Introduction

Renewable energies such as wind, solar, and tidal have recently experienced a remarkably rapid growth due to the increasing energy demand [1, 2]. However, the power generated from these sources always fluctuates, thus new energy storage devices are needed to be developed [3]. Supercapacitor has a lot of advantages such as high power density, long cycling life, wide temperature range [4–6]. It can release electric field energy through controlling units when needed, improve stability, and reliability of the system by the instantaneous power compensation. It has wide application prospects in clean power system, aviation, automobile, consumer electronics, and communication areas [4–7].

Currently, the main research of supercapacitor is developing electrode materials characterized by high energy density and steady performance [8–10]. Graphene (GN), a single layer of sp2-bonded carbon in a honeycomb crystal lattice, is considered to be a kind of new and high efficient electrode material with high electrical conductivities, specific surface area, and mechanical strength [11, 12]. However, capacitance performance of GN is usually poor as it is prone to conglomeration [13]. Hence, the application of GN alone as electrode material is limited. On the other hand, polyaniline (PANI) is the most outstanding conducting polymer among supercapacitor electrode materials because of its low cost of aniline monomers, better environmental stability, and simple synthesis process [10, 14]. Its unique doping behavior and bulk phase reaction mechanism ensure its good redox reversibility and high pseudocapacitance. Nevertheless, its comparatively low conductivity, low power density, and poor cycling stability limit the application of PANI in supercapacitor [15].

Taking advantage of their energy storage characteristics, combining GN with PANI should be effective to fabricate high-performance electrode materials. Indeed, Wu et al. [16] obtained chemically modified GN/polyaniline nanofibers (PANI-NFs) composite films by vacuum filtration. The film, free-standing and flexible, can be used as working electrode directly. As reported by Zhou et al. [17], GN-wrapped PANI-NFs were prepared by the self-assembly of negatively charged graphene oxide and positively charged PANI-NFs in an aqueous dispersion. Specific capacitance (C sp) of composite electrodes is up to 250 F g−1 at the current density of 500 mA g−1. One-step electrochemical synthesis was also used to produce GN/PANI composite films by Feng et al. [18], used in high-performance supercapacitor and H2O2 biosensor. Hao et al. [19] fabricated sulfonated GN/PANI nanocomposites by liquid/liquid interfacial polymerization. The C sp of composites is up to 763 F g−1 at a scan rate of 1 mV s−1, and its capacity retention is of 96 % after 100 cycles. Currently, main challenges for GN/PANI composites remain to relieve the agglomeration of GN, enhance its surface wettability in aqueous electrolytes, enlarge electrode materials effective specific surface area, and optimize microstructure of PANI.

Surfactants, unique amphiphilic molecules, can be adsorbed on solid surface. Steric hindrance from its long molecular chains can relieve the agglomeration of nanoparticle. Besides, anionic and nonionic surfactants with high surface activities could reduce the surface tension of liquid–liquid and solid–liquid. They can be employed as wetting agents to improve solid surface wettability. In addition, ordered surfactant aggregates, such as micelles, vesicles, and liquid crystals can be acted as micro reactors or soft templates to realize the regulation and control of nanomaterials morphology and structure. In recent years, several researchers have reported the surfactant applied in fabricating supercapacitor electrode materials [20–23]. It is found that surfactants have some influences on morphology, structure, and electrochemical performance of nano-materials.

Triblock copolymer poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol) (P123), a nonionic surfactant, has been widely used to fabricate ordered mesoporous materials such as SBA-15 [24]. In this paper, we report for the first time that P123 is employed as the assistant agent to prepare GN/PANI composites by one-step in situ polymerization in acid solution. We investigated the influence of P123 addition amount on microstructure and capacitive performance of GN/PANI composites. Analyses by SEM, TEM, low temperature N2 adsorption, together with electrochemical analysis showed that the composites microstructure and capacitance behavior can be controlled by P123 concentrations. Finally, a mechanism of interaction between P123 and composites was preliminarily proposed.

