cellulose nanofiber composite film for self-standing, flexible, conductive electrodes

Accepted Manuscript Title: Vacuum-Assisted Bilayer PEDOT:PSS/Cellulose Nanofiber Composite Film for Self-Standing, Flexible, Conductive Electrodes Authors: Youngsang Ko, Dabum Kim, Ung-Jin Kim, Jungmok You PII: DOI: Reference:

S0144-8617(17)30621-5 http://dx.doi.org/doi:10.1016/j.carbpol.2017.05.096 CARP 12382

To appear in: Received date: Revised date: Accepted date:

18-1-2017 24-5-2017 31-5-2017

Please cite this article as: Ko, Youngsang., Kim, Dabum., Kim, Ung-Jin., & You, Jungmok., Vacuum-Assisted Bilayer PEDOT:PSS/Cellulose Nanofiber Composite Film for Self-Standing, Flexible, Conductive Electrodes.Carbohydrate Polymers http://dx.doi.org/10.1016/j.carbpol.2017.05.096 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Vacuum-Assisted Bilayer PEDOT:PSS/Cellulose Nanofiber Composite Film for Self-Standing, Flexible, Conductive Electrodes Youngsang Ko, Dabum Kim, Ung-Jin Kim*, Jungmok You* Department of Plant & Environmental New Resources, Kyung Hee University, 1732 Deogyeongdaero, Giheung-gu, Yongin-si, Gyeonggi-do 446-701, South Korea *Corresponding Ung-Jin Kim and Jungmok You Tel.: +82-031-201-2626; Fax.: +82-031-204-8117 Email: [email protected] E-mail: [email protected]

Research highlights   

► PEDOT:PSS-CNF composite films were developed as a self-standing, highly flexible conductive film. ► A simple vacuum-assisted filtration was used to fabricate a bilayer PEDOT:PSS-CNF film. ► A bilayer PEDOT:PSS-CNF film exhibited good electrical conductivity and excellent flexibility

Abstract Sustainable cellulose nanofiber (CNF)-based composites as functional conductive materials have garnered considerable attention recently for their use in soft electronic devices. In this work, selfstanding, highly flexible, and conductive PEDOT:PSS-CNF composite films were developed using a simple vacuum-assisted filtration method. Two different composite films were successfully fabricated and then tested: 1) a single-layer composite composed of a mixture of PEDOT:PSS and CNF phases and 2) a bilayer composite composed of an upper PEDOT:PSS membrane layer and a CNF matrix sub-layer. The latter composite was constructed by electrostatic/hydrogen bonding

1

interactions between PEDOT:PSS and CNFs coupled with sequential vacuum-assisted filtration. Our results demonstrated that the resultant bilayer composite film exhibited a competitive electrical conductivity (ca. 22.6 S cm-1) compared to those of previously reported cellulose-based composites. Furthermore, decreases in the electrical properties were not observed in the composite films when they were bent up to 100 times at an angle of 180° and bent multiple times at an angle of 90°, clearly demonstrating their excellent mechanical flexibility. This study provides a straightforward method of fabricating highly flexible, lightweight, and conductive films, which have the potential to be used in high-performance soft electronic systems.

Keyword: Cellulose nanofiber; ; ; ; ; , Conductive polymer, PEDOT:PSS, Vacuum-assisted filtration, Composite film, Pattern

1. Introduction Flexible conductive films are crucial components for the development of next-generation soft electronic products such as implantable biosensors, flexible energy storage devices, wearable electronic devices, and solar cells (Chen, Xu, Wang, Qian, & Sun, 2015; Liu et al., 2015; Malti et al., 2016; Nataraj et al., 2009; Nataraj et al., 2009; Nataraj et al., 2011; Nataraj, Yang, & Aminabhavi, 2012; Singh & Kushwaha, 2013; Song et al., 2013). A common method to fabricate flexible conductive films is to use plastic substrates, such as polyethylene terephthalate (PET), polyethylene naphthalate (PEN), and polycarbonate (PC), because they are amenable to flexible mechanical support and mass production (Kang et al., 2007; Zhu et al., 2013). However, plastic substrates made from the by-products of oil industries have several major limitations, including low processing temperatures, large coefficients of thermal expansion, and poor recyclability. When using PET and PEN substrates, all processing steps should be performed below 110 °C and 160 °C, respectively. To make fundamental advances in next-generation soft electronic technologies, it is important to

