Boosting the Utilization and Electrochemical Performances of Polyaniline by Forming a Binder-Free Nanoscale Coaxially Coated Polyaniline/Carbon Nanotube/Carbon Fiber Paper Hierarchical 3D Microstructure Composite as a Supercapacitor Electrode.

Nanoscale polyaniline (PANI) is formed on a hierarchical 3D microstructure carbon nanotubes (CNTs)/carbon fiber paper (CFP) substrate via a one-step electrochemical polymerization method. The chemical and structural properties of the binder-free PANI/CNTs/CFP electrode are characterized by field emission scanning electron microscopy, transmission electron microscopy, Fourier transform infrared spectroscopy, and Raman spectroscopy. The specific capacitance of PANI/CNTs/CFP tested in a symmetric two-electrode system reaches 731.6 mF·cm-2 (1354.7 F·g-1) at a current density of 1 mA·cm-2 (1.8 A·g-1). The symmetric supercapacitor device demonstrates excellent cycling performance up to 10,000 cycles with a capacitance retention of 81.4% at a current density of 1 mA·cm-2 (1.8 A·g-1). The results demonstrate that the binder-free CNTs/CFP composite is a strong backbone for depositing ultrathin PANI layers at a high mass loading. The hierarchical 3D microstructure PANI/CNTs/CFP provides enough space and transporting channels to form an efficient electrode-electrolyte interface for the supercapacitance reaction. The formed nanoscale PANI film coaxially coated on the sidewalls of CNTs enables efficient charge transfer and a shortened diffusion length. Hence, the utilization efficiency and electrochemical performances of PANI are significantly improved. The rational design strategy of a CNT-based binder-free hierarchical 3D microstructure can be used in preparing various advanced energy-storage electrodes for electrochemical energy-storage and conversion systems.

Introduction

Supercapacitors are promising technologies because of their remarkable properties, such as ultrahigh output power density, fast charge/discharge (CD) capability, excellent cycling stability, and safety features.[1,2] The capacitance of a supercapacitor can be defined by using an electrical double-layer capacitor (EDLC) and a pseudocapacitor. The energy densities of pseudocapacitors are usually higher than those of EDLCs because pseudocapacitors store energy through the reversible faradaic reaction of their electrodes.[3,4] Therefore, the key point of preparing supercapacitors with high energy and power densities is to fabricate electrode materials with high pseudocapacitance and excellent electrochemical properties. In recent years, conducting polymers and their composites are widely used as pseudocapacitive materials because they can offer various oxidation states for redox charge transfer reactions to achieve significantly high energy densities.[5−10]

Polyaniline (PANI), as a kind of electricity-conducting polymer, has been extensively studied for supercapacitor applications because of its easy synthesis, low cost, and good pseudocapacitive performance. However, in energy-storage devices, a bulk PANI is ineffective given that its low available surface area leads to the low accessibility of electrolytes. Thus, the utilization efficiency of a bulk PANI is low because a large portion of the material becomes “dead” materials.[11−13] Moreover, the extensive swelling/shrinking of bulk PANI networks during long-term CD cycles results in a poor cycling life; thus, PANI has limited applications.[14−16]

To address the limitations of PANI, many researchers have focused on developing PANI materials integrated with highly conductive carbonaceous nanomaterials for the fabrication of high-performance electrode materials.[4,11,17−19] Yang et al.[20] deposited PANI on a single graphitized multiwalled carbon nanotube (MWCNT) grafted with poly(4-vinylpyridine) through in situ chemical polymerization to obtain a specific capacitance of up to 1065 F·g–1 and cyclic durability with 92.2% capacitance retention after over 1000 cycles. Jiang et al.[21] uniformly embedded PANI and graphene oxide (GO) between stacked CNT networks to achieve a synergistic effect and a specific capacitance of 729.3 F·g–1 at 1 A·g–1 in a three-electrode configuration. Using porous graphene as a current collector, Pourjavadi et al.[22] prepared a free-standing supercapacitor electrode with carbon fibers functionalized with PANI. The maximum specific capacitance of the electrode was 710 F·g–1 at a current density of 2A·g–1. Wang et al.[23] fabricated a whisker-like PANI on mesoporous carbon surfaces by in situ chemical oxidative polymerization. The electrochemical performance of the material was as high as 900 F·g–1 at a CD current density of 0.5 A·g–1, and the discharge capacitance loss was only approximately 5% after 3000 consecutive cycles. Khalid et al.[24] reported MWCNT-assisted PANI thin films on a gold-coated poly(ethylene terephthalate) sheet to increase accessibility for supercapacitive behavior. Cheng et al.(25) used electroetched carbon fiber cloth as an electrode for PANI deposition. PANI nanowires can reach a mass-normalized specific capacitance of 673 F·g–1 and an area-normalized specific capacitance of 3.5 F·cm–2. The PANI and carbon black composite synthesized by Wang et al.[26] demonstrated that carbon black as the secondary dopant of PANI exhibits high conductivity, extended conformation structure, improved porosity, high oxidation state, and reduced hydrolysis effect, and the synergistic effect between PANI and carbon black leads to superior capacitive performance. Xu et al.[27] synthesized PANI with carbon aerogel as a conducting filler to improve the electrochemical performance of PANI.

