Discharge Properties of Radical Polymer

Jun 10, 2010 - Takashi Sukegawa , Issei Masuko , Kenichi Oyaizu , and Hiroyuki Nishide ... Kenichi Oyaizu , Takeshi Kawamoto , Takeo Suga , and Hiroyu...
0 downloads 0 Views 271KB Size
Download PDF
J. Phys. Chem. B 2010, 114, 8335–8340

8335

Improving Charge/Discharge Properties of Radical Polymer Electrodes Influenced Strongly by Current Collector/Carbon Fiber Interface Sunao Yoshihara,† Hiroshi Isozumi,† Masanori Kasai,† Hisatomo Yonehara,† Yuko Ando,‡ Kenichi Oyaizu,‡ and Hiroyuki Nishide*,‡ Graphics Arts Laboratory, DIC Corporation, Tokyo 174-8520, Japan, and Department of Applied Chemistry, Waseda UniVersity, Tokyo 169-8555, Japan ReceiVed: March 4, 2010; ReVised Manuscript ReceiVed: April 30, 2010

Charge/discharge processes of organic radical batteries based on the radical polymer’s redox reaction are largely influenced by carbon fibers consisting in the composite electrodes to help electron transfer. To find the optimal structure of the composite electrodes, the dominant electron transfer processes were determined by ac impedance measurement of the composite electrodes. A strong correlation between the overall electron transfer resistance of the composite electrodes and the materials of the current collector suggests that the electric conduction to the current collector through the contact resistance should be crucial. It was also confirmed that the charge/discharge performance of the composite electrode was related to the overall electron transfer resistance of the composite electrode. These results indicated that the charge/discharge performance of the radical battery was dominated by the interfacial electron transfer processes at the current collector/carbon fiber interface and that the rate performance would be much improved by suitably designing the interfacial structure. 1. Introduction Recently, organic polymers with redox-active and chemically stable radical pendant substituents, bound to aliphatic backbones per repeating unit, have been paid much attention due to their specific characteristics based on their rapid and repeatable oxidation/reduction capabilities.1-6 A typical example of the pendant substituent in the so-called radical polymers is 2,2,6,6tetramethylpiperidinyl-N-oxy (TEMPO) groups, which is sufficiently robust as well as rapidly, reversibly, and stoichiometrically oxidized to the corresponding oxoammonium cation via chemical or electrochemical oxidation. Studies on the redox chemistry of the radical polymers have been stimulated by our recent approach to the “radical battery”,7-13 which is characterized by an excellent rate performance and potential capability of fabricating purely organic, flexible, paper-like, and transparent rechargeable energy-storage devices. A schematic of the radical battery structure and the mechanism of charge/discharge processes of the battery are shown in Scheme 1. Radical batteries are featured by the use of the radical polymers as the cathodeactive materials, together with anode-active materials such as Li metal sandwiching an electrolyte layer. The typical cathode consists of the radical polymer, a carbon fiber as the conductive additive, and a small amount of binders to enhance adhesivity to current collectors. The charge/discharge processes of the batteries are based on the radical polymer’s redox reaction, where the charging per repeating unit of the pristine neutral polymer accompanies ion transfer of surrounding electrolytes. The carbon fibers work as the electric conduction path to assist the repeatable oxidation/reduction processes throughout the radical polymer layer. The structure and composition of the polymer/carbon composite electrode are thus one of the most important factors and determine the overall characteristics of * Corresponding author. E-mail: [email protected]. † DIC Corporation. ‡ Waseda University.

SCHEME 1: Charge/Discharge Mechanism of the Radical Battery

the radical battery. However, previous studies to improve the battery performance focused on the properties of radical polymersthemselvesandconductiveagentsratherindependently.14-17 We forecast that determination of the process, most responsible for the electron transfer reactions, is the shortcut to improve the battery performance. In this paper, we report the electrochemical analysis on the radical polymer/carbon composite electrodes with a view to unravel the most influential step of the overall charge/discharge process, focusing on the electron transfer resistance18,19 obtained from impedance spectroscopy. We propose that four processes take roles in the electron transfer reactions within the polymer/carbon composite electrode and demonstrate that the charge/discharge properties of the composite electrode is dramatically improved by modifying the

