Environment-friendly wood fibre composite with high bonding strength and water resistance

With the growing depletion of wood-based materials and concerns over emissions of formaldehyde from traditional wood fibre composites, there is a desire for environment-friendly binders. Herein, we report a green wood fibre composite with specific bonding strength and water resistance that is superior to a commercial system by using wood fibres and chitosan-based adhesives. When the mass ratio of solid content in the adhesive and absolute dry wood fibres was 3%, the bonding strength and water resistance of the wood fibre composite reached the optimal level, which was significantly improved over that of wood fibre composites without adhesive and completely met the requirements of the Chinese national standard GB/T 11718-2009. Fourier transform infrared (FTIR), X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD) characterizations revealed that the excellent performance of the binder might partly be due to the amide linkages and hydrogen bonding between wood fibres and the chitosan-based adhesive. We believe that this strategy could open new insights into the design of environment-friendly wood fibre composites with high bonding strength and water resistance for multifunctional applications.


Introduction
With a rapidly developing economy, the demand for woodbased materials has been increasing in China, but China is a country that does not have large forestry resources. The juxtaposition of the increasing need for wood-based materials and the short supply has caused wide concerns for the people of China. However, there are large amounts of residual forest products being produced in China [1], most of which were dispersed or combusted. Using forest residue as a raw material

Preparation of chitosan-based adhesives and wood fibre composites
In this paper, the mass ratios of solid content in adhesive and absolute dry wood fibres were 1%, 2%, 3%, 4% and 5%. Approaches for preparing wood fibre composites were the same, while the amounts of chemical reagents used were variable (table 1). When the mass ratio of solid content in adhesive and absolute dry wood fibres was 1%, the chitosan-based adhesives and wood fibre composites were synthetized as follows: chitosan powder (1 g) was added into a beaker which contained distilled water X (34.48 g) and stirred until the chitosan powder was dispersed uniformly in distilled water. Meanwhile, acetic acid (0.67 g) was poured into a beaker containing distilled water Y (34.48 g) and stirred until a homogeneous acetic acid solution was formed. Afterwards, the acetic acid solution was poured into the beaker containing chitosan and distilled water under rapid agitation with a magnetic stirrer until a homogeneous chitosan solution was formed. Simultaneously, glutaraldehyde solution (0.5 g, 50 wt%) was diluted with distilled water Z (34.48 g). The chitosan solution served as the main agent, while the diluted glutaraldehyde solution served as the cross-linking agent. Before mixing the wood fibres and the chitosan-based adhesive, the wood fibres were dried in an oven at 80°C to reduce the moisture content from 18% to approximately 6%. Then the wood fibres were mixed in a high-speed mixer (Zhangjiagang Tongsha Plastic Machinery Co. Ltd) at 750 r.p.m. At the same time, the chitosan solution and glutaraldehyde solution were mixed up and stirred for a while. Then the adhesive was poured into the wood fibres immediately under strong agitation. Thereafter, the mixture containing wood fibres and adhesive was filled into a forming box (250 × 250 mm) and pre-pressed under 1 MPa of pressure to form a mat. Then the mat was placed between two parallel flat plates (400 × 400 mm) of a mechanically controlled oil-heated press to implement the final hot (170°C) pressing according to a pre-programmed hot-pressing schedule (figure 1). A 5 mm steel gauge was used to control the thickness of wood fibre composite. Finally, the edges of wood fibre composite were cut off by 30 mm using a table saw, resulting in a 220 × 220 × 5 mm board with a target density of 0.8 ± 0.02 g cm −3 . The wood fibre composite was conditioned at a relative humidity of 40% and room temperature for 2 days before bonding strength and water resistance tests. Wood fibre composites with mass ratios of solid content in adhesive and absolute dry wood fibres of 1%, 2%, 3%, 4% and 5% were designated as WFC 1, WFC 2, WFC 3, WFC 4 and WFC 5, respectively. Wood fibre composites without adhesive and the control group were designated as WFC 0 and CG, respectively.

