- Email: [email protected]
A facile and scalable approach to fabricating free-standing polymer—Carbon nanotube composite electrodes
Synthetic Metals 215 (2016) 35–40
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
A facile and scalable approach to fabricating free-standing polymer— Carbon nanotube composite electrodes Margarita R. Arcila-Veleza,1, Robert K. Emmetta,1, Mehmet Karakayab,c, Ramakrishna Podilab,c , Kryssia P. Díaz-Orellanaa , Apparao M. Raob,c , Mark E. Robertsa,c,* a b c
Department of Chemical & Biomolecular Engineering, Clemson University, Clemson, SC 29634, United States Department of Physics & Astronomy, COMSET, Clemson University, Clemson, SC 29634, United States Clemson Nanomaterials Center, Clemson University, Clemson, SC 29634, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 23 November 2015 Received in revised form 5 January 2016 Accepted 2 February 2016 Available online 15 February 2016
Nanoporous carbon materials are widely utilized in high-power supercapacitors due to their structural properties, chemical stability and conductivity, despite their limited energy density and charge storage capacity. Conducting polymers, on the other hand, possess high charge capacities; however, their application in commercial devices is hindered by degradation arising from their poor chemical and physical stability. Composites of carbon nanomaterials and conducting polymers have synergistic properties beneficial to supercapacitors, such as high capacitance and stability, but the limitations in scalable synthesis and polymer aggregation prevent widespread utilization. In this work, robust freestanding carbon nanotube (CNT)/electrically conducting polymer (ECP) electrodes are prepared using a simple dispersion filtration method, which can easily be scaled up. This process eliminates the use of binder, substrate or additional inactive weight. Composite CNT/ECP electrodes showed enhanced capacitance and charge capacity, achieving values up to 448 F/g and 84 mAh/g compared to 27 F/g and 10 mAh/g for pure CNT electrodes in aqueous electrolyte. Resulting symmetric cells exhibited energy and power densities of 5 Wh/kg and 283 W/kg, respectively, in aqueous electrolytes; and 12 Wh/kg and 744 W/kg, respectively, in organic electrolytes when using PEDOT/CNT electrodes. Given the process simplicity, relatively low cost and high throughput, the present composites have great potential for largescale manufacturing of CP/CNTs supercapacitor electrodes. ã 2016 Elsevier B.V. All rights reserved.
Keywords: Supercapacitor Conducting polymer Carbon nanotubes Scalable synthesis Filtration
1. Introduction Rapid growth in the demand for electrical energy storage continues to encourage the development of new materials and devices. With their ability to improve the performance of batteries or provide high-power, a recyclable and safe alternative such as supercapacitors have experienced tremendous growth in this industry. Supercapacitors, which typically comprise high-surface area, inert carbonaceous electrode materials, offer the benefits of high power density, rapid charge/discharge, and long lifetime. Even with these important advantages, carbon materials (e.g., activated carbon, carbon nanotubes, graphene) are limited by relatively low energy density as a consequence of their physical charge storage mechanism, electrical double layer capacitance
* Corresponding author at: Department of Chemical & Biomolecular Engineering, Clemson University, Clemson, SC 29631, United States. E-mail address: [email protected] (M.E. Roberts). 1 Both authors contributed equally to this work. http://dx.doi.org/10.1016/j.synthmet.2016.02.005 0379-6779/ ã 2016 Elsevier B.V. All rights reserved.
(EDLC), which is related to ion adsorption at the electrode surface. As we approach the maximum theoretical surface area in carbons, further increases in energy density are no longer attainable. On the other hand, electroactive materials such as metal oxides and electroactive conducting polymers (ECPs) can be prepared with relatively high surface areas, but are also able to store charge through Faradaic charge transfer in addition to EDLC. Metal oxides exhibit high Faradaic capacitance in aqueous solutions with good redox stability, but are limited by material cost, low operating voltage ranges, and the lack of scalable synthetic methods [1,2]. ECPs, such as polypyrrole (PPY), polyaniline (PANI), and poly (3,4-ethylenedioxythiophene) (PEDOT), provide a low-cost, large specific capacitance alternative to carbon materials and metal oxides [3–6] and can be easily prepared through various routes [7] from cheap monomers in large quantities. However, ECPs often suffer from poor cycle stability due to polymer over-oxidation stresses caused by volumetric changes that accompany the chargedischarge processes associated with ion insertion and removal [8]. To improve their performance and stability, significant efforts have
36
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
focused on creating micro or nanostructured ECP electrodes by modifying the synthesis processes, which enhances their mechanical and electrochemical properties [8–10]. Another approach to the same goal involves integrating a complementary material to reinforce and support the polymeric structure [11,12]. Carbon nanotubes (CNTs) are attractive for this purpose due to their excellent structural properties, electronic conductivity and chemical stability. CNTs improve the cycle life and mechanical properties of ECPs since CNTs adapt easily to changes in volume [13], and also introduce new electronic properties based on interactions between the two materials. Furthermore, the adaptation of such compositions helps overcome the low conductivity of ECPs in the neutral (or reduced) state and also increase electronic transport in thick electrodes [14]. Of equal importance, the addition of ECPs decreases CNT bundling and enhances the charge storage capacity compared to CNTs electrodes. In-situ chemical polymerization is the prevalent method for preparing ECP/CNT composites, which consists of oxidative polymerization of ECPs in a CNTsuspension and pressing the dry material into a pellet [15]. Through this approach, ECPs and CNTs are integrated on a molecular scale [16–20], increasing conductivity [21], although, homogeneity depends on the CNT dispersion and ECPs often prefer to grow on themselves rather than distributing homogeneously along the CNTs leading to polymer aggregation [17]. High specific capacitance values (500 F/g) [22] have been obtained by electrochemically polymerizing ECPs on CNTs due to strong interactions between the two materials [23]. Nevertheless, this method is limited to low amounts of polymer because as the film grows thicker it blocks electrolyte access to the CNTs. In melt compounding, CNTs are directly dispersed into a polymer melt; however, locally homogeneous states are difficult to achieve without breaking down the entangled CNTs [20]. A simple alternative to produce ECP/CNT composites involves mixing pre-synthesized ECP with CNTs [24,25], which is advantageous in terms of scalability and cost-effectivity [14]. Lota et al. [25] prepared composite electrodes by pressing a mixture of PEDOT and CNTs into pellets and obtained a single-electrode capacitance value of 95 F/g (at 2 mV/s), which possessed improved conductivity and cycle stability compared to PEDOT, albeit with significant diffusion limitations. In this work, we present a simple, low-cost approach to fabricating free-standing ECP/CNT composite electrodes using a sequential dispersion-filtration process. After ECP synthesis, ECP/ CNT dispersions (50/50 by wt.) are prepared in aqueous solutions containing sodium dodecyl sulfate (SDS) using ultrasonication, and then immediately filtered through a polyamide membrane with laboratory scale vacuum filtration. After electrodes are dried, robust, free-standing composite ECP/CNT paper electrodes are obtained, yielding high specific capacitance values up to 448 F/g (at 10 mV/s). In this process, no binder or pressure is needed to obtain efficient electrode performance. Considering the process simplicity, low cost and amenability to high throughput processing, our new approach to composite electrode fabrication provides a unique application to scalable, high power and energy supercapacitor electrode manufacturing.
synthesized by oxidative polymerization in aqueous HCl (1 M), using ammonium persulfate (APS) as an oxidant at 4 C for 2 h. PEDOT was synthesized using Fe(NO3)3 in an aqueous solution containing sodium poly(4-styrene sulfonate) for 6 h.
