Highly efficient supporting material derived from used cigarette filter for ...

Catalysis Communications 78 (2016) 1–6

Contents lists available at ScienceDirect

Catalysis Communications journal homepage: www.elsevier.com/locate/catcom

Short communication

Highly efficient supporting material derived from used cigarette filter for oxygen reduction reaction Gil-Pyo Kim 1, Minzae Lee 1, Hyeon Don Song, Seongjun Bae, Jongheop Yi ⁎ World Class University (WCU) Program of Chemical Convergence for Energy & Environment (C2E2), School of Chemical and Biological Engineering, College of Engineering, Seoul National University (SNU), Seoul 151-742, Republic of Korea

a r t i c l e

i n f o

Article history: Received 1 December 2015 Received in revised form 27 January 2016 Accepted 28 January 2016 Available online 29 January 2016 Keywords: Cigarette filter Porous carbon Recycle Carbonization Oxygen reduction Electrocatalyst

a b s t r a c t Bimodal porous nitrogen (N) doped carbon supported Pt composite was prepared as a catalyst for oxygen reduction reaction (ORR). The N-doped carbon (NCF) support was obtained via one-pot pyrolysis of the used cigarette filters. Physical characterizations and electrochemical tests proved that the presence of N dopant on the surface of the NCF not only provided highly dispersive active sites for the growth of the Pt nanoparticles but also the active centers for ORR itself. It was demonstrated that these combinative effects contributed on higher ORR activity and durability than those for the commercial carbon (Vulcan XC) supported Pt composites. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Developing proper catalysts for the oxygen reduction reaction (ORR) is the current issue for the industrial development of fuel cells [1]. During the past years, Pt has been used as the commercial electrocatalyst for the ORR due to its high activity [2]. Owing to the high cost and limited storage of Pt, however, decreasing the amount of usage with enhancing the catalytic activity of Pt catalyst has been come into a compulsory aim [3]. One of the available strategies to meet the demand is the increase of the active surface area of Pt by adopting appropriate supporting material (e.g., carbon material), which enables the high dispersion of Pt nanoparticles [4,5]. It has been reported that the catalytic performances such as onset potential, current density, and durability are susceptible to the type of catalytic supporting material [6]. The requirements for carbon material as a support are high surface area, good conductivity, and proper pore structure, which provide the uniformly distributed anchoring sites for Pt catalysts and facilitate the smooth electrolyte flux for effective catalytic reaction without hampering electron transfer [7]. Nitrogen (N)-doped carbon with bimodal structure, which consists of micropores and mesopores, is a proper candidate material for supporting the Pt nanoparticles [8]. Typically, the bimodal pore system can induce the high surface area, which confers an increased ⁎ Corresponding author. E-mail address: [email protected] (J. Yi). 1 Gil-Pyo Kim and Minzae Lee contributed equally to this work.

http://dx.doi.org/10.1016/j.catcom.2016.01.030 1566-7367/© 2016 Elsevier B.V. All rights reserved.

distribution of C–N catalytic center and provide the easier accessibility to active sites for reactants [9]. More importantly, it has been reported that the presence of N dopants on the carbon surface not only can donate excess electrons for the fine nucleation of Pt nanoparticles but also limit the mobility of Pt nanoparticles by metal–N interaction, which prevent them from agglomeration [10]. Moreover, the substituted N atoms are known to have the catalytic activity toward ORR [11]. Thus, it could be another contribution factor for exhibiting high catalytic performance. These physicochemical properties are beneficial to design an effective Pt-based electrocatalyst possessing highly dispersive Pt nanoparticles by inducing N species on the entire surface of the carbon supporting material. There have been several methods for the preparation of the N-doped carbon supporting material such as carbonization of organic material that readily containing N species, heat treatment of organics and/or carbon with N-containing gases, and utilizing N-containing chemicals with carbon sources together before the carbonization [12–15]. Among these, carbonization of organic material with ammonia (NH3) gas enables N doping within the carbon matrix and simultaneously allows formation of porous structure on the whole surface by a pyrolysis, which can be a powerful approach for achieving the desirable physicochemical properties for the carbon based electrocatalyst toward the ORR at one-step procedure [16]. In this study, we report on a simple and environmentally benign route to the scalable production of well dispersed Pt on hierarchical Ndoped porous carbon (Pt/NCF) via one-pot carbonization of used cigarette filters and direct reduction of Pt. This synthetic strategy possesses

