Enhancement Mechanism of Electrochemical Capacitance in Nitrogen

pubs.acs.org/Langmuir © 2009 American Chemical Society

Enhancement Mechanism of Electrochemical Capacitance in Nitrogen-/ Boron-Doped Carbons with Uniform Straight Nanochannels Taeri Kwon, Hirotomo Nishihara,* Hiroyuki Itoi, Quan-Hong Yang, and Takashi Kyotani Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan Received April 14, 2009. Revised Manuscript Received June 18, 2009 Anodic aluminum oxide (AAO) with uniform straight nanochannels was completely coated with pure, N-doped, or B-doped carbon layer. Their electric double layer capacitances are measured in aqueous (1 M sulfuric acid) and organic (1 M Et4NBF4/polypropylene carbonate) electrolyte solutions in order to investigate the capacitance enhancement mechanisms caused by N- or B-doping. Since the three types of carbon-coated AAOs (pure, N-doped, or B-doped) have exactly the same pore structure, the observed capacitance enhancement was ascribable to only the following factors: (i) better wettability, (ii) the decrease of equivalent series resistance, (iii) the contribution of space-charge-layer capacitance, and (iv) the occurrence of pseudocapacitance. From the measurements of the wettability and the electrical resistance of the coated AAOs together with the electrochemical investigation (the cyclic voltammetry, the galvanostatic charge/discharge cycling, and the impedance analysis), it is concluded that the pseudocapacitance through faradic charge transfer (factor iv) is the most important factor to enhance the capacitance by N- or B-doping. This can be applied to not only the present carbon-coated AAOs but also any other porous carbons.

Introduction

(ii)

An electric double layer capacitor (EDLC) is an electrical storage device that can be repeatedly charged and discharged like secondary batteries. EDLC has many advantages, such as high power density, long cycle life, and rare-metal free construction, compared with secondary batteries.1-3 However, the energy density of EDLC is much lower than that of secondary batteries. One of the possible ways to increase the energy density is the doping of heteroatoms, such as nitrogen and boron, into a porous carbon electrode. So far, many research groups have reported a positive effect of N-doping in both aqueous and organic electrolyte solutions4-14 and proposed the following possible mechanisms for the capacitance enhancement. (i) The improvement of electrode wettability, due to the increase in the number of hydrophilic polar sites (especially for aqueous electrolyte systems).13-16 *[email protected]. (1) Conway, B. E. Electrochemical Supercapacitors, Scientific Fundamentals and Technological Applications; Kluwer Academic/Plenum Press: New York, 1999. (2) Frackowiak, E.; Beguin, F. Carbon 2001, 39, 937–950. (3) Pandolfo, A. G.; Hollenkamp, A. F. J. Power Sources 2006, 157, 11–27. (4) Ania, C. O.; Khomenko, V.; Raymundo-Pinero, E.; Parra, J. B.; Beguin, F. Adv. Funct. Mater. 2007, 17, 1828–1836. (5) Frackowiak, E.; Lota, G.; Machnikowski, J.; Vix-Guterl, C.; Beguin, F. Electrochim. Acta 2006, 51, 2209–2214. (6) Jurewicz, K.; Babel, K.; Ziolkowski, A.; Wachowska, H.; Kozlowski, M. Fuel Process. Technol. 2002, 77, 191–198. (7) Kim, N. D.; Kim, W.; Joo, J. B.; Oh, S.; Kim, P.; Kim, Y.; Yi, J. J. Power Sources 2008, 180, 671–675. (8) Kodama, M.; Yamashita, J.; Soneda, Y.; Hatori, H.; Kamegawa, K. Carbon 2007, 45, 1105–1107. (9) Kodama, M.; Yamashita, J.; Soneda, Y.; Hatori, H.; Nishimura, S.; Kamegawa, K. Mater. Sci. Eng. B 2004, 108, 156–161. (10) Jurewicz, K.; Babel, K.; Ziolkowski, A.; Wachowska, H. Electrochim. Acta 2003, 48, 1491–1498. (11) Beguin, F.; Szostak, K.; Lota, G.; Frackowiak, E. Adv. Mater. 2005, 17, 2380–2384. (12) Hulicova, D.; Yamashita, J.; Soneda, Y.; Hatori, H.; Kodama, M. Chem. Mater. 2005, 17, 1241–1247. (13) Hulicova, D.; Kodama, M.; Hatori, H. Chem. Mater. 2006, 18, 2318–2326. (14) Lota, G.; Grzyb, B.; Machnikowska, H.; Machnikowski, J.; Frackowiak, E. Chem. Phys. Lett. 2005, 404, 53–58. (15) Shiraishi, S.; Mamyouda, H. Tanso 2008, 232, 61–66. (16) Kawaguchi, M.; Itoh, A.; Yagi, S.; Oda, H. J. Power Sources 2007, 172, 481–486.

