Investigation of a gel polymer electrolyte based on polyacrylonitrile and magnesium chloride for a redox capacitor

To supplement the ever increasing power demand in the present day world, supercapacitors have emerged as one of the suitable power storage devices. Redox capacitors have received a considerable attention due to factors such as high cycle ability, satisfactory specific capacitance and good stability. There are many reports on redox capacitors based on liquid electrolytes but these have several drawbacks. In this study, a magnesium chloride (MgCl2) based gel polymer electrolyte was prepared and characterized for its suitability to be used in a redox capacitor. The composition, 0.1 polyacrylonitrile (PAN):0.4 ethylene carbonate (EC):0.4 propylene carbonate (PC):0.125 MgCl2 (weight ratios) exhibits the highest room temperature conductivity of 4.08×10 -4 S cm -1 . Conductivity variation of this sample with the temperature shows Arrhenius behavior. The ionic transference number is 0.84. This shows that the GPE is an ionic conductor. Redox capacitors were fabricated with two polypyrrole (PPy) electrodes. They were characterized using Cyclic Voltammetry technique, Impedance Spectroscopy and Galvanostatic Charge Discharge test. Cycling at the scan rate of 10 mV/s within the potential window of -1.4 V to +1.4 V showed a specific capacitance of 120.35 F/g. Impedance data showed the electrochemical behavior of the device having resistive and capacitive behavior. Galvanostatic charge discharge test results exhibit average power density and energy density of 341 W/kg and 5 Wh/kg respectively.


INTRODUCTION
At present, there is a rising demand for portable power with the rapid growth of various applications including hybrid and electric vehicles, space crafts, memory backups, toys and many other devices.To meet this high demand, efficient and low cost batteries and supercapacitors are being developed all over the world as energy storage devices (Palacin et al., 2009;Abruna et al., 2008).Batteries can store energy but their power is low compared to supercapacitors.It is possible to define supercapacitors as power sources bridging the gap between batteries and conventional capacitors.Due to shorter time needed for full charging, high columbic efficiency and longer life than batteries, supercapacitors are widely employed in many applications today (Simon and Gogotsi, 2008;Yan et al., 2014).There are two types of supercapacitors namely redox capacitors and electrochemical double layer capacitors (EDLCs).Redox capacitors use electrodes made by conducting polymers or transition metal oxides (Ramya et al., 2013).They derive the capacitance from the storage of charge in the electrode material in response to a redox reaction.EDLCs use carbon based electrodes.The capacitance arrives from pure electrostatic attraction between ions in the electrolyte and the charged surface of an electrode (Zhong et al., 2015).In view of a larger number of charge discharge cycles, satisfactory specific capacitance and good stability, redox capacitors have become prominent.In addition, the easy and less complex methods of fabrication as well as good flexibility of conducting polymers have given the redox capacitors a global attention.Most reports on redox capacitors are based on liquid electrolytes (Fonseca et al., 2006;Uppugalla et al., 2014).

RESEARCH ARTICLE
Due to adverse effects arising from liquid electrolytes such as leakage, evaporation and dye desorption, nowadays attention has been focused on fabrication of redox capacitors with gel polymer electrolytes (GPEs) which exhibit liquid-like ionic conductivities while having mechanical properties similar to solid electrolytes (Tripathi et al., 2013;Hashmi et al., 2005).
In this study, redox capacitors were fabricated using a GPE based on polyacrylonitrile (PAN) with magnesium chloride (MgCl 2 ) and polypyrrole (PPy) electrodes and were characterized using Cyclic Voltammetry, Electro-chemical Impedance Spectroscopy (EIS) and Galvanostatic Charge Discharge (GCD) cycling.The composition of the GPE was fine tuned first to obtain a high conducting, free standing thin film and its nature of conductivity was also investigated.

EXPERIMENTAL Preparation of Gel Polymer Electrolyte samples
PAN (Aldrich, 99%), ethylene carbonate (EC) (Aldrich, 98%), propylene carbonate (PC) (Aldrich, 99%) and MgCl 2 (Aldrich, 98%) were used without further purification.First, the required amount of MgCl 2 was dissolved in EC and PC using magnetic stirring.When MgCl 2 was completely dissolved, required amount of PAN was added to the mixture and stirring was continued until a homogenous mixture was formed.Then, it was heated at 120 o C for 45 minutes inside a glass tube furnace.Resultant highly viscous mixture was pressed in between two well-cleaned glass plates.
Several samples were prepared with different PAN and MgCl 2 concentrations to find the composition that has the highest room temperature conductivity.EC and PC weight ratio was fixed as 1 : 1throughout the study.