2

develop a new set of materials that are easily processable, more flexible, sustainable, light-weight, sufficiently robust, and possess improved electrical properties. There is a growing demand for natural polymer-based nanocomposites for a variety of applications (Chen, Yang, Li, & Li, 2015; Jiang et al., 2015; Kushwaha, Avadhani, & Singh, 2015; Zhou et al., 2016). Among natural polymers, cellulose is the most abundant organic material derived from nature with a biomass production of 1.5 × 1012 tons per year; thus, it is a key source of many types of sustainable materials (Boissou, Vigier, Estrine, Marinkovic, & Jérôme, 2014; Hon, 1994; Hosoda, Tsujimoto, & Uyama, 2014; Hu et al., 2013; Salam, A Lucia, & Jameel, 2013; Tkalya et al., 2013). Over the past decade, cellulose nanofibers (CNFs), which are one type of cellulose-based nanomaterial, have been widely utilized in a variety of applications (including flexible electronics) due to their low cost, low density, high aspect ratio, high mechanical strength, and good thermal and chemical stability. Recently, various carbon nanomaterials (Aiyer et al., 2016; Chen et al., 2015; Jiang et al., 2015; Nataraj et al., 2009; Nataraj et al., 2009; Nataraj et al., 2011; Nataraj et al., 2012; Punetha et al., 2016) have been studied extensively as conductive materials for conductive nanocomposites. In particular, conductive composite films of CNFs and conductive polymers have garnered considerable attention for next-generation soft electronic devices (Mengistie et al, 2015; Salas, Nypelö, Rodriguez-Abreu, Carrillo, & Rojas, 2014; Valtakari et al., 2015; Zhou, Wang, Chen, & Xu, 2015). General strategies for fabricating CNF-based conductive composite materials include blending, coating, and doping of cellulose structures with conductive materials (Shi, Phillips, & Yang, 2013). However, they have intrinsic shortcomings related to their poor solubility, which makes the preparation of uniform composite materials difficult. A number of reports have been dedicated to developing conductive polymer-CNF composites because conductive polymers suffer from poor mechanical properties and processability. Several studies have demonstrated the fabrication of CNFconductive polymer composites based on either in-situ polymerization or ex-situ incorporation of conductive polymers in CNFs as biotemplates (Chen et al., 2015; Gou, He, Mo, & Zhao, 2015;

3

Kaitsuka, Hayashi, Shimokawa, Togawa, & Goto, 2016; Müller et al., 2016; Wang et al., 2016; Yu, Chen, Chen, & Liu, 2014; Zhang, Wu, Lu, & Zhou, 2015). However, in-situ polymerization reactions can damage the CNF matrix and are not easily controlled inside nanostructured matrices. To overcome these limitations, Park et al. recently reported the fabrication of a CNF-based conductive composite by the ex-situ incorporation of poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) into never-dried bacterial cellulose sheets (Khan, UI-Islam, Khattak, Ullah, & Park, 2015). The bacterial cellulose sheets were immersed in the PEDOT:PSS solution for around 12 h and then dried at room temperature. While no chemical reaction is required for the synthesis of these composites, this method is time-consuming and makes it difficult to precisely control the structure of conductive polymers in the CNF matrix. More recently, Jiang et al. reported the development of a heterogeneous composite of a synthetic fiber (aramid) and PEDOT:PSS (Li et al., 2016). This composite was made via vacuum filtration and had high stability and good flexibility. In this paper, to the best of our knowledge, we present the first attempt to develop a selfstanding, highly flexible, and conductive bilayer composite film. This film consists of a PEDOT:PSS membrane upper layer and a CNF sub-layer. We chose to use PEDOT:PSS and CNFs as the building materials and vacuum-assisted filtration (VF) as the manufacturing process for conductive composite film, because: (1) PEDOT:PSS, which is commercially available in aqueous dispersions, is environmentally stable, highly conductive, and allows for aqueous solution-based processing; (2) CNFs have excellent flexibility, low density, high mechanical strength, and good stability; and (3) VF has several advantages such as simple and rapid processing without a time-consuming and waterconsuming dialysis step, relatively low-cost process, and precise control over the layer-by-layer structure and component contents (Hyder et al., 2014). Our study is significant in that the fast, lowcost, scalable, and eco-friendly approach described here may expand the ability to create selfstanding, highly flexible, and conductive composite films for the development of next-generation soft electronic devices.

4

2. Materials and Methods 2.1. Materials Poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) and cellulose fibers (C6288, cotton linters) were purchased from Sigma-Aldrich and used without further purification. Dimethyl sulfoxide (DMSO) and sodium hydroxide (NaOH) were purchased from Duksan Pure Chemicals Company Co., Ltd. (Korea). Membrane filter paper was purchased from Advantec Co., Ltd. (Japan). 2.2. Preparation of cellulose nanofiber (CNFs) Cellulose fiber (25 g) was added to a 2% NaOH solution (1.2 L) and stirred for 3 h at room temperature. After washing the treated cellulose with distilled water up to pH 7, cellulose was loaded in 5 L of water and then high-pressure homogenization (Nano Disperser-NLM100, Ilshin Autoclave Co. Ltd., 25 passes at 1200 bar) was performed to produce CNFs. The CNF solution was stored at room temperature prior to fabrication of the composite films. Because of the good dispersion stability of the CNF solution, CNFs were not aggregated, even when kept at room temperature for a long time. 2.3. Preparation of two different conductive composite films (BF and SF) PEDOT:PSS was dispersed in deionized water at various concentrations of 0.4, 0.6, 0.8, and 1 wt%. According to the different PEDOT:PSS concentrations, DMSO (5 wt% of the PEDOT:PSS concentration) was added and thoroughly vortexed for several minutes. To fabricate a bilayer composite paper (BF) via two-step filtration, the CNF solution (5 mL, 0.42 wt%, average diameter of 27 nm as measured via SEM images), which was fabricated from cellulose fibers (CFs) using a highpressure homogenizer, was first poured into a PTFE filter membrane (0.2 μm pore size with 47 mm diameter, Advantec) and then vacuum-filtrated. Formation of the CNF layer was observed after 20

