DOI: 10.1002/cssc.201501141
Full Papers
Order of Activity of Nitrogen, Iron Oxide, and FeNx Complexes towards Oxygen Reduction in Alkaline Medium Yansong Zhu,[a, b] Bingsen Zhang,[b] Da-Wei Wang,[c] and Dang Sheng Su*[b]
In alkaline medium, it seems that both metal-free and ironcontaining carbon-based catalysts, such as nitrogen-doped nanocarbon materials, FeOx-doped carbon, and Fe/N/C catalysts, are active for the oxygen reduction reaction (ORR). However, the order of activity of these different active compositions has not been clearly determined. Herein, we synthesized nitrogen-doped carbon black (NCB), Fe3O4/CB, Fe3O4/NCB, and FeN4/CB. Through the systematic study of the ORR catalytic activity of these four catalysts in alkaline solution, we confirmed
the difference in the catalytic activity and catalytic mechanism for nitrogen, iron oxides, and Fe–N complexes, respectively. In metal-free NCB, nitrogen can improve the ORR catalytic activity with a four-electron pathway. Fe3O4/CB catalyst did not exhibit improved activity over that of NCB owing to the poor conductivity and spinel structure of Fe3O4. However, FeN4 coordination compounds as the active sites showed excellent ORR catalytic activity.
Introduction Fuel cells, as a clean and efficient power source, attracted significant attention during the last decades.[1] The cathodic oxygen reduction reaction (ORR) is the heart of fuel cell performance, and high-performance ORR electrocatalysts are essential for practical applications. It is critical to replace platinum-based electrocatalysts in order to reduce the manufacturing cost of proton exchange membrane fuel cells (PEMFCs). Metal-free and non-precious metal catalysts (NPMCs) with high ORR activity and stability are a major focus of PEMFC research.[2] In acid electrolytes, it is clear that pyrolyzed Fe/N/C and FeCo/N/C catalysts have good catalytic activity for ORR, that is, close to that of Pt/C, whereas metal-free catalysts fall behind.[3] However, in alkaline medium, both metal-free and non-precious metal catalysts show good catalytic activity. In 2009, Dai[4] reported that nitrogen-doped carbon nanotube arrays as metal-free electrocatalysts had better catalytic activity than that of the commercial Pt/C catalysts in alkaline medium. But the nitrogen-doped nanocabon materials are not [a] Dr. Y. Zhu School of Chemistry and Life Science Anshan Normal University Anshan, Liaoning, 114005 (China) [b] Dr. Y. Zhu, Prof. B. Zhang, Prof. D. S. Su Shenyang National Laboratory for Materials Science Institute of Metal Research Chinese Academy of Sciences Shenyang, Liaoning, 110016 (China) E-mail: [email protected] [c] Dr. D.-W. Wang School of Chemical Engineering University of New South Wales UNSW Sydney, NSW 2052 (Australia) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201501141.
ChemSusChem 2015, 8, 4016 – 4021
always metal-free catalysts; it was reported that 200 ppm Fe or Co impurities can significantly increase the ORR activity.[5] Despite the numerous studies about metal-free nitrogen-doped nanocarbon materials for ORR, such as carbon nanotubes, carbon nanospheres, mesoporous carbons, and graphene sheets,[6] some researchers suggested that the active sites could be iron oxides or Fe–N complexes in certain nanocarbon-containing electrocatalysts.[7] Although it is clear that all of the three kinds of active components (i.e., nitrogen, iron oxides, and Fe–N complexes) are active for ORR, the order of activity for ORR in alkaline medium remains elusive. Is there any difference in the catalytic activity and mechanism for ORR in alkaline medium between metal-free nitrogen-doped nanocarbon, iron, and Fe/N/C non-precious metal electrocatalysts? What are these differences? In order to answer these questions, we synthesized metal-free nitrogen-doped carbon black (NCB), Fe3O4/CB, Fe3O4/NCB, and FeN4/CB. Through the systematic study of the ORR catalytic activity of these four catalysts in alkaline solution, we were able to answer these questions.
Results and Discussion Firstly, inductive coupling plasma mass spectrometry (ICP-MS) was used to detect a trace amount of iron in carbon black for excluding the impact of iron. It is found that there is only 6 ppm iron in carbon black after treatment with hydrochloric acid. The content of iron is much less than the as-claimed metal-free electrocatalysts (200 ppm).[5] Ammonia gas was used as a clean nitrogen source to prepare nitrogen-doped carbon black at 950 8C. Based on X-ray photoelectron spectroscopy (XPS) analysis, nitrogen atoms were doped in the catalysts and the nitrogen content on the surface of NCB was 2.04 at. %. As shown in Figure S1 in the Supporting Informa4016
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers tion, the N 1s XPS spectra of NCB was deconvoluted into four components consisting of pyridinic N (398.6 eV), pyrrolic N (400.5 eV), quaternary N (401.2 eV), and oxidized pyridinic N (403.8 eV);[8] pyridinic N was the dominating group in the NCB catalyst. The electrocatalytic performances of CB, NCB, and Pt/C for ORR in 0.1 m KOH were measured by using linear sweep voltammetry (LSV), as shown in Figure 1. We clearly observed two
Table 1. ORR kinetic data recorded in alkaline medium for NCB, Fe–CB, Fe–N–CB, and Pt/C. Catalyst
Eonset [V]
E1/2 [V]
Tafel slope [mV dec¢1]
i0 [A cm¢2]
NCB Fe–CB Fe–N–CB Fe–Ph–CB Pt/C
0.847 0.789 0.849 0.971 1.043
0.765 0.682 0.746 0.918 0.934
62 83 88 36 86
1.5 Õ 10¢9 9.8 Õ 10¢9 6.5 Õ 10¢8 1.0 Õ 10¢9 8.7 Õ 10¢6
lytic performance of CB in alkaline medium (transforming ORR from a peroxide pathway close to a four-electron pathway), but NCB and Pt/C still have a large gap in the ORR catalytic activity, which is especially clear in Eonset difference of about 200 mV. The ORR catalytic activity of CB with iron oxide was also studied in alkaline medium. Fe–CB was synthesized by heattreating a mixture of carbon black and ferrous acetate at 950 8C in an inert atmosphere (Argon gas). The sample only contains iron species and carbon black. It can be seen that Eonset for Fe–CB is 0.789 V (in Figure 2 a). Though the value of Eonset for Fe–CB is shifted 21 mV more positive than that for CB, it is 58 mV lower than that for NCB. Based on the slope of the Koutecky–Levich plots, the number of electrons transferred in ORR catalyzed by Fe-CB (the value of n at 0.50 V)) is about 3.9,
Figure 1. Rotating disk electrode (RDE) voltammograms of CB, NCB, and Pt/ C in O2-saturated 0.1 m KOH. Scan rate: 5 mV s¢1, rotation rate: 900 rpm.