10.1021/jp1019526  2010 American Chemical Society Published on Web 06/10/2010

8336

J. Phys. Chem. B, Vol. 114, No. 25, 2010

interfacial structure that is relevant to the most important process, providing a design principle to fabricate radical batteries with excellent performances. 2. Experimental Section 2.1. Materials. All solvents were purchased from Kishida Kagaku Co. Poly(1-oxy-2,2,6,6-tetramethylpiperidin-4-yl vinyl ether) (PTVE), the radical polymer employed as the electrodeactive material in this report, was prepared as described previously by our group.20,21 A vapor-grown carbon nanofiber (VGCF-H) was purchased from Shouwa Denko Co. Poly(vinylidene fluoride) (PVdF) resin (KF#1300), employed as a binder resin in a N-methyl-2-pyrrolidone (NMP) based ink (i.e., a dispersion of the polymer/carbon nanofiber), was purchased from Kureha Chemical Co. (Carboxymethyl)cellulose (CMC) resin (CMC1360) and poly(tetrafluoroethylene) (PTFE) resin (D-210C) were employed as binder resins in water-based inks and were purchased from Daicel Chemical Industries Co. and Daikin Industries Ltd., respectively. An electrochemical grade solution of 1 M lithium bis(pentafluoroethanesulfonyl)imide (LiBETI) in propylene carbonate (PC) was purchased from Kishida Kagaku Co. All materials were used without further purification. 2.2. Preparation of Composite Electrodes. The polymer/ carbon composite electrodes were prepared by the stencil printing process on substrates or current collectors using several kinds of inks. We prepared two types of inks for our experiment, NMP-based inks and water-based inks. The preparative method for the composite electrodes from the NMP-based inks was as follows. A radical polymer, PTVE, was mixed with VGCF-H and PVdF in NMP by a relatively low-powered disperser to obtain the NMP-based inks. Nonvolatile content of each NMPbased ink was adjusted to 10%. The obtained inks were screen printed with a contact metal mask to give wet films with thickness around 180 µm on various substrates (i.e., current collectors such as an ITO/glass plate, a platinum plate, and a glassy carbon plate), and then dried under vacuum at 80 °C for 12 h to give several kinds of the composite electrodes of the same size (A ) 1 cm2) and thickness (about 20 µm) with varied composition, PTVE/VGCF-H/PVdF ) 10/80/10, 30/60/10, 50/ 40/10, and 70/20/10 (w/w/w). Hereinafter NMP-based inks with compositions of PTVE/VGCF-H/PVdF ) 10/80/10, 30/60/10, 50/40/10, and 70/20/10 (w/w/w) are abbreviated as N1, N2, N3, and N4, respectively. The composite electrodes from the water-based inks were prepared by a similar method with slight modifications as follows. The radical polymer PTVE was mixed with CMC and PTFE in water by a relatively high-powered disperser. The resulting paste was mixed with VGCF-H by the relatively lowpowered disperser to obtain the water-based inks. Nonvolatile content of each water-based ink was adjusted to 20%. The difference of the nonvolatile content between the NMP-based inks and the water-based inks was unavoidable to make inks in a similar viscosity range. The mixing equipments and the dispersing conditions were selected to disperse the polymers and the carbon nanofibers sufficiently in water without any significant breakage of carbon fibers. The obtained inks were screen-printed with a metal mask to give wet films with thickness around 180 µm on various substrates and then dried under vacuum at 80 °C for 12 h to give several composite electrodes of the same size (A ) 1 cm2) and thickness (40 µm) with varied composition, PTVE/VGCF-H/CMC/PTFE ) 10/ 80/8/2, 30/60/8/2, 50/40/8/2, and 70/20/8/2 (w/w/w/w). Hereinafter water-based inks with compositions of PTVE/VGCF-

