was analyzed by using an X-ray diffractometer (XRD, D/max-2500/PC, Japan). Cu–Kα radiation was operated at 40 kV and 40 mA with a scanning speed of 2°/min in the 2θ range of 5–40°. The crystallinity (CrI) was calculated with Eq. (7) based on the empirical method described by Segal et al. (1959).

$${\text{ CrI}} = \frac{{{\text{I}}_{{200}} - {\text{I}}_{{{\text{am}}}} }}{{{\text{I}}_{{200}} }} \times {\text{100}}\%$$
(7)

where I200 is the maximum intensity placed between scattering angle 2θ = 21° and 22°, and Iam is taken by using the minimum intensity placed between scattering angle 2θ = 18° and 19° (Segal et al. 1959; Ahvenainen et al. 2016).

Moisture sorption and dimensional stability test

Static moisture sorption and dimensional stability analysis: the saturated salt solution of NaBr was used to create a condition where relative humidity (RH) was 60% at 25 ℃ (Macromolecule 1958). Samples were subjected to the condition until reaching equilibrium state where mass and dimensions of wood were constant. Five replicates were prepared, and the moisture content and volume change were calculated with Eq. (8) to Eq. (11).

For NW:

$${\text{MC}}_{\text{ns}}(\% ) = \frac{{{\text{m}}_{3} - {\text{ m}}_{0} }}{{{\text{m}}_{0} }} \times {\text{100}}\%$$
(8)
$${\text{VC}}_{{{\text{ns}}}} (\% ) = \frac{{({\text{T}}_{{\text{3}}} \times {\text{R}}_{3} \times {\text{L}}_{3} ) - ({\text{T}}_{0} \times {\text{R}}_{0} \times {\text{L}}_{0} )}}{{{\text{T}}_{0} \times {\text{R}}_{0} \times {\text{L}}_{0} }} \times 100\%$$
(9)

For wood after treatment:

$${\text{MC}}_{\text{ts}}(\% ) = \frac{{{\text{m}}_{3} - {\text{ m}}_{t} }}{{{\text{m}}_{t} }} \times {\text{100}}\%$$
(10)
$${\text{VC}}_{{{\text{ts}}}} (\% ) = \frac{{({\text{T}}_{3} \times {\text{R}}_{3} \times {\text{L}}_{3} ) - ({\text{T}}_{t} \times {\text{R}}_{t} \times {\text{L}}_{t} )}}{{{\text{T}}_{t} \times {\text{R}}_{t} \times {\text{L}}_{t} }} \times 100\%$$
(11)

where MCns and MCts are the moisture content of NW and wood after treatment; VCns and VCts are the volume change of NW and wood after treatment; m3, T3, R3 and L3 are the mass, tangential, radial and longitudinal dimension of wood after moisture sorption;  mt, Tt, Rt and Lt are the oven-dried mass, tangential, radial and longitudinal dimension of treated wood prior to moisture sorption.

Dynamic moisture sorption analysis: samples were firstly preconditioned in about 45%RH by using saturated K2CO3 solution at 25 °C until a constant mass. Then samples were moved to a conditioning oven (DHS 225, YaShiLin Co., Ltd, Beijing) where RH changed sinusoidally between 45% and 75% at 25 °C in programmed steps according to a predetermined schedule. A thermo-recorder (TR-72Ui, T and d Co., Ltd, Japan) was placed near the samples to monitor the real-time RH (sensitivity ± 1%) and temperature (sensitivity ± 1 °C). The mass change of samples was measured by using an electronic analytical balance (ME104E, Mettler-toledo Co., Ltd, America) (sensitivity ± 0.1 mg). The real-time RH, temperature and mass change were automatically recorded by the connected personal computer (Yang et al. 2022). Five replicates were tested, and the moisture content of samples was also calculated based on the Eq. (8) and Eq. (10).

Contact angle (CA) measurement

The contact angles of distilled water droplets on oven-dried sample surface were measured with an optical contact angle meter (SL200KS, KINO, USA). Three replicates were prepared for each group, and two points were selected on the transverse section for contact angle test. Data after droplet deposited on the surface was recorded immediately and continually until 130 s.

Mechanical property analysis

Three-point bending test was processed to test the flexural strength and modulus of elasticity according to ASTM D4761-19. The speed was set as 1 mm/min and six replicates were tested for each group. The impact toughness test was based on the Chinese standard GB/T1843-2008 and GB/T1940-2009 (Yang et al. 2022). Six replicates were prepared and tested.

