Electrochemical Study of High Electrochemical ... - ACS Publications


Electrochemical Study of High Electrochemical...

0 downloads 123 Views 406KB Size

J. Phys. Chem. C 2007, 111, 227-233

227

Electrochemical Study of High Electrochemical Double Layer Capacitance of Ordered Porous Carbons with Both Meso/Macropores and Micropores Hirotoshi Yamada,*,† Haruka Nakamura,† Fumihiro Nakahara,‡ Isamu Moriguchi,† and Tetsuichi Kudo§ Faculty of Engineering, Nagasaki UniVersity, Graduate School of Science and Technology, Nagasaki UniVersity, and Joint Research Center, Nagasaki UniVersity 1-14, Bunkyo-machi, Nagasaki, 852-8521, Japan ReceiVed: June 22, 2006; In Final Form: September 5, 2006

Porous carbons with large meso/macropore surface areas were prepared by the colloidal-crystal-templating technique. The porous carbons exibited extremely high specific electrochemical double layer (EDL) capacitance of 200-350 F g-1 in an aqueous electrolyte (1 M H2SO4). The pore structure dependence of the capacitance was studied mainly by means of cyclic voltammetry and is discussed in detail. From the sweep rate dependence of the series resistance and capacitance, it was found that the ion-penetration depth at the porous electrode surface was finite and decreased with an increasing sweep rate. Peaks around the point of zero charge, which were observed in addition to typical rectangular voltammograms, were explained well by the potential drop in pores. The surface area dependence of the capacitance revealed that the contribution of the meso/macropore surface is as great as that of the plane electrodes and that only the part of the micropore surface adjacent to the opening mouths is effective.

1. Introduction Electrochemical double layer capacitors (EDLCs) have attracted considerable interest for their potential applications to energy storage devices, especially auxiliary high power sources of hybrid electric vehicles (HEVs) and fuel cell electric vehicles (FCEVs). Their outstanding properties, compared to batteries (Li ion and Ni-MH), are high power density, high reversibility, and long cycle life. Porous carbonaceous materials with large surface area (e.g., activated carbons) have been adopted as electrodes for gaining large capacities, that is, higher energy densities.1-9 Capacitance of activated carbons, however, does not increase proportionally to their surface area. The pore size of carbons is also one of the critical factors limiting capacitance: if pores are smaller than solvated ions, then the surfaces of those pores are hardly accessible to the ions. Recently, several groups have synthesized porous carbons by using nanostructured templates and have reported their EDLC performance.5-9 We have fabricated carbons with three dimensionally ordered interconnected pores (pore diameters: 16, 45, 80, and 120 nm) by using colloidal crystals as templates and have revealed that differential capacitance in aqueous solution depends linearly on the surface areas of the mesopores (2 nm < d < 50 nm, with d being the pore diameter) and the macropores (d > 50 nm) and that the contribution of the surface of the micropores (d < 2 nm) is negligible.8,9 The specific capacitance per unit surface area of meso- and macropores agreed well with the typical specific capacitance of carbon in aqueous electrolytes, 20 µF cm-2, and a very large capacitance of 190 F g-1 was obtained for 16-nm porous carbons. These facts indicate that the capacitance of porous carbon is affected by the surface of the larger pores. Due to the smooth transportation of electrolyte * To whom correspondence should be addressed. Phone/Fax: +81-958192861. E-Mail: [email protected]. † Faculty of Engineering, Nagasaki University. ‡ Graduate School of Science and Technology, Nagasaki University. § Joint Research Center, Nagasaki University.

