Facile Post-Carboxymethylation of Cellulose Nanofiber Surfaces for Enhanced Water Dispersibility.

To improve the water dispersibility of cellulose nanofibers without deteriorating the physical properties, it is necessary to develop methods that can selectively modify fiber surfaces. Herein, the reaction conditions for carboxymethylation of the surface of nanofibrillated bacterial cellulose were optimized using chloroacetic acid as an etherification agent. Carboxymethylation in a high-concentration alkaline solution (>5 wt %) in the presence of isopropanol caused the mercerization and carboxymethylation of not only the nanofiber surface but also the cellulose crystals within the nanofiber, resulting in nanofiber swelling and an increase in fiber width. In contrast, with a dilute alkaline aqueous solution (3 wt %), the nanofiber surface was successfully carboxymethylated without changing the inner structure. Furthermore, the morphology was not affected by the carboxymethylation reaction, and no fiber swelling occurred under these reaction conditions. When the substitution reaction proceeded only on the nanofiber surface, the maximum degree of substitution (i.e., the average number of carboxymethyl groups substituted per anhydroglucose residue in cellulose) was 0.091. After surface modification, the nanofibers became more negatively charged, which improved the dispersibility in water through electrostatic repulsion, resulting in a drastic increase in the transparency of the nanofiber dispersion. This method provides a general approach for the surface modification of cellulose nanofibers to increase water dispersibility.

Introduction

Cellulose nanofibers (CNFs) have attracted great interest for a wide range of applications because they are renewable, biodegradable, and have excellent mechanical and optical properties, high aspect ratios, and large specific surface areas.[1−3] CNFs are usually produced from wood pulp by intensive mechanical treatment, such as high-pressure homogenization.[4,5] Because of the high energy consumption associated with mechanical treatment, wood pulp is commonly chemically pretreated to aid in the breakup of cellulose fibrils. For example, 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) oxidization,[6,7] carboxymethylation,[7] phosphorylation,[8,9] and sulfonation[10] have been investigated as chemical pretreatment methods for CNF production. These pretreatment methods introduce ionic groups onto the pulp fibers, thus reducing adhesion between the fibrils and facilitating the breakup of fibrils during mechanical treatment. Furthermore, pretreatment can enhance the dispersibility of CNFs in water because the surface hydroxyl groups of the obtained CNFs are partially substituted by ionic functional groups.[6−12]

As an alternative source of CNFs, bacterial cellulose has attracted increasing interest because of its remarkable physical and chemical properties, biocompatibility, and ease of production methods.[13−15] Certain Gram-negative bacterial genera have been reported to produce bacterial cellulose, with the most common cellulose-producing strains belonging to the genus Gluconacetobacter.[13] The bacterial cellulose production process involves two coupled steps: polymerization and crystallization. Cellulose chains formed from glucose in the bacterial cytoplasm are secreted extracellularly. The chains crystallize as microfibrils, and subsequent consolidation of a certain number of microfibrils forms a gel-like membrane (pellicle) with a highly pure three-dimensional porous network of entangled nanoribbons of ∼100 nm in width.[16] The pellicle has low fluidity, moldability, and miscibility owing to its highly developed three-dimensional structure. Recently, water-dispersible bacterial cellulose was produced by culturing a cellulose-producing bacterium in a medium under aerobic agitation.[17,18] As the fiber width of the water-dispersible bacterial cellulose (∼20–40 nm) was smaller than that of cellulose obtained under static culture conditions, these dynamic culture conditions were considered to inhibit the consolidation of microfibrils during the formation of cellulose fibrils.[18] Although this water-dispersible bacterial cellulose, called nanofibrillated bacterial cellulose (NFBC), and CNFs have almost identical molecular structures in terms of the hierarchical structure of cellulose chains, their components are different. CNFs, which are generally produced from wood pulp, contain lignin, hemicellulose, and pectin, whereas NFBC is free of these components. In addition, NFBC has very long fibers (>15 μm in length) because it is directly produced by microorganisms without mechanical treatment that disrupts the fiber bundles, whereas CNFs are usually no more than a few micrometers in length.[19]

