Optimization of the transport behaviors of ions and electrons is the key for the property improvement of supercapacitor, which are essentially controlled by the design of hierarchical porous structure and electrical conductive backbone, from nanoscale to microscale, respectively. However, such design requirements are very difficult to be satisfied simultaneously, because the generation of porosity would result to the detrimental effects on the electrical conductivity of electrode. In this study, we propose to prepare a hierarchical porous supercapacitor electrode, with a novel 3-D highly porous (with pore size in the range of 50-100 nm) carbon nanotube sponges (CNTS) as a conductive substrate for the successively deposition of metal organic frameworks (MOF) and polyaniline. The porous structure of the sponge is beneficial for precursor penetration and uniform deposition of MOF and polyaniline (PANI) on to the nanotubes. The highly porous CNTS not only provides conductive highway for electrons, but also channels for ions quick diffusion. The coated MOF offers extra ion storage reservoir, while PANI further wire the insulating MOF together. In addition, the composite structure does not require any conductive additives or mechanical binders and delivers excellent capacitance coupled with flexible, compressive, and have relatively high specific capacitance.
I. INTRODUCTION
Electrochemical capacitors, also known as supercapacitors, represent a unique class of energy storage devices that bridge the gap between batteries and dielectric capacitors.1,2 With the outstanding comprehensive performance in terms of high power density, long cycle lifetime and safety, supercapacitors have been proved their important role in complementing or even replacing batteries in the energy storage field.2–5 The former store energy by physical adsorption of ions and electrons at the interface between the electrodes and electrolyte, and typically, activated carbon with high surface area is employed as electrode materials.5–7 The latter, on the other hand, store energy by redox reaction, and thus can significantly increase the energy storage capability at the price of decreasing power density and cycle life.8,9 Electrode structural design is the key step for the performance optimization for both of those two types of supercapacitors. Recently, advanced nano carbon materials, especially carbon nanotubes and graphene, have attracted great interest as supercapacitor electrodes, mainly because they have combined high conductivity, relatively high surface area and super mechanical property.2–4,10,11 However, these materials have little or no internal pores and even worse, they can easily agglomerate, which results to very limited accessible surface are to ions. In order to tackle with those problems, strategies such as unzipping carbon nanotubes to expose their inner pores, activation of graphene oxide to create nanopores, and assembling 3-Dimensional porous graphene aerogel and carbon nanotubes sponges to avoid agglomeration have been explored and show some effectiveness.12–15
Developing carbon materials with even higher surface area, and proper pores connectivity and size distribution is crucial to further increase the energy density of supercapacitors. Metal organic framework (MOF), one of the most porous materials with extremely high surface area, has great advantages in terms of structural control capability and versatility.12–17 The pore size and their connectivity can be controlled effectively and precisely by the proper selection of the ligand and metal. MOF has attracted enormous attention both from scientific and industrial areas, mainly for the application of gas adsorption.16 The investigation of MOFs used in the fabrication of supercapacitors, however, is still quite lacking. Previously, MOF was mainly employed as template or precursor for carbonization.18–21 For example, Takashi et al21 employed MOF as template for the polymerization and carbonization of a thin layer polyfury, which resulted to highly porous carbon materials with high specific capacitance. Jujiao et al22 integrated the properties of ZIF-8 and ZIF-67 to synthesis electrode materials with core-shell crystal structure, which has high surface area and relatively intact crystalline structure even after carbonization, resulting to very high specific capacitance. Recently, Deok et al23 tested the capacitive performance of typical MOFs and Jaeseok et al25 synthesized MOFs on the surface of dispersed carbon nanotube (CNT), followed by electrical wiring via further coating a thin layer of polyaniline (PANI). However, the employment of non-active materials, such as binders and conductive carbon black, decrease the energy storage capability at device level.