Yoshihara et al. H/CMC/PTFE ) 10/80/8/2, 30/60/8/2, 50/40/8/2, and 70/20/ 8/2 (w/w/w/w) are abbreviated as W1, W2, W3, and W4, respectively. Thicker films of the composite electrodes were obtained similarly by using the corresponding thicker metal masks. 2.3. Electrochemical Measurements of Composite Electrodes. Electrochemical measurements such as cyclic voltammetry, ac impedance spectroscopy, and charging/discharging experiments of the composite electrodes were carried out in a conventional three-electrode cell using a 1287 type potentiostatgalvanostat and a 1255WB type electrochemistry measurement system (Solartron). The temperatures of the three-electrode cells were controlled by a high-speed temperature changeable thermostat (ESPEC). All experiments, except those described in section 3.2.1, were carried out at 25 °C (vide infra). Platinum wire and commercial Ag/Ag+ immersed in a solution of 0.01 M AgNO3 and 0.1 M tetrabutylammonium perchlorate (TBAP) in CH3CN were used as the auxiliary electrode and the reference electrode, respectively. These electrodes were purchased from BAS Co. Composite electrodes with various compositions, prepared on various current collectors, were used as the working electrode. The formal potential of the ferrocene/ferrocenium redox couple was 0.45 V vs the Ag/Ag+ reference electrode. In cyclic voltammetry, the electrode potential was initially set at -0.2 V, swept up to +1.2 V and then down to -0.2 V, at a scan rate of 3 mV s-1. In ac impedance measurements, the frequency range was 105 to 10-2 Hz with an ac amplitude of 10 mV. The dc bias voltage was set at (Epa + Epc)/2 of each cells, at which 50% of the nitroxide radical units in the polymer were supposed to be oxidized to the cationic state (i.e., the oxoammonium cation). 3. Results and Discussions 3.1. Cyclic Voltammetry of the Composite Electrodes. Cyclic voltammograms recorded for the various compositions of composite electrodes on several kinds of current collectors prepared from the water-based inks or the NMP-based inks are shown in Figure 1. Composite electrode prepared on the ITO/ glass plate obtained from W1 and N1 gave reversible redox waves at (Epa + Epc)/2 ) 0.37 and 0.36 V vs Ag/Ag+ and peakto-peak separations (∆Ep ) Epa - Epc) at 0.40 and 0.33 V, respectively, as shown by the black curves in Figure 1a,b. A slight difference in the magnitude of the peak currents between the two voltammograms suggested some difference in the amount of the inks mounted on the current collectors, which was, however, sufficiently small to assume that the redox capacities were substantially identical for the two electrodes. The volumes of (Epa + Epc)/2 and ∆Ep, and the capability of the redox cycling were almost comparable, indicating that the redox property of the radical polymer was not affected by the ink media during the electrode fabrication process. A significant increase in the peak-to-peak separation and the magnitude of the peak current were observed as the content of the radical polymer increased and the amount of the carbon fiber decreased in the ink (Figure 1a,b). The effect of the ink composition was similarly observed for the composite electrodes prepared from the water-based inks and prepared from the NMP-based inks, which demonstrated that the decrease in the amount of the carbon nanofiber as the conductive additive to provide the electron transfer path lowered the redox performance while the increase in the amount of the radical polymer as the redoxactive materials led to the larger redox capacities of the electrodes. It has been shown22 that charge propagation in radical polymer layers is accomplished by way of an electron hopping through self-exchange reactions and that the overall redox

Charge/Discharge of Radical Polymer Electrodes

Figure 1. Cyclic voltammograms recorded for the various compositions of composite electrodes on several current collectors. All scan rates were 3 mV/s. (a) Composite electrodes on the ITO/glass plates prepared from W1 (black curve), W2 (red curve), W3 (blue curve), and W4 (green curve). (b) Composite electrodes on the ITO/glass plates prepared from N1 (black curve), N2 (red curve), N3 (blue curve), and N4 (green curve). (c) Composite electrodes on an ITO/glass plate (black curve), a platinum plate (red curve), and a glassy carbon plate (blue curve) prepared from W1. (d) Composite electrodes on an ITO/glass plate (black curve), a platinum plate (red curve), and a glassy carbon plate (blue curve) prepared from N1.

