Introduction

The deformation processing of wood can effectively promote its use in various applications, such as furniture, building, construction, automotive, and other daily necessities. Conventional methods of wood deformation include compression and bending processing using cell-wall deformations (Sandberg et al. 2012). These methods produce simple two-dimensional products, while retaining the cell arrangement and structure of the wood. On the other hand, reducing the element size of wood can efficaciously enable the formation of more complex products. Wood-plastic composites (WPCs), which are a mixture of wood powder and plastic, are widely used as building materials (Spear et al. 2015). However, they require an input of energy to miniaturize the wood, which results in the destruction of the original cell structure, such that the products do not retain the wood texture. Furthermore, in conventional processing, the element size of wood and the complexity of the product shape have a contradictory relationship. To date, there have been few successful reports on the formation of a complex shape from solid-state wood.

Recently, a new technique, wood flow forming (WFF), has been developed, which can form complex shapes with solid-state wood because it takes advantage of plastic flow deformation (Abe et al. 2020, 2021; Miki et al. 2014a, 2014b, 2017; Seki et al. 2016; Yamashita et al. 2009). The wood is compressed in a heated mold, which enables it to flow plastically owing to the shear sliding between wood cells to produce a final shaped product (Miki et al. 2017). The application of traditional plastic forming techniques for metal and plastic materials yields efficiently wood products with simple processes. Although these techniques can critically affect the original cell arrangement of wood with high compressive stress, the original wood cell wall structure can be maintained to obtain a unique cell-derived texture of the product (Miki et al. 2014a). However, to improve the deformability of wood during the forming process and the durability of the products, it is necessary to modify the wood by pretreatment before forming. The cell wall and compound middle lamella (CML) are modified by impregnating the wood with resin monomers (Miki et al. 2014b; Seki et al. 2016) and/or chemical modification (Abe et al. 2020, 2021). WFF has great potential for various applications; however, it requires high temperatures (> 100 °C) and pressures (> 50 MPa), making it energy-intensive and less productive. To solve these problems, this study focused on lignin, which acts as an adhesive and binds the polysaccharides (cellulose and hemicellulose) and cells together. The lignin content is higher in CML than in cell walls (Kumar et al. 2021). Moreover, the wood cells are dissociated at the interface between the cells (i.e., CML) during the WFF process (Yamashita et al. 2009). Therefore, lignin, especially in CML, has a significant plastic flow effect.

Delignification is a process that can remove lignin without breaking the cellular structure of the wood, and it has primarily been studied for pulping wood. However, in recent years, delignified wood (DW) (Kumar et al. 2021) has attracted attention as a functional material for use in transparent wood (Li et al. 2016, 2017, 2020), high-strength structural materials (Frey et al. 2019b; Jakob et al. 2020; Song et al. 2018), high-performance thermal insulators (Li et al. 2018), and thermal energy storage materials (Montanari et al. 2019). Deformation processing that takes advantage of the flexibility of DW has also been developed (Khakalo et al. 2020; Frey et al. 2018, 2019a). Frey et al. (2019a) reported that a completely delignified veneer in a water-swollen state exhibited significant deformability in the fiber direction. As the improvement on deformability of DW has been reported, the plastic flow deformation associated with WFF is potentially affected by delignification.

The thermal softening properties of water-swollen wood depend on the glass transition of lignin (Kojiro et al. 2008; Nakajima et al. 2009). When the molecular mass of lignin in wood decreases, its glass transition temperature also decreases, in turn significantly softening the wood (Nakajima et al. 2009). Therefore, it is expected that the changes in lignin due to delignification will promote the plastic flow deformation of wood and reduce the production energy required for WFF.