Langmuir 2009, 25(19), 11961–11968

The decrease of equivalent series resistance (ESR) of a capacitor cell by the increase of the carbon electric conductivity due to the introduction of electrons as a carrier.17,18 (iii) The occurrence of space-charge-layer capacitance in carbon by the increase of its electron density.1,15 (iv) The occurrence of pseudocapacitance through faradic charge transfer, because the nature of carbon becomes electron donor.11,12 Of course, in all actual cases, the observed capacitance may be more or less a combined result of these factors, and it is therefore not easy to reveal the contribution of each factor. In addition, carbon pore structure is inevitably altered during N-doping process and such change makes it difficult to assess the sole effect of N-doping, because the pore structure is also one of the crucial factors affecting EDLC performance. While N-doping introduces electrons into carbon as a carrier, B-doping may introduce holes, and in addition, B-doped sites would be more or less polar. Consequently, B-doping may also be effective to increase EDLC performance. Indeed, Shiraishi et al. and Cheng et al. have recently reported the enhancement effect by B-doping,19,20 but also in this case, its mechanism remains almost unknown. For the understanding of the mechanism of the capacitance enhancement by N- and B-doping, the use of doped and undoped carbons with exactly the same pore structure must be very effective. A hard template method21-23 is a powerful tool for the synthesis of carbons with uniform pore structure. Several researchers (17) Derradji, N. E.; Mahdjoubi, M. L.; Belkhir, H.; Mumumbila, N.; Angleraud, B.; Tessier, F. Thin Solid Films 2005, 482, 258–263. (18) Kim, J.; Choi, M.; Ryoo, R. Bull. Korean Chem. Soc. 2008, 29, 413–416. (19) Shiraishi, S.; Kibe, M.; Yokoyama, T.; Kurihara, H.; Patel, N.; Oya, A.; Kaburagi, Y.; Hishiyama, Y. Appl. Phys. A 2006, 82, 585–591. (20) Wang, D. M.; Li, F.; Chen, Z. G.; Lu, G. Q.; Cheng, H. M. Chem. Mater. 2008, 20, 7195–7200. (21) Kyotani, T.; Nagai, T.; Inoue, S.; Tomita, A. Chem. Mater. 1997, 9, 609– 615. (22) Kyotani, T.; Tsai, L. F.; Tomita, A. Chem. Mater. 1995, 7, 1427–1428. (23) Kyotani, T.; Tsai, L. F.; Tomita, A. Chem. Mater. 1996, 8, 2109–2113.

Published on Web 09/11/2009

DOI: 10.1021/la901318d

11961

Article

Kwon et al.

have synthesized N- and B-doped carbon replicas with uniform mesopores by using mesoporous silicas as templates5,7,20,24 and examined their EDLC performances. However, such carbons contain not only uniform mesopores but also a large amount of random micropores that significantly affect the capacitance behavior. Since it is almost impossible to control such a micropore structure in the mesoporous templated carbons, they are not suitable as a model porous material for the investigation of the effect of heteroatom doping. In this work, instead of the usual templated carbons, we use a carbon-coated anodic aluminum oxide (AAO) film as a model material. When an aluminum plate is anodically oxidized in an acid electrolyte, an AAO film is formed: its porosity consists of an array of parallel and straight nanochannels with a uniform diameter. We have demonstrated that an AAO film can be coated with thin carbon layer by chemical vapor deposition (CVD) of propylene, where the carbon deposition was surprisingly uniform and fully covered the whole surface of the film including the inner walls of the nanochannels (Figure 1).22,23 The carbon-coated AAO film (C/AAO) thus obtained is completely free of such random micropores as observed in the mesoporous carbons prepared by the template method. Due to the perfect carbon coating, C/AAO is electrically conductive and can be used as an electrode for EDLC.25 Moreover, nitrogen atoms can easily be introduced into the carbon layer through acetonitrile CVD.26,27 The resulting N-doped carbon-coated AAO film (N/AAO) has the same pore structure as C/AAO, and it is thus possible to exclude the effect of pore structure when their capacitance values are compared. By using the present CVD approach, it is even possible to prepare B-doped carbon-coated AAO film (B/AAO).28,29 Generally, the synthesis of B-doped porous carbons is much more difficult than the case of N-doped carbons. The previous report on the B-doping effect was therefore limited on B-doped carbon nanotubes (B/C is 0.009-0.011)19 and templated mesoporous carbons (B/C is 0.002-0.007),20 both of which contain a very small amount of boron. In contrast, the B/AAO contains a larger amount of boron in the carbon layer,29 thereby making the effect of B-doping more remarkable. In this work, we prepare C/AAO, N/AAO, and B/AAO, each of which has the same pore structure, and use these films as model porous materials for the investigation of the above-mentioned four factors (wettability, electric conductivity, space-charge-layer capacitance, and pseudocapacitance). From the obtained results, we discuss how each factor contributes to the capacitance enhancement in porous carbons in general.