Experimental

Preparation of GN/PANI composites and working electrodes

Aniline (ANI), ammonium persulfate (APS), hydrochloric acid (HCl) and ethanol (analytically pure) and conductive graphite (chemically reagent) were all obtained from Sinopharm Chemical Reagent Co., Ltd. Pluronic P123 (EO20PO70EO20, average relative molecular weight about 5800) was obtained from Sigma-Aldrich Ltd. GN (Conduction Type-GNSE1231) was supplied by The Sixth Element (Changzhou) Ltd. All raw materials were used directly without further purification. All solutions were prepared using deionized water.

Specific preparation procedures of GN/PANI composites are depicted as follows: Firstly, a certain amount of P123 was added to 20 ml anhydrous ethanol, mechanical stirred until completely dissolved. Secondly, 0.1118 g GN was dispersed into the above solution, stirring for 30 min to get the GN suspension. Thirdly, 20 ml 2 mol dm−3 HCl solution dissolved with 0.012 mol ANI was added to the suspension under constantly stirring. Lastly, APS solution (0.004 mol APS dissolved in 40 ml 1 mol dm−3 HCl) was added to the above solution to initiate polymerization. The synthesis lasted for 12 h, and GN/PANI composite powder was obtained after filtering, washing, drying, and grinding. Molar ratio of P123 to ANI (n P123/n ANI) was controlled to be 0, 0.0036, 0.0072, 0.0108, 0.0144, and 0.0180. The obtained composites were correspondingly denoted as GP-P0, GP-P1, GP-P2, GP-P3, GP-P4, and GP-P5. Pure PANI was synthesized in the absence of P123 and GN according to above experimental process.

The working electrodes were fabricated by the mixture of 85 wt% active materials, 10 wt% conductive graphite, and 5 wt% polytetrafluoroethylene (PTFE). PTFE was firstly dispersed in the ethanol under ultrasonic, followed by active materials and graphite. Then, the mixture was grinded adequately to form a paste. The paste was evenly smeared into a foam nickel mesh (1 × 1 cm2) and then dried in a vacuum oven at 60 °C for 8 h. Lastly, the electrode was pressed into a thin sheet under the pressure of 10 MPa to ensure a good electrical contact between active materials and foam nickel mesh.

Material characterization

Structure of GN/PANI composites was analyzed by X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and low temperature N2 adsorption. Thereinto, XRD, recorded from 10° to 80° (2θ angle), was carried out with Cu Kα radiation (λ = 0.15406 nm) on a Bruker D8 super speed X-ray diffractometer. FTIR spectra were recorded by a Bruker Tensor 27 spectrometer with KBr pellets from 4000 to 400 cm−1. XPS was performed on ESCALAB250Xi Photoelectron Spectrometer (ThermoFisher Scientific Ltd.) with Al Kα radiation as the excitation source. Low temperature N2 adsorption was conducted by automatic analyzers TriStar 3000 to characterize samples pore texture properties. Morphology of composites was observed by S-4800 II field emission scanning electron microscope (FESEM, 20 kV) and Philips TECNAI 12 transmission electron microscope (TEM) (Hitachi Ltd., Japan).

Electrochemical measurements

All electrochemical measurements were performed in 1 mol dm−3 KNO3 aqueous electrolyte by a three-electrode system (samples as the working electrode, activated carbon as the counter electrode, and saturated calomel electrode (SCE) as the reference electrode) at room temperature. Cyclic Voltammetry (CV) and Galvanostatic Charge–Discharge (GCD) tests were conducted by CHI660E electrochemical workstation (Chenghua, Shanghai China) within a potential range of −0.4 to 0.4 V. Scan rates and current densities were set from 5 to 100 mV s−1 and 500 to 5000 mA g−1, respectively. Electrochemical impedance spectroscopy (EIS) tests, with alternating current signal amplitude set at 5 mV, were carried out on AutoLab-PGSTAT30 (Netherlands) within the frequency range of 10−2 to 105 Hz.