Substitution of the surface hydroxyl groups of NFBC and CNFs with ionic functional groups is industrially important for controlling nanofiber dispersibility, viscosity, and transparency.[6−12] Commonly, ionic functional groups are introduced into cellulose molecules by substituting hydroxyl groups with carboxymethyl groups.[20] This substitution reaction generally proceeds via the etherification of cellulose with chloroacetic acid (CA) in a mixed solvent of sodium hydroxide aqueous solution and isopropanol.[21] Cellulose etherification involves two steps, namely, the treatment of cellulose with sodium hydroxide to form alkaline cellulose, which is often referred to as mercerization, and the etherification of alkaline cellulose with CA.[22] The formation of alkaline cellulose is necessary to improve the reactivity toward CA. Although isopropanol is not directly involved in etherification, its presence promotes a high concentration of sodium hydroxide around the cellulose fibrils, resulting in efficient etherification.[22−24] Carboxymethylated cellulose becomes water-soluble when the degree of substitution (DS; i.e., the average number of carboxymethyl groups substituted per anhydroglucose residue) reaches 0.4–0.5.[21,25] In general, commercially available sodium carboxymethyl cellulose, which is industrially synthesized from pulp using a similar method, has a DS in the range of 0.5–1.5.

Although cellulose carboxymethylation can be easily achieved in an alkaline solution/isopropanol mixed solvent, it is difficult to selectively carboxymethylate the cellulose molecules on the surfaces of NFBC and CNFs while maintaining a hierarchical structure and fiber morphology. Immersion in a high-concentration sodium hydroxide aqueous solution for the formation of alkaline cellulose causes nanofiber swelling, which allows the carboxymethylation of cellulose molecules to proceed within the nanofibers as well as on the surface.[26] Carboxymethylation within a nanofiber breaks hydrogen bonds between cellulose molecules, leading to a breakdown of the hierarchical nanofiber structure and deterioration of the physical properties. As a way to selectively carboxymethylate only the CNF surface layer, direct carboxymethylation of CNF dispersions with CA has been reported.[27] The surface layer modification in aqueous solution has the advantage of reducing CNF aggregation and preserving its characteristics as a nanofiber. However, the relationship between the carboxymethyl reaction conditions of CNFs and their degree of substitution has not been clarified. In addition, there is no report on the surface modification of NFBC. Therefore, a way to carboxymethylate only the cellulose molecules on NFBC and CNF surfaces is necessary to improve the dispersibility and transparency of nanofiber suspensions.

In this study, the selective carboxymethylation of the surface of NFBC was investigated to improve water dispersibility and transparency (Figure ). The reaction solvent, reaction temperature, reaction time, and CA concentration were optimized to determine the reaction conditions under which the surface cellulose molecules of NFBC could be substituted with carboxymethyl groups without causing structural changes within the nanofibers. In addition, the dispersibility of the carboxymethylated NFBC (CMNFBC) samples in water was evaluated using zeta potential and transmittance measurements and microscopic observations, and the relationship between the properties of CMNFBC and the DS was determined.

Figure 1

Preparation of CMNFBC by reacting neat NFBC dispersed in aqueous sodium hydroxide solution with CA.

Results and Discussion

Optimization of the Reaction Solvent for Preparing CMNFBC

Generally, to prepare carboxymethylcellulose from cellulose pulp, the cellulose fibers are first immersed in a high-concentration sodium hydroxide aqueous solution, which swells the fibers and converts them into alkaline cellulose. Subsequently, isopropanol and CA are added to the slurry to etherify the hydroxyl groups of cellulose.[24−26] In this study, to substitute the surface hydroxyl groups of cellulose fibers with carboxymethyl groups without changing the nanofiber structure, 12 CMNFBC samples (CMNFBC #1–12; Table ) were synthesized from neat NFBC by varying the solvent composition, sodium hydroxide concentration, CA concentration, reaction temperature, and reaction time. The molecular structures and fiber morphologies of the samples were analyzed to determine the optimal reaction conditions.

