Introduction

Cellulose nitrate (CN) is one of cellulose derivatives with inorganic ester groups and recognized as the first commercialized derivative. Since Braconnot first synthesized it by treating cellulosic materials with concentrated HNO3 in 1832, CN has been employed in various applications because of its unique properties (Saunders and Taylor 1990). The N content (wt.%), i.e., the degree of substitution (DS) with nitrate groups, influences the CN character. CN with an N content of > 12.2%, corresponding to DS = 2.32, is highly flammable and used as a propellant in smokeless gunpowder, dynamite, and rockets. Conversely, CN with an N content of < 12.2% is used in paints, lacquers, and plastics (Saunders and Taylor 1990; Alinat et al. 2014; Gismatulina et al. 2018). CN is also used as an adsorbent for biomolecules. Exploiting their high protein adsorption capacities, its membrane has been used for proteins transfer from polyacrylamide gels to CN membranes in “Western blotting” and the lateral flow immunochromatographic assay (Towbin et al. 1979; Burnette 1981; Fridley et al. 2013).

The morphologies of CN materials, including films, fibers, and particles, also influence their properties, such as combustion behavior and thermal stability. A material with sub-micron-scale CN particles displayed a higher burn rate than that of a bulk film of CN (Zhang and Weeks 2014). Similarly, nano- to sub-micron-scale CN fibers prepared via electrospinning exhibited lower activation energies of thermal decomposition than that of micron-scale CN (Sovizi et al. 2009). These results indicate that the down-sized CN materials can facilitate their properties, mainly because of the high surface area affects.

In our previous study, we reported the preparation and characterization of nitrated cellulose nanofibers (NCNFs) for the first time, as prepared via nitration of mechanically fibrillated cellulose nanofibers (CNFs) (Okada et al. 2021). A dispersion of never-dried CNFs was used as the raw material. Because CNF is immediately hydrolyzed under conventional nitration conditions using a mixture of H2SO4 and HNO3, we have tried the milder reaction condition. The pretreatment of the CNFs with dilute H2SO4 (30 wt.%) was essential in synthesizing the NCNF material in a high yield with a high DS. As a representative condition for preparing the NCNFs, at an H2O content of 50.8%, the highest N content of 13.7%, corresponding to DS = 2.83, and a yield of 28.3% were realized, which was equivalent to propellant-grade NC. The diameters of the NCNFs were approximately 30–50 nm. When they burned, the pressure increase rate and burning rate were 2- and 3.5-fold higher, respectively, than that of conventional CN, whereas their thermal stabilities were similar. NCNF is, therefore, expected to a novel energetic ingredient.

In this study, we investigated a method of nitrating dried CNFs for the first time, and a milder approach compared to the H2SO4/HNO3 system was developed. To inhibit acid hydrolysis of the CNFs, a weak acid was used instead of H2SO4. A mixture of acetic acid (AcOH) and acetic anhydride (Ac2O) employed. Because moisture content in the fiber is known to influence on reactivity (Iwamoto et al. 2019) and regioselectivity (Beaumont et al. 2021a) of esterification of cellulosic fibers, the nitration reaction process of dried CNF might be different from that of wet CNF. Therefore, the chemical structure of nitrated CNF was analyzed by various methods. Fourier transform infrared (FTIR) spectroscopy was used to determine the N content and DS of the nitrate groups, and solid- and gel-state nuclear magnetic resonance (NMR) spectroscopy revealed the crystalline structure and surface substitution of the produced NCNF material. Scanning electron microscopy (SEM) and specific surface area measurement were used to examine the structures of the NCNFs, and the thermal properties of the NCNF material were evaluated using differential scanning calorimetry (DSC).

Experimental

Materials

Cellulose powder derived from cotton (40–100 mesh, Toyo Roshi Kaisha, Ltd., Tokyo, Japan) and other chemicals (FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) were purchased and used as received.

Preparation of the CNFs via disk milling

Disk milling of the cellulose powder was conducted according to the reported method (Iwamoto et al. 2014). The cellulose powder was immersed in distilled H2O at a concentration of 5 wt.% overnight. The suspension was then ground using a disk mill (MKCA6-2, Masuko Sangyo Co., Saitama, Japan) equipped with two non-porous grinding stones. The rotational speed of the disk was 1800 rpm, and the cellulose suspension was repeatedly passed through the disks while the distance between them was decreased. The fibrillated cellulose suspension was centrifuged (14,400×g) and dispersed in tert-butyl alcohol (t-BuOH) several times to remove H2O from the suspension. The fibrillated cellulose in t-BuOH was then freeze-dried to yield the dried CNF sample as a raw material for use in subsequent surface nitration.

