Journal of The Electrochemical Society, 148 共10兲 A1130-A1134 共2001兲

A1130

0013-4651/2001/148共10兲/A1130/5/$7.00 © The Electrochemical Society, Inc.

Hybrid Supercapacitors Based on Activated Carbons and Conducting Polymers A. Laforgue,a,*, z P. Simon,a,* J. F. Fauvarque,a,** J. F. Sarrau,b and P. Laillerb a

CNAM, Laboratoire d’Electrochimie Industrielle, 75003 Paris, France CEAC-Exide, Research and Development Service, 92636 Gennevilliers Cedex, France

b

A new concept of hybrid supercapacitors has been designed with a conductive polymer as the positive electrode and activated carbon as the negative electrode. The chosen polymer was poly共4-fluorophenyl-3-thiophene兲 or P-4-FPT, whose good doping properties are well known. Several activated carbons were tested, and the results reported here were obtained with carbons from Spectracorp 共BP25 and YP17兲. The mass ratio of active materials was calculated to obtain the maximum cell voltage 共3 V兲. The system studied in our laboratory was 4 cm2 test cells which reached 48 Wh/kg of maximum energy associated with 9 kW/kg of maximum power 共considering the mass of the active materials兲. Industrial prismatic prototypes were then assembled, containing 60 cm2 electrodes. These prototypes reached a maximum energy of 7.5 Wh/kg 共total mass兲, and a maximum power of 250 W/kg. © 2001 The Electrochemical Society. 关DOI: 10.1149/1.1400742兴 All rights reserved. Manuscript submitted March 6, 2001; revised manuscript received May 28, 2001. Available electronically September 4, 2001.

Electrochemical capacitors 共also named supercapacitors or ultracapacitors兲 are nowadays among the most promising energy storage devices, due to their ability to deliver at the same time both high energy and high power.1,2 Carbon/carbon and metal oxide devices are the only ones currently produced on an industrial scale. Polymerbased supercapacitors have been intensively studied these last few years and have demonstrated significantly higher performances than carbon-based ones.3-10 Nevertheless, all these systems are assembled with polymers electrodeposited under very pure and soft conditions. Industrial production is then difficult to achieve. The use of composite electrodes in polymer-based devices could provide easier processing development. For this reason, the chemical synthesis of several polythiophenes and their processing into composite electrodes for energy storage devices were proposed in recent papers.11,12 The positive doping 共p-doping兲 processes of the chemical polymers were found to be as effective as for the electrochemical ones in terms of capacity, efficiency, and electrochemical stability. However, the negative doping processes of the polymers were not found to be stable enough to be used in energy storage devices. It was then decided to substitute the polymer with an activated carbon on the negative electrode. This system is then a new hybrid concept of supercapacitor. The positive electrode was composed of poly-4fluorophenylthiophene 共P-4-FPT兲, and several activated carbons were tested as negative electrodes. We report here the results obtained with carbons from Spectracorp, BP25 共130 farad/g in organic media兲 and YP17 共110 F/g in organic media兲. Experimental Cyclic voltammetry experiments were performed using an EG&G PARC model 273A potentiostat controlled by a computer using the software M270 共version 4.0兲. A three-electrode cell was used, with a platinum foil as the counter electrode and a Ag⫹/Ag electrode as the reference electrode 共consisting of a Ag wire in a solution of 10⫺2 M AgNO3 in acetonitrile兲. The potential of this reference electrode was found to be ⫹0.3 V/SCE in organic media. Galvanostatic cycling tests were carried out with a VMP potentiostat 共Biologic Technologies兲 able to record one point every 20 ms and allowing for control of the total cell voltage limits while following the potential of each electrode against the reference electrode. The monomer was synthesized by a coupling reaction of 3-bromothiophene and 4-fluorophenylmagnesium bromide in tetrahydrofuran 共THF兲 with NiCl2 共diphenylphosphinopropane兲 as catalyst.11,13-15 The yield for this reaction was more than 90% by weight.