2

G.-P. Kim et al. / Catalysis Communications 78 (2016) 1–6

the following desirable advantages: (1) recycling used cigarette filters as a carbon source for use in a Pt support, (2) developing the bimodal pore structure, which can provide large surface area and fast oxygen insertion and desertion and (3) simultaneous introduction of the N functional groups into this supporting material, which can play a key role in the formation of homogeneous nucleation sites for highly dispersed Pt nanoparticles and catalytic active sites toward the ORR. To the best of our knowledge, a study of the feasibility of preparing bimodal porous carbon material synthesized from used cigarette filters and its usage as a supporting material of Pt nanoparticles for ORR have not been reported yet. 2. Experimental 2.1. Preparation of catalysts As a feasible approach, we chose the world-widely consumed cigarette brand (Marlboro Light Gold produced by Philip Morris Int.) and popular brands (The One Orange and Bohem Cigar Mojito produced by Korea Tobacco & Ginseng Corp.) in our country, Republic of Korea together, as a model case. These industrial cigarette filters are largely composed of cellulose acetate (N95%), which reveals that no significant differences toward the origin of carbon source between each brand is chemically negligible [17]. The used cigarette butts from these brands were collected and filters were separated. N-doped carbon (NCF) and carbon without N-doping (CF) were obtained from heat treatments of the separated filters at 900 °C for 2 h under the NH3 and Ar, respectively. Identical mass of filters from each brand without any pretreatments (including washing) were used before the pyrolysis. NCF-supported Pt catalysts were prepared by the conventional impregnation method. Proper amounts of chloroplatinic acid (H2 PtCl6 ·6H2 O) solution (50 mL g − 1 ) were added into the 0.05 g of NCF powder dispersed deionized water (200 mL). Sodium borohydride (0.1 M, 30 mL) was added to each solution and stirred it for 12 h in order to reduce the Pt ions to Pt nanoparticles. The dispersion was filtered and dried at 60 °C for 12 h. The Pt contents for Pt/NCF

were examined to be 15.08 wt% and 20.46 wt%, which are denoted as Pt/NCF15 and Pt/NCF20, respectively. Pt/CF (21.99 wt%) and Pt/Vulcan XC (20.29 wt%) synthesized by using above method and commercial Pt/C (Alfa Aesar, 20 wt%) were compared as control samples, which are denoted herein as Pt/CF20, Pt/VC20, and Pt/C20, respectively. 2.2. Physicochemical characterization The morphology of the sample was characterized by JEM-3010 highresolution transmission electron microscopy (HR-TEM). An AXIS-HIS Xray photoelectron spectroscopy (XPS) was used to investigate the elemental species. The amount of N in the Pt/NCF20 was determined by using a CHNS 932 analyzer. N2 sorption isotherms were measured with a Micromeritics ASAP 2010 instrument. Atomic-distribution mapping was obtained by using HAADF-STEM (JEM-2100F, 200 kV). Pt contents were examined by JP/ICPS-7500 inductively coupled plasma with atomic emission spectrometry (ICP-AES). 2.3. Electrochemical measurements All electrochemical experiments were performed with an Iviumstat workstation in a three-electrode cell. Rotating disk electrode (RDE) was used for the measurement. A catalyst loaded glassy carbon electrode (GCE, 5 mm dia.), an Ag/AgCl electrode (KCl saturated) and a platinum electrode served as working, reference and counter electrodes, respectively. Catalyst inks were prepared by the following order. As-prepared catalysts (10 mg) were mixed with 20 μL deionized water for stable mixing with the binder in the next step. 57 μL of 5% Nafion and 800 μL of isopropyl alcohol were then inserted under vigorous stirring and sonication for both the 30 min. 7.0 μL of each ink was dropped onto the GCE and dried at room temperature. The linear sweep voltammetry (LSV) was performed between 0.4 and 1.2 V (vs. RHE) with a scan rate of 5 mV s−1 in O2 saturated 0.1 M KOH aqueous solution at various rotating speeds from 400 to 2000 rpm. The stability test was performed at 0.5 V (vs. RHE) for 6000 s.