solution in the interconnected pores, a large capacitance was retained even at high charge/discharge rates (with respect to 16-nm porous carbons, more than 134 F g-1 at 100 mV s-1 in 1 M H2SO4). These porous carbons not only are attractive from the viewpoint of applications like supercapacitors but also are scientifically interesting because a well-controlled structure facilitates a more simplified analysis and the large surface area emphasizes processes on the surfaces so that even small processes can be detected. In this report, we have fabricated carbons with a wide variety of pore sizes using colloidal crystals as templates and have investigated their capacitive properties in detail, mainly by means of cyclic voltammetry. 2. Experimental Methods Monodispersed SiO2 colloidal solutions, Spherica slurry 120 (with an average particle diameter of 120 nm), Cataloid SI-80 (80 nm), SI-45P (45 nm), S-30H (16 nm), and SI-30 (11.5 nm) were kindly supplied by Catalysts & Chemicals Ind. Co. Ltd., and 8-nm colloidal SiO2 was kindly supplied by Sumitomo Chemical Co. Ltd. Ordered porous carbons were prepared by following four steps based on the reported procedure.8,9 First, SiO2 colloidal crystals were fabricated by centrifugation of the colloidal solutions and subsequent drying. Then, the colloidal crystals were immersed in a mixed solution of phenol, aqueous formaldehyde, and a small amount of concentrated hydrochloric acid with a mass ratio of SiO2/PhOH/HCHO ) 1.0:0.8125: 0.222. After the immersion, phenol was polymerized together with formaldehyde in the interstitial space among SiO2 particles by thermal treatment at 400 K for 12 h in air. Next, the interconnected phenolic (novolac) resin was carbonized at 1073 K for 5 h in an Ar gas flow (100 mL min-1). Heating rates were varied in the range of 1-5 K min-1. Finally, SiO2 templates were removed from the carbon/SiO2 composites by an aqueous HF solution (46%) and the resultant porous carbons were dried under vacuum for 1 day. The carbons thus obtained

10.1021/jp063902g CCC: $37.00 © 2007 American Chemical Society Published on Web 12/05/2006

228 J. Phys. Chem. C, Vol. 111, No. 1, 2007

Yamada et al.

TABLE 1: Properties of Porous Carbons specific surface areac dSiO2a (nm)

rhb (K min-1)

atomic ratio H/C

Stotal (m2 g-1)

Smeso (m2 g-1)

Smicro (m2 g-1)

1 3 5 1 3 5 1 3 1 5 1 5

0.22 0.08 0.09 0.29 0.17 0.12 0.22 0.23 0.13 0.14 0.14 0.14

1508 1557 1139 1399 1094 1162 1656 1581 1214 1657 2090 1830

520 562 497 1041 894 898 1079 1093 1070 821 1964 1768

988 996 642 358 200 265 577 488 144 837 125 62

80 45 16 11.5 8

ad b c SiO2: particle size of template SiO2. rh: heating rate on carbonization. Total surface area, Stotal, and meso/macropore surface area, Smeso, were obtained from an RSPE plot and t plot, respectively. Micropore surface area, Smicro, was calculated by subtracting Smeso from Stotal.

are referred to as C[dSiO2]-rh, where dSiO2 indicates the average diameter of the template SiO2 particles and rh is the heating rate for carbonization. The removal of SiO2 from the carbon/SiO2 composites was confirmed by thermalgravimetry and differential thermal analysis (TG-DTA) on a TG/DTA6200 (Seiko Instruments Inc.; heating rate, 10 K min-1 in an air flow). To investigate the carbon structure, X-ray diffraction (XRD) patterns were recorded on a RINT-2200 (Rigaku, with irradiation of Ni-filtered Cu KR) and Raman spectra were measured on an RMP-210 (JASCO, with a 532-nm laser). The porous structure of carbons was observed by transmission electron microscopy (TEM, JEOL JEM-2010). Pore parameters were obtained by the analysis of nitrogen adsorption-desorption isotherms recorded at 77 K on a BELSORP-mini (BEL Japan, Inc.). The total specific surface area (Stotal) was determined by analyzing the RSPE plot based on the subtracting pore effect (SPE) method.10,11 The specific surface area of meso- and macropores (Smeso) was analyzed by using t plots.12 The specific surface area of micropores (Smicro) was estimated by subtracting Smeso from Stotal. The electrochemical double layer (EDL) capacitance of porous carbons was obtained as the differential capacitance of cyclic voltammograms conducted on electrochemical analyzers CV-50W (BAS Inc.) or S1287 (Solartron Analytical Inc.) at 25 °C. Electrochemical impedance spectra (EIS) were measured from 500 kHz to 5 mHz with an alternating current (ac) amplitude of 5 mV by using an impedance analyzer S1260 (Solartron Analytical Inc.) with the S1287. For each frequency sweep, the electrode potential was set and held at a measurement potential for 10-30 min until equilibration. For electrochemical measurements, porous carbons, mixed with poly(tetrafluoroethylene) at a mass ratio of 95:5, were pressed onto Au nets and were used as working electrodes. Ag/AgCl/sat’d KCl and Pt wires were utilized as reference and counter electrodes, respectively. A 1 M aqueous solution of H2SO4 was used as the electrolyte solution. Before each measurement, nitrogen gas was bubbled through the electrolyte for 30 min to eliminate dissolved oxygen. 3. Results and Discussion 3.1. Characterization of C[dSiO2]-rh. According to the elemental analysis, the atomic ratios of H to C of the obtained porous carbons were 0.1-0.2 (Table 1) and were much smaller than that of phenolic resin (0.93), which demonstrates the progress of carbonization. Raman spectra showed two bands around 1340 and 1580 cm-1 for all carbons, which are the so-

