Fabrication of ultrafine C-CNFs
Mass producible and affordable bleached pulp was selected as the starting material to prepare the unmodified CNFs and objective C-CNFs via the combination of chemical pretreatment and the mechanical nanofibrillation process. Generally, the CNFs from NaClO2 and KOH pretreatment do not have uniformly fine diameters and high zeta potentials and are then incapable of forming a homogeneous and long-standing stable suspension (Figure 2a-c). To pretreat bleached pulp to fabricate the objective C-CNFs, tailored H-DESs composed of CA (H-bond donor), ChCl (H-bond acceptor), and water (proton regulating agent) were designed and synthesized. Using the aspect ratio, carboxyl content, and yield of C-CNFs as evaluation indicators, various combinations of H-DES components in different ratios were tested to obtain an optimal formula. Results show that obtained C-CNFs exhibit finer diameters and higher carboxyl content with increasing molar ratio of CA to ChCl (Figure S1). Meanwhile, the C-CNFs suspension becomes more transparent, showing improved homogeneity and dispersion stability (Figure 2d). Note that excessive CA would cause excessive hydrolysis of cellulose, which results in a reduction in the length and DP of C-CNFs (Figure S1). For example, the lengths of C-CNFs from anhydrous DES (CA:ChCl=4:1) are shortened to around 1 μm. Therefore, the mass ratio of CA:ChCl at 3:1 is deemed an optimal ratio of these two components. The C-CNFs prepared with such a solvent show fine diameters of ~3 nm and good dispersibility (Figure 2e,f).
By adding an appropriate amount of water, the cost, H-bond acidity, and viscosity of H-DES can be reduced while maintaining C-CNF lengths greater than 2 μm, and the optimal mass ratio of CA:ChCl:H2O is determined to be 3:1:1 (Figure S2). The C-CNFs (Figure 2g-i) prepared with this solvent display more uniform and finer diameters and higher carboxyl content than the unmodified CNFs, and are comparable to those of the C-CNFs (Figure 2d-f) prepared with the anhydrous DES (CA:ChCl=3:1). The C-CNFs prepared in this condition also generate a high yield of 90%. The C-CNFs are primarily composed of single elementary fibers which show the statistically finest fiber diameter of 2.5 nm, longest fiber length of 10 μm, and highest aspect ratio of 4000 (Figure 2j). Three distinct peaks at 2θ = 16.2, 22.6, and 34.6, corresponding to the (110), (200), and (004) crystalline planes, are visible in all samples, indicating that the cellulose I crystalline structure was preserved following H-DES treatment (Figure S3). Therefore, the desired C-CNFs with high yield were successfully fabricated under the synergistic regulation of H-DES and subsequent ultrasonic nanofibrillation treatments.
Fourier transform infrared (FTIR) spectroscopy analysis reveals the reason for the high carboxylation rate of C-CNFs (Figure S4). Two adjacent carboxylic groups of CA are easily dehydrated to form the five-membered cyclic carboxylic anhydride,13 which has a negative effect on the carboxylation rate of cellulose. However, no cyclic carboxylic anhydride is detected in the FTIR spectrum of H-DES. We presume that the existence of ChCl featured with a CH2OH chain extending from one of the vertices can generate a steric hindrance effect, which hinders the formation of cyclic carboxylic anhydride. In this condition, more free carboxylic groups grafted to the cellulose improve the carboxylation rate of C-CNFs. Therefore, the C-CNFs prepared by the optimized H-DES have a high carboxyl content of 1.5 mmol/g (Figure S5), which is higher than most of the reported values in literature, while the width, length, and yield of C-CNFs are simultaneously far above those obtained by TEMPO-mediated oxidation and organic acid treatments (Figure 3a-c). Zeta potential is mainly associated with the carboxyl content of C-CNFs, and the zeta potential value of C-CNFs prepared using the optimized H-DES was about –47 mV, exceeding most of those achieved by conventional methods (Figure 3d). High carboxyl content and zeta potential are beneficial to the dispersion stability of C-CNFs in aqueous solution.
