Electrical double-layer capacitive properties of colloidal crystaltemplated nanoporous carbons

Electrical double-layer capacitive properties of colloidal crystaltemplated nanoporous carbons

Studies in Surface Science and Catalysis 156 M. Jaroniec and A. Sayari (Editors) 9 2005 ElsevierB.V. All rights reserved 589 Electrical double-layer...

683KB Sizes 0 Downloads 3 Views

Recommend Documents

No documents
Studies in Surface Science and Catalysis 156 M. Jaroniec and A. Sayari (Editors) 9 2005 ElsevierB.V. All rights reserved

589

Electrical double-layer capacitive properties of colloidal crystaltemplated nanoporous carbons I. M o r i g u c h i a'b, F. N a k a w a r a a, H. Y a m a d a a and T. K u d o a

aDepartment of Applied Chemistry, Faculty of Engineering, Nagasaki University, 1-14 Bunkyo-machi, Nagasaki, 852-8521, Japan bpRESTO, JST, 4-1-8 Honcho Kawaguchi, Saitama, Japan Nanoporous carbons with both micropores and meso- or macropores were synthesized by a SiO2 colloidal crystal-templating process. The SiO2 opal templates exclusively contributed to the formation of meso- and macropores in carbons. The electrical double-layer capacitance per surface area of the templated porous carbons was much higher than those of commercially available activated carbons with high surface areas. The specific capacitance per surface area originating from meso- and/or macropores of colloidal crystal-derived carbons was estimated to be 20+2 ktF cm -2. The mesoporous carbon with 16 nm pores showed the highest capacitance (190 F s among the non-activated porous carbons previously reported. In addition, the high capacities of porous carbons were maintained even at high chargingdischarging rate. 1. INTRODUCTION Nanostructured porous materials are attracted much attention from the viewpoint of a development of high performance catalysts and electrode materials. Especially a great interest has been devoted to the electrochemical capacitors such as electrical double-layer capacitors (EDLCs) and supercapacitors due to the increasing demand for energy storage devices which can be applied as high power sources for electric vehicles. The EDLC, based on the doublelayer capacitance at the electrode/electrolyte interface, was studied exclusively on porous carbon materials such as activated carbons with high surface area [1-3], and was already applied to practical memory backup systems in electric equipments. However, power and energy densities of the conventional EDLC materials are not enough for the high power sources at the present. Theoretically the higher the electrode surface area and the concentration of electrolyte, the higher value of electrical double-layer capacitance is expected. However, the proportionality between the capacitance and specific surface area of activated carbons are not observed in most of cases because micropores developed in the high surface area carbons cannot be accessible easily to electrolyte and thus will not contribute to the total capacitance of the materials [1,2]. In order to develop high performance EDLC carbon electrode materials, it is of importance to fabricate a porous structure so that the electrolyte solution and ions are accessible smoothly through pores to the large surface of electrode materials. In recent years, mesoporous carbons were successfully synthesized by some templating processes such as MCM-48 template method [4-9], and relatively high electric double layer capacitive properties were reported [6,10,11 ]. However, the porous structure is not optimized yet sufficiently for the electrochemical mass transport. We have already reported that macroporous carbons synthesized by a colloidal crystal-templating process showed a highly