called D-band and G-band, respectively. The intensities and widths of the peaks were almost the same for all the carbons. The XRD patterns exhibited very broad peaks around 24° and 44°, indicating poor crystallinity, which is typical for so-called hard carbons. X-ray photoelectron spectra of porous carbons indicated the presence of C-OH and COOH (and/or COOR) functional groups, details of which are given in the Supporting Information. Figure 1 shows TEM images of porous carbons. It is obvious that the pore sizes of the carbons agreed well with the sizes of colloidal SiO2 particles used. Carbons synthesized by using relatively large SiO2 particles (dSiO2 ) 80 nm) had an ordered porous structure, while the pore arrangement of carbons synthesized by using smaller SiO2 particles was disordered. N2 ad/desorption isotherms (Figure 2a) of all the carbons were classified into type IV, indicating the presence of both micropores and mesopores.13 For small C[dSiO2]s, the pore size distribution, obtained by the Barrett-Joyner-Halenda (BJH) method,14 showed an obvious peak around the diameters which correspond to the template SiO2 particle sizes as shown in the inset of Figure 2b. The specific surface areas of C[dSiO2]-5 are listed in Table 1. In Figure 3, Stotal, Smeso, and Smicro are plotted as a function of template particle size. For all heating rates, Smeso increased with decreasing template particle size, while Smicro decreased. Smeso of spherical pores would be in inverse proportion to the pore diameter (dp), following the equation: Smeso ) 6p/(1 - p)Fdp, where p is porosity (for an ideal hexagonal-closest-packing arrangement: p ) π/3(2)1/2 ) 0.74) and F is the density of the pore walls. In Figure 3, Smeso increased more moderately with decreasing dp than estimated based on the inverse relationship between Smeso and dp. This is ascribable to the decreasing porosity of porous carbons due to the disordered pore arrangement as observed in TEM images. Smeso tended to increase with slower heating rates especially for a smaller dp. It may also be interesting that Smicro values of C[11.5]-1, C[8]-1, and C[8]-5 are very small. These results may be related to the mechanism of carbonization (decomposition of precursor, release of gas, reorientation of carbon atoms, etc.), but further investigation of the dependence of the surface area on the heating rate is beyond the scope of this paper. 3.2. Cyclic Voltammetry. 3.2.1. EDL Capacitance. Figure 4 shows typical cyclic voltammograms of C[dSiO2], where the longitudinal axis is converted to the specific capacitance by dividing the current density by the sweep rate. Table 2 summarizes the apparent differential capacitance at 0.3 and 0.6 V vs Ag/AgCl/ sat’d KCl in anodic processes. The EDL

Double Layer Capacitance of Ordered Porous Carbons

J. Phys. Chem. C, Vol. 111, No. 1, 2007 229

Figure 1. TEM images of porous carbons C[dSiO2]-1 ((a) dSiO2 ) 80, (b) 45, (c) 11.5, and (d) 8 nm).

Figure 3. Template dependence of specific surface areas, Stotal, Smeso, and Smicro of porous carbons C[dSiO2]-1. Lines are guides to the eyes.