Nitration using the AcOH/Ac2O/HNO3 system

A mixture comprising arbitrary amounts of AcOH (5.0 mL, 87.3 mmol), Ac2O (5.0 mL, 52.9 mmol), and 60% HNO3 (1.0 mL, 21.9 mmol or 2.5 mL, 54.8 mmol) was stirred at 0 °C for 10 min. The mixture was then dropped onto 1.0 g of CNFs (6.2 mmol) and stirred at 0 °C. After the reaction, the mixture was diluted with distilled H2O (50 mL) and filtered. The precipitate was dispersed in 30 mL of aqueous 5% NaHCO3 solution and stirred at 100 ºC for 10 min to completely neutralize the dispersion. The precipitate was then collected via filtration and washed with hot H2O. The collected NCNFs were subjected to solvent exchange with t-BuOH and subsequently freeze-dried to prevent the aggregation of NCNFs during drying procedure. The dried NCNFs were then subjected to chemical and morphological analyses.

Measurements

FTIR spectroscopy

FTIR spectroscopy was performed using a Frontier FTIR spectrometer (PerkinElmer, Inc., Waltham, MA, USA) equipped with diamond/ZnSe attenuated total reflectance, and all spectra were processed using the PerkinElmer Spectrum IR software (version 10.6.1). The DS of nitrate groups was estimated according to a previously reported method (Okada et al. 2021). Before calculating the degrees of nitration, all spectra were normalized using the absorption peaks at approximately 1370 cm–1 derived from the C–H bending of cellulose (Kondo and Sawatari 1996; Colom and Carrillo 2002), whereas the background absorption at 1800 cm–1 was set to 0 A. Subsequently, absorption peak areas in the ranges 1190–930 (Acellulose), 900–770 (Anitrate 1), 1310–1220 (Anitrate 2), and 1710–1525 cm–1 (Anitrate 3) were observed. The measurement was perfomed 3 times in total, and the average of each peak area was determined. The relationship between Acellulose and Anitrate was plotted using 4 CN materials with different N contents, as determined via titration, and their linear fittings were calculated (Fig. S1). The N contents of the NCNF materials were then estimated using the fittings.

NMR spectroscopy

NMR spectroscopy was conducted using a 400 MHz DD2 NMR spectrometer (Agilent Technologies, Santa Clara, CA, USA), which was controlled using VnmrJ3.2 software (Agilent Technologies), and the data were processed using Mnova v14.2.2 (Mestrelab Research S.L., Santiago de Compostela, Spain). Solid-state NMR studies were performed with the 4 mm T3 solid probe HXY MAS Solids Probe 400 MHz WB (Agilent Technologies). Each sample was packed into a ZrO2 rotor with a diameter of 4.0 mm and spun at 10 kHz. Cross-polarization magic angle spinning (CP/MAS) 13C NMR spectra were recorded using the Agilent standard pulse sequence ‘Tangent CP’ with 1H and 13C 90° pulses of 3.5 and 2.0 μs, respectively, and a contact time of 7.0 ms. Scanning was repeated 10 000 times with a recycle decay of 5.0 s, and adamantane was used as an external reference, with its signal in the lower field adjusted to 38.5 ppm. Gel-state NMR spectroscopy was conducted using a 400 MHz NMR spectrometer (Agilent Technologies) equipped with a One probe. The CNF after nitration (~ 30 mg) was swelled in 700 μL dimethyl sulfoxide-d6 (DMSO-d6), and the obtained dispersion was packed in a φ 5 mm glass tube. The 1H and 13C-1H heteronuclear single quantum coherence (HSQC) NMR spectra were acquired using the Agilent standard pulse sequences ‘Proton’ and ‘Gradient HSQCAD,’ respectively. The central DMSO peaks were used as internal references (δC/δH: 39.5/2.49 ppm), and the NMR spectra of the CNFs were obtained via gel-state NMR spectroscopy (Saito et al. 2020).