The objective of this study was to elucidate the effect of delignification on the plastic flow deformation of wood. DW and untreated-wood (UW) samples with different molecular masses and amounts of lignin were prepared by subjecting them to delignification and varying the delignification time. Free compression testing was used to evaluate the deformability of the samples based on the stress at the flow starting point and the deformed cross-sectional area of the wood sample. The samples were characterized by attenuated total reflection-infrared (ATR-IR) spectroscopy, nuclear magnetic resonance (NMR) spectroscopy, Raman imaging analysis, and dynamic viscoelastic measurements. The effects of the amount, chemical structure, distribution, and molecular motilities of lignin on the plastic flow deformation of wood are discussed as well.

Materials and methods

Materials

Wood samples were successively cut in the longitudinal (L) direction from a block of sapwood of a Japanese cypress (Chamaecyparis obtusa) log collected from the Kiso region of Japan. The dimensions of the samples used for the dynamic viscoelastic measurements were 1 mm (L) × 30 mm (radial (R) direction) × 3 mm (tangential (T) direction). The samples for the other measurements (ATR-IR, NMR, Raman imaging analysis, and free compression testing) had dimensions of 5 mm (L) × 5 mm (R) × 5 mm (T). Prior to delignification, the wood samples were pre-treated with approximately 100 °C distilled water for 4 h, followed by methanol for 6 h to remove the low-molecular-weight components. The pre-treated samples were then dried at 50 °C for 18 h and 105 °C for 2 h to a relatively constant mass (m0). One side of the RT surface of some samples (5 mm × 5 mm × 5 mm) was microtome-finished prior to subsequent delignification.

Preparation of delignified wood samples

The pre-treated samples were delignified using an acetic acid solution containing 4 wt% sodium chlorite (NaClO2) (pH 3). The NaClO2 solution was impregnated into the pre-treated samples under vacuum, and the impregnated samples were treated at 45 °C for several reaction times (tr) of 10, 30, 60, 180, or 360 min. Then, the DW samples were washed several times with distilled water and stored in water at 20–25 °C Note that the DW sample with a tr longer than 360 min was brittle and fragile; therefore, it was considered unsuitable for WFF and was not evaluated in this study. In addition, the UW samples were impregnated with distilled water, instead of the NaClO2 solution, at 20–25 °C for 500 min or more (tr: 0 min).

The area in the RT cross section of the water-swollen DW and UW samples (sd and su, respectively) was measured, and the area increase rate (ART) of the water-swollen samples due to delignification was calculated using the following formula:

$$A_{{{\text{RT}}}} = \left( {s_{{\text{d}}} - s_{{\text{u}}} } \right)/s_{{\text{u}}} \times 100\, \left[ \% \right]$$
(1)

Some of the water-swollen samples were then dried at 35 °C for 24 h, at 50 °C for 18 h, and at 105 °C for 3 h to a relatively constant mass (mt). The mass loss (ML) due to delignification of the dried samples was calculated using the following formula:

$$ML = \left( {m_{{\text{o}}} - m_{{\text{t}}} } \right)/m_{{\text{o}}} \times 100\,\left[ \% \right]$$
(2)

Acetyl bromide method for quantification of lignin

The acetyl bromide method for quantifying lignin derived from wood is advantageous, because it is rapid, simple, and appropriate for small sample sizes (Hatfield and Fukushima 2005). The milled dried samples (5 mg) were placed in a glass reaction vial with 5 mL of 25% (v/v) acetyl bromide in acetic acid and 0.2 mL of perchloric acid (60%); the vial was subsequently sealed and heated at 70 °C for 30 min. After cooling, the contents of the vial were quantitatively transferred to a 100-mL volumetric flask containing 10 mL of 2 M NaOH and 25 mL of acetic acid, and the volume was made up to 100 mL with acetic acid. The UV absorption spectrum was measured with a UV–vis spectrophotometer (UV-3600; Shimadzu Corporation, Kyoto, Japan) at 280 nm. Ultimately, the lignin content (LC) was calculated using the following equation:

$$LC = A_{{\text{s}}} \times V/\left( {k \times W} \right) \times 100\left[ \% \right]$$
(3)

where As represents the absorbance of sample, k represents an absorption coefficient of standard lignin (= 20.09), V represents the volume of solution, and W denotes the weight of the sample.