Figure 2. (a) N2 ad/desorption isotherms of porous carbons with different pore sizes, C[dSiO2]-5 (dSiO2 ) 80, 11.5, 8 nm). Closed and open symbols represent data of adsorption and desorption branches, respectively. (b) Pore-size distribution for dSiO2 ) 16, 11.5, 8 nm.

capacitance increased with decreasing dSiO2, and a very large EDL capacitance of 200-350 F g-1 was observed. To the best of our knowledge, these are the largest ever reported for pristine carbon electrodes without pseudo-capacitance of Faradaic reactions. Details of the factors involved in obtaining a very large capacitance and the effect of the porous structure will be discussed in the following sections. 3.2.2. Sweep Rate Dependence. Figure 5 shows cyclic voltammograms of C[80]-5 measured at different sweep rates (solid lines). Rectangular voltammograms with small peaks were observed for slow sweep rates up to 10 mV s-1, and the shape became slim and leaf-like with increasing sweep rates. The change in the shapes of the voltammograms is caused by the finite time constant of the charging EDL capacitance. Here, the voltammograms were analyzed by using a simple circuit,

Figure 4. Cyclic voltammograms at 1 mV s-1 of porous carbons with different pore sizes, C[dSiO2]-5 (dSiO2 ) 80, 45, 11.5, 8).

consisting of a series connection of a resistor (Rs, electrolyte and electrode resistance) and a capacitor (Cs, EDL capacitor). For this simple circuit, the current I can be expressed as

For anodic sweeps: BI ) VCs + (I0 - VCs) exp(-t/RsCs)

(1-1)

For cathodic sweeps: AI ) -VCs + (I0 + VCs) exp(-t/RsCs)

(1-2)

where V is the sweep rate, t is the time from the beginning of an anodic or cathodic sweep, and I0 is the current at t ) 0. By

230 J. Phys. Chem. C, Vol. 111, No. 1, 2007

Yamada et al.

TABLE 2: Specific Capacitance of Porous Carbons specific capacitancea (F g-1) dSiO2 (nm)

-1

rh (K min )

C(0.3 V)

C(0.6 V)

1 3 5 1 3 5 1 3 1 5 1 5

223 201 122 330 241 295 271 248 241 271 290 313

203 200 122 224 195 232 200 204 200 218 255 227

80 45 16 11.5 8

Figure 6. Calculated series resistance, R, and EDL capacitance, C, as a function of sweep rate.

a Apparent differential capacitance at 0.3 and 0.6 V vs Ag/AgCl/ sat’d KCl on anodic processes in voltammograms.

Figure 7. Sweep rate dependence of EDL capacitance of porous carbons.

Figure 5. Sweep rate dependence of cyclic voltammograms of C[80]5 for 5, 20, and 100 mV s-1 (open symbols) and calculated voltammograms (solid lines).

repeating several sweeping cycles, the voltammograms converge on one closed curve as follows:

For anodic sweeps:

{

(

)}

(

)}

E - EL BI 2 ) Cs 1 + exp V 1+φ VRsCs

(2-1)

For cathodic sweeps:

{

EH - E AI 2 ) -Cs 1 exp V 1+φ VRsCs

(2-2)

where EH and EL are the higher and lower cutoff potential, respectively, and φ is given by

(3)

Figure 8. (a) Cyclic voltammograms of C[45]-5, repeated for 2000 cycles at 10 mV s-1. Those of the 1st, 20th, 100th, 500th, and 2000th cycle were picked out. (b) Cycle dependence of the capacity and the Coulomb efficiency.