Experimental Section Sample Preparation. Several AAO films were prepared by the anodic oxidation of an aluminum plate at a cell voltage of 20 V in 20 wt % sulfuric acid (10 °C) for 2 h. Then, the AAO films were separated from the aluminum substrate by reversing the polarity of the cell voltage. The thickness of the AAO films thus obtained was about 90 μm, and their pore size was about 24 nm. Note that the film thickness almost corresponds to the pore length. Most of (24) Kodama, M.; Yamashita, J.; Soneda, Y.; Hatori, H.; Kamegawa, K.; Moriguchi, I. Chem. Lett. 2006, 35, 680–681. (25) Jung, M.; Kim, H. G.; Lee, J. K.; Joo, O. S.; Mho, S. Electrochim. Acta 2004, 50, 857–862. (26) Xu, W. H.; Kyotani, T.; Pradhan, B. K.; Nakajima, T.; Tomita, A. Adv. Mater. 2003, 15, 1087–1090. (27) Yang, Q. H.; Xu, W. H.; Tomita, A.; Kyotani, T. Chem. Mater. 2005, 17, 2940–2945. (28) Yang, Q. H.; Hou, P. X.; Unno, M.; Yamauchi, S.; Saito, R.; Kyotani, T. Nano Lett. 2005, 5, 2465–2469. (29) Yang, Q. H.; Xu, W. H.; Tomita, A.; Kyotani, T. J. Am. Chem. Soc. 2005, 127, 8956–8957.

11962 DOI: 10.1021/la901318d

Figure 1. Schematic for the synthesis process of the carbon coated AAO: (a) AAO template and (b) carbon-coated AAO. the separated AAO films were ground down into small pieces and the particle size fraction between 25 and 50 μm was sieved out. To prepare three types of coated AAOs (C/AAO, N/AAO, and B/AAO), the resulting powdery AAO was subjected to the following CVD processes. For the preparation of C/AAO and N/AAO, propylene CVD (1.2 vol % under N2 flow at 800 °C for 2 h) and acetonitrile CVD (4.2 vol % under a N2 flow at 800 °C for 5 h) were conducted, respectively. In the case of B/AAO, propylene CVD (1.2 vol % under a N2 flow at 800 °C for 1.5 h) was first conducted on the powdery AAO, followed by a heat treatment (900 °C for 1 h in N2). Then, the second CVD (3.2 vol % benzene and 3.2 vol % boron trichloride under a N2 flow at 725 °C for 20 min) was conducted on the carbon-coated AAO. This second CVD step gave rise to the deposition of B-containing carbon layer on the already-deposited carbon layer. After each CVD, the coated AAOs were finally heat-treated at 900 °C for 1 h in an inert atmosphere. In order to obtain film-shaped coated AAOs, some of the separated AAO films were directly subjected to the above three types of CVD processes without grinding. Characterization. The uncoated and coated AAO powders were directly observed with a field-emission scanning electron microscope (FE-SEM; Hitachi, S-4800). Furthermore, the carbon layers deposited on the AAO nanochannels were mechanically liberated from the powdery AAOs and were observed with a transmission electron microscope (TEM; JEOL, JEM-2010). Nitrogen physisorption measurements were carried out at -196 °C with a volumetric sorption analyzer (BEL Japan, BELSORPmax) to characterize pore structures of the coated AAO powders. Specific surface areas were calculated by the Brunauer-EmmetTeller (BET) method. The total pore volumes were determined from the adsorbed amount at a relative pressure of 0.96. The micropore volumes were evaluated by using the DubininRadushkevich (DR) method. The mesopore size distributions were calculated by applying the Barrett-Joyner-Halenda (BJH) method to the adsorption branches of the N2 isotherms. For the estimation of the amount of the ultramicropores that cannot be detected by N2 physisorption at -196 °C, CO2 physisorption was carried out at 25 °C with the volumetric sorption analyzer. The surface composition and types of N or B functionalities in the carbon layers were characterized with an X-ray photoelectron spectrometer (XPS; Shimadzu, ESCA-3400). In order to estimate the affinity between the coated AAOs and each electrolyte solution (1 M sulfuric acid and 1 M Et4NBF4 in polypropylene carbonate (PC)), the contact angles of these solutions were measured by putting a droplet on each coated AAO film. For further understanding of water affinity, water vapor adsorption isotherms were measured with the volumetric sorption analyzer using the coated AAO powders. To evaluate the resistivity of the carbon layers, the electrical resistances of the coated AAO films were measured by a four-probe technique with a Loresta-GP MCP-T610 equipment (Mitsubishi Chemical Co.).

