Hybrid capacitors utilizing halogen-based redox reactions at interface between carbon positive electrode and aqueous electrolytes

Journal of Power Sources xxx (2016) 1e7

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Hybrid capacitors utilizing halogen-based redox reactions at interface between carbon positive electrode and aqueous electrolytes Shigeaki Yamazaki, Tatsuya Ito, Yuka Murakumo, Masashi Naitou, Toshiharu Shimooka, Masaki Yamagata, Masashi Ishikawa* Department of Chemistry and Materials Engineering, Faculty of Chemistry, Materials and Bioengineering, Kansai University, 3-3-35 Yamate-cho, Suita 5648680, Japan

h i g h l i g h t s
Aqueous HC is developed with Br-water treatment of activated carbon positive.
Conventional EDLC is changed into HC just by pretreatment and NaBr electrolyte.
HC shows no capacitance decay to 100,000 cycles with high coulombic efficiency.
The HC realizes 1.8 V operation in spite of an aqueous electrolyte.
1.8 V HC operation provides higher energy and power even than 2.7 V EDLC.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 January 2016 Received in revised form 2 April 2016 Accepted 4 April 2016 Available online xxx

We propose novel hybrid capacitors (HCs) with electrolyte-involved redox reactions of bromide or iodide species by pretreatment of an activated carbon positive electrode. The treatment is simple; impregnation of pores at an activated carbon fiber cloth (ACFC) as a positive electrode with bromine- or iodinecontaining water before cell assembly. The treated positive electrode is applied to a HC cell with a non-treated negative electrode of ACFC and its electrochemical performance is investigated by galvanostatic cycling and leakage current tests. Few studies on such “electrolytic” charge storage systems have provided acceptable capacitor performance because of inevitable self-discharge caused by diffusion of charged species form an electrode to the other one through an electrolyte. Nevertheless, our electrolyteredox-based HCs show excellent performance without undesirable diffusion of charged species. Moreover, the present HC utilizing a bromide redox system fulfills a practical cell voltage of 1.8 V in spite of an aqueous electrolyte system. This high voltage provides excellent energy density, which is 5 times higher than that in a conventional aqueous electric double-layer capacitor (EDLC), and 1.2 times higher even than that in a 2.7 V-class non-aqueous EDLC, while keeping high chargeedischarge rate capability. © 2016 Elsevier B.V. All rights reserved.

Keywords: Hybrid capacitor Electrochemical capacitor Aqueous electrolyte Activated carbon electrode

1. Introduction Electrochemical capacitors (ECs) often called supercapacitors, are significant energy storage devices that are considered as an object of important scientific and industrial development. Generally, ECs store much higher energy density than conventional dielectric capacitors, and provide higher power density than lithium-ion batteries [1,2]. Moreover, ECs have long cycle life (>106 cycles) that makes them maintenance-free energy storage devices

* Corresponding author. E-mail address: [email protected] (M. Ishikawa).

[3e5]. Therefore, ECs have been widely studied for various applications of electric equipment, e.g., automobiles, tramways, buses, cranes, forklifts, and wind turbines. The amount of energy density (E) accumulated in ECs is proportional to the capacitance (C) and the square of voltage (V) according to the following formula: E ¼ 1/2 C V

2

(1)

The capacitance depends essentially on electrode active material, whereas the operating voltage is limited by the stable voltage window of the electrolyte [6e9]. Although the limiting voltage of aqueous EC systems (the theoretical decomposition voltage of

http://dx.doi.org/10.1016/j.jpowsour.2016.04.021 0378-7753/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

