Fe₃O₄/carbon nanotubes/polyaniline ternary composites with synergistic effects for high performance supercapacitors

Abstract

This work studied the pseudocapacitive behavior of Fe₃O₄ and examined approaches to improve its electrochemical activity. Our results show that by optimizing the electrode architectures, we were able to construct ternary Fe₃O₄/CNTs/PANI composites that delivered excellent electroactivity and cyclic stability because of synergistic effects from each component. In this composite, CNTs were uniformly coated on the surface of Fe₃O₄ particles and were able to form conductive networks that electrically wired each particle, whereas PANI particles were able to bridge Fe₃O₄/CNTs composites and further ensure the overall mechanical integrity. Overall, with 8 wt% of PANI the electrodes were able to deliver a high specific capacitance of 260 F g⁻¹ at 0.5 A g⁻¹. The electrodes also had excellent cyclic stability and showed only 6% decay capacity after 1000 cycles.

---

1 Introduction

With the depletion of fossil fuels and increasing environmental crisis because of carbon dioxide emissions considerable effort has been devoted for developing efficient and sustainable energy storage devices, such as high-energy batteries, supercapacitors and fuel cells.^1,2 Among these energy storage technologies, supercapacitors have several unique advantages including fast charge–discharge rates, longer cycle life and higher power density.^3,4 These advantages make them attractive for a wide range of applications such as consumer electronics, memory back-up systems, energy management, public transportation, and military devices.^1,5 Supercapacitors can be categorized as electric double layer capacitors (EDLCs) and pseudocapacitors on the basis of their energy storage mechanism. Typical materials used for EDLCs are carbon materials that have high surface area and high electrical conductivity,^6,7 whereas materials for pseudocapacitors are mostly metal oxides/hydroxides (such as RuO₂, NiO and MnO₂)^8–10 and conducting polymers.^11,12 In principle, pseudocapacitors can deliver considerably higher specific capacitance than EDLCs, and therefore, have been under intensive investigation in recent years.^6,7,13

Among all of the pseudocapacitor materials, Fe₃O₄ is a particularly attractive material because of its considerably lower cost (compared with, for example, Ni, Co and Ru based materials), low toxicity, high theoretical capacity and relatively low operational potential, which make it suitable for use as an anode material.^14–16 However, Fe₃O₄ has low electrical conductivity and this significantly limits electronic/ionic transfer rate and results in poor capacity under realistic conditions.¹⁷ Low electric conductivity is a common problem for metal oxides and many approaches have been proposed to resolve this problem. In particular, an effective approach is to construct binary and/or ternary nanostructured electrodes by integrating metal oxide particles with carbon hosts (such as MnO₂ with graphene and carbon nanotubes).^1,7,18,19 This is because carbon materials can not only provide effective networks for electrical transport by bridging adjacent metal oxide particles, but also can contribute to the overall capacity through the EDLC mechanism.^17,20–23 On the other hand, it has also been discovered that the pseudocapacitance of metal oxides can be improved by adding an appropriate amounts of conducting polymers.^24–26 In this regard, polyaniline (PANI) is particularly attractive because of its easy preparation, good environmental stability and relatively high electrochemical activity.^19,27,28 However, this material is not stable during charge–discharge processes, and therefore, cannot be used as an active material alone. It should be noted that a combination of PANI with various porous carbon materials and/or transition metal oxides has been shown to be effective for improving its stability, as well as its capacity.²⁸ These previous observations and demonstrations suggest that it is possible to maximize the electrochemical performance of electrode materials through rational selection of functional materials and design of appropriate electrode architectures to enable the synergistic effects of different components.^29–31

In this work we aim to study the pseudocapacitive behavior of Fe₃O₄ and develop effective approaches to maximize its capacity for practical applications. Several combinations of materials, including Fe₃O₄ alone, Fe₃O₄/CNTs binary composites and Fe₃O₄/CNTs/PANI ternary composites, were tested in order to identify the most effective electrode architecture. Overall, we found that electrodes based on the ternary composites delivered the highest specific capacitance, better rate performance and longer cycle life that outperformed both the Fe₃O₄ particles and the Fe₃O₄/CNTs composites. We believe the improved electrochemical performance was due to the effective conductive path constructed by CNTs, as well as the synergistic effects of each component in the ternary composites. The network formed by CNTs can act as “conduction bridges” and enhances the electrical conductance between PANI and Fe₃O₄, as well as providing pathways for charge transfer in the electrode.