limiting current plateau in the LSV curve of CB, which indicate that ORR catalyzed by CB is not a four-electron process, but a two-electron reaction. However, the polarization curve of NCB has only one limiting current platform. The Koutecky– Levich plots (Figure S2 a, b) are drawn from the ORR curves based on the Koutecky–Levich equation: 1 1 1 ¢ ¼¢ þ I Ik 0:62nFAD2=3 v¢1=6 w1=2 cO2 where I is the current density, Ik is the current density in the absence of any mass-transfer effects, w is the rotation speed of the disk, F = 96 485 C mol¢1, A = 0.19625 cm2, D = 1.9 Õ 10¢5 cm2s¢1, c = 1.2 Õ 10¢6 mol cm¢3, and v = 1.0 Õ 10¢2 cm2s¢1.[9] According to the slope of the Koutecky–Levich plots, the values of n at 0.50 V (the number of electrons transferred) in the ORR catalyzed by CB and NCB were calculated to be 2.3 and 3.7, respectively. This suggests that the ORR catalyzed by CB transforms from a peroxide pathway to a four-electron pathway as a result of nitrogen doping. In addition, the onset (Eonset) and half-wave potential (E1/2) in the curve of NCB are 0.847 and 0.765 V, respectively. Comparing Pt/C (loading of 0.06 mgPt cm¢2) and NCB, the differences in the Eonset and E1/2 are 197 and 169 mV, respectively. From a Tafel ORR plot with NCB (Figure S2 c), the Tafel slope and exchange current density (i0) for the NCB catalysts were obtained (Table 1). The i0 value for NCB (1.5 Õ 10¢9 A cm¢2) is far lower than that for the Pt/C catalyst (8.7 Õ 10¢6 A cm¢2), and the Tafel slope of ORR for NCB (62 mV dec¢1) is close to that of the Pt/C catalyst (86 mV dec¢1). To summarize, the nitrogen doping can improve the ORR cataChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
Figure 2. (a) RDE voltammograms of CB, NCB, Fe–CB, and Fe–N–CB in O2-saturated 0.1 m KOH. Scan rate: 5 mV s¢1, rotation rate: 900 rpm. (b) CV of Fe– CB and Fe–N–CB in N2-saturated 0.1 m KOH. Scan rate: 10 mV s¢1.
4017
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers which is closer to a four-electron process than that by NCB. The addition of iron did not greatly improve the ORR catalytic activity of CB in alkaline medium, which is believed relevant to the type of iron species. The cyclic voltammetry (CV) curve of Fe–CB in Figure 2 b shows a pair of redox peaks at 0.586 and 0.821 V (the peak separation is 235 mV). The irreversible redox peaks are not assigned to the peaks of reduction/oxidation Fe3 + /Fe2 + in iron coordination compounds like [Fe(CN)6]3¢/ [Fe(CN)6]4¢, but to crystalline iron species. XRD was used to explore the phase composition in the catalysts (Figure 3). Fe–CB
Figure 4. (a) Low-magnification TEM image. (b) High-resolution TEM image. (c) High angle annular dark field scanning transmission electron microscopy (HAADF-STEM) maps of Fe–CB.
Figure 3. XRD patterns of Fe–CB and Fe–N–CB (JCPDS-Fe3O4 : 19-0629; JCPDS-FeO: 06-0615).
shows well-developed crystalline structures that can be assigned to Fe3O4. Moreover, the fine structure of Fe3O4 particles was further investigated by electron microscopy. As shown in Figure 4 a, the TEM image of Fe–CB shows that there are some large particles with the sized 30–40 nm in the catalyst. The high resolution transmission electron microscopy (HRTEM) image in Figure 4 b and energy dispersive X-ray spectroscopy (EDS) elemental maps in Figure 4 c further reveal that the particles are mainly Fe3O4. Based on the above analysis, it can be conformed that there was only one iron species, Fe3O4, in Fe– CB. Very recently it was hypothesized that the chemisorption of oxygen molecules is the rate determining step in ORR catalyzed by metal-free and non-precious metal catalysts.[7f, 8a, 10] And, compared with nitrogen, iron ions should reduce the lower activation energy of ORR owing to the better ability of the chemisorption of oxygen molecules on iron ions.[7a, 11] It is generally believed that reducing the activation energy can decrease the overpotential during the electrocatalytic process that would be mainly reflected in Eonset of the polarization curve moving close to the standard electrode potential (for ORR, Eonset shifts positively). However, iron ions in Fe3O4 do not exhibit better ORR catalytic activity in alkaline medium. We deduce that although iron ions in Fe3O4 have improved chemisorption of oxygen molecules than nitrogen, the poor conductivity and spinel structure of Fe3O4 nanoparticles limit the rate of electron transfer from the CB to active iron ions, which thus becomes the rate determining step in ORR. This may result in ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
the higher activation energy of ORR catalyzed by iron ions in Fe3O4 than that by nitrogen in NCB. Therefore, Eonset for Fe–CB is 58 mV lower than that for NCB. Note that the number of electrons transferred in the ORR catalyzed by iron is closer to a four-electron process than that by nitrogen, which may be the result of greater affinity toward chemisorption of oxygen in the former. In order to study the ORR catalytic activity of the catalyst containing nitrogen and iron, Fe–N–CB catalyst was synthesized by heat-treating a mixture of CB and ferrous acetate at 950 8C in ammonia. Through the analysis of XRD in Figure 3 and the CV curve of Fe–N–CB in Figure 2 b, it was concluded that there are two iron species, Fe3O4 and FeO. FeNx coordination were not present as the reversible redox of Fe3 + /Fe2 + was not observed in the CV curve of Fe–N–CB. From the HRTEM image and EDS elemental maps of Fe–N–CB (Figure 5), the black 100 ~ 200 nm particles are Fe3O4 nanoparticles and the distribution of nitrogen is uniform in the Fe–N–CB sample. This implies that Fe–N–CB is nitrogen-doped CB containing Fe3O4 and FeO, which enables the study the ORR catalytic mechanism when nitrogen and iron species simultaneously exist in the catalyst. As shown in Figure 2 a, Eonset of the Fe–N–CB polarization curve (0.849 V) is similar to that of NCB (0.847 V), whereas the limiting current density for Fe–N–CB is larger than that for NCB. Based on the slope of the Koutecky–Levich plots, the number of electrons transferred in the ORR catalyzed by Fe–N– CB (the value of n at 0.50 V) is about 3.9. To determine whether the ORR performance results from the catalysis of nitrogen or iron in Fe–N–CB, we synthesized xFe–N–CB, where x corresponds to 10, 20, 30, 40, 50, and 100 mg of ferrous acetate used during the preparation (see the Experimental Section for more details). Figure 6 shows the LSV polarization curves of the six catalysts in O2-saturated 0.1 m KOH. Eonset does not
4018
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 5. (a) Low-magnification TEM image. (b) High-resolution TEM image. (c) HAADF-STEM maps of Fe–N–CB.