Fig. 1. (a) Photograph of the used cigarette filters and Pt/NCF20 powder. (b) SEM image of the fiber of cigarette filter. HR-TEM images of the (c) NCF, (d) Pt/NCF20 and (e) Pt/VC20. Insets for (d) and (e) are the size distribution of Pt nanoparticles.

G.-P. Kim et al. / Catalysis Communications 78 (2016) 1–6

3

The kinetic parameters can be calculated based on the Koutecky– Levich (K–L) equations as follows [3]: 1 1 1 ¼ þ j jL jK 2=3

jL ¼ 0:62  F  D0

 ν −1=6  C 0  n  ω1=2

where j is the current density of catalyst measured from RDE, jL is the diffusion-limiting current density, jK is the kinetic-limiting current density, F is the Faradaic constant (96,485 C mol−1), D0 is the diffusion coefficient of O2 in 0.1 M KOH (1.9 × 10−5 cm2 s−1), ν is the kinematic viscosity of electrolyte (0.01 cm2 s−1), C0 is bulk concentration of O2 in electrolyte (1.2 × 10−6 mol cm−3), n is the number of transferred electrons, and ω is the angular velocity of disk electrode (rad s−1). 3. Results and discussion Fig. 1(a) shows the features of used cigarette filters and the assynthesized Pt/NCF20 powder. Fig. 1(b) shows the SEM image of the used cigarette filter fiber with the cross-sectional shape of alphabet “Y”. It is widely known that the industrially produced cigarette filters are composed largely of cellulose acetate. Overall morphology and size of NCF are irregular because the bulk carbon particle is physically grinded before the Pt deposition as shown in Fig. S1. The morphologies of NCF, Pt/NCF20 and Pt/VC20 were characterized by TEM as shown in Fig. 1(c-e), respectively. Some bright spots are appeared for NCF, indicating the developed pores within the particle after carbonization. These pores are randomly distributed in a whole surface, resulting in the development of bimodal pore structure. In the case of Pt/NCF20, Pt nanoparticles were highly dispersed on the surface of NCF without any agglomeration. In contrast, Pt nanoparticles for Pt/VC20 appeared to form larger particles at certain locations. The high agglomeration for the Pt/VC20 is caused by the formation of dangling bonds between adjacent Pt nanoparticles during the reduction procedure without any external forces. Furthermore, due to the intrinsic hydrophobicity and the chemically inert nature of Vulcan XC, few desirable sites for the growth of Pt nanoparticles are available, which results in a broad size distribution with aggregation of Pt. The average particle size of the Pt for Pt/NCF20 and Pt/VC20 were measured to be 3.85 ± 0.2 nm and 8.49 ± 0.2 nm, respectively. Based on the above results, it is clearly demonstrated that N-doping can permit uniform active sites for the growth of Pt nanoparticles due to the metal–N interactions [10]. In our case, the amount of N contents in the Pt/NCF20 was determined to be 5.6 wt% by using CHNS analysis. The pore characteristics of the samples were measured by N2 sorption isotherms. The isotherms of NCF, Pt/NCF20, and Pt/CF20 showed the type-IV, as shown in Fig. 2(a), with BET surface area of 2147.83 m2 g−1, 712.12 m2 g−1, and 329.33 m2 g−1, respectively. The isotherm of Pt/VC20 showed the type III with BET surface area of 161.94 m2 g− 1, indicating that the particle is physically formed as a multi-layered structure without having much pores. Hysteresis loops 1 = 0.45 to 0.8 for the NCF, Pt/NCF20, and Pt/ developed from P P− 0 CF20 are attributed to the coexistence of meso- and micropores. The smaller adsorbed pore volume for the Pt/NCF20 than that for the NCF implies that Pt nanoparticles are scattered on the whole surface of the NCF, and partially block some of the pores. Consequently, the decrease of pore volumes for micro- and mesopores are observed by the BJH pore distribution, as shown in Fig. 2(b). The larger amount of micropores for Pt/NCF20 than that for the Pt/CF20 is ascribed to the formation of micropores during the pyrolysis with NH3. It should be noted that the mesoporous morphology would provide more active sites for ORR [18]. XPS analyses were performed to observe the chemical composition of N dopant before and after the deposition of Pt on the NCF. The binding energy of chemically bonded N atoms for the NCF were 398.1 eV, 400.0 eV and 401.1 eV, which are assigned to pyridine, pyrrolic, and