Attenuated total reflection infrared (ATR-IR) spectroscopy.

The block-shaped sample (5 mm × 5 mm × 5 mm) in the dry state was cut in half in the L-direction, and ATR-IR measurements were performed near the center of the cut surface (RT surface, that is, near the center of the wood sample). The ATR-IR spectra were measured on a Nicolet 6700 spectrometer (Thermo Scientific Inc., Waltham, MA, USA) at a resolution of 4 cm−1 in the standard ATR mode; 32 scans were performed in the wavenumber range of 4000–700 cm−1.

Nuclear magnetic resonance (NMR) spectroscopy

Solid-state 13C NMR spectra were measured on a Varian 400 NMR system spectrometer (Palo Alto, CA) with a Varian 4 mm double-resonance T3 solid probe. A thin sample (approximately 0.1 mm) was cut out from near the surface, shredded, and then used for NMR measurement. The shredded samples were placed in a 4 mm ZrO2 rotor spun at 15 kHz within a temperature range of 20–22 °C. The 13C cross-polarization and magic-angle spinning (CP-MAS) NMR spectra were collected with a 2.6-μs π/2 pulse at 100.56 MHz for the 13C nuclei and an acquisition period of 40 ms over a spectral width of 30.7 kHz. Proton decoupling was performed with an 86 kHz 1H decoupling radio frequency with a small phase incremental alteration (SPINAL) decoupling pulse sequence. The 1H-13C cross-polarization for the spectrum acquisition was conducted with a 5.0 s recycle delay in 1024 transients using a ramped-amplitude pulse sequence with a contact time of 2 ms and a 2.6-μs π/2 pulse for the 1H nuclei. The amplitude of the 1H nuclei was linearly ramped down from 92.6% of its final value during the CP contact time. The 1H spin–lattice relaxation time in the laboratory frame (T1H) was indirectly measured by detecting the 13C resonance enhanced by the cross-polarization in the 13C CP-MAS sequence, and was applied after a π pulse to 1H nuclei using the inversion recovery method. The T1H analysis of the sample was carried out using the same solid-state probe used to obtain the 13C CP-MAS NMR spectra of the sample at the same contact time and acquisition period.

Raman imaging analysis

The block-shaped sample (5 mm × 5 mm × 5 mm) in the water-swollen state were sectioned to 25-μm-thick cross-sections with a rotary microtome (RMS; Nihon Microtome Laboratory, Inc., Osaka, Japan). These thin sections were placed between glass slides and coverslips with a drop of water. Thereafter, these sections were analyzed using a confocal micro-Raman system (LabRAM HR-800MX; Horiba, Ltd., Kyoto, Japan) equipped with a confocal microscope (BX41, Olympus, Tokyo, Japan), a 50 × air objective (LMPLFLN; NA = 0.5, Olympus, Tokyo, Japan), and a 514 nm argon ion laser (Series 532; Melles Griot Laser Group, Carlsbad, CA, USA). In particular, the incident laser power was 20 mW, and the Raman signal was detected by an air-cooled charge-coupled device (CCD) camera situated behind a 600 lines/mm grating. The mapping was recorded with an integration period of 0.3 s and a step size of 300 nm. Moreover, the LabSpec5 software (Horiba, Ltd.) was used for image processing, and the raw spectral data were baseline corrected to remove the background from the fluorescence.

