Creation and structural evaluation of the three-dimensional cellulosic material “White-Colored Bamboo”

2.1 Sample preparation

Current year culms of moso bamboo (Phyllostachys edulis) were taken from a plantation at the Fuchu campus of the Tokyo University of Agriculture and Technology. Bamboo blocks were obtained from the middle of the 9th internode. The size of blocks was 10 mm (tangential (T)) × 5–7 mm (radial (R)) × 20 mm (longitudinal (L)) (Figure 1a).

Figure 1:

Sample structure (a), cross-section of outer side (b) and inner side (c) observed by light microscopy combined with safranin and Astra blue staining.

2.2 Chemical treatments for removing non-cellulosic components

Bamboo blocks were treated by conventional delignification methods; each method has a different mechanism, as shown in Table 1. Alkaline treatment (T_S) was carried out at 95 °C in 5% sodium hydroxide solution for 2 h. Alcoholysis (T_A) involved the following steps: (i) bamboo blocks were immersed in a solvent of propylene glycol:48.5% sulfuric acid = 99:1 based on a weight ratio, and (ii) soaked samples were placed in a portable reactor (TVS-N2, Taiatsu Techno, Tokyo, Japan) and heated at 150 °C for 1 h. The Wise method (T_W) is a bleaching treatment carried out by using sodium chlorite solution with acetic acid (Wise et al. 1946). The samples were soaked in bleaching solution for more than 7 h at 75 °C with the addition of these reagents after every hour until the color of the samples was no longer changed. The obtained bamboo blocks were freeze-dried for following different analyzes.

2.3 FTIR spectroscopy

Bamboo blocks before and after the chemical treatments were cut from the outer and inner sides because the treatment progress was predicted to vary depending on the distribution of vascular bundles and parenchyma cells in the radial direction (Figure 1b, c). The chemical components of these specimens were measured by attenuated total reflection accessory with Fourier transform infrared (FTIR) spectroscopy (Spectrum Frontier, PerkinElmer, Waltham, MA, USA). The spectra were recorded from the range of 2000–500 cm⁻¹ at a resolution of 4 cm⁻¹ with an acquisition of eight scans for each specimen. After applying the ATR corrected function using the accompanying software package followed by baseline correction, the spectra were normalized based on the band at 1030 cm⁻¹ in the fingerprint region, whose absorbance was taken to be 1. Non-cellulosic components were evaluated from bands at 1730 cm⁻¹ and approximately 1510 cm⁻¹ which corresponded to hemicellulose (C=O stretching vibration in the acetyl group and carboxylic acid) and lignin (aromatic skeletal vibration), respectively (Faix 1991; Horikawa et al. 2019).

2.4 X-ray computed tomography

Dried bamboo blocks were scanned at 32 kV and 470 μA using X-ray computed tomography (X-ray CT; SkyScan 2211, Bruker, Billerica, MA, USA) to create 2D- and 3D-reconstructed images. A rotation step of 0.4 degrees and resolution of 10 μm was used. Additionally, vascular bundles of bamboo blocks were visualized by image processing.

2.5 X-ray diffraction analysis

X-ray fiber diffraction diagrams of the outer and inner specimens were obtained using a wide-angle X-ray diffractometer (R-AXIS RAPID, Rigaku, Tokyo, Japan) at a voltage of 32 kV and current of 250 mA to investigate the microfibril orientation. For these measurements, the longitudinal direction of the specimens was set along the vertical direction of the imaging plate, and the X-ray was perpendicularly irradiated to the tangential face of the bamboo block. The microfibril angles were calculated from the 200-plane reflection of cellulose I_β by Gaussian function fitting.

X-ray diffractograms were also obtained in reflection mode by employing an automated multipurpose X-ray diffractometer (SmartLab, Rigaku, Tokyo, Japan) at a voltage of 50 kV and current of 250 mA.

2.6 Transmission electron microscopy

After chemical treatments, a specimen was suspended in water and oxidized using the 2,2,6,6-tetramethylpiperidine-1-oxyl radical (TEMPO)/NaBr/NaClO system (Saito et al. 2006). This reaction was conducted at pH 10 and room temperature for 2 h. Next, the oxidation product was reduced by adding NaBH₄ and stirring for 3 h (Takaichi et al. 2014). After repeated washing and subsequent homogenization with a double-cylinder-type homogenizer (Hiscotron, Microtec, Chiba, Japan), individual bamboo cellulose microfibrils, dispersed in water, were obtained.

