Efficient separation of cellulose from bamboo by organic alkali

This study focuses on the effective separation of cellulose from bamboo through a two-step process. Several organic alkalies 2-pyrrolidinone, ethylurea, dibutylamine, N-methylformamide and tetramethyl guanidine were used to remove lignin and isolate cellulose from massive bamboo. The results showed that tetramethyl guanidine can effectively remove lignin and hemicellulose while retaining almost all the cellulose in the residual solid. The removal of lignin and hemicellulose achieved 86.0 and 84.0% after heating for 3 h at 150 °C, respectively. Subsequently, the final removal of lignin and hemicellulose increased to 91.5 and 93.8%, respectively, after a simple alkaline H2O2 bleach treatment. Interestingly, the loss of cellulose was very small after two-step treatments, and 96.9% of the component was still retained. The crystallinity increased from 69.8–75.2% in X-ray diffraction graphs due to the removal of lignin and hemicellulose. The Scanning electron microscopy images indicated that the diameter of cellulose bundles decreased from 80–100 µm to about 50 µm after organic alkali treatment, and then the fiber bundle was completely separated into a single long fiber with a diameter of about 10 µm after H2O2 bleaching. The Fourier transform infrared spectroscopy spectra confirmed the high selective removal of lignin and hemicellulose. Two-dimensional 1H-13 C Heteronuclear singular quantum correlation were analyzed to investigate the lignin structure and found that only the signals of –OCH3, Cγ–Hγ in β-O-4’ (Aγ) and β-β’ (Cγ) structures and C5-H5 in guaiacyl (G5) did not disappear after two-step treatment.


Introduction
With the consumption of fossil resources and the resulting environmental problems, biomass is considered to be an alternative kind of environment-friendly natural resources . As one of the main components of biomass, cellulose could be widely found in plants, some species of algae and biofilms secreted by bacteria (Moon et al. 2011). Since its sustainable potential and unique multidimensional structure, it is not only used in the paper industry, but also is attractive in oilfield applications, filler in rubber, nanogenerators, catalyst support, nanocomposite films and textile industry (Damay et al. 2019;Li et al. 2015;Liu et al. 2021). Bamboo is widely distributed in America, Asia and Africa. Due to its fast growth speed and short cycle, bamboo is regarded as an important source of renewable biomass, which is widely used in the papermaking industry (Silva et al. 2020;Muhammad et al. 2019). Compared with wood, bamboo has higher cellulose content, which is more conducive to the development and utilization of lignocellulosic biomass industry (Rusch et al. 2021).
Over the past few years, researchers have made considerable effort to separate cellulose from biomass. Since hemicellulose is easy to hydrolyze, the key problem of isolating cellulose from biomass is how to effectively degrade and remove lignin (Zhou et al. 2017). So far, the methods of removing lignin and isolating cellulose from lignocellulosic biomass including acidic aqueous solution, alkaline aqueous solution, ionic liquid, deep eutectic solvents and organic solvents were reported Lin et al. 2021;Guo et al. 2018;Ponnusamy et al. 2019). Chemical pulping is a typical chemical process of lignocellulosic delignification for the production of cellulose. Acid sulfite/sulfate and alkaline sulfate cooking were the main technical methods of cellulose separation in paper pulping (Aro and Fatehi 2017;Wanrosli et al. 2007). Although the removal of lignin was effective, the subsequent treatment of water pollution was quite troublesome. In order to solve this problem, the organic solvent method has attracted extensive attention due to the advantages of the high solubility of organic solvent to lignin and the recyclability (Zhang et al. 2016). Zhong et al. (2021) used 85.0% aqueous formic acid as solvent to obtained 38.9% yield of cellulose from corn stover for 30 min at 130 ℃. Ionic liquids (ILs) were widely used in biomass isolation ascribing to their high thermal stability and nearly non-volatility Liu et al. 2012;Halder et al. 2019). Financie et al. (2016) used a hydrophilic ILs [EMIM][DEP] to treat oil palm frond biomass followed by enzymatic delignification, and found that the contents of lignin and hemicellulose decreased from 24.0 to 8.5%, 20 to 12.1%, respectively. [Bmim][OAc] was reported to remove the lignin of 31.47% and xylan of 43.95% from corn stover . Recently, deep eutectic solvents (DESs) with physical and chemical properties similar to ionic liquids have also been used to dissolve lignin and separate cellulose from biomass (Tan et al. 2020). Bamboo residue was treated in a deep eutectic solvent composed of choline chloride and lactic acid in a molar ratio of 1:1, and the removal yields of 56% xylan and 26% lignin were observed for heating 90 min at 130 °C (Wu et al. 2021). Likewise, Wang reported a twostep treatment process of moso bamboo powder and obtain cellulose with a purity of 92.7% and 80.1% of delignification after hydrothermal treatment for 10 min at 200 °C, followed by lactic acid-choline chloride (10:1) DESs treatment for 6 h at 140 °C . Despite considerable efforts, the separation of cellulose from biomass currently encounters low cellulose purity and delignification, high loss of cellulose, high cost of recovery of solvents and environmental pollution .
Herein, in order to solve the above-mentioned problems in the process of separating cellulose from biomass, a two-step process was proposed as illustrated in Fig. 1. In the first step, a two-component organic alkali-water system was used for the treatment of block Bambusa emeiensis L. C. Chia & H. L. Fung (BECF) at 150 °C. This process can effectively dissolve lignin and hemicellulose into the mixed solvents and the solvents could be recovered and reused by vacuum distillation. The lignin and hemicellulose solids obtained by distillation can be further processed into high value-added products. This closed-loop treatment process produced no pollution. To further obtain higher purity cellulose, a simple alkaline H 2 O 2 bleaching treatment was used. Finally, a purity of 94% bamboo cellulose was obtained with only 3.1% cellulose loss.