Results and discussion

Structure and morphology characterizations

Figure 1a shows XRD patterns of samples. For PANI, crystalline peaks at 2θ = 15.3°, 25.3° and 20.2°, caused by periodicity perpendicular, parallel to the polymer chain and layers of polymer chains of alternating distance, respectively, can be attributed to (011), (200) and (020) crystal planes of conductive PANI-NFs [25–27]. Diffraction peaks of GN appear at 2θ = 25.9° and 42.6° correspond to the graphite-like structure (002) and (100) crystal planes, respectively [28]. Peaks of all composites are very similar likely the superimposed of PANI and GN, indicating that P123 has no effect on GN/PANI composites crystal structure, and GN and PANI has been combined successfully. Nevertheless, GN diffraction peak at 2θ = 42.6° is weaker, suggesting GN surface has been widely coated by PANI nanoparticles. FTIR spectra of samples are shown in Fig. 1b. Spectrum of GN is basically a straight line without characteristic bands, demonstrating that GN is a perfectly reduced state [29]. For PANI, several bands can be observed in its spectrum, suggesting the generation of doped PANI in its emeraldine salt (ES) form [17]. Characteristic bands of GP-P0 and GP-P3 are all similar to that of PANI, indicating the negligible effect of P123 on composites structure. Moreover, no characteristic band of P123 is observed in spectra of composites, implying that most surface P123 has been removed by washing process.

Fig. 1
figure 1

a XRD patterns and b FTIR spectra of samples

Full size image

The wide scan XPS spectrum (Fig. 2) indicates that GP-P3 is mainly composed of C, N, and O elements. In general, content of element can be calculated by peak area. Specifically, C/N ratio of GP-P3 surface is 7.3:1 while theoretical value of pure PANI is 6:1, implying that most of GN surface has been covered with PANI. Extra C may come from GN not covered by PANI. Exposed GN surface may be ascribed to the interaction between PANI particles. In addition, appearance of O may be attributed to H2O, residual P123 in composites or residual carboxylic groups of GN. C1 s spectrum can be deconvoluted into four Gaussian peaks, corresponding to C–C/C=C (284.8 eV), C-N (285.5 eV), C–O–C (286.4 eV), and C=O/O=O–O (290.0 eV), respectively [30, 31]. The presence of ether and carbonyl groups further confirms the existence of P123 and oxygen-containing groups of GN surface. Certainly, quantity of these impurities is little. On the other hand, N1 s spectrum can be decomposed into three different peaks, with binding energies at 400.0 (C=N–C), 400.5 (C-NH-C), and 404.7 eV (C-N+-C), respectively [31]. Thereinto, the existence of –N+– suggests the formation of doped PANI. XPS analysis is basically consistent with XRD and FTIR.

Fig. 2
figure 2

XPS spectrum of GP-P3 and core levels of C 1 s, N 1 s

Full size image

In Fig. 3a, brightness of GN sheets is uneven despite its good transparency, implying that parts of them are multi-layer and congregated together. Its agglomeration is very obvious in Fig. 4a. Pure PANI obtained in this work are all stumpy PANI-NFs, interweaved into a low voidage complex network (see Figs. 3b, 4b). If such PANI and GN with serious agglomeration are used as electrode material alone, the intercalation/de-intercalation of electrolyte ions can only be performed on the limited network surface. As a result, its utilization ratio and capacitance performance would be poor.