Figure 6. RDE voltammograms of Fe–N–CB with different iron contents in O2-saturated 0.1 m KOH. Scan rate: 5 mV ·s¢1, rotation rate: 900 rpm.
change with varied iron content and the limiting current densities are close except in the case of 100Fe–N–CB, which has higher current densities. In addition, the values of n at 0.50 V (the number of electrons transferred) in the ORR catalyzed by the six catalysts are all 3.9. This indicates that the iron in Fe– N–CB has little effect on the catalytic activity. We used these results to predict the possible ORR mechanism catalyzed by Fe–N–CB. When the overpotential is low at the beginning of ORR, nitrogen atoms in Fe–N–CB are the active sites for ORR as the activation energy of ORR catalyzed by iron ions in Fe3O4 is higher than that by nitrogen in NCB, as mentioned above (Scheme 1 a). This results in a similar Eonset for Fe–N–CB and NCB that does not change with the varied iron contents. However, when the potential scans negatively to 0.789 V, Eonset for Fe3O4/CB, the higher overpotential triggers the ORR on iron ions in Fe3O4. As a result, both iron ions and nitrogen atoms are active sites for ORR (Scheme 1 b, c). In addiChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
Scheme 1. (a) Schematic diagram for the ORR mechanism of Fe–N–CB at 0.847 V; (b) Schematic diagram for the ORR mechanism of Fe–N–CB at 0.789 V; (c) the LSV curve of Fe–N–CB (different regions with different active sites).
tion, it is common that the value of limiting current density is mainly related to the amount of active sites. Accordingly, the limiting current density for Fe–N–CB is larger than that for NCB because iron and nitrogen are both active sites for ORR when the current reaches the limiting current. Note that although iron ions and nitrogen atoms in Fe–N–CB catalyzes ORR together from 0.789 V, nitrogen is the man active site owing to the lower activation energy than that of iron ions. This is the reason that the limiting current densities of Fe–N–CB have no obvious change with varied addition of ferrous acetate from 10 to 50 mg. The limiting current density of 100Fe–N–CB significantly increases because there are many more iron ions as
4019
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers active sites for ORR compared with the other Fe–N–CB catalysts. Iron ions have improved ability toward chemisorption of oxygen molecules than nitrogen atoms, and ORR catalyzed by the former is closer to a four-electron process. As iron ions in Fe3O4 do not show the good ORR catalytic activity owing to the poor conductivity and spinel structure of Fe3O4, we believe that iron ions in iron species that have good conductivity and graphite-like structure serve as the ORR active sites with perfect catalytic performance. FeNx coordination compounds, are suitable iron species that fulfill these two requirements. Therefore, an iron phthalocyanine-based CB catalyst (Fe–Ph–CB) containing FeN4 was synthesized. The CV curve of Fe–Ph–CB in Figure 7 b shows a pair of reversible redox peaks at 0.785 and 0.860 V (the peak separation is 78 mV). We assign this to the reduction/oxidation (Fe3 + /Fe2 + ) of FeN4 in the catalysts. As shown in Figure 7 a, both Eonset (0.971 V) and the limiting current density for Fe–Ph–CB catalyst are the best among these Fe/N/C catalysts mentioned above. Also, based on the slope of the Koutecky–Levich plots, the number of electrons transferred in the ORR catalyzed by Fe–Ph–CB (the value of n at 0.50 V) is about 3.9. From the Tafel ORR plot, the Tafel slope and i0 for the Fe–Ph–CB are 36 mV dec¢1 and 1.0 Õ 10¢9 A cm¢2, respectively (Table 1), which are also better than the other catalysts. Unfortunately, the performance of the Fe–Ph–CB catalyst still falls short of Pt/C.
Conclusions We confirm that both nitrogen atoms and iron ions are the oxygen reduction reaction (ORR) active sites in alkaline medium. In metal-free nitrogen-doped carbon black (NCB), nitrogen doping can improve the ORR catalytic activity by transforming the reaction from a peroxide pathway to a four-electron pathway. However, compared with Pt/C and NCB, the difference in the onset potential (Eonset) of the polarization curve is ~ 200 mV. In addition to the number of electrons transferred, closer to a four-electron process, iron ions in Fe3O4 do not show improved ORR catalytic activity over NCB owing to the poor conductivity and spinel structure of Fe3O4. During the ORR process catalyzed by Fe3O4/NCB in which iron oxide and nitrogen atoms are present at the same time, the nitrogen atoms are the active sites for ORR when the over potential is low. The iron ions and nitrogen atoms can both serve as active sites to catalyze ORR when the potential shifts negatively, but the nitrogen atoms still accounts for the main part. The active sites in Fe/N/C, having ORR performance closest to that of Pt/C in alkaline medium, are the iron ions in the iron species, which have good electrical conductivity and a graphite-like structure. The activity of the catalysts is in the order FeN4 > N > FeOx in alkaline medium. We believe that the knowledge of the active site will lead the way to target-specific synthesis of highly active Fe/N/C catalysts for ORR in alkaline medium.
Experimental Section Catalyst synthesis Carbon black, typically a commercial carbon Ketjenblack EC 300 J (with Brunauer–Emmett–Teller surface area of about 801 m2 g¢1), was treated in 6.0 m HCl solution for 72 h to remove metal impurities. The acid-treated CB was heat treated at 950 8C in ammonia for 2 h, and nitrogen-doped CB, designated as NCB, was obtained. Ferrous acetate (0.05 g) was first dispersed with acid-treated CB (0.6 g) in 200 mL ultrapure water. After continuously mixing for 24 h, the suspension containing carbon and transition metal were vacuumdried using a rotary evaporator. By the subsequent heat-treatment at 950 8C for 2 h in an inert atmosphere (Argon gas) or in ammonia respectively, Fe–CB and Fe–N–CB were synthesized. Using the same synthetic procedure of Fe–N–CB in which the varied qualities of ferrous acetate (0.01, 0.02, 0.03, 0.04, 0.05, and 0.10 g) were added, xFe–N–CB where x = 10, 20, 30, 40, 50, and 100 were obtained, respectively. In addition, iron phthalocyanine (0.16 g) was dispersed with acid-treated CB (0.6 g) in 200 mL ultrapure water. After continuously mixing for 24 h, the suspension containing carbon and transition metal were vacuum-dried using a rotary evaporator. Then we obtained the catalyst Fe–Ph–CB.
Characterization
Figure 7. (a) RDE voltammograms of NCB, Fe–CB, Fe–N–CB, Fe–Ph–CB, and Pt/C in O2-saturated 0.1 m KOH. Scan rate: 5 mV·s¢1, rotation rate: 900 rpm. (b) CV of Fe–Ph–CB in N2-saturated 0.1 m KOH. Scan rate: 10 mV·s¢1.
ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
XRD analysis was carried out using a D/MAX-2500 PC X-ray diffractometer with monochromatized CuKa radiation (l = 1.54 æ). The patterns were obtained at a scan rate of 58 min¢1 with a step of 0.028. The XPS measurements were carried out in an ultra-high vacuum ESCALAB 250 set-up equipped with a monochromatic AlKa X-ray source (1486.6 eV; anode operating at 15 kV and 20 mA). A FEI Tecnai G2 F20 microscope equipped with high angle annular
4020
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers dark field (HAADF) and energy dispersive X-ray analysis (EDAX) detectors was employed to acquire the HRTEM and HAADF-STEM images and EDS under 200 keV.