2. Materials and methods
where I is the discharge current and m is the electrode mass. Both calculations for capacitance are within 5%. Energy density, E, is calculated from
2.1. Materials Multi-walled carbon nanotubes were purchased from CheapTubes (dia.=30–50 nm, length = 10–20 mm) and NanoTech Labs (dia.=70–80 nm, length = 700–800 mm), and used as received. Polypyrrole (PPY), polyaniline (PANI), and Poly(3,4-ethylenedioxythiophene) (PEDOT) were synthesized using previously reported methods [26–28]. Briefly, PPY was synthesized by oxidative polymerization using FeCl3 in methanol at 0 C for 6 h. PANI was
2.2. Electrode fabrication The ECP/CNT composites were prepared by dispersing CNTs with an ECPs in aqueous solutions containing 1 wt.% SDS using a tip sonicator probe with a power of 75 W for 15 min. Subsequently, the suspension was poured onto a polyamide filtration membrane (Whatman, 0.45 mm pore diameter) and filtered using a vacuum filtration setup. The filtrate on the supporting filter membrane was washed with distilled water several times, and then oven dried at 60 C for 8 h. The resulting films were peeled off the membrane yielding free-standing ECP/CNT paper-like electrodes with good mechanical properties [28,29]. 2.3. Electrode characterization The physical properties of the composite electrodes were investigated using scanning electron microscopy (SEM, Hitachi SU6600) and energy dispersive X-ray spectroscopy (EDAX). Composite electrodes were electrochemically evaluated as single electrodes using a potentiostat (Princeton Applied Research VersaSTAT 4) in aqueous 3 M H2SO4 solutions and in 1.5 M tetraethylammonium tetrafluoroborate (TEABF4) in acetonitrile (MeCN). A Pt mesh was used as counter electrode, Ag/AgCl (sat’d KCl) and Ag/Ag+ (0.1 M AgNO3 in MeCN) electrodes were used as aqueous and non-aqueous pseudo-reference electrodes, respectively. Electrochemical characterization included cyclic voltammetry (CV) at scan rates of 10–300 mV s1, electrochemical impedance spectroscopy (EIS) (0.5 V DC bias, 0.1–10000 Hz, 20 mV RMS) and galvanostatic charge/discharge cycles (C/D) at a current such that the discharge time is equal to 60 s. The reported specific capacitance values were calculated from CV at a scan rate of 10 mV/s using the complete electrode mass (ECP + CNT). 2.4. Capacitance, energy and power calculations All the specific values of the electrodes are reported on a per mass basis (gravimetric) with respect to the total mass of the electrode (CNTs + ECP). The specific capacitance, C, of single electrodes measured in 3-electrode cells were determined by taking the average capacitance from CV measurements (current divided by scan rate and mass) over the full voltage range (at 10 mV s1). The capacitance of symmetric supercapacitors was calculated using the total mass of both electrodes (which were approximately equal) using two methods: first, by taking the average capacitance from CV measurements, as described for 3-electrode measurements; and second, from the inverse of the galvanostatic discharge slope (dV/dt) in the linear range, according to C¼
1 I dV m dt
1 E ¼ C DV 2 2
ð1Þ
ð2Þ
where DV is the voltage difference from the charged to discharged state. The power density, P, is calculated from P¼
DE Dt
ð3Þ
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
where Dt is the time required to completely discharge the device. 3. Results and discussion Low (300) magnification SEM images (not shown) show fairly homogenous composite films with a few CNT bundles. In general, the introduction of ECPs into the MWNT papers created a more openly porous network (Fig. 1) amenable to ion access to the electroactive materials and enhanced the Faradaic processes. From the SEM at higher magnification (Fig. 1a, >6000), carbon nanotube agglomerates are observed sporadically throughout the electrode. With the addition of the ECP (Fig. 1 b–d), it can be seen how the polymer acts as a binder and bridges the gaps between the CNT islands, enhancing the material’s electrical properties. From Fig. 1b and d, the heterogeneous structure of PANI and PPY can be seen as well as these polymers tendency to aggregate in some areas forming clusters. In the PANI/CNT composites, the particles of 0.2–0.3 mm in size merge to form agglomerates with diameters of 2–6 mm. The PPY/CNT samples show a homogeneous distribution with uniform particle sizes of
37
0.1–0.3 mm, and unlike PANI, the individual particles did not fuse, although there is an uneven distribution of the polymer. PEDOT shows a webbed structure spread throughout the CNTs, suggesting the presence of a strong interaction between the two materials. Attempts were made to fabricate control electrodes using 10 and 20 wt% carbon black as the combined substrate and conductive agent. After filtration and drying, electrodes could not be removed from the filter membrane intact, but rather fell apart as a powder. In general, mixtures with higher wt% carbon black were less cohesive. As a result, the electrochemical performance of these materials was not be determined. Furthermore, these results validates the importance of using high-aspect ratio carbon nanotubes to maintain the mechanical robustness of free-standing ECP-CNT electrodes. To find an estimate of the mass ratios present in the composite materials, thermal gravimetric analysis (TGA) and EDAX were performed (Fig. 1e and f). The decomposition of CNT paper occurs between 570 C and 675 C, while the ECP/CNT composites show an initial loss of moisture due to the hygroscopic nature of ECPs, followed by various mass loss transition that include: structural
Fig. 1. SEM images of composite electrodes made of a) CNT, b) PANI/CNT, c) PEDOT/CNT, and d) PPY/CNT. e) TGA profiles of the composite materials. f) EDAX of PPY/CNT composite, where P is polymer and C is carbon (%wt.).