Preparation and Characterization of Electrode Sheets. The electrochemical performance for the coated AAO powders was evaluated with a three-electrode cell. A working electrode was prepared by mixing each powdery AAO with PTFE (weight ratio is 95:5) without any conductive medium like carbon black. The obtained mixture was cut and formed into a circular sheet with its mass ca. 5 mg (SEM images of the resulting sheets are presented in the Supporting Information). After being sandwiched in a Pt Langmuir 2009, 25(19), 11961–11968

Kwon et al.

Article

mesh, it was pressed at 20 MPa for several seconds to fix the electrode sheet to the mesh. A counter electrode was prepared by mixing activated carbon fiber (Kuraray Chemical Co., Ltd., FT-300-15) with PTFE and carbon black (the weight ratio is 90:5:5; total mass is ca. 10 mg). These two electrodes were immersed into an electrolyte solution (1 M H2SO4 aqueous solution or 1 M Et4NBF4 in PC) and placed in a sealed box under vacuum to impregnate the pores of the electrode materials with the electrolyte solution. The working electrode was then attached to the counter one through two sheets of a polypropylene separator (25 μm in thickness, #2400, Hohsen Corp.) or a cellulose separator (50 μm in thickness, TF-4850, Tokyo Sangyo Yoshi Co. Ltd.), and all of them were fixed by a specially tailored glass holder, thereby keeping the distance between the two electrodes constant at about 0.05-0.1 mm. Cyclic voltammetry (CV) and galvanostatic charge/discharge cycling (GC) measurements were carried out for each sample. The specific gravimetric capacitances were calculated from a cation (Hþ or Et4Nþ) desorption process in the corresponding GC curves. Impedance spectra were also measured with the three-electrode cell in the aqueous and organic electrolyte solutions over a frequency range from 100 mHz to 10 kHz with an amplitude voltage of 10 mV.

Results Structures of the Nanochannels and the Carbon Layers. As shown in Figure 2, uniform nanochannel structure in AAO was observed on the surface of each particle for not only the original AAO but also all the coated AAOs, suggesting that the inner surface of nanochannels in all powdery AAOs was uniformly coated with very thin carbon layer through each CVD method. An attempt was made to directly observe the carbon layers of the coated AAOs with TEM. For this purpose, the coated powders were further ground in a mortar. With this process, some of the carbon layers were partially liberated from the AAO particles, and then, the protruded carbon layers were observed. Figure 3 shows TEM images of such carbon layers, which are molded into a tubular shape of the AAO nanochannels. Note that each tube was embedded in an AAO particle before the grinding, and the internal tube surface corresponds to the inner surface of the nanochannels in the coated AAOs. The tube internal diameter is thus equal to the channel diameter of the corresponding coated AAOs. From Figure 3, the internal diameters of all the tubes and their carbon-layer thicknesses are found to be about 16 and 4 nm, respectively. Though a little thicker carbon layer was observed in very limited region of N/AAO, the internal diameter and the thickness in C/AAO and B/AAO were uniform in every part of the tubes. Since the heat-treatment temperature is not so high (900 °C), the carbon layers in all samples are poorly crystallized and are far from perfect graphite. B/AAO was prepared by the two-step CVD. The tube liberated from B/AAO (Figure 3c) therefore has double coaxial structure that comprises the outer carbon layer formed by the first CVD and the inner B-doped carbon layer formed by the second CVD.29 In order to estimate the thickness of each layer, a carbon-coated AAO was prepared only by the first CVD step, and the carbon-layer thickness was measured with TEM. As a result, its thickness was found to be about 2.8 nm. Thus, the thickness of the B-doped carbon layer can be estimated to be about 1.2 nm, as shown in Figure 3c. Figure 4 shows N2 adsorption/desorption isotherms and pore size distributions of the coated AAO powders. BET surface area and total pore volume were listed in Table 1. Though the BET surface area and the total pore volume of N/AAO are slightly smaller than those of the others (probably due to the thicker carbon deposition), all the coated AAOs basically have similar surface areas and pore volumes. In Figure 4a, all the samples show Langmuir 2009, 25(19), 11961–11968