Results and discussion

Morphology, mass and volume analysis

Figures 1 and 2 show the SEM analysis of microstructures and polymer distribution as well as the analysis of mass and volume changes, respectively. The honeycomb structure of poplar wood was mainly composed of cells such as vessel cells and fiber cells. For WP, the PVA distributed uniformly and formed films on the surface of cell walls. For WF, furfural resin was not only polymerized in cell walls but also in some cell lumina. Similarly, polymers also distributed in both cell walls and cell lumina for WPF. After delignification, many new pores appeared in cell walls, cell corner and middle lamella regions, consisting with previous findings (Liang et al. 2020; Kumar 2021). This indicated that lignin in wood was removed partially after delignification treatment. This result was further confirmed by the mass loss (8.8%) and volume decrease (2.2%) in Fig. 2. Compared to wood polymer composites without delignification, the polymer distribution was more uniform and almost no polymers existed in cell lumina of wood polymer composites after delignification. In addition, the new pores caused by delignification were filled with polymers. Some polymers were closly attached to the surface of cell walls, causing the reduction of cell lumina. As a result, higher mass gain and volume change were found for wood polymer composites after delignification. The negative impacts of PVA on the penetration of FA into cell walls reported previously (Wang et al. 2022) was also alleviated. WDPF had the highest mass gain and volume change, which reached about 95% and 12%, respectively. The results indicated that delignification facilitated more polymer uniformly penetrated into cells and caused higher cell bulking of wood polymer composites.

Fig. 1
figure 1

Microstructures and polymer distribution analyzed by SEM for different samples

Full size image
Fig. 2
figure 2

Mass change and volume change of WP, WF, WPF, WD, WDP, WDF and WDPF

Full size image

FTIR and XRD analysis

FTIR spectra for NW, WP, WF, WPF, WD, WDP, WDF and WDPF are shown in Fig. 3. Compared with the spectrum of NW, the intensity of peak at 1739 cm−1 representing the C = O was higher for WP. This could be ascribed to the acetyl group of PVA or the new C = O generation for the crosslinking between PVA and wood polymers (Mansur et al. 2008; Qiu and Netravali 2012), indicating the introduction of PVA into wood. The peak intensity at 1739 cm−1 also increased for WDP compared with that of WD, further illustrating the introduction of PVA into wood. Besides, the peak at 3436 cm−1 assigned to hydroxyl (–OH) stretching vibration decreased after PVA introduction, confirming the –OH of PVA interacted with functional groups of wood, such as the –OH or –COOH (Qiu and Netravali 2012). When compared the spectrum of NW and WF or WD and WDF, the peak at 1739 cm−1 shifted to the right, which was ascribed to C = O stretching vibration of γ‐diketone of open furan rings (Yang et al. 2022). This indicated the introduction of furfural resin into WF or WDF. Moreover, the new peak at 790 cm−1 due to skeletal vibration of 2, 5-disubstituted furan rings (Oishi et al. 2013) also confirmed the polymerization of furfural resin in wood. When compared WF and WPF or WDF and WDPF, the shift of peak at 1739 cm−1 and peak intensity decrease at 790 cm−1 indicated that PVA affected the self-crosslinking of FA. In addition, the peak at 1508 cm−1 and 1430 cm−1 due to lignin (Popescu et al. 2013) decreased obviously for wood after delignification treatment, indicating the removal of lignin. This result was supported by the former SEM analysis. For WDPF, the peak decrease at 3436 cm−1, the peak increase and shift at 1739 cm−1, the new peak at 790 cm−1 indicated that PVA and FA were both introduced into wood hierarchical structures. In addition, the peak at 1739 cm−1 was higher for WDP than WP, indicating delignification facilitated the impregnation of PVA into wood. The facilitation of delignification was further supported by the increased peak intensity at 790 cm−1 when compared WF, WPF with WDF and WDPF respectively. Consisting with previous studies (Elizabeth et al. 2006; Lin 2011; Kong et al. 2018), the WDPF had lower peaks at 3436 cm−1, 1739 cm−1 and 790 cm−1 than WDF, implying that PVA interpenetrated with FA molecules and interrupted the self-crosslinking of FA.

Fig. 3
figure 3

FTIR analysis for NW, WP, WF, WPF, WD, WDP, WDF and WDPF

Full size image

Diffraction patterns and crystallinity characterized by XRD for different samples are shown in Fig. 4. Similar to NW, three major diffraction planes of cellulose I named 110, 200 and 004 were found when 2θ was at about 16°, 22° and 34° for wood after treatments (Gong et al. 2014; Huang et al. 2019). This indicated that delignification or introduction of PVA and FA did not destroy the crystalline structure of wood. However, the crystallinity varied for different samples. PVA introduction increased the material crystallinity, since PVA could help tighten the cellulose molecules by forming hydrogen bonds with sorption sites and growing polymer films among micro-fibrils (Rao et al. 2019; Deng 2021). Delignification increased the concentration of cellulose and probably facilitated amorphous cellulose regions recrystallized or crystalline zones in adjacent fibrils co-crystallized (Wan et al. 2010). As a result, the crystallinity increased by about 12.7% for WD. Differently, the crystallinity of WF was reduced by about 13.5% due to the existence of amorphous furfural resin. The crystallinity of WDPF was lower than that of NW but still higher than WF and WPF by 8.5% and 3.5% respectively for the synergies of delignification and PVA introduction. Based on above SEM, FTIR and XRD analysis, it can be concluded that the cell environment was highly reconstructed. The schematic illustration of the wood polymer interactions and hierarchical structures of WDPF was shown in Fig. 5.