The calculated voltammograms are also plotted in Figure 5, with open symbols. The deviation of the measured curves from the calculated ones for slow sweep rates was large, mainly due to fluctuations in the observed voltammograms, and calculations could reproduce the experimental results well for fast sweep rates. Rs and Cs thus obtained are plotted in Figure 6. It is not surprising that both Rs and Cs decreased with increasing sweep rates. The decrease of both Rs and Cs indicates that the effective interface area decreases with an increasing sweep rate. When the sweep rate was slow enough, EDL capacitance could be charged on almost every interface. If the potential was swept

rapidly, then ions would be consumed only on shallow positions of pores and interfaces located deep in the pores would not be charged. Figure 7 shows the rate dependence of the capacitance of porous carbons obtained by fitting voltammograms as described above. The capacitance at 100 mV s-1 was smaller than that at 1 mV s-1. The decrease was 15% for C[45] and C[16] and 21% for C[11.5] and C[8]. These results indicate that the electrolyte transportation is rather smooth even in 8-nm pores. 3.2.3. Cycle Performance. Figure 8a shows cyclic voltammograms of C[45]-5 repeated for 2000 cycles at 10 mV s-1.

(

φ ) exp -

)

EH - E L VRsCs

Double Layer Capacitance of Ordered Porous Carbons

J. Phys. Chem. C, Vol. 111, No. 1, 2007 231

Figure 9. Nyquist plots of electrochemical impedance of C[8]-5 obtained at potentials 0.7, 0.5, 0.3, 0.1, and -0.1 V vs Ag/AgCl/sat’d KCl. The inset demonstrates Nyquist plots for the low-frequency region.

The porous carbon electrodes exhibited good cycleability with a subtle decrease in anodic current for 0.5-0.7 V for the early cycles (e100 cycles). The decrease may be related to deterioration of irreversible side reactions. Figure 8b plots the cycle dependence of the capacity and the Coulomb efficiency. The capacity deterioration was less than 3% for all the cycles and less than 0.8% for 400-2000 cycles. The Coulomb efficiency of the capacity was almost 100%, especially for g400 cycles, which demonstrates the high reversibility of the porous carbon electrode. 3.2.4. Shape of Voltammograms. In voltammograms of porous carbons with small pores, a pair of peaks was clearly observed around 0.3 V vs Ag/AgCl/sat’d KCl (Figure 4). The peak-causing phenomenon was very stable upon repeating cycles, as demonstrated in Figure 8a. The peak intensity became more significant with decreasing pore sizes. These kinds of peaks have often been ascribed to a Faradaic reaction or underpotential deposition on interfaces,15 that is, red-ox reactions of surface functional groups which were revealed by XPS (Fig. S1) (e.g., >CdO + H+ + e- / >C-OH).16 However, for the samples used here, such interfacial charge-transfer processes are supposed to contribute little to the capacitance. Elemental analysis which showed no correlation between the peak intensity and the composition of H and O also indicates a low likelihood of Faradaic processes. In Nyquist plots of EIS measured at several potentials (Figure 9), a semicircle was observed at a high-frequency region, which may often be ascribed to a Faradaic reaction.2 However, the semicircles were almost independent of the potentials, indicating other reasons for the semicircles than charge-transfer processes. The process giving rise to the semicircle will be discussed elsewhere. Figure 10a demonstrates the specific series capacitance (Cs) obtained from the following equation, Cs ) -1/2πfZ'', where Z'' is the imaginary part of the impedance and f is the measurement frequency, and Figure 10b shows Cs values at 0.01 Hz and those obtained from a cyclic voltammogram. The correspondence between both capacitances indicates that the peaks are not derived from diffusion-related processes but from a potentialdependent capacitance. With respect to why the peaks depend on both the pore size and the potential, it should be noted that the potential of zero charge (Epzc) was observed around 0.3-0.4 V vs Ag/AgCl/sat’d KCl. On the basis of this fact and the above results, the potential drop in pores is supposed to give rise to peaks: the solution potential inside pores is different from that outside pores. As is drawn schematically in Figure 11, for E ) Epzc (Figure 11a), the solution potential in pores, φs,pore, is the same as that of bulk solution, φs, and the total capacitance is given by the sum of the capacitance of the outer surface, Cout, and that of the pore surface, Cpore-surf. For E > Epzc (Figure 11b) and E < Epzc,

Figure 10. (a) Specific capacitance of C[8]-5 as a function of frequency. (b) Capacitance obtained from EIS and voltammogram.