SEM

Dried samples were placed on conductive tape and coated with Os using Os coater (Neoc-ST, Maiwafosis, Co., Tokyo, Japan) prior to observation. The time for Os vapor deposition was for 10 s, and the deposition rate was 0.5 nm/s rate. Therefore, the samples were coated with an approximately 5 nm-thick layer. The morphologies of the CNFs and NCNFs were observed using an S-4800 field-emission scanning electron microscope (Hitachi High-Tech Corp., Tokyo, Japan) operated at an accelerating voltage of 1.0 kV. To evaluate the fiber diameter, more than 50 fibers were observed for each CNF sample by using SEM data manager software (Hitachi High-Tech Corp., Tokyo, Japan). The fiber diameters were estimated by subtracting the thickness of the Os coating layer (10 nm) from the fiber widths.

Measurement of the specific surface areas of the CNFs and NCNFs

The specific surface areas of the CNFs and NCNFs were determined via Brunauer–Emmett–Teller (BET) analysis. As a pretreatment, samples were degassed under reduced pressure at 50 °C for 6 h using a vacuum degasser (BELPREP VAC II, MicrotracBEL Corp., Osaka, Japan). N2 adsorption isotherms were measured at −196 °C using a BELSORP MAX (MicrotracBEL), and the specific surface areas (SBET) were calculated based on the BET plots using the analysis software BELMASTER (MicrotracBEL).

DSC

The thermal stabilities of the NCNFs were evaluated via thermal analysis using a Pyris 1 DSC (PerkinElmer). NCNFs (0.5 mg) were packed into a 15 μL Al pressure vessel and heated from 50 to 300 °C at 10 °C/min.

Results and discussion

Nitration of the CNFs under mild reaction conditions

CN is synthesized industrially using an acidic mixture comprising concentrated H2SO4 and HNO3. H2SO4 acts not only as a solvent for cellulose, but also as a catalyst of the dehydration of HNO3 (Scheme 1a) (Li and Frey 2010; Sullivan et al. 2018). However, concentrated H2SO4 also causes severe acid hydrolysis of the CNFs (Okada et al. 2021). Therefore, overcoming the trade-off relationship is necessary to yield a high DS and suppress the acid hydrolysis of cellulose, and reducing the H2SO4 concentration resolved this problem in a previous study (Okada et al. 2021).

Scheme 1
scheme 1

Suggested mechanisms of cellulose nitration

Full size image

Using an alternative catalyst with a milder acidity is also effective in inhibiting acid hydrolysis. Several methods of the nitration of cellulose without using H2SO4 has been reported, e.g., nitration using HNO3 and H3PO4 has been known as a method of preparing highly substituted CN materials (Saunders and Taylor 1990). Nitrating cellulose using fuming HNO3 in an organic solvent (Yamamoto et al. 2006) or ionic liquid (Duan et al. 2022) is another example. An acidic mixture of AcOH, Ac2O, and 50% HNO3 is another system used in cellulose nitration (Barsha 1954; Urbański 1965). Though not used for cellulose, Koufaki et al. reported the successful nitrate esterification of acid-labile purine derivatives via acetyl nitrate using an AcOH/Ac2O/65% HNO3 mixture (Koufaki et al. 2012). As the respective pKa values of H2SO4, HNO3, H3PO4, and AcOH are less than −3, −1.4, 2.1, and 4.8 (Gutowski and Dixon 2006; Morales-delaRosa et al. 2014), the nitration reaction with AcOH/Ac2O/HNO3 may be the mildest of those described. Thus, we conducted the nitration of the CNFs under these reaction conditions.

The results of nitrate esterification of the CNFs are shown in Table 1. When 60% HNO3 was reacted with the CNFs by adding in an amount equivalent to 1.2 times the mole number of OH groups of cellulose, the product mass is > 100% relative to that of the raw CNF (Table 1, NCNF i–iii). In addition, the yield increased as the reaction time increased, whereas further increasing the amount of HNO3 used resulted in a reduced yield, likely due to the hydrolysis of the CNFs (Table 1, NCNF iv). The mass of the CNF material collected after using the AcOH and Ac2O mixture (Table 1, control) was lower than that of the raw CNF, which may be due to losses during purification. Thus, the increased masses of NCNF i-iii were due to nitration.

