Fabrication of conductive polyaniline nanomaterials based on redispersed cellulose nanofibrils

Dewatering or drying of diluted cellulose nanofibril suspensions is an effective way for reducing the costs of transportation and storage. In this study, poly(vinylpyrrolidone) (PVP) was introduced into a redispersing system of concentrated CNFs, and the obtained redispersed CNFs were used for fabricating CNF/PANI (polyaniline) nanomaterials by in situ polymerization method. The results showed that mechanical grinding with grinding gap of − 20 μm was an effective way to redisperse the concentrated CNFs, especially when the PVP was added in the redispersing process. The conductivity of the redispersed CNF/PANI film reached 1.08 S/cm and the specific capacitance reached 118.3 F/g (at 0.3 A/g), when the concentrated CNFs were redispersed with 5% PVP. During the polymerization process, PVP facilitated the PANI coating on the CNFs uniformly as steric stabilizer. This study provided a basis for the application of redispersed CNFs in conductive nanomaterials area.


Introduction
Cellulose nanofibrils (CNFs) as a kind of nanocellulose are usually disintegrated from cellulosic fibers by mechanical process after chemical or enzymatic pretreatment. Due to the high aspect ratio, entangled network structure, and abundant hydroxyl groups, CNFs have been attracted considerable attention in many fields, such as biomedicine, packaging, papermaking, cosmetics, food, conductive bio-nanomaterials, etc. (Wang et al. 2021;Perrin et al. 2020;Rosenau et al. 2019;Lourenço et al. 2019;Yu et al. 2017). Generally, the prepared CNFs are stored in a form of aqueous suspension and have a water content greater than 98% (Isogai, 2020). The low solid content of CNF suspension leads to high transport and storge costs, which limits the end-usage application of CNFs. In order to facilitate the commercialization of CNFs, many dehydration methods are applied to dry the CNF suspension including oven drying, freeze drying, spray drying, and supercritical drying (Peng et al. 2011). In the drying process, the agglomeration of CNFs was inevitable, and the drying strategies of adding additives (Kwak et al. 2019;Velásquez-Cock Abstract Dewatering or drying of diluted cellulose nanofibril suspensions is an effective way for reducing the costs of transportation and storage. In this study, poly(vinylpyrrolidone) (PVP) was introduced into a redispersing system of concentrated CNFs, and the obtained redispersed CNFs were used for fabricating CNF/PANI (polyaniline) nanomaterials by in situ polymerization method. The results showed that mechanical grinding with grinding gap of − 20 μm was an effective way to redisperse the concentrated CNFs, especially when the PVP was added in the redispersing process. The conductivity of the redispersed CNF/PANI film reached 1.08 S/cm and the specific capacitance reached 118.3 F/g (at 0.3 A/g), when the concentrated CNFs were redispersed with 5% PVP. During the polymerization process, PVP facilitated the PANI coating on the CNFs uniformly as steric stabilizer. This study provided a basis for the 1 3 Vol:. (1234567890) et al. 2017; Beaumont et al. 2017) or modifications (Yan et al. 2016;Wang et al. 2016;Eyholzer et al. 2009) of CNFs were proposed. For example, the replacement of water in CNF suspension with PEG had also been investigated to inhibit CNF agglomerations (Santmarti et al. 2020). However, dehydration of CNF suspension always leads to severe irreversible agglomeration due to the formation of hydrogen bonds, intermolecular forces, and capillary forces between adjacent fibrils (Wohlert et al. 2021;Peng et al. 2013;Quiévy et al. 2009). How to obtain the pristine structure and morphology of the individual CNFs in the dilute state after redispersing is a challenge for CNF end-use application (Sinquefield et al. 2020). Therefore, the development of low-cost, efficient, and non-destructive redispersing methods of concentrated CNF suspension is essential for commercialization of CNFs. In order to reduce the cost, concentration process have been proposed to obtain high solid content CNF suspension instead of dehydration process, and additives are usually introduced to CNF suspension, such as NaCl, PVA, CMC, t-BuOH, PVP, cyclodextrin, etc. (Sinquefield et al. 2020;Ding et al. 2019). These additives are adsorbed onto the surface hydroxyl groups of CNFs. When the CNF suspension is dewatering, the additives interfere with and block intrafibrillar bond formation. Poly(vinylpyrrolidone) (PVP) is a nonionic water-soluble polymer and usually used as dispersant Marani et al. 2015). Previous study showed CNF film had a relatively smooth surface by adding PVP due to the good dispersion of PVP in the CNF suspension . Besides, in the preparation process of some conductive polymers, based on the steric stabilizing effects of PVP the conductive polymers can grow and distribute uniformly (Ewulonu et al. 2020;Wu et al. 2014.).