Fig. 3
figure 3

TEM images of GN (a), PANI (b), GP-P0 (c), GP-P1 (d), GP-P3 (e), GP-P5 (f)

Full size image
Fig. 4
figure 4

SEM images of GN (a), PANI (b), GP-P0 (c), GP-P1 (d), GP-P3 (e), GP-P5 (f)

Full size image

In Fig. 3, brightness of GN sheets is weaker than that of GN/PANI composites, indicating PANI has widely loaded on GN surface. According to previous studies [32], oxidative polymerization of PANI contains two competitive processes: homogeneous nucleation in bulk solution and heterogeneous nucleation on GN sheet. The latter can be performed preferentially with the presence of GN. The lower right of Fig. 4e is the magnification of part 1, herein, a large amount of oligomer spherical PANI (OS-PANI) can be seen on GN surface. OS-PANI coated GN structure (OS-PANI/GN) is more obvious in Fig. S1. The lower left of Fig. 4e is the magnification of part 2, showing a relatively homogenously distribution of PANI-NFs on the surface of OS-PANI/GN. OS-PANI and PANI-NFs might be produced from heterogeneous nucleation and homogeneous nucleation, respectively. Structure of GN/PANI composites are basically the same as the sandwich structure of GP-P3, and the main difference is the distribution of PANI-NFs. Inset in Fig. 4c is the SEM image of GP-P0 at low magnification. PANI-NFs and GN are severely agglomerated. The inter-mixing degree of two individual components is low. The reason may be GN layers fold together without P123 and hence its surface utilization is low. In other words, the effective contact area between ANI monomers and GN surface is small, and most PANIs are generated from homogeneous nucleation. With the increase of P123 addition amount, shown in Figs. 3c–e and 4c–e, dispersity of PANI-NFs on OS-PANI/GN surface was improved. Inset of Fig. 3e is the HRTEM image. Obviously, distribution of PANI-NFs in GP-P3 is the most uniform and the inter-mixing degree of two single components is the highest, ascribing to the steric hindrance effect of macromolecule P123. However, excessive P123 will result in relatively low degree of compound mixing. Indeed, when the molar ratio of P123 to ANI is 0.018 (Figs. 3f and 4f), large amounts of PANI-NFs grow outward along the edge of OS-PANI/GN composites. It might be attributed to that P123 self-assembling into spherical or rod-like micelles with enormous size, absorbing on the surface of OS-PANI/GN. Electrostatic repulsion and steric hindrance of P123 force these PANI-NFs to generate only along OS-PANI/GN composites edge. P123 in this work acted as soft templates to control the morphology of GN/PANI composites, rather participated the chemical synthesis of composites. It can be further confirmed by the weight ratios of PANI to GN listed in Table S1. Weight ratios are basically the same though the composites were obtained with the assistance of different amounts P123.

Based on the data of D in Table 1, all samples can be considered as mesoporous materials (2-50 nm), conducive to electrochemical energy storage. Structure parameters of GN/PANI composites, especially S BET and S Meso, are lower than those of GN but closer to those of pure PANI, indicating that GN surface has been coated by a layer of dense PANI. It coincides well with SEM and TEM results. Furthermore, by comparing the textural properties of samples prepared with/without the assistance of P123, P123 is favorable in improving GN/PANI composites textural properties in general. With the increase of P123 addition amount, all parameters increase at first and then decrease. GP-P3 having the highest specific surface area, largest pore volume, and best average pore size, conducive to the intercalation/deintercalation of electrolyte ions and the improvement of electrodes capacitance performance. Therefore, 0.0108 can be ensured as the optimal molar ration of P123 to ANI in the preparation of GN/PANI composites.