Table 1

Reaction Parameters for the Preparation of CMNFBC Samples

CMNFBC sample	isopropanol content in the reaction solvent (vol%)a	NaOH concentration (wt%)a	temperature (K)	molar feed ratio of CA/AGUb	reaction time (h)
#1	50	10	333	10	3
#2	25	5	333	10	3
#3	0	3	298	10	24
#4	0	3	333	10	24
#5	0	3	343	10	24
#6	0	3	353	10	24
#7	0	3	363	10	24
#8	0	3	353	2	24
#9	0	3	353	5	24
#10	0	3	353	10	1
#11	0	3	353	10	6
#12	0	3	353	10	12

The total volume of the reaction solvent mixture containing aqueous sodium hydroxide solution and isopropanol was set to 300 mL, and the concentration of NFBC was set to 1.5 wt %.

Molar amount of CA per AGU in NFBC.

First, the carboxymethylation of NFBC was investigated in mixed solvents of sodium hydroxide aqueous solution and isopropanol. CMNFBC #1 and #2 were prepared in 50 vol % isopropanol in 10 wt % sodium hydroxide aqueous solution and 25 vol % isopropanol in 5 wt % sodium hydroxide aqueous solution, respectively, using a reaction temperature, reaction time, and CA concentration of 333 K, 3 h, and 10 times the AGU content in NFBC, respectively. The SSNMR spectra of neat NFBC and CMNFBC #1 and #2 are shown in Figure A. NFBC exhibited the typical SSNMR spectral pattern of native cellulose.[28,29] In contrast, the spectra of CMNFBC #1 and #2 showed carbonyl carbon resonance corresponding to carboxymethyl groups at 188–173 ppm,[30,31] indicating that the hydroxyl groups of NFBC were successfully substituted with carboxymethyl groups. For the carboxymethylated samples, the DS was determined from the integral value of the carboxyl carbon resonance relative to that of the C1 resonance at 110–96 ppm.[29] The DS values for CMNFBC #1 and #2 were 0.322 and 0.634, respectively.

Figure 2

Solid-state 13C NMR spectra (A) and XRD patterns (B) of neat NFBC and CMNFBC #1 and #2. In panel (A), the DS was determined from the integral value of the carbonyl carbon (C8) resonance relative to that of the C1 resonance, which was set to 1. In panel (B), the crystal planes of cellulose I and cellulose II are indicated in blue and red, respectively.

The effect of carboxymethylation on the inner hierarchical structure of NFBC was evaluated by XRD analysis (Figure B). The XRD pattern of neat NFBC showed the characteristic diffraction peaks of native cellulose at 14.8, 16.6, and 22.4°, which correspond to the (11–0), (110), and (200) crystal planes, respectively, of crystalline cellulose I.[31] The crystallinity index, which was determined from the ratio between the peak area corresponding to crystalline cellulose I and the total area,[31] was found to be 63% for NFBC. In addition to the diffraction peaks of cellulose I, the XRD patterns of CMNFBC #1 and #2 exhibited diffraction peaks at 14.4, 19.8, and 22.2° corresponding to the (11–0), (110), and (200) crystal planes, respectively, of crystalline cellulose II.[32] Based on the diffraction intensity, CMNFBC #1 contained fewer crystalline components than CMNFBC #2. However, the crystallinity indices of these samples could not be determined because of the ambiguity caused by the mixture of two crystalline allomorphs and an amorphous region. These findings indicate that the fibers swelled in the reaction solvent during the reaction process, resulting in a crystal transition from cellulose I to cellulose II via the formation of alkaline cellulose. Thus, the mercerization of NFBC occurred in 5 wt % sodium hydroxide solution in the presence of isopropanol.