Dynamic viscoelastic measurement

The temperature dependence of the loss tangent (tan δ) was measured by the tensile forced oscillation method using a thermomechanical analysis apparatus (TMA SS6100; Hitachi High-Tech Science Corp., Tokyo, Japan). The water-swollen sample (1 mm (L) × 30 mm (R) × 3 mm (T)) immersed in water was subjected to a temperature increase from 30 to 100 °C at a rate of 0.5 °C/min. The frequencies for the measurement were 0.01 Hz; the span was 18 mm in the R-direction; and the load amplitude was 70 ± 20 mN. Viscoelastic behavior is sensitive to the drying and heat history of the samples before the measurement (Kojiro et al. 2008). To unify the histories of the samples, they were heated to 100 °C, naturally cooled down to approximately 20 °C, and then subjected to measurements in the water-swollen state.

Free compression testing

A uniaxial compression test was carried out using a material testing machine (CATY TC-2kN-NS; Yonekura Mfg. Co. Ltd., Osaka, Japan) (Fig. 1). The testing machine equipped with a container enables horizontal compression tests of samples in a water-swollen state under confined heating conditions while recording load-stroke curves during hydrothermal compression. The temperature during compression (Tc) was controlled by heating the water with a cartridge heater immersed in a water pit.

Fig. 1
figure 1

Schematic of the experimental apparatus

Full size image

The container was sealed and heated until the temperature detected by the thermocouple stabilized to the target temperature (Tc). Then, the water-swollen sample was placed between the cylinders (diameter: 15 mm) under 2–3 N to ensure that the R was in the compression direction. The container was closed again to heat the sample to a constant temperature under saturated steam. Five minutes after achieving the temperature Tc, compression testing was conducted at a constant speed of 1 mm/min up to 3000 N of the compression load or until the maximum stroke of 5 mm of the testing machine. Tc was set to 40, 60, 80, and 100 °C in consideration of the decrease in the glass transition temperature of lignin due to delignification (Nakajima et al. 2009). Compression testing was conducted using three or more samples under the same conditions.

For one sample under each condition, the container was opened after compression testing, and the compressed sample was dried at room temperature to retain its shape while maintaining the stroke after testing, followed by further drying at 105 °C for 1 h. Then, the mass (md) and thickness (rc) in the compression direction were measured. The final compression ratio (Cd) was calculated as follows:

$$C_{{\text{d}}} = \left( {r_{{\text{b}}} - r_{{\text{c}}} } \right)/r_{{\text{b}}} \times 100\left[ \% \right]$$
(4)

where rb is the dimension in the R-direction of the water-swollen sample before the compression testing. The appearance of the compressed sample from the LT planes was observed using an optical microscope (VHX-970F; Keyence Corp., Osaka, Japan). The deformed cross-sectional area of the LT planes before and after compression testing was calculated by binarizing the captured image, and the area magnification (AMd) was calculated as follows:

$$AM_{{\text{d}}} = A_{{\text{c}}} /A_{{\text{b}}} \times 100\left[ \% \right]$$
(5)

where Ab and Ac are the areas of the deformed cross-sectional area of the water-swollen sample before compression and the dried sample after compression, respectively.

Results and discussion

Characterization of untreated and delignified samples

The mass loss and the area increase rate

Figure 2 shows the variation in mass loss (ML) (left axis, closed circles) and area increase rate (ART) (right axis, open circles) as functions of delignification time (tr). ML linearly increased with increasing delignification time. The ART also increased with increasing delignification time, indicating that delignification increased the wood dimensions in a water-swollen state. The increase in the dimensions of the water-swollen sample was caused by the swelling of the cell wall and CML. The total volume of the adsorbed water in the cell wall and CML exceeded the volume of lignin removed by delignification, which implied that the lignin removal increased the porosity and facilitated the water uptake.

Fig. 2
figure 2

Variation of mass loss (ML) and area increase rate (ART) with delignification time (tr)

Full size image

Lignin content

The variation in lignin content (LC) obtained from the acetyl bromide method as a function of delignification time (tr) is plotted in Fig. 3, which depicts the logarithmic decreases in LC with increasing tr. The LC of UW samples (tr = 0 min) was approximately 32%, which is consistent with previous reports (Pettersen 1984). After the longest tr (tr = 360 min), the LC decreased by approximately 13% from that of UW. The final value of ML expressed in Fig. 2 was very similar (13% of ML at tr = 360 min). Therefore, the contents of wood constituents other than lignin, such as polysaccharides, did not reduce after the delignification.