38
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
decomposition, elimination of dopant molecules, thermal oxidation, and crosslinking [30–32]. The final decomposition at 570 C is attributed to the oxidation of CNTs. From these TGA profiles, it was estimated that the overall compositions of ECP in each composite are 70%, 57% and 50% for PANI, PPY and PEDOT, respectively. Interestingly, it appears that the CNTs exhibited a lower oxidation temperature when composites were more uniformly mixed. It is plausible that the improved interaction between PEDOT/MWNT (compared to PPY/MWNT) prevents MWNT bundling which makes the CNTs more susceptible to oxidation. Since the composite material is not completely homogeneous, EDAX was performed at different spots in the samples. Fig. 1f shows the EDAX analysis of a section of a PPY/CNT sample. Areas where polymer seemed absent, elemental analysis confirmed that there was mainly carbon and the levels of oxygen were low indicating low water content in those areas. Where polymer was present, the levels of nitrogen (PANI, PPY) or sulfur (PEDOT) were higher, as well as the oxygen (water) levels, because ECPs absorb moisture, even when dried in vacuum for long times (>1 day). In the specific area evaluated with EDAX, PEDOT content ranged from 25 to 42 wt.%, due to its more homogeneous distribution along the sample. PANI and PPY were relatively localized because of their granular shape and tendency to aggregate, which resulted in compositions that varied from 7 to 70% and from 0 to 65%, respectively. Composite electrodes were electrochemically evaluated as single electrodes in aqueous 3 M H2SO4 solutions and in organic electrolytes of 1.5 M tetraethylammonium tetrafluoroborate (TEABF4) in acetonitrile (MeCN). Fig. 2a and b shows CV profiles of ECP/CNT electrodes in aqueous and organic electrolytes, respectively. It can be observed from the area of the CV profiles at 30 mV/s, that the specific capacitances of CNT electrodes in aqueous and organic solutions of 27 F/g and 48 F/g are greatly increased with the addition of the electroactive conducting polymers, which are comparable to the highest values reported in literature for ECP/CNT composites [13,25]. PANI/CNT electrodes
showed the highest specific capacitance values of 448 F/g in aqueous 3 M sulfuric acid (aq) and 276 F/g in 1.5 M TEABF4 in acetonitrile (org), which were similar to PPY/CNT (428 F/g aq, 315 F/g org). PEDOT/CNT electrodes exhibited a notably lower specific capacitance (118 F/g aq, 95 F/g org) due to its relatively low theoretical charge capacity of 57 mAh/g compared to 135 mAh/ g and 140 mAh/g for polypyrrole and polyaniline, respectively. It is also observed that although the active potential range of organic electrolytes is wider, the highest values of capacitance are obtained in aqueous electrolytes due to the higher conductivity of the electrolyte and smaller ion size. Nevertheless, there is an evident scan rate dependence of all the materials in both electrolyte types (Fig. 2c and d), which limits the material performance at very high charge/discharge rates. Compared to CNT electrodes, which display charge capacities of 10 mAh/g (20 mAh/g) in aqueous (organic) electrolytes, ECP/CNT composite electrodes exhibit improved charge capacities as high as 84 mAh/g (104 mAh/g) in similar electrolytes for PPY/CNT and PANI/CNT electrodes (Fig. 2 g and h). An increase of up to 8.4 times the CNT charge capacity is are result of the additional Faradaic energy storage in conducting polymers. Fig. 2e and f shows that the separation (DV), amplitude and width of the PANI redox peaks vary significantly with scan rate and not so prevalently for PEDOT and PPY (not shown), which is attributed to the (i) granular structure of PANI that adds resistance to the mass diffusion and charge transfer processes and (ii) more open structure of PEDOT and PPY. The fact that the CNT electrodes also present diffusion limitations in the CV profiles indicates that the characteristics are associated with the electrode thickness (125 mm), the presence of CNT bundles, and the random orientation of the CNTs [14]. As noted from the SEM images, this scan rate dependence is intensified for the ECP/CNT composites, because of polymer aggregates in some regions. The presence of the electroactive polymer is also evidenced by the increase of diffusion and charge transfer resistances (inset of Fig. 2g and h), where the CNT electrode shows the fastest ion diffusion, followed by the PEDOT/CNT electrode. It can also be seen that the PEDOT/CNT electrode presents the smallest charge
Fig. 2. Single electrode performance of CNT and ECP/CNT from 3-electrode cells. a) CV profiles in 3 M H2SO4 at 30 mV/s. b) CV profiles in 1.5 M TEABF4 in acetonitrile at 30 mV/ s. Scan rate dependence of capacitance in c) 3 M H2SO4 and d) 1.5 M TEABF4 in acetonitrile. Specific capacitance vs. scan rate for PANI/CNT in e) 3 M H2SO4 and f) 1.5 M TEABF4 in acetonitrile. Galvanostatic discharge curves are presented with Nyquist plots as the inset for electrodes measured in g) 3 M H2SO4 and h) 1.5 M TEABF4 in acetonitrile. Electrode performances are reported on a total mass basis (ECP and CNT).
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
transfer limitations confirming a favorable interaction between the two materials. The effective charge transfer resistance of CNT electrode is associated with the resistance of ion transport into the CNT electrode and into the double layer within the porous structure. The highest resistance was found in PPY/CNT and PANI/ CNT electrodes, which showed a bulky structure. The PEDOT/CNT electrode was quite similar to the bare CNT electrodes, presumably
39
due to the intimate interaction between materials, thereby preventing further increases in ion adsorption or transfer resistance. Symmetric cells of the composites were fabricated, as shown in Fig. 3a, using Celgard 3401 and Celgard 2325 as separators for aqueous and organic systems, respectively. Results for cells comprising each type of ECP/CNT electrode with aqueous
Fig. 3. Electrochemical properties of ECP/CNT symmetrical supercapacitors. a) Picture of the electrodes, separator and testing setup for a symmetric cell. Panels b through f describe data for ECP/CNT devices in 3 M aqueous H2SO4 (solid lines) and a PEDOT/CNT device in 1.5 M TEABF4 in acetonitrile (dashed lines) b) cyclic voltammetry profiles normalized by scan rate and mass; c) scan rate dependence of capacitance for devices described in panel b; d) Nyquist plots with the inset figure showing an expanded view of the high frequency region for devices; e) plots of normalized imaginary capacitance (C”/g) vs. frequency; and f) galvanostatic discharge curves with time normalized to mAhr/ g. All CV profiles were recorded at 10 mV/s and capacitances are reported on a total electrode mass basis (ECP and CNT).