As a conductive polymer, polyaniline (PANI) has attracted a lot of attention due to its low cost, easy synthesis, and conductivity. The redox reaction of PANI during charge/discharge process endows conductive material with pseudo capacitance, and thus enhances the capacitance (Itoi et al. 2018). However, neat PANI has poor film-forming capability limiting its application. The web-like CNFs with abundant hydroxyl groups could be chosen as a promising template for the polymerization of aniline monomer (Gopakumar et al. 2018;Yu et al. 2014;Rußler et al. 2011). In the process of polymerization, the hydrogen bonds, intermolecular forces, and capillary forces are induced the hydroxyl groups of CNFs and amine groups of PANI to form strong combinations, which ensures the continuous conductive network of CNF/ PANI composites (Wohlert et al. 2021;Zhang et al. 2019). Besides, the flexible CNF/PANI conductive nanomaterial had remarkable mechanical properties, lightweight and environmental stability, showing great potential application in electronic devices.
In this work, we developed a facile process for concentrating-redispersing CNFs aiming to efficiently utilize redispersed CNFs for synthesizing conductive polymer nanomaterials. CNF suspension was concentrated by a centrifuge and redispersed by a grinding process. PVP was used as dispersant in redispersing process, and the redispersed CNFs were further submitted to synthesize CNF/PANI conductive nanomaterials. PVP had a function of alleviating the agglomeration of PANI during the polymerization, which may facilitate the construction of continuous conductive network. This study can provide technical and theoretical support for the large-scale production and application of CNF based nanocomposites.

Materials
The commercial bleached softwood kraft pulp (BSKP) was obtained from a paper company (Guangdong, China) and used for the preparation of CNFs. Commercial endoglucanase (enzyme activity 7.3 IU/ ml) was acquired from Paper Chemical Co., Ltd (Jiangsu, China). Aniline, ammonium persulfate [(NH 4 ) 2 S 2 O 8 , APS], Congo red with a relative molecular weight of 696.66 and poly(vinylpyrrolidone) (PVP) with a relative molecular weight of 58,000 were purchased from Macklin Biochemical Co., Ltd (Shanghai, China). Citric acid, sodium citrate and hydrochloric acid were analytical grade.

Preparation of CNFs
The dry BSKP was immersed in water with a concentration of 3.0 wt% for overnight. Then the pulp was defibered by a lab Valley Beater (Voith Inc., Appleton, WI, USA) for 15 min. The enzymatic pretreatment was carried out based on previous report with some modifications (Huang et al. 2020). A certain amount of pretreated pulp was immersed in 50 mM citric acid-sodium citrate buffer solution (pH 5.2) to reach a concentration of 4.0 wt% and the endoglucanase loading was 9 mg/g substrate. The slurry was continuously agitated in a water bath at 50 °C for 2 h. After enzymatic pretreatment, the slurry was placed in a water bath at 85 °C for 15 min to denature the endoglucanase. The obtained pulp was washed with distilled water until the extruded water was natural and stored at 4 °C. The pretreated pulp was diluted to a concentration of 1.0 wt% and mechanically fibrillated using a super masscolloider (MKCA 6-2J, Masuko Sangyo Co., Ltd, Japan). The gap of discs was adjusted from 0 to − 100 μm. Finally, CNFs were obtained by passing grinder 15 times with a gap of − 100 μm at 2000 rpm.