AC Conductivity measurements
A circular shape sample with 14 mm diameter was cut from each GPE film.It was sandwiched between two stainless steel (SS) electrodes and assembled in a spring loaded brass sample holder.AC impedance measurements were taken from 2.5 MHz to 0.01 Hz frequency range using a computer controlled Metrohm Autolab M101 frequency response analyzer.Conductivity of GPE, σ was calculated using the Eq.(1).
where t is the thickness and A is surface area of the GPE pellet.R b was calculated using the resultant impedance plot and Non Linear Least Square fitting programme.
To study the conductivity behavior with temperature, impedance data were collected from 28 o C to 50 o C for the sample having the highest room temperature conductivity.

DC polarization Test
DC polarization test using blocking electrodes was performed to identify the contribution from ions and electrons to the conductivity.The GPE pellet was sandwiched in between two stainless steel (SS) electrodes and the current through the GPE pellet was measured as a function of time by applying 1.0 V DC bias potential.t i was calculated according to the Eq.(2).
where t i is the ionic transference number, i e and i T are stabilized current and initial total current respectively.

Preparation of redox capacitors
Pyrrole which was purchased from Aldrich and stored in the dark in a refrigerator was electrochemically polymerized in the presence of dodecylbenzenesulfonate (DBS -Aldrich) on fluorine tin oxide (FTO) glass plates by electrochemical polymerization using a computer controlled potentiostat (Metrohm Autolab101).The three electrode electrochemical cell used for electrochemical polymerization consisted of a working electrode (FTO), a Ag/AgCl reference electrode and a Pt counter electrode.Polypyrrole (PPy) films of the thickness of about 1.0 µm were prepared by applying a current density of 1 mA cm -2 .A GPE film was sandwiched in between two PPy electrodes each of area 1 cm 2 to fabricate the redox capacitors.

Characterization of redox capacitors
For the cyclic voltammetry tests, a three electrode electrochemical cell was used where one electrode of the redox capacitor was used as the working electrode and the other one as the reference and counter electrodes.Cycling was done by varying the potential window using a computer controlled potentiostat (Metrohm Autolab M101) at the scan rate of 10 mV/s.Specific capacity was calculated using the Eq.( 3) (Wang et al., 2013). (3) In the Eq. ( 3), ʃ IdV is the integrated area of the cyclic voltammogram, m is the mass of a single electrode, Δv is the potential window and S is the scan rate.
The redox capacitors were tested using EIS.Impedance measurements were taken using the Metrohm Autolab impedance analyzer over the frequency range 0.01 Hz -0.4 MHz at room temperature.
For GCD test, the potential response vs time was recorded at a constant current of 100 μA.Charging and discharging of the redox capacitor was done at the potential limits 1.4 V and 0 V. Following equation (Eq.( 4)) was used to calculate the discharge capacity (Wang et al., 2013).
where I is the constant discharge current, m is the single electrode mass, dt is the discharge time and dV is the potential drop upon discharging excluding IR drop.
Specific energy density (E s ) and specific power density (P s ) were calculated using the following equations (Eq.5 and Eq.6) (Wang et. al., 2013). (5) Here C is the specific capacitance, V is the charging potential and t is the discharging time.

Determination of the optimum composition
Variation of the room temperature conductivity with MgCl 2 concentration is shown in Figure 1.
It is seen that the room temperature conductivity increases with increasing MgCl 2 concentration and attains the maximum conductivity of 4.08×10 -4 S cm -1 .Further increase of MgCl 2 decreases the conductivity.It is known that conductivity of an ionic conductor is governed by several factors mainly by the concentration as well as mobility of charge carriers.
The initial increase of conductivity with salt concentration is evidently due to the increasing of charge carrier concentration.The subsequent decrease of conductivity after reaching the maximum value is very likely due to formation of ion pairs in the GPE which do not facilitate conductivity (Kumar and Munichandraiah, 2001).
. In order to change the composition of GPE further to obtain the highest conductivity, the polymer concentration was also varied.Conductivity variation with PAN concentration is shown in Figure 2. Room temperature conductivity increases with PAN concentration and reaches a maximum at 0.1 (by weight), followed by a decreasing trend thereafter.When increasing the PAN concentration, assistance for ion motion from polymer network may increase.This might be the reason for initial conductivity increase.As PAN concentration increases further, viscosity of the system also increases so that there will be a resistance for ion mobility (Sharma and Sekhon, 2006).Due to this, the conductivity may be reduced.
From these two studies, it could be concluded that the composition, 0.1 PAN:0.4EC : 0.4 PC : 0.125 MgCl 2 (weight basis) shows the highest room temperature ionic conductivity of 4.08×10 -4 S cm -1 .This composition resulted a thin and a flexible film free of bubbles and pinholes.
Conductivity variation of the highest conductive gel polymer electrolyte sample with temperature is shown in Figure 3.It has a linear behavior suggesting that conductivity variation with temperature follows the Arrhenius behaviour given by the following equation.
where σ 0 is the pre-exponential factor, E a is the activation energy for conduction, k is the Boltzmann constant and T is the temperature.

