Activated carbon derived from natural sources and electrochemical capacitance of double layer capacitor

Transcription

1 Indian Journal of Chemical Technology Vol. 20, November 2013, pp Activated carbon derived from natural sources and electrochemical capacitance of double layer capacitor M Natalia 1, Y N Sudhakar 2 & M Selvakumar 2, * 1 Norwegian University of Science and Technology, Faculty of Science and Technology, Trondheim, Norway 2 Department of Chemistry, Manipal Institute of Technology, Manipal , India Received 9 April 2012 ; accepted 11 June 2013 Carbon materials have been synthesized from areca fibres by burning them in an oven at 150 C, which is then gradually raised to 500 C for 1 h under flowing nitrogen at a heating rate of 10 C/min. The electrochemical double layer performance of these prepared carbon materials from natural sources is studied in 0.1 M Na 2 SO 4. Two types of electrodes have been prepared; first one with synthesized non-activated carbon and second with activated carbon using KOH and steam. Surface properties of the prepared carbon materials are studied using scanning electron microscopy. The surface area of the activated carbon is found to be 250 m 2 /g. The electrochemical properties of activated and non-activated carbons are evaluated by using cyclic voltammetry and AC impedance spectroscopy. Supercapacitors have been fabricated symmetrically using these electrodes and their electrochemical properties are studied. Specific capacitance as high as 47 F/g in 0.1 M Na 2 SO 4 electrolyte is obtained and it shows excellent cyclability with a coulombic efficiency of 95% at current density of 3 ma/g for 1000 cycles. The electrochemical performance of the high surface area carbon in the neutral electrolyte medium is found to be significantly high, and the reasons are discussed. Keywords: Areca fibres, Activated carbon, Double layer capacitor, Electrochemical capacitance, Supercapacitor Due to high requirement for energy and environmental problems, new techniques for energy storage are most desired. Greater use of natural resources is promoted, but to be able to fully use them, we require ability to store energy when it is available. Supercapacitors are a promising option, since they can be charged quickly and can be discharged quickly when the energy is needed. Supercapacitors, or electrical double-layer capacitors (EDLCs), were in use for several years now. Unfortunately, current EDLCs cannot store enough energy, commensurate to our needs. Therefore, new substrates need to be developed to increase energy density, without increasing the apparent area of an EDLC. Various carbonaceous materials such as activated carbon and carbon nanotubes have been currently investigated for use as electrode materials for EDLCs 1-3. Amongst those products, activated carbon is the most popular because of its abundance, cost effectiveness, and environmentally benign nature. EDLCs with activated carbon exhibit higher energy density than conventional capacitors. This is because of their extremely high surface areas 3-5. *Corresponding author. chemselva78@gmail.com Carbon can be activated either by a physical or chemical process. During physical process a large amount of internal carbon is removed to impart welldeveloped carbon structures, while in the chemical activation process, dehydrating agents are used, which influence pyrolytic decomposition, resulting in the formation of tar, which thus increases the carbon yield. There are different forms of activated carbon. The best choices for EDLC applications are activated carbon powder (ACP), activated carbon fibre cloth (ACF-cloth), and synthetic carbon aerogel. The characteristics of EDLCs are influenced by the specific surface area and pore size distribution, which is crucial for capacitance 6. The high surface area and the porosity of the activated carbon are the basic advantages to form an effective double-layer, which results in an EDLC with a high power density and long cycle life. Different activation processes might improve the capacitive behaviour of carbon materials. In the present study, activation of carbon derived from areca fibres is performed with KOH, where it acts as pore-forming agent, thereby improving the specific surface area of the material. The areca fibres are naturally occurring and give an advantage for preparing the activated carbon. Carbonization process is simple and environmentally benign. It is believed