40
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
electrolyte are presented in addition to the results for devices comprising PEDOT/CNT in organic electrolyte. Since PPY and PANI are only stable in the oxidized form, one electrode was polarized to 0.8 V and the other at 0 V in a 3-electrode cell prior to cell fabrication and testing. In these cells, the voltage difference is 0.8 V is the “fully charged” state, and 0 V in the “fully discharged” state, where both electrodes hold a charge of 0.4 V. Fig. 3b shows the device capacitive behavior at a 10 mV/s, in which cells comprising PANI/CNT showed the highest specific capacitance of 46 F/g, followed by PPY/CNT (20 F/g) and PEDOT/ CNT (17 F/g). The performance of PEDOT/CNT devices with organic electrolytes (upper axis of Fig. 3b) showed similar capacitance (18 F/g) as the aqueous electrolyte (lower axis), but over a much larger voltage range. The capacitance values calculated for symmetric capacitors in a two-electrode cell is ideally four times lower than a single electrode measurement performed in a three-electrode configuration; two electrodes in series give rise to twice the mass and half the total capacitance of the single electrode. Symmetric cells comprising PEDOT/CNT electrodes, which can be efficiently oxidized and reduced, are slightly less than the single electrode equivalent, likely due to slight asymmetries and slight misalignment of the electrodes within the cell. For symmetric cells with PANI/CNT and PPY/CNT, the performance is much lower than expected based on single electrode performance. This is because of two reasons; first, the conductivity of PANI (PPY) is very different in the oxidized and neutral state, and second, PANI (PPY) is not a suitable anode material (poor stability when reduced). During cell discharge, the cathode state changes from fully charged (oxidized) to partially charged and the anode state changes from no charge (neutral) to partially charged. Note that the voltage of the entire cell with both electrodes partially charged is 0 and 0.8 V when one electrode is charged and the other is not. Since the energy and power densities depend on the square of voltage, the use of an electrolyte that can withstand a higher voltage range is advantageous. Accordingly, devices comprising PEDOT/CNT in 1.5 M TEABF4 in acetonitrile exhibited power and energy densities of 744.3 W/kg and 12.2 Wh/kg, respectively, compared to the same material in aqueous electrolytes (3 M H2SO4), 139.6 W/kg and 3.1 Wh/kg (for 1 min discharge, current varies). As a comparison, symmetric devices comprising PANI/CNT and PPY/CNT showed power (energy) densities of 283.4 W/kg (4.8 Wh/kg) and 114.6 W/kg (2.3 Wh/kg), respectively, in aqueous electrolytes. Each device exhibited a scan-rate dependence in the capacitance due to the diffusion-limited behavior in both electrolytes (Fig. 3c), as expected from the performance of the individual electrodes. The highest scan-rate dependence was observed in the organic electrolyte, due to its lower ion conductivity evident from the Nyquist plot (Fig. 3d). Even though the resistance of the organic electrolyte is notably larger, the effect of the larger potential range prevails. In the aqueous system, devices comprising the various ECP/CNT composites showed similar charge transfer resistances, with the lowest coming from PEDOT/CNT, as stated earlier due the presumed strong interaction between the two materials. From the plot of imaginary capacitance vs. frequency (Fig. 3e), we found that cells with PEDOT/CNT had the shortest relaxation time constant [14], indicating that this device can work efficiently at high discharge rates. The PANI/CNT device shows the (i) slowest relaxation time constant due to polymer aggregation and redox kinetics, and (ii) largest charge capacity (8.1 mAh/g, Fig. 3f) due to its high theoretical charge capacity of 140 mAh/g. This result is a direct evidence of the inherent trade-off between charge storage and discharge rate/efficiency.
4. Conclusions In summary, we have developed a simple process for fabricating free-standing ECP/CNT electrodes that can be easily scaled to large size electrodes without the need for binder materials, substrate, or sophisticated processing techniques. Specifically, in this study, ECP/CNT composite electrodes (ECP = PPY, PANI and PEDOT) were prepared with a routine dispersion-filtration process for immediate use in existing capacitor technologies. Electrochemical analysis of these composites showed an increase of up to 16 times in specific capacitance and 8 times in charge capacity compared to pure CNTs. Furthermore, the electrochemical properties of composite electrodes and devices obtained with our production method are comparable to capacitors made with more elaborate and less scalable procedures. Given its simplicity and the increased capacitance of our composite electrodes, they are posited to promote scalable manufacturing of superior supercapacitors. Acknowledgements MER and AMR acknowledge support from NSF CMMI SNM Award #1246800. MER acknowledges partial support from 3M Non-Tenured Faculty Grant Award. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]
J.T. Kearns, M.E. Roberts, J. Mater. Chem. 22 (2012) 2392–2394. M.R. Arcila-Velez, M.E. Roberts, Chem. Mater. 26 (2014) 1601–1607. J.T. Kearns, M.E. Roberts, J. Mater. Chem. 22 (2012) 25447–25452. Y. Xu, J. Wang, W. Sun, S. Wang, J. Power Sources 159 (2006) 370–373. L.-Z. Fan, J. Maier, Electrochem. Commun. 8 (2006) 937–940. V. Gupta, N. Miura, Mater. Lett. 60 (2006) 1466–1469. R. Ramya, R. Sivasubramanian, M.V. Sangaranarayanan, Electrochim. Acta 101 (2013) 109–129. C. Peng, S. Zhang, D. Jewell, G.Z. Chen, Prog. Nat. Sci. 18 (2008) 777–788. A. Malinauskas, J. Malinauskiene, A. Ramanavi9 cius, Nanotechnology 16 (2005) R51. L. Pan, H. Qiu, C. Dou, Y. Li, L. Pu, J. Xu, Y. Shi, Int. J. Mol. Sci. 11 (2010) 2636–2657. E. Lahiff, C. Lynam, N. Gilmartin, R. O’Kennedy, D. Diamond, Anal. Bioanal. Chem. 398 (2010) 1575–1589. M. Ates, A.S. Sarac, Prog. Org. Coat. 66 (2009) 337–358. S. Bose, T. Kuila, A.K. Mishra, R. Rajasekar, N.H. Kim, J.H. Lee, J. Mater. Chem. 22 (2012) 767–784. M.R. Arcila-Velez, J. Zhu, A. Childress, M. Karakaya, R. Podila, A.M. Rao, M.E. Roberts, Nano Energy 8 (2014) 9–16. G.A. Snook, P. Kao, A.S. Best, J. Power Sources 196 (2011) 1–12. V. Khomenko, E. Frackowiak, F. Béguin, Electrochim. Acta 50 (2005) 2499–2506. E. Frackowiak, V. Khomenko, K. Jurewicz, K. Lota, F. Béguin, J. Power Sources 153 (2006) 413–418. L. Zheng, X. Wang, H. An, X. Wang, L. Yi, L. Bai, J. Solid State Electrochem. 15 (2011) 675–681. R.G.S. Goh, N. Motta, J.M. Bell, E.R. Waclawik, Appl. Phys. Lett. 88 (2006) 053101. H. Deng, L. Lin, M. Ji, S. Zhang, M. Yang, Q. Fu, Prog. Polym. Sci. 39 (2014) 627–655. Y. Long, Z. Chen, X. Zhang, J. Zhang, Z. Liu, Appl. Phys. Lett. 85 (2004) 1796–1798. J. Zhang, L.-B. Kong, B. Wang, Y.-C. Luo, L. Kang, Synth. Met. 159 (2009) 260–266. H. Zengin, W. Zhou, J. Jin, R. Czerw, D.W. Smith, L. Echegoyen, D.L. Carroll, S.H. Foulger, J. Ballato, Adv. Mater. 14 (2002) 1480–1483. Y.-T. Weng, N.-L. Wu, J. Power Sources 238 (2013) 69–73. K. Lota, V. Khomenko, E. Frackowiak, J. Phys. Chem. Solids 65 (2004) 295–301. S. Machida, S. Miyata, A. Techagumpuch, Synth. Met. 31 (1989) 311–318. J.E. Albuquerque, L.H.C. Mattoso, D.T. Balogh, R.M. Faria, J.G. Masters, A.G. MacDiarmid, Synth. Met. 113 (2000) 19–22. M. Lefebvre, Z. Qi, D. Rana, P.G. Pickup, Chem. Mater. 11 (1999) 262–268. Z. Qi, P.G. Pickup, Chem. Commun. (1998) 2299–2300. H. Meng, D.F. Perepichka, M. Bendikov, F. Wudl, G.Z. Pan, W. Yu, W. Dong, S. Brown, J. Am. Chem. Soc. 125 (2003) 15151–15162. N.H. Idris, J. Wang, S. Chou, C. Zhong, M.M. Rahman, H. Liu, J. Mater. Res. 26 (2011) 860–866. D.S. Dhawale, R.R. Salunkhe, V.S. Jamadade, D.P. Dubal, S.M. Pawar, C.D. Lokhande, Curr. Appl. Phys. 10 (2010) 904–909.