Chemical characteristics of C-CNFs
We further investigated the chemical structures of C-CNFs using a series of spectroscopic methods. In the Raman spectrum of C-CNFs, a new absorption peak at 1740 cm-1 corresponding to the carbonyl group (C=O) appears in Figure 4a, indicating the grafting of carboxyl groups onto CNFs. The Raman mapping images show that only hydroxyl groups (O-H) can be detected from the CNFs (Figure 4b), whereas hydroxyl groups and carbonyl groups (C=O) are manifested clearly from the C-CNFs (Figure 4c), demonstrating homogeneous distribution of carboxyl groups in the C-CNFs.32 X-ray photoelectron spectroscopy (XPS) confirms the chemical structure and state changes of C/O elements (Figure 4d). The C/O ratio of C-CNFs is higher than that of CNFs, suggesting that a plenty of carbon-oxygen groups has been grafted. Therefore, the C-CNFs have a new carbon atom binding form of O-C=O at 287.8 eV, compared to CNFs (Figure 4e,f). The signal of C=O stretching of carboxyl groups at 1735 cm-1 also appeared in the FTIR spectrum (Figure 4g). Furthermore, various C=O forms of the ester group (C7), methylene group (C8, C8*), quaternary carbon atom (C9), and carboxyl group (C10) are further revealed by 13C NMR spectroscopy (Figure 4h), proving the conjectural linkage structure.
The esterification of carboxyl groups onto cellulose leads to the crosslinking of C-CNFs. The degree of crosslinking can be assessed by the ratio of free carboxyl content (FCC) to total carboxyl content (TCC).18 FCC is the number of free carboxyl groups, TCC is the total number of free carboxyl groups (-COOH) and bonded carboxyl/ester groups (O-C=O) of CA.17 Ideally, only one carboxyl group of CA is esterified with cellulose, while the other two carboxyl groups remain free, and the FCC/TCC ratio under this condition is 2/3 (Figure S6), which indicates that no crosslinking has occurred. The FCC/TCC ratio of C-CNFs in this experiment is so close to this ideal value, which suggests that scarcely any crosslinking occurred between C-CNFs.
Physical and mechanical properties of C-CNFs
As shown in Figure 5a, freestanding and flexible films consisting of interwoven C-CNFs can be fabricated through a facile solution casting method. In comparison to conventional CNF and TEMPO-oxidized CNF films, the C-CNF film exhibits higher transmittance of 92% at 600 nm (Figure 5b) while maintaining good haze of 90% (Figure 5c), suggesting its use as a substrate material in emerging flexible electronics and optoelectronics.33-34 The C-CNF film has excellent mechanical properties, with a tensile strength of 115.4 MPa, which is 5.5 times higher than that of a CNF film (Figure 5d). The toughness, modulus, and elongation of the C-CNF film are 17.2, 4.5, and 3.2 times higher than those of the CNF film, respectively (Figure 5e,f). The C-CNFs with high aspect ratio and high charge also show superiority in preparing aerogels, and an aerogel made of 0.2 wt% C-CNFs shows ultralow bulk density of 0.002 g/cm3, high porosity of 99.86% (Figure 5g), superlow thermal conductivity (0.0289 W/mK, Figure 5h) close to air (0.025 W/mK), and outstanding compressive robustness (Figure 5i). Thermogravimetric analysis (TGA) reveals that C-CNFs has a thermal degradation temperature (Tmax=348 ℃) close to pristine cellulose (Figure S7). In addition, the co-existence of H-bonds and ester bonds between C-CNFs makes the C-CNF film/aerogel easily recyclable and rebuilt. The disposable C-CNF film/aerogel can be redispersed in water to become a uniform slurry and reformed into various types of products (Figure S8).
Scalable productivity of C-CNFs and derived materials
The preparation process of C-CNFs can be enlarged to a production scale of 100 L per batch (Figure 6a and S9), and the properties of C-CNFs from the large-scale production are almost the same as those from small-scale preparation. The C-CNFs from large-scale production can also maintain a homogeneous suspension state even after six months of storage (Figure 6b). Normally, the maximum solid content of CNFs is below 2 wt% (Figure 6c), while the maximum solid content of C-CNFs in this work reaches 10 wt% due to rich negative charges. The C-CNFs at 10 wt% solid content display a very fine and milky gel appearance (Figure 6d and Figure S10), with no obvious agglomerates observed. The production of such high solid content of C-CNFs can reduce the reagent cost and energy consumption, as well as facilitate the transportation. The H-DES used in this study to produce C-CNFs can be recycled and reused, which brings an economical and efficient advantage. Even after four cycles, the yield of C-CNFs is still 90% and the carboxyl content is 1.42 mmol/g (Figure S11). Therefore, this method has advantages of high efficiency, large-scale production, low cost, and recyclability, which shows its potential in industrial applications.
Based on the advantages of low cost and scaling up production of C-CNFs, we are capable of fabricating their materials from the laboratory to the pilot stage and maintaining their performance without degradation. For example, we have manufactured a transparent C-CNF film with a length of 4 m (Figure 6f) by solution casting and a room temperature drying process (Figure S12),35 and a large bulk C-CNF aerogel with dimensions of 150 cm × 25 cm × 1 cm (Figure 6e). Furthermore, as shown in the conceptual diagram in Figure 6g, the C-CNF films and aerogels and other derivatives can be produced on an industrial scale with the help of processing equipment.