Fig. 3
figure 3

Variation of lignin content (LC) with delignification time (tr)

Full size image

ATR-IR spectroscopy

Figure 4 shows the ATR-IR spectra of the UW and DW samples, which indicate that the chemical structure of the wood sample changes with delignification. The area of the peak at 1490–1530 cm−1, which corresponds to the skeletal vibrations of the benzene ring in lignin, decreased with increasing delignification time. This indicates the initiation of the reaction during the early stages of delignification, as the benzene rings of lignin start to disappear. This tendency is consistent with that of the LC (Fig. 3) calculated from the UV absorbance of the benzene rings. The absorbance peak at 1725–1750 cm−1 corresponding to the C = O stretch of lignin and hemicellulose initially increased (tr = 10 min) before decreasing in the later stages (tr = 360 min). The initial increase in this peak can be attributed to the reaction between NaClO2 and lignin, which indicates that the aromatic ring was cleaved (Li et al. 2017). On the other hand, the decrease in the later stage was due to the decrease in the amount of lignin. The initial structural changes of lignin due to delignification consisted of the elimination and cleavage of the benzene ring, followed by the elimination of the benzene ring.

Fig. 4
figure 4

ATR-IR spectra of the untreated-wood (UW, tr = 0 min) and delignified wood (DW, tr = 10, 30, 60, 180 and 360 min) samples in the dry state

Full size image

Solid state NMR measurements

Figure 5 shows the 13C cross-polarization (CP)MAS NMR spectra for each substituent in the UW and DW samples. Signals corresponding to biomass constituents in the wood were assigned based on our previous report (Nishida et al. 2014). The carbohydrates appeared as relatively large and sharp signals in the range of 60–110 ppm. However, most of the signals for cellulose and hemicellulose, except cellulose C4, overlapped with each other, and the crystalline and amorphous signals for cellulose C4 and C6 could be separately observed. The aromatic and olefinic groups in lignin appeared as broader signals in the range of 110–160 ppm, while the methoxy groups in lignin appeared as isolated signals at 56 ppm.

Fig. 5
figure 5

13C CP-MAS NMR spectra of untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 10, 30, 60, 180 and 360 min) samples in dry state

Full size image

During the early stages of delignification (tr = 10 min), the signal intensity of the OCH3 (56 ppm) and aromatic (110–160 ppm) groups in lignin rapidly decreased was observed. Meanwhile, the intensity of the C = O signal at a lower magnetic field (172 ppm) increased for 30 min, then gradually decreased as delignification progressed. The trends of the signal intensities for the aromatic and C = O groups in the 13C CP-MAS NMR spectra were similar to those in the ATR-IR spectra (Fig. 4). Therefore, in the first step of delignification, oxidation of the benzene ring of the guaiacyl unit on the surface of the lignin unit afforded two carboxyl groups at C3 and C4 positions (Hamzeh et al. 2008). Next, the cleavage of C–C and C–O bonds that provide bridging with other aromatic ring units at the inner site of the lignin resulted in higher ML of the delignified wood, because the oxidized portion was removed (Tarvo et al. 2010 and Qu et al. 2020). However, the signal pattern of the carbohydrates in the 13C CP-MAS NMR spectra barely changed as a result of delignification, indicating that the chemical structures of cellulose and hemicellulose remained unchanged as delignification progressed.