φs,pore is higher and lower than φs, respectively, and an additional capacitor appears at the pore mouth, Cpore-mouth. The capacitance Cpore-mouth is much smaller than Cpore-surf, because of the much smaller area of pore mouths in comparison to the pore surface area, and accordingly, the total capacitance decreases. It should be noted that the porous carbons used in this work exhibit two pore-size distributions, meso/macropores and micropores, which are derived from template SiO2 particles and from the carbonization of novolac resins, respectively. Accordingly, the potential drop in each pore is different. In the following section, contributions by meso/macropores and micropores to the capacitance will be discussed. 3.2.5. Capacitance of Meso/macropores and Micropores. To investigate the contribution of meso/macropores and micropores to the capacitance, it was assumed that the overall capacitance, CDL, consisted of both meso/macropores and micropores and could be given by

CDL ) cdl,mesoSmeso + cdl,microSmicro

(4)

where cdl,meso and cdl,micro are the specific EDL capacitance of the meso/macropores and the micropores, respectively.4 Dividing both sides of eq 4 by Smicro results in

Smeso CDL ) cdl,meso‚ + cdl,micro Smicro Smicro

(5)

On the basis of this equation, CDL/Smicro is plotted as a function of Smeso/Smicro in Figure 12, where C(0.6) and C(0.3) represent the capacitance at each potential obtained from the anodic current of cyclic voltammograms. For small Smeso/Smicro values, CDL/Smicro increased linearly with increasing Smeso/Smicro, and for large Smeso/ Smicro, the datapoints deviated from the trends and the slopes became moderate. cdl,meso and cdl,micro were obtained from slopes and intercepts of fitting lines for small Smeso/Smicro. The dependence of Smeso on CDL was negligible at both potentials (cdl,micro ) 4 ( 3, 2 ( 2 µF cm-2 at 0.6 and 0.3 V, respectively), while cdl,meso depended on the potential, that is, 22 ( 3 µF cm-2 at 0.6 V and 29 ( 3 µF cm-2 at 0.3 V. The EDL capacitance of a glassy carbon plate in aqueous H2SO4 is reported to be

232 J. Phys. Chem. C, Vol. 111, No. 1, 2007

Yamada et al.

Figure 11. Schematic of the potential drop in pores for (a) E ) Epzc and (b) E > Epzc. Left figures demonstrate potential profiles, and the right ones show connection of capacitance. φs, φs,pore, and φele represent the potential of bulk solution, pore, and electrode, respectively. Cout, Cpore-surf., and Cpore-mouth are capacitances on the outer surface, pore surface, and pore mouths, respectively.

or from tortuous and disordered pore structure of these carbons, which would be revealed by further study. The above results give hints for designing high-performance EDLC electrodes which exhibit both high-power density and high-energy density. We propose hierarchical porous structures which consist of two pore-size distributions: rather large pores (∼15 nm) that are interconnected for smooth electrolyte transportation, and on their pore walls, small pores (probably ∼2 nm) for large capacitance. Figure 12. Plots of CDL/Smicro vs Smeso/Smicro. C(0.6 V) and C(0.3 V) are the capacitance at 0.6 and 0.3 V vs Ag/AgCl/sat’d KCl, respectively.

18-20 µF cm-2,17 which is less dependent on the potential than the results presented here, for a rather concentrated solution.17,18 In contrast to the reported value, the obtained cdl,meso at 0.3 V is rather large and that at 0.6 V is comparable. These results indicate that the meso/macropores are large enough compared to the hydrated ions, and that the meso/macropores contribute at any potential, and additional parallel capacitance which depends on Smeso contributes to capacitance around Epzc. The finite ion-penetration depth into micropores, indicated in section 3.2.2., accounts for the additional capacitance. Only micropore surfaces which are adjacent to their opening mouths would contribute to capacitance. The number of opening mouths of micropores is supposed to be proportional to the meso/ macropore surface area, resulting in the dependence of the additional capacitance not on Smicro but on Smeso. Accordingly, eq 4 would be revised as

