Influence of current collector electrode on the capacitive performance of electrodeposited PANI: insight gained from frequency and time domain analysis

Abstract

This article focuses on the choice and effect of current collector electrode on the performance-indicating parameters of supercapacitors. The electrochemical properties of polyaniline (PANI) deposited on platinized fluorine doped glass (Pt) and fluorine doped glass (F : SnO₂) current collectors are evaluated in depth by galvanostatic charge/discharge (GCD), cyclic voltammetry (CV) and impedance spectroscopy (IS). The GCD results show that the Pt-based electrode has better supercapacitive behavior than the F : SnO₂ electrode. Scanning electron microscopy (SEM) results reveal that a porous granular array of PANI is induced on the Pt electrode, whereas a compact granular morphology is induced on the F : SnO₂ electrode. Results from the CV are used to describe the dynamic variation of capacitance as a function of the applied bias. CV results reveal a higher value of capacitance for the Pt-based supercapacitor, which confirms the higher pore filling by electrolyte ions. Impedance measurements lead to determination of parameters such as equivalent series resistance, rate capability of electrodes and AC conductivity of supercapacitors. The results obtained here could aid understanding of how simple time and frequency domain experiments can be strategically used for quantitative characterization of supercapacitors.

---

1. Introduction

Increasing demand for energy and concern about global warming has increased research interest in the field of energy generation from renewable energy sources.^1,2 However, most energy storage devices cannot meet the ever-increasing and urgent need for storing the excess energy generated by renewable sources. Supercapacitors have emerged as a new class of fast energy storage device, which can fill the gap between batteries and capacitors.^3,4 Supercapacitors are ideal for applications in portable electronic equipment, hybrid electric vehicles and standby power systems, where short-term peak power with high power density is required.^5,6 Supercapacitors can be classified into several types depending on specific capacity and power density: (1) electrochemical double layer capacitors (EDLCs), which work via the mechanism of charge separation at the electrode electrolyte interface with high power density, (2) pseudo capacitors, which work via the mechanism of reversible surface Faradaic redox reaction, with high capacitance and energy density and (3) hybrid supercapacitors (HSCs) and asymmetric supercapacitors (ASCs), which work via the principle of redox reaction at one electrode and electric double layer adsorption and desorption at the other electrode with high energy and power density.^7,8

A wide range of metal oxides, conducting polymers and carbonaceous materials have been studied and employed for supercapacitor applications.⁹ Conducting polymer, polyaniline (PANI) has attracted considerable attention because of its properties such as environmental stability, low cost, easy doping de-doping, multiple redox states and easy synthesis in various nanostructures.^10–12 Recently there has been a renewed interest in synthesis of PANI and its derivative for potential application in energy storage devices. Mi et al.¹³ synthesized nanofiber structured PANI for supercapacitor electrode application, with a specific capacitance of 428 Fg⁻¹ in H₂SO₄ electrolyte. Lie et al.¹⁴ synthesized PANI on titanium substrate, with a specific capacitance of 837 Fg⁻¹ in H₂SO₄ electrolyte. Chen et al.¹⁵ obtained a remarkable energy density of 84 W h kg⁻¹ for PANI nanotube grown using MnO₂ templates. Synthesis of PANI with carbonaceous material such as graphene, single wall carbon nanotubes and multiwall carbon nanotubes, various oxides of Ni, Co, In, Sn, Fe, Mn and so forth, is also being investigated to develop an economical electrode material with a high capacity for charge storage.¹⁶

The conventional supercapacitor configuration consists of four main sections: electrode, electrolyte, separator and current collector. The heart of the supercapacitor which determines the power delivery rate, efficiency and maximum extraction of stored energy at the electrochemical interface are current collector and morphology of the electrode. The aforementioned parameters are bias, frequency and temperature dependent, which provide critical guidance to determine and optimize the electrode morphology and current collector interface for fabricating and improving the performance of a supercapacitor. Accurate determination of the role of these entities on the performance of a supercapacitor often requires multiple experimental analysis techniques. Most of the characterization techniques used in this area simply aim to measure the overall storage and charge discharge mechanism of a supercapacitor.^17,18 Parameters such as transient response, dynamic behavior, net terminal impedance and influence of physical parameters on these responses are scarcely discussed in the available literature.

Keeping in mind these shortfalls, we fabricated supercapacitor electrodes with a configuration of platinized fluorine doped glass substrate/PANI (Pt/PANI) and fluorine doped glass substrate/PANI (F : SnO₂/PANI), while retaining the other mechanisms for both electrodes. Under these configurations, we undertook a detailed study (1) to determine the limits of traditional techniques and (2) on combined use of direct current (DC) and alternating current (AC) techniques to explore the voltage and frequency dependent dynamic trend of current collector–electrode interface, where the parallel contributions of the remaining components are described in terms of established mechanisms.