2 Experimental details

2.1 Materials

Raw carbon nanotubes were purchased from Chengdu Organic Chemicals Co., Ltd (95% purity, 10–20 nm diameter) and were purified/functionalized using mixed acid (concentrated H₂SO₄/HNO₃, volume ratio 3 : 1) by following a reported procedure.³² Aniline monomer was distilled under reduced pressure prior to use. All the reagents are of analytical grade and were used without further purification. The aqueous solution was freshly prepared with high purity water (≥18 MΩ cm) generated by a Gen Pure UV-TOC/UF ultra-pure water system (TKA, Niederelbert, Germany).

2.2 Synthesis of Fe₃O₄/CNTs composites and Fe₃O₄ particles

Fe₃O₄/CNTs composites were prepared using the solvothermal method.³³ The synthesis procedure is illustrated in Fig. 1. In a typical synthesis, 1.0 g polyvinylpyrrolidone (PVP), 80 mg CNTs and 40 mL ethylene glycol (EG) were mixed and stirred for 40 min. 1.35 g ferric chloride hexahydrate (FeCl₃·6H₂O) and 2.46 g sodium acetate (NaAc) were then added and the mixture was stirred for another 30 min. The mixture was transferred into a Teflon-lined stainless steel autoclave and reacted at 200 °C for 10 h. The reaction product was collected by centrifugation, washed 3 times with ethanol and dried at 60 °C for 6 h. For control purposes, uncoated Fe₃O₄ particles were also prepared without adding CNTs using the same method.

Fig. 1 Illustration of the synthesis process for the Fe₃O₄/CNTs composites.

2.3 Synthesis of PANI

PANI particles were synthesized using a method previously reported.³⁴ Typically, 0.2 mL aniline monomer was mixed with 0.07 mL concentrated H₃PO₄ and 15 mL water and the mixture was sonicated for 10 min to form an emulsion of aniline/H₃PO₄ complex. Then, 5 mL aqueous solution of ammonium persulfate (0.46 g) was added. The reaction was placed under ultrasonic conditions for 4 h and the temperature was controlled between 25 °C and 30 °C. The resulting precipitate was collected and was washed with water and ethanol. Finally, the product was dried under vacuum for 24 h.

2.4 Preparation of ternary Fe₃O₄/CNTs/PANI composites

The ternary Fe₃O₄/CNTs/PANI composites were prepared by directly mixing PANI particles with binary composites. For convenience, Fe₃O₄/CNTs powder was mechanically mixed with PANI in a container and the electrodes that contained 6 wt%, 8 wt% and 10 wt% PANI were designated as Fe₃O₄/CNTs/P₆, Fe₃O₄/CNTs/P₈ and Fe₃O₄/CNTs/P₁₀, respectively.

2.5 Materials characterization

Fourier transform infrared (FTIR) spectra were collected in a frequency range of 400–4000 cm⁻¹ on a Bruker TENSOR27 infrared spectrophotometer (BRUKER, Ettlingen, Germany). X-ray powder diffraction (XRD) patterns were recorded using a Bruker D8 advanced X-ray diffractometer with Cu Kα radiation. Scanning electron microscope (SEM) and transmission electron microscope (TEM) images were collected using a JEOLJSM-6700F Field Emission SEM and a JEOLJEM-100CX-II TEM.

2.6 Electrochemical measurements

All electrochemical measurements were performed in a three-electrode electrochemical cell with a saturated calomel reference electrode (SCE) and a platinum sheet (1 × 2 cm²) counter electrode using a CHI 760C electrochemical workstation. The working electrodes were prepared by a standard approach, by mixing 80 wt% composite material, 15 wt% acetylene black (AB) and 5% polyvinylidene fluoride (PVDF) with N-methyl pyrrolidinone to form a homogeneous slurry. The slurry was coated onto a piece of Ni mesh. Typical active material loading was approximately 1.0 mg cm⁻².

The electroactivity of different materials was studied using galvanostatic charge–discharge and cyclic voltammetry (CV) techniques. Electrochemical impedance spectroscopy (EIS) measurements were performed in the frequency range of 0.05 Hz to 100 kHz under an excitation of a sinusoidal wave of 5 mV at open circuit potential. The impedance data were fitted using ZSimpWin (ver. 3.10).

The specific capacitance (SC) (F g⁻¹), specific energy density (SE) (W h kg⁻¹) and maximum power density (P_max) (W g⁻¹) of the supercapacitors were calculated using the following equations:

where I is the discharge current (A), Δt is the discharge time (s), V is the potential change during the discharge (V), m is the weight of the active materials (g), U₀ is the potential at the beginning of discharge (after the ohm drop) and R is the equivalent series resistance (Ω).