Fig. 2. (a) N2 sorption isotherms and (b) BJH pore size distributions of the NCF, Pt/NCF20, Pt/CF20, and Pt/VC20.

quaternary N species, respectively, as shown in Fig. 3(a). The amount of pyridinic-N was decreased from 45.6 at.% to 31.6 at.% among all N contents after the Pt deposition as shown in Fig. 3(b). Since the pyridinic N is regarded as the most active sites for coordination with Pt ions, the ratio of pyridinic-N decreases [19]. The content of pyridinic-N for Pt/ NCF15 was examined to be 34.5 at.% for all N as shown in Fig. S2. The smaller changes (11.1 at.%) in pyridinic-N content for Pt/NCF15 than that for Pt/NCF20 implies that pyridinic-N dominantly participates in the formation of Pt growth. The amounts of oxygen atom for NCF, CF, and VC were examined to be 19.8%, 19.6%, and 22.3%, which is previously reported to promote the dispersion of Pt catalyst [20]. Hence, we conclude that Pt particles grow preferentially on the pyridinic-N [21]. Moreover, it was also found that the binding energies for Pt4f for Pt/ NCF20 negatively shift ~1.4 eV compared to those for Pt/CF20 and Pt/ VC20, as shown in Fig. S3, confirming that the interaction between pyridinic-N affects the electronic structure of Pt atom. Lowered binding energy in metal core level is related to the downshift of d-band center, which leads to the weakening the adsorption strength between the surface of Pt and the adsorbates [22,23]. Thus, the enhanced electrocatalytic activity toward the ORR can be achieved [24]. The presence of highly-dispersed Pt nanoparticles for the Pt/NCF20

4

G.-P. Kim et al. / Catalysis Communications 78 (2016) 1–6

Fig. 3. XPS spectra of the (a) NCF, and (b) Pt/NCF20. (c) HR-TEM (scale-bar included) and HAADF-STEM images of Pt/NCF20 with elemental mapping of C, Pt, N, and O.

was also confirmed by a HAADF-STEM elemental mapping analysis, which is good agreement with the dispersion of N dopants, as shown in Fig. 3(c). From these results, it was expected that the Pt/ NCF can exhibit the excellent electrocatalytic performance toward the ORR. LSV tests were performed to examine the electrocatalytic activity toward the ORR, as shown in Fig. 4(a). Since the relatively slow kinetics of ORR for carbon material in nature, which is strongly influenced by the 2e− reaction pathway, the limiting current densities of CF, NCF, and VC are lower than those of the Pt-based samples. Among these non-Pt samples, NCF exhibited higher limiting current density and the E1/2 than those for the CF and VC. This electrocatalytic superiority is mainly ascribed to the presence of N dopants, which may accelerate the ORR kinetic [25]. In particular, Pt/NCF20 exhibited a more positive ORR halfwave potentials, ΔE1/2 = 48, 48, 42, and 27 mV than those for the Pt/ CF20, Pt/NCF15, Pt/VC20, and the Pt/C20, respectively, as shown in Fig. 4(b). This indicates that the use of Pt/NCF20 can reduce the voltage loss (overpotential) compared to the other samples, leading to the improved cell efficiency. Although the amount of Pt loading for the Pt/ NCF15 was lower than those for the Pt/CF20, Pt/VC20, and Pt/C20, the