[2]
Electrochemistry measurements [3] [4] [5]
ORR activity and four-electron selectivity of catalysts were measured by a rotating disk electrode (RDE). RDE measurements were performed by Epsilon Electrochemical Station (PAR2273) in a conventional three-electrode electrochemical cell. To avoid any potential contamination of a non-precious metal catalyst by platinum, experiments were carried out using a graphite rod as the counter electrode. An Ag/AgCl electrode in 3.0 m KCl (0.220 V vs. natural hydrogen electrode, NHE) was used as a reference electrode. All potentials were later converted to the reversible hydrogen electrode (RHE) scale. Non-precious metal catalyst data were obtained at RT (ca. 25 8C) with a catalyst loading of 0.9 mg cm¢2 in 0.1 m NaOH at a rotating disk speed of 900 rpm. The data of Pt catalyst were recorded with a 20 wt % Pt/C at loadings of 60 mgPt cm¢2. The catalysts were ultrasonically dispersed in an isopropanol solution containing suspended 5 wt % Nafion solution for 30 min to form a catalyst ink, which was then applied to the glassy-carbon disk surface. The disk rotation rate was from 400 to 2500 rpm. CV characterization of the catalysts in the absence of oxygen was typically carried out in the potential range from 1.0 to 0.0 V at a scan rate of 10 mV s¢1 in nitrogen-saturated 0.1 m KOH at RT.
[6]
[7]
[8]
Acknowledgements [9]
The authors acknowledge the financial support from MOST (2011CBA00504), NSFC of China (21133010, 21203215, 51221264, 21261160487), and ‘Strategic Priority Research Program’ of the Chinese Academy of Sciences, Grant No. XDA09030103. Keywords: carbon · electrocatalyst · iron · nitrogen · oxygen reduction reaction [1] a) D. S. Su, G. Sun, Angew. Chem. Int. Ed. 2011, 50, 11570 – 11572; Angew. Chem. 2011, 123, 11774 – 11777; b) D.-W. Wang, D. Su, Energy Environ. Sci. 2014, 7, 576 – 591; c) G. Wu, K. L. More, C. M. Johnston, P. Zelenay, Science 2011, 332, 443 – 447; d) Y. S. Zhu, B. S. Zhang, X. Liu, D. W. Wang,
ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
[10] [11]
D. S. Su, Angew. Chem. Int. Ed. 2014, 53, 10673 – 10677; Angew. Chem. 2014, 126, 10849 – 10853. a) F. Jaouen, E. Proietti, M. LefÀvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T. Chung, C. M. Johnston, P. Zelenay, Energy Environ. Sci. 2011, 4, 114 – 130; b) G. Wu, M. Nelson, S. Ma, H. Meng, G. Cui, P. K. Shen, Carbon 2011, 49, 3972 – 3982; c) G. Wu, P. Zelenay, Acc. Chem. Res. 2013, 46, 1878 – 1889. M. LefÀvre, E. Proietti, F. Jaouen, J.-P. Dodelet, Science 2009, 324, 71 – 74. K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 2009, 323, 760 – 764. F. Jaouen, A. M. Serventi, M. LefÀvre, J.-P. Dodelet, P. Bertrand, J. Phys. Chem. C 2007, 111, 5971 – 5976. a) K. Ai, Y. Liu, C. Ruan, L. Lu, G. M. Lu, Adv. Mater. 2013, 25, 998 – 1003; b) S. Chen, J. Bi, Y. Zhao, L. Yang, C. Zhang, Y. Ma, Q. Wu, X. Wang, Z. Hu, Adv. Mater. 2012, 24, 5593 – 5597; c) L. Qu, Y. Liu, J.-B. Baek, L. Dai, ACS Nano 2010, 4, 1321 – 1326; d) T. Sharifi, G. Hu, X. Jia, T. Wgberg, ACS Nano 2012, 6, 8904 – 8912; e) Z.-H. Sheng, L. Shao, J.-J. Chen, W.-J. Bao, F.-B. Wang, X.-H. Xia, Acs Nano 2011, 5, 4350 – 4358; f) Y. Tang, B. L. Allen, D. R. Kauffman, A. Star, J. Am. Chem. Soc. 2009, 131, 13200 – 13201; g) S. Yang, X. Feng, X. Wang, K. Mìllen, Angew. Chem. Int. Ed. 2011, 50, 5339 – 5343; Angew. Chem. 2011, 123, 5451 – 5455; h) Y. Zhu, Y. Lin, B. Zhang, J. Rong, B. Zong, D. S. Su, ChemCatChem 2015, 7, 2840 – 2845. a) F. Jaouen, S. Marcotte, J.-P. Dodelet, G. Lindbergh, J. Phys. Chem. B 2003, 107, 1376 – 1386; b) M. LefÀvre, J.-P. Dodelet, Electrochim. Acta 2003, 48, 2749 – 2760; c) J. Maruyama, J. Okamura, K. Miyazaki, Y. Uchimoto, I. Abe, J. Phys. Chem. C 2008, 112, 2784 – 2790; d) C. M¦dard, M. Lefevre, J. Dodelet, F. Jaouen, G. Lindbergh, Electrochim. Acta 2006, 51, 3202 – 3213; e) H. Wang, R. Cote, G. Faubert, D. Guay, J. Dodelet, J. Phys. Chem. B 1999, 103, 2042 – 2049; f) L. Zhang, K. Lee, C. W. Bezerra, J. Zhang, J. Zhang, Electrochim. Acta 2009, 54, 6631 – 6636. a) P. H. Matter, L. Zhang, U. S. Ozkan, J. Catal. 2006, 239, 83 – 96; b) V. Perazzolo, C. Durante, R. Pilot, A. Paduano, J. Zheng, G. A. Rizzi, A. Martucci, G. Granozzi, A. Gennaro, Carbon 2015, 95, 949 – 963. a) S. Baranton, C. Coutanceau, E. Garnier, J.-M. L¦ger, J. Electroanal. Chem. 2006, 590, 100 – 110; b) K.-L. Hsueh, D.-T. Chin, S. Srinivasan, J. Electroanal. Chem. 1983, 153, 79 – 95; c) R. N. Itoe, G. Wesson, E. E. Kalu, J. Electrochem. Soc. 2000, 147, 2445 – 2450; d) D. R. Lawson, L. D. Whiteley, C. R. Martin, M. N. Szentirmay, J. I. Song, J. Electrochem. Soc. 1988, 135, 2247 – 2253. X. Hu, Y. Wu, H. Li, Z. Zhang, J. Phys. Chem. C 2010, 114, 9603 – 9607. a) G. Lalande, G. Tamizhmani, R. Cote, L. Dignard-Bailey, M. Trudeau, R. Schulz, D. Guay, J. Dodelet, J. Electrochem. Soc. 1995, 142, 1162 – 1168; b) K. Legg, F. Jona, D. Jepsen, P. Marcus, J. Phys. C 1975, 8, L492; c) S. Sun, L. Chen, H. Shi, Y. Li, X. He, J. Electroanal. Chem. 2014, 734, 18 – 24.