efficiency is improved by adding the carbon fiber as the conductive additive. As the ratio of the radical polymer in the composite electrode increased, the overall redox efficiency of the composite electrode decreased. While the theoretical capacity of the composite electrode (based on the amount of the polymer in the composite layer) increased as the content of the redoxactive materials increased by decreasing the amount of the carbon fiber, the obtained redox capacity was significantly smaller than the theoretically calculated or formula weight-based capacity (Table 1). Rational design of the polymer/carbon composite electrode thus requires consideration of balance of the redox efficiency and the capacity which trade off each other in the composite layer. Cyclic voltammograms of the composite electrodes prepared on the plates of ITO/glass, platinum, and grassy carbon revealed that the peak-to-peak separations significantly depended on the current collector although the composition of the composite electrodes were identical (Figure 1c,d). It was suggested that the current collector also influenced on the characteristic of the composite electrode, especially in the kinetic aspect. The largest redox capacity obtained with the ITO current collector suggested that the volume resistances and the work functions of the current collectors were not the critical factors (as those of ITO are not optimal among evaluated collectors) and was indicative of the presence of specific layer at the current collector/carbon fiber interface that dominated the overall redox properties. 3.2. ac Impedance Measurements of Composite Electrodes. In the composite electrode, the carbon fibers provide a conducting path by forming a network, and the radical polymer is located along the carbon network. And some radical polymers hold carbon fibers within themselves. The conducting path allows the repeatable redox processes of the radical polymer throughout the polymer layer, in which the redox efficiency of the composite electrode depends on the degree of dispersion of the carbon fiber and the radical polymer. A schematic repre-

J. Phys. Chem. B, Vol. 114, No. 25, 2010 8337 sentation of the cross-sectional profile of the composite electrode is illustrated in Figure 2. The charging/discharging process of the composite electrode is determined by four processes: (1) the radical site-to-site electron hopping by self-exchange reactions in the polymer layer, (2) the electrode reaction or heterogeneous electron transfer of the radical site at the carbon fiber interface, (3) the electric conduction in the carbon fiber network, and (4) the interfacial electron transfer from the carbon fiber to the current collector through the contact resistance. The structure and the properties of the composite electrodes are the most important factors that influence on the overall characteristics of the battery, because the charge/discharge processes of the batteries involve the radical polymer’s redox reaction in the composite electrodes. To elucidate the most important processes that dominate the battery performance among the proposed four processes, (1)-(4) with a view to improve the properties of the composite electrode, ac impedance analysis of the composite electrodes were investigated. ac impedance measurements are one of the most useful methods to clarify the rate-determining step of redox reactions of various composite materials. Up to now, ac impedance measurements have been used in various areas of research including the evaluation of the charge transfer resistance for the hydrogen evolution reaction in fuel cells,23 and the determination of electrochemical performance of anodes in lithium ion batteries.24,25 In Nyquist plots, the x- and the y-axes represent the real (Z′) and the imaginary part (Z′′) of the impedance spectra, respectively. Figure 3 shows the impedance spectra obtained for the three-electrode cell, in which the composite electrode, a platinum wire, and the Ag/Ag+ electrode are used as the working, the auxiliary, and the reference electrodes, respectively. The resistance of the electrolyte solution (Rs) appears at the realaxis intercepts in the higher frequency region. The resistance of the composite electrode (Rf) including the charge transfer resistance and the composite electrode/current collector interfacial resistance appears at the real-axis intercepts in the lower frequency region. In detail, the Z′-coordinate of the flexion point in the lower frequency region and that in the higher frequency region are defined to be Rs and R, respectively. And the difference in the value of Rs and R was assumed to be Rf. When the Nyquist plot has only one flexion point, Rs is equal to R, which means Rf is zero. Remarkably, Rs values were nearly constant (between 70 and 80 Ω) in all impedance spectra of the composite electrodes examined in the present study, which suggested that the propylene carbonate solution of 1 M LiBETI used in the impedance measurement had sufficiently high ionic conductivity of the lithium cation. The Nyquist plots for the composite electrodes on ITO/glass plates prepared from the water-based ink showed Rf values almost equal to zero (Figure 3a), indicating that the charge transfer resistances and the contact resistances at the composite electrode/current collector interface were small. As for the composite electrodes prepared from the NMP-based ink (Figure 3b), larger Rf were observed in some compositions of the composite electrodes. In particular, Rf became larger as the composition of the radical polymer increased (i.e., as the carbon fiber content decreased). The effect of the four processes on Rf shown in Figure 2 provided a diagnosis on the overall redox process (vide infra). 3.2.1. Effect of Radical Site-to-Site Charge Transfer Process. Dependence of electric conduction on temperatures in conducting polymers is expressed by the Mott-Davis model26-28 according to eq 1, where σ is the conductance, Eh is the energy barrier for the electron transfer, kB is the Boltzmann constant,

8338

J. Phys. Chem. B, Vol. 114, No. 25, 2010

Yoshihara et al.