5

min (Figure 1C) because the vacuum pulls the liquid down. Next, the PEDOT:PSS solution (2 mL; 0.4, 0.6, 0.8, and 1 wt% in water; with/without the DMSO pretreatment) was added on top of the CNF layer and then subjected to vacuum filtration for 30 min. After drying at 75 °C for 30 min, the resulting bilayer PEDOT:PSS-CNF composite film was peeled off of the filter membrane. To prepare a single-layer composite film (SF), the mixture solution of the CNF solution (5 mL, 0.42 wt% in water) and PEDOT: PSS (2 mL; 0.4, 0.6, 0.8, and 1 wt% in water; with/without the DMSO pretreatment) was strongly vortexed to form a stable dispersion. This suspension solution was then vacuum-filtrated on a PTFE filter membrane for 30 min. After drying at 75 °C for 30 min, the resulting single-layer PEDOT:PSS-CNF composite film (SF) was peeled off from the filter membrane. The thicknesses of the PEDOT:PSS layer on the BF and SF (PEDOT:PSS/CNF layer) were 3 and 15 μm, respectively. 2.4. Quantitative analysis of PEDOT:PSS weight percentages in BF and SF The weight of the pure CNF paper was measured after vacuum filtration for 20 min and dried overnight at 105 °C. After preparing and drying both of the composite films with various PEDOT:PSS concentrations (0.4, 0.6, 0.8, and 1.0 wt%), the weight percentages of PEDOT:PSS present in the resultant BF and SF were calculated using the fraction of the weight of the composite films and the weight of the pure CNF film. 2.5. Characterization and measurements The pure CNF paper, BF, and SF were observed via field emission scanning electron microscopy (FE-SEM, Hitachi S-4200, Carl Zeiss, model Merlin) and atomic force microscopy (AFM, XE-70, Park Systems) for structural and morphological analysis. For FE-SEM imaging, the specimen was fixed on a metal stub using carbon tape and was then coated with platinum (Pt) using a sputter coater (Q150RS, TESCAN KOREA). A FE-SEM operating at 3 Kv was used to obtain secondary electron images of the sample surfaces. The AFM images were obtained using the noncontact mode in air conditions with silicon cantilevers (PPP-NCHR 10M), as delivered by Park

6

Systems. The drive frequency of the cantilever was 288.68-296.62 kHz. Scanning sizes of 1.5 μm × 1.5 μm were implemented for at least three different areas of each sample. The rootmean-square (rms) roughness was determined over a 1.5 μm × 1.5 μm area using the Park Systems XEI software. To measure the resistance (Ω), composite samples were prepared at sizes of 2.5 cm × 2.5 cm. Measurements were carried out using a two-point probe method from -1 V to 1 V with an electrical measurement device (PGSTAT204, Metrohm Autolab). The sheet resistance was measured via a four-point probe (SP4-40085TFS, SIGNATONE) method using a sheet resistance tester (CMT-100S, Advanced Instrument Technology) with a threshold detection limit of 2 MΩ/sq. The conductivity of BF was measured by only considering the thickness of the PEDOT:PSS layer, while that of SF was measured with the total thickness of PEDOT:PSS and the CNFs. The film thickness was measured using a thickness measurement device (2109S-10, Mitutoyo, Japan). Thermogravimetric (TGA) analysis was conducted by a thermogravimetric analyzer (TGA N-1000, Scinco Co. Ltd.) with pure CNF paper, PEDOT:PSS, BF, and SF from 25 °C to 620 °C at a heating rate of 10 °C/min in a nitrogen flow of 50 mL/min. The tape test was carried out by attaching 3M tape to the BF and SF samples and then removing it. The results reported here are the average values of at least three experiments.

2.6. Bending, stability, and LED bulb test In the bending cycle test, BF and SF were bent with a bending angle of 180° and then stretched; this was repeated for 100 cycles. The resistance of the samples was measured by an electrical measurement device after 20, 40, 60, 80, and 100 cycles; these values were subsequently compared to the initial resistance value. In the stability test, the paper samples were placed in ambient conditions for 5 days. The resistances were measured by an electrical measurement device after 1, 2, 3, 4, and 5 days and then compared to the initial value. For the LED bulb test, electrodes were connected to the surface of the

7

samples in a folded or unfolded state. A potential was applied via an electrochemical measurement device using chronoamperometry (potential: 3 V). 2.7. Preparation of PEDOT:PSS patterned composite film A PTFE membrane filter mask with 7 dot patterns (5 mm in diameter) was fabricated with eyelet punch and then used to create PEDOT:PSS patterned composite films. After fabricating a CNF layer by VF, a filter mask was carefully placed on a CNF layer, and then PEDOT:PSS solution was poured onto a filter mask attached to the CNF layer before being subjected to vacuum filtration for 30 min. The filter mask is hydrophilic enough to tightly adhere onto the CNF layer. The PEDOT:PSS patterned composite film was dried at 75 °C in the oven and peeled off from the filter membrane.