The TEMPO-oxidized cellulose suspension was spotted onto a copper grid supported by a carbon film (ELS-C10, Okenshoji, Tokyo, Japan). Cellulose microfibrils were observed by negative staining with 2% uranyl acetate using transmission electron microscopy (TEM; JEM-1400 Plus, JEOL, Tokyo, Japan) at 80 kV.

3.1 Determination of optimal treatments for removing non-cellulosic components

After T_S and T_A, the color of the bamboo blocks became darker (Figure 2a), which may have resulted from the production of colored components by excessive decomposition of hemicellulose or modification of lignin. T_W also changed the color of the samples to brown on the outer side due to oxidation. FTIR spectra after each single treatment showed bands at 1730 cm⁻¹ (assigned to hemicellulose) or 1510 cm⁻¹ (assigned to lignin) or both decrease on the outer and inner sides (Figure 2b, c). However, each treatment alone could not completely remove the non-cellulosic components.

Figure 2:

Sample images (a) and fourier transform infrared (FTIR) spectra of bamboo specimens from outer side (b) and inner side (c) in each single treatment: (i) untreated, (ii) T_S, (iii) T_A, and (iv) T_W.

In the case of T_A-S and T_S-A, the colors of the bamboo blocks were similar to those of the samples after T_S and T_A (Figure 3a), where the band at 1510 cm⁻¹ remained (Figure 3b, c). After T_A-W the band at 1510 cm⁻¹ disappeared in the spectrum from the inner side with the progression of decolorization, whereas the outer side was not significantly changed compared to the untreated control (Figure 2b). For T_S-W, although the outer side remained colored, the bands at 1730 and 1510 cm⁻¹ disappeared. This indicates that the amounts of lignin and hemicellulose were below the FTIR detection limit. Therefore, both combinations of the 2 treatments were not completely effective.

Figure 3:

Sample images (a) and FTIR spectra of bamboo specimens from outer side (b) and inner side (c) when two treatments were combined: (i) T_A-S, (ii) T_A-W, (iii) T_S-A, and (iv) T_S-W. (ii), (iv) show longitudinal planes.

Finally, when T_S-A-W was conducted on the bamboo blocks, even the outer sides were completely whitened (Figure 4a, b). Furthermore, the FTIR spectra after T_S-A-W did not show bands at 1730 or 1510 cm⁻¹ on either side. The FTIR spectra after T_S-A-W exhibited spectra similar to that of cellulose purified from bamboo powder (Figure 4c) using a conventional method for lignified samples (Horikawa 2017). As a result, T_S-A-W was determined to be the optimal technique for selectively removing non-cellulosic components from bamboo blocks. Li et al. (2016) conducted sodium chlorite treatment alone and successfully developed delignified wood from Balsa, which is a low-density wood material. In contrast, bamboo is a higher-density material that cannot undergo complete delignification and decolorization using sodium chlorite alone (Figure 2). However, combining the three processes resulted in complete delignification and decolorization (Figure 4). In this study, the bamboo block after T_S-A-W was named as “White-Colored Bamboo”, and its hierarchical structure was further investigated.

Figure 4:

Sample images of front and back sides (a), longitudinal plane (b) and FTIR spectra (c) of purified cellulose from bamboo powder (i) and obtained from specimens (ii, iii). (ii) was from the outer side and (iii) was from the inner side.

3.2 Evaluation of hierarchical structure

3.2.1 Anatomical structure

The anatomical structures of untreated and white-colored bamboo were compared using X-ray CT (Figure 5). Although cracks and shrinkage caused by drying were observed, white-colored bamboo did not significantly change the shape and distribution of vascular bundles (Figure 5a, d). In addition, the 3D morphology (Figure 5b, e) and arrangement of vascular bundles (Figure 5c, f) were also unchanged. As a consequence, the parenchyma cells appeared to retain their original form.

Figure 5:

Cross-sections (a), (d) and 3D-reconstruction images (b), (c), (e), and (f) of bamboo blocks scanned by X-ray CT. (a–c) shows untreated bamboo while (d–f) are white-colored bamboo (after T_S-A-W). (c) and (f) are modified from (b) and (d) to visualize vascular bundles by image processing. The length in L and T directions differs from the actual sample size because of the measurement range.