3 Results and discussion

The morphology of the synthesized materials was studied using SEM and TEM. The Fe₃O₄ particles were sub-micron particles with diameters ranging from 100 nm to 600 nm as can be seen from the SEM image shown in Fig. 2a. Fig. 2b shows a typical SEM image of the PANI particles; these particles had a variety of shapes, and some of the particles were aggregated and formed flower-like structures with micron size. Fig. 2c shows a typical TEM image of the Fe₃O₄/CNTs composites. It can be clearly seen that CNTs were uniformly coated on the surface of the Fe₃O₄ particles and generated interconnected networks that could bridge adjacent Fe₃O₄ particles when used as electrode materials (Fig. 2c, inset). Thus, CNTs could not only act as a conductive agent, but also construct a conductive network. Fig. 2d shows a typical TEM image of the ternary Fe₃O₄/CNTs/PANI composites. It was observed that PANI particles were able to fill the gaps between adjacent Fe₃O₄ particles. The formation of such a well-organized structure might be obtained by the self-assembly/organization of particles with different sizes and morphologies upon drying. Shrinking and swelling of the PANI structure during cycling will form a loose network that exhibits a large elastic buffer space and further boosts the accessible capacity.^24,35 The interconnection and interactions between PANI, CNTs and Fe₃O₄ particles could effectively promote both electronic and ionic conductions throughout the composite materials, and therefore, could significantly improve their electrochemical performance, as discussed below.

Fig. 2 SEM images of (a) Fe₃O₄ particles and (b) PANI particles; TEM images of (c) Fe₃O₄/CNTs binary composites and (d) Fe₃O₄/CNTs/P₈ ternary composites. The inset image in (c) shows the effective conductive network constructed by CNTs.

The crystalline properties of the as-synthesized Fe₃O₄, PANI, Fe₃O₄/CNTs and Fe₃O₄/CNTs/P₈ composites were studied using XRD, and obtained results are shown in Fig. 3a. The XRD pattern of PANI powder has two broad and weak peaks ranging from 20° to 25°, which can be considered as the characteristic peaks of amorphous PANI.³⁴ The XRD patterns of Fe₃O₄ and Fe₃O₄/CNTs both exhibit six characteristic peaks at 2θ = 30.3°, 35.7°, 43.3°, 53.4°, 57.4° and 62.5°, which can be indexed to the (220), (311), (400), (422), (511), and (440) planes of face-centered cubic Fe₃O₄ (JCPDS card no. 19-0629), respectively.³³ On the basis of these results, the as-synthesized Fe₃O₄ and Fe₃O₄/CNTs composites had good crystal structure. The characteristic peaks of Fe₃O₄ and PANI can also be observed in the XRD pattern of the ternary Fe₃O₄/CNTs/P₈ composites.

Fig. 3 (a) XRD patterns of PANI, Fe₃O₄/CNTs and Fe₃O₄/CNTs/P₈ composites; (b) FTIR spectra of PANI, Fe₃O₄/CNTs and Fe₃O₄/CNTs/P₈ composites. The inset image in (b) shows Fe₃O₄/CNTs composites dispersed in water in the absence (left) and presence (right) of an applied magnetic field.

FTIR was used to analyze the surface chemistry of the PANI particles, Fe₃O₄/CNTs and Fe₃O₄/CNTs/P₈ composites. A typical spectrum of PANI is shown in Fig. 3b, and characteristic peaks located at ∼1582 cm⁻¹ and 1150 cm⁻¹ were observed. These peaks can be attributed to the characteristic vibration responses of the N=N and C–N bonds in PANI, respectively.³⁴ The spectrum of Fe₃O₄/CNTs composites is also shown in this figure and the composite showed peaks located at ∼590 and 450 cm⁻¹, which correspond to the Fe–O stretching resulting from the interaction of Fe–O–Fe in Fe₃O₄. In addition to these two peaks, a peak at 1600 cm⁻¹ was also observed and this peak can be assigned to the COO– stretching band from surface functionalized CNTs. The IR peaks from PANI and Fe₃O₄/CNTs composites are also present in the spectrum of the ternary Fe₃O₄/CNTs/P₈ composites (marked with a blue line). The inset image in Fig. 3b shows the response of the synthesized Fe₃O₄/CNTs composites under an applied magnetic field, showing that when a colloidal dispersion containing the composites was placed near a magnet the composites were quickly concentrated near the magnet, further demonstrating the successful formation of Fe₃O₄ particles.