Concentration and redispersion of CNFs
The obtained CNF (CNF ND ) suspension was added into a 10,000-mesh polypropylene fiber bag, and then concentrated by a centrifuge (PSB300N, China) for 15 min at the centrifugal force of 10,000 g. The 9.0wt% concentrated CNFs were diluted to 0.5 wt% with distilled water, and mechanically stirred at 300 rpm for 30 min labeled as CNF MS . The diluted CNFs were grinded by the super masscolloider with the disc gap of 0, − 20, − 50 μm for 5 times at 800 rpm, respectively. The dispersed CNFs obtained by grinding under different gaps was named as CNF R(0) , CNF R(-2) and CNF R(-5) , respectively. In addition, prior to mechanical stirring, 5% or 10% PVP (relative to the oven dried CNFs) was added to CNF suspension to investigate its role in the dispersion of CNFs. The samples added with 5% and 10% PVP were grinded with the disc gap of − 20 μm, and labeled as CNF R(-2)-5PVP and CNF R(-2)-10PVP .

Preparation of CNF films
The CNF films were prepared by vacuum filtration and subsequent hot pressing. The original and redispersed CNF suspensions were filtered through a 0.45 μm hydrophilic filter membrane (CA, Shanghai, China). After filtration, wet cakes were dried on a heating plate at 80 °C for 20 h with a loading of 2.5 kg.

Preparation of CNF/PANI conductive nanomaterials
The CNF/PANI conductive nanomaterials were prepared by in situ polymerization in the presence of APS as an oxidant and hydrochloric acid (HCl) as a dopant. Aniline monomer (0.25 g) was added into 5 ml 2 M HCl solution, and then dissolved in CNF suspension. The mass ratio of CNF/PANI was 1:1. The solution was stirred for 30 min at the temperature of 5 °C. The polymerization was initiated by adding 0.78 g APS (dissolved in 5 ml 2 M HCl). The molar ratio of APS/aniline was 1.25:1, and the final CNF concentration was adjusted to 0.3 wt%. After 90 min polymerization reaction, the dark green solution was filtered and washed with deionized water. The wet composites were dried as the previous CNF film process.

Characterization
Atomic force microscopy (AFM, Bruker, Germany) was used for analyzing the morphology of CNFs. The diameters of CNFs were obtained by measuring its width. 100-110 nanofibrils were measured by nanomeasure software 1.2. Transmittance of CNF suspensions (0.1 wt%) and films (60 g/m 2 ) were performed on a UV-Vis spectrophotometer (UV-1900, China) in the wavelength between 400 and 800 nm. Films were cut into 35 × 8 mm and were equilibrated at 23 ± 2 °C and 50 ± 2% RH for more than 48 h, and then tensile strength were performed with a span length of 15 mm at a stretched speed of 5 mm/min on a Tensile Compressive Universal Testing Machine (INSTRON 3342, USA). Fourier transform infrared (FTIR) spectra was acquired by a spectrophotometer (FT-IR 4700, Japan) in the wavenumber range 4000 to 500 cm −1 with a resolution of 4 cm −1 at 25 °C. The surface morphologies of CNF films and CNF/PANI nanomaterials were characterized by Field-emission scanning electron microscopy (FE-SEM, Merlin, Zeiss, Germany). Before FE-SEM test, a thin gold layer was sputtered on the samples and then operated at an accelerating voltage of 10 kV.
The specific surface area (SSA) was determined by the Congo red adsorption method, as reported by Wang et al. (2018). A certain amount of CNF suspension was mixed with Congo red solution and incubated at 60 °C for 24 h in an orbital shaker. To remove the unabsorbed Congo red, the suspension was centrifuged at 10,000 rpm for 15 min. The concentration of the supernatant was calculated from the standard curve. The absorbance of standard solution and supernatant were determined at 495 nm using an UV-Vis spectrophotometer (UV-2600, China). SSA (m 2 /g) was calculated according to the following Eq. (1), where N A is Avogadro's constant (6.022 × 10 23 mol −1 ), m 1 is the saturated absorption value (g/Kg), A d is the surface area of one Congo red molecule (1.73 nm 2 ), M ω is the molecular weight of Congo red, 696.66 g/ mol.
The zeta potential of CNFs was measured by a Fiber Potential Analyzer (FPA, Germany). Each sample was measured at a concentration of 0.5 wt%.
The water retention value (WRV) was measured by a centrifugal method (Gu et al. 2018). A certain amount of CNF suspension was wrapped in a 5000mesh nylon bag and placed in a centrifuged tube with a metal mesh in the middle to allow water to accumulate at the bottom of the tube. All samples were concentrated with the centrifugal force of 3000 g for 15 min by a centrifuge (Cence, H2050R, China). The centrifuged CNF gel was weighed before and after drying in an oven at 105 °C, respectively. The WRV was calculated from the following Eq. (2), where m wet and m dried are the weights of centrifuged CNF samples before and after oven drying, respectively.