Table 1 Results of the nitration of the CNF using the AcOH/Ac2O/HNO3 system
Full size table

Characterizations of the chemical structures of the NCNFs

FTIR spectroscopy

Figure 1 shows the FTIR spectra of the CNF and NCNFs. In the FTIR spectra of NCNF (i–iii) (Figs. 1b–d), absorption bands assigned to cellulose are observed at 3340 (ν O–H), 2900 (ν C–H), 1160 (ν C–O–C), and 1060 cm–1 (ν C–O at C3, ν C–C) (Oh et al. 2005). The FTIR spectra of the CNFs before and after treatment with AcOH/Ac2O mixture are in accordance (Figs. 1a, e), indicating that AcOH and Ac2O did not react with cellulose without HNO3. Concurrently, absorption bands representing the nitrate groups are observed at 1650 (ν NO2), 1280 (ν NO2), and 830 cm–1 (ν O–NO2) (Moore and McGrane 2003; Fernández de la Ossa et al. 2011). On the other hand, a small absorption peaks around 1750 cm−1 (ν C=O) were detected in Fig. 1b–d, indicating acetylation proceeded slightly as a side reaction (Colom et al. 2003). The peak areas of the absorption bands attributed to the nitrate groups increased as the reaction time increased. The N contents of the NCNFs were determined using the FTIR spectra (Okada et al. 2021). The N content of NCNF (iii) was 6.1%, representing the highest nitration, which was obtained using the AcOH/Ac2O/HNO3 system. Based on the N content, the DS of NCNF (iii) was calculated to be 0.88, which is lower than that of the NCNF reported in a previous study (Okada et al. 2021). Though the DS values are not so high, we will show the highly surface specific nitration on the CNF surface in the subsequent paragraphs.

Fig. 1
figure 1

FTIR spectra of a CNF, b NCNF (i), c NCNF (ii), d NCNF (iii), and e CNF treated with a mixture of AcOH and Ac2O

Full size image

Solid-state NMR spectroscopy

The solid-state NMR spectra of the CNF and NCNFs (i–iii) are shown in Fig. 2. The 13C CP/MAS NMR spectrum of the CNF before nitration (Fig. 2a) is consistent with that of Iβ-rich cellulose (Kono et al. 2002). After nitration, the C3 signal at 75 ppm decreased compared to the signals at 68–74 ppm, with two possible causes (Yamamoto et al. 2006). First, the C6 signal at 65 ppm was shifted to 72 ppm after nitration, and thus, the peak area at 68–74 ppm increased. Second, the C3 signal may shift upfield when the hydroxyl group at the C2 carbon is substituted with a nitrate group. The signal at 101 ppm, which is assigned to the C1 carbon in the anhydroglucose unit with the substituted C2-OH (Yamamoto et al. 2006), is observed in the spectra of the NCNFs (Figs. 2b–d). These results showed that nitration proceeded at C2 and C6. Additionally, slight increases in the signals at 78 ppm due to nitrated C3 are observed in the NMR spectra of NCNF (ii) and (iii), suggesting the introduction of NO2 groups at the C3 carbon atoms.

Fig. 2.
figure 2

13C CP/MAS NMR spectra of a the CNF, b NCNF (i), c NCNF (ii), d NCNF (iii), and e the NCNF prepared using the never-dried CNFs and H2SO4-HNO3 mixture

Full size image

The treatments of carbohydrates with nitric acid was reported to cause the oxidation of the primary hydroxy groups (Pigman et al. 1949). Also, nano-sized carboxycellulose materials are obtained after treatments of cellulosic fibers with nitric acid and oxidant (Sharma and Varma 2013, 2014; Sharma et al. 2017). Therefore, we investigated whether the oxidation of hydroxyl group of cellulose occurred. The 13C CP/MAS NMR spectrum of NCNF (iii) in the range of -5–205 ppm is shown in Fig. S2. Any obvious signals were not detected at 170–200 ppm, which is corresponding to carbonyl group (Sharma et al. 2017). This result indicated that the oxidation of hydroxyl group of cellulose negligibly proceeded in the nitration reaction with AcOH/Ac2O/HNO3. Additionally, the signal of acetyl group was not observed in the 13C CP/MAS NMR spectrum, suggesting that acetylation of CNF was minor reaction of this system.

In the 13C CP/MAS NMR spectrum of the CNF (Fig. 2a), the signals representing the C4 carbons of crystalline (C4D) and amorphous cellulose (C4U) are observed at 89 and 82 ppm, respectively. As nitration of the CNFs proceeded, the area of the peak representing C4D decreased, whereas that of the peak representing C4U increased. This indicated that the crystalline cellulose degraded as cellulose nitration proceeded. However, the signals assigned to cellulose Iβ was still detected after nitration for 6 h (Fig. 2d). Thus, NCNF (iii) still contained the crystalline structure of raw CNF.