A comparison of the fiber morphologies of CMNFBC #1 and #2 with that of neat NFBC (Figure and (Figure S1) revealed swelling and excessive carboxymethylation for the CMNFBC samples. The SEM (Figure ) and SPM (Figure S1) images of neat NFBC reveal distinct nanofibers, whereas blurred nanofibers with increased fiber widths were faintly observed in CMNFBC #1. In the case of CMNFBC #2, which had a lower DS (0.322), nanofibers were detected, but fiber swelling and adhesion between adjacent fibers could be observed. Owing to mercerization in the reaction solvents used to prepare CMNFBC #1 and #2, carboxymethylation occurred not only on the nanofiber surface but also within the nanofiber. Therefore, to carboxymethylate the surface of NFBC without changing the fiber structure, it is necessary to use a solvent that causes no mercerization.

Figure 3

SEM images of neat NFBC and CMNFBC #1 and #2.

Optimization of Reaction Temperature for Preparing CMNFBC

Next, CMNFBC #3–7 were prepared at reaction temperatures of 298, 333, 343, 353, and 363 K, respectively, in a 3 wt % sodium hydroxide aqueous solution without isopropanol (Table ) to avoid the swelling and collapse of the nanofiber structure. Because the reaction rate for the carboxymethylation of cellulose is extremely slow in the absence of isopropanol, the CA concentration and the reaction time were set to 10 times the AGU content in NFBC and 24 h, respectively. As shown in Figure , the SSNMR spectra of CMNFBC #3–7 gave DS values of 0.006, 0.035, 0.064, 0.091, and 0.091, respectively (Table ), indicating that some of the cellulose hydroxyl groups were substituted with carboxymethyl groups when the temperature was higher than 333 K, even in the absence of isopropanol at a NaOH concentration of 3 wt %. At temperatures above 333 K, the DS increased as the temperature increased, reaching saturation at 353 K. In addition, in the region of 60–110 ppm, the spectral shapes of the CMNFBC samples were almost identical to that of neat NFBC, indicating that most of the cellulose molecules in the CMNFBC samples were not carboxymethylated.

Figure 4

Solid-state 13C NMR spectra of CMNFBC #3–7. The DS was determined by the integral value of the carbonyl carbon resonance (C8) relative to that of the C1 resonance, which was set to 1.

Table 2

DS, Crystallinity Index, Selective Planar Orientation, and Surface Charge of CMNFBC #3–12 and Neat NFBC

CMNFBC sample	DS (SSNMR)	crystallinity index (XRD) (%)	selective planar orientation (XRD)	surface charge (zeta potential) (mV)
neat NFBC	0	67	0.37	–2.6
#3	0.006	67	0.34	–4.6
#4	0.035	66	0.30	–11.5
#5	0.064	64	0.28	–20.7
#6	0.091	64	0.23	–25.0
#7	0.091	63	0.22	–25.2
#8	0.004	66	0.34	–8.4
#9	0.067	64	0.26	–19.1
#10	0.003	64	0.33	–5.6
#11	0.058	63	0.26	–19.2
#12	0.088	63	0.24	–23.4

The XRD patterns of CMNFBC #3–7 were similar to that of neat NFBC, and no diffraction peaks corresponding to cellulose II were observed (Figure ). The crystallinity indices of these samples ranged from 63 to 67% (Table ), similar to that of NFBC (67%), suggesting that the inner structure of the nanofibers in these CMNFBC samples remained unchanged with carboxymethyl group substitution occurring preferentially on the nanofiber surfaces. As shown in Figure S2, the surface modification of the CMNFBC samples was confirmed by a decrease in the selective planar orientation (i.e., the ratio between the (11–0) and (200) peaks). The selective planar orientation values for CMNFBC #4–7 were 0.30, 0.28, 0.23, and 0.22, respectively, whereas that of neat NFBC was 0.37 (Table ). A similar decrease in the selective planar orientation was previously reported for CNFs and cellulose crystals with chemically modified surfaces.[18,19,31,33] The carboxymethylation of nanofiber surfaces is expected to hinder the formation of interfiber hydrogen bonds, resulting in CMNFBC samples with lower selective planar orientations.

Figure 5

XRD patterns of CMNFBC #3–7. The crystal planes for cellulose I are also indicated.