590 efficient electrical double-layer capacitive property [12]. In the present study, a correlation between the porous structure of colloidal crystal-derived carbons and electrochemical doublelayer capacitive properties was investigated in detail. 2. EXPERIMENTAL 2.1. Synthesis of porous carbons SiO2 colloidal solutions (Spherica slurry 120, Cataloid SI-45P and Cataloid S-30H) with the average particle diameter of 120, 45 and 16 nm were kindly supplied by Catalysts&Chemicals Ind. Co., Ltd. SiO2 opal crystals were obtained by a centrifugation of the colloidal solutions at 2000 rpm for 2 h and drying in vacuo for 1 day. 8 g of the opal was slowly immersed into a mixture solution of 6.5 g of phenol, 4.8 g of 37 wt% aqueous formaldehyde and a slight amount of 35wt% aq. HCI for l day. After separated from the solution, the opal was heated in an oven at 400 K for 12 h in air to produce a phenolic (novolac) resin network in the interstitial space of the SiO2 opal, and then was subjected to a carbonization treatment in an Ar gas flow (100 mL mini), the heating temperature of which was increased up to 1273 K with a heating rate of 5 K rain -1 and was kept at 1273 K for 5 h. The SiO2 template was removed from the composite by an HF etching with 46 wt% aqueous HF for 3 h and the obtained carbon material was dried in vacuo for 1 day. In the following, the carbons obtained by using the colloidal crystal templates were referred to as Carbon-x, where x indicates the average diameter of SiO2 particles. A non-templated carbon, denoted as Carbon-non, was also synthesized with the same manner from the phenolic resin directly. 2.2. Characterization

TG measurements of carbon/SiO2 composite and carbon materials were performed on Seiko Instruments Inc. TG/DTA6200 with a heating rate of 5 K min -~ in air. The carbonization state of porous carbons was investigated by elemental analysis (Perkin Elmer 240011 analyzer), X-ray powder diffraction (XRD) measurement (Rigaku RINT2200 diffractometer using CuKot radiation) and Raman spectroscopy (Renishaw Ramanscope System 1000, Ar Laser). The morphology of the porous structure of carbons was observed by transmission electron microscopy (TEM, JEOL JEM-100). Nitrogen adsorption-desorption isotherms of carbons were measured at 77 K on Micromeritics Co. Ltd., Gemini 2370. The total specific surface area was determined by the as-plot analysis using the subtracting pore effect (SPE) method [13]. The surface area originating from meso- and/or macropores was also analyzed by T-plots using a standard isotherm [14]. The surface area of micropores was obtained by subtracting the meso- and/or macroporous surface area from the total surface area determined by the Ots-SPE analysis. The electrical double-layer capacitance of the porous carbon was characterized by cyclic voltammetry in 2.0 mol dm -3 aqueous H2SO4 at room temperature using a three-electrode cell equipped with Pt counter and SCE reference electrodes and an electrochemical analyzer (Hokuto Denko Co. Ltd., HZ-3000). A mixture of the porous carbon (19 mg) and polytetrafluoroethylene (1 mg) with a weight ratio of 95:5 was pressed onto a gold mesh and was used as a working electrode. 3. RESULTS AND DISCUSSION 3.1. Characterization of porous carbons

The carbonization of phenolic resin with heating at 1273 K for 5 h in Ar atmosphere was confirmed by elemental analysis that showed the decrease in H/C atomic ratio with the heat-