Organic alkali treatment
The dried BECF was cut into small pieces with dimensions of about 1.5-2 cm (length), 3-5 mm (width) and 3-4 mm (thickness) to be used for the following experiments. A typical organic alkali treatment process was carried out as follows: 1 g bamboo pieces and 15 mL solution containing TMG, DMF and water were put into a 25 mL capacity autoclave and heated at 150 ℃ for 3 h. After cooling to room temperature, the solid residue was separated by filtration and then washed with water 3 times, and dried to obtain light yellow crude cellulose. Hemicellulose and lignin were dissolved in the brown black filtrate. The filtrate was distilled under reduced pressure to remove the solvent. After adding water, the brown lignin precipitated in the lower layer and hemicellulose was dissolved in the upper aqueous solution. After filtration and drying, brown powdered lignin was obtained.

Alkaline bleaching
The crude cellulose solid was subjected to alkaline bleaching treatment. The crude cellulose solid was put into a solution containing 3 wt.% sodium hydroxide, 6 wt.% hydrogen peroxide, 2 wt.% sodium silicate and 0.3 wt.% magnesium sulfate and heated at 60 ℃ for 5 h. Then filter to obtain white fluffy cellulose.

Preparation of solution
Solution A: 10 mL of 65% nitric acid solution was added to 100 mL of 80% acetic acid solution to obtain Solution A.
Solution B: 25 g K 2 Cr 2 O 7 was dissolved in 250 mL distilled water. 800 mL of concentrated sulfuric acid with a density of 1.84 was gradually added into the solution in ice water bath. The mixture solution was then stored in a glass jar in the dark.
Indicator A: 1.458 g phenanthroline and 0.695 g ferrous sulfate were dissolved in water. The solution was diluted to 100 mL and stored in a brown bottle.
Indicator B: 12 g CuSO 4 ·5H 2 O and 21 g tartaric acid were dissolved in a beaker with 400 mL of distilled water, and this solution was added to 400 mL aqueous solution containing 75 g Na 2 CO 3 . Then 0.89 g KIO 3 and 8 g KI were added to the mixed solution. The solution was transferred to a volumetric flask of 1 L and diluted with distilled water to the scale. Then the mixture was transferred to flask and heated to boil for 5 min. The mixture was left overnight at room temperature to cool. The clear solution was transferred and stored into a glass bottle.
Solution C: 60 g oxalate was dissolved in 800 mL distilled water and 70 mL of concentrated sulfuric acid with a density of 1.84 was gradually added into the solution. The solution was diluted to 1 L with distilled water.
Solution D: 40 g (NH 4 ) 2 SO 4 ·FeSO 4 was dissolved in distilled water and 20 mL concentrated sulfuric acid was added to the solution. The solution was diluted to 1 L with distilled water. The solution should be prepared within a week before use.