2

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7

water is 1.23 V at pH ¼ 7) is much lower than that of non-aqueous ECs, high power density and high specific capacitance are especially remarkable for aqueous ECs [10e12]. On the basis of their charge storage mechanisms, ECs are categorized mainly into electric double-layer capacitors (EDLCs) and pseudocapacitors [1,11]. In other words, the energy storage capacity of an EC arises either from electric double-layer charging or from pseudocapacitance. Electric double-layer charging is a non-faradic process; electronic charges are accumulated at a space charge layer inside electrode material while equivalent ionic charges are accumulated in the electrolyte close to the electrode surface. On the other hand, pseudocapacitance arises from some faradic reactions at electrode surface and generally shows characteristic capacitance behavior depending on the electrode potential [1,11]. To improve energy density of ECs, carbon materials [13e17], transition metal oxides [6,18e21], and conducting polymers [22e25] are well-known candidates for nonfaradic as well as faradic storage. Among the available metal oxides, RuO2 has the highest specific capacitance (~1000 F g1), but would have a prohibitive price as massive electrode material; a vehiclesized EC with RuO2 may cost more than $1 million [26]. Thus, to design advanced, reasonable ECs, one may find novel chargeedischarge mechanisms or new active materials. In our previous paper, we reported that an EC with the potentiostatic treatment at 70  C provides much higher capacitance than an EC without the treatment [27]; with an aqueous 3.5 mol dm3 NaBr electrolyte system, the treatment leads to a dramatic increase in the capacitance of a positive electrode. As a result, the EC with the treatment attains 2.4 times higher capacitance than an EC without the treatment in a high voltage region. In the later study [28e30], we found that the treated positive electrode may utilize not only double layer capacitance but also some redox reactions  involving Br species such as Br, Br 2 and Br3 in the electrolyte [28]. Moreover, we successfully proposed a non-aqueous EC utilizing an ionic liquid electrolyte containing Br; in this EC the positive electrode also involves Br species-based capacitance [29,30]. Thus these ECs can be regarded hybrid capacitors (HCs) with a partially faradaic positive electrode and a usual double-layer capacitancebased negative electrode. Although the above pretreatment is available, long duration is necessary for generating significant capacitance increase. Here we propose a novel aqueous HC utilizing electrolyte redox reactions of bromide ion species by modified treatment of an activated carbon electrode with a short time. This treatment with redox species is based on a simple method: bromine-water impregnation into positive electrode micropores before cell assembly. We also applied iodide redox reactions to such an electrolytic redox-based HC. Few studies [31e33] have so far demonstrated EC or HC utilizing an “electrolyte” charge storage system with high or acceptable performance because of a facile shuttle reaction or diffusion of redox species through an electrolyte between electrodes. In other words, reactive species on an electrode easily migrate to an opposite electrode unless they are effectively immobilized at an electrode. If one can design outstanding EC or HC involving electrolyte redox reaction systems, it may become a great progress of high-energy capacitor technology. In this context, this study is an attempt to realize such high-performance capacitor, which utilizes inexpensive redox systems without expensive nanomaterials nor raremetal oxides. 2. Experimental Activated carbon fiber cloths (ACFC: ACC-507-15, Nippon Kynol Inc., its specific surface area: ca. 1300 m2 g1) were used as electrode material. The treatment of ACFC with redox species was performed by impregnation with bromine water (Kanto Chemical

