Electrochemical Capacitances of Well-Defined Carbon Surfaces

Aleksandra Pacuła , Robert P. Socha , Małgorzata Zimowska , Małgorzata ... Minzhen Cai , Ronald A. Outlaw , Sue M. Butler , John R. Miller. Carbon ...
0 downloads 0 Views 242KB Size
9086

Langmuir 2006, 22, 9086-9088

Electrochemical Capacitances of Well-Defined Carbon Surfaces Taegon Kim,†,‡ Seongyop Lim,‡ Kihyun Kwon,† Seong-Hwa Hong,‡ Wenming Qiao,‡ Choong Kyun Rhee,*,† Seong-Ho Yoon,*,‡ and Isao Mochida‡ Department of Chemistry, Chungnam National UniVersity, Daejeon, 305-764, South Korea, and Institute for Materials Chemistry and Engineering, Kyushu UniVersity, Fukuoka, 816-8580, Japan ReceiVed May 16, 2006. In Final Form: June 12, 2006 Reported is the capacitive behavior of homogeneous and well-defined surfaces of pristine carbon nanofibers (CNFs) and surface-modified CNFs. The capacitances of the well-defined CNFs were measured with cyclic voltammetry to correlate the surface structure with capacitance. Among the studied pristine CNFs, the edge surfaces of platelet CNFs (PCNF) and herringbone CNFs were more effective in capacitive charging than the basal plane surface of tubular CNF by a factor of 3-5. Graphitization of PCNF (GPCNF) changed the edge surface of PCNF into a domelike basal plane surface, and the corresponding capacitances decreased from 12.5 to 3.2 F/g. A chemical oxidation of the GPCNF, however, recovered a clear edge surface by removal of the curved basal planes to increase the capacitance to 5.6 F/g. The difference in the contribution of the edge surface and basal-plane surface to the capacitance of CNF was discussed in terms of the anisotropic conductivity of graphitic materials.

To correlate the electrochemical capacitance with the properties of carbon, specific surface area and pore size distribution have been widely examined as the most common factors to govern the capacitance.1,2 However, such a correlation is not always consistent with various carbon materials. A carbon material with rather lower surface area exhibits higher capacitance compared to the conventional activated carbons with very large surface area.3 It has been believed, on the other hand, that the capacitances, especially per volume, in the practical capacitors are dependent on the wall surfaces of pores in the examined carbon materials.2a-c,3 In light of elucidating the relation between the wall surface structure inside a pore and capacitance, the electrochemical performances of the basal and edge surfaces of graphite have been measured to distinguish their capacitive contributions in a quantitative way.4 Although the results were diverse, it was clearly revealed that the graphene alignments around pores, and thereby the wall surface structures inside pores, appeared to influence the capacitance significantly. Such a divergence in the results, especially concerning the surfaces of carbon materials as exemplified as above, may come from the absence of a model material whose properties are controlled uniformly and characterized thoroughly. In this sense, carbon nanofibers (CNFs) can be ideal models in understanding the correlations between the microscopic structures of carbon materials and their macroscopic properties in various applications * To whom correspondence should be addressed. E-mail: ckrhee@ cnu.ac.kr; tel.: +82-42-821-5483; fax: +82-42-821-8896 (C.K.R.). Email: [email protected]; tel.: +81-92-583-7801; fax: +81-92-5837798 (S.H.Y.). † Chungnam National University. ‡ Kyushu University. (1) Nishino, A. J. Power Sources 1996, 60, 137-147. (2) (a) Nishino, A. In Technologies and Materials for EDLC and Electrochemical Supercapacitors; Nishino, A., Naoi, K., Eds.; CMC: Tokyo, 2003; p 129. (b) Lin, C.; Ritter, J. A.; Popov, B. N. J. Electrochem. Soc. 1999, 146, 3639-3643. (c) Kierzek, K.; Frackowiak, E.; Lota, G.; Gryglewicz, G.; Machnikowski, J. Electrochim. Acta 2004, 49, 515-523. (d) Shi, H. Electrochim. Acta 1996, 41, 1633-1639. (e) Teng, H.; Chang, Y.; Hsieh, C. Carbon 2001, 39, 1981-1987. (3) Lee, S. I.; Mitani, S.; Park, C. W.; Yoon, S. H.; Korai, Y.; Mochida, I. J. Power Sources 2005, 139, 379-383. (4) (a) Randin, J.-P.; Yeager, E. J. Electroanal. Chem. 1972, 36, 257-276. (b) Robinson, R. S.; Sternitzke, K.; Mcdermott, M. T.; Mccreery, R. L. J. Electrochem. Soc. 1991, 138 (8), 2412-2418. (c) Zaghib, K.; Nadeau, G.; Kinoshita, K. J. Power Sources 2001, 97-98, 97-103. (d) Randin, J.-P.; Yeager, E. J. Electroanal. Chem. 1975, 58, 313-322.