3. Results and Discussion In this study, a facile strategy is described for fabricating flexible conductive composite films based on PEDOT:PSS and CNFs. In contrast to previous studies that focused on either in-situ polymerization or ex-situ incorporation of conductive polymers inside the insulating CNF matrix, both of which resulted in heterogeneous composites, the goal of this study was to construct a bilayer conductive composite film composed of an upper PEDOT:PSS layer and a CNF matrix sub-layer. This is because the electrical conductivity of heterogeneous composites has been limited by the use of an insulating matrix. 3.1. Fabrication of PEDOT:PSS/CNF composite films Figures 1A show the chemical structures of PEDOT:PSS and CNFs, which can be stably suspended in aqueous solutions (Figure 1 B-i, ii). Additionally, the mixture of the PEDOT:PSS and CNF solution appeared to be stably dispersed in the aqueous solution (Figure 1B-iii). Figures 1C and D show schematic illustrations of the fabrication of the two different types of conductive composite films. The bilayer composite conductive film (BF) was fabricated in two steps (Figure 1C). In addition to the

8

BF sample described in Figure 1C, we fabricated a single-layer composite film (SF) for comparison. The SF sample was simply fabricated in one step with the mixture solution of PEDOT:PSS and CNFs via VF (Figures 1B-iii and D). Compared with the pure CNF paper (Figure 1C), both BF and SF showed an obvious dark blue color. These samples are large in size, smooth, flexible, and can be easily cut into any shape or size (without the brittleness of the PEDOT:PSS layer) by a blade or scissors. Figure 2 shows FE-SEM images of the CFs, CNFs, and two different samples of PEDOT:PSS-CNF composite films (BF and SF). As seen from the images of the CFs and CNFs (Figure 2A), the original CFs have an irregular shape and dimensions ranging from a few tens of micrometers to hundreds of micrometers before treatment with the high–pressure homogenizer. However, after homogenization, the CNF film exhibits continuous nanofiber network structures with a diameter in the range of 20-50 nm. Interestingly, as shown in the images of the top/bottom side of BF (Figure 2B), a smooth surface morphology without any fiber structure was observed on the top side while the traditional CNF morphology was observed on the bottom side. These images demonstrated that the PEDOT:PSS layer was separately deposited on top of the CNF layer, indicating a bilayer composite film. In contrast to BF, a novel morphology was observed on both the top and bottom sides of SF, indicating a single-layer composite film. We further studied and quantified the difference in the morphologies of CNF, BF, and SF by AFM analysis (Figure 3). Similar to the FE-SEM images, the surface of the CNF film shows an interconnected nanofiber morphology with a mean roughness (Rm) of 19.11 nm, while the top side of BF exhibits a much smoother morphology with an Rm of 5.06 nm. The novel surface morphology with Rm of 16.10 nm in SF might be the result of a thick overcoating of PEDOT:PSS on CNF structures. 3.2. Adhesion, thermal stability and electrical properties of BF and SF We investigated the adhesion strength and thermal stability of BF and SF by the tape test and TGA analysis, respectively. The adhesion of PEDOT:PSS and CNFs (BF and SF) appeared to be strong

9

enough for practical handling in future applications (Figure 4A). The stable adhesion between PEDOT:PSS and CNF might be due to the strong electrostatic and hydrogen bonding interactions, which originated from the hydroxyl groups of CNFs and the charged PEDOT:PSS (Khan et al., 2015; Li et al., 2016). Figure 4B shows a comparison of the weight loss of the two different composite films as a function of temperature along with the pure PEDOT:PSS and CNF samples. In all cases, gradual weight loss was observed between 50 and 110 °C due to the loss of adsorbed water molecules. The BF sample revealed little weight loss below 230 °C, indicating a wide processing/operating temperature range, which is critical for a variety of applications. Above 230 and 300 °C, there was sharp weight loss due to the decomposition of PSS dopants and CNFs, respectively (Liu et al., 2015). Compared with BF, the curve of SF exhibited a similar but somewhat faster weight loss process; this difference is likely due to its heterogeneous structure (Boutou et al., 2015; Schild, 1993). In general, DMSO has been used as a second dopant to improve the conductivity of PEDOT:PSS because it has the ability to induce the phase separation of excessive insulating PSS from the PEDOT:PSS domain and improve the crystallinity of PEDOT (π-stacking) (Crispin et al., 2006; Hohnholz, Okuzaki, & MacDiarmid, 2005; Reyes-Reyes, Cruz-Cruz, & López-Sandoval, 2010; Wang et al., 2015). Given the previous promising results for increasing the conductivity of PEDOT:PSS, we sought to explore the effects of DMSO on the electrical properties of composite films (BF and SF). Thus, we fabricated composite films with and without the DMSO pretreatment and then evaluated the sheet resistance (Figure 5). As shown in Figures 5A and B, the linear I-V characteristics revealed that the sheet resistances of composite films with the DMSO pretreatment (BF+DMSO and SF+DMSO) were significantly lower than those of the other composite films without the DMSO pretreatment (BFDMSO and SF-DMSO). The reason for the largest error bars on SF-DMSO is presumably due to the