Received: August 25, 2015 Revised: October 18, 2015 Published online on November 17, 2015
4021
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Order of Activity of Nitrogen, Iron Oxide, and FeNx Complexes towards Oxygen Reduction in Alkaline Medium Yansong Zhu,[a, b] Bingsen Zhang,[b] Da-Wei Wang,[c] and Dang Sheng Su*[b]
In alkaline medium, it seems that both metal-free and ironcontaining carbon-based catalysts, such as nitrogen-doped nanocarbon materials, FeOx-doped carbon, and Fe/N/C catalysts, are active for the oxygen reduction reaction (ORR). However, the order of activity of these different active compositions has not been clearly determined. Herein, we synthesized nitrogen-doped carbon black (NCB), Fe3O4/CB, Fe3O4/NCB, and FeN4/CB. Through the systematic study of the ORR catalytic activity of these four catalysts in alkaline solution, we confirmed
the difference in the catalytic activity and catalytic mechanism for nitrogen, iron oxides, and Fe–N complexes, respectively. In metal-free NCB, nitrogen can improve the ORR catalytic activity with a four-electron pathway. Fe3O4/CB catalyst did not exhibit improved activity over that of NCB owing to the poor conductivity and spinel structure of Fe3O4. However, FeN4 coordination compounds as the active sites showed excellent ORR catalytic activity.
Introduction Fuel cells, as a clean and efficient power source, attracted significant attention during the last decades.[1] The cathodic oxygen reduction reaction (ORR) is the heart of fuel cell performance, and high-performance ORR electrocatalysts are essential for practical applications. It is critical to replace platinum-based electrocatalysts in order to reduce the manufacturing cost of proton exchange membrane fuel cells (PEMFCs). Metal-free and non-precious metal catalysts (NPMCs) with high ORR activity and stability are a major focus of PEMFC research.[2] In acid electrolytes, it is clear that pyrolyzed Fe/N/C and FeCo/N/C catalysts have good catalytic activity for ORR, that is, close to that of Pt/C, whereas metal-free catalysts fall behind.[3] However, in alkaline medium, both metal-free and non-precious metal catalysts show good catalytic activity. In 2009, Dai[4] reported that nitrogen-doped carbon nanotube arrays as metal-free electrocatalysts had better catalytic activity than that of the commercial Pt/C catalysts in alkaline medium. But the nitrogen-doped nanocabon materials are not [a] Dr. Y. Zhu School of Chemistry and Life Science Anshan Normal University Anshan, Liaoning, 114005 (China) [b] Dr. Y. Zhu, Prof. B. Zhang, Prof. D. S. Su Shenyang National Laboratory for Materials Science Institute of Metal Research Chinese Academy of Sciences Shenyang, Liaoning, 110016 (China) E-mail: [email protected] [c] Dr. D.-W. Wang School of Chemical Engineering University of New South Wales UNSW Sydney, NSW 2052 (Australia) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201501141.
ChemSusChem 2015, 8, 4016 – 4021
always metal-free catalysts; it was reported that 200 ppm Fe or Co impurities can significantly increase the ORR activity.[5] Despite the numerous studies about metal-free nitrogen-doped nanocarbon materials for ORR, such as carbon nanotubes, carbon nanospheres, mesoporous carbons, and graphene sheets,[6] some researchers suggested that the active sites could be iron oxides or Fe–N complexes in certain nanocarbon-containing electrocatalysts.[7] Although it is clear that all of the three kinds of active components (i.e., nitrogen, iron oxides, and Fe–N complexes) are active for ORR, the order of activity for ORR in alkaline medium remains elusive. Is there any difference in the catalytic activity and mechanism for ORR in alkaline medium between metal-free nitrogen-doped nanocarbon, iron, and Fe/N/C non-precious metal electrocatalysts? What are these differences? In order to answer these questions, we synthesized metal-free nitrogen-doped carbon black (NCB), Fe3O4/CB, Fe3O4/NCB, and FeN4/CB. Through the systematic study of the ORR catalytic activity of these four catalysts in alkaline solution, we were able to answer these questions.
Results and Discussion Firstly, inductive coupling plasma mass spectrometry (ICP-MS) was used to detect a trace amount of iron in carbon black for excluding the impact of iron. It is found that there is only 6 ppm iron in carbon black after treatment with hydrochloric acid. The content of iron is much less than the as-claimed metal-free electrocatalysts (200 ppm).[5] Ammonia gas was used as a clean nitrogen source to prepare nitrogen-doped carbon black at 950 8C. Based on X-ray photoelectron spectroscopy (XPS) analysis, nitrogen atoms were doped in the catalysts and the nitrogen content on the surface of NCB was 2.04 at. %. As shown in Figure S1 in the Supporting Informa4016
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers tion, the N 1s XPS spectra of NCB was deconvoluted into four components consisting of pyridinic N (398.6 eV), pyrrolic N (400.5 eV), quaternary N (401.2 eV), and oxidized pyridinic N (403.8 eV);[8] pyridinic N was the dominating group in the NCB catalyst. The electrocatalytic performances of CB, NCB, and Pt/C for ORR in 0.1 m KOH were measured by using linear sweep voltammetry (LSV), as shown in Figure 1. We clearly observed two
Table 1. ORR kinetic data recorded in alkaline medium for NCB, Fe–CB, Fe–N–CB, and Pt/C. Catalyst
Eonset [V]
E1/2 [V]
Tafel slope [mV dec¢1]
i0 [A cm¢2]
NCB Fe–CB Fe–N–CB Fe–Ph–CB Pt/C
0.847 0.789 0.849 0.971 1.043
0.765 0.682 0.746 0.918 0.934
62 83 88 36 86
1.5 Õ 10¢9 9.8 Õ 10¢9 6.5 Õ 10¢8 1.0 Õ 10¢9 8.7 Õ 10¢6
lytic performance of CB in alkaline medium (transforming ORR from a peroxide pathway close to a four-electron pathway), but NCB and Pt/C still have a large gap in the ORR catalytic activity, which is especially clear in Eonset difference of about 200 mV. The ORR catalytic activity of CB with iron oxide was also studied in alkaline medium. Fe–CB was synthesized by heattreating a mixture of carbon black and ferrous acetate at 950 8C in an inert atmosphere (Argon gas). The sample only contains iron species and carbon black. It can be seen that Eonset for Fe–CB is 0.789 V (in Figure 2 a). Though the value of Eonset for Fe–CB is shifted 21 mV more positive than that for CB, it is 58 mV lower than that for NCB. Based on the slope of the Koutecky–Levich plots, the number of electrons transferred in ORR catalyzed by Fe-CB (the value of n at 0.50 V)) is about 3.9,
Figure 1. Rotating disk electrode (RDE) voltammograms of CB, NCB, and Pt/ C in O2-saturated 0.1 m KOH. Scan rate: 5 mV s¢1, rotation rate: 900 rpm.