2. Experimental

Preparation of the electrode

Fluorine doped glass (F : SnO₂) and platinized fluorine doped glass (Pt) substrates having 10 × 10 mm² area were purchased from Sigma Aldrich and Dyesol, respectively. Use of platinized F : SnO₂ substrate (having a particle size distribution varying from ∼25–32 nm) was suggested because of its high conductivity, low series resistance and high surface to volume ratio. Aniline purchased from Sigma Aldrich was purified by distillation and stored in a nitrogen glove box prior to use. A monomer solution of aniline in MilliQ water was prepared using analytical reagent grade hydrochloric acid (HCl). Platinum ring and Ag/AgCl (saturated KCl) from CH instruments served as counter and reference electrodes, respectively. A three electrode cell was used for electrodeposition of PANI. The cell and counter electrode were cleaned with freshly prepared 1 : 1 (v/v) H₂O₂/H₂SO₄ solution followed by ultrasonication in MilliQ water prior to each experiment. All the electrochemical experiments and measurements were performed using potentiostat CH Instruments 660D equipped with general purpose electrochemical software. A geometric surface area of 0.7 cm² was exposed to the electrolyte for both electrodes. A monomer solution prepared from 2% aniline in aqueous 1 M HCl (v/v) was used for electropolymerization, performed by cycling both the electrodes between −0.2 V and 0.76 V at a sweep rate of 0.015 V s⁻¹. Details of the polymerization process of PANI on a highly oriented pyrolytic graphite substrate are discussed in a previous article.¹⁹ After each experiment, the remaining electrolyte was soaked with filter paper and kept for drying. The microstructures of PANI on both the electrodes were produced by terminating the applied potential at initial value. To test the performance of PANI based solid state supercapacitor, approximately 10 wt% polyvinyl alcohol (PVA) solution in 20 ml deionized water was heated at 75 °C under constant stirring until a clear solution of PVA was obtained. A concentrated solution of H₂SO₄ was added and stirred gently for 45 min to obtain 1 : 1 PVA–H₂SO₄ solution.

SEM characterization

Electrodeposited PANI on Pt and F : SnO₂ substrates were placed onto circular adhesive carbon tape to be affixed to aluminum stubs. The morphology of PANI was investigated using Leo-s 440i scanning electron microscopy with an operating voltage range of 1–10 kV under ultra-high vacuum conditions.

Electrochemical measurements of electrode

All electrochemical measurements were performed in a two electrode cell with a fine layer of PVA–H₂SO₄ serving as an effective separator. Once the open circuit potential of the system was stabilized, galvanostatic charge/discharge, cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were performed in 1 M H₂SO₄ aqueous electrolyte. The constant current charge/discharge cycling was performed at room temperature within the potential window of 0–0.65 V with various current densities from 20 μA to 50 μA. The CV data were recorded at different scan rates within a voltage window of 0–0.6 V in 1 M H₂SO₄ electrolyte.

True analogue ramp mode rather than staircase mode, was used in all the CV measurements to avoid the problematic transient effect on capacitive current measurements.^20,21 For EIS measurement in H₂SO₄ electrolyte, AC signals of 10 mV amplitude at frequencies between 0.1 Hz and 100 kHz were used. Experimental Nyquist plots were collected at different DC voltages and validated using a frequency comparison method where simulations of the impedance behavior were performed on the basis of an equivalent circuit model using a program developed in MATLAB. Unless mentioned otherwise, all the potentials refer to Ag/AgCl/Cl⁻ (saturated) electrode.

3. Results and discussion

Morphological analysis

The morphologies of PANI films on platinized F : SnO₂ and F : SnO₂ were measured by SEM and are shown in parts A and B of Fig. 1, respectively.

Fig. 1 SEM images of (A) Pt/PANI with 1 μm scale bar and (B) F : SnO₂/PANI with 1 μm scale bar.

The obtained SEM morphology reveals that growth of PANI is homogeneous and granular in shape on both electrodes. The average grain sizes of PANI on F : SnO₂ and Pt are ∼300 nm and ∼90 nm, respectively. The approximate distribution and porosity of deposited PANI on both the electrodes were analyzed using ImageJ software, and the details are given in Fig. S1 and S2 of ESI.† The porous nanostructure of PANI on Pt has the advantage of providing a long and uninterrupted path for electron transport, which is favorable for supercapacitor applications.¹⁸ A noticeable distribution of large numbers of grains with almost uniform dimension on F : SnO₂ clearly indicates some chemical as well as electrostatic interaction, which is less prominent in PANI grown over Pt.

From the obtained morphology of PANI over Pt and previously reported results,²⁰ it is expected that PANI film possesses a higher ionic conductivity and capacitance because (1) the larger surface area means a higher number of sites for access of electrolyte; (2) of the formation of porous interconnected networks of PANI granules over the entire surface of the film. To confirm these results and to distinguish the contributions from different processes taking place simultaneously, time and frequency responses for both electrodes were analyzed.

Nucleation and growth of PANI

Fig. 2A and B shows the CV recorded during electropolymerization of PANI with a scan rate of 15 mV s⁻¹ in the potential range of −0.2–0.76 V at the Pt and F : SnO₂ electrodes, respectively. The obtained voltammograms for electropolymerization of PANI are similar to those reported previously by other authors.^22,23 The redox pair, i.e. oxidation peak at ∼0.23 V and the corresponding reduction peak at ∼0.05 V, is attributed to transformation of PANI from the reduced leucoemeraldine state to the partly oxidized emeraldine state.²⁴

Fig. 2 Cyclic voltammograms recorded during electropolymerization of PANI film: (A) on Pt electrode and (B) on F : SnO₂ electrode, First cycle of CV recorded during electropolymerization of PANI film: (C) on Pt electrode and (D) on F : SnO₂ electrode.

A weak or almost negligible anodic hump at ∼0.5 V indicates an insignificant formation of benzoquinone type degradation product. A rise in current during anodic scan at the terminating potential (∼0.76 V) is attributed to formation of the polyemeraldine state of PANI. During the electropolymerization process, the increase in peak current density with sweep segment is a measure of the growth rate of polymer.²⁵ To gain an insight into the different surface morphology (Fig. 1) and growth processes of PANI at the two electrodes, the first cycle of CV and total charge under anodic peak as a function of sweep cycle are shown in Fig. S2 and S3 of ESI,† respectively. The initial cycle of CV at the Pt electrode shows a higher current with the electroformation and reduction of the leucoemeraldine form of PANI than that of the F : SnO₂ electrode. This suggests the electrocatalytic effect of Pt nanoparticles means growth of PANI on Pt is more favorable than on the F : SnO₂ electrode. From the charge density plot at lower sweep cycles (inset of Fig. S3 ESI†), it is observed that charge density is much lower and increases at a very slow rate for the F : SnO₂ electrode. This may be because during the initial polymerization process the distributed nucleation site causes development of a short chain of PANI nanoparticles, whereas after ∼12 sweep cycles the charge density for F : SnO₂ electrode increases at the same rate as that of Pt electrode. This leads to the formation of a compact layer of PANI or enrichment at local sites. However, in the final polymerization process, the net charge densities for Pt and F : SnO₂ electrodes are 51 and 49 mC cm⁻², respectively. A small difference in the value of charge density for both the electrodes occurs because of the higher electroactivity of Pt nanoparticles.