such as capacitors. This particular choice could be justified, since the CNFs are well-defined and thoroughly characterized with transmission electron microscopy (TEM), scanning tunneling microscopy (STM) and Raman spectroscopy.5 Furthermore, it has been demonstrated that their surface structures were modified with a heat treatment and oxidation in a systematic and reproducible way, along with detailed characterization with TEM and STM.5a This unique preparative method rationalizes the use of surface-modified CNFs as model materials as well. In this work, presented are the electrochemical capacitances, measured by cyclic voltammetry, of typical CNFs and their manipulated ones. The CNFs, basically described by their diameters and graphene alignment, were chosen to correlate their capacitances with their surface structures, and the measured capacitance reflects a straightforward correlation between the surface graphitic structure and the capacitive performance. The carbon materials used in this study are classified into two groups: one is a group of pristine CNFs such as platelet (PCNF), herringbone (HCNF), and tubular (TCNF) types, whose graphene alignments are different from each other, and the other one is a series of PCNFs5a modified to manipulate their surface structures (see Supporting Information). Figure 1 shows TEM images of pristine CNFs and surfacemodified PCNFs. The surfaces of the pristine PCNFs and HCNFs were characterized by the graphitic edges, as the graphitic layers in the fibers were aligned in a way perpendicular and declined to the fiber axis, respectively, as shown in Figure 1a,b. In contrast, the TCNF, whose graphene alignment was parallel to the fiber axis, was dominantly wrapped by basal planes (Figure 1c). When the pristine PCNF (Figure 1a) was graphitized (GPCNF), the edge of the graphenes became covered with domelike basal planes, as clearly shown in Figure 1d. Ball-milling of the GPCNF (GPCNF-M) induced a transformation of the domelike basal plane and its alignment to reduce the interunit and intraunit tension caused by the nanosized curvature.5a The surface of the GPCNFM, however, remained covered with the domelike basal planes, as in Figure 1e. Hence, GPCNF and GPCNF-M were perfectly (5) (a) Lim, S.; Yoon, S. H.; Mochida, I.; Chi, J. H. J. Phys. Chem. B 2004, 108 (5), 1533-1536. (b) Yoon, S. H.; Lim, S.; Hong, S. H.; Mochida, I.; An, B.; Yokogawa, K. Carbon 2004, 42 (15), 3087-3095. (c) Yoon, S. H.; Lim, S.; Hong, S. H.; Qiao, W. M.; Whitehurst, D. D.; Mochida, I.; An, B.; Yokogawa, K. Carbon 2005, 43 (9), 1828-1838.

10.1021/la061380q CCC: $33.50 © 2006 American Chemical Society Published on Web 09/21/2006

Letters

Langmuir, Vol. 22, No. 22, 2006 9087

Figure 2. SEM images of an HOPG surface polished mechanically in a perpendicular direction to a basal plane. The magnification increases sequentially from a to d.

Figure 1. Typical TEM images of surfaces of well-defined CNFs and their corresponding models: (a) PCNF, (b) HCNF, (c) TCNF, (d) GPCNF, (e) GPCNF-M, and (f) GPCNF-NA. The white line in each image represents a scale bar of 5 nm.

covered by the basal plane without the edge of the graphenes being exposed. An oxidation of the GPCNF (GPCNF-NA), using 10% HNO3 at room temperature for 24 h, removed the curved basal-plane surface of the GPCNF to expose the edges (Figure 1f) more clearly than observed on the pristine PCNF, with a few remnants of protruding loop ends. Here, it was remarkable that the surface modification of the PCNF did not alter the fiber body. The surface modification processes described above were unique in terms of controlling the surface of PCNF to manipulate its physicochemical properties. Highly oriented pyrolytic graphite (HOPG) has been exclusively used to differentiate the effect of the edge and basal plane of graphitic materials.4,6 A specific reason for the particular use of HOPG was based on an assumption that a plane perpendicular to a basal plane surface would be an edge surface. In reality, however, a surface mechanically polished in a direction vertical to the basal plane of HOPG, as shown in Figure 2, was significantly far from an ideal surface of a graphite edge. In other words, the polished surface of protruding graphite layers was not free from exposure of the graphitic basal planes to an experimental environment, especially to an electrolyte. Therefore, the choice of well-defined CNFs as model materials in this work could be justified further. The left panels of Figure 3 show the cyclic voltammograms of the pristine CNFs (see Supporting Information for the electrochemical measurement). The voltammograms within the studied potential region were flat rectangles superimposed with a small broad reversible redox couple. Therefore, the capacitive behavior of the CNFs resulted mainly from electrochemical double-layer charging along with a negligible contribution of pseudocapacitance. The capacitance values of the pristine CNFs were measured from the currents in the cyclic voltammograms (6) Katagiri, G.; Ishida, H.; Ishitani, A. Carbon 1988, 26, 565-571.