Figure 2. FE-SEM images of the top views of the pristine and coated AAOs: (a) AAO, (b) C/AAO, (c) N/AAO, and (d) B/AAO.

typical type IV isotherms, indicating that they are mesoporous. Their pore size distributions are almost the same, having a peak at 16 nm (Figure 4b), which agrees well with the pore diameter estimated by the TEM observation (Figure 3). Moreover, each coated AAO has very small micropore volume (Table 1, column 4). We further tried to examine the presence of ultramicropores by CO2 adsorption at 25 °C, because such narrow pores cannot be detected by N2 adsorption at -196 °C. The CO2 adsorption measurement revealed that the pore volume of every sample was no more than 0.01 cm3 g-1 (Table 1, column 5). Thus, it has been confirmed that the coated AAOs do not have any narrow pores. It is also very certain that all the samples do not have any macropores, because their N2 adsorption isotherms never show any increase in the pressure region of p/p0 > 0.96 (Figure 4a). Consequently, we can conclude that these coated AAOs do not have any extra pores (ultramicropore, micropore, and macropores) except for their uniform-sized cylindrical mesopores. They are therefore suitable model porous materials to investigate the effect of surface N- or B-containing species on EDLC performance without considering the effect of pore structure. In order to investigate the hydrophilicity of the coated AAO powders, their water-vapor adsorption isotherms were measured (see the Supporting Information for the isotherm data). The pore volumes occupied by water (Vwater) at p/p0 = 0.95 are shown in Table 1. While the values of Vwater of C/AAO and B/AAO are smaller than the corresponding total pore volumes (Vtotal) detected by the N2 adsorption at -196 °C, Vwater of N/AAO is almost the same as its Vtotal. It is thus clear that N/AAO is more hydrophilic than C/AAO and B/AAO. Chemical forms of the nitrogen/boron species on the surface in N/AAO and B/AAO were examined with XPS. In the present coated AAOs, the thin carbon layer with a thickness of about 4 nm uniformly covers all the surface of each AAO particle (not only the outside but also the inside). Therefore, an XPS spectrum obtained from the external surface can be regarded as that from the inner surface of the nanochannels, as pointed out in the previous communication.26 From the area ratio of N1s to C1s signals corrected with standard XPS sensitivity factors, the N/C atomic ratio in N/AAO was calculated to be 0.075. The N1s XPS spectrum (Figure 5a) was divided into four components, pyridine (398.0 eV), pyridone/pyrrole (400.9 eV), quaternary N (401.3 eV), and oxidized N (403.7 eV);30 and this curve fitting suggests that (30) Raymundo-Pinero, E.; Cazorla-Amoros, D.; Linares-Solano, A.; Find, J.; Wild, U.; Schlogl, R. Carbon 2002, 40, 597–608.

DOI: 10.1021/la901318d

11963

Article

Kwon et al.

Figure 3. TEM images of the carbon layers liberated from (a) C/AAO, (b) N/AAO, and (c) B/AAO.

Figure 4. Nitrogen adsorption-desorption isotherms (a), and mesopore size distributions (b) calculated by the BJH method. Table 1. BET Surface Areas, Pore Volumes, and Resistivities of the Coated AAOs sample

a

2

-1

SBET (m g )

Vtotalb (cm3 g-1)

Vmicroc (cm3 g-1)

VCO2d (cm3 g-1)

Vwatere (cm3 g-1)

resistivity f (Ω cm)

C/AAO 32 0.09 0.01