The X-ray diffraction (XRD) measurement was performed on an X-ray diffractometer (X'pert powder, PANalytical, Netherlands), which was operated at 40 kV and 40 mA. The diffracted intensity of Cu Kα radiation was measured in a 2θ range between 5° and 40° at a scanning rate of 12°/min. The crystallinity index (CrI) was calculated by using the following Eq. (3) (Segal et al. 1959): where I 200 is the maximum intensity of 200 plane at 2θ around 22.5°, and I am is the minimum intensity at 2θ close to 18.5° which represents the amorphous regions. The crystallite sizes (D hkl ) of the crystal in 200 lattice planes were calculated according to Scherrer Eq. (4) (Yu et al. 2013;Ten et al.2012): where λ is the radiation wavelength (0.15418 nm), β is the width of half-maximum intensity in radians and θ is the diffraction angle.

Electrochemical tests
The conductivity of composite film was measured by a four-point probe (KDY-1, China). Cyclic voltammetry (CV) and galvanostatic charge/discharge (GCD) tests were carried on an electrochemical workstation (CHI660E, China). The composite film, Pt foil and Ag/ AgCl electrode were used as working electrode, counter electrode, and reference electrode, respectively. 2 M H 2 SO 4 was used as the electrolyte. CV tests were conducted in the potential window of − 0.3 to 0.7 V at different scan rates of 10, 20, 30, 50, 100, 200 mV/s. The specific capacitance (C Scv , F/g) of each sample can be calculated from the CV curve and the equation is as follows: where ∫ IdV is the integrated area of the CV curve, v is the scan rate (mV/s), m is the mass of active material (g), and ΔV is the potential window (V). The galvanostatic charge/discharge tests were performed in the potential range of from − 0.3 V to 0.7 V at different current density. The specific capacitance (C Sgcd , F/g) calculated from GCD curve is as follows: where I is the constant current from discharge (A), Δt is the discharge time (s), m is the mass of the active material (g), and ΔV is the voltage of discharge excluding voltage drop (V).

Effect of redispersion on the properties of CNFs
Specific surface area (SSA) was an effective method to characterize the redispersion of CNFs (Velásquez-Cock et al. 2017;Beaumont et al. 2017). In this study, SSA of CNFs was obtained by Congo red adsorption method. As seen in Table 1, CNF MS showed the lowest SSA of 87.39 m 2 /g, which indicated that mechanical stirring could not effectively redisperse the agglomerations of CNFs. In the concentration process of CNF suspensions, the fibrils could form irreversible bindings due to hydrogen bonds and intermolecular forces resulting in the SSA decrease of CNFs (Wang et al. 2021;Kwak et al. 2019). The SSA of CNFs was also affected by grinding parameters. The SSA of redispersed CNFs increased with the grinding gap decreasing. It was noteworthy that the SSA of CNF R(-5) (106.93 m 2 /g) was higher than that of CNF ND (104.66 m 2 /g). The results demonstrated that the grinding fibrillation method was an effective way in improving the SSA of redispersed CNFs.
In the process of CNF concentration, with the removal of water, the fibers gradually get together and form hydrogen bonds between fibrils. This leaded to fibril aggregation and influenced the swelling ability of CNFs. WRV was used to evaluated the effect of grinding treatment on the redispersion of concentrated CNFs. The WRV results of redispersed CNFs were shown in Table 1. The WRV of CNF MS was 531%, which was 68% lower than that of CNF ND (599%), which indicated that the concentrated CNFs maybe not totally redispersed by the treatment of mechanical stirring. However, WRV of redispersed CNFs significantly increased from 555 to 647% with the reduction of grinding gap. Grinding treatment was an effective way to redisperse the concentrated CNFs, and realized the concentrated CNFs had micro/nano scale and relatively uniform distribution again. This was consistent with previous research that WRV was depended on the fibrillation of fibers to some extent (Gu et al. 2018).