Analytical methods and instruments
The contents of cellulose, lignin and hemicellulose of raw bamboo and the treated bamboo were calculated according to the method provided in the literature (Qi et al. 2006;Hu et al. 2014). And the details are as follows: Cellulose: 0.05-0.06 g sample was added into 5 mL Solution A. The mixture was heated and kept for 25 min at 100 °C with stirring regularly. Then the mixture was centrifuged and the precipitate was washed with distilled water (10 mL × 3). The precipitate was dissolved with 10 mL Solution B and the solution was heated in boiling water for 10 min with stirring. After cooling to room temperature, the mixture was poured into a conical flask. After adding 3 drops of Indicator A, the solution was analyzed by titration to reddish brown with 0.1 N of Solution D.
Blank Control test: solution B was diluted from 10 mL to 15 mL, 3 drops of Indicator A were added, and then titrated by solution D.
The content of cellulose was calculated as Eq. (1) where x% is the percentage content of cellulose; a is the volume (mL) of consumed solution D in blank control test; b is the volume (mL) of consumed solution D for sample analysis; n is the weight of samples (g).
Hemicellulose: 0.1-0.2 g sample was added into 15 mL of 80% Ca (NO 3 ) 2 solution. The mixture was heated and kept for 5 min with stirring regularly at 100 ℃. Then the mixture was centrifuged and the precipitate was washed with distilled water (10 mL × 3). The precipitate could be dissolved by adding 10 mL of 2 N HCl solution and heated in boiling water for 45 min with stirring. The residue was centrifuged and washed three times with deionized water (10 mL × 3) and the rinsed aqueous solution was incorporated in the previous acidic centrifugal solution. 1 drop of phenolphthalein reagent and 2 N NaOH solution were added drop by drop to the solution until the solution turns purple red. The solution was then transferred to a 100 mL volumetric flask and diluted to scale. 10 mL of diluted solution and 10 mL Indicator B were taken into a large test tube and heated in boiling water for 15 min. After cooling to room temperature, 5 mL Solution C and 0.5 mL of 0.5% starch were added. Then the solution was titrated with 0.01 N Na 2 S 2 O 3 until the blue color disappears. The blank control test was also performed according to the above method. The content of hemicellulose was calculated as Eq. (2) where y% is the percentage content of hemicellulose; a is the volume of Na 2 S 2 O 3 solution consumed in blank control test (mL); b is the volume of Na 2 S 2 O 3 solution consumed for sample analysis (mL); n is the weight of sample (g).
Lignin: 0.05-0.1 g sample was added to 10 mL of 1% CH 3 COOH aqueous solution and the mixture was shaken for 5 min. The mixture was centrifuged and the precipitate was washed with 5 mL of 1% CH 3 COOH aqueous solution. Then the solid was soaked with 3-4 mL acetone for 3 min three times (3 mL × 3) and dried. 73% H 2 SO 4 aqueous solution of 3 mL was added to the precipitate with stirring. The mixture was left overnight at room temperature to dissolve the cellulose thoroughly. Then 10 mL distilled water was added and the mixture was kept in boiling water for 5 min. After cooling to room temperature, 10% BaCl 2 aqueous solution of 0.5 mL was added and centrifuged. The precipitate was washed with water twice (10 mL × 2). The precipitate was added to 10 mL Solution B and heated for 15 min in boiling water with stirring. After cooling, the mixture was poured into a conical flask. The precipitate was rinsed with distilled water and the washing liquid was combined into the solution in the conical flask. After adding 3 drops of Indicator A, the solution was titrated to reddish brown 0.1 N of Solution D. The blank control test was also performed according to the above method and the content of lignin was calculated as Eq. (3) where z% is the percentage content of lignin; a is the volume of consumed Solution D in blank control test; b is the volume of consumed Solution D for sample analysis; n is the weight of samples (g).
The crystallinity CrI of sample was calculated as Eq. (4) according to the literature (Ren et al. 2021;Li et al. 2009).
where I 200 is the maximum intensity of the lattice diffraction at around 22.5° 2θ, I am is the intensity of the amorphous cellulose at 18° 2θ (French 2014).
Scanning electron microscopy (SEM, JSM-7500 F, JEOL, Japan); Fourier transform infrared spectra (FT-IR, Nicolet NEXUS 670, America); Nuclear magnetic resonance (NMR, AM-400, Bruker, Switzerland) of samples were detected at 300 K using DMSO-d 6 as the solvent; X-ray diffraction (XRD, Rigaku Ultima IV, Japan) of samples were ground and laminated at room temperature and then made into a 2θ intensity vs. curve with a scan range of 5° to 40°; Planetary ball mill (YXQM-8 L) was employed to mill bamboo powder. WSB-2 Whiteness Meter was applied to measure brightness (ISO). Milled Wood Lignin (MWL) was obtained according to Bjorkman method. The procedures were carried out as: 10 g of the bamboo powder was extracted with tolueneethanol (2:1, v/v) for 24 h. Then, the sample was dried for 3 h at 70 ℃. The dried sample was milled for 8 h using a planetary ball mill at a speed of 250 rpm (15 balls of 2 cm diameter and 30 balls of 1 cm diameter). The ballmilled sample was heated in 96% dioxane for 24 h at 60 ℃ with reflux condensation (the solid-to-liquid ratio was 1:10 g/mL). The mixture was centrifuged and filtered, the residue was washed with dioxane three times and the supernatant was transferred together. The collected supernatants were combined at reduced pressure and then dried in a vacuum oven at 70 °C for 3 h to obtain MWL.
The experimental data are the average values of three experiments, and the standard deviation is less than 3%.