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
A facile and scalable approach to fabricating free-standing polymer— Carbon nanotube composite electrodes Margarita R. Arcila-Veleza,1, Robert K. Emmetta,1, Mehmet Karakayab,c, Ramakrishna Podilab,c , Kryssia P. Díaz-Orellanaa , Apparao M. Raob,c , Mark E. Robertsa,c,* a b c
Department of Chemical & Biomolecular Engineering, Clemson University, Clemson, SC 29634, United States Department of Physics & Astronomy, COMSET, Clemson University, Clemson, SC 29634, United States Clemson Nanomaterials Center, Clemson University, Clemson, SC 29634, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 23 November 2015 Received in revised form 5 January 2016 Accepted 2 February 2016 Available online 15 February 2016
Nanoporous carbon materials are widely utilized in high-power supercapacitors due to their structural properties, chemical stability and conductivity, despite their limited energy density and charge storage capacity. Conducting polymers, on the other hand, possess high charge capacities; however, their application in commercial devices is hindered by degradation arising from their poor chemical and physical stability. Composites of carbon nanomaterials and conducting polymers have synergistic properties beneficial to supercapacitors, such as high capacitance and stability, but the limitations in scalable synthesis and polymer aggregation prevent widespread utilization. In this work, robust freestanding carbon nanotube (CNT)/electrically conducting polymer (ECP) electrodes are prepared using a simple dispersion filtration method, which can easily be scaled up. This process eliminates the use of binder, substrate or additional inactive weight. Composite CNT/ECP electrodes showed enhanced capacitance and charge capacity, achieving values up to 448 F/g and 84 mAh/g compared to 27 F/g and 10 mAh/g for pure CNT electrodes in aqueous electrolyte. Resulting symmetric cells exhibited energy and power densities of 5 Wh/kg and 283 W/kg, respectively, in aqueous electrolytes; and 12 Wh/kg and 744 W/kg, respectively, in organic electrolytes when using PEDOT/CNT electrodes. Given the process simplicity, relatively low cost and high throughput, the present composites have great potential for largescale manufacturing of CP/CNTs supercapacitor electrodes. ã 2016 Elsevier B.V. All rights reserved.
Keywords: Supercapacitor Conducting polymer Carbon nanotubes Scalable synthesis Filtration
1. Introduction Rapid growth in the demand for electrical energy storage continues to encourage the development of new materials and devices. With their ability to improve the performance of batteries or provide high-power, a recyclable and safe alternative such as supercapacitors have experienced tremendous growth in this industry. Supercapacitors, which typically comprise high-surface area, inert carbonaceous electrode materials, offer the benefits of high power density, rapid charge/discharge, and long lifetime. Even with these important advantages, carbon materials (e.g., activated carbon, carbon nanotubes, graphene) are limited by relatively low energy density as a consequence of their physical charge storage mechanism, electrical double layer capacitance
* Corresponding author at: Department of Chemical & Biomolecular Engineering, Clemson University, Clemson, SC 29631, United States. E-mail address: [email protected] (M.E. Roberts). 1 Both authors contributed equally to this work. http://dx.doi.org/10.1016/j.synthmet.2016.02.005 0379-6779/ ã 2016 Elsevier B.V. All rights reserved.
(EDLC), which is related to ion adsorption at the electrode surface. As we approach the maximum theoretical surface area in carbons, further increases in energy density are no longer attainable. On the other hand, electroactive materials such as metal oxides and electroactive conducting polymers (ECPs) can be prepared with relatively high surface areas, but are also able to store charge through Faradaic charge transfer in addition to EDLC. Metal oxides exhibit high Faradaic capacitance in aqueous solutions with good redox stability, but are limited by material cost, low operating voltage ranges, and the lack of scalable synthetic methods [1,2]. ECPs, such as polypyrrole (PPY), polyaniline (PANI), and poly (3,4-ethylenedioxythiophene) (PEDOT), provide a low-cost, large specific capacitance alternative to carbon materials and metal oxides [3–6] and can be easily prepared through various routes [7] from cheap monomers in large quantities. However, ECPs often suffer from poor cycle stability due to polymer over-oxidation stresses caused by volumetric changes that accompany the chargedischarge processes associated with ion insertion and removal [8]. To improve their performance and stability, significant efforts have
36
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
focused on creating micro or nanostructured ECP electrodes by modifying the synthesis processes, which enhances their mechanical and electrochemical properties [8–10]. Another approach to the same goal involves integrating a complementary material to reinforce and support the polymeric structure [11,12]. Carbon nanotubes (CNTs) are attractive for this purpose due to their excellent structural properties, electronic conductivity and chemical stability. CNTs improve the cycle life and mechanical properties of ECPs since CNTs adapt easily to changes in volume [13], and also introduce new electronic properties based on interactions between the two materials. Furthermore, the adaptation of such compositions helps overcome the low conductivity of ECPs in the neutral (or reduced) state and also increase electronic transport in thick electrodes [14]. Of equal importance, the addition of ECPs decreases CNT bundling and enhances the charge storage capacity compared to CNTs electrodes. In-situ chemical polymerization is the prevalent method for preparing ECP/CNT composites, which consists of oxidative polymerization of ECPs in a CNTsuspension and pressing the dry material into a pellet [15]. Through this approach, ECPs and CNTs are integrated on a molecular scale [16–20], increasing conductivity [21], although, homogeneity depends on the CNT dispersion and ECPs often prefer to grow on themselves rather than distributing homogeneously along the CNTs leading to polymer aggregation [17]. High specific capacitance values (500 F/g) [22] have been obtained by electrochemically polymerizing ECPs on CNTs due to strong interactions between the two materials [23]. Nevertheless, this method is limited to low amounts of polymer because as the film grows thicker it blocks electrolyte access to the CNTs. In melt compounding, CNTs are directly dispersed into a polymer melt; however, locally homogeneous states are difficult to achieve without breaking down the entangled CNTs [20]. A simple alternative to produce ECP/CNT composites involves mixing pre-synthesized ECP with CNTs [24,25], which is advantageous in terms of scalability and cost-effectivity [14]. Lota et al. [25] prepared composite electrodes by pressing a mixture of PEDOT and CNTs into pellets and obtained a single-electrode capacitance value of 95 F/g (at 2 mV/s), which possessed improved conductivity and cycle stability compared to PEDOT, albeit with significant diffusion limitations. In this work, we present a simple, low-cost approach to fabricating free-standing ECP/CNT composite electrodes using a sequential dispersion-filtration process. After ECP synthesis, ECP/ CNT dispersions (50/50 by wt.) are prepared in aqueous solutions containing sodium dodecyl sulfate (SDS) using ultrasonication, and then immediately filtered through a polyamide membrane with laboratory scale vacuum filtration. After electrodes are dried, robust, free-standing composite ECP/CNT paper electrodes are obtained, yielding high specific capacitance values up to 448 F/g (at 10 mV/s). In this process, no binder or pressure is needed to obtain efficient electrode performance. Considering the process simplicity, low cost and amenability to high throughput processing, our new approach to composite electrode fabrication provides a unique application to scalable, high power and energy supercapacitor electrode manufacturing.
synthesized by oxidative polymerization in aqueous HCl (1 M), using ammonium persulfate (APS) as an oxidant at 4 C for 2 h. PEDOT was synthesized using Fe(NO3)3 in an aqueous solution containing sodium poly(4-styrene sulfonate) for 6 h.