Zeta potential is an important parameter to reflect the stability of colloidal dispersion. CNFs exhibited negative zeta potential, which was attributed the hydroxyl groups on fiber surface. Zeta potential of initial and redispersed CNFs were listed in Table 1. The zeta potential of CNF ND was − 34.3 mV. CNF MS had a zeta potential of − 26.4 mV, which indicated that the agglomeration had a negative effect on the absolute value of zeta potential. When non-ionic PVP was added in the system, the absolute value of zeta potential decreased significantly, indicating that PVP adsorbing onto the CNFs surface. The dispersion stability of CNF/PVP suspension was affected by the electrostatic repulsion between CNFs and PVP (Yang et al. 2017;Han et al. 2013). As seen in Fig. 1a, the CNF stability was enhanced as the redispersed condition became intense. When PVP was added in the redispersion process, the redispersed CNFs showed high stability especially with the addition of PVP increasing (Fig. 1a). PVP played the role as a dispersant and provided steric barriers between fibrils. The possible mechanism of PVP action on the stability of CNF suspension was suggested and shown in Fig. 1b.
The transmittance of CNF suspension was a typical representation for the dispersion of CNFs . Figure 2 showed the transmittance of CNF suspensions. The results showed that the redispersing property of CNF suspension was consistent with the transmission performance. The CNFs with small and uniform size was much stable, which reduced the light scattering and benefited for improving the transmittance. This result was also consistent with the previous reports (Zhu et al. 2013). AFM images were used to characterize the morphologies of initial and redispersed CNFs, and the diameter distribution of CNFs was analyzed. As seen in Fig. 3a, CNF ND with a mean diameter (MD) of 42.57 nm exhibited good uniformity. After concentration and mechanical stirring, CNF MS showed aggregation performance (Fig. 3b) due to the forming of intermolecular binding and hydrogen bonds between fibrils. The fiber bundles enabled CNF MS to have a high MD up to 97.28 nm. It indicated that the agglomerations between fibers were difficult to redisperse in the stirring process. The MD of CNF R(0) , CNF R(-2) and CNF R(-5) were 66.60, 50.69 and 42.27 nm, respectively. Figure 3c, d and e revealed that the agglomerations significantly decreased with the decrease of grinding gap. The grinding fibrillation process was benefit for forming uniform diameter distribution of redispersed CNFs. This also supported by the results of WRV (Table 1). CNF R(-2), CNF R(-2)-5PVP and CNF R(-2)-10PVP had similar diameter distribution. High addition of PVP resulted in the self-agglomeration, which was labeled by green arrows in AFM image of CNF R(-2)-10PVP sample (Fig. 3g). The chemical bonds of initial and redispersed CNFs were analyzed by FTIR. FTIR spectra of PVP, initial and redispersed CNFs were exhibited in Fig. 4a. The characteristic absorption peaks at 1665 and 1288 cm −1 were assigned to the stretching of C=O and C-N of PVP, respectively. A large absorption peak at 3400 cm −1 indicated that PVP contained absorbed water (Voronova et al. 2018). The peaks at 3329 and 1900 cm −1 were attributed to O-H vibration and C-H stretching of CNFs, respectively. Other typical cellulose peaks were observed at 2900 cm −1 (C-H stretching), 1427 cm −1 (-CH 2 and -OCH in-plain deformation), 1369 cm −1 (C-H deformation vibration) and 1051 cm −1 (C-O-C pyranose ring skeletal vibration). The spectrum of CNF R(-2) was similar to that of CNF ND , indicating that the chemical structure did not change in the redispersion process. The O-H vibration of CNF R(-2)-5PVP and CNF R(-2)-10PVP was shifted to 3333 cm −1 and 3336 cm −1 , respectively, which indicated that the new interaction between amide groups of PVP and hydroxyl groups of CNFs weakened the interfibrillar hydrogen bonds. In addition, the peaks at 1105 cm −1 are related to the C-O-C glycosidic ether linkages (Deepa et al. 2015). When PVP was added, the C-O-C group of CNF R(-2)-10PVP Fig. 3 AFM images and diameter distribution histograms of a CNF ND , b CNF MS , c CNF R(0) , d CNF R(-2) , e CNF R(-5) , f CNF R(-2)-5PVP and g CNF R(-2)-10PVP . The agglomerations of PVP were marked by green arrows showed a 1 cm −1 high-wavelength-shift to 1106 cm −1 , which may be attributed to the weakening of hydrogen bond between C-O-C groups and carbonyl groups (Gao and Jin 2018).