In this study, we used free standing carbon nanotube sponge (CNTS) as a 3-D current collector and substrate for the growth of core-shell MOF particles with very high surface area. ZIF-8 and ZIF-67 were sequentially coated on the surface of carbon nanotubes, forming a layer of MOF with core-shell structure on the surface of CNT, and forming hybrid CNTS/csMOF (CM) structure. The ZIF-67 not only provides mechanical strength to the ZIF-8 during carbonization and avoid its structural collapse, but also forming crystal bridges with carbon nanotube substrate and thus guarantee facile electron transport.16,24 Coating and wrapping a thin layer of conductive polymer, polyaniline, on CM, can further improve the electrical properties, as well as specific capacitance due to the redox characteristic of PANI. The obtained hybrid CNTS/csMOF/PANI (CMP) supercapacitor is flexible, compressive, and have relatively high specific capacitance. Figure 1a illustrates the synthesis process for CMP. As shown in Figure 1b, from macroscopy point of view, the electrode can be compressed as high as 80% without any noticeable residue strain. Even after 500 hundreds of compress at a strain of 50%, the electrode can still recover its original shape. Such mechanical behavior make this supercapacitors flexible and compressible.
(a) Schematic illustration of the synthesis process for CMP. (b) Photographs demonstrating the high degree of flexibility and resilience of the CMP composite electrode.
(a) Schematic illustration of the synthesis process for CMP. (b) Photographs demonstrating the high degree of flexibility and resilience of the CMP composite electrode.
II. EXPERIMENT
The porosity of CNTS is as high as 99%, and thus it can provide precursor considerable transportation channels to access the CNTs deep inside of CNTS.25 The CNTS were treated with HNO3 acid for 4 hours at 70 °C to make them hydrophilic, as we reported previously.12 The CMP based composite electrodes are prepared by three steps. Firstly, by direct immersing CNTS in the precursor solution of 2-methylimidazole (5.20 g) in methanol (120 ml) at room temperature, the pores in CNTS were quickly filled with the solution. This is mainly due to the capillary force induced by the nanosized pores in CNTS. Then, a solution of zinc nitrate hexahydrate (4.46 g) methanol (60 ml) was added to the above mixture under ultrasound and stirring for 15 min. Secondly, cobalt nitrate hexahydrate (4.38 g) was added to the methanol (60 ml) solution, and ultrasonically stirred at the same speed for 1 h. After several times centrifugation (10000 rpm) to obtain the CM. Finally, PANI was deposited, whereby the aniline monomer was polymerized in 1 M HCl solution using ammonium persulfate (The CM was immersed in 1 M HCl. And then an equal volume of precooled 0.2 M ammonium persulfate solution was added to the CM solution through drop-by-drop way. After that, the mixture was kept at 1-5 °C for 8 h for complete reaction).26 The products were rinsed with deionized water until the filtrate was neutral and freeze-dried for 24 h.
III. RESULTS AND DISCUSSION
As shown in Figure 2a, the pristine CNTS have open and porous structure with average pore size in the range of 50-100 nm, which can be well controlled by the CVD condition.12,13 Those hierarchical pores are formed by randomly distribution of one-dimensional CNTs in 3D space without any isolated pores, therefore, all pores would be available for ions. The flexible CNT could also provide buffer layer for the expansion/contraction of the pseudocapacitive PANI layer during charge/discharge process, promoting the cycling performance of electrodes. Since the CNTS was treated by acid before the coating of MOFs, the hydrophilic groups could act as nucleates for the growth of MOFs uniformly on each CNT. Indeed, as shown in Figure 2b, MOF crystals with average size of 300 nm were uniformly coated on the surface of CNTs. The enlarged SEM image (Figure 2c) further reveals that the coating of MOF crystal does not block the original hierarchical porous structures, manifesting the structural robustness of CNTS. TEM characterization (Figure 2d) again, confirms that MOF are firmly coated on the surface of CNTs, with average size of ∼200 nm. Such structure, consists of sub-micro sized pores formed by CNTs and nanoscale pores in MOFs, provides hierarchical pores for the effective transport of ions, greatly benefitting the supercapacitor performance.
(a) Scanning electron microscopy (SEM) image of pristine CNTS. (b) and (c) SEM image of CM. (d) Transmission electron micrograph (TEM) image of CM.
(a) Scanning electron microscopy (SEM) image of pristine CNTS. (b) and (c) SEM image of CM. (d) Transmission electron micrograph (TEM) image of CM.