The electrochemical properties of bare Fe₃O₄, Fe₃O₄/CNTs binary composites and Fe₃O₄/CNTs/PANI ternary composites were studied in 1 M KOH using CV, galvanostatic charge–discharge and EIS techniques. Fig. 4a shows a set of CVs of electrodes prepared with different materials as noted at 5 mV s⁻¹ using a relatively wide potential window of −1.1 to 0 V (vs. SCE). All of the current densities have been normalized based on the mass of electroactive material. The specific currents for CNTs and PANI were noticeably lower than that of the Fe₃O₄-based electrodes (Fig. 4a, inset), suggesting that both of them have low capacity alone and the pseudocapacitance from Fe₃O₄ was able to substantially improve the overall capacity. This also shows that all of the Fe₃O₄-based electrodes have characteristic pseudocapacitive responses related to surface redox-type reactions with considerably higher specific currents. The pseudocapacitance of Fe₃O₄ in KOH electrolyte may originate from the redox reactions between Fe²⁺ and Fe³⁺ with the intercalation of hydroxide ions shown as follows:^36,37

Fe^IIO + 2OH⁻ ↔ Fe^II(OH)₂ + 2e⁻

2Fe^IIO + 2OH⁻ ↔ (Fe^IIIO)⁺(OH⁻)₂(FeO^IIIO)⁺ + 2e⁻

Fig. 4 Comparison of CVs of (a) electrodes prepared with different materials at 5 mV s⁻¹, (b) Fe₃O₄/CNTs/PANI composites with different amounts of PANI at the scan rate of 20 mV s⁻¹, (c) Fe₃O₄/CNTs/P₈ composites at different scan rates, and (d) variation of specific capacitance at different scan rates. The inset image in (a) shows the CVs of PANI and CNTs at the scan rate of 20 mV s⁻¹.

Furthermore, the CV of the Fe₃O₄/CNTs electrode shows considerably higher specific current than the electrode made with pure Fe₃O₄. Such pronounced enhancement suggests that the binary composites may have more accessible sites and higher utilization of Fe₃O₄ due to better electrical conductivity with the presence of CNTs.³⁸ The CV of the ternary Fe₃O₄/CNTs/PANI composites shows a pair of broad and reversible redox peaks from the reactions of PANI and has the highest specific current among all of the materials studied, suggesting that a rational combination of these three materials is very effective for improving electrode performance. Fig. 4b presents a set of CVs for the Fe₃O₄/CNTs/PANI composites with different amounts of PANI at 20 mV s⁻¹. It is also evident that the capacitive currents do not change monotonously with the increase in PANI content, but strongly depend on the composition, suggesting strong synergistic effects between these three compositions. Overall, it was observed that the electrode assembled with 8 wt% PANI had the highest specific capacitance. The behavior of the ternary composites with 8 wt% of PANI under increased scan rates is shown in Fig. 4c. The overall shape of CV curves did not exhibit obvious distortions when the scan rate was increased from 5 to 50 mV s⁻¹. This indicates that this electrode had good reversibility and very rapid current response to the reversal of voltage at each terminal potential.³⁹ On the basis of these results it is evident that the ternary composite based electrodes delivered more ideal supercapacitive behavior.

Fig. 5a shows a set of typical charge–discharge curves of Fe₃O₄, Fe₃O₄/CNTs, and Fe₃O₄/CNTs/P₈ obtained with a potential range of −1.0 to 0 V at 0.5 A g⁻¹. The behavior of pure CNTs and PANI are also included for comparison (Fig. 5a, inset). The charge–discharge characteristics of Fe₃O₄/CNTs/P₈, Fe₃O₄/CNTs and Fe₃O₄ were similar, but the time required for fully charging or discharging the ternary composite was noticeably longer. The specific capacitance of different combinations of materials was determined using eqn (1) and the results are summarized in Fig. 5d. In good agreement with the CV tests, it was found that the ternary Fe₃O₄/CNTs/P₈ composites delivered the highest specific capacitance of 260 F g⁻¹, whereas the capacitances of Fe₃O₄/CNTs and Fe₃O₄ electrodes were 208 and 128 F g⁻¹, considerably lower, respectively. Furthermore, we also studied ternary composites with different amounts of PANI in order to identify optimal compositions (Fig. 5b); the specific capacitances of Fe₃O₄/CNTs/P₆, Fe₃O₄/CNTs/P₈ and Fe₃O₄/CNTs/P₁₀ in 1 M KOH at 1 A g⁻¹ are 156 F g⁻¹, 160 F g⁻¹ and 138 F g⁻¹, respectively. The specific capacitance of the ternary composites also strongly depends on the content of PANI (i.e. wt%) and the best performance was obtained with a loading of 8 wt%, which corresponds well with the CV results (Fig. 4b). It is interesting to note that the composite with 10 wt% PANI had an even lower SC than the composite without PANI. This decrease in activity might be because of the aggregation of PANI particles when their content is too high, which affects the overall interconnection between Fe₃O₄, CNTs and PANI and decreases the synergistic effect.