As shown by the SEM images in Figure , CMNFBC #4–7 maintained the same fiber morphology as neat NFBC, with clearly defined fibers and no obvious aggregation. However, in contrast to an average fiber width of 24.5 ± 1.7 nm for neat NFBC, those of CMNFBC #4–7 were slightly smaller (23.6 ± 2.6, 22.4 ± 2.4, 21.7 ± 2.2, and 21.6 ± 2.0 nm, respectively), decreasing as the DS increased (Figure ). SPM observations also confirmed a similar fiber morphology and a DS-dependent decrease in the average fiber width for the CMNFBC samples (Figure S3). It should be noted that the average fiber width for each sample determined by SEM was approximately 3 nm larger than that determined by SPM because the SEM samples were coated with platinum before observation. From these morphological observations, the average fiber width of the CMNFBC samples decreased as the reaction temperature increased, indicating that a higher reaction temperature promoted carboxymethylation of the nanofibers and that carboxymethyl group substitution proceeded from the surface layer. Because the DS of the samples was saturated at 0.091 (Table ), any further substitution led to the release of carboxymethyl cellulose chains from the nanofibers, resulting in a decrease in fiber width.

Figure 6

SEM images and fiber width distributions (n = 100) of neat NFBC and CMNFBC #4–7.

The zeta potential of neat NFBC was −2.6 mV, whereas those of CMNFBC #3–7 were more negative, varying from −4.6 to −25.1 mV (Table ). The surface charge of the CMNFBC samples decreased as the DS increased, indicating that the carboxymethyl groups were mainly distributed on the nanofiber surfaces. In addition, the negative surface charge was saturated at approximately −25 mV at a DS of 0.091; therefore, the surface hydroxyl groups of nanofibers are considered to be almost completely substituted by carboxymethyl groups when the DS reaches 0.091.

Water Dispersibility of CMNFBC

The dispersion states of CMNFBC #4–7 and neat NFBC in deionized water are shown in Figure A. The dispersibility of the CMNFBC samples in water was obviously enhanced compared with that of neat NFBC. The aqueous suspensions of CMNFBC #4–7 also exhibited higher transparency in the visible light region (350–750 nm; Figure B) than neat NFBC. The transparency of the CMNFBC suspensions improved as the DS increased. CMNFBC #6 and #7, which both had a DS of 0.091, exhibited the highest transparency, with a 10–14% increase in transmittance in the visible region as compared to the neat NFBC suspension (Figure B). Therefore, the substitution of carboxymethyl groups on the NFBC surface increased the negative surface charge, which led to electrostatic repulsion between the nanofibers and improved their dispersibility in water. In addition, the slight decrease in the average fiber width of the CMNFBC samples may suppress the scattering of transmitted light[33] and thus increase transparency.

Figure 7

Dispersion state (A) and transmittance (B) of neat NFBC and CMNFBC #4–7 in water.

Effect of CA Concentration and Reaction Time on Surface Modification

CMNFBC #6 and #7 were almost identical in terms of DS, average fiber width, crystallinity, and dispersibility in water. The only difference in the reaction conditions used to prepare CMNFBC #6 and #7 was the reaction temperature (353 and 363 K, respectively), indicating that the optimal temperature for the surface carboxymethylation of NFBC was above 353 K. Next, in an effort to improve the efficiency of synthesizing CMNFBC samples while maintaining the favorable properties of CMNFBC #6 and #7, the CA concentration and reaction time were optimized.

CMNFBC #8 and #9 were synthesized under the same conditions as CMNFBC #6, except that the CA concentrations were decreased to 2 and 5 times the AGU content in NFBC, respectively. As shown in Figure S4, the carbonyl carbon resonance was very weak in the SSNMR spectrum of CMNFBC #8, whereas that in the spectrum of CMNFBC #9 gave a DS value of 0.067. The decrease in the DS of CMNFBC #8 and #9 as compared to that of CMNFBC #6 confirmed the necessity of using a CA concentration of approximately 10 times the AGU content in NFBC.