Figure 6 shows the 1H spin–lattice relaxation times in the laboratory frame (T1H) of the dry samples. A minimal change was observed in the T 1H values at 65 and 75 ppm for tr = 60 min, but tended to increase after 60 min. For wood in the dry state, the spin–lattice relaxation for each wood constituent occurs via lignin, which has the shortest T1H value among the biomass constituents (Nishida et al. 2017). The increase in T1H of the DW samples at tr > 60 min was caused by the loss of lignin units with the short T1H values. In the early stages (tr ≤ 60 min), delignification occurred at the distal part of the polysaccharide chain, with minimal effect on the polysaccharides. With increasing tr (> 60 min), the reaction progressed toward the polysaccharide chain, resulting in an increase in the because of the reduced interactions between the polysaccharides and lignin.

Fig. 6
figure 6

1H spin–lattice relaxation time in the laboratory frame (T1H) of the untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 10, 30, 60, 180 and 360 min) samples in the dry state at 65 and 75 ppm

Full size image

Raman imaging analysis

The variations in the cellular-scale distribution of lignin with delignification were visualized by the Raman imaging technique (Fig. 7 (g–l)). Raman images were obtained by integration of the absorption bands between 1525 cm−1 and 1700 cm−1 (Fig. S1), assigned to the aromatic lignin vibrations (Segmehl et al. 2019). In the untreated sample (g), the lignin content was higher in CML than cell wall and the highest at the cell corners. The lignin distribution varied specifically in the early stages of delignification (tr ≤ 60 min), in which the lignin in CML located particularly at the cell corner decreased with delignification (g–j). With increasing tr (> 60 min), both lignin in CML and cell wall gradually reduced (j–l). Therefore, the delignification initiated at the distal part of the polysaccharide chain in the early stages (Fig. 6, tr ≤ 60 min) progressed from the CML rather than from the cell wall. Thereafter (tr > 60 min), the decrease in cell wall lignin (Fig. 7(j–l)) reduced the lignin interaction with polysaccharides that are abundant in the cell wall, as indicated by the results of T1H (Fig. 6).

Fig. 7
figure 7

Optical microscopy images (a–f) and Raman images based on intensity of spectral region assigned to aromatic lignin (1525–1700 cm−1, g–l) on the cross-section of the untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 10, 30, 60, 180 and 360 min) sample in water-swollen state. Squares in optical microscopy images (a–f) denote measured positions of Raman images (g–l)

Full size image

Dynamic viscoelastic measurement

Figure 8 shows the temperature dependence of tan δ for the water-swollen samples. The tan δ peak is attributable to the glass transition of lignin, and the peak temperature (Tg) in Fig. 8 corresponds to the glass transition temperature of lignin in the wood samples (Kojiro et al. 2008; Nakajima et al. 2009). As delignification progressed, the Tg shifted to the lower temperature and the shape of the peak broadened because of the varying molecular weight distribution. These trends correspond well with those by Nakajima et al. (2009). In general, the shift of Tg to a lower temperature indicates the reduction of molecular motility constraints, such as the decrease in molecular weight. In case of the wood in water-swollen state, the shift of Tg was caused by the reduction in the lignin molecular weight as well as the increased porosity and adsorbed water in the lignin molecules by delignification (Fig. 2). The results in Fig. 8 reveal that these modifications initiated during the early stages of delignification (tr = 10 min).

Fig. 8
figure 8

Temperature dependence of tan δ of untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 10, 30, 60, 180 and 360 min) samples in the water-swollen state

Full size image

Deformability of untreated and delignified samples

Figure 9 shows the relationship between the nominal compressive stress (σ) and the compression ratio (C) during compression testing at each Tc. The σ slightly increased to a C of approximately 60% in all samples, during which the cell lumens in the wood samples gradually closed because of the buckling of the cell wall. As C increases, σ significantly increase, followed by an inflection point at which the rate of increase in σ decreases (indicated by the arrow in Fig. 9). These compressive behaviors were similar to those observed in our previous report (Miki et al. 2017). Before the inflection point, the cell lumens were completely closed and the sample was consolidated, resulting in a rapid increase in σ. After the inflection point, the rate of increase in σ decreased because the wood sample plastically deformed in the unconstrained direction (L or T). The inflection point was also detected in the flat region for tr = 360 min at 100 °C (shown by (d) ▽); under this condition, plastic deformation occurred before the cell lumens were completely closed. The inflection points were not detected at 40 °C, 60 °C, and 80 °C for tr = 360 min and at 100 °C for tr = 180 min. This is because the inflection points due to the plastic deformation overlapped with the region denoting the rapid increase in σ.