Previous research work demonstrated the synergistic effects of the ideal EDLC behavior of nano carbonaceous materials and pseudocapacitive behavior of PANI-optimized electrochemical properties after the combination of those materials. However, most carbonaceous nanomaterials exhibit unsatisfactory properties, such as irreversible aggregation caused by strong van der Waals forces and surface group interactions and insufficient EDLC charge capacity due to the limited surface area. These drawbacks reduce the synergistic effects between PANI and carbon-based substrates. Moreover, in some powdered carbonaceous nanomaterials (such as CNTs, graphene, GO, and carbon spheres), assembling PANI/carbonaceous nanomaterial powder onto current collectors for the preparation of electrodes usually requires the use of binders and conducting additives.[28,29] These additional components significantly increase processing cost and substantially compromise electrode-level performance metrics.[30,31] Therefore, in order to fully exploit the advantages of the synergistic effect between PANI and carbon-based materials, a reasonable architecture design and performing structure-to-property tailoring should be considered.[32] These strategies are promising in considerably improving the electrochemical performance of PANI/carbonaceous composites. In our previous work, a binder-free approach for fabricating multidimensional and multicomponent nanomaterials with hierarchal 3D nano-architecture structures was designed and developed by directly growing a CNT forest on a 3D network carbon fiber paper (CFP) in a chemical vapor deposition (CVD) reactor.[33] The as-grown CNTs were coated uniformly on each individual carbon fiber forming a binder-free hierarchal 3D architecture structure. The resulting hierarchal 3D mesoporous matrix exhibited high conductivity and facilitated ion diffusion.[34]

In this paper, the features of the binder-free hierarchal 3D CNTs/CFP composite prepared previously rendered the matrix a suitable substrate for PANI modification. The electrochemical method was performed for the synthesis of forming nanoscale coaxially coated PANI/CNTs/CFP hybrid electrode materials. The binder-free hierarchal 3D microstructure CNTs/CFP not only provided a high reacting area for the high specific areal mass loading of PANI, but also promoted the formation of a nanoscale layer on the highly conductive and structurally stable CNTs/CFP substrate even with the high PANI mass loading. Consequently, the pseudocapacitance of PANI was fully utilized, and the cycling stability of PANI significantly improved.

Results and Discussion

Characterization of the PANI/CNTs/CFP and PANI/CFP Composites

The Fourier transform infrared (FT-IR) spectra of PANI/CNTs/CFP and PANI/CFP composites (Figure A) were used in analyzing the functionalized process and obtaining information on the chemical bond structures of the composites. A set of typical peaks corresponding to PANI appeared in the spectra of the PANI/CNTs/CFP and PANI/CFP composites. The peaks at 1567 and 1488 cm–1 in the spectrum of PANI/CNTs/CFP were due to the C=C stretching vibrations of the quinoid rings and benzenoid rings, respectively. The peaks at 1295, 1230, 1124, and 792 cm–1 in the spectrum of PANI/CNTs/CFP were attributed to the C–N stretching vibrations of aromatic amines, C=N stretching in the PANI, and in-plane and out-of-plane bending of C–H in aromatic rings, respectively.[18,37] Similar peaks were also observed in the FT-IR spectrum of the PANI/CFP composite.

Figure 1

FT-IR spectra (A), Raman spectra (B), and X-ray photoelectron spectroscopy (XPS) spectra (C–E) of PANI/CNTs/CFP.

Raman spectra were obtained for the further characterization of the structures of the samples. In Figure B, after PANI was coated on the CNTs/CFP substrate, the peak that appeared at 1168 cm–1 in the Raman spectrum was ascribed to the C–H bending of the quinoid ring of PANI. The peaks at 1348, 1491, and 1587 cm–1 were attributed to C–N+ stretching vibration, C=N stretching of the quinoid ring, and C–C stretching of the benzenoid ring of PANI, respectively.[37,38] Almost the same peaks appeared in the Raman spectrum of the PANI/CFP composite.

XPS spectra were obtained for further characterization of the doping level of as-deposited PANI on the CNTs/CFP substrate (Figure C–E). The XPS spectra of PANI/CFP are shown in Figure S1. In Figure C, three distinct peaks at around 531.2, 398.1, and 284.4 eV, and two weak peaks at around 230.8 and 167.9 eV in the XPS spectrum of PANI/CNTs/CFP correspond to O 1s, N 1s, C 1s, Cl 2s, and Cl 2p peaks, respectively. The presence of Cl indicated that PANI has been doped with the Cl– anion during cyclic voltammetry (CV) synthesis. The C 1s spectrum of PANI/CNTs/CFP is shown in Figure D. The four peaks of C 1s with binding energies at 284.7, 285.4, 286.3, and 287.7 eV were ascribed to C–C, C–N, C–O, and C=O, respectively.[39−41] The low-intensity peaks of the binding energy at 286.3 and 287.7 eV indicate that a few oxygen functionalities have been formed in the as-prepared PANI/CNTs/CFP composite. The N 1s core level spectra of PANI/CNTs/CFP were deconvoluted into four peak positions of 398.8, 399.6, 400.5, and 401.5 eV (Figure E). The four peaks of N 1s were fitted into the neutral imine nitrogen in quinoid groups (=N−), neutral amine nitrogen in the benzenoid groups (−NH−), protonated amine in the polaron state (−NH+•−), and positively charged imine in the bipolaron state (=NH+−).[42,43] The doping level represents the extent of oxidation or reduction of the polymer and is usually measured by the proportion of the dopant ions or molecules incorporated per monomer unit.[44] The doping level can be determined by the area ratio of the protonated components of the N 1s core level spectra to the total area of the N 1s core level spectra.[12,45] The maximum doping level achievable in the emeraldine form of PANI is 0.5.[46] The doping level of PANI/CNTs/CFP was calculated to be 0.48. The high doping level achieved in PANI/CNTs/CFP was attributed to the high positive charge density on nitrogen and the easy diffusion of the electrolyte ions through the PANI thin layers. The FT-IR, Raman, and XPS spectra indicated that PANI was successfully synthesized on the CNTs/CFP and CFP substrates.

