Cellulose nanocrystals isolated from sugarcane bagasse using the formic/peroxyformic acid process: Structural, chemical, and thermal properties

2.1 Materials

Sugarcane bagasse (SCB) materials were gathered from the Viet Nam National University − Linh Trung − Thu Duc metropolitan region in Ho Chi Minh City. Raw bagasse was first yellowish-white, then the stern section of the sugarcane stalk was removed, rinsed with water, dried in the sun, and then crushed into a fine powder. The compositional analysis of SCB was performed with the NREL protocol to determine structural carbohydrates and lignin (Sluiter et al., 2008). Sodium hydroxide (NaOH, 96 %), formic acid (HCOOH, 90 %), hydrogen peroxide (H₂O₂, 30 %), and sulfuric acid (H₂SO₄, 98 %) were supplied from Xilong Scientific Co. (China). All of these are commercial compounds that were used directly.

2.2 Isolation of cellulose

The three basic procedures in isolating cellulose from SCB were HCOOH acid treatment, peroxyformic acid (PFA) treatment, and bleaching with a NaOH/H₂O₂ mixture. First, the SCB powder was submerged in distilled water and boiled at 100 °C for 2 h to eliminate contaminants. The resulting mixture was filtered and desiccated for 8 h at 70 °C. The SCB powder was homogeneously blended with a 90 % HCOOH solution at a 1:15 ratio (SCB powder to 90 % HCOOH acid volume) at 80 °C for 120 min. The sludge was filtered, rinsed with hot water many times, and desiccated at 80 °C for 6 h. A solution of PFA, comprising 90 % HCOOH, 4 % H₂O₂, and 6 % H₂O, was then added to SCB powder. The PFA treatment was kept at 80 °C for 2 h. The solid product was filtered, extensively washed with distilled water, and dried at 80 °C for 12 h. Finally, the PFA-treated sample was bleached with a NaOH/H₂O₂ mixture. A 1 M NaOH solution was dropped into a PFA-treated suspension (4 wt%) until the pH reached 11. After that, 30 % H₂O₂, equal to 40 % of the solid mass in the rest, was slowly added. The mixture was then agitated for 1 h at 80 °C.

2.3 Production of cellulose nanocrystals (CNCs)

SCB cellulose was hydrolyzed with a 64 % sulfuric acid solution at 45⁰C for 30 min to make CNCs. The ratio of solid mass to acid volume was 1:15. The hydrolysis was quenched by adding a large amount of water (about 10 times the volume of the acid solution used) after the designed reaction time was reached. The diluted suspension was centrifuged at 6,000 rpm for 10 min, twice with distilled water and thrice with ethanol. After drying out at 80⁰C until the mass was unchanged, white powder CNCs were taken. The CNCs yield was calculated based on the initial weight of the raw SCB (Islam et al., 2017).

2.4 Characterization

3.1 Fourier transform infrared (FTIR) analysis

Fig. 1 depicts the FTIR spectra of natural bagasse, acid-treated, PFA-treated, bleached, and CNCs samples in the wavenumber range 4000 to 2000 cm⁻¹ and 2000 to 400 cm⁻¹, respectively. The spectra revealed the presence of functional groups in the structure of natural fibers. Spectral peaks and bands with wavenumbers in the 1270–1020 cm⁻¹ range were characteristic vibrations of the polysaccharide compound's C-O-C group (pyranose ring ether bond) (Mondragon et al., 2014; Lamaming et al., 2015). A broad spectral band at 3418 cm⁻¹, which corresponded to the region of stretching vibration of the hydroxyl (–OH) group, characterized the hydrophilic properties of plant fibers. The spectral peak at 2918 cm⁻¹ represented the stretching vibration of the C–H bond, presented in most organic components such as cellulose, hemicellulose, and lignin (Oun and Rhim, 2015). The raw bagasse sample's spectral peak at 1739 cm⁻¹ indicated the presence of an acetyl group, ester group, or carboxyl group of ferulic acid and p-coumaric acid in the hemicellulose component (Qua et al., 2011; Jiang and Hsieh, 2015; Lamaming et al., 2015). When HCOOH and PFA acid were added to SCB at 80 °C, they broke down the hemicellulose parts of SCB. In the meantime, the lignin component was flattened and moved to the surface, making the absorption peak at 1739 cm⁻¹ stronger in the PFA-treated sample than in the raw SCB sample.