2 NATALIA et al.: ACTIVATED CARBON DERIVED FROM NATURAL SOURCES 393 that the activation process of the carbon fibres increases the pore structure and modifies the morphology with a substantial increase in the specific surface area which is very critical for a better doublelayer capacitance. It has been reported 6 that activation of carbon using KOH provides a very large surface area of over 2000 m 2 /g. Graphitized carbons activated with KOH are reported to show a large capacitance of 40 F/g by weight in two electrode systems 5. Many different carbon materials from different sources were reported and activated for the use as an EDLC electrode, but only few studies have been reported on the synthesis of carbon from a natural precursor for EDLC applications 7-9. This inspired us to explore the opportunity of utilizing the carbon materials derived from areca fibre in a neutral electrolyte medium. The advantage of this activated carbon is the fact that it is prepared in an eco-friendly manner from renewable and natural resources. Process is cheap and easy, while obtained results are satisfying. This paper therefore reports the EDLC performance of carbon materials derived from areca fibres before and after activation using KOH. The structural and electrochemical properties of the carbon materials are also discussed. Experimental Procedure Preparation of activated carbon from areca fibres High quality areca nut shells were obtained from Karnataka, India. These shells were soaked, washed and dried. Then fibres were separated from outer and inner shells and divided into two sets. One set of fibres was placed in silica container for burning and then put in preheated oven to the temperature 150 C. The temperature was gradually raised to 500 C for 1 h under flowing nitrogen at a heating rate of 10 C/min. Fibres stayed in the oven at this temperature for approximately 1 h. Then they were removed for grinding and sifting. Obtained carbon powder was non-activated carbon (non-acs) sample. Another set of fibres was activated by treating with 0.1 M KOH solution for 3 days at 150 C wherein first and second day, KOH solution was freshly added in 1:4 ratio. These dried fibres were subjected to same burning process and so obtained carbon powder was further activated by passing steam under 16 lbs pressure for 20 min at 120 C by using autoclave. Surface area, pore size and pore distribution of the activated carbon (ACs) were tested using nitrogen absorption-desorption experiments at 77 K (Micromeritics ASAP 2010 surface area analyser) and calculated by the BET and BJH method, respectively. Fabrication of AC electrodes and symmetric supercapacitors The working electrodes were fabricated as follows: typically, the activated carbon (30 mg) and 1% of poly-(vinylidene fluoride) (PVdF) in N-methyl-2- pyrrolidone (NMP) were mixed to produce homogenous slurry, which was coated onto a 0.04 mm thick flexible stainless steel sheet (1 cm 1 cm). Finally, the fabricated electrodes were dried at 50 C in an oven for overnight. The loaded activated carbon was calculated to be ~0.6 mg (excluding carbon black and PVdF mass) on each electrode and thickness of the coated AC was found to be 0.04 mm by using a screw gauge. The symmetric supercapacitor (SC) was assembled using the prepared activated carbon electrode, separated by a polypropylene sheet and immersed in 0.1 M Na 2 SO 4. Surface morphology of the prepared two electrodes was examined on a JEOL JSM-6380LA scanning electron microscope. Electrochemical characterization The electrochemical performance was studied using cyclic voltammetry (CV), AC impedance at frequency range Hz, and galvanostatic charge and discharge measurements. CV studies were performed using a potentiostat/galvanostat (BioLogic SP-150, EC-Lab) in a three-electrode configuration using the stainless steel plate coated with activated carbon as the working electrode, Pt electrode as the counter, and saturated Calomel electrode (SCE) as the reference. CV studies were recorded between 0 volt and 1 volt in a 0.1 M Na 2 SO 4 neutral electrolyte at different scan rates. The specific capacitance was evaluated from the area of the charge and discharge curves of the CV plot. Galvanostatic charge/discharge experiments were performed in a similar setup as described above using same instrument, with a specific high current of 3 ma/g and between -1 volt and 1 volt. All the experiments were performed at room temperature. Results and Discussion SEM analysis Pore structure of areca fibres was studied using the SEM images (Fig. 1a and b) of activated carbon before and after treatment. As observed in the micrographs the carbon particle size decreases on activation, leading to better surface area of the sample. Number of pores and pore size on the electrode also increase, thus influencing capacitance.