onset potential (Eonset) of Pt/NCF15 was very similar to their values, meaning that 5 wt% of Pt can be saved when the NCF is used as a supporting material. Therefore, the superior electrocatalytic property for Pt/NCF20 is mainly ascribed to two major contributions: (1) high degree of Pt nanoparticles dispersion, which is attributed to the N-doped carbon support by providing abundant active sites for the facile formation of Pt nanoparticle without any agglomeration [20], and (2) the presence of N dopants, which contributes on the high catalytic activities even without loading an extra catalyst [26]. In particular, the highest electrochemical active surface area (EASA) for the Pt/NCF20 also indicates that the Pt–N interaction enhanced the production of high quality electrocatalyst (Fig. S4 and Table S1). The LSV analyses of Pt/NCF20 at various rotation speeds were performed to verify the kinetic property, as shown in Fig. 4(c). The increase of limiting current density for the Pt/NCF20 with respect to the rotation speed is due to the increased mass transfer at the surface of GCE and the reduced diffusion distance of O2 [27]. The ORR kinetic parameters of Pt/NCF20 were obtained from K–L plots by using Koutecky–Levich equation, as shown in inset of Fig. 4(c) [3]. The good linearity of K–L plot corresponds to the firstorder reaction kinetics against to the concentration of surrounding

G.-P. Kim et al. / Catalysis Communications 78 (2016) 1–6

5

Fig. 4. (a) LSVs of CF, NCF, VC, Pt/CF20, Pt/NCF15, Pt/NCF20, Pt/VC20 and Pt/C20 in O2-saturated 0.1 M KOH solution at a scan rate of 5 mV s−1 at 1600 rpm. (b) Comparison of E1/2 and onset potential (Eonset) for all samples. Eonset was determined by a potential just before reaching the 0.0 mA cm−2 of current densities from RDE. (c) RDE polarization curves of Pt/NCF20 at different rotating speeds at a scan rate of 5 mV s−1. The inset shows K–L plots at different electrode potentials. (d) The relative ORR cathodic current-time response for Pt/NCF20, Pt/ VC20, and Pt/C20, performed at 1600 rpm with 0.5 V (vs. RHE) for 6000 s.

oxygen. The transferred electron numbers (n) per oxygen molecule for Pt/NCF20 were calculated to be 3.85. These reveal that ORR occurred on the Pt/NCF20 is 4e− process. To further evaluate the practical viability for Pt/NCF20, the stability tests were performed as shown in Fig. 4(d). The Pt/NCF20 showed much superior stability (holds ~97.3%) than those of Pt/C20 and Pt/VC20 (hold ~84.6% and ~61.0%, respectively), during a 6000 s testing period. 4. Conclusions This work reports a green approach for the fabrication of Pt based electrocatalyst (Pt/NCF20) by recycling the used cigarette filters. Welldispersive N dopants on the surface of the NCF provided nucleation sites for Pt nanoparticles, resulting in higher Pt dispersion on the surface of the NCF than that of the commercially used carbon support (Vulcan XC). Bimodal pore size distribution and N dopants also contributed on the enhanced activity toward the ORR, resulting in the improved electrocatalytic activity of Pt/NCF20 toward the ORR. Therefore, it can be concluded that the NCF is a promising support material for the Pt nanoparticles toward the ORR catalysis as well as reducing environmentally threatened solid waste “cigarette butts”. Acknowledgements This research was supported by the Global Frontier R&D Program on Center for Multiscale Energy System funded by the National Research Foundation under the Ministry of Science, ICT & Future, Korea (NRF-2011-0031571) and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No. 2013R1A2A2A01067164).