Fig. 5 Galvanostatic charge–discharge curves of (a) electrodes prepared with different materials at 0.5 A g⁻¹, (b) Fe₃O₄/CNTs/PANI composites with different amounts of PANI at 1 A g⁻¹, (c) Fe₃O₄/CNTs/P₈ composites at different current densities, and (d) variation of specific capacitance at different discharge currents. The inset image in (a) shows the charge–discharge curves of PANI and CNTs at the current density of 1 A g⁻¹.

The rate performance of different electrode materials was also studied. Fig. 5c shows the galvanostatic charge–discharge profiles of the Fe₃O₄/CNTs/P₈ electrode at increasing current densities from 0.5 to 5 A g⁻¹. The specific capacitance was calculated and compared with Fe₃O₄/CNTs binary composites and pure Fe₃O₄ in Fig. 5d. The ternary composite had the highest specific capacitance and outperformed both of the other two materials under all of the current densities studied, and was able to deliver a high specific capacitance of 122 F g⁻¹ at 5.0 A g⁻¹. Recently, Fe-based electrode materials have been demonstrated to display excellent capacitive properties as promising high-performance materials in supercapacitors. When compared to the electrochemical performances previously reported for Fe-based materials, the specific capacitances obtained are comparable and directly reflect the effectiveness of the active materials (see Table 1). Nonetheless, we would like to mention that the scope of this work is to identify approaches to improve the performance of iron oxides and to demonstrate synergistic effects; we think our approach for the determination of specific capacitance should be effective.

Table 1 Comparison of the electrochemical performance of the Fe-based oxides in this work with previous reports

Materials	SC	Cell type	Electrolyte	Potential window	Processing of electrode	Ref.
Fe3O4/CNFs	135 F g^−1 at 0.42 A g^−1	3-ED	1 M Na2SO4	−0.9–0.1 V	Mixed with PVDF (5%), and pressed onto Ni foam	7
Fe3O4 nanoparticles	37.9 F g^−1 at 0.5 mA cm^−2	2-ED	6 M KOH	0–1.2 V	Mixed with AB (10%) and PTFE (5%), and pressed onto Ni grid	14
Fe3O4/CNTs	110.5 F g^−1 at 0.5 A g^−1	3-ED	1 M KOH	−1.0–0 V	Mixed with AB (10%) and PVDF (10%), and pressed onto Ni foam	17
Graphene/Fe2O3/polyaniline	638 F g^−1 at 1 mV s^−1	3-ED	1 M KOH	−1–0.1 V	Mixed with AB (10%) and PTFE (5%), and pressed onto Ni foam	24
Fe3O4/reduced graphene oxide	160.5 F g^−1 at 0.5 A g^−1	3-ED	6 M KOH	−1.1–0 V	Mixed with AB (10%) and PTFE (10%), and pressed onto Ni foam	35
Fe3O4 films	118.2 F g^−1 at 6 mA	3-ED	1 M Na2SO3	−1.2–0.2 V	Deposited onto stainless steel foil	40
Fe3O4–GNS	88 F g^−1 at 0.25 A g^−1	2-ED	1 M H2SO4	0–1 V	Mixed with KB (10%) and TAB (10%), and pressed onto stainless steel mesh	41
a-Fe2O3 nanotubes/reduced graphene oxide	69 F g^−1 at 10 A g^−1	3-ED	1 M Na2SO4	−1–0 V	Mixed with super P carbon (15%) and Kynar 2801 (15%), and pressed onto etched-copper foil	42
Fe3O4@SnO2 core–shell nanorod film	2.7 mF cm^−2 at 1.0 mA cm^−2	3-ED	1 M Na2SO3	−0.7–−0.2 V	Nanorod films were directly used as the working electrode	43
CNT/Fe3O4	129.3 F g^−1 at 2.5 mA cm^−2	3-ED	6 M KOH	−1–0.1 V	Mixed with AB (15%) and PTFE (5%), and pressed onto Ni foam	44
Fe3O4/graphene sheet	368 F g^−1 at 1 A g^−1	3-ED	1 M KOH	−1–0 V	Mixed with AB (10%) and PTFE (10%), and pressed onto Ni foam	45
N-doped graphene/Fe3O4	212 F g^−1 at 1 A g^−1	3-ED	1 M H2SO4	−0.2–0.8 V	Mixed with AB (10%) and PVDF (10%), and pressed onto graphite plate	46
Fe3O4/CNTs/P8	122 F g^−1 at 5 A g^−1	3-ED	1 M KOH	−1–0 V	Mixed with AB (15%) and PVDF (5%), and pressed onto Ni mesh	This work