Co., Inc., bromine content ratio: ~1%) or an aqueous 3.5 mol dm3 sodium iodide (NaI) solution dissolving 1 wt% iodine for the positive electrode (10 mm f, ~10 mg) before cell assembly. A negative electrode (10 mm f, ~10 mg) without the treatment and a glassfiber filter paper as a separator were concurrently immersed in the following electrolyte for 1 h under a reduced pressure. Employed electrolytes of the bromide redox and iodide redox systems were an aqueous 3.5 mol dm3 sodium bromide (NaBr) solution and an aqueous 3.5 mol dm3 NaI solution, respectively. Two types of HC cells were used in this study: (i) for evaluating polarization characteristics of positive and negative electrodes, a three-electrode asymmetric cell fabricated with the treated electrode as positive, a non-treated electrode as negative electrodes, and an Ag/AgCl reference electrode, (ii) a two-electrode asymmetric cell with the treated and non-treated electrodes was employed for usual electrochemical measurements. For comparison, we also assembled another three-electrode type cell containing a pair of electrodes without treatment and a reference electrode, as well as another two-electrode symmetric cell with non-treated positive and negative electrodes together with an aqueous 3.5 mol dm3 NaBr solution or an aqueous 1.75 mol dm3 H2SO4 solution. A typical chargeedischarge cycling test was carried out in a cell voltage range between 0 or 0.5 and 1.0 V at a current density of 1000 mA g1. The discharge rate capability of the cells was evaluated in a constant-current (CC) mode at various discharge current densities between 100 and 10,000 mA g1, immediately after their charging to 1.0 V in a CC mode at 100 mA g1. The leakage current test at a float voltage of 1.0 V was performed by using the cells charged to 1.0 V. The withstand voltage of the cells was determined by monitoring chargeedischarge curves at 1000 mA g1, with increasing the upper limit of charging voltage from 1.0 to 1.9 V with a 0.1 V-step every 3 cycles. All electrochemical measurements were performed with a Solartron model 1470E multi-stat electrochemical measurement unit at room temperature of 25  C. Before the present electrochemical measurements, 20-cycle formation was carried out at a current density of 100 mA g1. 3. Results and discussion Fig. 1(a) shows typical chargeedischarge profile at a current density of 1000 mA g1 for the HC cell utilizing the redox reactions of bromide species; this voltage corresponds to the potential difference between negative and positive electrodes. The present HC cell has the initial voltage of ca. 0.5 V because it consists of two different electrodes: the bromine water-treated positive electrode and the non-treated negative electrode. The discharge curve of the HC cell utilizing redox reactions of bromide species exhibits a negligible switching IR loss, which is similar to that of a symmetric EDLC cell with non-treated electrodes even at a high current density of 1000 mA g1. It is noteworthy that the chargeedischarge duration of the HC cell is about 2.6 times longer than that of a conventional EDLC cell. Moreover, the chargeedischarge coulombic efficiency of the HC cell at this current density is above 99.8%. This suggests that the HC cell can provide both high energy and good cycle performance. To understand each electrode behavior during cycling indicated in Fig. 1(a), the potential variation of both the negative and positive electrodes vs. an Ag/AgCl reference was observed and is shown in Fig. 1(b). An open-circuit potential (OCP) of the positive electrode in the HC cell is ~0.75 V vs. Ag/AgCl reference. This suggests that the bromine-treated electrode already has increased positive charges inside the electrode because of the presence of Br redox species as Br2 or Br 3 in the electrode micro-pores [34]. Throughout charging and discharging, the treated positive electrode keeps a stable

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7

80

Discharge capacitance / F g -1

1

Voltage / V

0.8 0.6 0.4 0.2

(a) 0

0

30

60

90

120

70 60 50 40 30 20 10 0

150

Time / second

0

2000

4000

6000

8000

4

1 10

-1

Current density mA g

0.8

Potential / V vs. Ag/AgCl

3

Fig. 2. Discharge capacitance as a function of current density for asymmetric HC cell utilizing redox reactions of bromide species with aqueous 3.5 mol dm3 NaBr electrolyte (C), symmetric EDLC cells containing non-treated electrodes with aqueous 3.5 mol dm3 NaBr electrolyte (B) and aqueous 1.75 mol dm3 H2SO4 electrolyte (:).

0.6 0.4 0.2 0 -0.2 -0.4 -0.6

(b) 0

30

60

90

120

150

Time / second Fig. 1. Voltage and potential vs. time curves during galvanostatic cycles of asymmetric HC cell containing bromine-treated positive electrode (ee) and symmetric EDLC cell containing non-treated electrodes (




) with aqueous 3.5 mol dm3 NaBr electrolyte at 1000 mA g1; (a) voltage between positive and negative electrodes, (b) each potential of negative and positive electrodes vs. Ag/AgCl reference.

potential at around ~0.75 vs. Ag/AgCl reference. This means that the positive electrode has much higher capacity based on additional redox storage of bromide ion species when compared to the negative electrode. The total voltage variation of the capacitor, therefore, mainly comes from the change in the negative electrode potential. The level profile of the treated positive electrode should be ascribed to the contribution of Br2 in bromine water; a reversible redox process involving Br species occurs in the electrode micropores. Fig. 2 shows the capacitance variation of the present HC and the symmetric EDLC cells containing an aqueous 3.5 mol dm3 NaBr solution and another symmetric EDLC cell containing an aqueous 1.75 mol dm3 H2SO4 solution as a function of current density. The discharge cell capacitance specified as gravimetric value of both electrodes (Ccell) is calculated from the following equation: Ccell ¼ (I  t)/(V  w)