10

heterogeneous PEDOT:PSS-CNF in the surface composition without DMSO treatment. In the case of composite films without DMSO pretreatment (BF-DMSO and SF-DMSO), SF-DMSO exhibited a lower sheet resistance compared to BF-DMSO. This might be because the CNFs mixed with PEDOT:PSS lessens the negative impacts of PSS by causing the phase deformation of the PEDOT:PSS domain in the filtration process. More importantly, BF+DMSO exhibited the lowest sheet resistance of 150±23 Ω/sq, while SF+DMSO showed a somewhat higher sheet resistance of 370±28 Ω/sq. To further examine the electrical properties of BF and SF as a function of the PEDOT:PSS contents, we fabricated composites with four different PEDOT:PSS solutions of varying concentrations. The PEDOT:PSS contents in the BF and SF could be tuned by simply changing the PEDOT:PSS concentration. As shown in Table S1, the weight percentages of PEDOT:PSS in both BF and SF were gradually increased to 48.7 and 50.0 wt%, respectively, by increasing the PEDOT:PSS concentration (0.4, 0.6, 0.8, and 1%). Interestingly, this quantitative analysis confirmed that both BF and SF were fabricated with very little loss of the initial amount of PEDOT:PSS during VF processing. We then evaluated the sheet resistance of these samples using current-voltage (I-V) curves (Figure 6). The linear I-V characteristics indicated that the sheet resistances of BF and SF gradually decreased with an increase in the PEDOT:PSS concentration (see Figures 6 A-C). As seen in Figure 6D, BF exhibited a conductivity that was more than 10 times higher than that of SF (ca. 22.6 S cm-1 vs. ca. 1.8 S cm-1). The dramatic difference in the sheet resistance and conductivity between BF and SF may be due to the formation of different PEDOT:PSS structures. In contrast to SF, the two sequential filtrations for BF lead to a homogeneous PEDOT:PSS layer that was separately deposited on the CNF layer (BF, see Figure 2B), which appears to be more electrically functional. More importantly, the resultant BF exhibited an electrical conductivity (ca. 22.6 S cm-1) that is competitive with previously reported conductivity values of cellulose-based composites (ca. 10-5 to 12 S cm-1, see Table S2) (Feng, Zhang, Shen, Yoshino, & Feng, 2012; Kaitsuka et al., 2016; Khan et al., 2015; Khan et al., 2015; Li et al., 2017; Qi et al., 2015; Shi et al., 2013; Xu et al., 2013; Zhang et al., 2015).

11

3.3. Flexibility and stability analysis of BF and SF To further demonstrate the mechanical flexibility of BF and SF, we measured changes in the electrical resistance during a bending test where both composite films underwent bending at an angle of 180° and subsequent stretching; this was repeated up to 100 times. As shown in Figure 7A, the electrical resistance is nearly unchanged before and after 100 cycles. The LED test confirmed that BF and SF were still electrically functional after the bending test. Even though both BF and SF were bent multiple times (four times) to nearly 90° at the same time, there was no significant change in the electrical resistance of BF and SF (Figures 7C and D). The electrical conductivities of these films were sufficient to turn on an LED bulb, even after the samples were folded eight times. In these flexibility tests, the LED light on BF was much brighter compared to the bulb on SF (see Figures 7A, C, and D), indicating the better electrical properties of BF. In addition to the excellent bending durability of BF and SF films, these films also revealed good stability in ambient conditions. As shown in Figure 7B, the initial electrical resistance of both films was almost unchanged after 5 days. It is worth noting that there were no obvious differences in the flexibility and stability results for BF and SF. 3.4. Patterning of PEDOT:PSS on CNF paper Patterning of conductive materials on flexible substrates is of considerable technological importance in a variety of applications (Kim, You, & Kim, 2010; Kim et al., 2016; Mu et al., 2015). Thus, we attempted to fabricate PEDOT:PSS patterned composite film. As shown in Figure 8, the PEDOT:PSS patterns with dot arrays of 5 mm diameter were successfully created on top of the CNF layer by using a filter mask. This patterning method does not require any surface modification of cellulose films such as wax printing. Overall, the analyses of the sheet resistance, conductivity, flexibility, and stability strongly indicate that a BF can function as a versatile, self-standing, highly flexible, lightweight, and highly conductive film for soft electronics systems.

12

4. Conclusions We developed a self-standing, flexible, lightweight, conductive bilayer nanocomposite film via simple vacuum-assisted filtration processing without a time-consuming dialysis step. By combining electrostatic/hydrogen interactions along with a vacuum force, nanocomposite films of PEDOT:PSS and CNFs were strongly assembled in an hour. A BF was successfully fabricated via two-step filtration. In comparison with SF, the BF samples exhibited an electrical conductivity that was more than 10 times as high. Notably, this composite film showed little change in its resistance when bent at an angle of 180° after 100 cycles or bent multiple times at an angle of 90° angle. This indicates the excellent mechanical flexibility of the sample, which is an important requirement for highperformance soft electronic systems. We envisage that the strategy described herein will enhance our ability to create highly efficient conductive films for a variety of applications. Acknowledgements This study was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) (No. 2015R1C1A1A01054258). References Aiyer, S., Prasad, R., Kumar, M., Nirvikar, K., Jain, B., & Kushwaha, O. S. (2016). Fluorescent carbon nanodots for targeted in vitro cancer cell imaging. Applied Materials Today, 4, 71-77. Boissou, F., Vigier, K. D. O., Estrine, B., Marinkovic, S., & Jérôme, F. (2014). Selective depolymerization of cellulose to low molecular weight cello-oligomers catalyzed by betaïne hydrochloride. ACS Sustainable Chemistry & Engineering, 2(12), 2683-2689. Boutou, A. K., Zoumot, Z., Nair, A., Davey, C., Hansell, D. M., Jamurtas, A., et al. (2015). The impact of homogeneous versus heterogeneous emphysema on dynamic hyperinflation in patients with severe