After CV synthesis, the typical scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images of PANI/CFP and PANI/CNTs/CFP were observed (Figure ). The original CNTs/CFP (Figure S2B) or CFP (Figure S2A) substrates were compared. The two substrates were coated with PANI (Figure A,D). PANI coated on each CFP fiber (Figure B,C) was not uniform on the CNTs/CFP surface (Figure F). The maximum thickness of the PANI layer on a CFP fiber reached up to ∼55 nm (Figure C), and the minimum thickness observed from the TEM image (inset picture of Figure C) was ∼6 nm. In comparison, the TEM image of individual PANI/CNTs (Figure F) showed that PANI was uniformly and coaxially coated on each CNT sidewall and the thickness of PANI ranged from 2 to 9 nm. The 3D network structure formed by the CNTs on CFP was clearly observed, though the CNTs conglutinated together after PANI polymerization (Figure E). The higher magnification of the bare CNTs/CFP substrate before and after the PANI deposition is shown in Figure G. Before the PANI polymerization, the CNTs were grown with a random orientation together with the formation of a 3D microstructure exhibiting a sufficient meso/macropore structure. The minimum distance between each CNTs is approximately 50 nm. The meso/macropore structure of PANI/CNTs is still visible after the electrochemical deposition of PANI, the minimum distance between each PANI/CNTs is approximately 30 nm.

Figure 2

SEM images of PANI/CFP (A, B) and PANI/CNTs/CFP (D, E), TEM images of PANI/CFP (C) and PANI/CNTs/CFP (F), and (G) the SEM images of CNTs/CFP before and after PANI deposition.

Generally, the electrochemical reaction provides a better method of synthesizing more pure PANI films with direct, simple, and an accurate control of the initiation and termination steps via facilely predesigning the experimental parameters, which can ensure the good control of the expansion degree of the as-deposited PANI film compared to chemical methods.[47−49] Comparatively, the electro-oxidation of aniline by the CV method can produce a more even polymeric film, which adheres on the surface of the substrate firmly.[50−52] Huang et al.[53,54] have systematically studied the relationship between the structures of PANI@ACNTs/Ti (or Al) foil composites and the supercapacitor performances. They proposed that through controlling the thickness of the PANI films to less than 11 nm by accurately controlling the electrochemical parameters for PANI deposition, there will be a sufficient distance between each PANI/CNT to maintain the opening nanopore structures of the composite and maximize the active surface area of PANI/CNTs for the ion diffusion to fully utilize the electrochemical capacitance of PANI/CNTs. Accordingly, in this work, the as-prepared PANI film can be controlled efficiently within a rational thickness range via the electrochemical polymerization method to ensure the hierarchical 3D microstructure of the PANI/CNTs/CFP composite as shown in Figure G. The preservation of the 3D network structure facilitated ion transport in the electrolyte solution during the supercapacitive reaction of PANI/CNTs/CFP.

Electrochemical Properties of the PANI/CNTs/CFP and PANI/CFP Composites

The electrochemical performances of the CFP/CNTs, CFP, PANI/CNTs/CFP, and PANI/CFP composites were first evaluated by CV measurements in the three-electrode system in 1 mol·L–1 H2SO4 electrolyte. The CV curves of the CNTs/CFP and CFP at a scan rate of 5–200 mV·s–1 are shown in Figure S3. The CV curves of the bare substrates and those two substrates with PANI deposited were recorded at a scan rate of 5 mV·s–1 as shown in Figure A. The rectangular shape CV curves with a small broad hump of CFP/CNTs and CFP indicated that the charge storage of those two substrates was mainly based on an ideal capacitive mechanism via the adsorption of electrolyte ions at the electrode/electrolyte interface and a fast surface redox reaction produced by the functional groups after electrochemical preoxidation. The areal specific capacitance of CFP/CNTs and CFP obtained from CV curves at 5 mV·s–1 was 98.4 and 32.3 mF·cm–2, respectively. Comparatively, a pair of obvious redox peaks was observed in the CV curves after synthesizing PANI on the CNTs/CFP and CFP substrates. Generally, the cyclic voltammogram of PANI is typically composed of three redox couple reactions over a wide range of potential, including the redox transition between leucoemeraldine (insulating) and protonated emeraldine (conducting), the transition between p-benzoquinone and hydroquinone, and the redox transition between the emeraldine and the pernigraniline.[55] Between the potential range of 0–0.4 V (vs Ag/AgCl), the prominent redox peaks observed in the CV curves of PANI/CNTs/CFP and PANI/CFP were mainly attributed to the conversion of PANI between leucoemeraldine and emeraldine.[56,57] The areal-specific mass loading of PANI on CNTs/CFP (about 0.54 mg·cm–2) was much higher than that on the CFP substrate (∼0.21 mg·cm–2). The specific capacitance (Csp) of PANI/CNTs/CFP at 5 mV·s–1 was approximately 816.6 mF·cm–2 (about 1512.2 F·g–1), whereas that of PANI/CFP was ∼265.3 mF·cm–2 (about 1263.3 F·g–1).