(2)

where I is the current value, t the discharge time, V the applied cell

voltage, and w is the mass of two electrodes. The discharge capacitance of the HC utilizing redox reactions of bromide species is much higher than that of the other EDLC cells. For example, the discharge capacitance of the HC cell utilizing redox reactions of bromide species and the symmetric EDLC cells with an aqueous 3.5 mol dm3 NaBr and an aqueous 1.75 mol dm3 H2SO4 at 500 mA g1 are 74.0, 29.1, and 51.3 F g1, respectively. Furthermore, the discharge capacitance of the HC cell utilizing redox reactions of bromide species has excellent rate capability; the present HC and symmetric EDLC cells with an aqueous 3.5 mol dm3 NaBr and an aqueous 1.75 mol dm3 H2SO4 at 10,000 mA g1 are 62.8, 14.7 and 42.4 F g1, respectively. This means that the asymmetric HC utilizing redox reactions of bromide species indicates excellent highrate discharge ability exceeding that of the other EDLC systems, suggesting that redox reactions of bromine species are fast enough. In addition, one may realize that in Fig. 2 some fluctuations are visible in the discharge capacitance, especially for the cell containing the pretreated positive electrode. Although this reason has not been clear, a possible origin is as follows: we applied a cutoff voltage to terminate cell charging. Usual capacitors are well controlled by this conventional method. However, as the present positive with the pretreatment has a quite flat profile, the total charging electricity is subject to a slight potential change of the positive at each cycle. Although the present HC cell has excellent discharge ability, there is still a critical question whether this system inhibits the diffusion of storage species at the positive into an electrolyte or not. Judging from the high coulombic efficiency, the present HC cell suppresses undesirable diffusion of reactive species at least in short duration. To confirm suppressing diffusion of reactive species with long duration, the leakage current of the asymmetric HC cell utilizing redox reactions of bromide species was measured at 1.0 V floating as shown in Fig. 3. Generally, leakage current is composed of two factors: the initial ionic current and an electronic current [35,36]. The ionic current is mainly due to delayed charging to noncharged parts of the electrode, which correspond to the highimpedance parts of electrode. This current portion should decrease in a short period. The electronic current is usually ascribed to minor redox reactions of electrolyte components as faradic