Table 1 Textural properties of Samples
Full size table

Capacitance behavior

In Fig. 5a, CV curves of GN are approximately rectangular, implying an ideal electric double-layer behavior [19]. A remarkable growth of current density response is observed with the increase of scan rates, demonstrating the excellent rate capability of GN [33]. For PANI (Fig. 5b), redox peaks can be attributed to the redox process of PANI between its leucoemeraldine form and polaronic emeraldine forms [25, 34]. With the increase of scan rates, the increase range of enclosed curves areas decreases, suggesting poor rate performance of PANI. Several more obvious redox peaks appear in Fig. 5c, indicating pseudocapacitive performance of PANI among GP-P3 is better than pure PANI with low voidage network structure. Hereinto, morphology of PANI is modified by P123. With the increase of scan rates, the current response of GP-P3 increases clearly, implying good rate properties. In addition, an anodic shift of oxidation peaks and a cathodic shift of reduction peaks on GP-P3 CV curves are due to the resistance of electrode [35]. In general, the larger the enclosed CV curves areas, the greater the C sp. From Fig. 5d, enclosed areas of GN/PANI composites are all greater than that of pure PANI, increasing at first and then decreasing with the increase of P123. GP-P3 obtains the largest enclosed areas, suggesting the highest C sp. It can be explained that surface utilization of GN is the highest and morphology of PANI loaded on GN surface is the most regular with the optimal P123 (n P123/n ANI = 0.0108). GP-P3 is most porous and it has the largest specific surface area. Consequently, static capacitance of GN and pseudocapacitance of PANI can be fully exploited to improve C sp and rate capability of GN/PANI composites.

Fig. 5
figure 5

CV curves of GN (a), PANI (b), GP-P3 (c) electrodes at different scan rates; of PANI and GN/PANI composites at 5 mv s−1 in 1 mol dm−3 KNO3 (d)

Full size image

In Fig. 6a, the deviation from linearity for curves of GN/PANI composites is due to the existence of pseudocapacitance [36], implying that capacitances of PANI and composites are mainly derived from pseudocapacitive behavior. The deviation degree of GP-P1, 2, 3, 4, and 5 are basically the same and are all lower than that of pure PANI and GP-P0 (Table S2), indicating the double-layer capacitance performance of P123-assisted composites is more remarkable, or rather, specific surface area of these samples is larger. GCD curves of samples produced with P123 are all more symmetric than those of pure PANI and GP-P0, indicating a better reversibility of charge–discharge reaction and a lower resistance of such samples [37]. Symmetry of GP-P3 curve is the best. By comparing discharge curves in Fig. 6b, P123 undoubtedly prolongs the discharge duration of electrodes. Moreover, with the increase of P123, the discharge time increases initially and then decreases. GP-P3 obtains the largest discharge time, or the largest C sp, which is consistent with the conclusion from CV curves. Within the potential window of −0.4 to −0.2 V, the discharge magnitude of composites decreased and the resistance increased, ascribing to the increasing conductivity from the continuous undoping process of PANI [36].

Fig. 6
figure 6

GCD curves (a) and discharge curves (b) of electrodes at a current density of 500 mA g−1

Full size image

C sp of PANI and a series of GN/PANI composites at different current densities are calculated according to the following equation:

$$ C_{\text{sp}} = I \times \Delta t/(m \times \Delta V), $$
(1)

where I is the charge–discharge current (A), Δt is the discharge time (s), m is the active material mass of working electrode (g), ΔV is the potential difference during discharge process (V), and C sp is the electrode specific capacitance (F g−1). Specific data are shown in Table 2. Due to the agglomeration of GN, a low C sp value 10.2 F g−1 is obtained at a current density of 500 mA g−1, while PANI and GP-P0 are, respectively, 102.3 and 118.5 F g−1. C sp of GP-P0 is larger than the sum of those of GN and PANI. This could be due to the synergistic effect of GN and PANI. That is, on the one hand, GN can provide PANI with an ideal conductive network, enhancing the degree of Faradaic reaction between the PANI and the electrolyte; on the other hand, PANI coated on GN surface can prevent the re-aggregation of GN sheets. Synergistic effect of GN and PANI is undoubtedly conducive to improve single component utilization ratio and hence to increase capacitance of composites. From the comparison of GP-P0- and P123-assisted composites, the addition of P123 significantly increased composites C sp values, confirming the positive effect of P123 on improving composites capacitance performance. With the continued increase of P123 addition amount, C sp value of composite materials first increases, reaching a maximal value of 215.8 F g−1 for GP-P3, before decreases. All C sp values in Table 2 decrease with the increase of the current density. This is because electrolyte ions do not have enough time to diffuse into deep pores of active materials at high current density, thus the insertion/de-insertion is limited on the surface or in shallow pores of electrode materials, resulting in the decrease of the utilization efficiency of composites [38]. As for rate capability, PANI is poor due to its pseudocapacitive behavior. GP-P0 with GN is slightly superior to PANI. For GP-P3, a high C sp 157.2 F g−1 is maintained at a current density of 5000 mA g−1, ascribing to the abundant pores suitable for electrolyte ions fast-moving (as shown in Table 1).