As shown in Figure 3a, the deposition of ZIF-8, ZIF-67 and PANI on CNTS significantly improves the capacitance. The cyclic voltammograms (CV) maintains approximately symmetrical rectangular shape over the range of scan rates employed, which is characteristic of the electrochemical double layer capacitance behavior. Therefore, we can infer that CMP can be used as a flexible electrode. Figure 3b shows the CV for the CMP over a range of scan rates from 2 to 50 mV/s. At the same time, the shapes of the CV curves are nearly rectangular and symmetrical at all scan rates tested. As shown in Figure 3c, the constant current charge/discharge test was performed at different current densities (1-10 A/g), and the charge/discharge curves were triangular. The specific capacitance of supercapacitor can be calculated according to the following equation:
where I (A) is the discharge current, Δt (s) is the discharge time, M (g) is the total mass of the active materials in each electrodes, and ΔV (V) is the voltage window. Specific capacitances of 342.5, 294.6, 230.8, 179.5 F/g at current densities of 1, 2, 5, 10 A/g. The results show that the electrode has excellent reversibility and good charge transport performance. As shown in Figure 3d, as the scan rate increases from 2 mV/s to 500 mV/s, the specific capacitance CNTS and CMP drop from 89 to 26 F/g, from 746 to 116 F/g. The energy density and power density are also important parameters affecting the electrochemical performance of supercapacitors. The energy density and power density of a supercapacitor can be calculated from the following equations:
where E (Wh/kg), C (F/g), ΔV (V), Δt (s) and P (W/kg) are the energy density, specific capacitance, the operating potential window, discharge time and power density, respectively. Figure 4 shows the relationship between the energy density and the power density. The CMP supercapacitor could obtain the highest energy density of 28.9 Wh/kg at a current density of 1 A/g (at a power density of 759.4 W/kg). Besides, the energy density was still 16 Wh/kg when the power density was increased to 8.23 KW/kg. The maximum energy density of the CMP supercapacitor is much higher than the values reported for other group.27,28
(a) CV for CNTS and CMP with a scan rate of 10 mV/s. (b) CV for CMP scans range from 2 to 50 mV/s. (c) Constant current charge and discharge curves of CMP at different current densities. (d) Specific capacitances based on CNTS and CMP at increasing sweep rates.
(a) CV for CNTS and CMP with a scan rate of 10 mV/s. (b) CV for CMP scans range from 2 to 50 mV/s. (c) Constant current charge and discharge curves of CMP at different current densities. (d) Specific capacitances based on CNTS and CMP at increasing sweep rates.
The Ragone plots of power density versus energy density for CMP supercapacitor.
This phenomenon is expected and is caused by diffusion rate limitations at such high current densities. The capacitance and good rate capability can be directly attributed to the structural porosity (CM) and uniform coverage of PANI. The conductive PANI layer also facilitates further improvement of capacitance by and laminating insulating MOF together. On other words, MOF are actually laminated between conductive CNT and PANI layers, potentially significantly increase the usage of pores in MOF. Because of the strong mechanical properties of CNT and structural robustness of CNTS, the developed hierarchical porous electrode is highly compressible.
IV. CONCLUSIONS
In summary, we have developed a new type of supercapacitor device that is deposited on 3-D high porosity CNTS using ZIF-8, ZIF-67, and PANI. The high mechanical properties of CNTS provide structural robust substrate for the successive coating of MOF and PANI layer. The sub-micro pores resulted from the randomly distributed CNTs, combined with nanoscale pores in MOF, the prepared electrode possess hierarchical porous structures, which is very important for supercapacitor performance. The conductive CNT and PANI layer can provide highway for the effective and efficient transport of electrons. The specific capacitance property of CNTS improved significantly by ZIF-8, ZIF-67 and PANI coating, with start specific capacitance increasing from 89 to 746 F/g, and can deliver a maximum energy density of 28.9Wh/kg. More importantly, the CMP composite is flexible, compressive, and have relatively high specific capacitance. Therefore, such sponges could enable flexible supercapacitor devices for flexible electronics applications.
ACKNOWLEDGMENTS
This work was supported by Hubei Provincial Natural Science Foundation (Project #2018CFB683), Hubei Provincial Department of Transportation (Research Project #2016-600-1-10), and Hubei Provincial Department of Construction (Construction Science and Technology Plan Project 2016-347-1-13). The technology disclosed in this paper has been protected by a Chinese patent.