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

4

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7

1 0.8

1

Voltage / V

Leakage current / mA

10

-3 EDLC w/odm Br-treatment, H2SO4 1.75 mol H2SO4 solution

0.1

0.6 0.4

EC NaBr cell HC w/aymmetric Br-treatment,

0.2

(a) symmetric cell EDLC w/o EC Br-treatment, NaBr

0.01

0

5

0

10

15

60

90

120

150

0.8 0.6

Potential / V vs. Ag/AgCl

current. This current portion tends to remain in a relatively long period. The leakage current for the asymmetric HC cell with the treated electrode is somewhat higher than that for the symmetric EDLC with the non-treated electrodes and an aqueous 3.5 mol dm3 NaBr in the initial period. However, this HC cell with the treated electrodes shows lower currents than the EDLC with the non-treated electrodes and an aqueous 1.75 mol dm3 H2SO4 over 15 h. In other words, a current due to leak processes for the HC cell with treated electrodes is smaller even than that for a conventional EDLC containing an aqueous 1.75 mol dm3 H2SO4. The ratio of ionic current to overall current would decrease with duration while the electric current ratio would increase. Thus the contribution of electric currents for some remaining redox reactions is the almost same among the tested EC cells. This means that the oxidation species, Br 3 and Br2, never diffuse from the electrode micro-pores to the counter electrode. We also verified the negligible amounts of diffused Br 3 and Br2 in the electrolyte by an UVevis spectrometric measurement. We also tried to apply iodide to electrolyte redox species because iodide redox reactions proceed in similar mechanisms to bromide redox reactions. Fig. 4(a) and (b) show the chargeedischarge profiles and each electrode potential variation of the negative and positive electrodes vs. Ag/AgCl reference at a current density of 1000 mA g1 for the HC cell utilizing iodide species, respectively, as a comparison with the bromine-utilized HC cell. The chargeedischarge time of the HC utilizing iodide redox reactions is somewhat shorter than that utilizing bromide redox reactions [Fig. 4(a)]. Nevertheless, the HC cell utilizing iodide redox also shows a negligible IR drop like the HC with bromide redox as well as the EDLC [Fig. 1(a)] even at a high current density of 1000 mA g1. The iodide system has a good coulombic efficiency of 99.9%. An OCP of the positive electrode in the HC cell with the iodide system is ~0.22 V vs. Ag/AgCl reference as shown in Fig. 4(b). This suggests that the iodine-treated positive electrode obtains positive charges based on iodide redox species as I2 or I 3 in the electrode micropores like the bromide system [31]. Our strategy of applying the pretreatment is to stabilize a positive electrode potential and keep the potential constant during

30

Time / second

Time / hour Fig. 3. Leakage current observed in asymmetric HC cell with bromine-treated electrode and symmetric EDLC cell containing non-treated electrodes with aqueous 3.5 mol dm3 NaBr electrolyte, and symmetric EDLC cell containing non-treated electrodes with aqueous 1.75 mol dm3 H2SO4 electrolyte at floating voltage of 1.0 V.

0

0.4 0.2 0 -0.2 -0.4 -0.6 -0.8

(b) 0

30

60

90

120

150

Time / second Fig. 4. Voltage and potential vs. time curves during galvanostatic cycles of asymmetric HC cells containing bromine-treated electrode with aqueous 3.5 mol dm3 NaBr electrolyte (ee) and iodine-treated electrode with aqueous 3.5 mol dm3 NaI electrolyte (———) at 1000 mA g1; (a) voltage between positive and negative electrodes, (b) each potential of negative and positive electrodes vs. Ag/AgCl reference.

whole cycling. If do not apply this, the positive potential becomes unstable, especially at the initial cycle, because the pristine positive electrode is intrinsically the same just as the negative electrode. This means that the positive-electrode working potential should significantly drift and the influenced negative should do so as well during cycling especially at the initial few cycles without the pretreatment, because what we can do is just to control the terminal voltage between the positive and negative as realistic cell operation. Thus, the present pretreatment is quite beneficial to stabilizing realistic cell performance. Table 1 summarizes discharge capacitances with various current densities for the asymmetric HC cells with the bromine- and iodine-treated positive electrodes and the symmetric EDLC cells with the non-treated electrodes. The HC cell utilizing bromide redox has not only the highest discharge capacitance but also excellent rate capability. The discharge capacitance of the HC cell with iodide at 100 mA g1 is the almost same as the EDLC cell with a H2SO4 solution at the same current density. Moreover, the cell with the iodide system has good rate capability, suggesting that the electrolyte redox reaction of iodide is also rapid like the bromide

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7

5

Table 1 Discharge capacitance (F g1) of asymmetric HC cells containing bromine-treated electrode with aqueous 3.5 mol dm3 NaBr electrolyte and iodine-treated electrode with aqueous 3.5 mol dm3 NaI electrolyte, and symmetric EDLC cells containing non-treated electrodes with aqueous 1.75 mol dm3 H2SO4 and aqueous 3.5 mol dm3 NaBr electrolytes. Current density (mA g1)