Table 2 Specific capacitances of PANI and a series of GN/PANI composites at different current densities in 1 mol dm−3 KNO3
Full size table

Figure 7a shows the SEM image of GP-P3 electrode before 1000 charge–discharge cycles. Because of the pressure of 10 MPa, the electrode surface is very flat. Materials on the surface should be PANI. After 1000 cycles, many cracks appeared at the surface (Fig. 7b). Two reasons could be account for this phenomenon: (1) the structure of the electrode is too dense, the continuous movement of electrolyte ions is easy to breakdown its surface; (2) the continuous swelling and shrinkage of PANI during the charge/discharge process resulted in a fragmentation of macromolecule chain. Appearance of cracks may lead to a decrease of capacitance. In Fig. 7c, a C sp value of 61.4 F g−1 is retained after 1000 charge–discharge cycles of PANI, with only 59.9 % capacity retained. Comparatively, GP-P0 exhibits a better cycling life with the capacity retention of 89.2 %. It can be ascribed to the synergistic effect of GN for inhibiting the structural failure of PANI. At a current density of 500 mA g−1, C sp value of GP-P3 is up to 215.8 F g−1 and its capacitance after 1000 charge–discharge cycles retains 91.8 %, increased by 82 and 2.9 % individually in comparison with GP-P0, suggesting P123 has a positive effect on improving composites cycle performance. Capacity retention of GP-P3 is higher than those reported values [16–18].

Fig. 7
figure 7

SEM images of GP-P3 pasted on Ni foam before (a) and after (b) 1000 cycles; c 1000 charge–discharge cycles stability curves of electrodes at a current density of 500 mA g−1; d Nyquist plots of electrodes. Inset is the amplified image of Nyquist plots in high frequency region and circuit diagram equivalent

Full size image

In Fig. 7d, the x intercept of Nyquist plots corresponds to equivalent series resistance (R s), including intrinsic resistance of active material, electrolyte resistance, and contact resistance between electrolyte and electrode [33, 38]. According to Table 3, R s of GN/PANI composites are all slightly lower than that of PANI (0.38 Ω), indicating that the synergistic effect between GN and PANI is helpful to reduce R s. Thereinto, R s of GP-P3 is the smallest. In high frequency region of Nyquist plots, diameter of the semicircle represents ionic charge-transfer resistance (R ct) of the layer on the electrode [39]. In general, the larger the diameter is, the higher the R ct will be and the poorer the conductivity is. From the inset in Fig. 7d, all semicircle diameters of composites are smaller than that of pure PANI, suggesting both GN and P123 are all conductive to decrease R ct and improve composites conductivity. Meanwhile, GP-P3 obtains a smallest R ct of 1.46 Ω (Table 3). In intermediate frequency area, electrodes reaction process is controlled by diffusion procedure [21]. Nyquist plots in this part evolve into an oblique line with the slope equal to 1. The more the slopes are closer to 1, the easier for the conduction of the diffusion process. Resistance in this part is named as Warburg impedance (W). Mesoporous structure and large pore size of electrode materials are favorable to decrease W [40]. According to previous analysis, GP-P3 possesses the most abundant mesoporous and the largest pore size compared to other composites. Therefore, W of GP-P3 should be the minimum, well consistent with Nyquist plots. However, in Table 3, W of GN/PANI composites are all larger than that of pure PANI, which could be related to the variety of PANI forms in composites [41]. At the low frequency part, there is a nearly vertical line. The more vertical the line is, the more ideal the capacitor for the supercapacitor will be [39]. Obviously, the deviation rate between pure PANI curve and vertical line is the largest, suggesting the assistance of P123 is helpful to optimize the capacitance behavior of GN/PANI composites. Curve of GP-P3 is the most vertical, indicating 0.0108 is the optimal molar ratio of P123 to ANI. In general, the higher the constant phase element (CPE) values, the quicker is the formation of electric double layer [42]. Thereinto, GP-P3 owns the maximum CPE among composites as is shown in Table 3.