* Electrochemical Society Student Member. ** Electrochemical Society Active Member. z

E-mail: [email protected]

The monomer 4-FPT was polymerized by a direct oxidation with FeCl3 as oxidant in chloroform or dichloromethane. This method has already been used to polymerize thiophene derivatives.11,16-21 The yield obtained for this reaction was more than 80% by weight. The polymer powder was then washed several times with methanol to remove residual chloride ions. Active materials were processed into composite electrodes in order to be tested. The polymer powder was mixed with acetylene black 共AB, electronic conductor兲 to ensure a good electronic percolation inside the electrode.11 The activated carbon powders were mixed with graphite 共SFG44 from Lonza兲 to obtain good electrode conduction. Then, the active materials were bound on stainless steel AISI 316L expanded current collectors with a binder developed by this laboratory, composed of 3 wt % carboxymethylcellulose 共CMC兲 and 2 wt % polytetrafluoroethylene 共PTFE兲. This binder allowed good adherence of the active materials on the current collector and good electrode accessibility to the dopant ions. Moreover, this binder allowed lamination of the active material pastes on the expanded current collectors, a process easily adaptable for industrial production. Results and Discussion Electrode optimization.—An electrode composition study was carried out with the polymer electrode. Figure 1 shows the cyclic voltamograms obtained with P-4-FPT electrodes bound with 3% CMC and 2% PTFE and different polymer contents in the electrode. Figure 2 shows the charges for doping and dedoping and the coulombic efficiency, Q discharged /Q charged , for different electrode compositions, calculated from the curves in Fig. 1. The polymer can be fully doped only when a high content of AB is present in the electrode. The capacity then reaches 40 mAh/g and 97% coulombic efficiency. When the AB content is decreased, so does the electronic percolation in the electrode, and some of the polymer does not participate in the doping process. However, for an AB content of 15% 共and 80% of polymer兲, the capacity for dedoping is still 30 mAh/g. When the AB content is decreased below this value, the capacity and the coulombic efficiency fall quickly, indicating a lack of electronic percolation in the electrode. To analyze the complete polymer doping, the characterization needs good electronic percolation. This will then be achieved using electrodes containing 30% of polymer and 65% of AB. However, to obtain the maximum capacity in a device, the dead mass brought by the conducting additive must be taken into account. Figure 2 also presents the doped and dedoped capacities 共from Fig. 1兲 balanced by the total mass of active materials (polymer ⫹ conductor ⫹ binders). The best compromise between the absolute capacity obtained from the polymer and the active material’s total capacity is an elec-

Journal of The Electrochemical Society, 148 共10兲 A1130-A1134 共2001兲

Figure 1. Voltamograms representing the positive doping of P-4-FPT for different electrode compositions: electrolyte 1 M NEt4 CF3 SO3 ; scan rate ⫽ 5 mV/s, m P-4-FPT ⫽ 25 mg. Electrode composition: X% P-4-FPT, 共95-X%兲 AB, 3% CMC, and 2% PTFE; SS AISI 316L current collectors 共4 cm2 兲.

A1131

Figure 3. Galvanostatic cycling curve of a positive P-4-FPT electrode in an electrolyte 1 M NEt4 CF3 SO3 in acetonitrile; m P-4-FPT ⫽ 15.8 mg; i ⫽ ⫾2 mA/cm2 .

The positive doping process starts at 0.5 V/Ref, and the potential linearly increases up to 0.9 V/Ref. The slope of the V(t) curve allows calculation of the pseudocapacitance during charging, 290 F/g. The dedoping process of the polymer extends from 0.9 to 0.5 V/Ref, where the end-of-discharge signal can be seen. A linear discharge is observed between 0.8 and 0.5 V/Ref. The pseudocapacitance during discharge, calculated from the slope of the V(t) curve, reaches 260 F/g. Charged and discharged capacities are, respectively, 31 and 28 mAh/g of P-4-FPT, which corresponds to a coulombic efficiency of 92%. The pseudocapacitance of the polymer is significantly higher than that of the best activated carbons, which develop 130 to 140 F/g in organic media. There is then a real advantage to the use of conducting polymers in supercapacitors to increase the energy density of these systems. Supercapacitor cells.—The principle of activated carbon/ conductive polymer hybrid supercapacitors is described in Fig. 4. When the system is charged, the polymer of the positive electrode is Figure 2. Doping and dedoping capacities and coulombic efficiencies from the voltamograms of Fig. 1. The inset represents the doped capacity divided by the total mass of the active materials 共polymer ⫹ conductor ⫹ binders兲.