Bonding strength and water resistance tests
The bonding strength and water resistance of wood fibre composite were evaluated through mechanical and dimensional properties (modulus of rupture, MOR; modulus of elasticity, MOE; internal bond strength, IB; thickness swell, TS and water adsorption, WA), which were measured according to the Chinese national standard GB/T 17657-2013. The MOR represents the rate of the bending moment and flexing modulus under the highest load, while the MOE represents the rate of the stress and strain of the load within the range of the elastic limit. The IB represents the rate of the highest damage tension perpendicular to the surface of the specimen and the specimen surface area. The TS and WA refer to the increase in the thickness and weight of a specimen soaked in water for 24 h at room temperature. The MOR and the MOE of the composites were determined by executing three-point static bending tests on specimens with a size of 200 × 50 mm at a cross-head speed of 5 mm min −1 . The IB was determined by pulling the specimen (50 × 50 mm) apart in a perpendicular direction. The specimens were sanded to take away the low-density surface area. Thereafter, metal fixtures were adhered on each side of the specimen using hot melting adhesive. A tensile load with a cross-head speed of 0.5 mm min −1 was applied on each metal fixture until failing happened in the specimen. The TS and WA were determined by measuring the thickness and weight of specimens with a size of 50 × 50 mm before and immediately after the 24 h immersing process. The MOR and MOE measurements were conducted on 12 repetitions, while IB, TS and WA measurements were conducted on eight repetitions.

Characterization of wood fibre composite
Scanning electron microscopy (SEM) images of the fractured wood fibre composites after IB tests were taken with a microscope (Quanta200, FEI, USA) operated at 12.5 kV. To make the samples conductive, sample surfaces were coated with a thin layer of gold using an ion sputter coater prior to the SEM analysis. FTIR spectra of wood fibres and wood fibre composites were collected using Nicolet Magna-IR560 E.S.P. with the ATR technique. The spectra were recorded at a resolution of 4 cm −1 by accumulating 32 scans. Surface chemistry of wood fibres and wood fibre composite was characterized by X-ray photoelectron spectroscopy (XPS) on a VG Thermo probe ESCA system (Thermo Electron, USA). The chemical shifts relative to C1 (284.3 eV) for C2, C3 and C4 were 1.5 ± 0.1 eV, 3 ± 0.1 eV and 4 ± 0.1 eV, respectively. X-ray diffraction (XRD) patterns were obtained using an X-ray diffractometer (D/max 2200, Rigaku, Japan) and the relative crystallinity (CI) was calculated according to Segal method [20] as follows: where I 002 represents the maximum intensity of both crystalline and amorphous region of the material and I am represents the intensity of the amorphous region of the material.