Effects of different organic alkalies on chemical composition
One way to separate cellulose from lignocellulose is that lignin and hemicellulose were sheared from cellulose and can be dissolved in a solvent, while cellulose exists in solid form. Acids or alkalis can cut off the linking group between cellulose and lignin or hemicellulose, thereby stripping and freeing lignin and hemicellulose. However, acids may also hydrolyze cellulose into soluble sugars, resulting in a large loss of cellulose yield. In this work, we use the strategy of combining alkaline organic solvent and water to separate cellulose from BECF. Several organic alkalies 2-pyrrolidone, NMF, Urea, dibutylamine and TMG were selected to investigate the effects on three components in bamboo. The small pieces of bamboo were put into an organic alkali-water solvent and heated for 3 h at 150 °C, and the residual solids were analyzed after filtering and washing with water. The contents of cellulose, lignin and hemicellulose in raw bamboo and the residual solids were listed in Table 1.
From Table 1, it can be found that the contents of cellulose, hemicellulose and lignin in raw BECF were 48.5, 18.9 and 29.4%, respectively. The presence of nitrogen-containing organic bases made the aqueous solution alkaline, which degrading lignin and hemicellulose. Under the conditions of this study, the organic alkalis 2-pyrrolidone, NMF, Urea, and dibutylamine caused obvious effects on the removal of hemicellulose, and the removal ratio ranged from 42.6 to 61.0%. However, the removal of lignin by these systems was weak, and the removal ranged from 5.4 to 27.9%. Obviously, the strongest organic alkali TMG exhibited the best efficient, and the removal of lignin and hemicellulose reached 75.7 and 86.9%, respectively, after heating for 3 h at 150 °C. Fortunately, the loss of cellulose was only 1.9% during this process.