Electrochemical characterization of electrodes: galvanostatic charge discharge

A comparative quasi symmetrical galvanostatic charge discharge plot of F : SnO₂/PANI and Pt/PANI electrodes within the potential range of 0–0.6 V at a constant current of 50 μA cm⁻² is shown in Fig. 3. Deviation in linearity of the charge discharge profiles suggests that both electrodes have pseudo capacitive characteristics. In Fig. 3 the net area under the discharge curve (specific capacitance) for the Pt/PANI electrode is ∼30% higher than that of the F : SnO₂/PANI electrode. The charge discharge curve for the Pt/PANI electrode was analyzed in two regions of potential in an attempt to identify the mechanism for this higher capacitance. The influence of these two regions on the performance indicating parameters and the physical origin in a supercapacitor will be discussed in the following sections.

Fig. 3 Galvanostatic charge discharge curves of Pt/PANI and F : SnO₂/PANI electrodes at a current density of 50 μA cm⁻².

Region I corresponds to the voltage range 0.4–0.6 V where a drop in potential across internal resistance takes place. Fig. 3 shows that in region I (0.4–0.6 V), the internal resistance R_s dominates the discharge curves of both electrodes. The internal resistance is governed mainly by a voltage-dependent physical mechanism originating at different layers or interfaces within the supercapacitor; a front contact made of Pt or F : SnO₂, carrier transporting interlayers, a charge storing PANI layer and the PANI/electrolyte interface significantly contribute to the value of R_s.⁹ The interface formed between the active (PANI) layer and front contact dominates the series resistance because of partial energy level alignment, which affects optimal charge transfer at the interface. A comparative examination of region I (Fig. 3) for both electrodes shows that resistive drop (iR_s) associated with F : SnO₂/PANI (∼0.45 V) is more substantial than with Pt/PANI (∼0.35 V). The Pt/PANI electrode has ∼1.5 times lower iR_s drop because of the higher conductivity of the Pt nanoparticle compared with F : SnO₂ glass, which increases the speed of charge transfer at the electrode, thereby possibly indicating this would be more proficient when used as a supercapacitor electrode.

Region II corresponds to the voltage range 0–0.4 V, where the combined effect of pseudo and electric double layer capacitance dominates the charge discharge profile of electrodes.^26,27 A comparatively longer charge discharge time (Fig. 3) obtained for the Pt/PANI electrode signifies a higher capacitance than the F : SnO₂/PANI electrode. Enhanced electrochemical performance of the Pt/PANI electrode can be attributed to the following: (1) a higher surface area coupled with a large number of pores in the Pt/PANI electrode facilitates deeper penetration of electrolyte ions and fast electron transfer through the electrode matrix, and (2) higher conductivity of platinized F : SnO₂ substrate greatly improves the performance of the Pt/PANI electrode essentially by reducing the iR_s drop.

Fig. 4 demonstrates the typical GCD plots of Pt/PANI and F : SnO₂/PANI electrodes at different current densities. The non-linear charge discharge profiles of both the electrodes deviate from the usual linearity, confirming redox and pseudo capacitive behavior of the electrode. The interfacial capacitance values of the Pt/PANI and F : SnO₂/PANI electrodes were estimated from the slope of discharge curve as , where I is the discharge current, S is the active area of the electrode (cm²), Δt is the discharge time and ΔV is the voltage window. Calculated interfacial capacitance values for both electrodes at different current densities are listed in Table 1.

Fig. 4 Galvanostatic charge discharge curves of (A) Pt/PANI and (B) F : SnO₂/PANI electrodes at different current densities.

Table 1 Electrochemical parameters of Pt/PANI and F : SnO₂/PANI

	Discharge current	F : SnO2/PANI	Pt/PANI
Interfacial capacitance (mF cm^−2)	50 μA	5.64	8.94
	40 μA	5.73	8.96
	20 μA	6.76	10.6
Series resistance (Rs)	—	44.4 Ω	41.9 Ω

---

Maximum values of interfacial capacitance are observed for the Pt/PANI and F : SnO₂/PANI electrodes at 20 μA current density, being 10.6 mF and 6.76 mF, respectively. A decrease in the value of interfacial capacitance with an increase in current density is observed for both electrodes, signifying electrode polarization at higher current density and a fast redox kinetic at the electrochemical interface.⁸ The obtained results suggest that the comparatively porous architecture of Pt/PANI facilitates ion diffusion at the electrochemical interface and fast electron collection at current collector, which effectively maximizes the capacitance of a supercapacitor. However, the GCD technique is not particularly useful for probing the following parameters of a supercapacitor: (1) transient load for switching application, (2) the net terminal impedance and (3) the dynamic trend of interfacial capacitance. Hence, DC (cyclic voltammetry) and AC characterization techniques (impedance spectroscopy) were coupled in the same framework of two bias regions (0–0.4 V) and (0.4–0.6 V) in an attempt to obtain conclusive evidence about the improved performance and aforementioned parameters of the supercapacitor.

Scan rate-dependent cyclic voltammetry

Analysis of leaky (non = Faradaic) and pseudo (Faradaic) capacitance formed at the electrochemical interface from CV at different potential scan rates is straightforward.^28–30 Fig. 5 shows the voltammetric responses of Pt/PANI and F : SnO₂/PANI electrodes in aqueous 1 M H₂SO₄ electrolyte as a function of scan rate within the optimized voltage window of 0–0.6 V vs. Ag/AgCl reference electrode.