Figure 2e shows the 13C CP/MAS NMR spectrum of the NCNF prepared using a previously reported method (Okada et al. 2021). The N content and DS of the NCNF are 12.6% and 2.5, respectively. As shown in Fig. 2e, the signals representing cellulose crystal Iβ are not observed, whereas signals assigned to cellulose trinitrate (CTN) I crystals are observed at 101, 78, and 72 ppm (Yamamoto et al. 2006). Therefore, nitrate groups might be introduced into the interiors of cellulose crystals in the NCNF prepared using the H2SO4/HNO3 mixture, altering the crystalline structure of cellulose. The differences in the NMR spectra suggest that the character of the NCNF prepared using AcOH/Ac2O/HNO3 may be quite different from that of the NCNF prepared using H2SO4/HNO3.

Gel-state NMR spectroscopy

NCNFs (i–iii) were well dispersed in DMSO-d6. Then the dispersions of NCNF were subjected to gel-state NMR spectroscopy (Saito et al. 2020). Figure 3 shows the HSQC-NMR spectra of the CNF before and after nitration. In the spectrum of the CNF (Fig. 3a), signals corresponding to cellulose are observed (Saito et al. 2020), except for that representing C5. In the spectrum of NCNF (i) (Fig. 3b), seven novel cross peaks are observed, although signals representing C2 and C3 of cellulose are also observed. In the spectra of NCNF (ii) and (iii) (Figs. 3c, d), all signals differ from those of the raw CNF. These signals were assigned to CTN, based on previous studies (Wu 1980; Yamamoto et al. 2006; Gismatulina et al. 2018). The signals might shift downfield because electron-withdrawing nitrate groups are introduced, and the signal assignments are shown in Table 2. Another cross-peak was detected at δH/δC 2.02/19.97 ppm, which is identified to CH3 of acetyl group (Kim et al. 2008), although its peak area was quite smaller than that of cellulose (Fig. S3).

Fig. 3
figure 3

HSQC-NMR spectra of a the CNF, b NCNF (i), c NCNF (ii), and d NCNF (iii). Red contours represent methyl and methine signals and blue contours are methylene signals

Full size image
Table 2 Assignments of the signals observed in the NMR spectra of the CNF and NCNF (iii)
Full size table

Although the DS of NCNF (iii) is 0.88, the signals of raw cellulose are not observed. In solution-state NMR spectroscopy, detecting the signals of compounds with low molecular mobilities, such as crystalline cellulose, is challenging (Kim and Ralph 2010). Thus, the nitrated moieties in NCNF (i)–(iii) may swell well in DMSO-d6, whereas the molecular mobility of the crystalline cellulose may be low in solvents. Based on the solid- and solution-state NMR studies, it was indicated that NCNF (ii) and (iii) might comprise two regions: highly substituted and crystalline cellulose regions.

Morphologies of the NCNFs

Figure 4 shows the surface morphologies of the CNFs before and after nitration, and the SBET of each sample is shown alongside each image. The SBET values were estimated from the N2 absorption isotherms (Fig. S4) by using the BET method. The fiber morphology of the CNF was retained after nitration, and the diameters of the fibers decreased as nitration proceeded. The distributions of fiber diameter of CNF and NCNF are shown in Fig. S5. Although it is an approximate value because of resolution of SEM image, the diameter of raw CNF was 10–120 nm and that of NCNF (iii) was 10–50 nm. Additionally, the SBET of the NCNF increases as the DS of nitration increases, reaching 258 m2/g, corresponding to the values of chemically-modified CNFs, such as maleic acid-modified CNFs (289 m2/g) (Iwamoto and Endo 2015). This suggests that the surface nitrate groups should promote nanofibrillation.

Fig. 4
figure 4

SEM images of CNF and NCNFs

Full size image

Mechanism of cellulose nitration in the AcOH/Ac2O/HNO3 system

NMR and IR spectroscopy, in addition to SEM, indicated that NCNFs prepared using the AcOH/Ac2O/HNO3 mixture display surfaces comprising CTN and cores comprising intact crystalline cellulose. These structures differ from those of NCNFs prepared using the H2SO4/HNO3 mixture, which exhibit CTN I structures. To explain this result, we focused on the functions of acids during nitration.