Close

Fig. 1

FTIR spectra of SCB at different stages of chemical treatment and CNCs.

Export to PPT

In all spectra, the peak at 1645 cm⁻¹ corresponded to the O–H bending of water adsorbed in the cellulose fiber (Mondragon et al., 2014; Li et al., 2016). The 1459 cm⁻¹ peak was representative of the deformation vibrations of CH₂ groups. The asymmetric stretching vibration of the C–H bond was visible at 1330 cm⁻¹. The peaks in the region 1200–950 cm⁻¹ were caused by C-O bond stretching vibrations. The −C-O-C group vibration of the pyranose ring produced a prominent band at 1047 cm⁻¹ (Arman et al., 2024). Comparing samples before and after bleaching revealed that the intensity of this band increased, indicating an increase in pure cellulose content. The peak intensity at 890 cm⁻¹ increased with each processing step in the samples' FTIR spectra. This peak was attributed to the glycosidic connections of the cellulose glucose ring (Lu et al., 2015).

After the PFA treatment, FTIR results showed that most of the hemicellulose and part of the lignin were removed. Lignin was one of the most abundant components in fiber and was the most challenging element to remove from natural fibers. Its presence hindered the −cellulose component's bonding surface. The fiber must then undergo a “bleaching” procedure, also known as delignification, to remove most of the lignin required to treat the fibers' component residues (Alemdar and Sain, 2008; Jonoobi et al., 2009). The FTIR study of the bleached and raw bagasse samples showed that the peak at 1739 cm⁻¹ had disappeared from the bleached sample, demonstrating that the bleaching process to eliminate hemicellulose and lignin was successful (Arman et al., 2024). The spectral peaks in the wavenumber range of approximately 1640–1650 cm⁻¹ were associated with moisture adsorption, which forms intermolecular hydrogen bonds between plant fiber constituents and airborne water molecules (Le Troedec et al., 2008).

In addition, the peaks at 1061 cm⁻¹ and 890 cm⁻¹ in the CNCs sample displayed strong C-O-C stretching vibrations, suggesting the presence of a pyranose ring and glycoside linkage. These results indicated that the CNCs are made up of pure cellulose. The hydrogen bonds in the cellulose molecule were responsible for most of the changes in the FTIR spectrum, meaning that during hydrolysis, the acid attack fractured the hydrogen bonds, drastically altering the number of intermolecular hydrogen bonds in cellulose molecules. The results demonstrated that the characteristic peak of the –OH vibration became more pointed and substantially increased in intensity (Laokul et al., 2015; Arman et al., 2024).

3.2 X-ray diffraction analysis (XRD)

XRD patterns were used to examine cellulose's crystalline structure. Fig. 2 shows the XRD patterns of the SCBs before and after they were treated with chemicals. All of the diffractograms showed two peaks at about 22⁰ and 15⁰. These peaks were likely related to the crystalline structure of cellulose I (French, 2014).

Close

Fig. 2

XRD pattern of SCB at various stages of treatment.

Export to PPT

Individual crystalline peaks were taken from the diffraction intensity profiles using the peak deconvolution method (Garvey et al., 2005; Park et al., 2010; Xing et al., 2018), as shown in Fig. S1. Peak deconvolution was the most conservative way to measure the crystallinity of cellulose with XRD analysis. Curve fitting required specific hypotheses, such as the shape and number of points. The Gaussian, Lorentzian, and Voigt functions were typically employed for deconvolution (Park et al., 2010). All of the peaks in this study fit with Gaussian functions, and the broad peaks at about 20⁰ and 40⁰ were thought to be from the amorphous input. An amorphous region was often seen in cellulose's 2θ range of 16–20⁰ (Park et al., 2010; French and Santiago Cintrón, 2013; Xing et al., 2018). This study also linked a broad peak around 35⁰ to the crystallinity component to match experimental results. As shown in Table 1, the CrI and the crystalline domain size were enhanced for the SCB treated with formic/peroxyformic acid.