Fig. 5 A DC cyclic voltammogram of the interface formed between (A) Pt/PANI (B) F : SnO₂/PANI film at different scan rates.

The cyclic voltammogram for both electrodes follows the description of i = i_r ± ν_dcC^′_i, where i represents the net current, i_r, ν_dc and C^′_i represents the Faradaic current, scan rate and scaled value of interfacial capacitance, respectively.^21,31 The electrochemical response current of the CV curves for both electrodes shows that the positive sweeps are asymmetric to their corresponding counterpart negative sweeps with reference to the zero current line.⁹

A prominent characteristic peak occurring during the positive and negative sweeps of CV for both electrodes is usually seen as evidence of pseudo capacitive behavior, and is distinct from the rectangular shape of an ideal capacitor with no resistance.³² A relative comparison of Fig. 5A and B shows that a higher Faradaic feature (net area) is observed for the Pt/PANI electrode, which is apparent from the porous nature of PANI over Pt nanoparticles (leading to higher surface area) compared with the compact nature of PANI over F : SnO₂ glass. A more porous and larger surface area offered by Pt/PANI film means there can be easy insertion and desertion of SO₄²⁻ ions into the polymer matrix leading to higher electroactivity. Whereas the weak redox activity of F : SnO₂/PANI is caused by difficulty in exchange and insertion of ions into the compact polymer matrix. Lower current values in the voltage ranges 0.35–0.6 V and 0.25–0.6 V during the negative sweeps for Pt/PANI and F : SnO₂/PANI, respectively, originate from the non-Faradaic feature of electrodes. The current measured within this potential window is governed mainly by charge/discharge of the double layer and is not affected by Faradaic oxidation or reduction of electrolyte ions. The non-Faradaic feature of both electrodes in this voltage range causes an iR_s drop in the GCD characteristics.

Furthermore, from Fig. 5 it is observed that as scan rate increases, the current under the curve rises, with a slight change in the CV shape. The CVs of both electrodes maintain their shapes within a selected range of potentials, even at a high scan rate of 0.2 V s⁻¹, indicating rapid current response of PANI films on reversal of voltage. Also, a slight shift in the redox peak for both electrodes with a change in scan rate confirms the active response of Faradaic side reactions.^20,33 For the Pt/PANI electrode, sharp response of current at each scan rate signifies good electroactivity and capacitive nature of the electrode. The increase in electroactivity at the film electrolyte interface is attributed mainly to easy mobility and insertion–desertion of ions through the polymer matrix and to the higher number of active sites offered by the working electrode. However, CV of previously reported PANI and PANI composite electrodes of similar configurations are different. There are likely to be several reasons for this difference possibly related to the different PANI morphology and structure, different preparation methods, chemical properties and the substrate.^25,34–36

To examine the linear equation i = i_r ± ν_dcC^′_i, a scaled version of obtained data from Fig. 5 is shown in Fig. 6. As the scan rate is increased from 0.005 to 0.1 V s⁻¹ starting from the outermost to the sequentially underlying plots, the voltammogram gradually reduces on both sides of the zero current axis for both electrodes.^31,37 This observation is attributed to the presence of inner active sites which cannot precede the redox transitions completely, probably because of a diffusion effect of protons within the electrode. Further reduction in the scan rate will drastically expand the outermost plot by enhancing this effect. However, in the present case, the effect of expansion of i/ν_dc with a reduction in scan rate for both electrodes (Fig. 6) cannot account for the pseudo capacitance. This is because evaluation of C_i from CV should be independent of scan rate. Under these circumstances, the interfacial capacitance is calculated using a linear relation of . The polarization resistance R_ct from CV and the electrolyte resistance R_s measured by EIS are used to evaluate the value of C_i. The accuracy of the DC measured capacitance depends on the relative values of R_ct and R_s. Fig. 7 shows the variation of C_i as a function of applied bias for Pt/PANI and F : SnO₂/PANI electrodes. The voltage-dependent feature C_i detailed for both electrodes follows the same trend, with a slight deviation in magnitude and bias offset.

Fig. 6 Scan speed normalized electrode current for (A) Pt/PANI (B) F : SnO₂/PANI electrodes recorded at different scan rates, with CV.

Fig. 7 DC voltage-dependent interfacial capacitance, C_i, obtained for both electrodes.

The observed values of C_i in the bias range (0–0.35 V) for both electrodes are in good agreement with those obtained from GCD measurement, and the reason for a lower magnitude of C_i for the F : SnO₂/PANI electrode is as discussed in the explanation of Fig. 5. Peaks in the bias range (0.35–0.55 V) for Pt/PANI and (0.35–0.47 V) for F : SnO₂/PANI pertain to conversion of fully oxidized pernigraniline to emeraldine, with involvement of protons in the redox reaction. Thus, the peak in the value of C_i can be attributed to the following: (1) the number of radical cations in the polymer chain is maximized, (2) an increase in the PANI electrolyte real surface, resulting in an enhanced number of available active PANI centers and (3) an increase in the population of quinoidal groups inside PANI. The involvement of protons in the redox reaction can be shown as:³⁸

The reduction in the value of C_i at higher bias is caused by decreasing conductivity as the polymer is deprotonated or the oxidation state changes towards either a fully oxidized or a fully reduced state.

The results obtained from DC voltammetry can be summarized as follows: (1) a larger value of capacitance is obtained for Pt/PANI compared with F : SnO₂/PANI, indicating excessive pore filling and charge transfer taking place at the PANI matrix; (2) a sharp drop in the electrode current at higher voltages on discharge results from the diffusion-limited mobility of electrolyte ions in the electrode pores;³⁹ (3) the capacitance of F : SnO₂/PANI is pseudo capacitive, mainly arising from the Faradaic reactions of PANI while the capacitance of Pt/PANI is overwhelmed by the electric double layer capacitance of Pt, along with the pseudo capacitance of PANI. In general, lower internal resistance is required for energy storage devices so that less energy is wasted by producing heat or voltage drop during discharge time.