Bonding strength and water resistance
The influence of the mass ratio of solid content in adhesive and absolute dry wood fibres on the performance of the wood fibre composites is shown in figure 2. It is worth noting that the trends for MOR, MOE and IB can be divided into two stages. During the first stage, there is a correlation to the mass ratio in the range from 0% to 3% and the MOR, MOE and IB increased from 9.16 MPa, 900.25 MPa and 0.14 MPa to 39.78 MPa, 3960.31 MPa and 1.86 MPa, respectively. The increase in bonding strength was remarkable and indicates that there is an optimal mass ratio of solid content in the adhesive and absolute dry wood fibres for the mechanical performance of the wood fibre composites. During the second stage corresponding to the mass ratio in the range from 3% to 5%, the mechanical performance of wood fibre composites deteriorated with increasing mass ratio. The change in bonding strength was reflected in a decline in MOR, MOE and IB from 39.78 MPa, 3960.31 MPa and 1.86 MPa to 32 MPa, 2679.72 MPa and 0.78 MPa, respectively. These results show that the mechanical performance of the wood fibre composite cannot increase with increasing mass ratio and will reach a limit at the mass ratio of 3%. As mass ratio increased from 3% to 5%, the adhesive solution concentration also increased. Thus, the chitosan content and glutaraldehyde content in the adhesive solution increased, which led to an adhesive solution with increasing viscosity and non-uniform distribution of wood fibres and adhesive [21]. The high chitosan content and glutaraldehyde content in the adhesive solution can also cause   premature curing of the adhesive [22]. The combination of non-uniform distribution of wood fibres and adhesive and premature curing of the adhesive were responsible for the deterioration of mechanical performance of wood fibre composites when the mass ratio was more than 3%.
The change trends of water resistance of wood fibre composites were similar to those of mechanical performance and can also be divided into two stages. The water resistance increased during the first stage correlating to mass ratio in the range from 0 to 3% and then declined in the second stage corresponding to the mass ratio in the range from 3% to 5% as reflected in the first increase and then decrease in TS and WA. The optimal TS and WA were 17.63% and 15.28%, respectively, both of which decreased significantly compared with that of WFC 0. The improvement of water resistance was partly attributed to the strong adhesion between wood fibres which were adhered closely to each other. The other reason was that chitosan shows good water resistance as it can prevent water from entering its molecular chains [16]. While the subsequent decline of water resistance was also ascribed to the non-uniform distribution of wood fibres and adhesive and premature curing of the adhesive, both of which caused the decrease in adhesion among wood fibres.
In summary, the MOR, MOE, IB, TS and WA all reached an optimal level when the mass ratio of solid content in adhesive and absolute dry wood fibres was 3%. The optimal mechanical performance and water resistance both improved significantly compared to WFC 0 and met the requirements for the Chinese national standard. WFC 3 bears favourable comparison with CG in terms of performance and  even shows much better IB than CG. These results demonstrate that an environment-friendly wood fibre composite with high bonding strength and water resistance could be fabricated through the appropriate addition of this chitosan-based adhesive. Figure 3 shows the FTIR spectra of wood fibres, WFC 0 and WFC 3. Compared to wood fibres, some changes were observed in the FTIR spectrum of WFC 0. The peaks at 1642 cm −1 and 1543 cm −1 corresponding to ester linkages decreased after hot pressing [22]. The ester linkages are components of hemicelluloses and extractives. Therefore, this decline shows that extractives and hemicelluloses were chemically changed during the hot-pressing process. The peaks at 1507 cm −1 , 1242 cm −1 and 780 cm −1 , which have been reported to be associated with C=C symmetric stretching vibration, C-O stretching vibration and CH out-of-plane vibrations corresponding to lignin, respectively, weakened after the hotpressing process [23,24]. These results indicate that the lignin was degraded. The peak at 1596 cm −1 , which belongs to superposition of C=C linkages of lignin and C=O linkages, strengthened after hot pressing [23,24], and this change implies the generation of some low molecular weight organics including furan, furfural and methyl-furfural due to the degradation of hemicelluloses, lignin and extractives. The formation of these low molecular weight organics resulted in the improvement of C=O linkages and was effective in the self-bonding of WFC 0 [25,26]. The peak at 1740 cm −1 in the FTIR spectrum of wood fibres corresponds to the association of COOH of hemicelluloses, while the FTIR spectrum of WFC 0 shows a peak at 1720 cm −1 corresponding to C=O linkages. The peak at 1720 cm −1 in the FTIR spectrum of WFC 0 was weaker than the peak at 1740 cm −1 in the spectrum of wood fibres [23]. The decline of this peak indicates degradation of hemicellulose and the shift can be attributed to the generation of hydrogen bonds associated with COOH groups. The peak corresponding to OH groups in the FTIR spectrum of wood fibres was at 3321 cm −1 , but it moved to 3332 cm −1 in the spectrum of WFC 0. The shift is due to the appearance of hydrogen bonds related to OH groups. Therefore, some hydrogen bonds between COOH groups and OH groups of wood fibres occurred after hot pressing. In summary, the newly formed low molecular weight organics and hydrogen bonds resulted in self-bonding of wood fibres and thus were responsible for the performance of WFC 0 [27].

Fourier transform infrared analysis
The mechanical properties and water resistance of WFC 0 were quite poor, while those of WFC 3 were excellent. This difference implies that the performance of WFC 3 relied on the self-bonding of wood fibres and adhesion between wood fibres caused by the chitosan-based adhesive. Therefore, elucidation of reactions between wood fibres and the chitosan-based adhesive can help to elucidate the bonding mechanism.
In comparison with WFC 0, the peak related to amide linkages at 1655 cm −1 strengthened in the FTIR spectrum of WFC 3 and this change implies a reaction between the amine groups of chitosan-based adhesive and carbonyl groups from wood fibres, which resulted in the formation of amide linkages [28]. The peak at 1596 cm −1 also strengthened in the FTIR spectrum of WFC 3. This peak was the superposition of the C=C linkages of lignin and the C=O amide linkages. This change is also a proof of amide linkages.    The peaks associated with OH groups at 3332 cm −1 in the FTIR spectrum of WFC 0 further shifted to higher wavenumbers at 3341 cm −1 in the FTIR spectrum of WFC 3 and this change indicates more hydrogen bonds between OH groups of wood fibres occurred [29]. These hydrogen bonds may exist between OH groups of wood fibres and OH and amine groups of the chitosan-based adhesive. The peak at 1720 cm −1 corresponding to C=O in the FTIR spectrum of WFC 3 was stronger than that of WFC 0, which was attributed to glutaraldehyde existing in the chitosan-based adhesives. The above FTIR analysis shows the formation of amide linkages and hydrogen bonds between wood fibres and the chitosan-based adhesives, which was essential for the excellent performance of WFC 3.