3 394 INDIAN J. CHEM. TECHNOL., NOVEMBER 2013 increasing adsorbed volume at low pressure (~0-0.2) is related to micropores, the desorption hysteresis at medium relative pressure (~ ) reveals the presence of mesopores and the small tails at a relative pressure near to 1.0 (>0.9) reveals the presence of macropores 10,11. The obtained Brunauer Emmett Teller (BET) surface areas, total pore volumes and average pore sizes are given in Table 1. Mesoporous and microporous surface areas and pore sizes of ACs are obviously increased from non-activated carbon to activated carbon. Among the prepared carbon materials, ACs show the high surface area (250 m 2 / g), large pore volume 0.24 cm 3 g 1 and the best pore size 2.33 nm, whereas non-acs show the surface area 200 m 2 /g, pore volume 0.2 cm 3 g -1 and the best pore size 1.88 nm. Figure 2b shows the pore size distribution of the ACs which is calculated by the Barrett Joyner Halenda (BJH) method. The maximum peak is observed for activated carbon (Fig. 2c) centred at 60 nm which infers that majority of pores are mesopores. Moreover, activated carbon displays the largest adsorbed volume (Fig. 2d) which indicates a higher surface area, larger pore volume, with better pore size. Fig. 1 SEM images of (a) non-activated carbon and (b) activated carbon Surface properties of the synthesized carbon materials are critical for the electrochemical performance when used as an electrode material in double-layer capacitors. Large pore size is easily accessible surface for the ions and it gives better capacitance. The fundamentals behind the chemical activation using KOH are known for different carbon materials. The pore generation in carbon materials treated with KOH is caused by the presence of oxygen in KOH, which results in the elimination of cross-linking and the stabilization of carbon atoms in the crystallites. Porosity, chemical and physical characterization of AC The N 2 adsorption-desorption isotherms, pore size distributions and the relation between pore size and adsorbed pore volumes are shown in Fig. 2a for the prepared activated carbon. Type IV isotherms are observed in Fig. 2b, which indicates the presence of different pore sizes from micro to macropores. The Cyclic voltammetry (CV) and AC impedance Typical CV curves for the areca fibre carbon are shown in Fig. 3A-D. The CV curves exhibit a typical capacitor behaviour showing a nearly rectangular shape for most scan rates. It is observed that the shape differs more from rectangle while voltage increases. Similar patterns have been observed for other activated carbons when the scan rate is higher than 50 mv/s The main idea behind the activation process using KOH and autoclave is to improve porosity of the carbon, thereby increasing the specific surface area. The specific capacitance values for the non-activated carbon at all scan rates are low because of the smaller surface area of the material. As expected, the activated carbons show an improved specific capacitance mainly because of the increased surface area. The specific capacitance values of the various electrodes and the supercapacitor have been calculated from the respective cyclic voltammetry using the equation 15 C= I/S, where S is the potential sweep rate and I, the average current. Capacitance of the carbon before activating is found to be 34 F/g at 5 mv/s and 22 F/g at 50 mv/s. In the case of the carbon activated using KOH and steam passing, the specific capacitance is 64 F/g at 5 mv/s, whereas it

4 NATALIA et al.: ACTIVATED CARBON DERIVED FROM NATURAL SOURCES 395 Fig. 2 (a) N 2 adsorption-desorption isotherms, (b) pore size distribution, and (c and d) pore width versus volume absorbed of activated carbon and non-activated carbon Sample Table 1 Surface area, pore volume, pore size and capacitance of the prepared activated Carbon Surface area Pore volume Pore size (S BET ), m 2-1 (V p ) a, cm 3 g -1 (d p ) a, nm Surface area Pore volume Capacitance Capacitance (S micropore ) b, m 2 g -1 (V micropore ) b, cm 3 g -1 (C sp ) c, Fg -1 (C sp ) c, Fg -1 AC Non-AC a Total pore volume calculated at P/P 0 = 0.99 and average pore diameter from BET method. b Microporus surface area and micropore volume from t-plot method. c Maximum specific