591 treatment from 0.93 (phenolic resin heated at 400 K) to 0.04 (Carbon-non; elemental compositions (wt%) were C 75.38%, H 5.82%, N 0.02% for the phenolic resin, and C 91.11%, H 0.32%, N 0.27% for Carbon-non. For the SiO2 opal-templated porous carbons undergone the pyrolysis at 1273 K and the subsequent HF treatment, the H/C atomic ratio was almost the same as that of Carbon-non; for example, H/C atomic ratio was 0.08 (C 88.94%, H 0.32%, N 0.04%) for Carbon-45. This indicates that the carbonization was carried out enough even for the template process. In addition, no SiO2 residue in the finally obtained carbon materials was confirmed by 100 % TG-weight loss with heating in air up to 823 K. On Raman spectra, G- and D-bands due to graphitic and amorphous phases appeared respectively at 1600 and 1350 cm 1 for all the carbonized samples (Fig. 1 shows the spectra of Carbon-non and Carbon-45 as representative examples). Broad XRD peaks were observed at 20/degree = 22 and 43 assignable to the diffraction from (002) and (100) planes of graphite crystal, respectively (Fig. 2) [15]. Assuming the carbons as the micrographitic materials, the crystallite sizes along c- and a-axes were estimated from Scherrer's equation to be 1.2 nm and 2.3 nm, respectively. Figure 3 shows TEM images of templated carbons, Carbon-120 and Carbon-16. Skeletal carbon frameworks consisting of spherical voids, the void size of which is comparable to the diameter of template SiO2 particles, were observed for Carbon-120 and Carbon-45. Spherical void arrays were also confirmed as can be seen in Fig. 3a. However, Carbon-/6 possessed a disordered carbon framework as shown in Fig. 3b. The structural disorderliness was reflected on the surface area as mentioned below. The porous structure of obtained carbons was also characterized from N2 adsorption-desorption isotherms (Fig. 4) by using T-plot and as-plot (SPE) analyses (Table 1). The inflection due to the capillary condensation into meso or macropores was observed on the templated carbons, and it shifted to lower relative pressure with decreasing the template SiO2 particle size in Fig. 4. Specific surface areas determined by the as-plots (SPE) analysis were actually adopted here as the total surface area because the BET method usually applied in the relative pressure of 0.1-0.3 overestimates the value of surface area of microporous materials [13, 16]. The total surface area of Carbon-non, 455 m 2 g-I, was mainly due to micropores since the presence of meso and macropores was not confirmed from the T-plot analysis. The SiO2 opal-templated carbons, on the other hand, possessed meso- or macropores, and the total surface areas were higher than that of Carbonnon, while the surface areas due to micropores were almost the same among the four carbons. The surface area originating from the meso- and/or macropores increased with decreasing the size of SiO2 particle used for the template. These results indicate that the SiO2 opal template

AD.ban d

G.band~,

~

~ W C

k

(100)

_c I

~ k / ~ . t..~ta_.,

!

2000

i

t

i

I

i

1500

i

i

i

I

i

1000

i

I

Carbon-45

i

500

RamanshiftI cm-1 Fig. 1. Raman spectra of Carbon-non and Carbon-45.

9

10

20

I

30

I

40

I

50

I

60

70

2t) / d e g r e e ( C u K ~

Fig. 2. XRD pattems of Carbon-non and Carbon-45.

592

Fig. 3. TEM images of (a) Carbon-120 and (b) carbon-16. contributed exclusively to the formation of meso and macropores in carbons. The value of specific surface area originating from meso or macropores of Carbon-45 was 2.8 times higher than that of Carbon-120. If the Carbon-non 455 n.d. 455 template opals with different sphere Carbon-120 662 262 400 size have the same closed packing Carbon-45 1130 737 394 state and the template structures are Carbon-/6 1358 980 378 ideally replicated on the inverted a determined from as plots, b determined from T-plots, opals, the ratio of specific surface c Sa. . . . -- Satotal- Sa......... ; surface area determined areas per weight (m2g -l) of inverted from the as-plot was used as Sato~ for the calculation opals must be theoretically consistent with the reciprocal of the ratio of particle diameters; the theoretical specific surface area can be expressed as 4rc~/(2~/2 d3x0.26xp), where d and p are pore diameter and carbon density in the wall, respectively. Comparing Carbon-120 and Carbon-45. the diameter ratio of template SiO2 spheres was 1/2.7 (= 45 nm/120 nm) and its 2500 reciprocal value was in very good agreement with the ratio of surface area originating from meso- and 2000' macropores, demonstrating an effective replication of the SiO2 opal structure on the porous carbons. The density of carbon constructing pore wall of ~ 1500 o Carbon-120 and Carbon-45 was estimated to be ca. -,,.., 0.55 g cm -3, which is much lower than the reported carbon density (1.8-2.25 g cm 3) [17]. The quite ~1000 low density of present carbons is ascribable to the z formation of a lot of micropores in the wall. On the other hand, the surface area of Carbon-/6 was much 500 lower than the value expected from the SiO2 particle a on non diameter (calculated mesopore surface area: ca. 1900 0 0.2 0.4 0.6 0.8 1 m 2 g-l) under the assumption that the carbon wall 0 Relative pressure, P / P o has the same carbon density as above discussion. This is due to the disordered porous structure as Fig. 4. N2 adsorption-desorption isotherms shown in the TEM image (Fig. 3b). of synthesized porous carbons. Table 1. Surface areas of colloidal crystal-derived porous carbons determined by as-plot and T-plot analyses. Sample Satotal , Sameso macro, Samlcro, m 2 g-1 a m 2 g-lb m 2 g-1 c

593 200 3OO

--.