Figure 3

Electrochemical characterization of PANI/CNTs/CFP and PANI/CFP electrodes. PANI was synthesized by the CV method. (A) CV curves of PANI/CNTs/CFP and PANI/CFP at a 5 mV s–1 scan rate. (B) Nyquist impedance spectra of the PANI/CNTs/CFP and PANI/CFP electrodes. (C) Schematic of the areal mass loading and the thickness of the PANI layers formed on CNTs/CFP and CFP substrates, respectively.

To identify the specific electrochemical active surface area (ESA) of the CFP/CNTs, CFP, PANI/CNTs/CFP, and PANI/CFP, the double layer capacitance method was utilized to calculate the double layer capacitances (Cdl), given that Cdl is directly proportional to ESA (for details, refer to the Supporting Information, Figures S4 and S5). Cdl corresponding to CFP/CNTs and CFP substrates was 33.8 and 8.2 mF·cm–2, respectively. Similar to the enhanced specific surface area of CNTs/CFP confirmed by nitrogen adsorption–desorption isotherm analysis (Figure S6), the enhancement in Cdl of CNTs/CFP was because of the availability of increased ESA for ion adsorption provided by the as-grown CNTs on CFP. This enhancement in the ESA of CNTs/CFP will ultimately lead to a high ESA for PANI, because a large electrochemical surface became available for the anilinium chloride monomer to undergo polymerization. After PANI deposition, the Cdl value was about three times higher for PANI/CNTs/CFP (101.1 mF·cm–2) compared to that for PANI/CFP (33.6 mF·cm–2). The enhancement in the Cdl of PANI/CNTs/CFP relative to that of PANI/CFP further proved the increased surface area through the CNT growth on CFP for more PANI deposition.[48−50] Therefore, the microstructure of CNTs/CFP not only provided efficient transport channels for electrons and ions but also provided a backbone with a high surface area for a relatively higher mass loading and ultrathin PANI deposition when compared to CFP and other conventional 2D structure substrates, as shown in the schematic picture in Figure C. Both of those two merits ensured highly efficient deposition and utilization of the as-prepared PANI to produce a high specific capacitance electrode. In addition, as mentioned in previous work, the total specific capacitance of PANI and the carbonaceous composite (Csp,total) included the contributions from both the faradic redox reaction (pseudocapacitance of PANI, Csp,PANI) and the double layer electrostatic storage (Cdl).[54] Thus, based on the total Csp,total (816.6 mF·cm–2) and the Cdl (101.1 mF·cm–2) of PANI/CNTs/CFP obtained above, Csp,PANI calculated by subtracting Cdl from Csp,total was 715.5 mF·cm–2. Therefore, the pseudocapacitive to capacitive contribution of the PANI/CNTs/CFP electrode at a scan rate of 5 mV s–1 was approximately 87.6%.

Electrochemical impedance spectra (EIS) are usually used in analyzing the characteristic transient features of electrochemical electrodes through a frequency response. As is well known, an ideal Nyquist plot consists of three regions that are dependent on the frequency range. In a high-frequency region, electrochemical behavior related to the interfacial charge transfer resistance of an electrode/electrolyte system is attributed to the capacitive behavior of electrodes.[47] The medium-frequency domain represents diffusion properties due to electrode porosities and surface states. At a high frequency, the nonzero intercept of the real axis indicates the equivalent series resistance (Rs) of a system, and Rs is mainly the combination of the bulk resistance of an electrolyte solution, the intrinsic resistance of an electroactive material, and contact resistance at the interface of the electroactive material and a current collector. The Nyquist plots for PANI/CNTs/CFP and PANI/CFP electrodes are shown in Figure B. The equivalent circuit (the inset picture in Figure B) in ZView software was used in fitting the obtained Nyquist plots. For those two spectra, a small semicircle was observed in the high-frequency region. The high-frequency intercept of the semicircle with the real axis representing Rs was 0.258 Ω·cm–2 for PANI/CNTs/CFP and 1.386 Ω·cm–2 for PANI/CFP. The same as the difference of Rs between PANI/CNTs/CFP and PANI/CFP, Rct of PANI/CNTs/CFP (0.048 Ω·cm–2) was lower than that of PANI/CFP (11.6 Ω·cm–2). The straight lines in the lower-frequency region of PANI/CNTs/CFP were more vertical to the real axis than those of PANI/CFP, indicating the better ideal capacitive behavior of the PANI/CNTs/CFP composite in the low-frequency region than PANI/CFP. The low Rs was attributed to the low contact resistance between PANI and the CNTs/CFP substrate. The 3D microstructure and the formation of a thin PANI layer promoted electrolyte ion transport and the charge transfer between the electrode and electrolyte.