---

EIS is an effective tool to analyze the internal resistance and the electron transfer rate at the electrode–electrolyte interface. Fig. 6 shows typical Nyquist plots for electrodes made with the three materials as noted; all of the plots showed a straight line at low frequencies and a semi-circle at high frequencies. Theoretically, the Nyquist plot for an ideal capacitor connecting its series with a resistor is a straight line parallel to the vertical axis. The straight lines for the spectra shown in Fig. 6; however, all had some degrees of deviation depending on the active materials, suggesting that the electrodes did not have ideal capacitive behavior. The capacitive loop in the high-frequency range is attributed to the relaxation process of the double-layer capacitance (C_dl) and the charge-transfer resistance (R_ct) due to Faradaic reactions. Overall, the impedance plots may be analyzed with the equivalent circuit shown in the inset of Fig. 6. In this circuit, R_s represents internal resistance, Q stands for a constant phase element to represent the C_dl, W stands for the Warburg impedance related to the transport process of electrochemical active species and C_L is the limit capacitance. The circuit element values were determined by fitting the impedance spectra and were listed in Table 2. The addition of CNTs and PANI is helpful in improving the capacitive behavior of Fe₃O₄. On the basis of EIS measurements, a calculation of the maximum power densities of Fe₃O₄/CNTs/P₈, Fe₃O₄/CNTs, and Fe₃O₄ based on eqn (3) gives 6.2 W g⁻¹, 6.0 W g⁻¹ and 5.3 W g⁻¹, respectively. Thus, the ternary composites exhibited the largest SC values and the improved electrochemical performance of the composite electrode could be because of the good conductivity of the carbon nanotubes and the synergistic effects between each component.

Fig. 6 Nyquist plots for various electrodes. The inset shows the equivalent circuit used to fit the impedance spectra.

Table 2 The calculated values of R_s, R_ct, W, Q and C_L through fittings of the experimental impedance spectra based on the proposed equivalent circuit in Fig. 8

	Rs (Ω cm^2)	Rct (Ω cm^2)	W (Ω^−1 s^1/2 cm^−2)	Q		CL (F cm^−2)
				Y (Ω^−1 s^n cm^−2)	n
Fe3O4	0.32	0.78	0.050	0.00030	0.78	0.0076
Fe3O4/CNTs	0.23	0.40	0.095	0.00015	0.89	0.0096
Fe3O4/CNTs/P8	0.25	0.42	0.098	0.000062	0.99	0.011

---

The cyclic stability of electrodes prepared with Fe₃O₄/CNTs/P₈, Fe₃O₄/CNTs and Fe₃O₄ was evaluated by a 1000 cycle galvanostatic charge–discharge test conducted at 5 A g⁻¹. The changes in specific capacitance as a function of cycle number are summarized and compared in Fig. 7a. The electrode with Fe₃O₄/CNTs/P₈ composites had excellent stability and its capacity only decayed by ∼6% after being cycled for 1000 times. In contrast, the capacity decays for the Fe₃O₄/CNTs and Fe₃O₄ electrodes were considerably higher and were 9% and 17%, respectively.

Fig. 7 (a) Cyclic stability tests for Fe₃O₄, Fe₃O₄/CNTs and Fe₃O₄/CNTs/P₈ composites measured at a current density of 5 A g⁻¹; (b) comparison of specific energy densities of various materials as a function of discharge current densities. The inset image in (a) shows typical charge–discharge curves for Fe₃O₄/CNTs/P₈.

On the basis of the results discussed above, it is evident that electrodes based on the ternary Fe₃O₄/CNTs/P₈ composites have the best performance metrics for practical applications. The excellent capacitive behavior of the ternary composites could be attributed to the synergistic effects between each component when constructed properly. Such effects could be better explained by the illustration shown in Fig. 8 in which interactions between each component are shown. In this composite, Fe₃O₄ is the main electrochemically active material and its redox process provides most of the capacity. CNTs could act as conductive bridges to connect individual Fe₃O₄ particles and improve the overall electrical conductivity.^38,47 PANI particles can alleviate the volume changes of the active materials during redox reactions and prevent the agglomeration/detachment of Fe₃O₄ particles.^24,48 Working together, the synergistic effect between Fe₃O₄, CNTs and PANI enables the ternary Fe₃O₄/CNTs/P₈ composites with greatly improved electrochemical performance.