Effects of different proportions of TMG, DMF and water
The solubility of solvent to lignin has an important influence on delignification. Previous studies by Raikwar et al. (2021) and Xue et al. (2016) have confirmed that DMF was an excellent solvent for the dissolution of lignin. In this study, DMF was selected as a cosolvent to increase the solubility of lignin. The effects of various volume proportions of TMG, DMF and H 2 O on the contents of cellulose, lignin and hemicellulose in residual solid were investigated after heating for 3 h at 150 °C. The results were illustrated in Fig. 2. As can be seen from the figure, when only water and TMG or DMF existed, the removal of hemicellulose and lignin were far inferior to those of the three solvents' coexistence. When the volume ratio of TMG: DMF: H 2 O arrived 7: 3: 5, the removal of hemicellulose and lignin were the highest, reaching 84.0% and 86.0%, and cellulose retention yield can reach 97.8%. On the basis of optimized experimental data, 7: 3: 5 was selected as the best ratio under this treatment condition.

Alkaline bleaching treatment
After the bamboo was treated with organic alkali, the remaining solid still contains 7.3% of lignin and 5.4% of hemicellulose, respectively. In order to obtain highpurity cellulose, an alkaline H 2 O 2 bleaching treatment was employed to remove the residual lignin. The remaining solid was put into a solution containing 3 wt.% sodium hydroxide, 6 wt.% hydrogen peroxide, 2 wt.% sodium silicate and 0.3 wt.% magnesium sulfate and heated at 60 ℃ for 5 h. Then filter to obtain white fluffy cellulose with a purity of 94% and a whiteness of 66.5%. As a result, after this two-step treatment, lignin and hemicellulose in bamboo were removed by 91.5 and 93.8%, respectively, and the loss of cellulose was only 3.1%.

Microscopic investigation of samples
The SEM images of bamboo powder and cellulose fibers after organic alkali and alkaline bleaching were shown in Fig. 3. The SEM images intuitively showed the microscopic morphology of fibers of raw bamboo and treated cellulose. It can be seen from Fig. 3a and b that in bamboo raw powder, about 15-20 individual fibers form fiber bundles with a diameter of about 80-100 μm, and the periphery was closely connected with hemicellulose and lignin. After high-temperature organic alkali treatment, the connected lignin and hemicellulose were reduced, and the diameter of the fiber bundle was reduced to about 50 μm (Fig. 3c), indicating that this treatment process can dissociate the fiber bundles. After the alkali bleaching treatment, no lignin and hemicellulose were found around the fibers, and the fiber bundles were completely separated into individual fibers with a diameter of about 10 μm (Fig. 3d), this indicating that the treatment can completely dissociate the fiber bundles to obtain long bamboo fiber.

XRD analysis
The XRD patterns of the natural BECF powder and the isolated cellulose were illustrated in Fig. 4a, b.  From Fig. 4, it can be observed that the natural bamboo exhibited characteristic crystal peaks at around 16° and 21.9°, which corresponds to the 110 and 200 lattice planes of cellulose I, respectively. For the isolated cellulose, the position of peaks appeared at 16° 2θ and 22.4° 2θ. The 200 lattice plane has a certain deviation and is closer to the position at 22.5°. This indicated that the crystal structure of cellulose was still in the form of cellulose I after treated with hightemperature organic alkali and alkaline H 2 O 2 bleaching. It can also be found that the crystallinity of cellulose increased from 69.8 to 75.2% through a two-step process. The difference in crystallinity was due to the remarkable decrease of hemicellulose and lignin, as the two components mainly existed in the amorphous regions (Li et al. 2009).

FT-IR analysis
FT-IR spectra analysis was performed to investigate the varieties of chemical functional groups after these treatments. Figure 5 showed the FT-IR spectra of alkaline bleaching cellulose (a), organic alkali treatment (b) and bamboo powder (c). In Fig. 5, the characteristic peaks of lignin appeared at 1729, 1634, 1453, 1422, 1240 and 833 cm − 1 . The peak at 2944 cm − 1 belonged to the C-H vibration of -CH 2 in lignin. The peaks at 1602, 1508 and 1422 cm − 1 were attributed to the skeleton vibration of benzene ring, and the peak at 1729 cm − 1 was due to C = O stretching. The peak at 1370 cm − 1 represents the phenolic hydroxyl groups. After organic alkali treatment, it was obviously seen that the characteristic peaks of hemicellulose and lignin disappeared, leaving the absorption peaks of cellulose, whose characteristic peak was at 1314, 1160 and 896 cm − 1 (Wei et al. 2016). Especially, the absorption peak at 1240 cm − 1 , which belongs to the C-O-C stretching vibration of the aromatic ether, disappeared. The absorption peak near 1314 cm − 1 , which was attributed to the in-plane bending of hydroxyl group, was obviously strengthened; this indicated a high selective removal of lignin and hemicellulose through organic alkali treating process, which was consistent with the chemical titration analysis above. After alkaline bleaching, the disappearance of the peak at 1631 cm − 1 and the appearance of the peak at 1640 cm − 1 may be due to the formation of ketone oxidation of hydroxyl group of lignin in the presence of hydrogen peroxide.