capacitance calculated at 2 ma cm -2 from three electrode system. d Maximum specific capacitance calculated at 2 ma cm -2 from two electrode system. decreases to 36 F/g at a scan rate of 50 mv/s. These results of non-ac and AC are compared in Fig. 3C, for the scan rate of 5 mv/s. It is observed that activated carbon has higher capacitance than non-activated carbon. The highest capacitance for supercapacitor with activated carbon obtained at 5 mv/s is 47 F/g. The main reason for the enhanced specific capacitance is high specific surface area, large pore volume (0.24 cm 3 g 1 ) and its pore size of 2.33 nm. The increase in the specific surface area leads to a significant increase in the double-layer capacitance. Further, the AC impedance studies (Fig. 4a) show that the charge-transfer resistance of carbon before activation is 130 ohms, when it s activated the charge-transfer resistance decreases to 80 ohms. This

5 396 INDIAN J. CHEM. TECHNOL., NOVEMBER 2013 Fig. 3 CV of (A) non-activated carbon, (B) activated carbon, (C) comparison of CVs of activated carbon and non-activated carbon, and (D) carbon-carbon supercapacitor on SS electrode at different sweep rates [(a) 50mVs -1 (b) 25 mvs -1 (c) 20 mv s -1 (d) 10mV s -1 (e) 5mV s -1 ] Fig. 4 AC Impedance spectroscopy response for (a) nonactivated carbon and activated carbon, and (b) carbon-carbon supercapacitor indicates that by activation there is an increase in the conductivity of the activated carbon. In all cases, a semicircle of large radius is obtained at high frequency region and a small straight line is obtained in the low frequency region. The charge-transfer resistance of the electrodes decreases at low frequencies due to a larger number of ions movement, this also reflects the result of cyclic voltammetry. Figure 4b shows a semicircle that is characteristic for supercapacitors impedance and a long straight line in the low frequency region. The semicircle results from the parallel combination of resistance, capacitance and the plot from linear region is because of Warburg impedance. In the low frequency region, the straight line part leans more towards imaginary axis and this indicates good capacitive behavior at low frequencies 16. From Figs 4 and 5, the double-layer capacitance of the active material 17 is calculated using the following formula: C dl 1 =... (1) 2π f R max ct where C dl is the double-layer capacitance of the material; f max, the frequency at maximum value of real part; Rct, the high frequency value-low frequency value. Table 2 shows double-layer capacitance of non- ACs, ACs and supercapacitor. Increase in capacitance values is attributed to an increase in accessible pores for the ions towards the electrode in the activated sample, which is due to smaller particle size of carbon, thereby showing large surface area as also

6 NATALIA et al.: ACTIVATED CARBON DERIVED FROM NATURAL SOURCES 397 supported by SEM images. As reported by Conway 4, electrode double-layer capacitance value of 6.9 µf cm -2 obtained from carbon fibre cloth of 0.51M (Et 4 NBF 4 ) in propylene carbonate is found to be similar to that observed in this study, but it has surface area of surface-area 1630 m 2 /g which is about six times greater than our prepared activated carbon (250 m 2 /g), implicating better pore size and accessibility for neutral electrolyte in the areca derived activated carbon. Comparatively, the obtained EDLC value and surface area are almost same to that of 7.9 µf cm -2 obtained from carbon black 4 in 1M H 2 SO 4 having surface area of m 2 /g. Time constant is very helpful to predict the supercapacitor whether it is applicable for AC or DC applications. Figure 5 shows real part of capacitance (C ) and the imaginary capacitance part of the supercapacitor (C ), in which there is appearance of large resistance peak at middle frequency region, indicating time constant τ o. It can be seen from the figure that when the frequency decreases, C sharply increases which tends to be less frequency dependent. This is characteristic of the electrode structure and the electrode/electrolyte interface 16. The detailed Table 2 AC impedance spectra Electrodes R ct, Ω C dl, µf cm -2 R s 10 2, Ω Non-activated carbon Activated carbon Supercapacitor