trode composed of 80% polymer, 15% AB, 3% CMC, and 2% PTFE. Cyclic voltammetry of the positive doping of P-4-FPT.— From the 30% polymer curve in Fig. 1, it can be seen that positive doping of the polymer begins at 0.3 V/Ref with a peak at 0.78 V/Ref. On the reverse scan, the dedoping peak can be seen at 0.55 V/Ref. The coulombic efficiency reaches 97% on the first cycle. The charged capacity is around 40 mAh/g, which corresponds to the one generally observed with ECPs. Galvanostatic cycling tests.—The positive doping of the P-4FPT was also studied by galvanostatic cycling in a three-electrode cell, with an overcapacitive activated carbon electrode as counter electrode. The 4 cm2 polymer electrode composition was still 30% P-4-FPT, 65% electronic conductor, 3% CMC, and 2% PTFE. Figure 3 presents the cycling curves of the P-4-FPT at ⫾2 mA/cm2 between 0.15 and 0.9 V/Ref.

Figure 4. Principle of an activated carbon/conducting polymer hybrid supercapacitor.

A1132

Journal of The Electrochemical Society, 148 共10兲 A1130-A1134 共2001兲

Figure 5. Schematic description of the two-electrode 4 cm2 test cells.

p-doped while a double-layer capacitance is charged at the surface of the negative electrode. A schematic description of the cell is prsented in Fig. 5. The active materials laminated on the expanded current collectors were placed between two PTFE plates and separated by two PTFE sheets 共Gore, each 25 ␮m thick兲. The system was then kept under constant pressure with the help of two stainless steel clamps. The reference electrode was placed between the electrodes, above the system, a hole in the PTFE plates allowing this electrode to be very close to the active materials. The cell was then placed in a box ensuring air tightness. The positive electrode was composed of 80% P-4-FPT, 15% AB, and 5% binders. The negative electrode composition was 80% activated carbon BP25, 15% graphite SFG44, and 5% binders. The electrolyte was a solution of 1 M NEt4 BF4 in propylene carbonate, according to the results described by Mastragostino et al.9 To make the system work properly, the polymer electrode has to cycle precisely in the p-doping domain of the polymer. The active mass ratio 共R ⫽ m polym /m carbon兲 was determined to be 1.6 to fit this condition and at the same time to lower the negative electrode working potential, in order to improve the total system voltage. Figure 6 presents the galvanostatic cycling curves of the positive electrode 共P-4-FPT兲, the negative electrode 共BP25兲 and the total cell voltage of a hybrid supercapacitor assembled in a 4 cm2 test cell. The positive electrode is kept in the p-doping potential of the P-4-FPT, between 0.4 and 0.9 V/Ref. At the same time, the negative electrode cycles from ⫺0.8 and ⫺2.1 V/Ref. The total system cycles then between 1.2 and 3 V. The P-4-FPT capacity was 28 mAh/g at the beginning of cycling, and its pseudocapacitance was 245 F/g which means that the in-

Figure 7. Picture of a prismatic hybrid supercapacitor prototype. Dimensions of the prototype 共mm兲: 59 ⫻ 26 ⫻ 165; V ⫽ 0.25 dm3 .

crease of polymer content in the electrode from 30 to 80% causes no loss in the capacity and only 10% loss in the pseudocapacitance. The capacitance of the carbon was 130 F/g. The internal cell resistance measured at 50 Hz was 12.5 ⍀ cm2 . The performance of the system was also measured in terms of maximum energy and maximum power, calculated with the following formula 2 E max ⫽ 1/2共 CU max 兲

共in W s兲 where C is the system capacity 共in farad兲 and U max the potential at the end of charge P max ⫽ U 20 /4R

Figure 6. Galvanostatic cycling of a 4 cm2 hybrid supercapacitor cell in an electrolyte 1 M NEt4 BF4 in PC; R ⫽ 1.6; m P-4-FPT ⫽ 75.3 mg; positive electrode: 80% P-4-FPT, 15% NA, 5% binders; negative electrode: 80% BP25, 15% graphite SFG44, 5% binders; SS AISI 316L current collector; i ⫽ ⫾5 mA/cm2 .