13

COPD assessed for lung volume reduction. Journal of Chronic Obstructive Pulmonary Disease, 12, 598-605. Chen, C., Yang, C., Li, S., & Li, D. (2015). A three-dimensionally chitin nanofiber/carbon nanotube hydrogel network for foldable conductive paper. Carbohydrate Polymers, 134, 309-313. Chen, C., Zhang, T., Zhang, Q., Feng, Z., Zhu, C., Yu, Y., et al. (2015). Three-dimensional BC/PEDOT composite nanofibers with high performance for electrode–cell interface. ACS Applied Materials & Interfaces, 7(51), 28244–28253. Chen, J., Xu, J., Wang, K., Qian, X., & Sun, R. (2015). Highly thermostable, flexible, and conductive films prepared from cellulose, graphite, and polypyrrole nanoparticles. ACS Applied Materials & Interfaces, 7, 15641−15648. Crispin, X., Jakobsson, F. L. E., Crispin, A., Grim, P. C. M., Andersson, P., Volodin, A., van Haesendonck, C., et al. (2006). The origin of the high conductivity of poly(3,4ethylenedioxythiophene)−poly(styrenesulfonate) (PEDOT−PSS) plastic electrodes. Chemistry of Materials, 18(18), 4354-4360. Feng, Y., Zhang, X., Shen, Y., Yoshino, K., & Feng, W. (2012). A mechanically strong, flexible and conductive film based on bacterial cellulose/graphene nanocomposite. Carbohydrate Polymers, 87(1), 644–649. Gou, H., He, J., Mo, Z., & Zhao, Z. (2015). Ultrasonic preparation of cellulose/Ag/polyaniline conductive composites and its electrical properties. Journal of Materials Science: Materials in Electronics, 26(9), 7295-7302. Hohnholz, D., Okuzaki, H., & MacDiarmid, A. G. (2005). Plastic electronic devices through line patterning of conducting polymers. Advanced Functional Materials, 15(1), 51-56. Hon, D. N.-S. (1994). Cellulose: a random walk along its historical path. Cellulose, 1, 1-25.

14

Hosoda, N., Tsujimoto, T., & Uyama, H. (2014). Green composite of poly(3-hydroxybutyrate-co-3hydroxyhexanoate) reinforced with porous cellulose. ACS Sustainable Chemistry & Engineering, 2(2), 248–253. Hu, L., Zheng, G., Yao, J., Liu, N., Weil, B., Eskilsson, M., et al. (2013). Transparent and conductive paper from nanocellulose fibers. Energy & Environmental Science, 6, 513-518. Hyder, M. N., Kavian, R., Sultana, Z., Saetia, K., Chen, P.-Y., Lee, S. W., et al. (2014). Vacuum-assisted layer-by-layer nanocomposites for self-standing 3D mesoporous electrodes. Chemistry of Materials, 26(18), 5310-5318. Jiang, L., Lu, Y., Liu, X., Tu, H., Zhang, J., Shi, X., et al. (2015). Layer-by-layer immobilization of quaternized carboxymethyl chitosan/organic rectorite and alginate onto nanofibrous mats and their antibacterial application. Carbohydrate Polymers, 121, 428-435. Kaitsuka, Y., Hayashi, N., Shimokawa, T., Togawa, E., & Goto, H. (2016). Synthesis of polyaniline (PANI) in nano-reaction field of cellulose nanofiber (CNF), and carbonization. Polymers, 8, 40. Kang, J.-W., Jeong, W.-I., Kim, J.-J., Kim, H.-K., Kim, D.-G., & Lee, G.-H. (2007). High-performance flexible organic light-emitting diodes using amorphous indium zinc oxide anode. Electrochemical and Solid-State Letters, 10 (6), J75-J78. Khan, S., UI-Islam, M., Khattak, W. A., Ullah, M. W., & Park, J. K. (2015). Bacterial cellulose–poly(3,4ethylenedioxythiophene)–poly(styrenesulfonate) composites for optoelectronic applications. Carbohydrate Polymers, 127, 86-93. Khan, S., Ul-Islam, M., Ullah, M. W., Kim, Y., & Park, J. K. (2015). Synthesis and characterization of a novel bacterial cellulose–poly(3,4-ethylenedioxythiophene)–poly(styrene sulfonate) composite for use in biomedical applications. Cellulose, 22(4), 2141-2148.

15

Kim, D., Kim, J., Ko, Y., Shim, K., Kim, J. H., & You, J. (2016). A facile approach for constructing conductive polymer patterns for application in electrochromic devices and flexible microelectrodes. ACS Applied Materials & Interfaces, 8, 22175-33182. Kim, J., You, J., Kim, B., Park, T., & Kim, E. (2011). Solution processable and patternable poly(3,4alkylenedioxy-thiophene)s for large area electrochromic. Advanced Materials, 23, 4168-4173. Kim, J., You, J., & Kim, E. (2010). Flexible conductive polymer patterns from vapor polymerizable and photo cross-linkable EDOT. Macromolecules, 43, 2322-2327. Kushwaha, O. S., Avadhani, C. V., & Singh, R. P. (2015). Preparation and characterization of selfphotostabilizing UV-durable bionanocomposite membranes for outdoor applications. Carbohydrate Polymers, 123, 164-173. Li, G., Tian, X., Xu, X., Zhou, C., Wu, J., Li, Q., et al. (2017). Fabrication of robust and highly thermally conductive nanofibrillated cellulose/graphite nanoplatelets composite papers. Composites Science and Technology, 138, 179-185. Li, Y., Ren, G., Zhang, Z., Teng, C., Wu, Y., Lu, X., et al. (2016). A strong and highly flexible aramid nanofibers/PEDOT:PSS film for all-solid-state supercapacitors with superior cycling stability. Journal of Materials Chemistry A, 4, 17324-17332. Liu, Y., Weng, B., Razal, J. M., Xu, Q., Zhao, C., Hou, Y., et al. (2015). High-performance flexible allsolid-state supercapacitor from large free-standing graphene-PEDOT/PSS films. Scientific Reports, 5, 17045. Malti, A., Edberg, J., Granberg, H., Khan, Z. U., Andreasen, J. W., Liu, X., et al. (2016). An organic mixed ion–electron conductor for power electronics. Advanced. Science, 3, 1500305.