The electrochemical performances at the same PANI mass loading on the same geometric areas of CNTs/CFP and CFP were analyzed, as shown in Figure A,B. PANI was electropolymerized through a galvanostatic (GAL) method, and the same specific areal mass loading of PANI was controlled at the same current density and polymerization time (Supporting Information, Figure S7). The areal-specific mass loading of PANI on CNTs/CFP and CFP was calculated to be about 0.34 mg·cm–2. The gravimetric specific capacitance on CNTs/CFP at the same mass loading of PANI was 998.3 F·g–1, whereas that calculated for PANI on CFP was 781.4 F·g–1 at a scan rate of 5 mV·s–1. Because the nucleation and growing process of PANI under GAL conditions is different from that of PANI prepared by the cyclic voltammetry method. The PANI film prepared by the CV method usually shows more fine porosity and the film is likely to have high quality than that deposited by the GAL method. Consequently, the electrochemical performance of PANI synthesized by GAL is not as good as the PANI prepared by the CV method.[58,59] The differences in properties of PANI deposited by GAL and CV methods were also reflected in the electrochemical impedance parameters. Although the areal specific mass loading of PANI deposited by the GAL method was lower, the Rs (0.58 Ω·cm–2) and Rct (7.99 Ω·cm–2) values of PANI/CNTs/CFP were higher than those of the PANI/CNTs/CFP synthesized by the CV method as described above. In comparison, the Rs and Rct values of PANI/CFP were 0.77 and 17.65 Ω·cm–2, respectively. TEM analysis showed that the average thickness of PANI on CNTs/CFP (Figure C) was approximately 6 nm, and the average thickness of PANI on CFP reached approximately 67 nm (Figure D). The higher Rct of PANI deposited on CFP than that of CNTs/CFP was attributed to the much thicker PANI film formed on the CFP substrate, which would not beneficial to utilize the electrochemical performance of as-prepared PANI more efficiently. Therefore, a thinner PANI film was more likely to be formed on CNTs/CFP than the CFP substrate at the same areal specific mass loading of PANI because of the higher ESA of the former substrate as mentioned above. Consequently, the electrochemical performance of PANI synthesized by the GAL method at the same areal mass loading on CNTs/CFP was improved because of the formation of a thinner PANI layer than that formed on the CFP substrate. Comparatively, a thinner layer of PANI enabled the full utilization of the pseudocapacitance of the as-prepared PANI.

Figure 4

Electrochemical characterization of PANI/CNTs/CFP and PANI/CFP electrodes with same specific areal mass loading of PANI obtained by the GAL method. (A) CV curves of CFP/CNTs, CFP, PANI/CNTs/CFP, and PANI/CFP at a 5 mV·s–1 scan rate. (B) Nyquist impedance spectra of the PANI/CNTs/CFP and PANI/CFP electrodes. (C) TEM of PANI on CNTs/CFP synthesized by the GAL method. (D) TEM of PANI on CFP synthesized by the GAL method.

The CV curves at different scan rates for PANI/CNTs/CFP synthesized by the cyclic voltammetry method are shown in Figure A. A pair of redox peaks could be observed in the potential range of 0–0.4 V (vs Ag/AgCl) at different scan rates. The redox reactions of PANI/CNTs/CFP could be explained by the reversible doping/dedoping reaction of the polymer chain during the charging/discharging process. The position of the anodic peak current gradually shifted to a more positive direction with the increase of the scan rate from 2 to 200 mV·s–1. This result might be because of the ion doping/dedoping process being the kinetic-controlling step for the ion exchange. To gain further insights into the kinetics of the PANI/CNTs/CFP composite, the capacitive and diffusion-limited elements from the total current response were quantitatively separated according to the relationship between the peak current (i) and the sweep rate (v).[60−62]where a and b are the adjustable parameters. The current response is proportional to the scan rate for a capacitive-dominated process (b = 1). However, the current response is proportional to the square root of the scan rate (b = 0.5), which is for a diffusion-controlled process. Figure B shows that the b values for anodic peaks were about 0.82, indicating that the capacitive-controlled behavior greatly contributed to the total current response of PANI/CNTs/CFP.

Figure 5

(A) CV curves of PANI/CNTs/CFP at different scan rates. (B) Curves fitted by the equation of log(i) = b log(v) + log(a) for the anodic peak current of PANI/CNTs/CFP in a three-electrode system. (C) CV curves of PANI/CNTs/CFP at 5 mV·s–1, where the shaded area represents the contribution of capacitive current. (D) Contribution ratio of capacitive and diffusion-controlled processes at different scan rates. (E) Specific capacitance as a function of scan rate for PANI/CNTs/CFP in a three-electrode system. (F) GCD curves of PANI/CNTs/CFP at different current densities.

According to the approach developed by Dunn et al.,[63] the current response at a fixed potential could be expressed as the combination of capacitive-controlled effects (k1v) and the diffusion-controlled effect (k2v1/2) to distinguish and quantify the capacitive contribution to the overall current response furtherwhere v is the scan rate (V·s–1), k1 and k2 are constants under a specific voltage, and k1v and k2v1/2 represent the currents from capacitive-controlled contribution and diffusion-controlled effect, respectively. In Figure C, the shaded region corresponds to a capacitive contribution of about 76.0% for PANI/CNTs/CFP at 5 mV·s–1. Furthermore, the capacitive-controlled contribution played a more dominant role in the total capacity with increasing scan rate (Figure D). In Figure E, the Csp retention was about 72.1% with the scan rate increasing from 2 to 150 mV·s–1, indicating the good rate capability of PANI/CNTs/CFP. The GAL CD (GCD) curves of PANI/CNTs/CFP were measured at current densities from 0.1 to 2 mA·cm–2, as shown in Figure F. The observed discernible plateau at small current densities due to the electrochemical redox processes of conductive PANI remained visible at a high current density of 2 mA·cm–2, which was consistent with the CV diagrams.