Table 3 Equivalent circuit diagram parameters of samples
Full size table

Function mechanism of P123

Figure 8 illustrates possible procedures of the generation of GN/PANI composites with P123. Accordingly, amphipathic macromolecule P123 is firstly adsorbed by electrostatic interaction on the surface of GN sheet to enhance its surface wettability and hence to improve its dispersity and uniformity in aqueous solution (Fig. 8a) [43]. Secondly, due to the steric hindrance of macromolecule P123, some surfaces of GN are unoccupied. Owing to the electrostatic force, positively charged ANI monomers are absorbed on negatively charged GN surface (Fig. 8b) [17]. Nucleation sites distributed heterogeneously on GN surface nucleate and generate OS-PANI at the beginning of polymerization [32]. Meanwhile, P123 is competitively displaced by OS-PANI particle, because the interactions of π-π conjugation, hydrogen bonding, and electrostatic force between GN and PANI [17] are larger than intermolecular forces between GN and P123. With polymerization reaction carried through, the initial PANI coated GN structure can be produced (Fig. 8c). Finally, homogeneous nucleation plays the leading role in polymerization [32], nucleation points are redistributed again on PANI surface. ANI monomers are polymerized to PANI-NFs along the surface of PANI spherical particle. Due to the steric hindrance of P123, distribution of PANI-NFs is uniform (Fig. 8d). When the addition amount of P123 is optimal (n P123/n ANI = 0.0108), the structure of GN/PANI composites will be almost the same to Fig. 8d. Consequently, the utilization ratio of GN and PANI, as well as capacitive performance of composite materials can be improved.

Fig. 8
figure 8

Illustration of possible procedures of the generation of GN/PANI composites with the assistance of P123

Full size image

Conclusions

GN/PANI composites were fabricated by one-step in situ polymerization in acid solution with the assistance of P123. According to FTIR and XRD tests, P123 on materials surface has been mostly removed by washing process and has little effect on GN/PANI composites crystal structure. XPS spectra indicate that GN surface has been covered with PANI. SEM and TEM demonstrate that PANI-NFs are loaded on OS-PANI/GN surface, indicating P123 indeed has impacted the morphology of GN/PANI composites. In particular, dispersity of PANI-NFs as well as the degree of the mixing between GN and PANI is the most optimal when n P123/n ANI = 0.0108. N2 physical adsorption data prove that GP-P3 possesses the highest specific surface area 37.6 m2 g−1, pore volume 0.19 cm3 g−1 and average pore size 20.4 nm among all composites. The same mass ratio of single components in GN/PANI composites and above characterizations suggested that P123 in this work major functioned as soft templates to control the composites morphology. Specifically, P123 could promote PANI to well-grown and well-dispersed on GN surface, enhancing the effective utilization of two single components. In addition, P123 ensured PANI with abundant pores, enabling a fast Faradaic reaction and providing a short ion diffusion path during the charge–discharge process. Combined with electrochemical tests, it can be confirmed that P123 is helpful to improve capacitive performance of GN/PANI composites. Comprehensive properties of GP-P3 are the best. In detail, specific capacitance of GP-P3 is up to 215.8 F g−1 at the current density of 500 mA g−1, increased by 82 % in comparison with GP-P0. Its capacity retention after 1000 charge–discharge cycles is 91.8 and 2.9 % higher than that of GP-P0. Additionally, equivalent series resistance (R s), charge-transfer resistance (R ct), and Warburg impedance of GP-P3 are the smallest among all composites. Applying P123 to the fabrication of GN/PANI composite electrodes is an effective way.