2. Materials and methods
where I is the discharge current and m is the electrode mass. Both calculations for capacitance are within 5%. Energy density, E, is calculated from
2.1. Materials Multi-walled carbon nanotubes were purchased from CheapTubes (dia.=30–50 nm, length = 10–20 mm) and NanoTech Labs (dia.=70–80 nm, length = 700–800 mm), and used as received. Polypyrrole (PPY), polyaniline (PANI), and Poly(3,4-ethylenedioxythiophene) (PEDOT) were synthesized using previously reported methods [26–28]. Briefly, PPY was synthesized by oxidative polymerization using FeCl3 in methanol at 0 C for 6 h. PANI was
2.2. Electrode fabrication The ECP/CNT composites were prepared by dispersing CNTs with an ECPs in aqueous solutions containing 1 wt.% SDS using a tip sonicator probe with a power of 75 W for 15 min. Subsequently, the suspension was poured onto a polyamide filtration membrane (Whatman, 0.45 mm pore diameter) and filtered using a vacuum filtration setup. The filtrate on the supporting filter membrane was washed with distilled water several times, and then oven dried at 60 C for 8 h. The resulting films were peeled off the membrane yielding free-standing ECP/CNT paper-like electrodes with good mechanical properties [28,29]. 2.3. Electrode characterization The physical properties of the composite electrodes were investigated using scanning electron microscopy (SEM, Hitachi SU6600) and energy dispersive X-ray spectroscopy (EDAX). Composite electrodes were electrochemically evaluated as single electrodes using a potentiostat (Princeton Applied Research VersaSTAT 4) in aqueous 3 M H2SO4 solutions and in 1.5 M tetraethylammonium tetrafluoroborate (TEABF4) in acetonitrile (MeCN). A Pt mesh was used as counter electrode, Ag/AgCl (sat’d KCl) and Ag/Ag+ (0.1 M AgNO3 in MeCN) electrodes were used as aqueous and non-aqueous pseudo-reference electrodes, respectively. Electrochemical characterization included cyclic voltammetry (CV) at scan rates of 10–300 mV s1, electrochemical impedance spectroscopy (EIS) (0.5 V DC bias, 0.1–10000 Hz, 20 mV RMS) and galvanostatic charge/discharge cycles (C/D) at a current such that the discharge time is equal to 60 s. The reported specific capacitance values were calculated from CV at a scan rate of 10 mV/s using the complete electrode mass (ECP + CNT). 2.4. Capacitance, energy and power calculations All the specific values of the electrodes are reported on a per mass basis (gravimetric) with respect to the total mass of the electrode (CNTs + ECP). The specific capacitance, C, of single electrodes measured in 3-electrode cells were determined by taking the average capacitance from CV measurements (current divided by scan rate and mass) over the full voltage range (at 10 mV s1). The capacitance of symmetric supercapacitors was calculated using the total mass of both electrodes (which were approximately equal) using two methods: first, by taking the average capacitance from CV measurements, as described for 3-electrode measurements; and second, from the inverse of the galvanostatic discharge slope (dV/dt) in the linear range, according to C¼
1 I dV m dt
1 E ¼ C DV 2 2
ð1Þ
ð2Þ
where DV is the voltage difference from the charged to discharged state. The power density, P, is calculated from P¼
DE Dt
ð3Þ
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
where Dt is the time required to completely discharge the device. 3. Results and discussion Low (300) magnification SEM images (not shown) show fairly homogenous composite films with a few CNT bundles. In general, the introduction of ECPs into the MWNT papers created a more openly porous network (Fig. 1) amenable to ion access to the electroactive materials and enhanced the Faradaic processes. From the SEM at higher magnification (Fig. 1a, >6000), carbon nanotube agglomerates are observed sporadically throughout the electrode. With the addition of the ECP (Fig. 1 b–d), it can be seen how the polymer acts as a binder and bridges the gaps between the CNT islands, enhancing the material’s electrical properties. From Fig. 1b and d, the heterogeneous structure of PANI and PPY can be seen as well as these polymers tendency to aggregate in some areas forming clusters. In the PANI/CNT composites, the particles of 0.2–0.3 mm in size merge to form agglomerates with diameters of 2–6 mm. The PPY/CNT samples show a homogeneous distribution with uniform particle sizes of
37
0.1–0.3 mm, and unlike PANI, the individual particles did not fuse, although there is an uneven distribution of the polymer. PEDOT shows a webbed structure spread throughout the CNTs, suggesting the presence of a strong interaction between the two materials. Attempts were made to fabricate control electrodes using 10 and 20 wt% carbon black as the combined substrate and conductive agent. After filtration and drying, electrodes could not be removed from the filter membrane intact, but rather fell apart as a powder. In general, mixtures with higher wt% carbon black were less cohesive. As a result, the electrochemical performance of these materials was not be determined. Furthermore, these results validates the importance of using high-aspect ratio carbon nanotubes to maintain the mechanical robustness of free-standing ECP-CNT electrodes. To find an estimate of the mass ratios present in the composite materials, thermal gravimetric analysis (TGA) and EDAX were performed (Fig. 1e and f). The decomposition of CNT paper occurs between 570 C and 675 C, while the ECP/CNT composites show an initial loss of moisture due to the hygroscopic nature of ECPs, followed by various mass loss transition that include: structural
Fig. 1. SEM images of composite electrodes made of a) CNT, b) PANI/CNT, c) PEDOT/CNT, and d) PPY/CNT. e) TGA profiles of the composite materials. f) EDAX of PPY/CNT composite, where P is polymer and C is carbon (%wt.).