Asymmetric HC system Bromide redox

Iodide redox

NaBr solution

NaI solution

72.0 74.0 74.0 62.8

57.5 58.2 58.0 46.2

system. Fig. 5 shows the discharge capacitance during extended cycling for the asymmetric HC cells utilizing redox reactions of bromide and iodide species, and the symmetric EDLC containing an aqueous 1.75 mol dm3 H2SO4 at a current density of 5000 mA g1. Throughout this long-term cycle tests, the asymmetric HC cells with bromine- and iodine-treated positive electrodes exhibit much higher discharge capacitance than the symmetric EDLC cell with a H2SO4 electrolyte. The discharge capacitances for the asymmetric HC cells utilizing bromine- and iodine-treated electrode retain 90.4% and 94.9% of their initial values even at the 100,000th cycle, while that for the EDLC cell continuously decreases to 64.1% at the same cycle. These results suggest that the present asymmetric HC can maintain essential properties as they are, even after long-term cycles, due to suppressing undesirable diffusion of reactive storage species. The positive electrode of the HC cell utilizing bromide redox not only operates at a relatively high potential of ~0.75 V, but also has good cycleability. On the other hand, the negative electrode of the same cell operates in a wide, stable potential region [Fig. 1(b)]. Therefore, the HC cell utilizing bromide redox can work with higher cell voltage if one operates the negative electrode to more negative potential. To know a maximum withstand voltage for the HC utilizing redox reactions of bromide species, we observed its chargeedischarge curves at 1000 mA g1, when the upper limit of

H2SO4 solution

NaBr solution

54.0 50.0 49.7 42.4

28.7 29.0 28.6 24.7

charging voltage was increased from 1.0 to 1.9 V with a 0.1 V-step every 3 cycles. Fig. 6(a) shows the chargeedischarge curves of the 3rd cycle in each voltage step for the asymmetric HC. Fig. 6(b) indicates each electrode potential of the positive and negative during the HC operation with increasing cell voltage corresponding to Fig. 6(a). The chargeedischarge curves show excellent reversibility up to a cutoff voltage of 1.8 V. In fact, the coulombic efficiency with

2

1.6

Voltage / V

100 500 1000 10,000

Symmetric EDLC system

1.2

0.8

0.4

(a) 0

0

100

300

400

500

Time / second

70 1

60 50

Potential / V vs. Ag/AgCl

Discharge capacitance / F g -1

200

40 30 20 10 0 0

4

2 10

4

4 10

6 10

4

4

8 10

5

1 10

Cycle number Fig. 5. Discharge capacitance as a function of cycle number for asymmetric HC cell containing bromine-treated electrode with aqueous 3.5 mol dm3 NaBr electrolyte (C), and iodine-treated electrode with aqueous 3.5 mol dm3 NaI electrolyte (A), and symmetric EDLC cell containing non-treated electrodes with aqueous 1.75 mol dm3 H2SO4 electrolyte (:). Chargeedischarge current density: 5000 mA g1, operation voltage: 0e1.0 V.

0.5

0

-0.5

(b)

-1 0

100

200

300

400

500

Time / second Fig. 6. Voltage and potential vs. time curves during withstand voltage test for asymmetric HC cell containing treated electrode with aqueous 3.5 mol dm3 NaBr electrolyte at 1000 mA g1; (a) voltage between positive and negative electrodes, (b) each potential of negative and positive electrodes vs. Ag/AgCl reference.

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

6

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7

cutoff voltages of 1.0e1.8 V is over 98%. The highest cutoff voltage of 1.9 V eventually becomes irreversible (broken curves in Fig. 6); the coulombic efficiency is under 60% probably because of hydrogen generation. Judging from Fig. 6(b), the chargeedischarge curves of the negative electrode with cutoff voltages until 1.8 V are essentially linear and have no obvious plateau positions. However, the charge curve of the negative electrode with the highest cutoff voltage of 1.9 V has a plateau region. This suggests that the highest voltage of the asymmetric HC cell utilizing bromide redox is essentially limited by the negative electrode behavior, because the positive electrode potential is always constant at about 0.8 V irrespective of the cell voltage variation thanks to stable bromine redox processes. The present cell operation demonstrates that the asymmetric HC cell utilizing bromide redox can provide high power density as well as high energy density because a high-voltage operation of 1.8 V is possible in spite of an aqueous electrolyte system. This high voltage provides excellent energy density, which is ca. 5 times higher than that of the present conventional aqueous EDLC with H2SO4. The discharge capacitance for the 1.8 V-based HC cell retains ca. 80% of the initial value at the 100,000th cycle. Fig. 7 shows the Ragone plots (the relationship between energy densities and power densities based on the total mass of both positive and negative electrodes) for the asymmetric HC cell utilizing bromide redox reactions with cutoff charging voltages of 1.0 and 1.8 V. For important comparison, we also show the Ragone plots of the 2.5 and 2.7 V-class non-aqueous EDLC with the same ACFCs and a conventional organic electrolyte: triethylmethylammonium tetrafluoroborate (TEMABF4) dissolved in propylene carbonate (PC). The present HC cell with 1.8 V operation shows outstanding performance; its energy density at the lowest power density corresponding to a current density of 100 mA g1 is about 1.2 times higher than that of a 2.7 V-class non-aqueous EDLC in spite of an aqueous electrolyte system. Moreover, the present HC provides not only higher energy density but also much higher power density when compared to the conventional non-aqueous EDLC. Such excellent and stable performance of the present HC systems demonstrates a promising alternative approach in the development of ECs.