共in W兲 where U 0 is the potential at the beginning of discharge 共after the ohmic drop兲 and R the internal resistance 共in ⍀兲. It represents the instantaneous power 共at the beginning of discharge兲. Considering the active masses 共polymer and activated carbon兲, the hybrid system delivered a maximum energy of 48 Wh/kg associated with a maximum power of 9 kW/kg. These performances are equivalent to those obtained with type III polymer-based supercapacitors 共30-50 Wh/kg and 3-10 kW/kg兲.22,23

Journal of The Electrochemical Society, 148 共10兲 A1130-A1134 共2001兲

A1133

Table I. The characteristics and performance of four test cells. Prototype Number of electrodes Current collector Separator 共PTFE兲, ␮m Positive active material Negative active material Ratio R (m ⫹ /m ⫺) Polymer weight 共g兲 Capacity, F Resistance 共1000 Hz兲, m⍀ Surface resistance, ⍀ cm2 Packaged weight 共g兲a Maximum energy, Wh Maximum power at 3 V, W Specific energy, Wh/kg Specific power, Wh/kg a

SC-01 2 ⫻ 11 SS AISI 316L 2 ⫻ 25 P-4-FPT 80% YP17 90% 1.68 41.9 2000 47 59 370.9 2.50 47.9 6.74 129.0

SC-02 2 ⫻ 24 SS AISI 316L 2 ⫻ 25 P-4-FPT 80% BP25 80% 1.7 40.1 2500 25 71 448.5 3.13 90 6.97 200.7

SC-03 2 ⫻ 25 SS AISI 316L 2 ⫻ 25 P-4-FPT 80% BP25 80% 1.58 37.6 2600 23 68 434.2 3.25 97.8 7.49 225.3

SC-04 2 ⫻ 24 SS AISI 316L 2 ⫻ 25 P-4-FPT 80% BP25 80% 1.65 39.7 2600 20 56 430.7 3.25 112.5 7.55 261.2

Packaged weight: total weight of the prototype 共case, stack, electrolyte, separator, ...兲.

Industrial prototypes.—Large prototypes were assembled in association with the CEAC-Exide Company on the basis of the results obtained on the 4 cm2 cells 共cf. Fig. 7兲. Table I lists the characteristics and performances of four prototypes. The electrode area was 60 cm2 共for each electrode兲. Positive electrodes were made with P-4-FPT, and the negative ones were activated carbon based. The electrolyte was a solution of 1 M NEt4 BF4 in propylene carbonate. Figure 8 shows the galvanostatic cycling curve of a prototype at 5 A between 1 and 3 V. The cell voltage during cycling was 3 V, which means that the mass ratio m polym /m carbon was well adapted regarding the cycling potential range of the positive polymer electrode; the positive/ negative active material ratio was thus validated for a larger scale. Prototypes SC-03 and SC-04 reached capacities as high as 2600 F, corresponding to specific energies around 7.5 Wh/kg 共considering the whole prototype mass兲. These are interesting values, which confirm the results obtained with the 4 cm2 cells. However, the resistances of the devices were high compared to the tests cells, limiting the power of the devices to around 250 W/kg. The increase in resistivity was about five times between the laboratory cells and the industrial prototypes, and was found to be caused by the lack of internal pressure in the prototypes. The adhesion of the active materials on the current collectors 共especially for the negative electrode兲 was then not strong enough to allow good conductivity. During galvanostatic cycling, the prototype capacities decreased, following the polymer deactivation.12,24-27 However, with the prototype SC-01, an interesting cylability test was achieved. Figure 9

presents the change in capacity of the prototype SC-01 during cycling. The initial capacitance is 2000 F but decreases during the first 100 cycles down to 1000 F. After this decrease, the capacity remains constant during the rest of the cycling test. The changes which can be seen around the 1000th cycle are due to charge-discharge profile modifications performed in order to stabilize the capacity variation. This capacity loss still has to be clarified. Meanwhile, the performance of the prototype during the last 8000 cycles from the point of view of the energy is still satisfactory, with a specific energy of 3.3 Wh/kg.

Figure 8. Galvanostatic curve of the prototype SC-04 at 5 A between 1 and 3 V.

Figure 9. Evolution of the capacity of SC-01 during galvanostatic cycling between 1 and 3 V.

Conclusions A new concept of hybrid supercapacitors based on both conducting polymers and activated carbons has been developed and tested with a polythiophene derivative and several activated carbons. The results obtained with 4 cm2 laboratory test cells has been confirmed at larger scale 共industrial prototypes兲 in terms of energy. However, the lack of pressure in the modules limited the power of the devices. To our knowledge, this is the first time that industrial prototypes of supercapacitors using polythiophene derivatives have been realized, reaching 7.5 Wh/kg of initial maximum energy and cycling at 3.3 Wh/kg for 8000 cycles.