limiting current plateau in the LSV curve of CB, which indicate that ORR catalyzed by CB is not a four-electron process, but a two-electron reaction. However, the polarization curve of NCB has only one limiting current platform. The Koutecky– Levich plots (Figure S2 a, b) are drawn from the ORR curves based on the Koutecky–Levich equation: 1 1 1 ¢ ¼¢ þ I Ik 0:62nFAD2=3 v¢1=6 w1=2 cO2 where I is the current density, Ik is the current density in the absence of any mass-transfer effects, w is the rotation speed of the disk, F = 96 485 C mol¢1, A = 0.19625 cm2, D = 1.9 Õ 10¢5 cm2s¢1, c = 1.2 Õ 10¢6 mol cm¢3, and v = 1.0 Õ 10¢2 cm2s¢1.[9] According to the slope of the Koutecky–Levich plots, the values of n at 0.50 V (the number of electrons transferred) in the ORR catalyzed by CB and NCB were calculated to be 2.3 and 3.7, respectively. This suggests that the ORR catalyzed by CB transforms from a peroxide pathway to a four-electron pathway as a result of nitrogen doping. In addition, the onset (Eonset) and half-wave potential (E1/2) in the curve of NCB are 0.847 and 0.765 V, respectively. Comparing Pt/C (loading of 0.06 mgPt cm¢2) and NCB, the differences in the Eonset and E1/2 are 197 and 169 mV, respectively. From a Tafel ORR plot with NCB (Figure S2 c), the Tafel slope and exchange current density (i0) for the NCB catalysts were obtained (Table 1). The i0 value for NCB (1.5 Õ 10¢9 A cm¢2) is far lower than that for the Pt/C catalyst (8.7 Õ 10¢6 A cm¢2), and the Tafel slope of ORR for NCB (62 mV dec¢1) is close to that of the Pt/C catalyst (86 mV dec¢1). To summarize, the nitrogen doping can improve the ORR cataChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
Figure 2. (a) RDE voltammograms of CB, NCB, Fe–CB, and Fe–N–CB in O2-saturated 0.1 m KOH. Scan rate: 5 mV s¢1, rotation rate: 900 rpm. (b) CV of Fe– CB and Fe–N–CB in N2-saturated 0.1 m KOH. Scan rate: 10 mV s¢1.
4017
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers which is closer to a four-electron process than that by NCB. The addition of iron did not greatly improve the ORR catalytic activity of CB in alkaline medium, which is believed relevant to the type of iron species. The cyclic voltammetry (CV) curve of Fe–CB in Figure 2 b shows a pair of redox peaks at 0.586 and 0.821 V (the peak separation is 235 mV). The irreversible redox peaks are not assigned to the peaks of reduction/oxidation Fe3 + /Fe2 + in iron coordination compounds like [Fe(CN)6]3¢/ [Fe(CN)6]4¢, but to crystalline iron species. XRD was used to explore the phase composition in the catalysts (Figure 3). Fe–CB
Figure 4. (a) Low-magnification TEM image. (b) High-resolution TEM image. (c) High angle annular dark field scanning transmission electron microscopy (HAADF-STEM) maps of Fe–CB.
Figure 3. XRD patterns of Fe–CB and Fe–N–CB (JCPDS-Fe3O4 : 19-0629; JCPDS-FeO: 06-0615).
shows well-developed crystalline structures that can be assigned to Fe3O4. Moreover, the fine structure of Fe3O4 particles was further investigated by electron microscopy. As shown in Figure 4 a, the TEM image of Fe–CB shows that there are some large particles with the sized 30–40 nm in the catalyst. The high resolution transmission electron microscopy (HRTEM) image in Figure 4 b and energy dispersive X-ray spectroscopy (EDS) elemental maps in Figure 4 c further reveal that the particles are mainly Fe3O4. Based on the above analysis, it can be conformed that there was only one iron species, Fe3O4, in Fe– CB. Very recently it was hypothesized that the chemisorption of oxygen molecules is the rate determining step in ORR catalyzed by metal-free and non-precious metal catalysts.[7f, 8a, 10] And, compared with nitrogen, iron ions should reduce the lower activation energy of ORR owing to the better ability of the chemisorption of oxygen molecules on iron ions.[7a, 11] It is generally believed that reducing the activation energy can decrease the overpotential during the electrocatalytic process that would be mainly reflected in Eonset of the polarization curve moving close to the standard electrode potential (for ORR, Eonset shifts positively). However, iron ions in Fe3O4 do not exhibit better ORR catalytic activity in alkaline medium. We deduce that although iron ions in Fe3O4 have improved chemisorption of oxygen molecules than nitrogen, the poor conductivity and spinel structure of Fe3O4 nanoparticles limit the rate of electron transfer from the CB to active iron ions, which thus becomes the rate determining step in ORR. This may result in ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
the higher activation energy of ORR catalyzed by iron ions in Fe3O4 than that by nitrogen in NCB. Therefore, Eonset for Fe–CB is 58 mV lower than that for NCB. Note that the number of electrons transferred in the ORR catalyzed by iron is closer to a four-electron process than that by nitrogen, which may be the result of greater affinity toward chemisorption of oxygen in the former. In order to study the ORR catalytic activity of the catalyst containing nitrogen and iron, Fe–N–CB catalyst was synthesized by heat-treating a mixture of CB and ferrous acetate at 950 8C in ammonia. Through the analysis of XRD in Figure 3 and the CV curve of Fe–N–CB in Figure 2 b, it was concluded that there are two iron species, Fe3O4 and FeO. FeNx coordination were not present as the reversible redox of Fe3 + /Fe2 + was not observed in the CV curve of Fe–N–CB. From the HRTEM image and EDS elemental maps of Fe–N–CB (Figure 5), the black 100 ~ 200 nm particles are Fe3O4 nanoparticles and the distribution of nitrogen is uniform in the Fe–N–CB sample. This implies that Fe–N–CB is nitrogen-doped CB containing Fe3O4 and FeO, which enables the study the ORR catalytic mechanism when nitrogen and iron species simultaneously exist in the catalyst. As shown in Figure 2 a, Eonset of the Fe–N–CB polarization curve (0.849 V) is similar to that of NCB (0.847 V), whereas the limiting current density for Fe–N–CB is larger than that for NCB. Based on the slope of the Koutecky–Levich plots, the number of electrons transferred in the ORR catalyzed by Fe–N– CB (the value of n at 0.50 V) is about 3.9. To determine whether the ORR performance results from the catalysis of nitrogen or iron in Fe–N–CB, we synthesized xFe–N–CB, where x corresponds to 10, 20, 30, 40, 50, and 100 mg of ferrous acetate used during the preparation (see the Experimental Section for more details). Figure 6 shows the LSV polarization curves of the six catalysts in O2-saturated 0.1 m KOH. Eonset does not
4018
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 5. (a) Low-magnification TEM image. (b) High-resolution TEM image. (c) HAADF-STEM maps of Fe–N–CB.