Fig. 4
figure 4

a XRD analysis and b crystallinity for NW, WP, WF, WPF, WD and WDPF

Full size image
Fig. 5
figure 5

Schematic illustration of wood polymer interactions and hierarchical structures of WDPF

Full size image

Surface wettability analysis

The surface wettability evaluated by the contact angle of distilled water on transverse sections is shown in Fig. 6. The contact angle of NW decreased quickly with time, and the contact angle decreased to about 13° after 23 s. For WP, it was found that PVA slightly retarded the decrease of contact angle with time and increased the contact angle due to the formation of PVA film. FA penetrated into wood and polymerized into hydrophobic furfural resin suggested by SEM and FTIR analysis. Consequently, the WF had a higher contact angle than NW, WP and WPF. The contact angle of WF almost kept constant at about 100° from 50 to 130 s. Delignification improved the arrangement of fibrils and increased overall crystallinity (Fig. 4). Gao et al. (2022) reported that the retained lignin aromatic skeleton after delignification partially contributed to the hydrophobicity increase of wood. As a result, the initial surface wettability decreased and the contact angle of WD was the highest before 30 s. However, lignin was relatively hydrophobic when compared with polysaccharides, and delignification increased the proportion of hygroscopic polysaccharides. Therefore, the contact angle of WD decreased fast relatively after 30 s. Generally, the WDPF had a relatively high and steady contact angle with time. The contact angle of WDPF reached 122° at 0 s and 103° at 120 s for the synergies of polymer film, hydrophobic furfural resin and delignification. This result indicated that WDPF had relatively steady and excellent hydrophobicity.

Fig. 6
figure 6

Contact angle of distilled water on transverse sections against time and contact angle at different time (0 s, 10 s, 30 s 60 s, 120 s) for NW, WP, WF, WPF, WD and WDPF

Full size image

Moisture sorption and dimensional stability analysis

Figure 7 presents the moisture sorption and volume changes at 20℃, 60%RH for different samples. Both PVA and FA introduction reduced the moisture content and volume changes. The PVA or hydrophobic FA resin distributed in both cell walls and cell lumina. Therefore, some water paths were blocked and the water accommodation was reduced, hindering the wood water interactions. Besides, the polymer filling caused cell wall bulking (Fig. 1, 2) reduced the hydroxyl group generation during wood swelling caused by moisture sorption. Due to the combined effects of PVA and FA, the WPF had the lowest volume change, which was lower than NW by over 71%. Delignification slightly increased the moisture sorption and volume changes for the removal of relatively hydrophobic lignin. However, delignification also facilitated more penetration of hydrophobic FA and PVA into wood (Fig. 5), causing higher cell wall bulking (Fig. 1, 2). Therefore, the moisture content and volume changes of WDPF were much lower than NW, and the decrease reached over 46% and 49% respectively.

Fig. 7
figure 7

Moisture sorption and volume changes at 20℃, 60%RH for different samples

Full size image

The dynamic moisture sorption against cyclic time for different samples is shown in Fig. 8. Moisture content against cyclic time for both NW and wood after treatments showed a similar sinusoidal changing trend. A time phase lag was visible between moisture content and RH change. Similar to the static moisture sorption, PVA and FA also reduced the dynamic moisture content. The WF had a lower dynamic moisture sorption than WP. Compared with NW, the average moisture content of WF and WDF in Fig. 9 decreased by over 43% and 36% respectively, though delignification increased the hydrophilicity and moisture sorption. The average dynamic moisture content of WDPF was higher than WPF but still lower than NW by over 48%. This could be explained by the following combined effects. The PVA and FA distributed in hierarchical structures of wood (Fig. 1, 3, 5), blocking water paths and reducing water accommodation. The hydrophobic FA resin and PVA polymer also formed dense films on surface of cell walls (Fig. 1), hindering the diffusion of water molecules. Besides, the cell wall bulking due to polymer filling (Fig. 1, 2) also reduced the hydroxyl group generation during swelling of wood under dynamic relative humidity changes. From the analysis of static and dynamic moisture sorption, it was concluded that WDPF had improved hydrophobicity and dimensional stability.