Energy density / W h kg-1

100 Br-in-aqueous (1.8V)

10

Br-in-aqueous (1.0V)

1

0.1 0.01

0.1

1

10

100

Power density / kW kg-1 Fig. 7. Ragone plots (relationship between energy density and power density) for asymmetric HC cell utilizing bromide redox reactions with aqueous 3.5 mol dm3 NaBr electrolyte, operation voltage: 1.8 (C) and 1.0 V (B), and for conventional nonaqueous EDLC with 1.96 mol dm3 TEMABF4/PC, operation voltage: 2.7 (-) and 2.5 V ( ).

▫

4. Conclusions We demonstrated novel HCs with chargeedischarge mechanisms based on an electrolyte-involved charge storage system, which utilizes redox reactions of bromide or iodide species. The pre-treatment is simple: bromine or iodine impregnation into pores at a positive electrode before cell assembly. The results reveal that the HC cells utilizing redox reactions of bromide species or iodide provide excellent performance as follows. (i) The discharge capacitance of the asymmetric HC cell with the bromide redox species is much higher than that of conventional non-treated EDLC systems. Moreover, the present HCs indicate excellent high-rate discharge ability, because redox reactions of bromide and iodide species are essentially fast. (ii) The discharge capacitance of the present asymmetric HC cells can retain over 90% of their initial value even after 100,000 cycles. (iii) In spite of an aqueous electrolyte system, the HC utilizing bromide redox realizes high cutoff voltage operation of 0e1.8 V. Therefore, the HC cell with redox reactions of bromide species can provide high energy density, high power as well as excellent cycle ability. Acknowledgements This work was financially supported by Grant-in-Aid for Scientific Research from MEXT (the Ministry of Education, Culture Sports, Science and Technology of Japan) No. 26288112 and by the Kansai University Grant-in-Aid for progress of research in graduate course 2014. References [1] B.E. Conway, Electrochemical Supercapacitors: Scientific Fundaments and Technological Applications, Kluwer Academic/Plenum Publishers, New York, 1999. [2] A. Nishino, J. Power Sources 60 (1996) 137. [3] A. Burke, J. Power Sources 91 (2000) 37. [4] R.J. Brodd, K.R. Bulluch, R.A. Leising, R.L. Middaugh, J.R. Miller, E. Takeuchi, J. Electrochem. Soc. 3 (2004) K1. [5] I. Tanahashi, A. Yoshida, A. Nishino, Bull. Chem. Soc. Jpn. 63 (1990) 3611.  