To investigate the electrochemical cycling performance of the composites, symmetric coin (SC) cells with PANI/CNTs/CFP as the electrode in 1 mol·L–1 H2SO4 were fabricated. Figure displays the electrochemical performance of the symmetric cells. As shown in Figure A, the redox peaks of PANI/CNTs/CFP were pronounced in the CV curves at different scan rates. The shapes of the CV curves for PANI/CNTs/CFP were almost unchanged when the scan rate increased from 5 to 200 mV·s–1, indicating a fast charge transfer process. The Csp retention at 150 mV·s–1 was 71.3%, indicating the good rate capability of PANI/CNTs/CFP. Figure B demonstrated that the GCD test of PANI/CNTs/CFP presented evident slope changes owing to the pseudoreactions of PANI. The specific capacitance of a single PANI/CNTs/CFP electrode remained at 731.6 mF·cm–2 (1354.7 F·g–1) at a CD rate of 1 mA·cm–2 (1.8 A·g–1). The cycling stability test (Figure C) indicated that the capacitance retention and the coulombic efficiency of PANI/CNTs/CFP SC were approximately 81.4 and 90.7% after 10,000 CD cycles at 1 mA·cm–2, which was higher than the capacitance retention of the PANI/CFP composite (Figure S8) and those of most of the PANI–carbonaceous electrodes reported in other literature (Table ).

Figure 6

Electrochemical characterization of PANI/CNTs/CFP in a symmetric two-electrode system. PANI was synthesized by the CV method. (A) CV curves of PANI/CNTs/CFP at a scan rate of 5 to 200 mV·s–1. (B) GCD curves of PANI/CNTs/CFP at a current density of 0.5 to 5 mA·cm–2. (C) Specific capacitance and the coulombic efficiency at 1 mA·cm–2 current density as a function of CD cycles for the PANI/CNTs/CFP symmetric capacitor.

Table 1

Literature Data Comparison of Specific Capacitance and Cycling Performance of PANI/Carbonaceous-Based Composites with Data from this Work

material	scan or discharge rate	potential range (V)	specific capacitance	capacity retention % (number of cycles)	preparation of PANI or electrode composite	refs
PANI/MC	1 A·g–1	0–0.9	1500 F·g–1a	83% (7000)b	c,d	(18)
PANI/3D graphene	10 mV·s–1	–0.15–0.8	1024 F·g–1a	86.5% (5000)b	c	(40)
PANI/CNTs/CFP/A-CFP	50 mA·cm–2	–0.2–0.8	626 mF·cm–2b	76.5% (5000)a	c	(64)
PANI/hierarchical graphene	400 mV·s–1	–0.1–0.9	601.4 F·g–1a	78.7% (1000)a	c,d	(39)
PANI/CNTs/carbon cloth	0.3 mA·cm–2	0–1.0	1275 F·g–1b	85% (1000)b	c	(65)
PANI/P4VP-g-GMWCNT	0.5 A·g–1	–0.2–0.85	1065 F·g–1a	92.2% (1000)a	c,d	(20)
PANI/porous graphene	4 mA·cm–2	–0.5–0.5	710 F·g–1a	96.8% (1000)a	c	(22)
			440 mF·cm–2b
PANI/GNR/CNTs	0.5 A·g–1	0–0.8	890 F·g–1b	89% (1000)b	c,d	(66)
PANI/CNTs/CFP	5 mV·s–1	–0.2–0.7	1512.2 F·g–1a		e and binder-free	this work
	1 mA·cm–2	0–0.9	731.6 mF·cm–2 (1354.7 F·g–1)b	81.4% (10,000)b		this work
	1.8 A·g–1

In a three-electrode system.

In a two-electrode system.

Chemical polymerization for PANI.

Slurry-casting process for preparing the electrode composite.

Electrochemical polymerization for PANI.

A comparison of specific capacitance and cycling performance between this work and those of previously reported carbonaceous based composites with PANI is summarized in Table . Comparatively, in this work, the hierarchical 3D microstructure carbonaceous substrate was synthesized by directly growing homogeneously distributed CNTs on the surface of CFP via the CVD method without the addition of a binder. Especially, the aggregation of the CNTs was avoided effectively during this one-step process. Second, compared with the chemical method, the electrochemical polymerization of PANI reported here was much faster, operationally simple, environmentally friendly, and provides a fine control of the initiation and termination steps for achieving purer and ultrathin PANI polymerization. The obtained results for specific capacitance and cycling performance in this work were comparable or superior to those of the PANI/carbonaceous based composites reported previously.