When cellulose is treated with the H2SO4/HNO3 mixture, H2SO4 may display two functions: (1) collapse the hydrogen bonding between cellulose molecules, increasing the reactivities of the OH groups of cellulose, and (2) dehydrate HNO3 to form nitronium ions, which are strong electrophilic agents (Scheme 1a) (Sullivan et al. 2018). When the penetration of reaction media into crystalline cellulose and nitration of cellulose occur simultaneously, nitrate groups should be introduced homogeneously throughout the fiber.

During nitration using the AcOH/Ac2O/HNO3 mixture, acetyl nitrate may be produced as an intermediate. Several types of intermediates are proposed in the AcOH/Ac2O/HNO3 nitration system, such as nitronium ions, N2O5, and acetyl nitrate (Bodor and Dewar 1969). Bodor and Dewar (1969) reported that acetyl nitrate is the most plausible intermediate, based on the results of semi-empirical calculations. When acetyl nitrate is an intermediate of cellulose nitration, the nucleophilic reaction of an OH group of cellulose with acetyl nitrate may be the driving reaction (Scheme 1b). As the structure of acetyl nitrate is bulkier than that of the nitronium ion, the penetration of acetyl nitrate into crystalline cellulose is inhibited, and thus, nitration of cellulose may proceed only at the surfaces of the CNFs.

Water molecules confined to the CNF surface has been reported to promote the surface-selective modification of CNF (Beaumont et al. 2021a, b). Therefore, water in the reaction medium could also affect on the surface selectivity of nitration reaction of CNFs.

Thermal properties of the NCNFs

The thermal property such as stability of CN remains a problem, even when it contains a low N content, because decomposition may induce spontaneous combustion (Sun and Qiu 2023). In addition, CN prepared using the AcOH/Ac2O/HNO3 system is reportedly unstable when acetyl nitrate persists (Barsha 1954). Therefore, the thermal properties of NCNF (iii) were investigated, and their DSC thermograms are shown in Fig. 5. Any thermal decomposition peak was not observed below 170 °C, suggesting that residual acid is sufficiently removed. Conversely, an exothermic peak was observed approximately at 201 °C, which is usually not observed in DSC curve of cellulose materials (Szcześniak et al. 2008). Thus, the exothermic peak at 201 °C indicates the thermal decomposition of cellulose nitrate. The values of heat of reaction (ΔH, J/g) of NCNF (iii) were 1.6 kJ/g, which was lower than that of NCNF prepared using H2SO4/HNO3. This might be due to relatively low N contents. The exothermic onset temperatures of decomposition of NCNF (iii) was 177 °C. As the thermal decomposition behavior of CN depends on its N content (Pourmortazavi et al. 2009), the pernitrate cellulose at the surface may be lower the thermal stability of NCNF (iii). Although these results indicated that NCNF (iii) might be not suitable for explosive applications, it was suggested that as a usage for paints, lacquers, and plastics, the NCNFs are stable under ambient temperature.

Fig. 5
figure 5

DSC thermogram of NCF (iii)

Full size image

Conclusion

Dried CNFs were nitrated using the AcOH/Ac2O/HNO3 system, which was milder than the conventional H2SO4/HNO3 system. The main conclusions drawn from the results are as follows:

  1. 1.

    The AcOH/Ac2O/HNO3 system afforded NCNFs with moderate DS values (0.4–0.8) and high collecting yields (> 100%) because of the sufficient suppression of the acid hydrolysis of the CNFs.

  2. 2.

    The NCNFs exhibited cellulose Iβ crystal structures, according to solid-state NMR spectroscopy, and the surface cellulose pernitrate structures were swollen in d-DMSO, according to gel-state NMR spectroscopy, revealing the core (cellulose Iβ)-shell (cellulose pernitrate) structures.

  3. 3.

    The NCNFs displayed morphologies comprising NF structures with decreasing diameters as the DS increased, and the specific surface area reached 250 m2/g.

  4. 4.

    The NCNFs were thermally stable at ≤ 170 °C.

Although the DS of the NCNF prepared using the AcOH/Ac2O/HNO3 system was lower than that observed in a previous study, the obtained material exhibited a unique core–shell structure, with potential for use in novel materials, such as coatings, lacquers, plastics, paints, and membranes.