16

Mengistie, D. A., Chen, C.-H., Boopathi, K. M., Pranoto, F. W., Li, L.-J., & Chu, C.-W. (2015). Enhanced thermoelectric performance of PEDOT:PSS flexible bulky papers by treatment with secondary dopants. ACS Applied Materials & Interfaces, 7(1), 94-100. Mu, J., Hou, C., Wang, H., Li, Y., Zhang, Q., & Zhu, M. (2015). Origami-inspired active graphene-based paper for programmable instant self-folding walking devices. Science Advances, 1, e1500533. Müller, D., Cercená, R., Gutiérrez Aguayo, A. J., Porto, L. M., Rambo, C. R., & Barra, G. M. O. (2016). Flexible PEDOT-nanocellulose composites produced by in situ oxidative polymerization for passive components in frequency filters. Journal of Materials Science: Materials in Electronics, 27(8), 8062– 8067. Nataraj, S. K., Kim, B. H., Yun, J. H., Lee, D. H., Aminabhavi, T. M., & Yang, K. S. (2009). Morphological characterization of electrospun carbon nanofiber mats of polyacrylonitrile containing heteropolyacids. Synthetic Metals, 159, 1496-1504. Nataraj, S. K., Kim, B. H., Yun, J. H., Lee, D. H., Aminabhavi, T. M., & Yang, K. S. (2009). Effect of added nickel nitrate on the physical, thermal and morphological characteristics of polyacrylonitrile-based carbon nanofibers. Materials Science and Engineering B, 162, 75-81. Nataraj, S. K., Kim, B. H., Yun, J. H., Lee, D. H., Aminabhavi, T. M., & Yang, K. S. (2011). Pore characteristics and electrochemical properties of the carbon nanofibres of polyacrylonitrile containing iron-oxide by electrospinning. International Journal of Nanotechnology, 8, 868-876. Nataraj, S. K., Yang, K. S., & Aminabhavi, T. M. (2012). Polyacrylonitrile-based nanofibers—A state-ofthe-art review. Progress in Polymer Science, 37, 487-513. Punetha, V. D., Rana, S., Yoo, H. J., Chaurasia, A., McLeskey, J. T., Ramasamy, M. S., et al. (2017). Functionalization of carbon nanomaterials for advanced polymer nanocomposites: A comparison study between CNT and graphene. Progress in Polymer Science, 67, 1-47.

17

Qi, H., Schulz, B., Vad, T., Liu, J., Mader, E., Seide, G., et al. (2015). Novel carbon nanotube/cellulose composite fibers as multifunctional materials. ACS Applied Materials & Interfaces, 7, 22404-22412. Reyes-Reyes, M., Cruz-Cruz, I., & López-Sandoval, R. (2010). Enhancement of the electrical conductivity in PEDOT:PSS films by the addition of dimethyl sulfate. Journal of Physical Chemistry C, 114(47), 20220–20224. Salam, A., A Lucia, L., & Jameel, H. (2013). A novel cellulose nanocrystals-based approach to improve the mechanical properties of recycled paper. ACS Sustainable Chemistry & Engineering, 1(12), 15841592. Salas, C., Nypelö, T., Rodriguez-Abreu, C., Carrillo, C., & Rojas, O. J. (2014). Nanocellulose properties and applications in colloids and interfaces. Current Opinion in Colloid & Interface Science, 19, 383396. Schild, H. G. (1993). Application of TGA/FTIR to the thermal degradation mechanism of tetrafluoroethylene–propylene copolymers. Journal of Polymer Science: Part A: Polymer Chemistry, 31, 1629-1632. Shi, Z., Phillips, G. O., & Yang, G. (2013). Nanocellulose electroconductive composites. Nanoscale, 5, 3194-3201. Singh, R. P. & Kushwaha, O. S. (2013). Polymer solar cells: an overview. Macromolecular Symposia, 327, 128-149. Song, M., You, D. S., Lim, K., Park, S., Jung, S., Kim, C. S., et al. (2013). Highly efficient and bendable organic solar cells with solution-processed silver nanowire electrodes. Advanced Functional Materials, 23, 4177–4184.