Electrochemical impedance spectroscopy

Electrochemical impedance spectroscopy (EIS) is one of the principal methods for examining the complex phenomena of electron interception and diffusion at the electrochemical interface.^40–42 An AC response model for elucidating the performance of a supercapacitor in the frequency domain is shown in Fig. 8, where the DC voltage 'V' is combined with frequency-dependent sinusoidal perturbation voltage 'Ṽ'. A Constant phase element (CPE) rather than capacitance is used to account for the frequency dispersion effects of the AC signal. The CPE is a non-intuitive circuit element used to quantify the real world capacitor response, and occurs because of the non-uniform distribution of reaction sites on the electrode surface.^21,31 The complex impedance of the CPE has the form of where, j = (−1)^1/2, Y₀ and n are frequency-dependent CPE parameters with n in the range of 0–1. For n = 1, the CPE takes the form of capacitance. A Warburg element , to account for the diffusion of ions across the polymer film, is introduced in the mid-frequency branch of the equivalent circuit in series with charge transfer resistance R_ct, where w is the Warburg coefficient. A parallel combination of interfacial resistance, R_i and C_i, is used to account for the Faradaic processes responsible for pseudo capacitive behavior. The various parameters involved in the equivalent circuit can be determined by fitting the net impedance with the obtained experimental EIS data.

Fig. 8 Electrical equivalent circuit employed to fit the impedance spectra obtained at different applied bias for both electrodes.

The EIS data are analyzed using complex Nyquist spectra (also called the Cole–Cole plot), which represent the frequency responses of imaginary and real components of impedance.⁴³ The maximum and minimum frequency of the impedance spectrum are represented by the left-most and right-most points on each plot, respectively. In general, a low frequency perturbation signal requires a relatively longer time for EIS measurement, during which the residual current may modify the electrode surface morphology. Thus, a frequency spectrum of 100 kHz–0.1 Hz was selected to maintain the validity of EIS measurements in the present study. The parameters selected for analyzing EIS data were optimized through multiple trials to ensure accurate and reproducible results, with a mean error of modulus less than 5% indicating that these fitting values are highly accurate. The best values of all parameters for the corresponding equivalent circuits from fitting the EIS data are listed in Table 2.

Table 2 The best fitting values of the equivalent circuit elements in ESI Fig. 4 for the impedance data shown in the main article

	Applied potential (V)	Rs (Ω cm^2)	Rct (Ω cm^2)	Y0dl (μF cm^−2)^n1	ndl	W (Ω s^−1/2)	Ri (Ω cm^2)	Y0i (mF cm^−2)^n2	ni	Error (%)
Pt/PANI	0	41.1	12.5	48	0.96	208	9	58	0.98	0.7
	0.1	41.1	10.5	44	0.96	185	2000	19	0.99	1.4
	0.2	41.2	10.1	32	0.96	172	4500	70	0.995	0.3
	0.3	41.2	9.4	20	0.96	172	4580	70	0.995	0.6
	0.4	41.9	30	84	0.82	250	—	—	—	1.2
	0.5	41.8	28 000	17	0.83	262	—	—	—	1.9
	0.6	41.8	38 000	15	0.86	296	—	—	—	2.2
F : SnO2/PANI	0	43.7	24.8	26	1	434	800	20	0.98	0.5
	0.1	43.7	24.1	25	0.96	454	2100	18	0.982	0.9
	0.2	43.7	22	30	0.96	400	4615	18	0.99	1.5
	0.3	43.9	20.6	25	0.96	666	4750	6	0.986	1.3
	0.4	44.3	4900	15	0.82	689	—	—	—	1.5
	0.5	44.3	78 000	15	0.81	712	—	—	—	2.8
	0.6	44.3	123 000	15	0.89	745	—	—	—	1.6

---

Fig. 9 shows that the fitting curve (solid line) and experimental data (symbols) match well with each other, implying that the equivalent circuit model reflects the electrochemical process occurring at the electrode. The complex Nyquist spectra obtained for both electrodes feature semicircles in the higher frequency region related to the double layer charging/discharging and R_ct. The high frequency semicircle is followed by a 45° segment typified by a Warburg element in the mid-frequency region. At very low frequency, a straight line parallel to an imaginary axis can be accounted for by employing R_i and C_i in parallel.^41,44 The point of intersection of real impedance (Z′) in the high frequency region represents the equivalent series resistance of the electrode/electrolyte interface.

Fig. 9 Nyquist plots for (A) Pt/PANI (B) F : SnO₂/PANI electrodes in region II. Inset shows the high frequency details of the impedance spectra obtained at 0 bias. Nyquist plots for (C) Pt/PANI (D) F : SnO₂/PANI electrodes in region I.