Fig. 8 Sketch highlighting the conductive network in ternary Fe₃O₄/CNTs/PANI composites.

4 Conclusions

This paper demonstrates that ternary Fe₃O₄/CNTs/P₈ composites assembled by proper mixing of PANI particles with binary Fe₃O₄/CNTs composites have high electrochemical performance because of the synergistic effects between each component. In this ternary composite, CNTs were uniformly coated on the surface of Fe₃O₄ particles and were able to form conductive networks that electrically wired each particle, whereas PANI particles were able to bridge Fe₃O₄/CNTs composites and further ensure good mechanical integrity. Overall, the ternary composite outperformed both the binary Fe₃O₄/CNTs composite and pure Fe₃O₄ and was able to deliver the highest specific capacitance of 260 F g⁻¹ at 0.5 A g⁻¹ and 122 F g⁻¹ at 5 A g⁻¹. The ternary composite also had excellent cyclic stability and its capacitance only decreased by 6% after 1000 cycles. The synergistic effects demonstrated in this work for improving the performance of Fe₃O₄ could have general significance for designing electrode architectures and could be readily applied for the design of other electrochemical energy storage systems, where high energy density and high power density devices are desired.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21373129). We would like to thank Dr Yingwen Cheng in Pacific Northwest National Laboratory (Richland, USA) for helpful discussions.