Fig. 9
figure 9

Relationship between the nominal compressive stress (σ) and compression ratio (C) during compression testing at each temperature during compression (Tc)

Full size image

The σ at the inflection point detected in Fig. 9 is the nominal compressive stress at the flow starting point (σy), which indicates the initial resistance of the samples to plastic flow. Regardless of temperature, the longer the delignification time, the smaller the σy. Therefore, the delignification process reduced the initial resistance to plastic flow.

Figure 10 shows photographs of the samples that were dried after free compression testing to preserve their shape. After compression testing, all the samples were able to maintain their deformed state after drying by pressure; they only deformed on T-direction, and not in the L-direction. Such anisotropy of plastic flow deformation is consistent with the phenomenon observed in our previous report (Miki et al. 2017). At high Tc, the sample that was delignification for a longer time exhibited the highest Cd of 91% (Fig. 10d), indicating that a considerably thin product could be fabricated from solid-state wood.

Fig. 10
figure 10

Photographs of the untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 360 min) samples whose shape was fixed by drying post-free compression testing. The final compression ratio (Cd) is shown in each photograph. The part surrounded by the dotted line of the circle shows the sample protruding from the cylinders of the testing machine

Full size image

Figure 11 shows an SEM image of the RT surface of the samples. None of the samples displayed cell wall destruction. Furthermore, evidence of mutual positional change between the cells was confirmed. The plastic flow deformation was mainly caused by the shear sliding phenomenon between the cells and at the boundary of the CML. The sample subjected to a longer delignification time (Fig. 11c, d) displayed several slip surfaces, and plastic flow occurred in units with a smaller number of cells. The UW samples (Fig. 11a, b) appeared to be significantly compressed prior to flow deformation, whereas the DW samples (Fig. 11c, d) exhibited lesser compression than UW. In addition, the delignification impacted the strength of linkages in the R- and T-direction; therefore, the samples exposed to longer delignification could slip in R- and T- direction without undergoing significant densification process, as depicted in Fig. 11. The cells slightly protruding in the L-direction were also observed, suggesting that delignification promoted slip deformation in the L-direction.

Fig. 11
figure 11

The SEM images of the radial-tangential surfaces of untreated wood (UW, tr = 0 min) and delignified wood (DW, tr = 360 min) samples after free compression testing

Full size image

Influence of delignification on plastic flow deformation of wood

The relationship between the ML by the delignification of wood samples and σy is illustrated in Fig. 12, which portrays the initial resistance of the wood samples toward plastic flow. More specifically, a higher Tc corresponds to a lower σy that was logarithmically reduced to a ML of 4%, thereby indicating a significant improvement of the plastic deformability in the early stages of delignification. Furthermore, the decrease in lignin from CML during the early stage of delignification (Fig. 7g–j) potentially reduced the linkage strength between cells, and consequently, decreased the initial resistance of the wood samples toward plastic flow.