Calculation of ionic transference number
The ionic transference number (t i ) was estimated by DC polarization test using two stainless steel (SS) electrodes.Figure 4 shows the current variation through the GPE with time.Ionic transference number was calculated using Eq.( 2) and it was 0.84.This is an indication for the predominantly ionic nature of the conductivity (Osman et al., 2012).

Characterization of redox capacitors
Figure 5 shows is the resulting cyclic voltammograms obtained for different potential windows when cycled at the scan rate of 10 mV/s.Mirror image current response of voltammogrnms on voltage reversal exhibits the good reversibility of the redox capacitor (Wu et al., 2010).Specific capacitance (Cs) was calculated for each window and it is shown in Figure 6.
With the increase of the potential window, Cs increases first and then, decreases after a maximum value of 120.35 F/g.The potential window that results the highest specific capacity is in the range -1.4 V to 1.4 V.In redox capacitors, capacitance is derived from redox reactions which are very fast and faradaic.When increasing the potential window, the probability for the occurrence of the complete redox reaction is high, so that, Cs can increase.However, further enhancement of the potential window will not support for redox reactions and therefore, Cs reduces.
The average value of C S was 120 F/g which is quite satisfactory compared to some reported values for redox capacitors based on GPEs (Gupta and Tripathi, 2013;Hashmi et al., 2005).Figure 7 shows the Nyquist plot resulted from impedance data.In general, a Nyquist plot of a conventional capacitor consists of a vertical line parallel to the imaginary impedance axis.For normal situations, that does not appear due to some other factors.Hence, it consists of numerous features at different frequency regions as can be observed in Figure 7.For the case of redox capacitors, at high frequency a semi circle appears representing the resistance of the bulk electrolyte.Second semicircle appears at intermediate frequency region due to the charge transfer resistance.With lowering frequency, the capacitive behaviors become dominant.A titled line can be seen due to the capacitance related to Warburg diffusion and it will be followed by another line having a steep rising behavior.It shows the capacitive behavior.The line having the steep rising behavior is not perfectly parallel to imaginary axis of impedance as for a typical capacitive behavior.It may be due to the surface roughness as well as non uniform active layer thickness (Prabhaharan et al., 2006).Figure 8 shows the results of galvanostatic charge discharge test.The linear shape suggests the capacitive behavior (Wang et al., 2013).
The energy density of the redox capacitor was 5 Wh/kg and the power density was 341.64 W/kg.The discharge capacity was about 50 F/g.This low value may be due to some iR drop due to possible interfacial contacts between the electrode and electrolyte.

CONCLUSIONS
GPE prepared with 0.1 PAN : 0.4 EC : 0.4 PC : 0.125 MgCl 2 yields a bubble free thin freestanding electrolyte film having a conductivity of 4.08×10 -4 S cm -1 .DC polarization test confirmed that the nature of the conductivity is predominantly ionic.When employed in a redox capacitor based on PPy electrodes, it showed an average specific capacity of 120 F/g for the scan rate 10 mV/s within the potential window -1.4 V to 1.4 V. From the Electrochemical Impedance Spectroscopy, Warburg type diffusion as well as capacitive behavior was observed.The average energy and power densities were 5 Wh/kg and 341.64 W/kg respectively.

Figure 1 :
Figure 1: Room temperature ionic conductivity variation of the gel polymer electrolyte with MgCl 2 concentration

Figure 2 :
Figure 2: Room temperature ionic conductivity variation of the gel polymer electrolyte with PAN concentration.

Figure 4 :
Figure 4: Variation of current through the symmetric cell with two SS electrodes under 1.0 V DC bias potential.

Figure 5 :
Figure 5: Cyclic Voltammograms obtained by varying the potential window of cycling of the redox capacitor (scan rate : 10 mV/s)

Figure 6 :
Figure 6: Variation of specific capacity with potential window of cycling.

Figure 7 :
Figure 7: Nyquist plot obtained for the redox capacitor.