TABLE 1: Theoretical Capacity and Redox Efficiency of the Composite Electrodesa ink radical polymer (%) carbon fiber (%) (Epa + Epc)/2 (V) Ep (V) theoretical capacity (mC) observed capacity (mC) redox efficiency (%) W1 W2 W3 W4 N1 N2 N3 N4

10 30 50 70 10 30 50 70

80 60 40 20 80 60 40 20

0.37 0.37 0.38

0.40 0.49 0.76

0.36 0.37 0.36

0.33 0.48 0.86

97 292 486 680 49 146 243 340

2.8 3.9 6.5 5.9 1.9 3.1 6.3 7.1

2.9 1.3 1.3 0.9 3.8 2.1 2.6 2.1

a The theoretical (i.e., the formula weight-based) capacity of the composite electrode is calculated from the amount of the polymer in the composite layer, where an increase in polymer/carbon ratio means an increase of the content of the redox-active materials. The observed capacity is given experimentally based on cyclic voltammograms. Cyclic voltammograms recorded for ink W1-N4 are showed in Figure 1.

Figure 2. Cross-sectional profile of the composite electrode, showing the four significant charge transfer processes: (1) the radical site-tosite electron hopping by self-exchange reactions in the polymer layer, (2) the electrode reaction or heterogeneous electron transfer of the radical site at the carbon fiber interface, (3) the electric conduction in the carbon fiber network, and (4) the interfacial electron transfer from the carbon fiber to the current collector through the contact resistance.

T is the absolute temperature, and σ0 has the information on the hopping distance.

σ ) σ0 exp[-Eh /kBT]

(1)

The parameter σ0 is a function of the temperature but is usually approximated to be constant. If the hopping efficiency between the radical sites in the composite electrode influences the overall electric resistance of the composite layer (Rf) evaluated by the Nyquist plots, the resistance is expected to change logarithmically with T according to eq 1. The temperature dependence of the Nyquist plots of the composite electrode prepared from N3 are shown in Figure 4, and the relationship between Rf and T are also shown. These results revealed that Rf simply increased upon heating but Rs decreased. These results suggested that Rf was not related to the radical site-to-site charge transfer process. 3.2.2. Charge Transfer at the Radical Site/Carbon Fiber Interface. Charge transfer between the radical site and the carbon fiber corresponds to the electron transfer at the polymer/ carbon interface. In the impedance spectra of all compositions of the composite electrodes prepared from the water-based inks (Figure 3a), Rf values were almost equal to zero, which indicated that the barrier for the interfacial electron transfer process was sufficiently low. Similarly, one could assume that the Rf for the NMP-based ink (Figure 3b) may not be influenced by the interfacial electron transfer process. However, the dispersed state of the radical polymer can be different in the two inks, because of the difference in the solubility of the polymer. Such a difference in the microstructure of the radical polymer and the carbon fiber can lead to the enhanced influence from the interfacial electron transfer process. Under these conditions, Rf

Figure 3. Nyquist plots of (a), (c) several composite electrodes prepared from W1 (black curve), W2 (red curve), W3 (blue curve), and W4 (green curve) and (b), (d) several composite electrodes prepared from N1 (black curve), N2 (red), N3 (blue curve), and N4 (green curve). Current collectors of all the composite electrodes were ITO/glass plates. The dry film thickness of the composite electrodes measured in (a)-(d) were 40, 20, 120, and 60 µm, respectively. (Film thicknesses of the composite electrodes measured in (c) and (d) were 3 times the thicknesses of those measured in (a) and (d) each.) Enlarged spectra are shown as insets.