38
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
decomposition, elimination of dopant molecules, thermal oxidation, and crosslinking [30–32]. The final decomposition at 570 C is attributed to the oxidation of CNTs. From these TGA profiles, it was estimated that the overall compositions of ECP in each composite are 70%, 57% and 50% for PANI, PPY and PEDOT, respectively. Interestingly, it appears that the CNTs exhibited a lower oxidation temperature when composites were more uniformly mixed. It is plausible that the improved interaction between PEDOT/MWNT (compared to PPY/MWNT) prevents MWNT bundling which makes the CNTs more susceptible to oxidation. Since the composite material is not completely homogeneous, EDAX was performed at different spots in the samples. Fig. 1f shows the EDAX analysis of a section of a PPY/CNT sample. Areas where polymer seemed absent, elemental analysis confirmed that there was mainly carbon and the levels of oxygen were low indicating low water content in those areas. Where polymer was present, the levels of nitrogen (PANI, PPY) or sulfur (PEDOT) were higher, as well as the oxygen (water) levels, because ECPs absorb moisture, even when dried in vacuum for long times (>1 day). In the specific area evaluated with EDAX, PEDOT content ranged from 25 to 42 wt.%, due to its more homogeneous distribution along the sample. PANI and PPY were relatively localized because of their granular shape and tendency to aggregate, which resulted in compositions that varied from 7 to 70% and from 0 to 65%, respectively. Composite electrodes were electrochemically evaluated as single electrodes in aqueous 3 M H2SO4 solutions and in organic electrolytes of 1.5 M tetraethylammonium tetrafluoroborate (TEABF4) in acetonitrile (MeCN). Fig. 2a and b shows CV profiles of ECP/CNT electrodes in aqueous and organic electrolytes, respectively. It can be observed from the area of the CV profiles at 30 mV/s, that the specific capacitances of CNT electrodes in aqueous and organic solutions of 27 F/g and 48 F/g are greatly increased with the addition of the electroactive conducting polymers, which are comparable to the highest values reported in literature for ECP/CNT composites [13,25]. PANI/CNT electrodes
showed the highest specific capacitance values of 448 F/g in aqueous 3 M sulfuric acid (aq) and 276 F/g in 1.5 M TEABF4 in acetonitrile (org), which were similar to PPY/CNT (428 F/g aq, 315 F/g org). PEDOT/CNT electrodes exhibited a notably lower specific capacitance (118 F/g aq, 95 F/g org) due to its relatively low theoretical charge capacity of 57 mAh/g compared to 135 mAh/ g and 140 mAh/g for polypyrrole and polyaniline, respectively. It is also observed that although the active potential range of organic electrolytes is wider, the highest values of capacitance are obtained in aqueous electrolytes due to the higher conductivity of the electrolyte and smaller ion size. Nevertheless, there is an evident scan rate dependence of all the materials in both electrolyte types (Fig. 2c and d), which limits the material performance at very high charge/discharge rates. Compared to CNT electrodes, which display charge capacities of 10 mAh/g (20 mAh/g) in aqueous (organic) electrolytes, ECP/CNT composite electrodes exhibit improved charge capacities as high as 84 mAh/g (104 mAh/g) in similar electrolytes for PPY/CNT and PANI/CNT electrodes (Fig. 2 g and h). An increase of up to 8.4 times the CNT charge capacity is are result of the additional Faradaic energy storage in conducting polymers. Fig. 2e and f shows that the separation (DV), amplitude and width of the PANI redox peaks vary significantly with scan rate and not so prevalently for PEDOT and PPY (not shown), which is attributed to the (i) granular structure of PANI that adds resistance to the mass diffusion and charge transfer processes and (ii) more open structure of PEDOT and PPY. The fact that the CNT electrodes also present diffusion limitations in the CV profiles indicates that the characteristics are associated with the electrode thickness (125 mm), the presence of CNT bundles, and the random orientation of the CNTs [14]. As noted from the SEM images, this scan rate dependence is intensified for the ECP/CNT composites, because of polymer aggregates in some regions. The presence of the electroactive polymer is also evidenced by the increase of diffusion and charge transfer resistances (inset of Fig. 2g and h), where the CNT electrode shows the fastest ion diffusion, followed by the PEDOT/CNT electrode. It can also be seen that the PEDOT/CNT electrode presents the smallest charge
Fig. 2. Single electrode performance of CNT and ECP/CNT from 3-electrode cells. a) CV profiles in 3 M H2SO4 at 30 mV/s. b) CV profiles in 1.5 M TEABF4 in acetonitrile at 30 mV/ s. Scan rate dependence of capacitance in c) 3 M H2SO4 and d) 1.5 M TEABF4 in acetonitrile. Specific capacitance vs. scan rate for PANI/CNT in e) 3 M H2SO4 and f) 1.5 M TEABF4 in acetonitrile. Galvanostatic discharge curves are presented with Nyquist plots as the inset for electrodes measured in g) 3 M H2SO4 and h) 1.5 M TEABF4 in acetonitrile. Electrode performances are reported on a total mass basis (ECP and CNT).
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
transfer limitations confirming a favorable interaction between the two materials. The effective charge transfer resistance of CNT electrode is associated with the resistance of ion transport into the CNT electrode and into the double layer within the porous structure. The highest resistance was found in PPY/CNT and PANI/ CNT electrodes, which showed a bulky structure. The PEDOT/CNT electrode was quite similar to the bare CNT electrodes, presumably
39
due to the intimate interaction between materials, thereby preventing further increases in ion adsorption or transfer resistance. Symmetric cells of the composites were fabricated, as shown in Fig. 3a, using Celgard 3401 and Celgard 2325 as separators for aqueous and organic systems, respectively. Results for cells comprising each type of ECP/CNT electrode with aqueous
Fig. 3. Electrochemical properties of ECP/CNT symmetrical supercapacitors. a) Picture of the electrodes, separator and testing setup for a symmetric cell. Panels b through f describe data for ECP/CNT devices in 3 M aqueous H2SO4 (solid lines) and a PEDOT/CNT device in 1.5 M TEABF4 in acetonitrile (dashed lines) b) cyclic voltammetry profiles normalized by scan rate and mass; c) scan rate dependence of capacitance for devices described in panel b; d) Nyquist plots with the inset figure showing an expanded view of the high frequency region for devices; e) plots of normalized imaginary capacitance (C”/g) vs. frequency; and f) galvanostatic discharge curves with time normalized to mAhr/ g. All CV profiles were recorded at 10 mV/s and capacitances are reported on a total electrode mass basis (ECP and CNT).
40
M.R. Arcila-Velez et al. / Synthetic Metals 215 (2016) 35–40
electrolyte are presented in addition to the results for devices comprising PEDOT/CNT in organic electrolyte. Since PPY and PANI are only stable in the oxidized form, one electrode was polarized to 0.8 V and the other at 0 V in a 3-electrode cell prior to cell fabrication and testing. In these cells, the voltage difference is 0.8 V is the “fully charged” state, and 0 V in the “fully discharged” state, where both electrodes hold a charge of 0.4 V. Fig. 3b shows the device capacitive behavior at a 10 mV/s, in which cells comprising PANI/CNT showed the highest specific capacitance of 46 F/g, followed by PPY/CNT (20 F/g) and PEDOT/ CNT (17 F/g). The performance of PEDOT/CNT devices with organic electrolytes (upper axis of Fig. 3b) showed similar capacitance (18 F/g) as the aqueous electrolyte (lower axis), but over a much larger voltage range. The capacitance values calculated for symmetric capacitors in a two-electrode cell is ideally four times lower than a single electrode measurement performed in a three-electrode configuration; two electrodes in series give rise to twice the mass and half the total capacitance of the single electrode. Symmetric cells comprising PEDOT/CNT electrodes, which can be efficiently oxidized and reduced, are slightly less than the single electrode equivalent, likely due to slight asymmetries and slight misalignment of the electrodes within the cell. For symmetric cells with PANI/CNT and PPY/CNT, the performance is much lower than expected based on single electrode performance. This is because of two reasons; first, the conductivity of PANI (PPY) is very different in the oxidized and neutral state, and second, PANI (PPY) is not a suitable anode material (poor stability when reduced). During cell discharge, the cathode state changes from fully charged (oxidized) to partially charged and the anode state changes from no charge (neutral) to partially charged. Note that the voltage of the entire cell with both electrodes partially charged is 0 and 0.8 V when one electrode is charged and the other is not. Since the energy and power densities depend on the square of voltage, the use of an electrolyte that can withstand a higher voltage range is advantageous. Accordingly, devices comprising PEDOT/CNT in 1.5 M TEABF4 in acetonitrile exhibited power and energy densities of 744.3 W/kg and 12.2 Wh/kg, respectively, compared to the same material in aqueous electrolytes (3 M H2SO4), 139.6 W/kg and 3.1 Wh/kg (for 1 min discharge, current varies). As a comparison, symmetric devices comprising PANI/CNT and PPY/CNT showed power (energy) densities of 283.4 W/kg (4.8 Wh/kg) and 114.6 W/kg (2.3 Wh/kg), respectively, in aqueous electrolytes. Each device exhibited a scan-rate dependence in the capacitance due to the diffusion-limited behavior in both electrolytes (Fig. 3c), as expected from the performance of the individual electrodes. The highest scan-rate dependence was observed in the organic electrolyte, due to its lower ion conductivity evident from the Nyquist plot (Fig. 3d). Even though the resistance of the organic electrolyte is notably larger, the effect of the larger potential range prevails. In the aqueous system, devices comprising the various ECP/CNT composites showed similar charge transfer resistances, with the lowest coming from PEDOT/CNT, as stated earlier due the presumed strong interaction between the two materials. From the plot of imaginary capacitance vs. frequency (Fig. 3e), we found that cells with PEDOT/CNT had the shortest relaxation time constant [14], indicating that this device can work efficiently at high discharge rates. The PANI/CNT device shows the (i) slowest relaxation time constant due to polymer aggregation and redox kinetics, and (ii) largest charge capacity (8.1 mAh/g, Fig. 3f) due to its high theoretical charge capacity of 140 mAh/g. This result is a direct evidence of the inherent trade-off between charge storage and discharge rate/efficiency.