Notes and references

1. Y. Hou, Y. W. Cheng, T. Hobson and J. Liu, Nano Lett., 2010, 10, 2727–2733.
2. G. Wang, L. Zhang and J. Zhang, Chem. Soc. Rev., 2012, 41, 797–828.
3. E. Frackowiak, Phys. Chem. Chem. Phys., 2007, 9, 1774–1785.
4. P. Simon and Y. Gogotsi, Nat. Mater., 2008, 7, 845–854.
5. C. Liu, F. Li, L. P. Ma and H. M. Cheng, Adv. Mater., 2010, 22, E28–E62.
6. A. L. M. Reddy, M. M. Shaijumon, S. R. Gowda and P. M. Ajayan, J. Phys. Chem. C, 2010, 114, 658–663.
7. J. Mu, B. Chen, Z. Guo, M. Zhang, Z. Zhang, P. Zhang, C. Shao and Y. Liu, Nanoscale, 2011, 3, 5034–5040.
8. T. Cottineau, M. Toupin, T. Delahaye, T. Brousse and D. Belanger, Appl. Phys. A: Mater. Sci. Process., 2006, 82, 599–606.
9. M. Toupin, T. Brousse and D. Bélanger, Chem. Mater., 2004, 16, 3184–3190.
10. D. W. Wang, F. Li and H. M. Cheng, J. Power Sources, 2008, 185, 1563–1568.
11. C. P. Fonseca, J. E. Benedetti and S. Neves, J. Power Sources, 2006, 158, 789–794.
12. H. Wang, Q. Hao, X. Yang, L. Lu and X. Wang, Electrochem. Commun., 2009, 11, 1158–1161.
13. Y. Chen, X. Zhang, D. Zhang, P. Yu and Y. Ma, Carbon, 2011, 49, 573–580.
14. X. Du, C. Wang, M. Chen, Y. Jiao and J. Wang, J. Phys. Chem. C, 2009, 113, 2643–2646.
15. C. Yang, J. J. Wu and Y. L. Hou, Chem. Commun., 2011, 47, 5130–5141.
16. J. Liang, Y. Huang, J. Oh, M. Kozlov, D. Sui, S. Fang, R. H. Baughman, Y. Ma and Y. Chen, Adv. Funct. Mater., 2011, 21, 3778–3784.
17. Q. H. Wang, L. F. Jiao, H. M. Du, Y. J. Wang and H. T. Yuan, J. Power Sources, 2014, 245, 101–106.
18. X. Zhao, L. L. Zhang, S. Murali, M. D. Stoller, Q. H. Zhang, Y. W. Zhu and R. R. Ruoff, ACS Nano, 2012, 6, 5404–5412.
19. J. Zhang and X. S. Zhao, J. Phys. Chem. C, 2012, 116, 5420–5426.
20. Z. S. Wu, D. W. Wang, W. Ren, J. Zhao, G. Zhou, F. Li and H. M. Cheng, Adv. Funct. Mater., 2010, 20, 3595–3602.
21. Z. Fan, J. Yan, T. Wei, L. Zhi, G. Ning, T. Li and F. Wei, Adv. Funct. Mater., 2011, 21, 2366–2375.
22. S. Chen, J. W. Zhu, X. D. Wu, Q. F. Han and X. Wang, ACS Nano, 2010, 4, 2822–2830.
23. R. Jiang, C. Cui and H. Ma, Electrochim. Acta, 2013, 104, 198–207.
24. X. Xia, Q. Hao, W. Lei, W. Wang, D. Sun and X. Wang, J. Mater. Chem., 2012, 22, 16844–16850.
25. D. Yan, Y. Liu, Y. Li, R. Zhuo, Z. Wu, P. Ren, S. Li, J. Wang, P. Yan and Z. Geng, Mater. Lett., 2014, 127, 53–55.
26. G. Ma, H. Peng, J. Mu, H. Huang, X. Zhou and Z. Lei, J. Power Sources, 2013, 229, 72–78.
27. Y. G. Wang, H. Q. Li and Y. Y. Xia, Adv. Mater., 2006, 18, 2619–2623.
28. J. Yan, T. Wei, Z. Fan, W. Qian, M. Zhang, X. Shen and F. Wei, J. Power Sources, 2010, 195, 3041–3045.
29. Y. Cheng, S. Lu, H. Zhang, C. V. Varanasi and J. Liu, Nano Lett., 2012, 12, 4206–4211.
30. C. Zhu, J. Zhai, D. Wen and S. Dong, J. Mater. Chem., 2012, 22, 6300–6306.
31. J. J. Xu, K. Wang, S. Z. Zu, B. H. Han and Z. X. Wei, ACS Nano, 2010, 4, 5019–5026.
32. J. Liu, A. G. Rinzler, H. Dai, J. H. Hafner, R. K. Bradley, P. J. Boul, A. Lu, T. Iverson, K. Shelimov, C. B. Huffman, F. Rodriguez-Macias, Y.-S. Shon, T. R. Lee, D. T. Colbert and R. E. Smalley, Science, 1998, 280, 1253–1256.
33. Q. Q. Xiong, J. P. Tu, Y. Lu, J. Chen, Y. X. Yu, Y. Q. Qiao, X. L. Wang and C. D. Gu, J. Phys. Chem. C, 2012, 116, 6495–6502.
34. X. Lu, H. Mao, D. Chao, W. Zhang and Y. Wei, J. Solid State Chem., 2006, 179, 2609–2615.
35. T. Qi, J. Jiang, H. Chen, H. Wan, L. Miao and L. Zhang, Electrochim. Acta, 2013, 114, 674–680.
36. Y. H. Kim and S. J. Park, Curr. Appl. Phys., 2011, 11, 462–466.
37. J. Pu, L. Shen, S. Zhu, J. Wang, W. Zhang and Z. Wang, J. Solid State Electrochem., 2014, 18, 1067–1076.
38. Z. S. Wu, G. Zhou, L. C. Yin, W. Ren, F. Li and H. M. Cheng, Nano Energy, 2012, 1, 107–131.
39. Z. Tai, X. Yan and Q. Xue, J. Electrochem. Soc., 2012, 159, A1702–A1709.
40. J. Chen, K. Huang and S. Liu, Electrochim. Acta, 2009, 55, 1–5.
41. K. Karthikeyan, D. Kalpana, S. Amaresh and Y. S. Lee, RSC Adv., 2012, 2, 12322–12328.
42. K. K. Lee, S. Deng, H. M. Fan, S. Mhaisalkar, H. R. Tan, E. S. Tok, K. P. Loh, W. S. Chin and C. H. Sow, Nanoscale, 2012, 4, 2958–2961.
43. R. Li, X. Ren, F. Zhang, C. Du and J. Liu, Chem. Commun., 2012, 48, 5010–5012.
44. D. Guan, Z. Gao, W. Yang, J. Wang, Y. Yuan, B. Wang, M. Zhang and L. Liu, Mater. Sci. Eng., B, 2013, 178, 736–743.
45. M. Liu and J. Sun, J. Mater. Chem. A, 2014, 2, 12068–12074.
46. L. Li, Y. Dou, L. Wang, M. Luo and J. Liang, RSC Adv., 2014, 4, 25658–25665.
47. G. Yu, L. Hu, N. Liu, H. Wang, M. Vosgueritchian, Y. Yang, Y. Cui and Z. Bao, Nano Lett., 2011, 11, 4438–4442.
48. Y. Liu, K. Huang, H. Luo, H. Li, X. Qi and J. Zhong, RSC Adv., 2014, 4, 17653–17659.

---