depends on the contact area between the radical polymer and the carbon fiber, which can be minimized by allowing the close contact of the radical polymer to the carbon fiber. Although the polymer/carbon-fiber contact area should have the maximum value amid the varied composite ratio, as shown in Figure 3b, Rf increased simply along the content of the radical polymer, which indicates that Rf was not related to the charge transfer process at the polymer/carbon fiber interface. 3.2.3. Electric Conduction in the Carbon Fiber Network. The electric conduction within the carbon fiber network mediates the charge transfer from most of the radical sites in the polymer to the current collector. When the layer thickness of the composite electrode is tripled, total distance of the electric conduction path should be tripled in the direction of the layer thickness. One can anticipate that Rf would also be tripled when the Rf value shown in the Nyquist plot is related to the electric conduction in the carbon fiber network. The layer thicknesses of the composite electrodes were varied with respect to those shown in Figure 3a,b, and the impedance measurements of these composite electrodes were carried out (Figure 3c,d). However,

Charge/Discharge of Radical Polymer Electrodes

Figure 4. Nyquist plots of composite electrodes prepared from N3. Impedance measurements of this composite electrode were carried out at temperatures of 20 °C (black curve), 30 °C (red curve), 40 °C (blue curve), 50 °C (green curve), and 60 °C (orange curve). The resistance data (Rs, R, and Rf) calculated from these spectra are shown in the inset. The current collector of this composite electrode was a glassy carbon plate. The dry film thickness of the composite electrodes was about 20 µm.

Figure 5. Nyquist plots of (a) composite electrodes prepared from W1 on various current collectors of an ITO/glass plate (black curve), a glassy carbon plate (red curve), and a platinum plate (blue curve) and (b) composite electrodes prepared from W3 on various current collectors of an ITO/glass plate (black curve), a glassy carbon plate (red curve), and a platinum plate (blue curve). The dry film thicknesses of the composite electrodes were about 40 µm.

the values of Rf shown in the Nyquist plots were almost the same as those in Figure 3a,b, which indicated that Rf was rather independent of the thickness of the composite electrode layer and was not related to the electric conduction in the carbon fiber network. It may be noted that new resistance components originating from the thick composite electrode layers were observed in the Nyquist plots in frequency regions lower than those for Rf (Figure 3d). This is an additional resistance factor that should be considered when the composite electrodes are made thicker to accomplish a larger battery capacity. 3.2.4. Electric Conduction across the Carbon Fiber/Current Collector Interface. Electric conduction across the carbon fiber/ current collector interface over the contact resistance was characterized by the impedance measurements of composite electrodes on different current collectors. The impedance spectra of composite electrodes recorded on various current collectors such as an ITO/glass plate, a platinum plate, and a glassy carbon plate were shown in Figure 5. The values of Rf obtained in the impedance spectra of the composite electrodes prepared from W1 and W3 depended on the current collectors. The composite electrodes on the ITO/glass plates gave smaller Rf than those on the platinum plates and the glassy carbon plates. The composite electrodes prepared from W1 gave larger Rf on the platinum plate than on the glassy carbon plate. However,

J. Phys. Chem. B, Vol. 114, No. 25, 2010 8339

Figure 6. Charge/discharge curves of the composite electrodes prepared from W1 on various current collectors of (a) an ITO/glass plate and (b) a glassy carbon plate. Charging/discharging rates were 1 C (black curve), 5 C (red curve), and 10 C (blue curve), respectively.