ero, F. Be guin, J. Power Sources 153 (2004) [6] V. Khomenko, E. Raymudo-Pinn 183. [7] S.A. Kazaryan, S.N. Razumov, S.V. Litvinenko, G.G. Kharisov, V.I. Kogan, J. Electrochem. Soc. 153 (2006) A1655. [8] A. Balducci, W.A. Henderson, M. Mastragostino, S. Passerini, P. Simon, F. Soavi, Electrochim. Acta 50 (2005) 2233. [9] A.S. Arico, P. Bruce, B. Scrosati, J.M. Tarascon, W.V. Schalkwijk, Nat. Mater 4 (2005) 336. [10] H.A. Andreas, B.E. Conway, Electrochim. Acta 51 (2006) 6510. [11] J.R. Miller, P. Simon, ECS Interface 17 (2008) 31. [12] S. Yamazaki, K. Obata, Y. Okuhama, Y. Matsuda, M. Yamagata, M. Ishikawa, J. Power Sources 195 (2010) 1753. [13] Y. Honda, M. Takeshige, H. Shiozaki, T. Kitamura, M. Ishikawa, J. Electrochem. Soc. 155 (2008) A930. [14] J. Huang, B.G. Sumpter, V. Meunier, Angew. Chem. Int. Ed. 47 (2008) 520. [15] B. Liu, H. Shioyama, T. Akita, Q. Xu, J. Am. Chem. Soc. 130 (2008) 5390. denas, A. Linares-Sol[16] P. Azaïs, L. Duclaux, P. Florian, D. Massiot, M.A. Lillo-Ro guin, J. Power Sources 171 (2007) 1046. ano, J.P. Peres, C. Jehoulet, F. Be [17] J. Chmiola, G. Yushin, Y. Gogotsi, C. Portet, P. Simon, P.L. Taberna, Science 313 (2006) 1760. [18] S. Trasatti, G. Buzzanca, J. Electroanal. Chem. 29 (1971) 1. [19] W. Sugimoto, H. Iwata, Y. Yasunaga, Y. Murakami, Y. Takasu, Angew. Chem. Int. Ed. 42 (2003) 4092. [20] C.-C. Hu, T.-Y. Hsu, Electrochim. Acta 53 (2008) 2386. [21] K. Naoi, S. Ishimoto, N. Ogihara, Y. Nakagawa, S. Hatta, J. Electrochem. Soc. 156 (2009) A52. [22] A. Rudge, I. Raistrick, S. Gottesfeld, J.P. Ferraris, Electrochim. Acta 39 (1994) 273. [23] K. Naoi, S. Suematsu, M. Hanada, H. Takenouchi, J. Electrochem. Soc. 149 (2002) A472. [24] M. Mastragostino, C. Arbizzani, F. Soavi, Solid State Ionics 148 (2002) 493.  ero, Y.H. Lee, F. Be guin, Carbon 47 (2009) 2984. [25] E.J. Ra, E. Raymundo-Pin [26] J.R. Miller, P. Simon, Science 321 (2008) 651. [27] T. Shimooka, S. Yamazaki, T. Sugimoto, T. Jyozuka, H. Teraishi, Y. Nagao, H. Oda, Y. Matsuda, M. Ishikawa, Electrochemistry 75 (2007) 273. [28] S. Yamazaki, T. Ito, M. Yamagata, M. Ishikawa, 218th Electrochem. Soc.

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021

S. Yamazaki et al. / Journal of Power Sources xxx (2016) 1e7 Meeting Abstract #312, Las Vegas, 2010. [29] S. Yamazaki, T. Ito, M. Yamagata, M. Ishikawa, ECS Trans. 41 (2012) 15. [30] S. Yamazaki, T. Ito, M. Yamagata, M. Ishikawa, Electrochim. Acta 86 (2012) 294. [31] G. Lota, E. Frackowiak, Electrochem. Commun. 11 (2009) 87. [32] G. Lota, K. Fic, E. Frackwiak, Electrochem. Commun. 12 (2011) 38. [33] K. Fic, M. Meller, E. Frackwiak, J. Electrochem. Soc. 162 (2015) A5140.

7

[34] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE, Houston, 1966. [35] M. Ishikawa, M. Morita, M. Ihara, Y. Matsuda, J. Electrochem. Soc. 141 (1994) 1730. [36] S. Yamazaki, A. Takegawa, Y. Kaneko, J. Kadokawa, M. Yamagata, M. Ishikawa, J. Electrochem. Soc. 157 (2010) A203.

Please cite this article in press as: S. Yamazaki, et al., Journal of Power Sources (2016), http://dx.doi.org/10.1016/j.jpowsour.2016.04.021