Because hydroxyl groups formed H-bonds between and within molecules, cellulose had ordered crystalline arrangements (Chirayil et al., 2014). The H bonding made it hard for the cellulosic chains to move around quickly and made the chains line up close to each other in an orderly way, which made them more likely to crystallize. In raw SCB, these crystalline domains were surrounded by amorphous parts like hemicelluloses, lignin, and pectin, which makes the SCB less crystalline. Combined with FTIR results, XRD patterns indicate that the crystallinity improved due to the effective removal of non-crystalline components using a series of chemical treatments (Abdul Rahman et al., 2017; Perumal et al., 2022b; Raza et al., 2022).

Acid and PFA treatments dissolved amorphous hemicelluloses and lignin, increasing CrI from 32.1 % to 42.0 %. Bleaching removed the remaining amorphous components, resulting in a CrI of 57.8 %. With their removal, a large number of crystalline domains can be realigned. The remaining underdeveloped regions were hazardous to acid attacks and exposed to unpredictable orientations. Hydronium ions (H₃O⁺) penetrated amorphous regions and promoted the hydrolysis of cellulose's glycosidic connections, resulting in the liberation of individual crystallites (de Souza Lima and Borsali, 2004; Park et al., 2010; Collazo-Bigliardi et al., 2018; Doh et al., 2020). After acid hydrolysis, a prominent and robust peak appeared in the XRD pattern, indicating a considerable 70.8 % increase in crystallinity. A higher crystallinity index was associated with increased stiffness of cellulose nanocrystals. As a result, when high cellulose nanocrystals were used as a reinforcement ingredient, the composite materials had better mechanical properties (Collazo-Bigliardi et al., 2018; Naduparambath et al., 2018; Melikoğlu et al., 2019).

3.3 Cellulose nanocrystals morphologies

Fig. 3 depicts digital photographs and SEM images of the SCB at various stages of treatment. After treatment with acid, SCB's color changed from light yellow to brown to a distinct white after treatment with PFA and bleaching. The apparent color changes were predominantly caused by eliminating hemicellulose, lignin, and other non-cellulosic components from SCB. Raw SCB appeared as long, rigid tubes (Fig. 3a), and longitudinal examination of SCB revealed that the stems were composed of many aligned fibers. There is a correlation between the presence of cementing materials, including lignin, hemicelluloses, waxes, and pectin, and the interconnection of fibers in the stem. SCB was treated in an acidic environment to dissolve hemicellulose, wax, and pectin partially. Rest and lignin swelled in the acidic environment. They migrated toward the surface, altering the surface morphology of acid-treated SCB (Fig. 3b). The sample color at this stage was a deeper reddish brown than the original because the phenol derivatives found in lignin absorb visible light. The PFA-treated stage mellowed the fiber by removing most of the hemicellulose and a small amount of lignin. Due to the oxidation effects of PFA and the cracking of polysaccharides, particularly hemicellulose and lignin, the morphology of PFA-treated SCB was substantially rougher than that of untreated SCB (Fig. 3c). SCB's hemicellulose, wax, and pectin preponderance were extracted through acid and PFA processing. Lignin, however, remained external to the fiber. Therefore, additional bleaching was necessary to remove lignin and acquire pure cellulose.

Close

Fig. 3

(a-e) Digital photographs and SEM images of the SCB at different stages of treatment; (f) EDX spectra of CNCs.

Export to PPT

The PFA-treated SCB was bleached in a NaOH/H₂O₂ mixture. The presence of NaOH accelerated the breakdown of H₂O₂ into H⁺ and perhydroxyl ions (HO₂^–). These ions attacked the double bonds on the chain and aromatic ring derivatives in lignin, breaking the outer lignin structure and releasing the cellulose component inside the SCB. After bleaching, the isolated cellulose fibers had a smoother surface and smaller diameter than the untreated SCB (Fig. 3d). It is known that cellulose fibers consist of cellulose crystal fibers and amorphous cellulose. The amorphous sections of bleached SCB were removed using sulfuric acid to obtain CNCs. Hydronium ions (H⁺) liberated by sulfuric acid degrade glycoside and ether bonds in the cellulose molecular chain of the amorphous region during this procedure. Therefore, the cellulose fiber's original structure was disrupted and fragmented into numerous fine ones (Fig. 3e). Fig. 3f demonstrates the EDX spectrum of CNCs, which confirms this material's high purity when only C and O were present.