calculation of the plotting technique have been taken from the literature 16,17. The importance Z(ω) data obtained from supercapacitor can be written under in its complex form: Z(ω) = Z (ω) + j Z"... (2) where Z (ω) and Z"(ω) are the real part and the imaginary part of the impedance respectively; j, the complex number (-1); and ω, the angular frequency. To describe the supercapacitor by using capacitance that is the functions of the pulsation ω noted as C(ω), it can be stated as C(ω) = C (ω) j C (ω)... (3) These C and C capacitance of the supercapacitor varies with frequency and can be defined as C (ω) = -Z"(ω)/ω Z(ω) 2... (4) C (ω) = Z (ω)/ω Z(ω) 2... (5) The low-frequency value of C (ω) corresponds to the capacitance of the cell and C (ω) corresponds to energy dissipation by an irreversible process that can lead to the movement of molecules. Using normalized reactive power Q / S % and active power P / S % versus frequency plot for the 1 cm 2 cell, time constant of the fabricated supercapacitor has been also calculated (Fig. 6). The complex power can be calculated using classical form, as shown below: S(ω) = P(ω) + j Q(ω)... (6) This equation comes directly from the definition of a complex number. P is called the active power (watt) and Q the reactive power. The use of complex capacitance from Eqs (3) and (4) leads to the following expressions: Fig. 5 Evolution of (a) real part and (b) imaginary capacitance vs. frequency for 1 cm 2 cells assembled with two electrodes containing activated carbon in 0.1 M Na 2 SO 4 Fig. 6 Normalized reactive power Q / S % and active power P / S % vs frequency (Hz) for the 1 cm 2 cell carbon-carbon super capacitor

7 398 INDIAN J. CHEM. TECHNOL., NOVEMBER 2013 P(ω) = ω C"(ω) V rms 2... (7) Q(ω) = -ω C (ω) V rms 2... (8) where V rms = V rms / 2 is maximum amplitude of the electrical signal. The relaxation time constant of the supercapacitor determined from these data is found to be 2 s. The time constant (τ o ) represents a transition for the supercapacitor between a resistive behavior for frequency higher than 1/τ o and a capacitive behavior for frequency lower than 1/τ o. Galvanostatic charge-discharge In order to evaluate the cycling stability of the synthesized carbon materials, galvanostatic chargedischarge studies were performed at a current of 3 ma/g between -1 volt and 1 volt in the 0.1 M Na 2 SO 4 electrolyte. The long cycle performance is shown in Fig. 7. The curve profile clearly indicates typical capacitor behaviour. The electrical parameters namely specific capacitance (SC), specific power (SP), and specific energy (SE) are calculated 18 using the following formulae: SC (Fg -1 ) = [I t]/[i m]... (9) SP (Wg -1 ) = [I I]/m... (10) SE (Whg -1 ) = [I t I]/m... (11) where I(A), I(V) t(s), m(g) respectively are dischargecurrent in amperes, voltage change after full charge or Fig. 7 Charge-discharge characteristics of carbon-carbon supercapacitor Cycle Specific capacitance, F/g discharge in volts, discharge time in seconds, and mass of active materials in grams. The results are presented in Table 3. There has been well-documented data on the electrochemical performance of the activated carbon supercapacitor electrodes 19. Nonetheless, most of the works have been reported using either an acidic electrolyte such as H 2 SO 4 or basic electrolyte such as KOH. It has been observed that the double-layer formation and the resulting capacitance show a strong dependence on the concentration of the electrolyte supported by the high surface area arising from the porous structure of the carbon. In the case of the acidic electrolyte, because of the presence of proton as the mobile species, better ionic diffusion is observed into the pores of the carbon, resulting in an enhanced EDLC performance. However, the main disadvantage is the high corrosion rate of the current collectors because of the extreme conditions of the electrolyte which limits the use of such an electrolyte system. In the case of RuO 2 (a potential pseudocapacitor material), the use of 5.3 M H 2 SO 4 is required to extract the full potential of the pseudocapacitance property 20. Here, both ion dissolution into the electrolyte combined with corrosion of the current collector have been reported to