18

Tkalya, E., Ghislandi, M., Thielemans, W., van der Schoot, P., de With, G., & Koning, C. (2013). Cellulose nanowhiskers templating in conductive polymer nanocomposites reduces electrical percolation threshold 5-fold. ACS Macro Letters, 2, 157-163. Valtakari, D., Liu, J., Kumar, V., Xu, C., Toivakka, M., & Saarinen, J. J. (2015). Conductivity of PEDOT:PSS on spin-coated and drop cast nanofibrillar cellulose thin films. Nanoscale Research Letters, 10, 386. Wang, L., Jiang, F., Xiong, J., Xu, J., Zhou, W., Liu, C., et al. (2015). Effects of second dopants on electrical conductivity and thermopower of poly(3,4ethylenedioxythiophene):poly(styrenesulfonate)-filled carbon black. Materials Chemistry and Physics, 153, 285-290. Wang, Z., Tammela, P., Huo, J., Zhang, P., Strømme, M., & Nyholm, L. (2016). Solution-processed poly(3,4-ethylenedioxythiophene) nanocomposite paper electrodes for high-capacitance flexible supercapacitors. Journal of Materials Chemistry A, 4, 1714-1722. Xu, J., Zhu, L., Bai, Z., Liang, G., Liu, L., Fang, D., et al. (2013). Conductive polypyrrole–bacterial cellulose nanocomposite membranes as flexible supercapacitor electrode. Organic Electronics, 14(12), 3331-3338. Yu, H., Chen, P., Chen, W., & Liu, Y. (2014). Effect of cellulose nanofibers on induced polymerization of aniline and formation of nanostructured conducting composite. Cellulose, 21(3), 1757-1767. Zhang, X., Wu, X., Lu, C., & Zhou, Z. (2015). Dialysis-free and in situ doping synthesis of polypyrrole@cellulose nanowhiskers nanohybrid for preparation of conductive nanocomposites with enhanced properties. ACS Sustainable Chemistry & Engineering, 3(4), 675-682.

19

Zhou, S., Wang, M., Chen, X., & Xu, F. (2015). Facile template synthesis of microfibrillated cellulose/polypyrrole/silver nanoparticles hybrid aerogels with electrical conductive and pressure responsive properties. ACS Sustainable Chemistry & Engineering, 3(12), 3346-3354. Zhou, Z., Peng, X., Zhong, L., Wu, L., Cao, X., & Sun, R. C. (2016). Electrospun cellulose acetate supported Ag@AgCl composites with facet-dependent photocatalytic properties on degradation of organic dyes under visible-light irradiation. Carbohydrate Polymers, 136, 322-328. Zhu, H., Xiao, Z., Liu, D., Li, Y., Weadock, N. J., Fang, Z., et al. (2013). Biodegradable transparent substrates for flexible organic-light-emitting diodes. Energy & Environmental Science, 6, 2105-2111.

Figure 1. Schematic illustration showing the fabrication of conductive composites based on PEDOT:PSS and CNFs via VF. (A) Chemical structures of PEDOT:PSS and cellulose. (B) Images of (i) PEDOT:PSS, (ii) CNFs, and (iii)

20

PEDOT:PSS/CNF mixture stably dispersed in aqueous solutions. Schematic illustrations of the fabrication of (C) the BF and (D) the SF via VA. The BF and SF are 4 cm in diameter.

Figure 2. SEM image of the (A) original CFs before high-pressure homogenization and CNF paper after highpressure homogenization, (B) a BF, and (C) a SF. The concentrations of PEDOT:PSS and CNF dispersed in aqueous solutions were 1 and 0.42 wt%.



21

Figure 3. AFM image of (A) pure CNF paper, (B) BF, and (C) SF. The concentrations of PEDOT:PSS and CNFs dispersed in the aqueous solution were 1 and 0.42 wt%. The measurement proceeded at a scan rate of 0.3 Hz and a range of 1.5 μm × 1.5 μm. Scale bar: 400 nm.

Figure 4. (A) Demonstration of the stability of the BF and SF via the tape test (samples were 4 cm in diameter). (B) Thermogravimetric analysis (TGA) plots of pure CNF paper and PEDOT:PSS and the two different types of composite films (BF and SF).

Figure 5. (A) Current-voltage (I-V) characteristics of the different types of composite films both with DMSO pretreatment (BF+DMSO and SF+DMSO) and without the DMSO pretreatment (BF-DMSO and SF-DMSO). The inset shows composite films without the DMSO pretreatment (BF-DMSO and SF-DMSO, current range from -

22

0.4 to +0.4 mA). (B) Sheet resistance of BF and SF with/without the DMSO treatment. The PEDOT:PSS (1 wt%)dispersed solution contained 5 wt% DMSO. The size of all composite films samples is 2.5 cm × 2.5 cm.

Figure 6. Current-voltage (I-V) characteristics of (A) a BF+DMSO and (B) a SF+DMSO as a function of the PEDOT:PSS concentration. (C) Sheet resistance of BF+DMSO and SF+DMSO as a function of the PEDOT:PSS concentration. (D) Conductivity of BF and SF with DMSO pretreatment and PEDOT:PSS 1 wt% concentration.

23

Figure 7. Demonstrations of the (A) mechanical flexibility and (B) stability in ambient conditions of the two different types of films (BF and SF). The changes in resistance were measured by an electrochemical device and compared to the initial value. The insets show (i) a bending angle of 180 o and (ii) LED emission after 50 folding cycles. Size: 1.2 cm × 1.2 cm. Current-voltage (I-V) characteristics of (C) BF and (D) SF with different folding times. The insets show the light emission of the LED bulb before and after folding of the film samples. Size: 5.5 cm × 3.0 cm.

24

Figure 8. Fabrication of PEDOT:PSS dot patterns on CNF film. (A) Schematic illustration showing the fabrication of PEDOT:PSS patterned composite films using a filter mask via VF. (B) Photograph images of PEDOT:PSS patterns with dot arrays of 5 mm diameter on composite film.

25