A1134

Journal of The Electrochemical Society, 148 共10兲 A1130-A1134 共2001兲

The work will now focus on the limitation of the capacity loss of the P-4-FPT during the first cycles, as well as the decrease of the series resistance by improving the electrode compression inside the module. Our aim is to reach internal resistance values close to the ones obtained with 4 cm2 cells 共around 10 ⍀ cm2 兲, which could enhance the power of the devices to more than 1 kW/kg. Acknowledgments The authors want to thank the European Community for financial support of this work within JOULE III project CT97-0047. Special thanks to the Italian partners included in the SCOPE project: University of Bologna 共Professor M. Mastragostino兲, ENEA’s Department of Energy 共M. Conte兲, and Arcotronics-Machinery Division 共S. Saguatti兲. CNAM assisted in meeting the publication costs of this article.

References 1. 2. 3. 4. 5. 6.

R. Ko¨tz and M. Carlen, Electrochim. Acta, 45, 2483 共2000兲. A. Burke, J. Power Sources, 91, 37 共2000兲. A. Rudge, J. Davey, I. Raistrick, S. Gottesfeld, and J. P. Ferraris, J. Power Sources, 47, 89 共1994兲. J. P. Ferraris, S. Gottesfeld, and A. J. Rudge, U.S. Pat. 5,527,640 共1996兲. J. P. Ferraris, M. M. Eissa, I. D. Brotherston, D. C. Loveday, and A. A. Moxey, J. Electroanal. Chem., 459, 57 共1998兲. H. Sarker, Y. Gofer, J. G. Killian, T. O. Poehler, and P. C. Searson, Synth. Met., 88, 179 共1997兲.

7. J. Giaccai, Y. Gofer, J. G. Killian, T. O. Poehler, H. Sarker, and P. Searson, U.S. Pat. 5,733,683 共1998兲. 8. C. Arbizzani, M. Mastragostino, and L. Meneghello, Electrochim. Acta, 41, 21 共1996兲. 9. M. Mastragostino, C. Arbizzani, R. Paraventi, and A. Zanelli, J. Electrochem. Soc., 147, 407 共2000兲. 10. J. C. Carlberg and O. Ingana¨s, J. Electrochem. Soc., 144, 61 共1997兲. 11. A. Laforgue, P. Simon, C. Sarrazin, and J. F. Fauvarque, J. Power Sources, 80, 142 共1999兲. 12. A. Laforgue, P. Simon, and J. F. Fauvarque, Synth. Met., Accepted. 13. K. Tamao, S. Kodama, I. Nakajima, M. Kumada, A. Minato, and K. Suzuki, Tetrahedron, 38, 3347 共1982兲. 14. T. Sone, M. Inoue, and K. Sato, Bull. Chem. Soc. Jpn., 61, 3779 共1988兲. 15. J. P. Monthe´ard, J. F. Delzant, and M. Gazard, Synth. Commun., 14, 289 共1984兲. 16. R. Sugimoto, S. Takeda, H. B. Gu, and K. Yoshino, Chem. Express, 1, 635 共1986兲. 17. X. Wang, G. Shi, and Y. Liang, Electrochem. Commun., 1, 536 共1999兲. 18. K. Meerholtz and J. Heinze, Electrochim. Acta, 41, 1839 共1996兲. 19. N. Sato, M. Rikukawa, K. Sanui, and N. Ogata, Synth. Met., 101, 132 共1999兲. 20. T. Olinga and B. Franc¸ois, Synth. Met., 69, 297 共1995兲. ¨ sterholm, and J. Korvola, Polymer, 33, 1559 21. V. M. Niemi, P. Knuuttila, J. E. O 共1992兲. 22. J. P. Ferraris, M. M. Eissa, I. D. Brotherston, and D. C. Loveday, Chem. Mater., 10, 3528 共1998兲. 23. Y. Gofer, J. G. Killian, H. Sarker, T. O. Poehler, and P. C. Searson, J. Electroanal. Chem., 443, 103 共1998兲. 24. U. Barsch and F. Beck, Electrochim. Acta, 41, 1761 共1996兲. 25. J. Wang, Electrochim. Acta, 42, 2545 共1997兲. 26. E. W. Tsai, S. Basak, J. P. Ruiz, J. R. Reynolds, and K. Rajeshwar, J. Electrochem. Soc., 136, 3683 共1989兲. 27. B. Rasch and W. Viekstich, J. Electroanal. Chem., 370, 109 共1994兲.