deteriorate the cycling performance of the capacitor. The aforesaid disadvantages motivated us to investigate the activated carbon performance in a neutral electrolyte such as 0.1 M Na 2 SO 4. Hseih 13 reported the electrochemical performance of carbon derived from melamine resin heat treated at different temperatures in two different electrolytes, 1 M H 2 SO 4 and 3 M NaCl. In the case of neutral electrolytes, the specific capacitance is mainly governed by the nonfaradic electrostatic sorption of ions at the double layer 11. Comparing the electrochemical capacitance of the carbon derived from melamine resin and that prepared from areca fibre in the neutral electrolyte medium, the specific capacitance and electrochemical performance are as follows. For the melamine resin based carbon, the best reported specific capacitance is 75 F/g at a low constant current at 20 mv/s 21. In the present case, the specific capacitance is 64 F/g at a Table 3 Charge-discharge characteristic nature of carbon-carbon supercapacitor Specific power W/g Specific energy, Wh/g Columbic efficiency, N% Equivalent series resistance, Ω Voltage drop, V

8 NATALIA et al.: ACTIVATED CARBON DERIVED FROM NATURAL SOURCES 399 constant current at 5 mv/s, which is found to be better at lower scan rates. The main difference between the carbon derived from melamine resin and areca fibre is the specific surface area. For the melamine resin based carbon, the maximum surface area reported is 345 m 2 /g. Thus, high surface area need not necessarily contribute to high capacitance for EDLC 22. Hence, the improved rate capability and enhanced long cycle performance in the present case are mainly due to the natural sources of areca fibres. Conclusion (i) Carbons from areca fibres treated with KOH and steam under pressure have Been successfully activated. The activated carbon shows enhancement in surface area of 250 m 2 /g and pore size of 2.33 nm, which leads to higher specific capacitance of 64 F/g at scan rate of 5 mv/s. (ii) These activated carbons materials exceed excellent electrochemical properties in comparison to that of non-activated carbon, and furthermore, these are prepared from natural and renewable sources. The specific capacitance of this activated carbon material by fabricating into supercapacitor shows a value of 47 F/g at scan rate of 5 mv/s and 95% coloumbic efficiency for first cycle at a current density of 3 ma/g. (iii) This study shows that it is possible to fabricate a supercapacitor from ecofriendly materials. Because of the green processing, the cost associated with the precursor and the simplicity in the activation process, this kind of capacitors has a bright future. Acknowledgement The one of authors (MN) acknowledges the International Association for the Exchange of Students for Technical Experience (IASTE) India for providing summer research internship. References 1 Ahmadpour A & Do D D, Carbon, 34 (1996) Ahmadpour A, King B A & Do D D, Ind Eng Chem Res, 37 (1998) Lua A C & Yang T J, Colloid Interface Sci, 290 (2005) Conway B E, Electrochemical Supercapacitors: Scientific Fundamentals and Technological Applications (Kluwer- Plenum Press), NewYork Frackowiak E & Beguin F, Carbon, 40 (2002) Tanahashi I, J Appl Electrochem 35 (2005) Guo Y, Qi J, Jiang Y, Yang S, Wang Z & Xu H, Mater Chem Phys, 80 (2003) Wu F C, Tseng R L, Hu C C & Wang C C, J Power Sources, 138 (2004) Liu Y, Xue J S & Dahn J R, Carbon, 34 (1996) Huang W, Zhang H, Huang Y, Wang W & Wei S, Carbon, 49 (2011) Xing W, Huang C C, Zhuo S P, Yuan X, Wang G Q, Jurcakova D H, Yan Z F & Lu G. Q, Carbon, 47 (2009) De Levie R, Electrochim Acta, 8 (1963) Hsieh C T & Teng H, Carbon, 40 (2002) Nian, Y R & Teng H, J Electrochem Soc, 149 (2002) A Bhat K D & Kumar S M, J Mater Sci, 42 (2007) Taberna P L, Simon P & Fauvarque J F, J Electrochem Soc, 150 (2003) A Ratner M A & Shriver D F, Chem Rev, 88 (1988) Rajendraprasad K & Munichandraiah N, Electrochem Solidstate Lett, 5 (2002) A Lee G J & Pyun S, J Korean Electrochem Soc, 9 (2006) Zheng J P, Cygan P J & Jow T R, J Electrochem Soc, 142 (1995) Subramanian V, Zhu H W & Wei B Q, Electrochem Commun, 8 (2006) Lozano-Castello D, Cazorla-Amoro D, Linares S A, Shiraishi S, Kurihara H & Oya A, Carbon, 41 (2003) 1765.