Generally, in active materials with a mass loading of less than 1 mg per cm2, a relatively higher loading of active materials is obtained, the higher capacitance value will be achieved.[67] The results above demonstrated that the areal mass loading of PANI on CNTs/CFP (0.54 mg·cm–2) obtained was higher than that on the CFP substrate (0.21 mg·cm–2) after the same CV polymerization process. However, the SEM and TEM images showed that PANI was deposited more uniformly and much thinner on the CNTs/CFP than on the CFP substrate. The specific capacitance, rate capability, and cycling stability improved significantly after the deposition of PANI on CNTs/CFP compared with those on the CFP substrate. This dramatic improvement mainly resulted from the differences among the microstructure characteristics of PANI/CNTs/CFP and PANI/CFP.

As shown in Figure S2, the as-prepared CNTs distributed uniformly on the CFP substrate is a typical hierarchical 3D microstructure and exhibited a higher specific surface area compared with the conventional CFP substrate. At a low mass loading, an ultrathin layer of PANI could be formed on both CNTs/CFP and CFP substrates and the as-deposited PANI would fully contribute to reversal of redox charge. Therefore, a low mass loading of PANI may not have strict requirement for the substrate. Comparatively, at a relatively higher PANI mass loading, the specific structure of CNTs/CFP still enabled the formation of an ultrathin PANI layer coated coaxially on CNT sidewalls. However, the PANI film deposited on the conventional 3D CFP substrate was much thicker than that on the CNTs/CFP substrate at the same higher specific areal mass loading. The thicker PANI layer on CFP may cause redox reactions mainly on or near the surface rather than in the bulk of the PANI film. A part of the bulk PANI on CFP generated “dead” materials, and thus, the utilization efficiency of PANI on the CFP substrate was low. Furthermore, owing to the disintegration and pulverization of the polymer caused by the insertion and expulsion of ions in the polymer structure, the thick PANI film coating on the conventional 3D porous structure CFP electrode was easily cracked and would further cause material exfoliation, leading to rapid capacitance degradation. By contrast, the ultrathin PANI layer retained on the hierarchical 3D microstructure CNTs/CFP substrate at a higher mass loading promoted the interaction between electrolytes and PANI, thereby increasing the utilization efficiency of PANI on the CNTs/CFP. Moreover, the hierarchical 3D microstructure of CNTs/CFP also enabled the facile insertion and expulsion of electrolyte ions into the entire PANI layer, thereby reducing stress on the polymer backbone and subsequently resulting in the high cycling stability of PANI/CNTs/CFP as compared with that of PANI/CFP.

Figure illustrates the mechanism of PANI/CNTs/CFP exhibiting an excellent electrochemical performance compared to the PANI/CFP composite at both the low and high mass loadings. In general, the experimental results above clearly show that the boosted electrochemical performance of the as-prepared PANI/CNTs/CFP compared to the PANI/CFP composite could be ascribed to the following features of the PANI/CNTs/CFP electrode, as illustrated in Figure : (1) the CNTs/CFP with a high surface area served as a strong mechanical backbone providing a high electrochemically active surface for the efficient deposition of the ultrathin PANI layer even at a high mass loading. (2) The binder-free hierarchical 3D microstructure of the CNTs/CFP composite provided enough space and efficient electrolyte transporting channels, thereby promoting the formation of an efficient electrode–electrolyte interface. (3) The ultrathin PANI layer coaxially coated on the sidewalls of CNTs enabled efficient charge transfer, shortened diffusion length, considerably improved the utilization efficiency of PANI, and efficiently alleviated the deterioration of PANI during the CD cycling process. The mechanism of enhancing the electrochemical performance of PANI by forming an ultrathin (nanoscale) film to improve the utilization of PANI and the prepared PANI/carbonaceous nanostructure composites strengthening the mechanical property and improving the conductivity of PANI are consistent with the previous reports.[53,55,68] Therefore, the hierarchical 3D microstructure of the prepared CNTs/CFP is an excellent substrate/scaffold for the efficient deposition and utilization of pseudoactive materials and affords a relatively high mass loading and capacitance without mechanical peeling as compared with a conventional 3D porous structure substrate.[69]

Figure 7

Schematic pictures for comparing the microstructures of CFP (conventional 3D porous structure)- and CNTs/CFP (hierarchical 3D microstructure)-based substrates used for the deposition of PANI.

Conclusions

In this work, the CNTs/CFP composite was successfully synthesized by the decomposition of an ethylene–hydrogen mixture with a Ni catalyst at 1023 K. PANI was synthesized successfully on the CNTs/CFP substrate and coated coaxially as an ultrathin layer on CNT sidewalls in a 0.1 mol·L–1 aniline and 1 mol·L–1 H2SO4 solution through the electrochemical method. The CNTs/CFP, given its hierarchical 3D microstructure, was an excellent binder-free substrate for efficient deposition and utilization of PANI for energy-storage applications. The areal specific capacitance of PANI/CNTs/CFP tested in a symmetric two-electrode system reached 731.6 mF·cm–2 (1354.7 F·g–1) at a current density of 1 mA·cm–2 (1.8 A·g–1). A free electrolyte ion movement and low equivalent series resistance were achieved, and the symmetric supercapacitor device demonstrated an excellent cycling stability performance up to 10,000 cycles and a capacitance retention of 81.4% at a current density of 1 mA·cm–2. The synthesis of CNTs can be easily scaled up. The other kinds of conventional 3D porous substrates, except the commercially available CFP, can be used in growing CNTs. The hierarchical 3D microstructure design developed in this study can be applied to the preparation of various advanced energy-storage electrodes for electrochemical energy-storage and conversion systems.