Figure 3. Cyclic voltammograms of well-defined CNFs in 0.5 M H2SO4 solution: (a) PCNF, (b) HCNF, (c) TCNF, (d) GPCNF, (e) GPCNF-M, and (f) GPCNF-NA. The scan rate was 10 mV/sec. Table 1. Electrochemical Double-Layer Capacitances of Well-Defined CNFs pristine CNFs

capacitance (F/g)

modified PCNFs

capacitance (F/g)

PCNF HCNF TCNF

12.5 ( 1.2 23.4 ( 2.9 4.5 ( 0.6

GPCNF GPCNF-M GPCNF-NA

3.1 ( 0.2 3.3 ( 0.7 5.6 ( 0.4

at 0.45 V and tabulated in Table 1. It is obvious that the capacitances of the edge surfaces of PCNF and HCNF were higher than that of the basal plane surface of TCNF by a factor of 3-5. The cyclic voltammograms of the surface-modified PCNFs, as shown in the right panels of Figure 3, evidently support the dependence of the capacitance on the surface structure of CNFs. When the edge surface of PCNF was changed to the domelike basal plane surface of GPCNF, the electrochemical double-layer charging current was dramatically reduced, as in Figure 3d. GPCNF-M, verified to have a basal plane surface, consistently showed a capacitive behavior similar to that of GPCNF. The capacitive current of GPCNF-NA, obtained upon the chemical

9088 Langmuir, Vol. 22, No. 22, 2006

oxidation of GPCNF, was recovered significantly, but not enough to reach the value of the pristine PCNF. The capacitances of the surface-modified PCNFs, as summarized in Table 1, were coherent with the results of the pristine CNFs in terms of surface structure. In addition, the voltammograms of GPCNF and GPCNF-M after an electrochemical oxidation at 1.6 V were comparable to the one in Figure 3f, which indicates that the domelike basal planes of the particular CNFs were electrochemically oxidized off to make the surfaces similar to that of GPCNFNA. It is noteworthy that, after an oxidation of GPCNF, regardless of the chemical or electrochemical method, oxygenated surface species appeared, as discernible with a broad redox couple at ∼0.3 V. Because the contribution of the pseudocapacitance to the total capacitance was still less than 10%, the recovery of the capacitance after the oxidation was significantly contributed by the readvent of a graphitic edge surface, that is, the electrochemical double-layer charging on the edge surface. The observed results support the idea that the higher capacitance values of the CNFs were evidently correlated with the presence of the graphitic edges at the surfaces of the carbon materials. Definitely, the higher capacitance was led from building up a higher electrical charge at the surface; thus, the graphitic edges at the surface were more efficient in charging the surfaces of the carbon materials under an electrochemical polarization. This particular high efficiency of the edges may come from the wellknown anisotropic conductivity, that is, the higher electrical conduction in the parallel direction to the basal planes relative to that in the perpendicular direction. Under a polarization condition, the π-electrons move more easily along the graphitic planes to the edges, so that more charge-build-up, and thus more capacitance, is effectively developed on the edge surfaces of the carbon materials such as PCNF, HCNF, and chemically oxidized GPCNF-NA. A close examination of the above results implies that the diameter of CNFs may be another factor influencing the electrochemical capacitance. The diameters of the modified

Letters

PCNFs were constant because one PCNF was modified in a consecutive way, so that only the effect of the surface structure was confirmed again. In the set of the pristine CNFs, however, the diameters of the CNFs were different: the approximate average diameters of the PCNF and HCNF were 150 and 200 nm, respectively. Phenomenologically, the larger diameter would lead to a larger capacitance under an identical surface structure, which would be the subject of further research. In summary, the capacitive behavior of well-defined carbon surfaces, using three types of pristine CNFs and a series of surfacemodified PCNFs, were presented to correlate with electrochemical capacitance. The CNFs with edge surfaces were more efficient in electrochemical capacitive charging than the CNFs with basalplane surfaces by a factor of 3-5. The particularly high capacitance values of the edge surfaces of CNFs were explained in terms of anisotropic electrical conductance in graphitic materials. One of the achievements of this work was to differentiate the contributions of the edge and basal-plane to the electrochemical capacitance of graphitic materials using welldefined CNFs. Also, it should be underscored that, for capacitor or capacitor support, it would be advantageous to use or develop carbon materials having more edge surfaces, even in the pores. Acknowledgment. The authors in Japan (S.L., S.-H.H., W.Q., S.-H.Y., and I.M.) acknowledge the financial support of the Japan Science and Technology Corporation (JST) through the CREST program. The work carried out in Korea was funded by the Korea Sanhak Foundation in 2005. T.K. was supported by a Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (KRF- 2005-213-C00027). Supporting Information Available: Experimental details in the preparation of the CNFs and the electrochemical measurement of capacitance. This material is available free of charge via the Internet at http://pubs.acs.org. LA061380Q