4. Conclusions In summary, we have developed a simple process for fabricating free-standing ECP/CNT electrodes that can be easily scaled to large size electrodes without the need for binder materials, substrate, or sophisticated processing techniques. Specifically, in this study, ECP/CNT composite electrodes (ECP = PPY, PANI and PEDOT) were prepared with a routine dispersion-filtration process for immediate use in existing capacitor technologies. Electrochemical analysis of these composites showed an increase of up to 16 times in specific capacitance and 8 times in charge capacity compared to pure CNTs. Furthermore, the electrochemical properties of composite electrodes and devices obtained with our production method are comparable to capacitors made with more elaborate and less scalable procedures. Given its simplicity and the increased capacitance of our composite electrodes, they are posited to promote scalable manufacturing of superior supercapacitors. Acknowledgements MER and AMR acknowledge support from NSF CMMI SNM Award #1246800. MER acknowledges partial support from 3M Non-Tenured Faculty Grant Award. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]
J.T. Kearns, M.E. Roberts, J. Mater. Chem. 22 (2012) 2392–2394. M.R. Arcila-Velez, M.E. Roberts, Chem. Mater. 26 (2014) 1601–1607. J.T. Kearns, M.E. Roberts, J. Mater. Chem. 22 (2012) 25447–25452. Y. Xu, J. Wang, W. Sun, S. Wang, J. Power Sources 159 (2006) 370–373. L.-Z. Fan, J. Maier, Electrochem. Commun. 8 (2006) 937–940. V. Gupta, N. Miura, Mater. Lett. 60 (2006) 1466–1469. R. Ramya, R. Sivasubramanian, M.V. Sangaranarayanan, Electrochim. Acta 101 (2013) 109–129. C. Peng, S. Zhang, D. Jewell, G.Z. Chen, Prog. Nat. Sci. 18 (2008) 777–788. A. Malinauskas, J. Malinauskiene, A. Ramanavi9 cius, Nanotechnology 16 (2005) R51. L. Pan, H. Qiu, C. Dou, Y. Li, L. Pu, J. Xu, Y. Shi, Int. J. Mol. Sci. 11 (2010) 2636–2657. E. Lahiff, C. Lynam, N. Gilmartin, R. O’Kennedy, D. Diamond, Anal. Bioanal. Chem. 398 (2010) 1575–1589. M. Ates, A.S. Sarac, Prog. Org. Coat. 66 (2009) 337–358. S. Bose, T. Kuila, A.K. Mishra, R. Rajasekar, N.H. Kim, J.H. Lee, J. Mater. Chem. 22 (2012) 767–784. M.R. Arcila-Velez, J. Zhu, A. Childress, M. Karakaya, R. Podila, A.M. Rao, M.E. Roberts, Nano Energy 8 (2014) 9–16. G.A. Snook, P. Kao, A.S. Best, J. Power Sources 196 (2011) 1–12. V. Khomenko, E. Frackowiak, F. Béguin, Electrochim. Acta 50 (2005) 2499–2506. E. Frackowiak, V. Khomenko, K. Jurewicz, K. Lota, F. Béguin, J. Power Sources 153 (2006) 413–418. L. Zheng, X. Wang, H. An, X. Wang, L. Yi, L. Bai, J. Solid State Electrochem. 15 (2011) 675–681. R.G.S. Goh, N. Motta, J.M. Bell, E.R. Waclawik, Appl. Phys. Lett. 88 (2006) 053101. H. Deng, L. Lin, M. Ji, S. Zhang, M. Yang, Q. Fu, Prog. Polym. Sci. 39 (2014) 627–655. Y. Long, Z. Chen, X. Zhang, J. Zhang, Z. Liu, Appl. Phys. Lett. 85 (2004) 1796–1798. J. Zhang, L.-B. Kong, B. Wang, Y.-C. Luo, L. Kang, Synth. Met. 159 (2009) 260–266. H. Zengin, W. Zhou, J. Jin, R. Czerw, D.W. Smith, L. Echegoyen, D.L. Carroll, S.H. Foulger, J. Ballato, Adv. Mater. 14 (2002) 1480–1483. Y.-T. Weng, N.-L. Wu, J. Power Sources 238 (2013) 69–73. K. Lota, V. Khomenko, E. Frackowiak, J. Phys. Chem. Solids 65 (2004) 295–301. S. Machida, S. Miyata, A. Techagumpuch, Synth. Met. 31 (1989) 311–318. J.E. Albuquerque, L.H.C. Mattoso, D.T. Balogh, R.M. Faria, J.G. Masters, A.G. MacDiarmid, Synth. Met. 113 (2000) 19–22. M. Lefebvre, Z. Qi, D. Rana, P.G. Pickup, Chem. Mater. 11 (1999) 262–268. Z. Qi, P.G. Pickup, Chem. Commun. (1998) 2299–2300. H. Meng, D.F. Perepichka, M. Bendikov, F. Wudl, G.Z. Pan, W. Yu, W. Dong, S. Brown, J. Am. Chem. Soc. 125 (2003) 15151–15162. N.H. Idris, J. Wang, S. Chou, C. Zhong, M.M. Rahman, H. Liu, J. Mater. Res. 26 (2011) 860–866. D.S. Dhawale, R.R. Salunkhe, V.S. Jamadade, D.P. Dubal, S.M. Pawar, C.D. Lokhande, Curr. Appl. Phys. 10 (2010) 904–909.