The crystalline structure of initial and redispersed CNF films was characterized by XRD. The XRD patterns of CNFs treated with different redispersing process were shown in Fig. 4b. The diffraction peaks of all samples at 2θ = 15.1°, 16.5°, 22.5° are corresponding to ( 110 ), (110), (200) lattice plane, respectively, suggesting that non-concentrated and redispersed CNFs are typical cellulose Ι crystalline form. All the samples had the similar CrI range from 68.34 to 71.75% and D hkl range from 3.02 to 3.20 nm ( Table 1). The results indicated that the concentration and redispersing process hardly affected the crystal structure of CNFs, and proved again the intermolecular and intramolecular binding between fibrils formed in the concentration process mainly occurred in the amorphous region of fibrils (Wang et al. 2021).

Properties of CNF films
The transmittance of CNF films (Fig. 5a) is also determined by the diameter distribution of CNFs. Fibril agglomerations were observed in CNF MS , CNF R(0) and CNF R(-2) films (Fig. 5e, f, g), leading to the transmittance of films decreasing. When 5% PVP was added to redisperse CNFs, the surface of CNF R(-2)-5PVP film (Fig. 5i, j) became smoother and more uniform than that of CNF R(-2) film and almost the same as that of CNF ND film (Fig. 5c, d) (Wang et al. 2021). Therefore, CNF R(-2)-5PVP film had higher transmittance than CNF R(-2) film. The addition of PVP improved the re-dispersibility of concentrated CNFs, and thus enhanced the light transmittance of the films.
The stress-strain curves of films prepared from initial and redispersed CNF suspensions were shown in Fig. 5b. The tensile strength was listed in Table 1. The films prepared by CNFs with high dispersibility degree showed high tensile strength. CNF R(-2) film had a tensile strength of 87.8 MPa. When the addition of PVP increased from 5 to 10% PVP, the tensile strength of films was increased to 103.3 ± 0.4 and 109.2 ± 0.2 MPa, respectively. It was expected that the dispersing effect of PVP further increased the tensile strength of redispersed CNF films . That was to say the PVP not only reduced the aggregations between fibrils in the CNF concentration process, but also contributed to the high strength of redispersed-CNF films.
The electrochemical performance of redispersed-CNF/PANI nanomaterials CNF/PANI flexible nanomaterials (Fig. 6b) were prepared by in situ polymerization. CNFs are supposed to be a promising template for loading PANI due to high surface area and luxuriant hydroxyl groups Zhang et al. 2019). As discussed above, the grinding gap of − 20 μm was chosen to be an appropriate process to redisperse concentrated CNFs. CNF R(-2) , CNF R(-2)-5PVP and CNF R(-2)-10PVP were used film (× 10,000), h CNF R(-5) film(× 10,000), i CNF R(-2)-5PVP film(× 10,000), j CNF R(-2)-5PVP film (× 20,000), and k CNF R(-2)-10PVP film (× 10,000) to prepare PANI nanomaterials. And the CNF ND was regarded as reference substance to study the effect of redispersion on the electrochemical performances of PANI nanomaterials. As seen in Fig. 6a, the conductivity of CNF R(-2) /PANI was 0.64 S/cm, which was lower than CNF ND /PANI (0.83 S/cm). The SSA of CNFs reduced due to irreversible aggregation between fibrils (Table 1), and the decrease in SSA resulted in the binding sites of polyaniline on CNF decreasing and affected the discontinuity of conductive routes, which was responsible for the reduction of conductivity. In addition, compared with CNF ND / PANI (Fig. 6c), CNF R(-2) /PANI (Fig. 6d) had an irregular and heterogeneous surface morphology due to fibril agglomerations, which indicated that noncontinuous conducting network reduced the conductivity. Redispersed CNFs containing PVP were used to fabricate conductive nanomaterials. A relatively homogeneous surface was observed in Fig. 6e for CNF R(-2)-5PVP /PANI. With increasing the addition of PVP to 10%, PANI was uniformly coated on the surface of CNFs. The CNF R(-2)-10PVP /PANI had a low conductivity of 0.75 S/cm, which may attribute to the discontinuous conductive routes caused by excessive non-conductive PVP. In the process of redispersion and polymerization, PVP has two functions: firstly, it was used as dispersant to stable the CNF suspension for further polymerization (Wu et al. 2014). Secondly, it acted as steric stabilizer to facilitate the uniform polymerization of PANI on the CNF surface (Fig. 6g). As seen in Fig. 6h, the hydrogen bonds between CNF, PANI and PVP might severe as traction force to form an integrated 3D network, which contribute to the formation of continues conducting routes. The conductivity of redispersed-CNF R(-2)-5PVP / PANI was as high as 1.08 S/cm, which is an order of magnitude higher than previous reports at the same PANI content (Gopakumar et al. 2018;Luong et al. 2013).