X-ray photoelectron spectroscopy analysis
The XPS analysis was used to identify the surface chemistry of the wood fibres, WFC 0 and WFC 3. It was observed that the O/C ratio of wood fibres was 0.28, while the O/C ratio of WFC 0 increased to 0.33 after the hot-pressing process (table 2). Combined with the results of FTIR analysis, it was inferred that the improvement of O/C ratio after hot pressing was attributed to the newly formed low molecular weight organics which are rich in oxygen-containing groups. The O/C ratio of WFC 3 further rose to 0.34 due to the addition of the chitosan-based adhesive, which contains an abundance of oxygen-containing groups [1]. The emerging N in WFC 3 was also due to the addition of the chitosan-based adhesive. Figure 4 shows the C 1 s spectra of wood fibres, WFC 0 and WFC 3 with high-resolution scans, which were deconvoluted into four components for all specimens: C1 (C-C/C-H/C=C) at 284.3 ± 0.1 eV,    [15,30]. Compared to wood fibres, C1 of WFC 0 decreased, while C2 and C3 increased. C1 only exists in the lignin and extractives, and thus the decrease of C1 was partly due to the degradation of lignin and extractives during pressing. The other reason was the relative increase of C2 and C3 due to the generation of oxygen-containing low molecular weight organics, such as furan, furfural and methyl-furfural. C4 of WFC 0 declined slightly compared to that of wood fibres and C4 only exists in hemicelluloses. This phenomenon is indicative of hemicellulose degradation. These results are consistent with the FTIR analysis.
By comparison with WFC 0, the C1 of WFC 3 decreased and C2 and C3 increased due to the addition of the chitosan-based adhesive. C4 of WFC 0 exhibited a downward trend due to the degradation of hemicelluloses, while C4 of WFC 3 went up. There were no C4 components in the chitosan-based adhesive itself (electronic supplementary material, figure S1). Therefore, the increase in C4 could be attributed to the newly generated amide linkages between wood fibres and the chitosan-based adhesives, which was beneficial for the composite properties of WFC 3 and aligns with FTIR analysis. Figure 5 shows the XRD diffraction patterns and the relative crystallinity of the wood fibres, WFC 0 and WFC 3. The crystal structure of cellulose in the wood fibres, WFC 0 and WFC 3 was similar and corresponded to the typical cellulose I pattern because both the diffraction peaks at 2θ = 16.5°and 22.6°could be observed in all three diffraction profiles [31]. Therefore, the hot pressing and addition of chitosan-based adhesive did not change the crystal structure. However, the relative crystallinity of these samples differed from each other. Crystallinity is a major factor that influences the mechanical and dimensional properties of wood-based materials. The influence of hot pressing, adhesives and other components of wood fibres on the crystallization behaviour and supermolecular structure of cellulose could be reflected through calculating the crystallinity. Compared with wood fibres, WFC 0 showed slightly increased crystallinity. The wood fibres were compacted closely and densely during the hotpressing process, and hence the possibility of generation of hydrogen bonds between OH groups of cellulose increased, which caused the improvement of relative crystallinity [32]. The self-bonding of WFC 0 partly relied on the hydrogen bonds due to pressure-driven fibril aggregation [33]. The degradation of hemicelluloses, lignin and extractives resulted in a relative increase in cellulose content, which could contribute to the relative increase in crystallinity.