The morphology of the three-dimensional images of dilute bleached SCB and CNCs suspensions were analyzed using TEM. As depicted in Fig. 4 (a,b), the fibers in the final bleached SCB suspension were nanoscale in diameter and had a high aspect ratio. The image displayed lengthy fibrils with diameters between 30 and 35 nm. The majority of fibers had an aspect ratio between 40 and 50. Fig. 4 (c, d) illustrates the TEM image of the CNCs obtained via acid hydrolysis. After treatment with acid hydrolysis, it was evident that the amorphous phase in cellulose nanofibrils had been effectively removed, resulting in a discernible length reduction. The CNCs had a diameter of 23 ± 2.7 nm, a length of 610 ± 18 nm, and an aspect ratio of 26.5 ± 6.7. This aspect ratio value was comparable to or even more significant than that reported in the scientific literature for CNCs isolated from sugarcane bagasse (Oliveira et al., 2016; Leão et al., 2017; Ren et al., 2023) (Table 2). The aspect ratio is crucial in determining the reinforcing capacity of nanofibers. The higher aspect ratio of CNCs suggested they could be used as a reinforcing filler in the film, increasing the biocomposites' mechanical strength (El Achaby et al., 2017). In comparison to other CNCs sources, the aspect ratios obtained in this study were significantly greater than those isolated from mengkuang leaves (10–20) (Sheltami et al., 2012), raw cotton linter (19) (Morais et al., 2013), rice straw (8.8–10.5), grape skin (5–10) (Hsieh, 2013), cotton linter (11–12) (Roohani et al., 2008), and waste cotton cloth (17 ± 15) (Wang et al., 2017), but lower than cellulose nanofibers derived from areca nut husk fibers (Chandra et al., 2016).

Close

Fig. 4

TEM image of (a, b) bleached SCB and (c, d) CNCs.

Export to PPT

In our study, the compositional analysis of SCB was observed to be 45 ± 1 % cellulose, 27 ± 2 % hemicellulose, lignin 20 ± 2 % with lesser quantities of ash (3–4 %), and extractives (1–2 %). The yield of CNCs extracted from SCB ranged from 37 % to 39 %, which was comparable to or even more significant than the yields obtained using alkaline and sodium chlorite treatments (Yue et al., 2015) or ultrasonic-assisted TEMPO-mediated oxidation and one-step APS oxidation method (Zhang et al., 2016). The difference in yield could be attributed to using an isolated cellulose process and acid hydrolysis conditions. This yield was comparable to Mengkuang leaves (28 %) and Okra fibers (30 %) (Sheltami et al., 2012; Fortunati et al., 2013). In contrast, the percentage yield of CNCs from SCB was greater than that of CNCs from barley straw and husk (12 %) (Fortunati et al., 2016), old corrugated container fiber (23.98 %) (Tang et al., 2015), sugarcane pith (23.84 ± 1.28 %) (Ren et al., 2023), and areca nut fibers (35 ± 0.2 %) (Ramesh and Radhakrishnan, 2022), and lower than the CNCs content of grape promace (38–46 %) and corn cob (65.33 ± 7.57 %) (Sartika et al., 2023). It indicated that both the cellulose source and isolation conditions influence the yield of CNCs. Utilizing mechanical cellulose pretreatment, acid hydrolysis with controlled hydrolysis conditions such as acid concentration, hydrolysis duration, and temperature, and post-treatment, including homogenization and ultrasonication, could have resulted in a higher yield of CNCs (Li et al., 2014).

3.4 DLS and zeta potential

The zeta potential of the CNCs suspension was −62.6 ± 2.3 mV (Fig. 5b). The negative charge was due to the addition of sulfate groups to the surface of CNCs during H₂SO₄ hydrolysis. This demonstrated that the treatment conditions induced the dispersion of CNCs in aqueous suspension, which was electrostatically stabilized by repulsive forces. The hydrodynamic size distribution of CNCs calculated using DLS was about 300.1 nm (Fig. 5a). The DLS analysis revealed a sharp distribution profile for the CNCs, indicating their uniform size. Similarly, recent research found that the CNCs isolated from SCB had a sharp and narrow distribution profile, suggesting that the CNCs crystals were more uniform in size and may have created a homogenous dispersion in water (Verma et al., 2021). Pandi et al. (2021) synthesized CNCs from cotton using ultrasound-assisted acid hydrolysis. The DLS research showed that the average hydrodynamic size of CNC was about 221 nm (Pandi et al., 2021).