CDL ) cdl,mesoSmeso + cdl,microSmeso‚R ) (cdl,meso + cdl,microR)Smeso (6) where R is a factor representing the contribution of micropores and is related to the effective ion-penetration depth. By assuming cdl,meso to be the typical value of 20 µF cm-2, cdl,microR is estimated to be about 0-5 µF cm-2 at 0.6 V and 5-10 µF cm-2 at 0.3 V. In Figure 12, the points of C[8] and C[11.5] are out of line at a large Smeso/ Smicro. This deviation may result from large electrolyte resistance in small pores, as indicated by the rate capability of the capacitance (see section 3.2.2). However, it is still unknown whether this limitation results from the pore size

4. Conclusions In this study, we fabricated porous carbons with monodistributed meso- or macropores (8-80 nm) by using colloidal crystals as templates. The obtained porous carbons showed very high EDL capacitance of 200-350 F g-1 in a 1 M aqueous H2SO4 solution. The pore dependence of the capacitance was investigated in detail, especially by cyclic voltammetry. The finite ion-penetration model was proposed to explain the simultaneous decreases in resistance and capacitance with an increasing sweep rate. Peaks around a potential of zero charge, which were more significant for carbons with smaller meso-/ macropores, were ascribed to additional capacitance contributed by a shallow part of the micropore surface adjacent to opening mouths. Acknowledgment. The authors are thankful to Mr. Hiroshi Furukawa for his technical support in experiments and TEM measurements. Supporting Information Available: Surface analysis of porous carbons was carried out by X-ray photoelectron spectroscopy, the results of which are demonstrated in parts a and b of Figure S1. This material is available free of charge via the internet at http://pubs.acs.org. References and Notes (1) Salitra, G.; Soffer, A.; Eliad, L.; Cohen, Y.; Aurbach, D. J. Electrochem. Soc. 2000, 147, 2486. (2) Lust, E.; Nurk, G.; Ja¨nes, A.; Arulepp, M.; Nigu, P.; Mo¨ller, P.; Kallip, S.; Sammelselg, V. J. Solid State Electrochem. 2003, 7, 91. (3) Lozano-Castello´, D.; Cazorla-Amoro´s, D.; Linares-Solano, A.; Shiraishi, S.; Kurihara, H.; Oya, A. Carbon 2003, 41, 1765.

Double Layer Capacitance of Ordered Porous Carbons (4) Gryglewicz, G.; Machnikowski, J.; Lorenc-Grabowska, E.; Lota, G.; Frackowiak, E. Electrochim. Acta 2005, 50, 1197. (5) Lee, J.; Yoon, S.; Oh, S. M.; Shin, C.-H.; Hyeon, T. AdV. Mater. 2000, 12, 359. (6) Zhou, H.; Zhu, S.; Hibino, M.; Honma, I. J. Power Sources 2003, 122, 219. (7) Fuertes, A. B.; Pico, F.; Rojo, J. M. J. Power Sources 2004, 133, 329. (8) Moriguchi, I.; Nakahara, F.; Yamada, H.; Kudo, T. Electrochem. Solid-State Lett. 2004, 7, A221. (9) Moriguchi, I.; Nakahara, F.; Yamada, H.; Kudo, T. Stud. Surf. Sci. Catal. 2005, 156, 589. (10) Sing, K. S. W. Carbon 1989, 27, 5. (11) Kaneko, K.; Ishii, C.; Ruike, M.; Kuwabara, H. Carbon 1992, 30, 1075.

J. Phys. Chem. C, Vol. 111, No. 1, 2007 233 (12) Lippens, B. C.; de Boer, J. H. J. Catal. 1965, 4, 319. (13) Pikunic, J.; Lastoskie, C. M.; Gubbins, K. E. In Handbook of Porous Solids; Schu¨th, F, Sing, K. S. W., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, Germany, 2002; Vol. 1, Chapter 2. (14) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373. (15) Conway, B. E. J. Electrochem. Soc. 1991, 138, 1539. (16) Barisci, J. N.; Wallace, G. G.; Baughman, R. H. J. Electroanal. Chem. 2000, 488, 92. (17) Kim, C.-H.; Pyun, S.-I.; Kim, J.-H. Electrochim. Acta 2003, 48, 3455. (18) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications, 2nd ed.; John-Wiley & Sons, Inc.: Hoboken, NJ, 2000; Chapter 13.