the composite electrodes prepared from W3 gave almost the same Rf values on the glassy carbon plate and on the platinum plate. This tendency was commonly observed in composite electrodes prepared from the water-based ink and from the NMP-based ink. The superiority of an ITO/glass plate as a current collector material would originate from the surface roughness of the current collector, and adhesion to the composite electrodes on the current collector, rather than the matching of the work function and the volume resistivity. Indeed, the surface roughness of these current collectors has been studied by many researchers.29-31 Clarification of the correlation between the superiority of the ITO/glass plate as the current collector material and the surface roughness of the current collector will be the topics of our future work. 3.3. Charging/Discharging Characteristics of the Composite Electrodes. The difference in the Rf values between the composite electrodes significantly reflected in the charging/ discharging behavior, especially for those prepared from W1 on an ITO/glass plate and those on a glassy carbon plate. To examine the influence of Rf on the charge/discharge properties, the charging/discharging experiments of those two composite electrodes were carried out (Figure 6). A distinct effect of Rf on the charge/discharge properties was observed for the composite electrodes even with the identical compositions. The charge/discharge experiments by constant current electrolysis of the composite electrode revealed that a larger specific capacity was obtained with the ITO/glass plate as the current collector than with the glassy carbon plate. These results suggested that the charge/discharge properties of the composite electrodes were strongly related to Rf of the electrodes observed in the Nyquist plots. The composite electrode with the small Rf showed a remarkably excellent charge/discharge property. 4. Conclusions Charging and discharging properties of the radical polymer/ carbon fiber composite electrodes was investigated by focusing on the four significant charge transfer processes: (1) the radical site-to-site charge transfer within the polymer layer, (2) the charge transfer at the radical site/carbon fiber interface, (3) the electric conduction in the carbon fiber network connecting the

8340

J. Phys. Chem. B, Vol. 114, No. 25, 2010

radical polymer sites and the current collector, and (4) the electric conduction at the carbon fiber/current collector interface. To determine the most effective process in the composite electrodes, which mainly consisted of the radical polymer and the carbon fiber, the impedance analysis of the electrodes was investigated. It was revealed that processes 1, 2, and 3 had less significant effects on the charge transfer resistance of the composite electrode but process 4 dominated the overall charge transfer process, based on the strong correlation between the overall resistance of the composite electrodes including the current collectors and the material of the current collector. The superiority of an ITO/glass plate as a current collector material would be would be originated by the surface roughness and adhesion to the composite layers, rather than their work functions and volume resistivities. We also confirmed that the charge/discharge performance of the composite electrode was related to the overall electron transfer resistance of the composite electrode including the current collectors. These results suggested that the charge/discharge performance of the organic radical battery was dominated by the interfacial electron transfer processes at the current collector/carbon fiber interface. A better performance of the battery is expected with electrode-active materials, which accomplishes a closer contact of the current collector and the carbon fiber. A strategy to improve the performance of the aluminum current collector was suggested, by mechanically pressing the carbon fiber to penetrate through the resistant layer, which is most likely an aluminum oxide. Further studies on the electron transfer process of current collector-composite electrode are being continued to improve the organic radical battery characteristics. Acknowledgment. This work was partially supported by Grants-in-Aid for Scientific Research (Nos. 19105003, 21655043, 21550120, and 21106519) from MEXT, Japan, and the NEDO Project on “Radical Battery for Ubiquitous Power”. We appreciate the discussion with Fundamental and Environmental Research Laboratory, NEC Co., and Functional Chemicals Research Laboratory, Sumitomo Seika Chemicals Co. This work was carried out at the Graphics Arts Laboratory, DIC Co., and the Advanced Research Institute for Science & Engineering, Waseda University. Supporting Information Available: Additional experiments, table of work functions and volume resistivities, Nyquist plots, and SEM image of the composite electrode. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) McNamara, C. A.; Dixon, M. J.; Bradley, M. Chem. ReV. 2002, 102, 3275–3300.