Figure shows the SSNMR spectra of CMNFBC #10–12, which were synthesized under the same reaction conditions as CMNFBC #6, except that the reaction times were decreased to 1, 6, and 12 h, respectively. Based on the carbonyl carbon resonances, almost no substitution of NFBC occurred with a reaction time of 1 h, whereas the DS increased to 0.058 and 0.088 when the reaction times were extended to 6 and 12 h, respectively. Because the DS of CMNFBC #12 was similar to that of CMNFBC #6, a reaction time of 12 h was sufficient for the carboxymethylation of the nanofiber surface.

Figure 8

Solid-state 13C NMR spectra of CMNFBC #10–12. The DS was determined from the integral value of the carbonyl carbon (C8) resonance relative to that of the C1 resonance, which was set to 1.

The XRD patterns of CMNFBC #8–12 (Figure S5) indicated that all the samples retained the crystalline cellulose I structure with crystallinity indices in the range of 63–66% (Table ), similar to neat NFBC. However, the selective planar orientation of these samples decreased with increasing DS, indicating that the carboxymethylation of the nanofiber surface inhibited the formation of interfiber hydrogen bonds. In addition, the zeta potentials of the aqueous suspensions became more negative with an increase in DS (Table ). These results for CMNFBC #8–12 confirmed that carboxymethyl group substitution occurred on the NFBC surface and improved the dispersibility.

Effects of Carboxymethylation Treatment

To improve the water dispersibility of CNFs, several methods have been devised for synthesizing CNFs with carboxymethyl-modified surface layers. Most methods involve the nonuniform carboxymethylation of raw pulp materials with subsequent mechanical treatment, such as high-pressure homogenization, to produce nanofibers.[7] For the carboxymethylation of pulp, a water-insoluble organic solvent such as isopropanol is generally used to improve compatibility with the alkali solution and achieve efficient carboxymethylation.[24,25] The mechanical treatment of these partially carboxymethylated pulps is very effective in reducing the cost of CNF production. On the other hand, excessive mechanical treatment includes the possibility of damaging and shortening the fiber length, decreasing aspect ratio, and nonuniform fiber widths.[7−12] In addition, The distribution of carboxymethyl groups may also depend on the mechanical treatment conditions in addition to the carboxymethylation reaction conditions.

In contrast, as revealed by the above optimization study, the use of CA in a 3 wt % sodium hydroxide aqueous solution without isopropanol allowed the surface of NFBC to be selectively carboxymethylated while maintaining the fiber morphology and hierarchical structure. To efficiently modify the surface layer, the optimal reaction conditions were determined to be a CA concentration of 10 times the AGU content in NFBC, a reaction temperature of 353 K or higher, and a reaction time of 12 h or more. Carboxymethylation under these conditions improved the dispersibility of NFBC without mercerization or swelling of the nanofibers.

Conclusions

In conclusion, it was found that the carboxymethylation of NFBC in a low-concentration alkaline aqueous solution results in the selective substitution of hydroxyl groups on the surface with carboxymethyl groups without changing the nanofiber structure. The obtained carboxymethylated NFBC nanofibers are expected to be used as transparent films and thickeners in medicine and food products, taking advantage of their long fiber length and microbial origin. In addition, the optimized reaction conditions determined in this study for NFBC could also be applied to other polysaccharide nanofibers such as CNFs and chitin nanofibers synthesized by various methods. The obtained carboxymethylated nanofibers exhibited improved dispersibility in water and increased the transparency of dispersions, which is expected to enhance the performance of nanofibers as dispersants and thickeners.

Experimental Section

Materials

All chemicals were of high-purity grade, and all solutions were prepared with deionized water. A 1 wt % suspension of the CM-NFBC[17] biosynthesized by Gluconacetobacter intermedius NEDO-01 (NITE P-1495) was provided by Kusano Sakko, Inc. (Japan). Before use, NFBC was thoroughly purified from the CM-NFBC using the following procedure. An 800 mL aliquot of the CM-NFBC suspension was concentrated by centrifugation at 16,000g for 10 min. The obtained pellet (dry weight: ∼8 g) was dispersed in aqueous 2 wt % sodium hydroxide solution by shaking at 150 rpm for 2 h, and then, the dispersion was centrifuged. After repeating the purification treatment three times, the pellet was washed with deionized water until it became neutral. Purified NFBC was suspended in deionized water at a concentration of ∼2.5 wt %. The precise concentration of the NFBC suspension was determined from the dry weight obtained for a 10 mL aliquot using a heating-type moisture analyzer (MS-70, A&D Co., Japan). Purified NFBC, hereafter referred to as neat NFBC, was used as a blank sample.