Irregular semicircles at higher frequency can be seen in the insets of Fig. 9A and B, caused by irregularity at the PANI electrolyte interfaces. The diameter of the semicircle for the Pt/PANI electrode is smaller than that for F : SnO₂/PANI, which confirms results obtained from the CV and GCD measurements, in that Pt has reduced the value of R_ct at the current collector. Besides the high conductivity of the Pt nanoparticle, another reason for lower R_ct of the Pt/PANI electrode is associated with the electropolymerization process during which aniline is oxidized into PANI. During the electropolymerization process, Pt nanoparticles may bind to available nitrogen sites in a PANI chain and form interchain linkages among several adjacent PANI chains. The reduced R_ct of the Pt/PANI electrode leads to shortening of the ion diffusion path, reflected by an almost vertical line at low frequency, whereas a slight deviation from a vertical line is observed for the F : SnO₂/PANI electrode. The fitting results of the elements in equivalent circuits for both electrodes at open circuit potential can be described as, first, series resistance, the values of R_s for Pt/PANI and F : SnO₂/PANI are 41.9 Ω and 44.4 Ω, respectively. As mentioned previously, R_s is derived from various factors, of which the resistance of the electrolyte solution is predominant.^31,37,43 A lower value of R_s for the Pt/PANI electrode suggests a better charge discharge rate performance compared with the F : SnO₂/PANI electrode. This fact becomes more prominent by comparing this result with GCD in Fig. 3. The second factor is R_ct, where the values of R_ct for Pt/PANI and F : SnO₂/PANI are 12 Ω And 21 Ω, respectively. Herein, the value of R_ct is mainly associated with dispersion of ions in the PANI matrix, and transportation and penetration of ions in the interior of the PANI matrix. A lower value of R_ct for Pt/PANI signifies a higher degree of freedom of the SO₄²⁻ ions in the Pt/PANI matrix compared with F : SnO₂/PANI. The third factor is w, where the values of w for Pt/PANI and F : SnO₂/PANI are 210 and 435 Ω s^−1/2, respectively. The higher value of w for F : SnO₂/PANI presents limitation of ion diffusion and indicates that the access of electrolyte ions to the active electrode surface has greater variation in ion diffusion path length from electrolyte to electrode surface. A similar conclusion was drawn from analysis of the CVs (Fig. 5). The fourth factor is C_i, where the values of C_i obtained for Pt/PANI and F : SnO₂/PANI are 11 mF cm⁻² and 4.5 mF cm⁻², respectively. A lower value of capacitance for F : SnO₂/PANI is consistent with the explanation of the CVs.

To examine the mechanism for the iR_s drop in GCD (Fig. 3), complex Nyquist spectra are shown for the same bias range (region I) in Fig. 9C and D for Pt/PANI and F : SnO₂/PANI electrodes, respectively. From the obtained Nyquist spectra at higher bias and Table 2, the values of capacitance become almost negligible while the values of resistance dominates = for both electrodes. Decrease in conductivity of the polymer because of deprotonation or change in oxidation state towards either a fully oxidized or a fully reduced state, causes a decrease in interfacial capacitance at higher bias. The values of resistance at higher bias are 45 kΩ and 140 kΩ for Pt/PANI and F : SnO₂/PANI, respectively, signifying a higher iR_s drop for the F : SnO₂/PANI electrode in the discharge characteristics. The above result reveals that the values of R_s, R_ct, C_i and w are dependent on the morphology and electrochemical activity of the electrode, which are essential for designing high performance supercapacitors.

The frequency dependence of complex impedance |Z| and capacitance C at different applied potential are shown in Fig. 10 and 11, respectively. The |Z| vs. frequency plots in the bias range of 0–0.15 V for both electrodes follow the same trend, but with differences in their absolute values.

Fig. 10 Frequency dependence of magnitude |Z| at different applied potential for (A) Pt/PANI electrode and (B) F : SnO₂/PANI electrode.

Fig. 11 Frequency dependence of capacitance at different applied potential for (A) Pt/PANI electrode and (B) F : SnO₂/PANI electrode.

The lower value of |Z| for the Pt/PANI electrode is again correlated to its lower intrinsic resistance. In the high frequency region, |Z| is weakly dependent on frequency and yields a value of slope very close to zero for both electrodes, which is characteristic of a pure resistor.⁴¹ However, in the low frequency region a slope of <1 for both electrodes signifies the characteristics of a pseudo capacitor. These observations indicate that the behavior of both electrodes changes from a simple resistor at high frequency to a pseudo capacitor at low frequency. The |Z| vs. frequency plots of both electrodes in the bias region of 0.4–0.6 V are essentially different from that of 0–0.15 V, indicating that the intrinsic charge storage mechanism is qualitatively different in these voltage regions. In the low frequency region |Z| tends to increase with the increase in bias, where the influence of |Z| on the discharge characteristics of supercapacitor is same as those discussed in context of iR_s in the supercapacitor electrode. The value of capacitance at different applied potential as a function of frequency was calculated using the equation Z′′ = (2πfC)⁻¹, where Z′′ is the imaginary part of the impedance, C is the capacitance and f is the frequency.

From the C vs. frequency plot, it is observed that the electrode reaches to full capacitance at low frequency and remains constant at a higher frequency. At low frequency, the value of capacitance for the F : SnO₂/PANI electrode tends to saturate, whereas a linear trend is observed for the Pt/PANI electrode. The value of capacitance obtained from EIS is almost of the same order as that obtained from CV at 5 mV s⁻¹. However, the difference between the absolute values of C obtained from the two different techniques is mainly a result of the different response characteristics of electrode to time and frequency domain, amplitude of perturbation signal and data sampled per frequency decade.^21,31,37 The C vs. frequency plots in the bias range of 0.4–0.6 V for both electrodes show similar behavior, where with the increase in applied bias, a decrease in the value of C is observed at low frequency.

Electrical property of electrode: AC conductivity

The charge transport mechanism in the conducting polymer is illuminated by investigating its frequency response. It is often preferable to investigate the frequency dependence of AC conductivity with the charge carrier systems.⁸ The AC conductivity (σ_ac) of supercapacitor electrodes was calculated from impedance data using σ_ac = σ_dc + ωε₀ε′ tan δ, where σ_dc, ω, ε₀, ε′ and tan δ are the DC electrical conductivity, angular frequency, permittivity of free space, dielectric permittivity and loss tangent, respectively. For insulating materials, if σ_dc ≪ ωε₀ε′ tan δ, then σ_ac = ωε₀ε′ tan δ. However, for conducting polymers σ_dc is large and dominates the value of σ_ac and hence σ_ac = σ_dc.