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.catcom.2016.01.030. References [1] M.K. Debe, Nature 486 (2012) 43–51. [2] I.E.L. Stephens, A.S. Bondarenko, U. Grønbjerg, J. Rossmeisl, I. Chorkendorff, Energy Environ. Sci. 5 (2012) 6744–6762. [3] E. Yoo, Y. Nagashima, T. Yamazaki, T. Matsumoto, J. Nakamura, Polym. Adv. Technol. 17 (2006) 540–543. [4] M. Kim, J.-N. Park, H. Kim, S. Song, W.-H. Lee, J. Power Sources 163 (2006) 93–97. [5] F.-D. Kong, G.-P. Yin, C.-Y. Du, S. Zhang, Y.-T. Qu, L. Du, Q.-Q. Xu, A.-X. Ling, Catal. Commun. 68 (2015) 67–72. [6] R. Siburian, Int. Res. J. Pure. Appl. Chem. 4 (2014) 541–550. [7] S. Zhang, S. Chen, J. Power Sources 240 (2013) 60–65. [8] F. Su, Z. Tian, C.K. Poh, Z. Wang, S.H. Lim, Z. Liu, J. Lin, Chem. Mater. 22 (2010) 832–839. [9] F. Pan, Z. Cao, Q. Zhao, H. Liang, J. Zhang, J. Power Sources 272 (2014) 8–15. [10] C.-L. Sun, C.-W. Pao, H.-M. Tsai, J.-W. Chiou, S.C. Ray, H.-W. Wang, M. Hayashi, L.-C. Chen, H.-J. Lin, J.-F. Lee, L. Chang, M.-H. Tsai, K.-H. Chen, W.-F. Pong, Nanoscale 5 (2013) 6812–6818. [11] J. Wu, L. Ma, R.M. Yadav, Y. Yang, X. Zhang, R. Vajtai, J. Lou, P.M. Ajayan, ACS Appl. Mater. Interfaces 7 (2015) 14763–14769. [12] F. Jaouen, M. Lefevre, J.P. Dodelet, M. Cai, J. Phys. Chem. B 110 (2006) 5553–5558. [13] R. Gadiou, A. Didion, D.A. Ivanov, I. Czekaj, R. Kötz, C. Vix-Guterl, J. Phys. Chem. Solids 69 (2008) 1808–1814. [14] M.M. Titirici, A. Thomas, M. Antonietti, J. Mater. Chem. 17 (2007) 3412–3418. [15] R.J. White, M. Antonietti, M.M. Titirici, J. Mater. Chem. 19 (2009) 8645–8650. [16] W. Luo, B. Wang, C.G. Heron, M.J. Allen, J. Morre, C.S. Maier, W.F. Stickle, X. Ji, Nano Lett. 14 (2014) 2225–2229. [17] B. Harris, Tob. Control. 20 (2011) i10–i16. [18] I. Kruusenberg, J. Leis, M. Arulepp, K. Tammeveski, J. Solid State Electrochem. 14 (2010) 1269–1277. [19] T. Palaniselvam, A. Irshad, B. Unni, S. Kurungot, J. Phys. Chem. C 116 (2012) 14754–14763.

6

G.-P. Kim et al. / Catalysis Communications 78 (2016) 1–6

[20] Z. Lei, M. Zhao, L. Dang, L. An, M. Lu, A.-Y. Lo, N. Yu, S.-B. Liu, J. Mater. Chem. 19 (2009) 5985–5995. [21] S. Sun, G. Zhang, Y. Zhong, H. Liu, R. Li, X. Zhou, X. Sun, Chem. Commun. (2009) 7048–7050. [22] H. Tsunoyama, N. Ichikuni, H. Sakurai, T. Tsukuda, J. Am. Chem. Soc. 131 (2009) 7086–7093. [23] L. Qiu, F. Liu, L. Zhao, W. Yang, J. Yao, Langmuir 22 (2006) 4480–4482.

[24] B. Hammer, J.K. Nørskov, Nature 376 (1995) 238–240. [25] M. Vikkisk, I. Kruusenberg, S. Ratso, U. Joost, E. Shulga, I. Kink, P. Rauwel, K. Tammeveski, RSC Adv. 5 (2015) 59495–59505. [26] P.H. Matter, E. Wang, M. Arias, E.J. Biddinger, U.S. Ozkan, J. Phys. Chem. B 110 (2006) 18374–18384. [27] W. Yang, T.-P. Fellinger, M. Antonietti, J. Am. Chem. Soc. 133 (2011) 206–209.