Preparation of CMNFBC

The neat NFBC suspension (1 g of dry weight, 6.2 mmol of anhydroglucose units (AGUs)) was concentrated by centrifugation at 16,000 for 10 min. Subsequently, the pellet was suspended in 300 mL of the reaction solvent (Table ) by stirring for 10 min. After repeating this solvent exchange procedure three times, the NFBC slurry was placed in a 500 mL glass flask equipped with a reflux condenser and a Teflon stirring blade. CA (2–10 mol/AGU; Table ) was added to the slurry, and the carboxymethylation reaction was carried out at various temperatures (298–363 K), controlled using an aluminum block thermostatic bath, with various reaction times (1–24 h), as listed in Table . After the reaction, the product was collected by centrifugation at 16,000g for 10 min and dialyzed against a stream of deionized water for 3 days to obtain the CMNFBC sample. Each purified CMNFBC sample was diluted with deionized water to a concentration of 0.5 wt % and then stored at 277 K before use. An aliquot of each CMNFBC suspension was freeze-dried for structural characterization.

Solid-State 13C NMR

Solid-state cross-polarization/magic-angle spinning 13C NMR (SSNMR) experiments were performed at 295 K on an AVIII500 spectrometer (1H frequency of 500.13 MHz) equipped with a dual-tuned 4 mm magic-angle spinning probe (Bruker BioSpin GmbH, Germany). The sample (80 mg) was packed into a 4 mm ZrO2 rotor, and the rotation frequency of the rotor was set to 10 kHz. The SSNMR spectra were obtained using a contact time, repetition time, and scan number of 2, 4, and 6144, respectively. The 13C chemical shifts were calibrated using the carboxyl carbon resonance of d-glycine (176.03 ppm) as an external standard.

X-ray Diffraction (XRD)

XRD patterns were recorded on a D8 ADVANCE diffractometer (Bruker AXS GmbH, Germany) equipped with a vertical goniometer (Cu/Kα, 0.1542 nm) at 295 K. The samples were analyzed at 40 kV and 50 mA in the 2θ range of 10–30° at a scan rate of 0.500°/min with a step size of 0.01°, where θ is the angle of incidence of the X-ray beam on the sample.

Morphological Observations

Field-emission scanning electron microscopy (SEM) was performed using a JSM-7500F electron microscope (JEOL Ltd., Japan). The freeze-dried sample was mounted on a brass stub using conductive carbon tape, coated with a thin layer (∼3 nm) of platinum by ion sputtering, and then observed at an accelerating voltage of 7 kV. Scanning probe microscopy (SPM) measurements were performed using an AFM5100N atomic force microscope (Hitachi High-Tech Co., Japan) with a silicone cantilever (SI-DF3P2, Hitachi High-Tech Co.). A drop (2 μL) of a neat NFBC or CMNFBC suspension (0.01 wt %) was placed on a newly cleaved mica plate, allowed to dry at 303 K in an oven for 30 min, and then analyzed. The fiber width distribution was determined using ImageJ software (Ver. 1.51w13, Wayne Rasband, National Institutes of Health, USA), with average widths calculated from 100 randomly selected points on the fibers in the SEM and SPM images.

Zeta Potential Measurements

The electrophoretic mobility of each CMNFBC or neat NFBC suspension (0.005 wt %) was measured using a Litesizer 500 analyzer (Anton Paar GmbH, Austria).

Water Dispersibility

The water dispersibility of each CMNFBC sample and neat NFBC was estimated based on the transmittance of the corresponding suspension (0.1 wt %) measured in the 340–750 nm wavelength range using an Evolution 201 UV–vis spectrometer (Thermo Fisher Scientific Inc., USA), with deionized water as the reference.