AC electrical conductivity of both electrodes as a function of applied frequency is shown in Fig. 12. Frequency-independent plateaus appear at low frequency for both electrodes, showing the behavior of DC conduction, which arises in electrodes from contributions of two components: (1) ionic or electronic conductivity, and (2) dielectric factor (ωε′), dependent on the extent of polarization of dipoles and accumulated interfacial charges. At critical frequencies of 60 Hz and 85 Hz for Pt/PANI and F : SnO₂/PANI, respectively, the plateau response starts to bend upwards and become proportional to the applied frequency. A comparatively higher value of critical frequency (f_c) for the F : SnO₂/PANI electrode signifies that the extra contribution of conductivity comes from the capacitive region, which provides less impedance at higher frequency. The increase in the conductivity at high frequency for both electrodes can be attributed mainly to the following: (1) formation of excess charge carriers, (2) charge moves in the crystalline region and this supports the presence of isolated polarons and bipolarons in this region and (3) the contribution of polarons moving along the shorter distance in the polymer chain.⁴⁵

Fig. 12 Frequency dependent electrical response of electrodes.

A lower value of offset in f_c for the Pt/PANI electrode signifies that Pt nanoparticles connect well with the PANI grains, and this causes development of the interconnected PANI network with the current collector, which helps to increase the rate of the electron tunneling pathway. At frequencies above f_c, the polarization effect becomes unimportant as the dipoles have insufficient time to align themselves in the direction of applied field. Thus the electric field at f > f_c causes a reduction in space charge accumulation and dispersion of dipoles in field direction, which reduce the extent of olarization. This is a necessary phenomenon at the Pt/PANI electrode in terms of constant energy density of supercapacitors. At the f_c of the Pt/PANI, the F : SnO₂/PANI electrode continues to operate in the DC conductivity region, where it behaves as a DC blocking capacitor. Thus, hopping of excited electrons through the PANI layer on Pt becomes easier at a comparatively lower frequency than in F : SnO₂/PANI.

The transient behavior of supercapacitors was analyzed in terms of complex capacitance analysis (CCA), which has emerged as an excellent technique for investigating bulk and interfacial electrochemical properties.⁴⁶ The capacitive response at low frequency has been widely employed for the study of pseudo capacitors. The imaginary part of complex capacitance C′′ (ω) for the Pt/PANI and F : SnO₂/PANI electrodes as a function of frequency is shown in Fig. 13.

Fig. 13 The imaginary part of complex capacitance as a function of frequency.

The imaginary part of complex capacitance represents a relaxation process at a supercapacitor electrode during ion transport. The complex capacitance C′′ for both electrodes was extracted from impedance data using the relation: C′′(ω) = Z′(ω)/ω|Z(ω)|², where Z′ (ω) is the real part of complex impedance Z (ω). The rate capability of an electric system (at which the electrode can discharge) is measured by the value of a characteristic parameter τ. The value of τ can be obtained from the peak frequency of the imaginary capacitance plot, by τ = 1/2πf.⁴⁶ The τ measured for Pt/PANI and F : SnO₂/PANI are 0.74 μs and 0.61 μs, respectively. A smaller value of τ for both the electrodes signifies that a higher output power can be delivered in a short span at output terminal from an electric system. As the time constant reflects the rate capability of an electrode, both electrodes will give poor performance at high charge discharge currents. Thus, although a higher value of capacitance is obtained for the Pt/PANI electrode, it has a poor rate capability at high frequency or large charge discharge currents. Therefore it was demonstrated that AC conductivity along with complex capacitance analysis is able to evaluate the transient behavior of supercapacitor electrodes, which is not possible using conventional GCD and CV measurements.

4. Conclusion

This article focuses on the choice and effect of current collector electrode on the performance-indicating parameters of supercapacitors. The GCD results show that a Pt-based electrode has better supercapacitive behaviour than a F : SnO₂ electrode. SEM results reveal that a porous granular array of PANI is induced on the Pt electrode, whereas a compact granular morphology is induced on the F : SnO₂ electrode. The dynamic variation of capacitance with the applied bias reveals a larger value of capacitance for Pt/PANI compared with F : SnO₂/PANI, indicating excessive pore filling and charge transfer taking place at the PANI matrix. The sharp drop in initial voltage on discharge results from the diffusion-limited mobility of electrolyte ions in the electrode pores, and the capacitance of F : SnO₂/PANI mainly arises from the Faradaic reactions of PANI while the capacitance of Pt/PANI is overwhelmed by the electric double layer capacitance of Pt along with the pseudo capacitance of PANI. The impedance measurement leads to determination of parameters such as equivalent series resistance, rate capability of electrodes and AC conductivity of supercapacitors.

Acknowledgements

The authors gratefully acknowledge Dr Narendra Chauhan, FCIPT, IPR, for the SEM measurements and DST (Project no. SR/S1/PC-44/2011 dated 04/07/2012) for the financial assistance. One of the author (K.P.) thanks DST INSPIRE fellowship program for junior research fellowship.