Close

Fig. 5

Hydrodynamic size analysis (a) of CNCs by DLS and zeta potential (b) of CNCs.

Export to PPT

3.5 Thermogravimetric analysis

To compare degradation characteristics at different processing stages, untreated SCB, acid-treated SCB, PFA-treated SCB, bleached SCB, and CNCs were thermogravimetrically analyzed. Table S1 compares SCB weight loss, maximum degradation temperature, and char yield at different treatment stages. The TG curves (Fig. 6a) depict the weight loss of samples between 25⁰C and 800⁰C. Due to various non-cellulosic and cellulosic components with varying decomposition temperatures, the thermal degradation of unprocessed SCB was a multistage process. Initial weight loss began at 39 °C and continued to about 120⁰C, caused by the evaporation of loosely bound surface moisture and intermolecular hydrogen-bonded chemisorbed water (Dai et al., 2018; Prakash et al., 2018). The subsequent stage of degradation was initiated by the thermal depolymerization of hemicelluloses and the disruption of glycosidic bonds in cellulose at 221.9⁰C. Lignin decomposed over a broader range of temperatures than cellulose and hemicelluloses (Chirayil et al., 2014).

Close

Fig. 6

TGA (a) and DTG (b) of the SCB at different stages of treatment.

Export to PPT

Consequently, the third stage of decomposition was most likely associated with lignin, the principal component of a cementing substance. The DTG of unprocessed SCB revealed two broad peaks, one at 242.4⁰C and the other at approximately 386.7⁰C, and continued to 400⁰C. The broader range of lignin breakdown observed may be attributed to the different thermal stabilities and cleavage temperatures of the oxygen functional groups in its structure (Brebu and Vasile, 2010).

TGA analysis of acid and PFA-treated samples revealed weight loss at temperatures below 150⁰C. At 80⁰C, the hemicellulose components of SCB dissolved when exposed to HCOOH acid and HCOOH/H₂O₂ acid. The lignin component was simultaneously swollen and migrated to the fiber surface. Due to lignin's hydrophobic properties, its migration to the material's surface reduced its capacity to absorb moisture. TGA results revealed a decrease in adsorbed moisture in two samples compared to the unprocessed SCB. In addition, the presence of flame-retardant lignin on the surface of two samples was responsible for the increase in the onset temperature of degradation of these materials.The degradation of acid-treated samples began at 367.3⁰C and reached its maximum rate at 397.3⁰C. The degradation of PFA-treated samples began at 360.8⁰C and reached its maximum rate at 390.5⁰C. After being treated with PFA, the SCB was bleached in NaOH/ H₂O₂ mixture. In the presence of NaOH, H₂O₂ broke down into H⁺ and perhydroxyl ions (HO₂^–). These ions attacked the double bonds on lignin's chain and aromatic ring derivatives, destroyed the outer lignin structure, and liberated the cellulose component within the SCB.

The DTG curve for bleached SCB (Fig. 6b) revealed that significant weight loss occurred between 360⁰C and 420⁰C, with the maximum degradation peak corresponding to cellulose degradation occurring at 395.8⁰C. The highest cellulose degradation peak for CNCs occurred at 250.0⁰C (Fig. 6b). The thermal stability of CNCs was lower than untreated and bleached SCB. This can be attributed to replacing –OH groups of cellulose with O-SO₃H groups during H₂SO₄ hydrolysis (Li et al., 2024a), the increased surface area of CNCs during thermal heating (Hemmati et al., 2018; Naduparambath et al., 2018; Melikoğlu et al., 2019; Singh et al., 2023), and the increased number of free ends in CNCs chains due to their reduced molecular weight (Kallel et al., 2016b; Luzi et al., 2019). Nevertheless, the residue content of CNCs surpassed that of the two samples as a result of the crystalline structure of CNCs (George et al., 2011) and the flame-retardant properties of sulfate ester groups formed during acid hydrolysis (Jiang and Hsieh, 2013; Li et al., 2024a). Similar pyrolysis behavior was reported in CNCs derived from several sources, including garlic straw (Kallel et al., 2016a), sunn hemp (Mahur et al., 2023), corn cob (Sartika et al., 2023), sugarcane pith (Ren et al., 2023), and walnut shells (Arbelaiz and Orue, 2023).