Fig. 8
figure 8

Dynamic moisture sorption for NW, WP, WF, WPF, WD, WDP, WDF and WDPF

Full size image
Fig. 9
figure 9

Average moisture content under dynamic condition for different samples

Full size image

Mechanical property analysis

Flexural strength and modulus of elasticity are both important load-bearing parameters for a construction material. Fig. 10 shows the flexural strength and modulus of elasticity for different samples. As expected, the flexural strength and modulus of elasticity of WPF slightly increased. This was because FA resin and PVA formed mechanical interlocking and restricted the motion of macromolecular chains, reducing the free volume of molecular sliding (Shen et al. 2021). Moreover, the PVA interacted with hydroxyl groups of wood and helped restrict the macromolecular motion (Qiu and Netravali 2012; Rao et al. 2019). Delignification increased the cellulose proportion and promoted the orderly arrangement of fibrils, leading to increased crystallinity (Fig. 4). Therefore, the flexural strength and modulus of elasticity of WD increased (Zhang 2011). In addition, delignification facilitated more penetration of PVA and FA into cell walls (Fig. 1, 5). The PVA and FA could infiltrate each other and form interpenetrating polymer networks (IPNs) (Sreehari et al. 2022; Amir et al. 2022). Consequently, the WDPF had the highest flexural strength and modulus of elasticity, and the increase reached 21.7 and 46.3% respectively.

Fig. 10
figure 10

Flexural strength (a) and modulus of elasticity (b) for different samples

Full size image

The fracture morphology and impact toughness for different samples are displayed in Fig. 11. Compared with NW, WF had no many fibers stretched out and its surface was relatively smooth. Consistent with previous results (Lande et al. 2004; Epmeier et al. 2004), the impact toughness of WF decreased, and the reduction reached over 46%. This was attributed to the dense cross-linking network structures of furfural resin (Lande 2008) and the inhomogeneous distribution of furfural resin in cells (Fig. 1), which tended to lead stress concentration under impact loading. Differently, the WP displayed scraggly fracture surface with some fibers stretched out, and the impact toughness increased by about 18%. This indicated the toughening effects of PVA on wood, which was confirmed by the result that WPF had a much higher impact toughness than WF. After delignification, many fibers stretched out and the impact toughness of WD increased by over 14%. Compared with WP, WF and WPF, the WDP, WDF and WDPF also had higher impact toughness respectively. This further indicated that delignification contributed to the improvement of impact toughness. Due to the synergistic effects of delignification and PVA impregnation, the impact toughness of WDPF was higher than WF by over 90%. The main mechanisms could be illustrated as follow. Delignification helped the microfibrils arrange more tightly and orderly and increased the crystallinity (Fig. 4). Therefore, the stress transfer between the cellulose microfibrils was promoted and the impact toughness was improved. In addition, PVA had flexible linear molecular structure, which benefited redistributing stress among molecules and alleviating the stress concentration as well as mitigating the development of destructive stress (Zheng et al. 2015; Li et al. 2016; Wu et al. 2012; Rao et al. 2019). The PVA could form interpenetrating polymer networks with furfural resin in wood hierarchical structures and reduce the self-crosslinking of furfural resin (Sreehari et al. 2022; Amir et al. 2022). Moreover, delignification facilitated more penetration of PVA into wood and helped PVA better exert its positive effects.

Fig. 11
figure 11

Fracture morphology and impact toughness of NW, WP, WF, WPF, WD, WDP, WDF and WDPF

Full size image

Conclusions

The synergies of introduction of PVA and FA into wood combined with delignification were proposed for processing poplar wood into WDPF with improved physical and mechanical properties. Compared to NW, the moisture and volume change of WDPF decreased by over 46% and contact angle kept over 100° after 120 s. The cell environment changes accounted for the improvement of hydrophobicity and dimensional stability. Delignification enhanced wood porosity and benefited modifier impregnation and uniform distribution in cell walls, causing higher cell wall bulking. Hydrophobic FA resin and PVA blocked water paths and formed films on surface of cell walls, reducing water accommodation and hindering wood water interactions. The flexural strength and modulus of elasticity were respectively improved by over 21% and 46%. The impact toughness of WDPF was improved by over 90% compared to WF. The mechanical strength improvement was mainly attributed to the combined effects of PVA-FA interaction and delignification. PVA with flexible macromolecular chains interacted with FA and wood, interrupting the crosslinking of FA resin. Delignification and PVA facilitated orderly arrangement of micro-fibrils and increased crystallinity, helping the improvement of mechanical strength. The fabrication and mechanism analysis of WDPF provided references for designing and upgrading low quality wood products.