NMR analysis
In order to investigate the effect of two treatment processes on lignin, the chemical structure information of both Milled Wood Lignin (MWL) and H 2 O 2 bleaching treated lignin (BTL) was analyzed by 2D HSQC. 2D HSQC NMR spectra of the lignin showed the predominance of lignin signals together with the remaining sugars. The side chain region (δ C /δ H 110 − 50/5.5-2.5 ppm) and aromatic region (δ C /δ H 150 − 100/8.5-5.5 ppm) of 13 C-1 H 2D HSQC NMR spectra of MWL and H 2 O 2 BTL were illustrated in Fig. 6. The main substructures of two kinds of lignin, including different units linked by ether and C-C bonds, and carbohydrates were depicted in Fig. 7.
As shown in Fig. 6, the side-chain regions (δ C /δ H 110 − 50/5.5-2.5 ppm) of NMR spectra of MWL and BTL both showed prominent signals corresponding to methoxyl (δ C /δ H 56.18/3.74 and 56.22/3.72 ppm). As seen in Fig. 6b, the spectra of MWL showed abundant chemical structure information, as observed in the previous reports (Wen et al. 2013;Hu et al. 2014). The C α -H α correlations in β-O-4' substructures were observed at δ C /δ H 72.82/4.92 and 71.55/4.72 ppm for structures linked to S and G lignin units.
The C β -H β correlations were found at δ C /δ H 81.82/4.02 ppm for β-O-4' S units. Likewise, The C γ -H γ and C 6 -H 6 correlations in β-O-4' appeared at δ C /δ H 60.71/3.51 and 128.36/7.32 ppm, respectively. Signals for phenylcoumaran substructures (B) were also detected in the spectrum of MWL, with their C α -H α , C γ -H γ correlations at δ C /δ H 89.16/5.30 and 68.02/4.15 ppm. A strong signal at δ C /δ H 72.70/3.51 ppm was corresponded to the resinols structures (C) with its C γ -H γ . Whereas, in the BTL spectrum (Fig. 6d), the signals of chemical groups were obviously different and there were new signals (δ C /δ H 100.45/5.11 and 72.01/3.60 ppm) appeared in BTL except for methoxyl and C γ -H γ still existed in alkylaryl ethers (A) and resinols (C), which may be caused by the residues of xylose and hexose in lignin due to hydrolysis of cellulose and hemicellulose.

Conclusion
In this study, a simple and effective method for separating cellulose from block BECF was proposed. The solvent can be recovered by distillation to obtain waterinsoluble solid lignin. This process was environmentally friendly and could produce almost no environmental pollutants. The organic alkali TMG exhibits the excellent removal ability on lignin and hemicellulose, and the addition of DMF and water improves the delignification effect. The diameter of cellulose bundles reduced from 80 to 100 μm to about 50 μm after organic alkali treatment, and then the fiber bundle was completely separated into a single fiber with a diameter of about 10 μm after H 2 O 2 bleaching treatment. NMR spectra were compared by MWL in detail, which showed that bamboo lignin was seriously degraded after two-step treatment. The final removal of lignin and hemicellulose arrived 91.5 and 93.8%, respectively. A purity of 94% bamboo cellulose was obtained with only 3.1% cellulose loss. The method proposed in this study can not only improve the paper pulping process but also provide an important technology for the preparation of dissolving pulp in the textile industry due to the effective separation of long fibers.