Figure 6. RDE voltammograms of Fe–N–CB with different iron contents in O2-saturated 0.1 m KOH. Scan rate: 5 mV ·s¢1, rotation rate: 900 rpm.
change with varied iron content and the limiting current densities are close except in the case of 100Fe–N–CB, which has higher current densities. In addition, the values of n at 0.50 V (the number of electrons transferred) in the ORR catalyzed by the six catalysts are all 3.9. This indicates that the iron in Fe– N–CB has little effect on the catalytic activity. We used these results to predict the possible ORR mechanism catalyzed by Fe–N–CB. When the overpotential is low at the beginning of ORR, nitrogen atoms in Fe–N–CB are the active sites for ORR as the activation energy of ORR catalyzed by iron ions in Fe3O4 is higher than that by nitrogen in NCB, as mentioned above (Scheme 1 a). This results in a similar Eonset for Fe–N–CB and NCB that does not change with the varied iron contents. However, when the potential scans negatively to 0.789 V, Eonset for Fe3O4/CB, the higher overpotential triggers the ORR on iron ions in Fe3O4. As a result, both iron ions and nitrogen atoms are active sites for ORR (Scheme 1 b, c). In addiChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
Scheme 1. (a) Schematic diagram for the ORR mechanism of Fe–N–CB at 0.847 V; (b) Schematic diagram for the ORR mechanism of Fe–N–CB at 0.789 V; (c) the LSV curve of Fe–N–CB (different regions with different active sites).
tion, it is common that the value of limiting current density is mainly related to the amount of active sites. Accordingly, the limiting current density for Fe–N–CB is larger than that for NCB because iron and nitrogen are both active sites for ORR when the current reaches the limiting current. Note that although iron ions and nitrogen atoms in Fe–N–CB catalyzes ORR together from 0.789 V, nitrogen is the man active site owing to the lower activation energy than that of iron ions. This is the reason that the limiting current densities of Fe–N–CB have no obvious change with varied addition of ferrous acetate from 10 to 50 mg. The limiting current density of 100Fe–N–CB significantly increases because there are many more iron ions as
4019
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers active sites for ORR compared with the other Fe–N–CB catalysts. Iron ions have improved ability toward chemisorption of oxygen molecules than nitrogen atoms, and ORR catalyzed by the former is closer to a four-electron process. As iron ions in Fe3O4 do not show the good ORR catalytic activity owing to the poor conductivity and spinel structure of Fe3O4, we believe that iron ions in iron species that have good conductivity and graphite-like structure serve as the ORR active sites with perfect catalytic performance. FeNx coordination compounds, are suitable iron species that fulfill these two requirements. Therefore, an iron phthalocyanine-based CB catalyst (Fe–Ph–CB) containing FeN4 was synthesized. The CV curve of Fe–Ph–CB in Figure 7 b shows a pair of reversible redox peaks at 0.785 and 0.860 V (the peak separation is 78 mV). We assign this to the reduction/oxidation (Fe3 + /Fe2 + ) of FeN4 in the catalysts. As shown in Figure 7 a, both Eonset (0.971 V) and the limiting current density for Fe–Ph–CB catalyst are the best among these Fe/N/C catalysts mentioned above. Also, based on the slope of the Koutecky–Levich plots, the number of electrons transferred in the ORR catalyzed by Fe–Ph–CB (the value of n at 0.50 V) is about 3.9. From the Tafel ORR plot, the Tafel slope and i0 for the Fe–Ph–CB are 36 mV dec¢1 and 1.0 Õ 10¢9 A cm¢2, respectively (Table 1), which are also better than the other catalysts. Unfortunately, the performance of the Fe–Ph–CB catalyst still falls short of Pt/C.
Conclusions We confirm that both nitrogen atoms and iron ions are the oxygen reduction reaction (ORR) active sites in alkaline medium. In metal-free nitrogen-doped carbon black (NCB), nitrogen doping can improve the ORR catalytic activity by transforming the reaction from a peroxide pathway to a four-electron pathway. However, compared with Pt/C and NCB, the difference in the onset potential (Eonset) of the polarization curve is ~ 200 mV. In addition to the number of electrons transferred, closer to a four-electron process, iron ions in Fe3O4 do not show improved ORR catalytic activity over NCB owing to the poor conductivity and spinel structure of Fe3O4. During the ORR process catalyzed by Fe3O4/NCB in which iron oxide and nitrogen atoms are present at the same time, the nitrogen atoms are the active sites for ORR when the over potential is low. The iron ions and nitrogen atoms can both serve as active sites to catalyze ORR when the potential shifts negatively, but the nitrogen atoms still accounts for the main part. The active sites in Fe/N/C, having ORR performance closest to that of Pt/C in alkaline medium, are the iron ions in the iron species, which have good electrical conductivity and a graphite-like structure. The activity of the catalysts is in the order FeN4 > N > FeOx in alkaline medium. We believe that the knowledge of the active site will lead the way to target-specific synthesis of highly active Fe/N/C catalysts for ORR in alkaline medium.
Experimental Section Catalyst synthesis Carbon black, typically a commercial carbon Ketjenblack EC 300 J (with Brunauer–Emmett–Teller surface area of about 801 m2 g¢1), was treated in 6.0 m HCl solution for 72 h to remove metal impurities. The acid-treated CB was heat treated at 950 8C in ammonia for 2 h, and nitrogen-doped CB, designated as NCB, was obtained. Ferrous acetate (0.05 g) was first dispersed with acid-treated CB (0.6 g) in 200 mL ultrapure water. After continuously mixing for 24 h, the suspension containing carbon and transition metal were vacuumdried using a rotary evaporator. By the subsequent heat-treatment at 950 8C for 2 h in an inert atmosphere (Argon gas) or in ammonia respectively, Fe–CB and Fe–N–CB were synthesized. Using the same synthetic procedure of Fe–N–CB in which the varied qualities of ferrous acetate (0.01, 0.02, 0.03, 0.04, 0.05, and 0.10 g) were added, xFe–N–CB where x = 10, 20, 30, 40, 50, and 100 were obtained, respectively. In addition, iron phthalocyanine (0.16 g) was dispersed with acid-treated CB (0.6 g) in 200 mL ultrapure water. After continuously mixing for 24 h, the suspension containing carbon and transition metal were vacuum-dried using a rotary evaporator. Then we obtained the catalyst Fe–Ph–CB.
Characterization
Figure 7. (a) RDE voltammograms of NCB, Fe–CB, Fe–N–CB, Fe–Ph–CB, and Pt/C in O2-saturated 0.1 m KOH. Scan rate: 5 mV·s¢1, rotation rate: 900 rpm. (b) CV of Fe–Ph–CB in N2-saturated 0.1 m KOH. Scan rate: 10 mV·s¢1.
ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
XRD analysis was carried out using a D/MAX-2500 PC X-ray diffractometer with monochromatized CuKa radiation (l = 1.54 æ). The patterns were obtained at a scan rate of 58 min¢1 with a step of 0.028. The XPS measurements were carried out in an ultra-high vacuum ESCALAB 250 set-up equipped with a monochromatic AlKa X-ray source (1486.6 eV; anode operating at 15 kV and 20 mA). A FEI Tecnai G2 F20 microscope equipped with high angle annular
4020
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers dark field (HAADF) and energy dispersive X-ray analysis (EDAX) detectors was employed to acquire the HRTEM and HAADF-STEM images and EDS under 200 keV.