References

1. M. K. Deshmukh and S. S. Deshmukh, Renewable Sustainable Energy Rev., 2008, 12, 235.
2. J. E. Garland, D. J. Crain, J. P. Zheng, C. M. Sulyma and D. Roy, Energy Environ. Sci., 2011, 4, 485.
3. L. F. Chen, X. D. Zhang, H. W. Liang, M. Kong, Q. F. Guan, P. Chen, Z. Y. Wu and S. H. Yu, ACS Nano, 2012, 6, 7092.
4. W. Liu, X. Yan, J. Lang, J. Pu and Q. Xue, New J. Chem., 2013, 37, 2186.
5. Y. Zhai, Y. Dou, D. Zhao, P. F. Fulvio, R. T. Mayes and S. Dai, Adv. Mater., 2011, 23, 4828–4850.
6. F. Su, C. K. Poh, J. S. Chen, G. Xu, D. Wang, Q. Li, J. Lin and X. W. Lou, Energy Environ. Sci., 2011, 4, 717.
7. F. Wang, S. Xiao, Y. Hou, C. Hu, L. Liu and Y. Wu, RSC Adv., 2013, 3, 13059.
8. S. Maiti and B. B. Khatua, RSC Adv., 2013, 3, 12874.
9. S. B. Kulkarni, U. M. Patil, I. Shackery, J. S. Sohn, S. Lee, B. Park and S. Jun, J. Mater. Chem. A, 2014, 2, 4989.
10. I. Kovalenko, D. G. Bucknall and G. Yushin, Adv. Funct. Mater., 2010, 20, 3979.
11. L. Shao, J. W. Jeonand and J. L. Lutkenhaus, J. Mater. Chem. A, 2014, 2, 14421.
12. W. Lin, K. Xu, M. Xin, J. Peng, Y. Xing and M. Chen, RSC Adv., 2014, 4, 39508.
13. H. Mi, X. Zhang, S. Yang, X. Ye and J. Luo, Mater. Chem. Phys., 2008, 112, 127.
14. G. R. Li, Z. P. Feng, J. H. Zhong, Z. L. Wang and Y. X. Tong, Macromolecules, 2010, 43, 2178.
15. W. Chen, R. B. Rakhi and H. N. Alshareef, J. Mater. Chem. A, 2013, 1, 3315.
16. S. Xiong, J. Wei, P. Jia, L. Yang, J. Ma and X. Lu, ACS Appl. Mater. Interfaces, 2011, 3, 782.
17. H. Gao, F. Xiao, C. B. Ching and H. Duan, ACS Appl. Mater. Interfaces, 2012, 4, 2801.
18. X. Chen, H. Wang, H. Yi, X. Wang, X. Yan and Z. Guo, J. Phys. Chem. C, 2014, 118(16), 8262.
19. D. Bhattacharjya and I. Mukhopadhyay, Langmuir, 2012, 28, 5893.
20. K. Pandey, P. Yadav and I. Mukhopadhyay, J. Phys. Chem. B, 2014, 118(11), 3235.
21. J. P. Zheng, C. M. Pettit, P. C. Goonetilleke, G. M. Zenger and D. Roy, Talanta, 2009, 78, 1056.
22. W. C. Chen, T. C. Wen and A. Gopalan, Synth. Met., 2002, 128, 179.
23. W. C. Chen, T. C. Wen, C. C. Hu and A. Gopalan, Electrochim. Acta, 2002, 47, 1305.
24. Z. Tang, J. Wu, M. Zheng, Q. Tang, Q. Liu, J. Lin and J. Wang, RSC Adv., 2012, 2, 4062.
25. M. H. Pournaghi-Azar and B. Habibi, Electrochim. Acta, 2007, 52, 4222.
26. W. Fan, C. Zhang, W. W. Tjiu, K. P. Pramoda, C. He and T. Liu, ACS Appl. Mater. Interfaces, 2013, 5, 3382.
27. J. Zhang and X. S. Zhao, J. Phys. Chem. C, 2012, 116, 5420.
28. J. P. Zheng, P. C. Goonetilleke, C. M. Pettit and D. Roy, Talanta, 2010, 81, 1045.
29. X. Wen, D. Zhang, T. Yan, J. Zhang and L. Shi, J. Mater. Chem. A, 2013, 1, 1233.
30. H. Wang, L. Shi, T. Yan, J. Zhang, Q. Zhonga and D. Zhang, J. Mater. Chem. A, 2014, 2, 4739.
31. J. Zheng, S. S. Moganty, P. C. Goonetilleke, R. E. Baltus and D. Roy, J. Phys. Chem. C, 2011, 115, 7527.
32. Y. Li, X. Zhao, Q. Xu, Q. Zhang and D. Chen, Langmuir, 2011, 27, 6458.
33. L. Bertoluzzi, I. Herraiz-Cardona, R. Gottesman, A. Zaban and J. Bisquert, J. Phys. Chem. Lett., 2014, 5, 689.
34. V. Gupta and N. Miura, Mater. Lett., 2006, 60, 1466.
35. V. Gupta and N. Miura, Electrochim. Acta, 2006, 52, 1721.
36. H. Li, J. Wang, Q. Chu, Z. Wang, F. Zhang and S. Wang, J. Power Sources, 2009, 190, 578.
37. S. S. Moganty, R. E. Baltus and D. Roy, Chem. Phys. Lett., 2009, 483, 90.
38. E. Song and J. W. Choi, Nanomaterials, 2013, 3, 498.
39. Z. L. Wang, X. J. He, S. H. Ye, Y. X. Tong and G. R. Li, ACS Appl. Mater. Interfaces, 2014, 6, 642.
40. K. Pandey, P. Yadav and I. Mukhopadhyay, J. Solid State Electrochem., 2014, 18, 453.
41. W. Sugimoto, H. Iwata, K. Yokoshima, Y. Murakami and Y. Takasu, J. Phys. Chem. B, 2005, 109, 7330.
42. H. Farsi, F. Gobal, H. Raissi and S. Moghiminia, J. Solid State Electrochem., 2010, 14, 643.
43. J. Bisquert, Phys. Chem. Chem. Phys., 2007, 10, 49.
44. C. Masarapu, H. F. Zeng, K. H. Hung and B. Wei, ACS Nano, 2009, 3, 2199.
45. A. B. Afzal, M. J. Akhtar, M. Nadeem and M. M. Hassan, Curr. Appl. Phys., 2010, 10, 601.
46. C. Yang, M. Zhou and Q. Xu, Phys. Chem. Chem. Phys., 2013, 15, 19730.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra09277e

---