3.6 Raman spectroscopy analysis

Raman spectroscopy correlates the polarizability of electrons in the molecular bond and provides information about the material's vibrational modes. Fig. 7a depicts the Raman spectra of the bleached SCB and CNCs; the obtained band pattern was comparable to that described in the literature (Li and Renneckar, 2011; Agarwal, 2014). The region between 250 and 1600 cm⁻¹ (Fig. 7b) was dominated by bands resulting from the COC, CH₂, CH, and OH groups' vibrations and the probable vibrations of CH-O and OH-O hydrogen bonds. 1474–1464, 1377, 1333, 1265/1284, 988, and 988 cm⁻¹ were the most distinctive bands associated with the molecular environment of methylene (–CH₂) and methine (–CH) groups. Broad bands at 1474 cm⁻¹ and 1464 cm⁻¹ were primarily attributed to CH₂ scissoring (bending) deformations and OH bending deformations at C-6. These bands provided information about the ultrastructure of intramolecular and intermolecular bonds within cellulose (Li and Renneckar, 2011).

Close

Fig. 7

Raman spectra of bleached SCB and CNCs (a) in the range of 200–3600 cm⁻¹ and (b–e) within specific band regions.

Export to PPT

The region between 1200 and 1300 cm⁻¹ was sensitive to glycosidic linkage orientation. In this region, a relatively weak band related to cellulose II was visible at 1265 cm⁻¹ (Fig. 7e). The vibrations of the CH₂ and CH groups related to the bands at 988 and 966 cm⁻¹ (Fig. 7d). C–O–C glycosidic linkage deformations were linked to 1118, 1095, 560–576, 517, and 490 cm⁻¹ bands. Ring deformations were visible between 576 and 1147 cm⁻¹ (Fig. 7c, e). The bands at 456 cm⁻¹, 432 cm⁻¹, and 424 cm⁻¹ corresponded to CCC and CCO ring vibrations. Bands at approximately 456 and 432 cm⁻¹ were characteristic of cellulose I, while bands at 424 cm⁻¹ were unique to cellulose II (Schenzel and Fischer, 2001). Agarwal (2014) demonstrated that the cellulose crystallinity index was associated with bands at approximately 380 cm⁻¹ (Agarwal, 2014). Other skeletal deformations typical of cellulose I included bands at 565 and 327 cm⁻¹, while those of cellulose II were at 576 and 351 cm⁻¹ (Fig. 7c). The 897 cm⁻¹ band corresponded to the bending vibration of the methine group at C-1 or the bending vibrations of HCC and HCO at C-6 in the pyranose unit (Schenzel and Fischer, 2001).

3.7 ¹³C NMR analysis

Fig. 8 depicts the spectrum of bleached SCB and CNCs. The region between 60 and 110 ppm contained cellulose carbon ¹³C NMR signals. Two peaks at 72.6 and 73.4 ppm corresponded to cellulose C-2 and C-3 carbon, respectively, while 75.5 ppm corresponded to cellulose C-5 carbon. The peak at 89.3 ppm corresponded to crystalline cellulose carbon C-4, while 105.5 ppm corresponded to cellulose carbon C-1 (Wikberg and Maunu, 2004). The C-6 signal at 65.7 ppm indicated crystalline presence (Vivian Abiaziem et al., 2019). The NMR spectra of the cellulose after chemical treatments (acid treatment, PFA treatment, and bleaching) revealed that most of the carbohydrates were eliminated. Hemicellulose removal was confirmed by the absence of the peak at 102 ppm attributed to the C-1 of the (1 → 4)-β-D-Xylp-2-O-(4-OMe-α-D-GlcpA) units in hemicellulose (Wen et al., 2011). As there were no peaks at 21, 56, or 173 ppm (Fig. S2) corresponding to hemicellulose and lignin (Idström et al., 2016), it can be concluded that the chemical treatments removed the carbons in the methyl and methoxyl groups in lignin as well as the carbons in the carboxyl groups. In conclusion, the spectrum of the two samples revealed that six carbon atoms caused each peak in the cellulose structure.

Close

Fig. 8

¹³C NMR spectra of bleached SCB and CNCs.

Export to PPT