[2]
Electrochemistry measurements [3] [4] [5]
ORR activity and four-electron selectivity of catalysts were measured by a rotating disk electrode (RDE). RDE measurements were performed by Epsilon Electrochemical Station (PAR2273) in a conventional three-electrode electrochemical cell. To avoid any potential contamination of a non-precious metal catalyst by platinum, experiments were carried out using a graphite rod as the counter electrode. An Ag/AgCl electrode in 3.0 m KCl (0.220 V vs. natural hydrogen electrode, NHE) was used as a reference electrode. All potentials were later converted to the reversible hydrogen electrode (RHE) scale. Non-precious metal catalyst data were obtained at RT (ca. 25 8C) with a catalyst loading of 0.9 mg cm¢2 in 0.1 m NaOH at a rotating disk speed of 900 rpm. The data of Pt catalyst were recorded with a 20 wt % Pt/C at loadings of 60 mgPt cm¢2. The catalysts were ultrasonically dispersed in an isopropanol solution containing suspended 5 wt % Nafion solution for 30 min to form a catalyst ink, which was then applied to the glassy-carbon disk surface. The disk rotation rate was from 400 to 2500 rpm. CV characterization of the catalysts in the absence of oxygen was typically carried out in the potential range from 1.0 to 0.0 V at a scan rate of 10 mV s¢1 in nitrogen-saturated 0.1 m KOH at RT.
[6]
[7]
[8]
Acknowledgements [9]
The authors acknowledge the financial support from MOST (2011CBA00504), NSFC of China (21133010, 21203215, 51221264, 21261160487), and ‘Strategic Priority Research Program’ of the Chinese Academy of Sciences, Grant No. XDA09030103. Keywords: carbon · electrocatalyst · iron · nitrogen · oxygen reduction reaction [1] a) D. S. Su, G. Sun, Angew. Chem. Int. Ed. 2011, 50, 11570 – 11572; Angew. Chem. 2011, 123, 11774 – 11777; b) D.-W. Wang, D. Su, Energy Environ. Sci. 2014, 7, 576 – 591; c) G. Wu, K. L. More, C. M. Johnston, P. Zelenay, Science 2011, 332, 443 – 447; d) Y. S. Zhu, B. S. Zhang, X. Liu, D. W. Wang,
ChemSusChem 2015, 8, 4016 – 4021
www.chemsuschem.org
[10] [11]
D. S. Su, Angew. Chem. Int. Ed. 2014, 53, 10673 – 10677; Angew. Chem. 2014, 126, 10849 – 10853. a) F. Jaouen, E. Proietti, M. LefÀvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T. Chung, C. M. Johnston, P. Zelenay, Energy Environ. Sci. 2011, 4, 114 – 130; b) G. Wu, M. Nelson, S. Ma, H. Meng, G. Cui, P. K. Shen, Carbon 2011, 49, 3972 – 3982; c) G. Wu, P. Zelenay, Acc. Chem. Res. 2013, 46, 1878 – 1889. M. LefÀvre, E. Proietti, F. Jaouen, J.-P. Dodelet, Science 2009, 324, 71 – 74. K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 2009, 323, 760 – 764. F. Jaouen, A. M. Serventi, M. LefÀvre, J.-P. Dodelet, P. Bertrand, J. Phys. Chem. C 2007, 111, 5971 – 5976. a) K. Ai, Y. Liu, C. Ruan, L. Lu, G. M. Lu, Adv. Mater. 2013, 25, 998 – 1003; b) S. Chen, J. Bi, Y. Zhao, L. Yang, C. Zhang, Y. Ma, Q. Wu, X. Wang, Z. Hu, Adv. Mater. 2012, 24, 5593 – 5597; c) L. Qu, Y. Liu, J.-B. Baek, L. Dai, ACS Nano 2010, 4, 1321 – 1326; d) T. Sharifi, G. Hu, X. Jia, T. Wgberg, ACS Nano 2012, 6, 8904 – 8912; e) Z.-H. Sheng, L. Shao, J.-J. Chen, W.-J. Bao, F.-B. Wang, X.-H. Xia, Acs Nano 2011, 5, 4350 – 4358; f) Y. Tang, B. L. Allen, D. R. Kauffman, A. Star, J. Am. Chem. Soc. 2009, 131, 13200 – 13201; g) S. Yang, X. Feng, X. Wang, K. Mìllen, Angew. Chem. Int. Ed. 2011, 50, 5339 – 5343; Angew. Chem. 2011, 123, 5451 – 5455; h) Y. Zhu, Y. Lin, B. Zhang, J. Rong, B. Zong, D. S. Su, ChemCatChem 2015, 7, 2840 – 2845. a) F. Jaouen, S. Marcotte, J.-P. Dodelet, G. Lindbergh, J. Phys. Chem. B 2003, 107, 1376 – 1386; b) M. LefÀvre, J.-P. Dodelet, Electrochim. Acta 2003, 48, 2749 – 2760; c) J. Maruyama, J. Okamura, K. Miyazaki, Y. Uchimoto, I. Abe, J. Phys. Chem. C 2008, 112, 2784 – 2790; d) C. M¦dard, M. Lefevre, J. Dodelet, F. Jaouen, G. Lindbergh, Electrochim. Acta 2006, 51, 3202 – 3213; e) H. Wang, R. Cote, G. Faubert, D. Guay, J. Dodelet, J. Phys. Chem. B 1999, 103, 2042 – 2049; f) L. Zhang, K. Lee, C. W. Bezerra, J. Zhang, J. Zhang, Electrochim. Acta 2009, 54, 6631 – 6636. a) P. H. Matter, L. Zhang, U. S. Ozkan, J. Catal. 2006, 239, 83 – 96; b) V. Perazzolo, C. Durante, R. Pilot, A. Paduano, J. Zheng, G. A. Rizzi, A. Martucci, G. Granozzi, A. Gennaro, Carbon 2015, 95, 949 – 963. a) S. Baranton, C. Coutanceau, E. Garnier, J.-M. L¦ger, J. Electroanal. Chem. 2006, 590, 100 – 110; b) K.-L. Hsueh, D.-T. Chin, S. Srinivasan, J. Electroanal. Chem. 1983, 153, 79 – 95; c) R. N. Itoe, G. Wesson, E. E. Kalu, J. Electrochem. Soc. 2000, 147, 2445 – 2450; d) D. R. Lawson, L. D. Whiteley, C. R. Martin, M. N. Szentirmay, J. I. Song, J. Electrochem. Soc. 1988, 135, 2247 – 2253. X. Hu, Y. Wu, H. Li, Z. Zhang, J. Phys. Chem. C 2010, 114, 9603 – 9607. a) G. Lalande, G. Tamizhmani, R. Cote, L. Dignard-Bailey, M. Trudeau, R. Schulz, D. Guay, J. Dodelet, J. Electrochem. Soc. 1995, 142, 1162 – 1168; b) K. Legg, F. Jona, D. Jepsen, P. Marcus, J. Phys. C 1975, 8, L492; c) S. Sun, L. Chen, H. Shi, Y. Li, X. He, J. Electroanal. Chem. 2014, 734, 18 – 24.
Received: August 25, 2015 Revised: October 18, 2015 Published online on November 17, 2015
4021
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