Fig. 12
figure 12

Relationship between the mass loss (ML) by delignification of wood samples and nominal compressive stress at flow starting point (σy). The results of reaction time (tr) of 180 min at 100 °C and tr of 360 min were not described because the inflection point could not be detected after the consolidation of the sample in Fig. 9

Full size image

The deformability was evaluated using the AMd calculated from the deformed cross-sectional area of wood samples. Figure 13a shows the relationship between ML and the AMd of the wood samples. The AMd increased with increasing Tc. The value of AMd was almost unchanged during the early stages of delignification (ML ≤ 4%); however, as delignification progressed, AMd remarkably increased. The value of AMd reached a maximum of 2.7 times for the longest tr of 360 min at the highest Tc of 100 °C. Furthermore, delignification caused a change in the density of the sample; therefore, the area per unit mass (Ad/md), which is an index of deformability, was calculated based on the area of the LT surface (Ad) and the mass (md) of the dried compressed samples. The relationship between ML and Ad/md of the wood samples is shown in Fig. 13b. The change in Ad/md depending on the Tc and the tendency toward ML were similar to those of AMd (Fig. 13a). In contrast to the tendency of σy in Fig. 12, the deformability evaluated from the deformed cross-sectional area did not increase during the early stages of delignification (ML ≤ 4%), but a significant increase was observed after a ML of 4% (tr > 60 min). Delignification with NaClO2 initially proceeds from the lignin-rich CML during the early stages of the reaction (Fig. 7g–j) and selectively softens the CML region. Then, the reaction and softening of the cell wall progresses during the later stages of the reaction (Fig. 7 (k, l), Xu et al. 2020). The results of T1H (Fig. 6) suggest that the lignin removal in the vicinity of the polysaccharide chains that make up the cell wall proceeds during the later stages of delignification (tr > 60 min), which can increase the flexibility of the cell wall. Those cell wall flexibility contributed to the increased deformability.

Fig. 13
figure 13

Relationship between mass loss (ML) by delignification and area magnification (AMd, a), and area per unit mass (Ad/md, b) of the wood samples

Full size image

A strong correlation is observed between the Tg (Fig. 8) and σy (Fig. 12), as shown in Fig. 14. Regardless of Tc, the σy linearly increased with Tg following the same slope. This indicates the enhancement of lignin molecular motility, which was influenced by the reduction in the lignin molecular mass and the increased amount of adsorbed water among the lignin molecules, significantly contributed toward the improved deformability at the initial state of plastic flow. Because the plastic flow of wood was shear failure originating from the CML (Fig. 11), structural defects in the CML generated during the early stages of delignification, such as decrease in lignin at the cell corners (Fig. 7), considerably affect the initiation of plastic flow.

Fig. 14
figure 14

The relationship between the peak temperature of tan δ (Tg) from Fig. 8 and the σy from Fig. 12. Tg is the average of three measurements

Full size image

Conclusion

We clarified the influence of delignification on plastic flow deformation due to the shear sliding between wood cells. The deformability was evaluated by the stress at the flow starting point and deformed cross-sectional area of wood samples obtained by free compression testing. The lignin content in the wood gradually decreased with delignification time. The ATR-IR and solid-state NMR spectroscopic analyses of delignification showed that the oxidative opening of guaiacyl ring on the surface of the lignin unit occurred in the early stages and the oxidized portion was released from the inner site of the lignin unit when delignification time was extended. Raman imaging analysis revealed that the lignin in CML decreased significantly, especially at the cell corner, followed by both lignin in CML and cell wall were reduced. In the early stages of delignification, the decrease in lignin in the CML reduced the stress at flow starting point. However, the deformed cross-sectional area of wood exhibited slight variations with delignification time during delignification in the early stages. In the later stages of delignification, the reaction possibly reached up to the vicinity of polysaccharide chains, causing larger deformations in the samples. Moreover, the stress at the flow starting point linearly increased with the peak temperature of tan δ corresponding to the glass transition temperature of lignin in water-swollen wood, regardless of the temperature during compression. The strategy of controlling the molecular mass and amount of lignin via delignification can yield more productive materials, as well as metal and plastic materials for achieving low-energy plastic flow deformation of wood. In this study, we examined delignified wood in the water-swollen state, but in the future, we plan to investigate the effect of other adsorbents, such as resin monomers, instead of water, on the plastic deformability of delignified wood.