Yoshihara et al. (2) Fulghum, T.; Taranekar, P.; Advincula, R. Macromolecules 2008, 41, 5681–5687. (3) Kanazawa, H.; Higuchi, M.; Yamamoto, K. J. Am. Chem. Soc. 2005, 127, 16404–16405. (4) Novak, P.; Muller, K.; Santhanam, K. S. V.; Haas, O. Chem. ReV. 1997, 97, 207–282. (5) Dijksman, A.; Marino-Gonza´lez, A.; Payeras, A. M. I.; Arends, I. W. C. E.; Sheldon, R. A. J. Am. Chem. Soc. 2001, 123, 6826–6833. (6) Yonekuta, Y.; Oyaizu, K.; Nishide, H. Chem. Lett. 2007, 36, 866– 867. (7) Nishide, H.; Oyaizu, K. Science 2008, 319, 737–738. (8) Nishide, H.; Suga, T. Electrochem. Soc. Interface. 2005, 14, 32– 36. (9) Yonekuta, Y.; Susuki, K.; Oyaizu, K.; Honda, K.; Nishide, H. J. Am. Chem. Soc. 2007, 129, 14128–14129. (10) Koshika, K.; Sano, N.; Oyaizu, K.; Nishide, H. Chem. Commun. 2009, 7, 836–838. (11) Suga, T.; Ohshiro, H.; Sugita, S.; Oyaizu, K.; Nishide, H. AdV. Mater. 2009, 21, 1627–1630. (12) Koshika, K.; Sano, N.; Oyaizu, K.; Nishide, H. Macromol. Chem. Phys. 2009, 210, 1989–1995. (13) Oyaizu, K.; Nishide, H. AdV. Mater. 2009, 21, 2339–2344. (14) Zhang, X.; Li, H.; Li, L.; Lu, G.; Zhang, S.; Gu, L.; Xia, Y. Polymer 2008, 49, 3393–3398. (15) Bugnon, L.; Morton, C. J. H.; Novok, P.; Vetter, J.; Nesvadba, P. Chem. Mater. 2007, 19, 2910–2914. (16) Nakahara, K.; Iwasa, S.; Satoh, M.; Morioka, Y.; Iriyama, J.; Suguro, M.; Hasegawa, E. Chem. Phys. Lett. 2002, 359, 351–354. (17) Oyaizu, K.; Suga, T.; Yoshimura, K.; Nishide, H. Macromolecules 2008, 41, 6646–6652. (18) Nakahara, K.; Iriyama, J.; Iwasa, S.; Suguro, M.; Satoh, M.; Cairns, E. J. J. Power Source 2007, 165, 870–873. (19) Kim, J. K.; Cheruvally, G.; Choi, J. W.; Ahn, J. H.; Choi, D. S.; Song, C. E. J. Electrochem. Soc. 2007, 154, A839–A843. (20) Suguro, M.; Iwasa, S.; Kusachi, Y.; Morioka, Y.; Nakahara, K. Macromol. Rapid Commun. 2007, 28, 1929–1933. (21) Suga, T.; Kasatori, K.; Nishide, H. Polym. Prepr., Jpn. 2004, 53 (2), 4824. (22) Oyaizu, K.; Ando, Y.; Konishi, H.; Nishide, H. J. Am. Chem. Soc. 2008, 130, 14459–14461. (23) Girishkumar, G.; Rettker, M.; Underhile, R.; Binz, D.; Vinodgopal, K.; McGinn, P.; Kamat, P. Langmuir 2005, 21, 8487–8494. (24) Yoo, M.; Frank, C. W.; Mori, S.; Yamaguchi, S. Chem. Mater. 2004, 16, 1945–1953. (25) Ruffo, R.; Hong, S. S.; Chan, C. K.; Huggins, R. A.; Cui, Y. J. Phys. Chem. C 2009, 113, 11390–11398. (26) Beadle, P.; Armes, S. P.; Gottesfeld, S.; Mombourquette, C.; Houlton, R.; Andrews, W. D.; Agnew, S. F. Macromolecules 1992, 25, 2526–2530. (27) Vijayashree, M. N.; Subramanyam, S. V.; Samuelson, A. G. Macromolecules 1992, 25, 2988–2990. (28) Epstein, A. J.; Bigelow, R. W.; Rommelmann, H.; Bibson, H. W.; Weagley, R. J.; Feldblum, A.; Tanner, B. D.; Pouget, J. P.; Pouxviel, J. C.; Comes, R.; Pron, P.; Kivelson, S. Mol. Cryst. Liq. Cryst. 1985, 117, 147. (29) Lei, H. N.; Metrot, A.; Troyon, M. J. Vac. Sci. Technol., B 1994, 12 (3), 1827–1831. (30) Heiduschka, P.; Munz, A. W.; Gopel, W. Electrochim. Acta 1994, 39, 2207–2223. (31) Armstrong, N. R.; Veneman, P. A.; Ratcliff, E.; Placencia, D.; Brumbach, M. Acc. Chem. Res. 2009, 42 (11), 1748–1757.

JP1019526