8 E

/ /

2OO LL

Carbon-16

180

Carbon-16/~

160

/ /

"7

100

Carbon-45 /

140 u_ "- 120

0

/

80

-200

Carbon-120

m

O

60

9

0

,

.

!

0.2 0.4 0.6 Potential / (V vs Ag/AgCI)

Fig. 5. CVs of synthesized carbons in 2 M aqueous H2SO4 in the potential range from -0.1 to +0.7 V (vs. Ag/AgC1) at a sweep rate of 1 mVs1. The arrows indicate the points adopted as the capacitance values.

crn-=

/

40

0.8

20-J:2pF

~/ /

-300 -0.2

i//~

100

-loo o

/0

/

O

/ /

2O

~lJ Carbon-non

i

00

i

' "

i

i

,

,

500

i

,

,

i

1000

,

1500

Total s u r f a c e a r e a / m 2 gq

Fig. 6. The capacitance (F g4)vs, total surface area plots of synthesized carbons.

3.2. Electrical double-layer capacitance of templated porous carbons All of carbons synthesized here showed rectangular-shaped CV curves due to capacitive charging and discharging to electric double-layer at 1 mV s-1 (Fig. 5). Weak cathodic and anodic peaks were additionally observed around 0.3 and 0.45 V vs Ag/AgC1, respectively, maybe implying a little contribution of faradaic process. The Carbon-non possessed a quite low capacitance (9.8 F g-I) and a typical specific capacitance of microporous carbons (2.2 ~tF cm -2) as reported [2]. In contrast, the templated porous carbons showed extremely high capacity in comparison with Carbon-non; capacitances at 0.3 V vs. Ag/AgC1 in the positive scan of Carbon-120, Carbon-45 and the capacitance at 0.45 V vs. Ag/AgC1 of Carbon-16 were 64, 127 and 190 F g-] (specific capacitance: 9.7, 11.2 and 14 laF cm-2), respectively. The capacitance values of Carbon-16 and Carbon-45 are higher than those of commercially available activated carbons with very high surface area (27-100 F g]) [2] and mesoporous carbons (95 F g-~) [ 10]. The surface areas originating from micropores are almost the same in the four carbons synthesized here (Table 1), thus the difference of capacitance is ascribable to the contribution from the surface of meso and/or macropores. As shown in Figure 6, the capacitance of carbons increases linearly with the total surface area and the intercept of the line to x-axis is about 400 m 2 g-I, which is good agreement with the micropore surface area of the templated porous carbons. In C~s-plots of N2 Carbon-16 adsorption isotherms, micropore filling (f-swing), which can be usually observed for microporous ~ ') 100 materials with the pore size below 0.7 nm, was confirmed for the present carbons. Since the hydrodynamic diameter of SO42- ion in aqueous solution can be estimated to be 0.41 nm from the 10 o Stokes equation, the micropore size is smaller than the twice of the ion diameter which is needed to form electrical double layers throughout at parallel-sided porous surfaces at 1 10 100 1000 S c a n r a t e / m V s -1 least. Therefore, contribution of the micropore surface to the capacitance would be very small. Actually it was reported that the small micropores Fig. 7. Sweep rate-dependent capacities of synthesized carbons. with the size below 0.7 nm could not be