Some particles were observed in the inner side of the untreated bamboo (Figure 5c) and in white-colored bamboo (Figure 5f). It has been reported that starch grains are most abundant on the inner side and decrease towards the outer side (Okahisa et al. 2007). Therefore, these particles may have been starch grains, suggesting that their quantity was reduced by the chemical treatments.

3.2.2 Orientation and crystalline structure of cellulose microfibrils

Figure 6 shows the X-ray fiber diffraction diagrams of the inner and outer sides of the specimens. The diffraction patterns differed between the outer (Figure 6a, b) and inner sides (Figure 6c, d) because of differences in the ratio of fiber and parenchyma cells in the radial direction, as previously reported by Okahisa et al. (2018). This trend on both sides was maintained before and after the treatments; thus, from a qualitative perspective, the microfibril orientation was undisturbed.

Figure 6:

X-ray fiber diffraction diagrams from the outer side (a, b) and inner side (c, d). (a, c) were obtained from untreated while (b, d) were acquired from white-colored bamboo blocks.

For more quantitative assessment, the microfibril angles were calculated from a fiber diffraction diagram obtained using 200-plane reflection. The microfibril angles of untreated bamboo were 2.2 and 4.0 degrees on the outer and inner sides, respectively (Table 2). The cell wall layer of fiber or parenchyma cell layer has a polylamellar structure (Liese and Köhl 2015). Particularly, the fiber cell wall structure has alternating thick and thin layers. Microfibrils in the former layer are orientated nearly parallel to the cell axis, whereas microfibrils in the latter layer are oriented nearly perpendicular to the cell axis (Parameswaran and Liese 1976). Lower microfibril angles on the outer side which contains densely packed with fibers are formed because of the longitudinal orientation of the thick layers. After the treatments, the microfibril angles were 1.9 and 3.4 degrees on the outer and inner sides, respectively (Table 2), showing similar tendencies to form fibers which were close to those of untreated bamboo. Hence, the microfibril orientation in the cell wall was maintained in the white-colored bamboo.

Table 2:

Microfibril angles of bamboo blocks calculated from the 200-plane reflection in Figure 6.

	Microfibril angles (deg.)
	Outer side of vascular bundles	Inner side of vascular bundles
Untreated bamboo	2.2	4.0
White-colored bamboo	1.9	3.4

X-ray diffraction analysis also revealed the crystalline structure of cellulose in the cell wall layer (Figure 7). Before and after treatment, bamboo specimens with the outer (Figure 7a, b) and the inner (Figure 7c, d) sides showed a typical diffraction pattern of cellulose I (Wada et al. 1997). The obvious sharpening of the diffraction peaks in the white-colored bamboo resulted from the removal of matrix components which led to a better microfibril orientation. Consequently, the chemical treatments used in this experiment did not cause crystalline modification.

Figure 7:

X-ray diffractograms from outer side (a, b) and inner side (c, d). (a, c) obtained from untreated, whereas (b, d) were acquired from white-colored bamboo blocks.

3.3 Morphology of cellulose microfibrils

The TEM imaging results confirmed that the TEMPO-oxidized cellulose microfibrils obtained from white-colored bamboo were uniformly individualized (Figure 8a). The microfibril width calculated from the magnified view was 3.77 ± 0.95 nm (Figure 8b). The width of approximately 3–4 nm is similar to that of terrestrial plants (Kuramae et al. 2014). In addition, cellulose microfibrils were much longer and not fragmented. Cellulose microfibrils obtained from terrestrial plants become fragmented to approximately 150 nm by acid hydrolysis (Horikawa et al. 2018; Nickerson and Habrle 1947; Nishiyama et al. 2003). Harsh alkaline treatments also decrease the length of microfibrils (Pavasars et al. 2003). However, these phenomena were not observed for white-colored bamboo. Thus, the morphology of cellulose microfibrils was preserved after treatment.

Figure 8:

Transmission electron microscopy (TEM) images of 2,2,6,6-tetramethylpiperidine-1-oxyl radical (TEMPO)-oxidized cellulose microfibrils obtained from white-colored bamboo: (a) overview and (b) magnified view.