X-ray diffraction analysis
Different from the increase in the relative crystallinity of WFC 0 compared to wood fibres, the relative crystallinity of WFC 3 declined slightly compared to WFC 0, which is partly attributed to the reduction of the possibility of hydrogen bonds between hydroxyls of cellulose caused by the amide linkages and hydrogen bonds between amino and hydroxyl groups of chitosan and hydroxyl groups of cellulose. The decrease in crystallinity might also be associated with the disruption of the surface of the crystalline regions by the cured chitosan-based adhesives [34], which agrees well with the earlier findings by other researchers [35]. In addition, the chitosan-based adhesive is amorphous [2]. Hence, the amorphous structure of the chitosan-based adhesive incorporated into cellulose also led to a reduction in crystallinity [36].
The XRD analysis agrees with the FTIR and XPS analyses and a combination of these results shows that the bonding mechanism of WFC 3 might be dependent on the amide linkages and hydrogen bonds between wood fibres and the chitosan-based adhesive. Figure 6 shows the fractured surfaces of WFC 0 and WFC 3 after IB tests. The original circular cell cavities of wood fibres were compacted during the hot-pressing process. But there were no signs of deformed cell walls, which indicates that wood fibre structures are maintained. The wood fibres in WFC 0 were randomly distributed and interlaced with one another and made it easy to form interlocks among individual wood fibres. The wood fibres in WFC 0 originally contacted closely to each other due to the hot-pressing process. However, as shown in figure 6, wood fibres were loosely contacted and some voids existed in the fracture surface of WFC 0. A pulling force divided WFC 0 into two parts and some wood fibres were pulled out directly from WFC 0 during IB testing. Some deformations of wood fibres due to the hot pressing were recovered after IB tests. These phenomena indicate that the adhesion between wood fibres was poor and did not prevent deformation recovery and wood fibres from pulling out.

Scanning electron microscopy image analysis
The wood fibre surfaces of WFC 3 were rougher than those of WFC 0 due to the uniform distribution of chitosan-based adhesive among the wood fibre surfaces. The wood fibres in WFC 3 were more compacted than those in WFC 0 and exhibited fewer voids. Moreover, there was even a torn fibre in WFC 3, which was pulled apart during the IB test. The torn fibre was snapped rather than pulled out completely, indicating strong adhesion among the wood fibres, which prevented wood fibres from pulling out and deformation recovery. Therefore, it can be concluded that the adhesion among wood fibres in WFC 3 was strong and ensured excellent mechanical performance and water resistance of WFC 3.
In summary, from the results of FTIR spectroscopy, XPS analysis, XRD patterns, SEM images and the previous works of ours and other researchers, the partial potential bonding mechanism of these wood fibre composites could be conjectured roughly ( figure 7). Firstly, wood fibres became closely and tightly interlocked under high pressure. Then hemicellulose, lignin and extractives in the wood fibres degraded and low molecular weight organics were generated during hot pressing. The further use of high pressure during hot pressing allowed for the cellulose, hemicelluloses, lignin and extractives to be pressed close to each other. The combination of high pressure and temperature allowed for the generation of hydrogen bonds among the different polymers. Meanwhile, amide linkages were formed between carbonyl groups of wood fibres and the amine groups of the chitosan-based adhesive. Hydrogen bonds between wood fibres and chitosan-based fibres were also generated. A combination of these interlocks, amide linkages, hydrogen bonds and low molecular weight organics contributed to the excellent performances of wood fibre composite. The whole bonding mechanism is very complicated and it is extremely hard to fully clarify all the bonding mechanisms in this paper. The bonding mechanism given above is only a part of the whole bonding mechanism and is of guiding significance and lays foundation for fabrication of the wood fibre composite and further research. There is still a lot of work to do with the bonding mechanism.

Conclusion
In conclusion, allowing for the shortage of wood-based materials and the negative effects of formaldehyde emissions in traditional wood fibre composites and inspired by chitosan, a simple approach was used that could fabricate environment-friendly wood fibre composites with high bonding strength and water resistance by using a chitosan-based adhesive and residual forest wood fibres. The obtained wood fibre composite with a mass ratio of solid content in adhesive and absolute dry wood fibres of 3% exhibited excellent MOR, MOE, IB, TS and WA up to 39.78 MPa, 3960.31 MPa, 1.86 MPa, 17.63% and 15.28%, respectively. These results were superior to the wood fibre composite without an adhesive and a commercially available product. Furthermore, the potential bonding mechanism analysis shows that the excellent performance of the chitosan-based adhesive was partly reliant on a combination of wood fibres interlocks, newly formed amide linkages, hydrogen bonds and low molecular weight organics. Because chitosan also shows outstanding antibacterial properties, the proposed method provides a simple strategy for designing and developing multifunctional environment-friendly wood fibre composite with excellent performance. Further research will continue to systematically investigate the bonding mechanisms and versatility of this wood fibre composite.
Data accessibility. Part of the datasets are provided in the main text and the others are uploaded as part of the electronic supplementary material.