CV measurement was used to evaluate the electrochemical performance of CNF/PANI nanomaterials in a three-electrode system. Figure 7a showed the CV curves of CNF ND /PANI electrode in a potential range of − 0.3 ~ 0.7 V at different scan rates. There were obvious redox peaks on the CV curves, suggesting ideal pseudo capacitances from PANI (Ruan et al. 2020;Han et al. 2019). The comparisons of CNF ND /PANI, CNF R(-2) /PANI, CNF R(-2)-5PVP /PANI and CNF R(-2)-10PVP /PANI nanomaterials were shown in CV curves at a scan rate of 30 mV/s (Fig. 7b). The C Scv of CNF ND /PANI at 30 mV/s was 14.3 F/g. After dewatering-redispersing process, the C Scv of CNF R(-2) /PANI decreased to 4.2 F/g at 30 mV/s due to the agglomeration. CNF R(-2)-5PVP /PANI composite electrode possessed the highest C Scv of 17.6 F/g at 30 mV/s. The excellent capacitive performance of CNF R(-2)-5PVP /PANI was ascribed to the uniform coating of PANI, which promoted the electrolyte contact and ion diffusion . The changes of C Scv at different scan rates were shown in Fig. 7c. When the addition of PVP increased to 10%, the C Scv of CNF R(-2)-10PVP /PANI (14.6 F/g) decreased. High PVP content leaded to discontinues conductive network, and impeded the charge transmission.
To further investigate the effect of redispersion on the capacitance performance of composite electrode, GCD measurement was carried out at various current density. As seen in Fig. 7d, the charge/discharge curves were asymmetric triangles, indicating the presence of pseudo capacitance caused by redox reaction of PANI (Liu et al. 2015;Chen et al. 2013). As the charge/discharge procedure, the time required gradually decreased with the increase of current density, which was similar to the previous report ). The C Scd of the CNF ND /PANI, CNF R(-2) / PANI, CNF R(-2)-5PVP /PANI and CNF R(-2)-10PVP /PANI electrodes were 83.6, 60.0, 118.8 and 97.8 F/g at the current density of 0.3 A/g, respectively. This agrees with the results of C Scv obtained from CV measurements. The capacitance of redispersed-CNF R(-2)-5PVP / PANI was higher than the previous reported CNF/ PANI/reduced graphene oxide composite (79.71 F/g) (Hsu et al. 2019). As the scan rate and current density increase, the C Scv and C Scd (Fig. 7c, f) decreased sharply, which was attributed to the hot-pressing, thereby reducing the porosity of nanomaterials. This further slowed down ion transport and reduced charge storage capacity at high scan rate or current density (Lay et al. 2016).

Conclusion
A mild redispersion method of concentrated CNFs and its application in PANI conductive nanomaterial have been developed in this study. The redispersing process is very essential for the specific surface area, water retention value, zeta potential value, and fibril diameter. The properties of CNFs redispersed by mild grinding process (− 20 μm) are almost the same with never-concentrated CNFs. PVP played an important role in improving the suspension stability of redispersed CNFs and mechanical strength of redispersed CNF film. The effect of redispersion on the electrochemical properties of CNF/PANI conductive nanomaterial was investigated. The concentration-redispersion process of CNFs had a negative effect on the electrochemical properties of conductive nanomaterials. However, CNFs redispersed by adding 5% PVP resulted in significant increase in conductivity and capacitance of nanomaterials, which ascribed to the steric stabilization effect of PVP. In summary, the redispersed-CNFs for the synthesis of PANI nanomaterial significantly facilitated the scalable fabrication and application of CNFs based conductive nanomaterials with high performance.