c.;b

o E

o

1

Illll

I

I

I IlJlJl

I

I

I lllfd

I

I

I llllt

594 accessible to electrolyte solution and they will not contribute to the electrical double-layer capacitance [2,18]. Considering above discussions, the real specific capacitance of carbons synthesized here was estimated to be 20+2 ~tF cm -2 from the slope of line in Figure 6. This value is close to the ideal specific capacitance of carbon discussed in references [ 1, 2]. Figure 7 shows the sweep rate-dependent capacitance of the four carbons synthesized here. The capacitances decreased gradually with increasing the sweep rate especially above 50 mV s-1, because the rate of charging-discharging to electric double layer cannot follow the fast sweep rate which causes a depression of the rectangular CV profiles. The depression of CV profiles would be due to a relaxation dependent on an RC time constant of the present electrode/electrolyte system [10]. However the high capacitance value above 100 F g-1 was confirmed for the Carbon-45 and Carbon-/6 upto 100 mV s-1. These results demonstrate that the colloidal crystal-derived nanoporous carbons are suitable as high rate mass transportable electrode materials. CONCLUSIONS The electric double-layer capacitance per surface area of the porous carbons, which were synthesized by the colloidal crystal-templating process, was much higher than that of commercial available activated carbons, and especially Carbon-16 possessed the highest capacitance among the non-activated porous carbons previously reported to our best knowledge. It was also found that the meso- and macropores generated in the porous carbons are suitable for a high rate transportation of electrolyte ions at the electrode interface. Therefore the colloidal crystal-templating process is useful for designing and tuning high performance EDLC carbon materials. ACKNOWLEDGEMENT This work was in part supported by a Grant-in-Aid for Scientific Research from Ministry of Education, Culture, Science, Sports and Technology of Japan, and The Science Technology Foundation of Ishikawa Carbon. The study made use of instruments (elementary analysis, XRD, and TEM) in the Center for Instruments Analysis of Nagasaki University. REFERENCES [1] E. Frackowiak, F. Beguin, Carbon, 39 (2001) 937. [2] D. Qu, H. Shi, J. Power Sources, 74 (1998) 99. [3] J. Gamby, P.L.Taberna, P.Simon, J.F.Fauvarque, M. Chesneau, J. Power Souces, 101 (2001) 109. [4] R. Ryoo, S. H. Joo, S. Jun, J. Phys. Chem. B, 103 (1999) 7743. [5] S. Che, K. Lund, T. Tatsumi, I. Iilima, S. H. Joo, R. Ryoo, O. Terasaki, Angew. Chem. Int. Eng., 42 (2003) 2182. [6] J. Lee, S. Yoon, T. Hyeon, S. M. Oh, K. B. Kim, Chem. Commun., (1999) 2177. [7] I. Moriguchi, Y. Koga, R. Matsukm'a, Y. Teraoka, M. Kodama, Chem. Commun., (2002) 1844. [8] Z. Lei, Y. Zhang, H. Wang, Y. Ke, J. Li, F. Li, J. Xing, J. Mater. Chem., 11 (2001) 1975. [9] J-S, Yu, S. Kang, S. B. Yoon, G. Chai, J. Am. Chem. Soc., 124 (2002) 9382. [10] H. Zhou, S. Shu, M. Hibino, I. Honma, J. Power Sources, 122 (2003), 219. [11] J. H. Jang, S. Han, T. Hyeon, S. M. Oh, J. Power Sources, 123 (2003) 79. [ 12] I. Moriguchi, F. Nakahara, H. Yamada, T. Kudo, Electrochem. Solid State Lett., 7 (2004) A221. [13] K. Kaneko, C. Ishii, M. Ruike, H. Kuwabara, Carbon, 30 (1992) 1075. [14] J.C.P. Broekhoff, B.G. Linsen, "Physical and Chemical Aspects of Adsorbents and Catalysts", Chap. 2, Academic Press (1970). [15] Y. Liu, J. S. Xue, T. Zheng, J. R. Dahn, Carbon, 34 (1996)193. [16] K. Kaneko, C. Ishii, Colloids and Surfaces, 67 (1992) 203. [17] D. R. Lide, 75th Edition of CRC Handbook of Chemistry and Physics, CRC press Inc. (1994). [ 18] S. Shiraishi, H. Kurihara, H. Tsubota, A. Oya, S. Soneda, Y. Yamada, Electrochem. Solid State Lett., 4 (2001) A5.