Experimental Section

Synthesis of the CNTs/CFP Composite

Argon plasma (OKSUN-PR24L) was utilized in the pretreatment of CFP pieces (CeTech) to etch the surfaces of material and improve their hydrophilic properties. The following conditions were used: an argon flow rate of 80 mL·min–1, an operating pressure of 30 Pa in the plasma reaction chamber, a radio frequency (RF) power of 100 W, an ambient temperature of 25 °C, and an operating time of 10 min[35] Metal catalysts were precoated on the etched CFP pieces by the wet impregnation method. The CFP pieces were immersed into an ethanol solution containing 0.1 mol L–1 NiCl2·6H2O for 5 min. After the impregnation of the metal catalysts, the CFP pieces were taken out and dried overnight at room temperature. The CNTs on the as-prepared CFP were allowed to grow in a CVD reacting tube. NiCl2-coated CFP pieces were found in the center of the quartz tube at 750 °C. H2 and Ar (1:4 volume ratio; Yuanzheng Technology) were used as the carrier and reducing gas, respectively, and the total flow rate was 500 sccm for 1 h. Thereafter, the H2/Ar mixture was replaced by a C2H4/H2 mixture (3:5 volume ratio; Yuanzheng Technology) for the growth of CNTs for 1 h.

Synthesis of the PANI/CNTs/CFP and PANI/CFP Composite

Before PANI deposition, the CNTs/CFP composite was electrochemically modified by applying a high potential (10 V, Princeton Applied Research, VersaSTAT MC) for 5 min in a 0.1 mol·L–1 Na2SO4 solution. The purpose was to render the surface of the composite hydrophilic.[12,36] Subsequently, PANI was synthesized by an one-step method on a conventional three-electrode system. An aqueous electrolyte solution of 1 mol·L–1 HCl (Sinopharm) with 0.1 mol·L–1 aniline (Sigma-Aldrich) was used for CV and the GAL electrochemical synthesis of PANI on the CNTs/CFP electrode. PANI synthesis was carried out at −0.2–0.8 V potential (vs an Ag/AgCl reference electrode) at a scan rate of 10 mV·s–1 for 15 cycles. In the GAL method, a current density of 2 mA·cm–2 was used for 360 s. After deposition, the as-prepared samples were washed with distilled water and dried at 60 °C in a vacuum oven for 12 h. For comparison, CFP was used as a substrate for PANI deposition through the same synthesizing process.

Characterization

The morphologies of the CFP and CNTs/CFP, PANI/CFP, and PANI/CNTs/CFP composites were investigated by field emission SEM (FEI Quanta 250 FEG) and high-resolution TEM (FEI TECNAI G2 12, Holland). Raman spectra were recorded on a HORIBA LabRAM HR using a 633 nm laser as the excitation source. FT-IR measurements were recorded at room temperature on a Thermo Scientific Nicolet iS5 FT-IR spectrometer. Surface chemistry and elemental composition of the PANI/CNTs/CFP and PANI/CFP samples were examined by XPS (Thermo ESCALAB 250XI) with Al Kα radiation. The Brunauer–Emmett–Teller (BET) specific surface area of CFP and the CNTs/CFP composite was obtained from nitrogen adsorption measurements at approximately −196 °C performed on a Micromeritics ASAP2460 instrument. Before the measurements, the samples were evacuated at 300 °C for 5 h.

Electrochemical Measurement

The electrochemical properties of the samples were measured on a three-electrode system in 1 mol L–1 H2SO4 solution, and a platinum mesh, an Ag/AgCl (sat.) electrode, and the PANI/CNTs/CFP (or PANI/CFP) samples were used as the counter, reference, and working electrodes, respectively. Symmetrical supercapacitors (SCs) were assembled by using the PANI/CNTs/CFP (or PANI/CFP), glass fiber papers (NKK separator, Nippon Kodoshi Corporation), and 1 mol L–1 H2SO4 aqueous solution as the electrode materials, separators, and electrolyte, respectively. Before assembling the SCs, the glass fiber paper separators were soaked in 1 mol L–1 H2SO4 solution for 0.5 h, and sandwiched between two of the above prepared electrodes. Subsequently, the above sandwich structure was put into the coin battery shells with the edge sealed.

CV curves, GCD curves, and EIS were obtained on a VersaSTAT MC electrochemical workstation (Princeton Applied Research). The EIS spectra were measured in the frequency range of 0.01–100 kHz with an alternating current (AC) perturbation of 10 mV. Operation stability was measured using a battery measuring system (Neware Battery testing system). The specific capacitance of the active materials or SCs was calculated using the following equations

In eq , Csp,single electrode (mF·cm–2 or F·g–1) is the specific capacitance of the CFP/PANI (or CFP/CNTs/PANI) composites calculated from the CV curves in the three-electrode system, and i, v, S, and V are the current (A), scan rate (mV·s–1), geometric surface area (cm–2) or the mass of PANI, and potential window, respectively. In eqs and 5, CSC (mF·cm–2 or F·g–1) is the total specific capacitance of the SC calculated from the CD curves, Csp,single electrode′ is the specific capacitance of a single electrode and calculated from the CSC of the symmetric supercapacitor system, and I (A), ΔV (V), S, and Δt (s) are the current, potential window, the geometric surface area or the mass of PANI, and the time of discharge, respectively.